I 
 
STUDIES ON THE REPRODUCTION AND 
 ARTIFICIAL PROPAGATION OF FRESH- 
 WATER MUSSELS •* * >* * 
 
 From BULLETIN OF THE BUREAU OF FISHERIES, Volume XXX, 1910 
 Document No. 736 : : : : : .' : : : : : : : Issued May 10, 1912 
 
 WASHINGTON GOVERNMENT PRINTING OPFICE 
 
STUDIES ON THE REPRODUCTION AND 
 ARTIFICIAL PROPAGATION OF FRESH- 
 WATER MUSSELS * s s * 
 
 From BULLETIN OF THE BUREAU OF FISHERIES, Volume XXX, 1910 
 
 Document No. 756 : : : : : : : : : : : : ; : : Issued May 10, 1912 
 
 7 a. 
 
 WASHINGTON :::::: GOVERNMENT PRINTING OFFICE 
 

 K 191? 
 
ll 
 
 STUDIES ON THE REPRODUCTION AND ARTIFICIAL 
 PROPAGATION OF FRESH-WATER MUSSELS 
 
 By George Lefevre and Winterton C. Curtis 
 
 Professors oj Zoology in the University of Missouri 
 
 i°5 
 
CONTENTS. 
 
 Page. 
 109 
 
 Introduction 
 
 I. Historical in 
 
 II. Reproduction 114 
 
 The marsupiurn 116 
 
 Use of the marsupiurn in classification 116 
 
 General structure of the marsupiurn 120 
 
 Internal structure of the marsupiurn 125 
 
 Phylogeny of the marsupiurn 135 
 
 Conglutination of the embryos 136 
 
 Stratification of unfertilized eggs 138 
 
 Abortion of embryos and glochidia 138 
 
 Breeding seasons 139 
 
 Long period of gravidity 141 
 
 Short period of gravidity 143 
 
 III. The larva 145 
 
 Structure of the glochidium 145 
 
 The hookless type 146 
 
 The hooked type 149 
 
 The Proptera or axe-head type 150 
 
 The larval thread 151 
 
 Behavior and reactions of glochidia 152 
 
 Reactions of hookless glochidia 153 
 
 Reactions of hooked glochidia 155 
 
 IV. The parasitism 156 
 
 Artificial infection of fish 156 
 
 Infections with hooked glochidia 158 
 
 Infections with hookless glochidia 160 
 
 Susceptibility of fishes to infection 162 
 
 Behavior of fishes during infection 163 
 
 Infection of fish in large numbers 164 
 
 Conditions necessary for successful infection 166 
 
 Duration of the parasitic period 167 
 
 Implantation and cyst formation 169 
 
 Metamorphosis without parasitism in Sirophiius 171 
 
 V. Attempt to rear glochidia in culture media 174 
 
 VI. Post-larval stages 175 
 
 Beginning of the growth period and life on the bottom 175 
 
 Juvenile stages and the origin of mussel beds 177 
 
 Rate of growth 179 
 
 Growth of mussels in wire cages 180 
 
 An artificially reared mussel 182 
 
 The origin and age of mussels in artificial ponds 184 
 
 107 
 
108 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 Page. 
 
 VII. Investigations on the upper Mississippi River 187 
 
 VIII. Economic applications 189 
 
 Protective laws 189 
 
 Selection and maintenance of a fish supply 190 
 
 The best seasons for infections 191 
 
 The mussel supply 191 
 
 Rearing and distributing young mussels 192 
 
 DC. Conclusion 193 
 
 Bibliography 195 
 
 Explanation of plates 198 
 
STUDIES ON THE REPRODUCTION AND ARTIFICIAL 
 PROPAGATION OF FRESH-WATER MUSSELS. 
 
 By GEORGE LEFEVRE and WINTERTON C. CURTIS, 
 Professors of Zoology in the University of Missouri. 
 
 J- 
 
 INTRODUCTION. 
 
 The threatened extinction in the upper Mississippi River and its more important 
 tributaries of those species of the Unionidse whose shells have been taken in enormous 
 numbers in recent years, both for the manufacture of pearl buttons and for the pearls 
 which they occasionally contain, has led the United States Bureau of Fisheries to under- 
 take an extensive investigation of the possibility of artificially propagating the com- 
 mercial species and of devising practicable means of restocking depleted waters which 
 present favorable conditions for their maintenance. The general direction of the inves- 
 tigation has been placed in the hands of the writers, who for several years have devoted 
 as much time as their regular duties have allowed to the work, in certain important 
 phases of which, however, many others have collaborated. 
 
 It was recognized at the outset that if the investigation was to be of any practical 
 value it must be wide in scope and must extend over a period of at least several years. 
 At that time much remained to be learned concerning the breeding habits and seasons 
 of the commercial species, the biological and physical conditions under which they live, 
 their distribution throughout the Mississippi Valley, and many other essential matters, 
 while it was yet to be discovered whether artificial propagation could be successfully 
 carried out. At the very inception of the work, therefore, a comprehensive plan was 
 outlined which was designed to include every subject that might bear even remotely 
 upon the central problem — the restoration of the exhausted mussel beds — and, although 
 many parts of this program have scarcely been touched, much progress has been made 
 in some of the more important lines. 
 
 The plan of work contemplated, besides a thorough investigation of the conditions 
 under which artificial propagation might be possible, a detailed study of the life history 
 and ecology of the Unionidse, with special reference to the geographical distribution of 
 the group throughout the Mississippi Valley, the breeding seasons and habits, the 
 
 109 
 
HO BULLETIN OF THE BUREAU OF FISHERIES. 
 
 physical conditions of the waters in which different species thrive and attain their 
 maximum growth, food supply, enemies and diseases, rate of growth and the influence 
 of environmental factors upon it, and the behavior of glochidia and fishes as parasites 
 and hosts, respectively. 
 
 The results that have already been obtained, although far from complete, will 
 serve as a basis for future investigations, while the lines of attack in the main prob- 
 lems have been definitely indicated. We have proceeded far enough to make it clear 
 that the ultimate end of the investigation is assured, and with adequate facilities for 
 the infection and care of large numbers of fishes and for the maintenance of the young 
 mussels during the early stages of growth following the metamorphosis, the final success 
 of the work can no longer be in doubt. The essential facts in the life history of the 
 Unionida are known; the breeding seasons and habits of the commercial species have 
 been sufficiently determined; the general conditions of infection and of the parasitism 
 of the larva have been learned experimentally; and the entire feasibility of artificially 
 propagating at least certain species of fresh-water mussels has been clearly demonstrated ; 
 while the requisite conditions for placing artificial propagation on a practical basis are 
 now thoroughly understood. 
 
 The writers' personal attention has in the main been directed to a study of the 
 conditions of reproduction in the group and the parasitism of the larva in their bearing 
 upon the problem of artificial infection of fishes with glochidia, while such phases of 
 the investigation as geographical distribution, systematic studies, and a number of 
 special ecological problems have been in the hands of other investigators. 
 
 At the recently established biological station of the Bureau of Fisheries at Fairport, 
 Iowa, while construction was still in progress, the work of propagating some of the 
 commercial species was inaugurated, and the excellent facilities of the station, which 
 has been especially designed for the purpose, are now being utilized by members of the 
 staff in attacking fundamental problems of both a scientific and an economic nature. 
 
 For the past five summers a number of field parties have been equipped and sent 
 out each year by the Bureau to collect fresh-water mussels and to obtain the fullest 
 
 Note. — It is a pleasure to state that a generous grant of money made by the National Association of Pearl Button 
 Manufacturers in the interest of the investigations enabled us to purchase a collection of books and pamphlets, dealing with the 
 literature on the Unionida?, which has been of invaluable assistance in the course of the work. To individual members of this 
 association, especially to Sir. J. E. Krouse, of Davenport, Iowa, Messrs. W. F. Bishop and Henry Umlandt, of Muscatine, Iowa, 
 and Mr. D. W. MacWillie, of La Crosse. Wis., we are indebted for many courtesies and for shipments of live mussels which they 
 have repeatedly secured for us. Many others have at times assisted us by sending us material, and in this connection we take 
 especial pleasure in thanking Prof. U. O. Cox, of the State Normal School at Terre Haute, Ind., who has kindly furnished us 
 on several occasions with valuable lots of gravid mussels from the Wabash River. 
 
 To a number of our students, who in various capacities have been of service to the investigations, we owe much, and among 
 them should be mentioned Miss Daisy Young, Messrs. Howard Welch, F. P. Johnson, W. E. Dandy, L. E. Thatcher, 
 and especially Mr. W. E. Muns, who acted as our assistant in this work for over two years. 
 
 Lastly, it is a pleasure to acknowledge our obligation to Mr. G. T. Kline, the biological artist of the University of Missouri, 
 who has contributed much to the value of our work by the beautiful and accurate drawings with which he has illustrated this 
 and previous papers published by us. 
 
 By permission of the Commissioner of Fisheries, we have had the privilege of publishing, in advance of this more detailed 
 report, the following papers of a preliminary nature: Experiments in the artificial propagation of fresh-water mussels 
 (Proceedings of the Fourth International Fishery Congress, Bulletin of the Bureau of Fisheries, vol. xxvm, 1908); The marsupium 
 of the Unionidce (Biological Bulletin, vol. xix, no. 1, 1910); Reproduction and parasitism in the Unionida? (Journal of Experi- 
 mental Zoology, vol. IX, no. 1, 1910); Metamorphosis without parasitism in the Unionidce (Science, vol. xxxm, no. S57, I9n). 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. I 1 1 
 
 possible data bearing upon their distribution, their habits, and the physical and bio- 
 logical factors of their environment, as well as information concerning the industries 
 which depend upon the mussel. Surveys of this character have now been carried out 
 on the Mississippi River and nearly all of its more important tributaries from Minnesota 
 to Tennessee, and as a result of these investigations an enormous amount of material 
 and information has been collected which, when examined and analyzed, will not only 
 have the greatest economic value, but will constitute one of the most important eco- 
 logical studies ever made on any group of animals. 
 
 I. HISTORICAL. 
 
 As has long been known, the Unionidae carry their young in the gills, which function 
 as brood pouches until the completion of the embryonic development. At the close of 
 this period the larva or so-called glochidium is fully formed and escapes from the egg 
 membrane while still within the gill. In some species the discharge of the glochidia 
 takes place at once, while in others they remain in the brood pouches for several months 
 without further change before being set free into the water. 
 
 The glochidium, long thought to be a parasite infesting the gills and known as 
 Glochidium parasiticum, was proved by Carus in 1832 to be the larva of the mussel 
 itself, although many years earlier Leeuwenhoek had given it the same correct inter- 
 pretation. In 1866 Leydig made the important discovery that the glochidium, after leav- 
 ing the parent, completes its development as a parasite on fishes. 
 
 The earliest observations of importance in the development of our knowledge 
 concerning reproduction in the Unionidae are those of Leeuwenhoek, made about 1695 ° 
 and recorded in the Arcana Naturae. During the two preceding centuries the belief had 
 gained ground that the mollusks had sexes like the higher animals, and this no doubt 
 helped to arouse a certain skepticism regarding the existence of any process of spon- 
 taneous generation among the representatives of this phylum. The observations of 
 Redi (1668), in disproval of spontaneous generation in insects, furnished collateral 
 evidence and appear to have been the direct incentive for Leeuwenhoek's examination 
 of the reproductive processes in certain mollusks, among others the fresh-water mussels, 
 and the discovery by Leeuwenhoek of eggs and sperm in these mollusks convinced him 
 that their reproduction must be effected by such means rather than by spontaneous 
 generation. 
 
 It is surprising to find how accurate were Leeuwenhoek's conclusions regarding the 
 general course of the development as far as the larval stage, later known as the glochid- 
 ium, and a survey of the subsequent literature shows that not until the work of Carus, 
 in 1832, were there published conclusions more in accord with the facts as now known, 
 nor a better summary of what we now term the embryonic period. The correctness of 
 these early observations, so far as they went, and of the conclusions drawn from them 
 have not been sufficiently recognized in most accounts of the literature, and for this 
 reason an explicit statement of their important features is desirable. 
 
 a The date of the publication referred to in the literature list is somewhat later, i jaa. 
 
I 1 2 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 Approaching the subject unhampered by any preconception in favor of the older 
 views, but rather with the belief that the conclusions of Redi would also hold for the 
 bivalves, Leeuwenhoek records, in the 83d and 96th letters of his Arcana Naturae, the 
 presence of separate sexes in Anodonta and Unio, as evidenced by the presence of eggs 
 and spermatozoa in separate individuals, and gives some account of the development. 
 That he clearly apprehended the main course of events is evident if we read his descrip- 
 tion of eggs found floating free in the fluid obtained by puncturing the upper part of 
 the foot upon either side, of similar eggs in more advanced stages within the outer gills, 
 and of various stages in the formation of the glochidial shell. Finally, he observed the 
 snapping of the valves, now so well known as a sign of the last stages in this embryonic 
 development, and upon seeing the rotation of the embryo in the egg membrane he 
 concluded that it must be unattached. He further observed that the individuals, 
 when ready for their egg laying (passage of eggs from ovary to gills), placed themselves 
 in spots where the water was shallow and where they were in direct sunlight — a fact 
 which seems to have been confirmed by other observers of the European species (Schier- 
 holz, 1888, p. 8, Unio and Anodonta). Observing the general similarity between the 
 bivalved larva and the adult, he seems never to have doubted that the glochidia, as 
 they were subsequently called, were the young of the mussel in which they were found 
 and therefore that these mollusks were viviparous, conclusions which so naturally fol- 
 lowed from all the facts that it is hard to see how convincing evidence could have been 
 manufactured for any other opinion. Upon removing these fully formed larvae and 
 setting them aside in dishes of clean water, with a view to observing their further 
 development, Leeuwenhoek met the stumbling block of all observers before the dis- 
 covery of the parasitism upon the fish was known, for the larvae lived but a short time, 
 soon becoming infested with a variety of animalcules, which he rightly concluded were 
 the immediate cause of their death. 
 
 These conclusions of Leeuwenhoek, so nearly in accord with our present knowledge, 
 were not entirely accepted, because they did not become known to some investigators 
 even a century later and because there was still a considerable recrudescence of the 
 older conception of spontaneous generation. The opinion of Poupart (1706) that these 
 mussels were hermaphroditic gained ground and dominated during the eighteenth 
 century, although the larvae, when found in the outer gills, were always regarded as the 
 young of the mussel until, in 1797, Rathke offered an entirely different explanation and 
 erected for them a new genus, Glochidium, and a species, parasiticum. According to 
 this explanation, which came to be known as the Glochidium Theory, it was supposed 
 that these multitudinous larvae were not the young of the mussels at all, but parasites with 
 which they had become infested. Since Rathke's theory attracted considerable attention 
 at the time and was later supported ardently by Jacobson (1828), and since it has given 
 us the term glochidium, we may note in passing the evidence upon which it was based 
 as stated by its later champion. 
 
 1. The form and organization of the little shells is entirely different from that of 
 the adult Unio and Anodonta. 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. I 13 
 
 2. They are of exactly the same form and size in the two genera and in the indi- 
 viduals of diverse size and age. 
 
 3. They are always of the same size and shape when they have reached their com- 
 plete development. 
 
 4. Their valves are of a consistency and hardness in no wise related to their size, 
 as should be the case were they the young of Unio and Anodonta. 
 
 5. Their development is not related to any season of the year nor to a certain age 
 of the animal in which they are found; that is to say, one finds in a single locality at 
 the same time individuals containing eggs, others with little bivalves, and some con- 
 taining even the fully developed organisms. 
 
 6. The enormous numbers which are found at one time in an individual are in no 
 wise proportionate to the number of the adults in any locality. 
 
 7. One can not conceive of organs so delicate as the gills being able to serve as a 
 sort of brood pouch, and there is no other example in the animal series of such a con- 
 dition, although these organs are often the seat of animal parasites. 
 
 Jacobson's statement is thus a curious jumble of half truths and of statements 
 which have since been shown to be entirely incorrect. 
 
 The importance attached to the dispute thus raised was so great that the Academy 
 of Sciences at Paris appointed two of its members, De Blainville and Dumeril, a com- 
 mittee with instructions to examine into and report upon the whole matter. This 
 report (De Blainville, 1828) presents an exhaustive review of the early literature and 
 details certain experiments performed by the committee with a view to testing the 
 matter by direct observation. These experiments, while tending to confirm the earlier 
 views of Leeuwenhoek, were insufficient for the complete overthrow of Rathke's Glo- 
 chidium Theory, for although the report was unequivocal in its conclusion that the 
 observations of all previous authors and the evidence advanced by Rathke himself did 
 not justify the Glochidium Theory, its lack of evidence from original observations 
 rendered it not entirely conclusive. Viewed in the light of our present knowledge, its 
 skillful and logical arraignment of Rathke's conclusions shows clearly the scant foun- 
 dation upon which the Glochidium Theory rested, but it was not until the work of 
 Carus (1832) that the question was finally set at rest. This author was able, in the 
 brightly colored eggs of Unio iittoralis, to see the passage of the eggs from the ovary to 
 the external gills and their development there to the mature glochidia, and thus to 
 prove beyond any doubt that the innumerable larvae which crowded the outer gills were 
 the young of the mussels in which they were found. 
 
 The paper by von Baer (1830) anticipated some of the points which Carus made 
 the more clear, and from this time on the serious difficulty for students of the embry- 
 ology was found in the failure to secure, either within the gills of the mussel, or upon 
 removal of the embryos to water, any developmental stages beyond the glochidium. 
 
 The period from Carus's paper (1832) to the date of the discovery by Leydig (1866) 
 of glochidia embedded upon the fins of fishes shows little progress toward a more com- 
 plete account of even the embryonic stages. De Quatrefages, who in 1836 described 
 
114 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 the glochidium as having a very complex structure and possessing many of the organs 
 of the adult mussel, made a distinctly backward step; and his account of hearts, 
 stomachs, livers, intestines, and aortas, all highly developed and double in each indi- 
 vidual, reminds one of the description of elaborate systems of organs in the infusoria 
 as given by Ehrenberg in his monograph published during the same year. Pfeiffer (1821, 
 taf. 11, fig. E) was the first to observe the minute outline of the glochidium at the umbo 
 of a young shell — a fact which, had it become generally known, would have saved Jacob- 
 son his defense of the Glochidium Theory. There remained, however, the unexplained 
 gap between the glochidium and such a stage of the young mussel, and this was filled 
 only by Leydig's discovery of the parasitism. With the clue thus given, the stages by 
 which the glochidium becomes the miniature adult, during the course of its parasitism, 
 were studied by Braun (1878), Schmidt (1885), Schierholz (1878 and 1888), and more 
 recently by Harms (1907-1909). All of these investigators obtained their material 
 in great abundance by the artificial infection of fish with the glochidia, and in their 
 several accounts the structure of the glochidium and the organogeny of the common 
 European species will be found very completely given. 
 
 The embryonic stages attracted new attention with the rise of cytological studies, 
 and the paper of Flemming (1875) was exhaustive for the period in which it was written, 
 although Lillie's more detailed and modern account (1895) of the cell lineage and the 
 formation of the glochidium in Unto complanatus and Anodonla cataracta has rendered 
 Flemming's paper of historical interest only, and has apparently left undone nothing 
 of importance in a description of the early stages in these species. 
 
 Further reference to the literature will be made as the several stages of the develop- 
 ment are discussed in the species we have followed. Since an excellent summary of the 
 literature, particularly that published since the paper by Carus (1832), may be found in 
 the work of Harms (1909), we omit further elaboration here. The report to the Paris 
 Academy (De Blainville, 1828) gives a good account of the literature for the earlier 
 period, and from this we have obtained a summary of the facts in such early papers as 
 have not been accessible. 
 
 II. REPRODUCTION. 
 
 The sexes are normally separate in the Unionidae, but in Anodonta imbecillis and in 
 a few other species of this genus the occurrence of hermaphroditism has been occasionally 
 recorded (cf. Sterki, 1898; Ortmann, 191 1). Although in the majority of the genera 
 of the Unionidce the sexes are indistinguishable externally, in a few, notably in Lamp- 
 silis, the shell of the female differs from that of the male in its greater convexity in front 
 of the posterior ridge and in more or less well-marked differences in the posterior outline 
 of the shell. In such cases the males and females may be readily assorted without 
 recourse to an examination of the soft parts. 
 
 At ovulation the eggs pass from the oviducts to the cloaca, and thence back into 
 the suprabranchial chambers, in which they are probably fertilized by spermatozoa 
 brought in by the respiratory current of water. From the suprabranchial chambers 
 they are conducted directly into those portions of the gills in which they are to remain. 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 115 
 
 Observations on the passage of the eggs from the ovaries to the gills are extremely 
 meager, and further information is needed concerning the factors involved in directing 
 the stream of eggs from the openings of the oviducts to their final resting place in the 
 water tubes of those regions of the gills which function as brood chambers. We owe 
 to Latter (1891, 1904) the most detailed account of this process which we have, and, in 
 lieu of any direct observations of our own on the subject, we may quote his interesting 
 description (1891) which is based upon Anodonta: 
 
 If a female be taken from the shell at this season (the spawning season) the eggs may be seen through 
 the transparent wall of the oviduct passing singly, but in a steady stream, to the genital aperture. Their 
 motion is due partly to " labour contractions " of the intrinsic muscles of the foot and partly to the cili- 
 ated lining of the oviduct itself. One by one the eggs issue from the genital aperture, whence they are 
 conveyed backwards by the abundant cilia which clothe the external surface of the nephridium. Along 
 the middle line of this surface there is a belt of especially long cilia which appear to be devoted to the 
 transit of the eggs; those dorsal and ventral to the belt work obliquely so as to keep the eggs in contact 
 with it. It is probable that the free dorsal border of the inner lamella of the inner gill plate is, under 
 normal conditions, applied to the visceral mass in this region so as to inclose a temporary tube, one of 
 whose walls is formed by the above-mentioned belt of specialized cilia. In the course of about 50 
 seconds an egg is thus swept back to the slit between the protractor muscle of the shell and the point of 
 fusion of the right and left inner gill lamellae; here they meet the stream of ova from the other side of 
 the body and so reach the exhalent current and the cloaca. 
 
 The process goes on for some 10 days or more in each individual and the number of eggs is immense 
 * * * probably half a million may be taken as a fair average. On reaching the cloaca * * * 
 their direction is reversed and they pass forward into the cavities of the right and left gill plates, which 
 serve as brood pouches. The method by which this change of direction is accomplished is not quite 
 clear. * * * I have, however, observed on several occasions a violent and sudden reversion of the 
 water currents such as would certainly be fully capable of carrying the eggs forward and into the latticed 
 recesses of the outer gills. This reversion is caused by the animal, firstly, closing all the ventral border 
 of the shell by means of the free edges of the mantle assisted by the flexible, uncalcified rim of periostra- 
 cum and leaving the siphons alone open, and, secondly, relaxing the adductor muscles so as to allow the 
 elastic ligament to make the valves gape apart. These actions cause the hydrostatic pressure within 
 the shell to be less than that of the water without and consequently there ensues a rush of water into the 
 shell through the open siphons. The whole procedure may be likened to a gulp and is achieved by 
 precisely similar physical forces. 
 
 This may possibly be the correct interpretation of the process, but additional 
 observations and experiments should be made for verification. Latter also attempts 
 to account for the fact that the eggs in Anodonta pass into the outer gill and not into 
 the inner, but his explanation is unsatisfactory and inadequate. It would be a matter 
 of the greatest interest to discover the mechanism which directs the eggs in the different 
 types of the marsupium into certain water tubes of the gills and not into others. 
 Special structural modifications must be correlated with the particular type as the 
 fundamental cause of these differences, and a very pretty problem is here presented in 
 the determination of such correlations. Since in the genus Quadrula all four gills 
 
 a It is to be remembered that this description is based upon the conditions as they occur in Anodonta, in which the inner 
 lamella of the inner gill is not fused to the visceral mass, and the inner suprabranchial chamber is consequently freely open to the 
 mantle chamber; in those forms, however, in which this lamella is fused for a part or all of its length, the eggs are received into 
 the anterior end of the inner suprabranchial chamber, into which the genital apertures open directly, and pass back through this 
 chamber to the cloaca. 
 
Il6 BULLETIN OK THE BUREAU OF FISHERIES. 
 
 become filled with eggs, a directive mechanism is probably absent in this genus, and a 
 careful comparison of the conditions in Ouadrula with the structure of the gills in 
 those genera in which only a portion of the gills is utilized as a brood chamber might 
 well furnish the clue to the discovery of a special mechanism in the latter. 
 
 While as a rule the great majority of the eggs, when a gravid gill is examined, are 
 found to be fertilized, different species differ markedly in the percentage of unfertilized 
 eggs present, and, in fact, a large proportion of the latter seems to be characteristic of 
 certain genera. In Lampsilis, Symphynota, Anodonta, and a number of other genera it 
 has been very unusual in our experience to encounter any considerable number of unfer- 
 tilized eggs, while, on the contrary, in Quadrula, Pleurobcma, and in some species of 
 Unio it is often true that even a majority of the eggs in a gravid female have failed of 
 fertilization; in fact, in these genera one expects to find a large percentage of such 
 eggs as the usual thing. 
 
 The entire embryonic development takes place in the gills of the female, and at 
 the close of this period the larva or glochidium is fully formed. The differences in the 
 length of time the glochidia are retained in the gills will be discussed later, but after 
 their liberation the completion of their development occurs while they are living as 
 parasites on the fish in all of the Unionidse, so far as known, except in the genus Stro- 
 phitus, whose glochidium, we have recently discovered, undergoes the metamorphosis 
 in the entire absenceof a parasitic stage. This extraordinary case will be referred to later. 
 
 As the embryology of the Unionidae has been described by Lillie (1895) in great 
 detail, and as Harms (1909) still more recently has published an excellent account of 
 the post-embryonic development, we shall omit all reference to the actual develop- 
 mental events, and confine ourselves to a discussion of those phases of the reproduction 
 and parasitism of the Unionidae in which we have been especially interested in connec- 
 tion with the problem of artificial propagation. 
 
 THE MARSUPIUM. 
 
 The term marsupium has been generally used to indicate those portions of the 
 mussel's gills into which the eggs are received from the suprabranchial chambers after 
 ovulation and which serve as brood pouches for the retention and nurture of embryos 
 and glochidia until the discharge of the latter. As no better name seems to be availa- 
 ble, we shall employ it in this paper. 
 
 USE OF THE MARSUPIUM IN CLASSIFICATION. 
 
 Since the extent to which the gills are specialized for this purpose varies in dif- 
 ferent groups of the Unionidae, Simpson (1900), in his "Synopsis of the Naiades," has 
 made use of the marsupium as the chief diagnostic character on which his classification 
 is based. Those groups in which the marsupium comprises the outer or all four gills 
 he designates as the Exorbranchiae, while those in which the inner gills alone receive 
 the eggs are distinguished as the Endorbranchiae. All of the European and North Ameri- 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 117 
 
 can species belong to the former group, while the latter contains forms that are found 
 chiefly in Asia, Australia, Africa, Central America, and South America. 
 
 As our observations have been confined to the Exobranchiae, reference will be 
 made only to this group, the following subdivisions of which are recognized by Simp- 
 son, each distinguished by special marsupial characters: 
 
 Tetragenae: Marsupium occupying all four gills. 
 
 Homogenae: Marsupium occupying entire outer gills. 
 
 Diagenae: Marsupium occupying entire outer gills, but differing from that of the 
 Homogenae in that the egg masses lie transversely in the gills. 
 
 Heterogenae : Marsupium occupying only posterior end of outer gills. 
 
 Mesogenae: Marsupium occupying a specialized portion in the middle region of 
 outer gills. 
 
 Ptychogenae: Marsupium occupying entire lower border of outer gills which is 
 thrown into a series of peculiar folds. 
 
 Eschatigenae : Marsupium occupying the lower border only of outer gills, but not 
 folded. 
 
 Simpson has established another group, the Digenae, for the genus Trilogonia, but 
 since its marsupium is constituted by all four gills (Sterki 1907), it should at least be 
 included in the Tetragenae, if not in the genus Quadrula, as Ortmann maintains (1909, 
 191 1). For a complete list of the genera occurring in each of Simpson's groups, refer- 
 ence may be had to his Synopsis (op. cit., p. 514-515). 
 
 These groups constitute Simpson's subfamily, Unioninae, his other subfamily, 
 Hyrianae (Hyriinae), coinciding with the Endobranchiae or those Unionidae whose mar- 
 supium occupies the inner gills only. In all of the Unioninae except the Heterogenae 
 and Digenae (Tritogonia) , according to Simpson, the sexes are indistinguishable exter- 
 nally. 
 
 It will be seen from the above classification that three general conditions exist 
 in the Unioninae, namely, one in which the marsupial adaptation involves all four 
 gills; one in which the entire outer gills only are utilized; and, lastly, one in which 
 some differentiated portion of the outer gills constitutes the marsupial region. It 
 would, accordingly, be a more logical procedure to make these general marsupial con- 
 ditions the basis of the classification and to recognize only three main groups corres- 
 ponding to the three general types of marsupium, to which the names Tetragenae, 
 Homogenae, and Heterogenae might be applied; and since all of the remaining forms 
 have a marsupium which may be readily regarded as a secondary modification of one 
 or another of the three types, they could be arranged in appropriate subgroups. If 
 this were done, the Diagenae would obviously fall within the Homogenae, while the 
 Mesogenae, Ptychogenae, and Eschatigenae would be placed under the Heterogenae, as 
 in all of the latter forms the marsupium is some specialized portion of the outer gills. 
 
 « Besides the Unionidae, a second family, the Mutelidse, is recognized by Simpson in his classification of the Naiades or pearly 
 fresh-water mussels. In these forms, which belong to Africa and South America, the marsupium is the inner gills only, and the 
 larva is not a glochidium but the so-called lasidium. The genera embraced in this family are not considered in the present account. 
 
Il8 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 Quite recently Ortmann (1910a, 191 1) has proposed an entirely different arrange- 
 ment cf the Naiades which is based upon a study of the anatomy and the larval charac- 
 ters of the fresh-water mussels of Pennsylvania. His system also lays especial stress 
 on the marsupial differentiations, but it involves a number of important modifications 
 in Simpson's classification which he maintains must be radically recast, in the light of 
 the facts which he has discovered, if it is to represent the natural affinities of the group. 
 
 It is not our purpose to present a critical discussion of the relative merits of the 
 two systems, as our only interest in this connection is concerned with the marsupium 
 as an accessory organ of reproduction, but as Ortmann has added a number of important 
 facts to our knowledge of this structure, it is necessary to state briefly the basis of his 
 classification so far as it has to do with the several marsupial modifications. In addi- 
 tion to the marsupial structure, he makes use in his arrangement of families, subfamilies, 
 and genera of a number of other characters which he considers of systematic value; 
 for example, the degree of fusion of the inner lamella of the inner gill with the visceral 
 mass; the dorsal aperture (supra-anal opening); the siphons; the differentiations of the 
 mantle edge; the structure of the glochidium; and shell characters. In contrasting his 
 arrangement with that of Simpson, however, reference will be made only to the 
 marsupium. 
 
 Confining himself to North American forms, he divides the Naiades into two 
 families, the Margaritanidse and the Unionidte. His discovery that in Margaritana 
 margaritijera there are no distinct interlamellar junctions in the gills, but only scattered 
 interlamellar connections, and consequently no definite water tubes, he considers of 
 sufficient importance to warrant him in creating a new family for this genus, Margari- 
 tanidffi, which he has thus sharply set apart from the remaining genera grouped under 
 the Unionidae, a procedure of doubtful wisdom." The fact that complete interlamellar 
 junctions are absent in Margaritana, which is further characterized by certain other 
 apparently primitive features, is of the greatest interest, but that these differences are 
 of sufficient significance to justify a separate family for Margaritana is not at all clear. 
 
 The Unionidae, after the removal of Margaritana, he divides into three subfamilies, 
 distinguished as seen below by definite marsupial characters: 
 
 1. Unioninae. "Marsupium formed by all four gills, or by the outer gills only; 
 edge of marsupium always sharp and not distending; water tubes not divided in the 
 gravid female." 
 
 This subfamily includes the following genera, which, however, he has recast to a 
 considerable extent by subtractions and additions of species: Quadrula Rafinesque 
 (including Tritogonia tuberculata) ; Rotundaria Rafinesque (established for Quadrula 
 tuberculata) ; Pleurobema Rafinesque (including Q. coccinca, pyramidata, obliqua, cooperi- 
 
 tt The condition described by Ortmann for Margaritana is quite similar to that which is found in the gills of Mytilus (cf. 
 Peck, 1877), in which complete interlamellar junctions are absent and the inner and outer lamella? are connected only by scattered 
 strands of subfilainentar tissue passing across the interlamellar space. This similarity in gill structure would argue strongly 
 for the primitive position of Margaritana among the Unionidse. In Lucina these interfilamentar junctions are larger and are 
 provided with blood vessels, while in Mytilus they are non-vascular. Ortmann does not state whether or not they contain blood 
 vessels in Margaritana. 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. Iig 
 
 ana); Elliptio Rafinesque (established for the North American species of Unto to dis- 
 tinguish them from the European). 
 
 2. Anodontinae. "Marsupium formed by the outer gills in their whole length, 
 distending when charged, and the thickened tissue at the edge capable of stretching out 
 in a direction transverse to the gill, but not beyond the edge (or only slightly so) ; water 
 tubes in the gravid female divided longitudinally into three tubes, with only the one in 
 the middle used as an ovisac, and closed at the base of the gill." 
 
 The following genera are grouped under this subfamily: Alasmidonta Say, Stro- 
 phitus Rafinesque, Symphynota Lea, Anodonloides Simpson, Anodonta Lamarck. 
 
 3. Lampsilinae. "Marsupium rarely formed by the whole outer gill, generally only 
 by or within the posterior part of the outer gill; edge of marsupium, when charged, dis- 
 tending, and bulging out beyond the original edge of the gill, generally assuming a 
 beaded appearance; water tubes simple in the gravid female." 
 
 The following genera are grouped together under this subfamily: Ptychobranchus 
 Simpson, Obliquaria Rafinesque, Cyprogenia Agassiz, Obovaria Rafinesque (including 
 Lampsilis ligamentind) , Plagiola Rafinesque, Paraptera gen. nov. (established for Lamp- 
 silis gracilis), Proptera Rafinesque (established for Lampsilis alata, purpuraia, lavis- 
 sima), Lampsilis Rafinesque (including Micromya jabalis), Truncilla Rafinesque. 
 
