THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY Editorial Board GARY N. CALKINS, Columbia University E. E. JUST, Howard University E. G. CONKLIN, Princeton University FRANK R. LlLLIE, University of Chicago E. N. HARVEY, Princeton University CARL R. MOORE, University of Chicago SELIG HECHT, Columbia University GEORGE T. MOORE, Missouri Botanical Garden LEIGH HOADLEY, Harvard University T. H. MORGAN, California Institute of Technology M. H. JACOBS, University of Pennsylvania G. H. PARKER, Harvard University H. S. JENNINGS, Johns Hopkins University F. SCHRADER, Columbia University ALFRED C. REDFIELD, Harvard University Managing Editor VOLUME LXXVI FEBRUARY TO JUNE, 1939 Printed and Issued by LANCASTER PRESS, Inc. PRINCE & LEMON STS. LANCASTER, PA. 11 THE BIOLOGICAL BULLETIN is issued six times a year. Single numbers, $1.75. Subscription per volume (3 numbers), $4.50. Subscriptions and other matter should be addressed to the Biological Bulletin, Prince and Lemon Streets, Lancaster, Pa. Agent for Great Britain: VVheldon & Wesley, Limited, 2, 3 and 4 Arthur Street, New Oxford Street, London, W.C. 2. Communications relative to manuscripts should be sent to the Managing Editor, Marine Biological Laboratory, Woods Hole, Mass., between June 1 and October 1 and to the Institute of Biology, Divinity Avenue, Cambridge, Mass., during the re- mainder of the year. Entered October 10, 1902, at Lancaster, Pa., as second-class matter under Act of Congress of July 16, 1894. LANCASTER PRESS, INC., LANCASTER, PA. / CONTENTS No. 1. FEBRUARY, 1939 PAGE PAPENFUSS, E. J., AND N. A. H. BOKENHAM The Fate of the Ectoderm and Endoderm of Hydra when Cultured Independently 1 HSIAO, SIDNEY C. T. The Reproductive System and Spermatogenesis of Limacina (Spiratella) retroversa (Flem.) v 7 REDFIELD, ALFRED C. The History of a Population of Limacina retroversa during its Drift Across the Gulf of Maine 26 BROWN, FRANK A., JR. Responses of the Swimbladder of the Guppy, Lebistes reticu- latus, to Sudden Pressure Decreases 48 TEWINKEL, Lois E. The Internal Anatomy of Two Phallostethid Fishes 59 KILLE, FRANK R. Regeneration of Gonad Tubules Following Extirpation in the Sea-Cucumber, Thyone briareus (Lesueur) 70 COSTELLO, HELEN M., AND DONALD P. COSTELLO Egg Laying in the Acoelous Turbellarian Polychoerus car- melensis 80 ZWILLING, EDGAR The Effect of the Removal of Perisarc on Regeneration in Tubularia crocea 90 MOORE, JOHN A. The Role of Temperature in Hydranth Formation in Tubularia 104 WOLFE, HAROLD R. Standardization of the Precipitin Technique and its Applica- tion to Studies of Relationships in Mammals, Birds and Reptiles 108 BARNES, T. CUNLIFFE Experiments on Ligia in Bermuda. VI. Reactions to com- mon cations 121 50652 iii iv CONTENTS No. 2. APRIL, 1939 PAGE HYMAN, LIBBIE H. Some Polyclads of the New England Coast, especially of the Woods Hole Region 127 KEPPEL, DOROTHY M. AND ALDEN B. DAWSON Effects of Colchicine on the Cleavage of the Frog's Egg (Rana pipiens) 153 WATERMAN, A. J. Effects of 2, 4-Dinitrophenol on the Early Development of the Teleost, Oryzias latipes 162 LOOSANOFF, VICTOR L. Effect of Temperature upon Shell Movements of Clams, Venus mercenaria (L.) 171 ROEDER, K. D. The Action of Certain Drugs on the Insect Central Nervous System 183 FRANK, JOHN A. Some Properties of Sperm Extracts and their Relationship to the Fertilization Reaction in Arbacia punctulata 190 DE LAMATER, ARLENE JOHNSON Effect of Certain Bacteria on the Occurrence of Endomixis in Paramecium aurelia 217 BRAUN, WERNER Contributions to the Study of Development of the Wing- pattern in Lepidoptera 226 MATTHEWS, SAMUEL A. The Relationship between the Pituitary Gland and the Gonads in Fundulus ... 241 WEISSENBERG, RICHARD Studies on Virus Diseases of Fish. II. Lymphocystis disease of Fundulus heteroclitus .... 251 WATERMAN, T. H., R..F. NUNNEMACHER, F. A. CHACE, JR. AND G. L. CLARKE Diurnal Vertical Migrations of Deep-water Plankton 256 HSIAO, SIDNEY C. T. The Reproduction of Limacina retroversa (Flem.) 280 CONTENTS v No. 3. JUNE, 1939 PAGE BIGELOW, HENRY B., AND WILLIAM C. SCHROEDER Notes on the Fauna above Mud Bottoms in Deep Water in the Gulf of Maine 305 DETHIER, V. G. Taste Thresholds in Lepidopterous Larvae 325 SAYLES, LEONARD P. Buds Induced from Implants of Nerve Cord and Neighboring Tissues in the Polychaete, Clymenella torquata 330 MORGAN, T. H. The Effects of Centrifuging on the Polar Spindles of the Egg of Chaetopterus and Cumingia 339 MENDOZA, GUILLERMO The Reproductive Cycle of the Viviparous Teleost, Neotoca bilineata, A Member of the Family Goodeidae. I. The Breeding Cycle 359 CLARKE, GEORGE L., AND DAVID D. BONNET The Influence of Temperature on the Survival, Growth and Respiration of Calanus finmarchicus 371 HARVEY, ETHEL BROWNE Development of Half-eggs of Chaetopterus pergamentaceus with special reference to Parthenogenetic Merogony 384 WELSH, JOHN H., AND HAROLD H. HASKIN Chemical Mediation in Crustaceans. III. Acetylcholine and Autotomy in Petrolisthes armatus (Gibbes) 405 COE, W. R. ' Sexual Phases in Terrestrial Nemerteans 416 BONNET, DAVID D. Mortality of the Cod Egg in Relation to Temperature 428 STILES, KARL A. The Time of Embryonic Determination of Sensoria and An- tennal Color and their Relation to the Determination of Wings, Ocelli and Wing Muscle in Aphids 442 CAUSE, G. F. Some Physiological Properties of Dextral and of Sinistral Forms in Bacillus mycoides Fliigge 448 Vol. LXXVI, No. 1 February, 1939 THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY THE FATE OF THE ECTODERM AND ENDODERM OF HYDRA WHEN CULTURED INDEPENDENTLY %R E. J. PAPENFUSS AND N. A. H. BOKENHAM 1 ^^•C^F^^ (From the Zoological Laboratory of the University of Cape Tozvn) It was recently found by Gilchrist (1937) that pieces of ectoderm of scyphozoan polyps, presumably the scyphistomas of Aur cilia, are able to regenerate small complete polyps ; while endodermal pieces are unable to regenerate although they may remain alive for several days. It is of considerable interest that a complete polyp can be regenerated from ectoderm alone as this phenomenon necessitates the formation of the endodermal layer from a tissue composed of cells of a completely different structure. In view of this fact it becomes of importance to know whether the newly regenerated tissue layer is formed by a trans- formation of certain of the differentiated cells ; or whether undifferenti- ated embryonic cells are activated to form anew the characteristic cell types of the lacking tissue, as Wilson and Penney (1930) found to be true for many cell types in regeneration from dissociated sponge tissue. Gilchrist did not study the origin of the endodermal layer in his regenerates but suggests that it may have been formed by the interstitial cells of the ectoderm. The present study was undertaken with the ob- ject of casting light upon this question. Hydra is a favorable organism for studying this problem as the endoderm can readily be separated from the ectoderm in the living condition (Papenfuss, 1934, pp. 227-228). Accordingly the two cellular layers were isolated and cultured inde- pendently. Both green and brown Hydra (species undetermined) were employed in each series of experiments. Attempts to Obtain Regeneration from the Ectodermal Layer In order to obtain a Hydra consisting only of ectoderm, it was first necessary to cut off the head of the specimen since it was not possible to remove the endoderm from the tentacles. The specimen was next everted and the endoderm removed. The Hydra now consisting only 1 The experimental phase of this study was carried out by E. J. Papenfuss while the histological part was done by N. A. H. Bokenham. 1 E. J. PAPENFUSS AND N. A. H. BOKENHAM of the ectodermal layer was then carefully examined under a higher magnification to make certain that fragments of endoderm had not re- mained behind. Hydras with buds were hot used as the endoderm could not be removed from these parts. In order to be certain that the entire endodermal layer is removed by this method, paraffin sections were made of individuals which had been so treated. Such sections show the specimens to be without endoderm. For the histological phase of the work, the specimens were fixed in formol-acetic-acid, stained in safranin or Heidenhain's haematoxylin, and fast green. The latter was found to be a good stain for the mesogloeal layer. After the endoderm had been removed from the Hydra which was to be cultured, the specimen was placed in a small Stender dish containing pond water. Filtered pond water was used throughout the experi- ments, and all the dishes and instruments were kept chemically clean. In order to ascertain if the conditions were favorable for the cultured tissue, a normal Hydra was placed in each dish. The data for this part of the paper were taken from 30 consecutive experiments. At the beginning of the study a few trial experiments were made which were not taken into consideration in the compilation of the results. In none of the experiments in which the ectoderm alone was cul- tured did a Hydra develop. The tissue always disintegrated, while the controls lived and appeared entirely normal. In 14 of the experiments the ectodermal tissue had completely disintegrated within 24 hours, in 8 of the experiments the tissue did not disintegrate until the second day, in 7 of the experiments a small piece of tissue persisted until the third day, and in one case a minute fragment of what appeared to be ecto- dermal tissue persisted until the seventh day. It may be noted that Gilchrist used only pieces of ectodermal tissue in his experiments while in the present work the entire ectodermal layer with the exception of the ectoderm of the head was used, as it seemed likely that regeneration would occur more readily from a large piece of tissue than from a small one. Since it was necessary in these experiments to remove the head of the Hydra and to evert the body in order to make the endoderm ac- cessible, a series of experiments were performed to determine whether a Hydra so treated can live and regenerate. The results from 35 ex- periments may be summarized as follows : In 30 of the experiments the Hydra lived and became normal; that is, the ectoderm again formed the external layer and tentacles developed. In the remaining five experi- ments, the Hydra died. In practically all the experiments in which the Hydra lived, the ectoderm was again on the exterior within 24 hours. INDEPENDENT CULTURE OF ENDODERM AND ECTODERM 3 The results of these experiments thus show that the majority of hydras will live under such conditions although an occasional one may die. It may be noted that in the experiments in which the Hydra did not live, the specimen was badly damaged while it was being everted. In the experiments in which the isolated ectodermal layer was cultured, such specimens were discarded as it was then difficult to remove the endoderm. From the first series of these experiments it may be concluded that the ectodermal layer of Hydra alone is not able to regenerate an indi- vidual, while the second series of experiments shows that the failure of the ectodermal layer to live and regenerate cannot be attributed to the treatment the Hydra had received prior to the removal of the endo- dermal layer. In order to determine whether the isolated endodermal tissue is capable of regeneration, the experiments set forth in the fol- lowing section were performed. Attempts to Obtain Regeneration from Endodermal Tissue It was not possible to obtain a complete endodermal layer by remov- ing the ectoderm as the latter is rigid and difficult to cut. It was accordingly necessary to evert the Hydra and to remove the endoderm by scraping it off with a fine glass needle. This tissue, however, dis- sociates very readily and comes off in small fragments. Since Peebles (1897) has shown that the smallest portion of a Hydra that is capable of regeneration is a sphere whose diameter is YG mm., it was considered necessary to fuse the fragments together so that a piece of tissue larger than this minimum size could be obtained for culturing. In the green species the fusion masses measured from 0.8 mm.-0.9 mm. in diameter and in the brown form they measured from 0.9 mm.-1.2 mm. in diame- ter. In a few experiments with both the green and the brown Hydra, the endodermal fragments from two individuals of the same species were allowed to fuse in order to obtain a larger mass of tissue. As the endodermal tissue of the green Hydra dissociates more readily than that of the brown, it was necessary to employ somewhat different methods for the two forms. When working with the green Hydra, the fragments were transferred by means of a fine pipette to a small hole in an agar bed. They were then placed as close together as possible so as to be in a position favorable for fusion. Only a small amount of water was allowed on the agar to prevent the fragments from floating apart. The slide was examined frequently and water added when necessary to prevent the tissue from becoming dry. If the hydras were fresh from the pond, complete fusion occurred in this form in approximately 15 E. J. PAPENFUSS AND N. A. H. BOKENHAM minutes. Hydras which had been in the laboratory for some time were unsuitable for this type of experiment unless they were from a well- established culture. When brown hydras were used it was found that larger fragments of endoderm could be obtained as the endoderm in this form dissociates less readily than in the green Hydra. It was then more practicable to place each individual fragment in contact with the other fragments as soon as it was removed from the Hydra. This was done with a fine glass needle. As much water as possible was withdrawn to prevent the fragments from floating apart. As the process of fusion takes longer in the brown Hydra, the slide was kept in a damp chamber, examined frequently, and water added when necessary to prevent the tissue from becoming dry. After the tissue had fused securely, the fusion mass was taken up in a pipette with a comparatively large opening and transferred to a small dish containing filtered spring water. A normal Hydra, to be used as a control, was placed in the dish with the endodermal fusion mass. This type of experiment is difficult to perform since the endoderm easily dissociates. The writers, however, carried through to completion 12 experiments. In no case was a new Hydra regenerated from the endodermal tissue, although the controls lived and appeared entirely normal. In the majority of experiments the fusion mass had com- pletely disintegrated within 24 hours while in a few cases minute frag- ments of the tissue could still be recognized after two or three days. The results of these experiments show that the endoderm of Hydra alone is not capable of regenerating an individual. It is of interest to note that similar fusion masses composed of both ectoderm and endo- derm will regenerate complete new hydras (Issajew, 1926; Papenfuss, 1932, 1934; Weimer, 1934). Discussion and Conclusions The object of the present study was to ascertain if a Hydra can be regenerated from either ectoderm or endoderm alone. The results of the experiments were entirely negative. Since Hydra has great regenerative abilities, being able to form a new individual from a small fragment of tissue (composed of both ectoderm and endoderm) only % mm. in diameter (Peebles, 1897), it is of considerable interest that a regenerate is not formed when practi- cally an entire cellular layer is cultured. In attempting to explain these results, the writers suggest that a regenerate did not develop because INDEPENDENT CULTURE OF ENDODERM AND ECTODERM 5 the differentiated cells of one tissue layer are unable to transform into the cell types characteristic of the lacking tissue layer, and the undiffer- entiated cells (the embryonic interstitial cells) do not become activated to regenerate this tissue. That differentiated cells do not transform into different cell types has been shown to be true previously by Wilson and Penney (1930), who found that in reunition of sponge tissue from dissociated cells, the specialized cells (the collar cells) which are carried over into the reuni- tion mass do not change their characteristics but become the collar cells of the new individual. Likewise in reunition in Hydra the specialized foot cells retain their identity and function as the foot cells of the new individual (Papenfuss, 1934). As the unspecialized interstitial cells of Hydra which are present in both the ectoderm and the endoderm are known to possess certain formative powers (Kleinenberg, 1872; Schneider, 1890; McConnell, 1932; and others), it seemed possible that these cells might regenerate anew the lacking tissue layer. However, judging from our results, it may be concluded that the formative powers of the interstitial cells are limited. According to Kanajew (1930), the interstitial cells play a subordinate role in regeneration in Hydra, the regenerated part, at least as regards the endoderm, being formed at the expense of the nearby differentiated tissue. In the sponge, however, the unspecialized mesen- chyme cells possess great formative powers, being able to reestablish very many of the characteristic cell types of a regenerate (Wilson and Penney, 1930). It should be noted that Ischikawa (1890) previously attempted to obtain Hydra regenerates from ectoderm alone. This author obtained ectodermal tissue by destroying the endoderm with acetic acid vapor. No regenerates were formed, however, and Ischikawa concluded that the interstitial cells were unable to give rise to the lacking tissue layer. Schulze (1918) suggests that the failure of the ectodermal tissue to regenerate was owing to the lack of nourishment normally supplted by the endoderm. Schulze's conclusion, however, loses support by the fact that isolated endodermal tissue also fails to regenerate, as was found in the present study. The results of our experiments are at variance with those of Gil- christ (1937), who found that pieces of ectoderm of a scyphistoma are able to regenerate complete new individuals. As the disposition of the endoderm of a scyphozoan polyp is somewhat complicated by the inva- sion of the ectodermal cells in the four taeniolae, it is possible that Gilchrist did not succeed in removing all the endoderm from the fragments. 6 E. J. PAPENFUSS AND N. A. H. BOKENHAM Acknowledgment The writers wish to thank Professor W. A. Jolly for the loan of a dissecting microscope and for granting them the facilities of the De- partment of Physiology where part of the work was done. They further wish to thank Dr. M. A. Pocock for aid in the collecting of material. BIBLIOGRAPHY 2 GILCHRIST, FRANCIS G., 1937. Budding and locomotion in the scyphistomas of Aurelia. Biol. Bull, 72 : 99. *ISCHIKAWA, C., 1890. Trembley's Umkehrungsversuche an Hydra nach neuen Versuchen erklart. Zeitschr. f. wissensch. ZooL, 49 : 433. ISSAJEW, W., 1926. Studien an organischen Regulationen. (Experimentelle Un- tersuchungen an Hydren.) Arch. f. Entw.-inech. Organism., 108: 1. *KANAJEW, J., 1930. Zur Frage der Bedeutung der interstitiellen Zellen bei Hydra. Arch. f. Entw.-mech. Organism., 122 : 736. KLEINENBERG, N., 1872. Eine anatomisch entwicklungschichtliche Untersuchung. Leipzig. McCoNNELL, C. H., 1932. The development of the ectodermal nerve net in the buds of Hydra. Quart. Jour. Micr. Sci., 75 : 495. PAPENFUSS, E. J., 1932. Experimental studies on fusion in Hydra (Abstract). Anat. Rec., 54: 54 (Suppl.). PAPENFUSS, E. J., 1934. Reunition of pieces in Hydra, with special reference to the role of the three layers, and to the fate of differentiated parts. Biol. Bull., 67 : 223. PEEBLES, FLORENCE, 1897. Experimental studies on Hydra. Arch. f. Entzv.-mech. der Organism., 5 : 794. SCHNEIDER, K. C., 1890. Histologie von Hydra fusca mit besonderer Beriicksich- tigung des Nervensystems der Hydropolypen. Arch. mikr. Anat., 35 : 321. SCHULZE, P., 1918. Die Bedeutung der interstitiellen Zellen fur die Lebens- vorgange bei Hydra. Sitsgsber. Gcs. Naturforsch. Freunde Berlin, Nr. 7 : 252. WEIMER, B. R., 1934. The physiological gradients of Hydra. III. Reconstitution of masses of dissociated pieces. Physiol. ZooL, 7 : 212. WILSON, H. V., AND J. T. PENNEY, 1930. The regeneration of sponges (Micro- ciona) from dissociated cells. Jour. Exper. ZooL, 56 : 73. 2 The authors did not have access to the papers marked with an asterisk but have seen the abstract of Kanajew's work in Biological Abstracts, Vol. 5 (II), 1931, p. 2952; and have read Schulze's (1918) account of the work of Ischikawa. THE REPRODUCTIVE SYSTEM AND SPERM ATOGENESIS OF LIMACINA (SPIRATELLA) RETROVERSA (FLEM.) 1 SIDNEY C. T. HSIAO (From the Biological Laboratories, Harvard University and the Woods Hole Oceanographic Institution,,- Woods Hole, Massachusetts) INTRODUCTION Liinacina retroversa was described as early as 1823 by Fleming un- der the name of :f Heterofusus rctroversus," but very little is known about its internal structure, its method of reproduction and its life history. Writing in 1924 — one hundred years later — Bigelow men- tioned that ' ' Nothing is known of the reproduction of Limacina retro- versa in the Gulf of Maine. . . ." The embryonic development and the reproductive system of several species of Pteropoda have been de- scribed but no exact information on the species L. retroversa is available. Van Beneden described the structure of the reproductive system of L. arctica in 1839 but, as pointed out by Meisenheimer (1905), he made the mistake of calling the ovotestis the ovary, the vesicula seminalis the testis and the accessory glands the " gland prostatique." Studies of the embryonic development of pteropods have been made by Gegenbaur (1853, 1855), Krohn (1856), Miiller (1857), Fol (1874, 1875) and Knipowitsch (1891). Meisenheimer (1905) described the comparative morphology of Pteropoda and compared the genus Liinacina with other Euthecosomata. He considered Limacina as a general group and made no mention of any species. Meisenheimer also appraised the earlier work and observations of Cuvier, d'Orbigny, Van Beneden, Gegenbaur, Souleyet, Pelseneer, Knower, etc. Bonnevie (1916) made further ob- servations on the reproductive organs of Cuvierina and recently Zarnik (1911) determined the chromosome number of several species of ptero- pods. Up to the present no one has studied the entire life history of any species of Pteropoda and it has been unknown whether any ptero- pods, like some other mollusks, change their sex. The reproductive system of Limacina retroversa, in particular, has not been studied. It 1 The name " Limacina retroversa " has been in use until 1930, when Sherborn pointed out that Cuvier (1817) used the vernacular name " Limacine " and he restored de Blainville's " Spiratella" in place of Lamarck's (1819) "Limacina." Thiele (1931) also used " Spircrtella." However, as " Liinacina" has the advantage of being more commonly used and better known it is retained in this paper. - Contribution No. 197 from the Woods Hole Oceanographic Institution. 7 8 SIDNEY C. T. HSIAO is the purpose of this paper to describe the reproductive system and spermatogenesis of Liinacina retrovcrsa as a basis for a report on the reproductive history of this species to appear in a subsequent paper. MATERIAL AND METHODS The material was collected by the research vessel " Atlantis " of the Woods Hole Oceanographic Institution during a series of cruises in the Gulf of Maine between June, 1933 and September, 1934. The speci- mens were caught by vertical hauls made with a 1.5-meter Heligoland net drawn from a level near the bottom up to the surface and were fixed on board the ship in 2 per cent formalin in sea water. Each specimen used for making permanent mounts was decalcified in 1 per cent acetic acid for a few minutes. In some cases the acid in the commercial formalin used in fixation had decalcified the specimens. The decalcified animal was then stained in toto for about one minute in a 0.1 per cent aqueous solution of Erhlich's methylene blue chloride to aid in orienting it in the paraffin block for sectioning. The stained speci- men was dehydrated in two grades of 1-4 dioxene in a calcium tower and directly imbedded in 53°-55° C. paraffin. Each individual was cut in serial longitudinal sections five micra in thickness which were mounted in order on slides. The vital stain was removed when the sections were hydrated after the removal of paraffin with xylene. Standard methods of staining with Heidenhain's iron haematoxylin and orange G were used in the final staining process. Stained sections were cleared in xylene and mounted in Canadian balsam. In order to reconstruct the different parts of the reproductive system sections from suitable individuals were drawn with a camera lucida. The tracings of every fifth or tenth section were made with a wax pencil on pieces of glass. Reconstruction of the reproductive system was made by superimposing the glass plates in order so as to determine the size and shape of each organ. These tracings were compared with Meisen- heimer's plates and descriptions (1905) of similar sections of the repro- ductive apparatus of Euthecosomata. Meisenheimer's terminology of the different parts of the regenerative system was followed. In the study of the spermatogenesis of Liuiaciua retrovcrsa the method for carmine smears was used in addition to the examination of permanent mounts stained in Heidenhain's haematoxylin. The speci- mens used for making carmine smears were collected in the summer of 1937 near Woods Hole. As soon as the animals were taken from water they were fixed for twenty minutes in a mixture of one part of acetic acid and three parts of absolute alcohol. Each Lhnacina was sepa- REPRODUCTIVE SYSTEM OF LIMACINA rated from other planktonic material and stored in 70 per cent alcohol. It was necessary to change the preservative several times in order to get rid of the precipitate brought down by alcohol from sea water. The gonads from several individuals were dissected out under a binocu- lar microscope and torn up into very small pieces. They were then stained with cytological aceto-carmine for a few minutes, covered with a coverglass and heated over a small flame without boiling. The slide with the material was then pressed between blotting papers to separate the cells and to remove the excessive stain. Such preparations showed the cellular structures very clearly and were used for spermatogenesis studies. Living animals have been kept in the laboratory for a week to ten days for observation of spawning and egg-development. OBSERVATIONS The Reproductive System The reproductive system of Limacina rctrovcrsa consists of a bi- sexual gonad, a gonoduct, accessory glands and a penis. The Gonads. — The bisexual gonad or ovotestis constitutes the pri- mary sexual organ. In a mature individual it is a spiral structure situ- ated posterior to the liver and occupies the posterior or last three or four whorls of the shell with whose cavity the smooth, non-lobated coils of the ovotestis conform (Fig. 1). In specimens preserved in formalin the gonad appears, after decalcification, granular and pinkish in contrast to the yellowish-green liver with which it is continuous. In living speci- mens the gonads can also be distinguished from the greenish liver through the transparent shell. In sectional view, the cortical and apical regions of the mature gonads are generally occupied by the female repro- ductive cells and the medullary portion by the spermatic tissues. The spermatic tissues are apparently arranged in tubules in the medulla of the ovotestis, but in well-developed males they may extend to the cortex with masses of mature spermatozoa obliterating the tubular arrangement of the reproductive cells. In large females, on the other hand, the mature ova may be found in the medullary region just before they pass into the hermaphroditic duct. The proportion of masculine and femi- nine reproductive cells varies with the season as well as with the size of the animal. Three different types of gonads can be distinguished. (A) Sex- ually undifferentiated gonads all of which occur in small-sized specimens ; (B) "pure male" gonads which belong to somewhat larger animals but never to animals greater than 1.1 mm. in diameter across the largest 0» /V^ O ^»<^ ' AR 10 SIDNEY C. T. HSIAO whorl of the shell; and (C) hermaphroditic gonads, or ovotestes, char- acteristic of all larger individuals. A. Sexually undifferentiated gonads. — In very small specimens the gonad is sexually undifferentiated in that no definite male or female reproductive cells can be recognized. This type of gonad corresponds with the "indefinite gonads" of Coe (1931) and Orton (1926). The whole gonad is small, nearly spherical and surrounded by the liver. The pregerminative cells, all of nearly equal size, form a solid mass which is enclosed within a thin fibrous Ancel's layer consisting of elongated cells FIG. 1. Camera lucida drawing of Limacina rctrovcrsa (X 30), showing the gonad and the rest of the body. C. Cavity between shell and ovotestis. F. " Flapper " or " foot." G. Gut. L. Liver. O. Ovotestis. S. Shell opening. with prominent nuclei. The cellular structure of the pregerminative cells is not easy to make out, for in specimens fixed with commercial formalin and stained with haematoxylin the nuclei are diffused and the structure not clear. In this, as in the other types of gonads, the fixative used tends to make the cells indistinct from each other so that they have the appearance of a syncytium. A sectional view of a young indefinite gonad is shown in Fig. 2, A. B. " Pure male " gonads. — All Limacina retroversa larger than 0.85 mm. in diameter show sexual differentiation, for either male or female REPRODUCTIVE SYSTEM OF LIMACINA 11 germ cells are discernible. But nearly all, if not all, of the very small sexually differentiated individuals are protandric. Half of the pro- tandric animals do not show any ovogenesis at all. Individuals pos- sessing this type of gonad correspond to Coe's (1933) "true males" or Orton's (1909) "pure males." In this description they will be called " pure males." The " pure male " type of gonad is confined to small-sized Limacina rctroversa. No individuals larger than 1.1 mm. in diameter show complete maleness in their gonadal organization. In Fig. 2, B a gonad without any feminine reproductive cells distinguishable from the male cells is shown. In this diagram a few clusters of sper- matozoa and a few groups of spermatids are seen while many of the other cells are undergoing active spermatogenesis. C. Hermaphroditic gonads. — Besides "pure males" sexually dif- ferentiated Limacina have gonads with various proportions of masculine and feminine types of tissue and show different signs of sexual activity. For convenience of description and comparison, these larger specimens, possessing ovotestes, can be divided into classes in terms of the per- centage of masculine or feminine tissue in the gonad. Individuals with more than 50 per cent of the tissue of their gonads consisting of mascu- line reproductive cells, showing spermatogenesis, are designated as her- maphroditic males and those with less than 50 per cent of the contents of their gonads made of male reproductive cells are called hermaphro- ditic females. On this percentage basis the gonads can be arranged from extreme masculine to pronounced feminine types. Gonads of individuals which are classified as hermaphroditic males are shown in Fig. 2, C and D. Figure 2, C shows the gonad of a mature hermaphroditic male with more than 75 per cent of its reproductive tis- sues consisting of masculine cells. This gonad occurred in a Limacina 1.2 mm. in diameter. In contrast to the "pure male " type of gonad, a peripheral layer of ovarian tissue is present, but it amounts to less than 25 per cent of the total reproductive cells in the organ. In the medul- lary portion clusters of mature spermatozoa are present while in the cortical region a small number of egg cells are developing. Individuals possessing this type of gonad may be of any size but most of them are comparatively small. They are smaller than the pronounced feminine type but larger than the " pure males." Figure 2, D shows the ovotestis of an individual belonging to the class having one-half to three-quarters of the gonad occupied by male tissue. In this ovotestis mature cells of both sexes are seen. At a point near the beginning of gonoduct, next to the retractor muscles, in the largest whorl of the spiral organ, mature ova and spermatozoa are seen ready to leave the sex gland. This simultaneous presence of ma- B 7843 37 9 FIGURE 2 REPRODUCTIVE SYSTEM OF LIMACINA 13 ture ova and spermatozoa near the gonoduct, or the presence of mature cells of both sexes along it, indicates that the animal is functioning both as male and female. Individuals of this type are called functional her- maphrodites, a name commonly used to designate this kind of repro- ductive individual among mollusks. Animals with less than half of the gonadal tissues consisting of male reproductive cells, which are classified as hermaphroditic females, can also be divided into two groups in a similar manner : (a) hermaphroditic females with gonads containing 50-25 per cent of male tissue, and (/?) hermaphroditic females with less than 25 per cent of male tissue in the gonad. These two types are represented in Fig. 2, E and F. As a general rule, gonads which are predominantly feminine are in a mature stage, containing mature ova and very often both mature ova and sper- matozoa. The Limacina possessing these gonads are generally larger than the masculine types. Part of a gonad with 25-50 per cent of its tissues consisting of male germ cells is shown in Fig. 2, E. This gonad is in a mature condition, for mature ova are found in the medullary portion of the ovotestis near the beginning of the gonoduct instead of in the cortical region where ovocytes of various stages of development are seen. Empty spaces are left in the gonad apparently by the spermatozoa which have vacated the FIG. 2. Drawings of different types of gonads of Limacina retroversa. A. Section of a complete gonad which is sexually undifferentiated. B. A " pure male " gonad in which no ovocytes are seen — section of a complete gonad. C . A hermaphroditic male gonad which has more than 75 per cent of its tissues made of male germ cells. Only the first or largest whorl is shown, the other two whorls being indicated by outlines. D. First or largest whorl of a hermaphroditic male gonad with 50-75 per cent of its tissues consisting of male germ cells. E. Right half of the first or largest whorl of a hermaphroditic female gonad which contains 25-50 per cent of male tissue. F. Right half of the first or largest whorl of a hermaphroditic female gonad with less than 25 per cent of its tissues made of male germ cells. 1. Nucleus of pregerminative cell. 2. Nucleus of spermatogonium. 3. Primary spermatocytes. 4. Secondary spermatocytes. 5. Spermatids. 6. Spermatozoa. 7. Ovocyte. 8. Ovum. 9. Outline of second whorl of ovotestis. 10. Outline of third whorl of ovotestis. 11. Outline of left half of first whorl of ovotestis. 12. Fibrous connective tissue surrounding gonad. 13. Retractor muscles. 14 SIDNEY C. T. HSIAO gonad previously. Figure 2, F show's a gonad which is even more pre- dominantly feminine than that shown in Fig. 2, E. Less than 25 per cent of this ovotestis is made up of masculine reproductive cells. In contrast to masculine types, the pronounced feminine hermaphro- • \ ' • 2*-T-r-« cJjjJrvCJ' -*ii» V*"** £*wA' ~:-J:---.«"&**rxi v A &$%$W FIG. 3. Photomicrographs of sections of gonad of Limacina retrovcrsa, show- ing hermaphroditic male and female gonads. A. Hermaphroditic male gonad in which more than 75 per cent of the tissues con- sists of male reproductive cells. B. Hermaphroditic female gonad in which 25-50 per cent of the gonadal tissues consists of male germ cells. C. Further enlarged view of a section of the second whorl of a hermaphroditic male gonad with 50-75 per cent male tissue, showing the germ cells in greater detail REPRODUCTIVE SYSTEM OF LIMACINA 15 dites possessing very little male tissue are limited to medium and, more generally, to large-sized Limacina. Another difference between' the ex- tremely male and female types is that while " pure males " are common, forming about one-half of the protandric males, " pure females " have never been observed. All female Limacina retroversa are hermaphro- ditic. FIG. 4. Diagram of the reproductive system of Limacina retroversa. 1. Ovotestis. 2. Hermaphroditic duct or gonoduct. 3. Receptaculum seminis. 4. Shell gland. 5. Albumin gland. 6. Vesicula seminalis. 7. Out-going duct. 8. Genital opening. Figure 3 shows some of the photomicrographs of sections of the gonad of Limacina upon which the diagrammatic drawings of Fig. 2 are based. Figure 3, A is a hermaphroditic male gonad with more than 75 per cent of its tissues consisting of male germ cells and Fig. 3, B, a hermaphroditic female gonad with 25-50 per cent of male tissue in it. Figure 3, C is a further enlarged view of a masculine gonad with 16 SIDNEY C. T. HSIAO 50-75 per cent male tissue showing the arrangement of the germ cells in greater detail. The Accessory Organs of Reproduction. — The accessory organs of the reproductive system are diagrammatically represented in Fig. 4, which is a composite picture based upon the reconstruction from serial sections of the whole animal such as those shown in Fig. 5. The gonoduct begins as a hermaphroditic duct for the passage of both ripe spermatozoa and ova. It opens into the ventral side of the anterior end of the spiral ovotestis and then turns to the right and passes forward, along the right side of the digestive track, to the anterior end of the body. It begins as a single tube with its wall composed of cuboidal cells, but later branches out into a pear-shaped structure (called "6" in Figs. 4 and 5). After meeting the accessory glands the gonoduct continues as the out-going duct. This part of the duct is much convoluted, leading to the external opening of the reproductive system on the right side of the mouth (8 in Fig. 4). This tube is com- posed of cuboidal epithelial cells with prominent nuclei and the cyto- plasm of these cells contains a large quantity of secretory granules. The accessory glands, like those of other pteropods, are very compli- cated structures. The largest gland, according to the terminology of Meisenheimer, is the shell gland. It is divided into three or four lobes occupying a great portion of the right side of the anterior part of the body of the animal. It opens into the hermaphroditic duct as well as the albumin gland (see Fig. 5, D and £). The shell gland is composed for the most part of large secretory cells, and tall, ciliated, columnar cells. The secretory cells are very abundant and obvious and with nuclei which are generally large and ellipsoidal, lying on the base of the cells away from the lumen of the gland, while the portion near it is filled with secretion. Perhaps the ovum stays longest in this gland on its way through the various parts of the reproductive system, for ova have been more often observed in this organ than in the others. The albumin gland (Fig. 5, B to D) is a snail-like coiled gland asso- ciated with the shell gland but smaller than it. A diagram of its outline is shown in Fig. 4. The cells of the albumin gland are largely secretory. Their inner edge, which faces the lumen, is filled with transparent secre- tion when seen in sections stained with iron haematoxylin. The vesicular seminalis (Figs. 4 and 5) is an elongated, pear-shaped organ connected with the hermaphroditic duct. It is lined with cuboidal epithelial cells. Inside this organ masses of mature spermatozoa are seen in the case of mature Limacina. The receptaculum seminis is connected with the basal portion of the out-going duct. It consists of two portions. The first portion is much REPRODUCTIVE SYSTEM OF LIMACINA 17 B I FIG. 5. Diagrams of longitudinal sections of the accessory organs of the re- productive system of Limacina retroversa (based upon camera lucida tracings from an animal cut into serial sections which are 5 micra thick). Figure A. Section No. 90. B. " " 100. C. " 115. D. " " 124. E. " 135. M. Mantle gland. P. Penis (in Figs. A-G). 3. Receptaculum seminis (Figs. C-7). 4. Shell gland (in all the figures). Figure F. Section No. 155. G. " " 165. H. " " 175. /. " 195. 5. Albumin gland (Figs. B-D} . 6. Vesicula seminalis (Figs. C-G). 7. Out-going duct (Figs. F and G). 18 SIDNEY C. T. HSIAO folded, with a thick wall composed of columnar cells, while the second portion is much thinner, being made of cuboidal cells and baggy in appearance (Figs. 4 and 5). Spermatozoa observed in the folded por- tion stain deeply with iron haematoxylin and are normal in structure, but those within the second portion stain only very lightly with the same B K N I'H.. 6. Spermatogenesis of Lhiiacina rctroi'crsa. REPRODUCTIVE SYSTEM OF LIMACINA 19 dye and show signs of degeneration and masses of disintegrating mate- rial are often observed. The penis is practically independent of the rest of the reproductive system both in position and in structure. In well-developed adults it is a very large organ lying dorsad to the rest of the genital structures in the anterior portion of the body near the mantle gland (P in Fig. 5). This organ is very much folded on its outer surface when not distended. It has its own secretory glands and retractor muscles. The spermatozoa which come out of the gonoduct run along the ciliated seminal groove towards the penis along the right side of the dorsal region of the " fin " or " flapper." GENERAL FEATURES OF GAMETOGENESIS The development of the male and female reproductive cells has been examined from fixed material cut into serial sections and stained with iron haematoxylin as well as from fresh specimens treated with the carmine smear technique. The steps in spermatogenesis have been ascertained, and it has been found that the maturation of the male germ cells is essentially the same as that of other gastropods. The pregerminative cells grow and multiply to form the primary spermatogonia (Fig. 6, A). The primary spermatogonia (Fig. 6, B) EXPLANATION OF FIGURE 6 A. Primary spermatogonium. B. Metaphase plate of primary spermatogonium showing the diploid number of 24 chromosomes. C. Secondary spermatogonium. D. Primary spermatocyte, beginning of leptotene threads. E. Primary spermatocyte, showing tetrad condition. F. Metaphase of first meiotic division, profile. G. Polar view of anaphase of first meiotic division. //. Early anaphase of second meiotic division, profile. /. Young spermatid. /. Chromosome plate of early anaphase of first meiotic division showing the hap- loid number of 12 chromosomes. K. Chromosome plate of early anaphase of second meiotic division, showing 12 chromosomes. L. A group of spermatids around a central cell. M. Same as L, with spermatids not completely enclosing the central cell. N. Metaphase plate of a somatic cell, showing the diploid number of 24 chromo- somes. O. Separate spermatids undergoing nuclear elongation, besides them a mass of discarded cytoplasm from other older spermatids. P. Cross-sections of spermatids showing the central canal. Q. Elongated spermatid, showing formation of tail filament. R. Spermatid further elongated than Q. S. Nearly mature spermatozoon with discarded cytoplasm beside it. T. Mature spermatozoon. 20 SIDNEY C. T. HSIAO divide by mitosis to give rise to the secondary spermatogonia (Fig. 6, C) which are smaller in size and with a higher nucleus-cytoplasm ratio than the primary. Generally the nucleus is placed on one side of the cell. Figure 6, B is a metaphase plate of a primary spermatogonium showing 24 chromosomes. The secondary spermatogonia differentiate into pri- mary spermatocytes by mitosis. The primary spermatocytes are identi- fied by the fact that they are nearly always grouped together in the gonad. At the first meiotic division the chromatin of the primary spermatocyte breaks up and forms loose masses and then changes into leptotene threads. These threads, as in some pulmonates, for example, Succina ovalis, Ostrea lurida, form such a compact mass that it is diffi- cult to follow the steps in conjugation and tetrad formation (that is, parasynapsis and diakinesis). Figure 6, D shows the beginning of the leptotene threads with the nucleolus still visible and Fig. 6, E shows a protetrad condition of the chromatin material after parasynapsis. The tetrads of early diakinesis in the form of circles and crosses are arranged on the surface of the nuclear membrane and take a very dark color in staining. A profile of the metaphase of the first meiotic division is shown in Fig. 6, F, while Fig. 6, G is a polar view of the anaphase of the same division. Figure 6, / is a much enlarged polar view of the early anaphase of this division showing the reduced number of 12 chromosomes. During telophase the chromosomes draw together and form a compact mass in the secondary spermatocytes. From the fact that a resting stage of the secondary spermatocyte is not seen it is inferred that this stage is most probably absent. This situation has been reported by Furrow (1935) for Valvata tricarinata. Cells undergoing the second meiotic division are smaller in size than those of the first. Figure 6, H is a profile of the early anaphase of the second meiotic division, and Fig. 6, K is a much enlarged view from one pole of the chromosome plate at an early anaphase of this division showing the haploid number of twelve chromosomes. During the telo- phase of this division the chromatin material becomes reticular as the nuclear membrane reappears and later condenses into a very dark stain- ing structure which eventually becomes the dark nucleus of the sperina- tid. In Fig. 6, / a newly-formed nucleus of a younger spermatid is shown. Two of the spermatids of Fig. 6, M are also in this condition. These figures show the formation of reticular structures within the nuclear membrane and the condensation of the chromatin material in forming the dense nucleus of the spermatid. The spermatids are generally arranged around a large degenerating cell which may be homologized with the " nurse cell " reported in most mollusks, but no cytoplasmic connection between it and the spermatids REPRODUCTIVE SYSTEM OF LIMACINA 21 has been observed (Fig. 6, L and M}. In spermiosis the darkly- staining nucleus elongates first at one point so as to assume an acutely ovate outline (Fig. 6, L). A few of the various shapes assumed by the nucleus during its metamorphosis into the head of the mature spermatozoon are shown in Fig. 6, 0. In association with the change of the nucleus, the cytoplasm elongated and the axial filament of the tail-piece appears (Fig. 6, Q). In the cross-section of an old spermatid of this stage the central canal through the nucleus can be seen as a light area in the center (Fig. 6, P). The spermatid is transformed into a mature spermatozoon by further elongation of the cytoplasm together with the growth of the axial filament and lengthening of the nucleus followed by a discarding of a greater portion of the cytoplasm (Fig. 6, Q to S}. The head-piece of the spermatozoa is slender and pointed at the anterior end. It is formed by the nucleus. The tail-piece con- sists of an axial filament and a very slender tail membrane formed by the cytoplasm. The middle piece of the spermatozoon is not clearly seen and its development could not be followed in our material. A mature spermatozoon (Fig. 6, T) is about 30 micra long and its head about half as long as its tail-piece. During spermiosis the spermatids move away from a spherical arrangement about a central " nurse cell " and become irregularly placed as seen by a comparison of Fig. 6, L and Fig. 6, O. When the spermatozoa are mature they are placed closely together side by side with their heads pointing in the same direction, but they do not form " sperm balls " or " sperm morulae." In no case has a " nurse cell " been observed to which the heads of the spermatozoa are attached as has been reported for some pulmonates and prosobranchiates. In specimens used in this study no atypical spermatogenesis has been observed. The only element in the gonad which might suggest apyrene spermatozoa is that shown in Fig. 6, 0. It consists of a mass of elon- gated, enucleated structures which stain very lightly. But it appears more probable that they are discarded portions of the cytoplasm of the transforming spermatids rather than apyrene spermatozoa for no steps comparable to the atypical spermatogenesis leading to the formation of these structures have been observed. The diploid number of chromosomes is 24. This number was first deduced from observations on the secondary spermatocytes which ex- hibit the haploid number of 12 chromosomes and was again checked against the chromosome plates of somatic cells. A metaphase chromo- some plate of a somatic cell drawn with a camera lucida from a carmine smear is shown in Fig. 6, TV and it may be compared with the chromo- some plate of the mitotic division of the spermatogonium shown in Fig. 6, B. The chromosome number wras further confirmed by the chromo- 22 SIDNEY C. T. HSIAO some plates of the second miotic division which also shows (Fig. 6, K) the haploid number of 12 chromosomes. It has been observed in young gonads (Fig. 2, B) that there were at first only a few clusters of spermatozoa in the whole organ. As the animal grows the number of clusters of spermatozoa increases and even- tually, in a mature Limacina, the mass of spermatozoa constitutes nearly all the male tissue in the gonad. It has been reported in other mollusks that each cluster of spermatozoa descends from one primary spermato- gonium. If this is also the case in Limacina, then it appears from this gradual increase of the number of clusters of spermatozoa with the B FIG. 7. Drawings of the female reproductive cells of Limacina retrovcrsa (Fleming). A. An oogonium. B. A growing ovocyte with homogeneous cytoplasm. C. A nearly mature ovum with a number of darkly-staining granules in the cytoplasm. growth of the animal that the spermatogonia do not all develop at the same time, but that they follow each other in their maturation. The female cells as a whole evidently develop slower than the male so that the animal generally assumes a protandric condition. The proc- ess of maturation of the female reproductive cells cannot be readily observed in our specimens of Limacina retroversa. When the different ovocytes in the ovotestis are compared it is seen that the growth of the cells is associated with the deposition in the cytoplasm of large granules which stain very dark with haematoxylin. The nucleus, on the other hand, takes stain less and less easily and eventually becomes much lighter as the ovocytes grow in size. A mature ovum measures 50 to 70 micra, REPRODUCTIVE SYSTEM OF LIMACINA which represents an increase of five to seven times in diameter over the first ovocyte. Figure 7 shows an oogonium together with a growing ovocyte and a nearly mature ovum. The change in size is quite notice- able. As in the case of spermatogenesis, the ova do not all mature simul- taneously, but only a few are fully developed at a time. When once started, oogenesis seems to continue throughout the remainder of life of the individual. Female germ cells before and during maturation are located on the peripheral region of the gonad, but ova, when mature, migrate to the central portion of the gonad and thence to the opening of the gonoduct (Figs. 2, D and 2, £). Fertilization probably takes place during the passage of the ova through the gonoduct. Female reproductive cells of different ages are present in the gonad, but the maximal number of mature ova observed in a single five micra section of one whorl of the ovotestis is not greater than ten to twenty. The largest number of ova observed in the gonoduct at the time of fixa- tion is about four or five. These observations, together with the fact that Limacina rctroversa kept in the laboratory lay a few eggs at a time, indicate that this animal does not spawn out after a brief period of intensive reproductive activity, but that spawning is more or less pro- tracted and continuous after it has once began. DISCUSSION AND SUMMARY From the above description of the maturation of the germ cells of Limacina retrovcrsa it will be seen that the process of spermatogenesis in this species agrees with those of other gastropod mollusks. Von Brunn (1884) first observed atypical spermatozoa and considered them to be rudimentary ova. Meves (1903) and Hertwig (1905), however, have described the steps in atypical spermatogenesis and tried to eluci- date the morphological and physiological significance of this process and its product. Reinke (1914) in Strobus bituberculatus and Kuschake- witsch (1910, 1911) in Conns and Vermctus, Gould (1917) in Crcpi- dula plana, and other authors, made similar studies. Recently Hick- mann (1931) described atypical spermatogenesis in Succina oval is, Furrow (1935) in Valvata tricarinata, Coe and Turner (1938) in Mya arenaria. But as far as it can be ascertained there is no atypical sper- matogenesis in Limacina retrovcrsa. This species seems to belong to the group of mollusks that do not show atypical spermatogenesis. Nurse-cells have been reported in many forms of mollusks during their maturation, but in Limacina rctroversa the arrangement of the spermatids and newly matured spermatozoa about a central cell may 24 SIDNEY C. T. HSIAO suggest the presence of nurse-cells though no cytological connection has been observed between these and the germ cells. Mature spermatozoa form balls or " sperm morulae " (according to Orton, 1926) in certain mollusks while in others, though of the same genus, they do not. Thus, Ostrea lurida and Ostrea virginica, in this country, have clustered and separate mature spermatozoa respectively. Limacina retroversa have separate mature spermatozoa. The accessory organs of the reproductive system of Limacina retro- versa agree very closely with the description given by Meisenheimer, who did not mention which species were used as representatives of the genus Limacina in his comparative studies. However, if Meisenheim- er's work on the secondary sexual organs is taken as a typical description of the genus Limacina, then Limacina retroversa may be considered as closely resembling the typical forms in the structures of the accessory sexual organs. BIBLIOGRAPHY BIGELOW, H. B., 1926. Plankton of the off-shore waters of the Gulf of Maine. Bull. U. S. Bur. Fish.. 40 (Part II) : 1 (1924). BONNEVIE, K., 1916. Mitteilung iiber Pteropoden. L. Beobachtungen iiber den Geschlechtsapparat von Cuvierina columnella Rang. Jena Zcitschr. Nat., 54 : 245. BOURNE, G. C, 1889. The generative organs of the oyster. (Abstract of a paper by Dr. P. P. C. Hoek.) Jour. Mar. Biol. Ass'n, 1 : 268. COE, W. R., 1931. Spermatogenesis in the California oyster (O. lurida). Biol. Bull., 61 : 309. COE, W. R., 1933. Sexual phases in Teredo. Biol. Bull., 65 : 283. COE, W. R., AND H. J. TURNER, JR., 1938. Development of the gonads and gametes in the soft-shell clam (Mya arenaria). Jour. Morph., 62: 91. FLEMING, J., 1823. On a reversed species of Fusus (Fusus retroversus). Mem. Werner. Soc., 4 : 498. FOL, H., 1874. Note sur le developpement des Mollusques Pteropodes et Cephalo- podes. Arch. d. Zool. E.vpcr. et Gen., 3 : 33. FOL, H., 1875. fitudes sur le developpement des mollusques. I. Sur le developpe- ment des pteropodes. Arch. Zool. Expcr. et Gen., 4: 1. FURROW, C. L., 1935. Development of the hermaphrodite genital organs of Valvata tricarinata. Zcitschr. f. Zellforschung., 22 : 282. GEGENBAUR, C., 1853. Bau der Heteropoden u. Peteropoden. Zcitschr. f. wiss. Zool.. 4: 334; 369. GEGENBAUR, C., 1855. Untersuchungen tiber Pteropoden u. Heteropoden. Ein Beitrag zur Anatomic u. Entwicklungsgeschichte dieser Thiere. Leipzig. GOULD, H. N., 1917. Studies on sex in the hermaphrodite mollusc (Crepidula). I. History of sex cycle. Jour. Exper. Zool., 23 : 1. HERTWIG, R., 1905. tiber das Problem der sexuellen Differenzierung. Verh. dcutsch. Zool. Ges.. 15: 186. HICKMAN, C. P., 1931. Spermatogenesis of Succinea ovalis Say, with special reference to the components of the sperm. Jour. Morph. and Physio!., 51 : 243. KNIPOWITSCH, N., 1891. Zur Entwicklungsgeschichte von Clione limacina. Biol. Ccntralbl, 11: 300. REPRODUCTIVE SYSTEM OF LIMACINA 25 KROHN, A., 1856. Beobachtungen aus der Entwicklungsgeschichte der Pteropoden, Heteropoden und Echinodermen. Mullcr's Arch. f. Anat., p. 515. KUSCHAKEWITSCH, S., 1910. Zur Kenntnis der sogenannten warmformigen spermien der Prosobranchier. Anat. Ans., 37 : 318. KUSCHAKEWITSCH, S., 1911. tiber die Entwicklung der spermien bie Conus mediterraneus Brug. und vermatus gigas Biv. Biol. Centralbl., 31 : 530. LOOSANOFF, V. L., 1937. Development of the primary gonad and sexual phases in Venus mercenaria L. Biol. Bull, 72: 389. MEISENHEIMER, J., 1905. Pteropoda. ll'iss. Ergebn. dc Deutsch, Tiefsee E.vp., Valdivia, 9. MEVES, F., 1903. Ueber oligopyrene und apyrene spermien und ihre Entstehung nach Beobachtungen an Paludina und Pygaera. Arch. f. micros. Anat., 61: 1. MULLER, J., 1858. Sur le developpement des Pteropoden. L'Institute, 26: 37. ORTON, J. H., 1909. On the occurrence of protandric hermaphroditism in the mol- lusc Crepidula fornicata. Proc. Roy. Soc., B, 81 : 468. ORTON, J. H., 1926. Observations and experiments on sex-change in the Euro- pean oyster (O. edulis). Jour. Mar. Biol. Ass'it, 14: 967. REINKE, E. E., 1912. A preliminary account of the development of the apyrene spermatozoa in Strombus and of the nurse-cells in Littorina. Biol. Bull., 22 : 319. REINKE, E. E., 1914. The development of apyrene spermatozoa of Strombus bitu- berculatus. Carnegie hist. Washington, Publ. No. 183, p. 195. SOULEYET, L., 1843. Observations anatomiques, physiologiques et zoologiques sur les mollusques pteropodes. Comfit. Rend. Acad. Sci., 17 : 662. STEPHAN, P., 1903. Sur les spermies oligopyrenes et apyrenes de quelques Proso- branches. Compt. Rend. Soc. Biol., Paris, 55 : 554. TESCH, J. J., 1913. Pteropoda. Das Tierreich, 36. THIELE, J., 1931. Handbuch der Systematischen Weichtierkunde. VAN BENEDEN, P. J., 1839. Memoire sur la Limacina arctica. Nonv. Mem. Acad. Roy. Sci. et Belles Lettrcs de Bruxelles, T. 14. VON BRUNN, M., 1884. Untersuchungen iiber die doppelte Form der Samenkorper von Paludina vivipara. Arch. Mikrosk. Anat., 23: 413. ZARNIK, B., 1911. Uber den Chromosomencyclus bei Pteropoden. Verh. dcutsch. Zool. Gcs., Leipzig, 20: 205. THE HISTORY OF A POPULATION OF LIMACINA RETROVERSA DURING ITS DRIFT ACROSS THE GULF OF MAINE ALFRED C. REDFIELD (From the Biological Laboratories, Harvard University, and the Woods Hole Occanographic Institution^ Woods Hole, Mass.} Geographical distribution in the sea is a hydrodynamic problem, in which the flow of. waters determines not only the " climate " of specific regions, but also the supply of organisms. to populate these regions. A pelagic population must drift with the waters in which it lives. In con- trast to littoral and terrestrial populations, there is no inherent fixity in its occurrence, save that determined by the current systems of the sea. The continuous drift of water through a given region not only raises questions regarding the maintenance of a population there (Damas, 1905 ; So'mme, 1933, 1934) ; it renders continuous observation on the natural history of the population singularly difficult, since the indi- viduals observed at any place one day are different from those observed at the same place on another occasion. Immigrant plankton which do not thrive or reproduce in the Gulf of Maine, but which are carried into its waters from offshore, were found by Bigelow (1926) to occur in a zone extending from the Eastern Channel along the eastern, northern and western sectors of the Gulf as though carried thither by the great cyclonic eddy which Huntsman (1924) and Bigelow (1927) have shown to dominate the circulation of the Gulf. Animals of arctic origin, Limacina hclicina, Mertensia ovum, Oikoplcura van Jioffcni, and Ptychogcna lactca were taken only in the eastern sector of this zone, apparently being unable to survive long enough to be borne further. Other immigrants, both of boreal and trop- ical origin, survived until carried as far as the offing of Cape Cod, but few were ever taken in the southern sector of the circuit. As his cruises were not taken with sufficient frequency, these observations yield no evidence concerning the velocity of the drift on which these invaders were borne, although its course was clearly enough indicated. The observations on the pteropod Limacina retroversa Fleming - de- 1 Contribution No. 196 from the Woods Hole Oceanographic Institution. - We have employed the name Limacina retroversa Fleming because of its current use in oceanography. Dr. W. J. Clench advises me that the generic name 26 HISTORY OF A POPULATION OF LIMACINA 27 scribed in this paper are of interest because during the greater part of the year it has been possible to follow the history of a rather definite population as it drifted across the Gulf of Maine, noting its rate of spread, its mortality, the growth and reproduction of its individuals, and its replacement by another population of the same species. Limacina rctroversa is a boreal form occurring from latitude 50° to northern Norway off the European coast and from latitude 34° to the southern part of Davis Strait in the western Atlantic. Bigelow consid- ered it one of the most characteristic of the permanent pelagic inhab- itants of the Gulf of Maine " where its numbers depend on local repro- duction and not on immigration from elsewhere." He also emphasizes the irregularity of its occurrence, an opinion with which I can agree since my observations are at some variance with his. The seasonal variation in the distribution of Limacina shown in Fig. 44 of Bigelow's monograph led me to suspect that its occurrence might be closely cor- related with the residual drift of the superficial water, since the captures in December and January extend along the northern and western shores of the Gulf, those from February to April occur over the western and southern sectors, while records for May extend from this zone north- easterly toward the Bay of Fundy. During the years 1933-1934 a systematic survey was made of the Gulf of Maine with the object of obtaining a more precise seasonal pic- ture of its hydrography and biology than was available from the ob- servations of Bigelow. Thirteen cruises were made by the research vessel " Atlantis " in the course of fifteen months, with the result that 684 hydrographic stations in the Gulf of Maine and its adjacent waters were occupied. Whenever weather permitted, a standard vertical haul was made at each station using a Heligoland larva net similar to that described by Kiinne (1933). The opening of the net was 143 cm. in diameter and it was made of No. 0 silk having 38 meshes to the inch. The net was drawn from a point near the bottom to the surface. The -catch of each net haul was preserved and from it the Limacina have been separated, counted and measured. Seasonal Distribution and Abundance The seasonal distribution and abundance of Limacina throughout the year, as shown by the number of specimens caught in each haul, is presented in Figs. 1 and 2. Limacina was scarce in the Gulf of Maine Spiratella has claims of priority arising from Blainville's description (Diet. Sci. Nat., ix, 407, 1817). No attempt has been made to distinguish L. retroversa from L. balca Moller since the distinctness of these species is in doubt (Bigelow, 1926, p. 116). 28 ALFRED C. REDFIELD APRIL LARGE SPECIMENS FIG. 1. Charts showing the distribution of Limacina retroversa in the Gulf of Maine in September and December, 1933, and in January, March, April and May, 1934. The numbers indicate the positions at which vertical hauls were made and the numbers caught per haul. The shaded area includes all hauls in which the catch was greater than the mean catch for the entire cruise. The numbers entered for April and May include only individuals attributable to population A as defined in text. HISTORY OF A POPULATION OF LIMACINA 29 in the fall of 1933. A dense population of these organisms appeared in the eastern side of the Gulf in December, 1933 and spread westward during the winter. By April these pteropods are concentrated in the western side of the Gulf ; the region to the eastward in which they first appeared being occupied by waters scantily populated. As the season advanced the size of the catch at each haul dwindled and the distribu- tion became more general. APRIL SMALL SPECIMENS FIG. 2. Charts showing the distribution of Liinacina rctroversa of all sizes in the Gulf of Maine in June and September, 1934, and of small individuals at- tributable to population B in April and May, 1934. These observations suggest that a huge swarm of Limacina drifted into the Gulf from the east in December and were carried westward by the cyclonic drift of water about the Gulf. The acceptance of this inter- pretation requires evidence on two points : ( 1 ) that the population moved not only in the direction but at a rate corresponding to the mass move- ment of the water, and (2) that a reasonable continuity exists between 30 ALFRED C. REDFIELD the population observed in December in the eastern region and that found in April and May in the western part of the Gulf. The Circulation of the Gulf The character of the circulation of the Gulf is adequately described by Bigelow (1927), who with Huntsman showed that the dominant cir- culation is a great anti-clockwise eddy around the basin of the Gulf, FIG. 3. Positions occupied by an area in which the salinity of the water at 50 meters depth exceeded 33°/oo during the period from September, 1933, to May, 1934. Arrows indicate the apparent direction of drift at each period. varying in velocity and in detail from season to season and complicated by subsidiary eddies. The great eddy receives accessions of water from the east, particularly in winter and early spring ; the area of inflow be- ing over the Northern Channel and the adjoining banks. Compensatory losses are occasioned by outflow across the Southern Channel and around the eastern end of Georges Bank. The circulation at the greater depths has a somewhat modified character owing to the confining features of the bottom contour. A report on the dynamics of the circulation of the Gulf, based on observations made during the cruises of 1933-34, is being prepared by Dr. E. E. Watson. HISTORY OF A POPULATION OF LIMACINA 31 While our vertical hauls have given no indication of the depths at which Limacina occurs, it seems probable that it will be carried with the more superficial circulation of the Gulf. Bigelow (1926, p. 121) states that " the most prolific depth zone may be stated as from 20 to 25 me- ters down to about 80, which corroborates Paulsen's (1910) generaliza- FIG. 4. Charts showing the relative positions of the area of high salinity at 50 meters depth and the distribution of Limacina between December, 1933 and April, 1934. The closed contours enclose the area in which S > 33°/oo- The hatched areas include catches greater than 100 specimens per haul. The size of each catch is roughly proportional to the area of the black circles. Open circles represent hauls in which no Limacina were caught. tion that Limacina lives chiefly shoaler than 50 meters in north Euro- pean waters, though it has occasionally been taken much deeper." That the sub-surface waters actually drift with velocity as well as direction comparable to that characterizing the Limacina population is indicated by a study of the distribution of salinity in the upper wrater levels throughout the period of the survey. An extensive area characterized 32 ALFRED C. REDFIELD by salinities higher than the surroundings occurred in the eastern part of the Gulf in September, 1933. Bigelow (1927, pp. 767, 768) noted a similar occurrence in 1913, 1914 and 1915. The position of this area gradually shifted during the following months, until March when it occurred in the southwestern part of the Gulf. Subsequently it could be traced with less certain continuity as it divided, apparently drifting in part out to sea, and in part in a northeasterly direction. The " saline pool " is most clearly limited in the data for salinities at 50 meters, at which depths it is enclosed by the 33°/oo isohaline. The successive FIG. 5. Photographs of representative catches of Limacina rctrorcrsa taken between December, 1933 and May, 1934 showing growth. The large specimens taken in May are from the western basin and represent the size attained by popu- lation A. The small May specimens were taken in the eastern part of the Gulf and are attributed to population B. Scale, 1 centimeter divided into tenths. positions of this isohaline are illustrated in Fig. 3, and correspond closely to the superficial circulation as deduced and illustrated by Bigelow (1927). The velocity of drift of this area, characterized by relatively high salinity, corresponds closely to that of the Limacina population which apparently followed close in its rear (Fig. 4). The observa- tions on the distribution of Limacina and the occurrence of water of S > 33°/oo at 50 meters depth agree in indicating a drift during the winter period of the sub-surface waters through a distance of 150 miles in four months or at a rate of about 1.25 miles per day. HISTORY OF A POPULATION OF LIMACINA 33 The Distribution of Size at Different Seasons The identity of the population sampled at different periods is indi- cated by an examination of the distribution of size among the specimens taken. These appeared to grow larger as the winter advanced (see Fig. 5). In April and May, however, large numbers of small specimens appeared in certain catches. Since the animals taken at different times and places can only be attributed to the same population if they display a reasonable relation in size, the distribution of size among the indi- viduals of each catch was determined. The largest diameter of the specimens was measured as they lay in random positions in a watchglass. Care was taken to avoid measurements along diameters exaggerated by the spreading of the foot as it protruded from the shell. The measure- ments were separated into 9 size groups of 0.3 mm. span. The number in each group was expressed as a percentage of the total catch falling within each group. Figure 6 shows the distribution of size among all the specimens taken during each cruise. In December size distribution is homogeneous ; the modal class size is 0.6-0.9 mm. with a uniform distribution of smaller and larger sizes ranging between 0.3 and 1.2 mm. In January the modal class size is the same, but the larger classes have grown at the expense of the smaller. By March a general increase in size is apparent, the modal class being 1.2-1.5 mm. and almost all specimens falling between 0.6 and 1.8 mm. The general shape of the polygon is unchanged as should be the case if the increase in size is due to uniform growth. In April the frequency polygon becomes skewed. Though the modal size is the same as in March, the larger classes have increased in number, relative to the mode, as demanded by growth. The skew is due to the appear- ance of significant numbers of very small (less than 0.6 mm.) indi- viduals which were not present in the March population. In May these small specimens increase greatly in numbers with the result that the fre- quency polygon is bimodal. The principal modal size is now 0.3—0.6 mm., being determined by the small individuals. A secondary mode occurs between 1.5 and 1.8 mm., apparently representing the mode of the original population which determined the mode during the previous periods. The observations made between December and May may be readily interpreted as follows. A homogeneous population of small individuals entered the Gulf from the east in December and grew continuously in size. In April a new population of small individuals appeared and by May these had increased in numbers so as to dominate the size distribu- tion. These may represent a generation of offspring from the original 34 ALFRED C. REDFIELD DECEMBER MAY JANUARY JUNE MARCH SEPTEMBER APRIL TOT AL FIG. 6. Frequency polygons showing the distribution of size in the total catch of Lima-cina obtained during each cruise. The lower right-hand polygon repre- sents the combined distribution for the entire series of cruises, corrected for variation in fishing effort. Ordinates, percentage of entire catch falling in each size group. Abscissae, size groups divided in intervals of 0.3 mm. diameter. Black areas are attributed to population A. HISTORY OF A POPULATION OF LIMACINA 35 population, or an invasion of a new group from outside the Gulf, or both. For convenience we will call the original population which ap- peared in December the A population; the new population which appeared in April the B population. The observations made subsequent to May are more difficult to interpret. The frequency polygon for June is again bimodal. The principal size class is now 1.2-1.5 mm. This is smaller than the mode of the A population for the preceding month and may be tentatively assigned to the B population, which may be expected to grow rapidly with the warming of the water. The largest size class, 2.1-2.4 mm., is relatively more abundant than at any previous time and probably con- tains the last survivors of the A population. A secondary mode occurs between 0.3 and 0.9 mm., which may represent a further production or a new invasion of young individuals, a hypothetical C population. The small specimens taken in June were caught a few miles off the coast of Cape Cod. Because of the remoteness of this position from the inflow into the Gulf, they are in all likelihood the offspring of the A population, which was richly represented in this region during the preceding cruise, rather than to a new invasion. Only five hauls were made during the June cruise and these all in the southwestern half of the Gulf. The data for this period are not as representative as those obtained at other times. The scanty and scattered catches made in September, 1934 yield a simple frequency polygon with maximal size class at 0.6-0.9 mm. The distribution of size is generalized and resembles that of the combined measurements of the total collections throughout the year. Apparently during the summer growth is rapid and reproduction and recruitment of the population from without the Gulf is more or less continuous so that separate broods or populations can no longer be distinguished. Dr. Sidney Hsiao has studied the reproductive history of the collections and finds that Litnacina produce eggs in small numbers continuously after reaching a size of 1.0 mm. It may be inferred that the offspring of one generation will vary greatly in age and after several generations their descendants may be represented by a complete assortment of sizes in which separate broods will be indistinguishable. The Distribution of Size in the Population of Different Sectors of the Gulf Let us return to the examination of the history of populations A and B with a view to determining whether they are actually homogeneous population's and to distinguishing whether population B arises through reproduction within the Gulf, or invasion from outside. 36 ALFRED C. REDFIELD The size distribution of representative catches made in different parts of the Gulf at each period has been examined. For this purpose the Gulf has been divided into seven sectors as indicated in Fig. 7. These sectors are arranged in the order in which the non-tidal drift of water may be expected to carry a pelagic population in the circuit of the Gulf. It should be noted, however, that much water will be carried from the FIG. 7. Chart of Gulf of Maine showing principal place names and the sectors into which the area is divided for analysis of population distribution. Contour encloses depths less than 100 meters. Yarmouth and Mt. Desert sectors directly to the Georges sector without penetrating the western sectors of the Gulf, as Fig. 3 indicates. Repre- sentative catches made at each period in each of these sectors have been selected and their size frequency polygons are given in Fig. 8. Figure 8 shows that from December through March the size distri- bution in different parts of the Gulf is homogeneous and that the in- HISTORY OF A POPULATION OF LIMACINA 37 crease in size noticeable in March is of general occurrence. The ex- tension of the population westward during the winter is apparent. The BROWNS YARMOUTH MT. DESERT SEGUIN MASS. CULTIVATOR GEORGES DECEMBER ^ r*"^n FEW NONE NONE NONE ""^^^^r~~r~rTt JANUARY • J FEW FEW NONE ^^^^^i I i ' i i ^ i n ^^^^^^ i i i i i i i ' i i i i i i i i i i i T i i '"i~i i ' i i i i ^"^^^^i"~T"~n~l MARCH J FEW L • d NO OBS. a ' PFT ""^^^^^r^ r~'l T"l "I — I "I "i | ' I "^^^^T^~l ^~r*^^r^^^^™r~l ^"~f™^^^^^^^^™1 I I I I I I 1 I ^ T~l APRIL "~^T~"^^^^T~n ( ( — ^^^^T~T~t I1 I r^^^^ T~| 1^ ^^^^^^^^ ~T~T~^ I MA Y — i — TT^ 1 ~H \\ ^ ' ' PI^^'F'^^^^W ^n ° I • I ~r^^^^~i ~~i r~T~"^^^^r~I JUNE NO OBS. NO OBS. NO OBS. 100 SEPTEMBER SO 1-2 24 NO OBS. FIG. 8. Frequency polygons showing the distribution of size in representative catches of Limacina from each sector of the Gulf of Maine during each cruise. No hauls were made in the sectors marked No Obs. Ordinates, percentage of entire catch falling in each size group. Abscissa, size groups divided in intervals of 0.3 mm. diameter. Black areas are attributed to population A. 38 ALFRED C. REDFIELD continuous occurrence of Limacina in the Georges sector is in accord- ance with the characteristics of the current system as noted above. In April the homogeneous A population appears in all the northern and western sectors of the Gulf. Small individuals representing the B popu- lation appear now in the eastern sectors, Browns, Yarmouth and Georges. In the outermost sector to the east, Browns, the B population is unaccompanied by any specimens of larger size, characteristic of the A population at this time. In May the small individuals of the B population are much more widespread, appearing abundantly in all sectors except the most westerly, Massachusetts sector. Only small individuals occur in the stations of the Browns and Yarmouth sector. The populations of the Mt. Desert, Cultivator and Georges sectors have a mixed, bimodal character. The duplex character of the population in April or May is brought out in Figs. 1 and 2, in which the numerical distribution of individuals of dif- ferent sizes is shown. Small specimens appear in numbers only in the water in the offing of the Eastern Channel in April and occur in small numbers along the course of the inflow through the Eastern Channel. In May they are generally distributed in the eastern part of the Gulf which is most accessible to inflowing water and are scarce in the western region in which the larger specimens of the A population are dominant. These observations justify the belief that the A population represents the invasion of the Gulf from the eastward by a homogeneous popula- tion which grows as it drifts westward. Reproduction in the Gulf The appearance of the B population in the eastern sectors, unaccom- panied by members of the A population and the extension of the smaller group westward during May makes it clear that the new population of small individuals which appears in the spring is due chiefly to a second invasion of Limacina into the Gulf from waters offshore and to the east- ward. Whether the small individuals which occur mixed with the A population in several sectors are the offspring of the A population is difficult to decide. Dr. Hsiao finds the A population to contain sexually active individuals during the spring, so that offspring from this popula- tion are to be expected. On the other hand, the relative scarcity in May of the B population in the Massachusetts sector where the A popu- lation was then most concentrated, argues against the possibility that the B population in neighboring sectors is in any important degree the offspring of the A population. It is noteworthy that when the A popu- lation first appeared in December it was dominated by individuals of HISTORY OF A POPULATION OF LIMACINA 39 small size and no individuals of large size occurred among them. This was also true in many hauls in which the B population appeared in April and May. This indicates that the parental generation does not survive after breeding long enough to be caught along with offspring which have grown to diameters of 0.3 mm. or greater. This consideration also sup- ports the view that the hauls in which large and small individuals occur together represent the mingling of populations of different origin rather than the simultaneous presence of a parental and filial generation. Growth Since the foregoing analysis of the collections appears to warrant the conclusion that the A and B populations are homogeneous entities, DEC. JAN. FEB. MAR. APR. MAY JUNE JULY FIG. 9. Diagram indicating the increase in diameter of the modal class of Limacina of populations A, B and C during the winter of 1933-34. Ordinates, size groups divided in intervals of 0.3 mm. diameter. Abscissae, time during which collections were made. The black areas represent the modal class size of each population. we may examine the growth rate of Limacina by comparing the size of the modal classes of these populations at different times throughout the year. In Fig. 9 the size of the modal class is plotted against the time the collection was made. Roughly speaking, the diameter of the A population doubles in about five months during the winter. At this time the temperature of the water column as a whole is below 7° C. The B population observed in May and June grows very much more rapidly as might be expected from the higher temperature of the surface waters at that time. 40 ALFRED C. REDFIELD Numerical Abundance and Mortality The A population is sufficiently established as a distinct and homo- geneous entity and was under observation for a sufficiently long period to warrant some statistical examination of its numerical abundance. It must be realized that the number of hauls made in each cruise is insuffi- cient and their distribution too irregular to allow great confidence in the outcome of the analysis of a matter of such complexity, but data of this sort are so rare and so difficult to obtain that one is warranted in the attempt to get the utmost out of it. TABLE I Statistics of the size of catches of Limacina retroversa made during the cruises of 1933-1934. Cruises No. of Hauls Total Catch Mean Catch per Haul Per Cent of Hauls with Limacina Present Per Cent of Hauls with Catch Greater than Mean Sept. 2-14, '33 34 104 3 44 21 Dec. 2-11, '33 22 3120 142 23 18 Jan. 8-13, '34 12 2685 224 66 33 Mar. 21, 29, '34 18 2535 141 100 55 April 17-23, '34*— Small. . . 24 141 6 71 33 Large 24 2575 107 84 24 Total 24 2716 113 100 23 May 21-June 3, '34— Small 44 4068 92 98 25 Large. . . . 44 1849 42 91 25 Total 44 5917 134 100 23 June 25-July 1, '34. . . 7 220 31 100 57 Sept. 17-27, '34. ... 26 507 20 92 20 " Two hauls made in the Eastern Channel on May 6, 1934 are included in the data for the April cruise. We may consider that all the individuals taken in December, Janu- ary and March belong to the A population. In order to eliminate as far as possible the B population individuals smaller than 0.9 mm. are elim- inated from the data for April and smaller than 1.2 mm. from that for May. The A population cannot be identified with assurance in the June collections. Increase in numbers can be due only to the invasion of the Gulf by new immigrants from offshore since the elimination of the B population precludes reproduction as a means of increase. Decreases in numbers may be due to water movements, carrying a part of the population out of the area under observation, or to mortality. Since the population first appears densely distributed at one side of the Gulf, some redistribu- HISTORY OF A POPULATION OF LIMACINA 41 tion of density between different points of observation is to be expected as the result of lateral mixing of the water. This may be expected to result in an increasing uniformity of distribution. Table I contains some statistics concerning the size of the catches made during the several cruises. An indication of the numerical abun- dance of Limacina in the Gulf as a whole at each period is given by the numbers in column 4 showing the average number taken per haul. Between December and January the numbers increase, after January there is a progressive decline in abundance. An examination of the distribution of Limacina from month to month (Fig. 1) warrants the conclusion that the invasion of the A population into the Gulf continues through December and January since large catches are made at these times in the region of inflow. The percentage of stations at which Limacina was taken (column 5) increases three-fold during this interval. The invasion appears to terminate in January, since following that period no considerable numbers of individuals referable to the A population were taken in the region of inflow into the Gulf. Invasion may conse- quently be excluded as a significant factor affecting the numbers of the population after January. The numbers of the population decline steadily from January through May, the mean catch of large individuals referable to popula- tion A decreasing to less than 20 per cent of the original in four months. Significant losses are attributable to the drift of water out of the Gulf. The continuous occurrence of catches containing Limacina in the Georges sector, which lies in the path of outflow through the Eastern Channel, and the concentration of the A population in the western and southern quarters of the Gulf in April and May, where it is subject to drift outward by way of the South Channel, favor such losses. It is probable that large losses in numbers also occur as the result of mortality. Measurements of the total volume of zooplankton caught throughout the Gulf indicate a general mortality of about 60 per cent during the winter period. The data in Table I indicate that the percentage of stations at which Limacina occurred is not less in April and May than in January. If the circulation into and out of the Gulf were rapid enough to account for the greater part of the loss in numbers, a greater decrease in the area of distribution would be expected. It may be argued that since the area of distribution does not decrease while the numbers caught per station decline five-fold, the decline is due to the death of a large part of the population. This argument, however, must be accepted with great re- serve, since the lateral mixing of water may cause small numbers of 42 ALFRED C. REDFIELD animals to be carried into regions in which they did not previously occur, and will tend to enlarge the area of occurrence. Another approach to the question of mortality may be made through an examination of the distribution of size in the catches as a whole throughout the year. In Fig. 6 the size distribution of the total catch is represented. This frequency polygon was prepared by averaging the percentages of the population represented by each size group during each period of collection, a procedure which corrects for the variation in " fishing effort " during different cruises, but does not correct for the irregularity of spacing of the cruises in time. The figure shows that the size most frequently caught is 0.6-0.9 mm. Smaller sizes are less frequent in part because both the A and B populations are carried into the Gulf half-grown, in part because they are more readily overlooked in sorting the catch and in the case of the smallest class, because many may pass through the mesh of the net. The larger size classes diminish progressively in their relative abundance, as would be expected if the population is subject to continuous mortality. The mode of the A popu- lation is 0.6-0.9 mm. in January, 1.5-1.8 mm. in May. Specimens of the latter size occur about one-third as abundantly as the former in the catches as a whole. This suggests a mortality of about 66 per cent in the time taken to grow from the former to the latter size, during four winter months. Such statistics, however, are not very convincing since the smaller classes are recruited by invasion and the chance of loss by drifting out of the Gulf increases with time, as does the size of the individuals of the population. Lateral Mixing The horizontal mixing of water masses, which has recently attracted the attention of hydrographers, may be expected to have consequences of importance in determining the distribution of organisms. If atten- tion is limited to the consequences of steady residual currents, one can only conclude that the pelagic inhabitants of a region such as the Gulf of Maine which is the seat of a cyclonic eddy must be constantly washed away by the flow of water through it. The rise and decline of the Litnacina population which we have observed supports this conclusion. The existence of lateral mixing, if of considerable extent, should tend to diminish the tendency of residual currents to transport all of the pelagic population out of such a region. By the transport of a part of the population backstream, or around and across the vortex of an eddy, an opportunity would be afforded for the endemic occurrence of the population in spite of the continuous drainage occasioned by the residual currents. HISTORY OF A POPULATION OF LIMACINA 43 The effect of lateral mixing must tend continuously to make the distribution of the population more homogeneous. After the introduc- tion of a limited water mass containing a rich population into a large unpopulated region, if lateral mixing alone were in operation one would expect the limits of the populated waters to become less distinct and the distribution to become more general. If samples were taken at random at any time during the mixing process, at first a limited number of sam- ples would contain numbers far in excess of the mean value for the col- lections as a whole, and the remainder would contain small numbers or none at all. As mixing proceeds, the poorly populated waters would become enriched at the expense of the densely populated regions. In both regions the density of population would change in the direction of the mean population for the entire region and eventually the entire distribution would assume a random character in which the modal popu- lation density would correspond with the mean. At this point half the samples would contain more than the mean number, and an equal num- ber less than this value. Subsequent mixing would tend only to diminish the variation of the numbers secured in the different samples. The catches of Limacina have been examined for such evidences of lateral mixing. Examination of the distribution of the population shown in Fig. 1 indicates that the populated region was sharply de- limited in December — no Limacina occurring in regions immediately in advance of the populated waters. In January moderate catches were made in regions well in advance of the main population, suggesting a gradual disintegration of the boundary zone by lateral mixing. In March the area of intense population covers more than half the Gulf and at the same time the numbers present at each station have diminished and become more uniform as required if lateral mixing is in effect. The same tendencies continue in April and May and are particularly evident when the distribution of small specimens (population B) is compared. While previous to March less than half the hauls have contained more than the mean catch for the Gulf as a whole, by March at least half the hauls are of this size — as required by random distribution. One feature of the distribution of large specimens in April and May is of particular interest. In the east central part of the Gulf, on the line between Cashes Ledge and Yarmouth, Nova Scotia, the catches of large specimens are on the whole greater in May than in April in spite of the fact that some mortality is doubtless occurring during this period and the movement is contrary to the expected drift of the water. As a further test to see if the distribution of population is becoming more homogeneous, the catches have been grouped into classes according 44 ALFRED C. REDFIELD to the number taken per haul and the number of catches in each class has been expressed as a percentage of the total number of catches made during the cruise in question. The result is presented in Fig. 10. The left-hand diagram shows the result when the size of the catch is noted DECEMBER DECEMBER 50 100 200 400 800 1600 FIG. 10. Frequency polygons showing the distribution of catches of different numerical abundance among the total number of hauls made during each cruise. Ordinates, percentage of entire number of hauls yielding catches of each size. Abscissae, left-hand diagrams, size groups divided in intervals according to number caught in each haul ; right-hand diagram, size groups expressed as multiples or fractions of the mean catch per haul during each period. numerically (and assigned to classes which for convenience increase in geometrical progression). The right-hand diagram represents the same data arranged in size classes which increase as multiples, or decrease as fractions, of the mean population of the region as a whole at the period of collection. The latter procedure eliminates the effect of losses HISTORY OF A POPULATION OF LIMACINA 45 through mortality. Both diagrams indicate the complete heterogeneity of the samples in December and January and the progressive approach of both the smaller and larger catches toward the mean value which reaches its culmination in April and May. Actually the situation observed is complicated by the drift of the population across the region and by the invasion of the eastern part of the Gulf by water poor in Limacina of population A in March and April. This inflow tends to decrease the homogenity of distribution and ac- counts for the increase in the proportion of catches in April and May which contained few or no Limacina. It also explains why the per- centage of hauls in which the catch was greater than the mean diminished in the months subsequent to March. These tendencies during April and May at least make it clear that lateral mixing is not sufficiently great to obliterate the effect of the residual current in depleting the Limacina population of the eastern portion of the Gulf. Because of the impossibility of separating the effects of mortality and of losses and gains through the residual movement of the water into and out of the Gulf from the apparent effects of lateral mixing, any attempt to evaluate the extent of the latter has been abandoned. Be- cause of the great influence which such mixing would have in enabling a population to maintain itself in a region from which the residual cur- rent would constantly tend to remove it, it seems desirable, however, to emphasize the possibility of such mixing and to indicate the way in which it appears to manifest itself. Discussion The evidence seems definite that the population of Limacina. which occurred in the Gulf of Maine in 1933-34 was not endemic, as Bigelow thought, but owed its origin to two separate invasions of young indi- viduals from offshore. Unlike the less successful immigrants of arctic or tropical origin, Limacina rctroversa thrives well enough to grow to sexual maturity and perhaps to reproduce within the Gulf and to be carried in considerable numbers to all its parts. For this reason Bigelow obtained evidence which led him to conclude the species was endemic. By following the history of these populations closely, however, it is shown that neither invasion was able to maintain its numbers in the Gulf — either because reproduction was not successful or was inadequate to balance the depletion due to mortality and the drift out of the Gulf. The two invasions consisted of dense swarms of small individuals of very uniform size, sharply separated by a period when practically no Limacina occurred in the region of inflow into the Gulf. The discon- 46 ALFRED C. REDFIELD tinuous character of these invasions no doubt accounts for the irregu- larity in occurrence of Liinacina in the Gulf noted by Bigelow. Why the invasion occurs in such definite periodic swarms can be answered only by studies carried out in the waters to the eastward from which these invasions undoubtedly come. It is an interesting question why the invading swarms are composed of individuals of such uniform size. In describing the population found in the Gulf in September, 1934, it was pointed out that animals of all sizes were mixed together and that successive broods can no longer be distinguished. Dr. Hsiao's studies indicate that the mature Liinacina continue to produce eggs during a considerable period of growth. Yet we found evidence to suggest that the parental generation does not sur- vive after breeding long enough to be caught along with its offspring at the time when these enter the Gulf. Dr. Hsiao has measured numerous catches of Liinacina taken over the continental shelf between Cape Cod and Cape Hatteras. Usually these catches are composed of individuals of uniform size, just as were those which invaded the Gulf. This riddle also can be answered only by closely timed observations in the offshore areas of production. The most conspicuous result of this study is the demonstration of the degree to which the occurrence of Liinacina in the Gulf of Maine de- pends upon the circulation of its waters. Damas in 1905 raised the question : How does the plankton of a given region maintain its character in the face of the continual circulation of the currents and how does a given species persist so as to possess a special geographic distribution? He concluded that there must exist a special zone or center of production in which adults abound and reproduce successfully and that to this re- gion circulatory currents serve to bring back periodically a proportion of the individuals which become entrained and dispersed by the con- tinual movements of the water. The circulatory system of the Nor- wegian Sea subsequently described by Helland-Hansen and Nansen (1909) appeared to supply just such a mechanism as was required to account for Damas' observation of the production centers of copepods in that region. S0mme (1933, 1934) has shown how such centers of pro- duction are established for Calanns finmarchicus and hyperboreits in the deep waters of the Norwegian fjords in winter, and has elucidated the mechanism which results in the release of the organisms from these re- gions at the time of reproduction. The observations on Liinacina have not revealed the presence of a center of production in the Gulf of Maine. They point to the existence of such regions offshore to the east- ward and are of interest rather in telling something of the fate of these animals, entrained in the movement of water, which are carried away HISTORY OF A POPULATION OF LIMACINA 47 never to return, yet for a while to occupy an important role in the ecol- ogy of other regions. Behind the geographical distribution of each spe- cies of plankton there must be a complex balance of biological and physical factors. Of the latter, flow of water appears to be paramount ; its consequences too frequently neglected. Summary A population of small specimens of the pteropod, Limacina rctro- versa, appeared in the eastern part of the Gulf of Maine in December, 1933. From collections made during the following 9 months informa- tion was obtained showing that the population was a homogeneous one, that its members grew to maximum size in 5 months, declining in num- bers as they did so. A second population of small individuals appeared in the Gulf in late spring, originating chiefly from offshore, but possibly in part being off- spring of the original population. These were unsuccessful in main- taining their numbers throughout the summer. In addition to the information on the life history of Limacina, the data indicate the rate of drift of the water in its circuit of the Gulf. It supplies also suggestive information on the dispersal of organisms through the lateral mixing of water. It emphasizes the dependence of pelagic organisms upon the current systems of the ocean and the diffi- culty involved in maintaining a permanent population in any one locality. REFERENCES BIGELOW, H. B., 1926. Plankton of the offshore waters of the Gulf of Maine. Bull. U. S. Bur. Fish., 40 (1924), Part II: 1-509. BIGELOW, H. B., 1927. Physical oceanography of the Gulf of Maine. Bull. U. S. Bur. Fish., 40 (1924), Part II: 511-1027. DAMAS, D., 1905. Notes biologiques sur les copepods de la Mer Norwegienne. Publ. de Circonstance, No. 22, Copenhagen. HELLAND-HANSEN, B., AND F. NANSEN, 1909. The Norwegian Sea. Report of the Norwegian Fishery and Marine Investigations. II. Part I. HUNTSMAN, A. G., 1924. Oceanography. Handbook of the British Association for the Advancement of Science, Toronto, pp. 274-290. KUNNE, CL., 1933. Weitere Untersuchungen zum vergleich der Fangfahigheit vershiedener modelle von vertikal Fischen den Plankton-netzen. Rapp. et Proccs-verb. dcs Reunions. Conseil perm, internat. pour I'explor. dc la mer, 83 : 19 pp. PAULSEN, O., 1910. Pterapoda. Bull. Trimestricl, Conseil perm, internat. pour I'explor. de la mer, Part I, pp. 52-59. S0MME, I. D., 1933. A possible relation between the production of animal plank- ton and the current-system of the sea. Am. Nat., 67 : 30. S^MME, J. D., 1934. Animal plankton of the Norwegian coast waters and the open sea. I. Production of Calanus finmarchicus (Gunner) and Calanus hyperboreus (Kroyer) in the Lofoten area. Report on Norwegian Fish- ery and Marine Investigations, IV, No. 9, pp. 3-163. RESPONSES OF THE SWIMBLADDER OF THE GUPPY, LEBISTES RETICULATUS, TO SUDDEN PRESSURE DECREASES FRANK A. BROWN, JR. (From the Department of Zoology, Nortlnvestcrn University) Most fishes maintain a density equal to that of the surrounding water. This is done through a very accurate control of the volume of their swimbladders by appropriate gaseous exchange between the swimblad- der and the blood. As the fish passes into deeper water the increased hydrostatic pressure compresses the bladder gases to the point that the fishes are no longer buoyed up in the water. Consequently, in order for the fishes to adapt themselves to the new pressure, they must put more gases into the bladder. Conversely, as the fishes rise in water the de- creasing hydrostatic pressure renders the gases in the bladder too buoyant and provision is made for the release of the proper amount. Many experiments have been performed in an attempt to determine the mechanism of control of the swimbladder. The great majority of these experiments have involved analysis of gases found in the swimbladder in different states of adaptation and under controlled experimental con- ditions. Although much valuable information has been obtained, our picture of the mechanism is by no means complete. Von Ledebur (1937) summarizes our present knowledge in this field and reviews briefly all the current theories to explain the secretion of gases into the swimbladder against the diffusion gradient and their re- moval. The variety of explanations indicates the need of more research before a definitive theory can be devised. The experiments to be described in this report were carried out upon the guppy, Lcbistes rcticulatus, a physoclistous fish. They demonstrate the responses of the swimbladder to decreased hydrostatic pressure and to situations in which the normal diffusion gradient, favoring passage of gases out of the bladder, is experimentally reversed and varied in steep- ness. Some interesting information has been obtained upon the mech- anism of gaseous exchange between the environment and swimbladder. The apparatus consisted of a two and a half gallon carboy. Through the stopper was projected a small bore glass tube which passed into a flask serving as an air cushion. To this flask was attached a mercury manometer and tubes which led to a vacuum pump on the one hand and 48 SWIMBLADDER RESPONSES TO PRESSURE DECREASES 49 a compressor on the other. Both of the latter tubes were equipped with stopcocks. The whole system was arranged so that the pressure within the system could be rapidly changed but always kept under complete control. The experimental animal was placed in the carboy completely filled with water which rose three or four inches into the small bore glass tube of the stopper. In a few experiments where the gases present in the water were to be equilibrated to each new pressure by shaking, a liter bottle half full of water was substituted for the carboy. With this equipment it was possible either to alter the pressure upon the fish without appreciably altering the gaseous concentration of the medium,1 or to subject the animal to atmospheric pressure in water which had been equilibrated with air under heightened or lowered pressure. With this same apparatus it was a simple matter to measure any changes in the total amount of gases in the bladder by the application of Boyle's law. The method was merely a refinement of the technique used by Evans and Damant (1928). From time to time the exact pressure to which the guppy was adapted was measured. As the amount of gases in the bladder increased, proportionately more pressure had to be applied in order to return the fish to the same density as the water, and con- versely, as the fish permitted gases to escape from the bladder, propor- tionate decrease in pressure was necessary to adjust artificially the fish density so that the fish would tend neither to rise nor sink. The validity of this method depends upon the assumption that the pressure of the bladder gases results solely from the external environment about the fish. Evans and Damant found this to be the case with the physoclistous fish they examined, the bladder gas not being held under pressure by any structure of the fish body. Some initial experiments were performed with completely normal guppies, but it was found difficult to make accurate readings of the exact pressure to which the fish was adapted at any given moment. This diffi- culty was a result of the fish's being able to maintain itself quite sta- tionary in the water through an appreciable range of pressures by very slight activity of the pectoral fins. Consequently, all experiments re- ported here were conducted with fishes whose pectoral fins were ampu- tated. This last procedure permitted equilibrium readings of consider- able accuracy in a few seconds of time. There has been no reason to suspect that this fin removal affects the swimbladder response in any manner. Such a method as the one used here may be applied to any physo- clistous fishes since in these fishes the change in gas quantity takes place 1 When not otherwise specified in the following experiments, the gaseous con- tent of the water was that resulting from equilibrium with air at room temperature and the atmospheric pressure (about 750 mm. Hg). 