Peruvian expeditions, 1311-1915. Publications no. 22. I f far t S Gato*^' . Y&Lii-^M^iEiasirirY- TRANSACTIONS OF THE CONNECTICUT ACADEMY OF ARTS AND SCIENCES Incorporated A. D. 1799 YOLUME 18, PAGES 291-345 APRIL, 1914 / Hepaticae : Yale Peruvian Expedition of 1911 BY ALEXANDER W.EVANS, Ph.D., EATON PROFESSOR OF BOTANY IN THE SHEFFIELD SCIENTIFIC SCHOOL OF YALE UNIVERSITY M YALE UNIVERSITY PRESS NEW HAVEN, CONNECTICUT 19 14. TRANSACTIONS OF THE CONNECTICUT ACADEMY OF ARTS AND SCIENCES Incorporated A. D. 1799 VOLUME 18, PAGES 291-345 APRIL, 1914 Hepaticae : Yale Peruvian Expedition of 1911 BY ALEXANDER W. EVANS, Ph.D., EATON PROFESSOR OF BOTANY IN THE SHEFFIELD SCIENTIFIC SCHOOL OF YALE UNIVERSITY M YALE UNIVERSITY PRESS NEW HAVEN, CONNECTICUT 1914 WEIMAR : PRINTED BY R. WAGNER SOHN CONTENTS PAGE INTRODUCTION . 295 SPECIAL PART 298 Marchantiaceae ....... . . 298 Plagiochasma ......... 298 Marchantia ........ . 299 Metzgeriaceae . ......... 299 Metzgeria ........... 299 Jungermanniaceae .......... 302 Plagiochila ........... 302 Lophocolea .......... 313 Radula ... 313 Porella 314 Microlejeunea ........ . 314 Taxilejeunea .......... 315 Dicranolejeunea .......... 316 Brachiolejeunea .... ..... 325 Marchesinia .......... 326 Omphalanthus .......... 326 Frullania 326 INDEX TO GENERA AND SPECIES . .... 345 V.— HEPATICAE: YALE PERUVIAN EXPEDITION OF 1911 By Alexander W. Evans, Ph.D. INTRODUCTION The Yale Peruvian Expedition of 1911, under the direction of Professor Hiram Bingham, carried on extensive explorations in the southeastern part of Peru. Although the main purpose of the Ex pedition was the accumulation of data on the Geology and Archeology of the region, collections of invertebrate animals and of plants were made by Professor Harry Ward Foote, one of the members of the party. Instead of attempting to collect indiscriminately he wisely restricted his attention to a few definite groups, and the group of the Hepaticae was among those selected. The material of this group comprises thirty-five packets, several of which contain an admixture of two or more species. Thirty-one species in all, representing fourteen genera, are in a condition to be identified. The specimens came from seven different localities, namely: Cuzco (one species), Huadquina (three), Lucma (five), Ollantaytambo (two), San Miguel (sixteen), Santa Ana (nine), and Urubamba (one). The following information about these localities has been furnished by Professor Foote. Cuzco, Ollantaytambo, Santa Ana, and Urubamba are all well-known towns, the last three being situated on the Urubamba River, which flows in a general northwesterly direction. Cuzco is a few miles distant from the river and nearer its source. San Miguel is a small district twenty or twenty-five miles below Ollantaytambo and is not to be confused with the town of San Miguel in the north western part of Peru; Huadquina is a large estate lying a few miles below San Miguel; while Lucma is a town on the Vilcabamba River, a small branch of the Urubamba entering from the west a few miles south of Santa Ana. Cuzco, at an elevation of 11,500 feet, is well above the forest line, although a single species of tree was observed there. Urubamba, Ollantaytambo, and Lucma, at elevations varying from 7,000 to 9,500 feet, are also above the true forest line, in spite of the fact that several species of trees are of occasional occurrence. San Miguel, Huadquina, and Santa Ana, at elevations varying from 3,000 to 6,000 feet are all in a region which is either forested now or has been in former times. The primary growth 296 Alexander W. Evans, contains a large proportion of trees, but the secondary growth con tains few trees and many bushes. There seems to be little tendency for the original species to reappear after deforestation, even if the land is left completely uncultivated. The entire region explored has a long dry season and suffers from lack of water. The hepatic flora of the Andes, except at very high altitudes, is rich and varied. It includes some of the largest and most conspic uous of the leafy forms, a great many smaller and less striking spe cies, and a fair proportion of thallose representatives. The number of endemic genera is exceedingly small, Myriocolea Spruce (with one species), Mytilopsis Spruce (also with one species), and Stephaniella Jack (with two species) being the only ones known at the present time. Most of the Andean species belong to large and widely distrib uted genera, such as Plagiochila, Radida, Poretta, and Fridlania. The Lejeuneae, also, as in most tropical regions, are numerous and diverse. Certain genera, such as Gymnomitrium, Marsupella, Lopho- zia, and Scapania, which are so richly developed in the arctic and alpine districts north of the tropics, are either absent altogether or very sparingly represented. In the present collection there are three thallose species (one Plagiochasma, one Marchantia, and one Metzgeria), nine Plagiochilae, five Frullaniae, and nine members of the Lejeuneae. The remaining five species belong to the genera Lophocolea, Radula, and Porella. Six of the species are apparently undescribed ; the others include a number of widely distributed species and a somewhat smaller number confined to the Andes. The geo graphical distribution will be more fully considered in connection with the individual species. The type-specimens of the new species will be deposited in the herbarium of the writer at New Haven, Connecticut. Several important works dealing wholly or in part with the Hep aticae of the mountainous portions of Mexico, Central America, and South America, have been published, although nothing has yet appeared which deals exclusively with the Hepaticae of Peru. The following works have been of especial service in the study of Pro fessor Foote's material, although numerous shorter papers have of course been consulted: Montagne, C. Florula Boliviensis. In d'Orbigny, A., Voyage dans l'Arnenque Meridionale. 72: 1-119, pi. 1-3. 1839. The Hepaticae treated in this work were, for the most part, collected by Alcide d'Orbigny during the years 1826-1833, in various parts of South America. Among the species discussed forty-five came from Bolivia, ten of which at that time were supposed to be endemic. Seven Hepaticae : Yale Peruvian Expedition of iprr. 297 of the Bolivian species are figured. The Florula is the first important contribution to our knowledge of Andean Hepaticae and is even now indispensable to the student. GOTTSCHE, C. M., LlNDENBERG, J. B. G., AND NEES VON ESENBECK, C. G. Synopsis Hepaticarum. Hamburg, 1844—1847. This is the first general account of the Hepaticae of the world and includes numerous species from various parts of the Andes. Gottsche, C. M. De Mexikanske Levermosser. Kongl. Danske Vidensk. Selsk. Skrift. V. Naturv. og Math. Afdel. 6: 97-380, pi. 1—20. 1863. This important work is based on the collections of F. Liebmann, made in various parts of Mexico but mostly in mountainous regions. About two hundred species are described, quite a number of which range beyond Mexico into Central and South America. Most of the plates, which are of a high degree of excellence, are devoted to the large and difficult genus Plagiochila. Gottsche, C. M. Hepaticae. In Triana, J., & Planchon, J. E., Prodromus Florae Novo-Granatensis. Ann. des Sci. Nat. Bot.V. 1: 95—198, pi. 1 7—20. 1864. Most of the species enumerated in this work were collected by Alexander Lindig, during the years 1859—1861, in the province of Bogota and in other parts of Colombia. With two exceptions the species figured belong to the genus Plagiochila. Spruce, R. Hepaticae of the Amazon and of the Andes of Peru and Ecuador. Trans. Bot. Soc. Edinburgh 15: xi + 588, pi. 1—22. 1884 —1885. This remarkable volume must always remain a model for the student of taxonomic hepaticology. It is particularly valuable for its account of the Lejeuneae. The work is based on collections made by its author during the years 1849—1862. The species figured are selected from the various subdivisions of the Hepaticae. Spruce, R. Hepaticae Bolivianae. Mem. Torrey Club 1: 113— 140. 1890. The collections forming the basis for this article were made by H. H. Rusby in eastern Bolivia. Jack, J. B., and Stephani, F. Hepaticae Wallisianae. Hedwigia 31 : 11— 27, pi. l— 4. 1892. The Hepaticae enumerated in this paper were collected by Gustav Wallis, between the years 1860 and 1878, some of them in the Philippines but the majority in Colombia and Peru. Stephani, F. Species Hepaticarum. 1898—1914. The latest work dealing nith the Hepaticae of the world. Four large volumes have already appeared and the fifth, which will probably be the last, is in course of publication. The work is issued under the auspices of the Herbier Boissier at Geneva, the earlier parts having been printed in their Bulletin and Memoires and the later parts as a Com plement to the Bulletin. 298 Alexander W. Evans, SPECIAL PART MARCHANTIACEAE I. Plagiochasma Lehm. & Lindenb. 1. Plagiochasma chlorocarpum (Nees & Mont.) Mont. Reboulia chlorocarpa Nees & Mont. Ann. des Sc. Nat. Bot. II. 5: 70. 1836. Plagiochasma chlorocarpum Mont. Fl. Boliv.; d'Orbigny, Voy. dans l'Amer. Merid. T: 59. 1839. Rupinia chlorocarpa Trevis. Mem. R. 1st. Lomb. III. 4 : 437. 1877. On earth, Ollantaytambo, 9,000 feet, July 31, 1911. The species of Plagiochasma are in need of considerable study before they can be considered well understood. The determination of the present specimens, therefore, must be regarded as provisional, more especially as the few capsules present are not in good condition. P. chlorocarpum is one of the species characterized by minute epider mal pores. The type material was collected by Bertero, in Chile, in 1828, and the species has been reported also in Gay's collections from the same country. Apparently no other stations have been recorded. Soon after the publication of P- chlorocarpum a second species of the genus, likewise based on specimens collected by Bertero in Chile, was proposed by Bischoff under the name P. validum.1 This species, which is recognized as valid by the authors of the Synopsis Hepaticarum and also, much more recently, by Stephani, has been reported from Bolivia as well as from Chile. Two Bolivian stations have been recorded : Sorata, Mandon, by Stephani,2 and the vicinity of Mapiri, Bang, by Rusby.3 If the descriptions given by Stephani of these two species are carefully compared no essential differences be tween them are apparent, and it is possible that P. validum represents a synonym of P chlorocarpum. Unfortunately the question can not be settled at the present time, although the writer has examined a specimen of Bischoff's plant in the herbarium of the British Museum and has also seen Bang's specimens in the herbarium of the New York Botanical Garden. The latter are apparently the same as those listed above from Peru, but the Bischoff specimen is sterile and so fragmentary that no definite conclusion can be drawn from it. 1 Handb. Bot. Term, und Systemk. 