 It will be seen by a comparison of the genera which Ortmann assigns to his three 
 subfamilies with the several groups of Simpson, that the most significant change intro- 
 duced by the former arrangement is the disruption of Simpson's Hoinogenaa and a 
 redistribution of its genera and those of the Digenas, Diagenae, and Tetragenae among 
 the subfamilies Unioninae and Anodontinae, the former receiving all of the genera con- 
 sidered bv Ortmann, except Alasmidonta, Strophitus, Symphynota, Anodonloides, and 
 Anodonta, which, by reason of the peculiar secondary division of the water tubes of the 
 gravid female in all of these genera, he insists should be placed in a subfamily by them- 
 selves. Apparently his grounds for the rearrangement are sound. In the Lampsilinae 
 are included all of the genera of Simpson's Heterogenas, together with those of the 
 Mesogenae, Ptychogenae, and presumably the Eschatigenae — a procedure which is in 
 harmony with the suggestion made above that the genera in which a differentiated 
 portion only of the outer gill functions as a marsupium should be grouped together. 
 
 The reader is referred to Ortmann's monograph for further details and for the con- 
 siderations which have led him to shift a number of species from one genus to another 
 and to establish certain new genera, while renaming others. 
 
 This system has the merit of being based upon a careful study of the anatomy of 
 the species with which he has been concerned, and he has clearly demonstrated the 
 fact that shell characters alone are not sufficient for a determination of true relation- 
 ships. To what extent his classification will replace Simpson's remains of course to be 
 seen, but in any future discussion of the matter the new facts brought to light by Ort- 
 mann in his study of the structural modifications of the marsupium must be reckoned 
 with. 
 
 18713° — 12 2 
 
120 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 GENERAL STRUCTURE OF THE MARSUPIDM. 
 
 In connection with our investigations on fresh-water mussels we have had occasion 
 to give quite a little attention to the anatomical and histological structure of the mar- 
 supium in a number of genera, and, furthermore, we have been particularly interested 
 in the changes that occur in the gills during the period of gravidity. We have already 
 published a brief account (1910b) of some of our observations on the marsupium, with 
 illustrations of the more important types, but, as Ortmann has since added a number 
 of new facts to the subject, it is advisable to present our results in greater detail and 
 with additional illustrations. For this purpose it will be more convenient to follow 
 Simpson's arrangement, and we shall refer to the species examined by us under the 
 several groups established by him. It will also be convenient in connection with the 
 description of the marsupium to refer somewhat incidentally to certain observations on 
 breeding habits, characteristics of the embryos, and related matters. The finer struc- 
 ture of the marsupium is reserved for a subsequent section of this report. 
 
 Tetragenm. — The marsupium in these forms comprises all four gills, a condition 
 which is undoubtedly the most primitive one among the Exobranchife. It is the con- 
 dition occurring in the genus Quadrula, in which, following Ortmann, we include Trilo- 
 gonia. We have encountered it in the following species: ebena Lea, heros Say, lachry- 
 ■mosa Lea, metanevra Rafinesque, obliqua Lamarck,™ plicata Say, puslulosa Lea, trigona 
 Lea, tuberculata Barnes (Tritogonia tuberculata) , and undulala Barnes. 
 
 No special structural modifications are present beyond the usual glandular folded 
 epithelium covering the surface of the interlamellar junctions which, as has been known 
 since the work of Peck (1877), are closer together in the marsupial than in the purely 
 respiratory gill. The gills when gravid, although somewhat distended and padlike in 
 appearance, never become swollen to the extent that is seen in many other genera. 
 Figure 5, plate vn, which is drawn from a gravid female of Quadrula ebena, illustrates 
 the typical appearance of the marsupium in this group, although the gills shown in the 
 figure are not as fully distended as is frequently the case. 
 
 In ebena&nd trigona the ovarian eggs and the embryos are frequently brilliantly colored 
 red or pink and when the marsupium is charged the color shows through the colorless 
 transparent walls of the gills, which present a striking appearance on removing the shell. 
 In all of the other species of Quadrula observed by us the pigmentation is absent, but in 
 ebena and trigona the color is found in a majority of the gravid females, the number of 
 such cases being somewhat greater in trigona (over two-thirds of all gravid females 
 examined in this species) than in ebena. The red pigment, however, whenever it occurs, 
 does not persist, but on the contrary totally disappears in the later stages of embryonic 
 development, and by the time the glochidia are fully formed no trace of it is left. We 
 have never seen a single case of a red or pink glochidium either in these two species of 
 Quadrula or in any other genus in which pigmented eggs and embryos occur. It is true 
 
 a Ortmann (op. cit., 1911, p. 330) states that only the outer gills serve as the marsupium in obliqua, and on this ground he 
 has removed the species to Pleurobema. If we have made no mistake in the identification of our specimens, our observations 
 on this species are not in accord with his. 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 121 
 
 that the marsupium may still be more or less deeply tinged with red, even when it con- 
 tains fully developed glochidia, but this is due to its containing a variable number of 
 unfertilized eggs, which do not lose the color, and not to the glochidia which are always, 
 as stated, entirely colorless. 
 
 The occurrence of unfertilized eggs is very common in all of the species of Quadrula 
 which have come under our observation, and their presence is more characteristic of 
 certain species than of others. They are quite rare in plicaia and pustulosa, for example, 
 less so in metanevra, common in ebena, while in trigona, in which they occur more fre- 
 quently than in any other species of Quadrula, they were found in a large majority of 
 cases. The number of unfertilized eggs in different females of a given species varied 
 from cases in which only a few such eggs were scattered among normal embryos all the 
 way to cases in which the marsupium contained no normal eggs or embryos at all. Eggs 
 which have not been fertilized, after remaining in the marsupium, become swollen and 
 stratified (see below) , frequently forming exovates and undergoing fragmentation before 
 final disintegration. 
 
 There seems to be a definite correlation between the presence of unfertilized eggs in 
 the marsupium and the occurrence of trematode parasites in the testis of the male; in 
 species like plicaia, in which unfertilized eggs were rare, only occasionally were the testes 
 infested with worms, but in trigona, for example, the trematodes were found in a large 
 number of males. It is not at all improbable that the amount of sperm available in a 
 given locality is greatly reduced as a result of the castration of males by this testis infest- 
 ing parasite. 
 
 The abortion of embryos and glochidia, which is so characteristic of the genus 
 Quadrula, and the significance of this peculiarity will be referred to later on. 
 
 Homogence. — The condition in which the entire outer gills only are utilized as a 
 marsupium is present in 1 6 genera, according to Simpson." We have verified its occur- 
 rence in Alasmidonta truncata Wright; Anodonta cataracta Say, grandis Say, implicata 
 Say; Arcidens confragosus Say; Plcurobcma cesopus Green; Symphynota complanata 
 Barnes, costata Rafinesque; and in Unio complanalus Dillwyn and gibbosus Barnes. 
 
 As has already been stated, Ortmann has disrupted the group, placing Plcurobema 
 and Unio in his subfamily Unioninae, while segregating Alasmidonta, Anodonta, and 
 Symphynota in his Anodontinas. This he has done chiefly because of a differentiation of 
 the ventral border of the marsupium and of a secondary division of the water tubes of 
 the marsupium in those genera included in the Anodontinae. These differences will 
 be referred to below. 
 
 The marsupium when filled with embryos or glochidia may be greatly distended 
 beyond its normal dimensions, and in this condition is an enormously swollen padlike 
 structure, with a smooth surface, filling a large portion of the mantle chamber. Figure 3, 
 plate vi, represents the gravid marsupium of Symphynota complanata, which may be 
 taken as typical of the Homogenae, although in Plcurobema and Unio the distension 
 is not so great. 
 
 a Margaritana is placed in this group by Simpson, but as it utilizes all four gills as the marsupium it should be included 
 with the Tetragenae. 
 
122 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 In Pleurobema w-sopus the eggs and embryos, like those of Quadrula ebcna and trigona, 
 are usually, but not always, colored red or pink, but the glochidia are invariably unpig- 
 mented. Unfertilized eggs in varying proportions are frequently found in this species 
 either mixed in with embryos at all stages of development or occurring alone; such eggs 
 always show a definite stratification of the egg substances. 
 
 Diageruz. — This group was established by Simpson to receive the genus Strophitus, 
 in which the marsupium occupies the entire outer gill and in external appearance is 
 similar to that of the Homogenae. But it is unique among the Unionidae in that the 
 embryos and glochidia are embedded in gelatinous cords (called "placentae" by Sterki, 
 " placentulae " by Ortmann), which lie transversely in the gills, whereas in all other cases 
 the egg masses are placed vertically, each one occupying an entire water tube. In 
 Strophitus, on the other hand, the cords are packed closely together, like chalk crayons 
 in a box, a variable number being contained in a single water tube, while the blunt ends 
 of the cords are distinctly seen through the transparent external lamella of the outer 
 gill. It should be stated that Ortmann (1910b, 191 1) has found that the discharge of 
 the cords is not through the lamellae of the gills, as Simpson (1900) has maintained, but 
 that it occurs in the usual manner through the supra-branchial chambers. A description 
 of the unique cords and the extraordinarily interesting life history of Strophitus is reserved 
 for a special section. 
 
 HcterogeruB. — In this group the marsupium occupies only the posterior portion of 
 each outer gill, varying in extent from about one-third to two-thirds of the entire length 
 of the latter. In young females the marsupium is shorter and not so fully distended as 
 in older ones. In fact, it is true of all Unionidae that the marsupium is less heavily 
 charged when the female is young. The differentiation of the posterior region is very 
 conspicuous even in the non-gravid female, as the marsupium is sharply marked off 
 either by a distinct fold or a notch from the anterior respiratory part, and, since it is much 
 deeper dorso-ventrally than the latter, it projects farther down into the mantle-chamber. 
 Its walls are also more membranous in appearance than are those of the respiratory 
 region, and after the discharge of the glochidia it is seen as a flabby collapsed pouch. 
 
 When gravid, the marsupium may be enormously swollen, the expansion being 
 greater along the ventral border than above, where, owing to its fixed position, it is inca- 
 pable of stretching. This greater ventral extension often causes the marsupium not only 
 to assume a fan-shaped form, which is so characteristic an appearance in Lampsilis, 
 but also to project forward under the respiratory portion, which in consequence becomes 
 sharply folded over on the outer surface of the marsupium. Not only is the marsupium 
 as a whole expanded in the way described, but each of its swollen water tubes is dis- 
 tended distally beyond the lower extremity of the interlamellar junctions so that the 
 ventral border becomes fluted or corrugated, as shown in figure 1, plate vi. This figure, 
 which illustrates the typical condition in the genus Lampsilis, is drawn from a gravid 
 female of L. subrostrata when fully charged with glochidia. The folded respiratory por- 
 tion of the gill, the fan-like expansion of the marsupium, and the corrugated border are 
 all clearly seen. 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 1 23 
 
 When the marsupium is less heavily charged, as in young females, the ventral 
 expansion may not be great enough to cause the conspicuous fold just described, and in 
 cases like this the marsupium, which may then appear kidney-shaped, is marked off 
 from the respiratory end merely by a notch by reason of its greater depth. Such a 
 case is seen in figure 6, plate vn which is taken from a gravid female of L. recta. 
 
 Simpson has included 14 genera in the Heterogense, only three of which, however, 
 have come under our observation, namely, Lampsilis (including Proptcra), Obovaria, 
 and Plagiola. We have recorded this type of marsupium in Lampsilis alaia Say, anodon- 
 ioides Lea, gracilis Barnes, higginsii Lea, Icevissima Lea, ligamentina Lamarck, luteola 
 Lamarck, recta Lamarck, subroslrata Say, and ventricosa Barnes; in Obovaria ellipsis 
 Lea; and in Plagiola elegans Lea and securis Lea. 
 
 No case of pigmented eggs has been encountered by us in this group, and unferti- 
 lized eggs in the marsupium are exceedingly rare. 
 
 MesogeruB. — This group is so designated by Simpson to include the genera Cy pro- 
 genia and Obliquaria, in which a variable number of enlarged water tubes in the middle 
 region of the outer gill are specialized as the marsupium, a larger anterior and a shorter 
 posterior portion of the gill retaining the ordinary respiratory character. We have 
 studied the condition in Obliquaria rcflexa Rafinesque and also in Cyprogenia irrorata 
 Lea, in which the structure of the marsupium is essentially the same, although the two 
 cases differ strikingly in general appearance. 
 
 The marsupium of Obliquaria rcflexa is shown in figure 7, plate vu. Here the modi- 
 fied water tubes, which project far down below the border of the rest of the gill, appear 
 enormously swollen when gravid and show a tendency to curve backward, the degree 
 of curvature becoming progressively greater in the tubes from the anterior to the pos- 
 terior end of the marsupium. A gradual decrease in the length of the tubes takes place 
 in the same direction. The tubes are slightly larger at their distal ends, so that their 
 form is somewhat club-shaped; this is seen more clearly in the shape of the egg masses 
 which form perfect casts of the cavities of the tubes (fig. 42, pi. xi). The corrugation 
 of the lower border of the marsupium is very conspicuous in the figure. The number 
 of water tubes comprising the marsupium in this species is not at all constant, but on 
 the contrary varies in the individuals examined by us from two to eight; according to 
 Simpson, they range from four to seven. During the breeding season each tube is entirely 
 filled with embryos or glochidia which adhere so firmly together that they form a mass 
 of tenacious consistency. 
 
 In Cyprogenia, the only other genus included in the group, the marsupium may 
 be regarded as a further development of the condition seen in Obliquaria. We have 
 observed it in but a single individual of C. irrorata, which was kindly sent to us by Dr. 
 R. E. Coker. This specimen, which contained fully formed glochidia, was collected 
 in the Cumberland River, Kentucky, in November, 1910. The tubes of the marsupium, 
 which present a most striking and unusual appearance, spring from near the middle 
 of the outer gill, are enormously elongated, and curved backward into a close coil, a 
 part of the coil passing under the posterior unmodified portion of the gill, as the tubes 
 
124 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 are turned slightly inward toward the median plane. The marsupium is well shown in 
 figure 8, plate VII. The distension of the marsupial water tubes begins at quite a distance 
 above the ventral border of the rest of the gill, as is seen in the figure. The anterior 
 respiratory portion is sharply separated from the rest of the gill by a cleft which extends 
 almost up to the level of the suprabranchial chamber. At first this was supposed to 
 be an artificial split, but as it occurs on both sides and its edges are perfectly smooth 
 and show no indication of injury, we have concluded that it must be a normal condition. 
 Unfortunately we have had no other specimens with which to compare it. 
 
 In our specimen the marsupium is slightly tinged with pink, the color being due to 
 unfertilized pigmented eggs which are scattered among the glochidia. Simpson speaks 
 of the marsupium as being purple. 
 
 The unusual form of the marsupium in Cyprogenia was originally described by 
 Lea (1827) in C. irrorata, but curiously enough he reversed the direction of the coil 
 in his figure, which appears to have been drawn from memory, as such a mistake could 
 hardly have been possible if he had had a specimen before him. a 
 
 Call (1S87) many years later described a similar marsupium in C. aberti Conrad, 
 which he very crudely figured. It is strange that, although he reproduces Lea's original 
 figure of irrorata by the side of his own, he makes no mention of the error in it. Judging 
 from Call's figure, the number of tubes in the marsupium of aberti is much larger than in 
 irrorata. He shows about 20, while Lea states that there are 7 or 8 in the latter, and 
 in our specimen there are 7. Simpson gives the number for the genus as 7-23. 
 
 Ptychogeruz. — This group contains a single genus, Ptychobranchus. The marsupium 
 occupies the lower half of the entire outer gill and is thrown into a series of folds, from 
 6 to 20 in number, according to Simpson. Each water tube of the marsupium is inflated 
 at its distal extremity to form a globular enlargement projecting beyond the interla- 
 mellar junctions — a condition which gives to the free edge of the gill the beaded appear- 
 ance so characteristic of the genus. This marsupium is well illustrated in figure i, 
 plate vi, which is drawn from a gravid female of P. phaseolus Hildreth. Seventeen 
 conspicuous folds, sharply demarked from each other, are shown in the figure, in which 
 the beaded border of the gill is also clearly seen. 
 
 Eschatigena?. — Simpson has established this group to receive the genus Dromus 
 in which the marsupium occupies the ventral half of the outer gill throughout the greater 
 portion of its length. We are indebted to Dr. R. E. Coker for several specimens of 
 Dromus dromus Lea, obtained from the Cumberland River in Kentucky in November, 
 1910, which have furnished the material for our study of this type of marsupium. Three 
 gravid females, all containing glochidia, were included in the lot. 
 
 As seen in figure 4, plate vii, the line of demarcation between the dorsal respiratory 
 portion and the ventral marsupial region is quite sharp and regular, owing to a constric- 
 tion of the gill where the two regions join. Below this line the gill is swollen to an extent 
 varying with the degree to which it is charged with glochidia. The anterior end of the 
 gill is not included in the marsupium and is sharply folded over on the outside of the 
 
 a We are indebted tu Mr. Bryant Walker, ot Detroit, fur having called our attention to this error in Lea's fiyure. 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 1 25 
 
 marsupium in this region. The depth of this fold varies with the fullness of the mar- 
 supium, as the greater is the distension of the latter the farther forward it is tucked 
 under the anterior respiratory region. Posteriorly the two portions of the gill are sharply 
 defined by a deep cleft, as shown in the figure. The surface of the marsupium is thrown 
 into an irregular series of low undulating folds which are more prominent in the more 
 heavily charged females. In two of the females the marsupium is a salmon pink, while 
 the third is colorless, but here, as in the other cases described in which glochidia are 
 present, the color is due to unfertilized eggs. 
 
 The record in our notes of the three females is as follows: 
 
 No. 1, small specimen, 44 by 39 mm. Marsupium colorless, only slightly distended 
 and not thrown into folds or undulations; no anterior fold, merely a notch; glochidia 
 colorless. 
 
 No. 2, larger specimen, 57 by 52 mm. Marsupium salmon pink, much fuller than 
 no. 1, and thrown into distinct folds; deep anterior fold; glochidia colorless, but many 
 pigmented unfertilized eggs and abnormal embryos mixed with them. (This is the 
 specimen from which the figure was drawn.) 
 
 No. 3, largest specimen, 58 by 55 mm. Marsupium with just a tinge of pink, more 
 heavily charged than either of the others and showing prominent folds or undulations; 
 deep anterior fold; glochidia colorless, and a few pigmented unfertilized eggs and 
 abnormal embryos present. 
 
 It is evident from this comparison that the smaller, and therefore presumably the 
 younger, females are less heavily charged than the larger and older ones; and, further- 
 more, that those changes in the gill which are the mechanical effects of gravidity, like 
 the folds, vary directly with the degree of distension of the marsupium. This conclu- 
 sion holds good for all the Unionidse which we have had an opportunity of examining, 
 and also applies to the experience of other observers. 
 
 The glochidia of Dromus dromus, which are excessively minute and of unusual 
 form, being kidney shaped, are referred to later. 
 
 INTERNAL STRUCTURE OF THE MARSUPIUM. 
 
 The marsupium of the Unionida? furnishes a beautiful illustration of a remarkable 
 diversity of form in the adaptation of an organ for a specialized function. One can not 
 study this structure in the North American Unionida without being forcibly impressed 
 with the great variety of detail which one and the same general adaptation is capable 
 of exhibiting. But whatever be the special direction which the modification has taken, 
 even in the most bizarre forms of the marsupium, like that of Cy progenia, there is never 
 any doubt as to the relation between the structural specialization and the function 
 which it is adapted to perform. The structural basis of the marsupium — one might 
 almost say the unit of structure — is the water tube, and it is from an investigation of 
 its finer structure and its relation to other tubes, similarly modified, that an under- 
 standing of the unionid marsupium is gained. The fundamental adaptation is a series 
 of compartments in the interior of the gills provided with a specialized glandular 
 
I2 6 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 epithelium lining the cavity and also with a mechanism in its walls which allows of 
 distension, often to an extraordinary degree. 
 
 The various types of marsupium are to be referred to differences in the manner in 
 which these compartments are associated to constitute the marsupium; to different 
 degrees to which the compartments are developed ; to differences in the modification of 
 the walls for the purpose of distension; and also to the development of special adapta- 
 tions in certain forms for increased aeration of the marsupium. Whether in the last 
 specialization the better aeration is needed for the gravid mussel, whose respiration 
 must be considerably interfered with when the entire outer gills are gorged with embryos, 
 as in Anodonta and Symphynota, or for the embryos themselves, is a question that is 
 discussed later, but from a comparison of the conditions existing in the different types 
 of marsupium it would seem that the respiratory modifications are primarily for the 
 adult and not for the embryos. The reasons for this conclusion should be reserved 
 until the internal structure of the marsupium has been described. 
 
 It is chiefly to Peck (1877) that we owe a correct interpretation of the structure 
 of the lamellibranch gill. It was he who first showed that the plate-like gills of the 
 higher forms, consisting each of an outer and an inner lamella, are formed by a series 
 of juxtaposed independent filaments, a fact that was essential to the later recognition 
 of a perfectly regular series of gradations throughout the lamellibranchs from the simple 
 ctenidium of the primitive Nucula to the complex double gill of the Unionidse. In the 
 least modified forms the filaments are straight, either plate-like or filamentous, but in 
 forms above these each filament becomes greatly elongated and bent upon itself to form 
 a compressed U or V, consisting of an inner and an outer limb. One limb, the inner in 
 the outer gill and the outer in the inner gill, is fixed above to the body wall, while the 
 other limb is free in the lower groups (Area, Mytilus), fixed in the higher (Unionidae), 
 although the inner limbs, forming the inner lamella of the inner gill, may not all be 
 fused to the body wall. The filaments constituting a lamella are interlocked either by 
 cilia or by interfilamentar junctions, and the gill may be further strengthened by inter- 
 lamellar junctions, which are either simple bars (Mytilus, Margaritana) or continuous 
 septa (Unionidae, except Margaritana). 
 
 In his study of the lamellibranch gill Peck described in much detail and with great 
 accuracy the structure of the gills of the Unionidae, and his account has furnished the 
 basis of all subsequent descriptions. The typical structure of the unionid gill is well 
 known. Each gill consists of two lamellae, an outer and an inner, composed of series 
 of juxtaposed filaments supported by chitinous rods and fused by the interfilamentar 
 junctions except where the inhalent ostia open into the interlamellar space for the 
 entrance of water. The dorsal edge of the inner lamella of the outer gill and of the 
 outer lamella of the inner gill is fixed to the body wall, while the outer lamella of the 
 outer gill is fused to the mantle (in Margaritana it is free posteriorly), and the inner 
 lamella of the inner gill is either free or more or less attached to the visceral mass (cf. 
 Ortmann, 1911). The two lamellae are continuous along the free ventral borders, and 
 thus form a flattened sac whose cavity opens above throughout its entire length into 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 1 27 
 
 the suprabranchial chamber; the four suprabranchial chambers lead posteriori)- into 
 the cloaca, which in turn opens to the outside water through the exhalent siphon. The 
 entire gill is subdivided by a series of close-set septa, the interlamellar junctions (except 
 in Margaritana) which separate the interlamellar space into a series of so-called water 
 tubes. Water in the mantle chamber is driven by the cilia guarding the ostia through 
 the lamellae into the water tubes, whence it passes into the suprabranchial chambers 
 and out through the exhalent siphon. The walls of the gill ate traversed by blood 
 vessels and lacunar blood spaces, and the current of water which passes through the 
 gill is a respiratory current. 
 
 The water tubes are lined by an epithelium which is ciliated, at least in some species, 
 on the inner faces of the lamellae, while it assumes a characteristic glandular nature on 
 the inner faces of the interlamellar junctions. The lamellae and the interlamellar junc- 
 tions are richly supplied with elastic and smooth muscle fibers, which are especially 
 highly developed in the junctions of the marsupial gills of the female — evidently in 
 adaptation to the great distensibility of which the latter are capable. In fact, the 
 purely respiratory and the marsupial gills exhibit a number of structural differences, 
 most of which were recognized by Peck (op. cit.) and which are undoubtedly to be 
 accounted for on the ground of the difference in function between the two kinds of gills. 
 Peck clearly described and figured the anatomical differentiation between the respiratory 
 and the marsupial gill in Anodonta, and pointed out, among other distinguishing marks, 
 the fact that the interlamellar junctions in the latter are not only thicker and wider 
 and are covered by a peculiar folded epithelium, but that they are set much closer 
 together. It will be well here to quote his description (op. cit., p. 59-60): 
 
 The interlamellar junctions in the outer gill plate (the marsupial gill) are, like the vertical vessels, 
 more numerous than those of the inner plate, occurring at intervals of seven filaments. They are long 
 ridges of dense lacunar tissue, running vertically from base to apex of the gill plate, and have a much 
 greater size, measuring more from one lamella to the other than those of the inner gill plate. In fact, 
 they are capable of very great extension, which takes place when the outer gill plate has its interlamellar 
 space occupied by the glochidian young of Anodon (pi. v, fig. 4). This great depth of the interlamellar 
 functions of the outer gill plate is their most remarkable feature, as compared with those of the inner 
 plate. It is accompanied by a different disposition of the vertical vascular trunks; for, whilst these 
 in the inner gill plate lie in the interlamellar junctions, in the outer gill plate they lie in the subfila- 
 mentar mass of concreted tissue at the line of origin of the great ridges which act as interlamellar junc- 
 tions. In consequence of this arrangement there are two vertical vessels in the outer gill plate to each 
 interlamellar junction, whereas there is only one to each junction in the inner plate. The arrange- 
 ment of these parts in the outer gill plate is no doubt correlated with its function as a brood pouch. * * * 
 The difference just noted between the outer and inner gill plates, due to the frequency of interlamellar 
 junctions and their relation to the vertical vessels, is accompanied by a further difference of form, which 
 is obvious when the sections given in plate v, figures 2 and 3, are compared. In the outer gill plate 
 the two lamellae are parallel to one another and of equal thickness. In the inner gill plate the outer 
 lamella is thicker than the inner, and its surface is thrown into a series of folds. 
 
 He figures very clearly the conditions described in both a non-gravid and a gravid 
 outer gill and also in the purely respiratory inner gill, and it is clear from his descrip- 
 tion that the peculiarities of the outer gill of the female are permanent differentiations 
 and are not merely present during gravidity. We have repeatedly observed the same 
 
128 BULLETIN OP THE BUREAU OF FISHERIES. 
 
 differences as described by Peck, not only in Anodonta but in a number of other genera, 
 and have also determined that the gills of the male are like the inner gill of the 
 female with respect to the frequency of the interlamellar junctions and the character 
 of their epithelium. 
 
 Peck's description has formed the basis of all of the textbook accounts of the 
 structure of the unionid gill, and two of his figures, showing the differences between 
 the inner and outer gills of Anodonta, are reproduced in Parker and Haswell's Text- 
 book of Zoology, volume i, page 638. 
 
 Ortmann (191 1) was evidently unacquainted with Peck's work, as he describes 
 essentially the same differences between the marsupial and respiratory gills but 
 without reference to Peck. He is the first, however, to show that the same differentiation 
 holds good throughout a wide range of genera. In this connection he states that he 
 "made a very important discovery, namely, that in all our Unionidce the anatomical 
 structure of the gills, which serve as marsupia, is permanently differentiated" (op. cit., 
 p. 283). He then describes in detail the points of difference, showing that in the 
 marsupial gill of the non-gravid female the interlamellar junctions, besides being more 
 numerous, are thicker and wider and are covered by an epithelium which is folded and 
 thrown into wrinkles, often of considerable proportions, whereas in the male and in the 
 respiratory gill of the female the epithelium is simple and unfolded (cf. Peck). "There 
 is no question," he says, "that this peculiar structure of the septa of the marsupial 
 gills is an adaptation to their function" — a conclusion long ago arrived at by Peck. 
 It should be stated that Ortmann has discovered another differentiating character 
 between the inner and outer gill, namely, a longitudinal furrow along the ventral border 
 of the inner gill which is entirely absent in the outer. This furrow is present in both 
 males and females. A similar furrow is figured by Peck for the gill of Mytilus, but 
 the figure in which it is shown is stated to be from the outer gill (op. cit., pi. iv, fig. 10). 
 
 Ortmann, in his careful study of the structure of the marsupium, has described a 
 number of constant differentiations, hitherto unrecognized, which distinguish the 
 several groups established by him in his system of classification. We are relieved, 
 therefore, of the necessity of a detailed description in this place, and reference may be 
 had to his interesting paper. It should also be stated that one of our former students, 
 Mr. J. L. Carter, is now engaged in making a comparative study of the unionid marsu- 
 pium in a large number of genera, and his investigation, which was undertaken pri- 
 marily for the purpose of following the changes, both anatomical and histological, 
 occurring in the gill from the pre-gravid to the post-gravid condition, is now well under 
 way. Although Ortmann's work has, in part, rendered this investigation unnecessary, 
 nevertheless Mr. Carter's study will contribute a number of facts, especially facts of 
 a histological character, which are not included in Ortmann's observations. 
 
 Only a brief reference here to the internal structure of the marsupium is called for 
 under the circumstances, and, since we shall need to compare our observations with 
 those of Ortmann, it will be a matter of convenience to refer to them under the three 
 subfamilies which he has distinguished. As we have not had an opportunity of exam- 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 129 
 
 ining the marsupium of Margaritana, we have nothing to add to Ortmann's description 
 of this genus, and shall confine ourselves to the Unionidaa as restricted by him. 
 
 Unioninee. — In this group there is, as Ortmann has shown, the least amount of 
 differentiation and the structure of the marsupium most closely approaches that of the 
 respiratory gill. Aside from the usual permanent differences, namely, the greater 
 frequency of the interlamellar junctions, their increased thickness and width, and the 
 folding of the glandular epithelium, there is little else to distinguish the marsupial from 
 the respiratory gills in this subfamily. Figure 50, plate xin, which shows a cross section 
 of two water tubes (w. t.) from the gravid outer gill of Quadrula ebena, represents the 
 typical appearance in the genera embraced in this subfamily. Only two embrvos are 
 drawn in the figure, although actually the water tubes are filled with them. The 
 interlamellar junctions (i. j.) are set very close together, at intervals of about five fila- 
 ments, and the marsupium is capable of onlv moderate distention. The epithelium 
 covering the inner surface of the lamellae is low and ciliated, while that of the inter- 
 lamellar junctions is high and glandular and exhibits irregular ridges and furrows. 
 The folds of the epithelium are always of course far more pronounced in the non-gravid 
 gill, as in this condition the interlamellar junctions are not stretched as they are when 
 the gill is charged with embryos. The throwing of the epithelium into folds and the 
 bending and crumpling of the septa themselves, when not under tension, is undoubtedlv 
 due to the elastic fibers which are wavy and wrinkled in the non-gravid gill, while they 
 are drawn out nearly straight when the marsupium is full. 
 
 When highly magnified, as in figure 64, plate xv, the epithelium, resting upon a 
 base of connective tissue and smooth muscle fibers and elastic fibers, is seen to be com- 
 posed chiefly of greatly swollen cells, whose vacuoles are filled with a clear mucus-like 
 colorless fluid. Scattered among these gland cells and seemingly often lying within 
 the vacuoles are seen several smaller and darker nuclei which are the nuclei of leuco- 
 cytes (1). In fact, there can be no doubt that the epithelium becomes infiltrated with 
 wandering blood cells from the underlying blood sinuses in the interlamellar junctions, 
 and many indications are present that seem to show that these cells actually wander 
 through the epithelium into the cavities of the water tubes, but what their ultimate 
 fate is, if this be the case, we are as yet unable to say. There is some evidence that 
 they are ingested by the mantle cells of the glochidia in species that carry the larvce 
 over the winter, like Lampsilis, but of this we can not be certain. 
 
 The above description of the epithelium of the interlamellar junctions will apply 
 in essential respects to the marsupium of all of the Unionidae that we have examined, 
 for the same characteristic histological structure is present everywhere. 
 
 Anodontince. — Ortmann has discovered in the genera which he places in this sub- 
 family a most remarkable differentiation which is evidently an adaptation for increased 
 aeration during the period of gravidity, as it totally disappears after the glochidia are 
 discharged and does not reappear until the onset of the next period. He describes the 
 condition as follows (1911, p. 324, 325): 
 
 Here each ovisac of the gravid female is not formed by a whole water tube, but only by a part of 
 it, the middle one, which is separated from two lateral canals by a folding up of the epithelium of the 
 
130 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 septa (interlamellar junctions). In addition, the ovisacs are closed above at the base of the marsupial 
 gill, thus forming a completely closed sac within each water tube. In one case (Strophitus) this sac 
 is again divided into secondary compartments. * * * This peculiar structure of the marsupial 
 gill is developed only in the gravid female, and is absent in the sterile (nongravid) female. These 
 characters are apparently connected with the prolonged breeding season, and the peculiar secondary 
 water tubes serve for the aeration of the embryos in the marsupium. 
 
 In a preliminary announcement of his new system of the Unionidte (1910a) he 
 briefly stated this discovery in the following words: 
 
 Water tubes in the gravid female divided longitudinally into three tubes, one lying toward each face 
 of the gill, the third in the middle; only the latter contains eggs or embryos, and is much larger than 
 the other tubes. This division into three parts is not present in the sterile female. 
 
 The statement of the presence of these lateral compartments of t he water tubes of 
 the gravid female, made in this brief form and without illustrations, misled us and 
 seemed at that time not to be in accord with our own observations on the marsupium 
 of Alasmidonla, Anodonta, and Symphynota, three of the genera included by Ortmann 
 in the Anodontinx. We had, it is true, seen narrow slit-like spaces lying opposite the 
 outer and inner faces of the water tubes, which were evidently not blood vessels, as the 
 ostia opened freely into them. We interpreted them as differentiations within the 
 lamellte themselves and supposed that they were merely collecting canals into which the 
 ostia opened from the outside and which led by irregular apertures on the other side into 
 the water tubes, as our sections showed here and there interruptions (now known to 
 have a different significance) in the inner wall of these canals. It did not occur to us 
 that these might be the lateral divisions referred to by Ortmann, as, in the sections of 
 the marsupium in which we had seen them, they appeared so evidently to lie wholly 
 within the lamc-llse. 
 
