to the acro- some of other animals. At the opposite end of the sphere a smaller vacuole develops (figs. 119, 124, 127-134); it is frequently difficult to see these two vacuoles plainly. Thus the sphere comes to lie behind the nucleus, and in a definite position, namely, 15 All these relations have been described as found in a particular testis, no. 282, in which a single sphere is the rule. But in certain other testes multiple spheres are frequent, so that sphere relations vary in different individuals of the same species. It is remarkable that in the large spermatocytes of follicles 1 and 3 no spheres were found. 776 THOS. H. MONTGOMERY, JR. to one side of the centriole, while at the opposite edge of the cen- triole the axial thread is rooted (figs. 127-134). By reason of the inclusion of two vacuoles of different sizes the sphere comes to show a bilobar form with an annular constriction, and just after the occurrence of the nuclear discharge it divides into two (figs. 185, 1386). The posterior portion (Sp.) becomes still paler and wanders into the tail of the sperm (figs. 185-139), often pass- ing a considerable distance behind the head (a greater distance than shown in any of the figures) ; there it becomes more and more indistinct, but I do not know whether it disappears entirely. The anterior portion of the sphere (Pf, fig. 135) suddenly becomes deep staining at the time of its separation from the posterior part, perhaps by addition to it of some of the material discharged from the nucleus. It separates completely from the posterior portion (fig. 136, Pf), then passes forward along the nuclear sur- face (fig. 137) until it arrives at the anterior end of the nucleus (figs. 138, 1389). There it becomes gradually elongated (figs. 140-144, 146) and thus comes to compose the lance or perfora- torium of the mature sperm (fig. 147). The mature lance is frequently thread-like and branched, nearly pseudopodial, and penetrates into the cytoplasm of the follicular nurse cells. I could not determine whether there is a cytoplasmic envelope around it and the nucleus of the mature sperm. This sphere of the spermatid resembles that of the spermato- cyte in optical appearance and in having no connection with centrioles; but differs in its elaboration of vacuoles, in its changes of form, in its division, and especially in having a close contact relation with the nucleus. The evidence is that these two spheres are dissimilar structures. D. The mitochondria These bodies had been figured in my paper of 1898 and described for the growth period of the spermatocytes, under the name of the idiozome mass. But I did not distinguish them from the idio- zome and sphere, nor did I possess suitably stained slides for their SPERMATOGENESIS OF EUSCHISTUS Gite more thorough analysis. In the present paper I am able to describe them in much greater detail, thanks to the use of stronger powers of magnification and especially to certain very successful iron haematoxylin preparations. In the spermatogonia no evidence was found of indubitable mitochondria. One sees there (figs. 1, 3) scattered granules in addition to the mitosome (mit) and the idiozome (7d), but these have the same optical appearance as the idiozome material and therefore are probably remains of a disintegrated idiozome of a previous generation: and on sections that show the mitochon- dria of the spermatocytes deeply stained, these granules of the spermatogonia always remain pale. Further, there are certainly no filamentous mitochondria (chondriokonts) in the spermato- gonia; from which we should conclude either (1) that there are no mitochondria at all in the spermatogonia, or else (2) that the mitochondria of these cells differ chemically and physically from those of the spermatocytes. In the earliest stages of the spermatocytes (fig. 5) are found merely similar pale granules. But shortly thereafter there arise delicate deeply staining threads in the vicinity of the idiozome (fig. 7). They do not undergo any marked increase up to the zygotene stage (figs. 7-30), and are often apposed to the idiozome but whether actually within it is hard to determine. Sometimes, however, deep staining granules do lie within the idiozome (fig. 30). There is no evidence in these earlier nor in later stages that they are produced by emigrated chromatin particles, in the sense that chromatin particles leave the nucleus bodily, and at first they are generally separated from the nucleus. Further, they are separated from the chromatin plate of the nucleus by the idio- zome. Just before the idiozome begins to degenerate they increase in length and number (figs. 31-35). Their rapid growth coincides more or less with the development of the sphere in the cell body, and with the conjugation of the autosomes in the nucleus (figs. 86-55), and they gradually come to extend into all regions of the cell body. Some of them are then usually on and within the sphere, others in contact with the idiozome, the most of them 778 THOS. H. MONTGOMERY, JR. separated from both of these bodies. Thus the mitochondria become scattered throughout the cell body but most abundantly in a perinuclear zone, as shown in figs. 42-55, 57-59, 61-63, 65, 66, 71, 84,85. So far as could be determined they are unbranched threads, though often twisted and angularly bent. What their number is was not ascertained, the drawings representing with fidelity only the majority of those present in a particular plane of some thickness and only those seen with the greatest distinctness. Thus they come to make a thick mesh. In general they are bent around the nucleus, and mitochondria that appear on the drawings to terminate against the nuclear membrane do not really end so but curve around its surface. When the centrioles become visible the mitochondria in their vicinity frequently converge towards them (figs. 57-59, 61-63, 65, 66, 71, 84, 85). Very often they are so closely looped upon themselves that the enclosure of such a loop may present the appearance of a granule with a sharp bor- dering line, and this illusion is the more striking in the cases where the mitochondria are faintly stained; such appearances misled me in 1898 into supposing the mitochondria to be a mass of large granules. But when they are deeply stained it can be determined that there are no large granules in the cytoplasm except the chromatoid corpuscles. During the prophases of the maturation divisions they become, for the most part, smooth, continuous threads, though some may retain the earlier appearance of chains of granules. Each thread is not always of even thickness throughout but frequently is irregularly dilated, as I have shown in the drawings. I have looked to see whether there occurs in the mitochondria any process analogous to the process of pairing of the chromo- somes, but have found no evidence of it. The stages just described were worked out mainly upon one preparation, for the reason that this was most excellent for the examination of the sphere. The stages next to be described have been drawn from the sixth follicle of testis no. 265 which exhibited more clearly than any other the mitochondria during the matura- tion mitoses. In this testis they also appeared uniformly of SPERMATOGENESIS OF EUSCHISTUS 779 greater diameter than in any other preparation, probably because they were less strongly destained.'® Fig. 86 represents a late prophase of the first maturation divi- sion, showing the greater number of mitochondria present, and a number of these lie close to the outer surface of the nuclear mem- brane (not within the nuclear cavity as the figure might indicate). In this and the next following stage (fig. 87) they show no such radial grouping around the centrioles as was to be seen in earlier stages (figs. 71 and 85). In fig. 88 the first maturation spindle is established, and now as in all later stages the mitochondria lie outside of the mantle fibers; their peripheral disposition is better seen in the polar view represented in fig. 91, where only those parts of them are drawn that le in the equatorial plane. Figs. 89, 92-94, show other cells of the same stage and demon- strate how manifold the arrangement of the mitochondria may be and how they do not appear to be influenced in any way by the spindle or the centrioles. It would seem possible to count them in the stages of the first maturation, for at this time they are unusually distinct and large. But this is a matter of great difficulty on account of their great length and irregular twisting; thus on several occasions I have spent several days drawing those of a single cell. Further these cells are generally too large to lie wholly within the plane of a section, therefore it is not always possible to determine whether apparent ends of mitochondria may not be truncations by the 16 The sections of testis no. 265 showed the mitochondria deeply stained only in follicles 1 and 6, those on the periphery which are most affected by the fixative (Flemming’s fluid) while in the intermediate follicles the mitochondria were only slightly stained, also in thicker sections they are better stained than in thinner ones. This illustrates how much a matter of accident it is to secure suitably stained preparations. Sometimes it happens that of two cells lying within the same cyst, one exhibits the mitochondria very distinctly while the other does not. Or again, certain of the mitochondria within a cell stain deeply and others faintly, or one mitochondrial thread may be distinct in one portion and pale in another, evidently dependent upon depth beneath the surface of the section. I have endeavored to select for drawing only those cells in which the majority of the mitochondria appeared deeply and uniformly stained, which has required the close comparison of a very large number of cells. 780 THOS. H. MONTGOMERY, JR. knife. The counts made on the mitochondria of these figures ‘ resulted as follows: Fig. 87, one hemisphere of a cell, about 10. Fig. 88, less than one hemisphere of a cell, 9. Fig. 89, nearly a whole cell, about 16-17. Fig. 92, nearly a whole cell, apparently 9. Fig. 93, one hemisphere of a cell, 11-12. Fig. 94, one hemisphere of a cell, 10. These counts give from 9 to 16 or 17 distinct and separate threads. These cells were selected as the clearest in my prepara- tions, the ones showing the mitochondria stained most deeply and uniformly, and of them that of fig. 92 was the most distinct. There now comes up the question how these behave during the division of the cell body, for they lie wholly outside of the spindle and do not fall under its dominion. Figs. 95 and 97 each show the mitochondria of about one hemisphere during the anaphase of the first maturation mitosis; a number of them lie more or less parallel to the spindle, others lie near the poles, but none are yet dividing. In the later stage of fig. 99 all the more strongly stained mitochondria of the cell are drawn, but there is still no division of them. In fig. 100, the only good case of this stage with mitochondria well stained, the more in- tensely stained ones have been drawn, and it can be distinctly seen that certain of those which lie parallel to the spindle are breaking into two at the equator, and that the others which lie removed from the equator are undergoing no division at all. It thus results that mitochondria which happen to cross the line of constriction of the cell body become there broken into two, and that the others pass bodily without division into one or the other of the second spermatocytes. This happens because they are quite unsymmetrical in arrangement and do not come under the influence of the spindle fibers. Further, it would be extremely improbable that a particular mitochondrial thread that does become divided would be pinched through in its exact middle portion. Again, the mitochondria often appear more numerous at one pole of the cell than at the other. (figs. 93, 97,99). Therefore during the first maturation mitosis there is no mechanism except the constriction of the cell body to divide the SPERMATOGENESIS OF EUSCHISTUS 781 mitochondria, consequently there is a variable and irregular portioning of them to the daughter cells, and it is a matter of chance how much mitochondrial substance goes into each of the latter. Fig. 101 shows them on a polar view of the equatorial plane of a second spermatocyte. Fig. 102 shows all the threads (12) of somewhat more than one hemisphere, and fig. 103 all (10) of one hemisphere of the same mitosis; in these metaphases most of the mitochondria are shorter than in the preceding division because most of them. had been divided then, but they lie quite as irregularly in the cell. Succeeding anaphases are exhibited in figs. 104, 107-109, illustrating that here the phenomena are essentially of the same kind as in the preceding mitosis, namely, those mitochondria that happen to lie across the eyuator become broken there by the cellular constriction while the others are not divided. In figs. 104 and 109 the mitochondria of only one hemisphere are drawn, but in figs. 107 and 108 all the mitochon- dria of the cell. ‘Though they are irregularly divided by the matu- ration divisions their relative amounts in the spermatids do not seem to be greatly different; the evidence is that each long mito- chondrial thread becomes divided in the second maturation mito- sis if not in the first. Curious protuberances of the spermatocytes are produced that would seem to indicate pseudopodial movements of the cell body during the maturation divisions. These are shown in figs. 88, 89, 92-95, and since all of these were drawn from cells floating free in the testicular cavity (except fig. 85) these cell processes are not due to the pressure of other cells. Similar protuberances are found upon second spermatocytes (figs. 102, 104, 107); and the left hand one of fig. 104 looks as though it might have persisted since the stage of fig. 94. Frequently mito- chondria extend into such protuberances, but not always; there- fore they cannot be considered the producers of them. Towards the close of the second maturation mitosis the mitochondria of each spermatid commence to fuse together (figs. 110, 111), forming an irregular mantle around the connective fibers of the spindle. Then those of each spermatid give rise JOURNAL OF MORPHOLOGY, VOL. 22, No. 3 782 THOS. H. MONTGOMERY, JR. to a true ‘Nebenkern’ (figs. 112, 113), lying at the distal end of the nucleus. This becomes spherical, stains more faintly than before, and differentiates into a central denser zone and a periph- eral (frequently vacuolated), more fluid layer (figs. 114-117). This mitochondrial Nebenkern then becomes irregularly pyri- form (figs. 118, 120), frequently with a concentric disposition of layers of different densities (fig. 119). Elongating still more it becomes divided into right and left halves (fig. 121), the axial thread lying between the two. One end of the now paired Neben- kern touches the nucleus near the centriole, while the other end grows backward into the tail of the sperm. Figs. 122 to 128 illustrate further stages of the growth of this paired Nebenkern, showing how it keeps pace with the growth of the tail and how its two halves become spirally twisted upon each other. Up into the adult sperm (figs. 127-147) the mitochondrial substance of the Nebenkern is found enclosing the axial thread; but whether it extends the whole length of the tail in the adult sperm was not determined. It. becomes gradually more pale in appearance, perhaps due to its becoming thinner in mass. It is certain, how- ever, that this mitochondrial substance extends a long distance into the tail and is not limited to a small middle part; indeed, the adult sperm shows no part comparable with the well defined middle piece of mammalian sperm. E. Discussion In cells of the spermatogenetic cycle a number of larger struc- tures are found in the cytoplasm, though generally of less com- plicated nature than in egg cells because most sperm cells do not appear to produce yolk. One of these is the mitosome (Platner, ’89; Spindelrestkérper, Meves, 97) which arises in the telophase of mitosis from spindle fibers (connective fibers) persisting at the distal end of the cell. It is most prominent in spermatogonia, where it is a dense body, often continuing into a subsequent prophase, and often fusing with mitosomes of neighboring cells. In Euschistus a mitosome SPERMATOGENESIS OF EUSCHISTUS 783 is not produced in the spermatocytes or the spermatids. There is much need of a careful study of it in the spermatogonia, for its genesis is by means fully established. A second body is a more or less spherical, lunar or cup-shaped structure that encloses the centrioles so long as they remain at the distal pole of the nucleus. Meves (1897) has named this the ‘idiozome,’ and later (’02a, p. 53) the ‘centrotheca.’ Previously it had been called variously ‘sphere,’ ‘attraction sphere,’ ‘archo- plasm’; but Meves pointed out that since it has no connection with spindle fibers, but disintegrates before the metaphase of mitosis, it is quite different from the attraction sphere of Van Beneden and the archoplasm of Boveri. A new and interesting relation in Euschistus is that the idiozome always touches the nucleus at that region where the chromatin composes a particu- lar peripheral chromatin plate; the only other writers who have found a similar chromatin plate are Pantel and de Sinéty (’06), but they described it as a thickening of the nuclear membrane (their fig. 10). In the spermatogonia this chromatin plate, and with it the idiozome, may lie at any pole of the nucleus, though both are always together; but in the spermatocytes the idiozome is invariably placed against the distal pole of the nucleus, where is also invariably the chromatin plate, and in these cells this chro- matin plate is composed of the ends of two or three chromosomes. Further, when the idiozome disintegrates, the chromatin plate of the nucleus also becomes disestablished. This constant local connection of these two structures suggests a nuclear influence in the formation of the idiozome. So soon as the centrioles wander apart from each other and pass out of the idiozome, the latter disintegrates; therefore it may be that the centrioles also have something to do with the formation of the idiozome, an idea treated specially by Buchner (710). In Euschistus no idiozome develops in the second spermatocytes or in the sperma- tids. In oogonia and the earlier stages of oocytes occur bodies quite comparable to the idiozomes of sperm cells, agreeing in position as well as in containing the centrioles; Van der Stricht (05, ’09) 784 THOS. H. MONTGOMERY, JR. has shown that in the oocyte of the bat they may persist up to the maturation mitoses.” In young oocytes of the guinea pig I have found it to have the appearance described by von Wini- warter. In animal spermatids there are two cytoplasmic bodies which may be combined or separate, and which have been variously known as Nebenkerne and spheres. Meves (’00) has shown that the Nebenkern of La Valette St. George is composed of granules (La Valette’s ‘cytomicrosomes’) which are comparable to the mitochondria of Benda, and the true Nebenkern is thus mivo- chondrial. The other ‘sphere’ of the spermatid Meves (’97, ’02) has called an ‘idiozome,’ homologizing it with the idiozome of a spermatogonium or a spermatocyte. I believe this homology to be a mistaken one, for this sphere of the spermatid behaves quite differently from a true idiozome in being vacuolar and especially in having no connection with the centrioles. There- fore the sphere of the spermatid should no longer be called an idiozome but rather the ‘sphere of the spermatid’ or the ‘spermat- idosphere.’ In Euschistus this sphere of the spermatid, as ina number of insects, undergoes rather complicated movements, then divides into two, and the anterior portion passes to a position at the apex of the sperm and there becomes the lance or perfora- torium; it thus produces the same structure as in most other flagellate spermatozoa. But I do not think it likely that this lance is of any mechanical service in penetrating the egg, for in EKuschistus it is an exceedingly delicate and often branched fila- ment, without any rigidity. Then Yatsu (’07) has shown that it is rather the shape of the nucleus of the sperm that is adapted for entering the egg. In Euschistus this lance enters into the cytoplasm of a follicular nurse cell; therefore it may rather have the value of a nutritive pseudopodium. In spermatocytes of Euschistus another sphere arises, at about the time of disintegration of the idiozome, and it appears 17 Compare especially Conklin ’02, Van der Stricht 704, ’05, 09, von Winiwarter 1900, 09, Bouin, ’05. SPERMATOGENESIS OF EUSCHISTUS 785 to have no relation to the idiozome of the spermatocytes nor to the sphere of the spermatids. It has no local connection with the idiozome, the nucleus or the centrioles, and would appear to be a structure entirely distinct from the idiozome, arising in a different place and at a later time; it has only a short persistence and disappears entirely before the maturation divisions. It is not clear whether this body has any homologue in oocytes, for it does not appear related to the vitellogenic layer (that portion of the yolk nucleus not comprising the idiozome). The other special inclusions of the cytoplasm of the sperm cells are those bodies first recognized by Benda (’98) as specific cell granules, and named by him mitochondria. Henking (’91) was the first to figure them in Hemiptera.!8 Good reviews on these bodies have been written by Meves (’00, ’02b, ’08b), Benda (03), Waldeyer (’01), Korschelt and Heider (’02), Goldschmidt (04). Their great interest lies in the fact that they appear to be constant components of the mature germ cells; and that they develop in embryonic cells into various filar structures such as connective tissue fibrils, myofibrils and neurofibrils, as shown especially by Meves (07a, ’08, 710), Duesberg (10), Korotneff (09), Hoven (10). Interest in them has culminated with the hypothesis of Meves (’08) that they represent an important hered- itary substance which has the same relation to the cytoplasm as the chromosomes to the nucleus, and this has received support from the discovery of Duesberg that they persist during cleavage cells in both animals and plants. And it may be noted that Benda’s original criterion of them, that they stain blue with his 18 Their synonymy is already rather confusing. By ‘mitochondria’ Benda intended granules taking a particular stain, and by ‘chondriomita’ rows of such granules enclosed in plasma threads. Meves (’07b) proposed the name ‘chon- driom,’ later replaced (Meves, ’08) by ‘chondriosome,’ to include both the single granules or mitochondria proper, and also the chains of such granules; for the latter he proposed (’07b) the name ‘chondriokonta,’ a chondriokont differing from a chondriomite in being a chain of granules not enclosed in a plasma thread. To corresponding chains Heidenhain (1900) gave the name ‘pseudochromosomes,’ while Goldschmidt (1904) and some of his followers include the mitochondria under the term ‘chromidia.’ 786 THOS. H. MONTGOMERY, JR. alizarine method, has come to be replaced by the criterion of their genetic origin and fate.!® Here I propose to discuss the mitochondria merely with rela- tion to their mode of origin, mode of division and history during the spermatogenesis, with reference especially to the phenomena in Euschistus. In Euschistus the spermatogonia contain a few granules that never stain as intensely as the mitochondria of the spermatocytes, and are never in the form of threads; it was not determined whether these are mitochondria or idiozome remains, but there is more indication of the latter origin. In various other animals, also, they appear to be few or absent in spermato- gonia, though described for such cells by Meves (’00, ’07), Gérard (09), Giglio-Tos (’08), Wassilieff (07), Lams (09), Dingler (’10), Holmgren (’02), Bouin (’05), Otte (07) and the Schreiners (’05)?° They are also few or absent in oogonia. But in Euschistus and other objects they are always much more abundant in the growth period of the germ cells, and this is certainly the time of their chief elaboration, a point noted also by Buchner (710). This is a rather important matter, and easily proven by comparing the . large mass of them in spermatocytes with their mass in spermato- gonia. The mitochondria thus have their main period of growth and multiplication in the growth period of spermatocytes and oocytes. Their later history, in maturer germ cells, embryonic and tissue cells is in general one of distribution and differentiation. This leads us to the question of their origin in the growth period. Some writers consider them essentially cytoplasmic structures, with no genetic relations to the nucleus; such are Meves, Duesberg, Bouin, Korotneff, Dingler, while Vejdovsky (and his results apply only to oocytes) regards them as portions 19 Probably many of the bodies described in spermatocytes as yolk globules will prove to be mitochondria, as well as many bodies previously confused with idio- zomes. However, Gross has distinguished in Pyrrhocoris yolk granules from mito- chondria. a 20 The large compact masses in the distal ends of spermatogonia, shown by Giglio-Tos in his fig. 1 and regarded by him as mitochondria, are clearly not such but mitosomes; he was misled by their taking the Benda stain, which shows how little service this stain is as a diagnostic. SPERMATOGENESIS OF EUSCHISTUS , 787 of spheres. But it must be said that none of these writers except Vejdovsky have paid particular attention to their origin. Pantel and de Sinéty leave the question open, though they show that the ‘pseudochromosomes’ arise in close contact with the nucleus. Another group of investigators (Dumez, Janssens, Popoff, Was- silieff, Goldschmidt, Buchner, Jorgensen) hold them to be of nuclear origin, for the following reasons: In the pachytene stage of the spermatocytes the chromosomes are frequently, though not in all objects, definitely oriented, radiating in a “bouquet stage’ towards the distal pole of the nucleus, and the idiozome lies at that pole in the cell body; near that pole of the nucleus the mitochondria make their first appearance. This position of the mitochondria, close to a particular pole of the nucleus, is taken to mean that they are produced there by some nuclear activity. Most of these writers maintain that they are engen- dered by actual emigration of chromatin particles out of the nucleus at that pole, as indicated especially by Janssens, Was- silieff, Jérgensen and Buchner. This idea of chromatin emi- gration has been particularly instigated by Goldschmidt’s view that the mitochondria belong in the class of chromidial forma- tions. In Euschistus we found in the resting spermatogonia a particular chromatin plate upon the nuclear membrane, and this is invariably at the point where the idiozome lies; but in these cells there are no demonstrated mitochondria present around the idiozome. In the spermatocytes of this species there is another chromatin plate, here produced by the ends of two or, three chromosomes, again always at the pole where the idiozome lies; in the close vicinity of this idiozome the first mitochondrial chains make their appearance. Pantel and de Sinéty found in Notonecta that the mitochondria arise between the idiozome and the nucleus, therefore in relation to both. In EKuschistus it is hard to make sure whether the mitochondria arise from the cytoplasm, from the idiozome or from the nucleus. But the fact is that they originate in the distal pole of the cell, close to the idiozome and the nucleus, therefore it is probable they are produced by either idiozome or nucleus or by a joint action of these. We saw previously that the idiozome probably 788 THOS. H. MONTGOMERY, JR. stands under the influence of the chromatin plate. In Eu- schistus there is no evidence that the mitochondria are produced by emigrated chromatin particles, and indeed the idiozome intervenes between them and the chromatin plate. Also in this species they certainly have no relation to the idiochromo- somes, which may lie upon any point of the nuclear membrane except the idiozome pole and they do not discharge any visible substance into the cytoplasm.2t Were the mitochondria of strictly cytoplasmic origin it would be difficult to explain why they always arise at a particular pole of the nucleus near the chromatin plate and the idiozome. Therefore it seems a better working hypothesis to conclude that they are produced either by some chemical interaction of idiozome and cytoplasm, or of nucleus and cytoplasm, which would be, in either case, an ultimate nuclear origin. It is interesting to note that their period of early development in the spermatocytes corresponds with the period of conjugation of the chromosomes, and the latter process may be the initial step in their production. This is all in agreement with the concept of the nucleus as the par- ticular formative center of the cell. After the idiozome of Euschistus has disintegrated, and the chromatin plate become disestablished, the mitochondria becomes scattered throughout the cell, and then they become larger and more prominent, evi- dently by autonomous growth. Thus it may well be that the nucleus directly, or acting through the idiozome, gives off fer- ments in small amounts to the cytoplasm and these ferments in their turn engender the mitochondria that later become self- perpetuating structures. Another point of interest is how the mitochondria become distributed in the maturation mitoses. In neither Euschistus nor other forms is there evidence of autonomous division of them in mitosis; they appear rather to become divided passively by the equatorial constriction of the cell; there is no other mechanism for their division, for they lie outside of the spindle and are 21 Both Buchner and Wassilieff hold the mitochondria to be produced by a pour- ing out of substance from the modified chromosomes, though Wassilieff’s results on Blatta have been contradicted by Morse. SPERMATOGENESIS OF EUSCHISTUS _ 789 little influenced by the centrioles. In Euschistus they lie quite irregularly in the spermatocytes, and generally each thread fails to divide in one of the maturation mitoses; further, it is a matter of chance at what point a mitochondrial thread becomes divided. They become irregularly divided in the two maturation divi- sions so that varying amounts of them become apportioned to the spermatids.22 That there is no accurate quartering of the mitochondria in a number of other species results from a study of the figures of various writers, such as the Schreiners (’05, Myxine), Meves (’00, Paludina, Pygaera), Gross (’06, Pyrrho- coris), Wassilieff (’07, Blatta). But in the bee and hornet (Meves, ’07, ’08), in Pamphagus (Giglio-Tos, ’08), in Blaps (Benda, ’03) and Stenobothrus (Gérard, ’09) they appear more evenly arranged around the spindles, and in these objects probably become more regularly divided. But there is good evidence that in certain cases it is a matter of chance how they become divided, in which regard they differ markedly from the chromosomes. The mitochondria in the spermatid of Euschistus coalesce to produce the true Nebenkern, which elongates and forms a pair of narrow bands lying on the sides of the axial thread and extending from the head probably to the end of the tail of the spermatozoon. A similar metamorphosis of the Nebenkern is known for flagellate spermatozoa in a number of invertebrates, while in the mammals the mitochondria engender a spiral fila- ment around the middle piece. There is now, however, consider- able evidence that in many forms of flagellate spermatozoa, such as those of mammals and molluscs, the whole tail enters the egg in fertilization, and is not left outside the egg (contrary to earlier observations); therefore it is probable all the mito- chondrial substance of the sperm enters the egg; much more substance, accordingly, than merely the chromatin of the head. Since then fertilization brings together the mitochondria of two parents, it now becomes of great importance to trace 22 T have previously intimated, (’10b) the possibility that the relative amount of the mitochondrial substance received might determine the sex-preponderance character of a sperm, a matter unfortunately very difficult to test. 790 THOS. H. MONTGOMERY, JR. the history of the mitochondria during the fertilization of the egg, and especially the behavior of those furnished by the sperm. PREFORMATION AND EPIGENESIS IN THE GERMINAL CYCLE, AND SEGREGATION OF THE GERM CELLS IN ONTOGENY The discussion of embryologists upon preformation and epigenesis has treated mainly the phenomena of somatic differ- entiation, the factors regulating the growth of the embryo from the egg. If it be not too premature to speak of a consensus of opinion reached in this discussion, it would be to the effect that epigenesis and preformation are not mutually exclusive, but that both probably proceed at the same time. Here it is my wish to call attention to the fact that in the his- tory of the germ cells may be recognized both preformation and epigenesis, whereby the germinal cycles offer a certain parallel to the somatic. In the spermatogenetic cycle a number of spermatogonial generations occur, during which the number of the chromosomes remains constant and no marked cytoplasmic specializations take place. But in the spermatocytes remarkable differences arise suddenly: the chromosomes group themselves into gemini, the mitochondria increase rapidly in amount, the whole cell becomes larger, and frequently the centrioles take on unusual forms and positions (as upon the cell membrane). Somewhat similar changes occur in the oocytes, and more complex ones, owing to the production of yolk substance. The spermatids often undergo a marked metamorphosis. As one reviews the sum total of these processes the conviction arises that there are here in the completest form both preformation and epigenesis. The chromosomes are on the whole the most stable parts, appar- ently continuous from generation to generation, and though they may pass through marked changes in forming the sperm nucleus, they later emerge, in fertilization, under the same forms that they had previously. On the whole they seem to be the particular preformed bodies of the germ cells. But this is not the case with certain other cell constituents. For in Eu- schistus, which seems to exemplify in main features the sperma- SPERMATOGENESIS OF EUSCHISTUS 791 togenetic relations of most insects, a true idiozome arises in the spermatocytes, mitochondria develop around it, the idiozome disintegrates and a sphere appears, thissphere disappears entirely and in the spermatid another sphere is produced that originates another new body, the perforatorium. It is clear that the first spermatocytes and oocytes are the most interesting cells of their respective cycles, for they exhibit the most significant processes—conjugation of chromosomes, reduction division, elaboration of mitochondria. In them the history of the cytoplasmic parts is markedly epigenetic. And another series of epigenetic changes is exhibited by the histo- genesis of the spermatozoon. There is then a parallel with the somatic cycle, which begins with an undifferentiated and terminates with a highly differen- tiated condition. The ripe ovum and spermatozoon are much more differentiated than the spermatogonium or oogonium; each of them enters, with the commencement of the growth period, upon its period of specialization. The recognition of this resemblance may throw light upon the problem of the segregation of the germ cells. By what process is it that certain cells of the embryo are held back from somatic differentiation to become germ cells? In other words, why do not all the cells become specialized? The answer, it seems to me, is to be sought in the distribution of the specializa- tions of the fertilized egg to the cells of the embryo. One set of specialized structures of the germ cells, the mitochondria, are now known to give rise to various important specializations of body cells. The mitochondria, that are elaborated, in greatest — part at least, during the growth period of the germ cells, persist from the fertilized egg into cleavage stages, and ultimately transform into various specialized fibrillar structures. With this in mind the setting aside of germ cells from body cells could be explained mechanically as follows: any cleavage cell which failed to receive mitochondria, or failed to receive particu- lar ones or a particular amount of them, would be incapacitated from engendering such somatic specializations, it would thereby become a germ cell. This might appear contrary to the idea 792 THOS. H. MONTGOMERY, JR. that the body cells become different from the germ cells by a mechanical process of chromatin diminution, as in Ascaris. But this is quite conformable to our argument, for in Ascaris those cells which become body cells are the ones that include the cast-off chromosome ends in their cytoplasm, and it will prob- ably be found that these ejected chromosome parts engender such cytoplasmic differentiations as characterize the body cells. Either this mechanism of the segregation of the germ cells may be admitted, or else one that would cause the mitochondria of prospective germ cells to remain unaltered or latent. For- tunately this is a matter that may be tested by observation. For if the first supposition be correct we should find that during cleavage an unequal distribution of the mitochondria occurs, just as we know occurs in the case of yolk granules. If the second be correct, then while the mitochondria are undergoing developmental changes in some of the cells, they should be found to remain relatively unaltered in others—the prospective germ cells. LITERATURE CITED BauMGARTNER, W. J. 1902 Spermatid transformations in Gryllus assimilis, etc. Kansas Univ. Sci. Bull. vol. 1. Benpba, C. 1891 Neue Mitteilungen iiber die Entwickelung der Genitaldriisen und iiber die Metamorphose der Samenzellen. Arch. Anat. Phys. 1898 Ueber die Spermatogenese der Vertebraten und héheren Everte- braten. Verh. Physiol. Ges. Berlin. 1903 Die Mitochondria. Ergebn. Anat. Entw., Bd. 12. Bereus, J. 1904 La formation des chromosomes héterotypiques dans la sporo- bd genése végétale. 1. La Cellule, tome 21. Briackman, M. W. 1903 The spermatogenesis of the myriapods. II. Biol. Bull., vol. 5. 1905a The spermatogenesis of the myriapods. III. Bull. Mus. Comp. Zool., Harvard, vol. 48. 1905b The spermatogenesis of the myriapods. IV. Proc. Amer. Acad. Arts and Sci., vol. 41. 1907 The spermatogenesis of the myriapods. V. Ibid. vol. 42. BonnEvig, K. 1906 Untersuchungen tiber Keimzellen. I. Jena. Zeit. Bd. 41. 1908 Chromosomenstudien. 11. Arch. Zellforsch. Bd. 2. SPERMATOGENESIS OF EUSCHISTUS 793 Borina, A. M. 1907 A study of the spermatogenesis of twenty-two species of the Membracidae, Jassidae, Cercopidae and Fulgoridae. Journ. Exp. Zool., vol. 4. Boveri, T. 1887 Zellen-Studien. 1. Jena. Bouin, P. 1905 Ergastoplasme, pseudochromosomes et mitochondria. Arch. Zool. Expér. (4) tome 3. Braver, A. 1893 Zur Kenntniss der Spermatogenese von Ascaris megalo- cephala. Arch. mikr. Anat., Bd. 42. Bucuner, P. 1909 Das accessorische Chromosom in Spermatogenese und Ovo- genese der Orthopteren, etc. Arch. Zellforsch., Bd. 3. 1910 Von den Beziehungen zwischen Centriol und Bukettstadium- Ibid., Bd. 5. Cuuss, G. C. 1906 The growth of the oocyte in Antedon, etc. Phil. Trans. Roy. Soc. London, vol. 198. Conxuin, E.G. 1902 Karyokinesis and cytokinesis, etc. Journ. Acad. Nat. Sci. Philadelphia, vol. 12. Coox, M. H. 1910 Spermatogenesis in Lepidoptera. Proc. Acad. Nat. Sci. Philadelphia. Davis, H. S. 1908 Spermatogenesis in Acrididae and Locustidae. Bull. Mus. Comp. Zool., Harvard, vol. 53. Dinater, M. 1910 Ueber die Spermatogenese des Dicrocoelium lanceatum Stil. und Hass. Arch. Zellforsch., Bd. 4. Duerssere, J. 1908 La spermatogénése chez le rat. Leipzig. 1910a Les chondriosomes des cellules embryonnaires du lapin et leur role dans la génése des myofibrilles. Arch. Zellforsch., Bd. 4. 1910b Observations sur la structure du protoplasme des cellules vége= tales. Anat. Anz., Bd. 36. 1910c Sur la continuité des éléments mitochondriaux des cellules sex- uelles et des chondriosomes des cellules embryonnaires. Ibid. DumeEz, R. 1902 Rapports du cytoplasme et du noyau dans l’oeuf de la Cytherea chione L. La Cellule, tome 19. Eisen, G. 1900 The spermatogenesis of Batrachoseps. Jour. Morph., vol. 17. Farmer, J. B. and Moorr, J. E. 8S. 1905 On the maiotic phase (reduction divisions) in animals and plants. Quart. Journ. Micr. Sci., vol. 48. Fick, R. 1907 Vererbungsfragen, Reductions- und Chromosomenhypothesen, Bastardregeln, Ergebn. Anat. Entw., Bd. 16. 1908 Zur Konjugation der Chromosomen. Arch. Zellforsch. Bd. 1. Foor, K. and Strosetu, E.C. 1909 The nucleoli in the spermatocytes and ger- minal vesicles of Euschistus variolarius. Biol. Bull., vol. 16. 794: THOS. H. MONTGOMERY, JR. Gérarp, P. 1909 Recherches sur la spermatogénése chez Stenobothrus biguttu- lus (Linn.). Arch. Biol., tome 24. Gieu10-Tos, E. 1908 I mitocondri nelle cellule seminali maschili di Pamphagus marmoratus (Burm.). Biologica, tome 2. GoupscumipT, R. 1904 Der Chromidialapparat lebhaft funktionierender Ge- webszellen. Ete. Zool. Jahrb., Bd. 21. GoOLDScHMIDT, R. und Poporr, M. 1907 Die Karyokinese der Protozoen und der Chromidialapparat der Protozoen- und Metazoenzelle. Arch. Protistenk. Bd. 8. GREGOIRE, V. 1904 La réduction numerique des chromosomes et les cinéses de maturation. La Cellule, tome 21. 1905 Les résultats acquis sur les cinéses de maturation dans les deux regnés. Ibid. tome 22. 1910 Les cinéses de maturation dans les deux regnés. Ibid., tome 26. GREGOIRE, V.et DeTonN. 1906 Contribution 4 1’ étude de la spermatogénése dans VOphryotrocha puerilis. Ibid., tome 23. Gross, J. 1904 DieSpermatogenese von Syromastes marginatus L. Zool. Jahrb., Bd. 20. 1906 Die Spermatogenese von Pyrrhocoris apterus L. Ibid., Bd. 23. Guyer, M.F. 1900 Spermatogenesis of normal and of hybrid pigeons. Chicage. HAckeR, V. 1895 Die Vorstadien der Eireifung. Arch. mikr. Anat., Bd. 45. 