I E) R.A R.Y OF THE U N I VEFIS ITY or ILLINOIS 560.97 C79 GEOLOGY ^jfJAN 29 1947 ON THE SUPPOSED CARNIVORA 0^^ THE EOCENE OF THE ROCKY MOUNTAINS. BY E. D. COPE. Extracted from the Proceedings of the Academy of Natural Sciences, Nov. 30, 1875.] PHILADELPHIA: COLLINS, PRINTER, 705 JAYN K STREET. 1875. 5-60.97 On the supposed Carnivora of the Eocene of the Rocky Moun- tains. — Animals which fulfilled the functions of the existing Car- nivora were abundant in North America during the Eocene period. The Wahsatch beds of New Mexico have yielded remains of more than a dozen species, which ranged from the size of a weasel to that of a jaguar. Investigation into the structure of these shows ?that while they differ in minor points among themselves, they agree in possessing characters which distinguish them from the true Carnivora. I have already pointed out,' that, in the genera ^ Ambloctonus, Oxysena, Stypolophus^ and Didymictis, the tibio- ^ tarsal articulation differs from that of the existing Carnivora^ and «^ suggested that these forms might prove to be gigantic Insec- ^ tivora. Further investigation has satisfied me that they cannot * be included in the order Carnivora^ and their systematic position Sf. proves to be of considerable interest. y, A greater or less part of the cranial chamber is preserved in \i specimens of Oxysena forcipata and Stypolophus Mans. In these animals it has a long, narrow form like that of the opossum, and in the first named, where the interior form can be seen, it is evi- dent that the cerebral hemispheres were small and narrow, and that the olfactory lobes were relatively large, and were entirely un- covered, projecting beyond the hemispheres. In Ambloctonus^ Didymictis, and three undetermined forms, the femur supports a third trochanter. In all the genera the ilium has a well-marked external anterior ridge, which continues from the acetabulum to the crest, distinct from the internal anterior ridge. The ilium has, therefore, an angulate or convex external face, as in Insectivora and Mar supialia, and does not display the usual expansion in a single plane of most of the placentals. In all the genera there is a strong tuberosity in the position of the anterior inferior spine, which is wanting in the Mammalia, except- ing certain Insectivora and Prosimiae,^ although it marks the posi- tion of the origin of the rectus femoris muscle in all types. The glenoid cavity of the squamosal bone is transverse, and well defined anteriorly and posteriori}', as in the Carnivora. Of the first series of carpal bones of the four genera named, I have been able to learn nothing, but in the genus Synoplotherium from the Bridger Eocene of Wyoming, which probably belongs to this group, the scaphoid and lunar bones are separate and not united as in the Carnivora. The above characters point to the Marsupialia or the Insec- ^ tivora as. the proper location for the flesh-eaters under considera- r^tion ; and the evidence is much more weighty in favor of the lat- ^ ter order as their true position. For in the genera Oxyaena and Didymictis the posterior part of the inferior border of the mandi- C/v bular ramus is not inflected as in Marsupialia., nor are the ante- ^ rior inferior iliac tuberosity and third trochanter seen in that or- der, while both exist in the Insectivora. ^ Cuvier describes-^ the tibia of Carnivora as follows: "Quant a ' Systematic Catalogue of the Vertebrata of the Eocene of New Mexico, — 1875, p. 7. ^ 2 See the figure of Solenodon by Peters, and Ohiromys by Owen. r\j ' Ossemens Fossiles, vii. p. 113. 2 la tete inf^rieure, tous les carnassiers se distingu^nt de Phomme par sa figure plus etroite du cote externe que le I'interne, et par sa division en deux fosses oblique, au moyen d'une arete arrondie qui repond a la partie de I'astragale. . . Le phoque I'a cepen- dant d'une forme tr^s-particuliere par I'excessif aplatissement de sa moiti^ superieure, et par sa facette particulaire infdrieure, qui est en concavitd simple et pen profounde." The astragalar articular face of tlie tibia in the genera above named is not divided into the two oblique fossae by " a rounded crest which is applied to the groove of the superior pully-shaped face of the astragalus." It is uninterrupted and more or less ob- lique in the transverse direction ; alwa3^s so at the posterior bor- der. The inner malleolar process is produced downwards, and rests in a concavity on the inner side of the neck of the astragalus. The astragalus, which I have seen in several of the species, pre- sents a corresponding trochlear face. That is, instead of a groove, it presents an open angle upwards, which separates the superior from the oblique internal face. The superior plane is flat, but is interrupted on the posterior side by a groove. This groove is the posterior extremity of that which divides the superior face of the astragalus in the higher Mammalia^ but here it contracts to a point and disappears next the fibular face just as it reaches the superior surface. The fibular face is vertical, and shares on its posterior part a large ligamentous fossa with the opposed part of the fibula. The distal end of the fibula is remarkably stout. This structure finds its counterpart in the internal half of the astragalus of the opossum. The arrangement permits a rotary movement of the astragalus and thus of the whole foot, on the tibia, the fibula with its fixed articulation with the astragalus, rotating on the tibia, as in thepedimanous Marsupialia. The flat- ness of the inner malleolus in some of the species indicates that the capacity for rotation was less in them than in others. This arrangement exactly reverses the extensive oblique fibulo-astraga- lar articulation seen in the opossum, the Petaurista^ Dai^yurus^ etc. Prof Owen, in describing the astralagus of the wombat {Phasco- lomys)^ says : " The upper articular surface for the tibia is as usual concavo-convex, the internal surface for the inner malleolus flat- tened, and at right angles with the preceding, but the outer articu- lar surface presents a triangular flattened form, and instead of being bent down parallel with the inner articular surface slopes away at a very open angle from the upper surface, receiving the articular surface of the fibula so as to sustain its vertical pressure. * * * This form of astragalus is also characteristic of the Koala, Petaurists, Dasyures, and the Pedimanous Marsupialia." In one species where the cuboid bones are preserved, it is evi- dent that the distal end of the astragalus articulated with this as well as with the navicular bone, although the facet of the astraga- lus is single and continuous. As the extensive transverse distal astragalar face is characteristic of all the species where it is pre- served, the contact of the cuboid and astragalus is probably com- mon to all of this division. The dentition of this group is consistent with its reference to 3 the sarcophagous Marsupialia or to the Insectimra. It has, how- ever, decided resemblances in the form pf the molars, and in the deficiency in the number of the inferior incisors, to sucli genera of Inaectivora as Mythomys and Solenodon, while in the large canines, it more nearly approaches Sarcophaga and Carnivora. I propose to include the genera Ambloctonus, Oxysena^ Stypolo- phiis^ and Didymictis in a special division under the name of Greodonta, This division may be regarded as a suborder of the Insectivora. It is possible that the genus Diacodon Cope belongs here also ; its species resemble Chiroptera in the inferior dentition, and are of small size. The genus Mesonyx^^ which I discovered in the Bridger beds of Wyoming, cannot be referred to the Greo- donta as here constituted, since the trochlear face of the astragalus is completely grooved above as in the true Garnivora^ and its distal end presents two distinct facets, one for the cuboid and the other for the navicular bones. It represents on this account a peculiar family, the Mesonychidse. To the Greodonta must be referred the genera Pterodon and Palaeonictis of De Blainville, from the French Eocene. This author and Pomel placed them in the Marsupialia^ but Professor Gervais remarks (Greologie et Paleontologie Fran9aise) that the evidence is insufficient for such a course. Here also doubtless belong supposed Garnivora from the Wyoming Eocene, stated by Marsh to be allied to the Viverridse. The remarkable type first introduced to the notice of paleonto- logists by Leidy, represented by the genera Anchippodus, Ecto- ganus, etc., has been looked upon as an order of Mammalia by Marsh, and termed the Tillodontia. He gives,** as its characters, the possession of claws, plantigrade feet with five toes, a third trochanter of the femur, and separate scaphoid and lunar bones. Also, that the dentition is characterized by " molars of the ungu- late type," small canines, and large scalpriform incisors in both jaws, faced with enamel and growing from persistent pulps as in the Rodentia. He says this order " seems to combine characters of the orders of carnivores, ungulates, and rodents." Except in the dentition, the definition above given applies to the Greodonta ; and an analysis of the dentition shows so many points of resemblance as to render it probable that they pertain to the same order of Mammalia. Also, except in the dentition, the characters given by Prof. Marsh do not differ from those of the Insectivora. The structure of the superior molars is not in- consistent with the same order, and the small canines and large incisors are even more like those of most Insectivora than are Greodonta. The singular form of these incisors, and their growth from persistent pulps, is rather characteristic of Rodentia. The transverse or tubercular premolars also distinguish this group from both the Greodonta and the true Insectivora. Tiie defini- tions of the order and sub-orders will then be as follows : — > Ann. Rept. U. S. Geol. Surv. Terrs., 1873, p. 550. 2 Amer. Journ. Sci. Arts, 1875, 331. 4 Insectivora. — Mammalia with small cerebral hemispheres which do not cover the olfactory lobes, nor the cerebellum ; with numer- ous clawed digits, and a third trochanter of the femur; with a transverse glenoid cavity for the mandible. Superior incisors normal, not growing from persistent pulps; canines large ; premolars compressed. Astragalus not grooved above, articulating with the cuboid as well as the navicular ; five toes on the hind foot ; Creodonta, Superior incisors large, growing from persistent pulps, and without enamel on the sides ; superior canines small when present ; premolars wide or tubercular ; Tillodonta. These suborders of the order Insectivora do not differ among themselves more than do those of the Mai^supialia^ and constitute a series of parallels with them. Thus the Creodonta resemble the Sarcophaga, the Insectivora vera the Entomophaga, and the TillO' dontathQ Rhizophaga^ tj^pified by Phascolomys. Tiie genera of the Creodonta differ as follows : — I. First and third inferior true molars without internal cusp. Last superior molar longitudinal ; last inferior molar carnassial ; Amhloctonus. II. Inferior carnassials with interior tubercle; no tubercular molar ; last superior molar transverse. Three tubercular carnassials;^ Stypolophus, Two tubercular carnassials ; Oxyaena, III. Inferior carnassial with interior tubercle ; a tubercular molar. One tubercular carnassial ; Didymictis. The number of toes on the hind foot cannot be certainly stated in all the genera, but in Stypolophus Mans and another species there were probably five, the inner being of reduced size. There is present in those species an en to-cuneiform bone which resem- bles that of Canis ; it is compressed, with one truncate concave terminal facet, and an internal oblique one at the opposite and proximal extremity. The form of the truncate articular face of the distal end indicates the existence of an inner metatarsal bone of moderate proportions, which probably supported a small hallux. This thumb could not be opposable as in the opossum. In general appearance the Creodonta differed from the Carnivora, in many of the species at least, in the small relative size of the limbs as compared with that of the head, and in some instances, as compared with the size of the hind feet. The feet were probably plantigrade, and the posterior ones capable of some degree of rotation. The probable large size of the rectus femoris muscle indicates unusual power of extension of the hind limb. They were furnished with a long and large tail. Probably some of the species resembled in proportions the 3Iythomys and Solenodon^ now exist- ing in Africa and the West Indies, but they mostly attained a much larger size. Published December 22, 1815. * For the meaning of these expressions, see Syst. Cat. Eoc, New Mexico, 1875, p. 6. ON THE PRIMITIVE TYPES OF THE Orders of Mammalia Educabili BY ON THE PRIMITIVE TYPES OF THE ORDERS OF MAMMALIA EDUCAIULIA. By PiiOF. E. D. Coi'E. {Head before the American Philosoijhical Society, April 18, 1873.) In the Proceedings of the American Philosophical Society, 1872, p. 554, tlie writer described a species of Quadrumanous Mammal untler the name of AnaptoiiiorphuH wmulus, comparing its dental and other characters with JSimia and Ilomo!^ In the American Journal of Science and Arts for November,* 1872, Prof. O. C. Marsh announced that he be- lieved that three genera previously described by him, viz., Thinolestes, Limnotheriuui and TehriatoleHtefi,\ were referable to the Quadrumana, saying that they "have the principal parts of the skeleton much as in some of the lemurs." Prior to either of these determinations, the author de- scribed a new genus and species as allied to Notharctus, Leidy, under the name of Toniitherium,X hut made no suggestion as to its ordinal posi- tion. On a re-examination of the last-named genus, I am satisfied that it also should be referred to the Quadrumana, and describe it as follows : TOMITHERIUM. Cope. Dental formula in an uninterrupted series. Last molars with five tubercles, others with four; all low and slightly alternating, the outer wearing into crescents. Canines quite small. Incisors very promi- nent, the median pair with transverse cutting edges. Symphysis coossi- fied, projecting in front. In the molars, the adjacent herns of the two outer crescents unite with the anterior outer tubercle ; the posterior outer is insignificant. There is a projection but no tubercle in front of the outer anterior tubercle. The premolars present but a single com- pressed conic crown ; the posterior, however, widened behind, and with a low tubercle. The first and second premolars are one-rooted (not en- tirely a generic character). I base the distinction between this genus and Notliarctus on the small canine, and the sub-horizontal position of the incisors ; believing that when other portions of the skeleton are studied, other differences will appear. The portions of the skeleton of the type species preserved are : the en- tire dentition of the lower jaw minus the crowns of the outer incisor, canine and first premolar ; the left ramus nearly complete, the extreme angle being wanting : the right humerus complete, with right ulna and radius, the latter lacking the distal extremity ; a large part of the left ilium ; the right femur nearly entire ; part of the left humerus, meta- /.^yir ^ Published Ociob.er 8th, 1872. t Published August 7th, 1872. X ftaftd a t tliu uica t iug OftptdjBibM CQ fc ht lOTOi *— ^^^^ /'^ rsy^ ■ - The mandibular rami are quite stout, but not very deep ; the sym- physeal portion long and oblique, and the coronoid and condylar portions elevated, with axis at right angles to that of the horizonal portion. The condyle is well elevated, and the coronoid process smallj; the dental foramen is half way between the margins of the ascending ramus, and opposite the bases of the crowns of the molars. The inferior margin of the jaw shows no tendency to inflection at a point immediately below this foramen, where it is broken off. The mental foramen is divided, the exits being at points opposite those between the premolars 1-2 and 2-3. The humerus has a round head directed backwards and a little out- wards. The tuberosities are rather small, of about equal size, and obtuse ; they enclose a short bicipital groove. The bicipital crests are very largely developed, and extend to the middle of the shaft enclosing an open groove between them. The external is narrow and most elevated, the internal more obtuse and directed inwards. The shaft is thus sub-triangular in section. The distal extremity is nearly at right angles to the axis of the proximal and is much expanded transversely. A large part of this expansion is caused by the truncate internal tuberos- ity, and by the less prominent external one. The latter is continued in a thin ala which only sinks into the shaft at» its middle. The condyles are small, the external the most prominent. There is a shallow olecranar fossa, and no coronoid, and hence no supercondylar foramen. There is an arterial foramen above the internal tuberosity. The ulna is compressed, and contracts rapidly to the extremity. The olecranon is broad and obtuse and the humeral cotylus oblique to the long axis. Tlje coronoid process is low. The shaft is remarkably curved from right to left (inwards). The radius has a discoidal head with central depression, and it was evidently capable of complete rota- tion. It exhibits a tuberosity and slight flexure below the head. The distal extremity has a horizontal triangular section with the apex inter- nal and truncate ; the shaft near it is quite flat. The left ilium is obspatulate and flat, widest at the convex crest, and slightly concave on the outer side. It is rather thin, and the impression for the sacral diapophyses is elongate. The inferior border thickens gradually to the acetabulum ; the superior is excised so as to form an open concavity. The right femur is remarkable for its length. Its shaft is flattened from before backwards, and without flexure. The great trochanter is large, and embraces a deep in-looking fossa. There is a flat tuberosity looking outwards just below, and the little tiK)chanter is a little below opposite to it. The condyles are sub-similar in size, the trochlear sur- face wide, but not flat, and the inner border thickened and considerably elevated. The femur is 1.75 times as long as the humerus ; it was scarcely longer, though a small piece is wanting from the shaft of our specimen. Bemarks. Having described the more important parts of the skeleton 3 preserved, 1 now proceed to consider its systematic position, and the order to wliic-li it should be referred. The first impression derived from the appearance of the lower jaw and dentition, and from the humerus, is that of an ally of the coati {Nasua). The humerus indeed is almost a fac-simile of that of Nasna, the only difference beinj^ a slight outward direction of the axis of the head. • The same bone resembles also that of many marsupials, but the flat ilium, elevated position of dental foramen, and absence of much inflection of the angle of the lower jaw, etc., render afidnity with that group highly improbable. The lengtli of the femur indicates that the knee was entirely free from the body as in the quadrumana, constituting a marked distinction from anything known in the Carnivora^ including Nasua. The round head of the radius indicates a complete power of supination of the fore foot, and is different in form from that of Carnwora, including Nasua ; and finally, the distal end of the radius is still more different from that of Nasua, and resembles closely that of Semnopithecus. We have, then, an animal with a long thigh free from the body, a foot capable of complete pronation and supination, and a form of lower jaw and teeth quite similar to that of the lower monkeys. The form of the humerus and its relative length to the femur are quite as marked as in some of the lemurs. The most marked difference is seen in the increased num- ber of teeth ; but in this point it relates itself to the other Quadrumana^ as the most ancient types of Carnivora and Ungulates do to the more modern ; e.g., Hycenodon to the former and Paloiosyops to the latter. In its special dental characters it shows a close resemblance to small types of the Eocene, which have been regarded as low Perissodactyles, as Hyo. psodus, etc. TOMITHERIUM ROSTRATUM. CopC. Proceedings of the American Philosophical Society, 1872, p. 470. This species was about the size of the prehensile tailed-monkey, so frequently seen in shows. The first and second premolar have but one root, the base of the second being about the size of the base of the canine. The latter are cylindric at base. The incisors form a parabolic outline, and have entire edges, the middle pair transverse ones. Enamel gener- ally smooth, premolars somewhat striate ; an indistinct inner cingulum. M. Length of entire dental series (straight) 0,044 symphysis mandibuli 020 Depth ramus at seqond molar 010 Length crown of " " 006 Width " " 0045 *' between two " " 014 " " canines 005 of ascending ramus above dent, foramen 016 Length of humerus 083 Diameter of head 013 4 M. Diameter of shaft at middle 0 0085 " " distal end, transverse 023 " " " antero-posterior 0078 Depth of olecranon 009 " " ulna at coronoid ; . . . .010 Diameter extremity of radius, proxhnally 009 " " " distally 010 Length of ilium from acetabulum 042 Width near crest 017 Length of femur preserved... 137 Width just below neck 017 " at middle Oil " extremity 019 " of trochlea 009 Longest chord of condyles and trochlea 019 The remains of this species were found together by the writer in the Bridger beds on Black's Fork, Wyoming. ANA.PTOMORPHUS. Cope. Proceedings of the American Philosophical Society, 1872, p. 554. This genus is represented by the left ramus mandibuli of a single species. The posterior portion is broken away, and the teeth remaining perfect are the P. M. 2, and M. 1 and 2. The ramus, though small, is stout, and deeper at the symphysis than at the last molar. What appears to be the dental foramen is nearly opposite the bases of the crowns of the molars. The mental foramen issues beneath the first premolar. Dentition of the ramus mandibuli, In. 2, C. 1, P.M. 2, M. 3, total, 16. It differs in many respects from these ; there is no interruption in the series near the canine, and the symphysis, though massive, is not coossified. Further details are, the last molar is three-lobed and elongated behind. The composition of the crowns of the preceding molars consists of four opposed lobes, which are very stout, and con- nected transversely by a thin ridge behind, or in close contact in front. The premolar tooth which is best preserved, is a perfect second, which, while having two roots, possesses a crgwn which stands almost entirely on the anterior, presenting a curved sectorial crest forwards and up- wards. The dentition is more decidedly quadrumanous in this genus than in the last, and it might be referred decidedly to Lemuridm were it not for the unossified symphysis. It no doubt represents a distinct group or family from Tomitherium, and one more nearly related to the existing types of Madagascar and South Africa. Anaptomorphus ^mulus. Cope, loc. cit. This species was about as large as a maraioset or a red squirrel. The enamel of the teeth is entirely smooth. MemuremenU. M. Length dental line 0.0148 " of last molar 0080 " ante-penult 0025 Width 0020 Length of three molars preserved 0070 From the Bridger Beds of the upper valley of Green River. On the Phylo(jeny of the Mammalian Orders. So much light is* thrown on this subject by the researches into the structure of the fossil mam- malia of the Eocene formation, that it seems opportune to call attention to the subject. I deem it demonstrated to a certainty, that the case with the mammals of this formation is the same as with the reptiles of the Trias, i.e., that the family types are all more generalized, and the orders not nearly so widely distinguished as in later periods of the world's history. The succession of forms which has terminated in the horse, has been clerfi'ly pointed out by Prof. Huxley, as well as the line which has given the world the beautiful order of the Artiodactyla ; but "the approximate lineal predecessors of the Prohoscidia, of the Ungulate animals as a whole, of the Quadnimana (including man), and of the Carnivora, have not been clearly pointed out. The genus Eohanleua has been shown* to be a Proboscidian which combines some important features of the Perissodactyla with those of its own order, thus standing in antecedent relation to the elephants, etc., of the present day. The number of such characters was shown to be some- what increased in Bathmodon, which therefore stands still nearer to the common point of departure of the two orders. This point is to be found in types nearer the clawed orders (Unguiculata), in the number of their digits (4-5), and in which the transverse and longitudinal crests of the molar teeth are broken up into tubercles more or less connected, either type of dentition being derived according as such tubercles are expanded transversely or longitudinally. We have several genera which answer this description so far as the teeth are concerned, but unfortunately the digits are unknown ; such are OUgotomus, Orotherium,, etc. The type of Tomitherium already described, evidently stands between Lemtirine monkeys and such small allies of Palcpotheriida^, with conic- tubercular teeth, and which abound in the Eocenes of Wyoming and France. The dentition of the two types is indeed but little different in, the Quadrumanous and Ungulate types respectively, being a continuous series of I. 2 or 3 ; C. 1, P.M. 3-4 ; M. 3 ; the canines but moderately developed. A comparison with Nama reveals no distant affinity. As above re- marked, the fore-limb presented a great similarity in this genus and Tomitherium. The teeth, though less numerous, in the molar series have * On the Short-footed Ungulata of Wyoming, page 3. 6 the cutting type anterior and tubercular posterior, in both genera. Noiharctus, Leidy, resembles Nasua still more than does Tomitheriutu,, and occurs in the same Eocene strata. Prof. Leidy originally regarded it as a Carnivore, and subsequently (Hayden's Survey Montana, 1871) placed it among Ungulates. He w^as probably nearly correct on both occasions, and that only a technical line will ultimately decide whether it be not a monkey.* But the genus which associates more definitely the orders Carnivora and Qimdrumana, is the Cercolejjfes, which F. Cuvierf placed between the two. Its two cutting premolars and thi ee true molars, with the co- ossitied rami of the mandible are truly Quadrumanous features, although it should on other grounds be regarded as a plantigrade carnivore. Sev- eral of the extinct genera of the Wyoming Eocene will prove to be allied to this form. Cercoleptcs does not, however, present us with the ultimate original type of the Garnivo'ra. Such a type must also generalize the seals, with their longitudinal, cone-bearing molars, and flat, fissured claws. Some of the seals also unite the scaphoid and lunar bones later in life than other Carnivora, henoe we would reasonably look for the division of these bones in their predecessors. The fiat-clawed genera of Wyoming:}: answer these demands. The genera Mesonyx and Synoplotherium presents us with a series of molar teeth wdiich repeat each other in form, are com- pressed below, and bear conical cusps. The jaws in the latter genus are slender, and the canines tend to the great development seen in many seals; but principally, the scaphoid and lunar bones are distinct, and the claws flat and widely fissured. The tympanic bone is more like that of the bear, and some seals, than that of the digitigrade Carnivora. These genera, though probably good swimmers, were well removed from the seals in the structure of the long bones of the limbs, and were probably remote in their ancestry. In Oligotomm, Orotherium, Ilyopsodus and similar forms, the conic tu- bercles of the lower molars have a slight alternation, and the posterior, which has a crescentoid section in wearing, inclines to connection with both the inner conic tubercles by low ridges. ^These ridges are fully devel- oped in Palmsyops so that we have a dental crest of two Vs, in the infe- rior molars. This in wearing produces the two crescents of Palceotherium. The addition of two tubercles on the inner side takes place in the higher forms, which terminates in the four crescent-bearing molars of the Rumi- nants. How this is done is proven later by examples from the maxillary teeth. In Orotherium vamcciense there is a tendency for the conic tuber- cles to be connected in pairs by low cross ridges. These ridges fully de- veloped produce the two cross-crests of Eyrachyus and Tapirus. In Bhi- nocerus the outer portion retains a crescentoid form, giving rise to an * Dr. Lock wood, of Rutger's College, in a recent number of the Popular Science Monthly, expressed serious suspicions of the Quadrumanous relationships of the Coati , little thinking at the time that the specimens to confirm his view were at that moment in the hands of palseon- tologists. t Dentes des Mammif«rs, p. .31. i See the Flat-clawed Carnivora of Wyoming, by E. D. Cope, April, 1873. 7 L-shaped crest. In Bathtnodou. diagonal ridges appear which would result in two Vs, as in PakeoHi/ops, were it not that both transverse and oblique elements of the posterior V tlisappear, leaving but one such in the middle and posterior part of the mandibular series. In Uintatherium the diag onal from the posterior crest never appears, leaving a transverse crest and a V on the true molars. In the superior molar series the flattening of the outer tubercles may proceed so far as to produce on wearing a confluence of the crescentoid surfaces. This is the case in OrotJierium sylvaticum in the mandibular series. In both Palceoi^i/ops and llyrachyuft those tubercles of the upper molars are confluent into two Vs (more or less open, when unworn). In the former the inner tubercles retain their primitive conic tubercular form, but in Palmotherium, Bhinoeerus, Lophiodon, Hi/racJiyus and Ta- pirus they elongate transversely so as to meet the corresponding outer tubercles (now crests) forming the familiar cross-crests of those genera. If the tubercles are alternate, they produce the oblique crest of Palceo- therium, if opposite, the cross-crest of Tapirus. If on the other hand the inner tubercles flatten like the outer, on wear- ing, we have the quadricrescentoid type of Anoplotherium and the Rumi- nants. But it is important to observe, that the lower types of Quadrumana and Garnivora present the quadrituberculate crown with tendency to flattening of the outer tubercles, as seen in these lowest Ungulata. In the Garnivora the sectorial toath is produced by the greater flattening and partial confluence of the outer tubercles, and the entire loss Qf the inner, the "heel" being in the dogs and cats, e.g. their only representa- tive. In the Quadrumanous families, including man, the primitive quad- rituberculate type of molars is preserved, the flattening of the outer tubercles being finally lost. It is to be observed that the lines of Ungulata, Quadrumana, and Gar- niDora, originate in plantigrade types, a state of things quite predomi- nant among the lower series, or Lissencephala. It is universal in Eden- tata and very usual in Rodentia and Insectimra. The lower forms of Marsupialia and all of the Monotremes present it. In the Marsupials, Rodents, Ungulates, and Carnivores we have series whose highest ex- pression is in the most highly digitigrade genera. The accompanying diagram is designed to express to the eye more clearly the propositions made above. By comparing it with a similar table published by Prof. Gill (Proceedings of the American Association for the Advancement of Science for 1871, p. 295), a close resemblance between the two may be observed, as well as certain differences. I wish to be understood that the genera named in it as ancestors, are to be regarded in the light of types of groups. There is no other mode of explaining the facts, than that in accordance with the law of "homolo- gous groups," i. e. that several genera of one group have undergone simi- lar modification into corresponding ones of a second group.* • See Origin of Genera, page 79, Prop. V. 8 Pal:EONtological Bulletin, No, 33. ON SOME MiMMALlA OF THE LOWEST EOCENE BEDS OF NEW MEXICO E. ID. OOIFE. {Extr. Am. Phil. Soc.) (Read before the Amer. Philosophical Society, Sept. 17, 1881.) FOR SALE BY A. E. FOOTE, 1123 BELMOJVT AVEJVUE, I Cope.] 484 [Sept. 17, On some Mammalia of the Lowest Eocene beds of New Mexico. By E. D. Cope. {Read hefore the American PJiilosojphical Society, Sept. 17, 1881.) Meson Yx NAVAJOvius, sp. no v. Smaller than the two known species, and with the crowns of the molars more compressed and the blades of the heels of the inferior series more acute. Molars seven, the first one-rooted. Last molar with a cutting heel like the others, and with the penultimate, with a rudimental anterior inner cusp. All the molars with an anterior basal tubercle except the first, second and third. No basal cingula. Principal cusp elevated and compressed, as in the premolars of Oxymna. Enamel minutely rugose. Mandibular rami and inferior canine teeth com- pressed, the angle of the latter not inflected. Length of inferior molar series M. .078 ; do. of premolar series .046 ; fourth premolar, length of base .010 ; elevation of cusp .008 ; second true molar, length .012, elevation .010 ; width of heel .005 ; depth of ramus at .020 ; diameter of base of crown of canine, vertical .009. Periptychus carinidens, gen. et. sp. nov. Creodontium. Char. Gen. No distinct sectorial teeth, the first and second true inferior molars simi- lar. They support a principal median cusp, a broad heel and a prominent anterior cingulum. The heel is more or less divided into tubercles ; the anterior cingulum is on the inner side, and represents the anterior cusp of a sectorial tooth. On the inner side of the principal cusp a cingulum rises, forming a flat internal tubercle. Last molar not smaller than the others ; premolars unknown. This genus belongs to the Amhlyctonidce with Amblyctonus and Palm- onyctis. It differs from both in the rudimental character of the anterior cusp, and from the former, in the presence of the internal tubercle. In Mesonyx the heel has a median cutting edge. Char. Specif. Parts of both mandibular rami and the shaft of a humerus represent this species. They indicate an animal of the size of the red fox, but much more robust. The mandibular ramus is rather shallow and thick, and the molars are not large. The heel of the penultimate supports three tubercles, of which the ex- ternal is the largest. The anterior cingulum supports a small cusp, and then rises to the internal tubercle, which is compressed. The sides of all the cusps are marked with distinct, well separated, vertical ridges. Each extremity of the internal cusp is connected with the principal cusp by a ridge. The first true molar has fewer cusps. Those of the heel are scarcely distinct, and form a border which rises prominently into the flat internal tubercle, which forms a narrow longitudinal blade. The anterior cingulum has no cusp and does not rise into the inner tubercle. The prin- cipal cusp has a strong entering groove next the inner tubercle. Length of crown first molar .0115 ; width of do. .006 ; elevation of do. .006. Length of second molar .011 ; width of do. .007 ; elevation of do. .0065. Depth of ramus at do. .020. The species is a good deal smaller tlian the Amblyctonus sinosus. 1881.] 485 [Cope, Trttsodon QUiviRENSTa, gen. et sp. nov. Char. gen. Dcrivod from the lower jaw. Probably only three premolars. True molars alike, consist- ing of three anterior cusps and a heel. The cusps are relatively small and the heel large. Of the former tlie internal is much smaller than the ex- ternal, and the anterior is rudimental, being merely a projection of the cingulum. The cutting edges of the large external cusp are obtuse. The heel is basin-shaped, and its posterior border is divided into tubercles, of which the external is a large cusp. The fourth premolar has no anterior inner tubercle, so that the anterior part of the crown consists of a com- pressed cutting cusp. The heel has two well-developed posterior cusps. The third premolar has a similar principal trenchant cusp, but a smaller heel. Canines large. This genus differs from Ilerpetotherium and Ictops in the simplicity of its fourth inferior premolar, and from Stypolophus am\ DeUathej^ium in the rudi- mental character of the accessory anterior cusps of the true molars, as well as in the three premolars. The rudimental anterior cusp of the true mo- lars, with the three similar true molars, separates it from Palmonyctis, and the presence of a conic inner cusp of the same indicates it as different from Amhlyctonus and Periptychus. It is not possible to state whether Triisodon must .be placed in the AmUyctoiiidm or not, on account of the absence of the superior molar teeth. This specimen of the type species of this genus is instructive as showing the succession of premolar teeth. Both the third and fourth premolars have temporary predecessors. The predecessor of the fourth premolar differs much from it in form, and is essentially identical in all respects with the true permanent molars. The crown of the predecessor of the third premolar is wanting, the roots only remaining in the jaw. The permanent third premolar was protruded before the permanent fourth. Which temporary tooth of Triisodon is homologous with the single one of the Marsupialia pointed out by Professor Flower?* As the additional permanent teeth of the placental Mammalia must have appeared later in time than the one already found in the implacentals, they must be those later protruded ; hence the fourth tooth in the jaw of Tri'isodon must be regarded as homologous with the fourth premolar of a placental, which is the last of that series to appear. If this be true, the tooth which follows the shed tooth of the Marsupials is not the fourth premolar, as supposed by Professor Flower, but the third premolar. This view is con- firmed by the fact that the milk tooth d'splaccd by the fourth tooth in Triisodon resembles in all respects the true molars, just as the permanent tooth occupying the same position does in Didelphys and some extinct eocene genera. This goes to show that this tooth, permanent in marsu- pials, is temporary in placentals, and that, in spite of its form in the for- mer group, it is the fourth premolar, and not the first true molar, as sup- posed by Professor Flower. Thus the posterior milk-molar of diphyodonts is a permanent tooth in the Marsupialia. * Transactions of the Royal Society, 1867, p. 031. Cope.] 486 [Sept. 17, Tliis observation confirms my conclusion that the Gredonta form n Qrou\^ intermediate between the Marsupialia and Carnivora. I may add that in Tri'isodon the inferior border of the lower jaw is not inflected posteriorly. Char, specif. — Size about that of the wolf. Inferior canine directed up- wards, its section nearly elliptic ; a faint posterior, no anterior cutting edge. Fourth premolar rather large, with an anterior basal cingulum which is angulate upwards, and is not continued on the inner side of the crown. Cusps of the heel each sending a ridge forwards, the internal lower, obtuse and descending to base of inner side of large cusp ; the ex- ternal larger, with an acute anterior cutting edge continuous with the cut- ting edge of the large cusp. True molars with an external, but no in- ternal basal cingulum. Border of heel with one large and three smaller tubercles, the former with, the latter without, anterior cutting edge. Ena- mel of all the teeth nearly smooth. All the cusps are rather obtuse. Measurements. — Length of inferior molar series : M. .080 ; long diameter of base of canine .013 ; length of true molar series .044 ; length of base of Prem. IV. .016 ; elevation of crown of do. .014 ; length of base of M. II. .016 ; width of do. in front .011 ; elevation of do., .014. The measure- ments of the jaw are not given, as the animal is not adult, the last molar not being yet protruded. From the lowest Eocene beds of Kew Mexico. DELTA-fHERiUM FUNDAMiNis, gcu. ct sp. uov. Char. Gen. Fam. Lep- tictidcE, agreeing with Ictops and Mesodectes in possessing an internal tu- bercle of the third superior premolar, but differing from both in having but one external cusp of the fourth superior premolar. Char. Specif. Repre- sented by the dentition of both maxillary bones minus the canines. The second premolar is convex on the inner face. The base of the third is a nearly equilateral triangle. The bases of the true molars are triangles, with the bases external. The internal angle supports an acute cusp, and has a posterior basal cingulum, which is very strong in the last three mo- lars. The two external cusps of the first and second molars are situated well within the base, which is folded into a strong cingulum. This cingu- lum develops strong anterior and posterior angles. This is the largest species of the family yet discovered. Extent of series of last six molars, M. .045 ; of true molars .026 ; diameters of fourth premolar, anteropos- terior .0074 ; transverse .0076 ; do. of second true molar, anteroposterior .0087; transverse .0100. This species was a fourth larger than the com- mon opossum, and very much resembles it in dental characters. CoNORYCTES COMMA, gen. ct sp. nov. Char. Gen. Allied to Mesonyx. Inferior canines not rodent-like, with conic crowns. Molars 3 — 3, the first one-rooted, the second two-rooted, the third with an anterior conic cusp and a posterior grinding heel. True molars consisting of two lobes, of subcylindric section, separated by deep vertical grooves. Enamel devel- oped on internal and external faces of crowns. Char. Specif Founded on a mandibular ramus which lacks the last molar, and has the crowns of the others worn. The external faces of the molars are much more ex- 1881.1 487 [Cope. posed than the interiuil, and arc somewhat contracted inwards. In the unworn crown there is a distinct anterior inner cusp, which is soon con- founded on attrition. The lieel of the hist premohir has a crescentic sec- tion, the internal horn the narrower. The anterior lobe is a robust cone. The base of the second and third premolar is oblique to the axis of the ramus outwards and forwards. It is possible that there is a minute first premolar filling the short space between the second and the canine. No cingula ; enamel obscurel3^ plicate, ramus robust. Length of molars minus the last .0405 ; length of base of first true molar .010 ; width of do. .009 ; elevation of crown do. .0055 ; length of base of fourth premolar .011 ; width of do. .008 ; elevation of crown of do. .0005. Anteroposte- rior diameter of base of crown of canine .010. Depth of ramus at first true molar .023 ; width of do. at do. .013. This genus differs from Estho- nyx in the form of the fourth premolar. In the latter the anterior lobe is compressed and trenchant. The species is larger than any of that genus, and nearly equal to the Ectoganus glirifonnis. Catathl^us rhabdodon, gen. et sp. nov. Char. Gen. With this genus I commence descriptions of several genera with bunodont denti- tion, which has some resemblance to that of some of the hogs. The one above named, with Miodamus, remind one of Tetraconodon Falc. and Lydd., in the enlarged proportions of their premolar teeth. I compare the genera as follows, introducing a i)robably perissodactyle form {Protogonia) for comparison : I. Third and fourth superior premolars one or two lobed externally, and with internal lobes. a. Superior premolars with two external lobes ; inferior fourth with two median cusps. Intermediate tubercles ; premolars not enlarged Phenacodus. aa. Superior premolars enlarged, generally with one external cusp. /9. A posterior internal cusp of superior molars ; Intermediate tubercles present ; last inferior premolar with inner cusp ; Catathlmus. Intermediate tubercles wanting, replaced by branches of an internal V ; no cusp on inner side of last inferior premolar Anisonchus. Intermediate tubercle present, connected with anterior inner by ridges ; inferior molars with Vs Protogonia. ft/S. No posterier inner cusp of superior molars. Intermediate tubercles present ; no inner lobe of last inferior premolar AEockmius. II. Superior premolars 1, 2 and 3 without inner lobe ; third with three external lobes (Pictet). Premolars compressed Dichohiine. In the genus Catathlfms the development of the preniolars is remarkable while the true molars are relatively small. The last three superior pre- molars have an elevated internal crescentic cingulum liomologous with Cope.] 488 fSept. 17, tlie inner lobe of the fourth superior premolar of the ruminants. The general character of the true molars is that of Phenaeodus. Parts of two or three individuals of this species have come into my possession, one of which includes nearly all of the molar dentition of both jaws. The external cusp of the superior premolars is compressed conic, and the in- ternal cingulum extends to its anterior base in the second, third, and fourth. The crown of the last true molar is about as long as wide, while that of the first is wider than long. Each supports seven cusps ; two subconic external, and one large median internal, which is connected by ridges with a small anterior and posterior median. Then there are a small anterior and posterior internal, making three internal. The internal crest is distinct from the principal cusp in the inferior premolars III and IV, but unites with it in the II ; it supports on the lY an anterior, a median and a posterior cusp, the latter forming part of the rather narrow heel. The true molars I and II have seven tubercles, the four principal ones, and three smaller, one anterior, one posterior, and one median. On the third the posterior forms a large heel. All of the molars, but especially the premolars, have the enamel thrown into sharp vertical parallel folds, in a manner I have not seen in any other mammal. Length of six superior molars .067 ; length of three true molars .029 ; length of base of third premolar .012 ; width of do. .012 ; width of base of first true molar .010 ; do. of third true molar .009 ; length of do. .010. Length base fourth inferior premolar .012 ; width do. .010 ; length of third true molar .0115 ; width of do. .009. The teeth indicate an animal of the size of the peccary. Anisonchus sectorius, gen. et sp. nov. Char Gen. This is derived from the superior P-m. IV and M. I and II, and from all the inferior mo- lars of three individuals. The superior teeth are accompanied by a ramus mandibuli, which contains alveoli of all the inferior molars, and the crowns of the P-m. IV and M. II. The leading characters have been given above. The inner posterior lobe is more prominent in this genus than in any of the others, and has a V-shaped apex. It projects further inwards than the anterior inner lobe. It is represented by a mere tubercle of the cingulum in Mioclcenus. In the lower jaw the last premolar is quite simple, consisting of a principal cusp, and a non-cutting heel. The second true molar has intermediate anterior and posterior cusps. The ge- nus difters from Pantolestes in the more numerous tubercles of the molars, and in the fact that the anterior inner tubercle of the true molars is not double. It may, however, be allied to that genus. Char. Specif. The fourth superior premolar covers a larger base than either of the true molars. The external cusp has a base extended antero- posteriorly, but the apex is conical, and there are no basal tubercles. The inner cusp has a crescentic base as in CatathlcBus, but the apex is nar- rowed and compressed conic. The external tubercles of the true molars are subconic, and do not develop any external ridges. They are connected by the crescentic slightly angular crest, whose apex forms the inner ante- J881.] 489 [Cope. rior boundary of the crown. This crest is not divitled into parts lioinolo- sious with the intermediate tubercles. The crowns of tlie M. I and II are surrounded by a basal cinguUmi, which in the M. I develops a tubercle at the anterior external angle. No internal or external cingulum on P-m. IV. Enamel nearly smooth. The ramus of the mandible is rather slender anteriorly. The P-m. IV is robust, and the cusp is behind the middle of the base of the crown. The heel is short and narrow, and has a raised border, connected with the base of the main cusp. The cusps of the second true molar are elevated and conic, the anterior external the highest, the others subequal. The base of the posterior pair is a little narrower than that of the anterior pair. There is no central tubercle as in CatcUhlmus rhabdodon, and no basal cingulum on cither tooth. Measurements. M. Length of three superior molars OIGO anteroposterior 0055 transverse 0070 anteroposterior 0052 transverse 0060 Diameters superior P-m. IV Diameters superior M. I Length of inferior molar series 0610 " " " true molar series 0160 ^ . . ( anteroposterior 0060 Diameters mferior P-m. IV < , i transverse 0040 . „ . -r-r (anteroposterior 0050 Diameters inferior M. II ^ , I transverse, 0040 Depth ramus at M. II 0090 A number of minor points will distinguish this species from those in- cluded among the Mesodonta, and especially those of Pcmtolestes, which it most resembles. The molar teeth are narrower behind, and the fourth premolar is larger. It is Miodmius sectorius, American Naturalist, Octo- ber, 1881, p. 831, MiocLiENUs TURGiDus, gen. et sp. nov. This genus differs from Catathlmus in the structure of the inferior premolars, which are without internal crest or cusp. The inner lobe of the superior premolars is less developed than that genus. In the present species the characters of Mioclmnus are best seen in the subconical tubercles of the premolars, particularly that of the heel of the fourth inferior premolar. In the other three species this heel is more of a crest and is connected with the princi- pal cusp by a low ridge. The four species may be characterized as follows : a. Cusps of last premolars conical in both jaws. Size medium. Last lower molar disproportionately small ; cusps low ; two anterior inner distinct ; true molars, .018 M. turgidus. aa. Fourth superior premolar with flattened external and conic inter- nal cusp ; inferior unknown. Cope.] 490 [Sept. 17, Size medium ; fourtli upper premolor equilateral ; all cusps acute ; true molars .01G5 M. subtrigonus. aaa. Cusps of last premolars compressed in lower jaw. Least. Second and third lower true molars subequal ; cusps, especially tlie internal, elevated ; anterior inner confluent into an edge ; true molars, .013 M. angustus. Largest ; cusps of inferior molars obtuse ; P-m. Ill .008, its heel short and small M. mancUhularis. Medium ; last inferior molar larger than penultimate ; true molars, .014 ; P-m. Ill .006 Anisonchus sectorius. Of M. turgidus there are two specimens ; and of M. subtrigonus, M. an- gustus and M. mandihularis one each. In the M. turgidus there are no cingula on the fourth premolar. It is wider than long, and the external face is a little flattened. The tubercles are conic ; the external has a small one at the anterior base, and a rudi- ment at the posterior base, and there is a low one on the posterior side at the middle. The second true molar is wider than the first. The tubercles are all round in section. Besides those already mentioned, there is a rudi- ment of a posterior inner on the first, which is represented by a cingulum on the second. The latter has basal cingula all around except on the inner side ; the same are visible on the first tme molar in a rudimental condition. Enamel nearly smooth. The inferior molars are of robust proportions. Their sizes are, commenc- ing with the largest : P-m. lY ; M. II ; M. I ; M. III. The last molar is only half as large as the penultimate. It has two anterior and an external lateral tubercles, and a heel. On the penultimate molar, there are two an- terior tubercles with a trace of anterior inner ; also a broad flat heel, with a low tubercle on the external side. The constitution of the first true mo- lar is identical. The fourth premolar has a rudimental heel consisting of a low tubercle only. The principal cusp is conic and is over the middle of the transverse diameter, and a little behind the middle of the antero- posterior diameter. No cingula. Enamel nearly smooth. Measurements. M. Maxillary bone. Length of base of P-m. lY, M. I and M. II 0175 ^. , -r. ^ anteroposterior 0055 Diameters base P-m. lY < . ^^p- i transverse 0065 ^. ^ , ^.^ T , anteroposterior 0060 Diameters base M. I { , ^^^^^ transverse 0070 anteroposterior 0060 . transverse 0095 Mandible. Length of bases of last four molars 0250 anteroposterior 0070 transverse 0055 \ Diameters base M. II Length of bases of 1 Diameters P-m. lY 1881.] 491 [Cope. Measurements. M. anteroposterior OOGO insvcrsc 0.000 anteroposterior 0055 transverse 0043 Depth of ramus at M. 1 0115 Thickness " " " 0085 T ( antei Diameters M. I < ^ i tranf Diameters M. Ill MioCLyENUs SUBTRIGONUS, sp. nov. Represented by a portion of a cra- nium anterior to the orbits and lacking the extremity of the muzzle, dis- torted by pressure. It exhibits nearly all of the molar teeth. The spe- cies differs from M. turgidus in the greater acuteness of all its cusps, and in the equilateral form of the fourth premolar. It is too large to belong to the M. angustus, which is represented by a mandible only ; and too small to be the M. mandihularis, whose maxillary dentition is unknown. The inner borders of the molar teeth are shorter than the outer, espe- cially in the last two molars. The last true molar is smaller than either of the others. The cusps are all subconical, but the internal is connected with the intermediate by ridges, which give it a triangular section. The latter form a V, homologous with that in Anisonchus, but not so distinct, and the intermediate tubercles are not lost in its branches as in that genus. The posterior inner lobe of that and other genera, is represented by a thickening of the cingulum. This cingulum extends entirely round the P-m. IV and M. I, and M. II ; the M. Ill is injured. The sides of the base of the P-m. IV are slightly concave. The enamel of all the molars is wrinkled. Measurements. M. Length of bases of last five molars 0285 -r,. ^ ^ , r. -n TTT ^ anteroposterior 0060 Diameters of base of P-m. TV < . Diameters of base of M. I Diameters base of M. II Diameters base of M. Ill transverse 0050 c anteroposterior 0060 ( transverse 0060 c anteroposterior 0060 t transverse 0075 anteroposterior 0040 transverse 0060 MiocLyENUS ANGUSTUS Cope, American Naturalist, 1881, October (Sep- tember 22d), p. 831. The least species of the genus, with the teeth about the size of Ilyopsodus paulus Leidy, but with more robust jaw. The molar teeth diminish in size regularly posteriorly from the P-m. IV. They all have three subequal posterior cusps which are less elevated than the ante- rior ones. The median is enlarged into a heel on the last tooth. The anterior are opposite, and the external is larger than the internal. There is no anterior internal. The external wears into an anteroposterior narrow grinding surface, which looks like a combination with an anterior median. The latter is, however, not separate on the least worn molars. The Cope.J 492 [Sept. 17, anterior outer cusp increases in size anteriorly, and is the large cusp of the P-m. IV. It sends a branch backwards on the inner side of the crown which forms the edge of the narrow concave heel. There are no cingula except a short one on the anterior corners of the base of the crown of the P-m. IV. Enamel obscurely wrinkled. Measurements. M. Length of posterior four molars 0180 Diameters of P-m. IV \ «°'<'™P°«t»io'- 0050 i transverse 0035 , r.TVT T (anteroposterior 0050 Diameters of M. I \ ^ ( transverse 0035 Diameters of M. II anteroposterior 0040 transverse 0032 Diameters of M. Ill \ anteroposterior 0045 ( transverse 0030 Depth ramus at M. 1 0110 Thickness " " " 0060 Phbnacodus puercensis, sp. nov. Three individuals. Last superior molar smallest ; first and second true molars with six tubercles, two ex- ternal, two median and two internal. A strong basal cingulum except on inner side. Inferior true molars besides the usual five tubercles, furnished with an anterior ledge with a tubercle at its interior extremity. A weak external basal cingulum. A little larger than the P. vortmani. Length of superior true molars M. .031 ; length of base of crown of M. Ill .006; do. of M. I .008; width of do. .008; length of base of crown of inferior M. Ill .0085 ; width of do. in front .006 ; depth of ramus at M. I .019. Phenacodus zuniensis, sp. nov. The least species of the genus, rep- resented by the mandibles of two individuals. The first and second true molars are narrowed in front, and there is no distinct anterior ledge, only a minute anterior inner tubercle. The external cingulum is more distinct and the enamel is wrinkled. The fourth premolar has a short base and the inner cusp is much smaller than the principal one ; it has a wide heel and an anterior basal tubercle. Length of true molars, M. .018 ; of last true molar .006; of base of first true molar .006; width of do. .004; depth of ramus at do. .011. Protogonia subquadrata, gen. et sp. nov. Fourth superior premolar with one external and one internal lobe. True molars with two external, two internal, and two intermediate lobes, both the latter connected with the anterior internal by a ridge. Supposed inferior true molars with two Vs with weak anterior branches ; last true molar with heel. This genus will enter the Chalicotheriidm of my system of Ferissodac- tyla,* if the feet are found to possess the requisite characters. It is allied, apparently, also to Hyracotlierium, but difiers in the Vs of the infe- *See Proceedings Amer. Philosoph. Society, 1881, p. 377-8. 1881.] 493 (Cope. rior molars, if they are properly identified ; and in the superior molars. The anterior transverse crest of that genus is represented in Protogonia, but not the posterior. This is replaced by a low ridge running across the course it pursues in Ilyraeotherium. The posterior median tubercle is also not found in the latter genus. Protogonia differs from Limnohyus in the subconic character of the external lobes of the superior molars. If the tu- bercles, excepting the posterior inner, should be converted into crescents, the genus Meniscotherium would be produced. Char. Specif. Probably two specimens ; one supporting three superior molars ; the other including damaged superior molars and the last two in- ferior molars. The animal was about the size of the red fox. The exter- nal cusp of the fourth superior premolar is flattened externally, and has a small lobe on its posterior edge. The inner tubercle is conic and is sepa- rated by a tubercle from the anterior base of the external. True molars without external ridges. The external cusps of the true molars are lentic- ular in section. The posterior inner cusp is in nearly the same antero- posterior line with the anterior, its section about equaling that of the intermediate cusps. The first and second molars have an external, an anterior and posterior, but no internal, basal cingula. The enamel is somewhat wrinkled where not worn. The heel of the last inferior true molar is elevated, and its worn surface forms the extended posterior branch of the posterior V. The posterior edge of the penultimate molar is elevated and curved forwards on the in ner side of the crown. The anterior cusp forming the angle of the V of this tooth, is higher than the posterior angular cusp, but the anterior limb descends rapidly as in Corypliodon. A weak antero-external, and postero- external cingula. Enamel wrinkled where not worn. Measurements. M. No. 1. Length of bases of three superior molars 025 anteroposterior 0066 Diameters of superior P-m. IV Diameters of superior M. I Diameters superior M. II transverse 0086 anteroposterior 0085 transverse Oil anteroposterior 009 transverse Oil No. 2. Length of bases of last two inferior molars 0225 -r,. , X- x> • 1 (anteroposterior 0114 Diameters of last mferior molar ^ ^ ^ ( transverse 0066 -r,. ^ i>- ^ • Ti/r TT (anteroposterior 0112 Diameters of inferior M. II < ^ . . ( transverse 0080 Depth of ramus at M. II , 0240 Thickness " " 0110 Meniscotherium TERR^RUBiiiE, sp. nov. My specimens of this species embrace the dentition of several individuals. Cope.] 494 [Sept. 17, The dimensions of the superior molars increase to the penultimate, while the external and posterior sides of the last molar are contracted, re- ducing its size. The external faces of the external Vs of the true molars are considerably impressed ; those of the premolars are nearly flat. The second premolar is two-rooted, and has a compressed crown, with- out either heel or cingulum, except a thickening of the posterior base. The base of the crown is triangular. The external plate of the third pre- molar is simple, and is connected with the internal cusp by a cingulum on the posterior base of the crown. The crown is transverse, and the inner tubercle rather small. The fourth premolar is much larger than the third. Its external plate is divided into two apices, which are not im- pressed. Their external faces are separated by a faint ridge, and are divided medially by a faint ridge. The anterior external angle is rather prominent. The anterior and a posterior cingulum extend to and round the inner base of the interior tubercle. Within the anterior external apex, is a well developed intermediate crest parallel to it ; and there is a corres- ponding crest within the posterior external apex. This one turns inwards at its posterior extremity, which is on the posterior cingulum. The anterior angle or horn of each external crescent of the true molars is very prominent. They are sections of short vertical ridges, which unite near the base of the crown, giving abruptness to the impression of the ex- ternal surface of the anterior lobe. The middle of each face has a faint median ridge. The two molars have an anterior basal cingulum, but no posterior or internal, excepting a trace between the bases of the internal lobes. The anterior intermediate crescent is quite parallel with the ex- ternal ; the anterior internal tubercle has a slightly V-shaped section. The posterior inner tubercle is quite confluent with an oblique intermediate crest, as in M. chamense. In the last true molar, as there is only one in- ternal tubercle, this crest is short, terminating at the posterior border. The last true molar is like the last premolar, except in its two impressed external crescents. A fragment of the right branch of the lower jaw supports two molars, and the alveoli of two others, all of which have two roots. These teeth are the four premolars, although the last one has the form of the first true molar. Should my surmise be correct, then the third premolar has nearly the same form and structure as the fourth. The anterior horn of its ante- rior V is not produced quite so for inwards as in the fourth tooth. At the point of junction of the adjacent horns of the Vs there is a slight antero- posterior extension, forming a median buttress of the inner side of the crown as in AncMtherium. The posterior horn of the posterior V is also incurved, as in that genus. The angles of the Vs of the inferior molars are rounded. The surfaces of the enamel of the teeth of both jaws is smooth. Measurements. Length of superior molars, less P-ni. I " of true molar series " of base of P-m. II M. .046 .028 .005 ISSl.] 495 [Cope. Measurements. M. T^. ^ TTT (anteroposterior 000 Diameters of base P-m. Ill -< * ^ ^ (transverse 007 " of base P-m. IV | ^^"teroposterior 008 (transverse 010 " of base ot M. II 0" (transverse 013 C vertical 005 inferior P-m. Ill (or IV) I anteroposterior .. . .007 transverse 005 Depth of ramus at same tooth 012 Thickness ramus at succeeding tooth 009 The Memscotherium terrmruhrcB differs from the M, cliamense in two features. The first is its superior size. The second is the flattened form of the external f\\ces of the true molars and the absence of the convexity of the external bases of the crown. My specimen of this species is from the red Eocene bed in Northwestern New Mexico, from the true "Wasatch horizon, or higher than that which produced the other species here described. It was found by my assistant, D. Baldwin. Remarks. As stated in my report to Lieut. Wheeler in 1877, no vertebrate remains had been found in the Puerco beds, which underlie the Wasatch in New Mexico, up to that time. It was therefore uncertain whether they form the top of the Cretaceous or the bottom of the Tertiary series. I have recently obtained evidence of the existence of Chnmpsosaurus in them, so that their position might be supposed to be in the Postcretaceous S3^stem. It is however quite possible that the species of Mammalia described in this paper were derived from the Puerco Formation. Their horizon is be- low the Wasatch, and they represent a different fauna from that of those beds. Attention has already been directed to this fauna in the pages of the Amer- ican Naturalist.* I have recorded the presence of the Creodont genera, Periptychus, Tri'isodon and Deltatherium, and of the saurian Champ- sosauniH. I have now added the genera Uyraeotherium and Meniscotherium,, and a number of new forms of considerable interest. These are the Creo- dont Mesonyx, a new genus allied to Esthonyx, and a series of genera and species with a suilline type of dentition, but whose affinities are by no means certain. This point cannot be determined until the characters of the feet are known. The facies of this fauna differs in several points from that of tlic Wasatch. Gorypliodon has not yet been discovered in it, and the flesh- eaters are very primitive. The suilloid genera are characteristic. * April, August and October, 1881. PkTNTRD SEPTEMnKll 30, 1881. ■* 9 Pal^ontolggical Bulletin, No, 34. CONTRIBUTIONS TO THE HISTORY OP THE YERTEBRATi OF THE LOWER EOCENE OF WfOMIHG AND NEW MEXICO, MADE DURING 1881. ID. oo:e=e- (Extr. Am. Phil. Soc) (Read before the Amer. Philosophical Society, Dec. 16, 1881.) FOR SALE BY A. E. FOOTE, 1^23 BELMO.KT AVEJVUE, 4^ ' but is well distinguished from the other cusps. The internal median cusp is small, the external median, large. The premolars are not so much larger than the true molars in this as in the typical species of the genus. The second and third are more elongate on the base than the fourth. The latter is also less compressed than those that precede it. It has a short wide heel, and a small anterior basal tubercle. In the second and third premolars the posterior edge of the principal cusp is sharp> and descends gradually to the posterior base of the crown. Both have small acute anterior basal tubercles. The first inferior premolar is one-rooted> and has a simple crown directed some- what forwards. It is separated from the second by a short space. The teeth anterior to this point are lost. Measurements^ M. Length posterior four superior molars. .-. ^ 0182 Diameters p.ni. IV | ^^^^'^P^'*^^^^' ' I transverse 0043 Diameters M. II | anteroposterior 0043 t transverse. » » , 0060 Diameters M. Ill | anteroposterior 0040 transverse . . . . » » 0060 Length of inferior molars » .0330 ** premolars 0193 P-m. Ill 0055 " P-m. IV 0045 Diameters M. L j anteroposterior 0040 I transverse 0033 Diametei^ M. Ill | anteroi>osterior .0053 transverse 0040 Depth of ramus at P-m. I .0047 M. II 0090 Length of astragalus 0102 Width of trochlea behind 0048 Diameters of cuboid j ^^"Sth 0070 I width of middle 0040 MlOCL^NUS ETSAGICUS, Sp. nOV. This, the largest species of the genus, is represented by the two rami of Cope.] 190 [Dec. 16, a mandible of an adult animal in good preservation. In their robust character the premolar teeth resemble those of the M. turgidus, but are not relatively so large, nor is the last true molar relatively so small, as in that species. The heel of the third premolar is obsolete, and that of the fourth is a wide cingulum. Neither exhibit an anterior basal tubercle, and in both the principal cusp is stout. The true molars widen posterior- ly to the anterior part of the last molar. The latter contracts rapidly to a narrow heel. The tubercles are all subconic, and the median ones of the last molar are small. There are no cingula, and the enamel is smooth. The ramus is not robust, and is of moderate depth. Its inferior border rises below the middle of the last molar tooth, and posteriorly. There is a "mental" foramen below the contact of the fourth premolor and first true molar. Measurements. M. Length of bases of six posterior molars 047 " " three premolars 024 P-m. II 009 P-m. IV 008 " " P-m. IV 005 Diameters basis of M. II j anteroposterior 0075 ^ transverse 0070 Diameter basis M. ni | ^^^^^^P^^^^^^^^ ^^^^ transverse 0070 Depth of ramus at P-m. II 0080 M. II 0140 This species is named from the Crow Indian name of the Big-Hom river, Etsagie. Concluding Remarks, The imleontologist who has examined the preceeding list, will readily perceive that it represents fully the Wasatach ftiuna, with little admixture of earlier or later forms. The only genus which belongs to the Bridger or middle Eocene, which occurs in the Big-Horn basin, is Pappichthys, The characteristic Bridger genera Hyracliyus, Palceosyops, Uintatherium, and the J'illodonta, are absent, and their place is taken by Phenacodus, Ilyra- cotherium, CorypJiodon and ToBniodonta, as in New Mexico. Several genera are, as elsewhere, common to the two horizons, and two species cannot be distinguished in the parts preserved. Such as Hyopsodus paulus and H. mcarius. A closer comparison may be made with the Wind- River group, on which I published a report in the Bulletin of the U. S. Geological Survey of the Territories.* The folloMing genera found in that formation have not been obtained from the Big-Horn. Protopsalis, Lambdotherium, Palceosyops, Hyrachyus.\ Genera of the Big-Horn not obtained from the Wind -River : Cynodontomys, Anaptomorphus ; Mesonyx, *1881, Feb. p. 201. t Since making my report on the Wind-River fauna, 1 have found the anterior part of the lower jaw of a species of this genus. 191 iCope. Deltatherium, Oxycena ; Manteodon, Ectacodon, Metalopliodon ; Anacodon, OUgotomus, Systcmodon ; Mioclmius. Three of these genera have been found in the Bridger, and five have been obtained in the lower Eocene of New Mexico. Five of the genera are new to science. An especial feature of the Big-Horn collection, as distinguishing it from those brought from other regions of the Wasatch formation, is the presence of numerous species of Phenacodus, and of new and rare species and genera of Coryphodontidce. II. The Fauna of the Catathl^us Beds or Lowest Eocene op New Mexico. A number of new species and genera from this horizon were described in my Paleontological Bulletin No. 33. The present paper adds a few to this list. Up to the present time no species of Coryphodon, and but few specimens of 5?/r«costerior transveree,, as well as a longitudinal median keel. The fragments of the supposed second individual include two large glenoid cavities with strong preglenoid crests, as in Mesonyx. As compared with Mesonyx, this genus differs in the V-shaped crest of the penultimate superior molar ; in Mesonyx it is represented by a simple cone. The last superior molar of Mesonyx is triangular and not transverse,, but the composition of the crown of that tooth, in Sareothraustes must be Diameters of inferior molar Measurements. f of cusps anteroposterior transverse .... .0070 .0052 .0110 .0065 Cope.] 194 [Dec. 16, known before the value of this character can be ascertained. If the view that Sarcothraustes has but three inferior premolars be correct, this charac- ter distinguishes it from Mesonyx, as do also the transversely expanded heels of the molars. The family MesoiiycMdm may be for the present re- garded as embracing the three genera of Sarc&thraustes, Mesonyx and Dissacus.^ Char. Specif. The penultimate superior molar has a strong posterior cingulum which commences within the line of the internal bases of the external cusps, and rises into considerable importance behind the internal cusp. There is also an anterior cingulum which does not rise internally, and which is continuous with a strong external basal cingulum. The latter passes round the posterior base of the posterior cone, and runs into the posterior branch of the internal Y. The posterior cone is smaller than the anterior cone, and its apex is well separated from the latter. The ap- pearance of this tooth is something like that of a carnivorous marsupial. The symphysis mandibuli slopes obliquely forwards, and is united by coarse suture. The ramus is stout and deep, as compared with the size of the molar teeth. The roots of the teeth are relatively large, especially those of the first two premolars. The crown of the canine is lost. The •first premolar points forwards, nearly parallel with the canine, and diver- gent from the second premolar. The crown of the second premolar is small and subconic, and has a rudimental heel, and no anterior basal tuber- cle. The first true molar resembles considerably that of Mesonyx. There is a small anterior basal tubercle on the inner side of the principal cusp. The expansion of the heel is transverse only, there being no longitudinal lateral edges or tubercles. The enamel is obsoletely, rather coarsely wrinkled. There are two rather large mental foramina ; the posterior be- low the anterior root of the first true molar, and the anterior below the l^osterior root of the second premolar. Measurements. M. Diameters of superior M. ii J i''"«'-oposterior externally .015 i transverse 024 Anteroposterior diameter of base of M. .iii 0095 Anteroposterior diameter base of crown of inferior canine 020 Length of bases of three inferior premolars 038 ( anteroposterior 019 Diameters inferior M. i. \ transverse 0095 1- vertical 0110 Depth of ramus at P-m. iii 0520 Width " " 022 * Amei-ieaii Xatuialist, Dec, 1881. ISSl.J 195 [(Jopc. ClIAMPSOSAUHUS rrKUCKNSIS, Sj). IIOV. I liiive already announced the discovery* of this Laramie genus in (lie Puerco beds of New Mexico, and described a species, C. amtralu, from that region. I now introduce two additional species from the same hori- zon. One of these is represented by a number of fragments which include three dorsal and four caudal vertebne of apparently one individual. They represent an animal of larger size than any of those heretofore referred to Champsostmrus, excepting the G. vaccinsulenHU. In all of the vertebne the neural arch is more or less codssified with the centrum, and the animal had probably reached its full size. One of the dorsal centra is split vertically and longitudinally, and shows the structure already figured by Leidy in the Ischyrosaurua antiquus \ Leidy. The surfiice exposed displays two diagonal lines of fissure cross- ing each other at right angles. Thej^ indicate cleai'ly the mode of origin of this amphiplatyan type of centrum. The centrum is first deeply am- phicoelous as in the Theromorphous reptiles of the Permian. The conical cavities are filled by the ossification of the remaining portions of the noto- chord, forming a conical body which always remains distinct from the re- mainder of the centrum. The articular faces of the dorsal centra are a little wider than deep, and the depth about equals the length of the body. They are not nearly so depressed as those of C. austraUs, and their outline is difl:erent. This is wider above and narrows below ; in both C. austraUs and C. saponensis the inferior outline is part of a circle. None of the dorsals preserved are keeled below. There is a fossa below the diapophysis wiiich has a subver- tical posterior boundary. The general surfiice (somewhat worn) does not display wrinkles near the articular faces. An anterior dorsal has a short compressed diapophysis with a narrow figure 8 articular surface, and its superior border is in line with the roof of the neural canal. The anterior caudals have subround articular faces ; the posterior are more oval and the bodies compressed. With greater compression, the length increases. Measurements. M. / anteroposterior 025 Diameters of an anterior dorsal } vertical 025 ' transverse 030 Height of costal facet of do 021 Diameters neural canal do. \ ^^'"^'c^^ ^07 t transverse 009 / anteroposterior 024 Diameters anterior caudal ^ vertical 021 ' transverse 021 ^ anteroposterior 025 Diameters posterior caudals } vertical 018 ' transverse 018 * American Naturalist, 1881. p. m\). t Transac. Anier. Philo.s. Soc. 1800. Cope,] 196 [Dec. 16, The typical specimen was found by Wm. Baldwin near the Puerco river, west of the Nacimiento mountain, New Mexico, in the typical locality of the Puerco formation. Champsosaurus saponensis, sp. nov. Represented in my collection by six cervical and several dorsal vertebr£e, one only of the latter with well preserved centrum, parts of ribs, and various other bones, whose reference is not yet certain. The cervical vertebrae include the os dentatum or centrum of the atlas. This shows its streptostylicate character in its distinctness from both the centrum and the free hypapophysis of the axis. Nevertheless it is more Crocodilian than Lacertilian in form. Its anterior face is transverse, with a little lip carrying forwards the floor of the neural canal, below which the face is leveled posteriorly. The inferior surface is narrow and transverse, as though adapted for the anterior part of the hypapopliysis of the axis. At each side it terminates in a prominent tuberosity, as though for the attachment of a cervical rib as in the Crocodilia. The anterior face is bounded posteriorly by a transverse groove which terminates in a fossa on each side. The posterior articular face of the os dentatum is wider than deep. The lateral angles of the superior face are rounded, and its median portion is concave. The axis displays a large facet for the hypapophysis. Behind it the inferior middle line is not keeled, but is coarsely wrinkled longitudinally. The posterior edge of the liypapophysial facet is the most prominent part of the inferior surface. The posterior articular face is deeper than wide. This is true of the faces of all the cervical vertebrae. The latter gradually increase in size posteriorly, and the dorsals become larger. The articular faces of all the centra are regularly rounded and not contracted below. The five cervicals are strongly keeled below ; the keel of the third centrum being split up anteriorly into narrow ridges. On the sixth the keel is more prominent and acute. The dorsal is not keeled. A trace of the parapophysis appears low down on the fourth cervical ; it rises and becomes prominent as a round tuberosity on the fifth and sixth. It ap- pears on the superior edge of the centrum of the dorsal vertebra, where it is connected with the diapophysis. It is near the middle of the length of the centrum, and not near the anterior border as in G. australis. The surfaces of the vertebrae are very smooth excepting where thrown into coarse wrinkles near the borders of the articular faces and near the hypapophysis. The edges of the articular faces are somewhat revolute on the sides in the cervicals, but not on the dorsal. They are impressed in the centre to a point, most strongly so as we pass forwards in the series. There is a fossa below the space anterior to the parapophysis of the dorsal vertebra, which is abruptly bounded below by a horizontal angle. A separate neural spine perhaps of a cervical vertebra, has the following form. It is stout, and is contracted rather abruptly at the apex from behind forwards. The section is broadly lenticular, angulate in 18M.] 197 [C<)1)0 front, and truncate behind. The posterior face has several longitudinal -wrinkles, including a median raised line, and there are some more irregu- lar wrinkles on the sides. Medfturoiicjits of vertchrfF. ]M. Anterior face of os dentatum f width 025 I depth (oblique) 012 Posterior face of os dentatum | ^^'^^^^h 020 i depth 01 H Length os dentatum above 014 ( , . c rde])th 022 \ posterior face . ^ Diameters axis ^ i. width 020 ( length 0185 Hypapophysial facet ()s dentatum | JjJ^ ^ length 022 Diameters fourth cervical < ^ . r depth 0225 ) anterior , /, , I width 022 lenuth 0215 Diameters sixth cervical < j^jj^^^j-i^^^. / depth 024. 1 width 0235 Hpaces between parapopliysis and diapophysis of do 0040 / length 02G5 Diameters of dorsal < ^j^t^^j-iQ^. depth 0260 ( I width 0265 Height of neural spine of?, from postzygapophysis 0210 Anteroposterior width of do. at base 0100 The portions of vibs are separated heads and shafts. The former are double and therefore cervical, and are quite large. If the shafts belong to them, the neck of this species must have been wide. The shafts are slender and are of dense bone. The section is oval at the middle, but towards the distal extremit}^ becomes tlattened and grooved and delicately line ridged on one side. The extremities of the long bones are A^ ithout con- dyles but have concave surfaces like those of the ribs. The bodies are ro- bust and angular. They may be abdominal ribs of unusual stoutness. From the Puerco beds, D. Baldwin. PuBTJHHED FEnnrAKY 20th, 1882. Paleontologieal Bulletin, No. 35. tup: CLASSIFICATION" OF THE UNGULATE MAMMALIA. {Rend before the American Philosophical Society, May 19, ISSS.) TO THE HISTORY OF THE VERTEBRATA OF THE PERMIAN FORMATION OF TEXAS. {Read before the American Philosophical Society ^ September 15, 1882 T) SYNOPSIS OF THE VERTEBRATA OF THE P=»XJEI^OO EOOEnSTE EIPOOI^:. {Read before the American Philosophical Society, October 20, 1882.) 02!T TUB SYSTEMATIC RELATIONS OF THE CARNIVORA-FISSIPEDIA. {Read before the American Philosophical Society, October 20, 1882. ) By PROFESSOR E. D. COPE. For Sale by A. E. Foote, 1223 BELMONT AVENUE, PHILADELPHIA. U. OF \U1 UK Cope.] 438 [May 19, Tht Classification of the Ungulate Mammalia. By E. D. Cope. {Read before the American Philosophical Society, May 19, 1882. ) In the present essay the osseous system is chiefly considered, and of this, the structure of the feet more than of any other part of the skeleton. The ungulata are here understood to be the hoofed placental Mammalia with enamel covered teeth, as distinguished from the unguiculate or clawed and the mutilate or flipper limbed, and the edentate or enamelless, groups. The exact circumscription and definition is not here attempted, though probably the brain furnishes an additional basis of it in the absence of the crucial, parietooccipital, calcarine fissures, etc. Suffice it to say that it is on the whole a rather homogeneous body of mammalia, especially distin- guished as to its economy by the absence of forms accustomed to an insectivorous and carnivorous diet, and embracing the great majority of the herbivorous types of the world. The internal relations of this vast division are readily determined by reference to the Characters of the teeth and feet, as well as other less im- portant points. I have always insisted that the place of first importance should be given to the feet, and the discovery of various extinct types has justified this view. The predominant significance of this part of the skeleton was first appreciated by Owen, who defined the orders Perisso- 1882.] 439 [Cope. dactyla and Artioddctylo. Professor Gill* has also used these characters to a large extent, but without giving them the exclusive weight that appears to me to belong to them. Other authors have either passed them by unnoticed, or have correlated them or subordinated them to other charac- acters in a way which has left the question of true affinity and therefore of phylogeny, in a very unsatisfactory condition. Much light having been thrown on these points by recent discoveries in paleontology, the results, as they appear to me, are here given. Fig. 1. Fig. 1.— Left anterior foot of Elephas a/ricanus (from De Blainville). Carpus. — It is well known that in the Perissodactyla and Artiodactyla, the bones of tlie two rows of the carpus alternate with each other ; that the lunar for instance rests on the unciform, and to a varying degree on the magnum, and that the scaphoides rests on the magnum and to some degree on the trapezoides and trapezium. It is also known that in the Proboscidea, another state of affairs exists ; i. e., that the bones of the two rows do not alternate, but that the scaphoides, lunar and cuneiform, rest directly on the trapezium and trapezoides, the magnum, and the unciform respectively. The preceding characters are sometimes included in the definitions of the respective orders. Further than this they have not been used in a systematic sense. Professor Gill says of the carpus of the Hyracoidea, " carpal bones in two interlocking rows ; cuneiform extending inwards (and articulating with magnum) ; * * * unciform and lunar separated by the interposition of the cuneiform and magnum." Professor Flowerf gives a figure which justi- fies these statements, but neither the one nor the other agree with my ♦Arrangement of the families of Mammals prepared for the Smithsonian Institution. Miscellaneous Collections 230. Nov., 1872. t Osteology of the Mammalia, p. 2(>C ; fig. 92. Cope. J 440 [May l:'. specimens. In the manus of a llyrax, capensU (from Vcrreaiix, Paris), I find the following condition of the carpus. The bones of the two series are articulated consecutively, and not alternately ; they do not interlock, but inasmuch as the magnum is a little narrower than the lunar, the latter is just in contact (anteriorly) with the trapezoides (centrale) on the one^ side, and the unciform on the other. My specimen agrees with Cuvier's figure of Ilyrax capensis in all respects. It is prol)al)lo that Professor Fig. 2.— Left anterior foot of P/ienaeodu.s prima'vus, one-third natural six.L (original). Fig. 3. — iiight anterior foot of Hf/rax eapensis ; (from Cuvier). (S'e. scapnloid bone; Z. lunar; cm. cuneiform:/), pisiform; tz. trapezium; td. trapezoides: m. magn tim ; . u n ei form . Flower has figured some other species under that name, which besides it- peculiarities, is of smaller size than the H. eapensis (see Fig. 3). In April, 1875* I described the manus of (Jorypliodon (Bathmodon), showing that the lunar was supported below b}'' the magnum and by parts of the unciform. This carpus has the characters of that of Hymx eapensis, with the last named articulation more extensive. This was the first description of the carpus of the Amhlyx>oda. In February, 1870, f Pro- fessor Marsh described the carpus of Uintotherium {Dinoceras), and asserted that the bones "form interlocking series." lie however states that "the magnum is supported by the lunar and not at all by the scaphoid," a state of things which does not belong to the inter- locking carpus. The trapezoides does not join the lunar, but the unci- form does so, as in Coryphodon. Professor Marsh's figure as to thearticu- * Systematic Catalogue of the vertebrata of the Eocene of Xew Mexico, p. (U. S. Geol. Survey W. of 100th 3Ier.). t Amer. .Tournal Sci. Arts, xi, p. 1G7 ; pi. vi., fig. 2. Ftg . '2. Fig. 3. 441 [Cope. Liioiis of the lUiig-iiuin docs not :ii;-rc'e Avith liis description, as it makes that bone articulate witli the scaphoid. The second description is how- y?ver correct, and the carpus is identical with that ot Coryphodov. (Fig. 4.) In the Am criam Naturalist, June, 1882,* I have sliown that the carpus I ilie Condf/larfhm is essentially like tliat of the Iff/raroldca. (Fig. 2.) B iG. 4. Fia. 5. Fig. 4 — Maniis of Coryphodon (original). The cuneiform is imperfect. Fig. 5.— Left posterior foot of Elephas indicus ; (from Cuvier). ca. calcaneum ; . astragalus; navicular ; rw. cuboid: ec. ectocuneiform : mc. uiesocunei- orm. Tarsus. — In tlie tarsus of the Perissodactyla and Artiodactyla it is well understood that the cuboid extends inwards so as to articulate with the astragalus, giving the latter a double distal facet. It is also well known that the astragalus of the Proboscidea has but a single distal articulation, that with the navicular. It is, however, true that the cuboid is extended inwards, but that it articulates Avith the distal extremity of the navicular instead of that of tlie astragalus. It was shoAvn by Cuvier that the astragalus of the Hyracoidea articulates with tlie jiavlcular only, and that the cuboid is not extended inwards so as to overlaid the latter. In 187B Marshf stated that the astragalus of the Amblypoda articulates with both cuboid and navicu- lar. Finally I discovered in 1881 ,X that the astragalus of the Gondylarthra trticulates with tlie navicular only and that the cuboid articulates with * Page 522, t American Journal Science and Art, January, 1873. t American Xatnralist, ISHl, p. 1017. Cope] 442 i<^^ the calcaneiim only. lu tlie tarsus then tliere are four types of articula- FiG. 6. Fig. 7. Fig. 6. — Left posterior foot of Phenacodus primcevus, one-third natural siz* (original). Fig. 7.— Right posterior foot of Hyrax capetisis (from Cuvier). Ca. calca- neum ; a. astragalus; n navicnlar; cu. cuboid : ecc. eetocuneiforni : mc. meso- cuneiform ; enc. entocuneiform. Fig. 8. Fio. 8.— Posterior foot of Coryphodon (original) 1882.] 443 [Cope. tion, wliich are typified in the Condylarthra, tlie Probosciclca, the Amhly- poda and the Artiodactyla respectively. (Fia:s. 5-0.) Fig. 9. Fig. 10. Fig. 9.— Hind foot of PoSbrotherium labiatum (original). Fig. 10.— Fore leg and foot of Hyracothertum venticolum (original). Orders. — From the preceding considerations we derive tlie following definitions of the primary divisions of the Ungulata, which should be called orders. In the first place I find the diversity in the structure of the carpus to be greater in the relations of the magnum and scaphoides,.than in the relations between the uncifonn and the lunar. In other words the trapezoides and magnum are mon; variable in their proportions than is the unciform. This is directly due to the fact that the reduction of the inner two digits is more usual than the l eduction of the external two. I there- fore view the relations of these bones as more characteristic. In the tarsus the really variable bone is the cuboid. It is by its extension inwards PROC. AMER. PIITLOS. SOC. XX. 112. 3d. PRINTED NOVEMBER , 1882. Cope.j 44A [May 19, that tlie additional facet of the astragalus is produced. Its relations will therefore be considered rather than those of the astragalus in framing the following definitions : Order I. Scaphoides supported by trapezoides and not by magnum, which supports lunar. Cuboid articulating proximally with calcaneum only Tfixeopoda. Order II. Scaphoides supported b}'^ trapezoides, and not by magnum, which supports lunar. Cuboid extended inwards and articulating with the distal face of the navicular Proboscidea . Order III. Scaphoides supported by trapezoides and not by magnum, which with unciform, supports the lunar. Cuboid extended inwards and articulating with astragalus AmUypoda. Order IV. Scaphoides supported by magnum, which with the unciform also supports the lunar. Cuboid extended inwards so as to articulate with the astragalus Diplarthra . The sub-orders are defined as folloAvs : I. TAXEOPODA. There are two, perhaps three sub-orders of the Taxeopoda; the Hyracoidea, the CondylartJira, and perhaps the Toxodohtia/-^ The Toxodontia are how- ever not sufflciently known for Una! reference.f The sub-orders are de- fined as follows : A postglenoid process ; no fibular facet of calcaneum, but an interlocking articulation between fibula and astragalus ; ungual phalanges trun- cate Hyracoidea. A postglenoid process ; no fibular facets on either calcaneum or astragalus ; a third trochanter of the femur ; ungual phalanges acuminate Condylarthra. There are a good many other subordinate characters which distinguish the GondylartJira, which will be given in my forthcoming volume iv of the Hayden Survey, on the Tertiary Yertebrata of Western America. II. PROBOSCIDEA. There may be two sub-orders of this order, the Proboscidea and the Toxodontia. I do not know the Carpus of Toxodon, but if it does not difi"er more from that of the elephants than the tarsus does ; it is not entitled to subordinal distinction from the Proboscidea. The sub-order of Pro- boscidea is defined as follows : A fibular articulation of the calcaneum; no postglenoid process ; no third trochanter of femur Proboscidea. * See my remarks on Toxodon, Proceedings Amer. Thilosopb. Society, 18Si, p. 402. t The consideial)Ie resemblance between the dentition of Toxodon and Hyrax miistnotbe overlooked. ISS-.'.] U5 III. AMBLYPODA. Ihe sub-orders of this order, as I pointed out in 1873, aro two, detinod as follows : Superior incisor teeth ; no ali-sphcnoid canal ; a third trochanter of femur ; Pantodonta. No superior incisors, nor ali-sphcnoid canal, nor third trochanter of femur ; Dinocerata . The difference between the Proboscidea and the Amhlypoda consists chiefly in that the navicular of the latter is shortened externally so as to permit the cuboid to articulate with the astragalus. The cuboid has the same form in both. The peculiar character of the navicular gives the a«;tragalus a ditlerent form. IV. DIPLARTHRA. This order is called by some authors the Ungulata, but that name is also used in the larger sense in which it is here employed. This appears to be its legitimate application, as the name should, if possible, be used for hoofed Mammalia in general, as its meaning implies. The two well known sub- orders are the following : Astragalus truncate distally ; number of toes odd, the median one the largest , Peris sodactyla. Astragalus Avith a distal ginghnnus ; number of toes even, the median two largest Artiodactyla. Phylogeny. — The serial arrangement of the bones of the carpus and tarsus seen in the Taxeopoda, is probably the primitive one, and we may expect numerous accessions to that order on further exploration of the early Eocene epochs. The modification seen in the more modern orders of Perissodactyla and Artiodactyla, may be regarded as a rotation to the inner side, of the' bones of the second carpal row, on those of the first. This rotation is probably nearly coincident with the loss of the pollex, as it throws the Aveiglit one digit outwards, that is on the third and fourth digits, rendering the first functionally useless to a foot constructed solely for sustaining a weight in motion. The alternation of the tw^o rows of carpals clearly gives greater strength to the foot than their serial arrange- ment, and this may probably account for the survival of the type possess- ing it, and the extinction ot nearly all the species of the type which does not possess it. Here is applied again the principle first observed by Kow^alevsky in the proximal metapodial articulations. This author shows that the types in which the metapodials articulate with two carpal or tarsal bones, have survived, while those in which the articulation is made with a single carpal or tarsal have become extinct. The double articulation is, of course, mechanically the more secure against dislocation or fracture. As regards the inner part of the manus I know of no genus Avhich presents a type of carpus intermediate between that of the Taxeopoda and Cope.] 446 [May 19, Amblypoda on the one hand, and the PeHssodactyla and Artiodactyla on the other. Such will ho\A-ever probably be discovered. But the earliest PeHssodactyla, as for instance Hyracotherium, Hyrachyus and Triplopus, possess the carpus of the later forms, Rhinocerus and 7 apirus. The order Amhlypoda occupies an interesting position between the two groups, for while it has the carpus of the primitive type, it has the tarsus of the later orders. The bones of the tarsus alternate, thus showing a decided advance on the Taxeopoda. This order is then less primitive than the latter, although in the form of its astragalus it no doubt retains some primitive peculiarities which none of the known Taxeopoda possess. I refer to the absence of trochlea, a character which will yet be discovered in the Taxeo- poda, I have no doubt. The To/xeopoda approach remarkably near the Bunotheria, and the unguiculate and ungulate orders are brought into the closest approxima- tion in these representatives. In fact I know of nothing to distinguish the Gondylarthra from the Mesodonta, but the ungulate and unguiculate characters of the two divisions. In the Greodonta this distinction is reduced to very small proportions, since the claws of Mesonyx are almost hoofs. Some of the genera of the PeriptycMdcB present resemblances to the Greodonta in their dentition also. The facts already adduced throw much light on the genealog}' of the Ungulate Mammalia. The entire series has not yet been discovered, but we can with great probability supply the missing links. In 1874 I pointed* out the existence of a yet undiscovered type of Ungulata, which was an- cestral to the AmUypoda, Prolosciden , PeHssodactyla and Artiodactyla, in- dicating it by a star only in a genealogical table. This form was discov- ered in 1881, seven years later, in the Gondylarthra. It was not until laterf that I assumed that the Diplarthra are descendants of the Ainhlypoda, although not of either of the known orders, but of a theoretical division with bunodont teeth.:}: That such a group has existed is rendered ex- tremely probable in view of the existence of the bunodont Prohoscidea and Gondylarthra. That the Taxeopoda was the ancestor of this hypothetical group as well as of the Prohoscidea, is extremely probable. But here , again neither of the sub-orders of this group represent exactly the ances- ^ tors of the known Amblypoda, which have an especially primitive form of the astragalus not found in the former. In the absence of an ankle- joint, the Amblypoda are more primitive than any other division of the Ungulata, and their ancestors are not likely to have been more specialized than they. It is probable that a third sub-order of Taxeopoda has existed which had no trochlea of the astragalus, which I call provisionally by the name of Platyarthra. * Homologies and Origin of Teeth, etc., .Journal Academy Nat. Science, Philada., 1874, p. 20. t Report U. S. Geol. Survey W. of 100th Mer., p. 282, 1877. X This hypothetical sub-order is called in the appended scheme, AmhlypoOa Hyodonta. 1882.] 447 [Cope. The preceding paragraphs wore written in May of the present year. On my return home, September 1st, after an absence of three months, I find that various parts of the slvcleton of Periptychm'* liave reached my mu- seum. On examination, I find that the astragalus of that genus fulfils the anticipation above expressed. It is withoat trochlea, and nearly resembles that of Elephas. As it agrees nearly with that of Phenacodiis in other re- spects I only separate it as a family from the Phenacodontidcc. One other type remains to be discovered which shall connect the Periptychidoi and the hypothetical Hyodonta, and that is a Taxeopod without a head to the astragalus, — unless, indeed, the ''Hyodonta " should prove to have such a head. I think the latter the less probable hypothesis, and hence retain the term Platyarthra for the hypothetical Taxeopod without trochlea or head of the astragalus. These relations may be rendered clearer by the following diagram : Taxeopoda. Condylarthra. Platyarthra .| f / \ Hyrjjcoidea. \ Pkoboscidea. Amblypoda. Hyodonta. ft Pantodonta. I Dinocerata. DlPLARTHRA. / \ Perissodactyla. Artiodactyla. Third contribution to the History of the Vertehrata of the Permian formation of Texas. By E. D. Cope. {Bead before the American Philosophical Society, September 15, 1882.) Since the publication of my second contribution to this subject, ^ I have described four additional species. These are, in Bulletin of the U. S. Geological Survey of the Territories ;§ Pantylus cordatus and Dimetrodon semiradicatus ; in the American Naturalist, || Eryops reticulatus Za- » See American Naturalist, October, 1882 tt Hypothetical. JPaleontological Bulletin, No. 32, Proceedings American Philosophical So- ciety, 1880; the plates, 1881. §Vol. vi, 1881, p. 79. 111881, p. 1020. Oope.] 448 [Sept. l'), tracJiys apicalis. Tlie last two M'erc not included in my catalogue of the Permian Vcrtebrata published previously* in the same year. The present paper adds some important points to this remarkable fauna, and explains the hitherto obscure relations of several genera . DIADECTID^l The pelvis and sacrum of a species of this grouj) are preserved in my collection, and they indicate further peculiarities of this group. The sacrum consists of two vertebrae only, and is thoroughly united with the pelvis by its transverse processes. The latter are decurved on the inner side of the iliac bones, and the sutures which distinguish them from the latter and from each other, are not serrate. The inferior arch is robust, but very narrow anteroposteriorly. The acetabulum is entire in every respect, so that it is probable that both pubis and ischium are united undistinguishably in the arch. The pubis is perforated by the usual in- ternal femoral foramen. The posterior edge is grooved, and it might be suspected that this marks the articulation of an ischium. The anterior edge is however grooved in the same way, so that the appearance is rather the position of muscular insertion. The spines of the sacral vertebrae are distinct, and have the usual form seen in JDiadectes. The two sacral vertebnxi and the absence of obturator foramen, are characters of the suborder Pelycosauria in which the latter differs from the Dicynodontia. I am still inclined to question whether the extraordi- nary characters of the cranio- vertebral articulation I have described, jus- tify the separation of the DiadectidcB as a third sub-order of the Theromor- pha, which I have called the Cotylosauria,\ or wiiether they are not due to the loss of a loosel}' articulated basioccipital bone. EDAPHOSAURUS Cope, genus novum. Apparently allied to Pantylus. Temporal fossjxi not overroofed; surfaces of cranial bones not sculptured. Mandibular and maxillary teeth subequal. Posterior half of the mandibular ramus expanded inwards and supporting numerous closely arranged teeth. Pterygoid, or perhaps an internal ex- pansion of the malar bones, supporting a dense body of teeth, correspond- ing to those of the lower jaw. Teeth subconical. The single species of this genus in my possession shows the following characters of systematic importance. An arch extends from the parietal plane posteriorly and downwards to the external base of the quadrate. The specimen is not yet in a condition to show how much of this is parie- tal, and how much squamosal or opisthotic. The proximal half of the posterior part of this arch is a distinct element, perhaps a transverse pro- cess of the supraoccipital. A distinct element connects the basioccipital on each side with the quadrate. The articular extremity of the latter has ♦American Naturalist Feb., 1881. t American Naturalist, 1880. p. 304. Kss-_M '±'±i7 [Cope. a deep iintoropostorior concave eiuitrgination. There is a flat bone ex- tending from it anteriorly which is apparently pterygoid rather than quadratojugal. The tooth bearing portion terminates opposite the middle of the basisphenoid. The occipital condyle is undivided, and the basisphenoid presents the usual two divaricating protuberances to the basioccipital. Edaphosaukus pogonias, sp. no v. Represented by the foUowsng portions of a skull ; basis cranii with por- tion posterior to the middle of the parietal bone ; left maxillary with dental plate , left mandibular ramus entire ; various flat bones undetermined. There is also a body which may be the atlas with its arch somewhat dislo- cated. These pieces are in part covered with a thin layer of the red deposit of the Permian bed in which they occur. The facial plate of the os maxillare is sub vertical, so that the orbit is lateral. The latter is rather small. The malar bone is narrow, and is continuous with the dentigerous bone of the palate. The latter has a thickened posterior edge, which commences below the anterior part of the orbit, and extends posteriorly to the middle of the basisphenoid. Thence the border turns forwards. Its anterior edge is below the anterior border of the orbit, and the general form is a longitudinal oval. The maxillary teeth are somewhat weathered and obscured by a thin layer of matrix. The posterior ones are compressed-conic ; the premaxfllaries are four in number on one side, and are more nearly conic, and have incurved apices. The median premaxillary suture is, however, not clearly defined, so that the number of premaxillaries remains uncertain. The centre of the prob- able nostril measures one-third the distance from the premaxillary border to the anterior edge of the orbit. There are eight rows of (?) pterygoid teeth at the posterior fourth of the series. The teeth are subequal and obtuse, increasing a little anteriorly. The mandibular ramus is robust, and the external face slopes inwardly and downwards. The external border rises a little above a few of the posterior teeth, but it is injured at the posterior of the coronoid process, so that its existence cannot be ascertained. The border then descends and turns inwards to the articulation, which is condyloid at its internal extrem- ity. The inferior edge of the anterior part of the ramus becomes a median ridge below the condyloid region, and terminates in a short, compressed angular process^. The symphysis is not coossified, and is convex down- w/'ards and forwards. The inferior part is subhorizontal, and forms the edge of a transverse plate which is separated from the vertical part of the ramus by a deep groove. The' inner vertical face of the ramus is strongly convex, as is the corresponding edge of the symphyseal suture The apices of the teeth are worn, but they were probably conic, the posterior gradually smaller and more obtuse. The interior face of packed teeth begins at the posterior two-fifths of the external series, and expands in- Cope.] 450 [Sept. 15, wards posteriorly. It contains six longitudinal rows opposite the ante- penultimate dentary tooth. All the bony surfaces are smooth. Measurements. M. Length of mandibular ramus (straight) 162 " symphysis of do. (straight) 038 " external dental series 077 Width of ramus at dental pavement 040 " skull at ends of 00. quadrata 138 " extremity of O. quadratum 024 " occipital condyle 018 Length of superior dental pavement 065 Width of basisphenoid posteriorly 029 The supposed axis vertebra is longer than wide, and the centrum is deeply excavated posteriorly. Anteriorly it appears to have lost a piece — the centrum of the atlas, which, while fitting it closel}^ was not co-ossified with it. There is a flat horizontal convex ala in the place of a diapophysis, and an obtuse median hypapophysial angle. The neural spine is compressed, except posteriorly, where it is transversely expanded, terminating above in a short obtusely accuminate apex. From this apex an obtuse rib passes down the median line, and disappears above the neural arch, where the spine is somewhat narrower. The postzygapophyses are well developed and look downward. Measurements of axis. M. Length of centrum below 020 Width, including diapophyses 035 Elevation of spine from postzygpophysis 038 Width of do., posteriorly 020 Remarks. — This interesting form is probably allied to Pantylus, which I have hitherto regarded as a Batracliian. The two genera may be placed in a special family of the Pelycosauria, to be called the Edapliosauridos. This family will be distinguished from the Glepsydropidm by the presence of more than one series of teeth on parts of the jaws. It is possible that Helodectes must be placed in it. ECTOCYNODON Cope. Paleontological Bulletin No. 29, p. 508. A species now before me resembles in generic characters the type of this genus, E. ordinatus. That species was described as having the canine tooth near the middle of the maxillary bone, while in the present one it is near the anterior part of it, as in some other genera. In the typical species, as in the species to be described, the cranial bones are sculptured, and the temporal fossoe are overroofed. The sculptured surface as well as the canine teeth distinguish Ectocynodon from Pariotichus Cope and Proco- lophon Owen, whicli^gcnera are otherwise related. I 1882.] [Coi)e. ECTOCYNODON AGUTI, Sp. nOV. This reptile is much larger than the Pariotichus brachyops, and the ante- ; rior part of the cranium has a different form. The general shape of tht; i head is much like that of a rodent mammal of the genus Daayprocta. It is rather wide at the temporal regions, flat above, and narrowed and com- pressed anterior to the orbits. The muzzle is narrowed and obtuse, and I the nostrils are terminal, and are lateral and a little anterior in direction. The maxillary alveolar edge is nearly straight, but the premaxillary edge, beginning below the posterior border of the uares, descends forward at an angle of 450, Viewed from the front, the premaxillary border is a festoon, strongly convex downwards, and below the anterior part of the nostril. The suture separating the premaxillaries is distinct. The orbits are of moderate size, as in an aguti, and invade the superior frontal plane in a slight degree. The frontoparietal fontanelle is rather large. The mandible is robust, and presents a short angle. It closes up behind the premaxillary lobate edge. Its teeth are concealed in the specimen. The maxillary teeth increase rapidly in size forwards. The premaxillaries commence smaller next the maxillaries, and increase in size to the first, which is a little larger than the anterior maxillary. The crowns are weathered away. The sculpture on the maxillary and malar bones con- sists of closely placed shallow fossae. On the posterior part of the frontals there are strong ridges radiating posteriorlyi and situated close together. Measurements. M. Length of skull to end of angle of lower jaAv 090 " frontoparietal fontanelle 056 orbit, above 026 " ramus mandibuli 082 Depth of skull at orbit , .033 " ramus " 019 Width of skull posteriorly 068 " " between orbits 017 " " between external nares 0105 Diameter of first premaxillary tooth 003 " second maxillary tooth 003 Six fossae of the malar bone 005 Seven grooves of the frontal bone 005 This species is much larger than the'.Ectocynodou ordinatus Cope, and the canine tooth has a more anterior position. Discovered by W. F. Cummins. DIPLOCAULUS Cope. Paleontological Bulletin No. 20. p. 187, Nov. 21st, 1877. Proceedings American Philos. Society, 1877, p. 187. This genus was characterized by me at the places cited, as follows: "Vertebral centra elongate, contracted medially, and perforated by the PROC. AMER. PHILOS. SOC. XX. 112. 3e. PRINTED NOVEMP.ER , 1882. < 'ope. J 452 ISept. 1"), foramen chordae dorsalis, coossified with the neural arch, and supporting transverse processes. Two rib articulations, one below the other, gen- erally both at the extremities of processes, but the inferior sometimes sessile. No neural spine nor diapophysis ; the zygapophysis normal and well developed." This diagnosis was derived from the vertebne of a single species from the Clepsydrops shale of Illinois, the D. salamandroides, and since that description was written, no additional specimens have come under my ob- servation. In the Catalogue of the Vertebrata of the Permian I placed the genus as the type of a family, the DiplocauUdm, among the Pelycosauria, I am now, however, through the energy of Mr. "W. F. Cummins, in pos- session of specimens of a number of individuals of a second species of Diplocaulus, found by him in the Permian beds of Texas. From them I derive that the genus and family must be referred to the Stegocephalous Batrachia. It is, however, exceptional among these in the fauna of which it is a member, in not belonging either to the Rhachitomi* or to the Em- bolomera, since the vertebral centra are not segmented, nor are the inter- centra present in any form. Under these definitions it must be referred to the suborder which includes Oestoceplialus, Ceraterpeton, etc., for which I have adopted Dawson's name Microsauria. The division includes genera with simple amphicoelous vertebral centra, and teeth without inflections of the dentine. The following characters must be added to Diplocaulus : Vertebrae with a more or less perfect zygosphen articulation ; centra shorter in the anterior than in the median part of the column ; axis and atlas solidly united by a long zygosphen, which is not roofed over by the zygantrum. Neural arch continued as a short tube into the foramen magnum. Atlas unsegmented, and, like the axis, without free hypapophy- sis. Cervical vertebrae not distinguished from dorsals, and with two- headed ribs. Orbit separated from the maxillary bone by the union of the lachrymal and malar. Either the malar, or more probably the quadratojugal, extends much posterior to the quadrate bone. It is bounded above by the squa- mosal, which extends anteriorly to the distinct postfrontal, thus covering over the temporal fossa. Posteriorly it extends into a long, free process, like the operculum of Polyodon ossified. This horn does not appear to consist of the epiotic as appears to be the c;ise in Oeraterpteon. The quadrate bone is extended very obliquely forwards and its extremity is divided into an hourglass-shaped condyle. In other words the condyle consists of two cones with apices continuous. The internal cone is the smaller, and its base is overlapped from before by a flat bone, probably the pterygoid. The cotyli of the mandible correspond. Mandible without angle ; symphysis short. The teeth are of about equal size, and are rather slender and with conical apex. Their surface is not inflected at any point. The superior series is^ * American Natunilist, 1882, p. :J.J4. mi.] 453 [Cope. double!, forming two lines between which the mandibular teeth close. This superior series stands near the external edge of the vomer, palatine and pterygoid bones successively. I have not been able to find any larger teeth in the jaws in this genus. Some fragments mingled with those here described, display such teeth, but I think they pertain to a species of another genus. I know nothing of the limbs of this genus. DirLOCAULUS MAGNICORNIS, Sp. nOV. The species is indicated by fragments of a number of crania, and por- tions of several vertebral columns. These were collected at two different localities by Mr. W. F. Cummins. The skull is very peculiar in the great extent of the parts posterior to the orbits as compared with the portion anterior to them. The posterior border not being complete, the proportions cannot be exactly given, but the part anterior to the orbits is two-thirds the length of the part extend- ing from their posterior border to near the base of the lateral horn, and one-fifth the distance from the orbit to the extremity of the horn. The part of the border of the orbit preserved indicates that the latter is of fair size. It is separated trom the maxillary border by at least its own diame- ter. The external nares are peculiarly situated. They are nearer the orbit than the end of the muzzle, and are close to the maxillary border, being separated from the mouth by a narrow strip of bone only. They are round, open nearly laterally, and are removed from the edge of the orbits by the diameter of the latter. The malar or quadratojugal bone is protuberant at the canthus oris and projects laterally beyond the mandible at its posterior part. It also pro- jects beyond the extremity of the quadrate bone. This border is continued as that of the external base of the horn, but the portion wiiich belongs to this element is soon distinguished from the superior element (squamosal) which composes the horn, by a groove. This groove is decurved, and bounds the apex of the element, which is a decurved, low tuberosity t The horn is produced backwards in a horizontal plane, forming a long fla triangle which contracts gradually with straight sides. The apex is nar- rowed, obtuse, and a little incurved. Near and at the extremity the horn is flat above and convex below. The mandibular quadrate cotylus consists of two fossse, which together form an approximate figure oo , of which the internal fossa is the smaller, and opens internally. The external one is nearly transverse. The supe- rior border of the ramus posteriorly is straight. The greater part of the superior aspect is occupied by a huge fossa which opens upwards. It is uncertain whether the horns meet at an entering angle on the middle line posteriorly or not, but the width of the base of the horn indi- cates that such is the case. The extremity of the muzzle is depressed, and is broadly rounded. The external surface of the skull is sculptured in the form of fossaj so distributed that the narrow ridges separating them do not form straight Cope.] 454 rsept. 15, lines, except in a few places on the superior face of the horn. This sculp- ture is. strongly impressed, and is of medium coarseness. It extends on the inferior face of the quadratojugal (?) posterior to the quadrate, and on the inferior side of the horn at the edges. It is most extended below from the interior edge, and for the terminal inch of the horn, is as well marked as on the superior face. Elsewhere the sculpture of the inferior side passes into punctie before disappearing. A groove marks the superior boundary of the maxillary bone, which divides when it reaqjies the superior surface. One branch descends behind the nostril, the other passes transversely across the lachrymal bone and shallows out before reaching the middle line of the muzzle. The mandible is even rougher than the superior sur- faces, and has a longitudinal groove below the dental line, to near the symphysis, where it runs out on the alveolar edge. The internal and ex ternal sides of the mandible posteriorly, are smooth. On the malar and other facial bones there are four fossa? in 9 or 10 mm. The atlas is peculiarly flattened above, the neural arch being a tube, without neural spine. Its anterior tubular prolongation is not long, and is deeply notched below. The condyloid fossa? are widely spread trans- versely and nearly flat, except that their surface is carried forwards on the neural tube. They are well separated below. There is a strong hypa- popliysial keel, which diminishes and runs out anteriorly. There are pre- zygapopliysial facets, but the postzygapophyses exist. Their superior edge is however carried posteriorly to form the sides of the huge embraciiio: zygantrum. These side processes, which I will call zygantropophyses, extend as far posteriorly as above the posterior end of the centrum of the axis, embracing almost the whole of the neural arch. There is another short median superior process, which notches the extremity of the zygos- plien. The side of the atlas between the postzj'gapopliysis and the con- dyloid facet is wrinkled, and the inferior face finely punctate. In the axis, the hypopophysis is a large ridge with a horizontal truncate edge. The costal heads of the diapophysis arc not split to the base of the latter and the superior is the more robust (extremities broken off). Cen- trum concave posteriorly, and on each side of hypopophysis with reticulate surface. A short zygantropophysis ; zygantrum not large. Exposed summit of zygosphen (nearl}^ equal neural arch) without neural spine. In both the axis and other cervical vertebrse, the superior diapophysis is con- nected with the zygapopliyses fore and aft, in accord with the shortness of the centra. In the more posterior vertebra? they become separated on account of the increasing length of the centrum. The third vertebra is like the axis, except in having a keel-shaped neural spine, and a short obtuse zygosphen continued from its base ante- riorly. With increasing length of centrum the diapophysis becomes longer,* and the hypapophysial ridge becomes wider, and coextensive with the in- ferior face of the centrum. It is separated by an angle from the sides iji the longer vertebne ; in those of intermediate length, the inferior face i- 1882.] 455 [Cope. convex. All of tliem retain the delicate lines and punctje of the inferior surface. The neural spine on the more elongate vertebne is a rather ele- vated keel, with horizontal superior edge. Its posterior extremity forms a wedge-like zygosphen. The zygantrum is a dee]) V-shaped cavity, open- ing posteriorh^ and not roofed over at any point unless for a small part of its fundus. Tlie zj^gapophyses are well spread, and have horizontal faces. Each of the columns of the diapopliysis sends a ridge forwards, which en- close a groove between them. Measure))} ciitst of m')'teh)'a>. M. Length of atlas below Olo " " at zygantropophyses 029 Expanse " " condyloid facets 034 ' ' of centrum atlas behind 0145 Depth of atlas at middle 010 Length of axis below 015 " " at zygantropophyses 016 Width of zygosphen above Oil Expanse of postzygapopliyses 024 Width of centrum posteriorly 012 Depth " " Length of centrum of another (No. Ill) 018 (No. IV) 023 Expanse of postzygapopliyses of do 018 Length of centrum of No. V 022 r^. ^ ^ TT * • 1 (Vertical 013 Diameters centrum V anteriorly } (.transverse 012 Expanse prezygapopliyses 021 Elevation of neural spine from centrum Oil r anteroposterior 023 ^Diameters centrum No. VI < vertical.* Oil ' transverse 013 The vertebne of this species are very much larger than those of the J). salama)id}'oides, and the diapophyses do not originate so low down on the centrum. Otherwise they are much alike. The cranium of the Illinois species is yet undetermined. The I). ))ia(jnicor)iis was discovered l)y W. F. Cummins. ACHELOMA. Cope, genus novum. Order Rhachitomi ; family Eryopidas,* differing from E)'yops in the ab- sence of notch of the posterior border of the skull between the epiotic and ({uadrate or squamosal bones, and in the absence of condyles of the hu- merus. Mandible without angular process. Teeth ot the jaws subequal, ratlier larger anteriorly ; some large ones on the , ( ^ ^ ^ I vertical 006 Length of rib of <5tli vertebra in advance of the vertebra measured 038 Width of rib distally 027 459 [Cope. ScdpuhLV arch. M. Length of scapuUi on anterior face 069 Width do. at antero-internal distal angle, transversely. .032 " of coracoid and epicoracoid at glenoid cavity, from edge of scapula 023 Length of epicoracoid and coracoid 037 " humerus 0C4 Width of shaft at middle OIG Diameters proximal end \ ^^"^ ^'^^ i short at middle 010 Diameters distal end | ^^^S • • • 039 I short at middle 010 Length ungual phalange 004 " second " 0075 " first " 0135 Width do. I P^'^^^"^^^^^^ Idistally 008 This species was discovered by Mr. AV. F. Cummins, to whom I dedi- cate it with much pleasure. ANISODEXIS Cope, genus novum. Class Batrachia ; order Rhachitomi ; family Eryopidte. Teeth on pre- maxillary, maxillary, and dentary bones^ of unequal lengths, some very large, others very small. Dentinal inflections straight, nearly reaching the pulp cavity. Cranial surfaces sculptured. Tliis genus diflfers from all the others of the Eryopidce, in the great and abrupt inequality of the teeth of the external series of the mouth, resem- bling in this respect some of the Saurians of this deposit, rather than the batrachia. Whether it possesses long palatine or pterygoid teeth such as most of the latter exhibit, is not rendered clear by the specimens, but ap' pearances indicate the presence of one near the anterior part of the maxil" lary. Mandibular series simple. Anisodexis iMBRiCAmus Cope, sp. nov. Founded on numerous fragments of the skull with jaws, and a verte- bral arch and spine found in connection with the remains of the Diplocau- JUS magnicorrda. These pieces indicate a larger species than the latter, and are nearly equal to the Eryops megaceph(dus. The jaws are not pre- served entire, but portions from different parts of the length display the dental characters. The sculpture of such parts of the superior surface of the skull is a coarse reticulation, coarser than in any other species known to me. Near the edges, some of the bones become smoother, and the ridges flatten into overlapping lamina3. The entire sculpture of the dentary bone is of this imbricate character, the apparent overlapping being from before back- TROC. AMER. PIIILOS. SOC. XX. 112. 3f. PRINTED NOVEMBER , 1882, Cope.] 460 [Sept. 15,. ' wards, and below upwards. This is totally different from what is observed in the other known species of Eryopidce, Trimerorliachidm, and Diplocau- lidce. The teeth are round in section, but become lenticular near the apex, developing low cutting edges. The basal grooves are fine, but distinct, and extend half way to the apex, or farther. One large, and one medium sized teeth stand on each dentary bone near the symphysis, and there are two similar ones at a point further back on the same bone. Near the an- terior part of the maxillary, below the ? nostrils, is a huge tooth, with a graduated series of small teeth posterior to it, and a very small one ante- rior to it. The neural arch of a vertebra has a well developed vertical spine. Its neurapophysis rested in an oval fossa of the centrum which probably was divided into pleurocentra. The prezygapophyses are very small, and look directly upwards. The postzygapophyses are much larger, and look obliquely outwards and backwards. The spine is not expanded at the summit, and is granular, as though it was protected by a cartilaginous cap. Its section is anteroposterioriy lenticular, M'ith acute edge (aitgle) posteriorly, and a very narrow truncate edge anteriorly. The latter is bounded below just above the root of the neural arch by two little fossa?. The posterior keel is bounded below by a corresponding single fossa. The posterior acute edge of the spine is dentate, and the surface on each side of it, is beveled with rabbeted surfaces as though for a coarse squamosal suture. But the appearance of suture is fallacious, and is simply due to contraction of the transverse diameter of the spine. The neurapophy sis- is much narrower anteroposterioriy than the neural spine. Measurements. M. Depth of maxillary bone at large anterior tooth 037 dentary at symphysis 025 near middle 021 Width " " " ' 015 Diameter of base of large maxillary tooth 010 " " small maxillary tooth 0035 Length " " " " 008 of large mandibular tooth near symphysis OIG Diameter of base of crown of do OOG Elevation of neural arch 037 C vertical 029 Diameters neural spine { . ^„ ( anteroposterior 019 I ^P^^ I transverse 012 Width neurapophysis anteroposterioriy. . 010 From Mr. W. F- Cummins' collections. I had thought at one time that this species might be referable to the ge- nus Leptophractus of the Coal Measures. No trace of the vertebrae of the Rhachitomous order has yet been found in that formation in this country, . nor have any of the Coal treasure genera of Batrachia yet been found ia 461 [Cope. tlie Permian of the United States.* It is not improbable that such occur- rence of genera may yet be substantiated, but the identification of an or- tler hitherto unknown in a formation, on uncertain characters, is not a safe proceeding. The vertebrte oi Leptophractus although not certainly known, are supposed to be of the Labyrinthodont type. Tlie teeth are much more compressed a'nd trenchant than in tlie present species, nor do there appear to be any long ones near the symphysis mandibuli. I consider the ques^ tion of reference to Lcptophractus to be still an open one. The family Eryopidm, thougli abundant in individuals, is not represented by many species. They are presumably as follows : Anisodexis wibricarius Cope. Acheloma cmnminsi Cope. Eryops reiiculatus Cope. Fry ops ferrkolus Cope (Parioxys oUm). Eryops megacephalus Cope. Actinodoii frossd rdi Gaudry . Zatrachys serratus Cope. Zatr<(chys apicalis Cope. But the occipital condyles are unknown in Acheloma and Zatrachys. I may add here that through the courtesy of Messrs. Scott and Osborne^ I have seen, in the Museum of Princeton College, vertebrae of some species of the Rhachitomi from Saarbriicken, along with Archegosaur us, with entire centra, from the same locality. Synopsis of the Vertehrata of the Paerco Eocene epoch. By E. D. Cope^ {Read before the American Philosophical Society, October 20, 1882.) REPTILIA. CROCODILIA. Crocodilus sp. Crocodilus sp. Crocodilus sp. TESTUDINATA. Plastomenus ? communis Cope. Dermatemys sp. Compsemys sp. Emys sp, * Peplorhina arctata Cope, from the Illinois Porrnian is not a PeplorJiina, but I Theromorph Saurian, Cope.] 462 [Oct. 20, CHORISTODERA. Champsosaurus amtralis Cope, American Naturalist, 1881, p. 090. Champsosaurus j^ueixensis Cope, .Proceedings American Philosophical Society, 1881, p. 195. Champsosaurus saponensis Cope, Loc. cit. 1881, p. 19G. MAMMALIA. MARSUPIALIA. Ptilodm mediaevus Cope, American Naturalist, 1881, p. 922. JPtilodm trovessartianus Cope, loc. cit. 1882, p. 686. Catopsalis foUatus Cope, loc. cit. 1882, p. 416. Catopsalis pollux Cope, loc. cit.. 1882, p. 685. Polymastodon tadensis Cope, loc. cit. 1882, p. 684. BUNOTHERIA. Taexiodonta. -Hemigaiius vuUuosus Cope, loc. cit. 1882, p. 831. Tce?iiolabis scalper Cope, loc. cit. 1882, p. 604. TiLLODONTA. Psittacotherium multifragum Cope, 1. c, 1882 p. 156. Psittacotherimn aspasim Cope, Proceed. Anier. Philosophical Society, 1882, p. 192, (1882). Mesodonta. Pelycodus 2^eUidens Cope, Proceeds. Amer. Philos. Soc. 1881, (1882) p. 151. Lipodeotes peUidens Cope, American Naturalist. 1881, p. 1019. HyopsodiLS acolytus Cope, sp. nov. This the least species of the genus, is also the oldest, being derived from the Puerco horizon. Parts of two individuals furnish the characters ot the inferior and superior true molars, and the fourth superior premolars. The species differs from those hitherto described in other characters than the minute size. One of these is the absence of posterior interior cusp, the heels of the first and second true inferior molars being bounded by a ridge only at this point, as in most of the species o^ Pelycodus. The last inferior molar is not smaller than the second, nor longer. The anterior cusps of all the molars a.re robust, so that on the first and second true molars they are separated by a shallow notch only. There is a rudiment of the anterior inner cusp on the first true molar but none on the second and third. The posterior external is obtuse and has a triangular section on all the molars ; a crest is continued from the heel of the third molar on the inner side of the crown half way to the anterior inner cusp. 1(IS2.J 463 [Cope. The Mkrosyopm sjyicrittiniK (Hirers from this species in its smaller size (true molars .008) and in the presence of posterior internal cusps of the true molars. The Ili/opnodm acolytus was found by Mr. D. Baldwin, in New Mexico. CllEODONTA. SdvcotJivaustes antiquus Cope, Proceeds. Amer. Philos. Soc. 1881 (1882), p. 193. Dif^tidcus carmfex Cope, Amer. Natst. Oct. 1882 (Sept.), p. 834. Dimacus navajociics Cope, loc. cit. 1881, p. 1019. Mesonyx navajovius Cope, Proceeds. Amer. Philos. Society, 1881, p. 484. IVi'isodon quivirensis Cope Amer. Nat. 1881, p. 667. lYiisodon heilprmianus Cope, Proceeds. Amer, Philos. Soc. 1881 (1882), p. 193. DeltatJierium fwidaminis Cope, Amer. Nat. 1881, p. 237 ; 1881, p. 337. Lipodectes penetrans, loc. cit. 1881, p. 1019. Deltatherium, baldwini Cope. This Creodont is known only from a portion of a right mandibular ramus which supports the two last premolars, and the first true molar with part of the second. It differs from the D. fundaminis in its materially smaller size, and in the forms of the teeth. The first true molar is a more robust tooth, and the basis of the posterior or heel crest is more rounded, and less angulate. The anterior inner cusp projects less anteriorly. The fourth premolar has a distinct anterior basal lobe which is w^anting in the D. fundaminis. Its heel is short and wide, and the posterior face of the principal cusp is flat, and there is a rudiment of an internal tubercle on its side. The second premolar is elevated and acute, has no anterior basal lobe, and has a very short wide heel, enamel slightly roughened. The animal was rather aged. Measurements. M. Length of P-m. ii and iii and M. ii Diameters m. i | ^^^^^^'^P^^^^^'^^^ ' • transverse , Elevation of crown of P-m. iii Depth of mandible at M. i , From the Puerco beds of N. W. New Mexico. Dedicated to Mr. D. Baldwin, the discoverer of the Mammalian Fauna of the Puerco beds, which is one of the most important in the history of American Palkion- tology. Deltatherium interruptum Cope. The smallest species of Deltatherium is, like the D. baldwini, only repre- sented by the anterior part of a riglit mandibular ramus, which supports the last premolar and the first true molar, with the bases of the other pro- .0160 .0058 .0040 .0052 .0180 Cope.] 4(34 [Oct. 20, molars and part of the canine. The canine is small and the first premolar in accordance with the generic character, is wanting. The second pre- molar is two-rooted,, The fourth has an elevated principal cusp, and a narrow heel on the inner side of the posterior base ; anterior base injured. The first true molar has very little sectorial character, and resembles the corresponding tooth of a Pelycodus. It differs entirely from that of the D. fundaminis in the possession of a well marked posterior internal cusp, which is connected by a ridge with the large internal lateral cusp of the heel. The anterior cusps of opposite sides sub-equal. A weak external basal cingulum on the anterior half of the crown ; no internal cingulum. Enamel of the tooth wrinkled. Measurements. M. Length of premolar series 0140 Elevation of P-m. iv 0040 DiametersofM.il ^^^^^^ posterior 0055 ^ transverse 0042 Depth of Ramus at P-m. 1. 0090 M 0113 On comparison with the D. fundaminis, the first molar tooth has the same dimensions, but the premolars are considerably smaller. The ramus is also shallower. Found by Mr. Baldwin in the Puerco beds of North- west New Mexico. Didymictis haydenianus, sp. nov. This creodont is represented by parts of the maxillary and mandibular bones of the left side, the former supporting the four, and the latter supporting the three last molars. The arrangement of the superior molars is much as in D. protenus, the fourth premolar being a true sectorial. The third premolar has no internal lobe, although the section of the base of the crown is narrowly triangular. It has anterior and posterior basal lobes, and a posterior lobe on the cutting edge. In the sectorial the median lobe is a good deal more produced than the posterior, though the two form together the usual blade. The anterior basal lobe is distinct ; and the internal is larger and is conic. The first true molar has the anterior external base of the crown produced. Its two external cusps are conic and distinct. The internal part of the crown is rounded and supports a conic internal tubercle, which is separated from the external cones by two small concentric tubercles. The second true molar is considerably smaller, and is transverse, its external border being very oblique. It has an acute internal lobe. The character of the species is well-marked in the inferior true molars. The first has thq form seen in other species of Didymictis. The heel is large, and with a median basin between lateral cutting edges. The two anterior inner cusps are of equal elevation and are near together ; the external is much larger. The last molar is elongate, but reduced in size. Its anterior three cusps, rudimental in other species, are here elevated, forming the triangular mass seen in the first true molar. Tlieyare not so 1882.J [Cope. elevated, however, as in that tooth, and thus not so much developed as in Oxymia, Stypolophm, etc. The fourth premolar has a median cutting edge on the short heel. Measurements. M. Length last four superior molars 023 " P-m. iii 0065 " iv 0085 Width " " 0050 i anteroposterior 0055 transverse 0088 oblique external 0072 ^. r anteroposterior 0027 Diameters M. , /^^p-p- I transverse 0055 Diameters inferior M. I r anteroposterior 007 t transverse .005 , . . T,r TT r anteroposterior .0055 Diameters inferior M. II ^ , \ transverse 003 Depth of ramus at M. II, (squeezed) 010 The peculiar characters of the last inferior molar distinguish this species from its congeners. The last superior molar is relatively smaller than in the D. protenus. In size this species is superior to the D. dawkinsitmus, and is smaller than the B. leptomylus. It is dedicated to the distinguished geologist Dr. F. V. Hayden. New Mexico, D. Baldwin. TAXEOPODA. CONDYLARTHRA. Periptychidse. Periptychus rliabdodon Cope. Catathlmis rJidbdodon, American Natur- alist, 1881, 829. Periptychus carinidens Cope, loc. cit. 1881, p. 337. Periptyclius ditrigonus Cope, sp. nov. This rare species is known from a right mandibular ramus, which ex- hibits part of the symphyseal suture, with the alveoli of the molar teeth, except the first. The only well preserved crown is that of the second true molar. The second true molar presents very peculiar characters, and the man- dibular ramus is shallower and thicker than in the two other species of Periptychus. The former has a wide external cingulum which is not present in the other species, and there are only six cusps instead of seven. These are peculiarly arranged. The anterior three are much as in P. rhabdodon, the anterior being not quite so far in- ternal as the posterior inner, close to it, and as large as the anterior external. The posterior three, are a posterior inner and posterior median as in P. rhabdodon, and a peculiarly placed posterior external. This is not Cope.] 4(5G [Oct. 20, opposite the posterior inner, but is anterior to such a position and inter mediate between the latter point, and the one occupied by the median tubercle in P. rhaModon. It is as large as the anterior external tubercle. All these tubercles are conical, and not connected by angles or ridges* The posterior external cusp leaves the cingulum wide posteriorly, and its edge develops some small tubercles. There are also some small tubercles at other points on the edge of the crown, but no other cingula. The enamel is not regularly ridged as in P. rhaModon, but has a rather coarse obsolete wrinkling. Measurements. M. Length from P-m. ii to M ii inclusive 052 Diameters of M.ii ( anteroposterior Oil t transverse 010 Depth of ramus at M. ii 022 Width of " 016 Depth of " " P-m. ii 019 From the Puerco formation of New Mexico, D. Baldwin, discoverer. Ha/ploconus Uneatus Cope, Amer. Nat. 1882, p. 417. Haploconus angustus Cope, Loc. cit. 1882, p. 418. Miodcmus angust'is Cope, loc. cit. 1881, p. 831. Haploconus xipliodon, sp. nov. This species is represented by a mandibular ramus, and perhaps by three rami. The one on which the species rests contains five molars, the middle one of the series broken, so that its form cannot be positively ascertained. It is probable that it is the first true molar, so that the animal exhibits the last true molar not entirely protru;led, and is therefore nearly adult, but there are some reasons for suspecting it to be j^oung. Thus the last inferior molar does not exhibit more of a heel than the second usually does, and the third supposed premolar is smaller than that tooth is in the other spe- cies, having nearlj^ the proportions of the second premolar. The teeth present may then be supposed to be the molars from the second to the sixth inclusive. But opposed to this view is the fact that the sup- posed third premolar has more the structure of that tooth in details, than that of the second, and the specimens accompanying, which have the tem- porary dentition apparently of the same species, present premolar teeth of a very different character. In any case the present specimen represents a third species of the genus, and I describe it at present as an adult. The third premolar has a simple compressed crown, about as high as the length of its base, and without anterior basal tubercle. It has a narrow triangular posterior face which is concave, and truncated by a cingulum below ; no heel proper, nor lateral cingula. The fourth premolar is an elongate tooth consisting of a compressed principal median lobe, an ante- rior lobe connected with it, and a heel. The latter has elevated posterior • and interior borders. A rudiment of an exterior border is seen in a narrow -J 467 [Cope. riili-c on the external side of the posterior f\ice of the principal lobe of the tootli. The sides of the premolars present rather distinct ridges, as in Peripty- chiis carinidem. The second true molar has two anterior and three poste- rior tubercles ; the latter close together, pointed and of about equal size. Of the anterior tubercles, the external is much the larger and more ele- vated. It is compressed and has a curved subacute anterior edge, which .extends mucli in front of the internal tubercle. There is no anterior inner tubercle, nor are there any cingula. The enamel of the sides of the crown presents a few vertical ridges. The last inferior molar only differs from the second, in the greater size of the median posterior lobe, which is never- theless smaller than in the two otlier species of Hdiiloconus. There is a mental foramen below the posterior edge of the second in- ferior premolar. Measurements. M. Length of last five inferior molars 0250 " third premolar 0050 fourth premolar 0066 second true molar 0050^ Width of second true molar 0032 Length of third true molar 0050 Depth of ramus at P-m. yi 0095 M. iii 0130 The two rami Avith the temporary premolars, exhibit the last true molar enclosed in the j aw. The third and fourth premolars are much like the fourth premolar of the specimen above described, but the fourth is a little more robust than that of the latter, which is very much like the third of the de- ciduous series. The space occupied by the supposed first premolar of the type specimen is too short for the fourth premolar of the deciduous series, otherwise it might be supposed to have occupied that position. The two .true molars resemble those of the type, excepting that the last one does not extend so far into the base of the coronoid process, and is in accord- ance with the position as number two in the series. The specimens were procured by Mr. D. Baldwin in the Puerco beds of New Mexico. Ilaploconm entoconus Cope, loc. cit. 1882, p. 686. Aiiisonclius coniferus Cope, loc. cit. 1882, October (September), p. 832. Aniss>ncJius f/illianus Cope. Haplocoiius gilUanus Cope, loc. cit., 1882, 1). 