BIOLOGICAL BULLETIN OF THE flDarine Biological laboratory WOODS HOLE, MASS. Editorial Staff E. G. CONK. MX — Princeton University. JACQUES LOEB — 7Yie Rockefeller Institute for !\Icdical Research. T. H. MORGAN — Columbia University. W. M. WHEELER — Harvard University. E. B. WILSON — Columbia University. Managing FRANK R. LILLIE — The University of Chicago. VOLUME XXV. WOODS HOLE, MASS. JUNE TO NOVEMBER 1913 PRESS OF THE NEW ERA PRINTING COMPANY LANCASTER. PA. CONTENTS OF VOLUME XXV Xo. i. JUNE, 191. v KITE, G. L. The Relative Permeability of the Surface and In- terior Portions of the Cytoplasm of Animal and Plant Cells. I WOODSEDALEK, J. E. Spermatogenesis of the Pig with Special Reference to the Accessory Chromosomes. SAFIR. SHELLEY R. A New Eye Color Mutation in Drosophila and its Mode of Inheritance 47 XEWMAX, H. H. Parthenogenetic Cleavage of the Armadillo Ovum 54 - Xo. 2. JULY, 1913. SHELFORD, VICTOR E. The Reactions of Certain Animals to Gradients of Evaporating Power of Air. A Study in Experi- mental Ecology 79 COTT, GEORGE C. Some Effects on Fund id us of Changes in the Density of the Surrounding Medium 121 CHAMBERS, ROBERT, JR. The Spermatogenesis of a Daphnid, Simocephalus vetulus ' .'T 134 Xo. 3. AUGUST, 1913. VESTAL, ARTHUR G. Local Distribution of Grasshoppers in Relation to Plant Associations 141 CHILD, C. M. The Asexual Cycle of Planaria I'data in Relation to Senescence and Rejuvenescence 181 HOLMES, S. J. Developmental Changes of Pieces of Frog Embryos Cultivated in Lymph 204 BERRY, S. STILLMAN. Nematolampas, A Remarkable New Cephalopod from the South Pacific 208 Xo. 4. SEPTEMBER, 1913. DUXGAY, NEIL S. A Study of the Effects of Injury upon the Fertilizing Power of Sperm ... 213 MORSE, EDWARD S. Observations on Living Solenomya (velum and borealis) 261 TASHIRO, SHIRO. A Chemical Sign of Life . . 282 in iv CONTENTS OF VOLUME XXV No. 5. OCTOBER, 1913. FAUST, E. C. Size Dimorphism in Adult Spermatozoa of Anasa tri.stis 287 SCOTT, JOHN XX". Experiments icith Tapeworms. I. Some Factors Producing Evagination of a Cysticercus 304 XX'AGNER, GEORGE. On a Peculiar Monstrosity in a Frog, 313 WHITNEY, D. D. An Explanation of the Non-Production of Fertilized Eggs by Adult Male-Producing Females in a Species of Asplanchna 3lS No. 6. NOVEMBER, 1913. WELLS, MORRIS M. The Resistance of Fishes to Different Con- centrations and Combinations of Oxygen and Carbon Dioxide. 323 TURNER, C. H. Behavior of the Common Roach (Periplaneta orientalis L.) on an Open Maze.. . 348 Vol. XXV. June, 1913. No. i BIOLOGICAL BULLETIN THE RELATIVE PERMEABILITY OF THE SURFACE AND INTERIOR PORTIONS OF THE CYTO- PLASM OF ANIMAL AND PLANT CELLS.1 (A PRELIMINARY PAPER.) G. L. KITE. (From the Department of Pathology and The Otho S. A. Sprague Memorial In- stitute, University of Chicago.) The permeability and osmotic properties of cells are explained on the assumption of the presence of surface and vacuolar plas- matic membranes which are of a semi-permeable or partially permeable character. It is generally held that the remainder of the protoplasm of a given cell plays a negligible role in both diosmosis and permeability. These assumptions are based on indirect evidence. The permeability of the internal cytoplasm of living animal and plant cells has not been investigated by direct methods. This paper gives the results of a study of the permeability of the internal cytoplasm and nucleus to dyes and crystalloids. The animal material used in this investigation included the eggs of Asterias, Cumingia, Chcetopterus, Nereis and the immature eggs of Necturus, Ameba proteus, Paramecium and the striped muscle and epidermal cells of Necturus. The plants selected were Saccharomyces, Mucbr, Saprolegnia, some five species of Spirogyra, Hydrodictyon, the manubrial cells of Chara, the leaves of Elodea, root-hairs of Vicia faba, Pisum, Hordeum and the parenchyma cells of Tradescantia. 1 The studies on marine eggs reported in this paper were carried on during the summer of 1912 at the Marine Biological Laboratory, Woods Hole, Mass., while occupying a table through the courtesy of the Director, Dr. F. R. Lillie, to whom I am indebted for many kindnesses. I 2 G. L. KITE. A number of methods were used in order to make the data comparative and to determine the error and limitations of a given method employed. The effect of various operations, such as partial dissections, and punctures, on the permeability and os- motic properties of cells was studied. Intracellular injections were made by Barber's method.1 THE PERMEABILITY OF ANIMAL AND PLANT CELLS TO ACID DYES. Incidental to this study, a number of facts bearing on Over- ton's so-called lipoid theory have been determined, and as animal physiologists have taken kindly to this notion some of these data will be given. Rhuland2 and others have shown a number of exceptions to Overton's3 conclusion that cells are impermeable to lipoid in- soluble acid dyes. By using a large number of species of animals and plants from widely separated genera and phyla, I have found it quite easy to discover cells that are freely permeable to many lipoid - insoluble acid dyes. Such well-known acid dyes as eosin and trypan red are good vital stains for Mucor and Saprolegnia. The lipoid-insoluble acid dyes used include: Trypan blue, which penetrates the eggs of Nereis and Chatop- terus, the root-hairs of barley and the Windsor bean, immature Necturus eggs, the peritoneal epithelium of Necturus, Ameba proteus, Paramecium, Mucor, and Saprolegnia. Trypan red : the eggs of Cumingia and Ch&topterus, the root- hairs of barley, the edible pea and the Windsor bean, Ameba proteus, Paramecium, Mucor, and Saprolegnia. Isamin blue: the root-hairs of barley and the Windsor bean. Analine blue: the eggs of Choetopterus, the root-hairs of barley and the Windsor bean, and Mucor. Acid fuchsin: The root-hairs of barley, the Windsor bean and Mucor. Acid green : Saprolegnia and Mucor. 1 Barber, Jour, of Inf. Dis., 1911, VIII., p. 348. 2 Jahr. Wiss. BoL, Bel. 51, H. 3, p. 376. 3 Jahr. f. Wiss. Bot., Ed. 34, 1900, p. 669. CYTOPLASM OF ANIMAL AND PLANT CELLS. 3 Acid violet: Windsor bean root-hairs, Saprolegnia, and Mucor. Beibricher scarlet: Barley root-hairs, the immature eggs of Necturus, the peritoneal epithelium of Necturus, Mucor, Sapro- legnia, and Paramecium. Indigo-carmin: Saprolegnia and Mucor. Ponceau, P. R.: the root-hairs of the Windsor bean, Ameba proteus, Paramecium, Saprolegnia, and Mucor. Indulin: root-hairs of Windsor bean, Ameba proteus, Sapro- legnia, and Mucor. Nigrosin: the peritoneal epithelium and the very immature eggs of Necturus, and Mucor. Eosin: Windsor bean root-hairs, Ameba proteus, Paramecium Mucor, and Saprolegnia. In recent papers Loewe1 states that even basic dyes are ad- sorbed by lipoids, when added to solutions of lipoids in organic solvents. THE INTRACELLULAR INJECTION OF DYES AND CRYSTALLOIDS. If indigo-carmin, methyl red, trypan blue, thiocarmine R., or azolitmin dissolved in sea-water be injected into any portion of the cytoplasm of the starfish egg, the sea water slowly diffuses into the surrounding protoplasmic gel, and finally the granular precipitated dye alone remains. An injection of indigo-carmin into the cytoplasm of an immature egg of Necturus results in a slight staining of the wall of the vacuole, while an injection of such widely different dyes, as indigo-carmin, trypan blue, and janus green (diethyl-safranin-azo-dimethyl-anilin), into the cytoplasm of the striped muscle cell of this animal results at most in a localized staining of the cytoplasmic gel immediately surrounding the mass of injected dye. The injection of indigo-carmin or nigrosin into the vacuole of Spirogyra results in a blue-green or light violet staining, respectively, of the entire cytoplasm. A blue-green staining of the cytoplasm of Hydrodictyon is affected by an intravacuolar injection of indigo-carmin. In fact, every acid dye that was injected into the vacuole of Spirogyra, Hydro- dictyon, the leaf cells of Elodea, and the parenchyma cells of Tradescantia penetrated, and stained the cytoplasm and usually the nucleus. 1 Loewe, Biochemische Zeitschrift, 1912, X'LIL, p. 150. 4 G. L. KITE. Particularly striking permeability phenomena are to be ob- served when dyes that do not penetrate Ameba protects are in- jected into the interior. Azolitmin, congo red, tropeolin ooo No. i, sodium alizarin sulphonate, and indigo-carmin were used for these injections. All the dyes enumerated were dissolved in salt and sugar solutions of different concentrations, and small and large doses were injected. It was found that these dyes diffuse quickly through the interior of A. proteus, if the concentration of the salt or sugar solution be not too high. A localized blue vacuole results when indigo-carmin, dissolved in distilled water, is in- jected into the ectoplasm of this animal. Usually in a short time the vacuole breaks into the interior and the dye rapidly diffuses. Sea water, distilled water, solutions of sodium chloride, po- tassium nitrate, and cane sugar have been injected into the interior of different types of cells. A small dose of distilled water is taken up by the surrounding cytoplasm of the starfish egg quite slowly. A vacuole of sea water requires a somewhat longer time to disappear, while a vacuole filled with hypertonic sea water increases in size. Hence, any portion of the cytoplasm of the starfish egg can exhibit the same general diosmotic proper- ties that are shown by the surface. It seems that this is also true for the cytoplasm of the striped muscle cell of Necturus, but on account of the very high viscosity of this substance the wall surrounding the injected fluid remains very irregular, and no accurate measurement of the volume injected could be made. Distilled water is absorbed from the interior of the muscle substance extremely slowly; and salt and sugar solutions more slowly or not at all. Doses of i m. sodium chloride and potassium nitrate and from .5 M. to 2 M. cane sugar diffuse through the cytoplasm of Ameba proteus when injected into the interior. The vacuoles that are formed by the injected fluid quickly collapse. Granules, fibrils, and globules can be produced in proteus by the injection of 2 M. cane sugar. The shrunken part of the cytoplasm does not readily take up water again. In this connection it may be added that the injection of mer- curv and isotonic salt solutions and the introduction of various CYTOPLASM OF ANIMAL AND PLANT CELLS. 5 foreign bodies into protoplasm usually lead to remarkably little reaction. THE EFFECT OF CHEMICAL TREATMENT AND OPERATIONS ON THE PERMEABILITY OF CELLS. Early in this investigation a general relation between per- meability and the degree of concentration of protoplasmic gels was noted. Several methods were devised for grading the con- centration of the surface and interior portions of the cytoplasm of various cells. The desired concentration gradient can be produced in the 'eggs of Asterias, Chcztopterus, Cumin gia,- and Nereis either by treating with very dilute acids, alkalies and saponin dissolved in sea water or puncturing or cutting with extremely fine Jena glass needles. It was proved that such operative treatment does not kill any portion of the egg. If the egg of Asterias be punctured, the acid dyes used penetrate the swollen area for varying depths, but never enter the normal unswollen cytoplasm. Under the conditions of my experiments only the swollen surface was penetrated and stained. It cannot be overemphasized that the concentration gradient experiment has proved an adequate test of the correctness of the membrane conception, at least for the eggs of Asterias, Cumingia, Chcetop- terus and Nereis. Dyes wrere selected which do not penetrate the normal egg. These dyes penetrate the swollen cytoplasm produced by the operation or chemical treatment, to varying depths, but never enter the unswollen cytoplasm. Some of the acid dyes do not penetrate the surface of the swollen area of a punctured egg, while others are stopped only by the unchanged cytoplasm. These results are exactly the converse of what would necessarily follow if the plasmatic membrane conception were really true. Moreover, it was noted that trypan red and erythrosin stain the surface of a .normal Chcetopterus egg, but even a great increase in concentration of the dyes does not pro- duce more than a surface staining. Here the surface is actually more permeable to these dyes than the interior of the egg. Again, the nucleus of the eggs of Asterias, Cumingia, Chcetopterus, and Nereis remain unstained when the eggs are placed in the best vital stains in use at the present time. Under this set of condi- 6 G. L. KITE. tions the nuclear membrane is impermeable to vital stains. This structure is a concentrated tough gel of relatively high viscosity and is not to be confused with hypothetical surface or vacuolar plasmatic membranes. If the concentration of a suitable vital stain be raised considerably or the nucleus be dissected out, staining occurs. Puncture of the walls and probably the cytoplasm of various types of plant cells has given unanticipated results. All the more common acid dyes enter the cells of such plants as Spiro- gyra, Elodea, Hydrodictyon, different root-hairs, and the paren- chyma cells of Tradescantia, following an extremely small puncture. Puncture of the walls of slightly plasmolyzed Spiro- gyra and Elodea cells is followed by protoplasmic staining when acid dyes are added. On the other hand, puncture of thoroughly plasmolyzed Spirogyra cells is followed at most by a slight surface staining, of the shrunken and concentrated cytoplasm, by the same acid dyes. During plasmolysis of Spirogyra a large amount of mucilag- inous material is poured out of its wall, and, as a result, this structure loses much of its rigidity, becomes softer, and permeable to many acid dyes. This striking change can be demonstrated by vital stains, dissection of the wall of thoroughly plasmolyzed cells, and by staining cells immediately after recovery from thorough plasmolysis. A very small cut in the wall of a manubrial cell of Chara gives even with trypan blue and trypan red only a local staining of the cytoplasm, which slowly spreads. THE DISSECTION OF CELLS. The chief cells that have formed the material for this study have been dissected by the use of adequate methods, and their physical properties, such as rigidity, viscosity, glutinicity, elasticity, tenacity, and colloidal state, have been determined. Although these data are too extensive to be given here, they form a part of the basis of my conclusions. CONCLUSIONS. i. The structural components of protoplasm vary greatly in their permeability to water, dyes, and crystalloids. CYTOPLASM OF ANIMAL AND PLANT CELLS. 7 2. Impermeability or partial permeability to water, dyes, and crystalloids is a property of all portions of protoplasmic gels. 3. The rate of penetration of protoplasm by dyes and crystal- loids is, in general, inversely proportional to the concentration of the living gel. 4. The best vital stains known penetrate such highly con- centrated protoplasm as the epithelial and striped muscle cells of Nee turns very slowly. 5. The interior portions of the cytoplasm of the starfish egg and probably the striped muscle cell of Necturus exhibit the same sort of osmotic properties as the surface. 6. The cell-walls, and not the protoplasm of many plant cells, prevent the entrance of dyes. SPERMATOGENESIS OF THE PIG WITH SPECIAL REFERENCE TO THE ACCESSORY CHROMOSOMES. J. E. WODSEDALEK, PH.D. ZOOLOGICAL LABORATORY, UNIVERSITY OF WISCONSIN. CONTENTS. I. Introduction 8 II. Material and Methods 9 III. General Arrangement of the Germinal Cells 10 IV. Spermatogonia 1 1 V. Primary Spermatocytes 13 1 . Resting Stage 13 2. Synizesis and Growth Period 13 3. Reduction Division 14 VI. Secondary Spermatocytes 15 1 . Dimorphism 15 2. Second Reduction Division (Equational) 16 VII. Spermatids 19 VIII. Development of the Spermatozoa 19 IX. Variation in Size of Mature Spermatozoa 25 X. Relation of the Accessory Chromosomes to Sex in the Pig 25 i. Dimorphism in the Number of Chromosomes of the Male and Female Germinal and Somatic Cells 28 XI. Summary 28 INTRODUCTION. In late years the occurrence of accessory chromosomes, in various degrees of complexity, has been recorded from diverse groups of invertebrates. Recently the field has been extended to the vertebrates, Guyer having shown their existence in the males of the guinea ('090), rooster ('09^), rat ('10), and man ('10); Newman and Patterson ('10) in the armadillo; Jordan ('n) in the opossum; Stevens ('n) in the guinea-pig; and King ('12) in Necturus. The existence of what appears to be such an element has also been reported in the ovary of the cat by Wini- warter and Saintmont ('09). In a recent paper Jordan ('13) states that the heterochromosomes are unquestionably lacking in the pig. I find, however, that the accessory chromosomes are very conspicuous in all of the pig material which I have studied. 8 SPERMATOGENESIS OF THE PIG. 9 Wilson ('09) found in Syromastes that half of the spermatids possess two more chromosomes than the remainder. It was predicted by him that in consequence the somatic cells of the female of this species would show two more chromosomes than the somatic cells of the male, and later the facts were found to be in exact accord with his predictions, the somatic cells of the female of Syromastes having been found to contain twenty-four, those of the male, twenty-two chromosomes. A similar condition has been found in other tracheates, but dimorphism in the number of chromosomes in the germinal and somatic cells of the two sexes among the vertebrates has thus far been only inferred, not actually demonstrated. The present study on the spermatogenesis of the pig was taken up at the suggestion of Professor M. F. Guyer, to whom I am much indebted for many helpful suggestions during the progress of the \vork. Points of special interest in this paper are as follows : 1. The presence of a distinct pair of accessory chromosomes. 2. The resulting dimorphic condition in the spermatozoa of the pig. 3. Dimorphism in the number of chromosomes in both the germinal and somatic cells of the male and female animals. 4. The abundance of large conspicuous interstitial cells in the testes. 5. The second reduction of the chromosome number in the secondary spermatocyte, which was found to be simply equa- tional. 6. The throwing off of a large mass of cytoplasm, containing one of the centrosomes, at the time of the final development of the spermatozoan . MATERIAL AND METHODS. I have been very fortunate in obtaining from several sources exceptionally good material for this investigation. The major part of the material studied was obtained from a vigorous Poland China boar about ten months old. The animal came from registered stock and was the property of the College of Agri- culture of the University of Wisconsin. The material was ob- IO J. E. WODSEDALEK. tained through the courtesy of Professor L. J. Cole. Immediately after the testes were removed from the live animal, pieces were placed in various kinds of fixing fluids, as Bouin's, Zenker's, Tellyesnicky's, and corrosive-acetic. Sections from various parts of the testes were made from four to twelve microns thick and while several methods of staining were employed, material fixed in Bouin's fluid and stained with Heidenhain's iron hematoxylin with acid fuchsin as a counter- stain, proved to be the most satisfactory. Delafield's hematoxy- lin with eosin counterstain was also used and gave favorable results. While material fixed in some of the other fluids men- tioned was fairly good, the first method proved to be so superior that the sections used in this study were almost wholly those prepared according to it. Other material, from a much older animal, was studied in sufficient detail to corroborate the results obtained through the study of the material secured from the younger and more vigorous animal. I also owe many thanks to Professor B. M. Allen for his generosity in placing at my disposal his embryological material of pigs of both sexes. The material in question, which was that used by Dr. Allen ('04) in his own researches on the embryology and development of the ovary and testes in mammals, was fixed in Flemming's fluid and stained with Heidenhain's iron hema- toxylin mainly. By examining a large number of these slides, I not only corroborated my previous count of the spermatogonial number of chromosomes, but was also able to make fairly con- clusive counts of the chromosomes in the somatic cells of the male, and in the oogonia and somatic cells of the female. With- out the use of his material this important phase of the problem could not have been worked out at this time. GENERAL ARRANGEMENT OF THE GERMINAL CELLS. The structure of the testes of the pig does not differ greatly from that of other well-known mammals, except for the presence of numerous masses of large well-defined interstitial cells which are scattered among the seminiferous tubules and comprise about one fourth of the entire volume of the testes (Fig. i). A detailed description of these interesting cells, which contain SPERMATOGENESIS OF THE PIG. II numerous mitochrondia, will be reserved for a separate paper. The germinal cells themselves are found in a great number of seminiferous tubules coiled throughout the interior of the testes. Sections of the tubules are found in groups of fifteen to twenty- five which are completely surrounded by walls of connective tissue. This connective tissue forms a sort of continuous net- work of walls throughout the testes and is rather thick in places, usually where three or four of the groups become appressed. Embedded in these walls are blood- and lymph-vessels which branch out into the masses of interstitial cells between the tubules. The arrangement of the cells in the tubules is similar to that of the other warm-blooded animals and the usual four types of germinal cells are present. The spermatogonia form a more or less regular layer of cells lying next to the wall of the tubule. These divide and give rise to two new cells, one or both of which may become the primary spermatocyte. The primary spermato- cytes are very abundant, especially in the spireme stage. After various changes and considerable growth the primary spermato- cytes divide and give rise to the secondary spermatocytes. The secondary spermatocytes divide in turn to form the spermatids, which transform directly into spermatozoa (Figs. 64-78). The material studied must have been in maximum activity at the time of fixation, for the four types of germinal cells were easily found in stages of growth and development. Mitotic stages of spermatogonia, primary and secondary spermatocytes, were very abundant, and frequently the entire field under the oil immersion lens was composed of cells in mitosis. Spermatozoa in all stages of development were very abundant, and clusters of the cells which were in the final stages of develop- ment could be seen attached to the long cylindrical Sertoli cells which often extend to the lumen of the tubule (Figs. 2 and 3). The Sertoli cells, which are quite abundant, are closely connected with the germinal cells and will be described in more detail later (Fig. 63). SPERMATOGONIA. The spermatogonia usually lie in a single layer next to the wall of the tubule though occasionally some of the cells are 12 J. E. WODSEDALEK. crowded out, thus forming a second layer which is always very irregular. There is considerable variation in the appearance and some variation in the size of different spermatogonia. Very frequently the cells are far apart, in which case they are flattened out on the tubule wall. The amount of cytoplasm is small and in some of the earlier stages the cell boundaries are very indistinct (Fig. 16). Later the nuclei assume a round shape and the cell wall becomes visible. The amount of chromatin material increases and the nuclei at this stage resemble very much the nuclei of the large interstitial cells both in appearance and size (Fig. 17). Two large nucleoli much alike in size seem to be present in all stages of these cells. Even in the testes of young embryos the two bodies are very conspicuous. Besides these large nucleoli a varied number of smaller, similarly staining bodies are usually present, two of which seem to occur most frequently, and evi- dently make their appearance first (Figs. 16 and 17). Later other small nucleoli appear. At present I am unable to attach any meaning to them as their numbers are so varied in the various cells. The two large nucleoli, however, are very constant, and I have been able to trace them throughout the entire spermato- genesis. From all appearances one is led to believe that these nucleoli and the accessory chromosomes are one and the same, a condition similar to that found by Guyer (*io) in man, and by others in some of the lower forms. Jordan ('n), however, has not been able to identify any such structures at this stage as the future accessory chromosomes in the opossum. At the conclusion of the resting stage numerous chromatin granules appear, which arrange themselves along fine threads in an entangled mass. The two nucleoli come in close contact and the nuclear membrane gradually disintegrates. Eighteen chro- mosomes appear in the late prophase of the spermatogonial division (Fig. 18). Sixteen of these are rod-shaped, variously curved and somewhat different in size. Two, which are always found close together, and frequently off to one side are slightly larger and oval in shape. The sixteen ordinary chromosomse or autosomes arrange themselves with the two accessory chromo- somes in the equatorial plate (Fig. 19). The accessories, like SPERMATOGENESIS OF THE PIG. 13 the autosomes, divide in this stage. A centrosphere containing a tiny centrosome can be observed in some cases, but its presence is not at discernible in these early cells as it is in the later stages. PRIMARY SPERMATOCYTES. i. Resting Stage. The primary spermatocytes which arise from the final sperma- togonial division in the early resting stage are usually somewhat smaller than the spermatogonia in the metaphase and prophase stages. The two large nucleoli are quite conspicuous and remain black in sections stained with iron hematoxylin even in very much destained material (Fig. 20). Usually two small nu- cleoli, stained the same way as the large ones, are plainly visible and occasionally only one, or more than two of these bodies appear. 2. S'ynizesis and Growth Period. An increase in the bulk of both the nucleus and cytoplasm begins after a brief period of rest. The chromatin appears to be arranged in much the same wray as it was during the resting con- dition of the spermatogonia, except that the linin fibers are coarser. The cytoplasm is composed of granular masses and clear, apparently liquid areas. The chromatin threads and nucleoli become massed at one side of the nucleus (Fig. 21). The nuclear wall expands and the clear area formed by the massing of the chromatin to one side becomes much enlarged. The clear areas in the cytoplasm decrease coincidentally with the increase of the liquid-like area in the nucleus and one is led to believe that the cytoplasmic fluid permeates the nuclear wall during the period of synizesis. During the collapse of the chromatin material, the nucleoli can be plainly seen especially in well destained sections (Fig. 21). They are usually in that portion of the chromatin mass which is nearest to the nuclear wall. Opposite this mass on the outside of the nuclear wall a centrosphere can often be seen, but the centrosome is rarely visible at this stage. The chromatin threads become arranged in a very much tangled mass of loops which later appear in about half the original number and fully twice as thick. (I have not been able to determine definitely whether 14 J. E. WODSEDALEK. the pairing takes place by telo- or parasynapsis.) The whole mass then moves toward the center (Fig. 22), the large clear area in the nucleus disappears, and the nuclear wall becomes spherical and clearly defined (Fig. 23). While the formation of the spireme is going on, the cytoplasm, too, increases greatly in volume and appears very granular. The cell now is fully twice as large as a spermatogonium and judging by the large number of cells that can be seen in the spireme stage one can safely conclude that they remain in that condition for some time. The spireme finally breaks up into U and variously shaped chromo- somes (Fig. 25). The two large nucleoli, which remain in full view throughout this stage, become oblong and can often be seen close together (Fig. 25). 3. Reduction Division. The primary spermatocytes when ready for division reveal ten chromosomes in the late prophase or early metaphase stage. The two accessories which are ordinarily off to one side (Figs. 8, 26, 27, and 29) can be recognized at a glance. The other chromo- somes are usually arranged in a ring. Judging from their large size and changed form, they are bivalent, representing the paired univalent chromosomes of the spermatogonium. That is, of the original eighteen chromosomes, sixteen have paired to form eight bivalents of the primary spermatocyte and two have remained unpaired as the accessory chromosomes. Sometimes, although ten chromosomes can be counted, it is difficult to tell just which are the accessories owing to overlapping. The chromosomes differ somewhat in size as can be seen in Figs. 26 to 29. Figs. 9 to 12 and 30 to 37 show the accessories in characteristic posi- tions in the metaphases of division of the primary spermatocyte. They always pass entire, side by side and in advance of the divided autosomes, toward one pole. This is possibly due to the fact that they are not retarded by division. The fact that the chromosomes immediately after divergence (Fig. 28) resume the appearance (except in size) that characterizes the univalent spermatogonial chromosomes, and also because the accessory chromosomes pass over entire to one pole in this division while they are halved in the next division, seems to indicate strongly that this is the reduction division. SPERMATOGENESIS OF THE PIG. 15 It is obvious that as far as chromatin content is concerned the division of the primary spermatocyte gives rise to two dissimilar cells, one of which receives eight chromosomes, and the other eight plus the two accessories or ten chromosomes (Fig. 38). Figure 39 is a drawing of one end of a late anaphase of such a division showing eight chromosomes, and Fig. 40 shows the other end which received eight of the ordinary chromosomes and the double accessory. Occasionally a small chromatin body is present in this first spermatocytic division (Figs. 28, 31, 32, 35 and 37). Figure 31 shows such a body passing to the same pole with the accessories, in advance of the other chromosomes. Figure 32 represents an earlier stage of much the same thing. In Fig. 35 it can be seen passing to the opposite pole, and Fig. 37 represents an extremely rare case where two such bodies are present, one somewhat larger, passing to either pole, even in advance of the two accessory chromosomes. Figure 28 shows the body outside of the main ring of chromosomes. While the small body can be seen fre- quently, as a rule no such an element can be detected, and while it may possibly be comparable to the small pair of chromosomes found so constantly in some of the Tracheata, my present data on its irregular occurrence and behavior do not permit a conclusion regarding its significance. SECONDARY SPERMATOCYTE. i. Dimorphism. The dimorphism of the secondary spermatocytes, which re- sulted from the last division, is again expressed in the resting stage that sometimes follows. Approximately half of them showed, under proper decolorization, two large chromatin nucleoli (Fig. 41) while in the others only the small nucleoli appeared (Fig. 42). The nucleoli retain the usual deep staining capacity, and as was true in the previous stages, even when all the other material is almost totally decolorized, the nucleoli remain very conspicuous. Frequently, both primary and secondary sperma- tocytes were found dividing in the same field, which fact seems to suggest that at times there is no intervening period of rest be- tween the two divisions, or that it is very brief. Figure 43 16 J. E. WODSEDALEK. shows the two resulting cells of a primary spermatocyte division which are still in close contact and both ready for dividing into spermatids. Resting stages, however, appear in abundance and in this stage, as was the case in the resting stage of the primary spermatocyte, the centrosome surrounded by a clear zone becomes large and very conspicuous (Figs. 41 and 42). 2. Second Reduction Division (Equation al] . Although the division of the primary spermatocyte gave rise to cells containing eight and ten chromosomes respectively (Figs 38, 39, 40), when these cells become ready for division half of them show four (Figs. 13, 14, 43, 45, 46) and the other half six chromosomes (Figs. 43, 44, 47). Thus a second pairing of the ordinary chromosomes, similar to that found by Guyer in the pigeon ('oo), corroborated by Geoffrey Smith ('12), man ('10), guinea ('090.), and chicken ('096), and by Jordan in opossum ('11), has evidently taken place so that there are four bivalents in each type of cell and the additional two accessory chromosomes in the one type. Stevens ('n) says that there is no such second synapsis or numerical reduction in the guinea-pig. Figures 45 and 46 show four large chromosomes in the metaphase stage of one type of secondary spermatocyte and Figs. 44 and 47 represent the other type in which six chromosomes appear, four bivalent autosomes plus the two accessories. The four chromosome group is evidently formed by the pairing of the eight chromo- somes of one type of cell resulting from the first maturation divi- sion at one pole, and six chromosome group is interpreted as being derived by the pairing of the eight chromosomes plus the two unpaired accessories at the opposite pole (Figs. 38, 39, 40). Guyer ('10) in speaking of the second conjugation of the chro- mosomes in man says : "Assuming that the respective chromosomes are more or less qualitatively differentiated, such a numerical reduction, however, by no means necessarily implies that there has also been a second qualitative reduction. Aside from the improbability of such a reduction, the general appearance of the divided chromosomes would not warrant this interpretation; for instead of the elon- gated univalent type as seen in the spermatogonia, or in ana- SPERMATOGENESIS OF THE PIG. l"J phases of the divisions of spermatocytes of the first order, the daughter chromosomes here retain the rounded appearance and increased size that is characteristic of the bivalent types (com- pare Figs, i, 10, and n, 13, 14, 15, 16 and 17). Thus while half of the spermatids receive five, and half seven chromosomes, in terms of univalence the numbers would in all probability be ten and twelve respectively." Jordan ('n) in speaking of the same condition in the opossum remarks as follows: "Similarity of form between the chromosomes of the first and second metaphase plates (i. e., double rods) suggests a similar manner of division: accordingly a second reduction. When one recalls, however, that a resting stage (Figs. 39 to 41) usually intervenes between the first and second maturation divisions, when the chromosomes pass through a reticular phase, the above conclusion is inadmissible; or rather, no definite conclusion respecting the character of the second division is justified. A double true reduction, suggested by the form of the chromosomes is contrary to our fundamental conceptions regarding the signifi- cance of chromosomes and need not, in view of the nature of the evidence, be seriously considered. Moreover, in the stage just preceding the brief resting phase of the spermatids (Figs. 57, 58 and 59) there occurs a resolution of the five chromosomes into nine and of the four into eight. This demonstrates that the true character of the second division is equational. The second numerical reduction involves a less close union apparently than the first, as a comparison of illustrations 29 and 43 will show. Again the fusion is sometimes incomplete to the extent of giving an occasional count of six chromosomes." My own belief that the real character of the second division in the case of the pig is simply equational is based on a number of facts. Even after the first maturation division the ordinary chromosomes are much larger than the spermatogonial univalents. This increase in size has evidently taken place during the growth period of the primary spermatocyte (compare Figs. 1 8 and 38). During the prophase of the secondary spermatocyte the chromo- somes apparently increase still more in size and four large bi- valent autosomes appear for division in the one type of cell and 1 8 J. E. WODSEDALEK. four plus the two accessories in the other type. Not only is the bivalent nature of these large autosomes conspicuous, but very frequently a quadrivalent character is discernible and they are much larger than the accessories (Figs. 44, 45, 46 arid 47). During the anaphase the bivalent nature of the chromosomes is often clearly visible at each pole. Figure 49 shows four bivalents and is a drawing of one pole of the dividing cell which received eight chromosomes during the first maturation division. Figure 52 is a drawing of one pole of the division of the type of secondary spermatocyte which received eight ordinary chromosomes and the double accessory. It will be seen that four of these are bi- valent in nature, while two are univalent, the two univalents being the results of the division of the two accessories (Figs. 44, 50 and 51) both of which have here divided for the first time since the spermatogonial division. The bivalent nature of the autosomes after the second spermatocytic division becomes even more conspicuous in cases where these chromosomes divide into two before they break up. Figure 56 shows an early spermatid cell which received the four bivalent autosomes and the two accessories. It can be seen that two of the autosomes have almost completely divided. The bivalent nature of the other two can also be plainly seen, wiiile the accessories retain their univalent appearance. Thus it can be seen that the spermatids produced by this second division do not receive four and six chromosomes respectively, but, four bivalents, or the equivalent of eight univalents, are present in the one type of cell, and four bivalents or eight univalents together with the two accessories in the other. The foregoing facts seem to indicate beyond doubt that the second maturation division is not a reduction but simply an equational one. Figures 50 and 54 represent an anaphase of division in a secondary spermatocyte showing two streaks of lagging chro- matic material. Although several such cases were observed ordinarily no such pronounced streaks occur (Fig. 51). Guyer found a similar condition in the secondary spermatocyte division in man and suggests that it may be the two accessories lagging behind. Although many accurate counts of the chromosomes in this second division were made, they were so massed together SPERMATOGENESIS OF THE PIG. 1 9 as to render a count impossible (Fig. 54). Not infrequently the chromosomes in the early spermatids become closely appressed before they begin to break up (Figs. 53 and 57). Sometimes the individuality of the chromosomes is apparent even after the nuclear wall begins to form (Fig. 58). SPERMATIDS. From the foregoing evidence it is obvious that there exists a dimorphism among the spermatids, one type containing eight and the other ten chromosomes after the last division. The chromosomes soon become irregular in shape and begin to break up, and the spermatids appear all alike except that immediately after the chromosomes disintegrate two nucleoli are visible in approximately half of the cells (Figs. 64 and 65). These retain for a short time the staining capacity characteristic of the nucleoli of the previous stages, but later disappear and the whole nucleus assumes a coarse granular appearance. The cells seem- ingly remain quiescent in that condition for some time as they are nearly always found in large numbers. The centrosome sur- rounded by a clear layer is again very conspicuous within the comparatively small centrosphere which lies close to the nucleus (Figs. 64 and 65). DEVELOPMENT OF THE SPERMATOZOA. Following the period of rest the spermatids begin to develop into spermatozoa. There is no perceptible change in the size of any of the parts, such as described by Jordan in the opossum. The first change to be detected is the extrusion of the centrosome, surrounded by a light area, from the small sphere (Fig. 66). It takes a position a short distance from the nucleus and soon begins to divide. The two new centrosomes move apart and one, somewhat rod-shaped, comes in contact with the nuclear wall but it remains connected by a mass of material with the other centro- some which assumes a disk shape and for a while remains in place (Figs. 67, 68 and 69). The disk is frequently perforated in the middle and has the appearance of a ring (Fig. 70). Simultaneously with the division of the centrosome, a portion of the sphere which remains in close contact with the nuclear wall 2O J. E. WODSEDALEK. migrates to a point directly opposite the centrosome in contact with the wall on the other side (Figs. 66, 67 and 68). The sphere in this process of migration is in such close contact with the nuclear wall that a slight depression can be detected in the latter. The depression is even more pronounced when the clear sphere becomes definitely fixed as the small acrosome (Figs. 68-75). In the guinea-pig the acrosome, which becomes nearly as large as the nucleus itself, according to Meves ('99), is also formed from the centrosphere or idiozome. While in the rat, according to Lenhossek, it is independently formed in the cytoplasm without relation to the preceding mitotic figure or the centro- somes. Meves ('97) also found that in the salamander the acrosome is formed from the idiozome which wanders around the nucleus to its anterior pole. McGregor's results on Am- phiuma ('99) agree in general with those of Meves, except that here the acrosome arises from only a part of the centrosphere, while a second smaller part passes to the base of the nucleus and forms the main part of the middle-piece. Coincidentally with the division of the centrosome and the migration of the small sphere the nucleus together with these structures moves to one side of the cell in the direction of the acrosome and soon prac- tically all of the cytoplasm is at the posterior end of the cell (Figs. 66, 67, 68, 69 and 70). The cell wall seems to persist as a thin mantle covering the head of the spermatozoan. The anterior part of the cell, bearing the acrosome, almost invariably points in the direction of the tubule wall while the mass of cytoplasm extends into the lumen (Fig. 2). Most of the chromatic material of the nucleus gathers at the center into a dense mass which has the same staining capacity as characterizes the chromosomes and the centrosome. Sometimes two or three masses are present (Figs. 73, 74 and 75). Soon after the division of the centrosome into a cylindrical anterior and a disc-shaped posterior body, a divergence of the two follows, but they remain connected by a streak of material (Fig. 69). The inner body upon coming in contact with the nuclear wall forms a depression in the latter and its anterior portion shapes into a disc which comes in close contact with the nuclear wall in the small depression at the posterior end of the SPERMATOGENESIS OF THE PIG. 21 nucleus. From this disc, which is apparently the end-knob, a cylindrical mass of material extends backwards uniting the two centrosomes (Figs. 68 and 69). A tiny filament extending back- ward from the middle of the posterior disc-shaped centrosome makes its appearance and is undoubtedly a continuation of the coarser filament which unites the two centrosomes (Fig. 69). As the tail grows longer the connecting filament becomes thinner, and the posterior centrosome, wrhich after division was disc- shaped and had increased somewhat in size, becomes trans- formed into a ring and it can be seen that the filament extends directly through it and continues backward as the tail (Fig. 70). The formation of the ring takes place simultaneously with the rapid growth of the tail and one is led to believe that the perfora- tion in the disc is partly due to the fact that the material formerly occupying that space goes to help in building up the axial filament. Shortly after the tail projects out of the cell the ring moves along the filament, backward, and soon swerves over to one side in the cytoplasmic mass (Figs. 71 and 72). Very frequently it can be seen a considerable distance away from the axial filament long before the latter is fully developed (Figs. 73, 74 and 75). This seems to indicate that the posterior centrosome takes no further part in the development of the filament and that the latter is mainly developed from the anterior centrosome, no part of which is discarded or thrown off. While most of the inner centro- some passes into the formation of the axial filament a part of it remains as the end-knob in the small middle-piece (Figs. 76 and 78), a condition similar to that found by McGregor in Am- phiuma ('99). The posterior ring-shaped centrosome, after moving away from the filament, sometimes divides (Fig. 73), but usually assumes a spherical shape (Fig. 74) and an interesting point in connection with this body is that during the final development of the spermatozoan it is invariably thrown off with a big mass of cyto- plasm (Figs. 74, 75 and 76). The casting off of a portion of the cytoplasm during the last stages of the developing spermatozoan has been described by Meves ('99) in the guinea-pig where it is closely similar to the process which occurs in the spermatozoid- formation in ferns; but the throwing off of a portion of the 22 J. E. WODSEDALEK. centrosome together with the mass of cytoplasm has not hereto- fore been, to my knowledge, recorded. The body usually stains as deeply as it does in the earlier stages for some time after the comparatively large mass of cytoplasm containing it is completely separated from the rest of the cell or young spermatozoan. In general, the behavior of the centrosomes in the development of the spermatozoan of the pig does not differ greatly from the conditions found in some of the other vertebrates. However, it might be well to briefly point out some of the differences. In the spermatids of the salamander, according to Meves ('97), the two centrosomes lie quite at the periphery of the cell and from the outer one grows out the axial filament. The two centrosomes leaving the idiozome by which they are first surrounded, now pass inwards toward the nucleus, the outer one meanwhile becoming transformed into a ring while the axial filament passes through it to become attached to the inner centrosome. The latter pushes into the base of the nucleus and enlarges enormously to form a cylindrical body comprising the main portion of the middle-piece. The ring divides into two parts, the anterior of which gives rise to a small body at the posterior end of the middle-piece identical with the end-knob. The other part of the ring wanders out along the tail and finally lies at the limit between the main part of the latter and the end-piece. In the pigeon, according to Guyer ('oo), the centrosome divides and moves out of the sphere and further away from the nucleus. The two new centrosomes move apart, but remain connected by a mass of material which later disappears and a very delicate fibril uniting the two centrosomes exists in its place. One of the centrosomes enlarges and transforms into a complete ring. The connecting fibril can later be seen passing from the smaller centrosome back through the ring and outside the cell. The centrosomes next approach the nucleus and as they draw near a slight invagination appears in the nuclear wall and the small centrosome moves into it. In the mammals, the work of Lenhossek on the rat ('98) and Meves on the rat, guinea-pig and man ('98, '99) gives a result similar to the condition found in the salamander. In all these mammals the young spermatids contain two peripherally placed SPERMATOGENESIS OF THE PIG. 23 centrosomes, from the outer one of which the axial filament grows out and the centrosomes later move toward the nucleus. In the pig the centrosomes are never peripherally located and division of the spermatid centrosome does not take place until it emerges, surrounded by a small clear sphere, from the main bulk of the centrosphere which is in close contact with the nucleus. It can be seen from the drawings of the later stages that the nucleus gradually elongates and flattens (Figs. 67-78). The large chromatic mass within the nucleus disintegrates and the particles become distributed at the periphery of the nuclear wall (Fig. 78). The tail envelope becomes apparent as the cell elongates and evidently develops from the cytoplasm (Figs. 74- 78). It later comes in contact with the axial filament and en- velops about half of its entire length. The remaining extremely thin portion of the tail is apparently the naked axial filament (Figs. 76, 77, 78). When the cells reach the stage represented in Figs. 71-73 they attach themselves in bunches to the large cylindrical Sertoli or nurse cells that often extend from the basement membrane a long distance toward the lumen of the tubule (Figs. 2 and 63). As the spermatozoa continue to develop a gradual decrease in the volume of the cytoplasmic mass of the Sertoli cells is noticeable. When the spermatozoa are apparently mature they abandon the Sertoli cells which at this time are very much collapsed. The spermatozoa after leaving the nurse cells do not pass directly to the lumen of the tubule, but remain scattered for some time among the masses of cytoplasm which were cast off by the same developing cells a short time before (Figs. 3 and 67). This cast- off material does not scatter about in the lumen, as one would suspect, but forms a sort of loose layer next to the cells following the maturing spermatozoa. The masses begin to disintegrate soon after they are discarded by the developing spermatozoan, and when stained with iron-haematoxylin numerous black bodies make their appearance in this material (Fig. 3). The spermato- zoa remain, with their bodies embedded in this layer, and tails extending into the lumen, until the mass almost completely dis- appears. Then they become free in lumen of the tubule and are ready to make their way out of the testes. During this period 24 J- E. WODSEDALEK. of attachment to the nurse cells and later to the discarded mass of cytoplasm the heads of the spermatozoa increase somewhat in size but retain the same staining capacity. The foregoing facts seem to show beyond doubt that the developing spermato- zoan derives nourishment not only from the nurse cells but also from the cast-off cytoplasmic material. When the sperm reaches maturity the acrosome disappears from view and the anterior edge of the head becomes well rounded. In addition to the parts mentioned, there appears very fre- quently a dense spherical" cytoplasmic mass represented in Figs. 67 to 75. As to the origin and fate of this body I am not entirely certain. It seems, however, that it is a portion of the centro- sphere, for in many of the early stages it was seen in the immediate neighborhood of the small body which later forms the acrosome (Fig. 67). Further evidence for this assumption is the fact that the combined mass of these two bodies seems to be equal to the size of the sphere in the stage immediately preceding (Fig. 66). The consistency of this body, too, is similar to that of the sphere. Figure 78 represents a mature spermatozoan of the pig which appears rather simple in structure and form. The entire nucleus of the spermatid has evidently developed into the head, which is oblong and flat (Figs. 70-78). The nuclear material breaks up into very fine particles which arrange themselves in a layer at the entire periphery of the head wall (Fig. 77). With Heidenhain's iron-hsematoxylin the head stains a sort of slate blue and is difficult to decolorize. A depression is present at the posterior extremity which is in contact with a small middle piece. The end-knob which is for some time attached to the nuclear wall within the small depression (Figs. 74, 75 and 76) finally breaks loose and passes to the posterior extremity of the middle piece, the greatest portion of which now remains very clear in appearance. This clear area is similar in appearance to the contents of the acrosome and one is led to believe that it is due to the presence of the substance which was seen in form of a clear sphere surrounding the centrosome after it emerged from the centrosphere and persisted about the two centrosomes, particularly around the inner one, after division (Figs. 66-69) SPERMATOGENESIS OF THE PIG. 60; sol 40 30 20 10 •m.m.-~ 10.5 11 US 12 12.5 13 135 14 14.5 15 15.5 16 FIG. i. Diagram showing the variation in size among four hundred mature pig spermatozoa. Figures at the left give the numbers of the individuals belong- ing to each type. Figures at the bottom give the lengths of the heads of the spermatozoa in millimeters, magnified two thousand times. VARIATION IN SIZE OF MATURE SPERMATOZOA. The spermatozoa of the pig vary considerably in size, and careful measurements revealed the fact that they may be arranged in two separate classes, one type being much larger than the other. Mature specimens, which were free in the lumen of the tubule and parallel to the objective, were selected at random, and outline sketches of four hundred heads enlarged (X 2,000) were made with the aid of a camera-lucida. The lengths of the sketches were then carefully measured and recorded in half millimeters. Figure I in the text shows the variation in size of the four hundred heads measured. It can be seen at a glance that two separate types of spermatozoa exist ; the greatest num- ber of the one kind measuring from 11.5 to 12 mm., and of the other type from 14 to 14.5 mm. The increased size of the latter is due presumably to the presence of the accessory chromo- somes. RELATION OF THE ACCESSORY CHROMOSOMES TO SEX IN THE PIG. Many investigators working on the problem of sex determina- tion are of the opinion that the sex of a zygote is determined at the time of fertilization, the spermatozoan in most cases carrying 26 J. E. WODSEDALEK. the determining factor. This view has been recently substan- tiated by the careful investigations of McClung, Stevens, Wilson, Morgan, Payne and others, who found that in many invertebrates such as certain insects, myriopods, arachnids and nematodes, two different kinds of spermatozoa are produced. The spermatozoa differ from each other by one chromosome, or by a group of chromosomes, called the accessory chromosomes, or X-chromo- somes. In some forms it was found that the spermatozoa each contain an accessory chromosome, but that in half the sperms this chromo- some was larger in its chromatin content than in the other. These are called the X and Y elements respectively. All the eggs produced (except in a few cases), on the other hand, contain the X-element. It has been shown in some of the Tracheata that whenever an egg is fertilized by a sperm containing the X-chromosome it develops into a female, while an egg fertilized by a sperm without the X-element, or by one containing the Y-element, gives rise to a male individual. In the vertebrates, Guyer was able to identify an X-element in such diverse forms as the guinea, rooster, pigeon, rat, and man, and other investigators have recently brought forth similar evidence. Patterson and Newman ('10) record its presence in the armadillo, Jordan ('n) in the opossum, Stevens ('n) in the guinea-pig, and King ('12) in Necturus. Up to the present in- vestigation, however, a dimorphism in the number of the chromo- somes in the male and female vertebrates has not been recorded. Guyer ('10) after speaking of the difference in the number of chromosomes in the somatic cells of the male and female Tracheata which possess the X-elements says: 11 In the light of these facts we should expect the somatic cells of man to contain twenty-two and of woman, twenty-four chro- mosomes. The tissues of the female have not yet been studied with this in mind. Flemming ('97) records the somatic number of chromosomes, determined from corneal cells, as twenty-four but unfortunately he does not record the sex of the subjects from which the material was obtained. If it were a female his count would bear out the interpretation given above." SPERMATOGENESIS OF THE PIG. 2J The behavior of the accessory chromosomes in the various stages of the spermatogenesis of the pig further substantiates the fact that two different kind of spermatozoa are produced in mammals. Since the dimorphic condition of the spermatozoa does exist, the question arose whether this dimorphism holds true in the chromosome number of the male and female somatic cells. With this in view I undertook an examination of sections of male and female embryological material supplied me by Professor B. M. Allen. A large number of these sections were studied and although mitotic stages were not very abundant in any one section, and the chromosomes were frequently bunched together rendering a count impossible, nevertheless, a number of counts of the chromosomes in the germinal and somatic cells of the two sexes were recorded. It was again found that the spermatogonial number of chromosomes is eighteen, and that the same number prevails in the somatic cells found in the mesonephros of the male embryos. Two of the chromosomes are usually somewhat larger (Fig. 59). Sections of the female material revealed a count of twenty chromosomes in the oogonia (Fig. 60), and the same number prevails in the somatic cells found in the mesonephros of the female embryos (Fig. 62). Four large chromosomes corre- sponding to the two accessories in the male can usually be de- tected in the oogonial cells. In the somatic cells of the female there are also four large chromosomes present corresponding to the two accessories in the male cells and the four accessories in the oogonia (Fig. 62). In a few cases ten large chromosomes were found in the early metaphases of division in the somatic cells of the female (Fig. 61). Two of these are considerably larger and are interpreted as the result of the pairing of the four accessories. The smaller chromosomes, eight in number, judging by their size, are evidently due to the pairing of the sixteen other chromosomes. The position of these cells would not warrant the supposition that they might be wandering germ cells. Furthermore, a similar condition was not observed among the oogonial cells in spite of the fact that mitotic stages occur far more frequently in the sections of the ovaries than they do in the mesonephros. In proportion to the large number of mitotic stages found in 28 J. E. WODSEDALEK. these long-continued searches, accurate counts of chromosomes in the cells of the various tissues were comparatively few. How- ever, all cases where a count was possible were recorded and although they varied somewhat, the above number of chromo- somes, eighteen and twenty, attributed to the male and female cells respectively, were the prevailing numbers. The foregoing facts are in accord with the expectations, and we have here in a vertebrate a condition verifying the results found in some of the lower forms. In view of these facts it is obvious that the eggs, carrying ten chromosomes, or half the somatic number, when fertilized by a spermatozoan containing ten chromosomes, give rise to an individual containing twenty chromosomes in its -cells, or a female. Those fertilized by the other type of spermatozoan, which contains only eight chromo- somes, give rise to individuals with eighteen chromosomes in their cells, which was found to be true in the male. The results of the present investigation, therefore, add support to the chromosome theory of sex determination, since they show that in the vertebrates, as well as in some of the lower forms, there exists a dimorphism in the number of chromosomes in the somatic as well as the germinal cells of the two sexes. It is highly probable that conditions similar to those found in the pig, as regards sex determination, exist in man and in the other vertebrates which possess the accessory or X-chromosomes. The resemblance in the behavior of the pair of accessories in the pig and their behavior in man is very striking and suggests that in all probabilities there exists a dimorphism in the germinal and somatic cells of man and woman. SUMMARY. 1. The usual four types of cells, the spermatogonia, primary and secondary spermatocytes, and spermatids are discernible in the spermatogenesis of the pig. 2. Large interstitial cells containing numerous mitochondria exist in great abundance and comprise about one fourth of the entire mass of the testes. 3. Numerous Sertoli or nurse cells are present and great numbers of spermatozoa can be seen in all stages of development in the various tubules. SPERMATOGENESIS OF THE PIG. 29 4. Two large nucleoli are present in the resting stages of the spermatogonia. These can be traced through the entire sper- matogenesis of the pig and are apparently correlated with, or are the same thing as the two accessories. Smaller nucleoli, usually two, are also present. 5. Eighteen rod-shaped chromosomes differing somewhat in size occur in all of the spermatogonia where a definite count could be made. Two of these, undoubtedly the accessories, can usually be seen to one side of the main mass of chromosomes. 6. During the spermatogonial division the accessories divide and occasionally pass to the poles in advance of the other chro- mosomes. 7. The last spermatogonial division gives rise to cells which through a process of growth become the primary spermatocytes. Both nucleus and cytoplasm increase greatly in size. Synizesis followed by synapsis occurs. 8. An apparently continuous spireme is formed which later breaks up into U-shaped and variously curved chromosomes. A centrosphere containing the centrosome is present. 9. The two large round nucleoli which remain very conspicuous throughout the process of growth of the primary spermatocyte come together and elongate during the late prophase. 10. Ten chromosomes appear for division in the primary spermatocyte of which eight are evidently bivalent and two accessory. The accessories are usually considerably to one side of the other chromosome. 1 1 . The two accessory chromosomes which are out of the main spindle pass undivided to one pole in advance of the other chromo- somes. 12. The primary spermatocyte division is evidently the reduc- tion division. It gives rise to two cells, one of which contains ten (eight autosomes plus the two accessories) and the other only eight chromosomes. 13. The chromosomes in these daughter cells are larger than the univalents of the spermatogonia and show some indications of bivalence. 14. The secondary spermatocytes formed by the last division ordinarily go into a resting stage. In half of these the two 3O J. E. WODSEDALEK. nucleoli are conspicuous and lacking in the others. In a few cases, at least, there is no resting period in the secondary sper- matocytes. A small centrosphere containing a relatively large centrosome is present. 15. During the metaphase of the secondary spermatocyte one half show four large bivalent chromosomes, and the remaining show the four large bivalents plus the two accessories or six chromosomes. A second pairing has apparently taken place but the division is simply equational as the four large chromo- somes often manifest a quadrivalent character. The accessories remain unpaired. 1 6. The quadrivalent nature of the autosomes in the secondary spermatocytes becomes all the more certain after division, as the chromosomes passing to the poles are of the bivalent nature. 17. The type of secondary spermatocyte which received eight chromosomes after the first maturation division gives rise to two spermatids each containing four bivalent or eight univalent chromosomes. The other type which received ten chromosomes after the first maturation division gives rise to two spermatids, each containing four bivalent or eight univalent chromosomes and the two accessories, each accessory having divided here for the first time since the spermatogonial division. 1 8. The dimorphic nature of the spermatids which develop into spermatozoa is further evinced by the presence of the two nucleoli in approximately half the cells. The centrosome is again conspicuous. 19. The first noticeable change in the transformation of the spermatid is in the centrosome. It emerges from the small sphere and divides into an anterior rod-shaped and a posterior disc- or ring-shaped body. 20. Both of the centrosomes contribute to the development of the axial filament. A portion of the anterior one persists as the end-knob, while the posterior body is finally cast off. 21. A portion of the sphere migrates to the opposite side and gives rise to the acrosome which disappears when the spermato- zoan is fully developed. 22. The nucleus of the spermatid passes to one side of the cell, elongates and flattens and forms the head of the spermatozoan. SPERMATOGENESIS OF THE PIG. 3! 23. The axial envelope which extends a little more than half the length of the tail is developed from the cytoplasm. 24. The half developed spermatozoa attach themselves to Sertoli cells where they continue to develop. 25. The cytoplasmic mass of the Sertoli cells decreases greatly in size as the spermatozoa continue to develop. Finally, when the spermatozoa are almost mature, a collapse of the nurse cells takes place when practically nothing but the nucleus and the cell wall remain. 26. An interesting event occurs in the final stages of the devel- opment of the spermatozoan when a large mass of cytoplasm together with the posterior centrosome is thrown off by the cell. 27. When the spermatozoa desert the Sertoli cells they do not pass directly into the lumen of the tubule but remain scattered with their heads embedded in the layer of the cast-off cytoplasm. 28. The cast-off cytoplasmic material is apparently used as food by the maturing spermatozoa and when practically all of the former disappears the fully developed spermatozoa become free in the lumen of the tubule. 29. Measurements of the mature spermatozoa reveal the fact that they are of two distinct sizes, a point no doubt correlated with the presence and absence of the accessory chromosomes. 30. The somatic cells of the male contain eighteen chromo- somes, a number corresponding to that in the spermatogonia. 31. The somatic cells of the female yield a count of twenty chromosomes as do the oogonia. 32. The foregoing facts add considerably to the support of the chromosome theory of sex determination, since they prove that in the vertebrates, as is true of some invertebrates, there exists a dimorphism in the germinal and somatic cells of the male and female. LITERATURE CITED. Allen, B. M. '04 The Embryonic Development of the Ovary and Testis of the Mammals. The American Journal of Anatomy, Vol. III., No. 2, pp. 89-153. Guyer, M. F. 'oo Spermatogenesis of Normal and of Hybrid Pigeons. Dissertation: Univer- sity of Chicago, 1900. 'oga The Spermatogenesis of the Guinea. Anat. Anz., Bd. XXXIV., Nr. 20. The Spermatogenesis of the Domestic Chicken. Anat. Anz., Bd. XXXIV., Nr. 22. 32 J. E. WODSEDALEK. / '10 Accessory Chromosomes in Man. Biol. Bull., Vol. XIX., No. 4. Jordan, H. E. 'n The Spermatogenesis of the Opossum (Didelphys virginiana) with Special Reference to the Accessory Chromosome and the Chondriosomes. Archiv fiir Zellforschung, 7. Band, i. Heft. '13 A Comparative Study of Mammalian Spermatogenesis with Special Ref- erence to the Heterochromosomes. Science, N. S., Vol. XXXVII., No. 946, pp. 270-271. King, Helen Dean '12 Dimorphism in the Spermatozoa of Necturus maculosus. The Anatomical Record, Vol. 6, No. 10. von Lenhossek, M. '98 Untersuchungen iiber .Spermatogenesis. A. m. A., LI. McGregor, J. H. '99 The Spermatogenesis of Amphiuma. J. M., XV., Suppl. Meves, F. '97 Uber Struktur und Histiogenese der Samenfaden von Salamandra. Ibid., L. '98 Uber das Verhalten der Centralkorper bei der Histogenese der Samenfaden vom Mensch und Ratte. Vehr. An. Gess., XIV. '99 Uber Struktur und Histogenesis der Samenfaden des Meerschweinschens. A. m. A., LIV. Newman, H. H., and Patterson, J. T. '10 Development of the Nine-banded Armadillo, from the Primitive Streak Stage to Birth; with Special Reference to the Question of Specific Polyembryony. Journ. of Morph., Vol. 21, No. 3. Smith, Geoffrey '12 Studies in the Experimental Analysis of Sex. Part 9 — On Spermatogenesis and the Formation of Giant Spermatozoa in Hybrid Pigeons. Quart. Journ. Micr. Sci., Vol. 58, part i. Stevens, N. M. 'n Heterochromosomes in the Guinea-Pig. Biol. Bull., Vol. XXL, No. 3. von Winiwarter, H., et Saintmont, G. '09 Nouvelles recherches sur 1'ovogenese et 1'organogenese de 1'ovaire des Mammiferes (Chat). Arch. Biol., Tom. XXIV. Wilson, E. B. '09 Studies on Chromosomes, IV., V. Journ. Exp. Zool., VI. 34 J- E. WODSEDALEK. EXPLANATION OF PLATES. PLATE I. (Fig. i, X20o; Figs. 2 and 3, Xsoo; Figs. 4-12, Xi,20o.) 1. Section of a pig testis showing the seminiferous tubules separated by wide areas of interstitial cells. 2. Section of a tubule showing bunches of developing spermatozoa attached to Sertoli or nurse cells. 3. Section of a tubule showing almost fully developed spermatozoa which have deserted the Sertoli cells and became scattered among the masses of cast-off cyto- plasm. 4. Resting stage of a spermatogonial cell showing two large nucleoli. 5. Primary spermatocyte in the synizesis stage. 6. Resting stage of a primary spermatocyte showing the two large nucleoli. 7. Spireme stage of a primary spermatocyte showing the chromatin threads and the two persisting nucleoli. 8. Late prophase of division in a primary spermatocyte showing eight chromo- somes in the main group, and the two accessories lying off to one side. The photograph does not reveal the full size of the accessories. 9 and 10. Side views of primary spermatocytes showing the cell ready for di- vision with the accessories lying just above the level of the regular equatorial plate of chromosomes. 11. Side view of a primary spermatocyte division with the two accessory chromosomes passing to the pole, side by side, in advance of the other chromosomes. 12. Side view of a primary spermatocyte division showing the two accessories at the pole while the autosomes are still in the equatorial plate undivided. 13. Two metaphases of secondary spermatocytes showing four chromosomes in the equatorial plate. The figure at the top is a side view, and the one at the bottom a polar view. 14. The cell at the top is one type of secondary spermatocyte which contains four chromosomes in the metaphase stage and the cell below is a spermatid resulting from the division of a type of secondary spermatocyte which received eight auto- somes plus the two accessories. 15. Resting stage of a spermatid showing two nucleoli. BIOLOGICAL BULLETIN, VOL. XXV. PLATE I. /*• . ** "• f *•"!*•• >* ' "' iTC«*fc *y>5 •;-«*i^ ¥ *^ k>. PA* •* k *^ c*«. * IfrA^fe ^Sl « I r**+ *** _e»*l^^ 'V*f-r v. ****V7-"' ^ .' « '^ ' **_ s ;^^^^ - -%?%.*•£.>•-'» J. E. WODSEOHLEK 36 J. E. WODSEDALEK. PLATE II. (All of the drawings on plates 2-6 were made with the aid of a camera lucida. X2.330.) 16. Early spermatogonial cell showing two large nucleoli. The cell boundary is quite indistinct. 17. Resting stage of a full grown spermatogonial cell showing two large and two. small nucleoli. 18. Late prophase of a spermatogonial division showing eighteen chromosomes' The two large oval chromosomes at the top are apparently the accessories. 19. Metaphase of division in a spermatogonium showing the two divided ac- cessories at the right. 20. Early resting stage of a primary spermatocyte showing the centrosome, two large, and several small nucleoli. 21. Primary spermatocyte in synizesis. 22. Primary spermatocyte following synizesis and synapsis; the entangled mass of threads moved toward the center of the nucleus and the threads appear to be twice as thick as those in the synizesis stage. 23. Primary spermatocyte following the synaptic stage. The clear area in the nucleus disappears and the nuclear wall becomes well developed. 24. Spireme stage of a primary spermatocyte showing increase in size of both the nucleus and the cytoplasm and also the two large nucleoli. 25. Breaking up of the spireme into U and variously shaped chromosomes. The two nucleoli become oval in shape and can always be seen close together in this stage. 26. 27 and 28. Late prophases of primary spermatocytes showing ten chromo- somes. The two oval accessories, which are smaller than the bivalent autosomes, are shown in characteristic positions. BIOLOGICAL BULLETIN, VOL. XXV. PLATE II. *<-: 16 mi ' "* .,.'•« .'-j 7-. ••'•-. ' .-X. ''. ." '.-^'V^X' "— • <£&#r 20 ' -";X ^mm^> .';,• ^^T^ • $ \;''v;:^ '•._;,/ 23 :>-x ^ .";'•'' --''"-'T • "'v-'-"''-"/.^ - ' J •'..•.;".•>•<• •'-- '"" 'A'.:X ''.-"' ""-^.^;.ii^fti^ . ..;• \'v : f^ ; ;•.. •••-•: ..< I "^">rvfc^ ''•}.;'>-' r ::-v:;'7 ' SS^ 26 ---...:„ -,^.-*- 27 28 J. I. WODSEDALEK 38 J. E. WODSEDALEK. PLATE III. 29. Late prophase of a primary spermatocyte showing ten chromosomes, the two accessories lying a considerable distance away from the regular equatorial plate. 30-37. Metaphase of division in primary spermatocytes showing the two accessory chromosomes in characteristic positions passing to the poles. Figures 31 and 32 show also a small body passing to the same pole with the accessories. Figure 35 shows it passing to the opposite pole; and Fig. 37 shows two such bodies, one passing to each pole, which is a case of extremely rare occurrence. BIOLOGICAL BULIETIN, VOL. XXV PLATE III. 29 g iintt A ^WW • ^^ v • . -Vv vS^T^-^'^VV-' 31 ••''•' • ' •' p •' i \"- '•• tJjLU'i^ f .