Talk:Journal of Morphology 22 (1911)

From Embryology



From the Department of Anatomy, University of Wisconsin



There has been an increasing amount of attention given in the last few years to a study of the origin and migration of the sexcells of the vertebrates. The number of forms in which this subject has been studied is being constantly extended. While much conclusive work has been done, upon the history of these cells in the elasmobranchs, and an equal amount in tracing them in the teleosts, up to this time, they have never been studied in the ganoids. This work was begun over two years ago, and was reported at the 1908 meeting of the Association of American Anatomists in New York. (Allen, '09.)

The material for the present work is obtainable in great plenty within less than half a mile of the grounds of the University of Wisconsin. Since the breeding habits of these two fishes have been thoroughly treated by other writers, it is not necessary to redescribe them. Telleyesniczky's bichromate-acetic fluid and Zenker's fluid have been used as fixing agents and have proved in every way satisfactory. One secret of obtaining good sections is to secure most thorough infiltration by placing the material in a solution of paraffine in turpentine. While turpentine has a bad reputation, no deleterious effects were noted in the course of the work. Paraffine sections were made without difficulty 7m and 10/^ thick, and were stained, for the most part, in haemalum and orange G. Heidenhain's iron haematoxylin stain was sometimes used for the later stages, but showed no superiority over the haem-alum. In the earher stages of development it could not be used at all, owing to the deep stain that it gives the yolk material.

With the abundance of material and the amount of time given to the work, it was possible to make a careful study of a large series of stages, much larger than it has been found necessary to use in the preparation of this paper.

It is not necessary to enter into a detailed account of the earlier work upon the origin of the sex-cells, because that has already been done in earlier writings. Since the author's articles on the sex-cells of Chrysemys and of Rana, some important papers have made their appearance, which, with one exception, bear out in a most gratifying manner the conclusions expressed by the writer in the two papers mentioned above and in somewhat less confident manner in the earlier writings of Wheeler, Woods and Beard.

These papers will be discussed in the light of the facts set forth in this paper in the last part of this article, since they are to be considered in a more or less controversial manner.


Lepidosteus, 4 '^nm. total length. Cells which appear to be sexcells lie in the ventral portion of the single layered gut entoderm. They can be but dimly distinguished from the other cells of the entoderm among which they lie. They have a more spherical shape than the other entoderm cells, never being flattened as the neighboring cells frequently are. Another difference lies in the fact that the sex-cells contain more and decidedly larger yolk spherules than do the adjoining entoderm cells. Unfortunately these differences are masked by the large quantities of yolk found in the entoderm at this stage. This is true to so great an extent than one can not be certain at this stage as to the identity of the cells in question. At this period the hind gut has a much greater diameter than it has at later stages. At a point onequarter the distance from its cranial to its caudal end it has a dorso-ventral dimension of .24 mm. and a transverse diameter of .20 mm.

Lepidosteus 6.8 mm. total length. In a similar part of the hind gut of a specimen 6.8 mm. in length the dorso-ventral dimension of the gut endoderm is .084 mm. while the transverse dimension is .056 mm. The total length of the hind gut in this later stage is 1.70 mm. as compared with .76 in the 4 mm. embryo. It is seen that there has been a very decided diminution in the diameter of the gut, and, furthermore, that this is out of proportion to the increase in length. It is correlated with a thickening of the gut wall, due to the drawing together of the component cells.

In the 4 mm. stage the gut entoderm was composed of a single layer of cells, while in the 6.8 mm. stage under consideration its lateral and ventral portions are made up of two, and in some places three irregularly arranged layers of cells, while the dorsal wall is made up of a single layer as in the 4 mm. stage. Two series chosen from several of this stage may be taken as showing typical differences. Both show an advance over the preceding stages in the greater ease with which the sex-cells may be distinguished. This is due to the contrast between the ordinary entoderm cells in which a considerable amount of the yolk material has been absorbed and the more rounded yolk-filled sex-cells. In neither embryo has the process of sex-cell migration commenced. This is clearly evident in one, while in the other there are a very few scattered yolk-filled cells of rather problematical character in the loose mesenchyme above and at the sides of the gut entoderm. One striking individual difference between • these two specimens is found in the fact that while in one the sex-cells have retained their primitive position in the ventral and lateral portions of the gut entoderm, in the other they have migrated up into its dorsal portions. Although it is somewhat difficult to establish with absolute certainty this migration from the ventral to the dorsal portions of the gut entoderm owing to the difficulty of distinguishing the sex-cells in the preceding stages, the individual differences in this regard observed in this stage, together with the fairly reliable observations upon earlier stages seem to point to an actual migration of this character.


Lepidosteus 7.3 mm. total length. In the single series of 7.3 mm. embryos the sex-cells in general still occupy the ventral and lateral portions of the gut entoderm, having come to lie in the dorsal wall at only a few points, especially toward the cranial end of the hind gut. A very few sex-cells are found to have migrated into the loose mesenchyme between the gut entoderm and lateral plates of mesoderm, occupying positions in it lateral and dorsal to the latter. These migrant cells are merely the precursors of a general migration which does not become conspicuous until the embryo has reached a length of 8.5 mm.

TABLE 1 Number of sex-cells in Lepidosteus





6.8 8.6

No count












9.3 C




133 125





















No count











Int. — Intestine.

Root — Root of mesentery.

R. — Right sex-gland. L. — Left sex-gland.

Lepidosteus, 8.6 vim. total length. Passing over several intermediate stages studied, the conditions found in a specimen 8.6 mm. long may be described. At this stage the lateral plates of mesoderm are just beginning to split and to form the coelomic cavities (fig. 7). The interval between the plates is filled with loose mesenchyme, of which that portion lying between the gut entoderm and the aorta will later be condensed by the apposition of the lateral plates of mesoderm in formation of the mesentery.


It will be seen by comparing the figures drawn to scale that the mesentery is at this stage not only relatively but actually much thicker than it is during later stages. The accompanying table serves to show the number of sex-cells and their distribution in certain stages.

In this specimen 73 sex-cells have already migrated out of the gut entoderm into the surrounding mesoderm tissues. While .most of them have migrated upward into the loose mesenchyme and splanchnopleuric mesoderm of the anlage of the mesentery, a few have passed laterally into the portions of the splanchnopleuric mesoderm that enter into the formation of the intestinal wall. While the migration of the sex-cells is seen to be well under way at this stage, the great majority of them still remain in the dorsal portion of the gut entoderm. Very few, indeed, are to be found in the ventral half at this stage. The sex-cells of the gut entoderm are easily distinguishable from the other entoderm cells; the latter have lost very nearly all of their yolk material and have become cylindrical in shape. These features stand out in sharp contrast to the large yolk content and spherical form of the sex-cells.

The migration of sex-cells from the gut entoderm into the mesenchyme dorsal or lateral to it may be clearly seen at a few points, as illustrated in figs. 7 and 8. They retain for the most part their spherical form, but cases like that shown in fig. 7 can be readily found. The shape of this sex-cell clearly indicates the mode of progression. They, undoubtedly, pass through the loose network of mesenchyme by an amoeboid movement, however slow or intermittent it may be.

In this stage sex-cells are found in the hind gut from its cranial end to within .2 mm. of the cloaca, a distance of 2.6 mm.

Lepidosteus 9.2 mm. long. The number of sex-cells that have migrated out of the gut entoderm is 136 in this specimen. The number of these is still increasing but solely by migration from the entoderm, since there is no evidence of division of the sexcells during these stages of sex-cell migration.

In this stage the coelomic cavities have appeared in the dorsal portion of each lateral mesodermal plate and the mesentery is


consequently much more clearly defined. In three specimens 9.3 mm. long the following counts of sex-cells outside of the gut entoderm were made:

A = 41 B = 311 C = 425.

A and C are extreme cases, indicating that the process of migration is an irregular one in point of time. The mesentery in specimen A is .46 mm. wide.

In specimen C those sex-cells destined to migrate out of the entoderm have for the most part already done so, while in A, an embryo of the same stage, the process is just beginning. The coelome is least developed in A and furthest advanced in C. This would indicate that the extent to which the migration of sex-cells has been carried on is correlated with the degree of development of the mesentery, resulting from the enlargement of the coelomic cavities. WTiile these three specimens belong, no doubt, to slightly different stages of development, they were very carefullj^ matched as to length, and are most certainly of the same age.

Lepidosteus 10.7 mm. total length. At this period of development, the mesentery is well formed, being much thinner (.18 mm.) than in the 9.3 mm. stage. This results in its possessing a denser texture (fig. 9). The great majority of the sex-cells are scattered through the mesentery, showing no definite arrangement; but lying for the most part in the mesenchyme enclosed between the somewhat denser splanchnic laj'ers of mesoderm. A few are found in the mesodermal layers of the intestine, while a fairly considerable number have remained in the gut entoderm. At this time such sex-cells as are found in any but the dorsal wall of the intestine, at its junction with the mesentery, are most probably destined to remain in their present positions. A few of the sex-cells have migrated to places immediately dorsal and lateral to the root of the mesentery. The latter may be considered to have reached the sex-gland anlagen, although their position relative to the root of the mesentery will be shifted, as we shall see, in the later stages, probably by a general shifting of the tissues in which they lie.


The number of sex-cells which have migrated from the entoderm is found to be 674. It can be fairly taken as the number destined to undergo migration from the entoderm in this particular individual. Those still remaining in the entoderm number 133.

A few scattered sex-cells are found as far forward as the cranial end of the hind gut. The latter is 3.41 mm. in length. Opposite to the cranial portion of the hind gut the sex-cells are rather sparse, increasing in number as one follows the series caudally. The}'- become most numerous a short distance caudal to a point two-thirds the distance from the cranial to the caudal end of the hind gut. The last one in the mesoderm is found at a distance of .46 mm. from the cloaca, and the last one in the entoderm lies at a point .27 mm. from the cloaca.

Lepidosleus 12 mm. total length. At this period migration of the sex-cells has progressed to the point where most of them have reached their final positions. They are still to be seen in the entoderm. This number (125) is quite close to that (133) of the similarly situated sex-cells of the preceding stage. The density of the mesodermal tissues surrounding the gut entoderm makes it seem quite unlikely that any more sex-cells could migrate into them from this source.

The distribution of the sex-cells is as follows:

392 outside of sex-gland anlagen.

Gut entoderm 125

Mesoderm of intestinal wall and Mesentery 104

Root of mesentery between sexgland anlagen 163

Right sex-gland 180l

Left sex-gland 179/ ^59 ^ sex-gland anlagen.

Total 751

The table may be allowed to speak for itself. The sex-gland anlagen grade into one another by an intermediate region at the root of the mesentery. More or less arbitrary limits had to be assumed to distinguish between these three regions. In later stages, illustrated by the 17 mm. stage, fig. 11, we shall see


that the sex-cells undergo lateral migration, either apparent or real, so as eventually to lie at some distance on each side of the median point.

The narrowest portion of the mesentery is at about one-quarter the distance from its origin to its insertion. Its minimum width, as measured here amounts to .028 mm., thus showing a great reduction as compared with the 10.7 mm. stage. This reduction in width is shared by the entire mesenterj', certain regions remaining broad only on account of the enclosed blood vessels. No doubt the migration of the sex-cells out of the mesentery is in large part responsible for this, but a considerable share of it must be ascribed to the fact that there has been a tendency for the tissues to become more compact.

The total length of the hind gut at this stage has reached 4.03 mm. Sex-cells are found in the entoderm at its cranial end, and from there extend to within 0.62 mm. of the cloaca. The distribution of sex-cells within the sex-gland anlagen is somewhat more restricted, since they extend from a point 0.31 mm. caudal to the beginning of the hind gut, to a point 1.00 mm. cranial to the cloaca. The}' are rather sparse at these two extremes.