2 : 56. pi. 56, f. 2753. 1842. 2 Bull, de l'Herb. Boissier 6 : 785. 1898. 3 Mem. Torrey Club 6 : 129. 1896. Hepaticae : Yale Peruvian Expedition of 1911. 299 The close relationship between P. chlorocarpum and P. peruvianum Nees & Mont, has already been commented upon by Montagne.1 P. chlorocarpum may be distinguished, however, by its dioicous in florescence and by its lack of apical innovations, the branching being either by forking or by means of lateral adventive shoots. In P. peruvianum the inflorescence is monoicous and apical innovations are frequently seen. The original material of P peruvianum was collected by d'Orbigny in Bolivia and not in Peru, and no additional collections have been reported. II. Marchantia L. 2. Marchantia lamellosa Hampe & Gottsche Marchantia lamellosa Hampe & Gottsche ; G. L. & N. Syn. Hep. 527. 1846. On a dry earth bank, Cuzco, 11,500 feet, July 6, 1911. The present specimens include two gemmiparous plants only and are referred to M. lamellosa on account of the conspicuous scales, which are deep purplish black in color and almost cover the ventral surface. Unfortunately the absence of sexual branches and of sporophytes makes a positive determination impossible. M. lamellosa was based on material collected by Moritz in Colombia. It has recently been recorded from Ecuador by Stephani, but apparently no other stations are at present known. METZGERIACEAE III. ffietzgeria Raddi 3. Metzgeria scyphigera sp. nov. On apple wood, Urubamba, 9,500 feet, July 17, 1911. The same species was collected by Barnes & Land at La Cima, Mexico, in lava fields (but on wood), at an altitude of 9,800 feet, on October 14, 1908, specimens received from W. G. Farlow. The Peruvian material may be designated the type. Yellowish green, sometimes tinged with blue, growing in depressed mats : thallus when normally developed prostrate, repeatedly dichotomous, well- developed branches about I mm. wide, mostly 0.9 — 1.5 mm. long between 1 Fl. Boliv. ; d'Orbigny, Voy. dans l'Amer. Merid. 72 : 60. 1839. 300 Alexander W. Evans, the forks, plane or slightly convex along the margins ; costa bounded both above and below by two rows of cells ; wings mostly fifteen to twenty cells wide, the cells thin-walled but with very minute trigones and occasional intermediate thickenings, averaging about 40 X 3° H- and about 3° 1" along the margins; hairs restricted to the margin and to the ventral surface of both wings and costa; marginal hairs some times abundant, sometimes very scanty, straight or nearly so, mostly 100 — 120^ in length ; alar surface hairs and costal hairs similar but a Uttle longer, the surface hairs never abundant and often absent altogether : inflorescence dioi- cous : ^ t>ranch (known only from the type specimen) obcordate, orbicular- obovate, about 0.15 mm. long, bearing a few straight marginal and surface hairs : $ branches (known only from the Mexican specimens) sparingly produced, globose, about 0.3 mm. long, usually smooth throughout : gemmae numerous, marginal, usually borne on more or less ascending and tapering branches, hmited in growth and destitute of hairs (except slime-papillae), orbicular or oblong- orbicular, about 0.15 mm. wide, concave, about six cells across, bearing a few rudimentary marginal cilia slightly displaced to the concave surface, one cell thick throughout and with a poorly defined stalk. (Fig. 1.) As pointed out by the writer a few years ago the gemmae in the genus Metzgeria1 often yield char acters of specific value, and some times make it possible to distinguish between closely related species. In M. scyphigera the gemmae are among the simplest that have been described. In their development they agree in all essential respects with the marginal gemmae of M. furcata (L.) Dumort., M. myriopoda Lindb., and M. oligotricha Evans, a marginal cell becoming directly the mother cell of a gemma without undergoing a preliminary division. Figure 1.— Metzgeria scyphigera Evans A. Tip of a gemmiparous thallus, X 40. B. Gemma at time of separation, X 60. The figures were drawn from the type specimen. 1 Vegetative Reproduction in Metzgeria. Ann. Bot. 24 : 271—303. f. 1—16. 1910. Hepaticae : Yale Peruvian Expedition of ipu. 301 A gemma at maturity consists of a circular or slightly elongated plate of cells with a single apical cell and a short and often indistinct stalk composed of two cells (Fig. 1, B). It is further characterized by being distinctly concave and by bearing a few rudimentary mar ginal hairs, slightly displaced to the concave surface. The germination of a gemma is very similar to that described for other species. The concave surface becomes attached to the sub stratum by means of the rudimentary hairs, which grow out into short rhizoids. The apical cell of the gemma then resumes its divi sions and gives rise to a flat strap-shaped thallus, which is at first no wider than the gemma but which gradually becomes wider and differ entiates a costa. In rare cases the young plant undergoes a branch ing by an apparent dichotomy before the costa makes its appearance. It is often quite impossible to determine where the gemma ends and the young thallus begins, but sometimes there is a constriction at the junction between the two. In some cases new marginal gemmae arise, either on the original gemma itself or on the young thallus while still in an undifferentiated condition. The gemmiparous branches (Fig. 1, A) tend to differ more or less from normal branches. They not only curve away from the sub stratum but totally lack hairs on both wings and costa. At the same time the wings grow narrower and narrower until they may be only three or four cells wide. With this tapering of a branch its growth soon comes to an end. These modifications resemble somewhat those found in the gemmiparous branches of M. fruticulosa (Dicks.) Evans, but apparently they never pass beyond the stages described and never reach the striking conditions sometimes seen in M. fruticulosa, where the branch may lose its wings altogether and become reduced to a radial costa bounded by cortical cells and growing at right angles to the substratum. Whether these apparent differences are actual or merely due to an insufficient supply of the new species can hardly be answered at the present time. In his Species Hepaticarum Stephani1 describes two dioicous South American species in which the scattered marginal hairs occur singly. They agree further with M. scyphigera in the costa, which is bounded both above and below by two rows of marginal cells and which bears hairs on its lower surface. One of these species is M.chilensis Steph., of Chile and New Zealand, and the other is M. aurantiaca Steph., of Brazil. In M. chilensis, however, the thallus is almost revolute, the wings are smooth on the lower surface, the cells measure 54 X 36 u, 1 Bull, de l'Herb. Boissier 7 : 937, 938. 1899. 302 Alexander W. Evans, and there are no trigones. In M. aurantiaca the wings bear scattered hairs as in M. scyphigera, but the thallus is strongly convex, the wings are only twelve cells wide, and there are no trigones. Stephani makes no mention of gemmae in any of his descriptions. JUNGERMANNIACEAE IV. Plagiochila Dumort. The genus Plagiochila is one of the largest genera of Hepaticae, nearly eight hundred species being recognized by Stephani in his re cent monograph.1 With but few exceptions these species are trop ical, and many of them seem to be exceedingly local in their distrib ution. In spite of their considerable size the species are extremely difficult to determine, the differential characters being often based on the type of branching present, the shape of the leaves, the number and peculiarities of their teeth, the size of the leaf -cells, and the nature of the local thickenings in their walls. The bracts and perianths also yield characters of importance, but the determination of speci mens must often be made without the assistance of these organs; in fact they are known in comparatively few species. It will be seen at once that most of the differential characters used are subject to more or less variation, and this of course adds appreciably to the difficulties of gaining a correct comprehension of the species. The present collection contains nine Plagiochilae in sufficient quantity to warrant determination; five of these are referred to described species, and the others are proposed as new. It is to be hoped that this course will not add to the confusion already prevalent in the genus. 4. Plagiochila alternans Lindenb. & Gottsche Plagiochila alternans Lindenb. & Gottsche; G. L. & N. Syn. Hep. 648. 1847. On earth, San Miguel, 5,000-6,000 feet, July 24, 1911; on a damp rock, Lucma, 7,000 feet, August 7, 1911. Plagiochila alternans was originally described from specimens collec ted by Liebmann in the province of Oaxaca, Mexico. It was after wards reported by Spruce from Bolivia, Rusby, and Costa Rica, Hyde, and by Stephani from Venezuela, Funck. The Peruvian specimens agree closely with the published descriptions and also 1 Bull, de l'Herb. Boissier II. 2—6. 1902— 1906., Hepaticae : Yale Peruvian Expedition of 1911. 303 with Gottsche's1 figures, except that the teeth on the leaves are more numerous, numbering from sixty to eighty on robust leaves instead of from twenty-five to forty. This is apparently due to the unusual development of the plants and is hardly sufficient to justify their separation as a distinct species. The present specimens bear a strik ing resemblance to large forms of P. asplenioidbs (L.) Dumort. and especially to the variety major; in the common northern species, however, trigones are always present in the leaf-cells and the teeth are shorter and less spine-like. 5. Plagiochila andicola Mont. & Gottsche Plagiochila andicola Mont. & Gottsche; Montagne, Ann. des Sc. Bot. IV. 6 : 187. 