-^-rv^^ K x^gSS ^&?K " ^L ' ^ •• . ." '. •• , • •* ^ -;•"•;' ''•' •*.;;•' C • I 33 jg$ XvT^J ,--• • ? ^%i^ ^- i "- "I .' '- -- - _ -75 J. E. WOOSEDALEK 40 J. E. WODSEDALEK. PLATE IV. 38. Late anaphase of division in a primary spermatocyte showing eight chro- mosomes at one pole and eight plus the two accessories at the other. 39. One end of a late anaphase of division in a primary spermatocyte showing eight chromosomes. 40. One end of a late anaphase of division in a primary spermatocyte showing ten chromosomes (eight plus the two accessories). 41. Resting stage of a secondary spermatocyte showing two nucleoli and a con- spicuous centrosome. 42. Resting stage of a secondary spermatocyte without the large nucleoli. 43. Two contiguous secondary spermatocytes of which one shows four bivalent chromosomes in metaphase of division, and the other four bivalents plus the two accessory chromosomes in metaphase of division. These two cells are no doubt the products of a division of a primary spermatocyte in which eight chromosomes pass to one pole and eight plus the two accessories to the other. 44. Early metaphase of division in a secondary spermatocyte which received the two accessories, showing six chromosomes in all. 45. Early metaphase of division in a secondary spermatocyte which did not receive the accessories, showing only the four large bivalents. 46. Late prophase of division in a secondary spermatocyte which did not receive the accessories. 47. Late prophase of division in a secondary spermatocyte which received the accessories, showing a total of six chromosomes. 48. Late anaphase of division in a secondary spermatocyte which received eight chromosomes, showing four chromosomes at each pole. 49. A spermatid showing four bivalent chromosomes which is apparently one of the resulting cells of the division of a secondary spermatocyte which received eight chromosomes. BIOLOGICAL BULLETIN, VOL. XXV. PLATE IV. A .; f •'•••' ; - :'y V - v 38 '•' '•'"',.!-. ' ' '- .' -V* --''**'. / -' -Jr^fMilf -: •'.'•• ,- / . v ' \:T;^rA^ '-, •. -^ ••*;:•''.•; '".'<-•; •-.: $ •. %&. ^>':> * * s's4i^'f-'!i'//^-:i,-'"O'-/ •.. , . '., Sii^^s^ 1 ''• ^^' • ' '•, '• - ,; ••- - ' L 1 J7.- • ^:±1 V::-:^.vXTT //'--.•• W - : • p, • 'mm ^i^ •'' :~:'S *^££:^ 4& •«%» S . 49 J E. »OOSEO«LEK 42 J. E. WODSEDALEK. PLATE V. 50 and 51. Late anaphase of division in a secondary spermatocyte which re- ceived the two accessories. Each of the latter divides in this stage. Figure 50 shows streaks of chromatin material uniting the two accessories. Usually no such streaks occur (Fig. 51). 52. One end of a late anaphase of a division in a secondary spermatocyte which received the two accessories. The latter are oval in shape and smaller than the bivalent autosomes. 53. A spermatid which received the accessories, showing a common arrangement of the chromosomes before they break up. 54. Division of a secondary spermatocyte which apparently received the ac- cessories. The chromosomes are so massed as to render a count impossible and the two streaks which rarely occur are probably due to retarded division of the accessories. 55. An early spermatid in which the chromosomes have broken up, while the two nucleoli remain in full evidence and are apparently the same thing as the accessories. 56. An early spermatid, which received the accessories, showing a characteristic breaking up of the bivalent autosomes into univalents. Two of the autosomes had divided almost completely while the other two also show signs of division. The accessories are univalent in nature. 57. A spermatid, which did not receive the accessories, showing a common ar- rangement of the four bivalent chromosomes. Very frequently these autosomes divide before they become massed together, as is the case of the autosomes in the other type of spermatids represented in Fig. 56. 58. A spermatid showing the reconstruction of the nuclear wall while the four bivalent chromosomes are still in full evidence. 59. Late prophase of a somatic cell, found in the mesonephros of a male pig embryo, showing eighteen distinct rod-shaped and oval chromosomes two of which are somewhat larger and are apparently the two accessories. 60. Late prophase of an oogonial division showing twenty chromosomes, four of which are longer and evidently the accessories. 61 and 62. Somatic cells, found in the mesonephros of a female pig embryo. Figure 61 shows a very uncommon condition revealing ten large chromosomes which apparently resulted from the pairing of the twenty chromosomes. Two of these are considerably larger and are interpreted as the result of the pairing of the accessories. Figure 62 shows the late prophase of division in a type of somatic cell occurring most frequently, and reveals a count of twenty chromosomes four of which are considerably larger than the others. 63. A full-grown Sertoli or nurse cell. BIOLOGICAL BULLETIN, VOL. XXV. PLATE V. x--.-:::;-:V''v.vs. dSEja ...^-r-1':'.^- feg •• •.•;-> .-:^ff - V:.,:4v-:V^^-.7-;-<:..^ .- %4< ••~&fpgr.'- .• .-. > •' •: . I W '.;.ffitV' .- -'-l-: ^^-iv^^f-K" -v4 •.«.•.-.• ••-,'•-. :'•••'. -.,-•. ! i^fjt *,- -.-, •. :' •-. -j" ••; .>.-. -. .^M.^ '••• •~' ' " • •' 'v :"i:; ^ ^liil | ::.-' HI- IP J. E. WODSEDM.EK 44 J- E. WODSEDALEK. PLATE VI. 64 and 65. Two contiguous spermatids, one without chromatin nucleoli, the other with two. The spermatids in general are about equally divided into these two classes. The centrosome is very conspicuous in this stage. 66. A spermatid showing the large centrosome, surrounded by a clear area, which had emerged from the sphere a short distance away. 67. A spermatid showing the centrosome in contact with the nuclear wall, and the migration of a portion of the sphere toward the opposite pole. 68. A spermatid showing the division of the centrosome, the position of the sphere as the acrosome, and the migration of the nucleus toward one side of the cell. 69. A transforming spermatid showing the separation of the centrosomes and the beginning of the axial filament. Also a dense spherical cytoplasmic body which is likewise present in some of the following stages. 70. A transforming cell showing a characteristic ring-like structure and position of the posterior centrosome with the axial filament extending through it. 71. A later stage showing the sloughing off of the posterior ring-shaped centro- some from the axial filament. 72. A somewhat later stage showing the posterior centrosome completely separated from the axial filament. 73. A cell showing the cast-off centrosome divided in two. 74 and 75. Still later stages of the developing spermatozoan showing the char- acteristic position of the posterior centrosome, and the formation of the axial envelope. 76. A spermatozoan almost fully developed showing the mass of sloughed off cytoplasm including the centrosome. 77. Side view of a mature spermatozoan showing the arrangement of the chro- matin at the periphery of the head wall. 78. A mature spermatozoan. BIOLOGICAL BUILETIN, VOL. XXV. PLATE VI. mm v-^viL v-J;;--' i . 75- >' .•"' 76 77 ' • •- •-'-•• •vv^:/o-«-jy • 78 J. t. WODSEOALEK A NEW EYE COLOR MUTATION IN DROSOPHILA AND ITS MODE OF INHERITANCE. SHELLEY R. SAFIR. A cross between a long-winged, vermilion-eyed female and a miniature-winged, red-eyed male yielded in the first generation red-eyed females and vermilion-eyed males, both long-winged. There appeared, however, in addition 15 long-winged males with a new eye color. This color is of the same general tone as the vermilion pink eye but is paler and somewhat creamier. A lighter area may also be seen encircling the darker area of the center. The origin of this new eye color was accounted for as follows: Since vermilion is sex-linked, all the sons must be vermilion whatever else they may be. Only males of the new eye color appeared, wherefore it was inferred that the something else added was also sex-linked. There were too many of them (15) to be explained each as a separate mutation, and it seems clear that the mutation itself occurred back in the stock from which the mothers were taken. If one or more of these Pi vermilion females were heterozygous for the new sex-linked factor we should expect in F! several of the double recessive males of the composi- tion cherry vermilion. (It was decided to call the new eye color cherry vermilion.) If then this eye color represents a double recessive cherry vermilion, when mated to red eye (the normal eye color) there should result in FI only red-eyed offspring. Then in F2, if both of these factors are sex-linked, there should appear four classes of males, viz., the double recessive cherry vermilion, and cherry and vermilion as the result of splitting cherry vermilion into its components, and also red eye color which is the double dominant. All the females should be red. That these results were actually realized can be seen below. f red 9 ............ 345 ( rl 0 ft \ I red 9 X cherry verm. 1 verm, cf ......... 158 I verm, cr 23i/ ,, { cherry verm, cr ... 158 The results bear out the view that the new eye color repre- sented the double recessive; vermilion plus some other sex- linked factor, viz., cherry, as the analysis shows: verm. 9 . .XCv — XCv cherry verm, cf Xcv — NEW EYE COLOR MUTATION IN DROSOPHILA. 47 Fi FEMALE. /XCv \Xcv verm. 9 F2 FEMALES. i MALE. XCv XCv XCv Xcv XCv vermilion 9 vermilion 9 XCv verm, cf F2 MALES. vermilion /•>«_! -71 I tVJOlli {J eosin 9 by cherry verm, cf : |cherry 9 [ eosin 9 • cherry 9 c? 119 I eosin no 154 cherry o71 20 > 76 cherry verm, o71 \ eosin verm, c71 18 The results may be accounted for in the following way: NEW EYE COLOR MUTATION IN DROSOPHILA. p eosin 9 cherry verm cf Fi FEMALE. XcVe XcvE cherry 9 GAMETES OF Fi 9 GAMETES OF Fi cf XcVe — XcVe XcvE Fi MALE. XcVe XcVe — F: FEMALES. XcVe XcVe eosin 9 eosin cf XcvE Xcve XcVE XcVe — F; MALES. XcVe eosin XcvE XcVe Xcve XcVe XcVE XcVe cherry 9 . n eosm 9 Cherry 9 XcvE Xcve cherry verm, c? , eosin verm, cr cherry In concluding I wish to express thanks to Dr. Morgan for furnishing me with material with which to carry on these crosses and for his helpful suggestions during the progress of the work. I have also been greatly helped by discussion of the theoretical side of the matter with Messrs. A. H. Sturtevant, H. J. Muller and C. B. Bridges. Mr. Bridges' suggestions and his final reading and criticism of the manuscript were directly responsible for the appearance of the paper in its present form. I must not forget to express my appreciation of the assistance of my friend Benjamin Schwartz. PARTHENOGENETIC CLEAVAGE OF THE ARMADILLO OVUM. H. H. NEWMAN. INTRODUCTION. In the year 1909 I began a general study of the ovary of the nine-banded armadillo in search of some explanation of the phenomenon of polyembryony. Among the first of the peculiar features to catch my attention were what appeared to be cleavage stages in certain types of ovarian ova. It was thought that this condition might have some bearing on polyembryony but further study and a review of the literature on kindred topics convinced me that these changes were in no sense to be interpreted as stages in normal development. They must, on the contrary, be interpreted as more or less abnormal and abortive attempts, on the part of ovocytes that have reached maturity but have been denied the normal culmination of their destiny in ovulation and fertilization, to develop embryos in spite of insuperable difficul- ties. I should hesitate to contribute another chapter to the already voluminous literature on the supposed parthenogenesis of mammalian ova during follicular atresia had I not at my command evidence of a very crucial character, which seems to me to demonstrate beyond controversy a certain amount of real parthenogenesis in a mammal. The ovaries studied are the same as those that formed the basis of a recent paper on maturation and fertilization in this species (Newman, 1912). Those ovaries, fixed in freshly made Zenker's fluid and stained by Bensley's copper chrome haema- toxylin process, give the clearest pictures. RESUME OF THE EVENTS OF MATURATION LEADING UP TO PARTHENOGENETIC CLEAVAGE. In order that the changes preceding cleavage may be under- stood it will be necessary to review the facts dealing with the maturation process which were described in a recent paper (Newman, 1912). The most striking peculiarities of this process 52 PARTHENOGENETIC CLEAVAGE OF THE ARMADILLO OVUM. 53 have to do with the reorganization of the cytoplasmic materials of the ovocyte. In this respect, and probably in many others, the armadillo bears a marked resemblance to the marsupials, espe- cially Dasyurus, the Australian native cat, whose early develop- g?, TEXT-FIG, i. A full-grown ovocyte, showing cytoplasmic organization, etc. Deutoplasmic zone (dz), formative zone (fz), germinal vesicle (gv), zona pelucida (zp) (X 410). dz TEXT-FIG. 2. A maturing ovocyte, showing the new reversed polarity. The ovocyte is placed with the animal pole upwards. The deutoplasmic zone (dz) occupies the animal pole, the formative zone (fz) occupies the vegetative pole. The polar spindle (/>s) lies in a tangential position at the equator of the ovocyte. Deutoplasmic granules (dg) lie in the center of the deutoplasmic mass. The zona pelucida (zp) is a dense envelope, without radiations. ment is so ably described by Hill (1910). So close is this re- semblance that any description given for one species would apply in almost every detail to the other. 54 H. H. NEWMAN. The full grown ovocyte of the armadillo has the structure shown in text-figure I. The cytoplasm shows two clearly defined regions, a peripheral zone of deeply staining homogeneous proto- plasm, the formative zone (f.z.~), and a central lightly staining, spherical mass of coarsely reticular or alveolar material, the deutoplasmic zone (d.z.), in which occur irregular masses of coarse and deeply staining granules (d.g.), the supposed equiva- lent of the yolk or deutoplasmic granules of the ova of marsupials. The germinal vesicle (g.v.} is flattened against the dense zona pelucida (z.p.} at the animal pole and in it the chromatin is seen as tetrads or in other forms characteristic of the prophases of maturation. Such an ovum has the appearance of being centro- lecithal. Coincident with the formation of the first maturation spindle a radical rearrangement of the two cytoplasmic zones occurs. The deutoplasmic sphere pushes its way to the surface at the animal pole and crowds the formative material to the vegetative pole where it assumes the form of a cap, thick at the pole and thinning out to a feather edge at the equator (see text- figure 2). The polar spindle, in a stable metaphase (p.s.), occupies a position as near the animal pole as possible without leaving the formative material and has its axis tangential to the periphery of the ovocyte though parallel with that of the main axis of the cell. This spindle is evidently insulated from the surrounding formative protoplasm by a capsule of non-staining material, and has no astral rays. A very large number of ovocytes were found in exactly this condition and it seems certain that the progressive changes normally come to a standstill at this point. The next step in normal development is ovulation, the stimulus of which initiates the completion of the maturation process. Such resting ovo- cytes occur in large follicles whose granulosa cells are intact and whose follicular fluid is homogeneous, abundant, and stains deeply with haematoxylin and allied stains. In any normal ovary during the period of cestus there are many ovocytes of this sort, all in practically the same condition and equally ready for ovulation. Probably such mechanical factors as size of follicle or nearness to the periphery determine which follicle shall rupture first. Fertilization of this one egg seals the fate of all the remaining PARTHENOGENETIC CLEAVAGE OF THE ARMADILLO OVUM. 55 ovocytes of the same generation in that they can never go through the process of ovulation and therefore cannot develop normally. A very small percentage of these ovocytes continue the matura- tion process to the extent of completing the first maturation division; a still smaller proportion (only three out of hundreds of cases examined) complete the second maturation; while the remainder, over 90 per cent., are struck by the processes of follicular atresia and either go into early cytolysis or enter upon a period of parthenogenesis. It is with the latter contingency that the present paper deals. ABSTRICTION OF THE DEUTOPLASMIC .MATERIAL AND ITS SUBSEQUENT FATE. The condition of equilibrium, just described and illustrated in text-figure 2, is normally disturbed only by ovulation, but, in the case of ovocytes whose normal history has been cut short by the fertilization of an ovum and the formation of a corpus luteum, the equilibrium may also be upset by the marked changes in the chemical environment incident to follicular atresia. The follicle in which atresia has set in rapidly decreases in size through the loss of follicular fluid, but, even before the amount of fluid has materially diminished, its chemical composition is clearly altered, for it no longer stains deeply with the ordinary cytological stains but remains pale gray under hsematoxylin treatment. Just what the nature of the change is I am unable to determine, but there can be no question as to its radical character. As atresia continues, the lumen of the follicle gradually disappears, partly through the loss of fluid and partly through the ingress of cells from the disintegrating granulosa layer of the follicle. With this radical alteration of chemical environment the ovocytes show marked changes, whose significance it is our problem to determine. The first act of the ovocyte entering upon parthenogenetic development is the abstriction of the deutoplasmic mass from the formative protoplasm. That this is a practically universal phenomenon can scarcely be doubted for there are in the present material literally hundreds of instances of it, and one can with confidence look for some stage of the process in every follicle 56 H. H. NEWMAN. that has reached the mid-period of atresia. Abstriction of the deutoplasm is essentially an act of rejuvenation on the part of a dying cell, in that the still living protoplasm is freed from the burden of the inert by-products of metabolism. The exact me- chanics involved in this purification process can scarcely be determined from the fixed material, but appearances seem to warrant the conjecture that the very fluid deutoplasmic material escapes from the cell membrane by the rupture of the latter and that, with the subsequent closure of the ruptured membrane, the deutoplasm becomes strictly extra-cellular and without effect, except in a secondary mechanical way, upon the living proto- plasm of the germ cell. Freed from its burden the renewed egg- cell rounds up into an approximately spherical form and ap- parently floats in the deutoplasmic fluid. It is of considerable interest to note that a similar abstriction of deutoplasmic material occurs as the initial step of normal development in Dasyurus. Hill shows that this discarded material plays no further part in development but merely occupies any available space within the persistent zona pelucida that is not occupied by the blastomeres. Finally it comes to lie in the cleavage cavity of the blastula and is probably gradually absorbed. In view of the striking similarity in the early history of the deutoplasm in the two species it is difficult to avoid the conclusion that the process, as described for these ovarian ova of the armadillo during atresia, is a close approximation of the first step in normal development, about which we know practically nothing at present. The deutoplasmic fluid, laden with its masses of deutoplasmic granules, usually gives the appearance of a multicellular body, but I am inclined to consider this condition as largely an artefact. It is not improbable that the strong fixing fluids used coagulate this somewhat viscous material in the form of many small rounded masses. In a few cases, as in Figs. 2 and 4, the deuto- plasm is only slightly broken, but in the majority of instances we find advanced fragmentation, as in Figs. I, 3, 5, etc. Whether the deutoplasm occurs in one or many fragments the deutoplasmic granules occupy a central position in the various pieces giving each fragment the illusory appearance of a nucleated cell. These are not however to be interpreted as in any sense true PARTHENOGENETIC CLEAVAGE OF THE ARMADILLO OVUM. 57 cells. At best they may be designated as cytoids. Other writers who have studied similar conditions in mammals have referred to the existence of multicellular conditions in which some cells possess nuclei and others do not. I suspect that, were the truth known, these fragments without nuclei would turn out to be of deutoplasmic origin. Very commonly these cytoids in our material form a pseudo-epithelial layer surrounding the true cells, as in Figs. 3 and 5, or less distinctly in Fig. I. In other cases they are distributed at random among the cellular products of cleavage (Figs. 7-11) and serve only to give a false appearance of cellular multiplicity and to confuse and obscure the pictures of cleavage. THE REESTABLISHMENT OF NORMAL INTERRELATIONS BETWEEN THE NUCLEUS AND THE CYTOPLASM AND THE RESULT- ANT FORMATION OF CLEAVAGE SPINDLES. The egg, much reduced in size but with a renovated cytoplasm, is now the seat of a renewal of nucleo-cytoplasmic exchanges, as evidenced by the appearance of an extensive system of astral rays at the two poles of the spindle. It will be recalled that the nucleus during the period prior to ovulation was a naked spindle, enclosed in an insulating capsule and entirely devoid of astral radiations. Such a spindle is evidently an isolated system and lies inert in the cytoplasm. The renovated cytoplasm of the egg from which the deutoplasm has been extruded now bears a different chemical relation to the nucleus and the change is shown in the disappearance of the capsule about the nucleus and in the appearance of typical astral radiations. That a renewal of metabolic relations between nucleus and cytoplasm is one of the most essential facts of normal fertilization has been recently maintained by Lillie (1911) on the basis of his studies of fertilization in Nereis, and there is reason to believe that this is the physiological explanation of parthenogenesis wherever the latter is found, whether normal or experimental. In this case it is my conviction that the sudden change in the chemical character of the cytoplasm restores the basis of life and growth to the egg and that cleavage is the natural consequence. In such cases as that shown in Fig. i the nucleus has returned to a 58 H. H. NEWMAN. resting state and has every appearance of undergoing healthy changes which should culminate in cell division by mitosis. It is impossible to state positively that such a nucleus has regressed from the spindle condition shown in text-figure 2, but that such is the case seems at present to be a reasonable conjecture. In eggs of this sort there is a characteristic zone of activity in the cytoplasm at a short distance from the nuclear membrane which may indicate active interchanges between nucleus and cytoplasm. The chromatin is in the form of a series of beaded threads resembling a spireme and therefore may be interpreted as in a very early prophase of the first cleavage. Soon the chromosomes begin to condense and the typical appearances of later prophases pass over into unmistakable mitotic figures like that shown in Fig. 4, which can be interpreted in no other way than as cleavage spindles. The so-called cleavage mitoses of other writers on this subject have been interpreted by opponents as belated maturation spindles which have been displaced and distorted by the abnormal conditions of atresia; but such an interpretation of the figures here described would seem very far- fetched in view of the fact that these spindles are so radically different from the naked, insulated spindles of maturation. I have never seen a vestige of astral radiations in a polar spindle, while the radiations in a number of such cases as that shown in Fig. 4 are as obvious and as extensive as those which I have observed in the cleaving eggs of annelids and other favorable material, and far clearer than similar appearances in fish eggs or other less favorable material. The spindles are frequently abnormal in form, being not uncommonly tri- or multipolar; and sometimes the poles are ill defined, as in Fig. 4, but the radiations are always very distinct. When the figure is multi- polar the chromatin distribution is very irregular and nests of nuclei (like those showrn in Fig. 12) are produced without any division of the cytoplasm. It is very common to find such multinucleate but unicellular eggs and I am inclined to attribute their existence to such irregular mitoses as that just mentioned. Certain considerations lead me to venture a conjecture as to the mode of origin of these multipolar spindles. It is not uncommon to find in ovocy tes of the type shown in text-figure 2 various steps PARTHEXOGENETIC CLEAVAGE OF THE ARMADILLO OVUM. 59 in the anaphases of the first maturation division, which might readily result in the formation of two nuclei without the extrusion of a polar-body. Conditions like those shown in Figs. 2 and 3 are not nearly so abundant as those like Fig. i.but they are about as common as are multipolar spindles and might readily result in the latter type of abnormality. Eggs with such paired nuclei have the general appearance of fertilization stages but cannot be inter- preted as such, unless the reunion of the polar nucleus with its sister, the nucleus of the secondary ovocyte, could be considered a sort of fertilization process. It is not difficult, however, to see how double nuclei of this sort could form just such multipolar spindles as that shown in Figs. 13 and 14. It will be readily noted from the figures that these abnormal spindles have a very much larger number of chromosomes than sixteen, the number seen in all clear maturation figures. This in itself militates against the interpretation of these phenomena as maturation processes, and is in accord with the conjecture that the figures may be the result of the cooperation of two nuclei derived from a maturation division. The bipolar spindles, such as that shown in Fig. 4 (see also Figs. 15 and 16), also contain more than the haploid number of chromosomes and cannot on that account be maturation figures. EARLY CLEAVAGE STAGES RESULTING IN THE FORMATION OF A FEW BLASTOMERES. In the course of these studies several very pretty two-cell stages have been encountered of which those shown in Figs. 5 and 6 are typical. The stage shown in Fig. 5 is a perfect two- cell stage in which both nuclei are in the resting phase. There are no other nucleated bodies within the zona pelucida, but a pseudo-epithelium of deutoplasmic cytoids forms a capsule about the two blastomeres. A clear case of a two-cell stage in which the second cleavage is well under way is shown in Fig. 6. The cleavage spindles are cut transversely in both cells, one of which shows an early and the other a later anaphase. After the second cleavage the subsequent divisions are less normal and the picture of blastomeric regularity is confused. The spindles lose distinctness, as a rule, and the chromatin 6O H. H. NEWMAN. begins to break down, giving appearances like those shown in Figs. 7 and 8, which are two sections through the same egg. Another example of the same condition is shown in Figs. 9 and 10, also taken from one egg. In Fig. 7 can be seen one cell with a fragmenting nucleus, a degenerating nucleus in another cell, and an incomplete mitotic figure with only a few small chromo- somes. In Fig. 8 occur three cells all of which show abnormal, though unmistakable, mitotic figures. Such an egg might be considered as a six-cell stage, with a prospect of reaching a ten- cell condition. There is every evidence, however, of approaching death and disintegration in such cases and one would not be inclined to look for much further developmental progress in such unpromising material. In the egg shown in Figs. 9 and 10 there are several imperfect spindles and two small but healthy nuclei. One might be somewhat more optimistic about the ultimate fate of cases of this sort, but, since these cases show about the maxi- mum of development in the present material, such optimism is scarcely warranted. Fig. II is introduced especially to show an unusually fine mitotic figure in one of several cells in an egg in which three other cells show more or less normal nuclei. This egg occurred in a follicle in a very advanced stage of atresia and it is rather surprising to find so pronounced a spark of life in a structure so nearly dead. In all of the cases cited and in nearly all in which similar phenomena were observed the zona pelucida was dense and quite intact, hence there can have been no invasions of stroma or follicular cells. So all nuclei found must be products of the division of the original germinal vesicle. In more advanced atresia, however, the zona begins to open up cracks which admit hordes of stroma cells and leucocytes that feed upon the dis- integrating egg material and form cell masses that have been interpreted by some authors as the products of embryonic development. In such cases the observer might be inclined to interpret such synthetic structures, which frequently have an epithelial structure, as tissues, equivalent to those formed in normal embryonic development. Leo Loeb ('n and '12) in recent papers interprets as placental tissues certain structures PARTHENOGENETIC CLEAVAGE OF THE ARMADILLO OVUM. 6l found under similar conditions in the guinea-pig ovary. Not having seen his preparations, but simply judging by his micro- photographic illustrations, I am inclined to think that, since he gets no intermediate or advanced cleavage stages, the cell complexes he discusses are open to the interpretation just given for similar phenomena in the armadillo, about whose synthetic origin there appears to be no doubt. Parthenogenetic develop- ment in all probability goes no farther in the armadillo than the stages illustrated in the figures. That true parthenogenesis begins and proceeds for a few steps seems assured, but it seems highly improbable, a priori, that development could long con- tinue in an environment so unfavorable as that afforded by follicles in which atresia has made such progress as we have seen. DISCUSSION OF THE LITERATURE ON PARTHENOGENESIS IN MAMMALS. In 1900 Bonnet in an able paper entitled "Giebt es bei Wir- beltieren Parthenogenesis?" reviewed all the literature dealing with parthenogenesis in vertebrates and gave much space to the question of parthenogenesis in mammals. He concluded that there was no incontrovertible evidence of this mode of develop- ment in any of the contributions dealing with the changes de- scribed as occurring in ovarian ova during follicular atresia. A considerable number of authors had described ovarian ova containing centrally situated mitotic figures which they had interpreted variously as cleavage spindles or as belated matura- tion figures. Multicellular masses within zona pelucida were described, in which some cells contained nuclei and others were without nuclei. Some writers considered this condition as a result of a degenerative fragmentation of nucleus and cytoplasm, and others were convinced that the cell mass was the product of parthenogenetic cleavage. Another class of writers called attention to the existence of various kinds of more or less com- plex ovarian dermoids, epithelioma and other teratoma, which by some were interpreted as the end product of the partheno- genetic development of ovarian ova. Judicially examining all of the evidence before him Bonnet concludes that all of the mitotic figures seen in ovarian ova are to be considered not as 62 H. H. NEWMAN. cleavage mitoses but as more or less abnormal maturation figures, that the so-called multicellular embryos are degenerative prod- ucts, and that teratoma and other kindred phenomena must be explained in some other way than as the products of partheno- genetic cleavage of ovarian ova. This conclusion is, I believe, fully justified by the evidence then available. Since this review and judgment of Bonnet several authors have reopened the question but have failed to agree. Spuler (1900) insists that the figures which he finds located in the center of ovarian ovocytes, after the extrusion of one polar body, are cleavage spindles, because normally ovulation occurs before this stage and the ovocytes do not develop thus far unless fertilized. He places especial emphasis on finding one ovum in which a centrally located spindle occurred in an egg that had two polar bodies. This, however, is probably a case similar to that described by the present writer (Newman, '12) as due to a precocious division of the first polar body, the spindle being merely the second maturation spindle. Spuler fails to add any material strength to the affirmative side of the question reviewed by Bonnet. Van der Strict (1901) working with the bat ovary also took the affirmative side of the question as the result of his discovery that ovocytes of the second order occasionally divided mitotically into two cells of approximately equal size. This division could not, he thought, be considered as an anomalous polar body formation, nor as a mitotic division due to degeneration, nor as fragmentation, but only as the beginning of parthenogenetic division. I am inclined to think that Van der Strict was dealing with phenomena closely akin to those described in the present paper for the armadillo, but he failed to make his evidence especially convincing. L. Loeb (1901) takes a position similar to that assumed by Spuler and Van der Strict and goes a step further in that he holds that the subsequent fragmentation of the egg material into nucleated and enucleated pieces is a progressive phenomenon akin to the parthenogenetic development of the egg of Chcctop- terns which, as Lillie has shown, develops into a larva of con- siderable complexity without any nuclear cleavage. Loeb figures PARTHENOGENETIC CLEAVAGE OF THE ARMADILLO OVUM. 63 one multicellular mass in which a. single cell is dividing mitotically, but the spindle is of the size and character of a polar spindle and lacks astral rays. Under the circumstances such a spindle would probably be capable of interpretation as the first polar spindle of an egg that had undergone degenerative fragmentation. Rubaschkin (1906), as the result of his studies of the guinea-pig ovary, takes a decidedly negative stand, maintaining that all mitotic figures occurring in ovarian ova during follicular atresia are more or less distorted or otherwise abnormal maturation mitoses, and in no sense the mitoses of embryonic cleavage. Multinuclear and apparent multicellular conditions are inter- preted by him as the result of degenerative changes brought about by the conditions incident to follicular atresia. Athias (1908 and 1909) once more reviews the whole situation and presents further data derived from the study of a number of species of mammals. After summing up the evidence pro and con he finds himself unable to reach a definite conclusion though leaning decidedly to the negative point of view. Pending an extensive program of investigation on the subject he prefers to reserve judgment. My own investigations herewith presented were begun, as stated, in 1909 and, as the reader will have discovered, support the view that a limited amount of parthenogenetic cleavage occurs but that development proceeds no farther than two or three cell divisions. SUMMARY. The evidence derived from a study of large numbers of arma- dillo ovaries demonstrates, I believe, that parthenogenetic cleavage takes place in atretic follicles of this species of mammal. Whether cleavage is preceded by maturation is not clear, but I am inclined to believe that no polar bodies are extruded in those ovocytes that are destined to undergo cleavage. There is strong reason to believe that the abstriction of the deutoplasmic material in these ovarian ova is equivalent to one of the steps in normal development, for a similar process occurs in the development of several marsupials, the ovogenesis of which shows the same history of deutoplasm formation and reorganiza- tion as that seen in the armadillo. 64 H. H. NEWMAN. There is no evidence that cleavage proceeds beyond the eight- cell stage, and even at that period there are many signs of advanc- ing degenerative processes. It seems certain, therefore, that, in the armadillo at least, no complex tissue masses such as ovarian teratoma or epitheliomata result from a continuation of the process of parthenogenetic cleavage. LITERATURE. Athias, M. '08 Sur les phenomenes de division de 1'ovule dans les follicules en voie d'atresie chez quelque Mammiferes. Bull, de la Soc. portugaise des Sc. naturelles, T. 2, Fasc. i. '09 Les phenomenes de division de 1'ovule dans les follicles de de Graaf en voie d'atresie chez le Lerot (Eliomys quercinus L.). Anat. Anz., Bd. XXXIV., No. i. Bonnet, R. 'oo Giebt es bei Wirbeltieren Parthenogenesis? Ergebn. der Anat. u. Ent- wickl., Bd. 9. Loeb, L. '01 On progressive changes in the ova in mammalian ovaries. Journ. of Medic. Research, Vol. 6. '05 Ueber hypertrophische Vorgange bei der Follikelatresie, nebst Bemerkungen iiber die Oocyten in den Markstrangen und iiber Teilungserscheinungen am Ei im Ovarium des Meerschweinchens. Arch. f. mikr. Anat., Bd. 65. 'n The parthenogenetic development of ova in the mammalian ovary and the origin of ovarian teratoma and chorio-epitheliomata. Journ. Amer. Medic. Assoc., Vol. LVI. '12 Ueber chorionepitheliomartige Gebilde im Ovarium des Meerschweinchens und iiber ihre wahrscheinliche Entstehung aus parthenogenetisch sich ent- wichelnden Eiern. Zeitsch. f. Krebsforschung, Bd. n, No. 2. Newman, H. H. '12 The ovum of the nine-banded armadillo. Growth of the ovocytes, matu- ration and fertilization. Biol. Bull., Vol. XXIII., No. 2. Rubaschkin, W. '06 Ueber die Veranderungen der Eier in den zugrunde gehenden Graafschen Follikeln. Anat. Hefte, H. 97. Spuler, A. .'oo Ueber die Teilungserscheinungen der Eizellen in degenerierenden Follikeln des Saugerovariums. Anat. Hefte, H. 50. Van der Strict, O. '01 L'atresie ovulaire et 1'atresie folliculaire du Follicule de DE GRAAF, dans 1'ovaire de Chauve-souris. Verhandl. d. Anat. Gesellsch., 15. Versamml., Bonn. HULL ZOOLOGICAL LABORATORY, UNIVERSITY OF CHICAGO, April 8, 1913. 66 II 1 1 \KWMAN EXPLANATION OF PLATES. PLATE I. FIG. i. Egg with large resting nucleus, showing resumption of chemical inter- action between nucleus and cytoplasm preparatory to the formation of the first cleavage spindle. The deutoplasm has been extruded and has been fixed in the form of numerous cytoids (cy) that almost surround the egg. The dense zona peludica (z.p.) encloses both egg and cytoids. The egg shows a zone of unmodified formative protoplasm (f.p.) and a precipitation ring (p.r.), where the changes between nucleus and cytoplasm are most active. Note that each cytoid has a centrally located, deeply staining, mass of granules, which are the deutoplasmic granules of earlier stages and are not to be interpreted as degenerating nuclei. (X 800.) FIG. 2. Egg in which the deutoplasm has just been abstricted and has not yet undergone fragmentation to any marked extent. Two nuclei (w1 and w2) lie side by side giving the appearance of a fertilization stage. It is probable that these are the product of the first maturation division, completed without the extrusion of a polar body. This could readily occur in cases where the maturation spindle occupies a central position. These two nuclei would doubtless form a multipolar spindle like that shown in Figs. 13 and 14 and would probably result in the formation of nests of nuclei, as in Fig. 12. Other labelling as in Fig. i. ( X 800.) BIOLOGICAL BULLETIN VOL. XXV. PLATE I. cy :" fp n -~_. M. H. NEWMAN 68 H. H. NEWMAN. PLATE II. FIG. 3. Egg with pseudo-epithelium of cytoids and with binucleated egg cell in the center. The explanations given for Figs, i and 2 will apply here. ( X 800.) FIG. 4. Egg showing mitotic spindle (ra.s.) with well-defined polar radiations running from nucleus throughout the cytoplasm. The rather blunt-ended spindle shows mitosis in an anaphase and there are many more than the reduced number of chromosomes, which one invariably finds in maturation spindles. Such a spindle cannot be other than a cleavage spindle. There are only two large cytoids (cyl and cy-). (X 800.) BIOLOGICAL BULLETIN, VOL. XXV. — n. >< :-A ...-•. ';-;•/ \ \y £' *mir-^:Tk~ + i » • . -.1 ^—\ \ / § 3 Ta 5 JO H. II. NEWMAN. PLATE III. FIG. 5. A two-cell stage of cleavage. The two cells are of somewhat unequal size, but the nuclei (nl and «2) are both in good condition and in the resting stage. A pseudo-epithelium of deutoplasmic cytoids surrounds the blastomeres. ( X 800.) FIG. 6. A two-cell stage passing into the four-cell condition. Two mitotic spindles are cut transversely to the axis. The spindle on the right (m.s°) is cut near the equatorial plate and shows only a few of the chromosomes. Other chro- mosomes are scattered through several sections. The spindle on the left (m.s1) is in a late anaphase and has two groups of chromosomes, one at each end of the spindle. The present section cuts through only one of these chromosome groups. Large numbers of cytoids occur in other sections. (X 800.) BIOLOGICAL BULLETIN, VOL. XXV. PLATE III. 71 J M. H. NEWMAN 72 H. H. NEWMAN. PLATE IV. FIGS. 7 and 8. Two sections through the same egg, showing one clearly defined mitotic spindle (m.s1), a cell containing three nuclei in the resting stage, one cell with what appears to be a degenerating nucleus («?), and one cell, shown in both sections, with an imperfect mitotic spindle and only a few chromosomes (ra.s2). Cytoids of all sizes are mixed in with the blastomeres. Such an egg seems to be in about a four-cell stage, but will probably not go much further. ( X 800.) BIOLOGICAL BULLETIN, VOL. XX/ PLATE 'V 7 H. H. NCWMAN 74 H. H. NEWMAN. PLATE V. FIGS. 9 and 10. Two sections through one egg. This seems to represent about an eight-cell stage. There are two cells with resting nuclei (n1 and «2) and three cells with more or less perfect mitotic spindles (m.sl-m.s*>). Cytoids are scattered among the blastomeres, giving the appearance of cell multiplicity. ( X 800.) BIOLOGICAL BULLETIN, VOL. XXV PLATE v. d m — - ms rn. 5- 77' IO H. H. NiWMAN 76 H. H. NEWMAN. PLATE VI. FIG. ii. A section through an egg in about a four-cell condition showing the most normal cleavage spindle (m.s) found in all the present material. Two other cells show unmistakable resting nuclei (nl and w2). ( X 800.) FIG. 12. An egg with many small nuclei in the formative protoplasm. Such a condition may have arisen through the formation of a multipolar spindle like that shown in Figs. 13 and 14. This is a very frequently found condition. (X 800). BIOLOGIOl BULLETIN, VOL. XXV. PLATE VI. - n" — d n II n r\ 12 H. H. NtWMAN 78 H. H. NEWMAN. PLATE VII. FIGS. 13 and 14. Shows the entire chromosome complex in a multipolar spindle which may have arisen from the fusion of two nuclei in eggs like those shown in Figs. 2 and 3. It will readily be seen that the number of chromosomes is very much higher than would arise from a spindle formed from a single nucleus. (X 1,600.) FIG. 15. An equatorial plate view of a first cleavage spindle of the bipolar type. The number of chromosomes is far larger than the haploid number char- acteristic of maturation figures. (X 1,600.) FIG. 16. Another example of the same phenomenon shown in Fig. 15. (Xi,6oo.) BIOLOGICAL BULLETIN, VOL. XXIV. PLATE VII 16 H. H NEWMAN Vol. XXV. July, i9i3. No. 2. BIOLOGICAL BULLETIN THE REACTIONS OF CERTAIN ANIMALS TO GRADIENTS OF EVAPORATING POWER OF AIR. A STUDY IN EXPERIMENTAL ECOLOGY. VICTOR E. SHELFORD. WITH A METHOD OF ESTABLISHING EVAPORATION GRADIENTS BY V. E. SHELFORD AND E. O. DEERE. CONTENTS. PAGE I. Introduction 79 II. A Method of Establishing Evaporation Gradients by V. E. Shelford and E. O. Deere So III. Material 84 IV. Experimental Results 85 1. Dry Air 85 2. Rapidly Flowing Air 97 3. Warm Air i oo 4. Death through Evaporation 102 V. General Discussion and Comparison 104 1. Rating of the Species 104 2. Comparison 105 3. Physiology of Water Withdrawal and Water Starvation 107 4. Importance of Evaporation Rate no VI. Summary 1 13 VII. Acknowledgments and Bibliography 114 I. INTRODUCTION. Aside from studies of the hygienic workers (Rubner and others), studies of mountain sickness, and other physiological phenomena of high altitudes and reduced atmospheric pressure (Cohnheim and others), little attention has been given to the effect of loss of water upon land animals. Since evaporation is determined by air movement, humidity, pressure, temperature, and indi- rectly by illumination, most of the so-called physical factors are measured in combination by instruments measuring evaporation. 79 8O VICTOR E. SHELFORD. A knowledge of the effect of evaporation upon animals is im- portant for the following reasons: (i) Because knowledge of the relationships of land animals to the surrounding medium is important from the standpoint of evolution and physiology, (2) because factors controlling distribution are effective in pro- portion to their effect upon the organisms concerned, and (3) because animals kept under laboratory conditions in experiments in behavior, genetics, etc., are often subjected to constantly changing atmospheric conditions and these changes may be sufficiently important to interfere with the results and the validity of the conclusions drawn. With these points in mind, the writer and Mr. E. O. Deere undertook to construct a piece of apparatus for the control of atmospheric conditions, but especially to establish experimental gradients and to test the reactions of animals to variations in the rate of evaporation. Various diffi- culties were experienced in getting the apparatus into working order and it was necessary for Mr. Deere to leave when the ap- paratus was ready for the control of humidity and some half dozen experiments had been performed. Nearly all of the experi- ments accordingly devolved upon the writer and we present the method of work only, as a joint contribution. II. A METHOD OF ESTABLISHING EVAPORATION GRADIENTS. BY VICTOR E. SHELFORD AND E. O. DEERE. The air supply was obtained by a compression pump as shown in Fig. i. A metal funnel (MF) covered with cheese cloth conducted the air to a pipe, 14 cm. in diameter and 340 cm. long, to a Beach-Russ Vacuum pump, no. i, run by a one-half horse-power motor. The air left the pump through a f-in. iron pipe, and entered a J-in. Crane oil separator, which removed the oil with which the pump is operated. In the pipe near the oil separator there was a pressure gauge, an air-cock opening to the exterior, and an automatic blow-off set for 5 Ib. pressure. From here the air passed through 15 m. of f-in. galvanized iron pipe, the central 5 m. of which was surrounded by a i|-in. iron pipe water-jacket, con- nected with the city water supply by means of a |-in. iron pipe, provided with a valve. By turning the city water supply into this jacket we were able to lower the temperature of the air when the outside air was warmer than the room. This was necessary because after the flows were divided they passed varying distances before reaching the tanks and if the temperature of the air was higher than that of the room, different amounts of cooling gave different temperatures in the experi- mental boxes. Several brass ground unions made it possible to take the entire line apart. The air current was divided into six parts by means of iron pipe Y bends, and reduced to J-in. with bushings. Each branch was provided with a REACTIONS OF ANIMALS. 81 •o +3 82 VICTOR E. SHELFORD. rising-stem straight-way valve, with hose end nipple attached by means of a j-in. brass ground joint union. The experimental boxes (Fig. lA) for the gradient experiments were designed by the senior author and were 6.5 cm. wide by 30 cm. long, and 2.5 cm. deep. The animals were confined in a portion of each box, 5 cm. wide, by a screen (8 meshes to the centimeter). Air was supplied to the boxes by means of three fish-tail burner- shaped introducers, opening into a narrow slit along the rear midway between top and bottom. The area of the slit in each introducer was approximately the same as the area of the 7 mm. brass tube which connected it with the pump. The air after leaving the fishtail burner-shaped introducer passed through the screen that confined the animals away from the slit, across the box, and out of the front through a screen similar to the first. The air tended to spread out at about the angle of the side of the introducer, and a small piece of metal was inserted between the rear wall and the confining screen, to deflect this part of the flow directly across the box. Each experimental box rested upon a movable board to each end of which a ring stand was fixed. Each ring stand bore a universal clamp, a small piece of Bessemer rod, and a single burette clamp used to hold a i candle-power incandescent lamp in any desired position. Our experiments were performed with the center of the in- candescent filaments 20 cm. from the bottoms of the experimental boxes, and one fourth of the distance from each end. These boxes were placed in a hood painted dead black, and provided with two curtains, one of which hung from above down- ward and contained a small slit for the observer, the other when in position came up from below to a point 10 cm. above the level of the lamp. This excluded practically all the faint light of the room, and the light was between the animals and the ob- server which made it less easy for them to see him. The control box (C) was supplied either with untreated air direct from the pump or with no air. The experimental box (E) and Fig. lA, when experiments with atmospheric humidity were being performed, was supplied with three kinds of air, wet at one end, dry at the other, and untreated or medium air in the center. The dry air was rendered dry by passage through three or more sulphuric acid filters, Fig. iB (see Shackell, '09). Each filter consisted of a Whitehall Tatum museum jar, 45 cm. tall and 9 cm. in diameter at the top. The covers were pieces of plate glass, 13 mm. thick and provided with two holes 25 mm. in diameter. They were clamped into place by ordinary hardware screw-clamps with a 6 cm. opening. The contact points were filed smooth and covered with rubber tubing. The air entered through a glass tube inserted in a rubber stopper in one of the two holes in the cover just mentioned. In order to conduct air to the bottom of the jar, the tube through which it entered was inserted into a second cork fitting inside a larger tube (Fig. iJB) so that when the glass plate was pushed into place, this stopper was pushed inside the glass tube so as to make an air-tight connection. The air left the filters from the top through a short tube inserted in the second rubber stopper. The filter jars were filled with crushed pumice stone, the pieces ranging from 0.5 mm. to 7 mm. in diameter, as indicated by the meshes of the screens used in separating and re- moving the finest and coarsest pieces. This was impregnated with crude sulphuric acid. After leaving the drying filters, the air passed through glass wool filters, made by filling ordinary i-liter, wide-mouthed wash bottles \vith glass wool. Tests of the air, made by bubbling the supply from these filters through a methyl-orange solu- tion for a period of 30 minutes, indicated that no sulphuric acid passed into the final delivery pipe. REACTIONS OF ANIMALS. The untreated air entered the final delivery pipe after passing through a '.glass wool filter. The wet air passed from the glass wool filter through two 2-liter\ as- pirator bottles filled with crushed pumice impregnated with distilled water. /To test the evaporating power of the three kinds of air, Livingston ('06, '08, 'ioa, 'iob, 'n ; see Abbe, '08) cup atmometers (evaporimeters) were used. Small Whitehall Tatum museum jars, similar to those used for the sulphuric acid filters, 19 cm. deep inside, and 9 cm. in diameter at the neck, were used to confine the atmom- eters. The atmometers were placed in position, after being stoppered with one- holed rubber stoppers with small tubes inserted. A large stopper with three holes was used to carry all the apparatus inside the jar. The tubes which connected them with burettes, where the amount of water evaporated was read off, passed through the central hole of the large stopper. Another tube which connected with the air supply passed through one of the side holes, parallel with the evaporimeter and reached within about 5 mm. of the bottom of the jar. The apices of the evaporimeter cups were from 25-30 mm. from the bottom. The third hole in the stopper was used for the exit pipe; the air from the supply passed upward to the bottom of the jar, which was inverted, and in returning to the exit tube, passed over the atmometer. The size of the jar was such that the velocity of the air over the atmometers was the same for a given flow, as across the cages. An atmometer thus enclosed was provided for each of the three kinds of air. For testing relative humidity (Fig. iC), two long chemical thermometers, graduated to o.i of a degree, were inserted inside of a glass tube 50 cm. long and 22 mm. in diameter. One of the thermometers was provided with a wick of absorbent cotton toweling, which dipped into a vial of water. Air from a supply pipe could be introduced into the tube below the bulbs. Two glass Y's were inserted between the filters and the experimental box, in each air line. The respective stems of these connected with the main air tube and with the experimental cage, as indicated in Fig. i . One arm of the Y of each line was connected permanently with one arm of the Y attached to the experi- mental box and supplied with a pinch-cock. The six free Y arms were supplied with free rubber tubes as shown in Fig. i. The three of these connected with the nearest filters could be joined to the observations tubes (Fig. i), or to the atmometers, or the thermometer tube. The other three (FT) could be used to connect the experimental cages directly with T of the control tube to give a rapid flow or when a coil of aluminum pipe submerged in hot water was interposed, to give warm air. These six free tubes made possible cross connections and the shift- ing of any kind of air to any section of the experimental cage. All open tubes were closed by means of pinch-cocks. For studying details of behavior and testing the ability of the animals to with- stand high rates of evaporation, three glass tubes, each with a total length of 21 cm. and an inside diameter of 32 mm., were used. These were connected with the free rubber tubing of the less frequently used arms of the Y's by means of funnels of the same diameter as the inside of the tubes. The stems of these were inserted into single-holed rubber stoppers. The large end of the funnel was cov- ered with a screen whose meshes were i mm. square, and the whole inserted inside the tube, with the stem projecting outside the rubber stopper for the attachment of the rubber tube. The funnel permitted the expansion of the air so as to practically fill the tube when it entered, and the screen prevented the small insects from entering the rubber hose. The air left each tube through the small funnel, similarly inserted into the other end. Since animals find difficulty in walking on curved glass sur- VICTOR E. SHELFORD. ai es, the lower third of each tube was filled with paraffin upon which sand was *-.f t ( 'd while the paraffin was still warm. From each of the tubes the air passed to an atinometer chamber, as previously described, so that the evaporation could be re- corded while the animals were being observed. In order to vary dryness only, it was necessary that the flows of the three kinds of air be the same. The flows were measured by collecting the air for a period of five seconds in a jar filled and inverted in a vessel of water. The pump commonly delivered air under sufficient pressure to force it through the filters and give a flow of from 12 to 1 6 liters per minute, an amount sufficient to change the air in the part of the cage immediately in front of each introducer (when all are flowing, in the entire cage) in a maximum of .6 of a second. It was possible to adjust the flows so that they were essentially alike, without collecting the gas. The boundaries of the rear walls of the cages were of solid metal and the slit was midway between top and bottom. A small triangle of thin paper 2 to 3 cm. long and about i cm. across the base was taken between the thumb and fore-finger by the point. With the hand held firmly by resting against the cage, the paper was placed in a position such that it was entirely in contact with the upper half of the rear wall without bending. The piece of paper was then lowered so that the broad end came in front of the slit and the deflection of the paper was noted and the flows adjusted until each gave the same deflection. The maximum flow was barely sufficient for effective experiments. The atmometers were those furnished by the Plant World. Three were selected with the standard .75 and restandardized by the careful adjusting of the flows until they were exactly alike. The evaporation was recorded for standard lengths of time when the flows of the medium and dry air were turned alternately over each of the atmometers for several periods of ten minutes and one period of an hour. It was found that the evaporation of one kind of air was the same no matter which evaporimeter was used. At the end of the experiments, the atinometer used with the wet air showed smaller evaporation due in part to sand that was accidentally blown out of the observation tube onto the atmometer. III. MATERIAL. The following species were studied: the yellow-margined millipede (Fontaria corrugate Wood), ground beetles (two species of Pterostichus) , the wood frog (Rana sylvatica LeC.), the red- backed salamander (Plethodon cinereus Green), the sticky sala- mander (Plethodon glutinosus Green) and several species of snails, all from moist forest situations, maximum evaporation is about 11.5 c.c. per day near the surface of the ground ; the common toad (Bufo lentiginosus Shaw), the digger wasp (Microbembex mono- don ta Say), the bronze tiger beetle (Cicindela scutellaris lecontei Hald.), the spiders (Geolycosa wrightii Em. and pikei Marx), all from dry sand ridges covered with cottonwoods and pines, a type of situation where the maximum evaporation per day is about 32.5 c.c. The animals were kept in the laboratory under REACTIONS OF ANIMALS. 85 as nearly natural conditions as possible. With the exception of Plethodon and Fontaria, they were kept only a few days. IV. EXPERIMENTAL RESULTS. i. Dry Air. The air used in the experiments was dried in the sulphuric acid filters described on page 82. The water vapor present after treatment depended upon the humidity of the original air, upon the temperature, the rate of flow, and the number of filters used. On account of variations in temperature and relative humidity from day to day, it is necessary to either practically saturate all the air with water vapor or to dry all of it and follow by standard treatment at a constant temperature if the same con- ditions are to be produced from day to day. Tables I. and II. (pp. 88, 96) show from 0.5 to 1.05 c.c. evaporation for 2O-minute periods (calculated from lo-minute exposures) and relative humid- ity of 9 to 20 per cent, for the treated air. The relative humidity of the air used ranged from 40 to 60 per cent, of saturation; the reduction in per cent, of humidity ranged from 34 to 52. The moist air was more constant and the evaporation is arbitrarily given as 0.02 c.c. per 20 minutes. This number is based upon a number of one-hour exposures of the atmometers, as readable results were not noted in lo-minute exposures. (a) Moist Forest Animals. i. Physiological Effect and Reactions. — With the exception of the snails the dry air was stimulating to all the animals tried. Plethodon cinereus wTas stimulated at once in the driest air and usually moved back and forth in the observation tubes. Activity sometimes alternated with short periods of coiling but movement was the rule and was usually increased during the first fifteen minutes when erratic movements occurred. These were usually followed by coiling or cessation of activity accompanying a dry appearance of the skin. The animals usually dried and shriveled without further activity, the ability to move being gradually reduced. In the medium air, when the rate of evaporation was low, they often behaved quite normally for a few minutes when 86 VICTOR E. SHELFORD. ceilings and activity began to alternate. This was followed by heightened activity and stiffening as before. In the gradient a negative reaction to air of more than a minimum evaporating power was clearly shown as indicated by Table I., and Chart I., Experiment n (p. 87). The salamanders appeared to sense the drier air at once as indicated by hesitation or by turning back when the change was encountered. The latter indicates that these animals have a sense of orientation in the gradient. They usually tried the driest air one or more times and the different trials were usually followed by turning when it was encountered again. They usually piled together in the moist air after 13 to 1 8 minutes and remained so for considerable periods. P. glutinosus is clearly more sensitive to dryness than is P. cinereus. While the former was clearly more stimulated in both observation tubes and gradient experiments than was the latter, stiffening due to drying appeared so early that in the gradient experiments glutinosus did not turn back as definitely as did cinereus. The wood frog was stimulated at once in the dry air; it showed agitation at first but very soon (a few seconds to five minutes) crouched close to the bottom, drew the legs close to the body, and partially or wholly withdrew and closed the eyes. After this had continued for a time, the frog usually hopped in the direction in which it was headed and if the disturbance was not relieved it repeated the crouching. While the frogs sometimes appeared to orient in the gradient, this capacity is poorly developed and the graphs are quite dif- ferent from those of the salamanders. The frogs showed a preference for the moist air and avoided the dry air mainly by random hops and a lesser tendency to hop in the moist air. The difference in the appearance of the animals in the different parts of the experimental cages was striking. In the air of highest and medium evaporating powers the withdrawal and closing of the eyes just referred to took place to a degree ap- parently proportional to the rate of evaporation. Here the skin was dry and dull. In the moist air, the difference after a few moments exposure, was striking. The skin glistened with moisture, the frog sat upright, the eyes were fully protruded and wide open, and the animals gave an impression of sagacity not REACTIONS OF ANIMALS. ll I l.iiiil.iml imTTrnnidiiiiriinlltiiiniiiiTliiiiil Cm i ill, mil mi I Imul ••^sas: s = aa I1"11!""1!""1!"1"111"!1""!""1!1""! 7(1111111111.11 I 1111111111.111 T, i m I illmn cd 'co 4-J •*-* 4_> a -H- o "o o 4-> cd o £ o a cd o CJ CU b« a cd IH CU 4-* CU '"o — •3 c CU CJ _o CU 4-) CO Q o "M ~ a CO CU •"^ CU .a cd cu IH CU IH o — •3 ~ LH j~l CO CU "> 3 CU O CU "^ •> ** CJ CJ CJ ^3 )H '4^ CO B CO t 4-» cd — t*^ cd r. _O a CU a o CU W> CU B Cu 13 ~ CO a 1 iments CU a HH CJ 4-> a y, CU o iidividu _> ui ^ ••H o. CU CU a CO 4-J +2 CU CU cd •o IH B •-H IH CU _O cd •"-" cd CU CJ a CU 'u CU a CO cd o ~ CO "^ a 'co CU M cd ~i. CO o a CO Id _ CU CU H CO cd "cu 3 • — i a 4-> 3 ^ 13 a .5 . * , T a o s: CU "o CJ CU a 'c3 M O CU CJ 13 o IH CU a 'a 2 CU •-* cd ^t CO £2 • — i CU — •^ a 1 .a 6 2 o ^ -. o CO cd 4-> anim J3 H *-M O CO "3 CN CU J3 CO CU cd o cu CO o cd CO S3 - cu CU M a •3 CU CO cd M M ^ ~ CO CU 4-> 3 •n CO CJ cu CU 4J a ~ CU CO 4-) cd CO CU _r^ ~ W CU CU CO CU £ CU "o _^ a CU a M .a . 4_> CU o •£ a 4J o "o CO CU CJ a E 3 a CU CU o a a o o CU cd CO CU ,0 CO cd "D g - •3 H CU CO 88 VICTOR E. SHELFORD. cu -S ^ d 0) )H CU 3 4-* en 1 J s 'o •*-• n 0 CU g p CN I* 0 -o •£ - .s eu R< ~ cd cu E £ ' en "a fe a | en en O E a "S .2 cu cd flj "3 & cu « cu a "o e« ^ CU < ^ -uT & "c 3 ;: ^ § . W U^ O g OJ IHH-, (-1 S o -c > CE) CJ V) J^ T3 . c t! ra 3 ^ 3 u C3 Ji "*~* .. S O *. > •3 s '-s .s •- 3 t— < en >, lit — • kH -5 -° ^ P o cd ~a g + J ^T "*"* u 3 " Z o -a -3 ^ - c J CU T3 C „. 4-» Co cu en b« •£ — 3 Q a C * J ^ trt i_, r w en o a cs « cu f~l f en <•• s iment cu o 4-1 . M g ^ ri CU ^ a Ofi . ^ cS .5 ^ S -r g CJ W J n o a CS cu ^ cu « 11 D O ~ - a oo O .•< o - js H ^£ 10 — H . . . '-U •+ r-* •- •« o S o H : * -§ en J-^ 3 o _cu •3 a OJ CU •a -p a § 1 & O CJ 15 "S s: IH 2 Su "o s O cu 3 -3 z S E en cu CJ .« ' B, «> ™ cS o cS 5 j* cu 4-» 2 1 fl'. ^ CN 5 •o * n 1 II cu | S G nj •^ QJ 4J en °1 u cu J 2 a cu "o js" <-fc-i 2l ' (5 O -*J &/1 •4-> rn S •;-> X H g ° O CJ cu O j 3 ri cu u- c cj £ ~ „: O WD E x ° "o ^ i« ^ cu cu T3 en cu Q • w a 3 cu bio 4-) CJ en "* 3 « cd d cS o cu u s -! . en n 3 a •^ " 73 X cu g -H • i CU — )H X ^ , 4-» .|~ 0 -C .3 £ .E H U O •* ro CN CN i -t 00 O o c*} •}S1OJ\[ O) ro o M M \o oo ro CN w o ro r*j ^ 0 \< 3 O r~ c O CN CN • 1- to ro O ro o rt Kg -O o D b £ H •jaXjQ oj SuipuodsajjO3 O ro F» ro w ro 10 M O O W IN CN M M •J34S1OJ^ O] Suipuodsajjo;") CN Tt t— TT M ^f 00 O O M M CN CN H H •)U9UiU3d ~X3 J° s-mjEjadmaj, \O *O CN M M CN CN CN CN CN c? CN CN M -^ CN l-l M CN CN CN CN CN CN CN CN CN H EXPERIMBNT. 'o S «*2 U.B2 •*a •^ O\ CN TT 1O to CN CN CN O 1O Tt 1-1 CN M M ro •uinip9j\[ Tf OO 1O M O\ M CN CN M t— IO\O 00 O CN 00 ^- CN M CN l-l CN CN •JSIOK CN ro ro 10 o oo O O\ r- oo o 00 M fO CN CN I O ro C^ oo \£ o rr M 5 t- «*• vO Turned Back From •J3AjQ 10 •* 10 CN 00 Hi - r^ o w O O M CN CN •J3js;oj\r CN OO H O O * o o o o o M CN d Relative Humidity in Per Cent, of Saturation. *- \ 0 OsvO \ CN M M o M o o\ M M 'O *O 1— 1 W t-- HH •uinipaj\[ | \0 M -t to \O to to 0 ss> IT) LO K •«™ I oo oo oo 1 o\ a o\ oo O\ 00 o o\ o\ 00 GO O O\ a Evaporation in c.c. Dur ing Experiment. ..a O 0 000 00 00 00 O\ vO 00 t> O O to O O O O oo \o t^ o\ to 10 to d 00 O O O o 0 0 O O 0 O O •uinipaj\; O O CN O >O ro O O O O to 10 10 ro ro ro ro ^ CN CN M d 00 0 0 O O O O O O O O O •»™ o o o o o CN 0 CN CN CO ro CN CN CN 0 0 O O O O O CN 0 d 00 000 o O O O O O O O oj a W P. cinereus (duplicate control). CU o s *«* "Sj s o &H &J cu — — tr < y * M cd * * *. 00 O I- O O -O < IO O M !O O M ca | REACTIONS OF ANIMALS. 89 w CONTROL. •[OJ1 -1103 jo ajnjcjaduiaj^ rj- *f Tf Q M O Ol CS W W IN C^ s o P) (4-1 o ||| •AJQ oj 3uipuods3.uo3 •^l" ro w O ^O ro 5- -^ Suipuodsajjo^ r* M co O O O M M M C4 M M M PI t- •JSIOJ\[ OJ O\ O 10 CO O w ro ^" ^" fO '-J* W ON ro 10 » r" C — o M •jajsioi^ oj SuipuodS9jao3 O O M PI «fr PI p) M M •juauiuad -xjj jo aan}BJ3dui3j_ ptprpTs^ pt p) O PJ EXPERIMENT. o •**a P) r» ON r PO M PI »• 0 0 ON * ro PI IO P) to •mnipaj\; PI M M M PI M ON M to M •JSIOJ\[ r^ r» to c i oo ^f 5 •* 10 V> o oo a i- Gfe _3 •jaXjQ HNHMm.m P) PI •J3jsioj\; M M M M O P> M Relative Humidity in Per Cent, of Saturation. •** O\ ON ON ON N M 10 M 1 •ujnipsj^ O O NO IO IO IO M 0 IO 10 •jsioj^ oo oo oo ON ON C\ oo oo ON ON oo ON Evaporation in c.c. Dur- ing Kxpcrinu-nt. *XJQ 0 O O O O O ON ON vO 10 ON\O 6 to 6 O O O O O O *uinip3j\[ 0 0 O ro ro P) IO IO ro O PI o 6 o o o 6 d •jsioj.^ PI W H P) P) P) O O 0 O O O PI O 6 O O O O O 0 "o u D. f. Fontaria HI Carahid •juarauadxg jo '"v^ ciM U PI NO PI ON ON M 10 O M cS 01 -± H VICTOR E. SHELFORD. REACTIONS OF ANIMALS. 10 I/O 3 O £ I O U 3 "> <£ IT) R O a V e ;- o fc£ Ov o c "° o i! S § u u o\ o 2 « c c £'§ a x W o LO "3 .b - < o >, U u n rt \O t-, 10 5 - c 5 fe .a i^ v a x W !iiMiliiHiliMiiTl,iiiliiiii1i,iiilMiiihi,iilii'ii.linMliiiriliiiiiliMMl(ni,T', jn.lniii'I him I 2=52 = i?a"a2 If 'TTN ' r~\^T\ i r=^\ r TTS T~\ r — -#T^ n""I'\ — ' nil. I. .. Ill lima = a - I I 'pm i pi ni^ ii ii ij ii ii 1 1 T''""I I" '"i \n\\ff--- I ; " " '7 1 1 • M 1 1 m 1 1 1 1 1 1 1 M 1 1 1 1 .- 1 1 1 , i . , i . 1 1 1 1 ! 1 1 1 1 1 ; . . i ii n 1 1 1 I. ........ nli, ml ..... I "l"'"lm"l ..... I ..... I 'i'i.liim.1 hiiiitiiiiiJuiiiNniM hiu — m' ' "'I i i" a 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 n 1 1 1 1 , i imTi n n p 1 1 1 I'IMJ '~i^i ' ijn_n "ffi i M 1*1 ii ii i~\ ^_i i ? SSs Jh •a S.b _cs a a tJD 41 C — •- 92 VICTOR E. SHELFORD. belonging to the anurans. In a few experiments with Chrophilus nigritus, it was found that this frog possesses a similar set of reflexes. The ground beetles (Pterostichus adoxus (i) and pennsylvanicus (2) were very sensitive to the dry air, P. pennsylvanicus par- ticularly so (Expt. 14, Chart V., p. 97). Here two specimens of pennsylvanicus tried the dry air a few times and then began to hesitate and turn back. The individual of adoxus was less active but after one trial of the driest air and one turning from the medium, came to rest in the moist air as did the pennsylvanicus after a number of trials and turnings and at the end of fifteen minutes. The loss of the stock of pennsylvanicus necessitated the repetition of the experiment with adoxns alone. In the second trial, the beetles when put in the center dashed lengthwise of the cage once and by chance all bunched together in the driest air, due to thigmotaxis and gregarious tendencies. Soon they became very much stimulated and single individuals dashed to the moist end of the cage and back with great speed. This bunching interfered with the reaction and at the end of the 20- minute period only two had worked out a preference for the moist air. Fontaria was clearly stimulated by the dry air after less than five minutes exposure in the observation tubes. The individuals in the medium air showed less activity than those in the driest. Individuals in the moist air remained quiet most of the time. In most cases in the gradients (Chart III., Expt. 56, p. 91) the Fontarias entered the driest air a number of times and then began to show greater activity in the dry end, to stay a shorter time there, to hesitate upon entering, and to turn back occasionally. Their reaction to the dry air was clearly negative as shown by a time preference for the moistest air, and the haltings and turnings even, though the ability to orient in the gradient seems poorly developed. Snails and slugs are not good for gradient experiments of short duration because of their sluggishness. The slug (Philomycus carolinensis Bosc.) was usually inactive in the moist air but quite active during the first half hour in the dry air. Here the ten- tacles were withdrawn in less than five minutes and remained so REACTIONS OF ANIMALS. 93 until death ensued. Inactivity or very slow movements were characteristic of the second, third and fourth half-hour periods. Marked reduction in size was evident in about two hours. They died after two and one half hours. Comparable results were obtained with several snails. Polygyra thyroides Say, when active at the time put into the observation tubes, behaved as follows. In the moist air activity continued intermittently. The animals retreated into the shell from time to time and usually remained stuck to the side of the tube for half an hour or more. In the dry air withdrawal into the shell followed in five minutes but partial extension sometimes con- tinued from time to time during the first half hour. In one individual, a fresh epiphragm was formed at the end of two hours and shrinkage and withdrawal into the shell continued during several hours of observation. Polygyra palliata Say behaved similarly. Active individuals put into dry air became inactive but more quickly than thyroides. When put into the tubes in an inactive state and with strong epiphragms no activity occurred in either medium or dry air. In the moist air the foot was pro- truded after 45 to 55 minutes and creeping began after 70 minutes. On one occasion specimens of several species were taken from the same stock jar and placed in the moist air together, with the following results: Polygyra palliata Say and P. thyroides Say became active in 10 to 20 minutes, Pyramidula alternata Say in 80 minutes, Polygyra fraudulenta Pit. showed no activity at the end of three hours and forty minutes but was found moving 14 hours later, a night having intervened. The experiments were carried far enough to show that activity may ordinarily be induced in faint light by air nearly saturated with moisture but it is clear that other factors are concerned because occa- sionally it is not induced and when so, frequently does not continue. (&) Sand Dune Animals. Of the sand area animals studied, the common toad is least characteristic of sandy situations because toads belonging to thfe same species are found in moist woods, and because toads of the dunes breed in the pools and not on the dunes. 94 VICTOR E. SHELFORD. Furthermore these toads are probably physiologically dif- ferent from toads of moister situations. The toads are the only sand animals used that clearly avoided dry air. The stimulation was less marked than that of the wood frogs and was not accompanied by striking or characteristic reflexes. Activity was greater in the controls of gradient experiments where uniform current was used than where still air was used (Chart II., Expt. 23, p. 90). In the gradient the toads were negative to drier air (Table III.) but turned back much less defi- nitely than did the salamanders. Rapid random movements appeared to be characteristic. The ability to orient in the gradient is poorly developed. The spiders (Geolycosa) appeared not to be affected by the moist air. When observed in the tubes, no differences between the individuals in the different conditions could be noted. In the gradient experiments (Table II.) a positive reaction to dry air was clearly shown when the spiders were induced to move about. It was necessary to select individuals of the same sex and of about the same size, as these animals manifested a very striking repulsion for one another and when one spider came near to another one or both darted away with great speed. Thus when one spider moved, three being present, more movement usually resulted and if none of the spiders was killed in combat the experiment resulted successfully. In many cases however, espe- cially when differences in size or sex occurred, some of the spiders usually were killed before the experiment ended. Experi- ment 29, Chart IV., shows a typical graph characterized by the erratic dashes made by two individuals when meeting. The digger wasps (Microbembex') were likewise slightly posi- tive (Table II.) to dry air, though their chief reaction was dig- ging (Chart V., p. 97). The digging reaction took place in the medium and moist air but not in the dry. There was no special activity in the dry air. In Chart V. the reaction of three indi- viduals in a gradient is shown; the crosses indicate that the wasp was digging at the half minute opposite which the cross appears. It will be noted that the crosses are all in the moist and medium sections. A few experiments were tried with grasshoppers from sand REACTIONS OF ANIMALS. 