As in the preceding stages, there is no clear evidence of division of the sex-ceUs, although one can not be absolutely certain upon this point. While at this time many are free from yolk material, others show but little diminution of it. It is true that the sexcells are often found arranged in clusters, but there is no evidence to show that these are due to repeated division of a parent cell rather than to a tendency for them to congregate through mutual attraction. ^Yhsit the nature of this attraction might be, we do not know; but it might well be akin to that influence which causes the sex-ceUs to migrate toward the sex-gland anlagen from their source. Similar clusters of sex-cells were found in earty stages in Chrysemys.

Lepidosteus 14-1 inm. total length. Little radical change is to be seen in this stage. The sex-cells were counted and gave the following results:


Gut entoderm 128

Mesoderm of intestinal wall and

mesentery 222

Root of mesentery between S. G.

anlagen 37

Right sex-gland 197 \

Left sex-gland 153 /

Total 737

387 outside of sex-gland anlagen.

350 in sex-gland anlagen.

There is a strikingly close correspondence between the results of the count hi this specimen and those in the preceding one. Attention may be called to the fact that in this specimen a materially greater number of sex-ceUs is found in the right sexgland than in the left. At the same time there is a very close correspondence in the total number of sex-cells that have reached the sex-gland anlagen as compared with the total number in the 12.0 mm. stage (359).

Lepidosteus 17 mm. total length. In this specimen those sexcells destined to occupy the sex-glands are seen to have migrated some distance to each side of the root of the mesentery, fig. 11. Their position relative to the root of the mesentery and to the Wolffian duct varies at different points along the sex-gland anlage. In the most cranial portion of the latter they lie just medial to the Wolffian duct. As one follows the sex-glands caudally, the sex-cells are found to lie closer and closer to the mesentery, being situated midway between the latter and the Wolffian duct in the middle region of the sex-gland anlage. The most caudally situated sex-cells lie close to the root of the mesentery.

In this and the succeeding stages the intestine had become so voluminous as to make the counting of the sex-cells in its walls very difficult and inaccurate. It is in fact not easy to distinguish them from the cells of the gut entoderm because of their rather small size and their entire lack of yolk material at this stage.

The total number of sex-cells in the sex-gland anlagen of this specimen is rather high, there being 235 in the left sex-gland and 262 in the right. The total number is 497.


The slightly greater number of sex-cells in the sex-glands of this specimen as compared with that in the previous ones is of little significance. It most certainly does not indicate that there has been any extensive division of them. In a previous work upon Chrysemys, (Allen '07), it was shown that there was an extreme amount of individual variation in the number of sex-cells. This variation in Lepidosteus is relatively slight compared with that observed in Chrysemys. In a specimen slightly older than this stage (18 mm.) there were 171 sex-cells in the right sex-gland, and 173 in the left one, the total number, 344, being not far from the average.

In these two stages, 17 and 18 mm., the sex-cells usually occur singly, although in places they are aggregated into clumps so thick as sometimes to show as many as five or six in a section of one of the sex-glands. Whether the sex-cells occur singly or in clumps, they are surrounded by peritoneal cells which contribute materially to the formation of the ridge-like anlage of the sex-gland.

Lepidosteus 21+ mm. total length. In a specimen of this length, fig. 12, there is no essential advance in the development of the sex-gland. There were 147 sex-cells in the right sex-gland, and 154 in the left one. The total number, 301, is distinctly below the average.

Comparison with other forms leaves no room for doubt as to the identity of these sex-cells. Since the aim of this paper is merely to trace out their origin, we will not follow them through later stages in their history, but will describe the conditions found in a specimen 110 mm., in length, fig. 13. A complete series of sections through the sex-gland region of this specimen was not made, so it is impossible to give a full account as to the number of sex-cells and general condition of the sex-gland at this time. In running through the series one is struck with the sparseness of the sex-cells. Never are more than two or three to be found in a single section, and often none at all. This would lead one to infer that there has been little or no multiplication of the sexcells even at this late stage of development.



A glance at table 2 shows that there is a general tendency to a reduction in the average size of the cell body in the later stages. This may be due to the absorption of the contained yolk material. There is no marked change in the size of the nucleus.


Dimensions of sex-cells of Lepidosteus







mm. 8.6














13.59 .






















Amia 4- fnm., total length. In the text figure A is shown a transverse section of an Amia larva of this stage. It will serve as a starting point from which we shall proceed to consider still earlier stages in tracing out the earliest phases in the origin and migration of the sex-cells. The section shown is taken "just anterior to the hind gut, the gut entoderm being clearly marked by its greater thickness and dorsal curvature. The cavity of the intestine at this point opens into the large sub-germinal cavity. The extra embryonic portions of the entoderm, i.e., those which do not form part of the anlagen of the alimentary tract and its appendages can logically be divided into four different regions : (1) The roof of the sub-germinal cavity which is distinguishable from the gut entoderm, as indicated; (2) The layer forming the floor of the sub-germinal cavity; (3) The peripheral layer of entoderm lateral to the sub-germinal cavity (peripheral entoderm) ; (4) The central yolk mass, or vitellus (vitelline entoderm). In the first three of these regions the cells are arranged in a single layer. They are characterized by the fact that the yolk spherules of the component cells are distinctly smaller than are those of



the vitellus, their diameter being from one-quarter to one-half of that of the typical spherules to the vitellus. In the latter cells are scattered a few of these smaller yolk spherules; but the distinction between the first three divisions of the entoderm and the vitellus is a very sharp one.

Perjph. C

Text figure A

In connection with this distinction it is interesting to note that the yolk spherules along the cleavage planes that cut through the vitellus are found to belong to this small type. It is easy to see that if the vitellus were cut up into cells as small as those comprising portions 1, 2, and 3 of the entoderm, the thickness of the layers of small spherules which form merely a border to the large cells would be so great as to comprise the entire body of the more finely divided ones. This difference in the size of the


yolk spherules is then probably associated with the difference in the size of the cells. The peripheral entodermal layer which we have designated as division three is interrupted lateroventrally by blood vessels lying in the mesoderm.

The lateral plates of mesoderm have long since broken away from the mesoblastic somites. Their inner margins lie at some distance to each side of the median line. While there is the slightest tendency in places for the splanchnic and somatic layers of mesoderm to split apart along the medial margins of the lateral plates, the remainder of the lateral plates show no indication of a splitting, even in the arrangement of the nuclei. It is, however, quite probable that such a plane of cleavage is already laid down. This is shown by the sex-cells (text fig. A) being imbedded in the lateral plates. When the somatopleure and splanchnopleure separate later, these will be found to lie in the coelomic cavity, being for a time merely adherent to the coelomic surface of the medial portion of the somatic mesoderm. One can fairly assume that during the period of migration, represented by fig. 5, the sex-cells push their way between the two layers of mesoderm following the potential cleft that separates them.

Text fig. A is very suggestive, as it shows sex-cells situated at intervals from a point just beyond that at which the roof and floor entoderm join the peripheral entoderm. The path of their migration is thus clearly marked out. In this figure it should be noted that the most laterally situated sex-cell lies in the entoderm, while all of the others are clearly imbedded in the lateral plates of mesoderm as already indicated. In but one or two of the many specimens examined were there any sex-cells found in the roof or gut entoderm. They arise in the peripheral entoderm from which they migrate into the lateral plates of mesoderm and through them to their medial borders, whence, as I shall later show, they pass into the sex-gland anlagen after the formation of the coelomic cavity.

The total number of sex-cells found in the mesoderm of the specimen of this stage was 87. Of these 40 were found on the right side and 47 on the left. Text fig. A will indicate their distribution.



Table 3 serves to show for purposes of comparison the numbers of sex-cells found in different specimens of Amia.

TABLE 3 Number of sex-cells in Amia calva







L. Total










None 1












4 1 11



















17 31




11 33




18 66




28 67




26 76




34 49



41 i 103






















40 ■


























































This stage is a convenient starting point from which to proceed in the study of earlier stages.


A7?iia 3.7 mm. total length. The conditions are, in the main, quite similar to those found in the 4 mm. stage. In one of the two specimens (B) in which the sex-cells were counted there were 92 sexcells in the mesoderm and 10 in the entoderm. Although this total number of 102 is greater than the number found in the 4 mm. stage (87), yet, as shown in table 2, no significance is to be attached to this on account of the great individual variation in the number of sex-cells observed, not only in Amia, but also shown by the author to be so obvious in the turtle, Chrysemys. In A of this stage, 72 sex-cells were found, 42 on the right and 30 on the left side.

Amia 3.5 mm. total length. Two larvae of this stage were studied. It was rather difficult to measure the specimens accurately, owing to the fact that the caudal portion of the body free from the yolk has a strong ventral bend. It can be straightened out only in later stages. The two specimens of this length were taken from the same nest and both are distinctly younger than the preceding, yet they showed decided differences from one another in the positions occupied by the sex-cells, probably owing to the fact that this, in all likelihood, is the period of their most active migration. In specimen A the sex-cells are quite numerous in the portion of the lateral plate of mesoderm, which lies immediately above the border of the subgerminal cavity. They occur in fair numbers in the mesoderm between this region and a point one-half the distance from this point to the median edge of the lateral plate of mesoderm. Only three were found nearer the median line than this. Of these, one had scarcely passed the midway point, one was still some distance from the median edge of the lateral plate, while one had actually reached that point.

In specimen B of this stage a large proportion of the sex-cells have reached the median edge of the lateral plate of mesoderm of each side. This is especially noticeable on the right side. The conditions in this specimen approach those described for the 4 mm. stage but do not show quite such an advanced condition, owing to the fact that a larger proportion of sex-cells are scattered along the outer portions of what we may call the sex-cell path. There



were noted two or three instances in which the sex-cells were migrating from the peripheral entoderm into the mesoderm.

Amia, 3 mm., total length; 155 hours. In a specimen of 3.0 mm. total length, the free caudal portion has but recently separated from the vitelline mass, and has attained a total length of .56 mm. By comparison with a number of embryos of 132, 137, and 147 hours old, the age of this embyro was estimated to be very close to 155 hours. This estimate was made by counting the number of sections passing through the posterior part of the embryo free from tJie yolk mass. Sufficient numbers of embryos were used to give a fairly accurate determination, there being seven specimens of the 147-hour, three of the 137-hour, and two of the 132-hour stages studied.


The numbers of sex-cells in each were as follows:




In A

39 59




In B


There were 49 sex-cells counted in the 3 mm., 155-hour embryo. This, it will be seen, is decidedly below the average and yet the number is greater than that found in the 5 mm. stage and in the much later 16 mm. specimens.

Only two of the sex-cells have migrated a very short distance along the lateral plate of mesoderm, beyond a point overlying the lateral boundary of the subgerminal cavity; the remainder of them all lie lateral to it. It will thus be seen that they show a much earlier phase of migration than that observed in the 3.5 mm. embryo, not only as regards the number that have migrated into the mesoderm, but likewise in the distance through which they have travelled in their journey in that layer toward the sexgland anlagen.

Amia, 11^7 hr. stage. That there is a great amount of individual variation in the rapidity with which this migration from the peripheral entoderm to the lateral plates of mesoderm is accomplished may be readily seen by referring to the numbers counted


in the mesoderm of seven specimens of the 147 hour stage. These specimens were all taken from the same nest and kept in the same dish, so there can be but very slight difference in their ages, due, if it exists, to the small difference in the time at which the eggs were laid. It will be seen that the total number of sexcells in the entoderm in these specimens varies from 22 to 76. The latter number is not only greater than that observed in the 3 mm., 155 hour stage, but almost equals that counted in many specimens of older stages after migration has been completed, as, for instance, the 11.4 mm. and 15 mm. stages (see table). In this stage clearly defined sex-cells can be seen in the peripheral entoderm just below the lateral plates of mesoderm, figs. 16 and 17. These cells are distinguishable from the other entoderm cells among which they lie, by the greater size of their contained yolk granules as contrasted with the small size of the yolk granules in the other cells that make up this layer. The difference is further marked by the more rounded form of the sex-cells. Comparison of these sex-cells in the peripheral entoderm shows them to be identical with other more clearly defined sex-cells in the mesoderm. Of this identity there can be no question, and it is equally clear, from a study of later stages, that these cells, having once migrated into the lateral plates of mesoderm, pass unaltered along the latter to come finally to rest in the sex-gland anlagen. There can be no doubt about the origin of the sex-cells from the entoderm. A number of cases were observed in which the sexcells were actually in process of passing from the peripheral entoderm into the lateral plates of mesoderm.