50 FRANK A. BROWN, JR. 0 <^ LO <— TTf Bin' rt f^j i- i — i be SWIMBLADDER RESPONSES TO PRESSURE DECREASES 51 only very slowly by secretion or diffusion of gases and hence the few seconds of sudden change of pressure for the purpose of determining the direction and extent of gas change cannot result in an abrupt change in gas content. Physostomous fishes such as the goldfish, on the other hand, almost instantly discharge gas bubbles by way of the mouth when the pressure is suddenly decreased. EXPERIMENTAL The first experimental series was designed to test the effect of changes in pressure upon the fish swimbladder response. Water was equilibrated with air at room temperature and at atmospheric pressure and placed in the carboy. A female guppy was permitted to adjust its density to that of the water in the new situation : a pressure of one atmosphere plus eighteen inches of water. This situation served as a starting point for each experiment. In as many separate experiments the pressure on the fishes was decreased by 75, 100, 125, 150, 225, 300, and 450 mm. Hg and increased by 300 mm. Hg. The responses of the fish to these changes are indicated in Fig. 1. The fishes responded to increased pressure by decreasing their den- sity gradually over the course of six or eight hours to the point of equal- ling that of the water. This density was maintained as long as the pressure was held constant. When the pressure was decreased by 75 or 100 mm. Hg, the fish increased its density to a value equal to the sur- rounding water. However, as the pressure was decreased more, by 125, 150, 225, 300, and 450 mm. Hg, the fish no longer adaptively adjusted its density. In- stead, its density decreased at a rate that was obviously a function of the amount of decrease in pressure. These results were confirmed by a second experiment in which only male guppies were used. With pressure decreases of more than 100 mm. Hg the density of the fish continued to decrease for several days at a nearly constant rate, or until the swimbladder became so large that the wall of the visceral cavity wras distended in a balloon-like fashion. When the pressure was de- creased by 300 or more mm. Hg such a state was reached in less than twenty-four hours. There thus appeared to be some level of pressure below which the fish was unable to adapt itself, and on the contrary, became more and more maladjusted. This pressure level appeared to be quite constant whether it was reached in one sudden pressure change, in two equal changes over several hours, or in three changes over a period of four days. Figures 2A and 2B illustrate results of this nature that were & 52 FRANK A. BROWN, JR. obtained. Furthermore, the rate of decrease in fish density in response to a decreased pressure appeared to be independent of the rate at which the value was approached. INTERPRETATION OF THE EXPERIMENTAL RESULTS The cause of the abrupt reversal of fish response from density in- crease to decrease as the pressure was decreased by values greater than '550 10 20 30 40 50 60 TIME (HOURS) 80 90 LJ CC CO LU CT 850 750 650 550 24 48 72 96 120 TIME (HOURS) 144 168 FIG. 2. (A) Graph showing the swimbladder responses to reduction in pres- sure when the reduction is carried out in a single step, in two steps over a period of about 20 hours, and (B) three steps over a period of about four days. about 120 mm. Hg was the first question to be answered. A guppy adapted to a pressure increase of 300 mm. Hg, when suddenly returned to the original pressure, responded by increasing its density to a value SWIMBLADDER RESPONSES TO PRESSURE DECREASES 53 again equal to that of water. On the contrary, a guppy directly sub- jected to a similar percentage decrease in pressure responded in an opposite fashion by decreasing its density. Therefore, a physically induced size increase of the swimbladder and any activity or response of the fish directly correlated with such increase would not totally ex- plain the situation. This eliminated any reflex associated with the eyes, fins, etc., as was indicated by Meesters and Nagel (1934) to be true over the normal pressure range. A brief additional experiment was performed which immediately eliminated any induced change in amount of gases dissolved in the water as stimulating the bloating phenomenon. A guppy in an unstoppered flask of water was placed in a vacuum desiccator. The pressure on the fish was then decreased by 300 mm. Hg. The fish commenced to bloat. The water now had a dissolved air content found in conjunction with a decreased pressure which stimulated the bloating of the fish. The flask was quickly removed from the desiccator and paraffin oil poured upon the water surface to prevent any re-diffusion of gases into the water at the restored pressure. If decreased concentration of any gases in the water was responsible for the bloating phenomenon, such bloating would be expected to continue. The fish in this situation increased its density to a value equal again to that of water. It was highly improbable that the fish was responding to pressure through the activity of any sense organ. The evidence indicated that if a postulated hydrostatic pressure sense organ was actually operating it displayed no signs of adaptation or fatigue (see especially Fig. 2B). A physical explanation thus appeared more plausible. Two further and decisive experiments were performed which demon- strated the character of this pressure effect. A guppy was placed in a bottle only partially filled with water. The water was thoroughly shaken with air at each reduced pressure. The fish adjusted its density to agree with that of the water for all the pressure decreases used. In one ex- periment the pressure was decreased by 300 mm. Hg. Again, when water was thoroughly shaken with air under a pressure increase of 500 mm. Hg, the guppy bloated when placed in this after the pressure had been restored to the original value. The rate of bloating was of the same order of magnitude as seen in response to a pressure decrease of 300 mm. Hg. A partial interpretation of the experimental results thus seems evi- dent. Except for the pressure exerted by the short column of water over the fish in the experiments, the gas in the bladder is of the same pressure as the atmospheric air at the water surface. Under reduced pressure, tension of gases in the water favors diffusion of gases into the 54 FRANK A. BROWN, JR. bladder and atmosphere alike. The greater the decrease in pressure, the more rapid the passage of gases into the bladder. Not up to this point explained, however, is the initial lower rate of bloating, sometimes even temporary gas output, which occurred when guppies were subjected to decreases in pressure sufficiently large to pro- duce the bloating phenomenon. The temporary output is seen in Fig. 1 (reduction in pressure by 125 mm. Hg) and in Figs. 2 A and 2B at de- creases in pressure by the amount 150 (when this was carried out in two steps) and 135 mm. Hg. This temporary initial activity, an ap- parent attempt of the fish at adaptation, commences strongly and then gradually over the course of two to five hours becomes wreaker until it is completely masked by the inward passage of gases. Either an out- wardly secreting mechanism which could secrete quite efficiently up to a point of fatigue, or simple diffusion alone, might account for the re- sponse. The former explanation can neglect the partial pressures of the various gases within the swimbladder. The latter explanation is based upon the assumption that at the beginning of the experiment the bladder possesses a higher percentage of CO2 than does air. Recent work of Meesters and Nagel (1934) has shown that a fish would, under the appropriate stimulation, decrease the amount of blad- der gas most rapidly when the initial percentage of CO2 in the bladder was high. Thus, fish were apparently able to eliminate this gas much more rapidly than either O2 or N2. The work of Jacobs (1932), con- firmed by Meesters and Nagel (1934), has demonstrated that a physo- FIG. 3. Responses of the swimbladder to sudden decreases in pressure fol- lowing immediately upon different experimental conditions of secretion and dif- fusion of gases to and from the bladder. A. Two experiments in which guppies are subjected to a pressure decrease of 300 mm. Hg in water equilibrated with air at the reduced pressure, then returned to atmospheric pressure in water equilibrated with the restored pressure. At the moment of adaptation to the restored pressure the fishes are subjected to a sudden pressure decrease of 300 mm. Hg. B. A single experiment in which a guppy was subjected to pressure increase of 260 mm. Hg and at the moment of adaptation subjected to a decrease to a point 300 mm. Hg below atmospheric pressure. C. Three experiments in which guppies long (20 hours to several days) adapted to a given pressure are subjected to a sudden pressure decrease to a point 300 mm. Hg below atmospheric pressure. D. Two experiments in which guppies are stimulated to decrease the gas content of their bladders and then suddenly subjected to a pressure decrease to a point 300 mm. below atmospheric pressure. E. The two upper graphs indicate the response of guppies, bloating as a result of a 300 mm. Hg decrease in pressure, when they are suddenly returned to atmos- pheric pressure. The lower graph shows the normal response of the fish upon increase in pressure of 300 mm. Hg and then, after some hours, return to atmos- pheric pressure. SWIMBLADDER RESPONSES TO PRESSURE DECREASES 55 1350 1150 950 750 u 550 D. 5 ^1350 r MI50 o 950- - B uj cc IT) V~> £15501- 1350 1150 950 750- D 10 15 20 25 30 35 40 10 15 20 10 15 20 5 10 15 20 5 10 15 20 TIME (HOURS) FIG. 3. 56 FRANK A. BROWN, JR. clistous fish, the perch, secretes into its swimbladder gas which consists of about eighty per cent CO2. These results may be applied to a physical interpretation of the observations that have been made upon the guppy. One would merely need to assume that the fish usually possesses in its bladder a partial pressure of CO2 several times that in air at the water surface. The pressure decreases of the experiments (never more than about sixty per cent) would still favor the passage of CO2 out of the bladder while one or both of the remaining two gases would diffuse inward. The magni- tude of the partial pressure of CO, inside and the pressure decrease would determine the rate of outward diffusion of this gas and this rate might even be great enough to make the volume of the bladder decrease. Meanwhile, the response of the secretory mechanism to the pressure decrease would be one of minimal activity. As the CO2 became depleted the rate of its outward diffusion would decrease correspondingly. The inward diffusion of the gases other than CO2 would also tend to decrease the partial pressure of CO2 in the bladder. Eventually, the partial pres- sures of all three gases would attain the same ratio as that of atmospheric air and then all gases would diffuse inward at rates to maintain this ratio. Indirectly, some evidence has been obtained to support the hypothesis just described. We would expect that increasing the proportion of CO2 in the bladder would permit a stronger opposition to the decrease in pressure, and a sufficient reduction in the bladder CO2 would eliminate it altogether. The following are experiments which give such support. Two male guppies were subjected to a pressure decrease of 300 mm. Hg in water brought into gaseous equilibrium at the reduced pressure. First there was a rapid increase in fish density and then after an hour or two a considerably slower one. These two rates were interpreted on the basis of the work of Meesters and Nagel (1934) to indicate an ini- tial rapid escape of CO2 and then the slower continued escape of other gases. The bladder was thus assumed to become relatively free of CO2 and also to contain a smaller quantity of the other gases than originally. The fish was then restored to the original pressure in water equilibrated with air at the restored pressure. The gas secreted into the bladder was assumed on the basis of Jacobs' (1932) results to be richer in CO, than that which was withdrawn. When the proper volume of gas for the new pressure was reached and before the amount of CO2 could decrease to equilibrium with a slower secretory rate, this fish was subjected to the sudden pressure decrease of 300 mm. Hg. This fish was able to oppose much more strongly the bloating associated with this decreased pressure than was the fish of the earlier experiments. Compare the results of experiments of this nature (Fig. 3A) with the results ob- SWIMBLADDER RESPONSES TO PRESSURE DECREASES 57 tained when guppies had neither appreciable gain nor loss in bladder gases over some days (Fig. 3C) but subjected to the same pressure decrease. The same general results were obtained in a slightly different man- ner. A guppy was stimulated to decrease its density by subjecting it to an increase in pressure of 260 mm. Hg. At the moment of adapta- tion to the new pressure the fish was subjected to a pressure reduction of 560 mm. Hg (Fig. SB). This was the equivalent of the usual re- duction in pressure of 300 mm. Hg. The reverse type of experiment was next performed. Guppies, for some hours adapted to increased pressures, were stimulated to let gases escape from their bladders. On the assumption that CO2 is the gas to pass out most rapidly these fishes should, after considerable increase in density, have no larger percentage of CO2 than is present in atmospheric air. And as expected the fish were at this moment unable to oppose in the slightest the bloating effect correlated with subjection to a sudden 300 mm. Hg pressure decrease. Figure 3D illustrates results of this nature. A final test for the hypothesis was made to ascertain that the gas present in the fish bloating rapidly under decreased pressure did not contain as large a percentage of CO2 as did the normally secreted gas. A guppy that had reached the maximum rate of bloating under the stimu- lus of the reduced pressure was suddenly restored to the original pres- sure. Under ordinary circumstances the initial response to such a decrease in pressure is a relatively rapid fish density increase which gradually or suddenly slows down. In the instance of this fish the initial response showed none of the usual initial rapidity and the whole process of return to the normal amount of gas for the pressure was relatively slow, as one would expect if the rate were determined by gases which were slow to diffuse (Fig. 3£). The normal rate of gas escape following secretion of gases into the bladder eighteen hours previously is also shown in the same figure for comparison. SUMMARY 1. A sudden and maintained pressure decrease from a value of one atmosphere to points between 625 and 300 mm. Hg results in an increase in gas content of guppy swimbladders and the rate of increase is an inverse function of the pressure. 2. This increase in gas content has been shown to be the result of lowering the swimbladder gas pressure, the gas tension of the body fluids then favoring passage of gases into the bladder. 58 FRANK A. BROWN, JR. 3. Guppies are able to oppose more or less successfully the inward diffusion of gases at first, but gradually in the course of two to five hours such opposition ceases. 4. The transitory attempt on the part of the fish at density adjust- ment in response to the pressure decreases is explainable in the presence of an initially high partial pressure of CO2 in the swimbladder. 5. The guppy appears to be unable to remove gases from the blad- der when the diffusion gradient does not favor such passage. LITERATURE CITED EVANS, H. M., AND G. C. C. DAMANT, 1928. Observations on the physiology of the swimbladder in Cyprinoid fishes. Brit. Jour. Exper. Biol., 6 : 42-45. JACOBS, WERNER, 1932. Untersuchungen zur Physiologic der Schwimmblase der Fische. II. Die Volumregulation in der Schwimmblase des Flussbarsches. Zeitschr. vergl. Physiol., 18: 125-156. MEESTERS, A., AND F. G. P. NAGEL, 1934. liber Sekretion und Resorption in der Schwimmblase des Flussbarsches. Zeitschr. vcrgl, Physiol., 21 : 646-657. VON LEDEBUR, ]., 1937. liber die Sekretion und Resorption von Gasen in der Fischschwimmblase. Biol. Rev., 12: 217-244. THE INTERNAL ANATOMY OF TWO PHALLOSTETHID FISHES LOIS E. TEWINKEL (From the Marine Biological Laboratory, Woods Hole, Mass., and the Department of Zoology, Smith College) INTRODUCTION The Phallostethidae are highly specialized little fishes found in fresh and brackish water in certain limited regions of the Malay peninsula and the Philippines. They are of general interest in that they appear to provide a unique example of change of structure and function in a typi- cal teleost organ, the pelvic fins of the male. The females are also modi- fied, but to a lesser degree. The fishes derive their name from a large, muscular copulatory organ, the "priapium," suspended below the head in the male and which contains, in addition to the ductus deferens, the terminal apertures of the digestive and excretory systems (Figs. 1 and 2). In the female, the organ outlets are also anterior but, since there is no copulatory organ, the apertures are located on the ventral surface of the gular region (Fig. 3). The phallostethids were first described and named by Regan in 1913 from a single species, but he notes a clear reference to their type by Duncker in 1904. Since Regan's work, twelve species have been dis- covered (Myers, 1928, 1937; Aurich, 1937). All are small, measuring from 14 to 40 mm. in length, and in 1935, after considerable study, Myers created for the group a new sub-order, Phallostethoidea, in the order Percesoces. Myers lists seven genera; Aurich groups them into three families and adds an eighth genus, Soleno phallus. Investigations on this group have been largely concerned with at- tempts to discover its taxonomic relationships, and a search has been made for possible homologies between the bony elements supporting the priapium and the bones of more typical teleosts. Regan (1916), Bailey (1936) and Aurich (1937) have described the osteology of the priapium in several species of phallostethids. Bailey and Aurich each conclude that the priapial bones represent greatly modi- fied pelvic fin and girdle elements, together with the first pair of ribs and, Bailey adds, possibly certain pterygials or rays of a portion of the pectoral fins. Aurich strengthens the argument that pelvic fin elements 59 60 LOIS E. TEWINKEL have been transformed into priapial structures by describing the presence of vestigial fin rays in two species of the genus Solenophallus, and in Gulaphdlus mirabilis. In all three species, he finds two or three delicate rays lying free in the skin behind the anus, and a second group of three rays inserted on the posterior extremity of the axial bone, a skeletal sup- port which extends throughout the length of the priapium (see Aurich Fig. 3). In Solenophallus, these latter rays become comb-like at the ends. The females of several species possess a pair of papillae on either side of the oviduct opening (Regan, 1913; Herre, 1925, 1926; Villadolid and Manacop, 1934; Bailey, 1936). These may be vestiges of pelvic fins, and it is thought (Bailey) that the minute bony supports found in them in Phenacostethus may be remnants of the pelvic girdle. Villado- lid and Manacop (1934) describe the development of the copulatory organ in Gulaphallus, showing it to arise from paired lobes surrounding the anus, on the ventral side of the throat. One of these lobes soon outgrows the other ; the two fuse and gradually produce the elongate, asymmetrical priapium. Brief descriptions of the soft parts of phallostethids have been pub- lished by Regan (1916) for Phallostethus dunckeri and Neostethus lan- kesteri, and by Villadolid and Manacop (1934) and Bailey (1936) for Gulaphallus mirabilis. There has been no detailed study of the gross and histological structure of the viscera, however, and it is to Dr. Hugh M. Smith that I owe the suggestion that such a study in these greatly modified forms would be of interest. Dr. Smith referred me to Dr. George S. Myers who supplied preserved material of two species, Phena- costethus sinithi, and Gulaphallus mirabilis, from the collection at the U. S. National Museum. I wish also to acknowledge with appreciation the help Dr. Myers has given me in sending me a number of papers on the phallostethids. MATERIAL AND METHODS Phenacostethus sinithi, described by Myers in 1928 from material collected by H. M. Smith in a fresh water stream in Siam, is the smallest phallostethid known (14-17 mm. total length) and is next to the minute Philippine goby, Mistichthys lusonensis (11-16 mm. total) in size. Gulaphallus mirabilis was described in 1925 by Herre as one of two species of a new genus which he had found in the fresh water streams of Luzon, the Philippine Islands. Its total length is not over 35 mm. The specimens at my disposal were fixed in formalin and alcohol and preserved in alcohol. Transparent preparations of the skeleton were made by staining with alizarin and clearing in KOH and glycerine. ANATOMY OF PHALLOSTETHID FISHES 61 These were studied and dissected under a binocular dissecting micro- scope and other unstained specimens were dissected in alcohol for a study of organ arrangement. Sketches were made with the aid of a camera lucida. Serial sections were cut in transverse, sagittal, and frontal planes, 7-1 2 //, thick, and stained in iron haematoxylin, alum haematoxylin and triosin, and in Mallory's triple connective tissue stain. In all cases, fixation was poor for an adequate histological study and much more satisfactory results could be obtained with properly fixed material. SKELETON The descriptions of almost the entire skeleton of P. smitJii and of G. mirabilis by Bailey (1936) and of the priapial skeleton of the latter species by Aurich (1937) give detailed information to which I can add nothing except to mention that Bailey did not notice the vestigial fin rays which Aurich finds in the priapium of G. mirabilis. After very careful examination, I have identified both the loose post-anal group of rays and the group articulated with the posterior end of the axial bone in that species, thus supporting Aurich (see Aurich, 1937, Figs. 3 and 8) . In the priapium of Phenacostethus, however, I have found no trace of fin rays. This fish is much smaller than Gulaphallus and the skeletal elements are correspondingly more delicate, which may in part account for the disappearance of such minute vestiges. PSEUDOBRANCH The pseudobranch, which Aurich says is lacking in phallostethids examined by him, including Gulaphallus, is present in both that species and in Phenacostethus in the form of two minute tufts (Myers, 1928) lying in a recess at the anterior dorsal end of the sub-opercular cavity. Microscopic study shows this organ to belong to type II of Granel's (1927) classification of teleost pseudobranchs, in that the epithelium of the branchial chamber surrounds each tuft separately, but does not ex- tend between the secondary lamellae of the tuft. Each primary lamella, or tuft, is furnished with an afferent artery and is supported by a pre- cartilage rod which is extremely delicate in Phenacostethus. The sec- ondary lamellae are close together and contain the large acidophile cells typical of the pseudobranch. VISCERA The general arrangement of organs in Phcnacostctlms is shown in Figs. 1 and 2. Comparison with Regan's (1916) figures of Phallo- 5 L AX 1 CT FIG. 1. Phenacostethus smithi d, internal organs in situ from the proctal side (X 16). Testis and portion of ductus deferens removed. The priapium is the large ventral appendage; several bones in it are outlined but Bailey's figures should be consulted for the complete skeleton. FIG. 2. Phcnacosthethits <$, internal organs in situ from the aproctal side. Abbreviations A anus ABL air bladder AX axial bone CT ctenactinium DD ductus deferens GA genital aperture GG gas gland GP genital papilla HT heart INT intestine K kidney L liver PR priapium PU sucker-like pulvinulus S spleen SC region of sensory canals ST stomach SV seminal vesicle T testis TX toxactinium UA urinary aperture UD urinary duct V valve in ductus deferens ANATOMY OF PHALLOSTETHID FISHES 63 stethus dunckeri and with Villadolid and Manacop's (1934) figures of Gulaphallus mirabilis will give evidence of the great general similarity in the visceral anatomy of the group. Digestive and reproductive organs are located almost entirely in the space anterior to the air-bladder and, together with the excretory ducts, these systems terminate in apertures situated on the ventral surface of the throat region. It has already been pointed out (p. 59) that this extremely anterior position is associated with the male copulatory organ or priapium which is suspended from the head. The outlets are asymmetrically placed on the priapium ; the anus and urinary duct open on one side, the proctal, while the ductus deferens ends in a genital papilla on the opposite or aproctal surface. As other workers have noted, this asymmetry is variable, that is, in some males the right side is proctal, in others, the left. In the female the apertures are mid-gular although no copulatory organ is present (Fig. 3). The following more detailed description applies to both PJicna- costetlius srnithi and Gulaphallus mirabilis unless attention is called to one species only. ENDODERMAL DERIVATIVES The digestive tract is simple, short, and without pyloric caeca; a typical carnivorous type, as is borne out by Villadolid and Manacop in their analysis of the food of Gulaphallus. The intestine makes one complete coil in Gulaphallus and a partial coil in Phenacostethus (Fig. 1) before turning anteriorly toward the anus. The liver is large and lies close to the stomach and intestine (mainly on the aproctal side in the male) ; no gall bladder was identified. The pancreas is diffuse tis- sue, lying in the bend between stomach and intestine, with considerable fat scattered among its cells. Perhaps because of unsatisfactory fixa- tion, no islet tissue is distinguishable. The spleen is partly embedded in the liver posterior to the stomach (Fig. 1). Thymus and thyroid glands are present. The large air bladder fills nearly one half of the body cavity. It lacks a pneumatic duct but is equipped in the anterior dorsal region with a vascular gas gland. NERVOUS SYSTEM The size of the brain is large in proportion to the skull as in other minute fishes (TeWinkel, 1935). It was impossible because of poor fixation to determine whether the brain and cord are modified in relation to the priapium. Neuromasts are scattered on the surface of the body, particularly on the head region in Gulaphallus; a few are present in Plicnacostcthus, but here too, fixation was imperfect. Four large canals, 64 LOIS E. TEWINKEL FIG. 3. Phcnacostcthus siniihi $, ventral view, anterior end (X 16). A. anus; HT, heart; N, neuromast; O, oviduct aperture; PAP, post anal papilla; UA, uri- nary aperture. FIG. 4. Portion of proximal coil of ductus deferens in P. sinithi. Ep, epi- thelium, showing in one cell probable position of nucleus ; dotted line represents depth of lighter area in cells ; GL, globules ; SP, sperm head. Drawn with camera lucida under oil immersion lens. FIG. 5. Portion of distal coil of ductus deferens from same section of P. smlthi as Fig. 4. Epithelial cells are much lower and globules between sperm heads much larger than in the preceding figure. (Same magnification as above.) ANATOMY OF PHALLOSTETHID FISHES 65 with patches of sensory cells (which are probably lateral line organs) on their ventral inner walls, extend longitudinally on the snout and above the eyes in both species. These occupy considerable space between the skin and skull, and occasionally a homogeneous mass, possibly a cupula (Denny, 1937), is present in contact with the sense organ. The cavity containing the inner ear is proportionally large as in Mistichthys. KIDNEY The kidney in both species is a median organ which lies dorsal to the air bladder and digestive tract and diverges anteriorly into two sections which extend from below the occipital region of the skull up to the posterior level of the gills (Fig. 1). In Phenacostethus, and possibly in Gulaphallns, the urinary ducts are paired upon their exit from the kid- ney a short distance in front of the air bladder. They soon fuse, how- ever, into a single duct which, in both species, passes between the in- testine and gonad to open in the male on the priapium, directly behind the anus (Figs. 1, 2), and in the female, posterior to the oviduct in the gular region (Fig. 3). Serial sections reveal one pair of pronephric glomeruli in both spe- cies. These lie dorsal to the heart and are directly attached to the dorsal aorta, which position is characteristic of the teleost pronephros. Poor fixation makes it impossible to tell whether the pronephric tubules are being replaced by lymphoid tissue, or whether they are functional, except the fact that the glomerular capillaries are full of blood cells, which suggests some function. The mesonephric glomeruli are smaller than the pronephric and are most numerous in that portion of the kidney anterior to the air bladder. The number of these glomeruli varies in Phenacostethus from 13 to 18, and in Gulaphallns, from 33 to 38, in the specimens where counts were made. In properly fixed material it would be more possible to determine, in so far as one could histologically, whether these fishes retain a functional pronephros in the adult, as is the case in Fierasfer, Zoarccs (Emery, 1880), Lepadogaster (Guitel, 1906) and Mistichthys (TeWinkel, 1935). The number of glomeruli in Phenacostethus is only slightly greater than that in Mistichthys (which has 12 to 13 on the average) and it would be of interest to discover whether the ratio of glomerular volume to body surface is in any way comparable in these two unrelated minute teleosts. REPRODUCTIVE ORGANS The ovary in the two species is single and lies along the anterior and ventral surface of the air bladder. In Phenacostethus there are on the 66 LOIS E. TEWINKEL average fifteen maturing ova and the largest of these measure approxi- mately 500 ft. Each ovum is surrounded by a thick porous zona radiata with surface filaments forming, but the method of filament attachment was not observed. The ovary of Gulaphallus contains about sixty ma- turing ova. The oviduct, in both species continuous with the ovary and doubtless formed by fusion of the mesovarium, is thick-walled and consists of a deep epithelial lining (possibly pseudo-stratified columnar) and a well-developed circular layer of smooth muscle. Its exterior opening lies just anterior to that of the urinary duct. The breeding habits and development of Gulaphallus have been de- scribed by Villadolid and Manacop (1934). They demonstrate that the fish is oviparous but that fertilization is internal. No sperm are present in the oviducts of either species examined by me ; the Gulaphal- lus material is immature and it is possible that Phenacostethus was col- lected at a time which was not the breeding season. The testis is an unpaired organ, on the aproctal side partly wrapped around the anterior wall of the air bladder (Fig. 2). Seminiferous tu- bules open anterio-dorsally into a common collecting tubule. Because the structure of the ductus deferens differs considerably in the two species, each will be described separately. In Phenacostethus all stages of spermatogenesis appear to be present in the testis. The ductus deferens is a long coiled tube, probably similar to that in Phallostethus dunckeri (Regan, 1913, 1916), forming a mass nearly equal to the testis in size. The tube is narrow in diameter for two or three coils, then increases to double its initial size, and finally extends into the priapium and into the genital papilla (Fig. 2). The epithelial lining of the tube is simple columnar, with tall cells in the narrower coils (Fig. 4), and cells about one-half this height in the large distal 'coils (Fig. 5), while that portion of the ductus deferens extending into the priapium is lined by low columnar epithelium. A valvular mechanism, greatly decreasing the lumen and lined by simple columnar epithelium, is present in the duct soon after it enters the genital papilla (Fig. 2). From this point the tube is slightly coiled, then dilates to form a seminal vesicle. Toward the tip of the papilla, the lumen is completely constricted and the wall seems to be composed of concentric layers of connective tissue, and very likely, some smooth muscle is also present. The ductus deferens is crowded with spermatozoa in all specimens of Plicnacostcthus studied. In the narrow proximal coils of the tube (Fig. 4), there appear among the sperms small homogeneous globules which are, in all probability, secreted by the tall columnar epithelial cells of the lining. The size of these globules increases by one-half by ANATOMY OF PHALLOSTETHID FISHES 67 the time the large distal coils are reached (Fig. 5) and this increase may be the result of fusion of smaller globules. The staining capacity of these bodies varies : in Mallory's triple connective tissue stain, they are purplish in the proximal coils, but distally, brilliant red; in iron haematoxylin, brownish proximally, and distally gray or black ; in alum haematoxylin arid triosin, uniformly pink. These globules may be gel- atinous spermatophores, but there seems to be no definite association between them and the sperms such as is described by Regan (1916) in two other phallostethids. Some spermatozoa may adhere to them, but there are quantities of free sperm in all parts of the duct. There is also a possibility that globules and sperm become united upon discharge, but study of living material is necessary to determine it. The testis in the male specimen of Gulaphallus which was sectioned is large but immature.. The ductus deferens, within the body cavity a very narrow duct probably without secretory function, dilates in the posterior portion of the priapium to form a seminal vesicle. The duct is of course entirely devoid of sperm so that the presence of sperm packets, affirmed by Bailey who evidently studied more mature speci- men's of this species, was not observed. (See PL 3, fig. 3 in Villadolid and Manacop, 1934). Study of the two phallostethids described in the preceding pages, in spite of the handicap of poorly fixed material, has shown that, histo- logically, there are no striking differences in the viscera that may be attributed to the priapium or to the displacement of the organ outlets to the throat region. The persistence of the pronephros, which may be functional in the adults, is of general interest because it occurs so rarely in teleosts ; it would be of value to examine other members of the group to determine whether that organ is always present in phallostethids. The great difference in the structure of the ductus defereris in the two species is undoubtedly related to family specialization. Possibly, the adult male Gulaphallus develops a glandular groove externally (as in N. lankesteri, Regan, 1916) to compensate for lack of abundant secre- tion on the part of the sperm duct ; such a groove was riot seen, how- ever, in the pre-adult male examined histologically. SUMMARY The anatomy of the soft parts of two phallostethid fishes, Plicna- costethus smithi and Gulaphallus m'lrabilis, has been studied. 1. Two groups of vestigial fin rays are present on the copulatory organs of Gulaphallus but are lacking on that of Phcnacostcthus. 2. A tuft-like pseudobranch, belonging to Granel's type II, is present in both species. 68 LOIS E. TEWINKEL 3. The viscera are located almost entirely anterior to the air bladder, and digestive, reproductive and excretory systems open externally on the sub-gular copulatory organ in the male and on the ventral surface of the gular region in the female. Liver, pancreas, spleen, thymus and thyroid are identified. 4. Neuromasts are scattered on the surface, and four canals, prob- ably of the lateral line type, are present on the heads of both species. 5. The kidney contains one pair of pronephric glomeruli which may be functional. There are 13-18 mesonephric glomeruli in Phena- costcthus, and 33-38 in Gulaphallns. 6. The ovary is single in both species and contains a relatively small number of ova. The oviduct, continuous with the ovary, is thick-walled and muscular. 7. The testis is a single mass. Seminiferous tubules open into a collecting tubule which continues as the ductus deferens. This duct forms a large coiled mass in Phenacostethus and is lined by simple columnar epithelium. Homogeneous globules, present among the sperms in the lumen, do not appear to form spermatophores. The ductus deferens in Gula phallus is very narrow within the body cavity, but di- lates to form a seminal vesicle in the copulatory organ. LITERATURE CITED AURICH, H., 1937. Die Phallostethiden. Mitt, der Wallacca Expedition, Zzveite Rcihe, Intern. Rev. ges. Hydrobiol. u. Hydrograph., 34 : 263-286. BAILEY, R. J., 1936. The osteology and relationships of the phallostethid fishes. Jour. Morph., 59 : 453-483. DENNY, M., 1937. The lateral-line system of the teleost, Fundulus heteroclitus. Jour. Comp. Neur., 68 : 49-65. EMERY, C, 1880. Fierasfer. Atti della R. Accad. Lincei, Roma, 3 ser., 7 : 167-254. GRANEL, F., 1927. La pseudobranchie des poissons. Arch. d'Anat. micr., 23 : 175-317. GUITEL, F., 1906. Recherches sur 1'anatomie des reins des quelques gobiesocides. Arch. Zool. Exper. Gen. (Ser. 4), 5: 505-700. HERRE, A. W., 1925. Two strange new fishes from Luzon. Philippine Jour. Sci., 27 : 507-513. HERRE, A. W., 1926. Four new Philippine fishes. Philippine Jour. Sci., 31 : 533-543. MYERS, GEO. S., 1928. The systematic position of the phallostethid fishes, with diagnosis of a new genus from Siam. Am. Mus. Novitatcs, No. 295, Am. Mus. Nat. Hist., New York. MYERS, GEO. S., 1935. A new phallostethid fish from Palawan. Proc. Biol. Soc. Wash., 48 : 5-6. MYERS, GEO. S., 1937. Notes on phallostethid fishes. Proc. U. S. Nat. Museum, 84: 137-143. REGAN, C. T., 1913. Phallostethus dunckeri, a remarkable new cyprinodont fish from Johore. Ann. Mag. of Nat. Hist. (Ser. 8), 12: 548-555. ANATOMY OF PHALLOSTETHID FISHES 69 REGAN, C. T., 1916. The morphology of the cyprinodont fishes of the sub-family Phallostethinae, with descriptions of a new genus and two new species. Proc. Zoal. Soc. London, 1 : 1-26. SMITH, H. M., 1927. The fish Neostethus in Siam. Science, N. S., 65 : 353-355. TEWINKEL, L. E., 1935. A study of Mistichthys luzonensis with special refer- ence to conditions correlated with reduced size. Jour. Morph., 58 : 463- 535. VILLADOLID, D. V. AND P. R. MANACOP, 1934. The Philippine Phallostethidae, a description of a new species, and a report on the biology of Gulaphallus mirabilis, Herre. Philippine Jour. Sci., 55 : 193-220. REGENERATION OF GONAD TUBULES FOLLOWING EXTIRPATION IN THE SEA-CUCUMBER, THYONE BRIAREUS (LESUEUR) FRANK R. KILLE (From the Marine Biological Laboratory, Woods Hole, and Martin Biological Laboratory, Swarthmore College, Swarthmore, Pennsylvania) Several accounts infer that the extensive powers of visceral regen- eration in the holothurians may extend to the gonads (Torelle, 1909; Deichmann, 1921 ; Bertolini, 1932) although they give neither precise data regarding the portion of the gon'ad involved nor even a gross analy- sis of the regenerated tissues. The present study considers the general problem of gonad regeneration in the holothurians with special reference to three specific points, namely, (1) the extent of the capacity to regen- erate gonadal tissue, (2) the presence or absence of germ cells within regenerated tubules and (3) the possible origin of any germ cells found within them. MATERIALS AND METHODS Previous experience demonstrated (Kille, 1931) that Thyone bri- arcus (Lesueur) is an extraordinarily good holothurian for experi- mental purposes. In contrast to the genus Holothnria which eviscerates through a tear in the cloaca, this genus casts off the entire anterior end of the body whenever it autotomizes the digestive system. With these parts lost, the animal is nothing but a dermo-muscular sac containing only a cloaca, respiratory trees, gonads, and mesentery. If left to itself, the circular body muscles at the extreme anterior end contract strongly so as to close off the body cavity from the sea. A healing process soon makes this closure permanent. However, if one makes a longi- tudinal cut at the anterior end equivalent to about one- fifth the length of the animal, many specimens may be turned inside out following their evisceration. In such a position the gonad tubules and the mesentery hang freely in the water. (Fig. 1, A, B, C). With the animal in this position, operation's on the gonad are possible. Most specimens can then be turned right side out with little difficulty. If the animal is re- turned to running water, a high percentage survive the operation (Kille, 1937). In some cases the circular muscles of the body wall contract so 70 REGENERATION OF GONAD TUBULES IN THYONE 71 strongly that the animal cannot be turned right side out. Such indi- viduals do not live more than a day or two. THE REPRODUCTIVE SYSTEM IN THYONE BRIAREUS (LESUEUR) A thick tuft of gonad tubules is found in each side of the dorsal mesentery as seen in Fig. 2. They originate from tissue localized within the mesentery in a region near the dorsal body wall (Fig. 2, gonad- basis) and about midway between the mouth and the anus (Fig. \, C}. These tubules empty their products into a common chamber located within the dorsal mesentery (gonad-basis chamber, Fig. 2). Anterior to the tubules this cavity continues as the lumen of a gonoduct which runs through the dorsal mesentery to open exteriorly at a point between the bases of the two dorsal tentacles. Each right and left half of the C FIG. 1. A, Thyone immediately after evisceration with a short longitudinal slit through the body wall at the anterior end (ventral surface to the right) ; B, Thyone being turned inside out with the aid of a blunt glass rod; C, specimen turned completely inside out. Cl, cloaca; G, gonad. (Mesenteries and muscula- ture of body-wall omitted.) Thyone gonad consists then of (1) a lateral portion of a chamber located within the dorsal mesentery (the gonad-basis) and (2) a mass of tubules which take their origin from the ventro-lateral wall of the cham- ber (gonad tubule, Fig. 2). The right and left tufts of tubules are symmetrically developed, there being but little variation in the number and size of tubules between the two sides. Within each tuft, however, one can distinguish tubules in various stages of development. The ex- 72 FRANK R. KILLE treme anterior ones are small, unpigmented, translucent, immature tu- bules. Such a condition has been described for many genera (Theel, 1901; Mitsukuri, 1903; Haanen, 1914; Deichmann, 1930, etc.). Pos- terior to these is a group of tubules which grade into the slender, more mature tubules possessing pigmented, opaque walls, rounded extremities and greater length. These long, opaque tubules make up the bulk of the gonad, yet the shorter, anterior tubules comprise at least a fifth of the total number of tubules present. It should also be mentioned that coelomic epithelium intestine FIG. 2. Diagrammatic representation of a cross-section through a piece of the dorsal body wall with the attached mesentery to show the relationship existing between body wall, mesentery, gonad and intestine. in the larger specimens collected during July and August there is an indication that posterior to the tuft of tubules still older tubules may have been lost. In the medium-sized specimens used in these experiments, one can say that the larger the specimen, the greater is the mass of the gonad for the tubules are not only larger but their number is greater (see Table I). Within two larger specimens whose body volume was approximately 140 cc., the count for the female was 400 tubules while the male pos- sessed 1035. REGENERATION OF GONAD TUBULES IN THYONE 73 All these data indicate that there is a continued anterior proliferation of the tubules from the gonad-basis. Histological analysis of the tu- bules supports this conclusion. The posterior tubules show various stages in gametogenesis though there are fewer mature germ cells pres- ent than in those tubules more anterior, which may be packed with well- developed gametes. The most anterior tubules, however, show no ripe gametes but enormous numbers of germ cells in the early stages of gametogenesis. At the present time, I am not prepared to give an ac- count of the seasonal variation in the gonad for all material discussed in this paper, whether normal or experimental, was collected in the months of July through October. TABLE I The total number of tubules found within a series of eleven specimens closely graded as to size. Size of Specimens Approximate Number Tubules in Each Tuft (Volume in cc.) 9 c? 200 250 4 130 4.5 110 5.8 107 6.6 150 8.0 160 10.0 12.0 185 13.0 170 14.0 200 19.0. , 170 RESULTS OF THE REMOVAL OF THE GONAD TUBULES EITHER WITH OR WITHOUT THE GONAD-BASIS Since it is possible that part of the gonad-basis is torn away with the gonad tubules whenever they are entangled in the digestive system during the process of evisceration, two types of operations were carried out. In Group I, the gonad tubules and the entire gonad-basis were extirpated from 6 female and 6 male specimens of medium-size. These were collected at Woods Hole in early August. In Group II, composed of a similar group of specimens, the gonad tubules were torn away from the gonad-basis by scraping the gonad area with the blunt back of a scalpel. This was done in order to produce a condition similar to that which occasionally exists after tubules have become entangled in a di- gestive system that is being eviscerated. Though this seldom happens in Thy one briar ens (Lesueur) at this season, it frequently occurs in certain Holothuria when the gonads are well-developed. 74 FRANK R. KILLE In both groups, the ability to feed indicated that the new digestive system was established within 15 to 22 days, which is the average period of time required for unoperated specimens (Kille, 1935) as well as for the controls to these experiments. Evidently this unusual procedure for a coelomic operation has no injurious effect. None of Group I, however, showed any signs of gonad regeneration. Eleven specimens were killed for examination at various intervals after the operation, ranging from 46 to 129 days. The other remaining specimen of the FIG. 3. Ventral view of the testes within a Tliyonc 63 days after the main mass of the gonad had been scraped away. For size comparison, the proximal quarters of 5 mature tubules of the original right testis are included on the reader's left. group was in poor condition at the end of the experiment and was dis- carded. In all cases the new digestive system and its associated struc- tures were well developed but the cut end of the gonoduct had healed blindly. On the other hand, the gonad-basis of each specimen in Group II proliferated either a right, or a left, or a right and a left tuft of mini- ature tubules. Evidently all of the very smallest tubules of the original group were not always rubbed off for some of the newly proliferated groups showed a few tubules so much larger and longer than the rest REGENERATION OF GONAD TUBULES IN THYONE 75 that it was clear that these belonged to the original gonad mass. Such groups were thus composed of a large number of newly proliferated tubules and, in addition, a smaller number of tubules which were present before the operation (Fig. 3, left gonad). This assumption is sup- ported by the microscopic examination of a number of specimens which were killed immediately after the operation. In some cases, one could identify very minute tubules which remained attached to the remnant of the gonad-basis. At least this remnant must certainly remain even after the most violent self-eviscerations of holothurians. The method of operation, however, frequently resulted in the re- moval of all the original tubules. Such an operation would easily ac- count for those newly regenerated tufts which are like the right testis in Fig. 3. All the regenerated tubules are small and equal in develop- ment. This condition stands in sharp contrast to what is found in a normal male of equivalent size and season. Typically there would be only a fourth of this number of minute tubules and they would show a graded series as to length. Further evidence that these tubules arose at the same time is furnished by their histological similarities. In each tubule, the germ cells are packed rather solidly within the tubule. In contrast to this, the longer tubules as seen in the left testis (Fig. 3) possess a germinal epithelium surrounding a definite central lumen. THE ORIGIN OF THE GERM CELLS WITHIN THE NEWLY PROLIFERATED TUBULES Since the microscopic examination of the newly proliferated tubules showed that they were packed with cells which were identical cytologi- cally with the primordial germ cells found within the tubules of the normal gonad, a careful study was made to determine the origin of these cells. Sections were prepared of normal gonads and of the gonad rem- nants left within operated but unregenerated specimens. Special atten- tion was given to the anterior region of the normal gonad where the immature tubules occur. Here the small, club-shaped tubules appear to be diverticula of the wall of the gonad-basis. Anterior to these, one finds a few, minute, knobbed protuberances of the wall, each of which possesses a solid core of germ cells. Apparently, each core originates from an independent center within the wall of the gonad-basis, for just anterior to the protuberances separate nests of germ cells are found in those tissues. In view of their size arid position, it is not surprising that they have never been described in this species. These nests of germ cells are evidently the equivalent of the " genital stolon " which is quite conspicuous in the dorsal mesentery of some holothurians (Theel, 1901 ; 76 FRANK R. KILLE MacBride, 1906) and from which the goriad-basis, gonad tubules and germ cells arise. Examination of remnants of the gonad-basis left within operated animals shows that nests of germ cells may or may not be left on either side as a result of the operation performed upon Group II. This would seem to correlate with the fact that new tubules packed with germ cells may be proliferated on either or both sides following such operations. It is assumed then that such residual nests of germ cells are the source of the germ cells found within the newly proliferated tubules. Though none of these 12 animals failed to proliferate tubules on at least one side, one would expect this to occur occasionally if this assumption is true. DISCUSSION Since only a small part of the gonad-basis is ever lost when the gonad tubules are torn out during a process of visceral autotomy, these results show that an eviscerated holothurian does not merely mature certain small, residual tubules but may actually proliferate an entirely new set of gonad tubules. The removal of the original set of tubules is, in reality, a removal of only that part of the goriad in which a portion of the germ cells are maturing. The remnant of the gonad-basis possesses a store of primordial sex cells which are then contributed to the new tubules as they are proliferated at a rate probably four times the normal. It is a case similar to that reported in vertebrates where growth of acces- sory or residual gonadal tissue may follow incomplete gonadectomy. It is apparent, therefore, that true " regeneration of the gonad " does not occur but merely a proliferation of tubules from a growth zone or area of the gonad-basis. In a period of 4 months, only a capacity for healing is demonstrated by the tissues of the gonoduct or of the gonad-basis when germ cells are absent. That the proliferation of tubules by the gonad-basis does not occur in the absence of germ cells is suggested by the facts that (1) a tubule stump (which contains no germ cells) never produces a new tubule, and (2) among the new proliferations, no sterile tubules have ever been found. It will also be noted in Fig. 3 that the tubules take their origin from an extremely localized area of the gonad-basis. This manner of origin suggests a dependence upon a restricted group of cells rather than upon a general capacity of cells widely distributed throughout the gonad-basis. It is of interest in this connection to note that in the chicken, Willier (1937) has obtained by means of chorio- allantoic grafting, male-like sex cords even though no primordial germ cells are present. A sexual gland without germ cells may also develop REGENERATION OF GONAD TUBULES IN THYONE 77 in the urodele under certain conditions (Humphrey, 1928). Further- more, Geigy (1931) has shown that in Drosophila the formation of a gonad is not dependent upon the presence of germ cells. By means of ultra-violet radiation he destroyed the primitive sex cells in the egg yet he obtained sterile ovaries and testes. This dependent relationship of the origin of tubules upon the pres- ence of germ cells in Thyone may be more apparent than real. A small group of cells at the anterior limit of the gonad area may retain a ca- pacity for tubule formation whereas cells at more posterior levels do not. The relation then might be one of position only, the removal of the germ cells which are in this region invariably involving a loss of these tubule- forming cells. Microscopically there is no evidence for such a differential capacity among the cells of the gon'ad-basis. The only striking histological difference along its antero-posterior axis is the presence or absence of germ cells. These results raise an interesting question relative to those holothu- rians such as H. parvula which undergo transverse fission in nature. In these species the gonad is located far anterior and therefore it is retained by the anterior portion. If the regenerative capacity of this genus is as limited as that of Thyone, then the posterior portions must continue their reproduction entirely by asexual means. While the occurrence of fis- sion has frequently been reported in certain species of holothurians (Crozier, 1917; Deichmann, 1921; Kille, 1936), there has been no demonstration that it is the sole means of reproduction for certain indi- viduals sterilized by the plane of a previous fission. Deichmann has reported (1921) 'feebly developed" gonads within regenerating por- tions of Holothuria (Actinopyga) difficilis. In a recent communication to the author this investigator states that these gonads were seen only within anterior portions. There is then no report in the literature that the posterior portions ever develop a gonad. The absence of any such natural evidence for the regeneration of the gonads from other tissues coupled with the data obtained through this direct experimental approach makes a strong case for the segregation of tissues with germinal po- tencies within a highly restricted region of the dorsal mesentery. As Stolte has recently (1936) pointed out, little is known concerning the cellular basis of regeneration within the echinoderm. Zirpolo (1928) believes there is a widespread totipotency among cells through- out the body of Asterias. On the other hand, studies on the genera Stichopus, Holothuria and Thyone leave us with only one established case to support such a generalization for Holothuria, namely, the forma- tion of a new gut epithelium from the cells of the mesentery in Stichopus regalis (Bertolini, 1930). In all other instances new tissues apparently 78 FRANK R. KILLE arise from remnants of the old with the one possible exception of the anterior gut epithelium within a regenerating Thyone (Kille, 1935, p. 93). Since holothurians show an unusual capacity for regeneration there has been a tendency to assume that the tissues (or a tissue) of this group possess a totipotency throughout the entire animal. In light of the evidence, this assumption is unwarranted. Apart from the prob- lem of organization, the unusual restorative powers are simply evidence of a normal capacity for cellular proliferation or migration and the maintenance of differentiation under new conditions. SUMMARY (1) The gonad in the sea-cucumber consists of two major portions: (a) a group of tubules in which germ cells are maturing and (£) a basal portion, the gonad-basis, from which these tubules arise. Failure to recognize this has probably led to the incorrect statements that these forms " will regenerate the gonad." (2) In order to test the extent of the capacity to regenerate gonadal tissue, a partial or a complete gonadectomy was performed upon Thy one briar ens (Lesueur). The gonad was exposed by turning the animal inside out after an induced autotomy of its digestive system and anterior end. (3) That this procedure had no very injurious effects is shown by the fact that approximately 95 per cent of the operated animals survived. These and the controls regenerated the autotomized parts as rapidly as did eviscerated but unoperated animals. (4) There was no regeneration of gonadal tissue following com- plete gonadectomy. (5) If only the tubules were removed, new tubules might arise from a restricted, anterior region of the gonad-basis at a rate four times the normal (at this season). (6) The only histological feature that distinguished this productive region from the rest of the gonad-basis was the presence of nests of germ cells. (7) The possibility that tubule formation is dependent upon the presence of germ cells was suggested by still other facts. No sterile tubules were ever found among the regenerated tubules. Furthermore, the sterile proximal portion of a tubule which sometimes remained after a tubule had been torn out, did not give rise to a new tubule. Finally, when the tubules were roughly torn out, that portion of the gonad-basis containing the germ cells might or might not be lost and this is corre- lated with the fact that after such treatment, a Thyone may regenerate either a right or a left, or a right and a left tuft of tubules. It is REGENERATION OF GONAD TUBULES IN THYONE 79 granted, however, that such evidence is not conclusive. Germ cell origin and tubule formation may be independently related to a totipotence pos- sessed by only a few cells in a restricted portion of the gonad-basis for which these experiments provide no test. LITERATURE CITED BERTOLINI, F., 19»30. Rigenerazione dell' apparato digerente nello Stichopus regalis, Publ. Stazionc Zoo/. Napoli, 10 : 439-447. BERTOLINI, F., 1932. Rigenerazione dell' apparato digerente nelle Holothuria. Publ. Stazionc Zoo/. Napoli, 12 : 432-443. CROZIER, W. J., 1917. Multiplication by fission in Holothurians. Am. Nat., 51: 560-566. DEICHMANN, E., 1921. On some cases of multiplication by fission and of coales- cence in Holothurians etc. Vidcnsk. Mcdd. Dansk. Naturhist. Foren., 73 : 199-206. DEICHMANN, E., 1930. The Holothurians of the western part of the Atlantic Ocean. Bull. Mus. Conip. Zoo/.. 71 : 43-219. GEIGY, R., 1931. Action de 1'ultra-violet sur le pole germinal dans 1'oeuf de Drosophila melanogaster (castration et mutabilite). Revue Suissc de Zoo/., 38: 189-288. HAANEN, W., 1914. Anatomische und histologische Studien an Mesothuria in- testinalis Ascanius und Rathke. Zeitschr. iviss. Zoo/., 109 : 185-255. HUMPHREY, R. R., 1928. The developmental potencies of the intermediate meso- derm of Amblystoma when transplanted into ventrolateral sites in other embryos : The primordial germ cells of such grafts and their role in the development of a gonad. Anat. Rcc., 40: 67-101. KILLE, F. R., 1931. Induced autotomy in Thyone. Science, 74: 396. \ KILLE, F. R., 1935. Regeneration in Thyone briareus Lesueur following induced autotomy. Biol. Bull, 69 : 82-108. KILLE, F. R., 1936. Regeneration in Holothurians. Annual Report Tortugas Laboratory, Carnegie Institution of Washington, 1935-36 : 85-86. ~ KILLE, F. R., 1937. Regeneration of the gonad tubules in Thyone briareus Lesueur following extirpation. Anat. Rec., 70 (Sup. 1) : 133-134. MACBRIDE, E. W., 1906. Echinodermata. Cambridge Natural History. 1 : 567. MITSUKURI, K., 1903. Notes on the habits and life history of Stichopus japonicus Selenka. Annot. Zoo/. Japan, 5 : 1-21. STOLTE, H. A., 1936. Die Herkunft des Zellmaterials bei regenerativen Vorgangen der wirbellosen Tiere. Biol. Rev., 11 : 1-48. THEEL, H., 1901. A singular case of hermaphrodism in Holothurids. Bihang. till K. Svcnska Vet. Akad. Handlings, Afd. IV, 27 : 1-38. TORELLE, E., 1909. Regeneration in Holothuria. Zoo/. Anzeig., 35: 15-22. WILLIER, B. H., 1937. Experimentally produced sterile gonads and the problem of the origin of germ cells in the chick embryo. Anat. Rec., 70: 89-112. ZIRPOLO, G., 1928. Rigenerazioni eterotopiche in Asterias tenuispina Lmx. Monit. sool ital, 39 : 20-21. EGG LAYING IN THE ACOELOUS TURBELLARIAN POLYCHOERUS CARMELENSIS1 HELEN M. COSTELLO AND DONALD P. COSTELLO (From the Hopkins Marine Station of Stanford University and the Zoology Department of the University of North Carolina) The few accounts of egg laying in the Turbellaria Acoela are brief and unsatisfactory. The Acoela have one or two genital pores, except Nemertoderma (Steinbock, 1931). The absence of oviducts in the Acoela (Bresslau, 1933) leaves the question of egg emergence unsolved. Von Graff (1905) concluded that the eggs of Otocelis rubropunctata, Polychoerus caudatus and Convoluta roscoffensis are liberated through the genital pore, which he considered the normal method for most Acoela. For species which possess rio oviducts he agreed with Weldon, Sabussow and Monticelli that the mouth might serve in egg deposition. Yet von Graff pointed out (p. 1960) that oviducts had been described with rea- sonable certainty for only Otocelis rubropunctata and Polychoerus cau- datus and that Amphiscolops langcrhansi possesses no formed passage. The evidence supporting von Graff's conclusions as to the role of oviducts and genital aperture was extremely meager. In his account of egg deposition in Otocelis rubropunctata (1904) he does riot mention the path of emergence of the eggs. Gamble and Keeble (1903) stated that in Convoluta roscoffensis " frequently the body breaks in two across the opening of the oviduct . . . ," i.e., presumably at the level of the genital pore. Gardiner (1895) is the only observer who has definitely stated that he witnessed the emergence of the eggs from the genital pore. He undoubtedly considered that the eggs of Polychoerus caudatus reach the female pore through the oviducts described and figured by Verrill (1893) and later by himself (1898). In these accounts the term ovi- duct was substituted for vitellarium, and the vitellaria were figured as uniting into a common duct leading directly to the female aperture. Mark (1892) found no connection between the vitellaria and the genital pore, and Lohner (1910) showed conclusively that this species possesses no oviducts. Lohner considered it probable that the eggs pass through a series of vacuoles in the marginal parenchyma and out by way of the female aperture. 1 This investigation was aided by a grant from the Society of Sigma Xi. The authors are indebted to Dr. C. E. McClung for reading the manuscript. 80 EGG LAYING IN POLYCHOERUS CARMELENSIS 81 Subsequent references to egg emergence in the Acoela are equally inconclusive. Von Graff (1911) concluded with reference to Anaperus gardineri that the eggs probably leave by way of the mouth or through breaks in the body wall. With regard to the entire group Bresslau (1933, p. 118) stated that the eggs probably pass through " einfache Gewebsliicken in das Zentralparenchym, um dann durch die Mundoff- nung nach aussen bef ordert zu werden. Gelengentlich vollzieht sich ihr Austritt noch direkter; durch Ruptur der Korperwand an irgendeiner Stelle." This latter alternative obtains in the case of Polychoerus car- melcnsis. a FIG. 1, a-e. Series of sketches to illustrate progressive stages of the process of egg laying. MATERIAL AND METHODS The reproductive system of Polychoerus carmelensis has been de- scribed and figured in a recent paper (Costello and Costello, 1938). Specimens about to deposit eggs were observed in situ by mean's of a hand lens or binocular microscope. Removal for examination invariably terminated the process. Twelve specimens were fixed during egg 82 H. M. AND D. P. COSTELLO laying, as follows: five in hot Heath's solution (50° to 60° C.), three in Lillie's, two in Worcester's (one " warm," one at 50° C.), one in Flem- ming's and one in Champy's solution. The subsequent study of hun- dreds of specimens has shown that hot fixatives are best for the general preservation of tissues. The specimens were sectioned individually and serially. B FIG. 2. Oblique sections of an egg-laying individual preserved in Lillie's fixative and stained by the Flemming tricolor method. A, through the side of the body; B, near the anterior end. To show jelly layers (/) of the egg mass, sur- rounding the worm, and the glands (#) which secrete the jelly. In A, the jelly is broken at several points (b) ; in B, the jelly mass is more complete. Magni- fication 43 X. OBSERVATIONS An individual about to deposit eggs assumes a resting position on the substratum. The anterior end of the body is at first elevated (Fig. la), as the animal sways slowly from side to side, and is then bent ventrally in progressive stages (Fig. \b). Meanwhile a jelly-like slime is secreted from the surface of the body and adheres to the substratum. As the anterior portion of the body continues the ventral curling, the posterior end relinquishes its attachment, and curls ventrally under the approaching anterior tip, forming a ball which is attached to the sub- EGG LAYING IN POLYCHOERUS CARMELENSIS 83 stratum only by the secreted jelly (Fig. Ic). During the curling, and subsequently, violent contractions are observable in the region of the vitellaria. The animal then begins to rotate, end over end, rather slowly within the gelatinous mass. During rotation, additional jelly is secreted by the surface of the animal (Fig. 2). Meanwhile, the eggs are emerg- ing from the body of the animal, but are, for the most part, hidden within its folds (Fig. Id). An occasional egg escapes laterally, and may then pass freely over the dorsal surface of the rotating animal (Fig. 5), within the jelly. Rotation continues for three or four minutes, after which the animal unrolls and works its way out of the jelly mass (Fig. Ic} through an opening so small as to greatly constrict the body. The eggs are left behind in the central cavity of the jelly mass, which con- tracts and forces the eggs into close proximity. The entire process of egg laying requires only five or six minutes. Egg laying occurs infrequently during the early portion of the day, most frequently shortly after sunset (Table I). That egg laying occurs also during the early night is evidenced by large numbers of egg masses found in the morning in late cleavage stages. The point of emergence of the eggs is concealed by the position as- sumed by the body during the process. Living individuals forcibly un- folded during egg emergence usually extruded no more eggs. In one case, however, an egg was observed to emerge through the ventral body wall at the level of the bursa. The fixed material showed very clearly the position occupied by the mature eggs immediately before extrusion, and the point at which the eggs are shed, although no individual egg was fixed in the act of emerg- ing. Of the twelve specimen's, one had shed all its mature eggs, three had shed none, the others, a portion. It has been noted (1938) that the oocytes move dorsally within the vitellaria as they mature. However, eggs undergoing maturation occupy the greater part of the dorso-ventral thickness of the vitellarium. Im- mediately before deposition the mature eggs are located medially and ven'trally between the body wall and the ventral surface of the bursa. The vitellaria are displaced laterally and the bursa is displaced dorsally, curving around the eggs as shown in Fig. 3. This arrangement was evidenced by the eleven fixed specimens which retained one or more mature eggs. Evidence as to the point of emergence of the eggs is limited to the nine specimens which had shed one or more eggs at the time of fixation. There was no indication of the emergence of eggs by way of the mouth and intestinal parenchyma, nor by way of the vagina. Three of the five best preserved specimens showed perfectly definite points of rup- 84 H. M. AND D. P. COSTELLO s ffl o -s: 13 s a/l S • H 1 t s § ' "XO o g " tM I \ \ I \ \ \ C \ ~~ I \ „ 0* \ X \ n ] \ \ *> 00 I \ •+ \ \ \ ~ lO aj \ \ \ - \ - ft, vo , « „ \ I \ \ ~ - \ \ \ \ \ ™ \ cs \ \ -\l\\\ - \ \ \ 7 \ \ \ I \ \ ts 1-H \ \ \ \ \ \ \ o \ \ \ - - \ cs 0 \ \ 3 \ 00 \ 00 I \ \ \ \ \ J, \ \ i C oo '" 4^j O H EGG LAYING IN POLYCHOERUS CARMELENSIS 85 ture of the ventral integument in the region of the bursa, with the eggs in such a position as to indicate their emergence at this point. The best two cases are illustrated by Figs. 4 and 5. In Fig. 4, the section passes through the flattened vagina (zr), the female aperture (§), the ventral protuberance in which two of the mature eggs appear, and <£V-,*;v • - . V 4s' W7^''" "" -•'>"*•' "^''^' FIG. 3. Oblique frontal section of an egg-laying individual, to show the median ventral position of the mature eggs (c), and the displacement of the bursa (bs) around the eggs. Hot Heath fixation, stained in Heidenhain's haematoxylin. Magnification 72 X. i[>, intestinal parenchyma; t, testis follicles; v, vagina; -rd, vas deferens ; vi, vitellarium. through the undisturbed mouth (//;,) and short pharynx (/>//)• The ventral integument shows a rupture (r) about 50 microns in width near one of the shed eggs (se). It is medial, extending through seven con- secutive sections (70 microns). A jelly layer (;') surrounds the shed eggs and is connected with the ventral surface of the body near the point of rupture. The unshed mature eggs converge toward the point of m -mf- '^sspn -Jo- *" • v ••" - -• *•"* ^&/" *i3v^V' ' "* ',:;': ;Siif^:'^%" fi^§V^'v!-;'''Vv'*'l*ji" \\A ./; - U^""^-^- ^^Vv' -•; W'^ . "•-- o -m :PK -t -se -r o -J -9 _v f^ ""7'Af ^'?;' -'•" ., r:i - ^^f^Jff^r. se_. - .. • -'. : ', -i^. !> •:' -;- r ^ '.^- ->; • ;.:.-: - '; - ' " > -^ , -v'¥^^>, ^%23fe::-k ^''^ ~^&£&'<$£&. - - -:.r •-/• r -S WSK^S " " •^"^-:i K^^?ifi :p%^i^i^^^^_ r / - i*-'-'^r":-^%sSrv , '. - - :.-^^%J S>' -ti^^: i '..- ^-'^d^[lo FIG. 4. Same individual as Fig. 3; twentieth section (0.20 mm.) ventral to that of Fig. 3. To show the rupture of the ventral integument (r), and the jelly strand connecting one of the two already shed eggs (se) with this point. The section passes obliquely frontally through the mid-region and nearly transversely through the mouth (in) and pharynx (/>/)). Magnification 50 X. 5, female genital aperture ; o, immature egg ; other abbreviations as above. FIG. 5. Oblique sagittal section of an individual fixed after the extrusion of all of its eggs. To show the broken ventral integument and disrupted bursa, also the undisturbed mouth and intestinal parenchyma, and the flattened vagina. Hot Heath fixation, stained in Heidenhain's haematoxlylin. Magnification 62 X. Ab- breviations as above. EGG LAYING IN POLYCHOERUS CARMELENSIS 87 rupture, as shown by the sections of the same individual represented in Figs. 3 and 4. The second specimen, represented in Fig-. 5, shows the relation of the bursa, ruptured integument and shed eggs. In this case also the condition of the mouth and the extremely flattened vagina pre- sent striking evidence that these structures are not concerned in egg deposition. The ventral integument in the region of the bursa is dis- turbed and lacks cilia. Bits of tissue arid jelly extend from the last shed egg to the region of the rupture. The bursa is greatly vacuolated, a portion of its tissue extending to, and apparently partially closing, the rupture. The disruption of the bursa results indubitably from the pres- sure of the eggs in their outward passage, combined with the violent muscular contraction's observable in this region throughout the egg-laying process. TABLE II Distribution of the numbers of eggs per egg mass deposited in fingerbowls by animals collected 8/20/37, from the third to eighth day in tlie laboratory. Numbers of Eggs per Egg Mass Total Numbers of Egg Masses 1-5 6-10 11-15 16-20 21-25 8/23. 23 3 14 3 5 72 4 34 13 18 27 3 19 19 12 3 0 2 5 2 1 0 0 0 0 126 10 69 40 37 8/24. 8/26* 8/27. 8/28. Totals 48 141 80 12 1 282 * Two-day interval. In the remaining seven specimens the exact point of rupture was not clear as a result of either the plane of sectioning, or of distortion following slow fixation. However, they correspond, in all essential features, with the two cases described above. If Polychoerus carmelensis regularly discharges its eggs through rupture of the body wall, evidence of this should be observable in mature specimens fixed at random. Histological examination has shown fre- quent irregularities in the epidermis, parenchyma arid bursa wall, in- cluding, in some cases, the interruption of epidermal ciliation. The general appearance suggests regeneration following recent disruption. A few miscellaneous observations regarding the breeding habits of Polychoerus cannelensis in the laboratory and in the tide-pools may be here recorded. Animals brought into the laboratory and placed in finger- bowls do not, as a rule, begin to deposit eggs (or continue egg deposi- 88 H. M. AND D. P. COSTELLO tion) until the third day thereafter. During the following three days egg clusters are deposited in abundance, after which they become less numerous. The individual worms do not usually deposit all of their eggs at once, as has been suggested for the closely related Polychoerus caudatus. This is attested by the fact that the number of egg masses obtainable from a single dish may exceed the number of worms present. Data relating to the numbers of eggs per egg mass are given in Table II. The figures represent the numbers of egg masses deposited in five fingerbowls during the first eight days in the laboratory. There appears to be a definite breeding season for this species in the Monterey region. Collections in 1937 were made at intervals from June 22 through July and August. In June no egg masses or young worms were found in the tide pools. By July 8 they were found in abundance on pebbles and shells and in the folds of the Ulva. Egg laying continued in the tide pools and aquaria through July and August, although in August there occurred an increasing number of apparently spent, large-sized individuals with reproductive system partially or com- pletely absent. Thus, in 1937, the breeding season began in late June or early July and declined toward the end of August. Less extensive observations indicated that the same situation obtained during the season of 1936. DISCUSSION The foregoing account of egg laying in Polychoerus cannelensis pro- vides evidence in support of the hypothesis of von Graff and Bresslau that the eggs of some of the Acoela may be discharged through breaks in the body wall. Convoluta roscoffensis is the only other species of this group in which egg laying has been described as involving rupture of the parent worm. In that species, however, the body of the worm is frequently broken completely in two at the time of egg deposition. The regenerative powers of the Acoela have been demonstrated by the experiments of Stevens and Boring (1905) and of Child (1907) for the Pacific Coast Polychoerns, and by Keil (1929) for the Atlantic Coast Polychoerns caudatus. The healing of injured surfaces occurs remark- ably rapidly in Polychoerus cannelensis. For example, when worms are cut from the posterior end almost the entire length of the body, the two almost separate halves are quickly brought into contact by their normal locomotor activities, and heal within about ten minutes. A small notch may remain at the posterior end. but frequently all trace of the cut is obliterated. For organisms which heal and regenerate so readily the emergence of eggs through rupture of the body wall, in the absence of oviducts, may be regarded as a natural and direct method. EGG LAYING IN POLYCHOERUS CARMELENSIS 89 SUMMARY The egg-laying habits of Polychoerus carmelensis are described, with especial emphasis upon the path of emergence of the eggs. The mature eggs pass from the two vitellaria ventrally and medially between the ventral integument and the bursa seminalis, and emerge to the exterior through a median rupture at the level of the bursa. The jelly mass is secreted by the entire surface of the animal during repeated rotation, head over tail. The eggs emerge during rotation, probably aided by violent muscular contractions observable at this time. The entire process of egg laying requires five or six minutes. LITERATURE CITED BRESSLAU, E., 1933. Turbellaria. Kukenthal-Krumbach Handbuch der Zoologic, 2 : 52-293. CHILD, C. M., 1907. The localization of different methods of form-regulation in Polychoerus caudatus. Arch. f. Entw.-mech., 23 : 227-248. COSTELLO, D. P., AND H. M. CosTELLO, 1937. Egg laying in the acoelous turbel- larian Polychoerus carmelensis. (Abstract) Anat. Rec., 70 (Suppl.) : 138. COSTELLO, H. M., AND D. P. COSTELLO, 1938. Copulation in the acoelous turbel- larian Polychoerus carmelensis. B'wl. Bull., 75 : 85-98. GAMBLE, F. W., AND F. KEEBLE, 1903. The bionomics of Convoluta roscoffensis, with special reference to its green cells. Quart. Jour. Micr. Sci., 47 (N. S.) : 363-431. GARDINER, E. G., 1895. Early development of Polychoerus caudatus, Mark. Jour. Morph., 11 : 155-176. GARDINER, E. G., 1898. The growth of the ovum, formation of the polar bodies, and the fertilization in Polychoerus caudatus. Jour. Morph., 15: 73-110. GRAFF, L. VON, 1904. Marine Turbellarien Orotavas und der Kiisten Europas. 1. Einleitung und Acoela. Zeitschr. f. zviss. Zool., 78: 190-244. GRAFF, L. VON, 1905. Turbellaria, Erste Unterklasse : Acoela Ulj. H. G. Bronn, Klassen und Ordnungen des Tier-Reichs. Bd. 4, Lief. 68-74, S. 1902-1984. Leipzig. GRAFF, L. VON, 1911. Acoela, Rhabdocoela und Alloeocoela des Ostens der Vere- inigten Staaten von Amerika. Zeitschr. f. wiss. Zool., 99: 1-108. KEIL, E. M., 1929. Regeneration in Polychoerus caudatus Mark. Biol. Bull, 57 : 225-244. LOHNER, L., 1910. Untersuchungen iiber Polychoerus caudatus Mark. Zeitschr. f. wiss. Zool., 95: 451-506. MARK, E. L., 1892. Polychoerus caudatus nov. gen. et nov. spec. Festschrift zum 70. Geburtstage R. Leuckarts. S. 298-309. Leipzig. STEINBOCK, O., 1931. Ergebnisse einer von E. Reisinger und O. Steinbock mit Hilfe des Rask-0rsted Fonds durchgefiihrten Reise in Gronland 1926. 2. Nemertoderma bathycola nov. gen. nov. spec. Vidensk. Medd. fra. Dansk. Naturhist. Foren., 90 : 47-82. STEVENS, N. M., AND A. M. BORING, 1905. Regeneration in Polychoerus caudatus. Jour. E.vper. Zool, 2 : 335-346. VERRILL, A. E., 1893. Marine planarians of New England. Trans. Conn. Acad., 8 : 459-520. THE EFFECT OF THE REMOVAL OF PERISARC ON REGENERATION IN TUBULARIA CROCEA EDGAR ZWILLING (From the Department of Zoology, Columbia University and the Marine Biological Laboratory, Woods Hole, Mass.) ' INTRODUCTION In 1903 Morgan observed that small pieces of Tubularia stems, 'which ordinarily regenerated hydranths at both ends, formed single hydranths at only one end when the other end of the stem was set upright in sand. Regardless of whether the distal or the proximal end was placed in the sand the exposed end was the one which formed the hydranth. Barth (193S£) extended these observations and showed that merely placing a glass capillary over the cut end of a stem would result in a, complete inhibition of regeneration at the covered end. He also observed that when the coenosarc terminated a few millimeters below the level of the perisarc no regeneration occurred. It was suggested that oxygen lack was the probable cause for this inhibition, and that, in the last-mentioned case, the perisarc did not permit sufficient Oo to reach the coenosarc for regeneration to occur. A logical experiment arising from these observations was one in which the perisarc was re- moved from a region along the length of the stem in order to expose the coenosarc to sea water. Morgan (1903) found that when long slivers of the stem of Tubularia (both perisarc and coenosarc) were cut out, the portions of the stem adjacent to this cut regenerated in the same way that two separated portions of a stem would regenerate, i.e., a hydranth formed at either side of the cut, with the oral ends facing the cut region. The two hydranths so formed were connected at the oral ends during the initial stages of regeneration but separated after a while. Goetsch (1929) briefly reported that when all cut ends of a stem of Pennaria were sealed with wax, the removal of some of the perisarc was sufficient stimulus to cause regeneration at the exposed region. Peebles (1931) attempted to get lateral outgrowths from stems of Tubularia by " wounding the perisarc " but failed to do so. The present work was an attempt to work out in some detail the consequences of exposing the coenosarc in Tubularia. These experiments were per- formed in a preliminary way in the summer of 1937 (when it was found 90 REMOVAL OF PERISARC AND REGENERATION 91 that exposing the lateral coenosarc did result in hydranth formation, especially if the ends were ligatured) and were repeated and quanti- tative data gathered in the summer of 1938. I should like, at this point, to thank Professor Earth for his many suggestions which were so help- ful throughout the course of these experiments. METHOD • Essentially the same technique as the one described by Earth (1938a) was used for selecting the stems, keeping them, and for performing the necessary calculations. Straight, unbranched stems were selected from colonies of long (8-9 cm.) stems. These were selected for uniformity in thickness, etc. The hydranth plus 3-5 mm. of the stem was removed and the required length of stem was cut from the region immediately proximal to this. The ends were then ligatured as desired and the peri- sarc was removed. This removal was effected by means of a pair of iridectomy scissors used under a binocular microscope. Ligatured stems were selected at random for the various groups in an experiment. The desired measurements (the area of the exposed coenosarc and the di- ameter of the stem) were made with an ocular mjcrometer. The stems were kept on a gauze platform in a dish where aerated sea water was circulating constantly. To insure uniformity of conditions all of the stems involved in any one experiment were kept together in the same dish. The rate of regeneration was obtained by measuring the length of the regenerant at the time it was constricted from the rest of the stem, and employing the time required for this to occur as t1 in the formula R = irrsL/tl (see Earth, 1938a) where r is the radius of the stem at the region involved, and L is the length of the primordium. In cases (Fig. 1 — 1 C) where the regenerant was not a complete cylinder the volume was obtained indirectly by employing a bit of rubber tubing as a model for the stem. A sector of the tubing equivalent in size and shape to the regenerant was cut out and weighed. It \vas found that this was from one-third to one-half the total weight of the tubing. The volume of the actual cylinder of stem was calculated using the longest length of the regenerant as L, and then the necessary one-third or one- half was subtracted from this figure. All of the data presented are averages for the various groups considered. Only stems showing re- generation within 72 hours after the beginning of the experiment were considered in the calculations. 92 EDGAR ZWILLING B B B FIG. 1. Diagrams to show the effect of increasing the size of the exposed area upon the morphology of the regenerants. One, 2, and 3 are the three size groups. A indicates the amount of perisarc removed. B shows the beginning of pigment deposition. C shows the constriction of the regenerant from the rest of the stem and the appearance of the tentacle buds. D represents the condition of the regenerant after it has emerged. REMOVAL OF PERISARC AND REGENERATION 93 RELATION BETWEEN THE AREA EXPOSED AND THE RATE OF REGENERATION After preliminary experiments indicated that the exposure of the coenosarc was a sufficient stimulus for regeneration, an attempt was made to see whether there was any relation between the area exposed and the rate of regeneration. If , as was suspected, exposure to oxygen was the cause for regeneration a greater amount of exposed tissue should result in a larger regenerant. This was found to be the case. Table I presents data from two experiments which illustrate this. EX- TABLE I Fifteen-millimeter stems. A, B and C are the three size-groups for the areas exposed. No., number of stems used; Diam.. diameter in micra; Area in square micra ; T,, time in hours; No. Reg., number of regenerants ; Rate, rate of regeneration. B )Y o fc (J ° 1-1 LJI a *l EX P.I NO. 10 9 10 DIAM. 769 789 802 AREA 55073 153793 281559 T, 43.7 48.1 47 NO. REG. 4 8 9 RATE 36 76 137 EXP.2 NO. 10 10 10 DIAM. 457 490 477 AREA 48152 132793 164962 T, 38.5 38.6 32.4 NO. REG. 7 8 9 RATE 71 89 129 periment 1 was performed early in June when the stems were thick and experiment 2 was a repetition of the experiment performed in July when the young thin stems prevailed. The results are essentially the same as far as rate of regeneration is concerned. As the area of the exposed region was increased, the rate of regeneration went up. Since the average time in which this regeneration occurred was approximately the same for all of the groups in any one experiment, it can be safely assumed that it was the amount of tissue converted in a unit of time that was increased as more coenosarc was exposed. (See Earth, 1938^, where both ti and L vary with the oxygen tension.) 94 EDGAR ZWILLING PLATE I t 1. A central regenerant (as in Fig. 1 — 1C) at the constriction stage. 2. A central regenerant (as in Fig. 1 — 3C) before constriction. The tissue is slightly contracted. 3 and 4. Single hydranths (as in Fig. 1 — 1) after emergence. 5-12. Various degrees of fusion of the two emerged hydranths. (10 was cut from the stem at the point of emergence.) 13. Looking down on two complete hydranths which are connected at the hypostomes. (Figs. 2 and 13 were photographed by Dr. A. N. Solberg.) (The author wishes to thank Mr. Jack Godrich for his assistance in preparing the graphic mate- rial for publication.) REMOVAL OF PERISARC AND REGENERATION 95 MORPHOLOGICAL AND POLARITY VARIATIONS When thicker stems were used (Table I, Experiment 1 and pre- liminary) in the above experiment it was found that in addition to a variation in the rate of regeneration there was a variation in the mor- phology and in the polarity of the regenerants which depended roughly on the amount of tissue exposed. As depicted in Fig. 1 and in Plate I, there were three main categories of regenerants. When the exposed area was small (Fig. 1, case 1 ; PI. I, 1,3 and 4), only a limited amount of the coenosarc immediately surrounding it was involved in regenera- tion. In these cases the tissue involved in hydranth formation did not extend completely around the stem but formed a curved disc. When the proximal tentacle buds appeared they radiated out from the exposed central area. The distal tentacles frequently formed during the later stages of regeneration or during emergence. Such regenerants emerged as complete single hydranths whose polarity, i.e., oral-basal axis, was at right angles to the original polarity of the stem. In most of the cases where a greater amount of coenosarc was exposed (Fig. 1, case 2) the tissue involved in regeneration extended completely around the stem but was not symmetrical. A longer portion of the stem was involved in hydranth formation in regions immediately adjacent to the opening, while less and less of the stem was involved the farther it was removed from the opening. Such regenerants formed cylinders with oblique ends and had the appearance of trapezoids when they were viewed from the side (Fig. 1, case 2, B and C}. These emerged as two more or less com- pletely fused hydranths (PI. I, 5 to 13). The extent of the fusion de- pended on the length of the region farthest from the opening. Where this was short the two hydranths were completely fused, as this increased in size the bases of the hydranths became separate but the oral portions were in common. All degrees of such fusion were encountered. The oral-basal axis in these cases was at various oblique angles to the orig- inal axis. In a few cases very large areas were exposed (the perisarc was removed from about one-half the surface of the stem for a length of 2-3 mm.) and a complete cylinder of coenosarc was involved in hydranth formation. The tentacle buds formed symmetrically around the stem (Fig. 1, case 3, C) and when the two hydranths emerged they were complete and mirror-images of each other. They remained fused at the hypostomal region and did not separate even when kept for 10 days. Here the original polarity was retained by the proximal regenerant and was completely reversed in the distal one. These results were similar to Morgan's cases (cited above) except that in his experiments the two hydranths separated. This difference is probably due to a mechanical factor, i.e., in Morgan's cases there was a thin bridge of tissue between 96 EDGAR ZWILLING the two hydranths and this probably broke very easily when the re- generants contracted, while in the present cases there was a thicker con- nection which was not broken. INJURY TO TISSUE In the light of Child's (1927&) experiments on Corymorpha, where laceration of the tissue (in this case not covered by a perisarc) was sufficient to cause regeneration, the work described above was open to the valid criticism that injury to the tissue and not exposure to the sea water was the stimulus for regeneration. The following experiment was performed to settle this question : Three groups of stems (Table II) TABLE II Fifteen-millimeter stems. A is the group in which the perisarc was removed without injury to the coenosarc. B is the group in which the perisarc was re- moved and the coenosarc was lacerated. C is the group where the coenosarc was lacerated and where the perisarc was replaced. Legend as in Table I. B U S Ju U — S » U — a — Xj NO. 9 10 10 DIAM. 725 745 738 AREA 189488 177805 187638 T, 45.5 47.4 — • NO. REG. 6 5 0 RATE 64.6 67.5 0 were selected and ligatured at both ends. The perisarc was removed very carefully in the first group and whenever there was any suspicion that coenosarc had been injured the stem was discarded. In the second group the coenosarc was deliberately lacerated after the perisarc had been removed. In the third group a three-sided flap was made in the perisarc and the underlying coenosarc was lacerated after the flap had been deflected. The flap of perisarc was then allowed to fall back into position — covering the lacerated tissue. Table II shows that the first two groups regenerated at approximately the same rates, while the third group did not regenerate at all. This shows conclusively that exposure to sea water is, in itself, a sufficient stimulus for the initiation of re- generation. REMOVAL OF PERISARC AND REGENERATION 97 DOMINANCE RELATIONS Since all of the experiments described above were performed on stems which were ligatured at both ends an attempt was made to see whether the open ends of a stem would affect the regeneration of an exposed region of that stem. Three experiments were performed in each of which four groups of 20 mm. stems were set up. In group A both ends were left undisturbed ; in group B the distal 2-3 mm. were ligatured so that there was no continuity between that region and the rest of the coenosarc ; in' group C the proximal 2-3 mm. were ligatured ; and in group D both ends were ligatured. Perisarc was then removed from the stem at various levels. In Table III the data for an experiment in TABLE III Twenty-millimeter stems. Showing the effects of the ends on a central re- generant. The openings indicated were 5 mm. from the distal ends. A, both ends open ; B, the distal end ligatured ; C, the proximal end tied off ; D, both ends liga- tured. Legend as in Table I. B C D V rts -fc I D — I un } L° Jk NO. 9 9 10 10 DIAM. 659 682 675 708 AREA 170053 145365 157562 192508 T, _ 36.3 34.3 39 NO. REG. 0 3 3 4 RATE 0 57.7 34.9 115.5 which the perisarc was removed from a region 5 mm. from the distal end is presented. Table IV contains the data for an experiment in which the perisarc was removed from a region which was 10 mm. from the distal end; and Table V contains the data when the perisarc was removed from a region which was 15 mm. from the distal end. It is apparent, when these data are examined, that the ends, especially the distal end, exert a strong inhibiting influence (dominance) on the cen- tral regenerant when these ends are in continuity with the tissue of the stem. When this continuity is broken by a ligature the inhibition is no longer present and the rate of regeneration of the central region in- creases greatly. The proximal end by itself does not exert nearly as much of an inhibiting influence as does the distal. In Tables IV and V it can be seen that when only the distal end was tied off the rate of re- generation was much higher than when' the proximal end was tied off. 98 EDGAR ZWILLING TABLE IV Twenty-millimeter stems. The opening is 10 mm. from the distal end. Rest of table as in Table III. B \ n ' 1 °— 3 3— to 1 ° h NO. 10 9 10 9 DIAM. 632 586 603 597 AREA 184778 179731 165931 139379 T, 44 56.6 52 50 NO. REG. 2 5 2 7 RATE 9.1 42.8 8.9 95.6 In fact the distal end exerted just about as much dominance by itself, in these cases, as did both ends combined. C in Tables IV and V is as low as A. In all three experiments there is a marked increase in the rate of regeneration when the inhibiting effect of both ends is removed. The one group which is in apparent discord with these data is group C in Table III. Here, despite the fact that the distal end is not liga- TABLE V Twenty-millimeter stems. The opening is 15 mm. from the distal end. Rest of table as in Table III. T B"T (Hi D * I°3d dc isfa * *^ ! - NO. 10 10 10 9 DIAM. 553 512 541 514 AREA 142567 117155 161352 111024 T, — 44.5 — 47.6 NO. REG. 0 6 0 9 RATE o 39.3 O 68.9 hired, the rate of regeneration of the central region is fairly high. This is probably due to the fact that the perisarc was removed from a region which was fairly close to the distal end (appr. 5 mm.). The same fac- tors responsible for the appearance of bipolar regenerants in small pieces of Tubularia are probably responsible for the high rate of regeneration of the central regenerants in this group. This has been explained (Morgan, 1901) as being due to the fact that these tissues are very REMOVAL OF PERISARC AND REGENERATION 99 close together and are probably metabolizing at about the same level and thus do not inhibit each other. Whether or not this is the case, the evidence still indicates definitely that the ligaturing of the ends will result in a great increase in the rate of regeneration of a central exposed re- gion and that the ends do exert a dominance over a central region when they retain continuity with such a region. Finally, an experiment was performed to determine whether domi- nance was exerted by a more distally placed exposed region over a more proximally placed exposed region. In Table VI we can see that TABLE VI Fifteen-millimeter stems. Both ends ligatured. A, the exposed region at the distal part of the stem ; B, the exposed region at the proximal end of the stem ; C, an exposed region at both ends. Legend as in Table I. B rgr ^fa T¥ — v- Otglisfa T nil D P NO. 10 10 10 DIAM. 503 514 521 538 AREA 149021 140021 163738 173309 T, 42.4 48.6 44.7 NO. REG. 5 6 7 0 RATE 41.3 45.9 60.2 0 this is definitely the case. In this experiment 15 mm. stems were used and both ends were ligatured close to the cut surfaces. The two ex- posed region's were separated by about 10 mm. The distal exposed area; though slightly smaller, completely inhibited the proximal one. Groups A and B in this table serve as the controls. DISCUSSION For the first few hours after its exposure the coenosarc of Tubularia is in a condition approximating the early indeterminate stages of an embryo (e.g., the ectoderm of an amphibian embryo). There is a tend- ency to continue its usual vegetative activity and produce more peri- sarc to replace that which has been removed. Sooner or later, however, what seems to be a competition between the tendency to retain the vegetative state and the tendency to form hydranth material begins to manifest itself. A visible evidence of this is the appearance of the pink pigment which typifies the first stage in regeneration. Now, de- 100 EDGAR ZWILLING pending on the existing conditions (size of the opening, dominance of another region, or inherent conditions within the stem), the coenosarc either becomes definitely determined to form hydranth tissue or retains its original condition. In the former case a hydranth forms and eventu- ally breaks through the thin perisarc which had been secreted in the earlier stages. In the latter the pigment disappears after a while and the coenosarc continues to produce more perisarc until the site of the removal cannot be distinguished from its surroundings. Occasionally, when the opening is very small or when another region is dominating it, the exposed coenosarc does not even begin to regenerate (i.e., no pink pigment is deposited) but continues to form more perisarc as though there had been no interruption. The analogy to amphibian material is made more striking when it is shown that there is a period after which the determination of the hydranth- forming material becomes irreversible. This has been shown by Peebles (1931) and by another method by the author in some unpublished preliminary work. The analogy breaks down, however, when the stimulus for the determination is considered. Although the primary stimulus for form and organ formation in am- phibian embryos may come from an outside source, the work of the past few decades has indicated that the immediate cause of determination of a given structure comes from adjacent tissues. In Tubularia, as well as in many other hydroids (Child, 1927a) the stimulus for hydranth formation comes from outside of the organism. (The various parts of the hydranth may be determined as in the Amphibia). Many ex- ternal agents have been cited (Child, 1929) as the stimulators for re- generation. In referring to the experiments in which stems of Tubularia were either placed end down in sand or in which the end was capped with a glass capillary, both Morgan and Barth suggested that an oxygen lack might be responsible for the ensuing inhibition. It was implied that the presence of sufficient O2 would allow regeneration to proceed. Accu- mulated wastes rather than O2 lack was not excluded as a possible cause of the inhibition. The present experiments do exclude this since the sea water in which the stems were kept was well aerated and was kept in constant circulation (the glass-capping experiments were repeated under these conditions and the same results were obtained). It seems likely, therefore, that the inhibition of regeneration' is due to oxygen lack ; and that when an adequate supply of O2 reaches the tissue it serves as a stimulus for regeneration. This idea is rendered more plausible when the effects of varying the oxygen tension on the rate of regenera- tion are considered. Barth (1938^) found that both the length of the regenerant increased and that the time in which constriction of the re- REMOVAL OF PERISARC AND REGENERATION 101 generant occurred decreased as the oxygen tension was increased. This indicates that as the only variable in these experiments, i.e., oxygen, was increased a greater amount of tissue was converted into hydranth tissue in each unit of time. This bears a definite similarity to the experiments cited above in which the size of the regenerant increased when more coenosarc was exposed to the influence of the sea water and suggests a similar causal relationship. The similarity between the action of oxygen and the exposure of the coenosarc is made more apparent by some un- published work of Dr. J. A. Miller. (Dr. Miller was kind enough to grant the author permission to mention some of his material in relation with the present work.) This work of Miller's concerns the effects obtained with varying the oxygen tension of the sea water in which small bits of Tubularia stem were kept. These small stems ordinarily yield bipolar regenerants. As the tension of the oxygen was increased, the number of incomplete bipolars was increased. This indicates that the various regions of the regenerant increased in size and that the central portion of the small stem was used for more distal structures instead of for the more basal regions which they would have formed in a lower concentration of oxygen. Since this increase in primordia size occurred equally at both ends, there was not enough material to form complete hydranths and two symmetrical incomplete forms re- sulted— more or less of the basal regions were missing. These results are, to some extent, the converse of the results obtained when the peri- sarc was removed from the stem. When the area exposed was small the regenerant was small and a single complete hydranth formed, — just as the bipolars were usually small complete hydranths when the oxygen tension was low. When the openings were sufficiently large more tissue was involved and the regenerants were enlarged and were more com- plete. When Miller increased the oxygen tension the bipolars became more incomplete for the same reason, i.e., the regenerants became en- larged but at the expense of their basal regions. The striking similarity between the effects of exposing the coenosarc and increasing the oxygen tension leads one to conclude that the former is also probably the effect of the oxygen in the sea water. The experiment where the perisarc was replaced helps to substantiate this. If the premise that an increased oxygen supply stimulates the coeno- sarc to form hydranth tissue is granted, then the results obtained in these experiments can be explained very readily. The coenosarc must be metabolizing at a certain level before hydranth tissue will form. When a small opening is made in the perisarc only the tissues immedi- ately around the opening get enough O, to raise them to this level. Since the basal portions of the hydranth (including the proximal ten- 102 EDGAR ZWILLING tacles) form at the lower level of activity, and since this lower level is symmetrical around the opening, a small but complete hydranth forms. But, since the region exposed to the oxygen is at the side, and the high activity level is around this, the new hydranth's polarity is at right angles to the polarity of the stem. The other form and polarity variations may be explained in the same way. As more oxygen is admitted into the stem, more of the coenosarc is activated to form hydranth tissue and, depending on the resulting gradient of oxygen, the two hydranths pro- duced are more or less fused. If a small region of the stem opposite the opening is activated the basal regions as well as the distal will be fused. If a greater region of this part of the stem is activated the basal portions of the hydranth will be separate. And as this region increases in length more and more of the hydranths become separated until two complete forms are produced. The work on dominance does not throw any new light on the factors involved in this phenomenon, but simply illustrates the inhibition exerted by the ends upon a central region. SUMMARY 1. Experiments in which lateral regions of the coenosarc of Tubu- laria were exposed indicate that the rate of regeneration in such regions is proportional to the area exposed to the influence of freely circulating sea water. 2. The morphology and polarity of central regenerants in thicker stems varies with the area exposed. Forms ranging from single com- plete hydranths, whose polarity was at right angles to the original po- larity, to two hydranths, fused to a greater or lesser extent, were ob- tained. The polarity of the double hydranths varied with the degree of fusion. All angles to the original polarity were obtained. 3. Injury to the exposed coenosarc is not a factor in the present cases. 4. The ends of a stem exert a definite dominance over the exposed central regions. This dominance may be removed by ligaturing the ends. When this is done the rate of regeneration of the central region increases greatly. 5. A more distally placed central exposed region inhibits another exposed region placed more proximally on the same stem. 6. It is suggested that the probable stimulator of regeneration in Tubularia is the oxygen in the sea water. The results on rate of re- generation and on morphological and polarity variations may be ex- plained by the amount of oxygen available to the coenosarc when areas of different size are exposed. REMOVAL OF PERISARC AND REGENERATION 103 LITERATURE CITED EARTH, L. G., 1938a. Quantitative studies of the factors governing the rate of regeneration in Tubularia. Biol. Bull., 74: 155. BARTH, L. G., 19386. Oxygen as a controlling factor in the regeneration of Tubularia. Physiol. Zool, 11: 179. CHILD, C. M., 1927a, Modification of polarity and symmetry in Corymorpha palma by means of inhibiting conditions and differential exposure. Jour. Exper. Zool, 47 : 343. CHILD, C. M., 19276. Experimental localization of new axes in Corymorpha with- out obliteration of the original polarity. Biol. Bull., 53 : 469. CHILD, C. M., 1929. Physiological dominance and physiological isolation in de- velopment and reconstitution. Arch. f. Entw.-mcch., 117: 21. GOETSCH, W., 1929. Das Regenerationsmaterial und seine experimentelle Beeinflus- sung. Arch. f. EnHv.-mech., 117: 211. MORGAN, T. H., 1901. Further experiments on the regeneration of Tubularia. Arch. f. Entw.-mech., 13: 528. MORGAN, T. H., 1903. Some factors in the regeneration of Tubularia. Arch. f. Entw.-mcch., 16 : 125. PEEBLES. F., 1931. Some growth-regulating factors in Tubularia. Physiol. Zool., 4: 1. THE ROLE OF TEMPERATURE IN HYDRANTH FORMATION IN TUBULARIA JOHN A. MOORE (From the Department of Zoology, Columbia University, New York City and the Marine Biological Laboratory, Woods Hole) Tubularia crocea (Agassiz) is a hydroid found on wharfs and piles in the Woods Hole region. It is abundant until the last part of July when the hydranths are dropped and the coenosarc retreats into the perisarc. The colonies remain in this condition until October when growth begins anew. The experiments reported in this paper show that temperature is an important factor in regulating this periodicity. Autotomy and resorption of hydranths under a variety of conditions have been studied by a number of investigators, for example, Baschlin (1932), Cerfontaine (1903), Cast and Godlewski (1903), Godlewski (1904), Huxley and deBeer (1923), Loeb (1900), Morse (1909), Riddle (1911), and Thatcher (1903). Morse (1909) made a special study of the cause of autotomy of the hydranth in Tubularia and con- cluded that "... temperature seems to be the only consistent factor involved in the decapitation. When the temperature is kept at about 10° C. or 15° C, the hydranths are retained, regardless of any other factors with the one exception of lack of oxygen, which we believe to be inoperative except under wholly artificial conditions." Sumner, Os- burn and Cole (1913) recorded the disappearance of Tubularia crocea during the summer from the harbor at Woods Hole but noticed that in cooler waters the colonies may remain in an active condition throughout this period. Elmhirst (1922) kept colonies of Tubularia in his aqua- rium for three years and noticed that hydranths appeared in midwinter and were lost in midsummer. Light and oxygen were thought to be the principal factors involved. If high environmental temperatures cause Tubularia to exist in a dormant condition, then keeping them in colder water should result in renewed growth. It was also thought that oxygen might be a factor since its importance in regeneration has been demonstrated by Barth (1937, 1938) and Miller (1937). The following experiments were performed to test the action of temperature and oxygen on dormant colonies of Tubularia. 104 TEMPERATURE IN HYDRANTH FORMATION 105 EXPERIMENTS On August 6, 1938 an experiment was begun to test the effect of oxygen and temperature on dormant colonies. Two 500 cc. flasks were placed in an 18.4° constant temperature bath. Into one of these, A, air was bubbled continuously at the rate of 1 liter per minute. The other, B, received oxygen at the rate of 1 liter per 11-14 minutes. These gases kept the water saturated and in rapid motion. Two similar flasks were placed in running salt water which varied from 23.2° to 24° for the duration of the experiment. (Sea water in laboratory tanks is from 0° to 1° warmer than where Tubularia grows.) One of these, C, received air at the same rate as A. The other, D, received oxygen at the same rate as B. Dormant colonies were placed in these containers. After 72 hours there were many hydranths in B, less in A, and none in C and D. At 96 hours nearly every free end of the B colony had a fully formed hydranth. In A they were not so numerous. D showed but two small hydranths (occasionally a hydranth is noticed on dormant colonies as they are collected), and C showed none. On August 18, 1938, colonies were collected and placed in flasks as in the previous experiment. The temperature of laboratory ocean water had fallen to 22° C. The experiment was terminated August 24, TABLE I See text for explanation. Flask Temp. 02 Total hydranths Hydranths/gm. colony °C. cc.fl. A 18.4 5.5 243 128 B 18.4 23.5 116 83 C 22. 5.0 0 0 D 22. 