95 i 111 ,i ill i, 1 1 !,, ,,,! , m i In i ill In ml In!, lit i n In nil IIIM iliiiiilli.Ml.il --•-,n--..»o;=H3s;2. * = «• 1 1 I T" ' " I ' ""I" '"I I" " T" " ' I 1 1 1 1 1 1 1 . i !" 1 1 3 I . I. II Ullll I.I I nil I Jj . . . . 1 O = E = i^; . I I I . . II , I I I , . I . . 1 1 1 1 1 1 1 1 1 1 1 1 1 1 ° = s 3 . : >« ; 2 = = Jlltllllillll IhlllllMlllillirlll O 9 a *' •- m. • 1 ..... |i""J!""X''il"jlll!iJ|lUirriUllIlil"J ..... 1 ..... im.MiJii'1 96 VICTOR E. SHELFORD. ters VI. — •o g O w CS cs M > HH 1 CO _OJ 3 03 E H c co • ^H -4-> 4) C CO CU O J3 4-» 1 60 _g S •3 cu "u •+J 3 _c _g CO O M | 0 • f 0 >. O M o O 4J CU 4-» CO «4-l "3 0 £ s 'S rt M rt l-( o C 3 CU < •a •a • o o JV £ 3 a CS 0 -i-> H 4-> 0 y «4-( 0 (4-1 c O _o CO -1_> _a _O 'C _ O rt QJ L^ cu •a u u IT] CU CD 4-> (U c« « ^ M C CU -*-! i | 3 c C/3 *E a •4^ M S CONTROL. •]OJJUOf) O f*O <*O ^f (M M W CS PI PI o o o PI PI M O PI o 1-1 o o> o P) PI PI M HI O . c • C rti "0 U <-" uJg IX •jaXjQ 01 . ro PI ro t1^ Pi ro ro PO Suipuodsgjjo^ n t»» CN t^- M ro ^f ro OO PI ro ro o\ PI O\ to O\ O O\ r*~; ^f ro PI M ro O P) Turned Back From •jaXjQ oj c^ O\ o ro M 00 O O O 00 H PI C t-t t- J\ 0 v. H HI M * •jgjsioj^j oj M 10 ro 10 o oo' O O O oo M Hi HI M EXPERIMENT. ,-SSSk W M W CS pi PI O O O O PI PI P) H O PI O M HI O O\ O PI PI PI PI HI M 0 N o i'gl •XJQ W O\ !>• JO MM M ? O O O\ PI to to ro O IO PI M HI 00 ^f HI rj- P) PI PI ro 0 •uimp9j/y[ o» O w 10 PI r^ ff>cc o M M PI M M P) V£ 5 O O TT ^ 0 •1S!oW O *O W O oo IO •* oo ro oo Pi Pi O tr> ro ro PI O O oo PI PI M ro ro HI ro -t P) u *H Q •jaAaQ O\ "O ON w M W M O f- PI O M « O oo t*~ vO ^C ^ P) •jajsioj^ H CS M M M M vO PI W O H M ro PI ' HI H j- oo HI to oo H H PI HI 4 M Relative Humidity in Per Cent, of Saturation. "AjQ O M O M 1 1 •™,™ 0 O 1 1 *iSTOJ.^ 00 oo 1 1 Evaporation in c.c. During Experiment. *XjQ r^ iri 10 O ^O t**» t*^ ON O to in o M r» o O\ oo oo t-- N O M O HI o o o o o O u O O o O O O O O O •™,™ PI PI P) IO CS PI PI Tt ro to O O PI TT ro ro PI ro P) O O Pi Pi PI PI ro ro Pi PI PI 6666 O O O O 0 0 O O 0 O O O '1SIO]\[ PI PI PI PI o o o o PI O Pi PI PI PI O O O O O PI PI PI PI P! PI O O O O O O 0 O O O 0 o o o o O o o o o o o yi 'o O. V. 4 M 0) Geolycosa CU Microbembex OJ •juguiugdxg jo -o^ ro PI ro i— < PI ^f ^j" ^D > 00 O\ O PI PI PI ro ro CU < PI Tf \O ^ ro ^3" PI CN PI PI ro ro V REACTIONS OF ANIMALS. Microbembex — Dry Air Pterostichus — Dry Air Experiment 33 Control 33 Experiment 14 Control 14 97 CHART V. Compares the reactions of a diurnal dune animal (Microbembex) and a nocturnal beech forest animal (Pterostichus). The two differ both in habitat preference and daily habits and the graphs are strikingly different. The crosses in tracings for Microbembex indicate that the animal was digging at the half minutes indicated opposite. m areas but they rested in one position, the depth of the cages not being sufficient to enable them to hop. The bronze tiger beetle (Cicindela lecontei Hald.) was tried and gave a negative reaction to air evaporating 1.2 c.c. in 20 minutes (3.6 c.c. per hour) and a positive reaction to air evaporating 0.52 c.c. per 20 minutes (1.56 c.c. per hour) produced by heat and current. They were so active that it was necessary to take readings of the number in each third every ten seconds. 2. Rapidly Flow Air. The rate of evaporation is markedly influenced by rate of flow, and particularly there is a marked difference between a barely measurable movement and a very slight breeze such as .52 meter per second (i.i miles per hour). Table III. brings 98 VICTOR E. SHELFORD. out these relations in a rough way. More accurate results were not easily obtainable because of the difficulty of controlling humidity and temperature and accurately measuring the flows. It will be noted that a doubling of velocity was accompanied by a doubling of evaporation only in the cases of .052 and .104 meter per second. A breeze of .208 meter per second can be distin- guished by the skin of the hand but usually lesser flows cannot. TABLE III. Showing the relation of evaporation to the rate of flow and to relative humidity under the experimental conditions, together with the relative rate of increase of evaporation and velocity. (0.52 meter per sec. equals i.i miles per hour, 0.68 equals 1.5; o.io equals 0.2.) The equipment is not accurate enough to make this more than a general guide. Pressure was not read. Approxi- Relative Ratios. Approximate Flow in Liters per Minute. Approximate Velocity in Meters per Sec. mate Evap- oration in c.c. pei- Hour. Tempera- ture in Degrees C. Humidity in Per Cent, of Satura- tion. Increase in Flow. Increase in Evapo- ration. 1.9 .OI2 •25 22.4 50 I 1.0 3-9 .026 .40 22.2 53 2 1.6 7-8 .052 .75 22.2 53 4 3-0 15-6 .104 1.50 22.2 53 8 6.0 31-2 .208 2.OO 22.2 54 16 8.0 62.4 .416 2.6O 22.2 53 32 10.4 78.0 .520 2.90 18.8 45 40 1 1. 6 15-6 .104 1.4 19.0 35 I I.O IO2.O .680 3.7 19.0 35 6.5 2.5 15-6 .104 1.6 18.0 45 i j 31-2 .208 2.3 18.0 45 2 i-4 Average : .104 1-55 20.1 49 I I.O .2O8 2.15 20.1 49 2 1-35 Compare with Schierbeck ('95), p. 221. Changes in rate of flow give greater difference in evaporation below .208 meter per sec. than above. Schierbeck's table gives accurate data from .88 m. per sec. to 4.23 m. per sec. for exposed water surfaces (see Livingston, G. T., '08, '09). (a) Physiological Effect and Reactions. The same species were studied and the same general physio- logical effects noted as where the difference in evaporation was due to dryness. Only slight evidence of mechanical stimulation occurred in the case of Fontaria in the gradient. The sala- manders showed the same kind of activity and symptoms of drying in the rapid as in the dry air. The animals pushed REACTIONS OF ANIMALS. 99 "8 o cfl 3 A ffl O w v> "' o o 0 • § 2 fco ss •3 • 2o 01 2 3 cd •O "-1 1) o> J5 -^ "*"* O *-" o. « IS Z! O of n3 en r : *j .2 •i I 03 (Tj t» M ^ S £ 03 O O > S>» S •2 x t lis " M .2.3 CONTROL. '[OJJU03 jo ajnjBjadtnaj, M CO M N M O) o\ M Ov t M N M PI M M N M 01 •—!/: C aj"C U £ t, u.|3 PH '4S!H CO M •3- M M to to M o HH \O N N co co rj- \O 00 CO tN M CO O M CO 0) O O) too •* •* •mnip3j\[ to r~- oj O H-t O M O O I> o 00 •* r— 10 oo M M M H O M d CO 01 \O oc M r~ M 10 CO P) (N t N •JAO'-[ r— M oo i— O 00 O\ r~ O. to M O 01 •* t~ CO O CO 10 Tf COO •* ^ 10 •* •* M T)- r~ r- LO O H CO CO M ro 3 rt ffl a •S8 1 H "HS!H M 00 10 rj- M M 0 to M M O M CO M O >o M to l-^ o\ o\ 0 •MOq 0 t N O o o O o o o O O O ^t Ox IO o o\ •* M 10 t-l M Relative Humidity in Per Cent, of Satu- ration. '43!H t— \O *f to N IO vO o o o O 10 M 0 t^ t o IO M IO t^ \O Tf >O M to r- r~ •«? r^ r^ •* •uimp3j\; I~- O ^r to M IO vO O o o O to 1-H O r^ •* o IO M 10 r^ \O ^ to M to t- t^ •* * r^ ^ •MOI r- oo •3- c\ vO vO O 00 O O\ 1 t^ t oo o\ t- oo •* O\ r-oo •* O\ Evaporation in c.c. During Experiment. "HS!H to t~- TT O\ 6 o M r^ 6 vo t 6 to r- •* o\ 6 6 M t— 6 IO o M IO c\ d M O IH 10 t^ O O M 6 M q M IO IO q q H M to q M •uinipaj^ to o M \O 6 r^ ro a o M O 01 >o M ^t t^ co o \o C\ *t r- O COO O4 •* t~. t^ CO CO t~ ro 'AVO'J +j + 4J + % + 4-1 V +1 +1 •ji 1 I Plethodon cinereus a; on 2 P. glutinosus v> X a K, (U M 03 Fontaria oo •* -4J a a ^ *^* X !^ a ^ u 's M 0 rt b, V oe nj 4 w V M cfl Geolycosa 0 6C rt •}U3UIIJ3dxj[ JO -0^1 O M to r~ o < M to PJ O IO t— 01 > < 3 X 4> £ 'hi OJ 4) CO o ^_ n .3 o X ^_* 4> M a J3 3 •3 to 4) "u ,3 3 ,j5 {- 1 cd to 4> | 3* £ 4) 4-» ~ 0 hi a 5 o o "cd to cd "o hi -4_> TO to CO ? cd cd to 4) .£> — hi cd ^j O CO cd £ u 4) hi "o | a C z O U — c? cd 0 "o CO p**^ CO a y w J _O cd 4) M cd ffl < h •o > £— i 41 cd 3 >, T3 2 41 a -— - ~ 1 41 4-) 1 £ - cu Q 4-> 0 CJJ hi were "M "cd | 2 hi 4) o hi 0 ^" 3 cu 4) O O 3 'o ~ to 3 •o TJ 0 *|~I 3 ^ cd O cd /. hi •4-J 0 E cd" 4) cu 41 "o -H* 4) 0 41 hi ** .3 >2 3 "^ Cu .0 C r& cd •a o 41 cd 0 *J £ M | difficult hi o "S a to 41 i M OJ 2 o "S 3 •3 £ a cu o CONTROL. ^c V E [H u- o c V O i_t u PH •UIJEM OJ 3ujpuods -3.1.103 •* i- HI M O ro Cv f ro ro Ti- ro O\ M ro -3- o •* ^ CS t~- ro 10 'i- 00 ro O O vo ro (N •O - •iun;p3j\[ oo oo »fr 00 o\ oo 0 oo oo r r oo 0\ oo oo 0\ 0 oo C\ oo oo oo O O O 00 O\ oo oo O O\ oo O Evaporation. • IUJB^ to vo 10 r- to 0 O O vO O O -o N 1O IO O vO to o o o \O O t~ O o o o 00 O O t^ O O O 0 0 o O O o o o o o o o o •uinipaj^ to vo ro ro 10 ro 10 10 T •* 10 •* O to ro •* t^. ro to to 10 ^ -=t rf IO "3- O VO ro •>*• t^ ro O O O O O O O O O O O 0 O O O O •JSIOJ\[ CM N O O N O N M O 0 M 0 N N 0 O N O M cq Hi < 10 10 r~ to t-t V < N TT vO t^ t^ t^ hi 41 < 10 o\ \O r- IM < I IO2 VICTOR E. SHELFORD. Pletkodon cinereus is quickly affected and over-stimulation and coiling appear within one minute while in the dry and rapidly moving air from two to ten minutes are required to bring about the same result. All the activities and symptoms of loss of water are the same as in the dry air (p. 86). In the gradients (Table VI.) one of the graphs shows less clear avoidance of the hot air due to over-stimulation and some loss of correlation in movement. Otherwise the reactions were entirely similar. No temperature experiments were performed with P. glutinosus but the difference in behavior in hot and dry air in the tubes was comparable to that of cinereus. Rana behaved exactly as in the dry air in the experiments, with only five degrees increase, but the graph of the gradient experiment with ten degrees difference is like that of Plethodon as a very clear orientation occurred. Fontaria showed the greatest difference when compared with the dry air. The activity was much greater in the hot than in the dry or moving air. Movement in the tubes was increased six times with an increase of 8° C. In the gradient a distinct orientation occurred, the animal turned back upon encountering the hot air. This was a decided difference from the reaction to dry and moving air (compare graphs of Chart II., p. 91). Some of the toads showed stimulation in the hot air. Activity was increased in the tubes and in the gradient experiments some of the individuals tended, to hop back and forth in the warm end (Chart II., Exp. 76) and only a weak negative reaction resulted. The detailed behavior of Geolycosa was not markedly affected by the difference in temperature and a slightly positive re- action was given. Cicindela lecontei was likewise positive in the gradient (see Table VII.) but in the tubes showed a greater tendency to "clean" the legs and antennae while in the hot air. 4. Death through Evaporation. All of the animals studied may be killed by loss of water. The results are given in Table VII. It will be noted that where records of size were preserved, the smaller animals died from loss of water much more quickly than the larger. This is perhaps due to the fact that the surface is greater in proportion to the REACTIONS OF ANIMALS. 103 CO •4_> a I 1 £ — a tn to 0) 1 co a; 3 •o o H Ol 01 * Ol 3 •3 V § V 0 -o-o ui w rr O t- 0) H M HO H M * * * * ro H Ol HI 6 6 01 C •uio ui qiSuaq O! O O 01^ 2 o o. W •sajn -UIJ\[ UI 3UIIJ •a u vo vo vo o to t>» \Q \^ ^* *^* co vo ro H Ol •* o CO ^O H 00 o 10 oo vo o 0 vo o O H M ro H O O 0 0 00 OO H H 0 00 oo HI 0 01 d O) § S O •3'0 UI o oo to t-- co oo r- >/•> ^^ Ov r- oo o CO ^)" O o o o 0 UOU^JOd'EA^ OO 01 VO H H 01 0 O Ol VO OO H O rj- m H 00 O 00 01 . "3 ; o .^N O M o J3 [f, U £: = : OJ g C -H) i> ° 1-1 - Q" M Q: ^ : : 4J M ca X Q •juaui M M 00 O\ Ol O\ o t^ r~ oo OvOO 00 10 co O\ O\ 3 ro O\ O\oo \ FIG. i. The islands remain distinct until their contents are transformed into free spermatids. The cytoplasm between the islands then disintegrates, producing large spaces in which lie the free sper- matids. The testis thus acquires a lumen which ever increases in size until the testis becomes a hollow sac full of ripe sperm. In the last stages, when the testis is a hollow sac whose walls consist merely of an epithelial coat, the giant nuclei are not to be found. However, there may be seen, scattered among the free sperm, irregular particles of disintegrating cytoplasm, possibly the remains of the rings of cytoplasm which formerly enclosed the islands. 136 ROBERT CHAMBERS, JR. The giant nuclei attain their greatest size in testes which are composed almost entirely of islands of spermatocytes in the various stages of maturation but in which no lumen yet exists. Simultaneously with the growth of the giant cell the nucleus and noticeably the nucleolus increase in size. A change also takes place in the staining reaction of the nucleus. The chro- matin network, which hitherto together with the nucleolus stained red writh safranin, loses that capacity and takes up light green, an acid stain. A well-grown giant cell thus pos- sesses a large nucleus with an enor- mous basic staining nucleolus and an acid staining nuclear network, the granules in the surrounding cyto- plasm staining red with safranin. Similar cells have been described, in literature on spermatogenesis, as • rudimentary ova. A significant fact, FIG. 2. however, which militates against such an interpretation, at any rate for the giant cells in the Simocephalus testis, is that they grow directly from spermatogonia and do not pass through the synapsis stage. The striking but superficial resemblance between these giant cells and growing oocytes is evidently due to the one function common to both, viz., that of an enormous growth in size. The ever-increasing size of the nucleolus during growth and its final dissolution in both types of cells favors the assumption that the nucleolus is intimately connected with cell growth. In the spermatocytes in Simocephalus where no growth occurs the spermatogonial nucleolus remains small during synapsis and early disappears. The same is true for Pandarus1 and for Cyclops? On the other hand, in spermatocytes wrhere growrth does occur, a growing nucleolus is described by Schmalz3 in an Ostracod. In this form the nucleolus grows during synapsis and during the sub- sequent growth period to disappear on the formation of the spindle for the first maturation division. 1 J. F. McClendon, Arch. f. Zellforsch., V., 1909. 2 R. Chambers, Jr., Univ. of Toronto Studies, Biol. Ser., No. 14, 1912. » J. Schmalz, A rch. f. Zellforsch., VIII., 1912. THE SPERMATOGEXESIS OF A DAPHXID. 137 The principal stages in the spermatocytic development of Simocephalus are indicated in Fig. 3. A resting spermato- gonium is shown in Fig. 3, a. The spermatogonial chromosomes appear to be slender more or less U-shaped rods. During metaphase (Fig. 3, b), they are too closely massed to be counted. I have no doubt, however, that they are considerably more than eight in number. Nuclei in synizesis (Fig. 3, c} show a decided FIG. 3. contraction of the chromatin threads. No growth occurs during this stage and no nucleolus is discernible. Intheprophase of the primary spermatocyte (Fig. 3, d), eight distinctly double rod-like chromosomes are evident. The chromosomes in this stage are very distinct being more or less regularly distributed just under the nuclear membrane and in over fifty cases counted the double chromosomes were constantly eight in number. 138 ROBERT CHAMBERS, JR. Fig. 3, e and /, show the primary spermatocyte in metaphase and telophase. The chromosomes then pass into an interkinetic resting nucleus, Fig. 3, g. The resolution of the chromosomes of the two daughter cells into a resting nucleus is not always synchronous. One may often find an interkinetic nucleus at one end of the telophasic spindle while the chromosomes at the other end are still massed in a densely staining body. There is no- doubt, however, that both ends pass into the resting state and form normal nuclei for one may find the entire contents of an island in interkinesis. And in still older islands all the cells pass into Metaphase II., leaving no cells behind. Fig. 3, h, repre- sents the nucleus of a secondary spermatocyte in prophase. The chromosomes are approximately eight in number. They are shown in Fig. 3, i, in metaphase. Fig. 3, k-o, show the cellular elements lying in part of the lumen of a maturing testis and a portion of the adjacent wall. The secondary spermatocytes, Fig. 3, k, are shown in telophase. At / are spermatids with vesicular nuclei. Their arrangement in islands is better shown in the upper part of Fig. 2. Upon the disintegration of the surrounding cytoplasm, the spermatids come to lie in the lumen of the testis. Here, Fig. 3, m, the nuclei contract somewhat and become more densely chromatic. In some of the spermalids, Fig. 3, n, n, the contents of the nucleus collects into a compact eccentric mass, which finally disintegrates and disappears. In other spermatids the nucleus remains vesicular and it is this second type only that is to be found in the distal end of the vas deferens of a mature testis. That approximately half of the spermatids degenerate is the impression gained by the examination of sagittal sections of entire testes. Lepeschkin,1 in a brief paper in Russian, kindly translated for me by Dr. M. Scholtz, of Cincinnati, on the spermato- genesis of the Daphnid, Moina rectirostris, describes occasional degenerating cells not only among spermatids but also among spermatocytes and spermatogonia. He speaks of the uniformity 1 W. D. Lepeschkin, Mem. Soc. Amis Sc. Nat. Anthrop. Ethnogr. Univ. Moscon, Vol. 98, Sect. Zool., Vol. 3, No. 9, 1907. THE SPERMATOGENESIS OF A DAPHNID. 139 of the cellular elements in young testes and of the formation of cysts enclosing spermatocytes, all the contents of a cyst being in the same stage of development. He does not describe nuclear changes, giving, as an excuse, the diminutive size of the cellular elements. His figures, showing degenerating cells, are not con- clusive for one might easily confuse synizetic and interkinetic nuclei and possibly different stages of the giant cells with de- generative appearances. Occasional abnormalities do occur in any gonad but in healthy normal specimens studied by me I have been unable to discover degenerative appearances except among spermatids. Two classes of spermatids, of which one only produces func- tional spermatozoa, are described by McClendon1 in his studies in the spermatogenesis of Pandarus sinnatus, a parasitic Copepod, a species which exists only in the sexual state. According to McClendon some of the spermatids are transformed into "nutri- tive spheres." The "spheres " are a constant feature in Pandarus and the proportion formed is very large. On the assumption that spermatids occur in male- and female-producing classes, this condition might possibly disturb the sex ratio of the species. This is not true for I have collected large numbers and have always found the males and females in approximately equal numbers. A personal study2 of the spermatogenesis of Pandarus sinnatus has convinced me that McClendon's "nutritive spheres" are derived not from spermatids but from spermatocytes during interkinesis. If, then, we assume post reduction for the sex- determining factor, a condition which obtains in most forms where a distinct accessory or "sex" chromosome occurs, the formation of the "nutritive spheres" will be from neutral cells and should, therefore, cause no disturbance in the ratio of male- and female-producing sperm. In Aphids3 two classes of sperm are produced in the male differing in the presence and absence of an accessory chromo- some. The class which does not possess the accessory chromo- 1 J. F. McClendon, Arch. f. Zellforsch., V., 1909. 2 R. Chambers, Jr. In press. 3 T. H. Morgan, Proc. Soc. Exp. Biol. and Med., 5, 1908, and Science, Vol. 29, 1909. W. von Baehr, Arch. f. Zellforsch., III., 1909. I4O ROBERT CHAMBERS, JR. some degenerates. By inference from other groups of insects this is the male-producing class. The other class produces functional sperm which, entering the egg, give rise to females only. These sperm correspond to the female-producing sperm of other insects. In Simocephahis no accessory chromosome is to be distin- guished, the sperm being apparently all alike in their chromo- some number. No means as yet have been found to distinguish two classes. The presence of degenerating sperm in the lumen of the testis does not necessarily prove the existence of two classes of sperm. However, as the functional sperm enter eggs which develop only into females, the assumption is permissible that these are the female-producing sperm and that the male-producing sperm are those which degenerate. Vol. XXV. August, /p/j. No. 3 BIOLOGICAL BULLETIN LOCAL DISTRIBUTION OF GRASSHOPPERS IN RELATION TO PLANT ASSOCIATIONS. ARTHUR G. VESTAL. CONTENTS. I. Introduction 141 Physiography of the Region 142 General Character of the Vegetation • 143 II. The. Plant Associations and the Grasshoppers which Occur Within Them 143 Associations of the Northeastern Conifer Province 144 Associations of the Eastern Deciduous Forest Province 145 Local Associations 147 Ruderal Associations 148 Arrangement of the Associations as Habitats 150 III. Summary of Habitat-Distribution of the Species 152 IV. General Discussion 155 The Assignment of Terrestrial Animals to Habitats 155 Estimation of Relative Numbers in Different Habitats 164 The Relation of the Animal to Plant and Animal Communities 166 Successional Relations 169 Geographic Relations 1 73 Seasonal Relations 1 75 V. Summary 177 VI. Acknowledgments and Bibliography 178 I. INTRODUCTION. During the summer of 1912, while at the biological station of the University of Michigan, at Douglas Lake, the writer found opportunity to study the relation between local distribution of grasshoppers and the plant associations. The group studied includes the three subfamilies Tryxalince, (Edipodince, and Acridiince, of the orthopteran family AcridiidcB. Habitat-distribution of Orthoptera has received considerable attention. Published accounts of Orthoptera contain the bulk of the data on habitats. The studies of Morse ('04), and of 141 142 ARTHUR G. VESTAL. Hancock ('n), classifying these insects on the basis of habitat, have been reviewed by Shelford ('126: 352). Biological surveys of certain regions, as the Michigan surveys, have included local distribution of grasshoppers (Adams, '06, '08; Ruthven, 'n; Hart and Gleason, '07). Studies of animal communities have included data on grasshopper distribution (Shelford, '120; Vestal, '136). In these two studies not all of the associations of the region were considered. In many of the above accounts the concept of the plant association as a habitat has appeared inci- dentally or not at all; this concept was used in a study of local distribution of birds, by Gates ('11). In the present study all the plant associations of the region have been included, and the plant association has been used as the index of the habitat. In general the results indicate that the important factors of local distribution are, in initial stages of development of the vegetation, physical conditions of the environment; in advanced stages, vegetational conditions; in either case, the character of the plant association is the index to local environmental conditions for grasshoppers. Collection data are in another paper (Vestal, '130). Physiography of the Region. Douglas Lake is situated less than twenty miles south of the northern tip of the southern peninsula of Michigan, in Cheboygan county. There are two main physiographic types, the first being old beaches and lake bottoms of the Nippissing and Algon- quin Lakes, and the second morainic in origin. The soil of the first type is almost pure sand; the relief is very slight. The morainic lands have typical rolling topography; they are of mixed composition, usually with loam or sandy loam surface soil. Streams are small and few, but on the whole the region is well drained, the soil being porous. Swamps occupy a very small proportion of the area, and are usually drained. Undrained swamps are few and small in the immediate vicinity. The lake is about four miles long, with irregular contour, and lies north of a chain of lakes which in recent geological time connected Lakes Michigan and Huron. Douglas Lake was itself, previous to this time, a part of the connection. Much of the shore line is sandy beach, with occasional sand-bars and spits, GRASSHOPPERS IN RELATION TO PLANT ASSOCIATIONS. 143 and in several places lagoons. Wave action is considerable on exposed parts of the shore-line, but the exposed sand has been very little acted upon by wind. Beach dunes occur sparingly and are of small area. General Character of the Vegetation. Much of this part of Michigan is covered with two general types of vegetation, which give rise to the common terms "pine lands" and "hardwood lands." The pine forests have been developed on the more sandy areas, the hardwoods on the loamy or clay moraines. "Hardwood lands" are of much greater agricultural value, and the more extensive farming districts are in morainic regions. The pine lands have been largely def crested , and fire has also been very prevalent, so that many of the original pine forests have been replaced by growths of the large-toothed aspen. In the immediate region of the biological station virtually no pine forest remains, though scattering growths of pitch pine are not rare along the beach, and a few pines are to be found scattered among the aspens. The aspen forest occupies more than one half of the entire area studied; the hardwood forest and hardwood clearings somewhat more than a fourth of the area, and other kinds of vegetation considerably less than one fourth. These other kinds of vegetation are grassland areas, cultivated fields, meadows and sedge growths, cedar bogs, and open peat bogs. The plant associations have been studied in detail by Gates ('13), with reference to a much larger area than has been covered in the present study. II. THE PLANT ASSOCIATIONS AND THE GRASSHOPPERS WHICH OCCUR WITHIN THEM. Associations of two vegetation regions or vegetation provinces (Gleason, '10: 42-45) occur within the area. The coniferous forests, the aspen forests, heaths and bogs of the region, typical of the northeastern states and much of Canada, represent the Northeastern Conifer Province. The deciduous forest (hard- woods) and the herbaceous and thicket growths of hardwood clearings represent the Eastern Deciduous Forest Province. While primarily consisting of these two geographic elements, the vegetation also includes local associations, particularly those 144 ARTHUR G. VESTAL. determined by local conditions of moisture; these are usually not characteristic of any one vegetation province; and ruderal associations, composed of introduced plants. Plant names used are those of Gray's Manual, seventh edition. Associations of the Northeastern Conifer Province. The culminating type of vegetation in the Northeastern Province is the balsam-spruce-birch forest, which is developed successively through lichen, heath, and different pine stages, in xerophytic situations; and through bog and bog-forest stages in water and wet ground. Only those associations are listed which are represented by distinct areas in the region studied. Cham cedaphne Association (Gates, '13: 57). — No grasshoppers were taken in the open bogs of cassandra. In Ontonagon county, northern Michigan, thickets of cassandra, alder, wax myrtle, high-bush cranberry, etc., are the habitat for Podisma glacialis .and Melanoplus islandicus (Morse, '06: 70). Thuja Association (Gates, '13: 66).- — The cedar or arbor- vitae growth known as Rees's bog, near the shore of Burt Lake, south of Douglas Lake, was studied. In this bog the peat soil is less than two feet deep, the substratum being sand or gravel. It is drained through the porous soil into Burt Lake. No bare soil is exposed, but partly moss-grown logs lying upon the surface are not infrequent. Trees are usually not more than twenty feet in height, but are close together, and cast a deep shade. Thuja occidentalis is the dominant species; Larix laricina, tamarack, and Picea mariana, black spruce, occur sparingly. The surface is a thick carpet of Sphagnum; many peat-bog plants, as Cornus canadensis, orchids, ericads, etc., are present, more abundantly in less shaded parts. The only grasshopper species taken in this bog. is Melanoplus islandicus. Aspen Association (Gates, '13: 77). — The extensive sandy pine lands of the region are nowr, as a result of fires and cutting, practically all occupied by the aspen forest. The dominant tree is the large-toothed aspen, Populus grandidentata. Other fre- quent species are paper birch, beech, red oak, hard maple, red maple, wrild cherry, Primus pennsylvanica, white pine, pitch pine. The undergrowth, quite similar to that of the pine forest, is chiefly composed of bushy blueberry plants, Vaccinium penn- GRASSHOPPERS IN RELATION TO PLANT ASSOCIATIONS. 145 sylvanicum, which are close together and usually less than six inches high, and of the taller and less numerous plants of the bracken fern, Pteris aquilina. Other plants are the bush honey- suckle, Diervilla lonicera, and several grasses and composites. Near the lake beach patches of bearberry, Arctostaphylos uva-ursi, and of dwarf juniper, Jimiperus horizontalis, form small heaths. Over all the open aspen growth, a considerable proportion of bare sandy soil is exposed, in the interspaces between plants. Dead leaves, partly decayed and partly burned stumps and logs, litter the surface. There are a number of bare roadways. The tree growth is irregular, being entirely absent In frequent local areas which vary in size from small unshaded patches be- tween trees to areas thirty meters in diameter. In such places there are indications that the undergrowth is practically inde- pendent of the trees. In the older tree growths the hardwood species have assumed control, indicating development into hard- wood forest. Ground conditions are more like those of closed forest. In the treeless parts of the association, in the bracken-blue- berry growth, Melanoplus angustipennis is the common grass- hopper species. Melanoplus atlanis and Camnula pelludica are common, and M. bivittatus occasional, along roadways. M. luridus is found sparingly in scattered aspen growths. Scirtetica marmorata occurs usually on or near the lichen-covered surfaces. M. fasciatus is more often found in the closed forest. On the sandy roads, and sparingly in the sandy interspaces between the plants, are found M. atlanis, M. angustipennis, Dissosteira Carolina, Spharagemon bolli, Circotettix verruculatus , Arphia pseu- donietana, Hippiscus tuberculatus. Associations of the Eastern Deciduous Forest Province. The most highly developed form of deciduous forest vegetation is the beech-maple or beech forest, well represented in the region. This develops, on dry soil, through herbaceous, thicket, and xerophytic oak stages, usually, followed by mesophytic red oak and maple stages. The following associations were studied in the Douglas Lake region: Herbaceous Associations (Gates, '13: 75). — The common fire- weed, Epilobium angustifolium, is the first plant to establish itself 146 ARTHUR G. VESTAL. on newly cleared or burned hardwood land. The clearings are soon overgrown with herbaceous plants, many of them intro- duced, as mullein, Verbascum thapsus. These burns and clearings vary considerably both as regards physical conditions and plant composition. They are only temporary stages in the process of reforestation. Usually they are dry and hot, being fully exposed to the sun, but often sheltered from wrind by surrounding forests. Grasshoppers are numerous, both in individuals and in species. In approximate order of abundance they are : Melanoplus atlanis, Camnula pellucida, Dissosteira Carolina, M. bivittatus, Sphara- gemon bolli, Circotettix verruculatus , M. lurid us, Chloealtis con- spersa, M. minor. Only the first two occur in any considerable abundance. Thicket and Bramble Associations (Gates, '13: 76). — The shrubby plants which replace the herbs in clearings are prin- cipally blackberry and raspberry, Rubns spp., red-berried elder, Sambucus racemosa, and young seedlings and shrubs of hard maple, Acer saccharum. These growths often form a dense tangle which is almost impenetrable. In dense parts of these thickets an occasional Melanoplus bivittatus would be seen upon a leaf at the top of a shrub. Where the ground could be seen, Melanoplus atlanis, Camnula pellucida, and Chloealtis conspersa were also to be found. Beech-Maple Association (Gates, '13: 71). — The beech-maple forest is dominated by two tree species, Fagus grandifolia, which occurs in places in nearly pure stands, and Acer saccharum. Tsuga canadensis, the hemlock, is also important in places. Ostrya virginiana, Betula lutea, and Tilia americana are infrequent species. In the deeply shaded parts of the forest, the under- growth consists mainly of young seedlings of Fagus and Acer, with small plants of Maianthemum canadense and Mitchella repens. partly hidden in the thick carpet of fallen leaves. In the sunlit spots the undergrowth is taller, with Acer pennsylvanicum and Sambucus racemosa. Many other forest plants occur. Stumps and logs are common, but no bare ground is exposed. The relative humidity is high. Exposure to sun and wind is at a minimum. Melanoplus islandicus is a characteristic grass- hopper species, of the deeply shaded parts of the forest. It is probable that Podisma glacialis variegata occurs also on shrub- GRASSHOPPERS IN RELATION TO PLANT ASSOCIATIONS. 147 bery. In more open parts of the hardwood forest, and near the edges of clearings, most of the grasshopper species that are common in clearings were also found, in very small numbers, however. Ravine Forest Association (Gates, '13: 66-70). — This associa- tion is best developed in the ravine occupied by what is locally known as Big Springs. A number of springs, fed by the under- ground drainage from Douglas Lake, to the north, form the head of Carp Creek, which runs south to Burt Lake. The ground is very wet, and is occupied by bog plants and deciduous forest undergrowth. Trees are Tilia americana, Tsuga canadensis, Fagus grandifolia, Ostrya virginiana, Betula lutea, Acer saccharum, Abies balsamea. Succession has proceeded from the coniferous bog forest toward the beech-maple association. Podisma gla- cialis variegata, recorded from Carp Creek, probably was taken in this ravine forest, Melanoplns islandicus probably also occurs in the same association. Local Associations. Dry Beach Associations (Gates, '13: 55). — Quite a number of the strand plants of the sandy margins of Douglas Lake belong to an assemblage which is characteristic of the sand shores of the Great Lakes. Three of these independently form associa- tions in the region. These are Elymus canadensis, Ammophila arenaria, and PotentUla anserina. Elymus forms a low, narrow dune about one eighth mile long, at the southeastern end of the lake. It and Ammophila are found just above the level of wave action at many places on the shore. Sandbars on the north shore were partly covered by Potentilla growth. In these beach associations the soil is pure sand, dry and shifting, with full exposure to sun and wind. The vegetation is dry and scanty. Grasshoppers typical of these situations are : Melanoplns atlanis, M. angustipennis, M. bivittatiis, Camnula pellncida, and Disso- steira Carolina. Two species not taken at Douglas Lake, Spharage- mon wyomingianum (Thomas), and Trimerotropis maritima (Harris), are very characteristic of beaches and dunes of the Great Lakes ; both are recorded by Shull from the Saginaw Bay region ('n: 225-226). Marsh Associations (Gates, '13). — Communities of marsh 148 ARTHUR G. VESTAL. plants bordering lakes, streams, and certain open bogs are usually composed of only one or a few plant species. These growths, depending upon local conditions of moisture, are frequently small in area, and different stations are not always alike. Condi- tions of shade and of soil are variable. No grasshoppers were found upon the marsh plants which grew standing in open water, although in dense growths individuals sometimes stray beyond the shore. The numbers of grasshoppers vary considerably in different stations. They are most numerous in tall, rather close, sedge or grass growths. Stenobothrus curtipennis is the characteristic grasshopper of littoral situations. Melanoplus atlanis, M. bivittatus, and M. differentialis are less frequently found. M. femur-rubrum, though not taken by the writer, is often found in such places. The Melanopli are more numerous in the higher and more open parts, while Stenobothrus occurs farthest out towards the water. Ruderal Associations. In the Douglas Lake region much of the native forest has been removed by cutting and burning, and its place has been taken by cultivated crops and weed growths. In addition many plant associations, though not destroyed, have been materially changed, and native animals have also been much affected. Secondary successions are principally due to interference by man. Most of the sandy land is now occupied by aspens as a result of destruction of the pines. One of the most important plants favored by artificial conditions is the blue-grass, Poa pratensis, which enters into nearly all ruderal growths. Other species of Poa probably occur in the region. Ruderal Dry Grassland Associations. — Abandoned fields, dry pastures, roadside growths, and modified aspen undergrowth are the common forms of ruderal growth in dry situations. They are necessarily very different in physical conditions and plant composition, depending upon differences in original status and subsequent modification. In certain large areas west of Douglas Lake, near Pellston, a growth of bluegrass has almost entirely replaced the aspen association. Near the lake bluegrass invades the aspens along roadways, and one can trace long abandoned roads by the presence of this plant. Other weed species are GRASSHOPPERS IX RELATION TO PLANT ASSOCIATIONS. 149 numerous. Grasshopper species of ruderal dry grassland are, in approximate descending order of importance : Melanophts atlanis, Camnula pelludica, M. bivittatus, Dissosteira Carolina, Arphia pseudonietana. All of these species, with the possible exception of the last, are more abundant in ruderal than in native vegetation. Sparsely Vegetated or Bare Soil. — The condition of bare soil is much more frequent, in humid climates, as a result of disturbance. Plowed fields and constantly trampled paths and roads are the commonest areas of bare soil, and these furnish suitable habitats for those species which normally rest on bare soil, not on the plants. The (Edipodince are of this habit to a large extent. The character of the soil is important for oviposition, and though the bare soil grasshoppers are more or less migratory, certain of them are found associated with particular types of soil. Hart gives the soil-preferences for a number of the Melanopli in a sand region in Illinois ('07: 214, 215). It is to be remembered that bare soil as a habitat is not sufficient; nearby vegetation is necessary, and in this region is always present. Where exten- sive areas of bare soil occur, grasshoppers are very rare except at the border. The insects are conspicuous in bare places, as roads, but are more abundant in the interspaces between plants, in open growths. Areas of bare soil differ from the sparsely vegetated areas merely in degree, and are really the same kind of habitat. In the region studied considerable bare soil is exposed along the beach, but in the aspen association roads are the typical areas of bare and sparsely-grown soil, as also in the hardwood district. Interstitial grasshoppers which are conspicuous in bare soil are: Dissosteira Carolina, Arphia pseudonietana, Circotettix verruculatus , Spharagemon bolli, and Hippiscus tuberculatus. Those of open grassland, frequently found on bare soil of inter- stices, and less frequently on bare soil of roads, are: Camnula pellucida, Melanoplus atlanis, M. bivittatus, M. angustipennis. Meadow Associations.- — Ruderal meadows and swales, like the native marsh and littoral associations, are variable in character. Probably the most common type is the bluegrass-white clover meadow. It is found in wet pastures and along boggy roads; it forms a very low, dense carpet, resembling a closely trimmed lawn, and probably always depends on constant cropping of I5O ARTHUR G. VESTAL. cattle. It is of common occurrence on the north shore of Douglas Lake, and near Munro Lake, several miles north. In this type of meadow Melanoplns atlanis and Stenobothrus curtipennis are of about equal abundance. M.femur-rubrum was not taken, but is more abundant in such places in other localities than is M. atlanis, M. differential is is typical in meadow habitats, and M. bivittatus is found sparingly. Arrangement of the Associations as Habitats. The associations of any particular region may be placed in groups with respect to several well-known criteria. One of the prevalent methods is to consider together all associations which are genetically related, which form a developmental series, each stage being succeeded by the next in order, until the ultimate or climax association is reached. Another mode of classification is geographic, placing together all associations of like geographic distribution, all typical of a definite climatic region, an ecological province. Another grouping is based upon growth-form : by this treatment associations dominated by plants of similar growth- form, indicating physiological likeness, would be considered similar, regardless of geographic or successional relationships. The common division of associations of a region into climatic or geographic, and edaphic or local, is based upon local distribution or habitat. The grouping into aquatic and terrestrial is a grouping based upon habitat in a larger sense: the medium of life is the important feature. Finally it is necessary to distin- guish between primary, original or native associations, and secondary, ruderal or cultural associations, wherever human influence has modified primeval conditions. These different groupings are not at all parallel, but form a complex. To orient an association with respect to others of the same region it is necessary to find its place in this complex. In a regional treat- ment it is usually necessary to consider all of these criteria. For different purposes, different bases of classification may be empha- sized, but in no case should one basis be confused with another. For the purpose of the present study, in which the plant associa- tions are considered mainly as habitats for grasshoppers, two bases of classification receive emphasis. The first is with respect GRASSHOPPERS IN RELATION TO PLANT ASSOCIATIONS. to habitat, the second with respect to growth-form of the plants. These two considerations determine the physical and vegetational conditions to which the grasshopper species are subjected. It is to be understood that grasshoppers are animals of the ground stratum and of the herbaceous stratum. In so far as these strata are similar as habitats in different plant associations, these associations may be treated together. Considerations as to whether associations are proximate or ultimate, geographic or local, native or ruderal, whether they belong to the Northeastern Province or the Deciduous Forest Province, are of interest only as they affect relative area of different associations within the region, and consequently the relative frequency and abundance of the grasshopper species of the different associations. In the following synopsis, the moisture factor is used for the division according to habitat, though other physical factors, such as soil, might have been used if it had been desirable to subdivide further. The synopsis arranges the associations of the region according to their similarity as habitats for grasshoppers. Herbaceous or grassland associations. In wet or moist habitats. Sedge and other littoral associations. Meadow associations. In dry habitats. Sparsely vegetated or bare soil. Beach-grass associations. Ruderal dry grassland. Hardwood clearings. Bracken-blueberry growth of open aspen forest. Forest associations. In wet or moist habitats. Cedar bog forest. Ravine forest. In mesophytic habitats. Closed aspen forest. Beech-maple (hardwood) forest. Among the dry grassland associations, variable conditions (resultants of both physical and vegetational agencies) are: texture and compactness of soil, humus content, proportion of bare surface; height, density, and general character of the vegetation. Sparsely vegetated grassland has a high proportion of bare surface; roadways and other small areas of bare soil are to be considered as part of an area of open vegetation. 152 ARTHUR G. VESTAL. It should be pointed out that conditions in the ground stratum of the aspen association are the same in the open treeless parts as among scattered trees. Exposure to sun and wind, as well as the vegetation, are almost identical in the two places. The open aspen forest is then, so far as its ground stratum is concerned, a grassland association. This is in accord with the results of Shelf ord ('i2a: 82), who found that ground stratum conditions in open forests in sand at the lower end of Lake Michigan lag behind in the succession from herbaceous to forest growth. In reality open forests of very scattered trees are usually mixed associations, dominated in places by trees, and in more open places by herbaceous plants. So far as ground conditions are concerned, the association is herbaceous, and the bracken-blue- berry growth is accordingly grouped with dry grassland. Ground conditions are not those of forest until the closed stage of the aspen association is reached. The lichen growth of open aspen areas, affording a locally different environment, may be regarded as a minor division of that association. A primary division in the above synopsis, "shrub associations,'' might have been added. The thicket and bramble associations were small in area, were little studied, and as the conditions in the ground stratum in thickets are much the same as those in the hardwood forests, with about the same grasshopper assemblage, this division is omitted for simplification. The grasshopper assemblages of these associations are pre- sented in summarized form in the third section of this paper, which here follows. III. SUMMARY OF LOCAL DISTRIBUTION OF THE SPECIES. The table which follows summarizes the habitat relations of the grasshopper species, and includes an approximate estimate of the numerical status of each species in each of its habitats. The method of numerical estimation is described on p. 165. It is necessary at this point to define the terms used to denote degrees of frequency, or regularity of occurrence, and abundance, or numbers of individuals per unit of area. Dominant species in a habitat, usually only one or two, have a very high abundance, and a high frequency, in that habitat. Abundant species are alsa Question-marks are inserted to IH cd a CU a 13 rt -*-» o a o 13 — - cu 12 z o cu '0 CJ 3 w 3 5 CO 3 PH cu CJ _^ CJ W +J o 3 g to i-3 in _! •o CU -*-> .2 'cj 1 O > J2 O ^ *-» Q cd 1 u •4-) E H cu ca •3 O EH O •a O to tn cd CU « ^ •a * +— i .HOPPE1 cu o cd a 3 _co u; >. '" w j w (3 « O a B3 mmediatel a 3 CO o fc. o ) , a ca cu to ffi a 0 cu ca H •3 CJ i-, 2 cu cu -C rt cu ^ H H CJ cd "3 _cy S a •3 3 3 ffl ca CO o g •o " — - N cd 2 K^ CO | O +-> ca g cu •— p i, 13 t/} cu MH 0 CU to 3 CU cu M CU IH 3 •3 U CJ U) 0 a D -*-» o cu "3 to ^3 ca o a abitat. £ CU ^ IH IH CU CJ H •3 a _y a . P co cd . O | o t*-" Q ctf CO « •5 fi ££ CO s- CJ o o o cL c w o T3 '5 r- ra *5 -*-» G3 W V 0, •^ B 1 ^ fe "o IH 13 IH 'u 'o 'o o o cj r>. O O O cr d- d o o cu — — — •c -r 'u 'u CJ CJ C C ft. rt. d1 d O CU jj- i! i w a c r*~* X ^ 1! 3 5 cu "O rH O ~ a 2 1 rrt Closed Aspens. — CJ <^. 0 rt W 0 g _co f a ^ ~ n cd — o a , o ca •z r~> - ;'^ 0 "•til II CO T ca . _i . CJ to ".CO '/". ^ to c O ta O O « O '•5 u a o 0 > o „ ca i ^ r >> C u v fc ^ 0) co 1 1%* 2 -22g 0 £0* ^ fV. £ ££ Q Q Si £ c a cu ca -a -2 ~ - "S a cd 2 .a 13 -3 cd o o - t «_ "I " -; i "O -. S O a° & 3 3*3 p 2 g a cc ° d - ~ a ca £ rt cu u ~ •~ ? ? ca p* cr c1 ~ w* to — £ cu ^ 3 1 Ofcfefcfefc 0 i il C3 < c3 ! •« ^ « X ^ 2 n-. *O • • - - • ~ "^ (n fe fe fc < •< ^ > "3 S ^ £ & 'a ca 5 £ a 2 8 O u 2 2 bo-S 0 "0 > Z Z 2 O d1 cu fc -u CU CU 5 g S . f & S » CO i li 5 . — — 3 c 2 3 jj I I I " — CJ a o ' — * CL CO Jj ^^ CJ '5 § ^ filif||i|{ff ~ > g 2 ? * CO CU >? V % 1-5 I a ^ . -— - **H 1 1 2 1 2 ^ o o o o o ^ o ~~ S i^2.^ *" -2 ^"^-2 liS3^S'-S,"3^3^-23 •-S »2 "^ ,i» ^^ ^^ *O V •— * fc^ "^ "^* ^ *^ "" ^~- ^ """ *** ^ ^ *^* ^ ^ 1 Frequently 2 Stations rep grasshoppers. I I '8 a _^ : ' ca o o CO § -^ s •— < r i n 154 ARTHUR G. VESTAL. frequent; they are usually to be found in the habitat in con- siderable numbers at any time. Frequent refers to grasshoppers occurring regularly in the habitat, though not always to be found, and seldom numerous when found. Occasional, occasionally found in the habitat, not frequent or abundant. Infrequent, not often found in the habitat. Accidental, occurring in a habitat rare or unusual for the species. Specimens are usually found singly and the abnormal habitat frequently adjoins one in which the species is more regularly found. The obvious facts shown by the table are : 1. Grasshoppers are more abundant, in species and in indi- viduals, in herbaceous or grassland habitats than in forest, and more abundant in dry than in moist or wet situations. 2. Certain species are much more restricted than others in range of habitats, and in accompanying range of toleration of physical and vegetational factors of the environment. 3. Although a species may be found over several associations, it is more abundant in one, or two, of these, than in others (Certain activities take place in more restricted habitats; chief of these restricted activities is the laying of eggs.) 4. No two plant associations have identical grasshopper assemblages. 5. No two grasshopper species have identical habit-preferences. These facts and others are considered in order in the general discussion. In general, with every change of habitat there is a change in the assemblage of grasshopper species. These replace one an- other in 'the various habitats, often with considerable overlap, but all can be arranged with respect to gradients of environmental conditions. Gradients of several factors which vary together may in general be said to run parallel with the development of vegetation. This development of vegetation is in part the result, in part the cause, of these changes of physical and vegeta- tional conditions. The variation of environmental conditions can be expressed graphically by lines representing gradients of the factors which change with development of vegetation.. Environmental condi- tions in each successional series may be represented by a line, GRASSHOPPERS IX RELATION TO PLANT ASSOCIATIONS. 155 at one end of which the conditions are those of sterile soil or water, environmental control being entirely physical; at the other end conditions are those of the closed, completely developed, climatic association, vegetational control of local environment being nearly complete. It is further possible to arrange the lines representing the different successlonal series within a region, so that relations between them may also be seen graphically. Fig. I (p. 156) is an arrangement of the gradients of physical and vegetational con- ditions in the Douglas Lake region, based on the development of vegetation, and accompanied by a representation of the selec- tion of these conditions by the various grasshopper species. IV. GENERAL DISCUSSION. The Assignment of Terrestrial Animals to Habitats. It is thought by some zoologists that the local distribution of most terrestrial animals is more or less haphazard, that there is no order in the distribution of animals into different habitats, or that, if there is order, the conditions of distribution are too complicated to be determined by any present methods. Shel- ford ('n: 591) points out several reasons for the prevalence of these and similar opinions. On the contrary, environmental relations of animals are now coming to be recognized as quite definite (Shelf ord, 'i2b: 333). The habitat, in the sense of abode for animals, is a particular combination of certain environmental conditions, physical and vegetational, and to some extent, animal, uniform over a certain area. More or less variability within this area is the rule, and we may consider the area of the habitat as larger or smaller, according to the degree of uniformity of conditions. In these areas of varying degree, plant and animal communities of various degree find their existence. Ecological classification, aside from its dealings with plant and animal individuals, has to do with the recognition and classification of these degrees of likeness and difference of environments and of plant and animal communities. (See Shelford, 'i2b: 355.) The habitat con- sidered most convenient in the treatment of local distribution of GRASSHOPPERS IN RELATION TO PLANT ASSOCIATIONS. 157 ;etational conditions within the Douglas Lak<- il stages developing from wet ground or open erophytic and hydrophytic series converge in top. Moisture content is successively higher nted in the region, as forest invasion would egion present conditions essentially those of per place in the gradient of grassland environ- hicket, closed grassland). In certain of the aspen forest. Such indirect transitions are 0 "cd 03 a y rH oi •d y C y CO y a y y 60 £ -J which are dotted in the extremes of the range. ies occur in equivalent stages of the different ;wo gradients; those restricted to aspen forest oo •--. ^ y 1^ o ^ * • — y cu 4-1 CU c 4-) CO y cu 60 y c ;> CO y y a y -o ^j > tH ~ £ a y ~ cu O .0 a o O CO c a .C M y c *" _c *J CO •c CO y '3 y y 1 cu .C — • o y .0 •~ •S 0) CO c 03 "o c 4-1 IH y 'co >. J3 a CU ~ cd '-t-. 0 a £ o Id E 1_^ CO C n 1 "a CU ^ cu ~ CO CO 03 60 r-1 4-1 rn .0 y cd 4-> y •o y y l*H O CO -t_> _CJ '•S 2 60 •G C cd CO .- 'y •n c 3 O 60 *-. •a 1 c o cu oi 7 ~ y o y C a successively greater vegetation would noi ain ruderal associatil ;ions are accordingly 03 z. = cu cd C _o '5o y CU t, as between beach- nt of coniferous fore rious gradients is sho' C y y CO _co 4J "3 CO CO o3 a y 03 CO 4-1 CO ,O y a CO OJ y a CO y a a 0 CO tion developi e vegetation IS 3 S CO "O CO cd 60 owever, cert hese associal _C — y y y •3 O C .22 r developme cd y CO y y "1 3 S '3 •o OJ •o O O •5 cd 03 — "o HH •*_> d Q 'y cd >H y y Id 60 y -M 'o S 4-) O CO y 60 S •d cd y CO y & y CO cu o3 CO -tJ c y y a CO y •a JZ O i i o y y CO J2 c IH O •3 cd a a c y c o y _c CO cu 60 03 CO E a "y Lationa 4-> CO CO cd 60 jd "co CO 2 60 "3 4-i o -M C y y 60 y H o "co CO cd 3_ y *S co mm radient jd "rt y •a y cu y CO •a _g CO y CO CO _y 60 "y y •4-t '£ S CO cd IH 'c 4_» "c o "y *o cu CO "y CO y 60 cd cn y c _o CO y •o "•L •3 cd "o y rence i 'y y a C y a CO cd y . I . Showing cd Conditions v max associatio o -*-» t development "u cd y c <4-l O •a cd I conditions. 4—* 'co C cd >H 4-> y J3 •4-) CO 4-> c y 'o •d •d y 'c :/, CO y 14 — tl y — proximate rang 3 y y O '3 J3 of vegetation. 4-> _60 r* 4-> u fe 0 u y "y £ K y > 1 cd 4_> z _y CO y i_, "o a < 13 CO •d O 4-> 1 "cd y 4-t o — y a a y S 60 a y y 4-> B | y cd ARTHUR G. VESTAL. grasshoppers within a region is the area marked off by the plant association, in which there is a general and usually recognizable uniformity of vegetation and of physical conditions. A species is to be assigned to a habitat if its occurrence in that habitat is a matter of regularity. In general, the animal species is seldom confined to a single habitat within a region. The determination of the more typical habitats selected by a par- ticular species involves estimation of relative numbers, and presents certain difficulties, which are discussed belowr. In a regional study of a group of animals taxonomically related, or of all the animals of the region, there are certain other difficulties which obscure the definiteness of habitat relations. Variation in Range of Habitats in Different Species. — The range of toleration of environmental conditions varies wddely among different species of a taxonomic group, and accordingly we find certain animals restricted to one or several similar habitats, while others range over many and different habitats with apparently little discrimination. Melanoplus atJanis, for example, occurs regularly in meadows, in dry open grassland, and in open forest. It is a species which can endure widely varying conditions. Upon examination, it is found that it is very much more numerous in certain habitats than in others, and that there are certain habitats in which it occurs only accidentally or not at all. In the Douglas Lake region, it was never seen in cassandra bogs, in bog or ravine forests, in closed aspen forest, and only infrequently or accidentally in closed hardwood forest. There is, then, a selection of habitats, and the local distribution of even . the most generally distributed forms is far from indiscriminate. Habitat- Distribution as Affected by Motility of Animals. — The habitat relations of plants are more evident than those of animals because the plant individuals are non-motile, while animal individuals can move about, and it is not to be supposed that they will necessarily stay within a single area of uniform conditions. They continually stray beyond the limits of conditions necessary for all or part of their activities. Among the grasshoppers, certain are much more motile than others, particularly the long- winged and active (Edipodince. These grasshoppers at times fly high and to considerable distances, and may be seen to traverse GRASSHOPPERS IN RELATION TO PLANT ASSOCIATIONS 159 areas of dry grassland, of marsh, and of forest. However, while at rest, very few will be found on trees, for example, or in deep shade, or in dense herbaceous vegetation. They will usually be seen in open dry vegetation or on bare soil. If found in other conditions, these will usually be not far distant from such open situations. Stray individuals are not common and are usually found singly. It is not surprising that stray individuals should be more frequent among species which are very abundant in the region, as Camnula pellucida, and among species of the more extensive habitats. The motility of animals is perhaps com- monly overestimated as a factor in width of local distribution. The daily itinerary of an animal is likely to be more circumscribed than is usually thought. In the case of the grasshoppers again, for most species the ordinary mode of progression is walking or crawling, rather than jumping or flying. Usually grasshoppers are noticed only when disturbed, or "flushed." Their behavior, jumping or flying when disturbed, is a special reaction to an approaching object. The ordinary activities are much less fre- quently observed. Birds, our most motile animals, have very definite habitat relations (see Gates, 'n). Though the actual number of observed occurrences of stray individuals in unusual situations may be large, these occurrences are very infrequent and exceptional when compared with occurrences in the normal habitat. Differences of Activity in Different Habitats. — Various activities of the animal may take place in different strata of one habitat, or in different habitats, separated horizontally. The stratum or habitat of greatest importance to the animal is the one in which the most narrowly limited activity takes place, and this activity is usually concerned with breeding (Shelford, '07, 'n).1 In the case of the Douglas Lake grasshoppers, eggs are laid just below the surface of the soil, or at least in the ground stratum, and proper conditions for oviposition are among the most important considerations which determine the presence of the particular species in the region, and the important consideration in deter- mining which habitat within the region is most essential to the 1 On p. 595 (Shelford, 'n) are given references to other authors, in which in- stances of breeding activities as being most narrowly limited are given for birds and for fishes. I6O ARTHUR G. VESTAL. species. The habitats in which few or no grasshoppers are found are those in which the soil is not suitable for egg-laying. In many animals the breeding period occupies a rather small part of the season of activity, and the adult animal may spend but a small proporlion of its active life in the breeding stratum or habitat. The habitat in which the species occurs most regularly and in the greatest numbers is the habitat in which the species is of greatest influence. It is the most important habitat of the species so far as plants and other animals are concerned. In associational studies, in which emphasis is laid upon relations between organisms, this habitat is most important. The animal may feed in several habitats, but principally in the habitat in which it is most frequent and abundant. The feeding activities, though of secondary importance (usually) in determining the presence of an animal in a particular habitat and a pardcular region, are of primary importance in relation to the communities of plants and animals of which the species becomes a member. In the study of all the habitats within a given region, emphasis would be placed on the habitat in which the species is most frequent and most abundant. The success of the species in the various habitats, as indicated by relative numbers, is a measure of the degree of correspondence between the environmental conditions actually furnished within the area of the habitat, and those required by the animal. The habitat in which the species is found most regularly and in the greatest numbers is the habitat in that region which most closely approximates the optimum environment for the species. The Different Activities in Climatic and in Local Habitats. — It so happens, in probably the majority of the grasshopper species of the Douglas Lake region, that the feeding habitats and the breeding habitats coincide more or less perfectly. In insects of incomplete metamorphosis, as Orthoptera, the possibility for the immature animals to correspond in mode of life rather closely to the adults is much greater than in insects in which changes from larva to adult are more radical. Grasshopper nymphs feed and hop about as do the adults. They cannot fly, but as flying more commonly does not take the animal into another habitat, this difference is of minor importance. Where the GRASSHOPPERS IX RELATION TO PLANT ASSOCIATIONS. l6l habitat in which the species is more frequent and abundant is of considerable extent within the region, as in climatic associations, soil conditions, or conditions for egg-laying, would be either uniform throughout the area of the habitat, or else suitable small spots would be scattered about within the area. There would then be no necessity for the females to migrate to a different habitat for oviposition. Thus most of the grasshoppers found in the aspen association would be able to lay eggs within its area, and the nymphs would find food and other necessary condi- tions in the same place as do the adults. In the beech-maple or cedar-bog forests, which individuals of Melanoplus islandicus probably rarely leave, suitable egg-laying sites, in this case wood as found in stumps or logs, are scattered about over the forest floor. It thus appears probable, in the extensive habitats provided in terrestrial climatic plant associations, that the corre- spondence between the conditions required for breeding and the conditions required for feeding and other ordinary activities, may be quite general. Conditions for egg-laying among insects, for example, may still be more restricted than conditions required for other activities, but the local variability of environmental conditions within the area of the plant association is sufficiently wide, usually, to include the more restricted conditions necessary for egg-laying, or whatever the most narrowly limited activity happens to be. If what is true in the case of grasshoppers is true in a large number of terrestrial animals, as seems likely, this means that the limits of the climatic plant association need not be passed, ordinarily, by a large number of the animal species, since all the necessary conditions are supplied within its area. The animal community of the area may be thus, in large measure, self-contained, and coextensive with the plant community. In local habitats, on the other hand, which are usually re- stricted in extent and consequently likely to present less varia- bility of environmental conditions within the area, it is less likely that all the conditions necessary for the animal species will be supplied within the area. The number of animal species which can find all the necessary conditions for existence within the area will be comparatively small; thus many of the species will be obliged to perform certain of their activities in other habitats. 1 62 ARTHUR G. VESTAL. Furthermore, environmental conditions in local habitats are likely to be the extreme, rather than the mean, conditions within the region. Thus sedge mar^h habitats in the Douglas Lake region have submerged or very wet soil; in the latter case, the growth may be so dense as to leave no exposed soil. These are extreme conditions for terrestrial habitats. The only grass- hopper of the region which was seen in such situations is the ectophytic Stenobotlints curtipennis. It is probable that this species deposits its eggs in higher and drier soil. Unlike most of the grasshoppers, its ordinary activities can be carried on in a very humid environment, and the presence of exposed soil is not necessary. The conditions necessary for breeding, however, do not correspond with the extreme condition of wet soil, and must be obtained outside the area of the extreme habitat. The local habitat is less likely to be self-contained than the extensive or climatic habitat. Difficulties Arising from Habitat Complications. — The habitat relations of animals sometimes are much less evident at the tension zone between two contiguous habitats, and there is the further complication that habitats are sometimes distinguished with difficulty. Boundaries will be invaded by animals of both habitats. The tension zone is much wider, and the confusion greater, between two rather similar habitats. As Shull points out ('n : 221), the determination of habitat relations is difficult in regions where the habitats are small in area and much inter- mingled. The tension zone presents one great advantage in that it allows determination of which species of one habitat range farthest and most frequently into the adjoining habitat. A graded series of species can be determined, which expresses the resultant of different factors entering into habitat selection of the different species. Unstable habitats are frequently indicated by the presence of mixed plant associations. These contain representatives from two or more plant communities. Thus near the beach of Douglas Lake, the bracken-blueberry growth is mixed with bearberry and juniper heaths, a few small clumps of beach-grasses, in certain spots, and in others by local growths of ruderal bluegrass. In such mixed areas the animal assemblage is not of the constant character seen in more uniform growths. GRASSHOPPERS IN RELATION TO PLANT ASSOCIATIONS. 163 Mixed conditions within the area of a habitat also tend to confuse. This is particularly well shown in open forest. Mela- no plus fasciatus and Chloealtis conspersa, of the species studied, are found in open forest with a number of other grasshoppers, but appear to be the only two for which forest conditions are really necessary. All the others are grassland species, and are present because of the grassland environment of the ground stratum. Grassland species may be found in very small sunlit grassy patches within a closed forest. These sunlit patches are not to be considered as part of the closed forest habitat. Prob- ably very few insects of the ground stratum are really typical of open forest. Roadsides and other local modifications are a source of error in assigning animals to a habitat. A collector traveling along the sandy roads through the bracken-blueberry growth of the aspen forest would see large numbers of bare-ground CEdipodince, almost all on the roads. He might never realize that these grass- hoppers are not particularly common in the undisturbed growth, and that they are animals of the roads and not of the bracken- blueberry growth. Furthermore, the habitat is not entirely uniform within its area. There are many extremely local environmental differences within the habitat, which influence the animal species as well as the plant species. Thus in scattered parts of the bracken- blueberry growth, occasional patches of lichen-covered surface are found. With these patches is associated the grasshopper Scirtetica marmorata. It does not occur over the entire bracken- blueberry growth, as does Melanoplus angustipennis , for instance. Within the habitat there are local differences in the degree of development of the vegetation. This is shown in grassland in the degree of openness of vegetation. Where considerable bare surface is exposed between the plants, bare-soil animals are found. In closed grassland, no animals of this habit occur. Differences of development are shown in the aspen forest. Among scattered trees the bracken-blueberry growth presents an en- vironment essentially that of grassland. In more compact tree growths ground conditions begin to approach those of the forest floor. Only in such situations is the short-winged grass- hopper Melanoplus fasciatus to be found. 164 ARTHUR G. VESTAL. Estimation of Relative Numbers in Different Habitats. The Distinction Between Frequency and Abtindance.—Many collectors do not distinguish between frequency of occurrence and abundance of individuals. To them a species is common if scattered individuals are seen frequently, or if numerous indi- viduals are seen infrequently. Frequency refers to regularity of occurrence in one or more habitats ; abundance is concerned with numbers of individuals per unit of area. The distinction is indispensable if we are to estimate numbers of individuals in different habitats. Thus many of the conspicuous (Edipodince, a number of which can be seen in flight at one time, but over a considerable extent, as far as one can see, are frequent species, but are not numerous as compared with certain inconspicuous Melanopli, a larger number of which may be found in any area within the habitat as large as a few meters square. These last are abundant and frequent. Difficulties in the Way of Numerical Estimation. — The writer is not acquainted with any very practicable method of estimating absolute numbers of insects in different habitats of a region. Absolute numbers of plants may be estimated under favorable circumstances, as may also the numbers of the larger or more sedentary animals. Following are some of the difficulties en- countered in estimating absolute numbers of terrestrial animals: In any one species, there are differences in numbers of individuals, in degree and kind of activity, in readiness with which it is per- ceived, in ease with which it is captured, or recognized if it is not taken, in time and in space. These conditions vary with the year, with the season, with weather conditions, and with the time of day; they vary in different habitats, in different strata, with different kinds of soil and against different backgrounds, and in different kinds of vegetation. There are many difficulties in the way of estimating even relative numbers of different species. Less abundant species are likely to be confused with more abun- dant kinds. There are differences in appearance, in conspicuous- ness, in degree of activity, in ordinary behavior and behavior upon being disturbed, with different species, which make some of these much more readily perceived, or recognized, or captured, than others. Differences in time of appearance and activity, and GRASSHOPPERS IN RELATION TO PLANT ASSOCIATIONS. 165 in place, in surroundings, may Make certain species appear more or less abundant or frequent than they really are. The method used by Shull ('n : 218) in estimating numbers of grasshoppers per unit of area is valuable, but does not distinguish the different species. His suggestion "to collect immense num- bers, and depend upon majorities to decide the usual habitat," besides being impracticable, is not satisfactory, for the numbers collected are not a true index of the numbers which actually occur by reason of the difficulties mentioned above. The Method Used in Determining Frequency and Abundance — The method used by the writer was to visit the various habitats, taking one or several grasshopper specimens of each species for verification, and to estimate abundance not from the numbers collected, but from the numbers observed per unit of area in each station, having in mind the considerations which tend to over- or underestimation of actual numbers.1 If in the bracken-blue- berry growth, a considerable number of Melanoplus angustipennis could be seen nearby, within a rod or two, and if this condition was practically uniform over the area of the association, as nearly as could be seen, an infinity sign was put down in the field notes opposite its name, indicating numbers in which it occurred, followed by the number collected, thus (oo, 2). If in twenty minutes walk through the aspens only four or five specimens of Scirtetica marmorata were seen, and two collected, the notes would appear thus (s, 2), the letter s representing several. If in a half-mile of sandy roadway, a considerable number of specimens of Spharagemon were seen, but never close enough together to be abundant, the numbers wrould be indicated by the plus sign (-f , i). If a species was taken in a particular kind of habitat in nearly all stations of this habitat visited, or if it was regularly 1 Facilities for determination were available at the biological station, and the species were identified as they were taken. With most of the species the writer had been acquainted, and the species soon became so familiar as to be recognizable at sight. In life there are many rather conspicuous peculiarities of color and of behavior which aid in recognition. Whenever there was doubt as to identity, as in the case of Melanopli particularly, specimens were caught and examined critically in the field, to avoid overlooking species of similar appearance. In all cases one or several specimens of each species were kept for verification. There is no reason why many familiar insects should not be recognized at sight as easily as birds are identified by ornithologists. 