686. Anisonchus sectorius Cope, Proc. Amer. Philos. Soc. 1881, p. 488, Mio- daenus sectorius, Amer. Nat. 1881, p. 831. Hemithlceus kowalevskianus Cope, Amer. Nat. 1882, p. 832. TIemithlcBUS opisthacus Cope. Mioclmnus opisthams, 1 c. 1882, p. 833. Conoryctes comma Cope American Naturalist, 1881, p. 829. I'ROC. AMER. PHILOS. SOC. XX. 112. 3g. trtnted novi^mber , 1882. Cope.] 468 [Oct. 2o, Conoryctes crassicuspis Cope. The posterior part of a mandibular ramus supporting the last two molar teeth indicates a second and larger species of the genus. The ramus is one-half deeper than that of the C. comma, and the second true molar is much larger than in that species. The last true molar is much smaller than the the penultimate, and consists of three anterior cusps and a longer heel. The former are obtuse, the external the longer, the internal equal, the anterior on the inner edge of the crown. The heel sustains a low conic tubercle. From the Puerco beds of N. W. New Mexico. Phenacodontidse. Protogonia, plicifera Cope, Amer, Nat. 1882, Oct. (Sept.), p. 83.3. ProtogoniasvhquadrataCoY)^, Proceedings Amer. Philos. Soc. 1881, p. 492. Phenacodus puercensis Cope, Proc. Amer. Philos. Soc. 1881, p. 492. PJienacodus zuniensis Cope, loc. cit. p. 492 ; loc. cit. 1881 (1882), p. 180. Pantolamhda hathmodon Cope, Amer. Nat. 1882, p. 418. Miodmnus turgidus Cope, Amer. Nat. 1881, p. 830. Mioclcenus mirdmus, sp. nov. This is one of the least mammalia of the Puerco fauna, exceeding by a little the Hyopsodus acolytus. It is represented by parts of two mandibles, which display all the true molars. As there are no premolars preserved, its reference to the genus Miodcemis is provisional only, but its true molars have the peculiar characteristics of those of the M. turgidus. The two anterior cusps of the true molars are higher than the heel, and they are united together to a point above the level of the heel. The sec- tion of both those of the M. ii is round ; that of the external one of the first is cresentic ; of the inner cusp, round. The heel is wide, and supports a cusp at the posterior external angle. It is bounded posteriorly, and on the inner side by a raised ridge, which gives with the cusp, on wearing a comma-shaped surface. A transverse ridge closely appressed to the ante- rior cusps connects them anteriorly. In one of the specimens there is a cingulum on the external side of the second inferior molar ; on the other specimen it is wanting. Enamel smooth. The mandibular ramus is rather deep and compressed, and displays an external ridge on the anterior border of the coronoid, which is not con- tinued downwards. Measurements {No. 2). M. Length of basis of true molars 0125 Diameters M. ^ | anteroposterior 0040 transverse 0035 Depth of ramus at M. ii 0073 From the Puerco beds of New Mexico. D. Baldwin. Miodmnus suhtrigonus Cope, Amer. Nat. 1881, p. 490, 491. Miodomus protogonioides Cope, loc. cit. 1882, Oct. (Sept.), p. 833. Miodcenus maiidihularis Cope, Amer. Nat. 1881, p. 830. MiodcBhus haldwini Cope, loc. cit. 1882, Oct. p. 833. 469 [Cope. GENERAL REMARKS. The preceding list of fifty-six species is doubtless sufficiently character- istic to enable us to form a pretty good idea of the Puerco fauna. Omitting six undetermined species of reptiles, we find the following peculiarities in the remaining forms. As already pointed out the three determined species of reptiles belong to a suborder, which has thus far been only found in the Laramie formation, or Cretaceous No. 6. This gives the Puerco at once a position below all the other tertiaries. The mutilate orders of mammals may be dismissed as being not likely to occur in a lacustrine formation. The orders of land Mammals are represented as follows : Monotremata 0 Marsupial ia 5 Rodentia 0 Chiroptera 0 Edentata 0 Bunotheria 15 Tseniodonta. 2 Tillodonta 2 Insectivora 0 Mesodonta 2 Lemuroidea 0 Creodoata ' .9 Taxeopoda 25 Hyracoidea 0 Condylarthra.^ 25 Proboscidea. 0 Amblypoda 0 Diplarthra 0 Carnivora. ....... ... . . . . 0 Quadrumana. 0 Totftl 45 The above list renders the peculiar facies of this fauna at once apparent. It is the only Tertiary fauna known, from which Perissodactyla are ab- sent. The absence of Amblypoda, one of the oldest types, is unexpected. The lack of Rodentia is remarkable, and perhaps only due to failure of discovery ; but if yet to be found, they must be very rare, and their absence is consistejit with their small representation in the "Wasatch beds above them. In the large number of Bunotheria, the Puerco agrees with the later Eocenes, but the order is here characterized by the small number of Mesodonta ; and the Lemuroidea are apparently absent. An especit),! feature of the fauna is the presence of five undoubted species of Mar- supialia of the family Plagianlacidae, which has its origin in the Jurassic 1 Cope.] 470 [Oct. 20, period, and extended through the Cretaceous. It is represented in the latter period in the Laramie by the genus Meniscoessus.* In the absence of a number of the existing orders of placental Mammalia, the Puerco agrees with other Eocene faunae. In the absence of all of the placental orders with convoluted cerebral hemispheres, this fauna is more primitive than any other Eocene fauna. The absence of all ungulata ex- cepting Taxeopoda, which have the most primitive foot structure, is further evidence of its primitive character. This is further increased by the pres- ence of the Marsupialia above mentioned. The general result is a mix- ture of Marsupial, and semi -marsupial forms, \Vith half lemurs, and a great expansion of the Hyracoid type. In more detail, the genera of Bunotheria may be compared with those of the period immediately following ; viz. : The Wasatch. One genus only of the Creodonta is common to the two epochs (Didymictis). Five of the species ^ maining are much like oppossums, and may be Marsupialia. The two genera {Deltatlierium and Triisodon) to which they belong, do not occur in the Wasatch. The remaining two genera, (three species) are peculiar to the Puerco, but represent a family (Mesonychidse) which occurs throughout our Eocenes. The two species of Mesodonta belong to genera of the Wasatch, one of them at least extending into the Bridger. The genera of Tseniodonta and Tillodonta are distinct from those of any of the later Eocenes, so far as known. Supplement on a new Meniscotheriumfrom the Wasatch epoch. Meniscotherium tapiacitis, sp. nov. The species now to be described is a good deal smaller than M. chamense, and, a fortiori, than the M. terrceruhrm. It is known to me from the nearly entire rami of a single mandible. These support the last five molars of one side or the other, and alveoli of two others and of the canine tooth. Two characters besides the small size, are observable in this jaw. First, the symphysis has not the shallow convex inferior outline in transverse section ; but is on the contrary angular, having sub vertical sides separate from a convex middle by a rounded angle. The symphysis is thus deeper than in M. terrcerubrm. Second, the crown of the third inferior molar tooth has partly the form of that of the second of the M. terrcerubrm. It is antero- posteriorly short, and has a short heel and no anterior basal lobe ; the sec- tion of the principal lobe is lenticular, and profile subconic. In M. terrcB- rubrcB this tooth is elongate, with well developed heel and anterior lobe. The alveolus of the canine is relatively larger than that of the M. terroefru- hrm. The coronoid process does not rise so close to the last molar tooth, nor so steeply, as in the latter species. The posterior recurvature of the internal extremity of the anterior limb of the posterior V of the true mo- lars is but little marked. * American Naturalist, 1882, p 830, Sept, 28th. m-i-i 471 [Cope. Measurements. M. Length of true molars on base 018 anteroposterior 006 transverse 0044 anteroposterior 0065 transverse 0038 Diameters M. ii Diameters M. iii •p.. . -r, ... f vertical 0045 Diameters P-m. lu < arteroposterior 004 Width of inferior face of symphysis 008 Depth ramus at P-m. iii .009 " M. iii .0103 This species was obtained by Mr. D. Baldwin from beds of probably lowest Wasatch age, in New Mexico, On the Systemaito Relations of the Camivora Fissipedia. By E. D. Cope, (Mead before the American Philosophical Society, October 20, 1882.) This order embraces the clawed mammalia with transverse glenoid cav- ity of the squamosal bone, confluent scaphoid and lunar bones of the carpus, and well developed cerebral hemispheres. It is well distinguished from all others at present known, but such definition is likely to be invali- dated by future discovery. Some of the Insectivora possess a united scapholunar bone, but the reduction of the cerebral hemispheres of such forms distinguishes them. The presence of the crucial fissure of the hemi- spheres is present under various modifications in all Camivora, while the parietooccipital and calcarine fissures are absent. The many types of existing carnivora fall into natural groups, which are ot the grade termed family in zoology. But the distinction of these from each other is not easily accompanished, nor is it easy to express their rela- lations in a satisfactory manner. The primary suborders of pinnipedia and fissipedia are easily defined. Various characters have been considered in ascertaining the taxonomy of the more numerous fissiped division. The characters of the teeth, especially the sectorials, are important, as is also the number of the digits. Turner* has added important characters derived from the foramina at the base of the skull, and the otic bulla, which Flow- erf has extended. Garrod^ has pointed out the significance of the number of convolutions of the middle and posterior part of the hemispheres. I have added some characters derived from the foramina of the posterior and lateral walls of the skuH.§ Mr. Turner also defines the families by the form and relations of the paroccipital process. * Proceeclings Zoological Soc, London, 1848, p. 63. t Loc. cit., 1869, p. .5. % Loc. cit., 1878, p. 877. § Proceedings Amer. Philosophical Society, 1880, p. • Cope. J 472 [Oct. 'Al, In studying the extinct carnivora of the Tertiary period, it has be- come necessary to examine into the above definitions, in order to de- termine the afQnities of the numerous genera which have been discov- ered. To take them up in order, I begin with the foramina at the base of the skull. The result of my study of tliese has been, that their importance was not overrated by Mr. Turner, and that the divisions of secondary rank indicated by them are well founded. Secondly, as to the form and structure of the auditory bulla. Although the degree and form of infla- tion are characteristic of various groups of Carnivora, they cannot be used in a systematic sense, because like all characters of proportion merely, there is no way of expressing them in a tangible form. For, if the forms in question pass into each other, the gradations are insensible, and not sensible, as is the case with an organ composed of distinct parts. The same objection does not apply so much to the arrangement of the septa of the bulla. The septum is absent in the Arctoidea of Flower (UrsidcB of Turner), small in the Cynoidea (Flower, Canidce Turner), and generally large in the ^luroidea (Flower, Felidm Turner). But here oc- curs the serious discrepancy, that in the Hysenidae, otherwise so nearly allied to the Felidae, the septum of the bulla is wanting. Nevertheless, the serial arrangement of the order indicated by Flower, viz. : commenc ing with the Arctoidea, following with the Cynoidea, and ending with the ^luroidea, is generally sustained by the structure of the auditory bulla, and by the characters of the feet and dentition, as well as of the cranial foramina. Turner's arrangement in the order, Ursidce, Felidae and Cani- dse, is not sustained by his own characters, and its only support is derived from Flower's observations on the external or sylvian convolution of the hemisphere of the brain.* There are three simple longitudinal convolu- tions in the raccoons ; in the civets and cats the inferior convolution is fis- sured at the extremities, while in the dogs it is entirely divided, so that there are four longitudinal convolutions between ttie Sylvian and median fissures. An important set of characters hitherto overlooked, confirms Flower's order. I refer to those derived from the turbinal bones. In the ursine and canine forms generally, the maxilloturbinal is largely developed, and excludes the two ethmoturbinals from the anterior nareal opening. In the Feline group, as arranged by Turner, the inferior ethmoturbinal is devel- oped at the expense of the maxilloturbinal, and occupies a part of the anterior nareal opening. These modifications are not, so far as my expe- rience has gone, subject to the exceptions seen in the development of the otic septa and molar teeth, while they coincide with their indications. The seals possess the character of the inferior group, or Ursidae, in a high degree. The characters derived from the paroccipital process are of limited ap- plication, as the study of the extinct forms shows. ♦Proceedings Zoological Society, London, 1869, p. 482. 1882.1 473 Cope. I would then divide the fissiped carnivora into two tribes as follows : External nostril occupied by the complex maxilloturbinal bone ; ethmo- turbinals confined to the posterior part of the nasal fossa ; the inferior ethmoturbinal of reduced size Hypomycteri. External nostril occupied by the inferior ethmoturbinal and the reduced maxilloturbinal Epimyctehi. While no doubt transitional forms will be discovered, the types at present known fall very distinctly into one or the other of these divisions. The characters are readily preceived on looking into the nares of well cleaned specimens. The Hypomycteri stand next to the Pinnipedia, since the maxilloturbinal bone has the same anterior development in that group. In searching for definitions of the families, it is necessary to be precise 68 to the definition of terms. The meaning of the word sectorial is in this connection important, since there are so many transitional forms be- tween the sectorial and tubercular tooth. A sectorial tooth then of the upper jaw, is one which has at least two external tubercles, which are the the homologues of the median and posterior lobes of the sectorial of the cat. By the flattening and emargination of their continuous edges, the sectorial blade is formed. One or two interior, and an anterior lobe, may or may not exist. In the genera of the Procyonidoe, except in Bassaris, the two external tubercles do not form a blade. The inferior sectorial tooth differs from the tubercular only in having an anterior lobe or cusp, which belongs primitively to the interior side. The inferior sectorial teeth with large heels, as in Viverridae and Canidae, I have called tubercular-secto- rials. The sectorial blade is formed by the union and emargination of the edges of the anterior and the principal external cusp. This blade is not well developed in the genus Cynogale and still less in the Procyonidm and Ursidce. The families are then defined as follows. Hypomycteri. I. No sectorial teeth in either jaw. Toes 5^5 . ^. Cercoleptidce . II. Sectorial teeth in both jaws, a. Toes 5-5 fi. No alisphenoid canal. True molars f Procyonidce, " " ^ MustelidcB. 1313. An alisphenoid canal. Molars quadrate, | Aeluridce. Molars longitudinal, f Ui'sidce, aa. Toes 5-4 or 4-4. Sectorials well developed, an alisphenoid canal Canidm, Cope. J 474 LOct. 20, Epimycteri. I. Molars haplodont. Toes 5-4 ; no alisphenoid canal ProtelidcB. II. .3Iolars bunodont, no sectorials. Toes 5-5 ; an alisphenoid canal Arctittidm. III. Molars bunodont, with sectorials. a. Otic bulla with septum. ^. Alisphenoid canal and postglenoid foramen, present. y. True molars well developed. Toes 5-5 Vive/rridcB. Toes 5-4. CynictidcB. Toes 4-4 , o Suricatidoi' yy. True molars much reduced. Toes 5-5 Cryptoproctidos. Toes 5-4 Nimramdoe. 1^13, No alisphenoid canal; post glenoid foramen rudimental or wanting. Toes 5-4 Felidce. aa Otic bulla without septum. No alisphenoid canal, nor post glenoid foramen : Toes 4-4.. Eycmidce. The genera of these families are the following : Cercoleptid^ ; Cercoleptes Neotropical. PB0CY0NID2E ; Procyon,^ Bassaricyon, Bassaris ; Neartic and Neotrop- ical. MusTELiD^ ; Melinse (two tubercles of internal side of superior sec- torial) ; Taxidea, Meles. Mustelinae, (one internal tubercle of superior sec- torials) / Enhydris, Pteronura, Lutra, Aonyx, Barangia ; Helictis, ZoriLla, Mephitis, Conepatus ; Mellivora ; Gulo, Galictis, Putorius, Mustela. JElurid^ ; Aelurus ; ^luropodm f Hyoenarctos. Ursid^ ; Eelarctos; Arctotherium ; Ursus ; Melursus. Canidje ; Megalotis\ ; AmpMcyon ; Thous, Paleocyon, Temnocyon,- Gale- cynus, Canis, Vulpes, Enhydrocyon, RycBnocyon, Tomarctus, Speothus, Synagodus, Dysodus, Oligohunis, Icticyon, Lycaon. Protelid^; Proteles. Ethiopian. Arctictid^ ; Arctictis. Indian. ViVERRiD^ ; Cynogale, Arciogale, Paguma, Paradoxui'us, Nandinia, Hemigale, Galidia, Prionodon, Genetta Viverricula, Viverra, Galidictis, Merpestes, Athylax, Calogale, Ichneumia, Bdeogale, Urm, Tosniogale, On- yehogale, Helogale, Rhinogale, Mwigos, Crossarchus, Ewplereg. Cynictid^ ; Gynictis, ? Ictitherium. SuRTCATiD^ ; Suricata ; Ethiopia. Cryptoproctid^ ; Procdurus ; Cryptoprocta. NiMRAViD^ ; Archcelurus, Mmravus ; yElui'ogale, Binictis, Pogonodon ; Hoplophoneus. * Including Nasua, which is not distinct. i This genus cannot be made the type of a family as is done by Dr. Gray, 1882.] 475 [Cope. Felid^ ; MachjBrodontinse ; Machmrodus, Smilodon ; Felinse ; Plethcel- urus (g. n.)* Gatolynx ; Felis ; Neofelis ; Uncia,\ Lynx, Cyncdurus. Hy^nid^, Hymnictis, Hycma, Grocuta. • » Type, Felis planiceps Temm. Char. Second (first) superior premolar two rooted ; orbit closed behind ; pnpil round. t Mr. Wortman has called niy attention to a character of this genus which confirms its separation from Felis, as I proposed in 1879. The maxilloturbinal bone is less complex in the genus Uncia, than in Felis, consistently with a less nocturnal habit, and less necessity for acute smell. Printed Nov. 11, 1882. ^aleontologieal Bulletin, No. 36. FIRST ADDITION TO the' Wna of the puerco eocene. (Read before the American Philosopldcal Society, January 5, 188S.) \ ON THE B I S OP THE ocEiE mmm mmmi m mmmi (Read before the American Philosophical Society, December 15, 1S82.) FOURTH CONTRIBUTION TO THE nSTORY OF THE PERMIAN FORIATIOH OF TEXAS. {Read before the American Philosophical Society, March 16, 1883.) By PROFESSOR E. D. COPE. For Sale by A. E. Foote, \ 1223 BELMONT AVENUE, PHILADELPPIIA. I 1883. ] 515 [Cope. First Addition to the Fauna of the Puerco Eocene. By E. D. Cope. {Read before the American Philosophical Society, Jan. 5, 1883.) There are fifty -five species included in my synopsis of the vertebrata of the Puerco epoch*. Ten of these are reptilia, the remainder mammalia. In the present paper a number of interesting additions are made. The typical specimens are figured in the fourth volume of the U. S. Geological Survey of the Territories, now in press. , Ophidia. Helagras pmsciFORMis, gen. et sp. nov. Char. gen. The generic characters are drawn from vertebrae only. These display a modified form of the zygosplien articulation, as follows : The roof of the zygantrum is deeply notched on each side of the median line so as to expose the superior lateral angles of the zygosplien. This separate median portion of the roof of the zygantrum torms a wedge- shaped body which may be called the episphen, It is surmounted by a tuberosity, which constitutes the entire neural spine. The latter is thus entirely different in form from that of other serpents. Articular extremi- ties of centrum round, the ball looking somewhat upwards. Costal arti- culation 8-shaped, the surfaces convex and continuous. Hypapophyses none on the two vertebrae preserved. Zygapopliyses prominent. Free diapophyses none. This genus is readily distinguished by the presence, now first observed, of the episphen in addition to the zygosphen ; and by the p^uliar form of the neural spine. We have now several vertebral articulations originally discovered in American vertebrata. These are the episphen as above, the hyposphen, which characterizes the Opistliocoelous Dinosauria (Sauropoda Marsh), and the Diadectidce of the Permian period ; and the zygantra- pophysis, which is present in the Diplocaulid family of Batrachia. CJiar. specif. A section of the vertebra at the middle is pentagonal, the inferior side slightly convex downwards. The lateral angle is the section of the angular ridge which connects the zygapopliyses, The episphen has a shallow rounded groove on its infero-posterior side, which is bounded by a projecting angle on each side at its middle. The episphen does not pro- ject so far posteriorly as the postzygapophyses, and the degree of its prom- inence diff'ers in different parts of the vertebral column. In one of the two vertebrae in my possession its prominence is small. The tuber- osity on its summit is a truncate oval with the long diameter anteropos- terior, and equaling two-fifths the length of the arch above. It is ele- vated above the rest of the median line, which is roof-like, with obtuse angle. The tubercular articular facet is entirely below the prezyga- pophyseal surface, but the tree part of the prezygapophysis extends well in front of it. It is distinguished from the capitular surface by a very slight constriction. A slight ridge extends from the capitular articulation * Paleontological Bulletin- No. 35, Nov. 11th, 1882. Cope.] 546 [Jan. 5, to the edge of the ball of the centrum. Below this the surface is slightly concave, and the middle line is gently convex. The latter terminates in an obtuse angled mark just in front of the edge of the ball. This edge is also slightly free from the ball. The capitular costal surfaces do not pro- ject inferiorly quite to the line of the inferior surface of the centrum. Measurements of a Vertebra. M. Length of centrum (with ball) 0070 ^ c vertical 0035 Diameters of ball ^ transverse 0040 Elevation of vertebra at episphen 0085 middle 0063 Width at prezygapophyses 0120 " tubercular costal faces 0105 •* of zygantrum 0058 Vertical diameter costal faces 0040 Transverse diameter tubercular costal face 0028 This snake was about the size of the black snake, Bascanium constrictor. It is an interesting species for two reasons. First, it is the oldest serpent known from North America. Second, in the imperfection of the zy.iian- trum we observe an approximation to the ordinary reptilian type of verte- bra, from which the ophidian type was no doubt derived. In the former there is no zygosphen or zygantrum. * Mammalia. Triisodon levisianus, sp nov. This creodont is represented by part of a right mandibular raums which contains the fourth premolar minus its principal cusp, and the first and second ttue molars, with the alveoli of the third. The ramus is deep, and probably belonged to an animal of about the size of the red fox. The molars have the structure most like that of the T. heilprinianus, especially an- teriorly. The principal anterior cusps are united together for most of their elevation, while the anterior inner is much smaller and lower, and is situ- ated between the middle and inner side of the anterior cusp. The heel is rather wide, and has a raised border. The external part of it is angular, ?ind is somewhat within the vertical line of the base of the crown. The fourth premolar differs from that of the type the genus, T. quivirensis, in having two acute longitudinal tubercles situated close together on the heel. The anterior masseteric ridge is very promijient. The masseteric fossa is strongly concave, but shallows gradually inferiorly. Its inferior border presents a low thickened ridge, which is recurved in front. This may be an individual character only. The inferior outline of the ramus is gener- ally convex, and does not rise much below the masseteric fossa. [Cope. Measurements. M. Length of last four inferior molars 0315 true molars 0230 / anteroposterior 0085 Diameters of M.i. I 0055 Length of P-m. iv. on base 0090 Depth of ramus at M.i 0200 Thickness " " 0085 This Triisodon is not only materially smaller than the T. heilprimanus, but differs in the characters of the heel of the inferior molars. In that species the internal border is tubercular ; in this one it is entire. The T. conidens and T. quivirensis difier in the arrangement of the anterior cusps. Dedicated to my friend, Henry Carvill Lewis, professor of mineralogy and geology in the Academy of Natural Sciences, Philadelphia. MiOCLiENUS FEROX, Sp. nOV. This new species is represented by three specimens. One of these in- cludes various separate teeth and a considerable portion of the skeleton ; a second includes loose teeth and a smaller number of bones of the skeleton ; and the third consists of a part of a mandibular ramus, which contains the three true molars. These indicate the largest species of the genus yet known, the first individual above mentioned being about the size of a wolf. The bones of the Miodmnus ferox enable me to refer the genus approxi- mately to its proper position in the system. Although we do not possess the corresponding parts of the Mioclcenus turgidus, the type' of the genus, it is probable, if not certain, that they agree in generic characters. The agreement in dentition extends to all the principal technical points, though the specific differences are marked. The skeleton is that of a creodont. Tlie unequal phlanges are compressed I (daws, and the metapodial bones have protuberant condyles. The astrag- alus has a simple head with convex surface, and the trochlea is a shallow open groove. I The tubercular dentition refers this genus to the Arctocyonidce.^ With ! this family it is accordingly placed provisionally. It differs from the known I fossil genera in the single tubercle of the internal part of the crown of the I superior molars. I The species M. brachystomus and M. etsagicus of the Wasatch epoch must now be removed from this genus. I have shown that the former is an Artio- dactyle. Now in technical points, the dentition of those species is identi- cal with that of Pantolestes Cope, as well as with Jfiodcenus. Although the ; skeleton of the type of Pantolestes, P. longicaudus of the Bridger Beds, is yet unknown, it is safe to suppose that it does not differ from that of the M. brachystomus. I therefore refer the two species first mentioned to Panto- lestes, and place that genus in the Artiodactyle sub -order. * For the dentition of this family sec Lemoine, Annales, Sc. Nat., 1878, July. 4 Cope.] ^48 [Jan. 5, Char, specif. — The canines are well developed, and have a robust root. The crown is rather slender and is very acute. It is rounded in front, but has an acute angle posteriorly. It is not grooved, and the enamel is smooth. The single-rooted first superior premolar is situated close to the canine, and behind it is a short diastema. I have the probable first true molar or fourth premolar. The external cusps are rather small, and are well separated from each other. The inner outline of the crown is rather broadly rounded. The internal tubercle is connected on wearing, with an anterior transverse crest which terminates near the inner base of the anterior external cusp in an intermediate tubercle. There is a posterior intermediate tubercle. There is a cingulum all round the crown excepting at the posterior intermediate tubercle. The second (? first) true molar is like the one just described, but has relatively greater antero-posterior width. In this tooth the cingulum extends all the way round the crown. There are but two inferior molars of this individual preserved, the second and third true. The former of these has a parallelogrammic outline with rounded angles. There are two posterior and two anterior rather large tubercles ; an anterior transverse ledge ; and a narrow external and posterior cingulum, the latter running into the internal posterior tubercle. The latter has a circular section, and is much smaller than the external posterior, which has a wide crescentic section. Of the anterior tubercles the anterior is much the larger, judging from its worn base. The third true molar is triangular in outline. Its crown includes two anterior and an external median tubercle. The inner and posterior parts of the crown form a wide shelf, with the internal edge denticulate. A weak external cingulum. Measurements of Teeth. M. Diameters base of crown of incisor \ anteroposterior 0045 c transverse 004 Tk- ^ v ^ . ( anteroposterior 0130 Diameters base crown of canine < ^ t transverse 0095 T.' . • Tv/r • ^ anteroposterior 0095 Diameters crown, superior M. i. ^ ^ t transverse 0120 / anteroposterior 0110 I transverse 0110 Diameters of inferior M. ii { anteroposterior 0120 1^ transverse OIOd / anteroposterior 0125 I transverse 0090 Diameters, M.? ii. Diameters of inferior M. iii. The second individual includes part of the superior walls of the skull. The fragment displays a high sagittal crest, which is fissured in front so as to keep the temporal ridges apart to near its anterior apex. The brain surfaces show small, smooth, flat hemispheres, separated by a constriction Irom the wide and large olfactory lobes. The navicular bone shows three well defined dislal facets, indicating probably five digits in the pes. The teeth of this specimen include a posterior superior molar, and an inferior 1883. 549 [Cope. third or fourth premolar, with other teeth. The premolar is like that of a creodont. Its principal cusp is a simple cone. To this is added a short wide heel, whose superior surftice is in two parts, a higher and a lower, divided by a median ridge. . A low anterior basal lobe, and a weak exter- nal cingulum. The third specimen belonged to an individual a little smaller than the other two. It includes the first inferior true molar, a tooth lost from the others. Its form is somewhat narrowed anteriorly, where it has two low, but well separated anterior inner tubercles, which form a V with the ex- ternal anterior. Specimen No. 1 is accompanied by fragments of vertebrae and limbs. The former are principally from the lumbar region, but fragments of the atlas remain. This vertebra is of moderate length, and the cotylus is somewhat oblique. The vertebrarterial canal is rather elongate, and its anterior groove-like continuation in front of the diapopliysis is not deeply excavated. The lumbar vertebrae are remarkable in the characters of their zygapophyses. These display subcylindric surfaces of the posterior pair, which indicates that the anterior ones are involuted, as in the specialized Artiodactyles and Perissodactyles of the later geological ages. Such a structure does not exist among carnivora, nor to my knowledge among creodonta, nor in any mammals of the Lower Eocene. I do not find it in Didelphys nor Phascolarctos, but it exists in a moderately developed degree in Sarcophilus. The articular surface forms more than half of a cylinder, and its superior portion is bounded within by an anteroposterior open groove. The surface within this is not revolute, as in Bos and 8us, but the articular surface disappears, as in Cervus. Eight such postzygapoph- yses are preserved, all disconnected from their centra. Two of them are united together. There are two other separated zygapophyses of smaller size, which have but slightly convex surfaces. One is probably a prezyg- apophysis of a dorsal vertebra. No centrum is preserved. Of the anterior limb there is a probable distal half of a radius. It is of peculiar form, and resembles that of SarcopJtilus ursinus more than any other species accessible to me. One peculiarity consists in the outward look of its carpal surface, which makes an angle of about 45^ with the long axis of the shaft. The oblinuity in 8. ursinus is less. The external border of the shaft in M. ferox is, however, straight, and terminates in a depressed tuberosity. Beyond this, the border extends obliquely outward to the carpal face, which it reaches at a right angle. The internal border of the shaft is gradually curved outwards to the external border of the car- pal face. Its edge is obtuse, while the external one is more acute for a short distance, and rises to the anterior (superior) plane of the shaft. The carpal face is a spherically subtriangular with rounded angles. It displays two slightly distinguished facets, one of which is superior, and the other is larger and surrounds it, except on the superior side. The internal mar- ginal projection, or "styloid process," is not so prominent as in S. ursi- nus, and is a roughened raised margin. Joining it on the inferior edge of Cope.] 550 [Jan. 5, the carpal face is another rough projection of the margin. Immediately opposite this, on the superior edge of the carpal face, is a rough tuberosity, which encloses a small rough fossa, between itself and the styloid pro- cess. Internal to it is a shallow groove for an extensor tendon of the manus ; then a low short ridge, and internal to that a wide shallow de- pression for other extensors. The carpal face differs greatly from those of SarcopJiilus and Didelpliys in having the inner portion wider than the outer, instead of the reverse, and in having no distinct styloid process. It indi- cates that the manus was turned outwards much more decidedly than in those genera. Of carpal bones the only recognizable one is the unciform. Its proximal articular surface rises with a strong convexity entad, and descends to an edge ectad. The metacarpal surface is concave in anteroposterior sec- tion, forming a wide shallow groove, extending in the direction of the width of the foot. Its two metacarpal areas are not distinguished. The entire first and second metacarpals, with the heads of the third and fourth are preserved. They considerably resemble those of Sarcophilus ursinus. The distal articulations are injured in both, but both display a sharp troch- lear keel posteriorly, which on the second extends nearly to the superior face of the articulation. The condyle is subround, and is constricted lat- erally, and at the base above. The second metacarpal is short and ro- bust, shorter than in Sarcophilus ursinus. The first is also robust, but is relatively longer, as it is three-quarters the length of the second. Its head is expanded, especially posteriorly, and the large trapezial face is subtrian- gular, with round apex directed inwards as well as forward. The poste- rior face of the head is notched ectad to the middle. On the external side of the head there is a vertical facet with convex distal outline, for con- tact with the second metacarpal. The head of the latter is narrow, and is concave between the sides. The concavity is bounded posteriorly by a raised edge. The anterior part of the proximal facet is decurved. The shaft is deep proximally, but on the distal half is wider than deep. The lateral distal fossse are remarkably deep and narrow, the condyle very much contracted. The head of the supposed third metacarpal is as wide as the sec- ond anteriorly, but narrows to the posterior third, and then contracts ab- ruptly to a narrow apex. The supposed external side of the head is per- fectly straight, and is continuous with the side of the shaft without inter- ruption. The entad side displays no facet, but has a depression below the head which adapts itself very well to the head of the first metacarpal. In fact, if the metacarpals just named second and third, exchange places, so that second is placed third and third second, the metacarpal series fits far better. The fourth fits the so-called second much better than the so- called third. This may therefore be the true order, although that first used agrees better with the carpus of Sarcophilus. The head of the so- called third is slightly convex anteroposteriorly, and is oblique laterally, descending a little to the inner side. The fourth metacarpal is wider an- teriorly than either the second or third. The inner edge is straight, while 1S83.] 551 [Cope. the outer is concave, the head being narrower before than behind. It has a lateral facet on each side ; the inner plane, the external concave in the vertical as well as in the anteroposterior direction. It thus approaches the form of a metatarsal, but is not so strongly excavated, nor is the head notched on either side. The unciform face is convex anteroposteriorly and plane transversely. The femur is broken up so that I cannot restore it. The head of the tibia is gone, but a considerable part of the astragalar face is preserved. This is transverse to the long axis of the tibia. It is narrowed anteroposteriorly next the fibular fticet. Malleolus lost. The shaft is robust, and does not expand distally for articulation with the astragalus. Three centimeters proximal to the distal end, the external side throws out a low, rough, ridge-like tuberosity. Above the middle the crest turns outwards, leaving the internal face convex. There is a broken patella, which has one facet much widw than the other. The astragalus has the trochlear portion a little oblique. That is, the in- ternal crest is a little lower than the external, and the inner face is a little sloping. The latter is impressed by a fossa above the posterior part of the sustentacular facet, which runs out on the neck. The trochlea has a shallow groove which is nearer the external than the internal crest, and which passes entirely round the posterior aspect to the plane of the inferior face of the astragalus. The groove for the flexor tendon is thus entirely en- closed, and issues on the inferior face at the posterior extremity of the groove which separates the sustentacular from the condylar facets. The external crest of the trochlea is less prominent posteriorly than the internal, thus reversing the relations of the superior part. The internal ridge be- comes quite robust, but does not flatten out and project sub-horizontally as in Oxymna forcipata. The fibular face is vertical ; neither its anterior nor posterior angles are produced. The neck is somewhat contracted (the in- ternal side is injured). The head is a transverse oval, strongly convex vertically, moderately so horizontally, and without flattening. A meso- cune'iform (or possibly ectocuntiform) bone is wedge-shaped in horizontal section, without posterior tuberosity, and its anterior face is a slightly ob- lique square. The narrower facet is oblique in the transverse sense. The metatarsals are represented, excepting the first and second. The only complete one is the fifth. The heads of the third and fourth are much like those of Oxymna forcipata, and of about the same size. Their anterior width is equal, and in both the external side is more oblique than the in- ternal. Both have a notch at the middle of the internal side, but they dif- fer in that the third has an open notch on the external side which is want- ing to the fourth. The lateral excavations of the external sides are deep and rather large, and thin out the anterior external edge. The lateral facets are correspondingly large on the fourth and fifth ; on the third meta- tarsal it is small, and a mere decurvature of the proximal surface. That of the fourth is longer proximo-distally than transversely. That of the fifth is about as long as wide, and presents more anteriorly ; or, to express it Cope.] 552 [Jan. 5, more accurately, the shaft and head present more outwardly than those of the fourth. The proximal, or cuboid facet is narrow anteroposteriorly, and is curved, the external side being concave. On the external side just distal to this facet, the head of the bone expands into a large outward-looking tuberosity, which is separated from the posterior tuberosity by a strong notch. Between it and the head proper, on the anterior face, is a large fossa. The entire form is something like that of the proximal extremity of a femur with head, neck, great trochanter and trochanteric fossa. A some- what similar form is seen in the corresponding bone of Oxymna forcipata. The shaft of the fifth metatarsal, is one-fifth longer than that of the second metacarpal (? 8d) above described. Its direction is straight, but it is some- what curved anteroposteriorly. Its section is subtriangular, the apex ex- ternal. The condyle is narrowed and sub-globular above, and spreads laterally behind, the external expansion being wide and more oblique. The keel is prominent, and is only visible from above (in front) as an angle. The distal extremities of some other metatarsals difier in being flatter at the epicondyles, and concave between them on the posterior face. The con- dyles are more symmetrical, and are bounded above on the anterior face by a profound transverse groove. Several phalanges are preserved, including part of an unguis. They are till depressed, and with well marked articular surfaces, of which the distal are well grooved, and the proximal notched below. The lateral areas of insertion of the tendons of the flexors are well marked on the edges of the posterior faces. An ungual phalange is much compressed at the base. The basal table is well marked, and has a frefe lateral edge. The nutritive foramen enters above the posterior extremity of this edge. No trace of basal sheath. Measurements oj No. 1. M. Length of atlas at anterior vertebrarterial foramen 0165 Expanse of postzygapophyses of a lumbar vertebra 0230 Diameter radius at middle of shaft 0100 Greatest distal width of radius 0220 Diameters carpal surface | ^^^^^^^^ ^^^^ I transverse 0185 / vertical (interiorly) 0130 Diameters of unciform < anteroposterior (greatest) 0140 ' transverse (in front) 0150 anteroposterior 0130 transverse 0120 Length of metacarpal 1 0310 Width metacarpal I at epicondyles 0110 r anteroposterior 0110 I transverse 0070 Length of metacarpal II (or III) 0400 Width do. at epicondyles 0120 / anteroposterior 0125 transverse 0075 Diameters head metacarpal I Diameters head metacarpal II \ Diameter head of M. Ill (or II) | 1883.] 553 [Cope. Measurements of No. 1. M. Diameters head of M. IV | anteroposterior. . 0130 Uransverse (at middle) 0070 Wiatli of patella near middle 0190 Diameters of tibia .07 M. from astragalus | ''^^^^^"^^^^^^^^^ ^J^^ Uransverse 0130 Anteroposterior width of astragalar face 0200 Total length of astragalus 0310 / length on groove 0210 Diameters of the trochlea } width above 0160 ' elevation externally 0130 Greatest width of astragalus below - 0225 Length anterior to internal crest of trochlea 0100 Diameters head of metatarsal III { ^^^^^^po^t^y^^^ • Uransverse (m front) 0110 Diameters head of metatarsal ly j anteroposterior 0140 C transverse 01 Oo / anteroposterior 0120 ' without tuberosity < ^ with lateral ' transverse } facet 0080 (without do.. .0040 transverse over all 0170 Length Mt. Y 0460 Width do. at epicondyles 0120 Width do. at condyle above 0065 Width of M. Ill or IV at epicondyles 0120 Width of proximal end of phalange 012 Length of smaller phalange (1st series) 0230 r vertical 0070 i transverse 0110 Ungual phalange, vertical diameter of cotylus 0090 Diameters head M. V Proximal diameter do- The specimen which has been partially described in the preceding pages as No. 2, has many pieces which are identical with those preserved in spec- imen No. 1. Among these may be mentioned the glenoid cavities of the squamosal bone. These display, besides the large postglenoid process, a well developed preglenoid ridge, as in Arctocyonidce, OxycenidcB and Meso- nychidoB. A large distal caudal vertebra of elongate form, indicates a long tail. An articular extremity of a flat bone is intermediate iii form be- tween the proximal end of the marsupial bone of Didelphys and that of Sarcophilus. Its principal and transverse articular surface is transversely convex, as in the latter {8. ursinus), but the lesser articular fiice is sepa- rated from it by an even shorter concave interspace than in the opossum. It has almost exactly the form of that of the latter animal. It is a short, flat cone, with two faces presenting on the same side, the one part of the concavity mentioned, the other flat and presenting away from it. This Cope.] 554 [Jan. 5, piece has a slight resemblance to the very peculiar head of the fibula in the oppossum, but is not like that of Sarcophilus ursinus. I, however, think it much more probably the proximal extremity of a marsupial bone. A supposed cuneiform is subtransverse in position, and resembles in gen- eral those of Oxycena and Esthonyx. It has the two large transverse prox- imal facets, the anterior one-quarter wider than the posterior. The distal facet (trapeziotrapezoidal) is simple. The navicular is much like that of OxycRTia forcipata, but is more robust. Its external tuberosity is flattened anteroposteriorly, and is produced proximally. The three distal facets are well marked, the median a little wider than the external, while the internal is subround, convex, and sublateral in position. The entocunei- form is a flat bone, with cup-shaped facet for the navicular, and narrow facet for the first metatarsus. This facet is transverse transversely, and concave anteroposteriorly. It shows (1), that there is a pollex; (2), that it is probably small ; and (3), that it was not opposable to the other digits, as is the case in the opossum. (4). It does not show whether the pollex has an unguis or not. Measurements JVo. 2, M. Transverse width condyle of mandible 0230 Anteroposterior width condyle of mandible (at middle) .0103 Diameters head oUsmarsupiil'^^^''^''^^^^ ^^^^ I anteroposterior 0068 Diameters cuneiform ( ^^^^^^^^ I anteroposterior 0115 r vertical in front .... , 0085 Diameters naviculars transverse 0180 ' anteroposterior (middle) 0110 /vertical at middle 0100 Diameters ectocuneiform ? anteroposterior (middle) . . .0140 ' transverse distally 0060 Two other bones of specimen No. 2 I cannot positively determine. The first resembles somewhat the trapezium of Sarcopliilus ursinus, and still more that of: Didelphys. I will figure it, as a description without identifi- cation will be incomprehensible. The next bone is of very anomalous form. It may be the magnum, which is the only unrecognized bone of importance remaining, or it may be a large intermedium. It has no re- semblance to the magnum of any mammal known to me. It was evi- dently wedged between several bones, as it has eight articular facets. Two are on one side ; the largest (convex and oval) is on one edge ; three are on one end, and two, the least marked, are on the other flat side, oppo- site to the first. Restoration. We can now read the nature of the primitive mammal MioclcBuus ferox, in so far as the materials above discussed permit. It was a powerful fiesh-eater, and probably an eater of other things than flesh. It had a long tail and well-developed limbs. It had five toes all around, and the great or first toe was not opposable to the others, and may have been 1883.] 555 [Cope. nidimental. The feet were plantigrade and the claws prehensile. The fore feet were well turned outwards. There were in all probability mar- supial bones, but whether there was a pouch or not cannot be deter- mined. These points, in connection with the absence of inflection of the angle of the lower jaw, render it probable that the nearest living ally of the Mioclcmus ferox is the Thylacynus cynocepJialus of Tasmama. The pres- ence of a patella distinguishes it from Marsupials in general. Its den- tition, glenoid cavity of the skull and other characters, place it near the Arctocyonidce. Should the forms included in that ftimily be found to pos- sess marsupial bones, they must probably be removed from the Creodonta and placed in the Marsupialia. This species is about the size of a sheep. The bones are stated by Mr. Baldwin, who discovered it, to be derived from the red beds in the upper part of the Puerco series. MlOCL^NUS BUCCULENTUS, Sp. nOV. A part ot the right maxillary bone which supports three molars indi- cates this species. The molars are P-m iv, M. i and M. ii. This series is characterized by the remarkably small size of the fourth premolar, and large size of the second true molar. The first true molar is intermediate. The fourth premolar consists of an external cone and a much smaller in- ternal one. There is a weak posterior basal cingulum. The reduced size of the internal cone suggests the probability that the third premolar has no. internal cusp,and that there may be but three premolars. In either case the species must be distinguished from Miodcenus. The first and second true molars have conic well separated external cusps, and a single pyramidal internal cusp. The intermediate tubercles are distinct. There is a posterior cingulum which terminates interiorly in a flat prominence. There is an anterior cingulum and a strong external one, which form a prominence at the anterior external angle of the crown. Enamel wrinkled. Measurements of Superior Molars. M. Length ot bases of P-m. iv M. i and ii 0180 Diameters P-m. / anteroposterior 0040 transverse 0046 Diameters of M. i / anteroposterior 0060 t transverse 0065 Diameter of M. ii | anteroposterior 0070 t transverse 0085 MlOCL^NUS SUBTRIGONUS Cope. This species has been known hitherto* from a palate with three molars. I am now able to give the characters of the inferior molar series, which have been found, by Mr. Baldwin, associated with the true superior molars. Of the latter it may be remarked that the second true molar is not so much * American Natiirallst, 1881, 490-1. Cope.] 556 [Jan. 5, longer than the first as in M. lucculentns, although the difference in size is very evident. The third is smaller than the first, and ovoid in outline, while the first and second are subquadrate. The external cusps are conic and widely separated and the intermediate areas are distinct. There is a cingu- lum all round the crown of the last two, and round that of the first except at the inner side, and at the anteroexternal angle. The last three inferior premolars are higher than long at the base, and are compressed and the apex acute. The posterior edge of the third and fourth is truncate, and simple. Each has a posterior cingulum which forms a narrow heel on the fourth. No other cingula. Of the true molars only the second is wanting. The form of these is like those of the M. ferox, with the cusps more prominent. The first only has trace of the anterior V ; in the others, the two anterior tubercles are opposite and connected by a short anterior ledge. The heel of the first consists of a basin bounded by these tubercles, of which the external is pyramidal and largest. The median posterior is small. The heel of the third is narrow and prominent, and the internal lateral tubercle is represented by a short raised edge. The enamel of all the molars is wrinkled, and the inner side of the premolars is grooved with the height of the crown. A weak external cingulum on M. iii. Measurements. M. Length of last three superior molars 0265 Diameters of M. i \ anteroposterior 0060 t transverse 0060 anteroposterior 0063 transverse 0072 r anteroposterior 0047 I transverse 0060 Length of last inferior molars 0340 Length of last three premolars 0140 Length of P-m. iv 0050 Elevation of P-m. iv 0050 Diameters of M. i | anteroposterior 0057 ( transverse 0042 Diameters of M. m ( anteroposterior 0070 t transverse 0035 Rather larger than the pine weasel, Mustela americana. MlOCL^NTJS COFtRUGATUS, Sp. UOV. This species is known from a right maxillary bone which contains the last four molar teeth, with parts of pelvis and other bones of one individual. This species is intermediate in size between the M. protogonioides and M. ferox, as the following measurements of the second superior true molar show : M. protogonioides. M. eorrugatus. M. ferox. Diameter, transverse Oil .0118 .015 anteroposterior 008 .010 .013 Diameters of M. Diameters of M. iii 1883 ] 557 LCope. The superior molars are more nearly quadrate than in the other species of the genus, owing to the better development of the posterior internal tubercle, which is, however, as in the others, a mere thickening of the posterior cingulum. It is wanting from the last superior molat. The cusps on the true molars are as in the M. ferox, small, and not large and closely placed as in M. protogonioides. The intermediate ones are nearly obsolete. The crowns are all entirely surrounded by a cingulum. The entire enamel surfaces wrinkled so as to be rugose, although the teeth are those of an adult and well used. The second superior molar is larger than the first, exceeding it in the transverse rather than the fore-and-aft diame- ter. The third is the smallest, and is of oval form with obliquely truncate external face. It is less reduced than in the M. turgidus. The fourth premolar consists of a strong compressed-conic cusp with three basal cusps of small size, viz., an anterior, a posterior, and an in- ternal. The last is the larger, though small, is formed like a heel, and is connected with the others by a cingulum. No external cingulum. Measurements. M. Length of last four molars 036 Diameters P-m. | anteroposterior 010 transverse 008 " M J / anteroposterior 010 I transverse 010 " M iii / anteroposterior 008 I transverse Oil From the Upper Puerco beds. Pantolambda bathmodon Cope, American Naturalist, 1882, p. 418. In describing this genus and species, I remarked, loc. cit., that they were "founded on a mandibular ramus, which supports the first true molar, and the last two premolars. The characters of these teeth remark- ably resemble those of Corypliodon. * * * It will be for additional material to demonstrate whether this genus belongs to the AmUypoda gi' Perissodactyla." A considerable part of the skeleton of this species having been recently sent me by Mr. D. Baldwin, I am able to throw much light on the affini- ties of this curious genus. In the first place, the phalanges (not ungual), show that the genus is ungulate. Secondly, the astragalus has a large distal facet for the cuboid bone. This proves that the genus cannot be referred to the Taxeopod order. The question as to whether it belongs to the Amblypoda or the Diplarthra would be decided by the carpus, but that part is unfortunately not preserved, and I have to rely on empirical indications for a provisional determination. Apart from the astragalus, the characters are those of the Gondylarthra rather than of the Perissodactyla, and it is therefore to be supposed that the carpus has also the characters of that order. This would Cope.] 558 I Jan. 5, place the genus in the Pantodonta, which has the carpus nearly that of the Taxeopoda, and the tarsus of the Diplarthra. The points of resemblance to the Condylarthra are the following : The ilium is narrow. The humerus has an.epltrochlear canal. The superior molar teeth have but one internal lobe. The resemblances to the Pantodonta are these : The cervical ver- tebrae are plane and short. The femur has a third trochanter. The pre- maxillary bone in dentigerous. The astragalar trochlea is as in the PeriptychidcB, and the Probosddia ; that is without groove, and slightly convex anteroposteriorly, thus differing from that of the Pantodonta. The dentition is especially like that of the Amhlypoda in general, and that of the superior series is unlike anything known in the Diplarthra. I propose to place this genus in the Amhlypoda for the present, next to the Pantodonta, but it cannot enter that sub-order on account of the form of its astragalus. The sub-orders of Amblypoda will be defined as follows : Astragalus with a head distinct from trochlea, with distal ar- ticular facets , Taligrada. Astragalus without head ; distal facets subinferior Pantodonta. In the sub-order Taligrada, the single family Pantolambdidce presents the following characters : Superior and inferior molars with the cusps developed into Vs. Post- glenoid process present ; postympanic and paroccipital not distinct. All the vertebrae with plain articulations. Humeral condyles without inter- trochlear ridge. Femur with third trochanter. Digits of posterior foot probably five. Metapodial keels small and posterior. Of this family Pantolambda is as yet the only known genus. Its leading characters are as follows : Canine teeth distinct ; dental series continuous. Superior molars all triangular, that is with a single internal cusp. External cusps of premo- lars unknown ; of molars two. Internal cusp V-shaped, sending its horns externally as cingula to the anterior and posterior bases of the external side of the crown, without intermediate tubercles. Inferior true molars with a crown of two Vs, the anterior the more elevated. Premolars con- sisting of one open V, with a short crest on a short heel, as in Coryphodon. Dental formula Pf ; C. { ; P-m. J| ; M. f ; the last inferior with a heel. A strong sagittal crest. Auricular meatus widely open below. Large postparietal, postsquamosal and mastoid foramina. Cervical vertebrae rather short ; other vertebrae moderate, the lumbars not elongate. A large tail. Humerus with large internal epicondyle. Femur with all the trochanters large. Ilium with the anterior inferior spine well developed. Metacarpals short, plantigrade. Phalanges of second series flat, and of subquadrate outline. The astragalus has a wide head, but no neck, as it is not separated from the tochlear portion by a constriction. It is as wide as the trochlear portion, but about one-third of its length ex- tends within the line of the malleolar face of the trochlear portion. The 1883.] 559 fCope. navicular face is flat, that of the cuboid bone is convex vertically, and one- half as long horizontally as the navicular, and only half as deep. These two facets are continuous with the sustentacular below. Interior to all of these, on the internal tuberosity of the head is a sub-round facet look- ing inwards, like tliat characteristic of the genus Bathmodon, but rela- tively larger. A continuous facet is seen on the adjacent edge of the navicular. The use of tiiese fticets is unknown. The brain case indicates small and nearly smooth hemispheres, extend- ing with little contraction into a rather large cerebellum. The olfactory lobes are produced anteriorly at tlie extremity of a rather long isthmus. If we consider the dentition alone, Pantolamhda is the ancestor of the CoryphodmtidcB. The history of the feet requires further elucidation. The Pantolambda bathmodon is about as large as a sheep. From the upper beds of the Puerco. MiXODECTES PUNGENS, gCn. Ct Sp. UOV. Char. Gen. — The position of this genus is uncertain, but may be near to Cynodontomys Cope, which I have provisionally placed among the*Pro- simise*. It is onl}^ known from mandibles, which have presumably the following dental formula. I. 0 ; C. 1 ; P-m. 4 ; M. 3. An uncertainty exists as to the proper names of the anterior teeth, which cannot be de- cided until the discovery of the superior series. For instance the formula may be ; 1. 1 ; C. 1 ; P-m. 3. The supposed canine is a large tooth, issuing from the ramus at the symphysis like a rodent incisor, and has an oval section, with long dia- meter parallel to the symphysis. The crown is lost from all the speci- mens. The second tooth is similar in form to the first, but is much smaller. It is situated posterior and external to the first. The next tooth is still smaller and is one-rooted. The third and fourth premolars have simple conic crowns, and more or less developed heels without cusps. The true molars are in general like those of Pelycodus ; i. e., with an anterior smaller, and a posterior triangle or V. The supplementary anterior inner cusp is quite small, while the principal anterior inner is elevated. The posterior inner is much more elevated than in the species of Pelycodus. Last inferior molar with a fifth lobe. This genus cannot be referred to its place without additional material, but the parts discovered indicate it to be between Pelycodus and Cynodon- tomys ; either in the Mesodonta or the Prosimim. I may here remark that in defining the latter genus I was in doubt as to the number of the inferior premolars. The discovery of the present genus renders it probable that it has three such teeth, and that the anterior two are each one-rooted. . Char. Specif. The mandible of the Mixodectes pungens is about the size of that of the mink. Its inferior outline is straight to below the second premolar, whence it rises upwards and forwards like that of a rodent. The anterior masseteric ridge is very prominent, but terminates below the * Paleontological Bulletin No. 34, p. 151. Cope.] 560 [Jan. 5, middle of the ramus. Inferior masseteric ridge much less pronounced. The inferior part of the ramus is robust below the base of the coronoid process, but there is no indication of recurvature of the, edge. Mental foramina two ; one below the front of the first true molar, and one below the second premolar. The oval base of the canine is not flattened on either side; that of the second tooth is flattened on the inner side. There is a great difference be- tween the sizes of the last three premolars. The fourth is twice as large as the third, and the second, judging from the space and the size of its al- veolus, was much smaller than the third, and the crown was probably a simple acute cone. The crown of the third is of that form, with the addi- tion of a short heel. The long axis of the base of the crown is diagonal to that of the jaw. The fourth premolar has a relatively larger heel than the third, bui it is shorter than the diameter of the base of the cusp. Its pos- terior edge is elevated. The cusps of the anterior pair of the true molars are elevated, but the interior is the most so. The supplementary one is not exactly in the line of the interior border of the crown. Each of the inner cusps are connected with the base of the external by a ridge, w^hich together form a V. The posterior base is nearly surrounded by a raised edge, which rises into cusps at the posterior lateral angles. Of these the internal is the more prominent. The edge connecting these cusps is slightly convex backwards, and evidently bears a part in mastication. The lateral borders of the last molar are somewhat expanded, and the fifth lobe is very short. No cingula on any of the teeth. Measurements. M. Length of dental series from " canine" exclusive 0265 " true molar series 0140 ( longtitudinal 0040 Diameters - canine" {t.^ns^erse 0030 Long diameter of base of "P-m. i" 0028 " " " P. m-ii 0017 -r,. ^ -D • r vertical 0055 Diameters P-m. iv ^ . ^^^^ C anteroposterior O050 Diameters M, ii (transverse... 0038 I anteroposterior 0050 Length of crown of M. iii 0060 Depth of ramus at P ni. iii 0090 M. iii 0100 HlXODECTES CRASSIUSCULUS, Sp. nOV. This mammal is represented by fragments of two mandibles from differ- ent individuals ; one less and the other more worn by mastication. The species differs from the last in its greater size, and in the relatively greater length of tlie last inferior molar. The length of the posterior four molars of the M. pungens equals that of the three true molars of the M. crassius- 1883.] 561 fCope. cuius; and the last true molar of the latter is half as long again as the pen- ultimate, while in M. pungens it exceeds it but little. The best preserved true molar is the second. Its most elevated cusps are the anterior and posterior inner, of which the anterior is subconic and more elevated. The anterior external cusp is crescentic in section, and sends crests to the supplementary, anterior, inner and the posterior anter- ior inner, both of which descend inwards. The posterior crest reaches the posterior base of the anterior inner cusp. The posterior external cusp is an elevated angle, sending crests forward and backwards. The former reaches the base of the anterior external cusp (not reaching the inner), while the latter passes round the posterior edge of the crown. As in M, pungens, it is convex posteriorly, and rises to the posterior internal cusp. In both species its appearance indicates that it performs an important masticatory function in connection with the su- perior molar. No cingula. Measurements. M. Length of bases of M. ii and iii ; (No. 3) 0125 base of M. iii ; (No. 3) 0070 Diameters crown M. ii; (No. 1) ^ anteroposterior ... .0056 i transverse 0050 Depth of ramus at M. ii ; (No. 1) 0100 Pekiptychus caeinidens Cope. Additional specimens of this species demonstrate that the last inferior molar has a different form from that of the P. rhdbdodon. While of the same length, it is narrower throughout, conformably with the smaller size of all the other molar teeth. Phenacodus calceolatus, sp. nov. This species is founded on fragments of the skull and limbs, with teeth, of a single individual. The teeth consist of two superior and four inferior molars of one side, and a smaller number of those of the opposite side. The teeth are of the size of those of the Phenacodus puercensis, and like that species, there is no median external cingular cusp of the superior molars. In these teeth the external basal cingulum is weak, but there is a strong anterior cingulum, distinct from any of the cusps. No internal cingulum. External cusps conical, well separated ; intermediate cusps rather large ; internal cusps rather large, close together, but deeply sepa- rated. The last superior molar is reduced in size. It has well developed anterior and posterior cingula, a weak external, and no internal cingula. The intermediate tubercles are rather large, and there is one large in- ternal tubercle. The heel of the last inferior molar is short, wide and rounded. The posterior tubercle is but little behind, opposite the posterior internal tu- bercle. The latter is separated from the anterior inner by a deep fissure, while the opposite side of the crown is occupied by a large median exter- Cope.] 562 [Jan. 5, 1883. nal cusp, which has a semicircular section. The large anterior cusps are confluent on wearing. No anterior cingulum in the worn crown. The crowns of the first and second true molars of the specimen are rather worn. They show that the posterior median tubercle is very indistinct and prob- ably absent. The bases of the smaller inner cusps are round, and on wear- ing unite with the larger external cusps. Of the latter the posterior is the larger. Anterior cingulum rudimental or wanting. No lateral or pos- terior cingula. The principal peculiarity of the lower dentition of this species and the one from which it is named, is the form of the third or fourth (probably third) premolars, both of which are preserved. They have a compressed apex, which descends steeply to the anterior base, with- out basal or lateral tubercle. The base of the crown spreads out laterally behind, and is broadly rounded at the posterior margin, so as to resemble the toe of a wide and moccasined foot. It is depressed, the surface rising to the apex from a flat base. Measurements. M. Diameters of second superior molar \ ^n™*!^-; ; Diameters of last superior molar \ l^Z^^:.::::. S Length of inferior true molars 0258 Diametersof M. -Z Diameters of theP.m.iii{-'-P—^^^ About the size of the P. puercensis. Note on the Mammalia of the Puerco and the Origin of the QUADRiTUBERCULATE SUPERIOR MoLAR. — It is now apparent that the type of superior molar tooth which predominated during the Puerco epoch was triangular ; that is, with two external, and one internal tubercles. Thus of forty-one species of Mammalia of which the superior molars are known, all but four have three tubercles of the crown, and of these thirty- eight triangular ones we may except those of three species of Peripiychus, which have a small supplementary lobe on each side of the median prin- cipal inner tubercle. This fact is important as indicating the mode of development of the various types of superior molar teeth, on which we have not heretofore had clear light. In the first place, this type of molar exists to-day only in the insectivorous and carnivorous Marsupialia ; in the Inseclivora, and the tubercular molars of such Carnivora as possess them (excepting the planti- grades). In the Ungulates the only traces of it are to be found in the molars of the Coryphodontidm of the Wasatch, and Dinocerata of the Proc ofAmerPlul Soc Vol XX.pa^e b63 PHENACODUS PRIMAFVIJB K T. Smclaiv&S 1. PHENACODUS PRlMAEVliS X. 2-3 .PERIPTYC} RHABDODON Vi . Dec. 16, 1882.] 563 [Cope. Bridger Eocenes. In later epochs it is chiefly seen only in the last supe- rior molar. It is also evident that the quadritubercular molar is derived from the tritubercular by the addition of a lobe of the inner part of a cin^ulum of the posterior base of the crown. Transitional states are seen in some of the Periptychidm (Anisonclius) and in the sectorials of the ProcyoriidcB. On the Brains of the Eocene Mammalia Phenacodus and Periptychus. By E. D. Cope. {Read before the American Philosophical Society, December 15, 1882.) PHENACODUS PRIM^VUS Cope. A cast of the cranial cavity gives the following as the general characters of the brain. The cerebal hemispheres are remarkably small, each one being less by one-quarter than the cerebellum. They are separated from the latter and from the large olfactory lobes by strong constrictions. The posterior one is 9ccupied by a thick tentorium. In like manner a wide groove for a robust falx separates the hemispheres above, a notch repre- sents the sylvian fissure, and the lobus hippocampi is quite large. The vermis of the cerebellum is quite distinct, and the lateral lobes are large. They are impressed laterally by the petrous bones as in various ruminants. The anterior columns of the medulla are not visible. There are traces of the convolutions on their hemispheres. The brain displays the following more special features. The olfactory lobes are as wide as long, and they diverge, having two external sides. In section they are triangular, presenting an angle downwards. The hemispheres are depressed, and wider posteriorly. They are well sepa- rated from each other and from the cerebellum ; so much so that it is quite probable that the copora quadrigemina are exposed. Their outlines are ^however not distinguishable on the flat surface which connects the hemispheres posteriorly. No further indication of sylvian fissure can be seen in the cast beyond an entering angle defining the lobus hippocampi anteriorly. The latter is prominent externallj^, and less so downwards. There are distinct indications of convolutions. There are three on each side above the sylvian convolution, and a fourth extends from the sylvian upwards and posteriorly below the posterior part of the third or external convolution. The sulci separating the convolutions are very shallow. The internal and external convolutions unite anteriorly, passing round the extremity of the median convolution. The space between this gyrus and the base of the olfactory lobe is only three millimeters. Cope.] 564 [Dec. 15, The cerebellum is larger than a single hemisphere. Its superior surface is somewhat flattened, and descends forwards ; the lateral boundary of this face is a projecting edge which rises behind to an angle of the vermis. The posterior face is shorter than the superior, and is vertical. It is sepa- rated by a space irom a very prominent lateral convolution, while the region of the flocculus is concave from the internal form of the ascending portion of the petrous bone. This concavity is open anteriorly. The base of the fifth pair of nerves is below its apex, and that of the sixth below the inferior extremity of the lateral convolution. The section of the medulla oblongata is a transverse oval ; its inferior face and that of the pons varolii, smooth. A deep fossa just anterior to the bases of the optic nerves. • Measurements of hrain". M. Length from vermis to olfactory lobes inclusive. , 070 " of olfactory lobes from above 015 " of hemispheres, from above 030 " of cerebellum from above 034 Depth of olfactory lobe 010 ** of hemisphere 023 " of cerebellum and medulla 026 *' of medulla at vermis 015 "Width of olfactory lobes at middle 030 " of hemispheres in front 044 behind 044 of cerebellum 036 " medulla at vermis 020 PERIPTYCHUS RHABDODON Cope. I have obtained a cast of the top and sides of the cerebral hemispheres, and the proximal portion of the olfactory lobes, from a skull of a Periptychua in which the teeth are preserved, and prove the species to be the P. rhab- dodon. The olfactory lobes are enormous, and the hemispheres small and very flat. The mesencephalon is entirely exposed. The cerebral hemispheres are very flat, and are only difibrentiated from the olfactory lobes, by a moderate contraction and depression, which forms the peduncle of the latter. Only the proximal part of the olfactory lobes is preserved, but this expands so as to be only a little narrower than the hemispheres. The peduncle has a ridge on the median line, and a shallow fossa on each side of it. The lateral outlines of the hemispheres diverge, and the widest part is posterior. There is no indication of sylvian fissure. The transverse sec- tion of the hemispheres would be a flat arch, but for the presence of a longitudinal oval protuberance on each of them, which dojiot quite touch the median line, and which have definite boundaries. If their limits determine the size of the cerebral hemispheres, then the latter are wider 1882.] 565 [Cope. than long, but they probably pass gradually into the mesencephalon be- hind them. These bodies remind one of the corpora olivseformia, and may represent the superior or median frontal convolutions. They are probably, however, not to be homologized with any convolutions, repre- senting rather the cerebral vault of the lateral ventricle. Posterior to them the flat surface descends gently without indication of copora quadri- gemina or other irregularity, and at a distance about equal to the length of the oval bodies, it begins to rise gently. The cranium is broken here, and no cast of the cerebellum was obtained. I may remark that the cranium from which this cast is taken is not crushed, and that it consists of parts of the parietal and squamosal bones only. The latter remain as far as the incurvature to the pterygoid pro- cesses in front of the glenoid cavity. Measurements of brain. M. Length from posterior rise to base of olfactory lobes 037 Length of oval bodies of hemispheres 018 Width of proximal part of olfactory lobe&. » 027 Width of olfactory peduhcles 021 Length from olfactory lobes to oval bodies of hemis- pheres...*. .005 Diameter of hemispheres at posterior part of oval bodies. .038 i>epth from sagittal crest to olfactory lobes .024 3EXPLAKATI0I^ OF PLATES. Plate I. Casts of the brain case of Pkenacodus primmvus Cope, natural size. Fig. 1. Lateral view. Fig. 2. Superior view. Fig. 3. Anterior view. Fig. 4. Posterior view, Plate II. Fig. 1. Brain of Phenacodus primceims, inferior view. Fig. 2. Cast of brain case of Periptychus rhabdodon, superior view* Fig. 3. Cast of brain case of Periptychm rkahdedon^ lateral view* Cope.] 628 [March 16, Fourth Contribution to the History of the Permian Formation of Texas. By E, D. Cope* {Bead before the American Philosophical Society, March 16, 1883.) PISCES. ECTOSTEORHACHIS CICERONIUS, Sp. DOV. The genus Ectosteorhachis Cope, is known up to the present time from ichthyolites, which do not exhibit the interior details of the structure of the skull. Several portions of crania having recently come into my hands, I am able to add some important features, and a new species, which I name as above. The base of the skull consists of ossified parachordals, which embrace the chorda dorsalis posteriorly and are continued for a short distance posteriorly as a tube. Anteriorly the chordal groove is open. Trabeculae not ossified. The cranial structure is an excellent illustration of a perma- nent embryonic type. Above and in front of the opening for the chorda, the neural canal enters the groove. The parachordals are subtriangular, presenting one angle forwards, and having the internal side that bounds the groove straight and longitudinally grooved. The anteroexternal side is oblique and nearly straight, and is overhung by the osseous roof of the skull. These characters are identical in both species. The E. ciceronius diflTers from the E. nitidus in having a narrower inter- orbital region, and in the possession of small tubercles of ganoine on the posterior parts of the superior surface of the skull. These are seen on the sides of the surface, and are quite small, not numerous, and •The third contribution can be found at page 447 Pioceedings ol the Society for 1882. 188S.3 629 [Cope. of various sizes and shapes. They resemble shining seeds. In E. nitidus these points are wanting, but there are rugosities on the postfrontal and pterotic regions of a radiating character, not found in E. ciceronius. Measurements. M. No. 1. Length of skull to occiput above (muzzle worn) 069 Interorbital width 014 No. 3. Length of osseous base of cranium (parachordal) 089 ** open median groove 023 Width of base at parachordals 036 " groove at apices of parachordals Oil " foramen notociiordae 0095 Found by Mr, W. F. Cummins. Gnathorhiza. serrata, gen. et sp. nov. This presumed fish is represented by some teeth which are processes of osseous bodies, which may be roots properly so called, or may be jaws. The osseous bases are shallow, and thickened on the free edge, which is directed obliquely aw9.y from the plane of the crown of the teeth. The teeth obtained are flat, and doubtless bilaterally symmetrical, though no complete pairs are preserved. The largest of these has a curved edge, and a branch extending posteriorly at right angles to it, joining it at a point at one side of its middle. The longer (and more curved) part of the convex edge, has two coarse angles ; the shorter part is finely denticulated, as is the transverse lamina. The principal edge is worn posteriorly by ase. The external convex face is marked by coarse and finer lines of growth, like those on corneous processes. A second form of tooth is not curved, but flat, 80 far as preserved. It has three coarse obtuse teeth. Two other toothed bodies resemble it. All the teeth are covered with brilliant ganoine on both sides. Measurements. M. Length of chord of larger tooth 010 " cross lamina 0055 Elevation of principal edge 006 with root 008 Thickness of root at base 003 The genus Gnathorhiza may belong to the Petalodont family, though I think it very doubtful. The characters of the roots of the teeth are more like those of sharks. BATRACHIA. Trimerorhachis bilobatus, sp. nov. Among the many specimens of animals of this genus which have passed through my handfi, I have not until now been able to select more than one Cope.J 630 [March 16 species, the 2. insignis. Mr. Cummins, however, now sends me parts of skeletons of four individuals, which present distinctive characters. Two of these include vertebral elements, and all embrace jaws and bones of the limbs and arches. The vertebrae present no important difference from those of T. insignis, but the surface of the intercentrum is not yet cleaned of a thin layer of matrix. The peculiar character of this species is most readily seen in the posterior portions of the mandibular ramus. The angle consists of two subequal tuberosities which are separated by a deep groove, instead of one prominent one. The external tuberosity is represented in the T. insignis by a small protuberance of the lateral enlargement of the external face of the ramus. The extremity of this tuberosity is in the T. bilobatus strongly honeycombed, and it is bounded below and externally by a groove which is faintly indicated in T. insignis. Above it, on the inner side, is another, shallow groove, from which it is separated by a sharp ridge. Both grooves are smooth. The superior one is wanting in T. insignis. The quadrate cotylus is more depressed externally than in T. insignis, thus making it more oblique. The internal fossa of the cotylus is not divided by a longi- tudinal groove, as it is in T. insignis The dental foramen is large, and is located as in the T. insignis. There is also an inferior longitudinal groove of the ramus as in that species. The surfaces preserved show that the sculp- ture is more marked in the T. bilobatus than in the T. insignis. Measurements. M. Depth of ramus at interior edge cotylus 026 Length • " from " " " 030 Width at " " " 017 *' of both tuberosities of angle 0125 Diameters of intercentrum' f anteroposterior Oil t transverse 021 Thickness of intercentrum 004 •The specimens described came from the same locality, and a different one from that which has produced the specimens of the T. insignis (Type No. 39, 1882). REPTILIA. ^ Pariotichus megalops, sp. nov. This reptile is known to me from a nearly complete, somewhat distorted cranium. A thin layer of matrix conceals the greater number of the teeth, so that the presence of canines cannot be demonstrated. Those which are visible are on the premaxillary and anterior parts of the maxillary bones. They are small, conic, slightly curved, acute and absolutely smooth. The muzzle is short and broadly rounded. The nareal opening is latero- superior, and is just above the principal convexity where the lores pass into the muzzle. Canthus nostralis rounded off. Interorbital region wide, convex in section, nearly plane anteroposteriorly, its width a little exceed- ing the diameter of the orbit. Orbit large and round, its diameter equal to 1888.J 631 [Cope. the length of the muzzle in front of it, obliquely measured, and one-half the distance from its posterior edge to that of the temporal roof (? squamosal bone). Posterior outline of skull above, truncate, surface slightly convex transversely. The premaxillary spines are short and wide, the nasals are also short and wide. The prefrontals and postfrontals form the superior edge of the orbit, excluding the frontals. The intercalaria (or ? pterotics) are very large; at the extemoposterior angle is a very small element in contact with the supra- occipital which may be the true intercalare. The supraoccipitals have considerable transverse extent, running out externally in narrow apices. All the bones of the cranium are sculptured in honeycomb fashion, the ridges radiating on some of the bones. That is, on the posterior parts ot the frontals and parietals and anterior part of the intercalare and squa- mosal. A groove follows the edge of the orbit, and turns inwards on the prefrontal bone, forming a rudimental lyra. External surface of mandible grooved below; superior part Concealed. Measurements. M. Width of skull between posterior angles. 018 Interorbital width 008 Axial length of skull 034 ** from muzzle to between centres of orbits. . .0096 Width of muzzle at nares 0095 Length from orbit to nostril 0035 Depth of skull posteriorly, to mandible 010 The superior part of the posterior region of the inner face of the dentary bone supports a patch of small obtuse teeth, which narrows forwards into the single row of the edge of the ramus. This patch is no doubt homolo- gous with that which is so largely developed in Pantylus. The surface of the cranium has been mostly weathered away in the type of Pariotichus, P. hrackyops, and I suspect that it is really sculptured and not smooth, as I originally stated. The P. megalaps differs from the P, braehyops in the larger orbit, the narrower interorbital space, and the smaller and more numerous teeth. Pariotichus and Pantylus and probably Eetocynodon must be referred to a special family, the P- 3 ? Hoplophoneus J Perissodactyla. Aceratherium, sp ~\ Aceratherium, sp y 3 Anchitherium, sp J Artiodactyla. Elotherium ramosum. Hyopotam la sp Oreodon, sp Oreodon, sp Oreodon, sp Leptomeryx, sp , Hypertragulus, sp . . . Total species 21 "Interesting features of the above catalogue are : The absence of Hyra- codon and Poebrotherium, so abundant in the beds of this age elsewhere ; the presence of fishes, not hitherto detected in them ; and the presence of the genius of tortoises, Trionyx. The latter genus has not hitherto been found in our Western lacustrine beds of later than Eocene age ; while they are abundant in our modern rivers. This discovery partially bridges the interval. The same is true of the fishes mentioned, which represent the order Nematognathi." PROC. AMER. PHILOS. SOC. XXl. 114. 2b. PRINTED OCTOBER 30, 1883 Cope.] 308 [Dec. 7, On the distribution of the Loup Fork formation in New Mexico. By E. I). Cope, {Read hefore the American Philosophical Society, Decemher 7, 1SS3.) In his report on tlie Geology of New Mexico to the Secretary of the Interior, by Dr. F. V. Hayden, in 1869, this eminent geologist described the Santa Fe marls in their principal physical features. In 1874, in my report to Capt. George M. Wheeler, U. S. Engineers, I showed that this formation is a member of the Loup Fork division of the Miocene Tertiary, a conclusion clearly deducible from the remains of vertebrata which it contains. An illustrated report on the latter was published in the fourth volume of the report of the United States Geographical and Geological Survey, W. of the 100th meridian, Capt. G. M. Wheeler in charge(1877). Since that time the writer has made several visits to parts of New Mexico not previously explored, and I am able to show that the Loup Fork for- mation has a much wider distribution in that Territory than has hitherto been supposed to be the case. In descending the Rio Grande, beds appear on the west side of the river, which strongly resemble those of Santa Fe. They extend along the eas^^ern base of the Magdalena mountains, and as far south as Socorro, in considerable extent and thickness. South of Socorro they appear, but less extensively. The eastern part of the plain which lies between the Rio Grande and the Mimbres mountains is composed of beds of this age where cut by the grade of the Atchison, Topeka and Santa Fe railroad, west of Hatch station. AVest of the Mimbres mountains the valley of the river of the same name is filled with debris of the bed of eruptive outflow which once covered the country, as far as traversed by the railroad from Deming to Silver City. Its age I could not ascertain. A great display of the Loup Fork formation is seen in the drainage basins of the heads of the Gila river. In traveling westward from Silver City, its beds first appear in the valley of Mangus creek, which enters the Gila "from the east. Crossing the Gila, the mail route to the west passes through the valley of Duck Creek, which flows eastwards into that river. Though bounded by eruptive hills and mountains and their outflows, the valley was once filled with Loup Fork beds, which have been extensively eroded, the principal exposures being on the north side of the valley, forming the foot hills of the Mogollon range. On the divide between the waters of the Gila and San Francisco rivers the formation rises in bluffs of 300 feel elevation. The descent into the valley of the San Francisco brings to light a still greater depth of this deposit. The valley which ex- tends from the canj'On which encloses the river south from the mouth of Dry creek to the Tulerosa mountains on thenorth, and between theMogol- lons on the east and the San Francisco range on the west, was once filled with the deposit of a Loup Fork lake. This mass has been reduced by the ■erosive action of the San Francisco and its drainage, to a greater or less ias3.] 309 [Cope. extent, as it has been protected b}' basaltic outflows or not. When so pro- tected, the river flows through comparatively narrow canyons. Where the outflow is wanting, the valley of the river is wider, and the Loup Fork formation remains as wide grassy mesas which extend to the feet of the mountain ranges. The age of these beds would have remained problematical but for the fortunate discovery by Mr. Robert Seip, of the skull of a species of Rhi- noceros of the typical Loup Fork genus, Aphelops. It is apparently the A. fossi'jer Cope, a species abundant in the Loup Fork beds of Kansas and Nebraska. It was found near tlie mouth of Dry creek in a conglom- erate bed of the formation. In the valley of the San Francisco the Loup Fork beds reach a thick- ness of 500 feet, and consist of sand, clayey sand, soft sandstone, and conglomerates of larger and smaller pebbles of eruptive material, having a near resemblance to those of the region of Santa Fe. Second Addition to the Knowledge of the Paerco Epoch. By E. D. Cope.'* {Read before the American Philosophical Society, December 7, 1SS3.) Recent collections from the formation above-named, include many finer specimens than have been previously obtained. Skulls of several species in calcareous concretions were received, so that their characters can be de- veloped more fully than heretofore. I mention especially Deltatherium fandaminis ; Periptychus rhabdodon and P. coarctatus ; Haploconus linea- tus ; H. entoconns ; Anisonchus sectorius ; Protogonia pUcifera ; Mioclmnus turgidus, M. ferox, M. mbtrigonus and M. cuspidatus, sp. nov. Some species hitherto rarely seen, prove to be abundant, as Hemithlceus kowalemkianus, Protogonia plicifera, Mioclaenus minimus and M. subtrigonus. With the additional species now described, the number ot Mammalia from the de- posit of the Puerco epoch amounts to seventy-four species. DiDYMICTIS PRIMUS, Sp. nOV. That the genus Didymictis existed during the Puerco epoch, has been already demonstrated by the discovery of the D. haydenianus Cope. This species is of aberrant form however, so that it remained to prove that the typical form had appeared so early in Tertiary time. This is now shown to have been the case by the discovery of the present animal, which is allied to the D. leptomylus of the Wind river and Wasatch epochs. The Didymictis primus is known from two maxillary bones with teeth, *The "First addition" appeared in tlie Proceedings of tlie American Philo- sophical Society for 1H83. beginning at page .545. Since that date I liave described in the Proceedings of the Philadelphia Academy, 1883, p. 108, the following spe- cies : Periptychus ro'irctatus, Fantolanibda cavirictuy, Zeiodon yracilis (g. n.) and Conoryctc.s ditrigonus. Cope.] 310 [Dec. 7, and a part of a mandibular bone with the last two molars in place, all be- longing to different individuals. The inferior sectoiial tooth is much like that of the D. leptomylus, but the tubercular is only two-thirds as long, and is not only absolutely, but relatively narrower posteriorly. It has the usual three cusps in a reduced condition. In the first superior true molar the external cusps are conical, and there is a small cusp between the ante- rior one and the produced anterior angle of the crown. There is an ante- rior intermediate tubercle, but no posterior one. The cingulum does not extend all round the inferior base of the crown, as it does in D. protenus. The sectorial has a distinct anterior basal conic lobe. The internal lobe is in transverse line with the last named, and is conical and not large. Measure7nents. M. Diameter inferior sectorial | anteroposterior 0138 transverse 0055 Diameter inferior tubercular | anteroposterior 0050 *. transverse 0033 Depth of ramus at M. i 0098 Diameter superior sectorial (No. i) j anteroposterior. . . .0110 t transverse 0060 Diameters superior sectorial (No. 2) { anteroposterior . 0050 transverse 0090 The fourth specimen is especially important as presenting almost the entire dentition including canines and incisors, and the anterior part of the skull from the line of the coronoid process of the mandible. The specimen shows that the species differs from the species of the Wasatch period with oval inferior tubercular, in the absence of the posterior cutting lobe of the third, and probably fourth inferior premolar. The corresponding superior premolars are also simple. The first premolars in both jaws are one- rooted. The canines are long and acute, and are directed vertically. Both have flat facets on their external (the only visible) faces : on the superior canine I count four lateral, and one nearly anterior. On the inferior I see three lateral and one nearly anterior. There are three small superior in- cisors, of which the first is the largest, and has a subconical crown. The infraorbital foramen is large, and is above the anterior border of the supe- rior sectorial. Measurements. M. Length of superior dental series to front of canine 041 " " crown of superior canine Oil " " superior true molars 0105 Depth of ramus at inferior sectorial 0090 In its simple premolars this species agrees with the D. 7iaydenia?iU8, and is more primitive than the Wasatch species. Tkiisodon rusticus, sp. nov. Founded on a portion of the mandible which supports the first two true molars and part of the last premolar. The species is of the type of 2\ 1883.] 311 [Cope. levisamis, but is much larger. I give here a synopsis of the species of the genus, so that its affinities may be better understood. In general, the genus Triisodon is characterized by the rudimental character in the infe- rior molars of the anterior cusp. It is thus like Ictops, but differs in having the fourth premolar different from the true molars and like the premolars. From Mioclmnus it difiers in having the anterior and posterior cusps of the inferior molars unequal ; the anterior forming together an elevated crest with two apices, while the posterior are low, and ou the borders of>a heel. I. Cusps of inferior molars compressed. Anterior cusp very low T. quivirensis. II. Cusps of inferior molars not compressed. Anterior cusp very low ; T. rusticus ; T. levisanus, and T. assurgens. Anterior cusp as high as other anterior cusps to which it is closely united. T. conidens and T. heilprinianus. In dimensions the T. rusticus is about equal to the T. quivirensis, thus exceeding the other species excepting the T. conidens. The interior ante- rior cusp is nearly as elevated as the exterior, and is united with it nearly to the apex ; the anterior cusp is a tubercle which projects forwards from its anterior base. The heel of the tooth is wide, and is rounded poste- riorly, and supports three tubercles, an external, a posterior and an inter- nal, all in contact with each other. On the second true molar the internal anterior tubercle presents a slightly projectiiig edge anteriorly and poste- riorly, which bounds a shallow vertical groove of the mass which repre- sents their united bodies. This is not apparent in the first . The enamel is smooth, but the animal is rather old. D. Baldwin, discoverer. Triisodon assurgens, sp. nov. This is the least species of the genus, and resembles in its inferior denti- tion the species of Diacodon. It is very much larger than the JD. alticuspis, the larger species of that genus, which is found in the Wasatch formation. The T. assurfjens is known from a mandibular ramus which supports the last four molars, the last premolar having lost its principal cusp. The peculiarity of the true molars is seen in their generally more produced character ; the anterior cusps are higlier and the heels are longer. The an- terior cusp is very small and basal ; the principal anterior cusps are united Measurements. C anteroposterior . . Diameters of m. i s transverse M. .0123 .0068 .0068 .0038 .0137 .007 .007 .0063 r anteroposterior Diameters of m. ii \ transverse Cope.] 312 [Dec. 7, to near tlieir free summits. There are the usual low marginal tubercles on the heels. That of the fourth premolar is a short simple edge. Measurements. M. Length of four molars on basis 028 '* " three true molars 0212 " "second true molar 008 Elevation of cusps of molars 0045 Length of last true molar .0067 Width of last true molar 0030 Elevation of last true molar in front 0035 Found by D. Baldwin. MiocL^NUS cuspiDATus, sp. nov. The species of this genus known to me are, with the present one, nine in number. They range in size from that of a rat {M. minimus) to that ol a wolf {M. ferox). The general osteological characters of the last named species are best known, and are described in the Proceedings of the Amer- ican Philosophical Society, 1883, p. 547. In two of the species the supe- rior dental series only is more or less known, and one species rests on mandibular dentition only. In the remaining seven species the dentition of both jaws is more or less known. The species may be arranged in groups as follows : I. The posterior heel of the second inferior molar bordered by a curved edge or crest. a. Posterior cingulum of superior true molars obsolete ; M. minimus, aa. Posterior superior cingulum weak ; M. turgidus. aaa. Posterior superior cingulum large, angulate ; M. corrugatus ; M. ferox. II. The posterior heel of the second inferior molar supporting a cusp. a. Posterior inner cusp of superior molars small, present on m. ii only ; 31". cuspidatus. aa. Posterior inner cusp large, present on m. i and m. ii ; premolars small, M. subirigonus ; premolars large, M. opisthacus [HemitJilceus mihi oUrn). III. Second lower molar unknown. J/, protogonioides, and M. mandih- ularis. The supposed M. baldwi/ti, resembles closely the species of Hemitlilceus. It is probable that two genera are here included under the head of Mio- clcenus. If the character is permanent, these will be distinguished as fol- lows : Third superior premolar with internal tubercle Mioclcenus. Third superior premolar without internal tubercle Oxydmnus. The species of Mioclcmvs are M. turgidus (type) ; and very probably M. opisthacus, minimus and M. suUrigonus ; but the diagnostic tooth has not been seen in them as yet. The species of Oxychmus are : 0. cuspidatus 1883.] 313 [Cope. and 0. corrugatm ; and very probably, 0. ferox. The position of tlie M. prolog onioide 8, 31. haldwini and M. mandibulai'is is uncertain, tliough the last two are probably Oxyclceni. The Miodmnus cuspidatus is distinguished among its congeners, by the transverse character of its superior molar teeth, that is, by the relatively smaller anteroposterior diameter as compared with the transverse ; and by the prominence and acuteness of their principal cusps. They thus stand at the opposite extreme of the genus from the 3f. turgidus, where the teeth are characterized by the robustness and obluseness of the cusps, although in the triangular basis of the second superior molar they agree. The ex- ternal cusps are compressed cones, and in contact at the base ; the inter- mediate tubercles are small and distinct. The internal cusp is large and prominent. The base of the fourth premolar is T-shaped, and is as long as wide. Its internal and exteijnal cusps are well developed. The cingu- lum of the true molars is complete all round on the last one, and on the two others except at the internal base, where it is interrupted. The second molar only displays a posterior inner tubercle of the cingulum, which is small, and does not give a truncate interior outline of the crown, charac- teristic of M. opisthacus, M. ferox, etc. On the ms. i and ii, the cingulum is expanded at the external angles of the crown, most so anteriorly. The an- terior expansion rises in a low cusp in the P-m. iv. The enamel is smooth. This species need only be compared with M. opisthacus and M. suUri- gonus, which are of about the same size. Passing by the differences already mentioned in the table, the fourth premolar has a difierent form from that of the M opisthacus. In the latter it is narrower and more trans- verse, and with larger conical cusps, much as in M. turgidus ; in the pres- ent species it has the trilobate outline seen in M. suhtrigonus. As to the latter species, the teeth are wide, and the cusps smaller and separated at at the base, and the cingulum is crenate and lobate, in a manner quite different from the smoothness and compactness of structure seen in the M. cuspidatus. Measurements. M. Length of base of last four superior molars .026 " " " three true molars 019 Diameters of P-m. iv / anteroposterior 006 ^ transverse 004 Diameters of m. i$ anteroposterior 006 c transverse 006 Diameters of m. h i anteroposterior 0064 t transverse 008 Diameters of m. iii | anteroposterior 0045 '-transverse 006 D. Baldwin, discoverer. Chriacus thuncatus, sp. nov. The genus Chriacus m. was characterized in the Proceedings of thr PROC. AMEIl. PHILOS. SOC. XXI. 114. 2n. PRINTED JANUARY , 1884. Cope.] 314 ' [Dec. 7, Academy of Philadelphia, 1883, p. 80, and two species were mentioned, C. pehidens (type) and C. angulatus. Tlie former of these is from the Puerco, the latter from the Wasatch formation ; the former is the larger species ; the latter quite small. I now add two species to the genus which are intermediate in dimensions between those already known. I. Posterior cingulum of superior molars with large tubercle. Large species ; C. pelvidens ; small species, C. truncatus. IT. Posterior cingulum with small tubercle ; small species ; G. angulatus. III. Posterior ciugulum without tubercle ; small species ; C. simplex. In the C. truncatus the posterior singular (inner) tubercle reaches the largest development, but is not present on the cingulum of the last supe- rior molar. The anterior cingulum is weak on that tooth and on the first true molar, but on the second it is thickened into a small anterior or inner tubercle. This with the posterior inner gives the crown a truncate internal outline, as is also the case in the C. pelvidens. The intermediate tubercles are distinct, and the external cusps are separate at the base. An external cingulum. The fourth premolar has a triangular base ; a single compressed external cusp, and a small acutely conical internal one. The internal tubercle is small and acute on the third premolar. The second pre- molar is small and probably one-rooted, and it is possible that there is no first premolar. The canine is directed vertically downwards, and the base of the crown is oval. Besides the considerably smaller size, the posterior internal cusps are relatively larger than in C. pelvidens. Measurements. M. Length of superior dental series including canine 039 Length of true molar series 014 Diameters P-m. iii anteroposterior 004 transverse 003 Diameters P-m. iv / anteroposterior 004 t transverse 005 Diameters M. ii | anteroposterior 005 transverse 0064 / anteroposterior 0033 ^ transverse 005 Diameters m. iii Two individuals from New Mexico, D. Baldwin. Chriacus simplex, sp. nov. This species is represented by a part of the left maxillary bone, which supports the true molars except a part of the last one ; and by parts of the mandible, with the first and second true molars, and perhaps one of the premolars. The true molars are about the size of those of the G. truncatus, but of verj^ different detailed structure, as already pointed out. The pos- terior cingulum is stronger than the anterior, but does not support a trace of a cusp, and they do not unite on the inner fMpeof the crown. External 188;{.] 315 [Cope. ciiigulum present. External cusps rather small, sei)arate. Intermediate cusps present ; V large and distinct. Enamel smooth. The inferior true molars support Vs ; in the second the anterior is smaller and is more elevated than the posterior. The latter is continued as a raised posterior, and partly interior border of the heel, v^ithout promi- nent cusp. Tlie crown has a distinct external and a very faint internal cingulum. In the supposed first true molar, the anterior V is more pro- longed anteroposteriorly as in the corresponding tooth of 31iocUmus ferox, etc., and the fourth premolar of Phenacodus j^rinuevus. The anterior cusp is the lowest. The heel supports three low cusps, of which the ex- ternal has a crescentic section, and the posterior is the smallest. It is probable but not certain that the fourth premolar has an internal cusp, as the tooth, presumably this one, is injured at that point. Should the internal cusp be absent, this species cannot be referred to Ghriacus. Measurements. M. Length of superior true molars 0135 f anteroposterior 005 Diameters of first true molars \ transverse 006 anteroposterior 0053 Diameters of second true molars transverse 007 , • w 1 f anteroposterior 0034 Diameters of third true molar < ^ . „ „ I transverse 006 ^. « ^ . ^ . . if anteroposterior. . .005 Diameters of first inferior true molar < , ^ _ _ i. transverse 0035 ^. ^ -, . ^ . X 1 /anteroposterior .0056 Diameters of second inferior true molar i , •-transverse 0043 D. Baldwin, discoverer. Tricentes crassicollidens, gen. et sp. nov. Char. gen. This genus is Chriacus with only three premolars in the su- perior, and probably inferior series. The canines are well developed, and lateral in position, leaving space for small incisors, thus differing from the genera of the Mixodeciidce, 3Iixodectes, Microsyops, ^ndCynodontomys, on the one hand, and from Necrolemur on the other. It has, so far as known, the dental formula of several genera of typical Lemuridse, but differs from these in the following points. The orbit is open posteriorly ; the inferior molars have the anterior triangle of three cusps ; and the fourth inferior premolar has an interior cusp. I have demonstrated the last mentioned characters on the ty.pe, T. crassicollidens only, but suspect its presence on some or all of the other species. In their details the superior true molars are like those of Mioclainus, as distinguished from those of Pelycodus. To this genus belongs the Mioclcenus subtrigoiuts, and probably, from the small size of its fourth premolar, the M. hucculentus. I add to these three a fourth, T. imcquidens, and remark that it is yet uncertain how many pre- molars are present in the Chriacus simjylex. Should the latter possess three only, it will be properly referred to Tricenles. Cope.] 316 [Dec 7, These species differ as follows : I. Posterior ciugulum of true molars i and ii, wide, rising into a small cusp. Length of true molars, M. .0155 crassicollidens. II. Posterior cingulum distinct, thickened inwards. Length of true molars (m. ii inferential) .0175, crowns narrowed, trans- verse lucculentus. Length of true molars .0170 ; crowns quadrate suhtrigonus.* Length of true molars .0135 ; crowns narrowed, transverse ( Chriacns) simplex. III. Posterior cingulum weak, disappearing inwards. Length of true molars.. 0105, crowns transverse except the third, which is very small ■ .inceguidejis. Char. Specif. The Tricentes crassicollidens is about tlie size of the Chriacus truncafus and resembles it a good deal. The latter has, however, a more transverse form of true molars, as compared with the present spe- cies, where the form is subquadrate. In the present animal the premolars are smaller, and if the third (second present) has an internal cusp, it is much more insignificant than in the C. truncatus. These two species and the Mioclcenus opisthacus resemble each other in the similar size, and in the true molars having the posterior inner cusp more distinct tlian in oilier spe- cies. They difler in the dimensions of their premolars, those of the M. opisthacus being the largest, and those of C. truncatus being intermediate in size. In the T. crassicollidens the anterior cingulum is also distinct. The external cusps are conic, and are well separated, and the internal V is dis- tinct. The internal cusp of the fourth premolar is small and compressed, so as to be transverse. The base of the third premolar is triangular and much longer than wide. All the superior molars, except the first premo- lar, are furnished with an external cingulum, which rises into a more or less distinct apex at its anterior and posterior angles. The first premolar is a simple cone. The alveolus of the canine tooth is of large size. The last true molar is not much reduced, and the first is as large as the second. This is not the case with the 2\ bucculentus, where the first is considerably smaller than the second. Measurements. M. Length of dental series to canine, exclusive 036 "diastema 006 " " premolar series 0143 " " true molar series 0152 ■p.. , • r anteroposterior 0043 Diameter of P-m. iv ^ ^ t transverse 0042 * There may be two species confounded under this name. A specimen figured in Vol. Ill of the final (4to) Report of the Hayden Survey, Plate XXIV, f, fig. 4 ihas four inferior premolars, all simple. 1883.] 317 [Core. Mea(surement8. M. ix. , f . r anteroposterior 0058 Diameter of :M. i ^ * _ t transverse OOoO Diameter of M. in j anteroposterior 0080 i. transverse OO-lo A pair of mandibular rami, found on the same day, and at or near the same phice, probably belong to the same species, if not to the same animal, they support all the teeth, but only the P m iv and the M. i and ii have yet been disengaged from the matrix. The P-m. iv is rather large and robust, and has a short wide heel, and an anterior cusp which leaves the main cusps half way to the apex, or at the same elevation as the internal cusp. The anterior three cusps of the true molars are elevated above the heel, and the anterior is nearly median, forms no blade with external ante- rior, and is smaller than the anterior internal cusp. The heel is well de- veloped, and its borders rise in two obtuse open Vs, whose apices look away from each other. The internal supports two cusps, the external, but one. No cingula ; enamel smooth. Measurements of inferior teeth. M. -r^. , e T> ^ \ f anteroposterior 0060 Diameters of P-m. iv < ^ C transverse OOdD Diameters of m. u | anteroposterior 0050 t transverse 0089 Length of bases of m. i and m. ii 0110 From Upper Puerco ; D. Baldwin. Teicentes in^quidens, sp. nov. This species is represented by two mutilated crania, obtained on the same day and near the same locality as the preceding species. One of ^liese, which I select as type, embraces the muzzle and palate anterior to the posterior border of the maxillary bone. Besides its inferior size, other characters distinguish this species. The simplicity of the superior molars is seen in no other, and the very reduced size of the third superior molar is not found in any of its allies. This is correlated with an oblique reduction of the maxillary bone behind, which gives the second true molar an oblique external border instead of the longi- tudinal one seen in the other species. The external cusps of the molars are conic, and are not in contact at the base. The internal cusp is also conic, and is larger than the external. The internal cusp of the fourth pre- molar is large. It is probable that the third premolar supports an internal cusp, as the crown base is as wide as long. The premolars are spaced in this species, as in the last, but the diastema is shorter than in the T. crassi- collidens, not exceeding the premolar interspaces. The external cingulum is quite weak. The canine alveolus is large. The incisors are wanting^ but the preraaxillary region is wide. The inferior dentition is unknown. Cope.] 318 [Dec. 7, Measurements of superior teeth. M. Length of dental series, including canine 0372 " from canine to m. i, exclusive 0130 Length of true molar series 0100 Diameter of P m. m | anteroposterior 0028 i transverse 0025 Diameters of P-m. iv / anteroposterior 0030 I transverse .0042 Diameters M. i i anteroposterior 0038 t transverse 0048 Diameters M. ii \ anteroposterior 0039 I transverse 0059 Diameters M. iu / anteropo^erior 0015 t transverse 0024 Upper Puerco ; D. Baldwin. Indrodon malaris, gen. et sp. nov. Char, gen. Family Anaptomorphidae, suborder perhaps Lemuroidea, as indicated by the dentition only. It differs from Anaptom'^rphus in three points. First, there are three superior incisors ; second, the first (third) premolar has no internal lobe ; and third, there is a distinct posterior in- ternal tubercle on the first and second superior molars. The animals of the Eocene period of the family of the Adapidm, may be- long to the Lemuroidea, but the evidence which I have derived from the feet of Pelycodus^ has led me to refer themf to the Insectivorous division of the Bunotheria, to the neighborhood of the Tup£Eidge and Erinaceidae. At the same time I retained provisionally the genera with three and two superior premolars in the suborder Lemuroidea, although the foot struc- ture of these extinct genera is yet unknown. I also indadvertently defined the Lemuroidea as having quadrituberculate superior molars, a character which I well knew to be wanting in various extinct and recent genera where they are tritubercular. Two families were proposedf for the Eocene lemuroids, which are defined as follows : Superior premolars three Mixodectidm. " " two Anaptomorphidce. The genera of the first named family are defined as follows : I. Canine teeth large and lateral, well separated. First superior premolar without internal lobe ; superior true molars tritubercular with cingula Tricentes. II. Canine teeth median in position or much reduced in size. «. Last inferior premolar without internal tubercle. Inferior premolars all one rooted ; canine and incisor small. .Necrolemur.X * Report of U. S. G. G. Survey W. of 100th Mer. G. M. Wheeler, iv, p. 140. t Proceedings Academy Natural Sciences, of Piiiiadelphia, 1883, p. 78-80. ; FlUiol Rech. Phosph. Quercy. 188.^] 319 [Cope. First premolars only one rooted ; canine small ; incisor very large Mixodectes,* aa. Last inferior premolar with internal tubercle. A very large ? canine ; first premolar only, one-rooted Microeyops.f A very large ? canine ; first and second premolars both one- rooted Cynoclontomys.X The genera of Anaptomorphidse, which on denial characters includes Indrodon, differ as follows : «. Incisors three. First superior incisor without inner lobe ; posterior inner tu- bercle present on first and second tubercle Indrodon. aa. Incisors two. First superior incisor with inner lobe ; no posterior inner tubercle on superior molars Anapiomorphus. The superior dental formula of Indrodon is I. ^ ; C. ^ ; P-m. ^ ; M. ^. The canine is compressed and acute ; the third premolar is compressed conic, and has two roots. The fourth premolar has but one external cusp. The external cusps of the true molars are conic and acute, and are con- nected with the internal cusp by ridges which form a V. Posterior inner cusp distinct on ms. i and ii, a part of the posterior cingulum. Intermedi- ate tubercles present, small. The superior incisors are well developed, and display no tendency towards the rodent type. A portion of lower jaw adheres to the skull, and may belong to the same animal. It supports the last two molars. These have two anterior, opposite, approximated cusps. The heel of the penultimate molar is rather large, and has a raised edge, which develops low tubercles at the angles. Char. Specif. The first and third superior incisors are a little larger than the second. Canine preceded and followed by diastemata, each of which is 1.5 times as long as the long diameter of the base of the crown. Premolars separated from each other and from the first true molar by in- terspaces half as long as the diastema. Neither tooth has any basal tubercles, but the posterior has a weak external cingulum, which is stronger posteriorly. The internal cusp of the same tooth is anterior, is acute and elevated. The superior true molars have a strong external cingulum, which rises into a small tubercle opposite the space between the external principal cusps. Of the latter, the anterior is a little more conic than the posterior, and both are well within the external border. On the last molar, the posterior external cusp is continuous with the external in- termediate tubercle, and forms a cutting edge within the posterior margin of the crown. The posterior inner tubercle is rather large, and projects further inwards than the apex of the anterior V on the second true molar, but not so far as in the species of Anisonchus and Haploconua. *Proceedings American Philosophical Society, 1883, p. 559. fLeidy Report U. S. Geol. Survey, Terrs. I. ; Cope, Pal. Bull., No. 34. Cope.] 320 [ Dec. 7, The surface of the cranium is too much obscured by cracks and fihnsol matrix to permit a view of the sutures and foramina. The face is wide, as the posterior part of the maxillary and the malar bone are expanded out- wards. I have not yet been able to ascertain the condition of the orbit posteriorly. The mandibular ramus is rather slender. Measurements. M. Length of dental series from posterior base of. i iii 0348 " " bases of superior incisors 0060 ** from i iii to P-m. iii, exclusive 0074 ** of premolars on maxillary bone 0060 " base of P-m. iii 0020 P-m. iv 0028 Width " 0038 T^. ^ . r anteroposterier 0030 Diameters m.i < ^ ^ I transverse 0033 T^. ^ .. f anteroposterior 0033 Diameters m. n < , * I transverse , 0040 Diameters m. iii | ^'"te'-Oposte"'"- 0030 1 transverse 0040 r^. , ' c • r anteroposterior 0033 Diameters inferior m. iw , ^ I transverse 0030 Depth of ramus mandibuli at m. ii 0070 The skull is about the size of that of the Bassaris astuta. D. Baldwin, discoverer. The discovery of this type in the Puerco formation is a fact of interest. In the shortening of its dental series it is the most specialized genus of the epoch, while the forms of its true molars are like those of the simpler Creodonta, and more specialized than those of Anaptomorphus, and the lemurs generally. In the simplicity of its premolars, however, it main- tains the general character of the Puerco fauna, and is more primitive than the forms just named. Its nearest ally of the Puerco yet known is Chriacus. Anisonchus agapetillus, sp. nov. This species is founded on parts of six mandibular mmi, none of which has more than four continuous molars in position, including the last. It is not entirely certain that these belong to a species of Anisonchus, because the superior molar teeth by which that genus is distinguished from Haplo- conus and Hemithlcuus, are wanting. The inferior molars have the ante- rior inner cusp moderately well developed, as in Anisonchus gillianus. The crowns of the true molars consist of two Vs ; of which the posterior base of the posterior one, is rendered irregular by the presence of a small posterior median tubercle. Of the anterior pair of cusps, the external is a little the more elevated, and the internal is more elevated than any ot the posterior ones. The internal posterior as well as the external posterior 321 [Cop*. cusp has a V-shaped section, because its anterior border is continued as an oblique ridge to the base of the anterior internal cusp. Internal cingula none ; a slight one on the external base of the large anterior external cusp. The heel of the third true molar is well developed, and rises into an acute cusp. That of the fourth premolar is short and flat. The anterior cusp of the same is basal and rudimental. This tooth is not enlarged as is usually the case in the Periptychida;, and it first here differs from these animals, and agrees with the unguiculate types in that its lateral faces are unequally convex. Measurements. M. Length of last four molars on base 014 " fourth premolar 0035 Elevation of " " 0038 Length of second true molar 0031 Width " " " (greatest) 003 Length of third " " " 004 Width " " " " 0028 Depth of ramus at second true molar 007 Anisonchus cophater, sp. nov. A mandibular ramus supporting three molars, two oJT them true, is all that I have seen of this species. Its proportions are the same as those of the A. agapetillus, that is, much smaller than the A. gilUanus, and the single premolar is much more like that of other species of the genus. The true molars differ from those of the A. agapetillus in two strong characters. First, the internal posterior cusp is inside the rim of the heel of the crown, that is, outside the bordering edge, and is therefore very distinct from the posterior median cusp. It is a sharp cone ; secondly, there is a cingulum extending from this cusp round the internal base of the internal anterior cusp. There is also one at the base of the external anterior cusp, which con- tinues to the heel only on the last inferior molar. The posterior heel is rela- tively wider, and the anterior Y relatively more contracted, than in the A. agapetillus. The anterior tubercle is moderately developed at the anterior base of the anterior V. The third or fourth premolar is equilateral, and larger than the true molars. It has a short apiculate heel, and a rudimental anterior basal tubercle. Measurements. M. r- /anteroposterior 0032 Dia™etersofm.ii]'~'^'— OO^O (yertical 0025 (posterior 0013 C anteroposterior ,0043 Diameters of P-m. iii or iv -j vertical (restored apex). . . .0040 . (transverse 0023 D. Baldwin, discoverer. Chirox plicatus, gen. et sp. nov. Cha/r. gen. These are known from three superior molars ; viz : the last PROC. AMER. PHILOS. SOC. XXI. 114. 20. PRINTED JANUARY , 1884. Cope.] 322 [Dec. 7, premolar, and the second and third true molars. The fourth premolar has two external, and one internal cusps, and the true molars have four cusps each. The cusps are of peculiar form. The second true molar resembles a convex body which has been divided by two cuts at right angles to each other, from which the quarters thus produced has spread away from each other subequally. The external faces of the cusps are convex. The apices are acute. The last superior molar is larger anteroposteriorly than transversely. The fourth premolar (supposed) is two-rooted. These molar teeth remind one of the inferior molars of Plilodus, through they differ much from them. The genus is probably nearer to Gatopsalis, and belongs to the Marsupial order. The presence of only two series of cusps in the superior molars, distinguishes it from these genera, which have presumably three series of such cusps. Lemoine has shown this to be the case in JSeoplagiaulax. Char, specif. The external cusps of the fourth premolar are flattened on the external side, and lean a little inwards. The internal cusp (proba- bly homologically the anterior) is opposite the anterior external, and has a convex internal face. Its apex is acute and compressed ; the apices of the external cusps are trihedral and acute. The cusps of the second true molars are more widely separated trans- versely than anteroposteriorly ; that is, the longitudinal fissure is wider than the transverse. The apices are all acute, the internal trihedral, the external more compressed. The transverse diameter of the last true molar is smaller than that of the second true molar, while the longitudinal is nearly the same. The crown projects convexly posterior to the posterior pair, and there is a small tubercle at the anterior base of the external anterior cusp. None of the teeth preserved display cingula. The bases of the crown are smooth, but the cusps are sharply and finely parallel-grooved on their external faces. Measurements. M. Diameters of P-m. iv Diameters of m. ii \ r anteroposterior , 0030 I transverse 0038 c anteroposterior 0033 t transverse 0035 ^ - ... r anteroposterior 0035 Diameters of m. iii i , ^aoa I transverse 0030 D. Baldwin, discoverer. Gatopsalis fissidens, sp. nov. This Marsupial is represented by a portion of the lower jaw which sup- ports the molar teeth. The first, which is probably the fourth premolar, is represented only by its single root, which fills a round alveolus near the anterior base of the first true molar. In size this species is intermediate between the small C. foliatus and the large G. pollux. The first molar is the longer and narrower, and the 1883.] 323 [Cope. second tlie shorter and wider, as in the known species. The first molar dillcrs from that of both the latter, in having the tubercles of one side sepa- rated nearly to the base. These tubercles are conic, and not flattened as in G. foUatus and C. pollux, and the two rows are separated by a distinct valley, as in the first named. There are five tubercles on one side, and four on the other side of the crown, and in addition, two small cusps at the an- terior extremity of each row, and another at the posterior extremity of one of the rows. These additional cusplets are not present in the other species. The last molar is relatively wider than in the other species. Its crown is a good deal worn, but there are probably more than two rows of tuber- cles, as there are some appendicular rows on one side of the crown at least. 3£easurements. Diameters m. i | ^^^^^'^P^^^^^^^ t transverse Diameters M.ii^^^t^^^P^^^^^'^^^ c transverse The Upper Puerco ; D. Baldwin. General remarks on the characters of the Mammalia op the Puerco Epoch. I have already called attention to the fact that the Mammalia of the Puerco epoch possess, with but few exceptions, superior molar teeth whose crowns include only three of the component tubercles of the normal mammalian molar, in a condition of full development.^ In the number of species of supposed placentals now known, sixty-seven, the proportion of species (1), with quadrituberculate superior molars is even smaller, being only four to sixty-three. The premolars display equally primitive characters, and to these I wish now to draw attention. 2. The presence of two internal tubercles of the fourth superior pre- molar is unknown as yet in the fauna. 3. The presence of two external cusps of the same tooth is known or inferred in only five species in the sixty-seven, and in two of the five it is of reduced size. 4. The presence of one internal cusp of the fourth superior premolar is demonstrated or inferred in all of the placental species, 5. The presence of the internal cusp of the third superior premolar is, on the other hand, only demonstrated in twenty-two species. In seven- teen it is wanting. Referring to the inferior premolars : 6. No species presents an internal cusp of the third premolar. 7. An internal cusp of the fourth premolar is present in only fourteen species. In twenty-nine species it is certainly wanting. * Proceedings of the American Philosophical Society, 1883, 562. American Jfaluralist, 1883, 407. M. .0135 .0050 .0090 .0075 324 [Dec. 7, 8. In no species of this formation is the fourth inferior premolar like a molar tooth. It is thus evident that the dentition of the mammalia of the Puerco fauna presents a much greater degree of simplicity than does that of the species of any of the later Eocene or other age. This result coincides with the results I have already obtained from a study of the structure of the feet, etc.* These may be summarized again as follows: 1. The species in which the number of toes is known, have them 5-5. 2. Those in which the feet are known are plantigrade. 3. No species is known to have interlocking carpal and tarsal bones, excepting the two species of Pantolainhda (carpus unknown). 4. No species is known to have well grooved astragalus (its presence is inferred in two species of Dissams). 5. No species is known to have a faceted radius or ulno-radius, adapted to the separate carpal bones of the proximal row. 6. In no species is the tongue in the metapodio -phalangeal joints devel- oped on the front of the metapodial bones. 7. The zygopophyses where known are all flat, except in some species (probably all) of Oxyclcenus, where they are simply convex-concave, and not doubly so. On the Trifuberculate Type of Molar Tooth in the Mammalia. By E. B. Cope. {Bead before the American Philosophical Society, Bee. 7, 1883.) It is now apparent that the type of superior molar tooth which pre- dominated during the Puerco epoch was triangular or tritubercular ; that is, with two external and one internal tubercles. f Thus, of sixty-seven species of placental mammalia of which the superior molars are known, all but four have three tubercles of the crown, and of the remaining sixty- five, all are triangular, excepting those of three species of Periptychus, and three of Conoryctes, which have a small supplementary lobe on each side of the median principal inner tubercle. This fact is important as indicating the mode of development of the various types of superior molar teeth, on which we have not heretofore had clear light. In the first place, this type of molar exists to-day only in the insectivorous and carnivorous Marsupialia ; in the Creodonta, and the tubercular molars of such Carnivora as possess them (excepting the plan- tigrades.) In the Ungulates its persistence is to be found in the molars of the Coryphodontidse of the Wasatch^and Dinocerata of the Bridger Eocenes. In later epochs it is occasionally seen only in the last superior molar. It is also evident that the quadritubercular molar is derived from the tri- tubercular by the addition of a lobe of the inner part of a cingulum of the * American Naturalist, 1883, p. 1056; Science, 1883, p. 275. t See American Naturalist, April, 1883, p. 407. 1SS3.] 325 [Cope. posterior base of the crown. Transitional states are seen in some of the Periptychidae (Anisoiichus), and in the sectorials of the Procyonidae. The tritubercular or triangular sviperior molar is associated with a corre- sponding form of the anterior part of the inferior molar, This kind of in- ferior molar* I have called the tubercular sectorial, and is very variable as to the degree of development of tlie sectorial cutting edge. The anterior triangle is formed by the connection by angle or crest, of the median and unterior internal crests with the anterior external. Its primitive form is seen in Didelphys, Pelycodus.Pantolambda and the Amblypoda generally ; in Centetes and Talpa; and in its sectorial form, in Stypolophus and Oxyiena, etc. The mechanical action of such teeth is as follows : Of course, it results from the form of the superior molars that the spaces between them are wedge-shaped, the apex external, the base opening to the palate. The base ot the triangular section of the anterior part of the inferior molar is inte- rior, and the apex exterior, and when the jaws are closed, this triangular prism exactly fits the space between the superior molars. The lower heel of the inferior molar receives the impact of the crown of the superior molar. Thus the oblique edges of the inferior triangle shear on the edges of two adjacent superior molars. The anterior parts of the inferior molars, and the superior molars, form an alternate dental series as distinguished from the prevalent opposed dentition of most mammalia. In so far it re- sembles the reptilian dentition. This primitive dentition has been modified in two directions ; viz. to form the grinding and the sectorial dentitions. As already remarked, the superior molars gradually acquire a posterior internal lobe, which produces the quadrituberculate type. This lobe, by opposing the anterior internal lobe of the next posterior inferior molar, precludes the entrance of the an- terior triangle of the latter between the two superior molars. Hence we find in the types which possess quadritubercular superior molars, that the anterior triangle of the inferior molar is not elevated, if present, as for in- stance in Rhinocerus. It is, however, more frequently atrophied, and dis- appears, forming the inferior quadritubercular molar so well known. On the other hand, as I have pointed out,f the anterior internal cusp of the triangle of the inferior molar may be more developed antero- posteriorly, giving the antero-internal edge of the triangle much greater obliquity than the postero-internal. In correspondence with this modi- fication, the superior triangular molar loses its equilateral character by the inore anterior position of its internal angle, thus elongating the posterior internal side of the crown. The latter thus fits the correspond- ing form of the triangle of the inferior molar, forming with it the shear of the sectorial tooth. *See Report G. M. Wheeler, D. Chief of Engineers on Explor. Surv. W. 100th Mer. Vol. IV, pt. li ; on the Creodonta. t On the origin of the sectorial tooth of the Carnivora, American Naturalist, 1875, Cope.] 326 [Dec. 7. In a former article, "On the Homologies of the Molar Teeth," etc., I traced the modifications of the superior and many of the inferior molars ot the ungulate mammals to a parent quadrituberculate type. In a subse- quent essay* I traced the origin of the inferior sectorial to a primitive five- tubercled, or " tubercular sectorial" type. Farther than this I did not go, and made no attempt to derive the few cases of triangular superior molars then known, nor the type of the superior sectorial. The revelations of the Puerco fauna show, that the superior molars of both ungulate and ungui- culate mammalia have been derived from a tritubercular type ; and that the inferior true molars of both have been derived from a "tubercular sectorial " type. Shall we look for the origin of the latter in a trituber- cular tooth also, i. e. tubercular sectorial without heel ; and will the crowns of the true molars of the primitive mammals alternate with, in- stead of oppose each other? This is a probable result of future discovery. * Journal Academy Natural Sciences, Philadelphia, March, 1875. Published January 2, 1884. Iileontologieal Bulletin, No. 38. SYNOPSIS OF THE PECIES OF OREODONTIM. (Mead before the American Philosophical Society, January 18, I884.) ON THE STRUCTURE OF THE SKULL IN THE lasmobrancli gents Diiriioiliis. [Read before the American Philosophical Society, March 7, I884.) BY PROF. E. D. COPE. FOR SALE BY No. 1223 BELMONT AVENUE, PHILADELPHIA. Jan. 18, im.] 603 fCope. Synopsis of the Species of Oreodontidve. By E. D. Cope. {Read before the American Philosophical Society, January 18, 188 4.) The tribe Ruminantia first appears in the White River Miocene period in North A.merican geological history. It is represented there by a num- ber of genera, which pertain to several family types. The most aberrant of these, the Oreodontidce, includes the largest number of forms, generic and specific. The Poebrotheriidce certainly embraces but few species, while a third group of genera, represented by Leptomeryx, which are inter- mediate between the Tragulina and Pecora, and should be perhaps regard- ed as aberrant Tragulidm, also includes a small number of species. The Oreodontidce constitute a family related to the Anoplotheriidm of the later Eocene, but representing a more specialized condition of the structure of the molar teeth, in the full development of the selenodont • type, which is rudimental in the Anoplotheriidm. Their feet, on the other hand, are less specialized than in the latter family. As a family, the Oreo- dontidce display very little tendency in their limbs to the specialized con- dition of the Ruminantia y but are more like those of the suilline groups, and, among recent families, of the Hippopotamidce. OREODONTIDJI], Dentition ; superior mcisors present ; molars selenodont. Cervicals with the transverse processes perforated by the vertebrarterial canal. No alisphenoid canal. Ulna and radius, and tibia and fibula distinct. Meta- podial bones four on each foot, with incomplete distal trochlear keels. Lunar bone not supported by magnum. Navicular and cuboid bones dis- tinct. The preceding synopsis of its characters should furnish a basis for the definite location of the Oreodontidce in the system. Dr. Leidy called its species Ruminating hogs, and created a family for Oreodon and the allied genera, under the name of Oreodontidce. This family is adopted by Prof. Gill who includes in it the Agriochoeridce of Leidy, and places it in his division Pecora, which is more comprehensive than the Pecora of Prof. Flower, being nearly identical with the Selenodonta of Kowalevsky. More precise expression of its affinity to the existing families is not given, ^ excepting to place it under a division "incertge sedis." As a selenodont type, this family is excluded from the Artiodactyla omnivora, and as having its metapodial bones distinct, it cannot be placed in any recent family excepting the Tragulidce. From this family it is distinguished by the distinct ulna and radius. We then turn to the ex- tinct families Poebrotheriidce and Anoplotheriidce. The former agrees with the Tragulidce excepting in its Cameloid cervical vertebrae, while the latter differs from the Oreodontidce in the structure of the feet. The Anoplotheriidce are didactyle in front, and tridactyle behind. The posterior foot has a well-developed second digit directed Cope. J 504 [Jan. IS, more inwards than the others, which it is supposed supported a nata- tory web. In the Oreodontidse all the feet are regularly tetradactyle.* The Anoplotheriidae differ also in the presence of an additional cusp on the inner side of the superior molars, accompanied by an imperfect de- velopment of one or both pairs of the internal crescents. In Anoplo- therium the internal crescents of the inferior molars are incomplete, and more or less represented by tubercles. In the Oreodontidm there are two pairs of fully developed crescents, and no internal tubercles. The de- tails of the structure express various affinities. The axis is intermediate between that of the suilline and ruminant Artiodactyla ; the other cer- vicals are suilline, while the remaining vertebrae are ruminant. The scapula is ruminant, not suilline ; while the humerus is like nothing but Anoplotherium. The radiocarpal articulation is intermediate between that of hogs and ruminants. The unciform supports the lunar bone. The sac- rum is ruminant, the ilium suilline. The femur and tarsus are much like those of the peccary. The genera of this family known to me are the following : I. Orbit incomplete ; last premolars in both jaws with two external crescents or Vs. Premolars three Coloreodon. Premolars four , Agriochoeriis. II. Orbit complete ; premolars four, the fourth with one external crescent. ^. No facial vacuities. Premaxillaries distinct ; otic bullae not inflated. . Oreodon. Premaxillaries distinct ; otic bullae inflated Eucrotaphus. Premaxillaries coossified; otic bullae inflated Meryeochoerus, aa- Facial vacuities present. Premaxillaries coossified, dentigerous ; vacuities prelachrymal only Meryehym. Incisors six above, persistent ; vacuities prelachrymal and pre- frontal ; nasal bones much reduced Leptauchenia. Incisors very few, caducous ; vacuities very large Gyclopidius. III. Inferior premolars three. True inferior canine functional; inferior incisors one on each side. Pithecistes. The number of species referred to these genera in the succeeding pages is as follows : Oreodon Eucrotaphus. . Merycochoerus. Merychyus . . . Leptauchenia . Gyclopidius. . . Pithecistes. . . . Agriochoerus . Coloreodon . . . 3 3 7 6 3 2 3 6 2 35 *I have observed this in the genera Oreodon, Eqcrotaphus, Merycochoerus, and Merychyus. 