 We were, however, in error, and our failure to recognize that these were really 
 divisions of the water tubes was due to the fact that the sections studied by us were 
 taken from near the ventral border of the gill, where the spaces are much narrower 
 and more slit-like, and also to the fact that at that time we had not happened to see 
 the lateral divisions in the process of being cut off from the water tubes during the 
 early stages of gravidity. Thinking that Ortmann had made some mistake in his obser- 
 vations, we unfortunately published a note (Lefevre and Curtis, 1910a) to this effect 
 and stated that no such division of the water tubes in the three genera referred to was 
 present. A more careful examfnation of our material, however, and a study of mar- 
 supia at different stages of gravidity showed us that Ortmann was entirely correct, and 
 we wish to express our regret at the overhasty publication of our note. The true facts 
 of the case are as Ortmann has stated them to be, although he has only very briefly 
 described the method of formation of the secondary septa which divide the lateral 
 compartments from the central portion of the water tube in which the embryos are 
 confined. Speaking of the origin of the septa, he says (191 1, p. 293): 
 
 In specimens where the eggs begin to go into the gills, this structure (the lateral divisions of the 
 water tubes) is sometimes not developed, but it appears soon, and the epithelial folds, which form the 
 secondary septa within the water tubes, begin to grow into the lumen of the water tubes, and the folds 
 of the opposing faces of the two septa finally unite in the middle. The point of union (cross section of 
 the line of union) is often distinctly seen in sections. 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH- WATER MUSSELS. 13I 
 
 At the outset of gravidity, vertical septa begin to grow out in all of the water 
 tubes of the marsupium from the surfaces of the interlamellar junctions close to the 
 inner and the outer lamellae of the gill. On each side of the gill one septum projects 
 posteriorly, while the other extends anteriorly, and the two meet halfway across the 
 cavity of the water tube. The free edges of each pair of opposed septa then fuse along 
 their entire extent from the ventral border of the gill to the supra-branchial chamber. 
 Specialized elongated epithelial cells forming a serrated border cover the free edge of 
 each septum, and, when the two edges meet, these cells interlock and fuse (fig. 56, pi. 
 xiv). In this way a space, quite narrow and slit -like below, but expanding gradually 
 toward the supra-branchial chamber, is cut off from the water tube on either side, lying 
 between the lamella and the large median division of the tube. As the septa unite, 
 the eggs become confined entirely within the large central space of the original water 
 tube, as Ortmann has stated, and it is this median division alone that functions as the 
 marsupial cavity. We shall speak of the lateral spaces as the respiratory canals, as 
 their function is undoubtedly to conduct a respiratory current of water to the supra- 
 branchial chambers. 
 
 In figure 57, plate xiv, one side of a water tube, with the adjacent portion of the 
 lamella, taken from a gravid marsupium of Anodonta cataracta, is shown in horizontal 
 section. The gill contains eggs in an early cleavage stage, only four of which, however, 
 are represented in the figure. The septa (s) are seen approaching each other, having 
 not yet quite met. In figure 51, plate xm, taken from the same species but not so 
 highly magnified as the last figure, the septa have fused and the respiratory canals 
 (r. c.) are completely shut off from the marsupial space (m. s.). In both of these figures 
 the sections were taken near the ventral border of the gill; had they been cut at a 
 higher level, the canals would be seen as much larger spaces. As is clearly shown in 
 figures 56 and 57, plate xiv, the ostia open freely into the respiratory canals, and water 
 must therefore enter the latter directly from the mantle chamber. The condition 
 here should be contrasted with that seen in figures 50 and 53, plate xm, which show 
 water tubes from the marsupia of Quadrula and Lampsilis, representatives of Ortmann's 
 Unioninae and Lampsilinse; here the ostia lead directly into the cavity of the tubes 
 (w. t.) which are not subdivided and the whole of which becomes filled with eggs. 
 Although it is not shown in figures 51, plate xm, and 57, plate xiv, the epithelium cov- 
 ering the outer wall of the canals, which is of course the lining of the lamellae, bears 
 cilia which probably aid in conducting the current of water toward the suprabranchial 
 chamber. Below, the canals are closed, and, since they are shut off from the marsupial 
 cavity after the fusion of the septa, but open freely above into the suprabranchial 
 chamber, there is but one course for the water to take — it must pass upward and 
 enter the suprabranchial chamber. The transition from the more or less flattened 
 epithelium lining the outer and inner walls of the respiratory canals to the large colum- 
 nar cells on the anterior and posterior surfaces is clearly seen in figure 57. 
 
 The same condition appears in figure 58, plate xiv, which shows one end of a canal 
 (the end marked X in the preceding figure) and the adjacent tissues, but under a higher 
 magnification. Among the columnar cells are seen numerous swollen mucus cells, 
 
132 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 which are similar to those occurring on the interlamellar junctions farther in. The 
 respiratory canals must be capable of expansion and contraction to a considerable 
 degree, as a rich supply of smooth muscle fibers, passing in both a vertical and a hori- 
 zontal direction, may be seen underlying the epithelium of the canals everywhere except 
 in the septum (fig. 58, pi. xiv). Large blood sinuses (b. s.) are found in the lamellae 
 just outside of the canals, as seen in this figure, which shows how close the blood must 
 come to the water within the canals (r. c). There can be no doubt that the water passing 
 through the canals is a respiratory current. 
 
 Although the respiratory canals open dorsally into the suprabranchial chambers, 
 the marsupial division of the water tubes is completely closed off from the latter, as 
 Ortmann has stated, by a roof which is developed in connection with the septa forming 
 the respiratory canals. The dorsal free border of each interlamellar junction at the 
 level of the suprabranchial chamber expands both anteriorly and posteriorly, but only 
 over the marsupial division of the tube. The anterior and posterior edges of these 
 umbrella-like expansions fuse with each other in exactly the same way as do the septa 
 already described, and, since they also become continuous laterally with the vertical 
 septa which separate the respiratory canals from the marsupial spaces, the latter 
 thereby come to be completely roofed over and do not open at all into the suprabranchial 
 chambers, unless the covering is broken. Of course, the formation of the roofing mem- 
 brane does not take place until after the marsupium is fully charged with eggs. Owing 
 to the gorged condition of the marsupium in these genera, the egg masses cause the 
 roof to bulge up into the suprabranchial chamber over the marsupial division of each 
 water tube, and on exposing the chambers the upper ends of the egg masses, covered, 
 however, by the delicate transparent roofing membrane, are seen protruding beyond 
 the dorsal boundary of the gill. In the drawing of Symphynota complanata (fig. 3, pi. 
 Vi), in which a portion of the suprabranchial chamber is exposed, the condition just 
 described is distinctly shown. 
 
 As Ortmann has described, the secondary division of the water tubes entirely 
 disappears after the discharge of the glochidia. The dorsal expansions of the inter- 
 lamellar junctions, which united to form the roof, give way along the original sutures, 
 and the glochidia are enabled to pass out; the septa separate in a similar manner, and 
 are gradually retracted, and when the marsupium returns to the resting condition no 
 trace of these structures is to be seen. 
 
 We have confirmed Ortmann's discovery of the respiratory canals in Alasmidonla , 
 Anodonta, Strophitus, and Symphynota. Figures representing the marsupial structure in 
 Anodonta cataracta (fig. 51, pi. xm; 57, 58, pi. xiv) have already been referred to. Figure 
 56, plate xiv, is a section taken from near the ventral end of a water tube in the gravid 
 marsupium of Alasmidonta truncata; the young embryos with which the marsupium 
 is filled are not shown. The respiratory canal (r. c.) at this level is quite small and 
 less slit-like than in Anodonta, but it widens out toward its dorsal end. The nuclei of 
 the interlocking cells where the edges of the opposite septa have fused are quite distinct 
 in the section. Figure 52, plate xm, shows a horizontal section from the gravid mar- 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 1 33 
 
 supium of Symphynota complanata at a stage when the glochidia are fully formed. In 
 this species, when the marsupium is fully charged, the interlamellar junctions are so 
 stretched that they become greatly reduced in thickness and appear quite membranous. 
 Figures 49 to 53, plate xm, showing a gravid water tube in Alasmidonta, Ouadrula, 
 Anodonla, Symphynoia, and Lampsilis, respectively, are all drawn under the same 
 magnification, and should be compared in order to observe the relative sizes of the 
 tubes in section in the several cases, as well as the different intervals between the inter- 
 lamellar junctions as shown by the number of intervening filaments in the lamellae. 
 
 Ortmann interprets the respiratory canals of the Anodontinae as an adaptation for 
 the better aeration of the embryos in the marsupium (191 1, p. 325). They are unques- 
 tionably a respiratory device, but for many reasons it would seem clear that they serve 
 primarily for the aeration of the blood of the gravid female and not of the embryos. It 
 is difficult to see how a membrane which shuts the embryos off from the water could 
 increase the facilities for aeration or why such a condition should be an improvement 
 as far as the embryos are concerned, over the marsupium in those genera where there are 
 no respiratory canals and the water comes into direct contact with the embryos. In 
 some of the species of Lampsilis {ligamentina , for example) the marsupium is as heavily 
 charged as in many of the Anodontinae, and the glochidia are also carried over the winter, 
 yet the respiratory canals are not present. In either case the embryos probably receive 
 an adequate amount of oxygen. But, on the other hand, it is not difficult to see that 
 the respiration of the gravid female might be seriously interfered with, when the entire 
 outer gill is gorged and swollen with glochidia and these same glochidia must remain in 
 the marsupium for months. In the Unioninae (Ortmann) the marsupium is gravid for 
 only a few weeks at the longest, and, furthermore, the gills are not so heavily charged, 
 while in the Lampsilince only a differentiated portion of the outer gill receives the embryos 
 and, although the marsupium may be heavily loaded and remain gravid over the winter, 
 the encroachment of the marsupial upon the respiratory function is not so extensive. 
 In these two subfamilies the need of a special respiratory device is, therefore, not as 
 great as in the Anodontinae. The close association of the maternal blood with the cur- 
 rent of water in the respiratory canals, as shown in figure 58, plate xiv, would add further 
 evidence for the view that the secondary division of the water tubes is an adaptation 
 for the better aeration of the blood of the gravid female, in correlation with the prolonged 
 period of gravidity and the interference with respiration by the excessive crowding of 
 the entire outer gill. 
 
 Reference should be made to the special conditions existing in Strophitus. Aside 
 from the formation of the respiratory canals in the manner peculiar to the Anodontinae, 
 Ortmann has briefly described a division of the marsupial cavity of each water tube by 
 the outgrowth of horizontal septa from the interlamellar junctions to form separate 
 closed spaces each one of which incloses a single " placentula " Referring to the peculiar 
 position of the "plaeentulae," which lie crosswise in the gill, he says (1911, p. 294) : 
 
 This arrangement is brought about by further outgrowths of the epithelial layers of the septa (inter- 
 lamellar junctions), which fill the spaces between two septa, or rather only the middle part, the ovisac, 
 
134 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 and thus the simple ovisac of A nodonta and other genera is here divided into a number of swollen, secon- 
 dary ovisacs, running transversely across the gill, each of which contains a short, more or less cylindrical 
 mass of eggs or embryos. * * * Also in Stro philus these structures are not present in sterile females, 
 and after the discharge of the glochidia they soon disappear. 
 
 We have observed this secondary division of the marsupial spaces in Strophitus 
 edentulus. 
 
 We have not studied in detail the histological structure involved in the peculiar 
 differentiation of the ventral border of the marsupium of the Anodontinse and have, 
 therefore, nothing to add to Ortmann's account (i 91 1 , p. 295) of the development of elastic 
 tissue in this region, which allows of the enormous stretching of the gill in these genera 
 when gravid. The lamellae appear to separate along the mid-ventral border, especially 
 in the middle portion of the gill, but are here connected by an elastic membrane which 
 closes the bottom of the water tubes, with the result that "the edge of the marsupium 
 in these forms does not appear sharp as in the Unto group, but blunt, rounded of], or 
 truncated." This distension of the ventral edge, which is much more conspicuous in 
 some genera than in others, is evidently a device to allow of a greater expansion of the 
 marsupium. 
 
 Lampsilince. — It will be recalled that Ortmann includes in this subfamily Simpson's 
 Heterogense, Mesogenae, Ptychogenae, and, although he does not refer to the genus 
 Dromus, he would probably also place the Eschatigenas here. We have already spoken 
 of the general external characteristics which distinguish the marsupia in these groups. 
 A great diversity of form is exhibited by the marsupium, but in all of the genera here 
 concerned certain features, which have been referred to, are possessed in common. 
 
 In all of the groups here considered the marsupium is formed by a varying number 
 of specialized water tubes in the outer gill, which are modified in different ways. In 
 most, the water tubes are utilized throughout their entire length, as in Lampsilis and 
 Obliquaria, but in other genera (Cyprogenia, Ptychobranchus and Dromus for example) 
 it is only the ventral portion of the tubes which retain the embryos. 
 
 The respiratory canals, which are present during gravidity in the Anodontinas, are 
 absent in the Lampsilina?, and the entire cavity of the water tubes in the marsupial region 
 becomes filled with eggs (fig. 53, pi. xm). The marsupium may show a high degree of 
 distension when charged, as is seen in many species of Lampsilis. It is in the Lamp- 
 silinae that we encounter the most capacious marsupial water tubes, the enlargement 
 reaching the maximum size in Obliquaria (fig. 7, pi. vn). In figure 53, plate xm, 
 which is drawn from a gravid marsupium of Lampsilis ligamentina, the characteristic 
 appearance of the water tubes in this genus is shown. The great antero-posterior 
 diameter of the tube (w. t.) is very noticeable, as the interlamellar junctions are repeated 
 at intervals of about a dozen filaments; the relatively large size of the tubes may be 
 readily appreciated by a comparison of this figure with figures 49-52, plate xm. The 
 interlamellar junctions, when the gill is fully charged, are stretched into thin mem- 
 branous septa (i. j.). 
 
 The dorsal free borders of the interlamellar junctions, while not forming a closed 
 roof over the water tubes as they do in the Anodontincc, in Lampsilis at least become 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 1 35 
 
 distended into rather conspicuous bulb-like expansions which greatly diminish the 
 openings of the tubes into the suprabranchial chamber, although their edges do not 
 fuse. 
 
 As the histological details of the structure of the marsupia in several genera belong- 
 ing to the Lampsilinae have been studied by Mr. Carter and will be described in his 
 forthcoming paper, a further account may be omitted here. 
 
 PHYLOGENY OF THE MARSUPIUM. 
 
 It is not without justification that a phylogenetic significance should have been 
 attached to the several types of the marsupium which occur in the Unionidae, for it would 
 seem clear that those forms in which the structure characteristic of the respiratory 
 gill is least modified, as in Quadru/a, are more primitive than those in which the special- 
 ization of the marsupium has gone much farther, as in Anodonta, Lampsilis, and many 
 other genera. 
 
 Simpson (1900) has considered these facts in some detail and concludes that the 
 oldest type of marsupium phylogenetically is that occurring in the Endobranchiae in 
 which the inner gills alone are used as brood chambers. It is a slight transition from 
 this condition to that presented by the Tetragenae with all four gills functioning for this 
 purpose. Basing his supposition largely upon shell characters and geographical distri- 
 bution, he further concludes that the Homogenae marked the next step in marsupial 
 differentiation, while the Heterogenae and all other groups in which a portion only of the 
 outer gills is modified for receiving the eggs are the latest product of the evolution of 
 the Unionidae. 
 
 That this series correctly represents the phylogenetic sequence in the appearance of 
 the marsupial modifications would seem to be borne out by the structural conditions 
 existing in the several types so far as we have examined them, provided that we assume, 
 with respect to the Homogenae, that genera like Plcwobcma and Unio, in which the mar- 
 supium is less specialized, are more primitive and therefore stand nearer the Tetragena 
 than such genera as Anodonta, Symphynoia , and others, which, as Ortmann has shown, 
 exhibit certain modifications evidently in advance over the marsupium of the former. 
 
 Ortmann (191 1), although he does not consider the Endobranchiae, has arrived at con- 
 clusions essentially similar to the above. He points out, however, that the absence of 
 complete interlamellar junctions in the gills of Margaritana would indicate that the new 
 family which he has created for this genus, Margaritanidae, is the most primitive group 
 of the Naiades, and this inference, as was indicated above, is further strengthened by the 
 fact that the simple gill structure of Margaritana is apparently similar to that of Mytilus, 
 which belongs to a lower group of lamellibranchs than the fresh-water mussels. 
 
 His conclusions concerning the sequence of his three subfamilies of the Unionidae may 
 be quoted (p. 328) : 
 
 Of the Lhn'onidie, the Unionince are certainly more primitive than the other two subfamilies, as is 
 evidenced by the simple character of the structure of the marsupial gills. The A nodontincc and Lanip- 
 18713°— 12 3 
 
136 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 siliiueaie more advanced, but they have advanced in different directions, and each has developed special 
 features of the sexual apparatus. Generally speaking, the Lam psilince contain the most highly advanced 
 types, as is shown, by the restriction of the marsupium to a part of the outer gill, and by the strong 
 expression of the sexual differentiation in the outer shell. Yet there are forms among the Anodontince 
 which show extremely complex structures (Stro phitus ) unparalleled in an)' other genus, and the peculiar 
 glochidiaof the Anodontince surely mark a high stage of development. 
 
 It is not necessary for our purpose to enter into a further discussion of the subject in 
 this place. 
 
 CONGLUTINATION OF THE EMBRYOS. 
 
 After extrusion of the eggs from the genital apertures, they are received intothesupra- 
 branchial chambers, and thence pass, as has already been described, into the water tubes 
 of the gills, eventually filling up those portions which function as the marsupium. In 
 a short time after entering the latter the eggs usually become conglutinated into masses 
 which are molded into the exact shape of the cavity of each marsupial water tube (Lefevre 
 and Curtis, 1910b). The masses are of course separated from each other by the inter- 
 vening interlamellar junctions of the gills. 
 
 Since it is a matter of convenience to have a word to apply to these compact masses 
 in which the eggs or embryos are held together, whether they be plate-like, club-shaped, 
 cylindrical, or of some other form, we shall employ the term conglutinate in referring 
 to them. Ortmann (191 1) has proposed the word placenta, which was introduced by 
 Sterki (1898) for the peculiar cords of Strophitus, but this is obviously misleading, as 
 there is no connection whatever between the masses and the maternal tissues. The 
 conglutinates vary greatly in different species in size and shape, and, since each is a cast 
 of the cavity of its water tube, they conform to the special conditions existing in the 
 several tvpes of marsupium. The commonest form is that of a flat plate, either elliptical 
 or lanceolate, being usually slightly blunter and thicker above and more pointed and 
 thinner below. Since we have already seen that the antero-posterior diameter of the 
 marsupial water tubes varies very much in different species, the thickness of the con- 
 glutinates must vary to the same extent. In Quadrula and L'nio, for example, in which 
 the interlamellar junctions are set close together, the conglutinates are very thin, being 
 not more than twice the diameter of an egg in thickness; whereas in Lampsilis, with its 
 much more capacious tubes, they may be three or four times as thick. In other words, 
 just as many eggs will lie abreast in a horizontal section of the marsupium as the antero- 
 posterior diameter of the water tube will allow. 
 
 This commoner lanceolate form of the conglutinate, differing, however, in size and 
 thickness, may be seen in the species of Quadrula, Plcurobcma, L'nio, and Lampsilis. In 
 figure 41, plate XI, two conglutinates of Lampsilis ligamentina are represented, one from 
 the flat side, the other on edge. An unusual form of conglutinate has been observed by us 
 in Quadrula metanevra; it is bifurcated and consists of two flat lanceolate masses which 
 are united for the upper third of their length, but free below. In those genera, however, 
 in which the form of the water tubes of the marsupium departs more widely from the 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 1 37 
 
 usual condition, the conglutinates are similarly modified. In Obliquaria reflexa, for 
 example, in which the marsupium consists of several elongated and distended water tubes 
 of tubular form, the conglutinates are large, slightly curved cylindrical masses of nearly 
 uniform diameter and generally blunt at each end. Three of them are shown in figure 
 42, plate xi; the one on the right was taken from the most posterior water tube of the 
 marsupium, which is not as long as the rest, and its conglutinate is correspondingly 
 shorter. The relation will be understood by reference to the figure of the marsupium 
 of this species (fig. 7, pi. vn). 
 
 There seem to be two methods by which the embryos are bound together to form 
 conglutinates — they may either be attached more or less firmly to each other by their egg 
 membranes, which are in this case of an adhesive nature, or they may be embedded in a 
 mucilaginous matrix of varying consistency. The former is by far the commoner condi- 
 tion and is seen in figure 17, plate viu, which is a detail drawn from one of the congluti- 
 nates of Obliquaria reflexa shown in figure 42, plate xi ; the immature glochidia with their 
 valves open are still contained within the membranes, which are closely adhering and by 
 mutual pressure are squeezed into a polyhedral form. In cases like this it is difficult to 
 determine whether there is a glutinous matrix between the embryos or not, but if any is 
 present, it must be in very small amount, since the embryos seem to be held together 
 solely by the adhesive surfaces of their membranes. In those cases, however, in which a 
 matrix is evident (Lampsilis), the embryos are not so closely appressed and are embedded, 
 more or less loosely, in a glutinous binding substance. This condition is illustrated in 
 figure 16, plate viu, which is a portion of a conglutinate of Lampsilis ligamentina seen 
 under higher magnification; as the matrix is transparent, it can not be shown in the 
 figure. 
 
 The conglutinates differ markedly in tenacity, for, whereas in some cases the mutual 
 adhesion is not strong and the masses consequently break up readily (Quadrula, Plcu- 
 robema, Unio, Lampsilis), in others (notably in Obliquaria) the embryos adhere so firmly 
 that they may be separated only with difficulty by teasing. 
 
 In still other species the embryos can not be said to form conglutinates at all, as 
 they are merely suspended in a slimy mucus which is not of such a consistency as to 
 enable the mass to maintain a definite form when removed from the gill. We have 
 observed this condition in Alasmidonla, Anodonta, and Symphynota, and Ortmann (191 1) 
 states that it also occurs in Anodontoides. 
 
 In most species (Quadrula, Unio, Lampsilis, Dromus) in which the conglutinates 
 are found, the adhesion exists only during the embryonic development and by the 
 time the glochidia are fully formed they are found to be free but for the mucus which 
 holds them more or less loosely together. In Obliquaria reflexa, however, the congluti- 
 nation persists, and the fully developed glochidia, still tenaciously adhering, are dis- 
 charged from the marsupium in the cylindrical masses already described (fig. 42, pi. xi) ; 
 even after lying in the water for some time they do not separate, and it has perplexed 
 us to understand how the glochidia of this species ultimately become attached to fish, 
 if they pass through a subsequent parasitic stage. Can it be that parasitism has been 
 
138 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 lost in Obliquaria as it has been in Strophitus, and that the metamorphosis takes place 
 while the glochidia are in the conglutinates? We have not yet had the material by 
 which to answer this question. 
 
 The relation of the embryos and glochidia of Strophitus to each other is so unusual 
 that its description is reserved for a special section (see below). 
 
 STRATIFICATION OF UNFERTILIZED EGGS. 
 
 It has alreaay been pointed out that not infrequently eggs pass into the marsupium 
 without being fertilized and remain there throughout the period of embryonic develop- 
 ment, as one may find them in the same gill with fully formed glochidia. In some 
 individuals we have found every egg in the marsupium in this condition. Such eggs 
 have been encountered chiefly in summer-breeding species, and they seem to be especially 
 common in Pleurobema and Quadrula, nearly every gravid female of which has been 
 found to contain at least some unfertilized eggs. After remaining in the marsupium 
 for a time such eggs generally become swollen and stratified into three distinct layers, 
 a heavier, often pigmented, mass at one pole, a clear or hyaline intermediate zone, and a 
 small granular cap at the lighter pole. As the eggs lie in a constant position in the 
 gills, which are placed vertically in the normal position of the animal, it can not be 
 doubted that the stratification is produced by gravity. It has not yet been determined 
 whether the substances which occur in these layers are the same as would be separated 
 out by centrifuging or not, but this is not at all unlikely. As many of the species of 
 mussels in which we have seen this condition, for example, Quadrula ebena, Q. irigona, 
 and Pleurobema asopus, have brightly colored red or pink eggs, the stratification is 
 quite striking, the pigment being always at the heavier pole, as it is invariably directed 
 toward the lower border of the gill. 
 
 ABORTION OF EMBRYOS AND GLOCHIDIA. 
 
 There has been a certain amount of discussion among the conchologists as to 
 whether or not the functioning of all four gills as a marsupium is a constant character 
 in Quadrula, and observations have been to a certain extent conflicting. Since Simpson 
 has made use of this feature in characterizing the group Tetragena?, some importance 
 has been attached to the apparent discrepancy in observations. 
 
 While examining mussels on the upper Mississippi River in the summer of 1908, 
 we observed a peculiarity of behavior in all of the species of Quadrula collected which 
 may account for the conflicting descriptions of the marsupium in this genus, and also 
 for the fact that in some species gravid females have never been observed at all. Every 
 species of Quadrula that came into our hands exhibited to a greater or less degree the 
 habit of aborting embr)'os and glochidia when taken out of the river, and if they were 
 not opened and examined at once upon capture they were generally found shortly 
 afterwards to be either partially or entirely empty. Some individuals discharged the 
 contents of their gills more readily and completely than others, the abortion involving 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 1 39 
 
 either all four gills or only the inner or outer ones, or, again, only a portion merely of 
 one or more gills. In the pre-glochidial stages, when the embryos are conglutinated, 
 the entire masses were discharged, while individuals were frequently seen in the act of 
 aborting their embryos or glochidia which were often expelled with considerable force 
 through the exhalent siphon. 
 
 This behavior was so characteristic of the genus that, in order to make a correct 
 determination of the condition of the marsupium, it was necessary to open quadrulas 
 immediately after taking them from the water. When this was done, all four gills 
 were invariably found to be charged on opening females which contained embryos in 
 pre-glochidial stages— that is, at any time before normal spawning had occurred. The 
 habit of readily aborting embryos when disturbed has also been observed by us in 
 Unio complanatus, which has been repeatedly seen in the act of discharging the contents 
 of the marsupium shortly after being placed in aquaria. In all likelihood it occurs in 
 other species of Unio, and it may possibly be characteristic of all forms in which there 
 is but little structural differentiation of the marsupium. We have, however, also 
 observed the discharge of embryos in Lampsilis ligamentiiia, but only after the gravid 
 females have been kept in the laboratory for some time. This species is apparently 
 very much less sensitive with respect to abortion than the quadrulas and Unio com- 
 planatus and only frees its gills of the conglutinates after long exposure to artificial 
 conditions. The premature extrusion is probably due to imperfect aeration of the water 
 and results from an effort on the part of the female to secure more oxygen; if this be 
 true, one would not expect to find it occurring so readily in those forms which have a 
 differentiated marsupium, like the Heterogense, since here the respiratory and marsupial 
 functions of the gills are not so intimately associated. 
 
 Both Schierholz (1S88) and Latter (1891) have referred to the occurrence of 
 abortion in Anodonta, but according to our experience it has never been encountered 
 in a single instance in either Anodonta or Symphynota, although gravid females have 
 been kept in tanks in the laboratory for weeks or even months. The presence of the 
 respiratory canals, which have been described as occurring in these genera during 
 gravidity, as well as the temporary membrane which roofs over the marsupial division 
 of the water tubes, might well account for the absence of abortion, or at least its rare 
 occurrence, in the forms in which these special conditions exist. The respiratory canals 
 doubtless lessen the evil effects of poor aeration, while the roofing membrane of the 
 water tubes would certainly offer some obstruction, as long as it was present, to a 
 liberation of the embryos. 
 
 BREEDING SEASONS. 
 
 In connection with our study of artificial propagation of fresh-water mussels, we 
 have found it necessary to collect data bearing upon the breeding seasons of a fairly 
 wide range of species, since the records of previous observers, for North American 
 Unionidae at least, have been insufficient to enable us to determine the full extent of 
 the seasons, especially in the case of some of the more important commercial species. 
 
I40 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 Although our observations have been largely confined to species occurring in the upper 
 Mississippi Valley and have been concerned primarily with species of commercial value, 
 we have continuous records throughout the entire year for a number of important 
 genera, and in every case the exact stage of development of the embryos has been 
 determined by microscopic examination. Many thousands of such observations have 
 been made, so that we are now in possession of detailed information dealing with the 
 duration and progress of the periods of gravidity obtaining in over a dozen genera of 
 the Unionidse. 
 
 We have fully confirmed the conclusion reached by Sterki (1895) that the North 
 American Unionidaj, with respect to their breeding seasons, fall into two classes, the 
 so-called "summer breeders" and "winter breeders" — a distinction, however, which 
 had previously been pointed out by Schierholz (1888) for European forms and fre- 
 quently recorded by later observers. The designation "winter breeders," however, is 
 not strictly appropriate, for in the species which belong to this group the eggs are 
 fertilized during the latter half of the summer and the glochidia, which are carried in 
 a fully developed condition in the marsupium throughout the winter, are not discharged 
 until the following spring and summer. In the case of the summer breeders, the eggs 
 are fertilized during late spring and summer and spawning as a rule is over by the end 
 of August. 
 
 In view of these facts, it would seem to accord better with the actual conditions 
 to separate the species with respect to the length of time that the glochidia remain in 
 the marsupium, designating them as those that have a "short period" and those with 
 a "long period" of gravidity, rather than to distinguish them as "summer breeders" 
 and "winter breeders," respectively, for with respect to the latter neither ovulation 
 nor discharge of the glochidia takes place in winter. This suggestion was made by us 
 in an earlier paper (1910b), and subsequently Ortmann (191 1) proposed the somewhat 
 awkward terms tachyttclic and bradyticlic (meaning quick-breeding and slow-breeding) 
 for Sterki's "summer breeders" and "winter breeders," respectively. 
 
 The breeding seasons as here defined are based upon data collected in the middle 
 and northern sections of the United States, and in the absence of adequate records from 
 higher and lower latitudes, it is impossible to say to what extent a colder or warmer 
 climate might affect the period of gravidity That it would have some influence can 
 hardly be doubted, although a distinction between a long and a short season will prob- 
 ably be found to hold true in general. 
 
 The breeding season is a generic character, for so far as our observations have gone 
 all of the species belonging to a given genus have essentially the same period of gravidity. 
 The prolonged period, furthermore, is correlated with the more pronounced structural 
 modifications of the marsupium which have been described above. 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 141 
 LONG PERIOD OF GRAVIDITY. 
 
 In the forms which fall into this category the eggs are fertilized, as has been stated, 
 during the latter half of the summer, from the middle of July to the middle of August, 
 and the glochidia, instead of being discharged when fully formed, are carried in the 
 marsupium until the following spring or early summer. In fact, in some cases the 
 close of one breeding period may overlap on the beginning of the next, as one may still 
 find in late July a few straggling females gravid with glochidia formed in the previous 
 autumn, while in other individuals of the species at the same time and in the same 
 locality the eggs are passing into the gills for the next season. This seems to be true 
 of several species of Lampsilis. We have encountered it in ligamentina, Conner (1909) 
 records it for radiata and nasuta, while Ortmann (1909) states that his observations 
 make it probable for venlricosa and luteola. Yet, as Ortmann observes, it is generally 
 true that an interval exists between the close of one period and the beginning of the 
 next. This interval, however, varies in length in different species, in some extending 
 from late spring until August, whereas in others it is of much shorter duration. It is 
 also to be noted that the discharge of glochidia does not take place in all of the indi- 
 viduals of a species at the same time, but on the contrary, spawning may extend over 
 a considerable period throughout the spring and early summer (cf. Ortmann, op. cit.). 
 
 All of the genera included in Simpson's Heterogena?, Ptychogenae, Eschatigenae, 
 and Diagense have the long period of gravidity, as do also a number of genera of the 
 Homogenae (Alasmidonia, Anodonta, Anodontoides, Arcidens, Symphynota) , while the 
 Mesogenae are represented in this group by Cyprogenia. These genera are embraced in 
 Ortmann's subfamilies Anodonlince and Lampsilince, and it should be noticed that in 
 all the gills show a high degree of specialization in adaptation to the marsupial func- 
 tion, a specialization which is undoubtedly correlated with the habit of retaining the 
 glochidia over a period of several months. 
 
 In the following list are given the species in which we have determined the long 
 period of gravidity : 
 
 Alasmidonta truncata. Lampsilis ligamentina. 
 
 Anodonta cataracta. Lampsilis luteola. 
 
 Anodonta grandis. Lampsilis recta. 
 
 Anodonta implicata. Lampsilis subrostrata. 
 
 Arcidens confragosus. Lampsilis ventricosa. 
 
 Cyprogenia irrorata. Obovaria ellipsis. 
 
 Dromus dromus. Plagiola elegans. 
 
 Lampsilis (Proptera) alata. Plagiola securis. 
 
 Lampsilis (Proptera) laevissima. Strophitus edentulus. 
 
 Lampsilis anodontoides. Symphynota complanata. 
 
 Lampsilis gracilis. Symphynota costata. 
 Lampsilis higginsii. 
 
 Ortmann (1909) has published some observations on the breeding seasons of the 
 Unionidas of Pennsylvania, supplemented by data from Lea and Sterki; his results in 
 
I42 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 all essential points agree closely with ours. He includes among "winter breeders" 
 several genera which we have not had under observation, namely, Truncilla, Micromya, 
 Ptychobranchus, and Anodontoidcs, while Arcidens, which we have recorded, does not 
 appear in his list. 
 
 There is given below a brief summary of our breeding records for the genera here 
 concerned. Although in many species we have examined hundreds of individuals and 
 have had them under observation continuously throughout the year, in others the 
 material has been more or less meager and observations scattered, but in most of the 
 forms the records have been adequate for a determination of the general limits of the 
 breeding season. 
 
 Alasmidonta. — Embryos from latter part of July to middle of August. No fully 
 formed glochidia have been seen, as gravid females have not been secured after August. 
 
 Anodonia. — Embryos from the middle of August to September; ripe glochidia from 
 earlv October to first of July. A distinct interim exists between close of one period 
 and beginning of next. According to Harms (1909), in European species of Anodonta 
 the eggs are fertilized about the middle of August, all of the individuals entering upon 
 the breeding season at nearly the same time, and by the middle of October almost all 
 of the females are gravid with glochidia. 
 
 Arcidens. — Glochidia in winter months. Only a few individuals secured. 
 
 Cyprogenia. — Glochidia in November. 
 
 Dromus. — Glochidia in November. 
 