1856. On live wood, San Miguel, 6,000 feet, July 24, 1911. The sterile type material of P. andicola was collected by Jameson at Quito, Ecuador, and apparently no other stations have been re corded. The present specimens agree with the published descriptions except in two particulars, the plants being a little less robust and the leaf-cells showing slightly larger measurements. According to the descriptions the stems are 7—15 cm. long, the apical leaf-cells measure 18 u, and the basal cells average 36 x 27 fi. In the Peruvian plants the most robust stems are only 4 cm. long, the apical leaf -cells average about 25 u, while the basal cells are about 45 x 28 a. These differ ences, however, are slight and the measurements compared are variable, so that they afford no adequate basis for a specific separation. Most of the Peruvian plants are sterile or reproduce by abundant propagula of the usual type. But a few show male inflorescences, which sometimes occur at the base of a branch and sometimes at or near the middle. In other words the male branch always produces normal leaves beyond the perigonial bracts. The latter, in the few cases observed, are in about six pairs and are closely imbricated. Instead of spreading obliquely, as do the normal leaves, they are sub- erect and only the apical portion is more or less squarrose. As in all typical species of Plagiochila the antical basal portion is strongly inflated and delicate in texture. The rest of the bract approaches a normal leaf in texture, except that the cells are a little smaller (averaging about 16 u m the apical portion), and the thickenings of the walls are less pronounced. When a well-developed bract is spread out it is seen to be broadly ovate and strongly unsymmetrical 1 Mex. Leverm. -pi. 7. 1863. 304 Alexander W. Evans, on account of the dilated basal portion. It measures about 1 X 0,9 mm., and the three to five blunt teeth are restricted to the apical region. Two or three antheridia are found in the axil of each bract. Closely related to P. andicola, as Gottsche notes, is the Mexican P. truncata Gottsche,1 originally described from specimens collec ted by Liebmann in the Sempoaltepec region and afterwards found by other collectors. In this species the apices of the leaves tend to be truncate, although in some of Gottsche's figures this condition is not pronounced, and the teeth are longer and sharper than in P. andicola. In P. subconvoluta Gottsche2, another Mexican spe cies known only from a single specimen found by an unknown collec tor, the leaves are much like those in P. andicola and P. truncata, but the postical bases are not dilated in such a way that they forms crests. 6. Plagiochila Binghamiae sp. nov. On rocks, Huadquina, 5,000 feet, July 30, 1911. Named in honor of Mrs. Alfreda Mitchell Bingham, a patroness of the Expedition. Yellowish or brownish green, not glossy, growing in loose mats : stems mostly 2 — 3 cm. long and 0.3 mm. in diameter, more or less rigid, ascending, simple or with an occasional branch, the branches obliquely spreading, almost invariably intercalary, similar to the stem but often with somewhat smaller leaves : leaves alternate, contiguous or loosely imbricated, widely spreading (at an angle of about 75 degrees), somewhat convex, ovate, 1.5 — 2.4 mm. long, 0.9 — 1.2 mm. wide, narrowly decurrent both antically and postically, antical margin somewhat revolute toward the base, slightly incurved to slightly outcurved, entire, postical margin more or less involute in decurrent portion and for a variable distance beyond, sometimes as far as the middle, outwardly curved but not dilated at the base, entire or with a single tooth in the outer part, apex broad and rounded with three or four scattered teeth, the latter irregular, mostly four or five in all, usually three or four cells long and three or four cells wide at the base, abruptly pointed ; leaf-cells averag ing 17^ in the apical region, 28 X 19," in the middle, and 32 X 23 /j, at the base, walls thin but with distinct triangular trigones, one or two of the sides usually bulging, intermediate thickenings few, oval, thickenings often con fluent, especially toward the base and along the margin ; cuticle smooth : underleaves minute, about 0.1 X 0.07 mm., deeply cleft almost to the base into two to four cilia: propagula numerous, with small leaves: remaining parts not seen. Fig. 2. Although most of the branches in this species are intercalary a terminal branch is occasionally produced and is of considerable 1 Mex. Leverm. 127. pi. 10. /. 1—4. 1863. 2 Mex. Leverm. 120. pi. 3, f. 1—10. 1863. Hepaticae : Yale Peruvian Expedition of 1911. 305 Figure 2.- -Plagiochila Binghamiae Evans theoretical interest . Apparently here, as in so many species of Plagio chila, terminal branches have been supplanted by intercalary branches, and the appearance of a terminal branch is to be looked upon as a reversion. It is perhaps associated with a vigorous growth and a retention of a juvenile condition, much as in the cases described by the writer under Cephalozia bicuspidata (L.) Dumort. and Lophocolea heterophylla (Schrad.) Dumort.1 Closely related to P- Bing hamiae is a sterile Ecuador plant collected by Spruce and described by him under the name P. hypan- tra.2 In this species terminal branches although infrequent are the rule rather than the excep tion. The leaves are similar in form to those of P. Binghamiae but the teeth, especially those in the apical region, are longer and sharper The leaf-cells are characterized by less robust thickenings. In P. patentissima Steph.,3 a Peruvian species collected by Lechler at St. Gavan and known to the writer from description only, the leaves bear from eleven to thirteen spines, those near the apex being especially large and acuminate. It differs further from P. Bighamiae in the size of the leaf-cells, those in the middle of the leaf measuring 54 X 27 u and those at the apex 36 X 27 u. 7. Plagiochila chinantlana Gottsche Plagiochila chiantlana Gottsche, Mex. Leverm. 12. pi. 5, f. 1—7. 1863. San Miguel, 5,000-6,000 feet, July 24, 1911. The Peruvian material agrees pretty closely with the specimens collected by Pringle at Patzcuaro, Mexico, and distributed by Under wood and Cook in their Hepaticae Americanae, No. 110. The species was first collected in the province of Oaxaca, Mexico, by Liebmann. According to Spruce it has been found also by Cross at San Sebastiano, Colombia, but Stephani does not record its occurrence outside of Mexico. A. Leaf, dissected from the stem, X 17. B. Cells from the middle of a leaf, X 300. C. Tooth from near the apex of a leaf, X 300. The figures were drawn from the type specimen. * Ann. Bot. 26: 33. .19x2. -Hep. .Amaz.-et And. 465. 1885. 3 Bull, de l'Herb. Boissier II. 2 : 877. 1902. 306 Alexander W. Evans, 8. Plagiochila flavescens Gottsche Plagiochila Guilleminiana 0 flavescens G. L. & N. Syn. Hep. 644. 1847. Plagiochila flavescens Gottsche, Mex. Leverm. 148. pi. 7, f. 1—3. 1863. On live wood, Lucma, 7,000 feet, August 7, 1911. Originally collected by Liebmann in the province of Oaxaca, Mex ico, and later reported by Jack and Stephani from either Colombia or Peru, in their account of Wallis's Hepaticae.1 The determination of the material from Lucma is based on the figures and description of Gottsche. The plants bear a strong resemblance to P. Guilleminiana Mont., a species listed below, but are somewhat more robust. The leaves, moreover, are relatively broader, the dilated portion at the postical base being very conspicuous and forming striking crests. The leaves measure about 4 mm. in length (including the dilated por tions) and about 3 mm. in width. Leafy propagula are present in abundance. 9. Plagiochila Footei sp. nov. San Miguel, 5,000-6,000 feet, July 24, 1911. Named in honor of Professor Harry Ward Foote, the collector. Pale green, varying to brownish green, not glossy, growing in loose tufts : stems mostly 3—6 cm. long and 0.35 mm. in diameter, more or less rigid, simple or sparingly branched, the branches all intercalary, similar to the stem but sometimes with somewhat smaller leaves: leaves subopposite, distant to subimbricated, obliquely spreading (at an angle of about 45 de grees), strongly convex, broadly ovate, about 2 mm. long and 1.6 mm. wide, narrowly decurrent both antically and postically, the antical margin extend ing a httle farther backward than the postical, antical margin slightly revo- lute toward the base, straight or somewhat outwardly curved, entire, postical margin strongly outwardly curved from the narrow decurrent portions, sharply dentate-ciliate, apex rounded with more crowded teeth than the postical margin ; total number of teeth mostly twelve to fifteen, unequal, acuminate, mostly three to six cells long and one to four cells broad at the base ; leaf-cells plane, averaging 22 u at the apical margin, 35 X 3° f in the middle, and 50 X 25 /j. at the base, walls thin but with minute triangular trigones with concave sides ; cuticle smooth : remaining parts not seen. (Fig- 30 Although at first sight the leaves in this new species appear exactly opposite, closer inspection shows that the decurrent bases on one side of the stem extend a little farther backwards than those on the other, and this is true both antically and postically. The line of insertion 1 Hedwigia 31 : 14. 1892. Hepaticae : Yale Peruvian Expedition of 1911. 307 here, as in allied species, is strongly and narrowly arched, and this peculiarity, associated with the convexity of the leaves, gives the plants a very characteristic appearance. As the leaves become dry they appear more and more revolute and spread from the axis at almost a right angle. When moist they spread more obliquely and show a strong tendency to come into contact postically. Two South American species are closely related to P. Footei. One is P- oresitropha Spruce,1 collected by Spruce in the Andes of Peru and by Rusby in Bolivia, and the other is P. mapiriensis Spruce,2 Figure 3. — Plagiochila Footei Evans A. Leaf, dissected from the stem, X 17- B. Cells from the middle of a leaf, X 225. C. Tooth from the postical margin of a leaf, X 223. The figures were drawn from the type specimen. collected by Rusby in Bolivia. According to Stephani P. mapiriensis is merely a form of the widely distributed P. rutilans Lindenb., but Spruce's description would indicate that he had another species be fore him: In P. oresitropha the general appearance is much the same as in P. Footei, the leaves being nearly opposite and spreading widely from the stem. Yet the species is somewhat more robust, and the leaves are larger and closer together. Spruce gives as measurements for the leaves a length of 2.5—3.25 mm. and a width of 2—2.75 mm., the ratio of width to length being a little higher than in P. Footei. The leaves show further differences in their teeth and in their cell structure. In P. oresitropha the antical margin bears five or six scattered teeth, only the basal portion being entire, while the teeth on the postical and apical margins are more crowded than in the new species. In fact the whole number of teeth on a leaf is approximately 1 Hep. Amaz. et And. 467. 1885. 2 Mem. Torrey Club 1: 131. 1890. Tbans. Conn. Aoad., Vol. XVIII. 21 April, 1914. 