505 lCk)pe. The present paper is chiefly devoted to the proper distinction of these species and genera or cranial characters only. Figures of all will be given in my volume which embraces this subject, in the Report of the U. S. Geological Survey of the Territories. Coloreodon ferox Cope, one-half natural size. Original ; from Report U. S. Geolog. Survey Terrs., vol. iii, F. V. Hayden in charge. OREODON Leidy. Proceedings Academy Philadelphia, 1857, p. 238. Ancient Fauna of Nebraska, Smithsonian Contrib. to Knowledge, 1853, p. 29. Extinct Mammalia Dakota and Nebraska, 1869, p. 72. Report U. S. Geological Survey Terrs., 1873, I, p. 201. Merycoidodon Leidy, Proceeds. Acad. Philada., 1848, p. 47 (nomen nudum). Cotylops Leidy, Loc. cit., 1851, p. 239. Premaxillary bones distinct from each other. Otic bullae not inflated. No lachrymal vacuity of the face ; nasal bones normal. Premolars four in both jaws. Dental formula L f ; C. | ; P-m. f ; M. f ; the series uninterrupted. Crowns of the molars robust, well distinguished from the roots. Grind- ing surface of the true molars simply selenodont, i. e., with but two pairs of crescents. Superior premolars composed of a single external compressed cusp with crescentic section, and internal cingula or crescent. The fourth premolar with a well developed internal crescent ; the first three with rudi- mental internal crescents in the form of basal cingula. Superior canines distinct. Inferior premolars of two kinds ; the first canine-like in form and function ; the others consisting of a single external cutting edge rep- Cope.] 506 [Jan. 18, resenting two crescents, of which the anterior has its posterior horn developed as an obliquely transverse crest directed inwards. Last true molar with a heel composed of two columns. In the superior temporary dentition the last premolar has the form of the first permanent true molar. The third premolar has five lobes, i. e., four crescents and an anterior odd one. The other temporary premolar resem- bles that of the permanent series. The last inferior temporary premolar has the three pairs of lobes usual in the Artiodactyla, and the two which pre- cede it resemble the corresponding permanent teeth. Says Leidy : * "The permanent true molars successively protrude and occupy their functional position before any of the deciduous molars are shed. The displacement of the latter by their permanent successors appears to begin with the eruption of the last of these, which is followed by those in advance. The first permanent premolar of the upper jaw appears to have protruded after the deciduous teeth, and occupied a position with them in the functional series, but remains after these are shed. " > The cranial characters which belong to Oreodon as a genus are the fol- lowing : Orbit completed behind ; temporal fossae separated by a sagittal crest. A lachrymal fossa, but no facial nor frontal vacuities. Premaxil- lary bones distinct from each other and from the maxillaries. Nasal bones well developed. Auditory bullae not inflated. The preceding dental and cranial characters have been pointed out by Leidy in his various palaeontological works. On account of the absence of the necessary material he was unable to give the characters of the remaining parts of the skeleton. These are of course necessary to a correct estimate of the affinities of the genus, and I will endeavor to add such information as my material will permit. This consists of numerous more or less complete skeletons found in connection with the skulls by myself in Colorado in 1873. Vertebrm. The cervical vertebrae are rather short, and the character of the articulation of the centra slightly opisthocoelous, and the articular faces are quite oblique. The axis is the longest vertebra ; the three last centra are subequal in length. In one of my series the seven cervicals are preserved. In all of these, excepting the seventh, the bases of the diapophysis are perforated by the vertebrarterial canal. In the sixth vertebra, the decurved parapophyses are especially robust. The axis and three succeeding centra display strong hypapophyses at their posterior ex- tremities, which are carried forwards as strong median keels. The odon- toid process is depressed so as to have a lenticular section ; it is not exca- vated above, but in my largest specimen the internal borders of the facets for the atlas are continued so as to enclose a short groove on each side at its base. In one smaller and immature specimen this is wanting. The vertebrarterial canal of the axis is enclosed as in the other cervicals. The canal for the second spinal nerve has a narrow roof, but there are no canals * Ancient Fauna of Nebraska, p. 4i, 1884.J 507 I Cope. for the succeeding pairs of nerves perforating the neural arches. The atlas is not very elongate. The base of the diapophysis has a perforating canal, which issues in a large inferior fossa. The vertebrarterial canal then per- forates the diapophysis upwards anterior to the middle of the base, and then soon enters the neural canal just posterior to the superior margin of the cotylus of the occipital condyle. The centra succeeding the cervicals increase gradually in length poste- riorly. Those of the anterior part ot the dorsal series are quite depressed, but the vertical diameter rapidly increases, so as to be equal to the trans- verse in some of the lumbars. A trace of the opisthocoelous articulation exists throughout the dorsals but is very little marked in the posterior centra. There are no hypapophyses on the dorsals, but on one of them, probably the third, the inferior and lateral faces are separated by a strong angle, which is strongest anteriorly, giving the articular face a subquadrate outline. The rib-bearing diapophyses are robust. On the posterior dor- sals the capitular and tubercular surfaces are confluent, forming a narrow facet on the anterior face of the diapophysis, in a manner not seen in Cervus elaphus or Sus scropha. The centra of the lumbars, after lengthen- ing, become shorter immediately in front of the sacrum. The vertical diameter of one or two posterior ones is less than that of the anterior ones. The greater number of the lumbars display a small compressed hypapophysis at their anterior extremity ; but this is wanting on the posterior ones. The neural arches of the dorsal and lumbar vertebras are nowhere perforated for the spinal nerves. The lumbar prezygapophyses embrace the articular faces of the poste- rior ones, which have a section of one side (below), the end (external), and a half the other side (above), of a transverse ellipse. The superior recurved surface does not appear. The sacrum consists of five vertebrae, with very depressed centra. The ilium is attached to the diapophysis of the first, and a small anterior por- tion of that of the second. That of the fourth is flat and free. The an- terior zygapophysis of the first displays a slight degree of the superior incurvature general in Artiodactyla. The caudal vertebrae were numerous, forming a long tail. The proximal ones are moderately depressed, while more distal ones with wide diapophysis and complete neural arch, are sub- cylindric, and more elongate. The number of vertebrae preserved in the most complete of my specimens, is as follows : Cv. D. L. s. Cd. 7 5 6 4 4 5 8 6 2 1 4 5 3 * * 7 8 6 5 3 An anterior, perhaps second, sternal segment is flat and subquadrate in outline, with large haemal articular face of the lateral margin anteriorly, and a small one posteriorly. No inferior carina. i'ope.] 508 [Jan. 18, The spine of the scapula rises abruptly from the neck as in Ruminantia, and the coracoid process is short and obtuse. The spine continues to the distal extremity, which is regularly convex. The most perfect innominata in my collection are deficient in the sym- physis. The form of the ilium is more that of a hog than of a ruminant. The peduncle is even stouter, and the superior border is abruptly expanded below the middle of the length of the bone. The superior and inferior borders are subparallel as in the hog, and not divergent as in the rumi- nants. The anterior edge is acute, and uninterrupted by an anterior in- ferior fossa or spine. The pubis is robust and transverse, and without prominent basal pectineal tuberosity. The incisura acetabuli invades the base of the pubis a little, but the ischium more extensively. The ob- turator foramen is quite large. The distal border of the ischium is ob- liquely truncated as in many other Artiodactyla, and more nearly re- sembles that of the peccary than any other recent form I have observed. The tuber proper is a convex edge, not thickened, and its superior edge is continued into a strong up-looking tuberosity. This region is not so robust as in most recent forms. The humerus of Oreodon is readily distinguished from that of recent Artiodactyla by several peculiarities. The greater tuberosity is large, ris- ing above the head ; and is incurved, terminating inwards in an acuminate apex. Its border at the base is thrown into an obtuse angle. The lesser tuberosity is small, and is well separated from the greater by a deep and wide bicipital groove. The deltoid ridge is distinct. The condylar ex- tremity is more transversely extended than in any recent Artiodactyle, owing to the fact the posterior interior distal tuberosity is placed interior to the trochlea instead of partially behind it, and that there is, in addition, an internal epicondyle not seen in the recent suilline or ruminant mem- bers of the order. The intercondylar ridge is strong, and wider than m most recent ruminants ; in the suillines it has nothing like such a develop- ment. Another peculiarity is the flange-like free border of the external trochlea, which is especially recurved at its superior part. The radius is distinct from the ulna throughout. The relation of the ulnar to the radiocarpal surface is posterior as well as exterior ; the com- mon suture of the two, making an angle of 45^ with the long axis of the radiocarpal surface. The head is a transverse oval, with the inferior face forming a regular curve without notch. Its articular surface is divided into three portions in adaptation to the internal and external humeral trochleae and the wide median ridge. The external face is beveled forwards above, to fit the flange-like projection of the external trochlea. The shaft of the radius is not very stout, and has a nearly equal transversely oval section to near the distal expansion. Here are wide grooves for the extensor tendons, one superior, the other obliquely exterior. The carpal articular facet has the general ungulate characters. The scaphoid facet is concave above, convex and condyloid below, and is only distinguished from the 1884.1 509 [Cope. lunar facet by a contraction of the anterior and posterior borders. There is no indication of distinguisliing ridge between the lunar and cuneiform facets. The posterior border at their junction is prominent, enclosing a fossa with the scaphoid condyle, which does not, however, excavate the intervening surface. The scaphoid condyle is not divided hy a ridge. The ulna gradually contracts distally from a robust olecranon. The shaft beyond the humeral cotylus has an oval section, with its long axis forming an angle of 45° to the perpendicular. The olecranon is short and compressed, its posterior border rising nearly as high as the coronoid process. The edges of the humeral cotylus are not flared beyond the shaft. In the carpiLS the unciform nearly reaches the scaphoid, which is sup- ported by the magnum and trapezium. Tlie great trochanter of femur is not produced beyond the line of the head, and is well recurved, enclosing a large fossa. The little trochan- ter is large. The fossa Ugamenti teris is submedian, subround and large. Distally, the patellar trochlear groove is quite elevated ; its lateral crests are of equal prominence, and nearly equal superior prolongation. The patellar groove is continued some distance above the crests, but there is no fossa in this region as in the hog. The popliteal fossa is well marked, and the condyloid articular surfaces are not entirely cut off from the rotular. The external linea aspera terminates first in a rugose muscular insertion, and then in a shallow fossa a short distance above the condyle. There is no crest nor deep fossa. This element is more like th* corre- sponding one in Dicotyles torquaius than in any other mammal. The patella is a short wide bone, with a large anteroposterior diameter. One extremity is acute, the opposite one truncate. The head of the tibia is also like that of Dicotyles. The spine is divided as usual, and not much elevated ; the crest is prominent, but is wide and truncate above at the head. It is not excavated as in Sus. The external tendinous notch is well marked. The external margin of the shaft does not display any sutural surface for the fibula. The surface of attachment of an external malleolus is distinct. The internal malleolar process is nar- row and is produced well downwards. The anterior intertrochlear angle is prominent ; the posterior only convex. The trochlege are deep, the outer being both the wider and the deeper. The astragalus presents well marked characters. The distal extremity displays the two usual parallel trochlese, which are separated by a pro- nounced angle. The cuboid trochlea slopes somewhat backwards, while the navicular is strongly concave. The tibial trochlese are unequal, the internal being smaller than the external. It is separated from the latter by a constriction which is well rounded and not angulate as in the hog. The external side of the astragalus displays a wide malleolar band, a wide posterior and narrow anterior calcaneal facets, and an undivided concavity intervening between the latter. On the inner side, the malleolar face Cope.J 510 [Jan. 18, descends to below the middle, as in Hypertragulus, and there is no verti- cal nor horizontal distal crest. The inferior calcaneal facet is undivided and not grooved, and does not extend over the internal border of the in- ferior side of the bone. It exhibits an acute border on the external side. The calcaneum is rather elongate, and the free portion is compressed and ■with obtuse margins above and below. The transverse astragalar pro- cess is not large and is not produced beyond its facet. The ascending plate is well developed and has a superior, uninterrupted convex facet for the fibula, with a narrow facet on its inner side. The inner distal astraga- lar facet extends the entire length of the cuboid facet. There is a longi- tudinal ridge on the external side of the distal end of the calcaneum. The navicular and cuboid bones are distinct from each other and from the ectocuneiform. The astragalar ligamentous fossa is in the naviculo- cuboid suture. The inferior proximal angle of the cuboid is produced posteriorly, and the peroneal process well forwards. The ectocuneiform is distinct, and much wider than long. The mesocuneiform is exterio- posterior in position, and the transverse diameters are small. It is pro- duced distally, overlapping the head of the second metatarsus. Ento- cuneiform wanting. The metapodial bones are entirely distinct. The lateral metatarsals are well developed. The second articulates with both the ecto- and mesocuneiform bones, by a proximal extremity which is laterally compressed. The third and fourth are subequal in width, and articulate exclusively with the ectocuneiform and cuboid respectively. The fifth metatarsus is compressed proximally, and the external part of its extremity articulates with a lateral fossa of the cuboid. The distal articular extremities of the metapodials are separated from the anterior face of their shafts by a transverse groove ; and they have a well marked articular fossa on each side. The trochlear tongue only exists on the posterior face, where it is prominent and compressed. It disappears on the middle of the distal end, and is wanting on the anterior face. The phalanges are depressed proximally, the penultimate ones distally also. The ungues are rather depressed and have convex external borders. There is a pair of sesamoid bones below the distal articular extremity of the metatarsals. History. The dental and cranial characters of this genus were fully described by Dr. Leidy in 1852, as already cited. In the Extinct Mam- malia of Dakota and Nebraska, published in 1869, Dr. Leidy added the following points in the osteology of the skeleton of the Oreodontidm (p. 72) : " What are supposed to be the bones of the forearm and leg are discrete, as in the hog, and the bones of the feet correspond in number with those of this animal." In 1873* Prof. Marsh confirmed these state- ments so far as regards the metacarpal bones, and added that "the navicu- lar and cuboid bones were loosely coossified or separate." The structure of the vertebrae, and of the greater part of the scapular and pelvic arches, ' Amer. Journ. Sci. Arts, p. 409. im.] 511 [Cope. with the carpus, tarsus and feet, with the exceptions above noted, are now described for the first time. This genus appears first in time in the known history of the family, and presents us with its primitive or least specialized characters, or those nearest the average condition of the ordinary primitive ungulate. Species. The species of this genus are difficult to discriminate from the evidence of crania alone, and their true number will remain uncertain until we can study entire skeletons. My material enables me to make some progress in this direction. After the removal of the forms with in- flated bullae to the genus Eucrotaphus, there remain the two species origi- nally referred to Oreodon by Leidy, the 0. culbertsoni and the 0. gracilis. To these Leidy subsequently added two others, the 0. afflnis, which is in- termediate in size between the two named, and the 0. hybridus, of larger size than either. As the condition of the otic bullae in the last is unknown, its generic reference is not certain. All these forms are from the White River epoch of Dakota, Nebraska and Wyoming. My material is largely from the White River beds of Colorado. I find from this region the true 0. gracilis and the 0. culbertsoni, abundantly represented. Besides these there is a form intermediate between the 0. gracilis and the 0. afinis, which is nearer the former than the latter. Of 0. gracilis there are two skulls complete ; of the form next larger, which I call 0. gracilis coloradoensis, two complete crania (one with skeleton), and a face with teeth. Of a form between the 0. affinis and the 0. culbertsoni, there are four skulls complete (two with skeletons) ; and of 0. culbertsoni proper, numerous parts of skulls with teeth, but none com- plete. No other regions which I have explored have produced these species ; not even the Ticholeptus beds, where they might have been rea- sonably expected to occur. The distinction of the previously known species will remain as Leidy has left it, with certain reservations in the matter of dimensions ; while I add two sub-species. Nasal bones obtuse posteriorly ; frontals little produced on either side of them ; true molar teeth not exceeding M. .035 in length ; canine and premolars .030 ; width of front .046. 0. gracilis. Nasal bones obtuse posteriorly, frontals little produced on either side of them ; true molar teeth not exceeding .037 in length ; canine and premolars .039 ; width of front at middle of orbits .046 0. coloradoensis. Nasal bones obtuse posteriorly, frontals little produced on either side of them ; true molar teeth not exceeding .038 ; front at orbits .057 in width 0. afinis. Nasal bones acute posteriorly ; frontal produced to an acute apex on each side of them ; molar teeth .040 ; front, .056. 0. periculorum. Nasals and frontals as last ; molar teeth .047 ; front, .050-}-. . 0. culbertsoni. Cope. J 512 [Jan. 18, From this table it may be seen that the passage from the small 0. gracilis to the large 0. culbertsoni is accomplished by a series of inter- mediate steps. That these extreme forms belong to one species cannot be admitted without evidence of more complete transition than we yet possess. As above remarked, groups of specimens represent each form and adhere to the definitions given with considerable fidelity. The largest of the specimens I refer to, the form 0. periculorum, however, reaches .042 in the length of the true molar teeth, and the smallest of the 0. culbertsoni measures .046. These I must consider as sub-species only. As regards the three remaining forms the length of the true molar series shows a complete gradation. The size of the cranium, as indicated by the interorbital width, is in the 0. affinis as large as that of the 0. culbertsoni according to Leidy, and the combination of characters presented by this form, would seem to entitle it to specific rank as suggested by Leidy. On the other hand the form coloradoensis agrees in interorbital width with the small 0. gracilis, differing from it in the greater length of the muzzle and of the cranium. But here, while the proportions of the premolar teeth distinguish the forms well, the length of the brain-case does not coincide exactly with the other measurements. The measurements of four skulls are as follows : 0. gracilis No. 1, length of skull M. 114.5 ; No. 2, .130. 0. coloradonesis No. 1, .129 ; No. 2, .135. Oreodon gracilis Leidy. Proceedings Academy Philada., 1851, 239 ; 1853, 392 ; 1854, 157 ; 1857, 89 ; Owen's Report Geolog. Survey, 1852, 550, PI. XI, figs. 2-3 ; PI. XIII, figs. 5-6. Ancient Fauna Nebraska 1853, p. 53, PI. V, figs. 3-4 ; VI, figs. 1-7. Extinct Fauna Dakota and Nebraska, 1869, 94, PI. VI, figs. 2-3, Abundant in the "White River beds of Dakota, Nebraska, Colorado and Wyoming, The two sub-species are distinguished as follows : Length of superior premolar series, M. .023 0. g. gracilis Length of superior premolar series, M. .029 0. g. coloradoensis. Oreodon gracilis gracilis Leidy. Dakota, Nebraska and Colorado. Oreodon gracilis coloradoensis Cope. Colorado. Oreodon affinis Leidy. Extinct Mammalia Dakota and Nebraska, p. 105 ; PI. IX, fig. 3. Probably from the White River beds of Nebraska. Oreodon culliertsoni Leidy. Owen's Report Geological Survey, 1852, 548, PI. X, figs. 4-6 ; XIII, figs. 3-4 ; Ancient Fauna Nebraska, Smithsonian Contrib. to Knowledge, 1853, 45 ; PI. II, III, IV, figs. 1-5, V, figs. 1-2, VI, figs. 8-11 ; Proceeds. U81.] 513 [Cope. Academy Philada., 1853, 392 ; 1854, 35, 157 ; 1857, 89 ; Bronn Lethsea Geognostica, 1856, 930. Extinct Fauna Dakota and Nebraska, 1869, p .86 ; PI. VI, fig. 1 ; VII fig. 2 ; IX, figs. 1-2. Merycoidodon culbertsoni Leidy, Proceeds. Acad. Phila., 1848, 47, PI. II ; 1850,121; 1851,239. Oreodon priscum Leidy, Proceed. Phila., Academy 1851, 238 ; Cotylops speciosa Leidy, Ibidem 239 ; Oreodon robustum Leidy, Ibidem 276. The White River epoch of Dakota, Nebraska, Colorado and Wyoming. The two sub-species are defined as follows : Length of superior true molar series from M. .040 to .042 0. c. periculorum. Length of superior true molar series from .046 to .050 0. c. culbertsoni. Oreodon culbertsoni periculorum Cope. This smaller race or sub-species has as yet only been found in the White River beds of Colorado and Wyoming. I do not detect any differ- ences between it and the Nebraska form other than those of size. The largest measurement of the 0. c. culbertsoni given in the, above table is derived from Leidy ; my largest specimen gives .047 as the length of the true molar series. Oreodon culbertsoni culbertsoni Leidy. Very abundant in the White River formation of Dakota, Nebraska, Colorado and Wyoming. EUCROTAPHUS Leidy. Proceedings Academy Philada., 1850, p. 92. Ancient Fauna of Nebraska, Smithsonian Contrib. to Knowledge, 1853, p. 56. Eporeodon Marsh, Amer. Journ. Sci. Arts, Vol. ix, 1875, p. 249. Premaxillary bones distinct from each other. Otic bulla swollen. No prelachrymal or nasal vacuities. This genus presents us with the first step in the series of modifications which the primitive form underwent with the advance of geological time. It appeared contemporaneously with the earliest representatives of the family, i. e., in the White River epoch, but in small numbers. In the succeeding or John Day epoch the genus Oreodon had disappeared, and the present form had multiplied enormously in individuals, if not in species. Subsequent to that epoch it is unknown. The greater number of the Oreodont remains found in Oregon belong to this genus. The Eucrotaphus jacksoni bore the same relation to the Oregon John Day fauna, as the Oreodon culbertsoni did to that of the White River epoch. The species of Eucrotaphus are distinguished as follows : I. Palatonareal border well posterior to posterior edge of maxillary bones. Cope.] 514 [Jan. 18, a. Infraorbital foramen above front of P-m. iii. Skull depressed, muzzle short ; paroccipital process behind bulla and not separated from it by grooves ; bulla grooved to apex for styloid ligament, etc. ; zygoma more robust. E. trigonocephalus. II. Palatonareal border in line with posterior edges of maxillary bones. aa. Infraorbital foramen above posterior part of third premolar. Paroccipital process behind otic bulla, the internal border of its base opposite that of the bulla E. jacksoni. Paroccipital process external to the middle of the otic bulla ; generally larger E. major. The name here employed for this genus is the one first given with a definition. The typical species, E. jacksoni, was widely distributed, and appears under several varietal forms and sizes, some of which have re- ceived names. Subsequently to the original description, Dr. Leidy added to the genus a second species, which probably belongs to the genus Agriochoerus. On this account Leidy inclined at one time to combine the two genera, but afterwards abandoned the idea. £u€rotaphus trigonoceplialiis, sp. nov. This distinct form is only known to me from a single skull of an old animal. In the character of its otic bulla it has resemblance to the species of Agriochoerus, while the maxillary part of the skull has the posterior position of a true Oreodon. The muzzle is rather depressed, and the premaxillary alveolar border is almost transverse. The position of the canine alveolus is swollen later- ally, and between it and the infraorbital foramen the side of the face is slightly concave. The expansion leading to the malar bone commences as the posterior slope of the concavity mentioned, and spreads laterally, without interruption, beginning to project beyond the superior alveolar border at the fourth superior premolar. In the E. jacksoni this is not apparent anterior to the first true molar. The top of the muzzle and the front are wider than in that species, and are gently concave in the transverse direc- tion. The anterior temporal ridges are well defined, and concave in out- line, uniting early to form a prominent sagittal crest. The malar bone is a little concave below the orbit. The malar process of the maxillary pro- jects downwards in an obtuse angle, opposite the penultimate superior molar. In E. jacksoni the malar is convex, and the tuberosity is opposite the last molar. The squamosal process is deeper than in the E. jacksoni, and sends a more robust apex into the malar bone, the apex not extending in front of the posterior border of the orbit. The supraoccipital crests are well developed, and project beyond the vertical plane of the condyles ; they continue into well marked posttemporal crests, as in the other species of the genus, as well as send an obtuse ridge downwards on each side to- wards the foramen magnum. The median supraoccipital plane disappears downwards in a wedge-shaped apex, which causes the transverse section 1884.1 515 [Cope. above the foramen magnum to be obtuse angulate instead of broadly flattened as in E. jacksoni. The mastoid crests are roughened and are vertical, but do not continue directly into the paroccipital processes, but are separated from them by a deep excavation of the external margin, due to the internal position of the base of the process. The long diameter of the base of the paroccipital process runs outwards and backwards, and it is attached to the bulla at the middle of the posterior extremity without any intervening grooves such as are seen in the other species of the genus. The bullae are ovoidal in anteroposterior section, the regularity interrupted, however, by the presence of a ridge on the exter- nal side directed posteriorly, enclosing a groove which is continuous with the stylohyoid fossa. The ridge continues into the inferior crest of the tympanic bone. The sphenoid bone is regularly convex in transverse sec- tion, while the basioccipital is concave on each side with a narrow median keel, which commences opposite the anterior edge of the paroccipital pro- cesses. The basicranial axis is not quite in line with the basifacial, but does not present such an angle with it as is seen in the species of Mery- cochoerus, where the skull is known to me. In this respect it agrees with the other species of the genus. The post glenoid processes are less promi- nent than in E. jacksoni, but have a base more widely extended outwards. The external border is very oblique, since the apex is narrowed. The glenoid region is more extended, both transversely and anteroposteriorly than in the E jacksoni. The anterior border is continued as an alisphe- noid angle which becomes prominent, and overhangs the foramen ro- tundum. The descending alisphenoid ridge commences within the anter- ior border of the foramen ovale. The pterygoid angle is anterior to the middle of the palatosplienoid wall of the nareal foramen, and in front of it the edge of the processus pyramidal is is marked by a shallow fossa or mark of insertion of the internal pterygoid muscle. The nareopalatal border is as far posterior to the line connecting the posterior edges of the maxillaries as the widtli of the second molar tooth. The palate is every- where nearly flat. The malar bones spread well away from the maxillaries on each side, the anterior border of the zygomatic foramen being a seg- ment of a circle. The squamosal pirt of the zygoma is more widely ex- panded than the malar part. In E. jacksoni the shaps of the zygomatic foramen is quite diflerent. Its anterior outline is interrupted by the pro- jection of the maxillary bone posteriorly, which gives its anterior outline a bilobate form. It is longer than wide in that species, and wider than long in the E. trigonocephalus. The infraorbital foramen is small. There are two lachrymal foramina ; one larger, within the preorbital border, the other smaller, below the tuberosity on the rim of the orbit. Tlie frontal foramina are separated by a space equal to one-fourth the entire frontal width. The supraorbital notches are wanting. The preorbital fossae are well marked, are distinctly defined above, and extend as far as the anterior border of the lachrymal bone. The orbit is round, and looks upwards as well as outwards and Cope.] 516 [Jan. 15, forwards, on account of the prominence of the zygomatic arch. There are two postparietal foramina, one below and behind, the other on the parieto-squaraosal suture. The mastoid foramen is not small. The incisive foramina are large, are longer than wide, and are separated by a rather wide isthmus. The palatine foramina are opposite the third premolars. There is a foramen immediately below the postfrontal process. The optic foramen issues posterior to the line of the posterior border of the orbit, and in front of the anteroinferior angle of the alisphenoid. The foramen rotundum is large and round, and is immediately below and within the ridge above mentioned, and is not overhung by a transverse ridge of the same, as in the species of Meryoochoerus known to me. The f. rotundum doubtless includes the f. sphenoorbitale. The f. ovale is smaller and is separated by a considerable interval from the f lacerum. The latter is subtriangular in form and is rather small, since the base of the otic bulla is in close sutural contact with the sphenoid and basioccipital for a considerable distance. The f. jugulare is sub- triangular in outline and is smaller than the f. rotundum. It is entirely distinct from the f. condyloideum, which is the size of the f ovale. No f. supraglenoideum. In comparing these foramina with those of the K jacksonl, a general resemblance is to be seen. The frontal fora- mina in that species are generally closer together than in ^. trigonocepha- lus, and the palatine foramen is generally opposite the fourth premolar in- stead of the third. The foramen magnum is slightly notched on it3 superior border in both. The posterior outline of the nasal bones is truncate ; it is more or less acuminate in all the specimens of B. jacksoni and B. major accessible to me. The prolongation of the frontal on either side of the na^ls is also short and truncate in this species, and narrow and acuminate in the B. jacksoni and B. major. The lachrymal is deeper than long ; in the species last named it is of variable size and form, but is usually as long as deep. There is no distinct ridge along the parieto-squamosal suture. The ali- sphenoid has a considerable contact with the parietal. The palatomaxil- lary suture is irregularly convex backwards on each side of the median line. It crosses the palate as in the B. jacksoni, at the front of the second maxillary tooth. The teeth are much worn, and the first and last true molars with several of the premolars have been lost, indicating the age of the animal. The incisors are small and hare round roots. The canines are large and of the usual form. The space between them and the first premolar is short. The fourth premolar is small. The second true molar isVider than long, and has no internal cingulum except between the lobes, and has a trace of anterior cingulum. Measurements. M. Axial length from occipital condyles to premaxillary border 187 Axial length from occipital condyles to postglenoid pro- cess • 031 1884.J 517 [(Jope. Measurements, M. Axial length from occipital condyle to postfrontal pro- cess 07G Axial length from occipital condyle to palatonareal bor- der 079 Axial length from occipital condyle to end of last molar. .091 Diameters of orbit \ ^^^^^^^^ ( horizontal 027 Depth malar bone at middle of orbit 016 " zygomatic process posteriorly to glenoid face 028 " skull (right angles to profile) at glenoid face 045 orbit 046 P-m. i 030 Elevation of occiput from foramen magnum 044 Width top of muzzle at preorbital fossse 040 at middle of supraorbital border 059 " " postfrontal process 075 " " malar below orbit. 110 " ** zygomatic process of squamosal 145 " of occiput at condyles 066 " " occipital condyles 039 " " palate at palatonareal foramen 028 at M. ii 032 " " canines 030 Length of superior dental series with canines 088 " *' premolar series 047 '* ** true molar series 036 Diameters canine at base | anteroposterior 009 ^ transverse 010 Diameters P-m. iv. j anteroposterior 009 t transverse 013 Diameters M. h. / Anteroposterior 014 i. transverse 018 The typical specimen of this species was found by Charles H. Sternberg on the North Fork of the John Day river. The horizon is probably some- what different from that of the true John Day epoch. Eucrotaphus jacksoni Leidy. Proceedings Academy Philadelphia, 1850, p. 92. Ancient Fauna of JSTe- braska, Smithsonian Contributions to Knowledge, 1852, p. 56, Plate VII, figs. 4-6. Oreodon hullatus Leidy, Extinct Mamm., Dakota and Nebraska, 1869, p. 106. Report U. S. Geol. Survey, Terrs. 1873, 1, p. 318. Oreodon occidentalis Marsh, Amer. Journal Sci. Arts, 1873 (May), p. 409. Epor- eodon occidentalis Marsh, Loc. cit., 1875, p. 250. Eucrotaphus occidentalis Cope, Bulletin U. S. Geol. Survey Terrs., V, p. 59. Comparison of numbers of crania from the White river and John Day Cope.j 618 [Jan. 18, formations fails to reveal any characters distinguishing them as more than one species. In fact the variation in various respects is greater among the individuals of the John Day epoch, than between those of the two epochs. This was by far the most abundant mammal of the John Day epoch while it appears to have been rare during that of the White River. Specimens differ in the size of the preorbital fossa irrespective of other differences. In some specimens it is wide and profound, including the lachrymal bone ; in others it is less extensive and is shallow, involving but part of the lachrymal. It is never wanting or obscure. For estimation of other characters, I select ten crania, nine from Oregon and one from Dakota, as expressing the greatest range of variation. Of these, three display a peculiarity in the form of the otic bulla. Instead of being con- tracted backwards in front, it is protuberant and full at its inferior anterior part. Five other crania, agreeing with these three in other respects, possess the normal form of bulla. In one cranium, which is rather more robust than the others,, the infraorbital foramen is a little posterior to its usual position, being above the anterior part of the fourth premolar. This tooth is also distinctly smaller than in other specimens of otherwise similar dimensions. The majority of specimens range nearly alike in dimensions, but there are forms distinctly larger and smaller, which may represent distinct species. This question can be better decided when the skeletons are known. I give three sub-species which are defined as fol- lows : Length of cranium M. .197; of molar series M. .086; long diameter of base of paroccipital process transverse ; its pos- terior base flat E. j. jacksom. Length of cranium M. .219 ; of molar series M. .091 ; paroccipi- tal process as above E. j. pacificus. Length of cranium, M. .235; of molar series, M. .099; paroc- cipital process strongly compressed, its posterior base an- gulate on the middle line. . . , E. j. Uptacanihus. The above measurements of length are made from the occipital condyles to the premaxillary border inclusive. The three forms may represent good species. The E. j. jackso?iiis of the size of the Oreodon culbei'tsoni ; the E. j. leptacanthus is larger than the E. major, while the E. j. pacificus is intermediate between the two. £ucrotaplius jacksoni Jacksoni Leidy. The typical specimen of the Oreodon bullatus Leidy agrees so nearly with the original type of Eucrotaphus jacksoni, that I cannot doubt their pert- inence to the same species. There are two specimens in the collection of the Philadelphia Academy, besides the last named, and at least one in the museum at Princeton. A specimen from the John Day, Oregon, cannot be distinguished from these. It agrees with Marsh's measurements and description of his Oreodon ocddentalis, and no doubt represents it. Its im.] 519 [Cope. identity with his 0. bullatus has already been surmised by Leidy (Report U. S. Geol. Survey Terrs., I, p. 318). Eucrotaphiis jacksoni pacificiis Cope. This form is materially larger than the last named, equaling in dimen- sions and resembling in general form the Eucrotaphus major Leidy, of the White River beds. It is no doubt the form which has been identified .under that name by Leidy in his report on John Day Fossils in the Report of the U. S. Geological Survey of the Territories, Vol. I. It is different from that animal in the form and position of the paroccipital process, as already pointed out. I have eight crania disengaged from the matrix which agree in dimensions and other characters assigned to this sub-spe- cies. In one of them the paroccipital process presents an approach to the form of that of the E. j. leptacanthus. A specimen from the White Buttes of Central Dakota agrees with those from Oregon in all the essential characters, and is the second one of the sub-species I have seen which is not Oregonian. I have many crania of this sub-species not yet entirely cleared of matrix. From John Day river and Crooked river, Oregon ; C. H. Sternberg and J. L. Wortman ; White river of Nebraska, Mus. Princeton. Eucrotaphus jacksoni leptacanthus Cope. This is the largest form of the genus, exceeding the typical E. major in the length of the skull by 23 mm. It is thus far represented in my collec- tion by two very perfect crania. There is considerable reason for antici- pating that this form will turn out to be a valid species. Besides the pecu- liar form of the paroccipital processes, the typical specimen presents the following characters : The frontal region is flatter than in the two other sub-species, and is concave on the median line in transverse section. This concavity is probably partly abnormal. The profile of the sagittal crest instead of pre- senting a gently convex outline, is concave, rising posteriorly. The lateral occipital crests instead of being angulate are truncate behind, and the in- ferior angle projects much beyond the vertical line of the occipital con- dyles. As this part is broken off in most of my specimens of the E. j. pacificus, I cannot decide as to its value. The inferior carina of the tym- panic bone extends forwards to contact with the internal extremity of the postglenoid process. It does the same in the Oregon specimen of E. j. jacksoni, and in the Dakota specimen of the E. j. pacificus. In two of the latter, from Oregon, where the part is cleaned, the keel does not extend so far forwards or inwards. The typical specimen is from the John Day beds of John Day river, Oregon, and was found by Jacob L. Wortman. Eucrotaphus major Leidy. Oreodon major Leidy, Ancient Fauna of Nebraska, 1853, p. 55, PI. TV, fig. 6. Proceedings Academy Philadelphia, 1853, 398 ; 1856, 164 ; 1857, 89. Cope.] 520 [Jan. IS, Extinct Mammalia, Dakota and Nebraska, 18C9, p. 99, PI. VII, fig. i ; VIII. Eporeodon major Marsh, Am. Journ. Sci. Arts, 1875, p. 250. I find this species to differ in the external position of the paroccipital process, as related to the otic bulla, from the E. jacksoni. I might add that it differs in dimensions from all excepting the E. jacksoni pacificus. In the E. jacksoni the base of the paroccipital process is in the same line as the interior base of the otic bulla. In the Oregon form of the E. major the base of the paroccipital process is much flattened, so as to be trans- verse, and its internal border is on the external side of the extremity of the large swollen bulla. This species differs also from the E. jacksoni in the median vertical carina of the occipital bone above the foramen magnum, a region which is in the E. jacksoni broadly flattened. Besides these points I do not notice any divergence from the E. jacksoni, with which it agrees in the various characters in which the latter differs from the E, trigonocephalus. The Nebraska and Oregon forms do not agree in all respects. Thus, while the dimensions of the dental series are the same in both, the frontal region is more elongate in the Oregon animal, giving greater length to the skull. The third superior premolar has a somewhat different form in the two. They may then be characterized as follows : Dental series M. .135 ; skull .224 ; third superior premolar, sub- triangular : E. m. major. Dental series M. .125 ; skull .240 ; third superior premolar sub- quadrate E. m. longifrons. Eiicrotaphiis major major Leidy. Known only as yet from the White River epoch of Nebraska and Dakota. Eiicrotapliiis major longifrons Cope. Known from a single skull from the North Fork of the John Day river, Oregon, found by Charles H. Sternberg. It may be observed here that the Oreodontidae of this locality are mostly distinct from the species of the John Day river proper. MERYCOCHOERIJS Leidy. Report U. S. Geol. Survey Terrs., I, 1873, p. 202. Bettany, Quart. Journ. Geol. Soc. London, 1876, p. 262 ; Cope, American Naturalist, 1884, p. 281. Leidy, Extinct Mammalia of Dakota and Nebraska, 1869, p. 110 (nomen nudum). Proceedings Academy Philadelphia, 1858, p. 24 (nomen nudum). As indicated in the analytical table at the head of this article, I can only distinguish this genus from Eucrotaphus by the confluence of the pre- maxillary bones. The position of the external infraorbital foramen can- not be regarded as furnishing generic characters, especially as it displays considerable variation and gradation. Some of the species are in this respect quite identical with species of Merychyus {M. superbus), while others 188#-.J 521 [Cope. possess the widely different position ascribed to this genus by Leidy. Few if any of the characters given by Mr. Bettany as those of the genus, can be regarded as other than characters common to several of its species. Perhaps the most important of these is the angle formed by the basifacial with the basicranial axis, by which the face is presented as much forwards as upwards. The species present considerable variety in form. The ge- nus embraces the largest species of the family, such as M. macrostegus, M. superhus, etc. The characters of the species are as follows : I. Foramen infraorbital above middle of fourth superior premolar ; pos- terior part of zygoma expanded ; palate moderately produced posteriorly. Squamosal part of zygoma less expanded anteriorly and with rounded border ; head elongated ; premaxillary bone not produced ; otic bulla larger, compressed, extending anterior to postglenoid process ; size large M. superhus^ Head shortened occipitally, so that a line drawn through post- glenoid and paroccipital processes makes 90° with the middle line ; malar bone openly grooved below orbit ; . angle of mandible obliquely truncate M. leidyi. Squamosal part of zygoma most expanded in front, and elevated behind, so that the cranium is as wide as from the paroccipi- tal process to the canine tooth ; its posterior angle rising to a level with the sagittal crest ; its inferior edge spread out- wards ; its superior edge truncated ; occiput not shortened ; malar flat below orbit ; postglenoid process marking front of bulla M. chelydra. II. Foramen infraorbitale above the first true molar. Palate greatly produced posteriorly. Squamosal part of zygoma much expanded, and with truncate edge ; malar bone robust, prominent ; skull, width equal length from condyles to first premolar ; maxillary produced anteriorly ; frontal plane, transverse diamond-shaped ; bulla si^jall, conical, posterior to anterior edge of postglenoid process M. macrostegus. Squamosal part of zygoma little expanded upwards or lateral- ly, edge rounded ; malar bone flat ; bulla large, extend- ing in front of postglenoid process ; front longitudinally diamond- shaped, decurved at orbit M. montanus. III. Foramen infraorbitale above anterior border of second true molar. Zygoma originating above second molar ; large ; incisors small (fide Leidy) M. rusticus. Zygoma originating above third true molar ; larger ; incisors large (fide Leidy) M. proprius. Of the above seven species, four are represented in my collection, sortie of them by a large amount of material. The latter are from the John Cope.] 522 [Jan« 18, Day and Ticholeptus Miocene horizons. The M. rusticus of Leidy is only known to me from the descriptions of that author. It is from the Sweet- ■^ater river, Wyoming, from a bed of probably Ticholeptus age. The M. proprius Leidy, also unknown to me by autopsy, is from the head of the Niobrara river, Nebraska, from a bed said by Hayden to be inter^ mediate between the Oreodon or White River and Procamelus, or Loup Fork horizons, and therefore probably ot Ticholeptus age also. The M. leidyi I only know from the description of Mr. Bettany. It is from the 4^ohn Day beds. Mr. Bettany also describes an M. temporalis, which I cannot distinguish from the M. auperbus Leidy. lUerycoclicerus super1>us Leidy. Oreodon superbus Leidy, Proceedings Academy Philadelphia, 1870, p. 109. Extinct Mam. fauna, Dakota and Nebraska, 1869, p. 211 ; Plate I, fig. 1 ; II, fig. 16 ; VII, figs. 7-11. M. temporalis Bettany, Quar. Journ. Geol. Soc, London, 1876, xxii, p. 269 ; PI. XVII. Of this fine species I have nine crania extracted from the matrix, and a good many not yet cleaned. As the specimen described by Leidy is in a very imperfect condition, the characters of the species, and even its generic position, have remained hitherto very obscure. As compared with the allied species, the M. superbus is slightly exceeded in size by the M. maerostegus and M. montanus. Its posterior zygo- matic expansion is less pronounced than in the M. maerostegus and M. chelydra, and its border is rounded, even when, as is sometimes the case, it is greatly thickened. In the first and last named of the above species, its border is separated by a distinct angle from both the internal and ex- ternal faces, forming thus a distinct truncate face which looks upwards. The otic bulla is larger than in the two species mentioned, and extends anterior to the postglenoid process. The nareal fissure extends well down towards the alveolar border of the premaxillaries, which are therefore more extensively separated than Leidy represents to be the case in the M. rusticus. The external face of the malar bone below the orbits is flat. The anterior extremity of the zygomatic process is not so progiinent as in M. chelydra, and is rounded instead of being flared out below, as in that species. The greatest width of the skull is at the glenoid surfaces, and not anterior to them, as in M. chelydra. In only one of seven crania, where the parts are preserved, does the posterior squamosal angle rise as high as the sagittal crest. I cannot detect any diflerence between the specimen described by Mr. Bettany as the type of his M. temporalis, and those of the M. superbus in my possession. The shallowness of the preorbital fossa described by Mr. Bettany is repeated in one of my crania, and its depth is very vari- able in the others. As regards the M. leidyi of Bettany, I have none ex- actly like it, although the type specimen does not differ much from the M. superbus, to judge from the figure and description given in the Quarter- ly Journal of the Geological Society, 1876, p. 270. The two distinctive 1884.] 523 [Cope. characters, which appear most tangible among those mentioned by Mr. Bettany, the shortness of the occipital region, as measured by the angle made by a line drawn through the postglenoid and paroccipital processes, with the middle line, and second, the grooved character of the sub- orbital part of the malar bone, are not found in any of my specimens of M. superbus. The anterior extremity of the squamosal process of the zygoma is protuberant in one of them, as in the M. leidyi. Another char- acter is suggested by Mr. Bettany's figure, but is not mentioned in the text. The angular border of the mandibular ramus extends obliquely forwards instead of being prominently convex as in the best preserved entire mandible of the M. superbus in my possession. Nevertheless in another specimen, where a good deal of the posterior border is preserved, the outline is nearly as oblique as in the M,. leidyi. The species, however, is distinct so far as now known. John Day epoch, Oregon, C. H. Sternberg and J. L. Wortman. Local- ities, John Day river, Bridge creek, and Camp creek of Crooked river. MerycoclKeriis leidyi Bettany. Quarterly Journal of the Geological Society of London, xxxi, 1875, p. 270 ; Plate XVIII. Defined and discussed under the preceding species. John Day epoch, Oregon ; Lord Walsingham. John Day river. Ulerycoclicerus clielydra, sp. nov. This species is known to me by a skull without mandible, which is entire, except that the extremity of the nasals and the border of the pre- maxillary bones are broken off. It is unfortunate that I have no second skull to confirm its characters, but my numerous specimens of the M. superbus, to which it is most nearly allied, do not present any approxima- tions which suggest transitions between the two. The striking character of this cranium is its great breadth at the tem- poral region, as compared with its length and other dimensions. The forms of the otic bulla differ from those of the M. superbus. One method of expressing the width of the skull is as follows. The point of the frontal bone which is equidistant from the supraoccipital notch and the external edge of the zygomatic arch, measured in a horizontal plane, is directly above the posterior or nareal palatal border, when the skull rests on the teeth. In the M. superbus, in the most robust examples, this point is above a point which is a good deal nearer to the line of the anterior edge of the glenoid surfaces than to the palatal border, and at least 30mm. posterior to the latter. That this relative shortness of the basicranial axis is not due to a shortening posterior to the glenoid surfaces, as is the case in M, leidyi Bett., is proven by the fact that a line drawn through the postglenoid and paroccipital process makes an angle of 90^ with the middle line, as in M, superbus. The muzzle is compressed and its superior surface is regularly rounded Cope.] 524 [Jan. 18, The side is divided by the gentle convexity continued, forwards from the malar region. Below this and above the premolars thB face is concave. Above it the preorbital fossa is well marked, though not deep, and gradu- ally fades out anteriorly. The interorbital region is flat, as in M. macro^te- gu8, and the supraorbital border is not decurved, as it is in M. superbus and M. montanus. The supraorbital and preorbital borders of the front are, however, not continuous as in M. macrostegus, though nearly in the same line, which they are not in M. superbus. The orbits are more oblique than in M. superbus, looking more upwards and forwards, and their verti- cal exceeds their transverse diameter. The malar bone though oblique, is more vertical than the orbit below the latter, and has an uninterrupted gently concave surface. The postorbital bridge is narrow, and consists one-half of the malar and one-half of the frontal bones. The inferior edge of the malar is thin and is slightly convex downwards, and passes behind the protuberant squamosal at a point behind the line of the postfrontal pro- cess. The anterior extremity of the squamosal is not protuberant below the orbit and only begins to rise gradually below the line of the postfrontal process. It then expands rapidly downwards and outwards in a strong curve, with its flat surface looking upwards as much as outwards. After making a short downward turn it rises steeply, contracting gradually in- wards, and presenting a convexity posteriorly, with its truncate edge looking outwards. Its apex is nearly on a level with the sagittal crest. The inner or descending edge of this process is concave, so that the apex overhangs a little the posterior outlet of the temporal fossa. The anterior temporal angles are strongly marked and unite into a sagittal crest. The edge of the crest is thickened, so that its section is a letter T. The supraoccipital bone presents a wide flat convexity above the foramen magnum, in distinction from the stronger convexity of M. superbus, and the still stronger of the M. macrostegus and M. montanus. As in the other species, the posttemporal (= lateral occipital) crests are only present at the upper half of the occiput. Between them there are two ligamen- tous or tendinous insertions, but no median keel. The exoccipital and posttympanic borders form a tuberosity below the meatus auditorius, which passes upwards into a short convex posttemporal crest. The paroccipital process nearly reaches the postglenoid by its anterior external edge. The tympanic is complete, is not keeled below, and extends itself as a lamina over the posterior side of the postglenoid process. The section of the basioccipital is open V-shaped. The inferior flat surface of the sphenoid is produced backwards in a wedge-shaped prominence to a line connect- ing the anterior edges of the paroccipital processes. It has the same form in M. macrostegus, but in three skulls of M. superbus, where it is visible, the apex of the w^edge does not extend posterior to the middle of the otic bullae. The bullae are small and subconical, and reach as far as the ante- rior edge of the postglenoid process.. In the latter the transverse diam- eter exceeds the anteroposterior, which exceeds the vertical diameter. This process and the otic bulla are of about equal protuberance. In four 1884.] 525 [Cope. crania of the M. superbus, where both are well preserved and exposed, the bulla is considerably more prominent than the postglenoid process. The glenoid surface is well-defined and equally wide at both extremities. The inferiorly presented surface of the zygomatic arch, is wider than in any of the other species, including examples of M. superbus of super'ioT dimensions in other respects. The surface is rugose. The length from a line connect- ing the median external columns of the last superior molar, to the poste- rior nareal border, enters three times into the distance from the latter to the border of the foramen magnum. In M. superbus it goes three to three and a half times ; in M. macrostegus and M. montanus once only. Be- hind the molars the produced palatal roof is more concave than between the last two true molars. The palate becomes then more concave (convex), and between the first premolars and canines becomes flat, and expands laterally. The nareal fissure is not much contracted between the pre- maxilliaries. The infraorbital foramen is above the anterior half of the superior fourth premolar, and is of moderate size. The frontal foramina are separated by a space which is less than half as M'ide as that which separates each one from the superciliary border. There is no supraorbital notch. The in- cisive foramina are large, are wider than long, and approach close to the bases of the canine teeth. The palatine foramina are minute or obsolete. The foramen ovale is isolated and is opposite the junction of the glenoid and postglenoid surfaces. The jugular foramen is isolated by the exten- sive contact of the otic bulla and the basicranial axis. Perhaps the condyloid foramen is included in it, as I do n.ot find it in the usual position. The animal is so old that no sutures are visible. The teeth are not all cleared from the matrix, which is hard and brittle. The first true molar is much worn. The first premolar is two-rooted, and is separated from the canine by a diastema equal in leijgth to the long dia- meter of its crown. Measurements. M. . Length from occipital qondyle to front of canine tooth. .300 " " " " " postglenoid process. . . .041 " " " " postfrontal process. .. .133 " " " " " palatonareal border . . .118 " " " " end of last molar 146 Diameters of orbit I ^^^^^^^^ ^^^^ 1 transverse 039 Depth of malar bone at middle of orbit 034 " " zygomatic process to glenoid face behind 088 Width of top of muzzle at preorbital fossa 043 at middle of supraorbital border C94 " " malar below orbit 160 *' " middle of zygomatic arch 254 ** of occiput at superior crests 050 Cope. J 526 [Jan. 18, Measurements. M. Elevation of occiput from foramen 084 Width of occipital condyles 063 Width of occiput at condyles 095 Depth of skull at right angles to profile at glenoid face. .095 orbit 087 " " " P-ml 075 Length of superior dental series with canine 159 " " premolar series ; 061 " " true molar series 065 Diameters M. i | anteroposterior 0180 transverse 0185 Diameters of canine | ^^^^^"P^^^^"^^ transverse 020 Diameters p.m. ii \ anteroposterior 0155 ( transverse 090 Width of palate at m. i >. .044 P-m. i 057 The typical specimen was found on the John Day river, Oregon, by Mr. J. L. Wortman. MerycoclKerus macrostegus, sp. nov. I have been able to discover in my collection as yet, but one cranium with entire mandible of this species. The very^arked characters of this skull are such that no farther evidence of its reference to a peculiar species is needed. Its affinities, as expressed in the analytical key which accom- panies the general discussion of this genus, are with the M. montanus. This is shown in the posterior positions of the infraorbital foramen, and of the posterior nares. As peculiar characters may be added the form of the frontal plane and of the otic bulla ; also the prolongation of both the premaxillary and supraoccipital regions, and the forms of the zygoma, the angle of the mandible, and the first inferior premolar tooth. The skull reaches a greater length than that of any species, excepting the M. mon- tanus, but is not nearly so robust as in the M. chelydra, resembling in this respect rather the M. superhns. The muzzle is compressed, and there is a decided concavity just above the second premolar, above which the surface is a little convex. Above the infraorbital foramen, the face is abruptly convex, the convexity slop- ing upwards to the base of the median ridge formed by the convex nasal bones. Behind this the side of the face is a plane which slopes outwards as it descends, which is only interrupted by the rather small, but well de- fined, preorbital fossa. The fossa is better defined in front than in the other species, but I do not know whether the character is constant. The front is a transverse diamond-shaped area, bounded posteriorly by the anterior temporal ridges, and anteriorly by the lines of the supraorbital borders 1884.J 527 [Cope. produced to their point of intersection with each other. Such point of intersection is above the second true molar in this species ; in M. superbus and M. chelydra it is above the posterior part of the second premolar. The area in these species enclosed by the lines in question is half as long again as wide, instead of wider than long by 18mm. This difference is partly caused by the greater prominence and flatness of the postorbital angle of the frontal bone in the M. macrosteyns, and the more anterior direction of the orbits, which I may add have none of the tendency to superior direction seen in M. chelydra. The wide triangular area thus enclosed on its external sides by the orbit and anterior temporal ridges, is perfectly flat. Such an area can hardly be defined in the other species, and the surface there is rounded and descending. The malar bone is deep, flat and a little oblique outwards, and the rim of the orbit projects a little, giving it a slight concavity. The orbit is deeper than wide. The anterior part of the zygomatic process of the squamosal is not protuberant below the orbit, but gradually rises outwards posteriorly, attaining its greatest expansion opposite the middle of the zygomatic foramen ; above, its course is for a time parallel with the middle line of the skull. The form of the zygomatic arch is more like that of M. chelydra than any other species, but it is not so much expanded, especially anteriorly. Its inferior and posterior surface is, however, widened, making an angle with the ex- ternal or marginal surface, which is in turn separated by an angle from the superior and anterior surface ; at the middle of the arch the superior surface has a width of 19mm., and the external a width of 23mm. The posterior angle rises to the plane of the summit of the sagittal crest, and the apex, which is less than a right angle, stands above the external base of the postglenoid process. The preglenoid border is not exactly at. right angles with the middle line, but makes a slight angle outwards and forwards. The long diameter of the zygomatic foramen is parallel with it. The ridge along the parietosquamosal suture is insignificent. The supraoccipital region is very prominent, and as in the other species of this genus is narrowed below by the disappearance of the posterior temporal or exoccipital crests. They are continued downwards and disappear, leav- ing a wide convex surface above the foramen magnum. This is separated; by the usual lateral fossa from the posterior temporal angles. The coossified mastoid and paroccipital processes much contract the auricular fossa below, but do not close it. The latter is contracted at the base of its terminal part, and is distally slender. The otic bulla is the smallest known in the genus, it is compressed and oval, and not produced beyond the postglenoid processes either forwards, backwards or down- wards, in this diftering much from the M. montanus. It is separated by wide and equal intervals from this process, the glenoid surface, and the basisphenoid. It sends a process backwards and inwards to a sutural junction with the basioccipital bone. The tympanic bone is flat below, and is united with the posterior base of the squamosal by a flat expansion. The postglenoid process is robust, and has the height and thickness equal, Cope.] 528 LJan. 18, while the width exceeds both. The basioccipital bone is prominently keeled on the middle line, so that the section is a V of a more compressed character than the section of the same in M. superbus. The median plane of the sphenoid is prominent, and is continued as a wedge with the apex opposite the posterior borders of the otic bullae. The palatine borders are parallel, except where they form on each side an open angle at the junc- tion of the descending process of the sphenoid, which is here directed for- wards. Its external border is distinct from that of the palatopterygoid plate, and makes a groove with it. The maxillary bone is not produced posterior to the notch on either side of the base of the posterior production of the palatine bones. The middle line of the latter is deeply concave opposite the former, and the palate is also especially concave between the first true molars. The palate is flat between the first and second pre- molars. The inferior surface of the squamosal process of the zygoma is roughened for the origin of the masseter muscle. The inferior edge of the malar comes from its inner side, and is narrow and with a median groove. Its inferior edge is continued as a ridge of the maxillary as far as opposite the anterior lobe of the second true molar. The maxillary bones are more produced anteriorly than in any of the other species. The apex of the nasal bones stands above the posterior border of the canine in this species ; above the anterior edge in M. superbus, M. clielydra and M. leidyi (fide Bettany). The posterior border of tlie nares is above the anterior part of the first premolar in the three species named, except M. chelydra where it is over the posterior edge of the canine : in 31. macrostegus it is above the posterior edge of the longer first premolar. The infraorbital foramen is large, and its posterior border is above the anterior root of the first true molar. The incisive foramina are large, and each one is a little longer than wide. The nareal opening contracts gradu- ally to its inferior apex. There is a considerable maxillary foramen op- posite the middle of the fourth superior premolars. The posterior nareal is not large ; its anterior outline is regularly concave. Its lateral (sphe- noid) borders reach to opposite the anterior faces of the postglenoid pro- cesses and bound the foramen ovale on the inner side. The latter is round, is rather small, and is opposite the middle of the postglenoid surfaces. The foramen rotundum on the other hand is large and vertically oval, and is bounded below by a transverse prominence of the base of the ali- sphenoid bone. It probably includes the sphenoorbital foramen, a foramen anterior to its inferior border probably communicating with the nareal chamber. The optic foramen is small, and is situated opposite the ante- rior two-fifths of the zygomatic fossa and a little above the line of the apex of the foramen ovale. The foramen lacerum is ovoid and not large. The posterior foramen lacerum is a transverse sigmoid, one extremity being the jugular foramen. The mastoid and postparietal foramina are of moderate and equal sizes. No postsquamosal or supra- or postglenoid foramina. The animal described is too old to exhibit sutures. 1884.1 529 fCope, The mandible possesses some distinctive characters. The angular border is not prominent posteriorly, extends forwards below, and projects below the general level of the inferior border of the ramus. Neither of these characters is observable in the only ramus of the M. superbus in which the lower part of this border is well preserved, but in some others of that species the superior part of the border is much as in M. macro- stegus. The inferior edge of the ramus is straight, but there is a descend- ing tuberosity of the symphysis which may be an individual peculiarity. The symphysis is very concave in profile, and the incisive border is pro- duced in accordance with the prolonged muzzle. In the M. 8up&rhus it is sometimes convex, sometimes a little concave, but not so much so as in this jaw. The coronoid processes are small and slightly everted. The inner ridge of its anterior base is more prominent than the exterior, and encloses a fossa with it. The masseteric fossa is not noticeable. There is one large mental foramen below the third premolar. The dental fora- men is large and oval, and when the mandible stands on a level surface is opposite the middle lobe of the third inferior molar tooth. In dentition this species is distinguished by the relatively large size of the premolar teeth, of which the first, second and third are two-roOted in both jaws. Both the first and second in the upper jaw have short diaste- mata anterior and posterior to them, the largest being behind the canine tooth, and nearly as long as the premolar's crown. All the teeth are a good deal worn in the specimen. One can see two internal cingula in- closing fossae oij the third premolar. The true molars increase in size rapidly posteriorly and the third has a well -developed external heel. The molars have no internal cingula ; these are present in five of seven skulls of the 31. superbus where these parts are cleaned. The most noteworthy point in the mandibular dentition is a very rudimental character of the in- ternal vertical ridge of the crown of the first premolar. The posterior fossa of the fourth premolar is closed,. and the anterior remains open, on wearing. In M. superbus both are closed in the specimen where visible. The anterior inner wall is represented in the second and third premolars by a cingulum. No cingula on the true molars. First premolar very ro- bust, its section lenticular. Meaturements. •■ M. Axial length from occipital condyles* to premaxillary border 345 Axial length from occipital condyles to postglenoid pro- cess 045 Axial length from occipital condyles to postfrontal pro- cess 188 Axial length from occipital condyles to palatonareal border.... 100 *The occipital condyles are brolcen off in the specimen, so I measure from the superior border of the foramen magnum, which is, in the other species, in the vertical line of the occipital condyles. Cope.] [Jan. 18, Measurements. M. Axial length from occipital condyles to end of last mo- lar 058 Diameters of orbit I ^^"^^^^^ ^^4 transverse 036 Depth malar bone at middle of orbit 037 " zygomatic process to glenoid face behind 077 " skull (right angles to profile) at glenoid face 088 " orbit 088 " P-m. i 068 Elevation of occiput from foramen magnum 084 Width top of muzzle at preorbital fossa 038 " at middle supraorbital border 109 '* " postfrontal process 137 " " malar below orbit .166 " middle of zygomatic arch 243 " of occiput at superior crests 050 " " ** condyles 101 Length superior dental series, with canine 177 " " premolar series 092 ** *' true molar series 083 Diameters canine I anteroposterior 013 ^ transverse 018 Diameters P-m. i j anteroposterior 017 <^ transverse 075 Diameters m. i I anteroposterior 019 «. transverse 0215 Diameters m. iii / anteroposterior 038 I transverse (at middle column) 029 Width of palate at P-m. i 061 ♦« " m. i 053 " " middle of zygomatic arch 047 Length of inferior dental series with canine 179 ** " premolar series 088 ** " true molar series 088 " of ramus to posterior edge 279 Depth of ramus mandibuli at condyle 124 " " m. iii posteriorly 073 " " *' m. i posteriorly 048 " " P-m. i (front) 015 Diameters inferior P-m.i { anteroposterior 019 transverse 0125 Diameters " p.„,. iv f anteroposterior 021 transverse 013 Diameters " j / anteroposterior .030 C transverse 014 18W.] 531 ICope. Diameters inferior m. iii Measurements. . ^ anteroposterior. t transverse M. .044 .018 This fine species is from the John Day epoch of tlie Miocene. The typical specimen was found by my assistant, Charles H. Sternberg, on Bridge creek, Oregon. Much credit is due Mr. Sternberg for his unwearied exertions in the cause of science, which have been continued through many occasions of risk and discomfort. Merycoclicerus montanus, sp. nov. This large animal is represented in my collection by a nearly entire skull with parts of both mandibular rami complete. Rami of another in- dividual give the entire dentition of the lower jaw except the incisors. A third individual is represented by a symphysis with premolars, ca- nines and incisors, and by various parts of the skeleton, including feet. Of the cranium mentioned, the muzzle to the preorbital fossa and the palate to the first true molar are wanting. The region of the larmier is lost, but the general resemblance of the species to the M. macrostegus in other respects, leads me to suspect that it is absent, and that the M. mon- tanus, is rightly referred to the genus MerycochcErus. This course is indi- cated by the structure of the superior molar teeth, which have the character of those of this genus, rather than that found in Merychyus. That is, the posterior internal crescent sends its anterior horn to the external wall of the crown, thus cutting off the posterior horn of the anterior crescent. Dr. Leidy has shown that the reverse is the case in the Merychyus major ; that is that the posterior horn of the anterior crescent reaches the external wall of the crown, cutting off the anterior horn of the posterior crescent. I have observed that this is also the case in the other species of Merychyus which have come under my notice. The posterior position of the infraorbital foramen and the greatly pro- duced palate distinguish this species from those of the John Day epoch, excepting the M. macrostegus, while in the M. rusticus and M. proprius, the infraorbital foramen is still further posterior. The palate of these species is unfortunately unknown. The part of the maxillary bone posterior to the infraorbital foramen is nearly flat, and the proximal part of the malar bone is also flat. The in- ferior edge of the latter is narrow and is marked by a groove which ter- minates anteriorly in a shallow fossa. The ridge continuous with this edge terminates above the anterior lobe of the second true molar. The zygoma as far as the anterior border of the glenoid cavity is slender, and not con- vex, but flat in every direction, nor is it decurved as in M. superbus. The zygomatic foramen is relatively much smaller than in that species. Its posterior or preglenoid boundary is not at right angles to the sagittal crest as in that species, but is oblique outwards and forwards at an open angle. The obtuse median edge of the zygoma looks upwards, not outwards as it does in M. superbus and M. macrostegus, and the superior expansion is Cope.] 532 [Jan. 18, opposite the internal extremity of the glenoid face, instead of the external as in M. superbus, or the middle, as in M. macrostegus. The border descending to the supraauricular " crest is thin and vertical in direction, and the superior angle stands above the middle of the postglenoid process, not external to it, as in the two species above named. The postglenoid process is robust and has a convex posterior face. The paroccipital pro- cess is long and acuminate. An external truncate ridge on the front of its base partially embraces the meatus auditorius, and curving forwards be- comes the anterior edge of the process, which is separated from the post- glenoid by but a narrow interval. The tympanic bone forms a tube more distinct from the surrounding regions than in the other species here de- scribed, and has a longitudinal inferior keel, which is not visible in the i/I superbus and M. macrostegus. It is separated at the meatus by but a short interval from the base of the postglenoid process. The supraauricu- lar and mastoid crests unite and form a short acute crest, which does not continue into a prominent posttemporal, but descends into a mere angle, which continues as a fine line to [the convexity of the true posttemporal crest above. The latter arises from the bifurcation of the sagittal crest, and after a strong convexity descends with its fellow to a narrow promi- nent convex ridge, which rises from the foramen magnum, Thus the oc- ciput on either side of this prominent middle line is deeply excavated, and the fossa is bounded on each side and anteriorly by the low posttemporal angle, and the more prominent mastoid tidge. There is no median keel. The median ridge of the occiput is more prominent and not so flat as in M. superbus, but is more as in M. macrostegus. The sagittal crest is well developed, and has a straight superior border, which is not thickened as in M. chelydra. The anterior temporal ridges are represented by an angle which is nearly right. The superior squamosal suture is marked by a prominent ridge. The front is gently convex transversely, and the supra- orbital border is more strongly decurved than in M. superbus, which are more so than in M. macrostegus. The basicranial axis makes a strong angle with the basifacial as in the other species of the genus, showing that the face was presented obliquely forwards, as in the peccary. The section of the basioccipital bone be- tween the paroccipital processes is V-shaped, owing to the presen<'.e of a strong median angle. In M. macrostegus this bone is similar, but in M. superbus it is much flatter, and there is a weak median keel. The sphenoid is in line with the occipital and has a broadly rounded-truncate inferior face. The otic bullse are large and compressed. They extend from the middle of the base of the paroccipital process to considerably in advance of the postglenoid process, and approach very near to the glenoid surface. The interval which separates them is small, equaling one-fifth the antero- posterior diameter of the bulla. This is very different from the M. macrostegus, where the space between the glenoid surface and the bulla, is equal to the anteroposterior diameter of the latter near the middle. As already pointed out, this species agrees with the species just named in the 188-l.J 633 [Copq. great prolongation of the palatal floor of the nareal cavities. The distance from the foramen magnum to the nareal border equals the distance from the latter to the line connecting the median external vertical crests of the last superior molars. In M. superbus the former measurement is two and one-half times as great as the latter. The mandible shows the nearer relationship to the M. macrostegue than to the M. superbus, in the anterior elongation and greater relative size of the premolar teeth. It agrees with the former in having the profile of the symphysis concave, and not convex as in M. superbus. It is less con- cave in my single specimen than in that of M. macrostegus. The posi- tion of the posterior extremity of the symphysis is below the middle of the third inferior premolar. The coronoid process is low, and of small size. Its compressed convex apex is directed at an angle of 45^ from the middle line outwards and forwards. Its anterior face soon widens out and the internal edge becomes much more prominent than the external, with which it encloses a shallow, subtriangular, subvertical fossa. The external border is continuous with the external alveolar border. The masseteric fossa is small and has no distinct inferior border, and does not descend below the level of the line of the middle molar teeth. The in- ferior border of the ramus is nearly straight. The inferior incisive alveo- lar border is much more strongly convex than in the M. superbus. The condyle has the posterior articular face on the inner side, as in other species. The infraorbital foramen is large and is above the anterior part of the first true molar tooth. The meatus auditorius is small. There are two postparietal foramina on the parietosquamosal suture. No supraglenoid or postglenoid foramina. There are two mental foramina, one not small below the anterior part of the first true molar, the other, quite large, below the posterior part of the third premolar. The dental foramen is situated on a level with the alveolar border and well posteriorly, its ante- rior border being a little in front of a line dropped vertically from the apex of the coronoid process. It is thus similar in position to that of M. macrostegus and different from that of M. superbus, where it is above the line of the apices of the molars, and is posterior to the line dropped from the apex of the coronoid. In the superior true molars, the size increases rapidly posteriorly. The third is relatively of more elongate form than the first, but the posterior external column is but little produced. The other vertical ridges are quite prominent. The external faces of the external lobes are nearly flat. Besides the relation of the adjacent horns of the internal crescents already mentioned, the posterior horn of the posterior crescent in the first and second molars is cut off* from the external wall of its own crown by the anterior horn of the anterior crescent of the crown next posterior. This does not exist in worn molars of M. superbus and M. macrostegus, but is observable in little worn teeth of the former. It does not look as though the character would disappear with wear in the M. montanus. The only Cope.] 534 [Jan. 18, trace of cinguluin on the superior molars is on the inner base of the ante- rior lobe, where it is weak, and in the interspace between the internal lobes, where it is a narrow tubercle. Enamel obsoletely vertically striate. It is wanting on the external side of the internal crescent, as Leidy has shown to be the case in certain species of Merychyus. The fifth lobe of the last inferior molar is well developed and has its two crescents separated by a groove. The adjacent horns of the external crescents are of about equal length. No cingula, except a trace on front and rear of crowns, and a tubercle between the bases of the external lobes. The fourth premolar has two fossae isolated, one anterior to and the other posterior to the principal apex, which is double, and anterior to the middle. Before wear, each of these fossae opens inwards. The crown of the third premolar has its inner face unequally divided by a crest behind the middle. Posterior to this the space is occupied on the inner side by two shallow fossae of which the posterior is the narrower. Anterior part of inner face of crown concave. One principal angular cusp. The second premolar has a compressed triangular crown with a long base, and a weak vertical ridge on the in- ternal side. The first premolar is a very robust tooth with a straight posterior border directed at 95° forwards, and is vertically truncate in the specimen by friction with the canine. Section of crown lenticular, rounded in front. Measurements. M. No. 1. Length from occipital condyle to postglenoid process. . . ;049 " " " " " postfrontal process... .135 Width of occiput at posttemporal crests .054 '* " " condyles 102 Elevation of occiput above foramen magnum .084 Length from foramen magnum to palatal border 060 Width between apices of otic bullae 042 Length from inferior m. iii to apex of coronoid process. .075 " of superior true molar series 084 Diameters m. i / anteroposterior . . 026 ^ transverse (at middle nb) 02d Diameters m. iii { anteroposterior. 084 I transverse (at middle nb) 025 Length of inferior true molar series 085 Diameters P-m. iv { anteroposterior. 0205 transverse behmd 015 Diameters m.i [*'"'='-''P<'^'«"" t transverse 016 Diameters m. iii { anteroposterior 040 trane verse 033 No. 2. Length of ramus mandibuli from incisive border to condyle (oblique) 280 535 [Cope. Measurements. M. Length of dental series (straight line) 191 " from hist molar to apex of coronoid 073^ " of premolar series 085 " " true molar series 084 " " second premolar on base 031 " " first premolar on base 0225 Depth of ramus at coronoid 044 " " end of m. iii 073 " " middle m. i 056 " P-m. i. vertically 034 The specimens of this species were found by Mr. J. C. Isaac in the Ticholeptus beds of Deep river, Montana, during his Expedition of 1880. ruerycochoeriis rusticiis Leidy. Report U. S. Geological Survey Terrs., 1873, i, p. 199, PI. Ill, figs. 1-3 ; VII, figs. 1-5 ; XX, figs. 9-81. Proceedings Academy Philadelphia, 1870, 109. The smallest species, characterized among other things by the closure of that part of the nareal fissure which separates the premaxillary bones below. According to Leidy's figure above quoted, the depth of the middle line of the undivided premaxillary is greater than the width of the bone, a state of things not approached by any of the species of this genus de- scribed in the preceding pages. The premaxillary in the M. proprius is not described. From the ? Ticholeptus beds of the Sweetwater river, Wyoming. Merycoclioerus proprius Leidy. Proceedings Academy Philadelphia, 1858, p. 24 ; Extinct Mammalia Dakota and Nebraska 1869, p. 110 ; PI. X. This large species represents the extreme form of the genus in the ante- rior position of its dental series as compared with the braincase. The zygomatic arch and infraorbital foramen are therefore more posteriorly placed than in any other species. The premaxillary bone is more promi- nent than in any other, and the incisor teeth have relatively larger dimen- sions. The size is about that of the M. superbus. I have not seen any other than the typical specimen. From the Ticholeptus beds at the head waters of the Niobrara river, Nebraska. lUERYCHYUS Leidy. Proceedings Academy Philad'a, 1858, p. 24, (nomen nudum). Extinct Mammalia Dakota and Nebraska, 1869, 115. Report U. S. Geological Survey Terrs, i, 1873, p. 203. Cope, American Naturalist, 1884, p. 281. Ticholeptus Cope, Bulletin U. S. Geolog. Survey Terrs., 1878, p. 380. Premaxillary bones codssified ; otic bulla swollen ; a vacuity between Cope.] 536 [Jan. IS, the maxillary, lachrymal, and nasal bones, or larmier. Nasal bones nor- mal. . First inferior premolar caniniform. This genus has not been defined prior to the present article, although some characters common to the species of the genus known to him, have been given by Leidy. As now defined it-is identical with genus Ticholep- ius Cope. This group was distinguished by the presence of a larmier, a character whose presence in the species of Merychyus has been hitherto unknown. It is not yet reported indeed as present in any of the original species of the latter, but I think that there can be no reasonable doubt of its presence there. A character found by Leidy in the M. major I find to be present in one or more of the superior molar t6eth in all the species. The posterior horn of the anterior internal crescent cuts oft the adjacent or anterior horn of the posterior internal crescent from contact with the inner side of the external wall of the crown. It is the anterior horn of the posterior internal crescent which reaches the external wall, in the genera Merycochoerus, Eucrotaphus and Oreoden. In Leptauchenia the arrange- ment is generally as in Merychyus ; see under the head of that genus. This genus is confined to the Upper Miocene beds, the Ticholeptus and Loup Fork epochs. In size the species range from medium to large, the M. major equaling any species of the family in dimensions. They are distinguished as follows : I. True molar teeth not prismatic. Infraorbital foramen above fourth premolar ; malar bone shal- low ; squamosal with superior zygomatic angle anterior ; true molars M. .043 M. pariogonus. II. True molar teeth more or less prismatic, a. Infraorbital foramen above third premolar. Larmier a slit ; front narrow M. arenarum leptorhynchus. aa. Infraorbital foramen above fourth premolar. /?. Zygomatic arch vertical, and with posterior angle small and rounded. Larmier triangular; front wide; true molar series M. .044; face convex M. arenarum arenarum. A^. Zygomatic arch expanded horizontally ; posterior angle strong, acute. Larmier large ; true molars .051 M. zygomaticus. yS/?. Zygomatic arch unknown. Facial plate generally concave ; true molars M. ,045 M. elegans. True (inferior) molars, M. .069 M. medius. True (superior) molars (m. iii inferred), M. .095 M. major. Of the above species, the M. arenarum and M. zygomaticus are known from entire skulls. In the first named, the foramen infraorbitale appears to be partly above the posterior edge of the third premolar, as well as above the anterior edge of the fourth. 188-1.] 537 I Cope. merycliyus areiiarum Cope, sp. nov. Sub species leptorliyn- chiis Cope. This species is represented by a sltull which lacks of completeness only the extremity of the muzzle and the angles of the lower jaw. Its size is about that of the Oreodon culhertsoni or of the Merycliyus elegans. The confluence of the premaxillary bones shows that the place of the species is with the last-named genus, and the sigmoid flexure of the masticating line of the superior dentition is a point of resemblance to the species of the same. The position of the external infraorbital foramen is one de- gree further posterior than in the species of Oreodon, and agrees with the position in two other species of Merychyus {M. arenarum and M. pariogonus), which is more anterior than in the other species of the genus. The foramen is in fact quite identical in position with that seen in most of the species of Eucrotaphus, to which genus the above named species must be regarded as the nearest in the genus to which they belong. As in other species of the genus, the malar bone is deeper and less prom- inent laterally than in those of Oreodon. The preorbital fossa is wider and shallower. The orbit is closed behind. The premaxillaries are convex in every direction, least so transversly. The fissure which separates them is quite narrow, and is separated from the alveolar border by a rather narrow isthmus of uninterrupted bone. At the canine tooth the direction of the surface becomes longitudinal by an abrupt turn, and the side of the face above the second premolar is un- interruptedly gently concave. The lateral convexity which bounds the preorbital fossa below, appears above the third superior premolar, and be- comes more prominent posteriorly as it passes into the flat surface of the malar bone. The anterior orbital border is prominent and thin, and does not develop a distinct tubercle, although its edge is roughened. The profile of the muzzle is a straight line descending gently from the interorbital region. Above the middle of the orbits the frontal bones are gently convex ; on the line of their anterior border, there is a concavity of the median line. The superior face of the nasal bones is flat, and is peculiarly narrowed, especially posteriorly, where the large preorbital fossae approach each other. The anterior temporal ridges are well marked, and after a gradual approach unite into a sagittal crest, which has a gently convex ris- ing profile. After the posterior bifiurcation of the latter, the convex posterior temporal crests do not project beyond the occipital condyles when the inferior edge of the lower jaw rests on a horizontal plane, as in so many other species of this genus and of its allies. These crests continue without interruption above the auricular meatus to the posterior base of the postglenoid process. As compared with the Oreodon culbertsoni, the postorbital part of the cranium is short ; it is also shorter than in any other species of Merychyus. Thus the length from the posterior border of the orbit to the convexity of the Cope.] 538 [Jan. 18, posterior temporal crest, is as long as from the former point to the anterior base of the first premolar. In the Oreodon culhertsoni, the same measurement is equal to the length from the same point to the anterior base of the third incisor. This shortening posterior to the orbit is seen to involve the zygomatic fossa as well as the region posterior to it. Thus the horizontal diameter of the orbit in the M. leptorhynchus is exactly equal to the distance between the posterior border of the same and the anterior edge of the glenoid cavity. The posterior part of the superior edge of the squamosal zygomatic process is thin and strongly convex. The apex of the convexity is above a point just anterior to the posterior border of the glenoid cavity. The posterior edge of the process is nearly vertical, and if continued would reach the middle of the base of the postglenoid process. The latter is compressed and rather elongate, and its convjex edge has considerable transverse extent. The paroccipital process is long and is flat on its posterior face. The postorbital process of the frontal is elongate wedge-shaped, with its truncate apex below joining a slight ele- vation of the malar bone, which is much less prominent than in Oreodon culbertsoni. It presents an angle outwards and forwards, as the orbital border. The anterior half of the zygomatic process of the malar bone is rounded- truncate below. The glenoid surface is plane transversely, and slightly convex, rising backwards, anteroposteriorly. The anterior border of the squamosal bone is not developed into a ridge. The frontal bone extends forwards on either side of the nasals, forming a narrow process above the lachrymal bones. It overlaps the superior edge of the maxillary, of which a narrow splint appears between it and the nasal. The nasals are rather narrow, and each has the posterior bor- der rounded. The latter fall above the middle of the first true molar tooth when the inferior edge of the mandible is horizontal. The lachry- mal bone has greater anteroposterior than vertical diameter, extending nearly to the line of the infraorbital foramen, or much in advance of its position in Oreodon culhertsoni, Eiicrotaphus jacksoni, or 3Ierycocharus superbus. The malar bone has a correspondingly large anterior extension, reaching to above the posterior part of the fourth premolar. It does not extend so far in the three species just named. The zygomatic process of the squamosal is more deeply received into the malar bone than in any of the three species mentioned, reaching to below^ the posterior third of the orbit. The larmier in this species is small, and its anteroposterior diameter is more than twice as long as the vertical. More than half of its inferior bor- der is formed by the maxillary bone. As it is exhibited in the specimen, its superior border is formed by the ascendiivg process of the maxillary bone ; whether this is overlapped by the laminar process of the frontal so as to bound the foramen, when in a perfect condition, is uncertain. The posterior edge of the larmier is the lachrymal bone. The external /orawen infraorbitale is on one side double. The supraorbital foramina form notches at the anterior edge of the supraorbital border. The frontal 539 [Cope. foramina are well separated from each other, as in the species of Meryco- chcerus. The space between them is about equal to that between each one and the superciliary border. There is a large postparietal foramen near the parielo-squamosal suture. If the supraglenoid foramen be pres- ent it is not distinguishable in the specimen. The orbit is rounded sub- quadrate, with the inferior anterior angle a little produced. The ascending process of the mandible is relatively elevated. The horizontal ramus narrows rapidly anteriorly, and the symphysis mandibuli is produced so as to rise at a very low angle. The alveolar portion is horizontal. The superior incisors are small and their apices are but little expanded, the external the most so. They are directed vertically downwards. The superior canine is quite small ; its crown exceeds in length that of the first premolar by but little, and is directed a little posteriorly as well as downwards. The roots of the first premolar are not as well distinguished as in many other species, and are united in their extra-alveolar part at least. The same is true of the second premolar. The apex of the cutting edge is in line with the anterior border of the crown ; the rest of the edge rises obliquely backwards. In the third premolar there is a slight bevel in front of the apex, which is much better developed on the fourth. These teeth are more truncate than the corresponding ones of the species of Oreodon and Eucrotaplius, and the larger species of Merycochcerus. The external faces of P-m. i and ii are convex ; that of P-m. iv is concave, but without the reverted vertical borders seen in Oreodon culhertsoni. The first true molar has long roots and a short crown. The last two molars have crowns of a more elongate character, with well developed anterior and middle ridges. The latter are not so prominent as those of the molars of the Merychyun zygomaticus. The inferior incisors are directed upwards at an angle of about 30^. They are similar and closely packed. The inferior canine is in close con- tact with the third incisor, from which it difiers in its larger, leaf-shaped crown. The inferior first premolar is a slender one- rooted caniniform tooth, with narrow crown and acute apex. The second premolar is one- rooted, and has a leaf-shaped crown, with acute-angled apex. The third is two-rooted, and has a wider and nearly symmetrical crown. The fourth is much larger, and its elongate crown laps inside of that of the third. Its low angular apex is median. The last inferior trut rounded ledge of the squamosal bone, which separates it from the postglenoid process. This ledge is much more devel- oped than in any other species of this family known to me. The bulla of the petrous bone is longer anteroposteriorly than transversely, and its anterior and posterior borders coincide with the anterior border of the postglenoid process, and that of the paroccipital process. The postglenoid process is robust, much as in the large species of Merycochoerus, and not compressed as in Merychyus leptorhynchus and M. irenarum. The zygo- matic arch is slender. The elevation of the posterior part of the zygo- matic process of the squamosal has a diflerent form from that seen in the species last named. It is angulate, not rounded. The position of the angle is different from that in M. zygomaticns in being more anterior, marking a point well in front of the anterior base of the postglenoid pro- cess. The border which connects the angle with the supra-auricular crest is then not vertical as in the species just mentioned, but is oblique, and it is also somewhat concave. The malar bone is shallow and stout, with truncate edge below. The squamosal process enters it to below the poste- rior third of the orbit. The postfrontal process is slender, and the post- orbital process of the malar is elongate, meeting the former opposite the middle of the orbit. It is thus longer than in any species of the family known to me. The frontal foramina are separated by an interspace equal to four-fifths the distance between each and the superciliary border. The parieto-squa- mosal suture ascends posteriorly in a nearly straight line to within M. .015 of the posterior zygomatic crest. The posterior squamosal suture then turns directly downwards, reaching the depressed portion of the crest where it bounds the huge mastoid fossa and foramen. The posterior part of the mandibular ramus, shows a regularly convex angular border commencing just below the condyle. The coronoid pro- cess is quite small and the short connecting edge between it and the con- dyle is not excavated below the level of the latter. The articular face of the condyle is directed upwards, and on the internal third, presents a face posteriorly also. The ramus diminishes rapidly in depth anteriorly. The masseteric fossa does not descend below the level of the second true molar, and is not sharply bordered anj'-where. The internal pterygoid fossa on the other hand occupies the entire inner face of the angle between the condyle and the inferior border, and anteriorly to the line of the last inferior molar tooth. The superior true molars have short crowns, as in Eucrotaphus and Oreodon. The anterior and median vertical ridges are very prominent, and the posterior vertical border of the posterior column projects to a slight Cope.] 544 [Jan. 18, extent posteriorly. Enamel smooth. The last inferior molar is not so disproportionately larger than the second as in M. leptorhynchus, arenarum and elegans; and with the second, has little of a prismatic character. No cingula. Measurements. M. Length from occipital condyle to postglenoid process. . . .047 " " " '* " postfrontal process. .. .101 Vertical diameter of orbit 030 Depth of malar bone at middle of orbit 012 " " zygomatic process at posterior angle . .024 Width at middle of supraorbital border 060 " " malar bones 090 " of occipital condyles .032 " " occiput at lateral crests 036 " " " condyles 061 Elevation of occiput with condyles 054 Depth of skull at glenoid surface 058 " " " " orbit, exclus. malar 054 " " mandible at condyle 075 " " " " coronoid 083 " '* " " posterior edge of m. iii 042 Depth mandible at middle of m. ii 028 / anteroposterior 016 I transverse 016 Diameters superior m. ii Diameters superior m. m | anteroposterior 020 I transverse 0155 f anteroposterior 015 I transverse 012 Diameters inferior m. ii Diameters inferior m. m f >^<^teropostenoT 0225 transverse 0115 A second specimen of this species consists of the occipital, parietal, and part of the frontal regions, with the right maxillary bone, and fragments of the left maxillary, of the mandible, etc. The latter demonstrates the position of the infraorbital foramen to be above the anterior border of the fourth superior premolar. The middle line of the occiput presents a keel on its superior half The basioccipital bone between the paroccipital pro- cess is expanded laterally, and is without median angle or groove. Between the bullae it is compressed, and its middle line forms a narrow truncation. Opposite the posterior third of the bulla, this surface ascends at an angle, and gradually widening, spreads into the general flattened convex inferior face of the sphenoid. The anterior part of the sagittal crest is a little better developed than in the typical specimen. The worn teeth indicate an old individual. The canine is large, and the first premolar has its roots well distinguished. The facial plate of the maxillary concave above second premolar. No appreciable diastema. 1881.1 545 [Cope. Measurements. Length of molar series " '* premolars on bases Width of canine posteriorly Diameters p.m. iv I ^^t^^'^l^^^^^^^'^'' I transverse, M. .081 .041 .010 .010 .012 Of this species I have but two specimens, which were obtained from the Ticholeptus beds of Deep river, Montana, by my assistant, J. C. Isaac. merycliyiis elegans Leidy. Proceedings Academy Philada., 1858, p. 24. Extinct Mammalia Dakota and Nebraska, 18G9, p. 118, PI. XI, figs. 1-11. Niobrara river, Nebraska. Herycliyus zygoma ticiis Cope. TicJioleptus zygomaticm Cope, American Naturalist, Feb. 1878. Bulle- tin U. S. Geolog. Survey Territories, 1878, p. 380. This species is peculiar in having the posterior expansion of its zygo- matic arch horizontal instead of vertical. It has a thickened external edge which continues into a strong posterior angle which projects behind the posterior margin of the postglenoid process. The auricular meatus is directed posteriorly in a way quite peculiar, resembling somewhat the position seen in some of the hogs. The malar bone is very prominent. The infraorbital foramen is above the contact of the third and fourth superior premolars. The larmier is large and its maxillary border descends posteriorly. In size this species is between the M. elegans and the M. medius. If my identification of New Mexican specimens is correct, this species diflers from the M. medius in the much less production of the premaxillary re- gion, besides the smaller size. Ticholeptus beds of Deep river, Montana ; J. C. Isaac. Merychyus medius Leidy. Proceedings Academy Philad'a, 1858, p. 25. Extinct Mammalia, Dakota and Nebraska, 1869, p. 119, PI. XI, figs. 12-14. Cope U. S. Expl. Surv. W. of 100th Mer., G. M. Wheeler, iv, pt. ii, p. 324. Niobrara river, Nebraska, Hayden ; Santa Fe, New Mexico, Cope. iwrerycliyus major Leidy. Proceedings Academy Philada., 1858, p. 26. Extinct Mammalia, Da- kota and Nebraska, 1869, p. 121, PI. X, figs. 15-16. This species, known hitherto from Leidy 's descriptions of four of the superior molars, is the largest of the genus, and perhaps of the family. More information regarding it is much to be desired. Headwaters of the Niobrara river ; from Loup Fork beds, according to Hayden. Cope. I 546 [Jan. 18, LKPTAUCHEWIA Leidy. Extinct Mammalia of Dakota and Nebraska, 1869, 122. Proceedings Academy Philad'a, 1856, 88, (nomen nudum), loc. cit. 1656, 163 (nomen nudum). As already remarked by Leidy, this genus is characterized by the pres- ence of enormous vacuities of the superior surface of the muzzle. The genus might be described as lacking the usual superior osseous wall of the nasal cavities and maxillary sinuses. The generic diagnosis is as follows: Otic bullae inflated. Four premaxillary teeth. Nasal bones excessively contracted, leaving a wide interspace between them and the maxillaries. Symphysis mandibuli coossified. This genus has but a short range in time, not having been yet found out of the Ticholeptus beds. It shows in its deficient ossification, and smaller size, that this line of the family was approaching its extinction, its deca- dence having already commenced in the genus Merychyus. The genera which follow in systematic order, Cyclopidius and Pithecestes, exhibit the last steps in the downward course. I. Infraorbital foramen above P-m. iii. " Three inferior incisors ; nasal sinuses to middle of orbit ; true molars .043 ; skull .135." (Leidy) L. major, "Nasal sinuses not extending so far posteriorly as in L. major; true molars .032 ; skull, .101." (Leidy) L. decora. "Nasal sinuses reaching to front of orbit ; true molars .020 ; skull .085." (Leidy) L. nitida. lieptaudienia major Leidy. Proceedings Academy Philad'a, 1856, p. 163 ; 1857, 89. Extinct Mam- malia, Dakota and Nebraska, 1869, p. 124, PI. XII, figs. 1-5. Tributaries of White river, Nebraska. Leptaucheiiia decora Leidy. Proceedings Academy Philadelphia, 1858, p. 88 ; 1857, p. 89. Extinct Mammalia of Dakota and Nebraska, 1869, p. 127, PI. XII, figs. 6-20. Tributaries of White river, Nebraska. Leptaiidienia iiitida Leidy. Extinct Mammalia of Dakota and Nebraska, 1869, p. 129 ; PI. XII, figs. 21-22. White Earth creek, Dakota, tributary of the White river. CYCLOPIDIUS Cope. Proceedings American Philosophical Society, 1877, p. 221. Brachy- meryx Cope, Ibidem, p. 220. Dental formula : I. f : C. | ; P-m. \ ; M. f . Premaxillary bones much reduced ; mandibular rami coossified. Otic bulla inflated. Prelachrymal vacuities present, and confluent with enormous nasal vacuities, which are due to the excessive reduction of the nasal bones. Orbit closed behind. 547 LCope. This genus is Leptauchenia without superior incisor teeth, and witli but two on each side below. I originally asserted the presence of superior incisor teeth, and it is true that there is in early life a minute tooth in each premaxillary bone, as indicated by the alveoli in a specimen which con- tains the full deciduous molar dentition. I have not seen the teeth them- selves, and it is evident that they are early shed. In an adult specimen of C. simus it seems that the alveolar portion of the premaxillary bone has been absorbed. The meatus auditorius externus occupies a more elevated position in this genus than in any other of the family. It is also directed somewhat posteriorly. There are postparielal foramina. The cerebral hemispheres are not large, and scarcely rise above the plane of the summit of the large cerebellum. Convolutions three on each side, weakly defined. The concavity of the superior border of the premaxillary bones, to- gether with their upward production, leads me to suspect that the exter- nal nares were superior in position. This is the indication of an aquatic habit of life, such as is led by the hippopotamus. Like that animal, the nostrils in Cyclopidius were probably valvular to prevent the ingress of the water. The animals probably passed much of their time in the water, and the nostrils could be brought to the surface for the purpose of respira- tion, while the remainder of the head and body remained concealed. The prominent rim of the auditory meatus suggests a similar valvular closure of the organ of hearing, and is also a provision for its easy approximation to the surface of the water when necessary. The milk dentition is like that of Artiodactyla in general. That is, in the superior series the third molar is more elongate and complex than its permanent successor, and the fourth is like the first permanent true molar in constitution. In the inferior series the anterior three teeth resemble the permanent premolars, while the fourth is trilobate. In the loss of the incisor teeth and the subprismatic molars, we observe in Cyclopidius the same evidences of specialization already known in other types of Ungulates. I know of but two species of Cyclopidius. Cyclopidius simus Cope. Proceedings American Philos(»phical Society, 1877, p. 221. Brachy- meryx feliceps Cope, Ibidem, p. 220 (immature). The specimens of this species in my possession embrace a complete skull with one zygoma and half of the brain-case wanting ; a left maxil- lary bone with all the teeth ; and three mandibular rami with dentition, all of adults. Of immature individuals, I have two muzzles with denti- tion of both sides, and six mandibular rami ; in all, parts of thirteen in- dividuals. The following description of the skull is taken from the speci- men first named, which is the type of the species. The cranium is wide and depressed, and the muzzle is short. The pro- Cope.J 548 [Jan. 18, file descends at the orbits into the nasal vacuities, which cause a deep ex- cavation of the facial plate of the maxillary region. The small nasal bones form a promontory below the level of the orbits, whose supe- rior borders are convex. The maxillary bones rise at the end of the muzzle, forming, probably, with the confluent premaxillaries, a sub- quadrate projection. The superior side of this process is concave on its interior aspect forming a curved suture of an expanded nasal bone. Its anterior edge is also concave on their inner side, as though adapted to a forward-looking nareal opening. This anterior border is produced downwards into a free conical process which bounds the canine alveolus in front. This I suppose is all that there is of the alveolar portion of the premaxillary bone. The corresponding part of the other side is lost. There is a well-marked preorbital fossa. Its supero-interior border bounds the huge nasal vacuity on each side. The nasal bones form a narrow promontory, with convex superior face, which extends a little beyond a line connecting the middles of the preorbital fossae. The vacuities exca- vate the frontal bones as for back as a line connecting the middles of the supraorbital borders. The frontal bone is thus of a ^- shape. The anterior temporal ridges are well defined, but do not reach the free edge of the frontal bone. Their union into the sagittal crest is gradual. The brain-case is moderately elongate, the postorbital process of the malar bone marking the middle of the total length. In profile the posterior part of the skull is nearly straight. The sagittal crest is gently convex, and is not so deeply bifurcated posteriorly as in most other forms. The posterior temporal crests are expanded laterally, and continue well developed to above the meatus auditorius, into the superior edge of the zygoma. They are not continued downwards on the occiput, as in most of the other genera of the family, but resemble the species of Merychyus more than any others in this respect. The temporal fossa has a wide floor, due to the lateral extension of the meatus auditorius, and the glenoid portion of the squamosal. The superior edge of the zygomatic process of the squamosal is little elevated, and is regularly convex. The process is not produced as far anteriorly as the posterior border of the orbit. The malar bone is remarkable for its depth, exceeding in this respect any species of the family yet known. Its external face slopes obliquely outwards below, but not very much, and is slightly and uniformly convex. Its inferior edge is thickened and descends anteriorly, and then thins and rises continuous- ly to the zygomatic process of the squamosal. The occipital aspect of the skull is wide and low. Its superior region is slightly convex and roughened on each side of the median line. From and below this valley, the middle line presents a sharp carina, which dis- appears in a narrow convexity above the foramen magnum. Between this convexity and the meatus auditorius, the surface is concave. The occipital condyle is small, and the exterior half is more extensive than the posterior half. The paroccipital process is large. Its base diverges from the occipital condyle, and is adherent by its anterior face to the otic bulla, 18S1.I 549 fCope. without intervening ridge. The posttympanic mass is broken away. It is inferior in position to the auricular meatus. Tlie latter, being directed posteriorly, is considerably produced behind the postglenoid process, leav- ing a wide postglenoid fossa. The postglenoid process is rather small, and its posterior face is entirely covered by the tympanic bone, while its interior edge is in close contact with the otic bulla. The bulla is of enor- mous size, and is a slightly compressed oval placed anteroposteriorly. It fills the entire space between the postglenoid process and the basicranial axis, and reaches anteriorly almost to the line of the anterior border of the glenoid region. The pterygoid process adheres to its internal wall for half its length, and it sends forwards on the external side of the ptery- goid, a narrow acuminate apex. The internal extremity of the glenoid cavity is concave, and the surface descends, forming a robust peduncle, as large as the postglenoid process, to which the anterior part of the otic bulla is attached. This is a character I have not seen in any other species of the family. A wide surface, continuous with that of the glenoid face, extends on the external side of the pterygoid ala of the sphenoid, to the angle where it unites with the pyramidal process of the palatine. It there termi- nates abruptly, but the external angle marks the end of a ridge, which ex- tends upwards and forwards to the postorbital process of the frontal. An- terior to this line the cranial wall is concave ; posterior to it, convex. The processi pyramidales are divergent, and have thickened and rounded infe- rior edges. The maxillary* bones are produced a little beyond their bases, leaving a notch between. The palatal surface is uniformly moderately concave. The incisive foramina are large ; the septa are wanting in my speci- mens, perhaps accidentally. The infraorbital foramen is above the middle of the fourth premolar tooth. The frontal foramina are further apart than in any other species of the family, being equidistant between the median line and the supraorbital border. There is an internal orbital foramen be- low the postorbital process, as in other species of the family. There are three postparietal foramina, two of which are on the squamosal suture. Below the anterior of these two is a large postsquamosal foramen. No supra or postglenoid foramina. The meatus auditorius externus looks equally externally and posteriorly. It is large and of oval out- line, the long diameter being parallel to the superior border, which is the usual suprameatal crest. Its tympanic or anterior border is very prominent, while the posterior border is a little less so. A posttym- panic tuberosity marks the middle of the inferior edge. Posterior to the meatus is the rather large mastoid foramen, which is above the in- ternal base of the paroccipital process. The basicranial bones being lost, the characters of the basal foramina are not determinable. The posterior nares are deeper than wide. The palatonareal border is a Gothic arch, of which the apex is opposite the posterior border of the last molar tooth. I perceive no palatal foramina. The median and posterior nasal sutures remain. The latter is a V with Oope.l 550 [Jan. 18, the apex opposite to the frontal foramina. Lambdoidal suture confluent. The malosquamosal suture marks the posterior edge of the posterior orbi- tal rim at the middle of the orbit. The parieto-squamosal suture has an in- ferior position in front. Opposite the front of the postglenoid process it converges inwards in line for the occipital bifurcation, and is continued as the parietooccipital suture, nearly to that point. The squamosal border, however, extends in a Z-form to the posterior temporal crest half-way be- tween the bifurcation and the meatus auditorius. It embraces an area of the posterior face of the skull, and the posterior half of the rim of the au- ricular meatus. The typical specimen presents only the alveoli of the canine and first premolar teeth ; otherwise the dentition is perfect. The crowns of the second and third premolars are obliquely quadrate in horizontal section, both a little wider posteriorly than anteriorly. This is due to the presence of a half crescent of the internal side, whose posterior horn is attached to the external wall, while the anterior is free. The external faces of these premolars is slightly convex ; of the fourth premolar is slightly concave. The first true molar is decidedly smaller than the second, and the second is smaller than the third. The external sides of the external columns are flat in the flrst true molar, but become more concave on the third. The anterior edges of the columns project; forming ridges; or in section, project- ing angles. No intermediate ridges, nor cingula. The third superior true molar has a prismatic crown, no roots being visible in either of the adult specimens, of which the typical one is rather old, as indicated by the wear of the teeth. In the latter specimen the roots of the second true molar are apparent, although the crown is elevated. The first true molar is not pris- matic, although the crown is not low. The specimen represented by the left maxillary bone contains the teeth which are wanting from the typical one. The section of the crown of the canine is a semicircle, the truncate face being posterior internal. It is not a large tooth, and is separated from the first premolar by a diastema equal to its diameter. The first premolar is one-rooted, the root with a groove on the internal side. The section of the base of the crown is a triangle, the faces being anterior, external, and posterointernal. Its inner face is concave above the base. None of the separate mandibular rami are complete, all lacking the angle and condyle. The former is full and round, judging from a frag- ment in my possession. The ramus diminishes regularly in depth for- wards. The sympliyseal region is short, and its anterior face is very steep, except at the alveolar region, where it is everted forwards. No trace of suture. The internal pterygoid fossa is large and strongly marked, so that the inferior edge of the ramus is inverted, so that the surface is convex ex- ternally. The last molar is placed somewhat obliquely. The first and second premolars are directed outwards and forwards, and the incisors directed forwards. There are two incisors on each side of the symphyseal line. They are very small and subcylindrical, and are closely packed between the canines. 188-l.J 551 [Cope. The canines are much larger, with cylindric root and flat, incisor-like crown. The first premolar is still larger, and is of about the same form as the canine, from which it is only separated by a slight divergence of the crowns. There are no diastemata. The second premolar has a compressed triangular crown, with a median ridge on the internal side. Its long diameter is diagonal, running outwards posteriorly. The long axis of the third premolar is similar, while the other teeth are more nearly in line. In the third premolar the fossa interior to the median internal heel is much deeper than that posterior to it. The corresponding fossa is still larger in the fourth premolar, while the crown has a heel in the form of a trans- verse curved crest, separated from the median heel on the inner side by a fissure. The true molars increase rapidly in size posteriorly, but not so abruptly as in the Pithecistes brevifacies. The internal crescents are very flat, and the posterior edges of their columns project moderately. The ex- ternal crescents are very convex. The prismatic character of the teeth in- creases much posteriorly, so that the roots of the third tooth are short, and the crown long. The enamel is minutely rugose. The third superior temporary molar has two pairs of crescents. The an- terior pair are, however, not so well developed as the posterior pair and the two valleys are soon obliterated by wear. The crescents are equal in the fourth temporary molar. The fourth permanent premolar is protruded at least as soon as the third true molar, sooner than the posterior column of the latter. In this it difiers from the Oreodon culbertsoni, where the last true molar is protruded first, and is a cotemporary of both the third and fourth deciduous molars ;* and the 0. gracilis, where the last true molar is a cotemporary of the third deciduous. In the inferior temporary dentition, the lobes of the last molar are sub- equal, the posterior one being a little the larger. The protrusion of the last true molar is also probably delayed until the shedding of the deciduous series, as in the superior series ; but my specimens are either very young or fully adult, and therefore I cannot demonstrate this point as fully as in the case of the superior series. Measurements of Skull. M. Length from condyle to front of canine inclusive. 117 " otic bulla (axial) 010 " palatonareal notch, 0575 " " " " anterior line of glenoid cavity. .038 Depth of occiput, including condyle 041 " at middle of orbit, exclusive of teeth 037 " " infraorbital foramen " " 016 " " premaxillary border " " 023 Width at " " above .022 " between orbits 038 * Leidy. Ancient Fauna of Nebraska, 1853, p. 51, PI. IV, figs. 1, 2. Cope.] 552 [Jan. 18, Measurements of Skull. M. Width malars below orbits 086 ** " zygomata at middle 092 " " auricular meatus 070 " of occipital condyles 0275 ' ' at middle of last molars inclusive 049 " " second premolars inclusive. . .030 r anteroposterior 023 Diameters otic bulla < transverse 018 (vertical 019 Diameters of nasal bones | ^^°Sth of fragment of. 024 t width at base 0125 Length of dental series 062 " ■ " premolar series 025 " " true molar 033 Diameters P-m. m i anteroposterior 007 Uransverse 0065 Diameters m. i / anteroposterior 0085 *- transverse 0085 Diameters m. m. / antereposterior 0146 <- transverse (greatest) Oil Depth of mandibular ramus at m. iii 031 " '*P-m. iv 019 Length of symphysis , 0245 " " premolar series 022 " true molar 035 " " of total dental series 063 Diameters P-m. iv 5 anteroposterior 0085 ( transverse 006 Diameters m. i \ anteroposterior 009 i transverse 0068 Diameters m. iii \ anteroposterior 016 (. transverse 0062 The second specimen with permanent dentition is of smaller size than the type, and the canine teeth are small. It may have been a female. The dental series, including the canine, measures M. 0.59 ; the premolar scries, 0.23 ; the true molars, 0.31. The number of specimens of this animal found in the restricted area of the Ticholeptus bed of Deep river, Montana, shows the former abundance of the species. It was probably gregarious, in the manner of the other Oreodontidse. We can depict it as seeking the swamps of the shore for its vegetable food, and spending much of its time in the water when not feed- ing. It was doubtless a good swimmer, and the characters of its feet will be sought for with interest for light on this point. The use of the huge superior nasal vacuity of the skull of this genus and Leptauchenia can 188-1.1 558 fCope. only be guessed. Perhaps it supported an inflatable bladder like that of the ciested seal, or a swollen muzzle like that of the saiga antelope. Cyclopidius emydiiiiis^ sp. nov. This species is represented in my collection by a nearly perfect cra- nium. It indicates an animal of about the same size as the C. simus. The differences between the two species may be enumerated in advance of the detailed description. Firstly, the external vertical ridges or crests of the true molars are directed obliquely forwards so as to overlap the ex- ternal wall of the anterior crescent much more extensively than in C. simus. (2) The crowns of the true molars have a relatively greater trans- verse diameter. (3) There is a peculiar process at the external base of the otic bulla, between the paroccipital and postglenoid processes, which may be called the subtympanic process. (4) There is no median occipital keel. (5) The maxillary bone is prolonged posterior to the last superior molar, which it is not in G. simus. (6) The oblique orbitosphenoid ridge is wanting. (7) The otic bullae are shorter and wider in their form. This character will require confirmation by examination of many individuals. The skull is singularly depressed and expanded laterally, so as to pre- sent an outline not unlike that of some river turtles. The orbits are in the anterior half, and look forwards and upwards, as well as outwards. The muzzle is short, so that its lateral borders approximate rapidly to a narrow truncate extremity. The maxillary borders do not contract quite so abruptly, and are visible outside of the canthus rostralis, when the skull is viewed from above. The brain-case is depressed, and is expanded posteriorly, and narrowed at the anterior line of the zygomatic foramina. The posterior temporal ridges are much expanded, forming a wide rim round the brain-case posteriorly, which is continued into the squamosal processes of the zygoma on each side. The anterior temporal ridges ap- proach each other very gradually on the middle line, and only reach the union into a sagittal crest a centimeter posterior to the frontoparietal su- ture. The edge of the crest is truncate, and it is not bifurcate posteriorly, as in most Oreodontidae. The occiput is broad and low, and differs in character from that of most other members of the family. Its posterior face is flat, only interrupted by a fossa on each side, just within the posterior edge of the meatus audi- torius externus. This edge is continued downwards into the external bor- der of a distinct mastoid process, which is also the external border of the occiput, deflected a little forwards. The paroccipital process is flat at the base, and is applied to the external half of the otic bulla. Its free extrem- ity is subround. The mastoid process forms a prominent ala of its exter- nal side, having a transverse width equal to that of the base of the par- occipital. Its inferior edge is truncate obliquely outwards and downwards to a subacute angle. The occipital condyles are relatively small. The external meatus of the ear looks outwards and backwards at an angle of 45^ to the middle line. The prominent edge of the mastoid pro-. Cope.] 554 [Jan. 18, cess is directly below its anterior border. Thus the tympanic bone is directed obliquely downwards and forwards. Posteriorly it is separated by a groove from the mastoid process. Anteriorly it is separated by a fossa from an osseous mass which occupies the space between it and the post- glenoid process. Before the skull was reconstructed from its fragments, this mass was observed to be entirely distinct from the postglenoid process, which it equals in height. Continuous with it, there descends another osseous body to near the line of the extremity of the mastoid process, with a truncate inferior edge, which is separated from the otic bulla by an open groove. The stylohyal ligament is probably inserted into a fossa at the anterior extremity of this groove. The postglenoid process is low and more extended transversely. The anteroposterior diameter is small. The glenoid surface is much extended transversely and terminates externally in a slight thickening. The zygomatic process of the squamosal bone is at first expanded horizontally and has a low convexity of the thin superior edge. Its vertically compressed portion is entirely supported by the ma- lar, and does not extend so far forwards as the anterior edge of the zygomatic foramen. The malar bone is remarkable for the depth of its suborbital portion, which fully equals the diameter of the orbit. Its infe- rior edge presents a thickened angle downwards below the anterior part of the last superior molar. Its superoanterior angle terminates in a promi- nent rib of the side of the face, which extends along the inferior edge ot the facial vacuity. Beneath the anterior part of the latter the face is con- cave. Above this concavity the ascending plate of the maxillary is con- vex in the vertical section, turning inwards at the apex to unite with the lateral part of the extremity of the nasal bone. The preorbital fossa is small and looks forwards and upwards. The otic bullae are larger than in any other Oreodontid. They are of a short oval form, somewhat truncate anteriorly and posteriorly. Thus they diflfer from those of C. simus, where they are elongate-oval. They only reach as far anteriorly as the middle of the internal extremity of the glenoid surface ; while in (7. simus they reach the line of the posterior outline of the zygomatic foramen. Thej^ terminate near the inferior inter- nal point, in a little acute osseous apex, which is smaller than in C. simus. The bullae approach so closely together that the bassioccipital is much nar- rowed, and the sides of its inferior surface are excavated so as to reduce the middle line to a narrow acute keel. The lateral excavations follow the posterior internal base of the bullae, leaving a median table, which is itself excavated by a shallow fossa, which extends from the median keel to the foramen magnum. The median keel disappears anteriorly. The sphenoid is protuberant downwards as a narrow convex rib, which rises and disap- pears in the presphenoid. The descending sphenoid ala forms the posterior boundarj' of the posterior nareal trough, and makes a strong angle with the pyramidal process of the palatine, which is turned outwards. The pterygoid squama terminates in an apex which points downwards and posteriorly towards the apex of the otic bulla. The palatonareal border is mi.\ 555 [Cope. V-sliaped, and is in line with the posterior edge of the maxillary bone. The latter projects beyond the last molar tooth as far as the anteroposterior diameter of the latter. It has no projection in the C. simus. There is no notch between the maxillary bone and the processes pyramidalis of the palatine. The palate is of nearly equal width from the last molar to the third premolar ; its roof is gently concave posteriorly ; nearly flat anteri- orly. The premaxillary bone is a narrow strip which rises nearly vertical- ly from its short alveolar border, and is curved outwards above in agreement with the expansion of the anterior edge of the maxillary, to which it is united b}'- simple suture. The nasal bones are of remarkable form. Together they enter the anterior part of the frontals in a V-shape, and extend forwards in a narrow shaft. Opposite the anterior borders of the orbits the shaft begins to widen gradually, and the surface to flatten, until they reach the posterior angle of the ascending part of the maxillary. Each one then expands outwards, terminating in a semi-disc which fits the concavity of the superior edge of the maxillary above mentioned. The entire shape of the nasal bones is that of a spade with a triangular apex to the handle, and the short blade at the opposite (anterior) extremity. The frontal bone is V-shaped, the angle posteriorly directed, and engaged between the parietal bones, and each branch terminating above each orbit. Narrow prolongations extend anterior and posterior to the orbit, joining the lachrymal and malar bones respectively. Its median suture is, like that of the nasal bones, well defined. The alisphenoid and parietal have extensive •connection. The parietosquamosal suture is horizontal in front ; it then gradually rises. It is not associated with a ridge as in some other species. The occipital forms the posterior five millimeters of the sagittal crest. The nasal opening is subtriangular, with the base above, and is directed unteriorly. The facial vacuities are enormous, and excavate the frontals to a point which make the anterior third of the orbit's diameter. They are •only separated on the median line by the very narrow isthmus of the nasal bones. The infraorbital foramen is above the anterior part of the fourth surperior true molar. The frontal foramina are small, and are not symmetrical. That of the left side is half-way between the median suture •and the superciliary border ; the other is nearer the superciliary border. No supraglenoid foramen. Postsquamosal present ; that part of the •cranial walls is injured. The anteroposterior diameter of the orbit ex- •ceeds the vertical. The auricular meatus is the largest known in the family, and it has a prominent border and regularly oval outline. Its long diameter rises posteriorly from the horizontal. It is more lateral- ly and less posteriorly directed than on the typical and only skull of C. simus. The foramen magnum has an openly angulate superior bor- der. Jugular, condyloid, and carotid foramina not obvious, owing to the ■close contact of the otic bulla with surrounding bones. Foramen ovale larger than the F. lacerum anterius, and external to it in position. F. ro- iundum still larger, inferior in position, bounded on the external side by a Cope.] 556 [Jan. 18, tuberous projection of the angle from the anterior edge of the glenoid sur- face. There is a deep fossa at the internal base of the postglenoid process, which possibly enters a foramen. No postglenoid foramen. Although the skull of the Gyclopidius emydinus is more robust than that of C. simus ; the length of the tooth-line is the same. The in- cisive edge of the premaxillary bone displays one empty alveolus, from which the single incisor was easily shed. The canine is not large, and the base of the crown has a regularly convex anteroexternal face ; apex lost. The diastema posterior to it is equal to its diameter. The crowns of the premolars are worn ; they are of about the size and proportions of the C. simus. The true molars differ, as I have already pointed out, in their greater transverse diameter, and the greater anterior prolongation of the anterior horns of the posterior external crescents. The deep notch which is enclosed between this fold and the wall of the crescent in front of it is filled with cementum. As to the form of the true molars, the transverse diameter of the first considerably exceeds its anteroposterior diameter ; in the C. simus the former diameter is equal to the latter. In C. emydinus the last true molar is as wide as its length without the heel ; in the C. simus, the transverse diameter is much less. In C. simus the heel is more prominent, and is recurved into a vertical ridge, which is wanting in the C. emydinus. In C. emydinus this tooth shows but little of the pris- matic character, as the roots are of usual length. The lower jaw of this species is not yet known. Measurements. M. Length of skull along base 129 Length from condyles to posterior edge of zygomatic foramen 042 Length from condyles to palatonareal foramen 063 " " " " line of last true molar 071 *' " occipital crests to line of orbits 074 " " " " facial vacuities 084 " " " " "ascending process of maxillary bones 115 Length from occipital crest to free end of nasal bones. .126 Elevation occiput, including condyles 045 " of front at middle of orbit, without molars. . .035 " " maxillary bone at P-m. iii 015 " " P-m. i .025 Width of skull at occipital condyles 0675 " ** " " superior edge of meatus auditorius.. . .057 " " " *' middle of zygomatic foramina 092 " " brain -case at middle of zygomatic foramina. . .029 " skull at orbits 083 " " ,, between orbits 047 " " muzzle at superior edge of nares 0215 1884.] [Cope. Diameter external nares MeamremenU. M. f vertical 014 1 transverse above 017 Diameter of a facial vacuity [ ^'^t^^oposterior 030 C tranverse 013 Diameter of orbit \ ant^^Posterior 023 ' vertical 019 .053 .026 vertical 0095 transverse 013 vertical 009 anteroposterior Oil [ vertical 025 Diameter of otic bulla \ anteroposterior 025 \ transverse 022 Width between canine teeth 008 last true molars 0285 Length of dental series 065 " true molar series 0343 " premolar 0254 Diameter of zygomatic foramen | f"^^^^^^^^^^^^^^ <- transverse Diameter of foramen magnum | Diameter of meatus auditorius Diameters P-m. ii anteroposterior 0056 transverse 0050 Diameters P-m. ^ anteroposterior 0070 (. transverse 0080 Diameters ™. j ^ anteroposterior 0075 ( transverse 0110 Diameters m. iii anteroposterior 017 transverse (with external rib). 012 The only specimen of this remarkable species knoM-n to me was found in the valley of Deep river, Montana, by my assistant, Mr. J. C. Isaac. The wear of the true molars shows that the animal was of full age, though not old. PITHECISTES Cope. Proceedings American Philosophical Society, 1877, p. 219. This genus represents the final term in the decadence of the once powerful and numerous family of the Oreodontidae. It is unfortu- nately established on a mandibular ramus only, and although some maxillary bones are referred to it with much probability, they are not preserved in such a way as to demonstrate the presence of the large nasal sinuses characteristic of Leptauchenia. I, however, suspect that they occur. The genus further resembles Leptauchenia in the coos- sification of the mandibular rami, and the reduction in number of the in- cisor teeth. In P, brevifacie-H there is but one inferior incisor tooth on each side. As reduction in the superior incisors usually precedes reduction in Cope.] 558 [Jan, 18, those of the lower jaw, I suspect that the former were absolutely wanting in this genus. If so, we have in the Oreodont line the same process of re- duction above, as has taken place in other lines of Artiodactyla at the latest or modern stage of their history. In Pithecistes the inferior canine is caniniform, and masticated in con- tact with the superior canine, owing to the great abbreviation of the sym- physeal region. The diagnosis of the genus is as follows : Inferior premolars three ; incisors one. Canine caniniform, masticating with the superior canine. No diastema. Symphysis coossified. Two species are referred to this genus without conclusive evidence as to the number of their premolars. It is probable that they have but three, since their superior fourth premolars are of reduced size and incomplete type of form. Pittiecistes torevifacies Cope. Proceedings American Philosophical Society, 1877, p. 219. Ticholeptus beds of Deep river, Montana. Discovered by J. C. Isaac. Pittiecistes decedens Cope, sp. nov. Established on a right maxillary bone, which contains the fourth pre- molar, the first and second true molars, and part of the alveolus of the third true molar. The last named tooth was not probably entirely pro- truded. This, with the moderate wear of the fourth premolar, indicates that the animal was fully grown, though young. The species differs from all the members of the family whose dentition is known to me in the small size and simplicity of structure of the fourth premolar. The internal crescent of this tooth bounds only the posterior three-fourths of the external wall, and therefore leaves the anterior edge of the latter free. It is, moreover, not very convex, and its edge is not so elevated as is that of the external wall. The latter is flat on the external side, and its anterior marginal angle corresponds with the point of junction of the anterior extremity of the internal crescent. The true molars have the anterior horns of their crescents prominent, being sections of well-de- veloped vertical columns. In this they differ from those of the P. hetero- don, where these ridges are very weak. The malar process of the maxillary bone is robust and prominent, and begins to expand opposite the first true molar. It presents a tuberosity downwards. The infraorbital foramen issues above the front part of the fourth premolar. Measurements. M. anteropost erior 006 transverse 005 Diameters P m. iv Diameter m. i \ anteroposterior 0087 ( transverse 0077 Diameters m- ii anteroposterior 0115 transverse 008 Ticholeptus beds. Deep river, Montana. J. C. Isaac. 