 Lampsilis. — Embryos from first of August to late September; glochidia from late 
 September to first of August. Our most complete record concerns this genus, several 
 species of which {anodontoidcs, ligamentina, recta, subrostrata, vcntricosa) we have 
 repeatedly had under observation continuously throughout the year. The gravid 
 period seems to be more extended in Lampsilis than in any other genus, for, although 
 June is apparently the month when the liberation of glochidia is at its height, some 
 females bearing glochidia may still be found, but in diminishing numbers of course, until 
 the beginning of August, a time when the next season is just setting in. Since ripe 
 glochidia may be obtained in abundance from October to July, inclusive, and since 
 Lampsilis furnishes several species of commercial value, the extended period of gravidity 
 in this genus becomes of the greatest importance in artificial propagation, as material 
 is available for the infection of fish throughout the greater part of the year. 
 
 Obovarta. — Glochidia during the fall, winter, and spring months. Spawning must 
 occur before June, as no glochidia have been encountered in June, July, and August, 
 although a number of females have been obtained during these months. 
 
 Plagiola. — Ripe glochidia during the winter and as late as the end of July; no 
 embryos- have been obtained. 
 
 Strophitus. — Embryos from late July to middle of August; glochidia from November 
 to middle of July. The interval between the seasons is very short, much shorter than 
 that observed by Ortmann (1909), who records an interim from May 22 to July 11. 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 1 43 
 
 Symphynota. — Embryos during August; ripe glochidia from late September to late 
 June. 5. complanata is a species which we have had on hand constantly for several 
 years, and we have followed it continuously through the year. Spawning is most active 
 in June. 
 
 SHORT PERIOD OF GRAVIDITY. 
 
 In the species having the short period of gravidity the entire breeding season is 
 confined to about four months, as it extends only from about the end of April to the 
 middle of August, and the glochidia are discharged as soon as they are fully developed. 
 It is highly probable, however, that the beginning of the breeding season is influenced 
 to a certain extent by temperature, for it would seem that ovulation may be postponed 
 for some weeks by cold weather at this time of the year. It was first pointed out by 
 Sterki (1895) that these summer-breeding forms are confined to a limited group of 
 genera, and Ortmann (191 1) has emphasized the fact that it is only the genera having the 
 least specialized marsupia that possess this apparently more primitive breeding season ; 
 these are the genera which constitute his subfamily Unioninse. Margaritana, unques- 
 tionably a primitive form, likewise breeds only in the summer. In all of these genera 
 the structure of the marsupium approaches most closely that of the respiratory gills; 
 none of the special modifications, so prominent a feature of the marsupium of other 
 genera, is present. There is apparently, however, one exception, for, as will be shown 
 below, our records indicate clearly that Obliquaria, which has a highly specialized marsu- 
 pium, is a summer breeder. 
 
 The following are the species which we have observed to have the restricted breeding 
 season : 
 
 Obliquaria reflexa. Quadrula plicata. 
 
 Pleurobema aesopus. Quadrula pustulosa. 
 
 Quadrula ebena. Quadrula trigona. 
 
 Quadrula heros. Quadrula (Tritogcmia) tuberculata. 
 
 Quadrula lachrymosa. Quadrula undulata. 
 
 Quadrula metanevra. Unio complanatus. 
 
 Quadrula obliqua. Unio gibbosus. 
 
 The following species, which do not appear above, have been determined by Ort- 
 mann (1909) to be summer breeders: Unio crassidens; Pleurobema clava and coccinea; 
 Quadrula kirtlandiana, rubiginosa, and subro'tunda. Our list, on the other hand, sup- 
 plements his by the addition of several species of Quadnda, for which data have pre- 
 viously been either entirely wanting or quite meager. 
 
 Obliquaria. — Since all of the forms which carry the glochidia over the winter have a 
 highly specialized marsupium, we should expect that Obliquaria, whose marsupium is 
 of such a nature, would also have the long gravid period. This expectation would be 
 further strengthened by the fact that the very closely related genus Cy progenia belongs 
 in the former group, as has been seen. It was therefore with some surprise that we 
 found O. reflexa breeding during the summer. Our record is as follows : Embryos from 
 
144 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 the latter part of May to July 9; glochidia from June 20 to August 8. This is a typical 
 record for a summer breeder, and there can be little doubt that the species must be placed 
 in this group. On the other hand, Sterki (1898, 1903) states that all forms which have a 
 differentiated marsupium carry their glochidia over the winter, and Ortmann (191 1) 
 includes Obliquaria in his Lampsilince, all of which he says are "bradytictic," although 
 specific reference to the breeding season of this genus is not made. Since, however, 
 we have not had an opportunity of observing the species during the fall and winter, 
 it is possible that it has the long period, although, if such is the case, its season begins 
 two months earlier than that of any other species in this class — a quite improbable 
 supposition. For the present, at all events, we must consider it a summer breeder. 
 
 Plcurobema. — Embryos from early June to early August; glochidia during July. 
 
 Quadrula. — Embryos from late May to middle of August ; glochidia from early June 
 to middle of August. Hundreds of females belonging to different species of this genus 
 have been examined throughout the rest of the year, but gravid individuals have never 
 been encountered except during the months indicated. 
 
 It should be mentioned that in the case of Q. heros Frierson (1904) has not found 
 this species gravid in Louisiana until October, when embryos were found. Young 
 embryos were again encountered in November and immature glochidia in January. He 
 concludes that hcros is an exception in the genus and is not a summer breeder. Our 
 observations on this species are very meager, but since we have found it bearing young 
 embryos in the latter part of May, they would seem not to be in accord with those of 
 Frierson. 
 
 According to Harms (1909), Margaritana, which breeds in Europe in July and 
 August, produces two successive broods during that time, from sixteen days to four 
 weeks, according to temperature, being required for the development of each. Although 
 we have not determined it beyond all doubt, our records strongly indicate that the species 
 of Quadrula also spawn twice during the season, first in June and July and again in July 
 and August. This, however, could not be definitely proven without a most extended 
 series of observations, and possibly not unless individual females were kept in aquaria 
 under close observation throughout the breeding season. 
 
 Unio. — Embryos from early June to early August; glochidia from middle of June 
 to middle of August. Conner (1907) records U. complanatus as beginning its breeding 
 season in April, and Lea (1863) found it gravid in May; but we have not had an oppor- 
 tunity of examining any species of the genus during these months. According to Harms 
 (1909) the breeding season of Unio in Europe begins early in March, or, if the weather 
 is cold, not until the end of May. 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH -WATER MUSSELS. 1 45 
 
 III. THE LARVA. 
 STRUCTURE OF THE GLOCHIDIUM. 
 
 As has long been known, two well-marked types of glochidia are found in the 
 Unionidse; one provided with a strong shell bearing a single stout hook at the ventral 
 margin oi each triangular valve; the other with no such hooks and a more delicate shell, 
 the valves of which are shaped like the bowl of a very blunt spoon. 
 
 A possible third type, which appears to be a derivative of the second, is seen in the 
 "axe-head" glochidium, originally described and figured by Lea (1858, 1863, and 1874) 
 in Lampsilis {Proptera) alata, Icevissima, and purpurata. 
 
 The first type is characteristically parasitic upon the fins and other external parts of 
 fishes from which scales are absent, the second upon the gill filaments. The occurrence 
 of these types in the genera which we have examined is shown by the following list: 
 
 Hooked glochidia: Hookless glochidia: Axe-head glochidium : 
 
 Anodonta. Cyprogenia. Lampsilis (Proptera) alata. 
 
 Strophitus. Dromus. Lampsilis (Proptera) laevis- 
 
 Symphynota. Lampsilis (majority of sima. 
 
 species). Lampsilis (Proptera) pur- 
 
 Obliquaria. purata. 
 
 Obovaria. Lampsilis capax. 
 
 Plagiola. 
 
 Pleurobema. 
 
 Quadrula. 
 
 Tritogonia. 
 
 Unio. 
 
 The axe-head glochidium occurs, so far as known, in only a few closely related 
 species which were generally included in the genus Lampsilis, but which, after being 
 first placed in the subgenus Proptera by Simpson (1900), have been elevated to the 
 genus Proptera by Sterki (1895 ar, d 1903), a change which has recently been approved 
 by Ortmann (191 1). The species long known to possess this axe-head glochidium are 
 Lampsilis {Proptera) alata, Icevissima, and purpurata, and recently Coker and Surber 
 (191 1 ) have described it for Lampsilis capax. 
 
 There is considerable diversity in size among glochidia even from the same genus, as 
 represented by the outlines in text figure 1 (A-o).all of which are drawn to the same scale, 
 the most striking cases being the difference between the two species of Plagiola (g and h), 
 and that between Lampsilis recta and gracilis (k and l). Harms (1909), who has studied 
 the exceedingly minute glochidia of Margaritana margaritijera, finds that they are 
 exclusively gill parasites, because their small size makes attachment elsewhere impossible. 
 
 The type of glochidium is constant for the genus, so far as our observations go, 
 save in the case of Lampsilis, as has just been mentioned. In some cases the shape is 
 also characteristic, as shown by Symphynota and Anodonta (a, b, and c), in which the 
 shell outline is a distinguishing feature. 
 
 In Dromus dromus the glochidium, which is of the hookless type (text fig. 1, m), is 
 greatly elongated antero-posteriorly thus presenting an interesting modification. 
 
146 
 
 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 THE HOOKLESS TYPE. 
 
 Since the greater part of our experimental infections with glochidia of the hookless 
 type have been made with our common species of Lampsilis, we have examined the 
 glochidia in this genus more extensively than any others and shall describe, as repre- 
 sentative of what has been observed, the hookless glochidium of Lampsilis subrostrata 
 which is shown in figures 13, 14, and 15, plate vin; and, since it is often necessary in 
 
 the practical work of in- 
 fection to examine the 
 glochidia alive in water 
 and to determine the 
 exact stage of their de- 
 velopment, we shall first 
 speak of their appear- 
 ance when in this con- 
 dition. 
 
 When examined 
 alive (fig. 13, pi. vni), 
 this glochidium exhibits 
 a shell which is compar- 
 atively firm in structure 
 and which may remain 
 unchanged by the water 
 even many days after its 
 living contents have been 
 destroyed. Evidence of 
 the shell's strength is 
 shown by the fact that 
 its shape remains un- 
 changed after the glo- 
 chidial muscle has caused 
 the lips of the shell to 
 bite deeply into a host's 
 tissue, and by the fact 
 that it is not easily broken 
 0.22 x0.19 mm. ky j-Qjjgh handling, as 
 
 when the glochidia are tumbled in and out of a pipette during the process of breaking 
 up the conglutinated masses. This strength is due to the carbonate of lime already 
 laid down in the shell and not to the cuticle, which is often referred to by investiga- 
 tors as though it were the sole constituent of the shell of the glochidium; for when the 
 carbonate of lime is dissolved by acid the cuticle becomes wrinkled and the shell parti- 
 ally collapsed. Viewed from the outside and closed (fig. 13, pi. vm), this shell of the 
 living glochidium exhibits a fine granulation over its entire surface and a distinct border 
 
 FlG. i. — Figures showing relative sizes and shapes of the shells of a series of glochidia 
 belonging to the following species: A, Sympkynota complanala, 0.30 X 0.29 mm.; B 
 S. coslata, 0.39 X 0.35 mm.; C. Anodcmta cataracta, 0.36 X 0.37 mm.; D, Lampsilii 
 (Proplcra) alata, 0.41 X 0.23 mm.; E, Quadrula mttanczra, 0.19 X 0.18 mm.; F. Q 
 pustulosa, 0.30 X 0.23 mm.; G, Plagiola elegans, 0.09 X 0.07s mm.; H. P. securis, 0.31 
 X 0.23 mm.; I. Quadrula ebcna, 0.15 X 0.14 mm.; J, Q. plicala, 0.21 X 0.20 mm.; K 
 Lampsilis gracilis, 0.085 X 0.075 mm.; L, L. recta, 0.24 X 0.20 mm.; M, DromUi 
 dromus, 0.19 X 0.10 mm.; N, Obliquaria reflcxa, 0.23 X 0.225 mm.; O. Vnio gibbosus 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OP FRESH-WATER MUSSELS. 147 
 
 around the free margin. At the hinge margin two denser areas may be observed, which, 
 when examined from the inner face of the valve, are found to be continuous with the 
 border around the free margin (fig. 13, pi. viii). The testwith acid shows that this entire 
 border is calciferous and that there is a thinner layer of carbonate of lime over the whole 
 surface and beneath the cuticle. This layer is often cracked, as one might break the 
 shell of a hen's egg, when preserved specimens are slightly crushed under a cover glass, and 
 it is then seen to be distinct from the cuticle which may wrinkle but does not break. Upon 
 the loss of the lime, the cuticle is no longer firm enough to preserve theshapeof the shell and 
 successful permanent mounts must therefore avoid acids at any stage of the preparation. 
 
 Along the ventral border of the shell is a flange, formed of cuticle only, and so 
 transparent that it is easily overlooked in a ventral view of the open glochidium (fig. 15, 
 pi. viii). Viewed laterally (fig. 14, pi. viii), this flange has at a certain focus the appear- 
 ance of a hook and may easily be mistaken for one when seen under a low magnifica- 
 tion. It is, however, a continuous flange, as shown in the figures, and not a hook; and 
 since its edge is very fine it must, when the glochidium closes its valves, cut into and 
 hold to a delicate tissue like that of the gill filament, thus performing much the same 
 function as the hook in the other type of glochidium. The general spoon-like character 
 of the valves is shown clearly by the figures. The adductor muscle is well seen in the 
 living specimen, being a conspicuous object from whatever angle it is examined. Viewed 
 laterally (fig. 13, pi. viii), or from the ventral aspect (fig. 15, pi. viii), the adductor is 
 seen to lie nearer the shell margin at one end of the hinge than at the other, a fact which 
 enables one to recognize at a glance the future anterior border of the shell. There is 
 also in this glochidium of Lampsilis subrostraia a slight difference in outline by which 
 these anterior and posterior borders of the shell may be distinguished (fig. 13, pi. viii), 
 while in the hooked type of glochidium (fig. 10, pi. viii, and text fig. 1, a, b, and c) 
 this difference is even more pronounced and one recognizes the anterior border of the 
 future adult by its slightly greater length. 
 
 The two outer pairs of sensory cells with their fine projections (fig. 14 and 15, 
 pi. viii) are readily seen in the living glochidium; the two inner pairs, in which the cells 
 project but a short distance from the surface, are more easily found in specimens which 
 have been properly preserved and stained. The position of the two outer pairs may 
 also be seen in the closed glochidium (fig. 13, pi. viii). Little can be seen of the rudi- 
 ments of the various organs of the adult without the careful staining of well fixed 
 material. In the living glochidium they appear as a slightly denser area on either side 
 of the median line and posteriorly to the adductor muscle (fig. 13, pi. viii). The cells 
 of the larval mantle (fig. 15, pi. viii), which occupy the greater part of the surface 
 exposed within the valves, appear in the living glochidium as a dense mass in which 
 cell outlines can not be recognized. 
 
 Further details in the structure of this glochidium can only be studied in specimens 
 which have been properly fixed and stained. After trying various reagents, we have 
 found that they may be stupefied in a few moments by the addition of several small 
 crystals of hydrochlorate of cocaine to the water in a watch glass, after which they 
 
148 BULLETIN OK THE BUREAU OF FISHERIES. 
 
 may be fixed with no serious shrinkage by using the solution of plain corrosive sublimate 
 obtained by diluting a saturated solution two or three times with water. Acids should 
 be avoided throughout the whole process. Alum cochineal, Delafield's hematoxylin, 
 and borax carmine, alone or with Lyon's blue, have been used as stains, each being 
 more suitable for the demonstration of certain structures. In this stained material the 
 shell shows a slight wrinkling of its ventral flange and is the only part not shown to 
 better advantage than in the living specimens. 
 
 The lateral pairs of sensory cells (fig. 14 and 15, pi. vm) are tall chimney-like 
 structures expanded at the base and terminated by several very fine motionless proc- 
 esses. A denser border where these processes are inserted in the cell is presumably due 
 to their continuation within the cytoplasm which has been observed in sections of these 
 and other glochidia. The two median pairs of sensory cells (fig. 14, pi. vm) project 
 only a short distance and have short processes. The anterior pair is located ventral 
 to the median portion of the larval adductor muscle, the posterior pair near the outer 
 ends of the rudiments of the adult organs (fig. 15, pi. vm). The designation of these 
 cells as "sensory" by all writers rests upon their structural features as described by the 
 earlier investigators, and upon the fact, recorded by Lillie (1895), of their staining 
 reaction with methylene blue. The actual connection of the cells with the larval muscle 
 fibers has been sought for by investigators, but never discovered. We have not 
 attempted a further demonstration of the function of these cells by the methods prac- 
 ticed in recent experimentation upon the protozoa and other minute organisms, although 
 such a study might yield some interesting results. 
 
 Lining the greater part of the surface between the valves, are the large cells com- 
 posing the larval mantle (fig. 15, pi. vm). They are filled with fine granules, which, 
 since these cells actually digest the tissue of the host during the early stages of the 
 parasitism, are probably the zymogen granules from which the digestive enzymes origi- 
 nate. The absence of these cells over the area of flexure ventral to the adductor muscle 
 will be noted in figure 15. In this area the ectoderm is thinner and there is no granu- 
 lation. The adductor muscle is composed of fibers having elongated nuclei and often 
 seen to branch toward the ends where they are attached to the valves. In a glochidium 
 of Lampsilis subrostrata, which has been carried over the winter in the parent gills and 
 which has therefore reached the highest stage of differentiation possible for this glochi- 
 dium, we can identify the rudiments of foot, stomodaum and enteron, and of the heart, 
 pericardium, and kidney, as described by Harms (1909) in his accounts of the structure 
 and organogeny in the hookless type of glochidium. Reference to figure 15, plate vm, 
 will make clear the following account of these rudiments. 
 
 In the median region, just posterior to the adductor, is a triangular area, the oral 
 plate; behind this a narrow band of closely set nuclei extending well out into the valves, 
 where it becomes wider. The ectoderm in the median part of this area becomes the 
 covering of the foot, while the deeper part of the area is endoderm, the rudiment of the 
 enteron. The lateral expansions of this general mass are mesodermal cells which are 
 closely applied to the endoderm and in which are found the rudiments of the kidney, 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 1 49 
 
 heart, and pericardium. A backward curve in the posterior outline at either side of 
 this mass appears to represent imperfectly developed lateral pits, from the outer borders 
 of which Schierholz (1888), Schmidt (1885), and Harms (1909) agree that the first 
 rudiments of the gills originate, and which are very conspicuous structures in the 
 glochidia of the hooked type. We have never observed any structure resembling the 
 larval thread or its rudiment in the fully formed glochidia of species of the genera 
 Lampsilis and Ouadrula, the glochidia of which we have studied most extensively; and 
 the larval thread is not present in functional condition in any of the species we have 
 studied from the genera listed on page 145, with the exception of Anodonta and Unto. 
 A discussion of this organ, which has heretofore been assumed to occur in all glochidia, 
 is given after the account of the hooked glochidium which follows. 
 
 THE HOOKED TYPE. 
 
 Our first infections were performed with the hooked glochidium of Anodonta 
 cataracta, which is essentially like the Anodonta type of glochidium described for Euro- 
 pean species, and which has been described in a detailed manner by Lillie (1895). Our 
 later work has been with the young of Symphynota complanata and 5. costata, the glochidia 
 of which resemble one another in structure, as shown by their outlines in text figure 1, 
 a and b, and figures 9 and 10, plate vm; so that here, as elsewhere noted in the case 
 of hookless glochidia, the outline appears to be a characteristic of the genus, which 
 enables one at once to distinguish the glochidia of Anodonta from those of Symphynota. 
 There is, however, a marked size difference between the glochidia of these two species 
 of Symphynota (text fig. 1, A and b). 
 
 In both Anodonta and Symphynota glochidia, the slightly greater length of one 
 border of the valve between hook and hinge is indicative of the future anterior region. 
 In most hookless glochidia there is a similar slight difference in the anterior and posterior 
 marginal outlines (fig. 13, pi. vm), but it is more difficult to detect, and in any case 
 the safest guide is the larval adductor muscle, which is always recognizably nearer the 
 anterior end, a position to be correlated with the location of the rudiments of the adult 
 organs in the posterior region. In the living glochidium of 5. complanata the shell 
 shows calcification beneath the cuticle and is marked as though the calcareous layer 
 were porous. 
 
 The external appearance of these hooked glochidia is like that shown for S. costata 
 in figure 10, plate vm. The hooks, with their spines, the fibers of the larval adductor, 
 and the sensory cells are seen when turned in profile view (fig. 9, pi. vin) ; but the 
 cellular structure is so obscure in living specimens that the rudiments appear only as a 
 denser area and even the fibers of the adductor muscle are not very distinct. There is 
 no sign of a larval thread or a thread gland, nor do sections of preserved glochidia reveal 
 such a structure. A conspicuous feature of the whole mass of glochidia in Symphynota, 
 as taken from the gill of the parent, is the thick, ropy mucus in which they are embedded. 
 This holds them so firmly together that when stirred up in a dish they remain suspended 
 and quite evenly distributed through the water, settling to the bottom only very slowly 
 
150 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 over a period of four or five minutes. During this suspension in the water the sucking 
 of a pipette will draw in glochidia over a wide area, as they are pulled by the invisible 
 strands into which the mucus has been divided. The significance of this mucus and the 
 absence of the thread gland are discussed under another heading of this paper. The 
 mucus is dissolved by the water in a short time, so that after 24 hours the glochidia are 
 found entirely free and snapping actively upon the bottom. We find that these glochidia 
 can be freed from the mucus by repeated washing, and that it is desirable to do this at 
 once if one wishes to keep them alive for the maximum period. When thus set aside 
 it is possible for them to remain alive for as long a time as two or three weeks. 
 
 In killing this glochidium we have used successfully crystals of chloral hydrate or 
 hydrochlorate of cocaine added to the water of a watch glass containing the glochidia, 
 and fixation with Merkel's fluid, or with weak corrosive sublimate, as described for 
 the hookless type. 
 
 Stained specimens show the same rudiments of stomodseum, enteron, and meso- 
 dermal structures, as described by Iyillie (1S95) and Harms (1909) for the glochidium of 
 Anodonta. The lateral pits are conspicuous and the cells of the larval mantle are well 
 developed laterally, though thinning out over the median part of the larval adductor, 
 where their boundaries are not clear and only a few nuclei are discernible. Sections show 
 two kinds of granules within the larval mantle cells, one staining deeply with iron 
 hematoxylin and the other with acid-fuchsin. Near each corner of each valve is a cell 
 which stains deeper than the rest and seems to contain more of the granules. The 
 significance of these six cells we can not determine. The sensory cells (fig. 9, pi. vm) 
 are slightly different in position from those in Anodonta. Lying along a line drawn 
 across from hook to hook are three large cells in line beneath the hooks and a smaller 
 one on either side between the larval adductor and the lateral pit. 
 
 THE PROPTERA OR AXE-HEAD TYPE. 
 
 This glochidium possesses hooks which are not homologous with those of the 
 Anodonta type and is to be regarded as more nearly related to the hookless forms, an 
 interpretation which is borne out by the fact that the "axe-head " can be readily imagined 
 as a modification of the glochidial outline seen in some species of Lampsilis, the glochidia 
 of which, like those of subrostrata (fig. 13, pi. vm), show some approach to a rectangular 
 form. Its four hooks are so arranged that those of one valve pass inside the opposite 
 ones, thus bringing the ventral margins close together and giving a very firm hold upon 
 the host's tissue. In other respects it does not show marked differences from the hookless 
 type, and the few experiments we have made with it indicate its attachment to the gills 
 rather than to the fins. 
 
 Recently Coker and Surber (191 1) have observed "an almost exactly similar 
 glochidium" in Lampsilis capax, while in Lampsilis (Proptera) larvissima they find an 
 axe-head glochidium which is of a somewhat different outline and lacks the hooks. 
 They point out the fact that in Lampsilis gracilis, a species which in its adult features 
 (form of shell) seems almost to intergrade with Icevissima, the glochidium is of the ordinary 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 151 
 
 hookless tvpe, although the outlines of the two glochidia are very similar when seen on 
 edge, as in their figures ia and 2a of plate 1. With respect to the significance of these 
 facts when applied "to a relationship between Iczvissima and capax," they conclude that 
 "there would be strong corroborative evidence in adult characters alone" for the closer 
 union of these three species, and this "in spite of the fact that larvissima and capax are 
 the two extremes in the degree of inflation." The similar degree of inflation of capax 
 and ventricosa offers, they believe, "only a striking instance of convergence in one 
 character." 
 
 THE LARVAL THREAD. 
 
 Our observations upon the occurrence of the larval thread (formerly erroneously 
 termed the bvssus) are of importance, since the current accounts in textbooks and 
 literature lead one to believe that this structure is a conspicuous feature of all glochidia. 
 Such an assumption is natural because the organ is conspicuous in the European ano- 
 dontas and unios and in the American species of these genera examined by Lillie (1895). 
 
 We find the larval thread present in the species of Unio and Anodonta which we 
 have been able to examine with care, and the thread is undoubtedly a characteristic of 
 these genera. We have never seen any sign of such a structure in the ripe glochidia of 
 the other genera, above listed, which possess hookless glochidia, nor in the hooked forms 
 of the genus Symphynota. Lillie (1895, p. 52) considers the thread a condensed excre- 
 tory product, which, accepting the account of Schierholz (1888), he thinks has also 
 become an organ which is of use in bringing the glochidium in contact with the fish. 
 This latter function is the one commonly ascribed to the thread. We have not studied 
 the pre-glochidial stages in the development of those species which show no thread- 
 gland in the mature glochidium, although it is important that this should be done with 
 a view to determining whether a homologue of the thread gland is present at any time. 
 We have, however, made repeated examinations of glochidia, either ripe or well along 
 in their development, in several species of Lampsilis, particularly in ligamentina , recta, 
 anodontoides , ventricosa, luteola, and subrosirata, and to a lesser extent in species of the 
 other genera mentioned, without finding any trace of the thread which is so conspicuous 
 a feature of the glochidium of Unio complanatus . 
 
 We have also examined the glochidia of Symphynota complanata many times with 
 the same negative results, and a smaller number of observations confirm this for S. cos- 
 tata. Since many species thus have no thread in any way functional for attachment to 
 the fish, the question arises whether the thread when present has as important a func- 
 tion in this respect as has been supposed. Our observations upon the glochidia of 
 Anodonta calaracta confirm the descriptions of Schierholz (1888) and others who have 
 studied the European species of A nodonta as to the tangling of the glochidia into masses 
 by means of their extruded threads, and in this genus the threads do seem effective in 
 drawing other glochidia into contact with the fish when a single one has become attached. 
 This is not, however, effective during the greater part of the period in which the glo- 
 chidium may remain alive upon the bottom, for the threads are dissolved within a day or 
 18713 — 12 4 
 
152 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 two and the glochidia then become entirely free from one another. When taken from 
 the parent gill the glochidia of Symphynota are entangled in a ropy mucu?, and this acts 
 in a manner similar to the threads of Anodonta, but it is usually dissolved after a few 
 hours in the water. In the ripe glochidium of U. complanalus the threads are extruded 
 immediately after the glochidia are removed from the parent and placed in water, and, 
 according to Harms (1907b, p. 819), the minute glochidia of Margaritana margaritifera 
 extrude their threads while still within the egg capsule. 
 
 When this extrusion has taken place in Unio complanatus the glochidia and broken 
 egg membranes become united into globular masses from which it is difficult to separate 
 individual specimens, and from observing such glochidia in contact with the fish we 
 are forced to conclude that they are not so likely to become attached to the gills or fins 
 as they are later, when they have been separated by the disintegration of the threads. 
 The glochidia of Lampsilis, which when fully ripe fall apart into masses of entirely 
 unconnected individuals, appear much better able to attach to the gills of fishes imme- 
 diately after their discharge from the parent. We believe, therefore, that the thread is 
 something to be gotten rid of rather than an organ of great importance in the attach- 
 ment to fish, and this is in agreement with Lillie's interpretation of this organ as an 
 excretory product. It is possible that some homologue of the thread exists in these 
 threadless glochidia, and a comparative study of the pre-glochidial stages might yield 
 material for interesting comparisons. 
 
 BEHAVIOR AND REACTIONS OF GLOCHIDIA. 
 
 At the time of spawning the glochidia, already freed from the egg membranes, and 
 usually held together in slimy strings, are discharged at irregular intervals. Being 
 heavier than water, they sink rapidly to the bottom, coming to rest with the outer sur- 
 face of the shell directed downward and the valves gaping widely apart. The belief was 
 formerly general that they "swim" about by rapidly opening and closing the valves, 
 after the manner of Pecten, and, in spite of frequent denials by Schierholz (1888), Latter 
 (1891), and others, the same statement is still occasionally encountered. In the recent 
 volume on Mollusca in the Treatise on Zoology, edited by Lankester, this inexcusable 
 error is repeated. "The glochidia," we are again informed, "swim actively by clapping 
 together the valves of the shell" (p. 250). They are, on the contrary, as is now well 
 known, entirely incapable of locomotion and remain in the spot where they happen to 
 fall, although it is true that they may exhibit from time to time spasmodic contractions 
 of the adductor muscle, which cause the valves to snap or wink, each contraction being 
 immediately followed by relaxation and opening of the shell. These movements of the 
 valves, however, are never so vigorous as to cause the glochidium to move from place 
 to place in the water. 
 
 The glochidia remain in this helpless situation until they die, unless they happen to 
 come in contact with the host on which they pass through the post-embryonic develop- 
 ment as parasites. The stimulus which causes the contraction of the muscle and results 
 in attachment to the host is, in the case of hookless glochidia, usually a chemical one, 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 1 53 
 
 but in that of the hooked forms it is mechanical. The latter may be readily imitated 
 and glochidia of this type made to grasp firmly the point of a needle or the edge of a 
 piece of paper by simply touching them between the open valves. When once closed 
 in this manner they do not relax, but remain attached to the object until they die. 
 
 The following statement made by Latter (op. cit., p. 56) has been frequently quoted, 
 especially in textbooks, but it has apparently never been verified or disproved. 
 
 The Glochidia are evidently peculiarly sensitive to the odor (?) [sic] of fish. The tail of a 
 recently killed Stickleback thrust into a watch glass containing Glochidia throws them all into the 
 wildest agitation for a few seconds; the valves are violently closed and again opened with astonishing 
 rapidity for 15-25 seconds, and the animals appear exhausted and lie placid with widely gaping shells, 
 unless they chance to have closed upon any object in the water (e. g., another Glochidium), in which 
 case the valves remain firmly closed. 
 
 Although it is not stated that the tail which caused such a commotion among the 
 glochidia had been cut off from the fish, it is probable that such was the case. We have 
 repeatedly tested glochidia in the same manner both with fins and gills of different 
 fishes, and, providing that a bleeding surface is not brought in contact with the water 
 containing the glochidia, absolutely no response on the part of the latter takes place. 
 The result, however, is much as Latter describes if a little of the fish's blood gets into the 
 water in the neighborhood of the glochidia, except that our experience has shown that 
 after snapping for a few seconds they come to rest in permanent closure. It therefore 
 seems possible that the contractions seen by Latter were due to the introduction of some 
 blood with the tail of the fish, as otherwise agitation of the glochidia under similar con- 
 ditions has not been observed by us. 
 
 Since the hooked and hookless glochidia, whose reactions to blood and to certain 
 salts we have studied, show important differences in their behavior, they are referred 
 to separately below 
 
 REACTIONS OF HOOKLESS GLOCHIDIA. 
 
 It was first observed that glochidia of the hookless type, in marked contrast with 
 the hooked forms, only occasionally exhibit spontaneous contractions and respond either 
 not at all or only sluggishly to tactile stimuli, and the question at once arose as to what 
 causes their closure when they become attached to fish. If the stimulus which brings 
 about a contraction of the adductor muscle in attachment is not a mechanical one, it 
 presumably is chemical in nature, but we were completely in the dark in the matter until 
 it was cleared up by the following experiments, the first of which were made with the 
 glochidia of Unio complanatus at Woods Hole, Mass. 
 
 When a small drop of blood of either the killifish, Fundulus diaphanus, or the white 
 perch, Morone americana, was placed over the glochidia contained in a small amount of 
 water in a watch glass, the effect was immediate and very striking. Every glochidium 
 was thrown into rapid and violent contractions, alternating with relaxations, the edges 
 of the valves either quite or nearly touching with each snap. Where the stimulus was 
 strongest — that is, immediately under the drop of blood — the glochidia exhibited two or 
 three strong contractions and then remained closed, but, proceeding outward to zones 
 
154 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 of diminishing intensity, the snapping occurred intermittently for from 10 to 50 seconds. 
 Here the contractions were quite rapid at first, one or two every second, but soon the 
 intervals became longer, until finally the activity was ended by the closure of the valves. 
 In some cases it was observed that after the first few snaps the muscle did not completely 
 relax, and each subsequent contraction caused the valves to describe a shorter arc. This 
 experiment was repeated time and time again, with invariably the same result, and it 
 was astonishing to see what a small quantity of the fish's blood was required to produce 
 the reaction. It should be emphasized, furthermore, that after the stimulus had caused 
 the final contraction of the muscle the valves remained permanently closed. 
 
 The experiment was later performed a great many times with the glochidia of Lamp- 
 silis ligamentina and subrostrata, and identically the same reaction was obtained with 
 the blood of several different fishes and that of the frog, Nccturus, and man. 
 
 Since the hookless glochidia, which are essentially gill parasites and, when taken 
 into the mouth of the fish lodge among the gill filaments, produce abrasions of the 
 delicate epithelium covering the latter, a more or less extensive hemorrhage from the 
 blood capillaries occurs, as may be readily seen from a microscopic examination. It is 
 therefore evident that blood exuding from the gill filaments in the immediate neighbor- 
 hood of the glochidia must have the same effect as in our experiments, and, by exciting 
 vigorous contractions of the adductor muscle, furnish an efficient stimulus in bringing 
 about a firm and permanent attachment to the filaments. It is true that hookless 
 glochidia will occasionally secure an attachment to the edge of the fins and other exter- 
 nal parts of the fish, but it is quite evident that they are not adapted to such locations, 
 as they rarely succeed in remaining there. It is possible that when they do become 
 attached to the fins the closure of the valves is due to the presence of blood on the latter; 
 but, since hookless glochidia occasionally close when touched repeatedly, the attach- 
 ment in these situations is probably brought about by a sluggish response to contact 
 with the edges of the fins. Their characteristic place of attachment, however, is the 
 gill filaments, and this definite reaction to the fish's blood constitutes a most striking 
 functional adaptation to the special habit of hookless glochidia as gill parasites. 
 