1 66 ARTHUR G. VESTAL. to be found in extensive areas representing this habitat, its frequency for that habitat was considered to be high. From the records which thus accumulated it was possible to arrive at a relative estimate of frequency and abundance for each species in its various habitats. Though the method is not free from error, it is the best which was available, and it is believed to be not far from a representation of actual conditions. Terms expressing degrees of frequency and abundance are dominant, abundant, frequent, occasional, infrequent, and accidental. These terms are used in the table of distribution on p. 153, and are there defined. The Relation of the Animal to Plant and Animal Communities. The Plant Association as an Index of the Habitat. — The area of the habitat as considered in this study has already been defined as the area conveniently marked by the extent of the plant asso- ciation. In the plant association or plant community physical conditions and vegetation are generally uniform. In the early stages of development of vegetation, local physical conditions dominate. In later stages the vegetation assumes the type determined by climatic conditions, and exerts nearly com- plete control over local physical factors. Thus the grass- hoppers in early stages of vegetation, as the CEdipodincB, most of which live in very open grassland, are more intimately associated with physical conditions, which to them are more important, while those of advanced stages of vegetation de- pend more upon vegetational factors, and less upon the char- acter of the soil. Melanoplus islandicus, associated with climax deciduous forest, is a species of the second type. The grass- hoppers as a group are most abundant in early stages of vegeta- tion in forest climates, while in grassland climates they occur abundantly in all stages. The plant association may thus be taken as the index of environmental conditions; it expresses the resultant of physical and vegetational conditions to which animals are subjected. Within the area of the plant association local variabilities in physical conditions are usually accompanied by local variabilities in the vegetation; the latter may exist inde- pendently. Grasshopper species are affected by these local differences. GRASSHOPPERS IN RELATION TO PLANT ASSOCIATIONS. 167 Vertical Distribution. — The distribution of terrestrial animals in space is both vertical and horizontal (Shelford, '126). Most of the grasshoppers of the Douglas Lake region belong to the ground and field (herbaceous) strata, as do the Acridiidce in general. With the exception of the forest species which oviposit in wood, the eggs are laid in the soil. Most of the species require bare soil, and many of their activities take place directly upon the surface. Stenobothrus is more commonly seen upon the plants, and certain species of Melanoplus live upon the plants part of the time. Podisma glacialis variegata is a shrub-inhabiting species. Grass- hoppers are in general typical animals of the ground stratum. Other families of Orthoptera are more typical of other strata. Horizontal Distribution. — Within a region an animal species will select habitats or associations in which conditions of the optimum environment are most closely approximated. The table on p. 153 indicates that no two grasshopper species of the Douglas Lake region select the same set of environmental condi- tions. Certain of the species are similar in distribution, but none are identical. They replace one another in different habitats, with some overlap, and can be arranged in series according to gradients of environmental factors (Fig. I, p. 156). Although a single species may be found in more than one association, it is not equally abundant nor equally regular of occurrence in these associations. Certain of the grasshopper species are typical of only one association, as Melanoplus fasciatus in closed aspen forest. Melanoplus atlanis, the most generally dis- tributed species, though abundant in five habitats, is most abun- dant in ruderal grassland, and very typical in such situations. Other grasshoppers are more characteristic of the other four associations in which Melanoplus atlanis is abundant. The various habitats in which a particular species may be found happen to be more or less similar in physical and vegeta- tional conditions. Data in the table on p. 153 indicate that those associations in which a grasshopper is found in common may agree only in containing certain or all of the conditions necessary for that species, and that there need be no successional or geographic relationships between the associations, nor is it a matter of concern whether the associations be native or ruderal, extensive or local. 168 ARTHUR G. VESTAL. The occurrence of a grasshopper species, or of several species, in two habitats or associations does not mean that these furnish similar environmental complexes for animals. It means simply that certain environmental conditions necessary for these par- ticular species are included among the conditions provided by these habitats. Thus in the Douglas Lake region, Melanoplus islandicus is found in beech-maple forest and in the cedar-bog forest. These two associations are radically different in many respects, as habitats for animals. The entire range of conditions presented within the area of an association, particularly if it be extensive, is likely to be considerably more inclusive than the range of conditions required by many animal species. Taxo- nomic groups of animals which are affected more particularly by conditions differing in the two habitats, will be represented differently in them. It follows that very little reliance can be placed on comparisons of habitats on the basis of the study of a taxonomic group, such as grasshoppers, except in respect of the particular conditions critical to the species of this group. Com- parisons of the entire animal communities of the two habitats would not be subject to this limitation, since nearly all of the environmental conditions within the two habitats would come into consideration. Within any one association the animal species may be dis- tributed generally throughout its area, as Melanoplus angusti- pennis in the bracken-blueberry growth; in certain instances it may be restricted to a part of the area characterized by a slight environmental difference; or it may occur in scattered parts of the association, characterized by scattered local differences, as Scirtetica marmorata in the lichen-covered patches within the bracken-blueberry growth. The Animal Environment. — In addition to the physical and vegetational influences upon the animal species, that of its animal environment must also be recognized. Direct effects of the animal-environment upon the animal species are probably greater than the indirect effects produced by modification of physical and vegetational environments. Among these latter more general effects of the animal community are the accumula- tion of organic remains, and particularly the effects of phy- GRASSHOPPERS IN RELATION TO PLANT ASSOCIATIONS. 169 tophagous animals upon vegetation. It is probable that animal environmental influences are greater than is commonly supposed, but that ordinarily they play a subordinate part as compared with physical and vegetational influences (cf. Shelf ord, 'i2a: 94). The grasshopper species is not only affected by the animal environment, but is itself a part of it. So far as the relations of the grasshoppers to the association and with other species is concerned, about all that is known, in general, is that they are among the most important of the plant-eating animals of grass- land associations, in which some of them are dominant species; and that they form the principal food supply for a comparatively large number of predaceous and parasitic enemies. It is not known to what extent different grasshopper species compete with one another. Differences in habitat and in time of activity may indicate removal of competition among certain species. Successional Relations. The successional relations of the Douglas Lake associations have been discussed by Gates ('13: 48), and a diagram illustrat- ing the successions is included. The work on the grasshoppers has not covered so wide an area, consequently many of Gates' associations are not well represented. Fig. I, on p. 156, will serve to illustrate successional relations of the associations in which grasshoppers were taken, and the changes in grasshopper species may also be seen, as one plant association is replaced by another. Initial stages in dry soil are shown at the lower left- hand part of the diagram; initial stages in wet habitats at the lower right. The two series converge at the top of the curve, in the climax beech-maple forest. The ordinary course of succes- sion from marsh associations is toward bog forest; from beach grass into pine forest. These successions are not shown in the diagram. The aspen association is the result of secondary succes- sion from pine forest, which is no longer well represented in the immediate region. Grassland associations between beach-grass and wet meadows are represented only by ruderal growths ; which have originated mostly by secondary succession. Closed grass- land is not represented, for invasion by forest occurs before it can develop. I/O ARTHUR G. VESTAL. Only the series leading from bare sand to climax forest through the aspen stages of burnt-over pine lands will be discussed in relation to grasshopper succession. The following table, taken from the distribution table on p. 153, and with the same notation, shows the grasshopper species of the various stages. Accidental and atypical occurrences are not recorded. It should be re- membered that the bare soil habitat is never extensive in the region, and that grasshoppers of bare soil depend also upon nearby vegetation. TABLE II. Bare Sand. Beach- Grass. Bracken- Blueberry. Closed Aspens. Beech - Maple. Dissosteira Carolina Freq.1 Freq Occas. Spharagemon boHi Freq.1 Occas. Occas. Circote'Mx verrucnlatus Freq. Occas. Occas. Arpliia pseudonietana Freq.1 Infreq. Infreq. Hippiscus litberculatits Freq. Infreq. Infreq. Camnula pellucida . . Freq.1 Freq.1 Freq.1 Melanoplits angnstipennis . . Melanoplus liiridus Occas. Freq. Dom. Occas. Scirletica marmorata. Infreq.2 Melanoplus fasciatus Infreq. Freq.1 Chloealtis conspersa ? Infreq. Podisma g. variegata > Melanoplus islandicus Freq. Mi-lnnapliis ntliniis . . . Freq.1 Abund. Freq. -> Infreq. From examination of the table we see that there is a successive change of species with development of vegetation, and that even when species are not replaced with changes in associations, their numbers are successively increased or decreased. The succession from bare sand to closed hardwood forest includes two successional series, the development from very open grassland growth to closed grassland, and the development of closed forest from open forest growth. The transition stage is the open aspen forest, in which trees are small and scattered. Ground conditions, dominated by the bracken-blueberry cover, are those of grassland which has not yet reached the closed stage. The change in ground conditions from those of grassland to those of closed forest is radical, as shown by the great difference 1 Frequently found, usually not in numbers, but in a few places approaching abundance. 2 More typical of scattered patches of lichen-covered surface; distribution in bracken-blueberry not continuous. GRASSHOPPERS IX RELATION TO PLANT ASSOCIATIONS. iyi between the grasshoppers of the bracken-blueberry and closed aspen stages. Climax beech-maple conditions have probably not yet been fully developed, in the immediate region, from the closed aspens of the pine lands. The beech-maple forest of the morainic areas of the region represent the ultimate condition of present closed aspen areas. The grasshoppers of the table show certain likenesses and differences in habitat-selection which may be correlated with their behavior characters. The first five species, all typical of very open situations, are active, motile forms, of strong and sustained flight, and are usually seen resting upon bare soil. They lay eggs usually in soil of loose texture. They are fre- quently seen on roads, and patches of bare soil, in the Douglas Lake region, and are abundant near the beach. They are fre- quently seen in the interstices between plants in open grassland, and become successively less numerous with the closing of the vegetation. Camnula pellucida is less like the typical bare-ground (Edipo- dincB in behavior. It is more variable in distribution, and though practically restricted to grassland, is more numerous in ruderal growths. Melanoplus angustipennis is an interstitial species, and in- creases in abundance with the closing of the vegetation until the sand is relatively stable, with the admixture of a little humus, though the soil is still loose in texture. Melanoplus luridus and Scirtetica marmorata have not been found in sufficient numbers to determine their status with satisfaction. The former appears to be a species of nearly closed grassland; the latter is more or less closely associated, in this region at least, with lichen surfaces that had not developed in earlier stages of herbaceous growth. Melanoplus fasciatus and Chloealtis conspersa, in this and other regions, are associated with dry open forests and forest borders. They are more frequently short-winged, and exhibit a departure from grassland behavior. The latter is known to deposit eggs in wood. Melanoplus islandicus is a shade-dwelling species of deep woods. It is flightless, and probably lays eggs in wood. Podisma 172 ARTHUR G. VESTAL. glacialis variegata has not been taken in the beech-maple forest, but probably occurs there. It differs from the other grass- hoppers of the region in being of the shrub stratum, rather than of the ground or herbaceous strata. The species in the table are arranged approximately in the order of succession of the associations in which they occur. Melanoplus atlanis, being more generally distributed, can hardly be assigned to a particular place in the series, and is placed at the bottom accordingly. It is most abundant in ruderal grassland, and is not at all common in forest. A parallel development of vegetation in sand is seen in a region in central Illinois. The herbaceous vegetation is sand prairie, is very open in the initial stages, and is replaced by xerophytic oak forest before a closed grassland is reached. The habitat- relations of the various grasshopper species are discussed by Hart ('07). A number of the species of the Douglas Lake region are here represented in habitats occupying equivalent stages in the successional series. The initial stages of herbaceous growth, characterized by a large proportion of bare surface, are accompanied by active, very motile grasshoppers which rest normally on the surface and which lay eggs in soil of loose texture. With the closing of the herbaceous growth grasshoppers of this habit gradually decrease in numbers, giving way to species which rest part of the time upon the vegetation, which are less motile, and which lay eggs in the less sterile types of soil of such situations. In sand regions of the forest climate, usually, forest invasion occurs before the herbaceous growth becomes closed. The ground conditions remain those of grassland until the forest approaches the closed stage, when the grasshoppers of open grassland are abruptly replaced by forest grasshoppers, which are less motile, usually flightless. They are fewer in species and individuals than grass- land members of the group, in part because of the fact that in advanced forest stages, in which the ground is almost entirely covered with dead leaves, the soil is generally inaccessible for egg-laying, and oviposition takes place typically in wood. GRASSHOPPERS IN RELATION TO PLANT ASSOCIATIONS. 1 73 Geographic Relations. The vegetation of the region, as has already been mentioned, is composed of two geographic elements, that of the Northeastern Conifer Province and that of the Eastern Deciduous Province. In addition there are certain associations, chiefly made up of plants which cannot be assigned to these vegetation regions. Comparison was made of the geographic distribution of grass- hoppers \vith that of the vegetation, and with their own local distribution. Most of the species are generally northern in distribution. Of these Podisma glacialis variegata, Melanoplus island-lens, and Chloealtis conspersa may be assigned to the Northeastern Province. Circotettix verruculatus, Melanoplus fasciatus, M. luridus, M. minor, Camnula pellucida, and Stenobothrus curtipennis range in both northeastern and northwestern coniferous regions, extending south to varying distances in both Appalachian and Rocky Mountains. Arphia pseudonietana does not range so far to the east as these preceding species. Arphia, Melanoplus luridus, M. minor, Camnula, and Stenobothrus occur also throughout the northern part of the Prairie Province. Spharagemon bolli and Scirtetica marmorata may be assigned to the Eastern Deciduous Province. Hippiscus tuberculatus is found in the northern parts of both prairie and deciduous forest. Melanoplus angustipennis is a prairie species, being found abun- dantly in sandy parts of the Prairie Province, ranging also west into the sage-brush country, in Utah. Dissosteira Carolina, Melanoplus atlanis, M. femur-rubrum, M. differ entialis, and M. bivittatus are of very wide geographic distribution, ranging over most of the United States and much of Canada. Certain species of the first group are rather sharply restricted to the coniferous forest regions, while others range well into the prairie and deciduous forest regions. Chloealtis and M. fasciatus range also into deciduous forest. M. luridus, M. minor, Camnula (in the northern prairie states), and Arphia range also into the prairie region. An excellent instance of the sometimes sharp boundary between two provinces is shown along the foothills of the eastern Rocky Mountains in Colorado. Here Chloealtis, IJ4 ARTHUR G. VESTAL. Circotettix, M. fasciatus, and Camnula are not found outside of the mountains, while Arphia, M. luridus, and M. minor are found on the plains. Stenobothrus bears no particular relation to climatic vegetation regions, as its presence is determined by local conditions of moisture. The occurrence of Melanoplus angustipennis at Douglas Lake, so far from the prairie region, seems at first very unusual. So far as the writer is aware, the species is not recorded from Michi- gan, though it must occur all along the Lake Michigan shore, and is known from Ontario. Its most necessary environmental condition is sandy soil, and this is well developed along most of the shore-line of the Great Lakes. The five species of most wide distribution geographically are those which have least definite habitat-preferences within a particular area. The extensive range, and the fact that these particular species are among those most important economically, as destructive to crops, is to be explained in terms of tolerance of widely varying conditions. The species which are restricted to forest habitats in the Douglas Lake district are restricted geographically to forest provinces. Certain of the species found in grassland or in open forest in the region studied range outside the forest provinces, being thus clearly independent of forest growth, while others appear to be restricted to forested regions, though found in open habitats. It is probable that Camnula pellucida and Circotettix verruculatus are restricted to the mountain areas of Colorado by radical changes in physical conditions from mountains to plains, rather than by changes in vegetation. In general, it may be said that grasshopper species associated with the climatic plant association have the same geographic range as this association. This range, when shared by many species, determines the ecological province. Where two similar climatic vegetation regions adjoin, as in the case of northeastern and northwestern conifer regions, the same grasshopper species may range over both, in similar associations. Certain species may also range locally into other provinces, in habitats locally approximating those of their own climatic association. Species associated with local habitats may be restricted to the province, GRASSHOPPERS IX RELATION TO PLANT ASSOCIATION 1 75 but more usually range over much wider areas (Shelford, 'n: 606). Species of very general local distribution within a re- stricted area are likely to be very generally and very widely distributed geographically, and are the species which most frequently invade ruderal and cultural growths, and which tend to replace other species with the spread of civilization. Species of closed associations show more evident local and geographic relation to the vegetation than species of open associations, in which local physical conditions dominate (cf. Shelford, 'i2a: 89, 90). Seasonal Relations. Distribution of animals in space is more or less affected by distribution of animals in time. In dealing with the latter it seems that seasonal relations and life-histories of animals are comparable to habitat relations and habitat-preferences, wrhen dealing with distribution in space. The growing season is not uniform in physical and vegetationa^ features of the environment. Seasonal changes are conveniently marked by changes in aspect of the vegetation. The successive changes in the environment for grasshoppers in the course of a season are somewhat different in forest associations from those in grassland. A field of study of seven weeks is not long enough to determine seasonal relations of grasshoppers, nor to determine what species of grasshoppers occur in that region. Clwrtopliaga mrldifasciata (De Geer), an early spring grasshopper almost certainly occurs at Douglas Lake. Melanoplus scudderi (Uhler) and Melanoplus punctulatus (Uhler) may reasonably be expected to occur within the region. Both are short-winged forest-inhabiting species of late summer and autumn. The time of adult activity of such species as were found, how- ever, is known; the species may be arranged according to life- history into the following groups: 1. Species which hibernate as nymphs, appearing as adults in spring, remaining active only during early summer, represented by Hippiscus luberculatus. 2. Species which hibernate as eggs, maturing very early in summer or late in spring, becoming scarce or disappearing by about September i. Melanoplus minor, M. bivittatus (sometimes persisting in numbers till frost), M. fasciatus (this species may more properly belong in the fourth group). 1 76 ARTHUR G. VESTAL. 3. Species which mature in spring or early summer, but which remain abundant until frost. Chlcealtis conspersa, Melanoplus angustipenni's, M. atlanis, M. femur-rubrum. The last two are known to be two-brooded, and M. angusti- pennis probably is also, in parts of its range. 4. Species which mature in early July, remaining active until frosts. This is the common grasshopper life-history. Stcnobothrus curtipennis, Camnula pellucida, Dissosteira Carolina, Spharagemon bolli, Scirtetica marmorata, Circotettix verruculatus, Melanoplus islandicus, M, luridus. 5. Late-maturing species, appearing from late in July to the middle of August, and remaining until frosts. Arphia pseudonietana, Podisma glacialis variegata, M. differentialis. The different timing of the period of activity of the species may be correlated with one or both of two kinds of seasonal difference in environmental influence: (i) Changes in physical and vegeta- tional conditions necessary for the various activities. Conditions of temperature, moisture, vegetation, etc., change so materially with the season as to present very different environments for animals at different times of the year. Grasshoppers of different habitat-preferences may perhaps occupy the same area at different seasons. Life-histories are affected by geographic variation of climatic conditions. The entire complex of behavior characters of the animal is intimately associated with its life- history (Shelford, '12^: 334). (2) Seasonal differences in an- tagonistic influence of other animals, particularly those of similar requirements. These influences among animals which eat the roots of the corn plant and the strawberry plant are discussed by Forbes ('09: 296). Different species of certain insect genera, .among them Arphia and Hippiscus, which are of similar habits and which occur in the same association, sometimes are active during successive parts of the season. One species of a genus may be completely replaced by another in a very short time, the second species continuing abundantly during the remainder of the season (Vestal, 'i3&). This replacement of one insect by another of similar habits is not uncommon among species of one genus, though not confined within genera. The differences in time of appearance among all the grasshopper species within a region, with the result that within some of the associations fewer species occur together at one time, may perhaps be ac- counted for partly in terms of antagonistic animal influence. The fact that grasshopper species of unique habits, which occupy GRASSHOPPERS IN RELATION TO PLANT ASSOCIATIONS. 177 certain habitats by themselves, so far as grasshoppers are con- cerned, have the most common type of life-history, hibernating in the egg stage and active from early July till frosts, suggests that antagonistic influence is not a factor in determining their life-histories. Such species are Stenobothrus curtipennis in marshy growths, and Melanoplus islandicus in deep forests. It appears that the species with least definite habitat-prefer- ences (Melanoplus atlanis, M. femur-rubrum, M. bivittatus) have also least definite life-histories. They mature early, remain active till frosts, and are sometimes two-brooded. Their wide distribution in time is probably due to the same cause as their wide distribution in space, namely, their lack of responsiveness to slight differences in environmental conditions. V. SUMMARY. Distribution of grasshopper species within the region studied bears evident relation to the plant associations. The plant association is the index to environmental conditions, and its extent marks the area of the habitat. In studies of local dis- tribution within a region, all the plant associations, including local and ruderal associations, should be considered, since other- wise certain habitat relations may be overlooked. As an environmental complex, the plant association is usually more inclusive than the total of conditions required by a particu- lar species, particularly if it is climatic or extensive. Associations may be similar as environments for grasshopper species if they agree merely in including physical and vegetational conditions critical to those species. Grasshoppers select habitats or asso- ciations in which favorable conditions are to be found, irrespec- tive of past history, extensiveness, geographic or successional relationships of the vegetation. There is very seldom any direct relation between grasshopper species and species composition of the plant associations, as few grasshoppers are selective feeders. Most of the grasshopper species are of the ground stratum, and soil conditions are essential. These grasshoppers are typical of open herbaceous associations, characteristic in initial stages of vegetational development in dry ground. Species of open forest 1/8 ARTHUR G. VESTAL. are mostly grassland forms, since ground conditions are essen- tially the same in both situations. In closed forest and moist herbaceous associations species and individuals are less numerous, are not confined to the ground stratum, and are more intimately associated with vegetational conditions. Grasshopper species can be arranged according to gradients of environmental factors (Fig. i, p. 156). Grasshopper succession is incidental to the development of vegetation. The change is not only one of species, but of habits as well. Grasshopper species have in general the geographic range of the types of associations which include the necessary physical and vegetational conditions. The ranges of many species are in agreement with the areas of vegetation provinces. Species of least definite local distribution are widespread geographically. Seasonal differences in time of activity of grasshopper species are marked. This is probably partly due to antagonistic influence of other animals. Seasonal and local distribution are interrelated. Species of indefinite local distribution have also least definite time distribution. UNIVERSITY OF COLORADO, April 5, 1913. VI. ACKNOWLEDGMENTS AND BIBLIOGRAPHY. The privileges of the University of Michigan Biological Station at Douglas Lake were extended to the writer during the summer of 1912, for which he is grateful to the Director, Professor Jacob Reighard. The writer is indebted also to Miss Alvalyn E. Woodward, who has furnished data concerning grasshopper species not taken by the writer; to Dr. H. A. Gleason, who has been of great assistance in the study of the plant associations; and to Dr. V. E. Shelford and to Dr. M. M. Ellis, who have read the manuscript. BIBLIOGRAPHY. Adams, C. C. '06 An Ecological Survey in Northern Michigan. Report from the University Museum, published by the State Board of Geol. Surv. as a part of the Report for 1905. Lansing. '08 An Ecological Survey of Isle Royale, Lake Superior. Report from the University Museum, published by the Geol. Surv. Lansing. GRASSHOPPERS IN RELATION TO PLANT ASSOCIATIONS. 179 Blatchley, W. S. '98 Two New Melanopli from Les Cheneaux Islands, Michigan. Psyche, 8: I95-*97- (Habitat of Melanoplus islandicus.) '02 The Orthoptera of Indiana. 27th Ann. Rep. Dept. Geol. and Nat. Re- sources: 123-471. Forbes, S. A. '09 The General Entomological Ecology of the Indian Corn Plant. Am. Nat., 43: 286-301. Gates, F. C. "u Summer Bird Life in the Vicinity of Havana, Illinois, in its Relation to the Prominent Plant Associations. Wilson Bull. No. 74: 1-27. '13 The Vegetation of the Region in the Vicinity of Douglas Lake, Cheboygan County, Michigan, 1911. Mich. Acad. Sci., 14: 46-107. Gleason, H. A. '08 The Ecological Relations of the Invertebrate Fauna of Isle Royale, Michi- gan. In Adams, 1908: 57-78. '10 The Vegetation of the Inland Sand Deposits of Illinois. Bull. 111. State Lab. Nat. Hist., 9: 23-174. Hart, C. A., and Gleason, H. A. '07 On the Biology of the Sand Areas of Illinios. Bull. 111. State Lab. Nat. Hist., 7: 137-272. Hancock, J. L. 'n Nature Sketches in Temperate America. Chicago. Lugger, O. '97 The Orlhoptera of Minnesota. Bull. No. 55, Agr. Exp. Sta. Minn.: 91-386. Morse, A. P. '04 Researches on North American Acridiidx. Carnegie Inst. Wash., Publ. No. 18: 7-55. '06 The Ecological Relations of the Orthoptera in the Porcupine Mountains, Michigan. In Adams, 1906: 68-72. '07 Further Researches on North American Acridiidce. Carnegie Inst. Wash., Publ. No. 68: 3-54. '08 Report on the Isle Royale Orthoptera of the 1905 Expedition. In Adams, 1908: 299-303. Rehn, J. A. G. '04 Notes on the Orthoptera of the Keweenaw Bay Region of Baraga County, Michigan. Ent. News, 15: 229-236, 263-270. Ruthven, A. G. 'n A Biological Survey of the Sand Dune Region on the South of Saginaw Bay, Michigan. Mich. Geol. and Biol. Surv., Publ. 4, Biol. Ser. 2. Scudder, S. H. '98 Revision of the Orthopteran Group Melanopli (Acridiidce), with Special Reference to North American Forms. Proc. U. S. N. M., 20: 1-421. Shelford, V. E. '07 Preliminary Note on the Distribution of the Tiger Beetles (Cicindela) and its Relation to Plant Succession. Biol. Bull., 14: 9-14. 'n Physiological Animal Geography. Jour. Morph., 22: 551-618. 'iaa Ecological Succession. IV. Vegetation and the Control of Land Animal Communities. Biol. Bull., 23: 59-99. 180 ARTHUR G. VESTAL. 'i2b Ecological Succession. V. Aspects of Physiological Classification. Biol. Bull., 23: 331-370. Shull, A. F. 'n Thysanoptera and Orthoptera (of the Saginaw Bay region, Michigan). In Ruthven, 1911: 217-231 (section on Orthoptera). Vestal, A. G. '133 Notes on Habitats of Grasshoppers at Douglas Lake, Michigan. Ent. News. (In press; collection records and dates.) 'i3b An Associational Study of Sand Prairie. (In press.) Walden B. H. 'n The Ruplexoptera and Orthoptera of Connecticut. Conn. State Geol. and Nat. Hist. Surv., Bull. No. 16: 41-169. Walker, E. M. '03 The Genus Podisma in Eastern North America. Canad. Ent., 35: 295-302. Woodward, A. E. 'n The Orthoptera Collected at Douglas Lake, Michigan, in 1910. Mich. Acad. Sci , 13: 146-167. THE ASEXUAL CYCLE OF PLANARIA VELATA IN RELATION TO SENESCENCE AND REJU- VENESCENCE. C. M. CHILD. I. THE LIFE CYCLE UNDER NATURAL CONDITIONS. During early spring in the region about Chicago, a planarian appears in temporary ditches and pools, particularly in those which are more or less filled with dead leaves. It is also often found in permanent bodies of water such as springs, permanent ponds and brooks, but seems to attain the greatest numbers in the temporary ditches and pools. The animal is apparently the species recently described by Stringer ('09) and named Planaria velata. The shape and proportions of the larger individuals are indicated in Fig. i. When the animals first appear soon after the ice melts they are mostly only 2-3 mm. in length and commonly light in color. They grow rapidly and soon the dorsal surface becomes very deeply pigmented so that they appear almost black. They are very active and their locomotion is much more rapid than that of most other fresh water planarians. During this period they react readily to meat of various kinds and can be collected in large numbers by placing pieces of meat in the water. In about four weeks they attain a length of 12-15 mm., their movements gradually become slower, they cease to react to food, become light gray in color from loss of pigment and sooner or later the pharynx disintegrates. Within a few days after these changes a process of division begins. As the worms creep about, the extreme posterior end adheres to the substratum and the rest of the animal pulls away and leaves it behind as a small fragment which becomes more or less spherical and within a few moments is covered with a slime which adheres to the underlying surface and hardens into a cyst. This process of division is repeated, often several times within a few moments, so that as the animal moves across the 181 182 C. M. CHILD. containing vessel it may leave behind it a series of such pieces. The pieces vary considerably in size, some being as large as 1.5 mm. in diameter, some only about 0.5 mm. The process con- tinues until half or two thirds or sometimes even more of the worm is separated into pieces and then the anterior region including the head may encyst without further division or in some cases dies. Under natural conditions the encysted pieces remain quiescent during the summer and the following winter and in early spring emerge from the cysts as minute, very active worms which at once begin to feed and grow and repeat the cycle. As I have determined by experiment, the encysted pieces are not capable of with- standing desiccation and it is probable that this fact is connected with the occurrence of the worms in ditches and pools partly filled with dead leaves. In such locali- ties even though the water disappears, the bottom under the thick layer of leaves is always more or less wet and the encysted pieces are not subjected to drying. During the last thirteen years I have collected these worms almost every year and have never found a single individual with mature sexual organs or even any indi- cation of sexual reproduction. Every year the active period ends with decrease in activity, cessation of feed- 1 ing, loss of pigment, fragmentation and encystment of the fragments. In this species then, under the conditions where it. occurs in this locality, development and growth result in a process of senescence, the individual breaks up into fragments which under- go regulation within the cysts to small whole animals, and these are to all appearances physiologically as well as morphologically young and are capable of repeating the life cycle. In short, senescence is followed in these animals not by death but by asex- ual reproduction and rejuvenescence. During a number of years I have kept a stock of these worms in the laboratory, have bred them through several asexual generations and have subjected them to various experimental conditions. The results of this asexual breeding and the experi- ASEXUAL CYCLE OF PLANARIA VELATA. 183 mental modifications of the life cycle will be discussed in another paper. II. THE PHYSIOLOGICAL RESISTANCE TO DEPRESSING AGENTS OF YOUNG AND OLD WORMS. In these experiments the method of comparing resistances which I have called the direct method was used. Here the depressing agent is used in sufficiently high concentration to kill the animals within a few hours and the occurrence of death is determined by disintegration of the worms which begins within a few moments after death. This method has been fully de- scribed in another paper (Child, '130). In that paper it was shown that with this method the animals with the higher rate of metabolism or more strictly of cell respiration are less resistant and therefore die and disintegrate earlier than those with the lower rate. Thus the differences in resistance enable us to compare the rates of respiration and so in a general way the rates of metabolism. In Table I. the first vertical column gives the length of time in the depressing agent in hours and minutes, the second the serial numbers of the lots of worms compared, and the columns I.-V. under "Stages of Disintegration" give the number of worms of each lot in each stage of disintegration at each time. As regards the five stages, I have found it convenient to distin- guish more or less arbitrarily these stages in the process of dis- integration, for disintegration usually appears first in certain definite regions of the body while other regions are still alive and show movement and it follows a more or less regular course (Child, 'i3«). The five stages are briefly characterized as fol- lows : I. Intact, no disintegration. II. Disintegration beginning, usually in head region. III. Body beginning to disintegrate but form still retained. IV. Margins disintegrated, form disappearing in consequence of swelling of tissues and separation of cells. V. All epithelium and pigment gone; swelling of tissues has extended to all parts and original form has disappeared. The distinction of these stages makes it possible to compare 1 84 C. M. CHILD. different lots of worms more closely than if only the time of complete disintegration were noted. In Table I. Lot I consists of ten worms 1.5-2 mm. in length, which had emerged from cysts within three or four days preceding and had been fed once with pieces of earthworm after emergence. Lot 2 consists of ten worms 13-15 mm. which had been raised in the laboratory from cysts with earthworm as food and were almost ready to encyst again. TABLE I. Series 77. KCN o.ooi mol. Length of Time in KCN. Lots. Stages of Disintegration. 1. II. III. IV. V. I 8 i I 1-30 2 10 I 5 i I 3 2.0O 2 10 I 4 i 5 2.30 2 10 I 2 2 i 5 3.00 2 10 I 3 7 3-30 2 6 3 i I 10 4-OO 2 3 5 i I 4-30 2 8 i i 5-oo 2 4 4 I i 5-30 2 i 4 3 2 6.30 2 3 7 7-30 10 It is evident at once from the table that the resistance of the worms of Lot I recently emerged from cysts is very much less than that of the large worms of Lot 2. Disintegration begins in ASEXUAL CYCLE OF PLANARIA VELATA. 185 Lot i after one and one half hours in KCN, while the worms of Lot 2 are still intact and slowly moving about. In Lot 2 dis- integration begins after three hours. All the worms of Lot I are completely disintegrated after three and one half hours, those of Lot 2 after seven and one half hours, i. e., the survival time of Lot 2 is nearly double that of Lot i. In other words the worms of Lot i have a much higher rate of metabolism than those of Lot 2. That the difference in size of the worms is not responsible for the difference in survival time Is evident for two reasons : first in these flattened elongated animals the surface increases almost as rapidly as the volume and second the time of beginning of disintegration (Stage II.) is much later in Lot 2 than in Lot I. The earliest stages of disintegration involve the external surface of the body and the surface of the large wrorms including the cilia remains alive for a much longer time than that of the small worms. Moreover, if the difference in size determined the dif- ference in survival time we should expect that this would be much greater since the small worms are only a minute fraction of the size of the larger. The difference in the rate of the meta- bolic processes affected by the KCN is the only factor which will account for the results (Child, '130). Unfortunately it has thus far been impossible to compare the worms emerging from the cysts with young worms hatched from eggs because I have never observed sexual reproduction in this species, but the difference in rate of metabolism between the small and large worms is similar to the difference known to exist in other forms between young animals sexually produced and old. It is, however, not necessary to use the extremes of the life cycle for comparison. Animals in various stages of growth may be compared and in all cases those which are nearer the stage when encystment occurs, i. e., those which are older as regards growth and development, show the higher resistance. In Table II. the survival times of a series consisting of five worms 5-6 mm. in length (Lot i) and five worms 11-12 mm. -(Lot 2) are given. 186 C. M. CHILD. TABLE II. Series 64, I., II. KCN o.ooi mo!. Length of Time in KCN. Lots. Stages of Disintegration. I. II. III. IV. V. I 2 3 2.30 2 5 I 3 I I 3.00 2 5 I i 2 I I 3-30 2 3 2 I I 4 4.00 2 i 4 I 5 4-30 2 4 I 5-30 2 5 6.30 2 4 i 7-30 2 5 In this series also the younger worms show less resistance, which signifies a higher rate of metabolism, but a comparison of Table II. with Table I. shows that Lot i of Table II. has a longer survival time than Lot I of Table I., i. e., wrorms of 5-6 mm. in length have a lowrer rate of metabolism than worms recently emerged from cysts. These facts show that a progressive de- crease in the rate of metabolism occurs during the growth of the animals. Those newly emerged from cysts have the highest rate, those which are full-grown and nearly ready to fragment and encyst have the lowest rate, w^hile intermediate stages show rates between these two extremes. These results have been confirmed by various other series with both KCN and alcohol. The small newly emerged worms die much earlier in all cases than the large worms. The differences in resistance to KCN, alcohol, etc., between young and old animals are the same in animals freshly collected from their natural habitat as in animals bred for one or more generations in the laboratory'. ASEXUAL CYCLE OF PLANARIA VELATA. I87 The results obtained in this way are further confirmed by the much greater activity of the small recently emerged animals. They move much more rapidly, are much more irritable and show a much higher rate of growth than the large animals. And finally the small worms from the cysts are capable, as noted in the preceding section, of repeating the life cycle. There can I think be no doubt that the worms emerging from the cysts are physiologically young and that they undergo a process of senes- cence as they grow in size. Evidently a process of rejuvenescence is associated in some way with the asexual reproduction which follows growth and development. III. EXPERIMENTAL REPRODUCTION. i. The Course of Experimental Reproduction. The process of reproduction of whole animals from pieces isolated by section is very similar to that in other planarians. Pieces from any region of the body and above a certain limit of size, wrhich varies somewhat with the region, are capable of giving rise to whole animals. As in other species of Planaria, the process consists in part of the outgrowth and differentiation of embryonic tissue from the cut surface and in part of redifferentiation of other tissues to a greater or less distance from the cut surface. In pieces of equal length the amount of anterior new tissue is greater and of posterior new tissue less in those from the anterior region of the body, while with increasing distance of the end of the piece from the head region the amount of anterior new tissue increases and 3 4 188 C. M. CHILD. that of posterior new tissue decreases. The development of the new head is more rapid in anterior than in posterior pieces. The position of the new pharynx is posterior to the middle in anterior and anterior to the middle in posterior pieces. Short pieces from the extreme anterior region frequently fail to develop a new posterior end. Fig. 2 shows a piece of this kind. Figs. 3, 4 and 5 show three pieces, the first from the anterior, the second from the middle and the third from the posterior region. The different amounts of new tissue produced are seen in the figures. All these graded differences, like those in Planaria dorotocephala (Child, 'lie), indicate the existence of a physiological gradient of some sort along the axis. As a matter of fact this gradient is essentially similar to that which exists in P. dorotocephala (Child, '12, 2. The Encystment of Artificially Isolated Pieces in Relation to Size of Piece and Region of Body. Pieces isolated by section may undergo the regulation to whole animals either with or without encystment. The frequency of encystment varies with region of the body from which the piece is taken, with the size of the piece and with the physiological age of the animal. The following records of series will serve to illustrate this. In these series a number of worms, ten, twenty or twenty-five, from the same stock and as nearly as possible of the same size and in the same physiological condition are cut into a number of as nearly as possible equal pieces, the corre- sponding pieces are placed together in one lot and results recorded for each piece. Since different numbers of worms are used in different series the results are given in percentages. Series ip, April ij, IQII. — Ten worms, full grown (12-14 mm.), but still feeding and deeply pigmented. Heads removed and remainder of body cut into two equal pieces, a, the anterior, and b, the posterior. Table III. show's the percentages of the pieces which develop into whole worms without encystment and of pieces which encyst soon after the operation and emerge, from a few days to several weeks later, as whole worms after regulation in the cysts. ASEXUAL CYCLE OF PLAXARIA VELATA. 189 TABLE III. No Encystment. Encystment. a go 10 b 60 40 Series 20, April ij, ign, — Ten worms from same stock, of same size and in same condition as Series 19. Heads removed and body cut into four equal pieces, a, 6, c, d, a being the most anterior. Table IV. gives the results. TABLE IV. No Encystment. Encystment. a 90 10 b 80 20 c 30 70 d 20 80 Series 27, April 77, ign. — Ten worms like those of Series 19 and 20 in size and condition. Heads removed and body cut into eight equal pieces, a-h, a being the most anterior. The results are given in Table V. TABLE V. No Encystment. Encystment. a 20 70 b 10 90 c 100 d 100 e 100 / 100 g 100 h 100 It is evident from these three series and abundantly confirmed by numerous others, first, that the frequency of encystment of pieces increases from the anterior to the posterior end of the body and second, that the frequency of encystment increases as the size of the piece decreases. In all these series the greater fre- quency of encystment in more posterior pieces is evident in greater or less degree. In Series 19 where the pieces represent halves of the body the percentages of encystments are small, in Series 20, composed of | pieces, they are larger except in the C. M. CHILD. most anterior piece and much larger in the two posterior pieces c and d which together equal b of Series 19. And finally, in Series 27 which consists of ^ pieces all the pieces encyst except 30 per cent, of a and 10 per cent, of b. When the pieces are cut still smaller all encyst. The frequency of encystment then shows in pieces of equal size a gradation from the anterior to the posterior end of the body and indicates the existence of some sort of a physiological gradient in the animal. Encystment may, however, occur in pieces from any region if they are sufficiently small, but in general anterior pieces must be smaller than posterior pieces to give the same frequency of encystment. This fact indicates that the physiological state of the piece differs in some way with its size. As a matter of fact this species possesses essentially the same sort of gradient in rate of metabolism as Planaria dorotocephala (Child, '130, 'i3c) and the relation between frequency of encyst- ment, region of the body and size of the piece depends upon the existence of this gradient and the changes in rate of metabolism which occur in pieces from different regions of the body and of different size after isolation. Further consideration of these points is postponed to another time. 3. The Frequency of Encystment of Pieces in Relation to Temperature. Series 5J, October 5, IQII. — Animals 9-10 mm. in length were selected from a stock which had been kept at a temperature of 20° C., the heads removed and the bodies cut into four equal pieces a-d. Lots of ten each of each of the four pieces were placed in three different temperatures, 10°, 20° and 28-30° C. Table VI. gives the results in percentages. It is evident at once from Table VI. that the frequency of encystment is greater with higher than with lower temperature, i. e., the higher the rate of metabolism in the pieces the greater the frequency of encystment. Numerous other series give the same results without exception, not only for pieces, but for whole worms. Worms which have been kept at a temperature of 20°, when placed in a temperature of 30° will often encyst entire while at 20° they remain active until they fragment and the pieces encyst, and at 10° many of them do not encyst at all. ASEXUAL CYCLE OF PLANARIA VELATA. TABLE VI. Temperature. Pieces. No Encystment. Encystment. Dead. a IOO T/-.O b 90 10 IO c 20 80 d IO 90 a IOO 20° b 70 30 c IOO d IOO a IOO 28-30° b c IOO IOO d IOO 4. The Frequency of Encysiment in Pieces in Relation to Age. In the very small young worms recently emerged from cysts pieces, unless very small, usually reproduce whole worms with- out going through a period of encystment. As the worms in- crease in size and become physiologically older the frequency of encystment increases until in worms which are almost ready to fragment and encyst naturally all pieces resulting from section usually encyst. The differences in this respect between half grown worms, worms which are about full grown but have not yet ceased to feed and still retain their dark color and worms which have stopped feeding and become gray in color are shown in the three series following. Series 47, September 21, IQII. — Twenty worms about half grown (7 mm. in length) wrere cut into four equal pieces, a-d. The percentages of regulation without encystment and of en- cystments appear in Table VII. TABLE VII. No Encystment. Encystment. Dead. a .... 0=> b QC ; c . . . t;? 4? d.. 2O 80 C. M. CHILD. Series 56 /, October 12, IQII. — Ten worms full grown but still dark in color and still feeding. Body cut into four equal pieces, a-d. Table VIII. gives percentages of encystments. TABLE VIII. No Encystment. Encystment. a 60 40 b 100 C 100 d ioo Series 58 /, October 13, IQII. — Ten worms, full grown, gray in color and no longer feeding. Body cut in four equal pieces, a-d. Table IX. gives percentages. TABLE IX. No Encystment. Encystment. a ioo b ioo c ioo d ioo The older worms show the greater frequency of encystment of pieces. The same results have been obtained in other similar series without exception. 5. The Physiological Condition of Animals Reproduced from Artificially Isolated Pieces. The animals reproduced from pieces isolated by section are physiologically young, whether a period of encystment occurs or not. In this respect they are similar to the worms produced from the pieces which separate and encyst naturally. Small pieces cut from the bodies of old worms and allowed to repro- duce whole animals show the same differences in rate of metab- olism from old animals as the worms emerging from cysts naturally produced. The differences in susceptibility to cyanide are essentially the same as in Table I. Moreover, these small worms arising from pieces of large old worms are capable of rapid growth if fed and of repeating the life cycle. As they grow the rate of metabolism, as indicated by their susceptibility to ASEXUAL CYCLE OF PLANARIA VELATA. 1 93 cyanide, decreases, the rate of growth and the degree of activity also decrease, they finally stop feeding, lose their dark color and give rise to cysts again and from these a new generation of young worms emerges. Stocks of animals produced from pieces have passed through this cycle repeatedly in the laboratory. The degree of rejuvenescence in this experimental reproduc- tion varies with the size of the piece. The smaller the piece, the more extensive the reorganization and the younger the worm which results. In all respects these results are essentially the same as those obtained with Planaria dorotocephala and described in an earlier paper (Child, 'n&). It is evident also that there is no essential difference in this respect between the process of fragmentation in old worms and the reproduction of young worms from the encysted pieces in nature and the process of experimental reproduction of animals from pieces isolated by section. In nature the fragmentation occurs only in old animals by a process characteristic of a certain stage of the life cycle. In experiment the pieces can be isolated at any stage of the life history and may be of any size. In both cases the reorganization, together with the period of starvation which is also a factor as will appear, brings about rejuvenescence and the worms thus produced are capable of repeating the life history from the stage at which they begin again to feed to the stage of fragmentation. IV. THE NATURE OF THE PROCESS OF ENCYSTMENT. It has been shown that the frequency of encystment of pieces increases writh rising temperature, with decreasing size of the piece, with increasing distance of the level of the piece from the head region and with advancing age of the animals. Pieces from any region of the body may encyst if the temperature is suffi- ciently high, if the pieces are sufficiently small or if the animal is sufficiently old. All of these conditions must have something in common as regards their effect upon the pieces since all produce similar results. What is this common factor? When a piece is cut from the body it is stimulated and its rate of metabolism increases. This is generally admitted but it can also be demonstrated by the cyanide method. The suscepti- 194 C- M- CHILD. bility to cyanide of a piece immediately after isolation is much greater than that of the corresponding region of the body in an uninjured animal of the same age and physiological condition. This greater susceptibility of the piece means that it has been stimulated by the act of isolation. After this sudden rise its susceptibility to cyanide decreases gradually during twenty-four hours or more and in small pieces may fall below that of corre- sponding regions in the uninjured animal (Child, 'i3&). This decreasing susceptibility means that the rate of metabolism in the piece is gradually decreasing as the stimulation resulting from section gradually disappears. The cyanide method shows further that the degree of stimula- tion increases as the size of the piece isolated decreases and also as the distance of the level of the piece from the head region increases. In other words smaller or more posterior pieces are more stimulated by the act of section than larger or more anterior pieces. And finally pieces cut from worms at a higher tempera- ture writhin certain limits are more stimulated and show a greater increase in rate than pieces from worms at a lower temperature. These relations between the degree of stimulation of pieces and the factors of size of piece and region of the body and various external conditions have been worked out completely for Planaria dorotocephala and the data will be presented in full elsewhere. Sufficient wrork has been done on P. velata to show that the relations are essentially the same as in P. dorotocephala, but since the work on the latter species furnishes the foundations for the conclusions and since the data for that species are in more com- plete form and will be published in a short time the evidence for the above statements concerning the degree of stimulation in pieces of P. velata is not presented in detail. So far then as region of the body, size of piece and temperature are concerned the frequency of encystment of pieces in P. velata runs parallel to the degree of stimulation by the act of section. Apparently the more the piece is stimulated by section the more likely it is to encyst. The process of encystment in this species consists in the rapid secretion over the surface of the body of a thick slime which ASEXUAL CYCLE OF PLANARIA VELATA. 195 soon hardens into a tough membrane and forms the cyst. It is a familiar fact that stimulation is often followed in the turbellaria by the secretion of a large amount of slime. That is exactly what occurs in these pieces and in this species the slime hardens and forms the cyst. Apparently then the encystment of pieces in Planaria velata is simply the result of a sudden stimulation. Any factor that increases the stimulation increases the frequency of encystment. As regards the greater frequency of encystment with advancing age of the worms, I have not been able to reach a definite con- clusion based on experiment, but my observations indicate that old worms secrete more slime on stimulation than young. Ap- parently the gland cells either increase in number or the quantity of the substance in them which produces the slime increases as the animals grow older. When the slime which produces the cyst first appears it is soft and an active whole animal is able to creep out of it without difficulty, but the pieces are much less active and do not succeed in escaping from it before it hardens. If the cysts are carefully opened with needles soon after they are formed and the pieces removed without injury or any great degree of stimulation they usually do not encyst again but develop into whole worms while free. But if they are injured or otherwise strongly stimu- lated they commonly encyst a second time. In short all the facts indicate that encystment of pieces is merely a result of the stimulation accompanying section. It is not an adaptation to conditions or a preparation for the future in any sense. The animals do not encyst because they usually live in temporary bodies of water but they are able to live under these conditions because they encyst. V. THE PROCESS OF FRAGMENTATION IN OLD WORMS. The process of fragmentation in nature is very evidently similar in character to the process of zooid-formation and fission in Planaria dorotocephala and P. maculata (Child, 'lie}. In consequence of increase in length of the body and the decrease in rate of metabolism as the animal becomes older the posterior regions of the body usually become to some extent physio- IQ6 C. M. CHILD. logically isolated from the dominant region (Child, 'na, 'nd). That the occurrence of fragmentation is connected with a decrease in the rate of metabolism and consequent physiological isolatiion of posterior regions is clearly indicated by the fact that fragmentation may often be induced, even in worms which are not full-grown, by suddenly lowering the temperature ten to fifteen degrees. In such cases fragmentation usually begins in the posterior region within a few days. The degree of isolation is not sufficient to permit development at once into a new individual but it is sufficient to permit some degree of independence in motor reaction, consequently, at some time when the worm is creeping the posterior end attaches itself and the rest of the body pulls away from it, as in P. dorotocephala. Apparently the greater part of the body in old fragmenting animals consists of a series of these small zooids for in most animals fragmentation continues until only the anterior third or fourth of the body together with the head remains. This anterior piece may then encyst or may undergo rejuvenescence without encystment and after some weeks give rise to a new posterior end, or in some cases it dies. The posterior zooids are present only dynamically and not mor- phologically, at least not visibly, and they are not to be thought of as absolutely fixed stable entities. When the animal is strongly stimulated it is able to control the whole length of the body and for the time being the posterior zooids may almost or quite cease to exist, only to reappear after the stimulation is over. When such zooids are established the regions at their ends must be subjected to constantly varying correlative conditions. Sometimes they may form a physiological posterior part of one zooid, at other times an anterior part of another and at still others a part of neither. Such changes in correlative conditions must tend to weaken and eliminate the existing structure in those regions since the development of such structure depends on a certain degree of constancy in correlative factors. In this way zones of structural weakness arise and these are the zones where separation occurs. Occasionally, either in consequence of weakness or perhaps because the physiological isolation of the posterior regions is ASEXUAL CYCLE OF PI.ANARIA VELATA. 197 insufficient the worm fails to fragment. In such cases parts of the body may become greatly elongated and a string of connected masses may arise. Figure 6 shows such a case. In the posterior region four distinct masses can be distinguished. These are connected by slen- der bands which are merely portions of the body greatly reduced in diameter. These four masses are connected with the anterior portion of the body by a long slender band resulting from the stretching of the middle region in consequence of the attempts of the head region to pull away from the attached posterior parts. These greatly elongated regions of the body consist of little more than the body-wall and muscles; the alimentary tract and the paren- chyma may be almost or entirely squeezed out of them. This animal finally became surrounded by a cyst in the form shown in the figure, but later the connecting strands apparently atro- phied, the pieces became entirely separate and each produced a whole worm. VI. THE DEVELOPMENT OF THE WHOLE ANIMAL WITHIN THE CYST. The development of the animal from the encysted piece, whether isolated artificially by section or by the natural process of fragmentation, is similar in all respects to the regulatory de- velopment of pieces which reproduce new wholes without en- cystment. This is showTn to be the case by the removal of the cysts from pieces at various stages of the process. In all cases the pieces are simply undergoing regulation. The process with- in the cyst may, however, be slower than in the unencysted piece, probably because the supply of oxygen within the cysts is less than in the water. The natural method of asexual reproduction in this species does not then differ essentially in any way from the process of experimental reproduction. The process of fragmentation gives 198 C. M. CHILD. rise to the same conditions in the piece as experimental isolation by section and the further history is the same in both cases. 7 8 Many teratological forms result from irregularities in frag- mentation or incomplete separation. The most common are partial duplications of anterior or posterior regions (Figs. 7 and 8) but various other forms appear. In Fig. 9, for example, a case is shown in which an incompletely separated posterior piece gave rise without encystment to two heads, a tail and two outgrowths of uncertain character, and Fig. 10 shows a case in which two worms with axes at right angles to each other are united by the middle regions of their dorsal surfaces. Ordi- narily the larger animal carried the other about on its back as in the figure, the ventral surface of the smaller worm being upper- most. Fig. 1 1 represents a case of so-called axial heteromorpho- sis and in Fig. 12 two heads appear at the posterior end of the larger individual and dorsal to them a tail. Evidently new ASEXUAL CYCLE OF PLAXARIA VELATA. 199 polarities arise very readily in the small pieces which result from fragmentation, probably because the pieces are so short t'hat the original axial gradient (Child, 'i3c) is practically eliminated and chance differences in the rate of metabolism in different parts of the piece are sufficient to establish new polarities. VII. CONCLUSION. In Planaria velata the individual very evidently undergoes a process of senescence as it grows and either experimental or natural asexual reproduction brings about rejuvenescence. More- over, the animal apparently returns to essentially the same physiological stage with each generation, for the species is able to persist without sexual reproduction and, as a following paper wrill show, numerous asexual generations have been bred in the laboratory without any indication of senescence of the stock. I have shown elsewhere (Child, 'lib) that the regulation of 2OO C. M. CHILD. isolated pieces of Planaria dorotocephala brings about rejuvenes- cence to a greater or less extent, according to the size of the piece,, the smaller piece giving rise to an animal which is physiologically younger than that produced by a larger piece. In that species starvation may also be a factor in rejuvenescence. Some experi- ments on the effect of starva- tion on Planaria velata will be described in another paper. At present it need only be said that the result is the same in both species. In my earlier paper on senes- cence the conclusion was reached that senescence results from the accumulation of structural prod- ucts of metabolism which con- stitute in one way or another obstacles to the chemical reac- tions. The processes of differentia- tion and growth undoubtedly ope- rate also in another way not considered in the earlier paper, to bring about a decrease in the rate of metabolism per unit of weight or volume. What we are accustomed to call the undiffer- entiated or embryonic cell repre- sents the general metabolic sub- stratum of the organism. Differ- entiation consists in the formation and accumulation of certain substances in the cell, some of which constitute more or less permanent structural features. At least certain of the substances composing these structural features are relatively stable under the usual physiological conditions and while certain chemical changes may occur in them, they are not broken down and elimi- nated to so great an extent as certain other substances. This relative stability must, in fact, be the basis of their persistence as elements of structure. The accumulation of these structural. 12 ASEXUAL CYCLE OF PLAN ARIA VELATA. 2OI substances within the cell brings about a decrease in the general metabolic activity per unit of weight or volume because it de- creases the proportion of the material involved in the general metabolic reactions to the inactive or less active material. The decrease in the proportion of the general metabolic substra- tum characteristic of the embryonic cell constitutes to some extent a histological criterion of the physiological change in the cell. In short, the decrease in rate of metabolism per unit of sub- stance, which is characteristic of development and senescence, is undoubtedly due in part to the fact that the proportion of the cell substance concerned in the general metabolic activity is decreasing and the proportion of less active or relatively stable substance is increasing. Changes in the size of the cell or in the size relations of nucleus and cytoplasm (Minot, '08) are not necessary factors in the result. To what extent the decrease in the rate of metabolism during senescence is due in a given case to actual decrease in the rate of chemical reaction and how far to a decrease in the proportional amount of chemically active or more active substance is often difficult to determine, but it is probable that in some cases, or «ven in some cells of the individual, the one factor and in others the other is the more important. As regards Pla naria velata, the facts are that the rate of metab- olism decreases during growth and development and increases when the substances previously accumulated are removed, either by regulatory reorganization, or by starvation. These facts show very clearly that in one way or another the accumulation of material in development decreases the rate of metabolism and its removal brings about an increase in rate. Senescence and rejuvenescence in this species consist essentially, I believe, in these changes. SUMMARY. i. After a period of growth and activity Planaria velata under- goes fragmentation from the posterior end forward, the frag- ments encyst and give rise by a process of regulation to whole worms of small size. 2O2 C. M. CHILD. 2. During the period of growth the worms are undergoing senescence, as the decrease in rate of metabolism indicates, but the small worms which emerge from the cysts are physiologically, as well as morphologically young, possess a high rate of metab- olism and are capable of repeating the life cycle. 3. In pieces isolated by section the frequency of encystment increases as the level of the piece becomes more posterior in the body, with decreasing size of the piece, with rising temperature and with increasing age of the animal. The facts indicate that encystment is the result of stimulation. The stimulation may result from section, from fragmentation, from a rise in tempera- ture or from other conditions. 4. The development of the encysted piece into a new whole animal is essentially the same process as the regulatory develop- ment of unencysted pieces. 5. This species is able to live for an indefinite number of gen- erations without sexual reproduction. Each new asexual genera- tion represents a return to essentially the same physiological and morphological stage. In other words, senescence leads to reproduction and the process of rejuvenescence in each asexual cycle carries the organism back to the same stage of youth. HULL ZOOLOGICAL LABORATORY, UNIVERSITY OF CHICAGO. REFERENCES. Child, C. M. 'ua Die physiologische Isolation von Teilen des Organismiis. Vortrage und Aufs. ii. Entwickelungsmech., H. XL 'lib A Study of Senescence and Rejuvenescence, Based on Experiments with Planarians. Arch. f. Entwickelungsmechanik, Bd. XXXI. , H. 4. 'lie Studies on the Dynamics of Morphogenesis and Inheritance in Experi- mental Reproduction. I. The Axial Gradient in Planaria dorotoce phala as a Limiting Factor in Regulation. Journ. Exp. Zool., Vol. X., No. 3. 'lid Studies, etc., II. Physiological Dominance of Anterior over Posterior Regions in the Regulation of Planaria dorotoce phala. Journ. Exp. Zool., Vol. XL, No. 3- 'ne Studies, etc., III. The Formation of New Zooids in Planaria and other Forms. Journ. Exp. Zool., Vol. XL, No. 3. '12 Studies, etc., IV. Certain Dynamic Factors in the Regulatory Morpho- genesis of Planaria dorotoce phala in Relation to the Axial Gradient. Journ. Exp. Zool., Vol. XII., No. i. '133 Studies, etc., V. The Relation between Resistance to Depressing Agents and Rate of Reaction in Planaria dorotoce phala and its Value as a Method of Investigation. Journ. Exp. Zool., Vol. XIV., No. 2, 1913- ASEXUAL CYCLE OF PLANARIA VELATA. 203 'i3b Certain Dynamic Factors in Experimental Reproduction and their Significance for the Problems of Reproduction and Development. Arch. f. Entwickelungsmech., Bd. XXXV., H. 4. '130 Studies, etc., VI. The Nature of the Axial Gradients in Planaria and their Relation to Polarity and Symmetry. Arch. f. Etwickelungs- mechanik, Bd. XXXVII., H. i. Minot, C. S. '08 The Problem of Age, Growth and Death. Stringer, Caroline E. 'og Note on Nebraska Turbellaria, with Descriptions of two New Species. Zool. Anz., Bd. XXXIV., No. 9. DEVELOPMENTAL CHANGES OF PIECES OF FROG EMBRYOS CULTIVATED IN LYMPH. S. J. HOLMES. In the course of some experiments on the behavior of the epithelial tissue of frog embryos when cultivated in lymph or plasma it was found that pieces of tissue sometimes underwent developmental changes. During the comparatively short time in which material was available for experimental work, oppor- tunity was not found for devoting as much attention to the changes in these pieces as was desirable. Nevertheless a few of the things observed were sufficiently striking and suggestive to be worthy of record, and it is hoped that the observations may be made more complete at a future date. The material used consisted of embryos and young larvae of Rana taken out of the jelly a short time before they wrere ready to make their natural exit. The jelly was placed for a few min- utes in an antiseptic solution, and the embryos were afterward cut to pieces in sterile Ringer's solution. Each piece was then mounted in a hanging drop of lymph or plasma of the adult frog, sealed up in a hollow slide, and kept in a cool place. Every few days the piece was transferred to a fresh medium. Changes in the ectodermic epithelium were quite manifest a few hours after the pieces were mounted, and in one or two days remarkable strands and sheets of ectodermic cells were to be seen extending into the culture medium. These structures and the behavior of their component cells are more fully described in another paper now in course of publication. But aside from the movements of the ectoderm, there were, in certain of the pieces, marked developmental changes of the internal parts. Irregular fragments of tissue frequently became more rounded in form; sometimes they put out lobes or processes of various kinds, and, in some cases, they increased noticeably in size. In a piece from the head region of a late embryo shortly before the rudiments of gills made their appearance, there were developed, 204 PIECES OF FROG EMBRYOS CULTIVATED IN LYMPH. 