1904 Bastardierung und Geschlechtszellenbildung. Festsch. f. Weis- mann. 1910 Ergebnisse und Ausblicke in der Keimzellenforschung. Zeit. indukt. Abstamm. und Vererbungsl., Bd. 3. Heiwennain, M. 1900 Die Zentralkapseln und Pseudochromosomen in den Samenzellen von Proteus. Anat. Anz., Bd. 18. HeNKING, H. 1891 Untersuchungen iiber die ersten Entwicklungsvorgingen in den Hiern der Insekten. Zeit. wiss. Zool., Bd. 51. Hertwic, O. 1890 Vergleich der Ei- und Samenbildung bei Nematoden. Arch. mikr. Anat., Bd. 36. Houmcren, N. 1902 Ueber den Bau der Hoden und die Spermatogenese von Silpha carinata. Anat. Anz., Bd. 22. Hoven, H. 1910 Sur l’histogénése du systéme nerveux periphériqle chez le poulet et sur le réle des chondriosomes dans la neurofibrillation. Arch. Biol., tome 25. JANSSENS, F. A. 1905 Evolution des auxocytes mAles du Batrachoseps attenua- tus. La Cellule, tome 22. SPERMATOGENESIS OF EUSCHISTUS 795 Jorpan, H. E. 1908 The spermatogenesis of Aplopus mayeri. Publ. Carnegie Inst., Washington. Jorgensen, M. 1909 Beitriige zur Kenntnis der Wibildung, etc., bei Schwim- men. Arch. Zellforsch., Bd. 4. Korotnerr, A. 1909. Mitochondrien, Chondriomiten und Faserepithel der Tricladen. Arch. mikr. Anat., Bd. 74. Korscuett, E. 1895 Ueber Kernteilung, ete., bei Ophryotrocha puerilis. Zeit. wiss. Zoll., Bd. 60. Korscuett, E. und Herper, K. 1903 Lehrbuch der vergleichenden Entwick- lungsgeschichte der wirbellosem Tiere. Jena. Lams, H. 1908 Les divisions des spermatocytes chez la fourmi (Camponotus herculeanus L.). Arch. Zellforsch., Bd. 1. Lenuossix, M. vy. 1898 Untersuchungen iiber Spermatogenese. Arch. mikr. Anat., Bd..51. McCuune, C. E. 1900 The spermatocyte divisions of the Acrididae. Bull. Univ. Kansas. McGiut, C. 1904 The spermatogenesis of Anax junius. Univ. Missouri Studies. No. 2. Maziarskr, S. 1910 Sur les changements morphologiques de la_ structure nucléaire dans les cellules glandulaires. Arch. Zellforsch., Bd. 4. Mepzs, G. 1905 The spermatogenesis of Scutigera forceps. Biol. Bull., vol. 9. Meves, F. 1896 Ueber die Entwicklung der minnlichen Geschlechtszellen von Salamandra maculosa. Arch. mikr. Anat., Bd. 48. 1897 Zelltheilung. Ergebn. Anat. Entw., Bd. 6. 1899 Ueber Struktur und Histogenese der Samenfiiden des Meer- schweinchens. Arch. mikr. Anat., Bd. 54. 1900 Ueber den von la Valette St. George entdeckten Nebenkern (Mitochondrienkérper) der Samenzellen. Ibid., Bd. 56. 19022 Ueber oligopyrene und apyrenen Spermien und tber ihre Ent- st hung, ete. Ibid., Bd. 61. 1902b Struktur und Histogenese der Spermien. Ergebn. Anat. Entw., Bde lar 1907a Die Spermatocytenteilungen bei der Honigbiene, ete. Arch. mikr. Anat., Bd. 70. 1907b Ueber Mitochondrien bezw. Chondrokonten in den Zellen junger Embryonen. Anat. Anz., Bd. 31. 1907c Die Chondriokonten in ihrem Verhiltnis zur Filarmasse Flem- mings. Ibid. 1908a Die Spermatozytenteilungen bei der Hornisse, ete. Arch. mikr. Anat., Bd. 71. 1908b Die Chondriosomen als Triiger erblicher Anlagen, etc. Ibid., Bd. 72. ~ 796 THOS. H. MONTGOMERY, JR. Monteomery, T. H., Jr. 1898 The spermatogenesis in Pentatoma up to the formation of the spermatid. Zool. Jahrb., Bd. 12. 1899 Cytological studies, with especial reference to the morphology of the nucleolus. Jour. Morph., vol. 15. 1900 The spermatogenesis of Peripatus (Peripatopsis) balfouri up to the formation of the spermatid. Zool. Jahrb., Bd. 14. 1901 A study of the chromosomes of the germ cells of Metazoa. Trans. Amer. Phil. Soc., vol. 20. 1903 The heterotypic maturation mitosis in Amphibia and its general significance. Biol. Bull., vol. 4. 1904 Some observations and considerations upon the maturation phenomena of the germ cells. Ibid., Bd. 6. 1905 The spermatogenesis of Syrbula and Lycosa, etc. Proc. Acad. Nat. Sci., Philadelphia. 1906 Chromosomes in the spermatogenesis of the Hemiptera Heterop- tera. Trans. Amer. Phil. Soc. N. S., vol. 21. 1910a On the dimegalous sperm and chromosomal variation of Euschistus, etc. Arch. Zellforsch., Bd. 5. 1910b Are particular chromosomes sex determinants? Biol. Bull., vol. 19. Morriti, C. V. 1910 The chromosomes in the oégenesis, fertilization and cleav- age of Coreid Hemiptera. Biol. Bull., vol. 19. Morse, M. 1909 The nuclear components of the sex cells of four species of cockroaches. Arch. Zellforsch., Bd. 3. Nicuots, M. L. 1910 The spermatogenesis of Euchroma gigantea. Biol. Bull., vol. 19. Orrr, H. 1907 Samenreifung und Samenbildung bei Locusta viridissima. Zool. Jahrb., Bd. 24. PANTEL, J. et SinftTy, R. pe 1906 Les cellules de la ligneé male chez le Noto- necta glauca L. ete. La Cellule, tome 23. PautmigR, F. C. 1899 The spermatogenesis of Anasa tristis. Jour. Morph., vol. 15, Supplement. PLATNER, G. 1899 Beitrige zur Kenntniss der Zelle und ihrer Teilungser- scheinungen. Arch. mikr. Anat., Bd. 33. Poporr, M. 1907 Eibildung bei Paludina vivipara und Chromidien bei Palu- dina und Helix. Ibid., Bd. 70. Ratu, O. vom 1895 Neue Beitrige zur Kenntnis der Chromatinreduction der Samen- und Eireife. Ibid., Bd. 46. Rickert, J. 1893 Zur Eireifung bei Copepopen. Anat. Hefte, Bd. 4. 1894 Die Chromatinreduction bei der Reifung der Sexualzellen. Ergebn. Anat. Entw., Bd. 3. SPERMATOGENESIS OF EUSCHISTUS 797 ScuHaFFNER, J.H. 1897 The division of the macrospore nucleus in Lilium. Bot. Gaz., vol. 23. ScurEIner, A. und K. E. 1904 Die Reifungsteilungen bei den Wirbeltieren° Etc. Anat. Anz., Bd. 24. 1905 Ueber die Entwicklung der minnlichen Geschlectszellen von Myxine glutinosa. Arch. Biol., Bd. 21. 1906a Die Reifung der minnlichen Geschlechtszellen von Salamandra maculosa (Laur.), Spinax niger (Bonap.), und Myxine glutinosa (L.). Ibid., Bd. 22. 1906b Die Reifung der Geschlechtszellen von Ophryotrocha puerilis. Anat. Anz., Bd. 29. 1907 Die Reifung der Geschlechtszellen von Enteroxenos oestergreni. Vid. Selsk. Skr. 1908a Die Reifung der Geschlechtszellen von Zoogonus mirus. Ibid. 1908b Zur Spermienbildung der Myxinoiden. Arch. Zellforsch., Bd. 1. Stevens, N.M. 1905 Studies in spermatogenesis with especial reference to the ‘accessory chromosome.’ Publ. Carnegie Inst. Washington. 1906 Studies in spermatogenesis. 11. Ibid. 1910 An unequal pair of heterochromosomes in Forficula. Jour. Exp. Zool., vol. 8. Stomps, T. J. 1910 Kerndeeling en Synapsis bij Spinacia oleracea L. Amster- dam. Stricut, O. vAN pER 1904 La structure de Voeuf des mammiféres. Arch. Biol., Bd. 21. 1905 Structure de l’oeuf ovarique de la femme. Bull. Acad. Med. Belg. 1909 Lastructure de l’oeuf des mammiféres (chauve-souris, Vesperugo noctula). 3me Partie. Mem. Acad. Roy. Belg., Bd. 2h. 2. VEJDOVSKY, F. 1907 Neue Untersuchungen iiber die Reifung und Befruchtung. Prag. Van Benepen, E. 1883 Recherches sur la maturation de l’ceuf, etc. Arch. Biol., Bd. 4. Waxpryer, W. 1906 Die Geschlechtszellen. in O. Hertwig’s Handbuch d. vergl. u. exper. Entwickelungslehre d. Wirbeltiere. Wauiacr, L. B. 1909 The spermatogenesis of Agalena naevia. Biol. Bull., vol. Wie Wassiuigrr, A. 1907 Die Spermatogenese von Blatta germanica. Arch. mikr. Anat., Bd. 70. Wicox, E. V. 1895 Spermatogenesis of Caloptenus femur-rubrum and Cicada tibicen. Bull. Mus. Comp. Zool., Harvard, vol. 27. JOURNAL OF MORPHOLOGY, VOL. 22, NO. 3 798 THOS. H. MONTGOMERY, JR. Witson, E. B. 1905 Studies on chromosomes. J. Journ. Exp. Zool., vol. 2. 1905b Studies on chromosomes. II. Ibid. 1906 Studies on chromosomes. III. Ibid., vol. 3. WrIntwarRTER, H.v. 1900 Recherches sur |’ovogénése et organogénése de!’ ovaire des mammiféres (lapin et homme). Arch. Biol., Bd. 17. WINIWARTER, H. v. et Sarnmont 1909 Nouvelles recherches sur l’ovogénése de l’ovaire des mammiféres (chat). Ibid., Bd. 24. Yatsu, N. 1907 A note on the adaptive significance of the sperm-head in Cere- bratulus. Biol. Bull., vol. 18. ZweIGER, H. 1906 Die Spermatogenese von Forficula auricularia L. Jena. Zeit., Bd. 42. PLATES All the figures have been drawn to the same scale with the aid of the camera lucida at the level of the base of the microscope, with. the Zeiss apochromatic immersion objective 1.5mm., and ocular 12; the original dimensions have been reduced one-fifth. The following abbreviations have been employed: c. centriole m. minute chromosome ch. c. chromatoid corpuscles mit. mitosome (spindle remains) D. larger idiochromosome pf. perforatorium d. smaller chromosomes pl. plasmosome id. idiozome sp. sphere PLATE 1 EXPLANATION OF FIGURES 1 and 8 from testis no. 103, all others from testis no. 282. 1 Penultimate spermatogonium, rest; (follicle 5). 2 The same, pole view of equatorial plate; (follicle 6). 3 Portion of a rosette of ultimate spermatogonia; (follicle 6). 4 Anaphase of ultimate spermatogonium; (follicle 4). 5 Later anaphase of the same; (follicle 4). 6-41 Successive stages of the early growth period of the first spermatocytes. All the cells from follicle 4 except the following: Figs. 23, 24 from follicle 5; figs. 10, 11, 15, 22 from follicle 6. 799 SPERMATOGENESIS OF AN HEMIPTERON T. H. MONTGOMERY JOURNAL OF MORPHOLOGY, VoL. 22, no. 3 4 5 PLATE 1 a..f e. | N T. H. Montgomery, del, PLATE 2 EXPLANATION OF FIGURES Successive stages of first spermatocytes, all from testis no. 282; all from follicle 4, except fig. 49 from follicle 5. 803 HEMIPTERON AN SPERMATOGENESIS OF T. H. MONTGOMERY 22, no. 3 VOL. JOURNAL OF MORPHOLOGY, PLATE 2 Tr, H. Montgomery, del. Mt Ds PLATE 3 EXPLANATION OF FIGURES Successive stages of first spermatocytes, figs. 74, 75, 82, 83 from testis, no. 286; 86-93 from testis no. 265; the remaining figures from testis no. 282. Figs. 74-83, 86-89, 91-93 from follicle 6, the remaining from follicle 4. 69-86 Later prophases of the first maturation spindle. 87-93 First maturation spindle. 807 SPERMATOGENESIS OF AN HEMIPTERON T. H. MONTGOMERY JOURNAL GF MORPHOLOGY, VOL. 22, No.3 PLATE 3 T. H. Montgomery, del. PLATE 4 EXPLANATION OF FIGURES 111 from testis no. 282, the others from testis no. 265. All the figures from follicle 6 except the following: 98 and 100 from follicle 4; 95-97, 110 from follicle 5. 94 Metaphase of first maturation. 95-100 Anaphases of first maturation, fig. 98 being a polar view of a daughter chromosomal plate. 101 Polar view of metaphase of second maturation. 102-104 Lateral views of second maturation spindles. 105-106 Polar views of daughter chromosomal plates of the second maturation spindle. 107-111 Lateral views of later second maturation spindles. 112 An entire spermatid (on the right) nearly completely separated from its sister (shown in part outline on the left). 113-116 Earliest stages in the histogenesis of the sperm. S11 SPERMATOGENESIS OF AN HEMIPTERON T. H. MONTGOMERY 115 eee JOURNAL OF MORPHOLOGY, VOL, 22, No. 3 PLATE 4 Tt. H. Montgomery, del. PLATE 5 EXPLANATION OF FIGURES Histogenesis of the sperm. All the figures represent lateral views, except 145, which illustrates a transection of a sperm bead of the stage of fig. 144; the entire spermatozoon is shown in 117-126, in the others only the head and the proximal part of the tail are represented. All the figures are of cells from follicle 6; 117-122 are from testis no. 265 ; 1385-137 from testis no. 116; the remainder from testis no. 282. 815 SPERMATOGENESIS OF AN HEMIPTERON T. H. MONTGOMERY JOURNAL OF MORPHOLOGY, voL. 22, no.3 5 PLATE T. H. Montgomery, del. THE LIFE HISTORY OF THE SCOLEX POLYMORPHUS OF THE WOODS HOLE REGION WINTERTON C. CURTIS From the Zoological Laboratory, University of Missouri THIRTEEN FIGURES CONTENTS TBGTOGINCHIOM a 75 tee eee te None tee Bene oo oasis o cee ho Ghee 819 Methodah.k tecnico. sea act se is eee Ce. EB dk ds i ie ee oe re 820 fhematare-of- the Scolex polymorphus!% 4.0.2)..¢04.. 02 04k soa oe ee 821 The normal occurrence of parasites in sand sharks taken at random....... 827 The infection of sand sharks with the Scolex polymorphus................. 829 MATTE eee es Sie ee mn Sh fl ok cee on a A ee 849 Diaberatireycited ssh a i ke Pats he chee: penis esa oe ah aed eee ee eee 848 INTRODUCTION In attempting to determine the life cycle of Crossobothrium laciniatum, a cestode found in the sand shark, (Carcharias lit- toralis) of the Woods Hole region, attention was at once directed to the cestode larva known as the Scolex polymorphus, which seemed not unlikely to be the young of this species. The studies here described were begun in the hope that these two forms would prove to be a single species, since both are admirable for labo- ratory purposes and have been so used by students at the Marine Biological Laboratory for many years. It appears, however, that such a relationship does not exist; for theresultsof the exper- iments give strong, though perhaps not entirely conclusive, evi- dence that the Scolex polymorphus develops into the species Phoreiobothrium triloculatum (Linton) and show conclusively that it cannot be the young of Crossobothrium laciniatum. 819 JOURNAL OF MORPHOLOGY, VOL. 22, NO. 3 820 WINTERTON C. CURTIS The material was collected and the experiments performed during the summers of 1903 and 1904 at the Marine Biological Laboratory and the United States Fisheries Bureau Laboratory at Woods Hole, Mass., and I am indebted to those in charge of these institutions for substantial aid in the prosecution of the work. METHODS The sharks used for infection and the squeteague from which the specimens of the Scolex polymorphus were obtained, were all taken during the summer months in the fish traps near Woods Hole. The sharks were marked by means of numbered copper tags, fastened to the dorsal fin and were kept in wooden fish cars about 5 x 6 x 14 feet, several of which were fastened together to form a float upon which much of the work could be conveniently carried on. At the outset it was necessary to devise some appa- ratus by which the animals could be held securely in a position convenient for any necessary operation and which could be manip- ulated with safety by a single person. The holder which was finally constructed consisted of a trough about four and one-half feet in length, formed by nailing two boards together and across their ends two shorter strips after the fashion of a farmer’s ‘pig- trough.’ The top of this trough was covered by a hinged lid which, when fastened down, left the head and tail exposed but held the body of the fish securely along the greater part of its length. The cross piece of one end was hinged to the surface of the float where the holder was being used and to the free end was hinged a support, which, when swung out, held the contrivance at any desired height as the free end was lifted toward the upright position. By working rapidly, it was possible, with the fish thus held securely, to complete the necessary treatment in a few moments, and any more elaborate apparatus providing for the irrigation of the gills seemed unnecessary, since the sharks gave no indication that their vitality had been impaired by this brief exposure to the atmosphere. In operation this holder was safe and in every way satisfactory. The shark was dipped up with THE SCOLEX POLYMORPHUS 821 a hand net and placed in the trough ventral side uppermost. Upon fastening its lid, the holder was at once raised to a conve- nient position and the operator could then quickly introduce any desired object into the mouth cavity. In feeding, the jaws were pried open, a piece of food placed in the mouth and after being guided past the gill arches was thrust down the oesophagus by means of a billet of wood. The oil of male fern and the calomel, used in the attempts to rid the sharks of their previous infection, were given in gelatin capsules having a capacity of 2cc. These were placed in a piece of galvanized iron pipe, which had been pushed down the oesophagus until it reached the stomach, and the capsule foreed down the pipe with a small wooden rammer. This proceeding, or the giving of food, could be accomplished with comparative rapidity and the sharks were often back in the ear within two or three minutes after being taken out. There was no sign of the regurgitation of food or drugs, though the bottoms of the cars were carefully inspected during the hours just after treatment, and both food and drugs were found in the digestive tracts of specimens which happened to be examined during the first few days. There can, therefore, be no doubt that both remained in the stomach when once introduced. Further details of methods used are given at the appropriate places throughout the paper. THE NATURE OF THE SCOLEX POLYMORPHUS Before describing the experiments in which the larval cestode known as the Scolex polymorphus was used to infect the sand shark, it may be well to offer a word of explanation regarding this form. The name 8S. polymorphus has been applied by Linton to a cestode larva found in a considerable number of fishes from the Wood Hole region. While most abundant in the squeteague (Cynoscion regalis) and the common flounder (Paralicthys den- tatus), he has found it in varying numbers in some twenty-eight other teleosts, the list of which is given on page 413 of his 1901 report. This larva, which I have represented in figs. 1, 2, 3,, 4, 12 and 13 of this paper, is frequently met with in a stage slightly 822 WINTERTON C. CURTIS : younger than the one here shown. In such a young condition it lives in the intestine of its teleost host, moving freely about and attaching itself by means of its four suckers. The slightly older stage which these figures represent and which was used in my experiments, is found in the cystic duct of some of these fishes, notably the squeteague and the common flounder. ‘The speci- mens which I used for infection purposes all came from the first | of these two hosts. The name Scolex polymorphus originated with Rudolphi (08) and has since been frequently applied to such larval forms taken from many different fishes, and by Van Beneden (’50) from crabs of the genera Carcinus and Eupagurus. Zschokke (’86), discussing such forms, showed that the mono-, bi- and triloculate types of bothria (figs. 2, 12 and 13 of the present paper), which can be dis- covered when any considerable number of individuals are exam- ined, are only developmental stages of the same form, and his results indicated that the 8. polymorphus with which he worked was the young of the genus Calliobothrium, and not of-Oncho- bothrium as had been previously suggested. These authors of course studied specimens from European waters. Monticelli (’88) in an extensive paper upon the S. polymorphus of the region about Naples, showed that the larvae, which he found in a large number of fishes, though most common in the flounders, were the young of Calliobothrium filicolle. This conclusion was based upon the close anatomical resemblances between the more advanced larvae and young specimens of C. filicolle and upon experiments in which a species of Torpedo, after being freed of all parasites by starving (a method which his experience and that of the collectors at the Naples station had shown to be effective), was then fed for a time upon specimens of Arnoglossus known to contain the 8. polymorphus. As a result of this, young specimens of the C. filicolle were obtained from the torpedoes so treated, and this taken with the anatomical resemblances seemed conclusive evidence. Monticelli has re- viewed the literature exhaustively and he gives (p. 89) a list of thirty-six cases in which authors have appended various specific names to the term scolex and made as many different species THE SCOLEX POLYMORPHUS 823 of this single form. He refers to these names as synonyms and apparently considers all the forms from whatever host to be the young of a single species. Although my experiments indicate that the 8. polymorphus of the Woods Hole region develops into Phoreiobothrium triloculatum and not into a species of Calliobothrium, a comparison of the larvae, as taken from the fishes about Woods Hole, with the de- seription which Monticelli gives shows that the two types are closely similar, probably almost indistinguishable. The myzo- rhynchus, bothria with one, two or three loculi, two faint red pigment spots in the neck region of some of the specimens, the general shape of the body and the characteristic movements, are apparently identical. Only in the case of the hosts they inhabit is there any very apparent difference, which is of course neces- sitated by the differences in the piscine faunas of two such widely separated regions. In his paper published in 1897, Linton suggested that the S.. polymorphus perhaps represents the young of a number of differ- ent cestodes, and also that none of the fishes (teleosts) in which he has found it is the true host of either larva or adult. He considers such larvae when found in teleosts to be ‘xencsites,’ or misplaced parasites, and thinks that the true intermediate hosts may be found among the species of crabs which frequent the feeding grounds of these fishes. In support of this view, he cites the fact that the existence of similar larvae in crustacea has been recorded by Van Beneden (’59). Should this hypothesis prove correct, we should have a case where the transfer from such a host to a teleost fish, while not fatal to the parasite, still presents conditions under which it can develop but a little way beyond the stage already attained. Basing his conclusion largely upon the presence of a median proboscis-like structure (the myzo- rhynchus) at the anterior end between the bothria, Linton (’97) expressed the opinion that our Scolex polymorphus is the young of the genus Echeneibothrium. This is not, however, as strong a clew as might seem, for in the adult Calliobothrium filicolle, to which Monticelli’s larvae developed, this structure is quite degenerate, though well marked in the larva, while in the adult 824 WINTERTON C. CURTIS of Calliobothrium leucartii and Calliobothrium verticillatum there is no trace of such a structure. In a later paper on the parasites from the fishes of Beaufort, North Carolina, Linton (’04) finds the Scolex polymorphus in many of the fish examined and speaks of the larvae as follows (p. 326): The larval cestodes, doubtless representing several genera, recorded in Parasites of the Woods Hole Region under the name of Scolex poly- morphus, were found in thirty-four of the fifty-nine Beaufort fishes examined. As at Woods Hole these forms are found not only in the alimentary canals of their hosts but also in the cystic ducts of several. They are almost never absent from the cystic duct of Cynoscion regalis. In all cases, where these worms have been obtained from the cystic duct and from the intestine of the same fish, those coming from the cystic duct are larger, plumper, and more opaque than those from the intes- tine. Some of the older larvae suggest the genera Calliobothrium, Acanthobothrium and Phoreiobothrium. Again, in speaking of the parasites of the sharp-nosed shark, Seoliodon terrae-novae, under Phoreiobothrium triloculatum, he says (p. 348): 1 scolex, no segments yet developed; length 2 mm.; hooks small. This specimen looks very much like some of the more advanced specimens of Scolex polymorphus which have occasionally been found, save that the bothria have assumed the characteristics of P. triloculatum. On page 359, under the parasites of the pipe fish, Siphostoma fuscum, he notes that the specimens of the Scolex polymorphus had ‘‘bothria with two costae and rudiment at anterior end, suggesting loculi which occur at the anterior end of bothria in Echeneibothrium and Acanthobothrium; no red pigment.” On page 407 under the parasites of the toad fish, Opsanus tau, he speaks of a specimen which was ‘“‘probably a young Callio- bothrium,”’ and of another which ‘‘had the characteristic bothria of Echeneibothrium and Rhinebothrium. Its prominent muscu- lar proboscis, (myzorhynchus), if retained in the adult would place it in the former genus.’’ Again, in another lot, “‘ The largest had bothria which resembled those of Calliobothrium and Acan- THE SCOLEX POLYMORPHUS 825 thobothrium, but without hooks.’’ And finally, others which had ‘‘red pigment, two costae, one specimen noted with rudi- mentary hooks (Calliobothrium or Acanthobothrium)” and in another lot a specimen is recorded with the ‘“‘rudimentary hooks and pigment spots.” Under the parasites of the sole, Symphurus plagusia, there are mentioned specimens which are ‘‘comparatively large, with two costae and red pigment like young Acanthobothrium, but without hooks.” Fig. 80, plate 12, shows a young specimen of Calliobothrium with rudimentary hooks, but otherwise much like the Scolex polymorphus. In view of these later observations of Linton and Monticelli’s results, one would expect the Scolex polymorphus from about Woods Hole to develop into one of the species of Calliobothrium found in this region, or perhaps some other species of the family Onchobothriidae, and this last is what I believe happens in the case of the larvae with which my experiments were performed. Since two species of the genus Calliobothrium (C. verticilatum and C. eschrichtii) have been found in our region, by Linton (99) who records this species from the dogfish (Mustelus canis), and since a considerable number of species belonging to most of the genera of the family Onchobothriidae have been described from Woods Hole by Linton, it would seem not at all unlikely that experiments made by feeding the Scolex polymorphus from the various teleosts to skates, dogfish and sharks might connect these larvae with other genera of the Onchobothriidae. Such experiments would be likely to give precise evidence for or against Linton’s belief that the larvae represent the young of more than one form and they might give us data for further consideration of the whole question of xenositism, which Linton suggests is the condition of these larvae when found in teleosts. In considering the possibility that the various forms of erus- tacea are the true intermediate hosts in which the development was begun, I have made a careful tabulation of the food of these fishes as recorded mainly by Linton (’99), but also in more detail for a smaller number of fishes in tne work of Peck (’95). This 826 WINTERTON C. CURTIS tabulation is not given since it is merely a compilation and the important point ascertained can be briefly stated; namely that crustacea of various sorts, shrimps, amphipods, isopods, crabs, etc., are an important food with almost all these fish. In cases like the squeteague and blue-fish, where they are not so important an item, it is noticeable that various crustacea-eating fish are a common food. This would account for the great numbers of the Scolex in the squeteague, which would then be like a sieve in which were retained many larvae which had come originally from crustacea through the medium of another fish. The flounder, Paralicthys dentatus, is the other fish in which Linton has found the Scolex most abundant. Its food consists of smaller fish and a large proportion of various crustacea, so that in this case the S. polymorphus might be obtained directly from crustacea, or indirectly from another fish. I have also tabulated the food of the fishes from which Linton has not recorded the Scolex to see whether crustacea form as large an element in their food supply, but no satisfactory facts can be gathered for the reason that the list of those containing the larva comprises the greater number of our smaller and more common fishes and because, as Linton expressly states, no systematic search has been made and hence the fact that the larvae have not been recorded from any fish may have little importance. I quite agree with Linton’s suggestion that this widely distrib- uted larva, though it does not resolve itself into several easily rec- ognizable types in the larval condition, may eventually be shown to represent the young of more than one cestode, and if I am cor- rect in my conclusion that the 8. polymorphus with which my experiments were made develops into Phoreiobothrium trilo- culatum, whereas the form with which Monticelli worked is the young of Calliobothrium filicolle, this may be the beginning of evidence which will give Linton’s interpretation a secure foun- dation and thus the name ‘polymorphus,’ which seems originally to have been given because an individual larva of this sort can assume such diversity of shape, may come to have a new signifi- cance from the existence of many species under a guise which does not show differences by which each may be recognized. . THE SCOLEX POLYMORFHUS 827 The close resemblance of our 8. polymorphus to the forms upon which Monticelli (op. cit.) worked, makes a discussion of the anat- omy superfluous beyond what is shown by my figures and their explanations which have been made quite full. The differences are only of a minor nature and hence this author’s account is adequate for the anatomy of our forms. THE NORMAL OCCURRENCE OF PARASITES IN SAND SHARKS TAKEN AT RANDOM A knowledge of the normal content of parasites found in the sharks, as collected, was important both for its bearing upon the results of treatment which attempted to rid them of all parasites, and upon the results of any artificial infection with young cestodes. There are very few sand sharks examined which have not some infection with Crossobothrium laciniatum, which is the only ces- tode parasite known to infect the digestive tract of this hostin considerable numbers. Some records by Linton (’99, p. 429) are here tabulated as quite representative of any dozen specimens taken at random. Table from Linton’s records bare | oe gers | oar alge 7) GOO aber RE et ges. | 1 20 Dilys 2 Sst SOOM ey Aaa ee eee ry ties cn Pee | 1 several PANT OUIS GN HO el OOO Mepawerts ani rama taicas aeeier | 1 numerous ANSI S G2 SO OMe «ate lta hie o 5 aegers lel: i 2 TSE LSS SOO) ean ere tami Meech wh: 4 op hs 1 i! ANI SUS tdi SS Oprrr cae ann eck. Che shag sho th: i 4 Aupast WS 1S Oot seem erica! aise sie ats a al 55 ANI CUISTOLO MS 90 ee pe ee eer cent e e 1 12 SFU ys UD OO Scare ee te eee ey aha | 1 47 uly 20; LODO os. Seppe eter ae, nee 1 16 August, L261 G00 Re er es en iin let etree | il numerous August 15 LOO cs oct m Ane SARS, | 1 106 From my own records, the results are similar, though the counts are usually higher because a careful search was being made for the small young specimens. The following table is from six sharks examined in 1904. 828 WINTERTON C. CURTIS | C. LACINIATUM DATE SHARKS ee oa | Adult | Young | TUly QO ee ee 1 34 | 50 July SO keane 1 20 | 10 AUgUStHl iy eeio uae weer 1 30 6 (AUGUST som sen cea cack 1 30 50 NOUS eon ea) le eas 1 oo | 25 AUpUSE yee ee eee a 1 50 | 30 During the years 1899-1910 this parasite has been used for study by the students in one of the courses given at the Marine Biological Laboratory at Woods Hole and we have always been able to obtain an abundant supply when several sharks were available. Sometimes the first shark opened has yielded all the material needed and it has never been necessary to examine more than four or five. Most of the actual counts recorded in my notes in 1903-04 were made upon sharks in which search was being made for specimens showing the early stages of proglottid forma- tion and for this reason the sharks recorded are perhaps those which seemed, when first opened, to have an abundance of the parasites. Linton’s records as given in the first table are there- fore more fairly representative. As a further example of their abundance, my notes record the examination on August 11, 1904, of ten sand sharks, taken in the traps on that date. Every one of the ten was infected and in only two cases was the number of the parasites noticeably small. Count was not made because it was evident that the amount of infection averaged substantially the same as that shown by Lin- ton’s record. From these data it is evident that one rarely finds a sand shark which has not some infection; and from the fact that the worms may be found in all stages of development, from the specimens just beginning to form segments to the large adults which are shedding motile proglottids, one may conclude that the source of the infection has been in contact with the sharks within a quite recent period, if, indeed, it is not acting upon them throughout the summer. THE SCOLEX POLYMORPHUS &29 THE INFECTION OF SAND SHARKS WITH THE SCOLEX POLYMORPHUS The first attempt at infection of the sand sharks with the Scolex polymorphus was made with fish which had been held without food for a period of three weeks, a treatment which Monticelli (83)found effective in ridding aspecies of Torpedo of its Callioboth- rium. In all, eleven fish were so treated and each was then given all the larvae obtainable from twelve squeteague, a dose which was estimated at not less than five hundred larvae for each shark. Each fish was fed at the time of the infection and in the three weeks after infection, during which they remained alive, each was fed four times. For food, the flesh of the squeteague was used, an amount about equal to one-third, or one-half the bulk of a good sized fish being used at a feeding; my guide in this matter being the size of the pieces of food commonly found in the stomachs of recently captured sharks, which had fed under natural con- ditions. Judging from the rate of digestion, as observed on sey- eral occasions, this amount of food was unnecessarily large. Such an amount once a week would be ample for sharks in cap- tivity. Moreover, the choice of food was not good; for one may be introducing almost any kind of a cestode larva by feeding the flesh of a teleost fish. In using for infection sharks which had not received any treatment, other than the three weeks’ starving above mentioned, I was, of course, aware that one might expect each one of them to contain the normal infection of C. laciniatum. It seemed, however, that if the S. polymorphus from the sque- teague did represent the larval form of Crossobothrium, it would develop readily in its normal host, and that the introduction of a very large amount of infection would perhaps give the fish thus treated so many young worms all the same size, as to show that they could only have come from the larvae introduced by the experimental infection. Three weeks after the infection these eleven sharks were killed and their digestive tracts carefully examined. Each contained adult specimens of C. laciniatum in numbers sufficient to indicate that all the sharks had their normal complement of parasites and 830 WINTERTON C. CURTIS therefore that the three weeks without food had produced no effect. In eight of the fish there were a considerable number of young specimens of C. laciniatum in all stages of proglottid formation, but as similar stages are commonly found in all sharks (Curtis, 03 and ’06), their occurrence here was no evidence that they had come from the introduced Scolex polymorphus, and the diversity of their stages made any such interpretation out of the question. In these eight specimens there were found in addition to the young and adult C. laciniatum, an unusual number of individuals of the species Crossobothrium angustum, which Lin- ton (99) p. 426), records as a frequent parasite of the dusky shark, Carcharinus obscurus, and the blue shark, C. milberti. Although present in greater numbers than I have found in any other lot of sand sharks, these C. angustum were of all stages from young to adult and there seemed, therefore, no evidence which would connect them with the larvae which had been introduced by my infection. In the whole number of sharks I found upwards of fifty young of another cestode, ali in about the same stage, and with well developed scolices and the segmentation into proglot- tids just beginning. Because of their conspicuous and charac- teristic bothria, these were at once recognizable as the young of the species Phoreibothrium triloculatum, a form which Linton has described from the dusky shark and which is represented in figs. 5, 6, 9, 10 and 11 of this paper. Unfortunately, my records give only the fact that each of these sharks contained some Phoreiobothria and fail to give their distribution in the individual sharks. The occurrence of this species in the sharks of this experiment would indicate, when taken alone, hardly more than that P. triloculatum is sometimes found in the sand shark, even though the only sand sharks in which I have found it are the ones pre- viously infected with the Scolex polymorphus. The fact that the individuals were all of about the same early stage is more import- ant, though not much stress can be laid upon it because of the failure of my records to state the distribution of these worms in the individual fish. I regard the results of this attempt at infec- tion, which was the only one I was ablesto carry through in the THE SCOLEX POLYMORPHUS §31 first summer of my work, as valuable only because they support the more satisfactory results obtained in the work of the follow- ing summer. A further discussion is, therefore, deferred until the results of the later work have been presented. Having been unsuccessful in the attempt to reduce the number of parasites by the process of starving, for as long a time as was available, if the sharks were to be used for any subsequent exper- iments, attention was directed at the beginning of my second summer’s work to the discovery of an effective method by which the sharks could be entirely freed of their cestode parasites. For this, I used the oil of male fern, one of the most powerful vermi- fuges used in human and veterinary practice. Following this practice, the dose of the oil was followed after an interval of from twenty-four to forty-eight hours by one of calomel. The manner of introducing these drugs into the stomach of a shark is explained in the earlier section of this paper where the general methods of work are given. In the tables upon the pages which follow will be found the detailed results obtained with the several lots of sharks treated in this manner. The necessary general explanations will be given in the discussion of the first table, while in the discussion of the subsequent tables I shall give only the facts which the tables show. TABLE 1 5 sharks, captured before July 2, kept without food until July 22 July 22, all sharks given 2 ec. each of oil of male fern July 23 the four surviving sharks given 1 ce. of calomel | ~ NO. GIVEN SHARK | CONDITION OF CESTODES IN = Dare IN THIS SERIES | SHARK AT DEATH | SPIRAL VALVE ETDS = : H wes July: 23eee eee eon Nome | Dying | 6 C. lacinia- | | | tum found | | | | dead | Julie 2Ots eee | Nos. 2, 3, 4 Killed | No signs of Mucous mem- ~ ? 3 d and 5 | | eestodes -— normal Table 1 shows the results obtained with five sharks which were starved from three to four weeks after capture and then given 832 WINTERTON C. CURTIS 2 ce. of oil of male fern, followed twenty-four hours later by about 1 ce. (measured dry) of calomel. The dates and other points of importance are shown in the several columns, under appropri- ate headings. For convenience of reference, each shark is given a number. The column which shows the ‘condition at death’ is inserted because specimens did not always survive the treatment and thus some of those examined had died from its effects. Such specimens are marked ‘dead’ while others which were killed because they seemed about to die are marked ‘dying.’ The specimens which are marked ‘killed’ are those which appeared to be in perfect condition’ when they were taken and killed for examination. The specimens found soon after death (‘dead’) and those which were killed when they seemed likely to die (‘dying’), presented data of some value, for the reason that under normal conditions the death of the shark is not followed at once by the death and consequent disintegration of the parasites. One finds that living cestodes can be obtained from untreated sharks, which have been dead in the water, or lying exposed to the air upon the wharf for five or six hours and I have often seen, in my work with the students at the Marine Biological Labora- tory, spiral valves left exposed to the air all day yielding an abundance of C. laciniatum which were still alive and seemingly about as active as if taken from a shark just killed, and these worms if put into sea water may live for as long a period as forty- eight hours. In no case of an untreated shark which, on being injured by rough handling in my cars or by the collectors when first captured, was killed before it died from the effects of this handling did I find the approach of death in the host killing the parasites. We may therefore conclude that when, in a shark which has died very recently, or in one which has been killed because death seems approaching, there are found dead Cestodes, the worms have been killed by the drugs and not by the actual, or approaching, death of the host. Such cases have for this reason a sufficient value to be considered in the series. The objections against them are, first, that they do not represent indi- viduals taken at random, and second, that it is of no value to show that the worms are killed in the sharks which do not sur- THE SCOLEX POLYMORPHUS $33 vive the treatment while other worms are not killed in the sharks which do survive. It should be remembered, however, that such non-survivors do not represent the weakest individuals in any lot, for the strongest were probably the ones which fought hardest when put in the holder and such very active sharks received rougher handling and perhaps they sometimes died from this cause. With these reservations, the specimens found soon after death and those killed when about to die, may be cited as showing the immediate effect of the drug upon the parasites. An important point, noted in some of the tables, is that about twenty-four hours after the dose of oil, that is, when the calomel was given, numbers of dead Crossobothria were squeezed from the anus as a shark was placed in the holder. Many entire worms of all sizes were thus obtained and these when placed in sea water slowly disintegrated, showing no signs of life, as they might have done if only stupefied by the oil. Of the five sharks here recorded, No. 1 was not in good condi- tion, though death did not seem near at hand, when it was killed the day after the oil was administered. It contained only dead C. laciniatum. Three days after the calomel was given the four remaining specimens were killed and examined, with the results which are shown in the table. They were all in good condition and the mucous membrane seemed quite normal. I may here state that my examinations of the spiral valves throughout this work have been most careful. In each case the outer wall was split longitudinally along one side and the cut continued down across the spiral folds to the opposite side. A similar cut was made across each fold along the middle of each half of the valve and the inner surface thus exposed as four parallel rows of trian- gular flaps, which were then examined one at a time under a lens. Where there was any such amount of chyme as to obscure the surface of the mucous membrane the valve was washed until clean and the washings examined in shallow dishes against a dark background. Table 2 shows the results in five sharks, which were given a heavier dose of the oil of male fern, and the calomel after an interval of forty-eight hours. After the dose of oil, shark No. 1 §34 WINTERTON C. CURTIS TABLE 2 5 sharks, captured before July 1, kept without food 3 to 4 weeks July 18, all sharks given 4 cc. each of oil of male fern July 20, the three surviving sharks given 1 cc. each of calomel DATE “IN THis SERIES | SHARE AT DEATH | GrrRAL VALVE REMARKS duilyel Oe eer No. 1 Dying 1 strobilla Traces of oil dead in stomach TUN AD), oth oa bee No. 2 Dead No signs of Digestive tract cestodes considerably decomposed July 21, .......00-| Nos.s,4and 5 Killed No signs of Mucous mem- cestodes brane in good condition TABLE 2a 1 shark, captured about June 15, and kept without food July 1, given several ec. of oil of male fern diluted with ether. | « 7 | oe NO. GIVEN SHARK | CONDITION OF | CESTODES IN EMARKS IN THIS SERIES SHARK AT DEATH | SPIRAL VALVE | a DATE | ee PT | sell oe hs | Uy FOR cee Killed | No signs of cestodes l seemed to be dying and was killed with the results indicated. No. 2 was found dead, but was a good deal decomposed and so the absence of worms may not mean much. The three surviv- ing specimens, which were killed twenty-four hours after the calomel, were without any cestodes. The results of these two experiments may be criticised on the ground that the sharks were not left alive long enough to show that more would not have died from the treatment, but my expe- rience has shown that when the animals survived this treatment for two or three days the subsequent mortality was not likely to be greater than among any other sharks kept in confinement. Some of the sharks noted in other tables were kept alive a much longer time. THE SCOLEX POLYMORPHUS 835 Before I began using the gelatin capsules a single shark which had been in captivity a few days was given some oil of male fern diluted with ether. How much actually got into the stom- ach I do not know, as some was spilled in pouring the mixture down the tube which was thrust into the oesophagus. No calo- mel was given. Five days later when the specimen was examined there were small traces of the oil still in the intestine and no worms were found. TABLE 3 20 sharks, captured July 4 to 7, kept without food ; Se - DATE A oFeeeant | GGueE Aemauch | coman yore REMARKS | (Oke July 25, all sharks given 2 ec. each of oil of male fern erent e hee * i = onl ol RO JulyaZ Ose ee No. 1 Dead 9 scolices and | Decomposition bits of strobi-| not noticeable le, all dead | in intestine duly 265.5553. ¢ No. 2 Dying 2 dead Size of worms | not recorded Julya262 . eee No. 3 Dying /20 dead and © Size of worms | in the faeces | not recorded July s268 sen eee | No. 4 Dead _4 dead and in| Decomposition faeces of intestine | not noticeable July 26, the sixteen surviving sharks given 1 ec. each of calomel July ihe 2 No. 5 Dead No worms Decomposition just begun Julya2 See eee eee No. 6 Dead No worms JULY 283959 aoe No. 7 Dying No worms Mucous mem- brane normal dpily:3 seer ee No. 8 Killed | No worms Mucous mem- brane normal August 2, the twelve surviving sharks each fed on shark’s flesh — { | 7 = : > | AURUSU Tiere Aces No. 9 | Dead No data | Considerably | decomposed } JOURNAL OF MORPHOLOGY, VOL. 22, No. 3 836 WINTERTON C. CURTIS TABLE 3—(Continued) August 15, the eleven surviving sharks each fed on shark’s flesh | DATE Sees Ses Pee cesae | eee) | ees August 28....... No. 10 Killed No worms AMPUStiese fas No. 11 Killed No worms | August 28....... No. 12 Killed 1 Crossobothrium laciniatum AUpUSt 2oo.5% 05: No. 18 Killed 1 Crossobothrium laciniatum AUSUSt eS sen | No. 14 Killed 2 Crossobothrium laciniatum | young 2 Crossobothrium laciniatum . | scolices 1 Crossobothrium angustum The six remaining sharks survived the treatment to this date and later, and were used for infection experiments. When examined they gave results as follows: Acoust Liane ser No. 15 Dead No worms | Badly decom- : | | posed August 22....... No. 16 Dead No data | Badly decom- | posed Ameust2ousen ec No. 17 | Killed No worms | Normal August 31....... No. 18 Killed 1 Crossobothrium laciniatum | medium size NURI Bly on ono - | No. 19 | Kalled 8 C. laciniatum large, scolices | | 5 C. laciniatum, small August 3l..2 =: I) No..20 | Killed 7 C. laciniatum, medium size 10 C. laciniatum, small Table 3 shows less satisfactory results than the foregoing, the net results being as follows: without parasites 11, found dead and so decomposed that no data were obtained 3; still infected 6. If the survivors alone are considered the results are not nearly so good, for out of ten surviving specimens (Nos. 8, 10, 11, 12, 13, 14, 17, 18, 19, and 20) we have only four (Nos. 8, 10, 11, and 17) which are entirely free of cestodes, while there are six (Nos. 12, 13, 14, 18, 19, and 20) which are still infected. Of these six, Nos. 12, 18 and 18 have but a single parasite, so the chances are that some parasites have been eliminated, but Nos. 14, 19 and 20 THE SCOLEX POLYMORPHUS 837 show so many that one could not fairly claim any reduction as a result of the treatment. When, however, those specimens which were killed when they seemed in bad condition are taken into account, it is evident that many worms must have been elimin- ated, and the results indicate the elimination of many of the parasites from this lot of sharks, probably of all of them in those sharks which died from the treatment, but they also indicate that the dose as here given cannot be relied upon. TABLE 4 12 sharks, captured before July 2, kept without food | vars AR an i gumonne | euanes July 28, each shark given 2 cc. oil of male fern July 29, each shark given 3 cc. calomel Augusts0)...... No. 1 Dying No worms Traces of oil and calomel in gut Aust: Wore No. 2 Killed No worms Gut normal ATICUStr ones No. 3 Killed No worms Gut normal Aust: 4-22 -ee No. 4 Killed No worms Gut normal Aumust 16oe ace " No. 5 Dead No data Decomposed Aurusth dd! . 