 Although the matter has not been exhaustively studied, it is in all probability the 
 salts of the blood that are responsible for these reactions. A series of experiments, 
 however, has been undertaken for the purpose of determining the reactions of glochidia 
 of this type to solutions of several different salts, and, although the investigation has 
 not yet been completed, a brief statement may be made here. Diluted sea water and 
 solutions varying in strength from 0.5 to 1 per cent of NaCl, KC1, KN0 3 , and NH 4 C1 
 have exactly the same effect as fish's blood, although the intensity of the reaction varies 
 somewhat in certain cases. Weak solutions of MgCl, and MgS0 4 , however, as would 
 be expected, inhibit contractions, and glochidia, after treatment with these salts, may 
 be killed in an expanded condition, if allowed to remain in the solutions for a sufficient 
 length of time. 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 1 55 
 REACTIONS OF HOOKED GLOCHIDIA. 
 
 The larvae of Symphynota complanaia, which are provided with stout hooks and 
 as a rule find permanent lodgment only on the fins and other external parts of the fish, 
 were used in studying the reactions of the hooked type of glochidium. In several re- 
 spects they differ from the hookless forms. When removed from the marsupium and 
 placed in water, they exhibit spontaneous contractions which occur at irregular and 
 rather long intervals, and this irritability may continue in the laboratory for a day or 
 two, or until the glochidia begin to disintegrate. Under such conditions the valves 
 are only partially closed at each contraction of the muscle, which, moreover, is never 
 strong enough to bring the points of the hooks into contact. It is followed at once by 
 relaxation of the muscle and the shell remains widely open until the next snap occurs. 
 
 Hooked glochidia, in striking contrast with the behavior of the hookless forms, 
 respond very actively to tactile stimuli, and, as has been stated, close completely and 
 immediately when touched with any object. This reaction must be the main factor 
 in bringing about their attachment to the fish's fins, when they are brushed over by 
 the latter while lying on the bottom. With glochidia like those of Symphynota com- 
 planaia the mere contact is sufficient to produce complete closure of the valves, and, 
 whether they are exposed to the fish's blood or not, attachment is possible as a result 
 of the tactile stimulus alone. They do react to blood, however, and exhibit a few 
 successive contractions, from 5 to 15, before final closure, but the way in which the 
 response occurs is quite different from that shown by hookless glochidia under similar 
 conditions. Instead of being thrown into violent and rapid snapping, the valves closing 
 and opening alternately, there is only partial recovery after each contraction, while 
 the valves are brought closer and closer together by a series of short jerks. The final 
 act of closing is interesting. As soon as the points of the hooks touch, the contraction 
 of the adductor muscle becomes continuous and the hooks are slowly bent inward 
 against each other. Under the steady pressure exerted by the muscle, aided probably 
 by the action of the myocytes, which have been described by Schmidt (1885b), the 
 spines on the outer surface are apposed and the hooks turned in completely between 
 the valves, the margins of which are brought together, if no object intervenes. It will 
 be readily understood that, owing to the turning in of the books, the spines are pressed 
 into the fish's tissues, when attachment to the host takes place, and a firm hold is thereby 
 secured. 
 
 When the glochidia of Symphynota complanaia were exposed to salt solutions, the 
 contractions produced were of the kind just described. KC1, KN0 3 , and NH 4 C1 in 
 solutions of 0.5 to 1 per cent caused a few successive jerks, the contractions being more 
 vigorous and closure occurring sooner with the stronger solutions. NaCl and Na 2 C,0., 
 in the same strength acted less energetically, and it was necessary to use a 2 per cent 
 solution to produce the same effect as was obtained with the weaker solutions of potas- 
 sium and ammonium salts. A 0.5 per cent solution of CaCl 2 produced no contractions, 
 while a 1 per cent solution after a latent period of 15 minutes caused either partial or 
 complete closure of the valves. MgCl 2 and MgS0 4 , in solutions of 0.5 and 1 per cent, 
 
156 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 inhibited contractions, and when the glochidia were allowed to remain in them they 
 finally died in the expanded condition. When the Mg salts, however, were used in 
 stronger solutions, closure of the valves occurred after a few spasmodic contractions. 
 
 IV. THE PARASITISM. 
 ARTIFICIAL INFECTION OF FISH. 
 
 In anv investigation which attempts to ascertain the facts of most importance for 
 the artificial propagation of a species, attention is at once directed to those points in the 
 life history where wholesale destruction of the individuals is most likely to occur. These 
 points of wholesale waste are usually to be found in the earlier part of the individual's 
 existence rather than during its adult life and are often preventable by artificial means. 
 In common with other animals which must overcome the chances of parasitism, the 
 Unionidae produce enormous numbers of eggs, the great majority of which are by virtue 
 of the brooding habit of the female mussel carried safely through their embryonic period 
 and discharged as glochidia. We have not attempted to estimate the numbers of 
 glochidia carried by full-grown adult females, but anyone who has seen them taken 
 from the gills knows that they must be numbered by the hundreds of thousands, or even 
 millions, and had these glochidia any great chance of survival and development to the 
 adult stage the supply of mussels would far exceed anything which has ever been known 
 in nature. When, however, the next stage of the larval history is sought for in nature, 
 it becomes apparent that we have reached a point in the life cycle where the destruction 
 and waste of individuals is wholesale and probably in excess of that which occurs at any 
 other stage. There is no evidence, save in the case of the species Strophitus edenlulus, 
 the metamorphosis of which we have discussed under another heading of this paper, 
 that any one of the Unionidae can pass beyond the glochidial stage without becoming a 
 parasite upon some fish, for the failure of glochidia to develop when left in water has 
 been observed by all investigators since Leeuwenhoek. 
 
 The large element of chance involved in this shift from parent to fish, which has 
 already been emphasized in our discussion of the glochidium, is again apparent when 
 fish are examined in nature with a view to determining the abundance of the parasitic 
 larvae under the conditions of natural infection, for all investigators agree that the para- 
 sites exist in numbers which are insignificant when compared with the masses of glochidia 
 which occur in the parent mussels. Only an occasional fish is found to be infected and 
 it thus becomes clear that the purely accidental nature of the infection makes necessary 
 the production of glochidia in such abundance as to overcome by sheer force of numbers 
 the chances of destruction. Fish become infected in nature by occasional glochidia, 
 but the chance that any fish will carry under natural conditions the number of glochidia 
 which our experiments have shown that individual fish are capable of carrying, when 
 artificially infected, is a negligible quantity. Here, then, we have the point of greatest 
 destruction in the life cycle of the Unionidae; and the point of attack for artificial propa- 
 gation is clear. The fish must be made to carry more glochidia. Under experimental 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 1 57 
 
 laboratory conditions it is found that a given fish may carry successfully a load of glochi- 
 dia so much in excess of what the same fish would ever be likely to carry in nature that 
 there is no reason why a single fish should not be made, under the conditions of artificial 
 infection, to do the work which a thousand fish perhaps could not do in the state of 
 nature. This has been from the first our main point of attack, and, with this in view, 
 we have studied the parasitism, first, by the infection of small lots of fish in aquaria 
 and, later, by the infection of fish in larger numbers in a hatchery. Other points in the 
 life cycle, as for example the stage immediately following the parasitism, may be found 
 by later work to be places of wholesale destruction; we are convinced, however, that 
 there can be no other where the mortality reaches such proportions as it does wheji the 
 countless glochidia are spread upon the bottom and left to the chance that will bring 
 them in contact with the parts of a fish's body suitable for their parasitism. 
 
 Throughout our experimental infections we have made use of small fish, usually 
 those under 6 inches in length, because such fish are more easily collected in numbers 
 and because we have not had proper facilities for the keeping of larger individuals. 
 Where small numbers of fish are used and each individual can be carefully watched, the 
 attainment of what may be termed an ' ' optimum ' ' infection in every case may be secured 
 with no great difficulty, and by following the methods practised by various investigators 
 ever since Braun (1878) and Schmidt (1885), we have obtained unlimited material 
 whenever necessary. If the glochidia are placed in shallow dishes and in water just deep 
 enough to cover all parts of the fish, the latter will usually keep the water sufficiently 
 agitated to insure a proper suspension of the glochidia and tolerably constant results will 
 follow. 
 
 It is very necessary that the glochidia be so distributed in the water as to come in 
 contact with the proper parts of the fish, and, in most cases, to guard against over rather 
 than under infection. Active fish, such as the rock bass (Ambloplites rupestris), and the 
 large-mouthed black bass (Micropierus salmoides) , are very favorable for gill infections, 
 since they keep the water so well agitated that the glochidia hardly settle to the bottom 
 at all, while their strong respiratory movements draw the suspended glochidia con- 
 tinually against the gills. With fish like the crappie (Pomoxis annularis), which when 
 undisturbed move about quietly and whose respiratory movements are less vigorous, the 
 water must be stirred to keep the glochidia suspended, or be so shallow that the fish are 
 always near the bottom. The smaller gill slit of the crappie is another factor which 
 makes for a very light infection in fish under 2 inches in length, since the glochidia reach 
 the gills by way of the mouth and not from the opposite direction. For fin infections, 
 sluggish fish like the German carp {Cyprinus carpio) need little attention, and the darters 
 (Ethcostoma cceruleum spectabile), which habitually rest upon the bottom for considerable 
 periods, become quickly loaded with glochidia upon both fins and gills; although, as we 
 shall see, the latter fish appears to be particularly adapted for ridding itself of the entire 
 infection. 
 
 In the account which follows, we are discussing the results obtained from the 
 infection of fish in small numbers and kept under careful observation in the laboratory. 
 
158 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 There is no reason for believing that larger numbers of fish would present any more 
 serious difficulties than are to be expected in the keeping of any fish in large numbers 
 within a restricted space; and, if one could insure as uniform and careful an infection of 
 the larger numbers, we have every reason to believe that such infections would prove 
 as successful as those here described. 
 
 INFECTIONS WITH HOOKED GLOCHIDIA. 
 
 For the infections with hooked glochidia, we have used principally Anodonta cala- 
 ractairom. Falmouth, Mass., the species studied by Lillie (1895). With these we infected 
 German carp under 6 inches in length and, unless otherwise stated, the following account 
 refers to this combination which gives typical results! A smaller number of infections, 
 made with Symphynota complanata and 5. costala upon carp and other fishes, are referred 
 to in a supplementary manner. The glochidia of A. cataracla become attached in large 
 numbers to the fins (fig. 19-25, pi. ix and x) and gills of the carp. They are also found 
 upon the other external parts which offer the condition of a soft scaleless epithelium like 
 that of the fins; thus, the region about the anus, the edge of the operculum, the lips and 
 in very heavy infections, even the soft area of the ventral surface between the mouth and 
 pectoral fins may become heavily loaded. Within the mouth cavity, the gill filaments 
 and also the gill bars and rakers become well covered. The glochidia which attach to 
 these mouth parts do not remain, for, although the fish may be carrying many of their 
 fellows upon its external parts, in about one week after the infection all glochidia have 
 disappeared from the gill filaments, which then become as clean as though never infected. 
 Scattered glochidia may remain upon the other internal mouth parts, for specimens are 
 occasionally seen well embedded and in advanced stages of their metamorphosis, but in 
 the main these parts also will become free of glochidia. 
 
 The general distribution upon the individual fins may be seen by reference to figures 
 19 to 25, plates ix and x, which show how great a proportion of the glochidia become 
 attached to the fin margins. If a fish is carefully watched, as its slight movements stir 
 up the glochidia during the infection, the latter are seen continually falling upon the 
 upper faces of the pectoral and pelvic fins. They may even be collected with a pipette 
 and heaped upon a motionless pectoral fin, remaining there for some minutes without 
 more than an occasional specimen becoming attached. The margin of the fin is so much 
 more favorable for attachment, that it is often thickly set with glochidia, when none are 
 found upon the fin surface, and this despite the fact that glochidia must, during infection, 
 strike against the surface of the fin many times for every time that one of them comes 
 in contact with a fin margin. It is, therefore, the margin of the fin for which this glochi- 
 dium is best suited, and, once fastened there, it is almost certain to remain and become 
 embedded by the growth of the host's epithelium. 
 
 Considered in a more detailed way and with reference to the parts of the glochidium, 
 we may explain this more frequent attachment to the margin as due to the fact that 
 when the glochidium strikes against any flat surface the sensory hairs are not stimulated 
 and the glochidium, which, as we have already shown in the case of the hooked forms, 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 1 59 
 
 responds principally to tactile stimulation, does not receive the stimulus to permanent 
 closure which is given by the presence of any foreign object inserted between the valves. 
 When a specimen does become attached to the surface of a fin, as is sometimes the case 
 (fig. 21 and 22, pi. ix, fig. 25 and 32, pi. x), it presumably gains its hold by catching 
 upon one of the ridges formed by the fin rays, for the hooks could hardly be used upon a 
 perfectly flat surface. Glochidia sometimes hold to the surface of a fin by a shred of 
 tissue, under which their hooks have caught, remaining there after all the neighboring 
 specimens are completely overgrown (fig 25, pi. x), only to be torn off later without 
 having caused any noticeable hypertrophy of the fin tissue. Figures 25 and 32, plate x, 
 show that glochidia may become overgrown either flat against the surface or upon edge, 
 and figure 24, plate ix, shows a young mussel leaving a surface attachment after a 
 parasitism of 74 days. 
 
 The behavior and reactions of glochidia are of course significant in connection 
 with the actual attachment when once the glochidium is brought in contact with a 
 suitable part of the fish's body and receives the normal stimulus to close its valves. 
 The bringing of the glochidium against just that part of the fish is a matter of the chance 
 distribution in the water. Hence the distribution of the glochidia to the several fins 
 is determined solely by the number likely to be brought in contact with a given part 
 of the body. Those fins which brush against the bottom are always the more heavily 
 loaded and the numbers elsewhere depend upon the extent to which the glochidia are 
 kept suspended in the water. The importance of the mucus for the glochidia of Sym- 
 phynota and of the larval thread for those of Anodonta and Unio in tangling the glochidia 
 into masses and drawing others against the fish when a single one has become attached 
 has probably been exaggerated, as explained in the section of this paper which deals with 
 the function of the larval thread. 
 
 Optimum infections, as we shall term those which are close upon the limit of the 
 number of glochidia which a fish can safely bring through the metamorphosis, often 
 show the glochidia very closely set one after another, as in figures 22 and 23, plate 
 ix, and figure 25, plate x, and several hundred may be safely carried by a fish 3 or 4 
 inches in length. Prolonged exposure causes so heavy an infection of the margins 
 (fig. 19 and 20, pi. ix) that the fin tissue appears unable to overgrow the mass of 
 glochidia, and they then remain attached without overgrowth for a week or more. 
 
 Figure 19, plate ix shows how on a part of the fin having no overcrowding normal 
 embedding occurred, while in the more crowded areas the glochidia were still uncovered 
 even seven days after infection. In the middle upper margin of this fin it would seem 
 that the overgrowth might well have taken place, for many cases like figure 25, plate x, 
 have been observed in which glochidia as closely set were properly embedded. The 
 failure of overgrowth in this region is probably due to the presence immediately after 
 infection of a greater number of glochidia many of which have since been detached. 
 In all cases of this kind a smaller number will finally become embedded than in an 
 infection where the fin has received more nearly the optimum load (fig. 21, 22, 23, pi. ix, 
 and fig. 25, pi. x), for the great majority drop off when the fin becomes so mutilated 
 
160 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 that bacterial or fungus infection sets in. These over-infections sometimes cause such 
 hypertrophy that the fins become swollen and the rays so drawn together that it is 
 impossible for them to spread out normally. Often the fins are raw and bleeding for 
 some days and show red areas within where the blood vessels have become abnormal. 
 The fish are likely to die from this or from the similar injury to their gills, and these 
 over-infections are unsatisfactory if one wishes to bring through their parasitism the 
 maximum number of glochidia. 
 
 The steps in the implantation of the glochidium by an overgrowth of the fish's 
 tissue may be seen in figures 21 and 22, plate ix, and figure 25, plate x. Figures 21, 
 plate ix, and 26, plate x, show the glochidium 3% hours after attachment to the 
 fish's fin. Most of the glochidia have bitten deep enough in from the margin to have 
 a good hold for their hooks. The beginning of the hypertrophy appears as a faint 
 mass of tissue, seen with its nuclei in the detailed figure 26, plate x. At the end of 
 12 hours the overgrowth is well advanced and sometimes, as in figure 27, plate x, shows 
 different stages even in neighboring glochidia. The ragged edge of the host's tissue 
 rises up crater-like about the glochidium, meeting above in a delicate mass, the nuclei 
 of which are shown. Figure 22, plate ix, shows that in 24 hours most of the glochidia 
 are more than half covered, whether upon the edge or the surface of the fins. At the 
 end of 36 hours (fig. 25, pi. x) optimum infections of the carp show all the glochidia 
 which have obtained a proper attachment well embedded, and from this time onward 
 the only change which is visible in whole mounts is a slight increase in the opacity 
 of the cyst, which renders the internal structure of the glochidium less distinct (fig. 23, 
 pi. ix). Some of our infections show embedding in as short a time as 6 hours (Sym- 
 phynota), and Harms (1909) gives 10 to 12 hours as the time which he observed in 
 Anodonta, so the time given for the figures above referred to is the maximum for hooked 
 glochidia which have been well located. Glochidia upon the fin surface become embed- 
 ded in a similar manner and are then in a very secure position (fig. 22, pi. ix, fig. 25 and 
 32, pi. x). 
 
 INFECTIONS WITH HOOKLESS GLOCHIDIA. 
 
 Our experiments in artificial infection with hookless glochidia have been more 
 extensive because this is the type of glochidium found in the species of mussels which 
 are of commercial importance. Species of the genus Lampsi/is (ligamentina, recta, 
 anodontoides, ventricosa, subrostrata, and luteola) have been the most frequently used, 
 but infections have also been made with several species of Quadrula and one of Unio. 
 The list of fishes employed as hosts for hookless glochidia is also more extensive and we 
 are, therefore, able to make statements which we know to be of wider application than 
 those made for the hooked glochidia. 
 
 When the same fish is used, the results for the several species of Lampsilis are 
 very uniform and we can thus discuss the parasitism of this genus as a whole; but we 
 do not find the same mussel giving uniform results with all species of fish. The glochidia 
 of this genus have been used successfully for the infection of blue-gill sunfish (Lepomis 
 pallidus), yellow perch (Pcrca flavescens), crappie, large-mouth black bass, rock bass, 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. l6l 
 
 the red-spotted sunfish (Lepomis humilis), and the green sunfish {Apomotis cyatiellus). 
 As with the hooked gloehidia, the infections have all been made upon fish under 6 inches 
 in length, upon which these gloehidia remain in numbers only on the gill filaments, 
 although during infection some may become attached to and even embedded upon fins 
 and other external parts. Harms (1908) concludes that the hookless type persists in 
 much greater numbers on the fins of small than of large fish, and that the hooked type 
 will survive upon the gills if large fish are used. It is doubtless true that the size of 
 the gills and fins is an important factor in determining the place of attachment for 
 each type, since the hookless form is better adapted for holding to a delicate surface 
 like a gill filament or a fine fin, while the hooked type seems likely to be easily torn 
 from such a surface. When the hookless form does once become established upon an 
 
 Fig. 2. — Rock-bass (Ambloplites rupestris) infected with gloehidia of Lampsilis ligamentina. About 2 
 carried through the metamorphosis by each fish in this infection. Note the large number c 
 
 00 were successfully 
 
 1 the gills. 
 
 external part, it will develop there without mishap, as shown by the figure of a hooked 
 and a hookless glochidium developing side by side upon the margin of a fin (fig. 29, 
 pi. x). Within the mouth cavity these gloehidia become attached to the gill bars and 
 iakers, if these parts are covered by a sufficiently delicate epithelium, though they are 
 always found in the greatest numbers upon the gill filaments. In most of our infec- 
 tions the filaments are more heavily infected toward their outer ends (fig. 43, pi. xi), 
 the distribution varying somewhat with the species of fish. For example, successful 
 infections of rock bass with Lampsilis ligamentina show about seven gloehidia upon 
 the distal third of the filament to one upon the proximal two-thirds; of large-mouth 
 black bass about 3 to 1, and of yellow perch about 1% to 1 — differences which are 
 probably due to some particular configuration of the mouth parts, which causes the 
 gloehidia to fall more upon one region of the filaments than another. 
 
1 62 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 In a fish which will carry a given glochidium successfully, over-infection of the 
 gills is easily accomplished and easily fatal, although species of fish differ greatly in the 
 amount of infection they are able to withstand without serious mortality. In one 
 of our most successful combinations (rock bass infected with Lampsilis ligament inn), 
 fish 4 inches in length were estimated to be carrying in the neighborhood of 2,500 
 glochidia, an average of more than two for every filament of the gills and yet there was 
 almost no mortality among the fish. A rock bass from this infection is shown in text 
 figure 2, which also illustrates the distribution of the glochidia on a single fish. In 
 this case the success of so heavy an infection is perhaps explained by the distribution 
 of the glochidia upon the gill filaments, for we found by count that there were about 
 seven near the tips to one on the proximal two-thirds of the filament, and thus the greater 
 part of every filament was left unchanged and in full functional condition, while in 
 other infections (large-mouth black bass with L. ligamentina) , where a much greater 
 proportion of the glochidia were upon the sides of the filaments, the mortality of the 
 fish was heavy, although the amount of infection was much less. A gill of the latter 
 fish infected with these glochidia is shown in figure 39, plate xi. The number estimated 
 for this fish, which was 4 inches in length, being only 450, is less than the optimum. 
 
 Implantation upon the filaments occurs in a manner similar to that of the hooked 
 glochidia upon the external parts, but much more rapidly. Figures 35, 36, 37, and 38, 
 plate xi, show the appearance at 15 minutes, 30 minutes, 1 hour, and 3 hours, respectively, 
 after infection, and our observations, showing that the cyst is completed within from 2 
 to 4 hours, agree with what Harms (1909) has found for gill infections. The prolifera- 
 tion will even continue after the gill has been cut from the fish and placed in a watch glass 
 for observation under the microscope (fig. 54 and 55, pi. xm). An immediate result of 
 the cyst formation is the obliteration of the lamella upon either side of the gill filament, 
 which thus becomes smooth and slightly swollen in the vicinity of the glochidium (fig. 43, 
 pi. xi). Figures 34 and 43, plate xi, show the general and detailed appearance of the 
 cysts and the diversity in the angles at which the glochidia are attached. 
 
 The older statement that the hooked glochidia are fin and the hookless gill parasites 
 finds, therefore, confirmation from our work, although it would be better to say that the 
 hooked attach most successfully to large strong margins like those of the fins, and the 
 hookless to soft and fine filamentous structures like the gills in fish of moderate size. 
 The reactions of the two types of glochidia to mechanical and chemical stimuli, with 
 respect to the part they play in attachment, have already been discussed. 
 
 SUSCEPTIBILITY OF FISHES TO INFECTION. 
 
 The susceptibility of different fishes to infection is a matter which has not been suffi- 
 ciently considered by any previous investigators. We have evidence that some species 
 are much less susceptible than others to one or the other type of glochidium, and that in 
 these cases any considerable infection is an impossibility. The most striking instances 
 of this are the German carp, certain minnows, and the darters. 
 
REPRODUDTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 1 63 
 
 In the case of the carp, while the fish is admirably suited to carrying the hooked 
 glochidia of Anodonta and Symphynota, we have never been able to secure a successful 
 infection of the gills with the hookless glochidia of the genus Lampsilis. The disappear- 
 ance of the hooked glochidia of Anodonta and Symphynota from the gills of the carp 
 may be due to the pulling away of these large and heavy glochidia from the delicate 
 gill filaments, as suggested in our consideration of the survival of the two types of 
 glochidia upon fins and gills, respectively. The disappearance of the hookless glochidia 
 of Lampsilis from both gills and fins of the carp can not be explained in this manner; 
 it suggests rather that there may be some reaction of the host's tissues comparable to the 
 processes which confer immunity against parasitic bacteria in higher vertebrates. With 
 minnows (Notropis cayaga and N. lutrensis) 2 to 4 inches in length, we have not been 
 able to secure any considerable infection with the glochidia of Symphynota complanata, 
 for, although they will attach in large numbers during infection, they all drop from the 
 fins and gills within a few days. The fins of these minnows are much more delicate than 
 those of the carp, and the explanation is perhaps that so large a glochidium is easily 
 torn away; but the large-mouth black bass has hardly a delicate fin, and for this fish we 
 have records of infections where no glochidia of 5. complanata became attached during 
 an exposure sufficient for the attachment of many to the gills. In this latter case, the 
 extreme activity of the fish must be considered as a factor which might keep the hooked 
 glochidia from attachment to the fins. 
 
 Darters (Etheostoma cozrideum spectabile) i^to2 inches in length can not be infected 
 successfully with the glochidia of Lampsilis, for although they may fasten so thickly to 
 the fins that many fish die during the first day after their exposure, the surviving fish 
 will slough off considerable portions of the fins and within a week show only the healed 
 and regenerating parts as an indication of their recent experience. The gill slits were so 
 small in these fish that only an occasional glochidium was found upon them. 
 
 Such cases as these are of great importance and should be followed up to determine 
 whether the simple mechanical conditions like over-infection, delicacy of fin, or con- 
 figuration of the mouth parts can give a satisfactory explanation ; or whether the histo- 
 logical changes of which the fish is capable, under stimulation by the glochidium, must be 
 regarded as the cause of its immunity. We have not carried out a sufficient number of 
 experiments to feel sure that the simpler explanations can be excluded. In any case, 
 it is interesting that fish like the minnows and darters, which live close to the bottom, 
 are not likely to become heavily infected by some of our most common glochidia. 
 
 BEHAVIOR OF FISHES DURING INFECTION. 
 
 The behavior of the fish during infection is a matter of some importance and has been 
 already mentioned in an incidental manner. The rock bass, large-mouth black bass, 
 and blue-gill sunfish, which are very active and which consequently exhibit powerful 
 respiratory movements, are well adapted to artificial infection, and the proper suspen- 
 sion of the glochidia in the water is secured by the movements of the fish alone. The 
 crappie, which are sluggish and easily killed by handling, require some special device to 
 
1 64 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 insure the optimum infection and are not well suited for work on a large scale because 
 of their behavior during infection. Fish which rest upon the bottom are sometimes not 
 so favorable as they might seem because they do not move about enough to keep the 
 glochidia in motion. While other features may be of greater importance, the behavior 
 of the fish as affecting the distribution of the glochidia in the water should always be 
 considered in deciding how useful any fish may be for purposes of infection. 
 
 INFECTION OF FISH IN LARGE NUMBERS. 
 
 The infection of fish in large numbers has been attempted with a view to determining 
 the feasibility of extending the methods described above to wholesale infections of fish 
 in a hatchery. As a result of two such attempts, we have no doubt that the successful 
 development of the methods needed for infection in connection with the artificial propa- 
 gation of mussels is only a matter of a little study in a properly equipped station. In 
 December, 1907, about 25,000 small fish, under 6 inches in length, were placed at our 
 disposal at the substation of the Bureau at La Crosse, Wis., and we were able on this 
 occasion to infect by wholesale methods about 12,000 blue-gill sunfish, 3,700 yellow 
 perch, 7,000 catfish, 2,000 crappie, 150 rock bass, 150 carp, and 100 roach. The greater 
 number of these fish were infected with the glochidia of Lampsilis ligamentina , and, 
 considering the fact that this was our first experience with so large a number of fish, the 
 results were satisfactory. Smaller lots were infected with the glochidia of L. anodon- 
 toides and L. recta, the results giving every indication that these two species are essen- 
 tially like L. ligamentina in the conditions of their development. The most successful 
 infections were obtained by placing from 100 to 200 fish in a common galvanized iron 
 washtub about two-thirds full of water. It was found that by adding to this body of 
 water the glochidia obtained from two or three specimens of Lampsilis, and, when it 
 seemed necessary, stirring the water by hand, tolerably constant results could be secured. 
 Our difficulties were with over- rather than with under-infection. It was also possible 
 to use the same tub a number of times without changing the water or adding to the stock 
 of glochidia. Infection was also attempted by lowering the water in the large retaining 
 tanks of the station to a depth of 4 inches and confining the whole number of fish which 
 had been held in the full tank to this much smaller body of water. This method was 
 found, in the absence of any attempt to keep the glochidia properly distributed through 
 the water, quite inadequate and it became necessary to reinfect these fish in the tubs. 
 
 The mortality of the fish in these experiments was decidedly in excess of what one 
 might expect for uninfected fish kept under similar conditions, a result clearly due to 
 the over-infection which is the one thing most to be guarded against. At the end of six 
 weeks some of the remaining fish were liberated in the west channel of the Mississippi 
 River at La Crosse, a locality which we then believed might be suitable for this species 
 of Lampsilis. 
 
 These infections were made under conditions of limited time and equipment and 
 were wholly tentative, the aim being to make a test of our methods on a large scale. 
 We revisited La Crosse a month after the infection, making careful examinations of the 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 1 65 
 
 fish and by shipping several hundred to Columbia were able to follow the development 
 of the glochidia under the conditions in our laboratory. The results were probably as 
 favorable as could have been expected under the circumstances. 
 
 In December 1908 a similar infection was attempted with about 6,200 large-mouth 
 black bass and 3,800 crappie in the station of the Bureau at Manchester, Iowa. Upon 
 this occasion the glochidia of Lampsilis ligamentina were again used in a majority of the 
 infections, similar results being obtained with L. anodontoides, recta, and ventricosa, which 
 were used for the minor infections. The black bass took the glochidia very readily and, 
 having had only a limited experience with this species of fish, we gave them an amount 
 of infection equal to that which had been carried successfully by the rock bass infected 
 at La Crosse in the previous experiments. The infection was estimated at from 2,000 
 to 2,500 glochidia to a fish 4 or 5 inches in length. This proved entirely too heavy for 
 the large-mouth black bass and the mortality among them amounted to about 55 per 
 cent in the 30 days they were under observation. By the third day after the infection 
 the hypertrophy of the gill tissue was so great as to be at once noticeable to the eye, and 
 this was clearly the cause of death. An infection of not more than 1,000 glochidia per 
 fish would have been more nearly the optimum load. 
 
 The crappie did not take the infection well despite longer exposure, the reason for 
 this being the size of their gill slits and their behavior as already discussed, and we do 
 not consider small fish of this species favorable for infection with any of the glochidia 
 from mussels which are of commercial importance. 
 
 Thirty days after these infections the surviving fish were liberated in the Maquoketa 
 River near Manchester, in a situation where the conditions were favorable for mussels 
 and where the presence of a dam below the point of liberation, together with the absence 
 of mussels of this species, made it seem possible that at some later period their appear- 
 ance in this locality might be traced to this experiment. We have never made any sub- 
 sequent examination of this stretch of the river with this in view, a thing which should 
 be done by one of the parties engaged in the field work of the mussel investigation. 
 
 These two experiments in the wholesale infection of fish, while disappointing in 
 some respects, give no indication of any insurmountable difficulties. It is fair to con- 
 clude that a little experimentation under hatchery conditions will make it as easy to 
 carry the glochidia through their metamorphosis in large numbers as we have found it 
 in small lots of fish kept in aquaria. The high mortality of the fish, being so clearly a 
 matter of over-infection, is a thing which can be guarded against without reducing too 
 greatly the load of glochidia which the fish may carry. It is then only a matter of dis- 
 covering the most suitable species of fish and finding out how best to handle them in 
 large numbers. 
 
 One thing which seems necessary for the rapid and uniform infection of fish in large 
 numbers is a device which will bring about a uniform distribution of the glochidia in 
 the water during the whole period of the fishes' exposure. Without something of the 
 sort it will hardly be possible to handle large numbers of fish with constant and uni- 
 form results. We have tried, though not very extensively, two means of effecting 
 
1 66 
 
 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 this. The first consisted of a two-bladed propeller fastened in the middle of the bottom 
 of a tub and rotated slowly, there being enough space in the water above the blades to 
 allow the fish room to escape the stroke. This device was not very satisfactory, but 
 as it was operated by hand and the blades roughly constructed, effective use might be 
 made of a more carefully adjusted mechanism of this type. A second and more 
 promising device consists of a branched system of iron pipes bored with many small 
 holes (text fig. 3), through which fine jets of water are forced out at the bottom of a 
 tank. The amount of pressure in these fine jets can be easily regulated from the main 
 supply pipe, and the height to which the glochidia will be driven from the bottom is 
 thus controlled. The tank may be allowed to overflow at the top and the glochidia 
 
 -<r-^S 
 
 Fig. 3. — Apparatus for keeping glochidia suspended in water while fish are being exposed to them for gill-infections. Tap 
 water entering at S issues in fine jets through the very small holes placed along the top and sides of the pipes on the bottom 
 of the aquarium, and an even distribution of glochidia throughout the water is thereby maintained. By regulating the 
 force of the water entering the pipes at S the glochidia are prevented from rising to the top of the aquarium and escaping 
 with the overflow. 
 
 prevented from being carried off in the overflow by so adjusting the force of the jets 
 that the glochidia will not rise quite to the surface. This device keeps the glochidia 
 suspended in a very uniform way, and it may prove to be just what is needed for the 
 uniform infection of large numbers of fish. 
 
 CONDITIONS NECESSARY FOR SUCCESSFUL INFECTION. 
 
 Three factors should be considered in attempting the infection of any species of 
 fish with glochidia, namely, the uniform suspension of the glochidia in the water, the 
 reaction of the glochidia when stimulated by mechanical or chemical contact with the 
 fish, and the reaction of the fish's tissues after the glochidium has become attached. 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 1 67 
 
 In any attempted infection of fish in large numbers, careful tests should first be made 
 upon a few fish in small dishes, with microscopic examination of the infected parts 
 from fish killed during the time of infection and for several days following, or until it 
 is clear that the glochidia have become safely established in their host's tissues. After 
 even limited experience one learns approximately the number of glochidia needed and 
 can determine roughly their suspension in the water by taking samples at random in a 
 pipette, which when held against the light shows clearly the individual glochidia. Dur- 
 ing infection it is possible to pick out individual specimens and by lifting up the oper- 
 culum of the living fish, examine the gills with a hand lens. The glochidia are then 
 seen individually and the progress of the infection can be watched. Fin-infecting glo- 
 chidia may be seen individually if a fish is placed in a small dish against a black back- 
 ground. 
 