559 [Cope. PitliecisteN lieterodoii Cope. Cydopidins Jieterodon Cope, Proceeds. American Philos. Society, 1877, p. 22. In this species the fonrth premolar lias the same form as in P. decedens, but the first true molar differs much in the more prismatic shape, and the absence of the external vertical ribs. It is quite possible that it does not belong to this genus. Ticholeptus beds of Deep river, Montana. J. C. Isaac. At^UIOCHCERUS Leidy. Proceedings Academy, Philadelphia, 1850, p. 121. Extinct Mammals Dakota and Nebraska, 1869, p. 131 (as family Agriocharidm). Orbit not closed behind. Fourth superior premolar with two external Vs. Fourth inferior premolar like true molars. Otic bulla inflated. Pre- maxillary bones distinct ; no vacuities in the facial bones. This genus commences cotemporaneously with the genus Oreodon, and persists longer, viz. : to the close of the John Day epoch. It represents a distinct line of succession from that which we have been considering, and one which contains but two known terms. Next to Agriochcerus comes, in this line, the genus Coloreodon Cope, which outlasted its predecessor so far as is yet known. It commenced with it in the John Day epoch, and con- tinuing into the North Fork beds, which are of later age, did not ap- pear later. This series Leidy regarded as a family distinct from the Oreo- dontidae. For the present I prefer the view of Gill, that it constitutes a subfamily, the Agriochoerinse. This genus presents us with one of the very few cases in the suborder Artiodactyla, in which the last premolar approaches (above) or accom- plishes (below) identity of structure with the true molars. This degree of complication was attained at the same period by both the equine and rhinocerontic lines of Perissodactyla, and all existing members of that order exhibit it. In the Agriochoeridae it made a beginning, but soon dis- appeared from the earth, and no Artiodactyle has developed such perma- nent premolars successfully since. In the characters of the skull this genus is less robust than the Oreodon- tidae ; but the general skeleton remains unknown. Five species have been described which are referable to this genus, and two others are now added. One of the former is without premaxillary or su- perior incisor teeth, and I therefore regarded it as representing a distinct genus under the name of Merycopater. It, however, appears that no speci- mens exist in our museums which exhibit this part of the skull in other species of the genus, so it is absolutely uncertain whether Agriochcerus possesses those teeth or not. The species may then be distinguished as follows : I. Otic bullae compressed, base anteroposteriorly ovoid. «. Foramen infraorbitale above junction of P-m. iii and iv. Cope.l 560 [Jan. 18, Front narrower; internal wall of fourtfi premolar not com- plete A. antiquus. Front wider ; skull shorter and higher ; internal wall of inferior P-m. iv complete ,.,..A. latifrons. aa. Foramen infraorbitale above junction of P-m. ii and iii. Front medially concave, laterally descending to orbits ; sagit- tal crest short A. irifrons. II. Otic bullae mammiform with triangular base. Front convex ; nasal bones acute posteriorly ; fourth inferior premolar complete ; infraorbital foramen above junction of P-m. iii and iv A. guyotianus. III. Otic bullae oblong, constricted at the middle. Infraorbital foramen above junction of P-ms. ii and iii ; front plane ; nasal bones truncate posteriorly ; postglenoid pro- cess robust A. ryderanus. Besides the above, Leidy has described an A. major* as near to the A. antiquus, but of larger size. Marsh has described a small species from the Uinta formation under the name of A. pumilus.\ Lydekker figures and de- scribes a superior molar tooth from India as probably belonging to this genus.:}: It is stated by him to have been found in the earlier pliocene for- mation. If this determination be correct, it represents the latest known species, as the A. pumilus of Marsh is the earliest. Owing to incomplete- ness in the descriptions of these species I cannot include them in the above synoptic table. Agrioclioeriis antiquus Leidy. Proceedings Academy Philadelphia, 1850, 121 ; 1853, 393 ; 1854, 157 ; 1857, 89. Ancient fauna of Nebraska, 1853, p. 24, PI, I, figs. 5-10. Bronn Lethaea Geognostica, 1856, 933 ; Leidy Extinct Mammalia Dakota and Nebraska, 1869, 132, PL XTII, fig. 4. White River epoch of Nebraska and Dakota. Agrioclioerus major Leidy. Proceedings Academy Philadelphia, 1856, p. 164; 1857. 89. 1 Eucro- taphus auritus Leidy, Owen's Report Geological Survey, 1852, p. 563, PI. XV, figs. 1-3. Ancient Fauna of Nebraska, 1853, p. 56 ; PI. VII, figs. 1-3. Bronn Lethaea Geognostica, 1856, 931. White River formation of Dakota and Nebraska. Agrioclioerus latifrons Leidy. Proceedings Academy Philadelphia, 1867, p. 32. Extinct Mamm. Dakota, Nebraska, 1869, p. 135, PI. XIII, figs. 1-3. White River epoch of Dakota and Nebraska. ♦Extinct Mammalia of Dakota and Nebraska, p. 134. t Amer. Journal Science and Arts, 1873, p. 250. JPaleontologica Indica. 188J.1 561 [Cope. Agrioclioerii§ trifrons sp. nov. This species is known to me by a single cranium of an immature indi- vidual. It lacks of perfection only the basioccipital, the pterygoid, and the alveolar border of the premaxillary bones. It retains the third and fourth deciduous premolars, while the third true molar is still in its alveo- lus, where it is exposed in place. Although the specimen is immature, its characters will not permit me to place it with any other species known to me. I have specimens of like age of the A. guyoiianus, and these are quite different. From A. ryderanus it differs in the form of its otic bulla, etc. The muzzle and front form a flat horizontal profile, while the parietal re- gion is convex. The profile descends gently to the supraoccipital border, or inion. The muzzle is compressed above and below the canine alveolus, and there is a concavity above the third and fourth premolars, and behind the foramen infraorbitale above this fossa the lachrymal region is con- vex. The nasal bones are lost, so that the form of their posterior suture cannot be ascertained. The frontal bones are gently concave in transverse section between two lines produced forwards from the anterior extremities of the temporal ridges, that is at the postorbital constriction of the cranium. These lines are represented by a rounded longitudinal angle, from which the frontal bone descends to the superciliary border on each side. A trace of this form is seen the A. ryderanus. The supraorbital borders diverge outwards and backvrards to the postorbital processes. These are prominent horizontally, and are abruptly decurved at the apex. The temporal ridges enclose an urceolate area, having a gentle convexity in their direction be- fore they unite at a point more posterior than in the other species, that is above a line connecting the anterior borders of the postglenoid processes. The malar bone is slightly concave on the external face, and is mode- rately deep, and not thick. The squamosal part of the zygoma is rather slender, and does not rise above the postglenoid process. Its superior edge continues without interruption into the posterior temporal crests, and so into the supraoccipital. The postglenoid process is like that of A. guyo- iianus, narrow and produced downwards. Paroccipital lost. The otic bulla is large, its anterior edge extending anterior to the postglenoid pro- cess. It is nearly twice as large as in M. guyoiianus, and extends much further forwards. It presents two flat sides, one external, the other out- wards aud forwards, and a convex side inwards and backwards. These sides meet at an angular edge below, which runs outwards and backwards. The sphenoid bone is convex between the bullae. Basioccipital lost. The palatonareal border is convex, and is opposite the middle of the second true molar. In the mature skull it would be probably more posterior. The palate is everywhere concave in transverse section. The frontoparietal suture is broadly convex, and is opposite the anterior edge of the glenoid surface, and 25 mm. in advance of the sagittal crest. The anterior processes of the bone on each side of the nasals are wide and truncate, and do not extend beyond the interior suture of the lachrymal Cope. J 562 [Jan. 18, bone. The latter is about as long on its superior sutures as it is deep at the orbit. It presents a distinct preorbital angle above a prominent tuber- cle. The occipito-parietal suture extends well forwards, 30mm. in advance of the crest. The squamosal does not reach to the lateral occipital crest. The infraorbital foramen occupies the position it has in the A. ryderanus. In a young specimen of guyotianus it has the same position as in the adult. The frontal foramina are about as far apart as each is from the supraorbital border. There is a postparietal foramen on the parieto squa- mosal suture, and there are three postsquamosals, two of them near together, and near the posterior suture, the other below the postparietal foramen. Foramen ovale oval, about as large as the F. rotundum, and separated from the foramen lacerum by the produced base of the inferior ala of the sphenoid bone. Palatine foramen opposite the third deciduous pre- molar. Superior canine teeth robust, bases of crown one-half lenticular, the pos- terior face truncate. A considerable diastema anterior to first premolar, and a short one behind it. Other teeth in continuous series. First and second premolars two-rooted ; absolutely simple. Third and fourth crown of first of usual form. First true molar smaller than second. Enamel minutely roughened. • Measurements. M. Length from supraoccipital crest to canine inclusive. . . .1^0 " " front of bulla 036 " " «c <. penultimate molar. .108 " orbit (axial) 105 " " front of orbit 126 " of sagittal crest 048 " " superior molars (last included) 076 *' " premolars 036 Diameters M.i^^^^^^-^P^^^^^^^^- transverse 014 anteroposterior 0165 transverse 020 Width of skull at postglenoid inclusive 077 " " '* middle of zygomas 050 " " fundus of canine alveolus 038 " between canines 020 " " second true molars 033 " of skull at postfrontal processes 076 " " " between anterior rims of orbits 066 Diameters M. ii The label from this specimen is lost. It is, however, from Oregon, and to judge from the color, from the true John Day epoch, rather than the North Fork bed. 1884.] 563 [Cope. A§:riocliceriis .guyotiaiiiis Cope. Hyopotamus guyotianus Cope, Proceedings American Philosophical So- ciety, 1878, p. 77. Merycopater guyotianus Cope, American Naturalist, 1879, p. 197, Proceeds. Amer. Philos. Soc, 1879, 375. Three crania, one with nearly entire mandible, and numerous fragments with mandibles, represent this species in my collection. It is the most abundant species of this genus in the John Day beds of Oregon. The cranium is of peculiar form. It is elongate from the orbits back- wards. The muzzle is elevated and compressed, so that the profile is hori- zontal, with subordinate irregularities. The occiput is therefore low as compared with the muzzle. The zygomata are rather slender, and are not expanded. The side of the muzzle is concave just below the superior border of the maxillary bone and above the fundus of the canine alveolus. The inferior part of the maxillary is concave from below the anterior border of the orbit to the line of the canine alveolus. The region above and anterior to the lachrymal bone is convex, leaving the flat nasal bones a little depressed. The frontal has a convex swelling on the middle line just posterior to the frontal foramina, from which point the surface slopes gradually and evenly to the supraorbital borders, and not in two planes, as in A. trifrons. At the front of the orbit the section of the frontal bone is convex at the sides and a little so at the middle. The supraorbital border is short and concave, not long and straight as in A. trifrons, and the postfrontal process is moderately prominent, and is not decurved. The anterior temporal ridges do not reach them. The former converge in nearly straight lines at an acute angle to a long sagittal crest. This in turn bifurcates into two very prominent posterior temporal crests, which over- hang the occipital condyles. The brain-case is an elongate-oval, and the olfactory portion is long and narrow, but not especially constricted at any one point. There is a prominent small tuberosity at the inferior part of the lachrymal bone ; above it the preorbital border is not defined as far as the beginning of the supraorbital. The postfrontal process originates be- low the anterior temporal surface which is continued along its posterior edge. The malar bone is concave on its external face. The zygoma is compressed, and has a long low superior convexity behind. Its crest con- tinues into a fine, low, posttemporal crest, which turns posteriorly above to its prominent posterior expansion above mentioned. The latter turn outwards at the apices, and send a low ridge downwards towards the occi- pital condyle. Below, the latter form a low angle on each side, which sep- arates a median from a lateral plane. Above, the occiput is deeply con- cave, and has a trace only of median keel. The basicranial axis is flat and rather wide between the otic bullae. The occipital condyles have distinct inferior boundaries which are sepa- rated by a flat interval. The posttympanic region is wide, and is bounded inferiorly by the deep styloid fossa. This is surrounded internally and pos- teriorly by the funnel-shaped base of the paroccipital process, which ex- tends first posteriorly as a longitudinal lamina, and then outwardly. Its Cope.] 564 [Jan. 18, edge terminates in a rough band which curves upwards and backwards to a point above the line of the occipital condyle. It is separated by a shal- low groove from the corresponding posttympanic ridge. The tympanic bone is not so long as in tlie species of Oreodontinse, and presents a tube- rosity externally. Like the paroccipital its base unites with the otic bulla. The bulla is small. Its base is extended towards the postglenoid process, but it is well separated from it, and does not reach the line of its anterior border. It presents a face anteriorly, and one inwards. The postglenoid process is narrow transversely, the depth and width being equal, and is elongate downwards. The coronoid process of the mandible is short, but has a base extended anteroposteriorly. The articular face of the condyle is convex anteropos- teriorly, and is extended downwards on the inner side behind. The hori- zontal ramus is slender, and has a straight inferior border. (The angle is broken away from this specimen.) The symphysis is oblique and nearly straight in profile. It is moderately elongate, and has the suture persist- ent. There is a tuberosity looking downwards from its posterior extremity, where it is rounded-compressed. The facial part of the lachrymal bone is longer than deep, and the lateral anterior part of the frontal is wide and obtuse, and extends anterior to the lachrymal. The nasals extend posteriorly to terminate in an acute angle which is above the anterior edge of the orbit. The frontoparietal suture extends across the space between the anterior temporal ridges at a point half way between the anterior border of the orbit and the anterior glenoid margin. The malomaxillary suture has no anteroinferior process. The mastoid forms a distinct mass between the exoccipital and squamosal. The sutures are largely coossified. The infraorbital foramen is above the contact of the third and fourth premolars. The space between the frontal foramina is about one-sixth the interorbital width. There is a large postparietal foramen on the parieto- squamosal suture, and there are two small postsquamosal foramina, in line above the posttympanic tuberosity. The mastoid foramen is small, and is not siAiated in a fossa of any extent, as is the case with the species of the Oreodontinse. There is a large foramen intermediate in position between that of the anterior condyloid and the jugular. Anterior and a little ex- ternal to it and slightly elevated between the confluent base of the paroc- cipital process and the otic bulla is another foramen, perhaps the jugular. Between the posterior base of the bulla and the basisphenoid, is a smaller foramen, probably the carotid. The other foramina are yet concealed by the matrix. The teeth do not differ in their form from those of other species of the genus. The second and third premolars have triangular bases, the second the narrower. The first has two roots. It is accidentally lost from one side, which circumstance led me to suppose at one time that this species has but three premolars above. The fourth premolar has its posterior ex- ternal V wdl developed, though a little smaller than the anterior. In the 188J.] 565 [Cope. specimen now described, the posterior internal rudimental cusp is quite well developed ; in the two other specimens now before me it is not so large. The superior canine is elongate, and not very robust, and its con- vex anterior border is directed partly posteriorly at the apex. The enameled portion of the crown is quite short. The premaxillary bones are narrow and weak, and are separated so as not to be in contact on the middle line. Its border displays two minute alveoli, from which teeth have been shed. I do not suppose that their presence is constant in the species. The external alveolus is twice the diameter of the internal. The inferior incisors are of normal number, but are very narrow, and much crowded. The canines arc very narrow, but are longer than the incisors. The first inferior premolar is more caniniform than in any other species of Oreodontidae known to me. The crown is a compressed oval in section, and is not expanded at its base. It is enameled to within 5 mm. of the alveolar border. A considerable diastema separates it from the second premolar. The description of the remaining teeth I take from a separate ramus of similar dimensions, as they are concealed in the type by their po- sition in juxtaposition with the cranium. The cusps of the true molars are pyramidal and acute, and entirely separate from each other. The external faces of the external are convex, their internal faces flat. The external laces of the internal are convex, the internal faces concave at the base, and convex near the apex. The anterointernal angle of the posteroexternal cusp extends to the base of the anterointernal cusp. The only difference between the first true molar and the fourth premolar, is that the anterior crest of the anteroexternal cusp is continued round to the front of the an- terointernal cusp, and to the internal side of the crown ; and the apices of the two anterior crests are separated by a shallow notch. The second in- ferior premolar has two roots. The heel of the third true molar is well de- veloped, and is convex posteriorly. Measurements. M. Length from occipital condyles to postglenoid process . . .038 " " " " " preglenoid border .. . .058 " " " " " postfrontal process. .104 " " " " " canine, inclusive 526 " " orbit to canine inclusive 085 " of mandibular ramus from condyle 176 " " symphysis mandibuli below 049' Width of occipital condyles inclusive 046 " '* occiput above 045 " " between otic bullse 016 " "at postglenoid processes inclusive 079 " "of skull above glenoid surfaces 100 " below orbits 099 " " " between orbits 068 " " " at fundus of canine alveoli 040 Cope.] 566 jjan. 18, Measurements. M. Width of skull between last upper molars, inclusive... .070 Depth of occiput to foramen magnum 043 " " " " basioccipital bone 058 " skull at last superior molar, exclusive 058 " " " " first premolar, exclusive 055 " " ramus mandibuli at front of M. iii 035 " " " " " " " P-m. iv 027 " " zygomatic arch above glenoid facet 027 Diameters of base of crown ( anteroposterior 009 of superior canine \ transverse 0085 Length of superior diastema 022 " " premolar series 039 " " true molar series 043 Width of premaxillary bones together 043 Diameters of base crown inferior P m. i ; anteropos- terior 0085 Length inferior diastema 015 " " last three premolars 031 " " true molars 046 " " last true molar 021 " " third premolar 009 fourth 012 Horizontal diameters of otic bulla \ anteroposterior. . . .019 t transverse 017 The specimens of this species which I have seen, are from the John Day river, and were obtained by Messrs. Sternberg, Wortman and Davis. The skull from which the above description was taken is the most perfect one of the genus Agriochoerus which has yet been found. Agriochoerus ryderanus Cope. Coloreodon ryderanus Cope, Bulletin U. S. Geological Survey Territo- ries, vi, p. 173. This species is represented in my collection by three nearly complete skulls without mandibular rami. While of the general size of the A. guyo- tianus, this species displays various well marked peculiarities. The most important of these are (1) the shape of the otic bulla, in which it differs from all other known Oreodontidae ; (2), the position of the infraorbital foramen, in which it resembles in this genus only the A. trifrons ; (3), in the form of the nasal bones posteriorly, in which it differs from the species where this part is known ; (4), in the form of the palatonareal border ; (5), in the form of the postglenoid process ; (6), in the outline of the sec- tion of the frontal bone. Agrioch(zrus ryderanus has the muzzle compressed laterally and flattened on top, as seen in the A. guyotianus and A. trifrons, and the side of the I884.J 567 [Cope. muzzle lias three distinct fossae. The largest of these is above the position of the fundus of the superior canine alveolus ; the second is below the fundus ; and the third is behind the position of the infraorbital foramen, and above the third and fourth premolars, and the first true molar. The lachrymal region is plane, and the nasals are flat. The frontal bone is nearly flat in section between the posterior borders of the orbits, but each is decurved to the lachrymal opposite the anterior border of the orbit. There is no indication of the three planes of the infraorbital region charac- teristic of the A. trifrons, nor of the median convexity of the A. guyotianus. The anterior temporal ridges commence about the middle of the width of each frontal bone, and unite after a shorter independent course than they have in A. guyotianus into a long, narrow saggital crest. This bifurcates posteriorly into two prominent lateral crests, which are directed down- wards and soon terminate, but which send forwards and downwards a delicate posttemporal crest. This passes without interruption into the supe- rior edge of the zygomatic arch. This arch is not expanded either upwards or laterally, and is rather weak. The external face of the malar bone is gently concave, and the inferior edge is rather wide, is truncate, and grooved along the middle. The occiput is deeply concave between the crests, and below them is gently convex. The superior edge of the fora- men is deeply notched at the middle, much as in M. guyotianus. The occipital condyles are large, and their inferoanterior angles are pro- duced horizontally for a short distance, forming short processes which are separated by a concavity of the basioccipital bone. The latter is plane be- low, but anteriorly develops a low meridian angle, which, widening on the sphenoid, causes its inferior face to be convex. The posttympanic element is distinguishable from the mastoid by a superficial groove, and a slightly free apex, and the mastoid from the paroccipital by a slight groove. The external base of the paroccipital extends but 5 mm. external to the line of the external border of the occipital condyles, and is therefore much less prominent than in the majority of species of Oreodontinae. The base of the paroccipital has a posterior and an anterior face, nearly at right angles with each other. The latter is continued into the pinched posterior promi- nence of the auditory bulla, and encloses on its external side, with the apex of the posttympanic, the deep stylohyoid fossa. The tympanic bone is represented by a tuberosity below the meatus, and a laminar expansion on the posterior face of the postglenoid process. The otic bulla's long axis is inwards, and a little posterior from the internal side of the postglenoid process, from which it is separated by a narrow interval. The bulla is con- stricted at right angles to its long axis, in two parts. The external part is subglobular with the side next the postglenoid process flattened. The internal part is roughened, displays a flat side posterointernally, and has an apical keel which extends posteriorly and a little externally into the base of the paroccipital processes. This form is not known in any other species of the family. The postglenoid process is more robust than in either M. guyotianus or M. trifrons. Its width and thickness are equal, Cope.] 568 [Jan. 18, and are a good deal longer than its height ; in the species named the height equals the other measurements. The pterygoid ala rises opposite the middle of the end of the glenoid surface, and the angle of its junction with the pyramidal process of the palatine is considerably in front of the middle of the trough of the posterior nares. Its edge posterior to this angle is shallowly grooved. The palatonareal border differs from that of any other species of the family known to me. It is acute in front, forming a Gothic arch, its apex being opposite the middle of the superior third true molar. In a young M. guyotianus, the only specimen of that species in which it is perfectly preserved, it is rounded, and extends to the posterior part of the second true molar. In an adult specimen, where the middle portion of the margin is lost, it extended at least as far forwards ; but its form is uncertain. The palate in the A. ryderanus is strongly concave throughout. The lachrymal bone has a different form from that of a A. guyotianus, more resembling that of A. latidens figured by Leidy. Its anterior superior angle is not produced, and its outline is a little deeper than long. The an- terior lateral prolongation of the frontal extends beyond it by nearly its width, and is wide, and terminates in an obtuse angle. The posterior edge of the nasals is broadly rounded, truncate at the middle, and is situated much in advance of the frontal foramina. The parietal is in contact with the alisphenoid. The squamosal does not extend beyond the vertical line from the base of the paroccipital process. The infraorbital foramen is above the anterior edge of the third superioi premolar, a position only seen elsewhere in the genus A. trifrons. The superior border of the orbit is concave and short as in A. guyotianus, and not straight and flat as in A. trifrons. The frontal foramina are above their middle, and their distance apart goes 4.5 times into the interorbital width. There is a large postparietal foramen on the parietosquamosal su- ture, and a large postsquamosal immediately below it. This arrangement differs from that seen in the other species here described, where there are two or three postsquamosals well posterior to the postparietal. Mastoid for- amen small. There are two palatine foramina on each side of the mouth, one opposite the posterior edge of the second premolar, and one opposite the posterior part of the fourth premolar. The anterior condyloid foramen is large. On one side is a small posterior condyloid, the only occurrence I have met with in the family. The foramen lacerum posterius is not divided into three foramina as in the A. guyotianus, but remains open as in the species of Eucrotaphus and Merycochoerus. It shows its nearer affinity to the first named species, however, in its triradiate outline ; and in the three grooves of the side of the bulla, which correspond to two of the three fora- mina. The /. lacerum anterius is not large, and is oblong in shape. The ovale is rather small, and is entirely bounded on the inner side by the pterygoid ala of the sphenoid. The /. rotundum is large and rather poste- rior. It is not bounded below by a transverse shoulder as is seen in the spe- cies of Merycochoerus, but is continued into a longitudinal groove, whose 188^.] 5G9 [Cope. external wall is longer than in any of the other genera of the family, ex- tending to a point half-way between the inferior edge of the foramen and the middle of the last superior true molar tooth. It is curved both inwards and downwards just posterior to the foramen. The superior molar teeth do not differ from those of M. guyotianus, M. antiquus and M. latifrons. The canines are very robust, and are separated from the first true molar by a considerable diastema. Measurements. M. Length from occipital condyles to line of postglenoid processes 046 Length from occipital condyles to line of preglenoid border 060 Length from occipital condyles to line of postfrontal process 110 Length from occipital condyles to line of canines, in- clusive 230 Length from orbit to canine inclusive 076 Width of occipital condyles inclusive 050 " " occiput at paroccipitals 031 " between otic bullae 015 " at postglenoid processes inclusive 093 " of skull above glenoid surfaces 110 " " " below orbits 094 " " " between orbits • 064 " " " at fundus of canine alveoli ; about 029 " " '* between last upper molars, inclusive... .080 " " palate at second true molars 033 " " ** " third premolars 031. " between superior canine 021. Depth of occiput to foramen magnum 038= " " " *• basioccipital 057 " " skull at last superior molar exclusive 050 " " " " first premolar, exclusive 046 " " zygomatic arch above glenoid facet 021 Diamelers base crown superior canine I ^^^^^^P^'^^''^^^- -014 I transverse 012 Length diastema to first premolar 0185 * ' premolar series 0365 " true molar series 045 Diameters P-m. iv (anteroposterior Oil transverse 014 Diameters m. i | ^^^^^^P^^^erior 014 <- transverse 015 Diameters M. iii / anteroposterior 018 ^ transverse 021 Cope.] 570 [Jan. 18, The skulls of this species came from the John Day bed of the John Day river, Oregon, and were found by Mr. J. L. Wortman. The species was established on an immature individual. The adults show that it belongs to this genus. COLOREODOX Cope. Proceedings American Philosophical Society, 1879, p. 375. Superior premolars three, the fourth with two external Vs, no facial va- cuities. The mandibles of the species of this genus are unknown, so that the character of the inferior dentition is unknown. The otic bullae are also destroyed in all the specimens, so that their character is unknown. In its reduced dental formula this genus represents one stage of that specialization which Owen has shown, has overtaken all the modern types of Mammalia. In this series this process seems to have stopped at this point, and not to have gone further, as the entire line has come to an end. The first superior premolar probably exists in a rudimental condition for -a short time, and is early shed. The same state of things has been found to exist as an abnormality on one side in the Agriochmrus guyotianus, and may be found again, but not so as to invalidate the characters of the genus Coloreodon. Two well-marked species of this genus have been described, which dif- fer as follows : .Smaller ; palatonareal border opposite posterior cusps of second true molar ; sagittal crest anterior, commencing opposite .optic foramina C. f&rox. Larger ; palatonareal border opposite posterior cusps of third true molar ; sagittal crest posterior, commencing opposite preglenoid border C. macrocephalus. Coloreodon ferox Cope. Fig. 1, p. 505. Proceedings American Philosophical Society, 1879, p. 375. The size of Oreodon culberisoni. Known from one skull from the North Pork of the John Day river, Oregon. C. H. Sternberg. Coloreodon macroceplialus Cope. Proceedings American Philosophical Society, 1879, p. 376. Size of the Eucrotaphus major. The typical skull is from the North Fork of the John Day riv^ar. A second skull, lacking all thej^rts posterior to the anterior origin of the sagittal crest, is undistinguwinable from the first. It was found at the " Cove " of the John Day river, Oregon. Both were obtained by Mr. J. L. Wortman. 1884.1 571 LCope GENERAL CONCLUSIONS. From what is now known of the history of the Oreodontidce, the following conclusions may be drawn. These are especially instructive as far as they go, since they involve the causes of the rise, great development, de- cadence and extinction of one of the best-marked types of Mammalia the world has seen. The history of this type involves more or less the history of the life of the North American continent during the Miocene epoch of Tertiary time. It moreover involves the laws which regulate the vital success of all types of life, and which express the causes of multiplication, of energy, of weakness, and of sterility. Two lines of the family, the Oreodontinm and the Agrioch(Brin(E, come to light simultaneously in geological time, the White River epoch, or the Oligocene. The latter is a higher type than the former in its more com- plex fourth premolars, while it is inferior in the non -closure of the orbits posteriorly. It may then be regarded as a parallel line. It has but two generic types, while the Oreodontinge present us with seven. So far as yet known, the Agriochoerinse did not continue as long as the Oreodontinse, as will be shown in tabular form below. In the progressive modifications of the Oreodontinse series, the first step was the inflation of the otic bulla (genus Eucrotaphus). This was suc- ceeded by the coossification of the premaxillary bones (genus Mer}-- cochoerus). These changes were accompanied by a regular increase in dimensions. The species of Merycochoerus are all of the largest size, and there are no small ones. The smallest species of Eucrotaphus are equal to the largest ones of Oreodon. The fourth genus Merychyus, while it loses none of the points already gained, shows a deficiency in its facial walls where vacuities appear. There is the greatest range of size here : with one species {M. major), as large as any of the Merycochoeri, we have another as large as the usual Eucrotaphi {M. eygomaticus), and several one degree smaller, or as large as the largest Oreodons. In the next genus the facial vacuities have attained to an enormous size. The premolar teeth become smaller, and the weakness of the narrow symphysis of the lower jaw is made up for by its coossification. The size is reduced from equal to the smallest Merychyi, to that of the smallest Oreodons (genus Leptau- chenia). In the next stage (genus Cyclopidius) the superior incisors dis- appear. Finally, the lower jaw is so reduced in front that it loses both incisors and premolars, in spite of its symphyseal coossification (genus Pithecistes). The species may be thus arranged in accordance with their distribution in time. White Biver Epoch. Oreodon gracilis ; O. affinis ; O. culbertsoni. Eu- crotaphus jacksoni ; E. major. Agriochoerus antiquus ; A. major ; A. latifrons. Jo7in Bay Epoch. Eucrotaphus jacksoni ; E. major. Merycochoerus Buperbus ; M. leidyi ; M. chelydra, sp. nov. ; M. macrostegus, sp. nov Cope. I 5»72 [March 7, Agriochoerus guyotianus ; A. trifrons, sp. nov. ; A. ryderanus. Coloreodon macrocephalus. North Fork of John Day River Epoch. Eucrotaphus trigonocephalus, sp. nov. ; E major. Coloreodon ferox ; C. macrocephalus. Ticholeptus Beds. Merycochoerus montanus, sp. nov. ; M. rusticus ; Mj. proprius. Merychyus arenarum, sp. nov. ; M, pariogonus, sp. nov. ; M. zygomaticus. Cyclopidius simus ; C . emydinus, sp. nov. Leptauchenia major ; L. decora ; L. nitida. Pithecistes brevifacies ; P. heterodon ; P. decedens, sp. nov. Loup Fork Beds. ? Merychyus elegans ; M. medius ; ? M. major.* The stratigraphic relations of these species may be represented under their generic heads in the following table : Oreodontincn. Oreodon Leidy Eucrotaphus Leidy.. MerycochcBrus Leidy Merychyus Leidy Leptauchenia Leidy . Cyclopidius Cope Pithecistes Cope AgriochmrincB. Agriochoerus Leidy.. Coloreodon Cope On the Structure of the Skull in the Elasmohranch genus Didymodus. By E. D. Cope. [Bead before the American Philosophical Society, March 7, I884.) The genus Diplodus was described by Agassiz from specimens of teeth from the European Coal Measures. In America, Newberry and Worthenf have described four species from the Carboniferous of Illinois and Ohio ; and I have reported two species from the Permian beds of Illinois and Texas. Recently Mr. Samuel Garman has described a shark, said to have been taken in the Japanese seas, under the name of Chlamydoselachus '0 CO 0 6 !^ 3 3 7 6 3 2 3 6 2 35 White River Epoch. John Day Epoch. ? North Fork Epoch. Ticholeptus Epoch. Loup Fork Epoch. 1 * The questions refer to the geological age. t Geology of Illinois, vol. li. Proc.Am.Philos. Sqc,N°I16.- 1884 T. Sinclair Son, Lit SKULL OF DIDVMODUS . 1S84.] 573 [Cope. anguineiis, whose teeth, as represented, do not differ generically from those of Diplodus. This is an interesting discovery, indicating that this genus, and not Ceratodus, is the oldest type of vertebrate now known in the liv- ing state. My collections from the Permian beds of Texas include not only numerous teeth, but jaws and crania. Among these I recognize two types of teeth, which I cannot distinguish from tliose of the D. compressus Newb., and D. gibhosus Agass. Whether these species belong to the same genus, is a question which I will discuss at the close of this article. I pro- visionally refer the D. compressus to a distinct genus, Didymodus, and will so call it in this article. The determination of the characters of this genus is a point of much interest The teeth resemble those of the existing sharks more than do those of any other genus of the Palaeozoic ages, but the antecedent im- probability of the modern type having existed at such an early period of the earth's history, is shown to be well founded by the present investiga- tion, which also throws much light on the question of the general phylo- geny of the fishes. I. Description. Twelve more or less complete crania of species of Didymodus are in my collection, and one set of jaws with small teeth and part of the cranium attached. One of the crania, unfortunately much broken, exhibits also some large teeth. All were found by the late Jacob Boll in the Permian beds of Texas. The skull of this species forms a continuum, which, however, displays distinct segments. First, however, as to the tissue of which it is composed. Both on the surface and in transverse fractures, it is more or less finely granular, the granules distinctly visible to the naked eye. These granules are composed of gypsum, as is also the matrix of a darker color in which they lie imbedded. Two hypotheses may be entertained regarding this structure. First, These granules may be regarded as the casts of coarse cartilage cells, and the matrix be in the place of the intercellular cartilage, replaced like the woody tissue in petrified wood. Second, The granules may be looked upon as replacements of osseous granules, such as cover the chondrocranium of most Elasmobranch fishes, while the matrix may be a replacement of the cartilage. The latter hypothesis is the more probable of the two, for two reasons : First, There is little probability of an unsupported chondrocranium retaining its form sufficiently long to per- mit the filling of its cells with a mineral deposit. Second, The granular type of ossification is well known in existing Elasmobranchs. It is only necessary to believe that the chondrocranium is penetrated by this kind of ossification. This state of things exists in the jaws also, which I de- scribe later. This structure has already been observed by Kner in the genus Pleuracanthus. The osseous cranium iS abbreviated anteriorly, and elongated posteriorly, CJope.] 574 [March 7, The orbit occupies part of the anterior third of the length. It is bounded in front by an obtuse preorbital process, and posteriorly by a laterally expanded and decurved postorbital process. The latter bears an articular facet on its posterior and inferior face. The top of the muzzle is exca- vated by a fontanelle which does not extend posterior to a line connecting the preorbital processes. There is a prominent cup-shaped occipital condyle. On each side of the cranium a short distance anterior to it, is a prominent process extend- ing outwards and a little backwards, which is excavated on its inferior side, but whose posterior side is decurved, so that the inferior concavity looks partially forwards. Into this cavity, and abutting against the decurved posterior edge, is a lateral process of the basal axial bone of the skull, which I take to be homologous with the lateral alse which occupy the same position in the sharks. Anterior to this junction no doubt the hyomandibular bone was suspended, for I suspect that it was articulated to a small condyle which is wedged into the fissure between the inferior and superior elements described, a centimeter anterior to their posterior extremities. This condyle is a distinct element of a subglobular form. The interorbital plane is continued posteriorly, bounded on each side by a depression which probably corresponds to the temporal fossa of higher vertebrates. The edges of this plane are thus well within the lateral borders of the cranium. The plane rises a little posteriorly, and is split into two narrow wedge-shaped processes, whicli project freely upwards and backwards. The rather short remaining part of the roof of the skull has a keel or sagittal crest on the middle line, which descends gradually to the foramen magnum. The base of the skull forms a continuum from the edge of the large occi- pital cotylus to the acuminate anterior extremity. The lateral basal alse are subcylindric, and are separated from the basicranial axis by a fissure for a short distance, and then unite with it. Two or three foramina ante- rior to this reunion, are in line with the defining fissure just mentioned. The basis cranii sends out a process on each side below the postorbital processes, giving a cross-shape to this part of the base of the skull. An- terior to this point it is free from other elements and contracts to an acuminate apex. The cranium is segmented, but a clean specimen is necessary to per- mit the straight sutures to be seen. In the first place, there is a distinct occipital bone, which includes exoccipital and basioccipital elements com- bined. The latter includes the large occipital cotylus, as in the Rhachi- tomous batrachian Trimerorhachis, and differs from the structure seen in the Lepidosirenid^e, where exoccipital elements only are present. The occipital extends but a short distance on the inferior face of the axis. It is preceded directly, and witliout imbrication, by a continuous axial ele- ment. If we recognize in the granular character of the tissue evidence of true ossification of the chondrocranium, we have here true continuous sphenoid and presphenoid bones. 188J.J 575 [Cope. Returning to the superior face of the cranium, we observe that the exoccipital elements form a wedge-shaped body, divided on the middle line by suture, with the apex forwards. Traces of this division are figured by Gegenbaur as present in Heptanchus.* Anterior to this the middle of the cranial roof is apparently occupied by another triangular bone with the base posterior and the apex anterior, and concealed beneath the free extremity of the element in front of it. The lateral sutures only are dis- tinguishable, appearing as grooves (fig. 2). This is the parietal bone. Ex- ternal to this and the occipital, and filling the space behind as well as an- terior to the postero lateral angle of the parietal, is the element which is produced outwards and backwards as already described. Were I describ- ing a true fish, this bone might be intercalare (epiotic) or pterotic. Perhaps it is both combined, or it may be the cartilage bone called by Giinlher, in Ceratodus, the "tympanic lamina." f The element anterior to the parietal is the cartilaginous representative of the frontal, and the fact that it terminates posteriorly in two free processes is significant of the true homology of the bones which terminate in like manner in the crania of the Lepidosirenidse.^ In this family and in the Ceratodontidse these bones are more or less separated on the middle line by the median pos- terior element. In Ceratodus the separation is wide ; in Lepidosiren the interval is uninterrupted, but narrow in front. In Protopterus these elements are in contact on the middle line, but diverge posteriorly. Bischoflf, Stanniusg and Giinther identify these elements with thefrontals in the genera they have described. Huxley || calls them supraorbitals, so that it becomes necessary to name the median posterior element a fronto- parietal, as a combination of two bones usually found distinct in fishes. The furcate structure of the frontal cartilage in Didymodus goes to show that the identification by Bischofi and Giinther is the correct one. There are also in this genus distinct paired membrane bones which do not take part in the bifurcation in question, and which appear to represent the frontals of Ceratodus. Each of these is a flat, subcrescentic supraorbital plate, which has a concave superciliary border. It is separated by a con- siderable interval from its fellow of the opposite side. Its anterior extremity is notched by a fossa which I suppose to represent the ante- terior (posterior in position) nostril. The ? frontal of the right side is dis- placed, and appears as a lamina lying on the frontal cartilage, showing that it is a membrane bone. From its relation to the nostril the question arises, whether it be not the homologue of the nasal. For hyomandibular bone, palatopterygoid arch, and mandibular arch, we have to rely principally on one specimen. On one of the skulls, two * Ueber den Bau des Schedels der Selachier, 1872, PI. I. / t Philosophical Transactions of the Royal Society, 1871, p. 511, indicated on the plates by the letter d. X Lepidosiren paradoza by Bischoff, Prof, in Heidelberg ; Leipsic, 1840. 2 Handbuch der Anatomic der Wirbeitbiere ; Rostock; Erstes Bach, die Fische. 1854, p. 49. U Anatomy of Vertebrated Animals, 1871, p. 145. * Cope.] 576 [March 7, curved rib-like bones lie parallel and divergent posteriorly on the right side of the frontal, in the temporal fossa. I cannot identify them. They are not present on the opposite side. As already described, there is a facet on the infero-posterior face of the postfrontal process. This in- dicates the point of articulation of the palatopterygoid arch, as it exists in the group Opistharthri of the sharks as defined by Gill, and as is clear- ly proven by the specimen now to be described. This includes the entire palatopterygoid and mandibular arches of one side, and the greater part of that of the opposite side, together with a considerable part of the right hyomandibular bone and probable ex- tremity of the ceratohyal. The anterior parts of both jaws support numerous small teeth, which closely resemble those described by Agassiz as belonging to his D. gibbosus. They differ from those of the JD. compres- sus in their smaller size. The palatine bones do not project much beyond the mandible, which, taken in connection with the form of the muzzle above described, renders it probable that the mouth was nearly terminal. In the palatopterygoid arch there is no noticeable separation or suture between the palatine and pterygoid elements. The inferior border of the palatine is swollen below the orbit ; its superior plate rises into a strong suborbital ala, which is concave externally, with thin superior edge. This edge rises posteriorly, giving the outline an elevated convexity, whose greatest upward prominence is above a point a little posterior to the middle of the jaw, and which probably articulated with the postorbital process of the cranium. Its surface gives indication of an articular sur- face appropriate to the corresponding one of the cranium. The superior border then descends rapidly to a vertical posterior border, which forms a somewhat prominent rim. This descends to the mandible, forming a regular ginglymus, the mandible bearing the cotylus. The mandible is rather robust ; its inferior edge is rather thin, and becomes incurved anteriorly. Its superior border is regular, except that it rises a little at the coronoid region, and is impressed, corresponding with a concavity of the surface, and arch of the border of the pterygoid region, just anterior to the posterior prominent ridge which forms its posterior edge. The hyomandibular bone is only exposed for its inferior half It issues from behind the palatopterygoid as a narrow shaft with obliquely truncate extremity. It is thus evident that the arrangement of the jaws is as in the two ex- ceptional existing genera, Hexanchus and Heptanchus. The external nostril already referred to, is a distinct, rather small fossa, on the lateral part of the superior face of the muzzle, near the extremity of the osseous portion. It is visible on both sides of the best-preserved specimen. It is continued forwards as a shallow groove. At the apex of the muzzle, is a fossa looking downwards, where roofed on each side by the ? nasal bones, which may represent the posterior nasal cavity. Or the latter may probably be represented by a lateral fossa just in front of the pre- orbital process. In either case it is evident that the nares are separated, 18S1.] 577 [Cope. and that the posterior one cannot be said to be within the oral cavity, as is the case in the known families of the Dipnoi. It is probable that there is a frontoparietal foramen at the posterior bifurcation of the frontal bones, corresponding to the conarium or pineal body of tlie brain. In a cranium broken across just anterior to the bifurcation^ a canal passing forwards and downwards is exposed. There is a foramen, or possibly only a deep fossa on each side of the middle line on the occipito sphenoid suture. The foramen magnum is rather small and opens upwards. Its border displays no articular surfaces. At the middle of a line connecting the posterior borders of the postorbital processes is a small shallow fossa, or probably foramen, from this there extends on each side backwards and outwards, a shallow groove apparently for a vessel, which terminates at the anterior one of three foramina already mentioned as in line with the fissure which distinguishes the lateral ala of the basicranial axis posteriorly. A similar groove connects the first and second of these foramina, and in one speci- men the groove from the median foramen joins this connecting groove. In front of the median foramen is a rather larger one on the median line, situated at the fundus of a short longitudinal groove. It is placed just posterior to a line connecting the preorbital processes. The grooves easily become obsolete by weathering. II. Affinities. In determining the systematic position of this animal, it will be con- venient to take a survey of the characters of the primary divisions of the fishes. In 1840 Bischoff published the first account of the osteology of Lepidosiren. In this description he called the frontal bones malars with a question, and the parietals frontoparietals. He described the skull as having an os quadratum. In 1854, Stannius in the Handbuch der Zoo • tomie* correctly determined the frontals and parietals, and stated further that the "lower jaw and hyoid bone articulate directly with continuous processes of the chondrocranium." This appears to be the first correct description of the cranial structure of the Dipnoi. In 1864, f Huxley re- stated the view of Stannius as to the nature of the mandibular articula- tion ; adopted the opinion of Bischoff that the frontal is a frontoparietal, and took a new position in calling the frontals supraorbitals. He also restates in general, the description of the skull of the Holocephali already given by Stannius. The system of Johannes Miiller, adopted by Stannius, was a great im- provement over preceding ones. It embraced, however, the error of in- cluding the Holocephali in the same sub-class (Elasmobranchi) with the sharks. This was adopted by Gill in 1861, ^ by Huxley in 1864 § and in 1871.11 All of these authors adopt at these dates the sub-class Ganoidea. *Erstes Buch, die Fische, p. 49. t Elements of Comparative Anatomy, p. 210. X Catalogue of the Fishes of the East Coast of North America, p. 24. § Elements of Comparative Anatomy. i The Anatomy of Vertebrated Animals, p. 120. Cope.J 578 [March 7, In 1871* the writer gave the following as the primary divisions of the sub- class Pisces : Holocephali, Selachi, Dipnoi, Crossopterygia, Actinopteri. The Holocephali was raised to an equivalency with the other sub-classes on account of the absence of distinct hyomandibular bone. The Dipnoi were defined by the median pelvic element, by the distichous arrangement of the segments of the pectoral and ventral fins, when present, on a me- dian axis, and by the supposed presence of a distinct hyomandibular bone. The latter definition must be abandoned, for though an ossification exists, it has been shown by Stannius, Huxley and Giinther, to be merely a de- posit in the continuous chondrocranium. The sub-class Crossopterygia was substituted for the sub-class Ganoidea of Agassiz and Miiller, as the latter was believed to have no actual existence as a division of fishes. After comparing the osteology of Polypterus, Lepidosteus and Amia, I remark (p. 320) "It is thus evident that the sub-class Ganoidea cannot be main- tained. It cannot be even regarded as an order, since I will show that Lepidosteus, Accipenser, and Amia, are all representatives of distinct orders. I hope, also, to make it evident that Polypterus should be elevated to the rank of a sub-class or division of equal rank with the rest of the fishes and with the Dipnoi, already adopted." The sub-class Ganoidea has not yet fallen into disuse, but there are strong symptoms that it will do so.f Among others I select the following extract from Huxley's paper on the ovaries of the smelt, published in 1883. ^: "As is well known, Lepidosteus presents an example of a Ganoid with oviducts like those of the higher Teleostei ; in Osmerus, on the other hand, we have a Teleostean with oviducts like those of the ordinary Ganoidei. It is tolerably obvious, therefore, that the characters of the female reproductive organs can lend no support to any attempt to draw a sharp line of demarkation between the Ganoids and the Teleos- teans. "Boas has recently conclusively shown that the same is true of the sup- posed distinctive character afibrded by the conus arteriosus ; and it has long been admitted that the spiral valve which has been described in the intestine of Chirocentrus is the homologue of that which exists in all the Ganoids, though greatly reduced in Lepidosteus. Indeed I am inclined to believe that the circular valve which separates the colon from the rectum in the smelt is merely a last remainder of the spiral valve. Thus, among the supposed absolute distinctions between the Ganoids and the Teleostei, only the peculiarities of the brain, and especially the so-called chiasma of the optic nerves, remain for consideration. My lamented friend Mr. Balfour, in the last of his many valuable labors, proved conclusively that the brain of Lepidosteus is, both in structure and development, a Teleostean * Proceedings Amer. Assoc. Adv. Science, p. 326. Transac. Amer. Philosoph. Soc, p. 449. fThe term ganoid can be used as an adjective to describe the scales already known by that name, and thus be preserved. X Proceedings Zoological Society of London, 1883, pp. 137, 138, 139. 18S-1.J 579 [Cope. brain. But it is singular that no one, so far as I know, has insisted upon the fact, not only that the Telcostean brain is essentially similar to that of the Ganoids, but that it is exactly in those respects in which the Ganoids and Teleostei agree in cerebral structure that they difler most markedly from the Plagiostomi and ChimtDroidei. " With respect to the chiasma of the optic nerves, the exact nature of that structure has not yet been properl)'^ elucidated either in the Selachians or in the Ganoids. But, whatever may come of such an investigation, the establishment of the existence of a true chiasma in the Ganoids, and of its absence in Teleosteans, can have but little bearing on the question of their affinities, since Wiederslieim has shown that a simple decussation of the fibres of the optic nerves, as in ordinary Teleosteans, takes place in many lizards." In 1877* I proposed the following primary divisions of the fishes, and have seen no reason to alter my views as to their value as a correct ex- pression of the affinities and diversities of this class of Vertebrata. The system difiers only from that of 1871 in the consolidation of the Crossop- terygia and Actinopteri into a single sub-class, the Hyopomata ; and in a few corrections of the definitions given. They are as follows : I. Suspensorium continuous with the cartilaginous cranium, with no hyomandibular. No rudimental opercular bone ; no maxillary arch ; pelvic bones present ; axial series of fore limb shortened, the deriva- tive radii sessile on the basal pieces ; axial series of hinder limb pro- longed in male Eolocephali. II. Suspensorium articulated with the cranium ; no maxillary arch ; no opercular nor pelvic bones ; bones of limbs as in the last ElasmohrancJii. III. Suspensorium rudimental, continuous with cranium, supporting one or more opercular bones ; cranium with superior membrane bones ; no maxillary arch ; a median pelvic element ; the limbs supported by segmented unmodified axes Dipnoi. lY. Hyomandibular and palatoquadrate bones articulated with cranium, supporting opercular bones ; a maxillar}'- arch ; no pelvic element ; axes of the limbs shortened, the derivative radii sessile on the basal pieces Hyopomata. In the definition of the Dipnoi, it is necessary to make the correction in accordance with the best observations on fresh specimens, above referred to, as I have not been able to determine the question from dried speci- mens in the Hyrtl collection. The suspensorium cannot be properly said to be articulated to the cranium in the sense in which it is said to be such in the Elasmobranchi. In the latter it is articulated by ginglymus ; in * Proceedings of the American Philosopiiical Society, 1877, p. 25; and in tlie Annual Reports of the Commissioners of Fisheries of Pennsylvania for 1879-80, p. 67 and 1881-2, p. 111. Cope.] 580 [March 7, the Dipnoi merely by suture or contact, with other cartilage bones. Its character is therefore more nearly that of the Holocephall than of the Elasmobranchi or the Hyopomata. In the light of the above considerations, to which sub-class must be re- ferred the genus Didymodus ? Does it possess a freely articulating hyo- mandibular bone, and maxillary, palatoquadrate and mandibular arches? The question must be primarily determined by these considerations, since the fins and their supports are unknown to us. The lateral posterior processes of the skull are in its superior plane, and their extremities do not present an articular facet' for the lower jaw. It is improbable that they were continued downwards as cartilage for the former articulation, as in the Holocephali and Dipnoi. Both from the presence of an articular condyle, and from the mechanical necessities of the case, I have little doubt but that there was a freely articulating hyo- mandibular bone. I have already described this element in fact as visible in a single specimen. The choice is thus limited to the Elasmobranchi and Hyopomata. It is decided in favor of the former by the absence of maxillary arch and of opercular apparatus. So then Didymodus is a shark, in spite of its peculiarities. Kner* speaks of the presence in the nearly allied Pleuracanthus {= Diplodus), of premaxillary and maxillary bones ; but this is no doubt a misinterpretation of the homologies, as he says they articulate with the lower jaio. In my jaws there is but one bone on each side, a palatopterygoid. In his researches on the structure of the skulls of sharks, Gegenbaurf shows the difierent methods of articulation of the palatopterygoid arch in the sub-class Elasmobranchi. In Heterodontus the palatopterygoid arch is attached to the skull throughout by its superior border, anterior to the orbit, but is free posterior to the orbit. In Hexanchus and Heptanchus it is free anteriorly, but articulates by its elevated posterior portion with the postorbital process. In the remainder of known recent Elasmobranchs it is free throughout, and merely in contact in front. These relations are also described by Huxley. X Brofessor Gill utilizes them as definitions of three (of four) primary divisions of the sub-class Elasmobranchi, § which he names the Opistharthri, (fam. Hexanchidae) ; Proarthri (Heterodon- tidge) ; Anarthi (sharks proper) ; and Rhinse (Squatinas). According to these definitions, Didymodus must be referred to the Opistharthri. The skull, however, presents other characters which must claim attention. Its *Sitzungsberichte Wiener Akademie, LV, p. 540. fUntersuchungen zur Anatomic der Wirbelthiere, Leipzic, 1872. J On the Anatomy of Ceratodus. Proceedings Zool. Society of London, 1876, p. 43-4, with figures. 2 Bulletin of the U. S. National Museum, No. 16, 1883, p. 967. Gills fourth group, Rhinse, does not appear to me to possess the value of the other three, nor are the " Raise" and "Pristes " more distinct. I therefore propose that the order Selachii, as defined in the following pages (of the sub-class Elasmobranchi), be divided into three sub-orders: Opistharthri, Proarthri and Anarthri, the lat- ter to include the true sharks, the Squatinee, the sawfishes and the rays. 1884.] ^Sl [Cope. reference to the Elasmobranchi is confirmed by the following characters : (1 ) The nares are not oral. (2) There is a large fontanelle on the summit of the muzzle. (3) There are processes corresponding to the lateral alae of the basicranial axis. In another character Didymodus differs from this and all other sub-classes of the Pisces. This is the penetration of the granular ossification through- out the chondrocranium. In the following characters it agrees with the Dipnoi : (1) The distinct exoccipital, parietal, and frontal elements. (3) The occipital cotylus. (3) The posterior bifurcation of the frontal cartilage. ^ In the following characters Didymodus resembles the Hyopomatous or , true fishes : (1) In the basioccipital bone with condyle. (2) In the ?os intercalare or pteroticum. (,3) The presence of a distinct element articu- lating with the proximal end of the hyomandibular. (4) The presence of membrane bones in the position of frontals. The characters above cited as constituting resemblances to the true fishes, will not, it appears to me, permit the reference of this genus to any of the divisions of sharks established by Prof. Gill. I therefore proposed a new order of the Elasmobranchi* for its reception, with the following name and definition. A basioccipital bone and condyle. Occipital, ? pterotic, and frontal bones distinct. Supraorbital (or nasal) bones present Ichthyotomi. The remaining Elasmobranchi, in which the above characters are want- ing, may be termed by way of contrast, utilizing an old name, Selachii. Were it not for the probable presence of the free hyomandibular bone, the order Ichthyotomi might be regarded, in the absence of knowledge of its limbs, as the possible ancestor of the Rhachitomous Batrachia. But as the Batrachia have no distinct suspensorium, or are, to use Miiller's con- venient term, monimostylic, their origin must still be sought for in some yet undiscovered type of Dipnoi. It is on the other hand very probable that the Ichthyotomi are the group from which the Hyopomata derived their origin. The distinct basioccipital with its two foramina, the superior origin of the hyomandibular, and the superior nostrils, all point towards the true fishes. The tribe of Hyopomata which must be their most im- mediate descendents, are the Crossopterygia, as I define that division. I must now compare the Ichthyotomi with such groups of the Hyopo- mata as they may be supposed to approach most closely. I begin by refer- ring to the marine eels of the order Colocephali. In 1871f I characterized this order as follows: "Parietals largely in contact; opercular bones rudimental ; the preoperculum generally wanting. Pterygoids rudimental or wanting ; ethmoid yery wide. Symplectic, maxillary, basal branchi- hyals, superior and inferior pharyngeal bones, all wanting, except the fourth pharyngeal. This is jaw-like, and is supported by a strong supe- rior branchihyal ; other superior branchihyals wanting or cartilaginous." * American Naturalist, 1884, 413. t Proceedings American Ass. Adv. Science, xx, pp. 328-334. Cope.l 582 [March 7, The statement "maxillary wanting," is in contradiction to the definition of the sub-class Hyopomata, which asserts the presence of those bones. Stannius* has asserted the absence of the "oberkiefer" in the eel; Giintherf describes their presence. As the absence of the maxillary bone would constitute a point of resemblance, if not aflBnity to the Elasmo- branchi, I have reexamined my material to determine the homologies oif the lateral dentigerous bone of the upper jaw of the eels. My specimens of species of the Colocephali include the following from theHyrtl collec- tion : Myrus 'vulgaris ; SphagehrancJius rostratus ; Moringua rcetaborua ; , Mnroina sp.; Murmna unicolor ; Murmna sp.; PoecilopMs polyzonus, and GymnomurcEna tigrina. The pterygoid bone exists in a rudimental condition in the OymnomurcBna tigrina, Myrus vulgaris, and one of the species of Mursena ; and whether lost in the preparation of the other crania or not, cannot be stated. In the Anguilla vulgaris the pterygoid bone is con- siderably larger, and extends to a point halfway between its base and the extremity of the muzzle. In the Conger vulgaris it extends still further forwards, reaching a transverse process of the anterior part of the vomer. No palatine bone appears. The premaxillary bone is not distinguished from the ethmoid in the Colocephali, nor in the Enchelycephali (Anguil- lidse, etc.). It is quite possible, therefore, that the external dentigerous bone or upper jaw, in both of these orders, may be the palatine, and the maxillary be wanting. The family of the Mormyridse appears to furnish the solution. In this group the structure and connections of the pterygoid bone are much as in Conger, and there are in addition distinct premaxillary and maxillary bones. It is clear that in this family it is the palatine, and not the maxillary bone, that is wanting. Similar evidence is furnished by the family Monopteridse. The definition of all four of the orders, Colocephali, Enchelycephali, Ichthyocephali and Scyphophori must, therefore, embrace this character. The Gymnarchidse agrees with the Mormyridse in this respect, and both families have the transverse process of the vomer which receives the pterygoid, as in the genus Conger.:}: The supposed resemblance to the sharks presented by the Colocephali is then not real, and the question as to the point of aflfinity of the Ichthyotomi to the true fishes remains open as before. I now refer to the remarkable characters presented by the deep sea fishes of the family Eurypharyngidae, as recently published by Messrs. Gill and Ryder. § These authors find the characters of the skeleton so remarkable, that they think it necessary to establish a new order for its reception, which they call the Lyomeri. The definition which they give is the fol- lowing : "Fishes with five branchial arches (none modified as branchi- ostegal or pharyngeal) far behind the skull ; an imperfectly ossified skull articulating with the first vertebra by a ba'sioccipital condyle alone ; only * Handbuch der Zootomie, Fische 1854, p. 76. t Catalogue Fishes, British Museum, vol. viii, p. 19. X These transverse processes are enormously developed in Gymnarchus. § Proceedings U. S. National Museum, Nov. 1883, p, 262. 1884.1 583 [Cope. two cephalic arches, both freely movable ; (1) an anterior dentigerous one — the palatine, and (2) the suspeusorial, consisting of the hyomandibular and quadrate bones ; without maxillary bones or distinct posterior bony elements to the mandible ; with an imperfect scapular arch remote from the skull ; and with separately ossified but imperfect vertebraj." M. Vaillant came to no conclusion as to the affinities of this group ; and Messrs. Gill and Ryder remark, "We are unable to appreciate any affinity of Gastrostomus to any Anacanthines, Pliysostomes, or typical Apods, nor does it seem to be at all related to Malacosteus, which has been universally considered to be a little modified Stomiatid." It is, how- ever, clear to me that the relationships of this family Eurypharyngidae are to the order Colocephali, and that they represent the extreme de- gree of the modification of structure which that order exhibits. In other words, the modification of the ordinary piscine type which is lound in the Anguillida3 (order Enchelycephali), is carried to a higher degree in the Colocephali, and reaches its extreme in the Eury- pharyngidae. The points of identity between the two groups last-named are so many, that it becomes desirable to ascertain whether they are susceptible of ordinal separation from each other. The characters above given to the order Lyomeri are in fact identical with those which define the order Colocephali, with a few possible exceptions. First, how- ever, I note that the supposed palatine arch, is probably the maxillary, as in the Colocephali, and that it is the palatopterygoid arch which is absent. The five branchial arches exist in the Colocephali, but the three anterior are rudimental, and the basal branchihyal bones of the fourth and fifth are closely united. There are, however, five arches. There is a ceratohyal arch in Muraena and Gymnomursena, but of very slender pro- portions. Whether this element is absolutely wanting in Gastrostomus, or whether the first branchial arch is its homologue, remains to be ascer- tained. Should the last two be coherent as in the Colocephali, we would then have the same number of hyoid arches in both, viz., six. The "im- perfectly ossified cranium " is shown in the detailed description given by Messrs. Gill and Ryder, to support the same bones which are found in the Mureenoid skull. The degree of ossification of the skeleton does not con- stitute a basis for ordinal distinction, if the same elements be present. For this reason the perforation of the vertebral centra by the remnant of the chordadorsalis does not seem to be of ordinal importance. In the more detailed description, there are a few charecters worthy of notice. First, "The notochord is persistent in the skull for half the length of the basioccipital." This indicates further the primitive condi- tion of the vertebral column, but scarcely gives basis for an ordinal defi- nition. Second (p. 266.), "The neurapophyses are slender, diverging (instead of convergent), cartilaginous distally, and embracing the neural sheaths on the sides, while by the neurapophyses is supported a membra- nous sheath which roofs over the nervous cord," etc. The nerual canal is well closed above in the Munenidtie, but in the Anguillidae it is largely Cope.] 584' [March 7, open above. The neurapophyses it is true unite, but at a distance above the neural cord, and as attenuated rods. Third, "There is no vomer de- veloped, but a triangular cartilaginous element pendent from the cranial rostrum affords attachment lor the palatine (read maxillary) element anteriorly," etc. This element probably exists in the Colocephali and similarly takes the place of the vomer, only differing in being ossified. I have been accustomed to regard it as the homologue of the bone called ethmoid in fishes. The character which distinguishes the Colocephali from the Encliely- cephali, now that their maxillary and palatine structure are shown to be essentially the same, is found in the hyoid apparatus. In the Enchely- cephali, the structure is as in ordinary fishes ; there is a glossohyal, and there are basihyals, and axial branchihyals, and superior pharyngeals. In the Colocephali all these elements are wanting, excepting the fourth supe- rior pharyngeal, which has the form of an antero posteriorly placed den- tigerous jaw, which opposes the lateral branchihyal of the fifth arch or, as it is generally called, the inferior pharyngeal. It is evident that the Eurypharyngidse are more similar to the Colocephali than to any other order in this respect also, but the description of these parts is not yet suffi- ciently detailed to enable me to determine what difference there may be in this respect, if any. The mobility of the quadrate bone on the liyo- mandibular cannot be regarded as of great systematic significance, although it is doubtless important in the economy of the fish. It is then evident that the Eurypharyngidte belong very near to, if not within, the order (Jolocephali. Towards the end of their description, Messrs. Gill and Ryder (p. 270), recognize this relationship, but deny that it indicates that this family is "from the same primitive stock as the Muraenids." I incline to the belief that it is the ultimate result of the line of development of which the Anguillidae form one of the first terms, and the Murcenidaj a later and more specialized one. It is therefore clear that the point of relationship oi the Ichthyotomi to the true fishes is not to be found in the Eurypharyngida3 or the Colo- cephali. In the following point Didymodus resembles Polypterus. The fossa above described asou each side of the basioccipital, is found in Polypterus. There it serves as a place ol insertion of a strong ligament on each side, which is attached externally to the epiclavicle, and serves to hold the scapular arch in its place. A similar structure exists in the Siluirda?, where the ligaments are ossified. It suggests for DidymOdus a scapular arch suspended more anteriorly than in sharks, possibly even to the skull. The genealogy of the fishes will then be as follows, first, however, it is to be understood that in asserting the derivations of one group from another, I mean that in accordance with the rule which I have termed "the doctrine of the unspecialized," the later type in each case is the descendant of the primitive and not the later sub-form of its predecessor. In this way is to be explained the apparent anomaly of regarding the 1S84.1 ■585 [Cope. notocliordal sturgeons as descendants of Crossoptery{»ia, whose modern representatives are osseous. The primitive Crossoptcrygia, and probably even the Actinopteri, were doubtless as cartilaginous as arc the existing sturgeons ; In this phylogen}'-, the Holocepliali, which have not differentiated a suspensorium. are regarded as the primitive fisheb, although the living representatives display some specialized character?, as, for instance, a membranous gill-cover which conceals the primitive slits. The line to the right continues the monimostylic character and passes into the reptiles, whose primitive types are also monimostylic, as Johannes Miiller called Ihem. In the later forms or streptostylicate reptiles of Miiller (Lacertilia, Ophidia), the quadrate becomes freely articulated.* In tlie left hand series, the Elasmobranchs immediately present us with the free suspensorium or hyomandibular, which is a well-known character of the remainder of the line, the modifications being the addition of sepa- rate elements, as the metapterygoid, "quadrate," and symplectic. The penetration of ossification into the chondrocranium of Didymodus, in regions not ossified in either fishes or batrachia (sphenoid and pre- sphenoid), and into regions not ossified in any vertebrate (frontal and parietal cartilages), may be, so to speak, only a local phenomenon, and not indicative of extensive phylogenelic consequences. For if it be so regarded, it evidently proves too much, giving affinities in the base of the skull to the reptiles, and in the roof exhibiting a character more highly developed than any known form of vertebrata. The Ichthyotomi include, so far as yet known, but one family, the Hybo- donlida3 of Agassiz. According to that author this family includes four genera, Hybodus. Pleuracanthus, Cladodus and Sphenonchus. It ranges from the coal-measures to the Jura inclusive. The genus Didymodus may be described as follows : Frontal plane well defined on each side by the temporal fossae, and ter- minating in two cornua posteriorly. Anterior nares on the superior sur- face of the muzzle. Supraorbital (or nasal) bones well separated on the median line and constituting the only membrane ossification Teeth with large lateral denticles. The species Didymodus compressus Newberry, may be defined as follows : Skull with massive walls. Form elongate, depressed, the orbit not ex- Actinopteri. Batrachia. / Dipnoi. / Holocephali. / * The phytogeny of the Reptilian series can be found in the Proceedings American Association Advancement of Science, xix, 1871, p. 213. The Batrachia are supposed to be their ancestors. Cope.l 586 [March 1, tending behind the anterior third of the length. Basicranial and basifacial axes in one line, flattened, the supraorbital border llat, concave on the edge ; postorbital processes obtuse, the temporal ridges commencing with thin posterior border, which they excavate. The ridges then turn, ex- tend parallel posteriorly, terminating in the horn-like processes already described, wath a slight divergence. The apices mark the posterior third of the length of the skull. The occipital condyle is wider than deep, and its superior border retreats forwards so as to cause its cup to look upwards. The exoccipital diameter at the foramen magnum is less than tliat of the basicranial axis, the osseous element of which, probably sphenoid, is re- curved on the sides to their middle. The sides of the latter expand a little to meet their lateral alai. Immediately above their contact is situ- ated the supposed condyle for the hyomandibular element. The basicranial axis is convex opposite the postorbital processes, from the bases of which a concavity separates it. It has a slight median groove at this point. It is much narrower than the interorbital width above. A short distance in front of the postorbital processes it begins to contract, and gradually reaches an acilminate apex. Superior to this apex, commencing posterior to it, the space between it and the supraorbital or nasal elemeats is occu- pied by a massive element (? etlimoid) which forms the floor of the nasal median fontanelle. The surfaces are smooth, but readily weather so as to be granular. The granules are subround, with flattened surface. Measurements of skull. M. Total length of skull to end of frontal bone (No. 1) 180 " " muzzle to orbit ; axial 024 " " " skull to postorbital process 058 " " " to apices of frontal cartilage 117 " " " " to ? pterotic apex (axial) 155 Width of skull at prefontals 045 " " " " supraorbital borders 055 " " " "? pterotic apices 088 " " occipital condyle 034 Depth " " 035 Jleasurements of jaws. Length of mandibular ramus from cotylus, inclusive. .145 Depth " mandibular ramus at cotylus 028 " middle 035 Length " palatopterygoid bone from cotylus, inclusive. .145 Depth " " "at postorbital articula- tion.: 071 Depth of palatopterygoid bone at orbit. 035 Length" ** " posterior to orbit 070 A second species has been brought to light by the researches of Mr. W. im.] 587 fCope. F. Cummins In the Permian beds of Texas. Parts of the jaws with two of its teetli are preserved. The lower jaw is distinguislied from that of the B. compressus by its small transverse as compared with its other di- ameters. The ramus is quite compressed, and is not thicker at the inferior edge than the superior, and is slightly concave on the inner side. Its ex- ternal face is nearly vertical. The angle is rounded forwards, and there is no angle behind the cotylus, which is raised above the superior line of the ramus. The cotylus is rather large, and has a shallow anterior supe- rior, and a posterior subposterior facet. There is no indication of a coro- noid process. The inferior edge of the ramus is swollen on the outer side, below the anterior border of the condyle, so as to mark with the thickened posterior edge of the ramus a fossa in the position of the mas- seteric. The teeth are pecular in the form of the root (Figs. 8-9). This part has no anterior projection, and the posterior portion is a flat, thin edged plate, wider than long. It carries a button, but no notch. There is a minute median denticle. The form of the root is thus very different from that of the tooth of the D. compressus (figs. 5, 7). Measurements. M. Depth of ramus at cotylus (vertical) 0G2 " " " 120 ram. anterior to cotylus. .048 Transverse diameter at the same point 009 Long diameter (oblique) of cotylus 031 ^ (anteroposterior Oil Diaireters of base of tooth | t,^„g^erse 037 Diameters of crown of lateral denticle anteroposterior .0048 transverse uOC I call this species Bidymodus platypternus. Should the name Didymodus be found hereafter to apply to species of Pleuracanthus, the latter generic name must be used for this species. nr. Historical. In 1837 Prof. Agassiz (Poiss. foss., iii, 66), described a spine which he believed to have belonged to a fish like the sting-rays, as Pleuracanthus Icevissimvs. The only example was obtained from the Dudley Coal field. In 1845 Prof. Agassiz (Poiss. foss., iii, 204), made known certain teeth, which he referred to sharks of the family of Hybodonts. Two spe- cies were distinguished, D. gibbossus and D. minutus. Both were obtained from the English Coal measures. In 1848 Prof. Beyrich (Berichte vernandl. k. Preuss. Akad. wiss. , 1848), proposed the generic name Xenacanthus for a German Carbonifer- ous form, referred to Orthacanlhus by Goldfuss (1847), but which ap- proached nearer to Pleuracanthus. In 1849 Dr. Jordan (.Tahrbuch fiir Min. u. Geol.. p. 843), described, under the name Triodus sessilis, a form subsequently ascertained to be identical with the Xenacanthus. Cope.] 588 f March 7, In 1857 Sir Philip de Malpas Gray Egerlon (Ann. and Mag. Nat. Hist., XX, 423), contended tliat tlie spines of Pleuracantlius belonged to the same fish as the Diplodus teeth, and that Xcnacanthus was likewise refer- able to the same type. In 1867 Prof. Kner (Sitzb. k. Akad. wiss. Wien, Iv, 540-584), published a memoir, illustrated by ten plates, in which he proved that Diplodus and Xcnacanthus were generically identical. In 1875 Messrs. St. John and Worthen proposed the genus Thrinacodus for the Diplodus incurvus and D. dapUcatus of Newberry and Worthen and the 7\ nanus St. J. and W., from Illinois. In 1883, in the Proceedings of the Philadelphia Academy (p. 108), I proposed the name Didymodus for the Diplodus compressus Newberry. In Science for 1884, p. 274 (March 7th), I called attention to the close re- semblance of the teeth of this genus to those of the recent shark, called by Garman Chlamydoselachus, and expressed my belief in the identity of the two genera. In the American Naturalist for April, 1884, p. 413, I gave a brief ab- stract of the characters of the skull of Didymodus, and proposed to regard it as the type of a new order to be callcvT the Ichthyotomi. In Science, 1884, p. 429 (April 11), Prof. Gill objects to the identification of the genera Didymodus and Chlamydoselachus ; on the ground of the dif- ferent forms of the teeth. He states that he doubts the pertinence of the two genera to the same Drder. He points out that the oldest name for Dip- lodus Ag. is Pleuracantlius Ag., and that the order Ichthyotomi had been already defined and named by Liitken, wiUi the name Xenacanthini. On these various propositions the following remarks may be made. (1.) There is no generic difference to be detected, in my opinion, be- tween the teeth which are typical of Diplodus Agass. and Thrinacodus St. J. and W. and the recent Chlamydoselachus. Difierences there are, but apparently not of generic value. The identification of the recent and ex- tinct genera rests, as far as this point goes, on the same basis as that of the recent and extinct Ceratodus. (2.) At the time of my proposal of the name Didymodus, I was not con- vinced that fishes of this type bore the spines referred to the genus Pleura- cantlius Ag. None of the authors cited figure any specimens which pre- sent both tricuspidate teeth and a nuchal spine. None of my ten speci- mens possess a spine. However, Kner describes two specimens as exhibit- ing both tricuspidate teeth and a spine, and Sir P. Egerton's statements {I. c), on this point are positive. So we must regard Pleuracanthus as the name of this genus, Avith Diplodus as a synonym. (?.) Diplodus being regarded as a synomym of Pleuracanthus, it follows that Chlamydoselachus Garm is distinct, on account of the different struc- ture of the dorsal fin, which is single and elongate in Pleuracanthus, ac- cording to Geinitz and Kner. The presence of the nuchal spine in Pleura- canthus is also probably a character of distinction, although we do not yet know whether such a spine is concealed in Chl:imydoselachus or not. 1884.] 589 [Cope. (4.) The identity of Didymodus (type Diplodus compressus Newberry) and Pleuracantlius, may now be questioned. None of the specimens are figured and described by the authors above cited, as displaying an occi pital condyle, or posterior frontal cornua. My specimens of Didymodus compressus do not exhibit teeth on the roof of the mouth, as Kner describes. There are no spines with the crania, although separate Pleuracantlius spines are not rare in the same beds. The teeth associated with the skulls, moreover, present a button on the superior side of the root ( Fig. 5). Agassiz figures teeth of this kind as belonging to the Diplodiis (jibbosus. St. John and Worthen make these teeth typical of Diplodus, and confer the name Thrinacodus on those without the button, a character which I do not think a constant one. The latter nanie is then probably a synonym of Pleuracan- tlius. The button-bearing teeth are figured and described by Kner as occur- ring scattered, and in a somewhat different horizon from that of the Pleu- racantlius specimens. In Germany, as in Texas, the button-bearing teeth are the larger. I suspect that the skull I have described represents a different genus from Pleuracanthus proper. This genus will not differ from Chla- mydoselachus Garm., in the lack of other evidence ; the teeth presenting only specific difference. (5.) Of course, a study of the anatomy of Chlamydoselachus, which I hope Mr. Garman will soon give us, may reveal differences between that genus and Didymodus ; but of these we know nothing as yet. (G.) The order Xenacautliini was proposed by Geinitz (Dyas) for Pleu- racanthus, on account of the supposed suctorial character of the ventral fins. This character is supposed by Kner to be sexual. In any case this division, whatever its value, must be subordinated to the order Ichthyo- tomi, as I define it. EXPLANATION OF PLATE. All the figures two-thirds natural size, except fig. G, which is one- half larger than nature. Fig. 1. Skull from above, right frontal bone displaced, and its anterior extremity broken off. Posterior apex broken from right frontal cartilage bone, a. Frontal or supraorbital bone, that of the right side displaced ; b, anterior nostril ; c. postfrontal facet for palatopterygoid ; d, frontal fissure. Fig. 3. Posterior'partof skull of another individual, from above ; a, occi- pital bone ; b, parietal ; c, a cornua of frontal bone. Fig. 3. Anterior view of fig. 2, displaying section of brain case ; a, frontal or parietal cartilage bone ; b, sphenoid ; c, brain cavity ; d, frontoparietal fontanelle ; e, hyomandibular condyle (? pterotic bone). Fig. 4. Anterior part of skull from below, of a third individual, display- ing orbits and postorbital processes. Fig 5. Tooih of Didymodus compressus Newb. , natural size, posterior view. Cope.] 590 [March 7, 1884. Fig. 6. Palatopterygoid and mandibular arches of a fourth individual from riglit side, with /m, hyomandibular. Fig. 7. Superior tooth of external row, without apices of two of the cusps ; from the palatine bone of the specimen represented in fig. 5 ; one- half larger than nature, anterior view. Fig. 8. Tooth of Bidymodm platypternus Cope, nat. size, from above posteriorly. Fig. 9. Tooth of a second specimen of Didymodus platypternus from below. Printed July 1, 1884. Paleontologieal Bulletin, No. 39. THE EXTINCT MAMMALIA OF THE VALLEY OF MEXICO. Read before the American Philosophical Society, May i6, 1884. ON THE STRUCTURE OF THE FEET IN THE EXTINCT ARTIO- DACTYLA OF NORTH AMERICA. Read before the American Philosophical Society, Aug. 15, 1884. FIFTH CONTRIBUTION TO THE Knowledge of tlie Fauna of tlie Permian Formation OF TEXAS AND THE INDIAN TERRITORY. Read before the American Philosophical Society, Aug. 15, 1884. By Professor E, D. Cope. IPOK. S^LE BIT J^. E. :F'00TE, 1223 BELMONT AVENUE, PHILADELPHIA. May 16, 188<.] 1 [Cope. The Extinct Mammalia oj the Valley of Mexico. By E. D. Cope. {Read, before the American Philosophical Society, May 16, 188 J^. ) The following study is based primarily on an examination of the speci- mens contained in the Museum Nacional of Mexico, which I was permit- ted to make through the kindness of the Director of the Departments of Geology and Mineralogy, Professor Mariano Barcena. Through the me- diation of the same gentleman, I obtained permission from Professor Antonio Castillo, Director of the School of Mines, to examine the corres- ponding material preserved in the fine museum of that institution. The knowledge derived from the study of the latter, reinforced the results I obtained from the study of the specimens of the Museum Nacional, so as to enable me to reach definite conclusions as to the definitions of various species which are represented in both collections. I wish to record the obligalions under which I have been laid by both of these distinguished gentlemen. I have, through their aid, been enabled to make a comparison between the pliocene fauna of Mexico, and that of Buenos Ayres, and that of Oregon. The species of the Pampean fauna contained in my private collection, are those exhibited by Messrs. Ameghino, Larroque and Brachet, at the Exposition of Paris of 1878. My Oregon material is derived from the explorations of my parties under Messrs. Sternberg and Duncan, and those of Professor Thomas Coadon of the University of Oregon, who kindly lent me his collection. The collections of the museums of the City of Mexico, above mentioned, are derived from the locality Tequixquiac, and the specimens referred to in the following pages are to be understood as having been derived from that locality unless otherwise stated. Tequixquiac is situated on the northern edge of the valley of Mexico, north of the City of Mexico and the town of Zimpango, and east of the gorge of Nochistongo. Cope.] 2 [May 16, GLYPTODON Owen. Glyptodon, sp. indet. A nearly complete carapace of this remarkable animal is mounted in the Museum Nacional, and a second, nearly as well preserved, is in the Mu- seum of the School of Mines. Jaws and teeth occur in the latter museum. The discovery of this genus at this extreme northern locality is due to Dr. Antonio Castillo. It was first announced by Dr. Mariano Barcena in the Revista Cientifica of Mexico, 1882, I, p. 3. The extension of this far southern genus to the latitude of Mexico during the Pliocene (Pampean) epoch, is entirely consistent with the further distribution of the great sloths and llamas to the United States at the same time. DIBELODON Cope. Mastodon pars, auctorum. Various attempts have been made to define as genera groups of species which are included within the limits of the genus Mastodon of authors. The first new name, Tetracaulodon, was introduced by Dr. Godman, who saw in the mandibular tusks of some individuals of the Mastodon ameri- canus Cuv., ground of its separation from the genus Mastodon, in which he believed those teeth to be wanting. This division was adopted by Dr. Grant and others, but has not been generally allowed. The next division was that proposed by Dr. Falconer, who, however, did not employ the names proposed by him in more than a subgeneric sense. He distinguished two series in the genus Mastodon. In one of these, the P-m. 3, and the Ms. 1 and 2 present three transverse crests, while in the other division these teeth present four such crests. To these divisions he gave the names of Trilopliodon and Tetralophodon respectively. The third attempt at division is that of Herr Vacek, who gives names to the two divisions of the genus in whicli the cross-crests are composed of tubercles or continu- ous ridges. These divisions he calls Bunolophodon and Zygolophodon respectively.* I will refer to these divisions in reversed order. Those proposed by Vacek cannot be regarded as genera, and their author did not use them as such. The tubercular crest passes into the straight crest by insensible stages. The divisions proposed by Falconer are more distinct, but not sufficiently so to represent genera. This may be understood by reference to the second lower molar of the Mastodon augustidens, which is, in some individuals, three crested, and in others four crested. Some other species present the same difficulty. On this point I quote the remarks of Dr. Lydekker :f "The foregoing survey of such a large series of Mastodon molars has led to the conclusion that the very regular ridge formula given by Falconer will not always hold good in regard to the true molars, * Vacek, Ueber Oesterreichische Mastodonten. Abh. der K. K. Geolog. Reich- anstalt, vii, Heft iv, "Wien, 1877, p. 45. t Geological Survey of India, Series x, Vol. i, pt. v, 1880, p. 256. 1881.1 3 [Cope. though ill the Indian species, at all events, it appears to be always con- stant in the milk-molars. We have seen that there is a tendency in the true molars of some of the Trilophodons (M. falconeri) to develop the talon into a fourth ridge, and in the Tetralophodons (M. latidens and M. sivalensis), a similar talon is developed into a fifth ridge, in the intermedi- ate true molars." M. humboldtii Cuv. (3/. andium Falc.*) shows a small fourth crest on the second true molar, according to Falconer. f The lower incisor teeth, on which Godman relied for the definition of his genus Tetracaulodon, were shown by Harlan, not to be constantly present in the Mastodon americanus. In fact, no adult specimen has been described in which two inferior incisors are present. The single one ob- served is very rarely found in adults, being a character more frequently found in the young. It is in this species a remnant of a character else- wiiere constant, which does not disappear quite so soon as the teeth of the whalebone whale, and superior incisors of the ruminant. But it is other- wise with other species referred to Mastodon. No specimens of the Mas- todontes angustidens, prodmtus and longirostris, are recorded, in which two inferior incisors are not present. For this reason the first and last- named were placed by Grant and others in the genus Tetracaulodon. Unfortunately this name was applied by its author to the M. americanus only, a species which cannot enter the genus furnished with a pair of per- sistent inferior incisors. It is also the type of Cuvier's Mastodon. J It thus unavoidably becomes a synonym of the latter. There is no doubt that the presence of a pair of persistent inferior in- cisors defines a genus as distinct from one in which there is not a pair of permanent inferior incisors. I agree, therefore, with Grant and others, in separating the Mastodons which present this character from the genus Mas- todon, under another generic head. I believe, also, that the presence or absence of a band of enamel on the superior incisors furnishes ground for the recognition of distinct generic groups, and would be so used in any other division of the Mammalia. It is often asked why it is necessary to multiply generic names on such grounds. My. answer is simply an ex- pression of the law governing the case, based on the supposition that when the species of animals and plants come to be fully known, the genetic series will be found to be uninterrupted, excepting by the presence or ab- sence of characters which appear or disappear during the growth of a set of individuals, which we on this account call a species, or refer to genus. The difference in the two cases consists in this : In the case of species, the characters are numerous and are matters of proportion, size, color, texture, etc., while in the case of the genus the character is single, and marks one step in the serial chain of structural modifications. In the case of the genus there is an actual addition or subtraction of some distinct * Paheontological Memoirs of Falconer, i, p. 100, pi. 8. t t Loc. cit., ii, p. 15. X Ossemens Fossiles, ii, p. 2.52, Ed. 18o4 : *' Ann. Mus., 1806, viii, 272," teste Leidy. Cope.] 4 [May 16, part or piece of the organism.* If now we fail to notice these points or steps, we must abolish all genera. If we define some and fail to define others, our practice ceases to have the uniformity of a law, and we aban- don the basis of scientific order. f One point, however, must be insisted on. In order that a character be usable for any purpose of definition, it must define. That is, it must belong to all the individuals referred to the species, genus, etc., defined, and must not be present in some individuals and wanting in others of those supposed to be defined by it. This being the case, adult animals only can be used for definition, as characters, es- pecially generic, are added from time to time up to maturity. Sometimes only one sex can be considered, since the adult characters are in certain cases never reached by one sex or the other. This is often the case with insects. Moreover, some latitude for exceptional variations must be al- lowed. Thus, the exceptional absence of the last molar in a dog does not invalidate the definition of the genus Canis, M. |. Of course, if all specimens of animals could be found, the definitions would all, or nearly all, be invalidated. But it is safe to assume that all the intermediate forms will not be found, so that the definitions of species will represent the state of our knowledge, and the results of the operations of nature in the preservation of individuals. The case is somewhat difierent with regard to generic characters. As these involve the addition or subtraction of some part, having definite dimensions, it is quite possible to say when the latter is present or absent. Characters of this kind present the appearance of abruptness of transition, to which I have referred in my paper "On the Origin of Genera," and which gave rise to the formulation, by Professor Hyatt and myself, of the "laws of acceleration and retardation." When such change prevails throughout all the individuals of one or more species, a new genus has its origin. As a matter of fact, the creation of generic modifications has been exhibited, in the history of life, by many individuals nearly contempora- neously. As the change involves hut one character, it offers a better oppor- tunity for the formulization of the laws of evolution, than in the case of specific characters, which are more numerous. The three genera of Elephantidae, of which mention has been made above, will then be defined as follows : Mastodon Cuv. Superior incisors without enamel band ; inferior incisors wanting. Type M. americanus. Dibelodon Cope. Superior incisors with enamel band ; inferior incisors wanting. Type D. shepardi. Tetrahelodon Cope. Superior incisors with enamel band ; inferior incisors present in the male at least. Type T. angustidens. To the genus Mastodon must be referred the following species. For * See " Origin of Genera," Proc. Acad. Pbilada., 1869, where this point is dis- oussed. t American Naturalist, 1884, July, p. 188^.J 5 [Cop( the dental characters of the Indian species I am indebted to Messrs. Fal- coner and Lydekker : Mastodon americanus Cuv., N". America. " fborsoni Hays, E. and S. Europe. " mirificiis Leidy, N. America. " /aZcowm Lydd., India. ** arvernensis C. & J., Europe. " sivalensis Falc, India. " . latidens Clift., India. Dibelodon shepardi Leidy, California, Mexico. ' ' tropicus Cope, Tropical America. " hiimholdtu Cuv., South America. Tetrabelodon angiistidens Cuv., India, Europe, N". America. " andium Cuv,, S. America, Mexico. " prodtictus Cope, SW. N. America. " euhypodon* Cope, N. America. " pentelici (jt^Vi(\YY, SE. Europe. " perimensis Falc, India. " pandionis Falc, India. ** turicensis\ Schinz, Europe. " campester Cope, N. America. " longirostrisKsLup, Europe. The condition of the inferior incisors is unknown in the Mastodon atti- cus Wagner, and M. serridens Cope, and M. proavus Cope ; and in some of the above species the presence of an enamel band on the superior incisors has not been established. I may add that I do not perceive how the so-called genus Stegodon can be distinguished, as at present, by the number of crests of the intermediate molars, and by the presence of cementum. It will probably be necessary to look for other characters in order to sustain it. Dibelodon shepakdi Leidy. Mastodon shepardi Leidy, Proceedings Academy Philadelphia, 1870, p. 98 ; 1872, p. 142. Mastodon obscurus Leidy part, Report U. S. Geol. Survey Terrs. I, p. 330, Plate xxi. This species was originally proposed on the evidence of a last inferior mo- lar tooth from Contra Costa county, California, and a part of a superior tusk from Stanislaus county in the same State. Dr. Leidy subsequently aban- doned the species. I however revived it in a synoptic table of the species of North American Mastodons in 1884. X The fossils of the Museum Nacional of Mexico, examined by me, included * American Naturalist, 1884, p. 525. t Von Meyer is my authority for the presence of mandibular tusks in this species, = M. virgaiidens Meyer, i American Naturalist, 1884 , p. 524. Cope.] 6 [May ]6, a well-preserved lower jaw of a Mastodon, which presents both rami, and both the last true molars, and the entire symphysis. In the collection of the Ecole des Mines I saw a palate with the second and third true molars of both sides in place, and the superior incisor teeth, or tusks. Other fragments of jaws, with numerous isolated molars, were seen in these col- lections and in that of the college of the city of Toluca.* From these specimens it is clear that the high valleys of Mexico were inhabited by a trilophodont mastodon, with a short decurved toothless symphysis like that of the Mephas primigenius, and with a band ot enamel on the superior incisor tusks. The molars have the character^ of those of the Mastodon andiutn of authors, and are of about the same size. The cross-crests are divided at the middle line only, and one half wears into a trefoil, while the other half wears into an oval, transverse to the long axis of the crown. The unworn crests are obtuse and not serrate ; and there are no accessory tubercles besides those forming the lateral lobes of the trefoils. The size of the ramus and of the teeth is about that of the M. angustidens, and smaller than that of the M. humholdtii. The last in- ferior and last superior molars have but four cross- crests and a small heel. This I verified on several specimens. A comparison of this species with those described, yields the following results : In the character of its molars it is identical with the M. andium, and differs from the M. humholdtii in the characters which distinguish the two species, as pointed out by Gervais.f That is, only one-half of each cross-crest wears into a trefoil, and the size is inferior. But it cannot be identified with the 2'etrabelodon andium, because, according to Falconer,:}: that species possesses a long massive deflected beak containing an incisor tooth. II It is true that the specimen figured by Laurillard in D'Orbigny's voyage dans I'Amerique Meridionale, PI. x, does not display a long bealc and tusk, although the symphysis is much more pronounced than in the present species. But that plate is made from a drawing, and may thus be of doubtful authority. If correct, it may represent the female, or, as Falconer suggests, the young of the T, andium. The last inferior molar figured by Dr. Leidy, 1. c, and formerly referred to a species under the name of Mastodon shepardi, has the character of the corresponding tooth of the Mexican species under consideration. The plate does not, how- ever, represent the specimen satisfactorily in one respect. The trefoils are not sufficiently distinct, on account of the faint representation of their basal lobes. These nearly block up the cross valley, a fact not to be de- rived from an examination of the plate, but which is clearly seen in a cast preserved in the museum of the Philadelphia Academy of Natural Sciences. * For the opportunity of examining the museum of tliis Institution I am much indebted to its President, Dr. Villada. t In Castelnau's Expedition, 1855; Recherches sur les Mammiferes Fossiles de I'Amerique Meridionale, p. 14. X Palseontological Memoirs, ii, pp. 226, 274. II The lower jaw figured by Falconer, Alem. i. p. 100, from Buenos Ayres, as M. audium is clearly M, humholdtii. 1884.] 7 [Cope. This specimen also agrees with tliose in the Mexican museums in the small number of crests on the last inferior molar : four with a short rudimental heel. Another specimen of apparently the same species is described and figured by Leidy as having been brought from Tambla, Honduras.^ This tooth is apparently anomalous in the contraction of the third cross-crest. The range of this species may then be given as extending from Califor- nia to the valley of Mexico, inclusive. A species apparently allied to the Dibelodon shepardi is the Mastodon serridens Cope,f of which the typical specimen was brought from south- western Texas. Premolar teeth of the same type were shown me by Professor Castillo, in the museum of the School of Mines. These came from a lignitic bed at Tehuichila, in the State of Morelos, of Loup Fork age. This epoch is indicated by the presence of the genera Protohippus and Hippotherium. The sharp, serrate edges of the crests distinguish the molar teeth from those of the D. shepardi, and as the species probably came from difierent horizons, they are probably distinct. A premolar mingled with those of D. shepardi, from the valley of Toluca, much re- sembles that of the M. serridens. DiBELODOX TROPICUS CopC, sp. nOV. Mastodon humholdtii? VonMeyer Palceontographica, 1867, Studien ueber das genus Mastodon, p. 64, PI. vi. Mastodon andium Leidy, Pro- ceedings Academy Philada., 1876, p. 38. A second species of Dibelodon inhabited the valley of Mexico, of larger size than the D. shepardi, and difiering somewhat in the dentition. Von Meyer describes and figures a ramus of a lower jaw, 1. c, brought by Herr Uhde from Mexico, which has, according to Von Meyer, no mandibular tusk, and probably a short elephantine symphysis. A very similar ramus, containing the last molar tooth, was presented to the Philadelphia Acad- emy of Natural Sciences by Dr. Isaac Coates, who obtained it from Tarra- pota, on the Huallaga river, in Eastern Peru. The extremity of the symphysis of this specimen is broken away, but enough remains to show that it was probably short, and that there was no inferior incisor. Reference to Von Meyer's figure shows that the last inferior molar has five well-developed cross-crests and a heel. The Peruvian specimen has the same character, the fifth cross- crest a very little more contracted than in Von Meyer's plate. Dr. Leidy describes the specimen as having four transverse ridges, besides a strong tubercular talon. But it seems to me that the talon is of such size as to be properly included in the cross-crests. On the same principle one might say that the Z>. shepardi has three cross- crests and a strong talon, as it has one less cross-crest than the D. trqpicus. The additional cross-crest, and the superior size, distinguish this form as a species from the D. shepardi. Von Meyer perceived these differences, and referred his specimen to the B. humholdtii. I am fortunately able to * Extinct Mammalia Dakota and Nebraslja, PI, xxvii, flg. 14. t American Naturalist, 1884, p. 524. Cope.] 8 [May 16, make a comparison of his plate and the Peruvian jaw, with a well pre- served jaw of the D. humholdtii, with perfect last molar and symphysis, from Buenos Ayrcs, in my collection. I am able fully to substantiate the characters already pointed out by Gervais, and to prove that the cross - crests of the molars form double trefoils, while those of the D. tropicus are like those of D. sTiepardi and the Tetrahelodon andium. The species last named is said by Falconer (loc. sup. cit.) to occur in Mexico, and speaks of having seen a well preserved lower jaw from the State of Tlaxcala. I have not met with it. The Mastodon americanus has not yet been found in Mexico. The most southern localities for the species known to me are Southern California, and near San Antonio, Texas. From the former region I possess a ramus with the last molar, presented to me by Mr. Scupham, of San Francisco ; / the other specimen was obtained from Mr. G. W. Marnock, of Helotes, near San Antonio, Texas. ELEPHAS Linn. Elephas primtgenius Blum. This species, of both the thick and thin plated varieties, was once very abundant in Mexico. I have received a series of teeth from Candela, in the State ot Coahuila, from Dr. Caspar Butcher, through my friend Dr. Persifor Frazer ; and Von Meyer has pointed out the occurrence of its re- mains in the valley of Mexico. The museums of Mexico contain very numerous portions of skeletons of this species, which prove that it was far more abundant than the species of Mastodon. Up to this time this locality is the southern known limit of its distribution on the American continent. APHELOPS Cope. Aphelops, sp. Aphelops ffossiger Co])e. Proceedings Academy Phil- adelphia, 1883, p. 301. The right half of the mandible, with part of the symphysis of a rhinoc- eros, was found in the valley of Toluca, sixty miles west from the city of Mexico, and Dr. Barcena sent me a photograph of it a year ago. I pub- lished a notice of it as above cited, in connection with remarks on a rhi noceros skull which I obtained on one of the heads of the Gila river in New Mexico. On my recent visit to the College of Toluca, I had, through the kindness of Professor Viilada, the opportunity of examining the jaw. Its characters do not differ much from those of the Aphelops fossiger Cope. It is considerably smaller, and has a very short diastema, but not shorter than in some jaws of the A. fossiger. The dimensions are as follows : Measurements. M. Length of ramus from base of canine 400 " " dental series with canine, less M. iii 235 " " molar series, less M. iii 200 " true molars, less M. iii 105 1834.] 9 [Cope. Measurements. M. Diameter of canine (transverse) 027 of P-m. ii 007 Depth of ramus at P-m. iii 070 " " at M. i 085 ** " " at front of M. iii 090 The matrix in wliicli tliis jaw was found, is much like the Upper Plio- cene material of Tequixquiac. It is therefore of probably later age than the true ApJielops fossiger, which is a characteristic Loup Fork species. Leidy describes (Extinct Fauna of Dakota and Nebraska, p. 230) a rhi- nocerus, probably an Aphelops, from California, under the name of R. hesperius. It is smaller than the Toluca specimen, but has a considerably longer diastema. Its geological horizon is uncertain. I mention here that rhinoceroses, probably of the genus Aphelops, ap- parently existed in North America during the Pliocene period. Bones of a species having resemblances to the A. fossiger have been sent me by my assistant, George C. Duncan, from the Equus beds of the eastern part of the Oregon desert. The genus has been hitherto supposed not to ascend higher than the Loup Fork, or Upper Miocene beds. These bones are accompanied by teeth of a peculiar Hippotherium unlike those of any spe- cies of the genus known to me from the Loup Fork Miocene. EQUUS Linn. The remains of horses are very abundant in the valley of Mexico,* and represent four species. In the determinatiou of these species it has be- come necessary to compare them with those hitherto found in North and South America. In making this comparison I exclude the species of Hippidium, which are all American, and whose molar teeth are easily dis- tinguished by the equality in size of the internal columns ; resembling in this respect the genus Protohippus. When the species of the genus Equus difier in the characters of their superior molar teeth, the diversity is to be seen in the size and form of the anterior internal column. The anteroposterior diameter of this column, as well as the integrity or emargiuation of the internal border of its section, varies according to the species. The infolding or the borders of the lakes has a value, but a: less constant one. The Equus caballus differs from all of the American extinct species, where the corresponding parts are pre- served, in the great elongation of the face, which is expressed in the greater lengths of the diastemata anterior and posterior to the canine tooth in both jaws. Other characters may be observed in the relative lengths of the limb bones, the form of the occiput, etc. It has been shown by Leidy, Riitimeyer and others, that it is not always practicable to distin- guish the species of horses by their teeth alone. A glance at Owen's * This fact has already been made known by Von Meyer, Palfcontographica, 1867, p. 70, and Owen, Transactions of the Royal Society, London, 1869. Copp.] 10 [May 16, plates of the dentition of the existing species of Equus*, shows the truth of this statement. Among tlie extinct species of Equus the range of va- riation is greater. Tlie following attempt at a discrimination of the species known to me, or so fully described as to be well known, must necessarily be regarded as provisional, until the skeletons are more fully recovered. American ex- tinct species only are introduced : I. Long diameter of anterior internal lobe of superior molars not greater than one third the long diameter of the crown. Borders of lakes crenate ; internal anterior lobe notched on the inner side so as to be bilobate; crowns a little curved ; large E. crenidens. II. Long diameter of anterior internal lobe more than one-third and not more than one-half the anteroposterior diameter of the crown. a Crowns more or less curved. Crowns wider than, or as wide as long ; enamel edges little folded £J. curvidens. aa Crowns straight or nearly so. /5 Diastemata longer. Crowns nearly square, enamel not very complex ; no facial fossa ; maxil- lary bone produced much beyond M. iii E. caballus. y5/S Diastemata shorter. y No facial fossa. Crowns nearly square ; enamel not very complex ; maxillary bone little produced behind last molar ; smaller E. hemionus; E. burcJielli; E. quag g a; E. zebra ; E. asinvs. Crowns longer than wide on face ; enamel little complicated ; face and maxillary unknown ; large E. ocddentalis. Crowns square ; enamel more folded than in other species ; face and max- illary unknown ; large E. major. Y A facial fossa. Crowns nearly square ; enamel less complex ; maxillary short posteriorly ; smaller E. andium. III. Long diameter of anterior inner lobe more than half that of crown of molar teeth. Crowns square ; enamel little complex (in Mexican specimens) ; diaste- mata and maxillary behind shorter ; no facial fossa ; large. ..E. excelsus. Crowns square ; enamel little complex ; smallest species E. barcencei. In using the above table it must be noted that gradations in the diame- ter of the anterior internal column (or lobe) exist, not only between indi- viduals of the same species, but between different teeth in the same jaw. This diameter is always greatest in the last superior molar, and the charac- ters of this tooth are such that they cannot be used in connection with the above table. * Philosophical Transactions, 1869. 1881.1 -L-L I Cope. Before describing the Mexican species I make some notes on the others embraced in the above list : Equus curmdens Owen. Of eight superior molar teeth from Buenos Ayres in my collection, two second premolars are perfectly straight, while the third true molar is the most curved. The other teeth exhibit different degrees of curvature. The area of the anterior internal column is not so flat on the inner side in any of them as in Owen's Plate (Voyage of the Beagle, Vol. i). My teeth have also a rather greater transverse diameter than Professor Owen's type. Equus caballus Jj. The common horse differs from all of the extinct species of the genus from American localities where the muzzle is known, in the greater length of the latter, with its diastemata, of both jaws, and in the greater prolongation of the maxillary bone posterior to the last true molar. Appropriately to the anterior position of the molar series, the fiicial ridge commences above the middle of the first true molar. In an Equus quagga in my possession the ridge commences above the middle of the last pre- molar. The basioccipital bone is more compressed than in any species of the genus known to me. Equus occidentalis Leidy. This species is represented in my collections by at least one hundred individuals, some of which have been lent me hy my friend Professor Thomas Condon of the University of Oregon. TXiej are nearly all derived from the Equus beds of the Oregon desert. Unfortu- nately there is no perfect skull. A few specimens from the same region I refer to the Equus excelsus, but as these are comparatively rare, I am safe in referring most of the bones to the other species. In these I find the following characters to separate the species from the Equus caballus : (1) The basioccipital bone is not compressed, and besides its inferior lateral angles it has a pair of lateral angles, one proceeding forwards from the inferior border of ea.ch. foramen condyloideum anterius. (2) The fossa en- closed between the paroccipital process and the basioccipital, is deeper, and has a raised border in front which separates it strongly from the plane of the petrous bone. This is not found in E. caballus. I verify it in three separate occipital bones of the E. occidentalis. (3) The astragalus and other bones of the feet are smaller than in E. caballus; the first named intermediate in size between that of the horse and that of the quagga. The cannon bones, when of the same length, are more slender. (4) The inferior canine issues in direct contact with the last incisor, without the diastema seen in the horse ; and the incisive arc is narrower and more produced. The symphysis is elongated not only forwards, but also pos- teriorly. The mental foramen is anterior to the bifurcation of the rami in E. occidentalis, posterior to it in E. caballus. Equus major Dekay. Dr. Leidy leads us to infer (Report U. S. Geol. Survey Terrs., Vol. i, p. 244), that this species differs from the E. occiden- talis, in the generally greater complication of the enamel folds. This I find to be the case in specimens from the Fish House, in the brick clay, near Philadelphia, and from the Big Bone Lick, Kentucky. Leidy Cope.l 12 [May 16, figures similar specimens from various parts of the Eastern and Southern States. Equus crenidens Cope, sp. nov. This large species of true horse is represented by molar teeth and frag- ments of jaws belonging to two individuals preserved in the Museo Na- clonal of Mexico, and to two others preserved in the Escuela des Minas- The typical specimen includes the three premolars of the upper jaw of an adult in perfect preservation. The species is primarily distinguished by the close and strong wrinkling of the enamel border of the lakes of the superior molar teeth. This wrinkling, or vertical plication, reminds one of what is seen in the Ele- jpJias indicus. This wrinkling is not found in the enamel edges which border the interior crescents on the inner side, nor in those bordering the internal lobes or columns. The borders of the lakes are not folded in the complex loops seen in Equus major Dek., but have the plainer looping seen in the Equus tan Ow. The grinding faces are nearly square. That of the second premolar is a rather shortened triangle, and less produced anteriorly than in the E. tau. The crowns of the third and fourth pre- molars are long and slightly curved. The measurements show that this is one of the larger species of horse. Measurements. M. Diameters of P-m. ii / anteroposterior 0430 t transverse 0305 Diameters of P-m. m 5 anteroposterior 0335 ' transverse 0340 Diameters of P-m. iv j anteroposterior , , .0310 i. transverse 0350 The crimping of the enamel of the lakes distinguishes this species from the others of the genus. From Tequixquiac. Equus tau Owen. Philosophical Transactions of the Royal Society, 1869 ; p. 565 ; pi. Ixi ; fig. 4. Of this species there are preserved in the Museum Nacional five superior molars, some of which belong apparently to one individual. In the Escuela des Minas, the series is a fine one. There are two skulls lacking the occi- put ; one skull lacking the occiput and muzzle ; parts of both maxillary bones with teeth, of one skull ; and a single maxillary bone with teeth, of a fifth skull. The specimen mentioned under the second head, has teeth and palate preserved, as in the figure given by Owen of his Equus conver- sidens, and I suspect it was from this specimen that the photograph was taken from which Professor Owen's figure and plate were made. It is pos- sible that his figure and description of the Equus tau were made from one of the maxillary bones mentioned under head three. I am not able to perceive the specific diflerences between these specimens. The character displayed 1881.] 13 [Cope. by Owen's E. conversidens, on which he relied to distinguish the species, may be the result of distortion. The maxillary bones of the type are loose and may be made to assume different angles to each other. The last superior molar is represented as unusually short by Owen. This appearance could be produced by the oblique angle of the aperture of the camera in photographing, due to its too anterior position. Be that as it may, I could detect no specific differences between the seven or eight spec- imens I examined. The Eqims tau is an average horse in all respects, presenting no very tangible characters by which to distinguish it from the existing species of the E. asinus and E. zebra group, so far as the parts which I examined go. It has the internal anterior column of the superior molar always less in diameter than half that of the crown of the tooth, and not characterized by any marked peculiarity. The borders of the lakes have an entering loop on each end of the inner border ; of these the adjacent ones are well marked, and the remote ones little marked. External to the adjacent loops the borders of the lakes are a little crenate. There is a small internal median loop of the internal enamel border at the notch. The crowns of the teeth are a little wider than«long, and they are not curved. The pal- ate notch reaches as far forwards as the posterior border of the second true molar, and the palatal foramen is opposite the front of the third true molar. The latter tooth is a little longer than the other true molars. The second premolar is short and robust. The diastemata are rather short, as can be seen by the appended measurements. Measurements. M. No. 1. Escuela des Minas. Length of precanine diastema 020 Length of postcanine diastema 074 Length of molar series 151 No. ^. Museum Nacional. Diameters of P-m. ii / anteroposterior 030 t transverse 024 Diameters of ?P-m. iii \ anteroposterior 024 c transverse 027 Diameter of ?P.m. iv \ anteroposterior 025 ( transverse 028 This species differs from the Equus andium Wagn., so fully described by Branco,* in the absence of a facial fossa. From Equus caballus it difiers in the short diastemata, and the little posterior production of the maxillary bone. How it differs from the species of the asinus section I do not yet know. Equus excelsus Leidy, Extinct Mammalia Dakota and Nebraska, 1869, p. 266 ; pi. xxi, fig. 31. *In Dames and Kaysei Palasontologische Abhandlungen, 1883, p. 110, Dr. Branco furnishes reasons for believing that the E. argentinus Burm. is the same species Cope.] 14 [May 16, A portion of a left maxillary bone supporting the true molars of a horse from the Oregon desert, received from Professor Condon, resembles closely the type specimen from Nebraska described by Leidy as above. Two skulls, in the two museums of Mexico already referred to, present the same dental characters. In identifying the Mexican with the Oregon and Ne- braska horses, I wish to be understood as making a provisional arrange- ment only, for unfortunately the cranium of the North American horse with this dentition is yet unknown. The uncertainty attending a dental identification being admitted, I proceed to the description. This species differs from the others, whose remains have been found in the valley of Mexico, in the elongate and flattened form of the lobe formed by the section of the anterior internal column of the superior molar. This long diameter generally exceeds the half of that of the crown of the • tooth by one-eighth the latter, and is rarely so short as one half of the same. The loops of the lakes are few, including only one near the pos- terior borders near the internal side and one on the anterior border of the posterior lake. There is generally a little loop at the notch between the two internal lobes. Crowns straight, second superior premolar elongate and acute. One of the crania is complete, lacking only the lower jaw, and the two third true molars. The other lacks all posterior to the palatal notch. From the former I derive the following characters : The apex of the nasal bones is above the superior canine tooth. The posterior border of the nares marks the middle of the anterior column of the third premolar. The infraorbital foramen is above the posterior edge of the second column of the fourth premolar. There are two notches on the anterior part of the superciliary border ; and there is a short exos- tosis on each side of the front, in line with the supraorbital border, in front of the preorbital border. Measuremenis. M. Length from superior edge of foramen magnum to in cisive border From posterior nares to incisive border Interorbital width Length of series of molar teeth ** precanine diastema " postcanine " Width of palate at third incisors " " canines, inclusive Diameters P-m. ii /anteroposterior i- transverse Diameters P-m. m ^ anteroposterior (. transverse Diameters M.iii ^anteroposterior ( transverse .565 .300 .166 .191 .022 .056 .093 .075 .0425 .0275 .032 .034 .0335 .029 ISSJ ) 15 [Cope. The internal anterior column of the superior molars is longer and flatter than in the specimens of the North American horse, but I do not feel at liberty to propose a new specific name for the Mexican animal. The absence of facial fossa and short diastemata throw it into the series of the asses. From all these the large flat internal column distinguishes it. The presence of the loop at the notch of the internal border in the Mexican specimens distinguish them from Leidy's type and from one of Condon's specimens. A second one of the latter has a small loop at the point in question. The absence of this loop is given by Leidy as characteristic of the E. occidentalis, but only a small proportion of my specimens of that species are without it. The Mexican specimens are from Tequixquiac. Equus barcen^t Cope, sp. nov. Two superior molars represent this species in the Museum Nacional. and two superior molars in the Escuela des Minas. A skull lacking all in front of the orbits inclusive, in the latter museum, probably belongs to the same species. This horse is distinguished from all the others here mentioned or des- cribed by its small size. In the characters of its superior molars it is like the Equus excelsus. The anterior internal column is flat, and its antero- posterior diameter is five-eighths that of the crown of the tooth. The prism is straight. The lakes have the margin but little looped ; the pos- terior notch of the anterior lake is trebled or triplex. The grinding face of the crown of the third superior molar is a little longer than the others. Measurements. M. Diameters of molar No. I \ anteroposterior 0215 i transverse 0230 Diameters of molar No. II \ anteroposterior 022 ( transverse 022 From Tequixquiac. I have dedicated this species to my distinguished friend Mariano de la Barcena, Professor of Geology in the National Museum and Director of the Meteorological Observatory of the City of Mexico. PLATYGONUS Leconte. Pla-tygonus ?compressus Leconte. A portion of the mandibular ramus of a species of peccary, apparently the above, was found at Tequixquiac, and is preserved in the museum of the College of Guanajuato. Dr. Alfredo Duges, the distinguished pro- fessor in the college, called my attention to the specimen, and gave me a cast of it. Its dimensions are similar to those of North American indi- viduals, as follows : Measurements. M. Diameters of M. i | anteroposterior 0145 <^ transverse 012 Cope.] J-O [May 16. Measurements. M. Diameters of M. n / anteroposterior 017 t transverse 014 HOLOMENISCUS, gen. nov. Under the head of this genus I give a synopsis of the results of my study of the extinct Camelidse of the American Pliocene epoch. I can compare the specimens from Buenos Ayres with those from Mexico and Oregon, and Branco and Owen have given detailed descriptions of speci- mens from Buenos Ayres and Mexico. From these sources I learn of the existence of the following generic forms of Camelidae. I omit Protolabis Cope,* and refer it to a separate family — the Protolabididse, on account of the presence of three superior incisors in each premaxillary bone, as in the primitive Ruminantia, combined with the presence of a cannon bone. I. Premolar teeth f . P-m. i separated by diastemata Procamelus, II. Premolar teeth f . P-m. ii below wanting PUauchenia. III. Premolar teeth |. Fourth inferior premolar triangular Cameliis. Fourth inferior premolar composed of two crescents, which enclose a lake (an inferior P-m. 3?) Palauchenia. Fourth inferior premolar composed of two crescents, with two posterior tubercles behind them ProtaucJienia. IV. Premolar teeth f- Fourth premolar below triangular AucTienia. V. Premolar teeth \ Fourth superior premolar composed of two crescents Holomeniscus. Fourth superior premolar consisting of a simple cone Eschatius. The position of this genus being determined as above, it remains to examine the material representing it, at my disposal. In 1873 Dr. Leidyf described a large species of llama from specimens from California, which include the entire inferior series of molar teeth, and one superior molar. The first inferior molar, properly the fourth pre- molar, has the crown partiall}^ worn, showing that it was opposed by a grinding tooth in the superior series. In the Museum Nacional of Mexico is preserved a complete mandibular ramus, containing all the teeth of one side of an animal smaller than Dr. Leidy's type, but having a general resemblance to it ; including the worn fourth premolar. In the collections of Professor Condon and myself from the Oregon desert, there are vari- ous isolated molars agreeing in measurements with Dr. Leidy's type, and belonging probably to the same species. In the Condon collection is part * Proceedings Academy Philadelphia, 1876, p. 145. t Report U. S. Geol. Survey Terrs., b\ V. Hayden, i, p. 255, pi. xxxvii, figs. 1-3. 1884.] 17 rCope. of a superior maxillary bone which contains the M. i and the alveolus of the P-m. iv, with the foramen infraorbitale anterius. The measurements of the M. i agree with those of the corresponding tooth of the lower jaw of Leidy's specimen. In the Museum of Mexico, there are preserved several superior true molars which also agree in dimensions with the correspond- ing teeth of the lower series of the type of the same A. hesterna of Leidy. The fourth superior premolar is wanting from this series. The fragment of maxillary bone in the Condon collection shows that this species had a large three-rooted fourth premolar. It is broken off at the anterior alveolus, but it is so attenuated at that point as to make it almost certain that there was no third premolar in front of it, as is found in the genus Auchenia. In further evidence of the existence of a genus characterized as above, by the absence of the P-m. ^, the jaw-fragment which represents the Auchenia vitakeriana* may now be cited. Holomeniscus mtakerianus Cope. Although I ascribed a third superior premolar to this species, I must now deny its existence in the adult animal. A slight fossa on the narrow alveolar ridge indicates the possible presence of a single-rooted rudiment of such a tooth in the young. In a comparison of this species with the Auchenia weddeUii Gervais, from the Pampean beds of Buenos Ayres, it is readily observable that the latter is a true Auchenia, with well devel- oped P-m. 3 in the upper jaw, and that it is of larger and more robust pro- portions than the E. mtakeriana. In the only well preserved lower jaw which I possess, there is a well developed P-m. iii, a tooth found only as an occasional accident in Auchenia lama (teste Owen Odontography). In the A. intermedia Gerv., from the same locality, this tooth is wanting from one ramus, while the other displays a shallow vacuity as though such a tooth had existed in infancy and had been shed. I therefore retain these species in Auchenia. Holomeniscus hesternus Leidy. Auchenia liesterna Leidy, loc. sup. cit. The existence of superior molars in the Museum Nacional of Mexico which agree with the corresponding teeth of the Californian and Oregonian llamas has been mentioned above. I give the dimensions of these teeth as follows : Measurements. M . Diameters M. i r anteroposterior 041 I transverse 033 ] ^ anteroposterior 041 transverse 040 Diameters M. ii I one individual. , Diameter M. iii \ /anteroposterior 053 j transverse 029 * Bulletin of the U. S. Geological Survey Terrs., 1878, p. 380. Cope.] 18 [May 16, Diameters P-m. iv Tliese molars are covered witli a layer of cementum, which is included in the measurements. The mandible, I am disposed to refer to a smaller variety of this species for the present. The well-worn fourth inferior premolar indicates that it could not belong to the genus Eschatius, where tliere is no opposing tooth in the superior series capable of producing such a result. The hook be- low the condyle is well developed in this jaw. The incisor teeth are narrow. The canine is small and is separated from the incisors by a dias- tema. The triturating surface of the fourth premolar is triangular, and includes a lake. The molars increase in size posteriorly. The mental foramen is large, and is situated behind a point below the canine. Measurements. M. Length of jaw from incisive alveoli to angle 415 Height at coronoid process 290 at condyle 318 *• ramus at M. i 070 " " middle of diastema 040 Length of symphysis 096 " from base of incisors to canine 043 *' canine to P-m. iv 092 " of all the molars 147 / anteroposterior 022 I transverse 013 Diameters M. i 5 anteroposterior 085 c transverse 019 Diameters M. ii j ^"'e^Po^'erior 042 (. transverse 019 Diameters M. iii i anteroposterior 048 I transverse. 01b From Tequixquiac. A cannon bone in Condon's collection, which may belong to ,this species, measures fifteen and a quarter inches in length. So far as the evidence goes it may as well have belonged to the Eschatius conidens. According to Leidy the cannon bone of the Auchenia calif ornica Leidy measures nineteen inches in length. A cannon bone of at least this size, with other bones of the skeleton, occurs in the museum of the School of Mines, and may belong to the Californian species. Whether that species is a true Auchenia or not remains uncertain, as the teeth are unknown. ESCHATIUS, gen. nov. This genus is well characterized by the reduction of ihe fourth superior premolar to a simple cone, in place of the usual double crescent charac- teristic of the Ruminantia generally. This is the greatest known reduc- tion of the premolar series in the Ruminantia, exceeding anything in the Bovidse, a family otherwise more specialized than the Camelidoe. If my 1884.1 19 [Cope. identification of mandibles be correct, there is but one inferior premolar, which is not prismatic, but has two divergent roots as in Auchenia. The crown is compressed. In any case this genus is distinct from Palauche- nia Owen, which is said to have the fourth inferior premolar composed of two crescents, somewhat as in the Protauchenia of Branco. There is also a simple conic third inferior premolar according to Owen, The type specimen of the type of Palauchenia, P. magna Owen, consists of isolated teeth put together in a bed of plaster of Paris. While there may be some uncertainty as to the position of the third premolar, I cannot agree with Professor Leidy* in the supposition that these teeth have been inverted by their describer, and really belong to the upper jaw. The specimen is preserved in the museum of the School of Mines, and I did not observe any second one. ESCHATIUS CONIDEKS, Sp. nOV. Primarily established on a superior maxillary bone, which contains all its teeth, whicli is preserved in the Museum Nacional of Mexico. I cannot distinguish from this individual another one which was found by Mr. C. H. Sternberg in the desert of Oregon, and whicli is represented by a good many fragments, including parts of both jaws. I describe the Mexican specimen first. The true molars increase rapidly in size posteriorly. The vertical ribs of the external anterior horns of the external crescents are very strong, and the external wall of the anterior crescent has a low rib on the me- dian line also. The posterior internal crescent of the last superior molar (which is not much worn) sends its anterior horn to the external wall, thus cutting ofi" the posterior horn of the anterior internal crescent. l^hQ foramen infraorUtale aw^mws issues above the anterior rib of the pos- terior crescent of the first true molar. The specimen is from Tequixquiac. The Oregon specimen includes a left maxillary and mandibular bones, with the roots or alveoli of the teeth remaining, together with numerous bones of the skeleton. As one or two teeth of the Holomeniscus hesternus are mingled with the other pieces, it becomes uncertain to which of the species some of the bones should be referred. This is the more difficult, as the superior molar teeth of the two animals are of nearly the same di- mensions. The probabilities are, however, that the greater number accom- pany the species represented by the jaws. I proceed to describe the latter. In the maxillary bone the single alveolus of the fourth premolar is close Measurements. M. Length of the four superior molars .126 .041 .024 .044 .051 " M. i t anteroposterior. Length of M. iii * Report U. S. Geolog. Survey Tei rs. I, p. 256. Cope,] 20 [May 16, to that of the anterior root of the large first true molar. Its section is a wide oval. The base of the second true molar is not longer than that of the first true molar. The external wall of the maxillary bone is broken so that the position of the infraorbital foramen cannot be positively ascer- tained. A narrow groove, which may be a part of the infraorbital canal, is exposed, and is continued forwards to a point anterior to the first pre- molar, where it probably issues. If this be a correct inference, its posi- tion is anterior to that observed in the Mexican specimen. The palatine foramen issues opposite the anterior root of the first true molar. In the Eolomeniscus hesternus this foramen issues opposite the fourth premolar's internal root. The fragment of mandible is the anterior part of the left ramus, includ- ing the premolar and half the symphysis. The fundi of the anterior alve- oli only are preserved. That of the canine is smaller than those of the in- cisor teeth, and is close to that of the external incisor. The mental fora- men is large, and is situated posterior to the mouth of the alveolus of the canine. The symphysis is not coossified. The alveolar edge of the dias- tema is narrow, and presents a narrow vertical parapet outwards, which makes an angle with the external convex side of the ramus. The inferior outline below the diastema is a little concave. The roots of the premolar are well separated. The crown is lost. The coronoid process, supposed to belong to the same species, is like that of the llama, near the condyle, and is quite elevated. It maintains its anteroposterior width to near the summit. Anterior edge rounded, the bevel extending on the external face towards its base. The posterior rotula of the condyle is median, and not on one side as in the llama and in the camel. The anterior part of the face presents forwards as in the llama, and is not so much expanded as in the camel. The petrous bone is as large as that of the camel, and has a more widely open styloid fossa, which is directed more inwards in the down- wards direction. The face also for the paroccipital process approaches much more nearly to its fundus than in either the camel or the llama. Measurements. M. Long diameter of alveolus of superior P-m. iv 009 M. i 036 M. ii 038 inferior P-m. iv 022 Length of inferior postcanine diastema 070 Depth of ramus at middle diastema 035 P-m.iv 045 It still remains to be ascertained whether this Oregon Eschatius belongs to the species that is found in the Pliocene beds of the valley of Mexico. Eschatius longirostris, sp. nov. This llama is known to me from a right mandibular ramus, which is broken off behind the last molar tooth, and which supports the symphyseal portion of the left ramus, less its external wall. In size this species is be- 1884.] 