308 Alexander W. Evans, forty. The leaf-cells differ but little in size in the two species, but the thickenings in the wall are much more pronounced in P. oresi tropha; the trigones have bulging sides and are often elongated at the angles, there are numerous intermediate thickenings (sometimes two on a long lateral wall), and there is more or less coalescence between the thickenings. In P. Footei, as noted in the description, the trigones have concave sides, and it may be added that intermediate thickenings are practically lacking and that coalescence apparently never occurs. In P. mapiriensis, known to the writer from description only, the leaves are relatively longer than in P. Footei, measuring about 2.5 mm. in length and only 1.1—1.4 mm. in width, and'the cells are smaller, averaging only 29 u in diameter ; in other respects the two species seem to be closely allied, the number of marginal teeth in P. mapirien sis varying from eight to fifteen. 10. Plagiochila Guilleminiana Mont. Plagiochila Guilleminiana Mont.; Lindenberg, Monogr. Hepat. Gen. Plagiochilae 152. pi. 33, /. 1-4. 1842. Plagiochila Haeckeriana Lindenb. & Gottsche ; G. L. & N. Syn. Hep. 644. 1847. Plagiochila oreocharis Spruce, Hep. Amaz. et And. 498. 1884. Plagiochila rhizophila Spruce, I. c. 495. 1884. On live wood, Santa Ana, 3,000 feet, August 4, 1911. The synonymy given above is taken from Stephani's Species Hepa ticarum.1 Accepted in this broad sense the species is variable and has a wide distribution. It was based on two Brazilian specimens, one collected by Raddi and the other by Guillemin, and was soon afterwards reported from Mexico in the collections of Liebmann. The type-locality of P. Haeckeriana is not given definitely by its authors but was apparently somewhere in the Andes, while P. oreo charis and P. rhizophila were based on material collected by Spruce, the first in Peru and the second in Ecuador. According to Stephani P. Guilleminiana has been found also in Guatemala by Wurr and in Colombia by Lindig. In the present specimens the stems are about 6 cm. long and are simply pinnate, the branches being few and irregular and apparently all of the Frullania type.2 The female inflorescences are borne on the tips of some of these branches. The leaves (Fig. 4) have a length 1 Bull, de l'Herb. Boissier II. 5 : 356. 1905. 2 See Evans, Ann. Bot. 26 : 4. 1912. Hepaticae : Yale Peruvian Expedition of ipn. 309 of about 3 mm., a width of about 1.8 mm. in the broadest portion, and of about 0.5 mm. in the apical region. They are, closely imbri cated and long-decurrent both antically and postically; when dissected off and spread out they are seen to be ovate-triangular and markedly unsymmetrical. In their natural position they spread widely from the stem, and those on the opposite sides lie almost in one plane, thus giving the shoot a flattened ap pearance. The antical margin is plane in the apical region and often along a part of the decurrent portion, otherwise it is more or less revolute and is usually slightly incurved. Occasionally it bears one or two small teeth in the outer part but in most cases is entire throughout its length. The postical margin is crispate and strongly dilated at the base, spreading backward from the stem at almost a right angle and thus forming a part of a postical crest. Beyond the di lated portion it extends almost in a straight line to the apex. On robust leaves the dilated portion sometimes bears a series of scattered spine-like teeth, but it is quite entire in many cases, and this condition seems to be the rule on the leaves of branches and of poorly devel oped stems . Beyond the middle the postical margin bears a few scattered teeth. The apex is broad and variously rounded, truncate, or acute, depending somewhat upon the number and position of its scattered teeth. The total number of teeth present on a leaf is usually between five and twenty, the higher numbers being found on leaves where the postical base is spinose-dentate. Most of the teeth are from three to five cells in length. On the perichaetial bracts the teeth are more numerous, numbering from twenty-five to thirty and are found not only along the whole postical margin but along Figure 4. — Plagiochila Guilleminiana Mont. A. Leaf, dissected from the stem, X 17. B. Cells from the middle of a leaf, X 300. C. Tooth from the postical margin of a leaf near the apex, X 225. The figures were drawn from the specimens collected at Santa Ana. 310 Alexander W. Evans, a considerable part of the antical margin as well. Many of these teeth are distinctly spinose in character. The leaf-cells have distinct triangular trigones with bulging sides. Toward the base occasional oval intermediate thickenings make their appearance, and, especially in the antical portion, there is a tendency toward a coalescence of the thickenings. The specimens just described agree in most respects with those distributed by Spruce under the name P. oreocharis, although the latter are a little less robust. The trigones in Spruce's plants, however, are a little larger, and the thickenings show a stronger tendency to coalescence throughout the leaf. Stephani comments on the great variability in the number of teeth found on vegetative leaves, but emphasizes the fact that the perichaetial bracts and the perianths present characters of more constancy. 11. Plagiochila pauciramea sp. nov. On damp earth and wood, Lucma, 7,000 feet, August 7, 1911. Yellowish to brownish green, not glossy, growing in loose tufts : stems mostly 5 — 6 cm. long and 0.45 mm. in diameter, flexuose, ascending, sparingly and irregularly branched, the branches all terminal, of the Frullania type, sometimes irregularly subdivided, obliquely spreading, similar to the stem but often shorter and with smaller leaves : leaves alternate, loosely imbri cated, obliquely spreading (at an angle of about 60 degrees), antical portion more or less convex near the base, otherwise plane or nearly so, ovate, about 3 mm. long and 2.3 mm. wide, long-decurrent antically, more shortly and narrowly decurrent postically, antical margin a little revolute near the base, slightly outwardly curved or nearly straight, entire at the base but bearing from two to four spinose teeth beyond the middle, postical margin narrowly revolute in the decurrent portion and slightly beyond, more strongly out wardly curved, entire at the base but bearing about ten spinose teeth scattered throughout the greater part of its length, apex broad and rounded with about five spinose teeth ; total number of teeth mostly fifteen to twenty, slender and acuminate, mostly four to ten cells long and one to four cells wide at the base, the larger teeth irregularly distributed but tending to be more numerous in the outer part of the leaf ; leaf-cells plane, averaging about 25 fi at the apex, 45 X 30 fi in the middle, and 55 X 30 fi at the base, walls appearing uniformly thickened on account of the fact that the thick outer walls extend down upon the vertical walls, the latter upon careful focusing appearing thin but showing small trigones with straight or slightly bulging sides and very rare oval intermediate thickenings ; cuticle smooth : underleaves about 0.15 mm. long and 0.12 mm. wide, divided almost to the base into four to eight slender cilia : remaining parts not seen. (Fig. 5.) Hepaticae : Yale Peruvian Expedition of ipu. 311 In spite of the few and irregular branches in P. pauciramea, it belongs to Spruce's section Frondescentes and has several relatives among the species already known from South America. P. montana Spruce, from Campana, Peru, is one of these, but in this species the leaves narrow more markedly toward the apex than in P pauciramea, the spines are shorter and broader, the walls of the leaf-cells do not appear uniformly thickened, and there are no conspicuous underleaves. In P amazo- nica Spruce, another allied species of the Amazonian Plain, extending up the eastern slopes of the Andes, the margin of the leaves is spinose throughout, even at the antical base, and the spines are more numerous, numbering about forty in all. In this species, further, the leaves are subtriangular in outline, both antical and postical margins tending, to straight in the outer part. Both P. montana and P. amazonica are known to the writer by Spruce. Figure 5.— Plagiochila pauciramea Evans A. Leaf, dissected from the stem, X 17. B. Cells from the middle of a leaf, those in the lower part of the figure drawn at a lower focus, X 225. C. Tooth from near the middle of the postical margin of a leaf, X 225. The figures were drawn from the type specimen. from specimens distributed 12. Plagiochila striolata sp. nov. On dead wood, Lucma, 7,000 feet, August 7, 1911. Yellowish to brownish green, not glossy, growing in compact tufts : stems mostly 1.5 — 2 mm. long, about 0.25 mm. in diameter, rigid, ascending, sparingly and irregularly branched, the branches all intercalary, obhquely spreading, very rarely subdivided, similar to the stems but usually with smaller leaves : leaves alternate, contiguous to loosely imbricated, obhquely spreading (at an angle of about 60 degrees), antical portion more or less convex toward the base, postical portion slightly concave, apical portion plane ¦312 , Alexander W. Evans, or nearly so, broadly ovate, about 1.5 mm. long and 1.2 mm. wide, slightly or not at all decurrent antically, shortly decurrent postically, antical margin more or less revolute (on well-developed leaves), slightly outwardly curved, entire, postical margin strongly outwardly curved from the short and narrow decurrent portion, sharply dentate-spinose, apex broad and rounded, simi larly dentate-spinose ; total number of teeth mostly ten to twelve, unequal, usually two to five cells long and one to three cells wide at the base, teeth in the apical region tending to be larger than the others ; leaf-cells plane, aver aging about 20 fj. in the apical region, 25 X 20 fj, in the middle, and 35 X 20^ at the base, walls with large triangular trigones usually with bulging sides, and (especially toward the base) occasional oval intermediate thickenings, coalescence between thickenings rare except near the margin ; cuticle minutely striolate-verruculose : remaining parts not seen. (Fig. 6.) Figure 6. — Plagiochila striolata Evans A. Leaf, dissected from the stem, X 17. B. Cells from the middle of a leaf, X 300. C. Tooth from near the middle of the postical margin of a leaf, X 225. The figures were drawn from the type specimen. Although roughened cuticles are frequently found in many genera of the Hepaticae they have rarely been associated with definite species of Plagiochila. Stephani, however, has recently described two South American species in which this peculiarity occurs. These are P. asperifolia Steph., collected by Puiggari and Ule in Brazil, and P. verrucosa Steph., collected by Funck and Schlim in Venezuela. Unfortunately these species are known to the writer from description only, but are evidently distinct from P. striolata. The first is somewhat more robust, the leaves being about 2 mm. long. It is further dis tinguished by its obliquely truncate leaves with three apical spines, the anterior one being usually much larger than the others. The second species is still larger, the leaf-cells are only 12 u wide at the apex and only 36 X 12 u in size at the base, while the cuticle is coarsely verruculose. The spines in this species are furthermore said to show a distinct increase in size not only toward the apex but also toward the postical base. Hepaticae : Yale Peruvian Expedition of ipn. 313 V. Lophocolea Dumort. 