2O5 in addition to several larger lobes, a number of finger-like proc- esses which very closely resembled the filaments of the gills. The central part of these projections was composed of connective tissue, and the epithelial covering consisted of a layer of columnar FIG. i. Piece from the head of a frog embryo drawn fifteen days after being cul- tivated in lymph, e, isolated ectoderm cells. ciliated cells. These outgrowths began three days after the preparation was made, and they were practically completely formed in a week. Fifteen days after the piece was mounted it had assumed the form shown in Fig. I. The cilia on the projec- tions were beating, and in fact they continued active for two more weeks, but there was little further change in the general form of the piece. When originally mounted the piece was opaque, and the cells composing it contained a large amount of yolk. After two weeks it had become nearly transparent; the larger part of the yolk in the cells had been assimilated, and a typical connective tissue had developed in the interior from the com- paratively unspecialized cells of the mesoderm. The piece was covered completely by ectoderm which, over most of the surface, was flattened, but assumed a cylindrical form on the more promi- nent projections. In the third week a layer of epithelium was sloughed off from a large part of the surface of the piece. Whether the finger-like processes represent gill filaments is more or less open to question. Processes of this kind failed to make their appearance from the hinder part of the embryos. 2O6 S. J. HOLMES. Should they be gill filaments they would afford a case of the self differentiation of an organ in the absence of certain stimuli, such as those afforded by the blood supply, which normally are associated with their formation. Fig. 2 represents a development from a cross section of a late embryo near the middle of the tail region. This piece, like the one just described, was at first opaque, owing to the large amount of yolk in the cells. A week later it became quite transparent, FIG. 2. A cross section from near the middle of the tail of a frog embryo drawn seven days after being cultivated in lymph. and numerous pigment cells became differentiated beneath the ectoderm. Externally it was completely covered by a layer of flattened epithelial cells. The connective tissue within the piece had become more specialized. At either end of the piece there appeared a prominent outgrowth, the one at the posterior end curving upward and forward. There were two large cavities lined with a single layer of much flattened endothelial cells. These cavities were much larger than any spaces in the original piece of the embryo, and not improbably represent closed and distended blood vessels. The notochord showed at each end an extension of cells which suggested that the organ was undergoing regeneration in both directions. These extensions lay entirely within the new outgrowths from the cut ends of the piece. Fig. 3 represents a developed fragment taken five days pre- viously from near the middle of an embryo. This piece contained some entoderm which is included in the dark pigmented mass PIECES OF FROG EMBRYOS CULTIVATED IN LYMPH. 207 near the center. At one side there was a thin vesicle filled with fluid. The ectoderm cells covering this vesicle were much flat- tened. Ten days later the whole piece was larger, owing to the increase in the size of the hollow vesicle which had come to 3 FIG. 3. FIG. 4. FIG. 3. A piece which developed from a fragment of a middle portion of a frog embryo. FIG. 4. The same piece as is shown in figure 3, but drawn ten days later. surround most of the rest of the fragment (Fig. 4). The pig- mented mass and the tissue around it had assumed a more nearly spherical form, and the whole structure had become more nearly spherical also. Apparently the chief factor in the form changes in this case was the absorption of fluid which caused the increasing distension of the thin walled vesicle. A similar proc- ess may account for the large spaces in the fragment shown in Fig. 2. In transferring the nearly spherical fragment to fresh lymph the thin wall of the vesicle partly collapsed, but it was subsequently distended to its previous form. During the time the piece was kept alive, which was about five weeks, it showed no other marked changes in form. UNIVERSITY OF CALIFORNIA, June 16, 1913. NEMATOLAMPAS, A REMARKABLE NEW CEPHALO- POD FROM THE SOUTH PACIFIC. S. STILLMAN BERRY, REDLANDS, CALIFORNIA. In a collection of Cephalopoda from Sunday Island, one of the Kermadec Group, which has been recently placed in my hands for investigation, is a new oegopsid possessing features of such general interest that the sender of the specimens, Mr. W. R. Oliver of Auckland, has kindly permitted me to offer the follow- ing brief preliminary account of the species in an American publi- cation pending the appearance of a full report on the collection in the Transactions of the New Zealand Institute. Nematolampas regalis new genus and species. The body is small, cylindro-conical in outline, and terminates in a not very sharp point posteriorly. The fins are very large in proportion to the body, their combined width being as great as the length of the latter, although they are not quite half so long. They are practically continuous with one another above, at least posteriorly, and their union with the mantle is very insecure in alcoholic material. On each side of the extreme posterior tip of the body and lodged in the angle between the fin and mantle is a small swelling containing a conspicuous heavily pigmented body of spherical outline which from its general appearance has undoubtedly a photogenic function. This entire mass with its contained organ is not very firmly attached and becomes readily dislodged when the fins have been loosened. The head is relatively large, short, and rounded. The eyes are very large. Bordering the ventral periphery of each eyeball is a longitudinal series of five beadlike photophores of a reddish color. Of these the central one is considerably the largest and so conspicuous that it is readily visible through the outer integu- ment. Both it and the smaller organs just adjacent to it are 20H NEMATOLAMPAS, A REMARKABLE NEW CEPHALOPOD. 2OQ circular-ovate in outline, but the terminal organs are com- paratively narrow and elongate. The arms are unequal in length, and in one instance, that of the ventro-lateral pair, this inequality is of a very extraordinary nature. This arm pair is in any case the longest and their basal portions much the stoutest, but furthermore each termi- nates in an exceedingly slender beaded filament which when straightened out is considerably longer than the entire remaining portion of the animal and is devoid of suckers, though the arm proper is normal in this respect. The entire arm bears a succes- sion of small but heavily pigmented photogenic organs scattered at various intervals along its outer margin. On the filament these appear as swellings or tubercles, often half as large in diameter as the filament itself, but on the basal portion of the arm they become rather deeply imbedded and are not easily seen except by transmitted light. Including its filament the better preserved arm of this pair carries a total of thirty-one photophores. The remaining arms are normal as regards their extremities, but those of the dorsal and dorso-lateral pairs bear each a single photophore near the tip. All the arms bear two rows of small suckers, but no hooks. The order of relative length is 3, 2, 4, i. The tentacles are about as long as the mantle, their clubs little expanded and armed with four rows of suckers. Each has a pale indistinct swelling in the stalk a short distance from the base and a similar one a little distance below the club. Except that the proximal one is situated considerably nearer the base of the stalk, these swellings occupy a region analogous to the position of the tentacular photophores described by Chun for Thaumatolampas (1903, p. 570, fig.; 1910, p. 59, pi. 1-4) and quite likely represent similar structures. As in Thaumatolampas also some of the more important luminous organs lie within the mantle cavity and in the living animal are visible only by reason of the transparency of the pallial tissues. These organs are eight in number and through- out are clearly homologous in the two genera. The two anterior are situated one on either side of the alimentary canal just back of the funnel and correspond to the anal organs of Chun. 210 S. STILLMAN BERRY. In the present form they are quite small. A little behind the middle of the body is a very conspicuous transverse series of five organs. The central one is unpaired. Close to it on each FIG. i. side is a very large flattened organ. The terminal organs are the smallest of the series and -situated one at the base of each NEMATOLAMPAS, A REMARKABLE NEW CEPHALOPOD. 211 gill. In the posterior part of the pallial chamber between the fins is a large unpaired organ. The medio-dorsal length of the mantle is 32 mm., the width of the same II mm. The length of the entire animal exclusive of the tentacles and the filaments of the third arms is about 57 mm. The length of the third arm pair is over 70 mm. From the above account it will be seen that this small squid possesses no fewer than ninety definitely and symmetrically arranged photogenic organs, and it may well be that there are even others which have escaped my search, as the opacity of the tissues in preserved material in many cases renders their detec- tion difficult. Those which I have been able to make out are shown somewThat diagrammatically in the accompanying drawing, and also in the following tabular statement : Ventral periphery of eyeball 10 Tip of dorsal arms 2 Tip of dorso-lateral arms 2 Ventro-lateral arms 62 + Tentacles 4 Within pallial chamber: Anal 2 Branchial 2 Abdominal 4 Posterior extremity of body J2 Total 90 No cephalopod heretofore described is at all comparable to this creature excepting the wonderful Thaumatolampas diadema Chun, already mentioned, and the Lycoteuthis jattai of Pfeffer ('oo, p. 161), both of which are now thought by the latter author to be based upon the same species ('08, p. 294). Chun, in the course of his amazing description of the luminous organs of T. diadema, says, "Unter allem, was uns die Tiefseetiere an wunder- vollen Farbungen darbieten, lasst sich nichts auch nur annahernd mit dem fast magischen Kolorit dieser Organe vergleichen. . . . Es war eine Pracht!" But to judge merely from the anatomy of Nematolampas, even Thaumatolampas must be outmarveled in life by this wonderful mollusk from the Kermadecs. 212 S. ST1LLMAN BERRY. LITERATURE CITED. Chun, Carl. '03 Aus den Ticten des Weltmeeres. Zweite Auflage. Large 8vo, Jena. '10 Dei Cephalopoden. I. Teil: Oegopsida. Wiss. Ergebn. deutsch. Tiefsee- Exped. Valdivia, Bd. 18, pp. 1—402, 2 pis. and 32 figs, in text. Atlas of 61 pis. Pfeffer, Georg. 'oo Synopsis der oegopsiden Cephalopoden. Mitteil. naturhist. Mus. Ham- burg, XVII., pp. 147-198. '08 Teuthologische Bemerkungen. Mitteil. naturhist. Mus. Hamburg, XXV.. pp. 289-295. Vol. XXV. September, 1913. No. 4. BIOLOGICAL BULLETIN A STUDY OF THE EFFECTS OF INJURY UPON THE FERTILIZING POWER OF SPERM. NEIL S. DUNGAY. CONTENTS. Page. I. Introduction 213 II. Materials and general methods 214 A. Occurrence 214 B. Collecting 214 C. Removal of germ cells 215 D. Culture methods 215 E. Controls 216 F. Treatment of sperm cells 216 G. Cytological methods 217 III. Outline of development of Nereis eggs 218 IV. Experiments 220 A . Upon Nereis 220 1 . Heat 220 2. Delay 227 3. Fresh water 229 4. Alcohol 230 5. Sodium hydroxide 231 6. Hydrochloric acid 232 7- Cold 233 B. Upon Arbacia 233 V. Cytological stud y 236 A . Normal .• 236 B. Experimental 237 VI. Discussion 244 A, The production of defectives 244 B. Fertilization 252 VII. Summary 254 I. INTRODUCTION. This paper deals with the results obtained by fertilizing normal eggs with spermatozoa injured by various methods. The results have a bearing upon some of the problems of fertilization. 213 214 NEIL S. DUNGAY. They also throw light upon the theories dealing with the produc- tion of defectives by the action of injurious agencies upon the germ cells previous to the time of fertilization. The experimental work was done at the Marine Biological Laboratory at Woods Hole, Mass., during the summers of 1911 and 1912. The results given are based upon observations made upon more than 600 cultures of living Nereis eggs and 200 cul- tures of Arbacia eggs. Samples from the various cultures were preserved for later cytological study, the results of which are also reported in this paper. II. MATERIALS AND GENERAL METHODS. A. Occurrence. In the evening from the time of darkening until about ten o'clock, sexually mature males and females of Nereis limbata swim freely at the surface of the Eel Pond at Woods Hole. Although they are most easily collected during the two weeks following the full moon, it is possible to find a few nearly every evening of the lunar month. The breeding habits have recently been described (Lillie and Just, '13) in a very interesting paper which gives in considerable detail the methods of catching and keeping the worms. B. Collecting. For my experiments the animals were caught in nets by lantern light, placed in separate finger bowls of sea water, and taken to the laboratory buildings at about 9:30 P.M. The worms were then washed in running sea water and the finger bowls were cleansed and refilled, covered with glass plates and placed in running water to keep cool. Early the next morning the accumulations of mucus were removed and the water was again changed. When carefully cared for in this manner, very few of the worms shed their germ cells during the night. The males practically never shed their sperm cells under these condi- tions. During one month in which a record was kept, over 90 per cent, of the females captured retained their eggs. By this method I have sometimes kept the females without shedding for as long as 60 hours, though, for experimental purposes, the eggs EFFECTS OF INJURY UPON SPERM. 215 which were retained so long were not used. The experimental work was usually started as early in the morning as possible so as to use the material in a fresh condition. A few experiments were carried out shortly after removing the animals from the Eel Pond. Since no better results were secured by working at night, the material was usually kept until morning when condi- tions for work were better. C. Removal of Germ Cells. The abundant eggs and sperm were removed by snipping the sides of the body wrall at intervals, with scissors. If no water is used the sperms push out as a white mass nearly free from liquid and the eggs may be secured as a thick greenish mass. As a rule, however, the operation was conducted under water. In all cases the usual precautions were taken to avoid the introduc- tion of stray sperm cells. The worms were thoroughly washed in a jet of salt water and, in the earlier experiments, in boiled salt water as well, a precaution which I afterwards found to be unnecessary since the sea water from the laboratory supply was repeatedly tested and found to contain no active Nereis sperm. The bodies of the animals used were quickly removed from the dishes of sperm and eggs so as to avoid, as far as possible, any contamination from body fluids, and every precaution was taken to avoid abnormalities due to outside influences such as strong sunlight, bacterial action, mechanical agitation, etc. D. Culture Methods. An effort was made to use a minimal quantity of sperm for the process of insemination, in order to avoid, as far as possible, errors due to polyspermy. In the case of the injured sperm, a large quantity was often necessary to secure a small percentage of fertilized eggs. Usually a small number of eggs was insemi- nated first in a small quantity of water to which a liberal supply of water was soon added. In this manner I was able to secure a sufficient number of fertilized eggs to make preservation possible. Up to the time when the larvae begin to swim, the water was changed several times. After this period it is very difficult to renew the water without losing many of the larva?. An effort 21 6 NEIL S. DUNGAY. was made to keep from crowding the developing eggs since this seems to be a source of abnormality. Any error which might creep in here was checked by the controls since they always contained at least as many eggs as were kept in the experimental cultures. E. Controls. Along with each experiment two or more control cultures were kept. One control consisted of a dish of sea water contain- ing a part of the eggs which were not inseminated. The other contained about the same number of eggs as was kept in each of the experimental dishes, inseminated with injured sperm; these were as closely followed during the period of observation as were the experimental cultures. In all experiments in which the sperm cells were subjected to the action of reagents it is impossible to avoid carrying over a trace of the reagent used, to the dishes of eggs. Control experiments were made in such cases by adding to the eggs at the time of insemination, as much of the reagent as was introduced into the experimental cultures. In nearly all cases the eggs and sperm used in the controls came from the same animals which furnished the materials for the experiments. F. Treatment of Sperm Cells. Attempts to injure the sperm cells as much as possible without preventing their attachment to the eggs proved unsuccessful in many cases. Either the sperm were killed or they were not sufficiently affected to give striking results. With improvement of the technique consistent results were secured, although these results indicated a great variation in the intensity of the injuries to the sperm cells, thus rendering it impossible to repeat an ex- periment with a complete duplication of results. The sperm cells were injured in various manners. Inasmuch as some methods gave promise of easier and more striking results or seemed less open to criticism, but few experiments were made along other lines in order to give more time to follow out the more desirable methods. Injury was attempted by soaking the sperm cells in distilled water or in diluted sea water ; by treating them with weak alcohol; by exposing them to the action of dilute acids or alkalis; by freezing; by heating; and EFFECTS OF INJURY UPON SPERM. 217 by keeping them so long that they were nearly dead. Since it proved to be so difficult to injure the sperm without destroying their motility it seemed to be desirable to try the effects of the X-rays or of the radium emanations; but as the proper appliances were not procurable at the time, I was unable to carry out this part of the work. An attempt was made to get results by using sunlight but the sperm cells gather together in such a way that they protect each other. No attempt was made to carry this method far although it might prove useful with proper technique. G. Cytological Methods. For later cytological study of the early stages, a large number of samples were taken from the experimental cultures and fixed in Meves's modification of Flemming's fluid, made as follows: 3.5 c.c. osmic acid, 2 per cent.; 15 c.c. chromic acid, 0.5 per cent.; 3 drops glacial acetic acid. The eggs were exposed to the action of this solution for about 40 minutes. Iron haematoxy- lin was used as a stain. The larval stages were preserved either in Gilson's mercuro-nitrate fixing fluid or in Bouin's mixture of formalin and picric acid. The fact should be mentioned at this point that it was found to be impossible for one person to make as careful observations and notes upon the living material, when preserving a series for later study, as were made when doing nothing else except following the developing eggs under the microscope. As a consequence the preserved samples are of two kinds, viz.: (i) a long close series with as much recorded data as the time allowed, and (2) irregular preservations made during extended periods of observation. Usually in getting a long series fixed it was neces- sary, especially in the earlier stages, to run at least two series at one time, since it is impossible to tell with certainty just how much the sperm cells used for inseminating have been injured. Again it should be noted that in the preservation of a long series, those samples first fixed often contain a greater proportion of the more normal eggs than those preserved later. This is due to the fact that the most normal sperm cells seem to initiate development earlier than do the others. Since the eggs which have formed jelly are more readily picked up by the pipette, 2l8 NEIL S. DUNGAY. they will tend to be taken from the cultures in greater proportions in the earlier part of the preserved series. An effort was made to overcome this by using several parallel cultures in different dishes but without entire success since the number of eggs from a single female is limited and since it does not seem wise to use eggs from two or more females, in the same experiment. III. OUTLINE OF DEVELOPMENT OF NEREIS EGGS. Although the maturation, fertilization, and development of the egg of Nereis have been described in more or less detail by others (e. g., Wilson, '92, Lillie, F. R., 'n, and '12) it seems desir- able to review' a few of the principal points at this time. The eggs, which vary considerably in size, are about 100 microns in diameter. Their irregular shape at the time of laying is soon lost and they take the form of a sphere slightly compressed in a polar direction. They are of a greenish color and moderately transparent. A polar view, which is the one usually obtained, shows the large germinal vesicle in the center of the egg and an imperfect double belt of la'rge oil drops, numbering approxi- mately from ten to twenty. Just beneath the vitelline membrane is a radially striated cortical layer which breaks down imme- diately following insemination, throwing off a thick envelope of jelly and leaving a conspicuous perivitelline space. The jelly is readily demonstrated by the use of India ink ground up in sea water, although it is nearly invisible in pure sea water. Since the unfertilized eggs normally lie in contact at the bottom of a dish and, upon insemination, become separated by the jelly, it is easy for the naked eye to determine whether or not a large part of the eggs have started to develop. A microscope, how- ever, is necessary to determine the condition of any individual egg. As a general rule unfertilized eggs do not form jelly upon standing undisturbed, at least not for several hours. Yet close examination usually reveals a mere trace of eggs which have formed this secretion. Inasmuch as strong mechanical agitation will induce jelly formation this is perhaps due to violent contact with the scissors upon opening the female, or possibly there may be a few eggs which are extraordinarily sensitive to stimulation. The latter view is rendered probable by the fact that the pro- EFFECTS OF INJURY UPON SPERM. 219 portion of eggs with jelly varies considerably among different lots, so much so that an occasional batch of eggs is found in which the majority form jelly. This jelly formation is probably not due to accidental insemination since fertilization cones are not seen and cleavage does not result. Yet to guard against possible misinterpretation an unfertilized control was kept for each experiment and examined along wTith the experimental materials. The active spermatozoon may be seen to attach itself to the egg. It immediately becomes quiescent and remains motionless until it is drawn into the egg. The outflow of jelly, which con- tinues for some fifteen minutes after insemination, now sweeps away all superfluous sperm cells, which have in the meantime lost their power of motion. Just beneath the attached sperm cell an attraction cone is pushed up by the egg cytoplasm until it unites with the membrane. It then gradually disappears, carrying with it the membrane so as to form a depression in which the sperm is now concealed. This all happens during the first 25 minutes after insemination. In the meantime the germinal vesicle has disappeared and the egg has assumed an irregular shape. About 45 minutes after insemination the egg resumes its spherical form and soon the sperm, after again coming prominently into viewr for a short time, is drawn into the egg. Shortly before the close of the first hour succeeding insemination the first polar body appears and is followed in a few minutes by the second. The polar bodies are easily seen at this time against the now yolk-free region of the pole. In a profile view they are very conspicuous. In the course of the second hour the first two planes of cleavage appear. The first plane, as has been stated by Just ('12), passes through the point of entrance of the sperm, dividing the egg unequally. The development proceeds rapidly from this point. Movement is often seen in twelve hours and fine trochophores are formed by the end of the first day. In the second day the animal begins to elongate and the larval setae grow out to a relatively great length. Naturally the time element, as given above, varies greatly with the tem- perature. 220 NEIL S. DUNGAY. IV. EXPERIMENTS. The sexual products of both Nereis and Arbacia were utilized for experimental purposes. However at the time when I could use Arbacia, the control cultures did not run as well as might be desired and so Nereis material was used for the greater part of the work. A. Upon Nereis. In the earlier experiments upon Nereis I subjected the sperm cells for a certain time to the action of graded strengths of various reagents. Later I found that the most satisfactory procedure is to keep the sperm cells, after removal from the body of the male, until they are nearly dead or to keep them for a few minutes at a temperature of about 44° C. The latter method seems to be less open to criticism since no substances of a possibly toxic action are transferred to the egg cultures by the process of insemination. All of the methods of experimentation which were used give essentially similar results, indicating that the action upon the sperm cells is probably not specific in any case. It seems rather that there is produced a general decrease in vitality which may manifest itself by a retarded development, by a high mortality or by the production of forms wrhich are more or less abnormal in structure or behavior. i. Heat. — Perhaps the most uniform results in the whole series of experiments which I have carried out were obtained through the exposure of the sperm cells to a certain degree of heat. This was accomplished by mixing the sperm cells with sea water, which was then placed in the bottom of a test tube, care being taken to prevent any of the liquid from touching the walls of the upper part of the tube. A thermometer was then introduced into the bottom of the test tube and was used to stir the liquid and to take the temperature readings every minute. The test tube was nearly immersed in a large beaker of warm water. The temperature was maintained in the bath thus formed, by means of an alcohol lamp or by adding to the bath warmer water from another container. In this manner it is possible to keep the temperature under reasonable control. EFFECTS OF INJURY UPON SPERM. 221 Particular care was taken to insure the perfect cleanliness of all pipettes. None of the liquid of the test tube was allowed to splash upon the sides of the tube so as to be above the level of the water in the bath. A large number of dishes of eggs were made ready and one or more of these were inseminated every minute until either all the sperms had been used up or the dishes of eggs had all been inseminated. In this way there were usually obtained one or more dishes of eggs which were inseminated at the critical time which occurred from 7 to 20 minutes after the sperm cells were placed in the warm bath. The variation in time is probably due to fluctuations in the temperature of the bath, small though they may be. This is borne out by some experiments in which slightly higher temperatures were used and in which the critical time came much earlier, so early in fact that it was difficult to catch the sperm cells at just the proper stage for use. Since it proved to be very difficult to determine in advance which culture would be the best for study, the preservation of a long series from the best -cultures is largely a matter of chance. But by preserving a large number of series for cytological examination, a few have been obtained which are satisfactory for study. The others, although they show essentially the same features, contain so many eggs which develop normally that the study of the affected eggs is very time consuming. In all, 23 experiments, involving several hundreds of cultures, were made by this method. Since the earlier work was carried out in a different manner, which gave me records of the course of development of the experimental cultures but no continuous series of preserved samples, the primary purpose of this set of experiments was to obtain material for cytological work. Ac- cordingly the records of these experiments are not as full as might be desired in the case of any one experiment. To illustrate the method of procedure the record of experiment 65 of July 4, 1912 is given. The eggs were removed from the female at 1:39 P.M. Sperm cells removed from male at 1:40 P.M. Sperm placed in bath at 1:47 P.M. Temperature of bath at end of successive minutes (read from centigrade ther- mometer in test tube) 43°; 44°; 43.4°; 43°; 44°; 44°; 43.8°; 43.3°; 222 NEIL S. DUXGAY. 43°; 43-7°; 43°; 43-7°; 43°: no later records. Samples of eggs from one female, numbered 65.1, 65.2, 65.3, 65.4, and 65.5, were inseminated with heated sperm at the end of n, 12, 13, 15, and 16 minutes respectively. In the first four about 95 per cent, of the eggs formed jelly immediately. Sample 65.5 gave about 80 per cent, with jelly. All of the above were discontinued. 65.6 was inseminated at 2:04 P.M. with sperm which had been heated for 17 minutes. Jelly formation began at once but took place in a gradual manner, some eggs failing to give off jelly until the end of half an hour. Eventually about 60 per cent, formed jelly. Of the eggs with jelly 50 per cent, failed to segment, 15 per cent died before gastrulation, 25 per cent, formed trochophore larvae later than did the controls, and 10 per cent, behaved in a seem- ingly normal fashion. Discontinued before the formation of the larval seta? because all had been preserved. Control of infertile eggs showed only a trace of jelly formation after 18 hours. The inseminated control developed normally. A series of preserva- tions from the experimental culture 65.6 was made at 10 minute intervals, beginning at 2:24 P.M. Experiment 63, July 3, 1912, was conducted in a similar manner and furnishes a somewhat fuller record of the later stages. 63.5 was inseminated at 10:36 with sperm which had been kept in a warm bath (about 44° C.) for 14 minutes. Over 70 per cent, formed jelly. First cleavage was observed at 12:15 P.M., 7 minutes later than in the case of the control series. At 1 :25 P.M. the control culture was nearly two cleavages ahead of the experimental culture. About 5 per cent, of the eggs in the experimental culture failed to segment after forming jelly. At 8:45 A.M. on the following day the experimental culture was very much less active than the control. About 5 per cent, had died in cleavage stages. 49 hours after insemination the control animals were in excellent condition and were sending out the second set of larval setee. The experimental culture showed clearly that the injury to the sperm cells may cause abnormalities in those eggs which seem to behave normally in the earlier stages. 50 per cent, of the developing eggs had not gone beyond a troch- ophore stage. Most of these died soon, though a few continued to live for at least another day without any change except an EFFECTS OF INJURY UPON SPERM. 223 increase of size. This growth is probably due to an increase in water content and gives the animals a dropsical appearance. 25 per cent, were as well developed as the control animals. The remaining 25 per cent, were in various stages from the trocho- phore to the formation of the second set of larval setae. Some of them seemed to be normal in every way except that their development was retarded. Others were less active. A small number, perhaps 5 per cent., seemed to be as far along as the controls but had no setae or were deficient in some other way. Most of these died within a few hours. The controls developed in a normal manner. The uninseminated controls showed no development beyond the formation of jelly in about 0.3 per cent. In all the cultures of the above experiment the conditions were made as good as possible by changing the water, removing dead eggs, etc. Without attempting to give other definite records, I may summarize briefly the different types of effect observed in the series of experiments using heat as a means of altering the sperm cells. (a) In some cases all of the jelly did not form and the peripheral alveoli of the egg were not entirely emptied. This occurred in from a trace to nearly 10 per cent, of the eggs in the experimental cultures. Some eggs formed a mere trace of jelly, some formed about half of the normal quantity, and some had nearly as much as normal. The different cases observed gave the impression that a larger stimulus caused a greater jelly formation. It seemed as though a sperm might in some way give only enough impulse to cause a part of the alveoli to be emptied, possibly due to a brief attachment to the egg. However no such egg was ever followed long enough to determine the cause of this partial jelly formation. (b) After the jelly is completely extruded and the polar bodies have been formed, cells are commonly found in which no further development takes place. A fewr eggs do this in many unin- seminated controls but there is no comparison in the frequency with which they may be found. In some experiments special care was taken to avoid mechanical agitation since violent mechanical stimuli may produce this effect. Yet the same 224 NEIL S. DUNGAY. results were found, indicating that maturation and jelly formation had been initiated by the sperm in some way, although later development failed to take place. In some experiments this has been seen in more than half of the eggs which formed jelly. All experiments which gave any indication of abnormality con- tained at least a few of this type of eggs. In those, cultures in which these eggs were present in considerable numbers I failed to see as many fertilization cones as usual. This suggests either that the eggs were infertile, due to the failure of the sperm to enter, or that the development was carried this far in response to some unknown stimulus given at the time of insemination. Cytological examination is necessary to determine this point. (c) Cytological examination of stained sections showed cases in which polar bodies appeared without jelly formation. No indication of this was seen in the living material as it was not supposed that it was possible. It would, of course, be difficult to find, even if such cases were known to occur. (J) Development sometimes stopped after the first cleavage or more rarely after the second. This occurred in less than 2 per cent, usually. It was usually associated with (e) an unsuccessful attempt to form a first cleavage plane. The behavior of such cases will be described under another heading in more detail. (/) In some cases, as many as 0.5 per cent, in one experiment, the first cleavage plane divided the cell into two equal parts. Such cases did not usually divide again. (g) In a very few egg cells the cleavage was uneven, forming 3, 5, or 7 cell stages which usually died without developing far. It is well known that uninseminated eggs may undergo a kind of pseudo-cleavage if allowed to stand for some hours but the cases mentioned here are not to be confused with this for several reasons. Their appearance is different. These forms appear much earlier than they ever do in the unfertile eggs. The control of unfertilized eggs never showed such pictures until much later and then they were of a different appearance. (7z) Other irregular forms of cleavage were seen, but very rarely. Since the later cleavage stages are more complex in appearance it is more difficult to find irregularities and they EFFECTS OF INJURY UPON SPERM. 225 may be of more frequent appearance than my notes indicate. Again it may be possible that such cases are sometimes present in eggs fertilized with normal sperm. I have no reason to believe that my control cultures contained such cases. (i) The rate of development in the experimental cultures often proved to be very irregular and was usually slower than that of the controls. For example, when the controls first have definite setse the experimental cultures have none. Segments are marked out earlier in the controls. Early cleavage planes are often delayed, but this may be due in some part to the failure of the weakened sperm to reach the egg as soon as it should. But it seems certain that, even allowing for this fact, the early stages are somewhat retarded in development. (j) In practically every experiment the controls lived longer than the experimental cultures and the death rate in them was not so much as one tenth as large. Although an effort was made to keep the controls under exactly the same conditions as the others there were usually fewer larvae in the experimental dishes (presumably a more favorable condition). Old eggs and organic materials were removed so as to eliminate any possible toxic action of products of decay. In some cases the water was changed from day to day but with little effect. It seems that the normal animals in the controls are able to withstand the artificial conditions incident to life in the laboratory for a much longer time than those in the experimental cultures, which are so weakened by the injury to the part which comes from the male parent, that they cannot exist under such conditions as are provided. (k) The trochophore stage seems to be a difficult one to pass. It frequently happens that this stage is permanent. In some cases at about the time when segments should be marked out and setae should appear, the animal becomes irregular in form or swells up and becomes transparent, due probably to the absorp- tion of water. In some cases the trochophore remains in its original condition without change. In most cases death takes place soon after elongation should take place but I have often kept the permanent trochophores for a day or more after they should have elongated. 226 NEIL S. DUNGAY. (/) When the larval setae should form, it often happens that none or only a few of them appear. In some cases this was true of more than half of the larvae. Very few experimental cultures showed less than 5 per cent, of such forms. There may be all grades of conditions from no setae, through a few scattering ones, weak ones, and abnormally directed ones, to nearly perfect forms. In cultures inseminated with normal sperm, very adverse external conditions may produce this effect or something, in a small number of the larvae, very much like it. But under the very best conditions that could be maintained this was always found in the experimental material, though never in the controls. Associated with this condition was the partial or complete lack of sensory appendages. (m) The body form is sometimes altered. It may be perma- nently bent toward one side by the asymmetrical development of a single segment. Often the body is covered with small wart like projections, giving it a peculiar roughened appearance. (n) The experimental cultures nearly always show a lack of vitality, as is evidenced by a lessened activity. The normal cultures are very active after reaching the motile stages. The others, at least in part, are usually more or less quiescent. The phototropic responses are much less uniform in the experimental cultures and do not take place so quickly. (0) In addition to the above list of abnormalities there is always present, unless the unfertile eggs are removed from the cultures, a number of other types which are confusing at first. As it is not always possible to remove every unfertilized egg during the first few hours, one is likely to find a variety of things which are easily misinterpreted unless they are very carefully studied. Unless such things have crept into my counts through an occasional oversight no consideration will be given them. Nearly all of the above list of abnormalities were present in most of the experimental cultures if they were not discontinued too early. The controls rarely produced any of them. The proportions of the different types naturally differed widely in the various experiments. If the sperm cells were injured to the limit, or nearly so, the abnormalities were largely confined to the early stages and death occurred before the later ones could ap- EFFECTS OF INJURY UPON SPERM. 227 pear, though a few individuals nearly always succeeded in reach- ing the later stages. If the sperm cells were not injured so much, the earlier stages were comparatively free from abnormalities and those of the later stages appeared, chiefly the permanent trochophores, the malformed bodies and the lack of setae. In these experiments, as in those to be described later, the number of eggs fertilized is inversely proportional to the strength and duration of action of the injurious agent, provided the pro- portions of eggs and sperm are constant. The degree of abnor- mality and the per cent, of abnormal forms is directly propor- tional to the strength and duration of action of the agent used. These statements, of course, are not to be taken as more than an approximation. No effort has been made to reduce the results to the form of an exact lawr. However the results of the series are consistent and conform nicely to the above generalizations. 2. Delay. — Striking results were obtained by keeping the sperm cells for some time before they were used for insemination. Naturally the extreme time which may elapse before the sperm cells become incapable of influencing the eggs varies greatly with the temperature, with bacterial conditions, and, I suspect, with the general vitality of the cells. The complexity of the conditions makes it extremely difficult to completely duplicate the results of any particular experiment. Oftentimes it happens that the results in experimental cultures are nearly normal and again development may not appear at all. Yet the results all seem to be consistent, the differences being in degree only. As in the set of experiments described under the head of heat, the abnormal types are undoubtedly due to certain degrees of injury to the sperm cells. The possibility of the results being due to toxic substances produced by bacterial action or sperm metab- olism during the period of delay, and introduced into the egg cultures at the time of insemination, seems to be rather remote. In order to minimize any such effect, the water was changed on the egg cultures several times shortly after insemination. One of the several extreme cases of effect upon early develop- ment is found in experiment 34 of July 15, 1911. The uninsemi- nated control remained unchanged. The fertilized control was inseminated at 2:11 P.M. At 3:15 P.M. the first cleavage was 228 NEIL S. DUNGAY. practically completed and many eggs were dividing the second time. At 3:25 the majority of the eggs were in a 4 cell stage. At 9:00 A.M. of the following day all were normal trochophores. The second day brought forth setae and succeeding days showed only a normal development. 34.1 consisted of fresh eggs and a large quantity of sperm which had been kept in a corked vial for 21 hours at room temperature. The sperm and eggs were mixed at 1:35 P.M. Less than 5 per cent, of the eggs formed jelly. At 2:59 P.M. there were many attempts to form a two cell stage. It will be noticed that this is over 20 minutes longer after insemination than the time at which the controls were nearly through with the first cleavage. Very few normal 2 cell stages resulted. Usually the constriction appeared in the proper location but a few cases were observed in which the attempted plane of cleavage would have divided the cell into equal parts, had it been completed. In the great majority of cases the con- striction lasted for a few minutes and then began slowly to dis- appear, the cell becoming irregular in outline. For example one cell was observed to be in the early process of first cleavage at 3:18 P.M. Slight progress was made for a minute and then no change took place until 3:23 P.M. when the furrow began to disappear. At 3:37 P.M. the furrow was invisible and there was no trace of the attempted cleavage. Irregularities in cleav- age were also seen. At 3:55 P.M. a 5-cell stage was noted. A 4-cell stage with two very large cells diametrically opposite to each other was observed at the same time. Another distinct 5-cell stage was seen. Another 4-cell stage, as seen from a polar view, had cells of equal size, the polar bodies appearing to be normal. At 3:40 P.M. and following, another case of at- tempted cleavage, similar to the one described above, was seen. At 3:44 P.M. a 3-cell stage was found. Many one cell stages which had formed jelly and thrown off polar bodies were now evident. It was not determined whether many of these had previously attempted to divide or not. Probably most of them had not. At 3:45 P.M. a y-cell stage was seen. At 9:00 A.M. of the following day most of the eggs with jelly were still in a one cell stage or in irregular early cleavage. Much of this irregular early cleavage is probably not a true cleavage but a EFFECTS OF INJURY UPON SPERM. 22Q budding. About 4 per cent, of the eggs were in a 2-cell stage of apparently normal form. A few trochophores, not at all active, were revolving feebly with a lateral displacement of perhaps twice their own diameter. They were abnormal in shape and appearance and though no cilia could be seen they were certainly present since the animals were motile. The whole culture died without developing farther. Other experiments gave similar results or produced fewer abnormalities in early development and lived to produce malformed larvae. Since on one day sperm which had been kept for 30 hours would cause most of the eggs in a culture to develop and upon the following day sperm only half as old might have no effect it seemed to be desirable to produce more uniform conditions for keeping the sperm. Accordingly the sperm cells were placed in an ordinary refrigerator in which the comparatively low tem- perature did not fluctuate so much as did that of the outer air. The best results were secured by using the sperm on the third day. By testing out the contents of a vial at intervals during the day it is possible to determine the proper time for starting an experiment. About 40 cultures obtained in this way have been under observation and have given reasonably uniform re- sults. All of the types of abnormal development described above under the head of heat were repeatedly found in this set of experiments. The results given by this set of experiments are very convincing to one who was able to follow the material closely. It is not known definitely whether the injury to the sperm was produced by toxic bacterial products, by the accumu- lation of metabolic products, or by the gradual weakening of the cell by its own metabolism. Whatever the cause, I have no doubt that the effects produced were brought about by insemina- tion with weakened sperm cells. j. Fresh Water. — Three experiments were made in which the sperm cells were placed in dilute sea water or in distilled water for short periods of time. Sperm from mixtures containing more than 20 per cent, of sea water produced normal development. The results produced by lower percentages of sea water and by distilled water were variable, some normal larvae appearing in cultures in which most of the eggs failed to segment. Permanent 230 NEIL S. DUNGAY. one and two cell stages were secured in large numbers. Attempts at first cleavage were also seen. Although so few experiments were made by this method that we do not get all the types of abnormality which were produced by the heat and delay, we have a clear indication that more work and better technique would produce very similar results. The work was dropped in order to take up more promising methods. 4. Alcohol. — In experiment 29 of June 29, 1911, the sperm cells were kept in a series of weak alcohols for 35 minutes. The un- fertilized control showed no jelly and no segmentation. The fertilized control was normal. Sperm which had been exposed to the action of 10 per cent, alcohol (made up with sea water) were incapable of producing any effect upon the eggs. 8 per cent, alcohol gave similar results. A few eggs, about 2 per cent., were fertilized by sperm cells which had been kept in 6 per cent, alcohol. Of the 15 eggs which were observed to form jelly, 2 did not segment, one divided only once, and the remainder were motile on the second day. Of the motile forms one half were abnormal in some way upon the fourth day. During the course of the whole experiment this culture was greatly retarded in development. Sperm from the 4 per cent, alcohol fertilized 25 per cent, of the eggs and, with the exception of a few unseg- mented eggs, all developed the power of motion upon the second day, though they did not develop so rapidly as did the controls. On the third day at least 3 per cent, were visibly abnormal in body form or in setse. The fourth day showed 10 per cent, to be abnormal. Some never formed any setse, some formed weak ones, and some were nearly perfect in this respect. Irregular bunches were often present upon the sides of the segments. The apparent age of the animals in the culture varied greatly. 2 per cent, alcohol did not prove to be so effective, for 60 per cent, of the eggs were fertilized and developed with fewer abnormalities than any of the cases given above. Other experiments show that sperm cells exposed to weak grades of alchool for a longer time give about the same results as those exposed to higher strengths for shorter periods. Early segmentation stages do not appear to be so much affected as the larval stages, though usually there were at least a few eggs which EFFECTS OF INJURY UPON SPERM. 231 failed to segment after jelly formation and maturation or which failed to segment more than once. There was in all cases a strong tendency for the experimental cultures to die earlier, to be less active, and to contain forms which were abnormal in body form or in setae formation. Permanent trochophores were also common. In this set seven experiments were completed, involving in the neighborhood of a hundred cultures. Several slightly different methods of procedure were used but the results in all cases were similar to those cited above. An extra control was carried in each experiment. This consisted of eggs and normal sperm to which was added at the time of insemination as much alcohol as was transferred to the experimental cultures along with the sperm used in insemination. In no case did this small quantity of alcohol cause this control to develop in an abnormal manner. This indicates that the abnormalities seen in the experimental cultures were not due to the presence of such minute quantities of alcohol in the egg culture. 5. Sodium Hydroxide. — In this group but few experiments were made. The sperm cells were exposed to the influence of different strengths of NaOH for a period of 20 minutes. The solutions used were made up by adding to 10 c.c. of sea water a certain number of drops of 0.75 per cent. NaOH solution. After exposure to one of these mixtures for 20 minutes the sperm cells were used for inseminating fresh eggs. Three controls were carried. The first, eggs only, showed no change. The second, eggs and normal sperm, developed in the usual manner. To the third, similar in composition to the second, was added the same amount of NaOH solution that was carried over to the eggs by the sperm in the experiment 40.4. This likewise developed in normal fashion, thus demonstrating that the trace of alkali necessarily introduced in the process of insemination of the experimental series was without any apparent effect upon early development. In table I. the results of a typical experiment are given. The early history of the experiment was not followed on account of a lack of time. The observed abnormalities were chiefly in the failure of the trochophores to develop, in which 232 NEIL S. DUNGAY. case irregular bunches appeared on the body which became enlarged and transparent. This appearance was not due to post mortem changes, for the animals sometimes continued to swim about slowly. Many times the setae failed to appear as has been described in other cases. In the case of 40.5 and 40.6 the smaller abnormality is probably due to the greater mortality, since the abnormal individuals are often not very resistant. The general impression given by this series is similar to that formed from the observations on the alcohol series. TABLE I. EXPERIMENT 40. July 21, 1911. All inseminated at 9:30 A.M. No. Drops NaOH. Per Cent. Jelly at 9:55 A.M. Time of First Cleavage. Per Cent. Segment. Later History. 40.1 I IOO — 10:38 A.M. IOO — Normal. 40.2 2 100 — 10:39 A.M. 90 5 per cent, abnormal in setae or in having permanent trochophore. 40-3 3 95 10:39 A.M. 90 About like 40.2. More weak setae. 40.4 4 60 10:43 A.M. 40 About like 40.3. and scattering 40.S 5 40 10:41 A.M. 15 Development slower. Many and scattering die. Living less abnormal than 40.4. 40.6 6 10 11:00 A.M. 5 Like 40.5 but more abnormal and scattering and more die. 6. Hydrochloric Acid. — The effects of acid upon the sperm cells were also observed, the experiments being conducted in the same manner as the alkali series. A number of drops of 9 per cent. HC1 were added to 10 c.c. of sea water and the sperm were placed in this solution for 25 minutes. All three of the controls were perfectly normal. Eggs inseminated by the sperm suspen- sion containing one drop of acid often stopped development in late cleavage. Others formed irregularly shaped permanent trocho- phores. Later over 50 per cent, of the larvae were abnormal in body form and character of the setae. The eggs inseminated by the sperm suspension containing two drops of acid frequently failed to cleave. A few permanent two-cell stages were produced. Some of the abnormal larval stages again appeared but not so large a proportion as in the previous case. More than two drops EFFECTS OF INJURY UPON SPERM. 233 of acid affected the sperm cells so much that the eggs inseminated by the sperm cells so treated were unchanged or merely formed jelly and polar bodies without segmenting. An attempt was made with weak NaOH solution to neutralize the acid in which the sperm cells were placed before insemination but this process so often destroyed the life of the sperm cells that it was useless for control purposes. However, the controls used seemed to be sufficient to show that the cause of the altered development does not lie in the presence of the acid in the experi- mental cultures of eggs. /. Cold. — Freezing the sperm was effected by placing them in sea water in a test tube and imbedding the tube in a mixture of powdered ice and salt. - 8° C. killed the sperm. - 2.2° C. gave a mush}' ice in the test tube. The sperm cells were not killed but caused a normal cleavage in the eggs. From 5 to 10 per cent, of the larvae were abnormal. A very few scattering abnormalities of all the kinds previously mentioned were ob- served. 1.6° C. did not seem to affect the sperm cells enough to alter the egg development. Since low temperatures are very difficult to handle by means of the crude methods employed extensive experiments were not carried out along this line. Refrigeration at a temperature considerably above freezing for longer periods was also used as a possible source of injury to the sperm cells. Since the effects produced are probably due as much to the delay as to the cold these experiments have been mentioned under a previous heading. B. Upon Arbacia. In July, 1911, a series of 18 experiments involving about 200 cultures was carried out, using the sexual products of the common sea urchin, Arbacia pimctulata. Along with some other investi- gators I had considerable trouble at that particular time in getting control experiments to develop in a normal manner. Accordingly the work wras not carried as far as it seemed desir- able at that time. Although the results were perfectly convinc- ing to one who could see the living material, the small percentage of abnormality in the controls might cause some to question the value of the published data. I will, however, give a brief survey 234 NEIL S. DUNGAY. of the work since it is in line with the work upon Nereis and since some of the features here described have not been previously recorded. Experiments were carried out by inseminating eggs with sperm which had been subjected to injurious conditions for a longer or shorter time. The injurious agents used were (a) sea water concentrated by boiling; (b) sea water diluted with distilled water; (c) sea water with from one to twelve per cent, of alcohol added; (d) sea water slightly acidified with HC1; (e) keeping the sperm cells from 6 to 24 hours after removal from the body; and (/) combinations of the above. In a general way these methods of treatment did not give specific effects but produced very similar results, although the intensity of the effects varied as was to be expected. In one case the controls were nearly as abnormal as the experimental cultures. But, taking the experiments as a whole, there can be no doubt as to the great differences to be found between the controls and the experimental series. The abnormalities produced by inseminating normal eggs with injured sperm may be briefly reviewed. In a few cases in which the sperm cells were exposed to extreme conditions, insemination gave a number of cases of membrane formation without develop- ment. Cleavage was found to be extremely irregular in some cases, producing fantastic forms, such as ciliated plates and rows of cells. The more normal cleavage was also found to be ir- regular but it is difficult to say that it was more irregular than it sometimes is in normal cultures. The development of the experi- mental series was often found to be retarded, especially after late cleavage. The early cleavage was not affected so much, though occasional indications of a slight retardation were observed. Curiously enough, in some of the experiments using alcohol, the experimental material was observed to be one or two cleavages ahead of the controls in the early segmentation stages, though the difference was soon in the other direction, if any was apparent. Prismatic forms were often observed in the controls when the experimental cultures were chiefly in the blastula stage. The death rate was much higher in all of the experimental series, the whole culture sometimes dying in various EFFECTS OF INJURY UPON SPERM. 235 stages from blastula to pluteus before any considerable number of deaths were observed in the controls. A very small number died during cleavage in stages beginning with the undivided egg. Blastulse whose cavities were more or less filled with cells, similar to the stereoblastulse so often described by experimental workers, were found in increasingly large percentages as the strength or duration of the injurious agent acting upon the sperm was in- creased. In some cases over 15 per cent, of the experimental cultures were of this form. They did not appear to develop farther but went to pieces. Evaginate gastrulse were formed rather uniformly throughout the experimental series and occa- sionally one or two were found in the controls. The proportion of this type of abnormality did not seem to show so close a corre- lation with the degree of injury to the sperm as did the stereo- blastulse. Other types of abnormal gastrulae were very common, this stage seeming to be a difficult one to pass. Some were observed to be much smaller in size than the normal type, perhaps due to a separation of the blastomeres in the early stages. Others were irregular in various ways. The most common type of this irregularity is what I have called in my notes "ragged." These forms do not possess such a clean cut appearance but are cpvered with minute irregularities and look as though they were going bad. The plutei were variable in the controls, but they were very much more so in the experimental cultures. The arms sometimes failed to form or appeared only as small buds. Usually however the arms were longer and more slender than in the controls. Several cases were observed in which the anal opening did not form. In all cases after the first appearance of motility a much larger number was found in the experimental cultures lying motionless on the bottom of the dish. If disturbed they quickly settled down to their former condition of rest. Some seemed to entirely lack the power of motion although this was not certainly determined. In general all of the results point toward a great disturbance in the development of all or nearly all parts of the organism. This disturbance seems to be produced by injury to the sperm cells used in insemination. An attempt to see if, in the case of the experiments using alcohol, the disturbance was due to a 236 NEIL S. DUNGAY. possible trace of alcohol carried over with the sperm in the process of insemination, was made by adding alcohol to the control cultures at the time of insemination. The addition in quantities comparable to the amount present in the experimental cultures was without effect. Table II. gives the principal features of a representative experiment, the minor features being omitted. TABLE II. EXPERIMENT A 2. July 4, 1911. Sperm removed at 9:00 A.M. from a male with testes about half soft. Two drops of thick sperm placed in 10 c.c. of each grade of alcohol at 9:15 A.M. and left there 5 hours and 20 minutes. Eggs about 2 per cent, immature. Used immediately after removal. All except A 2.0 (the uninseminated control) inseminated at 2:35 P.M. Discontinued July 7, 1911. No. Per Cent. Alcohol. Condi- tion at 3:3° P.M., July 4. Condition at 11:00 A.M., July 5. Condition at 2:20 P.M., July 5- Condition A.M., July 7. A2.0 O Nor- Normal. No cleavage. mal. No cleavage. Breaking down. A2.I 0 2-cell. Normal gastrulae. Late gastrulae. 10 per cent, ab- Trace abnormal. normal. A2.2 2 2-cell. Gastrulae 10 per Many abnormal late Mostly abnor- cent, ragged. gastrulae. Cili- mal. ated plates seen. A2.3 3 2-cell. Gastrulae 50 per As at 11:00 A.M. 60 per cent, very cent, ragged, i only more devel- abnormal. per cent, stereo- oped. blastulae. A few 2 cells. i morula. A2.4 4 2-cell. 40 per cent, abnor- As at 11:00 A.M. 80 per cent, ab- mal, i per cent. but more devel- normal. stereoblastulse. oped. A2.5 5 2-cell. Mostly abnormal As at 11:00 A.M. All dead. gastrulae. 4 per but more devel- • cent, stereoblas- oped. tulae. A2.6 6 No Trace swim. 5 per As at 11:00 A.M. All dead. cleav- cent, stereoblas- but more devel- age. tulae. oped. A2.-J 7 No None swim, n per Dead. cleav. cent, stereoblas- age. tulae. A2.8 8 None None fertile. fertile. V. CYTOLOGICAL STUDY. A. Normal. I have found it necessary to go into some of the cytological details which are to be found in the egg of Nereis after insemina- EFFECTS OF IXJURY UPON SPERM. 237 tion with normal sperm. These features have been described previously (Lillie, F. R., 'n, '12) but an outline of the cyto- logical changes is necessary for a proper understanding of the observations on the experimental material. My studies have been based upon my control series and upon slides loaned to me by Professor Lillie. The latter are especially fine and show clearly all of the conditions which he has described in his papers. Immediately after a spermatozoon becomes attached to the egg by means of its delicate perforatorium the contents of the coarse alveoli of the cortical layer of the egg begin to pass to the exterior where, in the course of fifteen minutes, they form a thick layer of jelly. A perivitelline space now occupies the position formerly held by the cortical layer of the egg. An entrance cone forms from the egg cytoplasm just beneath the spermatozoon and, about three quarters of an hour after insemination, draws the sperm head, in the form of a thread, into the egg, leaving the tail and middle piece outside. This usually occurs during the late anaphase of the first maturation spindle. The sperm head now begins to grow larger and, retaining its connection with the entrance cone, penetrates the egg protoplasm. The whole complex, as it passes to the yolk free area of the egg, rotates so as to bring the sperm nucleus ahead and nearer to the center of the egg. Meanwhile the sperm aster appears at the pole of the nucleus which is opposite the cone and soon divides, forming an amphiaster. Following the formation of the second polar body, the egg aster fades away and the egg chromosomes swell and form conspicuous vesicles, each one with a distinct chromatic nucleolus, which slowly fuse together to form the egg nucleus. By this time the sperm nucleus has increased in size and presents an appearance much like that of the egg nucleus, near which it has come to lie. The two nuclei now fuse and the asters of the first cleavage spindle, presumably derived from the sperm amphiaster, are seen on opposite sides of the cleavage nucleus. The process of cleavage then ensues. B. Experimental. In the experimental material it has proved difficult to establish a definite seriation in the cytological observations, since, as 238 NEIL S. DUNGAY. the observations on the living material indicate, in the experi- mental cultures different eggs start to develop at different times. This seems, at least in part, to be due to the relatively low activity of the injured sperm cells and to the consequent delay in attach- ment. Likewise it has been hard to get any satisfactory estima- tion of percentages of the various stages present in any preserva- tion. The reason for this lies in the fact that in most cases where marked results wrere obtained, only a small fraction of the eggs were fertilized and too few fertilized eggs were secured from any one series to get counts large enough to be significant in such complex conditions. The losses in sectioning and staining also add to the difficulties. Yet it has been possible to secure the main facts by going over a large number of slides and by com- paring the results of different experiments. Cytological study of the preserved samples from the experi- mental cultures of Nereis eggs shows the following classes of eggs. (a) Eggs not entered by sperm. No segmentation. 1. Polar bodies formed. Jelly not extruded. 2. Polar bodies formed. Jelly extruded. (b) Eggs entered by sperm. Segmentation. 1. First cleavage not completed. 2. Multipolar mitosis. Maturation incomplete. 3. Multipolar mitosis. Maturation completed. Poly- spermy. 4. Miscellaneous abnormal eggs. 5. Apparently normal eggs many of which would doubtless have proved to be defective had they been allowed to develop. Cytological study reveals nothing as to the cause of the ab- normalities occurring after the time of the first cleavage, with the single exception of the above mentioned multipolar mitosis. (a i) As is indicated by the records made from examination of the living material, by the India ink method, a partial or total lack of jelly formation is one of the striking things which is to be found in nearly all of the experiments. I have never seen this in normally fertilized eggs, either whole or in sections, EFFECTS OF INJURY UPON SPERM. 239 although egg sections from early preservations of the experi- mental series show in some slides as many as 5 per cent, of the maturing eggs with the alveoli of the cortical layer still wholly or partially filled. In most cases an attached sperm cell may be found outside the vitelline membrane. Since it is in an exposed position, not protected by jelly as in other eggs, its absence may be interpreted as due to its accidental removal after preservation. I have not found fertilization cones in sections which show any considerable amount of material still present in the cortical alveoli. Maturation goes on in an apparently normal fashion (Fig. i) though probably much slower than usual since the later maturation stages of this type are not seen until the normal maturation stages have been completed for some time. I have found most of the principal maturation stages, from the breaking down of the germinal vesicle to the formation of the vesicles of the female nucleus, and, with the exception of the cortical layer, they seem to be like the normal stages. The polar bodies either come to lie imbedded in the cortical layer, or are pushed entirely through it to the outside (Fig. 2). After the formation of the second polar body the chromosomes, in the usual manner, rapidly form a number of large chromosomal vesicles with the haploid number of chromatic nucleoli. But no indication of the sperm is to be seen within the egg. No trace of cytoplasmic radiations is ever seen within the egg after the maturation phenomena are completed and the sperm nucleus is not to be found unless the remnants of the sperm cell are still attached to the outside of the vitelline membrane. The chromosomal vesicles, not being able to fuse with the sperm nucleus, may now behave in one of several ways. They may scatter at random through the cytoplasm or they may retain their position beneath the polar bodies, in which case they usually form a more or less hazy, pcorly staining mass of irregular shape. In a few cases I have found about a dozen separate or partially fused vesicles containing in all about 26 chromatic nucleoli. At first I thought that this was a case in which the sperm had entered. But the entire absence of cyto- plasmic radiations, the irregular arrangement of the vesicles, and the determination of approximately the haploid number of vesicles indicates that we are dealing with a case of multiplica- 240 NEIL S. DUNGAY. tion of the chromatic nucleoli without true fertilization. In some cases the chromosomal vesicles seem to dissolve in the cytoplasm, leaving behind several small chromatic nucleoli. To sum up, the sperm attaches itself to the egg which slowly undergoes maturation but does not form jelly or segment. The sperm does not enter the egg and no fertilization cone is formed. After the formation of the second polar body, the chromosomal vesicles of the egg nucleus form and, failing to unite with a sperm nucleus, undergo degenerative changes. The spermato- zoon thus gives the egg a stimulus which induces only the forma- , tion of the polar bodies. In such cases the maturation process goes on in a normal manner with the exception of a retardation in the rate. This retardation is probably caused by the reten- tion of the jelly-forming materials in the cortical layer of the egg, which would tend to cut down the rate of exchange with the external medium. It is possible that we are dealing in this case with eggs which require a rather high degree of stimulation in order to produce jelly. Observation on unfertilized eggs shows that there is a considerable degree of variation in the amount of stimulus required to produce jelly formation. It may be that the injured sperm cells are incapable of giving so great a stimulus for jelly formation as do the normal ones and that some of the eggs can- not respond to the slight stimulus given, by emptying their cortical alveoli, although they may be able to form polar bodies. Whether the cause lies in the egg or in the sperm it is necessary, since these eggs develop normally if fertilized with uninjured sperm, to assume that the injury to the developmental process is produced by the action of external influences upon the sperm cell before the time of insemination. It is also clear that in the case of Nereis eggs maturation may take place without jelly formation. (a 2) In a much larger number of cases the eggs go a step farther than those described above. The living material shows eggs which undergo both maturation and jelly formation but which never segment. Although, in the earlier stages, a sperm cell may be seen to be attached to the vitelline membrane no fertiliza- tion cone is present. Examination of the sections of such eggs EFFECTS OF INJURY UPON SPERM. 