At this stage, sex-cells have a wide distribution in the peripheral entoderm, being scattered through a region extending from a point opposite to the region where the blood cells originated to the junction of the peripheral, sub-germinal and roof entoderm. In three specimens of the 137 hour stage, conditions are quite similar to the foregoing. In these embryos the number of sexcells ranged from 11 to 29. It will be seen that the maximum number of sex-cells counted in this stage is greater than the minimum number of the 147 hour stage, although in all three of these 137 hour embryos, the caudal end of the embryo, that part



that has been Hfted off the yolk, is decidedly shorter than in any of the 147 hour specimens.

Amia, 132 hr. stage. In four specimens of the 132 hour stage, the caudal end of the embryo was just ready to undergo separation from the yolk. Only in one of them had this really commenced, the separated portion having reached a length of but 20)U. Not one of these four specimens showed a single sex-cell in the mesoderm. There can be no question upon this point because they could be very readily detected if present. In the 137 and 147 hour stages those that migrated into the mesoderm stand out most clearly and sharply from the surrounding mesodermal cells. The points of difference between the two kinds of cells are very striking and unmistakable. The sex-cells on the one hand are large, spherical, have sharply defined boundaries, and are filled with large oval yolk grains; while the mesodermal cells are small, flattened, syncytial, and contain a very few minute yolk granules.

It is very much more difficult to trace the earlier history of the sex-cells in the peripheral entoderm, owing to the slight differences that may be taken as criteria in distinguishing them from the neighboring entoderm cells. Numbers of cells with all the characteristics of sex-cells are found just beneath the anlagen of the blood masses. This stage is just before the development of blood vessels within the embryo, and the blood-forming cells occur in the form of two sharply limited bands, one on each side of the embryo and at some distance lateral to it. Here and there, sex-cells are found in the peripheral entoderm, medial to these areas; but clearly defined cases of this sort are rather rare as compared with the large number seen in this region a little later in the 147 hour stage. It is quite likely that many of these sex-cells are overlooked at this stage owing to the fact that the neighboring entodermal cells contain rather large yolk grains at this time, while those seen in these cells in the 147 hour stage are much smaller than at this stage.

It is quite possible that the sex-cells may migrate medially in the entoderm from an entodermal source beneath the blood anlagen to various points between this region and the edge of the sub-germinal cavity. It is possible that a large proportion


of them may have developed in the peripheral entoderm throughout this entire extent. On the other hand, it is also possible that sex-cells may migrate up into this region from the central entoderm beneath.

We have traced the history of the sex-cells from the 4 mm. stage where they are readily identified by any one who has had any experience in observing these cells, back to the earliest stage at which they are distinguishable in the entoderm. We shall now follow them up to the period when they are enclosed in the definitely formed sex-glands and finally to the stage at which they are found to have begun to increase in number.

Amia 5 mm., total length. Passing from the 4 mm. stage to the next represented in our series, 5 mm., we find that the sexcells have made but little progress in their migration toward the median edge of the lateral mesodermal plates. The total number of sex-cells counted in this stage was surprisingly small, being 43 as compared with 87 in the 4 mm. stage. This difference in number is probably due to individual variation. The hind gut has materially lengthened, being 1.3 mm. in length, compared with .88 mm. in the 4 mm. stage. There has been a corresponding increase in the length of the region over which the sexcells are distributed. In the 4 mm. stage they extend from a point 0.06 mm., in front of the beginning of the hind gut, caudally to a distance of 0.35 mm. In the 5 mm. stage that we are considering, this region begins at the same point relative to the hind-gut and extends caudally for 0.50 mm., one isolated sexcell being found at a distance of 0.57 mm. behind the cranial limit of their distribution.

In the more caudal portion of this region the splanchnic and somatic layers of mesoderm have begun to separate to form the coelome. This separation does not at first lead to the formation of a continuous cavity but rather to a series of isolated, somewhat rounded cavities. Further caudad, the coelome becomes more and more completely developed, appearing as a large cavity on each side.

Amia 6 mm., total length. At this time the first sex-cells appear in the splanchnopleure just at the entrance of the hind gut.


The first sex-cells in the somatopleure are found in the sex-gland anlagen a short distance (0.04 mm.) behind this point. The sexcells are distributed somewhat irregularly from the cranial end of the hind-gut to a point 0.90 mm., caudad to this point and there are a few scattering sex-cells still further caudad than this.

The coelome is apparent as a continuous cleft on either side of the hind-gut along the entire extent of the region occupied by the sex-cells. The majority of the sex-cells are to be found in the dorso-medial extremity of the coelome, i.e., near the root of the mesentery. A few lie lateral and ventral to the intestine. The coelomic cleft has not as yet become wider than the diameter of the average sex-cell and we consequently see them usually bridging across it, fig. 18. In no case have they penetrated into the somatic mesoderm as we find them doing later. One sex-cell was found in the gut-entoderm, whither it may have migrated from the mesoderm. It is, on the other hand, quite possible for it to have migrated in the entoderm in the manner of sex-cell migration in the turtle. This is a point of minor significance and an occurrence which is at best very infrequent.

Amia, 7 mm., total length. Up to this time, the mesentery has been only potentially present, the two lateral plates of mesoderm being in contact above the gut-entoderm. Now, however, we find that it has begun to elongate and become thin. This is naturally correlated with the increase in the extent of the coelome, fig. 19. Two well defined sex-cells are found in the gut-entoderm, 0.06 mm., cranial to the opening of the hind-gut. These are to be interpreted in the same way as the cell in the entoderm mentioned above. The first sex-cell occurring in the mesoderm is found 0.08 mm. caudad to the beginning of the hind-gut. The sex-cells are distributed through a region extending from a point immediately back of the opening of the hind-gut to a point 1.05 mm. behind it, with a few scattering ones behind these. The total number of sex-cells is 74.

Amia, 9.1 stage. Sex-cells first appear .18 mm. cranial to the opening of the hind-gut. They extend from this point to a point 1.59 mm. caudad to this, giving a total extent of 1.67 mm. The total number of sex-cells counted at this stage amounted to


76. Of these all were in the sex-gland anlagen except three; one of which occurred in the gut-entoderm and two in the parietal peritoneum. I am inclined to consider it unlikely for these misplaced sex-cells to reach the sex-glands. One is struck, however, with the great difference in the relative number of misplaced sexcells in Amia as compared with Lepidosteus. This may be apparent rather than real, owing to the possibility that in Amia large numbers of them may have failed to migrate from the entoderm into the mesoderm during early stages. Owing to the difficulty of certainly distinguishing sex-cells in the entoderm from ordinary entoderm cells, it was quite impossible to make any count of those left behind in migration. All but a very few, however, that reach the mesoderm succeed, as we have seen, in reaching the sexgland anlagen. A considerable number of cells seen in the entoderm in later stages contain small yolk spherules and show other points of resemblance to sex-cells. In this stage the mesentery has become quite lengthened and the coelome very large. The sex-cells have penetrated into the root of the mesentery, fig. 20.

The sex-cells, with rare exceptions, still contain large quantities of yolk material. In these exceptional cases a finely granular appearance gives at least the suggestion of small unstained yolk spherules. The yolk appears in the shape of particles varying in size from small granules up to large lemon-shaped pieces quite as large as those with which the cells of the yolk entoderm are so completely filled.

Amia 11. j^ mm., total length. The sex-cells are fairly numerous over a region 1.85 mm. in length, beginning at a .point 0.06 mm. back of the yolk stalk and ending at a point 0.85 nam. cranial to the cloaca. Two isolated sex-cells are found caudad to the point named, one of them occurring very close to the cloaca. Their total number in this embryo is eighty- two. The sex-cells have much the same characteristics as in the previous stage.

This stage is marked by a decided increase in the length of the mesentery and by a decrease in the size of the yolk-sac, which is now but 0.7 mm. in diameter and is greatly hollowed out to form a portion of the intestinal wall.


While the sex-cells of the 9.1 mm. stage are imbedded in the mesoderm at the root of the mesentery and always close to the median line, they are found in the 11.4 mm. stage to occupy a position a short distance on each side of this point. Not only have they moved laterally, but they have also protruded into the body cavity, accompanied by a few mesoderm cells which are intercalated between them, fig. 21, and surround them with a thin peritoneal investment as well.

Amia 15 mm., total length. In this stage the sex-cells extend over a distance of 2.70 mm. in the caudad 0.50 mm. of which they are very sparse. The sex-glands protrude further into the body cavity than in the preceding stage, and the ligament of attachment becomes narrower. The genital ridge is very much lower in the gaps between sex-cells than it is in the sex-cell regions. In spite of the fact that it may be very low for quite a distance, it is continuous throughout. The genital ridges diverge quite widely at their cranial ends, approaching the median line at a point .4 mm. caudad to their point of commencement.

The sex-cells have almost uniformly used up their contained yolk material, although a few scattered ones are still closely packed full of them. The sex-cells in specimen A, numbered 28 on the right side and 49 on the left, the total number being 77. The number of sex-cells in specimen B was 38 on the right side and 45 on the left, the total being 83.

Amia 16 mm. long. In two 16 mm. larvae, conditions very similar to those of the 15 mm. stage were found. None of the sex-cells contained yolk material in a sufficiently large amount to be clearly recognizable. The striking thing about these two specimens is the very small number of sex-cells present, 33 in one case and 39 in another. There is no indication of degeneration or of a failure to migrate to the proper positions.- The case seems to be similar to one cited in Chrysemys, both being due to individual variation.

These two specimens were taken from the same brood and no doubt had the same parentage. Another 16 mm. specimen taken from a different brood showed 09 sex-cells, a number not very far below the maximum. From this fact, and from the



total absence of any indication of degeneration of sex-cells in these or earlier stages, I feel convinced that this small number does not indicate any tendency to degeneration of sex-cells.

Amia, 23.7 mm. total length. In the next stage studied, 23.7 mm., the sex-cells numbered 102. Here again there is no evidence of a change in the number of sex-cells originally present. The number, although somewhat high, is exceeded by some of the specimens of very much earlier stages. There is no evidence of sex-cell division nor of any degeneration.

Amia, 1^0 mm. total length. At this stage the sex-gland is elongated oval in transverse section. It has become bent over in such a way that the proximal edge is medial and the free edge


Dimensions of sex-cells of Amia

































lateral in position. The mesodermal cells have increased greatly in number. The peripheral cells have become arranged into a somewhat poorly defined layer, while the sex-cells lie in the interior of the sex-gland. No attempt was made to determine the time at which the sex-cells begin to divide, or to study the further, development of the sex-glands.

Measurements of the nuclei and cell bodies of the sex-cells gave the following averages, two diameters being measured in each of five sex-cells chosen at random in each stage.

Although the number of cells measured in each stage is hardly sufficient to justify one in considering these average dimensions to have any high degree of accuracy, I feel that we are quite justified in concluding from these figures that: (1) there is a fair decrease in the size of the cell-body as development proceeds, and (2) that there is a slight increase in the size of the nucleus. The decrease in the size of the cell-body is probably due to the absorption of the yolk material with which the sex-cells are so richly filled during the earlier stages. No good explanation to account for the slight apparent increase in size of the nucleus presents itself.