21 [Cope. twecD the Auchenia weddelU Gerv. and the Eschatins conidens, having just about the dimensions of the Camelus dromedarius or the Palaucherda magna Ow. It difters from the Eschatius conidens in the much longer inferior diastema, longer, coossified symphysis, and smaller true molar teeth ; the comparison being made with superior molars of the E. conidens. The alveolus of the inferior canine tooth is small, and is a short distance posterior to the third incisor, being separated by a short diastema. The mental foramen is very large, three times the size of that of the E. conidens, and its anterior edge is 20 mm. posterior to the canine alveolus. The alveolar parapet of the diastema is not so elevated as in E. conidens, but is distinct. The dentition shows that the animal is an old one. The fourth premolar has two divaricate roots, which spread nearly as far anteropos- teriorly as those of the first true molar. The crown is compressed. Apex broken. The crowns of the molars are worn ; that of the first to the roots. The heel of the third true molar is lost. Measurements. M. Width of mandible at inferior canines 027 Length of inferior postcanine diastema 110 " molar series 132 P-m. iv 027 M. i 029 M. ii 034 Width of " M. ii 022 Depth of ramus at middle diastema 043 P-m. iv 058 From the Oregon desert ; Professor Condon's collection. BOS Linn. Bos LATiFRONS Hai'lan. This species is represented by numerous remains, and must have been abundant in Mexico during the Pliocene epoch. On the structure of the feet in the Extinct Artiodactyla of North America. By E. D. Cope. {Read before the American Philosophical Society, August 15, I884.) The structure of the feet of a number of the Artiodactyles of the Ter- tiary beds of North America has already been described. In this paper I enumerate these, and add descriptions of some types which have been hitherto unknown. I commence with the Bunodonta. BUNODONTA. Pantolestes Cope. The structure of the tarsus only of this Eocene genus is known.* *Cope, Proceedings American Piiilosophical Society, 1881, p. 188. Pal. Bul- letin, Ho. 34. Cope. I 22 [August 15, The cuboid and navicular bones are distinct from each other and from the cuneiforms, and the ecto- and mesocunei'form are coossified. There are four metatarsals. The laterals (ii and v) are slender ; and the medians are distinct but appressed, their adjacent sides being flat- tened. This foot structure is remarkably advanced considering the early age, Wasatch Eocene, of the period of its existence, and the primitive, tritubercular bunodont character of the superior dentition. The seleno- dont types which appear first in our series of formations, the Oreodon- tidse of the White River low Miocene, present a much more primitive type of foot. The camel series is remarkable for the early and continued absence of the first and fifth metapodial bones. The first known of the line, Poebrotherium, from the White River beds, has only minute rudi- ments of them. It is probable the Pantolestes, or some member of the Pantolestidje, is an ancestor of Poebrotherium, with a number of lost types intervening. Elotherium Aym. The first information respecting the structure of the feet of this genus was furnished by Marsh.* He says "The radius and ulna were sepa- rate or very loosely united. The third and fourth metacarpals were nearly equal in size, and the second and fifth longer than the corres- ponding bones of the pes. In the latter the first digit was wanting, and the fifth rudimentary." This description leaves us in the dark as to the development of the second digit in the posterior foot and of the second and fifth in the anterior foot. The ambiguous language led me to infer that there are four digits of the anterior foot of the animal described by Marsh, and hence to separate it generically from Elotherium. The first definite information is derived from Kowalevsky, from his great memoir on the genus Anthracotherium.f He here states distinctly that the genus is bidigitate, but with small rudiments of the second and fifth metapodial bones. He shows also that the lunar is equally supported by the magnum and unciforum. In a memoir especially devoted to this genus^ he also shows that the cuboid, navicular and cuneiforms are dis- tinct, while the ecto- and mesocunei'forms are coosified, the entocunei'form being absent. The structure of the tarsus in this genus is then as in Pantolestes, and from this genus or one of the same family, Elotherium no doubt took its origin through intermediate genera. || Selenodonta. Oreodon Leidy. We owe to Leidy the following statement regarding the foot structure of this genus. § What are supposed to be the bones of the forearm and leg * American Journal Sci. Arts, 1873, p. 487. June, t PalfBontographica, 1873, p. 188, August ? X Loc. cit., xxii, N. F. II, 7, p. 415. II I have given the structure of the anterior leg and foot In Elotherium impera- tor, Bulletin U. S. Geol. Surv. Terrs., Vol v, p. 60. g Extinct Mammalia of Dakota and Nebraska, 1869, p. 72. 1884.] 23 [Cope. are discrete, as in the hog ; and the bones of the feet correspond in number with those of this animal. In 1873, Professor Marsli confirmed these statements as regards tlie metacarpal bones,* and added "that tlie navicu- lar and cuboid bones were loosely coosified or separate." In 1884f I gave a fall account of the structure of the limbs in this genus. I mentioned a peculiar feature of the carpus, viz.: that the os lunare is supported below by the inward extension of the unciform, so that the magnum is below the scaphoideum. I also showed that the cuneiforms are distinct, and that the entocunelform is wantmg. EucROTAPHUs Leidy. I have already stated that this genus is tetradactyle anteriorly and pos- teriorly.:}: I now add that the structure of the limbs and feet is in other respects like that of Oreodbn. • This is true of the inner extension of the unciform, so that the magnum is below the trapezoi'des. The inner side of the latter bone in the Eacrotaphus pacifism, is so excavated, that there was plainly a free trapezium of small size. In the posterior foot the entocunei'- form is wanting, and the mesocuneiform is distinct from the ectocuneiform. Merycochosrus Leidy. The first information of the foot structure of this genus is contained in my paper on the Oreodontidoe above cited. |I The fore and hind feet are there stated to be tetradactyle. I r^ow add that in the M.montanus Cope, the OS magnum is entirely below the scaphoid, and that there is a distinct trapezium. The posterior foot is constituted as in Eucrotaphus ; I also observe that the navicular has a peculiar little facet on its distal face near the front of the external edge. This fits a corresponding facet which forms the proximal surface of a ledge, which extends from front to rear on the inner side of the cuboid. In Eucrotaphus pacificus the arrangement is similar, excepting that the ledge of the cuboid is interrupted at the middle by a deep excavation. In Merychyus arenarum the cuboid is like that of MerycoclKBrus montanus in regard to this ledge. Merchyus Leidy. The limbs and feet in this genus are quite as in MerycochcBrus. The species which I have examined is the M. arenarum Cope. Leptomeryx Leidy. We possess as yet no information regarding the limbs and feet of this genus. It is therefore fortunate that I obtained in the White River bed of North Eastern Colorado, in 1870, a nearly entire skeleton of the L. emnsi\jQ\dj. The bones were all found close together, and belong to two individuals, and are without admixture of those of any other species. * Amer. Jour. Sci. Arts, p. 409 ; Marsh does not credit Leidy with his previous observations. t Proceeds. Amer. Phllos. Society, Pal. Bulletin, No. 38, pp. 508—10. X Loc. cit., p. 504. II Proceeds. Amer. Philos. Society, 1884, p. 501. Cope.] 2i [August 15^ From these, and inferentially from other specimens, is derived the curi- ous fact, that there are four distinct metacarpals, all supporting digits, while there are but two metatarsals, which are coosified into a cannon bone. This diversity between the limbs is unparalleled, although an approach to such a condition is seen in the peccary. In this animal, as is well known, there are four distinct digits in the nianus, while in the pes, the metatarsals are coossifled proximally, and the fifth metatarsal is re- duced to a scale. This difference between the two limbs is a further illustration of Mr. Ryder's statement that the posterior limb is in advance of the anterior in grade of development, for which I have endeavored to account by reference to the fact that it is the posterior foot which receives the greater number of impacts in progression. This is because the hind . limb is the principal propeller of the body. In accordance with the structure of the fefet, the fore-limb is much be- hind the posterior limb in the fixity of its parts. The ulna and radius are distinct ; the head of the latter a regular transverse oval. The distal extremity of the fibula is not coossifled with the tibia, but forms a sepa- rate bone, as in the Ruminantia. The lunar is mainly supported by the unciform, so much so that the front face of the magnum is not beveled to fit the former. Behind the face, the edge of the magnum is a little beveled for the lunar ; but the former bone lies almost entirely under the scaphoid. The trapezoides is coossifled with the magnum. No distinct trapezium. The cuboid and navicular are solidly united. The ecto- and mesocu- neiforms are distinct, and there is no entocuneiform. The second metatar- sal is represented by a flat oval bone which is borne on the underside of the projecting heel of the third metatarsal. The fifth is of smaller size, and is a scale imbedded in a depression of the posterior part of the side of the fourth. Ungues unilateral, trihedral and acute. Hypertragulus Cope. Remains of this genus are as abundant in the White River beds as are those of Leptomeryx, and like that genus I know but ilie one species, the H. calcaratus Cope. Unfortunately I have not been able to obtain bones of the skeleton connected with dentition from this formation, although numerous bones occur separately which probably belong to it. The genus is however abundantly represented in the John Day Miocene beds of Oregon, where Leptomeryx does not probably occur. At least no specimens of the latter are to be found in a collection of between one and two hundred individuals of this general type in my collection. I cannot distinguish the John Day species from the H. calcaratus, although the size is generally distinctly larger.* In other cases tlie size is the same. To the John Day specimens then I refer for the characters of the feet of this genus. * It is probably this species that is cited by Leidy as the Leptomeryx evansi iu the Report U. S. Geol. Survey Terrs. I, p. 216. 18SJ.1 25 LCope. The ulna and radius are coosified. The scaphoid and lunar facets of the radius are well distinguished by an oblique ridge. The carpus is un- known. The median metacarpals are separate ; whether the second and fifth are well developed I do not know, but suspect them to be so, as in Leptomeryx, since the third and fourth bear no adherent rudiments. The cuboid and navicular bones are united, while the cuneiforms are distinct from them and from each other, as in Leptomeryx. There are but two developed metatarsals, and these are distinct from each other. Thus the fore-limb in its uluo- radius exhibits a little advance over Leptomeryx ; while in the separate metatarsals it is behind the latter. Hypisodus Cope. This genus is remarkable for its prismatic dentition, being the only Artiodactyle presenting the character in the White River fauna.* It was probably well advanced in foot characters, but of these I know but little. Parts of two tarsi found with the jaws of the H. minimus Cope, are referred to the species on account of their very small size, and general cor- respondence. The cuboid and navicular are coossified. Their distal face, especially the navicular part, is so narrow transversely, that it is almost certain that the third and fourth metatarsals are coossified, and that the second and fifth are rudimental or wanting. There is no trace of facets for the latter on the naviculo-cuboid. PoflBROTHERiuM Lcidy. I have fully described the limbs of this genus in the Annual Report of the U. S. Geological Survey of the Territories for 1873f , as seen in the P. vilsoni Leidy, from the White River beds, and have confirmed them from a fine specimen of the P. steriibergi Cope, from the John Day or Middle Miocene of Oregon.:}: The characters are; ulna and radius coossified; trapezium and trapezoides present and distinct ; magnum supporting part of lunar. Two distinct metacarpals, ' scales representing the second and fifth ; navicular and cuboid bones distinct, as are the ecto- and mesocunei- forms ; entocunei'form wanting. Metatarsals two, distinct ; second and fifth represented by scales. Observations on the Phylogeny. I have maintained | that the selenodont dentition is a derivative of the bunodont, a proposition which seems unavoidable from a mechanical point of view. The testimony of palaeontology is also in its favor, since in America the oldest artiodactyle, Pantolestes, is bunodont. Kowalevsky in the phylogenetic table given in his monograph of Anthracotherium§ does * See Cope, Annual Report U. S. Geological Survey Terrs., 1873, p. 501, where the cuboid and navicular are stated to be united, t 1874, p. 499. J Bulletin U. S Geol. Survey Terrs. V, p. 59. B Journal Academy Natural Sciences. 1874. See also Ryder, The Mechanical Genesis of tooth forms, Proceeds. Academy Philada., 1879, p. 47. I 1873 (?4),p. 152. Cope.] 26 [August 15, not commit himself as to this point, but allows the development of the two types of dentition to appear to have been cotemporary and from some common origin. He then derives from such a common point of departure first, the Hyopotamidae, which first appear in the Eocene, and second, the ancestors of the Anoplotheriidae. From the Hyopotamidae he derives all the modern Selenodonta, exclusive of the Camelidae. The latter group he omits from his table, doubtless because his information on the subject was insufficient. The main line of origin of the Selenodonta is divided early in Miocene time, the genus Gelocus giving origin to the Pecora, and the genus Hyoemoschus to the Tragulina. In describing the characters of the genus Poebrotherium for the first time, I remarked as follows :* "The present genus is a Inore generalized type than Gelocus, and in its distinct trapezoid and distinct metacarpals represents an early stage in the developmental history of that genus. It also presents affinity to an earlier type than the Tragulidae which some- times have the divided metacarpals, but the trapezoides and magnum co- ossified. In fact Poebrotherium as direct ancestor of the camels, indicates that the existing Ruminantia were derived from three lines represented by the genera Gelocus for the typical forms, Poebrotherium for the camels and Hyaemoschus for the Tragulidae." These views being then established on sufficient evidence, it remains to make such additions as the facts cited in the present paper indicate. First in importance comes the place in the phylogeny of the Selenodonta, of the Oreodontidae. The peculiar inward extension of the unciform bone already ascribed to them, characterizes also among extinct forms the genus Leptomeryx, and probably Hypertragulus. Among recent ruminants it is only seen in the Tragulidae. f If we arrange these types in serial order we find the modifications of form to be generally identical with those of the other ruminant lines, in the coossifi cation of the bones of the legs and feet. This series may then be regarded as phylogenetic. The peculiar structure of the carpus of the Oreodontidae, puts them out of the question as ancestors of any type of existing ruminants other than the Tragulina. Whether they themselves can be traced to a five lobed, or to a four-lobed bunodont ancestor, remains an undecided question. It is not, however, probable that a five-lobed form has been intercalated in a series, both of whose extremities are four-lobed. If this be true, the Oreodontidae must be regarded as an ancestral type of Selenodonta, coequal with the Hyopo- tamidae, and it may well be questioned whether the latter can have been ancestors of the existing Ruminantia, whose molars are four-lobed. So the present investigation does not disclose the ancestral stock of the Pecora. In North America we have not progressed further in the solu- tion of this question than I reached in 1877,:): after a study of the genera * Bulletin U. S. Geol. Survey Terrs. Vol. 1, No. 1. p. 26, Jan., 1874. t Among Perissodactyles it occurs in Triplopus, Tapirus and the Rhinoceron- tidse. t Proceedings Amer. Philos. Soc. p. 228. 188 1.] 27 [Cope. Cosoryx Leidy, and Blastomeryx Cope. Iliad already* suggested that the former genus is the ancestor of the Cervidae, but subsequentlyf re- marked : " It is not probable this genus is the immediate ancestor of Cervus. from the fact that the molar teeth display in their prismatic form a higher degree of specialization than belongs to that genus. It is probable that the true ancestor combined the dental type of Cervus with the distinct roots and short crowns of the molars, with the type of horns here de- scribed." I at that time included a species {Cosoryx gemmifer Cope) in the genus, provisionally, which has the type of molars in question. Having discovered another, larger species, which has the same type of molars, I at once distinguished the provisional group in which I had placed the G. gemmifer, Blastomeryx, as a genus ; and in describing the species {B. borealis) observed as follows : "In brief, its molars differ from those of Cosoryx ("Dicrocerus") much as those of the deer differ from the molars of the antelope. While Cosoryx ("Dicrocerus") was probably the ancestor of Antilocapra, Blastomeryx was the ancestor of Cervus or Cariacus." This opinion expresses all the information I possess on the subject at present. It remains to ascertain ' the structure of the anterior feet in Hypisodus, which is the earliest genus of Ruminantia known to have prismatic molars. The following table will represent the views expressed in the preceding pages : ? Bovidse. Tragulidse. Camelidee. Hyopotamidae. Oreodontidse. Poebrotheriidse. * Proceedings Academy Philadelphia, 1874, p. 149. t Report Expl. Surv. W. of 100th Mer. U. S., G. M. Wheeler in charge, iv, pt. ii, Selenodonta. Quadritubercular Buno- /^donta. Tritubercular Bunodonta. (Pantolestidse. ) p. 349, 1877. Cope.] 28 [August 15, Fifth Contribution to the Knowledge of the Fauna of ilie Permian Formation of Texas and the Indian Territory. By E. D. Cope.* {Read before the American Philosophical Society, August 15, I884..) PISCES. CERA.TODUS FAVOSUS, Sp. DOV. This species is known to me from a piece of tlie lower jaw, which sup- ports a tooth. One extremity of the tootli is broken off, but from the cur- vature of its inner side, it is to be inferred that tlie portion lost is but small, probably including one of the three processes which the tooth possesses. The species may -be distinguished from those described by Agassiz, and from the existing species, by the great depth of the two emarginations of the external side. These enter the crovrn so deeply as reduce its width to dimensions no greater than those of each of the processes of the crown. The internal face is strongly convex, and one extremity is more strongly recurved than the other. The crown consists of a mass of coarse perpen- dicular simple calciferous tubules, which are enclosed in a rather thin layer of a dense substance which thickens downwards, and laps over the external ftice of the jaw bone. The external surface of this layer is vitreous. The walls of the tubules are of a dense and hard substance, of a darker color in the fossil, and the tubules are filled with a softer sub- stance, so that the grinding surface of the crown has the appearance of a small honeycomb. The diameter of the tubules ranges from 1. to .05 mm. The fragment of jaw is robust, is deeper than wide, and is strongly con- vex on the internal face. The internal inferior angle rises atone extremity above the level of the external inferior angle. The processes of the crown project freely beyond the bone, having rested on the cartilage which forms the external face of the jaw, as Giinther has shown to be the case in the G. forsttri. Besides the deep emarginations of the crown, the coarseness of the cal- ciferous tubules is a special character of this species. Measurements. M. Depth of jaw with tooth 019 " " without tooth 012 Width of crown at middle process 014 Probable length of crown 022 Found by Mr. W. F. Cummins. Agassiz did not record any species of this genus from below the Trias, but Fritsch has reported them from the Permian of Bohemia. BATRACHIA. Cricotus crassidiscus, sp. nov. Accession of additional material enables me to add several points to the * The "Fourth Contribution" will be found at page 628 of these Proceedings for. the year 18i3. V im.] 29 [Cope. knowledge of the osteology of this genus, and to distinguish satisfactorily three species. I have much pleasure in obtaining these additional facts, since everything relating to this curious genus is of interest. In the first place, the neural arches are not coossified to the centra, but are readily separated from them. Their basis of attachment forms, on each side of the median neural canal, an oblique triangular surface look- ing forwards and upwards, with the apex above and behind. The ease with which the neural arches separate accounts for the rarity of their occurrence on separate centra. They support the diapophyses at their lower border. Second, that the sacrum consists only of a centrum and an intercentrum, both of which take part in furnishing a concave facet for the attachment of the pelvis. Third, some of the ribs are two-headed, and their capitular articulation is with the posterior edge of the intercen- trum. Fourth, there is a hyposplienal articulation, as in the genera of Jurassic Saurians, Camarasaurus, Amphicoelias, etc., and in the Permian genus Empedias, among the Tlieromorpha. The hypantrum has, however, this peculiarity : that its sides are produced forwards into a process on each side below the prezygapophyses, each of which is subconical in form, but with the interior face excavated to receive the hyposphen, so that the sec- tion of the process is crescentic. These processes I have never previously observed. I call them hypantrapopliyses. I find them in the Gricotus hypantricus. The neural arches of the other species are either lost or in such close juxtaposition that I cannot see them. The species diflfer in part as follows ; the full characters can only be given in more detailed descriptions of more perfect specimens. I. Dorsal intercentra much narrowed or pinched above. Hypantrum unknown 0. heterocUtits. II. Dorsal intercentra equally robust above as below, or more so. Hypantrum unknown C. crassidiscus. Hypantrum with acute lateral processes C. hypantricus. The measurements of the C. crassidiscils are as follows : Measurements. Diameters of dorsal centrum behind vertical. . . transverse M. .025 .025 .013 .013 .010 .009 .025 .025 .009 .0095 .027 .029 .010 Base of neural Diameters of a dorsal intercentrum transverse. Diameters of coracoid transverse length j at glenoid face., ( at internal face. Cope.] 30 [August 15, It is probably this species which I have figured in the Proceedings of the American Philosophical Society,* and American Naturalist, f under the name of C. heteroditus. It is the most abundantly represented in my collection. In the specimen figured in the American Naturalist, the probable scapula is visible on both sides, but the coracoid is concealed by the pectoral scuta. CmcOTus HYPANTRicus, sp. nov. This Embolomere is probably represented by two individuals, which are of larger size than any species which have hitherto come under my no- tice, one of them very much larger. It is only the smaller specimen which is accompanied by the astragalus. Both of them display the hypan- trapophyses already mentioned in remarks on the genus under the head of G. crassidicus. As already pointed out in the key of species, the dorsal intercentra in the G. hypantricus are stout and not narrow above, but the thickness in- creases rather than diminishes upwards. They thus differ from the cor- responding intercentra in the G. heteroditus. In many of the dorsal in- tercentra the dense external layer which covers the inferior face continues upwards to an apex, the articular surfaces of the two ends meeting so as to exclude the former. This is also the case in the G. crassidiscus. The centra have the abbreviated form characteristic of the genus, and the for- amen chordae dorsalis is present, but is smaller than the 0. Jietervditus. The supposed astragalus is oblong ; proximal J border longer than the distal, which is separated by an obtuse angle from the ectad ; distal en- tad not reaching superior surface of bone, long, extending inwards below the revolute proximal part of the entad face, from which it is separated by a narrow oblique groove. Proximal and distal entad separated by notches of the two faces ; a ridge the length of the bone below. KEPTILIA. Clepsydrops leptocephalus, sp. nov. This species is represented by almost tlie entire skeleton, the principal deficiency bemg that of the scapular arch and the anterior limbs, with the * 1881, pi, li, figs a-b. t 1884, p. 39, pi. v and fig. 7. In pi. v, figs / and g represent the C. heteroditus. X I determine the ends of this specimen from a foot of Eryops. 3Ieasureinents. Diameters of centrum of indi- j anterposterior vidual with astragalus. 1 transverse . . . M. .015 028 .038 .029 .018 .038 Diameters adjacent intercentrum of do. anteroposterior .013 transverse 038 1881] 31 [Cope. phalanges of the posterior feet. The bones of the skull are mostly pre- served, but in a dislocated condition. They serve to demonstrate some of the characters of the genus and family. The quadrate bones of both sides are distinctly displayed. They are rather short, and articulate above by squamosal suture with the squamosal bones, which overlap them posteriorly. They narrow upwards, and are deeply grooved on the anterior face below. Each edge of the groove is produced forwards ; the external lor a considerable distance as an acumi- nate laminiform process, in the usual position of a quadratojugal bone. The production of the internal edge is shorter, and its extremity is verti- cally truncate. Its superior edge fits an incurvature of the superior edge of the pterygoid bone, and its internal face is applied to the external face of the latter. The pterygoid bone displays the subtriangular plate with dentigerous edges, such as I have already described as present in the species of Dime- trodon. In this species it is thinner and less massive than in any species of that genus yet known. This specimen enables me to locate it more precisely than heretofore. The pterygoids were probably placed much as I have represented them to be, in the Bmpedias molaris Cope (Proceedings American Philosoph. Society, vol. xix, p. 56, pi. v). They send inwards a subtriangular plate from each side, which approach each other on the median line without touching, and the adjacent edges are somewhat de- curved. The posterior edges are deeply concave on each side of the mid- dle line, and like the inferior edges, are dentigerous. The process for the quadrate extends outwards and backwards, and is thickened on its poste- rior edge, while its anterior edge, which is continued from the inferior edge of the posterior border, becomes very thin. The anterior production for the ectopterygoids extends outwards and forwards, leaving the anterior edge of the dentigerous plates as the concave posterior border of the large palatine foramina. The anterior production of the internal edge of the plate becomes very thin, and is broken in the specimen without showing articulation for the palatine. The squamosal extends both above and below its anteriorly directed zygomatic portion. The superior extremity shows squamosal suture for the parietal. The stapes is of large size. It consists of a stout rod terminating in a double extremity, something like the double head of a rib. The shorter head is expanded into a funnel shape. Near to it the shaft is perforated in the longer diameter by a foramen. The extremity of the other head is transversely truncate and is separated from the funnel by a deep notch. On the outer side of the fundus of this notch, a foramen penetrates the shaft obliquely and is continued into a canal which issues at the foramen first described. The distal end is truncated by an irregular sutural sur- face. In the specimen the bone lies behind the squamosal and quadrate bones, the simple extremity of the rod near the posterior edge of the quadrate. Cope.] 32 [August 15, The premaxillary bones are distinct. The teeth of that bone and of the maxillary are of unequal sizes. The axis has an expanded neural spine, and a diapophysis for rib articu- lation, but no parapopliysis or capitular fossa. The two latter features characterize all the vertebrae which follow, as far as the lumbar series. The column in the typical specimen is tolerably complete, with a break of uncertain, but probably not great length in front of the sacrum, and the loss of the distal part of the caudal series. Intercentra of rather small size are present throughout the series anterior to the sacrum. The inferior faces of the caudal vertebrae are yet concealed by matrix. The bases of the neural spines are compressed ; they were probably not elongate as in Dimetrodon, though they are unfortunately broken off, except that of the third cervicodorsal vertebra. Here the spine is short and truncate above, and rather wide anteroposteriorly. As in Dimetrodon there is no distinc- tion between cervical and dorsal vertebrce. The pelvis is well preserved, and has the characters already assigned to the G. natalis Cope.* The ilium has a process or narrowed continuation Avith parallel sides, directed backwards and upwards, and bearing a keel on the middle line on the internal side. The ischia are much produced posteriorly, and are separated by a notch on the middle line posteriorly. The head of the femur is expanded, including probably the homologue of the great trochanter of mammalia, and its arti'^ular face is crescentic, with obtuse horns. There is a trochanter below it on the posterior edge of the shaft. The condyles are inferior, and are separated by a deep groove above and a shallow one below. The articular faces of the two condyles are continuous, forming and oo -shaped figure. The proximal ex- tremity of the tibia is wider than the distal, and the articular face is unin- terrupted. That of the distal extremity is a transverse oval. Specific characters. While the vertebral centra of this species are rather short, the bones of the head are very much attenuated, and the jaws are long and slender. None of the four jaws is prefectly preserved, but the number of the teeth in the maxillary bone may be approximately fixed at thirty in a continuous series. One, and probably two of these, placed near the anterior part of the series, are larger than the others. They are placed at the position of the corresponding large maxillary teeth in Dimetrodon, but they do not display the dimensions seen in the species of that genus. To strengthen the jaw at this point, a rib rises from the thickened alveolar portion, and extends vertically on the inner side of the thin facial plate o# the bone. The facial plate is double, and each lamina, except at the rib, is not thicker than wrapping paper. The premaxillar}-- bones are robust, and are excavated postero-laterally for a very large nostril on each side. The spine is long. The alveolar edge bears five teeth, which are followed by a diastema. These diminish in size posteriorly, the first one being the largest, and equaling the large * Proceeds. Amer. Philos. Society, 1878, p. 509. 1881] 33 [Cope. maxillary teeth. The last two are quite small, less than the usual maxil- lary teeth. The dentary bones are very slender, and the distal end is somewhat thick- ened lo support two teeth larger than the others. These are the third and fourth from the extremity, and are not quite so large as the large teeth of the maxillary bone. The remaining mandibular teeth are small, and are not so much compressed as in the species of Dimetrodon. Many of them have only a posterior cutting edge, which is not denticulate. The apices are strongly turned backwards in the posterior part of the series. The posterior part of the dentary bone rises and carries some of the teeth with it. The surface of the free edge of the internal plate of the pterygoid bone is granular. The teeth on the posterior edge of the same are sub- conic, and in a single series. There are twenty-seven vertebrae in a continuous series, from and inclu- ding the axis. All bear diapophyses, and all are rib-bearing, except per- haps the last two, where they are of reduced size. They are more or less opposite the neural canal as far as the twenty-second centrum. On this vertebra the superior edge is on a level with floor of the canal, and pos- terior to this point the diapophyses rise from the centrum. Two sacral s and ten caudalsare preserved, and all have diapophyses and neural spines. The centra in this species are rather short, being as deep as long through- out the series, if measured at the middle. The edges are not undulate as in G. (Embolop/iorus) limbatus Cope. The intercentra are short and not extended upwards on the sides as in that species. Measurements. M. Length of quadrate bone 085 Width of condyle of quadrate bone (greatest) 037 Length from condyle of internal anterior process of do. .032 " external " " " .097 " of squamosal bone (vertical) 124 *' " pterygoid from palatal foramen 116 Width " " at middle 090 Length " internal dentigerous edge of do 070 " " posterior " " 051 " " maxillary bone posterior to canine brace 181 Thickness of " " at canine brace 020 Depth of " " nostril 016 Length of premaxillary bone (posterior apex restored).. .060 Width " " " atthirdtooth 022 Diameter of large (first) premaxillary tooth 008 " " " maxillary tooth (canine) 009 "small " " 006 Length of crown of last maxillary tooth 009 " " twenty-seven continuous cervico-dorsal ver- tebrae 855 Cope.] [August 15, Measurements. M. Length of two sacrals 65 " " ten caudals 260 (anteroposterior 034 vertical posteriorly 031 transverse posteriorly 030 Elevation of neural spine from centrum 071 arch " " 009 Width of postzygapophyses 030 Elevation of neural spine of fourth vertebra 058 Diameters ceo trum sixteenth vertebra ^ ^^^^.^«P«^^^^'«^ '^^^ vertical at end.. .035 Diameters end centrum seventeenth cen- f vertical 034 trum I transverse. .030 Expanse of postzygapophyses of seventeeth vertebra... .029 Diameters twentieth centrum vertical at end 031 anteroposterior 027 Diameters of twenty-ninth centrum ^ ^^^^^'^P^^^^"^^' ' • "^^^ ( transverse behind .035 Expanse of postzygapophyses of twenty-ninth vertebra. .024 Width of sacrum through fixed diapophyses 049 / anteroposterior . . .025 Diameters centrum twentieth caudal } vertical behind. . . .0265 C transverse " . . .0225 Expanse through diapophyses 047 Elevation of prezygapophyses (greatest) 039 / anteroposterior (apex of pubis re- Diameters of pelvis } stored) 235 ' vertical through acetabulum 123 Anteroposterior diameter of ilium at acetabulum 089 Depth of ischium at posterior edge of acetabulum 080 Length of ** from acetabulum 117 Length of femur. 179 Proximal diameter of fem„, | anteroposterior 075 t transverse (at middle) 0::iO transverse 038 anteroposterior 031 transverse 068 r external condyle. .031 Diameters shaft at middle Diameters of distal pnd ] anteroposteriors . I ^ I internal " .045 Length of tibia 150 froximal \ anteroposterior(middle) .040 ' * \ transverse 057 median anteroposterior. 019 ... .026 transverse 041 distal \ anteroposterior, 1884.] 35 [Cope. The typical specimen of this species was found by Mr. W. F. Cummins in the Permian beds of Northern Texas. CLEPSYDROrS MACROSPONDYLUS, sp. nOV. This species, like the last, much exceeds the C. natalis in dimensions. The bases of the neural spines are enlarged, so that it is probable that the spines were not elongate as in the species of Dimetrodon. Intercentra are present throughout the dorsal and caudal series of vertebrae. The den- tar}-- bone supports one or two large teeth near the extremity. These characters furnish the reasons for referring the species to the genus Clepsydrops. The individual by which the species is known, is represented by an axis vertebra, twelve continuous dorsal vertebrae ; nine other continuous verte- brae, of wliich three are lumbar, two sacral, and four caudal. Also by a part of the ilium, and by the greater part of a dentary bone. All of these speci- mens were found together, and possess an identical mineral appearance. That this reptile belongs to a distinct species from the C. leptocephalus is readily determined by the form of the dorsal vertebrae. The centra are a little longer than those of that species, but have a smaller vertical diam- eter. The latter is three-fifths of the former, while in the C. leptocepha- lus the two dimensions are reversed, the depth being a little in excess in corresponding parts ot the column. The dentary bone, on the contrary, is more robust than that of the C. leptocephalus, and supports, probably, a small number of teeth. The edges of the centra are not undulate or laterally flared. The cen- tra are strongly compressed, and in the anterior part of the column have an obtuse hypopophysial keel. The intercentra display equal width of the infei'ior surface; and are abruptly rounded at the extremities. The last one preserved is between the second and third caudal centra. It is shorter and M ider than the others, and does not display any trace of a chevron bone. The diapophyses are opposite the neural canal on the thirteen an- terior vertebrae preserved. Each one sends a horizontal rib forwards to the prezygapophysis, and another obliquely forwards and downwards which stops short of the edge of the centrum. These ribs enclose a fossa in front of the diapophysis. Posteriorly the anteroinferior rib grows more robust, and evidently supports part of the tuberculum of the rib. There is no facet for the capitulum until the antepenultimate vertebra of the anterior series is reached. Here and on the penultimate the anterior border is flattened into a facet, and on the last of the series, the facet marks the summit of a distinct tuberosity, which is produced by the cut- ting away of the border below it, to accommodate the intercentrum. The three lumbar vertebrae preserved are different from the dorsals in their greater abbreviation. This character is not unknown in other species of Pelycosauria. The centrum is contracted, but not compressed, at the middle. The diapophysis is altogether on the centrum, and supports no rib-facet. Its anteroinferior buttress is well developed, extending to the Cope.] 36 [August 15, margin of the centrum whicli is cut out below it for the intercentrum. The sacrum is rather robust. Its two vertebrae are not coossified, and support well developed neural spines, and a large free diapophysis for the ilium. The centra of the caudals, and their diapopliyses and neural spines are well developed. There is a fossa at the base of the spine on each side, in line with the zygapophysial surfaces, equidistant between them. The fragment of ilium is of appropriate size, and is quite robust. It displays the fossa for the sacral diapopliysis, and the acetabulum. The latter is remarkable for the prominence of the tuberosity on the superior border, which exceeds that of any species of Pelycosaurian known to me. The section of the ilium through it is triangular. The dentary bone is accompanied by the splenial to the middle of the symphysis. The latter is not very long. Its dentary portion turns up- wards. The ramus is quite robust, differing much from that of the 0. leptocephalus. It is broken off a little anterior to the tooth line, but the latter probably did not contain more than twenty-two teeth. These have anterior and posterior cutting edges, and are denticulate. The external face of the dentary is excavated by shallow, undulating, branching grooves. Measurements. _ M. Total length of vertebrie preserved 640 C anteroposterior 031 Diameters centrum of a f vertical behind diapophysis. . .019 dorsal vertebra [transverse/ t at middle 0115 Diameters neural arch / length with zygapophyses 041 of same vertebra I width at prezygapophyses 022 Diameters neural spine f anteroposterior 0145 of same vertebra. . . . ^ transverse behind 007 Diameter of intercen- / anteroposterior 0052 trum of do I transverse 023 ' anteroposterior 024 transverse at end 026 Diameters of a lumbar centrum ^ " " middle.. .023 j vertical behind arch. . .022 [ " at end 029 Length of sacrum 055 r anteroposterior . . . .024 Diameters of third caudal vertebra J vertical at end 023 [ transverse at end.. .022 Anteroposterior diameter of acetabulum 0325 Transverse diameter of ilium at tuberosity 0265 Length of dentary bone supporting twenty teeth 044 Thickness at twentieth tooth 0175 Depth ramus at second tooth 035 " " fifteenth tooth 039 183J.] 37 [Cope. The bones of this specimen are in excellent preservation. They were recovered by Mr. W. F. Cummins from the Peruvian beds of Texas. Edaphosaurus microdus, sp. nov. The genus Edaphosaurus Cope, was established on the E. pogonias Cope (Proceed. Amer. Philos. Soc, 1882, p. 448), which is represented by a specimen, which includes only a distorted cranium, with most of the parts preserved. The present species is represented by an individual of which I possess numerous vertebrae and ribs, and the dentigerous plates of both jaws. These are part of the dentary splenial in the inferior jaw, and the pterygoid or palatine of the superior. The specimen enables me to determine the characters of part of the vertebral column in the genus Edaphosaurus. In the first place the vertebrae possess enormously elongate neural spines, as in Dimetrodon. Next, the centra have a facet on the anterior edge above the middle for the head of the rib, as in a mammal. It is not re- peated on the posterior edge of any of the thirteen centra preserved. Thirdly, the ribs are only compressed proximally. Distally their section is a wide oval. The extremity is truncate and concave. The shaft is hol- low, the walls being thinnest distally. Specific characters. The grinding teeth of this species are about as numerous as in the E. pogonias, there being about seven in a transverse row on each plate. They are, however, less closely placed than in the typical species, and have more conic crowns. They do not form a pavement, as they are separated by wider interspaces. The centra are rather elongate, and the foramen chordce dorsalis is rather large. No intercentra are preserved, and if present they must have been very small, as the inferior rim of the centrum is not beveled to receive one. The neural spines have transverse processes which commence near the base, and project at intervals from the sides. The inferior ones are oval or subround in section ; those which succeed are more or less com- pressed. The extremities are enlarged fore and aft so as to be claviform in outline, but are compressed except where thickened by lateral tuberosi- ties. These are rarely symmetrical, one being larger and situated higher up, sometimes giving the apex an unsymmetrically bilobate form. Some- times they project at right angles to the terminal expansion. The shaft of the spine has a rather small medullary cavity, and this issues by an open mouth at the summit of the apex without constriction. This pecu- liar arrangement suggests a cartilaginous continuation of the spine which retains the nutritive artery of the medullary cavity. The anterior face of the shaft is grooved from the base for some distance upwards ; the posterior face is plane and then rounded above. Measurements. M. Diameters of inferior dental patch | ^^^^^^P^^^^''^^^ transverse 024 Cope.] 38 [August 15, Measurements. M. 'J dorsal centrum 1 I transverse ^. „ . f anteroposterior 0335 Diameters of a posterior j ^^^^.^^^ at end 026 at middle 015 Measurements of piece f length 133 of spine of same... J. diameter f anteroposterior 023 [at base. Uransverse 019 r vertical..!^* end 032 Diameters of median i " behind arch 025 anteroposterior 0465 dorsal , , at end, at flare 037 transverse i at middle 016 { Diameters of summit of ^^^^le [^^^^^^Vo^^^nor 032 Uransverse 032 The ramous character of the neural spines of this species is much like what is seen in the Dimetrodon crudger Cope. The rami in this species, however, retain their size upwards, and become compressed, a feature not seen in the D. crudger. The apices of the spines in the latter species are not dilated as in the E. microdns. Found by W. F. Cummins in the Permian beds of Texas. The posterior foot in Pelycosauria. — The foot-bones of the rep- tiles of the suborder Pelycosauria are abundant in the collections from the Permian formation, and I have examined my collection for specimens in which they are in normal connection, for the purpose of identifying them. I have been so fortunate as to find an entire tarsus, with the proximal parts of the metatarsi, in the skeleton which served as the type of my description of Clepsydrops natalis.^ The characters presented by this foot are no doubt present in all of the Clepsydropidae, which in- cludes the genera Theropleura, Dimetrodon, Embolophorus, and proba- bly others. Tarsal bones identical with those of the C. natalis were found with the original specimens of C. colleiiii and others of much larger size, accompany remains of species of Dimetrodon, or Embolophorus. The astragalus and calcaneum are large and well specialized bones, dis- tinct from each other and from the other tarsal elements. They do not resemble the corresponding bones of any known type of vertebrate, as will presently appear. The navicular bone is distinct, and the cuboid apparently consists of a single element. This depends on the interpreta tion given to a small bone on its posterior face, which is broken on its free edge, and maybe the head of the fifth metatarsus. There are three ele- ments in contact with the distal face of the navicular, which correspond with the thre6 mammalian cuneiforms. The space available for this con- tact seems hardly sufficient for the three elements present, one of which is out of position and on the inferior side of the carpus. This element * Proceedings American Philosopli. Society, 1879, 509. 1S84.] 39 [Cope. looks also from its free inferior side like an ungual phalange, but is flatter than is characteristic of this family. There are three metatarsals distal to the navicular, which are well accommodated with articular facets on the distal extremities of the three banes in question, so that their identifica- tion as the three cuneiforms, is probably necessary. The two remaining metatarsals are articulated, tlie fourth to the exterodistal facet of the cuboid ; and the fifth to the exterior side of the cuboid. The third, fourth and fifth metatarsals are directed at an obtuse angle posteriorly from the long axis of the astragalus. This structure is more mammalian than any form of foot yet known among reptiles, and agrees with the indications of mammalian character described as existing in the long bones of the limbs by Owen and by myself. The astragalus is an oblong bone with one long straight side, viz., that which is in contact with the calcaneum. This side has two facets for articulation with the calcaneum, which are separated by a groove, which forms a foramen when the two bones are in place. The proximal ex- tremity of the bone is much smiller than the distal, and is subround. The proximal half of the bone would be nearly cylindric were it not for the truncation caused by the calcaneal facet. The distal half of the bone is robust, and is surrounded on all sides by facets. These are the external or calcaneal, the distal or navicular, and the internal which is larger than the other two together. The first two are oblong and truncate, the navic- ular twice as large as the calcaneal, its transvere much exceeding its anteroposterior diameter. The internal facet already mentioned, covers the internal face of the distal half of the astragalus, which projects fur- ther inwards than the proximal half, rising abruptly from it. The facet is continuous with the navicular, and is at right angles to its plane. It widens proximally, and its proximal border is deeply notched. Its surface is convex from back to front, but not strongly so. In the astragalus of a species of Dimetrodon, it is divided by an angle into two facets, the two faces thus produced being nearly at right angles to each other. This in- ferior part of the facet is continued into a prominent border which is more or less roughened. A rounded tuberosity of the inferior face of the bone occupies the space between this border and the calcaneal border, so approaching the notch already described, as to cause a groove to proceed from it posteriorly and inwards. I described the corresponding bone in the Glepsy drops colUttii (Proceeds. Phila. Academy, 1875, p. 409) as a possible coracoid. The calcaneum has its postero-external edge broken in the specimen of Glepsydrops natalis described, but is probably a semidiscoid bone, with its straight margin applied to the astragalus. This margin presents a median flat elongate-oval facet, which is separated by grooves from a facet at each end. The proximal facet is the narrower, and passes by a curve into the proximal extremital facet, which is adjacent to the corresponding proximal facet of the astragalus. The distal internal facet is triangular and wider / Cope.] 40 [August 15, than long, and is separated by an angle only from the distal facet. The latter is a little more than a half circle in outline, and joins one bone of the second row, which I suppose to be the cuboid. The fact that it does not articulate with the second element in that row, leads me to suspect that the latter is the head of a fifth metatarsal. The external edge of the bone thins out more rapidly at the distal than at the proximal extremity. The cuboid bone is pentagonal in outline, and square in transvere sec- tion. It is not unlike that of the Amblypodous Mammalia. It has a transverse proximal facet, and two distal ones which meet at an angle about right. The fifth metatarsal is articulated with its posterior face ; and the fourth with the exterior distal face. The ectocuneiform articulates with the interior distal face. The navicular bone is subtriangular in trans- verse section, and with a subquadrate base articulating with the cuboid. Its longitudinal and anteroposterior diameters are about equal. The distal or metatarsal articulation of the entocuneiform is transverse and flat. The manner of articulation of the ankle-joint must have been different from the usual reptilian type. The proximal extremities of the astragalus and calcaneum combined are not too large to have received the distal ex- tremity of the fibula, so that the tibial articulation must be sought else- where. This may have been on the large distal facet of the anterior or inner face of the bone. A part of this facet looks upwards and probably supported the tibia, which was thus removed by a short space from that of the fibula. The down-looking part of the facet, which is more distinct in Embolophorus, must have articulated with a separate element. This may have been a spur, such as exists in the known genera of the Monotremata ; as the position is identical with that which bears this appendage in those animals. It is quite evident that an element additional to those known in the ordinary reptilian foot exists in the Clepsydropidae. The separation of the distal extremities of the tibia and fibula is not usual among reptiles, but it is common in the salamanders, where the os centrale comes between them. It is also evident that the subcylindric proximal part of the astragalus, which intervenes between the supposed tibial and fibular articulations, represents that bone. The metatarsals are directed obliquely backwards as well as outwards, as in Tachyglossus and Platypus. The following results may be derived from the preceding statements : (1) The relations and number of the bones of the posterior foot are those of the Mammalia much more than those of the Reptilia. (2) The rela- tions of the astragalus and calcaneum to each other are as in the Mono- treme Platypus anatinus. (3) The articulation of the fibula with both calcaneum and astragalus is as in the Monotreme order of mammals. (4) The separate articulation of the anterior part of the astragalus with the tibia is as in the same order. (5) The presence of a facet for an articu- lation of a spur is as in the same order. (6) The posterior-exterior direc- tion of the digits is as in the known species of Monotremata. Thus the characters of the posterior foot of the Pelycosauria confirm the 41 [Cope. evidences of Monotreine affinity observed by Professor Owen and myself in the bones of the legs, especially of the anterior leg. It remains a fact that with this resemblance in the leg there is a general adherence to the reptilian type in the structure of the skull. But this adherence is not so exclusive as has been supposed, as I will now endeavor to show. The structure of the columella auris in Clepsydrops leptoceph- ALUS. — As already briefly described above, this element is bifurcate at the proximal extremity. The shorter expanded extremity is the stapes proper. The oblique perforation of its base is a character which has not been hitherto observed in any reptile, not even in the allied form Hatteria (Huxley). If, as is probable, the perforation is homologous with the foramen of the mammalian stapes, we have here another point of resemblance to this class. The longer proximal branch of the columella has only half the width of the stapedial portion, and its long axis makes an obtuse angle with that of the latter. It is perhaps the ossified suprastapedial cartilage of Huxley, which that author states (Anatomy of Verteb rated Animals, p. 77) is not ossified in any of the living Sauropsida, Huxley supposes this cartilage to be the homologue of the incus, and remarks * that in a young Mammalian foetus "it appears exactly as if the incus were the proximal end of the cartilage of the first visceral arch." The columella now described resembles a rrb, of which the suprastapedial process resem- bles the head, and the stape-; the tubercle. If this process be the incus, the stapes is shortened as in the majority of Mammalia, unless the primi- tive suture between the two be longitudinal. The form and position of the true stapes give support to the view of Salensky, that it is not part of * true visceral arch, but is developed in the connective tissue surround- ing the mandibular artery. We see that in this Pelycosaurian it is not the proximal part of the arch, and surrounds the mandibular artery. The columella is divided into at least two distinct elements. This is clearly indicated by its abrupt truncation distally by a rough sutu- ral surface. If there is but one bone distad to the stapes, it is homol- ogous with the cartilage, which has been shown by Peters f to be distinct in Hatteria, crocodiles and various lizards. It is the trian- gular ligament of Cuvier. If the suprastapedial be incus, this ele- ment is malleus; and it is usually identified as such by the older anat- omists. In this structure we have evidence that the hypothesis that the articular and quadrate bones are homologous with the ossicula auditus is incorrect. The Pelycosauria will probably come under the head of "Sauropsides malleoferes" of Albrecht. We have here an approximation to the Mammalia in tvvo points : (1) The perforation of the head of the stapes ; (2) and the ossification of the incus, w^hich (3) is distinct from the malleus, thus furnishing homologues of the principal ossicles of the ear. It * Proceedings Zool. Society, London, 1869, p. 391. t Monatsberichte der Academic Sciences, Berlin 1868 (p. 592)— 1870. Cope.] 42 [August 15, is unnecessary to observe however, that this part of the skeleton does not resemble the corresponding part in the known Monotremes. Structure of the quadrate bone in the genus Clepsydrops. — The quadrate bone in Clepsydrops leptocephalus Cope, already described, is of highly interesting form. It consists of two portions, a vertical and a transverse, the latter much longer than the former. The vertical portion is wedge-shaped with the base fashioned into the condyle for the mandib- ular ramus. Its posterior face to the apex, is articulated with the large squamosal, which rises towards the parietal bone. The distal part of the quadrate is grooved anteriorly, and each edge sends a process forwards. The internal is short, and articulates with the pterygoid. The external is the long horizontal part of the bone already mentioned. It is compressed, and at the end is acuminate. Although the malar bone is out of place in the specimen described, examination of the skull of the Clepsydrops natalis, where it is preserved in position, shows that this horizontal ramus of the quadrate is nothing more than the zygomatic process of the squamosal bone of the Mammalia, forming with the malar bone the zygomatic arch. In the Pelycosauria there is hut one posterior lateral arch, as is demonstrated by many specimens ; hence, we have here a reptile with a zygomatic arch attached to the distal extremity of the quadrate bone. Important results follow this determination^. We have seen that, with Peters, we need no longer look to the auricular chain of ossicles, and especially to the incus, to And the holnologue of the os quadratum of the Yertebrata below the Mammalia. According to Albrecht the os quadratum is the homologue of the zygomatic portion of the squamosal hone. If this be true, in the process of specialization of the reptiles, the anterior or zygo- matic portion of the quadrate has been lost or separated as a quadrato- jugal bone, and the condylar portion extended, until it has reached the extreme length we observe in snakes. This determination of the character of the quadrate bone in the Theromorphous Reptilia is confirmatory of the theory broached by Albrecht.* Among many propositions novel to the science of osteology, none has been more unexpected than his assertion that the quadrate bone is the homologue of the zygo- matic and glenoid portion of the squamosal bone of Mammalia. This is in contradiction to the view held by many comparative anatomists from the day of Reichert to the present time. I made a study of these arches several years ago, which is published in the Proceedings of the American Association Adv. Science, Vol. xix, p. 18. Accepting the prevailing view that the quadrate bone is one of the auditory ossicles, I naturally homologized the superior arch of the rep- tilian skull, which articulates with the squamosal proper, with the zygo- matic arch, and looked upon the quadratojugal arch as an additional structure, connected with the peculiar development of the supposed incus. * Sur la valeur morphologique de I'articulatlon maudibulaire et des osselets de I'oreiile, etc , Bruxelles, Mayolez, 1883. 18S4.] 43 [Cope. Should Albrecht's determination of the homology of the quadrate bone prove to be correct, the quadratojugal arch is the zygomatic, and the superior arch becomes the accessory one. This being admitted, the Lacer- tilia cannot be said to have a zygomatic arch, and theTheromorpha do not possess their postorbito squamosal arch ; the diversity between the two orders being thus greater than has been supposed. The akticulation of the ribs in Embolophorus. — The ribs of the Theromorpha are two-headed. While the tubercular articulation has the usual position at the extremity of the diapophysis, the capitular is not distinctly, or is but partially indicated, on the anterior edge of the cen- trum, in Clepsydrops and Dimetrodon. In Embolophorus, as I showed in 1869, the capitular articulation is distinctly to the intercentrum. A second and larger species of that genus, recently come to hand, displays this character in a striking degree, since the intercentrum possesses on each side a short process with a concave articular facet for the head of the ribs. From the slight corresponding contact with the intercentrum seen in Dimetrodon and other genera, there can be little doubt that this is the true homology of the ribs in the order Theromorpha. The consequence follows from this determination, that the ribs of this order are intercentral and not central elements, and that they do not there- fore belong to the true vertebrae, thus agreeing with the chevron bones, with which they are homologous. It is also true that this type of rib-articulation approximates closely that of the Mammalia, where the capitular articulation is in a fossa excavated from two adjacent vertebrae. This is what would result if the intercen- trum were removed from a Theromorph reptile, and the head of the rib allowed to rest in the fissure between the centra left by the removal. It is well known that the double rib articulation of the other reptilian orders which possess it, viz.: Ichthyopterygia, Crocodilia, Dinosauria and Ptero- sauria, and in the birds, is different, the capitular connection being below the tubercular, on the centrum. Whether the capitular articulations and the ribs in these orders are homologous with those of the Theromorpha, remains to be ascertained. The origin op the Mammalia. — The relation of the characters of the Pelycosaurian suborder of the Theromorpha to those of the Mammalia may now be seen to be very important. I give a synopsis of the charac^ tera of these divisions parallel with those of the Batrachia contemporary with them, in order to give a clear idea of the reasons for believing that the Mammalia are the descendants of the Pelycosauria. The following table shows that the Mammalia agree with the Batrachia in two and part of another character ; with the Pelycosauria in six char- acters, and with other Reptilia in two characters. The Pelycosauria agree with the Batrachia in two and in parts of two other characters, and with other Reptilia in three characters, two of which (Nos. 2 and 3) are of prime importance. Of the characters in which the Pelycosauria agree Cope.] 44 [August 15, 02 O a a* . O &5 33. :i 3 c: a a J3 c3 O) 3 a . a 1) 03 a D a 'T O o ^ O T3 O g a a a; c3 o a CIS d 03 2 ^ 1 ^3 On:) a oj d a a ^ c3 ^ a •-3 »^ rr, a >y a S o 0) a « o S ^3 a 'TIS CD . 02 a ce 9. 2| a S d « - ° a5 ^ i3 a-;3 c >> So c3 ,a p. o a . o a £ d o s s a ^ --tea a >» a o o a o o 'o o O a p c3 1884.] 45 ICope. with the Mammalia, two are of first class importance (Nos. 1 and 5); three are of great but unascertained degree of importance (Nos. 4, 6 and 8), and one (No, 9) is of less importance. The two characters (Nos. 2 and 5) in which the Mammalia agree with the Batrachia, are of high importance, hut one of them is also a point in which the Pelycosauria agree with both (structure of the coracoid bone, No. 5). There is but one character, the distinctness of the quadrate bone, in which the Batrachia agree with the Reptilia in general. The preceding comparison renders it extremely probable that the Mammalia are descended from the Pelycosaurian Reptilia. The usual definitions have been invalidated, excepting that of the occipital con- dyles, but even this is not so absolute a character as has been supposed. In the gecko lizard, Uroplates, the occipital condyle is represented by the exoccipital pieces only, the basioccipital element being omitted nearly as in the Mammalia. Professors Huxley and Parker have declared it as most probable that the true ancestor of the Mammalia have been the Batrachia. It is evident that the Pelycosauria are in various respects the most Batrachian of the Reptilia, for they agree with them in three and parts of two other characters of the nine above enumerated. One of the latter is the structure of the posterior foot, which displays much less modification from the Batrachian type than that of the ordinary Reptilia. The first evidence of the resemblance of the Pelycosauria to the Mam- malia was empirical and not conclusive. This consisted in the characters derived from the long bones of the limbs. Professor Owen first called at- tention to this resemblance in the genus Cynodraco, which is a Theromorph reptile. I next pointed out corresponding peculiarities in the humeri of the American Theromorphs. I subsequently showed the resemblance between the pelvis of the Pelycosaur division, and that of the Monotremata. This was followed by a demonstration of the resemblance between the coracoid of the Pelycosauria and the Mammalia of the Monotrematous order, especially the family of the Platypodidse. The present article now adds that the structure of the posterior foot approaches near to that of the Monotremata ; and that the os quadratum and the ribs are essentially like the corresponding parts in all the Mammalia. The last three points are essential and fundamental. The three great distinctions between the Mammalia and Reptilia in the skeleton are : (1) in the quadrate bone ; (2) in the coracoid bone, and (3) in the occipital condyle. Of these the last only now remains, and this is weakened by the presence of the Mam- malian type in the geckotian lizard already referred to. The only inter- ruption in the series which has not yet been overcome is in the columella auris. No reptile is yet known where that element is divided into incus, orbicularis, and stapes, as in the Mammalia and some Batrachia (according to Albrecht). Of course the above comparison with the Monotremata con- siders the latter order in its proper ordinal definitions, and not in its special subordinate modifications now existing, the Platytidse and Tachyglos- sidae. Montremata dentition like that of the known Jurassic and Triassic Mammalia will doubtless yet be discovered in beds of those ages. \ Cope.] 46 [August 15, As this paper goes to press, the interesting announcement made at the meeting of the British Association for the advancement of science at Montreal may be referred to. Mr. Caldwell, the holder of the Balfour scholarship, telegraphs that he has discovered that the Platypus anatinus is oviparous, and that the egg is meroblastic. This confirms the hypothesis of descent from reptilian ancestors rather than Batrachian. Haeckel gives the segmentation as meroblastic, Studien zur Gastraea Theorie, Jena, 1877, p. 65. Note on the Tarsus. — I am just in receipt of an MS. from Dr. Baur, of New Haven, in which he presents an identification of the "internal navicular" bone of some rodents, and which probably existed in the ungu- late genera Pantolambda and Bathmodon. He identifies it with the tibiale, and denies that the astragalus includes that element, but that it consists wholly of the intermedium. This identification will also apply, though Dr. Baur in his manuscript does not make it, to the element which sup- ports the spur in the known Monotremata. It will also explain the nature of the element which occupies the same position in the foot of the Pel3'^co- sauria above described. The arrangement in this order of reptiles con- firms the conclusion reached by Dr. Baur, since the questionable element is here in direct contact with the tibial facet of the astragalus. Note on Phylogeny op the Vertebrata. — As my researches have now, as I believe, disclosed the ancestry of the Mammals,* the birds, f the reptiles, and the true fishes,:}: or Hyopomata, I give the following phy- logentic diagram illustrating the same. This will only include the lead- ing divisions. The special phylogenies of the Batrachia I and Reptilia,§ and some of the Mammalia ^ have been already given. The Mammalia have been traced to the Theroraorphous reptiles through the Monotremata. The birds, some of them at least, appear to have been derived from the Dinosaurian reptiles. The Repiilia in their primary rep- resentative order, the Theromorpha, have been probably derived from the Rhachitoraous Batrachia. The Batrachia have originated from the sub- class of fishes, the Dipnoi,:}: though not from any known form. I have shown that the true fishes or Hyopomata have descended from an order of sharks, ^: the Iclithyotomi, which possess characters of the Dipnoi also. The origin of the sharks remains entirely obscure, as does also that of the Marsipobranchi. Dohrn** believes the latter class to have acquired its * American Naturalist l!584, p. 1136. t Proceedings Academy Philadelphia, 18G7, 234. t Proceedings American Philosophical Society, 1884, p. 585. II American Naturalist, 1884, p. 27, § Pi'oceedings American Association for the Advancement of Science, xix, 1871, 203. ^ Proceedings American Philosophical Society, 1882, 447; American Naturalist, 1884, p. 261 and 1121. Report U. S. Geol. Survey W. of lOOth Mer., G. M. Wheeler, 1877, iv, ii, p. 282. ** Der Ursprung der Wirbelthiere u. d. Princip des Fuactionwechsels, von Anton Dohrn, Leipsic, 1875, p. 32. ri oc.Am.Philcs. Soc,N°il7. ISSl.l 4.1 [Cope. pre&;cnt character by a process of degeneration. The origin of the Verte- brata is as yet entirely unknown, Kowalevsky deriving them from tlie Tunicata, and Semper from the Annelida. Aves Mammalia Keptilia Selachii Ilvopomata Batrachia I \ Ichthyotomi Dipnoi Holocephali Leptocardi Marsipobranchi EXPLANATION OF PLATE. Fig. 1. Clepsydrops leptoceplialus Cope, right quadrate bone (Q) with condyle and zygomatic process (z) from the right, or external side. Pt, pterygoid bone of same side displaced so as to be in plane of quadrate, and to be seen from inferior side. One-half natural size. Fig. 2. Columella auris of the individual of Clepsydrops leptocepJialus represented in fig. 1 ; internal side. Fig. 2a external side ; 2b proximal extremity ; 2c distal extremity ; s, head of stapes ; Ecol. epicolumella ; d, distal articular surface, especially represented in fig. 2c. All figures are half natural size, excepting 2c, which is natural size. Fig. 3. Left half scapular arch of a Pelycosaurian, less clavicle and episternum, one-half natural size, sc, scapula ; cl, facet for clavicle ; cor, coracoid ; ec, epicoracoid ; open suture between coracoid and epicoracoid, indicating the immaturity of the animal. Fig. 4. Dorsal vertebra of a species of Embolophorus, one-half natural size; right side; a, from front; b, from below; ic, intercentrum; ca, capitular rib articulation. ! Fig. 5. Astragalus of individual figured in fig. 4, one-half natural z\ze ; j from below;, ca, ca, facets for calcaneum ; nn, do. for navicular ; tib, 2, \ do. for bone of spur, or os tibiale. 5«, same bone from external or cal- caneal border ; /, fibular facet. 56, same bone, proximal or fibular ex- tremity. Fig. 6. Left posterior foot of Clepsydrops natalis Cope, superior side, and Qa, inferior (plantar) side, two-thirds natural size, as, astragalus ; ca, calcaneum ; na, navicular bone ; cu, cuboid ; euc, mc and ecc, en- tocuneiform, mesocunei'form and ectocuneiform bones, respectively. I, II, III, IV, V, metatarsals. Tib 1, Probable tibial facet. In this specimen the calcaneum is displaced ; being turned backwards, so as to present its two astragalar facets (a«/) anteriorly. Printed November 20, 1884. {^Froni American Naturalist, Nov., Dec, 1884, Jan., iSSj.) PHE AMBLYPODA. ID. coifde:. POINTED January 6th, 1885. IIIO The Amblypoda. [November, THE AMBLYPODA. BY E. D. COPE. THE Amblypoda is that order of hoofed mammals in which the internal bones of the carpus are in linear series, the mag- num supporting a part of the lunar only, and not reaching the scaphoid ; and in which the tarsal bones are interlocked, as in the later types of hoofed Mammalia ; that is, the cuboid bone articu- lates extensively with the astragalus, as well as with the calcan- eum (see Figs. 1-2). Fig. I. — Right manus of Coryphodon. Original, from Proc, Amer. Philos. Soci- ety, 1882, p. 441. ^fT, scaphoid ; Z, lunar; Cu, cuneiform (imperfect); Tz, trape- zium; 7d?, trapezoides ; yJ^/i?, magnum (face broken) ; unciform. From a New Mexican specimen. See Report U. S. G. G. Survey W. of looth mer., iv, pi. LXI. Fig. 2. — Right posterior foot of a species of Coryphodon from New Mexico, one- half nat. size. From Report Expl. W. of 1 00th mer., G. M. Wheeler, IV, pi. LlX. Various other characters are associated with these, some of which may be coextensive with them, and therefore to be re- garded as characters of the order. Thus all the known species have five toes on all the feet, and a flat astragalus without trace of groove. The unciform bone is extensively in contact with the lunar, as in the Diplarthra. The hemispheres of the brain are of singularly small size, and are separated from the olfactory lobes and from the cerebellum by their crura. The feet are always short and plantigrade. Ordinal characters of wider significance are seen in the pres- 1 884.] The Amblypoda. nil ence of a postglenoid process ; the presence of enamel on the teeth, and the absence of any teeth growing from persistent pulps. The superior molar teeth are constructed on the tritubercular type. Mammalia presenting the above combination of characters first appear in the Puerco Eocene epoch, and continue through the Wasatch and Bridger epochs, and then disappear. They have not been found in later deposits. The Puerco species are small- est in dimensions, while those of the Bridger epoch are the largest of Eocene Mammalia, equaling in the size of some of their bones the largest of living terrestrial Mammalia. Besides their great size the species of the Bridger epoch prove the ancient exuberance of their growth-energy in the development of extra- ; ordinary processes and horns on the head and large and formida- i ble canine teeth. Three suborders represent this order. Each one is confined to a distinct horizon of the Eocene period, with one doubtful excep- tion. The characters are as follows : I Astragalus with a head; a third trochanter of femur; superior inoisors. . Taligrada. No head of astragalus; a third trochanter ; superior incisors Pantodonta. No head of astragalus, nor third trochanter, nor superior incisors Dinocerata. The Taligrada only occur in the Puerco epoch, the Pantodonta are confined to the Wasatch, and the Dinocerata belong to the Bridger epoch, one species having been found extending down- wards into the Wasatch, unless there be some error.^ The Taligrada. The Taligrada are represented by one family, and one genus, Pantolambda Cope.^ This form is interesting on various accounts. It furnishes the only known example of a tritubercular seleno- dont superior molar dentition. That is, each of the three cusps of which the crown of the superior molar is composed, is a well- formed V (Fig. 3 d). We shall see later on, that this form ex- plains the origin of the peculiar molars of some of the Panto- donta. Its astragalus is of remarkable form, which is interme- diate between that presented by the other members of the order, md the diplarthrous ungulates, as for instance, the rhinoceroses [Fig. 3, fig. d d' d"). It has the flat tibial face characteristic of ill the Puerco mammals, and of many of those of the Wasatch * This is the Bathyopsis Jissidens Cope. *See Naturalist, 1883, p. 406. III2 The Amblypoda. [November, epoch. The inferior molars are much like those of Corypho- don, but differ in certain details. The premolars are even more like those of that genus (Fig. 6). The humerus differs from that of all other members of the or- der in having an epicondylar foramen (Fig. 5 d). In this , , respect, and in its Fig. 3. — Bones and teeth of Pantolambda bathmodon , . Cope, two-thirds nat. size. From the Puerco beds of New large mternal Cpi- Mexico. Fig. a, part of maxillary and malar bones from coi^(;Jylg \\\^ form is below, showing true molars, all somewhat broken. Figs. _ ' (5 and cervical vertebrae, leftside; b' and e,hoY\t5 donta or Dmocerata. In of individual represented in figs. 3 and 4, two- the lower iaw of the P. ^^"^^^ distal half of humerus, 'posterior side; a', anterior side. Fig. (5, ilium cavirictus (Fig. 6) they are from inner side. Fig. metapodial bone from front ; , distal view ; d^ phalange ; c, unguis. somewhat recurved, fur- Original. nishing a formidable seiz- ing and holding weapon. The lower edge of the mandibular *0n this subject see Proceeds. Amer. Philos. See, 1883, p. 324. 1 1 14 The Amblypoda. [November, ramus is prominent anteriorly, leaving a concavity of the lateral face posterior to the root of the canine. This resembles a good deal what is seen in the cat-like genus Nimravus and allies, which are probably the ancestors of the saber-tooth tigers. The corre- sponding form in Pantolambda in the same way foreshadows the great development of mandibular flange, and of the superior canine, seen in the Dinoceratous group of this order. A 1884.] The Amblypoda, The homologies of the parts of the teeth of the different gen- era of the Amblypoda are not determined without considerable study. The greatest difficulties are presented by the superior molars, and in order to understand them it is necessary to study the inferior molars first. If we compare the inferior molars of the Pantolaynbda cavitictus with those of the Bath- yopsis fissidens (Fig. 7) (one of the Dinocerata) the following modifica- tions will be observed: (i) The two limbs of the anterior V in the true molar of the for- mer are represented in the latter by a single crest with two appress- ed apices at the inner extremity. These apices represent those of the two limbs of the V. (2) The posterior imb of the posterior V is much diminished in elevation, and does not join its anterior limb at the apex on the ex- ternal side of the crown. If we now examine the corresponding parts of another Dinoceratous Amblypod, a species referred by Mr. Osborn to the genus Loxolo- phodon (Fig. 8), we find the process carried still further. The crov/n has only two crests which approach sach other but do not join at the apex on the inner side. The 1 1 16 The Amblypoda, [November, pos":erior limb of the posterior V has disappeared, and a small Pig. 8. — Loxolophodon sp., mandible anterior to coronoid process, one-fourth nat. size ; from Bridger beds of Wyoming. From Osborn, memoir on Loxolophodon and Uintatherium. cusp represents the posterior crest of the anterior V. Fig. 9.— Teeth of Pantodonla, two-thirds nat. size. Figs, a-d, Coryphodon anax Cope ; from the Wasatch beds of Wyoming. Original, from Vol. ill Report U. b. Geol. Surv. Terrs. Fig. . Geol. Surv. Terrs., F. V. Hayden in charge. 1884.] The Amblypoda, II17 Let us now compare the superior molars of Pantolambda (Fig. 3) with those of Coryphodon (Fig. 9 a). The two external Vs of the former are represented in the latter by a cusp and a partial V, which form an oblique ridge on the external half of the posterior side of the tooth. This arrangement is already imitated on the last molar of Pantolambda, where one of the Vs has disappeared, and the one that remains is very oblique. The second V preserves its form better in the superior molar of Manteodon qnadratus (Fig. 9 /) ; while in the genus Metalo- phodon Cope, it is reduced to a cusp, forming the inner extremity of a transverse crest. (A figure of this genus will be given in the next number of the Naturalist.) The interior V of Panto- lambda is represented by a corresponding angle at the inner side of the crown in Coryphodon (Fig. 9 d), and the two ridges which connect it with the external base of the crown are well marked. Besides these there is developed on the anterior side of the base a strong cingulum. This is wanting on the posterior side, but is present in a feeble form in Pantolambda (Fig. 3). I have sometimes called these the su- perior and inferior anterior cin- gula respectively. In Corypho- don the posterior superior cin- g u 1 u m is unsymmetrically bulged near its interior end, and in Manteodon it develops a cusp at this point, the crown thus becoming quadritubercular (Fig. 9/)- If we .now compare these teeth with those of Uintatheri- um (Fig. 10) we have little diffi- culty in determining the presence of the two anterior cingula. The difficulty is to ascertain the homology of the posterior trans- verse crest. It is difficult to see in it the two external Vs ot Pantolambda confluent into a simple ridge, yet such it seems to be. This will be especially clear on comparison with the figure of Metalophodon. One is tempted to compare it with the poste- rier superior cingulum of Coryphodon, but the frequent presence Fig. 10. — Dinocerata teeth, one-fourth nat, size. Upper figures superior molars of Uintathermni leidianum, one-fourth nat. size. Lower figure, inferior molars of jaw of Loxolophodon represented in fig. 8. From Osborn, memoir on Uinta- theriumand Loxolophodon. Iii8 The Amblypoda. [November, of a cusp posterior to its internal extremity in the Dinocerata presents a further obstacle to such a determination. This cusp is probably homologous with the posterior internal tubercle of Man- FiG II. — Last three temporary molars, and first true molar partly protruded, of a Coryphodon from New Mexico, nat. size. Original, from Report U. S. G. G. Sur- vey W. of looth mer., G. M. Wheeler, Vol. iv,pt. ii. teodon (Fig. 9 /), and is, therefore, the posterior internal of the quadrituberculate molar type. The internal extremity of the pos- terior crest of the crown in the Dinocerata cannot, therefore, be that cusp, and that ridge is not a posterior transverse crest, such as is present in most Perissodactyla. It is interesting to observe that the fourth superior deciduous molar in Coryphodon (Fig. 1 1) has essentially the same structure as the true permanent molar of Pantolambda (Fig. 3). The origin of the Amblypoda has not yet been traced. What was known on this point, as well as on the question of the de- scendants of the order, was stated as follows in my paper on the classification of the ungulate/Mammalia, published in 1882:^ " As regards the inner part of the manus, I know of no genus which presents a type of carpus intermediate between that of the Taxeopoda and Amblypoda on the one hand, and the Perissodac- tyla and Artiodactyla on the other. Such will, however, proba- bly be discovered. But the earliest Perissodactyla, as for instance Hyracotherium, Hyrachyus and Triplopus, possess the, carpus of the later forms. Rhinoceros and Tapirus. The order Amblypoda occupies an interesting position between the two groups, for while it has the carpus of the primitive type, it has the tarsus of the later orders. The bones of the tarsus alternate, thus showing a decided advance on the Taxeopoda. This order is then less primitive than the latter, although in the form of its astragalus it no doubt retains some primitive peculiarities which none of the ^Palaeontological Bulletin No. 35. Proceedings Amer. Philosoph. Society, 1882, pJI The Amblypoda. 1119 known Taxeopoda possess. I refer to the absence of trochlea, a character which will yet be discovered in the Taxeopoda, I have no doubt. "The Taxeopoda approach remarkably near the Bunotheria, and the unguiculate and ungulate orders are brought into the closest approximation in these representatives. In fact I know of nothing to distinguish the Condylarthra from the Mesodonta, but the ungulate and unguiculate characters of the two divisions. In the Creodonta this distinction is reduced to very small propor- tions, since the claws of Mesonyx are almost hoofs. Some of the genera of the Periptychidae present resemblances to the Creodonta in their dentition also. "The facts already adduced throw much light on the genealogy of the Ungulate Mammalia. The entire series has not yet been discovered, but we can with great probability supply the missing links. In 1874 I pointed^ out the existence of a yet undiscovered type of Ungulata, which was ancestral to the Amblypoda, Pro- boscidea, Perissodactyla and Artiodactyla, indicating it by a star bnly in a genealogical table. This form was discovered in 1881, seven years later, m the Condylarthra. It was not until later^ :hat I assumed that the Diplarthra are descendants of the Ambly- poda, although not of either of the knov/n orders, but of a theo- •etical division with bunodont teeth.^ That such a group has existed is rendered extremely probable in view of the existence Df the bunodont Proboscidea and Condylartha. That the Taxeo- Doda was the ancestor of this hypothetical group, as well as of he Proboscidea, is extremely probable. But here again neither )f the suborders of this group represent exactly the ancestors of he known Amblypoda, which have an especially primitive form >f the astragalus not found in the former. In the absence of an Imkle joint the Amblypoda are more primitive than any other livision of the Ungulata, and their ancestors arc not likely to lave been more specialized than they. It is probable that a third iubo'rder of Taxeopoda has existed which had no trochlea of the iistragalus, which I call provisionally by the name of Platyarthra. "The preceding paragraphs were written in May, 1882. On ^ Homologies and Origin of Teeth, etc., Journal Academy Nat. Science, Philada., 874, p. 20. 2 Report U. S. Geol. Survey W. of looth mer., p. 582, 1877. ' This hypothetical suborder has been called Amblypoda Hyodonta. 1 1 20 The Amblypoda, [November, my return home, September ist, after an absence of three months, I find that various parts of the skeleton of Periptychus have reached my museum.^ On examination I find that the astragalus of that genus fulfills the anticipation above expressed. It is with- out trochlea, and nearly resembles that of Elephas. As it agrees nearly with that of Phenacodus in other respects, I only separate it as a family from the Phenacodontidae. One other type re- mains to be discovered which shall connect the Periptychidae and the hypothetical Hyodonta, and that is a taxeopod without a head to the astragalus — unless, indeed, the * Hyodonta ' should prove to have such a head. I think the latter the less probable hypothesis, and hence retain the term Platyarthra for the hy- pothetical taxeopod without trochlea or head of the astrag- alus." This group as at first defined had been already discovered, but in the above paragraph I added another definition, endeavor- ing to preserve an apparently useless name. The existence of Amblypoda Hyodonta is rendered almost certain by the discovery that the genus Pantolestes of the Wa- satch epoch is an artiodactyle with tritubercular bunodont supe- rior molars. The ancestral type of such a form must have been a tritubercular bunodont amblypod. Pantolambda is such a form with the tubercles modified into Vs. Moreover, such a type (Amblypoda Hyodonta) would be derived from a periptychid Taxe- opod with but little modification of the latter. A distinct facet- ing of the astragalus for the cuboid bone, and probably a change of the carpus by an articulation of the unciform and lunar bones, would be all that would be necessary. The discovery of Panto- lambda has increased the probability of such change having taken place in the hind foot, since the astragalus is intermediate in form between those of Coryphodon and Periptychus. This is particularly seen in the form of the head, which is not separated by a neck as in Periptychus, but is more prominent than in other Amblypoda (Fig, 3). Thus it is possible that Amblypoda Hyodonta will be found to have a short head of the astragalus, contrar)'- to my supposition above expressed. The headless astragulas of other Amblypoda and of Proboscidea is perhaps a derivative and not a primitive character. The following diagram gives expression to these views as to the phylogeny of the ungulates. It is the same that I published ^ See Naturalist for August, 1884, for figures. ' 1884.] The Amblypoda. 1121 < in 1882 in the paper above quoted, modified so as to be more in- I telligible : (To be continued}) 1 192 The Amblypoda, [December, THE AMBLYPODA. BY E. D. COPE. ( Continued from page 1121, November number,)'*' Fantodonta. 'Y'HERE is known as yet but a single family of this suborder, ^ the Coryphodontidae. Its representatives have been found in the lower lacustrine Eocene beds in Europe and North America in considerable abundance. About twenty species have been de- scribed, of which three have been found in England and France, and the remainder in the Rocky Mountain region of Nort America. They form a curious and interesting group of hoof Mammalia which did not survive the Lower Eocene time, exce in their probable descendants, the Dinocerata. The characte of the suborder have already been given in the Naturalist, pa 1 1 1 1. Five genera of the Coryphodontidae are known from deni characters. Two of these, Coryphodon and Bathmodon, ar known in their skeletal structure, the first-named very thoroughl The explanations of Fig. 7 (p. 1115) should read four-ninths nat. size; nottw thirds nat. size. 1884.] The Amblypoda. Bathmodon is nearest to the genus Pantolambda in its foot struc- ture. The astragalus has the same subquadrate outline (compare Figs. 3^ and 19^), and possesses, as in that genus, a facet on its internal anterior angle, for another bone. This character is found among recent Mammalia in certain rodents, where a separate bone, the internal navicular, articulates with the internal extremity of the astragalus. This is probably the case in Bath- modon and Pantolambda, but the separate bone has not been found. In Coryphodon it is evidently wanting, as the astragalus terminates internally in a hook-like apex, which gives it a very peculiar appearance (Fig. 2). Until I traced this astragalus to Coryphodon in 1873,^ the relation of this genus to the Dinocerata had not been suspected. FiG. Coryphodon elephantopus Cope, skull from left side, the left half of the >osterior part removed so as to display the small size of the brain cavity; two-ninths atural size. From Wasatch bed of New Mexico. Original, The small size of the brain in this family is well displayed in he section of the skull of Coryphodon represented in Fig. 12, yhere the cavity it has occupied is exposed. Its relations to the jkull are entirely different from those observed in all recent Mammalia excepting the elephants. As in them the diploe is epresented by large air chambers. The characters of the brain itself may be learned from Fig. 3. The exceedingly small size of the cerebral hemispheres t once arrests the attention, being much smaller than in the I * On the short-footed Ungulates of the Eocene of Wyoming. Proceedings Amer. bilosoph. Society. 1 194 The Amblypoda. [December, cotemporary genus Phenacodus/ and agreeing with Uintathe- rium, which has been described by Marsh. There is no trace of sylvian fissure nor of con- volutions. The middle brain is entirely exposed, and how much pertains to this and how much to the cerebellum re- mains uncertain. The ante- rior prolongation of the ante- rior pyramids is visible belov/ the middle brain, as though the pons varolii were wanting (Fig. np\ The characters of the long bones which are common to the members of this family may be observed in the Bath- modon pachypus, Figs. 14-15. The humerus shows the robust "•'^ ^ but not hooked tuberosities of Fig. 13. — Coryphodon elephantopus Cope, .1 v 1 ■• ^i. • 1 cast of brain cavity seen in fig. 12, about the head, and the smiple con- one-third nat. size. Fig. «, left side; b, dyles. The radius shows the above ; c, below, t, origin of trigeminus . nerve; anterior pyramids. Original, from Simple transverse head and Proceed. Amer. Philos. Soc, 1877. ^j^^ \^X.qx^\ distal facet. The femur (Fig. 15 ^) has well developed trochanters, including the third, and a fossa ligamenti teris. The ilium (Fig. 15 d) like that of animals with a large belly, is much expanded, and has a wide peduncle. The pubis and ischium are light. The five genera of Coryphodontidae differ as follows : I. Superior molars with two interior cusps. All the superior molars with a well-marked external posterior V Manteodon. II. Last superior molar with but one inner cusp or angle. a. Last superior molar with posterior external cusp. Anterior two molars with posterior external V Ectacodon- aa. Last superior molar without external posterior cusp, f Anterior two molars with posterior external V. Astragalus transverse, with internal hook and no facet Coryphodon. Astragalus subquadrate, with an internal facet and without internal hook, Bathmodon. ff First superior molar only, with posterior external V Metalophodon. * See Naturalist, 1884, p. 898, for figure. 1884.] TJie Amblypoda. The general relation of the teeth of these genera to those of other famines of the order has been discussed (supra p. 1117). Their relation to each other may be understood by comparison of Fig. (^a and /with Fig. 16. In the latter the anterior external V is marked a, and the posterior p. The posterior exterior angle of the latter is designated by the letter e. In Manteodon (Fig. 9 /) the posterior external V is observed to be well developed on the last superior molar. In Ectacodon (Fig. 16^) its posterior edge is represented by an external cusp only (e), the rest of the border Fig. 14. — Bathmodon pachypus Cope, bones of fore limb, one fifth nat. size, from Wasatch beds of Wyoming. Fig. a, left humerus from behind ; a\ proximal, distal views. Fig, i^, left radius from behind; b\ proximal, b'', distal views; pisiform bone. From Vol. ill Report U. S. Geo!, Survej', F. V. Hayden. being absent. In Coryphodon this border is sometimes trace- able (^7. anax, Fig. 9 a) or is wanting. In Metalophodon it is entirely absent, and is represented by a conical projection of the posterior crest only (Fig. i6<^p). The posterior V of the second superior molar is distinct in all except Metalophodon (Fig. 16 ^p), where it is represented by a crest only, as in the last molar of most of the species of Coryphodon. The succession is thus seen to consist of the gradual conversion of the external Vs into iigS The Amblypoda. [December, transverse crests, a process which is consummated in the Dino- cerata of the later Bridger epoch (see Fig. lo). Of Bathmodon two species are known, B. radians (Fig. i8) and B. pachypus (Figs. 14, 15 and 19). The latter is the larger. Fig. 15. — Bathmodon pachypus Cope, bones of individual figured in Fig. 14, one- fifth nat. size. Fig. a, left innominatum, somewhat distorted, internal view; by left femur from behind. Original, from Wasatch beds of Wyoming. equaling a large ox in dimensions. The crania of these species are unfortunately unknown. Both are from the Wasatch beds of Wyoming. Of Ectacodon Cope, but one species, the E. cinctus, has been I 1884.] 7he Amblypcaa. 1197 discovered. It is only known Irom the superior molar teeth fig- FlG. 16. — Superior molar series of Coryphodontidse, two-thirds nat. size, from the Wasatch beds of Wyoming. Original. Pig, Ectacodon cincHis Qo^q. Fig. Metalophodon testis Cope. ured in Fig. \6a. It is a large species, about equal to the Bath- in odon radians^ Fig. 17, — Coryphodon elephantopus Cope, skull, two-nintht> nat. size, from below. Original, from Report G. M. Wh eeler, U. S. G. G. Surv. W, of looih mer.. Vol. iv, pt. II. The greater number of species of the fanjily belong to Cory- phodon. In this genus the temporal fossae are lateral, as in modern ruminants, leaving a wide front with overhanging tern- The Amblypoda. [December, poral ridges (Figs. 20-21). The dental formula is the complete one of I. 3 ; Pm. \ \ M. f. The superior canines are very power- ful, and have three sides, giving a triangular section. In some of the species {C. molestus), this tooth is more compressed towards the apex, and the posterior face is narrow and concave, forming a groove. The inferior canines are also triangular in section, and the anterior angle is produced into an ala at the base (Figs. 9 c, 23 i). The neck and tail are of median length. The general appearance of the Coryphodons, as determined by 1884.] The Amblypoda. 1 199 the skeleton, probably resembled the bears more than any living animals, with the important exception that in their feet they were much like the ele- phants. To the general pro- portions of the bears must be added a tail of medium length. Whether they were covered with hair or not is, of course, uncertain ; of their nearest liv- ing allies, the elephants, some were hairy and others naked. The top of the head was doubtless naked posteriorly, and in old animals may have been only covered by a thin epidermis, as in the crocodiles, thus pre- senting a rough, impenetrable front to antagonists. Fig. 19. — Bathmodon pac hypus Co\>t, bones of foot of individual represented in figs. 14 and 15, one-fifth nat. size. Fig. a, astragalus from above; , inner side ; below. Figs, b , calcaneum. Fig. 20, — Coryphodon elephantopus Cope, skull figured in figs. 12, 17 and 21, two- ninths nat. size, from above. From Wasatch bed of New Mexico. Original, from Report U. S. G. G. Surveys and G. M. Wheeler, Vol. iv, pt. ii. The movements of the Coryphodons doubtless resembled those of the elephant in its shuffling and ambling gait, and may have been even more awkward from the inflexibility of the ankle. But in compensation for the probable lack of speed these animals were most formidably armed with tusks. These weapons, partic- ularly those of the upper jaw, are more robust than those of the Carnivora, and generally more elongate, and attrition preserved rather than diminished their acuteness. The size of the species varied from that of a tapir to that of an ox. I20O The Amblypoda. [December, We must suppose that the Coryphodons were vegetable feed- FiG. 21. — Coryphodon elephantopus Cope, skull from leftside, two ninths nat. size. From Wasatch bed of New Mexico. Original, from Report U. S. G. G. Survey W. of looth mer., Vol. IV, pt. ii. Y\G. 22. — Coryphodon latidem Cope, lower jaw, one-third natural size, from the Wasatch epoch of New Mexico. Fig. a, right ramus from internal side. Fig. both rami from above. Original, from Report U. S. G. G. Surveys W. of lOOth Mer., G. M. Wheeler in charge. This specimen has an anomalous premolar. 1884.] The Amblypoda. 1201 ers, but not restricted to any particular class of food. They were doubtless, to a large extent, like the hogs, omnivorous. Fourteen species of this genus have been described. They range in size from the dimensions of a tapir to those of an ox. In the absence of the bones of the skeleton the species may be distinguished by the inferior true molars, which are fortunately the parts most frequently preserved. The simplest form is that where the posterior crest of the posterior true molar is transverse, and there is no crest or cusp accessory to it on the inner edge of the crown, as in C. latidens (Fig. 22). A change in the form of this crest is seen in the C. ciinncristis , where it is curved forwards at the inner extremity so as to enclose a crescent-shaped valley. This species adds several other peculiarities of this tooth, as the presence of two oblique crests in front of the anterior cross-crest (Fig. 22). The superior incisors are angulate on the external face. The species was of about the same size as the C. latidens. Fig. 22. — Coryphodon curvia'istis Cope, portions of jaws with teeth, one half nat- ural size, from the Wasatch bed of Wyoming. From Report U. S. Geolog, Survey Terrs., Vol. ili. Original. Fig. a, right mandibular ramus with true molars and Pm. IV, from above. Fig, 3, penultimate superior molar from below ; c d, superior premolars from below ; e, superior incisors, external side ; f, superior canine, crown only; inferior incisors, external views; i, inferior canine, base of crown. The posterior cross-crest may send off the internal marginal crest at an angle, as in C. eocoeniis Ow., or C. obliquris Cope. The internal marginal crest may rise into a low tubercle, as in C. loba- tus or C. anax (Fig. 23 and gbd). In this case the posterior cross-crest may be very oblique, as in C. anax, thus giving the appearance of a heel or fifth lobe to the crown. This is indeed the character of such a crown, which if compared with that of the C. latidens only might suggest generic separation, but we 1202 The Amblypoda. [December, have every intermediate condition. Finally this internal ridge may develop into a conic cusp, as in the C. cuspidatus (Fig. 9 e). Of Metalophodon two species are known, the M. testis (Fig. 16 and M. armatiis Cope, both from the Wasatch beds of Wy- oming. Both species are of about the size of the Coryphodon latidens, and smaller than the C. anax. The skeletons are un- known. (To be continued.) 40 The Amblypoda. [January, {From the American Naturalist^ January, i88s.) THE AMBLYPODA. BY E. D. COPE. ( Continued from page 1202^ Vol. xviii.) DiNOCERATA. IN this suborder we have a series of mammals which . are in some respects the most remarkable that have ever existed. This is true whether we regard the bizarre appearance of their skulls, their dentition, so weak when compared with the bulk of their bodies, or the insignificant size of their brain. We only know them as yet from the Bridger or Upper Eocene formation of North America, with a species possibly from the Wasatch or Lower Eocene. The characters of this suborder have been already pointed out (Vol. XVIII, p. II 21). The differences from the Pantodonta are well marked, but the resemblances are such as to render it impos- sible to refer the Dinocerata to a different order. Their strong resemblances to the Proboscidia are generally admitted, but the few characters which distinguish them are of the first importance. These are, first, the very small size of the brain, especially of the cerebral hemispheres ; and second, the double distal articulation of the astragalus, where the facet for the cuboid bone is nearly as large as that for the navicular. Within the above definition there is room for much variation, which, however, the knowrt genera do not display. They agree in various points of minor importance. Thus there is no sagittal 1885.] The A mblypoda. 41 crest of the skull, the temporal ridges being lateral, and there is a great transverse supraoccipital crest. These crests are more or less furnished with osseous processes or horns. The middle pair of these (Fig. 29) consists in part of the maxillary bone, and stands in front of or over the eye. The nostrils are well roofed Fig. 24. — Loxolophodon cornuius(Zo'<^'t,'i\u\\{xQ>vcv right side, one-eighth natural size. From the Eridger bed of Wyoming. Original, from Report U. S. Geol. Sur- vey Terrs., F. V. Hayden, Vol. iii. Lower jaw restored from Osborn, Memoir on Loxolophodon and Uintatherium. From individual represented in PI. i. over by the nasal bones. There is always a diastema behind the canine tooth in both jaws. There is less difference between the premolar and molar teeth in the known genera than in the Panto- donta, and they all have the same pattern, although the origin of the pattern may be different in the two series. Thus in the upper 42 The Amblypoda. [January, jaw the crowns of the molars support two oblique cross-crests, which unite to form a V with the apex inwards. There is some- times an internal cusp or tubercle. The inferior molars consist essentially of an outer V and a heel ; the true molars differ in having the heel a little larger and more recurved on its posterior border, but it does not rise into a transverse crest as in the Coryphodontidae. Mr. Osborn shows that the inferior incisors in Loxolophodon are compressed and two-lobed. The known genera agree with the typical Proboscidia in the shape of the scapula with pos- terior expansion and apical acumination ; in the flat carpal bones; in the absence of pit for round ligament of the femur; in the flattened great trochanter, contracted con- dyles, and fissure-like intercon- dylar fossa of the same bone. Also in the short calcaneum or heel bone, which is wider than long, and rough on the inferior face ; in the five digits on both feet, and the wide peduncle and iliac plates of the pelvis and lack of angular production of the latter beyond the sacrum. In spite of these resem- blances, the Dinocerata are at one side of the line of descent of the mastodons and ele- phants (see Vol. xviii, p. 1 121, for phylogeny of the hoofed Mammalia). This is indicated not only by the structure of their feet, but by that of their teeth, which, as I have shown, constitute a survival of the tri- tubercular type which had been left behind by all other contempo- rary ungulates, and only survived in the flesh-eaters of the Bridger epoch. Fig. 25. — Uintatherium inhabileyizrsh, bones of feet, two-ninths natural size. Up- per figure anterior, lower figure posterior feet. From Bridger beds of "Wyoming. Slightly altered (lunar bone) from Marsh, Am. Journ. Sci. Arts, XI, PI. vi. TJie Amhlypoda. 43 The resemblance of the feet to those of Coryphodon may be readily seen by comparing Fig. 25 with Figs. 1-2 (p. iiio, Vol. xviii). The characters of the component parts are quite iden- tical. Professor Marsh has given us a figure of the cast of the brain chamber of the UintatJieriinn mirabile Marsh. It displays most striking peculiarities. These are: (i) The small size of the hemi- spheres; (2) the difficulty of distinguishing the cerebellum from the surrounding parts; (3) the large size of the olfactory lobes (Fig. 26). In all these respects there is a great resemblance to the brain of Coryphodon (Fig. 13). The hemispheres pass into the olfactory lobes by a gradual contraction of their outlines. They rise higher than, and then descend posteriorly towards the mesenceph- alon and cerebellum. The latter parts, as in Coryphodon, are not distinguished in the cast. The hemispheres are not convoluted, nor is there any sylvian fis- sure, according to Marsh's figures. This brain, as remarked by Marsh, is the most rep- tilian among the Mam- malia. One of the strongest confirmations of this statement, is the small size of the cerebellum. Owing to the imperfect character of the material which I have had the opportunity of examining, it is not possible to state the number of genera with absolute certainty. There are certainly three of these, and probably four. So far as present knowledge goes, they pertain to one family, which I have called the Eoba- siliidae. The three genera mentioned differ in the forms of the mandible ; the fourth has certain cervical vertebrae of a peculiar form, but the form of the mandible is unknown. I can only con- trast the genera as follows : Fig, 62. — UintatJieriiini mirabile Marsh, brain, one-third naU size. From Marsh, Amer. Jour. Sci. Arts, Vol. XI, PI. IV. 44 The Amblypoda, [January, Mandible unknown. Certain cervical vertebroe short and flat, as in Proboscidia Eobasileus. A A. Symphysis of mandible with four teeth on each side. a. Mandible without inferior expansion. Cervical vertebrae not very short; three premolars; lower incisors bilobate, LgxolopJiodon. aa. Mandible with anterior inferior expansion. Cervical vertebrae not short; three premolars Octotomus. aaa. Mandible expanded below, its entire length. Cervical vertebrae unknown; four lower premolars; incisors simple.. JBathyopsis. AAA, Symphysis of mandible with three or two teeth on each side. Mandible with very narrow symphysis UintatJierium. In probably a majority of the species the lower jaw has a deep flange on its inferior border below the canine teeth, which serves, like the corresponding structure in the saber-tooth tigers, to pro- 27. — Uintatheriiim leidwnum Osborn, skull left side, one-eighth nat. size; from the Bridger beds of Wyoming. From Osborn, Memoir on Loxolophodon and Uintatherium. tect the long superior canine tooth from lateral blows and strains (Fig. 27). In Bathyopsis this inferior expansion includes almost the entire inferior border of the ramus, giving an outline some- thing like that of Megatherium (Fig. 35). The genus Eobasileus was established on a species {E. pressi- cornis Cope) which is represented by a considerable part of the 18850 The Amblypoda, 45 skeleton, but without cranium or teeth ; hence most of its charac- ters remain unknown. The very short cervical vertebra which belongs to it serves to distinguish it from other genera. A sec- ond specimens [E. fiircatiis) found near the first, may belong to it ; it includes a fragmentary cranium, but unfortunately no cervical vertebrae. Its introduction into this genus is therefore purely arbitrary. The typical species is of large proportions, only second in size to the Loxolophodon cornutiis. Its limbs were more slender in their proportions. It is in this species that I find much evidence in favor of the presence of a proboscis of greater or less length. Should several of the other cervical vertebrae have been as short as the one preserved, it is evident that the animal could not pos- sibly have reached the ground with a muzzle so elevated as the long legs clearly indicate. In the species of the other genera, where the cervical vertebrae are longer, this may have not been the case. The bones of this species were discovered by the writer in an amphitheater of the bad lands of the Washakie basin, known as the Mammoth buttes, in Southwestern Wyoming. They were in greater or less part exposed, lying on a table-like mass of soft Eocene sandstone. A description of this remarkable locality is given in the Penn Monthly Magazine for August, 1872. The Eobasileus furcatus is principally represented by a skull in which the most important features have been preserved. As in all the species of Uintatherium in which the horns are known, these appendages stood in front of the orbits, it is probable that such was the case in the Eobasileus furcatus also. The muzzle is materially shorter and more contracted, and the true apex of the muzzle was not overhung by the great cornices seen in Loxolo- phodon cornutus. The occipital and parietal crests are much more extended in this species than in the L. cornutus^ so that in life the snout and muzzle had not such a preponderance of proportion as in that species. All the species of this genus were rather rhi- nocerotic in the proportions of the head, although the horns and tusks produced a different physiognomy. The known species of Loxolophodon Cope, are the largest of the order. Three species are kn.own to be distinct : the L. cornutus Cope, L. galeatus Cope, and L. spierianus Osborn. They differ in the form of the horns and in the shape of the occiput. 46 The Ainblypoda. [January, The cranium in this genus i§ elongated and compressed. The muzzle is posteriorly roof-shape, but is anteriorly concave and flattened out into a bilobed protuberance which rises above the extremity of the nasal bone. This extremity is subconic and short and decurved. A second pair of horn-cores stands above the orbits, each one composed externally of the maxillary bone, and internally of an upward extension of the posterior part of the nasal. Behind this horn the superior margin of the temporal fossa sinks, but rises again at its posterior portion, ascending above the level of the middle of the parietal bones. The occipital rises in a wall upwards from the foramen magnum and supports, a little in front of the junction with the superior and posterior crests bounding the temporal fossa, a third horn-core on each side. Fig. 28. — Loxolophodon spieriantis Osborn, skull rom left side, one-eighth nat- ral size ; from the Bridger beds of Wyoming. From Osborn, Memoir on Loxolo- phodon, etc. The three species may be distinguished as follows : Median horns triangular in section, with internal tuberosity, and above orbits ; occi- put narrow L. cornutus. Median liorns subquadrate in section w'ithout internal tuberosity ; occiput and nasal tubercles wide L. galeatus. Median horns subround and without tuberosity, in front of orbits ; occiput and nasal tubercles narrow L. spierianus. 1885.] The Amblypoda. 47 The LoxolopJwdon spicrianiis Osborn, was as large an animal as the two others, and had a very elongate skull with weak horns and narrow, high occiput. Its median horns are situated well ante- rior to the orbit, and its zygomatic fossa is remarkably small. It was discovered by the Princeton scientific exploring party at the '',amc locality that produced the other species, viz., the Mammoth buttes of Southwestern Wyoming (Fig. 28). In the L. cornutiis and L. galeatiis the tuberosities which stand Fig, 29. — Loxolophodon cornutus Cope, skull of individual represented in Plate i, one-eighth nat. size. Upper figure superior surface ; lower figure inferior surface. From Bridger Eocene of W^yoming. Original, from Report U. S. Geol. Survey of Terrs., F. V. Hayden in charge, Vol. ill. Owing to distortion of the specimen be- hind, the occipital condyles are too far apart in the figure. near the free extremity of the nasal bones are greatly developed, so as to represent a pair of cornices projecting upwards and for- wards over the narrow apex of the bones (Fig. 24). From above, the end of the muzzle in those spcices has a bilobate outline. They differ from each other materially in form ol the middle pair of horns. 48 The Amblypoda. [January, Mr. Osborn, of Princeton, has published a description of the lower jaw and teeth ot a species of Loxolophodon, which he identifies with the L. cornutus, which was derived from the local- ity and horizon of the species above mentioned (Fig. 8). They show that the descending flange of Uintatherium and Bathyopsis is only represented by a convex ridge on each side of the sym- physis. They point out the characters of the dentition, which are remarkable. The molars much resemble those of Bathyop- sis. The canines and incisors are alike in form, and in a contin- uous series. The crowns are compressed so as to be extended anteroposterior^, and are deeply emarginate, so as to be bilobed, the lobes with subacute edges. This form of incisors is unique, resembling only remotely the large median incisors of certain Insectivora (Fig. 30). Resem- blance to mammals of the same type may be traced in the molar teeth. We may ascribe to the Loxolophodon cormitus form and proportions of body Fig. 30. — Incisor and canine ... r i i i / tm teeth of left side of lower jaw Similar to those of the elephant (see Plate of Loxolophodon, one-fourth j\ j^e limbs, however, were some- natural size. From Osborn. ^ what shorter, as the femur (Fig. 31) is stouter for its length than in the E. indicus. It was intermediate in this respect between the latter species and the species of Rhi- noceros. The tibia is relatively still shorter. The tail was quite small. The neck was a little longer than in the elephants, but much less than in the rhinoceroses ; the occipital crest gave at- tachments to the ligainentum nuchcB and muscles of the neck, which must needs have been powerful to support the long muzzle with its osseous prominences, and to handle with effect the terri- ble laniary tusks. The head must have been supported some- what obliquely downvv^ard, presenting the horns somewhat for- ward as well as upward. The third or posterior pair of horns towered above the middle ones, extending vertically with a diver- gence when the head was at rest. The posterior and middle pair of horns were no doubt covered by integument in some shape, but whether dermal or corneous is uncertain. Their penetrating foramina are smaller than in the Bovidae. The cores have re- motely the form of those of the Antilocapra ajnericana, whence I suspect that the horns had an inner process or angle as in the 1885.] The Amblypoda. 49 prong-horn at present inhabiting the same region. The nasal shovels may have supported a pair of flat divergent dermal tuber- osities, but this is uncertain ; they are not very rugose. The elevation of the animal at the rump was about six feet, dis- tributed as follows, allowance being made for the obliquity of the foot: Inches. Foot 4,50 Tibia , 20.50 Femur 31-75 Pelvis 16.00 72.75 The anterior limbs were stouter than the posterior, judging Fig, 31. — Loxolophodon cormdus Cope, femur of individual represented in PI. one-ninth nat. size. From Bridger beds of Wyoming. Original, from Report U. S. Geol. Surv. Terrs., iii, F. V, Hayden in charge. from the proportions in various species, and were no doubt longer if of the Proboscidian character. This would give us the hj^po- thetical elevation at the withers : Inches. Leg 61.00 Scapula (actual) i 21.00 Neural spines (extremities) 7x0 Or 7 feet 5 inches , 89.00 50 The Amblypoda, [January, These measurements are made from the plantar and palmar surfaces, allowance being made for the pads. The obliquity of the anteroposterior axis of the anterior dorsal vertebra indicates that the head was posteriorly elevated above the axis of the dorsal vertebrcX. Owing to the lack of cervical ver- 1885.] The Amblypoda. 51 tebrae, the length of the neck cannot be determined. It may have been short, as in the Eobasileus pressiconiis^ or longer, as in the species of Uintatherium. The indications derived from the bones of the muzzle point to the attachment of a heavy upper lip. The numerous rugosities of the posttympanic and mastoid regions indicate the insertions of strong muscles. Some of these may have been adductors of large external ears. Fig. 33. — Uintatheruim mirabile Marsh, skull, one-eighth nat. size; upper figure trom front, lower figure from above. From Biidger Eocene of Wyoming. From Marsh, Amer. Jour. Sci. Arts, xi, PL 11. The inferior incisor teeth have no adaptation for cutting off vegetation. The mental foramen is small, but the small nutrient artery thus indicated is not adverse to belief in a prehensile under lip to make up for the uselessness of the teeth. The projecting nasal regions would prevent short lips from touching the ground . 52 The Amblypoda. [January, The posterior position of the molar teeth indicates use for a long, slender tongue. The species was probably quite as large as the Indian elephant, for the individual described is not adult, as indicated by the free- dom of the epiphyses of the lumbar vertebrae ; and fragments of others in my possession indicate considerably larger size. The very weak dentition indicates soft food, no doubt of a vegetable character, of what particular kind it is not easy to divine. The long canines were no doubt for defence chiefly, and may have been useful in puUing and cutting vines and branches Fig. 34. — Odotomus laticeps Marsh, lower jaw, one-eighth nat. size; upper figure left side, lower figure from above. From Marsh, Amer. Jotcr. Set. ArtSy XI, PI. V. of the forest. The horns furnished formidable weapons of de- fence. The anterior nasal pair might have been used for root- ing in the earth, if the elevation of the head did not render this impossible. This huge animal must have been of defective vision, for the orbits have no distinctive outline, and the eyes were so overhung by the horns and cranial walls as to have been able to see but lit- tle upward. The muzzle and cranial crests have obstri;cted the The Ambiypoda. 53 view both forward and backward, so that this beast probably- resembled the rhinoceros in the ease with which it might have been avoided when in pursuit. The genus Uintatherium Leidy, has the symphysis of the man- dible more contracted than in the other genera, and the number 'of its teeth correspondingly reduced.^ The type is the U. robus- tiini Leidy, a species which is known from the posterior part of a skull with a few molar teeth of both jaws, and a superior canine tooth of one individual; and by the greater part of the lower jaw of another. It is of smaller size than those referred to Loxolo- phodon, and also smaller than the U. leidianum Osb. (Fig. 27). Besides these two species four others have been described by- Marsh and referred to a genus Dinoceras, which is not yet known to be distinct from Uintatherium. The best known of these is the U. mirabile (Figs. 25, 26, 33), which has been well figured by- Marsh. It lacks a tubercle of the last superior molar which is present in the U. robustum. Its lower jaw is unfortunately un- known. A species described by Marsh as Dinoceras laticeps is of larger size than the D. robustum^ and Marsh figures its lower jaw (Fig. 34). It possesses four teeth on each side of the symphysis, as in Loxolophodon, but their form is not known. There is a deep flange of the lower edge of the ramus below the canine teeth, as in Uintatherium. As this form represents a genus clearly- distinct from either of /these, or Bathyopsis, I propose that it be called Octotomus. To this genus may belong some of the spe- cies now provisionally referred to Uintatherium. In these animals the nasal tuberosities are small, and do not overhang the apex of the nasal bones. The median horns are anterior to the orbits, and are of various degrees of development in the different species. The posterior horns vary in like manner (compare Figs. 27 and 33). The supraoccipital crest extends much further posteriorly in the U. mirabile than in some of the other species. In the genus Bathyopsis Cope, not only the incisors and canines, but also the molars are of the full number, i. e., I. ^ ; C. 1 ; Pm. ^; M. -g-. This, with the posteriorly extended expansion of the ramus of the lower jaw, distinguishes it from the other genera. But one species is known, the B. fissidens Cope, which ^See Cope, Prqceeds, Academy Philadelphia, 1883, p, 295. I 54 The Amblypoda. [January, was an animal probably as large as the Javan rhinoceros {^Rhi- nocerus sondaicus\ or rather smaller than the Uintatheriiun ro bus turn. The characters of the inferior molars in this and other genera of Dinocerata are very peculiar. In Bathyopsis they are con- structed on the plan of those of insectivorous marsupial and placental mammals, so as to lead to the suspicion that its food consisted of Crustacea, or insects of large size, or possibly of thin-shelled Mollusca. Fig. 35. — Bathyopsis fissidens Cope, mandible from right side, four-ninths nat. size. Specimen represented in Fig. 7. From the Wind River (Bridger) bed of Wyoming. Original, from Vol. iii, U. S. Geol. Survey Terrs., F. V. Hayden. The form of the ridges of the anterior part of the jaw of the Bathyopsis fissidens^ together with the remarkably large dental canal and mental foramen strongly suggest that the animal pos- sessed a large and perhaps prehensile lower lip. The lateral de- scending crests of the lower jaw must have affected the physiog- nomy curiously, especially when viewed from the front. In the history of the discovery of the various types of the Amblypoda, we have an illustration of the prevision which the palaeontologist may exercise as a legitimate inference from known facts. In closing his memoir on these animals (p. 44) Mr. Osborn remarks; "In the Upper Cretaceous or early Eocene lived a 1885.] The Amblypoda. 55 group of animals which were the common ancestors of the Dino- cerata and Pantodonta." This was written and published in t88i. In the following year, i8cS2, I discovered the Pantolambdidae in the lowest Eocene bed known in America. How well this family- fulfills the anticipations of Mr. Osborn may be seen by reference to the earlier pages of this essay on the Amblypoda (see Natu- ralist, Vol. XVIII, p. I II i). The tracing of the phylogeny of the Amblypoda from its I earliest to its latest representatives, has presented us with an inter- ! esting chapter in brain evolution. It has been asserted^ by Lartet, and repeated by Marsh, that there has been a continuous progress in the increase in the size and complexity of the brain in the Vertebrata, with the passage of geological time. This principle, as a whole, is confirmed by the results of my own studies. The Amblypoda constitute the sole exception known to me. The brain of the Pantolambda bathmodon, though of the same type as other Amblypoda, is relatively much larger than in its descendants of the Dinocerata and Pantodonta. It is a clear case of retrogression, and not of progression, in brain develop- ment. ^Comptes Rendus, June, 1868. 11 The C^eo 1884.] The Creodonta. 255 (^From the American Naturalist^ March, 1884.) THE CREODONTA. BY E. D. COPE, PRESENT knowledge justifies the generalization that, since the Eocene period, the mammalian fauna of the Northern hemisphere has diminished in the number of its species and gen- era. The Eocene fauna was richer than the Miocene ; the Miocene than the Pliocene, and the Pliocene was richer than the modern fauna. With this numerical diminution in species has come in- creased specialization of structure, which means both greater per- fection of mechanism and greater diversity of type.-^ The order of Carnivora is a universal and well-known factor of the mammalian life of the present period. It was equally so in the Pliocene and Miocene periods. When we come to examine the overflowing life of undoubted Eocene time, we can no longer find mammals which possess the essential characters of the order. The Carnivora are unguiculate gyrencephalous mammalia with a I coossified scaphoid and lunar bone of the carpus, called, there- j fore, the scapholunar bone. They have a grooved astragalus. ' No scapholunar bone has yet been found in ar>y Eocene mammal in North America, and it is doubtful whether any has been found in Europe. Nevertheless the Eocene fauna did not lack preda- tory flesh-eaters whose function was like that of the Carnivora of later periods, to restrain the undue increase of all other forms of life. Their variety was greater than that presented by their car- nivorous successors on the North American continent, and their numbers were proportioned to the general luxuriance of the life which furnished them subsistence. There were species whose size and powers of destruction equaled those of the bears, lions and tigers of modern times. Species of medium size abounded, 1 This generalization was published in the Report U. S. G. G. Survey W. of looth meridian, 1877, ly, p. 385. 256 The Creodonta. [March, while the smaller forms, representing in function the civets and weasels of to-day, were especially numerous. The systematic position of these animals has been difficult to determine with satisfactory precision, owing to the imperfect knowledge which we possess of their structure. Besides lack of scapholunar bone, they nearly all differ from the Carnivora in their ungrooved astragalus, and their greatly reduced and smooth cerebral hemispheres. Their position then can only be with the Marsupialia or the Insectivora. The superficial resemblances are often to the former order, where the carnivorous types Thylacinus, Sarcophilus, Dasyurus and the opossums, seem to present near affinities, if the structure of the teeth only is to be considered. Laurillard, DeBlainville and Gaudry have, at different times, as- signed to them this position. It is true, however, that no one osseous character, except the possession of marsupial bones, has yet been discovered which characterizes the order Marsupialia. It rests chiefly on the characters of its soft parts, especially of the brain and reproductive system. Most marsupials have the angle of the lower jaw inflected ; the Eocene flesh-eaters do not. The carnivorous marsupials generally have more than six incisor teeth ; the Eocene flesh-eaters have six or less. The marsupials generally have perforated palates ; the Eocene forms never dis- play the character. No marsupial bones have yet been found in the Eocene forms (with the possible exception of Mioclaenus), so that we cannot yet find the necessary reasons for placing our ex- tinct forms with the order which possesses them. On the other hand the brain was probably marsupial in its internal character, as it certainly was in its external character. Finally, the dis- covery of the character of the temporary dentition in the Eocene flesh-eaters has added force to the view that they cannot be mar- supials. M. Filhol has shown that in Hyaenodon there were three temporary molars, and I have proven that there were at least two in Triisod'on. There is, according to Flower, but one in the Marsupialia. It is, however, to be remembered that true marsupials, opossums, occur in the Oligocene in both North America and Europe (genus Peratherium Aym.), and that they resemble their Eocene predecessors very much, especially in the constitution and form of the molar teeth. The comparison with the Insectivora betrays no such funda- mental diversities as that with the Marsupialia. The differences 1884.] The Creodonta. 257 are of minor import, although in particular cases considerable. In spite of their often large size and evidently predatory habits, the Eocene flesh-eaters must be placed with the Insectivora. But as I have already pointed out,^ they have contemporaries which must go with them. These are two groups, the one with rodent, the other with edentate tendencies, the Tillodonta and Taenio- donta. The former of these is intimately allied to the living Chiromys of Madagascar, which itself is almost a lemur, by gen- FiG. I. — I-eft mandibular ramus of Triisodon quivirensis Cope, three-foutths nat. size, from the Puerco beds of New Mexico. Fig. a, external view, displaying the last temporary molar in place; h, the same from above; c, the same, internal side, the temporary molar removed and the permanent fourth true molar displayed in the jaw; d, the fourth permanent premolar viewed from above. Original; from Vol. IV, Report U. S. Geological Survey Terrs., F. V. Hayden in charge. eral consent. The whole of this assemblage I have regarded as an order of mammals to which I have given the name of the Bunotheria. In this order there are included six suborders, and ^ Report U. S. G. G Surveys W. of looth meridian, G. M. Wheeler in charge^ IV, p. 85, 1877. Proceedings Acadeniy Philadelphia, 1883, p. 77. 258 The Creodonta. [March, the carnivorous type of this series, the group now under discus- sion, I have called the Creodonta.^ Its definition is as follows : Neither incisor nor canine teeth growing from persistent pulps. Hallux not opposable. Superior true molars tritubercular or more simple. The only character which distinguishes the suborder from the Insectivora is the possession of tritubercular or more simple molars above. This is, as I have shown,^ a feature of more im- portance than has been hitherto supposed. The tritubercular molar is the primitive type from which the quadritubercular has been directly derived. It has furnished the starting point for both the carnivorous and the herbivorous dentitions, since it was common to both the clawed and hoofed mammals of the Puerco, or lowest known epoch of the Eocene period. So far as known, the Creodonta were all plantigrade, and had long tails, and were mostly five-toed. With the possible exception of Protopsalis tigri7ius^ they all had relatively larger heads and shorter legs than the majority of the Carnivora of the present period. This is true of all the recent Creodonta, none of which reach the dimensions of many of the fossil forms. The contents of the suborder Creodonta display, in their modi- fications, the usual range of simplicity and complexity consistent with the type, and the families may be arranged in some sort of phylogenetic order, in accordance with this principle. Neverthe- less a difficulty arises as to what are ancestral or primitive condi- tions, and what characters, if any, must be regarded as the result of degradation. As other parts of the skeleton are less frequently obtained, these considerations relate chiefly to the dentition. In my paper on the Homologies and Origin of the types of Molar Teeth, etc.,^ I pointed out that the cone was the simple form of tooth from which all others must have been derived. In the Mammalia this may have been modified in several ways sim- ultaneously, but two methods present themselves as the most cer- tain secondary primitive types. The first of these is that of the sim- ple premolar, where the cone is compressed, and is sooner or later followed by a horizontal extension or heel. This type persists in * For greater detail on this topic, see Proceedings Philadelphia Academy, 1883, P- 77- ^Proceedings American Philosophical Society, 1883, p. 324. Palseontological Bulletin, No. 37. 'Journal Academy Nat. Sciences Philada., 1874, March. 1 884.] The Creodonta. 259 the inferior true molars of the Mesonychidae. The second modi- fication consists in the addition of lateral cusps or spines to the simple cone. Such a type is seen in inferior molars of Spalaco- theriinn tricii^pidens Ow., of the Jurassic period, and in some of the incisors of Plcsiadapis tricuspidens^ Gerv. This form, by the shifting of the two subordinate cusps to the inner side of the principal one, will give a trituberculate molar of the lower series, an exaggeration of which is seen in the South African " mole,'* Fig, 2. — Mandible of Mesonyx ossifragus Cope, from the Wasatch epoch of the Big: Horn river, Wyoming, one-third nat. size. Original. From Report U. S. G. S. Terrs., Vol. ill. Chrysochloris. Tf a cingulum appears at the posterior base of such a tooth, we have a rudimental " heel," such as is seen in Centetes, and is still better developed in many creodont and mar- supial genera, forming the basis of the inferior sectorial tooth of the Carnivora. This, which I have called the " tubercular-secto- ^ Lemoine. Communication sur les Ossemens Fossiles tl. env. Rein.s, p. 9, pi. 11, fig. 13. The superior dentition of this specimen is represented in fig. 6. 26o The Creodonta. [March, rial," is especially well marked in Oxyaena (Fig. 3) and Stypo- lophus. All stages of diminution of the anterior of the two inner cusps of the inferior molars may be seen in the genera of Centetidae, as for instance in Ictops and Diacodon, until a true quadritubercu- late molar is reached as in Mioclsenus. In Miacis and Didymic- tis one tubercular-sectorial remains with one or more tuberculars. The process of developing a tubercular inferior molar from a tri- tubercular tooth, is the reverse of what has taken place in the Carnivora, where the tubercular-sectorial has, by loss of parts, Fig. 3. — Oxy!(^na forcipaia Cope, luanuiDic liom ihe Wasalch bed of New Mexi- co, one-half natural size. Original from Report U. S. G. G. Survey W. of looth mer. G. M. Wheeler, Vol. iv. The numbers indicate the approximate premolars and true molars. become a true sectorial. With these explanatory remarks I pre- sent the following definition of the families. I have formerly used the shape of the astragalus as a fainily definition. In the majority of the Creodonta its trochlear face is not grooved. In the Mesonychidae it is strongly grooved. In Mioclaenus and some species of Stypolophus, as well as in the moles, it is slightly grooved : I. Inferior molars premolariform. Inferior mtjlars consisting of a cone and heel Mesonychida. Inferior molars with a blade formed of anterior and median cusp«;, . . HycEnodontida ■ 1884.] The Creodonta, II. Inferior molars tritubercular without heel. No sectorial teeth ; tibia and fibula coossified Chrysochlorididce. III. Inferior molars tritubercular-sertorial or quadritubercular. a. No superior sectorial teeth. Tibia and fibula distinct; no zygomatic arch Centetidce. Tibia and fibula distinct ; zygomatic arch present Lcplictidce. Tibia and fibula coossified ; external tubercles of superior molars sub- conic, no clavicles Mythomyidce?- Tibia and fibula coossified ; external tubercles two Vs; clavicles TalpidcB. aa. First true superior molar sectorial. Tibia a^d fibula distinct; no tubercular molars Oxycenidce. aaa. Fourth superior premolar sectorial. Tibia and fibula distinct; true molars tubercular Miacidce. The number of genera and species embraced by these families is as follows. The extinct forms only are enumerated, and those of North America are specified : Families. Genera. Mesonychidse 4 Hyaenodontidse i Chrysochlorididse * Centetidse 18 Mythomyidoe * Talpida * Oxyaenidae , 3 Species. 7 10 44 N. America. 7 3 37 Miacidae , 80 63 Totals 28 The affinities of these families may be expressed in the follow- ing diagram, which may be regarded as an attempted phylogeny also: Insectivora Carnivora Talpidae Mythomyidae Miacidae Oxyaenidae Chrysochlorididae Centetidae Hyaenodontidae ^^^^ if, Mesonychidae Mesozoic forms. * This family is sometimes called the Potamogalidce. Mythomys i<;, however, the name first published, with a description, for the typical and only genus. 262 The Creodonta. [March, In geological time the Mesonychidae, Miacidse and Oxyaenidae are confined to the Eocene period. The Hyaenodontidae extend through the Upper Eocene and Oh'gocene or Lowest Miocene formations. The Centetidae are predominately Eocene, but in North America they are also found in the Oligocene. In recent times they are only known in the islands of Madagascar and Cuba. The Chrysochlorididae are modern and African, while the Talpidae commence in the Middle Miocene in Europe, and are as abundant in recent times in the Northern hemisphere. Mytho- myidae are only known as recent in Africa. These relations may be expressed as follows : Eocene. Miocene. Recent. Lower. Upper. Lower. Upper. Mesonychidae Hyaenodontidae... Chrysochlorididae. Centetidae Mythomyidae Talpidffi.. . Oxyaenidae Miacidae The distribution of the genera with relation to the American formations will be considered later. Mesonychid^. In this family the form of the inferior molars is not very differ- ent from that displayed by the last lower premolars of carnivor- ous Mammalia generally. The superior molars are of corres- pondingly simple structure, in like manner resembling the last premolars of the superior series of Carnivora and some other Mammalia. There are two types in the family. In the first, the astragalus is flat, as in most Creodonta (Fig. 5 a b). The only known genus is Amblyctonus, which I have, on account of its astragalus, placed in a separate family, the Amblyctonidae. This course may be justified by future discovery. The inferior molars also differ from those of Mesonyx in the development of the an- terior cusp, thus approaching the Hyaenodontidae (Fig, 4). The bones of AmbLyctomis sinosus Cope, are about as robust as those of the coyote. They have so far been found only in the Wasatch beds of New Mexico. M. Fischer has described an allied mam- 1 884.] The Creodonta. 263 mal from France under the name Apterodon gaudryi. He has not stated how this genus differs from Amblyctonus. Fig. 4. Fig. 5. Fig. 4. — Part of mandibular ramus of Ambiydonus sinosus Cope, with the last three molars ; upper figure the same, from above ; two-thirds nat. size. From the "Wasatch beds of New Mexico. Fig. 5. — Distal ends of tibia of, a-b, Ambiydonus swosus, and c d e, Oxycena morsitans Cope; b and d, distal views; two- thirds natu- ral size. P>om the W asatch beds of New Mexico. Original, from U. S. G. G. Expl. Surv. W. of looih mer., G. M. Wheeler. In the genus Mesonyx we have a structure of the astragalus found only elsewhere in flesh-eating mammals in the Hyaenodon.