 It is not difficult to determine by these means the optimum time for the exposure. 
 When 100 fish 5 to 6 inches in length are taken and the contents of a single marsupium 
 of a large Lampsilis is placed in an ordinary washtub, infections may be obtained some- 
 what as follows: Rock bass, exposed 30 to 40 minutes, 2,000 to 2,500 glochidia on gills 
 of each fish; large-mouth black bass, exposed 15 to 20 minutes, 500 to 1,000 glochidia on 
 gills ; crappie, exposed 20 to 30 minutes, 200 to 400 glochidia on gills ; yellow perch, exposed 
 20 minutes, 400 to 600 on gills; German carp (with Anodonta), exposed 30 to 40 min- 
 utes, 200 to 500 on fins. These figures are given as starting points for anyone attempt- 
 ing artificial infections and can not be taken as representing the results of precise deter- 
 minations of optimum infections for the fish in question, because the means for deter- 
 mining the numbers and distribution of the glochidia have been only approximate. 
 It will probably always be necessary, in the practice of artificial infection on a large scale, 
 to have the fish examined microscopically by a properly trained observer, and this 
 will be particularly true in the beginning of this work in hatching establishments, 
 because the practical details of artificial infection on a large scale have yet to be solved. 
 
 DURATION OF THE PARASITIC PERIOD. 
 
 According to the experience of previous observers, the duration of the parasitic 
 period varies inversely with the temperature of the water (Schierholz, 1888; Harms, 
 1907-1909). Although we have found this to be true in general, our experiments have 
 not shown so definite a relation between temperature and parasitism as has been 
 described by Harms, for example, and it is quite possible that other factors, which are 
 obscure, exert a modifying influence upon the length of time the glochidia remain on 
 the fish. Harms found that the glochidia of Anodonta completed the metamorphosis in 
 80 days at a temperature of 8° to io° C; in 21 days at 16 to 18 ; and in 12 days at 20 ; 
 while in the case of the hookless glochidia of Unio (which are gill parasites) the period 
 was 26 to 28 days at a temperature of 16 to 17°. He is inclined to attribute the some- 
 what longer time required for the metamorphosis of Unio to the fact that the glochidia 
 in this genus when discharged are in a less advanced stage of development than are 
 those of Anodonta — a difference that exists between all hookless and hooked glochidia. 
 i%7*3°— 12 5 
 
1 68 
 
 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 A few typical cases, selected from our records of infections are given in the accom- 
 panying table, which illustrates the far greater variability in the parasitic period than 
 that observed by Harms. 
 
 Table Showing Infections with Glochidia. 
 
 Experi- 
 ment. 
 
 Date. 
 
 Mussel. 
 
 Fish. 
 
 Expos- 
 
 Young 
 
 mussels 
 liberated. 
 
 Dura- 
 tion of 
 parasit- 
 ism. 
 
 Av. temp, 
 during 
 parasit- 
 
 HOOKED 
 GLOCHIDIA. 
 
 Dec. 3, 1909 
 Dec. 17, 1909 
 Jan. 7, 1910 
 Apr. 5,1910 
 
 Feb. 19, 1910 
 Mar. 6, 1909 
 
 Apr. 8, 1909 
 
 Apr. 13,1910 
 May 2, 1910 
 
 May 3, 1910 
 July 29. 1909 
 Aug. 5, 1908 
 
 Symphynota compla- 
 nata. 
 
 Apomotis cyanellus 
 
 do 
 
 Mm. 
 
 is 
 15 
 
 3° 
 
 9 
 10-15 
 
 10-15 
 
 8-15 
 
 50 
 
 7-14 
 
 30 
 
 Dec. 17-19 , , 
 
 Jan. 1-4 
 
 Jan. 18-21 
 
 Apr. 14-18. . . . 
 
 Mar. 5-12 
 
 Apr. 7-1 1 
 
 Apr. 27-May 1 
 
 May 2-8 
 
 May 15-26 
 
 May 17-25 
 
 Aug. 12-14 ■ ■ 
 Aug. 17 
 
 Days. 
 14-16 
 
 15-18 
 
 11-14 
 
 9-13 
 
 32-36 
 
 19-23 
 
 19-25 
 13-24 
 
 14-22 
 14-16 
 
 °C. 
 
 16.0 
 
 
 
 
 
 Pomoxis annularis. 
 
 Apomotis cyanellus 
 
 Pomoxis annularis. 
 Apomotis cyanellus 
 
 Apomotis cyanellus 
 
 
 
 
 IT S 
 
 BOOKLESS 
 GLOCHIDIA. 
 
 Lampsilis ligamentina . - 
 
 I7.8 
 
 
 
 
 do 
 
 Micropterus salmoides. 
 
 Apomotis cyanellus 
 
 Micropterus salmoides. 
 
 Apomotis cyanellus 
 
 .do 
 
 Micropterus salmoides. 
 Apomotis cyanellus 
 
 
 
 Lampsilis subrostrata. . . 
 Lampsilis ligamentina 
 
 Lampsilis subrostrata. . . 
 
 Unio complanatus 
 
 Quadrula plicata 
 
 18. I 
 
 
 18. I 
 
 
 18. 1 
 
 
 
 23 
 
 Micropterus salmoides . . 
 
 24.4 
 
 
 
 In the case of Symphynota complanata, which has hooked glochidia essentially like 
 those of Anodonta, the period varied from 9 to 18 days at average temperatures of 17. 8° 
 to 1 6° C, as compared with Harms's 21 days at practically the same temperature. 
 At lower temperatures, about io°, we have recorded a period of 74 days for 5. costata. 
 
 The absence of a close correspondence between the temperature and the duration 
 of the parasitism has been much more conspicuous in the case of hookless glochidia, which 
 have shown not only a remarkable range in the period but a considerable irregularity in dif- 
 ferent experiments made at about the same temperature. The shortest period recorded 
 by us was seven days in an infection of black bass with the glochidia of Lampsilis sub- 
 rostrata and L. recta in April when the average temperature during the parasitism was 
 20. 5 , but this unusual time was only observed in this one instance. A still more 
 remarkable case, but at the opposite extreme, was an infection of black bass and crappie 
 with the glochidia of L. ligamentina and L. recta which remained on the fish for 13 to 16 
 weeks. The infection was made in November and the young mussels were liberated 
 during a period of about three weeks in the following February and March; during the 
 parasitism the temperature varied from about 16 to 18. ° The cause of the extreme 
 duration in this case is not known, for in no other experiment at the same temperature 
 has the parasitism lasted for more than 25 days. 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 1 69 
 
 As may be seen in the table, with hookless glochidia (aside from the extreme cases 
 mentioned) the variation in the period has been from 12 to 36 days at average tempera- 
 tures ranging from 24. 4 to 17.8 ; but even at practically the same temperature the 
 difference may be quite marked, as in experiments no. 8 and no. 9. Experiment no. 6 
 should be noticed as being a case in which, contrary to expectation, quite a long period 
 (32 to 36 days) was recorded at 19.1 , whereas in other experiments (no. 5 for example) 
 the time was only 14 to 21 days at the lower temperature of 17. 8°. 
 
 It would seem clear that, although within certain wide limits the duration of the 
 parasitism is dependent upon the temperature of the water, nevertheless other factors 
 may enter into the case to either accelerate the metamorphosis or prolong it over a period 
 which is much longer than the usual duration of the parasitism. These factors would 
 seem to be associated with individual physiological differences in the interaction between 
 the fish and the parasite and are probably nutritive in nature, for on one and the same fish 
 some glochidia may remain several days longer than others. 
 
 As may be seen from an examination of the table, in which the period of liberation 
 is given in each experiment, not all of the young mussels leave the fish at the same time, 
 but, on the contrary, the liberation may occupy a week or more. Harms found that it 
 required from 5 to 6 days, the greater number leaving the fish during the middle of the 
 period. Our experience has usually been in accord with these observations, but we have 
 found the period to be somewhat more variable, from 2 to 1 1 days, or even much longer. 
 
 IMPLANTATION AND CYST FORMATION. 
 
 As has been described, the glochidium attaches itself to the fish by closing its shell 
 firmly over some projecting region which can be grasped between the valves, like the 
 free border of a fin or a gill filament. In so doing, a portion of the epithelium and 
 underlying tissue, including blood vessels and lymphatics and varying in amount with 
 the extent of the "bite," becomes inclosed within the mantle space of the glochidium. 
 This tissue early disintegrates into its cellular constituents, which are taken up by the 
 pseudopodial processes of the larval mantle cells, and, as Faussek (1895) has described, 
 are utilized as food during the early stages of metamorphosis. In figure 60, plate xv, 
 drawn from a glochidium six* hours after attachment to a fin, the disintegrated tissue, 
 consisting of loose epithelial cells, blood corpuscles, and fibers which lie scattered in the 
 mantle cavity, is seen in the process of being ingested by the mantle cells. Figure 61, 
 plate xv, shows a later stage, 24 hours after attachment, in which the detritus has been 
 entirely taken up, and the mantle cells are now heavily charged with food material. 
 
 Almost immediately after attachment proliferation of the epithelium begins as the 
 initial step in the formation of the cyst which eventually incloses the entire glochidium. 
 The overgrowth of the larva has been described by Faussek (1895) and Harms (1907-1909) 
 as a healing process on the part of the fish's tissues, resulting from the irritation caused by 
 the wound. The proliferation starts around the line of constriction produced by the 
 pressure of the edges of the valves on the epithelium, and, since the glochidium lies 
 between and prevents the immediate closure of the lips of the wound, the extending 
 
170 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 epithelium is forced to slide up over the surface of the shell on all sides, until the free 
 margins meet and fuse over the back of the larva, as may be understood by reference 
 to figures 59 to 61, plate xv, and 35 to 38, plate xi. 
 
 So rapid is the overgrowth, especially in the case of implantation on the gills, that 
 it would seem that something more than the mere mechanical irritation produced by the 
 glochidium is concerned in causing the proliferation of the epithelium. We have, 
 therefore, carried out a series of experiments with a view to determining whether or not 
 a chemical stimulus is provided by the larva, and by using various methods have studied 
 the action of glochidial extracts on the epithelium of both fins and gills. The results 
 have been entirely negative, although the question has by no means been settled by the 
 experiments which have been thus far attempted. By further improvements in the 
 technique, some of the difficulties involved in the investigation, which is still in progress, 
 may be overcome. 
 
 The process of implantation and cyst formation may be readily observed on the fila- 
 ments of an excised gill, which under favorable conditions will live long enough in a 
 dish of water to enable one to see the glochidium completely covered by the proliferated 
 epithelium. Figure 54, plate xm, drawn from the living excised gill, shows the distal 
 end of a single filament bearing a glochidium of Unio complanatus which has become 
 nearly covered by the walls of the cyst. In this case the gill was cut from the fish two 
 hours after the infection and the drawing was made an hour later; immediately after 
 the excision of the gill this particular glochidium was hardly half covered. The same 
 glochidium was kept under observation, and two hours later (five hours after the infec- 
 tion) the sketch was made which is reproduced in figure 55, plate xiii. By this time 
 the cyst, which is seen to have very thick walls, was completed, and formed a prominent 
 mass near the end of the filament. Shortly afterwards the tissues of the gill began to 
 disintegrate, but for at least three hours they remained alive and the proliferation of the 
 epithelial cells proceeded rapidly, the entire process of cyst formation taking place in a 
 perfectly normal manner. 
 
 The histological changes which the epithelium undergoes in the formation of the 
 cyst have been studied in this laboratory by Miss Daisy Young, and, as her results will 
 soon be published in detail, only a brief reference will be made in this place to the 
 essential points involved in the cellular changes occurring during implantation of the 
 glochidium. 
 
 Figure 59, plate xv, shows a very early stage, 15 minutes after attachment, in the 
 formation of the cyst on the fin of a fish which had been infected with the glochidia of 
 Symphynota complanata. The section is taken transversely through the glochidium 
 and the free border of the fin on which the parasite has a firm grip. The mass of 
 tissue, consisting of epithelial cells, connective tissue, and blood vessels in the mantle 
 chamber of the glochidium, is the edge of the fin which was inclosed between the valves 
 when attachment was effected. Already the proliferation of the epithelium is beginning 
 in the neighborhood of the constriction, where two mitoses may be seen on the right in 
 the figure. At the edges of the wound caused by the closure of the shell some of the 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. I 71 
 
 epithelial cells are undergoing degeneration, while on the left of the section quite a patch 
 of these cells is sloughing off, a not infrequent occurrence. The region of most active 
 growth and multiplication of cells is just below the line of constriction, and, as the cells 
 at this level increase in number, they appear to push those lying above them up over the 
 outside of the shell, so that the actual covering of the glochidium is due largely to this 
 mechanical gliding of the epithelium over its surface. Sections give no conclusive evi- 
 dence of amitotic division, while mitoses are generally abundant in the region of active 
 proliferation. An intermediate step in the process of implantation is illustrated in 
 figure 60, plate xv, less highly magnified than the last figure, which shows a glochidium 
 about half covered in six hours after attachment. The free edges of the cyst wall even- 
 tually meet over the dorsal side of the glochidium, where they then fuse. Figure 61, 
 plate xv, shows a case of complete implantation on a fin at the end of 24 hours; now the 
 epithelial covering is continuous and the glochidium entirely inclosed. The wall of the 
 cyst is seen at this time to be quite thick, but it usually becomes thinner later on as the 
 cells composing it flatten down. In the last two figures the mantle cells of the larva 
 clearly show epithelial nuclei and cell detritus which have been ingested. 
 
 In figures 62 and 63, plate xv, two stages are represented in the formation of the 
 cyst on gill filaments, taken at one hour and three hours, respectively, after attachment. 
 The glochidia are those of Lampsilis ligamentina . In figure 62, plate xv, the prolifera- 
 tion has made some progress, especially on one side, and three or four mitotic figures are 
 seen just below the glochidium and near the raw edge of the constricted epithelium. 
 A large mass of the tissues of the filament is also shown in the figure inclosed within the 
 mantle chamber of the glochidium. Figure 63, plate xv, represents a stage when the 
 process is nearly completed and the edges of the epithelial covering have met but not 
 yet quite fused. The cyst wall in this case is much thinner than that shown in figure 
 61, plate xv, but its thickness is quite variable. 
 
 In about one week after attachment, as a rule, the wall of the cyst begins to assume 
 a looser texture, the intercellular spaces becoming infiltrated with lymph, and from 
 this time on to the end of the parasitic period there is little further change in its 
 structure. 
 
 Before liberation of the young mussel, the valves open from time to time and the 
 foot is extended. By the movements of the latter the cyst is eventually ruptured, its 
 walls gradually slough away, and the mussel thus freed falls to the bottom. 
 
 Portions of the wall of the cyst often adhere to the shell after liberation, while, if 
 the young mussel has hooks, it may hang for a time by shreds of the fin in which the hooks 
 are embedded, as seen in figure 24, plate ix. 
 
 METAMORPHOSIS WITHOUT PARASITISM IN STROPHITUS. 
 
 In a brief paper (191 1) we have recently announced the discovery that in the genus 
 Sirophitus Rafinesque the metamorphosis takes place in the entire absence of parasitism, 
 and, since the life history of this form is without a parallel in the Unionida?, so far as is 
 known, reference may be made again to the interesting conditions which obtain in its 
 development. 
 
172 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 It has been known for a long time that in Strophitus the embryos and glochidia 
 are embedded in short cylindrical cords which are composed of a semitranslucent, 
 gelatinous substance, and that these cords, which are closely packed together, like chalk 
 crayons in a box, lie transversely in the water tubes of the marsupium. The blunt ends 
 of the cords are seen through the thin lamella of the outer gill, which in this genus, as 
 in Anodonta and others, constitutes the marsupium. The position of the masses of 
 embryos, while contained within the gill, is so unusual that Simpson in his "Synopsis 
 of the Naiades" established a special group, the Diagenas, for Strophitus — the only 
 genus of the family in which this peculiarity exists. In other genera the embryos are 
 conglutinated more or less closely to form flat plates or cylindrical masses, each one of 
 which is contained in a separate water tube and lies vertically in the marsupium. 
 
 So far as we are aware, Isaac Lea (1838) was the first to observe this interesting 
 arrangement which he described and figured, rather crudely to be sure, in Strophitus 
 undulatus (Anodonta undulata). In several subsequent communications (1858, 1863) 
 he added further details and illustrations, and also mentioned the occurrence of the 
 transversely placed cords, or "sacks," as he called them, in S. edentulus. He recorded 
 the former species as being gravid from September until March, and described the 
 extrusion of the cords from the female, as well as the remarkable emergence of the 
 glochidia from the interior of the cords after the latter have been discharged. 
 
 The sacks were discharged into the water by the parent from day to day, for about a month in 
 the middle of winter. Eight or ten young were generally in each sack, but some were so short as 
 only to have room for one or two. Immediately when the sacks came out from between the valves of 
 the parent, most of the young were seen to be attached by the dorsal margin to the outer portion of the 
 sack, as if it were a placenta. 
 
 The essential points in these observations have since been verified by other inves- 
 tigators. Sterki (1898), following the suggestion of Lea, has called the cords, which 
 differ strikingly from the conglutinated masses of Unio and other genera, "placentae," 
 thus indicating that he considered them to have a nutritive function. He also described 
 the extrusion of the glochidia, when placed in water, and their attachment to the cord 
 "by a short byssus thread whose proximal end is attached to the soft parts of the 
 young." He further states that the glochidia are inclosed in the placentae when the 
 latter are first discharged, and that after their extrusion they remain attached for some 
 time. 
 
 Strophitus edentulus, which Ortmann (1909) regards as identical with undulatus, is 
 a rare species in all of the localities in which we have collected mussels, and, until 
 recently, our only observations on this form were made upon a few gravid individuals 
 which were taken in the Mississippi River near La Crosse, Wis., during the summer of 
 1908. Mention has already been made of our records with reference to the breeding 
 season of Strophitus. 
 
 After verifying the main observations of Lea and Sterki, so far as was possible at 
 that season of the year, we examined the glochidia carefully with a view to determining 
 whether their subsequent life history would exhibit any peculiarities, as might be sus- 
 pected from their relation to the cords. At that time we did not observe the normal 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 1 73 
 
 discharge of the cords by the female; but we removed them from the marsupium, placed 
 them in water, and, after the glochidia had emerged (fig. 46, pi. xn), employed various 
 means to bring about their attachment to fish. None of these attempts, however, was 
 successful, although the fish were left in small dishes containing many cords for as long 
 a time as 12 hours. In the light of these results, which indicated the inability of this 
 glochidium to attach itself to fish, and in view of the fact that the cords so evidently 
 seemed to be a nutritive device, we felt it to be highly probable that in this species the 
 metamorphosis would be found to occur in the absence of parasitism — a prediction 
 which has been recently verified. 
 
 On February 6, 191 1, a single female of Strophitus edentulus, which had been kept 
 in the laboratory since the preceding November, was seen discharging its cords from 
 the exhalent siphon. The discharge continued until March 25, and during that time 
 the cords were thrown out in varying numbers from day to day. They measured from 
 2 to 10 mm. in length and about 1 mm. in diameter, although they became more or 
 less swollen after lying in the water for a time. Each cord contained from 10 to 24 
 glochidia arranged in an irregular row. In many cases the glochidia emerged from the 
 cords in a few minutes after the latter were discharged, and then usually remained 
 attached by the thread in essentially the same manner as has been described by Lea 
 and Sterki (fig. 46, pi. xn). The thread, which is apparently a modified larval thread, 
 is continuous at its distal end with the egg membrane, which generally remains embedded 
 in the cord; so intimate, in fact, is the union between the two that at times the mem- 
 brane, adhering to the thread, is dragged out of the cord when the glochidium is 
 extruded, in which case, of course, the glochidium becomes entirely detached from the 
 cord. 
 
 All attempts to infect fish with these fully formed glochidia were again unsuccessful, 
 even when the exposure was of long duration. Within a few days the extruded glochidia 
 died in spite of every effort to provide the most favorable conditions for their maintenance. 
 
 When the cords first began to be discharged, one of our students, Miss Daisy Young, 
 happened to notice that not all of the larvae were extruded, and that among those which 
 remained in the cords some had lost the larval adductor muscle, possessed a protrusible 
 foot, and showed other signs of having undergone the metamorphosis. Upon careful 
 examination this was found to be true, and it was discovered that these young mussels — 
 for such they undoubtedly are — are subsequently liberated by the disintegration of the 
 cord after having passed through the metamorphosis in the entire absence of a parasitic 
 period. We, therefore, have concluded that the emergence from the cords in the glo- 
 chidial stage is premature, due possibly to some change which has taken place in the 
 gelatinous substance surrounding them as a result of free contact with the water, or to 
 release from the pressure to which they are subjected while in the marsupium. It is 
 perfectly evident that these glochidia neither become attached to fish nor undergo any 
 further development; they have simply come out too soon and are lost. 
 
 The young mussels, on the other hand, which have developed inside the cords, when 
 liberated by the disintegration of the latter or removed directly by teasing, are found to 
 
174 BULLETIN OP THE BUREAU OF FISHERIES. 
 
 have reached as advanced a stage of development as is attained by any unionid at the 
 time it leaves the fish. They closely resemble the young of Anodonta at the close of the 
 parasitic period, and upon examination have been found to possess the following struc- 
 tures: The anteriorandposterioradductor muscles; the ciliated foot; two gill buds on each 
 side; a completely differentiated digestive tract, including mouth, esophagus, stomach 
 intestine, and anus ; liver ; the cerebral, pedal, and visceral ganglia ; otocysts ; the rudiments 
 of the kidneys, heart, and pericardium; while they also show a slight growth of the per- 
 manent shell around the margin of the shell of the glochidium (fig. 45, pi. xii). The larval 
 muscle has completely disappeared, although some of the mantle cells of the glochidium, 
 as well as the hooks of the shell, are still present. They crawl slowly on the bottom of the 
 dish by the characteristic jerking movements of the foot, after the manner of the young 
 of other species at a corresponding stage, although the valves of the shell gape more widely 
 apart and the foot is shorter and less extensible. We have not succeeded as yet in keep- 
 ing them alive for more than 10 days, but it is difficult in the case of any species to main- 
 tain young mussels of this age under laboratory conditions. 
 
 One of these young mussels after removal from the cord is shown in figure 45, plate 
 xii, in which many of the organs of the adult or their rudiments are clearly indicated. 
 A comparison will show that it is essentially as advanced in its development as the young 
 of Anodonta when it is liberated from the fish (cf. Harms's figures, 1909, and also our fig. 
 47, pi. xii, of Syinphynota coslata). 
 
 The conclusion is inevitable that we have here to do with a species which has no 
 parasitism in its life history, although the presence of hooks and other typical glochidial 
 structures would indicate that it has originated from ancestors which possessed the para- 
 sitic stage like other fresh-water mussels. The cord is undoubtedly to be interpreted as a 
 nutritive adaptation which arises in the marsupium during the early stages of gravidity, 
 since the young embryos are at first contained in an unformed viscid matrix and the cords 
 are a later product. 
 
 The whole history of this exceptional species warrants a more detailed study, and 
 Miss Young is now engaged in such an investigation. When her work is completed we 
 hope that it may include the entire course of development, the method of formation of 
 the cords, and the rearing of the young mussels during a much longer period than has thus 
 far been possible. 
 
 V. ATTEMPT TO REAR GLOCHIDIA IN CULTURE MEDIA. 
 
 Since the relation of the glochidium to the fish is essentially a nutritive one, it 
 seemed to us that it should be possible to rear the larvae through the metamorphosis 
 artificially, provided a suitable nutritive medium could be found, and accordingly a 
 series of experiments, with this object in view, were undertaken at our suggestion by one 
 of our students, Mr. L. E. Thatcher. Although the result has thus far been entirely 
 negative, we have not despaired of ultimate success, and, since the experiments are to be 
 continued, a brief mention of the methods employed may be made in this place. 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 1 75 
 
 It was natural to suppose that the blood of the fish would offer the most favorable 
 nutritive conditions for the development of the glochidia, and hence it has been used in 
 most of the experiments, which, moreover, have been made in the spring, when the water 
 in the laboratory was comparatively warm and the metamorphosis, if it had occurred, 
 would have taken place as rapidly as possible. 
 
 The glochidia of Lamps His ligamenlina and L. subrostrata were carefully removed 
 from the marsupium with a sterilized pipette and then repeatedly washed in distilled water 
 in order to obtain them as free as possible from bacteria and other organisms. A drop of 
 blood was next taken from a fish's heart and placed on a cover glass and a few glochidia 
 immediately introduced into it. The cover glass was then inverted over a hollow slide 
 containing a moist piece of filter paper, and the chamber sealed with vaseline. Every 
 precaution was taken to avoid contamination by bacteria. As soon as the glochidia 
 came into contact with the blood, of course they snapped shut in the manner already 
 described and in doing so inclosed some of the corpuscles, which it was to be presumed 
 would be ingested by the mantle cells. Although in some cases bacteria and infusoria, 
 probably introduced with the glochidia, appeared, in a majority of the cases the cultures 
 remained free from foreign organisms. In the latter event the glochidia lived for a few 
 days, but finally died without showing any indication of further development. Experi- 
 ments were tried with the blood of the frog and of Necturus, and also with extracts of 
 fish's tissues, bouillon and other nutritive media. In all, however, the results were 
 negative. The failure may possibly have been due to insufficient aeration, and experi- 
 ments are now being devised in which oxygen is to be introduced into the moist chambers, 
 and it is hoped that we shall yet succeed in rearing the glochidia in nutritive media 
 through the metamorphosis. 
 
 VI. POST-LARVAL STAGES. 
 
 BEGINNING OF THE GROWTH PERIOD AND LIFE ON THE BOTTOM. 
 
 The changes occurring during the parasitism and by means of which the glochidium 
 becomes transformed into the young mussel, ready for life on the bottom, are more prop- 
 erly described by the term development than by the word growth. The latter process 
 becomes the conspicuous feature only when the miniature mussel has left the fish. From 
 this time onward there are very few changes to which the term development may be 
 strictly applied; for, with the exception of the outer gill, all the important organs of the 
 animal have been laid down and have assumed something of their definitive structure 
 (fig. 47, pi. xn). 
 
 As soon as they are liberated from the fish the young mussels become quite active 
 and move about on the bottom of a dish by means of the foot (fig. 18, pi. vm, and fig. 48, 
 pi. xn), securing a hold by flattening the ciliated distal end against the bottom, and then 
 drawing up the body after the characteristic fashion of lamellibranchs. In these move- 
 ments the cilia of the foot play an active part ; they beat vigorously while the foot is being 
 extended, and apparently are effective in part at least in causing the protrusion. When 
 
176 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 the foot reaches its limit of extension, the cilia stop abruptly and remain quiet while the 
 forward movement of the body is taking place, only to resume their activity when the 
 extension begins again. Figure 18, plate viii, furnishes an excellent illustration of 
 the various positions assumed as the young mussels crawl about in their twisting, jerking 
 movements, and also shows the extent to which the shell has grown beyond the limits of 
 the glochidial valves by the end of the first week of free life. 
 
 In the great majority of forms, as appears from the work of other investigators and 
 our own observations, the mussel leaves the fish with only a very narrow margin of adult 
 shell protruding beyond the glochidial outline. The shape is still that of the glochidium, 
 although all other resemblances to this larval stage have disappeared. In the larva of 
 Symphynota costata this margin of the adult shell is so narrow, even after some days 
 upon the bottom (fig. 47, pi. xii), as not to protrude beyond the glochidial outline when 
 the young mussel is slightly contracted. Exceptions to this supposedly universal con- 
 dition have been observed by Coker and Surber (191 1) in the young of Plagiola dona- 
 cijormis and Lampsilis (Proptera) Icevissima — forms in which there is a considerable 
 growth of the definitive shell and presumably of the other organs during the parasitic 
 period. These cases are unique so far as known, but in view of the small number of 
 species which have been observed at all during this period of their existence other such 
 exceptions may be looked for. No data bearing upon the duration or other conditions 
 of the parasitic life are given in the paper in question, since the material studied was 
 from the gills of a fish which had been preserved after its infection under natural 
 conditions. 
 
 These stages immediately following the parasitism and until the mussels are about 
 20 mm. in length are less known than any others. They have seldom been found by 
 collectors, and the reasons for this are made clear by the work of Isely (191 1), to which 
 we shall presently refer. Pfeiffer first observed and figured in 182 1 a small shell having 
 the glochidial outline still visible at its umbo, and other cases have been recorded, 
 notably by Schierholz (1888). Such specimens were taken from nature and not from 
 mussels artificially reared. Indeed, no one has yet succeeded in following individual 
 specimens for more than a few weeks beyond the beginning of life on the bottom. 
 Recently Harms (1907, 1908, and 1909) has obtained these stages, by rearing, more 
 extensively than his predecessors and has figured (1907a, p. 81 1) the young of Anodonta 
 with a very substantial increase in size at an age of six weeks after the parasitism, 
 beyond which they could not be reared because of their destruction by small Crustacea. 
 He concludes that the latter constitute a serious danger to the life of the young mussel. 
 
 In our own work repeated attempts have been made to rear these stages to a size 
 which can be more easily handled, but without success. Specimens of Symphynota costata 
 (fig. 47, pi. xii) and of Anodonta cataracta have been kept alive in small dishes containing 
 green plants for a period of from one to two weeks after they had left the fish, and 
 Lampsilis ligamenlina and subrostrata for a period of six weeks. Little or no growth 
 was observed after the first week. The two species of Lampsilis formed a conspicuous 
 border of new shell during the first few days of bottom life (fig. 18, pi. vm, and fig. 48, 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 1 77 
 
 pi. xii) and then ceased growing although they continued to move actively about. 
 This would indicate that the difficulty lies in the lack of a suitable food supply. Crus- 
 tacea were not observed to play an important role, though we do not doubt the cor- 
 rectness of Harms's observations in this respect. 
 
 Figures 18, plate vm, 47 and 48, plate xii, will illustrate the appearance of the 
 young mussels at this period and an examination of figure 47 will show how extensively 
 the organs of the future adult have been laid down. Nothing remains to suggest the 
 glochidium save the shell, and structure and habit alike indicate that the organism is 
 now ready for a life on the bottom essentially like that of the adult. 
 
 JUVENILE STAGES AND THE ORIGIN OF MUSSEL BEDS. 
 
 For the sake of completeness, we shall discuss briefly at this point the present state 
 of our knowledge regarding the stages between the one last mentioned and that repre- 
 sented by the young mussels over 20 mm. in length, which are often found upon the 
 natural beds. In common with the experience of other collectors, we have seldom 
 found mussels under 20 mm. It would therefore seem clear that these early stages 
 are not at all common in localities where the slightly later stages and the adults are 
 found. Isely (1911) has published a preliminary note upon his study of this "juvenile" 
 period. We shall refer to his results rather fully, since there are no other recorded 
 observations which deal with these stages save in the way of incidental reference to 
 single specimens. This author states the problem by saying (p. yj) that: "Much diffi- 
 culty was experienced in finding young mussels for study and experimentation. I have 
 collected many specimens from the size of a nickel (20 mm.) to a quarter (24 mm.), but 
 mussels under the size of a dime (17 mm.) have been rare." The latter he terms the 
 "early juvenile" stages, including in this "the period following the time when the 
 mussel completes the parasitic stage and leaves the fish to lead an independent life 
 until it is about 15 mm. in length. This would cover, in most species, approximately 
 the first year of independent existence. Other periods may be designated as later 
 juvenile and adult life." He then reports the finding of 32 specimens in this early 
 juvenile stage representing four genera and nine species, as follows: (1) Lampsilis 
 luteola, two; (2) Lampsilis fallaciosa, one; (3) Lampsilis parva, four; (4) Lampsilis 
 gracilis, three; (5) Plagiola elegans, one; (6) Plagiola donaciformis , sixteen; (7) Anodonta 
 imbecillis, two; (8) Ptychobranchus phaseolus, two; (9) unnamed species, one. 
 
 All these specimens were found in places where the water was fairly swift, from 
 1 to 2 feet in depth, and on a bottom of coarse gravel, the particles of which were 10 
 to 25 mm. in diameter. They were anchored by the threads of a byssus gland "strong 
 enough to support the mussel in a rapid current" and capable of sustaining "the weight 
 of a number of small pebbles without breaking." 
 
 Here then, as Isely concludes, we have the clue to the habits and ecology of these 
 so little-known stages. The finding of representatives from so many genera and species, 
 both heavy and light shelled, under identical environmental conditions and the presence 
 of the functional byssus in all cases is pretty good evidence that this is the normal 
 
178 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 condition for early juvenile life in a wide range of forms. It is, moreover, interesting 
 to find in the Unionidffi, as in many other lamellibranchs (e. g., Mya and Pecten) a 
 functional byssus in the early stages, though there is no such organ in the adult. 
 
 As these results are very important and of convenience for reference in this paper 
 we may here quote Isely's conclusions in full. 
 
 The facts noted above are closely related, not only to the ecology of the juvenile mussel, but also 
 to the ecology of the adult. 
 
 1. They indicate the conditions essential for the most successful growth and early development 
 of the Unionidae. This kind of an environment gives a constant supply of oxygen and sufficient food; 
 is frequented by suitable fish; is free from shifting sand and silt accumulation. Those mussels that 
 drop from the fish in these favorable situations develop in large numbers, while the less fortunate, that 
 drop in shifting sand and silt, die early. 
 
 2. In the study of the ecological factors that are inimical to mussel life more attention should be 
 given to the consideration of the juvenile habitat. Absence of gravel bars and stony situations may 
 sometimes explain the scarcity of the Unionida? in certain streams and lakes where frequently water 
 content has been thought the chief unfavorable factor. 
 
 3. It is a well-known fact that in many streams certain stretches of mud bottom are found loaded 
 with mussels, while other areas, in the same stream, equally favorable from the standpoint of the habitat 
 of the adult mussels, have only scattering specimens. 
 
 This distribution of the adults may be explained by the assumption (which is fairly well established 
 by experimental study and will be discussed in a later paper) that the average mussel seldom travels far 
 up or down the stream from the place where it begins successful development. Stretches favorable for 
 juvenile development thus come to be the centers of dispersal in the streams where they occur. As 
 a result, areas of mud bottom near these favorable habitats become loaded with mussels by migration. 
 
 4. In the study of the life history of the Unionids we may consider the embryonic, the glochidial, 
 the parasitic, the early juvenile, and the adult as distinct periods for separate and special study. 
 