We can not consider this work as completed without making a comparison between the sex-cells and the other cells of the embryo. This subject will first be taken up in Amia where we have traced the sex-cells back to earlier stages than in Lepidosteus. It has already been pointed out that the sex-cells, as first seen in the peripheral entoderm, are to be distinguished only by the size and arrangement of the yolk spherules. The nuclei bear a close resemblance to those of surrounding cells of the same size, while the larger nuclei of larger cells show many points of similarity to them. In all except the earliest stages studied, these nuclei are quite rounded. The chromatin appears in the form of slender strands that take a peripheral position in the nucleus. There is invariably a plasmosome present and rarely two of them. In the 147 hour stage the nuclei of the sex-cells bear a resemblance not only to those of the neighboring cells but also to those of the gut entoderm. In fact, many nuclei of the mesoderm show similar characteristics.

After development has gone a little further, as in the 3.4 mm. and 4 mm. stages, the mesodermal nuclei and those of the gut entoderm are found to have become smaller and are more deeply stained than those of the sex-cells and peripheral entoderm. In all of these later stages, which include 5 mm., 6 mm., 9. 1 mm., 11.4 mm. and 16 mm. larvae, these differences are found to increase. Although the sex-cells undergo a migration from the peripheral entoderm into the lateral plates of mesoderm and through the latter to the sex-gland anlagen, they still bear a close resemblance


to certain cells of the peripheral entoderm. This not only involves a similarity of the nuclei but of the dimensions of the cell bodies. This is true even after the sex-cells and the corresponding cells of the peripheral entoderm have lost their yolk through absorption.

In the stage of 11.4 mm., the yolk mass has been greatly reduced (figs. 25 and 26). Only here and there about its periphery are cells to be found with well defined outlines. The great mass is syncytial, with large nuclei of varying size scattered here and there. While these nuclei of the vitelline mass are much larger than the sex-cell nuclei, they bear a close resemblance to the latter. The nuclei of the well defined peripheral cells are practically identical in size and appearance with those of the sex-cells.

While the similarity between sex-cells and between these two classes of cells is not so marked in Lepidosteus as in Amia, yet it appears to be equally true. In the 17 mm. stage (figs. 14 and 15) the yolk mass is still of fair size. There is a layer of peripheral entoderm that is largely made up of cells with clear boundaries, whose nuclei are similar to those of the sex-cells in respect to the presence and character of the plasmosome and in the form and distribution of the chromatin material. In many cases these nuclei are larger than those of the sex-cells; but many are found which are quite as small. These grade into the very large nuclei of the syncytial vitelline entoderm.

At this stage the tissues of the body have taken on their distinctive characters and their component cells have undergone in many cases a high degree of specialization. This emphasizes strongly the similarity between the sex-cells and the cells of the peripheral entoderm.

As we pass back to earlier stages, such as those of 9.3 mm., 5.9 mm., etc., we still find this similarity between these types of cells, although the nuclei of all the body cells tend to show greater and greater similarity to one another in the earlier stages. For instance, it becomes quite difficult to distinguish the nuclei of the gut entoderm cells from those of the sex-cells. Even the nuclei of the Wolffian ducts show quite a close resemblance to the sexcell nuclei during the early stages of development.


There are two ways of viewing the similarity that the sexcells of Amia and Lepidosteus bear to these cells of the peripheral entoderm. The well defined cells of the peripheral entoderm might be interpreted as sex-cells that have failed to migrate into the lateral plates of mesoderm. It would then remain to give an explanation of the resemblance that the nuclei of these cells bear to the nuclei of the vitelline entoderm and to account for the intermediate types of nuclei by which they grade into one another.

The other view of this problem is to consider sex-cells, peripheral entoderm cells, and vitelline entoderm cells as slightly differentiated blastomeres, dating from an early stage of development, and to consider the similarity that they bear to the cells of the peripheral entoderm as due to the fact that they too have remained in a relatively slightly differentiated condition. This view seems the more probable of the two. It is by no means a new one, having been advanced by Nussbaum in 1880.

It would be rash in the extreme to claim that the sex-cells might not differ in some essential chromosomal characters from the cells of the peripheral entoderm which they so closely resemble, and yet careful study has failed as yet to show any real differences. While such differences may exist, these cells all have much in common with one another.

In a recent paper by A. P. Dustin ('07), this author gives a new view of the origin and movements of the sex-cells of Triton alpestris, Rana fusca and Bufo vulgaris. Since his view is so greatly at variance with my own, it will be necessary to review this work in some detail. He begins with an account of the sex-cells of Triton, and stress is laid upon this form, the author showing a strong tendency to bring his studies upon Rana and Bufo into line with his work upon Triton.

He first recognizes the anlage of the sex-cells in the medial portions of the lateral plates of mesoderm in the 3 mm. larva of Triton. They occur only in the caudal half of the body and involve only those parts of the lateral plates of mesoderm lying medial to the Wolffian ducts. In the early stages these cells are filled with large yolk spherules and do not greatly differ from the mesodermal cells that surround them. At a later period the sex


cell anlagen are pushed together in the median line, between the aorta and the roof of the archenteron. They fuse into a median longitudinal rod of cells lying just above the dorsal root of the mesentery. By this time the sex-cells have lost their yolk material and have, to a large extent, assumed their definitive character. During these stages the number of the sex-cells has increased from one hundred to one hundred and fifty, occasional mitoses being observed. Soon after this stage of the median anlage (9 mm.) has been reached, the sex-cells migrate laterally to their final positions on each side of the root of the mesentery. At the stage of 14 mm., a large number of them degenerate, leaving only 60. A second generation of sex-cells soon begins to form from a source entirely different from the first, namely, from a transformation of ordinary peritoneal cells. Dustin is, in this regard, quite in accord with Bouin who expressed similar views regarding Rana. Dustin considers somewhat more briefly the corresponding stages in Rana and Bufo. Here he finds what he considers to be a substantially similar source of origin of the sexcells, namely the medial borders of the lateral plates of mesoderm. An incredible feature of his account is the statement that the lateral sex-gland anlagen contain no sex-cell at all comparable in size to those of the yolk-filled entoderm, at the period immediately prior to their union in the median line. Dustin would have us believe, nevertheless, that these selfsame sex-cells show a close resemblance to the entoderm cells immediately after this union of the lateral anlagen, and this in spite of the fact that both of these stages of development are so close together that the embryos upon which he made these observations were all of the same length. His own statement is as follows :

" Au moment ou les ebauches paires separees par une sorte de clivage des lames laterales du mesoblaste se sont rapprochees de la ligne mediane, les cellules sexuelles futures passent par une serie de transformations cytologiques a la suite desquelles elles auront presque les caracteres des cellules de I'hypoblaste vitellin. Les dimensions des corps cellulaires augmentent dans de fortes proportions; les grains vitellins deviennent beaucoup plus nombreux et plus volumineux; ils se colorent mieux par I'orange G. Par le fait de I'augmentation du nombre des plaquettes


vitellines, le noyau, souvent refoule a la peripherie de la cellule, presente a sa surface une serie d'encoches lui donnant un aspect herisse (p. 476).

He finds the number of sex-cells in Rana to increase gradually, from 75 in the 8 mm. stage to 90 in the 15 mm. stage, at which time sex-cells begin to be formed by the transformation of ordinary peritoneal cells. Simultaneously there is a degeneration of sex-cells which is overbalanced by this process of transformation.

In criticism of the above views I wish, first of all, to admit the possibility that Dustin may be perfectly correct in his account of the origin of the first line of sex-cells from the lateral plates of mesoderm in Triton. His account of this feature is circumstantial and rather convincing. His account of a transformation of peritoneal cells into sex-cells during later stages is by no means so easy of acceptation. His figures to demonstrate this are not convincing.

His counts of sex-cells are not given in any circumstantial detail and there is no indication as to whether the number of sexcells recorded for any given stage is the result of a count of the sex-cells in one specimen or in several. One can not be blamed for being skeptical of the value of such counts if made upon but one specimen of each stage, when so few stages are chosen to demonstrate general processes of degeneration and new formation. Such a process can only be established by a count of the sexcells of numerous specimens.

I wish to express my complete disbelief in the first appearance of the sex-cells in the lateral plates of mesoderm of Rana and Bufo in the manner described by Dustin. In my paper upon An Important Period in the History of the Sex-Cells of Rana pipiens" ('07) I showed that the sex-cells migrate upward from the median dorsal portion of the gut entoderm at the time when the two lateral plates are pushing together to the median line in the process of forming the mesentery. Attention was called to the resemblance that this process bears to an actual pinching off of the mass of sex-cells by the inner margins of the plates of


mesoderm. As pointed out in my article, the lateral plates of mesoderm, examined immediately before their approximation in the median line, show no cells which, as regards size or yolk content, in the least compare with the sex-cells.

It is especially gratifying to me to find support for my views in two recent papers. In one of these Kuschakewitsch ('08), referring to my paper of a few months before, stated: Der Verfasser hat die Abschniiring von Dotterzellen langs der dorsalene Sagittallinie des Dottersackes im hinteren Telle des Rumpfes beobachtet und die Theilname dieser Dotterzellen am Aufbau einer kompakten Mesenterial-anlage festgestellt, die Bouin (1900)) als ebauche genitale primordiale" aufgefasst hatte. Wie aus meiner Schilderung der entsprechenden Vorgange in der Normalreihe von Rana esculenta zu ersehen ist, kann ich die Angaben von Allen voUstandig bestatigen."

Another paper, appearing the same year (King, '08), gives an account of the origin of the sex-cells in Bufo lentiginosus which is in complete accord with the above, and states : "Allen's recent account of the origin of the sex-cells in Rana pipiens agrees essentially with what I have found in Bufo." Miss King finds no evidence in the course of development of any transformation of peritoneal cells into sex-cells as asserted by several writers among whom may be mentioned Bouin and Dustin. This is quite in accord with my observations upon Chrysemys ('06) in which the sex-cells were traced to the period of sexual maturity without finding any evidence of such transformation.

Miss May Jarvis ('08) in a paper upon The Segregation of the Germ-Cells of Phrynosoma cornutum" (preliminary note) finds the sex-cells to take their origin in the entoderm of the vascular area on all sides of the embryo, even cranial to it, and notes a few in the region of the brain. Her results are in their main features confirmatory of my own work upon Chrysemys. The following quotation from her paper is self-explanatory: Through the courtesy of Dr. Allen, I have been enabled to examine the more important stages in the migration of the germ-cells of Chrysemys; they are similar to my own material, as my conclu


sions, although differing from Dr. Allen's in details of early distribution and periods of migration, uphold his."

Rubaschkin ('08 and '09) in a couple of recent papers, has shown that the sex-cells of the rabbit and guinea-pig are first to be found in the entoderm at some distance on each side of the hind-gut and that they follow a path almost identical with that followed by the sex-cells of Chrysemys. These references to the coincidence of the views of other recent writers with my own are made to show that I do not stand alone in placing emphasis upon the entodermal origin of the sex-cells in the vertebrates. At the same time I wish, however, to disclaim any intention of making at this time a sweeping claim that the sex-cells of all vetebrates arise in the entoderm. Wheeler's work on Petromyzon ('99) shows that they may be included in the mesoderm at the time when that layer is split off from the entoderm. He has, however, pointed out their similarity to the entoderm cells and their dissimilarity to the mesodermal cells among which they lie.