^ Its distal extremity has two well-marked facets, one for the navic- ular bone and one for the cuboid (Fig. 9). The appearance is that of a perissodactyle mammal, and the astragalus of the Me- sonyx might well be taken for that of an ungulate of that sub- order. The tibial face of the bone is grooved, another point in which it differs from most of the Creodonta. Still another char- acteristic peculiarity is seen in the forms of the ungues. These are neither claws nor hoofs, but between the two, resembHng the corresponding part in some Rodentia. Their deeply fissured ex- tremities show that they are rather allied to claws than to hoofs (Fig. 10). The narrow navicular facet of the astragalus renders it probable that the inner digit or hallux is wanting, and that there are but four toes in the hinder foot. The form of the trape- zium shows that there are but lour toes on the anterior foot (Fig. 7). One more character completes the singular ensemble pre- sented by this genus. The zygapophyses of the lumbar vertebrae embrace each other as in the lower perissodactyles and artio- dactyles, a character not found in the Carnivora or Insect ivora. 'Fide Professor W. B. Scott. 264 The Creodonta. [March, Among the Creodonta I have only observed it elsewhere in the genus Mioclaenus,' but the vertebrae of many of these animals are unknown. It is wanting in Thylacinus and the opossum, and rudimental in Sarcophilus, among marsupials. Three species of the genus Mesonyx are known, M. ossifvagiis Fig. 6. — Mesonyx ossifragus Cope, skull anterior to postglenoid processes, inclu- sive, from below ; one-lhird natural size. From the Wasatch beds of Wyoming. Original, from Report U. S. Geol. Surv. Terrs., ill, F. V. Hayden. Cope, from the Wasatch Eocene, and M. lanms and M. obhistdens Cope, from the Bridger. The last named is the type. It is about the size of the wolf, and had more slender feet and claws than ^Proceedings American Philos. Society, 1883, p. 543. 1884.] The Creodonta. 265 the other two. The M. lanius was considerably larger than the M. oblusidens, equaling the black bear [Ursiis americanus) in size It had a large head, with a long, rather narrow and truncate, muzzle. The limbs were relatively smaller, not exceeding those of the black bear in length and thickness. The tail was long and slender as in the cats, while the claws were broad and flat as in the beaver. The molar, canine and incisor teeth of my specimen, as well as those of one in the Princeton Museum, are much worn by use. This is especially true of the canines of both, while the crowns of the molars of my Bitter Creek specimen are almost en- tirely worn away. The same peculiarity is to be observed in the specimens of the al- lied Amblyctonus sinosus, which I obtained in New Mexico. It is probable that these species chewed hard substances. The pe- culiar approach of the lower canines is a special modification for peculiar habits. ^ , , . . . - * Fig. 7. — Mesonyx lanius vi'hich I suspect to have been the devouring left anterior foot, one-half of the turtles which so abounded on land TT"-^ "'T ^Tu nr>t row of carpal bones of and in the waters of the same period. The ri^ht side of same individ- slender symphysis could most readily be r^i^''^^' .trBrSgf; introduced into the shell, while the lateral beds of Wyoming, pressure of the upper canines with the lower, would be well adapted for breaking the bony covering of those reptiles. The breaking of these shells in the attempt to masticate their contents would produce the unusual wear of the teeth observed. The Mesonyx ossifragii^ is the largest species, its skull exceed- ing that of the grizzly bear in dimensions (Figs. 2, 6, 8, 9 and 10) It was originally fo'und by the writer in the Wasatch beds of New Mexico, and was afterwards found by Mr. Wortman to be not un- common in the corresponding formation in Northern Wyoming. From material obtained by this gentleman, we can form a general idea of the form and proportions of the Mesonyx ossifragus. We can depict an animal as large as a large-sized American black bear, with a long, stout tail and a wide head as large as that of a grizzly bear. The fore limbs are so much shorter than the hind limbs 266 The Creodonta. [March, that the animal customarily sat on its haunches when on land. In walking, its high rump and low withers would give it somewhat the figure of a huge rabbit. Its neck was about as long as that of an average dog. Its tread was plantigrade, and its claws like those of various rodents, intermediate between hoofs and claws- The animal, to judge from its otter-like humerus, was a good swimmer, although there is nothing specially adapted for aquatic Fig. 8. Fig. 9. Fig. 8. — Anterior limb hones of the specimen of Mesonyx ossifragus represented in Figs. 2 and 6, one-tliird natural size. Fig. <2, humerus from front; b, distal view of distal extremity; right ulna and radius, external view ; the same from above. Fig. 9, — Posterior limb of Mesonyx ossifragus, the individual represented in Figs. 2, 6 and 8, one-third nat. size. Fig. a, femur from front; tibia from front; as- tragalus and calcaneum from above; d, the same, distal view. From the Big Horn region of the Wasatch epoch. life in the other bones of its limbs. Its teeth, on the other hand, are of the simple construction of the mammals which have a diet largely composed of fishes. We cannot but consider this animal 1884.J The C?rodonfa. 2.6; as one of the most singular which the Eocene period possessed. In size it was not exceeded by any other flesh-eater of the period, but was equaled by the Protopsalis tigrimts. Its anterior limbs were evidently relatively shorter than in the Mesoriyx lanius. The Sarcothraiistes antiqitus is a large ani- mal from the Puerco epoch of New Mexico, of about the size of the Mesonyx ossifragiis Its inferior molars have a wide heel as in Amblyctonus. The genus Dissacus Cope occurs in the same formation and locality Its inferior molars (Fig. ii) present an ac cessory cusp on the inner side of the prin- cipal cone. This constitutes the first step towards the tubercular-sectorial tooth of other families. There are two species, the D. navajovius, as large as the red fox, and D. carnifex Cope, of larger size. Fig. II. — Dissactis navajovius Co^Q, I'ight mandibular ramus, three-fourths natu- ral size ; a, external; superior views. From the Puerco beds of New Mexico. Original. Fig. io. — Parts of dioits of Mesonyx ossi- fragns, one-half natural size. Fig. a, metacar- pal ; b, metapodial, dis- tal end ; c, phalange ; d e f, an ungual pha- lange from above, be- low and side. Original, From the Wasatch beds of Wyoming. 344 The Creodoiita, [April, {From the American Naturalist, April, 1884.) THE CREODONTA. BY E. D. COPE. f Continued from the March number, p. 26y.) Hy^nodontid^. IN this family the anterior cusp of the inferior true molars is so developed as to form, with the median cusp, a true sectorial blade. The posterior cusp is rudimental, and in the last inferior molar wanting. There is no internal tubercle. In the superior molars, on the contrary, the anterior basal cusp has disappeared, and the posterior one is developed behind the middle one, like a blade. The superior molars, like the inferior, have no internal lobe. Professor W. B. Scott has studied the posterior limb and the brain-case of a species of Hyaenodon, which he describes as fol- lows. I quote from the advance sheets of his paper on this sub- ject, which he has kindly permitted me to use : " The hind limb is in essentials very like that of Mesonyx of the Bridger Eocene ; the femur has a decided third trochanter. The tibia is much like that of Mesonyx, and its distal end is characteristically Creodont in having its astragalus face almost 1 884.] TJie Creodonta. 345 flat, with only a very slight median ridge for the groove of the astragalus. The internal malleolus is very large. ** The astragalus is but slightly concave from side to side, much less so than in Mesonyx. " The foot is plantigrade, and the entire length of the calcan- eum rested on the ground. Five well-developed digits were pres- ent, terminating in short and stout compressed claws; very differ- ent from the peculiar depressed ungual phalanges of Mesonyx ; otherwise the resemblance of the foot as a whole to that of Meso- nyx is very striking. " The brain case attributed by Gervais to Hyaenodon must be- long to some other genus, or else our American species differ very radically from the French. In the American species the brain is relatively very small and simple, being but slightly larger than that of Thylacynus, to which animal Hyaenodon presents many interesting approximations in the structure of the skull and teeth. The cerebellum of Hyaenodon is entirely uncovered by the hemi- spheres, which in their turn seem to have but three straight longi- tudinal gyri, presenting the simplest type of the carnivorous brain." It is highly probable that this family is a derivative of a pentadactyl form of the Mesonychidae. Its appearance in time corresponds nearly with the disappearance of the latter. But one genus of this family has been thus far described, the Hyaenodon of Laizer and Parieu. Its dental formula is I. f; C. }; P-m. \ \ M. 3. The last three molars in both jaws are sectorials, and the last of these are the largest, and form the most effective shears for the dividing of animal tissues. The position of these teeth indicates a mouth fissured far posteriorly, and a correspond- ingly posterior position of the masseter muscle. This structure indicates a weak power of prehension of the canine teeth. This character is sustained by the frequently anteriorly directed infe- rior canines, and the generally slender mandibular rami. The Hyaenodons must be regarded as snappers, and not capable of holding on to a living enemy with much persistency.^ They were evidently weaker in all points of organism than the modern Car- nivora, which no doubt accounts for their extinction. Thirteen species have been described, all but three North American forms being French. The oldest of these, the H. parisiensis Kefst., is from the Upper Eocene or Oligocene, or the Paris Gypsum, but its reference to this genus is not yet certain. Gaudry, however, ^ See On the origin of the sectorial tooth of the Carnivora, American Natural- IIST, 1876, p. 171. 34^ The Creodonta. [Aprils states that Hyaenodon occurs in the Gypse. Seven species are described by Filhol as from the Phosphorites. Some of these (as H. leptorhynchus L. & P.) are elsewhere found in the Stam- pian, a Lower Miocene, but some of them are probably Upper Eocene.-^ The three North American species are from the White ^ Professor Gaudry thinks the Phosphorites include fossils of different Tertiary epochs. 1884.J Tlie Creodonta, 347 River or Oligocene horizon, and no species is known from later formations. The species range in size from that of the H. vul- pinus Gerv., which equals a red fox, to that of an American black bear, as the H. Jieberti Filhol, and H. horridus \^€\^y \2\ The latter species is from the bad lands of Nebraska. Leptictid/k.^ This family is very nearly related to the Centetidse, which are now living in Madagascar. The only character by which I dis- FlG. 13. — Stypolophus whiticc Cope, skull and part of posterior foot of two indi- viduals, two thirds natural size, P'igs. a-b from the Wasatch beds of the Big Horn river, Wyo, Figs, c-d from the basin of the Wind river, Wyo, Fig. c, internal side of part of right mandibular ramus. Fig. d, left tarsus minus cuneiform bones, from above. Original, from Vol. iir, Report U. S, Geological Survey Terrs., V. V. Hayden. tinguish it is by the presence of the zygomatic arch, a part of the skull which is absent in the Centetida; (Fig, 13). The Leptictidae ^ This family is included in the Centetidse in the first part of this paper, p. 261. 348 TJie Creodonta. [April, are no doubt the ancestors of the Centetidae ; and their later types, as Leptictis, approach the existing family in their dentition quite closely.^ The earlier types display great variety in their dentition, and give ground for distinguishing many genera. Two groups are easily recognized among the Leptictidae. In the first of these the last or fourth inferior premolar is a simple premolariform tooth, different from the inferior true molars, and without any internal cusp. In the second division the fourth in- ferior premolar is either like the true molars or approximates their form by the presence of an internal tubercle. To the latter group belongs the genus Chriacus, which from the slight devel- opment of the fourth inferior premolar (Fig. 14) approximates the first division. The genus may, however, be improperly referred to the Creodonfa. Fig. 14. — Chriacus angulatus Cope, right ramus of mandible and part of maxil- lary bone with teeth ; from the Wasatch beds of the Big Horn, Wyoming. Figs, a and d twice natural size. Figs, b-c natural size. Fig. a, first and second true molars from below; b, ramus from external side; c, the same, internal side; d, the same from above. Fig. 15. — MiocliEnus turgidits Cope, part of skull and lower jaw of one animal, two-thirds natural size. From the Puerco beds of New Mexico. Origi- nal, from Vol. Ill, Report U. S. Geol. Survey Terrs. There are seven genera of the first division of the family. These may be distinguished into two sections. In one of these there are three well-developed anterior cusps of the inferior true molars, forming a tubercular sectorial tooth ; in the other the anterior ^ Enough is no\v known of the mammalian fauna of Madagascar to convince us of its decidedly Miocene and, to some degree, Eocene character. The lemurs and Centetidae approximate nearly the Eocene types; theChiromys is near the Eocene Til- lodonta, while the closest allies of the carnivorous genus Cryptoprocta are found in the Lower Miocene of France and the Middle Miocene of Oregon. TheOligocene descendants of the Eocene types appear to have persisted in Madagascar. The rep- tiles are not African but are South American. 1884.] The Creodonta. 349 cusp is rudimental or wanting, and the tooth approximates more or less the quadrituberculate condition. In the latter subsection there are three genera. The first of these, Mioclaenus Cope, has the inferior true molars quadrituberculate and of equal elevation ; the first true molar may have an anterior or fifth tubercle. The external cusps of the superior true molars are distinct and coni- cal. In Triisodon Cope, the inferior true molars only are known. These have four cusps with a rudimental anterior fifth. They dif- fer from the corresponding cusps in all the other genera in being compressed so as to have fore and aft cutting edges. Diacodon Cope, is the third genus. Its superior molars are like those of Mioclaenus, but the two anterior cusps of the lower true molars are much elevated, the posterior are rudimental, and there^is a rudimental fifth in front. Mioclaenus presents the only truly quadritubercular lower mo- lars in the suborder. It is so far known only from the Puerco or oldest Eocene of North America. There are nine species known Fig. 1 6. Fig. 17. Fig. 16. — Mioclcenus corrugatus Cope, last four molars on maxillary bone from below, two-thirds natural size. From the Puerco beds of New Mexico. Fig. a, from below; b, from right side. FiG. 17. — Mioclcenus szibtrigonus Co'pQ, Y>^xi?, of maxil- lary and mandibular bones two-thirds natural size; from the Puerco beds of New Mexico; a, superior true molar teeth from below; b, left mandibular ramus, external side; do. inner side ; d, from above. From Report U. S. Geol. Survey Terrs., Vol. 111. so far, which range from the size of a mink {M. minimus) to that of a wolf {M.ferox) in the sizes of their jaws, but in the case of the M.ferox, of which a good deal of the skeleton is known, the body was relatively smaller. The species differ in the form of the third superior premolar, and in the robustness of their inferior premolars. The M. turgidus (Fig. 15) represents the type with robust premolars, and the M. subtrigonus (Fig. 17) those with more compressed premolars. 350 The Creodonta. [April, We can read the nature of the primitive mammal, Mioclceniis ferox, in so far as the materials permit. It was an effective flesh- eater, and probably an eater of other things than flesh. It had a long tail and well-developed limbs. It had five toes all around, and the great or first toe was not opposable to the others, and may have been rudimental. The feet were plantigrade and the claws prehensile. The fore feet were well turned outwards. There were, perhaps, marsupial bones, but this point is not yet certainly determined. The presence of a patella distinguishes it from mar- supials in general. Its embracing glenoid cavity of the skull, and form of the inferior molars, resemble those of the Arctocyonidae. This species is about the size of a sheep. The bones are stated by IV^r. Baldwin, who discovered it, to be derived from the red beds in the upper part of the Puerco series. The genus Tri'isodon includes as yet but one species, the T. quivirensis Cope, which is only known from the rami of the lower jaw. These bones are shorter and more robust than those of the coyote, and indicate an animal of perhaps the size of the, wolver- ine (Fig. I, p. 257). It was evidently strongly carnivorous in its diet, and was a capable biter. Its remains are from the Puerco of New Mexico. Diacodon includes seven species, two from the Wasatch Eo- cene, the others from the Puerco. The former are much the smaller (Fig. 18), while those of the Puerco vary in dimensions from the size of a common weasel (D. assiirgens) to that of a wolf (^D. co?iidens). The five spe- r r cies of the Puerco were formerly re- . — Jaws of species of Diacodon from the w^asatch ferred by me to the genus Triisodon, ^t.lf:/Atcop::'?igh; but are now more properly placed in mandibular ramus, inner side, Diacodon. The Only species in which twice natural size. Fig. . , 1 • ^.i n D. ceiatus, left mandibular the Superior molars are known is the D, ramus, natural size; c, same, conideus, where they are generally iden- twice natural size. Original, . , -it r -a/r- 1 from the Report u. s. G. G. tical With those of Mioclaenus. Surv. w. of looth mer., G. M. Qf the four genera of the first divi- Wheeler. ^ sion of the Leptictidae, which possesses tubercular sectorial teeth, but two are found in North America, while three of them have been discovered in Europe. The typical and most widely distributed genus is Stypolophus 1884.] The Creodonta. 351 Cope (Fig. 13). It has compressed premolars, except the fourth superior, which is conic with two basal lobes. In Proviverra Riitimeyer, this tooth is triangular and cutting. One species has been found in the Swiss Eocene. Quercitherium of Filhol is dis- tinguished by its very robust premolars. Like Proviverra it has but one rather small species. It is from the French Phosphorites. Stypolophus has the two cusps of the superior true molars close together, forming twin cusps, and they have behind them a heel, which is cutting. Two species have been discovered in the French Phosphorites, one of which, the 6*. cayltisi of Filhol, is preserved to us in the most perfect skull of a Creodont known. From it I have restored the skull of the 5. whitics (Fig. 13). It has elevated sagittal and posterior crests for the insertion of the temporal muscles, and the brain-case is very small. A cast of the brain displayed to Mr. Filhol the following characters : The hemispheres are small, and leave the cerebellum and the posterior ^,]?i\^iW7i\.\xxd\ size, from the Puerco forma- tion of New Mexico. Fig. d, maxillary bone with teeth, from below ; e, left ramus mandibuli, inner side; f, do., from above. All original, and from the Report U. S. Geol. Survey, Terrs., F. V. Hayden in charge, Vol. III. (Fig. 29), have a short subcircular tubercular molar, while that of 1884.] TJie Creodonta. 485 the other species is an elongate oval. The tendency in the genus has been to complicate the premolars and shorten the tubercular in the course of time. The smallest species is the D. daivkinsi- anus (Fig. 30), from both the Big Horn and Wind River beds, an abundant and acute-toothed species. The largest species is the D. altidens Cope, whose jaws are more robust than those of the coyote. I append the followmg table showing the distribution of the genera of Creodonta in the North American Tertiary forma- tions OLIGOCENE Amblyctonus Mesonyx Sarcothraustes . . Dissacus Hysenodon . . . . , Mioclsenus , Triisodon Diacodon Stypolophus Didelphodus . . . Chriacus Deltalherium. . . Ictops Mesodectes .... Leptictis Esthonyx Oxygena Protopsalis Patriofelis Miacis Didymictis Printed April 24, 1884. (^From the American Naturalist, May, iSSj.) THE LEMUROIDEA AND THE INSECTIYORA OP THE Eocene Period of Nortlj America. SIT ID. COI^E. Printed April 15, 1885. I 1885.] Lemur oidea and the Inseciivora of the Eocene Period y etc, 457 {From the American Naturalist, May, i88^.) THE LEMUROIDEA AND THE INSECTIVORA OF THE EOCENE PERIOD OF NORTH AMERICA. BY E. D. COPE. TWO distinct divisions are included in this article, because the material is not yet sufficiently complete to enable me to refer certain forms to the one rather than the other. The only charac- ters on which the osteologist can rely in endeavoring to distin- guish the two groups are these : First, the terminal phalanges of the Insectivora are compressed and curved, forming claws; while* those of the Lemuroidea and of most other Primates are more or less flat, and at the extremity rounded and depressed,^ or more or less like hoofs.^ Second, the hallux or inner toe of the posterior foot is opposable to the others, a character dependent on the form of the entocune'iform bone of the tarsus, which has in that ^See American Naturalist, April, 1885, where the Condylarthra are referred, with the Quadrumana, to the Ungulata. ' The marmosets are exceptions, having true claws. 458 The Lemur oidea and the Insectivora of the [May, case a rounded distal extremity, forming part of a cylinder directed more or less fore and aft, for articulation with the meta- tarsus or proximal element of the great toe. In the Insectivora this structure is wanting, the inner toe being fixed in a position parallel with the others as in the Carnivora. In the Lemuroidea the position of the thumb or pollex is less different from what is seen in the Insectivora, than is the case with the posterior foot. In the true lemur the thumb is but little opposable, except in the genus Chirogaleus and some others. The distal end of the tra- pezium bone of the carpus with which the thumb articulates, does not form a part of a cylinder in the Lemur or in the Tarsius. When the thumb becomes opposable in the monkey proper, the thumb facet of the trapezium is not rounded, but is wide and a little concave. It is not till we reach the man-like opposable thumb of the anthropoid apes that we find this bone presenting to the thumb a semicylindrical face like that of the entocunei- form bone of the posterior foot. The Condylarthra as I have pointed out,^ must be regarded asi a division of the order of Taxeopoda, along with the Hyracoidea, the lemurs, the monkeys and man. The difference between the hoofs of Phenacodus and the ungues of Lemur is too slight to admit of wider separation : and the other parts of the structure show an equal agreement. There is no trace of opposability of the hallux in Phenacodus however, nor any os centrale of the car- pus, characters which show that the suborder Condylarthra and Lemuroidea are distinct. In the pollex or thumb of Phenacodus, however, there is a distinct indication of opposability, though it is not so well developed as in the genus Lemur. The basal artic- ulation with the trapezium is narrow, but is directed partly fore and aft, so that the thumb looks inwards. Its power of flexure at the base has been slight, but the flexure at the base of the first phalange has been such as to make the end of the thumb quite ' opposable.^ From the Condylarthra then we trace the order Quadrumana on the one hand, and the hoofed orders on the other. In the following pages I will not attempt to distinguish which of the genera are lemuroid and which are insectivorous, since the ungual phalanges are yet unknown. An exception must be made in the case of the genus Pelycodus, where a single compressed 1 Naturalist, April, 1885. Primates and Taxeopoda are there regarded as nearly synonymous. 2 See Naturalist, 1884, Plate xxix, for the skeleton of Phenacodus. 1885.] Eocene Period of North America. 459 acute claw is known. This alone does not decide the question, since such a claw exists on the second toe of many Lemuroidea. These animals are readily distinguished into three divisions or families by the number of their premolar teeth. There are four such teeth in the Adapidae ; three in the Mixodectidae, and two in the Anaptomorphidae. In the Adapidae we have the most primitive type, and the one most nearly allied to the Condylar- thra, from which they were probably derived. In the Mixodec- tidae we have the dental formula of the existing lemurs, with a tendency in some of the genera to develop large cutting teeth in the position of incisors, thus approaching the aye-aye. In the Anaptomorphidae, on the other hand, we find a dental for- mula like that of the Simioidea and Anthropoidea, or higher Fig. I. — Adapts parisiensis Cuv., skull nat size, from the Phosphorites of Central France. From Filhol. monkeys and apes ; and in them we seem to get a hint of the derivation of these higher forms, and of man himself. The genera of the Adapidae are distinguished by various den- tal characters. Such are the presence of a second anterior-inner cusp of the inferior true molars ; the presence of an internal cusp of the fourth inferior premolar; the number of incisor teeth, and number of single-rooted premolars. The difference between the quadrituberculate and the quinquetuberculate inferior molar may be understood by reference to Fig. 2, where the teeth of the gen- era Hyopsodus (a) and Microsyops (^), which represent the two types, are placed side by side. In some genera, (Notharctus, Tomitherium, Figs. 4-5) the fifth cusp is present but weak. In others (Sarcolemur) it is repre- 460 The Lemuroidea and the Insectivora of the [May, sented by the anterior lobe of a twin or fissured anterior inner Fig. 2. — Inferior molar teeth of (a) Hyopsodus paulus Leidy, and Microsyops gracilis Marsh; enlarged four times linear, from above. Fig. ae, anterior external cusp; pe, posterior external cusp; az, anterior internal and pi, posterior internal cusp; 5 fifth or second anterior internal cusp. From Leidy, Report U. S. Geol. Surv. Terrs,, F. V. Hayden in charge. Vol. I. cusp. To simplify the understanding of these differences, I give the following table : I. Inferior molars qadrituberculate. Fourth inferior premolar with internal cusp : cusp on last molar opposite Hyopsodus Leidy Fourth inferior molar without internal cusp; cusps opposite Apheliscus Cope.^ Cusps of last molar alternate Opisthotomus Cope. II. Inferior molars quinquetuberculate. a. Anterior triangle not well developed on inferior molars. Fifth cusp separated from anterior inner by an apical fissure only. .Sarcolemur Cope. Fifth cusp separated ; canine distinct; one premolar one-rooted. . Notharctus Leidy. Fifth cusp well separated ; canine distinct ; two premolars one-rooted Tomitkej'ium Cope. Fifth cusp separated, low; canine incisor- or premolar-like Adapis Cuv. aa. Anterior triangle well developed on all the inferior molars. Canine distinct; one premolar one-rooted Pelycodus^ Cope. Jaws of four species of Hyopsodus are abundant in the Wa- satch and Bridger Eocene beds, and a species from the Puerco has been doubtfully referred to it. The ^^^C^^^^PP^^ best known species, the H. paidus Leidy ^^^^^^pr (Fig. 2), of the Bridger epoch, has the I jaws as large as those of a rabbit. The H. vicarius Cope, was smaller (Fig. 3). Nothing is known of the skeleton of any Fig. Z'— Hyopsodus vicarius specics of HyoDSodus. The only spe- Cope, jaws, from the Wind \ r a u r t a ' 'j- n \ River (? Bridger) Eocene of Cies of AphellSCUS {A. tnSldtOSUS CopCj Wyoming, natural size ; «,su- ^^s found in the Wasatch beds of New perior, inferior dental series. Original, from Report u. s. Mexico. It has large teeth in the posi- Geol. Surv. Terrs., iii. ^^^^ sectorials, and may be an aber- rant Creodont. Nothing is known of it but jaws. Two species * Of uncertain reference to this family and order. I88S.J Eocene Period of North America. 461 of Opisthotomus are known from the same horizon and locality, from teeth only. The O.flagrans Cope is, with the Adapis mag- niis Filh., the largest species of the family. Sarcolemur Cope in- cludes a single species from the Bridger beds, of the size of the Hyopsodus paulns. It has in its sharp dental cusps an effective biting apparatus. In Notharctus Leidy, the fifth lobe of the true molars begins to be apparent, though it is only present in the first molar, where it is represented by the in- jternal extremity of an I anterior crest. The ca- jnine in this genus (well developed IS Fig. 4.- Nothardus tenebrosus Leidy. «, mandible Onlv ^"^^"^ "g^t side, natural size. Fig. b, true molars from above, twice nat. size, linear. From Bridger one species is certainly bed of Wyoming. From Leidy, Report U. S. Geol. l^nown. ^^''^^y ^- H^y^^"' Vol, I. More of the skeleton lis known in the genus Tomitherium Cope than in any other one of Ithe family, and its rela- tionship to the lemurs was thus indicated at the time of its original de- scription in 1872. Un- fortunately the ungual phalanges remain un- known. As in Hyop- sodus and Pelycodus, there are but two infe- rior incisors in the low- er jaw, and these have transverse cutting edges, and are not pro- duced as m recent Y\G, ^.— Tomitherium rostrattim Cope, mandible lemurs. natural size; a, from left side ; ^, from above. Let- r , ' • ters as in Fig. 2. The first impression survey Tens., Vol. 462 The Lemur oidea and the Insectivora of the [May, derived from the appearance of the lower jaw and dentition, and from the humerus, is that of an ally of the coati (Nasua). The humerus indeed, is almost a fac-simile of that of Nasua, the only difference being a slight outward direction of the axis of the head. The same bone resembles also that of many marsupials, but the flat ilium, elevated position of dental for- amen, and absence of inflection of the angle of the lower jaw, etc., render affinity with that group highly improbable. Fig. 6.— Tomitkerium rostratum Cope, fore leg of animal represented in Figs. 5-7, nat size. Fi^. a, humerus; b, ulna; ^, radius, from front; b\ from side; c" ^ proximal end (artificially flattened below) ; c'^', distal end of radius. From Bridget beds of Wyoming. Original, from Report U. S. Geol. Surv. Terrs., Vol. ill. The length of the femur indicates that the knee was entirely free from the body as in the Quadrumana, constituting a marked distinction from anything known in the Carnivora, in- cluding Nasua. The round head of the radius indicates a com- plete power of supination of the fore foot, and is different in form from that of Carnivora, including Nasua ; and, finally, the distal I88s.] Eocene Period of North America. 463 end of the radius is still more different from that of Nasua,and resembles close- ly that of monkeys of the genus Semnopithecus. We have, then, an ani- mal with a long thigh free from the body, a manus capable of complete pro- nation and supination, and details of lower jaw and teeth quite similar to those of the lower monkeys. The form of the humerus and its relative length to the femur are quite as marked as in some of the lemurs. The most marked differ- ence is seen in the increased number of teeth; but in this point it relates itself to the other Quadrumana, as the most ancient types of Carnivora and Ungulates do to the more modern. This genus is allied to A dap is Cuvier, of the French Eocene (Fig. 8), but differs in the posses- sion of but two incisors on each side ; in Adapis there are three, according to Fil- hol. From that genus and Opisthotomus, it differs also in the structure of the last inferior molar, as exhibited in the analytical table. There are several species of Tomitherium, but the a Fig. 7. — Tomithef turn rostratum Cope, part of animal figured in Figs. 5-6, nat. size. Fig. fl, ilium inner side ; b, femur front ; do. posterior side. '464 The Lemuroidea and the Insectivora of the [May, best known is the T. rostratum from the Bridger formation of Wyoming. The following points may be gained by comparison with the skeleton of Lemur collaris (catalogue Verreaux). There is con- 'siderable resemblance in the details of structure of the molars from the third to the sixth, inclusive. Of course the anterior teeth differ widely in the two, and the last true molar of the Lemur has no heel. The principal difference in the humeri is seen in the superior size of the epi condyles of the 71 rostratum ^ and the rather more robust character of the shaft. The proximal half of the ulna is deeper, and the olecranon is not so wide in T. rostratum. The proximal part of the radius is very similar in the two species, but the distal extremity is in the T. ros- tratum less transversely extended, and thicker anteroposte- riorly. There is also much similarity in the ilia. The crest is more extensive in T. rostratum, and the inferior border is thinner at its proximal part. Towards the acetabulum the increase in width of this border is similar, and the anterior inferior spine is as prominent. The resemblance between the femora amounts to identity of character; that of the T. rostratum is more robust. The Mixodectidae include four and perhaps five genera. In Fig, 8. — Necrolemur antiquus Filh., skull natural size, from Phosphorites of Cen- tral France. From Filhol Rech. s. les Phosph. de Quercy. two of these the incisors have their usual position and space. One of these Tricentes Cope, has large canines well separated. It is uncertain whether the genus should not really be referred to the Creodonta.^ It contains three or four rather small species from the Puerco formation of New Mexico. In Necrolemur the canine is insignificant. One species of the genus, the N. antiquus Filhol, is known. It is represented by a cranium in excellent preserva- tion (Fig. 8) which has been fully investigated by Filhol. This able palaeontologist regards it as most nearly allied to the genus Galago now existing in Africa. It furnishes conclusive evidence of the former existence of lemurs in France. iSee Naturalist, 1884, p. 353. 1885.] Eocefie Period of North America, 465 Of the three genera with very large incisor (? canine) teeth, Mixodectes has the last lower premolar with a simple cusp. There are two species from the Puerco beds. The smaller of these, M, pimgens Cope (Fig. 9), is about the size of the kit fox. Its premolars are of irregular size. In the two other genera the fourth premolar has a second cusp on the interior side of the principal one. Both have the crowns of the inferior true molars composed of two triangles as in Mixodectes and Pelycodus. In Microsyops Leidy, there is but one one- rooted premolar. There are three spe- . , TTT .1 t -i-^ • 1 t t Fig. q. — Mixodectes pun^ens cies from the Wasatch and Bridger beds. Cope, lower jaw right ramus. The type is the M. frracilis Marsh (Fig. natural size. Fig. «^ from . / , , T 11 • above. Original, from Vol. Ill 2b)y from the latter. It was a small am- Report u. S. Geoiog. Survey mal, not exceeding a gray squirrel in dimensions. In Cynodontomys the premolar teeth are more reduced in size than in any of the allied genera, two of the three being one-rooted. The large ? incisor tooth has a correspondingly large devel- opment. The species was found by Mr. Wortman in the Wasatch beds of the Big Horn basin, Wyoming Terr. The most evident lemuroids yet found in America belong to the family of the Anaptomorphidae. But one genus is fig. lo.^Cynodontomys lat- certainly known to belong to it, Anapto- ^^^'^^ Cope, mandible, nat. Till ^'^^» from Wasatch beds of morphus Cope. The genus Indrodon^ Wyoming. Fig. a\ from resembles it in dental formula excepting ?!^°y% Original, from Report ^ ° U. S. Geol. Surv. Terrs., Vol. in the possession of three instead of two m, F. v. Hayden in charge, incisors. It embraces but one species, /. malaris, which was found by David Baldwin in the Puerco formation of New Mexico. Anaptomorphus was founded on the lower jaw of a small spe- cies, A. (Emulus Cope, which does not exceed that of a ground squirrel (Tamias) in size (Fig. 11). It agrees with a very few of the living lemurs (Indrisinae) in the number of its teeth, but it differs from them all in having short erect incisor teeth as in the higher ^ monkeys. The molar teeth known are a good deal like those of 466 The Lemuroidea and the Insectivora of the [May, the true monkeys in character, being quadrituberculate. The last premolar is quite different, having a compressed, simple, cut- ting crown. The ca- nine is quite small, and there is no diastema. The evidence furnished by this jaw was hap- pily supplemented by the discovery, at a later day, of an almost en- tire cranium of a close- ly allied species in the Fig. w.—Anaptomorphus amulus Cope, left ramus w^icatrli hf^H^ nf W^r of mandible, twice natural size, linear ; inner side; WaSatCH DedS Ot Wy- Q ; par^ of right mandibular ramus, natural size. Fig- 5a, right side; b, left do.; c, from above; d, posterior view, showing masseteric and dental foss?e. From the Lower Puerco of New Mexico. Original, from the Report of the U. S. Geol. Survey Terrs., F. V. Hayden, Vol. III. beds marks an important advance in 1884.] The Tertiary Marsupialia. 691 of one of the most interesting of living forms. On the other hand, the Polymastodontidae may well have derived their origin from the Tritylodontidae, which were also of herbivorous or granivorous habits. The family of the Plagiaulacidae^ is one of the most peculiar among those of the Mammalia, whether we consider its structure or its relations to geological time. Com- mencing in the Jurassic period, it persisted through the Creta- ceous to the Eocene. It then disappeared from view to remind us once more of its existence by its probable descendant, the ex- traordinary pouched lion of the Pliocene period of Australia, Tliylacoleo carnifex Owen (Fig. 9). The family exhibits the usual successional relation of its component genera. In this respect it repeats what I have already pointed out as a law of succession in placental Mammalia,^ a reduction in the number of premolar teeth. The following table exhibits these relations : I. Tubercles of superior molar crescentic. Fourth premolar serrate, not ridged Meniscoessus. II. Tubercles of molars subconic. a. Four compressed premolars below. Premolars serrnte, not ridged Ctenacodon. Premolars ridged and serrate Plioprion^ aa. Three compressed premolars. Premolars ridged Flagiaulax. aaa. Two premolars. Fourth premolar ridged Plilodits, aaaa. One premolar.* Fourth premolar ridged Neoplagiaulax. Fourth premolar smooth Liotovius.^ Of these seven genera but nine species are thus far known. Ptilodus and Ctenacodon have two species each ; and each of the others but one. Ctenacodon, Plagiaulax and Ph'oprion are Juras- sic ; Meniscoessus is Cretaceous, and the remaining three genera are Eocene Tertiary. The American genera are Ctenacodon Marsh (Fig. 7 ^), Meniscoessus Cope, and Ptilodus Cope. The first named is the most generalized of the family. The Menisco- essus conquistus Cope, has the distinction of being the only known mammal of the Cretaceous period. 1 This family is ihe equivalent of Marsh's " order" Allotheria. ' Bulletin U. S. Geol. Survey Terrs., vi, p. 168. Gen, nov., type Plagiaulax minor Falconer, * Number unknown in Liotomus. *Gen. nov., type Neoplagiaulax marshii Lemoine. 692 The Tertiary Marsuptaliu. [July, The genus first discovered was the Plagiaulax of Owen, of which the typical species was found in the Purbeck bed of the Isle of Wight, England. It was made the subject of a memoir by Falconer. Ctenacodon was next discovered in the Jurassic of Wyoming by Marsh in 1879. In 1880 Dr. Lemoine discovered the first Tertiary representative of the family, and in 1882 named it Neo plagiaulax eoccetius (Fig. 6). In November, 1881, I de- scribed the first American Tertiary form, which was discovered in the Puerco beds of New Mexico, under the name of Ptilodus Fig. 6. — Neoplagiaulax eoccpnus Lemoine, from the Cernaysian beds of Reims, France; mandibular rami and teeth of three individuals which are represented by the letters a, b and c. Figs. a. b, and c, much enlarged. Figs, a' , b' and c\ natural size. Ficr. from above. From Lemoine, Bulletin de Soc. Geol. de France, 1882, p. 249- medicevus (Fig. 8). Its presence in that formation, together with various other associated types, proved the near homotaxy of the beds explored by Dr. Lemoine near Reims with those of New Mexico.^ Up to this time the great Cretaceous period had remained a blank in the history of Mammalian life. European palaeontolo- gists had examined the fresh-water beds of this period for mam- malian remains without success. Among them, the gifted Kowalev- 1 See the Naturalist, 1883, p. 870. 1884.] The Tertiary Marsupialia. 693 sky, too soon lost to science, spent much time in the south of France among the formations which most nearly represent the American Laramie formation, but found no Mammalia. It remained for Mr. Wortman to crown a series of successful expedi- tions by the discovery of the Meniscocssiis conqnistiis in the Laramie formation of Da- kota, its loose teeth being found mixed with the teeth of dinosaurs and scales of gar-fishes. The characters of the molar teeth are highly appropriate to the geological age of the genus, the supe- rior molar resemblinhg bot that of the Jurassic Stereog- nathus and the Eocene Poly- mastodon. I have shown that the cut- FiG. 7. — Fig. a, Ctenncodon serratus Marsh, \ nat. size, from Marsh. Fig. molar of Stereognathus odiidctis, -f-, from Owen. Fig c, right fourth upper molar of Tritylodon longa, viis, from Owen, ^. Figs, d-f, Mcniscoessus ting tooth of the lower jaw conquistus Cope, I nnt. size. Fig. d, supe . -11 '"^^^ molar; Figs, e-f, superior fourth pre- m the genera with but one m„lari; g, humeral condyles of a smaller such tooth, as PtiloduS and species with jaw, f.mnd with the .T/^/^'jr^^mw.r. Thylacoleo, is the fourth premolar ;^ while the similar tooth in the existing kangaroo-rats, with which it has been compared, is a third premolar. In the living genera the fourth premolar resembles a true molar. It is necessary to remember this fact in the attempt to ascertain the phylogeny of the Multituberculata. This is not an entirely easy task, owing to the questions which arise as to the origin of the cutting premolars themselves. In general it is true of Mammalia that simple premolars precede the complex in time; but an exception to this rule is to be seen in the tritubercular superior sectorial tooth of some Creodonta and Car- nivora. Whether the premolars of this family are primitive or derivative is not as yet known. If they be primitive they may be direct modifications of the serrate teeth of the herbivorous Dinosauria or Theromorpha. The complex character of the pre- tnolars in the older Tritylodon suggests the possibility of the other alternative. The general history of the Plagiaulacidae con- *This tooth may possibly belong to a Saurian. 2 Naturalist, 1882, o. 521. 694 The Tertiary Marsupialia. [J"iy. firms the theory of derivation from complex premolars, and we observe in the later form, Thylacoleo, a simplification of the true molars also. The molariform fourth premolars in the existing Macropodidae confirm this view. In order to connect these latter with the ancestral form of the Plagiaulacidae, on a former occasion, I posited a theoretical form which should com- bine three of the cutting premolars of the extinct family with the molar- iform fourth premolar of the Macro- podidae. This I named Tritomodon.^ The discovery of Trity lodon has added confirmation to this view, at least as re- gards the prior existence of the molari- form premolars. The Polymastodon- tidae were probably derived from the Tritylodontidae by the usual process' of reduction of number of teeth and specialization of those that remain. Thylacoleo^ must be regarded as the type of a family distinct from the Plagiaulacidae, since it has but one true molar tooth in the upper jaw. The composition of that tooth is unknown, so that it is not certain whether the family Thylacoleontidae must be placed in the Multituberculata or Sar- cophaga. That it is a direct descendant of the Plagiaulacidae I think there is no doubt. The following phylogenetic scheme is similar to one I published in the Naturalist, 1882, p. 521, with some addition, and the removal of Polymastodon (Catopsalis) from the Plagfiaulacidae : Fig Cope, — Ptilodiis mediavus left mandibular ramus, nal. size ; from the Upper Puerco beds of New Mexico. Fig. external side; b, internal do.; c, superior view, \ nat. size. Orig- inal, from Report U. S. Geol. Survey Terrs., ill, F. V. Hayden in charge. ^Naturalist, 1882, p. 521. ^ Owen, Quar. Journ. Geol. Society, London, 1883, 1884 ] The Tertiary Marsnpialia. 695 Tritylodon Thylacoleo It appears from the preceding considerations that the dentition of the implacental Mammalia has had a history independent of that of the placental series so far as regards the herbivorous types at least. I have shown that the primitive types of the pla- cental series were tritubercular, and then quadritubercular, and then crested. In the herbivorous marsupials, on the contrary, we commence with multitubercular forms, and it is yet an open ques- tion whether these have had a quadri- and tritubercular ancestry or not. The Plagiaulacidae of the Jurassic period are of very small size, none of them exceeding in dimensions the house mouse. The same is true of the species of the Eocene period hitherto found in Europe. The American species are larger, the Ptilodus medicBvus equaling probably the Norway rat (Fig. 8), while the P. trove ssartianus is one-third smaller. The Meniscoessus conquis- tus is still larger, equaling about the Polymastodon foliatus. (Fig. 7). The arrangement of the crests of the fourth premolar in the species of Plagiaulacidae differs as follows : In Plagiaulax, Plio- prion and Neoplagiaulax this tooth is grooved. In Ptilodus the grooves have become so wide (Fig. 8) that the wide intervening ribs have become narrow keels. In Meniscoessus there are no keels, but the margin of the crown is serrate (Fig. 7 a c). TJic Tertiary Mai supialin . The Ptilodus medicBims further differs from the Ncoplagiaulox eocceniis in the more rodent-hke cliaracter of its incisor teeth. In the latter species these teeth resemble more those of the kanga- roos in their anterior direction. The diastema is longer in Ptil- odus, thus increasing the rodent resemblance. The fourth pre- molar is strongly serrate in the Neoplagiaulax, resembling in this al o the Mtsozoic types. The discussion between Professor Owen on the one side, and Fig. 9. — Thylacoleo carnifex Owen; skull from b^low, one half nat. size ; from the Pliocene beds of Australia. From restoration by Professor Owen in Geological Magazine, 1 883, p. 289. Messrs. Falconer, Krefft and Flower on the other, as to the nature of the food of Thylacoleo. is known to palaeontologists. From the form of the teeth alone. Professor Owen inferred the carniv- orous nature of the food of this genus, while his opponents in- ferred a herbivorous diet from the resemblance between the den- tition and that of the herbivorous Hypsiprymnus. I have pointed out (/. r.) that the comparison of Thylacoleo with Hypsiprimnus (I 1 884 ] The Tertiary Marsttpialia. 697 is weakened by two considerations : First, the cutting teeth in the two genera are not homologous ; second, the grinding series of molars, complete in Hypsiprymnus, is almost wanting in Thy- lacoleo. It evidently does not follow that because Hypsiprymnus is herbivorous Thylacoleo is so also. Professor Flower refers to the reduction of the molars in Thylacoleo as slightly compli- cating the problem, and concludes that the food of that animal may have been fruit or juicy roots, or even meat. It is difficult to imagine what kind of vegetable food could have been appro- priated by such a dentition as that of Ptilodus and Thylacoleo. The sharp, thin, serrate or smooth edges are adapted for making cuts and dividing food into pieces. That these pieces were swal- lowed whole is indicated by the small size and weak structure of the molar teeth, which are not adapted for crushing or grinding anything but very small and soft bodies. It is not necessary to suppose that the dentition was used on the same kind of food in the large and the small species. In Ptilodus medicevus the diet may have consisted of small eggs which were picked up by the incisors and cut by the fourth premolars. In Thylacoleo carnifex it might have been larger eggs, as those of the crocodiles, or even the weaker living animals. The objection to the supposition that the food consisted of vegetables, is found in the necessity of swallowing the pieces without mastication. In case it should have been of a vegetable character the peculiar premolar teeth would cut off pieces of fruits and other soft parts as suggested by Professor Flower, but that these genera could have been herbivorous in the manner of the existing kangaroos, with their full series of molars in both jaws, is clearly an inadmissible sup- position. Printed June 19, 1884. THE BATRACHIA OF THE PERMIAN PERIOD OF NORTH AMERICA. By prof. e. d. cope. From the American Naturalist, January, 188 4.. J 26 Batrachia of the Permian Period of North America. [January, {From the American Naturalist^ yanuary, 1884 ) THE BATRACHIA OF THE PERMIAN PERIOD OF NORTH AMERICA. BY E. D. COPE. THE class Batrachia holds an important position in the history of the Vertebrata, as the first member of that kingdom which occupied the land on the advent of the conditions suitable for air-breathing types. It thus stands in ancestral relation to the lines of the Sauropsida and Mammalia, and as the immediate de- scendant of the fishes. There are several orders of Batrachia, and they display remark- able diversities of skeletal structure. For the better understand- ing of these, I give the following table of their principal defi- nitions '} I. Supraoccipital, intercalary and supratemporal bones present. Propodial bones distinct. Vertebral centra, including atlas, segmented, one set of segments together support ing one arch , Rhachitomi. Vertebrae segmented, the superior and inferior segments each complete, forming two centra to each arch Embniomeri Vertebral centra, including atlas, not segmented; one to each arch Stegocephali. II. Supraoccipital and supratemporal bones wanting. Frontal and propodial bones distinct. a An OS intercalare. A palatine arch and separate caudal vertebrje Proteida^ aa No OS intercalare. A maxillary arch ; palatine arch imperfect ; nasals, premaxillaries and caudal ver- tebrae distinct Urodela, Maxillary and palatine arches distinct; nasals and premaxillaries united Gymnophiona. No maxillary or palatine arches ; nasals and premaxillary, also caudal vertebrae, dis- tinct Trachystomata ITI. Supraoccipital, intercalare and supratemporal bones wanting. Frontals and parietals connate ; propodial bones and caudal vertebrae confluent. Premaxillaries distinct from nasals ; no palatine arch ; astragalus and calcaneum elongate, forming a distinct segment of the limb Anura The animals of the division i are all extinct. Division 11 in- cludes the salamanders and their allies, with the worm-like Cce- cilians (Gymnophiona) ; while the third division embraces the frogs, toads, etc. ^This is partly derived from the table which I have given in Vol. II Palaeontol- ogy of the Geological Survey of Ohio, 1874, p. 352. See also Proceedings Phila- delphia Academy, 1868, p. 211. 1884.] Batrachia of the Permian Period of North America. 27 The characters displayed by the three divisions in question, indicate their relationships to be as follows : The orders of division i present in their cranial structure a greater resemblance to the limb-finned or crossopterygian fishes, than do either of the others. The third division is the most divergent from that type, and is in various respects the most specialized. This specializa- tion consists not only in a departure from the primitive Batrachia, but also from all other forms of Vertebrata. Its specialization is seen in the loss and coossification of various parts of the skele- ton. The Urodela display characters intermediate between the extremes of the class. Near them the Trachystomata (Sirenidse) are inferior by loss of parts of the skull, and of the pelvic arch, the result, probably, of a process of degeneration. The same is probably true of the Proteida, which have lost the maxillary arch of the Stegocephali, but retain their os intercalare.^ As regards the extinct orders, the primitive type is evidently the Rhachitomi, whose vertebral column displays an arrest o^ characters which are transitional in the higher Vertebrata. From this group the orders Embolomeri and Stegocephali have evi- dently been derived. We may then present the following gene- alogical table of the class Batrachia : Anura Trachystomata Embolomeri Stegocephali Rhachitomi As regards the connection of the class, as a whole, with other classes of Vertebrata, it is very probable that the extinct orders, as the Rhachitomi, were derived from some extinct form of Dip- noan fishes more or less related to the group of which the genus Ceratodus is a representative. In this type we have a persistent chorda dorsalis ; fins which present the type from which ambulatory ^ Of Cuvier; Epiotic of Huxiey, according to Vrolik. 28 Batrachia of the Permian Period of North Amenca. [January, limbs were derived ; a pelvis ; and a cranium nearer that of the batrachians than most other fishes present. The Crossopterygia are a little on one side of the parental stem, since they have no pelvis, and their limbs begin to show a beginning of that reduc- tion and specialization, which is carried to such an extent in the Actinopteri, or typical fishes. The Batrachia are supposed by Professor Huxley to have given origin to the Mammalia. There are many reasons to sustain this view, nevertheless no progress has yet been made by palaeonto- logical research in filling up the great interval which separates the Permian Batrachia from the Mesozoic Mammalia. It is also true that the limb bones of the Permian Reptilian order of the Theromorpha more nearly resemble those of the lowest Mamma- lia, the Monotremata, than do those of any other known forms. The Rachitomi. I proposed this name^ for a division of the Batrachia which predominated during the Permian period in both the old and new worlds. As stated above in the differential table, it is characterized by the primitive condition of its vertebral column. The cartilaginous chorda dorsalis was present in life, and the vertebral bodies are represented by ossifications of its sheath. In the Trimerorhachidae this ossification is superficial or cortical, while in the Eryopidae it penetrated more deeply into the chorda. The segments of the centrum are three in number, an inferior one or intercentrum, and a superior lateral one on each side, the pleurocentrum, as it is called by Gaudry. The lateral pieces support the neural arch, and on this account I have called them the centrum proper, as distinguished from the intercentrum. The neural arch is un- broken, and displays articular processes (zygapophyses) of usual form, and in some genera a large neural spine. The shoulder-girdle of these animals is remarkable for the small size of the coracoid element, resembling in this respect the salamanders, and approaching the mammals. There is probably a clavicle, as has been observed by Gaudry in Actinodon. He also finds the three thoracic shields of the Stegocephali (episterna and entosternum). I have such bones associated with the Ameri- can forms, but have not yet determined their species. The humerus has no head, but a band-like articular surface instead. 1 American Naturalist, 1882, p. 333. [884.] Batrachia of the Permian Period of North America. 29 [ts condyles, when present, are, on the other hand, well developed, md resemble those of salamanders, and of certain Mesozoic Hid lowest Tertiary mammals, as Meniscoessus of the Laramie, and Catopsalis of the Puerco. The terminal phalanges were not converted into claws, but , \ were flat and obtuse, as in existing Batrachia. The pelvic arch is inter- mediate between those of the Anura and salaman- . ders, and resembled close- hly that of the Pelycosau- i rian division of the Thero- Imorphous reptiles of the same age. The pubis and ischium are solidly united, ' without the intervention of an obturator foramen, and those of opposite sides form a boat-shaped body. This is suspended from the sacrum by a vertical I ilium, which rises from near the middle of each side. Its proportions are \.— Eryops megacephalusQo^^. Inferior • u 1 P^*^*- scapular arch of a different individual about as m the Salaman- from that represented in PI. ii; four-fifteenths ders. The femur has no Upper figure from front; lower, posterior view. head proper, and the dis- tal condyles are better distinguished than in any other Batrachia- and than in most Reptilia. The species of the Rachitomi are rather salamander-like in proportions, with relatively short legs and long tail, except in Eryops, where the latter was probably represented by a stump (see Plate 11). They had relatively large heads with wide gape, as in the frogs. None of the known species can be supposed to have had powers of leaping as in those modern animals. None of the known species had a formidable dentition, and in all of them the dentine is simply inflected, so that the section of a tooth presents a mass of closely-packed radii. The two families of this order are well distinguished by the 30 Batrachia of the Permian Period of North America. [January, form of the basioccipital bone. In the Trimerorhachidae its con- dyle is simple and concave, somewhat as in some fishes. The Ery-- opidae have the two condyles characteristic of Batrachia generally., This difference might be esteemed as of greater than family signifi-: cance, but it is less considerable than at first sight appears. The single cotylus-like basioccipital bone of the Trimerorhachidae is notched above, sometimes deeply, to receive the apex of the noto- chord. A corresponding notch on the inferior edge would, if present, divide the articulation into two surfaces, which would greatly resemble the condyles of Eryops. The latter are flat and look partly towards each other, and are evidently separated orig- inally by the fissura notochordce. To the Trimerorhachidae I have referred, with certainty, only the genus Trimerorhachis. I have been unable to learn the structure of the vertebrae in the European genus Archegosaurus. According to some authors they are simple, as in the Stego- cephali. Specimens of this kind, of the size of the Archegosaurus decheni, are found in the locality where the latter occurs, i. e.^ Saarbriicken in Alsace ; while rachitomous vertebrae from the same locality are of larger size, and resemble those of Eryops (Mus. Princeton, N. J.). I have called attention to the structure of the vertebral column of the Rachitomi from a mechanical standpoint.^ The notochord persists, the osseous elements about it in the sheath or skin, in the form of regular concave segments much like such segments as are cut from the skin of an orange, i. e., parts of spheres, having greater or less thickness according to the group or species. Now the point of divergence of these segments is on the side of the col- umn, the upper segments rising, and the lower segments expanding downwards. To the upper segments are attached the arches and their articulation, and the lower segments are like the segments oi a sphere. If you take a flexible cylinder, covered with a more or less resistant skin or sheath, and bend that cylinder sidewise, you will of course find that the folds of the surface will take place along the line of the shortest curve, which is on the side ; and, as a matter of fact, you will have breaks of very much the char- acter of the sutures of these vertebral segments. It may not be so| symmetrical as in the actual animal, for organic growth is sym-"^ metrical so far as not interfered with ; for, when we have two ^ 1 Science t 1883, p. 276. I 1884.] Batrachia of the Permian Period of North America. 31 forces, the one of growth and the other of change or interruption, and they contend, we still have in the organic being a quite symmet- rical result. In the cylinder bending in this way, of course the short- est line of curve is at the center of the side of the cylinder, and the longest curve is at the summit and base, and the shortest curve will be the point of fracture. And I presume it has hap- pened in the case of the construction of the segments of the sheath of the vertebral column that the lateral motion of the animal swimming has been the actual cause of the disposition of the osse- ous material in this form. A very good illustration of the effect of bending of a more or less flexible cylinder may be seen in the folds on the inner con- cave side of one's coat sleeve. In the accompanying figure the folds in the cloth represent the lines of flexure, or what would be Fig. 2. — Sleeve of a coat, showing folds produced by lateral flexure, which leave interspaces similar to the segments of a rachitomous vertebra. Thus i represents intercentrum,/, pleurocentrum, and n, neurapophysis. in a seat of ossific deposit, of interruption ; while the interspaces represent the segments bounded by such lines. The correspond- ence with the segments of the vertebrae of the Rhachitomi is remarkable. At the base we have the wide lenticular intercen- trum (viewed partly from below in the cut) ; between their lateral apices we have the pleurocentra, and above, another segment, which in the vertebra is the base of the neural arch. The view that the segmentation of the vertebral column is the result of lateral alternating strains, was proposed by Herbert Spencer (Principles of Biology, 11, p. 195). The present re- searches confirm this hypothesis in general. Mr. Spencer did not, however, specify the kind of segmentation to be expected from this process, and leaves it to be inferred that the segments will be cylinders of greater or less length. Thus he says (p. 205) " in a vertebral column of which the axis is beginning to ossify, the centrums consist of bony rings inclosing a still continuous rod of cartilage." And it is true that this is the primitive form in 32 Batrachia of the Permian Period of North America. [January, embryology, but whether true in any instance in palaeontological history yet remains to be ascertained. Trimerorhachis Cope. This genus presents the most imperfect vertebrae known in the order. It differs from ail others, including Archegosaurus, in the lack of a distinct neural spine. Its humeri do not display condyles, but had cartilaginous articular surfaces. The teeth are rather small Fig. 3. — Trimerorhachis insignis, skull from above ; one-half natural size. Smaller figure; muzzle showing sculpture and nares. and of equal size, except a large one or two inside the external series near the anterior part of the mouth. Two species are known, the T. insignis Cope, and 7! bilobatiis Cope, both from the Permian beds of Texas. Both were proba- bly of slender proportions, and had short weak lim.bs.^ The head of the T. insignis is wide, flat and rounded, and its superior surface is strongly wrinkled. The lyriform mucous groove does * This proportion is not certainly known. 1884.] Batrachia of the Permian Period of North America. 33 not extend behind the orbits. This was a very abundant species Fig, 4. — Trimerorachis insignis Cope; parts of skeleton, natural size. Fig. a basi- occipital, exoccipital and periotic bones, posterior view ; b, angle of mandible, ex- ternal view ; c, the same, posterior view ; d, part of vertebral column depressed by pressure, showing the intercentra (2) and the pleurocentra (/) ; . glenoid, and j^encrally lookiiii;- inon; downwards than outwards. I call this the subaqumnosal. Those foramina may bo arranged in four sets, as follows : I. Looking downwards ; Postglenoid. Subsquainoml. II. Looking outwards ; Pof^tsquamomL PoHtparietdl. III. Looking upwards ; Supra glenoid. IV. Looking backwards ; Mastoid. Some foramina of the same region are not necessarily connected with the sinus lateralis. Hyrtl, in his essay* on tlie arterial system of the Edentata, shows that a foramen near to the postsquamosal of the Tamandua" tetradactyla, gives passage to an " arteria diplo'etica," which is formed by the junction of tlie occipital branch of the carotid with an ascending branch of the temporo-maxillaris division of the carotid. The a. diplo'etica issues in a foramen which perforates the parietal bone on the orbital border. These two foramina may be called the/, diplo'eticum posterior and /. d. anterior, respectively. The former enters from the fundus of the same small fossa, which is also perforated in its superior portion by the /. postsquamosaie, the canal from the latter foramen taking the usual vertical direction. Still another foramen exists, which is, so far as my present knowledge goes, confined to the Monotremata and Marsupialia. It enters the posteri- or base of the zygomatic portion of the squamosal, and is directed forwards. In Tachyglossus it passes through the base of the zygoma, issuing in the base of the zygomatic fossa. In the Marsupialia it enters a fossa of the squamosal bone, which may or may not be partially filled with cancellous tissue. I call this the foramen postzygomaticum. I now give the results of my observations on the crania of the most im- ])ortant genera which I have observed, one hundred and sixteen in number. f Monotremata. Tachyglossus. The only foramina are the /./. diplo'etica anterior and pos- terior, and the postzygomaticum. The anterior half of the canal connect- ing the former two has no external wall. OrnithorhyncTius. Postzygomatlc large and passing tlirough the zygo- ma ; postsquamosal large ; no other foramina. Marsupialia. The types of this order generally have the postglenoid, and hardly over have the supraglonoid or postparietal. They are generally distinguished * Denkschriften Wiener Akademie, 1854, T. Ill, pi. 1. fMost of these are preserved in my private collection ; for alew I am indebted to the Museum of the Philadelphia Academy. Cope.] 454 [Feb. «. by the presence of the subsquaniosal, but in nypujwymnun and Macropui tliis foramen becomes the postsquamosal, through the faiUire of the post- zygomatic crest. It need not be confounded with another foramen also found in these genera, which enters above the mentu» auditorium externum, and communicates with the tympanic chamber, and which T call the su- pratpmpanic foramen. The subsquaniosal enters the sinous canal, and in Phase olarc.tos, where the postglenoid is wanting, constitutes its only exter- nal outlet. The order is further characterized by the presence of the post- zygomatic foramen. Phascolomys ; postzygomatic chamber enormous, extending above mea- tus. No foramina below, except supratympanic ; above, a supraglenoid and one or two postglenoids. Phascolarctos ; subsquaniosal and postzygomatic only ; the latter com- municating with an empty chamber, Macropus Hypsiprymnus ; postglenoid, postzygomatic, supratym- panic and postsquamosal ; the second communicating with a chamber filled with cancelli. Fjg. 1.— Skull of opossum ( DUielphys virgin iaiia) natural size, posterior view, parts of the right mastoid and squamosal bones removed. M, mastoid fora- men ; SBS, subsquamosai ; PG, postglenoid; PZ, postzygomatic foramen. Didelphys ; postglenoid, postzygomatic, subsquamosai and mastoid ; postzygomatic small. Dasyurus ; postglenoid, subsquaniosal and probably mastoid. I cannot find a postzygomatic. Phalangista ; postsquamosal and postparietal only; no postzygomatic nor supratympanic. Edentata. Tamandua ; F. f. diploetica anterior and posterior, and postsquamosal only. Basypus {(S-cinctus) ; postiilenoid (large), postsquamosal, mastoid, several postparietals, and a small subsquamosai. 1S80. ] 455 [Cope. Chlamydophoriis ; a iK)stgleii()ul only. Manis ; postzygonuitic only. MegaUmyx ; jiost squamosal and supratympanlc ; a closed fissure at in^si- tion of postglenoid. A large foramen below the usual position of mastoid. Bmdypus and Cholc^pns ; no foraniina. UODENTIA. In this oMer, so far as yet observed, the supi^glenoid and postparietal foramina are never present, while the ntastoid is rarely, and the sub- squaaiosal is generally, represented. The ridge connecting the zygoma with the inion being weak, the ditterence l)etween this foramen and the postsquamosal is less marked in this order than in the Marsupialia. It is, however, always on the inferior border of the squamosal bone. Lepus and Lagomys ; no foramina- Lagidmm ; no foramina. Cer^Jabes ; no foramina. Lagosiomus, Geomy s und Erithizon ; an enormous postglenoid without internal canal. Capromys^ (J^ceUfjenys, IScinrus, IlapledoiUuL, Hesperomys^ Mus ; jKJst- glenoid and postsquamosal only. Ilystrix^ Hydroclmrus, Neotoma ar»d Arvioola ; ^xistgleuoid and post- squamosal foramina confluent ; no others. Castor, Cynomys an-d Sperm&pMlus ; [>ostglenoid, postsqugiiiosal and mastoid. In«e€tivoiia. The foramina are very muck as iia the BedeMm im tke f-nialler forms, and as in the Garnwera in the larger. Blarina, Condglurd and Scnleps : postsquamosaS only. Erinaceus ; postglenoid and postsquamosal only. Mystomys the same, according to Allman's figures. Centetes ; postglenoid, postparietal and mastoid. Solenodon (from Peters' figures) ; postglenoid and postparietal. Chikopteka. In some members of tins order the foramina are, as In ntaity Carnivora, limited to tlM3 postglenoid and postparietal. ScoiopMlus {fuscus); postglenoid, postpaiieta! and mastoid. Pteropus ; postglenoid, subsquamosal and postsquamosal- Carnivora. In thi« ord-er tlxi fbraiaiilaa are few in numl>er, and arc very well defined. None of thean possess more than tliree, while the* spcvjaliaed forms, both terrestrial and acjuatic, do not possess them. Tric]':6cu8 and Arctecephalus ; no foramincijon postglenoid and postparietal only. Fig. 2. — Tenmocfjoii conrp/Kt'iis Cope, Lower Miocene of Oregon ; one-lialf natu- ral size. PP, postparietal foramen. Archcdurus^ Dinictis, Pogonodon, Iloplophoneun and MacJuBrodus (cere- hralis) ; postglenoid and postparietal only. Procyoii, Nasua and Bassaris ; postglenoid only. Canis, Vulpes and Urocyon ; postglenoid only. Yi'oevra, Mustela, Putorim and Mephitis ; postglenoid only, Felis {domentica) ; sometimes a minute postglenoid only ; sometimes none. Hycena, Uncia, Cynmlurus ; no foramina. PiiosiMi^:. Lemur, C/m-oynleus mu] Tarsius ; a postglenoid only. QUADRUMANA. Hapale ; postglenoid and postsquamosal. Geb'us ; jwstglenoid and postparietal. The latter is on the suture of the parietal and squamosal bones ; in Hapale pemcillata it is entirely within the squamosal bones. Ateles, OaUithrtx, Mycetes, Semnopithecm and Gynocephah^s ; no for- amina. Maeacus ; a small postglenoid. Troglodytes niger, gorilla ; a closed fissure, but no foramen postglenoid- euni. Homo ; no foramina in sixteen North American Indians examined of the Klamath, Bannock, Crow, Sioux and Cheyenne tribes. One postglenoid on one side in a South Australian. One or two mastoids are more usual, be- ing found in a good many specimens of all races. They are rarest in Hot- tentot negroes. Cktacea. Balcena, Beluga -dnd Mono do n ; no foramina. SlIiETs'IA. Halicore and Manatuif : a huge mastoid only. im). I * LCope. ITyracoidea. Ilyrax ; no foramina. PUOBOSOIDIA. Eleplias ; no foramina. PEIirSSODACTYIiA. In this order the number of the foramina ranges from very few to many. YiG. "(i—Aphelopsmcyalodm Cope, Loup Fork of Colorado ; one-sixth natural size; showing postparietal foramen. Ridnoeerm, ApJielops ; postparietal only. Fig. 4. Fig. 5. Skulls of Aphelops mfgalodm (Fig, 4) and A. fossiger ( Fig. .'3) from behind ; one- sixth natural size; showing absence of mastoid foramen. Tapirus; postparietal and a huge mastoid. Ancldtherium, Hippotlierium, ProtoMppus and Equus ; postparietal, postsquamosal, postglenoid and sui)raglenoid. In the last three genera ( 'ope.] 458 I Feb. H, the vessel issuing from the postsquamosal, grooves the petrous bone, leav- ing it at a point near that usually occupied by the mastoid foramen. In the second and last genera, and probably in the third, the sinous canal is protected by a bony crest in front, throughout its entire length. AllTIODACTYLA. Great diversities are found in this order, especially between the suilline and ruminant divisions. In the former, with the exception of the Hippo- potamidcB, there are no foramina ; in the Ruminantia the}' are more nu- merous than in any other order of the class. The Ruminantia are, like the equine Perissodactyla, characterized by the presence of the supraglenoid foramen ; to this the Gamclidoi and some others add the mastoid. The TraguUna must be excepted from this rule, for they have nothing but the post glenoid. Omnivora. Suft, Dicotyles and PhacocJimnis ; no foramina. Hippopotamus and Chceropsis ; postglenoid. postsquamosal, mastoid and a rudimental supraglenoid. Buminantia. Tragulus ; postglenoid only. Oreodon ; postparietal and mastoid. In one specimen of 0. culhertsoni from Colorado, I find a minute supraglenoid on each side ; in other speci- mens it'is wanting. Po'ehr other ium ; postparietal, postglenoid ; mastoid; a small supraglen- oid. SPG Fig. 0.— Skull of . Proeamehis occid'-ntalis T.eidy, Loup Fork of New Mexico; one-fourth natural size; sliowiug supmglenoid foramen, SPG. Procamelus, Camelus, Auclienia ; postglenoid, supraglenoid and mastoid. Bo% ; postglenoid and supraglenoid only. ArUilocapra ; postparietal, postglenoid, mastoid, and a large supraglen- oid. Giraffa ; poslglenoid, supraglenoid, postsquamosal and mastoid. Orens, Ool.% Cervus ; postglenoid, supraglenoid, postsquamosal, postpa- rietal and mastoid. ISSO.) 459 [Cope. From the preceding the following conclusions may be derived : (1) The sinous foramina furnish valuable diagnostic characters, and nuiy, with proper limitation, be used in systematic definition. (3) The primitive condition of the various mammalian orders appears to have been the possession of a limited number of these foramina. (3 ) The n)onotreme-marsupial line have developed a number of loramina in their own special way. (4) The Rodentia have chiefly developed those of the inferior part of the squamosal bone, if any. (5) The Carn'mora commenced with but few foramina, and have obliter- ated these on attaining their highest development. (6) The history of the Quadrumana is identical with that of the Gar- nivora. (7) The Perissodactyla present very few foramina in the lowest forms, and did not increase them in the line of the Rhinoceridce. In the line of the horses an increase in their number appeared early in geologic time, and is fully maintained in the existing species. (8) In the Omnivorous division of the Artiodactyla, time has obliterated all the sinous foramina. In the Camels an increased number was apparent at the same geologic period as in the history of the horses (White River or Lowest Miocene), and has been maintained ever since ; while the existing Pecora present a larger number of the foramina than any of the class of Mdmmalia. The only relation between these structures and the habits of the species concerned that can now be traced is, that the largest number of the foram- ina is found in the specialized vegetable feeders, while the smallest num- ber is found in omnivorous forms. I now give a synopsis of the distribution of the sinous foramina accord- ing to the foramina themselves. The /. /. diploetica, postzygomatica and mpratympanicum are not included, as their existence is restricted to the few types already mentioned. I. No foramina. Homo, Troglodytes, Cynocephalus, SemnopitJiecus, Mycetes, Cal- lithrix, Ateles. Uncia, Hymna, Arctocephalus, Trwliecus. Elephas, Hyrax ; Sus, Phacoch(Erus, Licotyles. Lepus, Lagidium, Cercolabes. CJiolcBpus, Bradypus. IT. Postglenoid only. a. Rudimental. Felis ; Phoca. aa- Developed. Chlamydophorm. Lemur, Ghirogaleus, Tarsius ; Macacus. Cope.] 460 [Feb. 0, Mustela, Putorius, Mephitis ; Canis. Vulpes, Urocyon; Viverra. Pr'ocyon, Nasua, Bassaris. Tragulus. aaa. Enormous. Lagostomus and Geomys. III. Subsquamosal only. Phascolarctos. IV. Postsquamosal only. Ornithorhyncus, Tamandua, Blarina, Oondylura, Scalops. V. Postparietal only. Bhinocerus, Aphelops. yi. Mastoid only. Halicore, Manatus. YII. Postglenoid and subsquamosal only. Hystrix, Hydrochmrus, Capromys, Ccelogenys, Sciurus, Haplo- dontia, Neotoma, Hesperomgs, Mus, Arvicola. VIII. Postglenoid and postsquamosal only. Brinaceus. Macropus, Hypsiprymnus. Hapale. IX. Postglenoid and postparietal. Ghiroptera sp. Temnocyoriy Enhydrocyon ; ArchiBlurus, Dinictis, Pogonodon, Hoplophoneus, Machcerodus. Cehus. X. Mastoid and postparietal. a. Mastoid small. Oreodon. aa- Mastoid enormous. Tapirus. XI. Mastoid, postglenoid and postsquamosal. Castor, Cynomys, Spermophilus. XII. Mastoid, postglenoid and subsquamosal. Dasyurus, Didelphys. XIII. Mastoid, postglenoid and postparietal. Scotophilus (fuscus). Centetes. Ilycenodon, Ursus, Arctotherium ; XIV. Supraglenoid and postsquamosal only. Phascolomys. XV. Supraglenoid and postglenoid only. Bos. XVI. Supraglenoid, postglenoid and mastoid. Procamelus, Camelus, Auchenia. XVII. Supraglengid, postglenoid, mastoid, postparietal. a. Supraglenoid small. Po'ebrotherium. 1880.] 461 [Cope. ««• Supraglcnoid large. Antilocapra. XVIII. Supraglcnoid, postglenoid, mastoid and postsquamosal. a. Supraglcnoid small. Hippopotamus, Clicbropsis. aa. Supraglcnoid large. Oiraffa. XIX. Supraglcnoid, postglenoid, postparietal and postsquamosal. a. Supraglcnoid small ;• mastoid not grooved. A7ichitherivm. ^. Supraglcnoid large, mastoid grooved. Hippotfierium, ProtoMppus, Equus. XX. Supraglcnoid, postglenoid, postparietal, postsquamosal and mastoid. Cervus, Oreas, Ovis. Printed March 6, 1880. DEPARTMENT OF THE INTERIOR. UNI 1 ED STATES GEOLOGICAL AND GEOGRAPHICAL SURVEY. F. V. HAYDEN, U. S. Geologist-in Charge. I.— Tlie Relations of the HORIZONS OF EXTINCT VERTEBRATA OF EUROPE AND NORTH AMERICA. II.— Observations on the FAUN^ OF THE MIOCENE TERTIARIES OF OREGON. BY E. D. COFE. EXTRACTED FROM THE BULLETIN OF THE SURYEY, Vol. V, No. 1. Washington, February 28, 1879. Art, II.— The Relations ot the Horizons of £xtmct Yertehrata of Europe and ]¥orth America. By £. D. Cope. The history of the succession of life uj)on any one portion of the earth^s surface is replete with matter for speculation. It shows us a series of faunse succeeding each other, each of which, in many instances, com- mences without previous announcement in the forms of older periods, and disappears without leaving representatives in later ones. With this basis of fact, which naturally enough has been furnished by the longest explored and best known portion of the earth, Europe, we turn to other lands with the hope of obtaining further light upon a subject so full of mystery. These types of life, did they originate in a single centre, from which they disseminated themselves and, if so, did each form originate in a region of its own or not ? Or, did the same types of generic structure appear at different points on the earth's surface independently j and, if so, whether cotemporaneously, or at different times I For a solution of these and similar questions, we naturally look to a comparison of the facts first established, with those obtained more recently by exploration in other regions. In this quest, no portion of the earth offers greater promise of results than America. As the second great continent, separated from the other by the greatest possible water sur- face, we anticipate the widest diversity in the character of its life-history. If the types of life have originated independently, we will find evidence of it by studying American palaeontology; if their origin has been through gradual modification, America should furnish us with many interme- diate faunae. The identification of the generic types of North American Yertehrata has now advanced to a point which renders such a comparison possible. Although the subject is in its infancy, the following pages will show that an important contribution to it can be now made. The compari- sons instituted in this i)aper commence with tlie coal-measures, and with the BatracMa of that period. As regards the palaeozoic fishes, I have not yet devoted that attention to them which is necessary for their discussion, and I refer to the papers of Newberry for several important identifications of genera as common to the two continents. The structure of the BatracMa of the coal-measures is not yet suffi- ciently well known to enable the most exact comparisons to be made, but close parallels, if not identities, of genera exist. Such are the BuU. V, 1 3 33 34 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [Yol.Y. OestocepJialus and Ceraterpeton of Ohio as compared with the Urocordylus and Cera terpeton of Great Britain. The Permian vertebrate fauna which I discovered in Illinois and Texas, exhibits close parallels, but not yet generic identity, in the two continents. Thus, the American Clepsydrops and Dimetrodon are near to the Deuterosaurus of Perm in Eussia, and the Lycosaurus of the mount- ains of South Africa. The Texan genus Parioticlius may, with further information, prove to be identical with Procolophon Ow. from the Tafel- berg. Humeri of the type discovered by Kutorga in Eussia, and by Owen in South Africa, are found in i^orth America, and the same re- markable type has been recently discovered by Gaudry in France. The peculiar type of Labyrinthodont vertebrae described by me under the genus Ehachitomus from Texas has been discovered by Gaudry in France. The present indications are that close similarity between the faunse of this period in Europe and America will be discovered. Never- theless, up to the present time no representatives of the striking Ameri- can forms Diadectes, Bolosaurus, JSmpedocleSj and Cricotus have yet been found in any other continent. As regards the Triassic fauna, it differs from that of the Permian in being better known in Europe than America. As marine Trias is little developed in ]N^orth America, so the vertebrate fauna of the Muschel- kalk has not been discovered in the latter country. It is otherwise with the Keuper. The characteristic genus of that epoch, Belodon, existed in America, and parallels, if not identity, exist in the genera Thecodonto- saurus and Falceosaurus. These are known in America from teeth only. The reptiles are accompanied in North America, as in Europe, by Stego- cephalous Batrachia, mostly Labyrinthodonts, but their generic af&ni- ties are yet unknown. The great Jurassic faunse are as yet but sparsely represented in North American palaeontology. The marine Vertehrata of the Lias are either unknown or are represented by a few provisional identifications of un- satisfactory fragments. We do not yet know any deposits in North America which contain the typical reptilian genera Plesiosaurus, leh- thyosauruSj Pliosaurus, and Bimorphodon^ or the fishes of the Bapediidw. This formation, so important in Europe, is almost omitted from the North American series. A few more characteristic fossils of the Eocky Mount- ain region represent the Oolite, particularly the Upper Oolite, while Teleosaurus and Steneosaunts^ and their allies, are not yet known from North American beds. Strata said to be included in the Dakota (which on the evidence of plants and invertebrate fossils has been placed at the bottom of the Cretaceous series) have produced a genus not ye>t distinguishable from Megalosaurus. This genus has not been identified beyond doubt from above the Oolite in England. From the same beds in the Eocky Mountain region come genera which nearly resemble the one from the English Oolite (Forest Marble) called by PhiUips, Cetiosau-i rusj and the genus from the Oxfordian of Honfleur, called by von Meyer,! No. 1.] COrE ON HORIZONS OF EXTINCT VERTEBRATA. 35 Streptospondyhis. Beyond this no comparisons can be made, and we therefore pass to the rich fauna of tlie Kimmeridge. North America cannot show such records of this epoch as have been found in Europe. There are no Arcliccopteryx^ Ehamphorhynchus, nor Fterodactylus ; no Lep- tolepis, ThrissopSj nor other of the numerous fishes of Solenhofen. The Omosaurns has, however, some very close relatives in the supposed Da- kota beds of the Eocky Mountains. No remains of that primitive Mar- supial fauna which occurs in the Purbeck have yet been detected in the Western Continent. A partial representation of the Wealden fauna of Europe is found in the beds of the Rocky Mountains mingled with the types of the Oolite and Kimmeridge already mentioned. The rela- tionships of this fauna to those of the European Jurassic series may be thus exhibited : American. European. Oamarasaurus Beds. Wealden. f Iguaiiodon, f Kypsilopliodon. Rypsilophodon. Hylceosaurus. f Cetiosaurus. Cetiosaurus. Camarasmirus, Eucamerotus.* Ornithopsis. AmpMcoelias. f GoniophoUs, Goniopholis. Ejinoieridge. Sypsirhoplms, Omosaurus. Caulodon. f Caulodon.f OXPORD. Upanterias, Streptospondylus. Oolite " Cetiosaurus,'''' f Megalosaurus. Megalosaurus. From the above table it will be seen how difficult it is at the present to parallelize the related beds of the Jurassic periods of the two conti- nents at the present time. All that can be said is that many types re- sembling | nearly those of different horizons of the European Jurassic are found to have lived together or near together in the Eocky Mount- ain region of North America. That the Cretaceous fauna of North America was the richest in the cold-blooded Vertebrata is indicated by the present state of discovery. The ocean of the interior of the continent deepened from the beginning * Chondrosteosaurus Owen. + Iguanodon prcecursor Sauv. t A near affinity has been shown by Professor Owen to exist between Eucamerotus and Camarasaiirus. Professor Owen believes these genera to be identical; but the neural spiness of the anterior dorsal vertebrsB are very different, being single in the former, and double in the latter. 36 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [Yol.Y. of the period until tlie epoch of the Mobrara, and then gradually shal- lowed until the elevations of the bottom began to divide the waters. The closing scenes of this great period were enacted amid a labyrinth of lagoons and lakes of brackish and fresh water, whose deposits form the beds of the Laramie epoch. The fauna of the deep-sea epoch, the Mobrara, is the best known. Here the remains of Pythonomorpha constitute its prevailing character- istic, while Elasmosaurus and PolycotyluSj with but few species, represent the numerous Sauropteryyia of Europe. Crocodiles were apparently wanting, while turtles and a peculiar group of Pterosauria were only moderately abundant. The fish fauna was very rich and varied. Here the Saurodontidcej like the Molluscous family of the Eudistes, appeared, and as soon disappeared, accompanied by the peculiar form, Eruiclithe^ and the family of Stratodontidce, The genera of Mount Lebanon, Lepto- tracJielus and Spaniodon, occur in this bed in Dakota 5 but the closest paral- lelism is exhibited with the Lower Chalk or Turonian of Western Europe. The general fades of the reptilian fauna is that of the Lower Chalk, and there is little doubt that several genera are identical in the two conti- nents, e. g. Elasmosaurus. The apparent peculiarity of the Chalk in America is the abundance of forms (four genera) of Pythonomorpha with numerous species, while but two genera have yet been found in Europe, and the i^resence of birds, with biconcave vertebrae, and teeth. This in- teresting type, which was first discovered by Seeley in the genus named by him EnaliorniSj and afterwards found by Marsh to possess teeth, has been found at a lower horizon in England, the Upper Greensand. But in England, France, and Westphalia occur the genera of fishes above mentioned, as Portheus, IchthyodecteSj Saurodon, Saurocephalus, Eri- sichthCj EmpOj Pachyrhizodus^ Enchodus^ Leptotraclielus^ etc. This close relationship of the horizons permits an identification, and it is the first instance which appears to me to be susceptible of satisfactory demon- stration. The next horizon of the Cretaceous which has yielded many verte- brate remains in I^orth America is the Fox Hills formation (including the Fort Pierre bed). Here the genus Mosasaurus appears in America, and is accompanied by the earliest crocodiles with procoelous vertebrae, and by numerous marine turtles which partake of the characters of both Chelydridce and CheloniidWj which I have called the Propleuridw. Beryx appears first here in America. The predominant genus of fishes is En- chodus, and the principal Binosauria are Bcdaps and lladrosaurus. This horizon has been parallelized with the Maestricht of Europe, and sev- eral genera are common to the two beds ; such are Mosasaurus and Enchodus. The genus Hadrosaurus^ and the family of tiu-tles I have called the Adocidw, remain undiscovered in Europe ; hence the identity of faunae cannot be established. The lacustrine beds, or summit of the American Cretaceous series, the Laramie of Hayden, present the remains of a populous fauna and a rich No. l.J COPE ON HORIZONS OF EXTINCT VERTEBRATA. 37 flora. The students of the paheobotany have declared this flora to be of Eocene, and the hiter portions of Miocene character, while the lacus- trine constitution of the strata has influenced the stratigraphic geolo- f»'ists to concur in the view that the formation should be arranged with the Tertiary epochs. That the fauna was of a mixed chanicter is the result of a study of its vertebrate fossils. The i)redominant type in i^^orth America was the JJinosauria, which were abundant in species and individuals, and this fact alone will suffice most palaeontologists as a reason for referring the epoch to the Cretaceous series. The genera of Dinosauria (PalwoscincuSj Cionodon, Diclonins, Monoclonius, Dysganus^ etc.) have not yet been found in any other part of the world. Mingled with them were species of crocodiles and turtles of indifferent character, while a number of other forms existed which had a limited range in time, and hence are important indicators of stratigraphic position. Such are the genera of fishes, Myledaplms Cope and CJustes Cope, which have been found also near Eeims, France, by Dr. Lemoine, in the Sables de Bra- cheux, which are regarded as the lowest Tertiary. Such is the curious Saurian type Champsosaurus (Cope) {Simwdosaurus Gerv.), and the tur- tle genus Compsemys Leidy, which Lemoine finds a little higher up in the series, in the Conglomerate of Cerny, which is in the lower part of the Suessonian. In France, a genus of the Laramie, Po??/f/iora^, extends into the Lignite or upper Coryphodon bed of the Suessonian. Thus the Laramie is intercalated by its characters between the Cretaceous period on the one hand and the Tertiary on the other, and its fauna includes genera and orders of both great series. These relations may be ex- hibited in tabular form as follows. I here include the faunae of the Sables de Bracheux and of the Conglomerate of Cerny as one, since both possess the types of the Laramie, while the horizon of the Lignite of Meudon, or the Suessonian, does not. Sables de Bracheux and Con- Laramie. GLOMERATE DE CeRNY. a. Tertiary. LopMochoerus. Flesiodapis. Fleuraspidotherium. Arctocyon, Clastes, Clastes, Peculiar. Champsosaurus, Compsemys. Myledaphus, Compsemys. Myledaphus. Scapherpeton. Champsosaurus, 38 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [VolY. Sables de Bracheux and Con- Laramie. GLOMERATE DE OeRNY. Cretaceous. Palceoscincus, Dysganus. Monoclonius, Diclonius. Cionodon. Lcelaps. Auhlysodon. If the Conglomerate of Cerny is the same horizon as the Conglomerate of Meudon^ we must add Coryphodon to the upper left-hand column, and probably Gastornis also. The result is clear that the French and Ameri- can formations together bridge most completely the interval between the Cretaceous and Tertiary series, as has been anticipated by Haydeu, in America, on geological grounds. It is also evident that another forma- tion must be added to the series already recognized in France, viz, the Laramie or Post-Cretaceous. This will be defined as the beds of the genera Climnpsosauriis and Myledaplms. In France, the presence of mam- malia will characterize the formation as a subdivision, for which it is probable that tlie name Thanetian must be retained ; while to the Ameri- can division, which is characterized by the presence of Dinosauria, the name of Laramie beds has been given. The Eocene fauna is so varied, especially in Europe, that it is neces- sary to compare the di\'isions separately, as in the case of the Cretaceous. Thus, the fauna of the Suessonian is quite as distinct from that of the Calcaire Grossier and Gypse (Parisian and Tongrian) in France as are those of the Wasatch and Bridger epochs in North America. I have already identified the Wasatch with the Suessonian or Orth- rocene, on account of the community of the following genera in the two continents: Goryphodonj Hyracotherium^ Amhlyctonus, Clastes^ and a form of birds close to Gastornis. I can now add Fheiiacodus^ Orothe- rium (Cope), and very probably Hyopsodus^ Adapts^ OpisthotomuSj and Frototomus. But, as above mentioned, in the lower beds of the Sues- sonian in France occur genera which are, so far as yet known, wanting in the Wasatch of America, but present in the beds of the Laramie. Such are two genera in the Conglomerate of Cerny, and four genera in the lower Sables de Bracheux. In the former bed, they are associated with the mammalian genera Lophiochceriis^ Plesiodapis^ Plem-aspidothe- rium, and Arctocyon ; and in the lowest, with a form referred with doubt to Hyracotherium. Thus the generalization may be made that the char- acteristic genera of reptiles and fishes of the Laramie of Xorth America are in America associated with Cretaceous Pinosaur ia, and not with Blam- malia; while in Europe they are associated vdth Mammaliaj and not with Dinosauria. In arranging the Laramie Group, its necessary position is NoA.] COPE ON HORIZONS OF EXTINCT VERTEBRATA. 39 between Tertiary .iiid Cretaceous, but on the Cretaceous side of the boundary, if we retain those grand divisions, which it appears to me to be desirable to do. The reasons for retaining it in the Cretaceous are two, viz : (1) because IMmmuria are a Mesozoic type, not known elsewhere from tlie Tertiary j (2) because Mammalia (should they be found in the future in the Fort Union) are not equal as evidence of Tertiary age, since they have been also found in Jurassic and Triassic beds. The parallel- ism of the American Wasatch with the Upper Suessonian of France is the second identification which may be regarded as provisionally estab- lished. The only discordant elements at present known are the Twnio- donta of the Wasatch, which have not been so far found in Europe, and the genus LopModon^ which is unknown in America. Above the Suessonian, a divergence in the characters of the European and North American faunas commences, and continues to be marked throughout the remainder of Tertiary time. So far as the Mammalia are concerned, the diversity between the continents was greater during the periods of the Upper Eocene and Miocene than at the present era. During these periods, a limited number of genera, common to the two continents, was associated with numerous genera in the one which did not exist in the other. As a consequence, our palseontological means of identification of the horizons are limited to a restricted list, and the task of applying a uniform nomenclature is, under the circumstances, diffi- cult. Another difficulty in the way of determining the place of the American beds in the European scale consists in the fact that the phy- sical history of the two continents during the Tertiary period appears to have been different. In America, the changes of level appear to have been more uniform in character over large areas. Each deposit has a wider geographical extent, and the fauna presents less irregular variation. In Europe we have a great number of comparatively restricted deposits, each of which differs from the others in possessing more or less peculiarity of fauna. After a study of these faunie, their natural ar- rangement in Europe into three series. Eocene, Miocene, and Pliocene, does not appear to rest on any solid basis. This is especially true of the distinction between the first two j and authors are at variance as to the point of demarkation between the last two. Thus, the Tongrian is the summit of the Eocene according to Eenevier, while Gaudry, with Fil- hol and others, places it at the base of the Miocene. One opinion is as well supported by facts, as now interpreted, as the other. Let us now consider the nature of the evidence on which we should rely in classifying faunae and the deposits which contain them. We are accustomed, at i)resent, to rely for our definitions upon all the faunal pe- culiarities upon which we can seize : the period of appearance of certain types; the duration of certain types; and the disappearance of certain types, depending on orders, families, and genera for the major divisions, and species at a given locaUty for the lesser. It is, of course, evident 40 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [VolY. that eitlier of the above-mentioned threh criteria are variable quantities, since discovery is constantly extending oui* knowledge of the distribu- tion of types. Hence the definitions are empirical and temporary. We must then, if we desire a stable system, examine the principles involved, and endeavor to discover definitions which stand on stronger founda- tions than those which we now possess. As a matter of fact, the old definitions of epochs and periods are con- tinually invahdated by new discoveries. As a matter of theory, this should be the case. To the believers in the doctrine of derivation, the obliteration of fau- nal distinctions is not a cause of surprise. Such await with confidence the day when complete phylogenies will be possible, and at present regard the interruptions in the succession of life as local only. Will the result then be, that palaeontology will cease to be available in the defini- tion of ages and of deposits? I answer no, on various grounds. In- terruptions in the succession of life in any given locality due to various causes have doubtless often occurred, and have left traces in the crust of the earth which are ineffaceable by discovery. But apart from this, one fact in this history is patent both to the friends and to the opponents of the doctrine of derivation. It is known that the world has witnessed, at every stage of its history, the extinction of some important type of life. Familiar examples are the Placodermi of palaeozoic time, the vari- ous reptilian groups of Mesozoic time, and the Amhlypoda of the Tertiary. Each minor subdivision of time offers its own record of persistences and extinctions of particular families and genera. Now, all departments of biology compel us to recognize the law of classification, that the order of forms is from the less to the more gen- eralized, from the simple to the more complex, and vice versa^ whether the lines of succession be those of descent or of creative order j and this law is true in time as well as in classification. It follows from this, that all types of life are, at the time of theii' appearance, less distinct and more general in their characters than they are later in their his- tory. It also follows, as a consequence of the principle of descent, which states that the types of one age have taken their origin from general- ized types of preceding ages, that there is no descent from the most specialized types 5 which is to say, conversely, that the genera, families, and orders whose extinction has been a marked feature of every geo- logic age have been the specialized types of those ages. We now have a clue to a basis of a definition for faunae, and hence for epochs, which discovery can safely build upon. The successive incre- ments of structure by which an important modification of animal type is introduced preclude the possibility of exact determination of the time at which such type may be said to have appeared. Even where such a point may be arbitrarily fixed, the type must then be less char- KoA.] COPE ON HORIZONS OF EXTINCT VERTEBRATA. 41 acteristicjilly represented than it is at the other limit of its existence, viz, the period of its disapi)earance. For these reasons I must regard the latter criterion as the true one in the discrimination of the subdivisions of geologic time, while the point of the appearance of types must be looked upon as of x)rovisional use only, and this quite independently of the changes which discovery will from time to time compel us to make in our knowledge of the distribu- tion of life in time and space. It must, however, be borne in mind that disapi)earance may be due to two causes: first, to extinction; and, secondly, to modification; a distinction which is entirely essential. The case of disappearance by modification is identical witli that of appearance by modification, and cannot be used otherwise in classi- fication. It is then the period of extinction of types to which I have reference. With these principles in view, we continue the comparison of the ex- tinct faunae of Europe and North America. If we take a general view of the Tertiary faunae, we find that the following well marked types representing families and higher groups have become extinct, and have left no living descendants or successors : Among Insectwora^ the Leptic- tidce in North America ; also the American Bunotherian groups Tccnio- donta and Tillodonta; alsotheJf(gso(^o?i^a of both continents; of Edentata^ Macrotlierium and Ancylothermm in Europe, and the Megatheriidce in North America ; among the Carnwora^ the Hycenodons and Proviverrce, with the Machwrodi ; of Ungtdata^ the entire order of Amhlypoda, which, however, doubtless disappeared in some of its members by mod- ification; but its only known suborders, the Pantodonta and the Binoce- rata, became absolutely extinct. Among Perissodactyla, both continents lost by extinction the CJialicotheriidw, which terminated in a great devel- opment in North America ; the genera Hippotlierium^ and Stylonus of the line of the horses, and the Bhinoceridce. Of Artiodactyla, two great divisions, representative of each other in the two continents, totally dis- appeared, viz, the Oreodontidw and the Anoplotlieriidce ^ to which must be added the Ryopotamidce. Of true ruminants, the most important type which has disajjpeared from both continents is that of the Camelidce, Of Suilline genera, Anthracotherium and JElotJieriimi may be looked upon as ha\ing left no persistent successors. Last of all, the Proboscidea retreated to the continents of the south. In view of the complexity of the European record, I first present the relations of the above mentioned phenomena as displayed in the sim- pler American system. As the present essay commences with the earliest periods, I exhibit the succession in descending order on the i^age. The horizons of the Tertiary which present distinct terrestrial faunae in North America have been named the Wasatch, the Bridger, the Uinta, the White Eiver, the Loup Fork, the Equus beds, and the Ohamplain. * Equus came through Frotohippus, the cotemporary of Hippotherium. 42 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [YolY. The types which became extinct* with the close of each of these epochs are the following : Wasatch. Gastornithidw Pantodonta, Bridger. Ba'enidce. Tillodonta. Mesodonta, Proviverra, Dinocerata, White Eiver. Leptictidce, Hycenodon, Chalieotheriidce, Hyopotamidw. Loup Fork. Ehinoceridce. Hippotherium, Stylonus, Oreodontidce, Equus Beds. Megatheriidce, Maclicerodus. Tapiridce. UlepJias. CamelidcB. The above table exhibits the present state of our knowledge : it wil doubtless be much extended by future discovery, but not otherwise mod ified. The numerous able writers on European vertebrate palteontology have more frequently recorded the appearance of types in defining thei faunal divisions than their disappearance. The following table is com piled from the writings of Gervais, Gaudry, Pomel, Filhol, Eenevier and others, but is not as complete as I would desire. SUESSONIAN. Falunian. Pantodonta, Ancliitlierium. Parisian (Bruxelhan, Bartonian, Anthracotherium, and Sestian). Palwochoerus, j PalceopMs (Bruxellian). Ccenotherium, ^ Proviverra. Oeningian. Pterodon. Ancylotherium. Ilesodonta. Dinotherium. Zophiodon (Bruxellian). Hippotherium, TONGRIAN. Aceratherium, Palceotheriidce. SUBAPENNINE. Chalicotherium, Mastodon, j Anoplotlieridce. Tapiridce. Elotlieriuni, Diluvial. Aquitanian. Hycena, Hywnodon. Machcerodus, Hyopotamus, Eleplias. Blimocerus. Hippopotamus. *Thi8 means, as already mentioned, the forms which left no direct successors in the Nearctic and Palaearctic Faunae. Ko. 1.] COPE ON HORIZONS OF EXTINCT VERTEBRATA. 43 Tlie above tables show that the history of maimualian life in the two continents presents many points of resemblance j but that there is a great ditliculty in correlating the epochs represented by the known fauniie. As regards the two i)rimary divisions, Eocene and Miocene, they have no special raison d/etre, as such faunaj as the Toiigrian and Oenin- gian are absolutely transitional in their character. More detailed com- parisons of the European and American fauna3 bring out many relation- ships not displayed by the above tables, and which I will now briefly consider. In the American Bridger, various genera of Mesodonta represent the few Adafpidcc of the Parisian, the genus Adapts* Ouv. being i)robably common to the two continents. The American AnaptomorphuSy a true Lemur, has been found by M. Filhol in the Phosphorites, and named Necrolemur. The characters of the numerous Camivora of the Bridger are as yet unknown. The Stypolophus of the Bridger is perhaps the Prototomus of the Wasatch, and this again has been discovered by M. Filhol t in France ; while a very similar, if not the same, genus has been discoA^ered in the Swiss Siderolitic, and named Proviverra. Hycmodon- tidw probably occur in the Bridger. Kowhere in Europe do we find the Dinoeerata and Tillodonta of the Bridger. Palceosyops is also unknown in Europe, but it plays the part in America of the Palwotlierium^ from which it does not greatly differ in structure. The latter genus is most largely developed in the Parisian, but is also characteristic of the Ton- grian. Ryrachyus is the American Lophiodon^ the difference between them being but slight : both are found in France ; the former in the Lower Parisian, the latter in the Phosphorites. Tapirulus | Gerv. is a genus common to the Bridger and to more than one horizon of the Parisian. The squirrel-like rodents of the Bridger are like those of the Parisian, but they are not confined to either epoch. The character which distinguishes the Parisian most widely from the Bridger, besides the absence of the Dinoeerata and Tillodonta, is the presence of numerous Selenodont Artiodactyla, as Xipliodon, Cwnotherium, Ampliimeryx, Ano- plotherium, etc. These are of primitive type, it is true j the Anoplotheriidce especially having probably four toes in the very short manus {Uurytherium), including the poUex, and three behind. They also display the character of a fifth crescent of the superior molars, which is wanting in the higher Selenodont types. But even these genera are absent from the Bridger. The ensemble is then, that the latter displays relationships backwards, or to the Suessonian, while the Parisian has a later fades, constituting an approach to the Tongrian and White Eiver. § The following table presents the relations of the Bridger fauna suc- cinctly, but it is much less complete than we hope to make it when its * Notharctus is undistinguistiablc from Adapis in inferior dental characters, tit is described as CynohycBnodon with two species. tGervais, 1850; Helaletes Marsh, 1872. $ See Ann. Rept. U. S. Geol. Surv. Terrs. 1873, pp. 461-2, where this view is proposed. 44 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [Vol.Y. numerous species now represented by catalogues of names are fully de- scribed. The Parisian is here regarded as including the divisions Briix- ellian, Bartonian, and Sestian (Gypse). Parisian. Bridger. Didelphys. fBidelpJiys. Vespertilionidw. VespertilionidcB Flesiarctomys. Plesiarctomys, Tillodonta. Mywnodontidw. Hycenodontidce. Adapts, Adapis. Anaptomorplius (Phosph.). Anaptomorphus. Dinocerata. Falceosyops, Palceotlierium. Lopliiodon. Hyracliyus (Phosph.). HyracJiyus. Taplrulus. Tapiruhis, Anthracotherium. Achwnodon, Cheer opotamus. DicJwhune. Anoplotheritcm, Xiphodon, AmpMmeryx, The rich Tongrian (Stampian) fauna is, according to authors, repre- sented in the Sables deFontainebleau, Puy en Yelay, Eonzon, Hempstead, and Oadibona in Italy. We find here Bidelphys in abundance, Hyoenodon, AmpMcyon^ Cynodon^ Palwotherium, Paloplotlierium^ Chalicotherium^ and Aceratherium. Of Artiodactyla, the Suillines are Anthracotherium and Motherium ; the Selenodonts, Hyopotamus and Gelocm. This list is the nearest known counterpart of that of the fauna of the White Eiver epoch of IS'orth America. To reproduce the latter, we must omit from the above catalogue the genera of Palceotheriidcey and replace them by the allied Chalicotheroid Menodus and Symhorodon, subtract Anthra- cotherium, and add the great body of the Oreodontidce. Then there are included in the White Elver fauna the higher Selenodont Artiodactyles of the Poebrotheriidce and Hypertragulidce, the corresponding types of which belong to the fauna of St. Gerand le Puy in France, or the Aquitanian epoch, which directly succeeded the Stampian. In Eu- rope we have here Dremotherium, Amphitragulus, Lophiomeryx, Dor- catherium ; in America, Leptomeryx, Hypertragulus, Kypisodus, and Poe- brotherium. It is curious that while Leptomeryx is also European,* it has not yet been found above the Phosphorites. Among Suillines, the Palceochcerus t of the Oregon White Eiver beds has also not been found * I think M. Filhol's Frodremotherium is identical with Leptomeryx. i Thinohyus Marsh appears to be the same. COPE ON HORIZONS OF EXTINCT VERTEBRATA. 45 below tlio Aquitanian in Europe. But the American Didelphys,* Hyce- nodon^ Amphici/on, Ulothcrnm, and Hyopotamus^ with tlie numerous Chalieotlieroid species, show ck^arly that the White Kiver fauna may be looked upon as a mixture of those of the Stampian and Aquitanian, the former of which is sometimes referred with reason to the TIi)per Eocene, wliile the latter is always left in the lowest Miocene. And the solution of this question of position as regards the White Kiverbeds appears to me to be at present by no means easy.t Accordhig to the system of Nau- mann, it should be called Oligocene. Although Artiodactyles witli Selenodont molars are far more abundant in both continents during this i^eriod than the last, a remarkable differ- ence is to be observed between them. Those of Europe still largely consist of the types witli five crescents, as represented by the numerous Hyopotami and Gcenotlieria^ while in America the modern four- crescent- bearing molar characterizes almost the entire suborder, the only excep- tion being two si)ecies of Hyopotamus. The following table will represent the relations of the White River fauna: Stampian and Aquitanian. White Eiver. Didelphys. Didelphys, Leptictidce. Protomyidce, Protomyidce.X Saccomyidce.^ Steneofiher, Steneofiher, Leporidce. Leporidce. Hycenodon. Hycenodon, A7np7iicyon, Ampliicyon, Canis, Canis. Temnocyon, Enliydrocyon, Gulo.W Gulo.W Dinictis. Machcerodus, Machcerodus. Clialicotheriidce, Palceotheriidce. Hyracodon, AceratJierium, Ancliitlierium, Elotherium. MotJierium, Palceochcerus. Palceochcerus, * Herpetotherium Cope ; Feratherium Aym. t See Ann. Report U. S. Geol. Surv. Terrs. 1873, p. 4C2, where the Wliite River beds are determined as Lower Miocene. t Ischyromys Leidy. $ Entoptychus and Pleurolicus Cope. WAmphictis Pom. 46 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [Yol.Y. Stampian and Aquitanian. AnthracotheriuM. White Eiver. Oreodontidw. Anoplotheriidce. Ryo;potamidce. Merycopater, Poehrotherium. Hypertragulus. LopMomeryx, AmpMtragulus. Leptomeryx (Phospli.). Dremotherium. Leptomeryx. Hypisodus, The Faliinian epoch includes in the large sense the Langhian, Helve- tian, and Tortonian divisions, embracing the rich deposits of the Orl^- anais, of Simorre, and of Sansan. We have here the true Miocene fauna, of which the following genera are characteristic : Edentata, Macrotlie- nMm;Prol)oscidea, Dmo^Aermm, Mastodon; Perissodactyla, Ancliitherium, Listriodon; Artiodactyla, Palceomeryx^ Picrocerus, Procervulus ; * Car- nivora, Amphicyouj Hycenarctos^ f Machcerodus ; Quadrumana, PUopi- thecus. The ancient genera AnthracotJierium and Ccenotherium continue throughout, and the existing genera Arvicola^ Lutra^ and Sus appear. The succeeding epoch, the Oeningian, including with it the horizons of Epplesheim and Pikermi, presents the additional genera Porcatlie- riunij Helladotlierivm^ several genera allied to Antilope, with Hippothe- rium^ the huge edentate Ancylotherium^ and the monkey Mesopithecus. It is from these materials that we must determine by comparison the American Loup Fork epoch, whose deposits are widely spread, and whose fauna is of well-marked character. Although called by my pre- decessors Pliocene in age, I have insisted that it should be referred to the Miocene series, and I think that the evidence to that effect, which I have produced, will be found conclusive. l!^evertheless, here, as in other American Tertiary horizons, the element of geographical peculiarity en- ters, and diminishes the number of identical types. Falunian. Loup Fork. Steneofiber, Steneofiber. Macrotherium. ) Ancylotlieriim, ] Amphicyon. Morotherium, AmpMcyonA Ccenohasileus. Tetralopliodon, Aphelops. Pinotlierium. Tetralopliodon, Aceratherium. Anchitlierium, Listriodon, *Gaudry, 1878; Dicrocerm Cope, 1874 (not Lartet); Merycodus et Cosoryx Leidy, nomina nuda. t Cdiiia ursinus Cope. No. 1.] COPE ON HORIZONS OF EXTINCT VERTEBRATA. 47 Falunian. Loup Fork. Hippotherium (Oeningian). Hippotherium, Ccenotheriim. Anthracotherium, Frotohippus. Hippidium.* Oreodontidce, Palwomeryx, Dicrocerus. Frocervulm, Frocervulus. Frotolabis. Frocamelm. Blastomeryx, The existing genera mentioned as found in the Falunian fauna are paralleled by the BicotyleSj Hystrix, and Mustela of the Loup Fork beds. It is evident that this latter horizon retains in its Oreodontidce the same traces of antiquity that the Falunian does in its Cwnotherium, but shows a more modern aspect in the omission of Anchitherium and its replacement by Hippotherium and Frotohippus^ and in the still more modern type Hip- pidium. Although but six genera of the two continents are determined as identical in the above table, yet others, which are facing on the same hue, are very nearly allied. Other differences are geographical. The fades of the Loup Fork horizon is then a compound of that of the Falu- nian and Oeningian, or Middle and Upper Miocene. In commenting on the above-described fauna in 1874,f I remarked that ^'the proper discrimination of American Pliocene remains to be accom- plished." It was not long after that date that material for making the identification of this horizon on this continent first came into my hands. This was derived from the superior Tertiary of Oregon, and includes a considerable number of species of fishes, birds, and Mammalia. I pub- lished a list of some of the species in 1878.| The character of the fauna from that region coincides with that which has from time to time been unearthed in the caves and other Eastern deposits to such an extent as to lead us to suspect that the differences between them are geographical only. In Europe, the Pliocene, or Subapennine, includes, according to D'Orbigny (1855) and Gaudry (1878), the Plaisancian and Astian, which are represented at the following localities : Plaisancian. — Montpellior; Casino (Tuscany). Astian. — Perrier, near Issoir, Coupet, Vialettc (Upper Loire), Chagny; English Crag ; part of deposits of the Val d'Arno. The characteristic of this fauna is the fact that the species belong mostly to existing genera, the chief exception being ^^/?290^/ierm?>^. The horses are chiefly represented by Equus. Common genera are ArctomySy * Fliohippus Marsh. t Report Lieut. G. M. Wheeler, IV, Palaeontology of New Mexico, 1874, p. 364. tBuU. Hayden's U. S. Geol. Surv. Terrs, iv, 1878, p. 389. Subapennine. 48 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [VolY. Lepus^ WephaSj Mastodon^ Tapirus, 8us, Cervus, Antilope, Bos, Canis, Machwrodiis, Felis, Ursus. In the Uquus beds of Oregon, a few extinct genera in like manner share the field with various recent ones, while not a few of the bones are not distinguishable from those of recent species. I give the following list, the extinct species being in italics : Mylodon sodalis. Thomomys (nr.) clusius. Tliomoinys talpoides. Castor fiber. Lutra near piscinaria. Auchenia magna. Auchenia vitalceriana. Cervus fortis. Canis latrans. ElepJias primigenius. Equus occidentalis. Equns major, Auchenia hesterna. The species derived from the cave formations of the Eastern States are more numerous, and differ from the Oregon fauna in many respects j yet the parallelism is close in the genera with the Equus beds on the one hand and the Pliocene of Europe and South America on the other. The differences distinguishing it from the Equus beds of Oregon are, however, such as compel me to regard it as a distinct division of the Pliocene, under the name of the Megalonyx beds. Megatherium (p). Mylodon (p). Megalonyx (p). Sciurus (s). Arctomys (s). Jaculus. Arvicola (s). Erethizon. Hydrochcerus (p). Machcerodus (sp). Mastodon (sp). Equus (sp). f Hippotlierium (s). Tapirus (s). Dicotyles (p). Cariacus (p). Bos (s). Castoroides. Lagomys (s). Lepus (s). Anomodon. Scalops. Arctotherium (p). Procyon. Canis (sp). Mustela (sp). In the above list, the extinct genera are marked in italics. There exists, as a marked feature of the jN^orth American Pliocene, to which I called attention several years ago,* a considerable representation of the fauna of the Pampean formation of South America : such are twelve genera, of which six are extinct genera, and four are peculiar to that formation and fauna. The genera found in the Pampean are marked(^), and those of the Subapennine {s). Id the list from the Oregon localities, Mylodon and Auchenia were observed to be the only distinctively Pampean genera. As a conclusion of the comparison of the American Equus beds in general with those of Europe, it may be stated that the number of identical gen- era is so large that we may not hesitate to parallelize them as stratigraph- ically the same. On the other hand, the agreement with the South American Pampean formation is so marked in some respects as to induce us to believe that the distinction is geographic rather than stratigraphic. Believing that the Pampean formation contains too large a percentage of extinct genera to be properly regarded, as it has been, as Postphocene or Quaternary, its characters, both essentially and as a result of the com- parison which I have been able to make, refer it properly to the Pliocene. *Proc. Acad. Phila. 1857, 156; Proc. Am. Philos. Soc. 1869, 178. l\0. 1.] COrE ON HORIZONS OF EXTINCT VERTEBRATA. 49 It appears, then, that the term riioceiie or Subapemiine is applicabk^ to the horizon of this IViuna in Europe and North and South America. Rl^SUME OF COMPARISONS. The conckisions to be derived from the facts enumerated in the pre ceding: pages are as follows : I. Portions of all the fauncT. of all the primary divisions of geologic time have been recognized on both the European and North American con tinents. II. Parallels requiring general identification of principal divisions of these faunaB may be detected. These are: the Coal-Measures; the Per mian ; the Laramie ; the Maestrichtian ; the Eocene ; the Miocene. III. Exact identifications of restricted divisions maybe made in a few instances only; such are the Turonian and the Niobrara ; the Suessonian. and the Wasatch ; the Eqiius beds and the Pliocene. It is not impossible that some of the relations mentioned in II will be- by the accession of further information, referrible to the list of exact comi)arisons in III. In all cases of identification it will be necessary to^ employ the name first proposed with definition, for the horizon, other names taking places as synonymes. But in the majority of strata it will be necessary to preserve the local names: thus those of Judith Eiver, Bridger, White River, and Loup Fork, applying to beds having no exact eqtlivalents in Europe, cannot be set aside for older ones, but must themselves be applied to corresponding faunal horizons elsewhere, should any such be found in future. And it will rarely happen that the minor subdivisions of such faunse will be found to have an extent sufficient to warrant their having other than local names. In the accompanying diagram the series of strata of Europe and North America, as determined by their palaeontology, are placed side by side for the purpose of comparison. Complete parallelism can only be predicated . of divisions of the first order separated by horizontal lines. Such rela tion is indicated by exact opposition of the areas representing the epochs in question. In giving the minor divisions of the European epochs I have generally restricted myself to those of the epochs which have American equivalents. Where there is no equivalent on one side or the other, the vacancy is represented by a diagonal line. In emi)loying names for epochs and their divisions, I have adhered to the law of priority as far as^ my knowledge of the literature allows.* I have given a few names to American formations, but only in instances where such had not been previously given. In such cases I have preferred employing the name of some characteristic genus of fossils, rather than one of local origin. * lu the European system I have been much aided by the atlas of Prof. Renevier of Lau- sanne, and by the writings of Woodward, Gervais, Hubert, Pomel, Gaudry, Filhol, etc. Bull. V, 1 4 50 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [VolY. WEST EUROPE. NORTH AMERICA. Astian. Pliocene. tiary. Pliocene. Equus beds. Plaisanciaii. Megalonyx beds. Oeniiigian. Oeningian. Loup Fork. Procamelus beds. Ticholeptus beds. Tortonian. Langhian. Falunian. White River. Oregon beds. White River. Aquitanian. Aquitanian. Stampian. Tongrian. Ter Sestian. Parisian. Uinta. 1 Uinta. Bartonian. Bruxellian. Bridger. Bridger. Suessonian. Suessonian. Wasatch. Green River. Wasatch. Tliauetian. Tliauetian. I . t) « t« 3 V 0 M V OB U ? Puerco. Puerco. Laramie. Fort Union. Judith River. Maestrichtian. Campanian. Santoixian. Seuonian. Cretaceous. Fox Hills. Fox Hills. Fort Pierre. Turonian. Cenomanian. Colorado. Niobrara. Carentonian. Fort Benton. Rliotomagian. Dakota. Dakota. Vraconian. Albian. Gault. Aptian. Rhodauian. TJrgonian. Urg.iptian. Hauterivian. TTealden. Valangian. c 'at IB es fa Camarasaurus beds. Purbeckian. Portlandian. Kimmericlgian. Portland. j Coralliau. 1 1 Oxfordian. 1 1 Bathian. 1 1 Upper Lias. 1 1 Lower Lias. No.i.] COPE ON HORIZONS OP EXTINCT VERTEBRATA. 51 WEST EUROPE— Continued. NORTH AMERICA— Continued. Rhaotic. Trias* Karnian. Konpor. K(^iipcr. Norian. Muschelkalk. Clep8ydrop.s shales. Thuringian. Permian. Carboniferous. Permian. Lodovian. Eryops beds. Coal-Measures. Coal period. Coal i)eriod. Coal-Measures. Conglomerate. Conglomerate. Mountain limestone. Mountain limestone. Eammcnian. Upper Devonian. Upper Devonian. Catskill. Chemung. Hamilton. Eifelian. Middle Devonian. Lower Devonian. Corniferous. Oriskany. Coblenzian. Lower Devonian. Ledburian. Upper Silurian. Silurian. Upper Silurian. Lower Helderberg. Ludlovian. Salina. Wenlockian. Niagara. Llandoverian. Caradocian. Lower Silurian. Lower Silurian. TTn fl Q ATI Llandeilian. Tremadocian. Trenton. X rimorciiai. Calciferous. Potsdam. 8 cs K ja u fa 52 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [VolY. The oldest of tliese I liave called tlie i^ryoj)."? beds, from the most abun- dant genus of Labyrinthodonts which is found in it. They contain, also, abundance of other Yertehrata^ none of which are higher than lieptilia (order Thero^norjoha), Avith plants, mollusks, etc. They consist of sand- stones, alternating with beds of red clay and coarse conglomerate and sphserosiderite, etc. They are chiefly distributed in i^^orthern Texas and Southern Indian Territory. The Clepsy drops shale, named by me in 18G5, forms a thin stratum, in Southeast Illinois and Southwest Indiana, consisting of black and rarely teddish carbonaceous shales and clays. These appear in some places to lie conformably on the Coal-Measures, to which they have been referred by previous geologists, but CoUett, Gibson, and others have shown that it is unconformable over considerable areas. It does not belong to the Coal-Measures. The Puerco marls were first observed by me in Kew Mexico in 1874, and afterwards found to have an extensive development in Southwestern Colorado, by Endlich, in 1875. He has referred them to the lowest place in the Tertiary series, but the absence of fossils renders it difficult to conclude whether they belong here or in the Laramie series. The Oregon White Eiver beds differ from those found east of the Rocky Mountains, although they contain a majority of the same genera, and many of the same species. They are wanting in the important genera Symhorodon and Memdus. To represent these genera, they have Dwo^ don J and, in addition, some peculiar genera of Bodentia, as Entoptyclms, PleurolicuSj and Meniscomys^ and the Suilline Palccochoerus. Among Car- nivora, the genus Enliydrocyon is, so far as known, characteristic of them. The Loup Fork beds are represented in the valley of Smith's Eiver, Montana, by a horizon which may be somewhat older than that hereto- fore known. The fauna presents us with the typical genera Frocamelus^ Hippotlierium^ Frotoliippus^ Mastodon^ and MerycochoeriiSj but, in addition, with the peculiar genera of OreodontidWj TicholeptuSj Cyclopidius, and PWiecisteSj* and with Euminants similar to Fcdceomeryx. These are wanting from the other parts of the formation, and I therefore name the two divisions the Ticholeptus and the Frocamelus beds. I have alreadj^ mentioned the Megalonyx beds as the equivalents in the east of I^^^orth America of the Equiis beds of Oregon and Califor- nia, but which present such important differences that they cannot be identified. The diflerences are displayed in the catalogues already given, the list of the Megalonyx fauna having been derived from the explora- tion of caves in Pennsylvania t, Virginia, and Illinois. The remains of this fauna are by no means found in caves only, but are found in swamps and Pliocene clays. The extinct genera characteristic of the Megalonyx beds are Megatherium^ Megalonyx^ Castoroides^ and Arctotherium ; the genera no longer living in Xorth America, HydrochceruSj Tapirus. * Proc. Am. Philos. Soc. 1877, p. 219. tLoc. fit. 1871, p. 73. iVo. i.J COPE ON HORIZONS OF EXTINCT VERTEBRATA. 53 In conclusion, it may be observed tliat the liicuniii in tlie series as pre- sented by one continent render us dependent on tlie otlier for the evidence necessary for the complete elucidation of the laws of the creation of ani- mal life. Phylogenies can be thus constructed wliich would otherwise be impossible, and the results of researches into the earliest types of Vertebrata become intelligible. Thus I have been able to prove, in sup- port of a thesis publish()d in 187-1, that the earliest Ungulate Mammalia were pentadactyle and plantigrade. I have also shown that the ankle- joint had not, in the primitive Mammalia^ the hinge-lilvc (character that it has in the later ones, but that it is without the interlocking superior articulation. The small size of the brain of early Mammalia, already pointed out by Lartet, has received extensive confirmation by the re- searches of Marsh, who has also shown the progressive increase in size of the whole body in various Mammalian lines. To these results I now add another, which is derived from the study of numerous Permian Ver- iebrata, viz, that the earliest land vertebrates had a persistent chorda dorsalis. COMPARISON WITH THE SCALE DERIVED FROM PALEOBOTANY. I now consider another kind of relation presented by the American and European horizons. I allude to the florae, for my knowledge of which I am necessarily dependent on the labors of others. I first ex- 'fiibit the determinations of the ages of the American formations already discussed, made by Mr. Lesquereux on the basis of the vegetable re- mains which they contain. I place by the side of these my own deter- minations of the ages of the same beds, as already related. The former are derived from the full memoir of Mr. Lesquereux in the Annual Eeport of the United States Geological Survey of the Territories for 1872, i)p. 410-417. It will be observed that there is a constant discrepancy between the two tables. Lesquereux. Formation. Cope. Miocene. Oligocene. Middle Eocene. Lower Eocene. Ui)per Cretaceous. If the determinations of Mr. Lesquereux be correct,* it is evident from the above that the vegetable life of North America reached its present condition one ei)och or period earlier than the higher Vertebrata, and that the nomenclature is thus thrown back by so much. It would ap- pear that the recent flora of North America is a period older than the *Tlie above x>aralle]s are well presented by Dr. Peale in his report to Dr. Haydeu, Ann. Rept. U. S. Geol. Surv. Terrs. 1874, p. 141 et scq. 54 BULLETIN UNITED STATES GEOLOGICAL SURVEY. WolY. fauna, i. e., has persisted longer than the latter by a certain length of geologic time. Applying the same reasoning to the past, I embodied the idea in reference to the Laramie period (^^Fort Union") in the state- ment that ^^a cretaceous fauna was then contemporary with a tertiary, flora"; and, later, tliat ^'an eocene fauna was contemjiorary with a inio- cene flora." It may have to be added that a miocene fauna was contempo- rary with a pliocene flora. Since Mr. Lesquereux has the support of the best palseobotanists of Europe, in his conclusions, it is useless to take the ground assumed by a few of my colleagues, that the former gentle- man has simply erred in his determinations. He gives us grounds for believing that he has not done so, by giving us the European standard by which his identifications are governed.* It is as follows : Lower limits not positively fixed ; largely developed in Italy. ? (Sub- apennine, E.D.C.) Miocene Oeningian; Mayencian; Aqnitanian. Oligoceue . . . Tongrian. Eocene Gypse of Aix; Alum Bay; Mt. Bolca; London Clay; Shei^pey; Gr^s of the Sarthe. ' Upper Landenian ; Sezanne (= Paniselian). Snessonian (Lignitic Soissonais; Sables de Braclieux) ; Lower Lan- Paleocene . . <( den i an. Hersian; Gelinden. ^ ^ Limestone of Mons, overlying nnconformably the Maestriclitian. This system, it will be observed, is almost exactly identical with that employed in the preceding pages as the standard of comparison for the Vertehrata. Yet it has resulted, from a most careful comparison of both faunte and florae of America with this standard scale, that two distinct palseontological series have to be adopted, the one for the vertebrate life and the other for the plants of the Western Continent. If this re- sult be accurate, and there appears to be no avoiding it, an explanation must be sought. There are only two possible ones : either the animal life of E"ortli America has lagged behind that of Europe by one period during past geologic time ; or, secondly, the vegetable life of America has been equally in advance of that of Europe during the same period. In other words, if the plant-life of the continents was contemporaneous, ancient types of animals remained a period longer in Xorth America than in Europe. If animal life was contemporaneous, plant-life had ad- vanced by one period in Europe beyond that which it had attained in ^^'orth America. In either case, the fauna! or the floral standard of esti- mation of geologic age of strata for ^^"orth America is a false one, since there can be but one standard of comparison for anything. But this great fact being understood, the evidence of each of the great depart- ments of life possesses its own intrinsic value. *Aun. Repoi-t U. S. Geol. Surv. Terrs. 1874, p. 285. Art* III.— Obiservations on the Faunae of the ]?Iioeene Tertiaries of Oreg^on. By E. D. €opc. A considerable number of Vertehrata, almost exclusively Mammalia^ have been described by authors from the White Eiver and Pliocene for- mations of Oregon. The descriptions are found in Professor Leidy's con- tribution to the Final Report of the United States Geological Survey under Dr. Hayden (Vol. I) ; in those of Professor Marsh in the American Journal of Science ; in a paper by Mr. Bettany in the Quarterly Journal of the Geological Society of London for 187 G ; and in a i^aper by myself (Paleontological Bulletni No. 30) in the Proceedings of the American Philosophical Society, published in December, 1878.* Having recently had the opportunity of inspecting a considerable amount of material from the horizons in question, I give a list of the species which I have observed. '/Ji^few new ones occur in collections received since the publication of my last paper, and are now described, together with some of interest from the Loup Fork beds of the same region. White Biver Fauna, TESTUDINATA. Stylemys oeegonensis Leidy. EODENTIA. Steneofiber gradatus Cope. Steneofiber ? NEBRAscENSis Leidy. Meniscomys hippodus Cope. Meniscomys multiplicatus Cope, Pleurolicus sulcifrons Cope. Entoptychus cavifrons Cope. Entoptychus planifrons Cope. Entoptychus crassiramis Cope. Pal^olagus haydeni Leidy. * See also the American Naturalist, December, 1878. 55 56 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [VolY. OARNIYORA. HOPLOPHONEUS BRACHYOPS Cope. Mach^rodus strigidens Cope. Enhydrocyon stenocephalus, gen. et sp. nov. Generic Characters. — Dental formula: I. O. |; Pm. f ; M. ^. The superior premolars consist of two ordinary and one sectorial; the first and second are both comi)ressed, two-rooted, and in the typical species with median lobe of posterior cutting edge. The two true mo- lars are transverse and tubercular. The three inferior premolars are all two-rooted, and with posterior lobe in the two known species. The heel of the sectorial is cutting, as in Temnocyon, and the internal tuber- cle is present. There is at least one inferior tubercular tooth ; speci- mens are injured so as not to display a second. In a nearly complete cranium belonging to the typical species of this genus, we observe the shortness of the facial part of the skull as com- pared with the length of the cerebral, and also the constriction of the skull behind the orbits. The zygomatic arches are robust and expanded, and the sagittal crest is high. The auditory bullae are inflated and thin- walled. The dentition of this genus refers it to the Canidce, but the form of the skull resembles that of Putorius vison and Lutra. Specific Characters. — The principal cusps of the inferior premolars present cutting edges, as does the median posterior lobe. In both third and fourth there is a small conic heel posteriorly, but an anterior basal tubercle on the fourth only. The sectorial is large and robust, and the heel is short, with an absolutely median cutting edge. The first tuber- cular is longer than wide, and presents a nearly median cusp in front, which is joined to a low one on the internal border of the crown. The superior canine has an obtuse cutting edge on the anterior and posterior borders of the inner side. The first (third) superior premolar is near to it, and is rather large, displaying a median cutting lobe and low posterior heel. The fourth is similar but larger. The sectorial is much worn in the only specimen where it is preserved ; it is rather short, and is widened anteriorly. The first tubercular is large, and has con- siderable transverse extent ; it is a little wider externally than inter- nally, and has much the form of the corresponding tooth in Canis. The second tubercular is transverse and small, not being much more than lialf the length of the first, and is situated in contact with it. Tlie cranium is remarkable for the anterior i)osition of the orbits, and the associated shortening of the face and lengthening of the parietal region. The orbits look somewhat forwards and very little upwards. The superciliary region is slightly prominent, and there is a prelachrymal concavity. The infraorbital foramen is moderate, and is situated mostly above the i)osterior part of the fourth premolar. The muzzle is flat No.l.] COPE ON MIOCENE FAUN^ OF OKEGON. 57 above, aud the nasal bones are wide, and are not emargmate above the osseous nares, as in many recent Carnivora. Posteriorly, the superior border of the brain-case descends, bat the parietal bones maintain a gently convex outline in their high sagittal crest. The sui)raoccipital region is elevated, and projects posteriorly. Measurements. Specimen No. 1. M. Antero-jiosterior diameter of second superior premolar 0.010 Antero-postorior diameter of third inferior premolar 0.013 Width of base of third inferior premolar 0.0065 Elevation of crown of third inferior premolar 0.010 transverse 0.010 ^ antero-posterior 0.021 Width of first tubercular 0.006 Diameter of inferior sectorial ) Specimen No. 2. Total length of cranium 0. 170 Width across zygomatic arches 0.114 Least width behind orbits 0.024 Depth of cranium with crest at otic bulla 0.070 Vertical diameter of orbit 0.025 Length from orbit to end of muzzle (axial) 0.040 Interorbital width 0.043 Width of muzzle above second premolar 0.018 Length of superior molar scries 0.051 Length of fourth x^remolar 0.012 Length of sectorial 0.016 Length of first tubercular 0.008 Width of first tubercular 0.015 Width of second tubercular 0.0085 The length of the skull is about that of the Coyote, but it is much more robust in all its proportions excepting the postorbital constriction. Discovered by Charles H. Sternberg in the Oregon White Eiver beds of the John Day Eiver region. Enhydrocyon basilatus, sp. nov. This Carnivore is represented by a mandible with coossified rami, which are broken off behind the sectorial teeth. The crowns of the latter and but one incisor and one canine tooth remain. The premolars and one canine are in good preservation. These portions indicate an animal of the same general character as the Enhydrocyon stenoceplialus^ but of larger and more robust propor- tions, and characterized by many dental peculiarities. These will be at once pointed out. The canine is directed upwards and a little outwards, and i)ossesses two obtuse ridges bounding the interior face. The third incisor is compressed and truncate superiorly and distally. The first (second) premolar is two-rooted, compressed, and trilobate. It consists of a principal cutting edge little elevated, and a small accessory lobe 58 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [YolY. at each extremity of the crown 5 its base is expanded posteriorly. The principal cusp of the third premolar is more elevated, and, besides the anterior and posterior tubercles, there is a basal posterior heel, which is continued as an exiDansion of the inner base of the crown. In the fourth premolar, the base of the crown is expanded, especially posteriorly ; the principal cusp has a nearly circular section at the base, and the posterior median lobe is a subconic tubercle standing on the middle of the heel. The sectorial is large and relatively rather narrow, but the details of its form are not ascertainable. Measurements, M. Length of dental series, including canine and sectorial 0.076 Length of the base of tlie sectorial 0.024 Length of the premolar series 0.037 Length of the fourth premolar , 0.016 Width of the fourth premolar 0.009 Length of the third premolar 0.013 Width of the third premolar 0.008 Length of the second premolar 0.009 Width between centres of crowns of fourth premolars 0.034 Length of symphysis 0.035 This species was probably of the dimensions of the Gray Wolf. Found by Mr. Sternberg in the same region as the U. stenocephalus, Temnocyon altigenis Cope. Oanis hartshornianus Cope Canis geismarianus Cope. Canis cuspigerus Cope. Canis lippincottianus Cope. Canis gregarius Cope. PEKISSODACTYLA. D^ODON SHOSHONENSIS CopC. Aceratherium pacificum Leidy. Anchitherium equiceps Cope. Anchitherium brachylophum Cope. Anchitherium longicriste Cope. ARTIODACTYLA. Elotherium imperator Leidy. Pal^ochgerus condoni Marsh. Pal^ochcerus pristinus Leidy. Pal^ochcerus sociALis Marsh. Merycopater guyotianus Cope. HoA.] COPE ON MIOCENE FAUN^ OF OREGON. 