2: No. 6 Killed 1 C. laciniatum, scolex AIOIStel en ee No. 7 Killed 3 C. angustum, small 3 C. laciniatum, small August 18, the five surviving sharks each fed on shark’s flesh 7 eee Be ee At : August 28....... | No. 8 | Dead No data | Decomposed | | August 29....... No. 9 Killed 1 C. laciniatum, small August 30....... No. 10 Killed 2 C. laciniatum small found at very anterior end of | spiral valve August 30....... J "Nott | Killed No worms AUpUStisOn se No. 12 | Killed 2 C. laciniatum, large and with ripe proglottids 1 C. laciniatum, small 838 WINTERTON C. CURTIS Table 4 shows five cases of entirely successful expurgation (Nos.1,2,3,4 and 11). Specimens Nos.9 and 10 had respectively one and two worms, while No. 12, since it has two large worms with ripe proglottids, does not justify the conclusion that the number of the parasites was even reduced, and No. 7 must be regarded in the same way. TABLE 5 12 sharks, captured July 25 to August 6, kept without food | | NO. GIVEN SHARK | CONDITION OF pete IN THIS SERIES SHARK AT DEATH | REMARKS | CESTODES IN | SPIRAL VALVE | August 17, all sharks given 2 cc. each of oil of male fern l AUgISt 135-4, seee | No. 1 Dying | No worms ~ | Dead worms | from anus in | handling August 18)... ). | No. 2 Dying |» Noworms | AUISUISt sl San oneee | No. 3 Dead No worms Dead worms | from anus in | | handling August 18, the nine survivors given 3 cc. each of calomel AUS USGOU see ee No. 4 Dying | No worms Gut normal August Sl... .. 2: No. 5 Kalled No worms Gut normal AUPUSt Ol. se. | No. 6 Killed No worms ' Gut normal August 31....... |) eNoso 7 Killed No worms | Gut normal AU CUStO lee | No. 8 Killed No worms | Gut normal August) shee se No. 9 Killed | No worms | Gut normal AMICUS hi eee ee No. 10 Active No data _ Escaped August 31....... No. 11 Killed 8 C. laciniatum, medium size Australes er No. 12 Killed | 4C. laciniatum, medium size Table 5 has twelve sharks, less one which escaped during the handling. Of these eleven individuals, Nos. 5, 6, 7, 8 and 9 were without any infection, Nos. 1, 2, 3 and 4 did notsurvive, but showed that the drug had killed the parasites. Nos. 11 and 12 have respectively eight and four specimens of C. laciniatum and hence must be counted as against the effectiveness of the treatment. THE SCOLEX POLYMORPHUS 839 It so happened that no shark, in which the Crossobothria sur- vived, was killed among the first in any lot; as may be seen by reference to the dates on tables 1, 2, 3, 4 and 5. Thus by the 11th of August my records showed thirteen sharks which were killed when in perfectly good condition, six others which were killed, when it seemed that they were likely to die, and seven others, which died from the treatment and in not one of these had I found a single living Crossobothrium. This seemed to demon- strate the effectiveness of the single treatment with the oil of male fern, which was therefore continued, the sharks which sur- vived it being kept for infection with the 8S. polymorphus. Later, when sharks began to appear in which some of the worms had survived, the season was so far advanced that there was neither TABLE 6 9 sharks, captured August 6 to 9, kept without food NO. GIVEN SHARE IN THIS SERIES CONDITION OF SHARK AT DEATH CESTODES IN E oI BL SPIRAL VALVE | “OREN ESS) August 18, all sharks given 2 cc. each of oil of male fern August 19, all sharks given 3 cc. of calomel August 19, dead Crossobothria from anus of one specimen in handling August 30... 7 | No. ill Dead No worms | Decomposed | a little August 30....... | No. 2 | Alive No data | Escaped August 30, ave the seyen survivors 3 cc. each of oil of male fern es i Te eae Be ES iceable d INO Worms No noticea September 1..... No. 3 Dea decomposi- | | tion : | iceable d No worms No noticeal September 1....-| Neem e decomposi- tion 1 Killed No worms Gut norma September 2...-. No. : aa ed No worms Gut normal September 2...-- No. : Killed No worms Gut normal Ssptember 2...-- No. : Killed No worms Gut normal September 2...-- No. ate Sirens Gut normal September 2...-. No. 9 840 WINTERTON C. CURTIS TABLE 7 Showing the results of infection experiments NO. | Sete eee Se KtanNe FED INFECTED FED REMARKS CAPTURE SERIES WITH O.M.F. July 6 .\No. 1] July 25 August 9 | August 10 ily Geers No. 2| August 1-3) August 4 | August 8 | August 10 July 6.....|No. 3| August 1-3] August 4 August 8 | August 10 Juliveleecer No. 4| August 1-3} August 10 | August 11 July 15.....No. 5| August 1-3} August 10 | August 10 Fed at time dil @es so - No. 6 | August 1-3 August 9 | August 10 | of infection JiUbY Weso0c No. 7 | August 1-3 August 9 | August 10 July 1.....{No. 8 | August 1-3} August 10 | August 11 Aiwlhy Wcoee No. 9 | August 1-3| August 10 | August 11 TABLE 7—Continued Showing the results as in tables 1 to 6 NO. GIVEN DATE Deere nee Soak CESTODES IN SPIRAL VALVE. REMARKS ABOVE DEATH August 12.... No. 1 Dead Many S. polymorphus found attached to the wall of the stomach and to the bit of the squeteague’s stomach in which they were placed when intro- duced. The gut was slightly decom- posed, but the larvae were still alive. August 16....| No.2 Dead No data. Badly decomposed August 31....| No.3 Killed Many small Phoreiobothrium trilocu- latum. See text p. 846 September 1..| No. 4 Killed 3 C. angusvum, small 3 C. laciniatum, small 7 Phoreiobothrium triloculatum. September 1..| No. 5 Killed 3 C. laciniatum, small. September 1..| No. 6 Killed 8 C. lacininatum, large with the long neck region 5 C. laciniatum, small September1..| No.7 Killed 10 C. laciniatum, small. 7 C. laciniatum, medium September 2..| No.8 Killed 3 C. angustum, small Many Phoreiobothrium triloculatum September 2.. No. 9 Killed 1 CG. laciniatum, scolex only. Many Phoreiobothrium triloculatum THE SCOLEX POLYMORPHUS 841 the time nor the material for trying a modification of the treat- ment upon any considerable number of sharks. Only nine sharks were differently experimented upon and the results obtained are represented in table 6, the net results of which are; without infection, 5; found dead but in sufficiently good condition to give reliable evidence that the worms had been killed by the drug, 3; escaped, 1. The results of all the expurgation experiments may be summar- ized in the table which follows: TABLE 8 Showing net results from all experiments in treatment with oil of male fern | SHARKS WHICH DIED UNDER THE TREATMENT, SURVIVING SHARKS | BUT IN WHICH THE WORMS WERE ABSENT OR HAD BEEN KILLED BY THE OIL OF MALE FERN TABLE REFERRED ol eae uf Cee eee | No infection | Still infected | mae sess er ‘Found dead Total | | 1 4 if 5 2 3 1 1 5 3 4 6 3 4 17 4 5 4 1 10 5 5 Pe 4 11 6 5 3 8 | f ; Peis 26 2 13 | 5 | 56 In this table, a few specimens which were found dead and so decomposed as not to yield reliable data are omitted. The speci- mens which survived treatment are of course most important and stand in the proportion of 26 for and 12 against the success of the attempted expurgation. Of the non-surviving sharks we - have 13 specimens taken while still alive and all showing that worms were killed, as a result of the treatment and some time prior to the death of the shark. When one compares the number of worms found in these treated sharks with the tables and records showing their prevalence in sharks which have had no such treat- ment, it will be quite evident that a large proportion of the parasites must have been eliminated even by the single dose. 842 WINTERTON C. CURTIS In those sharks which were given the double dose of o. m. f. (table 6) not a single cestode was discovered, but as only eight specimens were thus treated the number of trials is insufficient to show that even this treatment may be regarded as always effective. It is perhaps too much to expect a treatment that will be entirely effective in every instance, nor is such a treatment necessary, for a stray worm or two in one shark out of a dozen would still give us a result good enough for practical working purposes. Unsatisfactory though my results are, they do, I think, justify the belief that a little more experimenting along the line of repeating the dose one or more times would develop a method of treatment sufficiently effective for working purposes, i.e., will eliminate all the parasites, except in rare instances, without killing too many of the sharks. If such a method can be found, the sharks thus expurgated could be used in a variety of experiments. For example, by introducing young Crosso- bothria into a shark one might expect to find out more than we now know about the rate of growth and the maturing of pro- glottids in the cestoda, and another point worth examin- ing would be the truth of the current idea that it is the inability of the cestode to survive in the wrong host which limits the habitat of a given species of parasite to a single host, or to a few closely related hosts. This latter point might, indeed, be in- vestigated by the infection of sand sharks with larval forms other than the ones known to occur in that host, but where the normal infection is so great the examination of the specimens would be much easier if the host was first freed of all infection. An insufficient supply of squeteague during the latter part of the second season made it impossible for me to make more than a few infections after the method of treatment, as above described, had been established. Only nine sharks were so infected and none of these was killed until about three weeks after the infec- tion had been introduced. Hence the material is wanting to show my transitional stages between the 8. polymorphus and the young specimens of Phoreiobothrium (figs. 9, 10 and 11) which I believe to have come from this larva. This lack of early stages resulted not because I chose, having only a limited number of THE SCOLEX POLYMORPHUS 843 infections, to let them all run as long as possible, but because the sharks first infected were kept to run longest; while later infec- tions were planned for the earlier stages, an arrangement which would give the most extensive time results in a limited time. When, however, the time came for infecting the sharks from which to secure the earlier stages, the squeteague could not be obtained in numbers sufficient to yield the larvae for infection purposes. Of the nine specimens, two died in the cars and from one of these two no data were obtained, so there are only seven speci- mens from which entirely reliable conclusions can be drawn. The sharks used were all starved for three or four weeks after their capture and then given 2 ce. each of the oil, followed twenty- four hours later by about 0.5ce. of calomel. They were all specimens which had survived this treatment and the history of each individual previous to the time of the infection is given in the first half of table 7. Since there is no evidence connecting the genus Crossobothrium with the 8. polymorphus, this table may also be used to furnish data on the success of the attempt at expurgation. The first part of the table gives the dates of cap- ture and of the treatment with the drugs, between which events the sharks were kept in the cars without food. The dates at which they were fed are also shown for comparison with the dates of infection. Each shark is numbered, as in the previous tables, and these numbers are repeated in the second half of this table where the results of the examinations for parasites are tabulated in a manner similar to that followed in the tables for treatment with the oil. In introducing the S. polymorphus into these sharks I took the portion of the cystic duct containing the larvae from twelve squeteague, and placed these in the little sac obtained by cutting off the end of a squeteague’s stomach. This sac was turned wrong side out to avoid any possible injury from direct contact with the mucous membrane of the squeteague. The infection thus prepared was introduced into the shark’s stomach by pushing it through a piece of iron pipe having a diameter sufficient to admit such an object without undue pressure. For food, I used 844. WINTERTON C. CURTIS the flesh of another sand shark in preference to that of a teleost, since there is the minimum likelihood of finding any cestode larvae in the flesh of a large shark; whereas a teleost may, in addition to its normal parasites, contain almost any thing in the way of a ‘xenositic’ larva. All the sharks were fed shortly before or after their infection, as is shown in the first half of the table. The latter part of table 7 shows the results when -these same sharks were examined for parasites, and by reference to the num- ber given each specimen, one may follow any one fish through the two parts of the table. The only shark examined soon after the infection is No. 1, which was found dead. In this shark, specimens of the §. polymorphus were found adhering to the surface of the shark’s stomach and to the remains of the bit of squeteague’s stomach in which they had been introduced. Al- though this specimen was not found until decomposition was quite in evidence, these larvae still showed some slight movements and had therefore survived in the stomach for a period of three days. Shark No. 2 died during my absence from Woods Hole and was so badly decomposed when found that no data were obtained. In the spiral valve of specimen No. 3 there were a very large number of young Phoreiobothrium triloculatum (figs. 9,10 and 11). I collected and preserved some thirty-five of these worms, but this number represents only a small part of those present. Owing to their minuteness when only the deli- cate posterior end can be seen protruding from between the villi of the intestine and the tenacity with which their powerful hooks enable them to retain their hold, these larvae are often very difficult to detach from the walls of the spiral valve, though they may be large enough to be easily recognized. There must have been present in the shark many more than I collected; for taking into account the ones actually seen but not collected, I estimated at the time that there were a good many more than one hundred of these young worms in this single fish. A fact of per- haps more importance than their numbers is that in any one shark they were all in the same stage of development. Shark No. 4 shows three specimens of C. laciniatum which are to be regarded as worms which survived the expurgation treat- THE SCOLEX POLYMORPHUS $45 ment. Seven small specimens of P. triloculatum were collected, but in this case it was evident that the valve contained a much smaller number of these than the previous one. My notes taken at the time state that no more were seen in addition to those actually collected, although there may have been some toward the anterior end where the villi are longest. Specimens Nos. 5, 6 and 7 all present evidence against the effectiveness of the ver- mifuge. In 6 and 7 the numbers of surviving cestodes is so great that no certain effect from the drug is indicated. There is perhaps some significance in the fact that here, as in some other cases where the worms are found surviving the effects of the oil, the proportion of young to adult specimens is somewhat increased. This may indicate that the drug is more effective with the large worms which have long bodies extending among the folds of the valve than with very young ones which may often be almost buried among the villi and so escape the full effects of the drug. In this table there are recorded from specimens 4, 5, 6, 7, 8 and 9 a total of 48 surviving Crossobothria. Of these only 9 are beyond the period of segmentation into proglottids and from what we know of the proportions in which the young and old are commonly found it would seem that the drug has destroyed more adult than young worms. This con- clusion is of course based upon my belief that these young Crosso- bothria have not come from the introduced 8S. polymorphus. In sharks Nos. 8 and 9 the same condition was found as in No. 3, namely, so many specimens of small P. triloculatum that the counting them was an impossible task. Later when I examined very carefully the preserved specimens I[ found that the Phoreio- bothria from any one shark were of uniform size, but that when those from different sharks were compared there was some differ- ence in the size attained. For instance those from specimen 9 are almost twice the size of the ones from specimen 3. This differ- ence in size is noticeable only in the body portion of the larvae, the scolices being of very uniform size. The presence of the C. laciniatum and a few C. angustum, although there are more young specimens among them than one would expect to find in sharks taken at random, does not seem 846 WINTERTON C. CURTIS to indicate anything except the ineffectiveness of the attempt at expurgation. There can be no interpretation of the facts which would show that the 8. polymorphus had given rise to these Crossobothria. Despite the lack of intervening stages between the S. poly- morphus and the specimens of P. triloculatum, which I found in these sharks, I think there is some pretty good evidence that the latter have developed from the introduced 8. polymor- phus. First, the examination by myself, and also by Linton, of a large number of sand sharks at various times has never shown that young or adult specimens of Phoreiobothrium triloculatum are to be found as regular parasites of this shark, or as frequent ‘xenosites.’ I have never found this form in any sharks except those which I infected with the 8. polymorphus. It is very prob- able, however, that one might at any time find stray specimens of this worm, since the squeteague is a not uncommon food of this shark. The failure to find it in any of the sharks I have examined would indicate that it does not often survive, when introduced in nature along with the squeteague, and in my experiment only a small proportion of the larvae introduced have survived. Second, the specimens of P. triloculatum which I found in any one shark were of very uniform size, and unless we suppose that there is some limit to their growth, when in an abnormal environment (the wrong host), this would indicate that they all entered the shark at about the same time. In the third place this conclusion is also in line with the results of Monticelli (’88) who has shown that the S. polymorphus of European waters develops into the genus Calliobothrium, which belongs to the same family as Phore- iobothrium, a fact which is further discussed in the portion of this paper which deals with the nature of the 8. polymorphus. Attention should here be directed to one point which has per- haps suggested itself from the examination of the figs. 7-13. This is that the specimens of the young P. triloculatum (figs. 9 and 10) are considerably smaller than some of the specimens of the S. polymorphus, from which I suppose them to have devel- oped. This appeared to me at first a most serious objection, though upon further consideration it does not seem an insur- THE SCOLEX POLYMORPHUS 847 mountable one. The ragged and irregularly constricted ends of some of the specimens (figs. 7 and 8) suggest that part of the body is being lost, as in the case in the development of the strobila in those forms which have a typical bladder-worm stage. Again, while the body region is smaller than that of the larger specimens of the S. polymorphus (fig. 12) the scolex of the young P. triloculatum (figs. 9, 10 and 11) is considerably larger than that of the S. polymorphus. I should also add that fig. 12 represents one of the very largest of the larvae and has been killed under pressure to flatten the body and make it more suitable for a whole mount whereas the specimens of P. trilocula- tum were killed without pressure and the body has remained cylindrical. The 8. polymorphus reaches this size (fig. 12) only in the cystic duct and the smaller specimens present a lesser disparity in size. I am inclined to think that in the case of such large specimens portions of the S. polymorphus body may be moulted off just as in the case of the bladder portion in the cys- ticercus. One constant feature of the larger specimens of the S. polymorphus has, perhaps, some significance in this connec- tion. It is the occurrence of a denser region which terminates abruptly a little way behind the scolex (fig. 12). In a specimen stained and mounted whole after the carbonate of lime gran- ules, which occupy so much of the parenchyma in the body region, have been dissolved out, one can distinguish this region as having the same denser appearance as the tissue in the body of the young Phoreiobothria (figs. 10 and 12) or the region of proglottid formation in young specimens of C. laciniatum. It is possible that the scolex and this region of the S. polymorphus are the most important in the formation of the adult worm and that part or the whole of the body region may be lost. It seems probable that larvae like the 8. polymorphus have been derived from larvae of the cysticercus type and, if this be the case, it would not be surprising to have a part of the body region lost at this point in the development. 848 WINTERTON C. CURTIS SUMMARY Experimental infections of the sand shark (Carcharias littor- alis) with the cestode larva known as the Scolex polymorphus indicate that the larvae used for these experiments developed into the species Phoreiobothrium triloculatum. It seems clear that the common tapeworm (Crossobothrium laciniatum) of this shark cannot come from the Scolex polymorphus, even though this larval type may represent the young of a number of cestodes, a possibility which is referred to in the third section of this paper. Starving the sharks had no effect upon the cestodes, but by means of treatment with the oil of male fern followed by calomel. the great majority of the parasites wese eliminated before the sharks were artificially infected. It seems probable that by a little more experimentation a method of treatment could be secured, which, for eliminating these parasites, would be suffi- ciently effective for all working purposes. THE SCOLEX POLYMORPHUS 849 LITERATURE CITED BrRoNN, THImRREICH 1894-1900 Cestodes. Curtis, W. C. 1903 Crossobothrium laciniatum and developmental stimuli in the cestoda. Biol. Bull., vol. 5, no. 2, July, 1903. 1906 The formation of proglottids in Crossobothrium laciniatum. Biol. Bull., vol. 11, no. 4, Sept., 1906. Linton, Epwin 1886 Notes on the entozoa of marine fishes of New England, with descriptions of several new species. Rept. Commissioner of Fish and Fisheries for 1886. Washington. Published in 1889. 1887 Notes on the entozoa of marine fishes of New England. Part II. Ibid. for 1887. Washington. Published in 1890. 1897a Notes on larval parasites of fishes. Proc. U. S. Nati. Museum, vol. 19, Washington. 1897b Notes on cestode parasites of fishes. Ibid., vol. 20, Washington. 1899 Fish parasites collected at Woods Hole in 1898. Bull. U. S. Fish Commission for 1899. Washington. Published in 1900. 1899 Parasites of fishes of the Woods Hole region. Ibid. for 1899. Washington. Published in 1901. 1904 Parasites of fishes of Beaufort, North Carolina. Ibid. for 1904 Washington. Published in 1905. Montice.u, F. 8. 1888 Contributione allo studio della fauna elminthologico del golfo di Napoli. 1. Richerche sullo Scolex polymorphus. Mitth. zool. Stat. Neapel, Bd. 8, p. 85-152. Peck, James I. 1895 The sources of marine food. Bull. U. S. Fish Commis- sion for 1895. Washington. Published in 1896. Rupoupui, C. A. 1808 Entozoorum sive vermium intestinalium historia natur- alis. Vols. 1 and 2. 1808 and 1809. vAN BENEDEN, J. P. 1850 Les vers cestoides. Bull. de l’Academ. roy. de Belg. Tome 17, no. 1. WaGeENER, C. R. 1854 Die Entwicklung der Cestoden nach eigenen Untersuch- ungen. Nov. Act. d. k. Leop-Carol. Akad. d. Naturf. Bd. 24, suppl., Breslau. ZscHoKKE, F. 1886 Le development du Scolex polymorphus. Arch. d. se. phys. et. nat., 3 ser., Tome 16, Genéve. PLATE 1 EXPLANATION OF FIGURES 1 Outline of a living specimen of the Scolex polymorphus. The stippled areas just behind the four bothria show the location of the faint red pigment spots seen in some specimens. Magnified about 45 diameters. 2 Specimens of the Scolex polymorphus drawn from the living specimen to show the characteristic mode of attaching with all four suckers to the bottom of a dish in which they are being examined and the manner in which they fasten to one another. Magnified about 45 diameters. 3 and 4 Outlines of the Scolex polymorphus drawn from living specimens to show other characteristic shapes. The area occupied by the large and small vascular trunks of either side and the terminal vesicle are added to fig. 3 as they appear in a stained specimen. The filiform appearance shown in fig. 4 is often seen as the larvae draw themselves over a surface by the characteristic movements of their bothria. Magnified about 45 diameters. 4a From a specimen stained and mounted whole, showing a portion of the body region on a large scale. The cuticle (cu), the larger and smaller excretory trunks (wt) and the conspicuous longitudinal muscle fibres (mf) are shown. The stippling indicates the distribution of nuclei in the parenchyma as it appears in optical section. Along part of the upper margin are shown the minute projec- tions which occur upon the cuticle in the posterior part of the scolex region and for a short distance along the body, finally giving place to the smooth cuticle as shown in the figure. Magnified about 200 diameters. 5 Scolex of a young Phoreiobothrium triloculatum, taken from a sand shark artificially infected with Scolex polymorphus (see shark no. 3 of table 7). This view shows the division between two neighboring bothria along the mid-line of the figure. The characteristic pairs of hooks are shown attached along the line separating the upper and middle loculi of each bothrium. In this figure the subdivision of the posterior margin of the bothrium into three loculi, from which the specific name is derived, is not seen because the contraction of the lower mar- gin brings their surface at right angles to the general surface of the bothrium. The area on the mid-line of the figure toward the anterior end of the scolex, and marked by the closer stippling, appears in some specimens and may represent the ‘myzo- rhynchus’ of the Scolex polymorphus. Theminute spikes protruding from the neck region of the cuticle are shown in profile only. Magnified about 90 diameters. 6 The same as the last figure. One of the bothria is shown in front view and the three posterior loculi are expanded so as to show clearly. Magnified about 100 diameters. 850 LIFE HISTORY OF SCOLEX POLYMORPHUS PLATE 1 Ww. C. CURTIS JOURNAL OF MORPHOLOGY, VOL. 22. No. 3 DRAWN BY W. C. CURTIS 851 PLATE 2 f EXPLANATION OF FIGURES 7 and 8 The posterior end of a young specimen of Phoreiobothrium from a shark artificially infected with the Scolex polymorphus. These figures show the formation, either of psuedo-proglottids which are molted, or an irregular beginning of true proglottid formation like that of Crossobothrium laciniatum (Curtis, 06). The figure shows so much irregularity that the process does not seem lke true proglottid formation and the posterior end shows a peculiar outline which probably indicates that part of the specimen has been lost. Only the larger pair of the two water trunks is shown. Magnified about 70 diameters. 9 One of the largest specimens of Phoreiobothrium triloculatum which was secured from the artificially infected sharks (see shark no. 8 of table 7). This is one of the few specimens which showed a segmentation of the posterior end sufficiently regular for comparison with the formation of the posterior proglot- tids in Crossobothrium laciniatum. The scolex of the specimen is shown in out- line only. The area of the minute spikes is shown and back of this the two pairs of water trunks which in such a view are superposed throughout the length of the body, are here shown in successive regions. Magnified about 70 diameters. 10 and 11 Two of the smallest of the specimens of Phoreiobothrium trilocu- latum obtained from the artificially infected sharks (see shark no. 3 of table 7). The specimens represented in fig. 10 was ragged at the posterior end as though portions had been detached. Magnified about 70 diameters. 12 and 13° Show in more detail the features of the Scolex polymorphus. In fig. 12 the ‘myzorhynchus’ or proboscis-like structure is shown at the anterior end between the four bothria. The larger and smaller water vascular trunks are shown in successive regions and their convergence toward the posterior end where the vesesls become irregular and branched so that they can only be followed as an area of differentiation in the parenchyma. Each bothrium consists of a denser portion divided into three regions or loculi in the older specimens (fig. 12), and in those slightly smaller often showing only the line of division at the anterior end (fig. 13). In fig. 13 the opening of the terminal vesicle of the water tubes is seen in surface view, the posterior end of the specimen being slightly turned toward the observer. Magnified about 70 diameters. LIFE HISTORY OF SCOLEX POLYMORPHUS PLATE 2 W. CG. CURTIS yy JOURNAL OF MORPHOLOGY, VOL. 22, No. 3 DRAWN BY W. C. CURTIS THE LIMITS OF HEREDITARY CONTROL IN ARMADILLO. QUADRUPLETS: A STUDY OF BLASTOGENIC VARIATION H. H. NEWMAN anv J. THOMAS PATTERSON From the Zoological Laboratory, University of Texas (No. 108) FIVE TEXT FIGURES AND EIGHT PLATES CONTENTS MORO CtI ON: ~, {cha cckie wees Pace Cecio es oc eee ate eee eee 855 Morphology ‘of the imtegumentyedcc: 4c. se oles eee he ote ete cree ere ee 861 Meristic variation of the elements of the nine bands of armor............... 865 A Variation in the banded region asa whole.........:...-.---.--9--5- 865 B Comparative variability of males and females...................... 868 CP Variationeimolnesin Givi elvis ciind Sys ete rere sree rere 870 D_ Fraternal correlation in the banded region as a whole; an index of the limits‘of hereditanyaconmtrolyer sac. cre ats sons aa noe aie: 873 E Fraternal correlation in the individual bands.....................- 880 Atypical variation in the individual elements and in the bands.............. 883 A Scute ‘abnormalities’ in the species; their distribution and frequency. 883 B Hereditary control in connection with scute ‘abnormalities’......... 888 C Band ‘abnormalities’ in the species; their distribution and frequency. 893 D_ Hereditary control in connection with band ‘abnormalities’ .......... 897 Pairing, an intra-fraternal correlation; and its bearing on the problem of mnereditary COMUROl: stpvay se miaelsceheree chee electra sts Cie sine Aeon aie ce ee eee oe eee 905 Maesheredubary7combrolwot/Se xan. entenes acl cte sete cine ora seus eal Siete esters ea es 910 General Considerations eri by Gr ce sileya ech iets eee 911 SSUMIDINT RIYA. Sc rae eg a olen frac reNsconena tote eae oy ial tne ee ee eee lca ae 914 BH HORFAPINY issn ee eee re FE a EE Oe SE aC ee aR ae 917 INTRODUCTION To what extent or within what limits are the definitive characters of the individual determined at the time of fertilization and in how far are the minutiae of organic structure to be considered as the product of individual variability beyond the limits of hereditary control? No more fundamental question could well be raised, and none more difficult of solution. The question has recurred JOURNAL OF MORPHOLOGY, VOL. 22, No. 4 DECEMBER, 1911 855 856 H. H. NEWMAN AND J. THOMAS PATTERSON in various guises, but the underlying inquiry has remained un- changed. In one of its older forms the question is phrased: “What is the relative potency of nature and nurture in develop- ment?” In some of its more modern phases it appears in terms of ‘predetermination versus epigenesis,’ ‘blastogenic versus soma- togenic variation,’ and ‘heredity versus environment.’ In spite of the antiquity of the problem very little direct evi- dence has appeared for its solution. So far as we have been able to ascertain, the only facts that seem to throw any clear light on the situation are those furnished by cases of human ‘identical’ twins. As long ago as 1875 Galton showed his appre- ciation of the value of such data in his paper entitled ‘‘ The his- tory of twins, as a criterion of the relative powers of nature and nurture.’ In a subsequent paper (Galton, 92) he made use of finger-prints as criteria for distinguishing between twins and as a suitable character for testing the degree of likeness and unlike- ness of such pairs. That he had a deep insight into the under- lying fundamental problems involved in the situation is shown by his remark: “It may be mentioned that I have an inquiry in view which has not yet been fairly begun, namely to determine the minutest biological unit that may be hereditarily transmis- sible. The-minutiae in the finger-prints of twins seem suitable objects for this purpose.”’ In 1904, Wilder, interested in the same problem, presented a comprehensive review of the whole subject in his paper on ‘‘ Du- plicate twins and double monsters.’ His results are of great interest and will, we hope, assume a still greater significance in the light of the facts here presented. Wilder discovered a sur- prising degree of resemblance between the palm and between the sole patterns of duplicate twins and was able, he thought, to clas- sify twins as ‘fraternal’ or ‘duplicate’ on the basis of this resem- blance. In addition, the following phenomena were observed in the majority of cases, but were not universal: (1) A bilateral correspondence in the palms and soles of each individual of a set. ‘‘(2) A reversal of the finger patterns in either of the right or the left indices. LIMITS OF HEREDITARY CONTROL 857 ‘‘(3) Differences occur more frequently on the left side.” In his concluding paragraph, here quoted, are given in concise form the chief results and conclusions derived from his studies of duplicate twins: The influence of the germ-plasm and its mechanism (i.e., the direct control exercised by heredity) is exerted upon the friction-skin surfaces only so far as concerns the general configuration, i.e., the main lines, the patterns and other similar features; the individual ridges and their details (minutiae) are apparently under the control of individual me- chanical laws to which they are subjected during growth. Have we then arrived at the limit of the control of the predetermining mechanism beyond which mechanical laws are alone operative; and is it then possible to hold that the modifications in this latter field are the results of individual expe- rience, and that they are similar in the various members of a given species solely because of similar environment? To these and similar questions we can give no answer at present; yet it seems likely that in the general subject of palm and sole markings, not only in man but in other mam- mals as well, we have a set of easily observed and very significant data which may yield important results to future investigators. In addition to his data on palm and sole patterns Wilder fur- nishes us with rather elaborate physical measurements of four sets of duplicates. He realizes, however, that dimensional data are ‘‘far less determinative than are other characters employed, since they are liable to fluctuations through numerous causes, both internal and external, and it could hardly be expected that the similarities here would be very striking.’”’ Yet some most striking physical resemblances have been brought out by different authors. Vernon (’03), for example, gives the data for two pairs of identical twins, one of which, aged twenty-three, showed an average per cent difference of 0.28 per cent; while the other, aged twelve, a somewhat greater difference of 0.71 per cent. Weismann presents the data on one pair of twin brothers, aged seventeen, who showed a per cent difference of 2.2, nearly ten times that of Vernon’s first pair and about three times that of his second. Wilder’s measurements include a much wider range of charac- ters than do the others cited and are therefore probably of some- what greater value. Table 1 presents a brief summary of his data. 858 H. H. NEWMAN AND J. THOMAS PATTERSON TABLE 1 Showing physical statistics of Wilder’s twins ser Sue Gos es oe AGE IN YEARS Te eee de en eae 27 2.39 21.10 THA kira e kee eden 27 2.35 lyf lil TUIEVSY,. Set ar ee ewe 27 1.64 lO DA Ga a eeye, Gee SeRee ars et ee 50 1.76 ilefaaul Migane 85.5. 2.03 18.1 Although Wilder realizes that dimensional and other physical measurements ‘“‘should be taken during the younger life or at least before there is any marked difference in the experience,” yet he apparently considers that in all of his four sets, whose ages range from seventeen to twenty-one years, “these conditions are met with, as they are all those of young people.’’ Seventeen or more years of post-natal life would seem, however, to be sufficient for the operation of nutritional differences, some of which might tend to cause originally identical characters to diverge, and others, originally divergent characters to converge. It would seem to be inadvisable then to attempt to draw the line between nature and nurture on the basis of physical characters so subject to modification by environment, for a variation in nutrition alone might readily produce all of the differences shown in the sets of physical measurements cited. Nutritional differences doubtless manifest themselves even before birth, and hence would tend to vitiate the results of physical measurements taken on new-born duplicates even if such were available. Consequently it would seem advisable to limit investigation to those characters which reach a definitive condition at an early period and which are sub- ject to little or no modifications due to nutrition. The patterns of the friction ridges of the palm and sole are characters of this sort and should give highly reliable data as to the strength of hereditary control. But even this data, interesting and suggestive as it is, can be accepted only with a considerable amount of reservation; for it has, in common with all other data derived from LIMITS OF HEREDITARY CONTROL 859 a comparison of human identical twins, the fundamental weak- ness that it is based on an assumption which is clearly beyond the possibility of proof. Because certain twins exhibit a most strik- ing resemblance, it is assumed that they have been derived from a single fertilized egg. As Weismann (’02) expresses it, ‘“‘ Wir haben nun allen Grund, die erste Art von Zwillingen von zwer ver- schiedenen Hizellen abzuleiten, die letztere Art von einer einzigen, welche erst nach der Befruchtung durch eine Samenzelle sich in zwer Ever getheilt hat.” Were it possible in a number of cases to determine by exami- nation of the placental relationships of new-born twins whether they were duplicates or fraternals, and were it also possible to obtain data on these cases whose uterine history was known, we would have facts from which we could with confidence draw conclusions. Unfortunately, however, although monochorial twins have been observed at birth, no interest has been manifested in their resemblances. In view of the lack of satisfactory criteria ‘as to the origin of the twins investigated, the writers on these subjects have reasoned backwards from the facts of resemblance to the assumption of common origin, a procedure far from safe, but doubtless justified by circumstances. An arbitrary criterion is thus set up for the classification of twins; and those that come up to specifications are classed as ‘duplicates,’ while those that fail to meet the arbitrary requirements are relegated to the rank of ‘fraternals.’ One cannot but be impressed, as he reads Wilder’s monograph, with the author’s feelings of uncertainty as he attempts to classify certain pairs. The following extracts indicate his attitude: No. VII. This casehascaused me considerable trouble, owing to the preconceived notion that the marks ought to be found identical. The family emphasized the facial resemblance of these twins and when I first saw them they certainly looked much alike. One was, however, an inch taller than the other, and the facial resemblance, after a short acquaintance did not seem as great. . . . The case is plainly one of fraternal twins that resemble one another somewhat more than the average. No. XIII. According to personal appearance these should be dupli- cates. I have never seen them, but the one who took the prints wrote: ‘The Misses*—— are so similar in coloring, figure and features 860 H. H. NEWMAN AND J. THOMAS PATTERSON that even their best friends confuse them.’ It must be confessed, how- ever, that the differences in the formulae cannot be reconciled, and that the palms are, and remain, in respect to the main lines, very different. *In the light of the results presented below we are inclined to believe that Wilder was not justified in thus arbitrarily exclud- ing from the category of ‘duplicates’ such cases as those referred to, for we have found not a few sets of armadillo foetuses which exhibit greater differences than some of Wilder’s so-called ‘fra- ternals.’ Hence, although the results of his studies are valuable and highly suggestive, they are insecurely founded and therefore cannot be applied to the solution of the problem of the limits of hereditary control. The material which forms the basis of the present investigation consists of a collection of advanced sets of foetuses (removed from the uterus with all of their placental connections intact) of a species of mammal, Tatu novemcinctum, in which we have demonstrated conclusively the existence of specific polyembryony ; and hence all sets of embryos, whether strikingly similar or not, are known to be the product of the division of a single fertilized egg. The basic assumption involved in the case of human dupli- cate twins is thus obviated, and at the same time it is possible to eliminate the factor of a diverse post-natal environmental experience by examining unborn foetuses, whose inter-relation- ships are shown by their placental connections. In the scutes of the banded region we have characters little if at all subject, even during gestation, to environmental control. We plan in the present paper to present an intensive study of the phenomena of blastogenic variation as exhibited by these integumentary elements, limiting our present investigations to the well defined banded region, believing that the conclusions arrived at from the study of one region will prove to be generally applicable, and that what is true for one character or set of characters will be found to apply in a general way to the whole organism. In order that there may be no misunderstanding as to the kind of variates we are dealing with, it seems advisable to present in abbreviated form the results of a study of the morphology of the integument and of the variability of its elements as exhibited by LIMITS OF HEREDITARY CONTROL 861 a large sample of the species. Without this data one would scarcely be in a position to appreciate the degree of resemblance or difference that exist among the foetuses of the different sets. Owing to the existence of an extensive curio industry, making a specialty of baskets shaped from the shells of armadillos, there has been afforded an exceptional opportunity for gathering a large mass of data on the variability of the species. By availing our- selves of the large stocks of basket-shells in the hands of various dealers we have been able to examine 1768 individuals for scute and band ‘abnormalities’ and to count the scutes in the banded region of over 500 shells, including those of all males and females sent to us alive. MORPHOLOGY OF THE INTEGUMENT The integument is one of the most characteristic features of the anatomy of the armadillo. For the most part it consists of a series of bony plates which are arranged so closely together as to form an almost continuous armor, especially on the dorsal and lateral parts of the body. When attacked the animal is able to retract itself well within this shell-like structure, much after the manner of a turtle, and although the belly and legs do not possess an armor, in the strict sense of the word, yet even here the skin is studded with horny scutes and the feet are armed with powerful claws. Altogether the integument of the armadillo forms a protective structure of high efficiency in an otherwise defenseless animal. In our species, Tatu novemcinctum, five of the so-called armor shields described for armadillos are present. These are the ce- phalic, covering the front of the head; the scapular, overlying the shoulders; the thoraco-lumbar or banded region (sometimes called the movable zones), consisting of nine bands or incomplete rings; the pelvic, covering the hips; and finally the caudal shield, which consists of a series of rings surrounding the tail (fig. 17). The elements composing the armor exhibit in each of the shields a somewhat different and more or less characteristic arrangement; but since in this paper we are concerned with the study of varia- 862 H. H. NEWMAN AND J. THOMAS PATTERSON tion and heredity in but one of these shields, it is not necessary to enter upon a description of the elements of the others, except in so far as such an account would be of help in understanding the character of the elements in the particular region in question. The nine bands are united to one another by strips of flexible skin which permit considerable motion in this part of the armor. The first band is attached in a similar manner to the scapular shield, while the integument joining the ninth band to the pelvic shield may be present only toward the ends of the band, in which case the middle of the band is firmly united to the shield. The soft parts of the body lying directly beneath the strips of flexible skin are not exposed, because of the fact that the bands overlap one another for a distance equal to about one-third their width, that is, the posterior margin of any band overlaps the anterior third of the succeeding band. On account of this overlapping the banded region as a whole presents a distinctly testudinate appearance. ; Each band is composed of a number of elements, of which there are three kinds: (1) the thick, bony, dermal plates covering the under surface of the band and constituting its main body; (2) the thin, horny, epidermal scutes which cover the posterior exposed part of the band; and (3) the associated hair group. Each bony plate is oblong in outline, with its long axis consti- tuting the width of the band, and in the adult animal has an average width and length of 6 mm. and 30 mm., respectively. As we shall see later, the number of these plates varies in differ- ent bands as does also the number for the same band in different individuals, but in round numbers there are on the average about 62 plates to the band. The epidermal scutes, unlike the underlying bony plates, are of two types, which we may call primary and secondary. The first of these is represented by a wedge-shaped area having a slightly convex upper surface, and with its base forming a part of the posterior margin of the band (fig. 20). The base of each scute has two notches which are situated so as to divide the margin into three areas of about equal width. In reality the notches are but the mouths of hair pits located in the underlying bony plate LIMITS OF HEREDITARY CONTROL 863 and from which fine hairs extend. The plate also has two other marginal hair pits containing hairs, one situated well toward each corner of the base; but these emerge from beneath the scute, and consequently the margin of the latter shows no corresponding emargination, as in the case of the more centrally located pits. The secondary scutes are also more or less wedge-shaped, but, unlike the primary ones, have their bases directed toward the anterior margin of the band (fig. 20). In brief, their bases form the anterior limit of the exposed part of the band. The apex of the wedge is blunt and forms a small part of the posterior margin of the band. Through the median axis of the scute a faint groove extends from the middle of its base to a point near the apex. Upon the removal of the scute it is found that the groove is due to the depression of the suture between the two adjacent bony plates, directly above which the scute is situated. The secondary scutes are clearly composite structures; that is, they have been formed during the process of their evolution by the union of some three or four elemental units of regions such as are seen in the scapular and pelvic shields, which exhibit conditions more primitive than that of the bands. In addition to the four hair-pits already alluded to, several others belonging to the associated hair group appear on the upper surface of the plate. The most prominent of these is a row of six or seven extending along each side of the primary scute mark- ing. These pits possess very fine hairs which often extend above the surface of the band, especially in young animals. There are also faint indications of other hair-pits, both on the convex area of the plate as well as on its anterior unexposed third, but it is entirely beyond the scope of this paper to enter into a detailed description of them. It is sufficient here merely to note that each plate corresponds to a rather distinct hair area of other mammals, and consequently has a definite number of hairs included within its limits. ; It will be evident from the foregoing account that for each primary scute there is a corresponding plate in which is imbedded a definite group of hairs. A count of the primary scutes will therefore also give a count of the plates and of the hair groups. 864 H. H. NEWMAN AND J. THOMAS PATTERSON This whole complex will be, for purposes of brevity, designated the ‘scute,’ because the scute is the index of all of the elements entering into the complex. It has been suggested above that the bands have evolved from a more primitive condition of the integument—one in which the bony elements were not necessarily arranged into definite rows. This becomes obvious when one studies such adjacent parts as the scapular and pelvic shields. In each of these regions the general arrangement is much the same, but the transitional con- dition is more clearly brought out in the pelvic shield, and we shall therefore confine our account to this part. In the central part of the pelvic shield the bony plates are hex- agonal in outline, and are so closely crowded together thata solid bony structure is formed (fig. 16). At best the plates can only be said to be imperfectly arranged into rows. Toward the anterior margin of the shield, however, the serial arrangement into rows becomes more evident, and all of the plates show a distinct tendency to elongate in the antero-posterior direction. In the extreme anterior margin of the shield, or the part corresponding to a tenth band, they become distinctly oblong and greatly resemble those of the true bands. In many of them, however, one can still detect their hexagonal shape, although the anterior and posterior ends show but faint indications of their double- sided nature. Even in the last of the true bands, the ninth, the yosterior end of many of the plates is still in the form of an obtuse angle. On the upper surface of this same region of the armor the scutes show corresponding transitional conditions. The primary scutes are here slightly elevated above the more numerous second- aries and have their posterior ends capped with small white or unpigmented areas, giving to the entire pelvic shield a distinctly pebbled effect (fig. 21). As one passes forward on the shield there is noted a gradual change in the primaries from small poly- gonal areas to those with characteristic wedge-shaped outlines. This is particularly noticeable on the lateral aspects of the shield. In the typical regions of the pelvic shield the secondary scutes are more numerous than in the bands. In place of the single LIMITS OF HEREDITARY CONTROL 865 secondary we have here some four or five elemental units. In the neighborhood of the bands these gradually fuse together, and in the region immediately adjacent to the ninth band are usually typified by two pieces, one, a regular trapezoid in shape, forming the anterior part, and the other, triangular in outline, abutting against this posteriorly. In the true bands these two pieces fuse to form the characteristic secondary scute. A great deal of interesting data might be given concerning the much more primitive condition of the integumentary elements as seen on the ventral side of the animal, but it must be sufficient merely to suggest one or two of their more salient features. On the belly the horny structures only are present, and these are associated with a group of five or six hairs, oreven more. On the legs some few of the larger horny elements, especially those which have a tendency to be arranged into definite rows, are underlaid with true bony plates. Evidently the hair group is the most primitive element of the complex, and in connection with these elements have grown up the scutes and plates. MERISTIC VARIATION OF THE ELEMENTS OF THE NINE BANDS OF ARMOR A. Variation in the banded region as a whole It is our purpose to present in this section only so much of the results of our studies of the variability of the species as appears to be prerequisite for an understanding of the phenomenon of fraternal correlation. A concise tabulation of the distribution of the variates and a determination of the principal variation con- stants should serve the purpose in view as well as would a more detailed account. That the characters dealt with show a high degree of variability is obvious if one examine the array exhibited in table 2. The number of scutes in the banded region vary all the way from 517 to 625, a range of nearly 20 per cent. In connection with our studies on fraternal correlation it will be of value to bear in mind this high species range of variability. In table 2 is presented the array of individuals investigated, comprising 508 adults and shells, 866 H. H. NEWMAN AND J. THOMAS PATTERSON TABLE 2 Showing distribution of frequency of scutes in the banded region of 508 adults | | NUwens =| “FREQUENCY || NUSTOE OF | wrequency | “°MtEE,”” | FREQUENCY BL if 554 8 591 1 518 0 555 9 592 0 519 1 556 11 593 1 = | 1 557 10 594. 0 521 2 558 Pal 595 0 522 | 1 559 13 596 1 523 il 560 10 597 0 524 | 0 561 18 598 0 525 | 1 562 15 599 0 526 0 563 19 600 0 527 2 564 6 601 0 528 5 565 15 602 0 529 0 566 12 603 2 530 0 567 10 604 0 531 1 568 inl 605 0 532 3 569 8 | 606 0 533 3 570 12 any } 534 7 571 120 | ~~ «808 . 535 2 | 572 8 | 609 0 536 2 | 573 7 610 0 537 3 574 8 611 0 538 7 575 4 612 0 539 6 576 8 | 613 0 540 5 577 7 =) gp T6ie 0 542 5 579 7 616 0 BASH os | F 580 5 617 o 544 | 16 581 3 618 0 BAS 5 582 2 619 0 546 7 583 3 620 0 BAT | 18 584. 3 621 0 548 | 14 585 3 622 0 549 | 17 586 0 623 0 550 hia 587 4 624 0 551 | 7 588 0 625 I 552 | 12 | 589 3 553 | 15 | 590 2 LIMITS OF HEREDITARY CONTROL 867 ‘In grouping the variates shown in the table for purposes of seria- tion we have decided to fix the group size on as logical a basis as possible. The average range of variability of the sets of quad- ruplets is eight scutes. We shall therefore seriate the array in groups of eight for reasons which will become clear later. The polygon of variation obtained by this grouping represents a close 120 110 100 90 80 70 60 50 40 + N oO © o + N (=) co oO s N So wo N oa) Tt +t w o ~ ioe) oO fon) fo) -_ N N wo w wo nn w w w wo wo wo ive) Ke} oO o ' | ! 1 ' | ! 1 ! ) ' ' ! 1 ~ uw) oO - fon) ~ w on oO (op) > wo oo) - — N ~ wo ~ foo} (ep) fo) _ N ayy a | a «a * hats Bb wm tS 6 6 Fig. 1 Polygon of variation for the total number of scutes in the nine bands, as determined from a seriation of 508 individuals. Class range = 8 scutes. The solid line represents the observational and the broken line the theoretical normal curve. In this and the succeeding figure the abscissae refer to number of scutes and the ordinates to the number of individuals. approximation to a normal array as may be seen from a compar- ison of the observed and the calculated curve (fig. 1). The three important constants of the frequency polygon prac- tically coincide; the median being 558; the mean, 558.2; and the mode, calculated from these two constants, 557.6. On the ex- S68 H. H. NEWMAN AND J. THOMAS PATTERSON treme right of the polygon will be noted several extreme variates, ~ so far separated from the others as to be examples of discontinuous variation. The rejection of these extreme variates would render the three constants, mean, modeand median, practically identical, but it would seem inadvisable to take any liberties with the data. Rather we would prefer to accept a close approximation to the normal curve of frequency as an indication that we are dealing with a case of chance variation, little if at all disturbed by com- plicating factors. From the above array have been calculated the standard devia- tion, 14.89+0.31 scutes; and the coefficient of vanes 2.685 + ().32 per cent. In addition to these facts we are also led by analogy to infer that we have in the scutes of the nine bands of armor a variant which is inherited in the blended fashion. This inference is borne out by the evidence derived from an examination of the mothers of the various sets of quadruplets (table 6). Of the fathers we unfortunately know nothing. B. Comparative variability of males and females While the main mass of our statistical data came from an exam- ination of baskets made from the dried and shaped shells, a con- siderable number of individuals were identified as to sex. The arrays dealt with in this section consist of the scute counts of animals shipped to us alive, and of the advanced foetuses in our collection. The larger number of females is explained by the fact that our first interest was centered on the facts of develop- ment and hence we ordered only pregnant females. The number of individuals is, however, sufficiently large to furnish a basis of comparison between the sexes. Table 3 indicates the frequency distribution of the variates of the two sexes. It will be seen at a glance that the array of males here presented is decidedly more variable than that of the females. Part of the disparity may be due to the occurrence of one set of foetuses, all of which have a scute count of over 600; but even if we arbitrarily exclude this aberrant set we do not materially reduce the variability of the LIMITS OF HEREDITARY CONTROL 869 male array The mean of the female array will be seen closely to approximate that of the whole collection, while that of the males is several scutes higher. It is our impression that a larger collection of males would place the mean at about 559 or 560, which would indicate a slightly higher center of frequency for males than for females. Although the mode of the two sexes TABLE 3 Showing distribution of frequency of scutes in the banded region of 146 females and 81 males FrMa.es (146 LNDIVIDUALS) Mates (81 INDIVIDUALS) Re ecorEs. FREQUENCY Read FREQUENCY | pleas ie eres | Se aneee Beane tee, cet =e s 3 Bie sees 559 Bee le 520 1 567 1 520 1 560 2 e527 1 568 3 527 if 561 Gi) ess) 2: P5569 2 528 2 562 2 | 585 rl SO 3 532 1 563m |) 3) 540 1 (Ayana 3 534 2 564 | «4 541 1 572 3 539 2 565 5 542 1. ah oer! 1 540 1 566 4 544 Tee Sry 2 541 1 567 2 547 3 sj) 57g eee ee 542 1 568 1 548 4. 4) 9 2580) 1 543 3 569 2 550 2 587 1 544 2 571 Bey |) 55 2 50 Pieued 545 4 572 3 li 552 ee Wh 0K 3 546 5 573 Ae 558 5 | 607 1 FAR 1G) 574 3 554 1 621 1 BAS) Ul cee 575 1 555, Ui) g ae | B50: sie oo 4 577 1 558), Via | 551 4 578 3 J, 550s gy saat | 552 4 581 2 | 560 1 553 9 582 1 bi 561 ae | 554 3 583 1) 0p ABG20) eVemie2 | 555 6 585 1 56S) i Gale 556 5 589 1 564 1 557 5 590 1a aos 2 558 6 596 ae sale 1 | | Mean = 557.2 | Mean = 561.3 scutes Standard deviation=13.35 scutes Standard deviation = 17.71 scutes Coefficient of variation=2.39 per| Coefficient of variation = 3.15 per cent cent 870 H. H. NEWMAN AND J. THOMAS PATTERSON probably differs only to a slight extent, yet the standard deviation and coefficient of variation of the males are decidedly higher than those of the females, a condition that accords with our pre- vious knowledge of the comparative variability of the sexes. The possible significance of this variational dimorphism will receive attention in the discussion of the hereditary control of sex. C. Variation in the individual bands In view of our studies of fraternal correlation among the foe- tuses it appears to be necessary to determine the scute variability, not only of the banded region, but also that of each of the indi-_ vidual bands. We will thus be able to determine whether there is a greater or less variability in individual bands than in the whole region under investigation; and in addition we can compare the variability of any one band with that of another. In table 4 is TABLE 4 Showing distribution of frequency of scutes in the individual bands of 508 individuals FREQUENCIES PER BAND NUMBER OF SCUTES 1 | 2 3 4 5 6 7 8 9 5 Meee en Pe OO) t |o a | 0) SOc ago aor oe HBr cee SU ee ee. (ae aes 5 1 4 2 0 0} 0 56s Ca ener ae le ISe Oey | Mea al eas 3 ea Ly Reem NL Ns Tce he OF 4 267) a6y |) 20 ip 27 5 34) OA hae Cpe a ee a) 19 | 83-88 | 929) 58-41 a ie 59 36° | 65. | 7d] 60" P70" aa 55 6 |aue 60 Soe, Wane ee 7) 3) 102 |2f09) | 88" "116. |) 79 4) S40) 1G eS eee ey ae 7% || 86 [M04 |) 972, 80} GO, 1.49 mle TSP eet 62ers. een ne Oi 12-88 - Ge i eo 96 | 85 | 53 | 63 63.2 ans eee eee } $3: | 440 )040° | 61 |) 47°] «85 1025) 196" as Gtk CO creme Gl | 30. (95 | 86-7 1591s 57-4 So") “go, Boe Gy: cep aes eee 37 a W2e Gad | 17a ere Ger aia aaa 66). 5.5 0 Beeler PRM Big 2h 5) 224 Dial 4 34a Ose GT 2 See ae yee PP comet BS: 1 2) 21 oye Nmeser GSc 5. eee 2 Oe a al 1 1 ae 28 GON sive ee ee OF)! OAoa KOZ aro 3+) aaa | 56 BO. dj. =2 oo ee 0 0 024/440 0 Ol. @ 2h 9 hee IE dele eee 1 0 0 | 0 0 1 0 Br sions De 5 | 0.) 470 "0% | POM 0's tO emioenge tae a lsaat) 73 0 0 OF 10 0 0 OF eee —— | LIMITS OF HEREDITARY CONTROL 871 given in convenient form the data, band for band. The arrays for the individual bands are considerably more satisfactory for purposes of seriation than that for the banded region as a whole; for the number of classes is comparatively small and consequently the number of counts sufficient to give a smooth curve without any grouping of variates. As a sample of the type of curve derived from a seriation of the variates of the individual bands let us take that of band 1 (fig. 2). 100 Sy) 55) 57 58 59 60 61 62 63 64 65 66 67 68 69 70 7h: Fig. 2 Polygon of variation for the total number of scutes in the first band, as determined from a seriation of 508 individuals. Class range = 8 scutes. The solid line represents the observational and the broken line the theoretical normal curve. It will be noted that the mean, mode and median coincide and that the observed and the calculated curve are almost identical. The curves for the other bands are of the same type as that shown for band 1. These observations would seem to indicate clearly enough that in the distribution of the elements of the integument into bands we have a pure chance process controlled by fairly simple mechanical laws. JOURNAL OF MORPHOLOGY, VOL. 22, No. 4 872 H. H. NEWMAN AND J. THOMAS PATTERSON In order that it may be clear that the sample of the species presented by the sets of foetuses studied is fairly representative, it would be well to compare the variation constants of the foetuses with those of the large sample of the species here dealt with. Table 5 furnishes the means of comparison. This table reveals to the student of biometry a number of interesting conditions not strictly pertinent to the present inquiry, but doubtless worth noting briefly : 1. With a few minor exceptions, all of which fall within the limits of probable error, there is a gradual decrease of the mean, mode and median from the first to the fourth band and then a - gradual increase in these constants from the fifth to the ninth band. 2. The absolute variability, as indicated by the coefficient of variation is greater in each of the bands than in whole banded region. 3. The proportional range of variability is likewise greater for the individual bands than for the banded region as a whole. 4. The standard deviation may be considered for convenience to amount to two scutes. This will be worth remembering when we come to study the band correlation among the foetuses of the various sets. For the convenience of the reader it seems well to consider the subject of hereditary control, as it is exhibited in connection with TABLE 5 Showing the variation constants of the individual bands of 508 adults | ] . | | RANGE IN BANDS | MEAN fees MODE | aslo ba parce ea | press dl. eee | 62 +0.066 |. 62 62.0 2.223+0.047 | 3.59+0.0388 | 24.19 Di nceee | 60.55+0.065 61 61.9 2.217+0.046 | 3.66+0.038 | 23.3 Oekacees | 60.44+0.063 60 61.76 2.12+0.045 3.51+0.037 | 26.46 7 ae 60.23+0.062 60 60.92 2.08+0.044 3.45+0.035 24.9 Die aw 61.16+=0.063 61 61.64 2.11+0.045 3.46+0.085 | 24.54 Goaeeee 61.85+0.066 | 62 62.3 2.23+0.047 | 3.61+0.039 | 29.1 Aveo oats 8 | 63.09+0.066 63 62.82 2.22+0.047 | 3.52+0.037 | 20.6 cia eae | 64.43+0.068 | 64 63.14 2.24+0.048 3.49+0.036 26.39 9.......| 64.44+0.068 | 64 63.12 2.2 +0.046 | 3.57+0.038 | 23.27 LIMITS OF HEREDITARY CONTROL 873 the normal variability of the scutes in the banded region and in the individual bands, immediately following the variation data for the species, although it might appear to be more logical to complete the account of the conditions in the species before attack- ing that in the foetus sets. D. Fraternal correlation in the banded region as a whole; an index of the limits of hereditary control Having presented the facts of variation for the species, we are now in a position to investigate the conditions exhibited by the various ‘fraternities,’ which consist normally of ‘identical quad- ruplets,’ 2.e., of four individuals of the same sex enclosed in a com- mon chorionic vesicle. We have in our possession at the present time about thirty sets of foetuses in a stage of development suf- ficiently advanced to permit of the determination of sex and of the accurate enumeration of scutes. Of these some few are of especial value on account of their more or less atypical conditions; a few others are characterized by an atypical number of foetuses. There are, however, ten sets of each sex in which the number of foetuses, the arrangement of bands, etc., is sufficiently normal to permit of statistical treatment. These twenty sets furnish the material for the present study of correlation and the limits of hereditary control. The enumeration of scutes may be relied upon to be accurate, since they were all counted independently by both of the writers and all points of discrepancy carefully examined and settled according to mutual judgment. A full tabulation of the scute counts of the twenty fraternities is pre- sented in table 6. As in our previous papers the four foetuses are numbered as follows: Pair A { I. The lower individual of the right hand pair II. The upper individual of the right hand pair PairB {III. The upper individual of the left hand pair \ IV. The lower individual of the left hand pair Since each of these sets of foetuses is derived from a single fertilized egg it should be possible to determine exactly to what 874. H. H. NEWMAN AND J. THOMAS PATTERSON TABLE 6 A—Female sets . | | BANDS mn =eReee FOETUS AND 2 | MOTHE R* | 5 | by i) A Be N64, 278) ees | | as EE: ui Sassthnckesonieis emis ie eters 60 62 60 60 60 63 63 66 63 | 557 anintee tk Viel iota Soe a heictera eitopeeaeree 61 61 62 | 59 | 57 63 64 65 64 556 au Sethe span Seen | 60 60.) 5994) 5901 62 64s «63 63 64 553 ILIV). Casehioye ht Rios srarbuslteo the 60) | 59") 461 60) 63 62 60 64 62 | 561 Sethu Aen eABO le By) 3681 se 160%, (80%) 604 6 | 6h) cer enmneE Seok: wee EL stot orcarcic s vinihae Bettas 63 | 60 59 | 60 61 60 61 61 63 548 TET = «ci sevatasote avers tenee ee es | 64 61 58 60 61 | 61 61 65 63 554 1 RR Oca ec rian 61 62 59 | 61 | 60 63 64 66 62 | 558 Mts SORES Seether Ok 59 57 58 | 63 62 63 63 67 61 | 553 ha hee Pee UD wcasayehs e/a ace etsyopetete > Wea 61 59 59 60 60 61 65 63 ay |) GER PLT cet, cca ater eters 62 58 60 60 61 62 64 67 64 558 DSU. cepaiay eras eo eraih ave rare ty et 62 | 60 59 | 60 63 62 63 64 65 | 558 TE EAMES Sete TNE chee 63 61 63 60 61 63 | 63 65 67 566 98 1 aa Per ata abn catattan 61 63 62 63 63 61 62 65 65 | 5657 i on his MID Kose cee he wets) OL 62 61 61 61 61 65 64 66 562 WBE P Sie carota atateierazee ciceeacie 63 60 62 61 62 61 63 65 68 565 Mother: ss. .seer iene 62 62 59 61 62 60 64 63 64 557 To er ene eee) 63 62 61 59 63 61 62 66 64 561 99 Wise sconces Hdecs cows 63 61 63 61 61 62 64 64 65 564 Be: FREY Bay lt ELUTE Petco: Ste Pe ORR RES 63 62 62 62 62 61 64 66 65 567 TV sas, she ay ee ctaueervovenes 63 63 60 61 63 62 63 67 63 565 Mothers jae. crctie saynists 65 64 60 62 65 62 64 66 66 574 rere ies eh ee ead hae 60 62 61 61 60 64 62 62) | 554 119 AT Gil bis ssh dock 1s Nh wtacs aeere 63 60 59 60 60 60.5 64 66 62 554.5 a wr cas TS ea rote, shtee ara ste ieee OL 60 58 60 60 61 62 63 65 | 550 Ula sce hes oem 63| 59'| 60 | 59 | 61 | 60 | 61 |> 48) G20) ads Mother: ek ieee 62 58 60 61 59 61 62 63 62 | 548 ee NE ete Sen 62 | 62 | 64 | 61 | 60 | 62 | 64 | G& | (62> \i5er 121 TiS aS Node i otecerctins fe) aoceeete 61 62 59 62 63 62 63 67 66 565 Vaehaeeha ey: ELLY «pace ye velorabe.t en OL 62 61 60 61 64 65 67 63 564 WUE Sacrteaccene aera Maen 64 61 63 62 62 61 63 65 68 569 IMO GHEr*S rae dak 60 59 61 61 63 64 65 67 65 565 Arg oA ee ae ae 4 62 60 61 60 60 60 59 61 61 544 122 EE ea tctatatovoradee cera soneee ele 61 60 59 60 62 60 61 63 63 549 <1 Fence Eidos tere terc rotons sae AOL 61 61 59 59 59 61 62 63 546 L I Waa See Sie ae 62 62 60 61 60 58 59 64 62 548 Mother: 404.4..0snee ee 59 58 59 57 55 55 57 59 8) (rol "(Wesshcueselceds-neeec| GL or | easy Met | 88 |) Yeo!) 850) 1 ae eee 123 TT ort a oeete Pores arctan eerie 63 62 61 57 58 61.5 58 63 61 544.5 > ae il MTD ao heen trie, COL 60 61 60 61 61 | 59 62 60 545 {thie ces eee eee: 59 62 61 59 60 62 60 62 63-548 IMOthEnete ten ee 58 61 59 63 61 60 63 65 63 | 553 (iets ee ee oe eee AGONY SS ade Ste Gay.) Our] Glia le G2 ema Lunes Oa mmlmerse EDA es eteait Gottere octets anes 60 60 61 62 63 60 60 63 [pie | TLDs Re ep scee eto ee teen 61 59 60 60 59 61 61 65 62 | 548 IPA eramac dsc , A rudimentary em- bryo lies between m1 and Iv l Wane cere ctaceirtete seks 61 60 62 61 60 63 64 63 63 557 Mother sarees 58 60 60 61 61 62 64 65 64 555 *In some of our first collections the mothers were not kept, and hence the seute counts on these eannot be given. SERIES LIMITS OF HEREDITARY CONTROL 875 TABLE 6 B—Male sets BANDS n FOETUS AND < = = B| MOTHER = 1 2 3 4 5b 6 7 8 9 & i | i} ase cR erates eee 62 | 62 | 61 | 64 | 61 | 63 |. 65 | 66 | 67 | 571 |: eR ELD os , 62 | 62 | 63 | G4 | 64 | 62 | 65 | 65 | 67 | 574 Hide. SM ae ay | 62 | 61 | 61 | 64 | 62 | 63 | 64 | 65 | 66 | 568 Lae oe ne Re 63 | 61 | 60 | 61 | 63 | 62 | 65 | 67 | 66 | 568 | HRae A cient oe say | 68 | 62 | 60 | 61 | 63 | 65 | 65 | 67 | 685 | 571 TE ee Peete 64 | 64 | 63 | 60 | ai Gam iek | 67 | 66°! 574 TERE otra eee erase 63 | 60 | 62 | 61 | 63 | 65 | 65 | 66 | 65 | 570 trypan tts onal oie cet | 64 | 60 | 61 | 61 | 64 | 64 | 65 | 66. | 65 | 570 ihr Ber ae? at oie | 68 | 58 | 58 | 50 | 59 | 59 | 60 | 57 | 60 |.533 1F Ris AAR ee Teer | 6 | 59 | 56 )~56 | 59 | 58 57 | go | 61 | 527 {ke sepgobceee nonishe Sees } 60) 59) | 58) 61 |) 58) AS GOR hy ed =o: | 535 hee icin ane 59 | 58 | 57 | 57 | 56 | 56 | 59 | 58 | 60 | 520 (Mothers) in sis) | 58 | 56 / 60 | 56) 57 | Gt | 60 | 60 | 59 | 527 Ree See ese ea | 62 | 64 | 61 | 60 |-62 | 60 | 6o | 63 | 63 | 555 TE Shen EEE cs: 63 | 59 | 58 | 58 | Gi | 62 | 62 | 65 | 67 | 555 [ae Cree Ehe a aren | 60 | 59 | 58 | 58 | 60 | 61 | 60 | 65 | 67 | 548 Tg ere ete Be | 61 | 60 | 62 | 60 | 59 | 60° | 60 | 63 | 63 |.548 Mo tern aed ees 64 | 62 | 62 | 60 | 61 | 62 | 63 | 67 |. 65 | 568 | | ase ee ek ae Pe eB | etl, 4 62 | 61 | 62 | 63 | 61 | 65 | 66 | 565 Fi ae a eee a | 63 | 61 | 62 | 61 | 62 | 64 | 63-| 68 | 68 | 565 iTS ae een ee ied | et | 62 | 61_| 62 | 62 | 62 | 6s | os |. 65 | 565 be gear reve eee Gh) 6H, |) GO") 63+, 161 | 76263, Gaonhmenee abi ied Mest tow ceeaet SOF oe Ov) |) 6057 062i kodaman Osman Goma i ORO ahaa Ma, 60 |-62 | 61 | 56 | 57 | 57 | 62.| 64 | 68 | 542 Faia es nied aie. 63) | 58°} 60 | -56 | 59. 7) <60) emerges) oa 157 IVR renee tert oa. 62 | 57 | 58 | 58 | 58 | 62 | 61 | 66 | 66 | 548 PE Saree aah ed 64 | 62 | 63 | 62 | 59 | 62 | 65 | 68 | 67 | 572 Fa ea ine ee ie Sel 66 | 62 | 64°| 63 | 63 | 68 | 65 | 67 | 66 | 579 [DUR ee eters Se ees 64 | Gl | GL | 58%) 1627 1963) 6s 0/65 a 66) | 563 Te EE REARS Ke Aten tore 61 |, 64 | 61 | 62 | 62 | 62 | 65°) 68 | 65 | 570 IMopherss 95. asa 63 | 60 | 59 | 60 | 60 | 59 | 68 | 65 | 66 | 555 | fa Shale te ee tan ek sabe 64 | 62 | 60 | 59 | 59 | 60 | 62 | 63 | 62 | 551 lf Sec rate Me las | 62 | 62 | 59 | 50 | 50 | 62 | 62 | 64 | 63 | 552 Winnie etre eee 64 | 64 | 60 | 60 | 61 | 60 | 63 | 65-| 63 | 560 Use cee oe Ate ake tear ee 64 | 60 | 60 | 62) | 60 |) 62 |) 62) | 164 | 65, 559 (Mothers tracts )s.. | 63 | 59 | 60 | 60 | 59 | 60 | 60 | 63 | 66 | 550 He caasacon Aaa SAe eee 69.) 68) GGL 65. Wah | 67 Weer| 68.\| 4606 "Tingste ae 68 | 64 | 66 | 63 | 68 | 69 | 67 | 71 4 70 | 606 | isc a ee eee 70 | 69 | 68 | 66 | 67 | 67 | 72 | 71 | Z | 621 te iin eee 66 | 67 | 68 | 65.5 65-| 70-| 68 | 68 | 70 | 607.5 (Mother s.os20-.<.000 08 63 | 60 | 63 | 62 | 63 | 65 | 64 | 66 | 67 | 573 NUR ete nas cit 62 | 61 | 59 | 60 | 61 | 60 | 64 | 63 | 63_| 553 ERP ate od) we sas ae 62-| 62 | 58 | 58 | 61 | 62 | 61 | 63 | 63 | 550 Nas ae Aten ee aa | 61 | 62 ) 60 |~59 | 61. | 62.| 63 | 62 | 64 | 554 Uv TEE oe eon ee we 61 | 58 | 60 | 61 | 59 | 61 | 63 | 64 | 64 | 551 IMiGbher. ¢a2ce eco e. 63 | 62 | 61 | 59 | 62 | 62 | 64 | 63 | 66 | 562 876 H. H. NEWMAN AND J. THOMAS PATTERSON extent the definitive number of scutes is determined at the time of fertilization; for presumably, in so far as the four individuals of a set are alike, this similarity was determined before they became separate individuals, and, in so far as they differ, their differences are due to epigenetic factors operating after the sepa- ration. Should they prove to be exactly alike in the number and arrangement of scutes we would be warranted in claiming that hereditary control was absolute; if we do not find them exactly alike we may claim that hereditary control is not absolute but only partial. As an index of the strength of hereditary control we have decided to employ Galton’s coefficient of correlation, a con- stant indicating exactly what per cent of complete correlation or of absolute hereditary control exists. A short method for determining the coefficient of correlation, and one which seems especially devised for cases like the one in hand, was described by Harris (’09). This method appears to have been elaborated by Professor Pearson for use in cases of symmetrical correlation tables in which both variates have the same mean and standard deviation. The correlation table is made symmetrical by using each individual of each set first as a ‘subject’ and then as a ‘relative.’ In the case of our twenty sets of foetuses such a table requires over one hundred vertical and horizontal columns and would not be suitable for reproduction here. As a matter of fact it is scarcely necessary to make a cor- relation table in order to derive the desired constant. One need only determine the positive differences between the subject and relative classes and their frequency. In the present case we find the positive differences and arrange them as follows: Difkerencess. 0) “ly 2 3 45— 5O G6) i 8) 0) 10) a Aa ao Frequency 226 39° 10 16-13. 10° 9: 40. 7 34° A ad (OL a Sie Multiplying the square of each difference by its frequency and adding the products, we obtain the sum of the squares of the differences [Sv?]. Now since the negative differences are the same as the positive we may-double the above sum and divide by the number of cases, 240, and thus obtain the standard devia- tion of the difference squared (cv?), which is the only constant LIMITS OF HEREDITARY CONTROL 877 not already determined that is needed for our calculation. The formula used is as follows: 2 ox? 1 - 33.5 Substituting we get: r = (1 — 5 Te = 0.9348. We can use this coefficient of correlation as an index of the strength of hereditary control and can say concretely that, so far as the total number of scutes in the banded region is concerned, the individual is predetermined within certain narrow limits, or up to 93.48 per cent of complete determination. Beyond that point the variations in the number of scutes are due to differ- ences in epigenetic factors, whose nature we do not pretend to understand. No such close correlation as this has been determined for any of the ordinary blood relationships. The closest of all blood relationships, namely the fraternal, is represented by a correlation constant of only 0.4, a fact familiar to all who have read their Galton. Even homotypical parts of the same individual are cor- related only a little more closely than are brethren, a fact which leads Pearson to conclude that the fraternal relation is only a special case of homotyposis. In brief there appears to be no other inter-individualistic relationship so close as that which we have found to exist between the individuals of our sets of quadruplets. If we desire to find correlations comparable in closeness with that determined in our material, we must seek them among the closest of intra-individualistic relationships, such as that existing between the antimetrically paired organs of the same individual. As an example of such constants we may cite the correlation coeffi- cient between the right and the left sides of the carapace of Gel- asimus, which is 0.947, or that between the lengths of right and left meropodite of the first walking leg of the same species, which is 0.918. These constants are obviously of the same grade as that determined for the scutes of our quadruplets. From these facts we are doubtless justified in concluding that the four indi- viduals of each of our sets is morphologically the equivalent of a 878 H. H. NEWMAN AND J. THOMAS PATTERSON single individual and that we are dealing mae a special case of intra-individual variation. Were the phenomenon of specific polyembryony in need of any further support, the discovery that the degree of correlation among the foetuses of a set is the equivalent of that found to exist between the right and left sides of the same individual, would in itself constitute a demonstration of its validity. TABLE 7 Showing the variability of seven pairs of duplicate twins in terms of the coefficient of MENS INS eran { DESIGNATION OF TWINS | COEFFICIENT OF VARIATION WGISTIT aI Secs ah Oe eee eae UR RE iy aR ara eet A eee ae | ike WEEMOM BTCA Pea ye eke tee dee ert en ae ne ene erento | 0.22 MELMONSS (HID Hirk Lette heel bes Rrra ae eee es Cree | 0.34 Wal ier Sa Ret ih taeda ci eons earn, es a Re ne 1.438 Nall Ve Nes ech ie epee Rp Feet ems Mar vs eee Rm aA 1.65 VSG OTS SHI Ties, c Meee ee cee se PPS cee oo ey, Nee irae, Wiltlenesriny ic bcc alt Ars cit oP eee, OF Cano) REED Soden Ls iley IAW ETT earn Mee Sea ce ees nn Mrs cube aes Ba ae ee cms ee | 1.06 In order to be able to institute a comparison between the strengths of hereditary control exhibited by armadillo quadrup- lets and human duplicate twins it becomes necessary to employ some method which does, not involve the use of the coefficient of correlation, for the twin data is too scattering and diversified for our purposes. The usual method employed by the various writers for indicating the degree of resemblance between twins is that mentioned in the introduction, namely the per cent differ- ence method, but this can be used only in the comparison of pairs. A better method for our purposes involves the determination of the coefficient of variation for each pair or set and an averaging of these constants. Table 7 gives the data on the seven pairs of human identical twins referred to above. In table 8 are recorded the coefficients of variation determined for the twenty sets of foetuses. It will be noted that the degree of resemblance among the quadruplets is, on the average, nearly twice as close as that of thetwins. Thisis perhaps no more than we should expect, even Lay f LIMITS OF HEREDITARY CONTROL 879 if we were certain of the origin of each pair of twins, for the twins have doubtless been much modified by post-natal environmental experience. Incidentally it may be noted that the male sets exhibit a much higher absolute variability than the females sets, a fact which will be brought into discussion in a subsequent connection. Perhaps a more equitable comparison between human twins and the armadillo sets would be instituted if we were to consider the quadruplet set as consisting of two pairs of twins. Accord- ing to this method of comparison the difference between the two species is greatly increased, for the average of the forty pairs of armadillos is only 0.27 per cent, while that of the seven pairs of human twins is 1.62 per cent. In concluding the present account of the strength of hereditary control as it appears to be exercised in the case of the total num- ber of scutes in the banded region, it should be mentioned that further investigations now in progress, dealing with other regions of the armor and with certain dimensional variates, indicate TABLE 8 Showing the standard deviation of each of the sets of foetuses and indicating the comparative variability of male and female sets with respect to the total number of scutes in the nine bands. FEMALE SETS MALE SETS STANDARD COEFFICIENT - | STANDARD COEFFICIENT SET DEVIATION OF VARIATION SET DEVIATION | OF VARIATION (IN SCUTES) (IN PER CENT) (IN SCUTES) (IN PER CENT) DePee bm cere 2.38 0.438 La ee 2.48 0.48 A Fa, See es 3.69 0.66 As a, oa a 0.5 0.08 95 25 0.45 Peer SP ho le. 25386 1.10 eee a came ard he 15 0.26 D2 arene ono 0.63 OOM relat 72 2.16 0.38 LUGS eee 3.46 0.61 UGE es ee. 2.59 0.46 A. sees een 2.38 0.43 121 2.86 0.50 (20 Se eet 5.70 0.99 122 1.92 0.35 1s 7: eye oe 3 Sea 4.02 0.72 123 1.87 0.34 LBS ee the! 6.36 | 1.04 pe ech, 2.85 1.06 TSS eaebys ate. 1.56 0:28 880 H. H. NEWMAN AND J. THOMAS PATTERSON that the degree of correlation found to exist in the banded region © has its parallel in other regions; in fact some of the results already obtained seem to indicate the existence of higher degrees of cor- relation than any yet determined. E. Fraternal correlation in the ~ndividual bands In one of our previous papers (’10) we had occasion to refer to the subject of pairing in the following words: In this connection it should be mentioned that even where there is exact resemblance between the individuals of a pair in the total num- ber of scutes in the nine bands of armor, there is no perfect correspond- ence with respect to individual rows. The resemblance in total numbers of scutes is, however, a matter of more importance than the exact man- ner of their arrangement into rows, which is a secondary process. Further investigation has thrown light on this situation and we are now in a position to make a more satisfactory statement of the conditions referred to. By the application of statistical methods we have been able to compare the strength of hereditary control exercised over the total scute number and that over the arrangement of these scutes into bands. In order to do this it has been necessary to determine the correlation coefficient of each of the nine bands, using the same method which was applied to the whole banded region. The method pursued is illustrated in the case of band 1 (table 9). The same method of procedure was carried out for each of the nine bands and the data and results are seen in concise tabular form in table 10. Although the average coefficient of correlation for the individ- ual bands is high as compared with that found for any relation other than an intra-individualistic one, it is decidedly low as compared with that shown to exist in the case of the total number of scutes in the whole region. This would seem to indicate that there is here a much wider scope for the operation of epigenetic factors. Evidently the alignment of scutes to produce the bands is, to a large extent, a mechanical process, involving a certain amount of shifting due to pressure, etc. In the earliest stages of scute formation it is probable that the primordia of these elements are arranged somewhat after the fashion seen in the abdominal LIMITS TABLE 9 OF HEREDITARY CONTROL 881 Correlation and difference tables for the 20 fraternities, for band 1 Mean = 62.16 Standard deviation = 2.009 scutes 3.23 per cent Coefficient of variation | @IESSES: gis ack ceo 59 60 | 61 62 | 63 | 64 | 65 | 66 | 67 | 68 | 69 | 70 | Total Deviation rel. | seas | eee a ele ie) ee Devia- | Classes aenoindan | | of s | mean | | BO ce: ) =3 16 2 4 Si -8 12 60s eo aG. Leo 6 9 eal 4a 24 Mie oe Ps le (i OF 22a ae ia | 5 1 69 62 e. 216) 3 AE ee | SBE SMe 44 cae 0.84] 3 4 | 14 7 |16| 5 49 Caer 1.84 | i 5 4 | 5 | 10| 2 27 Gna: 2.84 | | 0 66. aks: 3.84. 