I do not seek to discredit the work of Dustin upon the sexcells of Triton; although his statements about the origin of the sex-cells in Rana and Bufo strike me as being very far from the mark, because they are so radically at variance with not only my own observations, but with those of King and Kuschakewitsch as well. Dustin, in his attitude toward the work of others, seems to consider that there must be a strict uniformity in all forms in both the place of origin and in the movements of the sexcells. He has apparently studied this problem first in Triton and at some length. His results, probably correct for that form, he has attempted to apply to Rana and Bufo as well, undeterred by the difficulties to which attention was called above. Dustin is quite ready flippantly to dismiss my work upon Chrysemys, because the results there expressed did not coincide with the views that he had formed regarding the origin of the sex-cells in Triton, Rana, and Bufo.^ The process of migration through the entoderm is so clear in Chrysemys, that it is unmistakable. The sex-cells are not only characterized by their larger size,

1 See postscript.


definite, rounded outlines and fine chromatin network, but by their large yolk content and the fact that they do not divide during the stages in dispute.

The sex-cells are migratory to a high degree. The path and time of their migration may vary greatly within a given group of animals, as illustrated by the case of Amia and Lepidosteus. While in the forms that I have studied they are first to be observed in the entoderm, I am quite open to conviction that in other forms they may migrate from this layer into the potential mesoderm before the two layers are separated, as shown by Wheeler in Petromyzon. It is even conceivable that they may lie, from the very beginning of development, in material destined to form mesoderm — that they may never have existed among cells actually or potentially entodermal. The more recent development of our work along these lines, however, most certainly tends to show that it is usual among the vertebrates for the sex-cells to first appear in the entoderm.


1. The sex-cells of both Amia and Lepidosteus have their origin in the entoderm. In Amia they are first distinguishable in the peripheral entoderm from the lateral angle of the subgerminal cavity to the anlage of the blood cells.

In Lepidosteus they are first seen in the ventral and lateral portions of the gut-entoderm, although analogy with Chrysemys leads us to assume that they may have migrated through the entoderm to these regions from more lateral anlagen, similar to those from which the sex-cells of Amia arise. In both forms, the sex-cells arise only in the region of the hind-gut. None were found at any considerable distance in front of it.

2. The path of sex-cell migration in Amia carries them out of the peripheral entoderm directly into the overlying lateral plates of mesoderm, along which they travel, to come to rest near the medial edges of the latter. These portions are destined to join above the intestine to form the mesentery. As the splanchnic and somatic layers of the lateral plates of mesoderm


split to form the coelome, the sex-cells adhere to the somatic layer at a point near the root of the developing mesentery — the sex-gland anlage. They later sink into the peritoneum of this region, which afterwards proliferates to form a long ridge — the sex-gland. Very few sex-cells fall by the wayside in this migration, practically all reaching the sex-glands.

3. In Lepidosteus the sex-cells, first seen in the ventral and lateral portions of the gut-entoderm, migrate to occupy a position in the dorsal portion of it, from which they pass dorsally into the loose mesenchyme that forms the substance of the developing mesentery. As the mesentery becomes more narrow and compact, owing to the increase in size of the body cavity, the sexcells migrate to its dorsal portion and laterally to the sex-gland anlagen. Roughly speaking, one-half of the total number of sex-cells reach the sex-gland anlagen, the remainder being distributed between the intestinal entoderm, the mesodermal layers of the intestine, the mesentery and the tissues at and dorsal to the root of the intestine.

4. The number of the sex-cells in Amia and Lepidosteus is a matter of individual variation for those periods of development during which they do not undergo division. The average number in Amia, after the period when the migration from the entoderm to the mesoderm has been completed, up to the latest stage in which counts were made, was found to be 75. In Lepidosteus it was 765, an average of 636 of these occurring in the mesoderm.

5. There is a close resemblance between the nuclei of the sexcells and of the yolk cells. This is especially true of certain cells of the peripheral entoderm, although these grade by gradual transition forms into the large nuclei of the vitelline entoderm. This is probably due to the fact that both types of cells have undergone but little differentiation in the course of development.


A few days before proof of this article came to hand, I received, through the courtesy of the author, a reprint of an article by A. P. Dustin, entitled, "L'Origine et I'Evolution des Gonocytes chez


les Reptiles," (Archives, de Biologie, 1910). This article deals with the origin of the sex -cells in Chrysemys marginata, the form which served as a subject for my own work of 1906. As noted above, Dustin in his paper "Recherches sur Torigine des gonocytes chez les Amphibiens" 1907, exhibited scant respect for my work on the sex-cells of Chrysemys. It was, no doubt, in large part, this feeling that prompted him to repeat my work. While he, no doubt, expected to find in this form a confirmation of his previously expressed views, he is led to substantiate completely my statements regarding the entodermal origin of the sex-cells. He traces them along the same migration path that I demonstrated four years before. For all this he now gives me full credit and support; but takes issue with my statements regarding the distribution of the sex-cells prior to their migration into the embryo, and, furthermore, claims to have -evidence to show that there is a new formation of sex-cells, due to a transformation of ordinary peritoneal cells. These points of controversy and certain other minor ones can not be considered here, but I promise a full discussion of them in another place. I may say that I am fully prepared to maintain my views upon all of the points at issue.

On my part, the work that I have carried on upon Necturus since this paper was written, has given me results quite similar to those at which Dustin arrived in his work upon Triton. I may say that preliminary studies have convinced me that the sex-cells arise in an essentially similar manner in Amblystoma. We then see that, in all three of these urodeles, the sex-cells arise from the inner edges of the lateral plates of mesoderm. I owe it to myself to call attention to the fact that I have at no time disputed the accuracy of Dustin's work upon Triton. While the evidence seems to me quite clear that this is the usual, if not the universal, mode of origin of the sex-cells among the urodele amphibians, I am ready to maintain with equal vigor the entodermal origin of the sex-cells in the aruran amphibians, at the same time admitting the possibility that exceptions to this apparent rule may be discovered. I do not feel however, that Dustin has proved his case in Rana fusca and Bufo vul


garis. The discussion of his work above gives the reasons for my position in this matter.

Not only does it seem probable that the sex-cells arise during early stages in the mesoderm of the urodeles, but this seems to be the case in the teleosts as well. The most recent and satisfactory support of this view is contained in the excellent paper of Dr. Gideon S. Dodds upon the Segregation of the Germ-Cells of the Teleost, Lophius, " in the Journal of Morphology, 1910. Here again, we must urge caution in forming a sweeping generalization from the facts thus far at hand. There is certainly a wide field for work in the study of the origin of the sex-cells of the vertebrates. It is a subject which should be approached in a spirit of broad toleration for the views of others. The sexcells are cells that retain their early embryonic character after the somatic cells have undergone specialization. It seems, from a number of observations made by different authors, that in most forms the sex-cells first make their appearance in the entoderm — the germ layer whose cells appear to maintain their primitive embryonic characters longer than do those of the other germ layers. At the same time, unimpeachable evidence shows that this apparently logical process is not universal, and I have at no time claimed that it is. The sex-cells, as show^n by Nussbaum, Eigenmann, Beard and others, do not belong to any one germ layer, but are, in a sense at least, independent of the somatic tissues. They are free to follow their own path in their travels from the place of origin to the sex-gland anlagen, where they finally come to rest. While this path is no doubt identical or similar in closely allied species and in more general divisions of the vertebrates, I do not feel that we are justified in attributing a high degree of phylogenetic importance to the different steps in the migration paths through which they travel.

I wish to express my indebtedness for the work of our departmental artists. Misses Hedge and Battey. I am indebted to Miss Hedge for the execution of diagrams 1-6 and for figs. 9, 10, 14, 15, 21, 22, 25 and 26; and to Miss Battey for figs. 11, 12, 13, 23, and 24. The remaining drawings are my own.



Allen, Bennet M. 1906 Origin of the sex-cells of Chrysemys. Anat. Anz. Bd. 29.

1907a A statistical study of the sex-cells of Chrysemys marginata, Anat. Anz. Bd. 30.

19076 An important period in the history of the sex cells of Rana pipiens. Anat. Anz., Bd. 31.

1909 The origin of the sex- cells of Amia and Lepidosteus. Anat. Rec, Vol. 3.

DusTiN, A. p. 1907 Recherches sur I'origine des gonocytes chez les Amphibiens. Arch, de Biologie, tome 23.

Jarvis, Mat. 1908 The segregation of the sex-cells of Phrynosoma. Biol. Bui., Vol. 15.

King, Helen Dean. 1908 The oogenesis of Bufo lentiginosus. Jour. Morph., Vol. 19.

KuscHAKEwiTSCH, S. 1908 Ueber den Ursprung der Urgeschlechtszellen bei Rana pipiens. Stzber. math. phys. Klasse, k. bayer. Akad. Wiss., Bd. 38.

RuBASCHKiN, W. 1907 Zur Frage von der Entstehung der Keimzellen bei Saugetierembryonen. Anat. Anz., Bd. 31.

1909 Ueber die Urgeschlechtszellen bei Saugetieren. Anat. Hefte, Bd. 39.

Wheeler, W. M. 1899 The development of the urogenital organs of the lamprey. Zool. Jahrbuch., Anat. Abth., Bd. 13.


Arch., Archenteron

Coel., Coelomic cavity

Ect., Ectoderm

Gut End., Gut entoderm

Int., Intestine

Lat. Mes., Lateral plate of mesoderm

Mes., Mesentery

Meson., Mesonephros

Myo., Myotome

Nolo., Notochord

P. Card., Post cardinal vein

Periph. End., Peripheral entoderm Roof End., Roof entoderm S. €., Sex-cells S. GL, Sex-gland

Sub-Germ. Cav., Sub-germinal cavity Sub-Germ. End., Sub-germinal entoderm Siv. Bl., Swim bladder Vit. End., Vitelline entoderm

Wolff. D., \ W. D., J

Wolffian duel




1 Diagram to show the migration path of the sex-cells in Chrysemys marginata.

2 Diagram to show the migration path of the sex-cells in Rana pipiens.



3 Diagram to show the migration path of the sex-cells in Lepidosteus osseus.

4 Diagram to show the last phase of the migration of the sex-cells in Lepidosteus osseus.

5 Diagram to show the migration path of the sex-cells of Amia calva.

6 Diagram to show the last phase of the migration of the sex-cells in Amia calva.





7 Transverse section through the hind-gut of an 8.6 mm. larva of Lepidosteus osseus. X 300.

8 Transverse section through the hind-gut of a 9.3 mm. larva of Lepidosteus osseus. X 300.

9 Transverse section through the hind-gut of a 10.7 mm. larva of Lepidosteus osseus. X 300.

10 Transverse section through the hind-gut of a 14.1 mm. larva of Lepidosteus osseus. X 300.

11 Transverse section through the hind-gut of a 17 mm. larva of Lepidosteus osseus. X 300.



Gelatin plates 1, 2 and 3 should have been numbered 3, 4 and 5








12 Transverse section of the rudimentary sex-glands of a 24 mm. larva of Lepidosteus osseus. X 300.

13 Transverse section of a sex-gland of a 110 mm. specimen of Lepidosteus osseus. X 300.

14 Part of a transverse section of a 17 mm. larva of Lepidosteus osseus, showing the reduced vitelline mass.

15 Detail drawing of a portion of the vitelline mass of the above section. X 300.

16 Transverse section through the region immediately lateral to the posterior portion of the sub-germinal cavity of a 147 hr. embryo of Amia calva. X 300. This shows the place of origin of the sex-cells.

17 Section passing similarly through another specimen of the same stage of Amia calva. X 300.

One sex-cell shown as it is pushing up into the mesoderm.