69 BucKOTAPiius SUPE1113US Leidy. EUCIIOTAPIIUS OCCIDENTALIS Maisll. Merycociicekus leidyi Bettauy. Merycochcerus temporalis Bettany. POEBROTnERIUM STERNBERGII, Sp. nOV. This Iviiiuiiiaiit is represeuted by a considerable part of the skeleton with both mandibular rami supporting the teeth, of one individual. The bones are all in close proximity, and sometimes in undisturbed relation, in a single block of stone. The species to which I give the above name presents the characters already ascribed to the genus Foehrotherium by Leidy as regards cranial features, and by myself as regards the rest of the skeleton. The third and fourth metacarpals are not coossified, and the second and fifth are not distinguishable. The preservation of the premaxillary bone in this species enables me to demonstrate the i)resence of superior incisor teeth, a character the presence of which I have heretofore only inferred. As compared with the P. vilsoniy the species differs in its superior size and greateir relative robustness. This is seen in the greater depth of the mandibular ramus, and the greater stoutness of the metapodial and other limb-bones. The last inferior molar tooth jjresents a character- istic peculiarity. The anterior external cusp is separated by a deep groove which divides the external side of the crown to the base from the succeeding cusp. It results that on trituration, the anterior exter- nal crescent is isolated, and does not communicate by its posterior horn with the succeeding crescent, as in P. vilsoni. The last premolar is more robust than that of the P. vilsoni^ the width of the half-worn surface being half the length of the tooth and enclosing behind an enamel fossa. In P. vilsonij this tooth is more compressed, and the fossa is represented by an open groove. The first inferior i)remolar occui)ies the middle of the diastema following the canine, instead of standing near the canine as in P. vilsoni. Measurements, M. antero-posterior 0.020 transverse 0.009 antero-posterior 0.013 transverse 0.009 Depth, of ramus at second molar 0.025 Length of ramus from third molar to extreme posterior edge 0.061 Length of metacarpus 0.178 Transverse x>roximal diameter of the two metacarpi 0.029 Greatest diameter of the head of the humerus 0.053 Antero-posterior diameter of the condyle of the femur 0.048 This species is named in honor of Charles H. Sternberg, the indefati- gable explorer of the fossil deposits of the West. BoocHOERUS HUMEROSUS, gen. et sp. nov. Generic Characters. — The species on which this genus is founded, is represented by a part of the skeleton, which is unfortunately not accom- Diameter of last molar Diameter of penultimate molar.. 60 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [VolY. panied by cranial bones or teeth. The characters are, however, suffi- ciently clear for purposes of determination. The great tuberosity of the humerus is produced beyond the head, and does not close round the bicipital groove. The intertrochlear ridge is well developed, and there is no internal epicondyle ; the external epicondyle is moderately devel- oped. On the carpal extremity of the ulno-radius, the facets of the scaphoid, lunar, and cuneiform bones, are distinguished by strong oblique ridges, and the last named is nearly in the horizontal line of the two others. In the carpus, the trai czoides is distinct, and the trapezium wanting. The unciform is in contact with the lunar. Metacarpals two, distinct from each other, with free rudiments of the second and fifth at their proximal extremities. Their distal keels confined to the posterior faces of their extremities. Phalanges depressed; ungues short, obtuse. In the above description is found a combination of characters not known to me to exist in any recent or extinct genus of Artiodactyla. Several of its features indicate affinity to the suilline division, Avhile others point to the Buminantia. The imperfect distal articulation of the metacarpals is characteristic of the extinct types Oreodontidce and Poebrotheriidcej and the two distinct metacarpals constitute the resem- blance to the latter family the stronger. The latter character is, how- ever, not inconsistent with the Omnivora^ and the depressed phalanges add to the weight of affinity in this direction. The distal extremity of the humerus is much like that of a peccary. The distal articular sur- face of the ulno-radius points, however, again to the Buminantia of the group Fecora, displaying a specialization quite in contrast with the primitive character of the metacarpo-phalangeal articulation. From these considerations it can be seen that it is not easy to affirm whether this genus possesses bunodont or selenodont dentition. If I may ven- ture an inference as to the affinities of the genus, I would suggest that it will be found to be as nearly allied to the Hypertragiilidce as to the Oreodontidce^ though not without suilline affinities. Specific Characters. — The only species of this genus known to me is of large size. It is represented by the greater part of a scapula and both anterior limbs and feet; by the pelvis, femur, and part of tibia, and by some vertebrae; all found in immediate association by Mr. Sternberg. These remains indicate an animal of the size of the Bhinocerus indicus. The animal is characterized by the massive proportions of the humerus as compared with the femur, and by the short, robust form of the meta- carpals. In the humerus, the external border of the great tuberosity is entire, and is not reverted, but descends backwards like the remainder of the surface. The apex of the great tuberosity is much recurved, rising steeply proximad of the head. The bicipital groove is deep. The lesser trochanter is large and simply conic ; its transverse extent is not great. External to its base is a small tuberosity, which is represented in Bos, but not in Dieotyles or Sus, The deltoid crest is very ijrominent, de- NoA] COPE ON MIOCENE FAUNAE OF OREGON. 61 Diameter of lieaclj scoiuling' to the middle of tlie leii^^tli of the humerus, before abruptly sinking to the shaft. Its continuation is very prominent as it crosses the axis of the shaft and becomes the anterior bounding ridge of the hiternal side of the distal extremity. The section of the shaft is thus siibtriangular at all i)oints, the obtuse apex of the triangles revolving from the external side proximally, to the internal distally. The external epicondyle is proximal to the condyle, as in IHcotyles^ but is more prom- inent than in that genus, and more as in Oreodori. It is the extremity of the external acute edge of the humerus, which, rising from the shaft at a point 90o j)osterior to the extremity of the deltoid ridge, turns for- wards to the external epicondyle. The condyles are transverse and not much contracted from side to side. Tlie intertrochlear ridge is sharper than in the species of Oreodon^ Dicotyles^ or Sus, and is continued round to the shaft anteriorly, as in Bos. The condyles otherwise resemble those of Dicotyles^ not being so contracted in their tree margins as in Bos, Measurements of the Humerus. M. Total length 0.500 Length from middle of head 0.425 Tw- - r -1 1 ^ antero-posterior 0.140 Diameter of proximal eiid< ^ ^ . ^ \ transverse 0.170 antero-posterior 0.100 transverse 0.100 Width of humerus near extremity of deltoid crest 0.130 Diameter of shaft iust below extremity of deltoid crest \ ^^^^^^ posteiior ^'^^f X transverse 0.065 Width at epicondyle 0.130 Transverse diameter of condyles 0.120 f intemally 0.070 Antero-posterior diameter of condyles^ at constriction 0.055 ( externally 0.050 This bone has about the size of the corresponding one of the RMno- cerus Indicus. The carpal extremity of the idno-radius is extended transversely. The cuneiform or ulnar articular face forms posteriorly two-lifths the entire extremity, and is oidy recurved in the external part of its posterior border, which is very concave. The ridge which separates it from the Innar surface is very oblique, following just outside of the ulno-radial suture, and contracting the cuneiform facet anteriorly. Distally and posteriorly it forms tlie external border of the posteriorly reverted lunar facet, bounding a deep fossa, which is posterior to the cuneiform facet on its inner side. The lunar facet widens behind at the expense of the scaphoid, so that the scapho-lunar ridge is even more oblique than the luno-cuneiform. This ridge disappears supero-anteriorly, and the lunar facet is recurved upwards, occupying the distal extremity of a strong median ridge of the ulno-radius. The reverted portion is almost a half ch'cle in outlme, and is partly continuous with the scai)hoid facet. The 62 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [YolY. latter is siibtriangiilar in outline, its apex being the point of conjunction of the scapho-lunar ridge with the internal border, just posterior to its greatest convexity. Its superior boundary is interrupted by the wide groove whicli separates the median ridge from the internal border of the distal part of the radius. But for the extent and transverse position of the ulnar portion of this articular face, it might be regarded as pertaining to a typical Euminant. The length of tlie carpus is about three-fourths its width, the proximal elements being larger than the distal. A feature of its anterior face is the close approximation of the inferior angle of the lunar to the superior angle of tlie third metacarpus, which allows the magnum and unciform a very slight contact. The external face of the scaphoid is chiefly lat- eral; its posterior border is a vertical, short tuberosity. The proximal face is abrux)tly decurved at the anterior outer angle, to meet the lunar. There are two separate oval superior lunar facets, and one narrow an- terior inferior one. The inferior face is nearly equally divided by a low cross ridge which fits a concavity of the posterior part of the mag- inim. The lunar is the largest bone of the carpus. Its x)roximal face is decurved anteriorly, posteriorly, and on each side, and is twice as long as wide at the middle. The postero-internal and antero-external angles are produced, the latter into a compressed process which articu- lates with the adjacent angle of the cuneiform. Besides this facet there is one other for the cuneiform, which occupies the posterior half of the inferior part of the outer side, and is separated from the superior edge by a deep groove. The unciform facet is in front nearly as wide as that of the magnum, but grows gradually narrower posteriorly. The facet for the magnum is concave, and grows very wide posteriorly, with the posterior internal angle produced downwards. The proximal facet of the cuneiform is very concave, tiie anterior and posterior borders being elevated, and the internal and external decurved; the latter prolonged a little backwards. Below this extremity on the external aspect is a fossa. The pisiform facet makes an angle of 90° with the ulnar, and extends behind and along the posterior edge of the latter to its' apex. The unciform facet is simple, and is in shape a right-angled triangle with convex hypothenuse. The posterior aspect of this bone is concave. The proximal face of the trapezoides is longer than wide, convex an- tcro-posteriorly, and subdiamond-shaped. The inferior face is narrow subdiamond-shaped, and has less than half the area of the proximal. There is a small round tuberosity on the posterior border, and no trace Measurements of the Ulno-radius. Transverse diameter, total . Transverse diameter of ulna Antero-posterior diameter of articular face j' externally j at middle of ulna M. 0.110 0.050 0.035 0.021 0.055 0.045 No. I.] COPE ON MIOCENE FAUNiE OF OREGON. 63 of facet for a trapezium. The anterior face of the mai^niiin is wider than long', and is divided into three phmes. Tlie i)roximal extremity is di- vided into two areas by a high antero-i)osterior keel. The inner area is the larger, and is bounded by the entire sui)erior border of the anterior face of the bone. The outer area, or the lunar facet, extends very ob- liquely downwards, most so in front, where it forms the external side of the magnum. It is interrupted by a large sinus, which leaves the pos- terior portion of the face narrow and transverse. Behind it is first ? fossa and then two short tuberosities, one above and external to the other. The inferior face is undivided, and is concave antero-posteriorly, and convex transversely. The unciform is the second bone of the car- pus in size. Its anterior face is broader than long, and is convex trans- versely. The inner face has in front a large trapezoidal facet for the third metacarpus, which is only separated from that of the lunar by the angle. The superior face is divided, by an angular ridge nearly parallel with the inner border, into two unequal faces for the lunar and cunei- form. The latter is half as wide posteriorly as long, and terminates an- teriorly in an obtuse angle. The distal face is undivided, but is recurved postero-externally, apparently offering a narrow facet for the fifth meta- carpus. This face nearly meets the cuneiform face posteriorly. Behind both the unciform is produced into a decurved, subconic tuberosity. Measurements of the Carpus/^ M. antero-posterior 0.066 Diameter of scaphoid ^. , ^ ^ longitudinal 0.048 Diameter of proximal face of scaphoid Diameter of distal face of scaphoid . . . antero-posterior 0.048 transverse 0.030 antero-posterior 0.048 transverse 0.023 , f anteriorly 0.047 Diameter of proximal face of lunar ^ *" ^ at middle 0.032 longitudinal 0.053 longitudinal 0.048 transverse 0.032 Diameter of cuneiform 5 antero-posterior, oblique 0.059 ^ transverse, behind 0.040 ( antero-posterior 0.040 \ transverse 0.044 c antero-posterior 0.040 \ transverse 0.031 Length of trapezoides 0.027 Diameter ofproximal face of trapezoides... \ antero-posterior 0.029 (transverse ^ 0.016 Diameter of distal end of trapezoides \ antero-posterior 0.020 C transverse 0.010 f ( anteriorly 0.029 longitudinal < , n /> n^n Diameter of magnum \ i externally 0.040 I transverse, posteiiorly 0.047 ^ antero-posteriorly, outer side 0.055 Diameter of anterior face of lunar Diameter of proximal face of cuneiform . Diameter of distal face of cuneiform * These measurements are always the greatest, and are axial, or in straight lines. 64 BULLETIN UNITED STATES GEOLOGICAL SURVEY. \VolY. Measurements of the Carpus — Continued. M. antero-posteriorly 0.046 transversely 0.031 Diameter of distal face of magnum \ antero-posteriorly 0.040 t transversely 0.048 { antero-posterior 0.080 Diameter of unciform ) transverse 0.053 V longitudinal, in front 0.040 Diameterof lunar facet 5 antero-posterior 0.047 ^ transverse (least) 0.016 Diameterof cuneiform facet \ antero-posterior 0.052 ( transverse (behind) 0.028 Diameter of 4tli metacarpal facet \ antero-posterior 0.044 ( transverse 0.046 As already remarked, this carpus displays resemblances to some re- cent types, and possesses some which are not known among living Artiodactyla. The inferior face of the scaphoid is narrower from side to side than in the Buminantia or 8us^ Dicotyles only approaching but not equalling it in this respect. The strong inferior keel of the lunar ex- ceeds that seen in any of the Buminantia or Omnivora. The pisiform facet is more oblique than in those groups. The posterior tuberosities of the magnum and unciform are larger than those of the genera of either group, while the distinct trapezoides, the slightly shortened mag- num and unciform, and slight decurvature of the cuneiform facet of the unciform posteriorly, are suilline characters. The metacarpals are robust, but flattened antero-posteriorly. The rudi- ments of the second and fifth are free, and the latter is the larger. The third has a considerable surface of contact with the unciform; its an- terior face displays two shallow fossae, one at each superior angle. The corresponding positions on the fourth metacarpal are occupied by two low bosses. Otherwise the surfaces of the shafts of these bones are uni- form. The phalangeal articular face is well reverted anteriorly and posteriorly, and is not bounded by a transverse depression anteriorly above. The carina is short, though prominent, and extends to the middle of the distal extremity. The lateral distal tuberosities are very low. Measurements of the Metacarpals, M. Length of M. Ill 0.210 Length of M.I V 0.190 transverse (total) 0.063 antero-posterior 0.055 T^- X TTT T X 11 1 transvei-se 0.053 Diameter of M. Ill distally < . ^ • / i n n^r ' antero-posterior (chord) 0.045 transverse 0.054 antero-posterior 0.045 transverse 0.043 antero-posterior (chord) 0.048 Length of M. II 0.035 Length of M.V 0.040 Diameter of M. V antero-posteriorly 0.021 Diameter of M. Ill proximally . Diameter of M. IV proximally Diameter of M. IV distally No. I.] COPE ON MIOCENE FAUNvE OF OREGON. 65 Tho 2)JiaJanf/cs are more depressed tlian in any genus of Artiodacfyla known to me, excepting Hipiwpotamiis. The proximal articular surface of the first is gently concave, witli the anterior border not produced. The shaft is not contracted, and is regularly convex above or anteriorly. The distal articular face is narrower above and not produced. The superior border of the proximal face of the second i)halange is i)roduced medially. The distal face is narrowed and produced upwards, so as to stand in high relief, from which it results that the middle of the shaft is very concave above. The external and internal borders of the inferior or i)osterior part of the distal face, are produced backwards, covering lateral basal ridges of more than half the length of the shaft, Avhich form the inferior border of lateral fossae. One unguis is preserved. It is distinct in form from that of Hippopotamus^ Siis, or DicotyleSj and re- sembles that of the llama. It is short, obtuse, and compressed. The external face is nearly plane fore and aft, and slightly convex vertically. The inner is convex fore and aft, and concave vertically. The profile descends steeply to the apex, the curve commencing but little beyond the base. The inferior face is at right angles to the interior face, and is moderately wide. Measurements of the Phalanges. M. Median length of first of M. IV 0.066 T, . T T , ( antero-posterior 0.042 Proximal diameter ^ ^^k- I transverse 0.05d T^. , , T , C antero-posterior (median) . . 0.030 Distal diameter ? ^ , ; ' „ ( transverse (greatest) O.Ouo Median length of second phalange 0.055 rx- X TIT . n ( antero-posterior 0.035 Diameter oi second phalange proximally < , ^ r. ° ^ \ transverse 0.045 Dijimeter of second phalange distally \ antero-posterior . * (transverse 0.041 Length of ungual phalange below 0.042 Proximal diameter of ungual phalange antero-jiosterior 0.032 transverse 0.025 femur is slender as compared with the humerus, and of moderate length. The great trochanter is produced, but not beyond the line of the convexity of the head, and is not much recurved. The expanse externally is about as great as that of the head internally. The tro- chanteric fossa is not large, and is cut off below by a plane surface at the base of the great trochanter, whose superior border forms a curved line connecting the great and little trochanters. The latter is large and projects well iuAvards. The fossa ligament i teris is large and central, having no connection with the border of the head of the femur. The posterior side of the shaft is fiat, and the anterior face regularly convex. The two faces meet •externally in a well-marked representative of the linea aspera. The rotular face of the femur is short and wide, with the borders somewhat oblique, and the inner edge is higlier than the outer at its proximal part. It is strongly convex from above downwards, and Bull. V, 66 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [Vol.Y. does not connect below with the condylar surfaces. Its entire inferior border is well defined and angulate. The condyles are well separated, and the intercondylar fossa is wide above or anteriorly. The external is a little the larger, and the internal is distinguished by the presence of a deep lateral fossa. The head of the tibia is characterized by a distinct bifid spine, and a low, wide crest, which presents an open, shallow groove forwards. The articular faces are of subequal width, but the external is shortened an- teriorly by the usual notch 5 it is also decurved posteriorly. Measurements of the Femur. M. Expanse proximally (greatest) 0.170 Expanse proximally at extremity 0.150 Diameter of liead 0.0G8 -p.. , r-ir., , -m^ antero.posteriorly 0.053 Diameter of shaft at middle < ^ 1 I transversely 0.060 Length of chord of rotular face ^ . . . 0.088 Width of rotular face 0.070 Expanse of condyles (greatest) 0.122 Greatest chord of distal end of femur 0.110 Measurements of the Tibia. M. ^. , r,^ J /. x-i • ( antero-posterior 0.130 Diameter of head of tibia < , I transverse 0.125 The form of the head of the tibia is much like the corresponding region in Oreodon culbertsoni; but the characters of the femur do not resemble those of that species, particularly as regards the distal ex- tremity. A peculiarity of the long bones of this species is seen in their very large medullary cavities. This is especially true of the humerus, whose walls are remarkably thin; those of the femur are thicker. This species was found by Mr. C. H. Sternberg in the John Day Eiver region. Leptomekyx evansi Leidy. Hypertragulus calcarattts Cope. This species is much more abundant in the John Day River deposit than the Leptomeryx evansi. The two genera represent a peculiar fam- ily, which I call the Hypertragulidce^ with the following characters : Selenodont Buminantia with an interrupted dental series, coossified ulna and radius, cuboid and navicular bones, and third and fourth me- tapodial bones. Only two continuous metapodial bones, their distal articular extremities not presenting a complete trochlear keel. Ko fibula. Premolars except the fourth, cutting. This family connects the TragiiUdcc with more typical Buminantia. It differs from that fiimily in the absence of the fibula and the external metapodial bones. From the typical Buminantia or Pecora, it differs No. 1.1 COPE ON MIOCENE FAUNAE OF OREGON. 67 in the incompleteness of the trochlear keel of the metapodials, and the tn^nchant character of the premolars, excepting- the last. I'he species of the preceding list which I have observed in other locali- ties are the following, which 1 procnred in the White Eiver beds of Eastern Colorado: Falwolagus haydeniy Canis hartHhormanus, Canis Uppincottianus^ Canis gregariuSj Leptomeryx evansi, Ilypertragiilus cal- caratKS. Professor Leidy lias recognized a immber of species as those previonyly found in the White Elver beds of Dakota by Dr. Ilayden. Loto}^ Forlc Fauna. Two new species were obtained by Mr. Sternberg at this horizon, which present characters of considerable interest. They are as follows : LuTRiCTis? LYCoroTAMicus, sp. nov. This Carnivore is represented by a left mandibular ramus, which con- tains alveoli and crowns of the canine and molars, excepting those pos- terior to the sectorial. These teeth have the formula, four i)remolars, of Mustela and of the Dogs, but the sectorial is much more like that of Lutra than that of either of the genera named. The heel of this tooth is long, and encloses a wide space transversely, while the sectorial portion is short and low, and includes a large internal tubercle. In the absence of the tubercular teeth, the generic reference is uncertain; but its char- acters agreeing, so far as they go, with the genus Lutrictis of Pomel, I refer it there provisionally. The first premolar only is one-rooted; the third is wide behind, develop- ing a low heel. The heel of the fourth is a little better developed, and there is a small anterior basal cutting lobe ; there is also a tubercle on the posterior cutting edge at the middle. The three cusps of the anterior part of the sectorial tooth are situated at the corners of an imaginary equilateral triangle. The heel continues the width of the crown, is wider than long, and is abruptly truncate behind. It supports a long cutting edge just within the external border, and a shorter one on the internal. The surface of the enamel is smooth. There are two mental foramina, one below the interval between the first and second premolars, the other beneath the anterior root of the third premolar. Measurements. M. Length of molar series without tubercular 0.0220 ^. ^ „ X • 1 ^ antcro-posterior 0.0066 Diameter of sectorial < ^ ^ ^^,r^ I transverse 0.0040 Length of heel of sectorial 0.0024 Length of fourth premolar 0.0045 Elevation of fourth in-emolar 0.0036 From the Loup Fork formation of Cottonwood Creek, Oregon; dis- covered by Charles H. Sternberg. Protolabis transmontanus, sp. nov. A nearly complete cranium, without lower jaw, of an adult animal, is 68 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [VolY. the basis of our knowledge of this species. It presents the characters of the genus in the following superior dental formula: I. C. 1 j P.m. 4 5 M. 3. The first premolar is situated in the middle of a long diastema, and a short one separates the canine from the third incisor. On comparison of this species with the P. JieterodontiiSj the tyi^e, and heretofore the only known species of the genus, various characteristic peculiarities may be observed, which will be noticed in the course of the description. It is considerably smaller than the P. heterodontuSj resem- bling in its dimensions the Procamelus occidentalis. The crown of the second superior incisor is directed forwards, and the cutting edge is oblique to the long axis of the tooth. The first incisor is equally large, and its alveolus occupies the apex of the premaxillary bone. In P. Jieterodo^itus^ the alveolus is smaller, and the apex extends considerably beyond it. The third incisor has a conic crown, with sub- round section. In P. heterodontus it is more robust, and is oval in sec- tion, with weak posteriDr cutting edge. The canine is less robust than the third incisor, and is about as far posterior to it as the latter is from the second incisor. The crown is slightly compressed, and is less robust than that of P. heterodontus. The first premolar is still weaker, and the crown is compressed ; the roots are only discrete at their extremities. It is situated a little more than one-third the distance between the canine ancj. second premolar behind the former. The second premolar is well developed, and is two-rooted. The third premolar is also large, with the grinding surface of the crown about half as wide as long. It has a strong internal basal cingulum, which on attrition encloses a groove-like fossa with the principal crown. The external face of the crown is gently convex between an anterior and a x)osterior ridge. The internal face of the crown is uniformly convex. The fourth x)remolar has both crescents well developed. Its grinding face is subsemicircular, and there are a strong anterior and a weak x)osterior external vertical ridge. In P. hete- rodontuSj the grinding surface of this tooth is more nearly subquadrate. The true molars are subquadrate in horizontal section, and have short croons, well distinguished from the roots. The anterior horn of each ex- ternal crescent is prolonged, constituting a section of a iirominent verti- cal external ridge of the crown at each point. The external sides of the columns are but slightly convex. The inner sides of the internal col- umns are strongly convex. The enamel borders of the lakes are abso- lutely simple, and there are no included enamel fossae. The posterior outer angle of the last superior molar is not produced. As compared with the true molars of P. transmontanus, those of the P. heterodontus are relatively smaller in transverse diameter. The mas- ticating surfaces of the crowns of the second and third are thus more elongate in outline. They are also rather inore prismatic, and the last two apparently occupied longer time in the process of protrusion. They are much larger than those of P. transmontanus. The foramen infraorhitale issues above the middle of the fourth pre- 1 ] COPE ON MIOCENE FAUNiE OF OREGON. C9 iiu)lar. In P. heterodontiis it issues above the anterior part of the first true molar. A sharx) angle separates the exterior and extero-inferior faces of the malar bone. Measurements. M. Length of dontal series fiom "base of Ih'st iucisor 0.257 Leugtli of iucisors on chord 0.028 Space between third incisor and canine - 0.006 Lengtli of crown of third incisor 0.013 Autero-posterior diameter of third iucisor 0.008 Length of interval between canine and first premolar 0.011 Length of interval between first premolar and second premolar 0.020 Length of three contiguous premolars 0.035 Length of third premolar 0.014 Width of third premolar (greatest) 0.0075 Length of true molar series 0.057 antero-i)osterior 0.018 transverse 0.018 Diameter of second true molar T.. , . 1 • 1 . 1 ^ antero-posterior 0.022 Diameter of third true molar < , ■ r, r,-,r, I transverse 0.019 Discovered by G. H. Sternberg in the Loup Fork beds of Cottonwood Creek, Oregon. ON THE Extinct Cats of America. BY E. D. Cope. From the American Naturalist, for December, 1880. ] THE AMERICAN NATURALIST. Vol. XIV. — DECEMBER, 1880. — No. 12. ON THE EXTINCT CATS OF AMERICA. BY E. D. COPE. IN following the general series of the Carnivora, we pass, as in other orders, from the generalized to the specialized, tj'pes. That we should begin with the /V^?rr than in Himantopus and Becurvirostra. The former is just half as long as the tibia, and seen in profile is almost straight. The crest of the tibia is very prominent, but is not produced proximally. The distal extremity of the tibia and proximal part of the tarsometatarse are so damaged as not to furnish satisfactory characters. There are five rectrices visible in the specimen. Those which are in all probability median are the longest, while the external tw:o are of equal length. This gives the outline a rather short wedge shape as the feathers lie closed. The expanded tail would be rounded with a shghfc median angulation. The extremities of the feathers are rounded, and 1^.8.1 COPE ON A WADING BIRD OF THE AMYZON SHALES. 85 lieir whole stmctvire is soft and dolicate. The length of the longest •eotrix is just about that of tlie tibia. Measurements. M. jength of the preiliac vortobrac . 010 jength of centrum of first V(irtebra) 0035 jengtli of sacrnm 021 jengtli of caudal vertebrfe on curve . 0145 jength of plowshare bone to apex 005 jength of ilium 024 jength of ilium to acetabulum ^ 012 jength of ischium from acetabulum 016 jength of pubis from acetabulum 019 Vidth between posterior angles of ilia 009 jength of femur 024 diameter of femur at middle 003 jength of tibia 047 Intero-posterior diameter at head 006 )iameter of shaft at middle 0027 Mameter of head of tarsometatarse 004 jength of median rectrix from plowshare bone .046 jength of external rectrix from plowshare bone 040 ^idth of i)ortion of tail preserved . 020 The strongly contrasted light and dark shades of color are not unfre- luently preserved in the insects of this formation. I suspect that had ihe rectrices of this species originally displayed the alternating white ind dark cross-bars characteristic of the Totani, some trace of them, ^ould be discoverable in the fossil, in spite of the fact that the entire feather is represented by carbon only. The brown tint of the specimen, )oth light and dark, is uninterrupted by pattern of any kind. Th^ tail is rather longer than in the Tringw, about equal to that of aany plovers and Totanij and shorter than that of Actiturus. The Gharadius sheppardianus was discovered near Florissant, Color- bdo, by Dr. G. Hambach, a skillful naturalist. I have named it in honor )f Edwin Sheppard, of Philadelphia, an excellent ornithologist and ikillful artist. \ ft I Art. TII.— On the IVimraYidaD and Canidse of the JUiocene Period. By E. ». Cope. In following the general series of the Carnivora, we pass, as in other orders, from the generalized to the specialized types. That we should begin with the Procyonidce and their allies is indicated by all the char- acters to be especially considered in the case. They have five toes on all the feet, and are plantigrade, resembling in these points all primitive Mammalia.* They have the original number of molar teeth, seven on each side, and of these none are distinctly developed sectorials. The condyloid and carotid foramina are distinct, and there is a postglenoid foramen. If, starting from this point of departure, we arrange the suc- ceeding families of Carnivora according to their resemblances and differ- ences in these respects, we have a tolerably consecutive series of divisions. Passing at present over the families Mustelidce, Viverridce^ Cryptoproc- tidce, and others with five toes on all the feet, we reach those in which the hind foot has lost a digit, leaving the number 5-4. These are the Pro- telidWy Canidw, and Felidce, We can take but one step further in this order ; that is to those species where the anterior foot has also lost a toe, which xjonstitute the family Hycenidce. The toes are therefore, here, 4-4. For the well-marked characters of the three families mentioned just before, I refer to another page, and i^roceed to define briefly the division which has been heretofore termed the Felidce. In doing so I am compelled to omit several of the characters generally employed to define that family, since I have found them to be wanting from various extinct genera. The only comprehensive definition which I can give is the following : Digits 5-4; sectorial teeth well developed in both jaws ; not more than one true molar tooth in the upper nor more than two true molar teeth in the lower jaw. Glenoid cavity grasping mandibular condyle anteriorly as well as posteriorly. Professor Gill, who has devoted much attention to the definition of the families of the Mammalia^ t gives the following skeletal characters * See Homologies and Origin of Types of Molar Teeth, of Mammalia Eclucabilia, Journal Academy Phila. 1874, March. t Arrangement of the Families of Mammals,-Smith8on. Misccll. Coll. 230, 1872, p. 5G. 1G5 166 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [Yol.Yl, in his diagnosis of the Felidce, and of the three comprehensive divisions within which he places it : " I. Skull with the paroccipital process ap- plied closely to the auditory bulla 5 the mastoid process small or obso- lete 5 external auditory meatus very short or imperfect. Div. A. Carotid canal minute and superficial or obsolete 5 condyloid foramen and foramen lacerum posticum debouching into a common fossa ; glenoid foramen minute or null. Os penis rudimentary. Subdiv. 1. Otic bulla divided by a septum into posterior and anterior chambers, communicatiug by a narrow aperture (Flower). Subdiv. a. Skull with no alisphenoid canal.'' All of the parts above mentioned I have found to be important in the defi- nition of the natural divisions of the Carnivora^ excepting those derived from the paroccipital and mastoid processes. But their condition in the extinct Carnivora which have been hitherto arranged with the Felidce^ and which resemble them very much in superficial characters, does not coincide with Professor GilFs definition. Thus, in the various American genera which are allied to Drepanodon^ the carotid canal is distinct from tliQ foramen lacerum posterms, and the condyloid foramen is also sepa- rated from it by quite a space. These are characters which belong to most of the Carnivora with five digits on all the feet. Further, the postglenoid and postparietal foramina are present ; also characters of the lowest Carnivora, as the bears and certain extinct dogs. Then, there is an alisphenoid canal, which is also found in bears, dogs, and the cat-like Crypto;procta. I cannot demonstrate that the otic bulla is divided, as the above diagnosis requires, in any of the fossil species. I have verified these characters on species of the following genera, of which I have well-preserved skulls: ArchceluruSj JSfimravus, DinictiSy Pogonodon (except those of the basal axis of the skull), and Hoplophoneus. Three genera as yet only found in Europe are similar in general char- acters, and probably agree with them. I allude to Procelurus Filh., JElurogale Filh., and Eusmilus Gerv. On the other hand, the genus^ Smilodon, which includes the American saber- tooths of Pliocene age^ agrees with the true cats in the points in question ; i. e., the alisphenoid* postglenoid, and postparietal foramina are wanting, the carotid foramen is^ either internal or wanting, and the condylar enters the jugular foramen at its mouth. This surprising condition of affairs makes it important to learn the characters to be found in the species of the longest-known genus^ Drepanodon, of the European beds. But although there are several good crania in European museums, T can find no description of their minute characters, and no mention made of their foramina. The proba- bilities are, on various grounds, that this genus agrees with JSmilodon in the latter characters. The reasons in favor of this supposition are the agreement in special dental characters, and the Pliocene age of the typi- cal species, I), cultridens. Whether the middle Miocene species of San- san and Epplesheim agree with this one in structure, is of course un- certain. Seven, and perhaps eight, genera, then, constitute a group to be dis- No.7.\ NIMRAVIDiE AND CANIDiE OF THE MIOCENE PERIOD. 167 tiiiguislied from the true Felidce, and, as it appears to me, as a distinct family. Should we ignore the characters adduced in this instance, we abandon at the same time the definitions of several of the other families of the order, and in fact throw the system into confusion. I have pro- posed to call this family the Nimravidce, and have contrasted it with the Felidcc in the following definition. Both are included in the division already defined on a preceding page. No distinct carotid foramen nor alisphenoid canal j condylar foramen entering the foramen lacerum posterius. No postparietal and gen- erally no postglenoid foramina Felidce, Carotid and condylar foramina entirely distinct from the foramen lacerum posturius^ an alisphenoid canal, and postglenoid and post- parietal foramina Nimravidce, NIMRAVID^. The dental characters of the Nimravidce are in general, those of the Felidce, the higher genera having the same dental formula. Descending the scale, the number of molar teeth increases at both ends of the series in the lower jaw, and anteriorly only in the upper, but the number of the true molars never exceeds J. The following table gives the defini- tions of the genera. I am unfortunately ignorant of the characters of the foramina in Procelurus and Pseudcelurus, as well as in Mlurogale and Eusmilus, I. Lateral and anterior faces of mandible continuous ; no inferior flange. a. Inferior sectorial with a lieel ; canines smooth. Molars f inferior sectorial with interior tubercle Procelurus. Molars f \', inferior sectorial without interior tubercle Fseudcelurus. II. Lateral and anterior faces of mandible separated by a vertical angle ; no inferior flange ; incisors obspatulate. a. No anterior basal lobe of superior sectorial ; inferior sectorial with a heel (and no internal tubercle) ; incisors truncate. Molars f ^ ; canine smooth Archcelurus. Molars | ^; canine denticulate ^lurogale. Molars | i ; canine denticulate Nimravus. III. Lateral and anterior faces of mandible separated by a vertical angle; an inferior flange ; incisors coni9 ; canines denticulate.* a. No anterior basal lobe of superior sectorial ; t inferior sectorial with a heel ; no posterior lobes of the crowns of the premolars. Molars Dinictis. Molars f i Pogonodon. Molars ^-^^t Hoplophoneus. Molars [| Eusmilus. It is readily perceived that the genera above enumerated form an unusually simple series, representing stages in the following modifica- tions of parts : (1) In the reduced number of molar teeth ; (2) in the * Gervais's figures of the canines of Eusmilus Udentatus represent no denticulations, but the figure is not clear. tRudimental in Hoplophoneus. 168 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [Vol. YI. enlarged size of the superior canine teeth ; (3) in the diminished size of the inferior canine teeth ; (4) in the conic form of the crowns of the in- cisors ; (5) in the addition of a cutting lobe to the anterior base of the superior sectorial tooth ; (6) in the obliteration of the inner tubercle of the lower sectorial, and (7) in the extinction of the heel of the same; (8) in the development of an inferior flange and latero-anterior angle of the front of the ramus of the lower jaw; (9) in the development of cut- ting lobes on the posterior borders of the larger premolar teeth. (1) The reduction in the number of molar teeth. The dental formula otFrocelurusis that of some Yiverridcesmd Canidce, and the reduction from this point to the end of the series is obvious. In Eusmilus^ as in Smilodon, the number of molars is less by one in the inferior series than in Lynx and JSTeofelis, where the formula is the smallest known among Felidce proper, viz : f \. (2) The enlarged size of the superior canine teeth. In Procelurus and Pseudcelurus the canines of both jaws are developed, as in recent Felidce. In Archceliirus the superior is the larger, but does not, relatively to the molars, exceed that of Felis. It is rather compressed in form and has a sharp cutting edge posteriorly. In Nimravus the superior canine begins to have the enlarged size of the saber-tooths, but its form is peculiar in the N. gomjplioduSj being spike- shaped rather than saber-shaped. We find the true saber shape first in Dinictis^ where it is compressed, and with a denticulate cutting edge on both front and rear. In Fogonodon it has reached a very large size, and it does not display much increase in this respect until we reach the last genus of the series, Fusmilus, where its proportions are enormous, almost as large as in the feline genus Smilodon^ where they appear to have been an inconvenience to the animal. (3) The diminished size of the inferior canines becomes evident in the lower genera of the third division (supra) of the N'imravidce, but is most decided in the highest genera, Hoploplioneus and Eusmilus. (4) The incisor teeth have the usual obspatulate or obovate outline in the genera of the first and second di- visions of the family, including Nimravus. They are conic in the true saber-tooths with flared \ow&y jaw, beginning with Binictis and ending with Eusmilus. (5, 6, and 7) The structure of tbe sectorials. The pres- ence of a heel and an inner tubercle of the lower sectorial are well-known characters of a majority of the Carnivora. In only the most highly or- ganized genera are they wanting, and among them are included all those of the Felidce that still exist. In the Nimravidce the inferior genera have both in a reduced degree, and they soon disappear as we ascend the scale. Thus, the inner tubercle is only i)resent in the species of Frocelurus^ Binictis^ and Roploplioneus. The heel, on the other hand, re- mains throughout the entire family. The anterior basal lobe of the superior sectorial has the same history, its absence being characteristic of the inferior Carnivores and of all the genera of Nimravidce except MoplophoneuSj where it is rudimental. It is Tvell developed in Brepano- don as in recent Felida^^ and is sometimes double in Smilodon. (8) The |»«.7.] NIMRAVID^ AND CANID^<: OF THE MIOCENE PERIOD. 1G9 levelopmont of the inferior flange and latero-antcrior angle of the man- libuhir ramns. There is a successive advance in the development of :hese characters, beginning with tlie second group, for in the first they ire wanting. The latero-anterior angle is developed in Archcelurus and illied genera, and is merely continued on the inferior border of the ^amus. In the third group it is much more acute, and is deflected down- ivards, forming the well-known flange of the saber-tooths. It is longest n the Eiismilus hidentatus Filh. (9) The highest genera of Nimravidce, \ g. RoplophoneuSj difter from the true Felidcc in the absence of the cut- ing lobes on the posterior edges of the crowns of the larger premolar :eeth. But, according to Filhol, these lobes are present in the generalized genera Procelurus and Pseudmlurus^ which are thus brought into a re- I lation with the Felidce not possessed by other Nimravidw. A characteristic perfection of the Felidce is seen in the genus Smilo- ion ; that is, the vertical direction of the ungual phalanges, by which the slaws become retractile. This is well disi)layed by the two si)lendid specimens of Smilodon necator from Buenos Ayres, which have been preserved,* Unfortunately, these phalanges have not yet been de. scribed in any species of the Nimravidw^ and it is not yet certain what their structure really was. Among the true Felidce the genus Cyncelurus displays a less degree of development in this respect than the other genera, the ungual phalanges lacking the proximal process below the articular facet. Such a condition is to be looked for among the less per- fect genera of Nimravidce. The succesion of genera above pointed out coincides with the order 3f geologic time very nearly. Those belonging to groups first and sec- 3ud belong to the Lower and Middle Miocene, except uFJlurogale, which is perhaps Upper Eocene, and Pseudcelurus, which is Middle Miocene. The genera of the first group of division third have the same Lower Miocene age, except Fusmilus, which has been found in the same forma- tation (Phosphorites) as the JElurogale. Drepanodon is Upper Miocene, and Smilodon is Pliocene. The relations of these genera are very close, as they differ in many Bases by the addition or subtraction of a single tooth from each dental series. These characters are not even always constant in the same species, so that the evidence of descent, so far as the genera are con- cerned, is conclusive. No fuller genealogical series exists than that which I have discovered among the extinct cats. As to the phylogeny of this family, there are flesh-eaters of the Eocene period which may well have been the ancestors of both the Nimravidce and Felidce.^ 1 have suggested that this i)osition is most appropriately held by the Oxycenidcej a family of several genera, which included the most formidable, rapacious mammals of that early period in both continents. * See American Naturalist, December, 1880, fig. 12. + See On the Genera of the Creodonta, by E. D. Cope, Proceed. Amer. Philos. Soc. July, 1880. 170 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [Vol.Yh The interval between them and the Nimravidce is, however, great, for in the Oxycenidw, when there is a superior sectorial tooth, the first true molar in the upper jaw is utilized instead of the last premolar, and the secondi true molar below is a sectorial as well as the first. Several intervening' forms must yet be found to complete the connection, if it have ever ex- isted. It is, however, very likely that the true Felidce were derived; from the genus Prowlurus, through Fseudcelurus, if indeed these two genera be not the primitive members of that family, for, as above re- marked, the evidence of their possession of the characters of the JVtm- ravidce has not yet been obtained. There can be no reasonable doubt that the genera Brepanodon and Smilodon in the Felidce are the descend- ants of Hoplophoneus and allied genera. In fact, the Nimravidce and JieZii^CB are " homologous groups having corresponding terms in the manner I foreshadowed as a general principle in 1868 (Origin of Genera). In looking for causes in explanation of the modifications of structure cited, one can easily discover that there is a close relation between the arrangement of the teeth and the mechanical laws involved in the per- formance of their function, that of seizing an active prey and of cut- ting up their carcasses into pieces suitable for swallowing. It is obvious that in the latter case the flesh-teeth bear the resistance and the masseter muscle is the power, and that the nearer these parts are together the better is the function performed. As a matter of fact, the sectorial teeth in modern Garnivora are placed exactly at the angle of the mouth, which is nearly the front border of the masseter muscle. Both the muscle and the teeth have, however, moved forwards in con- nection with the shortening of the jaw behind. This has been due to the necessity of bringing the power (masseter) nearer to another point of resistance, viz, the canine teeth. In the early carnivores (aa Hycenodontidce) the long jaws supported more numerous teeth than in any modern families, and the fissure of the mouth was probably very wide. The canine teeth were evidently very ineffective weapons. The animals probably only snapped with their jaws, and did not attempt to lacerate or hold on, as do the cats. The dogs of to-day are long-jawed, and they snap in a manner quite distinct from anything seen among the cats. The only dogs that hold on are the short-jawed bulldogs. So in the use of the canines, we have the ground of the shortening of the jaw behind and before, and the consequent change of structure, which resulted in the modern perfected Felidce. ^0.71 NIMRAVIDiE AND CANID^ OF THE MIOCENE PERIOD. 171 The following list shows the number and distribution of the species )f the Nimravidcc. The position of a cross on a line indicates an inter- uediate geological position. Proselurus iulieni Filh Proselunis lemanensis Filh Pseudajlurus hysenoides Blv PaeudiBlurua edwardsi Filh Pseudajlurus intrepidus Leidy — Paeudailuius sivalensis Lydd Archailurus debilis Cope ^lurogale intermedia Filh ,Elurogale acutata Filh Nimiavus goraphodus Cope N^imravus confortus Cope Dinictis felina Leidy Dinictis cyclops Cope Dinictis squalidena Cope Pogonodou platycopia Cope Pogonodon brachyopa Cope Hoplophoneus oreodontis Cope . . . Hoplophoneus primaivus Leidy. . . Hoplophoneus occidentalis Leidy. Hoplophoneus cerebralis Cope Eusmilua bidentatus Filti Upper Eocene. Lower Miocene, Upper Miocene. Pliocene. DESCRIPTIONS OF NEW SPECIES. KlMRAVUS GOMPHODUS Sp. nOV. Nimravus hrachyops Cope, Proceed. Academy Philad. 1879, p. 170, not Machcerodm brachyopa Cope, Proc. Am. Phil. Soc. 1878, p. 72. This carnivore is represented by parts of three individuals ] one of them by a nearly complete skull. The species is rather larger than, the average Uncia concolor. Measurements of sJculL M. Axial length from occipital condyles to premaxillary border . 20d Axial length from inion to premaxillary border 220 Axial length from premaxillary border to canine tooth 017 Axial length from premaxillary border to anterior border of superior sectorial.- . 066 Axial length from premaxillary border to posterior extremity of maxillary bone. . 097 Axial length from premaxillary border to postglenoid process 162^ Length of nasal bone from nasal notch 065 Length of sagittal crest from inion 082^ Width of premaxillary bone (greatest) 019 Width of each nasal bone at middle 008 Width of each frontal bone at middle of orbit 028 Width of each frontal bone at postfrontal angles 035 Width of skull at anterior part of zygoma 098 Width of zygomata at temporal fossa Ill Width of skull at meatus auditorius .* 074 Width of skull between apices of paroccipital processes 054 Width of occiput at middle .044 Width of foramen magnum 024 172 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [Vol. Elevation of occiput above forarnen 036 Width of cliin at base 022 Width of chin at summit 028 Depth of chin .040 Depth of ramus at diastema 027 Depth of ramus below last premolar 031 Length of ramus 157 Elevation of condyle 033 Elevation of coronoid process 071 Length of superior canine 045 ^ „ , ^ . . ( antero-posterior ^ 016 Diameters oi base of superior canine < ( transverse 008 Anteroposterior diameter at middle 010 Distance from canine to third premolar 016 Length of molars, including third premolar 045 Length of base of third premolar 018 Elevation of cusp of third premolar 013 Length of base of sectorial 025 Elevation of cusp of sectorial 015 Width of tubercular 009 Elevation of inferior canine 024 Antero-posterior diameter of inferior canine at base 012 Length of inferior diastema 022 Length of inferior molar series 063 Length of third premolar 0175 . Elevation of third premolar 0175 Length of fourth premolar .020 Elevation of fourth premolar 015 Length of sectorial 025 Elevation of median cusp of sectorial 016 . The characters of this species will be fully detailed in my final reporl to Dr. Hayden, now passing through the press. From the John Day Eiver, Oregon. IS'IMRAVUS CONFERTUS Sp. nOV. This species is as yet represented by a mandibular ramus only. It ifl one-third smaller than that of J^. gom2)hodus. The inferior border of the ramus is broken off, excepting for a space below the diastema. The general form is narrow, as in N, gomphoduSj and there is a projecting ledge along the inner base of the sectorial similar to that seen in the latter species. The angle separating the side from the front of the ramus is rather stronger than in JV. gompJiodus, but there is no indication of an inferior flare. The diastema is shorter than in the typical species, its length equalling that of the base of the third (first) premolar; in N. gomijJiodus it is half as long again. The symphysis is correspondingly shorter, ceasing a little in advance of, and at the pos- terior border of, the inferior canine tooth, while in N. gomphodus it con- tinues for one diameter of the canine behind its posterior border. The crown of the inferior canine tooth is directed backwards, and its iiro.7.J NIMRAVIDiE AND CANIDiE OF THE MIOCENE PERIOD. 173 ll errato cutting edge is preseuted almost entirely inwards. The interno- nterior face of the crown is flat, and has a low shoulder at the base. The iiolars have the proportions of those of N. gomphoduSj differing only in heir smaller size, which is very apparent, as can be S€en by the measure- iients. The first (third) premolar is a little longer on the base than high^ las no anterior tubercle, and has a short cutting basal heel. The fourth )remolar has subequal anterior and posterior basal cutting lobes, and the )ase is longer than the elevation of the median cusp. The sectorial tooth las a short cutting heel, but no trace of inner tubercle. The anterior lobe s as long as the median, but not so high. It overlaps the fourth premo- ar as far as the base of the median cusp, incisor teeth are pre- lerved in the specimen. Tubercular small. Measurements of sTcull, Depth of ramus at diastema 020 Depth of chin 027 ilevation of inferior canine 016 Diameter of inferior canine at base 010 Liength of inferior diastema 014 licngth of inferior molar series 053 Lengtli of third premolar 014 Slevation of third premolar 010 ■jength of fourth premolar 016 Elevation of fourth premolar 013 Length of sectorial 022 Elevation of median cusp of sectorial 015 One specimen, from the John Day Yalley, Oregon, found in the Truckee 'ormation, by J. E. Wortman. DOLOREODON RYDERANUS Sp. nOV. Eepresented by a nearly complete skull, without lower jaw. These ndicate the third and smallest species of the genus. The specimen be- onged to an adult animal, as indicated by the condition of the last two nolar teeth. Besides the small size, two characters may be cited as distinguishing Ms species from those already known. First, the temporal ridges con- verge very gradually, so that the sagittal crest does not appear anterior to the line of the otic buUse, posterior to which point the skull is broken ibove. Second, the face is constricted immediately posterior to the po- jition of the fundus of the alveolus of the canine teeth. The position of this alveolus is prominent, and occupies the superior half of the maxil- 'ary bone, which is excavated beneath it. This excavation is bounded behind by the infraorbital foramen. The lachrymal bone presents an ingle into the orbit. The latter is open posteriorly, but the opposing processes approach each other. The zygomata are slender, v. The enamel of the molars is slightly wrinkled. 174 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [Toivi. Measurements of sTcull, M. Total length from left side of inion to front of canine 147 From same to end of maxillary bone 075 From same to palatal notch , 087 Length of diastema Oil Length of molar series 059 Length of true molars , 035 ^. ^ ( antero-posterior 013 Diameters of second true molar < ( transverse 015 ^. ^ , . -. i antero-posterior 015 Diameters of third true molar < ( transverse 019 The linear measurements of this skull are three-fourths that of C. feroxj and three-fifths that of G. macrocephalus. Found by J. L. Wortman, on the John Day Eiver, Oregon. Dedicated to my friend John A. Eyder, of Philadelphia. Pal^ochcerus platyops sp. nov. Established on a nearly complete skull, which lacks the muzzle ante- rior to the third premolar teeth, the lower jaw, and parts of the zygomata. The last superior molar is not protruded, and it is probable that the fourth deciduous premolar still remains in the jaw. The size of this species exceeds that of any other member of the genus- The postorbital process is also more elongate, so as to inclose the orbit to a greater extent than is usual in Palceochcerus. The temporal ridges are strong, and rise into a convexity above the anterior part of the tem- poral fossa. From this point they converge gradually to form the sagit- tal crest, which has a truncate edge. Between the ridges the surface is concave, the basin widening forwards as far as a line passing through the posterior third of the orbits. The frontal bone rises steeply from the orbit to the edge of this basin, and is regularly convex in front of it. The profile descends forwards, so that the section at the infraorbital, foramen is broadly convex, and not compressed, as in the species of Palceochcerus from the Truckee beds of Oregon. The emargination of the palate extends as far forwards as the line of the anterior border of the second true molar. The posterior border of the infraorbital foramen is above the middle of the anterior root of the fourth premolar. The post- glenoid, mastoid, and postparietal foramina are present ; the last named rather small and in the inferior part of the parietal bone. The base of the fourth premolar is remarkably extended antero-pos- teriorly. Its crown has a posterior basal cingulum, and is in contact with those of the molars anterior and posterior to it. The crown of the first true molar is narrowed inwards, the anterior border being more oblique than the posterior. Both of these borders have a wide cingu- lum. There are two large external cusps, with antero-posterior continu- ous edges, two small median cusps, and a larger internal cusp. No internal or external cingula. The second true molar is also narrowed inwards, but less so than the first, and the posterior border is the only O.7.] ^IMRAVID^ AND CANIDvE OF THE MIOCENE PERIOD. 175 blique one. There are wide posterior and anterior cingula, and five usps arranged as in the first molar. The internal cusp is relatively irger in this tooth. Measurements of skull. M. iength from occipital condyles to anterior border of last premolar 186 iength from occipital condyles to palatal edge 135 iengtli from occipital condj^les to line of anterior border of glenoid fossa 051 Vidth of occipnt above 046 Vidth at posterior origin of zygomata 071 Vidth between orbits (least) 077 Vidth at middles of first molar teeth 098 Vidth between first molar teeth 040 jlevation of occiput, with condyles 082 Elevation of muzzle at front of fourth premolar 043 Hameters first molar | antero-posterior 021 ( transverse 016 ( antero-posterior 022 Hameters second molar | ^^^^^^^^^^^ This species is easily distinguished from those which have beeu dis- ;overed heretofore, by its large size, by the peculiar form of its molar eeth, and by its flattened muzzle. The specimen above described was liscovered by Gapt. Emmett Crawford, Third Cavalry, U. S. A., on the ipper waters of the Big Cheyenne Eiver, Dakota, in a bed of the White iiver formation. It was presented to me by Dr. William H. Corbusieur, U. S. A., to whom my especial thanks are due. Protolabis prehensilis sp. nov. ; This camel is supposed to have existed on the evidence of portions of ;he mandibles of two individuals. These include the symphyseal por- tion, and one of them the ramus as far posteriorly as the first true nolar, inclusive. These remains indicate a robust species of the size of :he Procamelus angustidens, or between the P. gracilis and P. robustus. The most marked peculiarities of this species are the following : The 3auine and first premolar are very robust, and the latter is one-rooted ind with an oval section at the base like the canine. The second pre- Qiolar is also one-rooted, while the third and fourth are two-rooted, and liffer very little in size from each other. The first true molar is abruptly larger, although narrowed in front. The root of the second premolar is round and of robust proportions, its diameter being one-half linear that of the first. The roots of the incisors are robust, that of the first being rather larger than that of the third. The symphysis terminates a little behind the line of the first premolar. The mental foramen is unusually extended in the antero-posterior direction . Measurements of sJcull. M. Length of symphysis (No. 1) 080 Length of bases of incisors and canine 043 .Diameters of canine^ antero-posterior 016 ( transverse ^ Oil 176 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [Yol.YJ M. Diameters of premolar (i) 5 antero-posterior 01- ( transverse 00! Length of last three premolars on base (No. 2) 03 Length of base of fourth premolar 01; Length of base of first true molar 01! Length of first diastema (No. 2) 01' Length of second diastema (No. 1) 02( Depth at second diastema (No. 1) 03; Depth at first true molar 04f The generic position of this species is uncertain. I place it provis ionally in Protolahis on account of the large development of the infe rior incisors. Its robust canine and first premolar, and small third anc fourth premolar distinguish it from any of the species described. Found by E. H. Hazard, in the Loup Fork beds of Southern braska. EUMYS LOCKINGTONIANUS Sp. nOV. This rodent is represented by a nearly perfect skull, which is withoui lower jaw. Its specific characters separate it widely from the U. elegani Leidy and U, nematodon Cope. It is considerably larger than either and the temporal ridges are very obsolete and do not unite posterioi to the orbits, as in U, elegans, resembling in this respect the U. nemato don. The parietal region is wide and flat above. The interorbital regioi is only moderately contracted. The muzzle is rather short as compared with the total length of the skull. The interorbital region is gentl} convex above, and the top of the muzzle is flat. The zygoma is quitt slender, and the otic bullae are large and prominent. The notch of th( palate extends as far forwards as the posterior part of the last superioi molar. The infraorbital foramen is very large and round. The anterior face of the superior incisor is nearly plane, and it h marked by a weak groove near the inner and a strong groove near the external border. In elegans this face is convex and without grooves The molars are rather small for the size of the skull ; their crowns ar( worn by use. The third is subround in section, and its diameter is about half that of the first j the latter has the anterior odd lobe quite small. . Measurements. Total length of skull OSSC Length (axial) to front of orbits 014( Length (axial) to palatal notch 019( Length (axial) to first molar 012( Width at otic bullae 020r Width at middle of zygomata 022( Width of interorbital space 006( Width between first molars OOSf Length of molar series 007( Length of first molar 003^ Width of superior incisor 001' This species is dedicated to my friend W. I^". Lockington, the well known naturalist of San Francisco. -I NIMRAVIDiE AND CANIDiE OF THE MIOCENE PERIOD. 177 UlUUUS BALLOVIANUS Sp. IIOV. This squirrel is tlio second species of its geuus supposed to occur in jie Truckee beds of Oregon, and the third Sciurus obtained thus far from 16 Lower Miocene or Oligocene of the West. The typical specimen for- Imately includes the cranium, with both rami of the mandible, so that |s reference to the genus Sciurus rather than to Gymnoptychus is assured, ike the latter genus, the infraorbital foramen is reduced to a slit, but, alike it, there is but one internal tubercle of the crowns of the superior lolars instead, of two. The skull is flat above, and the interorbital space is also flat, and is imarkably wide. Temporal ridges none. Muzzle short and narrow, alate wide, its posterior notch extending as far forwards as the last iperior molar. The ascending ramus of the mandible originates oppo- te the anterior part of the last inferior molar. The massateric fossa itends to opposite the anterior border of the second inferior molar. The lental foramen is near the superior border of the posterior part of the liastema. The second and third superior molars, the only ones pre- irved, have two cross-crests and a strong anterior cingulum. The sternal extremities of the cross-crests are little elevated, and there are 0 other cingula. The inferior molars have basin-shaped grinding faces, ith a lobe at each angle. There is a small tubercle between the lobes f the inner and outer pairs. The incisors of both jaws are much com- ressed, strongly convex in front, and, in the lower jaw at least, without jiulpture. Measurements, (M. engtli of skull to orbit 0090 ^idth between orbits 0090 ^idth of muzzle 0047 '^idtb between last molars 0040 engtb of superior dental series 0054 iameters of second molar ^ antero-posterior 0015 c transverse 0016 'idth of superior incisor 0013 eDgth of mandibular ramus 0150 levation of ramus at coronoid , 0080 engtli of diastema 0030 engtli of inferior dental series 0070 epth of ramus at second molar 0045 This species is much smaller than the Sciurus vortmani, from the same orizon of Oregon. The type specimen was discovered by Mr. L. S. >avis, of Mr. Wortman's party. The name is given in honor of Mr. r. H. Ballon, of Chicago, a naturalist and journalist. CANID^. Species of this family were very abundant during the Miocene period 1 North America as in Europe. Those of the Lower and Middle [iocene epochs belong to genera allied to, but distinct from, Cams; bile those of the Upper Miocene (Loup Fork) and later horizons, per- 12 GB 178 BULLETIN UNITED STATES GEOLOGICAL SURVEY. {VolYJ tain to the latter genus, with few exceptions. The characters of thi genera are as follows : I. Molar formula i f . Humerus with epitrochlear foramen Amphicyoh II. Molar formula | f . Humerus with epitrochlear foramen. Inferior sectorial heel trenchant Temnocyoh Inferior sectorial heel basin-shaped Galecynm Humerus without epitrochlear foramen. laferior sectorial heel basin-shaped Cani III. Molar formula f si . Heel of inferior sectorial trenchant Enhydrocyoi IV. Molar formula 1 1. Heel of inferior molar basin-shaped Icticyo) V. Molar formula | \. First inferior molar two-rooted Hycenocyoi To these genera I refer nineteen species of the American Miocenes. Amphicyon Lartet. Bulletin Soci^td Gdologique de la France, 1836, vii, 217-220; BlainvilL Comptes-Rendus, 1837, v, 434 ; L'Institut, 1837, v, 18-19 ; Blainville, Oste* graphic, ix, Subursus, 78-96. Dental formula : 1. f ; 0. i j Pm. f ; M. |. The true molars of the si perior series all tubercular j the last two of the inferior series also tube cular. First inferior true molar a sectorial, with an internal tuberc) and a heel with a superior groove, bounded by raised borders. Hi merus with an epitrochlear arterial foramen. Much is yet to be desired in the elucidation of the characters of th genus, especially of the American forms, which are less abundant an of smaller size than those of Europe. The typical species, Ampliicyc major Blv. was the largest, equalling a bear in size. It is derived froi the Miocene of Sansan, and a smaller form of it is found, accordin to Pomel, at San Gerand-le-Puy. Other species are derived from tl latter locality, and all are typical of the Miocene formation in Europ In the " Mio-pliocene" of India a single species has been discovered, tl A. palceindicus of Lydekkier. Three species occur in the Lower and Mi- die Miocene of North America, the largest of which about equals tl wolf in size. On account of the large development of the inferior tube cular teeth, I have suspected that the Canis ursinus Cope, from tl Loup Fork group of Kew Mexico, would prove to be an Amphicyon, ■ so, it is the only representative of this genus in our Upper Miocene. The three American species differ as follows : The A, cuspigerus small, not exceeding the kit-fox in dimensions. The A. Jiartshornian is about the size of the coyote, and has rather smaU tubercular molai especially of the lower series. The A, vetus is a little larger, but h.i the tubercular molars disproportionately larger than those of the > liartsliornianus. ^0.7] NIMRAVID^ AND CANID^ OP THE MIOCENE PERIOD. 179 Temnocyon Cope. Paleontological Bulletin, No. 30, p. 6, Decombor 3, 1878 ; Proceedings Ameri- can Philosophical Society, 1878, p. C8. Dental formula : I. j| j 0. | ; Pm. | ; M. f . Two molars in each jaw tubercular. Inferior sectorial with well-developed heel, which is keeled with a cutting edge above. An internal tubercle of the same. A post- glenoid, but no postparietal foramen. Humerus with an epitrochlear arterial foramen. The characters on which I rely at present for the discrimination of this genus from Ganis are two. The first is the presence of a cutting edge on the superior face of the heel of the inferior sectorial, in place of a double row of tubercles surrounding a basin. When well developed, these characters present a broad contrast, but indications of transitional forms are not wanting. Thus, in some extinct Canes the internal crest of the heel is less elevated than the external, which is the homologue of the single crest of Temnocyon, and in some specimens of Temnocyon coryphceus there is a cingulum on the inner side of the median keel, which represents the internal crest of Canis. Secondly, the epitrochlear foramen of the humerus, a character common to all of our Lower Miocene Canidce yet known. The keel of the sectorial, which defines this genus, is simply a repeti- tion on that tooth of the heel which belongs to the posterior premolar teeth of many Carnivora. It finds resemblances in such Eocene forms as Mesonyx and Palwonyctis. Among recent Canidce it is apparently unknown, and is very rare in other groups. The Cynodictis crassiros- ' tris Filhol, from the French Phosphorites, strongly resembles the species of Temnocyon in generic characters. Three species of the genus are known to me. They may be distin- guished as follows. A fourth species, T. josepM, is provisionally placed w^ith these : t. First superior tubercular molar with a wide median fossa, bounded within by a tubercle. Length of superior molar series from canine, .070 ; of true molars, .0215. T. altigenis. Length of molar series from canine, .067 ; of true molars, .014. T. wallovianus sp. nov. [I. First superior tubercular molar with narrower basin, bounded within by a V-shaped crest. Length of dental series from canine, .055 ; of true molars, .014 coryphceus. Length of dental series from canine, .051 ; of true molars, .013 ; muzzle narrow, zygomas wide T. josephi sp. nov. All of the above species have been derived from the Truckee Miocene )eds of Oregon. I, however, anticipate the discovery of these or other ipecies of the genus in the White Eiver beds of Dakota and Colorado. 180 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [Foi.Vl, Galecynus Owen. Quarterly Journal Geological Society London, 1847, iii, 54-60. — " Cynodon Aymard, Annales Soci6td du Puy, 1848, xii, p. 244. — Cynodictis Bravard et Pomel, Notice sur les Ossemens Fossiles de la Debruge, 1850, p. 5. — Cyo- therium Aymard, Ann. Soc. d'Agric. du Puy, 1850, xiv, p. 115"; Bronn. Dental formula: I. f; 0. Pm. f • M. |. Inferior sectorial with internal tubercle, and with a heel with raised or tubercular internal and external borders. First premolar in both jaws one-rooted. A post- glenoid but no postparietal foramen. Humerus with an epitrochlear arterial foramen. This genus, which is abundantly represented by species and Individ- uals, existed during the Upper Eocene epoch in Europe (in the Phos- phorites), and also during the White Eiver or Oligocene in ^forth America. As the structure of the feet of the numerous species from these epochs is not yet known, and, therefore, some doubt as to their correct generic reference may still exist, I only regard the genus as a certain inhabitant of IS'orth America during the Truckee or Middle Miocene epoch. This is indicated by the Galecynus geismarianus, where the number of the toes on the posterior foot has been ascertained. All the species of the genus from Eocene and Lower Miocene beds, as well as most of those of the Loup Fork epoch, are characterized by the relatively small size of their sectorial teeth. In this they resemble the AmpMcyons, Temnocyons^ and other forms of Canidw of the same period, and differ from such true Canes as C. ursinus, C. scevuSj and C. haydeni, which display the enlarged sectorial teeth of the existing species of the genus. Of course there is every gradation in this respect between tLe two types. In the older species the internal tubercle of the inferior sectional tooth is more largely developed than in the later ones, thu.« approaching some of the species of Yiverridce, where it is still more largely developed. As in other characters, there are gradations in this also, so that neither in it nor in the relative size of the sectorials do I find ground for the separation of the species in question from the genus Canis, as has been proposed in the case of some of the species in Europe. Through the kindness of M. Filhol, I possess jaws of a number of the species found by himself and others in the Phosphorites of Central France, including the Canis Maunus^ the type of the genus Cynodon o\ Aymard. These agree very nearly with the species of dogs from the American Miocene beds as to generic characters. Professor Owen, in the paper above cited, proposed to distinguish the genus Galecynus ovi account of the greater length of the pollex as compared with that found in the existing species of Canis. This character appears to me to be of an unsatisfactory nature, owing to the fact that gradations in the length of a digit are difficult to express with precision in other than j specific sense; and the gradations may certainly be expected to occur.. I find in the G. geismarianus a character which separates the genu5 from CaniSj viz, the presence of the epitrochlear foramen of the hu merus. In this point it agrees with Anvphicyon and Temnocyon, I ar Iio.7.\ NIMRAVIDiE AND CANIDiE OF THE MIOCENE PERIOD. 181 range cotemporary and generally similar species under the same gen- eric head, as the most reasonable course in the absence of direct evi- dence. The American species of Galecynus, then, may be arranged as follows : I. Smaller species with little or no sagittal crest. * Temporal ridges uniting close behind orbits ; otic bullf© small. Small ; no external ridge on inferior sectorial G. gregariua Cope. Temporal ridges uniting early ; otic bullae large. . Larger ; no external ridge on inferior sectorial ; teeth robust. |p G. geismarianua Cope. Smaller ; an external ridge on lower sectorial ; teeth more robust. G. latidens Cope sp. nov. ** Temporal ridges not uniting anteriorly ; otic bullfe large. K Least ; muzzle narrow ; superior tuberculars wide ; no external ridge on in- ^m- ferior sectorial G. lemur Cope. HYiENOCYON Cope. Paleontological Bulletin, No. 31, 1879, p. 3 (Dec. 24); Proceedings American Philosophical Society, 1879, p. 372. This genus re^ts on the characters furnished by a single species, which is represented by but few remains. Its family position is doubt- ful, and my reference of it to the Ganidce is only provisional. It may, so far as the evidence goes, be a member of the Mustilidce or even of the Felidce, Dental formula: G.|; Pm. M. p\. Last superior molar rather narrow, transverse. Inferior premolars all two-rooted, and with well- developed posterior cutting lobe. Inferior sectorial large, with heel. Probably no inferior tubercular tooth. The characters above given agree with those of Icticyon in the su- perior series, but differ in the inferior in the absence of the Pm. I. and the M. II. The only known species is the Hycenocyon hasilatus [JEnhydrocyon hasi- lotus Cope olim.), from the Truckee beds of Oregon. Icticyon Lund. Kongl. Danske Vidensk. Selsk. Afhandl. naturvidensk. og math. Classe, ix, Deel, 1842 (October, 1841) fide Burmeister; Van der Hoeven Wissen. en natuurk. Verb, der Koninkl. Akademie Amsterdam, Deel iii ; Burmeister Erlauterungen zur Naturgesch. Brasiliens, 1856, 2. — Cynalicus Grayj Ann. Magaz. Nat. Hist. London, xvii, no. 112, 293. — Melictis Schinz. Revue et Magaz. de Zoologie, 1848, 176, fide Burmeister. The dental formula is, I, f j C. ^ ; Pm. f ; M. J. The single superior tubercular molar is similar in general to that of other Canidce. The in- ferior sectiorial has an internal cusp and posterior heel, the latter with a low cutting edge on one side. Inferior tubercular well developed. One existing and one extinct species have been found in Brazil, the latter in the caves. I described a species from the Miocene which I can- not separate from them generically. This is the Icticyon crassivuUus Cope (Proceedings Academy Philadelphia, 1879, p. 190). Art. Till.— On the Tertebrata of the Wind River £oeene Beds or Wyomingr. By E. D. Cope. The Wind Eiver, the principal source of the Big Horn, rises in the Wind Eiver Mountains, in Western Central Wyoming, and flows through 1 bad-land region for a hundred miles. This region was explored by Dr. F. V. Hayden in 1858, who makes the following observations re- specting it (American Naturalist, 1878, p. 831) : Along tlie east side of the Wind Eiver Mountains, and filling up the Upper Wind River Valley, is a great thickness of Tertiary strata that has been weathered into ^^ery remarkable forms, and which are known in the West as ''bad lands". The jtrata are most beautifully variegated with various shades of pink or brick-red color, io that they sometimes remind one of the Jura-Trias red beds. This formation was lescribed by me in 1859 in detail, and named the Wind Eiver group. It covers a jroad area in this region, extending from the source of Wind Eiver to the Sweet WqXgt Mountains, south, more than one hundred miles, and west an average width of me to five miles. The aggregate thickness of this group cannot be less than 5,000 eet. On the west side of the Wind Eiver Mountains no formations older than the vV asatch group are found. This group rests, doubtless, on the Archaean nucleus, in- iliuing at the base 5 to 10 degrees. All the older sedimentary rocks have been entirely iwept away from the granites for a distance of 100 miles ; while on the opposite or 'ast side, all the corresponding strata are visible, from the Silurian to the Cretaceous The Wasatch beds cover a large part of the Green Eiver Valley, especially about its iources. During the past summer I sent a party into the Wind Eiver Basin, mder direction of Mr. J. L. Wortman, already well known from his lumerous important i^aleontological discoveries in Oregon. This gen- ieman made a thorough exploration of the bad lands, and probably ob- ained all the fossils found on the surface in the region. The following ist of forty-five species shows that the collection embraces nearly all of be characteristic types of the American Eocene, and that twenty-six ipecies are new to science. Among the most remarkable of these I nay cite the large flesh-eater Protopsalis tigrinuSy the largest of the Eocene i)eriod yet known, and the Amblypod, Bathyopsis fissidens, an inportant addition to the forms of that peculiar order. Mr. Wortman's explorations were not accomi)lished without accident, 18 having lost most of his outfit on his first crossing of the Wind Eiver. Che bad lands form a most forbidding region, mostly waterless, and at elevation which is nnfavorable to the sparse vegetation which is I )ermitted by the dryness of the climate. 183 184 BULLETIN UNITED STATES GEOLOGICAL SURVEY. IFoLVI PISCES. 1. Glastes sp. Scales of this genus are moderately abundaiifc. 2. Pappichthys sp. Vertebrae of this genus occur in the collection. LAOERTILIA. 3. Placosaurus. A species probably of this genus is not rare j and vertebrae indicat two or three species of lizards. TESTUDINATA. Tortoises are not abundant j a portion of the plastron of a specie probably of — 4. Dermatemys— Being the only determinable fragment procured by Mr. Wortman. CROOODILIA. 5. Crocodilus sp. Not very common. EODEOTIA. 6. Plesiarotomys bucoatus Cope. Three individuals. 7. Plesiarotomys delioatissimus Leidy. Four individuals. 8. Plesiarotomys delicatior Leidy. Eight individuals. CHIEOPTEEA. 9. Vesperugo anemophilus Cope. American Naturalist, 1880, p. 745. Eepresented by the anterior part of a skull without lower jaw. D( tition: I. C. Ij Pm. 2-, M.S. Posterior molar narrow, its postei external Y rudimental; first and second molars subequal. Fourt premolar elevated and acute, with an external basal cingulum j ^ecoi premolar simple, acute. Profile steeply elevated behind orbital regioi less steep in front of it ; zygomas wide. Length from interorbital regio to above canine alveolus in front, .010 j interorbital width, .005 ; widt of zygomas, .012 j width between outsides of last molar teeth, .010 j lengt of molar series, .OO85 length of true molars, .004. BUNOTHEEIA. ] T^NIODONTA. 10. Calamodon cylindrifer sp. nov. The only individual of this species discovered by Mr. Wortman represented by fragments of the jaws, with several teeth, both loos( iro.8.] COPE ON EOCENE VERTEBRATA OF WIND RIVER. 185 and imbctlded in matrix. The former show that the molars have hut one root. The latter include the large rodent-like incisors in a frag- iiuMitary condition, and a nearly comi)lete tooth intermediate in charac- ter between the flat-banded teeth and the molar teeth of the known species of Calamodon. It may occupy an intermediate position in the jaw, but I do not know of any appropriate place for it in the mandible of Calamodon arcamcenus. I think there is little doubt the individual belongs to a species with narrower teeth than any of those of the two species already named. The characteristic tooth in question is nearly cylindric, and the part preserved is quite long and slender. Its grinding surface is worn con- cavely, as in the flat teeth of the known species of Calamodon. The enamel is in two bands, one wider than the other, and each of equal width throughout. The space of cementum separating them on one side is nearly twice as wide as that on the other. The cementum layer is not so thick as in the species of the genus hitherto described. The shaft of the tooth is slightly curved, and the wider band of cementum is on the inner side of the curve. Ileasurements. M. Widtli of enamel of large incisor 018 Length of shaft of cylindric tooth 041 ( antero-posterior Oil Diameters of grinding surface of cylmdric tooth ^ transverse 010 INSECTIVORA. 11. ESTHONYX ACUTIDENS Sp. nOV. The largest species of the genus, and represented by two individuals. The first of these includes the last molars of both series and an anterior true molar j the second includes most of the dentition of one maxillary bone, the last true molar being probably the only tooth miss- ing. Four of the molars of this specimen are in place, and three are loose. Under the circumstances, I estimate seven molars, of which the fourth premolar is like the first true molar, and the third premolar has its internal lobe very much reduced. The two preceding premolars have one root, and short, compressed, and acute crowns. The second is abruptly very much smaller than the third, and is close to it ; the first is close to the second, and is a little larger. The canine is larger still, and is somewhat compressed. Externally viewed, it looks like the i canine of a carnivorous mammal; but viewed from within, it displays marked peculiarities. It has here a median rib, separated from the fore and aft edges of the crown by a groove. This ridge is without enamel, and the edges are produced and very sharp. The enamel of the exter- nal face extends twice as far towards the base as on the interior side. The enamel of this tooth, with that of the premolars, is wrinkled ; that of the molars is smoother. 186 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [VolYX, The details of the inferior teeth preserved do not differ much from those of the U. bisulcatuSj excepting that the heel of the last true molar is much more produced. The U. acutidens is considerably larger than either of the species ot the genus heretofore described. Measurements. '\ No. 1. M. , vertical 0065 Diameters of last inferior molar ^ anteroposterior 0130 ( transverse 0064 ( anteroposterior . 0095 Diameters of a true molar | transverse 0074 c anteroposterior 0097 Diameters oflast superior molar > ^^ansverse ... .0130 No. 2. Length of five superior molars preserved 0410 Length of premolar series 0325 Length of bases of Pm. I and II 0125!! ( anteroposterior . 0097 Diameters of Pm. III| transverse 0098 ( anteroposterior 0086 Diameters of first true molar ^ transverse 0133 Anteroposterior width of base of crown of canine 0080 Transverse width of base of crown of canine , 0050 12. ESTHONYX SPATULARIUS CopC. American Naturalist, 1880 (Nov. 25), p. 908. Represented by five molar and premolar and two incisor or canme teeth, apparently belonging to one individual. These are about the size of those of U. hisulcatus, but present several differences of detail. Thus, the basin of the heel of the last inferior molar is not obliquely cut off by a crest which extends forwards from the heel, but is surrounded by aii elevated border, which rises into a cusp on the external side. The incisor-canine teeth are more robust than those of U. Msulcatus, one of them especially having a spoon-shaped crown, with the concave side divided by a longitudinal rib, on which the enamel is very thin. The enamel descends much further down on the external than the internal side of these teeth. The rodent-like tooth does not accompany the speci- men. Length of base of last inferior molar, ,009 j width anteriorly, .005 j length of crown of canine-incisor No. 1, .009; width at base, .005,* length of crown of second canine-incisor at base, .012 ; width, .006. MESODONTA. \ 13. Hyopsodus paulus Leidy. Numerous specimens. 14. Hyopsodus vicarius Cope. ^b'' Less abundant. 4' ^0.8] COPE ON EOCENE VERTEBRATA OF WIND RIVER. 187 15. Pelycodus jarrovii Cope. A jilw fragment supporting the last two molars presents the charac* jQTS and dimensions of this species. The genus Pelycodus differs from Tomitherium in that the second premolar resembles the later ones in laving two roots instead of having but one root like the first, as is found n the latter genus. :6. Pelycodus tutus Cope. Tomitherium tutum Cope, Report Expl. Surv. W. of lOOth Mer. under Capt^ Wheeler, iv, pt. ii, p. 141. Eepresented by numerous specimens. 7. Pelycodus nunienum sp. nov. Fragmentary jaws of six individuals of this species were found by It. Wortman. They indicate a species intermediate in dimensions be- ween the P. tutus and P. frugivoruSj which is further defined by the orm of the last inferior molar. The best preserved ramus supports all the teeth posterior to and in- luding the third premolar. The last-mentioned tooth has an elevated cute crown, without any anterior basal tubercle, and a very short pos- erior heel. The fourth premolar is very stout ; its cusps are not much levated, and the heel is short. The anterior basal tubercle is quite mall. All of the true molars have a second cusp in front of the anterior iterval, but it is quite small, excepting on the first, where it is more^ istinct. The external crescents of all the molars are well defined, but le posterior does not inclose the crown behind with an extension of its orn. The last molar is a little longer than the others, and its posterior order is produced into two cusps. A simple raised border is found here |i P. frugivorus, 1 Measurements, M. ength of molar series from third premolar, inclusive . . - i , . 0228- sngth of true molars 015O ( anteroposterior 0050' lameters of first true molar | transverse 0038 ( anteroposterior 0065 lameters of last true molar ^ transverse 0040 ijpth of ramus at Pm. Ill 0095 5pth of ramus at last true molar 0095 J. Pantolestes secans sp. nov. Represented by the adherent rami of a mandible, on both of which le posterior four molar teeth are preserved. The species is about the size of the P. chacensis, and hence larger than le P. longicaudus. It differs from both in the proportions of its teeth,, id especially in the large size and sectorial character of the fourth pre- olar. The length of the latter exceeds a little that of the third true olar, while in the other species it is shorter. This length is caused by :e extent of the anterior basal tubercle and posterior heel. The latter 188 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [YolV] is entirely surrounded by a cingulum, and its median line is elevatei into a blade, which is continuous with the posterior edge of the princi pal cusp. Both edges of the anterior tubercle are also trenchant. Tht two cusps of the anterior inner tubercle of the first and fourth molar are well developed, but on the second molar there is but one cusp. Thi is probably a character to be relied on in distinguishing the species froii the P. chacensis. E'o external basal cingula j enamel smooth. As is the case with the species of Pantolestes already known, the I secans seems to have been rare. 19. MiCROSYOPS SPEIRIANCJS CopC. American Naturalist, 1879, p. 908. Established on a portion of a mandibular ramus which contains th< three true molars in perfect preservation. As the number of premola teeth is unknown, its refer^ce to this genus is provisional only. Thi last true molar has the form of that of the M. gracilis Leidy. It is dis tinguished by its very small size, since it is considerably less than th' H, vicarius {H, ? minusculus), and by the equality in size of the molar,'^ The heel of the third molar is very small, and the two cones of the inne side of the crowns of all the molars are acute. The external crescent are very well defined, the anterior sending a horn round the anterio extremity of the crown. The posterior is connected with the corrc sponding internal tubercle by a median conic posterior tubercle. Lengtl jot true molar series, .008 j length of second molar, .0026 ; width of se( ond molar, .0022 ; length of last true molar, .0025 j width of last trn molar, .0016 ; depth of ramus at second molar, .0043. Dedicated to m. friend Mr. Francis Speir, of Princeton, JST. J., who, in connection wit) Messrs. Scott and Osborne, has made important additions to our know) .edge of the Eocene Vertehrata. \ 20. MiCROSYOPS GRACILIS Leidy. I Eepresented by numerous jaws. f 21. MiCROSYOPS SCOTTIANUS Sp. nOV. A nearly entire left mandibular ramus is all that I have seen of thi species. The crowns of the fourth and sixth molars furnish the onl dental characters available, but the number and forms of the bases < the others are readily ascertainable. The ramus of the jaw is more slender than in 31. gracilis, and the la^ true molar has quite a different form. Instead of being shorter than i Measuremeyits. Length of last four molars Length of fourth premolar Elevation of fourth premolar Length of last true molar Depth of ramus at first true molar M .021 .000 .004 .005 .007 0.8 ] COPE ON EOCENE VERTEBRATA OF WIND RIVER. 18^ Uicd species, this tooth is rather longer, evidently in consequence of a ell-developed heel. The fourth premolar has a strong inner tubercle, Qd no anterior cusp or cingulum. Its heel has an elevated posterior order, enclosing a fossa with the principal cusps. No external or inter- al cingula. Third premolar with two roots. Alveolus of the second, ^rge and apparently simple j it is filled with matrix. Canine large, irected forwards, and occupying all the space between a short dias- 5ma and the symphysis. The latter extends posteriorly to below the aterior part of the third premolar. The ramus is compressed and laintains an equal depth to the end of the molar series. Its inferior order descends below the coronoid process, and is not incurved, but le external face is convex. The anterior masseteric ridge is well larked, descending to below the middle of the ramus. Masseteric issa flat. Mental foramen below the third premolar. Measurements, ength of the fragment of ramus 0435 Bngth of dental series without incisors 0280 C antero-posterior 0040 iameters of canine [ ^^^^^^^^^ 0025 sngth of premolar series 0100 engtli of fourth premolar 0040 Idth of fourth premolar behind 0027 ength of true molar series 0136 sngth of last true molar 0052 idth of last true molar anteriorly 1 0030 epth of ramus at third premolar 0090 epth of ramus at last molar 0090 This species is dedicated to ray friend Prof. William B. Scott, of the ollege of ISTew Jersey. CREODONTA. 5. MlACIS CANAYUS Sp. nov. Established on the mandibular rami of two individuals, which display e roots and some of the crowns of all the teeth exclusive of the in- sors. The root of the canine indicates that the crown is of large size and mpressed at the base. The first premolar is one-rooted, and is sepa- ted from the second by a short diastema. The second has two well- stinguished roots, which are separated from those of the third by a dias- ma like that in front of them. Posterior to this there are no diastemata. ' le second root of the fourth premolar is much larger than the anterior. ' le sectorial, though the largest tooth, is of but moderate dimensions ; :} heel supports two posterior tubercles. The first tubercular is a lit- ii shorter. It has a raised border, and the anterior part two angu- tubercles. The second tubercular is a very small tooth, but has two 3ots, the posterior of which is posterior to the anterior border of the J cending ramus. 190 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [rolYlJ^ According to Leidy's measurements, this species is about the size of his M. vorax of the Bridger formation. That species has, like the two others of that horizon, a second tubercular tooth with only one root. Measurements. Length of dental line posterior to canines . 0440 Length of premolar series 0250 Length of base of fourth premolar 0065 Length of base of sectorial 0085 Length of base of first tubercular 0060 X/ength of base of second tubercular 0040 Depth of ramus at second premolar 0150 Depth of ramus at second true molar 0100 This species was probably about the size of the gray fox. 23. MiACis BREViROSTRis sp. nov. This species differs from those of the Bridger epoch in the same way that M, canavus does, i. e., in the biadicate last inferior molar. Its dimen- sions are intermediate between those of M. edax and M, vorax^ hence a little smaller than those of the M. canavus. This difference is partially seen in the shortening of the premolar series of teetli. They are closer together than in the M. canavus^ and the roots are larger. The sec- torial tooth is shorter. The fourth premolar has a low anterior basal <}ingulum ; the posterior part of the crown is robust. The first tuber- cular molar is wide, and consists of a basin-shaped heel and a short anterior portion which is more elevated. The latter consists of two €usps, which are connected by an anteriorly convex ledge, but there is no third anterior tubercle as in M. ^arvivorus. The ramus is quite robust, and the basis of the canine tooth is unusually large. Mental foramina are below the anterior parts of the second and fourth premolars, respect- ively. Last inferior molar small. Measurements. * I Length of molar series 038C Length of premolars - 020C Length of base of fourth premolar 006G Length of base of sectorial 0072 Length of base of first tubercular 0048 Length of base of second tubercular 0042 Depth of ramus at second premolar 0140 Depth of ramus at second true molar 0140 24. DiDYMIOTIS ALTIDENS CopC. American Naturalist, 1880, Oct. p. 746. Eepresented by several specimens. The species is larger than the D. protenuSy or about equal to the coyote; but the tubercular molar is relatively smaller, and has the three anterior cusps better developed. The heel of the tubercular sectorial is longer and the three cusps more Vo. 8.1 COPE ON EOCENE VERTEBRATA OF. WIND RIVER. 191 alevjitod than iu T>. protenus. Diameters of latter tooth : Length antero- posteriorly, .015 j length of heel, .000 j elevation of external side of 3rowu anteriorly, .015 ; width at same point, .009. Length of crown of buberuiar, .009 ; width of same, .006 ; elevation anteriorly, .005. 35. DiDYMICTIS LEPTOMYLUS Gope. American Naturalist, 1880, Doc. p. 90H. Represented by the posterior three inferior molars. These indicate a species of smaller size than the D. protenus^ with the tubercular molar relatively narrower, and perhaps longer. The anterior part of the latter lias the three cusps well defined and close together, and behind them is m oblique longitudinal cutting edge. The middle of the posterior margin rises into a tubercle. The anterior cusps of the tubercular sec- torial are elevated j the heel has a strong external cutting edge and internal ledge. Length of tubercular sectorial, .009 j width of same, 005 ; length of tubercular, .007 ; width of same in front, 26. DiDYMICTIS DAWKINSIANUS Sp. nOV. This flesh-eater is represented by more or less imperfect mandibular rami of three individuals. The most complete of these lacks only the portions posterior to the coronoid process, and those anterior to the first premolar, and supports all the teeth excepting the first and second premolars. The premolars are all two-rooted excepting the first. The base of the fourth premolar is considerably longer than that of the bbird. Both of these teeth have a short posterior heel, and above it a cutting lobe. The fourth has a well-marked anterior basal tubercle, rhe heel of the sectorial is relatively short, and the anterior portion of the tooth elevated. The anterior and inner cusps are high, and about iqual, but the external cusp is much higher. The external border of the leel is more elevated than the inner. The tubercular molar is elongate, md has a small triangular anterior portion somewhat elevated, in slight i'esemblance to the sectorial tooth. This portion consists of two opposite 3usps and a lower one in front of the anterior inner, which connects ;v^ith the external by an anterior ledge. The posterior portion has a ubercle on the external side, besides a posterior elevation. The ramus s rather slender, and the masseteric fossa is bounded by a prominent idge in front, but fades out below. The measurements show this to be the smallest species of the genus, 3eing much less than the J). leptomylus. Measurements. M. -engtli of dental series, including first premolar liengtb of premolar series.. w lremolar. Enamel wrinkled ; no external cingulum. Second premolar vith a very short heel with an acute tubercle. Length of molar series, 064; of true molars, .040; of last true molar, .016; depth of ramus at econd premolar, .020 ; at third true molar, .030. Seven individuals in he collection, one with complete series of maxillary teeth. These in- 198 BULLETIN UNITED STATES GEOLOGICAL SURVEY. rFoZ.VL j elude four premolars (the first one-rooted), so that the formula M. |f is that of Fachynolophus Pom. rather than of Lophiodon. 1 refer another species to the same genus, by analogy, as it agrees in the dentition of ! the inferior jaw. 39. Hyracotherium angustidens Cope. Apparently an abundant species. There are three sizes which I refer here, which may represent different species, but this cannot be deter- j mined without better material : A. Depth of ramus at last premolar or first true molar, .0120 j length of crown of first true molar, .0070; length of last true molar, .0100. Lower jaw of one specimen. B. Depth of ramus, .0140 ; length of first true molar, .0065 ; of last molar, .0100. One lower jaw. 0. Depth of ramus, .0155 ; length of first true molar, .0075 ; of last true molar, .0100. Two individuals. Portions of lower jaws of three other individuals in the collection are apparently referable to the H. angustidens, 40. Hyracotherium venticolum sp. nov. Hyracotherium vasacciense Cope, American Naturalist, 1880, p. 747 ; not Report Expl. Surv. W. of lOOth Mer. iv, p. 264. Eepresented by an entire skull, with some bones of the skeleton, of one individual. In general, this species is to be distinguished from its near ally, the H. vasacciense, by the slender mandibular ramus. The depth of this bone is about equal to that found in the larger varieties of the R. angus- tidens, but the teeth are much larger, having the proportions of those of: the H. vasacciense. This remark applies especially to the last inferior molar. The inferior canines form part of an uninterrupted series with thei incisors. The superior canine is separated from the superior incisors by a diastema. The first premolar in both jaws is isolated. The second superior premolars have two cusps, and an internal ledge posteriorly. The third and fourth superior premolars are similar, the fourth display- ing a little larger transverse diameter. The true molars are of subequal dimensions. Their external cusps are subconic. All the molars except the first and second premolars are entirely surrounded by a basal cin gulum, which rises into a low cusp at the anterior external angle of th< crown. The third inferior premolar has its two median cusps wel3 separated and a wide posterior heel. The heel of the last premolar i> wider, but carries no internal cusp. The external cusps on all the teetl wear into well-defined Ys. The posterior five molars have an externa basal cingulum, but no other. The mandibular ramus is compressed. The ascending ramus rise almost vertically a short distance posterior to the last molar. The sym physis is narrow, and extends to below the middle of the first premolar The infraorbital foramen opens above. .8] COPE ON EOCENE VERTEBRATA OF WIND RIVER. 199 Measurements* M. ngtli of consecutive superior molars 04500 ngth of diastema between Pm. I and II 00350 ngth of second premolar 00700 idth of second premolar posteriorly 00500 ameters fourth premolar ^ anteroposterior 00700 i transverse 00900 ameters second true molar J anteroposterior 00850 c transverse 01120 ngth of entire inferior premolar series 05800 ngth of entire inferior true molar 00282 ameters first true molar J anteroposterior 00800 I transverse 00600 ameters last true molar 5 anteroposterior 01200 ( transverse 00650 pth of ramus at fourth premolar 01650 ptli of ramus at third true molar in front 01650 The lower jaw of a second individual agrees with the type in its pro- rtions. . Hyracotherium craspedotum Cope. American Naturalist, 1880, p. 747. Size of S. tapirinum, but the tubercles of the inferior molars are not nnected by cross-crests, and they all possess a strong external basal igulum, which also extends round on the posterior base of the I and II le molars. Heel of fourth premolar with a diagonal ridge j two ante- ►r cusps well separated, and no tubercle in front of them* Second emolar with narrow heel ; last true premolar with wide heel. Length molar series, .056 j of true molars, .033 ; of last molar, .014 ; depth of tnus at second premolar, .018 ; at last true molar, .023. This is the largest species of the genus found in the Wind River beds, brts of two individuals were obtained by Mr. Wortman. . Orotherium vintanum ('^ Marsh") Cope. Report Expl. Surv. W. of 100th Mer. iv, p. 255. A portion of the mandible of a single specimen, containing the charac- [istic fourth premolar and other teeth. ARTIODACTYLA. A species of this order is represented by an astragalus. This is the st indubitable evidence of the existence of this order during the asatch epoch that I have seen. The following species are referred re provisionally only, as no part of their skeletons is known. . Phenacodus vortmani Cope. Hyracotherium vortmani, American Naturalist, 1880, p. 747. This species is represented by portions of mandibles of four individ- As. One of these supports the i)osterior four molars, another the pos- rior two, and another the last premolar and first true molar. 200 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [Vol.Vl These indicate an animal of much smaller size than the P. primcevm but with a similar constitution of the molar teeth. The teeth suppori four conic cusps, which do not incline to fuse transversely, as is geuer ally the case in Ryracotherium. Those of the posterior pair are sepa rated by a tubercle, and a rudimental tubercle stands behind the notch between those of the anterior pair. The posterior median tubercle is developed into a heel on the last molar. On prolonged wear, the re suiting pattern represents two Ys, the posterior limb of the posterioi being regularly convex backwards. There are no cingula on any of the molars. The ramus is rather robust and is not deep. Measurements. No. 1. Length of bases of posterior Ave molars 0440 Length of bases of true molars 0250 Depth of ramus at Pm. Ill 0160 Depth of ramus at M. Ill in front 0190 No. 2. ^ Diameters of M. II \ anteroposterior 0080 C transverse .... 0070 Diameters of M. Ill \ anteroposterior 0080 ( transverse - 0055 The characters of the typical specimen are as follows : The jaw frag- ment indicates an animal of about the size of the Hyracotherium craspe- dotum, but with the opposite cones of the inferior molars not united by cross-crests. There is a tubercle between the posterior pair of the first in- ferior true molar. The anterior tubercles of the fourth premolar are close together, and there is a strong cusp anterior to these. No basal cingulum on this tooth. Length of molars 3 + 4 + 5, .025 ; depth of ramus at Pm. IY, .018. 44. Phenacodus prim^vus Cope. A fragment of a Tower jaw, with teeth. 45. Phenacodus teilobatus Cope sp. nov. A lower-jaw fragment, supporting the three true molars, of one indi- vidual, represents this species. It is of the dimensions of the P.i)rim(B' vus, and displays the same general constitution of the teeth. The only difference noticeable is an important one. The anterior internal tuber- , cle is accompamed by two others of less elevation — the one immediately i anterior, the other immediately posterior to and not deeply separated from it. The internal face of these tubercles slopes obliquely outwards on the second and third true molars. The external tubercles have their external faces sloping inwards on all the true molars. There are no cingula. yo.8.] COPE ON EOCENE VERTEBRATA OF WIND RIVER. 201 Measurements. M. Length of true molar series , 0390 .L If n J. ± 1 C anteroposterior 0107 Di^imoters of first true molar < . ^ ^ i transverse • »0097 ^ /..I.. J J 1 (anteroposterior .OlSO Diameters of third true molar < , • ^ x aiaa ( transverse in front 0100 GENERAL OBSERVATIONS. Until the faunae of the Wasatch and Bridger epochs are better known it will not be possible to fix the relation which that of the Wind Biver beds holds to them. It is, however, quite evident that in some respects it differs from both, both by what it possesses and what it lacks. In the following lists these differences are displayed so far as they relate to genera. The left-hand column represents the Wasatch, the middle the Wind Eiver, and the right hand the Bridger. Wasatch. Wind River. Bridger. Clastes Clastes. Clastes. Pappichthys. Pappichthys, Crocodilns, Crocodilus, Crocodilus, Plesiarctomys. Plesiarctomys, . Plesiarctomys, Myops, Yesperugo. ? Vesperugo, Galamodon, Calamodon, Estlionyx, Estlionyx, Ectoganus, AncMppodus, Ictops. fictops, Mesonyx, Miacis. Miacis, Didymictis, Didymictis, ? Oxycena, ? Pachycena, 1 Protopsalis, ? Stypolophus. Stypolophus, Stypolophus, Microsyops, Microsyops, Pantolestes, Pantolestes, Pantolestes, Pelycodus, Pelycodus, Tomitherium, Anaptomorphus. Coryphodon, Coryphodon, Bathyopsis. Uintatherium, Palceosyops, Palceosyops, Lamhdotherium, If fPacliynolophus, Pachynolophus, Syracotherium, Hyracotherium, ? Orotherium, Orotherium. ? Hyrachyus, Phenacodus. Phenacodus, 202 BULLETIN UNITED STATES GEOLOGICAL SURVEY. [F^.jvi. From the above it is evident that the Wind Eiver fauna -in- cludes genera which have been hitherto restricted to either the WS,- satch or Bridger lists. Of these, especially Bridger genera, there are six 5 that is, six genera which have not yet been detected in beds of the Wasatch epoch. On the other hand, there are nine genera which have been found in Wasatch beds and not in those of the Bridger. Three important Wasatch genera have not been found in the Wind Eiver formation, while seven of the characteristic genera of the Bridger are not included in the list of those of the Wind Eiver. The result, imper- fect as it is, indicates a considerably greater conformity to the Wasatch epoch than to the Bridger in the faunal characters of the Wind Eiver beds, and points to the confirmation of Dr. Hayden's views as to the identity of the two epochs. The new species above described will be fully illustrated in the fourth volume of the,3eport of the United States Geological Survey of the Ter- ritories, F. YT'Hayden in charge, now passing through the press. NATURAL SCIENCES OP PHILADELPHIA. 203 On the Crocodilian genus FEBOSUCHTI S. BY EDWAKD D. COPE. Characters. — Toes 5-4, with claws two — three. No osseous nasal septum or bony eyelid. Belly protected by series of osseous plates, as well as the back. All the genera of Crocodiles hitherto known as living are characterized by the possession of three claws on the fore foot. The pres^-nt therefore oifers a remarkable exception. The free fingers and half webbed toes, and the bony abdominal buckler, together with the cartilaginous nasal septum are points of strong resemblance to Jacare (Gray, including Cienwn Gray), but it differs from these creatures in the lack of bony orbital plate. In specific cha- racters it differs from those of this genus, which it most resembles, as J. n i- gra, in the absence of a transverse bony ridge between the orbits. Another feature of importance is the relation of the canine teeth of the lower jaw to the upper. On one side this tooth is received into a notch, as in Cro'bodiles, on the others it enters a pit of the maxillary bone, within the border of the same, as in Alligators I This remarkable combination may be abnormal even in this species, but this cannot be now ascertained, as it rests at the present time 00 a single specimen only. As its affinities are rather more alliga- torial, I am disposed to anticipate that the dental arrangement of the latter type will be most common. Perosuchus fuscus Cope. Char, specifics. — Nuchal plates in a cross row of six ; cervicals in four cross-rows, all of four plates except the last of two. Dorsal plates six — in a few eight in each transverse row. No posterior crest on arm or leg. Tail short with remarkably low crest. Muzzle broad, flat, without any ridges ; its width at the eighth tooth entering 1-4 in length from end of muzzle to anterior margin of orbit. Description. — The specimen in the Museum of the Academy is young, measuring only 2 feet 5 inches in length. Of this the skull measures to the margin of the supra-occipital 2 in. 10 5 lin. ; and the tail from the vent 13 in. 1 line. From groin to heel 3 in. 2-5 lines, and the hind foot 2 in. 7-5 lines. The muzzle is a broad ovate, the sides rather more convergent anteriorly than in the Alligator mississippiensis. There is a thickening in front of each orbit, and between them on the middle line another, which together enclose two shallow concavities. Superciliary margins raised, the cranial table quite flat. The margin of the quadrato-jugal bone projects strongly. The scales of the limbs are all smooth and those of the dorsal region with very low keels. The sides have four longitudinal rows of ovate scales sepa- rated by scarcely defined smaller ones. The abdominal plates are longer than broad, and are in twelve longitudinal rows. Dorsals in seventeen transverse series from interscapular to crural region. The lateral crests of the tnil are only obtuse keels ; they unite on the thirteenth annulus behind the vent inclusive. Color above dark brown, almost black on the upper sur- face of the head. The tail is paler^ of a light olive brown. Lower surface everywhere bright yellow, including the entire lower jaw and margin of the upper. Eyelids and a band through ear yellow, the former with a black spot above. Remarks. — This interesting addition to our knowledge of the Reptilia was made by Schuite Buckow, of New York, while on a visit to the interior part of the course of the Magdelena River, in New Grenada. This naturalist has also enriched our collections with other interesting vertebrata of that region, both living and dead. 1868.] 204 PROCEEDINGS OF THE ACADEMY OF Sept. 1st. The President, Dr. Hays, in the Chair. Twenty-two members present. Sept. Sth. The President, Dr. Hays, in the Chair. Twenty-eight members present. Sept. 15th. The President, Dr. Hays, in the Chair. Twenty-eight members present. The following papers were presented for publication : Extinct Mammalia of Dakota and Nebraska, including an account of some allied forms from other localities, together with a synopsis of the Mammalian Remains of North America. Illustrated by twenty- eight plates. By Joseph Leidy, M. D. Preceded by an introduc- tion on the Geology of the Tertiary Formations of Dakota and Nebraska, accompanied by a Map. By F. V. Hayden, M. D. Notice of American Species of Ptychodus. By Joseph Leidy, M. D. Synopsis of the Extinct Batrachia of North America. By Edw. D. Cope. _ Dr. Leidy read a letter from Mr. B. Waterhouse Hawkins, pro- posing to erect in the Museum, at his own expense, the fossil remains of the Hadrosaurus to their natural relations in the figure of that great Dinosaur, in accordance with Dr. Leidy's descriptions in his Monograph of the Cretaceous Reptiles of the United States. On leave being granted, the following resolutions, offered by Dr. J. L. LeConte, were adopted : That the Academy accept the proposition of Mr. B. Waterhouse Hawkins, to erect in this Hall, at his own expense, a restoration of the skeleton Hadrosaurus. That the thanks of the Academy be respectfully tendered to Mr. Hawkins for his liberal offer, and that the Curators be instructed to furnish to him every facility in the use of specimens in the Museum, w^hich the most liberal interpretration of the By-Laws will permit. Sept. 22d. Dr. Bridges, in the Chair. Twenty-four members present. Sept. 29th. The President, Dr. Hays, in the Chair. Twenty-four members present. [Sept. NATURAL SCIENCES OF PHILADELPHIA, 205 Dr. D. G. Brinton was elected a member. On favorable report of the Committee, the paper of Dr. Leidy, presented Sept. 15th, entitled " Extinct Mammalia of Dakota and Nebraska," etc., reported in favor of its publication in the Journal. On favorable report of the Committees, the following papers were ordered to be printed : Notice of American Species of PTYCHODUS. BY JOSEPH LEIBY, M. D. The Cestraciont genus of fishes Ptychodus, so far as known, is confined to the Cretaceous Formations. Remains, consisting of teeth, I have had the opportunity of inspecting from Alabama, Mississippi and Kansas, and although reported to exist in the Cretaceous Formation of Delaware, I have not met with them from that locality nor from the Green Sand, of correspond- ing age, of New Jersey. The following list comprises all the specimens of American Ptychodus teeth I have had the opportunity of examining. Ptychodus Mortoni. Agassiz, Poissons Fossiles III. (1833-43), 158, Tab. 25, figs. 1 — 3; copied in figs. 773, 773a, of Dana's Manual of Geology. Palate bone of a fish ? Morton : Syn. Org. Rem. Cret. Group. (1834), pi. xviii, figs. 1, 2. The teeth of Ptychodus Mortoni I have seen only from the cretaceous forma- tion of Alabama and Mississippi. Morton, in the work above noticed, figures a tooth, but does not mention the locality from which it was obtained. Agassiz, in his Poissons Fossiles, gives a good representation of a tooth of this species, from the Green Sand of America, in three views, figs 1 — 3, Tab. 25. Dixon, in his Geology of Sussex, represents two small teeth, (figs. 6, 7, pi. xxi), which he refers to the same species. Though exhibiting some resem- blance in character to the American teeth, I think a further comparison is necessary to establish their specific identity. The teeth of Ptychodus Mortoni are well defined in character, and in com- parison with teeth of well recognized European species are almost generic in their peculiarity. Though exhibiting some variety, their likeness presents a distinct specific uniformity. Their size of course varies greatly with age and the relative position they occupied with one another in the mouth of the fish. Viewed from above, the crown is reniform in outline, the long diameter being transverse ; the incurvature posterior. The crown rises in the form of a cone with a more or less obtuse summit. The sides of the crown slope to the base and frequently more or less abruptly expand, laterally approaching the latter. The back partis occupied by a wide triangular sinus for the reception of the fore-part of the crown of the tooth which was situated in front of it when the teeth were contained within the mouth. The border of the crown is thick and rounded and dips beneath. At the sinus it is prominent. The summit of the crown presents a prominent crucial ridge, more or less distinct in different specimens. From the cross numerous ridges of about the same thickness diverge upon the sides of the cone, branching in their course, multiplying and becoming finer, and ultimately conjoining upon the base in a fine reticulation extending to the borders of the crown. The coarser ridges vary in their pro- portionate length in different specimens. The reticulation of the base is most extensive laterally, occupying usually half the breadth of the space between the summit and border. It also occupies the sinus, and is least developed at the fore-part of the crown. The width of the crown approaches double the 1868.] 206 PROCEEDINGS OF THE ACADEMY OF fore and aft measurement, and the height is usually little less than the latter. The root partakes of the form of the outline of the base of the crown, but is more square and is flat or transversely concave below. Twelve specimens of teeth of Ptychodus Mortoni^ from the Cretaceous For- mation of Alabama, belonging to the Yale College Museum, have been sub- mitted to my examination by Mr, William M. Gabb. Among them occurs the largest tooth of the species I have seen, and larger than any on record. It is labelled as having been derived from Perry Co,, Alabama, The fang and parts of the lateral and back borders of the crown are broken away. la the perfect condition the crown has measured a little over two inches in transverse diameter, one inch and a quarter antero-posteriorly, and ten lines in height. The crushing surface is porportionately less prominent at the centre than in the smaller teeth attributed to the same species, and is more uniformly convex, or less expanded laterally at the base. The borders of the posterior sinus also are less abrupt or defined. The unworn summit presents a crucial ridge, of which the lateral radii are most distinct and directed pos- tero-laterally. From the crucial ridge, numerous ridges, equally prominent, diverge, branch in their course and ultimately conjoin in a fine reticulation at the base af the crown. This reticulation has the greatest breadth at the sides of the crown and is least developed at the fore-part. Eleven teeth from Uniontown, Alabama, exhibit a gradation in size from less than three-fourths that of the above described specimen down to one little more than a fourth of its diameter. The specimens present a remarkable sim- ilitude throughout. Some are proportionately wider fore and aft than others, and the smallest are porportionately higher than the largest ones. The outline of the base of the crown is reniform, with the relation of the longer and shorter diameters varying. The largest specimen has the crown an inch and a half wide, a little over three-fourths of an inch fore and aft, and about half an inch in height. The sides of the crown expand at the base laterally; the fore-part forms nearly a uniform slope, and the back surface slopes to the sinus, which forms a broad triangular depression. The fang is fourteen lines wide, seven lines fore and aft, and three lines in depth. The crown of a median sized tooth of the series, unworn, measures scant 14 lines wide, 7^ fore and aft, and 6J high. The smallest specimen has the crown 7 lines wide, 4^ fore and aft, and an equal height. Its base laterally appears more abruptly expanded than in the others. Most of the specimens are unworn and exhibit the characteristic ridges of the crown in a striking manner. In three specimens the coarser ridges are resolved into the reticulation much earlier or nearer the summit than in the others. In one specimen the crown is smooth or totally devoid of ridges, presenting the same appearance repre- sented in figs, 4, 5, pi. XXX, of Dixon's Geology of Sussex, and described as " nascent or incomplete teeth of Ptychodus. Seven specimens of teeth in the Museum of the Academy, from Alabama, exhibit the same characters expressed in the description of those above. They all present an unmistakeable specific likeness, though varying in the propor- tions of their diameter. The largest specimen has the crown 16 lines wide, 11 lines fore and aft, and 8 lines high. The root is an inch wide, 7 lines fore and aft and nearly 3 lines thick, A second specimen, with the crown 16 lines wide and 9 lines fore and aft, has been proportionately lower than the former. Its summit is worn away, leaving an exposed circular disk of vaso- dentine 4 lines in diameter. Two specimens in the Museum of the Academy, presented by Prof. Joseph Jones, are from Green Co., Alabama. The larger is perfect and unworn. The crown is scant 14 lines wide, by 7 lines fore and aft, and 5 lines high. The root is 11^- lines wide, 4^ fore and aft, and 2 lines thick. Two specimens in the Museum of the Academy, presented by Dr, Wm, Spill- man, are from Columbus, Mississippi. They present the same character as the Alabama specimens. The larger specimen has the crown 20 lines wide, 10 lines [Sept. NATURAL SCIENCES OF PHILADELPHIA. 207 fore and aft, and has perhaps been about 8 lines high. The summit is worn off, leavincf an exposed flat circular surface of vaso-dentine half an inch in diame- ter. The root is 15 lines wide, G lines fore and aft, and three lines thick. The smaller specimen consists of an unworn crown 11^ lines wide, 1 lines fore and aft and 5J lines high. Ptychodds occidentalis, n. s. The Museum of tlie Academy contains a specimen consisting of the crown of a tooth of a species of Pfyc/iodus ditlering from any other previously known. It was obtained by Dr. John L. LeConte, in association with other remains of fishes, from an ash-colored rock of the Cretaceous series, a few miles east of Fort Hays, Kansas. The tooth is remarkable, especially from the comparatively near approxima- tion of its diameters, the width transversely and fore and aft and the height approaching one another more nearly than in any other species. The fore- part of the crown is somewhat injured and the root is broken away. The transverse diameter of the crown at base is 14 lines ; the fore and aft diameter has been about an inch ; and the height is also an inch. In shape the crown is a blunt cone with the sides sloping evenly to the base and to the posterior sinus. The latter is a triangular concavity about two- thirds of the breadth in height. The direction and arrangement of the ridges of the crown are much like as in the European Ftychodus decui-rens, but the principal ridges crossing the crown transversely are finer and the intervals much narrower, indeed the space occu- pied by a pair of ridges with their interval in P. decurrens would accommodate three ridges with a pair of intervals in P. occidentalis. Descending the sides of the cone the ridges branch as in P. Mortoni, and at the basal half of the crown form a reticulation much as in P. decurrens. At the back of the summit of the crown the principal ridges continue their transverse or parallel course until near the upper part of the sinus, into which as they descend they are resolved into a fine reticulation. The fore-part of the crown is occupied by a reticula- tion formed by the descent, convergence and division of the more anterior principal ridges. From the description it will be observed that the tooth holds an interme- diate position in anatomical character to those of Plychodus 3fortom, and P. dec wrens. Three small teeth, found by Dr. Le Conte in association with the latter, re- semble, in their proportions and in the proportionate size and arrangement of the ridges of the crown, the teeth of P. decurrens, but perhaps may belong to the same species as the large tooth above described. The larger of the three specimens is perfect, but has the summit of its crown worn off". The crown measures 7 lines transversely, 6 lines fore and aft, and has been from 4 to 5 lines high. The root is 6 lines wide, 4J lines fore and aft, and 2^ lines thick. Comparatively coarse ridges cross the crown transversely, curving forward laterally and ending in a marginal reticulation, ^j^nching ridges descend in front from the foremost of the transverse ridges, aHpiliewise end in a marginal reticulation. The sinus is occupied by a finer reticulation joined by fine ridges descending from the summit and sides of the crown. The smallest tooth, like- wise perfect, has the crown 4^ lines wide, 3-^- lines fore and aft, and 2^ lines high. Three additional specimens associated with the former ones, are the smallest teeth of Ftychodus 1 have seen, but I suspect that they belong to the same species. They are transversely ellipsoidal in outline at the base of the crown, and this appears as a low cone elevated at tlie inner third and with a broad expanding base. The sinus is situated at the inner posterior third. The sur- face of the crown is crossed with transverse ridges which form a narrow retic- ulation at the border. The largest of these small specimens is 3^ lines trans- verselv, 1| fore and aft, and f of a line high from the root. The smallest tooth is 2^ lines wide, 1| fore and aft, and ^ a line from the root. 1868.] 208 PROCEEDINGS OF THE ACADEMY OF Ptychodus polygyrus. Agassiz : Poissons Fossiles III, 156 ; Dixon: Geology of Sussex, 1850, 363. Gibbes : Jour. Acad. Nat. Sci., 1849, 299, pi. 42, figs. 5, 6. Dr. Gibbes, in the work above noticed, figures two teeth, from the cretaceous formation of Alabama, which he refers to Ptychodus polygyrus. They clearly bear a close likeness to specimens of the European species of that name. A single specimen of a tooth, accompanying the Alabama specimens be- longing to the Yale College collection, resembles the teeth of the European Ptychodus polygyrus. The crown is nearly square or transversely oblong, with the fore and back borders nearly straight, and the lateral borders convex. The crushing surface is moderately convex and is crossed transversely by ten coarse acute ridges, separated by similar intervals. The borders of the sur- face, including the posterior sinus, are occupied by comparatively fine vermicu- lar and interrupted ridges, appearing like granulations. The coarse ridges are nearly straight, and at the end rather abruptly resolve themselves into the finer vermicular ridges of the border. From European specimens of the teeth of P. polygyrus and P, latissimus, this tooth appears especially to differ in the proportionately greater degree of fineness of the bordering vermicular ridges or granulations of the crown. Its measurements are as follows : Width of crown 13 lines; fore and aft 11 lines; height 5^ lines; width of fang 8 lines ; fore and aft 6^ lines ; thickness 3 lines. Of other species of Ptychodus^ Agassiz mentions teeth of P. mammillaris, found in the excavation of the Delaware canal, and preserved in the Museum at Paris. (Pois. Fos. Ill, 151.) I have seen no specimens of that species from an American locality. The vomer is double, and usually bears teeth in this class ; the premaxillary is usually double ; Amphiuma and Spelerpes b e 1 ii are exceptions. Teeth never planted in deep alveoli. There are six orders, as follows : Caudal vertebrae and frontal bones distinct. Inferior pelvic elements not confluent. 0. o. maxillaria, prefrontalia, palatina and pterygoidea wanting ; nasalia present. Ethmoid,* two lateral pieces, each forming part of palate. Mandible toothless, condyloid. No " postorbital and su^yitemporal bones." First pair ceratohyals d^Act. Caudal vertebra and frontal bones distinct. Inferior pelvic elements not confluent. 0 o. maxillaria, prefrontalia and nasalia wanting; palatina and pterygoidea present. Ethmoid,* a vertical plate on each side the cerebral lobes. Mandible toothed, teeth pleurodout.f * Erroneously called orbitosphenoids by me, Jour. Acad. 1866 (on Anura). fThe statement made by Dr. Gray that the teeth of Necturus are canaled, as in venem- ous serpents, by a channel entering at the base and issuing below the tip, appears to the writer to be of doubtful accuracy. No other opening exists in the teet.'i oif Necturus m a- Synopsis of the Extinct BATRACHIA of North America. BY EDWARD D. COPE, A.M. BATRACHIA. Trachystomata. Peoteida. [Sept. NATURAL SCIENCES OF PHILADELPHIA. 209 Ceratoliyals, first ])air connate. No postorbital aud supertemporal bones. Urodela. Usual cranial bones present, but pterygoids reduced or wanting. No " postorbital or supertemporal bones." Caudal verJfc'ae and frontal bones distinct. Ethmoid a^W^ical plate on each side. Mandible dentigcrous, teeth pleurodont. Inferior pelvic elements horizontal, in contact, no osseous pubis ; ilium sus- pended to a sacral rib. (Mostl}' no quadratojugal.) Gymnophidia. Usual cranial bones present and distinct, including frontals and pterygoids. Caudal vertebrae distinct. No postorbital or supertemporal bones.''* Ethmoid an annulus surrounding cerebral lobes. ^ Mandible dentigerous ; teeth anchylosed by their bases.f (A quadratojugal.) Stegocephali. Usual cranial elements distinct, including frontals and pterygoids, and add- ing " postorbitals and supertemporals." Caxidal vertebrae ? Ethmoid normal. Inferior pelvic elements distinct. Mandible dentigerous ; teeth with anchylosed bases, or (A quadrato-jugal.) Anura. Frontal and parietal confluent, nasals wanting or rudimcntal ; other cranial bones present. Postorbital, supratemporal and usually nasals wanting. Ethmoid an annulus (usually complete above) surrounding cerebral lobes. Caudal vertebra represented by an elongate compound style. Inferior elements of the pelvis consolidated into a single vertical mass ; ilium attached immediately to sacral vertebra. Quadratojugal. STEGOCEPHALI. Xenorhachia. The vertebral centra not ossified ; ? the dentition pleurodont ; teeth simple ; ? no branchial hyal bones. ? Occipital condyles. c u 1 a t u s than the emargination at the base of the root occupied by the growing crown of the successional tooth, as in other Batraehia. If the structure described by Dr. Gray exists, it is in a species as yet unexamined by American zoologists. Professor Winchell, of Ann Arbor, confirms my observation. In my Synopsis of higiier groups of Batrachia (Journ. Acad. Nat. Sci. 1806), I stated that Amphiuma possesses minute scales. Gray, in 1850 (Catalogue Brit. Mus.), makes the same statement, which Dumeril (18(53. Oatal. Mus. Paris) contradicts. I must accord with Prof. Dumeril, since a subsequent examination has convinced me that they do not exist. The specimen in which the appearance of scales was presented was mislaid at the time of writing, and I find it was due to numerous free portions of the true derm, which are con- tinuous with the attached portions. * When the temporal fossa is overarched it is by expansion of the maxillary and quadrato- jugal. (Stannius says "squama temporalis."} fThe teeth of Csecilia are compressed with a trenchant posterior edge, which is crenate, after the manner of Megalosaurus, Carcharias, etc. Thus to the numerous genera of Saurians and Selachians possessing this ciiaraeter, must be added a Batrachian. 1868.] 210 PEOCEEDINGS OF THE ACADEMY OF MiCROSAURIA. Vertebral centra ossified ; no branchial hyoids ; teeth simple or with slightly inflected enamel, pleurodont. Occipital condyles. Ganocephala. Vertebral centra cartilaginous ; branchial hyoids present ; t|g||| with slightly inflected enamel, anchylosed by their bases. No ossified occqPIl condyles. Labyrinthodontia vera. Vertebral centra osseous ; no branchial hyoids ; teeth with much inflected enamel, anchylosed by their bases. Occipital condyles. Xenorhachia. This order I proposed for the reception of the genus Amphibamus Cope, in 1865. I proposed to regard, as one of its characters, the existence of opistho- coelian V9|tebr£e. Such impressions were observed in the matrix in Avhich the fossil was^reserved, as to induce a belief in the existence of such vertebrte, and the existence of these in a well ossified condition, in the apparently nearly allied genus Pelion Wyraan, strengthened such belief. There were actually, however, only osseous neural arches present, and I am now decidedly of the opinion that the vertebral centra were either cartilaginous or annuliform, as in Archegosaurus. AMPHIBAMUS Cope. Proc. Acad. Nat. Sci. Phila. 1865, 134. Amphibamus grandicbps Cope, Proc, Ac. N, Sci. Phila. 1865, p. 134. Palaeon- tology 111. State Survey. Tab. Carboniferous ; Lower Coal Measures, Morris Co., Illinois. MiCROSAURIA. This sub-order was established by Prof. Dawson, for small lizard-like verte- brates from the Coal Measures, which he thought presented points of affinity to, or should be under the Saurian reptiles, at the same time recognizing Ba- trachian characteristics. After examining the evidence brought forward by Prof. Dawson, it appears to the writer that the Saurian characteristics are analogical only, and not in- dicative of true affinity, and that these creatures form, in fact, a series closely resembling or parallel with what was probably an immature stage of the Labyrinthodontia. They are in fact Labyrinthodonts, with simple or very slightly inflected enamel of the teeth, and with the extent of the exostosis of the cranial bones much reduced. This character has been much overrated by some authors. In the Dendrerpeton obtusum Cope the grooving and pit- ting exists only on the posterior parts of the cranium, and gradually disappears anteriorly. In the Alligator mississippiensis the same is the case. The points in which they have been said to resemble the Lacertilia, are, 1st, the dermal scales ; 2d, the parietal fontanelle ; 3d, possession of ribs. All of these features belong to the Labyrinthodontia ; the Xenorhachia also had scales. On the other hand, the two occipital condyles, indicating the existence of a parasphenoid bone, distinguishes it at once from all the Allantoid verte- brata, and the form of the vertebrae is very Batrachian. In the Lacertilian families of Gecconidae and Hatteriidoe only avc have biconcave vertebraj, but the concavities are comparatively^ shallow, and the vertebrae less constricted medially than in the Microsauria. Those of the latter are much like those of Salamanders, according to Prof. Dawson's figures. The bones figured as pelvic are uulike those of any Batrachia or Lacertilia known to the writer. But until those of the Labyrinthodontia are discovered, [Sept. NATURAL SCIENCES OF PHILADELPHIA. 211 •we Ciuinot assert that they difFor from tlie latter. The long spatuliform ele- ments figured as pelvic are, perhaps, scapubv, which are of not very different tyi)e in the Tracliystomata, Proteida, and the Ganocephala. The only species included in this tribe in which inflections of the enamel have been described, is the Dendrerpeton acadianum, and here it is only at the base. It is, however, not impossible that this genus should not be asso- ciated with Hylerpeton, Oestocephalus, etc. The genera Urocordylus, Cera- terpeton, Leptcrpeton, Ophiderpeton and others recently described by Prof. Huxley, also belong here. PELION Wyman. In litteris. Raniceps Wyman, Amer. Jour. Sci. Arts, 1858, 158. Not of Cuvier (Pediculati). • Pelion lyellii Wyman. Raniceps lyellii Wyman, 1. c. This animal differs from the genus Amphiliamus in the well-ossified verte- bral axis ; no remains of a tail with elevated neural spines exist in the type specimen, and no ventral scales are seen in it. Middle Coal Measures, Jeff'erson Co., Eastern Ohio. HYLONOMUS Dawson. Hylonomus lyellii Dawson, loc. cit. viii, 167, The Joggins. Nova Scotia Coal Measures. Hylonomus aciedentatus Dawson, 1. c. viii, 258. Coal xMeasures ; with the last. Hylonomus wymanii Dawson 1. c. viii, 270. Coal Measures, Nova Scotia ; with the last. PARIOSTEGUS Cope. This genus is represented by a large part of the cranium of a Batrachian from the triassic coal measures of Chatham Co., North Carolina. If not a Ba- trachian, it could only belong to a ganoid fish, but though some of its charac- ters are somewhat ichthyic, it lacks the following important elements of the ganoid structure, i. e., free post- and suborbital bones ; postnareal cavities ; branchiostegal, and arched branchihyal bones. On the other hand it has a large preorbital, bounding the frontal and maxillary to the nares, and th6 inner border of the orbit, as in Stegocephalous Batrachia ; also a postorbital element contributing to the formation of an extended supratemporal roof. Contrary to what has been found the case in most genera of Stegocephali, the maxillary appears to extend posteriorly to a free termination, as in modern Salamanders, and the supra-temporal bone presents a very prominent, obtuse, arched margin. This margin extends from the orbits on each side, and is in- curved towards the posterior part of the cranium. There is therefore no quadratojugal piece. The maxiliary and mandibular pieces are slender, flat bones, as in Menopoma ; the form of the posterior or articular portion of the latter cannot be ascertained from the specimen. The more or less exposed part of the median region of the latter exhibits a succession of shallow transverse notches, enclosing thirteen obtuse elevations. The former resemble rudimental lateral alveolae for minute pleurodont teeth. A few other similar minute ribs, and perhaps a minute curved cone without sculpture, are the only other indications of dentition. The bones of the upper surface of the cranium are most readily interpreted by reference to those of Menopoma. A pair of narrow nasals, acuminate be- hind, penetrate between the frontals as far posteriorly as the posterior margins of the orbits. The suture between these is very distinct, and entirely straight. The preorbitals extend to above the orbits, and there appear to cease with a trans- verse suture. Between these and the nasals a broad triangular element eaters 1868.] 