21.3 31 38 r --»•«< the colonies removed from the flasks and the number of hydranths care- fully counted under the binocular microscope. The oxygen tension in each flask was determined by the Winkler Method. Excess water was removed from the colonies and their weight determined. The results are shown in Table I. DISCUSSION Morse (1909) found that Tubularia disappeared when the tempera- ture of ocean water rose to 20° C. We observed that in the summer of 1938 it disappeared when the water temperature was 21°. From Morse's experiments and those here reported it seems highly probable 106 JOHN A. MOORE that autotomy of hydranths and the existence of the colonies in a dormant condition is due to high temperature. Conversely, hydranths will be produced in abundance at a few degrees below 20°. North of Cape Cod where the water does not, rise above the critical temperature for Tubularia, colonies flourish throughout the summer months (Morse, 1909). This existence of coelenterates in a state of dormancy in regions where conditions during part of the year do not permit active growth has been pointed out by Broch (1925), who called attention to its similarity to spores in plants. It is doubtful if oxygen in the concentrations normally present in the sea plays a significant role in the disappearance or reappearance of active colonies of Tubularia. However, in flask D (Table I), kept at 22°, a number of hydranths were formed. The oxygen tension was far above that ever present in the ocean. On bright days the oxygen tension of water bathing the dormant colonies has measured 6.1 cc. per liter, higher than in either flasks A or C (Table I) yet there is no renewed growth as long as the temperature is over 20° C. In the experiment conducted at 24° saturation with oxygen failed to produce a significant number of hydranths. It appears that at temperatures just above the critical point a high oxygen tension may stimulate some hydranth formation but this effect is not noticed at still higher temperatures. The author wishes to thank the Department of Zoology of Columbia University for providing facilities for investigating this problem at Woods Hole and Professor L. G. Earth for his helpful suggestions. SUMMARY When ocean water rises above 20-21° C. in the Woods Hole region the hydranths of Tubularia crocea are lost and the coenosarc retreats into the perisarc. The colonies remain in this dormant condition until the temperature drops in autumn. Growth is then resumed. The dormant colonies can be stimulated to produce hydranths in three days by keeping them at 18.4° C. It is thought that temperature is the principal agent in the disappear- ance and reappearance of active colonies. LITERATURE CITED EARTH, L. G., 1937. Oxygen as a controlling factor in the regeneration of Tubu- laria. Biol. Bull, 73: 381. EARTH, L. G., 1938. Oxygen as a controlling factor in the regeneration of Tubu- laria. Physiol Zool, 11: 179-186. TEMPERATURE IN HYDRANTH FORMATION 107 BASCHLIN, K., 1932. Histologische Untersuchungen iiber Ruckbildungserscheinun- gen an Siisswasser- und Meereshydroiden. Zoo/. Jahrb., Abt. f. Allg. Zool. u. Physio!,, 52 : 223-294. BROCH, H., 1925. Biogeographie vom Tier aus. Naturwiss., 13 : 447-452. CERFONTAINE, P., 1903. Recherches experimentales sur la regeneration et 1'hetero- morphose chez Astroides calycularis et Pennaria Cavolinii. Arch, dc Biol., 19: 245-315. ELMHIRST, R., 1922. Cyclic conditions and rejuvenation in hydroids. Nature, 109 : 208-209. CAST, R., AND E. GODLEWSKI, 1903. Die Regulationserscheinungen bei Pennaria Cavolinii. Arch. f. Entiv.-mcch., 16: 76-116. GODLEWSKI, E., 1904. Zur Kenntnis der Regulationsvorgange bei Tubularia mes- embryanthemum. Arch. f. Entu>.-mech., 18: 111-160. HUXLEY, J. S., AND G. R. DEBEER, 1923. Studies in dedifferentiation. IV. Resorp- tion and differential inhibition in Obelia and Campanularia. Quart. Jour. Micr. Sci., 67 : 473-495. LOEB, J., 1900. On the transformation and regeneration of organs. Am. Jour. Physiol, 4 : 60-68. MILLER, J. A., 1937. Some effects of oxygen on polarity in Tubularia crocea. Biol. Bull, 73 : 369. MORSE, MAX, 1909. The autotomy of the hydranth of Tubularia. Biol. Bull., 16: 172-182. RIDDLE, O., 1911. On the cause of autotomy in Tubularia. Biol. Bull., 21: 389- 395. SUMNER, F. B., R. C. OSBURN AND L. J. COLE, 1913. A biological survey of the waters of Woods Hole and vicinity. Bull, Bur. Fish., 31 : (1911), Part 2: 566. THACHER, H. F., 1903. Absorption of the hydranth in hydroid polyps. Biol. Bull, 5: 297-303. STANDARDIZATION OF THE PRECIPITIN TECHNIQUE AND ITS APPLICATION TO STUDIES OF RELA- TIONSHIPS IN MAMMALS, BIRDS AND REPTILES HAROLD R. WOLFE (From the Department of Zoology, University of Wisconsin, Madison, Wisconsin) Serologic techniques are now used in studies of relationships among bacteria, plants (see review by Chester, 1937) and animals. The types of tests commonly employed are the complement fixation, the agglutina- tion and the precipitation reactions. The particular type of study gov- erns the test employed and there are advantages to each one. Although it is claimed that the precipitin reaction is not as sensitive as other serologic methods, its simplicity and accuracy make it advan- tageous in studies of animal relationships where large numbers of tests are necessary. The inconsistency of results, as shown in serologic literature, is possibly not due to the unreliability of the precipitin reac- tion but rather to the differing techniques of the workers who make the tests. While the techniques that were first employed in the precipitin reaction were important and gave interesting results, they have been much improved in recent years and need still further standardization. There should be a uniform method of measuring the end-point (titer), the chemical and physical factors should be constant and the methods of injection should be standardized. The precipitin reaction is the one that has been most commonly used in the study of animal kinship, though agglutinins, responsible for blood groups, have been employed in the case of apes and man. Compara- tively little research has been done on the serological relationships of reptiles or of birds. Nuttall (1904) stated that there was " a similarity of the blood constitution of all birds." He found differences in degree of reactions but these were so slight that a study like that made with mammalian bloods was not possible. Sasaki (1928) claims that he was able to distinguish the Japanese duck (Anas domestica erecta) from the Chinese goose (Anser cygnoides) with untreated sera and could distin- guish the Japanese duck from the Muscovy duck (Cairina moschata) by specialized sera. These results would be more reliable were the possible error of the tests determined. Graham-Smith (1904 — see Nuttall, Sec- tion VIII) found a closeness of relationship between the Crocodilia and 108 SEROLOGICAL RELATIONSHIP STUDIES 109 Chelonia and between the Ophidia and Lacertidia. The reactions indi- cated that the former groups stood more closely related to the Aves. Ehrhardt (1929) reviewed the whole field very completely. Some important research has been done recently showing the significance of the tests in determining uncertain relationships. Boyden and Noble (1933) applied the reaction to a study of relationship among several amphibians and clarified the status of these forms. Eisenbrandt's (1938) extension of these methods to chiefly parasitic helminthes is of value in showing that lower animals with only relatively small amounts of circulating proteins may also be successfully studied. Zuckerman and Sudermann (1935) reported on "Serum relation- ships within the family Cercopithecidae." As part of their conclusions they state, "... their findings support the view . . . that the serum- precipitin test is of limited value in tracing phylogenetic relationships." Their conclusions from their data are not justifiable, however, since these data were obtained by a technique which is questionable. Magnus (1908) noted that antisera specificity was lessened by con- tinued injections of an antigen. Wolfe (1935, 1936) and Wolfe and Baier (1938) emphasized the importance of the effects of injection meth- ods on the specificity and aspecificity of the serum precipitin tests. It was shown that the type of response is greatly influenced by the method of injection, the specificity being enhanced by limiting the amount of protein injected. With this technique they found greater distinctions among more closely related forms and less " extra-group " reactions. These results suggested that even greater specificity might be possible if the amount of protein injected was reduced to a minimum. It is known that exceedingly small quantities of protein may cause the formation of precipitins in rabbits. Hektoen and Cole (1932) were able to produce antisera of fairly high titers in rabbits by injecting as low as .00029 gram of an ovalbumin solution. The purpose of this research is: (1) to determine whether precipitin specificity can be increased by limiting the quantity of protein injected into rabbits to a minimum necessary for antibody production; (2) to discover whether a serological relationship study on birds and reptiles is practicable; (3) to attempt to further standardize precipitin technique. METHODS AND MATERIALS Healthy adult rabbits of mixed breeds and weighing approximately three kilograms were injected intravenously with serum proteins of ten species of mammals, five species of birds and one species of reptiles. Listed in Table I are the scientific and common names of the animals 110 HAROLD R. WOLFE TABLE I Antigens used in experiments Scientific Name Common Name Grams protein per 100 cc. serum Homo sapiens * Pans satyrus Macacus rhesus Macacus sinicus Papio hamadryas Lasiopyga mona Cebus capunicus Bos laurus * Bison bison * Capra hircus * Ovies aries Odocoileus virginianus Tamiasciurus hudsonicus * Sciurus carolinensis * Sciurus niger Citellus tridecemlineatus Tamias striatus Marmota monax Rattus norvegicus Felis domesticus Canis familiaris Callus domesticus * Columba livia * Strix varia Larus argentatus Ardea herodias Mergus merganser Anser anser Anas boschas Alligator mississippensis Phrynosoma cornutum Chelydra serpentina * Chrysemys picta Emys blandingii Clemys insculpta Thamnophis sertalis Man Chimpanzee Rhesus macaque Bonnet macaque Hamadryas baboon Mona guenon Capuchin monkey Ox American buffalo Domestic goat Domestic sheep White-tailed deer Red squirrel Gray squirrel Fox squirrel 13-lined ground squirrel Eastern chipmunk Ground hog White rat Cat Dog Chicken Pigeon Barred owl Herring gull Great blue heron American merganser Domestic goose Domestic duck Alligator Texas horned toad Snapping turtle Painted turtle Blanding turtle Wood turtle Common garter snake 6.843 6.677 5.425 6.926 5.530 8.616 7.198 6.825 7.982 6.593 6.051 5.843 6.409 8.750 6.560 5.096 5.139 6.260 7.052 6.825 6.364 4.676 3.133 4.070 3.070 3.550 3.800 4.387 4.676 5.158 4.389 3.654 3.446 2.299 3.550 5.405 Used for production of antisera and test antigens, others only for test antigens. from which the blood proteins were secured. The amount of protein per 100 cc. was determined on the whole serum by the Folin-Wright (1919) modified macro-Kjeldahl method. Corrections were not made for non-protein nitrogen. Table II records antisera production data for some of these species. The quantity of protein injected and titers are based on protein calculated from total nitrogen. SEROLOGICAL RELATIONSHIP STUDIES 111 The amount of protein injected per animal in the first series has been as low as .000786 gram. This is an exceedingly small amount and not always do such quantities injected into rabbits yield potent antisera. When a second series of injections was given the dosage was usually about one-half the initial series dosage. At present a standardized in- jection technique is being followed. Rabbits are injected with .001 gram TABLE II Antisera production data Rab- bit No. Serum protein injected Injected dil. of solution Total protein injected Bled — days after last injection Titer (dil.) cc. grams 205 Red squirrel 6.0 (2%) .00768 12 512,000 215 1.2 (2%) .001536 12 512,000 136 3.2 (1 1) .102544 14 51,200* 217 Gray squirrel 1.0 (1 1000) .001 10 0 1.0 (1 1000) .001 12 16,000 0.75 (1 1000) .0075 8 512,000 223 2.0 (1 1000) .002 10 128,000 125 2.25 (undil.) .19687 10 51,200* All Ox 1.5 (1 500) .003 2 0 4 0 6 16,000 9 1,024,000 A 9 Buffalo 1 mg./kil. .0034 10 256,000 A10 Goat 1 mg./kil. .0034 10 128,000 185 Human 5.75 (2%) .00786 12 51,200* 77 3.0 (undil.) .20531 12 102,400* 209 0.6 (2%) .00082 13 128,000 221 2.0 (1 : 1000) .002 10 512,000 225 Chicken 2.0 (1 : 1000) .002 12 0 0.6 (1 : 500) .0012 7 512,000 228 0.75 (1 : 1000) .0015 8 1,024,000 229 Pigeon 1.0 (1 : 500) .002 10 512,000 222 Snapping turtle 1.0 (1 : 500) .002 10 0 0.6 (1 : 500) .0012 7 256,000 * This notation indicates a serial dilution from a 2 per cent standard — all the rest are from standard solutions containing .001 or .002 gram per cc. of protein per kilogram body weight distributed over three injections. The injections are made on alternate days. When additional series are necessary the amount administered is one-half that of the initial series given in two injections on alternate days. Ring tests were used exclusively in determination of the titer. The test consisted of having a duplicate series of serological tubes containing 112 HAROLD R. WOLFE 0.5 cc. of antigen in serial dilution with 0.1 cc. of antiserum layered be- neath the antigen. The titer was determined after incubation in a water bath at a temperature of 37.5° ± 1° C. The possible error in reading the reactions was ± 1 tube. Several experiments consisted of readings at intervals during the sixty minutes in order to study the progression of the titers. Unless definitely specified, the antisera were used undiluted or 1.0 cc. was diluted with 0.25 cc. of a buffered physiological saline solution. This small dilution did not have any effect on the strength of titers in the experiments. In the earlier experiments the serial dilutions were made from stand- ard test antigens which were 2 per cent solutions. The latter were made by diluting 1.0 cc. of the undiluted blood writh 49.0 cc. of buffered saline giving a 1 : 50 dilution. Since the protein content of the antigens varied TABLE III Anti-human sera Antigen serum proteins of 77 185* 209 221 Man . . 102,400 51,200 128 000 512 000 Chimpanzee 25,600 12,800 128,000 256,000 Rhesus macaque 12,800 1,200 0 64,000 Bonnet macaque 6,400 1,600 0 64,000 Baboon 25,600 400 0 256,000 Mona guenon 12,800 400 128,000 Capuchin monkey 16,000 Rodentia 0 0 Artiodactyla 0 0 0 4,000 Carnivora 0 0 1,000 * Titers from a 2 per cent solution of undiluted antigen. from 4 per cent to 9 per cent the titers need adjustment. In the later experiments all serial dilutions are made from standard solutions that contain either .002 or .001 gram protein per cc. thus making all titers comparable. EXPERIMENTAL DATA Effects of Amount of Antigen Injected on the Specificity of the Antisera Produced One of the main purposes of this paper is to show the variance in the degree of the homologous and heterologous precipitin reactions when one factor, the injection procedure, is varied. Antisera 77, 185, 209, 221 are anti-human sera whose reactions are recorded in Table III. Serum No. 77 was produced by injection of .21594 gram protein while the rabbit producing No. 209 received only .00082 gram of protein. (Reference SEROLOGICAL RELATIONSHIP STUDIES 113 should be made to Table II for exact detail on quantity of material in- jected in the experiments to be described.) Of the four antisera 209 is by far the most specific, giving reactions only with chimpanzee and human blood antigens. No. 185 is more specific than 221 even though the latter received one-third the amount of protein given the former. Injections of the more minute quantities (similar to 185 and 209) usually gave reactions illustrated by these two and more rarely reactions like 221. On the other hand, when 3 cc. of undiluted serum was in- jected, the most usual response was like that given by No. 77. Sera 77 and 221 yielded primate group reactions but No. 185 re- sulted in a sub-grouping showing that the blood proteins of man and chimpanzee are more nearly related to each other than to the blood of the monkeys. Serum 209 was an extremely specific one, giving reac- tions with only chimpanzee arid the homologous antigens. Such antisera would be of great value to the medico-legal workers, as well as to the student dealing with man's closest relatives. The reactions of three anti-red squirrel and three anti-gray squirrel sera are shown in Table IV. Serum 136 gave a Sciuridae group reac- TABLE IV Anti-squirrel sera Antigen serum proteins of Anti-red squirrel Anti-gray squirrel 136* 205 215 125* 217 223 Red squirrel 51,200 25,600 51,200 12,800 51,200 51,200 0 400 400 512,000 0 0 0 0 0 0 512,000 128,000 128,000 0 0 0 0 2,400 51,200 51,200 2,400 6,400 3,200 0 200 400 64,000 512,000 256,000 4,000 8,000 0 4,000 4,000 0 128,000 2,000 0 0 0 0 0 0 Gray squirrel Fox squirrel 13-lined gray squirrel. . . . Chipmunk Ground hog Rat Artiodactyla Primata * Titers from a 2 per cent solution of undiluted antigen. tion with no sub-grouping evident, but No. 125, an anti-gray squirrel serum, reacted much more weakly with the blood of red squirrel and other sciurids but could not be differentiated from fox squirrel blood. Both these sera were produced by injection of a much larger quantity of protein than were the other four listed in the table. The latter all show that a definite difference exists between red squirrel blood and that of the fox or gray squirrel bloods. Sera 215 and 217 subdivide the Sciuridae reactions into three sub-groups. Serum 205 is very specific and gave no cross reactions at all and No. 223 allowed for distinction between the 114 HAROLD R. WOLFE TABLE V Anti-aves sera * Antigen serum protein of Anti-chicken Anti-pigeon 225 228 229 Chicken 512,000 4,000 16,000 8,000 0 8,000 8,000 0 2,000 0 0 0 0 1,024,000 256,000 256,000 256,000 0 0 0 0 0 0 512,000 0 0 0 0 0 0 0 0 Pigeon Goose Duck Merganser Great blue heron Barred owl Herring gull Snapping turtle Painted turtle Alligator Common garter snake Cryst. egg albumen of chicken * Fowl and reptile data are from a thesis by Miss Bernice Cohen done under the supervision of the author and used with her permission. two closely related forms of gray and fox squirrels. This latter result is very rarely secured by the precipitin test with native antisera. The variation in results that were secured with different dosages do not subtract from the value of the precipitin test but, on the other hand, give it greater possibilities. The differences in reactions seem to imply that a distinction can be made among genera within a family and possibly TABLE VI Anti- snapping turtle Antigen-serum Protein of No. 222 Snapping turtle 256,000 Painted turtle 64,000 Blanding turtle 16,000 Wood turtle 16,000 Alligator 0 Horned toad 0 Common garter snake 0 Aves (7 species) 0 Cryst. egg albumen of chicken 0 species within a genus as well as among larger groups. Similar results have been reported previously but there now seems to be a method whereby these results can be secured consistently. It would be ad- visable to have many antisera for each group studied in order to insure significant results. It might be stated that it is believed that the variability of results is not attributable to the innate characteristics of the test but is probably SEROLOGICAL RELATIONSHIP STUDIES TABLE VII Effects of Length of Incubation Period and of Dilution of Antisera 115 Serum Undiluted Dilution 1 : 3 1 : 5 Incubation Period — Minutes 1 5 10 30 60 60 60 And- buffalo— A 9 Buffalo 0 Ox 0 Sheep 0 Goat 0 Deer 0 64,000 64,000 0 0 0 128,000 128,000 0 0 0 128,000 128,000 4,000* 4,000 32,000 256,000 256,000 64,000 128,000 128,000 512,000 512,000 4,000 4,000 4,000 128,000 128,000 1,000* 2,000* Anti-goat — A 10 Goat 0 Sheep 0 Buffalo 0 0 0 0 0 0 0 64,000 64,000 0 128,000 64,000 0 0 1,000 Ox 0 0 0 0 0 Deer 0 Anti-beef — All Beef 64,000 Buffalo 64,000 Sheep 0 Goat Deer 0 128,000 128,000 0 0 512,000 512,000 0 0 1,024,000 1,024,000 16,000 64,000 1,024,000 1,024,000 128,000 128,000 256,000 512,000 256,000 0 0 0 512,000 256,000 * Readings not clear. due to the antibody producer, in this case, the rabbit. It would be an ideal condition were one able to secure a group of rabbits with similar potentialities for antibody production. Tests ivith Ainan and Reptilian Blood Sera There is a meagerness of serological data with respect to reptilian and avain bloods probably due to the conception that the animals within each of these groups are very closely related to each other and are not suitable for study by means of the precipitin reaction. The data listed in Tables V and VI contradict this contention. There are indications that the reactions of anti-avian and anti-reptilian sera produced by using very small amounts of proteins may be useful in confirming present morphological data and thus aiding in the clarification of disputed relationships. 116 HAROLD R. WOLFE Effects of Incubation Period and Dilution of Antisera on Titers Previous workers have considered such factors as the length of in- cubation period and the effect of dilution of the antisera on the homolo- gous and heterologous reactions. The usual method is to make readings after 1 hour incubation at a temperature of 37° ± 1° C. when the ring test is used and the antisera are used undiluted or diluted to various degrees. Table VII shows the reactions of three anti-Artiodactyla sera which illustrate both time and dilution effects. These antisera readings at 5, 10, and 30 minutes usually showed much greater specificity than did the readings at 60 minutes. Serum A 9 lost its distinguishing charac- teristic at 60 minutes so that the reactions of buffalo and ox could not be distinguished from deer and goat. It should be noted that antisera A 10 and All show that beef or buffalo can be markedly differentiated from either goat or sheep bloods even at 1-hour readings. Heretofore these great differences have not been shown (with rare exceptions) when undiluted and untreated antisera against the serum proteins were used. By using very small quantities of blood for injections, results similar to, or showing greater differentiation, are consistently secured. The dilution of antisera with physiological saline is first noted to have an effect on the antigens more distantly related (according to taxonomical schemes). These antisera became much more specific by diluting them as low as 1:3 (2 parts saline and 1 part antiserum). With beef antiserum All the reactions of sheep, goat and deer anti- gens disappeared at antiserum dilutions of 1 : 3 while the buffalo antigen titer remained similar to the homologous reaction. Such low dilutions show these effects only when the antisera have been produced by the methods described; that is, injections of minute quantities of protein's. DISCUSSION Data are presented that show how a greater uniformity of results may be obtained with the serum precipitin reaction. These data further supplement previous work of the author in regard to the specificity and aspecificity of the test. The consistency of the specificity of the reaction with the whole serum proteins of mammalian, avian and reptilian bloods is greater than heretofore reported with untreated rabbit antisera. A distinction of the bloods of such closely, related forms as ox and sheep, red squirrel and gray squirrel, several species of turtles and of a number of birds was possible. SEROLOGICAL RELATIONSHIP STUDIES 117 The cause of this greater specificity is believed to be the small quan- tity of protein injected. A standardized quantity is now used; the amount injected in the first series is .001 gram of protein (based on total nitrogen) per kilogram body weight and distributed in three injec- tions. If the titer of the antiserum resulting is insufficient the animal is reinjected with .0005 gram per kilogram body weight given in two injections. This amount injected is thought to be approximately the smallest amount of the antigen used that will regularly cause antibody formation. The results are not always similar and therefore it seems necessary to produce a number of antisera against a particular antigen in studying relationships and choose those which have the desired specificity. With the present procedures employed the antisera against a particular kind of antigen fall into groups. The reactions of the different groups differ quantitatively but not necessarily qualitatively. Not only are relation- ships shown but the degree of such relationships can be studied with greater accuracy than before. Morphology does not allow for such a type of approach to the classification of animals. These differences can best be understood by referring to the tables on rodent or primate reactions. The high specificity secured with a few of the antisera even when used untreated should interest the worker concerned with the medico- legal aspects. Antisera that give reactions only with very closely related bloods should be of much value and no doubt the evidence so secured would be readily accepted by our courts. Such antisera are compara- tively easy to produce if the amount of protein injected is reduced to the approximate minimum. The titer of the reaction at 1 hour incubation did not show the marked differences that were shown after incubating for shorter pe- riods. This factor, the length of incubation period, has not been used with any uniformity for distinguishing closely related proteins. When antisera are produced by the injection of very minute quantities of pro- tein the homologous blood and very closely related ones give ring tests with high titers after 1, 5, 10 or 20 minutes incubation and after that period the blood of more distantly related forms becomes positive. These reaction's, when properly performed, are usually very distinct and consistent. Attempts were made to use readings after incubation of more than 1 hour. These readings are not so clear-cut and not consistent and so are considered unreliable. The second in vitro factor which enhanced the specificity of the re- action, that is, dilution of the antisera, showed a direct correlation with 118 HAROLD R. WOLFE the results secured by studying the period of incubation. Boor and Hektoen (1930) and Hektoeri and Boor (1931), using antihemoglobin sera, found that dilutions with 3 parts of saline or normal rabbit serum eliminated extra specific reactions. Satoh (1933) demonstrated that dilution of 1 : 10 or less affects the titer of an antiserum when it is pro- duced by few injections with a small quantity of material, and dilution of 1 : 25 or more affects the reactions when the antisera result from injection of larger quantities given in a greater number of injections. In Satoh's report the dilution seemed to have a similar effect upon the heterologous and homologous reactions. Wolfe (1933) reported that dilutions of 1 : 2 make the readings of the " ring " clearer but that dilu- tions of 1 : 10 do not always eliminate extra-group reactions. These antisera were produced by injection of a relatively large quantity of antigen. The antisera that were used in the present work were easily affected by small dilutions with buffered saline. The lowest dilution to have an effect was a 1 to 1 (1 part saline) dilution but with some antisera it was necessary to dilute with 2 or 3 parts of saline in order to show decrease in reaction. The dilution always lowered or eliminated the heterologous titer first and one could thus verify the dissimilarities between such closely related bloods as ox and sheep, or man and monkey. But even a higher dilution could not distinguish consistently between ox and buffalo or man and chimpanzee for the dilutions had similar effects upon both titers. The data presented in this paper are fairly consistent. The methods reported are easily followed and it seems they insure a uniformity of results. The different approaches used resulted in a partial solution of the primary problem concerned ; namely, the degree of similarity of the blood proteins of some of the more closely related mammals, birds and reptiles. It seems very possible that a reliable serological relationship study of birds and of reptiles can be accomplished. Additional data may aid in clarifying the " taxonomic confusion " of birds. Finally — a plea is made to the workers using the serologic method, especially the precipitin technique. It is desirable to obtain uniform and reliable results and one should use the methods that are best adapted to the particular problem. The earlier workers' contribution was and is invaluable but not necessarily the only one that is of use. A modifica- tion of the earlier methods is in order and only when one uses the best possible method for the particular study will the results be consistent and therefore biologically significant. SERO LOGICAL RELATIONSHIP STUDIES 119 SUMMARY 1. Antibodies were produced in rabbits by injecting very small quan- tities of serum proteins of mammals, birds and reptiles. 2. Antisera produced by injection of a minimum quantity of antigen were the most specific as determined by the " ring " method. 3. Occasionally the serum proteins of such closely related animals as ox and sheep or gray squirrel and red squirrel could be distinguished. 4. Antiserum dilutions as low as 1 : 3 at times eliminated the heter- ologous reaction. 5. The " ring " appeared in the homologous and very closely related protein solutions earlier than in the more distantly related ones when incubated at temperatures of 37° ± 1° C. 6. It was not possible to distinguish between the buffalo and ox serum proteins or between the goat and sheep serum proteins. 7. Serological relationship studies of birds and reptiles seem feasible. 8. The procedures have not been entirely new but the necessity for uniformity and standardization' of methods has been emphasized. LITERATURE CITED BOOR, ALDEN K., AND LUDVIG HEKTOEN, 1930. Preparation and antigenic proper- ties of carbonmonoxide hemoglobin. Jour. Inf. Dis., 46 : 1. BOYDEN, ALAN, AND G. K. NOBLE, 1933. The relationships of some common Amphibia as determined by serological study. Am. Mus. Novitates, No. 606: 1. CHESTER, K. STARR, 1937. A critique of plant serology. Quart. Rev. BioL, 12: 19-46; 165-190; 294-321. EHRHARDT, ALBERT, 1929. Die Verwandtschaftbestimmungen mittels der Im- munitatsreaktionen in der Zoologie und ihr Wert fur phylogenetische Untersuchungen. Ergcbn. it. Fortschr. der Zool., 7 : 279. EISENBRANDT, LESLIE L., 1938. On the serological relationship of some Helminths. Am. Jour. Hyg., 27: 117. FOLIN, O., AND L. E. WRIGHT, 1919. A simplified macro-Kjeldahl method for urine. Jour. Biol. Chcm., 38: 461. HEKTOEN, LUDVIG, AND ALDEN K. BOOR, 1931. The specificness of hemoglobin precipitins. Jour. Inf. Dis., 49 : 29. HEKTOEN, LUDVIG, AND ARTHUR G. COLE, 1932. Precipitinogenic action of minute quantities of ovalbumin. Jour. Inf. Dis., 50: 171. MAGNUS, WERNER, 1908. Weitere Ergebnisse der Serum-Diagnostik fur die theoretische und angerwandte Botanik. Ber. dcut. Bot. Gesell., 1908, 26a: 532. NUTTALL, GEORGE H. F., 1904. Blood Immunity and Blood Relationship. Cam- bridge University Press. SASAKI, KIYOTSUNA, 1928. Serological examination of the blood relationships between wild and domestic ducks. Jour. Dept. Agr. Kyushu Imperial Univ., 2: 117. SATOH, TAKEO, 1933. Ueber Prazipitintiter und " Prazipitingehalt." Zcitschr. f. Immunitatsforsch. u. exp. Therapie, 79: 117. WOLFE, HAROLD R., 1933. Factors which may modify precipitin tests in their ap- plication to Zoology and Medicine. Physiol. Zool., 6 : 55. 120 HAROLD R. WOLFE WOLFE, HAROLD R., 1935. The effect of injection methods on the species speci- ficity of serum precipitins. Jour, of Immun., 29: 1. WOLFE, HAROLD R., 1936. The specificity of precipitins for serum. Ibid., 31 : 103. WOLFE, HAROLD R., AND JOSEPH G. BAIER, JR., 1938. Comparison of quantitative precipitin techniques as influenced by injection procedures. Physiol. Zool, 11: 63. ZUCKERMAN, S., AND ANN E. SuoERMANN, 1935. Serum relationships within the family Cercopithecidae. Jour. Exp. Biol., 12 : 222. EXPERIMENTS ON LIGIA IN BERMUDA VI. REACTIONS TO COMMON CATIONS T. CUNLIFFE BARNES (From the Osboni Zoological Laboratory, Yale University, and the Bermuda- Biological Station) Five preceding papers of this series (Barnes, 1932, 1934, 1935, 1936, 1938) have described the effect of salts and other factors on the littoral isopod Ligia baudiniana. The present report deals with the negative reaction of Ligia to filter paper moistened with salt solutions and the survival in sea water containing a higher percentage of K, Ca, or Na. Reaction to Filter Paper Moistened with Salt Solutions Large filter papers (diameter 25 cm.) were cut in two, each half saturated with a different solution and placed in a covered flat dish in a photographic dark room having a light of 61.9 foot candles directly above the center of the dish. Five isopods (freshly collected) were re- leased in the dish and their distribution on the two halves was observed at five to fifteen-minute intervals for a period of one to two hours. The dish was rotated 90° after each observation to eliminate any unsuspected source of orientation. With distilled water on both sides the distribu- tion was approximately equal. It is known (Barnes, 1938) that the isopods tend to collect on the distilled water side when the other half of the paper is moistened with sea water and the present experiments were designed to test the salts separately. The aversion for sea water is shown by specimens previ- ously kept in air on seaweed moistened with sea water (Barnes, 1938) and by specimens washed rapidly in distilled water (Barnes, 1935), but it was found that prolonged immersion in distilled water destroys this reaction. Thus in the present experiments of 168 specimen's previously immersed separately in 100 cc. of distilled water for an average of half an hour, 82 collected on the sea water side and 86 on the distilled water paper. For % M NaCl vs. distilled water, the ratio is the same as for sea water vs. distilled water, and in both cases the negative reaction no longer occurs at 50 per cent dilution (Table I). The aversion for KC1 paper 121 122 T. CUNLIFFE BARNES was still evident in a dilution of 10 per cent (see Table I). The isopods showed a positive reaction to CaCL paper compared with distilled water paper and an " all-or-none " reaction to CaCl2 occurred when the other half contained NaCl. An even distribution occurred between CaCU and LiCl paper and the negative reaction to LiCl compared with distilled water was less than that to the other salts. Survival in Modified Sea Water The toxic action of Na, Ca and K was studied by increasing the proportion of each salt in sea water. Specimens were tested separately in 100 cc. of mixtures of sea water and isotonic salt solution (Table II). TABLE I Reaction of Ligia to filter paper saturated with salt solutions. (The animals were tested in groups of five.) Treatment of each half of paper Total number of isopods on each half Ratio 5/8 M NaCl vs. distilled water 71 85 73 83 1 37 26 31 56 61 96 105 22 49 127 120 77 72 63 147 82 81 82 99 104 55 22 63 1 1 1 1 4 1 4 4 4 4 4 t 1 1 1.78 1.41 1.05 0.86 63.00 3.97 3.15 2.61 1.46 1.62 1.08 0.52 1.00 1.28 75 per cent 5/8 M NaCl vs. distilled water 50 per cent 5/8 M NaCl vs distilled water 25 per cent 5/8 M NaCl vs distilled water 5/8 M NaCl vs 3'5 M CaCl2 8 100 per cent 5/8 M KC1 vs distilled water 75 per cent 5/8 M KC1 vs. distilled water 50 per cent 5/8 M KC1 vs distilled water 25 per cent 5/8 M KC1 vs distilled water 10 per cent 5/8 M KC1 vs distilled water 5 per cent 5/8 M KC1 vs. distilled water — — M CaClj vs. distilled water 8 — M CaCl2 vs. 5/8"M LiCl 8 5/8 M LiCl vs. distilled water The isopods showed a high tolerance for sea water containing excessive sodium or calcium, but increasing the KC1 content is rapidly fatal. The toxicity of Na is a smooth function of the concentration. The Ca curve is erratic and the K curve falls abruptly for sea water mixtures containing over %6 M KC1 (see Fig. 2). Survival in Oxygenated Solutions The wide variation in survival times in a given salt solution suggests REACTIONS OF LIGIA TO COMMON CATIONS 123 1:4.0 - FIG. 1. Negative reaction to filter paper containing salt solutions. Ordinates: ratio of the number of isopods on the salt solution side to the number on the dis- tilled water side. Abscissae: percentage concentration in cc. of 5/8 M salt. Solid circles: KC1. Open circles: NaCl (see Table I). that other factors besides the direct chemical effect of the ions may be involved. To eliminate effects of asphyxiation oxygen was bubbled TABLE II Survival of Ligia in modified sea water Parts of salt solution in 100 cc. of modified sea water Average survival Maximum survival Number of specimens 9 per cent 5/8 M NaCl . . hours 69.8±15.2 hours 136 9 50 per cent 5/8 M NaCl . 59.4 ±4.5 168 39 71.4 per cent 5/8 M NaCl 30.9 ±3.5 79 20 80 per cent 5/8 M NaCl . ... 20.7 ±0.8 24 37 90 per cent 5/8 M NaCl . ... 8.0 ±0.03 11 43 10 per cent 5/8 M KC1 . . 44.7 ±5.6 112 19 11 percent 5/8 M KC1. . 36.5±5.5 48 8 20 percent 5/8 M KC1. . . 10.2 ±0.7 17 19 30 per cent 5/8 M KC1 . . 3.7±0.4 6 8 40 per cent 5/8 M KC1 s'±o 5 5 50 per cent 5/8 M KC1 0.7 ±.08 1 10 3 5 30 per cent — M CaCl2 68.5 ±10.8 264 25 8 50 per cent — M CaCl2 34.2 ±5.7 242 24 8 70 per cent — M CaCl2 39.1 ±6.7 254 19 8 90 per cent — M CaCl2 38.8±2.9 62 17 8 124 T. CUNLIFFE BARNES 20 40 60 80 100 FIG. 2. Survival times of isopods in sea water in which the concentration of one ion is increased. Ordinates : average survival in hours. Abscissae : con- centration of 5/8 M KC1 and NaCl and of 3.5/8 M CaCL in cc. per 100 cc. of modified sea water. Solid circles : KC1 ; open circles : NaCl ; triangles : CaCL. through the 100 cc. of salt solution in several cases, but did not affect the survival times in distilled water, % M NaCl or % M KC1 (Table III). TABLE III Survival in oxygenated solutions Medium Average survival Maximum survival Number of specimens Distilled water hours 40±0.2 hours 6 8 5/8 M NaCl . . 8.7±0.6 11 4 5/8 M KC1 1.6±0.1 3 7 Discussion It is of interest to note that the aversion for sea water is similar to aversion for NaCl and in both cases the negative reaction ceases at half dilution. The quantity of KC1 in sea water is below the threshold for the negative reaction to this salt, while CaCL has no repellent action when compared with distilled water. It was suggested (Barnes, 1938) that the salts on the paper stimu- late the isopods to greater movement which compels them to collect on the salt-free side. The pronounced aversion for KC1 and positive reac- REACTIONS OF LIGIA TO COMMON CATIONS 125 tion to Ca support this hypothesis. The rather precise relation between the concentration of NaCl and KC1 and the magnitude of the response may also be cited. It is unfortunate that our quantitative knowledge of chemical stimulation is very meager compared with the extensive studies of stimulation by light and gravity. The ratio of effect of K vs. distilled water to Na vs. distilled water is 2.2, somewhat greater than the ratio of ion'ic mobilities, i.e., 1.5. Hopkins (1932) has shown the importance of ionic mobility in stimulation by salts. The reaction of the animals when each side of the substratum is treated with a different salt is not an additive effect of the reactions to each salt compared with distilled water paper. Thus the ratio for CaCL vs. distilled water is 1 : 0.52 and NaCl vs. water is 1 : 1.78 but the ratio for CaCL vs. NaCl is as high as 63.0 : 1. This striking difference probably results from the combined action of the inhibiting Ca and stimulating Na on the leg movements. The differences cannot be explained by toxicity alone. Thus Li, the most toxic of the ions, gives a 1 : 1 ratio with Ca. In the case of sea water vs. distilled water the salt requirements of the animals must also be considered. Thus if depleted of salt by a half hour's immersion in distilled water the negative reaction to sea water disappears. The role of the flushing mechanism for the gills by which water rises between the last pair of legs is not known. Several animals were observed with the legs in position for the capillary conduit, but there was not sufficient solution on the filter paper to flush the gills. The specific effect of Na, Ca, and K is indicated in the toxicity curves for sea water containing increasing concentrations of each ion. The smooth relation between concentration and toxicity for Na suggests a gradual modification of a normal process in the animal. The very steep KC1 curve indicates severe toxic action and the lack of correlation with concentration in the case of calcium is characteristic of a surface re- action. It was shown (Barnes, 1938) that the protective action of Ca in hypotonic sea water also resembles a surface reaction. The experiments in which oxygen was introduced into the solutions help to show that the reported survivals of immersed animals are true salt effects. This is especially important in the case of K which para- lyzes all gill movement. It is probable that oxygen would increase the survival in less toxic solutions in which the animals live for several days. Summary 1. The negative reaction of Ligia to filter paper moistened with % M NaCl is similar to the negative reaction to sea water. In both cases dilu- tion by 50 per cent destroys the effect. 126 T. CUNLIFFE BARNES 2. The aversion for filter paper containing KC1 and NaCl is approxi- mately a rectilinear function of the concentration. 3. The results can be explained by the assumption that the ions stimulate the legs and the greater activity results in a distribution on the salt-free side of the substratum. 4. When presented with a " choice " between filter paper containing CaCl2 and NaCl, there is an " all-or-none " aversion for the NaCl. 5. The survival of Ligia in sea water containing increasing con- centrations of Na, K and Ca is described. 6. The survival in distilled water or in solutions of single salts is not affected by oxygenation. The writer is indebted to Dr. J. F. G. Wheeler for many courtesies. LITERATURE CITED BARNES, T. C, 1932. Salt requirements and space orientation of the littoral isopod Ligia in Bermuda. Blol. Bull., 63 : 496. BARNPS, T. C., 1934. Further observations on the salt requirements of Ligia in Bermuda. Biol. Bull, 66: 124. BARNES, T. C., 1935. Salt requirements and orientation of Ligia in Bermuda III. Biol. Bull, 69: 259. BARNES, T. C., 1936. Experiments on Ligia in Bermuda. IV. The effects of heavy water and temperature. Biol. Bull., 70 : 109. BARNES, T. C., 1938. Experiments on Ligia in Bermuda. V. Further effects of salts and of heavy sea water. Biol. Bull., 74 : 108. HOPKINS, A. E., 1932. Chemical stimulation by salts in the oyster, Ostrea vir- ginica. Jour. Exper. Zool., 61 : 13. Vol. LXXVI, No. 2 April, 1939 THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY SOME POLYCLADS OF THE NEW ENGLAND COAST, ESPECIALLY OF THE WOODS HOLE REGION LIBBIE H. HYMAN (From the American Museum of Natural History, Neiv York City} This report is based on polyclads collected at Woods Hole, Massa- chusetts, Mt. Desert Island, Maine, Connecticut and Long Island, New York. I believe it includes practically all of the species expected to occur in the neighborhood of Woods Hole. The Woods Hole speci- mens were personally collected in the summer of 1935 or were loaned for identification by Mr. George Gray, Curator of the Museum of the Marine Biological Laboratory. The Peabody Museum of Yale Uni- versity kindly loaned some specimens collected by Verrill along Massa- chusetts and Connecticut shores. The specimens from Mt. Desert Is- land, Maine, were personally collected in the summer of 1937. The specimens from Connecticut and Long Island were presented by Dr. Herman Spieth and Mr. M. D. Burkenroad. Ten species are represented in the material; none of these is new but most of them are interesting as old species badly in need of rework- ing. Each species has been carefully studied whole and in serial section to determine its status according to modern taxonomic standards. It is gratifying to me to be able to clear up the taxonomic position of some of the old species of Girard and Verrill. The scheme of classification here adopted is taken from Bock's in- valuable treatise (1913) which is fundamental to all modern work on the group. Before proceeding to the taxonomic descriptions, it seems desirable to give a brief account of the reproductive system of polyclads, since this system is of paramount importance in many genera. Polyclads are hermaphroditic with numerous testicular and ovarian follicles scattered throughout the parenchyma; there are no yolk glands. Fine vasa ef- ferentia connect the testes to the paired vasa deferentia, expanded, thin- walled, usually sinuous canals, packed with sperm. These separately or after a union enter the male copulatory apparatus, which typically consists of seminal vesicle, prostate or granule vesicle, and penis. The seminal vesicle is a rounded, oval or curved sac with a thick muscular 127 128 LIBBIE H. HYMAN wall. In some cases (Stylochus in part, some species of Notoplana) the terminal parts of the vasa deferentia acquire muscular walls and entirely resemble the seminal vesicle, so that a tripartite seminal vesicle results (Fig. 1). Certain glands known as granule (German workers) or prostate (English workers) glands are constantly associated with the male apparatus of polyclads. They may open directly into the ter- minal part of the male canal but more often are associated into a definite rounded or oval body with a muscular wall, termed the granule or pros- tate vesicle. The granule glands are often included inside this vesicle or may open into it from the exterior (extracapsular glands of Bock) or both. The granule vesicle may form a direct part of the male canal so that the sperm must pass through its lumen as in the Leptoplanidae (Fig. 5) ; this condition of the vesicle will here be termed "interpolated " (" eingeschaltet " in German). In some acotylean families and in the Cotylea in general the granule vesicle is " free," that is, connects by a duct to the male canal (Fig. 1). The function of the granule glands is unknown but the secretion is ejaculated with the sperm and must be of importance in their physiology. After passing through or receiving separately the granule vesicle, the male duct enters the penis proper, typically having the form of a projecting conical muscular papilla. It is frequently provided with a long or short hard stylet, sometimes curved or coiled, and may then itself be much reduced. The penis papilla may project directly into the male atrium or antrum or it may be at the inner end of a long canal leading from the atrium; this canal is termed by Bock the penis pocket (Fig. 5). When a penis pocket is present there may be at the place where the pocket widens into the male atrium, a penis-like projection, the penis sheath. In the family Planoceridae there is no proper penis but instead a cirrus, armed with teeth or spines. Penis here is limited to a male copulatory organ which is simply pro- truded by muscular action ; cirrus means a copulatory organ which is everted when used so that the lining is turned to the outside. The ripe eggs commonly collect in paired thin-walled expanded canals, the uteri, which in the family Leptoplanidae unite in front of the pharynx. The uteri, separately or after a union, open into the common unpaired female canal which then proceeds to the genital pore. This canal from the genital pore to the entrance of the uteri is commonly muscular and receives innumerable cement glands ; Bock terms it vagina since it must serve for copulation in most polyclads. It is also the passage by which the eggs reach the exterior. Near the genital pore it may be widened into a cement pouch into which most or all of the cement glands open. The cement glands secrete the material by which the eggs are fastened into ribbons or masses and stuck to objects. In many poly- SOME POLYCLADS OF THE WOODS HOLE REGION 129 clads the female canal continues inward proximal to the entrance of the uteri and terminates in a small or large spherical, oval, or tubular sac, now known as Lang's vesicle (Fig. 5). This appears to be a seminal receptacle since sperm are often seen in it. The male apparatus is always anterior to the female apparatus. The genital pores are commonly separate but may be united. Paired or multiple male apparatuses are not uncommon. SUBORDER ACOTYLEA LANG 1884 Definition. — Polyclads without a sucker behind the female genital pore ; pharynx ruffled ; copulatory apparatuses behind the middle of the body; tentacles when present of the nuchal type; uteri anterior to the female genital pore. SECTION CRASPEDOMMATA BOCK 1913 Definition. — Acotylea with marginal eyes, in addition to the usual cerebral and tentacular eyes. FAMILY STYLOCHIDAE STIMPSON 1857 (EMEND. BOCK, 1913) Definition. — Craspedommata with a band of eyes along the whole or part of the body margin ; female or both genital pores very near the posterior end; granule vesicle free; tentacles generally present; body usually oval, thick, firm and opaque. GENUS STYLOCHUS EHRENBERG 1831 (EMEND. LANG 1884) Definition. — Stylochidae wTith male and female pores close together in the posterior fourth of the body ; tentacles well-developed ; pharynx large with well-marked folds; prostate vesicle large, chambered, with extracapsular gland cells ; seminal vesicle single or tripartite ; Lang's vesicle absent. STYLOCHUS ZEBRA (VERRILL) 1882 Syn. Stylochopsis zebra Verrill 1882. Stylochus zebra (Verrill) 1892. Description. — The external features of this familiar Woods Hole polyclad have been adequately described by Verrill (1892). The body is of elongated oval or oblong form, obtuse at each end, of thick and firm texture, reaching a length of 30-40 mm. The dorsal surface has a color pattern of alternating yellowish or white and brown or chocolate cross-bars of which the most anterior and posterior ones are V-shaped. Near the anterior end 'are two rounded tentacles filled with eyes and between the tentacles are the paired cerebral groups of eyes. These *HAi 130 LIBBIE H. HYMAN are loose elongated groups extending behind and before the level of the tentacles. The margin is completely encircled with a band of small eyes, wider in front, diminishing caudally. Study of sections of the copulatory apparatus proves for the first time that this species was correctly placed by Verrili in the genus Sty- lochus. The uteri and vasa deferentia course alongside the posterior part of the very large, much ruffled pharynx behind the posterior end of which occurs the copulatory apparatus. The rear ends of the vasa deferentia expand and acquire muscular walls as they ascend to join the seminal vesicle proper, producing the three-parted type of seminal vesicle characteristic of several species of Stylochus (Fig. 1). The central lobe of this or seminal vesicle proper continues posteriorly as the ejaculatory duct which narrows and becomes less muscular, receiving near the penis the duct of the prostate vesicle. The latter is a large oval chambered sac with thick muscular walls penetrated by the necks of extracapsular granule glands (Fig. 1). The erect position of the prostate vesicle is unusual for the genus ; the vesicle is commonly hori- zontal. After receiving the prostatic duct the ejaculatory duct continues through the penis, a simple conical projection into the small male atrium. The latter opens below by the male genital pore. Shortly behind this occurs the female genital pore, from which the muscular ciliated vagina extends dorsally, then curves posteriorly and downward to terminate where it receives the uteri (Fig. 1). A Lang's vesicle is absent as in the genus in general. The color pattern and the erect prostatic vesicle serve to distinguish this species from others of the genus. Distribution. — Region of Woods Hole, on wharves and pilings and along shores; also found by Verrili at other places in Vineyard Sound and in Long Island Sound, near New Haven, Conn., in dead shells of Sycotypus occupied by hermit crabs. STYLOCHUS ELLIPTICUS (GIRARD) 1850, NEW COMB. Syn. Planocera elliptica Girard 1850. Stylochopsis littoralis Verrili 1873. Stylochus littoralis (Verrili) Lang 1884. Eustylochus ellipticus (Girard) Verrili 1892. I am fortunately able to give a description of this species and estab- lish its correct systematic position. There are available two vials of Verrill's collecting, a specimen in the Woods Hole material, and a num- ber of individuals presented by Dr. Spieth and Mr. Burkenroad whom I here wish to thank for their trouble. Study of this material proves that SOME POLYCLADS OF THE WOODS HOLE REGION 131 the animal is a typical member of the genus Stylochus and hence that Verrill was in error in erecting a new genus Eustylochus for the species (Verrill, 1892). Eustylochus thus becomes a synonym of Stylochus. Description. — 5\ ellipticus (Fig. 2) is of oval or elliptical form and may reach a length of 20-25 mm. ; specimens as small as 4 mm. were found to have the male system fully developed but the female system was immature. The body is flat but somewhat thick and opaque with undulated margins. The color is stated by Verrill and others to be various shades of cream, yellow, or brown, veined or reticulated with a lighter shade, or freckled with golden brown on a lighter ground color; the periphery is clear and translucent and there is usually a light cream to brown stripe in the posterior middorsal region. There are two nuchal tentacles filled with eyes ; these are probably elongated and pointed in life but in the preserved specimens were rounded and some- what sunk into depressions. Between and in front of the tentacles occur the cerebral eyes, subject to much variation. Not infrequently these occur in two pairs of groups of two eyes each (Fig. 2), but often there are three to five or more (up to nine were seen) eyes in each of the four groups and there may be a few scattered eyes farther back over the brain. Specimens have also been seen in which the cerebral eyes were not definitely arranged in four groups but formed a loose band on each side. A band of eyes occurs along the margin of the anterior third to half of the body; these eyes are larger in the mid-anterior region and diminish in size and number around the sides of the body. In many specimens no eyes could be seen along the margin of the posterior half or two-thirds of the body but in some individuals there are a few small eyes scattered along this region so that the entire margin in such cases is encircled with eyes. The pharynx is of moderate size, smaller proportionately than in Stylochus zebra but well folded. The branches of the digestive tract radiate to the periphery making a reticulum through numerous anasto- moses between adjacent branches. Study of the copulatory apparatus in serial sections (Fig. 3) shows that the species is a typical Stylochus and that no grounds exist for separating it into a distinct genus. The genital pores are very close to- gether near the posterior margin. The coiled vasa deferentia run along- side the pharynx and their distal ends turn dorsally and become mus- cular as they join the seminal vesicle proper. There is thus some de- velopment of the tripartite condition of the seminal vesicle seen in several species of Stylochus, but the central lobe is here much larger than the lateral lobes formed by the ends of the vasa deferentia. The central lobe tapers posteriorly into the ejaculatory duct which runs beneath the 132 . LIBBIE H. HYMAN prostatic vesicle to open into the penis lumen. The prostatic vesicle is a very large oval sac with a chambered interior and a thick muscular wall penetrated by the necks of extracapsular granule glands which discharge into the chambers of the interior. The prostatic vesicle has the hori- zontal orientation characteristic of the genus. Its duct joins the ejacu- latory duct in the interior of the penis and the common duct so formed continues to the penis tip (Fig. 3). The penis papilla is a moderately muscular conical projection armed with a short stylet situated in the lumen. The occurrence of a stylet appears to be unusual in the genus Stylochus. The expanded male atrium closely surrounds the penis papilla and exits below by the male genital pore. The lining of the male atrium and the base of the penis papilla are filled with glandular secretion. The large female pore, situated immediately behind the male pore, is lined by a very tall epithelium also having a glandular border of eosino- philous secretion. The vagina accompanied by cement glands proceeds dorsally and forward as a narrow tube which near the dorsal body wall turns sharply backwards and terminates by receiving the uteri. Distribution. — Stated by Verrill to occur from New Haven, Conn., to Casco Bay, Maine, under stones in shallow' water and tide-pools ; not uncommon now at Cold Spring Harbor, Long Island, and along the shores of the west end of Long Island Sound where it joins the East River. Apparently scarce along the Connecticut shore, where four speci- mens were taken under a rock at Indian Neck, by M. D. Burkenroad. One specimen lent by Mr. Gray came from the bottom of a light-ship, off Woods Hole. Verrill's specimens from the Peabody Museum were as follows: one vial containing six animals, labelled Savin Rock (this is For all figures the numbers listed below have the same connotations. 1, marginal eyes, 2, tentacles, 3, tentacular eyes, 4, cerebral eyes, 5, brain, 6, pharynx, 7, vasa deferentia, 8, enlarged muscular terminations of the vasa defer- entia, 9, seminal vesicle, 10, prostate vesicle, 11, glandular chambers of the prostate vesicle, 12, extracapsular granule glands, 13, penis papilla, 14, ejaculatory duct, 15, prostatic duct, 16, stylet, 17, male genital pore, 18, male atrium, 19, penis pocket, 20, female genital pore, 21, vagina, 22, cement pouch, 23, cement glands, 24, en- trance of the uteri into the vagina, 25, uteri, 26, stalk of Lang's vesicle, 27, Lang's vesicle, 28, cirrus, 29, retractor muscle of the cirrus, 30, protractor muscle of the cirrus, 31, musculo-glandular fold of the vagina, 32, main intestine, 33, penis sheath, 34, female genital atrium, 35, granule glands, 36, common genital pore, 37, pouch from genital atrium. PLATE I FIG. 1. Copulatory apparatus of Stylochus zebra, sagittal view. FIG. 2. Stylochus cllipticus, dorsal view of whole mount. FIG. 3. Stylochus ellipticus, sagittal view of copulatory apparatus. SOME POLYCLADS OF THE WOODS HOLE REGION 133 PLATE I 134 LIBBIE H. HYMAN near New Haven), under stones, Oct. 1871; and one vial containing one animal labelled Stylochus littoralis, New Haven, Oct. 1872. SECTION SCHEMATOMMATA BOCK 1913 Definition. — Acotylea without marginal eyes ; commonly with paired cerebral and tentacular clusters of eyes close together and far back from the anterior margin. FAMILY LEPTOPLANIDAE STIMPSON 1857 (EMEND. BOCK) Definition. — Schematommata with flat thin elongated bodies ; pros- tatic vesicle, when present, interpolated ; male apparatus directed back- wards ; definite tentacles usually absent but may be present ; dorsal sur- face colored, generally some shade of brown; uteri confluent in front of the pharynx. GENUS NOTOPLANA LAIDLAW 1903 Syn. Leptoplana (in part). Definition. — Leptoplanidae with true seminal vesicle and chambered prostatic vesicle into which the ejaculatory duct penetrates ; male ap- paratus close behind the pharynx, far removed from the posterior mar- gin ; penis with or without a stylet ; eyes in four clusters ; tentacles gen- erally absent or rudimentary. It has been necessary to limit the genus Leptoplana into which it was formerly customary to throw almost any leptoplanid species to the first recognizable species ascribed to Leptoplana, namely L. tremellaris (O. F. Miiller) 1774, and to other species which can be shown to be similar to this species. It unfortunately happens that the anatomy of L. tremellaris is somewhat aberrant and hence very few species remain in the genus Leptoplana. I know of none on the Atlantic coast of the United States but one occurs in Puget Sound and some may exist on the California coast although most of the Pacific coast species called Leptoplana must be transferred to other genera. With this limitation of the genus Lep- toplana, the majority of the typical leptoplanids fall into the genus Notoplana, a genus of many species, all so much alike externally that identification is practically impossible without recourse to serial sections of the copulatory complex. Only one species of Notoplana is so far known from the Atlantic coast of North America but many occur on the Pacific coast. SOME POLYCLADS OF THE WOODS HOLE REGION 135 NOTOPLANA ATOMATA (O. F. MlJLLER) 1776 Syn. ? Lcptoplana variabilis (Girard) Ver»ill 1892. Description. — This species has already been well described by Bock (1913) and hence I shall be brief. The species, seen at Mt. Desert Island, Maine, may reach a length of 28 mm. and has the typical ex- ternal aspect of the genus. The elongated oblanceolate body is broadest across the anterior fourth and then tapers to the rounded posterior end (Fig. 4). Young specimens are even more tapering than fully grown ones. The color is grayish brown above, more or less flecked and streaky. There are the usual four clusters of eyes, two rounded ten- tacular clusters representing the sites of the aborted tentacles and con- sisting of 6-10 large eyes and a few small ones; and the elongated cere- bral clusters extending forward from the brain and composed of about 15-40 eyes of various sizes, but smaller than the largest of the tentacular eyes. The number of eyes increases with age so that young specimens have few eyes. The whole anatomy is typically leptoplanid with a central elongated ruffled pharynx, radiating anastomosed digestive branches, uteri en- circling the pharyngeal pocket, and vasa deferentia looping backwards across the stalk of Lang's vesicle (Fig. 5). The details of the copu- latory apparatus have been discussed by Bock (1913) who presents a sagittal view and I give a ventral view (Fig. 5) made from a pressed live specimen. The vasa deferentia enter separately the angles of the rounded muscular seminal vesicle from which the ejaculatory duct pene- trates more than halfway into the lumen of the spherical thick- walled chambered prostatic vesicle. From this the elongated curved penis pocket leads to the male atrium. A very small penis papilla next the prostatic vesicle bears the moderately long curved stylet which occupies the penis pocket and often protrudes from the male genital pore. There is a penis sheath. The wide muscular vagina continues beyond the entrance of the uteri forward, then bends backwards as a narrow stalk terminating in an enlarged Lang's vesicle (Fig. 5). Bock is of the opinion that Lcptoplana variabilis is identical with Notoplana atomata. While the former is undoubtedly a species of Noto- plana, I find some points in Verrill's description of it (1892) which do not fit my observations on N. atomata. Thus Verrill gives the shape of L. variabilis as oblong or elliptical while N. atomata is always wedge- shaped, often markedly so, and describes the penis stylet as much longer and more coiled than I have seen it. I have not seen any other shape in a number of specimens examined than the simple bend figured. 136 LIBBIE H. HYMAN Locality. — Notoplana atomata appears to be widely distributed along shores of the North Atlantic coast from Scandinavia to Maine ; under stones, Mt. Desert Island, Maine, and probably northward. GENUS EUPLANA GIRARD 1893 Syn. Discoplana Bock 1913. Definition. — Leptoplanidae of somewhat elongated form, without a prostatic vesicle ; genital pores close together ; tentacles lacking ; penis unarmed, mostly small or even absent ; Lang's vesicle present or absent. Type : Enplana gracilis (Girard) 1850. I regret the necessity of making Bock's genus Discoplana a synonym of Euplana but the circumstances leave me no other course. EUPLANA GRACILIS (GIRARD) 1850 Syn. Prosthiostomum gracile Girard 1850. Prosthiostomum gracile Girard, Verrill 1892. Euplana gracilis Girard 1893. A number of specimens of a poly clad considered unnamed by mem- bers of the invertebrate staff at Woods Hole were secured in 1935. Study of serial sections revealed that the animal belonged to Bock's genus Discoplana and for some time the form was regarded as a new species of Discoplana. Then Verrill's figure of Prosthiostomum gracile Girard was recognized as identical with the poly clad, in question (Ver- rill, 1892, p. 497). The question then arose whether the species actu- ally is Girard's species, as Verrill thought. It is my opinion that there is very little doubt of this, although Girard was entirely mistaken in placing the animal in Prosthiostomum. Later, 1893, Girard recognized this error and created for the species the new genus Euplana. This genus appears to be quite valid under the rules and therefore I must regretfully declare Discoplana Bock 1913 a synonym of Euplana Girard 1893. The correct name of the animal is then Euplana gracilis (Girard).1 PLATE II FIG. 4. Notoplana atomata, from life, dorsal view, Mt. Desert Island, Maine. FIG. 5. Dorsal view of copulatory apparatus of Notoplana atomata as seen in live pressed specimen. FIG. 6. Euplana gracilis, from life, dorsal view, Woods Hole, Mass. FIG. 7. Another view of the anterior end of Euplana gracilis, from life. FIG. 8. Sagittal section of the copulatory apparatus of Euplana gracilis. 1 If the identity of the species under consideration with Prosthiostomum gracile be questioned, then Discoplana would again become a valid genus and the animal would be a new species of Discoplana. SOME POLYCLADS OF THE WOODS HOLE REGION 137 PLATE II 138 LIBBIE H. HYMAN Description. — E. gracilis is of slender elongated shape, resembling a fresh-water planarian (Fig. 6). The anterior end is at times somewhat pointed (Fig. 6), at other times rounded (Fig. 7). From the level of the brain the body tapers gradually to the obtuse or rounded posterior end. The species is small, with a maximum length of 8 mm. and of yellowish gray or brownish gray coloration. The eyes are the most dis- tinctive external feature and readily serve to distinguish this species from all other North American polyclads. There are typically six eyes on each side, four in a lengthwise row, representing the cerebral clusters, and two eyes obliquely placed close together, representing the tentacular clusters. Specimens have been seen with some slight irregularity of eye number, such as one or two additional eyes, but the described arrange- ment is the rule. There are no marginal ocelli as supposed by Verrill. The small, slightly ruffled pharynx is situated far forward, a position explaining the mistake in placing the species in Prostliiostomwn. Be- hind the pharynx the median intestinal trunk is conspicuous in whole mounts (Fig. 6). From this trunk lateral branches extend to the periphery. The main feature of the copulatory apparatus (Fig. 8) is the absence of the prostatic vesicle. The male and female genital pores are situated close together some distance behind the pharynx at about the middle of the body. Behind the pharynx the expanded thin-walled vasa deferentia unite to a but slightly muscular seminal vesicle which tapers directly to the male genital pore (Fig. 8). There is no penis papilla and no pros- tate vesicle nor are any extracapsular granule glands discernible in the available material. The female pore shortly behind the male pore leads into an expanded cement pouch into which open great numbers of cement glands (Fig. 8). The vagina then curves posteriorly and ventrally and terminates where it receives the two uteri. There is no Lang's vesicle. The uteri are confluent in front of the pharynx, then course along the sides of the pharynx behind which they expand considerably before join- ing the vagina. E. gracilis closely resembles in its sexual anatomy several other species of Euplana (==Discoplana) differing chiefly in the total absence of a penis papilla. Distribution. — Found in abundance in the Eel Pond at Woods Hole, Mass., on wharves and pilings among masses of hydroids and sponges ; apparently formerly common in Boston Harbor and along the Connecti- cut shore. Neotype. — One whole mount ; paratype, one set of sagittal sections, deposited in the museum at Woods Hole. SOME POLYCLADS OF THE WOODS HOLE REGION 139 GENUS STYLOCHOPLANA STIMPSON 1857 (EMEND. BOCK) Definition. — Leptoplanidae with true seminal vesicle and non-cham- bered prostate vesicle, the latter not penetrated by the ejaculatory duct ; tentacles present or absent ; eyes in paired cerebral and tentacular groups ; separate or fused genital pores, which may be near the posterior margin. In general aspect the species of this genus closely resemble those be- longing to Notoplana. The essential difference between them lies in the prostate vesicle which is not chambered and not penetrated by the ejacu- latory duct. STYLOCHOPLANA ANGUSTA (VERRILL) 1892 Syn. Leptoplana angusta Verrill. [not Stylochoplana angusta (Verrill) Palombi 1928]. The finding at the Peabody Museum of a vial containing three speci- mens which are undoubtedly " Leptoplana " angusta has enabled me to determine the systematic status of this species. It was found to be a leptoplanid which fits best into the genus Stylochoplana although not entirely conforming to the present conception of this genus. It seems preferable, however, to expand the definition of the genus than to create a new genus for the species. In 1928 Palombi identified a Stylochoplana species obtained by the Suez Canal expedition as Leptoplana angusta, and called it Stylocho- plana angusta (Verrill). This identification is erroneous as the species in question differs decidedly from the latter. Professor Palombi has been informed of the necessity of creating a new specific name for his species. Description. — S. angusta is stated by Verrill to be of elongated elliptical form, thin, with flexible undulated margins, rounded anterior end, and notched posterior margin. Figure 9 is a drawing of a whole mount made from one of Verrill's specimens ; the other two were sec- tioned sagitally. The maximum length is given by Verrill as 12-16 mm., the breadth 4—6 mm., and the color as various shades of light brown. The cerebral and tentacular eyes form a continuous band on each side (Fig. 9), of which the posterior group of larger eyes on either side of the brain undoubtedly represents the tentacular eyes while the band of smaller eyes extending forward from these are the cerebral eyes. The narrow ruffled pharynx is somewhat posterior in position and this presumably accounts for the very posterior location of the copulatory complexes. 140 LIBBIE H. HYMAN There is a large common genital pore close to the posterior margin shortly in front of the notch. This continues into a genital atrium which curves anteriorly receiving the vagina in its dorsal wall and the penis in its anterior wall. The small male copulatory complex (Fig. 10) conforms to that of the genus in general. The vasa deferentia after a union enter the anterior end of the elongated oval seminal vesicle. From this the curved ejaculatory duct runs to the prostatic vesicle. This is of oval form with a thin muscular wall and the interior lined by granule glands. No extracapsular granule glands were seen in the sections ; naturally after sixty years in preservative the animals are not in very good histological condition. From the prostate vesicle, the ejaculatory duct continues backward and enters the small pyriform penis which projects into the male portion of the atrium as a slight conical eminence. The very long vagina runs from the roof of the atrium far forward and then after a short backward bend opens into a large spherical Lang's vesicle (Fig. 10). The uteri join the backward bend so that the stalk of Lang's vesicle is very short. The vagina is slightly muscular along its entire course and is accompanied throughout by cement glands. I am unable to determine from my material whether the uteri are confluent in front of the pharynx but Verrill's statements indicate that this is the case. In the whole mount, the uteri are much distended with eggs (Fig. 9) and extend anteriorly beyond the brain, a condition unusual in the Leptoplanidae. The two sets of sagittal serial sections are identical except for one very puzzling feature. In one of them, the larger of the two, the common genital atrium continues dorsally behind the vaginal entrance as a wide pouch (Fig. 10), which appears to open on the dorsal surface, presumably into the posterior notch. The specimen is unfortunately imperfect at this place so that I cannot be sure of the actual presence of an opening there ; but all appearances indicate that this is the case and I have so drawn the region in question (Fig. 10). The second specimen (Fig. 11) shows no trace whatever of such a pouch although it appears to be quite mature sexually. More material will be necessary to settle the matter of the occurrence and relations of the pouch in question. It may represent a ductus vaginalis, as in Bock's genus Ceratoplana (Bock, 1925). Sty. angusta differs from the typical members of the genus in the far posterior location of the genital pore and the forward course of the vagina. The first point is not of so much importance since Bock (1924) found a Japanese species which he assigned to Stylo choplana and which also has a very posterior common genital pore. Kato (1937) reports SOME POLYCLADS OF THE WOODS HOLE REGION 141 10 PLATE III FIG. 9. Stylochoplana angusta, dorsal view of whole mount, neotype. FIG. 10. Sagittal view of the copulatory apparatus of Stylochoplana angusta, from one of the sets of serial sections, showing problematical pouch from genital atrium dorsally. FIG. 11. Terminal part of the copulatory apparatus of Stylochoplana angusta, _ from the second set of serial sections, showing absence of a pouch. 142 LIBBIE H. HYMAN another Japanese Stylochoplana having the same feature. But in these species, the vagina, despite the posterior situation of the genital pore, makes the usual backward bend, so that Lang's vesicle lies well behind the genital pore. The situation found in Sty. angusta with the long anterior extension of the vagina and the location of Lang's vesicle much anterior to the male apparatus appears to be unique. One could con- sider it a basis for the erection of a new genus but I prefer to leave the species in Stylochoplana, at least for the present. Distribution. — Stylochoplana angusta was found by Verrill at Prov- incetown, Mass., 1879, on the bottom of a whaling vessel recently ar- rived from the Carolina coast. The polyclad probably came from this region, especially as it was associated with other animals said by Verrill to be southern. The vial from the Peabody Museum contained a label on which was printed: Off Cape Cod, U. S. F. C. 1879, and written: Leptoplana, vessel bottom. Neotype. — Whole mount designated as neotype; one set of sagittal serial sections, designated as paratype; both deposited in the Peabody Museum, Yale University. GENUS HOPLOPLANA LAIDLAW 1902 Syn. Planocera, Group B, Lang 1884. Definition. — Schematommata without a true seminal vesicle or penis papilla; ends of the vasa deferentia enormously enlarged into thick- walled muscular tubes serving as accessory seminal vesicles ; penis stylet fastened directly to the prostate vesicle ; tentacles present ; with cerebral and tentacular eye clusters, the latter closely embracing or in the tentacle bases ; no Lang's vesicle ; shape oval or rounded. In 1902 Laidlaw rightly removed some of Lang's species of Plano- cera to a new genus Hoploplana, which does not even belong to the family Planoceridae. It also does not entirely fit into the family Lepto- planidae, so that Bock (1913) left it as an appendix to the latter. It would seem to me desirable to place Hoploplana in a family by itself despite the fact that its affinity to the Leptoplanidae is indicated by the finding in Puget Sound of a species of Notoplana, N . segnis Freeman 1933, which has also very enlarged, highly muscular terminations of the vasa deferentia. HOPLOPLANA GRUBEI (GRAFF) 1892 Syn. Planocera grubei Graff 1892. This species, common on the Sargassum, has been adequately de- scribed and figured by von Graff (1892). Three specimens, brought SOME POLYCLADS OF THE WOODS HOLE REGION 143 to me for identification by Dr. Hadley of the Invertebrate staff at Woods Hole, were taken from the Sargassum in Vineyard Sound, August 3, 1938. The coloration was stated to consist of gray reticulations on a pinkish-brown or flesh ground. M. D. Burkenroad of the Peabody Museum, who has collected a number of specimens of this species from the Sargassum in the Gulf of Mexico and North Atlantic, reported the color as consisting of livid white reticulations on a brown ground. It is clear that von Graff's conception of the color, based on preserved mate- rial, is erroneous. Preserved specimens give no idea of the original coloration. A colored figure of this species made by Dr. J. F. G. Wheeler of the Bermuda Biological Station is being published in my article on the polyclads of Bermuda and the Sargassum (Hyman, 1939). HOPLOPLANA INQUILINA (WHEELER) 1894 Syn. Planocera inquilina Wheeler 1894. Hoploplana inquilina (Wheeler) Bock 1913. This species, familiar at Woods Hole, was rightly removed by Bock to the genus Hoploplana. Wheeler presents an excellent figure of the entire animal but his account of the internal details leaves much to be desired. I have therefore deemed it necessary to make a study of serial sections. Description. — The expanded thin-walled vasa deferentia packed with sperm course alongside the pharynx near the posterior end of which they alter into thick-walled muscular accessory seminal vesicles (Fig. 12). These, after reaching a level behind the male pore narrow, ap- proach the median line, and fuse to a narrow duct, the common vas deferens. This runs anteriorly penetrating into a round somewhat muscular organ which Wheeler considers the penis. It is, in fact, the prostatic vesicle and this mistake accounts for Wheeler's inability to find such a vesicle. It is not, to be sure, a very typical prostate. It has an excessively thick wall, containing muscle fibers coursing parallel to the surface contour and in the center a small area of granule glands radiating around the central small ejaculatory duct. The granule glands are best seen in frontal section (Fig. 12) as they are oriented parallel to such a section. Scattered glands also seem to extend into the thick wall of the prostate vesicle. After passing through the glandular re- gion the ejaculatory duct terminates in the base of the stylet which constitutes the entire penis (Fig. 13). It is directed ventrally in the available specimens, extending into the male atrium. The latter has quite a muscular wall and is lined by an extremely tall columnar epi- thelium which bulges from the genital pore (Fig. 13). 144 LIBBIE H. HYMAN It will be seen from the foregoing description that the male apparatus of H. inquilina accords completely with that found in other members of the genus. The vasa deferentia form accessory seminal vesicles, there is no true seminal vesicle, and the prostate vesicle bears directly the penis stylet without the intervention of any penis papilla proper. The female copulatory apparatus consists simply of the vagina which extends dorsally from the genital pore receiving the secretion of the thickly strewn cement glands. The vagina is somewhat muscular, and lined by tall slender epithelial cells. After making a backward bend, the vagina constricts and beyond this constriction the shell glands cease and the lining epithelium alters to large vacuolated cells. This portion of the vagina resembles histologically a Lang's vesicle but morpho- logically is not such since the uteri open into its ventral end (Fig. 13). Distribution. — Hoploplana inquilina has been recorded only from the vicinity of Woods Hole where it inhabits the mantle cavity of the large gasteropod Sycotypus. FAMILY PLANOCERIDAE Lang 1884 (EMEND. BOCK 1913) Definition. — Schematommata with a cirrus instead of a penis ; cirrus lined with spines, hooks, or ridges eversible to the exterior ; with tenta- cles ; with tentacular and cerebral eyes ; copulatory complexes imme- diately behind the pharynx ; with true or accessory seminal vesicles ; prostate vesicle free or interpolated. GENUS GNESIOCEROS DIESING 1861 - Syn. Pelagoplana Bock 1913. Definition. — Planoceridae with wedge-shaped pellucid bodies; with tentacles containing eyes ; cerebral eyes also present ; with a true seminal vesicle and interpolated prostatic vesicle; cirrus armed with parallel toothed bands; vagina with a powerful musculo-glandular fold; Lang's vesicle transverse. Type, Gnesioceros sargassicola (Mertens) 1833. PLATE IV FIG. 12. Frontal view of the vasa deferentia and prostate vesicle of Hoplo- plana inquilina, showing area of granule glands in the center of the vesicle around the ejaculatory duct. FIG. 13. Sagittal view of the copulatory apparatus of Hoploplana inquilina. FIG. 14. Gnesioceros verrilli, dorsal view, whole mount, Woods Hole speci- men. 2 In 1861 Diesing created the genus Gnesioceros for two species : " Planaria " pellucida and "Planaria" sargassicola both described by Mertens in 1833. Pla- naria pellucida belongs to the genus Planocera and was removed to that genus in 1844 by Oersted. Diesing did not indicate which of the two species he considered typical of the genus Gnesioceros. I hereby remove pellucida from the genus and declare sargassicola to be the type of Gnesioceros, thus making the genus valid. SOME POLYCLADS OF THE WOODS HOLE REGION 145 14 PLATE IV 146 LIBBIE H. HYMAN GNESIOCEROS SARGASSICOLA (MERTENS) 1833 Syn. Planaria sargassicola Mertens 1833. Stylochus sargassicola Ehrenberg 1836. Planocera sargassicola Oersted 1844. Stylochus pelagicus Moseley 1877. Gnesioceros mertensi Diesing 1861. Stylochus sargassicola (Mertens) Lang 1884. Stylochoplana sargassicola (Mertens) Graff 1892. Pelagoplana sargassicola (Mertens) Bock 1913. This is the common polyclad of the Sargassum found on the floating weed in various parts of the world, common in the Gulf of Mexico, Caribbean, and North Atlantic. I took several specimens alive from the Sargassum at Bermuda and Burkenroad collected hundreds of speci- mens during an expedition to study the 'biology of Sargassum inhabi- tants. Alive the animal is pellucid with small brown spots and an expanded anterior end tapering to a pointed posterior end. It has been figured and described by von Graff (1892) and I have given an account of the structure of the cirrus (Hyman, 1939). A specimen of this species was taken from the Sargassum in Vineyard Sound, August 3, 1938. GNESIOCEROS VERRILLI, NEW NAME Syn. Imagine oculifera (Girard) Verrill 1892. (not Imagine oculifera Girard 1853). In 1882 Verrill found in Quisset Harbor a single specimen of a polyclad which he assigned to Girard's species Imagine oculifera. This identification is clearly erroneous since Girard plainly states that Imagine oculifera has a band of eyes completely encircling the margin and hence is a stylochid while Verrill's figure plainly shows a planocerid. Two vials of specimens were obtained from Mr. Gray which I am confident are identical with Verrill's specimen. Examination of these animals proved that they belong in the genus Gnesioceros. As it is necessary to give a specific name to Verrill's species, I propose to call it verrilli. Description. — G. verrilli is very similar in all respects to G. sargas- sicola. The graceful body is anteriorly expanded (Fig. 14), being widest at about the level of the brain, and from there tapers gradually to the pointed posterior tip. The species is small, around 6-8 mm. in length. The preserved specimens were colorless but Verrill gives the color as carmine red. As this color was stated, however, to be limited to the intestinal branches, it is nearly certain that it resulted from SOME POLYCLADS OF THE WOODS HOLE REGION 147 ingested food and that the animal is in fact, like G. sargassicola, pellucid with small brownish marginal spots. There are two elongate nuchal tentacles, each containing several eyes, most of which are in the tentacle bases. Between the tentacles, extending some distance behind and in front of them, are two loose elongate groups of cerebral eyes, of about 8-12 eyes each (Fig. 14). From the long, narrow, slightly ruffled pharynx a rich net of anastomosing intestinal branches radiates to the periphery. The copulatory complexes behind the pharynx are very similar to those of G. sargassicola which have been described by Graff, 1892, and Hyman, 1939. The vasa deferentia enter separately the ventral angles of the pyriform seminal vesicle (Fig. 15) which ascends and narrows to the ejaculatory duct opening into the anterior end of the prostate vesicle. The latter is an elongate structure divided into chambers by cross-partitions and covered with a thick muscular coat which operates the cirrus. The prostate vesicle continues directly into the cirrus, a curved beak-like structure composed of hard yellow material, presum- ably not chitin but an albuminoid. The structure of the cirrus was described and figured in another place (Hyman, 1939) and hence a brief statement will be made here. The cirrus possesses a median groove and thus resembles a conch shell. It is made of parallel toothed bands (Fig. 16) which can be pulled into the interior along the groove by the action of the retractor muscle fastened along one side of the prostate vesicle. The protractor muscle on the other side of the prostate vesicle is attached to the central part of the bands. When it contracts, it pulls the bands to the outside where they take a horizontal course. The cirrus is situated in a pocket which continues into the male atrium and so to the male genital pore. The female genital pore is shortly behind the male pore. From it the ciliated, slightly muscular vagina extends dorsally, then curves back- ward and ventrally and receives the uteri. Beyond this union the female canal continues as the stalk of Lang's vesicle and opens into the latter. Lang's vesicle is unique in the genus Gnesioceros in taking a transverse course, so that it embraces the rear and sides of the female genital pore in a somewhat crescentic shape. The genus is further notable for the presence in the vagina of a powerful musculo-glanclular fold, consisting of a heavy muscle mass to which are attached numerous gland cells. In G. sargassicola this fold is ring-shaped, completely encircling the vagina ; but in G. verrilli it is well-developed only in the anterior wall of the vagina, the side facing the cirrus, and but slightly present and more diffuse in the posterior vaginal wall (Fig. 15). The epithelium over the main muscle mass is reduced to a cuticularized membrane. In G. sar- 148 LIBBIE H. HYMAN gassicola the vaginal epithelium from the musculo-glandular fold to the genital pore is also cuticularized and thrown into lengthwise ridges; but this is not the case in G. verrilli. These differences in the vaginal structure constitute the chief dis- tinctions between G. sargassicola and G. verrilli. Other differences are : the pharynx appears to be narrower and less ruffled in G. verrilli than in G. sargassicola and the prostate vesicle narrower and longer relative to the cirrus region in the former than in the latter. Neotype. — It has seemed desirable to designate a neotype in the form of a whole mount deposited in the museum at Woods Hole ; and a set of serial sections is also presented to this institution. Distribution. — Taken from the eel grass, Devil's Foot Island, Woods Hole, Mass., December, 1930 and October 28, 1931; also found by Verrill in Quisset Harbor, Buzzard's Bay, on sandy bottom in four or five fathoms, September 4, 1882. SUBORDER COTYLEA LANG 1884 Definition. — Polyclads with a sucker behind the female genital pore (a few exceptions); pharynx ruffled to tubular; copulatory complexes generally in the anterior half of the body; prostatic vesicle, when pres- ent, always free ; uteri generally behind the female genital pore ; tentacles when present of the marginal type ; eyes on the anterior margin, on the tentacle bases when these are present. FAMILY EURYLEPTIDAE LANG 1884 Definition. — Cotylea with anteriorly directed tubular pharynx; male apparatus below or immediately behind the pharynx, directed forward, with a free prostate vesicle and armed with a stylet ; with or without uterine glands; tentacles vary from well-developed pointed marginal folds to none ; with paired cerebral clusters of eyes and paired marginal clusters which are on the tentacles when these are present. GENUS EURYLEPTA EHRENBERG 1831 (EMEND. LANG 1884) Definition. — Euryleptidae with well-developed pointed tentacles; main intestine with very few (up to about five) lateral branches; male PLATE V FIG. IS. Sagittal view of the copulatory apparatus of Gnesioceros verrilli. FIG. 16. Prostate vesicle and cirrus of Gnesioceros verrilli. FIG. 17. Tentacles and eye distribution of Eurylepta maculosa, specimen col- lected by Verrill. FIG. 18. Sagittal section of the same specimen as Fig. 17, showing pharynx and female copulatory apparatus ; because of damage the male apparatus, to be expected immediately behind the pharynx, is missing. SOME POLYCLADS OF THE WOODS HOLE REGION 149 PLATE V 150 LIBBIE H. HYMAN apparatus beneath the posterior end of the pharynx ; uterine glands mostly one pair or none. EURYLEPTA MACULOSA VERRILL 1892 A single specimen of this species was found at the Peabody Mu- seum; after the eye arrangement had been drawn (Fig. 17) the animal was cut into sagittal sections but these were not very satisfactory. However, it seems desirable to describe the findings. Description. — The species is said by Verrill to have a thin, change- able, elliptical or oblong body, 10-12 mm. long, with thin often undu- lated margin. The anterior margin is upfolded to make two fairly long bluntly pointed tentacles which bear eyes on their anterior faces (Fig. 17). There are, as also noted by Verrill, no eyes at or near the tentacle tips. There are elongated paired cerebral groups of eyes which are continued forward to the margin between the tentacles by scattered clusters of eyes. The preserved specimen was distinctly mottled with brown and white ; Verrill gives the color as pale, yellowish or pinkish white, irregularly specked and mottled or veined with purplish or brown. The serial sections show a typical euryleptid structure (Fig. 18) with short tubular pharynx directed forward. From the pharynx the main intestinal trunk runs backward in the middorsal line and has about three pairs of wide lateral branches, thus conforming to the genus Eurylepta. To either side of the intestinal trunk is a large thin-walled uterus stuffed with eggs. No uterine glands have been found. The region immediately behind the pharynx where the male copulatory ap- paratus would be expected is unfortunately broken and no trace of the male copulatory apparatus has been found. The female apparatus (Fig. 18) appears to be completely present. The female pore, shortly behind the pharynx, leads into a small expanded atrium from which a narrow vagina, receiving the innumerable cement glands, proceeds dorsally. The great mass of the cement glands almost obscures the female duct. The vagina widens into a small chamber from which a duct runs posteriorly and receives the uteri. There appears to be little doubt that the species was correctly placed in the genus Eurylepta. Distribution. — Found by Verrill on piles, in mud, and among algae at low tide at Woods Hole and vicinity, stated to be uncommon. The Peabody Museum specimen was labelled : Eurylepta maculosa V. cotype, mud, Woods Hole, Aug. 2nd. Specimen. — The set of sections made from the specimen has been returned to the Peabody Museum. SOME POLYCLADS OF THE WOODS HOLE REGION 151 Whole mounts or preserved specimens of most of the species de- scribed in this paper (except those belonging to the Peabody Museum) have been deposited in the Museum at Woods Hole. In some cases a set of serial sections has also been presented to this institution. SUMMARY 1. Stylochus zebra (Verrill) 1882 is shown to have been correctly placed by Verrill in the genus Stylochus. 2. Eustylochus ellipticus (Girard) 1850 is shown to be a typical member of the genus Stylochus, its name thus becomes Stylochus ellip- ticus, and Eustylochus becomes a synonym of Stylochus. 3. Notoplana atomata, usually erroneously called Lcptoplana vari- abilis, is the most common polyclad of the Maine coast. 4. Prosthiostomum gracilc Girard 1850 has been rediscovered at Woods Hole. The later name Euplana gracilis which Girard gave to this species is valid and Discoplana Bock 1913 becomes a synonym of Euplana Girard 1893. 5. Specimens of Leptoplana angusta Verrill 1892 have been found at the Peabody Museum, Yale University. The species is placed in the genus Stylochoplana although it differs somewhat from the typical members of the genus. The name is then Stylochoplana angusta (Ver- rill). 6. As noted by Bock (1913), Planocera inquilina Wheeler 1894 must be transferred to the genus Hoploplana and the correct name of the species is Hoploplana inquilina. What Wheeler considered to be the penis is in reality the prostatic vesicle. 7. Two common and well-known Sargassum polyclads have been taken on the Sargassum in Vineyard Sound ; these are Hoploplana grubci (Graff) 1892 and Gnesioccros sargassicola (Mertens) 1833. Pelago- plana Bock 1913 becomes a synonym of Gnesioceros Diesing 1861. 8. Another species of Gnesioccros, G. verriUi, new name, occurs around Woods Hole ; it was seen by Verrill in 1882 and erroneously identified by him as Imagine oculifcra. 9. A single specimen of Eurylcpta inaculosa Verrill 1892 was found at the Peabody Museum; some details of its anatomy are described. ADDENDUM Since this manuscript was sent to press there have appeared three articles on Atlantic coast polyclads by A. S. Pearse and associates (Proc. U. S. Nat. Mus., 86, Jour. Elisha Mitchell Sci. Soc., 54, and Bull. Mt. Desert Is. Biol. Lab. 1938). I am unable to agree with many of the identifications, taxonomic placing, and anatomical findings in these ' 152 LIBBIE H. HYMAN • articles but detailed criticism will be withheld until a study of Pearse's specimens now under way has been completed. LITERATURE CITED BOCK, S., 1913. Studien iiber Polycladen. Zool. Bidrag, 2 : 31. BOCK, S., 1924. Eine neue Stylochoplana aus Japan. Arkiv. Zool., 16: 1-24. BOCK, S., 1925. Planarians. Parts I-III. Papers from Dr. Th. Mortensen's Pacific Expedition 1914--1916. No. 25. Vidensk. Medd. jra Dansk naturh. For en., Vol. 79: 1. DIESING, K. M., 1861. Revision der Turbellarien. Abtheilung : Dendrocoelen. Sitz'ber. math, naturw. Klasse Akad. Wissensch. Wien, 44 (Abtheilung) I) : 485. FREEMAN, D., 1933. The polyclads of the San Juan region of Puget Sound. Trans. Am. Micros. Soc., 52 : 107. GIRARD, CH., 1850. Several new species of marine planariae of the coast of Mas- sachusetts. Proc. Boston Soc. Nat. Hist., 3: 251. GIRARD, CH., 1853. Descriptions of new nemerteans and planarians from the coast of the Carolinas. Proc. Acad. Nat. Sci. Phila., 6: 365. GIRARD, CH., 1893. Recherches sur les Planaries et les Nemertiens de 1'Amerique du Nord. Ann. Sci. Natur., Ser. 7, 15 : 145. GRAFF, L. VON, 1892. Pelagische Polycladen. Zeitschr. wiss. Zool., 55: 189. HYMAN, L. H., 1939. Polyclad and acoel Turbellaria from Bermuda and the Sargassum. In press, Bingham Oceanogr. Foundation. KATO, K., 1937. Polyclads collected in Idu, Japan. Japan. Jour. Zool., 7: 211. LAIDLAW, F. F., 1902. Marine Turbellaria. Fauna and Geography of the Mai- dive and Laccadive Archipelagoes. I, Part 3. LAIDLAW, F. F., 1903. On a collection of Turbellaria Polycladida from the Straits of Malacca (Skeat Expedition, 1899-1900). Proc. Zool Soc. London, 1903, Part 1, p. 301. LANG, A., 1884. Die Polycladen. Fauna und Flora des Golfes von Neapel, Vol. 11. MERTENS, H., 1833. Untersuchungen iiber den inneren Bau verschiedener in der See lebender Planarien. Mem. Acad. Imper. Sci. St. Petersbourg, Ser. V, Sci. math. phys. nat., Vol. 2. MULLER, O. F., 1776. Zoologicae Danicae Prodromus seu animalium Daniae et Norvegiae indigenorum characteres, nomina et synonyma imprimis popu- larium. Havniae. PALOMBI, A., 1928. Report on the Turbellaria. Zoological Results of the Cam- bridge Expedition to the Suez Canal. 1924. Trans. Zool. Soc. London, 22 (Part 5) : 580. STIMPSON, W., 1857. Prodromus descriptionis animalium evertebratorum, quae in Expeditioni ad Oceanum Pacificum Septentrionalem, a Republica Federata missa Cadwaladero Ringgold et Johanne Rodgers ducibus, observavit et descripsit. Proc. Acad. Nat. Sci. Phila., 9: 149. VERRILL, A. E., 1873. Report upon the invertebrate animals of Vineyard Sound and the adjacent waters, etc. U. S. Comm. Fish and Fisheries, Comm. Reft, for 1871-72. VERRILL, A. E., 1882. Notice of the remarkable marine fauna occupying the outer banks off the southern coast of New England, No. 7, and of some addi- tions to the fauna of Vineyard Sound. Am. Jour. Sci., 24 : 360. VERRILL, A. E., 1892. Marine planarians of New England. Trans. Conn. Acad. Arts and Sci., 8 : 459. WHEELER, W. M., 1894. Planocera inquilina, a polyclad inhabiting the branchial chamber of Sycotypus canaliculatus, Gill. Jour. Morph., 9: 195. Early references not listed here will be found in Lang (1884). EFFECTS OF COLCHICINE ON THE CLEAVAGE OF THE FROG'S EGG (RANA PIPIENS) DOROTHY M. KEPPEL AND ALDEN B. DAWSON (From the Biological Laboratories, Raddiffe College, and Harvard University) The most striking biological effect of the alkaloid drug, colchicine, is its capacity, when applied in the proper concentration, to arrest mitosis in the metaphase (Dustin, 1934, 1935). Under certain conditions, especially in plants, this drug has also been shown to be effective in pro- ducing polyploidy, doubtlessly due to its ability to inhibit to a greater or lesser degree the orderly sequences of the mitotic process (Blakeslee and Avery, 1937; Nebel and Ruttle, 1938). There have been a great many studies of the effect of physical and chemical changes in the external environment on the developing egg of the frog. Some of the early results obtained, especially by centrifuging (O. Hertwig, 1897), and by exposure to high temperatures (O. Hert- wig, 1898), rather closely parallel the effects produced by treatment with colchicine. In the present report the observations are limited to the changes induced by relatively short exposures to the drug applied almost immediately after fertilization. Methods The eggs of Raua pipiens were obtained and inseminated by the method of Rugh (1934). After a quarter of an hour the eggs were flooded with water and in another quarter of an hour were separated by dissection but the jelly was not removed. They were then put into the colchicine solutions. In later experiments the solutions were made up and used in the dark, except for a red light, since colchicine is affected in solution by light (Lits, 1936). The experiments were conducted at 8° and 20° C. Description In the first series of experiments, the colchicine solutions were made up in concentrations ranging from 1 : 1000 to 1 : 1,000,000 and the eggs were left in the solutions continuously. In the 1 : 1000 solution all cleav- age activity was^uppressed, and eventually disintegration set in. In the 1 : 10,000 solution the first cleavage took place as in the controls, the second cleavage set in at the normal time but showed some irregularities. 153 154 D. M. KEPPEL AND ALDEN B. DAWSON Soon after this, the membranes of most of the eggs seemed to be tightly stretched, appearing as though an abnormal amount of water had been absorbed. The cleavage grooves became more and more shallow. No further cleavage took place, and four and a half hours after fertilization no evidences of cleavage were visible except the presence of light streaks on the animal hemisphere. These streaks, which were due to the absence of superficial pigment, coincided with the position of the cleavage fur- rows. They appeared to have been caused by the movement of pigment deeper into the egg in association with the development of the clefts. In the 1 : 100,000 solution of colchicine, the first, second, and third cleavages frequently took place as in the controls. Shortly after the initiation of the third cleavage, many of the eggs showed diminished grooves, the others continued dividing as in the controls. On the follow- ing day, some of those in which the grooves were fading showed no trace of cleavage except the light streaks, while others had the animal pole region largely cellular with the vegetative hemisphere uncleaved. The proportion of cellular to non-cellular material varied considerably. These were all dead on the third day. In the 1 : 1,000,000 solution, the early cleavage was like that in the controls, no fading being evident except possibly on the vegetative halves. The second day embryos varied from completely normal to those with a small cell cap in the region of the animal pole, the rest of the egg being uncleaved. These embryos also, even those that appeared normal, were dead on the third day. It is evident from these preliminary experiments that the eggs varied considerably in their susceptibility to colchicine, and that with a con- tinuous exposure it was relatively toxic. A 1 : 100,000 exposure for one hour was found to give a fair number of abnormal forms without too high mortality and even a concentration of 1 : 10,000 for the same length of time was also satisfactory, although the mortality was somewhat greater. Most of the eggs in both these concentrations gave rise to normal embryos. The following studies, then, were made on selected individuals showing the characteristic abnormalities. It is extremely difficult to interpret fully many of the abnormalities obtained by the treatment with colchicine. It was not possible to secure an orderly and progressive series of developmental stages as in normal development. The wide range of susceptibility of the eggs to the drug and the diverse effects obtained make it necessary to consider each egg individually and attempt to relate the conditions in earlier cleavages to the morphological patterns found in the more advanced embryos. In general the segmentation of treated eggs becomes meroblastic, with the EFFECTS OF COLCHICINE ON CLEAVAGE 155 cap of cells produced in the region of the animal pole resting upon a more or less completely undivided yolk. Early Cleavage Reference has already been made to the fading of cleavage. Ob- servations on living eggs and sections show that it takes place first at the vegetative pole, and sometimes a furrow extending halfway down the egg is left (Figs. 1, 2, and 3), while in other cases no trace of the cleavage remains, or only a shallow depression may persist at the surface of the animal pole. In some in which no external groove is left a cleft inside the egg can be seen in sections. In rare instances the first cleavage may persist even on the vegetative hemisphere, though subsequent cleav- ages are restricted in their persistence to the animal hemisphere. The nuclei may go on dividing although there is no cytoplasmic division. In several cases in which the early cleavage furrows had completely dis- appeared, a curious sort of delayed or secondary cleavage appeared at about the time when the controls were in the blastula stage (three days at 8° C.). In such cases a few relatively small cells were outlined on the surface, and contained one or more large vacuolated nuclei. More similar nuclei appeared in the main body of the egg. The significant result of the arrest or retardation of the early mitosis and the secondary fading of the early cleavage furrows, especially in the vegetative hemisphere, is the production of an early blastula in which definitive cells tend to be restricted to the animal hemisphere. The de- gree of restriction of cell cleavage is variable. In extreme cases only a small cap of cells is formed but all gradations from this condition to almost complete cleavage of the yolk of the vegetative half may be encountered. Blastitlac In many embryos with only a small cellular area in the region of the animal pole a stage comparable to the blastula is not attained. The cells divide progressively but do not assume any regular pattern. They are usually loosely grouped above the undivided yolk mass without the formation of a definite segmentation cavity (Figs. 4 and 6). This con- dition was observed most frequently in embryos which had developed at the lower temperature, 8° C. In others, a modified blastula, consisting of a cellular roof which is separated more centrally from the uncleaved yolk by a segmentation cavity or blastocoele, is formed. This cavity apparently may be sec- ondarily enlarged by the progressive vacuolation of the adjoining yolk mass (Fig. 7). Below the cavity are the free nuclei of the yolk, becom- 156 D. M. KEPPEL AND ALDEN B. DAWSON ing larger, more abnormal, and more irregularly spaced toward the vege- tative pole. They are usually surrounded by dense accumulations of pigment. In embryos in which approximately half of the material is cellular the blastula is frequently more complete with a cellular floor of variable thickness. The roof is often differentiated into two layers comparable to the outer, epidermal and inner, nervous layer. Occasionally, however, part of the animal half may fail to undergo segmentation while a con- siderable amount of the yolk is cleaved." In such cases the degree of primary segmentation that may have occurred in the yolk is difficult to determine. A study of living animal-hemisphere embryos showed that the first two cleavages could often be seen on the surface of the vegetative pole and subsequently faded. Whether or not these cleavages EXPLANATION OF PLATE Figures 1 to 9 are photomicrographs taken at several magnifications. The magnification was varied in order to bring out significant details or to include sufficient area to show relationships. In all instances the solutions of colchicine \vere applied for one hour, beginning thirty minutes after insemination. PLATE I EXPLANATION OF FIGURES FIGS. 1 and 2. Modification of early cleavage; 4 hours, 20 minutes after in- semination ; temperature 20° C. ; colchicine, 1 : 10,000. FIG. 3. Early cleavage ; 5 hours, 20 minutes after insemination ; temperature 20° C. ; colchicine, 1 : 10,000. The plane of section is oblique to the polar axis giving a false impression of the degree of cleavage of the vegetative hemisphere. FIG. 4. External view of late cleavage showing restriction of cleavage to the animal half of the egg ; 72 hours after insemination ; temperature 8° C. ; colchicine, 1 : 100,000. FIG. 5. Partial embryo ; 48 hours after insemination ; temperature 20° C. : colchicine, 1 : 100,000. From sections it appears that the first cleavage persisted and then cell division was restricted in one blastomere more than in the other. FIG. 6. Late atypical cleavage; 72 hours after insemination; temperature 8° C. ; colchicine, 1 : 100,000. Cleavage is sharply restricted to a cap of cells in the region of the animal pole. Large, numerous yolk-nuclei with concentrations of pigment about them are scattered through the uncleaved yolk. FIG. 7. Animal-hemisphere blastula ; 48 hours after insemination ; temperature 20° C. ; colchicine, 1 : 100,000. Only the roof of the blastocoele is cellular. The floor is composed of unsegmented yolk which is highly vacuolated. The cellular configuration (on the right) possibly represents the beginning of gastrulation by involution. FIG. 8. A transverse section of the upper half of an abnormal gastrula ; 48 hours after insemination ; temperature 20° C. ; colchicine, 1 : 100,000. Neural plate, notochord, mesoderm and an incomplete layer of entoderm are present. The under- lying yolk is undivided and ventrally an area of degeneration is seen. FIG. 9. An area from the cellular roof of a blastula; -18 hours after in- semination; temperature 20° C. ; colchicine 1:100,000. The nuclei of this embryo are of two distinct sizes. The smaller are of normal size. The larger may be due to tetraploidy. EFFECTS OF COLCHICINE ON CLEAVAGE 157 \. " *• „ 6 N .«*. <* &- ^ 8 '.'