1 2 1 te ee eae Cues 4.84 | 0 (ee ioe 5.84 1 be eile assy G0 Ls 6.84 (ea 1| Miers ROmeeety WY aed | 1 iu 3 Matal so )ee. i2 | 24 | 69 | 44 | 49\/ 27) 0| 6| 0| 3] 3] 3| 240 v = positive differences of x and y Coat woe aan de Ie Meat G “7 tg. eGu. ima en O: | Maen e 4 3 3 0 Ord 0 0 Ory AOE Selim nOiea cts ser oe 1 Di Oy] O00. \) OCR aa ae, 22| 14 | 14 6 0 1 Oia Oni lye Cie ee |S(v)2 Ieee, cle 0 0 0 0: jc Oule 400) nue 389 16| 5 0 0 0 0 Ov tach ae pa 10); 0 } 2 0 0 OF Oe PO | i% Od Oyu 6 0 0 0 On | 80 Oni) 20. ee 1 Ie hie | | | 25 Ol OO Ca s5..| | | | 373 aaa |) a aris | eae Ol Lees : | | | 0 | 39 | | | | | 68 | 882 H. H. NEWMAN AND J. THOMAS PATTERSON TABLE 10 Showing coefficients of correlation and other constants in each of the nine bands of armor BAND MEAN ) ox? av anaes oN rE) ty IC ane Oe A ees 62.16 4.090 3.10 0.6201 DR BOS Cm ater tr CA re eee 61.03 4.712 3.94 0.5820 Shea Seu A Pee 60.05 5.887 3.80 0.6769 Le Reged ree Ae sie eee scl See Shs ay Ae 60.45 3.972 3.93 0.5051 Dee ALORS Mv te ee ee €0..97 4.828 3.41 0.6469 Gis cote yen een es ae 61.73 4.970 2.52 0.7458 hae pd dich ON Oak AAO RES 62.70 5.735 | 3.30 0.7123 SS LG eg | MOE Be EE = WNP te eS ORD te 64.40 3) 1153) | 3.75 0.6363 NRG AA Re Veg cy ts Ee ee ie 64 .22 5.875 3.66 0.6880 IMME ait wy. Ve taney ects toe cee Pe bs Yate eee ae ee ae 0.6459 region, where the rows are far from straight and where it is not always possible to assign scutes to their respective rows. At such a stage in the development of the banded region we may assume that the ultimate position of certain scutes is determined only in a general way, and whether or not they come to be aligned with one or the other of two adjacent bands is determined by fac- tors beyond the limits of hereditary control. Considering that the scutes are not only determined rather sharply for the whole region, but that, within this region, their primordia would also be regionally distributed under the influence of hereditary control, the coefficient of correlation is no larger than we should expect. But the correlation is so low as to preclude the possibility of any strong hereditary control being exercised over the assignment of scutes to particular bands. ‘This conclusion is in close harmony with that expressed by Wilder in connection with his studies of friction ridge patterns in human twins, and which is quoted in the introduction to the present paper. We might well ask with him: “Have we then arrived at the limit of hereditary control of the pre- determining mechanism beyond which mechanical laws are alone operative?’ Perhaps we are in a position to give with somewhat greater assurance than was Wilder, an affirmative answer to the question. LIMITS OF HEREDITARY CONTROL 883 ATYPICAL VARIATION IN THE INDIVIDUAL ELEMENTS AND IN THE BANDS A. Scute ‘abnormalities’ in the species; their distribution and . frequency (1). Double scutes. In the individual scutes and in the bands certain so-called ‘abnormalities’ frequently occur, and while these may be but the expression of teratological phenomena, yet we believe that some of them at least are the result of more deeply seated factors and have a real phylogenetic significance, and con- sequently are worthy of consideration in a paper on variation and heredity. We shall speak of them as atypical variations. The atypical variation of the scutes is expressed in at least three different ways, one of the most frequent of which we shall designate as the ‘double scute.’ In this type two contiguous primary scutes are fused along their adjacent sides, and the double structure thus formed has five or six hairs at the free or posterior end (figs. 6a, 7a, 8a). Evidently there is a suppression of two or three hairs consequent upon this union. All stages of fusion have been observed in our material, and in the figures just referred to is seen a series of three, taken from one specimen showing different degrees of the process. The nature of the origin of these double scutes is made clear when the region affected is examined from the under side, where the double bony element is always clearly expressed (figs. 6b, 7b, 8b). Sometimes the bony elements are of equal size and occupy the full width of the band, but more frequently one of the plates is greatly reduced, as if there had been a tendency to crowd it out (fig. 8b). The double scutes occur in about ten per cent of the animals— or to be more exact, in a total of 516 individuals examined for this particular point 51 showed double scutes. Ordinarily there is but one of these to an animal, but four exceptions to this have been found, as follows: one with four double scutes, one with three, and two with two each. For the purpose of locating exactly the various scutes, we have chosen to begin their enumeration always on the left margin of the band. Thus scute ‘10’ of band 5 would be the tenth scute 884. H. H. NEWMAN AND J. THOMAS PATTERSON counting from the left margin of this band. Since it is obvious that the double scute is the product of what was originally two distinct elements, it has been given the value of two inour collected data, though we designate it by the number corresponding to its first or left-hand element. This point can be made clear by a reference to the numbering in figs. 6 to 10. The double scutes are generally distributed over the bands as can be seen in table 11. There is, however, a tendency for them TABLE 11 Showing distribution of ‘double scutes’ LATERAL DISTRIBUTION BANDS DISTRIBUTION IN BANDS j Left | Middle | Right RMR OS Steg: Mee hee ek. | 2.407 NAO 3: 24, 31149 4 | 3 1 71 eee ae enter? | 10, 10, 14, 15, 27, 29, 57 4 2 1 SW yer be. aes che a 12, 25, 27, 29, 41, 53, 57, 59 Cea Pe 4 ee aa Ae erie | 3, 19, 19, 23, 54, 54, 55, 60 A gigs tad Mn 4 Race MEE el oe | 2, 2, 2, 32, 44, 53, 57 Se aa 3 Ginrel ee sck see. 2h. 16; 2h, 27. OBA SbanTa os ara 18S 3 | (ea teen Pe rte 3, 12, 24, 50 2 ad) eee! 1 Becatee ts We ee ec NOS 2, 15, 36 Pg ete et! 0 ie a Pere he As 36, 65 Oy Neen 1 Pobels. im. Ak i.cess | 56 cases | 22 | 16 18 Nore, in the left-hand half of the table the numbers in the horizontal rows lying opposite the bands designate the first elements upon which the double scutes fall. Thus in the first band eight animals had these scutes, falling on ele- ments 2, 4, 7, etc., respectively. to be localized in the anterior two-thirds of the banded region, or in bands one to six. Perhaps of more importance is their lateral distribution, which can be shown by dividing each band into three equal parts—two of which represent the left and right thirds and one the middle third—and determining the number of double scutes falling within each of these general divisions. The data thus collected is shown in the right half of the table, from which it is clear that there is a tendency for these scutes to be localized toward the margins of the armor. They fall most frequently LIMITS OF HEREDITARY CONTROL 885 on scute ‘2’ (six cases) on the left, and scute ‘57’ (four cases) on the right. However, it should be pointed out that there are but four cases of exact coincidence (one involving three specimens) by which we mean.double scutes falling at the same point in the same band in two or more individuals. (2). Incomplete scutes. The second type of atypical scute vari- ation is just the opposite to that of the preceding, in that it is prob- ably the product of a splitting of what was originally a single scute, and results in the formation of an incomplete element. This type appears so rarely that not very much importance can be attached toit. In the 516 specimens examined only three cases of incomplete scutes were found. These occur as follows: Speci- men one, a female, between scutes 52 and 54 of band 6 (the small plate is counted as a whole one); specimen two, a shell, between scutes 14 and 16 of band 8; specimen three, also a shell, between scutes 56 and 58 of band 5. The first and second of these are sketched in figs. 9 and 10, respectively. The most interesting point brought out is that in each case one of the scutes lying adja- cent to the incomplete one has but three hairs instead of the typi- cal four, while the small scute has in each case but a single hair associated with it. The presence of three hairs in an adjacent scute would not alone be cogent proof that the primordium of such a scute had split to give rise to the incomplete element, as will be evident when the next section of the paper is reached. However, all of the other relations in these specimens point toward a fission process. (3). Three-hair type of scute. The third type of atypical variation is perhaps the most fundamental of all, because it in- volves a distinct change in the morphological unit of the scute. In this type the bony plate underlying the scute has but three hairs at its posterior end instead of the customary four. It was first detected in a shell which had an unusually high number of scutes in the bands; in fact the one representing the most extreme case of high numbers that we have so far found. When the shell is first observed the attention is at once attracted by the narrow- ness of the scutes, and upon examining them the interesting fact is revealed that the majority show the three-hair type (fig. 22). 886 H. H. NEWMAN AND J. THOMAS PATTERSON We have studied this shell in considerable detail and have compiled the data secured in table 12. Here it will be noted that 391 of the 625 scutes have but three hairs each, and further- more that there is a strong tendency for the 3’s to be distributed on the lateral parts of the shield. However, the most striking fact revealed in the table is indicated in the fourth column, where the total number of hairs for each band is given. Bands 2 and 8, which show the extremes in scute variation, have exactly the same number of hairs, while the two extremes in hair variation, bands 7 and 9, are but six hairs apart in their totals. TABLE 12 Shell 2, with 625 scutes showing distribution of 3-hair and 4-hair scutes DISTRIBUTION OF 3’5 AND 4’S IN BANDS | LATERAL DISTRIBUTION OF 3’ BANDS 3-Hair Type | 4-Hair Type | Number of | ‘Left | Middle Right ee = —— | a | lense Ae ae 49 | 22 | 235 | 19 16 14 Din aS Tee 34 | 33 | oA | la 12 8 Sy ie 37 | 31 | 225 Pas 5 14 A | 38 | 30 | ode) Ul Teac Bias 14 aks 38 30 | DOA Wy 19's Volt ee 13 Be 49 op Dare Tala Ooh Sail) grat 22 7 SR u SU ion bd 41 27 Fe os 7 16 Be ea 58 15 23 +08) JO1e Veet 21 Dia tae a ral 47 7 Oe ee mate er eh: 16 otal 20a. alee | 2109 | 176 77 | 138 The question naturally arises: Can the high number of scutes exhibited by this specimen be accounted for by the fact that so many of them are of the three-hair type? Undoubtedly it can, but the affirmative answer necessitates the assumption that the integumentary primordium out of which each band arises contains a definite number of hair follicles, which are later distributed into groups of 3’s and 4’s according to the propensities of the formative scutes. Itis evident upon this assumption that a ten- dency for the scutes toform about groupsof four hairs would result in the production of a shell having a low number of scutes, while LIMITS OF HEREDITARY CONTROL 887 a tendency to form about groups of three would produce just the opposite result. A reference to table 13,in which the total counts on five specimens are recorded, indicates that, although the ani- mals may vary considerably in the total number of scutes, yet TABLE 13 Showing proportion of 3-hair and 4-hair scutes in a small sample of the species | NUMBER | NUMBER SPECIMEN | BE Oe cea | pane | rome | LEFT MIDDLE RIGHT Shell 2...... | 625 391 234 | 2109 AB 77 138 2) ie | 577 15Qe 418) 2149 70 29 60 DOI a8 ss | 545 7 AGS 2103 33 9 35 2 7 ee | 534 19 FGF Me PO) fi 2 10 9204 2 ve. | 529 26 503 | 2091 13 Os ee there is very little difference in their total hair counts. In the two extreme cases, shell 2 and female 204, there is a difference of 96 scutes, but in total hair counts they vary by only 18 points, which is a very inconsiderable difference when one considers that there are between three and four times as many hairs as scutes. Undoubtedly data on a larger number of animals would reveal a much greater variation; but the fact remains that, however variable the species may be with respect to the scutes (and their associated plates), it is comparatively stable with respect to the hairs of a given region. A statistical study of the hairs might be made, but their enumeration is difficult; for, on the one hand, not a few of the adults possess armors from which many of the hairs have been lost by abrasions through contact with rocky dens, and on the other hand, even advanced foetuses do not have hairs sufficiently developed to allow a census of them being taken. For this and other reasons it has seemed best to confine our studies to the variability and heredity of the scutes, and not to attempt a compilation of statistics on hair counts, which at best could be only imperfect. It is perhaps sufficient then merely to have indicated the line along which the evolution of the species is apparently directed. The fact that in all of the regions showing a primitive condition of the scutes, as in the case of the belly, JOURNAL OF MORPHOLOGY, VOL. 22, No. 4 888 H. H. NEWMAN AND J. THOMAS PATTERSON the hair groups are larger than in the more highly differentiated region like that of the bands, is prima facie evidence that here the direction of evolution is from the four-hair to the three-hair type, with a consequent increase in the total number of scutes. These high numbered specimens with their many scutes of the three-hair type are of further interest because they are probably to be regarded as mutations, in the sense that they represent dis- continuous variations. This in part can be made clear by an inspection of table 1. At the lower end of the series therein rep- resented the specimens occur with a frequency sufficient to be explained on the basis of an ordinary fluctuating variability, and this is also true of the other end of the series up to the specimen with 596 scutes; but beyond this point specimens are of rare occur- rence and consequently are separated by wide gaps. However, neither the rarity of their occurrence nor their wide separation is sufficient to place such specimens in the category of mutations, because there is always the possibility that a larger number of the species would furnish variates enough to fill up the gaps. If we would therefore explain these as saltations we must look to another source of information, namely, to their behavior in inherit- ance. This topic will receive attention in a succeeding section. B. Hereditary control in connection with scute ‘abnormalities’ (1). Double scutes. Attention has been called to the com- parative rarity of double scutes in the species and to the fact that in only a very few cases do double elements fall in the same place inthesame band. Itisrarely the case even that the same general region of a band of more than two specimens shows the peculiarity in question. When, therefore, we find in as many as three out of four members of a set of foetuses a double scute in almost precisely the same locality, we are forced to the conviction that even these elements, which might be defined in the words of Galton as ‘the minutest biological units,’ are predetermined in the undi- vided germ cell. Two sets of foetuses show the conditions most clearly and they are described in detail below: LIMITS OF HEREDITARY CONTROL 889 Set 121 (Mother without any double scutes) Foetus 1. No double scutes Foetus 1. In band 2, scute 29, double Foetus 111. In band 2, scute 28, double Foetus tv. In band 3, scute 28, double It is to be noted that in three of the four individuals the double element occurs practically in the same spot. In view of the fact that, even in perfectly normal sets, the number of scutes in the same row varies widely among the individuals, it would hardly be expected that a predetermined doubling could strike the same numbered scute in any two members of a set. The regional hereditary control in this case is therefore somewhat more accu- rate than we would expect to find it, in viewof the laxity of hered- itary control in the matter of the arrangement of scutes into bands. The fact that in one of the foetuses the double scute occurs on the third band, while in two other foetuses the doubling affects the same or contiguous scute of the second band, only goes to bear out the idea that the process of scute alignment is to a large extent a mechanical process, involving the shifting backward or forward of the individual primordia under the influ- ence of pressures. It is only to be expected then, in the light of these considerations, that we should find a case now and then where a readily recognizable element like a double scute should indicate by its position that it may have been shifted from one row to another. On account of the small extent of its ‘abnormalities’ this set has been used in the study of correlation (table 6, A) as well as for an example of a peculiar type of band arrangement, (p. 874). Set 128 (Mother without atypical scutes of any kind) Foetus 1. In the last row of the scapular shield, scute 30 is double; scute 2 of band 2, double; scute 10 of band 3, double Foetus 11. In the last scapular row scute 28, double; an incomplete scute like that shown in fig. 9, occurs in band 6 between scutes 2 and 3 Foetus. Inthe last scapular band scute 14, double; scute 49 of band 6, double Foetusitv. Scute 31 of band 4, double 890 H. H. NEWMAN AND J. THOMAS PATTERSON In this set the following points are noteworthy: 1. In none of the specimens examined in the statistical study of the frequency and distribution of double scutes was a case of doubling found in the last scapular row; hence the likelihood of the conditions described being due to coincidence is extremely remote. 2. The location of the double scute is evidently not so rigidly defined as in the former case, since the double scute in foetus 11z is rather far separated in position from that of either of the other foetuses. In the two individuals of the natural pair A, of the left hand side (right in the figure), however, the position of the double scute is almost precisely the same. One can detect the difference in position only by counting the scutes. Careful examination of the two right hand individuals in fig. 24 will make this clear. In the subsequent discussion of the phenomenon of pairing this circumstance will receive further attention. 3. It is very unusual, as was indicated in the general discussion of scute ‘abnormalities,’ for an individual to have more than one double scute. When, therefore, three of the four quadruplets have two or more double scutes and the other has one we are inclined to suspect that they are all predetermined. 4. In this and the last set the ‘three-to-one’ proportion is shown in several ways: (a) In set 121 three show a double scute and” one lacks it; (b) in set 123 three have the scapular double scute and one lacks it; (c) in the same set one of the four has three double scutes, one has only one double scute, and one of the four has an incomplete scute in place of a double scute. Many other cases of a similar kind are noted in connection with both normal and atypical sets. The following points although possibly of no real significance should be noted: (a) In foetus tv, the adjacent of foetus 1, occurs a double scute in the same position in its band as that which occurs in the scapular region of foetus 1, but four bands posterior to the latter. (b) In foetus 11 occurs an incomplete scute occupying the same position in its band as does one of the double scutes of foetus 1, but situated four bands posterior to the latter. (c) Tn foetus 111 a double scute of the sixth band is in position almost LIMITS OF HEREDITARY CONTROL 891 exactly the ‘mirror image’ of the double scute in the third band of foetus 1. These facts may indicate various degrees of imperfect heredi- tary control in connection with the localization of these minute ‘abnormalities,’ but if this be the case we are dealing with a state of affairs too complex to admit of solution with the present material. There are several other sets which exhibit scute ‘abnormalities.’ These may for convenience be brought together into a compact table (14). Here the number of the double or incomplete ele- ments in the various columns refers to the position of each in the band; ‘D’ indicates double scute and ‘S’ incomplete or split scute: Although in the members of these sets the double or incomplete elements would appear to occur almost at random, so far as their position within the nine bands is concerned, there are some cases in which, within a given pair at least, the hereditary control is fairly obvious. Table 14 further indicates that a close morpho- logical relationship exists between the two types of ‘abnormality’ and that they may readily be imagined to have a common hered- TABLE 14 Showing the distribution of double and split scutes in the 20 sets of foetuses BAND SET FOETUS | im re eae 7 mn a3 1 2 3 4 5 6 7 8 9 Tis prey eee 5D | | | sh ante } Tiles | 58D | | Eyer es | yee oe emer ie | iIneeion wees | | 46D | I | 9D | | | DIGE ect II | | 25D | 1100 ee A he | 9D | ja entree eee | 35 a eae Eeeaeren ey | | | | 13D I ae 25D | 54D | | Be ode FE matt | 60D [Us Witieenec vce a | | | GOS sistas ate TIE... ee. | 11D | 892 H. H. NEWMAN AND J. THOMAS PATTERSON itary basis. It would seem to be profitless to attempt any detailed analysis of the conditions in the sets tabulated, but there are some points worth study that have not been mentioned. Every case in which two or more individuals show a scute ‘abnormality’ is almost undoubtedly a case involving some phase of hereditary control, for on the basis of chance there would be very few sets more than one individual of which would have a double scute. In concluding this account of the hereditary control of double scutes attention might be called to the fact that in this connection we have excellent material for testing Galton’s ‘‘minutest bio- logical unit capable of hereditary transmission.’? Whether or not these minute pecularities of the scutes are directly inherited from the immediate parents we cannot at present determine, but we are certain that the conditions are blastogenic in the sense that they are predetermined in the fertilized egg. If this be the equiv- alent of hereditary transmission we may rightly claim to have found just the sort of unit that Galton was looking for. It seems unlikely that any smaller unit could be predetermined. (2). Three-hair type of scute. In only one set of foetuses (set 135) have we found that peculiar condition described as a possible discontinuous variation or mutation. Although we are unable to count the hair primordia in the foetuses, we are confident that the great majority of them are of the three-hair type. This assumption is justified by the fact that the scutes have the same narrow appearance that is seen in adults in which the three-hair type prevails, and by the additional circumstance that we have never found an individual with over 600 scutes in which the three- hair type of scute did not largely predominate. As an additional piece of evidence in favor of the idea that the condition in question is a mutation, it seems highly probable that it is inherited in the exclusive fashion and is dominant. If we suppose that the foe- tuses show a blend between the mother, an individual with 573 scutes, and the unknown father, it would necessitate the positing of a male parent with a scute number far in excess of any we have found to occur in the species. That the lower scute numbers, which are made up from shells showing comparatively few three- LIMITS OF HEREDITARY CONTROL 893 hair scutes, are inherited in the blended fashion is shown by a comparison of the counts of the mothers and those of the foetuses. In nearly every case the latter show marked divergences from the former, which would seem to indicate a total lack of exclusiveness about the inheritance. Attention might also be called to the fact that in this set three individuals have almost identically the same number of scutes while the fourth has decidedly more. C. Band ‘abnormalities’ in the species; their distribution and frequency The atypical variations in the bands consist primarily of extra or supernumerary bands or parts of bands, and show a rather marked degree of regularity in that they occur repeatedly in about the same regions of the armor. They are found most frequently in the first or second band, or in both, and may occasionally be seen in the region of the eighth or ninth band, but rarely appear in bands three to seven. Their frequency has been determined from a study of 1768 specimens, in which 60 are abnormal, or about 3.4 per cent of the total number of individuals examined. On account of the apparent nature of their origin, as well as for convenience in description, we have chosen to consider these ‘abnormalities’ under three headings, as follows: (1) Fusions, in which parts of two bands have united to form a single structure; (2) Splittings, in which a series of elements of a band have divided transversely to produce a double row; (3) Additions, in which a band or part of band, either from the scapular or pelvic shield, has been added to the thoraco-lumbar shield, thereby increasing the banded region, and in case of the contribution or an entire band, producing a ten-banded animal. (1). Fusions. This kind of variation is found in 31 of the 60 ‘abnormal’ specimens, or in more than fifty per cent of the cases. It usually expresses itself in one of two ways. In one type two adjacent bands have their scutes at one side of the armor united to form a single structure. This type may be spoken of as ‘uni- lateral,’ in contrast to the second or ‘bilateral’ type, in which the 894. H. H. NEWMAN AND J. THOMAS PATTERSON fusion of the two bands occurs on both sides of the armor. A very good example of this latter type is shown in fig. 11 in which seven scutes are involved on the left side and four on the right. The unilateral type occurs 25 times in the 31 cases of fusion, and appears 15 times on the left side and 10 on the right. It is distributed as follows among the several bands: 18 times between bands 1 and 2; 3 times between bands 8 and 9; and once each between bands 2 and 3, 4 and 5, 5 and 6, 6 and 7. In almost every case of unilateral fusion the scutes at the extreme lateral margins of the bands are involved, and in only a few specimens is it situated well within the margin (e.g., in fig. 18). The six remaining cases of fusion are all of the bilateral type, and in each instance the two fusions are practically symmetrical, both as to position and extent. This is true even when they are situated toward the median line of the armor. All the bilateral fusions observed are between bands one and two. (2). Splittings. The second kind of atypical band variation is the splitting, which as already stated consists of a transverse division of the elements to give rise to two bands. It is not always an easy matter to determine, especially in complex cases where two bands seem to be greatly confused, whether we are deal- ing with a fusion or a splitting; but a rather safe, though not uni- versal, rule to follow is to count the elements in each row of the double regions, and if these be equal in numbers it is safe to decide that one is dealing with a splitting (fig. 12). Using this as our principal method of determination we have classified sixteen specimens as belonging to this group. Five of these are unilateral (three on the left and two on the right) and eleven bilateral. One of the most striking features of the latter type is the strong tendency to be almost bilaterally symmetrical. For example in fig. 12 is shown one in which the splitting begins in the seventh scute on each side, and the specimen fails to show symmetry in its bilaterality only in having the splitting slightly less extensive on the right than on the left side. Another point worthy of note is the fact that splittings are confined almost en- tirely to the first band; thirteen of the fifteen cases occurring here. LIMITS OF HEREDITARY CONTROL 895 The fourteenth and fifteenth cases appear in the sixth and ninth bands, respectively. Several of the cases of splitting are rather complicated, but not to such an extent as to render their classification uncertain. One of these is shown in fig. 13, and it will be noted that while the ‘abnormality’ is clearly of the bilateral type, yet it is slightly com- plicated by having an additional splitting just to the right of the left-hand split. (3). Additions. The final class of band variations is additions, by which is meant the adding of a band or part of band from either of the shields lying adjacent to the banded region. Six cases of this type have been observed, and five of these are from the seapular shield. An excellent example of this type is seen in fig. 19. The right half of the posterior row of scutes has dropped down from the shield and forms a perfect band on this side of the armor. One can follow the scutes of the added half-band across to the left side where they clearly form the posterior row of the scapular shield, so no question can exist regarding the origin of such ‘abnormalities.’ It is easy to imagine how a completion of the dropping would resuit in adding an entire new band to the banded region, and the four or five cases of ten-banded animals that we have observed may have acquired their extra band in this manner. However, it is possible that additions may take place from the pelvic shield; in fact, our sixth case of additions has probably come to exist in this way, because the pelvic shield of this particular specimen is foreshortened. It should be noted also that a ten-banded animal could be produced by a complete splitting of one of the nine regu- lar bands. In four of the seven cases of band variations, not classified in the above groups, not sufficient data were taken to permit an . exact determination of their characters; they are merely recorded as being ‘abnormal.’ Two of the remaining three are almost exactly the same, except that one is in the first band and the other intheninth. The latteris sketched in fig. 14. The ‘abnormality’ is located approximately in the middle of the band, and consists of a region of four double scutes. At first one would be inclined 896 H. H. NEWMAN AND J. THOMAS PATTERSON to suggest that it had been produced by a splitting of four scutes, but upon a closer inspection it will be seen that the upper row is closely associated with the left half of the band, and the lower row with the right half. In the light of the ‘theory of concres- cence’ one is tempted to suggest that it has been brought about through a failure of the embryonic primordium to affect a perfect meeting in this region, and consequently the inner ends of the two half-bands have slipped past each other for a short distance. In this connection attention should be called to a certain rather rare atypical condition which sometimes appears either in the ninth band or in the first row of the pelvic shield. The peculi- arity consists of a ‘jog’ at the middle point of the band or row, and in all probability has come about in a manner like that sug- gested just above. A specimen showing this condition in the first row of the pelvic shield is seen in fig. 23. The final ‘abnormality’ to be considered is shown in fig. 15. It really is a double peculiarity, in that both the eighth and ninth bands are involved. In the eighth band there are three small primary scutes lying just anterior to numbers 53-56 of the main scutes of the band. These small scutes do not affect the bony plates and must therefore be looked upon as very rudimentary in character. In the ninth band a somewhat similar ‘abnor- mality’ is seen, lying at scutes 52-56 of this band. It consists of four small scutes, which however affect the bony plates, as can be seen in the sketch of the under surface of this region of the band (fig. 15 6). It cannot be said to be a splitting because there are five plates in the lower or main row and only four above. In concluding this section of the paper two facts brought out in the foregoing account should receive especial attention, because of their direct bearing on what is to follow. (1) In all those cases of atypical variations that we have designated as bilateral there is a very strong tendency toward symmetry in the affected. regions both in position and extent. This is particularly clear in specimens classed as splittings, as the citation of a few cases will make evident: (a) In shell no. 8 band 1 has 57 scutes, and the two splittings occur in scutes 6-14 on the left and 44-51 on the right—were the right-hand split in scutes 44-52 instead of LIMITS OF HEREDITARY CONTROL 897 44-51, it would be a perfect example of bilateral symmetry; (b) in shell no. 21, in which band 1 has 59 scutes, the splits are in scutes 8-13 and 46-52—here again a change of one point on the right side would make a bilaterally symmetrical ‘abnormality’; and (c) in shell no. 13 there are 63 scutes in band 1, and the splits are in scutes 4-16 and 48-59— a close approximation to sym- metry. These conditions remind one of Wilder’s results in dupli- cate twins, already referred to in the introduction. (2) The second point to be emphasized is this, that although the atypical variations in the bands show a marked degree of regularity in occurring so frequently in the first and second bands, yet they display a great diversity in the fact that scarcely any two of them are exactly alike. This diversity for the species is striking when considered in the light of the fact that in 1768 specimens examined but two sets of coincidences have been found, and these occur in the simplest type of ‘abnormality.’ Since many of the shells upon which these data were taken are from the same locality it is not improbable that some of these cases are either brothers or sisters, and not coincidences. The fact that only two sets of these supposed cases of coincidences, and these of the simplest type (very slight fusions), have been found in 1768 individuals will serve to emphasize the importance of fraternal correlation as brought out in a subsequent section of the paper. D. Hereditary control in connection with band ‘abnormalities’ Considering the comparative rarity of band ‘abnormalities’ in the species we have been fortunate to secure a collection of sets of foetuses, among which appear examples of practically all of the types of malformation described above. As in the case of meristic variation in the normal scutes and in the matter of double scutes the precision of hereditary control is much more marked in some cases than in others; in some the conditions are fairly simple and in others highly complex. The sets are described in detail and discussed separately, in so far as the special condi- tions of each case are involved. The general significance of many of the observations can be discussed to advantage only after all 898 H. H. NEWMAN AND J. THOMAS: PATTERSON the data has been presented and the underlying problems pre- sented for discussion. Set 64. (Mother normal). Foetus 1 appears at first sight to be perfectly normal, in so far as band arrangements are concerned; but examination reveals the presence of ten full bands. In view of the fact that all of the other members of the set show more or less extensive regions of splitting in the first band we are forced to conclude that the extra band in this individual has been pro- duced by a process of band splitting carried to a completion. Fig. 3 Diagrams of the ‘abnormal’ bands in the four foetuses from female no. 64. In this as in the two succeeding figures, the Roman numerals I-1v refer to the individual embryos from which the sketches were made, while the arabic indicate the number of scutes in each of the regions to which they are adjacent. This is really only a bilateral expression of the condition seen in the right hand half of the first band of foetuses m1 anditv. That there are 64 scutes in the first half-band and only 62 in the second might seem to militate against the idea of splitting, but there occur in our collection several undoubted cases of splitting where the number of scutes in the two series produeed by the split are unequal. The split first band of foetus 1 is shown in fig. 3, I. LIMITS OF HEREDITARY CONTROL 899 Foetus 1 shows a condition entirely different from that de- seribed for its partner, namely a bilaterally symmetrical, regional splitting of the first band, beginning four scutes from the margin on each side and involving 15 double rows of scutes in each case. The condition is diagrammatically shown in fig. 3, 1. Foetus 11 exhibits a decidedly asymmetrical regional splitting of the first band, being a mixture of the two pure types shown in foetusestand iz The left half of the band is identical with that of foetus 11, while the right half is the duplicate of that in foetus 1 (hie 3, Er): Foetus Iv is identical with foetus 111, in so far as the splitting is concerned, but shows in addition to the latter a fusion of the anterior row of the completely split half with the opposite half of the last scapular row of scutes. This condition has been ob- served in the shells of several adults and has been interpreted as a case of the incipient addition of a band to the banded region by means of a ‘drop-down’ from the scapular shield. The present condition could hardly be interpreted in that way in view of our knowledge of the conditions seen in the other members of the set. It would appear that we have here a case of an epigenetic process, involving a secondary fusion of two unrelated half bands, the right half of the first band and the left half of the last seapular row (fig. 3, Iv). This is in some ways the most extraordinary set in the collection and suggests a number of theoretical considerations for general discussion. The main points that should be noted while the facts are fresh in mind are as follows: 1. The splitting process involving the first band occurs in all members of the set, which would indicate that this much was predetermined. 2. The regional splitting involves in all four cases (twice in foetus 11) exactly the same number of scutes, 15 in each ease; and these are located precisely the same distance from the margin every time. The precision of hereditary control is, in this regard, nothing short of marvelous, since it is perfect. 3. There are evidently two distinct expressions of the splitting tendency, the complete and the incomplete. The distribution of 900 H. H. NEWMAN AND J. THOMAS: PATTERSON the two types is quite impartial in so far as their frequency is concerned, for, out of the possible eight lateral halves of the four foetuses, four are occupied by each type. It would appear from this that each was equally strongly predetermined; but the exer- cise of hereditary control in the distribution of the two types among the four foetuses is somewhat haphazard and reminds one strongly of a pure chance combination of two elements selected in pairs, like, for example, the Mendelian ratio of F 2’s, D-2Dr-r. 4, Apart from the ‘secondary’ fusion of one of the split half bands with the scapular shield, the members of the natural pair B (11 and tv) are strikingly identical. Evidently hereditary con- trol within the pair is more accurate than in the whole set. The explanation of this condition must be discussed in the subsequent section on pairing. Set 96 (Mother normal). This case is somewhat simpler than - the last in that it involves only one type of ‘abnormality,’ namely a simple fusion of the first two bands. We have decided to call the condition a fusion, because, including the two which are united, there are only nine bands. ‘The strictly marginal charac- ter of the band unions would also serve to indicate a fusion, for we have found no cases of incipient marginal splitting in our exam- ination of adult shells. Foetus 1 shows a unilateral fusion of comparatively slight extent, involving only 5 scutes and confined to the right side. There are 57 free scutes in band 1 and 58 in band 2 (fig. 4, 1). Foetus 11 shows a bilateral fusion of small extent, involving 7 scutes on the right and 4 on the left. There are 51 free scutes in band 1 and 49 in band 2 (fig. 4, 11). Foetus 11 shows an extensive unilateral fusion, involving 21 scutes and confined to the right side. ‘There are 40 free scutes in band 1 and 39 in band 2 (fig. 4, m1). Foetus Iv shows the same condition as does foetus III except that there are 19 fused scutes instead of 21. There are 41 free scutes in each of the first two bands (fig. 4, Iv). The following points may be noted: 1. The fusion in the first two bands occurs in all individuals but differs in its extent and in the unilaterality or bilaterality of its LIMITS OF HEREDITARY CONTROL 901 expression. Evidently a marginal fusion of bands 1 and 2 is predetermined, but here the predetermining influence ceases to operate and epigenetic factors of some sort determine whether the character shall have a unilateral or a bilateral expression and to what extent the fusion is to be carried in each case. It will be noted that hereditary control is much less precise in this case than in set 64. 