212 PEOCEEDINGS OF THE ACADEMY OF on each side, not attaining the probable position of the nostrils. Each is di- vided by a longitudinal groove, which is probably a suture, and which would then divide the frontals from the parietals. The frontals would then divide the parietals entirely, as they do in Menopoma for the anterior half of their length. This would give the frontals a narrow form, acuminate in front, and bounded behind by a regular coarse, zig-zag transverse suture. The cranium behind this point is rugose, and the surface not well preserved, and it can only be said that two irregular grooves converge to a point between the posterior extremities of the frontals, like the boundaries of the supraoccipital. The posterior boundary of the cranium with the condyles cannot be readily deter- mined. When the postorbital roof bone is raised up, the meeting of two gular dermal bones, as I interpret them, is seen. One of these is a plate directed backwards and*outwards, bearing minute radiating lines on its upper surface. It meets a similar flat plate directed forwards and outwards, with similar lines radiating to the circumference. The inner margins of these plates were not seen. The orbits are remarkably small, and situated probably near the middle of the longitudinal measurement of the cranium. The external nares are n-ot defined, but symmetrical depressions in the position they usually occupy in Salamanders are distinct. The general form recalls Menopoma, particularly the necessarily small eyes. A slender curved bone with a slightly dilated and truncate extremity, lying by the cranium in connection with the mandible, is like a branchihyal of that genus. Nevertheless it cannot be positively assigned to this genus, as numer- ous scales of cycloid fishes are on the same block. Pariostegus myops Cope. The surfaces of the cranial bones are little sculptured ; there are small tu- berculiform elevations on the parietals and more numerous ones on the pre- orbitals. The postorbitals show the strongest markings of elongate pits which radiate to their circumference, leaving a smooth obtuse border. The nasals present a series of small warts at a little distance on each side of their com- mon suture, and transverse to it. The surface of the maxillary is marked with longitudional grooves and shallow pits. No suture separating maxillaries and premaxillaries can be traced with cer- tainty, though the bones of the jaw are interrupted at the usual place of suture, opposite the nostril. Measurements. Lin. Length of specimen (including mandible) 18 Width between outer convexities postorbitals 17 " inner borders orbits 11 " of same without preorbitals '. 8 " of nasals at middle 2-5 " orbit 1-5 Length of frontal nasal premaxillary , 11. " supposed branchihyal 12. The name is derived from the roof-like postorbitals with free lateral margin. Locality. — Coal bed of the Keuper Triassic, Chatham Co., North Carolina. The species was discovered by Prof. Jos. Leidy, who handed it to me for de- scription. It is in the Museum of the Academy of Nat. Sciences of this city. DENDRERPETON Owen. Journal Geol. Soc. London, 1853, p. 81. In the form of the cranium this genus differs from Brachydectes and Ophi- derpeton, much as Menopoma does from Amphiuma. Two species appear to have left their remains in the coal measures at Linton, Ohio. [Sept. NATURAL SCIENCES OF rillLADELPHIA. 213 Dendrerpeton obtusum Cope. This species is known by a partially preserved cranium. The superior sur- face is exposed, the outlines of the jaws and orbits are well preserved, with the occipital condyles. The os-quadratum is directed obliquely backwards, and the angle of the mandible extends to a line a little behind that of the occipital condyles. The zygomatic arch exists in a position similar to that in which it may be seen in a few genera of Anura, as Discoglossus and Pelobates. It ex- tends downwards and forwards from the suprasquamosal to the maxillary region, but whether it is homologically squamosal or malar, the specimen can- not show. The postorbital is present as well, and with the last, and the su- pratemporal forms the bony roof of the temporal fossa. A piece which may be the pre- and postfrontals combined, borders the inner superior margin of the orbit; it widens posteriorly, where it has contact with the parietal, etc., and narrows in front. Supra-occipitals form together a broad triangle on the upper plane of the cranium, of less extent than the adjoining supratemporal. The latter elements are pitted, and towards their margins radiate grooved. These sculpturings grow less on the margins of the supratemporal, and tlie portions of the surface of the more anterior element remaining are so slightly marked as to give the impression that the sculpturing in this species is much less than in others of the genus. A few beaded ridges are all that remain on a few of the parietals and postorbitals ; the maxillaries have a slightly stronger sculpture seen in a few spots. The general form of skull is elongate behind, and much shortened in front of the orbits. The orbits are thus altogether in front of a line equally dividing the cranium transversely, while in the D. a c a d i a n u m Ow. they are in the middle of the skull. The outline of the muzzle in one species is thin, broad, rounded, as in the Menopoma allegheniensis, while in the latter it is ovate and produced. The parietal bones extend to opposite the posterior margins of the orbits, are then gradually contracted and form an acuminate prolongation on each side the wedge-shaped frontals. The prefrontals are thickened on each side the front, behind the external nares. The sutures defining the frontals ?.,nteriorly, the nasals, and the premaxillaries behind, cannot be made out. The median longitudinal suture is a marked and zigzag one, and can be seen as far post- eriorly as the anterior margin of the orbits. The external nostrils are large and opposite the inner margin of the orbit on each side. This separation of the nares is associated with a greater transverse extent of the premaxillaries than in some of the genera. These have been set with numerous teeth, judg- ing by their small impressions ; no larger ones have left traces, and no traces of any on the maxillaries. The teeth of the genera before described are all much larger relatively, indicating still further the diversity between them. A fragment of mandible remains, but without teeth or external surface. It shows a large internal canal. Measurements. Lines. Total length cranium 25-5 Width cranium 3 lines behind orbits 24 between orbits 7-5 " " nares 5- " " occipital condyles 2-2 " of supraoccipital bones 6 " of right parietal 6 Kxtent of premaxillaries 8-7 Length orbit , 6 From the Coal Measures at Linton, Columbiana Co., Ohio, (West Pennsylva- nia Basin). Discovered by Dr. John S. Newberry. Another cranium accompanies the collection which belongs to a species dis- 1868.] 214 PROCEEDINGS OF THE ACADEMY OP tinctfrom the last. The muzzle is not so broadly rounded, and the premax- illary teeth are relatively much larger. The sculpture is more delicate, with the ridges more acute. The orbits and nares are not defined. The maxillary is well preserved for a length of an inch ; its teeth are smaller than the premax- illaries ; I count four in a line ; crown simple conic. External surface of maxillary not very strongly sculptured. This species cannot be referred to its genus without further material. I therefore do not name it, hoping to avoid the unworthy practice of some, who give prospective names — to be applied to other peoples future discoveries, and the like, Dendrerpeton acadianum Owen, Quart. Journ. Geol. Soc, x, 1853, 81. Daw- son loc. cit. Coal Measures : Joggins of Nova Scotia. Dendrerpeton owenii Dawson, Canadian Naturalist and Geologist, viii,161. Coal Measures : as the last. HYLERPETON Owen. Hylerpeton dawsoni Owen, Journ. Geol. Soc. Lond., 1862, 241. Dawson, Canadian Naturalist and Geologist, viii, 272. Carboniferous Coal Measures. The Joggins, Nova Scotia. BRACHYDECTES Cope. This genus is indicated by two rami of a mandible, and a portion of a pre- maxillary only. These, when compared with those of Oestocephalus and Dendrer- peton from the same locality, and with others described by authors, are so much stouter, i. e. shorter and more elevated, that they evidently pertained to a genus not hitherto known. The genus further differs from Oestocephalus in having the teeth of equal size to the posterior parts of the series, that is to the base of the elevated coronoid process. The teeth are elongate cylindric cones, with their acute tips turned a little posteriorly. The fractured ones display a large pulp cavity. The three premaxillaries preserved are similar but without curvature of the tips. They do not exhibit striae or any other sculpture. So far as the remains known go, the genus is nearer Hylerpeton than any other. The latter does not give any indication of the very elevated coronoid process of Brachydectes, though the external portion of the dental bone in that region being lost, little can be said about it. Prof. Owen's plate indicates a ramus whose depth at the last tooth enters 8J times the total length. In our species this depth enters about five times. There are at least nine teeth in the Nova tScotian species ; seven in the present one. Brachydectes newberuyi Cope. This species is represented by one nearly perfect ramus mandibuli, one den- tary bone, and one premaxillary probably not complete. The dentary bone appears to have been attached by suture to the articular and angular, as its free margin has very much the outline of that suture in Amphiuma and lizards. 1 he coronoid process would also seem to be a part of the same bone as in Amphiuma and Menopoma, and not composed of a coronoid bone as in lizards. It rises immediately behind the last tooth, and displays no suture. The lower portion of the dentary is prolonged into an acute angle. This is separated by a deep and wide concavity from the superior posterior prolonga- tion, which is obtuse and rises at once into the coronoid process. Teeth on this dentary seven ; the same number is on the preserved ramus ; this number I suspect to be complete or nearly so. The teeth terminate at the obvious termination of each ramus, which is it is true slightly obscured. The teeth are the longest of the Microsauria in relation to the depth of the ramus, equal- ling the largest in Ophiderpeton. They are doubtless exposed, as are some of [Sept. NATURAL SCIENCES OF PHILADELPHIA. 215 those of the last named genus, by the splitting away of the outer parapet of the deutary bone. As no traces of alveoli have been thus rendered visible, I suspect the dentition to have been pleurodont, as in existing Batracliia. No external surface of the mandible remains, but there arc no impressions of sculpture on the matrix. A little external face of the premaxillary displays none. Measurements. Lines. Preserved length of ramus (imperfect) 11. Depth at Lxst tooth 2. Length of exposed tooth 1-7 " deutary 7-5 Depth at corouoid 3-5 at first tooih 1-3 SAUROPLEURA Cope. Neural and haemal elements of the caudal vertebrae elongate, distally dilated and grooved, attached by contracted bases. Ventral aspect defended by a series of oblique dermal ribs on each side, which meet anteriorly on the median line. Limbs distinctly developed. Ribs long, well developed. Scales none. No dermal bones have been discovered, nor are any portions of the cranium known. This genus is allied to the Urocordylus of Huxley, recently discovered in the coal measuresfn Leinster, Ireland. It differs only in the presence of elongate lizard-like ribs (whence the name), and in the absence of " oat-shaped scales" of the upper surfaces. It is a matter of much interest in American palfeontology that this remark- able type should be found to occur in our coal measures. It was first announced by Dr. Newberry at the meeting of the American Association for the advance- ment of Science for 1867 (see Proceedings, p. 144), as a supposed Urocordylus, occurring with Ophiderpeton. He mentioned at the same time the discovery of the ganoid Diuichthys Newb. The forms discovered by Dr. Newberry have an interesting relation to those of Ireland, such as types of the present period frequent!}^ present. The genera may be thus parallelized ; where no representation exists, we niay look forward to a future discovery to supply the present want : Ceraterpeton Huxl., represented by Urocordylus Huxl., " " Sauropleura Cope. Lepterpeton Huxl., " " Dolichosoma Huxl., " " Molgophis Cope. Ophiderpeton Huxl., " " Oestocephalus Cope. Brachydectes Cope. Herpetocephalus Huxl., ? ? Dendrerpeton Ow. Of the American genera, Sauropleura and Oestocephalus exhibit the peculiar ventral dermal armature of Urocordylus and Ophiderpeton, while Molgophis does not possess it, nor Dendrerpeton, if our species truly belongs there. The museum of Columbia College, New York, contains portious of two spe- cies of Sauropleura, but both unfortunately represented by portions only of the vertebral column. These are, though closely resembling the species dL^scribed by Prof. Huxley, sufficient to demonstrate marked generic distinction. This is further established by the remains of the trunk of a third, and larger species, whose relationships can be shown to lie within this genus. This individual has been spread over a surface of the coal slate, exhibiting ventral armature, dorsal region with ribs, and anterior and posterior limbs. Of skull and caudal vertebrae nothing remains. The dermal riblets are arranged as in Urocordylus, i. c, in parallel lines di- rected obliquely forwards and continuous on the median line, forming there a 1868.] 216 PROCEEDINGS OF THE ACADEMY OF chevron, directed forwards. The striae are not so closely placed as in U. p e c- t i n a t u s , but are separated by grooves wider than themselves. The humerus, ulna, and radius are rather stout, and of a size relative to the l>ody, as in common types of existing Sauria ; the ulna and radius separate. There is no carpus, but five well-developed digits have phalanges in the follow- ing numbers, commencing on the inside : 3 — 4 — 5 — 6 — 5. The last phalange of the second is obscured, and it is not positive that the number is as given ; it is more probable than that it should have been 3. The outer toe has been more slender than the others ; the distal phalanges of all the toes are short conic, as in Salamanders. Thus this form differs much from Amphibamus, where the numbers are 3 — 3 — 4 — 5 — 4, showing a lower developement of limbs. The ribs are long and curved, as in Eeptiles, and judging by their distances the vertebra are short ; the latter are not well defined, but there is no indica- tion of prominent spines of any kind. The pelvic bones and portions of those of the hind limbs are present, but so obscure and confused as not to be made out. Enough remains to show that the hind limbs are considerably larger than the anterior. Sauropleura digitata Cope. This species had a length of body about equal to that of a fully-grown Cha- maE'leo vulgaris of the largest size, or of a half-grown Menopoma. Thirteen ribs on one, and several on the other side, are preserved ; where they terminate, probably at the pelvic region, some small or rudimental ribs p^ject from the two or three first caudals. Three ribs and their interspaces extend over five lines. The humerus is broken, but its length can be clearly made out to be seven lines ; it has no condyle, and is dilated at both extremities. The ulna and radius are distinct, truncate, hollow, and dilated at the ends. Length of ulna 5-1 lines, distal width 1-8 lines. Carpus not ossified. The fourth toe is considerably longer than the others, the fifth is next, and reaches the basal third of the antepenult phalange of the fourth ; the third is very little shorter ; the first is not quite so long as the first two of the third. The bones of the hind limb are not readily distinguished. They are evidently much longer and larger than the anterior; no part of a foot is preserved. This form is probably allied to Urocordylus. It has relatively much stronger ribs in relation to the vertebne than we have seen in that genus, and there is no evidence of the existence of the peculiarly formed spines of the vertebrae characterizing the latter. The limbs are relatively much stronger than in Ophiderpeton, and it lacks the peculiar dermal armature of that genus. vSauropledra pectinata Cope. This species is represented by portions of the vertebral columns of four indi- viduals. In two of these, vertebral centra are discoverable ; in one quite defi- nitely. They are slightly constricted medially, and without ridge or process. The neural and haemal spines of superior and inferior lines are similar, and in the specimens undistinguishable. The dilated portions form nearly equi- lateral triangles, which stand on moderately short pedicels. They are weakly ridged, and each ridge is prolonged into a narrow acute tooth, beyond the margin, of which eleven ma}' be counted on one of the best preserved. The longitudinal striae are terminated near the pedicel by two others, which cross obliquely from each side, and, meeting, present an appearance similar to an overlapping of each margin. The edges of the spines form a continuous line. As in the other species, there are no indications of other processes, nor of dermal scales. The smallest of the specimens shows that in front of the region furnished with the peculiar spines described, the body is furnished with a mass of bristle- or hair-like scales. The grooved neural sjjines are slightly displaced anteri- orly, and the bristle-like mass looks like a continuation of their striae, and it is not easy to find any line of demarkation between them. The serrate spines are [Sept. NATURAL SCIENCES OF PHILADELPHIA. 217 lontinued further forward on one side than the other. These linear scales were arranged as in other genera, in lines which converge forwards to the me- dian line. They are somewhat obscured in the specimen, but it can be de- termined that they are continuous on the median line. Whether this is the posterior or anterior portion of the body cannot positively be determined from the specimen ; it is, however, most likely the posterior, for near the posterior portion of the striate surface a weak pair of limbs is given off on each side. On the right, a moderately stout ? femur is followed by a broken tibia and fibula, and by five slender, closely appressed metatarsals. The last arc about two- fifths as long as the space between them and the femur; beyond them a few- slender phalanges are moderately distinctly defined. The tibia is more distinct on the left, but no tarsus or phalanges ; some of the metatarsals are preserved here also. Length of limb to end of metatarsals equal to five vertebra; in juxta- position,~measured along the edges of the neural spines. The limb has been slender, especially the hand. The above specimen enables me to assign, as the ventral armature of this species, a closely packed series of V-shaped grooves, which lie in connection with an obscure vertebral column, on the block containing one of the typical specimens of this species. They are not continuous with any of the series ex- hibited on other parts of the block ; some of these at least are the doublings of the slender animal, and this ventral portion has been displaced. The grooves are perhaps the impressions of h;x;mapophysial rods, vastly more numerous, however, than the number of vertebra ; perhaps they are rather the dermal armature. Huxley figures a portion of this as on the block with UrocordyluS' wandesfordii, but does not refer it to its precise relation to the animal. A few well-developed ribs are preserved with this portion, — the only ones I can refer to this species. The vertebra^ are partly enclosed in matrix, partly impressions. The neural spines, though expanded antero-posteriorly, are less, elevated than in the caudal region, and have left no traces of their character- istic ribs or serration. The number of spines in the type specimens is six in a half-inch ; in the smallest, just described, ten in the same distance. The height of the spine in the former is 1-15 lines. As the characters of this species are most determinable, 1 regard it as thft type of the genus Sauropleura. Sauropleura remex Cope. This species is larger than the S. pectinata, and about equal to the Uro- cordylus wandesfordii Huxl. The caudal spines differ from both in the greater attenuation of the haemal series, and the presence of a basal lamina on the neural. It is represented by a portion of the vertebral column three inches in length. In this space may be counted twenty-four vertebrae Such of the latter whose outlines are visible display centra, characteristic of the genus; their terminal concavities conic, with apices meeting medially; zygapophyses present; and their length a little greater than their depth. The characteristics of the species are the remarkable length and slendernesa of the fan-shaped neural and haemal spines, and the absence of an acute serra- tion on their margins. In this species the spines have a larainiform expansion at the base in their planes. In the other species here described these spines are not only relatively broader and more fan-shaped, but they are acutely ser- rate on the margin and constricted at the base. In S. remex the dilated neural spines are a little more than three times as long as they are distally wide, while the haemal spines are a little narrower. The neural spines stand about the middle of the centrum. The basal half is furnished with an anterior ala, which leaves the anterior margin rather ab- ruptly and extends to the next spine in advance. It returns gradually to the centrum and is separated from the articular face of the latter by a notch. A 1868.] . 15 218 PROCEEDINGS OF THE ACADEMY OP similar ala exists on the posterior margin of the neural spine, which extends for a shorter distance above the base, and is narrower than the anterior. Each spine presents a median groove on its surface, which extends halfway to the base or further ; on each side of this are some three other grooves, which extend but a short distance ; surface otherwise smooth. The ends of the grooves slightly notch the truncated end of the spine. The hremal spines are on the posterior portions of the centra, and in con- tact with the anterior part of the basis of those succeeding. They are without the dilatations of the neural spines, and are directed rather more obliquely backwards. They are similarly grooved, though without that so distinctly median, seen in the neural series. Both neural and haemal spines become larger towards the posterior part of the vertebral column. There appear to be no zygapophyses nor diapophyses, nor rudiments of ribs. The centra are rather stout, and somewhat constricted medially. There are no traces of dermal armature of any kind. Measurements. Lines. Length of a posterior centrum 1-2 Depth " " " 1- Length neural spine of adjoining vertebra 4-4 Basal width 1-4 Median width -9 Distal width 1-1 Length of a more anterior neural spine 4-3 Distal width " " " 1-5 *' anterior haemal spine 4 <• width " c 1-4 From the Coal Measures, the Western Pennsylvania and Ohio Bituminous Basin, at Linton, Columbiana Co., Ohio, near the Ohio River. Prof. J. S. Newberry. (ESTOCEPHALUS Cope. This genus is known from a single species as yet. As before remarked, it represents in many respects the Ophiderpeton of Huxley, and has been alluded to by Dr. Newberry as the same. It, however, differs markedly in the narrow lanceolate form of the head, with probable accompanying peculiarities of de- tail, and in the presence of limbs, which have not been found in the Irish genus. The form of the head is somewhat nearer that of Lepterpeton HuxL, but the limbs of the American genus have as yet been seen as one pair only, and very small, while in Lepterpeton there are two pairs, which are large. The general form of the body of Oestocephalus is more snake-like. In more detail, we have an elongate lanceolate head with little or no sculp- turing of the external surface of the bones. The angles of the mandibles are much prolonged backwards as in Archegosaurus and frogs, and the well de- veloped ribs commence but a short distance behind the head. The vertebrae are slender, and furnished with well developed diapophyses. A pair of s;^- metrical bones, whose impressions are seen posterior to the occipital region, look like ceratohyals or small scapulae, and one of them is continuous with a second piece, which occupies the place of a humerus. A third piece follows, which is probably ulna ; no radius or manus is preserved. This then is a rudimental fore limb, situate very close to the head. The skin has been occu- pied by a great number of closely packed, curved, spine-shaped scales. They have occupied the ventral integument passing from the median line of the belly outwards and posteriorly, having acute tips which may have penetrated the skin on each side ; whether-such tegumentary spines protected the back can- not now be determined. [Sept. NATURAL SCIENCES OF PHILADELPHIA. 219 OEKTOnEPHALUS AMPHIUMINUS Cope. This species is representcil by the imperfect crania and anterior portions of the bodies of two individuals. They indicate an animal of tht^average size of the Amphiuma means. The extremities of the vertebrw are deeply concave, but the centra are so long as to prevent the concavities entering more than one-fifth of the latter, each. The diapophyses are behind the middle, and are broad, curved back- wards, and acuminate, as in Amphiuma. The centra have a prominent median line below, with a longitudinal concavity on each side. Five of them a little exceed an inch in length. Neural spines are nowhere visible. The humerus is longer than the scapula and is considerably dilated distally ; the scapula slightly dilated at its superior extremity. The dermal armature commences immediately behind the head, and forms a band of 14 lines in width ; measuring across the spine-like scales, in a width of a line, four cylinders may be counted. The external portions are carried backwards, the interior nearly straight, those of the anterior more delicate than the posterior. The head is wedge-shaped with regularly acuminate sides. The top of the cranium is somewhat broken in the specimen ; the portions preserved are smooth, and the longitudinal suture is distinct for a considerable distance. The angle of the mandible is produced considerably behind the occiput and is enlarged and rounded. The end of the muzzle is broken away, and the region of the orbits so fractured as to render their precise location uncertain. The superficial layer of the cranial bones is nowhere clearly visible, so tlmt it can- not be ascertained whether it is sculptured or not. The quadrate Bone pro- jects well posteriorly. Some fragments indicate small cylindric teeth, as in Amphibamus, butthej'^are not characteristic. Measurements. Lines. Length cranium without muzzle ih-Z Width " posteriorly 11-5 Length scapula 2-1 " humerus 2-5 " of sixth vertebra from skull 3 Extent diapophyses 3-5 Width centrum 1-5 This species was discovered by Prof. Jno. S. Newberry, at Linton, Eastern Ohio, in the slate of the coal measures. Mus. Columbia College. The characters of the genus are further shown by a part of another indi- vidual in the same coal slate matrix. The cranium and anterior portion of the vertebral column only are preserved, the latter so much injured as to render the vertebral characters very obscure. As in the other, the bristle-like scales extend along the dorsal region to near the cranium. The anterior two-fifths of the ventral side shows a large number of small oval scale-like bodies, which belxjnged undoubted]}' to the animal and were probably dermal scales. They are, however, neither regular in form or position. Close behind the head two or three long bones of the fore limbs have been exposed. They are slender, and similar to those of the last specimen. The cranium, though without the muzzle, shows its long wedge- shape. The maxillary bone cannot be distinguished, nor can the orbits be made out; one ramus mandibuli is pretty well preserved ; it shows no coronoid pi ocess. Twenty-one teeth may be counted on a portion a little more than one-third its length. The anterior eleven of these are abruptly longer and stouter than the others. They are,' exce^ a few most anterior, in pairs, i. e. with a slight va- cancy between every two. The larger ones were broken at the bases, exhibit a moderate pulp cavity ; the smaller, a large one extending to near the tip. Several, though not all of the larger teeth, displa}^ a shallow groove on the ex- 1868.] 220 PROCEEDINGS OF THE ACADEMY OP ternal face to near the tip, which is probably owing to pressure, and a partial crushing. The points of the larger teeth are rather abruptly acute, and turned abruptly backwards. A portion of their increased length (-25) is to be attrib- uted to the splftting off of the external dentary margin, and the exposure of the roots. No alveoli are shown, and the dentition is probably pleurodont, 'with anchylosis of expanded base as in true Labyrinthodonts. MOLGOPHIS Cope. This genus is established on remains represented by three specimens, which are two series of dorsal vertebrae with ribs, and a series of caudals. One of the dorsal series embraces sixteen vertebrae, the other fourteen ; the caudal series, twenty-two. From its serpentine form this genus may be compared with the Dolichosoma of Huxley, though a close relation does not exist between them. In the Irish genus the series of caudal vertebrae is quite short, and the ribs are short and but little curved. In Molgophis the tail has been like that of an elongate ser- pent, and the ribs are as well developed as those of many reptiles. Though no limbs or arches can be certainly found, a rather quadrate, paral- lelogrammic piece, about as long as the diameter of a vertebra, may be a femur. This is, however, very doubtful. The characters of the genus are, a long serpentine body, without dermal armature, so far as discoverable ; the vertebrae large and broad, with very prominent zygapophyses and moderate neural spine, those of the caudals with- out narrowed bases (and grooved or serrate edges, most probably). ? Limbs and cranium unknown. This genus differs from Urocordylus in its caudal vertebrae, and from Ophi- derpeton in its dorsals; the latter, in their zygapophyses projecting laterally, resemble those of Amphiuma. It differs from Sauropleura in the absence of ventral dermal bands and in the longer body, without indication of limbs. The size of the vertebrae would indicate a body of the size of a rattlesnake, (C. h o r r i d a), and therefore too large for the species named Brachydectes n e w- b e r r y i. The ribs are long, and though the head is not bifurcate, there appear to be both tubercle and head on the dilated extremity. They show themselves, where crushed, to have had a large median vacuity. Molgophis macrurus Cope. The neural arches, viewed from above, have a posterior V- shaped outline, from the fact that the broad zygapophyses meet on the median line, and spread out distally over the broad anterior ones adjoining. The latter appear to be somewhat concave, and to border the former exteriorly as well as inferiorly. The base of the neural spine extends to the posterior emargination, but not quite to the anterior. The breadth of the dorsal vertebra above is equal from the emargination behind to the anterior margin of the anterior zygapophyses. The caudal series must have been very long, as there is very little diminu- tion in the size of the vertebrae throughout the series preserved. They present much the same form as the dorsals, but are more contracted medially, and the zygapophyses have a more transverse direction. There may indeed be a dia- pophysial element beneath these, but the two cannot be distinguished if so. They are connected by longitudinal impressions indicating the existence of the tendinous bands in the longitudinal muscles seen in Amphiuma, or the osseous spicules in the same situation in birds. The neural spines, indicated by their narrow bases, occupied the lengths of the neural arch, and remind one of Amphiuma. # The ribs are long for a Batrachian, but shorter than in a reptile. They are well curved, chiefly near the proximal extremity. The longest I can find, measured by a chord, equals two vertebrae and two-fifths. Three vertebrae [Sept. NATURAL SCIENCES OF PHILADELPHIA. 221 measured along the median line above equal eleven lines ; one of these is 3-6 lines in width above ; width of a (?) posterior caudal 3 1. This animal has been, like Amphiuma, a snake-like Balrachian, but probably of even more elongate form. How near its affinities to this genus may be, cannot be ascertained, owing to want of important parts of the skeleton, but it differs in the important feature of the large, well developed riba. LABYRINTHODONTIA. DITCYOCEPHALUS Leidy. DiCTYocEPHALUS ELEGANS Lcidv, Proc. Acad. Nat. Sci., 1856, 256. Emmons Geology North Amer,, p. 59.' Tab. 31. Triassic Coal Measures, Chatham Co., N. Carolina. CENTEMODON Lea. Cbntemodon sulcatus Lea, Proc. Acad. Nat. Sci., Phila., 1856, 78. Triassic Shales near Phoenixvilie, Chester Co., Penn. BAPHETES Owen. Baphetes planiceps Owen, Quart. Journ. Geol. Soc. Lond. x*, 1853, (xi, notes). Carboniferous Coal Measures of the Joggins, Nova Scotia. EUPELOR Cope. Gen. nov. Char. Teeth subcylindric, with large pulp cavity at the basis only ; external surface without grooves ; dentine divided by numerous flat vertical laminse of a dense substance, probably enamel, which radiate from very near the pulp cavity to the external enamel layer. The species on which this genus depends was originally described by the writer as a Mastodonsaurus. The latter genus, however, exhibits external grooves where the inflections of enamel enter and separate the dentine. These inflections, as is well known from the figures and descriptions of Professor Owen, are more or less convoluted, some of them very highly so. The laminae of the teeth of the Eupelor cannot be looked upon as inflections of enamel, but rather as branches. They are exceedingly thin, and our sections do not demonstrate them to be double. If they are double they are very much more attenuated than the external enamel stratum. They may be distinguished in a section of the wall of the pulp cavity at the base of the root as well as else- where. EuPELOR DCRUS Cope, Mastodonsaurus durus Cope. Proceed. Acad. Nat. Sci., Phila., 1866, 249. From the Triassic Red Sand Stone near Phoenixvilie, Chester Co., Penn. On AGAFHELTJS, a genus of toothless Cetacea. BY EDW. D. COPE. During the autumn of 1866 a whale was cast ashore on the Long Beach, Ocean Co., N. J., opposite Westocunk, on the other side of Little Egg Harbor, near the residence of Wm. A. Crane. A recent visit to the spot fur- nished me with the means of determining the species to which this monster of the deep belonged, although not with the completeness desirable, as the tide had a short time previously taken off" the most bulky part of the carcass. Thus the cranium, cervical and dorsal vertebrae, with the first ribs, the most import- ant portions for its identification, were lost. There were preserved, however, the mandibular arch, ear-bone, one scapula and both fins, numerous ribs, many 1868.] 222 PROCEEDINGS OF THE ACADEMY OF lumbar and caudal vertebra?, with the baleen from one side of the maxilla. These portions, with a few prominent points dependent on the observations of Wm. A. Crane, serve to indicate a species not only new to our fauna, but new to modern science. The evidence of my informant, as that of an old and ex- perienced coaster and waterman, and one familiar with the appearance of our cetaceans, confirmed by his sons and by the specimens preserved, so far as they Avent, I consider reliable. That the species should have remained unde- scribed until the present time will not appear surprising to those who read carefully Gray's recently issued " Catalogue of Cetaceans," or Eschricht and Reinhardt " Om Nordhvalen," Copenhagen, 1861. The scapula preserved is low and elongate, with well-developed acromion and coracoid process. It is evidently of the type of Bala^noptera and Physalus ; the ulna and radius relatively less elongate than in Sibbaldius 1 a t i c e p s and b o r e a 1 i s , being 1 -5 as long as the humerus, thus resembling Physalus. The four fingers, with the second much the longest, form a fin of the type of these genera. The ear-bone is much more compressed than in Physalus an- t i q u o r u m or Sibbaldius 1 a t i c e p s . The mandibular ramus is rather massive, moderately curved, and with a more elevated coronoid process than in any Avhale that I have seen. The greatest peculiarity is in the form of the lum- bar and anterior caudal vertebrae ; they are of a much more elongate form than any I have seen ox found figured, excepting those of the Bahenoptera r o s - trata (as figured by Gaimard in Voyage de la Recherche), which, however, are relatively shorter. Those of the present species are of greater length than transverse diameter, the lumbars most elongate ; all furnished with an acute hypapophysial keel and concave sides, and entirely transverse diapophyses. This peculiarity is consistent with the account of my informant, who stated the animal to have been of an unusually elongate and slender form. When it came ashore it had perhaps been dead ten days; the flukes and muscular region as far as the third caudal vertebra had been devoured, probably by sharks and killers, and the abdominal region much lacerated ; the edge of a fin preserved was slit by the teeth of some carnivorous enemy. The measurement from the end of the muzzle to the end of the third caudal was 35 feet, which may be re- duced to 33 feet axial. Up to this point the dorsal line was, according to my informant, entirely smooth, without knob or fin, or scar of one ; hence I sup- pose the fin (if present) to have been situated as in Sibbaldius, &c., at the pos- terior fourth of the length, and not as in Baltenoptera, on the posterior third. It may then be safely assumed, bearing in mind the form of the vertebrje, that ten feet of the whale's length had been removed, making in all 43 feet. That the species attains over 50 feet is probable, as the present individual was quite young, the epiphyses separating from the vertebra? with the greatest ease. The slender form of the animal is corroborated by the slenderness and slight curvature of the ribs, one attached beneath the scapula, probably the second, being nar- rower than the corresponding ones in Sibbaldius. 1 therefore think it most probable that in this form the anterior ribs are single-headed. The baleen is peculiar; throughout the length of the maxillary bone it no- where exceeded one foot in length, and the Avidth of the band, or length of the base of each plate, four inches. It is of a crearay-Avhite ; the fringe very coarse, white, and resembling hogs' bristles. The proportions in most respects present a contrast to those of Physalus species, and Sibbaldius species. While the cranium and fin of the Physalus antiquorum are of about equal length, the latter is four-sevenths the for- mer in the present species. In the Physalus the cranium enters the length 4-7 times ; in Sibbaldius 1 a t i c e p s 4 06, and in the present species 6-6 times ; in Biilaenoptera ro strata 4 5 limes. In general features this Cetacean seems to be an intermediate form of the toothless whales ; and an additional feature, which depends, on the observation of my friend W. Crane, and in which 1 cannot conceive it possible that he should be mistaken, indicates still more conclusively that it pertains to a genu.? [Sept. NATURAL SCIENCES OF PHILADELPHIA. 223 not before characterized. The whale was first driven on shore on its back, and the guhir and thoiacic regions were seen to be entirely without ridges or plicae of any kind, but as smooth as any other part of the body, or as the throat of a right whale, Baluena c i s a r c t i c a Cope, which is not uncommon on the same coast. ^ This my informant told me was the species known among the whalers as the "Scrag AVhale." Though this name is indefinite when applied by whalers of different nationalities, it is probably used with accuracy by those accustomed to any particular region. At any rate I have little doubt that this is the spe- cies called by the same name by Dudley, who in 1725 wrote an account of the whales known by the whalers of the coasts of New England. He says it is near the right whale (B. c i s a r c t i c a) in figure also ; " is near akin to the fin-back, but instead of a fin upon its back, the ridge of the after part of its back is scragged with half a dozen knobs or knuckles. He is nearest the right whale in figure and quantity of oil. His bone is white, but won't split." This is published, with an account of the other species known, in the 33d volume of the Philosophical Transactions. He mentions particularly the fin-back and hump-back whales, describing the deep folds of the chin, throat and sides of those genera. There can be little doubt that his " scrag whale " had a smooth throat like the Bakenre, and not a plaited one like the Balaenopteras and their allies. By the preceding account it has been shown that the species has but four slender fingers at the carpus ; hence it is obviously the type of genus in- termediate between Balaena and Megaptera, not hitherto recognized, — furnished, however, with the scapula of Balaenoptera. Captain Atwood, a resident of a part of the peninsula of Cape Cod, Mass., who is a good observer of the life of the ocean, thus writes of the scrag whale in J. A. Allen's Catalogue of the Mammals of Massachusetts, in the Proc. Bos- ton N. H. Soc. for 186« : *' Scragg. — A species of whale known by this name, and nearly allied, if not identical with the right whale, is sometimes taken here. It is the opinion of many of our whalemen that they are not a distinct species, but are the young right whale that lost its mother while very young, and has grown up without parental care, which has caused a slight modification. The most prominent feature is on its dorsal ridge ; near the tail there are a number of small pro- jections or bunches, having some resemblance to the teeth of a saw. It has no dorsal fin or hump on its back." Additional evidence of the existence of this genus has been furnished by the Smithsonian Institution. In accordance with recommendations and directions furnished by the writer, Wm. H. Dall, the enterprising director of the West Coast Scientific Exploring Expedition, originally commanded by Dr. Kennicott, sent to the Institution drawings and descriptive notes of the grey whale of the coasts of Upper and Lower California. The writer has also examined an al- most complete set of whalebone, with some other portions of the same species, in the museum of the Essex Institute, at Salem, Mass. The baleen is similar in character to that of the present species, but presents specific differences. The notes of Capt. Dall indicate a long-finned, smooth-throated whale, with a flat-pointed head like a fin-back, and no dorsal fin, but a series of knobs on the posterior region of the back. That it in all respects conforms to the generic type of the Atlantic species, can be determined from the description which follows. The Atlantic species was named from Dudley's description by the com- piler, Erxleben, without his adding to our knowledge of it, Balffina g i b b o s a. I will follow Dr. Gray in adopting this name. The latter author, in his excel- lent Catalogue of Seals and Whales in Brit. Mus., refers it, on the basis of the same description, to Balaena, with doubt. Genus AGAPHELUS Cope. Fingers four, elongate. Cervical vertebrae ? Lumbar and anterior caudal 1868.] V 224 PROCEEDINGS OF THE ACADEMY OF vertebrae longer than their greatest diameter. Dorsal fin wanting. G-ular and pectoral region without folds. Scapula with well developed acromion and coracoid. Baleen narrow, short. Agaphelus gibbosus Cope. # Scrag Whale, Dudley, Philos. Trans, xxxiii., 250, and of The Whalers. Balctna ffibbosa Erx\ehen, Systema Mammaliura 610 (from Dudley), and after him of Gmelin, Bonnatere, Lacep^de, Virey, Gerard, Desmarest & Fischer. Gray, Catal. Brit. Mus. 1850, p. 18, and 18(36, p. 90. Agaphelus gibbosus Cope, Proc. Ac, N. Sci. Phila., 1868, 159. Balsenoptera rostrata Cope, Proc. Ac. Nat. Sci., Phila., 1867, 147. Ft. In. Total length (estimated) of young 43 Length to third caudal vertebra 33 Length of cranium (estimated) , 6 10 " mandibular ramus (in curve)......... 6 " pectoral limb 4 Width of " " 15 Length of humerus 11-5 " radius and ulna 17 Posterior margin of scapula. 14 Length of coracoid from glenoid cavity 3-3 " glenoid cavity , 6-3 Mandible, length from condyle to coronoid 13-5 " depth at coronoid 8 5 " " 2-5 feet from coronoid 4-6 The form of the mandibular ramus is peculiar, and more like that of the Balsenoptera rostrata than any other. It is triangular in section, having an inferior angulated ridge, and a broad, slightly convex, superior face, instead of their usual ridge. Such a ridge leaves the coronoid process, but soon turns inwards to form the inner outline. Width of the superior face 3-5 inches. The coronoid process is quite elevated, and turned outwards. In the fresh animal the lower lip included the upper all round. The laminae of whalebone are placed on a base having a sigmoid flexure. Greatest depth of the gum 1 in. 3 lines Within each principal lamina are two supplementary lamina, the intermediate being the narrower, the inner triangular, its intermediate bristles arising from the gum. The bristles of the supplementary plates are longer and finer than those of the outer ; in the latter, three series of bristles are en- closed between very thin enamel plates. All the laminae are thin, five in an inch, and split transversely straight; white cream-colored, with a purplish shade near the centre of the base. The ulna is slender, but furnished with a prominent round- ed and flattened olecranon, which is prolonged into a thin cartilaginous plate, formed like the diapophysis of a vertebra, and in the plane of the ulna ; this structure appears to have been ossified in the Sibbaldius bo real is Fisch., as figured by Dubar In the Agaphelus gibbosus it occasions an abrupt angulation near the basal third of the inner margin of the fin. In the scapula, the coracoid is in its plane, but the larger acromion diverges outwards. The anterior caudal vertebrae are more elongate than the lumbosacral, less depressed, and with the centra in every way larger. All are sharply keeled on the median line below, with a concave face between the keel and the base of the diapophysis. The caudal and lumbosacral diapophyses are obspatulate, the anterior becoming narrower. The neural spines of the lumbar vertebrae are much elevated, concave above both before and behind, the zygapophysis measuring a point considerably below the middle. In. Third (?) caudal (not perforate) length centrum 7-3 depth 6 width ,. 6-5 [Sept. NATURAL SCTENCES OF PHILADELPHIA. 225 height neural spine zygapopliysis length (iiapophysis (from front base) Lumbosacral 1 ; length centrum ^'^ depth 53 width 6-3 height neural canal 2-5 width 1-5 height neural spine 15-3 zygapoi)hyses 5-5 length diapophysis V-5 greatest width do 5-3 Lumbosacral 2 (more anterior) length centrum 6-3 Lumbosacral 3 (anterior) height centrum 4-3 width " 6 length di, ectrimerous ; ^, ecdimerous ; .4c, actinotrichia; ^r, fin-rays; jSa, baseost ; ^z, axonost; 7c, intercentrum. ports certain " fin-rays," or it may include only hair-like rods or " actinotrichia." The latter characterize primitive types. In the case of the median fins it is characteristic of primitive fishes to have the baseosts and axonosts articulated with the neural spines of the vertebrje. In modern fishes the axonosts are not so articulated, and do not correspond with them, while the baseosts are rudimental or wanting. (Fig. 3.) Remarkable modifications take place in the successive evolution of the caudal fin. The primitive condition of the vertebral column in the tail of a fish is straight, with spines equally divergent above and below. This is tlie diphycercal tail (Fig. 4, .4). It persists in some modern fishes ; 12 e. g., the eels. In the next stage the axis is turned up at the end, and the spines of the inferior side are spread out fan-like, while those of the superior side are crowded together. The dermal part of the fin may develop an angle on the inferior spines (hsemal spines), the result being a two-lobed fin, in which the upper lobe is much larger than the lower. This is the heterocercal tail (Fig. 4, B, C, and D). In later forms the hiemal spines grow larger and tiie neural spines smaller, and the axis grows shorter by the abortion of the terminal centra. As a result, the lobes of the caudal fin become equal. In the most fully developed types the haemal spines fuse into a fan-shaped bone which supports the fin. This is the homocercal caudal fin (Fig. 3, E). In its growth it generally passes through the diphycercal and heterocercal stages before reaching mnturity. Fig. 4 (over). Fig. 4.— Tails of fishes ; A, diphycercal tail of Polypterus ; B, diphyheterocercal tail of Coccos- teus ; C, caudal fin of a shark (Centrina); I), end of vertebral column of trout {Salmo fario) ; E,. end of vertebral column of barbel (Barbus fluviatilis) ; ep, Epural ; hy, hypural bones. Subclass L— HOLOCEPHALI. « But one order of this subclass is known : A single external branchial fissure ; actinotrichia present ; basilars, axonosts, and neural spines articulating with 14 each other; pectoral fin pluribasal, with three axonosts and numerous basilars ; veutrals with elongate axonosts and basilars ; CJiimceroidei, In all known Chimj^roidei the teeth are large paired bodies, one pair in the lower and two in thcvupper jaws, composed of coarse vascular dentine. These contain columns varying in number and shape, consisting of coarse tubes wdth calcareous walls which terminate on the masticating surfaces. Notochord persistent, the ossifications, when present, consisting of delicate rings, more numerous than the neural arches. In the existing forms the males have claspers like those of sharks. Chimseroidei appear in the Corniferous Limestone in Ohio, Wisconsin, and Iowa, and in the Eifel Limestone in Rhine Prussia. They are found in Triassic and Jurassic beds in Europe, and abound in the Cretaceous of Europe and Nprth America, and New Zealand They extend through the Cenozoic beds of Europe and North America, and a few species still exist. The known families are the Ptyctodontidse (of doubtful reference), the Squalorajidie, Myriacanthidie, and Chimse- ridse, the last named only still remaining alive, iu three genera. Subclass II.— DIPNOI. Two orders of this subclass are known, as follows : Paired fins rudimentary or absent ; pelvic elements (so far as known) double, lateral ; body more or less protected by plates ; Arthrodira. Paired fins archipterygial (i. e., unibasal) ; i3elvic elements fused on the middle line; body with- ■"1* out bony plates (Fig. 5) ; Sirenoidea. The Arthrodira are represented by three or four families, the Coccosteid?e, the Asterosteidse, and the Mylostomidie, all restricted to the periods of the Devonic system. Their remains abound in these horizons of Europe and North America. They possess behind the head a segmented _ dorsal plate and a compound ventral plate Ym. 5.— Ceratodus forsterii Krefft; or plastron. The anterior lateral superior 51.teoL1°t°S'f'''i%r'act?n^ plates articulate with the posterior lateral trichia. From wiedersheim. cranial elements by a hinge-like jomt. Notochord persistent. These forms have some resemblance to the Agnatha Antiarcha, but they have a well-developed mandibular 15 arch, and distinct pelvic elements. Their teeth are processes of the edges of the jaws. The SiRENOiDEA commence in the Carbonic system and continue to the present day, when representatives of two genera, Ceratodus and Lepidosiren, exist in the fresh waters of the Southern Hemisphere. The families are four, the Dipteridse, Phaneropleuridje, and Cter.odontidse where the skull is covered with small tessellated plates, and the Lepido- sirenidse, where the bones of the skull are few and large. The Dipteridae and Phaneropleurida3 are confined to the Devonic, and the Ctenodontidse to the Carbonic. The Lepidosirenida3 commence in the Permian (Texas) and continue to the present time. The teeth of Sirenoidea are plates with radiating ridges, or else processes of the jaws. Subclass III.— JELASMOBRANCHII. The sharks are divided into three orders, as follows : Paired fins uuibasal and archipterygial ; a basioccipital element ; Ichthyatomi. Paired fins pluribasal (with three axonosts) ; no basioc- cipital ; Selachii. Paired fins sessile, with a single skeletal element of uncertain homology ; Acanthodii. Fig. 6. — Pleuraeanthus lueiut Ag. From the Coal Measures of Alsace, From Sauvage. In the IcHTHYOTOMi the notochord persists, and the centra are represented b}' segments ; the neural spines, axonosts, and baseosts are continuous, and the dermal rays are actinotrichia (that is, hair-like and more numerous than the baseosts). There are are two families, the Pleuracanthidse and the Cladodontidse, which differ in the form of the pectoral fin. In the former they are biserially pinnate (Fig. 6), while in the latter they are uniserially pinnate (branches (baseosts) on one side of the axis (axonosts) only). The Pleuracanthidse have teeth with 16 two principal equal cusps, while the Cladodontidae have a principal median cusp, with or without smaller lateral cusps. The Ichthyotomi are all confined to the Carbonic system. The Selachii (sharks and rays) present two lines of relation, or sub- orders, which differ as follows : Vertebrae, when developed, having the concentric laminae predominating over the radiating laminae ; anal fin absent ; Tedospondyli. Vertebrae when developed, with the radiating laminae predominating over the concentric; anal fin present ; Asterospondyli. In the Tectospondyli the majority of living types have the body de- pressed, so that the branchial fissures are on the inferior surface. This Fig. 1 .—Heptanchus griseus, left pectoral fin, pluribasal type ; SB, scapular arch ; NL, forameu ; Pr, propterygiura ; Ms, mesopterygiuni ; 3lt, metapterygium ; a, b, axis of metapterygium , i?a, basilars ; FS, actinotrichia. From Wiedersheim. type is seen in the skates, saw-fishes, and rays. In the Asterospondyli, on the other hand, the branchial fissures are lateral. There are several families of Tectospondyli, which appeared at different periods of geo- logical time. They are as follows : I. Teeth in the usual jaws. a. The crowns of the teeth closely overlapping each other like shingles. Edges of crowns forming a grinding face ; (1) Petalodontid(Z. aa, The crowns of the teeth not overlapping ; i3. Summits of crowns forming a grinding surface. 17 (2) Psammodoniidcd ; (3) Rajidm ; (4) Ehmobatidce; (5) Trygonidiz ; (6) Myliohatidm (eagle rays). /5/3, Summits of crowns elongate cusps. Body not flattened ; (7) Spinacidce. II. Teeth in a produced muzzle.* (8) Pristiophoridce ; (9) Pristidce (saw-fishes). These families have the following range in geological time : Plistocene 1 2 3 4 5 6 7 8 9 Neocene . Eocene . . Cretaceous Jurassic. . Triassic . . Carbonic . Devonic . Siluric . . Ordovicic . Cambric . Huronic . The families of the Asterospondyli differ as follows : I. Teeth molariform.* Teeth separate ; (1) CestradontidoB. Teeth confluent ; (2) Cochliodontidce, II. Teeth with elevated cusps. (S) Hexanchidce ; (4) Scylliidce; (5) Lamnidce. (6) Carchariidce ; * Except 8ome Cestraciontidse. 18 The distribution of these families in time is as follows : 1 3 1 3-4 5 6 Siluric From these tables it appears that the Selachii with grinding teeth are prior in geological age to those with teeth especially appropriate to a carnivorous diet. The latter reached their greatest perfection, as well as size and numbers, in the Neocene system, some of the Carcharodons having been probably seventy feet in length. • All the families of Elasmobranchii are common to the Old and New Worlds. The order Acanthodii possesses a specialized type of paired fins, more like those of the higher fishes than is found in other Elaemo- branchii. The vertebral axis is notochordal, and the fins do not display any actinotrichia or rays. The males have no claspers. Tail hetero- cercal. In the known genera the fins are supported by a large spine in the anterior border of each ; and the integument is covered with small quadrate granules, which also extend over the fins. The teeth are 19 absent or minute, or consist of a single compressed triangular cusp. The known species are of rather small size. The Acanthodidie has but one dorsal fin, while the Ischnacanthidae and Diplacanthidse have two. In the Diplacanthidaj only have clavi- cles been observed ; in the other two families they are wanting. Among Diplacanthidse the genus Climatius is remarkable in having a series of spines on each side between the ventral and pectoral fins, which indicate the position of the primitive lateral fold, from which the paired fins are supposed to have been derived. Diplacanthidse are Devonic, as are also Ischnacanthidie. The Acanthodidse belong to the Devonic and Carbonic. Subclass IV.— TELEOSTOMI. * There are four superorders of the Teleostomi or true fishes, which differ in the structure of the fins : I. Median fins each with a single bone representing axonosts. .Paired fins unibasal ; Rliipidopterygia. II. Median fins with numerous axonosts. Paired fins with baseosts; pectorals with axonosts, which are distinct from baseosts ; Crossopterygia. Paired fins wdth baseosts ; pectoral fins with axonosts and baseosts confounded ; pluribasal; Podopterygia. Pectoral fins only with baseosts, these confounded with axonosts, and pluribasal ; Actinopterygia, Superorder Rhipidopterygia. The orders of Rhipidopterygia are the following. They all have actinotrichia in place of fin-rays. I. Paired fins with the basilars arranged on each side of the median axis, or archipterygial. Median fins with basilars ; Taxistia. II. Paired fins with the basilars arranged fan-shaped at the end of the short axis. Median fins with basilars ; Rhipidistia, Median tins without, caudal fin with, basilars ; Adinistia, The Taxistia includes but one family, the Holoptychiidse, which is of Devonic age. The Rhipidistia includes the Tristichopteridse, from the Devonic and Carbonic ; the 0.steole])idse, from the same ; and possibly the Onychodontidse, which are Devonic. The Actinistia includes the single family of the Coelacanthidie, which appears in the Lower Carbonic and ranges to the Upper Cretacic, in both Europe and America. (Fig. 8.) In all of the Rhipidopterygia the tail is either heterocercal or diphy- cercal, and the chorda dorsalis persists. The scales have a layer of ganoine, which extends also on the head. The latter has a well- defined, persistent transverse suture separating the parietal from the frontal elements. 20 Superorder Crossopterygia. The superorder Crossopterygia includes two orders, as follows : Baseosts and axonosts well developed ; actinotrichia ; no fin-rays ; pectorals ? unibasal ; Haplistia. Baseosts rudimental ; fin-rays ; pectorals tribasal ; Cladistia, But one family is included in the Haplistia, the Tarasiidse, from the Lower Carbonic of Scotland. The Cladistia are represented by a family which is not known in the fossil state, the Polypteridse of the rivers of Africa. The vertebrae in this genus are ossified and biconcave. Superorder Podopterygia. The superorder Podopterygia has also two orders. They are thus defined : Branchiostegal rays present ; Lysopteri. No branchiostegal rays ; Chondrostei. Fig. S.— Undina penicillata Munst. (Coelacanthidae) ; one-third natural size, showing rhipidoij- terygian and actinistious types of fins; j, jugular plates; 6, scales of Undina acutidens. From Zittel. In these orders the notochord is persistent, and there are either acti- notrichia, or fin-rays which are more numerous than the baseosts. Tail heterocercal or diphy cereal. The Lysopteri includes four families, which differ as follows : I. Tail heterocercal. Teeth acute, external ; Palceoniscidce. Teeth obtuse, on palate and splenial ; PlatysomidcB. No teeth ; ChondrosteidoR. II. Tail diphycercal. Teeth present ; scuta on body ; Belonorhynchidce. 21 The Paleoniscidae range from the Devonic to the Jurassic inclusive ; the Platysomidte belong to the Carbonic "exclusively ; the Belono- rhynchidie to the Trias ; and the Chondrosteidaj to the Jurassic. The Chondrostei are degenerate representatives of the Podopterygia. They are deficient in various normal ossifications, and have an addi- tional series of membrane bones in the middle line of the skull. The two families are the Accipcnseridie, or sturgeons, and the Polyodontidse, or paddle-fishes. Both are represented at the present day in the northern regions of both hemispheres, and both appear first in geologi- cal time in the Eocene system. Superorder Actinopterygia. In this superorder we have the finally specialized type of the true fishes. This consists in the abbreviation of the skeletal parts of the true fins, so that the basilar elements become sessile on the scapular. Coincidentally with this result, the fin-rays of the median fins become distinctly developed and articu- lated each with its corresponding baseost or axonost. Fishes of this superorder have, for the most part, liomocercal tails, the superior lobe being contracted to the size of the inferior lobe, and the vertebrae ossified ; but in some of the lower and older types both heterocercal tails and notochordal vertebrae still remain. Diphycercal tails also continue well along in the ascend- ing scale. fall into FiG.9.—Salmo fario L : left shoulder-girdle*, Cm, posttemporal ; B 1, epiclavicle ; D, clavicle ; D 2, postclavicle ; S, scapula ; Co {CI), coracoid ; Ea, basilars; L, scapular foramen ; HS, FS, fin- rays. The Actinopterygia two tribes : Ventral fins abdominal ; a ductus pneumati- cus ; no spinous dor- sal fin ; parietal bones not usually separated by supraoccipital ; scales usually cycloid ; Malacopteri, Ventral fins usually thoracic or jugular; no ductus pneumaticus ; usually a spinous dorsal fin ; parietal bones usually separated by the supraoccipital ;• scales usually ctenoid ; Acanthoj^teri. In the cartilaginous fin-rays and posterior ventral fins the Mala- copteri most nearly approach the fishes of the superorders considered in the preceding pages, than do the Acanthopteri. The persistence of the communication between the swim-bladder and the gullett (ductus pneumaticus) is another indication of this affinity. Accordingly we find representatives of the Malacopteri in older periods than we do the 22 Acanthopteri. They are abundant in the Jurassic and Cretaceous and later periods, while the Acanthopteri are very rare first in the Cre- taceous, and are abundant first in the Eocene. The orders of the Malacopteri are the following : I. Median fins with actinotrichia. Basilars of median fins well developed ; notochord persistent ; (1) Docopteri* II. Median fin-rays equal to and articulating with axonosts. a, Vertebrae complex, the pleurocentra and intercentra distinct. Anterior vertebrae similar to others ; (2) Merospondyli. aa, Vertebrae with centra and intercentra both complete on part of the column at least ; amphicoelous ; Anterior vertebrae similar ; (3) Halecomorphi. aa«, Vertebrae (intercentra) opisthocoelous. Anterior vertebrae similar ; a precoracoid arch and a coronoid bone ; (4) Ginglymodi. aaaa, Vertebrae (intercentra) amphicoelous. /? ^, Precoracoid arch. Y, No symplectic bone. Pterotic simple ; anterior vertebrae modified, and with ossicula auditus ; parietals not distinct (5) Nematognathi. Pterotic annular, including a cavity which is closed by a distinct bone ; anterior vertebrae simple, without ossicula auditus ; parietals distinct ; (6) Scyphophori. yy, A symplectic bone. Anterior vertebrae coossified, and with ossicula audi- tus ; pterotic simple ; parietals distinct ; (7) Pledospondyli. Anterior vertebrae not modified ; pterotic simple ; parietals distinct ; (8) Isospondyli. No precoracoid arch. Y, Scapular arch suspended to cranium. '5, A symplectic. Anterior vertebrae and pterotic simple ; parietals separated by supraocciptal ; (9) Haplomi. Anterior vertebrae modified ; parietals not separated ; (10) Glanencheli. dd, No symplectic. Anterior vertebrae simple ; a preoperculum and pala- tine arch ; (12) Ichthyocephali. YY, Scapular arch free from cranium. (J, A symplectic bone. Hyoid arches and pectoral baseosts developed ; (13) Holostomi. Baseosts fused into a sinale cartilage ; (11) Xenomi. * The position of this order is uncertain. No symplectic. Opercular bones and five osseous branchial arches, with ceratohyal ; (14) Enchelycephali. Opercular bones and one osseous branchial arch, ceratohyal ; (15) Colocephali. No opercular bones, nor ceratohyal, nor osseous branchial arches ; (16) Lyomeri. The families of the Physostomi are as follows : DocoPTERi ; Dorypterida). Merospondyli ; Sauropsidie ; Pycnodontidie ; Stylodontidie ; Sphsero- dontidse ; Macrosemiidie. Halecomorphi ; Amiidaj (dog-fishes). GiNGLYMODi ; Lepidosteidie (bony gars). IsospONDYLi ; Dapediidie ; Lepidotidie ; Aspidorhynchidse ; Saurodon- tidse (Fig. 10) ; Osteoglossidie ; Heterotidse ; Galaxiidse ; Clupeidse (herring) ; Chirocentridse ; Salmonidse (salmon) ; Thymallidae (grayling) ; Alepocephalidse ; Gonorhynchidje ; Sauridse ; Lutodir- idie ; Aulopidie ; Elopidie; Albulidse ; Hyodontidie ; Notopteridse. AcTiNOCHiRi ; Pelecopteridse. Plectospondyli ; Characinidse ; Sternopygidse ; Cobitidse ; Cyprini- dse (carp) ; Catostomidie (suckers). ScY^PHOPHORi ; Mormyridie ; Gymnarchidse. Nematognathi ; Siluridse (cat-fishes) ; Hypophthalmidje ; Aspredinidse. Haplomi; Esocidse (pike); Stratodontidse ; Umbridae; Cyprinodon- tidse ; Amblyopsidse (blind fishes). Xenomi ; Dalliidse. Glanencheli ; Gymnotid^e (electric eels). IcHTHYOCEPHALi ; Monoptcridie. HoLOSTOMi ; Amphipnoidse. Enchelyxephali ; Nemichthyidse; Anguillidse (eels); Congrid^e (eels) Synaphobranchidse ; Simenchelyidse. Colocephali ; Mursenidie (eels). Ly'OMERI ; Saccopharyngid?e; Eurypharyngidse. Fig. 10.— Portheus molossus, Cope ; isospondylous fish from the Upper Cretacic of Kansas, one-twentieth natural size. Original. The geological range of the orders of the Malacopteri is as follows : 24 CD • CO j T— 1 t-H O 1 Oi 1 1 ! 1 i 1 1 : I 00 1 1 • 1 1 1 i t— CD 1 CO 1 1 i i ! i 1 1 Plistocene . 1 Neocene. . Eocene . . Cretaceous Jurassic. . Triassic . . Carbonic . Devonic . Siluric . . Cambric . Huroriic . 25 The orders of which no record is given are only known in the living state. The Plectospondyli represent the highest form of the Malacop- teri, while the Haplomi give the connection with the Acanthopteri through the Percesoces of the latter. The Isospondyli offer the con- nection with the Merospondyli. The Apodal line (the eels), including orders Nos. 12 to 15, is a degenerate one, ending in the greatly degen- erate deep-sea forms of the Lyomeri. The orders of the Acanthopteri are the following : A. Scapular arch suspended posterior to the cranium. Maxillary bone distinct ; no interclavicles ; epi- branchials aud pharyngeals present ; inferior elements distinct ; (1) Opisthomi. A A. Scapular arch suspended to cranium by a posttemporal bone, a, Ventral fins abdominal. Branchial arches developed, third superior pharyn- geal enlarged ; gill-fringes linear ; no inter- clavicles ; (2) Percesoces. Epibranchials and superior pharyngeals reduced in number ; interclavicles ; gill-fringes linear ; (3) Hemibranchii. Epibranchials and superior pharyngeals wanting; gill-fringes in tufts ; (4) Lophobranchii. aa, Ventral fins thoracic or jugular. /5, Anterior (spinous) dorsal fin expanded into trans- verse laminae, sessile on cranium. Cranium normal ; branchial bones present ; (5) Discocephali. y3/5, Spinous dorsal fin not tranversely expanded. Y, Posttemporal projecting freely from skull. First vertebra united by suture to cranium ; inter- calaria united behind supraoccipital ; basilar pectoral bones elongated ; (6) Pediculati. Posterior cephalic region normal ; the anterior twisted so as to bring both orbits on one side ; inferior pharyngeals distinct ; (7) Heterosomata. Cranium normal, premaxillaries usually coossified with maxillaries behind, and the dentary with the articular ; pharyngeal bones distinct ; (8) Plectognathi. Cranium normal ; bones of jaw distinct ; (9) Percomorphi. yy. Posttemporal an integral part of the skull. Cranium normal ; bones of jaws distinct ; pharyn- geals separate ; (10) Craniomi. The families of the preceding orders are as follows : Opisthomi ; Mastacembelidse ; Notacanthidse. Percesoces ; Opheocephalidoe ; Mugilidse (mullets) ; Atherinidse ; Sphyraenidse (barracuda) ; Scombresocidte (soft gar). Hemibranchii ; Pegasidse ; Gasterosteidse (stickle-back) ; Fistular- iidae ; Centriscidse ; Amphisilidse ; Dercetidie. 26 LoPHOBRANCHii ; Solenostomidse ; Syngnathidse (pipe-fishes) ; Hippo- campidje (sea-horses). DiscocEPHALi ; Echeneididse (remoras). Pediculati ; Antennariidse ; Lophiidse (fishing frog). Heterosomata ; Pleuronectidse (flat-fishes). Plectognathi ; Triacanthidse ; Balistidse (trigger-fishes) ; Tetro- dontidse (bladder-fishes) ; Diodontidse ; Ostraciidse. Percomorphi ; (Anacanthini) Ophidiidse ; Gadidse (cod) ; Macru- ridse ; (Haj^lodoci) Batrachidse (toad-fishes) ; (Cyclopteroidea) Cyclopteridse ; (Scatophagoidea) Scatophagidse ; (Epilasmia) Acroneuridse ; Chsetodoutid^e ; (Rhegnopteri) Trichidiontidse ; (Distegi) Scorpsenidse ; Cottidse (sculpins) ; Blenniidse ; Gobiidse ; Platycephalidse ; Rhamphocottidse ; Agonidse ; Heterognathidse ; Gerreidse ; Carangid^e (pompano) ; Sillaginidse ; Pristipomatidse ; Scienidse (maigres) ; Sparidse ; Percidse (perch) ; Berycidse ; Scombridse (mackerel) ; Trichiiiridse ; Xiphiadidse ; (Labyrin- thici) Osphromenidse ; Anabantidse ; (Pharyngognathi) Embio- tocidfe ; Cichlidse ; Labridse ; Scaridse. Craniomi ; Triglidse (gurnards) ; Dactylopteridse. The table shows that no order of the Acanthopteri has become extinct. The Percomorphi display the greatest numbers and impor- tance at the present time. The Lophobranchii and Plectognathi are degenerate types. The geological distribution of these orders is as follows : 27 Huronic Cambric Siluric Triassic Jurassic Eocene Neocene Plistocene .... 1— » to CO 1 1 1 Ox i 1 OS -J I 00 eo O 28 Class III.— BATRACHIA. There are three subclasses of Batrachia, as follows : Basioccipital, supraoccipital, intercalary, and supratem- poral bones present ; propodial bones distinct ; Stegocephali. Basioccipital, supraoccipital, and supratemporal bones wanting ; propodial bones distinct ; no urostyle ; Urodela. Basioccipital, supraoccipital, intercalare, and supratem- porals wanting ; frontals and parietals connate ; propodial bones connate; lumbosacral vertebrae united into a urostyle ; Salientia. Subclass I.— STEGOCEPHALI. There are four orders of Stegocephali. «, One occipital cotyloid articulation. Vertebral bodies represented by basal and lateral ele- ments (inter centra and centra) ; Ganocephali. aa, Two occipital condyles. Vertebrae represented by distinct and incomplete inter- centra and ceutra (pleurocentra) ; atlas segmented ; Bhachitomi. Centra and intercentra complete, making two vertebral bodies to each neural arch ; Embolomeri. No centra ; intercentra, ea.ch supporting a neural arch ; Microsauri. Fig. 11. — Trimeror/iMc'iis insignis Cope ; parts of skeleton, natural size ; a, occipitalsegment with single cotylus; b,c, posterior part of lower jaw ; d,e, series of verteljrae; d, side view, depressed ; e, view obliquely from above, both showing the rhachitomous structure ; i, inter- centra ; p, pleurocentra. Original. From the Permian of Texas. 29 The Stegocephali are only known from the Carbonic and Triassic systems. They are all more or less notochordal, but a considerable range in this respect is found in the Microsauri. In some of them the notochord is cut off by the vertebral ossification, while in others (Branchiosaurus) the ossification is a mere sheath round the chorda. In the Ganocephali and Rhachitomi the vertebral centra are repre- sented by a segment beneath each branch of the neural arch, the pleuro- centrum (Fig. 11, jo); and these rest below on a median inferior piece, the intercentrum (Fig. 11, i)- Embolomeri each of these segments is developed so as to form a disc, so that there are two ver- tebral bodies to one arch. In the Microsauri the intercentra have been fused with the pleurocentra, so that the body consists principally of the FiQ. 12— Trimerorhachisinsignis Cope; skull. From the Permian bed of Texas. Original. former, and the same structure is characteristic of the remaining Batra- chia. In the Keptilia, on the other hand, the intercentra gradually disappear, being represented in most of the types in the cervical and caudal regions only. In the latter they support the chevron bones. Of the Ganocephali two families are known, the Trimerorhachidse (Figs. 11-12), without, and the Archegosauridse with neural spines of the vertebrae. They occur in the Coal Measures and the Permian of Europe and North America. The Rachitomi possesses but one family, the Eryopidse, which often reached a large size (Fig. 13). These are confined to the Coal Measures. 30 If the Labyrinthodontidse belong to this order they range also to the Trias inclusive, in both continents. The Embolomeri includes one family, the Cricotidse, which belongs to the Permian of Europe and North America. (Fig. 14.) The MiCROSAURi embraces the following families : Branchiosauridse ; Hylonomidse; Molgophidse; Phlegthontiidse. The smaller forms are mostly from the Coal Measures, while some large ones occur in the Trias. Some of the Hylonomidse come from the Permian of North America and Europe. They were mostly of small size, and their ver- FiG. vs.— Eryops megacephalus Cope ; skull, froiu above, one-fourth natural size. From Per- mian bed of Texas.. From Cope. tebrse exhibit very various degrees of ossification. In some the bodies are completely ossified, while in others the ossification forms a super- ficial layer. In the Labyrinthodontidie the ossification is more com- plete. The dentine may be entire, or deeply inflected, so as to form straight or labyrinthic folds. 31 Fig. li.—Cricotus heteroclitus Cope ; one-third natural size. From Permian of Texas, a, head from above ; 6, part of belly from below. From Cope. Subclass XL— URODELA. There are three orders of Urodela. a, An OS intercalare. Palatine arch and vomer present ; Proteida. aa, No OS intercalare. A maxillary arch and vomers ; Pseudosauria. No maxillary arch or vomers ; Trachystomata. Under the Proteida the only family known is the Prote'idse (Fig. 15). The Pseudosauria embraces the following families: Cryptobran- chidse; Arablystomidse ; Hynobiidse; Plethodontidse ; Thoriidse ; Des- mognathidae; Salamandridse ; Pleurodelid^e; Amphiumidse; Caeciliidise. The Trachystomata includes only the family of the Sirenidse. A possible member of the Proteida occurs in the Coal Measures, but certainly known members of the order are not known in a fossil state. Of Pseudosauria fossil forms are first found in the Laramie in America, and they are not uncommon in the Neocene in Europe, Trachystomata are not known fossil. 33 Subclass III.— SALIENTIA. But one order of this subclass is recognized. It is thus defined : Vomers and palatopterygoid arch present ; Anura. The Anura has the families arranged under the following suborders : Internal nostrils opening together on the middle line ; no tongue ; coracoids connected by a cartilage on each side ; Aglossa. Internal nostrils separate ; a tongue ; coracoids con- nected by a separate cartilage on each side, one overlapping the other ; Ardfera. Internal nostrils separate ; a tongue ; a single median cartilage connecting all the coracoids ; scapular arch free ; Firmistemia. As in Firmisternia, but scapular arch articulated to skull ; Gastrechmia. (Aglossa) : Xenopidse ; Pipidje. (Arcifera) : Discoglossidje ; Bufonidse ; Dendrophryniscidse ; Astero- phrydidse ; Pelodytidse ; Scaphiopidse ; Hylidse ; Cystignathidse ; Amphignathodontidse ; Hemiphractidse. (Gastrechmia) : Hemisidae. (Firmisternia) : Engystomidse ; Phryniscidse ; Dendrobatidse ; Cophy- lidse ; Dyscophidse ; Colostethidse ; Ranidse ; Ceratobatrachidse. Remains of Anura have been found in the Jurassic beds of Colo- rado, but to which suborder they pertain is unknown. They next appear in the Eocene of Wyoming. Well-defined forms are found in the phosphorites of France, which include both Arcifera and Firmis- ternia. They continue to the present day. The time relations of the orders of Batrachia are represented in the preceeding table. Class IV.— MONOCONDYLIA. There are two subclasses of Monocondylia. Anterior limbs ambulatory, with numerous carpal and meta- carpal bones ; two aorta roots ; integument consisting partly of scales ; JReptilia. Anterior limbs volant, with the carpals and metacarpals more or less coossified and reduced in numbers ; integument consisting in part of feathers ; one aorta root ; Aves. Approximations between these subclasses exist at various points. Thus the Dinosaurian reptiles resemble birds in the structure of the posterior limbs and pelvis ; while among birds the Saururse approach reptiles in the structure of the manus and of the caudal vertebrae. The definitions above given are, however, not violated. 34 Anura. Trachystomata Pseudosauria Proteida Microsauri Embolomeri 1 1 Rachitomi 1 Ganocephali Plistocene . . . Neocene .... 1 Eocene .... Cretacic .... Jurassic .... Triassic .... Carbonic . . . Devonic .... Siluric .... Ordovicic . . . Cambric .... Huronic .... 35 Subclass I.— REPTILIA, (1) Ichthyopterygia. (2) Testudinata, Nine orders of Reptilia are known. I. The quadrate bone united with the postorbital bars by suture. A. Cranium with one postorbital bar. (Synaptosauria.) a, A paroccipital bone. A supratemporal bone ; ribs two-headed on cen- trum ; carpals and tarsals not distinct in form from metapodials; No supratemporal ; sub- and postpelvic ossifica- tions ; interclavicle and clavicles separated from and below scapular arch ; ribs one- headed ; coracoid large, free ; aa, No paroccipital bone. Ribs mostly two-headed, capitulum intercentral ; clavicles and interclavicles forming part of shoulder girdle ; no sub- or postpelvic bones ; pelvic elements below plate-like; obturator foramen small or none ; Ribs one-headed ; scapula triradiate ; no clavicles ; coracoid large, distinct; no sub- or post- pelvic bones ; AA. Cranium with two postorbital bars. a, No paroccipital bone ; (no supratemporal). Ribs two-headed ; no interclavicle ; external digits greatly elongate to support a patagium ; Ribs two-headed ; no interclavicle ; acetabulum perforate ; feet ambulatory ; no patagium ; Ribs two-headed ; an interclavicle ; acetabulum closed ; feet ambulatory ; no postfrontal bone ; Ribs one-headed ; an interclavicle ; acetabulum closed ; feet ambulatory ; (8) Bhynchocephalia. II. The quadrate bone loosely articulated with the postorbital bars and only proximally. (Streptostylica.) The quadrate bone in contact only with adjacent elements ; no paroccipital ; supratemporal present ; ribs one-headed ; one or no post- orbital bar ; (9) Squamata. The geological relations of these orders ^re as follows ; (3) Theromora, (4) Plesiosauria, (Archosauria.) (5) Ornithosauria. (6) Dmosauria. (7) Crocodilia. 36 ^ Plistocene . 1 2 3 4 5 0 7 8 9 Neocene Eocene . . _ Ofptflpic Jurassic ? Triassic . . Kjcii. UUUIL/ • Siliiric . Ordovicic . Cambric . Huronic . The above table shows that the Theromora is the only Paleozoic order. Its members appeared towards the close of the Carbonic system, and spread over the whole earth before the end of the Paleozoic ages. Their remains have been found in New Mexico, Texas, Illinois, France, Germany, Russia, and the Cape Colony of Africa. They include carnivorous forms and those with a molar dentition for crushing hard substances (Fig. 17). They display some points of strong resemblance to the Mammalia. Thus in the Theriodontia the posterior foot is closely similar to that of the^Monotrematous order. In several families the humerus strongly resembles that of the same type. The scapular arch in the Procolophonina resembles that of the same type. In the Theriodontia the coracoid bone is reduced as in the salamanders. In the Diadectid^e there are teeth on the vomer as in the same. The humeral condyles are much like those of salamanders. The vertebrae are biconcave and mostly notochordal, in all Theromora, thus resembling the Batrachia. Thus the Theromora are at once more nearly allied to 37 the Batrachia on the 'one hand and the Mammalia on the other than any other reptiles. The size of the known species ranges from that of a rat up to that of a lion. The order Theromora includes five suborders. I. Palate closed, except posteriorly. A temporal foramen ; Placodonta. II. Palate open anteriorly for nares. A. The coracoid bone large, reaching sternum. Dentition abundant ; pubis and ischium plate-like ; ribs one-headed ; Proganosauria. Fig, 16.—Lys(rosaurusfrontosus Cope ; an Anornodont from South Africa, one-third natural size. 38 Fig. 17. — Empedias molaris Cope; inferior view of skull, oue-third natural size. A Cotylo- eaurian from the Permian of Texas. Four or five sacral vertebrae ; centra not notochordal ; no intercentra; dentition imperfect or wanting ; ob- turator foramen minute ; Anomodonta. AA. The coracoid reduced, not reaching sternum. Ribs two-headed ; two or or three sacral vertebrae ; centra generally notochordal ; intercentra generally present ; dentition al)undant ; Theriodonta. Ribs single-headed ; temporal fossa overroofed ; dentition abundant ; intercentra ; Cotylosauria. The Placodonta includes the Placodontidse only. The Proganosauria, the Mesosauridae, the Procolophonidse, Proterosauridse, and Rhyncho- sauridae. Cotylosauria includes the Pariasauridse and the Diadectidse (^'s-m . ... The Theriodonta embraces the families of the Clepsydropidse, Pario- tichidie, and Bolosauridse. The Anomodontia includes the single family of the Dicynodontidse (Fig. 16), and perhaps the Endothiodontidse. The geographical distribution of these suborders is as follows : 89 N. America. S. America. Europe. Africa. Placodontia ; X Proganosauria ; XXX Anomodontia ; X X Theriodonta ; X XX Cotylosauria ; X XX Of the above families the Placodontidse and the Proterosauria occur in the the Trias. The others are confined to the Permian. The Plesiosauria embraces the following families : Plesiosauridie, Nothosauridie, and Lariosauridie. The last two families are only known from the Triassic of Europe. The Plesiosauridie range from the Jurassic to the Cretaceous, inclusive, and include many species. They are all adapted for aquatic life, and were sometimes formed like the Cetacea ; but others had very long necks, which could extend the head to great depths or reach the surface when the body was sunk. Some of the species reached a length of forty or fifty feet. Their remains have been found in North and South America, Europe, and New- ZcHland. The order Testudinata presents four subordinal modifications, as follows : I. No descending processes of the parietal bones. Vertebrae and ribs free and separated from a bony exoskeleton ; no descending processes of the parietals ; Athecce. II. A carapace and plastron, and descending process of parietals. a, Sacral and caudal ribs articulating with neural arches only. Neck bending in vertical plane, last cervical articulat- ing with first dorsal by zygapophyses only ; pelvis not anchylosed ; marginal bones wanting or rudimental ; Trionychoidea. aa, Sacral and caudal ribs articulating with body of vertebrae only. As the last ; but marginal bones present and con- nected with ribs, and last cervical and last dorsal vertebrae articulating by bodies ; pelvis not anchy- losed to plastron ; Cryptodira. Neck bending in horizontal plane, the last cervical and first dorsal vertebae articulating by bodies ; pelvis anchylosed to carapace and plastron ; mar- ginal bones present and connected with ribs ; Pleurodira. The Athecce includes the single family of the Dermochelydse. The Trionychoidea includes only the Trionychidse. The Cryptodira embraces the Cheloniidse, Testudinidse (Fig. 18), Cinosternidae, Dermatemydidae, Chelydridae, Baenidae, and Adocidae. The Pleurodira includes the Pleurosternidae, Sternothaeridae, Pelome- dusidae, Pleisiochelydidae, Chelydidae, and Carettochelydidae. The earlier tortoises are intermediate in character between the Cryp- todira and Pleurodira, and they have continued side by side to the 40 present day. The Pleiirodira are, however, now confined to the Southern Hemisphere. The Trionychoidea first appear in the Creta- ceus, and continue as the so-called soft-shelled turtles of fresh water. Fig. 19. — Ichthyosaurus tenuirostis Conyb. ; one-sixth natural size. Fiom the Lias of Wiirt- emburg. From Doederlein. 41 the Eocene, and were then and still are one species now exists, the " leather-back families of have the Ichthyo- bones of The Athecse first appear in marine in their habits ; but turtle." The order Ichthyopterygia embraces the sauridi« and Mixosauridte. The Mixosauridie the fore arm and leg distinguished by their form from those of the carpus and tarsus, and they thus approach more nearly other reptiles. They are the oldest family inhabiting the Euro- pean waters during the period of the Trias. In the Ichthyosauridie all the bones of the limbs have the same form, except the humerus and femur. They continued until the close of the Cretaceous. They were the especially marine Reptilia, having the shape and habits of the Cetacea. They reached the length of twenty-five feet during the Cretaceous period. Their remains have been found in all parts of the world except Africa, but they were more numerous in the oceans now cover- ed by Europe than elsewhere. Some of the forms had but few (Mursenosaurus) and others (Baptanodon) no teeth. This typi- cal genera had numerous teeth grooved at the base. The vertebrae are short and biconcave (Fig. 19). Two families enter the Ornithosauria, viz., the Pteranodontidse and the Pterodac- tylidse. The last-named family is furnished with teeth, while the first-named includes species which have a toothless beak like that of abird. The Pterodactylidse range from the Trias to the Cretaceous inclusive. The earlier forms had a long tail, and these continued through the Jurassic (Fig. 20). Some of the later forms had an edentulous beak in front of the toothless portion. The true Pterodactyles had a much abbrevia- ted tail and a long neck. The Ornithosauria were the flying order of reptiles, comparing with the other orders much as bats compare with other orders of Mammalia. The DiNOSAURiA include the great land reptiles of the Mesozoic systems. The order embraces no species adapted for flight, and none adapted for a life in which movement was made by paddles or limbs modified for swimming. The order includes the largest land animals Fig. 20. — Dimorphoda macronyx Ow. from the English Trias. B, sternum. From Owen. ^ - - _ -J" aide''onl^t7mh'?.T/i'^i*'ii' ^«'^«™""^^ ««;>^emt.* Cope; cervical vertebra, from ab^^lnd ^Ij^a.- 7 » natural size; ^, do , dorsal vertebra, from behind, one-tenth natural size - C. Amphiccelias alius Cope ; femur, from behind, one-tenth natural size "aiurai size , 43 that have ever lived, and their remains have been found in all except the Australian continents. The Dinosauria embrace two suborders, as follows : Pubic elements directed downwards ; Saurischia. Pubic elements directed backwards ; Orthopoda. The Saurischia were mostly of carnivorous habits, while the Orthopoda were herbivorous. Both suborders commence in the Trias, and close with the end of the Postcretaceous. To the Saurischia belong two families, the Cetiosaurida) and the Megalosauridse, the former supposed to be herbivorous, the latter carnivorous. The former have robust inferior pelvic bones, and teeth with spoon-shaped crowns. The latter have long, slender pubis and ischium, and sharp, knife-shaped teeth. To the Cetiosauridse belong the largest known Vertebrata with ambu- latory legs ; the Camarasaurus siipremus Cope measuring seventy feet in length, and the Amp hicce lias Jragilissimus Cope being considerably larger. These creatures had the dorsal vertebrae hollow, and probably penetrated in life by branches from the lungs, while the tail vertebrse and limb bones are solid. The former served as floats and the latter as anchors as they walked on the sea bottom. This family is only known from the Jurassic of Europe and North America (Fig. 21). The Megalo- sauridse ranged throughout all Mesozoic time. Some of the species were quite small, and others were gigantic, being the most dangerous car- nivorous animals that ever existed. Many of them were of kangaroo- like form, and in some of them the vertebrae were mere hollow shells. In most of them the limb bones were hollow. The Orthopoda include the Scelidosauridae, Hypsirhophidae, Agathau- midse, Camptosauridse, Iguanodontidse, and Hadrosauridse, all supposed to have been of herbivorous habits. Representatives of these families have been found in both Europe and North America, excepting the Scelidosauridse, the members of which are so far only knowai from Europe. They all include large forms, but the most gigantic are members of the last four families. These had an additional bone, the predental, which formed a toothless extremity of the lower jaw, while the Agathaumidse add a corresponding toothless bone in front of the premaxillary. In all the families except the Agathaumidse the suc^ cessional teeth appear on the inner side of the base of the functional teeth, as in lizards. In the Agathaumidse they appeared under the middle of the base, as in crocodiles. In all except Hadrosauridse one row was in functional use at a time ; but in Hadrosauridse two or three rows were used at once. Some of the species of the latter had as many as 2,000 teeth arranged in four magazines, one in each jaw (Fig. 22). The Agathaumidse mostly had formidable horns on the head. In Agathaumas Cope the longest horn was at the middle of the nose ; in Polyonax Cope the horns were over the eyes. Agathaumas sylvestre Cope reached a length of forty feet, and had the legs of sub- equal length. Many of the Hadrosauridse were of kangaroo-like pro- portions. 45 The Cko(x>dilia includes three suborders, as follows : ^ Posterior nostrils opening in front of palate ; nasals very short; premaxillary very large; external nostrils posterior, nasals short ; Parasuchia. Nasal bones very long, separating the small premaxil- laries ; external nostrils anterior ; Pdeudo-iuchia.^ Nareal canal underroofed to behind larynx ; no clav- icle; pubis excluded from acetabulum ; external nostrils anterior ; Eumchia. The Parasuchia and Pseudosuchia are restricted to the Trias of both continents. The most important family of the Parasuchia is the Belo- dontidie (Fig. 23). The Pseudosuchia include the Aetosauridse, the members of which were completely incased in an armor of bony plates. To the Eusuchia belong the true crocodiles. They commence in the Jurassic with the Teleosauridne and Goniopholididie, which are followed in the Cretaceous by the Crocodilidse and Gavialidse, which still exist (Fig. 24). The Khynchocephalia include two suborders : Vertebrae amphicoelous ; axis undivided ; Sphenodontina. Vertebrae amphiplatyan ; dentatum of axis separate ; Choristodera. To the Sphenodontina belong the Protorosauridse and Paleohatteriidse, from the European Trias, the Homoeosauridae, from the European Juras- sic, and the Sphenodontidse, which are represented by one species now living in New Zealand. They are all of terrestrial habit, and of small or medium size. The only family of the Choristodera is that of the Champsosauridae, which is found in the Postcretaceous of North America and the Eocene of Europe. The species were aquatic, and reached the size of some of the modern caymans. The dentine is inflected at the base of the teeth, and the limbs were paddle-like. This order furnishes the starting point for the great Archosaurian series. The Squamata is an extended group, which is represented by three suborders, which are defined as follows : Alisphenoid modified as epi pterygoid or wanting, leav- ing brain-case open ; parietals flat ; an intercla- vicle and clavicle ; teeth with dentinal roots : Lacertilia. Epipterygoid present ; parietals decurved, partially enclosing brain-case ; no clavicle nor interclavicle ; teeth with osseous roots ; Pythonomorpha. No epipterygoid ; brain-case enclosed in front ; no clavicle nor interclavicle ; no fore limbs ; teeth rootless ; Ophidia. The three suborders of Squamata first appear in the Cretacic sys- tem. The Lacertilia and Ophidia still exist, but the Pythonomorpha did not survive Mesozoic time. * Position as yet doubtful. Fig. 23.— Belodon buceros Cope ; skull, one-fourth natural size. From th« Triaa of the Rocky Mountains. ^, the side ; 5, below; C, above. Original. 47 Fig. 2i.—Crocodilus acer Copei; skull, one-third natural size. From the Eocene of Utah. Original. The Lacertilia embraces the following superfamilies : I. Cervical vertebrae very numerous, more than nine. Teeth pleurodont ; (1) Dolickosauria. II. Prootic not produced beyond arched body ; teeth acrodont ; olfactory lobes not underarched ; two suspensoria. No clavicle nor interclavicle ; no columella ; tongue papillose, its extremity sheathed ; (2) Ehiptoglossa. A clavicle proximally simple; an anchor-shaped interclavicle ; a columella ; tongue papillose, not sheathed ; (3) Acrodonta. III. Prootic bone not produced beyond arched body ; dentition pleurodont ; olfactory lobes not underarched ; two suspensoria. A clavicle proximally simple; an anchor-shaped interclavicle ; a columella ; tongue papillose, not sheathed ; (4) Iguania. IV. Prootic bone not produced beyond arched body ; dentition pleurodont, or nearly so ; two suspensoria. a, Clavicle simple proximally ; olfactory lobes not under- arched by frontal. Interclavicle cruciform ; tongue papillose , (5) Diploglossa. aa, Clavicle proximally simple ; olfactory lobes underarched by frontal. Vertebrae proccelous ; tongue smooth ; (6) Tliecaglossa. 48 Vertebrae amphicoelous ; tongue papillose ; (7) Uroplatoidea. aaa, Clavicle proximally expanded ; olfactory lobes under, arched by os frontale. Tongue papillose or smooth ; (8) Nyctisaura. aaaa, Clavicles, when present, expanded proximally ; olfac- tory lobes not underarched. Vertebrie procoelous ; tongue scaly ; (9) Leptoglossa. V. Prootic bone produced beyond asched body ; one suspen- sorium (= supratemporal, wanting) ; pelvic arch rudimentary or wanting. Frontal bone underarching olfactory lobes ; supra- occipital gomphosis internal ; no orbitsphenoid ; (10) Anielloidea. Frontal bone underarching olfactory lobes ; supra- occipital gomphosis external; an orbitsphenoid ; (11) Opheosauri. The families of these superfamilies are the following : Dolichosauria ; Dolichosauriidse. Bhiptoglossa ; Chamseleonidse. Acrodonta; Agamidse. Iguania ; Iguanidse, Anolidse. Diploglossa ; Zonuridse, Pygopodidse, Anguidse, Xenosauridse, Helo- dermidie. Thecaglossa ; Varanidse. Uroplatoidea ; Uroplatidse. Nydisauria ; Eublepharidee, Gecconidse. Leptoglossa ; Xantusiidse, Teidse, Lacertidse, Gerrhosauridse, Scincidse, Acontiidse, Anelytropidse. Annielloidea ; Aniellidae. Opheosauri; Chirotidse, Amphisbsenidse, Trogonophidse. Fig. 25.— Skull of Iguana luberculata Linn ; external side view, natural size. The recent forms of the Lacertilia are much more numerous than the known fossil forms. The geological distribution of the latter is shown in the following table : 49 Huronic . Cambric . Ordovicic . Siluric . . Devonic . Carbonic . Triassic . Jurassic . Cretacic . Eocene . . Neocene . Plistocene . OS <3\ a> <» i 00 o 50 Fig. 2&.—Clidastes propython Cope; skull, one-half natural size. From the Upper Cretaceous of Alabama. The Dolichosauria are only known from European formations. Some of the species are snake-like in form. Acrodonta are best known from the same region, but one species is known from the American Eocene. It may belong to the Rhipto- glossa. The extinct Iguania and Lepto- glossa are all European as far as known, while Diploglossa are known from both continents. The Pythonomorpha (Fig. 26) include two families, the Mosasauridse and the Plio- platecarpidse. The Mosasauridse were the predominant type of sea-saurians during the Cretaceous period in North America, and they were common in Eu- rope and in New Zealand, and some species have been found in Brazil. The Plioplatecarpidse are only known from the Upper Cretaceous in Europe. Some of the species of Mosasaurus and Liodon reached a length of fifty feet. Their limbs were short, inflexible paddles, and the pelvic bones were very slender and feeble. The Ophidia include the following superfamilies : A. Supratemporal intercalated in the cranial walls. (Angiostomata.) a, No ectopterygoid ; palatines bounding choanse posterior- ly; ethmoturbinal forming part of roof of mouth ; rudi- ments of a pelvis. {Scoleco- phidia.) Maxillary bone fixed to prefrontal and premaxillary ; a pelvis ; (1) Catodonta (blind snakes). Maxillary bone vertical and free from all others ; no pelvis ; (2) Epanodonta (blind snakes). aa, An ectopterygoid ; palatines not bounding choanse pos- teriorly. Maxillary bone free, horizontal ; (2) Tortridna. A A. Supratemporal attached scale- like to cranial walls, produced freely posteriorly ; ectoptery- goid present. (Eurystomata.) 51 Maxillary bone bone horizontal, in contact with the j)re- maxillary, and furnished with solid teeth ; no rudi- ments of pelvis ; (4) Asinea (harmless snakes). Maxillary bone horizontal, thickened in front, and not reaching premaxillary, and bearing a perforate tooth ; (5) Proteroglypha (cobras). Maxillary bone vertical, not reaching premaxillary, ar- ticulating with the pre- frontal by a ginglymus, and to the ectopterygoid without imbrication, and bearing a perforated tooth; (6) Solenoglypha (vipers and pitvipers). The families embraced by these superfamilies are as follows : Catodonta; Stenostomidse. Epanodonta ; Typhlopidse. Tortrieina ; Tortricidae, Uropeltidse. AsineoL ; Xenopeltidse (Fig. 27), PythoBidse, Boidse, Charinidse, Achro- chordidse; Nothopidse, Colubridae. Proteroglypha; Hydrophidse, Najidse, Elapidse, Dendraspididse. Solenoglypha ; Causidse, Atractaspididae, Viperidse, Crotalidse. The paleontology of the snakes is very imperfectly known. The oldest known genus, Symoliophis Sauv., from the Neocomian of France, may be either one of the Catodonta or Epanodonta. The oldest forms of Asinea (harmless snakes) are Pythonidse and Boidse ; they occur in Fig. 21 —Xenopeltis unicolor Reinwt. Recent, from Siam; skull, natural size. Original. (Asinea.) both continents. The largest are the species of Paleophis Owen, which occur in the Eocene of New Jersey and England ; they equaled the largest existing pythens in dimensions (Fig. 28). The venomous snakes appear in the Neocene in both Europe and North America. The geological distribution of the superfamilies of snakes is as follows : 52 Plistocene . 1 1 2 3 4 0 6 Neocene Eocene . . Cretacic . Jurassic Triassic Carbonic . Devomc . biluric . . Ordovicic . Cambric . Huronic . Fig. 2S.—A, PalcBophis lUtoralis Cope ; B, P. halidanus Cope ; vertebrae. Both from the Eocene of New Jersey. 53 Subclass II.— AVES. There are two superorders of the birds, as follows : Metacarpal and carpal bones all distinct, the digits with ungues ; caudal vertebrae numerous, unmodified ; clavicles united; pelvic elements distinct; teeth present ; Saururce, Fig. 29. — ArchoBopteryx lithogrc^hica Wagn. From Middle Oolite of Bavaria, much reduc«d. From Dames. 54 Metacarpal and carpal bones reduced in number, coossi- fied; ungues wanting or single; caudal vertebrae reduced in number, the terminal areas usually coossified ; Eurhipidurce. The Saurur^e includes but one order, which is defined as follows : Vertebrae biconcave ; feathers arranged in one series on each side of the caudal vertebrae ; teeth present ; Ornithopappi. To this order but one family belongs, viz., the Archseopterygidae. This family is represented by one genus, Archseopteryx from the Oolite (Middle Jurassic) of Solenhofen, Bavaria. It was furnished with long feathers on the wings, adapting it for flight, and the long lizard-like tail had a row of long feathers on each side of it. It is the oldest known type of birds, and is an important one as showing the near connection of birds with reptiles. (Fig. 29.) The EuRHiPiDUR^ include four or five tribes which differ as follows : Ischium free from ilium posteriorly ;* palate dromse- ognathous (i. e., maxillopalatines articulating with vomer which is between them, palatines not articulating directly with sphenoid rostrum); no teeth ; Ischium free from ilium posteriorly ; palate ? dro- mseognathous ; teeth ; Fib. 50.~-J>inomis parviis Owen. From Plistocene of New Zealand. From Owen. iZ, ilium; it, ischium ; pp, pubis; st, sternum ; B, tarsometatarsus. * With some apparent exceptions. (1) Ratit(B. (2) Odontolcce. 55 Ischium free from ilium posteriorly ; palate ? not dromfeognathous ; teeth ; (3) Odontotormoe, Ischium coossified with ilium posteriorly; palate not dromseognathous ; feathers distributed in area, those of the wings much differentiated ; (4) Euornithes. Ischium coossified with ilium ; no teeth ; feathers universally distributed and not differentiated on wings ; (5) Impennes. The Ratitse are the ostriches and their allies ; the Odontolcse and Odontotormse are toothed birds ; the Impennes are the penguins, and the Euornithes are the remaining, or typical birds. The geological distri- bution of these tribes is as follows : 1 2 3 4 6 Carbonic 56 The tribe ODONTOLCiE includes but one order. Teeth in a grove ; sternum without keel ; wings rudi- mental ; pelvic bones free posteriorly ; vertebrse saddle-fehaped ; Dromceopappi. The Drom^opappi has but one family, the Hesperornithidse, which have been only found so far in the Upper Cretaceous of Middle North America. They have many of the characters of the divers (order Cecomorphse). To the ODONTOTORMiE One order only is referred. It is thus characterized : Teeth in sockets ; sternum keeled ; wings well devel- oped ; ischium and pubis free posteriorly ; verte- brse biconcave ; Pteropappi. The family of the Ichthyornithidse is the only one known to belong to the Pteropappi. Its members are known from the Upper Creta- ceous of North America and the Lower Cretaceous of England. To the Impennes but one order belongs. This is the Ptilopteri. Ilium not anchyiosed with sacrum ; bones of wing not foldable on each other ; metacarpals not separated ; hallux directed forwards ; feathers scale-like ; vertebrse opisthocoelous ; Ptilopteri. The Ptilopteri includes the recent family of the Aptenodytidse or pen- guins. They first appear in the Upper Eocene of New Zealand. No species, living or extinct, have been found in the Northern Hemisphere. The Ratit^ includes the following orders : Sternum without keel ; clavicles ; wings rudimental ; Struthiones. Sternum without keel ; no clavicles ; wings rudimental ; Apteryges. Sternum with keel ; clavicles ; wings rudimental ; Gastornithes. Sternum with keel ; clavicles ; wings functional ; Orypturi. The families belonging to these orders are the following : Struthiones ; Struthionidse (ostriches), Rheidse, Casuariidse, Dromaeidse, Dinornithidse (Fig. 30), jEpiornithidse. Apteryges ; Apterygidse (kiwis). Gastornithes ; Gastornithidse. Orypturi ; Crypturidse (tinamus). Of the above families the Struthionidse are remarkable in having the pubes forming a ventral symphysis. The Rheidse are equally remark- able in having a ventral symphysis of the ischia. The geological range of these orders is as follows : 57 Plistocene .... Struthiones Apteryges Gastornithes Crypturi Ordovicic .... Extinct Struthiones are known from all the great regions of the Old World, but none from the New. The extinct Apteryges come from the Australian realm only. The Gastornithes are known from the Eocenes of both Europe and North America. Gastornis has persistent sutures of the skull, and a tooth-like process on the upper jaw. G. edwardsii was nine feet high. The species of Dinornis from the Plistocenes of New Zealand were all of large size, some of them reaching twelve feet in height. One species is known from Australia. The EuoRNiTHES include numerous suborders, which are defined as follows : I. Maxillopalatines united across the middle of the palate. (Desmognathce). A. Four toes directed forwards (pamprodactylous). Toes webbed ; no basiptery golds ; (1) Steganopodes. 58 Toes free ; vomer unossified ; no basipterygoid pro- cesses; (2) Colididei. AA, Three toes directed forwards * Short basipterygoid processes; toes generally webbed ; prsecocial ; - (3) Chenomorphce, No basipterygoid processes ; bill and legs slender ; toes generally free ; altricial ; (4) Herodii. Bill and claws hooked ; toes free ; vertebrse saddle- shaped ; altricial ; (5) Accipiires. Bill hooked; toes free; vertebrse opisthocoelous ; rostrum movably articulated with skull ; basip- terygoids ; (6) Heterospondyli. Toes free ; vertebrse saddle-shaped ; rostrum fixed ; (7) Coccygomorphce. AAA. First and fourth toes directed backwards (zygodac- tylous). Rostrum freely articulated with the skull; ver- tebrse opisthocoelous ; (8) Fsittad. AAAA. First and second toes directed backwards (hetero- dactylous). Basipterygoids present ; heteropelmous ; (9) Trogonoidei. II. Maxillopalatines not united across the palate ; vomer narrowed and acute in front. {Schdzognathce.) A. Toes three forwards (anisodactylous). Schizorhinal ; toes webbed ; (10) Cecomorphce. Toes free ; legs long ; feathers with after shaft ; prsecoces; (11) Grallcs. No basipterygoids ; lachrymal bones coossified with rostrum ; toes free ; (12) Opisthocomi. Toes free ; hallux rudimental ; (13) GallincR. Toes free ; hallux well developed ; two carotids ; (14) Pullastros. Toes free ; hallux well developed ; one carotid artery ; basipterygoids ; (15) Micropodioidei. A A. First and fourth toes directed backwards (zygodac- tylous). No cseca coli ; no interclavicle ; one carotid ; (16) Picoidei. III. Maxillopalatines not united on median line ; vomer single, truncate, and excavated in front. {JEgithognathce.) A. Toes three in front (anisodactylous). Toes free ; hallux well developed ; tarso metatarsus with five tendinous canals ; basipterygoids wanting or rudimental; sternum with two notches ; no cseca coli ; one carotid artery ; (17) Passeres. AA. Four toes directed forwards (pamprodactylous). Toes free ; no basipterygoids ; sternum entire pos- teriorly ; tensor patigii brevis muscle attached to a tendon which extends to the manus ; no cseca ; (15) Micropodioidei. The time distribution of the Euornithes is as follows, as far as known : * Except Cuculidse and Rhamphastidse, which are zygodactylous. 59 Huronic . Cambric . Ordovicic . Siluric . . Devonic . Carbonic . Triassic . Jurassic . Cretacic . Eocene . . Neocene . Plistocene M 1 M Ilk. Ox " 00 CO o I-* to CO »4 60 The cormorants, pelicans, and boobies of the Steganopodes appear in the Eocene. The Eocene representatives of the Chenomorphse are primitive flamingoes, the true ducks and geese not appearing before the Neocene. The Eocene Accipitres are Falconidae, the owls not appearing before the Neocene. The kingfishers are the earliest repre- sentatives of the Coccygomorphse appearing in the Eocene ; the remain- ing families are not known prior to the Neocene. The Phasianidse of the Gallinse appear in the Middle Eocene. It is thus evident that the majority of the families of the Euornithes are not known prior to the beginning of the Neocene system. The families of the Euornithes are as follows : Steganopodes; Phsetonidse, Fregatidae, Pelecanidaa, Sulidse, Phalacro- coracidse, Plotidse. Chenomorphce ; Palamedeidse, Anatidse, Phoenicopteridse. Herodii; Ibididse, Ciconiidse, Balsenicipitidse, Ardeidse. Accipitres ; Cathartidse, Falconidae, Pandionidse, Strigidse. Psittaci; Psittacidae. Trogonoidei; Trogonidae. Cecomorphce ; Colymbidae, Heliornithidae, Alcidae, Laridae, Procellariidae. OralloB ; Chionidae, ThiAocoridae, Glareolidae, Dromadidae, Charadriidae, Otididae, Eurypygiidae, Rhinochetidae, Cariamidae, Psophiidae, Gruidae, Rallidae. Opisthocomi ; Opisthocomidae. GaUincB; Tetraonidae, Phasianidae. Pullastrce ; Cracidae, Megapodiidae, Pteroclidae, Dididae, Columbidse. Colioidei ; Coliidae. Heterospondyli ; Steatornithidae. Coccygomorphce ; Cuculidae, Coraciidae, Alcedinidae, Upupidae, Musoph- agidae, Todidae, Momotidae, Bucerotidae, Rhamphastidae, Capri- mulgidae, Bucconidae, Indicatoridae. Micropodioidei ; Cypselidae, Trochilidae. Picoidece; Picidae, Yngidae. Passeres. This order is divided into five superfamilies, as follows : Tensor patagii brevis picarian ; Menuroidei. ' desmopelmous ; Eurylosmoidei. bronchiotracheal ; ) m • j • schizopelmous. i ^ tracheal; schizo- 1 et • -j - ^ ' Y ormicaroidet. pelmous. j ^ Acromyodian, schizopelmous ; Passeroidei. The families of these superfamilies are the following : Menuroidei; Menuridae, Atrichornithidae. Eurylcemoidei ; Eurylaemidae. Tyrannoidei ; Xenicidae, Philepittidae, Pittidae, Tyrannidae, Cotingidae, Phytotomidae. Formicaroidei ; Conopophagidae, Pteroptochidae, Formicariidae. Tensor patagii brevis passe- rine; syrinx. Mesomyodian, 64 Passeroidei; Alaudidse Motacillidse, Timaliidse, Liotrichidse, Musci- capidae, Turdidse, Cinclidse, Troglodytidse, Chamseidse, Hirun- dinidse, Artamidse, Laniidse, Paridse, Paradisiidse, Corvidse, Stur- nidse, Melipliagidse, Nectariniidse, Certhiidse, Ploceidse, Tanagridse, Icteridse, Fringillidse. The primary characters of the Mammalia are seen in their limbs and in their teeth. The general characters of the skeleton may be learned from Fig. 30 (JElurodon scevus Leidy). The mutilate type of limbs is seen in Fig. 37 (Cetotherium cephalus Cope). The phalanges are con- nected by integument and form an inflexible paddle, and in the typical forms the elbow is also inflexible. The difference between the ungui- culate and ungulate phalanges may be seen by comparing Fig. 30 and Fig. 48 with Figs. 31 and 42. The compressed curved form of the former, adapted for prehension, is easily distinguished from the flat generally wide type seen in the latter, which is adapted for support. The different carpal and tarsal types of the Ungulata are seen in Figs. 31 and 32. A is the taxeopodous ; B, the proboscidian ; C, the am- blypodous, and D and E the diplarthrous, represented by the Perissodac- tyla (D) and Artiodactyla (E) respectively. Fia. 34.— Diagrams of types of mammalian dentition^ from Osborn ; 1, haplodont ; 2, proto- dont; 3, triconodont; /I, tritubercular superior and inferior ; 5, tritubercular superior, tubercu- losectorial lower ; 6, quadritubercular superior, quinquetubercular inferior; 7, quadritubercu- lar, both jaws. The molar types of dentition in historical and developmental order to the quadritubercular are represented in Fig. 34. Forms up to No. 5, inclusive, predominate in the Unguiculata ; and 6 and 7 with their deri- vatives predominate in the Ungulata. The derivatives of Nos. 6 and 7 are formed by the developmeut of ridges or crests which connect the cusps longitudinally or transversely, forming various patterns, two of which are represented in Figs. 53 and 58. Such types are termed lophodont. Those in which the crests and grooves between them become respectively elevated and profound are termed ptychodont (Figs. 41 and 56). Two subclasses are known to belong to the Mammalia : An interclavicle ; a large coracoid articulating with Class v.— MAMMALIA. the sternum ; Prototheria. i 65 ffo interclavicle ; coracoid very small, co^5ssified with scapula ; not reaching sternum ; Eutheria. The Prototheria have an existing order, the Monotremata, and it is apposed, with much probability, that two orders which appear in the ['rias and continue until the Eocene, inclusive, belong to it. If this lassification is admitted, the Eutheria have their first representatives Q the Postcretaceous, and their latest are now numerous. Of the Prototheria, there are prob- bly three orders of which species are :nown, but the location of the two first numerated below is not certain. ncisors reduced ; molars with com- pressed cutting crowns, and un- divided roots ; Protodonta. ncisors much enlarged ; molars with tubercular grinding surfaces, and distinct roots ; Multituherculata. \o true teeth at maturity ; Monotremata, The families are the following : \otodonta ; Dromatheriidse. fultituberculata ; Tritylodontidse, Pla- giaulacidse, Chirogidse, Polymastodont- idse. fonotremata ; Ornithorhynchidse. The geological range of these orders is shown by the table on next age. Fig. 35.— Piilodus medicevus Cope ; mandibular ramus ; a,b, \ ; c, |. Original. From Puerco bed of New Mexico. Fig. SQ.—Chiroz plicalus Cope ; palate and molar teeth, from below, 3-2 natural si»e. From nerco bed of New Mexico. Original. 66 Plistocene . . Protodonta M ult itub erculata Monotremata Neocene . . . Eocene . . . Cretacic . . . Jurassic . . . Triassic . . . Carbonic . . Devonic . . . biluric . . . Cambric . . . Huronic . . Species of Protodonta have been found thus far only in the Triassic beds of North Carolina. Their molar teeth bear some resemblance to those of some of the corresponding teeth of the Theromorous reptilian genus Dimetrodon. The oldest Multituberculata appear in the Trias of South Africa in the Karoo beds, and in the Upper Trias (Keuper) of Wiirtemburg (Europe). They are more abundant in the Jurassic of England and of the Rocky Mountain region of America, and in the later Cretaceous of North America and of Patagonia. In the latest Cretaceous (Puerco) of New Mexico the largest forms occur, which equal an adult kangaroo. The last of them are found in the lowest Eocene in the north of France. Extinct and recent Mono- tremata are restricted to the Australian realm. The EuTHERiA are represented by the following numerous orders : I. Marsupial pelvic bones (generally) ; palate perforated ; (vagina double ; placenta wanting ; corpus callosum rudimental ; cere- bral hemispheres small.) (Didelphia.) 67 One deciduous molar tooth ; (1) Marsupialla. II. No marsupial bones ; palate generally entire ; (one vagina ; placenta and corpus callosum well developed.) (Monodelphia.) A. Posterior limbs wanting, or represented by minute rudi- ments ; anterior limbs oar-like. {Mutilata.) Elbow joint inflexible ; carpals discoid, and, with the phalanges, separated by cartilage ; lower jaw without ascending ramus ; (2) Cetacea. Elbow joint flexible ; carpals and phalanges with close articulations; mandible with ascending ramus ; (3) Sirenia. AA. Posterior limbs present ; ungual phalanges compressed and curved on one or all the feet.* ( Unguiculata.) jS, Carpal and tarsal bones generally in linear series. y, Teeth without enamel ; no incisors. Limbs ambulatory ; hemispheres small ; (4) Edentata. YY, Teeth with enamel ; incisors present. No postglenoid process; mandibular condyle not transverse ; mastication proal ; limbs not volant ; hemispheres small ; (5) Glires. Anterior limbs volant ; hemispheres small ; (6) Chiroptera. A postglenoid process ; mandibular condyle trans- verse ; mastication orthal ; no scapholunar bone ; f hemispheres small, smooth ; (7) Bunotheria, A postglenoid process ; limbs not volant, with a scapholunar bone ; mastication orthal ; hemi- spheres larger, convoluted ; (8) Camivora. /5/5, Carpal and tarsal bones alternating ; faceted. Anterior limbs prehensile ; mandibular condyle and mastication transverse ; (9) Ancylopoda, AAA. Posterior limbs present ; ungual phalanges not com- pressed and hooked. J ( Ungulata.) /5, Carpal, and usually tarsal bones in linear series. § Limbs ambulatory ; teeth with enamel ; (10) Taxeopoda. /S/S, Carpal bones alternating externally ; tarsals in linear series. Limbs ambulatory, median digits longest ; teeth with enamel; (11) Toxodontia^ /5/9/S, Tarsal bones alternating ; carpals linear or reversed diplarthrous. Cuboid bone partly supporting navicular, not in contact with astragalus ; no canine teeth ; (12) Proboscidea.. i3fij3l^, Both tarsal and carpal series more or less alternat- ing ; the distal row inwards. Os magnum not supporting scaphoides ; cuboid sup- porting astragalus; superior molars trituber- cular; (IS) Amblypoda, Os magnum supporting scaphoides ; superior mo- lars quadritubercular ; || (14) Diplarthra, * Except Mesonychidse, some Glires, and posterior feet of some Edentata, f Except Talpa and Erinaceus. X Except in the Hapalidaj. § Except in Dendrohyrax. || Except Pantolestidtc. 68 The geological range of these orders is as follows : ec 1 • O ! 1 1 C5 - 1 1 i 1 to i ! 1 ! ! ) i j Plistocene Neocene § o W Triassic Carbonic Siluric Ordovicic 69 Incisors Incisors The Marsiipialia fall into two suborders, which differ as follows : Polyprotodontia. Diprotodontia. The Polyprotodontia are carnivorous and insectivorous in their habits, and their present range is Australasia and the two Americas. They are the older of the suborders, being represented in the Eocenes of Europe and North America by species of opossums. Genera pro- bably allied occur in the Postcretaceous of North America. No large forms are known. The Diprotodontia are now restricted to Austra- lasia, and the extinct forms belong to the Plistocene of the same region. These include some very large kangaroos, with still larger animals of the genera Nototherium and Diprotodon. The Diprotodon australis Owen was as large as a rhinoceros, but though allied to the kangaroo, was typical of a different family. The feet were plantigrade, and the fore limbs were larger than the hind limbs. The Cetacea are represented by three suborders, which differ as follows : External nostrils at middle of muzzle ; temporal fossae elongate approximated ; teeth ; Archceoceti. External nostrils at base of muzzle ; temporal fossae short, lateral ; teeth ; Odontoceii. External nostrils at base of muzzle ; temporal fossae short, lateral ; no teeth, but transverse horny laminae on the upper jaw ; Mysticeti. Fig. 37. — Peratherium fugax Cope ; anterior p From Lower Neocene (VVhite River bed) of Colorado Original. cf art of skull. X 2. Fig. Gaudry. -Zeuglodon cetoides Owen ; skull, much reduced. From Eocene of Alabama. From The Archaeoceti are few in number, and are chiefly represented by the gigantic Zeuglodon. They are not known before or after the Eocene system, nor out of the geographical area of the Northern Hemisphere. The Odontoceti first appear in the older Neocene, and the oldest family, the Squalodontidse, have the posterior teeth two- 70- rooted, as in the Zeuglodons. The other families with one-rooted teeth were contemporary with them, and are represented in the present seas by numerous types, as the dolphins, sperm whales, etc. The Mysticeti, or whalebone whales, appear in the Lower Neocene, and are still abundant. They include the largest vertebrates. (Fig. 39.) SiRENiA first appear in a very generalized family in the Eocene beds of some West Indian Islands. In the Neocene of Europe and North America forms approaching more nearly to the existing types are not rare, especially in the south of Europe. At present species exist on the shores of all the continents in the warmer lati- tudes. (Fig. 40.) The Edentata have been generally restricted to the Southern Hemisphere, although during the Neocene they ranged as far north as the ^gean Sea in Europe, and to latitude 46° in North America during the Plistocene. They first appear in the Eocene of Pata- gonia, and were extremely common in the Neocene throughout tropi- cal America, where they are still abundantly represented. A few species still remain in the Ethiopean and Paleotropical geographical realms. The Megatheriidse of the South American Neocene were of large and gigantic size, the largest species pertaining to the genus Megatherium, which ranged in North America to South Carolina. The Glyptodontidse were covered with an immovable carapace consist- ing of bony tesserae, somewhat like that of the armadillos. They varied in size from that of a sheep to that of a rhinoceros. Species were abundant in the Neocene of South America, when they also ranged north to Texas, Florida, and Kansas. The most ancient (Eocene) Edentata display traces of enamel on the teeth. The Glires are found in the strata of all regions, but in reduced numbers in the Eocenes in all countries except temperate South America, where they were very abundant at that age. There are four suborders, which differ as follows : I. Fibula not articulating with calcaneum ; ankle and elbow not tongued and grooved ; one pair of incisors in upper jaw. Incisive alveolus not passing into center of ramus of lower jaw; fibula distinct; Hystricomorpha. Incisive alveolus penetrating ramus ; fibula distinct ; Sciuromorpha. Incisive alveolus penetrating ramus ; fibula coossified with tibia ; Myomorpha. II. Fibula articulating with calcaneum ; ankle and elbow tongued and grooved ; two incisors in upper jaw. Inferior incisor penetrating ramus ; fibula coossified ; Lagomorpha. The Hystricomorpha (porcupines, cavies, etc.) are abundant in the Eo- cene of South America, and are present in the Upper Eocene of France. They are more abundant in the Neocene of South America and France, and are present in the Upper Neocene of North America. At present they are cosmopolitans, excepting Australia, but they chiefly abound in SJouth America. (Fig. 41.) The Sciuromorpha only occur fossil in the yland. Original. (Cetacea.) Fi(i. -10, — Halicore il ii gong Ciiv. ; skeleton, much reduced. From Australasian seas. From Cuvier. iSirenia.) 1 Fig. A2.—Plesiarctomys delicatissimus Leidy ; skull. From Eocene of Wyoming. Original. 72 Northern Hemisphere, and are the principal Glires of the Eocene. From that period they became more abundant in both North America and Europe, reaching their highest expression in the beaver, and spread into South America. (Figs. 42-43.) The Myomorpha (mice, rats, etc.) have a few supposed representatives in the Eocene of Patagonia, but only begin in the Neocene in the Northern Hemisphere. The dor-mice are peculiar to European beds, and the pocket-gophers to North American. They cover the earth, including a few forms in Australia, at the present period. (Fig. 44.) The Lagomorpha (rabbits) appear first in the lowest Ne- ocene of North America (White Kiver), and continue to the present day. They appear in Europe at about the same time, but not in South America until the Plistocene. (Fig. 45.) The Chiroptera (bats) includes two primary divisions at the present time, the insectivorous (Animalivora), and the frugivorous (Frugivora), which differ as follows : Crowns of molar teeth with Vs ; those of opposite jaws interlocking in mastication ; Animalivora. Crowns of teeth obtuse, not interlocking in mastication ; Frugivora. The Frugivora (flying foxes) present a degenerate dentition, and they are not known in the fossil state. Animalivora, on the other hand, appear in typical forms in the Lower Eocene (Wasatch) in North America, and the Upper Eocene (phosphorites) in France. The BuNOTHERiA present four suborders, which are characterized as follows : Molars 8 or more; incisors several, with closed roots ; Pantotheria. Molars 7 or less; incisors with closed roots, not enlarged ; fibula distinct ; otic bulla developed ; Creodonta. Molars 7 or less ; canine generally small, incisors generally enlarged ; fibula generally united with tibia ; Insectivora. Fig. 43.—Plesiarctomys delicatissimus Leidy ; a, hu- merus; 6, proximal ends of ulna and radius; c,d, tibia, distal end; e,/, astragalus; all natural size. From Eocene of Wyoming. 73 Molars 7 or less ; incisors few, growing continually from open roots ; Tillodonta. The Pantotheria are the most ancient of the Eutherian Mammalia, appearing rather abundantly in the Jurassic of Europe and North America. They are all of small size. The Creodonta represent Fig. H. — Bnioptychus crassiramis Cope; a from the Lower Neocene of Colorado. Original; pocket-gopher from the Middle Neocene of natural size, o, anterior half of skull, from below ; Oregon, natural size. Original, c, lower jaw b,c, lower jaw; d, tibia and fibula, from front; from above. e, do., from below. the Carnivora in the Postcretaceous and Eocene systems. They have (except in one family) more than one sectorial molar, and these are true molars, and not premolars. They have (except in one family) strong ca- nines. They range in size from that of an opossum (Stypolophus sp.) to that of a grizzly bear (Hemipsalodon sp.). They are found in the horizons mentioned in North America, Europe, and South America. A few species (Hysenodon, etc.) remain over in the Lower Neocene (White River). (Fig. 46.) Undoubted Insectivora appear in the Upper Eocene of France, but supposed members of the suborder occur in the Lower Eocene of the same. In America none are certainly known prior to the Lower Neocene (White River). They are rare in all formations in America, but they are abundant in the Middle Neocenes of Europe, and in later beds, where forms of moles, shrews, and hedgehogs are abundant. The enlargement of the incisors seen in the Insectivora reaches its extreme in the Tillodonta, where they grow from persistent pulps, as in the Glires. These animals first appear in the Postcretaceous of North 75 America,, and are not uncommon in the Eocenes of America, with one form in the Eocene of Switzerland. Later than that age they are unknown. (Fig. 46.^) The Creodonta. are probably the ancestors of the Carnivora and Insectivora, and the Tillodonta of the Glires. The Pantodonta are ancestors of the Creodonta. The time relations of the Bunotherian order are expressed in the following table : Plistocene .... Pantotheria Creodonta lusectivora Tillodonta Neocene .... ? ? Devonic Ordovicic .... True Carnivora are definitely known from the beginning of the Neocene, and it is probable that they already existed in the Upper Eocene in Europe. There are two suborders, which differ as follows : Digits distinct ; posterior limbs free ; Fissipedia. Digits united into paddles by integument ; hind limbs partly enclosed in general integument ; Pinnipedia. The Fissipedia have their closest connection with the Creodonta, where the latter exhibit exceptionally a dentition like the dogs, in the 76 family Miacidse. From the dogs development has pursued two princi- pal lines. The one has tended to an omnivorous diet, and has termina- ted in the bears. The other has tended to an exclusively carnivorous diet, and has ended in the cats, which display the characters of the order in the greatest perfection. True bears appear in the Plistocene, Fig. 48. — Nimravus gomphodns Cope, two-fifths natural size ; left side of skull. From Miocene of Oregon. 3-4, premolars; 1-2, true molars. and cats (Fig. 48) and hyaenas in the Upper Miocene, in the Northern Hemisphere. In the Southern Hemisphere cats do not appear until the Plistocene (Fig. 49), and hyaenas do not occur in the Western Hemisphere. Dogs are very abundant throughout the Neocene, except in the southern continents, where they do not appear till the Plistocene (Fig. 31). The Pinnipedia appear later in geologic time than the Fissipedia, no forms being known of an age prior to the Middle Neocene. The earliest forms are of two types related to the true (earless) seals and walruses respectively. Of the primitive affinities of the Pinnipedia nothing is known, but they are suspected to have had connection with the Creodonta. The Ancylopoda is a group which includes but few species, which have been found in India, Europe, and North America. They pos- sibly appear in Europe in the Upper Eocene, and occur in the succes- sive stages of the Neocene, beyond which they did not continue. Their characters are very peculiar, including long fore legs and short hind ones, with claws and digits like sloths, but dentition like perissodactyle ungulates. The largest species equaled a grizzly bear. 79 The order Taxeopoda includes the following suborders : I. ? No clavicle. Astragalus not interlocked laterally -with the tibia ; fibula not articulating with calcaneuni ; head of astragalus rounded ; canines ; no anapophyses ; Condylarthra. Astragalus not interlocked with tibia ; fibula articulating with calcaneum ; astragalus head flat ; canines ; Litopterna, Astragalus interlocking at side with tibia ; fibula not articulating with calcaneum ; head of astragalus flat ; no canines; Hyracoidea. II. Clavicles present. Incisors growing from persistent pulps ; anapophyses ; Daubentonioidea.^ Incisors with closed roots ; anapophyses ; Quadrumana. Incisors with closed roots ; no anapophyses ; Anthropomorpha. The Condylarthra are characteristic of the Postcretaceous in North America, and the Lower Eocene of both continents. They have not yet been detected in the Soutliern Hemisphere. Their characters con- nect inseparably the Ungulata and Unguiculata divisions, since the carpus and tarsus are like those of the latter, while the ungues are hoofs or semihoofs. The molars are tritubercular, quadritubercular or lophodont. The best known genus is Phenacodus (Fig. 50), which is the synthetic type of all ungulates. In the known Condylarthra there are five digits on all the feet. The Litopterna are only known, so far, from the Cenozoics of South America, ranging throughout the entire series. The teeth present the variations seen in the Perissodactyla, some being bunodont and some lophodont, but the bunodont forms are tritubercular, like the lowest Condylarthra. This group displays a remarkable reduction in the digits as in the Perissodactyla, passing to three and one digit, as in the horse line. Fig. 51. — Necrolemur antiquus Filh. ; a lemur from the Upper Eocene of France, natural siee. From Filhol. The Hyracoidea are only known in the recent state in Africa and Western Asia. Their molar dentition is lophodont, while the incisors approximate somewhat the rodent pattern. In like manner Dauben- tonioidea are only known as recent in Madagascar. They are allied to the lemurs. Quadrumana appear in the Lower Eocenes in North America and Europe as lemurs (Fig. 51). True monkeys do not appear until the •The horny coverings of the terminal phalanges of all but the first digit in the only known genus, are claw-like. 80 Middle Neocene in the Old World, while they are absent from North America, and occur in the Plistocene of South America. The Anthro- pomorpha (Fig. 52) appear first in the Middle Neocene of France and the Upper Neocene of India, but man does not appear until the Plistocene in both America and Europe. The exact stage of the Plistocene at which man's remains have been found is not clearly ascertained, but he was contemporary with various extinct species and genera of Mammalia in both Europe and North and South America. The history of the Taxeopoda may be expressed as follows : x iisiocentj . 1 2 3 4 5 6 Neocene Eocene . Cretacic Jurassic Triassic . . Carbonic . Devonic . Siluric . . Ordovicic . Cambric . Huronic . The ToxoDONTiA are confined to the South and Central American Continents. They appear first in the Eocenes of Patagonia and con- tinue through the Plistocene system. These later representatives of the genus Toxodon reached the size of the largest species of rhinoceros. In the Mesotheriidse the feet are pentadactyle, and there is a clavicle ; in the Toxodontidse the posterior foot is tridactyle, and there is no clav- icle. The teeth are lophodont, and early began to be prismatic, and 81 Fig. 5:i.—Coryphodon elephantopus Cope ; skull, from below, 2-9 natural aize. Original, the Wasatch Eocene of New Mexico. 84 to grow from persistent pulps. This appeared first in the incisors, some of which resemble somewhat those of rodents, and appeared later in the molars. The Amblypoda are, on the other hand, restricted to the Northern Hemisphere, and to the Postcretaceous and Eocene systems, where their remains abound both in Europe and North America. They are the only order of Ungulata with superior molars constructed on a tri- tubercular basis, and the inferior molars on the tuberculosectorial pattern. The earlier forms of the Puerco are the smallest ; those of the Lower Eocene (Suessonian, Wasatch) are next in size (Fig. 53) ; while those of the Middle Eocene (Bridger), which have been found in North America only, are often of gigantic size (Dinocerata) (Fig. 54), and have the cranium adorned with horny processes. The order Proboscidia is unknown prior to Middle Neocene time in the northern part of the Eastern and Western Hemispheres, and late Neocene or Plistocene time in South America. In other parts of the Southern Hemisphere they are unknown, although one of the two living species {Elephas ajrica7ius) is restricted to Africa. The ear- liest forms of Proboscidia are Dinotheria and Mastodons (Fig. 55), which reached huge dimensions. Elephants appear in late Neocene times in India, and spread in later epochs over North America and Europe. In America their remains are not known from south of the Fig. 56. — jElephas indicus Ij.; superior molar tooth, reduced. From Tomes. valley of Mexico. The hairy mammoth (Elephas primigenius Blum.) was a contemporary of prehistoric man in the Old World, and proba- bly in the New. The DiPLARTHRA iucludc the most specialized types of Mam- malia as regards the structure of the skeleton, dentition, and digestive system, but they are inferior to the anthropomorphous suborder of the Taxeopoda in the structure of the brain. There are two suborders, as follows : Feet with the third digit of predominating dimen- sions; distal end of astragalus not forming a ginglymus ; Perissodactyla. Feet with the digits 3 and 4 predominating and sub- equal ; astragalus with a distal ginglymus ; Artiodactyla. The Perissodactyla exhibit various series of forms in which the toes diminish in number, from four in front and three behind, to one on all the feet (the horse). They may be considered under two heads with 86 respect to the structure of their superior molars, viz. : first those in which the external wall of the crown does not form two Vs (Rhino- cerontoidea), and those in which such Vs are present, with the angles directed inwards (Equoidea). Each series includes several families, mostly extinct ; in the former the rhinoceros (Fig. 58) and tapir, and in the latter the horse, are still living. The earliest forms belong to the former division, and some of them (H3^racotherium) have the B Fig. 59.—Aphelops megalodus Cope, one-sixth natural size ; B, inferior view of cranium Original. Hornless rhinoceros from the Loup Fork Neocene, Colorado. superior molars almost quadritubercular (Figs. 34-6). They appear first in the Lowest Eocene. Three-toed horses first appear in the Lower Neocene, and one-toed horses in the Upper Neocene. The Artiodactyla are represented by a great variety of forms, which differ primarily in their dental characters. Thus the oldest type (Pan- 87 tolestoidea) has tritubercular superior molars. Of the remainder, one sericvS (Suoklea) have the mohirs quadritubercular, or more highly tubercular. The remaining types have the tubercles of the molars more or less flattened on one side, so as to give crescentic figures on section (Fig. 59), and are hence called selenodont. In some of these there are five crescents (Anthacotheroidea) ; in the others there are only four. The latter may have most of the premolars simple (Cameloidea) or complex (Booidea). Some types of the former and all of the latter Fig. 59.— Protolabis transmontanus Cope ; skull, one-third natural size. A Cameloid from the Ticholeptus bed of Oregon, a, from left side ; b, from below. Original. lack superior incisor teeth, and have the cuboid and navicular bones coossified. The Booidea culminate in forms with horns or bony processes of the skull, which may be permanent (Bovid^e) or annually shed (Cervidse). These divisions appear in geological time in the order of structural modification as here mentioned. This appearance is repre- sented, together with the divisions of the Perissodactyla, in the following table : 88 O C o o =3 a o o c 89 The above table inchules only tlie most important of the super- tamilies of the Diplarthra. Those of the Perissodactyla are equally distributed on both sides of the Northern Hemisphere, but only modern forms of the Equoidea appear in South America and in late Neocene time. Pantolestoidea have been so far found in North America only, and the other superfamilies are relatively rare in that continent, except the Cameloidea, which are more abundant than in the Old World. This and the Booidea onl}'' appear in South America in the late Neocene. Fig. 60.— Anterior feet of Artiodactyla, with both series of carpals, except in No. 4. From Kowalevsky. No 1, Hippopotamus ; 2, Hyopotamus ; 3, Dorcatherium ; 4, Gelocus; 5, Cervus. As in the Perissodactyla there is a reduction of the digits in most of the lines of the Artiodactyla, which reaches its extreme in the Cameloidea, and in the Booidea. In the most specialized types of these super- families but two digits remain, and the metacarpals of these are fused into a single bone, the " cannon bone." This structure first appears in time in the latest Neocene. (Fig. 60.) The families embraced in the orders of Eutherian Mammalia are the following : Marsupialia ; (Polyprotodontia) ; Triconodontidae, Amphitheriidse, Myrmecobiidse, Dasyuridse, Didelphidie, Peramelidse ; (Diproto- dontid); Phascolomyidse, PhalangistidsG, Tarsipedidse, Diprotodon- tidse, Macropidse, Thylacoleonidse. Cetacea ; (Archceoceti) ; Zeuglodontidas ; ( Odontoceti) ; Squalodontidje, Platanistidse, Physeteridaj, Delphinidse, (3Iystacoceii) ; Balsenidse. SiRENiA ; Prorastomidie, Halitheriidse, Manatidie, Halicoridse, Rhy- tinidae. 90 BuNOTHERiA ; (Pantotherid) ; Amblytheriidse ; Creodonta) ; Meso- nychidae, Esthonychidie, Arctocyonidse, Miacidse, Hysenodontidse, Leptictid^, Centetidse ; (Lisectivora) ; Galeopithecidse, Tupseidae, Solenodontid^e, Macroscelididse, Talpidse, Adapisoricidse, My- thomyidae, Scalopidte, Chrysochloridse, Erinaceidse, Myogalidse, Soncid2e ', (Tillodonta) ; Tillotheriidse, Ectoganidse, Stylodontidae. Edentata ; Orycteropodidse, Manidse, Bradypodidse, Megatheriidse, Myrmecophagidse, Dasypodidae, Glyptodontidse. Glires ; (Sciuromorpha) ; Sciuridse ; (3fyomorpha); Dipodidse, Muridse, Myoxidte, Saccomyidse, Microtidse, Lophiomyidae, Bathyergidse ; (Hystrlcomorphce) ; Paradoxomyidse, Hystricidse, Echinomyidse, Octodontidse, Capromyidse, Caviidse, Chinchillidse ; (Lagomorpha) ; Lej^oridse. Chiroptera ; (^Animalivora) ; Phyllostomidae, Desmodontidae, Rhino- lophidse, Noctilionidae, Vespertilionidse, Emballonuridse ; {Frugiv- ora); Pteropidae. Carnivora ; (Fissipedia) ; Cercoleptidae, Procyonidae, ^luridae, Can- idae, Bassarididae, Mustelidse, Protelidae, Arctictidae, Viverridse, Cynictidae, Suricatidse, Cryptoproctidae, Nimravidae, Felidae, Hyae- nidae ; (Pinnipedia) ; Phocidse, Otariidae, Odobaenidae. Ancylopoda ; Chalicotheriidae. Taxeopod A ; ( Condylarthra) ; Periptychidae, Phenacodontidae, Menis- cotheriidae ; {Litopterna) ; Proterotheriidae, Macraucheniidae, Astra- potheriidae ; (Hyracoidea) ; Hyracidae ; {Daubentonioidea) ; Chi- romyidae ; ( Quadrumana) ; Mixodectidae, Adapidae, Anaptomor- phidae, Tarsiidae, Lemuridae, Hapalidae, Cebidae, Cercopithecidae ; (Anthropomorpha) ; Simiidae, Hominidae. ToxoDONTiA ; Atryptheriidae, Interatheriidae, Protoxodontidae, Meso- theriidae, Xotodontidae, Toxodontidae. Proboscidia ; Dinotheriidae, Elephantidae. Amblypoda ; (Taligrada); Pantolambdidae ; (Pantodonta) ; Cory- phodontidae ; (Dinocerata) ; Uintatheriidae. DiPLARTHRA ; (Perissodactylo) ; Lophiodontidae, Triplopidae, Caeno- pidae, Hyracodontidae, Rhinoceridae, Tapiridae, Lambdotheriidae, Menodontidae, Palaeotheriidae, Eqiiidae ; (Artiodadyla) ; Pantoles- tidae, Eiirytheriidae, Anoplotheriidae, Dichobunidae, Caenotheriidse, Anthracotheriidae, Xiphodontidae, Suidae, Hippopotamidae, Mery- copotaraidae, Dichodontidae, Oreodoiitidae, Poebrotheriidae, Proto- labididae, Camelidae, Eschatiidae, Tragulidae, Moschidae, Bovidae, Cervidae. ONTA. BY E. D. COPE.