* i PLATE I 158 D. M. KEPPEL AND ALDEN B. DAWSON actually passed completely through the yolk was not determined, but it seems fairly safe to conjecture that they had not. The formation of the cellular floor of the blastocoele may have occurred simultaneously j ~* with cleavage of the cells nearer the animal pole but there is evidence that in many cases the cleavage of the yolk occurs secondarily, probably through the progressive organization of cells from a region comparable to the germ-ring of normal embryos. As far as it is possible to judge from sections, the process of cellulation occurs around free nuclei di- rectly adjacent to previously formed cells. In these regions the free nuclei are more likely to be normal in size and shape than those at a greater distance from the zone of definitive cells, i.e., nearer the vege- tative pole. The cells formed are at first more or less spherical and sometimes extracellular yolk remains between them. In many instances there was considerable disturbance of the distribu- tion of pigment especially at the margin of the polar cap. Frequently the polar pigment was carried down as irregular streamers over the vege- tative hemisphere leaving an irregular border between the animal and vegetative halves. In other instances the local disappearance of pigment on the surface of the animal hemisphere appeared to be due to its being carried deeper into the egg. Degenerative changes may be involved. Gastrulae It seems probable in the embryos in which segmentation is restricted to a cap of cells at the animal pole, that gastrulation could not occur. However, even in these, there is a tendency for the marginal cells of the cap to show a regional separation from the uncleaved yolk suggesting the initiation of gastrulation. In more complete blastulae, gastrulation may set in between the cleaved and uncleaved material. More often the yolk is cellular in the immediate region of the imagination. However, the evidence on the process of gastrulation is not conclusive. Externally the blastopore lip is straight and does not seem to be extended laterally very far around the circumference of the egg. In cases in which the yolk is not cleaved at the point of gastrulation the process seems to be limited to simple involution with the cells migrating to form a more or less continuous layer under the roof of the blastula. These cells are usually comparatively small and heavily pigmented, resembling those of the superficial epidermal layer. There does not seem to be any super- ficial down-growth in the region of gastrulation but laterally and an- teriorly there is considerable overgrowth of a single layer of epidermal cells over the undivided yolk. In embryos with cleaved yolk at the point of gastrulation there is usually a sac-like imagination which extends progressively into the bias- EFFECTS OF COLCHICINE ON CLEAVAGE 159 tocoele and rests on the upper surface of the undivided yolk. In such cases the blastocoele is eventually obliterated and a small archenteron is established. Later Development Many embryos die after the initiation of the modified gastrulation, probably because of their inability to carry the process to any degree of completion. However, several survived in which gastrulation ap- peared to be represented primarily by involution. In these a disc-like blastoderm was developed which consisted of a neural plate with rudi- mentary neural folds, notochord, tissue occupying the position of dorsal inesoderm and a thin, sometimes incomplete layer of underlying ento- derm (Fig. 8). A careful study of serial sections fails to yield any evi- dence of invagination. Cauclally the epidermal cells, forming a free margin, rested directly upon the uncleaved mass of yolk. The yolk-filled tissue lateral to the notochord is interpreted as being derived by the progressive cellulation of the yolk adjacent to its junction with the mar- gin of the blastula roof. The cells of the notochord and entoderm were relatively small and heavily pigmented. This, however, does not consti- tute an irrefutable argument for their origin by involution as the dis- tribution of pigment is very irregular in these abnormal embryos. Fur- thermore, even in normal embryos the cells of this region contain con- siderable pigment, possibly related to their metabolic activity and more rapid multiplication. Usually the segmentation of the animal hemisphere occurs uniformly and the partial embryos tend to be symmetrical. In some cases, how- ever, the restriction of cleavage is unequal and asymmetrical embryos are produced. Such an embryo is illustrated in Fig. 5. It appears that in this case the first cleavage persisted and then the cell division of the region of the animal pole on one side was restricted more than on the other. Embryos do not survive beyond the neural fold stage. Nuclear Size In about half of the animal hemisphere embryos cells were found that were much larger than normal. Those generally occurred in groups and might be either epidermal or deeper-lying cells. They had corre- spondingly large nuclei (Fig. 9). Mitoses were observed but they were seldom favorable for chromosome counts due to the dense concentration of pigment granules about the spindle figure. Polar views adequately demonstrated that the number of chromosomes in the large cells was greater than normal but the exact ratio could not be established. It would seem quite probable that these represent tetraploid conditions. In 160 D. M. KEPPEL AND ALDEN B. DAWSON several embryos, one in particular, sectioned at late cleavage, the cells were of varied sizes, and the mitoses were very irregular. Eccentric, distorted monasters often at the sides or corners of the cells were fre- quent. Such conditions probably explain the variation in nuclear size. In some cases, especially in eggs which had developed at 8° C., seg- mentation was restricted to a few superficial cells. In these, the ad- jacent area was characterized by regularly spaced, small centrospheres with associated asters and surrounded by pigment granules. Many of these were not associated with chromosomes or nuclear material and have been interpreted as cytasters. Yolk and Free Yolk-nuclei In the uncleaved portions of the egg the yolk underwent regular changes which were probably degenerative in character. The yolk plate- lets disappeared locally and large vacuoles were formed as the yolk products went into solution. Progressive coalescence of vacuoles pro- duced large fluid-filled spaces which were distributed irregularly through the non-cellular portion of the egg (Fig. 7). The free nuclei, especially in the areas more distant from the cleaved portion of the animal pole, became progressively larger and more ir- regular. In the earlier cleavage stages they appeared to undergo normal mitosis but in the blastula and later stages the mitoses became progres- sively abnormal and large, irregularly-lobed nuclei appeared. These nuclear complexes were usually surrounded by dense aggregations of pigment granules. Some nuclei are vacuolated and stain very lightly while others are massive and stain densely. These giant nuclei, however, are not characteristic only of colchicine-treated embryos. They were also observed by Hertwig (1897) in the meroblastic segmentation of the frog's egg under the influence of centrifugal force. Summary The treatment of the developing frog's egg with colchicine (concen- trations from 1 : 10,000 to 1 : 100,000) in amounts sufficient to retard mitosis or cause a temporary arrest of the process, results in a varying proportion of the eggs exhibiting a meroblastic type of cleavage. The degree of restriction of cleavage to the animal-pole region varies from the formation of a small cap of cells to a condition in which cleavage may include almost the entire egg. The level of development and completeness of the embryo are directly correlated with the extent of restriction of cleavage. Such animal-hemisphere embryos may not be capable of attaining even a modified blastula stage. In many blastulae EFFECTS OF COLCHICINE ON CLEAVAGE 161 only the roof of the segmentation cavity is cellular. The process of gastrulation is always greatly disturbed and in extreme cases may be limited to a modified involution. Frequently, neural plate with begin- ning neural folds, notochord, dorsal mesoderm and entoderm are dif- ferentiated. Embryos rarely develop beyond this level. In many embryos there are scattered areas in which the cells are large and possess correspondingly large nuclei. These are tentatively interpreted as the result of tetraploidy, a phenomenon commonly pro- duced in plants by colchicine. The free yolk-nuclei are characteristically abnormal. LITERATURE CITED BLAKESLEE, A. F., AND A. G. AVERY, 1937. Methods of inducing doubling of chromosomes in plants by treatment with colchicine. Jour, Hered., 28 : 393. DUSTIN, A. P., 1934. La radiotherapie et la chimiotherapie envisagees a la lumiere des travaux sur les chocs caryoclasiques. Jour, des Sti. Med. Lille, No. 49, p. 561. DUSTIN, A. P., 1935. L'action de la colchicine sur les tumeurs malignes. Leeuw- enhoek Vereeniging-IVieme Conference, Amsterdam, p. 7. HERTWIG, O., 1897. Uber einige am befruchteten Froschei durch Centrifugalkraft hervorgerufene Mechanomorphosen. Sitsungsber. konig. Preuss. Akad. Wiss., Halbband 1, S. 14. HERTWIG, O., 1898. Ueber den Einfluss der Temperatur auf die Entwicklung von Rana fusca und esculenta. Arch. mikr. Anat., 51 : 319. LITS, F. J., 1936. Recherches sur les reactions et lesions cellulaires provoquees par la colchicine. Arch. Internat. Medecine Experim., 11: 811. NEBEL, B. R., AND M. L. RUTTLE, 1938. The cytological and genetical significance of colchicine. Jour. Hered., 29 : 3. RUGH, R., 1934. Induced ovulation and artificial fertilization in the frog. Biol. Bull, 66 : 22. EFFECTS OF 2, 4-DINITROPHENOL ON THE EARLY DEVELOPMENT OF THE TELEOST, ORYZIAS LATIPES A. J. WATERMAN 1 (From the Thompson Biology Laboratory, Williams College, Williamstown, Mass.) The stimulating action of 2, 4-dinitrophenol on metabolism and oxygen consumption has been demonstrated for many organisms both embryonic and adult (Shoup and Kimler, 1934, luminous bacteria; Bodine and Boell, 1938, grasshopper embryos; Root and Etkin, 1937, toadfish; and others). It raises the metabolic rate of female rats (Halpern and Hendryson, 1935) but does not accelerate development in amphibian embryos (Cutting and Tainter, 1933; Dawson, 1938; Buchanan, 1938). In sea urchin eggs this and other nitro- and halo- phenols simultaneously increase respiration and block cell division (Krahl and Clowes, 1938). The physiological action is discussed by Bodine and Boell (1938). In this investigation a study has been made of the effects produced by 2, 4-dinitrophenol on the development of the egg of the teleost, Orysias latipes, (a) when continued exposure is begun at late cleavage to optic vesicle stages, (b) when the chorion is and is not pricked, and (c) when the egg is exposed to a lethal concentration for various periods of time. This substance is highly toxic. In Rana pipiens (Dawson, 1938) it either retards, produces abnormalities or completely arrests develop- ment depending chiefly on the concentration. Among the abnormalities produced are: persistent blastulae, arrested early gastrulae, persistent yolk plugs and exogastrulae. These results are somewhat comparable to what others have obtained in amphibia with high temperature, thy- roxine, x-ray exposure, and dilute Ringer's solution. Buchanan (1938) found that such respiratory effects as this substance may have do not compensate for the retarding effect of low temperature on the develop- ment of Amblystoma. The effects on gastrulation are especially in- teresting because of the numerous studies of exogastrulation which have been made in both amphibia and echinoderms. 1 This work was aided by a grant from the Rockefeller Foundation. 162 DINITROPHENOL AND FISH DEVELOPMENT 163 MATERIALS AND METHODS Orysias latipes (Aplocheilus latipes) is a small oviparous cyprino- dont teleost, measuring about one inch in length, which is also known as the Japanese Medaka. This fish breeds well in the balanced aquar- ium and can stand low temperatures. Its small, colorless, transparent eggs are laid almost daily over long periods and remain attached to the cloacal region of the female until they are brushed off on the aquarium vegetation. The attachment occurs by the numerous processes peculiar to the egg chorion. Females lay from two or three to as many as twenty-five or more eggs at one time. If the eggs are regularly re- moved, they will continue to lay for several months. Otherwise the presence of the eggs appears to retard and later to inhibit egg produc- tion. Development is rapid and hatching occurs about fourteen to eighteen days after laying. Several factors may be responsible for this variation in hatching time, i.e. temperature possibly, inherent variation in developmental rate of embryos from the same and different females, and the time the eggs are laid in the reproductive period. At the end of the first day after laying, the eggs are in a late cleavage stage. The optic vesicle stage is reached during the second day and on the third to fourth days the heart is beating and differentiation is well advanced. For the present study a 1 per cent stock solution of 2, 4-dinitrophenol was made according to the method of Halpern and Hendryson (1935), and sufficient amounts added to spring water to make the desired solu- tions. Concentrations of 1 : 10,000 up to 1 : 1,000,000 were used, but the most significant results appeared at concentrations from 1 : 40,000 up to 1 : 200,000. For the most part the embryos remained in the solu- tions for the entire experimental period. Eggs were also placed in a known lethal concentration for the late cleavage stage (1 : 10,000) and samples removed at intervals thereafter to fresh water. Since the egg is surrounded by a chorion, several studies were made in which the chorion was pricked. All experiments were made at room temperatures. Exposure was begun at four developmental stages ; early and late cleav- age, closure of the blastopore, and optic vesicle. EXPERIMENTAL In general the most conspicuous effect upon development is a re- tardation in both growth and differentiation which is proportional to the length of exposure and to the concentration of -the drug, but there is variation in different individuals of a culture as some are more sus- ceptible than others. No evidence of any stimulation was seen either in the developmental rate or in the hatching time nor were any cases 164 A. J. WATERMAN of exogastrulation obtained. The results obtained in one experiment are shown in Table I which is a typical example of the many that were TABLE I Effect of 2, 4-dinitrophenol on early development of Orysias latipcs (Late cleavage stage) Average Tem- Experi- ment Concen- tration Age in days Number dead heart -beat (No. of sees, for Heart- beat (low and high) perature when exam- Comments Hatching time 50 beats) ined 21 Control 7 None 28.7 24 25° C. 1 in 18 days 34 1 in 21 days 5 in 25 days 1 : 40,000 7 None in 8 35.2 33 25° C. Markedly re- All dead 13 42 tarded. Few days after have no circula- laying. tion and rudi- mentary heart. 1 : 80,000 7 1 in 14 31.2 27.5 25° C. 3 retarded. 2 dead in 22 34.5 others like con- days, all dead trol. in 25 days, none hatched. 1 : 120,000 7 None in 15 29 26 25° C. 6 retarded 2 in 19 days 32 2 in 2 1 days 1 in 22 days 2 in 25 days All died soon after hatching. 1 : 160,000 7 Nonejn 1^ 28.6 26.5 25° C. No retardation 1 in 18 days 31 8 in 19 days 2 in 2 1 days 1 in 25 days 2 died 1 : 200,000 7 1 in 7 28.4 27 25° C. No retardation 1 in 18 days 30 5 in 25 days tried. Concentrations greater than 1 : 10,000 inhibit all further develop- ment beyond possibly a few cleavages, while in a solution of 1 : 160,000 development is normal and the embryos hatch about the same time as the control. Although development occurs in concentrations greater than 1 : 90,000, it is very retarded, causing early deaths and no hatching. The effect of the drug is cumulative with time and retardation becomes increasingly evident with continued exposure to strong concentrations. These observations confirm those of others on the high degree of tox- icity of this substance. Except for the difference in the effective concentrations and the formation of certain abnormalities, the effects of 2, 4-dinitrophenol on the development of this fish closely duplicate those found by Dawson on the frog, and by Solberg (1938) for Fitudulus embryos exposed to X-rays. The survival of the embryos is directly correlated with their DINITROPHENOL AND FISH DEVELOPMENT 165 sensitivity to the drug. It is generally difficult to duplicate the results of one experiment with eggs taken from the same or different females on successive days. Another comparable observation is that temporary developmental inhibition does not necessarily cause irreversible injury, providing the exposure is not too long. Both the controls and the experimental animals were run at the same temperature. This was always that of the laboratory and it varied between 24° and 27° C., with an average around 25°. This slight vari- ation probably did not influence the action of the drug since Dawson found that in the frog the results were apparently not greatly modified by temperature. However, a large proportion of his animals showed arrested and abnormal development at 6° to 12° C. This is far below the temperature range of the present study. In Bufo, Buchanan found that the toxicity is higher at 21° than at 6° C. The heart and extra embryonic vascular system are particularly sen- sitive. Since the rate of heart beat corresponds roughly to the con- centration and length of exposure, it serves as a check on the other effects observed. However, no stimulation was observed. Concentra- tions greater than 1 : 130,000 progressively slow up the rate and ampli- tude of the heart beat and also provoke structural irregularities. De- formities of the heart lead to interrupted and poor circulation which in turn provokes other abnormalities as development progresses. These embryos die early. In concentrations greater than 1 : 40,000 the extra- embryonic circulation never becomes established and the heart is rudi- mentary although other embryonic differentiations may take place, i.e. nervous system, sense organs, body form, and sometimes somites. Blood corpuscles develop in all embryos but the quantity is less in the more retarded types. Blood vessels are also affected as, in the extra- embryonic area especially, some may be distended with blood while others remain unconnected. Deformities in the myotomes and a reduction of their number are of common occurrence. The number may especially be reduced in the tail. Here the irregularities bring about shortening and curious shapes and positions of the tail. It often becomes bent at odd angles, twisted, curved or shortened. The tail fins may fail to develop but typical paired appendages are usually present in less abnormal cases. Embryos of these types did not hatch. The nervous system and sense organs exhibit a variable response to the drug. In the stronger solutions some development stops at a neural tube stage while other embryos show the presence of the optic vesicle with but slight indications of the brain. In others the nerve cord is shortened or deformed. Often the neural canal is filled with a mass 166 A. J. WATERMAN of cells. Different regions of the brain are often missing, enlarged or abnormal on one side or the other. Very little effect is observed on the ears. The optic vesicles often fail to invaginate or show a variable amount of pigmentation. A size difference is also noticed as well as a malformation of the lens. No particular type of teratological embryo is produced by dini- trophenol, but the production of uniform abnormalities is fairly con- sistent. Gastrulation either occurs normally, is retarded, or is entirely inhibited. If the exposure is not too long a few of the retarded embryos TABLE II Sensitivity of different developmental stages to a lethal solution of 2, 4-dinitrophenol (1:10,000) Stage Length of exposure before transfer to fresh water Development during exposure 4 hrs. 10 hrs. 21 hrs. 29 hrs. 45 hrs. Early cleavage Retarded in size and de- velopment. Small optic vesicle or rudimentary nerve tube. Some dead. Dead None Late cleavage Retarded in size and de- velopment. Small optic vesicle stage. Small optic vesicle stage. Dead None Closure of blastopore Slight re- tardation in some. Smaller and thinner em- bryos. Slight pigmentation of eyes. No develop- ment beyond optic vesicle or young non-pig- mented optic cup stage. Rudimentary heart. No circulation. No develop- ment beyond young optic vesicle stage. Dead None. Perhaps slight in- crease in cell number. Optic vesicle No effect Slight re- tardation. Slower heart rate. Marked re- tardation. No heart ac- tion. 96 hrs. exposure did not inhibit development. Yes, but at a slower rate. The above results are those seen after the embryos had been in fresh water for 3-4 days following exposure. are able to complete gastrulation and to continue development in fresh water. Otherwise they give cases of persistent yolk plug which also occur in some instances during the experiment. Subsequent develop- ment is very atypical and death soon ensues. The inhibited blastulae do not increase in size and development goes no farther except for a few irregular cleavages especially noticeable at the periphery. None of these are able to gastrulate in fresh water. Unlike the frog, no living cases of exogastrulation were seen. DINITROPHENOL AND FISH DEVELOPMENT 167 The results of the study of the relative sensitivity of several develop- mental stages to dinitrophenol are summarized in Table II. They show that the sensitivity decreases at least from the early cleavage stage to the optic vesicle stage and in the latter case development is able to con- tinue, but more slowly, in a lethal solution which quickly kills the earliest stage tested. DISCUSSION The resistance of fish embryos to the experimental conditions de- scribed in this study varies at different stages. At the optic vesicle stage they are much less sensitive to 2, 4-dinitrophenol than at the early cleavage stage. Solberg (1938) has likewise observed this to be true of Fundnlus embryos exposed to X-rays. The earlier in development that the embryos are treated, the greater are the deformities. During the gastrula stages there is again a slight increase but thereafter the sen- sitivity declines rapidly and larger doses are required to modify de- velopment. Changes also occur in the sensitivity of different organs of the fish embryo to X-rays and dinitrophenol. As differentiation progresses the organs become more resistant ; the time depends on the time of appearance and the rate of differentiation. There was no special difference noted in the adjustment of early and late cleavage, or even older stages to the experimental solutions. Except for the question of sensitivity, early and late stages gave almost comparable effects. Dawson (1938) found this the case in his study of the effects of dinitrophenol on the development of Rana piplens when treatment was begun at the 2-cell and early blastula stages. However, under the influence of high temperature, Hoadley (1938) found that the behavior of frog eggs is different immediately after the first cleavage from what it is at the 128-cell stage. There was a more complete ad- justment to the new conditions on the part of the first group. The suppression of cell division by dinitrophenol appears to be a common phenomenon. First reported (Krahl and Clowes, 1938) for sea urchin eggs, it has subsequently been found to occur in the frog (Dawson, 1938) and in the present study on the fish. At first this effect is hardly noticeable but with continued exposure mitosis is almost sup- pressed. The rate of this disturbance depends on the concentration of the drug. At one extreme development may be stopped at such an early stage as the optic vesicle, while at the other quite typical embryos may develop. However, the latter are noticeably smaller than the con- trol, die sooner, and never hatch. Solberg (1938) finds that following X-ray exposure, the mitotic index of Fundulus embryos changes which accounts, at least in part, for the changes in sensitivity. The close 168 A. J. WATERMAN parallelism of the two changes during early development seems to him to be more than coincidental. The chorionic membrane of the Oryzias egg is permeable to dini- trophenol but some difference is noted in the effective concentrations between cases of whole membranes and those in which the chorion is pricked. Even further differences might be noted if this membrane were entirely removed. The drug evidently penetrates slowly and higher concentrations and longer exposures are necessary for the intact mem- brane. This may account for the difference in the effective concentra- tions reported for the frog and Bufo. In vertebrates, studies of exogastrulation have been limited thus far to the amphibia. The period of gastrulation is a critical one. Hoadley (1938) has shown that under the influence of supra-maximum tempera- tures early gastrulae and very late blastulae are unable to complete gastrulation and die. Early blastulae form giant blastulae which never undergo gastrulation. Many other environmental conditions are known to affect or modify amphibian gastrulation (cf. Dawson, 1938) ; dilute Ringers solution, X-ray exposure, thyroxin. Dawson obtained per- sistent blastulae, arrested early gastrulae, persistent yolk plugs and exo- gastrulae through the use of 2, 4-dinitrophenol. These results show that exogastrulation in amphibia is no more of a specific reaction to any particular agent than has been found to be the case among echino- derms (Child, 1936). They have all followed instances of develop- mental disturbance and arrest. In the present study no cases of living exogastrulae were seen. Either gastrulation was inhibited entirely and the persistent blastulae remained as such until cytolysis and death occurred or gastrulation was retarded or stopped during the process. Cases of persistent yolk plug occurred occasionally. In all these interrupted types, the origin and the differentiation of the organ rudiments and even body formation were restricted. There was no indication of a differential effect on gastrulation alone. It would have been interesting to see \vhat form exogastrulation would take in a telolecithal egg. Certain types of abnormalities occurring in the embryos under the influence of 2, 4-dinitrophenol have been reproduced in fish by other agents and in the amphibia by various agents. For example, in Fuii- dulus heteroclitus Solberg (1938) obtained comparable effects with X-rays on the development of the circulatory system, myotomes, nervous system, etc. Reference should be made to Dawson (1938) for work on the frog embryo. In general no special type of teratological embryo was produced but results were fairly consistent. As Dawson says, none of the effects obtained by treatment with dinitrophenol can be inter- DINITROPHENOL AND FISH DEVELOPMENT 169 preted as being produced specifically by the drug but are such as might follow any treatment which disturbs the orderly processes of develop- ment and differentiation. If the increase in metabolic rate is one of the factors involved it is difficult to correlate this idea with the results secured by X-rays and by other agents tested on amphibia. From the present study it would appear that 2, 4-dinitrophenol does not stimulate gastrulation or development in Oryzias latipes and, in the effective concentrations employed, it acts like any toxic agent. As such it retards, inhibits and disturbs normal orderly developmental processes in proportion to the concentration and length of exposure. Its effects are cumulative with time. Various concentrations which did not pro- duce any observable defects caused no reduction in the hatching time. This confirms in the fish the observations of others that the drug does not accelerate development or metamorphosis in amphibian embryos (Cutting and Tainter, 1933; Dawson, 1938; Buchanan, 1938). Dawson suggests that the increase in the metabolic rate may be one of the factors involved in the effects produced by 2, 4-dinitrophenol on the frog embryo since they resemble those caused by thyroxin (Bau- mann) and high temperature. According to Bodine and Boell (1938) the toxic effect may be due to the increased accumulation of toxic metabolic products of which the organism is unable to rid itself. In this connection it is interesting to note the conclusion of Lindahl (1936) from his study of the effect of SO4-deficient sea water on the develop- ment of Paracentrotus lividus (the European sea urchin), since re- sults were secured which in some cases are comparable to those of the frog and fish. SO4-deficient sea water produces persistent blastulae, exogastrulae, ectoblastulae with cilia, etc. He considers that hydro- carbon metabolism dominates at the animal pole while protein metabolism dominates at the vegetal pole. The toxic action of this deficiency is caused by the accumulation of toxic phenolic derivatives coming from protein metabolism. SUMMARY A study has been made of the effect of 2, 4-dinitrophenol on the development of four embryonic stages of the teleost, Oryzias latipes, i.e. early and late cleavage, closure of the blastopore and optic vesicle stages. Both the sensitivity of the whole organism and constituent parts to the drug decreases with age and differentiation. The general effect is cumulative and proportional to the concentration and length of exposure. In addition to inhibitory and retardative effects, other de- velopmental abnormalities include those of the heart and blood vessels, myotomes, nervous system, body shape, etc. 170 A. J. WATERMAN Concentrations of 1 : 10,000 up to 1 : 1,900,000 were used, but the most significant results appeared at concentrations from 1 : 40,000 to 1 : 200,000. When the chorion is pricked there is some difference in the effective concentrations. No evidence of any stimulation was seen either in the developmental rate or in the hatching time. Gastrulation was either inhibited or retarded and no examples of exogastrulation were seen. If exposure has not been too long, recovery takes place to a more or less extent in fresh water. Some of the results of the present study have been duplicated in the amphibia by various agents and in fish by X-rays. In general no special type of teratological embryo was produced but the results are fairly consistent and reproducible. It does not seem possible to at- tribute the effects to a specific action of this drug and hence to increased metabolic and respiratory rates alone, unless the change produces toxic products of which the embryo is unable to rid itself. LITERATURE CITED BODINE, J. H., AND E. J. BOELL, 1938. The influence of some dinitrophenols on respiratory metabolism during certain phases of embryonic development. Jour. Cell. Comp. Physiol., 11: 41-64. BUCHANAN, J. W., 1938. Developmental acceleration following inhibition. Jour. E.vper. Zool, 79: 109-128. CHILD, C. M., 1936. A contribution to the physiology of exogastrulation in Echinoderms. Arch. f. Entw.-mech., 135: 457-493. CUTTING, C. C., AND M. L. TAINTER, 1933. Comparative effects of dinitrophenol and thyroxin on tadpole metamorphosis. Proc. Soc. Exper. Biol. and Med., 31 : 97. DAWSON, A. B., 1938. Effects of 2, 4-dinitrophenol on the early development of the frog, Rana pipiens. Jour. Exper. Zool., 78: 101-110. HALPERN, S. R., AND I. E. HENDRYSON, 1935. Comparative effects of dinitrophenol and thyroid on pituitary-gonadal complex of female rats. Proc. Soc. Exper. Biol. and Med., 33 : 263-265. HOADLEY, L., 1938. The effect of supramaximum temperatures on the develop- ment of Rana pipiens. Growth, 2 : 25-48. KRAHL, M. E., AND G. H. A. CLOWES, 1938. Physiological effects of nitro- and halo-substituted phenols in relation to extracellular and intracellular hy- drogen ion concentration. Jour. Cell, and Comp. Physiol., 11 : 1-20. LINDAHL, P. E., 1936. Zur Kenntnis der physiologischen Grundlagen der De- termination im Seeigelkeim. Ada Zoologica, 17: 179-365. ROOT, R. W., AND W. ETKIN, 1937. Effect of thyroxine on oxygen consumption of the Toadfish. Proc. Soc. Exper. Biol. and Med., 37: 174-175. SHOUP, C. S., AND A. KIMLER, 1934. The sensitivity of the respiration of lu- minous bacteria for 2, 4-dinitrophenol. Jour. Cell and Comp. Physiol., 5 : 269-276. SOLBERG, A. N., 1938. The susceptibility of the germ cells of Oryzias latipes to x-radiation and recovery after treatment. Jour. Exper. Zool., 78 : 417-440. WATERMAN, A. J., 1937. Effect of salts of heavy metals on development of the sea urchin, Arbacia punctulata. Biol. Bull, 73: 401-420. EFFECT OF TEMPERATURE UPON SHELL MOVEMENTS OF CLAMS, VENUS MERCENARIA (L.) l VICTOR L. LOOSANOFF (From the U. S. Biological Laboratory, Mil ford, Connecticut) INTRODUCTION Knowledge of the effect of temperature on the physiological activi- ties of mollusks is essential to the solution of many problems of shell- fisheries. Galtsoff (1928), in his studies of the effect of temperature on oysters, demonstrated its bearing on the problems of oyster culture and sanitary control of the oyster industry. His studies and those of Nelson (1921) and Hopkins (1931) added greatly to our knowledge of this phenomenon in relation to oysters, but little or nothing is known about other Pelecypoda. This study deals with the effect of tempera- ture upon the shell movements of the hard-shell clam (Venus mer- cenaria L.). In Pelecypoda, respiration and feeding are influenced by three factors, namely, the rate of activity of cilia of the gill epithelium, changes in size of the ostia, and by the movements of the shell-valves. It has already been shown (Galtsoff, 1928) that low temperatures induce hibernation in oysters. During the hibernation period the animal is unable to feed because of the disturbances of the gill mechanism, and because of closure of shells. It is of interest, therefore, to determine the temperature at which an animal enters into the hibernating stage and whether or not there is a definite correlation between the shell move- ments of a mollusk and the temperature of the surrounding water. Experiments on the effect of temperature upon shell movements of the clam (V . mercenaries) were begun on November 17, 1934, and con- tinued until July 29, 1935. They were conducted in a large outdoor concrete tank, 10 by 20 feet, at the laboratory of the U. S. Bureau of Fisheries in Milford, Conn. The tank was so constructed that the water entered it only at the last third of each flood tide and flowed out only partially at the ebb. Thus, the water in the tank was partly renewed twice every twenty-four hours. At low water stages the tank retained approximately 4,000 gallons of water. Because of this arrangement rapid changes in the temperature and salinity of water in the tank were rather uncommon. A roof over the tank protected it from the hot sun 1 Published by permission of the U. S. Commissioner of Fisheries. 171 172 VICTOR L. LOOSANOFF and heavy rainfalls, facilitating still further the maintenance of a com- paratively even temperature and salinity. METHODS Altogether 399 records from 47 different clams were obtained, each representing the movements of a clam's shell-valves during a 24-hour period. Usually the records of two clams were taken simultaneously. Clams were attached to a recording apparatus for varying periods of time ranging from one to eighteen days. During the course of this work the water temperature in the experimental tank varied from - - 1.0° -f -9 — -r B FIG. 1. Apparatus employed in studies of shell movements of clams. Description in text. to 28.0° C., thus covering almost the entire range of temperature to which clams are subjected under natural conditions. The temperature of the water was recorded continuously by Brown's recording ther- mometer which was checked frequently against thermometers certified by the U. S. Bureau of Standards. Each clam was immobilized by imbedding one of its shell-valves in a mixture of sand and cement. The animals were connected to the lever of a recording apparatus in such a way that each movement of the shell was recorded on a chart. The apparatus (Fig. 1) consisted of two main TEMPERATURE AND SHELL MOVEMENTS OF CLAMS 173 parts, a foundation for keeping an experimental animal in a desired position (A), and a Foxboro recorder (5). An animal (/) imbedded in a block (g) of sand-cement mixture was placed on a large round concrete base (r) and connected to a recording instrument by a rod (d) and a chain (a) which was attached to a hook (o) of the recording apparatus. A flat piece of cement with a hole in the center was sealed to the free shell-valve of a clam, and on it rested a fine metal pin (e). By moving the nut (i) along the bent rod (h) the weight of a struc- ture connecting the animal to the recording apparatus could be very accurately counterbalanced. This arrangement made it possible to elim- inate any unnecessary pressure which would be exerted upon the clam. The rod (/) was connected to a writing pen (£) which recorded each movement of a clam shell upon the chart. By moving the hook (o) along the rod (/) the record of valve movement on a chart could be increased or decreased, as desired. In these experiments the distance between the experimental animal and the recorder was about 10 feet. The depth of the water over the clams varied from 4 to 7 feet depending upon the stage of tide. While the time recorder has already been employed by other inves- tigators in studying the shell movements of mollusks (Galtsoff, 1928), the method of holding the animal in a desired position and attaching it to the recorder, as described in this paper, is a new one. The main advantage of using the apparatus described here, in pref- erence to those employed by other investigators, is that it considerably simplifies the experimental work by rendering the handling and replace- ment of animals very easy. By pulling up the cord tied to the ring (b) of the foundation unit, the entire lower portion of the apparatus can be quickly raised from the bottom and the animal examined, or new animals substituted for the old ones. All manipulations in placing the animal in position and connecting it to the recording machine are per- formed out of water. The use of the chain (a) eliminates long rods, which were commonly used by other workers. At each change of chart on the recording apparatus the experimental animal was tapped lightly to compel it to close its valves tightly. After this was achieved, a base line («) was drawn on the chart by rotating it 360°. Having a distinct base line, every movement of a shell, no matter how slight, could be distinguished on the chart (m). RESULTS Table I and Fig. 2 show the percentage of time the clams remained open at different temperatures. While subjected to very low tempera- tures ( — 1.0° to + 1.9° C.) all the clams remained completely closed. 174 VICTOR L. LOOSANOFF Altogether 56 records were obtained at such temperatures. At tem- peratures ranging from 2.0° to 2.9° C., with the exception of one case when an animal was open for about 5 hours, all the clams were inactive. A very slight increase in clam activities was noticed at temperatures ranging from 3.0° to 4.9° C. At the latter temperature the majority of experimental animals remained completely closed, but a few opened on several occasions, bringing the average of open time to 9 per cent. A sudden increase in clam activities was noted at 5.0° to 5.9° C., when the percentage of time open increased to 35 per cent. Such a pro- nounced increase in shell activities indicates that for many clams the critical hibernating temperature lies somewhere between 5.0° and 6.0° C. TABLE I Number of 24-hour records of shell activities obtained at each temperature ranging from —1.0° to 28.0° C., per cent of total time, and average time shells re- mained open at each temperature during 24-hour period. Temperature No. of records Per- centage of total time shells opened Average time shells opened during 24 hours Temperature No. of records Per- centage of total time shells opened Average time shells opened during 24 hours °C. hours minutes ° C. hours minutes -1.0- 0.1.. 18 — — 14.0-14.9 8 88 21 7 0.0- 0.9. . 8 — — 15.0-15.9 16 81 19 26 1.0- 1.9.. 30 — — 16.0-16.9 6 78 18 43 2.0- 2.9.. 26 1 0 14 17.0-17.9 12 84 20 10 3.0- 3.9.. 30 4 0 58 18.0-18.9 6 89 21 22 4.0- 4.9.. 32 9 2 10 19.0-19.9 4 76 18 14 5.0- 5.9.. 22 35 8 24 20.0-20.9 12 86 20 38 6.0- 6.9.. 6 29 6 58 21.0-21.9 6 90 21 36 7.0- 7.9.. 12 65 15 36 22.0-22.9 10 86 20 38 8.0- 8.9.. 22 67 16 5 23.0-23.9 4 78 18 43 9.0- 9.9.. 22 74 17 46 24.0-24.9 16 69 16 34 10.0-10.9.. 4 88 21 7 25.0-25.9 9 71 17 2 11.0-11.9.. 8 85 20 24 26.0-26.9 6 83 19 55 12.0-12.9.. 26 84 20 10 27.0-27.9 4 89 21 22 1 "? 0—1 3 0 14. 71 1 7 9 i\J.\J 1O..7. . Irt / 1 i / £ Total 399 At 7.0° to 7.9° C. all the clams were open some of the time, bringing the average time open to 65 per cent. From 8.0° to 10.9° C. a further increase in shell activities was shown. At 10.9° C. the animals remained open 88 per cent of the total time. As the water warmed from 11.0° to 27.9° C. no definite correlation between the temperature of the water and the length of time the shells remained open could be detected. Between these two points, the percentage of open time fluctuated be- tween 69 and 90. TEMPERATURE AND SHELL MOVEMENTS OF CLAMS 175 The individual differences displayed by the clams used in these experiments were quite significant. For example, while some of the animals exposed to temperatures of 5.0° to 5.9° C. remained open for the entire period of twenty-four hours, others did not open at all. At low temperatures the clams may remain completely closed for several days. For example, one clam had its shell closed from noon on December 16, 1934, until noon on January 3, 1935, a period of 18 days. The temperature of the surrounding water during that period fluctuated from — 1.0° to 6.5° C. Another clam remained completely 100 90 § 80 CL, O 70 w S 60 S 50 i 40 U 30 S 20 10 I I i i i r T i r T i r I -20246 8 10 12 14 16 18 20 22 24 26 28 TEMPERATURE °C FIG. 2. Percentage of time clams remained open at different temperatures during 24-hour periods. closed from March 8 to March 19, 1935. In several other instances the animals' shells remained closed for various periods extending from 3 to 6 days. In every one of these cases, the record of shell activity showed no deviation from the base line, thus indicating that not even a slight movement of the shells took place during such a prolonged period. Somewhat similar observations were made by Galtsoff (1928) on one specimen of 0. virginica which remained tightly closed for 67 hours when left in cold water of a temperature varying from 0.5° to 1.6° C. Galtsoff (1928), in his work on O. virginica, could not find a definite correlation between the effect of temperature and the time the shells of 176 VICTOR L. LOOSANOFF oysters remained open. This was due, probably, to the fact that in his experiments the temperature of the water varied within the compara- tively narrow range of 13.0° to 22.0° C. No systematic observations were performed at low temperatures. Similar conclusions could have been reached in this work if the observations had been confined to higher temperatures only. As has been mentioned already, there was no correlation found between the time the clam shells were open and the temperature increase from 11.0° to 27.9° C. If, however, the shell movements are examined at temperatures ranging from 0.0° to 11.0° C., the correlation between the rise in temperature and the gradual increase in duration of openness of the valves is noticed. This is especially evident within the range of 3.9° to 10.9° C., in which the average period of openness increases from 4 to 88 per cent (Fig. 2). Such an increase indicates that as soon as the water temperature gradually rises from the hibernation point to 11.0° C., the factors or conditions controlling the opening of the shells approach the optimum. The present study, however, does not provide definite information as to the exact temperature at which the optimum conditions for open- ing of clam shells are reached. Within a 11.0° to 27.9° C. temperature range the shells are open from 69 to 90 per cent of the total time, but the time the shells of the animals remain open does not increase pro- portionally to the increase in temperature. The highest percentage of time open is recorded at temperatures 21.0° to 22.0° C. when the clams remained open 90 per cent of total time, or 21 hours and 36 minutes per 24-hour period. The average period of time the shells remained open at temperatures from 11.0° to 27.9° C. is 19 hours and 35 minutes. This figure closely resembles those obtained by some other investigators in their studies of oysters. Nelson (1921), basing his conclusions on the records of 3 oysters (O. virginica) kept under observation for 21 days, states that the animals remained open an average of 20 hours per day. Galtsoff (1928), from more numerous observations, concluded that the average period of time the shells of oysters remained open is 17 hours and 7 minutes per day. Hopkins (1931), working on O. lurida, found that the oysters were open and presumably feeding over 20 hours per day. From the study of the movements of clam shells and from observations of other investigators on oyster shell movements, it appears that under favorable conditions these pelecypods keep their shells open as long as possible. Experimenting with 0. lurida, Hopkins (1931) came to the conclu- sion that it is not so much the existing temperature of the water which determines how long the shells remain open as it is the changes in temperature which occur. This sensitivity of oysters to temperature TEMPERATURE AND SHELL MOVEMENTS OF CLAMS 177 changes varies in an inverse manner with the temperature of the water. Thus, a small drop in temperature causes closure of the shell if the temperature is well below the optimum, but produces no effect if near the optimum. In clams no such effect of sudden changes in temperature could be detected. For instance, on December 2, 1934, between 2:00 and 3:00 P.M., the water temperature in the experimental tank rose A.M. 12 2 4 6 TIME OF DAY P. M. 8 10 12 2 4 6 8 10 12 CLAM NO. 6 12 3 T*C DECEMBER 2, 1934 T°C 14 10 0 DECEMBER 20,1934 OCTOBER 22, 1934 SHELLS CLOSED OPENED H- HIGH WATER STAGE L - LOW FIG. 3. Records of shell activities of clams and temperature of surrounding water during 24-hour periods. A sudden change in temperature (December 2 and 20, 1934) failed to cause the specimens to exhibit rapid closing or opening of shells. Shells closed and opened when temperature remained unchanged (October 22, 1934). No correlation between the tidal stages and shell movements could be observed. H — high-water stage, L — low-water stage. from 9.5° to 13.5° C, and two hours later decreased to 9.5° C. (Fig. 3). Both of the experimental animals, whose shells were open, remained ap- parently undisturbed. In another case, on December 20, 1934, the water temperature in the experimental tank increased from 3.0° to 6.5° C. within two hours and then quickly dropped to 2.0° C. Neither of the two experimental animals showed any noticeable activities. 178 VICTOR L. LOOSANOFF It appears that in Venus mercenaries the mechanism controlling the opening and closing of the shells is less sensitive to minor temperature changes than that of the oyster. In clams, the closing or opening of the shells is very often performed when the temperature of the sur- rounding water remains at the same point for some time. On the other hand, the closing or opening often occurred at decreasing temperatures, while again, in numerous other instances, these phenomena took place at increasing temperatures (Fig. 3, October 22, 1934). As far as can be judged by examining hundreds of the records obtained in this work, small changes in the temperature of the surrounding water do not di- rectly influence the shell movements of clams. During periods of openness the adductor muscles of the clam do not remain in the same position but relax and contract, indicating the changes in tonus level. Contraction of the muscles, although quite pronounced, seldom results in the complete closing of the shells. The data ob- tained help to answer the question whether there is a definite peri- odicity in contraction and relaxation of clam muscles, and, if so, whether such periodicity is affected by exposing the animal to different tempera- tures. Figure 4 represents the records of the shell movements of two experimental animals, Nos. 14 and 16, exposed to different temperatures ranging from about 0.0° to about 20.0° C. and having intervals of about 4.0°. The temperature of the water during these experiments fluctuated not more than ± 1.0° C. Each record represents the activities of the clam for a period of 12 hours, from 6:00 P.M. until 6:00 A.M. The night period is chosen in preference to daytime because the changing of charts in the recording apparatus, usually made in daytime, slightly dis- turbed the animals. By taking the night half of a daily record, more reliable information is available. The muscles controlling the shells of clams relaxed and contracted, sometimes at very brief, and at other times at prolonged intervals (Fig. 4). It is difficult, however, to find a definite periodicity throughout any of the 12-hour periods shown in Fig. 4. Al- though during several brief intervals the movements of the shells were of a definitely periodic type and exhibited a rhythmic character, more often they occurred as unsystematic and inconsistent. In this respect the shell activities resemble those of oysters, in which, according to Hop- kins (1936), partial closures of the shells are sometimes periodic, some- times only occasional and unorganized. Clams whose shell activities are graphically shown in Fig. 4 present an opportunity of comparing the behavior of two individuals subjected to identical environmental conditions. With the exception of the records obtained at the temperature of 16.0° C., all other records of the shell activities of these two clams were obtained simultaneously. At the tern- TEMPERATURE AND SHELL MOVEMENTS OF CLAMS 179 FIG. 4. Records of shell movements of two clams Nos. 14 and 16, exposed for 12-hour periods to different temperatures ranging from 0.0° to 20.0° C. Base line is indicated by series of dots. H — high-water stage, L — low-water stage. 180 VICTOR L. LOOSANOFF perature of 0.0° C. both animals remained closed, thus exhibiting uni- form behavior. At 4.0° C. one clam remained closed whereas the other was open part of the time. At higher temperatures both clams were active. The records show that there were considerable individual dif- ferences in the behavior of these animals although they were subjected to the same external environmental conditions. Such variability may probably be explained as due to internal causes. In these experiments no definite correlation between the stage of the tide and shell movements of the experimntal animals could be detected. This is demonstrated in Figs. 3 and 4 which show activities of clams at different tidal stages. The failure of the experimental animals to react TABLE II Percentage of time the shells of clams remained closed at daytime and night when exposed to different temperatures. Temperature classes Number of records Number of clams Average percentage of daylight closed Average percentage of night closed 0.0- 4.9° C 7 3 67.0 74.0 5.0- 9.9°C 24 6 36.0 8.0 10.0-14.9° C 27 5 20.0 9.0 15 0-19 9° C . 40 6 18.0 23.0 20 0-24.9° C. . . 46 9 23.0 18.0 25.0-28.0° C 19 5 23.0 20.0 in a definite way to tidal changes is attributed to the fact that the tanks, where the animals were kept, are so constructed that they are filled at flood tide only and retain part of this water during the ebb stage. Therefore, fluctuations in temperature and salinity of water in the tank were not as great as those in the adjoining harbor. The material on hand permits the answer to the question whether there is a correlation between opening and closing of clam shells and light and darkness. Nelson (1921) suggested that such correlation does exist for oysters. His conclusions were based upon oysters lying on a natural oyster reef subjected to the usual changes in temperature, sa- linity, pH and other factors found in open coastal waters. Galtsoff (1928), working in the laboratory under the fairly stable conditions of a relatively constant water supply and lesser differences between diurnal and nocturnal illumination, found no correlation between periods of clo- sure and daylight hours. Webb (1930), working on O. edulis, found that the onset and termination of daylight have but little influence on the behavior of oysters as judged by their valve movements. Hopkins TEMPERATURE AND SHELL MOVEMENTS OF CLAMS 181 (1931) observed the diurnal variation in the amount of time oysters (O. lurida) remained open, but this, according to his opinion, could be directly correlated with temperature fluctuation. In the present study 163 complete (24-hour period) records were examined. All other rec- ords showing that the animal was either open or closed for the entire period of 24 hours were not taken into consideration. The period of daylight was taken as the time between sunrise and sunset. After aver- aging all the data it was found that the clams kept their shells closed during 25 per cent of the total daylight, and 19 per cent of the total time of darkness. Thus, it appears that whereas there was no diurnal vari- ation in shell activities of the clams, the animals were closed for a some- what longer period in daylight than in darkness. Records, mentioned above, were obtained between April 10 and July 29, 1935, thus covering the period of 110 days during which the environ- mental conditions of clam existence were gradually changing. In that span of time the temperature of the water increased from 3.7° to 27.5° C. Such a change in temperature, from below hibernation point for the majority of the clams to the maximum temperature of the year, offered an opportunity to determine whether the ratio between time of shell closure at daylight and at night varied with the rise in temperature as the season progressed. To answer this question all 163 records were arranged in six temperature classes, 5.0° apart. In referring to Table II, which shows the results obtained, it should be remembered that be- cause the number of records is different for each temperature-class the direct comparison between classes is somewhat difficult. However, by studying this table it will be noted that the ratio of time the shells re- mained closed in daylight or at night varied considerably with the tem- perature. Such variation is especially significant at the temperature ranging from 5.0° to 14.9° C. At these temperatures the animals had their shells closed for a much longer time in daylight than at night. At present no explanation can be advanced for the occurrence of this phenomenon. SUMMARY 1. A new apparatus, by means of which the shell activities of many bivalve mollusks can be measured and recorded, is described. 2. The analysis of 399 daily records of shell activities of 47 clams, subjected to temperatures ranging from - - 1.0° to 28.0° C., showed that the length of time which the animals remain open partly depends upon the temperature of the surrounding water. 3. For the majority of clams hibernation begins soon after the water temperature decreases to 5.0° and 6.0° C. At lower temperatures the 182 VICTOR L. LOOSANOFF clams may remain completely closed for very long periods. No shell movements were exhibited, and no disposal of ejecta occurred. 4. Within the temperature range of 3.9° to 10.9° C., the average period of openness increased from 4 to 88 per cent of the total time, showing a correlation with the rise of temperature. 5. There was no correlation between the duration of openness of the clam shells and the temperature increase from 11.0° to 27.9° C. Within this temperature range the shells were open from 69 to 90 per cent of the total time, but the percentage did not increase simultaneously with the increase of water temperature. 6. The highest percentage of time open was recorded at tempera- tures 21.0° to 22.0° C., when the clams remained open 90 per cent of the total time, or 21 hours and 36 minutes per 24-hour period. 7. Small changes in the temperature of the surrounding water did not influence the shell movements of clams. 8. There appeared to be no definite periodicity in the clam shell movements. During brief intervals, the shell movements may be of a periodic type and exhibit a rhythmic character, but generally they appeared to be unsystematic and inconsistent. 9. There were considerable individual variations in the behavior of clams kept under identical environmental conditions. 10. Under the conditions of the experiments no definite correlation between the stages of tide and the shell movements of the animals could be detected. 1 1 . The animals were closed for somewhat longer periods in daytime than in darkness. LITERATURE CITED GALTSOFF, P. S., 1928. Experimental study of the function of the oyster gill and its bearing on the problems of oyster culture and sanitary control of the oyster industry. Bull U. S. Fish., 44: 1928 (1929) : 1-39. HOPKINS, A. E., 1931. Temperature and the shell movements of oysters. Bull. U. S. Fish., 47 : 1-14. HOPKINS, A. E., 1936. Activity of the adductor muscle in oysters. Phvsiol. Zool, 9 (4) : 498-507. NELSON, T. C., 1921. Report, Dept. of Biol., New Jersey Agric. College Exp. Sta., for year ending June 30, 1920 (1921) : 317-349. Trenton. WEBB, MARSHALL H., 1930. An apparatus for recording the valve movements and the extrusion of dejecta of oysters. Jour, du Conseil, 5 (3) : 361-382. THE ACTION OF CERTAIN DRUGS ON THE INSECT CENTRAL NERVOUS SYSTEM K. D. ROEDER TUFTS COLLEGE, MASSACHUSETTS The purpose of this paper is to describe the effects of certain drugs on general reflex activity in insects, and to compare the pharmacology and physiology of the insect and vertebrate nervous systems. The litera- ture reveals few positive observations of the effect of injected drugs on insects, though in general it seems that insects are less responsive than other invertebrates. MATERIAL AND METHODS The material for these experiments consisted of mature female pray- ing mantids (Mantis religiosa) and mature cockroaches (Periplaneta americana) . The various drugs used were made up in .6 per cent saline shortly before injection. The drugs were introduced by injection into the head capsule with a small syringe and fine hypodermic needle. Since the insect circulation is rather sluggish, there is only slight general dis- tribution of an injected solution, and the site of injection determines to some extent the effect of a drug. An injection of .01 cc. into the head capsule appears to reach the supraesophageal and to some extent the subesophageal ganglion, but has little or no effect on the rest of the nerve cord. Records were kept in the form of written observations and 16 mm. moving pictures. Little attempt will be made to present quanti- tative results, since it is difficult to inject an exactly measured dose into an unanaesthetised insect, and there is no satisfactory way of quantify- ing the resultant change in behavior without giving false impressions. EXPERIMENTAL Strychnine The normal posture and degree of activity of the mantis are variable, and have been described at length (Roeder, 1937). Injection of .02 mg. strychnine causes a slight drop in tonus and decrease in antennal vibra- tion. This effect is very transitory, lasting only a few minutes. Larger doses of from .05 to .1 mg. result in a pronounced departure from the normal within 3 to 5 minutes. First the antennal vibrations become 183 184 K. D. ROEDER slower and the movements of the mouth-parts very sluggish. With large doses movements of the head appendages cease entirely. The tonus of the trunk muscles drops, and the insect takes on a crouching posture owing to the sagging of the long prothorax. The degree of locomotor activity varies, though at first there is an increase, the insect walking slowly and restlessly. The movements of the legs are coordinated, and identical with normal locomotor movements. Owing to insensitivity of the head appendages and the crouching posture, the mantis is unable to climb up or over any obstacle, and since it moves in a straight line, it eventually becomes wedged in a corner or under some object. Mechani- cal stimulation of the antennae or photic stimulation of the eyes fails to produce any response, and the whole head appears to be quite insensitive. These symptoms : a drop in tonus, increased locomotor activity, and paralysis of the head appendages also follow destruction of the cerebral ganglia, therefore the effect of strychnine on the mantis central nervous system is entirely depressant. Experiments with cockroaches yielded essentially similar results. The insects became very quiet, and there was little mouth-part or antennal movement. Both mantids and cockroaches recovered completely from the effects of strychnine within an hour. The effect of strychnine on these insects is then in direct contrast to its well-known effect on the vertebrate spinal cord, where it supposedly lowers thresholds between neurons in the cord, and is capable of turning an inhibitory process into an excitatory one. Many invertebrates show a comparable reversal of inhibition (Knowlton and Moore, 1917; Moore, 1918; Crozier and Federighi, 1924; and Crozier, 1927, 1930), though strychnine is recorded as having little or no effect on insects (Crozier, 1922; Crozier and Pilz, 1924). Many other papers are summarised in a review by Poulsson (1920). The only vertebrate organ which is de- pressed by strychnine is the heart (Poulsson. 1920). Pilocarpin This drug belongs to a group which has a specific stimulating effect on parasympathetic effectors (Kuntz, 1934). In vertebrates it increases salivary secretion, produces vasodilation, has a vagus-like effect on the heart, and increases contractility in the stomach and intestine, while its action is blocked by atropin (Dixon and Ransom, 1924). No record of its effect on the vertebrate central nervous system could be discovered. If pilocarpin is injected into the head capsule of a praying mantis, it produces a state of great excitation. The minimum dose producing observable effect is about .01 mg. while doses of from .02 to .1 mg. bring about a change in behavior within thirty seconds. The walking legs become flexed, the prothorax raised, and the head ventrally flexed. DRUGS AND THE INSECT NERVOUS SYSTEM 185 apparently due to an increase in tonus of the ventral neck muscles. There is an increase in amplitude and rate of antennal movement, and the mandibles are opened and shut, producing a clicking sound. There are spasmodic movements of the raptorial legs, consisting of flexions and extensions, and occasionally the whole body is swayed rapidly from side to side. All reflex activity increases except locomotion. Only three out of thirty insects showed any tendency to walk, either spontaneously or on stimulation. Recovery occurs rapidly and the insect is normal within an hour. High muscle tone and inhibition of locomotor activity are typical of an insect with the right and left cerebral ganglia separated. This is in- terpreted as being due to increased activity of the tonus and inhibitory locomotor centres when released from the inhibitory effect of the contra- lateral ganglion (Roeder, 1937). Since pilocarpin when applied to the cerebral ganglia of an intact insect has the same effect as median division, it must have an excitatory action upon these centres. This can be sub- stantiated by removing the cerebral ganglia prior to injection. An insect so treated walks rapidly in a crouching position. Since the cerebral centres are now absent the pilocarpin exerts its excitatory effect upon the locomotor centre in the subesophageal ganglion with consequent activity. Pilocarpin has a comparable effect on cockroaches. A solution of the drug was injected into the head in doses of approximately .01 cc. A solution of 1 : 500 produced immobility and apparent paralysis but solu- tions of 1 : 500 to 1 : 5,000 caused intense excitation. It is impossible to give an adequate description of the behavior because so many things occur simultaneously. The mouth-parts and antennae are in continual movement, the head is moved from side to side and ventrally flexed, and cleaning movements are made with the legs, which are often ab- normally extended. Occasionally there are violent cramps and spasms involving the whole of the body musculature and very rapid side-to-side swaying movements similar to those observed in treated mantids. More dilute solutions of 1 : 5,000 to 1 : 25,000 either have no noticeable effect or produce only a restlessness and general increase in activity. The above symptoms appear within one minute of injection and last for about thirty minutes. Recovery is usually complete. The excitation produced by pilocarpin seems to indicate a similarity between the nervous systems of these two insects and the vertebrate parasympathetic. This similarity leads to an investigation of the response to acetylcholine and eserine. Acetylcholine and Eserine Recent work has shown that acetylcholine acts as mediator between parasympathetic nerves and autonomic effectors, and between pre- and 186 K. D. ROEDER postganglionic fibres in the autonomic ganglia. Naturally produced acetylcholine is destroyed very rapidly by the enzyme esterase, but it can be protected from destruction by a drug, eserine. By its protective action eserine renders autonomic effectors and postganglionic fibres more sensitive to introduced acetylcholine, and, by preventing the immediate destruction of naturally produced acetylcholine, can prolong the post- ganglionic response to a preganglionic stimulus (Cannon and Rosen- blueth, 1937). For these experiments cockroaches only were used, as mantids were not available. Acetylcholine hydrochloride was dissolved in .6 per cent saline, and injected into the head, in .01 cc. amounts. Strengths from 1 : 500 to 1 : 25,000 were used but in no case was any change in behavior noted. Since the acetylcholine was unprotected by eserine its effect may have been only transitory, and therefore would have escaped atten- tion. In order to prolong the effect, if any, cockroaches were first in- jected with eserine. This drug had an unexpected effect, which was comparable with, but more pronounced than, that of pilocarpin. Eserine alone in dilutions of 1 : 250 to 1 : 500 produced an almost instantaneous spasm and immobility from which the insect did not recover, while solu- tions of 1 : 500 to 1 : 2,500 caused enormous reflex activity, which con- sisted of rapid antennal quivering, continuous mouth-part activity, move- ments of the head, great increase in extensor leg tonus and general spasms. Usually the insects did not walk, but one or two were so active that no observation could be made. Solutions more dilute than 1 : 5,000 either had no effect or produced only a slight restlessness. It can be seen that this strong eserine effect made it impossible to test the effect of acetylcholine protected by eserine. Though the pres- ence of acetylcholine in the nervous system of these insects has not been demonstrated, the action of eserine could be explained by its presence. Atropin The specific action of this drug is to block transmission of impulses in autonomic ganglia and to parasympathetic effectors. It also renders these structures insensitive to parasympatheomimetic substances such as acetylcholine and pilocarpin (Cannon and Rosenblueth, 1937). From the foregoing experiments it might be expected that atropinised insects would fail to respond to pilocarpin, and would also show a decreased reflex activity. The answer to the second question is somewhat doubtful. Atropinised cockroaches remain very inactive for long periods, though they are able to respond if disturbed. Cockroaches were first brought into a condition of great activity with .01 cc. of 1 : 500 pilocarpin, and DRUGS AND THE INSECT NERVOUS SYSTEM 187 then subjected to an injection of atropin. This failed to decrease the pilocarpin effect to a noticeable extent. If the insects were first injected with atropin and then given an injection of 1 : 500 pilocarpin, they showed either normal or decreased activity in eight out of ten cases. Two insects showed a very brief phase of excitation, lasting a few sec- onds, and then became inactive, but in no case was there the prolonged and intense reflex activity which occurs if such a dose of pilocarpin is administered alone. Therefore, previous treatment with atropin pro- tects cockroaches against the pilocarpin effect, just as it does in the ver- tebrate parasympathetic. CONCLUSIONS The following conclusions seem justified : (1) The effect of strychnine on the reflex activity of the mantis and cockroach indicates that the central nervous systems of these insects and vertebrates differ in certain respects. (2) The action of pilocarpin, eserine, and atropin on the reflex ac- tivity of these insects indicates that there is a pharmacological similarity between their nervous systems and the vertebrate parasympathetic. It must be recognized that the experiments outlined above are sug- gestive rather than conclusive, since the methods used are in many ways unsatisfactory. First, there is no way of knowing the exact concentra- tion of drug actually reaching the cerebral ganglia, partly because it is not an easy matter to inject small exact amounts into a struggling insect, and partly because of the sluggish and casual nature of the insect circula- tion. Second, there is no way of telling exactly what structures are being reached by the injected drug. From information available on the functions and interrelations of the cerebral ganglia in the mantis (Roecler, 1937) it seems highly probable that these structures are being directly affected. The great activity of the mouth-parts indicates that the subesophageal ganglion is also affected to some extent, though it is hard to say to what extent the drugs are affecting sense organs and neuro-muscular junctions. Third, observation of changes in the general reflex activity is a complex and unsatisfactory criterion. However, a satisfactory technique avoiding some of these difficulties has been devel- oped, and it is hoped that more extended conclusions can be submitted in a later paper. The similarity of the insect central nervous system and the vertebrate parasympathetic immediately suggests the presence of a chemical sub- stance such as acetylcholine, which would serve as mediator in the transfer of impulses from neuron to neuron in the central nervous sys- tem. No proof of the presence of such a substance is presented here, 188 K. D. ROEDER though the action of eserine is suggestive. Acetylcholine unprotected by eserine produces no noticeable effect when injected, but an insect is ca- pable of such rapid and brief responses that one would expect very rapid destruction or removal of any substance mediating the transfer of impulses. The question as to whether these findings apply generally to insects cannot be given an adequate answer. It has been reported that strych- nine produces little or no excitation in caterpillars (Crozier and Pilz. 1924), while pilocarpin raises sensory thresholds in the same insects (Crozier, 1922), while crustacea (Clarke and Wolf, 1932; Viehoever and Cohen, 1937) and many invertebrates respond to strychnine in a vertebrate-like manner. It would seem that a generalization with respect to insects must await a further investigation. SUMMARY 1. Adult praying mantids and cockroaches were injected in the head with solutions of various drugs, and the changes in general reflex activity were recorded. 2. Strychnine caused a decrease in reflex activity in both insects, large doses causing complete cessation of antennal and mouth-part move- ment. This is contrary to its effect on the vertebrate spinal cord. 3. Pilocarpin and eserine bring about an increase in mouth-part and antennal movement, head movement and extensor leg tonus, and produce spasmodic twitching and general contraction of the body musculature. In the case of pilocarpin this effect is prevented by atropin. 4. Acetylcholine alone has no detectable effect on activity. Its effect, if any, when injected with eserine is masked by the excitatory effect of eserine. 5. It is concluded that, in their reactions to drugs, the nervous sys- tems of the two insects studied show little similarity to the vertebrate central nervous system, but considerable similarity to the vertebrate parasympathetic. LITERATURE CITED CANNON, W. B., AND A. ROSENBLUETH, 1937. Autonomic Neuro-effector Systems. The Macmillan Company, New York. CLARKE, G. L., AND E. WOLF, 1932. The mechanisms of tropistic reactions in Daphnia and the strychnine effect in Daphnia. Jour. Gen. Physio!., 16 : 99-105. CROZIER, W. J., 1922. " Reversal of inhibition " by atropine in caterpillars. Biol. Bull., 43 : 239-245. CROZIER, W. J., 1927. Galvanotropism and " reversal of inhibition " by strychnine. Jour. Gen. Physiol, 10: 395-406. CROZIER, W. J., 1930. Reversal of galvanotropism in echinoderms. Jour. Gen. Psychol, 3 : 268-276. DRUGS AND THE INSECT NERVOUS SYSTEM 189 CROZIER, W. J., AND H. FEDERIGHI, 1924. Suppression of phototropic circus move- ments of Limax by strychnine. Jour. Gen. PhysioL, 7 : 221-224. CROZIER, W. J., AND G. F. PILZ, 1924. Central nervous excitation by alkaloids in insects. Am. Jour. PhysioL, 69: 41-43. DIXON, W. E., AND F. RANSOM, 1924. Pilocarpin, physostigmin, Arecolin. Gifte welche bestimmte Nervenendigungen erregen. Handbuch der Experi- mentellen Pharmacologie 2 (2) 746-785. Herausgegeben von A. Heffter. Julius Springer, Berlin. KNOWLTON, F. P., AND A. R. MOORE, 1917. Note on the reversal of reciprocal inhibition in the earthworm. Am. Jour. PhysioL, 44: 490-491. KUNTZ, A., 1934. The Autonomic Nervous System. Lea and Febiger, Phila- delphia. MOORE, A. R., 1918. Reversal of reaction by means of strychnine in planarians and starfish. Jour. Gen. PhysioL, 1 : 97-100. POULSSON, H., 1920. Die Strychningruppe. Handbuch der Experimentellen Pharmacologie 2 (1). Herausgegeben von A. Heffter. Julius Springer, Berlin. ROEDER, K. D., 1937. The control of tonus and locomotor activity in the praying mantis (Mantis religiosa L.). Jour. E.vpcr. ZooL, 76: 353-374. VIEHOEVER, A., AND I. COHEN, 1937. Mechanism of strychnine. Am. Jour. Pharm., 109: 285-316. SOME PROPERTIES OF SPERM EXTRACTS AND THEIR RELATIONSHIP TO THE FERTILIZATION REACTION IN ARBACIA PUNCTULATA x JOHN A. FRANK (From the Marine Biological Laboratory, Woods Hole) Since the early work of Fieri (1899), numerous attempts have been made to isolate substances from sperm which could be shown to function in normal fertilization. Extracts and filtrates of sperm, prepared by a wide variety of methods, have consistently proved inactive as partheno- genetic agents. The literature of this subject has been reviewed by Sampson (1926), who points out that any positive results reported may be ruled out on the basis of technical or interpretive errors. Largely because of these discouraging results, the search for the hypothetical " fertilizing substance " of sperm has been abandoned during the past decade. The purpose of this investigation was to study certain properties of sperm extracts obtained by heating and filtering sperm suspensions of Arbacia. The experiments deal with (1) the agglutinating action of sperm extracts on Arbacia eggs, (2) the relation of such extracts to egg-water (" fertilizin " of Lillie, 1913, etc.), (3) the effects of sperm extracts on the fertilizing power of sperm, the fertilizability of eggs, and the development of fertilized eggs, and (4) the action of sperm extracts as parthenogenetic agents. MATERIALS AND METHODS The experiments were carried out during the summers of 1934, 1937, and 1938 on the gametes of the sea-urchin, Arbacia punctulata. Eggs and sperm were obtained by the method described by Just (1928). After cutting out the lantern and peristome with scissors, the coelomic fluid was washed away with sea water and the urchin placed aboral side down over a Stender dish and allowed to shed. Mixed " dry " sperm from several males was kept in a test tube until required and mixed shed eggs were washed and placed in a small volume of sea water. Where large amounts of material were needed or when the 1 A thesis submitted in partial fulfillment for the degree1 of M.D., Yale Uni- versity School of Medicine, New Haven, Conn. 190 PROPERTIES OF ARBACIA SPERM EXTRACTS animals shed poorly, the ripe gonads were removed with forceps, washed, and strained through cheesecloth. All instruments and glass- ware were rinsed thoroughly in tap and sea water before use to prevent chance contamination of eggs with sperm. In preparing jellyless eggs, shed (" normal ") eggs were treated with a mixture of 1.4 cc. N/10 HC1 in 50 cc. sea water, washed, and ex- amined in Chinese ink to insure the absence of jelly (Lillie, 1915fl). This procedure removes the jelly from 90-100 per cent of the eggs without harming them. Sperm extracts were prepared as follows: Sperm suspensions, made up in 10-cc. portions by adding a measured volume of dry sperm to sea water, were heated one minute over a small Bunsen flame. The al- buminous sperm coagulum was removed by filtration through No. 5 \\liatman filter paper. The filtrate, or sperm extract, was considered to be- equal in concentration to that of the sperm suspension from which it was obtained. Thus a 10 per cent sperm extract was prepared by beat ing and filtering 1 cc. sperm in 9 cc. sea water. In preparing and determining the fertilizin concentration of egg- water, Lillie's method (1914) was followed with slight modifications. Fresh unfertilized Arbacia eggs, which bad stood at least fifteen min- utes in a small volume of sea water, were lightly centrifuged and the supernatant fluid ( egg-water ) removed with a pipette. The presence of fertilizin was detected by adding a drop of egg-water to a drop of fresh 1 per cent sperm suspension and observing reversible agglutina- tion of the sperm under the microscope. By making successive half- dilutions of the egg-water with sea water, a dilution is reached at which the sperm agglutinate for three to six seconds but above which no agglutination occurs. This dilution, by definition, contains one fer- tilizin unit and the fertilizin concentration of the original egg-water is thereby determined. For example, if an egg-water agglutinates sperm at 1 '1600 dilution but not at 1/3200, the undiluted egg- water contains 1600 fertilizin units. THE EGG- AGGLUTINATION REACTION \Yhen a few drops of Arbacia eggs are added to a small quantity of sperm extract in a tube, a striking reaction occurs. As each drop falls it sets into a dense red mass which remains suspended from the surface of the solution by a thin ribbon of coagulated eggs (Fig. 1, A ). When the tube is shaken the ribbons coalesce to form a loose clot of eggs which usually floats at the surface of the extract (Fig. 1, B). A similar quantity of eggs, shaken in heated water as a control, forms 192 JOHN A. FRANK t PLATE I Microscopic appearance of the egg-agglutination reaction. All eggs photo- graphed in Chinese ink. FIG. 2. Normal .-irbacia egg in heated sea water surrounded by a wide zone of clear, colorless jelly. X 430. FIG. 3. A similar egg after five minutes exposure to a 5 per cent sperm extract. Note that a distinct agglutination membrane has appeared at the periphery of the jelly. X 430. FIG. 4. A similar egg after thirty minutes exposure to a 5 per cent sperm extract. The agglutination membrane has become denser but the jelly immediately in contact with the cortex remains unaffected. X 430. IMG. 5. Strands of agglutinated jelly connecting eggs in 5 per cent sperm extract. X 215. PROPERTIES OF ARBACIA SPERM EXTRACTS 193 a homogeneous suspension which slowly settles (Fig. 1, C). Subse- quent agitation of each tube causes an increase in the density of the egg- masses in the extract, while the eggs in sea water redistribute them- selves through the solution and again settle to the bottom of the tube. We will call this phenomenon the " egg-agglutination reaction," and the inciting substance present in sperm extracts, the egg-agglutinating sub- stance. These terms are used in a purely descriptive sense without implying an analogy to immunological agglutinations. FIG. 1. Gross appearance of the egg-agglutination reaction. All photographs natural size. (A) Five drops Arbacla eggs in 3 cc. of 5 per cent sperm extract. Note the drops of agglutinated eggs suspended from the surface of the solution. (B) Ten drops of Arbacia eggs in 3 cc. of 5 per cent sperm extract. The tube was shaken once. The eggs have formed a large dense mass which floats at the surface. (C) Ten drops of Arbacla eggs in 3 cc. heated sea water. The tube was shaken once after adding the eggs. Observe that the ova are homogeneously suspended and have begun to settle as shown by the clear area immediately below the surface. The details of the egg agglutination reaction may be observed micro- scopically by making an egg suspension in Chinese ink which reveals the colorless jelly layer, or chorion, surrounding each ovum. By means of a capillary pipette, sperm extract is introduced into a drop of this mixture mounted on a slide beneath a raised coverslip and observed under high power. As the extract comes in contact with an egg, a faint black line of adhering ink particles immediately appears at the periphery of the jelly. This line rapidly increases in thickness until the chorion is encircled by a dense membranous structure which will be termed the " agglutination membrane." 194 JOHN A. FRANK The microscopic appearance of eggs treated with sperm extract is shown in Plate I, Figs. 2-5. A normal egg is surrounded by a wide zone of clear colorless jelly with poorly-defined margins (Fig. 2). On treating such an egg with sperm extract, a distinct agglutination mem- brane forms at the periphery of the jelly (Fig. 3). which gradually increases in density (Fig. 4). Particles of jelly freely suspended in the solution also agglutinate to form a dense interlacing network con- necting adjacent eggs (Fig. 5). With these changes the jelly becomes very adhesive, causing the ova to stick firmly together in clusters when the coverslip is moved about. Woodward (1915) states that the jelly of Arbacia eggs placed in " boiled sperm suspension " swells and becomes sticky so that the eggs adhere to each other and to the container. This observation was repeatedly confirmed during the present study with the exception that swelling of the jelly was never noted, either in boiled sperm suspensions or in sperm extracts. In some instances the chorion retains its original width but it generally shrinks markedly, as shown by a decrease in the space be- tween agglutination membrane and cortex after prolonged exposure to the solution. The egg-agglutinating substance affects only that portion of the jelly with which it comes in contact and incomplete agglutination membranes involving one-half or less of the circumference of the chorion appear on those eggs which have been only partially exposed to the extract. In general, the density and rate of formation of the agglutination membranes are directly related to the concentration of extract. A 4 per cent sperm extract clots the eggs immediately into a solid mass sur- rounded by a thick unbroken line of agglutinated jelly ; as the extract decreases in concentration, fewer eggs are involved, the reaction is slower, and the membranes are thinner until below about 0.5 per cent both gross and microscopic reactions become negative. The egg-agglutination membranes grow tougher and more resistant to tension with time. After a few minutes exposure to a 2 per cent extract, the coverslip may be gently agitated without altering the circular shape of the membranes and the eggs move about freely in the unaffected central jelly. When greater tension is applied by rapidly agitating the coverslip, the membrane ruptures and the ovum may then be torn away leaving the membrane firmly anchored to the slide with its frayed edges waving about in the solution. The egg-agglutination reaction is irreversible. Eggs which have once agglutinated remain firmly massed together in spite of transfer to sea water and eventually cytolyse /;/ situ. Vigorous shaking increases PROPERTIES OF ARBACIA SPERM EXTRACTS 195 the size and density of the egg-clusters and leads to rapid destruction of the ova. The jelly of Arbacia eggs binds the egg-agglutinating substance and removes it from the extract. This is illustrated by a typical experiment : 1 cc. shed eggs was added to 3 cc. of 20 per cent sperm extract and the agglutinated egg-mass was filtered off. The filtrate weakly agglu- tinated 1 cc. eggs, after which the solution was again filtered but this filtrate failed to agglutinate .5 cc. eggs. Three cc. of the same extract diluted with 2.5 cc. sea water as control strongly agglutinated eggs. The egg-agglutinating substance present in the 20 per cent extract has, there- fore, been fixed by the jelly of 2 cc. eggs. Sperm extracts not only act on the jelly but also on the cortex of Arbacia eggs. To study this, jellyless eggs were divided into two lots, one of which was fertilized with fresh sperm. Samples from each lot were fixed in Bouin's solution (without acetic acid) or 2 per cent osmic acid and then washed thoroughly in sea water. Jellyless eggs treated with sperm extract form clusters which are too minute to be reliably detected grossly so a microscopic method was used: a drop of jellyless eggs was mixed with a few drops of extract on a slide and a glass needle was moved through the mixture. The appearance of distinct egg- clusters represented a positive agglutination test. Numerous experiments demonstrated that the cortex undergoes cer- tain changes in sperm extracts causing the eggs to fuse so firmly together that bits of cytoplasm may be ripped out of the ovum by agitating the egg-clusters. In addition the fertilization membranes of fertilized jelly- less eggs adhere to each other and may be pulled completely away from the zygote. These changes take place equally in unfertilized and ferti- lized jellyless eggs whether living or dead. The structure of the cortex, however, does not appear to be visibly altered by exposure to sperm ex- tracts. Controls of the same eggs in heated sea water remained singly spaced in all cases, except for occasional small clusters caused by the use of acid in removing the jelly (Lillie, 191 5a). Identical experiments on normal eggs showed that the jelly of fer- tilized and unfertilized eggs, both living and dead, agglutinates in sperm extracts. Immature ovocytes, with or without jelly, agglutinate as readily as mature ova. To sum up : ( 1 ) The egg-agglutination reaction is not dependent on living protoplasm, since dead ova agglutinate as readily as living ones. (2) Immature, unfertilized, and fertilized eggs agglutinate equally, indi- cating that the developmental state of the ovum plays no part in the reaction. (3) Sperm extracts agglutinate jellyless eggs by directly af- fecting the cortex which becomes markedly adhesive. 196 JOHN A. FRANK One of the first questions to arise was whether the egg-agglutinating substance could be extracted from tissues or cells of Arbacia other than sperm. Extracts of Arbacia shells, lanterns, peristomes, ovaries, eggs, and coelomic fluid, prepared by heating 4 cc. of each tissue one minute in 10 cc. of sea water, were found to be without effect on ova. Extracts of immature testes or ripe testes washed free of sperm were likewise inert. The egg-agglutinating substance is thus highly tissue-specific and can be extracted solely from sperm. Moreover, sperm extracts act spe- cifically on Arbacia eggs and will not agglutinate blood corpuscles, sper- matozoa or any other cells of the male or female sea-urchin. The egg-agglutination reaction is not confined to the gametes of Ar- bacia and a certain degree of cross-agglutination exists between eggs and sperm extracts of related and distant species. Preliminary experiments on the gametes of Nereis and Echinarachnius gave the following results : (1) Echinarachnius sperm extracts agglutinate ova of Echinarachnius and Arbacia. (2) Nereis sperm extracts agglutinate Arbacia eggs but do not agglutinate fertilized or unfertilized Nereis eggs. In Nereis, jelly extrusion occurs after fertilization, hence both fertilized and un- fertilized ova were tested for agglutination. Just (1922) found that Nereis sperm boiled in sea water causes a small percentage of jelly for- mation, maturation, and differentiation without cleavage of unfertilized Nereis eggs. (3) Sperm extracts of Arbacia agglutinate Echinarach- nius eggs. Such extracts will not agglutinate Nereis eggs but cause jelly extrusion and maturation (without cleavage) of unfertilized Nereis eggs. The chemical specificity of the egg-agglutination reaction is less dis- tinct than the biological. Arbacia eggs are agglutinated by certain salts, foreign sera (Robertson, 1912), acids and alkalis. PHYSICAL AND CHEMICAL PROPERTIES OF THE EGG-AGGLUTINATING SUBSTANCE These experiments were devised to study certain general physical and chemical properties of the egg-agglutinating substance and do not repre- sent exact quantitative studies. Each experiment was repeated several times with different materials and the results were consistent in each case. The egg-agglutination tests were standardized by adding three drops of fresh eggs to 2 cc. of sperm extract in a test tube and shaking the tube lightly. In certain experiments the agglutinating strength of the ex- tracts was rated as follows : PROPERTIES OF ARBACIA SPERM EXTRACTS 197 0 = = No agglutination. All eggs homogeneously suspended. 1 = = Few thin shreds of agglutinated eggs with the majority freely sus- pended. = One or more loose clusters of agglutinated eggs with some ova freely suspended. 3 == Single dense clot of agglutinated eggs with no eggs free in the solu- tion (maximum). Appearance, Specific Gravity, and pH of Sperm Extracts Extracts of clean shed sperm are colorless while extracts of testis sperm have a light purple tint ("purple X" of Glaser, 1914). Dilute sperm extracts are clear; those of high concentration (40 per cent) are faintly opalescent. The specific gravity of a 50 per cent extract is 1022, which about equals that of sea water (av. sp. gr. = = 1028). The pH of a 60 per cent sperm extract in sea water, measured with the Coleman glass electrode, is 7.74 ; that of a 1 per cent extract is 8.02. The slight relative acidity of the 60 per cent extract probably results from CCX liberated by the thick sperm suspension prior to extraction. These val- ues approximate the pH of sea water at Woods Hole (av. pH = 8.00) . Relation of Temperature to the Extraction and Stability of the Egg- aggh i tin a ting Substan c e To determine whether the egg-agglutinating substance can be ex- tracted from sperm at low temperatures, 10 cc. sea water was heated in a waterbath to a given temperature, a measured amount of mixed shed sperm was stirred into the sea water and kept for one minute at the desired temperature, after which the solution was filtered. Under these conditions it wras found that sea water below about 57° C. will not ex- tract the egg-agglutinating substance. This can be extracted in small amounts from a 20 per cent sperm suspension at 57° and from a 1 per cent suspension at 68° C. Higher temperatures extract greater quan- tities of the substance from sperm. There appears to be an inverse relation between sperm concentration and the critical temperature of extraction. The rate of loss of the egg-agglutinating substance on standing is a function of the temperature of the extract : 50 cc. of a 10 per cent sperm extract was divided into two equal parts, one of which was aged at room temperature (26° C.) and the other in an icebox (7° C). The results of tests on each solution at various times are given in Table I. The agglutinating strength of the extract kept at 7° C. remained constant for twenty-five hours and then gradually diminished and dis- 198 JOHN A. FRANK appeared after forty-five hours. The same extract at 26° C. lost its ac- tivity in less than thirteen hours. In a similar experiment, a 1 per cent extract lost its agglutinating power in less than twelve hours at 26° C, and in twenty hours at 7° C. showing that the egg-agglutinating sub- stance disappears more rapidly from dilute extracts. We may conclude that this substance is preserved best at low temperatures. The egg-agglutinating substance is highly heat-resistant and only slowrly destroyed by boiling : 25 cc. of a 20 per cent sperm extract was boiled for several hours in a beaker, the volume being kept constant by addition of distilled water. The agglutinating strength of the extract remained high (3) for four hours, then gradually decreased, and dropped to zero after five and one-half hours of boiling. TABLE I Egg-agglutinating strength of a 10 per cent sperm extract aged at 7° C. and 26° C. (3 = maximum) Temperature Age of Extract 7° C. 26° C. hours 0 3 3 13 3 0 25 3 38 2 41 1 45 0 Dialysis and Filtration Thirty cc. of a 30 per cent sperm extract was divided into three equal parts (A, B, and C), each of which was placed in a Thomas Diffusion Shell (No. 4471). Extract A was dialysed against running sea water in a beaker, B was suspended in a cylinder containing 10 cc. sea water, and C was kept as a control. After twelve hours each extract was tested for its activity. If dialysis had occurred, one would expect a decrease in the concentration of the egg-agglutinating substance in extracts A and B and the appearance of this substance in the dialysate of extract B. By dilution of each solution with sea water, ho\vever, it was found that the concentration of the substance in A and B exactly equalled that of the control, C. The dialysate of B was totally inert. Numerous similar ex- periments proved conclusively that the egg-agglutinating substance is non-dialysable. Also it will not pass through a Berkf eld filter regardless of the concentration of sperm extract. These facts suggest that it is a colloidal substance of large molecular size. PROPERTIES OF ARBACIA SPERM EXTRACTS 199 Source of the Egg-agglutinating Substance The sperm heads of Arbacia are composed chiefly of nucleic acids and " Arbacin," a substance having properties in common with both histones and protamines (Mathews, 1897). In view of the relatively minute amounts of lipids which are confined largely to the sperm tails, it seemed likely that the egg-agglutinating substance might be a protein or protein derivative. Sperm extracts were found to be negative to the following tests for proteins : Millon's, Adamkiewicz, biuret, xanthopro- teic, and ninhydrin. No precipitate forms on adding saturated (NH4)2- SO4 or 70 per cent alcohol and no reduction occurs with Benedict's solu- tion or bismuth subnitrate. Hence if proteins or carbohydrates are pres- ent in sperm extracts, their concentration is too minute to be detected by ordinary chemical tests. The lipids of sperm may be readily separated from the proteins by alcohol-ether extraction. The procedure used was adapted after Bloor's method for extraction of total lipids from blood serum (Peters and Van Slyke, 1932, p. 495). Ten cc. of mixed shed sperm was shaken in 150 cc. alcohol-ether mixture (three parts 95 per cent alcohol plus one part ethyl ether) and the mixture boiled by immersion in a waterbath. The solution was made up to 200 cc. with alcohol-ether, cooled, and filtered through fat-free filter paper. The rubbery pink protein residue was washed thoroughly in sea water, heated one minute in 10 cc. sea water and filtered. The filtrate strongly agglutinated eggs. The honey-colored lipid extract was evaporated to dryness on a waterbath, the residue was dissolved in 100 cc. petroleum ether, filtered, and re-evaporated to dryness. This procedure yielded a small quantity of yellow waxy material which was shaken in 10 cc. sea water, heated one minute, and filtered. The filtrate had no effect on eggs. To determine whether heat was the factor responsible for liberating the egg-agglutinating substance in the above procedure, shed sperm was shaken in cold alcohol-ether solution, after which the protein precipitate was washed in sea water, shaken vigorously in 10 cc. sea water, and filtered. The filtrate markedly agglutinated eggs proving that heat is not essential for extraction. To sum up, the egg-agglutinating substance is derived from the pro- tein residue (head) of the spermatozoon and is not found in the lipid fraction. Chemical Composition of the Extraction Medium Sperm extracts made in tap or distilled water will not agglutinate eggs and if the residue from a sperm suspension, which has been heated 200 JOHN A. FRANK one minute in sea water, is reheated in distilled water, the extract is likewise inert. By substituting sea water for distilled water, in the above procedures, the egg-agglutinating substance is readily obtained. These facts suggested that certain salts or ions, present in sea water, are essential for the extraction of the substance. A thorough study of this problem requires chemical analyses of various extraction media and is beyond the scope of this investigation which is concerned with reac- tions occurring in sea water. The experiments cited below, therefore, represent preliminary studies: Ten per cent sperm suspensions were made up in each of the fol- lowing solutions and extracted by heating one minute and filtering : 1. Van't Hoff artificial sea water solutions each of which lacked a single salt present in normal sea water : CaQ2, MgSO4, MgCL, KC1, and NaCl. 2. Pure solutions of each of the above salts in distilled water. 3. Pure solutions of Mg, Ca, K, and Na acetate in distilled water. The reagents were all made up isosmotic with sea water and the pH varied from about 7.2-9.0. (The viscosity, ionic concentration, and molarity could not be controlled.) None of the solutions visibly in- jured ova or sperm, or agglutinated eggs. It was found that an active sperm extract agglutinated eggs washed in any of the above media to the same degree as eggs washed in sea water. Shed eggs in sea water were therefore used in the egg-aggluti- nation tests. From the results, listed in Table II, the following conclusions may be drawn : (1) Sea water lacking magnesium (MgCL, MgSO4 or both) will not extract the egg-agglutinating substance from sperm whereas pure solutions of MgCl2 or MgSO4 will. Magnesium is thus essential to the extraction process. (2) Sea water lacking calcium gave variable results ; out of 6 tests, 3 were weakly positive and 3 negative. A pure solution of CaCl2 readily extracts the egg-agglutinating substance. Calcium, therefore, appears to be necessary in sea water for the extraction of this substance. (3) Sperm extracts made in sea water which lacks either NaCl or KC1 agglutinate eggs. Extracts made in pure NaCl or KC1 solutions have no effect on eggs. These salts evidently play no part in the ex- traction of the egg-agglutinating substance in sea water. (4) If acetate ions are substituted for chloride ions in pure solutions of the salts of sea water, it is found that Mg and Ca acetate will extract the egg-agglutinating substance from sperm but Na and K PROPERTIES OF ARBACIA SPERM EXTRACTS 201 acetate will not. The egg-agglutination reaction is therefore not de- pendent upon the presence of any particular anions. To sum up, the bivalent cations, magnesium and calcium, are the only ions of sea zvater which are necessary for the extraction of the egg-agglutinating substance from sperm. Both of these ions must be present in sea water but each will act alone when present in sufficient concentration in distilled water. Extraction of Sperm in Acid and Alkaline Sea Water In these experiments .10 N HC1 and .10 N NaOH were diluted with varying quantities of sea water and the solutions used in preparing 5 per cent sperm extracts. TABLE II The agglutination of eggs by 10 per cent sperm extracts prepared in various isosmotic salt solutions. Three drops of eggs added to 2 cc. of each extract, -f- = positive tests. 0 = negative tests. All solutions prepared by Marine Biological Laboratory Chemical Room. Extraction Solution Egg-agglutination * Artificial Sea Water (Van't Hoff ) + * Ca-free Sea Water Q p * Mg-free Sea Water 0 * MgSO4-free Sea Water 0 McCl2-free Sea Water 0 KCl-free Sea Water + * NaCl-free Sea Water + .34 M CaCl2 + * .37 M MgCl2 + *MgSO4-7H2O + .53 M KC1 0 .52 M NaCl 0 Ca Acetate + Mg Acetate + K Acetate 0 Na Acetate 0 * Indicates extracts on which fertilizin-inactivation tests were run with results parallel to those for egg-agglutination tests. Lillie (1915a) states that acid sea water heavily agglutinates and later cytolyses Arbacia eggs. The present study shows that when sperm is extracted in acid sea water of above .003 N HC1 content, the agglu- tinating effect of the acid hides the action of the egg-agglutinating sub- stance. These substances, however, have different actions since HC1 agglutinates and later cytolyses eggs whereas the egg-agglutinating sub- stance agglutinates but does not kill the ova. Sperm extracts in sea water of or below .003 N HC1 content agglutinate eggs. Sea water of 202 JOHN A. FRANK this degree of acidity has no effect, which proves that the agglutination is here caused by the egg-agglutinating substance alone. Haas (1916), Irving (1926), and Kapp (1928) have shown that the addition of .10 N NaOH to sea water precipitates most of the magnesium and some calcium as hydroxides. In the present study it was found that alkaline sea water containing more than .01 N NaOH loosely flocculates Arbacia eggs, probably due to precipitation of calcium and magnesium hydroxide in the egg jelly. This precipitate is removed when alkaline sea water is heated and filtered (as in preparing sperm extracts) and eggs added to the filtrate remain singly spaced. It is thus possible to determine the highest concentration of alkali in sea water which will permit the extraction of the egg-agglutinating sub- stance from sperm. This was found to be about .012 N NaOH sea water; above this degree of alkalinity no agglutination takes place; below it the extracts agglutinate eggs. The previous section has shown that sea water lacking Alg or Ca will not extract the egg-agglutinating sub- stance from sperm. According to Irving (1926), .017 molar NaOH precipitates 19 per cent of the total Ca and 16 per cent of the total Mg from sea water. A possible explanation for the action of alkaline sea water in extracting sperm is that sufficient Mg and Ca are removed from sea water by concentrations of alkali above .012 N NaOH to render it inert as an extracting medium. Below this degree of alka- linity, however, enough ionized Mg and Ca are present to allow extrac- tion of the egg-agglutinating substance. The pH of the solutions varied from 3.2 (.003 N HC1 in sea water) to 9.5 (.012 N NaOH in sea water). Since sperm extracts made in either of these solutions agglutinate eggs, the egg-agglutination reaction takes place over a wide range of hydrogen-ion concentrations. THE INACTIVATION OF FERTILIZIN BY SPERM EXTRACTS As is well known from the classical studies of F. R. Lillie (1913- 1919), mature unfertilized eggs of Arbacia secrete a substance into sea water which reversibly agglutinates sperm of the same species. Lillie named this substance " f ertilizin " and concluded that it functions in fertilization as an essential chemical link in the union of egg and sperma- tozoon. For further details concerning the fertilizin theory of fertiliza- tion see the recent review of Just (1930). Since fertilizin is present in high concentration in the jelly of Arbacia eggs, it seemed probable that sperm extracts might react in some way with this substance, which proved to be correct. When sperm extracts are mixed with egg-water, the latter loses its capacity to ag- glutinate sperm; some substance in the extract has neutralized the PROPERTIES OF ARBACIA SPERM EXTRACTS 203 fertilizin. For description purposes we will call this " fertilizin-inactiva- tion " and the hypothetical substance responsible for it the " fertilizin- inactivator." To study the relation of fertilizin-inactivation to the concentration of sperm extract, 2 cc. of successive half-dilutions of an 8 per cent extract was placed in a series of test tubes and 2 cc. egg-water of known fertilizin content was added to each tube. This procedure was repeated with egg-waters of descending fertilizin content, the concentra- tion of sperm extracts remaining constant. The highest dilution of egg-water was mixed with an equal quantity of heated sea water as a control. A drop of each mixture was tested for fertilizin by its ability to agglutinate a drop of fresh 1 per cent sperm suspension ; absence of sperm agglutination constituted a positive fertilizin inactivation test ; agglutination of sperm, indicating the presence of free fertilizin in the mixture, was considered a negative fertilizin-inactivation test. In Table III, which summarizes the results of three experiments, a change from 4- to 0 in each horizontal row represents the end-point at which the extract in the corresponding vertical column becomes too dilute to inactivate all of the fertilizin present. For example, 400 fer- tilizin units are inactivated by a 4 per cent but not by a 2 per cent sperm extract. In each experiment at least three determinations were made at each of these end-points with different sperm suspensions as indicator in order to reduce error caused by variation in materials. Note (Table III) that an 8 per cent extract inactivates 800 fertilizin units whereas a .06 per cent extract only inactivates 10; a .1 per cent extract inactivates 20 but not 40 fertilizin units. The controls agglu- tinated sperm in all instances. Sperm extracts thus inactivate fertilizin in roughly quantitative proportions, the capacity for inactivation varying directly with the concentration of extract. Certain incidental phenomena were noted during the course of this experiment. When a drop of concentrated sperm extract plus egg-water is added to a drop of spermatozoa, the latter form a dense, highly active ring around the introduced drop but sperm agglutination does not occur. Sperm extracts therefore react only with fertilizin without altering the " sperm-activating and aggregating " substances of egg-water, which confirms Lillie's statement (1914) that these substances are distinct from fertilizin. Furthermore, it was noted that on shaking concen- trated sperm extract plus egg-water mixtures, a light purple, flocculent precipitate appears and the egg-water, which previously was purple, immediately becomes colorless. The precipitate floats at the surface of the solution and is probably formed by agglutination of sub-microscopic jelly particles present in the egg-water. 204 JOHN A. FRANK It may be pointed out here that fertilizin does not appear to be essential to the egg-agglutination reaction, since fertilized jellyless eggs washed free of fertilizin agglutinate as readily as unfertilized jellyless eggs which secrete large quantities of this substance. TABLE III Combined results of three experiments on the inactivation of fertilizin by sperm extracts. Upper horizontal column gives concentration of sperm extracts; vertical column gives concentration of egg- water in units of fertilizin. For each test, 2 cc. egg-water was added to 2 cc. extract and the mixture tested for its ability to aggluti- nate a 1 per cent sperm suspension. + = positive tests (no sperm agglutination); 0 = negative test (sperm agglutination). Concentration of Sperm Extract: 8% 4% 2% 1% 0.5% 0.2% 0.1% 0.06% 0.03% Fertilizin Units Fertilizin-Inactivation Tests + + 0 0 0 800 4- 4- 0 0 0 + + 0 0 0 400 + J 0 0 0 0 0 0 0 + + + 0 0 0 200 -)- 4. 4- 0 0 + + 0 0 100 J I I 0 0 0 0 + _l_ 4. 0 0 80 + 4- 4- 0 0 + + 0 0 0 + + 4. 