13. Lophocolea diaphana Spruce Lophocolea diaphana Spruce, Hep. Amaz. et And. 426. 1884. On wet earth, San Miguel, 6,000 feet, July 24, 1911. The specimens are sterile and fragmentary, and their reference to the present species is open to some doubt. L. diaphana was collected by its author at several localities in Ecuador and was afterwards distributed in his exsiccatae. In the set examined by the writer the leaf-cells show minute trigones, although these are not mentioned in Spruce's description. In the specimens from San Miguel there are no trigones, but this might very well be owing to the poor develop ment of the plants. VI. Radula Dumort. 14. Radula andicola Steph. Radula andicola Steph. Hedwigia 23: 114. 1884. Radula viridi-aurea Spruce, Hep. Amaz. et And. 324. 1885. On rocks, San Miguel, 5,000 feet, September 2, 1911. A few fragments only but apparently referable to this species. The plants are sterile except for the presence of discoid gemmae, which are similar to those found in R. complanata (L.) Dumort. R. andicola was based on three specimens, the first collected by Krause in Ecuador, the second by Birschall in Venezuela, and the third by Deventer in Brazil; the original material of R. viridi-aurea was collec ted by Spruce in Ecuador. In his recent synopsis of the genus Radula Stephani1 reduces Spruce's species to synonymy, and ascribes a still wider distribution to the plant, quoting it also from Cuba, Costa Rica, Colombia, Peru, and Bolivia. 15. Radula ramulina Tayl. Radula ramulina Tayl. Jour. Bot. 5: 374. 1846. On rocks, San Miguel and vicinity, 5,000—6,000 feet, July 24 and September 3, 1911. A large and conspicuous species, pale yellowish in color and growing in depressed mats. The original material was collected by Jameson in the province of Pichincha, Ecuador. Spruce collected it repeatedly in the same country, and it has also been definitely reported from Bolivia. Stephani gives no precise localities for the species but states that it is abundant in the Andes. 1 Sp. Hepat. 4: 174. 1910. 314 Alexander W. Evans, VII. Porella L. 16. Porella arboiea (Tayl.) Trevis. Madotheca arborea Tayl. Jour. Bot. 5: 379. 1846. Porella arborea Trevis. Mem. R. 1st. Lomb. III. 4: 407. 1877. On earth, San Miguel, 5,000-6,000 feet, July 24, 1911. The specimens are sterile but agree pretty closely with those collected by Spruce at Tunguragua, Ecuador, and distributed in his exsiccatae. The original material of the species was found by Jameson in the province of Pichincha, Ecuador. According to Stephani it is abundant in the Peruvian Andes. 17. Porella squamulifera (Tayl.) Trevis. Madotheca squamulifera Tayl. Jour. Bot. 5: 378. 1846. Porella squamulifera Trevis. Mem. R. 1st. Lomb. III. 4: 407. 1877. On wood, Lucma, 7,000 feet, August 7, 1911. The specimens, which are sterile, have been carefully compared with those collected by Spruce at Pangor, Ecuador, and distributed in his exsiccatae. Although the Peruvian material does not agree very closely with Spruce's specimens in general habit, no structural differences could be demonstrated, and it therefore seems advisable to refer them to P. squamulifera, at least provisionally, rather than to try to separate them as a distinct species. The type material of P. squamulifera was collected by Jameson in the province of Pichincha, Ecuador, and according to Stephani the species occurs in the Andes of Peru. VIII. Microlejeunea (Spruce) Jack & Steph. 18. Microlejeunea bullata (Tayl.) Evans Lejeunea bullata Tayl. Jour. Bot. 5: 398. 1846. Microlejeunea bullata Evans, Mem. Torrey Club 8: 164. pi. 21, f. 20-29. 1902. On rocks, Santa Ana, 3,000 feet, August 4, 1911. The specimens are sterile and in very small amount but are appar ently correctly determined. M. bullata is a widely distributed spe cies, originally described from material collected by Guilding on the island of St. Vincent. It is now known also from South Carolina, from Florida, from several additional islands in the West Indies, and (according to Spruce) from numerous localities in Ecuador. Hepaticae : Yale Peruvian Expedition of 1911. 315 IX. Taxilejeunea (Spruce) Schiffn. 19. Taxilejeunea debilis (Lehm. & Lindenb.) Schiffn. Jungermannia debilis Lehm. & Lindenb.; Lehmann, Pug. Plant. 4: 51. 1832. Lejeunea debilis Lehm. & Lindenb. ; Nees & Montagne, Ann. des Sc. Nat. Bot. II. 5: 60. 1836. Montagne, Fl. Boliv. in d'Orbigny, Voy. dans l'Amer. Merid. 72: 63. fl. 1, f. 2. 1839. Omphalanthus debilis Lehm. & Lindenb. ; G. L. & N. Syn. Hep. 306. 1845. Lejeunea (Taxi-Lejeunea) debilis Steph. Hedwigia 29: 12. 1890. Taxilejeunea debilis Schiffn. Bot. Jahrb. 23: 579. 1897. On earth, San Miguel, 5,000-6,000 feet, July 24, 1911. A pale and delicate species growing in loose tufts. Widely distributed in the American tropics from the West Indies, through Mexico and Colombia, to Peru and Bolivia. The original material came from the island of St. Vincent, collector's name unknown. The specimens in the present collection have perianths and are in excellent condition. 20. Taxilejeunea florida (Spruce) Steph. Lejeunea (Taxi-Lejeunea) florida Spruce, Hep. Amaz. et And. 221. 1884. Taxilejeunea florida Steph. Sp. Hep. 5 : 467. 1914. Ollantaytambo, 9,000 feet, July 21, 1911. This species was based on material collected by Spruce in the Andes of Ecuador. Unfortunately the Peruvian specimens lack perianths, but they show both male and female inflorescences and agree closely with Spruce's full description. 21. Taxilejeunea pterogonia (Lehm. & Lindenb.) Schiffn. Jungermannia pterogonia Lehm. & Lindenb.; Lehmann, Pug. Plant. 6: 44. 1834. Omphalanthus pterogonius Lehm. & Lindenb.; G. L. & N. Syn. Hep. 306. 1845. Omphalanthus subalatus Lindenb. & Gottsche; G. L. & N. Syn. Hep. 747. 1847. Lejeunea (Taxi-Lejeunea) pterogonia Spruce, Hep. Amaz. et And. 216. 1884. Taxilejeunea pterogonia Schiffn.; Engler & Prantl, Nat. Pflanzen- fam. Is: 125. 1895. On earth, San Miguel, 5,000-6,000 feet, July 24, 1911. 316 Alexander W. Evans, The specimens are sterile but are apparently referable to this species. It was originally described from Peruvian material in the herbarium of Kunze. It is now known to have a wide distribu tion in tropical America and has been reported from Mexico, Colombia, Brazil, and Ecuador, as well as from Peru. X. Dicranolejeunea (Spruce) Schiffn. Several genera of the Lejeuneae have recently been discussed at length by the writer in connection with the Hepaticae of Porto Rico. Since Dicranolejeunea has not yet been detected on that island the present opportunity is taken of calling attention to its peculiarities. So far as known the genus is entirely tropical in its distribution. In his Species Hepaticarum Stephani recognizes three species from Africa, three from Asia, and twenty-two from North and South America, thus showing its preponderance in neo tropical regions. The species flourish best at relatively high alti tudes, altough a few have been reported from near the sea level. The bark of trees is the favorite habitat for the majority of the species; a few grow on the leaves of woody plants or on rocks; while others form more or less compact tufts on banks rich in humus. Sometimes a species is able to grow on several substrata. Under favorable conditions the plants show a yellowish or brownish pigmentation. Apparently in all cases the stems are at first pros trate, and oftentimes this habit is retained through life. In other cases the prostrate stems give rise to erect, ascending, or pendulous branches, or secondary stems, which become subdivided and con tinue their growth indefinitely. Under these circumstance it may become difficult or impossible to demonstrate the prostrate stems at all. A similar distinction between a primary prostrate axis and secondary stems occurs in many genera of the Hepaticae, such as Porella, Mastigolejeunea, and Bryopteris. Sometimes it represents a constant feature of an entire genus; sometimes, as in Dicranolejeunea, it is restricted to certain species. The branching in Dicranolejeunea is of two types, just as in Stictolejeunea, Bryopteris, and most species of Brachiolejeunea. The Frullania type, in which a leaf without a lobule is situated at the base of the branch, occurs on robust shoots and is associated with vigorous growth (Fig. 7, B; 8, C). The Radula type, in which the leaf at the base of the branch bears a distinct lobule, occurs on older shoots, especially where the development of re productive organs brings about a retardation or cessation of growth. Hepaticae : Yale Peruvian Expedition of ipn. 317 Apparently the subfloral innovations are invariably of this type (Fig. 7, A; 8, A). In branches of the Frullania type the first underleaf of the branch is either undivided or bifid and is displaced in such a way that the branch seems to arise in its axil. The first leaf on such a branch is rudimentary and very small but the suc ceeding leaves show normal features. The leaves vary from loosely to closely imbricated (Figs. 7, 8). As the plants become dry the lobes tend to assume a suberect position and often wrap themselves around the stem, very much as in Mas- tigolejeunea and Brachiolejeunea. When water is supplied the lobes become more or less flattened out and form an angle of forty-five degrees or more with the axis. In general outline the lobes are ovate and narrow to an acute, obtuse, or even rounded apex; they are plane or slightly convex, and in certain species the postical margin tends to be revolute; the margin is either entire or sharply and irregularly dentate. The lobule shows a more or less arched keel and is strongly in flated. In some cases the inflated portion involves the entire lobule, the free margin being involute except in the region of the sinus, where the opening into the water-sac is situated. In other cases the inflated portion lies along the keel and the free margin is appressed to the lobe. In the various species examined by the writer the free margin has normally borne two teeth, and these may be distinguished as the apical tooth and the proximal tooth respectively, the proximal tooth being situated between the apical tooth and the basal end of the margin. In poorly developed leaves the proximal tooth may be indistinct or even obsolete. The