2. The ‘three-to-one’ proportion once more presents itself in that three members of the set have the fusion on the right side only, while one has it on both sides. Fig. 4 Diagrams of the ‘abnormal’ bands of the four foetuses from female no. 96. 3. There is a distinct pairing on the basis of the extent of the fusion, pair A (1 and 1) showing the character in a much less extensive form than pair B (ar and Iv). Within the pairs the resemblance is very close. Set 101 (Mother normal). The conditions seen here are in many respects equivalent to those described for set 64. All four foetuses show a rather advanced stage of splitting in the first band. At first we thought that foetuses 11 and Iv were quite 902 H. H. NEWMAN AND J. THOMAS PATTERSON normal, but further examination reveals our mistake, for each of these two individuals possesses ten well defined bands. The condition seen in one of the foetuses is so obviously a splitting that we are compelled to interpret the extra band in each of these foetuses as having been produced by a complete splitting of the first band. Foetus 1 shows an extensive bilaterally symmetrical and some- what complex regional splitting of the first band. . On both sides, beginning 6 scutes from the margin, there occurs a splitting involy- ing 9 scutes, and in the central part of the band there is another TTT COCA Pete 17 Fig. 5 Diagrams of the ‘abnormal’ bands of foetuses t and 11 from female no. 101. split region of 17 scutes. Separating the median from the two lateral splittings are paired unsplit regions, each involving 8 scutes. In the ninth band there is another band peculiarity, consisting of a ‘jog’ in the middle of the band similar to that-sshown in fig. 23. There are 34 scutes to the left of the break and 31 to the right. The region involving the fusion is illustrated dia- grammatically in fig. 5, 1. Foetus 11 shows a more advanced stage of the splitting in that the process has gone on to completion on the right side and has extended to within 6 scutes of the margin on the left side (fig. 5, 11). Asin foetus 1, there occurs here also a ‘ jog’ in the ninth band, which is approximately the ‘mirror image’ of that in the partner individual, in that there are 34 scutes on the right of the break and 30 on the left. LIMITS OF HEREDITARY CONTROL 903 Foetuses 1 and tv both show a completed splitting of the first band into two, thus producing ten bands; and they also lack the irregularity in the ninth band. Attention is called to the following points: 1. All four members of the set show an extensive splitting of the first band. The splitting was predetermined in so far as the location within a certain band is concerned, but the extent or manner of its expression appear to have been beyond the limits of hereditary control. 2. There was evidently a fairly rigid hereditary control in the matter of the various limits of the incomplete splittings, as one may judge by the facts that the two sides of foetus I are exactly identical and that six scutes constitute the marginal unsplit' region in every case where the splitting is incomplete. This would indicate a precision of hereditary control almost as remarkable as that noted in set 64. 3. There is a pairing of foetuses on the basis of the extent of the splitting, pair A (1 and 1) showing an incomplete splitting, and pair B (im and Iv) a complete one. Pairing is again evident in the matter of the irregularity in the ninth band, pair A (1 and 11) showing it, pair B (ar and tv) lacking it. 4. The ‘three-to-one’ proportion appears again in that three individuals are bilaterally symmetrical and one is unilateral in the expression of the splitting process. Set 118 (Mother normal). This is the only set in our collec- tion where only one member of a set shows a band ‘abnormality.’ Foetuses I, 11 and Iv are normal. Foetus m1 shows a fusion of small extent between the first two bands, involving only three scutes at the left hand margin. The extent of the ‘abnormality’ is so slight that it seems scarcely worth while to discuss the possible bearings of thecondition. It would appear best to consider the case as one of incipient fusion, which may have arisen epigenetically in one of the four individ- uals. In the next generation it might have been inherited in a more pronounced form. There is, however, the alternative explanation which is developed in the discussion of the theoretical causes of pairing. JOURNAL OF MORPHOLOGY, VOL. 22, NO. 4 904 H. H. NEWMAN AND J. THOMAS PATTERSON Set 121 (Mother abnormal). This is one of two observed cases of the direct inheritance of an atypical condition from the mother. In this case the mother had a right marginal fusion of the last two bands, involving 4 scutes. Foetus I is normal. Foetus 11 shows a ‘jog’ in the ninth band similar to that shown in fig. 23 and in set 101. There are 33 scutes on each side of the break and hence the latter is median. Foetus 111 is normal. Foetus Iv shows a ‘jog’ in the ninth band just like that in foetus 11 except that there are 33 scutes on the right and 35 on the left. The ‘abnormality’ in the mother is entirely different in char- acter from that shown by the two foetuses, but in that it is uni- lateral and involves the ninth band in both mother and foetuses, there would seem to be ground for believing that there is some genetic connection between the peculiarities seen in the two generations. It is quite possible that the parental and filial conditions may have no hereditary connection, but there must nevertheless be some predetermining mechanism controlling the occurrence of the same peculiarity in two of the members of a set. It should be noted that the two diagonally placed foetuses seem to be paired with respect to the band irregularity, and that the same individuals were noted as paired on the basis of scute counts. This condition strengthens the hypothesis that occasion- ally there may occur such a shifting of the blastomeres as to bring about a false pairing in the arrangement of the quadruplets with- out interfering with their inherited potentialities. For a more detailed discussion of this situation reference is made to one of our former papers on this subject.! Set 123. Mother shows a deep crease in the scapular shield, between the ninth and tenth scapular rows (counting from the posterior margin of the shield). This crease is apparently due to the suppression of part of a row of scutes. All four foetuses show the same peculiar crease in precisely the same region. The photograph (fig. 24) will serve to emphasize 1 Newman and Patterson, 1910. This Journal, vol. 21, no. 3, p. 401. LIMITS OF HEREDITARY CONTROL 905 the striking identity of the four quadruplets with respect to the peculiarity on question. This is the only undoubted case of the direct transmission of a definite seute peculiarity from the mother to the offspring which we have encountered. In this case it seems clear that hereditary control and hereditary transmission are equally precise, and it appears probable that most of the cases we have described are inherited either from the unknown fathers or from some recent ancestor. PAIRING; AN INTRA-FRATERNAL CORRELATION; AND ITS BEARING ON THE PROBLEMS OF HEREDITARY CONTROL The most difficult question that confronts us in attempting to gain an insight into the operation of the predetermining mech- anism concerns not so much the resemblances between the quad- ruplets as their differences. Why are they not all exactly alike and why are there closer resemblances between some individuals of a set than between others? It has been shown that, with respect to their uterine connections, the four foetuses are definitely paired, in that two of them are attached to each of the lateral placental discs of the chorionic vesicle. Embryological investigations have revealed indications of a much earlier and more intimate pairing. With very few exceptions the resemblances between the foetuses are strictly in accord with their placental pairing, the closest resemblances occurring between the two individuals attached to the same placental disc. This is not a mere matter of arbitrary judgment, but is based on a comparison of the inter-pair and intra-pair correlation. By the use of the difference method previously employed the two constants have been determined as follows: Inter-pair coefficient of correlation = 0.9257 Intra-pair coefficient of correlation = 0.9517 It will be noted that the correlation between the individuals of opposite pairs (inter-pair correlation) is lower than that deter- mined for the whole set (.9348), while the correlation between the individuals of the same pairs is decidedly higher; in fact this is probably the highest correlation constant determined for any 906 H. H. NEWMAN AND J. THOMAS PATTERSON organic relation. How surprisingly alike are the members of the pair may be realized by an examination of table 6. A glance down the column of totals will reveal the frequency with which exact identity exists between the paired individuals and how rarely such is the case between individuals of opposite pairs. It might be claimed that any method of pairing would give a closer correlation than that derived from a treatment of the whole set. That this claim is without justification is readily shown by the experiment of pairing the foetuses in different ways and deter- mining their correlation constants. There are two other possible ways of pairing the quadruplets. We may put together the adja- cent members of opposite pairs, associating foetus I with Iv and II with 111, or we may pair them diagonally, associating 1 with 11 and 1 withtv. The result of the experiment is shown below: Correlation coefficient, true pairs (I-11) and (111-1v) = 0.9517 Correlation coefficient, false pairs (I-1v) and (11-111) = 0.9270 Correlation coefficient, false pairs (1-111) and (11-rv) = 0.93894 It will be noted that the correlation constants for both of the false pairs differ only to a slight extent from that of the whole set; which shows conclusively that there is no intra-fraternal correlation on the basis of either of these artificial groupings. The pairing relation is brought out even more strongly in the study of atypical scute and band conditions. In each set dealt with reference has been made to the instances of pairing as they occurred, and in several cases it was shown that pairing was exhib- ited in several different ways in the same set. No more convinc- ing evidence of the reality of pairing could be asked for than is afforded by the conditions seen in sets 64, 96 and 101. The fact of pairing would seem then to need no further dem- onstration; but its underlying causes and its relation to the mechanics of hereditary control are problems that require much consideration. It must be frankly admitted that so far we have failed to prove conclusively that each foetus in a set is the product of a single blastomere of the four-cell stage. The youngest embryonic vesicles appear to be still single individuals so far as any visible LIMITS OF HEREDITARY CONTROL 907 demarkation ‘of embryonic primordia is concerned; but the absence of a visible line of separation between the quadrants of the vesicle does not preclude the possibility that each quadrant may be formed exclusively, or nearly so, of cells derived from one of the first four blastomeres. It seems very likely, in fact, that the cleavage products of each blastomere would continue to occupy the relative position held originally by the parent cell, in spite of the various complex developmental processes that ensue, just as the cells derived from the first two blastomeres, in many organ- isms, retain their bilateral positions and go to form the right and left sides of the individual. To this extent then we are probably justified in tracing back each of the quadruplets to one of the first four blastomeres. If this be granted it results logically that each pair must be the product of one of the first two blastomeres. On this basis alone are we able to offer any reasonable explana- tion of the observed phenomenon of pairing or to attempt an answer to the question: Why should there be a closer resemblance between paired than non-paired individuals? Our original explanation of the condition naively assumed that the first cleav- age would divide the fertilized egg into two somewhat unequal parts, and that the second cleavage would divide these half eggs into quarters more nearly equal than were the halves. In brief we assumed that the first cleavage gave products more variable than the second. Is there any basis for such an assumption? Is there, as development proceeds, a progressive decrease in the variability, real or potential, of daughter cells? It has been dis- covered from a study of the intra-individual variability of certain plants, notably Ceratophyllum, (Pearl, ’10), in which whorls of leaves are successively produced by-the apical bud or growing point, that the leaves of the first whorl are the most variable, least closely correlated, and that later ones vary less in an orderly progression. Similarly, if we consider that in the first cleavage the armadillo ovum divides itself into two potential individuals and that each of these in turn divides into two more individuals, we would expect to find a decreasing variability among the like parts produced at each successive division. In this case, however, the production of quadruplets destroys the twins and we can dis- 908 H. H. NEWMAN AND J. THOMAS PATTERSON cover the variability of the products of the first division only by determining the inter-pair correlation in our sets of quadruplets. This constant should be lower than that determined between the individuals of the several pairs, the intra-pair correlation. This is exactly what we have done, and the results, as given, are in accord with the hypothesis. An interesting test of the validity of this explanation of pairing might be made in connection with the foetuses of Tatu hybridum, one of the South American arma- dillos, which has most commonly eight or more polyembryonic foetuses. Here the production of individuals has gone at least one step further than in our species, and it should be possible to find out whether the products of the last division are more closely correlated than are the paired individuals in our species or than the products of the previous division. It is to be hoped that some investigator to whom the material is available may see fit to satisfy our curiosity on this point. If our hypothesis should prove to be well founded we shall have furnished an interesting illustration of the law of decreasing variability with the production of like parts, which is probably a corollary of the more general law that variability decreases progressively with advancing development, a law made clear by Vernon in his book on ‘‘ Variation in animals and plants.” What are the logical consequences of accepting such an explan- ation of pairing, and what light may the ideas expressed throw on the problem of the mechanics of hereditary control? Two courses are apparently open. We may consider the fertilized armadillo egg as heterogeneous in structure, so that its four quad- rants, which occupy the positions of the four blastomeres, bear the different materials which determine the differences between the four foetuses. On this basis it would be difficult to explain the paired relation, and still more difficult to accept the necessary consequences of the assumption that, where certain atypical conditions occur in all four foetuses, the physical basis of the ‘abnormality’ must have been repeated four times over in as many quadrants of the egg. It would involve too severe a strain on one’s credulity to ask him to believe that in sets 121 and 123, for ~ LIMITS OF HEREDITARY CONTROL 909 instance, the double scute primordium occurred separately in three of the quadrants of the uncleaved oosperm. The alternative hypothesis involves the assumption that the differences are due to the lack of complete accuracy in the bilateral distribution of the hereditary materials (probably chromosomes). During the process of cleavage by means of the mechanism of hereditary transmission, whose visible operations are probably intimately associated with the mitotic figure, the various materials which condition the development of the definitive structures are distributed more or less unequally to the two daughter cells. The next cleavage involves another unequal distribution of mate- rials, but the inequality is lessened; and presumably, as the cleav- age process continues, either the distributing mechanism becomes more exact through practice or else the material distributed becomes progressively more homogeneous, and hence less apt to produce variability. The Mendelian-like ratios which we have noted so often may be no more than fancied parallelisms, but we have been unable to dismiss the phenomenon without some speculation as to its possible significance. The true Mendelian ratio of ‘three-to- one’ is the result of the segregation of grand-parental characters, a fact which has suggested the thought that we may have here a condition involving in some way, which we are at present unable to understand, the interaction of dominant and recessive ancestral characters. The only alternative explanation of the presence of a character in three individuals of a set and its absence in the fourth involves the assumption of latency, whose aid we hesitate to invoke, because we fear that its generous assistance has already been presumed upon over much. After all there is still much to explain in connection with the problem in hand and we can scarcely expect to be able to solve some of its mysteries with- out the aid of breeding experiments. In spite of the many difficul- ties that confront us in connection with attempts to keep these animals alive in confinement and to have them breed under exper- imental control, we feel that breeding is imperative and that success will come in time. 910 H. H. NEWMAN AND J. THOMAS PATTERSON THE HEREDITARY CONTROL OF SEX The only character that seems to be rigidly controlled—which does not vary at all in the members of a given set of foetuses— is the character of sex. Whether the set shall be male or female is apparently settled at the time of fertilization and has its physi- cal basis presumably in a dimorphism of the spermatozoa. Sex having been determined, there is no room for individual variation with regard to this character, unless there be such a thing as a more or less pronounced maleness or femaleness. If degrees of sex do exist they no doubt express themselves in terms of fertility. Along with the primary character of sex are predetermined all of the secondary sexual characters, but these are not so rigidly controlled. Probably no more highly variable characters exist than those that are associated with sex. In the armadillo the two sexes are remarkably alike with regard to somatic characters. So similar are the two sexes that we have never been able to tell them apart without examining the genitalia. A statistical study of the scutes has, however, revealed a sexual dimorphism in the scute numbers. The mean number of scutes in males is a little higher than that in females, but the difference is one of only four or five scutes in the whole banded region. A more pronounced dimorphism exists in the comparative variability of the two sexes. Not only in the species, but within the confines of the various fraternities, the males are decidedly more variable than the females. Is there any structural dimorphism of the hereditary materials that could be held to be in ahy way corre- lated with this variational dimorphism? A preliminary study of the spermatogenesis of this species has revealed that a dimorphism of the spermatozoa probably exists. If the results of a more extensive study support this tentative conclusion, it will be possible to bring the question of the pro- duction of male and female sets into line with the general conclu- sions already reached on sex-determination in the other forms in which a dimorphism of the sperms brings about a balanced chromosome complex in the female and an unbalanced one in the male. Can there be any underlying connection between the bal- LIMITS OF HEREDITARY CONTROL 911 anced chromosomal relations of the female and variational stability on the one hand, and between the unbalanced condition of the chromosomes in the male and its variational instability on the other? If such a connection exist we have, in addition to sex, another character whose physical basis may in some way be asso- ciated with the dimorphism in the chromatin content of the fer- tilized egg. A further suggestion might be made in this connection. It seems to be an established fact that the male sex is the more highly specialized, for males depart more widely from the juvenile type than do females. Might it not be possible that the unstable equi- librium of the male chromatin complex lies at the basis of the higher specialization of the male; for a condition of instability would involve an increased potentiality for progressive change? Is there any more or less justification for these suggestions than there is for the generally accepted idea that sex is in some way causally associated with the presence or absence of the accessory chromosome, or its equivalent? GENERAL CONSIDERATIONS In his chapter on blastogenic variation Vernon quotes from Weismann the statement that ‘‘the individual is determined at the time of fertilization, or, in other words, the individuality of the organism results from the fact that the germ-plasm is composed of the paternal and maternal ids which are brought together in the egg cell.” As evidence of the validity of this statement the author brings up the facts about human identical twins and on the basis of these facts claims that ‘‘ heredity is potentially decided at the time of fertilization.” In the present paper we have shown that the individuality of the organism is not precisely determined at the time of fertiliza- tion, but that the characters are hereditarily controlled only within certain limits. These limits we have been able to define with respect to a number of different characters. Our results are based on the assumption that the degree of divergence shown between the four foetuses is the index of the 912 H. H. NEWMAN AND J. THOMAS PATTERSON variational potential of the fertilized egg. We assume that if the egg which produces a given set of quadruplets could be made to produce just one individual, this individual would have a potential range of variability equal to that exhibited by the set of quadruplets. Whether or not this assumption is justified depends to some extent on whether our ideas about the operations of the mechanism of hereditary control are sound. If one attrib- ute the differences between the foetuses to the fact of the unequal distribution of the hereditary materials, he would seem to be forced to the conviction that, were a single foetus to develop, there would be no chance for the operation of such a factor, and hence the heredity would be settled at the time of fertilization. This conclusion does not appear to be so necessary when we consider that, even where the egg produces only one individual, it must still divide into two, four, etc., blastomeres, and that each division affords an opportunity for an unequal distribution of hereditary materials. In the first cleavage, instead of producing two dis- tinct individuals, the egg divides its material into two bilateral halves, which are destined to produce the right and left sides of the definitive body. Now it has been shown in another connec- tion that the correlation between the antimetrically paired organs of the same individual is of the same grade as that shown to exist between the quadruplets. This fact simply confirms the idea that the variability of the sets of foetuses gives a reliable measure of the variational possibilities of the fertilized egg. Specific polyembryony doubtless furnishes a special case of intra-indi- vidual variability, in which the original individual breaks up into several strictly homologous and independent parts. We have realized from the beginning of our work on this subject that the results obtained might appear to some biologists to be explicable on the basis of similar or different environmental influ- ences operating during gestation. For these reasons we have been careful to select for study structures little if at all likely to be influenced by environmental factors. It will doubtless be claimed by some that the foetuses are almost identical because they have developed under almost identical conditions, and that the slight differences are the result of equally slight differences in LIMITS OF HEREDITARY CONTROL 913 position, nutrition, etc. That the environmental factor cannot be seriously considered is shown when particular cases are exam- ined. What kind of an environmental stimulus, for example, could operate to produce such a definitely localized, minute peculiarity as the double scutes in the individuals of sets 121 and 123? One could hardly imagine that a slight difference in the amount or character of the maternal nutriment would produce such a condition, nor could any mechanical factor, such as posi- tion, pressure or contact with the amnion, be held accountable for so definitely localized a character occurring in several individuals. Again, in the matter of atypical band arrangements, it would appear equally absurd to attribute resemblances or differences to environmental factors. Consider, for example, the conditions in set 64. Here the region of incomplete splitting is so definitely localized and involves so fixed a number of scutes that one would have to posit some kind of environmental influence which would be able to cover just so many scutes and place itself just so far from the margin in every case. It would also be necessary to explain why two of the four individuals should show the effects of the influence bilaterally and why the other two should show it unilaterally. These and many other cases that might be examined point to the untenability of the position taken by the proponents of the efficacy of environmental factors, and strongly fortify the position here taken, that the characters dealt with are purely of blastogenic origin. We conclude then that we have without question made an advance in the direction of determining the limits of hereditary control. We have shown with what degree of exactness the num- bers of certain integral variates, the scutes, may be predetermined and how much room is allowed for the play of epigenetic factors. We have demonstrated that the alignment of scutes into bands is very largely controlled by mechanical factors. We have indi- cated the degree of exactness with which certain ‘abnormalities’ may be hereditarily controlled, and how small a biological unit is capable of predetermination. One cannot but be impressed, however, with the diversity of conditions seen in the different sets. In some of the normal sets 914 H. H. NEWMAN AND J. THOMAS PATTERSON it appears that hereditary controlis almost perfect, as, for example, in set 4, where there is a difference of only one scute among the four foetuses; in other sets, 97 for example, the variation among the foetuses is so great that, were they not found to be enclosed in a common chorion, one would hardly believe them to be related. The same conditions prevail in connection with the atypical scutes or bands. In some sets the ‘abnormality’ is repeated in the dif- ferent individuals with the utmost fidelity of detail, while in others neither the extent of the region affected nor its location is at all rigidly defined. In a word, one can speak definitely as to the limits of hereditary control for only one set at a time. What appears to be true for one case does not apply exactly to another. Even here, then, where one would expect the phenomena of vari- ation and heredity to exhibit almost diagrammatic simplicity, we find a high degree of complexity and lack of uniformity; and one is again compelled to acknowledge that nature is baffling in her manifoldness of expression and in her freedom from the trammels of exact laws. SUMMARY 1. The data derived from human identical twins cannot serve as a criterion for the determination of the limits of hereditary control, for two reasons: (a) ‘The origin of the two individuals from a single fertilized egg is assumed from the fact of resemblance. (b) The comparison between the twins is made only after years of post-natal life. 2. Armadillo quadruplets furnish a reliable substitute for human identical twins, because: (a) The phenomenon of specific polyembryony has been demonstrated for the species. (b) The unborn foetuses, with all placental connections intact, are used. (c) The scutes of the nine bands of armor, which are the objects of the present investigation, are elements that reach their defini- tive number and arrangement long before birth, and hence are excellent for the study of heredity. 3. A study of the morphology of the integument reveals the fact that the integumentary unit is a complex element made up of a bony plate, a horny scute and a well defined hair group. LIMITS OF HEREDITARY CONTROL 915 These are so closely associated and so definitely inter-related that a study of one element furnishes an index of the variability of all. For convenience, the external element, the scute, is chosen for statistical study. When the term ‘scute’ is used the whole com- plex is to be understood. 4. A statistical study of the species variation in the nine bands reveals the facts that we are dealing with a highly variable char- acter which fluctuates according tothelawsof chance. The mean, mode and median practically coincide, and the observed and the theoretical variation curves are very similar. 5. Males are decidedly more variable than females with respect to the characters studied. 6. The variability of the bands taken separately is proportion- ately greater than that of the banded region as a whole. In each band, however, the variation in the numbers of scutes appears to take place according to the laws of chance. 7. The twenty sets of normal foetuses furnish an ideal array for the study of fraternal correlation. Taking each set as a fra- ternity, and dealing with the total number of scutes in the banded region, a correlation coefficient of .9348 is obtained. This is taken ‘as an index. of the strength of hereditary control with respect to the character in question. The only correlation constants at all comparable with this are those derived from a study of the anti- metrically paired organs of the same individual. This fact con- firms the idea of the polyembryonic origin of the quadruplets, and shows that, morphologically, we are dealing with four parts of one individual. 8. The correlation coefficients determined for the individual bands are comparatively so low that the conclusion is reached that the process of scute alignment is largely mechanically determined and hence beyond the limits of hereditary control. 9. A study of the atypical variation in the banded region shows that there are several types of scute ‘abnormality,’ double scutes, split scutes and the three-hair type of scutes; and also several types of band ‘abnormality,’ fusions, splittings and additions. All of these conditions are comparatively rare and highly diver- sified in detailed expression. 916 H. H. NEWMAN AND J. THOMAS PATTERSON 10. Practically all of the types of atypical variation are found in the present collection of foetuses. They are shown to be pre- determined, in some cases, with remarkable precision, and in other cases, only in so far as their general character and location are concerned. 11. A further statistical study of pairing serves to demonstrate the truth of this relation. A mechanistic interpretation of pair- ing is offered, involving the idea that the differences in the four foetuses of a set may be due to the inexactness of the distribut- ing mechanism in cleavage. The paired condition is thought to be an illustration of the general law that variability decreases with the production of like parts. 12. Sex is the only character absolutely predetermined, but the suggestion is made that the greater variability of males, both in the species and within the several sets, may be associated with the lack of balance in the chromosome complex. 13. In reply to the possible objection that the variability of the four foetuses of a set is not necessarily an index of the possible range of variability of the ovum, it is argued that the variation between the right and left sides of the body of a single individual is of the same grade as that shown to exist among the quadruplets, and hence, from the variational standpoint, the single foetus is on the same footing as the quadruplets. 14. The conditions described are shown to be inexplicable on the basis of varying environmental factors acting during gesta- tion. 15. The lack of uniformity in the different sets of foetuses, with respect to the limits of hereditary control, leads to the realization of the complexity of the problem and to an acknowledgment that we are still far from a complete understanding of the factors involved. LIMITS OF HEREDITARY CONTROL 917 BIBLIOGRAPHY Daveneort, C. B. 1904 Statistical methods. New York. Gauton, F. 1875 The history of the twins, as a criterion of the relative powers of nature and'nurture. Journ. Anthrop. Inst. 1892 Finger-prints. Macmillan, London. Harris, J. ArrHuR 1910 A short method of calculating the coefficient of cor- relation in the case of integral variates. Biometrika, vol. 7, pp. 214-218. Newman, H. H. anp Patrerson, J. THomas 1909 A case of normal identical quadruplets in the nine-banded armadillo, and its bearing on the prob- lems of identical twins and of sex determination. Biol. Bull., vol. 17, no. 3, August. 1910 Thedevelopment of the nine-banded armadillo from the primitive streak stage to birth; with especial reference to the question of specific polyembryony. Jour. Morph., vol. 21, no. 3. Peart, RayMonp 1910 Intra-individual variation and heredity. Proc. Seventh Intern. Zool. Congress. Prarson, K. 1909 Determination of the coefficient of correlation. Science, N.S., vol. 30, pp. 23-25. Vernon, H. M. 1903 Variation in animals and plants. London. WeIsMANN, A. 1893 The germ-plasm. London. Witper, H.H. 1904 Duplicate twins and double monsters. Amer. Jour. Anat., vol. 3, no. 4. 1908 The morphology of cosmobia; speculations concerning the sig- nificance of certain types of monsters. Amer. Jour. Anat., vol. 8, no. 4. 918 H. H. NEWMAN AND J. THOMAS PATTERSON 57 10 b 9 10 a Figs. 6-8 These figures show the upper and lower surfaces of three double plates in various degrees of fusion. X 3. Figs. 9 and 10 Two specimens showing incomplete plates. x 3. 919 LIMITS OF HEREDITARY CONTROL i ( ct T FX ‘surygyds jeseyRyrq & Surmoys ueumoeds & Jo puvq 4say oy, ZI Ro Ay | X ‘spueq puodses pue ysig oy} usamyoq woIsny Jo odAy [wsoquytq & YIM usuMdeds y IT oes | al! No. 4 JOURNAL OF MORPHOLOGY, VOL. 22, ‘~ X ‘e0UedSe10U0N Jo ssovoid oY} Sutnp vipiourtid 4Jo] pus 4YSIA oy} Jo Suryoour yoojrodumt uv Aq poonpoid usoq oavy AvuUIsIyy, ‘“puvq oyy Jo yutod e[pprur oy} ye Ajo}vuNxoIdde SULA] UOISaL o[qnop JAYS B YIM puBq ae OL FL SW xX ‘pueq ojeid -MOdUT JOYS & Suronpord Ur pozfNses svy YOrYM 4yor oy} uo y1[ds [eUOTyIpps UP SI o1OY} YeYY ydooxe ‘Burpoderd oy} 04 iepronig El ‘SIT vl HY 920 LIMITS OF HEREDITARY CONTROL 921 | ae. Fig. 15a Right-hand ends of the 8th and 9th bands of a shell showing two slight ‘abnormalities,’ for a description of which see text. X 2. Fig. 15b Under surface of the affected regions of the preceding. X 3. Fig. 16 Under surface of a portion of the anterior half of the pelvic shield. Note the hexagonal shape of a majority of the bony plates. These are all per- forated by one or two small holes through which blood vessels and nerves pass. Ke THOMAS PATTERSON NEWMAN AND J. Isl dale 922 LUERLehete Pig, LANAALAAL Aah. PULL ERPERERS thet es, eRe REREPT Pemee egg ety FFE HeVVT tis #445, imal. iving an TVig.17 Snap shot of a 1 LIMITS OF HEREDITARY CONTROL 923 bithaas ekg 5 40 if WL ee hac aia iw fig. ‘ Mvtwisiee ithe Peis cans cdl re ily beker PANG PAM Lek bi ide aig SEAS “eh BUR ih e erred. Ae ROME ney te bhdaua ss Laan?” “MAE Fig. 18 A shell showing a fusion between bandsland2. 1. Fig. 19 A shell showing an addition to the banded region from the scapular shield <= 20 21 Fig. 20 Photograph of a portion of the left side of a shell. This shows the con- dition of the primary and secondary scutes. X 3. Fig. 21. Photograph of the upper surface of the pelvic shield (also part of the banded region), to show the pebbled effect. X 3. 924 LIMITS OF HEREDITARY CONTROL 925 . ANE S o oe Pe: create enatrste 23 Fig. 22 Left side of the shell with the high count of scutes (625), many of which are of the 3-hair type. A number of these can be made out in the photograph, especially in the first and second bands. 3. Fig. 23 The middle portion of the pelvic shield of a specimen showing a ‘jog’ in the region of the ninth band. X 3. 926 H. H. NEWMAN AND J. THOMAS PATTERSON Fig. 24 Dorsal view of the anterior ends of the litter of embryos from female no. 125. The embryos all show a well marked crease in the middle of the scapular shield. The mother also had this same peculiarity. Note the double scute in the middle of the last row of the scapular shield of each of the two embryos lying on the right. These two embryos are members of one pair. X 2. EXPERIMENTS, ON THE CONTROL OF ASYMMETRY IN THE DEVELOPMENT OF THE SERPULID, HYDROIDES DIANTHUS! CHARLES ZELENY SEVEN FIGURES INTRODUCTION The following experiments? were made as part of a study of the factors controlling asymmetry in the Serpulid, Hydroides dianthus. The operations were performed on young individuals in which asymmetry was just starting to develop and in which the adult mechanism for reversal of the opercula was not yet present. The results are interesting not only in connection with the prob- lem of the cause of the original asymmetry and of the reversal of the opercula in adults but also as bearing on the extent of agree- ment between regeneratory and ontogenetic stages. A more definite formulation of the problems follows. 1. The first functional operculum of young animals is developed before the first rudimentary operculum. In the absence of any special mechanism in the form of a rudimentary operculum does reversal of position follow removal of the fanctional organ at this stage? What bearing does the behavior following such opera- tions have on the question of the origin of the original asymmetry and the cause of reversal of the opercula in adults? 2. The first functional operculum of the young is different in type from that of the adult. Does the regenerated operculum 1 Contributions from the Zoological Laboratory of the University of Illinois, No. 8. 2 The experiments were performed at the Biological Laboratory of the Bureau of Fisheries at Woods Hole, Mass., during July and August, 1909. Iam indebted to Dr. Francis B. Sumner, the director, and to Professor Raymond C. Osburn, acting director during a part of my stay, for many courtesies. A preliminary report of the experiments was given before the Central Branch of the American Society of Zoologists at the lowa City meeting, April 8, 1910. 