0 0 40 4- 4- + + 0 + + 0 0 4. + + 0 20 4- + 4. + 0 + + 0 0 0 10 I : t t 0 PROPERTIES OF THE FERTILIZIN INACTIVATOR We have described two effects of sperm extracts, one on eggs and the other on fertilizin. To determine whether these effects are related PROPERTIES OF ARBACIA SPERM EXTRACTS 205 to each other, simultaneous tests for the egg-agglutinating substance and fertilizin-inactivator were run on sperm extracts. The fertilizin-inac- tivation tests represent a part of the previously described experiments dealing with the properties of the egg-agglutinating substance and the results of the former will, therefore, be briefly summarized without repeating experimental details. The fertilizin-inactivation tests were standardized by adding .3 cc. sperm extract to .3 cc. egg-water of known fertilizin content and testing the mixture for agglutination on a fresh 1 per cent sperm suspension. The inactivating strength of a sperm extract was considered equal to the maximum number of fertilizin units which it neutralized. The fertilizin-inactivator is markedly tissue-specific and cannot be extracted from any tissue or cells of Arbacia except sperm. It has long been known that blood and tissue extracts of Arbacia will not combine with fertilizin but that extracts of washed unfertilized eggs contain a substance, " anti-f ertilizin," which neutralizes minute amounts of fertilizin (Lillie, 1914). Because anti-fertilizin combines only with fertilizin without affecting the sperm-aggregating and activating factors of egg-water, it acts like the fertilizin-inactivator, but the latter is much more potent in neutralizing fertilizin. TABLE IV Fertilizin-inactivating strength of a 10 per cent sperm extract aged at 7° C. and 26° C. Inactivating strength given as maximum number of fertilizin units inactivated by extract. Age of Extract Temperature 7° C. 26° C. hours 0 400 400 50 13 400 25 200 38 , 100 41 50 45 25 Low temperatures preserve the fertilizin-inactivator as shown in Table IV, which gives the data from a single experiment. After thir- teen hours at 26° C. the inactivating strength of a 10 per cent extract dropped from 400 to 50 ; after twenty-five hours at 7° C. the inactivat- ing strength fell from 400 to 200 and then gradually decreased to 25 in forty-five hours. The fertilizin-inactivator is quite thermostable and is gradually destroyed by boiling. The inactivating strength of a 20 per cent sperm 206 JOHN A. FRANK extract fell from 400 to 50 after two and one-half hours at 100° C. and remained at 50 after five and one-half hours of boiling. It can be obtained from the protein residue but not from the lipid fraction when sperm is extracted with hot alcohol-ether and it is ex- tractable from sperm shaken in cold alcohol-ether. It will not dialyse or pass through a Berkfeld filter and is not present in sperm extracts made in tap or distilled water. In the previously described experi- ments with artificial sea water solutions, fertilizin-inactivation tests were run on a few of the extracts (marked with a (*) in Table II) and the results paralleled the egg-agglutination tests in each case. Because the egg-aggglutinating substance and fertilizin-inactivator in general showed parallel behavior in all experiments they will be referred to from now on collectively as the " active principles " of sperm extracts. SECRETION OF ACTIVE PRINCIPLES AND THEIR RELATION TO THE FERTILIZING POWER OF SPERM Filtrates of fresh or old sperm suspensions which have been passed through Whatman filter paper and the supernatant fluid of centrifuged sperm suspensions will not react with fertilizin or eggs. The active principles are found only in sperm extracts and are not secreted by sperm into water. It was found that extracts of sperm which have been aged until dead are inert. The fact that the active principles can only be extracted from fresh sperm suggested that they might be related to the fertilizing power of sperm. Experiments were therefore undertaken to determine whether the loss of fertilizing power of aging sperm suspensions bears any relation to the loss of active principles from extracts of these sus- pensions. Sperm suspensions of various concentrations (8 per cent, 4 per cent, 2 per cent, and 1 per cent) were prepared by adding measured amounts of mixed shed sperm to 80 cc. sea water in a series of Erlen- meyer flasks. The sea water was sterilized in an autoclave to lessen bacterial growth which shortens the life of sperm (Cohn, 1918). The flasks were closed with sterile corks and aged at room temperature (26°-28° C.) for three days. At intervals 4 cc. of each suspension was heated thirty seconds, filtered, and the extract divided into two equal parts which were tested for the active principles. In the fertilizin-inactivation tests, in order to insure complete inac- tivation of fertilizin by each extract at the start of the experiment, descending concentrations of fertilizin were added ; 200 fertilizin units were added to the 8 per cent, 100 to the 4 per cent, 50 to the 2 per cent and 25 to the 1 per cent sperm extract. PROPERTIES OF ARBACIA SPERM EXTRACTS 207 Simultaneously with these tests, the fertilizing power of the sperm was measured by adding one drop from each sperm suspension to three drops of fresh eggs in 10 cc. sea water and counting the percentage of fertilized eggs in each dish two hours later. TABLE V The relation between loss of fertilizing power of aging sperm suspensions and loss of active principles from extracts of these suspensions. Horizontal rows give simultaneous tests made at intervals recorded at left. A — egg-agglutination tests. B = fertilizin inactivation tests; the figures in brackets are the number of fertilizin units added to each extract. C = fertilizing power of sperm suspensions expressed as the percentage of eggs fertilized by each suspension. -f- = positive tests; 0 = nega- tive tests. Age of Sperm Suspensions A B C Sperm Extracts Sperm Extracts Sperm Suspensions Egg-agelutination Tests Fertilizin-inactivation Tests Percentage of Eggs Fertilized (200) (100) (50) (25) 8% 4% 2% 1% 8% 4% 2% 1% 8% 4% 2% l% hours 24 + + + + + 0 + + + + + 0 + + + 0 0 0 0 + 0 0 0 0 + + 0 + + 0 + + + + + 0 + + + 0 0 0 + + + 0 0 0 100 100 100 100 100 26 100 99 100 93 13 0 100 100 100 12 0 0 100 89 94 11 3 0 30 36 . . 49 56 72 The data are recorded in Table V. Column A lists the egg-agglu- tination tests, column B the fertilizin-inactivation tests, and column C the fertilizing power of each suspension. Each horizontal row thus records comparable tests at various times. For the first thirty-six hours the active principles are present, in general, in all extracts except for the 1 per cent which has lost its capacity to agglutinate eggs. Through- out this period the sperm suspensions fertilize 89-100 per cent of the eggs. After forty-nine hours, both egg-agglutination and fertilizin- inactivation tests are negative in the 2 per cent and 1 per cent extracts and at this time the fertilizing power of the 2 per cent and 1 per cent sperm suspensions has dropped to 12 per cent and 11 per cent respec- tively. The active principles have disappeared in all extracts after seventy-two hours and the fertilizing power of each suspension has dropped to zero except for the 8 per cent suspension which fertilizes 208 JOHN A. FRANK 26 per cent of the eggs. At the close of the experiment living sperm were present in each suspension (except the 1 per cent) as shown by clear cytoplasm of the sperm heads and slight motility. Many sources of error are present in an experiment of this sort which cannot be adequately controlled. The necessity for testing the same solutions for three days introduces unavoidable contamination and the condition of the eggs and sperm used for testing varies widely in different urchins from day to day (Goldforb, 1929). Nevertheless, it is clear that the loss of fertilizing power of aging sperm suspensions parallels the loss of active principles from extracts of these suspensions. THE EFFECT OF SPERM EXTRACTS ON FERTILIZATION AND DEVELOPMENT Effect of Sperm Extracts on Both Gametes In an extensive series of experiments, normal and jellyless eggs were fertilized in sperm extracts (.1 per cent to 8 per cent), transferred to sea water after exposure to the extracts for from thirty seconds to three hours, and at various times counts were made of the percentage of eggs developing. Controls consisted of the same eggs fertilized in heated sea water. Fertilization and development in extracts below about 2 per cent concentration equals that of the controls but above this concentration marked inhibition occurs. The majority of ova remain unfertilized and no fertilization membranes appear despite the presence of numerous active sperm at the cortex. Those eggs which are acti- vated usually cleave normally but cease developing at or before the non-motile blastula stage if allowed to remain in the extracts. When the embryos are transferred to sea water they develop into normal plutei. These effects occur in both jellyless and normal eggs but the inhibition is greater in the latter, probably due to agglutinated jelly which physically prevents escape of the blastulae from the egg-masses. In these experiments both germ-cells were simultaneously exposed to sperm extracts and the block to fertilization could be due to the action of extracts on the ovum alone, spermatozoon alone or on both gametes. We will now consider the action of extracts on each gamete. Effect of Sperm Extracts on the Fertilizing Power of Sperm Sperm suspensions were made in sperm extracts and fresh eggs fertilized with these mixtures : 1 per cent sperm suspensions were pre- pared in .5 per cent to 8 per cent extracts and after exposure of sperm to the extracts for various times, one drop from each suspension was used to fertilize three drops of fresh eggs in 20 cc. sea water. The per- PROPERTIES OF ARBACIA SPERM EXTRACTS 209 centage of fertilized eggs was counted two hours after insemination. A 1 per cent sperm suspension in boiled sea water served as control. TABLE VI The inhibiting effect of sperm extracts on the fertilizing power of sperm. One per cent sperm suspensions prepared in sperm extracts of various concentrations. Figures give percentage of fresh eggs fertilized by one drop from each suspension after exposure of sperm to the sperm extracts for times given at left. Control = 1 per cent sperm suspension in boiled sea water. Concentration of Sperm Extracts 8% 4% 1% 0.5% Control Time Sperm Exposed to Sperm Extracts Percentage Eggs Fertilized 5 seconds 18 1 68 _ 15 seconds 3 2 20 92 45 seconds 8 3 17 100 60 seconds 7 5 18 92 15 minutes 3 3 32 12 30 minutes 3 1 5 4 99 60 minutes 1 1 1 1 99 From Table VI, which gives the result of one experiment, it is seen that the percentage of eggs fertilized by sperm suspensions in extracts of above 1 per cent drops greatly after fifteen seconds at which time the 4 per cent suspension fertilizes only 2 per cent of the ova. After fifteen minutes exposure, the fertilizing power of the .5 per cent sus- pension has dropped from 92 per cent to 12 per cent ; after thirty minutes the percentage of eggs fertilized by each suspension has fallen almost to zero. The controls fertilized 99 per cent of the ova after one hour. Numerous motile sperm were seen at the surface of the ova but fertiliza- tion membranes did not appear. Sperm extracts, then, block fertiliza- tion by a rapid inhibiting effect on sperm which lose their fertilizing power but retain motility. The loss of fertilizing power is directly pro- portional to both the concentration of extract and the time of exposure. Effect of Sperm Extracts on the Fertilizability of Eggs These experiments were run on jellyless eggs in order to permit di- rect contact between cortex and sperm extract and to eliminate any me- chanical interference with fertilization due to agglutination of the egg jelly. Twenty drops of jellyless eggs from a single female were placed in 10 cc. sperm extracts (.1 per cent to 8 per cent) and at intervals ten 210 JOHN A. FRANK drops of eggs were transferred from each extract to 10 cc. sea water. The ova were fertilized immediately after transfer by two drops of a 1 per cent sperm suspension (made up fresh at the start of the experi- ment and used throughout) and the percentage of eggs developing was counted two hours later. Twenty drops of the same eggs in boiled sea water were treated in the same manner and served as control. TABLE VII The inhibiting effect of sperm extracts on the fertilizability of eggs. Jellylcss eggs added to sperm extracts of various concentrations, transferred to fresh sea water, and immediately inseminated with fresh sperm. Figures give percentage eggs fertilized after exposure to the extracts for the times given at left. Control = same eggs in boiled sea water. Concentration of Sperm Extracts 8% 2% 0.5% 0.1% Control Time Jellyless Eggs Exposed to Sperm Extracts Percentage Eggs Fertilized minutes 15 20 24 15 21 2 57 45 17 21 8 91 90 89 87 66 91 98 96 97 93 100 92 30 60 90. 120 The data of a typical experiment are given in Table VII. After two hours exposure to extracts of above 2 per cent, less than 10 per cent of the eggs are fertilized. At this time less inhibition occurs in the eggs exposed to the .5 per cent extract since 66 per cent are fertilized. The .1 per cent extract equals the controls (91-98 per cent fertilized). We may conclude that the inhibiting effect of sperm extracts on the fertiliza- bility of eggs is due to an effect on the cortex and is directly related to the concentration of extract and length of exposure.2 Effect of Sperm Extracts on Fertilized Eggs and Embryos We have seen that fertilization and development of eggs fertilized in sperm extracts are inhibited when the gametes are exposed to extracts 2 Glaser (1914) stated that eggs which have been treated for two hours with " boiled sperm infusion " remain unfertilized when heavily inseminated. He be- lieved this inhibition to be due to " purple X " since boiled sperm infusions from which the color had been removed did not block fertilization. Woodward (1915) recorded similar results. In the present study it was found that the presence of coloring matter did not appear to influence in any way the inhibiting action of sperm extracts on the fertilizability of eggs. PROPERTIES OF ARBACIA SPERM EXTRACTS 211 during their union in fertilization. The inhibiting effect of sperm ex- tracts also occurs before the union of the germ-cells, involving egg and sperm individually. We will now consider the action of sperm extracts on eggs after normal fertilization. Jellyless and normal eggs were inseminated in sea water, transferred to sperm extracts (8 per cent to .1 per cent) ten minutes after fertiliza- tion, and some of the eggs were transferred hack to sea water after exposure to the extracts for from thirty seconds to one hour. Controls were the same eggs in sea water. The following observations were made : Three hours after fertilization — Normal cleavage in all extracts equal to that of controls. Seven hours after fertilization — Development of eggs in extracts of above 2 per cent concentration stopped at non-motile blastula stage. Healthy motile blastulae found in remaining extracts and controls. Twenty-three hours after fertilization — Embryos in 8 per cent extract have cytolysed at the blastula stage. Rare plutei in the 2 per cent extract with majority cytolyzed at blastula stage. Healthy plutei found in remaining extracts and controls. Throughout the series the ova transferred to sea water developed into normal plutei, indicating that the inhibitory effect of sperm extracts is reversible. In the concentrated extracts normal eggs were inhibited to a slightly greater degree than jellyless eggs due to agglutinated egg-masses preventing escape of the blastulae. When a drop of sperm extract is added to a drop of free-swimming plutei in Chinese ink, the embryos are paralysed at once on entering the extract and fall to the bottom. Peristaltic movements of the gut and feeble ciliary motion are the only indications that the embryos are alive. Under high power the cilia are stuck together by their tips and the plutei are surrounded by clumps of ink particles enmeshed in the agglutinated cilia. Due to the increased stickiness of the cilia, the plutei adhere to- gether in large clusters when a glass needle is moved through the solu- tion. This phenomenon occurs in plutei which have developed from completely jellyless eggs and so is not the result of agglutination of particles of egg jelly by the sperm extract. In attempting to explain these findings it must be realized that ex- tracts of whole sperm undoubtedly contain numerous unknown sub- stances but the action of the egg-agglutinating substance accounts for many of the above facts. We have seen that eggs fertilized in sperm extracts or placed in extracts after fertilization in sea water proceed 212 JOHN A. FRANK normally to the non-motile blastula stage and then abruptly stop devel- oping. It is likely that the cilia, which are pushed forth from the sur- face of the blastula at this time, are immediately agglutinated by the extract with the result that the blastulae are unable to break out of the fertilization membranes and die in situ. When embryos are transferred to sea water from the extract, the cilia are able to function, motility follows, and development continues normally. Paralysis of motile em- bryos by sperm extracts is probably also caused by agglutination of the cilia. PARTHENOGENESIS WITH SPERM EXTRACTS Repeated attempts were made to initiate development in unfertilized Arbacia eggs by exposing them to sperm extracts. Both normal and jellyless eggs were immersed in extracts of various concentrations (.5 per cent to 100 per cent) for from five seconds to twenty-four hours and then transferred to sea water. In other experiments eggs were treated with mixtures of egg-water and sperm extracts, extracts from which the active principles had been removed by boiling or aging, dia- lysed extracts, etc. The results were entirely negative in all instances. It is clear that immersion of eggs in sperm extracts is not sufficient stimulus for development. Perhaps local application of sperm extracts may be effective. DISCUSSION We have seen that there are two major effects of sperm extracts, one upon the surface of the egg and the other a reaction with f ertilizin, which paralleled each other under various experimental conditions. It is not possible to determine whether these effects represent the activity of two distinct substances or two properties of a single substance since the indi- cators for each are different; eggs for the egg-agglutinating substance and fertilizin for the fertilizin-inactivator. It will be assumed that a single " active substance " is present in sperm extracts which reacts with eggs and fertilizin. The question of chief importance relating to this study is whether the active substance functions in fertilization. This substance is extract- able only from sperm, agglutinates Arbacia eggs and embryos but has no effect on Arbacia sperm or blood corpuscles, and combines specifically with fertilizin which is believed to play an important part in fertilization. This marked degree of tissue specificity between eggs, fertilizin, and the active substance suggests that the latter is related to fertilization. Because sperm loses its fertilizing power before loss of motility, Lillie (1915&) postulated that sperm contains a " fertilizing substance" PROPERTIES OF ARBACIA SPERM EXTRACTS 213 which combines with fertilizin and thereby activates the ovum. Such a substance, if present in sperm extracts, must theoretically (1) combine with fertilizin and neutralize its capacity to agglutinate sperm, (2) be lost with the fertilizing power of sperm, and (3) activate unfertilized eggs. Since sperm extracts inactivate fertilizin in quantitative propor- tions and loss of fertilizing power of sperm suspensions parallels loss of the active substance from extracts of these suspensions, two of the above criteria are fulfilled. On the other hand, sperm extracts will not activate eggs, which would tend to disprove the similarity of the fertiliz- ing and active substances. It is possible that the active substance is non-diffusible and hence cannot permeate the cortical membrane, local application is necessary to cause activation, or a penetrating force (sperm) is essential to carry the fertilizing substance into the cytoplasm. Fertilization might then result from injecting sperm extracts into the cortex or increasing the permeability of the vitelline membrane. The block to fertilization caused by sperm extracts is additional evidence against the identity of the active and fertilizing substances. Assuming that fertilizin is necessary for activation, the inhibiting effect of sperm extracts on the fertilizability of eggs may be due to the neutralization of fertilizin at the egg surface by the active substance of sperm ex- tracts. Possibly agglutination of the peripheral layer of the cortex physically prevents sperm penetration. Lillie (1919, p. 219) suggests that there is a high degree of chemical specificity in the union of egg and sperm and Just (1930, p. 331), de- fending Lillie's use of immunological terms in his fertilizin theory, states, "... The biology of fertilization has more points in common with immune reactions than with any other biological phenomenon." The active substance of sperm extracts resembles bacterial agglutinins in certain respects. The following properties of bacterial agglutinins are taken from Wells (1929) and Kolle and Hetsch (1935) : (1) Agglu- tinins disintegrate in sera when preserved, are non-dialysable, pass only incompletely through Chamberland filters, and are not destroyed by ex- tracting serum with lipoid solvents. (2) Electrolytes are necessary for agglutination reactions, which will not occur in distilled water. The cation and not the anion of the added salt is of importance. The reac- tions take place over a wide range of hydrogen-ion concentration (pH 3.7-9.0). (3) The specificity of agglutination reactions is less marked than that of other immune reactions due to the complex structure of cells which permits similar proteins, carbohydrates and lipids to occur in cells of different species ; different species of bacteria will thus react with a single agglutinin. (4) Dead bacteria agglutinate as readily as living ones and the agglutination of living bacteria does not kill them. In each 214 JOHN A. FRANK of the above properties there is a remarkable similarity between the ag- glutination of bacteria by sera and the agglutination of eggs by sperm extracts. Certain differences are also found ; thus all known agglutinins are destroyed at or below 70° C. whereas the egg-agglutinating substance resists boiling for hours. One of the theories concerning the mechanism of agglutination is that the agglutinin reduces the electronegative charge which in part keeps bacteria (or other cells) in suspension, thereby rendering them more susceptible to the precipitating action of salts. We have pointed out that egg-agglutination occurs only in the presence of certain cations (Mg or Ca) and according to Heilburnn (1923), the surface of Arbacia eggs is negatively charged. Possibly the active substance of sperm extracts re- duces this negative charge at the surface of the ovum and permits the cations of the electrolyte to precipitate the eggs. The agglutination membrane would then form by precipitation of dissolved colloids at the periphery of the jelly, which has been " sensitized " to the precipitating action of salts by the active substance. The active substance appears to be related to fertilizin in that both substances are non-dialyzable, thermostable, colloidal, and do not give the usual protein tests. It is likely that the active substance and fer- tilizin belong to a class of chemical compounds related to, if not identical with, bacterial agglutinins. In conclusion, the remarkable tissue specificity of the active substance, its ability to combine selectively with fertilizin, its agglutinating action on eggs and embryos, its relation to the fertilizing power of sperm, and finally, the analogies between immune reactions, fertilization, and the properties of the active substance, all suggest that this substance plays a definite role in the fertilization reaction. • SUMMARY 1. Sperm extracts of Arbacia, prepared by heating and filtering sperm suspensions in sea water, irreversibly agglutinate eggs of the same spe- cies. The reaction is characterized grossly by a dense coagulation of the ova and microscopically by the appearance of an " agglutination mem- brane " at the periphery of the egg. jelly. Sperm extracts agglutinate both jelly and cortex of unfertilized, fertilized, living or dead ova. 2. The " egg-agglutinating substance " is fixed by the jelly and re- moved from solution. It is found only in extracts of sperm and cannot be extracted from any other cells or tissues of the sea-urchin, is not secreted by sperm into sea water, and is not present in extracts of old sperm. The egg-agglutinating substance is colorless, non-dialysable, PROPERTIES OF ARBACIA SPERM EXTRACTS 215 colloidal, highly thermostable, disintegrates on standing, is preserved best at low temperatures, and does not give the usual protein tests. On ex- tracting sperm with hot or cold alcohol-ether, the egg-agglutinating sub- stance is found in the protein residue but not in the lipid extract. Mg and Ca ions are the only ions of sea water which are essential to the extraction of the egg-agglutinating substance, no specific anions taking part in the process. Sperm extracts made in distilled or tap water are inert. 3. Sperm extracts inactivate the sperm-agglutinating power of Ar- bacia fertilizin, the capacity for inactivation varying directly with the concentration of extract. The " fertilizin inactivator " has many proper- ties in common with the egg-agglutinating substance. Loss of fertiliz- ing power from aging sperm suspensions parallels the loss of the egg- agglutinating substance and fertilizin inactivator from extracts of these suspensions. 4. Sperm extracts block fertilization in Arbacia by a direct effect on each gamete : (a) Sperm suspensions in sperm extracts rapidly lose their fertilizing power as measured by their ability to fertilize fresh eggs in sea water. The extracts thus block fertilization by a direct action on the spermato- zoon. (b~) The fertilizability of jellyless eggs is greatly decreased after ex- posure to sperm extracts, which exert an inhibitory effect on the cortex. (5) Arbacia embryos are instantly paralysed and agglutinated by sperm extracts. Eggs which are fertilized in extracts cease developing before the motile blastula stage but this effect is partially removed by transferring the eggs to sea water. 6. All attempts to activate Arbacia eggs by immersion in sperm, ex- tracts have been unsuccessful. 7. It is suggested that the agglutination of eggs and the inactivation of fertilizin by sperm extracts represent two effects of a single active substance extracted from sperm. Theoretical and experimental evidence is presented that this substance probably plays a definite role in the fer- tilization reaction. I am deeply grateful to Dr. L. G. Earth of the Department of Zo- ology, Columbia University, for his invaluable advice and friendly en- couragement during the planning and carrying out of this work. LITERATURE CITED COHN, E. J., 1918. Studies in the physiology of spermatozoa. Biol. Bull., 34: 167-218. 216 JOHN A. FRANK GLASER, O., 1914. On auto-parthenogenesis in Arbacia and Asterias. Biol. Bull., 26: 387-409. GOLDFORB, A. J., 1929. Variation of normal germ cells. Studies in agglutination. Biol. Bull., 57 : 333-349. HAAS, A. R., 1916. The effect of the addition of alkali to sea water upon the hydrogen ion concentration. Jour. Biol. Chan., 26: 515-517. HEILBRUNN, L. V., 1923. The colloid chemistry of protoplasm. I. General con- siderations. II. The electrical charges of protoplasm. Am. Jour. PhysioL, 64: 481-498. IRVING, L., 1926. The precipitation of calcium and magnesium from sea water. Jour. Mar. Biol. Ass'n, N.S., 14: 441-446. JUST, E. E., 1922. The effect of sperm boiled in oxalated sea-water in initiating development. Science, N.S., 56 : 202-204. JUST, E. E., 1928. Methods for experimental embryology with special reference to marine invertebrates. Collecting Net, 3 (No. 2) : 7. JUST, E. E., 1930. The present status of the fertilizing theory of fertilization. Protoplasms, 10: 300-342. KAPP, E. M., 1928. The precipitation of calcium and magnesium from sea water by sodium hydroxide. Biol. Bull., 55 : 453-458. KOLLE, W., AND H. HETSCH, 1935. Experimental Bacteriology in its Applications to the Diagnosis, Epedimiology, and Immunology of Infectious Diseases. Seventh German Edition. Revised by J. Eyre. Vol. 1, 592 pp. Mac- millan, New York. LILLIE, F. R., 1913. Studies of fertilization. V. The behavior of the spermatozoa of Nereis and Arbacia with special reference to egg-extractives. Jour. Expcr. Zool., 14: 515-574. LILLIE, F. R., 1914. Studies of fertilization. VI. The mechanism of fertilization in Arbacia. Jour. Expcr. Zool., 16 : 523-590. LILLIE, F. R., 1915a. Sperm agglutination and fertilization. Biol. Bull., 28 : 18-33. LILLIE, F. R., 19156. Studies of fertilization. VII. Analysis of variations in the fertilizing power of sperm suspensions of Arbacia. Biol. Bull., 28 : 229- 251. LILLIE, F. R., 1919. Problems of Fertilization. The University of Chicago Press, Chicago. LOEB, JACQUES, 1913. Artificial Parthenogenesis and Fertilization. The Univer- sity of Chicago, Chicago. MATHEWS, A., 1897. Zur chemie der spermatozoen. Zcitschr. f. physiol. Chemie, 23: 399-411. PETERS, J. P., AND D. D. VAN SLYKE, 1932. Quantitative Clinical Chemistry. Vol. II. Methods. Williams and Wilkins Co., Baltimore. PIERI, J.-B., 1899. Un nouveau ferment soluble : L'ovulase. Arch, de Zool. Exper. ct Gen., 7: (Notes et revues) xxix-xxx. ROBERTSON, T. B., 1912. On the isolation of oocytase, the fertilizing and cytolysing substance in mammalian blood-sera. Jour. Biol. Chcm., 11: 339-346. SAMPSON, MARY M., 1926. Sperm filtrates and dialyzates. Their action on ova of the same species. Biol. Bull, 50: 301-338. WELLS, H. G., 1929. The Chemical Aspects of Immunity. Second Edition. Am. Chem. Soc. Monograph Series, No. 21. Chemical Catalogue Co., New York. WOODWARD, ALVALYN E., 1915. Note on the nature and source of " purple X." Biol. Bull, 29 : 135-137. EFFECT OF CERTAIN BACTERIA ON THE OCCURRENCE OF ENDOMIXIS IN PARAMECIUM AURELIA1 ARLENE JOHNSON DE LAMATER (From the Zoological Laboratory, Johns Hopkins University) It was at first believed that endomixis occurs in Paramecium aurclia at rather regular intervals, but detailed observations have shown that there is much variation in the length of the interendomictic period. Different stocks show differences in the average interval between endo- mixes: thus Sonneborn (1937) showed that, under identical conditions, in the stock R the mean interval between endomixes was 18.3 days (66.3 fissions), while in Woodruff's long-lived stock (W) the mean interval was 46.3 days (110.4 fissions). Environmental conditions have been shown to alter the interendomictic period. Woodruff (1917) showed that sudden alterations in the cultural conditions bring on endomixis earlier than it would otherwise occur. Jollos (1916), by various changes in temperature and bacterial content of the culture medium, caused great changes in the length of the interendomictic period, from a minimum of 3 days to a period sufficiently great for 168 fissions to occur. Young (1917) increased the frequency of endomixis in Paramecium aurelia by alterations in the age of the culture medium and by changes in tempera- ture. In Paramecium caudatum, Chejfec (1930) showed that the per- centage of individuals undergoing endomixis in mass cultures was greater in a medium containing the bacillus coli than in hay infusion. In the following study of Paramecium aurelia the frequency of en- domixis and length of the interendomictic period were compared in cul- ture media containing diverse bacteria, in different concentrations. Spe- cial attention was directed to the question whether it is primarily the unfavorable nature of conditions that brings on endomixis or increases its frequency. A branch of Woodruff's long-lived stock of Paramecium aurelia was employed in the experiments. All lines of descent were derived from a single individual which underwent endomixis at room temperature during the first five-day period of the experiment (see Fig. 1). There- after all cultures were kept in a constant temperature box at about 25° to 26° C., except during the daily transfers. 1 I wish to thank Dr. H. S. Jennings and Dr. T. M. Sonneborn for their help and criticism. 217 218 ARLENE DE LAMATER Five types of bacteria were employed in the different culture media, namely Flavobacterium brunneum, Bacillus niger, B. cereus, B. pro- digiosus, and B. coli. FIVE DAY PERIODS I 5 10 15 20 i i i i i i i i i i i i i I I i I i i I i i i FLAVOBACTER I UM I I I I 100 FLU NIGER H II 1 1 I CEREUSH 2 I I lOONIGIER I 3 12 PROD.H 100 CEREUS COLI H COLI I I III IOOFLAVOBACTERIUM I FL. I NIGER I I 2 7 3 I I 1 CEREUS I 21 II FL. I I I I I FL. PROD. I I I FL. I FIG. 1. Diagram of the course of the experiments. Each interval from left to right represents a five-day period. The horizontal lines show the number of five- day periods through which each culture (of 12 lines each) continued. The small numbers below the horizontal lines in certain periods give the number of lines (out of the 12) that underwent endomixis in that period. The heavy vertical lines show the source of the cultures; thus the culture " Niger I " was derived from the Flavobac- terium culture at the end of its fourth five-day period : it continued through the seventeenth five-day period. Flavobacterium, Flavobacterium brunneum; Niger, Bacillus niger; Cereus, B. cereus; Prod., B. prodigiosus; Coli, B. coli. Designations preceded by 100 indicate cultures in which the concentration of bacteria was 100 times as great as normal ; designations not preceded by 100 indicate the normal concentration. I and II indicate Groups I and II, mentioned in the text. BACTERIA AND ENDOMIXIS IN PARAMECIUM 219 The standard culture medium was that commonly employed in this laboratory. It consists of infusion of desiccated lettuce leaves, into which is introduced the green alga Stichococcus bacillaris, in the pro- portion of three 2 mm. loops of the alga (from 20 to 22-day-old slants) to 20 cc. of the filtered fluid. To this is added a 1 mm. loop of Flavo- bacterium bmnneum from a one-day-old agar slant. (For details see Sonneborn, 1936.) Since Poramcciwn aurelia flourishes well in this medium, it was employed as a norm or standard for comparing the results with the other bacteria. For the other bacteria the medium was prepared in the same manner, with the substitution of equal amounts of the other bacteria for Flavo- bacterium bmnneum. The medium is buffered to prevent the produc- tion of different pH by the diverse bacteria. In some of the experi- ments a nutritive medium was employed in which the quantity of bac- teria was 100 times as great per unit volume as in the normal or standard medium. The purpose was to determine the results of preponderance of bacteria of a certain type, so that it was not considered essential to carry on the experiments under purely aseptic conditions ; this would, for experiments of this nature, be practically impossible. Tests were made at intervals to determine whether there was marked contamination. No cross-contamination of the different bacterial types employed was ob- served in any case. There were three types of common laboratory contaminants; when these became at all abundant in the cultures (up to 30 per cent) such cultures were rejected. For determining how far the effects of certain bacteria in certain concentrations are harmful or unfavorable, the following criteria were used. An unfavorable effect may be shown by (1) a slowing of the fission rate, (2) an increase in the proportion of deaths occurring, (3) incomplete division, resulting in the production of partially united chains of individuals. Accordingly, observations on these points were recorded for each different type of culture (see Table I). The occurrence of endomixis in the different media was detected by the daily staining of samples from each line. The general plan of the experiments is shown in Fig. 1, on which is likewise indicated the number of lines in which endomixis occurred in particular five-day periods. In Table I are summarized the general results of culture in the different media. The following account of results is to be read in connection with the figure and table. From a single individual that had just undergone endomixis, twelve lines were developed as quickly as possible, by cultivation in the normal Flavobacterium medium described above. This culture of 12 lines in Flavobacterium was maintained throughout the 22 five-day periods of 220 ARLENE DE LAMATER TABLE I A summary of the effects of different bacterial media upon endomixis, the number of deaths and of chain-formation in Paramecium aurelia Group Tota num- ber of line- days Mean fis- sions per line per five- day perioc Total num- ber oi death Num- ber oi death per 100 line- days Total num- ber of chains formec Num- ber of chains per 100 line- days Tota num ber of lines in endo mixis Num- ber of lines in endo- mixis per 100 line- days Flavobacterium brun neum Normal Concentration .... 100 Times Normal (Group I) . . 1232 748 10.38 5 27 10 31 .81 4 14 2 26 .16 3 48 5 3 .41 40 Group I Returned to Normal 360 9 13 3 83 3 .83 3 .83 100 Times Normal (Group II) . . 236 609 9 3.81 9 3.81 0 0 Bacillus niger Normal Concentration (Group I) . . 746 8.23 6 .80 0 0 23 3.08 Group I Returned to Flavobacterium 291 11.92 0 0 0 0 0 0 Normal Concentration (Group II) 590 8.91 9 1.53 0 0 18 3.05 100 Times Normal 240 9.82 3 1.25 0 0 8 3.33 Bacillus cereus Normal Concentration (Group I) 750 8.69 4 .53 0 0 11 1.47 Group I Returned to Flavobacterium 220 8.88 0 0 0 0 0 0 Normal Concentration (Group II) 297 8.93 3 1.01 0 0 12 4.04 100 Times Normal 225 4.61 7 3.11 7 3.11 0 0 Bacillus prodigiosus Normal Concentration (Group I) . . 77? 8.88 7 .91 7 .91 2 .26 Group I Returned to Flavobacterium 207 8.48 0 0 2 .97 0 0 Normal Concentration (Group II) .. 197 9.42 4 2.03 4 2.03 1 .51 Bacillus coli Normal Concentration (Group I) . . 658 9.43 15 2.28 15 2.28 4 .61 Normal Concentration (Group II). . 359 8.92 8 2.23 10 2.79 0 0 BACTERIA AND ENDOMIXIS IN PARAMECIUM 221 the experiments; it is represented by the upper horizontal line in Fig. 1. Cultures employed in the experiments with other media were branched off from this normal culture in particular five-day periods, as shown in Fig. 1. In each of the different media a group of 12 lines was kept in progress during a number of five-day periods, as indicated in the figure. The number of " line-days " given in column 2 of Table I is the number of lines of descent (usually 12, but with occasional irregularities), mul- tiplied by the number of days during which the lines were in progress. In some cases two different groups of 12 were tested at different times in a particular medium; these are designated as I and II in the figure and table. The main features of the diverse experimental cultures are as follows. FLAVOBACTERIUM BRUNNEUM : NORMAL OR STANDARD CONCENTRATION The 12 lines in this medium were kept in progress for 22 five-day periods, which covered the entire time of the experimental cultures. Total number of line-days, 1232. Mean fission rate, 10.38 per line per five-day period. Endomixis had occurred at the beginning of the cul- ture ; it occurred again in but four of the lines, in periods 16, 18, 20 and 21 (Fig. 1) — after intervals respectively of 75, 87, 97 and 103 days (168, 180, 212 and 224 generations). The remaining 8 lines did not undergo endomixis during the rest of the 110 days of the culture. The interendomictic periods are thus very long (as was before known to be the case in this stock). Thus, as the table shows, the total number of lines found to be in endomixis during this part of the experiment was but 5 (including that in the first generation), so that the number of endomictic lines per 100 line-days was but 0.41 (Table I). In this medium the total number of deaths was but 10, or in the proportion of 0.81 per 100 line-days. Only 2 cases of united chains (incomplete division) occurred, thus but 0.16 per 100 line-days. High concentrations of Flavobacterium. — At the beginning of the seventh 5-day period a group of 12 lines (Group I) was branched off and cultivated in the same medium, but with the concentration of bac- teria 100 times as great. They were continued here till the end of the 19th period. As Table I shows, this decreased the fission rate to about one-half its former rate (5.27 in place of 10.38 per five-day period) : multiplied the death rate by five and the proportion of incomplete fissions (chains) by about 22. Furthermore, the size of the individuals was notably decreased, and the contained crystals became larger and more numerous. Thus this high concentration must be considered distinctly unfavorable as compared with the normal. Seven days after the trans- ARLENE DE LAMATER fer (in period 8, Fig. 1) two of the 12 lines underwent endomixis. A third line was in endomixis in the seventeenth five-day period : almost simultaneously (in periods 16 and 18) two lines were in endomixis in the normal concentration, so that no special significance can be attributed to this. Thus the change to the unfavorable high concentration apparently induced endomixis in 2 of the 12 lines. At the fourteenth period a new group of 12 lines was branched off from the normal concentration to the concentration 100 times as great (100 Fl II, Fig. 1) : it was continued through 4 periods. The unfavor- able effects were shown about as before (Table I), but endomixis wras not induced. Restoration to normal concentration after culture at the high con- centration:— In the case of Group I cultivated at the hundred-fold bacterial concentration, after seven five-day periods (at the 14th period of the experiment as a whole), a group of 12 lines was restored to the more favorable normal concentration. Fission rate was restored to nor- mal, the proportion of deaths and of chains was greatly decreased (Table I). The normal size of the animals was restored after five days. Endomixis occurred in 3 of the lines in periods 17 and 18, but this cannot be attributed to the change to the normal concentration, since in the 100-fold concentration, as well as in the original culture in normal conditions, lines were in endomixis at practically the same times (in periods 16, 17, 18, Fig. 1). BACILLUS NIGER At the beginning of period 5, and again at the beginning of period 13, derivatives of the normal control group in Flavobacterium were washed and transferred to a medium containing Bacillus niger in place of Flavobacterium (see Fig. 1, " Niger I " and " Niger II "). The niger medium roughly corresponded in concentration to that of Flavobacterium (one 1-mm. loop to 20 cc. of culture fluid). In both cases there was a sharp lowering of the fission rate in the first period after transfer, but this was soon partially recovered, the average fission rate being 8.23 (Group I) and 8.91 (Group II) as compared with 10.38 in the entire Flavobacterium culture medium, so that these differences are hardly significant. As Table I shows, none of the other criteria indicate that B. niger is less favorable than Flavobacterium. Yet endomixis was at once induced in B. niger — in 7 lines in Group I, in 11 lines in Group II (Fig. 1). In all there were 23 lines in endomixis in Group I of B. niger, 18 in Group II — as compared with but 5 in the entire course of the Flavobacterium culture. The number of endomictic lines per 100 BACTERIA AND ENDOMIXIS IN PARAMECIUM line-days was in B. niger a little over 3, as compared with 0.4 in Flavo- bacterium (Table I). In Group I, 9 lines underwent a second endo- mixis while in B. nigcr, and the interendomictic interval was short, ranging from 24 days (44 generations) to 48 days (79 generations), as compared with 75 to 103 days in Flavobacterium. Similar relations are found in Group II of B. niger. It appears clear, therefore, that culture in B. niger greatly increases the frequency of endomixis, and that this is not a result of unfavorable conditions, for B. niger appears otherwise as favorable as Flavobac- terium. When derivatives from Group I in B. niger were restored to Flavo- bacterium, after 40 days in B. niger, there was after a few days a rise in fission rate. No endomixis occurred in those thus restored to the normal medium. High Concentrations of B. niger When individuals of Group II in B. niger were transferred to a concentration of B. niger 100 times as great as normal, there was no appreciable change in fission rate, mortality, or in number of imperfect divisions. Endomixis, however, was increased, 8 lines undergoing en- domixis in the 4 periods of this culture (Fig. 1). This agrees with the fact previously noted, that the presence of B. niger in the culture medium increases the frequency of endomixis, and that this increase is not the result of the unfavorableness of B. niger. BACILLUS CEREUS Two groups of 12 lines were transferred from Flavobacterium to a normal concentration of B. cereus, at periods 5 and 13 of the experiment (Fig. 1). Fission rate was very slightly decreased, and the animals were on the average a little smaller in size than in Flavobacterium, but mortality did not increase and there were no imperfect fissions, forming chains. Thus this bacillus can hardly be considered appreciably harmful. Yet, as in B. niger, endomixis was definitely increased in frequency. There were eleven endomixes in the 12 lines of Group I, 12 in those of Group II, while in the same periods the lines in Flavobacterium under- went but 4 endomixes. Seemingly B. cereus is less effective than B. niger in producing endomixis, since none of the lines in B. cereus under- went a second endomixis, while a number of those in B. niger did so, though the number of periods was the same for the two types of culture. Restoration of lines of B. cereus culture to Flavobacterium did not result in the production of endomixis. 224 ARLENE DE LAMATER High Concentration of B. cereus Derivatives of Group II in normal B. cereus were transferred at the nineteenth period to a concentration of B. cereus about 100 times as great (Fig. 1). This decreased the fission rate by about one-half, in- creased the mortality, and the formation of abnormal chains, and caused the production of large crystals in the cytoplasm. Conditions were clearly unfavorable, but endomixis was not induced. BACILLUS PRODIGIOSUS At periods 5 and 13, two groups (I and II) were derived from the normal control cultures in Flavobacterium, and were cultivated in the normal concentration of B. prodigiosus. Fission was slightly lowered and the number of abnormal chains slightly increased (Table I) ; also a few individuals without macronuclei were observed. Thus the medium containing this organism must be considered slightly unfavorable. No increase in the frequency of endomixis was produced. BACILLUS COLI Groups I and II, derived in the usual way from the Flavobacterium cultures at the beginning of periods 9 and 13, showed little change in fission rate, but there was a considerable increase in the proportion of deaths, and of abnormal divisions (chains). Also numbers of indi- viduals without macronuclei occurred. In Group I, four cases of typi- cal endomixis occurred (Fig. 1). In addition, there occurred in most of the lines irregular fragmentation of the macronucleus. This began in both groups a few days after the transfer to B. coll, and continued for about 10 days. The fragments varied from day to day in size and number. They were scattered through the cytoplasm, or sometimes were in pockets of the macronucleus. The macronucleus was frequently present in two pieces ; often it exhibited signs of activity, in the pos- session of projections of various sizes. Some of the stages closely resembled figures presented by Diller (1936, p. 37) for the process of hemixis in his types B and C. This fragmentation was in most cases not followed by the typical climax stage of endomixis, though it was in one of the lines. In other cases, after the disappearance of the frag- ments, the macronucleus remained normal. Such fragmentation oc- curred only in media containing Bacillus coli. SUMMARY 1. In the long-lived stock of Woodruff, cultivated in a standard medium containing a normal concentration of Flavobacterium brunneutn, BACTERIA AND ENDOMIXIS IN PARAMECIUM 225 endomixis occurred only at very long intervals, of 75 to 109 days, or more. Increase in the quantity of Flavobacterium to 100 times the normal concentration made the conditions very unfavorable, but hardly increased the incidence of endomixis. 2. Substitution of Bacillus niger for Flavobacterium, in the same proportions, brought on endomixis and increased its frequency, although it did not make the conditions unfavorable. Increasing by 100-fold the concentration of B. niger caused a further increase in the frequency of endomixis, but still without making the medium unfavorable. 3. Bacillus cereus was not appreciably unfavorable in its action on Paramecium, yet like B. niger it increased the frequency of endomixis. Its tendency to induce endomixis appears to be slightly less than that of B. niger. Increasing the concentration of B. cereus 100-fold made conditions very unfavorable, but did not increase the tendency to endomixis. 4. Bacillus prodigiosus is slightly unfavorable as compared with Flavobacterium, but it does not increase the frequency of endomixis. 5. Bacillus coli is unfavorable as compared with Flavobacterium. It produces a tendency to fragmentation of the macronucleus, usually not followed by complete endomixis. 6. Thus in these experiments there is no correlation between the unfavorable effect of certain bacteria and their tendency to produce endomixis. Certain species (Bacillus niger and B. cereus} are not unfavorable to Paramecium aurelia yet (in this stock at least) they much increase the frequency of endomixis. Certain other conditions that are unfavorable (particularly high concentrations of Flavobac- terium) do not increase the frequency of endomixis. LITERATURE CITED CHEJFEC, M., 1930. Zur Kenntnis der Kernreorganisationsprozesse bei Para- maecium caudatum. Arch. f. Protist., 70: 87-118. DILLER, W. F., 1936. Nuclear reorganization processes in Paramecium aurelia, with descriptions of autogamy and " hemixis." Jour. Morph., 59: 11-67. JOLLOS, V., 1916. Die Fortpflanzung der Infusorien und die potentielle Unster- blichkeit der Einzelligen. Biol. Centralbl, 36: 497-514. SONNEBORN, T. M., 1936. Factors determining conjugation in Paramecium aurelia. I. The cyclical factor: the recency of nuclear reorganization. Genetics, 21 : 503-514. SONNEBORN, T. M., 1937. The extent of the interendomictic interval in Paramecium aurelia and some factors determining its variability. Jour. Exper. Zool., 75 : 471-502. WOODRUFF, L. L., 1917. The influence of general environmental conditions on the periodicity of endomixis in Paramecium aurelia. Biol. Bull., 33 : 437-462. YOUNG, R. T., 1917. Experimental induction of endomixis in Paramecium aurelia. Jour. Exper. Zool., 24 : 35-53. « ^fc CONTRIBUTIONS TO THE STUDY OF DEVELOPMENT OF THE WING-PATTERN IN LEPIDOPTERA WERNER BRAUN (From the Department of Zoology, University of California, Berkeley, California) INTRODUCTION The mechanism of pattern determination in Lepidoptera wings has long been a favorite model for the study of gene action. Every study of developmental processes is concerned with the study of development of manifold structure, which finds its best object in the study of pattern, e.g. the pattern of a butterfly-wing. Goldschmidt tried the first success- ful analysis of the factors responsible for the development of this pattern and expressed it in his theory of different velocities of developmental processes. Although most parts of his analysis could be experimentally checked and extended by Kiihn and his school, the decisive observations ("relief-stage") of Goldschmidt's work were either not found, mis- interpreted, or doubted by several investigators. An additional and more extended investigation of those stages was therefore necessary and will be furnished in this paper.1 PREVIOUS EXPERIMENTS ON THE DEVELOPMENT OF THE WING-PATTERN Goldschmidt's Investigations In his book " Physiologische Theorie der Vererbung " (1927), Gold- schmidt contributed an extensive analysis of the pattern problem. A study of the development of the individual wing furnished the most decisive point in this analysis. Goldschmidt (1920, 1923) ob- served that the later wing-pattern is already completely represented in the young pupal wing, long before pigmentation takes place. It can be made visible by detaching unpigmented pupal wings from the body and allowing them to dry. After a short time a relief was visible on the former homogeneous-looking wing. Later white scales became erected and formed the raised regions and later dark scales collapsed and formed the depressed regions. The whole relief represented the later wing-pattern in every detail. He observed this " relief-stage " in Platysamia cecropia L., Lymantria, Thais 1 Acknowledgment for help and advice is due to Professor R. Goldschmidt and to the United States Government for assistance rendered through its Works Progress Administration Official Project No. 465-03-3-192. 226 DEVELOPMENT OF WING-PATTERN 227 and several other species. Goldschmidt assumed that in certain parts of the pattern the scales are still soft bags filled with blood, which collapse when dried out. In other districts, however, the scales are already chitinized and do not collapse when dried out. Because the later dark scales are consequently softer at this point of development than the later white scales, he concluded that later dark scales exhibit a slower development than later white scales. He arrived at the conclusion that the different velocities of development for the different parts of the wing were primarily responsible for the formation of wing-pattern. According to his theory the different velocities of development start with the mosaic of the epidermis cells and their product, the scales which develop with different speed in different parts of the wing. A section through the young pupal wing in a certain stage would accordingly show us different parts of the wing in different stages of development. At the same time different chemical substances will be present in the body, which if deposited in the scales will determine the color. These substances, like tyrosine, carotine etc., can be end-products of metabolic processes and a special production of them for the coloration-processes is not necessary. The deposition of these substances in the scales can be dependent on a certain colloidal stage of the chitin. If at a certain stage in the development tyrosine is present, it will only be deposited in those scales which represent a certain condition of the chitin at this moment, which means only a certain part of the pattern. Other parts of the pattern will react according to their stage of chitin-development and the pigmentation substances present. The development of a very complicated pattern can thus be explained by a very simple mechanism. Goldschmidt (1920, 1927) then tries to trace these processes back in development and discusses the determination-points for these differences in developmental velocities (" sensitive period," determination-stream, Liesegang phenomenon). The Investigations of Kit Jin and His Co-workers Kiihn and his co-workers tried to attack the same problems and furnished further experimental material. They worked with the flour- moth, which shows a pattern consisting of dark and light pattern-ele- ments which are partly symmetrical (see Fig. 1). The various parts of the pattern do not only contain different amounts of pigment but the shape and size of the scales in various districts differ also. Therefore, the determination of the pattern is here not only a process of distributing different amounts of pigment but a morphogenetic process as well. Feldotto (1933) determined the sensitive periods for this pattern and showed that heat-treatment at the pupal age of 12-72 hours was WERNER BRAUN able to modify the development of the different parts of the pattern. Each system of the pattern has its particular time of maximal modifica- tion at different stages of development. Ktihn and von Engelhardt (1933) cauterized the young pupal wing and thus were able to show that a determination-stream spreads over the wing from 48-60 hours of pupal age. A period of particularly distributed cell divisions can be considered as the first reaction to the process of determination. These cell divi- sions were first described by Kohler (1932). He reported that from 36-84 hours of pupal age a great number of cell-divisions are visible which have their maximal distribution in districts of subsequent dark pattern elements while subsequent light pattern districts show few cell- divisions. The distribution of cell-divisions therefore resembles the subsequent pattern of pigmentation and Kohler consequently called it mitosis-pattern. A mitosis-pattern seems to be the predecessor to the later wing-pattern. Braun (1936) examined the mitosis-pattern more closely with special regard to its role in the development of the wing-pattern. He showed that this mitosis-pattern is visible from 36-148 hours of pupal age. Subsequent dark pattern-districts always represent maxima of mitotic divisions, while the minima are in later light pattern-districts during this period. The mitosis-pattern spreads in a wave-like fashion over the wing from the proximal base to the distal margin. Two such mitosis waves are distinguishable during the period of mitotic division. The first one goes over the wing from 36-84 hours, and represents cell- divisions of hypodermis-cells and the first division of the scale-building cells. The second wave starts at 84 hours and ends at 148 hours and represents the second division of the scale-building cells. When the first proximal divisions occur we find no mitosis in distal parts of the wing, with the exception of the cell-divisions of the later marginal scales, which start developing earlier and continue at a faster rate throughout the whole wing-development. Furthermore, it could be shown that this mitosis-pattern is the first sign of a completed determination of the wing-pattern. Then an attempt was made to investigate the role of this mitosis- pattern in the development of the wing-pattern. It could be shown that the districts of mitosis-maxima (presumptive dark districts) repre- sent districts of greater intensity of cell-division. Throughout the stages of development following the mitosis-period, later dark districts show a greater number of cells per unit than later white districts. The size of cells in the presumptive dark districts is accordingly smaller. This observation could be checked on the adult wing also, where one finds ^•j DEVELOPMENT OF WIXG-PATTEKX 229 more scale-building cells per unit in dark districts than in light districts. In addition it was found that all scales grow out simultaneously. From these observations it was concluded that a different intensity of cell- j division was primarily responsible for the formation of the pattern. The principle of different velocities of development was discarded in favor of the principle of different intensity of cell-division as the pri- mary factor in the development of a pattern. Kohler and Feldotto (1937) again tried to discard the existence of different developmental velocities for Lepidoptera in a recent paper. Their argument is based on their failure to find the scales of different districts in different stages of growth during experiments on scale-development. Furthermore, it had not been possible to observe the relief stage in Epliestia. In response to these views expressed by the E/»/;(\vf /(/-workers. Gold- schmidt showed in recent publications (1938) that the observations made on Epliestia are in perfect harmony with his discoveries. The fact that we find more cell-divisions in certain districts during the mitosis-period means that the cells in these districts have not yet reached the end of their multiplication period. These districts are the later dark districts, they show a retarded development not only in the relief-stage but at this early stage as well. The cells in later white districts show less cell-divisions because they have reached the end of their multipli- cation period earlier and therefore differentiate faster than the later dark cells. (An investigation concerning the number of divisions which a cell undergoes in later dark or in later white districts would probably prove this explanation. Such an investigation is now under way.) The observation of the simultaneous growing out of scales does not interfere with this theory, because the different speed of differentiation of the scales in the different districts does not necessarily mean different growth. The process of growth might be nearly identical for all wing areas, while the different speed of development should be noticeable by differences in the histological structure of the scales, which would lie hard to examine at this early stage. To test these views it was necessary to undertake a more detailed investigation of the relief stage, where the different velocity of develop- ment for the different pattern districts was demonstrated so clearly. Furthermore, an attempt had to be made to find additional evidence for the different speed of development at stages previous to the relief stage. The relief-stage had to be found in Eplicstia, where the mitosis-period had been found. Finally the mitosis-period had to be observed in other Lepidoptera, where the relief-stage had been demonstrated previously. The results of these investigations, which fully confirmed the validity of Goldschmidt's views, are described in the following section. 230 WERNER BRAUN NEW INVESTIGATIONS ON THE DETERMINATION OF THE \YING-PATTERN Material and Methods Ephcstia kitliniclla Zeller, Platysauiia cccropia L. and Papilio a/a.v L. were the objects of the experiments to he described. Ephcstia cultures were kept according to the method described by Kiihn and Henke. Cccropia and Papilio pupae were obtained in the fall and stored in a refrigerator, from which they were put into an incubator at any time desired. They started development as soon as they were placed in a warmer temperature. All pupae developed at 25°. Fifty pupae of Platysauiia cccropia, fifty pupae of Papilio and approximately one hun- dred and fifty pupae of Ephcstia were used for these investigations. The Appearance of the Relief Stage As pointed out before, Ephestia-workers have never been able to find the stage in the pupal-development which is decisive for the theory of different developmental velocities, namely the relief stage. There- fore the first step was to hunt for this decisive stage in the development of the flour-moth. In order to become acquainted with the stage in question, the author repeated Goldschmidt's (1920) experiments on the Cccropia moth and Papilio. The observations confirmed Goldschmidt's results completely. Platysainia cccropia L. — The wings were removed from the pupae at a time when no pigmentation had yet taken place, the best time being one to two days before pigmentation. They were then placed on a slide and allowed to dry. A relief appeared which showed subsequent white parts erect, subsequent dark parts collapsed. This stage is rela- tively short, approximately 3-5 per cent of the whole pupal stage, and was observed on ten Cccropia wings. The relief of the Cecropia-w'mg is especially clear in the distal parts where the dark and yellow half-moon is located on the pigmented wing. (For excellent photos of this stage see Goldschmidt's 1920 paper.) Papilio aja.v L. — The same stage was easily found in Papilio aja.r. Here too the relief-stage takes 3—5 per cent of the complete time of pupal development. It is very clear all over the wing, corresponding to the mature wing-pattern, and appears approximately twenty minutes after the wing has been put on a slide for drying. The subsequent light parts of the wing are indicated by clearly erected parts, the subsequent dark pigmented districts of the wing are all collapsed. This stage could be observed in fourteen wings which were removed from the pupae one to two days before pigmentation started. The different unpigmented pat- tern districts of the wings were visible even without drying. If the light DEVELOPMENT OF WING-PATTERN 231 was shone in a certain way on the still attached and living wing, the dis- tricts of subsequent dark pigmentation and light pigmentation were clearly distinguishable by what appeared to be a difference of consistency of the areas in question. This difference can be demonstrated even more clearly, if one puts the wing into water. Again the districts of subse- quent different pigmentation can be easily distinguished on the completely unpigmented wing. The action of two time factors for this period could be observed in Papilio. (1) At the moment pigmentation started, the relief disappeared in pigmented parts, while unpigmented parts showed the relief distinctly. (2) It is known that the hind wing always shows a faster development than the fore wing (e.g. pigmentation). Conse- quently the relief was already present on the hind wing, while all scales were still soft on the fore wing. The relief of the hind wing disap- peared earlier because pigmentation set in earlier. (Pictures of the relief stage in species closely related to Papilio can be found in Gold- schmidt's 1923 paper.) Ephestia k. — The corresponding stage was found without much dif- ficulty in Ephestia at eight to nine days of pupal age. That it had never been noticed before is surprising. The wing of the pupa was easily detached after the surrounding chitin was broken and carefully removed. Then the wing was placed on a slide and allowed to dry. After about fifteen minutes the two later dark bands appeared clearly as collapsed parts on the dried wing (Figs. 1 and 2). The action of an additional time-factor could be observed on these wings. They showed that the relief-stage proceeds like many other processes in the development of the flour-moth (mitosis-pattern, pigmentation) in a wave-like fashion from the proximal base to the distal margin of the wing. In early stages the relief of the proximal band only was visible, while in distal parts all scales were collapsed (Fig. 2 a). A little later the whole wing showed the relief, scales both of the proximal and distal bands being collapsed (Fig. 2 &). The marginal scales were an exception. They always ex- hibited a faster development, first noticeable during the mitosis period, and they were already erected in distal " white " parts in the early relief- stage. As in Papilio, the districts of subsequent different pigmentation were visible on the undried wing if the light was reflected in a certain way, as well as when the wing was put into water. These observations were based on a study of approximately sixty pupae. The Cliitinization of Scales during Pupal-development as Proof of the Different J7elocity of Development Sections. — The fact that no regional differences in the speed of out- growing scales could be detected has been mentioned before (Kohler, 232 WERNER BRAUN 1932; Braun, 1936; Kohler and Feldotto, 1937). It was necessary to make sections of the wings in the relief stage to see if erect scales show a morphological difference at this stage. Such sections were made from FIG. 1. Pigmented pupal wing of Eplicslin kiihniella. Pi, Pn proximal dark bands, Di, Dn distal dark bands, Mi, Mil middle-spots. FIG. 2(7-6. Relief-stage of the pupal wing of Eplicstia kiihmclla. a, early relief-stage showing the proximal band clearly, />, later relief-stage showing proxi- mal band, distal band and five marginal spots. X 9. FIG. 3. Papil'w uja.r. X -/3. DEVELOPMENT OF WING-PATTERN 233 eight Cccropia wings (through the district of the " half-moon "), from twelve EpJicstia wings, and six Papilio wings, but no morphological differences in the stage of development of collapsed or erect scales of different forms were noticed. In many Lepidoptera white and dark areas are composed of scales of different forms, correspondingly later dark parts showed larger scales on the Cccropia wing than later white parts. Consequently the different speed of differentiation of the scales in the different pattern districts does not manifest itself by different stages of growth. Growth seems to be the same for all parts of the wing. The difference therefore has to be sought in different velocities of the chitini- zation process of the later white and dark scales at this stage as sug- gested by the soft condition of later dark areas. Chemical Tests. — In fact, such a difference could be detected by chemical reactions. There is not much known about the chemistry of chitin, but P. Schulze ( 1922 ) described a differential reaction for hard and soft chitin. The object is first saturated with iodine for a short time, blotted with filter paper, and finally covered with a solution of ZnCL. The reaction will show very soon but disappears after some time. Hard chitin will show a violet-brown color, soft chitin, a light brown. Ten unpigmented Cccropio wings were treated accordingly and showed a distinct reaction. Subsequent white parts appeared dark ; subsequent dark parts appeared light. This experiment exhibits clearly that the different parts are found in different stages of chitinization. The same experiment was repeated with six pupae of Papilio and twenty pupae of Eplicstia and always resulted in a clear pattern negative. Even the sections which showed no morphological difference reacted distinctly to this chitin test. The chitin reaction worked during the whole relief stage. It would have been desirable to check this test by making the same experiment with an adult pigmented wing. In this case all scales should show the violet-brown reaction. However, it was not possible to extract the pigment from an adult wing, which would have been neces- sary in order to obtain a bleached wing on which the reaction would be visible. Several other chitin reactions were tested, but none of them gave as clear a result as did the iodine-zinc reaction. Test with Polarized Light. — After it had been proven that the dif- ferent speed of development actually can be shown to be the different velocity of chitin hardening, an experiment to test the refraction of the differently chitinized scales was tried. Adult scales are doubly refract- ing, and it was hoped that one could find a stage where one kind of scales was singly refracting. Scales from different regions were investigated under crossed Nicols, but no clear results were obtained. It was men- 234 WERNER BRAUN tioned before that the scales from different regions have different forms, and this difference may add to different refraction effects. Tyrosine Reactions. — The next experiment was an attempt to pro- duce premature pigmentation artificially in order to watch the reaction of the different regions. Studies on the chemistry of melanin formation by several authors have shown that two components are necessary for pigmentation: a chromogen and an oxydase. Oxydase (e.g. tyrosinase) is always present in the blood of Lepidoptera as may be easily shown, while the chromogen is supposed to enter the wing at a definite stage of development (the time of pigmentation). Oxydase and oxygen together produce a melanin pigment. If the blood of Lepidoptera contains tyro- sinase it should be possible to produce a dark pigment on the yet un- pigmented wing by providing the necessary chromogen, e.g. by soaking the unpigmented wing in a solution of tyrosine. ( 1 ) Papllio a/a.r. — An unpigmented pupal wing of Papilio at the time of the relief stage was put in a saturated solution of tyrosine. At the stage in question the pupal wing is always folded in a definite manner, which allows for the subsequent expansion of the wing after hatching. The wing is compressed in length and width to one-third of its actual size. This results in a wave-like appearance of the epithelium, with the scale- bearing epithelium as the crest and the epithelium without scales as the trough. (These folds do not interfere with the observation of the relief-stage on the pupal wing.) After immersion in the solution for four hours at 30°, the wing exhibited the following characteristics : it was completely stretched like the wing of the imago and showed the complete pattern of an adult pigmented wing; subsequent dark parts were dark, subsequent light parts were white (unpigmented) (Figs. 3 and 4). Mi- croscopical examination of wings which had been treated with tyrosine revealed that the pigment is actually deposited in the scales. It is, how- ever, not as black as the normally deposited pigment in the wing of the hatched butterfly. This experiment shows that a certain viscosity of the chitin is necessary for the deposition of a pigment in colloidal solu- tion. After a scale has reached a certain point of hardening it cannot deposit any more pigment. The tyrosine can react with the tyrosinase in the still soft scales of subsequent dark parts, but no reaction can take place in the hardened scales of subsequent light parts. The tyrosine ex- periment demonstrates clearly how the different speed of hardening in different regions leads to the pattern of pigmentation. Pupal wings of different age were treated accordingly to investigate the length of the stage during which a tyrosine reaction is possible. Very young and soft pupal wings, those on which the scales have just started to grow, show after treatment with a tyrosine solution a light DEVELOPMENT OF WING-PATTERN 235 gray color all over the wing with very faded contrasts of dark and light areas. The veins, however, and the marginal scales show a dark black- pigment (Fig. 5). A control wing, dried at the same stage, shows all scales collapsed. In a later stage a dried-out wing will still show all scales collapsed, but if it is treated with tyrosine, a pale pattern will become visible, and subsequent dark parts will show black reaction, while subsequent light parts will not show any reaction with tyrosine (Fig. 6). With the help of the tyrosine reaction we thus can trace back the dif- ferential velocity of chitinization to a very early point in the scale de- Frn. 4a-!>. Pupal wings of Papilio aja.v after treatment with tyrosine ca. 1 day before pigmentation begins to set in. a. fore wing and hind wing X l1/^, b, part of the fore wing X 21/£. FIG. 5. Pupal wing of Papilio tija.r after treatment with tyrosine ca. 2 days before pigmentation begins to set in. X 2%. FIG. 6. Pupal wing of Papilio aja.v after treatment with tyrosine ca. 3 days before pigmentation begins to set in. X 2l/2- velopment. Long before the relief stage we are able to produce the later pattern, which means that before wye can see the roughly morpho- logical difference of the chitinization in the relief stage, we are able to demonstrate it by chemical means. The chitin of subsequent white scales either hardens earlier and does not allow the tyrosine to penetrate, or at these moments the white scales do not contain sufficient blood which, as carrier of the oxydase, is necessary for the reaction with the tyrosine. The first explanation seems more plausible. All the wings which showed the tyrosine reaction still have to be examined microscopically more carefully to complete our knowledge of the details of this process. But 236 WERNER BRAUN the fact remains that from a very early stage, the different velocities of the different districts can be shown long before we can observe collapsed or erected scales after exposure of the wing to air. The tyrosine reac- tion can be performed until normal pigmentation sets in on the wing and always gives a clear positive of the pattern. Eighteen wings of Papilla were used for these observations and Papilla aja.v has proved to be an excellent object for these studies. (2) Ephcstia k. — The tyrosine reaction was tested on thirty flour- moth wings but did not show the bands distinctly, probably because the parts of the pattern of the Ephcstia wing are not clear and limited enough. It has been mentioned above that the pupal wings which are folded when removed from the pupae will always unfold after they have been placed in tyrosine. No other solution tested produced this phenomenon. Pupal wings from several species, among others pupal wings of Drosoph- ila, were put in a tyrosine solution, and they always unfolded after one or two hours. This fact might be of great help for other embryo- logical work in insects. Dissolution of the Scales. — The next test was based on the following consideration. The scales show a distinctly different chitinization at the relief stage. It is easy to remove scales from the collapsed or erect parts. If the scales from both parts could be dissolved, the harder scales should need a longer time for dissolving. Mature scales on the wing of a hatched butterfly should show the same degree of chitinization in all parts, and white and dark scales should need the same time for dissolution. Such an experiment could furnish another proof that the different chitinization during development is really an expression of the differential speed of some developmental processes. All our ex- periments have a greater meaning if it can be proved that all scales are chitinized equallv on the adult wing. The experiments proved that this idea was right. Papilla aja.v scales from subsequent light and dark parts were dis- solved in concentrated sulfuric acid. Subsequent dark scales start to dissolve as soon as touched by the acid. Subsequent white scales first show air bubbles when put into H.,SO4, and then start to dissolve slowly (ca. 12 minutes). If the same experiment is repeated with scales from a mature pigmented wing, the white scales will show the same reaction as they did during the relief stage. This means they were already mature at this age. Dark pigmented scales, however, now show a longer resistance to the acid and start to dissolve only after hours of treatment. The prolonged time of the dark scales as compared with that of the white scales is probably due to the pigment deposition in the dark scales. DEVELOPMENT OF WING-PATTERN 237 Scales from the same wing which were utilized for the observation of the relief-stage were used for this experiment. The same results were obtained in fifteen successive tests. The Mitosis-pattern in Cecropia and Papilio ajax In order to construct a general picture of the elements important for the pattern formation it was finally necessary to show that all the stages discussed (mitosis-pattern, different chitinization, relief-stage) are of general occurrence. Only the general presence of all stages, in par- ticular relief-stage and mitosis-period, allows us to correlate them. The mitosis-pattern of Ephestia has been clearly demonstrated and closely investigated previously (Braun, 1936). The same kind of mitosis- pattern has been found now in ten wings of young Cecropia pupae, and follows the same principles as in Ephestia. (The youngest Papilio pupae which we had in our laboratory were already too old to show the period of cell division.) These observations confirmed that the mitosis-pattern, as well as the relief stage, is a general stage in the development of the wings of the Lepidoptera. DISCUSSION The new evidence gained by these experiments confirms Gold- schmidt's views completely. The tyrosine reaction, the relief stage and its chitin reactions, and finally the test of dissolving scales, leave no doubt that we are dealing with different velocities of development for the different parts of the pattern. This difference turned out to be a difference in the hardening or maturing of chitin. Under these cir- cumstances it is only logical to interpret the mitosis-pattern in these terms too. We have seen that all the stages of development for dif- ferent parts of the pattern are generally present in such different species of Lepidoptera as Ephestia, Papilio ajax and the Cecropia moth. Ephestia-workers neglected Goldschmidt's views on the basis of not find- ing different districts of the pattern in different stages of development, but they only considered growth, which does not seem to be an exact indicator for the different velocities. The indicator for these velocities seems to be of a purely histological nature. We have been able to dem- onstrate such an indicator in the process of chitinization of the scales. In interpreting the mitosis-pattern in the same terms as the subsequent stages of development we are aided by the similarity of time factors in both stages. The mitosis-period and the relief stage proceed wave-like proximally to distally over the wing. In both periods we find the mar- WERNER BRAUN ginal scales advanced in development as compared with all the other scales. Based on former investigations and our new experiments we can describe the development of the pattern in Lepidoptera as follows : In the very young pupal wing, districts of different visible as well as physico-chemical structure are present, partly produced by the localiza- tion of tracheas, veins and different surface conditions, perhaps also by processes in which a Liesegang phenomenon is involved. At a certain time in development a determination stream (or more than one) spreads over the wing and is distributed pattern-like according to the chemico- physical conditions present in the substrate and dependent upon the points of origin and directions of the stream or streams. In the case of Ephestia it can be shown that the subsequent dark districts are deter- mined in the areas where the determination-stream stops. At any time before the occurrence of the determination-stream the subsequent pat- tern can be altered by extreme environmental conditions like heat treat- ment. Different parts of the pattern can be altered at definite limited periods, the " sensitive periods." Different phenotypes or genotypes are produced by the change of time-action of the determination-stream. The determination-stream determines districts with different speeds of differentiation. These different velocities of development are first visible in the mitosis-period which follows the period of the determination- stream. During this mitosis-period more cell-divisions are visible in later dark districts than in subsequent white districts, thus forming a mitosis-pattern, equal to the later pattern of pigmentation. The cells of the subsequent dark parts, which show a more intensive division process, have not yet reached the end of their multiplication period, in contrast to the subsequent white parts which enter their period of differentiation first. The mitosis-pattern spreads over the wing in the form of a mitosis-wave from the proximal base to the distal margin of the wing. Two such mitosis-waves can be observed. The first one is composed of cell-divisions of the hypodermis-cells and the first divisions of the scale-building cells. The second one is composed of the second cell- divisions of the scale-building cells. The different velocity of develop- ment for the different parts of the pattern can be observed again some time before pigmentation sets in. It can be made visible by the tyrosine reaction, the relief stage and its chitin reactions, and by the test of dis- solving scales. These tests show that the subsequent dark parts, which have already been retarded during the mitosis-period, develop slower up to the time of pigmentation. The velocity of hardening of the chitin is different for the different parts of the pattern, and this difference finally leads to the pattern of pigmentation. Since a pigment is de- DEVELOPMENT OF WING-PATTERN 239 posited in the chitin in colloidal solution, it can only be deposited as long as a scale has not reached a certain point of hardening. At a certain time of development pigment is present in the body and the subsequent dark parts, being still soft at this time, will deposit pigment ; the subse- quent white scales on account of their faster development, which has led to a harder chitinization, cannot deposit any pigment at this point. In this way the different velocities of differentiation of different areas of the wing throughout the pupal development starting from the time of the determination stream will lead to the pattern of pigmentation. SUMMARY 1. Goldschmidt's observations and ideas on the problem of different velocities of development for different parts of a pattern are described. The experiments of Kiihn and his school and their negative views in regard to the idea of different velocities of developmental processes are summarized. 2. New observations confirm the existence of the " relief-stage " in the pupal wing of Platysamia cecropia and Papilio ajax, as well as in Ephestia k., where it has not been previously observed. 3. It was observed that the relief-stage proceeds over the pupal wing from the proximal base to the distal margin in a wave-like fashion as is the case with the mitosis-pattern. The relief stage and the mitosis- pattern appear on the hind wing earlier than on the front wing. 4. No difference in growth of the scales was observed in sections through the wing during the relief-stage. 5. With the help of a chitin-reaction it could be shown that during the relief-stage the subsequent light parts of the wing were more chitin- ized than the subsequent dark parts. 6. A test with polarized light during the relief-stage gave no clear results. 7. With the help of artificial pigmentation (tyrosine-reaction) a com- plete pattern can be produced on the still unpigmented wing long before the relief-stage could be observed. 8. This tyrosine-reaction shows that a certain condition of the chitin is necessary to deposit pigment in the scales. Subsequent light parts of the wing are more chitinized at the time of pigmentation than subsequent dark parts and therefore cannot deposit any pigment. This experiment proves that the different velocities of development of the different parts of the pattern lead to the pattern of pigmentation. 9. By dissolving scales from subsequent dark and light districts in H.>SO4 during the relief-stage and dark and light scales on the mature 240 WERNER BRAUN wing, it could be actually shown that the scales are differently chitinized at the time of pigmentation and equally chitinized on the mature wing. 10. The similarity of processes during the mitosis-period and the relief stage and their general appearance in different species of Lepi- doptera allow us to correlate them. Both stages express the different velocities of developmental processes for different parts of the wing, which finally produce the pattern of pigmentation. LITERATURE CITED BRAUN, WERNER, 1936. tjber das Zellteilungsmuster im Pupenfliigelepithel der Mehlmotte Ephestia kuhniella Z. in seiner Beziehung zur Ausbildung des Zeichnungsmusters. Arch. f. Entw.-mcch., 135 : 494. FELDOTTO, W., 1933. Sensible Perioden des Fliigelmusters bei Ephestia kiihniella Zeller. Arch. f. Entw.-mech., 128 : 299. GOLDSCHMIDT, R., 1920. Untersuchungen zur Entwicklungsphysiologie des Fliigel- musters der Schmetterlinge. Arch. f. Entw.-mech., 47 : 1. GOLDSCHMIDT, R., 1923. Einige Materialien zur Theorie der abgestimmten Reak- tionsgeschwindigkeiten. Arch. mikr. Anat., 98 : 292. GOLDSCHMIDT, R., 1927. Physiologische Theorie der Vererbung. Berlin. GOLDSCHMIDT, R., 1938. Physiological Genetics. McGraw-Hill Book Co. KOHLER, W., 1932. Die Entwicklung der Fliigel bei der Mehlmotte Ephestia kuhniella Zeller, mit besonderer Beriicksichtigung des Zeichnungsmusters. Zeitschr. Morph. u. Okol. Tiere, 24 : 582. KOHLER, W., u. W. FELDOTTO, 1935. Experimented Untersuchungen iiber die Modifikabilitat der Flugelzeichnung, ihrer Systeme und Elemente in den sensiblen Perioden von Vanessa urticae L., nebst einigen Beobachtungen an Vanessa io L. Arch. Jul. Klans-Stiftg., 10: 313. KOHLER, W., u. W. FELDOTTO, 1937. Morphologische und experimentelle Unter- suchungen tiber Farbe, Form und Strucktur der Schuppen von Vanessa urticae und ihre gegenseitigen Beziehungen. Arch. f. Entiv.-mech., 136 : 313. KUHN, A., 1932. Zur Genetik und Entwicklungsphysiologie des Zeichnungsmusters der Schmetterlinge. Nach. Ges. Wiss. Gottingen, Math.-physik. KL, p. 312. KUHN, A., u. M. VON ENGELHARDT, 1933. liber die Determination des Symmetrie- systems auf dem Vorderfliigel von Ephestia kuhniella Z. Arch. f. Entw.- mech., 130: 660. KUHN, A., u. K. HENKE. Genetische und entwicklungsphysiologische Unter- suchungen an der Mehlmotte Ephestia kuhniella Zeller I— VII. Abh. Ges. Wiss. Gottingen, Math.-physik. KL, 15 : H. 1 (1929); VIII-XII, 15: H. 2 (1932) ; XIII u. XIV, 15: H. 3 (1936). SCHULZE, P., 1922. tiber Beziehungen zwischen pflanzlichen und tierischen Skelett- substanzen und iiber Chitinreaktionen. BioL Zcntralbl., 42 : 388. STOSSBERG, MARGARETE, 1937. Uber die Entwicklung der Schmetterlingsschuppen. BioL Zentralbl, 57: 393. SUFFERT, FRITZ, 1937. Die Geschichte der Bildungszellen im Puppenfliigelepithel bei einem Tagschmetterling. BioL Zentralbl., 57 : 615. THE RELATIONSHIP BETWEEN THE PITUITARY GLAND AND THE GONADS IN FUNDULUS SAMUEL A. MATTHEWS (From the Thompson Biological Laboratory, Williams College, and the Marine Biological Laboratory, Woods Hole) INTRODUCTION From the extensive work on mammals, birds, reptiles and amphibians it is clear that in these forms the anterior lobe of the pituitary gland is an essential factor in the control of the sex cycle. In teleost fishes, how- ever, there is little evidence concerning the role of the pituitary body. Houssay (1930, 1931) found that the injection of saline suspensions of the hypophyses from large fish (Micropogon opercularis} into small ones (Cnesterodon) was followed by the expulsion of eggs in from one to three days, although saline suspensions of muscle or saline solution alone produced no effect in control animals. This expulsion of eggs, however, preceded normal spawning by only 15 days. Cardoso (1934) injected saline suspensions of the pituitary bodies of Pimelodus clarias into other individuals of the same species. In both sexes he noted an increase in weight of the gonads over those of control animals, an effect much greater in the female. Pereira and Cardoso (1934), working with Prochilodiis, injected saline suspensions of pituitary glands of the adult into females. In all cases ovulation occurred, varying from 24 to 96 hours after injection, although the injection of suspensions of nervous tissue into control animals did not result in ovulation. More recently von Ihering (1935) repeated these experiments, injecting suspensions of the pituitary glands of Hoplius malabaricus into two species of As- tyanax. Mating activity was stimulated with an increase in size of both ovary and eggs over the untreated controls. When injections were made into Prochilodus of both sexes eggs and spermatozoa were obtained. Pituitary-like substances such as prolan and extracts of the pituitary gland have likewise been used in fishes with some success. Calvet (1932) placed young Petromyzon in aquaria containing pregnancy urine and found that the ovary increased markedly in size over that of the controls. Damas (1933) injected 30 lampreys with pregnancy urine and obtained expulsion of eggs in all cases. He also employed extracts of the pars anterior of the pituitary gland with similar results. Boucher, 241 242 SAMUEL A. MATTHEWS Boucher and Fontaine (1934) injected the urine of pregnancy into eels and reported enlargement of the gonads in both sexes and evidence of the maturation of spermatozoa in the testis. On the other hand, wholly negative results were obtained by Koch and Scheuring (1936) after in- jecting Phoxinus with both prolan and an extract of the anterior lobe of the pituitary gland. EFFECT OF INJECTION OF PITUITARY EXTRACTS In an effort to determine whether or not the pituitary gland of Fun- dulus heteroclitus is concerned in the control of the sex cycle two meth- ods were employed, the injection of pituitary glands or extracts of them, and extirpation of the pituitary body. The results of the injection method have been disappointing. In male individuals no effect which can be ascribed to the material injected has been observed. The follow- ing is a brief summary of the results of several experiments. Follow- ing the successful use by Burns and Buyse (1933) of extracts of mam- malian pituitary glands in stimulating precocious sexual development in Amblystoma, injections of a crude, alkaline extract of the whole pituitary gland of sheep were made into the body cavity of Fundulus.1 In the first series records were obtained from 28 males so injected, with a num- ber of uninjected controls. While spermatozoa were obtained in a number of cases they were also obtained in control animals, developing in both groups after the individuals had been maintained in the labora- tory for some time. This appearance of spermatozoa was possibly a result of the change from the lower temperature of the Delaware river in which they were collected to the higher temperature of the laboratory. At any rate, that the appearance of spermatozoa in these males was not a result of the injections was shown by further experiments. Twenty- four males were divided, 12 placed in an aquarium in a constant tem- perature room at 5° C. and the other 12 kept in an aquarium in which the temperature of the water varied from 18° to 22° C. with an average temperature during the experiment of 20°. Since the constant tem- perature room was illuminated by daylight the light conditions in the two rooms were fairly comparable. Half the animals in each room were injected daily with 0.1 cc. of sheep pituitary extract. Records were obtained on 18 of these animals. In the warm room the testes were activated and spermatozoa appeared in both uninjected and injected ani- mals in all cases after 7 days. In the cold room no spermatozoa ap- peared in either injected or in control animals. Microscopic examina- tion of the testes showed that those of the injected animals were com- parable to those of the uninjected controls, those of both groups in the 1 1 wish to express my thanks to Dr. Oliver Kamm of the Research Labora- tories of Parke, Davis and Company for this material. PITUITARY GLAND AND GONAD IN FUNDULUS 243 cold room showing less activity than was the case in the testes of animals from the warm room. In females a more varied injection program was employed in an attempt to produce ripe eggs. The following materials were injected : whole pituitary extract of sheep, Antuitrin S (Parke-Davis), extracts of fish pituitary glands (Fundulus and Mustelus), and freshly ground pituitary glands of Fundulus. Records were obtained from 35 injected females. Only 4 of these, animals injected in March and early April, delivered ripe eggs. The earliest case of this occurred on March 16 after the animal had been in the laboratory for 1 1 days and had received 3 injections of 0.15 cc. each of sheep pituitary extract. The other 3 cases delivered eggs late in April (April 20-28) about one month before eggs were obtained from normal animals (May 27). While these 4 cases suggest the possibility that injection of pituitary extracts into Fundulus in early spring after the ovocytes have grown to a certain size may cause them to mature more rapidly than they normally do, certainly the number of cases is too small to be of much significance. In the 31 other cases no ripe eggs were obtained and microscopic examination of the ovaries showed no significant differences between them and those of control animals. EFFECT OF HYPOPHYSECTOMY The second method employed to study the influence of the pituitary gland on the gonad cycle in Fundulus was the removal of the gland. The operation was performed essentially in the manner previously de- scribed (Matthews, 1933) but with several modifications. A single, short mid-ventral incision was made in the branchiostegal membrane, reaching anteriorly as far as the base of the tongue. Through this in- cision the tongue was dissected free from the tissues of the floor of the mouth. The tip of the tongue was then pulled backwards and held at right angles to the ventral surface of the head with the left hand, its wedge shape serving to hold the wound open. After cutting through the mucous membrane in the roof of the mouth the parasphenoid bone was cut through on three sides, bent to one side of the median line to expose the hypophysis, and then replaced after the pituitary body had been removed. If the original incision in the branchiostegal membrane was of the proper length the tongue could be tucked through the incision and the free tip would hold it in place, thus closing the wound in the branchiostegal membrane without sutures. Healing in these cases was much more satisfactory than in the earlier operations. This operation was carried out at two periods of the year. One group of animals was operated on in March and April, the other in October, November and December. The mortality, as usual, was high. 244 SAMUEL A. MATTHEWS Of 176 operated animals weight records and fixed gonads were ob- tained from only 73, — 33 males and 40 females. When the gonads from these animals were examined differences were noted between them and TABLE I Percentage of total body weight formed by the testis in hypophysectomized Fundulus as compared with normal and operated controls I Normal and Operated Controls Hypophysectomized Animals Date Killed Days Post- Operative GW/BW (per cent) Date Killed Days Post- Operative GW/BW (per cent) NOR Oct. 10. 0.22 Oct. 19 10 0.21 Oct. 10. 0.31 Oct. 19 13 0.19 Oct. 10. 0.27 Oct. 26 13 0.22 Oct. 24 18 0.26 Oct. 26 20 0.19 Oct. 28 21 0.25 HYCNov. 17. 39 0.76 Nov. 7 32 0.21 Nov. 27 48 0.37 HYCDec. 3. 53 0.71 Dec. 3 51 0.33 HYCDec. 3. 56 0.49 Dec. 17 66 0.27 Dec. 24 73 0.43 HYCMar. 12. 12 0.46 Mar. 7 5 0.46 Apr. 20 5 1.05 Mar. 11 9 0.77 HYCMay 5. 206 1.20 May 4 10 1.48 HYCMay 5. 206 2.01 May 6 12 0.81 May 6 12 1.62 May 6 13 0.43 May 8 14 0.24 May 10 16 0.38 May 10 16 0.54 May 12 18 0.29 May 13 19 0.44 May 7 22 0.67 May 14 29 0.41 May 16 31 0.21 HYC June 11. 48 1.26 June 11 48 0.30 NOR June 11. 1.96 June 11 49 0.28 June 11 55 0.17 June 11 55 0.08 June 11 57 0.28 June 11 57 0.17 May 5 206 0.15 those of controls, particularly in the males. These changes were ap- parent in the weight of the testis relative to the total body weight, and in both gross and microscopic appearance. Table I summarizes the way in which the testis weight changes after PITUITARY GLAND AND GONAD IN FUNDULUS 245 hypophysectomy. As has been shown (Matthews, 1938) during the normal seasonal cycle the testis forms the smallest percentage of the body weight from September to December and the greatest in May and June. When the pituitary gland is removed in October there is no sig- nificant difference in the weight of the testes of experimental animals as compared with those of controls even after 73 days. In the one animal which survived until the next breeding season, however (206 days), the difference in the proportion of the body weight formed by the testis as compared with that of control animals was striking. Obvi- ously the testis in this animal had not enlarged with the onset of the breeding season as did that of the control. In the series from which the pituitary gland was removed in March and April and the animals killed during the next breeding season in May when the testis normally reaches its greatest size, the same difference between the weight of the testes of control and of hypophysectomized animals was noted, though in these cases they had been hypophysectomized only 48 to 57 days pre- viously. How soon after operation this difference in weight is notice- able cannot be accurately determined from this series due to an insuf- ficient number of animals killed early in the experiment. It is clear, however, that there is no difference up to about 10-12 days after re- moving the hypophysis. After 13 or 14 days, however, the proportion of the total body weight formed by the testis is consistently and notably less than in control animals. In gross appearance the differences between the testes of hypophy- sectomized and control animals are clear from Figs. 1 and 2. In place of the large whitish testis of the control animal that of the hypophy- sectomized individual is smaller and has the gray, translucent appear- ance of the testes of sexually inactive animals killed in late fall. The microscopic appearance of the testes of these animals also shows differences between control and hypophysectomized individuals. Al- though the testes of animals operated on in October show no significant weight differences as compared with those of controls, the microscopic structure of these testes is different. The testes of control animals killed in October show a large number of primary spermatogonia, a number of secondary cells, including both secondary spermatogonia and primary and secondary spermatocytes which are multiplying to form cortical cysts, as well as a number of spermatids and a few spermatozoa (Fig. 3). In two animals killed during mid-October from which the pituitary glands had been removed 10 and 13 days previously the pri- mary spermatogonia are as numerous as in the controls and are dividing in both cases as indicated by the presence of mitotic figures. The strik- ing differences between these two testes and those of control animals 246 SAMUEL A. MATTHEWS killed at the same time lie in the reduced number of cells of the second type and in the small number of mitotic figures in these as compared with those in the control testes (Fig. 4). In fact, the cysts, which are prominent in the control testes, are small and poorly defined in those from the hypophysectomized individuals. Moreover, spermatids and spermatozoa are extremely scanty. The testes of these two hypophy- sectomized animals present the picture of multiplication of spermato- gonia without subsequent maturation of sperm cells. If this lack of maturation of spermatozoa is due to the lack of the pituitary gland, then it might be expected that such differences would be more marked in those hypophysectomized animals allowed to live on into the normal breeding season when maturation of sperm cells is most rapid. That such is the case is indicated in a number of such cases. Obviously the removal of the pituitary gland must precede the killing of the animal by a sufficient number of days for the effect on the testis to be noted. Thus in none of the animals hypophysectomized in March and April and killed from five to twenty-two days later are any differences observable in the testes of the operated and control animals. In fact, one of these animals killed on April 20, five days after removal of the pituitary gland, showed a definitely active testis (Fig. 5). The differences be- tween this testis and that of one of the same series of operated animals killed on May 16 are well marked (Fig. 6). In the latter case the only cells present are primary spermatogonia and a few spermatids and spermatozoa. The differences between the two are such as might be explained by assuming that on April 15, when the pituitary glands were removed, rapid multiplication and maturation of the spermatogenic cells were well under way in preparation for the next breeding season. In the animal killed five days after removal of the hypophysis the lack of the gland had not as yet produced any effect on the testis. Twenty-six EXPLANATION OF PLATE I FIG. 1. Testis of control animal which had its pituitary gland exposed but not removed on October 10 and was killed May 5. X 1.4. FIG. 2. Testis of animal hypophysectomized on April 15 and killed June 11. X 1.4. FIG. 3. Cross-section of testis of normal animal killed October 10. X 194. C, cortical cyst containing spermatocytes ; PS, primary spermatogonia ; S, sper- matids. FIG. 4. Cross-section of testis of animal hypophysectomized October 9 and killed October 19. X 194. PS, primary spermatogonia ; M, medulla of testis. FIG. 5. Cross-section of testis of animal hypophysectomized April 15 and killed April 20. X 80. C, cortical cysts containing cells which are dividing rapidly; SZ, spermatozoa. FIG. 6. Cross-section of testis of animal hypophysectomized April 15 and killed May 16. X 80. PS, primary spermatogonia; SZ, spermatozoa. PITUITARY GLAND AND GONAD IN FUNDULUS 247 PS~---. . c~M-;^-m J tJrillP . • . .• . ^ * .*.«,T-**5'7 sz -*£•• ^•^- *?.*' J^J'^f.'t *$&*< Mmi^s^ff •.-• -*5*.* - - T V-^M vS'jJEii*1 * -51^ *£.*£& '$$ • '.'--^^RsiSffi r^i .^7-^rv^5»» •S*:^/.---L-^^ •^•-.'•>*^% PS sz PLATE I 248 SAMUEL A. MATTHEWS days later, however, only a few spermaticls and spermatozoa remain from the mass that had already been produced. Evidently after the lack of the pituitary gland has had time to manifest itself, further maturation of spermatogenic cells is either very much retarded or per- haps altogether lacking. The most striking examples of this change in the testis following hypophysectomy were found in those animals which were allowed to run well over thirty days after removal of the pituitary gland and then killed close to or during the breeding season. Seven of these cases were observed with 5 control animals killed at the same time. These animals were killed from 48 to 206 days after removing the hypophysis. PS M -- 8 PLATE IT Fi<;. 7. Cross-section of testis of control animal which had its pituitary gland exposed but not removed on April 24 and was killed June 11. X 80. C, cortical cysts with cells which are dividing rapidly ; ST., spermatozoa. Fir.. 8. Cross-section of testis of animal hypophysectomized April 24 and killed June 11. X 80. I'S. primary spermatogonia ; M, medulla of testis. With the exception of minor individual variations, differences in the microscopic appearance of the testes of these animals as compared with those of controls are much the same as those already described. The testis of all the hypophysectomized animals show only primary sperma- togonia, a few scattered secondary spermatogonia and either very few spermatids and spermatozoa or none at all (Figs. 7 and 8). It might be noted, however, that in all these animals the primary spermatogonia are dividing, even in the one case from which the pituitary gland had been removed 206 days previously. The data for hypophysectomized females are at present very scanty. As in the male, there is little change in weight of the female gonad in PITUITARY GLAND AND GONAD IN FUNDULUS 249 those cases killed in the fall months and differences in the microscopic structure of the ovary of hypophysectomized as compared with control animals are hard to detect. Although more females were operated on than males the mortality of these animals was particularly high as the breeding season approached. As a result records on hypophysectomized females killed during May and June have been obtained on only a few animals. The oldest of these in post-operative age was killed only 55 days after removal of the hypophysis and records were obtained on only two others which were killed more than 30 days following hypophy- sectomy. In all three cases, however, the ovary wras lighter relative to the total body weight than in control animals, several of which ran 206 days after exposing but not removing the pituitary gland. An ex- ample is provided in case HY 4/17-6/11 in which the ovary formed 1.52 per cent of the body weight 55 days after removing the hypophysis as compared with HYC 10/10-5/5 (206 days post-operative, 4 indi- viduals) in which the average weight of the ovary constituted 2.87 per cent of the body weight. It is obvious, however, that more cases from which the hypophysis has been removed for more than 30 days prior to the normal breeding season must be obtained before the effect of the loss of the pituitary gland on the ovary may be accurately determined. SUMMARY Injections of pituitary extracts into male Fundulus have been prac- tically without stimulating effect on the testis. In females records were obtained from thirty-five cases of which only four delivered eggs earlier than did control animals, one case delivering ripe eggs six weeks, the other three cases about four weeks before the normal breeding season. After removing the pituitary gland the gonads undergo regressive changes as compared with the controls, changes which are particularly striking in the testis. When the pituitary gland was removed in the fall and the animals were allowed to run until the next breeding season in May and June the testes failed to enlarge as they normally do. Mi- croscopically such testes showed numerous primary spermatogonia but very few spermatocytes and practically no spermatozoa. When the pituitary gland was removed in March and April, after the testis had already enlarged somewhat, fourteen to twenty-one days after operation the testes had become notably smaller than in control animals, and showed only primary spermatogonia with a few spermatids and spermatozoa. These results indicate that the pituitary gland exerts a controlling influence on the seasonal cycle which the testis of this teleost fish exhibits. This influence is apparently of greater importance in maturation than in proliferation of the germ cells. 250 SAMUEL A. MATTHEWS LITERATURE CITED BOUCHER, S., M. BOUCHER, AND M. FONTAINE, 1934. Sur la maturation provoquee des organes genitaux de 1'anguille. Compt. Rend. Soc. BioL, 116: 1284. BURNS, R. K., AND A. BUYSE, 1933. The induction of precocious maturity in the reproductive tract of recently metamorphosed female salamanders, by an extract of the mammalian hypophysis. Anat. Rec., 58 : 37. CALVET, J., 1932. Action du lobe anterieur d'hypophyse chez divers Vertebres (Lamproies, Oiseaux). Compt. Rend. Soc. BioL, 109: 595. CARDOSO, D. M., 1934o. Relations entre 1'hypophyse et les organes sexuels chez les poissons. Compt. Rend. Soc. BioL, 115: 1347. CARDOSO, D., 19346. Relagao genito-hipofisaria e reproducao nos peixes. Archiv. Instit. Bio I., 5: 133. DAMAS, H., 1933. Note sur 1'apparition naturelle et provoquee des caracteres sexuels chez la lamproie. Bull. Soc. Roy. des Sci. de Liege. No. 4, p. 94. HOUSSAY, B. A., 1930. Accion sexual de la hipofisis en los peces y reptiles. Revista Soc. Argentina BioL, 6 : 686. HOUSSAY, B. A., 1931. Action sexuelle de 1'hypophyse sur les poissons et les reptiles. Compt. Rend. Soc. BioL, 106: 377. VON IHERING, R., 1935. Die Wirkung von Hypophyseninjektion auf den Laichakt von Fischen. Zool. Anseig., Ill: 273. KOCH, W., AND L. SCHEURING, 1936. Die Wirkung von Hypophysenvorderlappen- hormon auf den Laichakt von Fischen. Zool. Anzeig., 116: 62. MATTHEWS, S. A., 1933. Color changes in Fundulus after hypophysectomy. BioL Bull., 64: 315. MATTHEWS, S. A., 1938. The seasonal cycle in the gonads of Fundulus. BioL Bull., 75 : 66. PEREIRA, J., JR., AND D. M. CARDOSO, 1934. Hypophyse et ovulation chez les poissons. Compt. Rend. Soc. BioL, 116: 1133. STUDIES ON VIRUS DISEASES OF FISH II. LYMPHOCYSTIS DISEASE OF FUNDULUS HETEROCLITUS l RICHARD WEISSENBERG -' (From the Effiin/ham B. Morris Biological Farm of the W is tar Institute, Bristol, Pa.) At the meeting of the American Society of Zoologists at Indi- anapolis, December 1937, the author presented a paper and a demon- stration concerning " intracellular parasitism in fish producing a gigantic growth of the infected cells." 