24! completely confirms the observations made upon the living material. The sperm cell becomes attached to the vitelline membrane, the processes of maturation and jelly formation follow in a normal manner, and the animal pole of the egg becomes free from yolk. But no fertilization cone is formed and no attach- ment granules appear at the end of the perforatorium (Fig. 4). The egg protoplasm just beneath the attached sperm shows no more change than that at any other part of the egg. The per- foratorium, which in cases of normal attachment becomes wider and more easily seen (Fig. 5), remains as a very delicate process which cannot be followed beyond the vitelline membrane. Large granules, which stain deeply in the iron heematoxylin, soon appear in large numbers in the peripheral parts of the egg and also around the egg nucleus. In later stages the sperm head becomes larger and somewhat less dense and often disappears entirely. This is presumably due to secretions which are thrown off from the egg, as has been demonstrated by F. R. Lillie ('12). No evidence of the presence of a sperm within the egg can be found. At the close of the second maturation period the egg chromosomal vesicles begin to swell and may unite with each other to a greater or less extent (Fig. 3). Cytoplasmic parts of the mitotic figure never appear although there are sometimes traces of chromosome formation within the vesicles in the earlier stages (Fig. 6). In some cases the vesicles separate and scatter through the cytoplasm where they may dissolve and leave behind a number of deeply staining chromatic nucleoli. More often the vesicles remain in place and break down into a mass of poorly staining debris (Fig. 7) in which, in the earlier stages, chromosome-like forms may be visible. Two or three hours after insemination the cytoplasm frequently begins to bud off small pieces filled with deeply staining granules. This usually appears first. near the equator. Later the whole cell may break up into parts which more or less resemble blastomeres, although close examination in the living state reveals their nature. In sections it is easily seen that the parts do not possess nuclei. It is clear that the injured sperm has given the first stimulus of fertilization. But, although the initial stages have been pro- duced, the effects necessary for further development have not 2_}2 NEIL S. DUNGAY. been called forth within the egg. The achromatic parts of the mitotic figure are entirely absent although there are indications that the chromosomes may undergo a part of the changes pre- liminary to first cleavage The effects produced by the attach- ment of the sperm seem to be nearly the same as those secured by centrifuging the eggs or by giving them any mechanical stimulus (Lillie, 'n). It is of interest to find that a few of the eggs of the type described above show the presence of two or more sperm cells attached to the vitelline membrane. This has been found in several cultures in which many eggs showed no evidence of having come in contact with a sperm cell. Apparently the weakened sperm cells are not able to promptly call forth in the egg the reaction which prevents the attachment of more than one sperm. The partial jelly formation which was observed in the living material is often associated with the failure of the sperm to enter the egg. However cases have been seen in which the sperm entered and caused first cleavage to take place. The farther history is not known but, since preservations made from some- what later stages show no traces of such eggs, it is probable that the cortical alveoli are soon. emptied. (b i) A very interesting observation upon the living material has been recorded above. In many experiments a few eggs seemed to have nearly completed the process of first cleavage when the cleavage plane ceased to advance and gradually faded away. Cytological examination fails to pick out these eggs until the early telophase stages of first cleavage. The earlier processes seem to go on in a normal manner. In most cases the chromo- somal vesicles fail to unite properly and the plane of cleavage fades away. Usually it leaves behind a change in the arrange- ment of the yolk granules and its position is occupied by a some- what lighter staining band in which small granules are less abun- dant. Parts of the sharp line which marks the completed cleavage plane in sections may be left behind, especially those parts near the animal pole of the egg (Fig. 8). The chromosomal vesicles sometimes scatter widely and partially dissolve and sometimes they remain near each other. Very often the region near the vesicles becomes filled with deeply stainirg granules. In one case EFFECTS OF INJURY UPON SPERM. 243 a cell was found in the late prophase of second cleavage, though without a first cleavage plane. In this case the nuclei have been able to divide again though the cytoplasmic division was never completed. In most cases the cells which fail to finish the first cleavage never go any farther. (b 2) A very few eggs were found in each of the experiments which exhibited the above described attempted cleavage, in which multipolar spindles were found. Unfortunately my pre- served material contains very few of these eggs and the sections of some of them were lost or badly broken in the process of preparation. There are two different kinds of eggs which show this condition. One kind has formed two polar bodies. There are very few of these eggs which are in a condition good enough to work upon. Although the cultures from which the eggs were taken were more than 75 per cent, unfertilized these eggs seem to show evidences of polyspermy. This is borne out by the fact that, in a few cases, two sperm cells were seen to be attached to a single living egg in cultures similar to those from which these eggs were taken. Although it is conceivable that multipolar figures might arise from causes other than polyspermy, it is very unlikely. I regard the few observations of polyspermy in the living cultures as good evidence that the few cases under consideration are due to the entrance of more than one sperm cell. The observations men- tioned under (a 2) above also point in this direction. (b 3) The other kind of multipolar figures is found in eggs which show defective maturation. In one case an egg was found in which the first polar body had formed. Just beneath it lay 14 chromosomal vesicles and in the center of the section about 20 vesicles lay near a strong cytoplasmic radiation. The presence of the radiation probably indicates that a sperm nucleus has fused with the egg nucleus, although the number of vesicles which can be made out is below the diploid. Another sperm cell lies in contact with the vitelline membrane, probably attached to it. Although this is not a case of multipolar mitosis in itself it is possible that the 14 vesicles near the animal pole, which represent the second polar body, may enter into the cleavage process and aid in the production of multipolar figures. One or 244 NEIL S. DUNGAY. two cases of such figures in eggs with but one polar body have been noted in late cleavage stages. Other eggs may fail entirely to produce polar bodies. In these eggs from six to nine poles form and the spindles are often very complicated and produce an irregular distribution of the chromosomes, which, in the earlier figures, take the form which is found on the second maturation spindles. It has been shown (Lillie, F. R., 'n) that the formation of the first polar body may be prevented by the use of the centrifuge. In such cases multipolar figures of various kinds are formed. Clearly my case is due to the sup- pression of the polar bodies. I have not seen similar cases in sections of normal series though, in such small numbers as are found in the experimental series, it is possible that they may have been overlooked. (b 4) A few scattering abnormalities have also been found but not in sufficient numbers to be of any significance. Two cases in which the sperm and egg nuclei fused together and failed to develop farther have been seen. Two cases of prophase figures from second cleavage were found in which the spindles lay at right angles to each other in different planes. Other abnormal conditions were also found. All of these cases are open to the interpretation that the egg was abnormal before fertilization and are not worth description. Careful search was made for eggs showing abnormalities at the period of entrance of the sperm. It seemed likely that the stages of fusion of the egg and sperm nuclei would contain evi- dence of abnormality but close examination failed to show anything of the kind, except in the case of the two eggs already mentioned. Since a kind of parthenogenetic cleavage might be produced by the partial action of the sperm cell, search was made for eggs exhibiting the haploid number of chromosomes in the first cleavage. But in all cases the number was clearly diploid. VI. DISCUSSION. A. The Production of Defectives. — For a long time it has been assumed by many that the offspring of parents who habitually take alcohol or other drugs or who work in lead are very apt to be in some way weak or defective. Some of the unfortunate EFFECTS OF INJURY UPON SPERM. 245 conditions in the offspring which different men have associated with parental alcoholism are insanity, epilepsy, feeble-minded- ness, cretinism, macrocephaly, lack of self-control resulting in criminality or in over-indulgence in alcoholic drinks, malforma- tions, retarded development, poor health, and the like. Medical men often state that there is every indication that a great number of abortions, stillbirths, and births of defectives are definitely connected with times of conception corresponding to periods of drunkenness, or drug craze, or of working in lead. So much has been written on both sides of the question, especially with refer- ence to alcohol, that it is impracticable to attempt to give any exhaustive review of the literature. Unfortunately, also, so much of the writing is the product of prejudice and of a desire to establish some point, regardless of the evidence, that it is difficult to separate the wheat from the chaff. There are three theories to be mentioned in connection with the relatio^ of alcoholism to the production of defectives, (i) Since the excessive use of alcohol unquestionably produces un- desirable effects upon the body of the consumer, many believe that the offspring of an alcoholic parent may in some way inherit certain weaknesses thus produced. But it seems that at present we have no reason to believe that specific somatic alterations of the parent, produced by drugs or otherwise, are inherited by the offspring. Consequently any such theory may be rejected at once. (2) Some writers go so far as to assert that there is no direct causal connection between parental alcoholism and defectiveness in the offspring. They maintain that alcoholism, insanity, feeblemindedness, epilepsy, etc., are often merely different forms of expression of a single weakness. In many cases, as stated by W. Branthwaite, H. M. Inspector under the Inebriates' Act ('08), alcoholism is undoubtedly caused by some weakness which is present in the family. There seems to be sufficient evidence to show that this statement is true. Yet we cannot accept the theory that this hereditary weakness entirely explains the rela- tions under discussion. (3) Very few writers have even ventured to suggest that the defective conditions appearing in the offspring of alcoholics 246 NEIL S. DUNGAY. may be due in part to the direct action of the alcohol upon the germ-cells ; in case the father alone be alcoholic that his spermato- zoa may be so affected as to induce abnormal development of the ovum fertilized by him. Until recently there has been little evidence to show that such a theory is tenable and more evidence is desirable. The data presented in this paper demonstrate that the sperm cells of Nereis and Arbacia may be so affected by alcohol and by other methods of treatment, that their union with normal eggs will produce an abnormal development. As is shown below, there is evidence of a similar nature in other forms. In man the conditions are so complex that they are hard to analyze, especially since we may not resort to direct experiment. Alcohol poisoning may take place in utero, during the nursing period, or even later. Or it may be that the same depressing conditions which caused the parents to drink may react similarly upon the offspring. Many other difficulties also confront the investigator who seeks to solve the problem by the methods of the past. It seems clear that such methods will never give the solution. Definite biological expeA/nent upon the lower animals must form the chief basis for any conclusions which we may make in the future. With this thought in mind the present series of experiments was undertaken in the hope that some definite evidence might be presented upon the subject. Although there may sometimes be a suspicion that the data relating to man have been selected with a view of establishing a point rather than seeking the truth, it is worth while to recall a few cases because of their interest in the present connection. Several European investigators (e. g., Schweighofer and Bez- zola) are frequently quoted in the literature as having found that there is a definite relation between the time of the greatest number of stillbirths, abortions, and births of mental defectives, and the great feast seasons, during which much alcohol is con- sumed. From a statistical standpoint such statements have been severely criticized, especially by Pearson and Elderton ('10). There is also little evidence to show that the effect of the alcohol was not exerted upon the developing embryo. Since the effect of the alcohol taken by the mother may be either upon EFFECTS OF INJURY UPON SPERM. 247 the egg or upon the developing organism, the most desirable evidence comes from cases in which the male germ cells alone are exposed to the supposedly unfavorable conditions. Saleeby ('n) quotes Galton as having given the three following cases. A man who had normal children became a drunkard and his later children were all imbeciles. A healthy woman who had by a drunken husband five sickly children who died in infancy, later married a healthy man and produced normal children. A man with two healthy children acquired the cocaine habit and engendered two idiots. Schweighofer gives a case of a normal woman who had three normal children by a sound man. She later married a drunkard. Of the three children from this union one had infantilism, one was a drunkard, and one was a degenerate. In a third marriage she again bore healthy children. Paul ('60) tells of the children of lead workers. From 32 preg- nancies, the father alone being exposed to the lead poisoning, there resulted 12 abortions, stillbirths, and premature labors, and 20 living births. Of the living 8 died during the first year, 4 during the second, and 5 during the third. Paul states that the influence of the lead is as real as in the cases where the mother is exposed, though perhaps the effects produced are not so great. The observations quoted above are among the best on record. Many other similar observations can be collected by anyone who thinks it worth the time. A rather full bibliography is given by Hoppe ('12). In the light of recent work these facts are very interesting. Although in some cases the remarriage serves as a control, the lack of data concerning the previous family histories is a defect sufficiently serious to warrant us in questioning any conclusions which may be drawn from these data alone. So far as I have searched there are no observations upon man which meet with rigorous scientific requirements. Elderton and Pearson ('10) conclude from the statistical study of English school children that the offspring of alcoholic parents are slightly brighter, heavier, and less diseased than those of sober parents and that epilepsy and tuberculosis are of no more frequent occurrence than among the children of non-alcoholics. 248 NEIL S. DUNGAY. They also record a higher mortality among the children of alco- holics and conclude that the more resistant survive. The publication of this paper has been followed by widespread criti- cism. The details of the individual cases are not given. We do not know what may be included under the terms "alcoholism" and "intemperate." We should know the condition of the chil- dren who were not in school, due perhaps to lack of ability or to ill health, the relation of the time of conception to periods of drunkenness, and whether the mother, the father, or both, were alcoholics. In view of the lack of so much desirable evidence and of the heterogeneity of the materials investigated it seems that the question as to the general applicability of the conclusions reached is at least an open one, the more so since the experi- mental evidence seems to be directly opposed in most cases. Additional light is given by the researches of Nicloux ('oo) who proves that alcohol may reach the ovaries and testes of mammals and that these organs take up considerable quantities of the drug. Alcohol is also present in the seminal fluid very shortly after it is taken into the stomach. Accordingly it seems to be possible that the sperm cells may be injured. Bertholet's work ('09) in a sense confirms that of Nicloux, since he finds that testicular atrophy is common among alcoholics. His observa- tions lead one to think that sterility should be much more com- mon than it is among drunkards. There are a number of published observations upon the lowrer animals. Mairet and Combemal ('88) foumd that a dog treated with absinthe for 8 months and paired with a normal female gave 12 young. 2 were born dead and the others all died within II weeks after birth. The small numbers and the lack of adequate control make this experiment indecisive. Bardeen ('07) found that toad eggs, when fertilized by sperm cells wrhich had been previously exposed to the X rays, developed abnormally. Since the question involved is wider than the alcohol question these results are significant. O. Hertwig ('10, 'n) and G. Hertwig ('12) obtained similar results by the exposure of the sperm cells of various animals to radium and also by injecting methylene blue into the dorsal lymph sac of the male frog some days before the sperm cells were used to EFFECTS OF INJURY UPON SPERM. 249 fertilize normal eggs. In this work it was found that the stronger action upon the germ cells produced earlier and more profound alterations in the offspring. Stockard ('12) furnishes a set of well-planned experiments upon guinea pigs which, when tested, produced normal offspring. The males were intoxicated by inhaling alcoholic fumes. As a result of 24 matings of alcoholic males with normal females he reports no results or early abortions in 14 cases, 5 stillborn litters, and 5 living litters containing 12 young. Of these, 7 died shortly after birth. The remaining 5 "are unusually small and very shy and excitable animals." The parents remained in good health throughout the experiment. Although the number of matings is rather small there is a clear indication that the action of the alcohol upon the male germ cells is the cause of abnormal results. Considered in connection with the work of Bertholet and of Nicloux this is very strong evidence. The experiments of Nice ('12) do not seem to entirely agree with the experiments just quoted. His mice, both male and female, were fed on milk and crackers, to which 2 c.c. of 35 per cent, alcohol was added daily. They were also furnished with drink in the form of 35 per cent, alcohol. The fecundity of the mice was greater than that of the control series, though the mortality was n.i per cent, in the offspring in the experimental series and zero in the controls. None of the young were de- formed. Nicotin, caffein, and tobacco fumes gave similar results. In the absence of farther experimental data it seems fair to as- sume that the germ cells of the mice are in some way less sus- ceptible to these drugs, though the comparatively high mortality indicates that there was an effect. Gager ('08) subjected pollen grains to the action of radium and secured a marked change in the plant resulting from pollina- tion of a normal plant. Some of the effects persisted through several generations. My owyn experiments upon Nereis and Arbacia demonstrate that injury of the sperm cells by several methods may produce a series of abnormalities which may appear in various stages from the time of insemination up to late larval stages. It is not known whether, if the animals were kept under favorable condi- 25O NEIL S. DUNGAY. tions for life and growth, other abnormalities would occur in later stages, but since experimental cultures which are apparently normal in the earlier stages frequently develop abnormalities in the larval condition, it is very probable that other changes would come to the surface in still later stages. In the Nereis experiments there can be no doubt as to the cause of the abnor- malities observed. The controls, using the sexual products from the same animals, exclude the possibility of abnormalities being present in the lines used. The very large numbers handled demonstrate that the results are not due to the chance out- cropping of hidden defects. The series composed of Arbacia material, if standing by itself, might be questioned. But the general type of results is so similar to that obtained by Bardeen, the Hertwigs, Stockard, and myself on other forms that there can be no question as to their applicability. The lack of specificity in the action of the agents used is remarkable. At first thought it seems strange that acid, alkali, alcohol, heat, delay, and other means should produce similar results. It is clear that the action must be much the same in all the cases which I have recorded as well as in those recorded by the authors cited. At present we can neither assign a definite reason for the lack of specificity in the results nor tell how the abnormalities are produced. I can only suggest that the sperm cells are affected in such a manner that their vitality is lessened, or in other words, their rate of metabolism is lowered. It is known that the rate of metabolism in the normally fertilized egg rises rapidly after the time of fertilization and continues to rise for some time. Child ('n) has shown that in the regulation* of pieces of Planaria the type of structure formed may be definitely controlled through changes in the rate of metabolism produced by means of low temperature, anaesthetics, carbon dioxide, etc. He finds that, in a general way, regions of normally low rate, such as posterior regions, are most affected. Any particular process of morphogenesis seems to require a certain minimal rate of metabolism for its normal completion. It is possible that we are dealing with a similar case. The injured sperm cell may be unable in some cases to give to the egg a sufficient stimulus to raise the rate of metabolism to a point EFFECTS OF INJURY UPON SPERM. 25! where it is able to draw in the sperm head. In other cases the sperm head enters and the rate may be increased greatly but not enough to cause the normal completion of many processes. In some cases the rate may be so low that cleavage cannot take place in a normal manner. Since nearly all of the forms investigated showr abnormalities at the period of gastrulation it may be that the minimal rate necessary for normal gastrulation is often not reached by the eggs in the experimental cultures. In Nereis at the period of elongation fcllo\\ing the tro.h ->j: hore stage, there is again a product'on of abnormalities in the posterior region. Although the evidence is very fragmentary at present it seems that the lack of specificity in these experiments is perhaps capable of explanation upon the basis of lowered rate of metab- olism produced by injury to the sperm. Farther research will be necessary before wre can do much more than venture a guess as to the solution of the problem. We are safe in concluding from the observations upon man and the higher mammals and from the experimental work upon mam- mals, amphibians, annelids, echinoderms, and the higher plants, that it is possible to injure the male germ cells by the application of external forces so as to produce a change in the next generation at least. We cannot say just how long this change may persist. The work of Gager upon plants indicates that changes so pro- duced may persist through several generations. Tower's work upon the female germ cells of Leptinotarsa indicates that the changes may persist or may gradually fade away. In all proba- bility the results obtained by the methods which I have used will prove to be for the most part transitory, although there is no reason to believe that some of them may not persist. There is, then, good reason to believe that some drugs, such as alcohol and cocaine, are a detriment, not only to the consumer but also under certain conditions to his offspring. Since alcohol appears in the seminal fluid very shortly after being taken into the stomach, there is good reason to believe that a man, intoxi- cated for the first time, even, may beget offspring which will be in some degree defective. It is also possible, though not demon- strated, that, in addition to drugs taken voluntarily into the system, the products of abnormal metabolism may exercise g. 252 NEIL S. DUNGAY. similar influence. It is even conceivable that nervous states may be able through alteration of somatic metabolism, to affect the germ cells. Many other possible causes of abnormality might be mentioned. But speculation is useless at this time and experimentation is needed. A large field for investigation is opened up and the results of experiments in this field cannot fail to be of interest to the student of eugenics. The sociological application of the observations here recorded is sufficiently obvious. B. Fertilization. — The results of these experiments are com- pletely in accord with those given by F. R. Lillie ('n and '12). So far as my observations extend they are practically identical with those given by Lillie. He finds that if the attached sperm cell is removed by centrifuging, the processes of jelly formation and maturation go on in a normal manner, though cleavage does not result. My experiments demonstrate that if the sperm is injured so much that it fails to enter the egg, essentially the same results are secured. This fact again supports the view expressed by Loeb ('09), Lillie ('11 and '12) and Bataillon ('12) to the effect that at least two factors are involved in the process of fertiliza- tion. The first is concerned with membrane formation and, in itself alone, is insufficient. Certain cytoplasmic changes such as the rearrangement of the yolk granules are also produced by the initial stimulus. Since slight stimuli cause jelly formation and maturation, pricking as in Bataillon's experiments is probably sufficient to produce these changes. Certainly it is difficult in Nereis to see any action of the sperm beyond the attachment of the perforatorium, which is responsible for the early changes in the egg. In Nereis it is evident that maturation may take place in the absence of membrane formation if the stimulus given by the sperm is sufficiently light. The second factor has to do writh the internal stimulus. It is a difficult matter to determine at just what stage fertilization is complete. In all cases which I have observed, the formation of a fertilization cone is succeeded by the entrance of the sperm head and by the formation of the first cleavage spindle. The internal stimulus is not yet com- pleted, even in the case of fertilization by a normal sperm cell at the time when the fertilization cone is formed. This is shown in EFFECTS OF INJURY UPON SPERM. 253 Lillie's experiments by the removal of the sperm cell at this stage. Since the formation of attachment granules and a fertilization cone is not in itself necessary for membrane formation, matura- tion and other visible cytoplasmic changes, as is shown by my experiments, and since their presence without the sperm head, does not lead to a greater visible change than does their absence, there is good reason to believe that they function solely in the attachment, penetration, and revolution of the sperm head. After these functions have been performed they cannot be traced much farther. They are probably dedifferentiated and behave as ordinary cytoplasm. I am inclined to think that in Nereis the mere penetration of the egg cytoplasm by the sperm head is insufficient although the observations on this point are not entirely satisfactory. The experiments of Ziegler ('98) and Wilson ('03) support this view. Ziegler succeeded in producing a constriction in the egg of the sea urchin in such a way that it separated the egg nucleus from the sperm nucleus. The part containing the sperm nucleus segmented. The remainder failed to segment, but gave indication that the presence of the sperm nucleus was not without effect, by dissolving and reappearing several times. Wilson cut the eggs of Cerebratulits in two, shortly after the penetration of the sperm cell. When the cut separated the two nuclei the part containing the sperm nucleus segmented and the other part formed polar bodies but refused to cleave. Even in cases where the germ nuclei of Nereis copulate there is not necessarily a complete stimulus to development. F. R. Lillie thinks that the partial sperm nuclei produced by centri- fuging at the proper time do not cause normal cleavage. Some- times a partial cleavage results and sometimes cleavage ceases in the two-cell stage. In my experiments many eggs either attempted to cleave and failed or stopped after cleaving once. This indicates that the mere presence of the germ nuclei is insufficient as an internal stimulus for development. The normal interchange between the nucleus and cytoplasm is not neces- sarily brought about. Although the nucleus of the sperm succeeds in producing aster formation yet the rate of metabolism is so low that normal cleavage cannot take place or can take place 254 NEIL S. DUNGAY. only once or twice. Although we have no direct evidence as to the comparative rates of metabolism of eggs fertilized with normal sperm and those fertilized with injured sperm it seems possible that a difference exists. The retardation in development, the higher mortality in the later stages, and the differences in photo- tropic response indicate that this is likely. VII. SUMMARY. 1. Nereis eggs, inseminated with sperm cells which have been injured by one of several methods, may fail to develop in a normal manner. 2. In some cases the egg does not form a fertilization cone, attachment granules are lacking and the sperm head is not drawn into the egg but remains outside attached to the vitelline mem- brane. There are two classes of such eggs, neither of which segment, viz.: (a) Those which slowly undergo maturation with- out forming jelly, (b) Those which form both jelly and polar bodies. 3. Those eggs in which the sperm head enters, form the first cleavage spindle in an apparently normal fashion but may fail to complete the division of the cytoplasm. Those which com- plete this division may develop abnormalities in later stages. 4. The eggs of Arbacia exhibit a similar series of abnormalities when fertilized by weakened sperm cells. 5. There is no indication of specificity in the action of the agents used in injuring the sperm cells. 6. These experiments demonstrate that eggs fertilized with sperm cells injured by alcohol and by other means may produce abnormal forms. Taken in connection with the demonstration by others that the germ cells of mammals may be exposed to injurious conditions this has an important bearing upon the relation of alcoholism to the production of defectives. 7. There seem to be at least two factors involved in the process of fertilization. The one has to do with membrane formation and certain other changes, the other with the internal stimulus. 8. In Nereis the presence of the two germ nuclei within the egg is not necessarily sufficient as an internal stimulus for normal development. EFFECTS OF INJURY UPON SPERM. 255 I am indebted to Prof. Frank R. Lillie for suggesting this line of work and for his many kindnesses during the progress of the work. LITERATURE CITED. Bardeen, C. R. '07 Abnormal Development of Toad Ova Fertilized by Spermatozoa Exposed to Roentgen Rays. Jour. Exp. Zool., Vol. 4. Bataillon, E. '12 La parthenogenese des amphibiens et la " fecondation chimique " de Loeb- Ann. des. Sci. Nat. Zoo., Ser. 9, Tom. 16. Bertholet, E. '09 Ueber Atrophie der Hoden bei chronischen Alcoholismus. Centralbl. f. allgem. Path., Bd. 20. Branthwaite, W. '08 Inebriety, its Causation and Control. Brit. Jour. Inebriety, Jan., 1908. Child, C. M. 'n Experimental Control of Morphogenesis in the Regulation of Planaria. Biol. Bull., Vol. 20. '12 The Process of Reproduction in Organisms. Biol. Bull., Vol. 23. Elderton, E., and Pearson, K. '10 A First Study of the Influence of Parental Alcoholism on the Physique and Ability of the Offspring. Eug. Lab. Mem., X., Dulan, London. Gager, C. S. '08 Effects of the Ray of Radium in Plants. Mem. New York Bot. Garden, IV. Hertwig, G. "12 Das Schicksal des mit Radium bestrahlten Spermachromatins im Seeigeli. Eine experimentell-cytologische Untersuchung. Arch. f. mik. Anat., Bd. 79. Hertwig, O. "10 Die Radiumstrahlung in Hirer Wirkung auf die Entwicklung tierischer Eier. Mitteilung vom 15. Juli, 1909. Sitzungsber. der Konigl. Preuss. Akad. d. Wissensch., XL '10 Neue Uintersuchungen u'ber die Wirkung der Radiumstrahlung auf die Entwicklung tierischer Eier. Mitteilung vom 28. Juli, 1910. Sitzungsber. der Konigl. Preuss. Akad. d. Wissensch., XXXIX. '11 Mesothoriumversuche an tierischen Keimzellen, ein experimenteller Beweis fur die Idioplasmanatur der Kernsubstanzen. Mitteilung vom 6. Juli, 1911. Sitzungsber. d. Konigl. Preuss. Akad. d. Wlss., XL. 'n Die Radiumkrankheit tierischer Keimzellen. Ein Beitrag zur experi- mentellen Zeugungs- und Vererbungslehre. Bonn. Fr. Cohen. Hoppe, H. '12 Die Tatsachen ueber den Alkohol. Ed. 4. Munich. Just, E. E. '12 The Relation of the First Cleavage Plane to the Entrance Point of the Sperm. Biol. Bull., Vol. 22. Lillie, F. R. 'ii Studies of Fertilization in Nereis. I. and II. Jour. Morph., Vol. 22. '12 Studies of Fertilization in Nereis. III. and IV. Jour. Exp. Zool., Vol. 12. '12 The Production of Sperm Iso-agglutinins by Ova. Science, N. S., 36. 256 NEIL S. DUNGAY. Lillie, F. R., and Just, E. E. '13 Breeding Habits of the Heteronereis Form of Nereis limbata at Woods Hole, Mass. Biol. Bull., Vol. 24. Loeb, J. '09 Die chemische Entwickelungserregung des tierischen Eies. Berlin, Julius Springer. Mairet et Combemale. '88 Influence degeneration de 1'alcool sur la descendance. Comptes rend. hebd. des Sci. de 1'acad. des Sci., Tom. 106. Nice, L. B. '12 Comparative Studies on the Effects of Alcohol, Nicotine, Tobacco Smoke, and Caffein on White Mice. Jour. Exp. Zool., Vol. 12. Nicloux, M. 'oo Passage d'alcool ingere dans quelques glandes et secretions genitales. Comptes rend. Soc. de. Biol., Tom. 52. Paul, C. '60 Etude sur 1'intoxication lente par les preperations de plomb, de son influence sur le produit de la conception. Arch. gen. de. med., Tom. 15. Pearson, K., and Elderton, E. '10 A Second Study of the Influence of Parental Alcoholism on the Physique and Ability of the Offspring. A Reply to Medical Critics of the First Memoir. Eug. Lab. Mem., XIII., Dulan, London. Saleeby, C. W. 'n Parenthood and Race Culture. Moffat, Yard & Co., New York. Stockard, C. R. '12 An Experimental Study of Racial Degeneration in Mammals Treated with Alcohol. Arch. Int. Med., Vol. 10. Wilson, E. B. '92 The Cell-lineage of Nereis. Jour. Morph., Vol. 6. '03 Experiments on Cleavage and Localization in the Nemertine Egg. Arch, f . Entw'mech., Bd. 16. Ziegler, H. E. '98 Experimented Studien iiber die Zelltheilung, II. Arch. f. Entw'mech., Bd. 6. 258 NEIL S. DUNGAY. EXPLANATION OF PLATES. PLATE I. All figures were drawn with the camera lucida with Leitz apochromat 2 mm. oil immersion objective, and Zeiss No. 6 compensating ocular, c, cortical layer. FIG. i. Egg fixed i hour and 20 minutes after insemination with sperm which was injured by heating to 43-44° C. for 8 minutes. No jelly has formed. Periph- eral alveoli still intact. Anaphase of first maturation. FIG. 2. Reconstruction of 2 sections of an egg from same culture as Fig. i, fixed i hour and 50 minutes after insemination. Both polar bodies have been formed. Peripheral alveoli still unemptied. Egg nucleus in form of scattered vesicles, only part of which are seen in this figure. In all, 14 vesicles and about 28 nucleoli are indicated. Sperm cell (in another section) is still attached to vitelline membrane. FIG. 3. Egg fixed i hour and 15 minutes after insemination with sperm which was injured by heating to 43-44° C. for 17 minutes. Jelly was formed and both polar bodies are present. Second polar body shown. Egg nucleus in form of vesicles, 13 of which are indicated in the various sections. Sperm cell (in another section) is still external. Strongly staining granules around periphery of egg. eiOLOGICAL BULLETIN, VOL. XXV. PLATE I. ^X^ '£$£££:$:jWffi;$\-:i .•".'••I'?-' •'•• - ' '.'• /£ %$$^ty;!$ij$i •y.v/\">*:;".-V;-; ssfc Sii8S6^ISi ^fe»^[ i^SlttS» •-^••JB •^;:?;:^:V^^':^:^:l^.^ &£££ 7 c 260 NEIL S. DUNGAY. PLATE II. FIG. 4. Egg, from same culture as Fig. 3, fixed i hour and 5 minutes after insemination. Sperm attached to vitelline membrane, no attachment granules, and no fertilization cone. Perforatorium still delicate. Jelly and both polar bodies have formed. Vesicles of egg nucleus as in Fig. 6. FIG. 5. Egg from control culture 43 minutes after insemination, showing fertilization cone, attachment granules, and sperm with thickened perforatorium. FIG. 6. Egg from same culture as Fig. 3, fixed i hour and 27 minutes after insemination. Jelly and both polar bodies have formed. Sperm, in another section, attached to vitelline membrane. Vesicles of egg nucleus showing chromo- some within. FIG. 7. Egg from same culture as Fig. 3, fixed i hour and 36 minutes after insemination. Jelly and both polar bodies formed. Egg vesicles degenerating. Sperm (in another section) attached to vitelline membrane. FIG. 8. Egg fixed 2 hours and 17 minutes after insemination with sperm removed from male and kept at room temperature for 21 hours. This egg has not succeeded in completing the division of the cytoplasm and the chromosomal vesicles, about 28 on each side, are scattering. Strongly staining granules around the chromosomal vesicles. B.OLOGICAL BULLETIN VOL. xxv. mm PL,TE II. ''Rik-M-iiit- -; v:AV? 7 OBSERVATIONS ON LIVING SOLENOMYA (VELUM AND BOREALIS). EDWARD S. MORSE. The genus Solenomya, represented by a few species, is widely distributed throughout the world. It has been found on the east and west coast of North America, in West Africa, the Mediterranean, the Canaries, Australia and New Zealand. The few species agree in almost every detail but vary greatly in size. Our two species, 5. velum and 5. borealis, are one inch and three inches long respectively. S. bartschi from Manila is eight and one half inches in length. The species vary greatly in the proportional number of individuals. 51. velum is at times thrown up by thousands on our beaches, S. borealis is rarely found. 5. grand is, which departs more widely from the type than the others, is known by two specimens and a few fragments, while the gigantic 5. bartschi is represented by a single specimen. The various species have without exception a long semi-cylindrical shell, rounded at both ends, a long and almost straight hinge margin. They all have radiating ribs with an apparent inter- space in the series in which however a faint rib may be detected. They all have a highly polished periosteum which runs far beyond the border of the true limy shell, and to which the mantle with its curious system of muscles is closely adherent. The radiating ribs are continued through the overlapping periosteum by a marked thinning of the substance and Professor Drew discovered that when the periosteum folds inward as the valves close these thin areas fold in a plaited manner. In the dry state these interspaces often split and the substance being very brittle it is hard to preserve unbroken in cabinet specimens. The various species vary from a light yellowish brown to a dark brown and even a tar color. The shell within is grayish blue or lead color in S. borealis, purplish white in S. velum. Pelseneer8 has shown that Solenomya has a very primitive form of gill and Mitsukuri7 had previously studied the gill of 261 262 EDWARD S. MORSE. Nucula and Yoldia and showed also the primitive condition of the gills in these genera. Pelseneer has united Solenomya, Nucula, Leda and Yoldia in an order under the name of Proto- branchia, forming the lowest order in Lamellibranchs. It is interesting to observe that, while other genera of this low order vary greatly among themselves and from each other, Solenomya remains fixed in character despite the wide variation in size among the species and their wide distribution in space. With the exception of the incomparable work of Deshayes,2 that of Stempell11 and Pelseneer8 on Solenomya togata and that of Drew3 on Solenomya velum there have been very few obser- vations on the anatomy or habits of the living creature, and most of the records are derived from accounts of previous observers, and these were in many cases incorrect. Gould says the foot of the animal protrudes behind, whereas it protrudes in front; that the edge of the mantle opening is fringed and that two of the fibrils are larger than the others, which is incorrect, as regards our species and that the palpi are triangular, whereas they are long, narrow and semi-tubular. Stimpson's description, though brief, is without error. Professor George N. Perkins,9 in his " Molluscan Fauna of New Haven," a memoir teeming with details concerning the soft parts of Mollusca and their habits, has given one of the best accounts of Solenomya ever published in this country. Carpenter, in his lectures on Mollusca, says the mantle is closed in front, whereas it is wide open, and that there is a tail on each side of the external opening. Cooke mentions the two long tentacles springing from the siphonal orifice. Stempell figures in S. togata two papilla? much longer than the other siphonal processes. There is a puzzling discrepancy in the figures of the siphonal tentacles of Solenomya togata as portrayed by various authors. Philippi10 figures 61. togata with short siphonal papillae resembling a rosette. Pelseneer,8 in his contributions to the study of Lamellibranchs, figures this species without a trace of long siphonal tentacles. In his introduction to the "Study of Mol- lusca" he gives a figure of S. togata with a pair of siphonal tentacles 17 mm. long. This figure is however copied from Deshayes'2 "Natural History of Mollusca." It is probably OBSERVATIONS ON LIVING SOLENOMYA. 263 from this figure that Carpenter, Cooke and others have derived the long tail or tentacle mentioned by them. The excessive length of the two tentacles figured by Deshayes suggests the curious retractile tentacle in Yoldia limatula as described by Brooks.1 A careful examination of S. velum however showed no trace of such a sense organ. Stempell,11 who studied this species at Dohrn's laboratory and published an account in 1899, figures two siphonal tentacles 3 mm. in length. Are these differences the result of variation? Is there more than one species of Solenomya in the Mediterranean? Have the drawings been made from alcoholic specimens? The mystery can only be cleared up by a study of the living creature. In the following observations of Solenomya velum I have been greatly indebted to Mr. William F. Clapp, of Cambridge, and Miss Marjorie Newell, of Gloucester, for the material herein described. Mr. Clapp brought me thirty-nine living specimens of S. velum, young and nearly full grown. These were collected April 27, on Round Flat, a sandy mud area in Duxbury Bay. They were all buried from six to nine inches below the surface. Mr. Clapp believed they were simply inhabiting abandoned worm holes. Although powerful diggers he thinks they ordinarily dig but an inch or two deep. If they do dig to the depth he found them he believes they must occupy the same burrow for a considerable length of time, for in every case he observed the sides of the hole were discolored, closely resembling worm holes. I placed three specimens in a jar of sandy mud and they soon burrowed to the bottom of the vessel, leaving three sharply defined round holes on the surface of the mud. Drew says: "Solenomya lives in rather hard mud, frequently very sandy mud, and, I think, keeps its burrow more or less open." The holes which Mr. Clapp observed and which he cautiously sug- gested might be abandoned worm holes were undoubtedly holes made by Solenomya. Verrill says in regard to S. velum that it is "occasionally found burrowing in the pure fine silicious sand near low-water mark, about two inches below the surface, but its proper home is in shallow wrater beyond low-water mark, and it is perhaps most abundant when there is mud mixed with sand and it also lives in soft mud.' 264 EDWARD S. MORSE. The manner of burrowing is peculiar. In one case an indi- vidual buried itself siphon end downward, and for three days remained in this position with its disk-like foot level with the surface of the mud. In every case the individual rested on the bottom of the dish, ventral region uppermost and valves widely open, limited only by the closed portion of the mantle which is drawn tense by the distended valves. When placed on the surface of sand or mud it soon pushes itself backward by means of its foot which, assuming a pointed tongue-like shape, is thrust forward and downward lifting the anterior end and thus depress- ing the posterior end. Others buried themselves and went to the bottom of the dish, where they remained in a horizontal position. Whether they went head first or tail first was not observed. Others buried themselves in the mud and remained out of sight for hours with no burrows communicating with the surface. Drew states that Nucula delphinodonta buries itself in the mud with no surface communication. The strong alternate movements of the sides of the siphonal area may be related to the habit of burying itself posterior end downward. As a further proof that Solcnomya buries itself siphon end downward may be cited the condition of a young Solenomya borealis preserved in alcohol for dissection. I found adhering to the anterior end of the shell close to the margin three colonies of a 'species of sponge. Here was an evidence that the creature had not only been buried siphon end downward, but that the anterior end had been slightly protruded above the level of the sand and had remained in that position long enough for the accumulation of foreign growth. The periosteum is so exceedingly smooth and polished, and the creature is so active in its swimming habits that the adhesion of foreign growth would hardly be looked for.1 An examination of a number of shells showred little evidence of wear or burial at either end. The pedal opening is so large and the activities of the foot are so incessant that more water is admitted anteriorly than posteriorly and the siphonal opening has little to do so far as conveying water and food to the gill cavity is concerned. 1 Dr. Drew inform? me that in digging this species from the mud many of them float on the surface of the water, the periosteum repelling the water as if oiled. OBSERVATIONS OX LIVING SOLENOMYA. 265 If burrowing posterior end downward is a common habit of Soknomya it forms an exception to all lamellibranchs that burrow. So far as I know all lamellibranchs that burrow in sand, mud, wood or stone penetrate the substance anterior end first, and in that position they rest. Indeed the foot is the primary instrument used in effecting this penetration. Solen- omya is unique in this respect. As before stated the creature, resting on its back, thrusts out its foot in a pointed shape, presses the bottom of the dish, then immediately withdrawing it at the same time expands the fimbriated disk which unfolds in a graceful manner, and in that expanded condition swings back and forth a few times in the pedal opening which is widely distended. The overlapping periosteum then folds abruptly within the shell, which closes at the same time as the thick foot is drawn in between the polished wralls of the periosteum. The creature then falls over on its side and remains in that position until the valves again open which is almost immediately. The method of swimming has been accurately described by Drew. The act consists in thrust- ing out the foot, promptly expanding it and then suddenly with- drawing it, at the same time closing the shell and expelling the water from the siphonal end. These motions are often repeated a number of times without the animal moving at all. When these movements are made with sufficient vigor, however, the animal seems to leap or dart in the water, anterior end forward, going three or four times the length of the shell at each leap. Stimpson says: 'The thinness of the shell enables the animal to make surprising leaps and I have seen it leaping or swimming about the water for some time without touching the bottom. The leap is performed by suddenly drawing in the umbrella- shaped foot, at the same time that water is expelled from the posterior opening by the closing of the valves." I counted the number of darts made by different individuals and, though the specimens I had were probably enfeebled by their long trans- portation in a small glass vessel, I found the following result— 22, 24, 31 and 36 darts respectively were made before the creature fell to the bottom of the dish. The 36 darts were made by a young individual. The darts w'ere vigorous, and wrere made at the rate of 90 to 100 a minute. 266 EDWARD S. MORSE. The violent activity of the animal in its rapid darting through the water, the repeated thrusting out of the heavy pedunculate foot and vigorous closing of the valves accounts for the necessity of the continued absorption of food as indicated by the rapid accumulation of flocculent fcecal matter in the dish. The thirty-nine specimens, young and nearly adult, collected April 27, were placed in a white enamelled pan and showed no signs of increased enfeeblement for a week or more. The water was changed often. At the end of three weeks they were all dead, the young ones surviving the longest. The individuals showed no sen- sitiveness to the obstruction of light, nor did they align themselves in any special way in relation to the lighted window. They did not seem to have the sensitiveness of other lamelli- branchs; the jarring of the table did not cause them to close, though agi- tating the water or touching them with a needle ever so slightly prompted them to close their shells, which remained closed for a few seconds, when they slowly opened and repeatedly thrust out the foot. In fact they became very lively after agitating the water. As one views a specimen from below with the valves wide open and the closed part of the mantle stretched like a drumhead from the margin of the periosteum, the calci- ned portion of the shell is hardly in view. The periosteum, ex- tended as it is when the valves are open, shows no marked shoulder or channel at its junction with the thin membranous mantle, though in section a slight break is seen. The pedal opening is wide and extends backward nearly to the center of the body, where the free mantle edges come together, forming an elongated oval opening when distended, and through this opening FIG. i. View of animal from below. OBSERVATIONS ON LIVING SOLENOMVA. 267 the thick stocky foot is seen occupying the whole space (Fig. i). At the point of junction of the two sides of the pedal opening there are crowded together muscle fibers, producing a silvery appearance, and at this point muscle fibers that run parallel to the borders of the pedal opening — which Drew described as in the nature of sphincter muscles — cross to opposite sides and continue their course posteriorly. May not the origin of the peculiar sharply defined cruciform muscles in Solecurtus, Tagelus, Macoma and Tellina be traced to these separate crossing muscle fibers in some primitive lamellibranch like Solenomya? FIG. 2. FIG. 3. FIG. 4. FIG. 2. Pedal opening showing white area, W. FIG. 3. Side view of animal. FIG. 4. Papilla on margin of pedal opening. The edge of the pedal opening is at times reflected and at its junction in the median line a few blunt papillae or tubercles appear along the edges to nearly half the length of the opening. These are so aligned that when the two edges come together in closing the tubercles interlock. In Fig. 2 the tubercles are seen with the pedal opening nearly closed. They vary in size, are con- tractile and stand out at a sharp angle from the border so that they are distinctly seen when the animal is viewed from the side (Fig. 3). Until nearly adult these tubercles are translucent, but in the oldest specimens they become opaque white, re- sembling white kid, and appear hard and horny until touched when they partly retract. They are covered with a transparent epithelium (Fig. 4). This peculiar whitening not only covers the tubercles but whitens the mantle border from which they spring. A blotch of white is also seen between the tubercles and the keen edge of the periosteum, which at this point is also 268 EDWARD S. MORSE. touched with white. Even the base of the foot behind has a small blotch of white showing only when the foot is greatly elongated. This coloration in all exposed parts of this region is certainly a curious feature. It is as sharply defined and localized as if one with a paint brush had delicately touched with white every exposed part of the animal within this limited area. It is true the external shell is also touched with white at the anterior end. In the young a dot of white is seen in the first three or four transparent interspaces of the free periosteum. In older specimens these white spots are drawn out in broken lines; not in all specimens, however. In older specimens anteriorly the FIG. 5. FIG. 6. FIG. 5. Glands from side of pedal opening. FIG. 6. Glands from anterior mantle margin. whitening is sometimes seen on the dorsal side of each rib. This peculiar coloration is apparently correlated with the attitude of the animal in nature. Resting on its back the ventral region of the animal shows this whitening. Are the white markings at the anterior end of the shell another indication that the creature more often or habitually buries itself siphonal end downward? Along the anterior half of the mantle, as it springs from the edge of the periosteum, are seen a number of what appear to be short white lines with glandular enlargements. They begin a little in advance of where the mantle parts, forming the pedal orifice, and are separated by a considerable space (Fig. 5), becoming crowded toward the rounded anterior border (Fig. 6). They spring from just under the edge of thev periosteum and OBSERVATIONS ON LIVING SOLENOMYA. 269 point in the general direction of the median line, but are disposed irregularly in the membrane, some inclined forward and others backward, but most of them are at right angles to the mantle border (Fig. I, G). Under a higher power these bodies have the appearance of setigerous follicles. Stempell figures one in his memoir of 5. togata as a simple alveolar border gland. His figure shows a much more bulbous gland and the thread-like extension is not shown as in those here figured (Fig. 5). It was some time before I determined that there was no protuberance beyond the surface of the mantle, so strong was their resemblance to setigerous follicles. FIG. 7. FIG. 8. FIG. 7. Anterior view of animal. FIG. 8. Anterior end from below: A, anterior adductor muscle; F, foot in sec- tion; H, hepatic follicles; M, mouth; P, palp-appendages; R, pedal retractors. The margin of the mantle becomes thickened as it borders the anterior end of the shell and ends abruptly just before the shells meet dorsally. Just in front of where the shells meet above are two short unpaired tentacles or tubercles in the median line and then ten or more on the edge of the rounded anterior edge of the mantle on each side diminishing in size, however, toward the ventral region, where they cease (Fig. 7). Writers have described the whole free edge of the mantle as bearing these papillee and even shown them in figures. Looking at this region from below the abrupt termination of the thickened mantle margin and the papillae are shown (Fig. 8). A triangular mus- cular layer seems to connect the two thickened portions of the mantle dorsally. Also a narrow band of fibers runs from the 270 EDWARD S. MORSE. abrupt terminations of the mantle converging centrally. Fig. 8 shows the relation of these various parts in front of the anterior adductor. As the animal becomes enfeebled the mantle (or, what would be a more proper name, the ventral membrane) ruptures, ex- posing the dark-colored gills below. There is a median suture in this ventral membrane and in one rupture the suture became dislocated, showing there was a strain upon it. In another case a small rupture appeared on each side of the median suture. That this membrane limits the expansion of the valves is shown by cutting the membrane, when the valves immediately open FIG. 9. FIG. 10. FIG. 9. Various attitudes of foot, and ruptures in ventral membrane. FIG. 10. Dorsal view of hinge margin. wider in much the same way as when the adductors are severed in other lamellibranchs. Fig. 9 represents the various ruptures, and at the same time illustrates different attitudes of the foot. In the adult the pedal opening extends back nearly half way to the center of the body. In the young it extends a little over a third back. Viewing the shell from above, the exposed ligament presents some curious features which I have tried to represent in Fig. 10. In the vicinity of the umbones a narrow elongated wedge-shaped substance, black in color, is seen, behind which a brown oval ligament appears, split at its posterior end in which is wedged a white substance. From the junction of the two valves anteriorly a narrow wedge-shaped substance narrowing posteriorly is peculiar in being a polished white. In the young OBSERVATIONS ON LIVING SOLENOMYA. 27 1 the periosteum is light yellowish, growing darker as age advances. The hypobranchial gland appears as a mass of globular cells con- taining short rod-like bodies. A number of graceful attitudes are shown in the movements of the pedal disk which I found difficult to figure. The fimbriated edge shows as distinct short tentacles w th grooves running down outside, corresponding to the ten- tacles and corresponding lines or grooves in the disk running to the median line resembling a diminutive actinoid coral Fungia. At other times the edge of the disk appears serrated. These number thirty in all (Fig. n). In the young these resemble long papillae. FIG. ii. FIG. 12. FIG. ii. Dorsal view of pedal disk. FIG. 12. Side view of foot, showing palp-appendage and branchia. The gills are most peculiar in Solenomya. As Pelseneer has shown, there is only one gill on each side having an upward and downward series from a central base. This line dividing the upper and lower series runs from the dorsal anterior portion of the gills to nearly a ventral point posteriorly (Fig. 12). The filaments are distinctly separate and are beautiful in their regularity. The filaments are arranged on the support from which they spring directly opposite each other. Each filament is supported by a chitinous rod in one series, slightly bent at the end — firm yet elastic. The same chitinous framework is seen in the gills of Nucula proximo, and Yoldia limatula, as shown by Mitsukuri, Drew and others. Mitsukuri,7 in describing the chitinous framework supporting the gills in Nucula proximo,, says: "Whether it is really formed of chitin I do not know, but as previous writers have described the substance as of that nature it will be convenient to use the term ' chitinous support' for the present." The gills are colored a light brownish purple, their posterior terminations are free and lighter colored (Fig. 13). 272 EDWARD S. MORSE. A view of the posterior end of the animal reveals a single long siphonal opening which often changes its outline, at times being an oblong, oval slit, but usually the upper part, corresponding to the anal opening, is much larger, giving the aperture a double- gourd-shaped form, the upper bulb being larger. Above the anal region, just where the two shells meet, is an unpaired tubercle, or tentacle which is contractile. Stempell figures a similar FIG. 13. Termination of branchia. Q -\^-wr- C FIG. 14. Siphonal opening: .4, B, from above; C, D, from behind; E, from behind at an angle. tubercle in 5. togata, and in another form to be described later a very long tentacle is seen. Just below this tentacle a pair of blunt tubercles arise which are very contractile and appear at times as two and even three pairs of tubercles. The sides of the anal opening are without tubercles. The narrow branchial opening is bordered on each side by from six to eight short tubercles or papillae hardly varying in size. These are also contractile. A few attitudes of the siphonal opening are shown in Fig. 14. A view of the siphonal opening from the side shows no pro- trusion as in the siphonal expansion of siphonated lamelli- branchs. Indeed there is no appearance of a siphon, simply an OBSERVATIONS ON LIVING SOLENOMYA. 273 opening with papillae about it and these quite unlike the long pointed papillae of the siphonated lamellibranchs or even the truncate papillae of Saxicava, though morphologically they are the same. On viewing this region from above the papillae appear bunched together or spread apart, the result of the sudden dilatation or partial contraction of the opening. The sides of the siphonal opening are in constant motion in and out, though this motion is alternate and rhythmic. The proboscidiform foot swings from one side to the other within the shell, like an elephant's trunk and this motion may be related to the alternate siphonal movements; this correlation, however, was not ob- served. Just below the siphonal opening a slight depression is seen on the mantle on each side connected by a transverse depression forming a symmetrical figure. Stempell shows a marking of the same nature in S. togata in the form of an angle with the apex down- ward. A view of the animal as it rests in the usual position on its back with valves widely apart will reveal the attitude and behavior of the palpi (Fig. 15). Cutting the ventral membrane will give a clearer view of these parts. The palpi are adherent for a short distance to the side of the foot but are free beyond, and extend backward and downward. These organs are long, slender, semi-tubular with ends much larger and the whole structure delicate and diaphanous. No line or angularity in section indicates in any way that this semi-tubular appendage has come about by the adhesion of two palpi. The palpi of other lamellibranchs are more or less triangular in shape with their upper edges partly united. In Solenomya velum, the long tube-like appendages encircle the base of the foot with their dilated extremities resting directly on the compact sides of the B FIG. 15. Anterior end from below: .4, anterior adductor muscle; B, branchia; F, foot! P, palp-appendages resting on branchia. 2J4 EDWARD S. MORSE. gill plates, and in this position are in gentle and constant motion with occasional side jerks. After watching these browsing movements for awhile one becomes convinced that the behavior of the ends of the palpi is that of feeding. Even the peculiar jerks at times suggest that some larger morsel has been caught. The protean shapes the appendage assumes are shown in Fig. 16. The strong ciliary action of the gills is continually sweeping along particles of organic matter toward the feeding appendages which gather the stuff and convey it by ciliary action to the mouth. The movements of these appendages may be distinctly seen through the ventral membrane and at no time are they at rest. The continual ingestion of food is in- dicated by the great quantity of excreta which is voided. In this connection it is interesting to note that Drew6 in his memoir on the life habits of Yoldia limatula, another member of the Proto- branchia, describes long appendages to the palpi which extend backward beyond the posterior end of the shell and rest on the sand. These appendages are semi-tubular and being ciliated "rapidly elevate the mud which is full of living organisms and finally pass it between the palpi," and thence, of course, to the mouth. Mitsukuri, having described the palpi in Nucula proximo,, says: "At their posterior end there are two remarkable struc- tures. One of these is a hood-like structure which is the posterior prolongation of the united upper edges of the inner and outer palpi. The other, lying immediately below the first, is a long tentacular appendage. It is a hollow tube, open, however, along a line on its posterior aspect, and having its cavity con- tinuous with the space between the two palpi. As it has been seen protruded with the foot outside of the shell, and since, in alcoholic specimens, a great deal of dirt and sand is found along its length and between the palpi from its base to the mouth, it is no doubt a food-procuring organ, probably sending a constant stream of nutritive matter to the mouth by means of its cilia. It is interesting to notice in connection with this appendage that in Nucula, the gills, unlike those of ordinary lamellibranchs, must be practically useless for obtaining food, as will be evident OBSERVATIONS ON LIVING SOLENOMYA. 275 from the following description of them," etc. Drew4 finds these palp-appendages in Nucula delphinodonta functioning as food collectors. The creature is immersed in mud out of sight and only when placed in shallow mud was he able to observe its behavior. The gills are certainly very small in Nucula. In Solenomya, however, the gills are very bulky, filling nearly half the mantle cavity, are highly ciliated and as we have seen the palp-appendages rest directly on their anterior surfaces. There can be no question that these appendages in Solenomya are strictly homologous with the palp-appendages in Yoldia, Nucula and allies. The gills, however, are the food accumulators from which as we have seen the palp-appendages, collect the material for nutrition. The mouth is difficult to make out in the living creature. In two instances I have observed a slightly brownish line marking the position of the mouth. While there are many features in common between Solenomya and the other members of the order Nuculidse as seen in the posterior position of the umbones, the primitive gills, the palp- appendages, the fimbriated disk-like foot, the highly polished periosteum; the absence of true palpi in Solenomya might be considered an evidence that this peculiar form stands lowest in the order Protobranchia. Some years ago Miss Newrell discovered at Annisquam a colony of what was supposed to be Solenomya velum. A few wrere brought home alive, but circumstances were not favorable at the time for detailed study and only a fewr drawings were made. The shell is apparently identical with that of S. velum herein described, yet the character of the siphonal opening or rather the appendages surrounding that aperture are widely different. The single siphonal orifice is ano-branchial as in the other species of Solenomya. Above this orifice is a very long unpaired tentacle with broad base. On each side of the anal region is a broad flap bearing five tentacles, the lowest one small, the next one long slightly curved upward with three minor twigs springing from its side, the next three smaller and diminishing in size. Below this process is a wider flap supporting nine tentacles of nearly uniform length. Between the anal and branchial processes viewed from the side is a white band! I hesitate giving these 276 EDWARD S. MORSE. details for fear some systematist will instantly announce a new genus or family. Nevertheless, no figure nor description of Solenomya yet published has shown anything approaching these remarkable processes. The unpaired dorsal tentacle is short in .S. togata (Fig. 18), even shorter in 5. velum (Fig. 19), while in I his form (Fig. 17) it is exceedingly large with a broad base. The FIG. 17. Solenomya sp.? Annisquam. FIG. 1 8. Solenomya togata. FIG. 19. Solenomya velum. pedal opening shows the same tentacles starting from the pos- terior end of the opening. The same alternate movements of the siphonal opening were observed. The shell as before re- marked differed in no respect from that of S. velum, and I am puzzled at the remarkable difference in structure of the siphonal opening. The question arises whether the Annisquam form may not be the young of 5. borealis. This can only be decided by securing a living specimen of this rare form. Professor Drew,5 in his observations on the habits, anatomy and embryology of members of the Protobranchia, alludes to the great diversity in structure and concludes that the proto- branchia have been derived from a primitive type. The re- semblance he finds in certain stages of the embryo of Yoldia and OBSERVATIONS ON LIVING SOLENOMYA. 2J7 Nucula to those of certain other mollusks is interesting in the fact that these mollusks are in every case low and aberrant forms. He says: "The most striking peculiarities in the develop- ment are connected \vith the formation and disappearance of the tests. Outside of the group, so far as I have been able to learn, Dondersia is the only other mollusk whose embryos are known to be provided with similar tests." Dondersia is now regarded as belonging to the Amphineura, but for awhile was looked upon as related to the worms. Drew remarks that the young embryos of Dentalium bear certain resemblances to the embryos of Dondersia, Yoldia and Nucula. Here again a resemblance is seen to an aberrant group of mollusks wThose affinities were for a long time obscure. Drew also says that a somewhat similar resemblance is noticeable in the case of the embryo of Patella. Again a resemblance is found to a group whose characters are archaic. Dall, in speaking of the Doco- glossa, says the various forms manifest what may be termed a peculiar persistency of immaturity when compared with other groups of gasteropods. Korschelt, Patten, Fisher, Lankaster, Pelseneer and others testify to the primitive characters of Patella. In my paper on "An Early Stage of Acmcea," I hava collected a number of extracts from the above authors in regard to these low characters. Pelseneer agrees with Stempell that the characteristic features of Solenomya represent the oldest living group of the lamelli- branchs. He calls attention to the absence of a protractor pedis, extensive overgrowth of ventral mantle edges, peculiar develop- ment of the branchi-anal siphon, the form of shell and absence of interlocking teeth, with position of ligament, remarkable overlapping of periosteum, prismatic structure of limy shell, rudimentary form of mouth lips and mouth appendages, almost without winding of intestines, excessive elongation of the ven- tricle, the presence of pericardial glands, position, size and structure of gills, separation of kidneys without cross com- munication. Interesting features will doubtless be revealed when the embryology of Solenomya shall have been made known, and with the abundance of one species on our coast, S. velum, we hope before long light may be thrown on the subject. 278 EDWARD S. MORSE. APPENDIX — NOTES ON LIVING Solenomya borealis. Since the above observations were made on Solenomya velum, and the matter in type, the opportunity has been given me of studying living specimens of the large species Solenomya borealis. My friend, Major John M. Gould, collected a number of living specimens of various sizes from a young one, measuring 9 mm., to full-grown individuals, measuring 83 mm. These were obtained from dredgings in six to seven fathoms of water by a huge dredging machine engaged in deepening a channel in Portland harbor, Maine. The work of collecting, as may be imagined, was done under desperate conditions. Jumping in between avalanches of mud and water, glancing, raking and jumping out again! In this rapid recorinaisance Mr. Gould observed that the creatures were buried in holes and got the impression that they were buried posterior end downward, the broad, light-colored disk-shaped foot projecting from the holes. This attitude, as before remarked with regard to S. velum, is contrary to the behavior of all lamellibranches that bury them- selves wholly or partially. I am still in doubt about the matter. It is interesting to note, however, that when the creature was placed on coarse material it behaved in precisely the same way as 5. velum. In every case resting on its back it thrust out its foot, raised the anterior end of the body as if endeavoring to thrust the posterior end into the mud. In this connection, it is interesting to observe that the anterior ends of the shells of older specimens were covered with films of slime while the posterior ends were clean and polished. In only one instance has the animal been seen to swim or dart through the water though both young and old were specially observed for this behavior. The foot was often thrust out as in S. velum, but not so often nor with such energy. No lateral swing of the foot was observed. The periosteum appeared oily as in S. velum and repelled the water, and Dr. Drew informed me that this feature was so marked in 6". velum that when he dug them from the mud they sometimes floated on the water. The character of the siphonal end upon which marked specific differences will be established was quite different from S. velum and it can now be positively stated that 5. velum and 51. borealis OBSERVATIONS ON LIVING SOLENOMYA. 