apical tooth is usually large and sharp, being two or three cells long and one or two cells wide at the base (Fig. 7, D, E) ; sometimes the tip cell becomes divided by a longitudinal wall (Fig. 8, G). The hyaline papilla is situated at the proximal base of this tooth but displaced to the inner surface of the lobe, usually two cells from the margin. Beyond the apical tooth the long and shallow sinus extends to the end of the keel. The proximal tooth is invariably shorter and less complex than the apical tooth (Fig. 8, H.). In many cases both teeth bend inward toward the lobe, making it im possible to determine their true structure without dissection. In certain respects the lobules just described agree in structure with those found in Brachiolejeunea, Ptychocoleus, and Odontole- The leaf-cells (Fig. 8, F) have thin walls, small trigones, and occasional intermediate thickenings, much more numerous in 318 Alexander W.' Evans, certain species than in others. Sometimes the trigones are so poorly developed that it is difficult to demonstrate them. The underleaves are undivided and vary in outline from oblong to orbicular or even reniform. They are attached by an arched line, the base is cuneate and often decurrent, and the margin is quite entire and either plane or more or less revolute. The rhizoids when present arise from a very rudimentary disc. According to Spruce the inflorescence in Dicranolejeunea is typically paroicous, although he admits that a dioicous infloresc ence sometimes occurs. Stephani describes paroicous, autoicous, and dioicous species. In all the last, however, he admits that the male inflorescence is unknown. The great difficulty in demon strating the vestiges of antheridia in paroicous species after the sperms have been discharged makes it possible that some, if not all, of the species described as dioicous may in reality be paroicous. It should also be remembered that certain monoicous species occa sionally produce individuals in which antheridia but no archegonia are present. The female inflorescence is sometimes borne on a leading branch and sometimes on a more or less abbreviated branch. In almost every case two subfloral innovations are developed below a primary inflorescence, although occasionally only one is present (Figs. 7, 8). The innovations are usually short and are often terminated by secon dary inflorescences. These in turn may develop new innovations upon which tertiary inflorescences may arise, but no cases have been observed in which inflorescences of a higher order have occurred. The innovations are typically in pairs, even in connection with secon dary or tertiary inflorescences, but single innovations are more fre quent in such cases than in connection with a primary inflorescence. The perichaetial bracts (Figs. 7, F— K ; 8, 1— K) are usually distinctly smaller or at least narrower than the leaves. The lobes show a stronger tendency to be toothed, and dentate bracts occur in certain species where the vegetative leaves are entire. The lobules are rarely well developed. In most cases they are exceedingly minute and may even be reduced to a hyaline papilla. The bracteoles are usually larger and broader than ordinary underleaves, but they show less tendency to be toothed than do the lobules of the bracts. The perianth (Figs. 7, 8) yields several of the more important characters of the genus. It is distinctly compressed and the sharp lateral keels bear dentate or spinose wings variable in width. The antical surface is sometimes plane, sometimes slightly concave, and sometimes bears a very low keel without either wing or teeth. Hepaticae : Yale Peruvian Expedition of ipn. 319 The postical surface, theoretically at least, bears two low dentate or spinose keels, but these are often coalescent so that only a single median keel appears to be present, and the teeth or spines may be few or absent altogether. The male inflorescence in paroicous species occupies a variable length of the fertile branch below the archegonium. The perigon- ial bracts differ scarcely if. at all from the ordinary leaves, except for the presence of the single antheridium enclosed within the lobule. The perigonial bracteoles agree in all respects with the underleaves on ordinary vegetative shoots. In autoicous species the male in florescence occupies a branch of variable length. The bracts are monandrous, closely imbricated, and much smaller than ordinary leaves. The reduction in size affects the lobes most strongly, these being but little larger than the lobules. The bracteoles are likewise smaller than ordinary underleaves and extend the whole length of the inflorescence. When compared with the other genera of the Lejeuneae, Dicrano lejeunea is seen to be especially close to Odontolejeunea, in its restricted sense.1 In both genera the lobes of the leaves tend to be more or less toothed, although species with entire leaves are included in Dicranolejeunea; in both genera the free margin of the lobule bears one or more teeth, the number being two in all the known species of Dicranolejeunea and varying from one to four in Odontolejeunea ; in both genera the hyaline papilla of the lobule is situated at the base of the apical tooth on the surface turned toward the lobe; in both genera the leaves lack ocelli, and the leaf-cells have pigmented walls with local thickenings. They agree further in their undivided under leaves, entire in Dicranolejeunea and either entire or variously toothed in Odontolejeunea; in the small or obsolete lobules of their perichaetial bracts ; and in their compressed perianths with dentate wings on the narrow lateral keels. It is difficult, indeed, to separate the genera by characters which are both constant and significant, although each comprises a group of closely related species. It should be noted, however, that the branching in Odontolejeunea is always of the Radula type, that the subfloral innovations are always single, that the peri gonial bracts are typically diandrous, and that the radicelliferous discs on the underleaves are well developed. In Dicranolejeunea both the Radula and the Frullania types of branching occur, the sub floral innovations are typically in pairs, the perigonial bracts seem to be constantly monandrous, and the radicelliferous discs are either 1 See Evans, Bull. Torrey Club 31 : 183. 1904. 320 Alexander W. Evans, absent altogether or. very rudimentary. Odontolejeunea is further distinguished by the occurrence of specialized caducous branches, which multiply the plants vegetatively; in Dicranolejeunea no method of vegetative reproduction has been observed. Of course the species of Dicranolejeunea in which secondary stems arise from a prostrate primary stem differ markedly in habit from all the species of Odonto lejeunea, in which the plants always cling closely to the substratum, but they differ in the same way from the species of Dicranolejeunea in which a prostrate habit is retained through life. Two species of Dicranolejeunea occur in the present collection. One is the type of the genus, D. axillaris (Nees & Mont.) Schiffn.; the other is apparently undescribed. 22. Dicranolejeunea axillaris (Nees & Mont.) Schiffn. Lejeunea axillaris Nees & Mont. Ann. des Sc. Nat. Bot. II. 5: 59. 1836. Lejeunea (Dicrano-Lejeunea) axillaris Spruce, Hep. Amaz. et And. 139. 1884. Dicranolejeunea axillaris Schiffn.; Engler & Prantl, Nat. Pflanzen- fam. I3: 128. 1895. San Miguel, 5,000-6,000 feet, July 24, 1911; a few fragments only, mixed with Plagiochila chinantlana. This common and variable species was originally described from Bolivian specimens collected by d'Orbigny. Its known range extends from the West Indies and Mexico, through Central America and the Andes, into Chile. It has been recorded also from the Galapagos Islands. Since the published descriptions are not altogether in accord, even on essential points, a full account of the plant will ' perhaps be useful. According to Spruce J the leaves are contiguous to subimbricated, the apex of the lobe bears three or more sharp teeth, the margin of the lobule is unidentate, the leaf-cells are thickened at the angles (and rarely in between), the inflorescence is paroicous, and the post ical surface of the perianth is either smooth or else bears one or two cilia near the apex. According to Stephani2 the leaves are im bricated, the apex of the lobe is broadly acuminate-acute and a few teeth are borne below the apex, the lobule is obliquely truncate at the apex with an acute angle, the leaf-cells have no trigones, the in florescence is monoicous, the small androecia, sessile on the stem, 1 Hep. Amaz. et And. 139. i? 2 Sp. Hep. 5 : 164. 1912. Hepaticae : Yale Peruvian Expedition of ipn. 321 bear from four to eight pairs of bracts, and the postical keel of the perianth is smooth. From the study of a series of specimens from various localities it has become evident that some of these discrepancies are due to Figure 7. — Dicranolejeunea AXILLARIS (Nees & Mont.) Schiffn. A. Part of a secondary shoot, showing a primary inflorescence with peri anth and the bases of two innovations, postical. view, X 17. B. Part of a secondary shoot, showing the base of a branch of the Frullania type, postical view, X 17. C. Two leaves, antical view, X 17. D. Part of a lobule, show ing the two marginal teeth, X 225. E. Apical tooth of another lobule, X 225. F. — H. Bracts and bracteole from a primary inflorescence, X 27 I. — K. Bracts and bracteole from a secondary inflorescence, X 27. L. Trans verse section of a perianth in upper third, X 40. The figures were drawn from specimens collected by Spruce in the Andes of Ecuador and distributed in his exsiccatae. * the variability of the species, but that others can hardly be explained in this way. The writer finds that' the leaves are sometimes rather closely imbricated (Fig. 7, A— C) and that sometimes they scarcely overlap at all. The lobes taper gradually and are sharply pointed 322 Alexander W. Evans, at the apex. Sometimes, except for this apical tooth, the margin is quite entire, but it is much more usual for the lobe to bear from one to five supplementary teeth in the vicinity of the apex. These teeth tend to be more numerous along the antical margin, but some of them are occasionally situated on the postical margin. In most cases the apical tooth is distinctly larger than the others, but this is not always the case, the antical tooth next the apex sometimes approach ing or equalling the apical tooth in size. Under these last circum stances the apical tooth is less conspicuous than on most leaves. The true lobular features can not always be made out without dis section. If this is done the lobules are found to conform pretty closely with Spruce's description, although it should be noted that neither he nor Stephani