 These results of Isely's are clearly of very great importance in the problem of arti- 
 ficial propagation and it is to be hoped that his observations may be greatly extended 
 in the near future. The number of different species which he has found is a most 
 promising sign that he is on the right track, and we may hope that we shall soon reach 
 a satisfactory understanding of this stage of the life cycle hitherto so little known. 
 
 At this point a word regarding the formation of beds may be opportune. It is a 
 familiar fact that many species are most likely to be found congregated in beds which 
 in some of the larger streams must have contained, before the shells came into commer- 
 cial use, numbers of mussels which are hardly conceivable. Elsewhere in the stream 
 the mussels are found scattered and wandering over the bottom. In the absence of any 
 indication that the individuals of a species are in some manner attracted to one another, 
 the simplest explanation of the formation of beds would be the same as that given in 
 other cases of this sort. The conditions of food supply, current, character of bottom, 
 etc., must differ considerably, and we may reasonably suppose that some places present 
 the optimum conditions over an extended area and that in such a place a bed may be 
 formed. As the mussels wander over the bottom they may by chance enter such an 
 area of optimum conditions and will then move about less actively or come to rest, 
 because in the absence of unfavorable conditions there is no stimulus to continued loco- 
 motion. The result is that individuals which enter are likely to remain and more keep 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OP FRESH-WATER MUSSELS. 1 79 
 
 coming in. This kind of an explanation has been offered, by the students of animal 
 behavior in recent years, to account for the formation of aggregates in a great variety of 
 the lower organisms ; and it appears the most reasonable one in such cases as the one in 
 hand, where there is no evidence that the gregariousness is due to a definite recognition 
 of the presence of other individuals. 
 
 RATE OF GROWTH. 
 
 It has been quite generally believed, by those investigators who have given their 
 attention to this matter, that the mussel shell grows during the warmer months of the 
 year and that in winter there is no appreciable addition to its margin. When growth 
 begins again in the spring, the winter's rest has left a mark which appears as a dark 
 line on light-colored shells or as a deeper groove in others where the color is not so con- 
 spicuous. Finer lines may be found between these rings of growth, but the latter, like 
 the rings of a tree, mark the years. It is certain that these more conspicuous lines or 
 "rings," as we may term them, indicate an alternation of growing and resting periods in 
 the formation of the shell. It is not entirely certain that a single growth period must 
 always correspond to a single year; for, when any lot of shells is carefully examined, 
 some will be found in which the "rings" are distinct and strongly suggestive of an annual 
 increment, while others of the same size may not show these rings in any such distinct 
 fashion, and one is forced to conclude either that the annual rings, if such they be, are 
 not always clearly to be seen or that some mussels may grow at a very different rate 
 from others. The examination of any considerable number of shells leads to the belief 
 that even if the annual-ring theory can be proved conclusively the rings are often not 
 sufficiently distinct from the intervening lines to give an unquestionable record of the 
 age. 
 
 Assuming that these rings, when clearly seen, do represent years, it would seem that 
 the shell grows very rapidly during the first few years of the mussel's life and after that 
 much more slowly. To judge from the lines alone, we should say that many of the large 
 Quadrula shells had reached one-half their size in ten or a dozen years and then taken 
 forty or fifty for the remainder, so closely set are their later rings of growth; and that 
 shells of these species can not reach the most desirable commercial size in a less period 
 than twenty or thirty years. Since these are regarded as the best of all button shells, 
 the outlook may seem discouraging, because, like hardwood timber, the best shells take 
 too long to grow. 
 
 The "ring theory" if proved would not, however, make the situation so discourag- 
 ing as might seem from the species of Quadrula; for we have in some members of the 
 genus Lampsilis shells which are almost if not equally desirable, and such evidence 
 as we have from the rings indicates that shells like these may reach a commercial size 
 in a very few years and that even forms like the quadrulas may become marketable 
 within a period of four or five years. 
 
 In a recent paper, Israel (191 1) has reported his conclusion that there is no winter- 
 rest period and that more than one ring may be formed in a single year. This statement 
 
l8o BULLETIN OP THE BUREAU OF FISHERIES. 
 
 is based upon the examination of the shell margin in mussels collected at various seasons 
 of the year and of mussels which had been placed in wire inclosures on the bottom of 
 the stream after having been accurately measured. The results from these plantings 
 were fragmentary because of the accidental destruction of most of the inclosures. In 
 one case, however, he found specimens which "when placed in the inclosure in August, 
 1909, and measuring 18 mm. in length, had reached, at the time of their examination in 
 June, 1910, a length of 26 mm." He reports that other similar investigations are in 
 progress, the results of which we shall await with interest. 
 
 Since no accurate observations on the rate of growth of fresh-water mussels have 
 ever been made, we have attempted to secure definite data bearing upon this problem. 
 The data obtained are derived from two entirely different lines of observation, as indi- 
 cated by the headings of the sections which follow, and although meager they show 
 that with better facilities it should not be difficult to follow individual mussels from the 
 juvenile to the adult stages, and thus to determine their rate of growth in an accurate 
 manner. 
 
 GROWTH OF MUSSELS IN WIRE CAGES. 
 
 While engaged in mussel investigations at La Crosse, Wis., during the summer of 
 1908, we collected a number of young clams (fig. 68, pi. xvii) belonging to 16 different 
 species, and after weighing and measuring them accurately they were distributed in wire 
 cages, which were then anchored by long wires in midstream to the piers of a bridge over 
 the west channel of the Mississippi River opposite La Crosse. One hundred and sixty- 
 three small mussels, belonging to the following genera and representing both thin and 
 thick shelled forms, were planted out in this manner: Alasmidonta, Anodonta, Lampsilis, 
 Obliquaria, Obovaria, Plagiola, Quadrula, and Unio. 
 
 Some of the cages contained only a single specimen of each species represented in it, 
 in which case an absolute identification would be possible, should the cage be recovered 
 later, while, if two or more individuals of a species were put in a cage together, only 
 specimens of practically the same size were selected. In the latter case it would of 
 course be impossible to subsequently distinguish an individual mussel, and only the 
 average rate of growth could be determined for the individuals present. It was assumed 
 that mussels of the same size and under the same conditions would grow at practically 
 the same rate. 
 
 These plantings were made at intervals from June 29 to August 10, 1908. An 
 opportunity did not present itself to make an attempt to recover the cages for over two 
 years, but in November, 1910, Dr. R. E. Coker, who knew of the experiment, made a' 
 search while on a visit to La Crosse and was fortunate enough to find 2 of the 11 cages 
 planted by us in 1908. One of the cages was deeply buried in the mud and all of the 
 mussels in it were dead; as they showed little or no growth, they were evidently killed 
 shortly after the planting. In the other cage, however, 6 living mussels were found, 
 as follows: 3 Lampsilis ventricosa, 1 Obovaria ellipsis, 1 Quadrula solida, 1 Anodonta 
 imbecillis . These 6 mussels, with the exception of the specimen of Obovaria ellipsis, 
 were readily referred to definite individuals as recorded at the time the cage was set out. 
 The comparative measurements and weights are given below. 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. l8l 
 
 June 29, 1908. November 15, 1910. 
 
 Lampsilis venlricosa: 
 
 (1) 45 by 30 mm., 16 grams 85 by 65 mm., 129.85 grams. 
 
 (2) 47 by 32 mm., 15 grams 81 by 57 mm., 115.5 grams. 
 
 (3) 47 by 30 mm., 16.5 grams 96 by 67 mm., 145.2 grams. 
 
 Obovaria ellipsis: 
 
 (1) 52 by 52 mm., 59.1 grams 57 by 55 mm., 74.6 grams. 
 
 (The identification of this specimen is somewhat uncertain.) 
 Quadrula solida: 
 
 (1) 35 by 36 mm., 27 grams 45 by 46 mm., 46.3 grams. 
 
 Anodonta imbecillis: 
 
 (1) 30 by 25 mm., 8 grams 61 by 28 mm., 13.3 grams. 
 
 In each case, the first measurement is the greatest antero-posterior length of the 
 shell, and the second the distance from the top of the umbo to the ventral margin taken 
 approximately at right angles to the lines of growth. An interesting and important fea- 
 ture of these specimens is the fact that the original margin is clearly indicated by a con- 
 spicuous line on the shell of each, and as the measurements within this line correspond 
 with the original measurements, the identification is made sure for each individual. 
 
 We quote below an analysis of the results sent us by Dr. Coker.who made the second 
 series of measurements after the recovery of the cages : 
 
 Lampsilis venlricosa. — They have increased in length by 34 to 39 mm. and in height by 25 to 37 
 mm., and they now weigh approximately 7, 8 and 9 times as much, respectively, as when first put out. 
 Furthermore, the added area of shell is divided by a conspicuous dark ring and a less distinct ring which, 
 one is tempted to assume, represent the periods of cessation of growth during the two winters. If 
 such an interpretation is made, the growth was accomplished chiefly during 1908 and 1909, while during 
 the present year (1910), the mussel having reached adult size, the growth has been considerably less. 
 
 Increase in size stated by percentage (present measurements compared with original measurements). 
 Period, June 29, 1908, to November 15, 1910, 2 years, 4J months: 
 
 Length. Height. Weight. 
 
 Specimen no. 1 per cent.. 188 217 812 
 
 Specimen no. 2 do... 172 178 770 
 
 Specimen no. 3 do. . . . 204 223 880 
 
 The proportion of increase is slightly greater in height than in length, and the coefficient of increase 
 in weight is, as might be expected, something like the cube of the coefficient of increase in either 
 dimension. 
 
 Obovaria ellipsis. — The specimen has probably gained very little in length or height but materially 
 in weight. It was nearer its adult size, is doubtless a slower growing species, and has probably gained 
 in weight by increase of thickness of shell. But we are not so sure of the identity of this specimen. 
 
 Quadrula solida. — Has gained nearly 30 per cent in length and height and 70 per cent in weight. 
 
 Anodonta imbecillis. — Has more than doubled in length, with negligible increase in height, while 
 it has increased 66 per cent in weight. This is particularly interesting as showing a marked change 
 in form from the young to the adult. 
 
 Text figure 4, a and b, represents outline sketches of two of the three specimens 
 of L. venlricosa described above, showing the exact size of each after the completion of 
 the growth in the fall of 19 10; the line marked a is the margin of the shell at the time the 
 planting was made in 1908; while lines b and c are the two successive rings indicating 
 cessation of growth. The two areas inclosed between these lines, representing the two 
 chief periods of growth which have occurred, are not of equal extent in the three speci- 
 
1 82 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 mens. In a they are of about equal width, while in b the second area is much greater 
 than the first. The area between line c and the margin of the shell is in all three cases 
 very narrow, showing that, as the mussel approaches the adult size, further increase in 
 the shell must take place very slowly. The recovered specimen of Q. solida shows only 
 one broad area of growth, and a very narrow one around the margin. This mussel was 
 relatively much nearer adult size when put in the cage than the specimens of ventricosa. 
 Dr. Coker comes to the following conclusion with respect to the age of the specimens 
 of L. ventricosa: 
 
 They are very significant, as they show clearly that growth is much more rapid than is generally 
 suspected. Considering what the growth has been since the cages were put out, it is fair to assume that 
 the specimens had only one year's growth at that time. That is to say, they were glochidia in the spring 
 of 1907, and, since they must have been carried in the gills of the mother over the preceding winter, 
 their complete age at this time (Nov. 15, 1910) is a little over four years. 
 
 Their age since the metamorphosis would therefore be about three years. Their 
 probable history, on the above assumption, is as follows: 
 
 1. Eggs fertilized in August, 1906. 
 
 2. Glochidia discharged in spring or early summer, 1907. 
 
 3. Liberated from fish in summer, 1907. 
 
 4. Collected at age (since metamorphosis) of about one year and placed in cages 
 June 29, 1908. 
 
 5. Recovered and remeasured, November 15, 1910. 
 
 The rate of growth of these individuals is probably typical of the genus Lampsilis, 
 and the experiment indicates at least that commercial mussels may reach a marketable 
 size in three years from the time they leave the fish. With the heavier shelled species 
 (those of Quadrula, for example) the rate of growth is probably slower and a longer 
 time must elapse before they are large enough for commercial use. 
 
 These experiments, meager as they are, are quite significant and furnish the first 
 definite data, so far as we know, relating to the rate of growth of fresh-water mussels. 
 With the proper facilities and the opportunity of examining the mussels at closer in- 
 tervals, similar plantings could readily be made and exact information obtained on 
 the growth of all the important species. To prevent the cages from being buried in 
 the sand or mud would seem to be the chief precaution that should be taken in future 
 experiments of this kind. 
 
 AN ARTIFICIALLY REARED MUSSEL. 
 
 Another experiment, although it does not throw light upon the question of the rate 
 of growth in nature, might be mentioned in this connection on account of its significance 
 for the problem of artificial propagation. A lot of black bass which had been infected 
 with the glochidia of Lampsilis ligamentina, ventricosa, and recta at Manchester, Iowa, 
 on December 2, 1908, were brought to Columbia, Mo., and placed in a large tank con- 
 taining sand. The fish were left in the tank, where the young clams were allowed to 
 fall off in the hope that some would survive and be later recovered. The sand was 
 examined at intervals thereafter but never thoroughly, as the chance seemed very slight 
 that any of the young clams were still living. On December 26, 1910, however, a single 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 183 
 
 small individual of Lampsilis ventricosa was found alive and active in the sand of the 
 same tank. There can be no 
 doubt that it was derived 
 from the infection referred 
 to, as no young clams of this 
 species had ever been in the 
 laboratory, and no subse- 
 quent infections were made 
 in that tank. The exact size 
 of this young mussel was 41 
 by 30 mm. on December 26, 
 1910. It is still alive, but as 
 late as June, 191 1, it was 
 practically of the same size. 
 Since it is over two years old, 
 it is evident that it is quite a 
 dwarf, and, had it been reared 
 under favorable conditions, 
 it undoubtedly would have 
 been much larger by this time. 
 The tank in which it has 
 spent all of its life is supplied 
 with tap water, which is 
 obtained from deep wells and 
 contains little that a mussel 
 could utilize as food, and its 
 small size is undoubtedly due 
 to the fact that it has been 
 underfed from the beginning. 
 The shell shows no indication 
 whatever of lines of inter- 
 rupted growth, but this is 
 only what might have been 
 expected, as the mussel has 
 never been exposed to low 
 temperatures. It is evident 
 therefore, that it has been 
 growing continuously, but 
 very slowly, throughout its 
 entire life. 
 
 This individual, however, 
 is of no little interest, as it is 
 the first fresh-water mussel actually reared artificially from the glochidium, and in a sense 
 18713 — 12 6 
 
 Fig. 4. — Two individuals of Lampsilis ventricosa recovered on November 15, 1910, 
 after having been confined in a wire cage in the Mississippi River for two 
 years and four and a half months. The line a is the original margin of the 
 shell at the time of planting. June 29, 1908. and the lines b and c represent the 
 "rings" which are due to the periods of cessation of growth. Natural size. 
 
1 84 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 furnishes a demonstration of the feasibility of artificial propagation. Had the food supply 
 in the tank been adequate, it would now be a mussel of about two-thirds the adult size. 
 
 THE ORIGIN AND AGE OF MUSSELS IN ARTIFICIAL PONDS. 
 
 A second line of evidence bearing upon the rate of growth has been obtained in 
 connection with an examination of certain artificial ponds in the vicinity of Columbia, 
 Mo. In this region it is customary for the farmers to construct, for the watering of cattle, 
 ponds in which water is held the year round by the impervious clay soil. We have 
 examined many of these small bodies of water and have records of the approximate, if 
 not the exact, dates of their construction. In 12 of these ponds, the ages of which 
 are from 5 to 40 years, we have found specimens of Lampsilis subrostrata and Unio 
 tetralasmus in some numbers, and in two of the ponds the mussels are present in very 
 great numbers. 
 
 The occurrence of the mussels in the different ponds has been considered, first, 
 with a view to the question of their original introduction into a given pond, and, second, 
 their rate of growth. The first of these two considerations will be discussed here as a 
 matter of convenience, although it should more properly be considered in a section 
 dealing with the introduction of mussels into favorable localities. 
 
 As to their origin in the ponds, we find the facts interesting because it is quite clear 
 that a majority, if not all of the ponds, must have been stocked with mussels which 
 were first introduced as parasites upon fish. The significant facts in this connection 
 are: That we have never found a pond containing mussels but no fish, although there 
 are a number of ponds containing fish in which we have thus far failed to discover any 
 mussels, and that none of the ponds have outlets or other immediate connections with 
 streams in which the mussels occur, but are situated, for the most part, on high ground 
 far from the watercourses, making it impossible that the mussels could have worked 
 their way into these bodies of water by any ordinary process of migration. Since it is 
 very unlikely that persons have introduced adult mussels into so many places by intent 
 or accident, the mussels must have appeared in these ponds by natural means and the 
 most probable of these is their introduction while parasites upon the fish with which 
 the ponds were stocked. The transportation of small individuals attached to the mud 
 on the feet of birds or of terrestrial animals, so often suggested as a means of dispersal 
 in a case like this, is a possible mode of origin, although it seems hardly a probable one 
 in view of the excellent chance the mussels would have of being introduced while still 
 parasites. 
 
 One of the above ponds, which is about 40 by 60 feet in area and 10 feet in depth, 
 is particularly interesting since it contains great numbers of Lampsilis subrostrata and 
 also of the sunfishes (Lepomis humilis and Apomotis cyanellus), which we have found in 
 our laboratory experiments to be very favorable hosts for the glochidia of this mussel. 
 The mussels are of all sizes and the pond has existed for many years. We do not know 
 its exact age nor how long ago fish were introduced. The mussels were first discovered 
 in 1907 and have ever since been found in abundance. Their success is doubtless due, 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 1 85 
 
 in large part, to the abundance of a fish favorable for their parasitism. Nothing in 
 these specimens, nor in what we know of the history of this pond, gives a clue to the age 
 of the mussels. 
 
 Another pond has great numbers of Unio tetralasmus . This pond was constructed 
 in 1 901 and during the first year was stocked with fish (the exact species unknown). 
 In 1907 it contained a great many mussels as long as 4 inches, and since that year the 
 largest individuals have slightly exceeded this size, which is near the maximum as we 
 know it for this species. It is inconceivable that these unios were introduced as adults, 
 for they are present in great numbers, and the farmer who owned the land was astonished 
 to find them there four or five years after the pond was established, because it was near 
 the entrance to his dooryard and he knew that no one had introduced mussels in any 
 such numbers and that there was no watercourse connecting the pond with any creek 
 in which mussels occurred. These mussels evidently came as parasites upon the fish 
 with which this pond was stocked during the first year and they had reached a length 
 of 4 inches in a period of five years. The abundance of the adults when the pond was 
 six years old and the presence of some smaller specimens made it seem that more than 
 one generation was represented, and hence some may have reached this size in a shorter 
 time. The shell of Unio tetralasmus is light and is by no means a good button shell. 
 Still it is not an impossibility, commercially speaking, for we have been assured by one 
 of the leading button manufacturers, Mr. J. E. Krouse, of Davenport, Iowa, to whom 
 we sent shells from which buttons were cut, that a marketable button could be made 
 from them and would be made if there were no other shells available. 
 
 The appearance of Lampsilis subrostrata and Unio tetralasmus and no other species 
 in all the ponds examined suggests the question, why have these two species and no 
 others become established? If they were introduced as glochidia infecting fish, is it 
 likely that the different lots of fish placed in so many ponds were infected solely with 
 the glochidia of these two species? It seems much more probable that other mussels 
 were introduced in the parasitic stages and that they w r ere not able to survive long 
 upon the bottom of these ponds. We have introduced large adult specimens of Ouadrula 
 metanevra and Symphynota complanata into one of the ponds in question and found 
 some of them still alive after two years. This pond had a very soft mud bottom well 
 covered with a layer of black muck filled with the soft coal soot from the smoke of a 
 neighboring power-house chimney and seemed unsuitable for any variety of mussel. 
 It had become, in spite of this, well stocked with Lampsilis subrostrata and is the pond 
 referred to in detail in a previous paragraph. The survival here of these specimens of 
 heavy shelled mussels for a period of two years shows that the adults are not at once 
 killed even by unfavorable conditions, and we are therefore inclined to believe that 
 when these species are introduced into the ponds on fish their destruction occurs in the 
 early juvenile stages. 
 
 If a small body of water can be so fully stocked by the scant infection of glochidia 
 obtained by fish in nature, we should be able to introduce mussels like these into a pond 
 far more effectively by the use of fish which had been artificially infected and to rear 
 
1 86 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 them to adult size within a short term of years. Accordingly, we have attempted the 
 introduction of Lampsilis ligamentina into one of the ponds where no mussels had ever 
 been found by placing in the pond several hundred fish well infected with the glochidia 
 of this species; but several examinations of the mud and silt from the bottom, made 
 during the 1 8 months following, have failed to show anything as a result of the experiment. 
 
 The conclusions drawn from these observations are encouraging because they 
 indicate, first, that other species, like those of the genus Lampsilis, whose shells are of 
 excellent quality for the best of buttons, may be reared to commercial size in about 
 the same length of time, and, second, that restricted localities can be stocked with 
 mussels by the introduction of fish infected with glochidia. The members of the genus 
 Lampsilis have shells which are evidently not much heavier than the shell of Unio 
 tetralasmus, a fact which better fits them for life upon soft bottoms where there is little 
 current, and in such localities they often occur. They move about more actively than 
 the heavier shelled species and this, doubtless, enables them readily to seek out the 
 most favorable food conditions in any body of water, instead of remaining long in one 
 place where the conditions are very stable, as do the heavier shelled species. The 
 study of any mussel which can live in small ponds like those in question and from which 
 button shells can be obtained should be followed up with care, since the extensive 
 culture of mussels would be a far simpler matter in ponds than in any stream where 
 high and low water and the shifting of the bottom might so largely interfere with the 
 most carefully located beds. For this purpose the species of Lampsilis which give 
 good button shells would seem the most desirable, because they are better adapted for 
 the conditions and because our planting experiments indicate that they reach a market- 
 able size in a shorter time than the quadrulas. 
 
 We feel that there is nothing discouraging in what is at present known regarding 
 the rate of growth under the average natural conditions. Moreover, it should be 
 remembered that in most invertebrates where the growth rate has been studied this 
 may be modified to an astonishing degree by the food supply and that the actual size 
 of an individual furnishes no trustworthy clue to its age. It is not at all unlikely that 
 proper study of the food and other conditions necessary for the maximum rate of growth 
 will enable us to obtain shells of commercial size in even slow-growing varieties within 
 a reasonable number of years. To judge from the supposed annual rings of specimens 
 taken in nature, Quadrula ebena may take from 20 to 30 years to reach, under natural 
 conditions, the size which is most desirable. The question whether this is a necessity. 
 or only a result of the poverty of food conditions which most mussels meet in nature, 
 is one which must wait upon the proper scientific analysis of the mussel's food and rate 
 of growth in this and other species, and there is no problem in connection with the 
 attempted artificial propagation which has more pressing importance. 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 1 87 
 
 VII. INVESTIGATIONS ON THE UPPER MISSISSIPPI RIVER. 
 
 A brief reference may here be made to certain field studies which were carried on 
 in connection with our mussel investigations during the months of June, July, and 
 August, in 1908, on the upper Mississippi River. The Bureau of Fisheries put at our 
 disposal for this purpose its substation, a small building provided with tanks and running 
 water, at La Crosse, Wis., and also its steamboat, the Curlew, which not only furnished 
 us with living quarters, but was of invaluable service for transportation from place to 
 place on the river (fig. 65, pi. xvi). The boat, which is ordinarily used in the work of 
 reclaiming young fish from the overflow of the river during the floods which occur in the 
 spring and early summer, is equipped with aerated tanks, seines, and other apparatus 
 and provided us with what was essentially a floating laboratory. With these facilities 
 much was accomplished that would have otherwise been impossible. In addition to the 
 usual crew of the Curlew, the party consisted, besides ourselves, of Messrs. W. E. Muns, 
 Howard Welch, F. P. Johnson, and W. E. Dandy, students in the University of Missouri, 
 who served as assistants. 
 
 The primary object of the expedition was a determination of the breeding seasons of 
 the commercial species of mussels as far as possible at that time of the year and an 
 examination of the depleted mussel beds in the upper Mississippi River, which have 
 been all but destroyed as a result of the ravages of the mussel fisheries. 
 
 With a clamming outfit of our own (fig. 69, pi. xvn), consisting of a flat-bottomed 
 skiff and "crow-foot" dredges — the usual apparatus employed by the mussel fishermen — 
 we were able to secure thousands of mussels, which were examined microscopically for 
 the purpose of determining their sex and the stage of development of the embryos. The 
 data thus obtained furnished a mass of detailed information, especially with respect to 
 those species which breed in the summer, but as they are incorporated in the account 
 already given of the breeding seasons, there is no need to refer to the subject again. 
 
 The planting of young mussels in cages for a determination of the rate of growth 
 was also made during this summer, with the result as described in a preceding section. 
 
 Some attempts were made to infect fish with glochidia, but this phase of the work 
 was greatly interfered with by the high water of the river, which remained at flood stage 
 unusually late in the summer of 1908 and made the seining of fish very difficult. Some 
 infections, however, were carried out with the glochidia of a few summer-breeding species, 
 the fish being retained in the tanks at the La Crosse station throughout the parasitic 
 period and the duration of the parasitism determined. 
 
 A thorough survey of the mussel beds from Winona, Minn., to Lansing, Iowa, was 
 made, and records taken at each locality where mussels were collected. No large beds 
 at all were discovered, and in every instance where mussels were found indications of the 
 ravages worked by the clammers were apparent. An account of the distribution of the 
 species throughout this section of the Mississippi River and their relative abundance is 
 not presented here, as the results of our observations in these respects will be incorpo- 
 rated in the work of the several field parties which have been engaged in the study of 
 
1 88 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 the geographical distribution of the Unionida? throughout the Mississippi Valley under 
 the direction of the Bureau of Fisheries during the past four or five years. 
 
 While working in the neighborhood of La Crosse, we made a careful investigation of 
 the west channel of the river at this locality, with a view to determining whether places 
 of this nature presented favorable conditions for experimental rearing of young mussels. 
 As is usually the case with the accessory channels of the river in this region, the west 
 channel at La Crosse is dammed across its head for the purpose of confining the water 
 in the main channel, and, although at high-water stages of the river the dam is sub- 
 merged, during the greater part of the year the volume of water in the channel is greatly 
 reduced and the current retarded. These dams, however, are never tight, and a greater 
 or less quantity of water constantly seeps through them. A thorough study of this 
 channel showed that it contained very few mussels indeed, and of those species that 
 were found living in small numbers under these conditions, the majority belonged to 
 Lampsilis, ventricosa being by far the most abundant form. Whenever a channel of 
 the river is dammed, the slackening of the current causes an enormous sedimentation to 
 take place, and in these "sloughs," as such obstructed channels are called, sand and 
 mud bars and shoals have been formed to an extent varying with the length of time since 
 the dam above them was built. The more sluggish species of mussels, like the quadrulas, 
 are especially ill adapted to these conditions and are frequently buried and destroyed 
 by the deposits of silt in the river, an occurrence of which we found abundant evidence. 
 With the more actively moving and burrowing species, as those of Lampsilis, the case 
 is different, for apparently they may adjust themselves more readily and by their far 
 greater ability to move from place to place they may avoid the danger of being buried. 
 We found little evidence that the quadrulas, for example, move about at all, while, on 
 the contrary, the tracks of slowly wandering individuals belonging to the species of 
 Lampsilis were everywhere conspicuous on the sandy bottoms of the shallow sloughs. 
 
 An interesting case of the destruction of mussel beds in situ by sedimentation is 
 shown in figure 70, plate xvn, which is a photograph taken on the bank of a slough, near 
 Muscatine, Iowa, which was exposed by a gully washed out by rains and cut directly 
 through an extinct mussel bed. The photograph shows the surface of the cut where the 
 mussels are exposed as they lie embedded in the muddy bank. The bed is buried under 
 about a foot of mud, and it is interesting to note that the valves of the mussels are closed 
 and lying together in pairs. The latter fact proves conclusively that this is not an old 
 shell heap, for the valves of the shells would be found scattered and separated in that 
 event, but a mussel bed which had once existed in the river near the bank. It was 
 probably buried under the deposits of sand and mud which followed the building of the 
 dam across the head of the slough. An investigation of the species represented in the 
 bed showed that they all belonged to Quadrula, being chiefly ebena, pustulosa, and trigona, 
 while not a single individual belonging to Lampsilis could be found in it. It is probable, 
 as already stated, that it is the sluggish species, like those of Quadrula, that are the prin- 
 cipal sufferers in catastrophies of this nature, and are caught and smothered in the process 
 of sedimentation, while the propensity to wander possessed by the more active species 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH -WATER MUSSELS. 1 89 
 
 enables them to move out into deeper water when the deposit of silt becomes a menace. 
 The result of our study of the conditions obtaining in sloughs like the west channel 
 at La Crosse, which are closed by dams at their heads, proves conclusively that such 
 waters afford a very unfavorable habitat for mussels, and that therefore they are not 
 adapted to experimental uses. 
 
 VIII. ECONOMIC APPLICATIONS. 
 
 It may not be inadvisable to discuss briefly certain applications of the results 
 obtained in the foregoing investigations to the practical work of artificially propagating 
 fresh-water mussels on a commercial basis. It must be emphasized at the outset that 
 the ultimate object of the investigations — the restocking of depleted waters with com- 
 mercial species of mussels — is not dependent for its realization solely upon the success 
 of rearing mussels artificially from the glochidia, but that other methods of attaining the 
 same end may be employed which are of equal, if not greater, importance. 
 
 PROTECTIVE LAWS. 
 
 Much can undoubtedly be done by securing the passage of laws by State legislatures 
 for the closing of certain streams or sections of streams against all clamming for a period 
 of years of sufficient length to allow of a natural increase of the mussels; by laws pro- 
 hibiting the use of the ordinary "crow-foot" dredge, which takes immature and adult 
 individuals indiscriminately, ° and by laws prohibiting the discharge of sewage and 
 factory refuse in the neighborhood of mussel beds. By these and other protective 
 measures of a legal nature, a great deal might be accomplished in the way of conserving 
 the supply of mussels in the more important waters, but, since in the case of many rivers 
 the control is in the hands of two or more States, the passage of such laws would require, 
 to be effective, similar action on the part of several legislatures, and such cooperation 
 might not be obtained without the greatest difficulty. 
 
 The utter futility of laws which would establish a closed season of the year against 
 clamming is apparent in the light of our knowledge of the breeding seasons of the 
 Unionida. We have already seen that there is no month in the year when some species 
 are not bearing embryos or glochidia, and as species of commercial value are found in 
 both groups — those with the long and those with the short period of gravidity — a 
 closed season at any time would be of little or no avail. Several species of Lampsilis, 
 for example, which bear embryos or glochidia from August to July, furnish valuable 
 shells for the pearl-button industry, while the species of Ouadrula and other summer 
 breeders, gravid from May to August, supply shells of the best quality. Any law then, 
 designed to relieve the situation, which prohibits the taking of mussels during a sup- 
 posed breeding season is based on ignorance of the facts, for the entire year is the breed- 
 
 o Mussels caught on a hook of the ' ' crow-foot ' ' are generally so badly injured internally in the process that, even if they are 
 afterwards thrown back into the river, the majority probably die. A special form of hook has been devised by Mr. J. F. 
 Boepple which is so constructed that small mussels can not be caught by it. The use of some such selective apparatus should 
 be required by law. 
 
190 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 ing time of the Unionidae. A law, however, which would close a river or large section of 
 a river for a period of five years or more would be most beneficial, for in that time much 
 could be accomplished both by artificial and by natural means to restore normal conditions. 
 Even artificial propagation, unaided by certain protective measures, could hardly be- 
 come effective on however extensive a basis it might be carried on, for unless some 
 means can be devised for saving the young mussels it is difficult to see how much head- 
 way could be made against the destruction of the supply. It therefore becomes of vital 
 importance not only to make illegal the use of any apparatus which will catch or injure 
 young mussels, but to see that the law is rigidly enforced. 
 
 Certain requisite conditions for the artificial culture of fresh-water mussels, based 
 upon our knowledge of their life history and habits, may now be briefly referred to. 
 
 SELECTION AND MAINTENANCE OF A FISH SUPPLY. 
 
 Although only a comparatively few kinds of fishes have been thus far used in our 
 experimental infections, and doubtless as our experience widens many more will be 
 found to be favorable for the purpose, success has been attained chiefly with the black 
 basses, rock bass, and the sunfishes. All of these fishes have proved to be extremely 
 resistant to the injurious effects of gill infections (practically all of the commercial 
 species of mussels have hookless glochidia, which are gill parasites) ; to be able to carry 
 large numbers of glochidia through the parasitic period; and to be easily kept in confine- 
 ment — three necessary conditions for the success of propagation. It is to be hoped 
 that other fishes will be found to be equally useful, but at present those just mentioned 
 afford the most promising material for the work. As has already been shown, some 
 species of fishes are very easily killed even by light gill infections, while others, accord- 
 ing to our experience, have resisted all attempts to bring about permanent implantation 
 of glochidia on their gills. The latter is particularly true of German carp and catfishes. 
 
 Fortunately, the basses and sunfishes can be obtained in large quantities without 
 serious difficulty. In the reclamation work conducted by the Bureau of Fisheries 
 along the upper Mississippi River, immense numbers of young bass are annually seined 
 from the sloughs and "lakes" into which they are carried when the river rises over its 
 banks during the flood stages of early summer. When the water recedes these young fish 
 are caught outside the banks of the river, and only the small fraction of them which is 
 reclaimed in the seining operations is saved from the wholesale destruction (fig. 67, 
 pi. xvi). There is no limit to this supply of material for the work of mussel culture, 
 and doubtless extensive use will be made of it at the Fairport station. 
 
 Even more valuable for the purpose are the species of sunfishes which we have used 
 (probably other species of the same group are equally good), for, besides being just as 
 resistant and as readily infected as the black bass, they are more easily kept and are 
 less subject to disease in confinement. An adequate number of breeding ponds, in which 
 sunfishes could be left to multiply naturally, would insure a large and constant supply 
 of these fish for artificial infections. 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 191 
 
 THE BEST SEASONS FOR INFECTIONS. 
 