279 are distinct species. It is surprising, however, to see how closely they resemble one another (Fig. 20). At first sight S. borealis seemed to be an enlarged S. velum. The color of the foot and ventral membrane varies. In 61. borealis the foot is a light brown with a tinge of purple, darker on the ventral keel, with a darker blotch near the ventral portion. The papillae on the edge of the FIG. 20. Side view of 5. borealis, natural size. pedal disk were longer and sharper than in S. velum though the same in number. The tubercles on the edge of the mantle open- ing were the same in number and position, standing out at right angles from the ventral membrane. Their color, like the mem- brane, was brownish, the tips of the tubercles being dark brown. No trace of a white pigment-like color on ventral membrane FIG. 21. FIG. 22. FIG. 21. Siphonal opening. FIG. 22. Anterior end from below: F, foot; P, palp-appendages resting on the branchia; B, branchia. or tubercles was observed in either young or old. The palpi were not so long though the ends were more foliated (Fig. 22). Their movements were not so active. The palpi rested on the anterior portion of the gills and were evidently engaged in secur- 280 EDWARD S. MORSE. ing food as described in S. velum. Alternate movements of the side of the siphonal opening were observed as in S. velum. The anterior aspect of the creature was much the same, the two median dorsal tubercles, one behind the other, and the anterior edge of the mantle bearing papillae. The posterior aspect was quite different from S. velum and resembled more closely S. togata of the Mediterranean as figured by Stempell. The resemblances are seen in the large size of the median dorsal tubercle, the few small tentacles in pairs, succeeded by a pair of very long tentacles, then a number of very minute tentacles bordering the upper edges of the siphonal opening followed by a number of tentacles increasing in size to the lower portion of the opening (Fig. 21). All of these tentacles were retractile, the upper series appearing like round tubercles when contracted- The longer tentacles surrounding the orifice were in constant action, a bcnding-in movement as if grasping. In the collection of fifteen specimens of S. borealis the indi- viduals varied in length from 9 mm. to 83 mm. Omitting the smallest specimen the others were easily arranged in four distinct series of sizes as follows : ist Series. 2d Series. 3d Series. Full Grown. 24 mm. 39 57 73 29 mm. 40 59 77 31 mm. 44 77 31 mm. 78 83 Twenty-eight specimens of S. velum from Sandwich were in the same way readily grouped in series. One measured 7 mm. in length, 6 averaged 10 mm., 8 averaged 12 mm., 5 averaged 13 mm., 6 averaged 15 mm. and 3 averaged 19 mm. From this one might infer that Solenomya is not an annual. In conclusion I wish to say that this brief examination of the external features of S. borealis has convinced me that my sup- position that the Annisquam individual might be the young of S. borealis is incorrect. Furthermore that a study of a more advanced specimen of S. velum, at Woods Hole showed a nearer approach to the Annisquam specimen. I have never seen a minute specimen of 6*. velum and Professor Drew informs me that he has never found one with eggs. The protoconch of Solenomya OBSERVATIONS ON LIVING SOLENOMYA. 28l will show interesting features and the embryology of this low form will be of great importance. EXPLANATION OF FIGURES. Note. — Only the largest specimens averaging 19 mm. were studied. The various degrees of magnification of the figures here given may be understood in a general way by realizing that the entire length of the animal was 19 mm. BIBLIOGRAPHY. 1. Brooks '74 On an Organ of Special Sense in the Lamellibranchiate Genus Yoldia. Proc. Am. Asso. Adv. of Sci., 1874. 2. Deshayes '44~'48 Histoire naturelle des Mollusques. (Exploration scientifique de 1'Al- gerie.t Paris, 1844-1848. 3. Drew 'oo Locomotion in Solenomya and its Relatives. Anat. Anzeiger, XVII. band, No. 15, 1900. 4. Drew '01 Life History of Nucula delphinodonta. Quart. Journ. Micr. Sci., XLIV., 1901. 5. Drew '99 Some Observations on the Habits, Anatomy and Embryology of Members of the Protobranchia. Anat. Anzeiger, XV., 1899. 6. Drew '99 Yoldia limatnla. Mem. Biol. Labor. Johns Hopkins, Univ., IV., 1899. 7. Mitsukuri '81 On the Structure and Significance of Some Aberrant Forms of Lamelli- brancheate Gills. Quart. Journ. Micr. Sci., XXL, 1881. 8. Pelseneer '91 Contribution a 1'etude des Lamellibranches. Arch, de Biol., XL, 1891. 9. Perkins '69 The Molluscan Fauna of New Haven. Proc. Soc. Nat. Hist., Vol. 13, 1869. 10. Philippi Uber das thier von Solenomya Mediterranea. Arch. Naturg., Jg. I, Vol. I. 11. Stempell 'oo Zur Anatomic von Solenomya togata. Zool. Jahrb. (Anat. u. Ontog.)^ XIII., 1900. A CHEMICAL SIGN OF LIFE.1 (From the Laboratory of Biochemistry and Pharmacology, The University of Chicago; and Marine Biological Laboratory, Woods Hole, Mass.) SHIRO TASHIRO. From the earliest time to the present, our criterion of life has been connected with the changes which are brought about by death. The locomotive power of living matter is one of the first things that disappears when it dies. Thus the idea of move- ment erroneously led the simple minds of our ancestors to believe that wind, fire, thunder and water had life. By close study of the locomotion of living matter, however, we have gradually traced this part of living phenomena to irritability of tissue— the property characteristic to living tissues only. This proto- plasmic irritability is not only the potential of living matter, to give rise to physical changes when a stimulus is applied to it, but the fundamental and the only characteristic which is common to all the living as long as they possess the power to perform their own functions. And it is this property, according to Professor Mathews,2 that not only is the most probable point of attack of natural selection but is also one of the main factors which determines phylogenetic development of organisms. The presence of this irritable property in tissue — the universal sign of life — cannot easily be determined in all living tissues. The most common way of determining irritability is the physical changes brought about by a stimulus. This physical change, however, is not an unfailing indication of a response of the tissue against stimulation. Several tissues in the animal, such as nervous tissues, and an abundance of examples in the plant kingdom, do not manifest at all any visible change, when stimu- lated. Not only those living do not show such mechanical 1 I wish to thank Dr. F. R. Lillie, through whose kindness I was occupying a table in the Marine Biological Laboratory, Woods Hole, Mass., at the time when the apparatus for these experiments was made. 2 Mathews, Amer. Naturalist, XLVII., p. 94, 1913. 282 A CHEMICAL SIGN OF LIFE. 283 changes while functioning, but several non-living matters constantly undergo physical changes when disturbed by external conditions. Therefore it has been customary for the biologist to decide whether or not matter is living by several other ac- companying changes during the performance of a function. Such changes have been measured by rise of temperature, pro- duction of CO2, histological variation before and after the stimulation, and electrical response. None of these functional changes when taken individually can be considered as a specific sign of life. Although Herzen claims that under certain conditions of local narcosis, the nerve fiber may give an action current although no muscular contraction follows, and O. B. Ellison recently demon- strated by the use of cinchonamine hydrochloride the absence of negative variation without abolishing excitability, yet according to Waller the presence of life can be demonstrated by an electrical change. In his admirable book on the "Signs of Life,"1 he states that chemical change is a sign of life and an electrical change is a sign of a chemical change; therefore the electrical change is the sign of life. It is very interesting to note in the case of a dry seed, he could measure equally well quantitatively the different electrical changes according to the different ages which characterize the different degrees of vitality of the seed, in spite of the fact that he could not detect any chemical change which usually produces CO?. He concludes, however, that it Is possible, or rather certain that our method of chemical investigation is not refined enough to reveal to us the smallest and most infinitesimal change that may be going on in apparently dry or dormant seed. It was this conclusion of his that sug- gested to me the desire to make an inquiry to ascertain whether or not I could find some new method by which an easy sign of life may be observed. When I had constructed a new apparatus,2 which can detect COa as small as o.ooooooi gram, the measurement of irritability became much simpler, for with it we established a few new facts which deal with the fundamental nature of protoplasmic irrita- bility. The idea that irritability in general is closely associated 1 Waller, "Signs of Life," New York. 2 Am. J. of Physiol., XXXII, p. 137, 1913. 284 SHIRO TASHIRO. with chemical phenomena has been set forth long ago. Pro- fessor Mathews, in his paper on "Animal Oxidation," expressed the idea that cellular respiration must be the fundamental process of all organic activity around which all functional phenomena are intimately connected. Later in his study on the action of ether and other anaesthesias on an anaerobic tissue he confirmed his idea that all the substances that affect irritability must neces- sarily attack the tissue respiration first. With the aid of the new apparatus, the direct evidence for his hypothesis has been brought forth in connection with the study of metabolism of the nerve fiber, in which three fundamentally interesting facts have been established. In the first place, the most excitable tissue of all the protoplasm, the nerve fiber, is constantly undergoing chemical changes, giving off CCV In the second place, when this tissue is stimulated COa production is greatly accelerated, giving more than double the amount. Finally, the rate of COa is greatly influenced by conditions such as anaesthesia, which are known to affect tissue irritability, showing a direct relation between respiration and excitability.1 Before I had concluded, from these facts, that all the irritable tissues must respire and should give off more COo when stimu- lated, a crucial experiment was done on a dry seed. Waller's conclusion has been fully realized. A living dry seed not only gives off COo quantitatively proportional to its weight, but also gives off more COo when stimulated, a phenomenon true to living seeds only. These facts enable me to propose a new sign of life, namely, a chemical sign. The criterion is simple. If we are given a tissue which gives more COz when stimulated, the tissue must be alive; it is excitable. A discovery of a remarkably simple method of stimulation made this chemical sign of life much more easily practicable to all living tissues. Simple mechanical crushing of the living tissue is the new method. I have already argued elsewhere2 that the phenomenon that the nerve gives off more COz when crushed, is due mainly to an extreme stimulation and that it is characteristic of living, excitable tissues only. Therefore without any attempt to settle the question as to how CO2 production is increased, we can use 1 Am. J. of Physiol., XXXII., pp. 107-136, 1913. 2 Am. J. of Physiol., XXXII., p. 121, 1913. A CHEMICAL SIGN OF LIFE. 285 this method as a means to denote protoplasmic irritability, as long as we can abolish this sign of life by rendering the tissue unexcitable. The nerve, dry seeds, including wheat, oats, rice and several other living tissues from animal and plant kingdoms, if treated with ether, or killed, can never be made to produce more CO2 by crushing, whereas, a tissue however little CO2 it may produce normally will give off more CO2 when crushed, provided it is living. From these general findings, I conclude that this method can be used to detect vitality of protoplasm under normal conditions. The details of this simple means of finding the sign of life are as follows: The biometer is used. This is an apparatus which has two respiratory chambers, each furnished with a cup in which a drop of barium hydrate can be introduced. With it, the comparative outputs of CO2 as small as o.ooooooi g. from the two tissues can be estimated simultaneously.1 Two dry living kernels of the same seed, with equal weights, are chosen. One is placed in the right chamber, and the other is crushed and placed in the left. After the necessary cleaning of the apparatus with CO2- free air, a drop of the barium hydrate is introduced upon each cup in both chambers. By watching the drop, it would become obvious that the crushed seed is giving off more CO2 than the uninjured, as indicated by the speed of formation and quantity of precipitate of the barium carbonate. Not only such distinc- tion between two can be observed, within a few seconds, in the case of the tissue, which normally gives off a comparatively large quantity of CO2, but also dead tissue whether or not it gives off CO-2 without crushing, will never produce more CO2 when mechanically smashed. In other words the phenomenon of production of more CO2 by crushing is characteristic only of a living tissue. By killing the seeds, and repeating the experi- ment, such a conclusion can easily be confirmed. With these facts, I am proposing a new sign of life, namely, a chemical sign of irritability. It is a measurement of CO2 due to stimulation. It is not exhalation of CO2 which is character- 1 The detail of this apparatus is described in Am. J. of Physiol., XXXII., p. 141, 286 SHIRO TASHIRO. istic of life, for there are many dead matters which give off CO? continuously, but it is the increase of CO2 production when crushed which is certainly a phenomenon associated with excit- able tissue only. Hence, when a tissue gives off more CC>2 when mechanically crushed, or injured, it is a sure sign that this tissue is living. f Vol. XXV. October, 1913. No. 5 BIOLOGICAL BULLETIN SIZE DIMORPHISM IN ADULT SPERMATOZOA OF ANASA TRISTIS.1 E. C. FAUST. I. INTRODUCTION. The work on which this paper is based was undertaken with the object of determining the existence or non-existence of size dimorphism in the mature spermatozoa of Anasa tristis. Sper- matogenesis studies of this form have shown that one half of the spermatozoa receive one more chromosome each than the other half. A corresponding difference in size of the mature sper- matozoa may therefore be expected. The study was carried on during the academic year 1912-1913, under the direction of Professor Charles Zeleny, to whom my thanks are due for his many valuable suggestions. II. MATERIALS AND METHODS. I. General Account. — The material used in this study was collected from two localities, Urbana, 111., and Bradford, N. H., during the month of October, 1912. The testes of these indi- viduals were teased out in normal salt solution and made up into temporary and permanent smear preparations. Smears of adult spermatozoa were prepared by the aceto- carmine method and by the osmic-haematoxylin method. In addition intra miam staining was tried, but without success. The first mounts were made up as aceto-carmine preparations. Although this affords an excellent preparation for the study of spermatogenesis, it is not a good method to use in the study of adult spermatozoa. Curling is frequent. Distortion by swelling 1 Contributions from the Zoological Laboratory of the University of Illinois, under the direction of Henry B. Ward, No. 23. 287 288 E. C. FAUST. occurs more than is desirable. In the head of the spermatozoon no inner chromatic rod is differentiated from the cytoplasmic envelope. Moreover, it often takes several days to measure a considerable number of individuals, so that temporary prepara- tions, such as those obtained with aceto-carmine, are objection- able. For this same reason intra vitam staining was abandoned. For the most of the preparations Delafield's haematoxylin was used and proved much more successful than Schneider's aceto- carmine. The testis was taken out of the individual and placed in normal salt solution. It was then transferred to a slide that had been previously coated with a thin film of albumen fixative. The spermatozoa were teased out on the slide and examined in a drop of some fluid, such as Ringer's, to determine their motility or non-motility. After most of the solution had been allowed to evaporate so that only a bit of moisture was left around the spermatozoa, the preparation was fixed in osmic fumes for thirty seconds and then stained in Delafield's haematoxylin. This method was found to give a maximum number of straight spermatozoa with an inner chromatic rod well differentiated. The microscope used in the measurements was equipped with a Leitz 2 mm. oil immersion objective and Zeiss compensating oculars. A large number of individuals was measured and a curve of variability plotted. In every individual it was the length of the inner chromatic rod which was determined. A description of this chromatic rod occurs later under the descrip- tion of a mature spermatozoon. 2. Sources of Error. — The probable sources of error that need special consideration are (i) imperfect technique in preparations; (2) immaturity of the spermatozoa; or (3) incorrect measure- ment. (i) Error Due to Imperfect Technique in Preparations. — Several smears of mature spermatozoa were prepared by the aceto- carmine method. Measurements made from them were not uniform, due to faulty technique. Many of the sperm heads had been curled or shrunken in fixation. Lack of an exact tip or base of the chromatic rod could have given rise to an error of from 1^1 to 5/x in measurements of length. Distention and distortion were common. On account of the faulty prepara- DIMORPHISM IX SPERMATOZOA OF AXASA. 289 tions, the results obtained from them have not been included in the data. In the aceto-carmine preparations it was found that drawing off the excess stain by a bit of filter paper as recommended, carried along with it many spermatozoa. In fact, preliminary measurements of such preparations showed a prevalence of larger spermatozoa on the side toward which the drawing was being made. Precautions were taken to eliminate this fault by fixing the spermatozoa to the slide with a thin film of albumen fixative. Also osmic fumes were used in preference to a liquid killing agent and Delafield's hsematoxylin, instead of some more complicated stain, in order that all the spermatozoa might be preserved. This osmic-hsematoxylin combination has been found particularly effective in producing straight spermatozoa with well-marked chromatic elements. In the preparations from which data were compiled from ninety-five to ninety-seven per cent, of the spermatozoa were sufficiently straight for meas- urement. (2) Error Due to Immaturity. — When preliminary mounts were made about the first of October, 1912, the material was found to be too immature for the purpose of this work. All stages in spermatogenesis were present from the spermatogonium to the young spermatids, but no adult spermatozoa were found. One of the problems confronting the writer in determining suitable preparations for study was the recognition of a truly mature spermatozoon. In general, the attenuated character of the head, with the chromatic rod (nuclear element) extending almost its entire length, is sufficient for the recognition of a mature spermatozoon. On that basis all preparations wrhere more than five per cent, of the individuals were immature were discarded. In order to satisfy himself more fully of the maturity of the individuals measured, the author examined all of them in Ringer's solution, which demonstrated their motility. In this fluid the developing spermatids and immature spermatozoa were quiet, while the fully ripe spermatozoa were active for a con- siderable length of time. However, motility is not to be taken as a general criterion of maturity. In the house-fly, Musca domestica, for example, immature forms are motile. Again, these 290 E. C. FAUST. spermatozoa were similar in all respects to those obtained at Woods Hole in August, 1913, from adult males of this species. Because of their structure and specific reaction the author is assured that all spermatozoa entering into consideration in the data are mature individuals. The description of a mature spermatozoon becomes necessary at this point. As far as the writer has been able to discover no one has described a chromatic rod writh cytoplasmic sheath for the mature spermatozoon of A nasa tristis. Paulmier's description and figures show no such structure. Meek's figure for the ' ' penultimate stage" of maturing spermatozoa of Stenobothrus viridulus shows something of the nature of this chromatic rod (Plate III., Fig. 36). FIG. i. Mature spermatozoa, (a) Entire spermatozoa; (b) straight sperm- head, the inner chromatic rod of this type was measured; (c) curved sperm-head, a characteristic type not measured. The author's preparations agree with Paulmier's description of the mature spermatozoon olAnasa tristis in having no middle piece dis- tinguishable, so that the tail connects directly with the posterior end of the head. The head has the chromatic elements com- pacted into a long attenuated rod of nearly constant diameter throughout its entire length. This chromatic rod is enveloped by a cytoplasmic capsule. The latter is also attenuated, and has a constant outer diameter. The length of the chromatic rod ranges from 24^ to 36^ and the width from ^ n to f /JL. Although the width is too minute for practical measurement, in general, it seems to be proportional to the length. The cytoplasmic DIMORPHISM IN SPERMATOZOA OF ANASA. 291 sheath is two to three times as thick as the chromatic rod. It extends 2n to 3/z anterior to the anterior end of the chromatic rod, and is continuous posteriorly with the sheath of the axial filament. Considerable observation has shown that the fre- quency curve for chromatic rod length does not differ from the Frequency. Length in M. 22.68 23.76 24.84 25.92 27.00 28.08 29.16 30.24 31.32 32.40 33-48 34-56 35-64 FIG. 2. Bimodal curve of variability of chromatic-rod length for ir. Greater number of spermatozoa grouped around the upper mode,[M". Frequency. Length in n. 22.68 23.76 24.84 25.92 27.00 28.08 29.16 30.24 3i.32T32.4O 33.48 34.56 35.64 FIG. 3. Bimodal curve of variability of chromatic-rod length for il. Greater number of spermatozoa grouped around the lower mode, M'. curve for the entire head. Unlike the adult spermatozoa described by Paulmier ('99) and Stevens ('05), the author's preparations show a distinct region of cytoplasm around the chromatic rod, and fail to demonstrate that the sheath is entirely, or even mostly "used as food material for the growth of the tail sheath," as Paulmier described it (p. 254; also Figs. 56 and 57). The author's haematoxylin preparations show no acrosome or E. C. FAUST. centrosome such as Paulmier described for this form. This difference might be claimed to exist because of the different staining capacities of Delafield's hsematoxylin, used in the author's preparations, and iron-alum hcematoxylin which Paul- mier employed. However, Meek's preparations of the "ante- penultimate stage" of Stenobothrus, which were stained with the iron-alum haematoxylin, show a distinct chromatic rod and sheath, but no acrosome and a diminishing centrosome. Frequency. Length in M. 22.68 23.76 24.84 25.92 27.00 28.08 29.16 30.24 31.32 32.40 33.48 34.56 33.64 FIG. 4. Bimodal curve of variability of chromatic-rod length for ir and il combined. Number of spermatozoa around both modes approximately equal. Almost bilaterally symmetrical. Certain spermatozoa of unusually large size were observed. These are the so-called "giant spermatozoa." Such "giant spermatozoa" have been described by Henking ('91), and Wilcox ('95) f°r various species, and by Paulmier ('99) for Anasa tristis. Paulmier described those of "double and quadruple" the normal size, and attributed the phenomena "to the non-completion of one or both of the spermatocyte divisions." The author has observed only the double-sized forms. They are found in about one per cent, of the measurements. These spermatozoa are about one and one fourth times the length of the average normal spermatozoon, and contain about twice the amount of chro- DIMORPHISM IN SPERMATOZOA OF ANASA. 293 matic material of the normal forms. Smith's "giant sper- matozoa" of hybrid pigeons, which are twice the normal length, do not appear to belong to this category. In a few cases the author has observed two mature chromatic rods of normal length within a single cytoplasmic membrane. Evidently these were cases where the chromatic elements had divided normally, but Length in M. 22.68 23.76 24.84 25.92 27.00 28.08 20.16 30.24 31.32 32.40 33.48 34.56 35.64 35.72 FIG. 5. Bimodal curve of variability of chromatic-rod length for 2r. Greater number of spermatozoa around the upper mode, M". Frequency. 45 40 35 30 25 20 IS 10 Length 'in m. 22.68 23.76 24.84 25.92 27.00 28.08 29.16 30.24 31.32 32.40 33.48 34-56 35.64 FIG. 6. Bimodal curve of variability of chromatic-rod length for 2!. Greater number of spermatozoa around the lower mode, M'. where the cytoplasm had failed to divide, so that the two chro- matic elements had developed together side by side within the same sheath. It is interesting to note that such "twin" sper- matozoa are always of different lengths, varying from 1.5,11 to 3/1 from each other. "Giant spermatozoa" of this same type have been observed by the author in smear preparations of the house- fly, Musca domestica. 294 E. C. FAUST. The " giant spermatozoa ' ' from the author's preparations had no definite range of variability but blended into the upper reaches of the curve for the normal spermatozoa. The most extreme " giant forms," as well as those that are less extreme, might be considered as wide variants in the frequency distribution of normal forms. If the average normal length is taken, and a "giant spermato- zoon" of twice that chromatic volume is computed, then such a form would still fall within the upper reaches of the normal frequency distribution. It is evident, then, that such forms Frequency. Length in M. 22.68 23.76 24.84 25.92 27.00 28.08 29.16 30.24 31-32 32.40 33-48 34-56 35-64 36.72 FIG. 7. Bimodal curve of variability of chromatic-rod length for 2r and 2! combined. Number of spermatozoa around both modes approximately equal. Almost bilaterally symmetrical. can be considered as "giant spermatozoa" only when an arbitrary upper limit is placed to the normal frequency curve and these wide deviants considered as twice the volume of average sper- matozoa. (3) Error Due to Incorrect Measurement. — It is evident that, for the purpose of'measurement, the inner chromatic rod must be well differentiated both anteriorly and posteriorly. Aceto- carmine preparation failed in this respect, while the Delafield's DIMORPHISM IN SPERMATOZOA OF ANASA. 295 brought it out clearly. Because the posterior limit of the chromatic rod is more easily located than the anterior limit, all measurements were made from the posterior end forward. Another item to be taken into consideration is that the sper- matozoon may not always be in a perfectly horizontal position on the slide, and hence all parts of it may not be in focus at once. Length in n. 22.68 23.76 24.84 25.92 27.00 28.08 29.16 30.24 31.32 32.40 33.48 34.56 35.64 36.72 FIG. 8. Bimodal curve of variability of chromatic-rod length for 3r. Greater number of spermatozoa around lower mode, M'. Length in M. 22.68 23.76 24.84 25.92 27.00 28.08 29.16 30.24 31.32 32.40 33.48 34-56 35-64 36.72 FIG. 9. Unimodal curve of variability of chromatic-rod length for 3). M" the only mode. This may be due to a downward or upward bending or curving of the anterior end of the head, or to a tilting of the entire head, whether curved or straight. To preclude any danger of incorrect measurement all such spermatozoa were recorded separately and only entered in the data with those which were distinctly crooked. 296 E. C. FAUST. The total of all spermatozoa, curved, crooked and tilted, which were not measured amounted to three to five per cent, of the numbers measured. Had this group been all of a critical length they would not have been sufficient in a single case to change the fundamental character of the curve of variability. An example of such a "curved" sperm-head is shown in Fig. I, c. A question arose as to how many spermatozoa must be meas- ured before the curve is fairly constant and dependable. It has been observed that with one hundred fifty or two hundred indi- viduals the main features of the curve are outlined; that the curve fluctuates in its details up to three hundred individuals; that four hundred individuals are enough to give a curve of variability which is accurate and dependable. Even a measure- ment of seven hundred ten individuals, which was made in one case (number ir), failed to change the character of the curve from that secured for the first four hundred of the seven hundred ten individuals. In addition to these precautions, several other checks were made in measurement. One preparation, number ir, was measured three times with three different magnifications; the curves for all three measurements were similar. All measure- ments were made with the writer's same eye, but were checked by other research students in the laboratory. Finally, only the best of light was used while the measurements were being made. III. DATA OBTAINED. The data presented were obtained from eight smear prepara- tions taken from four pairs of testes, fixed and stained by the osmic-Delafield method described above. These preparations are designated as I right (ir) and I left (il), 2 right (2r) and 2 left (2!), 3 right (3r) and 3 left (3!), 4 right (4r) and 4 left (4!). Taken as a whole the measurements show very clearly that there is a distinct size dimorphism in the spermatozoa. Individual testes, however, differ considerably from each other in details. Preparations ir and il. — These spermatozoa when examined in Ringer's solution were very active. From one thousand to two thousand individuals were present on each slide preparation, comprising the total number of spermatozoa from each testis. DIMORPHISM IX SPERMATOZOA OF ANASA. 297 With ocular no. 8, seven hundred ten individuals of ir were measured. This number has been shown to be sufficiently large to plot a reliable frequency curve. Spermatozoa selected were taken as random samples from all parts of the slide. The variability curve is shown in Fig. 2. It shows two distinct modes at the lengths 28.08^1 and 30.24^, with a point of depression midway between at 29.16^. Not only are the two modes equally distant from the intermediate low point, but the entire curve is Frequency. Length in ^. 22.68 23.76 24.84 25.92 27.00 28.08 29.16 30.24 31-32 32.40 33-48 34-56 35.64 FIG. 10. Bimodal curve of variability of chromatic-rod length for 3r and 3! combined. Number of spermatozoa around upper mode, M", is greater. approximately bilaterally symmetrical. This same preparation was measured again with a no. 2 ocular and yet again with a no. 12 ocular. In the use of no. 2 ocular five hundred twenty- four individuals were measured; in the use of no. 12 ocular four hundred twenty-nine measurements were recorded. Both of these give curves agreeing closely to the bimodal curve of the first measurement. With ocular no. 2, the magnification was hardly sufficient to correctly place the modes. Ocular no. 8 was found most satis- 298 E. C. FAUST. factory and used in all the following measurements. For all these data each measurement was recorded in sequence and later incorporated into a polygon plot. All crooked or curved indi- viduals were entered separately and not used except as a check. Frequency. Length in n. 22.68 23.76 24.84 25.92 27.00 28.08 29.16 30.24 31.32 32.40 33.48 34-56 35.64 FIG. ii. Bimodal curve of variability of chromatic-rod length for 4r. Greater number of spermatozoa grouped around the upper mode, M". Frequency. Length in M.. 22.68 23.76 24.84 25.92 27.00 28.08 29.16 30.24 31.32 32.40 33.48 34-56 35-64 FIG. 12. Bimodal curve of variability of chromatic-rod length for 4!. Greater number of spermatozoa grouped around the upper mode, M". DIMORPHISM IN SPERMATOZOA OF ANASA. 299 They amounted to 5.22 per cent, of the total number measured. Such a per cent, is not sufficient to affect the dimorphic character of the curve, even if it represented spermatozoa all of a critical length, i. e., 29.16^. Length in M. 22.68 23.76 24.84 25.92 27.00 28.08 29.16 30.24 31.32 32.40 33.48 34-56 35-64 FIG. 13. Bimodal curve of variability of chromatic-rod length for 4r and 4! combined. Number of spermatozoa around upper mode, M", is greater. Four hundred individuals from il, chosen at random, were measured. This curve also shows a clear dimorphism. It is plotted in Fig. 3. Preparations 2r and 2\. — Of the right testis 2r, four hundred eighty individuals were measured; of the left testis 2l, four hundred individuals were measured. The plots of these curves, Figs. 5 and 6, show that they are strikingly similar to those of 3OO E. C. FAUST. ir and il. It will be noticed, however, that 2r has 30.78^ as a mode of higher frequency, wrhile 2! has 27.54/4 as a mode of higher frequency. Preparations jr and jl. — Of 3r only three hundred individuals were measured, because this comprised the total number that was present in the testis. Of 3! five hundred individuals were measured. The curve of 3r is shown in Fig. 8. It is bimodal in character, but with a greater number of individuals grouped around the lower mode. The plot for 3! is shown in Fig. 9. It is unmistakably unimodal in appearance, with the mode at 30.78^, the length where the upper mode of the preceding bimodal curves occurred. Comparison with the bimodal curves shows plainly that it is the upper half of a bimodal curve, such as that shown in Fig. 2. Preparations 4r and 4!. — In these preparations four hundred spermatozoa from each testis were measured. 4r is shown in Fig. n, and 4! in Fig. 12. These are both bimodal and show the dimorphic nature of the spermatozoa, although they differ considerably from the bilaterally symmetrical type. Preparations made at Woods Hole in August, 1913, from adult males were measured to check the foregoing data. When plotted as a variability curve, they, too, yielded a bimodal curve, \vith modes at the same lengths as those in the preparations Ir and il. IV. DISCUSSION. The data presented in the previous section show very distinctly the dimorphic character of the spermatozoa of Anasa tristis. On the basis of the chromatic-rod length they fall into two classes, those with an extra amount of chromatin, and those without such an additional amount. Whatever differences and vari- ations there are among the several curves, the differentiation into two groups remains proved. Such minor deviations as occur are probably due to secondary factors not yet discovered, and do not directly affect the bimodal grouping. This pronounced dimorphic character of the spermatozoa of Anasa tristis must bear some fundamental relation to the di- morphism of chromosome number demonstrated by spermato- genesis studies. It is altogether probable that the mature DIMORPHISM IN SPERMATOZOA OF ANASA. 30 1 spermatozoa grouped around the higher mode of the curve of variability are those possessing the accessory chromosome, and that those grouped around the lower mode lack this accessory chromosome. The approximate equality of the two groups (see Fig. 2), moreover, strengthens the view that the two groups are produced as spermatids in equal numbers. However, some of the curves do not show this approximate equality, but have marked digressions from a perfectly bilateral symmetry. They may have been produced as spermatids in equal numbers, and the departure from a bilateral grouping may be due to a difference in nourishment; or this deviation may be due to a slightly earlier ripening of the one group and their ejection from the testis. Another interesting point brought out by the data is the relation between the right and left testis of the same individual. If the curves of ir and il, 2r and 2!, are examined carefully, they are seen to approach bilateral symmetry even more closely when right and left testes of the same individual are plotted together than when each is plotted separately. Such combinations are produced in Figs. 4 and 7. The former is the combination curve for ir and il; the latter is the combination curve for 2r and 2!. In the combination ir and il, of eight hundred individuals plotted an actual difference of only eight is found between the two sides, or a difference of only one per cent. In the combi- nation 2r and 2!, a difference of only five exists between the two sides of the curve, or a difference of only five eighths of one per cent. On the other hand combination curves for the testes of numbers 3 and 4 show no such close bilaterality. A probable explanation for this discrepancy is offered in the relation between the type of curve and abundance of spermatozoa. In the speci- mens numbers I and 2 there was a great abundance of sper- matozoa in each testis. Because four to five hundred individuals were found to be reliable as a basis for a curve of variability, the entire number was not measured, but random samples were taken for measurement. In numbers 3 and 4, on the other hand, a different condition existed. For each of these there were not more than five hundred present in each testis. This abundance or scarcity of spermatozoa may have a very vital connection with the fundamental characters of the curve. 3O2 E. C. FAUST. In the cases of most evident bilateral symmetry (in numbers ir and il, 2r and 2!), the preparations were made up early in November, when the spermatozoa were still in abundance. In the cases where the curves do not show this bilateral symmetry (in numbers 3r and 3!, 4r and 4!), such a maximum number of mature individuals was not found. In one testis, at least, in each of these pairs (3r and 4r) the majority of the spermatozoa had been ejected from the testis. This fact was shown by the emptiness of the testis upon first examination. We may there- fore provisionally consider the normal condition as one in which there is a maximum number of mature spermatozoa, and the bilaterally symmetrical curve of variability as the best expression of such a norm. It might be assumed that this dimorphism in length is due not to any internal factor which marks out two groups of mature spermatozoa, but represents merely a sudden growth in length whereby the maturing spermatozoa pass over the low inter- modal point very rapidly, so that few are "caught" in that place. There are definite reasons why such a theory cannot be accepted. The spermatozoa in the upper and lower reaches of the curve show by their structure that they are equally mature. This is shown not only by their general size and shape, but by their staining reaction which is very different from that of immature spermatozoa. Again, immature forms when they do occur are just as likely to occupy the low intermodal portion of the curve as the mature ones. In fact immature forms have been recorded for such a critical length. Such an objection must be put aside as unjustifiable. V. SUMMARY. I. The work presented in this paper has proved the presence of size dimorphism in the adult spermatozoa of Anasa tristis. Measurements of the length of the chromatic rod of several hundred spermatozoa from each testis, plotted as a frequency curve, have demonstrated two modes, one between 27. 54^ and 28.08/1, and the other between 30.24/1 and 30.78/4, and the data offer adequate support for the conclusion that length dimorphism exists. DIMORPHISM IN SPERMATOZOA OF ANASA. 303 2. This dimorphism in length of adult spermatozoa probably bears a fundamental relation to the dimorphism in chromosome number revealed by spermatogenesis studies. VI. BIBLIOGRAPHY. Henking, H. '91 Untersuchungen iiber die ersten Entwicklungsvorgange in den Eiern der Insecten II. Zeit. f. wiss. Zool., Bd. L1V., 1892. Meek, C. F. U. 'n The Spermatogenesis of Stenobothrns virtdulatiis, with Special Reference to the Heterotropic Chromosome as a Sex Determinant in Grasshoppers. Jour. Linn. Soc., V., 32, pp. 1-22. Paulmier, F. C. '99 The Spermatogenesis of Anasa tristis. Jour. Morph., XV., Sup., pp. 223- 272. Smith, G. '12 On Spermatogenesis and the Formation of Giant Spermatozoa in Hybrid Pigeons. Quar. Jour. Micr. Sc., V., 58, pt. i, pp. 159-. Stevens, N. M. '05 Studies in Spermatogenesis. Pt. I. Carnegie Inst. Wash., Pub. No. 36. Wilcox, E. V. '95 Spermatogenesis of Caloptenus femur-rubrum and Cicada tibiccn. Bull. Mus. Comp. Zool., XXVII. EXPERIMENTS WITH TAPEWORMS. I. SOME FACTORS PRODUCING EVAGINATION OF A CYSTICERCUS .1 JOHN W. SCOTT. During the past winter the opportunity came to me to try a series of experiments upon the bladderworm stage of Tcenia serrata, one of the common dog tapeworms. The work was made easy because an abundance of material could be secured from the cottontail rabbit in this vicinity. As I have found no record of previous experiments of this kind, it has occurred to me that a brief account of them might be of general interest. One series of experiments in particular I shall here present for consideration. One of the methods used makes it possible to study the process of evagination in the living animal, and to readily secure evaginated cysticerci without the formality of passing them into the intestine of the host. Tcenia serrata is interesting historically from the fact that it was the species first used to demonstrate the characteristic life history of a cestode. Kiichenmeister took the bladderworms, known as Cysticercus pisiformis Zeder, from the body cavity of hares and rabbits and fed them to dogs. In the course of two to three months he found they transformed and developed into the adult Tcenia serrata, so that proglottides were detached and lost in the fseces. When the eggs of these forms were fed to hares or rabbits, after the second day of the experiment, minute whitish cysts were discovered in the liver tissue. Subsequently, in about thirty days these parasites left the liver and developed into the full grown Cysticercus pisiformis. It is to be noted that this life history involves a parasitism in two entirely different hosts. In general, in the first of these hosts the young tape- worm reaches a stage in development known as a Cysticercus, and usually comes to rest in some particular tissue or part of the host's body. But in order to continue development it must be 1 Contribution No. 5 from the Zoology Laboratory of the Kansas State Agri- cultural College. 304 EXPERIMENTS WITH TAPEWORMS. 305 transferred to the intestine of a second or definitive host. If this transfer is not made development stops, after a time death ensues*, and degeneration of the young tapeworm takes place. However, if it reaches the intestine of the definitive host, barring accidents, it evaginates the already formed scolex, attaches itself to the mucous membrane by hooks or suckers, grows to maturity, reproduces hermaphroditically, and to complete the life history all that is needed is for the eggs to reach again the particular kind of individual that may serve as an intermediate host. This is of course a matter of common knowledge. In order to fully understand the experiment it will be best to call to mind something of the structure of a cysticercus and the transformation which takes place when it reaches the in- testine. Fig. i is a drawing made to show the general structure of a mature cysticercus taken from a half-grown wild rabbit. It is still enclosed within the cyst. When taken from the ab- dominal cavity a cysticercus consists of a whitish, elongated bladder-like structure filled with a watery fluid; at the smaller end is a dense, compact mass of tissue. Examination shows that this more solid portion of the bladderworm is due to an invagi- nation of the anterior end as shown in Fig. 3. Within the cavity of the invagination is found the apparatus, the scolex, which serves for attachment to the walls of the intestine. In this case the scolex is provided with both hooks and suckers. A few hours after a cyst has been eaten by a dog, a young tapeworm, already transformed, may be taken from the intestine. In this process the invaginated portion evaginates, the scolex is everted, and the bladder digests or is separated and lost in the faeces. Fig. 2 is a drawing made from an evaginated cysticercus that was treated with artificial digestive juices, and Fig. 4 is an enlarged drawing of the scolex of the same. Now if Cysticercus pisifonnis is kept in physiological saline (0.7 per cent. NaCl) it never or rarely evaginates. Fifty-five specimens were left in such a solution for three days, in fact until many had died, and at the end of that time only five had evaginated. Besides this occasional evagination, one may produce eversion of the scolex by careful physical manipulation. Braun and Lu'he state that this can be readily done in the case of 306 JOHN W. SCOTT. Cysticercus tenuicollis. 'The parasite should be held just below the head-cone in the finger and thumb of one hand, while a regular pressure from within outwards is exerted with the fingers upon the head-cone. The head-cone will lengthen and if the manipulation is repeated several times, it will turn inside out and will hang down as a flat, wrinkled, contractile band upon the surface of the bladder. If pressure is now exerted upon the free end, the head will finally become everted, springing out with a sudden jerk." I have found the process of evagination is much more difficult to produce in the smaller Cysticercus pisiformis, and it is not believed that the contraction of stomach or intestine of the host could have any important role worthy of consideration in producing the eversion of the scolex. We must therefore look upon the process of eversion as a response to chemical rather than to physical stimuli. For when fed to a dog practically all cysticerci may evaginate under favor- able circumstances. In young, previously uninfected, dogs the percentage rose as high as 90 to 100 per cent, of infection. In old dogs the percentage of infection is not so high. Whether this is due to a sort of immunity dependent upon previous infection, or to some change in the strength of the digestive juices, I am unable to say. Since it is highly improbable that physical stimuli have any important role in producing eversion, it was decided to try the plan of stimulating these animals with artificial digestive juices. The following formulae were used as a basis in preparing the solutions used in these experiments : 1. For artificial gastric juice: Water 100.0 parts Hydrochloric acid 0.2 (dog, 0.3-0.5 part) Pepsin o.i 2. For artificial pancreatic juice: Water 100.0 parts Sodium chloride 0.6 Sodium carbonate 0.2 (dog 0.4 part) Pancreatin 0.2 My solutions did not include all of the elements found in the natural product, but the calcium chloride and the potassium EXPERIMENTS WITH TAPEWORMS. 307 phosphate are found in such small amounts in gastric juice that they were considered negligible. That they are not important factors in this experiment was verified by the results. For a similar reason diastase, potassium phosphate and lipase were not included in the artificial pancreatic juice which I used. Two interesting questions had arisen in this connection: (1) Why is the young parasite not digested in the stomach? (2) What are the factors which influence or produce evagination? The answer to the first question is that as a matter of fact some of the cysticerci probably do meet the fate suggested, and that all react negatively to the gastric juice in such a way that its full effect is hindered or prevented. The answer to the second question will come out as we proceed. In one of my experiments twenty cysticerci were removed from their cysts and ten were placed in each of two stender dishes. On these was poured a solution consisting of 0.4 per cent, of hydrochloric acid in which was dissolved a small amount of scale pepsin. Upon the addition of this liquid the bladderworms contracted strongly, especially at the invaginated end, and remained perfectly quiet, though they had previously been somewhat restless. The room temperature was about 90° Fahrenheit. Within an hour the bladders were digested enough so that they began to come apart. The cysticerci were left in this solution three and one half hours, and in all this time they showed no signs of activity, though the loose bladders were well digested. The acid solution was then removed and replaced by a solution containing the chief elements of artificial pancreatic juice as described above, leaving out the pancreatin. Within ten seconds several cysticerci had become active, and in a short time a few of them had evaginated. They were restless, however, and very soon all again drew in the scolex. Some pancreatin was now added to the solution; by the time this was dissolved most of the cysticerci were active and in the course of a few minutes ten out of twenty specimens had evaginated, though some of these had hardly completed the process. The dishes were then left over night. The next morning all were completely everted and most of them were relaxed in an excellent condition for preservation or study. 308 JOHN W. SCOTT. The preceding experiment indicates the probable grave danger that the parasite undergoes in passing through the dog's stomach. It also illustrates the quick response that the bladderworm gives to favorable or to unfavorable surroundings, and the response itself is a reaction to a chemical stimulus. After a considerable number of preliminary experiments in which the main facts of this paper were established, I planned a more complete series, one of which I shall give as typical of the entire set. The earlier experiments were conducted at room temperature. First, I made up some solutions which contained the principal elements of gastric and pancreatic juices, the formulae for which have been given. Second, several solutions were prepared con- taining different combinations of certain of these elements. In most cases the solution had as nearly as possible the same concentration of each component that is found in the natural digestive fluids of the dog. This is shown in the third column of Table I. Equal amounts of these solutions were next put in stender dishes and six well-developed and carefully selected cysticerci were placed in each solution. All stender dishes were then placed in a large water-bath in which the temperature varied less than one degree from 37^° C. I was unable to examine this experiment until four hours and fifty minutes later, and this allowed more time than is ordinarily required for stomach digestion. The general results of this first examination are given in the fourth column of the table. It will be noticed that in lots 2, 3 and 10, in which hydrochloric acid alone was used, there was not a single case of evagination in the eighteen speci- mens. The same was true for lot I where sodium chloride was used. The best result was produced by the artificial pancreatic juice, lot 8, in which five out of six evaginated. The next best results were found in the enzyme solutions, lots 4, 7, and n, though the reaction to the alkaline solution of sodium carbonate might be considered just as good, lot 6. The dishes were again examined the next morning, 21 hours after the beginning. This allowed considerably longer time than that required for complete digestion, and therefore the results shown in the last column do not come out in such contrast as they do when examined some- what earlier. For subsequent work has shown that cysticerci EXPERIMENTS WITH TAPEWORMS. 309 when about to die will sometimes evert the scolex under un- favorable conditions. TABLE I. To SHOW THE RESULTS OF AN EXPERIMENT MADE TO DETERMINE THE EFFECTS OF CERTAIN SOLUTIONS UPON THE PROCESS OF EVAGINATION. ALL MATERIAL WAS KEPT IN A WATERBATH AT ABOUT 37^° C. In lots 9, 10 and n the solutions were changed after four hours and fifty minutes. Lot No. No. Blad- deruorms Used. Solution Used. Result After 4 Hrs. 50 Min. Result After 21 Hrs. I 6 0.7% NaCl None evaginated i evaginated 2 6 0.2% HC1 None evaginated i evaginated 3 6 0.4% HC1 None evaginated i evaginated 4 6 0.2% pepsin sol'n i evaginated 2 evaginated 5 6 HzO — 100 pts. HC1— 0.4 pt. Pepsin — 0.2 pt. i evaginated Bladders all di- gested i completely di- gested Bladder all digested 6 6 0.4 sod. Carb. sol'n i evaginated i almost evaginated 4 evaginated 7 6 HzO — 100 pts. 0.2% pancreatin 2 evaginated 4 evaginated (slightly acid) 8 6 HiO — 100 pts. NaCl — 0.6 pt. Sod. carb. — 0.4 pt. Pancreatin — 0.2 pt. 5 evaginated 5 evaginated i not evaginated (young) 9 Sol'n lot 5 followed by Sol'n lot 6 o evaginated 6 evaginated 10 6 Sol'n lot 3 followed by Sol'n lot 6 o evaginated S evaginated i evaginated, partly ii 6 Sol'n lot 4 followed by Sol'n lot 7 2 evaginated 3 evaginated i partly evaginated A study of the table shows the following special results: Lot I verifies previous observations, though the salt solution is here raised to body temperature. Therefore, we may conclude that temperature alone appears to have no important influence in producing evagination. From lots 2 and 3 it is also clear that hy- drochloric acid interferes with or at least does not help eversion. The pepsin solution slightly aids the process as shown by a com- parison of lots 4, 5, and n. The pancreatin solution is a more favorable medium (lot 7). However, preceding a pancreatin solution by a pepsin solution does not seem to increase the 310 JOHN W. SCOTT. amount of evagination over that when it is used alone (lot n). Lot 5 shows that artificial gastric juice is harmful to the cysticerci' for it tends to digest them and undoubtedly interferes with evagination (cf. also lot 9). This last effect is due to the acid. Artificial pancreatic juice as used in lot 8 is more effective than either of its elements used separately (cf. lots I, 6, and 7). But we find that artificial pancreatic juice produces its best results when preceded by treatment with artificial gastric juice. In lot 9 this treatment gave 100 per cent. An almost equally good result was obtained in lot 10 when treatment with an acid (hydrochloric) was followed by an alkali (sodium carbonate). This experiment or parts of it were repeated several times, but the results were invariably similar in character. Other experiments like these were tried upon the cysticerci of T. serialis. The results were so much like those described that there is no need to give details here. The cysticerci of this form are more apt to be digested by the gastric juice, probably due to their smaller size. While much remains to be done in this connection, it is believed that one may safely draw the following general conclusions: 1. Treatment with artificial gastric juice followed by immersion in artificial pancreatic juice furnishes a ready and efficient means of producing the evagination of cysticerci of T. serrata. This is an easy method of obtaining material for preservation and for the study of the living parasites. It is best to use the gastric juice in a diluted form if one wishes to preserve the bladders intact. 2. In producing evagination the most important factors in artificial pancreatic juice are the alkali, sodium carbonate, and the extract of pancreatin. But to obtain the best reation to these stimuli previous treatment with hydrochloric acid is required. 3. Treatment with an acid followed by an alkali (lot 10) gives rather better results than treatment with pepsin followed by pancreatin .(lot n). Hence the sodium carbonate appears to be a more important factor than pancreatin in producing evagi- nation. 4. The cysticerci are not digested in the stomach of the dog because they do not evaginate in a harmful acid medium. The EXPERIMENTS WITH TAPEWORMS. 3! I acid causes them to contract into as dense a mass as possible. This negative response to an inorganic acid is a fundamental and very general characteristic of protoplasm. For example, a very small amount of free acid produces fatal results upon pro- tozoa, developing eggs, spermatozoa, and growing cells in general. "Sour" soils are unfavorable to the growth of most plants. Indeed, this experiment adds one more bit of evidence that one of the primary functions of the acid in the gastric juice is to kill bacteria and other living tissues that reach the stomach. 5. Finally these experiments suggest a way in which one may possibly determine why many tapeworms have a specific defini- tive host. Not that they have acquired a particular love for that host, but that the special host furnishes the right stimulus at the right time to call forth the proper reaction in the cysti- cercus. In another paper I have shown that these bladder- worms will not produce tapeworms when fed to pigs, and there is one instance on record where a man swallowed five of them, without harmful results. We shall, no doubt, soon be able to explain much of this peculiar and specific behavior of parasites on the basis of a direct response to favorable or unfavorable physical or chemical stimuli. This will bring some of the little understood phenomena of parasitism into line with the brilliant work that has been done in recent years on the behavior of free-living forms. BIBLIOGRAPHY. Braun, M. '06 Animal Parasites of Man. 3d Ed. Wm. Wood and Co., New York. Braun, M., and Liihe, M. '10 Practical Parasitology. Wm. Wood and Co., New York. Kiichenmeister, F. H. '51 Vorlaufige Mittheiling (iiber Cysticercus pisiformis der Kaninchen). Ztschr. f. klin. Med., Bresl., V. 2, p. 240. Neumann, L. G. '05 Parasites and Parasitic Diseases. 2d Ed. Revised and edited by James Macqueen, London. Scott, John W. '13 The Viability of Certain Cysticerci in Pigs and in Young Dogs. Science, N. S., Vol. XXXVII. , No. 946. (Abstract.) 312 JOHN W. SCOTT. EXPLANATION OF PLATE I. FIG. i. Diagrammatic drawing of cysticercus of T. serrala enclosed in cyst. (X4-) FIG. 2. An evaginated cysticercus of about the same size as the preceding. This specimen was treated with diluted artificial gastric juice followed by artificial pancreatic juice. ( X 4.) FIG. 3. Optical section of the anterior end of the cysticercus shown in Fig. I, with a rather complicated invagination. Drawn with aid of camera lucida. (X 18.) FIG. 4. Appearance of the scolex after evagination; parts slightly altered in position by pressure of cover-glass. Drawn with aid of camera lucida. ( X 46.) BIOLOGICAL BULLETIN, VOL. XXV. PLATE I. JOHN W. bCOTT. ON A PECULIAR MONSTROSITY IN A FROG. GEORGE WAGNER, DEPARTMENT OF ZOOLOGY, UNIVERSITY OF WISCONSIN. About three years ago a student, Mr. F. L. Conover, brought me a remarkable specimen of Rana pipiens, which he had found among the specimens at the biological laboratory of the Madison High School. The frog was alive. It measured fifty-two milli- meters body length and was apparently in full vigor, its death being caused by an overheating of my office at a time when I had to be unexpectedly absent for several days. The remarkable character of this specimen consists in the presence of three extra limbs extending caudad from the region of the sternum, two on the right and one on the left side. It is very evident that they are attached to an extra basal piece overlying the sternum. All three are more or less imperfect, so that I find it impossible to determine whether they represent arms or legs, or both. Each has four malformed digits, and each shows a thickening indicating the location of the joint between the proximal and distal parts of the limb, the elbow or knee. All are pigmented on the surface away from the frog, with no pigment on the opposite surface. The accompanying photograph (Fig. 2) and radiograph (Fig. i) of the specimen illustrate better than any verbal description the actual con- ditions present. Monstrosities in nature involving the hind legs are not un- common in the frog, but those involving the fore limbs are apparently rare. The only one known to me is one figured by Sutton ('92, p. 112, Fig. 60), involving the presence of an extra leg on the left side. There is a case in the human being, often quoted, shown in Fig. 3; I have been told about another closely similar case exhibited in this country several years ago, but I have not been able to get exact information concerning it. Dr. F. R. Lillie kindly called my attention to a very important paper by Tornier (105) which I would otherwise have overlooked. 314 GEORGE WAGNER. Tornier, by an ingeniously chosen cut, separated the upper one quarter or so of the Anlagen of both hind limbs in Pelobates. The result was in several cases, at least, the formation of the FIG. i. normal pairs of limbs plus two complete extra girdles each with a pair of limbs. Less completely successful experiments resulted in fewer and less fully formed supernumerary parts. But the monstrosities in all cases resembled closely those found in nature, and the close approximation and exposed situation of the Anlagen of the two hind limbs make accidents in nature which may be compared with this artificial "accident" at least not improbable. Tornier 's work, it seems to me, gives us a very satisfactory explanation of such forms. Is it equally applicable to the forelimbs? The location of the Anlagen under the opercle makes experimentation here much more difficult, and no such experiments have been carried out, A PECULIAR MONSTROSITY IN A FROG. 315 so far as I know. One might expect that by operating on one of the Anlagen a supernumerary arm, such as that of Button's case, could be produced. It is an experiment that some skilful operator should attempt to perform. But the Anlagen of the two limbs are here not closely approximated ; they are far apart. A simultaneous operation on both is not possible, and an accident in nature equally affecting both is in the highest degree improb- able. It appears much more probable to me that in the frog here described we have a case quite parallel to that in Fig. 3, which is FIG. 2. properly designated, according to Schwalbe ('07), as a Thor- acopagns parasiticus . The generally accepted, and to me very probable, explanation of these forms is that they started with a 316 GEORGE WAGNER. monstrosity consisting of two complete embryos united at the breast bone. For some reason or other one gains a start in development. As their circulatory systems are united (as is believed, and in many cases demonstrated) this stronger indi- vidual (autosite) soon drives its blood into the other (parasite), FIG. 3. (After Adami, after Wintersohn.) the heart of which then degenerates through disuse. This means a less adequate blood supply, atrophy, and a lesser or greater absorption by the stronger individual. It is easy to see that the later in the embryonic period the inequality begins the more extensive in development will the parasite be. I have not attempted to seek all the cases of this sort in man. Approximate parallels of that shown in Fig. 3 must be, to judge from the literature, quite rare. As to the lower vertebrates I have not been able to find a record of any case closely similar to this. It is an interesting fact to note that a frog could grow A PECULIAR MONSTROSITY IN A FROG. 317 to maturity with so very material an impediment. It makes one ask just how much of a variation from the normal is needed to enable natural selection to perform its cruel wrork? ZOOLOGICAL LABORATORY, UNIVERSITY OF WISCONSIN, July 31, 1913. BIBLIOGRAPHY. Adami '10 Principles of Pathology. 2cl edition. Vol. I. Philadelphia, Lea and Febiger. Schwalbe '07 Die Morphologic der Missbildungen des Menschen und der Tiere, u.s.w. II. Teil. Jena, Fischer. Sutton, J. B. '92 Evolution and Disease. London, Walter Scott. Tornier, G. '05 An Knoblauchkroten experimentell entstandene uberzahlige Hinterglied- massen. Archiv f. Entwickl. Mechanik, 20, pp. 76-124. AN EXPLANATION OF THE NON-PRODUCTION OF FERTILIZED EGGS BY ADULT MALE-PRO- DUCING FEMALES IN A SPECIES OF ASPLANCHNA. D. D. WHITNEY. It has been observed by various investigators in the study of certain of the rotifers that in order for a male-producing partheno- genetic female to develop fertilized eggs the female must pair with the male soon after leaving the egg while she is quite young and small. No one has observed the real reason or necessity for this early pairing. While working with a species of Asplanchna in the summer of 1908 at Cold Spring Harbor, New York, and again in the present summer it has been possible to watch the pairing of the two sexes under the microscope. This species is one that has very large individuals, probably as large as any species of rotifer. Both the female and the male are very transparent, thus making it possible to observe all the internal organs, including the ovaries and the eggs of the female, and the organs of the male with the living sperm in the testis. The sperm are very large and can be easily seen with the. one third objective. When the male and female pair the two individuals come into contact with each other and the male keeping the head pressed against the body of the female bends the body and assuming the shape of a letter U brings the posterior end into contact also with the body of the female. Then the copulatory organ of the male is forced through the cuticle of the female into the body cavity, like a hypodermic needle, and the sperm passing through it are ejected into the body cavity of the female. The sperm immediately become active and, on account of their large size, can be easily observed swimming around in the body cavity of the female. This hypodermic injection of sperm may be made on any part of the trunk region of the female. 318 EGG PRODUCTION BY ADULT MALE-PRODUCING FEMALES. 319 Many observations were made in the pairing of different young females with males. All young females observed were seen to be injected with active sperm from the male. Each female was then placed in a separate watch-glass and allowed to mature and to produce eggs or viviparous young. Some of these young females which were seen to receive living sperm into their body cavities developed parthenogenetic young daughter females ana others developed fertilized thick-shelled resting eggs. It has been shown by Maupas,1 Lauterbaum,2 Whitney,3 and Shull4 that if the young male-producing female pairs with a male thick-shelled fertilized eggs will be produced, but it has never been demon- strated that in pairing the female-producing females also receive sperm into their body cavities in the same manner as do the male- producing females. Larger and more mature parthenogenetic female-producing females as well as larger and more mature male-producing females were placed with males and many observations made. The male and female individuals come into contact with each other and the male assumes the same position as with the young females. The male makes several attempts to pierce the cuticle of the female with the piercing copulatory organ but is unsuccessful and in some instances sheds the sperm out into the water on the outside of the body of the female. These sperm are active for a few seconds, but soon become inactive and, owing probably to the injurious effects of the water, soon die. This failure of the male to inject sperm into the body cavity of both the adult female-producing female and the adult male-producing female was observed many times and in every case it was a failure. In some instances, the sperm were not shed into the water by the male. This seemed to depend more or less upon the condition of the male. If it had been isolated alone for several hours and was young it would shed the sperm out into the water when in contact with the female but if it was old and partly spent it would not shed the sperm out into the water. 1 Maupas, M., C. R. Acad. Sci. Paris, T. CXI., 1890. 2Lauterborn, Biol. Centralb., XVIII. , 1898. 3 Whitney, D. D., Jour. Exper. Zool., V., 1907. * Shull, A. F., Jonrn. Exper. Zool., VIII., 1910. 32O D. D. WHITNEY. This ability of the male to pierce through the cuticle of young females and the failure to pierce through the cuticle of adult females is due, probably, to the increased thickness or compact- ness, or both, of the cuticle of the older females. As the females grow and become older their cuticle, which is supposedly of more or less chitinous material, becomes more difficult to penetrate. It is highly probable that if the sperm could be injected into the body cavity of the adult male-producing females that the eggs would be fertilized and thick-shelled resting eggs would be produced. Males do not seem to be able to discriminate either between each other or young females of either kind or mature females. They attempt to pair with any other individual with which they come into contact. Males have been seen to inject sperm into the body cavities of each other. Many males from a general mixed culture often contain a few free sperm actively swimming around in their body cavities. These sperm have been placed in their body cavities by other males. Thus it happens that more or less of the sperm of a male may be wasted by being injected into either young female-producing females or into other males, or even shed out into the water in attempted impossible copulation with adult females. SUMMARY. 1. Both kinds of young females, female-producing and male- producing, pair with male individuals and receive active sperm into their body cavities. 2. The young female-producmg females having been injected with living sperm by the male mature and produce partheno- genetic daughter-females but the young male-producing females having been injected with living sperm by the male mature and produce thick-shelled fertilized eggs. 3. Male individuals pair and inject active sperm into the body cavities of each other. 4. The mature female-producing females and the mature male-producing females pair with male individuals but do not receive sperm into their body cavities because the copulatory EGG PRODUCTION BY ADULT MALE-PRODUCING FEMALES. 321 organ of the male is unable to pierce through the thickened cuticle of the adult female. 5. Much of the sperm of the males may be wasted by being injected into young female-producing females or into other male individuals or by being shed into the water in attempted copula- tion with adult females. BIOLOGICAL LABORATORY, WESLEYAN UNIVERSITY, MlDDLETOWN, CONN., August 12, 1913. Vol. XXV. November, 1913. No. 6 BIOLOGICAL BULLETIN THE RESISTANCE OF FISHES TO DIFFERENT CON- CENTRATIONS AND COMBINATIONS OF OXYGEN AND CARBON DIOXIDE. MORRIS M. WELLS. PAGE I. Introduction II. Plan of Experimentation . 324 III. Apparatus IV. Reactions of the Fishes - 329 V. Resistance of the Fishes -332 VI. General Discussion • 33§ 1. The Breeding Behavior of the Adults 33$ 2. Resistance of the Eggs and Fry . 34O 3. Resistance and Reactions of Young Fishes 34* 4. Resistance and General Behavior of Adults 34* VII. Summary • 345 Acknowledgments and Bibliography. I. INTRODUCTION. In a recent article (Shelford and Allee, '13) entitled 'The Reactions of Fishes to Gradients of Dissolved Atmospheric Gases" the authors discussed the physiological effects of gases in solution, on fishes, and presented a table showing the results of some preliminary experiments upon the resistance of fishes to low oxygen. They also report some experiments on the resistance to high concentrations of carbon dioxide. It is the purpose of this paper to report some further experi- ments, which have been carried on in the same laboratory, with the purpose of determining what position, varying concentrations of oxygen and carbon dioxide, hold, in the physiology and thus the ecology and economy of fishes. The experiments herein reported were undertaken at the suggestion of Dr. V. E. Shelford, and have been carried on in his laboratory, with the use of part of the same apparatus (gas control) that was used in the experi- ments referred to above. 324 MORRIS M. WELLS. II. PLAN OF EXPERIMENTATION. The method followed throughout in these experiments has been, to introduce the fishes into water containing a constant and fatal concentration of oxygen, carbon dioxide, or combination of the two, and then to observe and record the reactions of the fishes, during the time which elapsed between introduction and death. This interval, between introduction and death, was found to be the only time that could be accurately determined for every species used. For this reason, it has been taken as the basis for the data upon which the conclusions of the paper are based. Shelford and Allee give, in their table of resistance to low oxygen, the time consumed between introduction and the turning of the fishes upon their backs. This method was tried in the experiments herein described, but it was found that with the species and conditions used, the "turning time" as they call it, did not, at least in so far as could at this time be determined, represent a definite and comparable time in the succumbing reaction. For example, it was found, that in many instances, the loss of equilibrium was a gradual and not a sudden process, an illustration being the case of the rock bass (Ambloplites rupestris). It was found that when two or more species were compared, first, with regard to turning time, and second, with regard to dying time, the comparisons did not show the same relations, in the majority of cases. Furthermore the catfishes and darters often died in a normal upright position, and displayed at no time a reaction that might be taken as the turning point. It should be stated that Shelford and Allee observed their fishes under conditions which varied considerably from those of these experiments. They confined them in standing water, the temperature of which was not constant ; the waste products were allowed to accumulate; and the gas concentrations varied as a result. In the following experiments, the fishes were observed in running water of constant temperature, and very slightly varying gas concentrations. For the determination of the death point, the best criterion that presented itself was cessation of movement. To make certain that this point might safely be taken as the death point, individuals of different species were, from time to time, removed RESISTANCE OF FISHES. 