927 928 CHARLES ZELENY resemble the one removed or does it develop directly as an oper- culum of the adult type? 3. Further, since the first operculum is a modification of a branchia and therefore originally has a respiratory function, is its removal followed by regeneration, first as a branchia which only later develops the opercular modification, or is the opercular modification regenerated directly? For a description of the opercula in adults and for experiments on reversal, reference is made to two former papers (Zeleny ’02, 705). The development of the opercula in the young is treated in the second of these (pp. 38 to 54). A brief outline of the neces- sary points must suffice here. ASYMMETRY IN ADULTS In the adult Hydroides dianthus there is a large functional operculum or tube plug on one side of the body, either right or left, and a small rudimentary operculum on the other side (fig. 1). The functional operculum (fig. |, /”) has a stout stalk end- ing in a hard chitinous enlargement consisting of a serrated cup, from the centre of which rises a circlet of curved spine-like pro- jections. The genus Hydroides is characterized by the presence of these two separate circles of serrations or projections in its functional operculum. ‘The rudimentary operculum (fig. 1, /) is a small bud-like structure corresponding in position exactly with the functional operculum of the opposite side of the body. Near the base of each opercular stalk there is a well defined line, the breaking line or breaking joint. When the cup of the functional operculum is removed, the rudimeatary operculum of the opposite side develops into a func- tional operculum similar to the one which had been injured. The stalk of the inj:1red operculum meanwhile drops off at its break- ing joint and a rudimeutary operculum develops in its place. There is thus, as a final result of the operation, a reversal in posi- tion of the two opercula. This reversal may be repeated several times. CONTROL OF ASYMMETRY IN HYDROIDES 929 AK ) YY yy / Lf ) / Ss) \ \\ S VAY Hy Vy Yh? Vif WOH WW \ Uy S Yi 4, LS \ i / SMH JL I ff / / Le ff f, / : / Uy, tf, 7 ci S4i tt LOBES B Fig.1 Hydroides dianthus. A, dorsal view of right-handed specimen, showing relations of parts. Ends of branchiae and functional operculum not given (6); B,C, diagrams of anterior surface of head of left-handed and right-handed speci- mens (X6); D, branchia viewed from inner surface (X25); #, rudimentary oper- culum (X 30). fF, functional operculum (X 30). The process is very similar to the reversal of the two chelae in Alpheus as described by Przibram (’01 to ’07), Wilson (’08), Zeleny (’05) and Stockard (10). In Alpheus there is a larger so- called ‘snapping’ chela on one side of the body and a smaller ‘cutting’ chela on the other side. These chelae differ from each other both in size and in structure. When the snapping chela is 930 CHARLES ZELENY removed, the cutting chela changes into a snapping chela and in place of the former snapping chela a cutting one is developed. In Hydroides the presence of the functional operculum in some way inhibits the growth of the rudimentary operculum into a functional one. Likewise in Alpheus the presence of the snap- ping chela inhibits the change of the cutting chela into a snapping chela. It is evident that the rudimentary operculum in the one case and the cutting chela in the other need only the proper stim- ulus or removal of an inhibition to develop at once into organs like their mates. These smaller organs are therfore in one sense merely stages in the development of the others. Various suggestions as to the explanation of this inhibition have been made. In Hydroides the injury to the large operculum pro- duces activity on both sides of the body, bringing about the immediate growth of the rudimentary into a functional operculum and astart in the same direction to form a rudimentary operculum on the side of the injury. The functional operculum is reached only on the side with the earlier start, the presence of the one functional operculum restraining the possible development of another functional one. This view is supported by the result obtained when the head of Hydroides is removed. In this case, as also in a part of the cases in which the two opercula are removed without injury to the body proper, a functional opercu- lum is developed on each side of the head though the size of each is usually reduced. In such a case the two opercula get an equal start and neither one is able fully to restrict the development of the other. The cases of similar chelae in the adult Alpheus and Homarus may be explained in a similar way by coincident development. Under the ordinary conditions of removal of both chelae the ad- vantage of greater blood-supply and probably other features lies oa the side of the stouter snapping or crushing chela. Therefore reversal does not occur in such cases. When, however, there is a secondary advantage of just the proper strength occurring to the side of the former smaller chela two equal chelae may develop. The probability of this explanation is strengthened by the fact that in those cases in which two equal snapping or crushing chelae CONTROL OF ASYMMETRY IN HYDROIDES 931 were obtained by Wilson (’03) and Emmel (’08) there was an addi- tional factor, difference in time of removal of the two chelae, in- jury of the nerves of the chelae or the removal of the walking leg or legs on the side of the former more slender chela (see Stockard ’10, discussed below). Wilson has suggested that the reversal may be under nervous control and he made a study of it in Alpheus by cutting the nerves going to the chelae. His results, as he himself recognized, are, however, not conclusive, since other influences, such as disturbed blood-supply, are not eliminated. Wilson also suggests that the snapping chela may develop always on the side of the body which has the greater amount of material to start with and which there- fore has the greater body of nutrition directed to it. Stockard has tested this hypothesis as applied to the whole group of append- ages on the two sides, by removing the walking appendages on the side of the cutting chela in the case in which the snapping chela is removed. The operation leaves the greater mass of ap- pendage ‘material on the side of the chela removed. Neverthe- less, practically all of the Alphei showed reversal as in cases without removal of walking appendages. Variations of these experiments showed the same negative result. The removal or non-removal of other appendages on either side has no proved relation to the phenomena of reversal. It should however be said that Stockard’s experiments do not test the essential point of the hypothesis of difference in amount of material or effective mechanism for growth in the form of blood-supply, etc., between the stumps of the two chelae themselves. The walking legs may not be directly concerned in the difference of activities of the two sides at the first chela level and the hypothesis of difference of materials at the first chela level be unaffected thereby. Further- more, if there be a correlation between these different levels, is it proper to assume, as Stockard has done, that the smaller amount of nutritive activity will be on the side regenerating other append- ages. As a matter of fact, the opposite assumption is indicated by some experiments (Zeleny ’09), and Stockard’s experiments themselves, as far as they show any difference in results between the two cases, point in the same direction (Stockard ’10). JOURNAL OF MORPHOLOGY, VOL. 22, No. 4 932 CHARLES ZELENY ASYMMETRY IN YOUNG INDIVIDUALS The adult asymmetry of Hydroides is preceded by a symmet- rical stage. With the development of a functional operculum the first asymmetrical phase is assumed, followed in turn by a normal reversal in position and a change in structure to the adult type. One or more additional reversals may then take place. It was shown that these reversals occur in nature and furthermore it is probable that they may come periodically without being pre- ceded by any definite injury to the functional operculum other than the wear of ordinary use. The free-swimming larva of Hydroides, after attachment to a solid object, secretes a tube around its body and develops bran- chiae on its head. There is a definite stage in which.the branchiae are all alike and symmetrically arranged with respect to the median line (fig. 2). Additions to these branchiae take place at the lower edge of each lateral group. At a stage slightly older than the one shown in fig. 2 the young serpulid closely resembles in its branchial characters an adult of the genus Protula (Fritz Miiller ’64). There is no trace of an opercular modification 1a any of the branchiae. The branchia next to the dorsal one on the left side then be- gins to form a terminal cup with a single row of serrations (figs. 3 and 4). The branchial filaments are still present and the respir- atory function is evideatly still retained along with the new tube- plug function. The opercular branchia at this stage may be compared with the functional operculum of adults of the genus Apomatus, in which also the opercular enlargement is on a stalk bearing branchial filaments though the enlargement itself is dif- ferent in structure. Apomatus has in addition a rudimentary operculum of the same type as the functional, 7.e., borne on the end of a branchia which retains its respiratory filaments (fig. 6, A, B,C, D,). It is important to note that the operculum in Hy- droides first appears on the left side of the body while in adults it is sometimes on the right and sometimes on the left. During the following period the branchial filaments disappear from the opercular stalk and at the same time the corresponding CONTROL OF .ASYMMETRY IN HYDROIDES 933 Ls Fig. 2 Young Hydroides dianthus with six branchiae (individual C, no. 2780). The most dorsal pair has not yet developed its pinnules. The levels of the cuts made in the operation are indicated by C. Fig. 3 Hydroides dianthus, age 34 days; stage with four pairs of branchiae. The next to the dorsal one on each side is shown in full. The left one shows the beginning of the knob of the functional operculum. The right one drops off later and the first rudimentary operculum is developed from its stump. 934 CHARLES ZELENY branchia of the opposite side drops off, the break appearing at the base of its stalk. In its place a bud-like rudimentary oper- culum is developed (fig. 5). There is then on the left side a func- tional operculum with a naked stalk and a terminal cup’ bearing —SJ q = 4 y Fig. 4 Hydroides dianthus, age 26 to 28 days but further advanced than the in- dividual shown in fig. 3. Stage with four pairs of branchiae. Figure shows the three most dorsal pairs. The ventral pair is directed away from the observer and is not shown. The opercular knob of the left side is notched and its stalk still retains the respiratory pinnules. The branchia next to the dorsal one of the right side has eight pinnules but differs from the other branchiae, except the oper- cular one, in the absence of new pinnule buds. A, level of cut in individual A, (no. 2778); B, level of cut in individual B, (no. 2779); Ba, regenerating functional operculum one day after operation in individual B. a single row of serrations, and on the right side a rudimentary operculum. The characteristics are those found in the adults of the genus Serpula except for the fact that in Serpula the func- tional operculum may be either right or left (fig. 6, £). CONTROL OF ASYMMETRY IN HYDROIDES 935 After a period of activity in this stage lasting several days the functional operculum drops off. The loss is accompanied by the development of the rudimentary operculum into a functional operculum not like the one that had been lost but more compli- cated in structure, with two distinct and qualitatively different X Fig. 5 Hydroidesdianthus. Dorsal view (X 38). Onthe left side is functional operculum of Serpula type with naked stalk and one row of serrations. On right side is rudimentary bud developed from the base of the cast-off second branchia of that side. rows of serrations. This operculum is similar to that of the adult Hydroides. In place of the operculum that had dropped off a rudimentary operculum is developed. The condition is now like that of the adult except that some adults have the functional oper- culum on the left and others on the right. It is therefore neces- 936 CHARLES ZELENY sary to assume further reversal or reversals of position during later development and perhaps throughout adult life. The original asymmetry is apparently always of the same na- ture, 2.e., the functional operculum always develops on the left A Fig. 6 A, B, C, D, Apomatus ampullifera. Adult. A, dorsal view showing functional and rudimentary opercula and branchiae. Pinnules not represented (X 5): B, tip of non-operculate branchia (x 17). C, tip of rudimentary operculum (X 17): D, tip of functional operculum (X 17): £#, distal portion of functional oper- culum of Serpula vermicularis (xX 19). side and from a particular branchia. Adult individuals differ in character because some have had an odd and some an even number of reversals. CONTROL OF ASYMMETRY IN HYDROIDES 937 EXPERIMENTS ON ASYMMETRY IN YOUNG INDIVIDUALS The first functional operculum appears before the first rudi- mentary (fig. 4). During the stage in which the opercular stalk retains its branchial filaments there is as yet no modification of the branchia of the opposite side. It ends in a point just like that of other branchiae. Buds of new pinnules are however absent (figs. 3 and 4). There is at this time no perfected mechanism for reversal as in the aduit. The removal of the functional opercu- lum, or at an earlier stage the removal of the end of the branchia which later develops the opercular enlargement, was accomplished in several individuals (figs. 2and 4). Ia only a few of them, how- ever, was the operation clean cut and‘the further history of the experiment followed. The character of the material did not allow experimenting with narcotization. It was therefore necessary to keep watch of the young animals under a binocular microscope, holding a needle knife blade above the opening of the tube. With exceptional good fortune it was possible to remove the terminal cup ofthe operculum in a few instances without seriously injuring the rest of the animal, though in most cases the neighboring branchia of the same side was also injured. There was no reversal of opercula, though interesting develop- ments followed. In place of the removed functional operculum a new one like the one removed was developed. There was no loss of the old stalk, the regeneration taking place directly at the cut surface (fig. 4, Ba). It might have been expected that, in case regeneration occurred, the new operculum would at once grow out as one of the Hydroides type as found after the first natural reversal. This, however, did not occur. The regenerated structure therefore repeats the stage of the removed structure and does not pass on to the next ontogenetic stage (fig. 7, A). The effect upon the branchia of the opposite side is also inter- esting. The terminal part of the branchia develops a small knob, approaching in character a rudiméntary operculum of this species but formed from a group of cells which in normal growth never develop opercular enlargements, but are, in fact, lost when 938 CHARLES ZELENY this branchia drops off to make place for the rudimentary oper- culum developing in its place (fig. 7, B, C). A branchia with a small knob-like rudimentary operculum at its end is thus formed. This condition is never found in normal development in this species but is a normal feature of the adults of the genus Apomatus which have also functional opercula with stalks bearing branchial filaments (fie. 6:4, B,C 1D): B q E ), Fig. 7 A, regenerated functional operculum of Serpula type. Individual A (no. 2778), on August 3rd. B, branchia next to the dorsal one of right side as modi- fied, following operation on the corresponding branchia of the left side. Indi- vidual A, (no. 2778) July 31. C, enlarged tip of B. D, rudimentary operculum on left side of individual A as developed after the functional operculum of the Serpula type had dropped off, August 14. E, functional operculum of Hydroides type as developed on the right side of individual A after the functional operculum of the Serpula type on the left side had dropped off, August 14. G It might have been expected that the removal of the functional cup would accelerate the breaking off of the corresponding branchia of the other side of the body and be followed by the development of a normal rudimentary operculum. As a matter of fact, the acceleration of opercular development is found, but in a way entirely different from the normal. eo hap. top a ; sartertalis \ > fab undod AS x Be FS AY Ree + Lapfouls cordis ~~ Deanathontia pulmonis. @T Pulmo cozdis flabellii eft x carne mollt z serea creat? fpnmecoagnls “~~ fis eft fm Lottantini. £e eft cépofit? pulmo ex carne moll pamenlis 2) Pa cu vale. fezer vena A harp i atextro Veritriculo cozdis poztat far ~\_ chibanteftfigurar’. gn pre pofterion marozé bns logitudine > tn pteant Sas figura oulmonis Fig. 8 Sketch of the heart from Peyligk, 1499 gentes.’ In the figure that I have photographed the iris of the eye is black, while in that reproduced by Sudhoff from the copy in the Leipzig library the iris of the right eye is indicated by a white circle. The book is provided with text-figures of separate organs copied from Peyligk’s treatise. The figures, however, are not printed from the same blocks; they are nearly identical but care- ful inspection will show slight differences in the lines. Fig. 10 shows the sketch of the liver with a part of the text. It is the same as the corresponding figure in Peyligk but is not quite so carefully engraved. 966 WILLIAM A. LOCY Gregor Reisch. In the 1504 edition of Reisch’s Margarita Phil- osophica there is an illustration (see fig. 11) showing new details in internal anatomy of the thorax and abdomen. Although the anatomy is very crude it is an improvement over the correspond- ing figures of Peyligk and Hundt. The kidneys and bladder are represented in a more nearly normal position. The lungs are Fig.9 Visceral anatomy from Hundt’s Antropologium, 1501 divided into lobes; the liver, stomach, spleen and intestines are still very untrue to nature. The nomenclature of the period is shown in the names attached to the organs, the lung, ‘pulmo,’ the heart, ‘cor,’ the liver, ‘epax,’ the kidney, ‘ren,’ the bladder, ‘vesica,’ ete. In 1503 a similar picture had appeared in the edi- tion of the Margarita Philosophica from the printing house of ANATOMICAL ILLUSTRATION 967 PU. cxip.natura ls bific. 4.10 fi, Duris apantbue ~ Burcéns,y,can,c, * cattus i vino rlity. KG fert.ariftologis vi. Opbor?.rlviij. {plenetict, Sple ij.apbort.cxrit, - utumn? tnimic? eft fplenert | bet put pulmoné.q: pay pos tabe Scum Alber.rt9.de gia libo,c. vij,tn fi, bus tins caput. ty. : Par eft membrii o2gas Lonfantinys | nici pumifanguisine Tenapote £par? figura ftruméralitcr gnarius, ie ; arnefanguine a villis nervoy vacua tvenulis otertus magniocaui.rubicundy.et luz bricofum, 11 dextro latere sialis collocati ftoacbo fuprepofiti_ve Gy 7 det Siete puma aoe d1geftionc , ~. Lpar eflemembii organic ds — ae HUS.CHE coe pinerfitas formax In ptib? a : toto. *flon em gliber ps ei? inre Sy gralie sud capt adr landbl Lametl penute tpi? epar intri Nicolus.fmo,¥. fice aftitucces.ocanttate zgibbo BraC,V.C.). fitaté babedt eandé ci toto. né tn - ft gfa nomé totiue merencur, &par auc latine dz tecur quafita Bute? riiti, tercr Tena chilis Fig. 10 Part of page from Hundt’s Antropologium, 1501 John Schott. In this earlier picture the ureters are not shown, and the intestine is represented as connected with the bladder. I am indebted to Wieger’s treatise for the sketch of 1504, that appeared in the Margarita Philosophica published by Grieninger. This figure was reproduced in other texts and the original of this cut (fig. 11) is in Brunschwig’s Destillirbuch. Leonardo da Vinci. With Da Vinci we come to the one man who, before Vesalius, showed independence in observation and 968 WILLIAM A. LOCY S PHISICVM “™ 2 a 5, = g3 mn. “a \ Y Fig. 11 Visceral anatomy from Reisch’s Margarita Philosophica, 1504. This cut from Brunschwig’s Destillirbuch, 1512 (after Wieger) notable fidelity to nature in his sketches. Although the larger number of his anatomical drawings were made about 1510 they were not fully published until 1898 and 1901. They bear inter- nal evidence of having been made from actual dissections. It is well known that he became associated with Della Torre who projected a treatise on anatomy for which Leonardo was to supply the drawings. Nevertheless, Da Vinci had studied anatomy inde- pendently before his association with Della Torre, and he contin- ANATOMICAL ILLUSTRATION 969 ; baeerd ¢ el ahem brent pret tert earl j i i saree VE TEE eee perre T, Pe Witiad Fi bake itl A nue | eo kL eaekat ke apy at itt) a sla lien eee ; _ greet * - \ Ane Gt AG) wun . logs AMEN alirg pire? llewree Vay on! Ye che wht e® Vag ep Se Rae ll ee & : 7 ey * t “ f } J ant bed Siero wiethser amomdoy ap ofp Recon) og re “ra map [[t necare sre] exes afer sen Eo rane Nanp abep aie “NF crs] 2 a> tie Raab nathan ag nrin) ive] afenet ane) Nef Sted) oorery Omi fr vat Vmpay ane + UV) Aad arn % Lyedn bef atna naw ali 2? ™, 7 i Ape pr vetese wun aad py arta) ai +i wien tl wifpnee or sav ry oem n'y» Aalyg asjonnd mbaly foil h ete vie Fs ead a Vb and : i AWS ety. i) were ternnead) vin oe nel vay are : a" poter wre Dope rt alpina ctl \) oN acces ; PST re. | see) ee ow ened cred] all sho sti am srhetn > nverbow i Wife poy! Fig. 12 Anatomical sketches from I Manoseritti di Leonardo da Vinci, 1510 (after the Paris facsimile edition) 970 WILLIAM A. LOCY ued to make dissections and anatomical sketches after the death of the latter in 1506. We may assume that his method was im- proved by intimate collaboration with a professional anatomist, but we must recognize that this extraordinary man was a master unto himself. The association with Della Torre was not merely that of an artist working under an anatomist who exposed the parts and required sketches made under his direction. It was 24 fo oat ab 7 ot eens or hE np ating Hay pe [aad oat et ‘ ead eqahee ction P (ROE me eed PINNED ote aiinp Ale allen se] pine? ofa) Wap ciidn tala rte ona pairs) ut Afarfprray alo : 5 eee agen NP aslibs ato aoe: 3 an om? one} Aged Area ale ae git ty wD be Fig. 13 Anatomical sketch by Leonardo da Vinci, 1510 rather the codperation of two anatomists, one of whom was gifted with great powers of artistic delineation. Antonio de Beatis had it from Leonardo’s own lips, about 1510, that he had dis- sected not less than thirty human bodies, both male and female. Leonardo projected a comprehensive work on anatomy of which he speaks in his History of Painting, and also in his manu- script notes. The notes and drawings bear testimony that this treatise was not designed merely for artists but was to be, as well, a work for medical students and for the professional anatomist. ANATOMICAL ILLUSTRATION 971 The working drawings and notes for this projected work are pre- served as a part of the manuscript collection in the Royal Library at Windsor Castle. They were published in Paris in 1898 as Foglio A of Leonardo’s Manoscritti. This sumptuous volume Poa ep alciode Boars gethyrad Fig. 14 Anatomical sketch by Leonardo da Vinci, 1510 contains 245 anatomical sketches, reproduced as fac-similes both as regards the sketches and the paper upon which they are drawn. The notes are translated into French. His other anatomical sketches, also in the library at Windsor Castle, were published raved! sien mene arate fom PR om Pore l Eola A 1 ee piel om ap yans | 0 Spent 3 ¥ hie Syewwed ena) é F ; Yale wpaba Pe ver od fat ud Her hd mole vats s] id ates a wrrsiq Yt wet ntenle ots Dae enon ter arenes six =e allewpen | 43 WILLIAM A. LOCY Sif anes pst: i © saabid eat) hee (ane) “Byrr a ahd hd fat et} othe} Nn Anatomical sketch by Leonardo da Vinci, 1510 ANATOMICAL ILLUSTRATION 973 in eight additional volumes in 1901. This does not include the volume on the anatomy of the horse. The range of the drawings is astonishing; the entire collection embraces more than 750 separate sketches, some of them being several times repeated. The notes accompanying the sketches, always written from right to left, are, usually, descriptions of the figures, but, sometimes, are general reflections regarding the plan of his projected book. That he read anatomy is evident since he specifically corrects some misstatements of Mundinus. Leonardo placed great reli- ja: Fi vari ce ce DR en eer diee © port * x <~ vader Obed ig asap . ' ; AVA? can Ye orae] eq ed ee log artiaw df’ «Jada danke) mee?) *P> hae J f 4 - — > 2 in oO Fig. 16 Sketches of sections of the brain by Da Vinci, 1510. ance on good figures, declaring them to be essential to the under- standing of anatomy. Some of his delineations of muscles have been so frequently reproduced that they are well known, but it is not so generally known that he made deep dissections of all kinds including the viscera and the brain. The reproduction of a few of Leonardo’s sketches will serve to show their quality, and will at once reveal the fact that they are totally different from any other sketches of the period. These drawings are not made from anatomical descriptions, but from 974 WILLIAM A. LOCY 9 “CEllaye, aed Oscoronale offaperne ty Sti ame “ off ra alereferine oH ‘——- 7 ! gph mrt ft OF 6 Adintory Cofte vere et mein : in O6 femoris -O8 Thi-e¢ Pre fubAnchie Veearia offa ‘.)\ anvil) Fig. 17 Plate of the skeleton published by J. Schott, 1517 (after Wieger) original observations. The drawings speak for themselves, ac- cordingly brief comments will suffice. Fig. 12 shows a page with four separate figures and manuscript notes. The heart and chief blood vessels of the thorax and abdo- men, and the sketches of the lungs are obviously made from actual dissections of the human body. ANATOMICAL ILLUSTRATION 975 méfdie 3 digelerté pS sf ctena3i3 Oreafi.declariert 111 Dciwe “Abteilung def bonps m0 003 hirns cellen Fig. 18 Anatomical sketches from Phryesen’s Spiegel der Artzney, 1517. This cut from a Dutch edition of 1519 976 WILLIAM A. LOCY eine x NRL i ‘ beeper TE EET p97 200 EP 2 J 0 HE JEat A COC Od Ly x =a ee ef Ne Fe } 4 q eee ange tariacii ampliti \ mis Addition. 0 bus fuper anaco | we | ORD BS mia mudini ona cunt certu ciufdé in pittiniizvey | initoz€ Eedacto,, SOC OO Sie 1 & cree 2S Mae, Sa hg eS i ti Z Mis : a ais ‘ ~. RS LRT NA ST OE ITT EIT EET SNESSS LSS SSS SSS See p hay, ALY 4 AaYs: Wk aes AL OTT OITA =~ \ ESI ae ¥ $f ecmitaon uaa ORT G OL TD AE ftp eee SS TS eee SS Fig. 19 Title page of the Commentaries on Mundinus by Berengarius, 1521 ST ANATOMICAL ILLUSTRATION aetna ne ED 7 eo | AS, EF ane eee = m Ra FF CON <4 f vo scare ceive Fa ly snes \ t & DE ANATOMIA Te DAD: Se 4 ay Z > CS 6 s COLA OER PLO BaP eX Aan Be AG Fig. 20 Figure of the skeleton from the Isogoge Breves of Berengarius, 1523 Fig. 21 Posterior view of the skeleton from the Commentaries on Mundinus Cin ifta fi gura uvident offa ptis poz fterioris hois uidéf ét due calue:in Gru dextraurcol ronalis omif fura’d €i pte fupiori: uf & fagittalis: dé imediozut ét laude comif? fura: g éi pte _iferiori:a Jate - fib? ét nidet due cémifiu reatme fupra- noijateinana b& tomia cranei 5 g funt fupia emifuras fq | mofas seek Ppe'aures:i5 ifte funt fere ifenfibiles: a finiftris é alia caluazin G ui dé€ madibu/ le:& pars ¢6/ miffure coro nalis: &'due comiffure in Esk fra fagittalé ad uni Jatus exiftentes: & uf unt os de duob’ offib*® of Berengarius, 1521 WILLIAM IX LOCY ~ ERS S cay ANATOMICAL ILLUSTRATION 979 Fig 13 shows the chief blood vessels of the neck and the adja- cent region. Fig. 14 shows a deep dissection of the blood vessels of the thigh. | In fig. 15 we have a rather comprehensive dissection of the thorax and abdomen with the alimentary canal removed. The original of this figure is 1383 x 22 inches. A limited number of drawings can give no adequate concep- tion of Leonardo’s work in anatomy. His sketch of the stomach and intestine is a good drawing of the relative size and the normal arrangement of these viscera. In the delineation of muscles it is not merely the superficial layers that engage his attention, he shows details of the arrangement of the tendons on the toes and fingers, a number of cross-sections of the leg at different levels, the muscular architecture of the heart, etc. Among his many pictures of the bones, he correctly draws vertebre from various aspects, and the bones of the fore-arm in pronation as well as in other positions. He made sketches of the dissection of nerves. His figures on generation show uteri opened, with contained foetuses, and the placental connection. Before leaving his work, however, we should have one of his sketches of the brain as shown in fig. 16. Here one sees, on the left, a median sagittal section, and, on the right, a horizontal section. These sketches show fairly the extent to which the brain had been dissected up to the year 1510. Other contemporary or nearly contemporary artists, as Michael Angelo, Raphael, and Albrecht Diirer, made anatomical sketches, but not so comprehensive as those of Da Vinci, and the details regarding which it is not necessary to consider. Johannes Schott. In 1517 there appeared from the publishing house of John Schott at least two anatomical plates, one repre- senting a skeleton and the other a sketch of the internal anatomy of the body. The picture of the skeleton is shown in fig. 17. It is still very crude in its execution, but in some particulars is an improvement on the earlier printed figures. The skull is better drawn than in the plates of Helain and Griininger (figs. 3 and 4), but it still shows the spurious ‘os laude, sive capitale.’ JOURNAL OF MORPHOLOGY, VOL, 22, No. 4 980 WILLIAM A. LOCY There are other marked deficiencies as in the arm and wrist, where the carpal bones are enumerated as eight, but are not drawn, ete., etc. The sketch of the internal dissection was published in several texts as in the Phryesen of 1518 (see Chievitz, p. 90), 1529, etc. Choulant reproduces a similar but not identical figure, also from Phryesen, the plate of which bears the date 1517. Phryesen. (Fries, Friesen, etc.) In 1518 there appeared in Strassburg the Spiegel der Artzney of Laurentius Phryesen, con- taining two plates. The one is a copy of the Griininger skeleton (see fig. 4), and the other a visceral anatomy, surrounded by six figures of the anatomy of the brain and one of the tongue. This cut appears in the different editions of Phryesen with some modi- fications. Fig. 18 shows a copy of this plate from a Dutch edi- tion of Phryesen dated Strassburg, 1519. The original woodcut is 54 x 72 inches. The edition of 1529 contains another picture of the visceral dissection,—the same as shown in Chievitz, fig. 31,— that lacks the marginal sketches of the brain, and is also somewhat different in other details. The figures of the brain in the Spiegel der Artzney, except for those of Leonardo, are a new departure in anatomical illustrations. Jacobus Berengarius Carpensis (Carpus). Berengarius has often been heralded as the greatest anatomist between Mundinus and Vesalius, and, if we except Da Vinci, the assignment of this rank to him is perhaps justified. Whatever may be said of his alleged dissection of more than one hundred bodies, the illustra- tions of Berengarius are not original, nor are they based on good observation. They bear resemblance to sketches in the manu- scripts of the fourteenth century and to printed pictures in earlier publications, as the Conciliator differentiarum (1496), Margarita philosophica (1504), ete. As Has already been said, we find that all sketches of the period, with the sole exception of those of Da Vinci, show interrelationships with manuscript illustrations as well as with earlier printed figures. As Roth has pointed out, the anatomical writings of Berengarius are compilations without credit being given to the original sources, and there is inharmony between his text and the illustrations,—a circumstance that is, at times, adverted to by himself. It is altogether likely that the ANATOMICAL ILLUSTRATION 981 cuts were inserted by his publisher from such pictures as were available. The first anatomical publication of Berengarius was an exten- sive series of commentaries on Mundinus. In this the text of Mundinus is printed in larger type, and the forty commentaries in smaller, but so extensive are the annotations that the book is brought up to a thick quarto volume of 1056 pages. This book, published at Bologna in 1521, is rare and a cut of the title page is shown in fig. 19. The size of the original is 42 x 7 inches; the border is red and the enclosed printing black. His commen- taries contain at times corrections to Mundinus, and show the results of some observations mixed with dialectic compilations from the earlier writers. In the 21 illustrations the dependence on tradition is very marked. He soon branched out for himself and wrote an introduction to anatomy, designated Isogoge breves, etc., which was first pub- lished in 1522 and followed by a modified edition in 1523 which is the only one well known. In addition to a copy of his commen- taries of 1521, I have had for examination three editions of the Isogoge breves; that of 1523, Bologna, 4°, 80 leaves with 23 wood- cuts; an edition of 1535, Venice, 4°, 63 leaves, 19 plates, and a small pocket edition (22 x 44 inches letter-press), dated 1530, and containing 24 figures. The illustrations are wretched copies of those of the edition of 1523, the increase in their number, by one, is owing to the separation of two figures that appear on one plate in the larger edition. This appears to have been a relatively _ cheap edition for students. More than one-half the illustrations of the commentaries are reproduced in the Isogogee of 1523 and new ones are added. Most of the plates in the edition. of 1523 are provided with an ornamental border, added to a double line boundary, while the plates of the commentaries of 1521, are limited by a single line border. Roth reproduces a full-size figure of the skeleton from the Isogogee of 1523, but his plate lacks the ornamental border. Fig. 20 is a representation of the skeleton from the Isogoge of 1523 in which the ornamental border has been retained, but the marginal description, present in the original, has been cropped 982 WILLIAM A. LOCY off. This curious figure has 13 ribs, widely expanded pelvis and a spurious fissure in the frontal bone. The posterior view of the skeleton, shown in fig. 21, is taken from the commentaries of 1521. It has a single-line border and the marginal note has been retained. The basin-like pelvis appears more fantastic than in the preceding figure. The skull shows two spurious furrows on the parietal bones, the presence of which seem to have confused Berengarius. The two best illus- trations in Berengarius are those of the bones of the hand and of the foot. The close resemblance of these pictures to drawings of Leonardo da Vinci gives ground for the suspicion that they were In some way based upon his sketches. Although this is a mere conjecture, these two figures are on a different plane of accuracy from any other illustrations in the Isogoge breves. Dryander (also known as Johann Eichmann). This professor of anatomy at Marburg published, in 1537, an Anatomiae h. e. corporis humani dissectionis pars prior, etc., illustrated by 20 plates that were based on dissections. I have not seen a copy of this work, but have examined his edition of Mundinus and other earlier writers, published in 1541, which contains most of these earlier figures, some new ones, and 18 figures copied from Berengarius. The copy at my disposal contains 45 figures, one plate of which is repeated. Some of Dryander’s illustrations are a considerable improvement on those of Berengarius. Fig. 22 shows his sketch of the alimentary tube, the original woodcut being 44 x 6inches. The drawing of the caecum and the vermi- form appendix shows that it is based on observation, but the figure is not so good as that of Da Vinci of the corresponding parts. Walther Hermann Ryff. In 1541 appeared Ryff’s Anatomi with a very long and cumbersome title. This book, of which I have examined a copy in Dutch and one in German, was published after the first plates of Vesalius (1538) and before the appearance of the famous Fabrica (1543). It, with the Dryander mentioned above, lies on the border line of pre-Vesalian illustrations of anat- omy. One of Ryff’s illustrations of the arterial circulation, repro- duced in fig. 23, gives a fair idea of the appearance of his sketches. ANATOMICAL ILLUSTRATION 983 Other anatomical illustrations of the pre-Vesalian period em- brace the very rare copper plates of Canano, showing bones and muscles of the arm. Choulant says that, prior to 1543, one book \ " \ iiss \ WW: at \) mA & = S > G S& LSS 4 iS VF SIS Z Veo eS \\ Ss Zw OSS Lary 4 = S : S ‘lis ISS. SN N AW RNS Wy ave IN SS Ui NN wEXNS = ie : \S f ba SS % 00, eazi crap Se S Z SS oS < = =WVv Zw tiny, \\\\ oo RE = S ZR Ly > AS aA VSqve =f SS Ss =_ ype =e Sa SY LS OP yl, BWIO"™R ON WT x) AIS =) ZN 2 N wy AN N = S % Wes; Z, =, : IRS S ee Z \V" aa a | § 4 Y << » Fig. 22 Anatomical sketch from Dryander’s edition of Mundinus, 1541 of the work was published, containing 27 illustrations, but the work was never completed. Between 1536 and 1543 there were several plates of anatomical figures placed on the market. These so-called ‘Fliegende Blatter’ include the six Tabule anatomic 984 WILLIAM A. LOCY Mnacowit /Conwafaceur ond belGrabung. i. S “oF = - Fig. 23 Anatomical sketch from Ryff’s Anatomi, 1541 of Vesalius that appeared in 1538 and were a forerunner of his great work of 15438. There were also during the period other anatomists of more than usual insight, as Achillini, whose anatomical treatises were ANATOMICAL ILLUSTRATION 985 not illustrated, and, therefore, do not properly come under con- sideration here. Summary. The chief printed illustrations of anatomy before Vesalius may now be chronologically arranged, omitting different editions of the same work with slight modifications of the figures: 1491. The arrangement of the viscera in the human female in Ketham’s Fasciculus Medicine. 1493. The skeleton of Richard Helain. 1493(?). A demonstration of visceral anatomy in the Meler- stat edition of Mundinus. 1496. The abdominal muscles, in the Conciliator differen- tiarum. 1497. Plate of the Griininger skeleton. 1497. The wounded man with internal anatomy in Brun- schwig’s Chirurgie. 1499. Anatomy of the three body cavities, together with fig- ures of the separate organs, in Peyligk’s Compendicsa. 1501. Similar illustrations in Hundt’s Antropologium. 1503-’04. Organs of the thorax and abdomen in Reisch’s Margarita Philosophica. 1510 (?). Leonardo da Vinci, more than 750 sketches of human anatomy; not, however, published. 1513. Mundinus, zodiacal signs and rough sketch of the heart. 1517. Plates of visceral anatomy and of the skeleton published by Johann Schott. 1518. Skeleton and visceral anatomy in Phryesen’s Spiegel der Artzney. 1521, ’22-’23. Berengarius, commentaries on Mundinus and Isogogee breves. 1536. Fliegende Blatter. 1538. Six Tabule Anatomice of Vesalius. 1537. Dryander, Anatomiae, corporis humani, etc. 1541. Dryander, edition of Mundinus, with 45 illustrations. 1541. Ryff, Anatomi. 986 WILLIAM A. LOCY The survey of these printed sketches of anatomy, covering a century-and-a-half before Vesalius, bringsinto noticethe relatively slow progress. While we remember that this is the period of the awakening of the scientific spirit, still, the drama of intel- lectual progress does not unfold as rapidly as we might expect. Why, after the revival of dissection under Mundinus, and why, especially, after the introduction of printing, was there not more rapid progress? Some seek to find an answer in the difficulty of getting material for dissection and others in the opposition of the church, but the thing that held anatomical science in check, was not so much the lack of opportunity to dissect as the mental habit of the time. The disposition to dissect was not especially strong. That internal hunger for the analysis of nature at first- hand was not of dominating insistence. The effects of tradition and of education had to be overcome, and the gradual assimila- tion of new methods and new ideas was necessarily slow. Those who would have done better under gifted and inspired leaders were perplexed and too closely bound by the mental habit of the time to map out and follow an independent course. Thus, the retarding influence was generic rather than specific. Independent spirits of great originality were rare then, as now, and it seems natural that the habit of imitation should have so long perpetu- ated anatomical sketches of poor quality. Da Vinci was the only man whose product exhibits great originality and independence. His anatomical work was on the plane of that of Vesalius but his sketches were not printed until long after. The practice of dissection by medical men was not so actively opposed by the church as is generally supposed. A superficial reading of the bull of Pope Boniface, de Sepultis, issued in 1300, has led to the statement that it was directed against the practice of dissecting for scientific purposes, but it was, in reality, a pro- scription of the practice of dismembering the bodies of dead Crusaders, in order that their bones might be more readily trans- ported home for burial in consecrated ground. The practice of plagiarism was widespread during this period. Publishers and authors engaged in it in a wholesale way; both sketches and text were commonly copied without credit being ANATOMICAL ILLUSTRATION 987 given. The ethics of the rights of intellectual property were unrecognized. The earliest printed sketches were derived. from manuscript sources and, these, in turn, were based upon the tra- ditional anatomy, chiefly of Galen and his commentators. Now and then a touch of original observation was added to the tradi- tional figures but they were not perfected. _Dependenceon author- ity was still the deep-seated method of the intellectual life, and the rise of independent observation was slow. But, the better intellects were opposing it, and with all these limitations the light of the renaissance was breaking. Dependence on authority was giving way, and, finally, thanks to the work of his predecessors, Vesalius was able to establish a new method based on observation and reason. With the publication of his Fabrica in 1543, there was ushered in the era of good illustrations of anatomy. The prevailing mental habit of the time was now at least partly over- come and the era of independent observation was started. 