18 Transverse section through the hind-gut of a 6 mm. larva of Amia calva. X 300.

19 Transverse section through the hind-gut of a 7 mm. larva of Amia calva. X 300.



Periph. EncL.jg.^


«ir.-J?>Vsw B,, ^f:'^ ^1^ r^x




20 Transverse section through the hind-gut of a 9.1 mm. larva of Amia calva. X 300.

21 Transverse section through the hind-gut and sex-gland anlage of an 11.4 mm. larva of Amia calva. X 300.

22 Transverse section through the young sex-glands of a 16 mm. larva of Amia calva. X 300.

23 Sketch to show the orientation of the sex-glands in q 40 mm. specimen of Amia calva as seen in transverse section.

24 Detail drawing of the sex-gland as seen in above sketch. X 300.

25 Drawing to show the orientation of the much reduced vitelline mass of an 11.4 mm. larva of Amia calva.

26 Detail drawing of a portion of the vitelline mass indicated above. This shows the resemblance that certain cells of the peripheral entoderm show to sexcells of this stage. X 300.






From the Laboratory of Experimental Pathology of the University of Pennsylvania, and from the Pathological Laboratory of the Barnard Skin and Cancer Hospital, St. Loxds, Mo.

In the course of an experimental investigation into the causes of the cychc changes taking place in the uterine mucosa and into the factors underlying the formation of the maternal placenta in mammals, we observed that cyclic changes in the structure of the ovary correspond to the uterine cycle. It has of course been known that at certain times ovulation takes place in the mammalian ovary, and furthermore, changes have been described as occurring in the ovarian follicles of certain mammals in connection with copulation and during pregnancy; but the cyclic changes taking place in the ovary 'quite independently of copulation and of pregnancy and merely dependent upon ovulation have, as far as we are aware, not yet been recognized. While we know of no publication dealing with the cyclic changes in the ovaries in general, a valuable study of the changes taking place during pregnancy in two species of Insectivores and in one species of Lemurid has been made by C. H. Stratz.i This author comes to the conclusion that in the period following copulation all the ovarian follicles become atretic; that during pregnancy small follicles are formed but also become atretic before they can develop; that only towards the end of pregnancy the follicles begin to grow to a considerable size, and that they reach the stage of maturity during the puerperium.

Stratz was not in a position to determine in an exact manner the time elapsed since the last copulation of the animals the ova ^ C. H. Stratz: der geschlechtsreife Saugethiereierstock. Haag. 1898.



ries of which he examined. He also seems to have examined a relatively very limited number of ovaries of animals during the different stages of pregnancy, and furthermore he studied only certain parts of each ovary. A methodical study of ovaries of non-pregnant animals was not undertaken. While his observation that after copulation all follicles become atretic is approximately, but not altogether correct, as far as its general validity is concerned, in the guinea pig the processes taking place in the ovaries during the subsequent stages differ from the conditions described by Stratz in the case of Tupaja, Sorex and Tarsius.

Furthermore Stratz does not recognize the essential factor upon which the cyclic changes in the ovaries depend. The conclusions in the last chapter of his publication show this clearly.

He summarizes as follows: If we find all follicles atretic, the animal has been pregnant. If at the same time a new corpus luteum is present, we have to deal with an early stage of pregnancy. If we detect some normal follicles, besides numerous atretic follicles and a new corpus luteum, we have to consider a puerperal condition of the animal. A large number of atretic besides a few normal follicles also* suggests a puerperal state.

These general conclusions are not justified; the changes of the follicles do not, as Stratz assumes, depend upon pregnancy, and if we should attempt to use the criteria given by Stratz in the case of guinea pigs and mammals in general we would be liable frequently to make mistaken diagnoses. Notwithstanding, these necessary criticisms, the work of Stratz is very valuable and it advanced to a considerable extent our knowledge of the ovaries.

Since his publication no more detailed investigation into the processes taking place in the ovaries under various conditions has appeared, as far as we are aware. Within recent years, however, the question has been raised whether a new ovulation can take place during pregnancy.

We limited our investigations to the study of the ovary of the guinea pig. We examined several hundred pairs of ovaries of animals in which the period of the sexual cycle at which the ovaries


were obtained had been ascertained. In each case the entire ovary was cut into serial sections.

During the progress of our work new problems arose and an accident made it impossible for us to re-examine all our material in order to answer several questions which were raised at a later stage of our investigation. We especially regret our inability to determine the existence of follicles which were ready to rupture, in certain cases in which these data would have been of considerable interest. Our work is therefore incomplete in some respects. We expect, however, very soon to be able to supplement our present work, wherever necessary.


The condition of the ovaries of a guinea pig in the last days of pregnancy is as follows: there are small, medium sized and large follicles without degeneration of granulosa cells. In other large follicles various stages of granulosa degeneration are present. Many follicles show further advanced stages of atresia, in which connective tissue grows into the follicular cavity. Especially numerous are the last stages of atresia in which the zona pellucida is directly surrounded by very cellular connective tissue. Mitoses are seen in the granulosa cells of the well preserved follicles. We also find here a few mature follicles which are characterized by an increase in cytoplasm of the granulosa cells. These follicles are large; their cavity is very wide. The nuclei of the granulosa cells are not as densely packed in these follicles as in the ordinary large follicles, this peculiarity being due to the marked development of the cytoplasm. They can be easily recognized in sections stained by haemotoxylin and eosin, inasmuch as they appear stained more reddish, in contradistinction to the ordinary large follicles in which the blue color of the nuclei predominates, while in the mature follicles the red stain of the cytoplasm is a distinguishing feature. In these mature follicles the number of mitoses is very much smaller than in the ordinary large follicles. With the increase in the quantity of cytoplasm and the relative decrease in the nuclear material,


the cell proliferation is diminished. The number of mitoses is usually very small, or mitoses may be absent in such follicles. Another characteristic feature is the relative lack of degeneration of the granulosa in these follicles. While the ordinary large follicles degenerate in the large majority of cases, the granulosa cells becoming karyorrhectic, as soon as the follicle attains a certain size; the mature follicles are very much more resistant. The changes in the granulosa cells described above and which lead to the transformation of an ordinary large follicle into a mature red-staining follicle, and simultaneously to a decrease in cell proliferation of the granulosa and to a diminished karyorrhexis of the granulosa cells, probably produces a decrease in cell metabolism, and this decrease in cell metabolism stands perhaps in a causal relation to the decrease in cell multiplication and to the greater resistance of the granulosa cell. A slight degree of degeneration of the granulosa may even occur in the mature red-staining follicles; a few of the central granulosa cells may degenerate; and in one case we observed even a fargoing degeneration of the granulosa in a mature follicle. It becomes therefore probable that these mature follicles also degenerate, if ovulation does not take place. This transformation of an ordinary large follicle into a mature follicle takes place only to a limited extent; the large majority of the follicles degenerate before they have reached the stage of full maturity. This holds good even in the case of guinea pigs before delivery, in which a rupture of follicles will soon take place.

The corpora lutea of pregnancy which, at the time at which we examined the ovaries, were approximately fifty-six to sixty-four days old and which had formed soon after copulation, show already some retrogressive changes in the lutein cells. A considerable number of the vessels entering the corpora lutea have a very thick wall consisting of several rows of cells. A large number of the vessels, however, have merely an endothelial lining. In many of the vessels no lumen is visible, the circulation through the corpus luteum being evidently not very active; some of the capillary vessels have, however, a widely open lumen. The quantity of the connective tissue in the centre of the corpus luteum is small.


on account of the previous proliferation of lutein cells which encroached more and more upon the space originally filled by the connective tissue. The corpora lutea are large. The lutein cells show signs of degeneration ; they are finely vacuolar and may have a foamy appearance ; a certain number of cells take less eosin and appear therefore pale. Many cells have a sharply defined, redstaining outline. The nuclei also show changes; they are frequentquently deformed, indentated ; or they are round, vesicular, but stain less with haematoxylin ; they appear somewhat karyolytic. Mitoses could not be seen in the lutein cells. The degree of retrogressive changes may vary in different corpora lutea even in the same ovary.

We see therefore that even before delivery and before a new ovulation has taken place, degenerative changes set in in the corpora lutea, and it accords with these retrogressive changes that mitoses are absent or at least very rare in such corpora lutea.

Besides these corpora lutea of pregnancy we may find in such ovaries 'yellow bodies' representing the last stage of retrogression of corpora lutea. In the corpora lutea which were transformed into such yellow bodies, degeneration must have set in approximately sixty to sixty-five days ago. These 'yellow bodies' have the following structure: In their centre and periphery we find hyaline connective tissue; between these two zones of hyaline connective tissue a relatively small number of degenerated large lutein cells is enclosed, in which, during the process of retrogression, a large amount of yellow pigment was produced.


In the period directly following delivery the condition of the ovaries, as far as follicles and corpora lutea are concerned, is approximately the same as in the period preceding it. The growth and degeneration of the follicles still continue to take place, and in follicles in which the granulosa has completely or almost completely degenerated an ingrowth of connective tissue and complete atresia of the follicles occur. The retrogressive changes in the corpora lutea also progress, but at a slow rate, and on the whole the



corpora lutea are not very different from those found in the preceding period. This description holds good for instance for ovaries of a guinea pig extirpated ten minutes after complete delivery.

Soon after delivery (usually within a few hours) the guinea pig is ready for a new copulation and ovulation, and after ovulation changes take place in the follicles which will be described later.

The corpora lutea of the preceding pregnancy undergo no very marked changes within the next two days after delivery, although vacuolization of the lutein cells and degenerative changes in the nuclei show probably a slight advance ; the lutein cells do not stain as well with eosin and appear pale. If copulation take place soon after delivery, a rupture of the mature follicles occurs within the succeeding six or ten hours; but if copulation be prevented by isolating the female, ovulation frequently occurs, but does not need to take place within thirty-six hours after delivery. In several cases in which an actual copulation was prevented, in which however the male was in contact with the female for a short time after delivery, the rupture of the follicles and the formation of new corpora lutea took place in the usual way. The changes in the new corpora lutea within the first two days after delivery are the same as those described in a previous paper.^

In three cases the lower part of the uterus or the vagina of guinea pigs were tied completely or incompletely towards the end of pregnancy. This procedure led to the death of the fetuses, followed by expulsion of the dead fetuses in a case in which the occlusion had been incomplete. In another case the animal was killed by chloroform six days after the application of the ligature, and the fetuses were found dead; furthermore autolysis of the placenta had set in. In these cases especially the periphery of the corpora lutea of the preceding pregnancy showed vacuolization of the lute tein cells. The nuclei were shrunken or somewhat chromatolytic. Notwithstanding the degenerative changes visible in the corpora lutea, no new ovulation had taken place. From these and other observations it follows that delivery as such does not lead to far 2 The formation of the corpus luteum in the guinea pig. Journal American Medical Association, February 10, 1906.


going changes in the ovaries; that merely a slow progress takes place in changes which had set in before delivery. We furthermore see that without copulation a spontaneous ovulation does not need to take place after delivery, notwithstanding the degenerative changes in the corpora lutea; that ovulation can, however, occur without copulation, and this seems to be the rule, if the male had been in contact with the female for some time after delivery, a copulation having been made impossible during this period of contact.


This description applies to ovaries of guinea pigs which had copulated a few hours previously, in which an ovulation had however not yet taken place — ovulation usually taking place approximately six to ten hours after copulation. In another case we examined the ovaries of a guinea pig that was ready for copulation ('in heat') in which, however, an actual copulation had been prevented by occluding the vagina by means of a strip of plaster.

The condition of the follicles in these ovaries was similar to the condition found in ovaries preceding and immediately following delivery; we find good follicles of small, medium and large size; mitoses are present in the granulosa of such follicles. The majority of the large follicles however show more or less degeneration of the granulosa, with the exception of the few large follicles which progressed to complete maturity; they showed the cytoplasmic changes described above. In these as well as in some other well preserved large follicles the theca interna appears somewhat hyperemic. We also find the various stages of connective tissue ingrowth and of the subsequent diminution in the size of the follicles ('connective tissue atresia') which we described in the case of the other ovaries. In this case we do not find corpora lutea of a preceding pregnancy, but corpora lutea of an ordinary ovarian period, not accompanied by pregnancy. These corpora lutea are much smaller than those of pregnancy ; their lutein cells show vacuolization, indicating the beginning of retrogressive changes. Notwithstanding these retrogressive changes an occasional mitosis


can still be found in lutein cells. The corpora lutea of the second last ovulation have in the meantime been transformed into yellow bodies. Processes of degeneration have therefore set in in the corpora lutea of non pregnant as well as of pregnant guinea pigs before ovulation. These beginning degenerative changes do however not prevent the occurrence of a few mitoses in the corpora lutea of previously not copulated animals, while in the degenerating corpora lutea of pregnancy we have so far not been able to detect the presence of mitoses in lutein cells.