makes any mention of the proximal tooth. The apical tooth of the lobule is short and rather sharp, being tipped by a single cell or by two superimposed cells; in other words it is usually two cells long and one or two cells wide at the base (Fig. 7, D, E) . Just below the tooth four or five cells are usually connected with the basal cell or cells. The proximal tooth is short and consists of little more than a single projecting cell. Apparently trigones are always present, but they may be very inconspicuous with concave sides. Even when better developed they are less striking than in most members of the Lejeuneae Holostipae. With regard to the inflorescence it is extremely difficult to demon strate antheridia or their vestiges after the sperms have been set free and especially after perianths and sporophytes have developed. It is only on very young fertile shoots, where the innovations have just begun to grow, where the archegonium is still unfertilized, and where the perichaetial bracts and bracteole are only partially developed, that the antheridia show clearly. In typical cases there about eight pairs of perigonial bracts just below the female inflor escence. As might be expected delicate and etiolated plants some times develop antheridia but no archegonia. This is shown particu larly well in material collected by A. Stewart on Albemarle Island, one of the Galapagos group. In some plants from this collection certain stems or leading branches bear a series of perigonial bracts and then continue their growth as ordinary vegetative shoots. This may perhaps be considered an example of reversion, due to the fact that the conditions are not very favorable for the development of reproductive organs ., Spruce's description of the perianth and bracts (Fig. 7, F— L) is unusually full and accurate. The postical keel of the perianth, as he remarks, is low and usually smooth. But this appearance is not Hepaticae: Yale Peruvian Expedition of ipn. 323 altogether constant. In some cases it bears. long cilia, similar to those on the lateral wings . These ciha are never numerous— from one to three in most cases — and, when more than one, often occur in two rows, representing the angles of the two coalescent postical keels. In most cases the postical ciha are close to the apex, but they may be situated as far back as the middle of the perianth. Apparently they are never borne on wings. In D. hypoacantha (Spruce) Steph.,1 known only from Pallatanga, Peru, where it was collected by Spruce, D. axillaris has a very close ally. In fact the validity of D. hypoacantha as a species may be considered doubtful. The two species are of about the same size and agree closely in their leaves, underleaves, bracts, and brac teoles. The leaves in D. hypoacantha, to be sure, occasionally have more teeth, as many as eight in extreme cases, but usually the number is no higher than in D. axillaris. According to Spruce the inflores cence in D. hypoacantha is paroicous, but Stephani describes it as dioicous. No antheridia could be discovered in the material studied by the writer so that neither of these statements could be confirmed. In separating D. hypoacantha Spruce relied particularly on characters drawn from the perianths. The form, the beak, and the lateral wings are much the same in the two species, but a typical perianth in D. hypoacantha bears narrow spinose wings along the two coalescent postical keels. These are shorter than the lateral wings but resemble them in other respects. Unfortunately this typical condition seems to be very rare. Frequently only one of the postical wings is devel oped, or else there are no wings at all and the spines grow out directly from the keels. In extreme cases the number of spines may be reduced to only two or three, and the perianths then bear a strong resemblance to those of D. axillaris. 23. Dicranolejeunea rotundata sp. nov. On live wood, Santa Ana, 3,000 feet, August 3, 1911. Brownish green, not glossy, growing in depressed mats : stems about 0.15 mm. in diameter, sparingly pinnate, the branches obliquely to widely spreading, similar to the stem but sometimes with slightly smaller leaves, never microphyllous : leaves imbricated, obhquely spreading, the lobe slightly convex and falcate, ovate, mostly 0.9 — 1 mm. long and 0.65 — 0.75 mm. wide, arching part way across the stem, antical margin outwardly curved from the base to the broad and rounded or very bluntly pointed apex, postical margin straight or more or less outwardly curved, margin entire throughout ; lobule 1 Sp. Hepat. 5 : 166. 1912. Trans. Conn. Acad., Vol. XVIII. 22 April, 1914. 324 Alexander W. Evans, ovate, 0.45 X 0.25 mm., more or less inflated in basal and carinal portions, keel straight or shgtly arched, free margin curved, slightly involute at the base, apical tooth blunt, three or four cells long, usually tipped with two Figure 8.— Dicranolejeunea rotund at a Evans A. Part of a floriferous shoot, with a primary inflorescence and two sec ondary inflorescences, postical view, X 17. B. Secondary inflorescence with one innovation, the underleaves dissected away, postical view, X 17. C. Part of a shoot showing the base of a branch of the Frullania type, post ical view, X 17. D. Unfertilized secondary inflorescence, the bracteole dissected away, postical view, X 17. E. Two leaves, antical view, X 17. F. Cells from the middle of a lobe, X 225. G. Apical tooth of a lobule, X 225. H. Proximal tooth of a lobule, X 225. I, J. Bracts from a primary inflorescence, X 17. K. Bracteole from another primary inflorescence, X 17- L. Transverse section of a perianth in the upper third, X 17. The figures were drawn from the type specimen. superimposed cells or with two cells side by side, proximal tooth shorter but usually tipped with two superimposed cells ; lobule often poorly developed, with indistinct teeth or even entire on the free margin ; cells of lobe plane or nearly so, averaging 15 ,u at the margin, 35 X 27 p in the middle, and 40 X 3° /" at the base, thin walled, trigones small, triangular, intermediate Hepaticae : Yale Peruvian Expedition of ipn. 325 thickenings narrowly oval, infrequent except near the base : underleaves contiguous to loosely imbricated, plane, broadly orbicular, about 0.45 mm. long and 0.55 mm. wide, cuneate at the base, broad and rounded or truncate at the apex, entire throughout or slightly crenulate near the base : inflores cence apparently dioicous (male inflorescence unknown) : Q inflorescence borne on the stem or a leading branch, innovating on both sides, the inno vations soon floriferous, the secondary inflorescences innovating on one or both sides with sterile innovations ; bracts unequal in size, those of the prim ary inflorescence suberect, the lobe of the larger one ovate, 1.2 X 0.9 mm., gradually narrowed to a blunt point, entire, the lobe of the smaller one more narrowly ovate, 0.8 X 0.45 mm., subacute, entire, lobule a smaU basal fold, maximum size about 0.45 X 0.07 mm. but in extreme cases re duced to a hyaline papilla; bracteole similar to the underleaves but relati vely narrower, about 0.75 X 0.8 mm., bracts and bracteoles of secondary flowers smaller but otherwise similar to those of the primary flower ; perianth shghtly exserted, obovate in outhne from a cuneate base, 1.1 mm. long and 0.9 mm. wide (in secondary inflorescences only 0.9 X 0.8 mm.), truncate at the apex with a short beak, antical surface concave, mostly smooth, rarely with a spine or two in the apical region, postical surface with a broad rounded keel bearing a few scattered spines in the apical portion, lateral keels bearing broad wings extending from the apex sometimes to beyond the middle, sharply and irregularly spinose, the spines usually six to ten on each wing, mostly three to six cells long and two cells wide at the base but very variable : capsule (immature) about 0.35 mm. in diameter. (Fig. 8.) The present plant is one of those species in which a prostrate habit is retained through life and is remarkable for its entire, broad, and usually rounded leaves. Its closest ally is perhaps D. phyllorhiza (Nees) Schiffn., a species of the American tropics known to the writer from descriptions only.1 This species, which shows the same habit of growth as D. rotundata, is considerably more robust, the leaf-lobes being 2.25 mm. long and 2 mm. wide, according to Stephani's des cription. It is further distinguished by its relatively broader leaves, by its autoicous inflorescence, by the obovate-oblong lobes of its perichaetial bracts, and by its spatulate bracteoles. XI. Brachiolejeunea (Spruce) Schiffn. 24. Brachiolejeunea densifolia (Raddi) Evans Brachiolejeunea bicolor (Nees) Schiffn.; Engler & Prantl, Nat. Pflanzenfam. I3: 128. 1895. Brachiolejeunea densifolia Evans, Bull. Torrey Club 35 : 158. 1908.2 1 See Stephani, Sp. Hep. 5 : 162. 1912. 2 A full synonymy of the species may be found here. 326 Alexander W. Evans, On wood, San Miguel, 6,000 feet, July 24, 1911. On rocks, Santa Ana, 3,000 feet, August 4, 1911. A widely distributed species in tropical America, its known range extending from Mexico to Brazil and Bolivia. It flourishes best at high elevations. The writer has already pointed out the fact that the West Indian records for the species are based on incorrect deter minations.1 XII. Marchesinia S. F. Gray 25. Marchesinia brachiata (Swartz) Schiffn. Marchesinia brachiata Schiffn.; Engler & Prantl, Nat. Pflanzen- fam. I3: 129. 1895. On earth, San Miguel, 6,000 feet, July 24, 1911. This species has recently been fully discussed by the writer and its extensive range commented upon.2 It is known at present from Mexico and the West Indies; from Venezuela, Colombia, and Brazil; from Ecuador, Peru, and Bolivia; and from the Galapagos Islands. XIII. Omphalanthus Lindenb. & Nees 26. Omphalanthus filiformis (Swartz) Nees. Omphalanthus filiformis Nees; G. L. & N. Syn. Hep. 304. 1845. On rocks, Huadquina, 5,000 feet, August 1, 1911. On wood, Santa Ana, 3,000 feet, August 4, 1911. As in the case of the preceding species Omphalanthus filiformis has recently been described at length by the writer.3 It has an equally wide distribution, being known from several of the West Indian Islands, and, on the mainland, from Mexico to Bolivia. It has been collected also on the Galapagos Islands. XIV. FruUania Raddi Over seven hundred species of FruUania are recognized by Stephani,4 the vast majority being tropical. Many of them are of considerable size and show a striking reddish or purplish pigmentation, so that they readily attract the attention of collectors. In certain respects the genus shows a high degree of specialization, although it 1 Bryologist 15 : 62. 1912. 2 Bull. Torrey Club 34: 546. pi. 32. 1907. 3 Bull. Torrey Club 34 : 18. pi. 3, f. 1-9. 1907. 4 Sp. Hep. 4: 316 — 686. 