988 WILLIAM A. LOCY BIBLIOGRAPHY The full titles of the printed books from 1491 to 15438, containing anatomical cuts, and listed in the body of this paper as ‘Sources,’ are often long and cumber- some. They will be found in the Catalogue of the Surgeon General’s Library, in the lists of printed books in The British Museum, in Hain’s Repetorium Biblio- graphicum, in Haller’s Bibliotheca Anatomica, etc. The other books used chiefly as references are: Batt 1911 Andreas Vesalius the Reformer of Anatomy. Baas 1889 Outlines of the history of medicine. New York. Cuievitz 1904 Anatomiens historie. Copenhagen. : CHOULANT 1852 Geschichte der anatomischen Abbildungen. Leipzig. Horr 1904 Die Anfange der Anatomie bei den alten Kulturvélkern, in Ab- handl. zur Ges. der Medizin. Breslau. Paget 1898 Geschichte der Medizin. Berlin. PuscHMANN 1902-05 Handbuch der Geschichte der Medizin. Jena. Roto 1892 Andreas Vesalius Bruxellensis. Berlin. Supuorr 1907-08 Various articles in Studien zur Geschichte der Medizin, Leipzig; and in Archiv fiir Geschichte der Medizin. Leipzig. Torry 1903 Geschichte der Medizin, in Puschmann’s Handbuch. Jena. WeINDLER 1908 Geschichte der gynikologisch-anatomichen Abbildung. Dres- den. Wiecer 1885 Geschichte der Medizin und ihrer Lehranstalten in Strassburg vom Jahre 1497 bis zum Jahre 1872. Strassburg. MINIMAL SIZE REDUCTION IN PLANARIANS THROUGH SUCCESSIVE REGENERATIONS S. J. HOLMES It is a familiar fact that very small pieces of fresh water plan arians will regenerate and give rise to minute individuals closely resembling the original form. I have endeavored to ascertain how far reduction of size in Planaria maculata may be carried without causing a failure to give rise to a normal individual. With very small pieces of an adult planarian there is a large pro- portion of cut surface which produces an injurious effect and there are also various mechanical impediments to regeneration; these facts, combined with the specialized condition of much of the tissue, conspire to restrict the regenerative capacity of the parts. In order to eliminate somewhat these factors the device was re- sorted to of subjecting the animals to a number of successive divisions. A planarian was cut into fifteen or twenty pieces; when these had regenerated into small planarians they were again cut into several pieces. These regenerated into still smaller individuals which in turn were divided, the process being con- tinued until forms were reached which were so small that com- plete regenerations were no longer obtained. In this way, when the minimal size limit was approached, regeneration became very slow and many of the pieces lived for months without restora- tion of the missing parts. Asa general rule it may be said that the smaller the piece the more slowly the restorative processes take place. In this way the proportion of cut surface was reduced, the tissue kept in a more plastic condition, and the whole process of regeneration made easier. Eugene Schultz has studied the reduction in size of planarians from starvation. He found that planarians could be reduced in this way to one-tenth or one-twelfth of their original size. A study of the size of cells of various kinds showed that there was 989 990 S. J. HOLMES little reduction in size either of the cells or their nuclei as a result of starvation; the diminution in the size of the body was produced mainly by their reduction in number. Various organs suffered unequally in this process. Organs of copulation, sex ducts and vitellaria were among the first to disappear, the eyes degenerated, and there was a marked reduction in the number of parenchyma cells. The muscle cells suffered little decrease and the number of muscle bands remained unaltered; there was little reduction in the nervous system. The male sex cells were among the least altered. Cells of the intestinal epithelium and the outer ectoderm, while reduced in number, were not reduced dispropor- tionally to the body as a whole. The study of small regenerated planarians was undertaken in order to ascertain how far the various organ systems would suffer on account of reduction in size and how far the results might be parallel with the effects of starvation. Through successive regen- erations it is possible to carry the reduction very much farther » than can be done by the withdrawal of food. While starvation may reduce the animal to one-tenth or one-twelfth its original size, by the method of successive regenerations it may be reduced to ;ooo OF yso0 its original size. Many of these very minute animals had practically the same form as the adult. Several specimens were sectioned and careful measurements were made of several kinds of cells and compared with measurements of corresponding cells of individuals of ordinary size. Ectoderm cells, parenchyma cells, cells of the intestinal epithelium were of the same size as in the larger worms. The muscle cells, while less in length, were nearly as thick as in the larger worms. The nuclei were also not reduced in size and therefore bore the usual relation to the size of the cells. The gonads, sex ducts, copula- tory apparatus and vitellaria could not be found. The muscu- lar system is well developed, the outer layers being present and only a little thinner than in normal individuals. The number of dorsoventral strands in a cross-section is not more than about one-fourth that of larger specimens and they contain fewer fibers. The alimentary canal has but a very few short branches. The cells are not shrunken as occurs in starved individuals and MINIMAL SIZE IN PLANARIANS 991 cross sections of the diverticula appear much as in the larger worms except for the reduction in the number of cells. The size of the brain and the diameter of the nerve cords bear about the same ratio to the rest of the body as in large individuals. Ina few cases there was but one eye instead of two and this was not a median one as sometimes occurs in small individuals but was in the position of one of the lateral eyes. The eye is of normal size, in relation to other parts and it has essentially the usual structure, but there is a great reduction of the number of the retinal cells. The pharynx, which bears about the same relative proportion to the body as in larger planarians, is composed of the same epi- thelial and muscular layers. The relative proportion of the par- enchyma and digestive organs is little altered in the smaller indi- viduals. The relative thickness of the outer epithelium is how- ever much greater, since it is composed of but one layer of cells which have the same size in the large and the small planarians. Pigment cells occur sparsely scattered over the dorsal surface and appear of enormous size in relation to the rest of the body. On the whole, the small individuals are strikingly like the larger ones in general form and the relative proportions of the systems of organs. | | Observations were made on the movements and reactions of these minute forms. Methods of locomotion, exploring movements of the head, reactions to light and contact, responses to mechan- ical stimuli, righting movements, and various other activities, even down to the most delicate details, were carried out in practi- cally the same way as In individuals of normal size. These facts indicate how effectively the functional unity of the organism is maintained notwithstanding the enormous reduction in the num- ber of its cells. One factor which probably determines the mininal size which may be attained is the fact that the size of the cells cannot be reduced and there must be a certain number of kinds of cells to preserve the physiological unity of the organism. There must be nerve cells, muscle cells, parenchyma cells, epithelium, etc., if the planarian is to be a planarian. The work of the organism, like 992 S. J. HOLMES that of a factory, can be performed on a large or a small scale, but as there are many functions to be discharged, and as one cell cannot do the work of another, a point naturally has to be reached somewhere when a further reduction of the number of cells brings operations to a standstill. Matters might work out, however, in a different way by effecting a general simplification of structure, such as occurs in the reduction of Hydra. This simplification of structure, which has been compared to a reversal of embryonic development, does not proceed in the planarian very far. The loss of sex ducts and associated organs is of doubtful significance, since these parts often atrophy at certain periods in adult indi- viduals. In attempting to carry reduction below a certain size the cut ends of the pieces heal over and there is little further change; the organism does not transform itself into a simple embryonic stage, and we cannot with safety speak of the reversal of developmental processes (if it be really such) beyond perhaps a few retrogressive steps. rs THE GEOTROPISM OF PARAMOECIUM EK. H. HARPER From the Zoological Laboratory, Northwestern University FIVE FIGURES Are there any free-swimming organisms which are oriented to gravity by means of the shape or contents of the body in the same way as the axis of orientation of certain eggs is determined by the difference of specific gravity of the opposite poles? The body of Paramoecium caudatum suggests the possibility of its being oriented to gravity by the difference in buoyancy of the two ends. The posterior end is broader and the anterior end indented by a deep groove. Is the body of Paramoecium ‘stern- heavy,’ and if so, does it account for the ordinary negative geo- tropism of the animal, the anterior end having a tendency nor- mally to point upward? Various explanations have been pro- pounded for the geotropism of Paramoecium, such as difference in water pressure at different depths, difference in pressure on the lower and upper surfaces and finally Lyon,! ascribes the geotropic response to the internal st muli of heavy particles, making the body of Paramoecium the analogue of the statocyst of higher forms. To test the writer’s hypothesis, it was thought it might be possible to increase the difference in specific gravity of the two ends by making the animals ingest heavy particles. These would at first lhe toward the posterior end near the mouth, making the body more ‘stern-heavy,’ and so possibly increasing the negative geotropic tendency. ‘Lyon, E. P., 1905. On the theory of geotropism in Paramoecium: Amer. Journ. Physiol., vol. 14, pp. 421-482. 993 994. E. H. HARPER EXPERIMENTS WITH SUBSTANCES OF HIGHER SPECIFIC GRAVITY The experiments were carried out as follows: The water used was ordinary tap water boiled to drive away gases. , > ON Fig. 7. Diagram V. Scheme of the life history of Adelia, Monas, etc. toma would be graphically shown in diagram IV. Insuchformsas Adelea, Monas and Actinophrys among the protozoa, in no algae so far as I know, reduction occurs during the fusion of the gametes; the life history is given in diagram V (text fig. 7). Nowwhich of these life histories most nearly approximates prim- itive conditions? The question involves a discussion of the rela- tion in phylogeny of the alternation of generations, sexuality and reduction. The evidence already given shows that the latter two phenomena antedate the alternation of generations, for both are found among simple animals and plants that have not-achieved the alternation. THE SPERMATOPHORE IN ARENICOLA 1038 SEXUALITY AND REDUCTION Many biologists are inclined to regard sexuality and reduction as causally related. Thus Davis says ‘‘Chromosome reduction as a physiological process seems to be a corollary of sexual nuclear functions.” (Bot. Gaz., vol. 43: p. 109.) If either were the effect of the other, we should expect reduction to bear a fixed relation to the sexual act. As itis we have seen that it may occur before, during or after such act. It would be better to say, then, that reduction is an adjunct of sexuality and both are probably corollaries of some fundamental cause yet to be determined. It has been suggested by Biitschli and later writers that the union of the gametes is a process of rejuvenescence. In other words, the sexual act is induced by exhaustion of the organism. We know that chromidia are due to degeneration produced by exhaustion and otherwise (Dobell, ’07). Further, a very simple type of reduction has been found in some protozoa (Monas, Adelea, etc.) to consist of the extrusion of chromidia. The very incomplete state of our knowledge regarding reduction, chromidia and sex, particularly among the lower plants and animals, makes the proposal of even a working hypothesis premature. I call attention to the exhaustion of the organism through repeated asexual divisions as a possible cause of both sexual union and reduction, merely to illustrate the possibility that both are corol- laries of some underlying single cause. About the only justifiable conclusion regarding the phylogenetic relation between the sexual act and reduction is that they arose as adjacent phenomena. This seems probable, since in primitive forms, both plants and animals, they occur close together in point of time. It would seem, on a priori grounds, too, that if a sexual union occurred in forms that had been producing asexually, a reduction in the amount of chromatin would promptly occur if it had not occurred previous to the act, so as to restore customary conditions as speedily as possible. 1034 ELLIOT ROWLAND DOWNING THE PRIMITIVE ANIMAL TYPE Botanists seem agreed that the significance of the persistent gametophyte in the higher plants is that it represents a return to a primitive type. The conclusion seems a proper one. We would similarly expect the gametozoic generation, then, to rep- resent a primitive animal type. The spherical spermatophore, arising from a spermatogonium as a spore might cleave, must find its counterpart in the phylogeny of the higher animals as the thallus-like gametophyte of the higher plants indicates a thal- line ancestry for them. Possibly the Volvocales approximate most closely to such an hypothetical ancestral form. The group is one that has been appropriated by many zoologists as showing marked animal affinities, and is admittedly a widely aberrant type by all botan- ists. Volvox reproduces asexually; certain cells of the central cavity, the so-called parthenogonidia, reproduce by fission, and the offspring divide and subdivide, the products cohering in spher- ical masses to form colonies like the parent. These are freed when the parent colony disintegrates. Volvox also reproduces sexually. Certain cells differentiate as eggs and come to lie in the interior of the colony. Other cells, similarly discharged to the interior, develop within them numerous sperm. The sperm break out of their containing cells and fertilize the eggs. The oospore remains inactive for some time as a resting spore but finally develops a colony like the parent. My preparations are not yet sufficiently clear to permit of working out all the chro- matin changes, but lam quite positive that reduction occurs in Vol- vox, 4s it does in animals, during gametogenesis. In the Dicyemidae, looked upon by many, zoologists as the connecting link between the protozoa and the metazoa, the proc- ess of reproduction is suggestively similar, except that the asex- ually produced young bore out from the parent to an independent existence instead of awaiting the death of the parent. The center of the dicyemid is occupied by a cell, in which, besides the nucleus, there are one to several embryo cells from which the asexual individuals arise. These cells are evidently the homologues of THE SPERMATOPHORE IN ARENICOLA 1035 those parthenogonidia found in the Volvox colony. Max Hart- mann, in describing the reproduction of the Dicyemidae, speaks of this central cell as the agametangium because in it develop the asexual, or as he calls them, the ‘agamic’ individuals. After several generations of these agamic individuals, there arise the sexual or the gametic individuals. They arise from the embryo cells (Keimzellen of Hartmann) by a process of cleavage very sim- ilar to the development of the agamic individuals, except that in the female the reproductive cells are split off early. The female as well as the male sexual individual grows within the asex- ual parent. The eggs are developed from the embryo cells of the female. They are freed from the gametic individual into the so- called agametangium of the agamic individual by the death of the gametic form. The egg matures by forming polar bodies and is fertilized by sperm which have discharged from the male gametic individual by its death and disintegration. To harmonize Hartmann’s description with the language I have used in giving the alternation of generations in the Arenicolidae I have only to change his terms slightly; (I have adopted the ter- minology proposed by Beard, cited later). Call his agamic in- dividuals the sporozoon and the Keimzellen spores; his gametic individual, the gametozoon; then the dicyemids make concrete our conception of how the alternation of generations of the higher forms arose from such simple ones as Volvox, even though the line of ascent may not actually have passed through the Dicye- midae. Imagine that in a Volvox-like form the sexual colony or gameto- zoon develops eggs and sperm before it is discharged from the colony which reproduces asexually and the condition of the dicye- mids is practically achieved. Now imagine that a regular alter- nation of generations is established instead of the intercalation of an occasional sexual generation in the midst of the dominant asexual reproduction of the Dicyemidae and a condition is estab- lished that needs little if any modification to give the alternation of generations as we find it in the Arenicolidae, as follows: The adult Arenicola, a sporozoon, corresponds to the sporophyte colony of Volvox or the agamic individual of the dicyemid. At 1036 ELLIOT ROWLAND DOWNING some time in its life history cells are developed within its central cavity—the primitive germ cells or spore mother cells. These develop spores which we know as spermatogonia or oogonia. Such spores, in Volvox, develop new ¢olonies which may be sex- ual and which are freed when the parent disintegrates. In the Dicyemidae, if a gametic individual develop from such spores (or Keimzellen) it is retained in the parent where it disintegrates to free its egg or sperm. In a word the gametozoon degenerates prematurely within the sporozoon. Now in Arenicola, I take it, the development of the spermatogonia into the spermatophores is the development of the gametozoa. Gametozoon degeneration begins before its development is complete and the sperm are pro- duced by a short cut. Instead of developing an individual, with- in which some cells form the clusters of sperm, its cells form the sperm cluster immediately. In this genus, as I have already shown is the case in Hydra, the sperm seems to form within the spermatid, reminiscent perhaps of the primitive condition found in Volvox of forming the sperm within the cell. For the sake of my theory, I should like to agree with Tann- reuther in the spermatogenesis of Hydra. He claims that the spermatozoa develop ‘in groups, each group enclosed within a single cell or cyst. But the clearness of my own preparations seems to negative completely such an interpretation. I can only reiterate what I have already published on the spermatogenesis of this form. It is too bad, for Hydra would make even a better transition form than it does, between such a spermatogenesis as we have in Volvox and the typical animal spermatogenesis in which the spermatid transforms into the sperm in its entirety rather than developing the sperm within it. Hartmann, in the paper on the Dicyemids, describes briefly and figures a development of the fertilized egg that is very similar to the development of the asexual spores. The same parallelism is noted in many botanic papers, the oospore developing much as an asexual spore does in its early stages. To find, then, in Arenicola that the spermatogonia or asexual spores develop to produce the spermatophore or gametozoon in a manner homol- ogous to the development of the fertilized egg is to strengthen the position taken that the spermatophore is the gametozoon. THE SPERMATOPHORE IN ARENICOLA 1037 THE GAMETOZOON, A 2X FORM Botanists, reasoning on the basis of the fact that the disappear- ing gametophyte in the higher forms possesses the haploid num- ber of chromosomes, have been led to assume, not only that the gametophyte represents a reversion to the primitive type, but also that this primitive plant possessed the « number of chro- mosomes. Reasoning in an analogous manner, we should be forced to assume, that, since the gametozoon possesses for most of its life history, the 22 number of chromosomes, so the prim- itive animal type did not have the reduced number. This dis- tinction between plants and animals has long been recognized. “So far as groups of plants above the thallophytes are concerned, the period of chromosome reduction has been found to be always associated with sporogenesis and never with gametogenesis as in the case of animals.’”’ (Yamanouchi; Polysiphonia, p. 43.) The difference may help to trace the gradual separation of the plant and animal types in the course of evolution. Upon this distinction, for instance, Dobell rests his belief in the plant affin- ities of the Phytomonadina (The structure and life-history of Copromonas subtilis, p. 112). The discovery already mentioned that Volvox has reduction occurring during gametogenesis would justify, in a measure, the classification of the form as an animal rather than as a plant. THE COMMON PROTOTYPE Presumably plants and animals have come from a common an- cestor. Now in all higher animals reduction occurs near the close of the gametozooic generation. It occurs, in all higher plants, near the close of the sporophyte generation. In the com- mon ancestor it must have occurred at a point between these two extremes, possibly during or in close connection with the con- jugation of the gametes. Such a possibility is rendered probable from the fact that, in thallophytes and protozoa, the reduction occurs at variable times in the life histories, usually as an adjunct of the union of the gametes, as if that variation were dominant which later becomes fixed in the two prevailing plant and animal types. 1038 , ELLIOT ROWLAND DOWNING ORIGINAL POSITION OF REDUCTION If reduction originally occurred in the primitive common ances- tor somewhere closely adjacent to the union of the gametes, the reduced number of chromosomes would exist for only a short period and might not occur at all. The major part of the life history of the forms would possess the somatic or diploid number. This is now the case for most animals and for many of the algae as already shown. In plants, then, the place of reduction in the life history has been shifted; the phenomenon has been postponed. In animals it occurs much nearer its original position. Reduction shifted Since in practically all animals and in many algae, the phenom- enon of reduction occurs before or during conjugation of the gametes, the preponderance of evidence appears in favor of such a position in the primitive common ancestors of plants and ani- mals. It is all the animals and many algae against the higher plants in favor of such an hypothesis. Text fig. 7, then, might nearly represent primitive conditions. Evidently following the gamete-bearing generation with its definite number of chromosomes would come a spore-bearing generation with the same number of chromosomes. This is true now in the Conjugales, Coleo- chaete, etc., and I believe in Volvocales and the animals. True, in Conjugales, Volvocales, etc., there are several spore-bearing generations following each other in succession. But if, for any reason, the alternation of generations be established by the omis- sion of all but one of these spore-bearing generations, there would be left the gametophyte and sporophyte generations as I conceive them to exist in Arenicola, only that the reduction has shifted from a position like that of fig. 7 of the text to a place before the conjugation of the gametes. For plants the shift in position has been, in the opposite direction—a shift that is seen progressing in text fig. 6 and completed in fig. 5. In Chamberlain’s theory of the alternation of generations in animals, he maintains that the shift in the animal group has THE SPERMATOPHORE IN ARENICOLA 1039 been in the same direction as in the higher plants. As pointed out by Coulter and Miss Pace, the end result attained in plants by the gradual reduction of the gametophyte generation to the point of complete extermination, in such forms as Pandanus, is entirely similar to the condition found in animals. Because the end results are similar is not prima facia evidence that the means of achievement in the course of evolution have been the same; it is very evident that in the higher plants such a condition as is found in Cypripedium, etc., is the result of a reduction of a gametophyte generation with the x number of chromosomes, for all steps in the process are evidently traceable. But nowhere in the animal kingdom, not even among the protozoa, is there any evidence of a corresponding gametozoic generation with a reduced number of chromosomes. REDUCTION AND TETRAD-FORMATION It is true that there is a striking similarity between the formation of the tetraspores in most plants and the development of the spermatids from the spermatocyte. It is rendered doubly suggestive by the fact that reduction occurs during both processes. Yet, even in the plants themselves, we are not warranted in con- cluding that all cells in which reduction occurs are homologous. Reduction occurs without tetraspore. formation in a sufficient number of cases, as in Lemanea, Chantransia, etc., to show that there is no fundamental phylogenetic association involved. The customary appearance of a fourfold division at time of reduction may be based on some fundamental! property of carbon compounds, possibly on the tetravalent condition of carbon itself; in which case it would not be strange to find tetrad formation common in plants and animals without assuming that, when occurring, a morphological homology is indicated. BEARD’S HYPOTHESIS In the preceding pages I have noted one hypothesis of the alter- nation of generations in animals—that proposed by Chamber- lain, and I have given my reasons for discarding it. Another 1040 ELLIOT ROWLAND DOWNING hypothesis of the alternation of generations in animals has been proposed and vigorously advocated by Beard. He recognizes an antithetic alternation of generations in animals. He, too, identifies ‘‘the primary germ cells as the equivalents of the spore- mother-cells of plants.’”’ He regards the larva or phorozoon as the asexual generation, the homologue of the sporophyte. He derives the gametozoon, the adult animal, from it by apospory. He is forced to conclude, then, ‘‘ that the final reduction of chromo- somes has been deferred to a later portion of the life cycle in metazoa as compared with plants.” The same objections apply to Beard’s hypothesis as to Cham- berlain’s, namely, that there is no evidence in fact of the succes- sive steps in the postponement of reduction in animals similar to that so constantly apparent in plants. So far as we know, the process always occurs closely adjacent to the union of the gametes, continuing, in the higher animals, in much the same position in the life cycle that it occupies in the lower animals and plants. Whereas, in plants, it is evidently shifted from this primitive position and the successive steps of the shift are traced with some degree of certainty in living forms. Furthermore, it seems to me, an impossible task to articulate his theory with what we know of the reduction phenomena and the development of the protozoa and mesozoa. He says: The sexual generation of plants is at best a miserable failure from the morphological point of view. . . . The higher one ascends the smaller it bécomes until in the higher plants it has almost reached the vanishing point, without, however, being able to disappear entirely. In the animal it is the larva, the phorozoon, or asexual generation which makes the bravest show in the lower metazoa; . . . In the higher forms it becomes reduced. But how will a relation be established between the larva of the metazoa and the asexual generation of the protozoa? The one . should pass over into the other. It seems unwise to adopt a theory which demands a hiatus at this point, when it is easy to blaze a possible, uninterrupted trail along which evolution may have proceeded by way of such forms as the Volvocales and the Dicyemidae, if we adopt the hypothesis I have proposed. THE SPERMATOPHORE IN ARENICOLA 1041 BIBLIOGRAPHY ALLEN, Cuas. E. 1905 Die Keimung der Zygote bei Coleochaete. Ber. deutsch. bot. Gesells., Bd. 33: p. 286. DeBarry, A. 1878 Ueber apogame Farne und die Erscheinung der Apogamie im Allgemeinen. Bot. Zeit., Bd. 36: pp. 449-487. Berarp, J. 1902 Heredity and the epicycle of the germ-cells. Biol. Centralbl., vol. 22: pp. 321-328, 353-360, 398-408. Bearp, J. anD Murray, J. A. 1895 On the phenomena of reproduction in ani- mals and plants. Ann. of Bot., vol. 9: pp. 441-468. Birscut1, O. 1876 Studien tiber der ersten Entwicklungsvorginge der Hizelle, der Zelltheilung und die Conjugation der Infusorien. Abh.d.Senckenb. naturf. Gesell. Fr. a. M., Bd. 10: pp. 213-452. Ca.xins, Gary N. 1895 The spermatogenesis of Lumbricus. Jour. Morph., vol. 11, no. 2: pp. 271-302. CHAMBERLAIN, C. J. 1905 Alternation of generations in animals from a botanic standpoint. Bot. Gaz., vol 39: pp. 187-144. 1910 Nuclear phenomena of sexual reproduction in Gymnosperms. Am. Nat., vol. 44: pp. 595-603. Cuitp, Cuas. M. 1900 The early development of Arenicola and Sternapsis. Arch. f. Entw. der Org., Bd. 9: pp. 587-717. CouuterR, Joun M. 1908 Megaspores and embryo sacs. Bot. Gaz., vol. 45: pp. 361-366. Coutter, Barnes, Cowies 1910 Text book of botany. Chicago. DanceEarpD, P. A. 1898 Sur les Chlamydomonadinées. C. R. Ac. Sci., Paris, Tome 127: p. 736. Davis, B. M. 1905 On the plant cell, vi and viz. Am. Nat.. vol. 39: pp. 449- 500 and 555-600. 1909 Origin of the Archegoniates. Am. Nat., vol. 43: pp. 107-111. Dosett, C. C. 1907 Physiological degeneration in Opalina. Q. J. Micr. Sci., vol. 51: pp. 633-646. 1908 The structure and life-history of Copromonas subtilis. Q. J. Micr. Sci., vol. 52: pp. 75-120. 1909 Chromidia and the binuclearity hypothesis. Q. J. Micr. Sci., vol. 53: pp. 279-326. Downline, E. R. 1905 The spermatogenesis of Hydra. Zool. Jahrb., Abt. f. Anat. u. Ont., Bd. 21: pp. 379-426. 1909 The connections of the gonadial blood vessels and the form of the nephridia in the Arenicolidae. Biol. Bul., vol. 16: pp. 246-258. 1042 ELLIOT ROWLAND DOWNING Farmer, J. B. anp Diasy, L. 1907 Studies in apospory and apogamy in ferns. Ann. of Bot., vol. 21: pp. 161-199. FauveL, P. 1899 Arenicola ecaudata. Mem. Soc. Sci. Nat. Cherbourg., Tome 3l: pp. 101-186. GaMBLE, F. W. ano AsHwortsH, J. H. 1900 The anatomy and classification of the Arenicolidae. Q. J. Micr. Sci., vol. 43: pp. 419-569. HarTMANN, Max 1906 Untersuchungen iiber den Generationswechsel der Dicy- emiden. Brussels. Also in Mem. of Royal Belgian Acad., N.S., vol. 1. Hertwie, R. 1896 Ueber Kerntheilung, Richtungskérperbildung und Befruch- tung von Actinosphaerium eich., Abh. d. k. bay. Akad. d. Wiss., Miinchen, 11 KI. 19: pp. 1-104. Hoyt, W. D. 1909 Alternation of generations and sexuality in Dictyota dicho- toma. Bot. Gaz., vol. 49: pp. 55-57. Karsten, G. 1909 Die Entwicklung der Zygoten von Spirogyra jugalis Ktzg., Flora, vol. 99: p. 1. Line, Raupw# 8. 1905 The structure and development of the nephridia of Aren- icola cristata Stimpson. Mitth. a. d. zool. Stat. z. Neapel., Bd.17: pp. 341-405. LoBianco, 8. 1899 Mitth. a. d. zool. Stat. z. Neapel., Bd. 13: p. 484. Meyer, Ep. 1901 Studien iiber der Kérperbau der Anneliden, ur. Mitth.a.d. zool. Stat. z. Neapel., Bd. 14: pp. 247-585. Mercatr, Maynarp M. 1908 Opalina. Its anatomy and reproduction, with a description of infection experiments and a chronological review of the literature. Arch. f. Protist. Bd. 13: pp. 195-374. NERESHEIMER, E. 1907 Die Fortpflanzung der Opalinen. Arch. Protistenk., Suppl. 1: pp. i-42. Pace, Lutu 1907 Fertilization in Cypripedium. Bot. Gaz., vol. 44: pp. 353- 374. ProwazEk, 8. von 1901 Kerntheilung und Vermehrung der Polytoma. Ost. bot. Zeitschr., Bd. 51: p. 51, ete. 1901 Flagellatenstudien. Arch. Protistenk. 2: pp. 195-212. ScoenK, H. 1908 Ueber die Phylogenie der Archegoniaten und der Characeen. Engler’s Bot. Jahrb. Bd. 42: pp. 1-87. ScHaupinn, F. 1896 Ueber die Copulation von Actinophrys sol. Sitzber. Akad. Wiss., Berlin. Bd.1: pp. 83-89. STRASBURGER, Ep. 1904 Die Apogamie der Eualchemillen und allgemeine Gesichtspunkte die sich aus ihr ergeben. Jahrb. wiss. Bot., Bd. 41: pp. 88-164. SrepLecki, M. 1899 Etude cytologique et cycle evolutif de Adelea ovata Schnei- der. Ann. Inst. Pasteur, Tome 13: pp. 169-192. THE SPERMATOPHORE IN ARENICOLA 1043 TANNREUTHER, G. W. 1909 Observations on the germ-cells of Hydra. Biol. Bull., vol. 16: pp. 205-209. WituraMs, J. Luoyp 1904 Studies in the Dictyotaceae. Ann. of Bot., vol. 18: pp. 140-160 and 183-204. Wotrsk, J. J. 1904 Cytological studies in Nemalion. Ann. of Bot., vol. 18: pp. 607-630. YAMANOUCHI, SHiaho 1906 The life history of Polysiphonia. 42: pp. 401-449. 1908 Apogamy in Neiphrodium. Bot. Gaz., vol. 45: pp. 289-318. Bot. Gaz., vol. JOURNAL OF MORPHOLOGY, VOL. 22, No 4 PLATE 1 EXPLANATION OF FIGURES 1 Latero-dorsal view of the second left nephridium of A. cristata. X 15. 2 Latero-ventral view of the second left nephridium of A. crist2ta. X 15. 3 Latero-dorsal view of the third left nephridium of A. grubii. X 15. 4 lLatero-ventral view of the fourth left nephridium of A. marima. X 15. 5 Latero-dorsal view of the third left nephridium of A. claparedi. X 15. 6 Latero-ventral view of the third left nephridium of A. ecaudata. X15. The bladder end shows a latero-dorsal view, but the rest of the nephridium is twisted by the weight of the gonad. 1044 THE SPERMATOPHORE IN ARENICOLA PLLIOL ROWLAND DOWNING JOURNAL OF MORPHOLOGY, VOL. 22, NO. 4 1045 PLATE 1 PLATE 2 EXPLANATION OF FIGURES 7 Section through the testis of A. cristata in October. It corresponds to the dotted portion of text fig. 1. The cells are some 7.5 uw in diameter. 8 Section through a testis of A. cristata in February. It corresponds tothe dotted portion of text fig.2. Fibrous degeneration and phagocytosis are apparent. 9 Section through the testis of A. cristata; numerous spermatophores are forming. 1046 9 he Poe oe +h oF 4 ey a Ree ; ie agi THE SPERMATOPHORE IN ARENICOLA . PL. ‘EH 2 . ELLIOT ROWLAND DOWNING s by > ek t * ene ae) WOO OAE YS OSCE AG OK ACOA aig © A VSG OOS : F ay oe LOXE I re Ne per. | a LA } \@) ( a ANSE oe} OMe re) “A J Pee , { ir Joon \ &/ ie ° oe Wiiexk Y ») fe) ® 2 A eden PLATE 3 EXPLANATION OF FIGURES 10 A spermatophore from the body fluid of A. cristata. These cells are the sixth generation of spermatogonia from the primary spermatogonia. Each divides to form the last spermatogonial generation. The cells have an average diameter of 3.5u. All are in the equatorial plate stage. xX 340. 11 A spermatophore from the body fluid of A. cristata. The cells are the sper- matids Just after the division of the spermatocytes of the second division. They are in the early anaphase. X 340. 12 A mature spermatophore of A. cristata. Smear preparation fixed in vom Rath’s fluid and stained in saureviolett and iron haematoxylin. X 625. 18 Body fluid of A. cristata under low power, showing spermatophores in various stages (a, b, ce), and coelomic cells (d), leucocytes and chloragogue cells. > 100. 14 A phagocyte from testis of A. cristata in April. xX 930. Note the ingested cell. 15 An early stage in the formation of a spermatophore, from the body fluid of A. cristata. Diameter of the spermatophore 32 «4; average diameter of the cells An (0. 1048 THE SPERMATOPHORE IN ARENICOLA PLATE 3 ELLIOT ROWLAND DOWNING JOURNAL OF MORPHOLOGY, VOL. 22, NO. 4 1049 PLATE 4 EXPLANATION OF FIGURES 16 Section of a more typical spermatophore. X 1250. 17 A giant spermatogonium. > 1800. 18 A giant spermatogonium dividing, the two-cell stage. 19 A giant spermatogonium dividing, the four-cell stage, polar view. 20 A giant spermatogonium dividing, the eight-cell stage, side view. ] A giant spermatogonium dividing, the eight-cell stage, polar view. 2 A giant spermatogonium dividing, the sixteen-cell stage, polar view. 23 A later stage in segmentation of a spermatogonium, a spermatophore from the body fluid of A. cristata. 24 The ‘invagination’ of the forming spermatophore in A. cristata. 25 Section through the saucer-shaped spermatophore of A. cristata. The cells are spermatids, 1} X 4yu. 1050 THE SPERMATOPHORE IN ARENICOLA _ ELLIOT ROWLAND DOWNING JOURNAL OF MORPHOLOGY, VOL. 22, No. 4 1051 PLATE 4 ery SUBJECT AND AUTHOR INDEX LLEN, Bennet M. The origin of the et ar of Amia and Lepidosteus. poe ond Lepidosteus. The origin of the sexes Ampelophila. The effects of inbreeding and selec- tion on the fertility, vigor and sex-ratio of Dro- sophila. 123 Anatomical illustration before Vesalius. 945 ANDREWS, E. A. Male organs for sperm-transfer in the crayfish, Cambarus affinis; their struc- ture and use. 239 Animal geography. Physiological 551 Ant-colony as an organism. The 307 Arenicola and a theory of the alternation of genera- tions in animals. The formation of the sperma- tophore in 1001 Armadillo quadruplets: a study of blastogenic vari- ation. The limits of hereditary controlin 855 Asymmetry in the development of the serpulid, Hydroides dianthus. Experiments on the con- trol of 927 IOGRAPHY: Charles Otis Whitman XV Birds. On the olfactory organs and the sense of smell in 619 Blastogenic variation. The limits of hereditary control in armadillo quadruplets:astudy of 855 AMBARUS affinis; their structure and _ use. Male organs for sperm-transfer in the crayfish 239 Cell-division. The action of salt solutions followed by hypertonic sea-water on unfertilized sea- urchin eggs and the réle of membranes in mitosis. The physiology of 695 Changes in the relative weight of the central nervous system of the leopard frog. On the regular sea- sonal 663 Chemistry of the white and yellow yolk of ova. On the formation, significance and 455 Chickens. The transplantation of ovaries in 111 Cuitp, C. M. The regulatory processes in organ- isms. 171 Chromosomes. A review of the chromosomes of Nezara; with some more general considerations. Studies on 71 Coelenterate ontogeny. Some problems of 493 Control in armadillo quadruplets: a study of blas- togenic variation. Thelimitsof hereditary 855 Control of asymmetry in the development of the ser- pulid, Hydroides dianthus. Experiments on the 927 Craig, WALLACE. Oviposition Induced by the male in pigeons. 299 Crayfish, Cambarus affinis; their structure and use. Male organs for sperm-transfer in the 239 Curtis, WINTERTON C. The life history of the Scolex polymorphus of the Woods Hole region 821 Cyclosalpa affiinis (Chamisso). The growth and dif- ferentiation of the chain of 395 AVENPORT, C. B. The transplantation ae ovaries in chickens. 111 Differentiation of the chain of Cyclosalpa affinis (Chamisso). The growth and 395 Donaupson, Henry H. On the regular seasonal changes in the relative weight of the central nervous system of the leopard frog. 663 Downine, Evtior Rowrianp. The formation of the spermatophore in Arenicola and a theory of the alternation of generationsinanimals. 1001 Drew, Giuman A. Sexual activities of the squid, Loligo pealii (Les). 327 Drosophila ampelophila. The effects of inbreeding at selection on the fertility, vigor and eee ° GGS and the réle of membranes in mitosis. The action of salt solutions followed by hypertonic sea-water on unfertilized sea-urchin. 695 Euschistus. The spermatogenesis of 731 ERTILIZATION in Nereis. Studies of 361 Frog. On the regular seasonal changes in the relative weight of the central nervous system of the leopard 663 ASTROPODS. The mechanism of locomo- tion in 155 Geography. Physiological animal 551 Geotropism of Paramoecium. The 993 Growth and differentiation of the chain of Cyclosalpa affinis (Chamisso). 395 Guinea pig. The cyclic changes in theovary of ee 37 ARGITT, CHartes W. Some problems at coelenterate ontogeny. 493 Harper, E. H. The geotropism of Paramoecium 993 Hemipteron, Euschistus. The pate rae at an Hereditary control in armadillo quadruplets: a study of blastogenic variation. The limits of 885 Ho.umes, S.J. Minimal size reduction in Planarians through successive regenerations. 989 Hydroides dianthus. Experiments on the control of asymmetry in the development of 927 | pees selection onthe fertility, vigor and sex ratio of Drosophila ampelophila. The effects of 123 My hares Myrtue E., W. E. Rirrerand. The growth and differentiation of the chain of Cyclosalpa affinis (Chamisso). 395 | Pater The origin of the a of Amia and Linun, Frank R. Biography: Charles Otis ae man XV Studies of fertilization in Nereis, 361 Liturn, Rautpw’S. The physiology of cell-division. IV. The action of salt solutions followed by hypertonic sea-water on unfertilized sea-urchin eggs and the réle of membranes in mitosis 695 Locomotion in Gastropods. The mechanism of 155 Locy, Wirt1am A. Anatomical illustration before Vesalius 945 Logs, Leo. The cyclic changes in the ovary of the guinea pig 37 Loligo pealii (Les). Sexual activities of the squid 327 1053 1054 EMBRANES in mitosis. The action of salt solutions followed by hypertonic sea-water on unfertilized sea-urchin eggs and the role fe) Mitosis. The action of salt solutions followed by hypertonic sea-water on unfertilized sea-urchin eggs and the réle of membranes in 695 MornxHavs W.J. The effects of inbreeding and selection on the fertility, vigor and sex-ratio of Drosophila ampelophila. 123 MontTGoMERY, JR., THos. H. The spermatogenesis of an hemipteron Euschistus 731 EREIS. Studies of fertilization in 361 Nervous system of the leopard frog. On the regular seasonal changes in the relative weight of the central 663 Newman, H. H., and J. Taomas PaTtTERSON. The limits of hereditary controlin armadillo quadru- plets: a study of blastogenic variation 855 Nezara; with some more general considerations. A review of the chromosomes of 71 LFACTORY organs and the sense of smell in birds. 619 Ontogeny. Some problems of coelenterate 493 Organs for sperm-transfer in the crayfish, Cambarus affinis; their structure and use. 239 Ova. On the formation, significance and chemistry of the white and yellow yolk of 455 Ovaries in chickens. The transplantation of 111 Ovary of the guinea pig. The cyclic changesin the 37 Oviposition induced by the male in pigeons. 299 ARKER, G. H. The mechanism of locomotion in gastropods 155 PaTTERSON, J. THomas, H. H. Newman and. The limits of hereditary control in armadillo quadruplets: a study in blastogenic variation 855 Paramecium aurelia and Paramecium caudatum 223 Paramoecium. The geotropism of 993 Physiological animal geography. 551 Physiology of cell-division. IV. The action of salt solutions followed by hypertonic sea-water on unfertilized sea-urchin eggs and the réle of membranes in mitosis. The 695 Pigeons. Oviposition induced by the male in 299 Planarians through successive regenerations. Min- imal size reduction in 989 EGENERATIONS. Minimal size reduction in Planarians through successive 989 Regulatory processes in organisms. 171 RIpDLE, Oscar. | On the formation, significance and chemistry of the white and yellow yolk of ova 455 Ritter, W. E.,\and Myrtte BE. Jonnson. The growth and differentiation of the chain of Cyclo- salpa affinis (Chamisso). 395 | | INDEX So polymorphus of the Woods Hole region. The life history of the 821 Sea-urchin eggs and the réle of membranes in mitosis. The action of salt solutions followed by hypertonic sea-water on unfertilized 695 Selection on the fertility, vigor and sex-ratio of Dro- sophila ampelophila. The effects of inbreeding and 123 Sense of smellin birds. On the olfactory organs and the 619 Serpulid, Hydroides dianthus. Experiments on the control of asymmetry in the development of the 927 Sarin of Amia and Lepidosteus. The origin a the Sex ratio of Drosophila ampelophila. The effects of inbreeding and selection on the fertility, vigor and 123 Sexual activities of the squid, Loligo pealii (Les). 327 SHELFORD, Victor E. Physiological animal geog- raphy. 551 Size reduction in planarians through successive regenerations. Minimal 989 Spermatogenesis of an hemipteron, Euschistus. The 731 Spermatophore in Arenicola and a theory of the alter- nation of generationsinanimals. The formation of the 1001 Sperm-transfer in the crayfish, Cambarus affinis; their structure and use. Male organs for 239 Squid, Loligo pealii (Les). Sexual activities of the 327 Strona, R.M. Onthe olfactory organs and the sense of smell in birds. 619 RANSPLANTATION of ovaries in chickens. i Nearer The limits of hereditary control in armadillo quadruplets: a study of blasto- genic 855 Vesalius. Anatomical illustration before 945 ; EIGHT of the central nervous system of the leopard frog. On the regular seasonal changes in the relative 663 WHEELER WiLLIAM Morton. The ant-colony aa an organism. 307 Whitman, Charles Otis: Biography. Xv Witson, Epmunp B. Studies on chromosomes. VII. A review of the chromosomes of Nezara; with some more general considerations. 71 Wooprurr, LoRANDE Loss. Paramecium aurelia and Paramecium caudatum. 223 OLK of ova. On the formation, significance and chemistry of the white and yellow 455 ELENY, Cuarurs. Experiments on the con- trolof asymmetry in the development oi the serpulid, Hydroides dianthis. 927 ie peas yk oats es ee eit Spi > Tens at ig z ee pe ii app ae ae te ate ee te ee = Se P a - ~ ' pS =e _—, a ee < mina ge ee MGS ag > oy are ate —— Se en a ~~ So Se se SOS aes gl ae ee ee ee nS ~ as = ~ = zi ee 4 tat Bia ne = epee : ~~ —_ = = er 2 Ao _ — y artes el “ - _ = ne —— ———s < a = — = —_ a = oot ees il ee Oe ee - ee ~ _ ~ ~ - ~ ~ a ne io - om Sapo moe a a ee aes mm eae pr a. Peay ~ ft lh lg POPE ILE PIO a ee ea fo eth dj ca nr airs nia Ce iz 45 Aes wr AON Fy Asha da, = i iQ - ai ' wt n - fad : “4 “it SOU VE RIN eect ing Nive te shares temas tasch Tas nee te ate Mee Tht at Ah aa 2 Sine na ne niet inti n Ten Fhe WD's SANE TRO San cee MRD aA CT LARA we ad RS Nene a SS gS SE