3 This paper also dealt with parasitic protozoa as well as with the lymphocystis virus disease and the peculiar structure of the lymphocystis cells. With reference to this latter, Dr. Roland Walker, Biological Laboratories, Rensselaer Polytechnic In- stitute, Troy, New York, sent to the author sections through a tumor of the common salt water killifish Fitndulits hctcroclitns, which he had preserved in August, 1937, at the Marine Biological Laboratory, Woods Hole, supposing that this tumor represented something similar to the demonstrated hypertrophied fish cells. Studying these preparations I could easily verify that the tumor in- deed represents a lymphocystis tumor, never described hitherto on Fnndnlns heiroditus. I am greatly indebted to Dr. Walker, who de- cided to leave the description of this interesting observation entirely to me. The significance of this discovery is not represented so much by the finding of a new host of lymphocystis disease, but by the fact that this fish is known to be easily kept under laboratory conditions. Therefore, it can be hoped that lymphocystis disease could be culti- vated in the laboratory for further experimental studies if sufficient at- tention were paid to a new appearance of this virus disease on Fnnditlus heteroclitus. The writer estimates from the series of sections that the tumor probably had a diameter of 2 to 3 mm. The localization of the tumor was in the membrane of the tail fin. The preserved fish was the only specimen among 200 Fiindnlus hctcroclitns on which Dr. Walker ob- 1 Studies on virus diseases of fish. I. Lymphocystis disease of the Orange filefish (Alcutcra schoepfii) appeared in the Amcr. Jour. Hygiene, Vol. 28, No. 3, 455-462, November, 1938. - Member of The Wistar Institute of Anatomy and Biology, Philadelphia, Pa. 3 Abstract in Anat. Rec., 70: No. 1, Suppl. No. 1, 68. 251 252 RICHARD WEISSENBERG served such a tumor. The tumor, preserved with Benin's fluid and stained with Harris' hematoxylin, consists of large cells of diameters of 135 to 225 microns, located in the connective tissue of the fin membrane. A POL-V B ..... FIG. 1. A. Two lymphocystis cells of Fundnlns hctcroclitus. X 210. B. Upper part of the upper cell with higher magnification. X 600. n, nucleus ; e, nucleolus ; i, inclusion bodies ; m, membrane ; j, strongly stained granules in the cortex of the lymphocystis cells ; c, small connective tissue cells ; [>, pigment cell ; i), vessels. These large cells (Fig. I A) contain in their cytoplasm the peculiar reticular inclusion bodies which represent the most conspicuous struc- tural element of lymphocystis cells. The inclusions (i) appear in the sections as separate coils, but are probably part of a continuous net- work, similar to those found in the lymphocystis cells of the perches LYMPHOCYSTIS DISEASE IN FUNDULUS 253 Stizostcdion). Figure IB shows with higher magnification their basophilic framework representing a reticular or alveolar struc- ture, strongly stained with hematoxylin. The meshes or alveoles are filled with a paler tingible substance. These inclusion bodies are so characteristic that the diagnosis of lymphocystis cells is assured without further ado although the second typical structure of the lymphocystis cells, the cell-membrane, shows a remarkably poor development. The cell-membrane appears in Fig. I A It-S FIG. 2. Lymphocystis cell of Fitndnlits hctcroclitns, the cytoplasm of which has withdrawn by shrinking from the membrane (;»). n, nucleus; c, nucleolus ; i, inclusion bodies ; s, strongly stained granules in the cortex of the lymphocystis cell ; c, small connective tissue cells. X 600. only as a contour (in) and also with higher magnification (Fig. IB), the configuration of the membrane is not very clear. But cells, in which the cytoplasm has withdrawn from the membrane by shrinking (Fig. 2), show distinctly that the membrane is represented by a cuticula of double contour, evidently homologous to the thicker membranes of the lymphocystis cells of other fish. The nuclei contain in their peripheral zone a fairly large amount of chromatic substance as bars and fields of granules. One or two nucleoli 254 RICHARD WEISSENBERG ( c) are to be seen which, as it seems, can give origin to buds (Fig. I A ). Just beneath the membrane of the lymphocystis cells a layer of irregular granules (s) , strongly stained with hematoxylin, could be observed, which, as a rule (Fig. I A), have only developed on those planes of the cells which face the surfaces of the fin membrane. Something similar was seen by the author hitherto only on some degenerating lympho- cystis cells of Plcuroncctcs flcsns, but there in the whole peripheral zone of the cells. Because it would be very important for the further research on lymphocystis virus to obtain a stock of lymphocystis-affected fish which might be easily kept in aquaria, the writer attempted to pursue in 1938 the trail found by Dr. Walker. Unfortunately, it seems that the oc- currence of lymphocystis disease on Fumiulits hetcroclitus is very rare. Five hundred and fifty specimens of this species, preserved at the Marine Biological Laboratory, Woods Hole, in August 1938, were thoroughly examined without discovering any other case of lymphocystis disease. Further, there was no success in attempting to transmit to Fnndiilns hetcroclitus the lymphocystis disease of the wall-eyed pike perch, Stizo- stcdio)i z'itreitin, fresh material of which could be procured from the Great Lakes. Corresponding experiments to transmit the Stizostedion- disease to the freshwater killifish, Fiindulns diaphanus, gave the same negative result. The method applied in these experiments consisted in dispersing an emulsion of Stisostedion-lymphocystis tumors into the aquarium and in feeding small pieces of the tumors. The Fundiilns- fish of both species were very eager to swallow the Stizostedion-\ym- phocystis cells, but none of 16 F. heteroclitus and of 12 F. diaphanus became infected. Concerning this result, the following facts may be taken into ac- count : ( 1 ) The transmission of lymphocystis disease under laboratory conditions was hitherto only successful by keeping diseased and healthy fish together or by the method applied in the Stizostedion-Fundulus- experiments which attempted to imitate the natural conditions of in- fection. (2) This method gave, as a rule, positive results in high percentage in all experiments in which the applied tumor material originated in the same kind of fish to which the treated fish belonged [See experiments of Weissenberg in Europe on Accrina ccnnia (1914)4 and Plcuroncctcs flcsus (1921)5 and in this country on Stizostedicm 4 Weissenberg, R., 1914. Uber infectiose Zellhypertrophie bei Fischen (Lvm- phocystiserkrankung ) . Sitzumjsbcr. d. K;il. f'rcnsx. Akad. d. IViss., Pliysik. •mathcm. KL, 30: 792-804. 5 Weissenberg, R., 1921. Lymphocystiskrankheit dcr Fische. Handbnch der Pathogcncn Protozoen hcraitsgcyebcn ron T. Prowazck-Nollcr, 3: 1344-1380. LYMPHOCYSTIS DISEASE IN FUNDULUS 255 vitreum (still unpublished)]. (3) The lymphocystis disease of Acerina cernua was discovered by Weissenberg on a Baltic Sea tribe of this species. A freshwater tribe of the same species showed in an infection experiment a lower susceptibility than the original tribe (Weissenberg, 1914, I.e., p. 802). (4) Experiments to infect fishes of other species, genera or families by keeping them together with lymphocystis-affected fish has hitherto always given negative results. But these observations (Joseph, 1918; 6 Weissenberg, 1921, l.c, p. 1348) referred to fish on which lymphocystis disease is not observed in nature. The negative result in the aforesaid experiments to transmit lym- phocystis disease from Stizostedion vitreum to Fundulus heteroclitus and diaphanus, in which experiments a certain susceptibility for the disease could be supposed at least for F. heteroclitus on account of the described case of Woods Hole, is in accordance with the interpretation that different kinds of lymphocystis viruses are to be distinguished which are adapted to different fishes. It would be very desirable to pay attention to further cases of lym- phocystis disease of Fundidus heteroclitus and to keep affected speci- mens together with healthy fish of this species for cultivating a stock of lymphocystis-affected fish, which could easily be kept under labora- tory conditions and therefore would be well adapted for the further study of this interesting virus disease in this country. SUMMARY The first and until now solitary case of lymphocystis disease in Fundidus heteroclitus is described. Fundidus heteroclitus and F. di- aphanus proved not susceptible to the virus of the lymphocystis disease of Stisostedion vitreum in infection experiments. 6 Joseph, H., 1918. Untersuchungen iiber Lymphocystis Woodc. Arch. f. Protist., Bd. 38. DIURNAL VERTICAL MIGRATIONS OF DEEP-WATER PLANKTON TALBOT H. WATERMAN, RUDOLF F. NUNNEMACHER, FENNER A. CHACE, JR., AND GEORGE L. CLARKE (From the Woods Hole Oceanographic Institution1 and the Biological Labora- tories, Harvard University) INTRODUCTION Diurnal vertical migrations have long been known to play an im- portant part in the lives of pelagic organisms. The numerous previous studies on these migrations in oceanic animals have been limited almost entirely to plankton living relatively near the surface. It was the purpose of the present investigation to continue and extend the obser- vations made by Welsh, Chace, and Nunnemacher (1937) on the di- urnal vertical migrations of deep-water animals. More specifically, our objectives were (1) to ascertain the vertical extent of these migrations for the bathyplankton,2 (2) to determine the greatest depth in the sea at which these migrations still occur, and (3) to discover the relations of these phenomena to the changes in submarine illumination with time and depth. HISTORICAL CONSIDERATIONS The earliest knowledge of diurnal vertical migrations of marine organisms arose from the practical observations of fishermen. Thus man has known from time immemorial that under certain conditions if he wishes to obtain a good catch of herring, his nets should be set deep in the water during the daytime and shallower at night. Such obser- vations were abundantly confirmed and extended for shallow-water animals by the work of a great many marine biologists (an extensive literature is cited by Rose, 1925; Russell, 1927; and Clarke, 1933&). As far as the plankton of deeper water was concerned, a small amount of information was also already at hand by quite an early date. For example Rang (1828) observed that the pteropod, Hyalocylis striata, among others, could be taken at the surface of the sea only after sunset and before sunrise and sank into deeper water during the day. Recently Stubbings (1938) has found that this species descends during 1 Contribution No. 205, Woods Hole Oceanographic Institution. 2 See Ekman ('35, pp. 451-77) for a comprehensive general discussion of the bathypelagic fauna. 256 DIURNAL BATHYPLANKTON MIGRATION 257 the day to 200-250 m. so that for at least part of the day it is a bathy- pelagic organism, i.e., one which lives in the intermediate depths rather than near the surface or the bottom of the sea. But the significance of this and other similar early observations was not appreciated until long afterwards since the existence of a bathyplankton had not been realized until the latter part of the nineteenth century when the data of the various great oceanographic expeditions were analyzed. Murray (1885, p. 218) concluded from his experiences on the " Challenger " that, " the great majority of plankton organisms live at various depths down to and even deeper than 100 fathoms during the day . . . and only come to the surface at night. . . ." Other investi- gators came to similar general conclusions (e.g., Chun, 1890, pp. 82-4; Brauer, 1906, p. 337; 1908, p. 232). Brauer was among the first to point out the possible relation of these migrations to such interesting biological problems as bioluminescence, type and size of eye, and pig- mentation in the bathyplankton.3 However, it remained for Fowler (1905, 1909) first to demonstrate by means of numerical data for schizopod and ostracod crustaceans the diurnal vertical migrations of moderately deep-living species. Soon afterwards Murray and Hjort (1912) showed clearly that such migrations were undertaken by or- ganisms which were definitely " deep-sea " animals during the daytime ; these investigators also emphasized the correlation of these phenomena with the general biology of the sea, sustaining and extending some of Brauer's suggestions by means of their more complete data. The further inquiry into the exact and quantitative aspects of the subject presented considerable technical difficulties since for this purpose it is necessary to know accurately where the plankton has been caught, which is impossible when working with open nets. Hence the precise determination of the vertical extent of diurnal migrations and of the various depths at which they occur had to await the development of efficient deep-sea equipment for sampling populations at known depths without contamination from other strata. Since the invention of the Sigsbee gravitating deep-sea trap (1880) and Palumbo's closing net (Chierchia, 1885, p. 81), many oceanographers have devised such apparatus ; yet it was only very recently that reliable equipment has been available for these purposes. METHODS The observations were carried out in July, 1937, at " Atlantis " Station 2894 (39° 06' N., 70° 16' W.) which is in continental slope 3 For recent investigations of these problems see Welsh and Chace, 1937, 1938. 258 WATERMAN, NUNNEMACHER, CHACE AND CLARKE water about 300 miles east of Cape May, N. J. This position was chosen for several reasons: (1) The slope water, except for the super- ficial layers, is relatively free from currents (see Iselin, 1936, p. 11). (2) The station was far enough from both the Gulf Stream and coastal waters to be free of their influence. (3) The water was sufficiently deep to prevent benthonic organisms from confusing the results. (4) Data from the continental side of the Gulf Stream might offer some interesting points for comparison with results previously obtained (Welsh, Chace, and Nunnemacher, 1937) in the far more transparent waters of the Sargasso Sea. In order to keep the ship in about the same geographic position the hauls were made in various directions while the work was being done. The maximum deviation from the original station was about 14 miles (see Table I). TABLE I Position of the ship at various times while the work was being carried out. first position in the table is that designated as Station 2894. The Date Time Latitude Longitude July 18th 10 P.M. 39°06' N. 70°16' W. 19th. 5 A.M. 38°57' 70°30' 19th . . . Noon 39°01' 70°30' 20th .... Noon 39°06' 70°33' 21st Noon 39°00' 70°30' 22nd Noon 38°59' 70°30' Horizontally hauled closing nets were used, modified from those em- ployed by Leavitt (1935, 1938) as described by Welsh, Chace, and Nun- nemacher (1937). These were 2 m. in diameter at the mouth and 9 m. long ; the sides were of stramin, 6 threads to the centimeter, with a silk inset, about 20 threads to the centimeter (54 to the inch), in the tail. While the nets were open, the speed of the ship relative to the surface water was kept close to 2 knots. However, the actual speed of each net in relation to the water layer through which it was moving was not known. Since the water at various depths was probably moving at different velocities and possibly even in different directions, the results in such a situation might be very misleading as to the actual population distributions. But in the present case the fact that the hauls were made in various directions precludes the origin of systematic errors from this source. Furthermore, the evidence from the catches for a regular diurnal migration in the various organisms indicates that the relative DIURNAL BATHYPLANKTON MIGRATION 259 movements of the water were not of sufficient magnitude to confuse the results. The depth at which a net was fishing was estimated by means of the trigonometric relation between wire angle and the amount of cable out. At 1,000 m. this estimate was probably accurate within about 10 per cent, being more accurate in shallower and less so in deeper tows. Thirty-nine catches in a series of 16 hauls were made with the nets fishing for 2 hours at various depths and times throughout more than 2 diurnal cycles. Each entire haul took 4 hours so that 6 were made in a 24-hour period. Tows were made at 100, 200, 400, 600, 800, 1,000 and 1,200 meters, calculated depth, and since the bottom was at 2,860 m., the population thus sampled was purely pelagic. Although the closing nets worked very well in general, 4 of the 39 hauls failed; these are marked in Fig. 1 by asterisks. Figure 1 represents graphically the times and depths of all the clos- ing net hauls made at this station and shows that various depths were sampled on succeeding days. During the morning of the third day on station a series of tows at 100 and 1,200 m. was begun, but the sea became too rough to continue towing (force 5 on the Beaufort scale) so only one haul was made before work was stopped. In an attempt to fill in the gaps in our series we made tows at appropriate depths at noon and midnight on July 22, i.e., about 36 and 48 hours after the continuous series of hauls had been abandoned. Throughout the periods when the plankton hauls were being made, a record was kept of the light intensity on deck ; furthermore, two series of measurements of submarine illumination were carried out. The same submarine and deck photometer cases were employed and the same procedure followed as previously described by Clarke (1933a, 1938a).4 However, the Westinghouse " photox " cells in his apparatus were replaced by emission photoelectric cells of the " photronic " type (manu- factured by the Weston Electrical Instrument Corp.) since the former had been found to be somewhat variable. To measure temperature changes, a small thermometer was installed inside the case of the deck photometer, beside the cell, where it could be read through the pho- tometer window by raising the opal glass momentarily. The tempera- ture inside the photometer case during the day was found to vary between a minimum of 19° C. and a maximum of 35° C. Correction factors throughout this temperature range were determined from tests carried out in the laboratory at the end of the cruise, in collaboration with 4 In making these measurements the shadow of the hull is avoided by heaving the ship to with her stern toward the sun and suspending the sea photometer from the end of the mizzen boom (compare criticism by Poole, 1938). 260 WATERMAN, NUNNEMACHER, CHACE AND CLARKE RELATIVE LIGHT o in ill I I Illllll INTENSITY AT SURFACE -0 5 8 8 STORMY WEATHER - 28HRS OMITTED 0 8 o o o o o o o o o 0 <\j • S.S . t/> u. rt o re o. 01 O fc -o 2 CL» flj ^ •*-* U r" ^d o i: CM ->-> 5 <*H "^ i^ -^^ rt rt o O rt oj *o h. *o en 264 WATERMAN, NUNNEMACHER, CHACE AND CLARKE number of them was captured. Sixty per cent of the 153 specimens in the hauls were taken at 1,000 m. and none were caught at any time of day above 800 m. It is unfortunate that the 600 m. net on the midnight of July 20-21 failed to fish so that we do not know whether this prawn was present at this crucial point or not. It we assume on the basis of evidence afforded by some of the other crustaceans, as we have done above for Gennadas, that the population was about 200 m. lower in the water on July 22 than on the days preceding the storm of July 21, \ve can conclude that Hymenodora probably was present at 600 m. on the midnight of July 20-21. Owing to the lack of hauls deeper than 1,000 m. during the middle of the day, the lower boundaries of five out of the seven histograms are indeterminate. Indecisive as the present data may be, it is probable, nevertheless, that this animal was undertaking diurnal vertical migrations of about 200 m. amplitude. Parapasiphae sulcatifrons Smith (Fig. 3B) was the next most nu- merous decapod crustacean in our catches after Gennadas and Hy- menodora. Fifty-nine per cent of the 123 specimens 5 taken occurred at 800 m. and 10 per cent of them at 1,000 m. Furthermore, this prawn was never caught at 200 m. at any time of day and was taken at 400 m. only at midnight. Thus the general vertical distribution of this species was appreciably deeper in the water than that of Gennadas but not of Hymenodora. A diurnal vertical migration of Parapasiphae also occurred. The downward movement was essentially accomplished in the four hours from midnight to 4 A.M. and the upward migration apparently started slowly in the late afternoon although 75 per cent of its extent was ac- complished more rapidly after 8 P.M. The fact that these animals, too, were considerably lower in the water on July 22 is shown by the differences between the histograms for the first and second midnights in Fig. 3B. Only one of the six species of Sergestes, S. arcticus Kroyer, taken at this station occurred in sufficient numbers for our present purpose. An examination of the distribution in the catches of the 104 specimens of carapace length of 10 mm. or over (Fig. 3C) shows that the older stages of this prawn were living shallower in the water than either Parapasiphae or Gennadas since the center of the average vertical dis- tribution of the population was just above 600 m. rather than about 800 m. as it was for the other two forms. These Sergestes, moreover, 5 Immature, post-larval, and larval specimens were included in this count as well as in the figure. Although the immature specimens were not identified as to species, they were included here as P. sulcatifrons since all individuals whose devel- opment was sufficiently advanced to permit identification belonged to this species. HYMENODORA GLACIALIS 1200 I2M 4 R 8 I2N PARAPASIPHAE SULCATIFRONS 85 I2M 12 M I2M SERGESTES ARCTICUS 12 M 4 R 8 I2N 4 85 12 M D ACANTHEPHYRA PURPUREA 1200 12 M I2M FIG. 3. Decapods : data condensed as in Fig. 2. The numerals to the right of the histograms represent the numbers of specimens taken. These figures in- clude : A, larval, post-larval, and immature stages ; B, all stages ; C and D, adults. 266 WATERMAN, NUNNEMACHER, CHACE AND CLARKE also exhibited an extensive diurnal vertical migration of at least 200 and probably as much as 400 m. The absence of significant numbers of 'the species in the hauls made for the first midnight of Fig. 3C was probably due to the fact that most of the population was in less than 200 in. of water at this time. Some circumstantial support is given this explanation by the presence of some of these decapods at 100 m. at 4 A.M. and by the position of the popu- lation shown by the histogram of the second midnight in Fig. 3C since the stock of Sergestes by analogy with observations on several of the other Crustacea would be lower in the water by about 200 m. than it would have been at the same time under the conditions of the preceding days. There is some evidence from an analysis of the catches that those Sergestes arcticus of carapace length less than 10 mm. live at and mi- grate to slightly shallower levels than the larger individuals shown in the figure and discussed above, but hauls closer together than ours would have been necessary to determine this relation quantitatively. Only 68 Acanthephyra pur pur ea A. Milne-Edwards (Fig. 3D) were taken in our hauls, and these had a rather patchy distribution. Never- theless, there is some definite evidence for the occurrence of a diurnal vertical migration of at least 400 and possibly 600 m. Among the more primitive malacostracans two species of euphau- siids, two species of mysids, and two species of amphipods were caught in sufficient numbers to provide adequate samples for the present analysis. The relatively enormous number of 180,000 specimens of the eu- phausiid, Neinatoscelis megalops G. O. Sars (Fig. 4/4), was taken. The 1,543 adults among these were living mainly between 400 m. and the surface although 12 of them were taken in the only 1,200 m. haul. From the evidence of the very patchy occurrence of all the Neinatos- celis, one would conclude that this species swarms markedly, which may account for the extremely small numbers caught in the noon hauls. Clearly, however, a diurnal vertical migration at least 200 m. in extent was being accomplished by the adult Newia-toscelis, shown in the figure. The other schizopods taken in significant numbers were also migrat- ing vertically, as much perhaps as 600 m. in the case of Thysanopoda acutifrons Holt and Tattersall (Fig. 4B~), and 400 m. in those of the mysids, Boreomysis inicrops G. O. Sars (Fig. 4C), and Eucopia un- guiculata (Willemoes-Suhm) (Fig. 4D). A comparison of the histo- grams of the first and second midnights in Figs. 4B, C, and D, indicates that these organisms, as well as several of the previously discussed Crustacea, were 200 m. deeper in the water after the storm of July 21 DIURNAL BATHYPLANKTON MIGRATION 267 NEMATOCELIS MEGALOPS 1200 12 M 4 R 8 I2N 85 12 M THYSANOPODA ACUTIFRONS 1200 12 M I2M C oJ BOREOMYSIS MICROPS 12 M I2M EUCOPIA UNGUICULATA I2M I2M FIG. 4. Euphausiids and mysids. A, large adults ; B and C, adults ; D, all stages. See Fig. 2 for further explanation. 268 WATERMAN, NUNNEMACHER, CHACE AND CLARKE than they were on the previous two days. It is furthermore noteworthy that the Thysanopoda (Fig. 4B), as well as the Gennadas (Fig. 2), Para- pasiphae (Fig. 3 A], and Acanthephyra (Fig. 3D) apparently began to move slowly upwards in the water almost as soon as they had reached the lowest point of its downward excursion in the morning; this slow upward migration was succeeded at about 8 P.M. by a more rapid move- ment which brought the animal to the highest point of its migration about midnight. CYPHOCARIS ANONYX 1200 I2M 4 R 8 I2M VIBILIA PROPINQUA 1200 I2M 4 R 8 I2N 4 85 I2M FIG. 5. Amphipods. Adults. See Fig. 2 for further explanation. The two amphipods caught in sufficient numbers, Cyphocaris anonyx Boeck (Fig. 5 A) and Vibilia propinqua Stebbing (Fig. SB}, also show a diurnal vertical migration of about 400 m. amplitude. In the case of Vibilia, however, the considerable amount of swarming which is ap- parent obscures this migration somewhat. DIURNAL BATHYPLANKTON MIGRATION 269 DISCUSSION Since the hauls were made in such a way that our data give us in- formation about points separated in effect by 200 m. and 4 hours, a detailed analysis of the diurnal vertical migrations is clearly impossible. Nevertheless, our results show that all of the crustaceans investigated were accomplishing vertical migrations 200 to 400, and possibly 600 m. in extent, which is in general agreement with the results of recent investigators. For instance, certain of the larval stages of euphausi'ids taken by Fraser (1936, p. 154) were apparently moving upward 200-250 m. to reach the surface at night. Stubbings (1938, p. 27) found that species of the pteropods, Creseis and Hyalocylis, underwent a diurnal migration of 200 m. vertical extent and that some of the other pteropods may have been migrating even further. The copepod, Pleuromamma robusta, was shown by Mackintosh (1934, p. 92) to exhibit the most pro- nounced vertical migration of all the plankton animals of the antarctic surface waters; this crustacean rose into the upper 100 m. around mid- night and descended apparently below 600 m. during the day. Among the malacostracan crustaceans Hardy and Gunther '(1935, p. 240) ob- served that Euphausia frigida and E. triacantha underwent a diurnal vertical migration of at least 200 m. and that the amphipod, Para- themisto gaudichaudi, migrated a vertical distance of about 60 m. in the course of its diurnal movement. Since the latter animal was taken within the upper 100 m. at all hours, this observation is similar to those made on other shallow water plankton. Clarke (1934a, Figs. 2 and 3), for example, showed that Calanus finmarchicus adults and copepodid stages IV and V exhibited a diurnal migration of about 100 m. The rate of the diurnal vertical migrations may be roughly cal- culated from our results. If the speeds of downward movement of the centers of the populations for the various species, as shown in the figures, are taken to represent the average rates of movement for the organisms concerned, these are found to vary from 29 m. per hour for Nematoscclis to 125 m. per hour for Thysanopoda. The average for the eleven species of crustaceans is 67 m. per hour. The only com- parable data are those of Hardy and Gunther (1935, p. 240) mentioned above for the migration of the two species of Euphausia in which the upward migration apparently occurred at a rate of 100-200 m. per hour. In some of these organisms a slower upward movement during the day preceded the more rapid upward migration shortly before midnight. This phenomenon has been observed in shallow-water zooplankton by a number of workers and has been explained on the basis of a gradual 270 WATERMAN, NUNNEMACHER, CHACE AND CLARKE light adaptation of the animals (see Russell, 1931, p. 402; Kikuchi, 1938, p. 37). In spite of this slow upward trend, however, our ob- servations suggest that the upward migration in the early evening hours was as rapid as the downward movement in the early morning. A comparison of the swimming speeds of the animals in the plankton with the rates of these diurnal vertical migrations would be of great interest. Unfortunately, there are no data available for these or com- parable plankton organisms. Measurements have been made, however, of the rates of locomotion of several copepods and crustacean larvae (Russell, 1927, p. 221; Welsh, 1933). The latter found that Centra- pages would swim towards a light of an intensity that might be found 20-30 m. below the surface on a clear day at a rate of 82 m. per hour (actually measured over a path only 10 cm. long). As Parker (1902), Gardiner (1933), and Seiwell and Seiwell (1938) have shown, gravity alone causes copepods to sink at a rate of about 24 m. per hour. Thus, if any organism to which the above quoted rate of sinking is applicable were migrating downward 400 m. in 8 hours, it would sink 192 m. merely through the effect of gravity and would have to swim downward through the remaining 208 m. at an average rate of 26 m. per hour. For the upward trip the gravitational effect would add the 192 m. to the distance to be swum (see Clarke, 1934fr) so that the animal would have, in effect, to swim upward at a rate of 592 m. in 8 hours or 74 m. per hour. To consider it in another way, the organism in order to accomplish the downward migration would have to be ac- tively swimming towards the bottom at the experimentally determined rate of 82 m. per hour for only about one-third of the 8-hour period, whereas in order to move upward to the original level in the same length of time it would have to be swimming towards the surface at this rate during 7 of the 8 hours. As the pelagic malacostracans such as those discussed in this paper are known to be very active swimmers, they could presumably accom- plish such vertical migrations with greater ease than the animal postu- lated above on the basis of data obtained from copepods. The Crus- tacea in our analysis are larger, moreover, and, other things being equal, would sink more rapidly through the water than the copepods so that an even greater proportion of their downward movement may be ef- fected by the mere absence of any persistently oriented swimming activity. Many experiments have shown that several environmental variables, such as light and temperature, will affect the locomotor activity of zooplankton, but, as Clarke (1934&) pointed out, these factors under natural conditions would seem to work in the wrong direction to explain distribution as it is found. DIURNAL BATHYPLANKTON MIGRATION 271 Since the great majority of previous workers have recognized that light is the most significant single environmental factor in diurnal vertical migrations, it is interesting to correlate these phenomena with the changes in the intensity of illumination at various hours of the day (Fig. 1). During the first hour after daylight hecame measurable, its intensity in- creased at an exceedingly rapid rate. Similarly, the illumination dimin- ished at an equally rapid rate during the hour preceding the last measure- ment of the day. During the four hours in the middle of the day the illumination remained relatively constant and such changes in intensity as resulted from different degrees of cloudiness were of a much lower magnitude than those of early morning and late afternoon. The most striking point which appears in a comparison of the bio- logical and the light intensity data is that much of the migration appar- ently occurred when there was no light more intense than starlight even at the surface. Before sunrise most of the animals had already de- scended about half of their total downward distance, and several, such as Boreomysis, Eucopla, and Parapasiphae, had reached the deepest point in their migration by sunrise. In most cases, furthermore, the greater part of the upward migration must have been accomplished between the time of sunset and midnight. It is clear from these facts that the vertical migrations of these organisms cannot be explained simply on the basis of a negative photo- tropism which overshadowed a negative geotropism in the daytime and left it with a free hand at night. Similarly the hypothesis that the animals aggregate by means of a tropistic mechanism (Michael, 1911; Russell, 1934) in a region of optimum light intensity and follow the latter as it moves vertically at various hours of the day apparently will not explain the situation here. Such an hypothesis is further weakened by the lack of evidence in the present case for a general vertical scattering of the population throughout the water column at night when it would be free of the concentrating effect of light. Whether we assume that external environmental or internal physio- logical factors are controlling the diurnal migrations, the intensity of illumination penetrating from the surface must play an important role in these phenomena since it is the only daily environmental variable in deep water. It may act in combination with other external factors, such as temperature and gravity, since many investigators have shown experimentally that interactions of such factors may cause a tropistic reversal (Kikuchi, 1938, pp. 35-6). Or it may merely keep in phase cyclic internal physiological conditions of the organisms which them- selves regulate these migrations (see Russell, 1927, p. 225; Welsh, WATERMAN, NUNNEMACHER, CHACE AND CLARKE Chace, and Nunnemacher, 1937, p. 195). Many cases are known (Welsh, 1938, p. 136) of the persistence of internal diurnal rhythms under constant external conditions, but one would expect that in nature at least an occasional action of the external factor to which it owes its origin, in this case light, would be necessary to maintain the synchrony indefinitely. A second problem is posed by the fact that most of the Crustacea here considered were about 200 m. lower in the water on the midnight of July 22 than they were at the same hour two or three days before. Furthermore, the maximum illumination at the surface on July 22 was 66 per cent greater than the maximum on July 19. Previous studies have shown that certain organisms appear to live at depths where the light intensity is of a particular magnitude, i.e., an optimum intensity for the organism in question ; these organisms may, as a con- sequence, appear in shallower waters on dull days or in high latitudes (Russell, 1933, p. 569; Murray and Hjort, 1912, p. 664-6); thus the light intensity itself may have been responsible for the observed differ- ence in distribution. It is clear, howrever, that the amount of illumination could not ex- plain so great a change in distribution. Since actual measurements down to 84 m. (Fig. 6) showed that the light is reduced by a factor of 10, for every 25 m. of water, the observed change in light intensity at the surface would result in a difference of only about 5.5 m. in the depth at which any given intensity of light would be found. Even in the clearest sea water yet measured (see below) this change in light intensity at the surface v/ould result in a maximum change of only 50 m. in the depth to which any particular intensity of light would penetrate. An alternative causal or contributory factor to the observed lower position of the plankton on July 22 may have been the storm of the previous day, but it is not likely that a relatively mild disturbance of this sort at the surface would be appreciable at a depth of over 200 m. Furthermore, Hardy and Gunther (1935, p. 268) conclude from their observations that, "... the state of the sea has little or no bearing on the number of organisms in the surface layers. . . ." A more important point is the fact that simultaneous measurements on the horizontal patchiness of the plankton were not made. Conse- quently we cannot tell whether or not the observed difference might rather be due to the fact that on the later day we were sampling a different population with a different vertical distribution pattern from that of the previous days (Russell, 1933, p. 569; Hardy, 1936). One of the most important questions raised by previous work on the diurnal vertical migrations of bathypelagic organisms is how deep DIURNAL BATHYPLANKTON MIGRATION 273 down in the water do such migrations occur. Most of the crustaceans showing diurnal vertical migrations discussed in this paper were living at depths of 600-800 m. during the four hours during the middle of the day. If, as pointed out above, we recognize the necessity of light in one way or another for the vertical migrations of these animals, we must assume that at some time during the day the organism is affected 001 i i PERCENTAGE OF SURFACE LIGHT Q05 Ql Q5 5 10 i i i 1 1 1 1 i i i i 50 100 0 1 1 1 1 10 20 30 40 50 m 60 70 80 90 FIG. 6. Relation between depth and irradiation expressed as a percentage of the light just over the surface (logarithmic scale). Station 2894. Lat., 39° 06' N. ; Long., 70° 16' W. July 22, 1937. Sky : clear. Sea : moderate, white caps. Series 450 measured by photronic cell with opal disc. Series 451 measured by photronic cell with "green" filter combination (Schott-Jena BG-18 and GG-11) and opal disc. by light. Clearly, under any given physiological conditions, there is a minimum intensity of illumination which would be just sufficient to affect any particular animal ; in the open sea there would be a definite maximum depth beyond which this minimum light intensity would never penetrate even at noon on the brightest days. A diurnal vertical migra- tion in which illumination from the surface was supposed to have any m T) .10 m 451 K=.062 450 274 WATERMAN, NUNNEMACHER, CHACE AND CLARKE controlling effect would thus be impossible below this depth. As pointed out above, however, vertical migrations could be temporarily maintained under constant conditions by a diurnal physiological rhythm within the organism. Thus, if unusual conditions such as a storm should cut down the incident light so that deep-living animals did not come under the influence of light for several days, there would not be a wholesale loss of the population through continued downward migra- tion or random scattering into unusually great depths (see Russell, 1930), but the diurnal migrations would continue more or less as usual under the guidance of the physiological rhythm. In view of these facts the question of how much light penetrates to the depths at which the migrating Crustacea are living during the day- light hours is one of great interest. The results of our two series of light penetration measurements, shown in Fig. 6, indicate that the trans- parency of the water was relatively uniform down to 84 m., the limit of measurement, with an average value of the extinction coefficient 6 of £==0.092 for the green component of daylight. Hence, the intensity of illumination was reduced to one-tenth its value by passage through 25 m. of water. Assuming that the transparency of the water remained the same below the depth of our deepest measurement and extrapolating we find that the illumination at 800 m. would be 10~32 times its intensity at the surface. But if we assume that the water below 100 m. is as clear as any sea water which has been measured, we may on this basis calculate the maximum amount of light possible at any depth. Since the extinction coefficient varies enormously with wave-length, the actual result reached will depend on the part of the spectrum considered. In the clearest water yet known the intensity of light in the green wave-lengths (Clarke, 193Sfr) was reduced to one-tenth by 61 m. of water. Thus if the water below 100 m. at our station was as clear as this, at 800 m. the intensity of illumination would be 10~15 times the value at the surface. In the clearest water yet known the blue wave-lengths were reduced to one-tenth their intensity in 230 m. (Clarke, 1933a).7 If water of like transparency existed below 100 m. at our station, the illumination of 800 m. would be a little less than 10"r times the surface intensity;8 this value is between 100 and 1,000 times higher than the threshold intensity for human vision. 6 In the equation I/Io = e~kL, where L is the thickness of the water layer in meters through which the intensity is reduced from I0 to /. 7 Basing our calculations on the most transparent strata only. 8 Assuming the same reductions in the upper 100 m. for the blue component as for the green, which is known (Oster and Clarke, 1935) to be approximately true in this region of the Atlantic. DIURNAL BATHYPLANKTON MIGRATION 275 We have some evidence, however, that conditions as favorable for penetration as this latter case for blue light did not exist at our slope water station since on a bright, calm day Beebe (1934, p. 276) observed through the window of the bathysphere that at a depth of 610 m., the water appeared absolutely dark to the human eye. But Beebe was not able to look upward in the water so that he could not tell at what depth the downward component of the light had become indistinguishable from the blackness below.9 On the other hand, the water of the Sar- gasso Sea, where Beebe made his observations, is characteristically (Clarke, 1936a) nearly twice as transparent in the upper 100 m. for the green region of the spectrum as was the water at our present station. It would be of great interest to compare the amounts of light present at 800 m., where migrations were definitely going on, with the visual sensitivity of these Crustacea. Unfortunately, no experimental studies have yet been made on vision in any deep-sea animals. Nevertheless, from previous work on crustaceans and other arthropods with com- pound eyes, it is possible to draw two extremely general conclusions : first, that arthropods are more sensitive to the blue wave-lengths of light than to the green (see Waterman, 1937, p. 464 for table sum- marizing pertinent data) and second, that for these organisms the minimal light intensity for response is of the order of magnitude of 10~6 to 10~7 millilamberts (Crozier, 1938, personal communication).10 An intensity of illumination of 10~6 millilamberts is equivalent to about 10~10 times that of full noon sunlight and is the value used by Clarke (1936&) as a threshold intensity for determining the depth at which fish can see. A crustacean with a visual threshold of such a magnitude in the green wave-lengths certainly could not be stimulated by this region of the spectrum when the intensity of the latter is re- duced to 10~15 times that at the surface. Since this would be the il- lumination present at 800 m., when we assume that the water below 100 m. was as clear as any sea water that has been observed, we may conclude that under these conditions the crustaceans could not be re- 9 Presumably this directional effect would still be appreciable at great depths since directional differences in intensity were recorded photographically by Hel- land-Hansen (Murray and Hjort, 1912, p. 252) at 500 m. and since more recent work of Whitney (1938, p. 211) has demonstrated that in optically homogeneous waters the ratio of the intensities of light passing downward to those of the light scattered upward is a constant (see also Pettersson, 1938). Of interest in this connection is the occurrence of upwardly (dorsally) directed "telescopic" and stalked eyes in a number of different families of deep-sea fishes of the " twilight zone" of the seas (Brauer, 1908, p. 233) which would suggest that this directional difference in the light intensities is of biological significance even at considerable depths. 10 Any value obtained for a sensory threshold will, of course, depend largely on the exact experimental conditions under which it was determined. 276 WATERMAN, NUNNEMACHER, CHACE AND CLARKE spending to light from the surface. But if the crustacean had an equally low visual threshold for blue light, our calculations have shown that in the clearest water observed there would be almost 1,000 times as much light present at 800 m. as would be necessary for the thresh- old of vision. Under such maximum conditions even at 1,000 m., where the immature Hymenodora in our catches possibly gave some indications of a diurnal vertical migration, more than 100 times the necessary amount of light would be present. The threshold intensity for vision, 10~10 of full sunlight intensity, would occur at a depth of nearly 1,500 m. under such conditions. Thus, the water below the surface layers could be, as it probably was, considerably less transparent than the clearest sea water ever ob- served and still permit sufficient visible light for affecting the reactions of bathypelagic organisms to penetrate to depths of about 800 m. The preceding discussion makes it clear that several important as- pects of this problem must be investigated more thoroughly before a satisfactory solution will be possible. These are in brief: (1) actual measurements of the transparency of the water at greater depths and of the relative intensities of the light in various directions; (2) more detailed work on the maximum depth at which diurnal vertical migra- tions occur; (3) measurements of the photosensitivity of deep-sea animals and correlations of the maximum wave-length sensitivity to the wave-length of the most penetrating part of the spectrum ; (4) ex- perimental studies on deep-sea animals designed to elucidate the nature of any diurnal physiological rhythms and their relation to the organisms' behavior in the sea. SUMMARY 1. A study was made with the aid of closing nets of the diurnal vertical migrations of bathypelagic organisms at a station in continental slope water of the western North Atlantic. 2. While the hauls were being made a continuous record of the light intensity at the surface was kept. 3. The penetration of light into the upper 84 m. of water was directly measured photometrically; the average extinction coefficient for green light was k = 0.092. 4. All of the malacostracan Crustacea (to which the detailed results presented in this paper are limited) which occurred in sufficient numbers for analysis exhibited diurnal migrations 200 to possibly 600 m. in vertical extent. 5. The speed of vertical movement in these migrations varied from 24 to 125 m. per hour among the various crustaceans. DIURNAL BATHYPLANKTON MIGRATION 277 6. A considerable part of the migrations took place while the light intensity even at the surface was no greater than starlight. 7. Several Crustacea living at 800 m. during the day showed exten- sive diurnal vertical migrations. 8. It is concluded, however, that whether the migrations are regu- lated by external environmental or by internal physiological factors, at some time of day the organisms concerned are affected by light penetrat- ing from the surface. 9. Calculations made from the light penetration data indicate that the amount of light probably present during the middle of the day at the depths where the animals were migrating was adequate to support this conclusion (8). LITERATURE CITED BEEBE, W., 1934. Half Mile Down. Harcourt, Brace. New York. BRAUER, A., 1906. Die Tiefsee-Fische. I. Systematischer Teil. Wiss. Erg. Deutschen Tiefsee-Exp. ..." Valdivia" Bd. 15, Lief. 1. BRAUER, A., 1908. Die Tiefsee-Fische. II. Anatomischer Teil. Ibid., Bd. 15, Lief. 2. CHIERCHIA, G., 1885. Collezioni per studi di scienze naturali fatti . . . dalla . . . " Vettor Pisani." Riv. Marittima, Sept. 1885. CHUN, C., 1890. Die pelagische Thierwelt in grossen Meerestiefen. Verhandl. Gesell. Deut. Naturf. u. Aerste, 63 (Pt. 1) : 69-85. CLARKE, G. L., 1933a. Observations on the penetration of daylight into mid- Atlantic and coastal waters. Biol. Bull., 65: 317-37. CLARKE, G. L., 19336. Diurnal migration of plankton in the Gulf of Maine and its correlation with changes in submarine irradiation. Ibid., 65 : 402-436. CLARKE, G. L., 1934a. Further observations on the diurnal migration of copepods in the Gulf of Maine. Ibid., 67 : 432-55. CLARKE, G. L., 19346. Factors affecting the vertical distribution of copepods. Ecol. Monogr., 4: 530^0. CLARKE, G. L., 1936a. Light penetration in the western North Atlantic and its ap- plication to biological problems. Rapp. et Proces-verb. Con. Perm. Int. E.rpl. Mer, 101 (Pt. 2, No. 3) : 3-14. CLARKE, G. L., 19366. On the depth at which fish can see. Ecol., 17 : 452-6. CLARKE, G. L., 1938a. Seasonal changes in the intensity of submarine illumination off Woods Hole. Ibid., 19 : 89-106. CLARKE, G. L., 19386. Light penetration in the Caribbean Sea and in the Gulf of Mexico. Jour. Mar. Res., 1 : 85-94. EKMAN, S., 1935. Tiergeographie des Meeres. Akademische Verlagsgesellschaft. Leipzig. FOWLER, G. H., 1905. Appendix on the vertical distribution and movement of the Schizopoda. Trans. Linn. Soc. Lond., Ser. 2, 10 : 122-9. FOWLER, G. H., 1909. Biscayan plankton collected during a cruise of H.M.S. "Research," 1900. Pt. 12: The Ostracoda. Ibid.: 219-336 FRASER, F. C., 1936. On the development and distribution of the young stages of krill (Euphausia superba). Discov. Reports, 14: 1-192. GARDINER, A. C., 1933. Vertical distribution in Calanus finmarchicus. Jour. Mar. Biol. Ass'n, N.S., 18 : 575-610. HARDY, A. C., 1936. Observations on the uneven distribution of oceanic plankton. Discov. Reports, 11: 511-38. «*.•»-**• 1 • »-*\! WATERMAN, NUNNEMACHER, CHACE AND CLARKE HARDY, A. C., AND E. R. GUNTHER, 1935. The plankton of the South Georgia whaling grounds and adjacent waters, 1926-1927. Ibid., 11 : 1-456. ISELIN, C. O'D., 1936. A study of the circulation of the western North Atlantic. Papers in phys. oceanogr. and metcorol. Mass. hist. Tech., Woods Hole Oceanogr. hist., 4 (No. 4) : 1-101. KIKUCHI, K., 1938. Studies on the vertical distribution of the plankton Crustacea. II. The reversal of phototropic and geotropic signs of the plankton Crus- tacea with reference to vertical movement. Rcc. Oceanogr. Works in Japan, 10: 17-41. LEAVITT, B. B., 1935. A quantitative study of the vertical distribution of the larger zooplankton in deep water. Biol. Bull., 68: 115-30. LEAVITT, B. B., 1938. The quantitative vertical distribution of macrozooplankton in the Atlantic Ocean basin. Ibid., 74: 376-394. MACKINTOSH, N. A., 1934. Distribution of the macroplankton in the Atlantic sector of the Antarctic. Discov. Reports, 9 : 65-160. MICHAEL, E. L., 1911. Classification and vertical distribution of the Chaetognatha of the San Diego region. Univ. Calif. Publ. Zoo/., 8: 21-186. MURRAY, J., 1885. In Reports of the " Challenger " E.vped. Vol. 1, Narrative of the Voyage, p. 218. MURRAY, J., AND J. HJORT, 1912. The Depths of the Ocean. Macmillan Co. London. OSTER, R. H., AND G. L. CLARKE, 1935. The penetration of the red, green, and violet components of daylight into Atlantic waters. Jour. Opt. Soc. Amer., 25 : 84-91. PARKER, G. H., 1902. The reactions of copepods to various stimuli, and the bear- ing of this on daily depth migrations. Bull. U. S. Fish. Comm. 1901, 21: 103-23. PETTERSSON, H., 1938. Measurements of the angular distribution of submarine light. Rapp. et Proces-vcrb. Con. Perm. hit. E.vpl. Mer, 108 (Pt. 2, No. 2) : 9-12. POOLE, H. H., 1938. Review of W. M. Powell and G. L. Clarke, 1936. Jour. Con. Perm. Int. Expl. Mer, 13 : 111-3. RANG, M., 1828. Notice sur quelques Mollusques nouveaux appartenant au genre Cleodore, et etablissement et monographic du sous-genre Creseis. Ann. des Sci. Nat., 13 : 302-19. ROSE, M., 1925. Contribution a 1'etude de la biologic du plankton ; le probleme des migrations verticales journalieres. Arch. Zo'61. Exper. Gen., 64: 387-542. RUSSELL, F. S., 1927. The vertical distribution of plankton in the sea. Biol. Rev., 2 : 213-62. RUSSELL, F. S., 1930. Do oceanic plankton animals lose themselves? Nature, 125: 17. RUSSELL, F. S., 1931. The vertical distribution of marine macroplankton. X. Notes on the behaviour of Sagitta in the Plymouth area. Jour. Mar. Biol. Ass'n, N.S., 17: 391-414. RUSSELL, F. S., 1933. On the biology of Sagitta. IV. Observations on the natural history of Sagitta elegans Verrill and Sagitta setosa J. Miiller in the Plymouth area. Ibid., 18 : 559-74. RUSSELL, F. S., 1934. The vertical distribution of marine macroplankton. Ibid., 19: 569-84. SEIWELL, H. R., AND G. E. SEIWELL, 1938. The sinking of decomposing plankton in sea water and its relationship to oxygen consumption and phosphorus liberation. Proc. Amer. Phil. Soc., 78: 465-81. SIGSBEE, C. D., 1880. Description of a gravitating trap for obtaining specimens of animal life from intermediate ocean-depths. Bull. Mus. Comp. Zobl., 6: 155-8. DIURNAL BATHYPLANKTON MIGRATION 279 STUBBINGS, H. G., 1938. Pteropoda. John Murray Exp. Sci. Rep., 5 : 15-33. WATERMAN, T. H., 1937. The relative effectiveness of various wave lengths for the photokinesis of Unionicola. Jour. Cell, and Comp. Physiol., 9 : 453-67. WELSH, J. H., 1933. Light intensity and the extent of activity of locomotor mus- cles as opposed to cilia. Biol. Bull., 65 : 168-74. WELSH, J. H., 1938. Diurnal rhythms. Quart. Rev. Biol, 13: 123-39. WELSH, J. H., AND F. A. CHACE, JR., 1937. Eyes of deep sea crustaceans. I. Acanthephyridae. Biol. Bull, 72 : 57-74. WELSH, J. H., AND F. A. CHACE, JR., 1938. Eyes of deep-sea crustaceans. II. Sergestidae. Ibid., 74: 364-75. WELSH, J. H., F. A. CHACE, JR., AND R. F. NUNNEMACHER, 1937. The diurnal migration of deep-water animals. Ibid., 73 : 185-96. WHITNEY, L. V., 1938. Transmission of solar energy and the scattering produced by the suspensoids in lake waters. Trans. Wisconsin Acad., 31 : 201-21. THE REPRODUCTION OF LIMACINA RETROVERSA (FLEM.) SIDNEY C. T. HSIAO (From the Biological Laboratories, Harvard University, and the Woods Hole Occanographic Institution?- Woods Hole, Massachusetts} INTRODUCTION The structure of the reproductive system of Limacina retroversa and the process of spermatogenesis of this species have been described in a previous paper (Hsiao, 1939). It was shown that very small animals are sexually undifferentiated. The gonads of sexually differentiated but small-sized individuals are protandric and many of them are often with- out female tissue, that is, " pure male," while those of the bigger animals are essentially hermaphroditic, showing different proportions of male and female tissues. Some individuals function as male, others as female and still others both as male and female at the same time. It is the pur- pose of this study to examine the relationship between size and sexual phases and the time of the occurrence of the latter so as to determine their sequence in the reproductive history of this form and to compare it with other mollusks whose sexual history is known. DESCRIPTION OF MATERIAL The material was collected by the research vessel " Atlantis " during a series of cruises in the Gulf of Maine between June, 1933 and Sep- tember, 1934. Specimens collected at the following periods were used : Date Designated as December 2-1 1 , 1933 December specimens January 9-13, 1934 January specimens March 21-28, 1934 March specimens April 17-May 13, 1934 April specimens May 21-June 2, 1934 May specimens June 25-July 1, 1934 June specimens September 17-24, 1934 September specimens The specimens were caught by vertical hauls made with a 1.5-meter Heligoland larva net drawn from a level near the bottom up to the sur- face and fixed on board the ship in 2 per cent formalin in sea water. 1 Contribution No. 206 from the Woods Hole Oceanographic Institution. 280 REPRODUCTION OF LIMACINA RETROVERSA 281 Limacina rctroversa were later sorted out from the other species and specimens were selected from one to three stations from each cruise so as to make a series covering nearly the whole range of size. Each indi- vidual was measured with shell-opening facing the observer under a binocular microscope with a calibrated ocular micrometer. The diameter of the largest whorl of the spiral body of the animal, excluding the pro- truding margin of the shell-opening (A-B in Fig. 1 ) , was measured and used to designate the size of the specimens studied. FIG. 1. Camera lucida drawing of the shell of Limacina retroversa. A B: maximum diameter of the shell used for comparison of size. X 20. From a study of the size distribution of Limacina rctroversa from 1933-34 and the movement of the water masses, Redfield (1939) has shown that at least two different populations entered the Gulf of Maine during the period of observation. The first group, called "population A," invaded the Gulf during December, 1933 and January, 1934, and later in the spring a second group, " population B," entered the basin during April and May from the same general direction. The positions of the hydrographic stations from which specimens used in this study were taken are shown in Fig. 2. In this map the solid circles with the abbreviations of the months beside them show the position and time of collection of the material from the population of Limacina retroversa, which has been identified as " population A " by Redfield. The posi- tions of the stations providing the material fall within the region where large numbers of " population A " were found (as shown by Redfield's map), that is, in or near the center of density of this population in the different months. On the same map (Fig. 2) the open circles, with the abbreviations of the months in parentheses under them, indicate the sta- tions in April and May from which specimens were selected which repre- 282 SIDNEY C T. HSIAO sent " population B." From these specimens the reproductive history of the species is worked out. The following table (Table I) shows the number of individuals used from each cruise grouped into size classes with a class interval of 0.3 mm. FIG. 2. Map of the Gulf of Maine showing the hydrographic stations from which samples of Limacina retroversa are selected for microscopic study. • : stations for " population A." O : stations for " population B." The method employed in making serial sections of these specimens has been described in a previous paper. The following observations are based upon a comparative study of these serial sections. RELATION BETWEEN SIZE AND SEX The relation between size and sex in Limacina retroversa has been examined from two different angles: (1) The gonadal types and size, and (2) the relation between functional types and size. REPRODUCTION OF LIMACINA RETROVERSA 283 Gonadal Types and Size The following morphological types are seen among Limacina. (a) Sexually undifferentiated individuals all of which are small-sized; (&) " pure males " which are nearly completely male in constitution and are somewhat larger; and (c) hermaphrodites, which have various propor- tions of male and female tissues in the ovotestis and are generally of large size. The occurrence of the different morphological types of in- dividuals in Limacina retroversa, in order of size, is shown in Table II. Sexually Undifferentiated Individuals. — Of all the Limacina retro- versa sectioned the largest individual which has a gonad still in a sexu- ally undifferentiated condition measures 0.85 mm. in diameter. Indi- TABLE I Table showing the number of individuals from each period of collection which were used in this study. The individuals are separated according to size into groups with a class interval of 0.3 mm. • Total Size groups in mm. num- ber Month of cruise ob- tained 0 to 0.3 to 0.6 to 0.9 to 1.2 to 1.5 to 1.8 to 2.1 to during 0.3mm. 0.6 mm. 0.9mm. 1.2 mm. 1.5 mm. 1.8 mm. 2.1mm. 2.5 mm. month Dec. ... 3 13 9 8 33 Tan.. 13 12 4 3 32 Mar 4 4 7 10 1 26 Apr. . 5 9 5 6 1 26 May 4 11 7 6 3 6 2 1 40 June 4 3 6 2 6 5 2 28 Sept 1 6 8 7 4 3 29 viduals larger than this size all show definite gametogenesis. All speci- mens with a diameter less than 0.6 mm. are sexually undifferentiated. Hence, the process of sexual differentiation must commence during the period when the animal is between 0.6 mm. and 0.85 mm. in diameter. This is brought out more clearly when the percentage of sexually un- differentiated individuals is plotted against size as shown in Fig. 3. "Pure Males."- -These are animals with gonads which contain little or no noticeable amount of ovarian tissue. They correspond to Coe's " true males " or Orton's " pure males." These individuals are confined to comparatively small-sized groups. Table II, column 8, shows the sudden appearance and the steady but more gradual decrease of " pure males " among animals arranged in the order of increasing size. Among larger animals this type of individual gives place to those which have 284 SIDNEY C. T. HSIAO various proportions of feminine reproductive cells in the gonads. These are the hermaphrodites described in the next paragraph. On the aver- age, about 22 per cent of the specimens less than 1.2 mm. in diameter belong to the " pure male " class. Many of these males are functional at the time of fixation as shown by the presence of mature spermatozoa in the gonoduct. Animals with Various Proportions of Male and Female Tissues -in the Gonads. — In addition to the sexually undirferentiated animals and the " pure males " a small number of animals less than 1 mm. in diameter TABLE II Proportion of different morphological types of Limacina retroversa separated into groups without regard to the time of collection. Animals smaller than 1.2 mm. in diameter are separated into 0.1 mm. classes, those larger than 1.2 mm., into 0.3 mm. classes. 1 2 3 4 5 6 7 8 9 10 Size in mm. Total num- ber Un- differ- enti- ated Pure male >50% male 50% male Per- cent- age of <50% male 0.4 3 3 0 0 0 100 0 0 0 0.5 3 3 0 0 0 100 0 0 0 0.6 13 8 3 2 0 62 23 16 0 0.7 15 5 7 3 0 33 47 20 0 0.8 24 3 8 13 0 13 33 54 0 0.9 12 0 3 7 2 0 25 58 17 1.0 14 0 2 8 4 0 14 57 28 1.1 15 0 2 10 3 0 13 67 20 1.2 to 1.49 44 0 0 34 10 0 0 77 23 1.5 to 1.79 34 0 0 26 8 0 0 77 23 1.8 to 2.09 19 0 0 11 8 0 0 58 42 2.1 to 2.39 4 0 0 2 2 0 0 50 50 2.4 to 2. 7 2 0 0 1 1 0 0 50 50 may have gonads containing various proportions of male and female tis- sues at the same time. Of all the animals greater than 1.1 mm. in diam- eter no undifferentiated or " pure males " were found. They all contain spermary and ovarian tissues in the same gonad. Following the criteria described in a previous paper these hermaphroditic animals can be classi- fied as hermaphroditic males or hermaphroditic females according to whether they have more than 50 per cent or less than 50 per cent of their gonadal tissues consisting of male reproductive elements. The propor- tion of these two kinds of individuals in the different size groups is REPRODUCTION OF LIMACINA RETROVERSA 285 shown in Table II. Column 5 shows the number of individuals with more than 50 per cent of their gonadal tissue made up of masculine ele- ments, excluding those which are already listed in column 4 as " pure males." It shows the number of hermaphroditic males only. The per- centages of hermaphroditic males of this class are given in column 9. Column 10 shows the percentages of hermaphroditic females based on the numbers shown in column 6. A comparison of columns 9 and 10 will show that the minimal size of the hermaphroditic females is much greater than that of the hermaphroditic males. This occurrence of hermaphro- ditic males as well as " pure males " in smaller-sized animals when com- pared with the hermaphroditic females indicates that the male repro- ductive tissue is developed when the animals are smaller than is the case with the female cells of reproduction. Owing to the small number of individuals in each size group among the larger animals the data in the lower half of Table II have been re- arranged so that the class interval is three times as large as in the upper half. It will be seen that among animals smaller than 1.2 mm. the pro- portion of hermaphroditic males increases steadily while among those larger than this size the percentages decrease from 77 to 50. On the other hand, there is only a small proportion of hermaphroditic females among the smaller animals, but among the larger ones the percentage increases from 23 to 50. Approximately speaking, before the animals reach 1 mm. in diameter the proportion of males increases very rapidly, indicating an early proliferation of the male reproductive cells giving rise to the types of males whose gonads are predominantly masculine in appearance. After the animals grow to a size greater than 1.5 mm. in diameter the female reproductive cells overtake the male in development, thus bringing about a decrease in the proportion of hermaphroditic males after the 1.8 mm. size group and an increase in the predominantly femi- nine type. The decrease in the proportion of animals which are pre- dominantly masculine continues until the animals are about 2 mm. in diameter, from which stage onward the two sexual types are equally numerous (Fig. 3). The Relation between Functional Types and Size In addition to the above morphological types, we can analyse our ma- terial into functional types. From the structure and the place of oc- currence in the body of mature germ cells each Liniacina's state of re- productive activity at the time of fixation can be deduced. Thus, when mature spermatozoa are seen in the gonoduct, or at the base of the penis, the individual is obviously functioning as a male and can be classified as an active male. When mature ova are found in the ovotestis the indi- 286 SIDNEY C. T. HSIAO vidual can be considered as a mature female, while the presence of ripe eggs in the gonoduct indicates that the individual is a spawning female. The simultaneous presence of mature spermatozoa in the gonoduct and ripe ova in or out of the gonad is an indication that the individual is a functional hermaphrodite, to use Coe's terminology. The occurrence of these functional types is shown in Table III. 100 60 60 40 HERMAPHRODITIC FEMALE ° ( HERMAPHRODITIC MALE SEXUALLY UNDIFFERENTIATED 2.4 2.1 1.8 1.2 U N CO •0.9 -0-6 0-5 FIG. 3. Diagram showing the relation between size and various morphological types of gonads of Limacina retroversa (expressed in percentages of the total of each size group). It has been observed that some individuals as small as 0.7 mm. in diameter may function as males. But the majority of specimens actively functioning as males only are between 1.2 mm. and 1.8 mm. in diameter (column 4). In larger-sized groups their place is taken by the females and individuals functioning both as male and female simultaneously. The proportion of mature females with ripe eggs in the gonad, on the contrary, increases with the increase of size. The class with the REPRODUCTION OF LIMACINA RETROVERSA 287 largest number of mature females, as shown in Table III, is the 2.1 to 2.4 mm. group. The smallest Limacina with mature ova is more than 1 mm. in diameter. At the time of fixation many mature females were in the act of spawning, as evidenced by the presence of ripe ova at vari- ous points along the gonocluct. Roughly speaking, about half of the mature females in each size group were in the act of spawning when fixed. The proportion of the third functional type, functional hermaphro- dites, in each size group is shown in column 6 of Table III. The chief occurrence of this type of animal is among the larger-sized specimens. It will be seen from Table III that although both specimens of the 2.4— 2.7 mm. size group are functional hermaphrodites, the number of indi- viduals used is so small that no great importance can be attached to the TABLE III Proportion of different functional types of Limacina retroversa expressed as percentage of each size group irrespective of the time of collection. Class interval: 0.3 mm. 1 2 3 4 5 6 7 Size groups (mm.) Total number Undiffer- entiated Active male Mature female Functional hermaphro- dite Without signs of activity 0.3-0.6 6 per cent 100 per cent 0 per cent o per cent 0 per cent 0.6-0.9 52 30 17 0 0 53 0.9-1.2 1.2-1 5 41 44 0 o 39 48 15 9 5 14 41 29 1.5-1.8 43 o 47 18 21 14 1.8-2.1 19 0 5 42 10 43 2 1-2.4 4 o 0 50 50 0 2.4-2 7 2 o 0 0 100 0 deduced percentage of this type of animal. Until more large specimens are studied it seems better to leave this group out in our examination of the relation between the functional types and size. In the 2.1-2.4 mm. size group mature females and functional hermaphrodites are equally numerous. When the gonads of these functional hermaphrodites are examined and classified in terms of the proportion of masculine and feminine tissue in them the following relationship is seen. Type of gonad: Percentage of specimens With more than 75% male tissue 57% With 50%-75% male tissue 33% With 25%-50% male tissue . 10% With less than 25% male tissue 0% SIDNEY C. T. HSIAO Most functional hermaphrodites have more than 75 per cent of the gonadal tissue made of male reproductive cells, one-third of them have a 50 per cent to 75 per cent male type of gonad, while no Limacina with less than 25 per cent male tissue in the gonad, that is, no very pronounced female type of individual, functions as hermaphrodite. Since the gonad of this species of pteropod changes from predominantly male to pre- dominantly female type as the animals grow (see Fig. 3), this relation between the proportion of functional hermaphrodites and the amount of masculine tissue in the gonad can be interpreted to mean that animals possessing " more than 75 per cent male " type of gonad later have their female germ cells developed while the male phase of reproduction is still functioning, thus giving rise to a large proportion of hermaphrodites functioning as male and female at the same time. By the time the gonads change over to the predominantly feminine condition the male phase of sexual activity also disappears and hence few or no functional hermaphrodites are seen among individuals with a pronounced feminine type of gonad. Figure 3 summarizes the facts about the relation of size to sexuality. Sexual differentiation commences during the time when the animals are between 0.6 and 0.9 mm. in diameter. The male germ cells proliferate in smaller-sized individuals giving rise to the " pure males " and her- maphroditic males before the animal reaches 1.2 mm. in diameter. The proportion of " pure males " decreases with increase in size. As the size increases further, the proportion of hermaphroditic males decreases, while that of the hermaphroditic females becomes greater and greater until the two sexual types are equal in number among animals greater than 2 mm. in diameter. Functionally the situation is summarized in Fig. 4. The male phase starts among smaller Limacina than the functional female and hermaph- roditic phases. The highest proportion of actively functioning males is found among animals 1.2 to 1.8 mm. in diameter. The large propor- tion of mature females and functional hermaphrodites occurs among larger animals replacing the male phase in those size groups. Females become mature after they have attained a diameter greater than 1 mm. It is probable that mature females and functional hermaphrodites are equally numerous among animals greater than 2 mm. in diameter. The percentage of each of these three functional types among the different size groups is shown in Fig. 4. * SEASONAL CHANGES IN THE REPRODUCTIVE CYCLE When the condition of the reproductive system of individuals col- lected at different times of the year is examined comparatively it is found REPRODUCTION OF LIMACINA RETROVERSA 289 that, as a general rule, in winter individuals are small protandric males and that later hermaphrodites and functional males and females appear. We shall first take up the morphological types of sexual individuals and the question of the proportion of these sexual types in different months, then the appearance of different sexual activities at various times of the year and lastly the question of the reversal of sex predominance. It has o/ /o 100 80 60 40 20 0 100 80 60 40 0 MATURE FEMALE r u n i> 1 1 vj n M L. HERMAPHRODITE . . r-r~ 100 FUNCTIONAL MALE 80- 60 • 4O . T \J 20- O r — . .9 2.1 27 MM. SIZE FIG. 4. Distribution of the three functional types of Limacina rclroversa among various size groups. Functional males occur among smaller-sized animals than functional hermaphrodites and mature females. been shown by Redfield (1939) that " population A " can be clearly dif- ferentiated from " population B " from December up to May and after that time there is a good deal of mixing of populations with " population B " predominating over the descendants and remnants of " population A " and new recruits into the Gulf. In the following description data from December to May refer to " population A " and those after May, to the mixed populations. 290 SIDNEY C. T. HSIAO Change in Gonad Structure Young sexually undifferentiated individuals were found among the specimens taken in each month except March and April. Figure 5 shows the percentage of the individuals taken each month which are sexually indefinite. During the winter months of December and January "popu- lation A," which entered the Gulf then, contained 14—16 per cent of sexually undifferentiated animals. In March and April these animals drifted to the western portion of the Gulf and grew larger, as shown by Redfield, and there remained no sexually undifferentiated individuals. After May, a portion of the animals examined, 19-20 per cent, were 100 = " POPULATION A" = "POPULATION B" 30 H 20- DEC JAN MAR APR MAY JUN SEP FIG. 5. Histogram showing the percentage of sexually undifferentiated indi- viduals in each month. " Population A " solid black, " population B "shaded with straight lines. About 15 per cent of the material caught (in the winter) from " population A " and 20 per cent from " population B " are sexually undifferentiated. Absence of immature material in March and April indicates that members of " popu- lation A " are all differentiated by March. again sexually indefinite. This coincides with the entrance into the Gulf of a new population in the east and the spawning of " population A " in the west portion of the basin. As far as " population A " is con- cerned, sexual differentiation is completed by the month of March and all the individuals of this population were either developing in the direc- tion of males or hermaphrodites. Animals which were sexually differentiated have been divided into male or female type according to whether they had more or less than 50 per cent of their gonadal tissue made of male reproductive cells. The proportions of these male and female types among sexually differentiated Limacina collected each month are shown in Fig. 6. In this figure the REPRODUCTION OF LIMACINA RETROVERSA 291 material before May refers to " population A " and that after May to the mixed populations, with " population B " as the predominating con- stituent. It will be seen that in December and January nearly all the sexually differentiated Limacina were predominantly male. After these months the proportion of predominantly male individuals became re- duced : in April they constituted only 40 per cent, while in May the two sexual types were equally numerous. After May the proportion of predominantly male animals became more numerous again among the mixed populations. 100 100 FIG. 6. Proportions of hermaphroditic males and hermaphroditic females among sexually differentiated Limacina rctrovcrsa in each month. Unshaded area represents hermaphroditic females, shaded areas hermaphroditic males. Area shaded with continuous lines refers to " population A," that with broken line, to " population B." J1/? refers to the ratio of sex predominance in each month. When the change in the percentage of the two sexual types among sexually differentiated Limacina from month to month as shown in Fig. 6 is calculated into ratios of sex predominance (the number of pre- dominantly male individuals over those predominantly feminine) a change from a high value to unity is seen. These values are appended at the bottom of Fig. 6. It will be seen that in December and January this value is very high, but in May it is equal to one, while among the mixed populations the value is again in favor of the male type in the later part of the year. 292 SIDNEY C. T. HSIAO In Fig. 7 are shown the proportions of " pure males," animals with little or no feminine tissue in the gonad, and " less than 25 per cent " male (very pronounced feminine) types among the sexually differenti- ated individuals in each month. About 45-55 per cent of sexually differ- entiated Limacina in December and January were " pure males," the rest being hermaphroditic males. In March the proportion of l< pure males " was reduced greatly and from April onward this type of individual was not seen any more. On the other hand, pronounced feminine indi- viduals with less than 25 per cent of the gonadal tissue made of male germ cells were not seen before May. It will also be noticed that loo-, 7 fo 80 H 60- 40 20- DEC JAN MAR APR MAY JUN SEP FIG. 7. Monthly appearance of " pure males " and hermaphroditic females with less than 25 per cent of their gonadal tissue consisting of male germ cells. Histograms before April refer to " population A " and those after April to " popu- lation B" with mixed populations. '' Pure males" represented by unshaded areas; " less than 25 per cent male " type of hermaphroditic females represented by shaded areas. the occurrence of this type of extreme female in the later part of the year has been always low in comparison to the other types. Change in Reproductive Activity The reproductive activities of Limacina rctrovcrsa show a parallelism with the morphological changes associated with the advance of time. During December and January when " pure " and hermaphroditic males constituted nearly all the groups of sexually differentiated individuals no hermaphroditic females with mature ova in the gonad were seen in the specimens examined, while spermatozoa, on the other hand, were evidenced as being liberated by the fact that mature male germ cells were REPRODUCTION OF LIMACINA RETROVERSA 293 found in different parts of the gonoduct. It is in March, however, that individuals with mature ova in their ovotestes began to appear. This type of individual continued to increase in proportion until May. In May animals had been collected and fixed during their spawning activity. Later in the year all the different types of functional individual were seen. The occurrence of these types of functional individual in the various months is shown in Table IV. TABLE IV Distribution of functional types of Limacina retroversa, expressed in percentages, in different months. 1 2 3 4 5 6 Month Number studied Active males Mature females Spawning females Functional hermaphrodites Dec 24 per cent 63 per cent o per cent o per cent o Jan 26 58 o o o Mar 22 55 5 o o Apr 25 32 20 o 8 May 23 17 43 30 26 June 21 19 38 19 19 Sept 26 19 27 12 35 It will be seen from column 3 of this table that individuals ready to liberate spermatozoa are seen throughout the year, but that the proportion is higher in the earlier part of the year than in the later and that the reduction of the number of active males is accompanied by an increase in the number of individuals functioning as females (columns 4 and 5) and those acting as hermaphrodites (column 6) after April. A com- parison of columns 4 and 5 will show that spawning is found after the mature females were observed in March. It has been found in the section on the relation between size and sexuality that about half of all the mature females examined were fixed during the act of spawning. From column 5 of Table IV it is seen that in May about three-quarters of the mature females were fixed in this stage, while in June the pro- portion of spawning females amounted to half of the matured ones, and in September, less than half. This indicates a higher spawning activity in May than in the other months. The fact that no spawning individuals were seen among the mature females in April does not exclude the pos- sibility that egg-spawning took place in this month but that the indi- viduals did not happen to be fixed with eggs in the gonoduct and selected for examination in this study. We may consider spawning to start be- 294 SIDNEY C. T. HSIAO tween April and May and continue to the fall with May as the most important period for breeding. From the above examination of the reproductive system of Limacina in terms of time it appears that among " population A," which entered the Gulf of Maine first, all the individuals wrere sexually differentiated by the beginning of March. In winter, a great majority, if not all, of them were " pure males." The number of such " pure males " decreased in March and they were not seen in April, giving place to individuals with a lesser degree of masculinity, to functional hermaphrodites and eventually to hermaphroditic females with very little masculine tissue in their gonads. The proportion of male tissue in comparison with the female in the gonad decreased as the animals grew. By April 40 per cent of the specimens collected had the feminine tissue so well developed as to warrant their being classified as hermaphroditic females. Animals with " less than 25 per cent male " gonads occurred after the " pure males " had disappeared from the collections. In May the two sexual types were equal in number. After May the bulk of the species in the Gulf was derived from the recently arrived " population B " which mixed with the descendants and remnants of " population A " and other new recruits. Among the mixed population sexually undifferentiated indi- viduals were seen again and all the structural types were present. The functional phases of reproduction are summarized in Fig. 8. More than half of the protandric males started to function in winter and the male phase continued into spring and summer as it became gradu- ally replaced by the female and functional hermaphroditic phases of re- production. Females began to attain maturation in March and the pro- portion of mature females increased rapidly in April and May. Spawn- ing of ova commenced between April and May, being highest in the latter month. The largest proportion of functional individuals, includ- ing the male, female, and hermaphroditic types, is seen in May when half of them were spawners of eggs and the other half were males and hermaphrodites in function. These individuals were seen in the months from June to September also. It appears that May is the most impor- tant period for spawning though the reproductive activity was carried on into the autumn. Change in Sex Predominance A change in the value of the ratio of the sexes during growth is generally used as one of the indications of a change of sex. As the term " sex ratio " is not very appropriate for monoecious forms where both male and female reproductive tissues are present in the gonad, REPRODUCTION OF LIMACINA RETROVERSA 295 it seems better to speak of sex predominance and ratio between male and female predominance. The decrease in the value of this ratio from small to large-sized animals may be due, in general, to any one of the following causes : (a) Sexual dimorphism such that the males are habitu- ally smaller than the females, (fr) differential mortality, that is, more males than females die as the animals grow, and (c) sex reversal or an actual change in the condition of the reproductive system as the animals grow. There is no evidence from our material either to show that there 100 80- OEC JAN MAR APR MAY JUN SEP I 1 = NO DEFINITE SIGN OF FUNCTION 222 = FUNCTIONAL MALE |B * MATURE FEMALE E23 - FUNCTIONAL HERMAPHRODITE FIG. 8. Monthly distribution of the three functional types of Limacina retroversa. is a sexual dimorphism or to indicate a differential mortality which may be interpreted as the cause of the appearance of a higher proportion of males among smaller-sized Limacina and a greater number of females among larger ones. The third alternative, that a change in the pre- dominant sexuality takes place as the animals grow, however, has much evidence in its favor. In the first place, it has been shown above that there is a parallelism between the changes in the ratio of sex predomi- nance with reference to size and that with reference to the advance of the season : male types dominate among smaller-sized Limacina and dur- 296 SIDNEY C. T. HSIAO ing the winter months, and this excess of males decreases as the animals become larger and also as the season advances, and eventually the two types become equal among still larger-sized specimens and at the height of the breeding season in May. This parallelism indicates a change in the sexuality, both in structure and function, of the individuals as they grow. In the second place, it has already been shown in a previous paper that there is no clear line of demarcation between male and female among Liinacina, but that hermaphroditic male and female merge into each other and that a continuous series can be made out. As the size of the animals increases, a greater and greater portion of the gonad appears to contain female tissue. :< Pure males " are confined to small-sized animals and in the winter months. The continuous change in the proportion of male and female reproductive tissues in the gonad culminating in many functional hermaphrodites and females with less than 25 per cent of the tissue of their sex glands made of masculine elements argues in favor of a reversal of sex predominance in this species. Furthermore, there is no " pure female " among Limacina: all the hermaphroditic females are large and apparently derived from male types. Thirdly, a comparison of the size distribution of the different sexual types in the various months also shows that there is a change in sex predominance. In Fig. 9 the proportions of (a) sexually undiffer- tiated young, (b) individuals which are predominantly masculine and (c) those which are predominantly feminine, have been indicated in the histograms prepared by Redfield showing the distribution of size among all the specimens taken during each cruise. It shows the earlier occurrence of protandric males and later appearance, in March and April, of female types. The appearance of the female type among larger size groups suggests that they are transformed from the pro- tandric individuals rather than from the very small undifferentiated ones. From these morphological male and female types of Limacma func- tional hermaphrodites have also been indicated in the same figure (Fig. 9). These functional hermaphrodites increase from a very small pro- portion in April to a maximum in May and then decline in June. This change in the relative proportion of this type of individual supports the idea mentioned above that protandric males develop ovarian tissue while they are still functioning as males and eventually become functional females. Finally, if the modal class of each month's collection is taken as representative of the total catch, the phenomenon of change of sex predominance can be brought out when the different modal size groups are compared. Analysis of the size frequency distribution showed that REPRODUCTION OF LIMACINA RETROVERSA 297 MAY .3 .6 .9 [2 15 1.8 2.1 24 M A .9 1.2 1.5 IB 2.1 2 4 MM. .3 .6 .9 1.2 15 IB 21 24 MM. -- SEXUALLY UNOIF TED. - A — • = "POPULATION A" B = "POPULATION B" .3 .6 .9 1.2 1.5 (.8 2.1 24 MM. FIG. 9. Distribution of different types of Litnacina rctrovcrsa among various size groups in each month. The proportion of each sexual type is added to the histogram showing size distribution prepared by Redfield (1939). 298 SIDNEY C. T. HSIAO there is a single mode from December to April (Redfield, 1939). This modal class shifts from smaller to larger size during this period, showing a general growth of the members of the population. The sexual con- dition of the gonads of the modal class shows that not only growth, but sexual change has also taken place. In the modal class of the December material all the differentiated individuals are males and over 60 per cent of them have more than three-quarters of their gonadal tissues com- posed of male reproductive cells. Half of the males are ready to lib- erate spermatozoa. In January all the individuals of the modal class are predominantly male and 65 per cent of them are ready to discharge the product of their sex glands, thus showing an increase in the pro- portion of sexual differentiation with growth and an increase in the number of individuals capable of inseminating others. But it is in March that 11 per cent of the modal class are seen to have more female tissue in their gonads ; the rest are still all predominantly male, half of which are ready to discharge or are discharging spermatozoa. The proportion of females increases to one-half of the modal class in April and some of the individuals begin to function both as male and female simultaneously. Of the two modal classes in May, the larger one shows, in the condition of the reproductive system, a continuation of the general changes of the previous months. For there is a further reduction of the male type and an increase of the functional hermaphro- dites and actively functioning females. In this month, when the egg- laying individuals are added together, that is, spawning females plus functional hermaphrodites, they make up the whole of the modal class (see the right half of the histogram for May in Fig. 9). Thus if the modal class found by Redfield is taken as representative of the material collected, observations on the sexual condition of these groups give a consistent story of the change of sex predominance and function. RELATION BETWEEN SEXUAL PHASES AND DISTRIBUTION When the sexual condition of populations A and B are compared several interesting differences are seen. The first characteristic of the sexual phases of " population B " in contrast to " population A " is that sexual differentiation seems to commence earlier with reference to size. In " population A " individuals as large as 0.85 mm. in diameter have been seen as undifferentiated sexually, while the largest undifferen- tiated individual of "' population B " is only 0.6 mm. in diameter. In addition to the earlier differentiation of the gonads, a second difference is seen in the manner of differentiation between the two populations. Instead of the male elements in the gonads always taking the lead in REPRODUCTION OF LIMACINA RETROVERSA 299 development so as to bring about protanclry, some individuals among " population B " have both types of sexual cells developed at nearly the same time in addition to protandry, so that proterogyny and hermaphro- ditism appear. In view of this type of development, small-sized mature hermaphroditic females and functional hermaphrodites may also be interpreted as derived directly from young undifferentiated individuals with the protandric stage omitted. In " population A " such a situation is not seen. The third point of difference is that the development of the female elements of the gonads of individuals in " population B " is also earlier. This can be seen when the minima of the size of the individuals which achieved maturation of ova in the two populations are compared. The minimal size of Liinacina with mature ova is 0.8 mm. in diameter in April and 1 mm. in May among members of " population B," while among ;' population A " the minimal size of mature females is 1.3 mm. in April and 1.67 mm. in May. It is possible that these small mature females in " population B " did not pass through a protandric stage of noticeable length, though other specimens of the same population have a distinct young male phase preceding the female. In " population A " the smallest functional hermaphrodite is 1.5 mm. in diameter in April and 1.6 mm. in May, while during the same months functional hermaphrodites as small as 1.1 mm. in diameter are found among " population B." In Fig. 9 the right portion of the May histogram which has been attributed to " population A " by Redfield is very similar to that of April in the distribution of sexual types. In both cases hermaphroditic females occupy the large size groups with functional hermaphrodites between them and the small-sized, predominantly male individuals. On the other hand, the left portion which represents " population B," has a size distribution comparable to those of the winter polygons but the sexual phases are quite different from them. Here the hermaphroditic females and functional hermaphrodites appear among protandric males in small size groups, which in the previous months only contain pro- tandric and " pure " males. All these differences in the sexual condition of the two populations seem to point convergingly to one underlying factor, namely, that dur- ing the warmer part of the year sexual differentiation takes place sooner than in the colder time and the sexual phases may be telescoped together when differentiation is speeded up in the summer. This is in line with the previous observations made on Ostrea by authors on both sides of the Atlantic (Coe, 1932, 1936; Needier, 1932; Orton, 1926; Sparck, 1925). However, this does not exclude the possibility that there is a 300 SIDNEY C. T. HSIAO difference in the origin of " populations A and B " which may account for some of the observed differences in sexual development. COMPARISON OF THE REPRODUCTION OF LIMACINA RETROVERSA WITH OTHER MOLLUSKS As the reproductive history of no other Opisthobranchiata has been described, it is not possible to compare the reproduction of Limacina retroversa* with that of the members of the same genus or family or even order. Only in a general way can the sexual sequence of this form be compared with that found among other mollusks. In the winter months Limacina which entered the Gulf of Maine are of small size and sexually undifferentiated. As they grow this initial phase is followed by protandry in which nearly all the individuals are predominantly male. The protandric phase is succeeded by a her- maphroditic phase during which the animals transform from a more masculine to a more feminine condition. In the final phase the indi- viduals are predominantly female in structure and generally function as hermaphrodites, liberating both ova and spermatozoa, or as females only. This type of sexual transformation conforms in a general way with that found among gastropods. The sequence approximates most closely to that of the prosobranchiate gastropods such as Valvata tricarinata worked out by Furrow (1935) and Patella vulgata by Orton (1928). But the time required for the transformation of sex is different. Orton ob- served that the male Patella transformed into female at the age of one year, while Furrow found that in Valvata the different sexually func- tional phases followed each other very closely and the protandric phase was very short. However, in Limacina retroversa the change in sex predominance seems to take place during a period of two or three months. For instance, in one population which entered the Gulf of Maine in December and January, the protandric males were transformed into females from March to May. In the warmer part of the year this transformation required a shorter time for its completion. SUMMARY 1. In the winter months sexual differentiation commences when the animals are between 0.6 and 0.85 mm. in diameter. In the summer sexual differentiation starts when the animals are smaller. 2. The early proliferation of the male reproductive cells leads to a protandric phase. The " pure males " give place to hermaphroditic males and later to hermaphroditic females as the size of the animals becomes greater than 1 mm. in diameter. After the specimens are 1.5 REPRODUCTION OF LIMACINA RETROVERSA 301 mm. in diameter the female reproductive cells overtake the male in development, bringing about a decrease in the proportion of hermaphro- ditic males and an increase of the opposite sexual type. This change continues until the two types are equal, when the animals are 2 mm. or more in diameter. 3. Functionally the male phase starts among smaller Limacina than do the functional female and hermaphroditic phases. The highest pro- portion of actively functioning males is found among animals 1.2 to 1.8 mm. in diameter. Among larger animals the functional hermaphro- ditic and functional female phases replace the male phase. Hermaphro- ditic females begin the functional phase after they are greater than 1 mm. in diameter. 4. Members of a population which entered the Gulf of Maine in the winter were all sexually differentiated by the beginning of March. Nearly all of these individuals developed into protandric males, about half of which were " pure males." The number of such " pure males " decreased in March and disappeared in April, being replaced by animals with greater predominance of feminine tissue in the gonads. By April about 40 per cent of the specimens were hermaphroditic females and in May the two sexual types were equal. 5. The male functional phase began in winter and continued into spring and summer. Hermaphroditic females matured in March and their proportion increased in April and May. Spawning of ova com- menced by the end of April. The largest proportion of functional individuals of all types was seen in May. 6. Reasons are advanced for the belief that the change in sexual constitution of the populations is due to a reversal of sex predominance of the individuals. 7. The sequence of the sexual phases of Limacina retroversa agrees in a general way with those found for other members of the gastropod group of mollusks, particularly that of Valvata. This study has been carried out under the helpful direction of Professor Redfield to whom the writer wishes to express his indebtedness. BIBLIOGRAPHY AMEMIYA, I., 1929. On the sex change in the Japanese common oyster, Ostrea gigas Thunberg. Proc. Imper. Acad. Tokyo, 5 : 284. AWAITI, P. R., AND H. S. RAI, 1931. Ostrea cuc-.illata (Bombay oyster). Indian Zool. Mem., 3:1. BOUCHON-BRANDELY, M., 1882. On the sexuality of the common oyster (O. edulis) and that of the Portuguese oyster (O. angulata). Artificial fecundation of the Portuguese oyster. Bull. U. S. Fish. Com., 2 : 339. 302 SIDNEY C. T. HSIAO BURKENROAD, M. D., 1931. Sex in the Louisiana oyster, Ostrea virginica. Science, 74: 71. COE, W. R., 1931. Sexual rhythm in the California oyster (Ostrea lurida). Sci- ence, 74: 247. COE, W. R., 1932a. Sexual phases in the American oyster (Ostrea virginica). Biol Bull, 63 : 419. COE, W. R., 19325. Development of the gonads and the sequence of the sexual phases in the California oyster (Ostrea lurida). Bull. Scripps. Inst. Ocean. Tech. Ser., 3: 119. COE, W. R., 1933. Sexual phases in Teredo. Biol. Bull., 65: 283. COE, W. R., 1934a. Alteration of sexuality in oysters. Am. Nat., 68: 236. COE, W. R., 19345. Sexual rhythm in the pelecypod mollusk Teredo. Science, 80: 192. COE, W. R., 1935a. Influence of external stimuli in changing the sexual phase in mollusks of the genus Crepidula. Anat. Rec., Suppl., 64 : 80. COE, W. R., 19355. Sequence of sexual phases in Teredo, Ostrea and Crepidula. Anat. Rec., 64: 81 (suppl.). COE, W. R., 1936a. Environment and sex in the oviparous oyster Ostrea virginica. Biol. Bull., 71 : 353. COE, W. R., 19365. Sequence of functional sexual phases in Teredo. Biol. Bull., 71 : 122. COE, W. R., 1936c. Sexual phases in Crepidula. Jour. Exper. Zool., 72 : 455. COE, W. R., 1936c?. Sex ratio and sex changes in mollusks. Manifestation P. Pelseneer Bruxelle, p. 69. FURROW, C. L., 1935. Development of the hermaphrodite genital organs of Val- vata tricarinata. Zeitschr. f. Zellforsch., 22 : 282. GEM MILL, J. F., 1896. On some cases of hermaphroditisrn in the limpet (Patella) with observations regarding the influence of nutrition on sex in the limpet. Anat. Anzeig., 12: 392. GEMMILL, J. F., 1900. Some negative evidence regarding the influence of nutrition on sex. Communicat. Millport. Mar. Biol. Stat., 1 : 32. GOULD, H. N., 1917. Studies on sex in the hermaphrodite mollusk Crepidula plana. II. Influence of environment on sex. Jour. Exper. Zool., 23: 225. III. Transference of the male producing stimulus through sea-water. Jour. Exper. Zool., 29: 113. GRAVE, B. H., AND J. SMITH, 1936. Sex inversion in Teredo navalis and its rela- tion to sex ratios. Biol. Bull., 70 : 332. GUTSELL, J. S., 1926. A hermaphroditic viviparous oyster on the Atlantic coast of North America. Science, 64 : 450. HSIAO, SIDNEY C. T., 1939. The reproductive system and spermatogenesis of Limacina (Spiratella) retroversa (Flem.). Biol. Bull., 76: 7. LOOSANOFF, V. L., 1937. Seasonal gonadal changes of adult clams, Venus mer- cenaria. Biol. Bull, 72: 406. NEEDLER, A. B., 1932«. Sex reversal in Ostrea virginica. Con. Can. Biol. and Fish., 7 : 283. NEEDLER, A. B., 19325. American Atlantic oysters change their sex. Prog. Kept. All Biol Sta. and Fish. Ex p. Sta., 5: 3 ORTON, J. H., 1909. On the occurrence of protandric hermaphroditisrn in the mol- lusc Crepidula f ornicata. Proc. Roy. Soc., Ser. B. 81 : 468. ORTON, J. H., 1919. Sex-phenomena in the common limpet (Patella vulgata). Nature, 104 : 373. ORTON, J. H., 1922. The phenomena and conditions of sex-change in the oyster (O. edulis) and Crepidula. Nature, 110: 212. ORTON, J. H., 1926. Observations and experiments on sex-change in the European oyster (O. edulis). Jour. Mar. Biol Ass'n, 14: 967. REPRODUCTION OF LIMACINA RETROVERSA 303 ORTON, J. H., 1928. Observations on Patella vulgata. Jour. Mar. Biol. Ass'n, 15 : 851. ORTON, J. H., 1931. Observations and experiments on sex-change in the European oyster (O. edulis). Jour. Mar. Biol. Ass'n, 17: 315. PELSENEER, P., 1894. Hermaphroditism in Mollusca. Quart. Jour. Micr. Soc., 37: 19. PELSENEER, P., 1926. La proportion relative de sexes chez les animaux et par- ticulierement chez les mollusques. Mem. Acad. Roy. Belg. Sci., 8: 11. REDFIELD, ALFRED C., 1939. The history of a population of Limacina retroversa during its drift across the Gulf of Maine. Biol. Bull., 76 : 26. SPARCK, R., 1925. Studies on the biology of the oyster. Kept. Dan. Biol. Sta., 30: 1. YOUNGE, C. M., 1926. Protandry in Teredo norvegica. Quart. Jour. Micr. Soc., 70: 391. The Standard for Microscope Glass Gold Seal Microscope Slides and Cover Glasses Crystal Clear • Non-Corrosive • Will Not Fog Microscopic work demands glass of unusual clarity. Gold Seal Sb'des and Cover Glasses are made from glass practically free from alkali. They attain a precise and uniform thinness of plane surface. Therefore, Gold Seal offers an unusual degree of crystal clarity. Further, Gold Seal is guaran- teed against corrosion, fogging or any imperfection. Specify Gold Seal Slides and Cover Glasses. MADE IN U. S. A. CLAY-ADAN3 C0 44 EAST 23.D STREET, NEW YORK/ A Perfect Illustration LANCASTER PRESS, Inc. LANCASTER, PA. Or the lack of it, may make or mar a scientific paper. * For 65 years we have specialized in THE EXPERIENCE we have making reproductions by the Helio- type process of the most delicate gained from printing some pencil and wash drawings and photo- graphs; and by the Heliochrome proc- sixty educational publica- ess, of paintings and drawings in • i f* i color. tions has fitted us to meet Ask the editor to whom you submit the standards of customers your next paper to secure our esti- mates for the reproduction of your who demand the best. illustrations. • The Heliotype Corporation We shall be happy to have workers at the MARINE BIOLOGICAL LABORATORY Est. 1872 172 Green St., Jamaica Plain, write for estimates on journals or Boston, Mass. monographs. Our prices are moderate. Vol. LXXVI, No. 3 June, 1939 THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY NOTES ON THE FAUNA ABOVE MUD BOTTOMS IN DEEP WATER IN THE GULF OF MAINE1 HENRY B. BIGELOW AND WILLIAM C. SCHROEDER (From the Woods Hole Oceanographic Institution, Woods Hole, Massachusetts) INTRODUCTION It has long been known that the waters just above bottom, in the mud-floored depressions of the western side of the Gulf of Maine at depths of 70-90 meters, are the home of an abundant population of the large edible shrimp, Pandalus borealis, accompanied by the smaller spe- cies, P. montagui, P. propinquus and Dichelopandalus leptoceros (Rath- bun, 1883, p. 142; 1884, p. 819). After the introduction of the otter trawl to the Gulf of Maine waters, small amounts of these shrimps were landed from time to time in New England ports, and experimental trawling by the General Sea Foods Corporation in 1927-1928 again revealed the presence of P. borealis and its companion species in com- mercial quantities in the deep bowl just north of Cape Ann (Birdseye, 1928). But exploitation failed to follow, although Hjort's explora- tions of 1897 and 1898 (Hjort and Ruud, 1938) had long before led to the development of an extensive fishery in the other side of the Atlantic, which has gradually expanded to all suitable localities around the coasts of Norway. Investigations by Hjort and Ruud (1938) and by Broch (1935) had also shown that P. borealis is but one member of a distinctive faunal community, some of the members of which live on bottom, others in the waters close above. Information from various sources had similarly shown that the Pandalus of the Gulf of Maine have as companions the same species of flat fishes, of rose fish (Sebastes}, and of rockling (Enchelyopus), as in Norway, as well as several gadoids (Merluccius, Urophycis), which while peculiar to American waters, are at least similar to gadoid members of the Norwegian shrimp community. Hjort and Ruud's successive explorations had further demonstrated that Pandalus occurs in greatest abundance in situations where the bottom waters are sufficiently quiescent for organic debris to settle to the bottom. 1 Contribution No. 210 from the Woods Hole Oceanographic Institution/^' AfcV /~OVi3 2HC ON w* r LI Mt V 306 HENRY B. BIGELOW AND W. C. SCHROEDER The presence of similar faunal assemblages living under conditions presumably the same in the two sides of the Atlantic opens interesting problems of environmental control. The Woods Hole Oceanographic Institution, therefore, welcomed the opportunity, in August, 1936, to make a further survey of the Gulf of Maine shrimp grounds with " Atlantis," under Professor Hjort's personal supervision, and with the assistance of Captain Hagbart H0ium, an experienced Norwegian shrimp trawler. The purpose of the investigation being primarily to explore the abundance of Pandalus and of associated fishes in a favorable environ- ment, most of the hauls were made in the bowl north of Cape Ann and west of Jeffrey's Ledge (Area A, Fig. 1), where shrimps were already known to be plentiful, and which resembles the Norwegian shrimp grounds in its fjord-like topography and considerable depth. For com- parative purposes, hauls were also made in the northeastern and south- western branches of the open basin of the Gulf (Areas B, C, Fig. 1). Corings of the bottom for chemical analyses were obtained at thirteen of the stations. The catches were so encouraging from the fisheries standpoint, that the exploration of the local shrimp stock was continued during the following two months by local fisheries interests — again under the super- vision of Professor Hjort and of Captain Holum. The present report on the regional distribution of the more prominent members of the Pandalus community is a sequel to Hjort and Ruud's (1938) preliminary report, and to Wai ford's (1936) discussion of the prospects of the shrimp fishery that appears now to be in the process of development. Methods The hauls (with two exceptions) were made with otter trawls of the sort used in the Norwegian shrimp fishery, brought by Professor Hjort from Norway for the purpose, the distinctive feature of which is that the footrope works about six inches above the bottom to avoid catching mud, with a sweep rope suspended below (for detailed de- scription, see Walford, 1936). The success of the cruise was largely due to Captain H0ium's skill in their use. One trial was also made with the footrope scraping the bottom. The mesh-size of the Nor- wegian nets was 1% inches (stretched) throughout. But four sizes of trawl were employed, their footropes respectively 35, 76, 82, and 117 feet in length, and the hauls varied in duration from 30 to 90 minutes, at an average speed of 1.5 knots. It has therefore been necessary to adjust the catches to a common standard, for comparative purposes, FAUNA ABOVE MUD BOTTOMS IN DEEP WATER 307 a 60-minute haul with 82-foot trawl being chosen as the procedure most often followed during the cruise. Unfortunately, this adjustment is only a rough approximation, because the precise outline assumed by an otter trawl depends on a variety of factors such as speed, angle of the AREA A STA. 49 - 56 FIG. 1. Chart of western part of the Gulf of Maine, showing locations of " Atlantis " stations 2644-2664, August 18-23, 1936. boards, and the type of bottom over which it is being drawn. The present calculation is based on the assumption that the breadths of sweep of trawls with footropes of different lengths are in the same ratios as semicircles of corresponding perimeters. 