1910 — 1911. Hepaticae : Yale Peruvian Expedition of ipn. 327 retains a primitive type of branching. In the present collection five species are in a condition to be determined, a sixth species being ster ile and doubtful. Of the five species the first three noted below belong to the subgenus Chonanthelia of Spruce, the fourth to the sub genus Trachycolea, and the fifth to the subgenus Thyopsiella. The three species of Chonanthelia are among the most widely distributed Frullaniae of the American tropics but have never been illustrated. Since they are of interest from various points of view the writer takes the present occasion for discussing them at length. The species of Trachycolea has already been figured and described by the writer in another connection,1 while the species of Thyopsiella is still in need of further study. 27. FruUania gibbosa Nees FruUania gibbosa Nees; G. L. & N. Syn. Hep. 411. 1845. On rocks, Santa Ana, 3,000 feet, August 4, 1911. The author of F. gibbosa based his species on a series of specimens from the West Indies, Guiana, and Brazil. In the later pages of the Synopsis Hepaticarum it is recorded also from Mexico and Colombia. Spruce extended its known range into Peru and commented on the curious fact that it behaved as a weed and always grew in the neighbourhood of human habitations.2 Stephani notes its occurrence in Alabama and in Chile and states also that it is very common in tropical America, without giving definite localities. The material from Santa Ana belongs to Spruce's variety densissima, which he suggests may be specifically distinct. In the opinion of the writer, however, the differential characters assigned to this variety represent exaggerations of the features characteristic of the species and are neither marked enough nor constant enough to warrant a specific separation. The first specimen of F. gibbosa quoted by Nees von Esenbeck was collected by Swartz on tree-trunks in Jamaica and may be re garded as the type of the species. It represents a part of the material upon which Swartz based his Jungermannia obscura,3 a composite species which is no longer recognized. A portion of this type material, preserved in the Lindenberg herbarium at Vienna (No. 6979), has been examined by the writer; it is unfortunately very fragmentary but seems referable to F. gibbosa as at present understood. The same herbarium contains specimens from Colipa, Mexico (Liebmann, 1 Trans. Conn. Acad. 10: 14. pi. 6. 1897. 2 Hep. Amaz. et And. 22. 1884. 3 Fl. Ind. Occid. 3: 1869. 1806. 328 Alexander W. Evans, No. 6947) British Guiana (Campbell, No. 6948), French Guiana (Leprieur, No. 6977), Barbados (Herb. Hooker, No. 6952), and Merida, Colombia (Moritz, Nos. 6949, 6950). These specimens are among those mentioned in the Synopsis Hepaticarum, and all clearly represent the same species. C. Wright's specimens of F. gibbosa from Cuba, distributed in Hepaticae Cubenses ; a large series from Jamaica, collected by Harris, Underwood, Maxon, and the writer; and Spruce's specimens from Tarapoto, Peru, distributed in Hepaticae Spruceanae, have likewise been examined. All of this material is clearly F. gibbosa, but the specimens from the Bahama Islands, recently reported upon by the writer,1 differ somewhat from the others and perhaps belong elsewhere. With regard to the habitat of the species it occurs both on bark and on rocks, frequently in exposed situations, and some of its pecul iarities seem to be associated with a xero phytic environment. The plants cling closely to the substratum and the stems branch frequently. Sometimes the branches are variable in length and become irreg ularly subdivided ; sometimes they are more or less definitely Hmited in growth and give the shoots a plumose appearance. Under the latter circumstances the branches tend to be simple or else develop only one or two short branches of a higher rank. In color the plants vary from a greenish yellow to a deep dull brown and are often tinged with whitish, owing to the fact that the margins of the leaf-lobes die and bleach out through exposure. The leaves are very closely imbricated, especially in the variety densissima, the imbrication affecting not only the lobes but the lob ules as well (Fig. 9, A) . The lobes are convex when dry but are more or less squarrose; when moist the apical portion spreads still more widely from the axis, making the surface of the lobe somewhat con cave, but this appearance is never so pronounced as in F. squarrosa (R. Bl. & N.) Dumort., a species which resembles F. gibbosa at first sight and which grows in similar localities. The lobes in the present species are broadly orbicular, measuring about 0.9 mm. in length and 1.2 mm. in width; they arch considerably beyond the axis and are rounded at the apex. At the base a distinct auricle is developed and this may be plane or more or less crispate (Fig. 9, D). This auricle is usually so large and so strongly dilated that it covers over the line of attachment of the lobe. The margin of the auricle is cren- ulate from projecting cells, although the rest of the lobe is quite entire. 1 Bull. Torrey Club 38: 220. 1911. Hepaticae : Yale Peruvian Expedition of ipn. 329 The lobules (Fig. 9, A-C) measure 0.6-0.7 mm. in length and 0.45-0.6 mm. in width when well developed. In normal cases the upper part develops an inflated water-sac, well rounded above and with an oblique mouth. In some cases the inflated portion occupies the entire sac and the mouth is wide open ; in other cases the inflated Figure 9. — Frullania gibbosa Nees A. Part of a branch bearing a paroicous inflorescence with perianth, postical view, X 27. B. Two stem-leaves and an underleaf, the stylus of the leaf on the right forcibly reflexed, postical view, X 27. C. Three stem- leaves and an underleaf, postical view, X 17. D. Stem-leaf, antical view, X 17- E. Cells from the middle of a lobe, X 225. F. — H. Bracts and brac teoles of a single inflorescence, F. the innermost series, G. the second, and H. the third (perigonial), X 17- L Transverse section of a perianth in the upper third, X 27. C. was drawn from a specimen collected by the writer at Content Gap, Jamaica, in September, 1906; the other figures, from the specimens collected at Santa Ana. 330 Alexander W. Evans, portion is restricted to the upper and outer parts of the sac, the free margin is appressed, and the opening is reduced to a small hole at the outer end. The apex of the lobule, where this hole is situated, never forms a projecting beak as in so many species of Chonanthelia. The water-sac, in fact, bears considerable resemblance to those found in Trachycolea. The lower part of the lobule, usually occupying in F. gibbosa half the length or even less, is the so-called appendiculum of authors. In the present case it is less well developed than in most species of Chonanthelia and is variable in form, tending perhaps to be triangular. Along the inner edge the appendiculum is attached to the lobe by a short keel, which spreads more or less widely away from the axis or, in rare cases, is subparallel with it. The outer margin of the appendiculum, extending from the end of the keel to the outer end of the water-sac, is sometimes straight and sometimes more or less strongly curved. In most cases it is well covered over by the lobe but occasionally extends slightly beyond it. The margin is apparently never toothed in any way but may be very slightly cris- pate. In some cases the lobule does not form a water-sac but con sists of an ovate lamina, which is either plane or slightly canalic ulate and which tapers to a blunt point. The stylus is remarkably developed and sometimes approaches the lobule itself in size. It varies from an ovate to a broadly reniform expansion and is sometimes plane but much more frequently cris- pate. Sometimes the stylus is appressed to the axis and is directed toward the apex of the shoot (Fig. 9, C, on left) ; sometimes it is com pletely reflexed by the auricles of the underleaves (Fig. 9, C, on right) ; and sometimes a portion only is reflexed, while the rest occupies the normal position. Under the last conditions the stylus is sometimes bifid. Apparently the broad expansion just described represents an extreme development of the inner basal portion of the stylus. The reason for this idea is that, between the expansion and the lobule, a slender tooth-like structure, tipped with a hyaline papilla, can be discerned (Fig. 9, B, on right). In most species of FruUania the entire stylus consists of a structure of this character. Although the stylus in F. gibbosa is so conspicuous and affords one of the most distinctive features of the species, it is not mentioned by Nees von Esenbeck, and its characteristics are not brought out by Stephani. Gottsche,1 however, emphasizes its importance, and Spruce describes it at some length. According to his account it attains its highest development in the variety densissima, where 1 Ann. des Se. Nat. Bot. V. 1: 166. 1864. Hepaticae : Yale Peruvian Expedition of ipn. 331 it is often more or less adnate to the auricles of the underleaves. The writer has been unable to demonstrate this adnation, even on robust plants, although it might perhaps be hazardous to maintain that it never occurred. In any case, however, the adnation is too inconstant to be considered a differential character of importance. The underleaves are about as large as the lobes and agree with them in being densely crowded. They are broadly orbicular or reniform and are attached by a slightly arched line. The two sharp apical teeth are separated by a lunulate sinus and are usually less than one fourth the length of the underleaf. The teeth are commonly connivent and tend to be more or less twisted or con torted, while the margin of the sinus is often revolute. Somewhere between the bottom of the sinus and the line of attachment the underleaves form a short and rounded tubular projection, at the tip of which the rhizoids are developed. At the base the under leaves are strongly auriculate on each side. The auricles are much like those on the lobes. They are often more or less crispate and may be so dilated that they partially cover over the line of attach ment of the underleaves. Sometimes the crispate nature of the auricles is so exaggerated that incomplete water-sacs are formed. It will at once be noted that the auricles of the underleaves, the auricles of the lobes, and the expanded portions of the stylus are all very similar structures and that through them the capacity of the species for holding water by capillarity is greatly increased. And it is precisely these structures which attain their highest development in markedly xerophytic localities. The cells of the lobe average about 23^ along the margin, 33 u in the middle, and 43