325 and placed in fresh tap water, as soon as all movement had ceased. In none of these cases was there recovery or even further movement. However when a fish was removed while the movements were still faintly visible, it usually recovered and became normal once more. The nearness to the death point at which a fish might still be resuscitated varied with the species and with the individual. This variation was not fully determined. It is probable that if errors were made in deter- mining the death point, they were upon the side of exceeding the actual time, for in nearly all cases, a fish was not regarded dead until no movement had been visible for from two to five minutes. III. APPARATUS. The apparatus used in the experiments consisted of the gas control apparatus, already mentioned,1 and of three large wide- mouthed bottles, with connections, etc. In brief the gas-control apparatus consists of a series of perforated pans, of boilers, and coolers. The water may be made to pass through the pans into the boilers and through the coolers, in the order named, or it may be turned directly into the boilers and on into the coolers. If gas is to be introduced, it is injected through a gas introducer, between the first and second coolers. In these experiments, the only gas introduced was carbon dioxide. This gas was intro- duced from the cylinders in which it is purchased. According to Shelford and Alice's analyses, the gas in these cylinders contains 99.4 per cent, carbon dioxide and .6 per cent, nitrogen (different cylinders may vary slightly). Variations in the oxygen con- centration were obtained by treatment with the apparatus. The very low concentrations (.1 c.c. per liter) were obtained by boiling the water vigorously in the apparatus, and in order to do this it was necessary to start with the hot tap water of the laboratory. Because it was necessary to use the hot tap water in some experiments it was thought best to use it for all, and this plan was followed except in experiments with high oxygen (10 c.c. per liter). In these cases the cold tap water was used. In either case the temperature at which the water was used was the 1 For lull description of gas control apparatus, together with drawings, see Shelford and Alice ('13), p. 214. 326 MORRIS M. WELLS. same, for after the water had passed through the coolers in the apparatus, its temperature was the same as that of the cold water running through the coolers. In these experiments the temperature differed slightly on different days but was constant between 3 and 5 degrees centigrade. By analyses, which are given in Shelford and Alice's paper, it has been shown that there is little chemical difference between the hot and cold tap water of the laboratory. At the time the analyses were made, however, it was found that the tap water after it had passed through the gas control apparatus, showed a slight increase in magnesium content. By later analyses, I have found that this increase was due to the accumulation of mag- nesium containing scale, in the boilers. To obviate this differ- ence, I have kept the boilers free from any considerable accumu- lation of scale, and although no further analyses have been made since the beginning of the experiments, I do not think it likely that the effect of the apparatus upon the water has been any greater than the normal daily fluctuations of the water itself, excepting, of course, the differences that have been intentionally produced. After passing through the gas-control apparatus, the water was led into a 2-liter wide-mouthed bottle (see Fig. i). The introducing tube led the water to the bottom of this bottle, while the exit tube reached half way to the bottom. Thus all gas bubbles were retained in this first bottle and not allowed to pass into the experimental bottles. From bottle A, the water passed into a larger (8-liter) bottle, B. This was the first experimental bottle. From bottle B, the water was led into still another bottle, C, which was the second experimental bottle. An experiment was conducted as follows. The gas-control apparatus was started and allowed to run from two to four hours, so that the gas concentrations might be regulated and made constant. The time required for this varied with what was to be done to the water. When consecutive tests, five to ten minutes apart, showed the concentrations to be those wanted, and constant, the hose connected with the outlet of the apparatus was slipped over the inlet tube of bottle A , the outlet tube of A connected with the inlet tube of B, etc. After the stream of RESISTANCE OF FISHES. 327 water had filled all the bottles and had been running through them for some time, during which time the corks of all the bottles had been sealed with modeling clay, it was again tested for concentration and constancy of dissolved gas. The samples were usually collected at the exit tube of bottle C. If the tests showed the desired concentration to be present and constant, fishes were quickly introduced, the corks rapidly replaced and reseated, and observations begun. Because of the fact that at FIG. i. Showing the arrangement of the catch bottle A and the experimental bottles B and C. In .4 the inlet and outlet tubes are so arranged that bubbles of undissolved gas are retained in the bottle. In B and C the inlet tubes lead the water to the bottoms of the bottles. The outlet tubes reach only to the bottoms of the corks. This insures thorough mixing of the water and thus conditions throughout the bottles are the same while the water is flowing through. the beginning of an experiment one set of fishes usually de- manded the entire time of the observer, but one set was intro- duced at a time. The second set was introduced and observed after the first set had passed through the first and more vigorous stages of the succumbing process. In order to make certain that the water as it passed through bottle B was not affected to a measurable extent, by the presence of the fishes, a number of tests were made of samples collected at the exit tube of bottle B. None of these tests showed any 328 MORRIS M. WELLS. measurable difference in gas concentrations, as compared with samples taken at the exit tube of bottle A, and it is not probable that the salt content was appreciably affected by the fishes. The method of analysis that has been used throughout is a slight modification of the standard methods of the Public Health Association.1 In all cases, the collections of samples were made in narrow-necked bottles, through which the water was allowed to run for several minutes, with the delivery tube reaching to the bottom of the bottle. The bottles were kept corked as much as possible until after all the chemicals had been added. These precautions were essential, for in the case of low oxygen a sample may show 2-3 c.c. per liter, if collected and titrated in an open 250 c.c. graduated cylinder, while if tested as above, it may be found to contain but from .I-.I5 c.c. per liter. For the carbon dioxide tests, a special bottle was devised. With this bottle, the cork was removed only to run in the sodium carbonate in titrating. This was always done immediately after the collection was made. The water flowed through the bottles at the rate of from 500 to 600 c.c. per minute. This changed the water in the larger experimental bottle, once in about 15 min. and in the smaller once in 6 min. In both experimental bottles, the introducing tube led to the bottom, while the exit tube reached just to the under side of the cork. Finally, in order that the temperature of the water might not vary in the experimental bottles, a stream of cold tap water was kept running over each throughout the experiment. The number of fishes used in an experiment varied with their size. A smaller number was usually placed in the smaller bottle. The number in any single experiment varied from 2 to 10 and about 160 were killed, in all. About 25 experiments were performed; the shortest occupied from two to three hours and the longest five to seven days. The fishes used in the experiments were taken in one of the small creeks near Chicago and were for the most part experi- mented upon before they had been in the laboratory aquaria for more than two or three days. In no case had they been in captivity for longer than three weeks. In the laboratory they 1 See Jour. Infectious Diseases, '05, supplement No. i. RESISTANCE OF FISHES. 329 were kept in running tap water and there was practically no mortality. In all some fifteen species were used. In a few instances, few individuals were available and in these cases the data will not be presented unless particularly suggestive. IV. REACTIONS OF THE FISHES. The experiments were of six kinds, this number resulting from the different combinations of gases, those being selected which it seemed would be most likely to give significant results. They were (i) low oxygen (0.1-0.15 c.c. per liter) and high CO2 (35-50 c.c. per liter) ; (2) low oxygen and alkaline wTater (0.5 c.c. N/2O HC1 made neutral); (3) low oxygen and medium CO2 (15-17 c.c. per liter); (4) low oxygen and low CC>2 (about I c.c. per liter); (5) medium oxygen (3-4 c.c. per liter) and high CO?; and (6) high oxygen (8-10 c.c. per liter) and high CO;;. All the concentrations used are found at times in natural waters. Birge and Juday ('n, pp. 144-169) report concen- trations of carbon dioxide for Wisconsin lakes, which vary from about -7 c.c. to nearly +35 c.c. per liter. They also report concentrations of oxygen varying from o to 15 c.c. per liter. On page 89 (/. c.) they report concentrations of carbon dioxide as high as 47.8 c.c. per liter, in the ground waters of the state. Alice (unpublished) in tests of the water in the series of ponds along the south shore of Lake Michigan has found 40 c.c. per liter of carbon dioxide. In analyses of the water of these same ponds, I have found concentrations of oxygen varying from I c.c. to 12 c.c. per liter. It is probable that the extreme concentrations, as low oxygen (o.i c.c. per 1.) and high carbon dioxide (35-50 c.c. per 1.) are not in any sense general conditions, but concentrations not so extreme are of frequent occurrence. Birge and Juday state that conditions of low oxygen and high carbon dioxide may exist for long periods in the deeper waters of the Wisconsin lakes, but the influence of such concentrations upon general fish distri- bution is probably not so great as that of smaller but more widely distributed concentrations. The presence of high and low concentrations of carbon dioxide is affected by many factors such as vegetation in the water, 33O MORRIS M. WELLS. character of surrounding soil and incoming water, depth of the water, season of year, daily temperature, animals present, decaying organic matter, rainfall, exposure of the surface of the water to winds. \Yith so many factors affecting the gas con- centrations, they will vary greatly even within the same body of water, at any given time. Birge and Juday (/. c.) have found this to be especially true in small deep lakes, for different depths. In my owTn analyses of the waters of the Chicago region, I have found horizontal differences that are also somewhat striking. For instance the lower end of a fifty-foot pool in a small creek may contain 5 c.c. per liter of oxygen less than the w^ater of the upper end. Such variations are common and their influence upon fish distribution must be marked. In describing the reactions of the different fish species, it will be possible to give a general account which will stand in the main for all the types of experiments. Given in chronological order, the reactions were as follows: Upon introduction into the treated water, the activity of the fishes was usually increased for a short time. This increase was due in part to the handling, but also in part to the stimulation of the water, since fishes introduced in the same manner into normal water were not so active. The greatest increase in activity came with introduction into the low oxygen (0.1-0.15 c.c. per 1.) water. The period of activity usually lasted for only 5 to 10 seconds at which time some of the fish occasionally lost their equilibrium (e. g., small- mouth bass), or they sank to the bottom or floated in the water in an upright position. In all of the experiments, and especially in the low oxygen, the opercular movements at once became much more vigorous and continued so up till near the death point. The time period between introduction and the appear- ance of signs of loss of equilibrium varied with the combination and concentration of the gases present. In general this period varied directly as the lowness of the oxygen concentration or the highness of the carbon dioxide. The fishes, upon sinking to the bottom, or upon becoming quieter in the water above, rested for a longer or shorter period and then began to "nose" more or less actively about the bottle. Gulping occurred almost from the beginning. This is a normal RESISTANCE OF FISHES. 33! reaction to some extent but in the experiments the vigor and frequency of the gulping was very noticeable, and gas bubbles were often ejected at the gulps. The fishes now passed through a period of alternating rest and activity. During the early- phases of this period, activity was in the ascendency; later rests occupied most of the time. The resting periods lengthened with the gradual loss of equilibrium. The first sign of "stag- gering" was a falling to a vertical position with the head up. The fish gradually sank to this position and at first made swim- ming movements which caused it to resume a horizontal position. Later the restoratory movements were not made until the fish came in contact with the bottom. Still later the contact stimulus failed to cause a response and the fish came to rest on the bottom, lying either on its side or on its back. The gill movements were usually still strong and regular. The fish now lay on the bottom, making only occasional swimming movements. In the low oxygen experiments, the movements of the fishes often became convulsive and uncoordinated before equilibrium was lost, and as time passed, this phenomenon became more and more noticeable. Carbon dioxide, in the concentrations used, did not cause much lack of coordination, but it has been demon- strated that very high CO2 (100 c.c. or more per 1.) will quickly produce this effect. The effect of the carbon dioxide in the concentrations used wt;s that of an anaesthetic which stimulates in small amounts, or at first, but later anaesthetizes. In this connection it was noted that the fishes that were the first to succumb to the carbon dioxide were usually those that dis- played the greatest activity. As the fishes became more and more overcome, the gills developed irregularity in movement; gulps and extra large opercular movements became frequent; later the gills became more or less distended and stiffened, and the movements in- frequent. Thus in cases where the gill movements averaged 9-10 in 10 seconds at first, there might now be an average of one or less in the same period. At about this time the gill movements took on a jerky mechanical appearance which continued until they stopped altogether. In nearly all cases, after the gill movements had stopped, life was still indicated by movements of the fins and 332 MORRIS M. WELLS. especially the pectorals, which twitched or waved somewhat regularly for several minutes after the opercular movements had ceased. With the final cessation of movement upon the part of the fins, the fish was considered dead. This final cessa- tion of all movement was often preceded in the low oxygen experi- ments, by a sudden violent paroxysm, during wThich, the fish "scooted" about the bottle in a blindly convulsive manner. V. RESISTANCE OF THE FISHES. The resistance of the fishes varied with the individual, with the species, and with the size. No reliable data were obtained with regard to individual variation, because of lack of knowledge of the previous history of the animals. Specific and size differ- ences in resistance were however great enough to cover up, to a considerable extent, the unknown individual factor and are to this extent indicatory. With regard to the size (i. e., weight) in the same and different species, it was found that small fishes showed a greater resistance per unit of weight than did the larger ones. At the same time, the larger fishes, because of their excess weight, usually lived longer per individual. These two sets of reactions are clearly shown in the tables which follow. TABLE I. SHOWING THE COMPARATIVE RESISTANCE OF LARGE AND SMALL FISHES OF THE SAME SPECIES TO Low OXYGEN (.i-.is c.c. PER LITER) AND MEDIUM CARBON DIOXIDE (15-17 c.c. PER LITER). Species. No. Fish. Av. Length in Cm. Av. Weight in Grams. Av. Dying Time per Fish in Min. Av. Dying Time per Gram in Min. £ $ a 7. - l/^ ^O ro jsd amix SuiXg N 10 O *-O 01 ^O oi r^» O\ |-( ON O O O\ O Ov *-* O^.O^O OS N ro ^t ro ^1" l^> OO c £ .-•a1-; \fi ^-* •sajnuij\; u; jad auqx SuiX(I "^- Tf O O O t* 00 01 i-O O*, *O ^O ^^ O M uo ro ^ ro •sainuij^ ui qsijj jad auiix SuiXfj IT) IT} IO O4 W t^- IO •O OC O O "^ O 2C 00 O ON O\ r- ON 01 • MM w ro ^ oo Id1 O :f ^ i—3 o •sajnui[\i ui qsi j jo uiujf} jad auiix SuiAQ OOrONinr-M^xoro M ^r^^o row O^oo •sajnuij\; ui qsijj jad ainixSuiAQ OOOOOt-OWO O o o ioo\oo r-Tj-io l/-)Ot>-O400O O CO 13 ^ C - f rt C ° c .2 j' I-J V-UC O •sa]nuij\[ ui IJSI j JO UlEJf) jad auj;x SuiAsed minnow (Pimephales notatus ub (Hybopsis kentuckiensis) .... i shiner (Notropis cornutus) .... ss (Amblopiites rupestris) illhcad (Ameiurus melas) 11 *? 1 ' I" i— ' )>__i " O G O ^ ~~ o C CJ P J~i — •— C O _ii 336 MORRIS M. WELLS. young were averaged with the adults, the fishes were divided into two groups upon the basis of weight. The dividing line between the groups fell more or less naturally at about 5 grams. Upon calculating the data for the two groups thus formed, it was further found that each group illustrated much the same specific relations, and for this reason but one group, that of the larger or adult fishes, will be discussed. In Table IV. is indi- cated the resistance of a number of species to different experi- mental conditions. Table IV. illustrates the following points: (i) the relative specific resistance of the fishes to the six artificial environments taken singly (vertical columns, except last two) ; (2) the relative specific resistance of the fishes to the environments taken together (last two vertical columns) ; (3) the efficiency of each environ- ment as a death-producer for the species used; (4) the antagon- istic action of oxygen and carbon dioxide when in the same solution (vertical columns 5 and 6) ; (5) that an optimum carbon dioxide concentration probably exists for the fishes in question (columns 2 and 4); and (6) that of the concentrations used, the low oxygen was more detrimental to the fishes than the high carbon dioxide (columns I and 3). With regard to the relative resistance of the different species to the environments taken singly, attention is called to the fact that the order varied with the different combinations. This indicates that fish species vary in their resistance to any one factor, the species that is more resistant in one environment being less so in another. The relative resistance of the species to the environments taken as a whole is shown in the last two columns. The order was obtained by averaging the resistances to the single environments. Reading from right to left along nearly any horizontal row of figures will show an increasing resistance of the species to the different environments as arranged in the table. The most fatal combination used was low oxygen (0.1-0.15 c.c. per 1.) and high carbon dioxide (35-50 c.c. per 1.), and the least fatal, high oxygen (8-10 c.c. per 1.) and high carbon dioxide. That this would be the result has already been inti- mated by Shelford and Alice ('13). If in the table, column 5 be compared with column 6, it will be RESISTANCE OF FISHES. 337 seen that the presence of the larger amount of oxygen increased the resistance of the fishes to the high carbon dioxide. Hill and Flack ('io) report this same antagonistic reaction between the two gases and decide from some rather conclusive experiments that the partial pressure of oxygen influences both the higher and the lower limit of carbon dioxide which can be endured by the organism. A comparison of the second and fourth columns of the table will show that the fishes died more quickly in water that was slightly alkaline than they did in water that was slightly acid. The exact meaning of this difference is not clear. It may be that the alkaline water has some detrimental local effect upon the gills, or it may be that the neutrality mechanism (Henderson, '13) of the fishes is more quickly affected for the worse in the presence of IOWT oxygen, when the water is slightly alkaline than when it is acid. At any rate the results indicate that the fishes have a carbon dioxide optimum. This agrees with Bottazzi's investigations for tissues, as quoted by Jerusalem and Starling fio). They quote him as saying that "possibly a certain partial pressure of carbon dioxide in the liquid or blood, bathing the tissues, may represent the most favorable condition for the exhibition of the tissues' activity, and that in every tissue it ought to be possible to find an optimum tension of carbon dioxide at which the tissue would do its best work." A comparison of the different columns in Table IV. will show that the more effective factor in producing death was the low oxygen. One good illustration of this will be noticed in the case of the black bullhead (Ameiiinis melas] which occurs in columns one and four. In these columns are presented results for two sets of conditions, which differed only in the concentration of the carbon dioxide, the other factor being low oxygen in both cases. It will be seen that the bullhead lived nearly as long in the high carbon dioxide experiment as it did in the low. This indifference of the fishes to variations in carbon dioxide under these conditions is not so clear with the other fishes, but it seems to be generally true that o.i c.c. per liter of oxygen will produce death sooner than 50 c.c. per liter of carbon dioxide. It should be noted however (Shelford and Alice, '13) that most fishes 338 MORRIS M. WELLS. detect and react to carbon dioxide even in quite low concentra- tions quicker than they do to low oxygen. It has already been stated that higher concentrations of carbon dioxide (e. g., 100 c.c. per 1.) are more fatal than o.i c.c. per liter of oxygen. VI. GENERAL DISCUSSION. There are a number of questions relating to the physiology of fishes that are touched upon by the results of the foregoing experiments, but no attempt will be made to discuss them at this time, for the data are far too incomplete. There are, how- ever, certain ecological and economic bearings, which may be taken up briefly, with some profitable results. From the experiments described in this paper, and from those of other workers (Shelford and Allee, '13; Ransom, '66) we may conclude that, in general, the distribution and at times the existence of fishes depend upon two things, namely, (i) the resistance of the fishes to any condition or set of conditions in the environment, that may vary so as to become harmful; and (2) the reaction of the fishes to any such varying condition or set of conditions. Of these two factors, namely, resistance and reaction, neither can be said to be all-important in any environment, and in most environments the two are inextricably woven together. In the following discussion no attempt will be made to separate the two, but the factor of reaction and behavior will be emphasized over that of resistance, for it seems to me that future investi- gation must show fully, what previous investigation has already indicated (Shelford, 'n; Shelford and Allee, '130), that the reactions of fishes to the conditions of the environment are more vital in determining their distribution and persistence than is their power of resisting adverse conditions. With regard to behavior and resistance, the fish life cycle can be broken up into four rather distinct periods: (i) the breeding behavior of the adults; (2) the resistance of the eggs and fry; (3) the behavior and resistance of the young fishes; and (4) the resistance and general behavior of the adults. i. The Breeding Behavior of the Adults. — It is generally known that the eggs and fry are stages of relatively low resistance. It RESISTANCE OF FISHES. 339 is therefore a matter of much importance that the adults cho<>M a location for depositing the eggs that will enable them to develop normally, and which will prove efficacious for the first stages of growth of the fry. That the adults make the selection successfully in a majority of cases is undoubtedly true and the fact that the conditions at the breeding grounds may be very different from those of the normal habitat does not seem to interfere with the choice. The conditions necessary for early stages of different species vary widely. Thus the adults of two or more species may live in the same general habitat up to the 1: reeding season when the process of selection of suitable breeding grounds often results in their becoming widely separated. This results in the eggs of the different species being deposited, and the young hatched, under conditions which may differ greatly, but which in most instances prove to be the best for the development of the first stages of each species. By what process the adults of each species are enabled to select from the great variety of combinations of conditions that combination that is best suited to the development of their own eggs and fry is a matter for investigation. It is probably true that, to a large degree, the choice is a result of reactions, chemical and otherwise, to the factors of the environment. Everman ('98) states that in Louisiana the blue catfish (Ictalurus furcatus] and goujon (Leptos olivaris] are influenced in their movements by the temperature of the water. During the winter they come farther down the river, where the water is warmest, and in the summer run further up stream or retire to the deeper waters. In some recent experiments with fishes in temperature gradients I have found that many species are very sensitive to slight differences in the temperature of the water, detecting and reacting to differences as small as i to 2 degrees C. Gurley ('02) thinks that temperature together with salinity are the factors by which the salmon and other fishes which come from salt into fresh water to breed are enabled to find the freshwater streams. Green ('09) states that salmon, when coming into fresh water during the breeding season, often spend from 2 to 4 days swim- ming back and forth in the brackish water before passing on into 34O MORRIS M. WELLS. the streams. Some fishes select certain kinds of bottom for nesting sites, and in doing so probably react to light, temperature and dissolved gases. What particular factors are most important in the selection of breeding sites is not at present clear; it will probably be found that many factors are acting in most cases, but that some few factors are of such common importance in the breeding reactions of large groups of fishes that they may be used as an index to the breeding behavior of such groups. In connection with the breeding behavior of fishes there are, besides the general reactions involved in the selection of suitable situations for depositing the eggs, a great number of more specific reactions which are often of the utmost importance in the suc- cessful rearing of the new generation. Such reactions are familiar and are illustrated by those of nest-building, aerating the eggs by different devices, guarding the eggs, etc.1 2. Resistance of the Eggs and Fry. — In spite of the reactions of the adult fishes, which tend to protect the eggs, it is true that of the many thousands that one female may deposit, but very few ever reach the adult stage (Paige, '08) and in most cases the greatest mortality comes at the period when the eggs and fry are developing, (Anthony, R., '08). At this time the relative resistance of the fishes is very low, and many of the eggs probably do not develop because of too great variation in the surrounding conditions; others may develop into abnormal forms which are unable to survive the juvenile period. Ransom ('66) has shown that a certain concentration of oxygen is necessary in the development of fish eggs ; he has also shown that carbonic acid arrests development and may result in the production of abnormal forms which never reach the free swim- ming stage. Milner ('72) observed that a supply of oxygen was necessary in the successful shipping of eggs, and Loeb ('12) states again the results of some experiments performed upon the eggs of the sea-urchin, in which he found that these eggs can develop only in the presence of free oxygen, that if the oxygen is withdrawn development stops, but begins again if the oxygen is readmitted. He also states that the process of fertilization 1 See citations in Shelf ord, "Ecological Succession," III., 'n. RESISTANCE OF FISHES. 34! probably results in a decided increase in the oxidations of the egg. This increase in the oxidations would tend to make the eggs more sensitive to adverse conditions, especially in low oxygen water, and it is very probable that eggs must at times be subjected to severe conditions because of this. 3 and 4. Resistance and Reactions of Young and Adult Fishes.— Of the importance of the resistance powers of fishes in general, in waters where there is no escape from the harmful conditions, there can be no doubt. Also the relative resistance of young and adult fishes is a matter of consequence. In the experiments, described in this paper, it was shown that concentrations of detrimental factors, which are long in affecting adult fishes, may prove fatal to small fishes of the same species, in a comparatively short time. From this it will be seen that, ecologically and economically, the fact that young fishes are more resistant per unit of weight cannot be considered of great importance. With regard to the resistance of fishes to environmental com- plexes it is possible to separate environments into two general classes, namely, (i) environments whose conditions are constant or nearly so from season to season and year to year (large bodies of water) ; and (2) environments where the conditions fluctuate considerably from season to season or from year to year (smaller lakes and streams). If we consider fishes in their relation to these two types of environments, with resistance as the factor in the foreground,. wre must look upon the matter of fish existence and distribution as determined by the ability of fishes to withstand the constant conditions in class I, or the varying conditions in class 2. This then forces us to the conclusion that the matter of fish existence in a given area is one of little flexibility. If, however, we con- sider the behavior factor as important, it will be seen at once that the matter of fish persistence is now determined by the reactions of the fishes to the two sets of conditions, and becomes at once flexible, because of the fact that the fishes may avoid adverse conditions by moving out of them. That fishes will react in this way in natural waters is no longer a question of doubt. Both adult and young fishes are sensitive to detrimental concentrations of oxygen and carbon 342 MORRIS M. WELLS. dioxide, as was clearly shown in the experiments described in this paper, and that they will turn back from such concentrations has been shown by Shelf ord and Alice ('13) and in some later experiments by myself. With regard to the comparative sensitiveness of young and old fishes, Wiegelt ('85) found that young fishes were more sensitive to ammonia than adults, and Shelford and Alice found the same relation to be generally true for their fishes, in the cases of oxygen and carbon dioxide. Such being the case, in natural lakes and streams fishes would generally find a way out of areas of adverse conditions (Green, '09) if the conditions did not appear too rapidly, and the first to become active in seeking escape would be the young fishes. Many complications suggest themselves in a consideration of the ability of fishes to persist in a given environment. One will be considered as suggestive. What will be the result of planting a species of fishes in an environment, in which by its combined resistance and behavior reactions, the species is able to reproduce, but at a great disadvantage, as for instance high egg and fry mortality? This question is tied up with the matters of acclimatization, adaptation, etc. There are two main possibilities. First, the adverse conditions may result in the resistance of each suc- ceeding generation being raised; or second, the result may be a lowering of the specific resistance, with each succeeding genera- tion. Lowering the resistance of the species will soon lead to extermination; raising it may result in the species' becoming prolific in waters that at first threatened extermination. Whether fishes adjust themselves to conditions is not yet clear. There is evidence that they may adjust, in the fact that many lower forms are able to do so. Wood Jones ('12) states that corals of the same species show great variation in structure in different environments; Moore ('97) states that very young oysters will become acclimated to new conditions when older ones will not, and ('08) he further states that "commercial sponges are very susceptible to the influences of the environment and when transplanted from one place to another speedily change in character." Whether fishes will so adjust is a matter for investigation and especially since, so far as is known, fishes are not particularly plastic at any stage. RESISTANCE OF FISHES. 343 To determine the efficiency of a given body of water as a fish-producer requires the solution of a large number of problems such as the above, and many of these problems demand that the relative importance of resistance and reaction, in fish distribution and existence, be known. In the discussion thus far, resistance pf fishes has been shown to be of considerable importance in fish survival, under certain conditions and at certain stages of development. That fish reaction is the more important factor, however, is indicated largely by one fact, namely, adverse condi- tions in natural or artificial fish environments are seldom so general or so sudden in appearance that the fishes cannot escape them, at least to an extent sufficient for sitn'ii'al, by making the proper reactions, and observation and experimentation indicate that fishes, in the majority of cases, make such reactions. It seems pretty well established that fishes regulate their own distribution, and thus indirectly their existence, largely through their reactions to the environmental factors which they encounter. In other words, fishes seldom put their resistance powers to the test, so long as there is a way out of the disturbing conditions. It follows that resistance of fishes becomes a life and death matter, only when adjustment cannot be made by a behavior reaction, and such situations are rare in the life cycles of most fishes. On the other hand, fish reactions are an important factor even in the smallest and most enclosed bodies of water, provided the waters are able to support fishes at all. In ponds, for instance, the fishes may be able to withstand the stagnancy of the dry season, by coming to the surface, gulping the surface film, burrowing into the bottom, etc. (Kendall, '10). In some fish environments, vertical differences in conditions are most important in their effect upon fish reactions. In small lakes and streams horizontal conditions are important. In such waters there are areas of vegetation, muddy bottom, sand}' or gravelly bottom, ripples, pools, etc. Where such conditions are present there is considerable opportunity for the selection of habitat, upon the part of the fishes, and that they exercise such choice has been demonstrated by Shelford ('no). As a result of the choosing, we find the darters, for example, in the swift-flowing rock-bottomed ripples, while the suckers in general 344 MORRIS M. WELLS. are to be found only in the muddy-bottomed pools. Moreover the two groups of fishes seldom encroach upon one another's environments, even though they be directly adjoining as in streams where ripples and pools alternate. If fishes are kept to their own environments by rather definite reactions to conditions, then should the conditions begin to change, as for instance, the pool begin to fill with resulting changes in dissolved gases, temperature, light, current, etc., the fishes will move away and will not settle down until they reach, more or less accidentally perhaps, but also by definite reactions at times, another set of conditions, that is similar to that which they formerly inhabited. Furthermore, the con- ditions might change slowly, or but a little, so that the young fishes and the adults of the more sensitive species would be the first and perhaps the only ones to leave. In this case the result would be the slow, partial and perhaps complete depopulation of the area, provided the adverse conditions prevented the entrance of other fishes from the outside. Shelford and Allee ('130) found that certain species of fishes will turn back quite definitely from concentrations of carbon dioxide as low as 5-7 c.c. per liter, and from oxygen as high as .7-1 c.c. per liter. This being the case, the above illustration need not be considered as at all hypothetical, for concentrations such as those indicated may be found in certain parts of nearly any system of rivers and lakes. Often the adverse conditions are seasonal in occurrence, or they may appear only when certain factory and sewage wastes are introduced into the waters (Marsh, '07), but whatever the time or cause, the result must be the partial or complete depopulation of the area so long as the adverse conditions continue. It should be noted in this connection, that small variations (e. g., 5-10 c.c. CO2 per liter) from the normal, probably in many instances, produce in the long run effects similar to those pro- duced by greater variations (e. g., 25 c.c. CO2 per liter) in rela- tively short periods. From the standpoint of ultimate persistence of the fishes, it makes little difference whether they die within an hour, a week, a month, or do not die at all, but merely stop reproducing successfully, the final result must be the same, RESISTANCE OF FISHES. 345 i. e., the disappearance of the fishes from the area so long as other fishes do not come in from the outside. Furthermore the fishes might continue to reproduce more or less normally, with- out preventing depopulation, if the adverse conditions caused the young fishes to leave the area. \Ye can say then that in the long run the fish population of an area will vary directly as the concentration of adverse conditions, or that (i) the moving out of the fishes from the area, and (2) the turning back of outside fishes which tend to enter, will at some concentration of adverse conditions inevitably result in complete depopulation. Furthermore, so long as such reactions are possible, the only part played by resistance will be that the least resistant fishes, since they are usually also more sensitive, will move out or turn back first. In neither case need the matter of dying time be taken into consideration. SUMMARY. 1. Fishes die from lack of dissolved oxygen or excess of dis- solved carbon dioxide. 2. Oxygen in large amounts (10 c.c. per 1.) antagonizes the detrimental effect of high carbon dioxide (50 c.c. per 1.). 3. The action of detrimental concentrations of carbon dioxide and of oxygen is first to stimulate, and if detrimental enough, to later cause coma. 4. Low oxygen (o.i c.c. per 1.) in alkaline water causes death sooner than low oxygen in slightly acid water. This suggests that the fishes have a CO2 optimum. 5. The resistance of the fishes to fatal concentrations and combinations of oxygen and carbon dioxide varies with the individual, with the species, and with the size (i. e., weight). 6. Small fishes are more resistant per unit weight than are large ones. This fact is not particularly important ecologically or economically. 7. Ecologically and economically, it does not matter whether the fishes in a given body of water die within a minute, an hour, a week, or do not die at all but merely fail to reproduce success- fully; the final result must in any case be the same, i. e., the disappearance of the fishes from the area, unless new stock be constantly added. 346 MORRIS M. WELLS. 8. Fish resistance is important in certain enclosed bodies of water, and at certain periods of the life cycle (egg and fry stages), but the more important factor in fish distribution and survival is the reaction of the fishes to the environment. ACKNOWLEDGMENTS AND BIBLIOGRAPHY. I am indebted to Dr. V. E. Shelford for suggestions during the preparation of this paper. Anthony, R. '08 The Cultivation of the Turbot. Bull. Bur. Fish., Vol. XXVIII., 1908, Doc. No. 686, pp. 861-870. Besana, Guiseppe '08 American Fishes in Italy. Bull. Bur. Fish., Vol. XXVIII., pp. 847-854, Doc. No. 695, 1908. Dannevig, G. M. '08 Apparatus and Methods Employed at the Marine Fish Hatchery at Flode- vig, Norway. Bull. Bur. Fish., Vol. XXVIII., Doc. No. 680, 1908. Forbes, S. A., and Richardson, R. E. '10 The Fishes of Illinois. Natural Hist. Surv. of 111., Vol. III., St. Lab. Nat. Hist. Greene, Chas. W. '09 The Migration of Salmon in the Columbia River. Bull. Bur. Fish.. Vol. XXIX., Doc. No. 743. Gurley, R. R. '02 The Habits of Fishes. Am. Jour. Psychology, Vol. 13, pp. 408-425. Henderson, Lawrence J. '13 The Fitness of the Environment. New York. Hill, Leonard, and Flack, Martin '10 The Influence of Oxygen Inhalations on Muscular Work. Jour. Physi- ology, London, 1910, No. 5. Jerusalem, R., and Starling, E. H. '10 On the Significance of Carbon Dioxide for the Heart Beat. Jour. Physiol., Vol. XL., No. 4- Kendall, Wm. C. '10 American Catfishes: Habits, Culture, and Commercial Importance. Bur. Fish. Doc. No. 733. Kincaid, Walter, S. '08 New Methods of Transporting Eggs and Fish. Bull. Bur. Fish., Vol. XXVIII., pp. 1037-1039, 1908, Doc. No. 706. Knight, A. P. '03 Sawdust and Fish Life. Kingston M. Quart., 1902-3, VII., No. 3. Loeb, Jacques '13 The Mechanistic Conception of Life. Chicago. Milner, James W. '72 Report of the Fisheries of the Great Lakes. Report of \J. S. F. C. for 1872 and 1873, pp. 28-32. RESISTANCE OF FISHES. 347 Moore, H. F. '97 Methods of Oyster Culture with Notes on Clam Culture. Rep. Fish. Comm., 1897, pp. 263-340. '08 The Commercial Sponges and the Sponge Fisheries. Bull. Bur. Fish., Vol. XXVIII., Part I., pp. 403-511. Paige, Chas. L. '08 A Method of Cultivating Rainbow Trout and other Salmonids. Bull. Bur. Fish., Vol. XXVIII., pp. 781-787, 1908, Doc. No. 677. Ransom, W. H. '66 On the Conditions of the Protoplasmic Movements in the Eggs of Osseous Fishes. Jour. Anat. and Physiol., Vol. I., 1867, pp. 237-245. Rosenberg, Albert '08 Experience in Abating Disease among Brook Trout. Bull. Bur. Fish., Vol. XXVIII., pp. 941-945, Doc. No. 694. Shelford, V. E. 'na Ecological Succession. I. Stream Fishes and the Method of Physio- graphic Analysis. BIOL. BULL., Vol. XXI., No. i, June, 1911. 'nb Ecological Succession. II. Pond Fishes. BIOL. BULL., Vol. XXL, No. 3, August, 1911. 'nc Ecological Succession. III. A Reconnaissance .of its Causes in Ponds with Particular Reference to Fish. BIOL. BULL., Vol. XXII., No. i, December, 1911. Shelford, V. E., and Alice, W. C. '13 The Reactions of Fishes to Gradients of Dissolved Atmospheric Gases. Jour. Exp. Zool., Vol. 14, No. 2, February, 1912. Smith, Hugh, M. '08 The Transplanting of Fish. What has been done by the Fish Commission. Scientific Am. Suppl., Vol. 66, pp. 190-192. Stevenson, C. H. '97 The Restricted Inland Range of the Shad, Due to Artificial Obstructions, and its Effect on Natural Reproduction. Bull. Bur. Fish. Doc. No. 379. Titcomb, John W. '08 Fish Cultural Practices in the United States. Bull. Bur. Fish., Vol. XXVIII., pp. 699-757. Townsend, C. H. '07 The Cultivation of Fishes in Natural and Artificial Ponds. Rep. N. Y. Zool. Soc., pp. 89-112, 1907. von Pirko, Franz '08 Naturalization of American Fishes in Austrian Waters. Bull. Bur. Fish., Vol. XXVIII., pp. 977-982. BEHAVIOR OF THE COMMON ROACH (PERIPLANETA ORIENTALIS L.) ON AN OPEN MAZE. C. H. TURNER, SUMNER HIGH SCHOOL, ST. Louis, Mo. TECHNIQUE. The maze (Fig. i) used in these experiments was constructed out of a sheet of copper 13.5 inches long and twelve inches wide. Each runway is one inch wide and has neither sides nor top; and between each adjacent runway there is a space one and a half inches wide. This maze contains four blind alleys ; two which are straight (Fig. I, A, B), one which is L-shaped (Fig. I, 8 DE), and one which bears three culdesacs (Fig. I, 9 10 n C 6 a F 9 D E 10 11 i i_ j 2 in. FIG. i. A diagram of the maze used. I is the platform on which the roach is placed. C is the point from which the incline departs. The numerals 1-7 indicate the direct passageway to C. A is blind alley I., B is blind alley II., 8 DE is blind alley III. and g FG 10 n is blind alley IV. of the tables. 348 BEHAVIOR OF THE COMMON ROACH. 349 FG). When in use the maze is supported, in a horizontal posi- tion, by glass pillars which rest in wide pans of water (Fig. 21. These pillars were made by inserting glass stirring rods in the corks of wide-mouth bottles. When thus supported, the maze was about eight inches above the surface of the water which extended beyond it, in all directions, to a distance of eight to twenty inches. To facilitate the taking of accurate notes the parts of the stage were labeled as indicated in Fig. i . FIG. 2. Photo showing the maze and the upper portions of its support as arranged for use. These experiments were conducted, in the summer time, in an out-of-doors insectary, the whole north wall of which, except a narrow strip for a door, was constructed of wire netting. The three other walls were without either windows or doors, hence this north opening was the only source of light. Except in the few special cases mentioned in the body of this article, the maze was always arranged with the side /I on the north and parallel to that side of the house. As originally planned the experiments were to test the ability of the roach to learn to go by the shortest route from the portion of the maze marked "/" to a dark box placed at some definite place on the maze. A few preliminary experiments demonstrated that the box did not make a satisfactory goal. Some insects on reaching the box would enter it and remain therein; others would 350 C. II. TURXKR. climb up lo the top of it, meander around and then descend and go to some other part of the maze. Hence a cardboard incline, r\ tending from some definite part of the maze to the jelly glass that had served as a cage for the roach, was substituted for the dark box. This same series of experiments taught me that the best place for the incline to leave the maze was the point labeled "C." To be of value the incline should be located where the roach, in roaming about the maze, is sure to accidentally discover it. It was found that, in almost all cases, a roach, on reaching the junc- tion of 7, C and 8, turned into 8 and, after passing into D and often into E, retraced its steps and entered C; hence an incline placed at E would have been too easy a problem. It was noticed that roaches seldom entered 9; hence to place the incline at the extremity of either 1 1 or G would be making the task too difficult. C seemed the golden mean. The roaches used as the subjects of these experiments were isolated in jelly glasses and given an abundance of food. No roach was used in an experiment until it had been in its jelly- glass cage for several days. This was done to get the roach accustomed to confinement and to my presence. The roaches used were females varying in size from young ones ten milli- meters in length to full-grown individuals. The roach to be tested was always placed on 1. This was done in two different ways. Sometimes the hind leg of a roach was grasped in a pair of forceps and the roach placed on /; at others, by means of forceps, the roach was transferred from its glass cage to a narrow cylindrical beaker. This glass was then covered with a small piece of paper and the whole inverted on the maze at I. The paper was then removed, and, as soon as the roach quieted down, the cylindrical glass was removed. If a roach ran the maze three times in succession without making an error, it was considered to have solved it. After each trial, the surface and the edges of the maze were painted with alcohol to remove any odor that might have been deposited by the roach. For convenience, the errors made were grouped under two heads: those which resulted in the roach falling into the water and those made by the roach while remaining on the maze. BEHAVIOR OF THE COMMON ROACH. 351 I'nder the first head there are three subdivisions: (i) rushes into the water without attempting to run the maze, (2) falls into the water after the roarh had begun to run the maze, and ($) jumps into the water. Under the second head there are two subdivi- sions: (i) entering blind alleys, and (2) moving in the wrong FIG. 3. Learning curves constructed from the average reactions of ten roaches. The spaces from right to left represent successive trials. A, the number of times the roach rushed into the water before beginning to run the maze. B, the number of times the roach fell into the water after beginning to run the maze. C, the number of times the roach jumped into the water. D, the total number of times the roach entered blind alleys. E, total number of errors made. F, total minutes required to run the maze. 352 C. H. TURNER. direction along the right path. From the time the roach entered a blind alley until it returned to the right path was counted one error. If a roach which was moving in the wrong direction along the right path paused from time to time, each movement after a pause was counted an error. Movements, other than dashes into the water, which were made by a roach on / before it had entered the runway I were not counted; but once the roach had entered the runway i, even though it returned to the starting platform /, all incorrect movements made by it were counted errors. In my first experiments I arranged for the successive trials of each roach to come at intervals of several hours. This was done because I had an idea that experiments conducted at short intervals would produce fatigue effects which would vitiate the work. However, I soon found that the best results were obtained where the successive trials came at intervals of about half an hour and were continued throughout the working hours of a day. Towards the end of a long hot day such fatigue effects as a pronounced slowness of movement and the lapsing into errors that the roach had formed the habit of omitting would appear; but, on the whole, the repetition of trials at short intervals was much more satisfactory than the other method of experimenting' DISCUSSION OF THE EXPERIMENTS. Upon being placed on the maze for the first time, a roach almost invariably rushes off into the water. Upon being re- placed on the maze, it usually repeats the performance; some, however, do not rush into the water a second time. Sooner or later it stops rushing into the water and begins to move around in search of some other means of escape. It moves to and fro along the runways, if paths with neither sides nor top may be called runways, enters blind alleys, occasionally falls into the water,1 makes its toilet one or more times, perhaps engages in a few acrobatic stunts, and finally, by accident, discovers the incline and passes down it to the glass cell that is its home. The first time the roach is placed on the stage, this performance 1 After it had fallen into the water, the roach was always replaced on the maze at the place from which it had its fall. BEHAVIOR OF THE COMMON ROACH. 353 consumes from fifteen to sixty minutes. The next few trials the roach behaves in practically the same manner, hut makes fewer and fewer mistakes. Finally, after a prolonged series of trials, in spite of frequent lapses, the mistakes are gradually eliminated and the roach runs the maze, without making any errors, in from one to four minutes. As the roach moves along, the anten- nae are waving almost incessantly, as though seeking stimuli. The mistakes are eliminated so gradually that this may be con- sidered a trial and error type of learning, if one may use that expression without predicating the absence of sensations and feelings. Such, in brief, is the behavior of the common roach on the maze; but one is impressed by the variations displayed. These variations in behavior are of twro types: differences due to age and modifications due to individuality. The older roaches usually move much more slowly and much more carefully than the younger ones. Roaches from ten to twelve millimeters long usually move so rapidly that they might well be called frisky. The slower gait of the older roaches is not due to feebleness, for they are fleet enough when placed on the floor of a room; but to what, in human beings, WTC call caution. As a result of this difference in speed, the younger roaches, on their initial trial, usually run the maze much more quickly than the adults; but, in doing so, make many more mistakes. The variations due to individuality are the ones that are especially impressive. Some roaches, with humped backs, move sedately along in the middle of the runways, pausing at each corner to explore upward, outward and downward with their antennae: others trot along in the middle of the runways, at first usually falling into the water at each angle, but later slowing up at the corners, exploring with their antennae and moving onward. Some roaches, with their bodies extended until they are practically flat, drag themselves along so slowly that it taxes the patience of anyone who happens to be watching them; others, moving sometimes slowly and sometimes rapidly, always keep the claws of the legs of one side of the body in contact with the edge of the runway. When such a roach turns a corner it usually, although not invariably, crosses to the opposite side of the runway and clings to the edge with the legs of the corre- FIG. 4. Learning curves of a female roach 17 mm. long. The spaces from right to left represent successive trials. A, the number of times the roach rushed into the water before beginning to run the maze. B, the number of times the roach fell into the water after beginning to run the maze. C, the number of times the roach jumped into the water. D, the total number of times the roach entered blind alleys. E, total number of errors made. F, total minutes required to run the maze. BEHAVIOR OF THE COMMON ROACH. 355 spending side of its body. Some roaches pause at practically every point on the side of the runways and explore upward, outward, and downward with their antennae, and then, retracing their steps, explore the same territory over and over again; yet other roaches pause only here and there to make explorations. Some roaches, and this is especially true of roaches from ten to fifteen millimeters in length, move along part of the time on top of the maze and part of the time suspended from its edge. Had the maze bee.. Constructed of cardboard these same roaches would have made part of their journey suspended from the bottom of the maze. Some roaches pause from time to time to make their toilet : others never once pause for such purposes. Some few roaches, after making several attempts to find an exit from the maze, stop trying and act as though they have given up all hope of succeeding; others, after failing to find a means of escape, attempt to jump to freedom. These jumping roaches were always returned to the maze at the point from which they jumped. After one to many jumps had proved failures, these roaches usually stopped jumping and proceeded to solve the maze in the right way. These variations were not inflexible instinctive responses, for the same roach did not always behave the same way at all trials. JUMPING ACTIVITIES AND WILL. Although this jumping activity results in a plunge into the water, it resembles neither the dashes into the water made by a roach on being placed on the maze for the first time nor the falls into the water by roaches that are trying to run the maze. The roach pauses at the edge of the maze and explores outward and downward with its antennae. It acts as though it were trying to see something at a distance and then, after a pause, makes what an athlete would call a broad jump. Many roaches displayed this jumping behavior, but some were more prone to jump than others. I experimented with one roach which, on its initial trial, made ten jumps from the maze; usually from a different point each time. This jumping attitude is so characteristic that one can always predict when a roach is likely to jump. I say likely to jump instead of going to jump; because, after a roach has once 356 C. H. TURNER. jumped into the water, the jumping attitude does not always result in a spring. To see a roach, which has learned to avoid rushing off of the maze into the water and which will struggle hard to avoid slipping from the edge of a runway into the water, halt, reach outward and upward with its antennae, act as though it were trying to see what is beyond, pause and then jump is food for much thought. Have we not here a conflict of impulses and is not the jumping or the refusing to jump the resultant of this conflict? Is not such a resultant what the human psychologists call an act of will? Whenever I behold the jumping behavior, I am impressed with a feeling that the roach is experiencing a conflict of impulses and hence is displaying a will. In most cases the jumps were made from some one of the outer edges of the maze; but, in a few cases, the jump seemed to be aimed at a definite point. On two occasions a roach jumped and landed on the cork of one of the bottles that supported the maze. At another time, a roach jumped from the runway 2 to blind alley G and traveled along that, over zigzaging pathways, to the main trail at C and thence to the incline and down it to its cell. On its next trial, the roach attempted to make this same leap; but only its forefeet touched G and it fell into the water. On two other occasions I noticed roaches attempt to jump from one runway to another; but they always failed to secure a foothold. ACROBATIC FEATS. The broad jumps described above are not the only acrobatic activities of the roaches that ran the maze. As mentioned above, the roaches frequently moved along suspended from the edge of a runway. With the three feet of one side clinging to the edge of the maze and with the three feet of the other side braced against the under side of the runway, any of the roaches could progress for a short distance, and the young roaches, those from ten to twelve millimeters in length, would run along in this position and return to the upper side at pleasure. In this inverted position, the young roaches could even turn around without returning to the upper side of the runway. Letting go of the edge of the maze with the first and second feet of the off side and catching hold of the edge with the third foot of the A r: i FIG. 5. Learning curves of a female roach 21 mm. long, ^he spaces from right to left represent suc- cessive trials. A, the num- ber of I times the Broach rushed into the water be- fore attempting to run the maze. B, the number of times the roach fell into the water after attempting to run the maze. C, the num- ber of times the roach jumped into the water. D, the total number 'of times the roach entered blind alleys. E, to- tal number of er- rors 'made. F, totaT minutes re- ' \ I 5 n\ I I i j I W ?fl 10 P i quired^to run maze. the : r ! '; _L_ \i N 1 r I t L 1 ffl 1 i H II • , * i I ; i ; ; : i » i ; i i I \ I '• f I \ i 1 ; iffl ! ; rn E\ f : 1 1 / \| : • 1 <7 . 1 • '. \ * i. * \ > i 1 • \ || \ : / Ix \ . » : \ i \\ Aj & \ / '-. i' » > , \ A W- 'S \ I 5 , % C* "s 1 f g \/ 3J1^ »L \ / / K, / 358 C. H. TURNER. inner side, the roach would swing around on the underside of the maze until it could catch hold of the edge with the first and second feet of what had been the inner side. It then would remove the third foot of what had been the outer side and move along in the opposite direction. It was a common thing for roaches ten to twelve millimeters long to rest suspended beneath the maze with the claws of the third pair of legs clinging to the edge of a runway and with the other feet braced against its undersurface. Frequently a roach was noticed making its toilet while sus- pended from the maze by one foot. On one occasion I observed a roach, about seventeen millimeters long, which was hanging suspended from the maze by its right third leg, brace itself by bringing the left third leg in front of the other and clinging to the edge of the maze with both feet. TOILET-MAKING HABITS. I stated above that frequently movements to solve the maze are interrupted by toilet-making activities. Since most people look upon roaches as nasty things, this toilet-making behavior is a surprisingly interesting instinct. The mouth-parts, the first pair of legs and the third pair of legs are the instruments used by the roach in making its toilet and each has its special work. In cleaning the head and the base of the antennae the first pair of legs are used in much the same way that a cat uses her forelegs in washing her face. One of the flexed legs is rubbed downward over the head and the base of the antenna? one or more times and then cleaned by the mouth-parts. This may be repeated several times. As would be naturally expected, the antennae, which seem to be the most important sense organs of the roach, are cleaned oftener than any other part. If you are standing a short distance from a roach that is resting on the maze, you will notice an antenna suddenly bend downward into the mouth. Slowly it straightens itself while the constantly moving mouth-parts remove the dirt. The next moment the antenna of the other side suddenly bends downward and is treated in the same way. One day a rather keen observer, who was visiting my insectary, BEHAVIOR OF THE COMMON ROACH. 359 was sitting about a yard from a roach that was making her toilet. Suddenly he exclaimed: "What powerful muscles those slender antennae must contain." Those antennae do contain muscles which are used in waving them in search of stimuli; but those muscles are far too weak to bend the antennae into the mouth of the roach. Had that visitor looked a little more keenly he would have seen the roach bow its head, dart one of its forelegs forward, catch in its bend the antenna of the other side and bend it downward to the mouth. Sometimes, after a plunge into water, the antennae become so wet that they are held together by capillary attraction and extend forward and upward like the horn of a unicorn. In that position they are out of the range of both of the forelegs. It is interesting to watch the energetic and unavailing efforts of such a roach to clean its antennae. Standing on its second and third legs with the front part of the body elevated, it moves first one foreleg and then the other after the antennae in such rapid succession that it resembles a gymnast taking arm exercises. The palps of the maxillae and of the labium are cleaned by the mouth-parts; but each palp is flexed into the mouth by its own muscles. The body assumes a characteristic attitude while the legs and the ventral side of the body are being cleaned by the mouth- parts. Supported by the three legs of one side and by the hind leg of the other, the roach, with the side supported by one leg elevated at the expense of the other, reaches underneath her body and gives her legs and it a good cleaning with her mouth- parts. Again one is reminded of the behavior of a cat making her toilet. To clean the dorsal surface of the abdomen, the roach uses first one hind leg and then the other as a scraper or brush. With these same legs the cercopods are thoroughly cleaned. In this cleaning process the spines on the legs are quite serviceable. SENSATIONS. Plans have been formed to test rather thoroughly the senses of the roach; but, since it will be a long time before the tests can be completed, it is thought best to publish the following pre- liminarv account. 360 C. H. TURNER. Tactile. — The tactile sense of the roach is remarkably keen. Even a slight jar to one portion of the maze is responded to by a sudden more or less prolonged halt on the part of the roach on a distant point on the maze; and this is true when neither antennae nor palpi are touching the maze and even when the antenna? have been amputated. Olfactory. — On the maze itself no conclusive evidence was obtained of the use of this sense. True the antennae were almost continually waving in space, but it was impossible to determine whether they were seeking olfactory or tactile stimuli. The trap used in capturing the roaches for these experiments was similar to the Graham roach trap described by F. L. Washburn in the Journal of Economic Entomology for June, 1913; but I used an eight-ounce bottle instead of an Erlenmeyer flask and no hairs were placed around the apex of the smaller cone. A trap baited with dry oatmeal flakes would capture practically no roaches; but one baited with oatmeal steeped in stale beer invariably captured large numbers. It seems reasonable to assume that it was the sense of smell that enticed the roaches into the trap. Auditory. — Up to now my notes on the auditory sense are exceptionally non-committal. On the maze roaches seem to pay no attention to sounds produced continuously. For ex- ample, a loudly ticking clock was placed two feet from the maze; but to its sounds the roaches made no responses whatever. To certain suddenly produced sounds they responded by halting suddenly, to others they made no response. For example: one day while some tinners were fixing the guttering of a nearby house, to certain noises made by the tin the roaches were quite responsive; but, when, by means of a small bell, I attempted to demonstrate that they respond to suddenly produced sounds, the roaches made absolutely no external responses. By means of a Galton whistle and other methods an attempt is being made to solve the problem; but, up to now, no satisfactory solution has been reached. Vision. — That the roach possesses vision of some kind is "certain, and at times a roach would act as though it were able to distinguish objects at a distance. Recall roaches jumping from one runwav to another — a distance of an inch and a half- BEHAVIOR OF THE COMMON ROACH. 36! and roaches jumping from the maze to the top of a nearby bottle. Yet I could bring a pencil lo within a centimeter of the head of a roach without causing a response unless I touched one of the antennae. CONCLUSIONS. 1. By arranging the trials at intervals of half an hour, a roach may be taught, within a day, to run the maze. 2. The gradual manner in which it eliminates its errors would cause one to say the roach learns to run the maze by the trial and error method; yet, in so doing, it utilizes sense stimuli. This is evidenced by the careful manner in which it examines (often over and over again) the corners and edges of the maze and the space adjacent thereto. 3. At times the roach acts as though experiencing the emotion the psychologists call will. 4. Although the effects of training persist for a long time, yet the memory of the roach is poor; for after an interval of twelve hours marked lapses were noticed. 5. In its toilet-making activities the behavior of the roach resembles very much the toilet-making activities of the cat. 6. In their behavior on the maze roaches display marked individuality. 362 C. H. TURNER. TABLE I. Errors which Errors Due to Movements Caused the Roach <>n the Maze. to Fall into the ju * 0 Water. '3 u' N 55 C"^ B •o — '• By Entering v v u tn 3 M Blind Alleys. c .C N " B _3 .C C/J ^ B 3 '" jz OJ N V a< 1 3 .2 UJ3 ,_J ^""^ .5? rT" 1^ h i- *ti 4J si g- 2 o § ^ ^ ^^ HN 4J « "5.2 -?os K^. 1-1 __ S HH 6 S V S i< ^H OJ *J 3 1) o Z > 0 „ fc ^1 J3 ^-1 E 3 H rt ££ /c •H i Aug. 7, 3:55 P.M. 0 0 0 I o 5 6 I o 0 I 19 60 2. Aug. 8, 8:02 A.M. i6j 3 0 I I 13 18 o I i 2 24 33 3 Aug. 9, 6:54 A.M. 23 3 0 0 0 24 27 i 0 2 3 24 46 4 Aug. 10, 8:08 A.M. 24 J- 0 O I o 3 4 o 0 0 o o 7 S Aug. io, 6:53 P.M. io! I O o 0 3 4 0 0 0 0 i 7 6 Aug. ii, 7:52 A.M. 13 I O I o 3 5 2 0 0 2 8 i-1 7 Aug. ii, 2:05 P.M. 6i o o 1 o S 6 I 0 0 I 4 9 8 Aug. ii, 3:22 P.M. 13 o o I I 0 2 0 2 2 4 IO 21 9 Aug. 12, 7:23 A.M. 16 I I o o IS 17 o 8 8 16 4 15 IO Aug. 12, 8:06 A.M. i 2 0 I o 23 26 0 2 2 4 9 32 ii Aug. 12, 9:16 A.M. 4 o 0 I o 0 I 0 0 0 o o 5 12 Aug. 12, 10:18 A.M. i I I 0 o 7 9 I 0 I 2 3 IO 13 Aug. 12, 12:46 P.M. 2? 0 0 o o i I I 0 I 2 4 4 14 Aug. 12, 12:54 P.M. 5 0 0 o o 4 4 0 O O O 2 3 IS Aug. 12, i :20 P.M. 1 0 0 I o o I 3 O o 3 4 2 16 Aug. 12, 1:33 P.M. 1 0 0 I o o I 2 O 0 2 2 4 17 Aug. 12, 2:29 P.M. I o 0 0 0 o o O O 0 O 2 7 18 Aug. 12, 2:58 P.M. 1 2 0 o 0 0 o o O O o 0 4 i 19 Aug. 12, 3:34 P.M. !_ 0 o 0 o I i o o o 0 I 2 20 Aug. 12, 4:01 P.M. J 0 o 0 o I i o o 0 0 I 2 21 Aug. 13, 6:04 A.M. 14 0 o I 0 o i o o 0 0 2 3 22 Aug. 13, 6:45 A.M. 3 1 I o 0 0 o i o o o 0 I 2 23 Aug. 13, 7:08 A.M. 1 0 o 0 0 I i o o 0 0 O 2 24 Aug. 13, 8:03 A.M. I I 0 0 0 o i 1 o 0 0 0 0 I This is a compilation of the reactions of the first ten roaches examined. Throughout the first part of this series of experiments a long rest was given after each trial. BEHAVIOR OF THE COMMON ROACH. 363 TABLE II. Errors which Errors Due to Movements Caused the Roach . V on the Maze. to Fall into the 4) c Wat;r. )o N 'Si "U — ' O TO By Entering 4) [fl V U OJ i .£ t/1 E * 3 uJ Ci u OJ a g c 3 5 gj -Q " 11 OJ N rt * o Date. S"3 <; <," •oa 0-S o o, o u en S o (J €M «.£ ? ? c5 M H •^ = x- 1 Is C 3 (A a •x 1 3 3 " - « 00 O v bo _O bo C *-~ a o x — X c S. . H ja - .— •— '], ° e H V X -H be bib be M ~ ^ x ^_ 3 _S £ QJ 0 "~ E fo E M .1 i s£.| 1 — I o <« 3 ej JH — * two C -C H jj; O S3 ^ o ^5 -£ " 1-1 1— I „• E L — o a u."5fc OJ 1) u "s Z 1 o Z i — i - 0 ~ V •gsa 3 3 c 3 H HH r z I Aug. 8, 9:30 A.M. 0 I O i i o 3 2 10 0 12 13 45 2 Aug. 14, 5:46 A.M. 1301 I 0 0 0 4 5 2 2 0 4 9 48 3 Aug. 14, 7:04 A.M. II 0 o i 0 i 2 O 0 0 0 0 4 Aug. 14, 7:31 A.M. 1 0 o 0 2 2 0 0 o 0 0 7 5 Aug. 14, 8:03 A.M. 1 o o o 0 O 0 0 O o 0 0 7 6 Aug. 14, 8:40 A.M. A 0 0 I 0 O I O S 0 5 4 12 7 Aug. 14, 9:02 A.M. I o 0 o 0 O 0 o 0 0 0 I 4 8 Aug. 14, 9:37 A.M. 1 2 o o o o 0 0 o I 0 I 0 3 9 Aug. 14, 10:14 A.M. 3 o 0 0 0 O O o 2 o 2 0 6 10 Aug. 14, 10:42 A.M. 1 o 0 0 0 2 2 0 0 0 o 3 6 ii Aug. 14, ii :i8 A.M. 1 2 I o o 0 O I 0 O 0 0 0 6 12 Aug. 14, 11:45 A.M. | I o 0 o 2 3 0 o o 0 o 2 13 Aug. 14, 12:16 P.M. ^ 0 0 I o 0 i 0 o 0 o 2 4 14 Aug. 14, 12:46 P.M. 1 2 0 o I o 0 i 0 0 o o 2 3 15 Aug. 14, 1:08 P.M. i I 0 I o 0 2 0 o 0 0 2 3 16 Aug. 14, 1:58 P.M. I o 0 0 o O 0 o o o 0 I 3 17 Aug. 14, 2:28 P.M. ^ 0 o o o 0 0 0 0 o 0 0 2 18 Aug. 14, 3:07 P.M. 1 2- o o o o o 0 0 0 0 0 6 3 19 Aug. 14, 3:50 P.M. 3 4 o 0 0 o o O o o o 0 i 4 20 Aug. 14. 4:24 P.M. i I 0 0 0 o I o o o 0 i 4 21 Aug. 15, 5:30 A.M. 13 I o o I 0 2 I 0 0 I 2 4 22 Aug. 15, 5:46 A.M. 1 o o 0 o o 0 0 0 0 0 0 2 23 Aug. 16, 6:35 A.M. 24! o I 0 0 o I 6 0 0 6 2 12 This table records the reactions of a female roach 21 mm. long. Throughout this series of experiments the rest period between the experiments was short. 364 C. H. TURNER. TABLE III. 2% Errors Due to Movements on Errors which Caused the O rt.S E = the Maze. Roach to Fall into the Water. ui° (A "5 'C Due to Entering Blind Alleys. OJ J: *" c L- (A g £ o C/J C 2 Is H •S C o 3 "£ "« ,-, 'T' f*"1 £ <= - r, L "-I 3 L, .£ « <_ 2 j~ rt B) > '35 tft 0) i!ljs H M M •-•(J = u 01 J3.1J C^P a 0 cu ~ 'he aj llfl O .C . In ._ ai U ^ i-t a o £2 go 1« o 3 °~ " bi .^ £^ E1*' s£^ H 1"°™ C £ cCQ ~ 3 •- C S g^ E- £~ itS W If QJ O _Q tfl E fo , <-." " ' H > - 2 Ky j-. 1-. J "7 "^ " O. ^_ £ a HH « M > 2 Ife ^2 gir £?*" J3 "" H := a ** 2. c/; o H I O 2 O 2 o 18 22 2 3 I 6 13 37 2 i8j 2 0 3 I 9 15 I i 2 4 6 27 3 16 2 o I 0 10 13 2 2 I 5 7 23 4 165 I o I 0 6 8 O I I 2 7 26 5 115 0 o I I 6 8 0 0 O O 4 17 6 5 I o I 0 7 9 0 I 0 I 3 12 7 4* I 0 2 o 7 10 I I O 2 9 13 8 15 0 0 I o 2 3 I 2 o 3 5 IO 9 9^ I 0 0 o 9 10 o 2 o 2 2 II 10 3i I o 0 0 6 7 0 2 2 4 7 II ii i I o 0 1 2 4 0 I I 2 4 12 12 I I o 0 o 6 7 0 0 0 0 2 5 13 if 0 o I o 0 i o O 0 0 2 5 14 i 0 0 I 0 I 2 0 0 0 O 2 3 15 i I o I o I 3 I 0 0 I 3 3 16 3 4 o 0 0 0 I i 0 0 0 0 i 4 17 3 o o o 0 I i 0 0 0 o i 5 18 1 2 0 o 0 0 I i 0 0 o 0 5 2 19 1 2 0 0 0 0 I i 0 0 0 o i 4 20 1 2 0 0 0 0 I i 0 0 0 o 2 4 21 1 2 0 0 o 0 0 o 0 0 o o 2 5 This table records the reactions of a young female roach 1 7 mm. long. Through- out the first part of the series the rest period between two trials was long. BEHAVIOR OF THE COMMON ROACH. 365 TABLE IV. Errors which Errors Due to Movements Caused the Roach 4) on the Maze. to Fall into the a 3 O. a o 2 H C3 V K-S G J3 . II |§ — c ex e 3 K_ a o o •a ^" CJ '7. f Date. H! **" jS ^ ~ 1 ^ J-- v o M.h •- " ' 00 2 '" '•' i>,= Ui 3 OJ 'o I £•- '^ °5 *s M JyQ v c c ** S a cd u r^ j£ 0 3 3 5 o .£ H H 00 o" M u M E'c C'H ;=? ,0 Rs « *o X ti bio H •So H ~f, 'v--S j_^ f"1 "oji ft) OJ OJ JH E E W a^ 1 , QJ uX fc J3 h ^ t- rt £ i "I -G *"" • ,_4 .— C _c J Q 3«,S (^ | I"H _• " S- a" u 6^ 1 'Ts o 11 fc o & i. V.VV J3 H H 1-1 s -C S H| H H 1 Aug. 8, 9:13 A.M. o 0 O I 0 13 14 I 2 2 5 9 14 2 Aug. 9, 8:19 A.M. 23 I O I i 5 8 I 2 I 4 6 26 3 Aug. 9, 1:43 P.M. s§ 0 0 o o o 0 I 0 0 I 2 3 4 Aug. 10, 6:57 A.M. I?4 I 0 I 0 9 ii 0 3 3 6 18 57 5 Aug. 10, 11:18 A.M. 4l 0 0 I o 5 6 0 o o o 2 21 6 Aug. 10, 2:52 P.M. 3l I 0 o 0 8 9 0 i 0 i I II 7 Aug. 10, 4:26 P.M. ij 2 0 3 0 17 22 I 0 o i 23 22 8 Aug. 10, 6:30 P.M. 2 I 0 I o 6 8 0 o o o 3 6 9 Aug. ii, 5:40 A.M. IIj I 0 I o 13 15 0 o 0 o 3 18 10 Aug. ii, 7:00 A.M. ij I 0 I 0 7 9 0 I I 2 5 12 ii Aug. ii, 8:23 A.M. i^ 3 o I I 13 18 2 o 2 4 IO 23 12 Aug. ii, 10:54 A.M. 25 2 o I o 5 8 0 0 0 0 5 7 13 Aug. n, i :23 P.M. ' 2\ O 0 I I i 3 0 0 0 0 0 7 14 Aug. ii, 2:00 P.M. \ I 0 I 0 i 3 O o O o 2 2 IS Aug. ii, 2:29 P.M. j \ O I I o 5 7 0 o O o 3 5 16 Aug. ii, 3:12 P.M. f I 0 I 0 8 10 0 0 0 0 i 7 i? Aug. ii, 3:46 P.M. \ 0 o o 0 8 8 o o 0 o i 9 18 Aug. ii, 4:37 P.M. f I o I o 5 7 0 o o o 2 2 19 Aug. 12, 6:02 A.M. 13^ I o I 0 7 9 0 0 0 0 6 22 20 Aug. 12, 6:30 A.M. 1 2 o 0 0 0 3 3 0 4 0 4 3 12 21 Aug. 12, 6:56 A.M. £ I 0 I o 2 4 o 4 o 4 6 13 This table records the reactions of a female roach 10 mm. long which after 21 trials had displayed no ability to learn the maze. It learned to avoid falling into the water; but displayed no ability to avoid making errors on the runways. It finally gave up attempting to learn the maze and tried to jump to freedom. The roach was accidentally killed at the close of the 2ist experiment. Its reactions are recorded because it was the most unapt roach that I had. /& * MBL WHOI LIBRARY WH 17JU H