308 HENRY B. BIGELOW AND W. C. SCHROEDER The chemical procedure, for which our thanks are due to Dr. S. A. Waksman, consisted of analyses of the upper half-inch of each core, for carbon, by the chromic acid oxidation method, and for total nitrogen by a semi-micro-Kjeldahl method. These were carried out at the New Jersey Agricultural Experiment Station at New Brunswick, New Jersey. Measurements were also made at Woods Hole of percentage loss on ignition, but showed considerable irregularity. Since this loss does not TABLE I Catches of the more prominent species at each station, adjusted as above (p. 306). Numbers italicized are derived from approximate counts. Station Depth Amounts of Pandalus in liters Numbers of Sebastes Numbers of Mer- luccius Numbers of Hippo- glossoides Numbers of Glypto- cephalus Numbers of Uro- phycis tenuis Numbers of other fisht 2644 200 5 108 15 0 0 11 6 2645 183 2 16 8 0 2 3 4 2646 144 2 432 240 2 0 65 4 2647 228 1 3 1 0 0 7 11 2649 164 32 150 700 12 10 32 2 2650* 175 24 224 338 4 1 36 0 2651* 168 11 131 262 3 2 6 1 2652 160 71 168 196 13 14 39 1 2653f 155 168 107 20 47 24 28 34 2654 158 126 183 840 13 35 11 2 2655 168 100 150 225 13 3 40 2 2656 177 56 283 351 63 12 23 7 2657a 120 2 100 160 5 0 20 1 26576 120 3 103 67 17 8 40 2 2658a 162 16 400 467 9 1 67 4 26586 163 50 450 600 10 5 0 3 2659a 164 48 150 600 24 12 50 6 26596 164 27 188 700 13 9 70 3 2660 192 1 250 300 26 18 100 14 2662 198 1 270 100 0 2 0 5 2663 183 4 833 200 5 0 47 39 2664 215 1 67 47 0 0 12 7 * Net damaged, calculations assume that only half the catch was saved. t In this haul, the footrope scraped the bottom. | See Table II. represent the organic fraction alone, but also such water as is driven off from the clay fraction at high temperatures, these values are not included in the present discussion. Qualitative Composition of the Community The most interesting aspects of the catches from the qualitative standpoint are that the number of species taken in abundance at any given station was so small (Table I), and that the same few species FAUNA ABOVE MUD BOTTOMS IN DEEP WATER 309 dominated, in all three areas visited, as well as at intervening locations. Among the catches of shrimps, for example, only occasional specimens of other related species, or of Pasiphaea, were noted among the more plentiful P. borealis. In fact, the latter so predominates among the shrimp population in deep water in the Gulf that a sample taken at random from catches made at depths greater than 160 meters in the Jeffrey's Bowl, in the summer of 1938, consisted of 738 specimens of P. borealis, but only 9 of all other pandalids combined. In shoaler water, however, in that general vicinity, the relationship is the reverse, as off Casco Bay in 30-40 meters, and at a neighboring station in about 90 meters, where P. propinquus greatly predominated over P. borealis TABLE II List of "various" fishes other than those recorded in Table I, trawled on bottom in Jeffrey's Bowl and in the open basin of the Gulf of Maine, during August, 1936. Species Jeffrey's Bowl Open basin Myxine glutinosa X X Squalus acanthias X X Raid senta X Ram stabuliforis X X Raid scabrata X X Clupea hdrengus X X A rtediellus uncinatus X Cyclopterus lumpus X Lumpenus lampetraeformis X Cryptacanthodes maculatus X * Pollachius virens x Gddus callarias X x Melanogrammus aeglefinus x Urophycis chuss X X Urophycis chesteri X Enchelyopus cimbrius X X Brosme brosme X Macrourus bairdii X Lophius am-ericanus X (39 to 0 and 39 to 1 respectively) at that same time. This is in line with Rathbun's (1883, p. 143) statement that while P. montagui and Dichelopandalus Icptoceros are often associated with P. borealis over soft bottoms, they are not only the most plentiful in shoaler water, but occur on gravelly and sandy bottoms as well as on mud. Our catches of fishes, in August, 1936, were similarly dominated by a group of 5 species, in combination, namely: by rose fish (Scbastcs marinus), silver hake (Merluccius bilinearis) , hake (Urophycis tennis) and flat fishes (Glyptocephalus cynoglossus and Hippoglossoides plates- soides), and to such an extent that the total catch of about 11,900 fish 310 HENRY B. BIGELOW AND W. C. SCHROEDER included only 183 specimens of all the other species combined, that are included in Table II. The total list of 24 species together with a few others that have been picked up in deep trawlings previously (Argentina, Triglops, Cottunculus, Aspidophoroides, Liparis, Pholis, Leptoclinus, Zoarces, Lycenchelys) includes the great majority of the members of the Gulf of Maine fish fauna that are to be expected on the floor of the deep troughs, whether as regular inhabitants or stragglers. This is a very much less varied assemblage than inhabits the corresponding depth zone along the offshore slope of the Gulf of Maine and of southern New England (for complete list of records there, see Goode and Bean, 1896). Regional Distribution Regional segregation of the catches (Table III) shows that shrimps were taken in much greater average abundance in the deepest part of the Jeffrey's Bowl (Area A} and on the northern slope of its northern rim (Station 2658, 163 meters) than in either branch of the main basin of the Gulf (Areas B, C}. TABLE III Regional distribution of average catches. Area Hauls Shrimp Sebastes Mer- luccius Flat fish Uro- phycis Various fishes Total fishes Number Liters Numbers Numbers Numbers Numbers Numbers Numbers A 7 60 187 416 28 26 2 649 B 2 1 260 200 23 50 10 543 C 6 2 243 92 2 34 12 383 St. 2658 2 33 425 534 13 0 5 977 From the environmental standpoint, the outstanding feature of this Bowl is the fjord-like character given to its deepest parts (170-192 meters), by the shallow ridge of Jeffrey's Ledge to the south and off- shore, and on the north, by a sill with a maximum depth of about 134— 137 meters (Fig. 2). This enclosure appears to be more effective, so far as concerns the activity of circulation of the bottom waters, than the topography of the sill might suggest, for temperature averages some 1-2° colder in midsummer in the bottom of the trough than at corre- sponding depths in the open basin to the east of the barrier, only a few miles distant. Hjort and Ruud (1938, p. 103) have already pointed out that the mud was richest in carbon and in nitrogen where large catches of shrimps were made. This is corroborated by the following comparisons of the catch of shrimps with the percentages of organic carbon and of nitrogen FAUNA ABOVE MUD BOTTOMS IN DEEP WATER 311 in the corings (Table IV), the average being 4.02 per cent of organic matter for the group of stations where the catch of shrimps was more than 30 liters, but only 3.05 per cent for stations yielding 5 liters of shrimps, or less. 70 W -43 70 W FIG. 2. Submarine topography of the region between Jeffrey's Ledge and the western coast-line of the Gulf of Maine at depths greater than 100 meters. At one station (2662), it is true, where the percentage of organic matter was high, the catch of shrimps was very small ; regional irregu- larities are, in fact, to be expected, in the nature of the case. But in no instance was the mud poor in organic content where shrimps were 312 HENRY B. BIGELOW AND W. C. SCHROEDER abundant. While the correlation is not so close between the catches of shrimps and the percentage of nitrogen, the latter averaged 0.338 per cent for the rich and 0.234 per cent for the poor shrimp stations. In other words, the average percentage of organic carbon was about one- third higher, and that of nitrogen about one-half higher where catches of shrimp were relatively large than where they were small. Unfortu- nately, the data do not afford so satisfactory a comparison from the regional standpoint, because the mud was sampled at only two of the stations in the deeper part of the Jeffrey's Bowl. At each of these (Stations 2649, 2654), however, the percentage of organic carbon closely approached the maximum found at any station, while the percentages of nitrogen (0.295 and 0.309 per cent) were at least somewhat above the TABLE IV Percentage of organic matter and of nitrogen in dry mud and catch of Pandalus. Station Shrimps Organic matter* Nitrogen liters per cent per cent 2654 126 4.1 0.309 2659 37 av. 3.7 0.195 2658 33 av. 4.2 0.553 2649 32 4.1 0.295 2644 5 2.6 0.311 2663 4 2.6 0.115 2657 2.5 av. 2.4 0.201 2645 2 3.7 0.217 2647 1 3.1 0.205 2660 1 3.3 0.301 2662 1 4.0 0.310f 2664 1 2.8 0.217 * Carbon X 1.724. f Determination made at Woods Hole. average for the series as a whole (0.262 per cent). It is interesting that the station that was the richest of all, both in organic carbon and nitrogen (Station 2658) was at least moderately productive of shrimps, though located north of the sill. Thus, it seems sufficiently established, both for Norwegian waters and for the Gulf of Maine, that in summer Pandalus is abundant only in situations where organic matter is plentiful on the sea floor below ; and by available evidence this applies to the males throughout the year. But the fact that considerable catches of ovigerous individuals were landed by fishermen from depths of 70-90 meters only, in the northwestern side of the Gulf, during the winter of 1938 suggests that the large females tend to work inshore into shoaler and more turbulent situations during the egg-bearing period. FAUNA ABOVE MUD BOTTOMS IN DEEP WATER 313 It seems that we are dealing here with a food relationship, for while no direct observations are available as to the diet of Pandalus in the Gulf of Maine, Wollebaeck (1903, see also Hjort and Ruud, 1938, p. 60) found the stomachs of this shrimp in Norwegian waters to contain animal remains of various origins, among a mass of indeterminate mud, evidence that while it feeds on bottom, its diet consists of the remnants of dead organisms or perhaps of bacteria and protozoa, rather than of the mud itself. This is in line with Waksman's (1933) demonstration that the organic matter in the mud is largely in the form of humus, and hence not likely to be readily digestible by the higher animals. The catches of Merlucdus also averaged considerably larger in Area A and at the neighboring station (2658) than in either of the other areas, though there is no reason to suppose that the type of bottom is of direct influence in this case, but rather that the significant parallel- ism is between the number of fish and of shrimps, as discussed below. Sebastes, on the other hand, averaged somewhat less plentiful in the Jeffrey's Bowl (Area A) than over the rim of the latter (Station 2658) or in the open basin (Areas B and C}. But the difference in average catch from area to area was so small for pleuronectids and for Urophyds that it may not have been regionally significant. Among the fishes of minor numerical importance, Enchelyopus cimbrius and Urophyds chuss were encountered at so many stations (respectively 13 and 12 hauls out of 22), that they appear practically universal over the deep bottoms both of the Jeffrey's Bowl and of the western branch of the open basin of the Gulf, though the maximum catch was only 9 specimens in the one case, 10 in the other. Enchelyopus was also taken in the eastern basin, and failure to take U. chuss there in our few hauls is no evidence that it does not actually occur there. Odd specimens of all other species included in Table II, and listed on page 309, have also been taken in the bottom of the Jeffrey's Bowl at depths greater than 160 meters on one occasion or another, with the exceptions of Triglops, Argentina, Aspidophoroides, Pholis, Leptoclinus, Brosme, Urophyds chesteri, Macrourus, Lophius, haddock and Ameri- can pollock. These are also to be expected there, either because plenti- ful in somewhat shoaler water in the near vicinity, or because already known to occur in deep water elsewhere in the Gulf. Arctic species in particular, such as Cottunculus, would be as likely to maintain a small population in the low temperatures of the Bowl as anywhere else to the west of Cape Sable. In short, there is nothing in the available data to suggest that the sill responsible for the fjord-like character of this Bowl gives any distinc- tive character to the fish fauna of its bottom waters, as contrasted with 314 HENRY B. BIGELOW AND W. C. SCHROEDER that of the open basin of the Gulf to the eastward. At most, certain species that have their center of abundance over the slope outside the Gulf (Macrourus, Urophycis chesteri, Argentina} may be progressively less numerous westward, until they reach the trough only as stragglers, if at all ; whereas certain others may be most plentiful there, whether be- cause of cold water affinity, or to utilize the particular food supply pro- vided by the local abundance of Pandalus, as appears to be the case with Merluccius. Urophycis tenuis has long been known to feed chiefly on shrimp (see Bigelow and Welsh, 1925, p. 450, for summary of its feeding habits) ; in fact, all but one of the specimens of this species that were opened during our cruise contained more or less remnants of large Pandalus. But the following tabulation (Table V) fails to suggest any TABLE V Numerical relationship of fish to shrimps. Shrimps, liters 126 100 71 56 54 48 32 27 24 16 11 5 U. tenuis, numbers 11 30 40 76 0 50 34 75 36 67 6 23 Sebastes, numbers 183 ISO 168 783 450 150 150 188 ?.?4 400 131 108 Flat fish, 2 sp., numbers . . 48 16 25 75 15 36 22 22 6 10 5 0 Merluccius, numbers 840 225 196 351 600 600 700 700 338 467 262 15 Shrimps, liters 4 3 2 2 2 i i l 1 U. tenuis, numbers 60 40 4 86 20 14 100 0 1Q Sebastes numbers 833 103 16 432 100 3 ?so ?70 67 Flat fish, 2 sp., numbers . . 5 25 2 2 5 44 2 2 0 Merluccius, numbers 200 67 8 240 160 1 300 100 47 correlation between the numbers of Urophycis and the amounts of shrimp taken, for the average catch of the former averaged about the same for the nine stations which yielded 24 liters or more of shrimps, as for the ten stations that yielded 5 liters or less (34 and 37 Urophycis, respectively), while the largest and the smallest catches of Urophycis were each made at a station where shrimps were very scarce. The catches of Sebastes seem also to have been independent of those of shrimps, the average being almost exactly the same for the group of stations where shrimps were plentiful (20 liters or more), as for the FAUNA ABOVE MUD BOTTOMS IN DEEP WATER 315 group where they were scarce (216 and 218 Sebastes, respectively) ; while the largest and smallest catches of Sebastes (833 and 3) were both made at stations yielding only 4 liters and only 1 liter of shrimps, respectively. Neither do the present data yield any information as to the extent to which Pandalus, when abundant, serves as prey for Sebastes, the great majority of the latter having regurgitated the stomach contents before they were brought on board. A rough parallelism does, however, appear, between the abundance of Merluccius and of shrimps, illustrated by the facts that the catch of the former averaged four times as great (494) for the group of stations 10 8 5 10 15 20 25 30 35 LENGTH IN CM. FIG. 3. Length-frequency distribution of 270 specimens of Sebastes, Station 2662, smoothed once by running average of three. The broken curve marks the lower limit for adequate sampling by the nets employed. where shrimps were abundant as where they were scarce (114) ; that it was only where the catch of shrimps was less than 6 liters that the catch of Merluccius was as small as 100; and that no hauls where shrimps were so scarce yielded as many Merluccius as the average for the series as a whole (306). To judge from the predaceous nature of this fish and from its independence of the bottom, we are dealing here with a predator-prey relationship, though our present data fail to yield direct information in this regard for Merluccius any more than for Sebastes, and for the same reason. Flat fishes of the two species, combined, also averaged considerably 316 HENRY B. BIGELOW AND W. C. SCHROEDER more plentiful (av. 30 flat fish) where shrimps were in more than moderate abundance than where the catches of shrimps were 5 liters or less (av. 9 flat fish). But there being no reason to suppose that either of these pleuronectids preys to any extent on shrimps, the parallelism in this case suggests that the same muddy bottoms which provide the rich- est feeding for Pandalus are also rich in the small invertebrates on which Hippoglossoides and Glyptocephalus chiefly subsist. While we have no precise data, in this respect, for the Gulf of Maine, such is cer- tainly true in Norwegian waters, where soft bottoms in similar situations 10 8 x o Se UJ O J 'C :*V« '^"u THE EFFECTS OF CENTRIFUGING ON THE POLAR SPINDLES OF THE EGG OF CHAETOPTERUS AND CUMINGIA T. H. MORGAN (From the William G. Kcrckhoff Laboratories of the Biological Sciences, California Institute of Technology, Pasadena, California) The factors involved in the movement of the maturation spindle to the pole of the egg are obscure and have been little commented on. In many eggs it has been shown that the mitotic rays may extend through almost the entire polar hemisphere. In such eggs the axis of the spindle is described as at first horizontal, or possibly somewhat oblique. It then turns with one end directed to the pole, but whether there is one particular end that takes this position or whether either end may so orient is unknown. It is conceivable, if there is a predetermined end, that it may trace back to the last oogonial division, or if not, that a cytoplasmic difference may determine the direction of its turning. At present there is no evidence for either of these assumptions. In some respects the eggs of Chaetopterus offer excellent material for the study of the conditions that lead to the movement of the meiotic spindle to the pole. The eggs can be procured in any quantity, and the same individual will supply material for two or more weeks. When re- moved from the parapodia the germinal vesicle is present in the eggs. It breaks down within a few minutes and the polar spindle forms, which moves within ten minutes to the pole of the egg where it remains until the egg is fertilized. The first polar body is given off 20 to 22 minutes after the egg is fertilized, and the second polar body about ten minutes later. Since all the eggs are in the same stage an opportunity is given to study the effects of centrifuging on the spindle at any stage desired. If the egg is centrifuged at the time when the wall of the germinal vesicle is disappearing, and centrifuging is continued during the time when the polar spindle is being formed deep in the egg, the spindle may be prevented from passing to the pole or to any other point on the sur- face. Whether it does or not may depend in part on the centrifugal force employed; in part on the orientation of the egg during centrifug- ing ; in part on the destruction of the peripheral rays of the mitotic fig- ure ; and in part on the redistribution of the contents of the egg. If the egg is centrifuged after the polar spindle has reached the pole 339 340 T. H. MORGAN its position there or its change of position will depend on the orientation of the egg with respect to the centrifugal force. If the egg is cen- trifuged during the time of extrusion of the polar body the result may again depend on its position with respect to the centrifugal force and, theoretically, at least, on whether the spindle is in anaphase or telophase or whether the second polar spindle is in process of formation. Again, if the egg is centrifuged after the first polar body has been given off, the second spindle may remain in place or be driven into the interior. If the latter occurs the question arises whether it will return to the surface and give off the second polar body, or remain in the in- terior of the egg. These alternatives may depend on the location of the pole and also on whether the egg is elongated by the centrifugal force, or remains spherical or nearly so. The first problem then is to find out whether the eggs do or do not orient on the machine with the polar hemisphere turned toward the center of rotation. In Chaetopterus the orientation of the eggs on the machine has been discussed by Whitaker and Morgan (1930), Wilson (1930), and Morgan (19376, 1938). It has been found that there is a tendency for most of the eggs to orient more or less with the polar hemisphere toward the center of rotation, but in a considerable number of cases the pole may make any angle with the centrifugal axis up to as much as ninety degrees or even more. I have suggested (1938) that the failure of many eggs to orient perfectly may be due to the contents of the eggs beginning to respond to the centrifugal force more quickly than the egg as a whole. The best evidence for this is found when the eggs are centrifuged after the first polar body has been extruded. Under these circumstances the polar body frequently does not lie at the centripetal pole, but to one side of it. In such cases there can be no question of a change in the position of the true pole. If the eggs are strongly centrifuged before the spindle has reached the pole there is the possibility that the presence of the oil cap may interfere with the ex- trusion of the polar body in the oil field. This possibility has also been examined. If the centrifugal force is not too strong the oil field ex- tends as a broad thin cap over the centripetal pole, and the spindle may pass through its center, or in most of the eggs at one side. In other eggs the spindle lies outside the cap. On the other hand, if the centri- fuging is strong, or over a longer time, then the oil cap is smaller in circumference and extends deeper into the egg. If, as seems to be the case, fewer polar bodies are given off in its center, this is in part due to the small arc it covers, but possibly the oil may then interfere with the passage of the spindle through it. In such cases the spindle may pass to the surface outside the oil cap, which may mean that it does not come EFFECTS OF CENTRIFUGING ON POLAR SPINDLES 341 off exactly at the true pole but near it. There is no serious objection to such an assumption, but it is difficult to prove because other evidence shows that the egg fails in many cases to orient with the pole exactly in the axis of rotation. The way in which the stratification of the materials of the egg takes place has been little discussed by those who have used the centrifuge to separate the materials of the egg. If it is assumed that at first the oil, the clear cytoplasm, the yolk and the pigment are equally distributed in the egg, they may be considered as passing by or through each other ac- cording to their specific weights. On the other hand, these differences may be of such a kind that one of them, the oil for instance, may move more quickly to one end than does the yolk or the pigment to the opposite end, while the clear protoplasm comes to occupy the middle zone. De- terminations of the so-called absolute viscosity of the egg, by measure- ments of the time it takes to bring about a visible stratification of these materials, are open to the criticism that some of the materials reach their positions sooner than do others. For example, the yolk and the pig- ment of the egg of Cumingia may take a longer time to be separated clearly than does the oil and its underlying clear zone. If the clear zone is the matrix from which the other materials are separated, then, I sup- pose, it is this matrix that is responsible for the viscosity, but I am not sure that this is the correct interpretation. However, in most cases reported, comparisons are made between eggs in different stages of maturation, and such comparisons may be more nearly valid. When the ripe ovarian eggs, or eggs just set free, are killed in Flemming and later sectioned and stained there is found a ring of black granules in the protoplasm around the nucleus. The nucleoplasm re- mains unstained in iron haematoxylin. After the germinal vesicle breaks down this region stains blue. If the ovarian eggs are killed in aceto-carmine, sectioned, and stained in iron haematoxylin fewer dark granules are seen. Whether there are two kinds of granules, or, in the latter case, fewer are stained is uncertain. If the egg is on the centrifuge when the germinal vesicle breaks down the larger granules move along the periphery towards the centrif- ugal end forming there a semicircle in section (Fig. 39). They are imbedded in a somewhat homogeneous layer that stains yellow in eosin which is thickest around the antipole outside the granules. In the cen- ter of the outer hemisphere there is a clear spherical area, sometimes having a flocculent appearance slightly stained by eosin. In it are a few dark granules. This clearer sphere does not appear to be due to failure of the preservatives to penetrate into it. Its origin is obscure. It looks as though it were the clear nucleoplasm, but that part of the egg is now 342 T. H. MORGAN occupied by a finely granular stained zone (as it is also in the uncentri- fuged egg). Over this clear sphere there is a crescent of very finely granular substance that stains very dark in iron haematoxylin (Fig. 40). The chromosomes are sometimes imbedded in this crescent or lie just outside it. The eggs of some other species of animals have a more protoplasmic layer around and beneath the pole, while the yolk lies mainly in the anti- polar hemisphere. If such eggs are centrifuged, and orient on the ma- chine, the oil present in the clear region or in the yolk must pass through the protoplasmic region while the yolk simply becomes more condensed in the opposite hemisphere, leaving the clear zone nearly in its original place. But if the egg is not so oriented at the time when the redistribu- tion begins there is an alternative possibility. The protoplasmic layer may move around one side of the egg, and the yolk around the opposite side. The polar spindle, if not attached to the surface, may then be carried along with the protoplasmic layer. When removed from the centrifuge, the stretched protoplasm may move partly back to the original polar field, and the polar body be given off at or near the pole. I have shown this to be the case in the egg of llyanassa, and there is some evidence that such a slipping around may take place in the centri- fuged egg of the frog, even in response to gravity in an inverted egg. In Crepidula a similar movement may take place although Conklin's figures do not give any evidence of such a rotation. In llyanassa the rotation of the interior of the egg was best demon- strated either by holding the egg in an inverted position in a small tube or by imbedding inverted eggs in gelatin (dissolved in sea water) that was solidified before placing on the machine (Morgan, 1935). In the latter experiment the question may arise whether the gelatin itself affects the result. The same question has arisen in the course of the present experiments when eggs were centrifuged on top of a gum arabic solu- tion, or in or above an " isotonic solution " of cane sugar or raffinose. In fact, these methods were abandoned after the first experiments be- cause it was found that some of the solutions not being quite isotonic caused the egg to shrink slightly and possibly inhibited, for a time, the movement of the spindle, or interfered with its movements on the centrifuge. Even when the eggs come to lie on top of a gelatin or gum arabic solution they lie in a gradient where the sea water mixes with the gum arabic, etc. This effect will be considered later. It suffices here to say that the stratification of the eggs can be as perfectly brought about when they lie on the gum as in sea water alone, but in most cases these eggs do not elongate or break apart as in sea water, and if kept too long in the solution the extrusion of the first or second polar body may be suppressed. EFFECTS OF CENTRIFUGING ON POLAR SPINDLES 343 Centrifuging the Egg during the Dissolution of the Nuclear Membrane By means of the centrifuge it is possible to prevent the polar spindle from moving to the pole; or later, after it has moved to the pole, to carry it back into the interior of the egg. Under these conditions there are two alternatives. The chromosomes may then separate, but whether the mitotic figure will be capable of dividing the egg into equal or un- equal cells, or whether the double or quadruple set of chromosomes re- mains in place and takes part later in the cleavage division can only be determined in each kind of egg by examining it. The other possibility is that during the first or the second polar division one end of the spindle, attached to the surface, may remain in place and the other end be carried deeper into the egg (as Conklin has described for the egg of Crepidula}, and bring about a division sometimes into two large cells. In Chaetop- terus, and also in Cumingia, I have not succeeded in doing this — the spindle and its attached chromosome either remaining at the surface after centrifuging, or moving as a whole deeper into the egg. Different eggs probably behave differently in this respect. In Ilyanassa the polar spindle in " inverted eggs " is carried along in the protoplasm at or near the surface of the egg while in Crepidula the spindle is driven through the center of the egg or at least it appears to do so judging by Conklin's drawings, although he does not specifically state this to be the case. This kind of result may depend to what extent the eggs orient on the machine; for, if some or many of them lie obliquely, when centrifuging begins it is to be expected that the lighter and heavier parts may slip around each other rather than interpenetrate. Both in Ilyanassa and in Crepidula a spermatozoon has entered the egg while in the oviduct, hence the polar spindle is activated. On the other hand, Chaetopterus does not give off its polar bodies unless the egg is fertilized. Such eggs must be fertilized in order to activate the mitotic division. The egg can be activated by the sperm either before it is centrifuged, when the conditions are more nearly comparable to the eggs of the other animals, or it can be first centrifuged and then ferti- lized. Both tests have been made. It is also possible to break apart the eggs on the centrifuge and obtain large or small tops. It might seem that the likelihood of division by means of the polar spindle would be thus facilitated, but this is found not to be the case. Even very small tops may give off small polar bodies. When many eggs are centrifuged in sea water they make a compact mass at the bottom. In squirting them loose, with a long large-mouthed pipette, some will, no doubt, be broken apart if they have become elongated and have a narrow neck, but it is surprising how many even 344 T. H. MORGAN of the latter remain intact. Unless the membrane is broken the eggs do not greatly elongate, but if the centrifugal force is strong enough the membrane breaks at the yolk region and the yolk squeezes out leaving the rest of the egg surrounded by the shriveled membrane. Sometimes more, sometimes less, of the egg escapes, and it may be pinched apart at almost any level. The oil, too, after heavy centrifuging, may pinch off, and either float freely or remain inside the membrane. In order to avoid breaking the eggs apart and still elongate them for an examination of the subsequent behavior of the polar spindle, they were sometimes centrifuged over a sugar solution heavier than sea water or over gum arabic. This method avoids packing the eggs together at the end of the tube as they do in sea water. The packing can be avoided by adding only enough eggs over the sugar or gum to form a single layer on the surface. It was found, however, that the sugar, and the gum also that was used, bring about some shrinkage of the egg and that the movements of the spindle may be retarded or even stopped. This method had, therefore, to be abandoned since in many of the ex- periments it was desired to move the spindle and not interfere with its activity, for, if it is held back, along with the rest of the egg, it will, when the eggs are returned to sea water (where they quickly become more or less spherical), proceed under most conditions either to come to the surface and produce polar bodies or else skip that stage if held as long as an hour on the centrifuge, hence defeating the purpose of the experiment. Formation of Polar Spindle Since the condition of the egg at the time of centrifuging had a bear- ing on the results, it became necessary to prepare a normal series of stages for comparison. When the egg of Chaetopterus is removed a large nucleus is present in the center of the egg that has an irregular contour. In the living egg its outlines disappear in a few minutes and in the course of 11 minutes a clear oblong area appears at the pole. Stained preparations show that the chromosomes have not yet reached the pole at the time when the clear area is first formed. They lie near the center in a cluster or clump for 9 minutes. Nine tetrad chromo- somes are present, one smaller than the others. At 12 minutes they lie in a spindle that is moving to the surface. Centrifuging the Eggs in Sea Water Several different experiments were made in which the time of cen- trifuging and the force applied to eggs in different stages of polar body formation were tested in order to find out whether the polar spindle, if EFFECTS OF CENTRIFUGING ON POLAR SPINDLES 345 driven into the egg, would cause the egg or an egg-fragment to divide by means of the polar spindle, or return to the surface. In some of the experiments the unfertilized eggs were centrifuged at the time when the chromosomes had just emerged from the germinal vesicle. The chromosomes are then near the center of the egg in the polar hemisphere and it might be supposed that they could be held there during the time that it takes for the spindle to form and move to the pole. This is in fact what sometimes happens. If then the eggs are removed from the centrifuge and fertilized an opportunity might seem to be given for the buried spindle to divide the egg, but in no case did this happen. The chromosomes either remained in the interior and took part in the later division of the egg, or, under some conditions, moved to the surface and gave off one (usually) or two polar bodies before the cleavage came on. The alternative, viz., to fertilize the eggs at the time when the polar spindle has formed and then centrifuge them, was also tried. Under these conditions it seemed possible that the activated spindle might di- vide the egg while on the centrifuge, or later, if taken off at the proper intervals ; but again the results were negative. The spindle moved to the surface and gave off one or two polar bodies, or remained in the egg and the chromosomes took part later in the cleavage. Several attempts were made to drive the polar spindle (after ferti- lization) into the interior at the time when it was about to divide to produce the first polar body but, as a rule, the spindle and its attached chromosomes either remained at or near the surface, or if driven into the interior, it failed to divide the egg at the time of polar body forma- tion. Since, in previous experiments with Ilyanassa, it had been found, though very rarely, that the second polar spindle may divide the egg (if it is elongated at that time), a number of experiments were made with this stage but without success. In nearly all cases stained preparations were made of the eggs in order to check up the records of the living eggs. These confirmed the observations on the living eggs, but added very little further informa- tion except with respect to the location of the chromosomes. It is obvious, then, if the eggs fall, to some extent, at random on the centrifuge, that the spindle (and chromosomes) would have to move across the diameter of the egg or through its center or along one side in all cases where the pole of the egg happened to be at the centrifugal pole, since the chromosomes come to lie in the protoplasmic zone of the centrifuged egg, or near it, or in the oil field. On the other hand, in the reverse situation the spindle (and chromosomes) would, if it moved at all, shift to the interior. In all intermediate positions the chromo- 346 T. H. MORGAN somes would already lie at or near the level toward which their specific gravity tends to locate them. It has been shown in previous papers that the egg of Chaetopterus on the centrifuge tends to orient on the machine with the true pole more or less turned towards the center of rotation. Hence, in the majority of cases, there would be little tendency to shift the level of the spindle, that had already reached the surface, to a new position. In fact, this corresponds with the observations on preserved eggs centrifuged at that time. Practically no spindles were found at the apex of the oil but not infrequently at its border. Also none were ever found in the yolk that occupies about half of the egg. If a spindle happened to be in what becomes the yolk hemisphere after centrifuging (in an egg completely inverted), it would be driven into the lighter half of the egg. Though it has not been possible to induce the polar spindles to act as a division center for the whole egg, or even for a part of an egg, the result is not due to a destructive action on the spindle, since it may after centrifuging move to the surface and give off one or two polar bodies. The failure is also not due to the egg not being elongated at the time when the spindle division is due, since very many cases of elongated eggs were observed. It may be said then that the failure of the polar spindle to divide the egg of Chaetopterus is due in part to the tenacity with which the spindle adheres to the surface after it has once reached the surface (except as explained above), or, in other cases, to the strong tendency for the spindle to come to the surface. But granted that these conditions occur in most of the cases, it still remains unexplained why the polar spindle, if held inside the egg during the time of its normal formation, is unable to divide the egg, or a fragment of the egg. In this connection, it should be recalled that the division of the egg of Crepidula or of Ilyanassa to form the so-called giant polar body is a relatively rare event. Details of Centrifuging (1) The eggs were fertilized 2 minutes after removal, and 5 minutes later were centrifuged at 2080 r.p.m. at 19% cm. from center for one-half hour. When removed (Figs. 1, 2) a group of chromosomes was present either in the center of the stained zone or at one side. The eggs were not much elongated. No polar bodies had yet been given off. Later the eggs rounded up and a few showed polar bodies. One hour and 5 minutes after fertilizing about half of the eggs had di- vided. Comment: The polar spindle had not reached the surface (after 7 minutes) when the eggs were centrifuged. They were kept on the machine for a half hour during which time the first polar body was overdue. The stained preparations show that the chromosomes had come to the surface on the machine, but the polar body was not extruded. It is probable that the formation of the polar spindle had been interfered with. Even after removal few polar bodies were found. (2) Three minutes after removal the eggs were fertilized and kept for 13 minutes. They were then centrifuged for 5 minutes at 1110 r.p.m.; for 5 minutes EFFECTS OF CENTRIFUGING ON POLAR SPINDLES 347 at 2080 r.p.m. ; and for 5 minutes at 2750 r.p.m. The eggs were somewhat elongated (Fig. 3). No polar bodies had been given off. A group of chromosomes sur- rounded by a clear area was present in the stained zone. The center of the yolk was more translucent than the periphery. Some of the eggs gave off one or two polar bodies 20 minutes after the normal eggs. Later many eggs divided, most with one, a few with two polar bodies ; but in some eggs no polar bodies were found. Comment: The polar spindle must have reached the surface (after 16 minutes) before the eggs were centrifuged. After 15 minutes on the centrifuge no polar bodies had been given off although overdue about 10 minutes. Later some eggs showed a polar body 10 minutes after those of the normal control. It appears that the polar spindle had retreated or had been driven into the interior, but later, in some of them at least, it came to the surface and gave off one or two polar bodies when returned to sea water. (3) Five minutes after removal eggs were fertilized and ll1/^ minutes later were centrifuged at 2080 r.p.m. for 5 minutes ; then for 5 minutes at 2750 r.p.m. When removed the eggs were somewhat elongated (Fig. 4). The chromosomes are at the surface of the stained zone. A spindle is not evident. The centrifuging (after 16^ minutes) was too late or not strong enough to dislodge the chromosomes from the surface. The eggs showed polar bodies (one, or two?) 12 minutes after removal. The eggs had become more rounded. Antipolar lobes came in soon after this (Figs. 5, 6, 7). Comment: These eggs, in preparations, showed usually only one polar body, or none at all. The lobes in most eggs were in the yolk region, but it is noticeable that in many eggs the antipolar lobe came partly in the stained zone (Figs. 5, 6, 7), which means that such eggs had not oriented on the machine. Half of the eggs (those in the opposite tube) were kept on the centrifuge (at 2750 r.p.m.) for 7 minutes longer. They are more elongated, and the chromosomes are still attached to the surface (Fig. 8). Eggs of this lot were preserved 17 minutes after removal when some of them were dividing. The dividing eggs have one polar body ; those not yet divided showed none in nearly all cases, although the large asters for the first division were present. Comment: The eggs were cen- trifuged about 4 minutes before the first polar body was due. They were on the machine for only 10 minutes. Polar bodies came off 10 minutes after removal, generally only one. Eggs of the same lot were centrifuged for 7 minutes more. They were more elongated and the chromosomes were at the surface. One polar body came off, but whether this corresponds to the first or the second normal polar body can not be stated. In either case the number of chromosomes after fertili- zation should be three-fold. (4) After standing for 5 minutes the eggs were fertilized and then, after standing for 10 minutes more, were centrifuged at 1700 r.p.m. for 30 minutes. The eggs were somewhat elongated (Figs. 9, 10), but no polar bodies had been given off, at least from most eggs. A clear nucleus was found in some eggs (Fig. 9) and a bunch of chromosomes in others (Fig. 10). A few began to divide 5 minutes later. In some eggs, where the constriction for the first furrow is appearing at the top of the stained area, the oil is pinched off as a separate globule (Figs. 11, 12, 13, 14), and if it happens to contain the polar body nucleus on the surface it resembles a large polar body. But there can be no doubt that this is an event taking place long after the polar division took place. The eggs in the opposite tube were centrifuged for 8 minutes longer (Figs. 15, 16). They were more elongated ; some gave off a polar body in or near the oil, others did not. A nuclear vesicle, or a group of chromosomes, lies in the stained zone. Comment: These eggs were centrifuged about 5 minutes before the polar body was due and kept on the machine for 30 minutes longer. No polar bodies were present, at least in most eggs, and a nucleus or a collection of chromosomes was present in the interior. Evidently the chromosomes had retreated from the pole, but the polar spindle did not divide the egg while on the machine. (5) After 7 minutes in sea water the eggs were fertilized, and stood for 10 348 T. H. MORGAN minutes in sea water when they were centrifuged at 1110 r.p.m. for 40 minutes. When removed the eggs were elongated and, some at least, had given off a polar body (or two) (Figs. 17, 18). They began to divide 10 minutes after removal (Fig. 19). Comment: These eggs were centrifuged about 3 minutes before the polar body was due. Some eggs gave off a polar body while on the machine. They divided 10 minutes after removal (Fig. 20), without polar bodies or some- times with only one. (6) Eggs, after standing 5 minutes, were fertilized and stood for 10 minutes. They were centrifuged at 2080 r.p.m. for 25 minutes and for 15 minutes at 2750 r.p.m. They were much drawn out and broken ; many middles were present. Eggs in one of the tubes were centrifuged for 12 minutes more. These too were much drawn out and many broken. The cylindrical eggs contracted and 12 min- utes later some of the whole eggs showed the oil pinched off (Fig. 21). Some of the middle also showed a constriction (Fig. 22). This was 10 minutes at least before the cleavage was due. Without an examination of stained preparations the large clear globules containing oil might seem to be giant polar bodies, but prep- arations show they are without nuclei and quite clearly due to partial fusion of top and middle. Comment: The eggs were centrifuged 5 minutes before the polar body was due, and at a higher speed than the preceding for 40 minutes. One or two polar bodies had been given off, and after removal a few eggs showed the oil region constricted off, but since these bodies did not contain a nucleus and often had one or two polar bodies at the surface, it is clear that these bodies are due to the protoplasm pinching in two, and not due to a division by the polar spindle. (7) Eggs were kept in sea water for 3 minutes, then fertilized and after 5 minutes were centrifuged at 1110 r.p.m. for 3 minutes; at 2080 r.p.m. for 2 minutes; and at 2080 r.p.m. for 15 minutes. The eggs were elongated (Figs. 23, 24). The chromosomes were at the surface in the stained area, generally in the more faintly stained neck, sometimes in two groups. Rarely a polar body was found, though, as a rule, no polar bodies had been given off. The eggs were contracted 8 minutes afterwards (Figs. 25, 26), but no polar bodies were present. Later some of the eggs divided without polar bodies ; other eggs had not divided at this time. Com- ment: The eggs were centrifuged 8 minutes after removal, hence about 12 minutes before the polar body was due and probably before the spindle had reached the surface. They were on the centrifuge for 20 minutes. The polar body was over- due when removed. The preparations show the polar chromosomes at the surface but only rarely a polar body had been given off. (8) Eggs stood 5 minutes in sea water and were then fertilized and after 8 minutes more were centrifuged at 2080 r.p.m. for 10 minutes. The eggs were oval. Preparations show a group of chromosomes in the center of the egg in the stained zone. No polar bodies were observed in the living egg, or in the preparations. Most of these eggs cleaved at the same time as the normal. In the opposite tube the rest of the eggs were centrifuged at 2750 r.p.m. for 10 minutes longer. These were more elongated. The chromosomes were still in the center of the stained zone (rarely at the surface). No polar bodies came off, and no giant polar divisions were observed. Comment: The eggs were centrifuged 13 minutes before the polar body was due. The spindle must have been at or near the surface when the eggs were put into the machine. Centrifuged for only 10 minutes the eggs were not much elongated. The chromosomes were found in the preparations in the interior of the stained zone. No polar bodies came off after removal and the eggs cleaved at the normal time. Other eggs of the same lot, centrifuged 10 minutes longer, gave the same results. (9) The next set was not fertilized. After standing 6 minutes the eggs were centrifuged for 10 minutes at 1110 r.p.m.; then for 10 minutes at 2750 r.p.m. (Figs. 27, 28). The eggs were, for the most part, separated into middle and bottom ; the oil was off in most cases. In the middles the chromosomes were found at the surface (in clumps) (Figs. 27, 28). Some middles were drawn out EFFECTS OF CENTRIFUGING ON POLAR SPINDLES 349 PLATE I into long strands. In the opposite tube the eggs were taken off after 20 minutes. They were elongated and fewer were broken apart (Figs. 29, 30, 31). They were then fertilized at 9.51, i.e. after two minutes. Seven minutes later they were round. Eleven minutes later a polar body was not present in most eggs. The eggs began to divide after 23 minutes (Figs. 32, 33, 34, 35) but did not show any polar 350 T. H. MORGAN bodies as a rule. Comment: This set of eggs was not fertilized before centri- fuging. The eggs were put in the centrifuge 6 minutes after removal. At the time the germinal vesicle was just breaking down. They were centrifuged for 20 minutes. Later, when fertilized, they did not give off a polar body, or perhaps rarely. The meiotic spindle did not bring about division of the eggs. ( 10) Unfertilized eggs stood in sea water four minutes ; they were then centri- fuged for 15 minutes at 2080 r.p.m. The eggs were much elongated. Fertilized, they contracted, some into two parts (as shown in Fig. 36). Later, some of them had a polar body at the side, and later still divided with an antipolar lobe. Eggs in the opposite tube were centrifuged at the same rate for 10 minutes longer. They were more elongated, and many consisted of two bulbs connected by a narrow neck ; others were broken apart. The eggs were fertilized, and in the course of 10 min- utes most of them had contracted into an oval form; but a number of them con- tracted into two spheres joined by a narrow neck. Preparations of these showed that the two parts were not the result of a division, as shown by the position of the chromosomes in the outer sphere and the presence of a polar body at the sur- face (Figs. 36, 37, 38). Comment: This experiment shows that even when the eggs were elongated at the time of fertilization the meiotic spindle did not divide the egg. If by chance it had happened to lie in the neck, it is possible that condi- tions there might have led to the separation of the anaphase plates, but if this should happen the division of the protoplasm into two parts would not be due to the meiotic figure but rather to the contraction of the distal and proximal parts, or bulbs of protoplasm and the breaking of the connecting strand. (11) Unfertilized eggs were put into the centrifuge tubes one minute after removal when the germinal vesicle was still intact, and rotated for 5 minutes at 1110 r.p.m. The contents had begun to be stratified and the germinal vesicles had disappeared. After killing in picro-acetic, the sections were stained in iron haema- toxylin and saffranin. A broad oil cap is present, followed by a clear zone (stippled in Fig. 39). The centrifugal end of the egg has a clear faintly coagu- lated reddish sphere. A deeply-stained, blue crescent of finely granular material lies over it. Centrifugal to this clear area is a half circle of large, deeply-stained granules. The outermost layer is stained red and is without obvious granules. The scattered chromosomes lie in the clear zone, about the middle of the egg, and just above the blue crescent. A large black nucleolus lies somewhere in the same region. Beneath the egg-membrane there is a layer of minute red granules. I do not know what material makes up the blue crescent, perhaps chondriosomes. The deeply-stained black granules may be called yolk, which leaves unnamed the red outer layer of the distal end in which the black granules are imbedded. Some of the remaining eggs were fertilized 5 minutes later. The extrusion of polar body was little delayed, if at all, in these eggs. It appeared most often at the edge of the oil field or even in its center. Later most of the eggs divided, but most of them showed only one polar body in the preparations. Half of the same lot of eggs were centrifuged for 5 minutes more at 1110 r.p.m. (Fig. 40). They gave nearly the same picture as before (Fig. 39), except that the chromosomes are collecting around an aster, and the nucleolus has disap- peared. The oil area is smaller and, after fertilization, the polar body area lies outside the oil field. The extrusion of the polar body was delayed (8 minutes). Later some eggs were found to have two polar bodies. (12) Unfertilized eggs of another lot were put at once in the centrifuge and rotated for one-half hour at 2080 r.p.m. Many eggs were broken. The preserved, intact eggs, sectioned and stained in Delafield's haematoxylin, showed four distinct layers with the chromosomes arranged around a clear body, probably the nucleolus, in the center of the inner stained zone, or at its surface. No mitotic spindle was found in either case. The rest of the eggs were centrifuged one-half hour longer. The conditions were practically the same as in the preceding set. In both cases some eggs were later fertilized. One or two polar bodies came off, and later most tops and whole eggs divided. EFFECTS OF CENTRIFUGING ON POLAR SPINDLES 351 It will be noticed that when the eggs are centrifuged at once the germinal vesicle breaks down, setting free the chromosomes in the middle layer of the egg. Remaining on the centrifuge for one-half to one hour the group of chromosomes either stays in place, or is carried or moves to the surface. The spindle must develop later in some eggs, at least, if they are fertilized, since a polar body may be given off. In other cases the chromosomes remain in the interior, possibly because the egg itself has passed to a later stage of its development, or because the meiotic spindle does not develop, or because if it does, it fails to carry the chromo- somes to the surface. Also to be taken into account is the fact that the mass of eggs is crowded together at the end of the centrifuge tube where a lack of oxygen may slow down the development. In such cases the whole egg would be expected to be held back, and then resume its processes at the stage where it was stopped. (13) The eggs were allowed to stand for 15 minutes in sea water. The polar spindle would then be at or near the surface. The eggs were centrifuged for 15 minutes at 2080 r.p.m. Most of them were killed in Flemming, sectioned, and stained in haematoxylin and eosin. As shown in Fig. 41 five zones are easily recog- nized in the somewhat elongated egg. The oil cap is deep; beneath it lies a stained granular zone. The spindle lies in this zone near the surface. Next comes a less stained zone of finely granular material. This is followed by a zone of larger granules that run in streaks as though moving outward. Then follows a zone of large granules or lumps buried in a less stained material (probably the red stained material of eggs killed in picro-acetic stained with eosin). The other half of the same lot were centrifuged 15 minutes more at 2080 r.p.m. The eggs were more elongated (Fig. 42) and showed the same five layers as those of Fig. 41. The distal dark layer is more nearly fused with the outer black layer, but the preparation is so black that the distinction between these two layers is more difficult to see. Later, the centrifuged eggs of this set were fertilized after 40 minutes. Some eggs showed polar bodies at one side (after 15 minutes), then became pear-shaped ; a lobe appeared and the eggs divided normally at the normal time. Neither the elongation nor the stratification interfered with normal cleavage. A Double Egg One giant egg was found in a series of sections of centrifuged eggs (Fig. 43). It is here drawn to the same scale as the others, but owing to imbedding etc., it is more shrunken. Its double size is shown by comparison with other normal-sized eggs (Fig. 44) on the same slide. The normal eggs show polar spindles in anaphase with 9 chromosomes in each half -plate (Fig. 45). The chromosomes of the giant are scat- tered in the center of what appears to be a large bipolar spindle (Fig. 46). There are about 18 chromosomes, which is double the normal number, but if they have already divided this would not be the case. There is no real evidence, however, that they have divided as yet, and moved apart. Fused Eggs Fused eggs were found in two preparations (Figs. 47-51). The eggs had stood for 3 minutes in sea water, were then fertilized, and after 13 minutes were centrifuged at 1110 r.p.m. for 2 minutes, at 2080 r.p.m. for 5 minutes and at 2750 r.p.m. for 15 minutes. The eggs were 352 T. H. MORGAN 31 51 EFFECTS OF CENTRIFUGING ON POLAR SPINDLES 353 about twice as long as wide. Amongst them were three " fused eggs " (Fig. 50). The yolks are completely fused but the protoplasmic parts are still separated and each contains a polar spindle. These eggs were killed 28 minutes after fertilization. There is one other case in the same preparation in which five eggs are fused, at least they have a large common yolk, the protoplasm and oil of each making different angles with the yolk. The fused eggs must have been stratified before fusing, since the stained zone and oil cap make different angles with the yolk. There is one other preparation (Fig. 51) that has one double egg. It had been treated in much the same way as the last, but its history is not entirely clear. Centrifuging the Eggs of Cuiningia When the eggs of Cuiningia are ejected from the siphon there is present a well-developed spindle that lies somewhat excentrically in the interior of the egg in a more or less horizontal position or else pointed toward the pole. It remains there until the egg is fertilized and then moves to the pole. When the unfertilized egg is centrifuged the spindle remains in the clear zone (see below). After removal to sea water and fertilization, the polar bodies are given off without respect to the strati- fication, which means that the eggs are little, if at all, oriented in the tubes. The eggs are faintly pink; some lots have more pigment than others. Most of the pigment lies in a broad zone in the outer region of the egg. When centrifuged the oil is quickly driven to the centripetal end. When centrifuged for a longer time (one hour) at a higher speed, the pigment is driven to the end opposite the oil leaving a broad clear middle zone that fills most of the egg. If this zone is carefully ex- amined it is found to consist of a clear zone beneath the oil and a faintly yellowish zone next to the pigment. There is a rather sharp boundary line between these two if the eggs have been sufficiently centrifuged. If the living egg is stained, by adding a little methyl green to the sea water, the zone next to the pigment becomes slightly bluish-green, in- creasing the difference between it and the protoplasmic zone. In 1908, 1909, and 1910 I described the results of centrifuging this egg. Four zones were recognized, but the difference between the proto- plasmic and yolk zone was not clearly distinguished and what is iden- tified as yolk in the sections (1910, Plate II, Fig. N-U) is the pigment plus probably some yolk also. In 1934 Costello recognized four zones identifying the yolk and the pigment as two separate zones. Stronger centrifugation than I had employed was used by Costello, vis. 6000 gravity. Sections of eggs killed immediately show a small clear vesicle just 354 T. H. MORGAN beneath the surface not previously recognized. Its origin and nature was not determined, but in a set of eggs that had stood three-quarters of an hour in sea water there was observed a clear cap on the eggs sur- rounded by a faint ring of pigment. Whether this cap is derived from the clear vesicle is not known. This cap had been seen in 1910 and recognized, but such eggs may have stood for some time in sea water. Sections were made of normal eggs and of those after centrifuging. They were killed in Flemming or in Bouin solutions and stained in haematoxylin and different dyes. The normal egg, after Flemming, contains a number of deeply-stained granules that are scattered through the protoplasm around the spindle, except at one point, the pole of the egg towards wrhich the spindle is pointed. When the living egg is centrifuged these granules are driven to the centrifugal pole where they fill a segment of the egg that is red or pink in color. This region stains yellow after certain killing and staining reagents, and the granules in it are difficult to see and perhaps tend to fuse together. In the earlier paper (1910, Plate II) this region is spoken of as the yolk. Sometimes black stained pigment granules lie more or less scattered in this yellowish " yolk " depending on methods of preparation. Whether there is pres- ent in addition to the pigment granules another material in this part of the centrifuged egg in which the dark granules are imbedded is uncertain. The region next to this is distinguished from the clear proto- plasmic region only with difficulty in the living centrifuged egg. In sections of stained eggs it can also be distinguished from the protoplasm zone. Whether it is " yolk " material in the ordinary sense, or another kind of substance, or yolk of a somewhat different nature, remains uncertain. When the living centrifuged eggs are put into Flemming solution after centrifuging the pigment segment immediately becomes brown. The oil segment at the opposite end more slowly becomes brown, but later jet black. It holds this color throughout the treatment with the reagents subsequently used (water, alcohol, xylol, etc.), but the pigment section loses its dark stain and may stain a light yellow if picric acid is present or stains red in eosin. In some of my earlier papers (1910) the pigment region of the centrifugal hemisphere of the living egg is colored red. These eggs were not centrifuged as long or at as great a speed as were those mentioned above, where the pigment is driven further into the centrifugal pole to form there a red segment. On the other hand, in some of the stained sections (Plate II) there is a yellow seg- ment that is spoken of as yolk and no dark granules are indicated. As stated above, this " yolk " is either the pigment granules that have EFFECTS OF CENTRIFUGING ON POLAR SPINDLES 355 coalesced in the preserving reagent, or a substance in which they are imbedded. Drawings of other sections of eggs (in that paper) that were prob- ably preserved in another reagent (Flemming), show the pigment gran- ules scattered in the protoplasm before centrifuging; and after centri- fuging they are represented, as here, driven to the centrifugal region. The presence of a large mitotic spindle in the interior of the Cumingia egg suggested that the egg might be a suitable object to examine whether the spindle could be induced to divide the egg, producing a " giant polar body." However, since this spindle and its equatorially arranged chro- mosomes do not become active until a spermatozoon enters the egg, it becomes necessary to fertilize the egg either before or after centrifuging to activate the spindle. Both procedures have been used, but with two possible exceptions no such division was brought about. Eggs centrifuged before fertilization, removed to sea water and then fertilized give off at least one, and in most cases probably both polar bodies. Eggs centrifuged after being fertilized give off polar bodies either on the machine or after removal. The length of time in the centrifuge tube was also taken into account, since eggs that were not lying in the axis of centrifuging or even inverted would be expected to behave differently if the centrifugal force was sufficient to keep the polar spindle from moving to the true pole. If, on the other hand, the pole should lie at the more centripetal end of the protoplasmic zone, the movement of the spindle and its surrounding rays might even be facilitated while on the centrifuge. In the earlier experiments that I made (1910) on the egg of Cumingia I found that the polar spindle may push its way through the yolk or pigment or oil and give off the polar bodies at the surface. This means that neither of these regions presents a serious obstacle to the movements of the spindle, although statistical evidence is lacking as to whether to some extent there may be some difficulties in reaching the pole when the yolk underlies it. More- over, in these cases the centrifuging was low, and not long enough to separate visibly the pigment and yolk fields. If centrifuged at a higher rate, or for a longer time, the outcome might have been different. A few examples may be cited as follows: Unfertilized eggs were centrifuged for 15 minutes at 1664 r.p.m. at 10% cm. from the center of rotation. When removed the eggs were not elongated or very little so. Fertilized after 5 minutes many eggs were seen to be giving off a polar body after 15 minutes at the side of the clear zone. No giant polar bodies were formed. A few eggs about twice the size of others were recorded in this set. Another set was fertilized and stood for 12 minutes in sea water and 356 T. H. MORGAN was then centrifuged for 10 minutes at 1664 r.p.m. They were spheri- cal or columnar, and a few showed a dent at the side. After 6 minutes some eggs extruded a polar body. Only a few eggs divided; also the controls were not normal. One set centrifuged as above was then fertilized. Eight minutes later polar bodies were observed, some of them exceptionally large. No second polar body was observed 7 minutes later, but 5 minutes later one polar body was seen in some eggs and rarely two. A deep constriction was observed in some eggs at this time. The eggs divided 10 minutes later. The following cases apply to eggs centrifuged after fertilization. Since the polar spindle is active only after fertilization it is here rather than in the preceding cases that we might hope to bring about a division of the egg by means of the polar spindle, especially when the eggs were kept on the centrifuge during the active phase of the spindle. Four minutes after fertilization the eggs were centrifuged for 20 minutes at 1664 r.p.m. The first polar body is expected about 20 min- utes after fertilization. Two minutes after removal an egg was found divided into two nearly equal parts. This is one of the two cases ob- served that might be interpreted as due to the polar spindle dividing the egg. Nine minutes later polar bodies were observed in some eggs. Eleven minutes later the eggs divided. In the control, normal eggs, the first polar body came off 21 minutes after fertilization. There was, then, some delay in polar body extrusion in the centrifuged eggs. Another set stood for 10 minutes after being fertilized. They were then centrifuged for 12 minutes at 3260 r.p.m. When removed the first polar body had been given off while in the centrifuge tube in some eggs. Later, when the eggs were divided, one, two or no polar bodies were found. Another set was fertilized and allowed to stand until the first polar body was out (18% minutes). This was an attempt to find out whether the second polar spindle was more efficacious in dividing the egg than the first one. The eggs were centrifuged for 28 minutes at 1664 r.p.m. They were then a little oval, and some showed a slight constriction. No divisions were present. These few experiments are only exploratory. The negative results, with two possible exceptions, can not be understood to exclude the pos- sibility that under other conditions the polar spindles would divide the egg, but the experiments were not promising. Morgan and Tyler, in 1935, centrifuged the eggs of Urechis under conditions similar to those described here. In one respect the egg is different from those of Chaetopterus and Cumingia; the germinal vesicle EFFECTS OF CENTRIFUGING ON POLAR SPINDLES 357 remains intact until the eggs are fertilized ; it then breaks down, and the dent in the egg disappears. Eggs centrifuged before fertilization give off the first polar body at the true pole and produce normal embryos. If centrifuged after fertilization the chromosomes are driven towards the pigment side of the egg (centrifugally), and the polar bodies are given off somewhere in that hemisphere ; if they are given off at all. The evidence indicates that these eggs fall at random on the machine, hence in many cases the polar bodies are given off at other regions of the egg than that of the true pole, as Conklin also found in Crcpidula. The evidence also shows that the first and second cleavage planes pass through the secondary pole, but these eggs do not produce normal embryos. It can not be said, therefore, that a new polarity has been imposed on the egg, in so far as " polarity " is an essential feature of development. When the eggs are on the centrifuge at the time when the chromo- somes are set free from the germinal vesicle and the polar spindle is in process of formation, separate asters may appear in the same regions as the chromosomes. Later a spindle may develop between two asters and these are the eggs that give off polar bodies. When, on the other hand, the chromosomes remain in the egg, most of them are collected on the cleavage spindle; but some may remain outside attached to the astral rays. The abnormal development of the embryo may, in part, be due to the irregular distribution of the chromosomes. If eggs are centrifuged after the extrusion of the first polar body, the second polar spindle may be displaced from the pole, but does not bring about the division of the egg. Whether it returns to the pole or gives off a polar body at some other point was not definitely settled, although in a few cases two polar bodies were observed at different points on the surface. In none of these experiments were the eggs kept long enough on the centrifuge to pass by the time when the polar bodies are formed in the normal egg. In this respect the experiments with Chaetopterus and Cumingia cover somewhat different grounds. BIBLIOGRAPHY CLEMENT, ANTHONY C, 1935. The formation of giant polar bodies in centrifuged eggs of Ilyanassa. Biol. Bull, 69 : 403. CONKLIN, EDWIN G., 1915. Why polar bodies do not develop. Proc. Nat. Acad. Set., 1 : 491. CONKLIN, EDWIN G., 1916. Effects of centrifugal force on the polarity of the eggs of Crepidula. Proc. Nat. Acad. Sci., 2: 87. CONKLIN, EDWIN G., 1917. Effects of centrifugal force on the structure and de- velopment of the eggs of Crepidula. Jour. Exper. Zool., 22: 311. LILLIE, FRANK R., 1906. Observations and experiments concerning the elementary phenomena of embryonic development in Chaetopterus. Jour. Exper. Zool, 3: 153. 358 T. H. MORGAN LILLIE, FRANK R., 1909. Polarity and bilaterality of the annelid egg. Experiments with centrifugal force. Biol. Dull., 16 : 54. LILLIE, FRANK R., 1909. Karyokinetic figures of centrifuged eggs ; an experi- mental test of the center of force hypothesis. Biol. Bull., 17: 101. MORGAN, T. H., 1908. The effects of a centrifugal force on the eggs of Cumingia. Science, 27 : 446. MORGAN, T. H., 1909. The effects produced by centrifuging eggs before and dur- ing development. Anat. Rec., 3 : 155. MORGAN, T. H., 1910. Cytological studies of centrifuged eggs. Jour. Exper. Zool, 9 : 593. MORGAN, T. H., 1927. Experimental Embryology. Columbia University Press. New York. MORGAN, T. H., 1933. The formation of the antipolar lobe in Ilyanassa. Jour. Exper. Zool., 64 : 433. MORGAN, T. H., 1935o. Centrifuging the eggs of Ilyanassa in reverse. Biol. Bull., 68 : 268. MORGAN, T. H., 1935&. The separation of the egg of Ilyanassa into two parts by centrifuging. Biol. Bull., 68 : 280. MORGAN, T. H., 1935c. The rhythmic changes in form of the isolated antipolar lobe of Ilyanassa. Biol. Bull, 68 : 296. MORGAN, THOMAS HUNT, 1936. Further experiments on the formation of the antipolar lobe of Ilyanassa. Jour. Exper. Zool., 74: 381. MORGAN, T. H., 1937a. The behavior of the maturation spindles in polar frag- ments of eggs of Ilyanassa obtained by centrifuging. Biol. Bull., 72 : 88. MORGAN, T. H., 19376. The factors locating the first cleavage plane in the egg of Chaetopterus. Cytologia, Fujii Jubilee Vol. p. 711. MORGAN, T. H., 1938. A reconsideration of the evidence concerning a dorso- ventral pre-organization of the egg of Chaetopterus. Biol. Bull., 74: 395. MORGAN, T. H., AND ALBERT TYLER, 1935. Effects of centrifuging eggs of Urechis before and after fertilization. Jour. Exper. Zool., 70: 301. MORGAN, T. H., AND ALBERT TYLER, 1930. The point of entrance of the sperma- tozoon in relation to the orientation of the embryo in eggs with spiral cleavage. Biol. Bull., 58: 59. TYLER, A., 1930. Experimental production of double embryos in annelids and mollusks. Jour. Exper. Zool., 57 : 347. WHITAKER, DOUGLAS, AND T. H. MORGAN, 1930. The cleavage of polar and anti- polar halves of the egg of Chaetopterus. Biol. Bull., 58 : 145. WILSON, EDMUND B., 1929. The development of egg-fragments in annelids. Arch, f. Entiv.-mech., 117: 179. WILSON, EDMUND B., 1930. Notes on the development of fragments of the fer- tilized Chaetopterus egg. Biol. Bull., 59: 71. THE REPRODUCTIVE CYCLE OF THE VIVIPAROUS TELEOST, NEOTOCA BILINEATA, A MEMBER OF THE FAMILY GOODEIDAE x I. THE BREEDING CYCLE GUILLERMO MENDOZA (From the Department of Zoology, University College, Northwestern University) INTRODUCTION Much attention has been devoted recently to the reproductive cycles of viviparous teleosts. Some of the recent work has been concerned with the histological phases of the gonad during reproduction (Scott, 1928; Bailey, 1933; and Turner, 1933, 1938&) but more attention has been devoted to different aspects of the breeding cycle. In this respect, much of the work has been concerned with members of the family Poe- ciliidae (order Cyprinodontes) because of the large number of vivi- parous species available and because of their relative abundance (see Turner, 1937a for recent work on the poeciliids). With the exception of three papers (Turner, 1933, 1937&; Mendoza, 1937) nothing more has been reported on the family Goodeidae from the central plateau of Mexico, a family which contains twenty-six viviparous species. In view of the dearth of information on the goodeids, this investigation was undertaken to study the female reproductive cycle of Neotoca bitineata in order to throw some light on the reproductive activities of this viviparous family. The information obtained in this investigation involves three phases of the reproductive activity of Neotoca: (1) the breeding cycle, (2) the germ cell cycle during gestation, and (3) the somatic cycle of the ovary. The present paper is devoted entirely to the phenomena associated with breeding as observed in the laboratory: the breeding season, the lengths of the brood intervals, the sizes of broods, superfetation, fertilization, and color variations in the female. The periodic variation of the germ cells in the ovary and the gross and microscopic changes which the ovarian soma undergoes during gesta- tion will be dealt with in later papers. 1 I wish to express my appreciation for the advice and guidance of Dr. C. L. Turner under whose direction this investigation was undertaken. 359 360 GUILLERMO MENDOZA MATERIALS This form was chosen because of the number of females available, their convenient size, and because of the rapidity of brood production. Specimens of Neotoca used during the investigation were raised in the laboratory. A record of over eighty females was kept during the course of the investigation ; approximately thirty-five of these were killed and serial sections made of the ovary for microscopic examination. No single female was kept during the entire period of observation, which lasted from the fall of 1934 to the spring of 1937, since specimens frequently died or were killed during critical periods of the year and were replaced by new females. The individual females under observa- tion always were isolated with one or two males, the isolation tanks con- sisting chiefly of one- or three-gallon aquaria. All tanks were filled with seasoned water and abundantly stocked with plants. With the exception of a few aquaria, all were kept in the same room, subject to a fairly constant room temperature of about 70° F. and the normal, daily fluctuation of light. Electric lights were turned on but infre- quently at night so that any possible effects of artificial illumination on the reproductive cycle were negligible. Records were kept of the date of isolation of the females, frequency of brood production, number and size of young, and the condition of the maternal gonad at time of death. In most cases, when females were killed a unilateral dissection of the ovary was made and measurements of the young on that side were taken ; the opposite side of the ovary was left intact for sectioning. In a few females, the entire ovary was sectioned without the unilateral dissection. BREEDING SEASON The following description of the breeding season is based solely on laboratory observations ; there has been no study of the females in their natural habitat such as was performed by Scott on Fitsroyia lineata (1928) and by Turner on Brachyrhaphis episcopi (1938«). Table I shows the preponderance of broods appearing during spring and early summer. Series A consists of older females isolated during the early fall of 1934. Of the nineteen females isolated, only six had October broods and only one of the six started further gestation that fall. The other five specimens already had started a period of sexual in- activity or rest destined to last until the following spring. The remain- ing females of this series were allowed to live through the fall and winter but they produced no more broods ; nevertheless, the following spring all but one of the females started to breed again. Evidence of a period of THE BREEDING CYCLE OF NEOTOCA BILINEATA 361 reproductive inactivity over the winter is obtained from the fact that over half of the females in Series A had no fall broods. Additional evidence of this winter inactivity appeared constantly during the investigation. Early in 1935, the individual females of Series B were isolated and records of brood production were started. It is evident from Table I that but few broods were produced during January and February ; how- ever, almost all females produced broods during March and April. Dur- ing May there was a decrease in the number of broods followed by a further decrease during June and July. The latter month marks the termination of brood production. Females that were alive after July bore no young during the summer, fall or winter. The females of Series C were isolated during the fall of 1935. With the exception of one specimen which had young during December, no other broods were produced until March. It is evident again that broods appeared predominantly during spring and early summer. July again marks the termination of brood production. Series D comprises a small group of females isolated during the fall of 1936. Of about one dozen females isolated, only three had had young by February of 1937 at which time the records were discontinued. For some reason not understood, many deaths were recorded during the months of May, June, and July whereas deaths during other times of the year were most infrequent. The marked mortality evident in the females so jeopardized the regularly breeding specimens that frequently they were sacrificed during July and preserved rather than to risk death and subsequent loss of the tissues for histological examination. If it had not been desired to get a complete histological record of the ovary during gestation, females would have been permitted to live as long as possible so as to get a continuous record of breeding throughout various seasons. It is possible then that despite the deaths of some females, few would have produced a small number of broods during summer, fall, and winter. The total number of broods produced during each month over a period of three years was as follows : Jan. Feb. Mar. Apr. May Jun. Jul. Aug. Sept. Oct. Nov. Dec. 6 10 19 29 18 14 10 0 0 4 2 1 In summary it may be said that brood production occurs chiefly dur- ing spring and early summer whereas the greatest number of broods appears during the month of April. The months of August and Sep- tember show no brood production. This may be caused in part by the fact that most females had been killed or had died by the latter part of July. Among the females that remained, however, no brood production 362 GUILLERMO MENDOZA was observed. It is very likely that, had older females been used, more fall broods would have been recorded. BROOD INTERVALS Some viviparous fishes such as Poecilistes plcurospilns and Heter- andria formosa (Turner, 1937a) are characterized by a marked decrease TABLE I The complete record of breeding from the fall of 1934 to the early spring of 1937. The number of the different females is merely for purposes of identification. The columns marked "Brood I," "Brood II," etc., indicate the dates of the first, second broods etc.; the dates are expressed as fractions, the numerator indicates the month of the year, the denominator indicates the day. The columns marked "Interval" include the interval of days between successive broods. Days are recorded on which the females were killed or died naturally; the asterisk following the date of death indicates that the female was bearing young at the time. Female Brood I In- terval Brood II In- terval Brood III In- terval Brood IV Female killed Female died 1 10/1 11/12* 2 10/8 11/12 3 10/11 11/15 4 10/22 11/19 1934 5 11/7 11/21 6 11/14 11/21 Series 8 2/6* A 9 2/6* 10 2/6* 11 2/6* 12 2/6* 1935 13 2/6* 14 2/6* 15 2/23 16 2/15 3/4* 19 3/9 4/1 1 1/8 62 3/11 5/7 2 1/29 50 3/20 15 4/4 5/10 3 2/2 44 3/17 40 4/26 39 6/4 7/4 4 1/1 5 5/29 6 2/10 71 4/21 48 6/8 7/9 7 2/13 55 4/8 4/26 Series 9 2/15 45 3/31 41 5/11 1935 B 10 2/15 17 3/3 5/17 11 3/4 43 4/16 41 5/27 17 3/11 45 5/4 18 3/10 66 5/16 6/3 19 3/21 40 4/30 40 6/9 40 7/17 7/29* 20 3/16 41 4/26 44 6/9 40 7/19 7/24 21 3/23 36 4/28 44 6/11 7/17 22 3/23 42 5/4 36 6/10 40 7/20 8/5* 23 3/27 43 5/9 48 6/26 7/21* THE BREEDING CYCLE OF NEOTOCA BILINEATA 363 TABLE I (continued} Female Brood I In- terval Brood II In- terval Brood III In- terval Brood IV Female killed Female died 24 3/27 51 5/17 54 7/10 7/21 25 4/1 41 5/11 44 6/24 26 4/19 7/20 27 4/5 42 5/17 44 7/1 7/21* 28 4/6 6/4 29 4/6 31 5/7 30 4/9 48 5/27 31 4/9 7/12 32 4/9 45 5/24 Series 33 4/16 6/4 1935 B 34 4/15 46 6/1 7/15 35 4/20 36 4/25 40 6/4 45 7/19 7/20 37 4/24 5/21 38 4/26 40 5/28 43 5/21 44 5/13 46 6/3 47 7/20 47 7/19 8/11 48 3/23 60 5/23 7/8 49 12/1 120 4/29 64 7/2 7/8 56 4/23 49 6/11 6/19* 57 3/1 55 4/25 Series 59 6/1 39 7/9 7/16 C 60 3/23 43 5/6 1936 63 5/6 66 6/20 6/20 67 3/7 72 4/8 49 5/27 7/16 73 3/1 41 4/17 6/14* Series 76 1/15 D 77 1/23 1937 • 81 1/12 in length of brood intervals with the advance of the breeding season. Intervals between broods decrease from as much as twenty-six to twelve days in the former and from forty-one to three or four days in the latter. In Neotoca-, however, as the breeding season advances there is no ap- parent increase in the rapidity of brood production. Furthermore, the average length of the brood intervals occurring during different months of the breeding season show no significant difference in length through- out the season (Table II). 364 GUILLERMO MENDOZA Most brood intervals are regular and fluctuate around a normal fixed period. Sixty-eight per cent of the 51 recorded intervals show a maxi- mum variation of 9 days (39—48 days) whereas fully 51 per cent of the intervals vary but 5 days (40-45 days). The numerical average for TABLE II Average length of brood intervals throughout the breeding season. Time Occurrence of Intervals Average Length of Intervals days February-March 43.6 March-April 43.8 April-May 47.3 May-June 43.4 June-July 45.8 the entire group is 45.0 days. Although most intervals are fairly regu- lar, a few variations ranging from 15 to 71 days were recorded. As an example of regularity in brood production, the following case is cited (Table III). It involves the parallel breeding behavior of four females (Series B) occupying adjacent tanks. TABLE III The breeding record of four females in adjacent tanks. Female Brood I Int.* Brood II Int. Brood III Int. Brood IV 19 March 21 40 April 30 40 June 9 40 July 19 20 March 16 41 April 26 44 June 9 40 July 19 21 March 23 36 April 28 44 June 11 22 March 23 42 May 4 36 June 10 40 July 20 * Interval in days. In addition to the above example of parallel breeding, several other instances led the writer to suspect that breeding was unusually plentiful on certain days, as if there were regular waves of brood production among several females. Analysis of the brood record resulted in the following interesting conclusions. There occurred three distinct groups of females, with members of each group having parallel breeding cycles. The breeding period for each group normally lasted about ten days and appeared again approximately forty-five days later. The number of females in each group varied from four to eight individuals, each female producing two or more broods. Each group retained its identity despite the fact that members within a group did not produce their broods in any regular order. The groups furthermore retained the same relative THE BREEDING CYCLE OF NEOTOCA BILINEATA 365 order in their breeding periods throughout the season: ABC, ABC, ABC, AB, A. The groups were not composed entirely of females start- ing brood production during the same month; some females started breeding later in the season and yet their cycle corresponded with one of the group cycles already established. Of a total of 80 broods pro- duced during 1935 between January 1 and July 21, 58 fit into this regular cyclic phenomenon; 13 broods could not be accounted for in this scheme since they were first and only broods ; finally, only five females producing a total of 9 broods participated in irregular cycles which failed to fit into the above rhythmic activity. The breeding records other than those of 1935 were too scanty to permit such an analysis. It is true that the selection of groups was made arbitrarily. There is probably nothing that determines rigidly the number of days during which members of a group will produce young, but segregation of the females was made possible because, regardless of the order of brood TABLE IV Increase in size of broods during the breeding season. Brood Total number of broods Total number of young Average number of young per brood First 53 302 6.0 Second 30 240 8.0 Third 15 163 10.8 Fourth 4 52 10.4 production within the units, the breeding period of one group did not overlap with that of another group. Regularity was maintained re- markably well during the season for a phenomenon which frequently is considered to be sporadic. Finally, the determination of these groups is not significant in itself ; rather, the importance lies in that they demon- strate markedly the regularity of the breeding behavior of the females of this species. It is true, obviously, that there are minor differences in the rate of breeding among the individual females. ( SIZE OF BROODS The present analysis of the size of broods is based on records ob- tained from 53 females which produced a total of 103 broods. On the basis of the information available, it is evident that the broods increase in size during the season. This increase is apparent in the general averages of all first, second, third, and fourth broods (Table IV). 366 GUILLERMO MENDOZA It is emphasized that the above evidence concerns in the main part young females producing their first broods ; consequently, the increase in the average number of young per brood for successive broods actually im- plies that with growth of the females, broods become larger. Although the third and fourth broods were not as numerous as would have been desired, together they indicate the tendency for more numerous young in the third and fourth broods. The difference between the last two broods obviously is too small to be significant. It is to be expected that older females would have produced somewhat larger broods. Only a maximum of four broods is recorded for any one female. It is likely that females have more than four broods during their active reproduc- tive life, but it is doubtful if they have more than four or five broods during one season since the interval between broods is so long. SUPERFETATION An outstanding characteristic of the reproductive cycle of Neotoca is the absence of superfetation. This absence is concluded from three sources : the brood interval, the evacuation of young, and examination of the ovary itself. Broods are spaced far apart and are regular in their appearance. Furthermore, there is no evidence that normally only a part of the young are evacuated at one time. Generally, young are known to be born within minutes of each other. In only two females was there evidence of a phenomenon resembling superfetation. In both cases the first brood was small and was followed in two weeks by another small brood. Since it is improbable that the young were completely developed in that short a time, it is more likely that they were part of the first brood but for some unknown reason were not expelled along with the others. Since this phenomenon failed to appear elsewhere no significance was attached to its occurrence. Examination of many ovaries, both by gross and microscopic methods, failed to reveal the presence of more than one brood in the ovary at one time. Normally, all embryos in an ovary were in the same stage of development. It is true that slight irregularities in size occurred but they were seldom great. Therefore, it may be concluded that each brood is distinct in Neotoca because of the brood interval, the simultaneous evacuation of all young, and because examination of the ovaries in a large number of females revealed but one brood in the ovary at one time. FERTILIZATION In viviparous fishes producing annual broods, copulation and fertili- zation may occur at the beginning of the breeding season or, as occurs THE BREEDING CYCLE OF NEOTOCA BILINEATA 367 in Cymatogaster aggregates (Eigenmann, 1892), copulation may take place several months before the breeding season, in which case the sperm are stored in the ovary until time for fertilization. Among vivi- parous forms producing several broods during one season, the time of copulation may vary. In forms such as the poeciliids, there is a con- tinual introduction of spermatozoa by the male. However, storage of sperm has been reported for Xiphophorus helleri (van Oordt, 1928; Bailey, 1933) and has been determined experimentally in Heterandria fortnosa by Turner (unpublished data). In such cases, spermatozoa are stored and are capable of fertilizing several sets of ovocytes over a considerable period of time. Normally, however, it appears that the storage is not necessary because of the continual introduction of fresh spermatozoa by the male. There is no evidence that such a storage phenomenon occurs in Neotoca; in this species, copulation and fertiliza- tion occur concurrently or nearly so. Spermatophores which have been reported for some viviparous teleosts likewise do not occur in this species. The actual time during which fertilization occurs has been determined as follows. Examination of 7 ovaries with young in late stages of gestation failed to reveal spermatozoa in the ovary or any indication of fertilization. Three females killed at intervals of 5, 6 and 7 days after the birth of young likewise failed to reveal signs of fertilization. How- ever, 5 females killed at intervals of from 7 to 15 days after the birth of young revealed new embryos in different stages of segmentation. Spermatozoa were evident in almost all of these ovaries. Although it is possible, of course, that fertilization occasionally may occur at a dif- ferent time, it is concluded from the above evidence that normal fertili- zation occurs about the seventh day after the release of a previous brood. The details of the cytology of fertilization will be considered in another paper. FEMALE COLORATION Peculiar coloration of the females during certain phases of the breed- ing cycle has been reported by Scott for Fitsroyia. Although the regu- larity and the significance of these color changes have been questioned by some investigators, reference is made herein to a peculiar change in the color of the females. During the first year it was noticed frequently that the abdominal region of the female became a vivid blue whereas normally it is a pale gray. Furthermore, it was noticed that females with parallel breeding cycles assumed this coloration approximately at the same time. The coloration normally lasted a few days and then dis- appeared. If the female became excited, the color could disappear in a few seconds to return later in a gradual manner. Weekly observations 368 GUILLERMO MENDOZA during the second breeding season indicated that in non-breeding fe- males, the coloration appeared very irregularly whereas among the breed- ing females some semblance of regularity was observed. In most cases, the color appeared on an average of 10 days after the birth of a previous brood and 33 days before the birth of the following brood. It is sig- nificant that two virgin females, isolated from the time of birth, showed the same coloration simultaneously, at frequencies comparable to the cycles of breeding females. Microscopic examination of the virgin gonads, fixed during the color phase of the female, revealed the ovarian soma in a very active, turgid state similar to that of an ovary during early gestation. Thus it may be that this color is correlated in some way to a periodic activation of the ovarian soma, whether or not the female is exposed to males. Furthermore, the time of appearance of the color in the breeding females correlates closely with the time of fertilization, thereby suggesting that there occurs periodically in the ovary an inherent activation usually accompanied by color changes in the female and evidently facilitating copulation and fertilization. DISCUSSION Compared with the breeding activities of other viviparous teleosts, Neotoca most closely resembles the poeciliids which in general breed actively during the spring and summer and but little during fall and winter. This seasonal wave of reproductive activity occurs in several genera including Mollienisia, Xiphophorus, Gambusia and others. These poeciliids and Neotoca obviously stand in contrast to viviparous forms such as Cymatogaster, Fitsroyia and others which have but an annual brood. As has been suggested by other investigators (Kuntz, 1913; Turner, 1937a), the writer also is of the opinion that the recurrent broods during the season are comparable to several annual broods rather than to one extended annual brood, as suggested by Barney and Anson (1921) for Gambusia. The former opinion is shared by the writer since in Neotoca there is a definite and complete activation of the ovary in- cluding separate copulation and fertilization for every brood. Whether the laboratory breeding cycle is similar to that of the females in their natural habitat is unknown. The only information available is that obtained at the time the original stock was collected during March by Dr. C. L. Turner. Reproductive activity was present but not at a maximum. It has been shown for Gambusia affinis, at least, that the breeding records of the females in the laboratory and in the field correlate closely. On the other hand. Turner (1938a) has found that Brachyrhaphis episcopi, a poeciliid, fails to show this alternate high and low level of reproductive activity but breeds continuously throughout THE BREEDING CYCLE OF NEOTOCA BILINEATA 369 the year in its normal habitat in the Panama Canal Zone. It is more likely that of the two, Neotoca resembles Gambusia more closely since the temperature and light factors are not quite as constant in its habitat on the Mexican plateau as is true for Brachyrhaphis. It is probable that the information on the breeding cycle of this species will apply to most goodeids although there is one exception known, Goodea luitpoldi, a form which is larger, matures more slowly, has greater brood intervals, and larger young. The cause of the high mortality evident among the females during early summer is unknown. It is unlikely that death normally follows the production of one or two broods as has been reported for another viviparous form, Comcphorus baicalcnsis (Dybowski, 1902). It seems more reasonable to assume that the females are at a low physiological level because of their reproductive activities and hence they are more susceptible to minor changes in food, temperature, or perhaps even to minute organisms in the water. Of these, temperature may prove to be the most important. It has been shown that the breeding cycle remains quite constant throughout the season, differing from forms such as Poecilistes and Heterandria. It is indeed difficult to explain the cause of such a regular production of broods. It is not likely that external conditions are of major importance in the establishment of the regular cycles in Neotoca. However, it may be that food and temperature especially help to bring about the birth of young on some particular day or interval of days if the young are ready or nearly ready for birth. The regular cycles of brood production very likely are governed by some internal, inherent mechanism, a mechanism which is inherited in the species as a whole and probably expresses itself through hormonal control. Once brood production starts, the inherent periodicity expresses itself usually in a regular manner. However, in the absence of careful experimentation, it is unwise to discuss further the relative importance of external or internal factors in the determination of such a marked regularity in breeding activity. Since the embryos remain in the ovarian lumen until time of birth, the ovary of Neotoca may be compared to the uterus of mammals. There is further similarity in that marked periodic activation of the ovarian soma occurs not only in breeding females but also in virgins. A third parallel exists in that fertilization evidently occurs during this period of activation. Consequently, in view of the above comparisons, it is sug- gested that there occurs in the female of this species a period similar to that of oestrus in the mammalian female. 370 GUILLERMO MENDOZA LITERATURE CITED BAILEY, R. J., 1933. The ovarian cycle in the viviparous teleost Xiphophorus helleri. Biol. Bull, 64 : 206. BARNEY, R. L., AND B. J. ANSON, 1921. The seasonal abundance of the mosquito- destroying top-minnow, Gambusia affinis, especially in relation to fecundity. Anat. Rec., 22: 317. DYBOWSKI, B., 1902. Life-history and young stages of the " Fat-Fish " of Lake Baikal. (Abstract) Jour. Roy. Micr. Soc., p. 167. EIGENMANN, C. H., 1892. On the viviparous fishes of the Pacific coast of North America. Bull. U. S. Fish. Com. (1894), 12: 381. KUNTZ, A., 1913. Notes on the habits, morphology of the reproductive organs and embryology of the viviparous fish Gambusia affinis. Bull. U. S. Bur. Fish. (1914), 33: 177. MENDOZA, G., 1937. Structural and vascular changes accompanying the resorption . of the proctodaeal processes after birth in the embryos of the Goodeidae, a family of viviparous fishes. Jour. Morph., 61 : 95. SCOTT, M. I. H., 1928. Sobre el desarollo intraovarial de Fitzroyia lineata (Jen.) Berg. Anal. Museo Hist. Nat. de Buenos Aires, 34., (Ictiologia, pub. num. 12). TURNER, C. L., 1933. Viviparity superimposed upon ovo-viviparity in the Goodei- dae, a family of Cyprinodont teleost fishes of the Mexican plateau. Jour. Morph., 55 : 207. TURNER, C. L., 1937o. Reproductive cycles and superfetation in poeciliid fishes. Biol. Bull., 72 : 145. TURNER, C. L., 19376. The trophotaeniae of the Goodeidae, a family of viviparous cyprinodont fishes. Jour. Morph., 61 : 495. TURNER, C. L., 1938a, The reproductive cycle of Brachyrhaphis episcopi, an ovo- viviparous poeciliid fish, in the natural tropical habitat. Biol. Bull., 75 : 56. TURNER, C. L., 19386. Histological and cytological changes in the ovary of Cyma- togaster aggregatus during gestation. Jour. Morph., 62 : 351. VAN OORDT, G. J., 1928. The duration of the life of the spermatozoa in the fer- tilized female of Xiphophorus helleri Regan. Tijdschr. d. Ned. Dierk. Vereen., 1:1. Reprinted from BIOLOGICAL BULLETIN, Vol. LXXVI, No. 3, 371-383, June, 1939 Printed in U. S. A. THE INFLUENCE OF TEMPERATURE ON THE SURVIVAL, GROWTH AND RESPIRATION OF CALANUS FINMARCHICUS GEORGE L. CLARKE AND DAVID D. BONNET (From the Biological Laboratories, Harvard University, and the Woods Hole Oceanographic Institution'1} Experiments which have been in progress at the Woods Hole Oceano- graphic Institution on the nutrition of the ecologically important copepod, Calanus fimnarchicits, in relation to the food cycle of the sea have shed light on the types of food suitable for this species and on the rate of feeding of which it is capable under different conditions (Clarke and Gellis, 1935; Fuller and Clarke, 1936; and Fuller, 1937). Since the temperatures to which Calanus is subjected off our coasts vary greatly with the season and with the locality, it was desired to extend the investigation of the survival and growth of this species in the laboratory to include a larger range of temperature than had been possible pre- viously. In the earlier experiments survival of the copepods had gen- erally been poor, and Gross (1937) has expressed the opinion that even short intervals of temperature change, such as are entailed by the removal of the culture dishes from the constant temperature bath for inspection, are seriously detrimental to animals in culture. In the experiments about to be described we therefore proposed not only to test the effect of a variety of temperatures on growth but also to compare the survival of animals kept continuously at constant temperatures with that of other individuals subjected to periodic changes in temperature. Calanus was shown to exist in the waters off Woods Hole through- out the year 1935-36 and to undergo reproduction during the spring months (Clarke and Zinn, 1937). Yet food organisms were not suffi- ciently abundant in these waters to meet the nutritive requirements which we have assumed for Calanus even at the maximum rate of feed- ing observed in the laboratory, if diatoms alone are considered (Fuller and Clarke, 1936; and Fuller, 1937). Food in the form of diatoms appeared to exist in only one-tenth the necessary concentration. This discrepancy has led to such suggestions as the considerable use of the nannoplankton by the copepods and this possibility should be investi- gated further, but the difficulty could be removed equally well if the 1 Contribution No. 214. 371 372 GEORGE L. CLARKE AND DAVID D. BONNET value assumed for the food requirements of Calanus were shown to be ten times too high. The value used was that given by Marshall, Nicholls and Orr (1935) and was calculated from the amount of oxygen taken up by a given number of copepods in unit time. The confirmation of this work was therefore of the greatest importance and we proposed in the present undertaking to repeat the measurements of these investigators, using their own method, and to check them by employing a more precise method. Furthermore, we desired to ascer- tain whether differences in temperature exert any significant effect on the oxygen requirement of Calanus and its application to the nutrition of this species. SURVIVAL AND GROWTH Material and Methods For the first group of experiments, which was conducted at Woods Hole during the summer of 1937, copepods were obtained from off- shore localities in the vicinity of Cape Cod. In the second group of experiments, which was undertaken at the Harvard Biological Labora- tories in Cambridge, material was obtained off Gloucester. In all cases the copepods were collected by towing a scrim net for a few minutes, and were placed in jars which were kept cold during the trip to the labora- tory. Immediately upon arrival the animals were transferred by means of a pipette to covered culture dishes and these were set in baths whose temperature was maintained constant to within 0.5° C. In all the ex- periments except those involving a change of temperatures (see below) the copepods were transferred each day by pipette to freshly prepared culture dishes (each containing 250 cc. of -water) which had been brought to the proper temperature. The sea water used in the Woods Hole experiments was taken freshly from the harbor (except for the experiments of Table II) ; that used in Cambridge had been collected at Nahant several months pre- viously and allowed to stand in carboys. In both instances the water was Berkefeld filtered (except for a few tests at Woods Hole in which unfiltered harbor water was used) and then thoroughly aerated. In most of the experiments a known concentration of food was pro- vided by adding a measured amount of a " persistent "• culture of either the diatom, Nitsschia closterium (Plymouth strain) or the green flagel- late, Platymonas subcordifonnis - to each culture dish. The amount of organic, matter represented by these forms and also by the diatom, Rhizosolenia sp., which was abundant in Woods Hole waters during the 2 Kindly identified for us by Professor G. W. Prescott of Albion College, Michigan. EFFECT OF TEMPERATURE ON CALANUS 373 summer, may be judged from the following determinations of nitrogen kindly carried out for us by Dr. T. von Brand : Nitsschia 2.5 yN per 106 cells3 Platymonas 12.7 " " " " Rkisosolenia 96.0 " " " " To ascertain whether the number of animals present or the concen- tration of food organisms was sufficient to affect significantly the dis- solved oxygen or the pH of the culture medium, tests were carried out on a representative series of culture dishes as indicated in Table I. The TABLE I Tests of pH and 62 in culture dishes with different numbers of copepods and of food organisms. Concentration Number of PH O* cc./liter c*-atyA v cells/cc. Calanus Initial After 24 brs. Initial After 24 hrs. 0 0 8.09 7.86 4.25 5.67 0 10 8.09 7.81 4.25 6.42 30,000 0 8.11 9.09* 4.68 5.94 30,000 10 8.11 7.73 4.68 6.65 300,000 0 8.17 8.10 4.81 6.49 300,000 10 8.17 8.00 4.81 6.37 * This value shown by other tests to be atypical. results show that sufficient O2 was present in all containers and that the pH was only slightly altered. The slight increase in O2 content and decrease in pH in the control was due to incomplete temperature equi- librium initially. Experiments The first set of experiments was designed to ascertain whether the change in temperature which results from the removal of the culture dishes from the constant temperature bath for daily inspection and transfer of animals has a deleterious effect on the copepods. For this purpose groups of Calanus and also of Centropages typicus — another copepod common in the Woods Hole region — were divided into two parts and placed in culture dishes in constant temperature baths main- tained at either 8° C. or 13° C. One culture dish in each group was left undisturbed in the bath as a control but others were removed for varying lengths of time as indicated in Table II. For the duration of 3 The reason why this value is larger than that found the year before and reported by Fuller (1937) is not known. Variations of this sort must be taken into consideration in calculating the potential food value of each species. 374 GEORGE L. CLARKE AND DAVID D. BONNET the experiments the survival of Calanus was better than that of Centra- pages but in neither case was survival in the control dishes significantly better than that in the dishes which were subjected to a daily tempera- ture change. We conclude, therefore, that contrary to the suggestion of Gross (1937), the daily inspection at room temperature is not harmful in our experiments with these copepods. Our next set of experiments was designed to test the effect of the concentration of food organisms on the survival and growth of Calanus at temperatures of 7° or 8° C. For this purpose culture media were prepared in which the concentration of food organisms differed by TABLE II Effect of change of temperature on survival. Comparison of number of deaths in cultures removed from constant temperature bath for daily periods of varying lengths with number of deaths in cultures not removed. Two hundred and fifty cc. Berkefeld filtered sea water from tap in each culture dish, left unchanged, and no food added. Experiments carried out in July, 1937 at Woods Hole. Initial no. of animals Temp, of Period culture Temp, of Dura- tion No. Species in each culture dish constant tempera- ture bath removed from bath each day culture when replaced of experi- ment of deaths °C. °C. days Calanus (Adults)* 10 8.5 30 minutes 14.0 8 1 « a 10 8.3 not removed 8 3 Calanus (Stage V)f 15 8.0 60 minutes 16-19 7 0 a a 15 8.0 not removed 7 1 Centropages (Adults)* 20 8.5 30 minutes 12-15 8 17 <> n 22 8.5 120 16-21 8 15 II It 20 8.2 not removed 8 12 li II 20 13.3 30 minutes 15-18 8 18 11 II 20 13.3 120 20-21 8 20 II tl 20 13.0 not removed 8 18 * Animals obtained at Whistle Buoy at western entrance to Vineyard Sound. t Animals obtained at South Channel, east of Nantucket. very large amounts (Table III). Ten to twenty-five Calanus were introduced into each culture dish and their condition noted each day for the duration of the experiments (2 to 4 weeks). The daily transfer of animals to freshly prepared culture media prevented an appreciable change in the concentration of the food from being brought about by either the grazing of the animals or the multiplication of the food organisms. Accumulation of metabolites was similarly avoided. In this set of experiments no copepods were observed to moult successfully but certain individuals reached the moulting stage and died during the casting of the shell. As in previous investigations, the EFFECT OF TEMPERATURE ON CALANUS 375 number of animals which began the moulting process may be taken as an index of growth. The reliability of this procedure is strengthened by the fact that low numbers of " deaths in moult " are usually accom- panied by high numbers of " deaths not in .moult " and vice versa (see Table III). On this basis our results reveal the best nutritional condi- tions in those cases where 30,000 cells/cc. of Nitzschia, or more, were provided, but a concentration as high as a million and a half cells/cc. appeared harmful The moulting stage was attained (i.e. " deaths in TABLE III Effect of concentration of food on survival of Calanus. Temperature of con- stant temperature bath for experiments in July and August: 8° C., in September: 7° C. Each group of animals placed in 250 cc. of culture medium. Animals in copepodid Stage V. Experiments conducted at Woods Hole in 1937.