In connection with ovulation certain far reaching changes take place in the ovaries. All follicles, with exception of very small ones, degenerate. These changes set in with ovulation, or they may perhaps start somewhat earlier, namely, simultaneously with those processes that bring about ovulation. As we have pointed out above, the general degeneration of the follicular granulosa which we find directly after ovulation cannot yet be observed before ovulation. This sudden degenerative process is quite independent of copulation; we found that it can be produced through ovulation without a preceding copulation. We discovered experimental means through which we can produce a spontaneous ovulation without a preceding copulation. Such an ovulation is followed or accompanied by the same degeneration of the granulosa. Moreover, if we keep a number of female guinea pigs separated from the males and if we examine their ovaries after various periods of isolation, we find occasionally ovaries in which the rupture of follicles had taken place a few days before. In this case also the typical follicular degeneration takes place independently of a preceding copulation.

Six and a half hours after a preceding copulation the ovaries showed, besides the presence of newly ruptured follicles, the following changes in the follicles : All, with the exception of very small follicles, show granulosa degeneration; in the large majority of the follicles almost the whole granulosa is found in a process of degeneration. We also find folHcles in the process of connective tissue


atresia. Similar conditions are found in other ovaries at the same period.

Twenty-two hours after copulation some granulosa cells are found degenerated even in small follicles, (follicles having a small cavity) ; these degenerated granulosa cells are dissolved.

Similar changes take place in ovaries of guinea pigs in which ovulation followed delivery. In a guinea pig in which copulation took place two hours after delivery and in which the ovaries were examined seventeen hours after copulation, only a few quite small follicles without granulosa degeneration were found ; in the large and also in the medium sized follicles much granulosa degeneration had taken place, the central granulosa cells degenerating first. Almost no entirely good follicles were left. As soon as the interna becomes exposed, phagocytic cells (rounded off interna cells) peneirate into the follicular cavity and these cells take up debris of the granulosa. The degeneration of the granulosa cells is as usual followed by ingrowth of connective tissue.

In other ovaries the granulosa may be degenerated to a great extent, but some remnants may still be left. Especially the granulosa cells of the discus proligerus survive usually the rest of the granulosa. We find of course various stages of connective tissue atresia besides the degeneration of the granulosa. From these observations it follows that the onset of degeneration of the granulosa must be extremely rapid.

If we extirpate the corpora lutea, from two to eight days after copulation a new spontaneous rupture of follicles takes place in most cases approximately from thirteen to fifteen days after the previous copulation, even if the female had been kept entirely isolated during the whole period following the extirpation of the corpora lutea. This early spontaneous ovulation is accompanied by the same follicular degeneration which we described above.

It is an interesting problem, whether an artificially produced rupture of a follicle, with the subsequent development of a corpus luteum, is accompanied by the same acute follicular degeneration. Several years ago we made experiments in which we pricked or cut the surface of ovaries of guinea pigs which were either 'in heat,' without however having copulated, or which copulated a few hours previously, or which had in some cases copulated from three to six


days previously. In only one case did we find a young corpus luteum the origin of which could reasonably be attributed to the cutting of the ovary and to the artificial ru^pture of a follicle. In this case an animal had been used which showed the first symptoms characteristic for the period of heat. Three days after the cuts had been made the ovaries were examined. One young corpus luteum was found in the cortex of the ovary. Blood and connective tissue were found in the center of the corpus luteum; connective tissue and vessels grew into the corpus luteum, which was very small. In this ovary we found good follicles of small medium and large size; we also found large follicles with beginning and wdth further advanced granulosa degeneration, and with beginning ingrowth of connective tissue. In as much as in no case of spontaneous rupture the follicles were found in a similar condition at that period after the rupture, it is very probable that we have in this case to deal with an artificial rupture of follicles and that such an artificial rupture of follicles is not accompanied by the rapid degeneration of the follicular granulosa.

On the basis of our previous results we can easily understand, why in all probability we succeeded in one case only in causing an artificial rupture of a follicle. Such an experiment does not promise to be successful, unless we have the chance of opening a mature follicle, and such an opportunity exists only at periods of very short duration.

In these ovaries we find usually two or three generations of corpora lutea; namely:

1. The young corpora lutea, developing in the recently ruptured follicles. These corpora lutea we have described elsewhere in their development up to the sixth day.

2. Corpora lutea that had formed at the time of the preceding ovulation, which had not been followed by pregnancy in female guinea pigs which had been kept separated from males. These corpora lutea are therefore in all probability approximately nineteen to twenty-eight days old. They show signs of beginning retrogression. Their lutein cells are more or less vacuolar, especially in the periphery, where the vacuolization usually begins ; gradually the vacuolization progresses to the central part. In the center of


the corpus luteum we find a relatively small amount of fibrous tissue. We not only find capillary vessels but also vessels the wall of which consists of two coats penetrating the corpus luteum.

These corpora lutea begin to shrink very soon, and three days after the new rupture they are usually smaller than immediately after the ovulation. Notwithstanding the degenerative processes which are apparent in these corpora lutea, it is not uncommon to to find still mitoses in the lutein cells of such corpora lutea within the first twenty hours after the new rupture of follicles has taken place. At a later period mitoses were not seen in this series. The mitoses appear in the relatively well preserved, but they may be present even in somewhat vacuolar lutein cells. It is possible that occasionally mitoses occur also in endothelial cells of the capillaries.

3. The third generation is represented by yellow hyaline bodies. They are the remnants of corpora lutea that formed forty or more days ago.

If we examine ovaries of young guinea pigs, two and a half to three months old, we may find only the first, or the first and second generations of corpora lutea, but yellow bodies may be lacking.

We see therefore that preceding and following the rupture of new follicles in non-pregnant animals, processes of degeneration have begun in the corpora lutea of the preceding ovulation, and that notwithstanding such processes of degeneration, mitoses may occur in such corpora lutea for a short period following the new ovulation. These corpora lutea which are not accompanied by pregnancy are much smaller than the corpora lutea of pregnancy and they shrink more rapidly. The absolute diminution in size is more rapid than in the retrogressing corpora lutea of a preceding pregnancy. Concerning the relative rapidity of retrogression (the percentage decrease in size, the full size of the corpora lutea being taken as the standard), we cannot make any definite statement, not having carried out any measurements.

The mode of retrogression is the same in both ordinary corpora lutea and in those of pregnancy. The vacuolization begins in the periphery, where it becomes most marked, and from here it proceeds into the interior of the corpus luteum.



In one case the ovaries of a guinea pig were examined which on examination had previously been found to be in a well developed stage of pregnancy, but which had aborted about three to four days previously. The four corpora lutea showed signs of degeneration. The lutein cells were vacuolar in the periphery, in the center the cells stained pale red with eosin, the vesicular nuclei showed a diminution in the amount of chromatin. The cell outlines were very sharp, staining red with eosin. In the center there was dense connective tissue and many blood vessels had very thick walls.

Follicles of small, medium and large size, with well preserved granulosa, were present. A few mature, red staining follicles without mitoses or degeneration in the granulosa were also found. Many other large immature follicles showed various stages of granulosa degeneration. There were of course also present various stages of connective tissue atresia.

We see therefore that abortion is not followed by or associated with marked changes in the follicles. Whether the mature follicles which we found in these ovaries matured as a result of abortion, or whether the mature follicles were present before the onset of abortion, we cannot state with certaintj^, although it is more probable that maturation of the follicles followed abortion. We also note the beginning retrogressive changes in the corpora lutea. But in this case also we cannot be sure that the degenerative processes had not set in before the abortion had commenced.


Six days after an ovulation we find in the ovaries on the whole the following condition of the follicles: There are well preserved follicles of small and medium size, with mitoses in the granulosa cells. A limited amount of granulosa degeneration is found only in rare instances. In such follicles mitoses are absent or their


number is decreased. Follicles in an advanced state of connective tissue atresia are frequent.

The character of the follicles at this period -of the sexual cycle is the same in cases in which the last ovulation was preceded by delivery, in which, therefore, in the previous period of the sexual cycle a pregnancy was present, and in other cases in which the previous period of the sexual cycle had not been comphcated by pregnancy. We see therefore that within six days quite small follicles, possessing only a very small follicular cavity, grow and reach medium size. During this period the granulosa of medium sized follicles did not degenerate, and no large follicles had as yet developed. We find therefore principally, besides the follicles with preserved granulosa, follicles in an advanced state of connective ,^issue atresia.

Six days after ovulation we find the corpora lutea of the last generation (corpora lutea six days old, as follows : The center of the corpus luteum is filled by a more or less loose connective tissue. Mitoses are present in the lutein cells as well as in the endothelial cells of the capillaries. Almost all the vessels have a capillary character. They penetrate into the central connective tissue. At that period vessels with two coats (intima and muscle coat of the media) can be observed for the first time, although they become more frequent at a somewhat later period.

In guinea pigs in which a pregnancy and delivery preceded the last ovulation, the corpus luteum of the preceding pregnancy shows marked signs of degeneration. Especially the peripheral cells are frequently coarsely, while the more centrally situated cells are more finely vacuolar; but even in the latter the protoplasm stains less with eosin and the nuclei are slightly chromatolytic ; the cells appear distinctly pale. The vessels are very thick and at certain places in the periphery the connective tissue of the neighborhood seems to begin to grow into the peripheral parts of the corpus luteum.

The ordinary corpora lutea of the second generation (not accompanied by pregnancy) show marked vacuolization; they diminish in size and in one case yellow pigment developed in a few of the vacuolar cells. Therefore in the course of five to eight days


since the beginning of degeneration the retrogressive changes have much advanced. The retrogressing corpora lutea of pregnancy of the corresponding generation are much larger at this period than the ordinary corpora lutea.

In a certain number of ovaries we also find a further (third) generation of retrogressing corpora lutea, represented by yellow bodies.

One corpus luteum deserves especial mention. In an ovary of a guinea pig which had ovulated approximately four and a half days before, five corpora lutea were found, four of which showing the typical structure. In the fifth of these corpora lutea, however, the lutein cells were arranged in the shape of glandular ducts. This condition has perhaps been produced through a dissolution of the central cells. Otherwise the corpora lutea in this ovary were normal.

The same typical changes in the folhcles noticed in ovaries of this period after a preceding copulation and ovulation are also found in ovaries in which a spontaneous ovulation took place independently of a preceding copulation. As we stated above, such a spontaneous ovulation can be produced through an early excision of the corpora lutea. The same follicular changes take place also in pregnant animals in which, through an excision of the corpora lutea about six to eight days after copulation, a spontaneous ovulation is produced approximately thirteen to fifteen days after the beginning of pregnancy, without the pregnancy being interrupted.

We see therefore that these cyclical changes in the ovaries are essentially independent of copulation and of pregnancy and are directly connected only with ovulation.


At this stage of the sexual cycle we find good follicles of small, medium and large size with no, or only very little, granulosa degeneration. We also find follicles in connective tissue atresia. We see therefore that in approximately eight days follicles origin ally


very small have reached a large size. The new (eight days old) corpora lutea grow actively during this period and show frequent mitoses in lutein cells. The corpora lutea of the preceding ovulation (second generation) continues to shrink and show marked vacuolization of the lutein cells. If the second last ovulation were accompanied by pregnancy, the retrogressing corpora lutea were still larger.