 It has already been stated that the duration of the parasitic period of the mussel is 
 inversely proportional to the temperature of the water. This fact is obviously import- 
 ant for mussel culture, since the longer the fish have to be kept while carrying the glo- 
 chidia the greater is the loss from disease and other causes. The loss not only involves 
 the fish but the potential mussels which they are nourishing as well. It therefore be- 
 comes desirable to reduce, as far as possible, the length of time that the infected fish 
 must be retained, and this we have seen depends upon the temperature. Late spring 
 and summer, consequently, are the seasons when the maximum efficiency from arti- 
 ficial infections should be obtained, for in the warmer water at that time the duration 
 of the parasitism will be at the minimum — about two weeks or even less. The glo- 
 chidia of Lampsilis are available all through the spring and as late as July, while those 
 of Quadrula can be obtained during the summer months, and most of the commercial 
 species of mussels fall in these two genera. Of course infections can successfully be made 
 in the fall and winter and the duration of the parasitism reduced by keeping the water 
 artificially warmed, but the difficulty of maintaining the fish alive under these con- 
 ditions is greatly increased. 
 
 THE MUSSEL SUPPLY. 
 
 By far the greater number of species of commercial value, as has already been stated, 
 belong to the genera Lampsilis and Quadrula, and, as both of these genera are widely 
 distributed, practically all of the mussel-bearing streams of the Mississippi Valley 
 may be drawn upon for a supply of material for cultural purposes. We have found 
 that living mussels may be shipped even long distances with little or no mortality, 
 especially in cool, weather, and it is therefore possible to obtain breeding material from 
 places at quite a distance from the station where the infections are to be made, should 
 the local supply be inadequate. We have had on several occasions large numbers of 
 gravid mussels shipped from Terre Haute, Ind., to La Crosse, Wis., to Manchester, 
 Iowa, and to Columbia, Mo., with scarcely the loss of an individual, and have successfully 
 used the glochidia obtained from them in infecting thousands of fishes. 
 
 According to our experience mussels thrive very well in confinement, in small ponds 
 and laboratory tanks, and that without any special attention to a food supply. We 
 have for years been keeping both pond and river forms alive in the laboratory for months 
 at a time in tanks containing a few inches of sand on the bottom and supplied by tap 
 water. Under such conditions mussels have frequently been retained in the laboratory 
 from the fall to the following summer. It should therefore be an easy matter to keep 
 mussels for breeding purposes in ponds with natural bottoms in any quantity desired, and, 
 if the ponds are fed with river water, a natural food supply should be present in abundance. 
 
 Since, as has been pointed out above, the species of Quadrula, Unio, and other sum- 
 mer breeders abort their embryos and glochidia with astonishing ease when disturbed, 
 it will be necessary, when making infections with the glochidia of forms exhibiting this 
 peculiarity, to collect the material at a time prior to the fertilization of the eggs and to 
 
192 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 allow them to enter upon the breeding season after being placed in the ponds of the 
 station. We have had females of different species of Quadrula become gravid in the 
 tanks of the laboratory after they had been held in confinement for weeks or even months, 
 and therefore no difficulty should be encountered in obtaining a supply of glochidia 
 from these forms under the conditions mentioned. 
 
 REARING AND DISTRIBUTING YOUNG MUSSELS. 
 
 After the fish have been infected, one of two things may be done in distributing 
 the young mussels resulting therefrom: Either the fish, after having been retained in 
 tanks or ponds until near the end of the parasitism, may be taken to the stream which is 
 to be restocked and the clams allowed to drop off there, or the liberation may take place 
 in ponds where the young mussels may be reared until they are of considerable size, 
 say until they are a year old, and then distributed as desired. Both methods might be 
 used successfully, but in the first case it is to be supposed that only a very small pro- 
 portion of individuals thus liberated would succeed in reaching maturity, as they would 
 be exposed to the same destructive agencies as are encountered under natural conditions. 
 The difficulty and expense of transporting the infected fish, the mortality among the 
 fish themselves resulting from shipment, and the subsequent loss of large numbers of 
 the young mussels are considerations which lead one to regard this method as not an 
 efficient one. It should be stated, however, that in using this method of distribution 
 it would not be necessary to liberate the fish and thus lose them for subsequent infections, 
 for they could be confined in wire-bottomed fish cars set out in the streams, and after 
 the mussels had all fallen off and dropped through the bottoms of the cars the fish could 
 be returned to the station. This would of course involve a very large amount of labor 
 and much expense. 
 
 It would, therefore, seem to be a far more effective practice to retain the young 
 clams in ponds with natural bottoms until they could with safety be liberated in the 
 streams. After infection, in this event, the fish could be set free in these ponds at once, 
 and allowed to remain there throughout the parasitism of the glochidia, at the close of 
 which they could be seined out and made to do service again. Supplied with river 
 water, the ponds should furnish an adequate amount of food for a practically normal 
 rate of growth of the young mussels, which at the end of a year at latest should be of 
 sufficient size to be placed in favorable localities in the rivers. When ready for dis- 
 tribution, the water in the ponds could be drawn off and the juvenile mussels raked 
 carefully from the sand or mud. If properly packed, it should be possible to ship 
 them in large numbers to considerable distances. It is only reasonable to suppose 
 that a large proportion of the mussels thus reared would reach maturity after distribu- 
 tion, and it is certain that the number coming through would be far greater than would 
 be the case if the first method should be pursued. 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 1 93 
 
 IX. CONCLUSION. 
 
 Of course, many practical details essential to success will have to be worked out 
 before the artificial propagation of fresh-water mussels will have passed beyond the 
 experimental stage, for the efficiency of the work from an economic point of view will 
 doubtless depend upon the satisfactory solution of certain problems in technique, which, 
 although secondary in character, are nevertheless a prerequisite of success. 
 
 However much is yet to be done — and it should be clear that the work is far from 
 completion — the entire feasibility of artificial propagation has been demonstrated beyond 
 the shadow of doubt. Besides filling in the gaps, some of them important ones, in the 
 results already obtained, certain fundamental phases of the mussel investigations 
 remain practically untouched. Chief among these is an exhaustive study of the physical 
 conditions of the waters as affecting the growth of mussels: The relation between the 
 mineral content of the water and shell formation; the relation between the character of 
 the bottom, whether rocky, sandy, or muddy, to the habits of different species; and the 
 relation between the rapidity of current to the life of the mussel and the kind of shell 
 which it secretes. These and many other interesting problems of a similar nature 
 await solution. 
 
 The immense mass of data that have been collected by the Bureau of Fisheries 
 with respect to geographical distribution of species and their relative abundance through- 
 out the Mississippi Valley has not been digested, yet the results which will be derived 
 from a careful analysis of this information will have a fundamental economic bearing 
 upon mussel culture. It is essential to know the centers and limits of distribution of at 
 least the more valuable commercial species for the purpose of effectively conducting 
 the operations in restocking streams and of avoiding useless labor in attempting to 
 establish a species where the chances of its survival would be slight. 
 
 The whole problem of the food of mussels is as yet untouched. Not only are we 
 ignorant of the specific food forms among the micro-organisms upon which mussels 
 depend, but we do not know whether different species, or rather species living under 
 different physical conditions and species possessing different habits, utilize different 
 food forms. The possibility of artificially rearing cultures of the unicellular organisms 
 used as food — when we know what these forms are — for enriching the water in which 
 young mussels are retained before distribution should be determined, for it is undoubtedly 
 true that results of the greatest practical importance and interest would be derived 
 from such an investigation. 
 
 Very little is known at present respecting the enemies and diseases of fresh-water 
 mussels, yet the importance of information of this nature can not be overestimated. 
 Especially should we know the relative susceptibility of different species to parasitic 
 diseases, and whether certain species are immune against the invasion of parasites 
 which in the case of other forms constitute serious enemies. 
 
 A most fascinating and valuable field of investigation lies open in the study of the 
 causes of pearl formations, for since these concretions are due, in part at least, to the 
 
194 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 presence of parasites, the possibility of producing them at will offers an interesting 
 opportunity for experimental study. 
 
 The Unionida?, in short, are a group of animals which, for the great variety of 
 problems, both scientific and economic, presented in their unique life history, their 
 structure, functions, and habits, their many interesting adaptations, and in their 
 economic relations, is scarcely excelled by any other invertebrates except the insects. 
 At present we may be said to possess only an introduction to a knowledge of the family, 
 and the writers of this paper will feel amply repaid for their labor if they have succeeded 
 in exposing some of the problems which here lie open for investigation and at the same 
 time in laying the foundation for the artificial culture of fresh-water mussels. 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 195 
 
 BIBLIOGRAPHY. 
 
 Baer, C. E. von. 
 
 1830. Ueber den Weg den die Eier unserer Susswassermuscheln nehmen um in die Kiemen zu 
 gelangen. Archiv fur Anatomie und Physiologie, bd. 7, p. 313-352. 
 Blainville, Ducrotay de. 
 
 1828. Rapport fait a l'Academie des Sciences de Paris sur un memoire de M. Jacobson. Annales 
 des Sciences naturelles, t. 14, p. 22. 
 Braun, M. 
 
 1878a. Ueber die postembryonale Entwicklung unserer Susswassermuscheln. Berichte der 
 physikalisch-medicinischen Gesellschaft zu Wiirzburg, Mai-Heft, p. 24-27; also in 
 Jahrbuch der deutschen malakozoologischen Gesellschaft, jb. 5, p. 307-319. 
 1878b. Ueber die postembryonale Entwicklung unserer Susswassermuscheln (Anodonta). 
 
 Zoologischer Anzeiger, jg. 1, p. 7-10. 
 1884. Ueber Entwicklung der Enten- oder Teichmuschel. Sitzungsberichte der Dorpater 
 
 Naturforscher-Gesellschaft, bd. 6, p. 429-431. 
 1889. Die postembryonale Entwicklung der Najaden. Nachrichtsblatt der deutschen malako- 
 zoologischen Gesellschaft, jg. 21, p. 14-19. 
 Call, R. E. 
 
 1887. Note on the ctenidium of Unio aberti Conrad. American Naturalist, vol. 21, p. 857-860. 
 Carus, C. G. 
 
 1832. Neue Untersuchungen iiber die Entwickelungsgeschichte unserer Flussmuschel. Nova 
 Acta Physico-medica Academiae Caesareae Leopoldino-Carolinae Naturae Curiosorum, bd. 
 16, p. 1-87. 
 Coker, R. E., and Surber, T. 
 
 191 1. A note on the metamorphosis of the mussel Lampsilis laevissimus. Biological Bulletin, 
 vol. 20, p.. 170-182. 
 Conner, C. H. 
 
 1907. The gravid periods of Unio. Nautilus, vol. 21, p. 87-89. 
 
 1909. Supplementary notes on the breeding seasons of the Unionidae. Nautilus, vol. 22, p. 
 in, 112. 
 Faussek, V. 
 
 1893. Biologische Studien. I. Ueber Parasitismus und Viviparitat. Russkoje Bogatstwo, bd. 1. 
 1895. Ueber den Parasitismus der Anodonta-Larven in der Fischhaut. Biologisches Central- 
 
 blatt, bd. 15, p. 115-125. 
 1901. Ueber den Parasitismus der Anodonta-Larven. Verhandlungen des V. internationalen 
 Zoologen-Congresses (Berlin), p. 761-766. 
 
 1903. Parasitismus der Anodonta-Larven. Memoires de l'Academie des Sciences de St. 
 
 Petersbourg, viii ser., classe physico-mathematique, t. 13. 
 
 1904. Viviparitat und Parasitismus. Zoologischer Anzeiger, bd. 27, p. 761-767. 
 Flemming, W. 
 
 1874. Ueber die ersten Entwicklungserscheinungen am Ei der Teichmuschel. Archiv fur 
 
 Mikroskopische Anatomie, bd. 10, p. 257-293. 
 
 1875. Studien in der Entwicklungsgeschichte der Najaden. Sitzungsberichte der kaiserlichen 
 
 Akademie der Wissenschaften (Wien), in. abth., bd. 71, p. 1-132. 
 Frierson, L. S. 
 
 1904. Observations on the genus Quadrula. Nautilus, vol. 17, p. in, 112. 
 
196 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 Harms, W. 
 
 1907a. Ueber die postembryonale Entwicklung von Anodonta piscinalis. Zoologischer 
 
 Anzeiger, bd. 31, p. 801-814. 
 1907b. Zur Biologie und Entwicklungsgeschichte der Flussperlmusehel (Margaritana margari- 
 
 tifera Dupuy). Ibid., bd. 31, p. 814-824. 
 1907c. Die Entwicklungsgeschichte der Najaden und ihr Parasitismus. Sitzungsberichte der 
 
 Gesellschaft zur Beforderung der gesammten Naturwissenschaften zu Marburg, p. 79-94. 
 
 1908. Die postembryonale Entwicklung von Unio pictorum und Unio tumidus. Zoologischer 
 
 Anzeiger, bd. 32, p. 693-703. 
 
 1909. Postembryonale Entwicklungsgeschichte der Unioniden. Zoologische Jahrbiicher, Abteil- 
 
 ung fur Anatomie und Ontogenie, bd. 28, p. 325-386. 
 ISELY, F. B. 
 
 1911. Preliminary note on the ecology of the early juvenile life of the Unionidae. Biological 
 Bulletin, vol. 20, p. 77-80. 
 Israel, W. 
 
 191 1. Najadologische Miscellen. Nachrichtsblatt der deutschen malakozoologischen Gesell- 
 schaft, p. 10-17. 
 Jacobson, L. L. 
 
 1828. Undersogelser til naermere Oplysning af den herskende Mening om Dammuslingernes 
 
 Fremarling og Udvikling. Kongelige Danske Videnskabernes Selskabs Skrifter, 
 
 Naturvidenskabelig og Mathematisk Afdeling (Kjobehavn), 1828, p. 251-297; reprinted in 
 
 Bidrag til Bloddyrenes Anatomie og Physiologie, heft 1, Kjobenhavn, 1828, p. 249-362. 
 
 Latter, O. H. 
 
 1891. Notes on Anodon and Unio. Proceedings of the Zoological Society of London, p. 52-59. 
 1904. The natural history of some common animals. Cambridge. 
 Lea, Isaac. 
 
 1827. Descriptions of six new species of Unios, etc. Transactions of the American Philosophical 
 
 Society, vol. 3, p. 259-273. 
 1838, 1858, 1863, 1874. Observations on the genus Unio, together with descriptions of new genera 
 and species, vol. 2, 6, 10, 13. Philadelphia. (Originally printed in 
 Transactions American Philosophical Society and Journal Academy of 
 Natural Sciences, Philadelphia. ) 
 Leeuwenhoek, A. van. 
 
 1722. Arcana Naturae Detecta, t. 2, epist. 83, and t. 3, epist. 95 and 96. Leyden. 
 Lefevre, G., and Curtis, W. C. 
 
 1908. Experiments in the artificial propagation of fresh-water mussels. Proceedings of the 
 Fourth International Fishery Congress (Washington), Bulletin of the Bureau of 
 Fisheries, vol. xxvm, p. 617-626. 
 1910a. The marsupium of the Unionidae. Biological Bulletin, vol. 19, p. 31-34. 
 1910b. Reproduction and parasitism in the Unionidae. Journal of Experimental Zoology, vol. 
 
 9, p. 79-115. 
 1911. Metamorphosis without parasitism in the Unionidae. Science, vol. ^$, p. 863-865. 
 Leydig, F. 
 
 1866. Mittheilung iiber den Parasitismus junger Unioniden an Fischen in Noll. Tubingen, 
 Inaugural-Dissertation. Frankfort a. M. 
 Lillie, F. R. 
 
 1895. The embryology of the Unionidae. Journal of Morphology, vol. 10, p. 1-100. 
 1901. The organization of the egg of Unio, etc. Ibid., vol. 17, p. 227-292. 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 197 
 
 Ortmann, A. E. 
 
 1909. The breeding season of Unionidae in Pennsylvania. Nautilus, vol. 22, p. 91-95 and 99-103. 
 
 1910a. A new system of the Unionidae. Ibid., vol. 23, p. 114-120. 
 
 1910b. The discharge of the glochidia in the Unionidse. Ibid., vol. 24, p. 94, 95. 
 
 1911. A monograph of the Najades of Pennsylvania. Memoirs of the Carnegie Museum (Pitts- 
 burgh), vol. 4, p. 279-347. 
 PECK, R. H. 
 
 1877. The minute structure of the gills of lamellibranch Mollusca. Quarterly Journal of Micro- 
 
 scopical Science, vol. 17, p. 43-66. 
 Pfeiffer, C. 
 
 182 1. Naturgeschichte deutscher Land- und Susswasser-Mollusken. Weimar. 
 Poupart, F. 
 
 1706. Remarques sur les coquillages a deux coquilles, et premierment sur Ies Moules (Anodontes). 
 Memoires de l'Academie des Sciences de Paris, p. 51-61. 
 
 QUATREFAGES, A. dE- 
 
 1835. Sur la vie intrabranchiale des petites Anodontes. Annates des Sciences naturelles, t. 4. 
 
 1836. Memoire sur la vie intrabranchiale des petites Anodontes. Ibid., t. 5, p. 321-336. 
 RaThke, J. 
 
 1797. Om Dammuslingen. Naturhistorie Selskabets Skrifter (Kjobenhavn), t. 4, p. 139-179. 
 Schierholz, C. 
 
 1878. Zur Entwicklungsgeschichte der Teich- und Flussmuschel. Zeitschrift fur wissenschaft- 
 
 liche Zoologie, bd. 31, p. 482-484. 
 1888. Ueber Entwicklung der Unioniden. Denkschriften der kaiserlichen Akademie der 
 Wissenschaften (Wien), Mathematisch-naturwissenschaftliche Classe, bd. 55, p. 183-214. 
 Schmidt, F. 
 
 1883a. Vorlaufiger Bericht Tiber Untersuchungen der postembryonalen Entwicklung von Ano- 
 
 donta. Sitzungsberichte der Dorpater Naturforscher-Gesellschaft, p. 303-307. 
 1885b. Beitrag zur Kenntniss der postembryonalen Entwicklung der Najaden. Archiv fur 
 Naturgeschichte, jg. 51, p. 201-234. 
 Simpson, C. T. 
 
 1900. Synopsis of the Naiades, or pearly fresh-water mussels. Proceedings of the United States 
 National Museum, vol. 22, p. 501-1044. 
 Sterki, V. 
 
 1895. Some notes on the genital organs of Unionidae, etc. Nautilus, vol. 9, p. 91-94- 
 
 1898. Some observations on the genital organs of Unionidae, etc. Ibid., vol. 12, p. 18-21 and 
 
 28-32. 
 1903. Notes on the Unionidae and their classification. American Naturalist, vol. 37, p. 103-113. 
 1907. Note. Nautilus, vol. 21, p. 48. 
 
BULLETIN OF THE BUREAU OF FISHERIES. 
 
 EXPLANATION OF PLATES. 
 
 [Drawings by G. T. Kline.] 
 
 PLATE VI. 
 
 Fig. i. Gravid female of Ptychobranchus phaseolus. Actual length 96 mm. 
 Fig. 2. Gravid female of Lampsilis subrostrata. Actual length 50 mm. 
 Fig. 3. Gravid female of Symphynota complanata. Actual length 170 mm. 
 
 PLATE VII. 
 
 Fig. 4. Gravid female of Dromus dromus. Actual length 57 mm. 
 Fig. 5. Gravid female of Quadrula ebena. Actual length 98 mm. 
 Fig. 6. Gravid female of Lampsilis recta. Actual length 122 mm. 
 Fig. 7. Gravid female of Obliquaria reflexa. Actual length 55 mm. 
 Fig. 8. Gravid female of Cyprogenia irrorata. Actual length 38 mm. 
 
 PLATE VIII. 
 
 Fig. 9. Hooked glochidium of Symphynota costata, anterior end view. For measurements see text 
 figure 1. 
 
 Fig. 10. Hooked glochidium, as above. Lateral view of living specimen. 
 
 Fig. 11. Axe-head glochidium of Lampsilis (Proptera) alata, anterior end view. For measure- 
 ments see text figure 1. 
 
 Fig. 12. Axe-head glochidium, as above. Lateral view. 
 
 Fig. 13. Hookless glochidium of Lampsilis subrostrata, lateral view. For measurements see text 
 figure 1. 
 
 Fig. 14. Hookless glochidium, as above. Posterior end view. 
 
 Fig. 15. Hookless glochidium, as above. Ventral view. 
 
 Fig. 16. Detail of a conglutinate of Lampsilis ligamentina. The glochidia, still inclosed in the 
 membranes, are less crowded together than those of figure 17, and are embedded in a mucilaginous 
 matrix. 
 
 Fig. 17. Detail of a conglutinate of Obliquaria reflexa, showing the membranes closely pressed and 
 adhering together. 
 
 Fig. 18. Young mussels (Lampsilis ligamentina) one week after liberation from the fish, showing 
 various positions assumed in crawling, the ciliation of the foot, and the new growth of shell. 
 
 PLATE IX. 
 
 Fig. 19. Fin of a carp about 3 inches long, 7 days after infection with glochidia of Anodonta cata- 
 racta, showing complete failure of the overgrowth of fin tissue in all places where the glochidia are greatly 
 crowded. See explanation in the text, p. 159, of the conditions along the upper margin. 
 
 Fig. 20. Tip of an over-infected fin, as above, 12 hours after infection, showing no appreciable over- 
 growth because of the crowding. The shadows represent glochidia upon the under surface. 
 
 Fig. 21. Pectoral fin of a carp, as above, 3^ hours after infection; an optimum infection. 
 
 Fig. 22. Ventralhalf of caudal fin of acarp, asabove, 24 hours after infection; an optimum infection. 
 
 Fig. 23. Tip of fin, as above, 32 days after infection. The shadows represent glochidia upon the 
 under surface. The infection is less than the optimum. The glochidia were well overgrown and all 
 alive when the fish was killed. 
 
 Fig. 24. Young Symphynota costata, attached by only a shred of tissue and about to drop from the 
 fin after a parasitism of 74 days. 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. I 99 
 
 PLATE X. 
 
 Fig. 25. Fin, as above, 36 hours after infection with glochidia of Anodonta cataracta, showing com- 
 plete overgrowth of all glochidia which have become properly attached. 
 
 Fig. 26. Glochidium of A. cataracta upon fin margin of carp, 3>- 2 ' hours after infection. Prolifera 
 tion of cyst just beginning. 
 
 Fig. 27. Glochidia, as above, upon fin margin of carp, showing different stages of cyst proliferation, 
 even in neighboring glochidia. 
 
 Fig. 28. Glochidia, as above, 24 hours after infection. 
 
 Fig. 29. Hooked and hookless glochidia (.4. grandis and L. recta) embedded and developing on a 
 fin margin. 
 
 Fig. 30. Glochidia of A. cataracta upon fin of carp, 3 days after infection, showing the cyst com- 
 pletely formed. 
 
 Fig. 31. Glochidium of A. cataracta upon fin of carp, developing normally after a shift of 90 degrees 
 from the position first taken. 
 
 Fig. 32. Two glochidia of .4. cataracta, overgrown after 36 hours upon surface of a carp's fin. 
 
 Fig. 33. Glochidium of .4. cataracta 35 days after infection. The metamorphosis is more advanced 
 than in figure 30 and the rudiments of the foot and other organs have assumed greater size. 
 
 PLATE XI. 
 
 Fig. 34. Three gill filaments of the rock bass infected with glochidia of Lampsilis ligamentina. 
 The metamorphosis of the glochidia has hardly begun, although they have been attached for 28 days. 
 
 Fig. 35, 36, 37, and 38. Stages in the formation of the cyst surrounding a hookless glochidium 
 (Lampsilis ligamentina) upon a gill filament of the black bass. Taken at 15 minutes, 30 minutes, 1 hour, 
 and 3 hours, respectively, after infection. The transverse lines on the filaments indicate the lamellae. 
 
 Fig. 39. Anterior gill of a black bass infected with glochidia of L. ligamentina, showing distribution 
 upon the gill as a whole and the appearance of the cysts. 
 
 Fig. 40. Gill of yellow perch, as above. 
 
 Fig. 41. Two conglutinates of Lampsilis ligamentina removed from the marsupium. One is shown 
 from the flat surface, the other on edge. Actual length 17 mm. 
 
 Fig. 42. Three conglutinates of Obliquaria reflcxa removed from the marsupium. Actual length 
 17 mm. 
 
 Fig. 43. Part of a gill of black bass infected with glochidia of L. ligamentina, showing the distribu- 
 tion and orientation of the glochidia in an infection above the optimum for this fish. Only the row of 
 filaments toward the observer is shown. 
 
 PLATE XII. 
 
 Fig. 44. Symphynota costala, dissected from fin of carp 47 days after infection. The anterior end 
 is to the left. Rudiments of foot, digestive tract, liver diverticula, and the first gill buds are recognizable ; 
 also the hooks and the degenerating adductor of the glochidium. Compare with figure 47. Actual size, 
 
 o-39 °y °-35 mm - 
 
 Fig. 45. Stropkitus edentulus, from a living specimen which had completed its metamorphosis 
 without parasitism and which was actively crawling about on the bottom. Seen from the ventral side. 
 The anterior and posterior adductors are well developed and within the foot the pedal ganglia and litho- 
 cysts may be seen. Two gill buds are found on either side. Sections show that the internal organiza- 
 tion is as far advanced as that of the young mussels shown in figures 47 and 48. X 106. 
 
 Fig. 46. A single cord discharged from the marsupium of Slrophitus edentulus, showing glochidia 
 extruded and others still within the cord. Xi3-5- 
 
 Fig. 47. Symphynota costata, a young mussel which had been crawling upon the bottom for 6 days 
 after a parasitism of 68 days. The very narrow margin of the adult shell has been drawn within the 
 18713 — 12 7 
 
200 BULLETIN OF THE BUREAU OF FISHERIES. 
 
 valves. The gloehidial shell and its hooks are still in evidence. In other respects the young mussel 
 shows most of the features which are characteristic of the adult. The anterior end is to the right. 
 Anterior and posterior abductors, anterior and posterior retractors, digestive tract divided into esopha- 
 gus, intestine and stomach with its large diverticula, cerebral, pedal, and visceral ganglia, lithocysts, 
 three gill buds, palp rudiments, the heart and pericardium will be recognized by their resemblance to 
 the adult organs. Sections show the rudiments of the kidneys. From a stained and decalcified speci- 
 men. Actual size, 0.39 by 0.35 mm. 
 
 Fig. 48. Lampsilis ligamentina, a young mussel 1 week after the close of the parasitic period. The 
 margin of the shell is extended well beyond the gloehidial outline and shows the first lines of growth. 
 More calcification has rendered the shell so opaque that the internal organs are no longer visible with- 
 out decalcification. Stained specimens and sections show about the same degree of organization as in 
 the Symphynola larva of figure 47. The foot with its cilia is shown extended and attached to the bottom 
 preparatory to drawing the mussel forward. From a living specimen. Actual size, 23 by 20 mm. 
 
 PLATE XIII. 
 
 Fig. 49. Alasmidonta truncata. Horizontal section of a water tube of gravid marsupium, taken 
 near ventral border of gill. The respiratory canals (r. c.) are small clefts, indistinctly shown under this 
 magnification (cf. fig. 56); the marsupial space contains young embryos. 
 
 Fig. 50. Quadrula ebena. Horizontal section of two adjacent water tubes (w. t.) of gravid mar- 
 supium containing young embryos. 
 
 Fig. 51. Anodonta cataracta. Horizontal section of a water tube of gravid marsupium, showing 
 respiratory canals (r. c.) and marsupial space (m. s ); the latter contains young embryos. 
 
 Fig. 52 Symphynota complanata. Horizontal section of a water tube of gravid marsupium, show- 
 ing respiratory canals and marsupial space; the latter contains glochidia. Note the thin, stretched 
 interlamellar junctions. 
 
 Fig. 53. Lampsilis ligamentina. Horizontal section of a water tube (w. t.) of gravid marsupium 
 containing young embryos. Note the thin, stretched interlamellar junctions (i. j.). 
 
 Figs. 54-55. Two stages showing process of implantation of a glochidium of Unio complanaius on 
 a filament of a gill excised 2 hours after infection. Figure 54 is taken 3 hoars after attachment, while 
 55 is the same glochidium drawn 2 hours later. The greater part of the cyst was formed while the gill 
 was in a watch glass. 
 
 PLATE XIV. 
 
 Fig. 56. Alasmidonta truncata. Horizontal section through portion of lamella and water tube 
 of gravid marsupium, showing respiratory canals (r. c.) near ventral border of gill; each canal is sep- 
 arated from the marsupial space by a septum (s). The interlocking cells, forming the suture in the 
 septum, are clearly seen. 
 
 Fig. 57. Anodonta cataracta. Section similar to last, but taken before fusion of folds (s), which 
 are seen not quite touching. The septum is formed by their fusion. Eggs contained in the marsupial 
 space are in an early cleavage stage. 
 
 Fig. 58. Anodonta cataracta. Region marked X in last figure, highly magnified, showing glandular 
 epithelium of respiratory canals (r. c), adjacent blood sinus (b. s.), and histological structure of sur- 
 rounding tissues. Note the muscle fibers. 
 
 PLATE XV. 
 
 Figs. 59-61. Transverse sections of glochidia of Symphynota complanata, taken 15 minutes, 6 hours, 
 and 24 hours, respectively, after attachment to edge of fish's fin, showing three stages in formation of 
 cyst. In 59 proliferation of epidermis is just beginning; in 60 glochidium is about half embedded; 
 while in 61 formation of cyst is completed. In 59, which is more highly magnified than the other two, 
 and in 60 several mitoses are shown in region of proliferation. In 60 cellular detritus from enclosed 
 edge of fin is being ingested by mantle cells of glochidium. 
 
REPRODUCTION AND ARTIFICIAL PROPAGATION OF FRESH-WATER MUSSELS. 201 
 
 Fi.gs. 62-63. Transverse sections of gloehidia of Lampsilis ligamentina, taken 30 rpinutes and 3 
 hours, respectively, after attachment to gill filament. In 62 the development of cyst has made con- 
 siderable progress, while in 63 the cyst wall is practically completed. In 62 several mitotic figures 
 are seen in the epidermis where multiplication of cells is taking place. 
 
 Fig. 64. Highly magnified section of a portion of the glandular epithelium of an interlamellar 
 junction in the gravid marsupium of Quadrula ebena, showing the large mucus cells and the nuclei of 
 several leucocytes (1) with which the epithelium has become infiltrated. 
 
 PLATE XVI. 
 
 Fig. 65. Station of the Bureau of Fisheries at North La Crosse, Wis., and steamer Curlew, used 
 in mussel investigations during summer of 1908. 
 
 Fig. 66. Interior of station at North La Crosse, equipped as a laboratory. 
 
 Fig. 67. Seining young black bass near La Crosse in a "lake" which had been filled by the over- 
 flow of the Mississippi River during the early summer. The fish thus obtained were artificially infected 
 with gloehidia. 
 
 PLATE XVII. 
 
 Fig. 68. Dredging for young mussels in a slough near La Crosse. 
 
 Fig. 69. The clamming outfit used in the mussel work on the Upper Mississippi River. The two 
 "crow-foot" dredges, with the mussels still clinging to the hooks just after a haul, are seen resting on 
 the stanchions. 
 
 Fig. 70. An old mussel bed near Muscatine, Iowa, buried under a foot or more of sand and mud 
 but exposed in cross section by a gully washed out by rains. The mussels are seen in situ embedded 
 in the wall of the gully. 
 
Bull. U. S. B. F., 1910. 
 
 Plate VI. 
 
Bull. U. S. B. F., 1910. 
 
 Platk VII. 
 
 Fig. 4. 
 
 Fig. 7. 
 
 Fig. S. 
 
Bull. U. S. B. F., 1910. 
 
 Plate VIII. 
 
 Fig. 9. 
 
 Fig. 13. 
 
 Fig. 10. 
 
 Fig. 14. 
 
 Fig. 15. 
 
 Fig. 1 
 
 Fig. 17. 
 
Bull. U. S. B. F., 1910. 
 
 Plate IX. 
 
 
 Fig. 24. 
 
Bull. U. S. B. F., 1910. 
 
 Plate X. 
 
 Fig. 27. 
 
 Fig. 28. 
 
 Fig. 30. 
 
 Fig. 29. 
 
 Fig. 31. 
 
Bull. U. S. B. F., 1910. 
 
 Plate XI. 
 
 Fig. 41. 
 
 Fig. 42. 
 
 Fig. 43. 
 
Bull. U. S. B. F., 1910. 
 
 Plate XII. 
 
 Fig. 45- 
 
 Fig. 44. 
 
 *P ' 'fa 
 
 Fig. 46. 
 
 Fig. 47. 
 
 Fig. 48. 
 
Bull. U.S. B. F.. 1910. PiATE XIIL 
 
 .^om^mmmw^ r.c, 
 
 re. 4lilP%^ ^ff||||§Cl 
 
 ^F *1 r- VT I fw.t.f 
 
 •—4@^^aiA A I J! 
 
 
 1) 
 
 '^0^^J^B^ 
 
 Fig. 50 
 
 Fig. 49- 
 
 Fig. 51. 
 
 Fig. 52. 
 
 Fig. 53. 
 
 
 
 Fig. 54. 
 
 Fig. 55. 
 
Bull. U. S. B. F., 1910. 
 
 Plate XIV. 
 
 Fig. 57. 
 
 Fig. 58. 
 
Bull. U. S. B. F., 1910. 
 
 Plate XV. 
 
 •®l 
 
 ftp SG I 
 
 fee- s n ©,,fc '■ ; .# 
 
 P"->, £* 
 
 rs ■'- Y ^ft?'---*J e > %<*''& &\ .'Sift- <3<D 
 Fig. 59. 
 
 '*>*».&' 
 
 Fig. 64. 
 
Bull. U. S. B. F., 1910. 
 
 Plate XVI. 
 
 Fig. 67. 
 
Bull. U. S. B. F., 1910. 
 
 Plate XVI I. 
 
 **>.'& 
 
 •'J^^r ^^H 
 
 
 
 
 *s 
 
 a f_3KS*^^lfi£l5i«f**Tf 
 
 
 
 
 ■■ v^5 
 
 
 
 
 tB ^t_ ' " 
 
 B 
 
 
 
 / * 
 
 
 
 
 
 
 p^ 
 
 ~*~~~~^ 
 
 WKZ-- _ 1~Z 
 
 
 
 
 mfO^~~d—£ 
 
 
 Fir. 70. 
 
LB |\| '12