* Food organism Cone, of food Exper. begun Exper. concluded Initial no. of cope pods Deaths in moult Deaths not in moult cellslcc. Nitzschia 300,000 July 11 Aug. 9 15t 8 5 11 150,000 11 ii 15 10 4 ii 30,000 0.2 — till i i i i till lilt 0 10 TEMPERATURE 15 20°C FIG. I. Relation between temperature and amount of oxygen consumed by stage V Calaniu. Dashed lines give the limits of all determinations with the Dixoh-Haldane respirometer. Experiments carried out at Woods Hole during August and September, 1938. grease or vaseline. The respiration flasks were completely immersed in a constant temperature bath controlled to 0.5° C In the experiments which follow 6 to 30 animals were placed in 10 to 20 cc. of water in one of the respiration flasks. This flask had been painted black to exclude light since Marshall, Nicholls, and Orr reported that differences in illumination exerted a marked influence on the rate of respiration. Equilibrium between, the liquid and the gas above it was attained by attaching the entire apparatus to a Warburg shaker. Since continuous or rapid shaking resulted in poor survival of the copepods and might influence abnormally the rate of respiration, the EFFECT OF TEMPERATURE ON CALANUS 379 cn z o o o o shaking was limited to the 5 or 10 minutes immediately before each reading and was confined to an amplitude of 2 cm. at a frequency of 100/min. Readings were ordinarily made every 20 minutes for 4 or more hours. The change in volume for each period was observed and reduced to cc. dry gas at N. T. P. Blank trials, which were run at intervals in exactly the same way but with no animals, showed no variation in volume over periods as long as 4 hours. In certain cases water in which copepods had been living and which presumably contained 0-06 0.05 a 0.04 LU 0.03 0.02 001 0.00 0 TIME IN HOURS FIG. 2. Progress of two typical experiments on consumption of oxygen by stage V Calanus. Closed circles : 6 animals at 16.8° C. Open circles : 28 animals at 13.0° C. a characteristic number of bacteria was used for the blank tests. Thus assurance was provided that in the experiments neither microorganisms nor other extraneous agents were interfering appreciably with the results. Experiments Two measurements of the oxygen consumption of Calanus were made using the Winkler method. The first of these was carried out at 5.5° C. using 25 animals confined in 200 cc. of sea water for 4 hours and yielded a value of 0.35 cc. O,/hr./l,000 copepods. In the second 380 GEORGE L. CLARKE AND DAVID D. BONNET case 75 animals in 200 cc. were tested at 15.5° C. for 4 hours and were found to have consumed oxygen at the rate of 0.98 cc. O,/hr./l,000 copepods. The first of the values is somewhat higher and the second is considerably higher than those obtained by Marshall, Nicholls, and Orr (1935), as may be seen in Fig. 1. The experiments using the Dixon-Haldane respirometer comprised 17 tests carried out at 7 temperatures ranging from 2.5° C. to 16.8° C. In each test from 4 to 12 readings were made during the course of the run and these were plotted against time, as in the two typical experiments shown in Fig. 2. The fact that these curves and all the others which have been used in the calculations exhibited only a slight deviation from a straight line shows that the consumption of oxygen was sensibly con- stant over the period of each experiment. Marshall, Nicholls, and Orr reported a drop m oxygen consumption during the first 10 or 15 hours after the capture of the copepods but in our experiments, since measure- ments were not begun until the day following capture, this period of decreasing respiration was avoided. From the average slope of each curve of oxygen consumption a value was obtained which was converted into cc. CX/hr./l,000 copepods and the results plotted in Fig. 1. It will be observed that a considerable variation exists among the several tests at the same temperatures. Evi- dently respiration in Calanus can vary between wide limits and this con- clusion is borne out by the observations of the earlier investigators. However, there is no doubt as to the general magnitude of the oxygen requirement of Calanus in the present experiments. The influence of temperature on respiration is indicated by the upward slope of the dashed lines in Fig. 1 enclosing all the points. The values increased from an average of about 0.32 cc. (X/hr./l,000 copepods at 5.5° C. to about 0.91 cc. O2/hr./l,000 copepods at 16.8° C. Although our measurements give a somewhat higher average rate of respiration at all temperatures than was obtained by Marshall, Nicholls, and Orr, they are of the same order of magnitude. Furthermore, there is no indication of a possible lower oxygen requirement for the Woods Hole copepods since every one of the present values was higher than those obtained by the British in- vestigators. DISCUSSION The foregoing experiments not only have confirmed and illuminated the early work at Woods Hole, but in addition have raised a number of new questions. It is clear that both the amount of food and the tem- perature have an important influence on the growth of Calanus and on EFFECT OF TEMPERATURE ON CALANUS 381 its survival.5 The situation is further complicated by the fact that under conditions which tend to promote growth, survival may be poor because of the likelihood of death during the moulting process. The provision of food improves growth and survival, but the exact amount of food seems not as important as originally supposed (cf. Clarke, 1939). Al- though a minimum was ordinarily required for growth, in certain ex- periments at low temperatures some moulting was observed in the entire absence of food. And in at least one case a very high concentration of food was found to be harmful (see also Lucas, 1936). It is a striking fact that a large amount of moulting took place and good survival was observed in our experiments conducted in April and May (Table IV) whereas poor survival and no successful moulting was the rule for the tests made during the summer and early autumn (Table III) and for the summer experiments of Fuller (1937, Table V, tem- perature 13° C.).6 In all cases Nitzschia was used as the food organism in at least some of the cultures, the temperatures of the cultures were not widely different, and exactly the same procedure was followed. For the April and May tests copepods were obtained from a point off Gloucester but in two of the other cases the experimental animals were procured from a location which was also north of Cape Cod although considerably to the south of Gloucester.7 If the place of origin of the copepods is immaterial, it might be suggested that the time of year at which the ani- mals were collected could have an important bearing on the laboratory results. According to the observations of Clarke and Zinn (1937) and of Fish (1936) the second of the two yearly broods of Calanus grows to copepodid stage V during the summer and remains in that stage until the following January or February. It seems possible that in nature feeding may be reduced to a minimum and mortality may be especially high during the late summer and that animals collected during this period might exhibit poor survival and refuse to grow even under the most favorable laboratory conditions. Added support for this view is derived from the fact that in our experiment begun September 11 (Table III) no successful moults occurred and no animals were observed even to attempt to moult. 5 In fact Fish (1937) is of the opinion that the presence of suitable tempera- tures is the most important factor in the production of zooplankton in the Gulf of Maine and the Bay of Fundy. 6 In the earlier summer experiments of Clarke and Gellis (1935) and of Fuller and Clarke (1936) growth and survival were better, but the different methods used make comparisons inappropriate. 7 The following ocean temperatures were observed : Off Gloucester, April 26, 1938, 9.6° C. at surface. Cape Cod Bay, August 14, 1937, 18.3° C. at surface, 10.2° C. at 15 m., 8.63 C. at 30 m. South of Cape Cod temperatures were generally higher. 382 GEORGE L. CLARKE AND DAVID D. BONNET In all of our experiments at Woods Hole and especially in those described above the copepods were noticed to be heavily parasitized im- mediately after death. An animal which appeared perfectly healthy one day would be found dead on the next day with its body filled with opaque whitish or colored masses. Possibly parasites gain a foothold within the copepod and remain unnoticed until they suddenly reach some vital tissue. From this initial infection the rapid and obvious spread of the parasite to all parts of the body could be accomplished after death within a relatively short interval. Since Jepps (1937) reported that a large proportion of Calanus population was found to be infected when collected, it is possible that the inroads of parasites may be an important element underlying the fluctuation in the abundance of this species both seasonally and annually. The experiments on respiration show that the values which have been used for the oxygen consumption of Calanus and hence its food require- ment, are of the correct order of magnitude. In fact, our measurements indicate that an even larger amount of food per day is required by these organisms than was assumed by Marshall, Nicholls, and Orr. There- fore, the discrepancy between the nutrient required and the number of diatoms available for food gapes wider than ever. According to Harvey (193X) Calanus can filter out the larger species of diatoms much more effectively than the smaller forms. He found that 100 times the volume of water may be swept free in an hour when Lauderia or Dityliitm were provided as food as when Nitzschia was used. However, in nature the larger diatoms are not as abundant as the more minute species, and a careful investigation of the amount of food per cc. of sea water which the scarcer but more bulky types would provide is seriously needed. Another point which needs study is the " micro-distribution "of the phy- toplankton and nannoplankton. Although uniformity in abundance may be found in a series of one-liter water samples taken from an area, the distribution within each, liter may be highly irregular. If the food or- ganisms occur in clusters, and if a copepod can find and stay in a local though minute zone of abundance, sufficient nutriment could be obtained without the necessity of filtering an excessive volume of water. Such explanations as the foregoing may resolve the discrepancy between the daily food requirement and the average abundance of potential food organisms. SUMMARY Laboratory tests on cultures of Calanus in relation to the ecology of this species showed that: (a) removal of the culture dishes from con- stant low temperature to room temperature for daily periods as great EFFECT OF TEMPERATURE ON CALANUS 383 as 120 minutes was not harmful to the copepods ; (b) growth was poorer at 3° C. than at 6 to 9° C. but survival is better at the lower tempera- tures ; (c) both growth and survival decreased regularly in passing from experiments conducted in the spring to those in the autumn. Measurements of the respiration of Calamis using the Winkler method and the Dixon-Haldane respirometer showed that the magnitude of the oxygen requirement for our animals is of the same order as f or those tested by Marshall, Nicholls, and Orr, and possibly is higher than previously reported. The discrepancy between the estimated food re- quirement of Calamis and the average abundance of diatoms therefore still exists but certain possible explanations are discussed. REFERENCES CLARKE, G. L., 1939. The relation between diatoms and copepods as a factor in the productivity of the sea. Quart. Rev. Biol., in press. CLARKE, G. L., AND S. S. GELLIS, 1935. The nutrition of copepods in relation to the food-cycle of the sea. Biol. Bull, 68: 231-246. CLARKE, G. L., AND D. J. ZINN, 1937. Seasonal production of zooplankton off Woods Hole with special reference to Calanus finmarchicus. Biol. Bull., 73: 464-487. DIXON, MALCOLM, 1934. Manometric Methods. Cambridge University Press. FISH, C. J., 1936. The biology of Calanus finmarchicus in the Gulf of Maine and Bay of Fundy. Biol Bull, 70: 118-141. FISH, C. J., AND M. W. JOHNSON, 1937. The biology of the zooplankton popula- tion in the Bay of Fundy and Gulf of Maine with special reference to production and distribution. Jour. Biol Board of Canada, 3: 189-321. FULLER, J. L., 1937. Feeding rate of Calanus finmarchicus in relation to environ- mental conditions. Biol Bull, 72: 233-246. FULLER, J. L., AND G. L. CLARKE, 1936. Further experiments on the feeding of Calanus finmarchicus. Biol Bull, 70 : 308-320. GROSS, F., 1937. Notes on the culture of some marine plankton organisms. Jour. Mar. Biol Ass'n, 21 : 753-768. HARVEY, H. W., 1937. Note on selective feeding by Calanus. Jour. Mar. Biol Ass'n, 22 : 97-100. JEPPS, M. W., 1937. On the protozoan parasites of Calanus finmarchicus in the Clyde Sea area. Quart. Jour. Micros. Scl, 79 : 589-658. LUCAS, C. E., 1936. On certain inter-relations between phytoplankton and zooplank- ton under experimental conditions. Jour, du Conseil Int. pour I'explor. de la mer, 11 : 343-362. MARSHALL, S. M., A. G. NICHOLLS, AND A. P. ORR, 1935. On the biology of Calanus finmarchicus Part VI. Oxygen consumption in relation to en- vironmental conditions. Jour. Mar. Biol Ass'n, 20: 1-28. DEVELOPMENT OF HALF-EGGS OF CHAETOPTERUS PERGAMENTACEUS WITH SPECIAL REFERENCE TO PARTHENOGENETIC MEROGONY ETHEL BROWNE HARVEY (From the Marine Biological Laboratory, Woods Hole, and the Biological Laboratory, Princeton University) It has been shown in previous studies on sea urchin eggs (1936, 1938) that the non-nucleate fractions; obtained by centrifugal force, may be activated with parthenogenetic agents and may cleave and de- velop, without nuclei, to the blastula stage. This I have termed partheno- genetic merogony. Five species of sea urchins have given similar re- sults. I had hoped that the parthenogenetic merogone of another class of animal might develop further than the blastula, and for this reason have made the present study. The egg of the worm, Chaeiopterus, fulfills the two requirements necessary for parthenogenetic merogony. First, it can be broken apart readily by centrifugal force so that one can obtain non-nucleate fractions in abundance, and second, the normal egg can easily be made to develop parthenogeneiically by treatment with salts. Although my chief concern has been the development of the parthenogenetic merogones, a study has also been made of the fertilized merogones for comparison, as well as of the whole centrifuged egg and the white (nucleate) halves both fertilized and parthenogenetic. Although some of my observations may seem to repeat the earlier ones of Lillie (1906, 1909) and Wilson (1929, 1930), the present study is on fractions of a definite size and known nuclear and cytoplasmic composition whereas previous observations have been made on fragments of irregular size and composition and of assumed nuclear content. Certain, differences in my results as com- pared with previous work may be attributed to this difference in material. Other discrepancies are. due to the £i eater centrifugal force used in these experiments, but particularly to the fact that the eggs have been centrifuged in a medium of cane sugar and sea water. The sugar changes the surface of the egg, and in fact sometimes acts on these eggs as a parthenogenetic agent, similar to KC1. A sugar solution was used as a medium because it can be made, by the addition of the proper amount of sea water, of the same specific gravity as the eggs (as well as, by proper dilution, of the same osmotic pressure), so that the eggs remain suspended during centrifugal ion s and are free to break into two 384 DEVELOPMENT OF HALF-EGGS OF CHAETOPTERUS 385 halves. It also permits the eggs to orient with the lighter area, con- taining the polar spindle, toward the centripetal pole of the centrifuge, whereas in sea water alone the heavy eggs are quickly thrown to the bottom of the centrifuge tube and are held in the position in which they are thrown. Although no special study has been made of the relation of polarity of the egg and the cleavage planes to the stratification and plane of breaking caused by the centrifugal force, there seems no doubt from observations with the centrifuge microscope that the eggs in the sugar solution do orient with the polar area toward the centripetal'axis of the centrifuge. It may be mentioned in passing that the Arbacia egg in the 12-cell stage, when the colorless micromeres appear, also orients with the micromeres toward the centripetal pole of the centri- fuge ; this can be readily observed with the centrifuge microscope. The lack of pigment granules makes these cells lighter than the other cells, which are pigmented. In the Chaetopterus egg, the clear polar region of the egg when ready for fertilization (Photograph 18) is lighter than the granular region and therefore the egg orients as stated above. When the eggs are centrifuged immediately on laying, in the germinal vesicle stage (Photographs, 17, 3), there is no visible lighter area, and there is no apparent orientation. This is, I think, in line with Morgan's studies (1937, 1938); he has definitely determined that the germinal vesicle is not eccentric even in eggs still in the parapodia; it therefore cannot cause orientation. Stratification and Breaking (Photographs 1-16) The Chaetopterus egg stratifies readily. In fact, simply leaving the eggs undisturbed over night in isotonic KC1 produces a very nice strati- fication (Photograph 1). This is true also for the Arbacia egg to a slight extent. The Chaetopterus egg, both unfertilized and fertilized, stratifies also readily with centrifugal force (Photograph 2). It strati- fies similarly to the Arbacia egg except that there is no pigment; the stratified egg is almost exactly like that of another sea urchin, Sphae- r echinus granularis. Some photographs of the stratified eggs of the annelid, Chaetopterus, and of the echinid, Spaer echinus, are very similar and might easily be mistaken the one for the other except that there is a little more oil in the Chaetopterus egg (Cf. Photographs 4 and 5). The unfertilized eggs of Chaetopterus were usually in these experiments centrifuged after the polar spindle had formed; it was found difficult to stratify them well (Photograph 3) and almost impossible, with the forces used, to break them apart in the very short interval elapsing 386 ETHEL BROWNE HARVEY between laying and the breaking of the germinal vesicle. The oil goes to the light pole, then there is a clear layer, then a band of mitochondria which stains purple with methyl green, and then the yellow yolk granules at the heavy pole. (Photographs 4, 6). The nucleus, or rather the first polar spindle, always lies in the clear zo^ie under the oil. The eggs are very heavy and are usually thrown to the bottom of the centrifuge tube in the same sugar solution in which Arbacia eggs are nicely sus- pended. The result is that the egg breaks irregularly ; usually a small oil cap is thrown off and the rest of the egg breaks into irregular frag- ments (Photographs 13, 14 of eggs in the centrifuge microscope slide). This is also the case when the eggs are centrifuged in sea water alone. The earlier experiments of Lillie and Wilson and most of Morgan's experiments were done in this way. It is possible, however, to keep the eggs suspended during centrifu- gation by using a minimum of sea water with the eggs, and a large amount of the isotonic sugar solution. In this way, after centrifuging PLATE I Stratification and Breaking of Eggs The photographs were all taken of living eggs with a Leica camera, and for the most part with a water immersion ( X 40) lens. The magnification on the plates with the exception of Plate II is approximately 175 X ; Plate II is approxi- mately 250 X. The smaller photographs were taken with a low power (X 10) lens and are here magnified about 65 X. The times (i.e., minutes, hours) given are derived from the data recorded at the time of photographing. PHOTOGRAPH 1. Chaetopterus egg kept overnight undisturbed in isotonic KC1. PHOTOGRAPH 2. Egg centrifuged (after fertilization) for comparison with Photograph 1. PHOTOGRAPH 3. Egg centrifuged immediately on laying; shows unbroken germinal vesicle ; this broke immediately afterwards. PHOTOGRAPH 4. Stratified Chactopterns egg ; oil, clear layer, mitochondria and yolk. Polar spindle is in the clear layer under the oil. PHOTOGRAPH 5. Stratified egg of Sphaercchinus granularis for comparison with Photograph 4; same layers. PHOTOGRAPHS 6-9. Breaking apart of stratified egg of Chaetopterus into two halves of uniform size. PHOTOGRAPH 10. White half-eggs, containing oil, clear layer, mitochondria and a little yolk ; also polar spindle. Note uniformity of size. PHOTOGRAPH 11. White half -egg showing band of mitochondria; stained with methyl green. PHOTOGRAPH 12. Yellow half-egg, containing yolk. PHOTOGRAPHS 13-15. Eggs immediately after removal from the centrifuge microscope, and still in the centrifuge microscope slide. PHOTOGRAPHS 13, 14. Eggs supported at bottom of slide. Oil cap pulls off and egg breaks irregularly. PHOTOGRAPH 15. Eggs suspended in sugar solution. Ready to break into halves of uniform size. PHOTOGRAPH 16. A group of eggs just removed from the centrifuge tubes, showing regularity of breaking and uniformity of size of halves. The larger black spheres are the whole eggs viewed from above. DEVELOPMENT OF HALF-EGGS OF CHAETOPTERUS 387 PLATE I _ : ^^ ( » • 10 12 13 14 15 t » 16 Photographs 1-16 388 ETHEL BROWNE HARVEY for five minutes at 10,000 X g., the eggs stratify, become elongate, then dumbbell shape, and then break into spheres, quite regular in size and quite similar to the halves of the Arbacia egg (Photographs 6-9, 16; 15 in the centrifuge microscope slide). The Chactoptcrns egg breaks, so that the lighter or white half contains oil, clear layer, nucleus (or polar spindle), mitochondria and a little yolk (Photographs 9, 10, 11). The heavier yellow half contains only yolk (Photographs 9, 12). The measurements for the whole and half eggs are : Whole egg Diameter 94 ^ Volume 434900 n 3 White half " 83 p. " 299400 M Yellow half 64 p. 137300 p. The white half has over twice the volume of the yellow half. Many batches of eggs have given similar figures. Normal Development (Photographs 17-31) As is well known, especially from the early work of Wilson (1883) and Mead (1895, 1897, 1898 a, b), the early development of the egg is as follows. The egg is laid in the germinal vesicle stage (Photograph 17) ; after resting in sea water a few minutes, the germinal vesicle PLATE II Normal Development of the Chactoptcrus Egg PHOTOGRAPH 17. Immature egg immediately on laying; germinal vesicle still intact in center. PHOTOGRAPH 18. Egg 15 minutes after laying ; polar spindle in white area at upper pole. PHOTOGRAPH 19. Fertilized egg, showing crinkled or fluted fertilization mem- brane. Twenty-five minutes after fertilization. PHOTOGRAPH 20. Egg flattened where first polar body is being extruded. Fifteen minutes after fertilization. PHOTOGRAPH 21. Egg rounded again after first polar body has been extruded; note fluted membrane near polar body. Twenty-five minutes after fertilization. PHOTOGRAPH 22. Egg flattened again at time of formation of second polar body. Thirty minutes after fertilization. PHOTOGRAPH 23. Same egg 1 minute later ; it has again rounded out. PHOTOGRAPH 24. Pear-shaped stage. Thirty-four minutes after fertilization. PHOTOGRAPH 25. Same egg 2 minutes later ; beginning of polar lobe. PHOTOGRAPH 26. Same egg 4 minutes later ; 40 minutes after fertilization. First cleavage beginning, with polar lobe still present. PHOTOGRAPH 27. Same egg 20 minutes later ; first cleavage complete and polar lobe withdrawn ; note polar bodies near furrow at upper pole. PHOTOGRAPH 28. Four-cell stage, 1 hour, 40 minutes after fertilization. Note characteristic Brechungslinie or cross-furrow. PHOTOGRAPH 29. Eight-cell stage, 2 hours after fertilization. PHOTOGRAPH 30. Blastula, just before swimming; 4% hours after fertiliza- tion. PHOTOGRAPH 31. Trochophore, 1 day after fertilization. PLATE II 29 Photographs 17-31 390 ETHEL BROWNE HARVEY breaks and the first polar spindle forms. The polar region is seen in the living egg as a clear area near the surface; it is fully formed (at 23° C.) 10-20 minutes after laying (Photograph 18). The egg remains in this condition, with the first polar spindle in the metaphase, indefinitely unless fertilized. If fertilized at this stage, in about 15 minutes1 the first polar body appears (Photograph 20). The fertilization membrane is the same, according to Whitaker (1933), as the vitelline membrane which lifts from the surface. It is peculiar and characteristic for this egg and quite different from the fertilization membrane of sea urchin eggs. It is crinkled or fluted especially in the region of the polar bodies (Photographs 19, 21). The egg becomes markedly flattened at the pole at the formation of the first polar body (Photograph 20) ; then becomes spherical (Photograph 21). Then the second polar body is formed, again with a slight flattening (Photograph 22) — about 30 minutes after fertilization; and the egg once again rounds out (Photo- graph 23). Soon after this (35 minutes after fertilization) the egg becomes pointed or pear-shaped with the narrow end toward the polar bodies (Photograph 24). After again rounding out, the characteristic polar lobe is formed (about 5 minutes after the pear-shape) at the op- posite end (Photograph 25). This remains till after first cleavage which begins about 40 minutes after fertilization (Photograph 26). The first cleavage is very unequal giving a small cell and a large one into which the yolk lobe is absorbed (Photograph 27). At the second cleavage, l^o hours after fertilization, the small cell divides equally and the large cell unequally so that there are two small cells, one large cell and one intermediate in size (Photograph 28), a form of cleavage quite typical for many annelid and mollusk eggs. The next cleavage is shown in Photograph 29, taken 2 hours after fertilization. The spiral type of cleavage which occurs in Chactoptcrus is difficult to follow in the living egg, so that no attempt has been made to photograph the later cleavage stages of the normal egg, nor to follow them in the experimental work. A normal blastula is shown in Photograph 30, taken just before swim- ming, which occurs about 5 hours after fertilization ; and a one-day trochophore, in Photograph 31. 1 The times for the various stages are quite variable in the Chactoptcrns egg even with constant temperature. It differs in this respect from the Arbacia egg in which, in the height of the season, the stages occur with almost clock-like regularity at any given temperature. Late in the season (September, October) there is considerable delay in the Arbacia egg irrespective of temperature. Eggs may be obtained through November from Arbacias which have been brought in before the first part of August and kept in the aquaria in running water. Animals brought in from the sea after August 15 have practically all shed their eggs. DEVELOPMENT OF HALF-EGGS OF CHAETOPTERUS 391 Normal Egg, Parthenogenetic The parthenogenetic development of the normal uncentrifuged egg by the addition of KC1 to the sea water has been described by Mead (1898 b), Loeb (1901) and Lillie (1902, 1906), and in detail by Allyn (1912). The peculiar feature is the amoeboid form of the egg and the frequent occurrence of very large nuclei without cell division. The later embryos tend to stick together and form complex multiple organisms as described particularly by Loeb (1901). This is probably due to the effect of the KC1 solution on the surface of the cells. Unequal two- and four-cell stages of the characteristic annelid type also occur, and the parthenogenetic egg may give rise to apparently normal swimmers just like the fertilized egg. Some equal two-cell stages also occur. My best results for parthenogenesis have been obtained by putting the eggs into a solution of 1 gram KC1 + 100 cc. of sea water for 30 minutes, then transferring them to sea water. However, I have found other proportions and times of exposure to work better with some batches of eggs. The different batches respond quite differently to the KC1 solutions and I could get no dependable standard solution. U'liolc Egg Ccntrifngcd, Then Fertilized (Photographs 32-42) The centrifugcd whole egg, which has been elongated by the cen- trifugal force, becomes spherical again quite quickly after the force is removed, and this whether the egg is fertilized or not. It is thus strik- ingly different from the sea urchin egg which becomes spherical if left unfertilized, but retains its elongate or dumb-bell shape if fertilized or artificially activated, all through the cleavages up to the swimming blastula. After fertilization of the centrifuged egg, in Chaetopterus, the polar bodies usually come off near, but not on top of, the oil cap (Photograph 32) ; if the oil cap has been centrifuged off, however, they usually come off exactly at the centripetal pole (Photograph 33) ; the oil cap apparently interferes with the protrusion of the polar bodies. This is shown also by the results of centrifuging the egg immediately on laying, while the germinal vesicle is still intact, as shown in Photo- graph 3. When, after removal from the centrifuge, the germinal vesicle breaks, it dissipates the oil in clumps (Photograph 34). The polar spindle forces itself to the center of the oil, which now forms a circle of clumps around it (Photograph 35). After fertilization, the polar bodies come off in the center of the ring of oil clumps. The fertilized centrifuged egg usually passes through the changes in shape characteristic of the normal fertilized egg as described above, in- 392 ETHEL BROWNE HARVEY eluding the formation of the polar lobe. It often divides unequally and typically in the first cleavage (Photograph 36), and the second cleavage as well as later ones (Photograph 39) may he of the charac- teristic annelid type as has been described by Lillie (1906) and Wilson (1929). Some of the cleavages, however, especially in strongly centri- fuged eggs, are irregular and atypical. The first division plane is usually perpendicular to the stratification (Photograph 36). It may, however, pass parallel with the stratification, separating the egg into a clear and a granular zone (Photograph 37). These zones may develop separately just as in the centrifuged Arbacia egg (Harvey, 1932) ; the upper clear area usually divides into cells but the lower yolk half is characterized by nuclear without cytoplasmic divisions (Photograph 38). Many swimming, apparently normal, larvae have arisen from the fertilized centrifuged eggs (Photograph 40). But the blastulae tend to gather together in groups and fuse with each other and with the white halves, PLATE III PHOTOGRAPHS 32-42. Eggs cciilrifui/ed, then fertilized. Photographs 43-46. Eggs centrifuged then treated u'itli KCI ; partheno genetic. PHOTOGRAPH 32. Polar bodies at side of oil cap; 23 minutes after fertilization. PHOTOGRAPH 33. Polar bodies at centripetal pole where oil cap has been centrifuged off. Twenty-five minutes after fertilization. PHOTOGRAPH 34. Egg was centrifuged immediately on laying (see Photo- graph 3) ; germinal vesicle breaks after removal from centrifuge and scatters oil in clumps. PHOTOGRAPH 35. On formation of polar spindle a few minutes later, the clumps of oil are forced into a ring. PHOTOGRAPH 36. Normal cleavage of centrifuged egg; cleavage plane at right angles to stratification. Note polar lobe. Fifty minutes after fertilization. PHOTOGRAPH 37. Cleavage plane parallel with stratification. One and one- quarter hours after fertilization. PHOTOGRAPH 38. Later stage of Photograph 37 ; upper cells have cleaved several times, lower cell has not cleaved, but the nuclei have divided. One and three-quarter hours after fertilization. PHOTOGRAPH 39. Normal development of centrifuged egg. Two and one-half hours after fertilization. PHOTOGRAPH 40. Same, four and three-quarter hours after fertilization; just before swimming. PHOTOGRAPH 41. Fusion of whole centrifuged and white half-blastulae. Seven hours after fertilization. PHOTOGRAPH 42. A pair of these, high power. PHOTOGRAPH 43. Two-cell stage of parthenogenetic centrifuged egg; typical unequal blastomeres ; cleavage plane perpendicular to stratification. Two hours after treatment. PHOTOGRAPH 44. Same, equal cells. Note polar lobe. PHOTOGRAPH 45. Four-cell stage, parthenogenetic. Three hours after treat- ment. Compare with Photograph 74, of the yellow half-egg fertilized. PHOTOGRAPH 46. Amoeboid form of parthenogenetic centrifuged egg, 5 hours after treatment. Compare with parthenogenetic white half, Photograph 58; with fertilized yellow half, Photograph 75 ; and with parthenogenetic merogone, Photo- graph 96. DEVELOPMENT OF HALF-EGGS OF CHAETOPTERUS 393 PLATE III i 37 38 39 • 35 40 42 46 Photographs 32-46 394 ETHEL BROWNE HARVEY forming complex organisms (Photographs 41, 42) just as is charac- teristic of the parthenogenetic normal eggs. This agglutination must be due to the effect of the sugar solution (in which the eggs have been centrifuged) on the surface of the eggs. It cannot be due entirely to the lack of a fertilization membrane, since eggs from which fertilization membranes have been removed by shaking, and which were kept in sea water, without sugar, do not agglutinate. Centrifuged Egg, Pai'tJicnogenctic (Photographs 43-46) The centrifuged egg will develop parthenogenetically by the addition of KC1 to the sea water, in exactly the same way as the uncentrifuged. Polar bodies may be given off often while still in the KC1 solution, and there occur some typical normal cleavages (Photograph 43). But there is a tendency toward equal first cleavage (Photograph 44) and most of the eggs soon become amoeboid and develop with large nuclei and few cell divisions like the uncentrifuged parthenogenetic eggs (Pho- tographs 45, 46). The blastulae tend to fuse together just as do the parthenogenetic normal eggs and the fertilized centrifuged eggs. De- velopment is slower in the parthenogenetic eggs, and swimmers occur in 8 hours instead of 5. This has been found true also of parthenogenetic normal eggs of Cliactoptcnts as well as of many other eggs. iriiitc Half-egg, Fertilized (Photographs 47-55) The white half-egg, which contains the polar spindle, can be ferti- lized, and may develop exactly like the normal whole egg. The egg passes through the flattened and pear-shape phases and the polar bodies are usually given off near the oil cap ; these are usually larger than in the normal jegg (Photograph 47). The egg usually divides unequally in the first cleavage (Photograph 48, lower egg) and the large cell unequally in the second cleavage giving the typical four-cell stage of two small cells, one large and one intermediate in size (Photograph 49, lower egg). The nuclei are especially striking owing to the lack of granules ; these obscure the nuclei in the living whole eggs. There is a tendency for the egg to become amoeboid even before the first division, so that it is not possible to determine if there is a polar lobe or not (Photograph 50). There are often protuberances from one of the blastomeres resembling polar lobes and these occur also in the four-cell stage (Photograph 51). There is also a tendency toward equal first cleavage. Normal blastulae and swimming trochophores are often DEVELOPMENT OF HALF-EGGS OF CHAETOPTERUS 395 formed exactly like those of the whole egg, except that they are white instead of yellow and are smaller (Photograph 52, inset, upper left). But frequently, the hlastulae have irregular sized cells and large nuclei like the parthenogenetic whole eggs and they tend to become amoeboid (Photograph 52). The blastula shown in Photograph 53 was both amoeboid and ciliate. These white blastulae often stick together in clumps and fuse with each other and with the whole blastulae present in the same lot (Photographs 54, 55, 41, 42). As pointed out above, this is probably due to the effect of the sugar solution on the surface. The development of the fertilized white halves takes place at about the same rate as the whole eggs ; swimmers occur in 5 hours. The striking thing is that the white half-eggs often cleave, at least in the early stages, exactly like the whole eggs in the very peculiar and characteristic pattern typical of the annelid egg. This has been found also to be the case in many irregular fragments studied by Lillie (1909) and Wilson (1929). White Half-egg, Parthenogenetic (Photographs 56-62) The white half-eggs may be activated by KC1 similarly to the whole eggs ; some have been activated and have produced swimmers without further treatment than the sugar solution in which they were centri- fuged. They may throw off polar bodies often while still in the KC1 solution (Photograph 56) and may cleave in the typical annelid fashion (Photograph 57) , at least to the four-cell stage. But they, like the whole parthenogenetic eggs, often divide equally at first cleavage, and they usually become amoeboid and develop with irregular cells and large nuclei (Photographs 58-62) ; these irregularities are much more pro- nounced in the parthenogenetic than in the fertilized white halves. Development is also delayed as is usual with parthenogenetic eggs. Yellow Half -egg. Fertilized (Fertilized Merogone) (Photographs 63-84) The yellow half-egg, which contains no nucleus, may be fertilized and the fertilization membrane is lifted off. I have observed no polar bodies of which I could be sure, though there are often clear amoeboid protrusions simulating polar bodies. One would not, of course, expect polar bodies in this half-egg, since the polar spindle is always in the other half. There is also no polar lobe of which one can be sure. There is usually no cleavage of this half-egg, but it becomes amoeboid and may live in this condition for several days. In some eggs, I have 396 ETHEL BROWNE HARVEY found, several hours after fertilization, a nucleus, from the sperm, and later two nuclei, and much later several nuclei of different sizes — that is, nuclear division without cell division (Photographs 63-65). such as is characteristic of the heavy red halves of the Arbacia egg (Harvey. 1932). However, several eggs have given a typical first cleavage of two unequal cells (Photographs 66, 67). Several eggs have divided into two equal cells, and in one case these divided again equally giving four equal cells — not typical for this form (Photographs 68, 69). These four cells, later on, fused, cleaved again, hecame amoehoid and soon went to pieces (Photographs 70-75). In general a failure to cleave and a tendency to become amoeboid are characteristic of the fertilized yellow halves or fertilized merogones. Xo blastulae or cili- ated structures from the fertilized yellow halves have been observed. A series of photographs (76-84) will show the extreme amoeboid ac- tivity of one of these eggs in the two-cell stage. These photographs PLATE IV U'liitc halj-cggs, fertilized (47-55) and parthenogenetic (56-62) PHOTOGRAPH 47. White half fertilized, with polar bodies. Twenty-five min- utes after fertilization. PHOTOGRAPH 48 (lower egg). Typical two-cell stage. One and one-quarter hours after fertilization. PHOTOGRAPH 49 (lower egg). Typical four-cell stage. One and one-half hours after fertilization. Note Brechungslinie or cross-furrow as in Photograph 28. In upper egg, all the oil is in one blastomere, and the blastomeres are not quite typical in size. PHOTOGRAPH 50. Just before first cleavage, amoeboid at position for polar lobe. Fifty minutes after fertilization. PHOTOGRAPH 51. Four-cell, amoeboid. One hour after fertilization. PHOTOGRAPH 52. Amoeboid blastulae with large irregular nuclei, and also normal white blastula (inset, upper left). At the upper right is a normal whole blastula (centrifuged). Five hours after fertilization. PHOTOGRAPH 53. White blastula, amoeboid and also ciliated. Six and one-half hours after fertilization. PHOTOGRAPH 54. White blastulae fused together. See also Photograph 41. Six hours after fertilization. PHOTOGRAPH 55. A pair of fused blastulae, high power. See also Photo- graph 42. PHOTOGRAPH 56. Parthenogenetic white half with two polar bodies. One hour after activation. PHOTOGRAPH 57. Typical two-cell stage of parthenogenetic white half. Two hours after activation. PHOTOGRAPH 58. Parthenogenetic white halves, amoeboid. One and one-half hours after activation. Cf. with Photograph 46 of whole centrifuged egg, par- thenogenetic. PHOTOGRAPHS 59, 60. Parthenogenetic white half blastulae, amoeboid, with irregular nuclei. Four hours after activation. Cf. with Photograph 52, fertilized white halves. PHOTOGRAPHS 61, 62. Blastulae both ciliate and amoeboid. Twenty-four hours after activation. Cf. Photograph 62 with Photograph 53, fertilized white half. DEVELOPMENT OF HALF-EGGS OF CHAETOPTERUS 397 PLATE IV 56 49 ^^? - 57 59 61 Photographs 47-62 398 ETHEL BROWNE HARVEY were taken at short intervals ; after two hours the egg ended up in a two-cell stage, much as it was in the beginning. The meaning of this is not known, but perhaps it points to a similar condition, with regard to the cell surface, for amoeboid motion and cell division. Yellow Half-egg, Partheno genetic {Parthenogenetic j\Icrogonc) ( Photographs 85-96) The yellow half -eggs, containing no nucleus or polar spindle, may be activated artificially by KC1. These are the parthenogenetic mero- gones. After activation, a thick fluted membrane is lifted off exactly like that of the normal fertilized egg, and quite characteristic of this PLATE V 3Y//tw half -eggs, fertilized. (Fertilized incroi/oiies) PHOTOGRAPH 63. Sperm nucleus in yellow half-egg. Two hours after fertilization. PHOTOGRAPH 64. Same egg with two nuclei, a few minutes later. PHOTOGRAPH 65. Another yellow half-egg with several nuclei but no cell division. Five hours after fertilization. PHOTOGRAPH 66. Twro-cell stage ; cells unequal. PHOTOGRAPH 67. Typical two-cell stage. Two hours after fertilization. PHOTOGRAPHS 68-75. Successive photographs of the same egg showing cleav- age and fusion of blastomeres, and amoeboid character. PHOTOGRAPH 68. Two equal cells. Two hours after fertilization. PHOTOGRAPH 69. Five minutes later, 4 equal cells. PHOTOGRAPH 70. Ten minutes later. Blastomeres begin to fuse. PHOTOGRAPH 71. Three minutes later. Possibly polar lobe. PHOTOGRAPH 72. One minute later. Three cells; two have fused. PHOTOGRAPH 73. Six minutes later. Four (or 5) separate cells again. PHOTOGRAPH 74. Five minutes later. Fusion of blastomeres, and egg be- coming amoeboid. Compare with Photograph 45, of the whole centrifuged egg, parthenogenetic. PHOTOGRAPH 75. One-half hour later, 3 hours after fertilization. Amoeboid. Compare with Photograph 46, of whole centrifuged egg, parthenogenetic; and with Photograph 58 of parthenogenetic white half and with Photograph 96 of parthenogenetic merogone. PHOTOGRAPHS 76-84. Successive photographs at short intervals showing amoelxiid character of a two-cell stage of the fertilized yellow half -egg. PHOTOGRAPH 76. Two-cell stage, almost equal blastomeres. Fertilized at 9.30 A.M.. photographed at 10.50 A.M. PHOTOGRAPH 77, at 11.35 A.M. PHOTOGRAPH 78, at 11.40 A.M. PHOTOGRAPH 79, at 11.50 A.M. PHOTOGRAPH 80, at 12 noon. PHOTOGRAPH 81, at 12.10 P.M. PHOTOGRAPH 82, at 12.20 P.M. PHOTOGRAPH 83, at 12.21 P.M., two-cell stage, slightly unequal. PHOTOGRAPH 84, at 1.10 P.M., two cells almost equal and quite similar to the original. Photograph 76. Note large nucleus. DEVELOPMENT OF HALF-EGGS OF CHAETOPTERUS 399 PLATE V 63 68 72 76 79 82 69 73 65 66 67 70 71 75 77 78 ,'<-/ ^ 80 81 83 84 Photographs 63-84 400 ETHEL BROWNE HARVEY particular egg (Photograph 85, cf. Photographs 19. 21). They then usually become amoeboid and develop no further, though they may live for several clays. There is no definite polar lobe, but amoeboid proc- esses sometimes simulate it. Some cases of first cleavage have been ob- served, however, and these were of both the equal type and the unequal type with varying degrees of inequality (Photographs 86-90) ; these also became amoeboid and there was no further development (Photographs 91-93). The amoeboid activity of the parthenogenetic merogone is shown in a series of photographs taken at intervals of a few minutes (Photographs 9-4—96). The parthenogenetic merogone, then, resembles the fertilized merogone in its tendency to become amoeboid and in its limited cleavage, though the fertilized merogone goes perhaps one step further in development. (Compare Photographs 86-96 with Photo- graphs 66-84). Discussion A study of the CJiactoptcnts egg has shown that the non-nucleate fraction of the egg of another type of animal than the sea urchin can be activated artificially. The parthenogenetic merogone of Chactoptcrus PLATE VI Yellow half-eggs, parthenogenetic. (Parthenogenetic mcrogones) PHOTOGRAPH 85. Yellow half, parthenogenetic, with characteristic fluted fer- tilization membrane. One hour after activation. At right upper corner is part of a whole cell, also with fluted membrane. Compare also with Photograph 19, of the normal whole egg. PHOTOGRAPH 86. Two-cell stage, unequal. Two hours after activation. Com- pare with Photograph 66, of fertilized merogone. PHOTOGRAPH 87. Typical two-cell stage. Two hours after activation. Com- pare with Photograph 67, of fertilized merogone. PHOTOGRAPH 88. Cleaving into two almost equal cells. Five hours after activation. PHOTOGRAPH 89. Two equal cells. Six hours after activation. Compare with Photograph 68 of fertilized merogone. PHOTOGRAPH 90. Two equal cells. Five hours after activation. Compare with Photograph 76. Clear protrusion looks like a polar body. PHOTOGRAPHS 91, 92, 93. Amoeboid forms of older parthenogenetic merogones. Twenty-four hours after activation. Compare with Photograph 46 of whole centrifuged egg, parthenogenetic, and with Photograph 75, of fertilized merogone. PHOTOGRAPHS 94-96. Successive photographs of the same parthenogenetic yellow half-egg, taken at very short intervals to show amoeboid character. Eight- een hours after activation. Compare with Photographs 76-84 of the fertilized merogone. Also compare Photograph 96 with Photograph 75 of the fertilized merogone, and with Photograph 46 of the parthenogenetic whole egg, and with Photograph 58 of the parthenogenetic white half. PHOTOGRAPH 94. Activated at 5 P.M. July 21, photographed at 11 A.M.. July 22. PHOTOGRAPH 95. At 11.02 A.M. PHOTOGRAPH 96. At 11.03 A.M. DEVELOPMENT OF HALF-EGGS OF CHAETOPTERUS 401 PLATE VI \ ' 85 87 88 89 90 91 92 93 94 95 96 Photographs 85-96 402 ETHEL BROWNE HARVEY goes through only one cleavage and thus does not develop as far as that of the sea urchin. This is not due, however, to the lack of a nucleus, since the same fraction fertilized does not develop much further. The unfertilized Chaetopterus egg is perhaps more highly organized than the unfertilized sea urchin egg. The parthenogenetic development of the normal Chaetopterus egg is peculiar in that usually there is no orderly cleavage with the accompany- ing nuclear divisions, but instead, amoeboid forms with irregular cells and large nuclei. It is not surprising, therefore, that there is no orclerl\ development of the parthenogenetic merogone, and that it likewise, be- comes amoeboid. It may be that the concentration of heavy granules into this half of the egg is a factor also in preventing further develop- ment. This is suggested by the experiments of Whitaker and Morgan (1930), who obtained polar and anti-polar halves of the Chaetopterus egg by cutting with a needle. These halves have, of course, a mure even distribution of granules than the halves obtained by centrifuging. They found that after fertilization, both of the halves would cleave un- equally like the whole egg, and that the anti-polar half would form a polar lobe ; there were certain irregularities, however. But the anti- polar half obtained by cutting develops more normally, at least in the early stages, than the anti-polar half obtained by centrifuging. It is of interest, too, that the parthenogenetic merogone of the Sphaer echinus (jranidaris egg, whose stratification is so similar to that of Chaetopterus (Photographs 4 and 5) also becomes amoeboid and does not cleave regularly; the parthenogenetic centrifuged whole egg of this form like- wise becomes amoeboid (see Photographs 53-56 of my 1938 paper). The amoeboid activity which is characteristic of the parthenogenetic egg or its fraction both in Chaetopterus and Sphaer echinus, may be largely due to the action of the parthenogenetic agent on the surface of the egg. Amoeboid activity may be another expression of the condition of the cell surface characteristic of cleavage just as gelation seems to be characteristic of some of the interior cytoplasm at this time, in the formation of asters. That the early development is dependent on the cytoplasm rather than the nucleus is brought out again in the development of the white halves of the Chaetopterus egg. These, whether fertilized or partheno- genetic, may cleave according to the peculiar and characteristic annelid pattern. It seems remarkable that these fractions, and even very much smaller fragments, according to Lillie and Wilson, may cleave just like the whole egg. This certainly leads one to the conclusion that it is the matrix or clear substance in the egg, rather than nuclei or visible granules that is important in early development. DEVELOPMENT OF HALF-EGGS OF CHAETOPTERUS 403 Attention may be called to the frequent occurrence of equal cells in the first cleavage instead of the typical unequal cells. In the present ex- periments, there is a tendency to equal division in parthenogenetic whole eggs, both normal and centrifuged and both in the white and yellow halves both fertilized and parthenogenetic. Other observers have noted the same thing in certain of these cases. In the Cwningia egg, I found that pressure on the egg after fertilization had the same effect (Browne, 1910) and similar results have been obtained by Tyler (1930) for the Chaetopterus egg; he found also that cold, heat, ultra-violet rays and anaerobiosis would produce the same effect. One would assume that an unequal division of the two cells is a more specialized form, and that whatever mechanism is responsible for it, is put out of gear by certain experimental conditions. It seems to be due, at least in many cases, to an action especially on the surface of the cell, rather than on the interior. Summary 1. The unfertilized Chaetopterus egg (94^) may be stratified and broken into unequal halves of a uniform size by centrifugal force. The nucleate white halves (83 p,) contain oil, clear layer, mitochondria and a little yolk. The non-nucleate yellow halves (64 fji) contain only yolk. 2. The centrifuged whole egg may develop, both fertilized and parthenogenetic, similarly to the normal uncentrifuged egg. Certain peculiarities have been noted. 3. The white halves, both fertilized and parthenogenetic, may cleave according to the typical annelid pattern. But there is a tendency toward equal first cleavage and amoeboid form and later, development with irregular cells and large nuclei ; the blastulae tend to fuse together ; these peculiarities are much more pronounced in the parthenogenetic than in the fertilized white halves. 4. The yellow halves, both fertilized (= fertilized merogone) and parthenogenetic (= parthenogenetic merogone) usually lift off a char- acteristic fluted fertilization membrane and pass through only one cleav- age and become markedly amoeboid. 5. Early development without nuclei (parthenogenetic merogony) is established for the annelid, Chaetopterus, in addition to the five species of echinoderms previously studied. The development of the partheno- genetic merogones of Chaetopterus does not go as far as that of the sea urchin. 404 ETHEL BROWNE HARVEY LITERATURE CITED ALLYN, H. M., 1912. The initiation of development in Chaetopterus. Biol. Bull, 24: 21. BROWNE, E. N., 1910. Effects of pressure on Cumingia eggs. Arch. f. Entzv.- mech., 29 : 243. HARVEY, E. B., 1932. The development of half and quarter eggs of Arbacia punctulata and of strongly centrifuged whole eggs. Biol. Bull., 62 : 155. HARVEY, E. B., 1936. Parthenogenetic merogony or cleavage without nuclei in Arbacia punctulata. Ibid., 71 : 101. HARVEY, E. B., 1938. Parthenogenetic merogony or development without nuclei of the eggs of sea urchins from Naples. Ibid., 75 : 170. LILLIE, F. R., 1902. Differentiation without cleavage in the egg of the annelid Chaetopterus pergamentaceus. Arch. f. Entw.-mech., 14: 477. LILLIE, F. R., 1906. Observations and experiments concerning the elementary phenomena of embryonic development in Chaetopterus. Jour. E.rper. Zool, 3: 153. LILLIE, F. R., 1909. Polarity and bilaterality of the annelid egg. Experiments with centrifugal force. Biol. Bull., 16 : 54. LOEB, J., 1901. Experiments on artificial parthenogenesis in annelids (Chaetop- terus) and the nature of the process of fertilization. Am. Jour. Physio!., 4 : 423. MEAD, A. D., 1895. Some observations on maturation and fecundation in Chaetop- terus pergamentaceus, Cuvier. Jour. Morph., 10 : 313. MEAD, A. D., 1897. The early development of marine annelids. Ibid., 13 : 227. MEAD, A. D., 18980. The origin and behavior of the centrosomes in the annelid egg. Ibid., 14: 181. MEAD, A. D., 1898&. The rate of cell division and the function of the centrosome. Biological Lectures delivered at the M. B. L. of Woods Hole, 1896-1897. Ginn & Co., Boston. MORGAN, T. H., 1937. The factors locating the first cleavage plane in the egg of Chaetopterus. Cytologia, Fujii Jubilee Volume, p. 711. MORGAN, T. H., 1938. A reconsideration of the evidence concerning a dorsoventral pre-organization of the egg of Chaetopterus. Biol. Bull., 74 : 395. TYLER, A., 1930. Experimental production of double embryos in annelids and mol- lusks. Jour. Exper. Zool., 57 : 347. WHITAKER, D. M., 1933. On the rate of oxygen consumption by fertilized and unfertilized eggs. IV. Chaetopterus and Arbacia punctulata. Jour. Gen. Physiol, 16 : 475. WHITAKER, D. M. AND T. H. MORGAN, 1930. The cleavage of polar and anti-polar halves of the egg of Chaetopterus. Biol. Bull., 58 : 145. WILSON, E. B., 1883. Observations of the early developmental stages of some polychaetous annelides. Studies from the Biol. Lab., J. H. U., 2: 271. WILSON, E. B., 1929. The development of egg-fragments in annelids. Arch. /. Entw.-mcch., 117: 179. WILSON, E. B., 1930. Notes on the development of fragments of the fertilized Chaetopterus egg. Biol. Bull., 59 : 71. CHEMICAL MEDIATION IN CRUSTACEANS. III. ACETYL- CHOLINE AND AUTOTOMY IN PETROLISTHES ARMATUS (GIBBES) JOHN H. WELSH AND HAROLD H. HASKIN (From the Bermuda Biological Station for Research, Inc., and the Biological Laboratories, Harvard University) Acetylcholine has been shown to be present in nervous tissues of crustaceans in considerable amounts (Welsh, 1938; Smith, 1939) and to have an excitatory action on the decapod heart (Welsh, 1939 a and b), but, thus far, its action elsewhere in crustaceans has not been demon- strated (see Katz, 1936, regarding acetylcholine and crustacean skeletal muscle). Autotomy, or the casting of legs by crustaceans, is a well-known phenomenon resulting from a unisegmental, reflex stimulation of the autotomizer muscle of a leg (Wood and Wood, 1932). Hence central transmission at synapses, peripheral transmission at myoneural junc- tions, and conduction along fibers of the reflex pathway must occur in this process. If acetylcholine is normally involved in the transmission of nerve impulses in the autotomy reflex, it should be possible to obtain evidence by appropriate injections of acetylcholine and of drugs which are known to affect its rate of destruction and its action. That stimu- lating and inhibiting substances may play a part in autotomy in Porcel- lana and Uca was suggested by Hoadley (1934, 1937). Abramowitz and Abramowitz (1938), while testing a series of drugs for their effects on the chromatophores of Uca, observed that the injection of acetyl- choline caused autotomy. The present paper records observations made on autotomy in an anomuran of Bermuda. The experimental work was done at the Ber- muda Biological Station and was aided by grants from the Milton and Porter Funds of Harvard University. Materials and Methods The animal used was Petrolisthes armatus (Gibbes), of the family Porcellanidae. The members of this family are said to autotomize with the greatest ease and after the least stimulus of any of the decapod crustaceans (Wood and Wood, 1932). P. armatus occurs in consid- 405 406 J. H. WELSH AND H. H. HASKIN erable abundance under stones between tide levels in Bermuda. They were captured and handled with care in order to prevent autotomy, and experiments were performed, whenever possible, on animals which had been in the laboratory less than twenty-four hours. Since Petrolisthes may be caused to autotomize a leg by grasping it gently with forceps, this was the method of stimulation employed. Individuals were placed in small containers rilled with sea water. Each leg was seized, held for approximately one second, and, if not dropped, was then released. Petrolisthes has three pairs of legs which are de- veloped for walking, and a pair of chelae, so a series of eight stimuli could be given the animal without repeated stimulation of those append- ages which failed to autotomize on the first trial. Legs were grasped in random order and the order was varied from animal to animal. After obtaining results on one hundred normal animals to serve as controls, fifty animals wrere injected at the base of one of the last walking legs with 0.05 cc. of a perfusion fluid prepared according to Pantin (1934). As the average body volume of the animals used was only 0.3 cc., this amount of injected fluid caused a considerable dilution of the blood, but this dilution and the act of injection had no evident effect on autotomy. The drugs employed were always given in the volume of fluid used in the injected controls. These drugs were acetylcholine chlo- ride (Hoffman-LaRoche), eserine (physostigmine) sulphate (Merck), atropine sulphate (Merck), and adrenalin chloride (Parke-Davis). Fresh solutions of these substances were prepared daily and the neces- sary dilutions were made just before using. This is especially important in the case of acetylcholine which is rapidly hydrolyzed in an alkaline solution. Results Control Experiments. — The results obtained on uninjected control animals are given in Fig. 1. The first leg l stimulated by grasping was dropped by 54 of the 100 animals, or it may be said that there was 54 per cent autotomy. The second, third and fourth legs stimulated showed increasing tendency to autotomize up to 85 per cent. Then an abrupt drop in the number of autotomies appeared until the eighth, or last leg, autotomized in only 21 per cent of the trials. This increasing tendency to autotomize up to a certain point, followed by a decreasing tendency, is not due to individual differences in the autotomizing mechanisms of different legs, for with such differences alone a varying, random order of stimulation would give a straight line parallel to the abscissa. It 1 No distinction will be made between walking legs and chelae and both types of appendages will be spoken of as legs. ACETYLCHOLINE AND AUTOTOMY IN ANOMURANS 407 can be explained, however, by assuming that each stimulus to a leg, whether the leg is dropped or not, affects the tendency of the other legs to autotomize. Up to the fourth stimulus the effect is excitatory and after the fourth stimulus it is inhibitory. Fifty Petrolisthes injected with 0.05 cc. of " crab Ringer " and stimulated ten to fifteen minutes after injection autotomized in a manner similar to the uninjected controls (Fig. 1). The line which is drawn through the data shown in Fig. 1 may be called the normal curve of autotomy for Petrolisthes, when stimulation 100 o I- o 50 UJ o ct Uj Q. O CONTROL -UN INJECTED • CONTROL- INJECTED LEGS STIMULATED FIG. 1. Percentage of legs dropped by 100 normal animals when each leg was stimulated once, and the order of stimulation was varied from one animal to the next, is shown by the open circles. Results obtained on 50 animals injected with 0.05 cc. of " crab Ringer " are shown by the closed circles. The line drawn through the data has been called the " normal curve " of autotomy. is as described. An obvious change in the " normal curve " produced by the injection of some substance would give an indication of the effect of that substance on the process of autotomy. Acetylcholine. — The injection of various amounts of acetylcholine in preliminary tests indicated that 25 gamma of acetylcholine - was inter- mediate between a lethal dose and one whose effects had largely disap- peared between the time of injection and the time of stimulation. In 2 The amount of substance injected will be expressed as the weight of the salt even though the salt is not indicated. The volume of fluid was always 0.05 cc. 408 J. H. WELSH AND H. H. HASKIN most cases larger amounts caused a number of legs to be dropped during the injection process, or shortly after the injection was made. The legs dropped were almost always on the side on which the injections were made.3 In 20 per cent of the animals the injection of 25 gamma of acetylcholine caused one to three legs to be dropped without further stimulation. Fifty animals, each of which was injected with 25 gamma of acetyl- choline and its legs stimulated approximately five minutes after injection, yielded results which are shown in Fig. 2. The first leg to be grasped too o K- o Ul o 1.59 d-arginine . . . 28 9 / } 21 2 ) } ^/-arginine. . . 14 6 \ 1.98 13 2 1.61 ; J acid only a small increase of the number of cells at the beginning of cultivation can be observed, and the further growth soon stops. The experiments have shown that on 1-aspartic acid both the smooth sinistral (LS) and the smooth dextral (DS) strains of Bacillus mycoides grow better than on the racemic dl-aspartic acid. Among the amino acids used by us the only substance giving unlimited growth of Bacillus mycoides was arginine. The experiments were made at 24° in test tubes with liquid medium of the following composition : K2HPO4 — 0.05 per cent, MgSO4 — 0.03 per cent, Nad — 0.03 per cent, glucose — 1 per cent and arginine — 0.05 per cent. Each number given in Tables I and II is a mean value for five parallel observations. In all cases the number of cells in 1/160 cu. mm. was counted 48 hours after the beginning of the experiment. Both the dextral and sinistral forms of Bacillus mycoides, in both twisted and smooth modifications, grow better on natural d-arginine 2 This hypothesis is further supported by recent observations of Bruckner and Ivanovics (Zeitschr. f. physiol. Chem., 247: 281, 1937). They observed that in the composition of the capsule of some bacteria an unnatural isomer of glutamic acid participates. DEXTRAL AND SINISTRAL FORMS OF BACILLUS 461 than on the racemic dl-arginine. Since racemic arginine in weak watery solution consists of the dextrorotatory and of the laevorotatory isomers, the lessened growth of Bacillus mycoides on this preparation may be explained by the fact that the unnatural laevorotatory isomer, entering into the composition of racemic arginine, is less appropriate for growth than the usual dextrorotatory isomer of arginine. In this relation the properties of dextrally and of sinistrally twisted strains coincide with one another. Oxidation of Glucose by De.vtral and Sinistral Strains of Bacillus mycoides at Different Temperatures Experiments with the oxidation of glucose were made with the Warburg technique at three temperatures : 22°, 25° and 28° C. Bacteria growing on the solid medium at 20° were removed with a platinum loop TABLE II The growth of smooth sinistral (LS) and of smooth dextral (DS) forms of Bacillus mycoides on optically isomeric arginines. Amino acid Number of cells in LS Growth on (/-arginine if growth on Asymmetric structure — > The spiral growth of metabolism of the cell wall £ Forces directly inducing the spiral twist (turgor accord- ing to Castle) One may conjecture that the inversion of the direction of spirals in Bacillus mycoides is related to optical inversion of some secondary sub- stance in metabolism. The latter brings on the inversion of some struc- 464 G. F. CAUSE tures in the cell wall, and during the interaction of these with the forces inducing the spiral twist the inversion in the direction of the spiral growth of the cells is brought about. In conclusion it is to be pointed out that the optical inversion of the secondary substances determining cellular form is judged from indirect evidence, and it is consequently highly desirable to check this conclusion by direct chemical methods. SUMMARY 1. The investigation of different physiological properties of dextral and of sinistral forms of Bacillus mycoides shows that in the basic fea- tures of their metabolism these strains coincide with one another. Both in sinistral and in dextral strains the natural dextrorotatory arginine proves to be more suitable material for growth than the racemic arginine. Both the sinistral and the dextral strains oxidize the natural dextrorota- tory glucose equally well and with the same temperature coefficient. 2. The specific difference between the dextral and the sinistral strains is not consequently connected with the inversion of the basic proto- plasmic system, but with differences in some secondary substances de- termining the cellular form. On the basis of the study of the phenome- non of heat injury in the growth of morphologically inverse dextrally twisted strains of Bacillus mycoides the hypothesis is advanced that some organic substance, participating in the determination of the cellular form, has undergone an optical inversion of configuration in the dextral strains. REFERENCES ARKWRIGHT, I. A., 1930. Variation. A System of Bacteriology. Vol. 1, p. 311. BOYCOTT, A., C. DIVER, S. HARDY AND F. M. TURNER, 1929. The inheritance of sinistrality in Limnaea peregra. Proc. Roy. Soc., Ser. B., 104: 152. CASTLE, E. S., 1936. A model imitating the origin of spiral wall structure in cer- tain plant cells. Proc. Nat. Acad. Sci.. 22: 336. CASTLE, E. S., 1936. The origin of spiral growth in Phycomyces. Jour. Cell. Compar. Physiol, 8: 493. DOOREN DE JONG, L. E., 1933. Ueber Bacillus mycoides und den Pleomorphismus. Arch. Mikrobiol, 4 : 36. CAUSE, G. F., AND N. P. SMARAGDOVA, 1938. An analysis of growth of the yeast Torula utilis on optically isomeric leucines and valines. Biol. Zhurn, 7 : 399. (In Russian with English summary). GERSBACH, A., 1922. Ueber die Wendigkeit der Kolonieauslaufer des Bacillus mycoides ("Isomerie" bei Bakterien). Zentralbl. Bakt., Abt. I, 88: 97. HASTINGS, E. G., AND H. SAGEN, 1933. The effect of physical environment on the cultural character of Bacillus mycoides. Jour. Bact., 25 : 39. JOHN-BROOKS, R. S., 1930. Morphology. A System of Bacteriology. Vol. 1, p. 104. KOLTZOFF, N. K., 1934. Genetics and physiology of development. Biol. Zhurn., 3: 420. (In Russian.) DEXTRAL AND SINISTRAL FORMS OF BACILLUS 465 KONONENKO, E., 1935. On the basic forms of Bacillus mycoides Flugge and their mutual relations. Microbiol., 4:4. (In Russian.) LEWIS, I. M., 1932. Dissociation and life cycle of Bacillus mycoides. Jour. Bact., 24: 381. LEWIS, I. M., 1933. Secondary colonies of bacteria with special reference to Bacillus mycoides. Jour. Bact., 25 : 359. LUDWIG, W., 1932. Das Rechts-Links-Problem im Tierreich und beim Menschen. Berlin. Springer. LUDWIG, W., 1936. Bestimmung und Vererbung der Asymmetrie-form. (Rechts- Links-Problem.) Verh. Zool. Ces., 38: 21. NEEDHAM, J., 1934. Morphology and biochemistry. Nature, 134 : 275. OESTERLE, P., AND C. A. STAHL, 1929. Untersuchungen iiber den Formenwechsel und die Entwicklungsformen bei Bacillus mycoides. Zentralbl. Bakt., Abt. II, 79: 1. PRINGSHEIM, E. G., UND J. LANGER, 1924. Zur Entwicklungsphysiologie der Kolonien des Bacillus mycoides Flugge. Zentralbl. Bakt., Abt. II, 61 : 225. ROBERTS, J. L., 1938. Evidence of a rotational growth factor in Bacillus mycoides. Science, 87 : 260. STAFF, C., UND H. ZYCHA, 1931. Morphologische Untersuchungen an Bacillus mycoides. Arch. Mikrobiol., 2 : 493. INDEX ANATOMY, internal, of two phallo- stethid fishes, 59. Aphids, time of embryonic determina- tion, sensoria and antennal color, determination of wings, ocelli and wing muscle, 442. Arbacia, fertilization reaction, and some properties of sperm extracts, 190. Autotomy and acetylcholine in crusta- cean, 405. TDACILLUS mycoides, physiological properties of dextral and sinistral forms in, 448. Bacteria, and endomixis in Paramecium aurelia, 217. BARNES, T. CUNLIFFE. Experiments on Ligia in Bermuda. VI. Reactions to common cations, 121. BIGELOW, HENRY B., AND WILLIAM C. SCHROEDER. Notes on the fauna above mud bottoms in deep water in the Gulf of Maine, 305. BOKENHAM, N. A. H. See Papenfuss and Bokenham, 1. BONNET, DAVID D. Mortality of the cod egg in relation to temperature, 428. — , - — . See Clarke and Bonnet, 371. BRAUN, WERNER. Contributions to the study of development of the wing- pattern in Lepidoptera, 226. Breeding cycle, Neotoca bilineata, 359. BROWN, FRANK A., JR. Responses of the swimbladder of the guppy, Lebistes reticulatus, to sudden pres- sure decreases, 48. ("^ALANUS fmmarchicus, tempera- ture and survival, growth, and respiration, 371. Centrifuging, effects on polar spindles of Chaetopterus and Cumingia eggs, 339. CHACE, F. A., JR. See Waterman, Nunnemacher, Chace and Clarke, 256. Chaetopterus egg, effects of centrifuging on polar spindles, 339. pergamentaceus, development of half-eggs with reference to partheno- genetic merogony, 384. Clams, temperature and shell-move- ments, 171. CLARKE, G. L. See Waterman, Nunne- macher, Chace and Clarke, 256. — , — . — ., AND DAVID D. BONNET. The influence of temperature on the survival, growth and respiration of Calanus finmarchicus, 371. Cleavage, of frog's egg, as affected by colchicine, 153. Clymenella torquata, buds induced from implants of nerve cord and neigh- boring tissues, 330. Cod egg, mortality in relation to temper- ature, 428. COE, W. R. Sexual phases in terrestrial nemerteans, 416. Colchicine, effects on cleavage of frog's egg, 153. COSTELLO, HELEN M., AND DONALD P. Egg laying in the acoelous turbel- larian Polychoerus carmelensis, 80. Crustacean, autotomy and acetylcholine in, 405. Cumingia egg, effects of centrifuging on polar spindles, 339. £)AWSON, ALDEN B. See Keppel and Dawson, 153. DE LAMATER, ARLENE JOHNSON. Effect of certain bacteria on the occurrence of endomixis in Paramecium aurelia, 217. DETHIER, VINCENT G. Taste thresholds in lepidopterous larvae, 325. Dextral and sinistral forms, Bacillus mycoides, physiological properties, 448. Drugs, effect on insect central nervous system, 183. gCTODERM, of Hydra, cultured independently, 1. Egg laying, in Polychoerus carmelensis, 80. Embryonic determination of sensoria and antennal color, time of, and determi- 466 INDEX 467 nation of wings, ocelli, wing muscle in aphids, 442. Endoderm, of Hydra, cultured inde- pendently, 1. Endomixis, and effect of bacteria on, in Paramecium aurelia, 217. "P RANK, JOHN A. Some properties of sperm extracts and their relationship to the fertilization reaction in Arbacia punctulata, 190. Fundulus heteroclitus, lymphocystis dis- ease of, 251. — , pituitary gland and gonads, 241. USE, G. F. Some physiological properties of dextral and of sinistral forms in Bacillus mycoides, 448. Growth, and temperature, in Calanus, 371. Gulaphallus mirabilis, internal anatomy, 59. Gulf of Maine, deep water fauna above mud bottoms, 305. Guppy, swimbladder responses to sudden pressure decreases, 48. "LJARVEY, ETHEL BROWNE. Devel- opment of half-eggs of Chaetopterus pergamentaceus with reference to parthenogenetic merogony, 384. HASKIN, HAROLD H. See Welsh and Haskin, 405. HSIAO, SIDNEY C. T. The reproduction of Limacina retroversa (Flem.), 280. . , — . — . The reproductive system and spermatogenesis of Limacina (Spiratella) retroversa (Flem.), 7. Hydranth formation, and temperature, Tubularia, 104. HYMAN, LIBBIE H. Some polyclads of the New England coast, especially of the Woods Hole region, 127. INDUCTION, of buds from im- plants of nerve cord and neighboring tissues, in Clymenella torquata, 330. Insect central nervous system, drugs, effect on, 183. T7"EPPEL, DOROTHY M., AND ALDEN B. DAWSON. Effects of colchicine on the cleavage of the frog's egg (Rana pipiens), 153. KILLE, FRANK R. Regeneration of gonad tubules following extirpation in the sea-cucumber, Thyone briar- eus (Lesueur), 70. T EBISTES reticulatus, swimbladder responses to sudden pressure de- creases, 48. Lepidoptera, larvae, taste thresholds, 325. — , wing-pattern, development of, 226. Ligia, reactions to common cations, 121. Limacina, reproductive system and spermatogenesis, 7. - retroversa, history of population, during drift across Gulf of Maine, 26. — , reproduction of, 280. LOOSANOFF, VICTOR L. Effect of tem- perature upon shell movements of clams, Venus mercenaria (L.), 171. Lymphocystis disease of Fundulus heter- oclitus, 251. A/fATTHEWS, SAMUEL A. The re- lationship between the pituitary gland and the gonads in Fundulus, 241. MENDOZA, GUILLERMO. The reproduc- tive cycle of the viviparous teleost, Neotoca bilineata, a member of the family Goodeidae. I. The breeding cycle, 359. Merogony, parthenogenetic, and de- velopment of half-eggs of Chaetop- terus pergamentaceus, 384. Migrations, diurnal vertical, of deep- water plankton, 256. MOORE, JOHN A. The role of tempera- ture in hydranth formation in Tubularia, 104. MORGAN, T. H. The effects of cen- trifuging on the polar spindles of the egg of Chaetopterus and Cumingia, 339. Mortality and temperature in Calanus, 371. — , in cod egg, 428. XTEMERTEANS, terrestrial, sexual phases, 416. Neotoca bilineata, breeding cycle, 359. NUNNEMACHER, R. F. See Waterman, Nunnemacher, Chace and Clarke, 256. 468 INDEX QRYZIAS latipes, effects of 2, 4- dinitrophenol on early development, 162. pAPENFUSS, E. J., AND N. A. H. BOKENHAM. The fate of the ecto- derm and endoderm of hydra when cultured independently, 1. Paramecium aurelia, bacteria and occur- rence of endomixis, 217. Perisarc, removal, effect on regeneration in Tubular ia crocea, 90. Petrolisthes armatus, autotomy and acetylcholine, 405. Phenacostethus smithi, phallostethid fish, internal anatomy, 59. Pituitary gland, and gonads in Fundulus, 241. Plankton, deep-water, diurnal vertical migrations, 256. Polychoerus carmelensis, egg laying, 80. Polyclads of the Woods Hole region, 127. Precipitin technique, standardization and application to relationships in mammals, birds and reptiles, 108. O ANA pipiens, cleavage of egg, effect of colchicine, 153. REDFIELD, ALFRED C. The history of a population of Limacina retroversa during its drift across the Gulf of Maine, 26. Regeneration of gonad tubules following extirpation in the sea-cucumber, 70. — , Tubularia crocea, effect of removal of perisarc, 90. Reproduction of Limacina retroversa, 280. — , cycle of, in Neotoca bilineata, 359. Respiration and temperature, in Calanus, 371. ROEDER, K. D. The action of certain drugs on the insect central nervous system, 183. QAYLES, LEONARD P. Buds in- duced from implants of nerve cord and neighboring tissues in the polychaete, Clymenella torquata, 330. SCHROEDER, WILLIAM C. See Bigelow and Schroeder, 305. Serological techniques, standardization, for studies of relationships, 108. Sex, phases in terrestrial nemerteans, 416. Spermatogenesis and reproductive sys- tem of Limacina, 7. Sperm extracts, properties of, and fertil- ization reaction in Arbacia, 190. STILES, KARL A. The time of embryonic determination of sensoria and anten- nal color and their relation to the determination of wings, ocelli, and wing muscle in aphids, 442. Swimbladder, responses to sudden pres- sure decreases, in guppy, 48. HPASTE thresholds in lepidopterous larvae, 325. Temperature and growth, in Calanus, 371. Temperature, and hydranth formation in Tubularia, 104. - mortality of cod egg, 428. - respiration, growth and survival in Calanus, 371. - shell-movements of clams, 171. TEWlNKEL, Lois E. The internal anat- omy of two phallostethid fishes, 59. Thyone briareus, regeneration of gonad tubules, 70. Tubularia crocea, regeneration, effect of perisarc removal, 90. — , temperature and hydranth forma- tion, 104. WENUS mercenaria, temperature and shell-movements, 171. ATERMAN, A. J. Effects of 2, 4- dinitrophenol on the early develop- ment of the teleost, Oryzias latipes, 162. WATERMAN, T. H., R. F. NUNNEMACHER, F. A. CHACE, JR., AND G. L. CLARKE. Diurnal vertical migrations of deep- water plankton, 256. \VEISSENBERG, RICHARD. Studies on virus diseases of fish. II. Lympho- cystis disease of Fundulus hetero- clitus, 251. WELSH, JOHN H., AND HAROLD H. HASKIN. Chemical mediation in crustaceans, III, 405. Wing-pattern, development of, in Lepi- doptera, 226. WOLFE, HAROLD R. Standardization of the precipitin technique and its application to studies of relation- ships in mammals, birds and reptiles, 108. Woods Hole region, polyclads of, 127. WILLING, EDGAR. The effect of the removal of perisarc on regenera- tion in Tubularia crocea, 90. Volume LXXVI Number 1 THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY Editorial Board GARY N. CALKINS, Columbia University E. E. JUST, Howard University E. G. CONKLIN, Princeton University FRANK R. LlLLIE, University of Chicago E. N. HARVEY, Princeton University CARL R. MOORE, University of Chicago SELIG HECHT, Columbia University GEORGE T. MOORE, Missouri Botanical Garden LEIGH HOADLEY, Harvard University T. H. MORGAN, California Institute of Technology M. H. JACOBS, University of Pennsylvania G. H. PARKER, Harvard University H. S. JENNINGS, Johns Hopkins University F. SCHRADER, Columbia University EDMUND B. WILSON, Columbia University ALFRED C. REDFIELD, Harvard University Managing Editor FEBRUARY, 1939 Printed and Issued by LANCASTER PRESS, Inc. PRINCE & LEMON STS. LANCASTER, PA. Biological Materials For a number of years the Supply Department of the Marine Biological Laboratory has had excellent success in furnishing Living Marine Materials. We can supply groups of Molluscs, Crustaceans, and various Coelenterates in season, as well as the following bal- anced aquaria: Set I. Live Marine Material for five-gallon Aquarium. Includes the following: 5 gallons Sea water 1 small Thyone 2 small Metridium (Sargartia may be substituted) 2 small Mytilus 2 small Starfish 2 Littorina 1 Hermit Crab Green algae (Ulva when 2 small Sea Urchins available) Price per set $5.00 Set II. Live Marine Material for ten-gallon Aquarium. Includes the following: 10 gallons Sea water 3 small Sea Urchins Littorina or Urosalpinx 4 small Metridium 2 small Thyone 1 Nereis 2 small Starfish ., .. „ , 1 Red Starfish (Henricia) 2 small Mytilus 1 small Crab Green algae (Ulva when Balanus Astrangia available) 2 Hermit Crabs Price per set $10.00 We assume responsibility for shipments going as far west as the Mississippi and as far south as Georgia, and guarantee to deliver the animals in good condi- tion between November 1st and March 1st. Carboys only may be returned for full credit if sent by prepaid express as an empty container. Catalogues sent on request Supply Department MARINE BIOLOGICAL LABORATORY Woods Hole, Mass. No. A-196 DISSECTING SETS This illustrates one of the many dissecting sets which comprise our complete stock. Our NEW catalog No. 125 describes and illustrates twelve sets varying from a simple set for the elemen- tary student to an elaborate one for the research technician. We will gladly send you a catalog upon request. Also the Largest Variety of DISSECTING INSTRUMENTS— MICRO SLIDES and COVER GLASSES - MICRO SLIDE BOXES and TRAYS - MAG- NIFIERS—INSECT PINS--RIKER MOUNTS -- MUSEUM JARS — PETRI DISHES - RUBBER TUBING, ETC. There are also available separate catalogs on Charts, Models, Specimens and Preparations covering the fields of: Human and Comparative Anatomy, Physiology, Neurology, Zoology, Botany, Embryology, Entomology, Ecology, etc. THE BIOLOGICAL BULLETIN THE BIOLOGICAL BULLETIN is issued six times a year. Single numbers, $1.75. Subscription per volume (3 numbers), $4.50. Subscriptions and other matter should be addressed to the Biological Bulletin, Prince and Lemon Streets, Lancaster, Pa. Agent for Great Britain : Wheldon & Wesley, Limited, 2, 3 and 4 Arthur Street, New Oxford Street, London, W.C. 2. Communications relative to manuscripts should be sent to the Managing Editor, Marine Biological Laboratory, Woods Hole, Mass., between June 1 and October 1 and to the Biological Lab- oratories, Divinity Avenue, Cambridge, Mass., during the re- mainder of the year. INSTRUCTIONS TO AUTHORS Preparation of Manuscript. In addition to the text matter, manuscripts should include a running page head of not more than thirty-five letters. Footnotes, tables, and legends for figures should be typed on separate sheets. Preparation of Figures. The dimensions of the printed page (4i4x7 inches) should be borne in mind in preparing figures for publication. Draw- ings and photographs, as well as any lettering upon them, should be large enough to remain clear and legible upon reduction to page size. Illustrations should be planned for sufficient reduction to permit legends to be set below them. In so far as possible, explanatory matter should be included in the legends, not lettered on the figures. Statements of magnification should take into account the amount of reduction necessary. Figures will be reproduced as line cuts or halftones. Figures intended for reproduction as line cuts should be drawn in India ink on white paper or blue-lined coordinate paper. Blue ink will not show in reproduction, so that all guide lines, letters, etc. must be in India ink. Figures intended for reproduction as halftone plates should be grouped with as little waste space as possible. Drawings and lettering for halftone plates should be made directly on heavy Bristol board, not pasted on, as the outlines of pasted letters or drawings appear in the reproduction unless removed by an expensive process. Methods of repro- duction not regularly employed by the Biological Bulletin will be used only at the author's expense. The originals of illustrations will not be returned except by special request Directions for Mailing. Manuscripts and illustrations should be packed flat between stiff cardboards. Large charts and graphs may be rolled and sent in a mailing tube. Reprints. Authors will be furnished, free of charge, one hundred re- prints without covers. Additional copies may be obtained at cost. Proof. Page proof will be furnished only upon special request. When cross-references are made in the text, the material referred to should be marked clearly on the galley proof in order that the proper page numbers may be supplied. Manuscripts should be returned with galley proof. Entered October 10, 1902, at Lancaster, Pa., as second-class matter under Act of Congress of July 16, 1894. RECORD THE Results of Research PERMANENTLY B&L Type R Photomicrographic Equip- ment used in conjunction with your micro- scope makes it possible for you to keep a permanent record of your microscopical findings, easily and economically. Type R Equipment (illustrated with the B&L Microscope GGB) consists of a B&L Type H Camera using 5x7 plates and an illuminating unit mounted on a common support. It is a convenient and adaptable arrangement that may be used horizontally or vertically. It is equipped with a compound shutter. For gross photography a rack and pinion focus- ing mount for focusing Micro Tessar lenses may be mounted on the front board. The camera may be quickly swung aside when it is desired to do visual work. For complete information write Bausch & Lomb Optical Co., 707 St. Paul Street, Rochester, N. Y. BAUSCH &• LOMB ....WE MAKE OUR OWN GLASS IO INSURE STANDARDIZED PRODUCTION FOR TOUR GLASSES INSISI ON B « L ORIHOGON LENSES AND B > L fRAMES . . . MADE IN U. S. A. The Standard for Microscope Glass Gold Seal Microscope Slides and Cover Glasses Crystal Clear • Non-Corrosive • Will Not Fog Microscopic work demands glass of unusual clarity. Gold Seal Slides and Cover Glasses are made from glass practically free from alkali. They attain a precise and uniform thinness of plane surface. Therefore, Gold Seal offers an unusual degree of crystal clarity. Further, Gold Seal is guaran- teed against corrosion, fogging or any imperfection. Specify Gold Seal Slides and Cover Glasses. 44 EAST 23.p STREET, NEW IMPORTANT NOTICE TO SUBSCRIBERS / IBRARIES and individuals desiring to complete sets or runs of the BIOLOGICAL BULLETIN will be able to ob- tain certain numbers and volumes at reduced prices. In order to equalize stock, a number of special offers are being made. Prices will be furnished on request. Please address communications to: SECRETARY, THE BIOLOGICAL BULLETIN, Marine Biological Laboratory, Woods Hole, Massachusetts A Perfect Illustration Or the lack of it, may make or mar a scientific paper. For 65 years we have specialized in making reproductions by the Helio- type process of the most delicate pencil and wash drawings and photo- graphs; and by the Heliochrome proc- ess, of paintings and drawings in color. Ask the editor to whom you submit your next paper to secure our esti- mates for the reproduction of your illustrations. I he Heliotype Corporation Est. 1872 172 Green St., Jamaica Plain, Boston, Mass. LANCASTER PRESS, Inc. LANCASTER, PA. •6 THE EXPERIENCE we have gained from printing some sixty educational publica- tions has fitted us to meet the standards of customers who demand the best. We shall be happy to have workers at the MARINE BIOLOGICAL LABORATORY write for estimates on journals or monographs. Our prices are moderate. CONTENTS Page PAPENFUSS, E. J., AND N. A. H. BOKENHAM The Fate of the Ectoderm and Endoderm of Hydra when Cultured Independently 1 HSIAO, SIDNEY C. T. The Reproductive System and Spermatogenesis of Limacina (Spiratella) retroversa (Flem.) 7 REDFIELD, ALFRED C. The History of a Population of Limacina retroversa during its Drift Across the Gulf of Maine 26 BROWN, FRANK A., JR. Responses of the Swimbladder of the Guppy, Lebistes reticu- latus, to Sudden Pressure Decreases 48 TEWINKEL, Lois E. The Internal Anatomy of Two Phallostethid Fishes 59 KILLE, FRANK R. Regeneration of Gonad Tubules Following Extirpation in the Sea-Cucumber, Thyone briareus (Lesueur) 70 COSTELLO, HELEN M., AND DONALD P. COSTELLO Egg Laying in the Acoelous Turbellarian Polychoerus car- melensis 80 ZWILLING, EDGAR The Effect of the Removal of Perisarc on Regeneration in Tubularia crocea 90 MOORE, JOHN A. The Role of Temperature in Hydranth Formation in Tubularia 104 WOLFE, HAROLD R. Standardization of the Precipitin Technique and its Applica- tion to Studies of Relationships in Mammals, Birds and Reptiles 108 BARNES, T. CUNLIFFE Experiments on Ligia in Bermuda. VI. Reactions to common cations. 121 Volume LXXVI Number 2 THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY Editorial Board GARY N. CALKINS, Columbia University E. E. JUST, Howard University E. G. CONKLIN, Princeton University FRANK R. LlLLIE, University of Chicago E. N. HARVEY, Princeton University CARL R. MOORE, University of Chicago SELIG HECHT, Columbia University GEORGE T. MOORE, Missouri Botanical Garden LEIGH HOADLEY, Harvard University T. H. MORGAN, California Institute of Technology M. H. JACOBS, University of Pennsylvania G. H. PARKER, Harvard University H. S. JENNINGS, Johns Hopkins University F. SCHRADER, Columbia University ALFRED C. REDFIELD, Harvard University Managing Editor APRIL, 1939 Printed and Issued by LANCASTER PRESS, Inc. PRINCE & LEMON STS. LANCASTER, PA. The Biology of the Cell Surface How Does Life Reveal Itself? This is a timely book which will appeal to all who look with interest upon the mani- festation of life in animals and in man. The biologist, whatever his special interest, at some time or other is concerned with the development of the egg. The author presents from a purely biological point of view a thesis which sets a new goal for biology, the science of life. He unravels the problems of animal development, exposes them singly, defines them, and relates them to the activity of the cell surface and to the larger ques- tions : What is life and how does life reveal itself ? Dr. Just, an experimental embryologist of thirty years experience, has a peculiar talent for handling living eggs and observing vital processes. This talent together with his rare analytical mind has made him known in biological circles throughout the world. He has also an exceptional ability to express abstract truth with simplicity and clearness and thus related to human experience. In this book he brings his readers into an arena of conflicting biological thought, expressing himself with such clearness that all can follow his argument. BY ERNEST EVERETT JUST 42 Illustrations (116 Figures) Some in Colors. Tables, Bibliography. 392 Pages. Washable Fabric, $5.50 P. BLAKISTON'S SON & CO., Inc., Philadelphia, Pennsylvania Slide number E13.78 of sections of whitefish blastula is an outstanding preparation in several respects. When you use this slide to teach mitotic division, there will be no time lost hunting for the figures, for each slide car- ries many diagrammatically clear figures of every stage of mitosis. Hundreds of teachers consider it far superior to Ascaris, Tradescantia or onion root-tip. Sample slides of this preparation will be forwarded at your request. GENERAL BIOLOGICAL SUPPLY HOUSE (Inc.) 761-763 EAST 69TII PLACE CHICAGO, ILLINOIS No. A-196 DISSECTING SETS This illustrates one of the many dissecting sets which comprise our complete stock. Our NEW catalog No. 125 describes and illustrates twelve sets varying from a simple set for the elemen- tary student to an elaborate one for the research technician. We will gladly send you a catalog upon request. Also the Largest Va riety of DISSECTING INSTRUMENTS— MICRO SLIDES and COVER GLASSES - MICRO SLIDE BOXES and TRAYS - MAG- NIFIERS-- INSECT PINS-- RIKER MOUNTS -- MUSEUM JARS PETRI DISHES - RUBBER TUBING, ETC. There are also available separate catalogs on Charts, Models, Specimens and Preparations covering the fields of: Human and Comparative Anatomy, Physiology, Neurology, Zoology, Botany, Embryology, Entomology, Ecology, etc. EAST 23«D STREET THE BIOLOGICAL BULLETIN THE BIOLOGICAL BULLETIN is issued six times a year. Single numbers, $1.75. Subscription per volume (3 numbers), $4.50. Subscriptions and other matter should be addressed to the Biological Bulletin, Prince and Lemon Streets, Lancaster, Pa. Agent for Great Britain: Wheldon & Wesley, Limited, 2, 3 and 4 Arthur Street, New Oxford Street, London, W.C. 2. Communications relative to manuscripts should be sent to the Managing Editor, Marine Biological Laboratory, Woods Hole, Mass., between June 1 and October 1 and to the Biological Lab- oratories, Divinity Avenue, Cambridge, Mass., during the re- mainder of the year. INSTRUCTIONS TO AUTHORS Preparation of Manuscript. In addition to the text matter, manuscripts should include a running page head of not more than thirty-five letters. Footnotes, tables, and legends for figures should be typed on separate sheets. Preparation of Figures. The dimensions of the printed page (4^4x7 inches) should be borne in mind in preparing figures for publication. Draw- ings and photographs, as well as any lettering upon them, should be large enough to remain clear and legible upon reduction to page size. Illustrations should be planned for sufficient reduction to permit legends to be set below them. In so far as possible, explanatory matter should be included in the legends, not lettered on the figures. Statements of magnification should take into account the amount of reduction necessary. Figures will be reproduced as line cuts or halftones. Figures intended for reproduction as line cuts should be drawn in India ink on white paper or blue-lined coordinate paper. Blue ink will not show in reproduction, so that all guide lines, letters, etc. must be in India ink. Figures intended for reproduction as halftone plates should be grouped with as little waste space as possible. Drawings and lettering for halftone plates should be made directly on heavy Bristol board, not pasted on, as the outlines of pasted letters or drawings appear in the reproduction unless removed by an expensive process. Methods of repro- duction not regularly employed by the Biological Bulletin will be used only at the author's expense. The originals of illustrations will not be returned except by special request. Directions for Mailing. Manuscripts and illustrations should be packed flat between stiff cardboards. Large charts and graphs may be rolled and sent in a mailing tube. Reprints. Authors will be furnished, free of charge, one hundred re- prints without covers. Additional copies may be obtained at cost. Proof. Page proof will be furnished only upon special request. When cross-references are made in the text, the material referred to should be marked clearly on the galley proof in order that the proper page numbers may be supplied. Manuscripts should be returned with galley proof. Entered October 10, 1902, at Lancaster, Pa., as second-class matter under Act of Congress of July 16, 1894. BIOLOGICAL ABSTRACTS Now published in FIVE MONTHLY SECTIONS in addition to the present complete form One or more of these sections should appeal to every scientist working in Biology. ABSTRACTS OF GENERAL BIOLOGY will include General Biology, Biog- raphy-History, Bibliography, Evolution, Cytology, Genetics, Biometry and Ecology. Priced at $4. ($4.50 Foreign.) ABSTRACTS OF EXPERIMENTAL ANIMAL BIOLOGY will include Ani- mal Physiology, Nutrition, Pharmacology, Pathology, Anatomy, Embryology, Animal Production and Ecology. Priced at $9. ($4.50 Foreign.) ABSTRACTS OF MICROBIOLOGY AND PARASITOLOGY will include Immunology, Bacteriology, Viruses, Parasitology, Protozoology and Helmin- thology. Priced at $5. ($5.50 Foreign.) ABSTRACTS OF PLANT SCIENCES will include Phytopathology, Plant Physiology, Plant Anatomy, Paleobotany, Systematic Botany, Agronomy, Horticulture, Forestry, Pharmacognosy, Pharmaceutical Botany and Ecology. Priced at $6. ($6.50 Foreign.) ABSTRACTS OF ANIMAL SCIENCES will include Paleozoology and Helmin- thology, Systematic Zoology, Economic Entomology and Ecology. Priced at $6. ($6.50 Foreign.) Subscribers to one or more of these sections will receive the Indexes to the whole of BIOLOGICAL ABSTRACTS. (Foreign subscribers add 50 cents per section for postage.) Send your order now! BIOLOGICAL ABSTRACTS, University of Pennsylvania, Philadelphia, Pa. IMPORTANT NOTICE TO SUBSCRIBERS / IBRARIES and individuals desiring to complete sets or runs of the BIOLOGICAL BULLETIN will be able to ob- tain certain numbers and volumes at reduced prices. In order to equalize stock, a number of special offers are being made. Prices will be furnished on request. Please address communications to: SECRETARY, THE BIOLOGICAL BULLETIN, Marine Biological Laboratory, Woods Hole, Massachusetts Biology Materials PRESERVED SPECIMENS for Zoology, Botany, Embryology, and Comparative Anatomy LIVING SPECIMENS for Zoology and Botany including Algae, Protozoan and Drosophila Cultures, and Animals for Experimental and Laboratory use. MICROSCOPE SLIDES for Zoology, Botany, Embryology, Histology, Bacteriology, and Parasitology. Catalogues promptly sent on request Supply Department MARINE BIOLOGICAL LABORATORY Woods Hole, Mass. CONTENTS Page HYMAN, LIBBIE H. Some Polyclads of the New England Coast, especially of the Woods Hole Region 127 KEPPEL, DOROTHY M. AND ALDEN B. DAWSON Effects of Colchicine on the Cleavage of the Frog's Egg (Rana pipiens) 153 WATERMAN, A. J. Effects of 2, 4-Dinitrophenol on the Early Development of the Teleost, Oryzias latipes 162 LOOSANOFF, VICTOR L. Effect of Temperature upon Shell Movements of Clams, Venus mercenaria (L.) 171 ROEDER, K. D. The Action of Certain Drugs on the Insect Central Nervous System 183 FRANK, JOHN A. Some Properties of Sperm Extracts and their Relationship to the Fertilization Reaction in Arbacia punctulata 190 DE LAMATER, ARLENE JOHNSON Effect of Certain Bacteria on the Occurrence of Endomixis in Paramecium aurelia 217 BRAUN, WERNER Contributions to the Study of Development of the Wing- pattern in Lepidoptera 226 MATTHEWS, SAMUEL A. The Relationship between the Pituitary Gland and the Gonads in Fundulus 241 WEISSENBERG, RICHARD Studies on Virus Diseases of Fish. II. Lymphocystis disease of Fundulus heteroclitus 251 WATERMAN, T. H., R. F. NUNNEMACHER, F. A. CHACE, JR. AND G. L. CLARKE Diurnal Vertical Migrations of Deep-water Plankton 256 HSIAO, SIDNEY C. T. The Reproduction of Limacina retroversa (Flem.) 280 Volume LXXVI Number 3 THE BIOLOGICAL BULLETIN PUBLISHED BY THE MARINE BIOLOGICAL LABORATORY Editorial Board GARY N. CALKINS, Columbia University E. G. CONKLIN, Princeton University E. N. HARVEY, Princeton University SELIG HECHT, Columbia University LEIGH HOADLEY, Harvard University M. H. JACOBS, University of Pennsylvania H. S. JENNINGS, Johns Hopkins University E. E. JUST, Howard University FRANK R. LILLIE, University of Chicago CARL R. MOORE, University of Chicago GEORGE T. MOORE, Missouri Botanical Garden T. H. MORGAN, California Institute of Technology G. H. PARKER, Harvard University F. SCHRADER, Columbia University ALFRED C. REDFIELD, Harvard University Managing Editor JUNE, 1939 Printed and Issued by LANCASTER PRESS, Inc. PRINCE & LEMON STS. LANCASTER, PA. NEW BIOLOGY TEXTS The Phylum Chordata By H. H. Newman $3.60 This new text, by an inspiring and experienced teacher, sets the pace for a memorable course in vertebrate zoology or comparative anatomy. It spreads before the student the fascinating and amazing history of the vertebrates in such a way that he may see the broad significance of the data he observes in his laboratory work, and may carry away from his study a sharpened facility for observation and an enriched understanding of the vertebrate world around him. ^Biology of the Vertebrates By II. E. 11'altcr Revised Edition $4.00 (probable; The new edition of this widely used text provides a full, up-to- date course of study on the vertebrates, outstanding for its effec- tive organization and for its interesting style of presentation. In addition to complete material on vertebrate evolution and anatomy, there is a useful introductory survey of some of the important features of the sciences most intimately related to comparative anatomy — -ecology, palaeontology, cytology, embry- ology, etc. Elements of Plant Pathology By I. E. Mcllnis and G. C. Kent. $4.50 (probable) The excellent choice of material, the unusually effective and interesting presentation, and the unique illustrative equipment all combine to make this new book the outstanding text on the market for introductory courses in plant pathology. Attention is focused on parasitism — on living organisms in disease processes. The diseases chosen to illustrate general principles are widely representative, and information about them thoroughly up to date. The book is illustrated with 258 diagrams and photographs, 9Urr of which are original. THE MACMILLAN COMPANY • 60 FIFTH AVE., NEW YORK No. A-196 DISSECTING SETS This illustrates one of the many dissecting sets which comprise our complete stock. Our NEW catalog No. 125 describes and illustrates twelve sets varying from a simple set for the elemen- tary student to an elaborate one for the research technician. We will gladly send you a catalog upon request. Also the Largest Variety of DISSECTING INSTRUMENTS— MICRO SLIDES and COVER GLASSES - MICRO SLIDE BOXES and TRAYS - MAG- NIFIERS-- INSECT PINS--RIKER MOUNTS -- MUSEUM JARS PETRI DISHES - RUBBER TUBING, ETC. There are also available separate catalogs on Charts, Models, Specimens and Preparations covering the fields of: Human and Comparative Anatomy, Physiology, Neurology, Zoology, Botany, Embryology, Entomology, Ecology, etc. 44 EAST 23u STREET, NEW YORK THE BIOLOGICAL BULLETIN THE BIOLOGICAL BULLETIN is issued six times a year. Single numbers, $1.75. Subscription per volume (3 numbers), $4.50. Subscriptions and other matter should be addressed to the Biological Bulletin, Prince and Lemon Streets, Lancaster, Pa. Agent for Great Britain: Wheldon & Wesley, Limited, 2, 3 and 4 Arthur Street, New Oxford Street, London, W.C. 2. Communications relative to manuscripts should be sent to the Managing Editor, Marine Biological Laboratory, Woods Hole, Mass., between June 1 and October 1 and to the Biological Lab- oratories, Divinity Avenue, Cambridge, Mass., during the re- mainder of the year. INSTRUCTIONS TO AUTHORS Preparation of Manuscript. In addition to the text matter, manuscripts should include a running page head of not more than thirty-five letters. Footnotes, tables, and legends for figures should be typed on separate sheets. Preparation of Figures. The dimensions of the printed page (4%x7 inches) should be borne in mind in preparing figures for publication. Draw- ings and photographs, as well as any lettering upon them, should be large enough to remain clear and legible upon reduction to page size. Illustrations should be planned for sufficient reduction to permit legends to be set below them. In so far as possible, explanatory matter should be included in the legends, not lettered on the figures. Statements of magnification should take into account the amount of reduction necessary. Figures will be reproduced as line cuts or halftones. Figures intended for reproduction as line cuts should be drawn in India ink on white paper or blue-lined coordinate paper. Blue ink will not show in reproduction, so that all guide lines, letters, etc. must be in India ink. Figures intended for reproduction as halftone plates should be grouped with as little waste space as possible. Drawings and lettering for halftone plates should be made directly on heavy Bristol board, not pasted on, as the outlines of pasted letters or drawings appear in the reproduction unless removed by an expensive process. Methods of repro- duction not regularly employed by the Biological Bulletin will be used only at the author's expense. The originals of illustrations will not be returned except by special request. Directions for Mailing. Manuscripts and illustrations should be packed flat between stiff cardboards. Large charts and graphs may be rolled and sent in a mailing tube. Reprints. Authors will be furnished, free of charge, one hundred re- prints without covers. Additional copies may be obtained at cost. Proof. Page proof will be furnished only upon special request. When cross-references are made in the text, the material referred to should be marked clearly on the galley proof in order that the proper page numbers may be supplied. Manuscripts should be returned with galley proof. Entered October 10, 1902, at Lancaster, Pa., as second-class matter under Act of Congress of July 16, 1894. BIOLOGICAL ABSTRACTS Now published in FIVE MONTHLY SECTIONS in addition to the present complete form One or more of these sections should appeal to every scientist working in Biology. ABSTRACTS OF GENERAL BIOLOGY will include General Biology, Biog- raphy-History, Bibliography, Evolution, Cytology, Genetics, Biometry and Ecology. Priced at $4. ($4.50 Foreign.) ABSTRACTS OF EXPERIMENTAL ANIMAL BIOLOGY will include Ani- mal Physiology, Nutrition, Pharmacology, Pathology, Anatomy, Embryology, Animal Production and Ecology. Priced at $9. ($4.50 Foreign.) ABSTRACTS OF MICROBIOLOGY AND PARASITOLOGY will include Immunology, Bacteriology, Viruses, Parasitology, Protozoology and Helmin- thology. Priced at $5. ($5.50 Foreign.) ABSTRACTS OF PLANT SCIENCES will include Phytopathology, Plant Physiology, Plant Anatomy, Paleobotany, Systematic Botany, Agronomy, Horticulture, Forestry, Pharmacognosy, Pharmaceutical Botany and Ecology. Priced at $6. ($6.50 Foreign.) ABSTRACTS OF ANIMAL SCIENCES will include Paleozoology and Helmin- thology, Systematic Zoology, Economic Entomology and Ecology. Priced at $6. ($6.50 Foreign.) Subscribers to one or more of these sections will receive the Indexes to the whole of BIOLOGICAL ABSTRACTS. (Foreign subscribers add 50 cents per section for postage.) Send your order now! BIOLOGICAL ABSTRACTS, University of Pennsylvania, Philadelphia, Pa. IMPORTANT NOTICE TO SUBSCRIBERS / IBRARIES and individuals desiring to complete sets or ^~* of the BIOLOGICAL BULLETIN will be able to ob- tain certain numbers and volumes at reduced prices. In order to equalize stock, a number of special offers are being made. Prices will be furnished on request. Please address communications to: SECRETARY, THE BIOLOGICAL BULLETIN, Marine Biological Laboratory, Woods Hole, Massachusetts The Standard for Microscope Glass Gold Seal Microscope Slides and Cover Glasses Crystal Clear • Non-Corrosive • Will Not Fog Microscopic work demands glass of unusual clarity. Gold Seal Slides and Cover Glasses are made from glass practically free from alkali. They attain a precise and uniform thinness of plane surface. Therefore, Gold Seal offers an unusual degree of crystal clarity. Further, Gold Seal is guaran- teed against corrosion, fogging or any imperfection. Specify Gold Seal Slides and Cover Glasses. MADE IN U. S. A. CLAY-A0AMS COL' 44 EAST 23>D STREET, NEW YORK A Perfect Illustration LANCASTER PRESS, Inc. LANCASTER, PA. Or the lack of it, may make or mar a scientific paper. •fr For 65 years we have specialized in THE EXPERIENCE we have making reproductions by the Helio- type process of the most delicate gained from printing some pencil and wash drawings and photo- graphs; and by the Heliochrome proc- sixty educational publica- ess, of paintings and drawings in color. tions has fitted us to meet Ask the editor to whom you submit the standards of customers your next paper to secure our esti- * f f mates for the reproduction of your who demand the best. illustrations. The Heliotype Corporation We shall be happy to have workers at the MARINE BIOLOGICAL LABORATORY Est. 1872 172 Green St., Jamaica Plain, write for estimates on journals or Boston, Mass. monographs. Our prices are moderate. Biology Materials The Supply Department of the Marine Biological Labora- tory has a complete stock of excellent material for summer school work, including the following: PRESERVED SPECIMENS for Zoology, Botany, Embryology, and Comparative Anatomy LIVING SPECIMENS for Zoology and Botany including Algae, Protozoan and Drosophila Cultures, and Animals for Experimental and Laboratory Use. MICROSCOPE SLIDES for Zoology, Botany, Embryology, Histology, Bacteriology, and Parasitology. Catalogues promptly sent on request Supply Department MARINE BIOLOGICAL LABORATORY Woods Hole, Mass. CONTENTS Page BIGELOW, HENRY B., AND WILLIAM C. SCHROEDER Notes on the Fauna above Mud Bottoms in Deep Water in the Gulf of Maine 305 DETHIER, V. G. Taste Thresholds in Lepidopterous Larvae 325 SAYLES, LEONARD P. Buds Induced from Implants of Nerve Cord and Neighboring Tissues in the Polychaete, Clymenella torquata 330 MORGAN, T. H. The Effects of Centrifuging on the Polar Spindles of the Egg of Chaetopterus and Cumingia 339 MENDOZA, GUILLERMO The Reproductive Cycle of the Viviparous Teleost, Neotoca bilineata, A Member of the Family Goodeidae. I. The Breeding Cycle 359 CLARKE, GEORGE L., AND DAVID D. BONNET The Influence of Temperature on the Survival, Growth and Respiration of Calanus finmarchicus 371 HARVEY, ETHEL BROWNE Development of Half-eggs of Chaetopterus pergamentaceus with special reference to Parthenogenetic Merogony 384 WELSH, JOHN H., AND HAROLD H. HASKIN Chemical Mediation in Crustaceans. III. Acetylcholine and Autotomy in Petrolisthes armatus (Gibbes) 405 COE, W. R. Sexual Phases in Terrestrial Nemerteans 416 BONNET, DAVID D. Mortality of the Cod Egg in Relation to Temperature 428 STILES, KARL A. The Time of Embryonic Determination of Sensoria and An- tennal Color and their Relation to the Determination of Wings, Ocelli and Wing Muscle in Aphids 442 GAUSE, G. F. Some Physiological Properties of Dextral and of Sinistral Forms in Bacillus mycoides Fliigge 448 LIBRARY UH 17IU J