The third generation of corpora lutea was represented by atretic yellow bodies the age of which varied approximately between forty-eight and ninety-five days.


We fmd good follicles without granulosa degeneration of small, medium and large size, besides various stages of granulosa degeneration and of connective tissue atresia, early stages with beginning ingrowth of connective tissue included. In the granulosa of well preserved follicles mitoses are present as usual.

At this stage — ten days after ovulation— the ovary presents again its normal aspect. The follicles have grown to a large size and undergo the ordinary retrogressive changes. The ten to eleven days old corpora lutea are well developed; in the centre a relatively small amount of connective tissue is present. Mitoses in the lutein cells are usually frequent ; they occur perhaps also in endothelial cells of capillaries. The large majority of the vessels have a capillary character, but occasionally a vessel is seen with a double coat of cells. Marked signs of degeneration are absent, but a few slightly vacuolar lutein cells may occasionally be seen.

The second generation of corpora lutea, originating in the second last ovulation, are small vacuolar bodies with much connective tissue and thick vessels. If, however, this second last ovulation had been followed by pregnancy, the retrogressing corpora lutea of the previous pregnancy are as yet much larger; the lutein cells have become very vacuolar; many thick vessels are present. In some of the vacuolar lutein cells yellow pigment appears.

A third generation of corpora lutea is represented by yellow bodies. They are, however, not found in all ovaries.


In this series of animals pregnancy had been prevented after a preceding copulation, either by ligaturing the tubes within the first two days after copulation, or by making long incisions into the uterus approximately four to six days after copulation.

The ovaries were also examined in a certain number of other guinea pigs of this period in which pregnancy existed. The accompanying pregnancy does not produce any marked change in the ovaries and the preceding description applies on the whole equally well to these ovaries.


In this series of animals pregnancy was prevented in the same manner as in the series of animals examined ten to eleven days after ovulation. The follicles have approximately the same character as in the previous period. We see the same varieties of follicles. Small foUicles grow and become large and, after having reached this stage, or even at a slightly earlier stage, granulosa degeneration sets in with consecutive connective tissue atresia. In the granulosa of well preserved follicles numerous mitoses are present, and mitoses may even be found, if a slight amount of granulosa degeneration has taken place. The corpora lutea of the last ovulation (I generation) show more generally the beginning of vacuolization, especially in the periphery of the corpus luteum; but on the whole the corpus luteum is still well preserved and usually mitoses are found in some of the lutein and occasionally in cells belonging to blood vessels.

In the center we find connective tissue with thin spindle-shaped nuclei, and a number of vessels with walls consisting of several rows of cells penetrate into the central connective tissue. In some of the lutein cells the protoplasm appears dense and stains deeply with eosin. It appears probable that in such cells the nucleus had started to divide by mitosis, but degenerative processes seem to have set in and interrupted the process of the mitotic division. We are however not certain that this interpretation,


which would perhaps agree with an opinion expressed by Regaud and Dubreuil,^ is correct.

The second generation of corpora lutea is represented by small vacuolar bodies with relatively much connective tissue and thick vessels. These atretic corpora lutea originated at the time of the second last ovulation and are therefore approximately thirty-three to forty days old. If this second last ovulation had been followed by pregnancy, the corpora lutea of this period are still much larger than the corpora lutea of the corresponding generation without an accompanying pregnancy; but a considerable shrinking of these corpora lutea has also taken place. The vessels are to a great extent collapsed. The lutein cells are finely or coarsely vacuolar, take less stain, still possess nuclei and a distinct cell wall, staining with eosin. The third generation of corpora lutea is again represented by yellow bodies. They are not present in all ovaries, but are found especially in the ovaries of older guinea pigs. Occasionally the degenerating corpora lutea of the second generation may also be absent.

In guinea pigs in which the last ovulation was followed by pregnancy, the condition of the follicles is very similar. The corpora lutea of the first generation, however, are large and show frequent mitoses in lutein cells, occasionally also in lutein cells the periphery of which is vacuolar. There are possibly also mitoses present in the endothelial cells. The retrogressing corpora lutea of the second and third generations are in pregnant animals of a similar character as those described in the ovaries of guinea pigs of the same period without an accompanying pregnancy.


Pregnancy had in most cases been prevented by the same means which were used in the preceding stages. In a few instances in which pregnancy had occurred an early abortion followed. The follicles exhibit on the whole the same character as in the preceding stage ; we find good foUicles of small, medium and large size,

3 C. R. Soc. Biol., 54. 1908.


and follicles in various stages of granulosa degeneration and of connective tissues atresia. We may also find large mature follicles. In how many cases these latter are present, will still have to be determined. In such animals a rupture of follicles is imminent.

In three guinea pigs a spontaneous ovulation had taken place at this period, notwithstanding the absence of male guinea pigs. In such cases young corpora lutea were found and, accordingly, a condition of the follicles characteristic of a period directly following ovulation. In the large majority of cases however a spontaneous ovulation did not take place in ovaries at this period of the sexual cycle. In such cases the folUcles showed the character described above.

The corpora lutea of the first generation, which originated as a result of the last ovulation, show more or less signs of beginning retrogressive changes as indicated by fine or coarse vacuolization of the lutein cells. The intensity of this degenerative change varies is different ovaries. On the whole the retrogressive changes seem to be more marked in the nineteen days than in the sixteen days old corpora lutea; but variations seem to occur, even in corpora lutea of the same age. The vacuolization is usually most marked in the periphery and progresses toward the center. Other lutein cells are still more solid and mitoses in lutein cells can be seen in the majority of the corpora lutea of this period. In cases in which mature follicles are present and a spontaneous rupture of follicles is therefore soon to be expected, the corpora lutea show much vacuolization; but here also mitoses are still present in lutein cells.

In some cases the retrogressive changes are still further advanced and a connective tissue capsule may appear in the periphery of the corpus luteum. The marked vacuolization of peripheral lutein cells may be accompanied by a diminution in the lumen of blood vessels. Vessels with coats consisting of several rows of cells are seen regularly in these corpora lutea. The connective tissue in the center of the corpora lutea is usually dense and relatively small in amount.

In those cases in which a new spontaneous ovulation had taken place the vacuolization of the corpora lutea had still further pro


gressed and under such circumstances mitoses were no longer present in them.

The corpora lutea of the preceding (II) generation, originating in an ovulation that took place at least thirty-seven days ago, are sometimes represented by small bodies which are surrounded by a thick connective tissue capsule; much fibrous tissue is found in the center and the lutein cells between these two zones show very large vacuoles. The vessels remaining in such structures have very thick cellular walls. In other cases some yellow pigment appears in such vacuolar cells and in still other cases we see only yellow, atretic bodies. It is probable that the latter structures are found in cases in which a still longer time has elapsed since the preceding (second last) ovulation. There may of course have occurred a longer interval than twenty days between the last and second last ovulation.

When the second generation was represented by a corpus luteum of pregnancy, the retrogressive changes were also marked, shrinking of the corpus luteum and vacuolization of the lutein cells are pronounced, but such corpora lutea are still considerably larger sixteen to nineteen days after the completion of pregnancy than ordinary corpora lutea of the corresponding generation. Some of the vacuolar cells may show a yellow pigmentation. In such ovaries we may find a still older generation of retrogressing corpora lutea present, represented by yellow atretic bodies which owe their origin to an ovulation that took place more than a hundred days ago; and if the last named (third last) ovulation were followed by a pregnancy, this ovulation may have taken place approximately one-hundred and fifty days ago. Not in all animals are so many generations of corpora lutea found ; especially in young animals (two to three months old only one generation may be present.

If the last ovulation that took place fifteen and one half to nineteen days ago were followed by pregnancy, the follicles in the ovaries of pregnant animals of this period do not show any marked difference from the follicles of non-pregnant animals at the corresponding period after ovulation. In both cases we find good follicles of various sizes and the different stages of retrogression of


follicles which we mentioned above. In the ovaries of pregnant animals of this period we may also find mature follicles, the granulosa cells of which have more cytoplasm that stains red with eosin. Such follicles show less granulosa degeneration and a decrease in the number of mitoses is visible in the granulosa cells. Some degeneration of granulosa cells may however occur in these follicles and their further fate will still have to be determined.

The corpora lutea of pregnancy (first generation) are well preserved. Fine vacuoles may however be present, especially in the peripheral lutein cells. Mitoses are also present. They do not show such pronounced signs of retrogression, as occur in corpora lutea of non-pregnant animals of this period.


At this period the proportion of animals in which a spontaneous ovulation had taken place, notwithstanding the separation of females and males, is much greater than in the preceding period. Among twenty-two guinea pigs a spontaneous ovulation had taken place in eight, while in the fourteen other females no rupture of follicles had as yet occurred. In at least one and possibly in more of these fourteen guinea pigs a rupture was however imminent, as indicated by the presence of mature, red-staining follicles. In those animals in which ovulation had taken place within the last few days the follicles were in the condition corresponding to that stage after ovulation. The corpora lutea that originated as a result of the ovulation twenty to twenty-six days previously showed marked degeneration; the cells were vacuolar; in one case the lutein cells formed a hyaline material in which the vesicular nuclei were imbedded. Mitoses were present in only one case, in which the rupture had taken place apparently within the last twenty-four hours, but even vacuolar cells may divide mitotically. Many blood vessels have thick cellular coats and the blood vessels in general do not seem to be patent.

In all the other guinea pigs in which a new rupture of follicles had not yet taken place the follicles behave approximately in the same manner as in the previous stage; we see follicles of various


sizes without granulosa degeneration/ and follicles of large and also of medium size in various stages of granulosa and connective tissue atresia.

In the ovaries of the guinea pig in which a spontaneous rupture of follicles was imminent, the twenty-two days old corpora lutea also showed the signs of early degeneration; some of the cells were still good, but the majority were vacuolar.

In the guinea pigs, in which a spontaneous ovulation had not yet taken place, the corpora lutea of the last ovulation were also in a process of degeneration, which was especially marked during the later stages, twenty-four to twenty-six days after ovulation: here the vacuolization was very pronounced, and occasionally connective tissue began to grow into the periphery of the corpus luteum. The vessels of these corpora lutea were very thick. In some other ovaries, especially in those examined twenty and twenty-one days after ovulation, the number of relatively well preserved cells was still greater. On the whole the number of mitoses found in lutein cells at this period is distinctly diminished.

The older generations of corpora lutea are represented by atretic yellow bodies, which are however not present in all animals. In one case a corpus luteum was present that originated as a result of an ovulation that took place approximately ninety-three days before and was accompanied by pregnancy. In this cases twentyseven days after delivery very little of the lutein tissue was left, the blood vessels had very thick coats, and the fibrous tissue of the remnant of the corpus luteum was very prominent.

If the ovulation which took place twenty to twenty-seven days before were followed by a pregnancy, no new spontaneous ovulation took place, 'the conditions of the follicles was the same as in those guinea pigs in which the last ovulation was not followed by pregnancy and in which no new spontaneous ovulation had as yet taken place. The corpora lutea of pregnancy of this period showed much less vacuolization, although a slight amount of it may have been present, especially in the periphery of the corpus luteum. Mitoses were more common in these corpora lutea of pregnancy than in the ordinary corpora lutea of the same period. Their size was also greater.



In regard to the ordinary corpora lutea and the corpora lutea of pregnancy of previous generations, the same retrogressive changes which were described above in the ovaries of non-pregnant guinea pigs of this period, were found in pregnant animals.

We see therefore that the condition of the corpora lutea indicates the condition of the follicles, and conversely the condition of the follicles indicates the history of the corpora lutea. At a certain time (approximately ten days) after the ovulation a certain equilibrium is reached between the growth and the degeneration of the follicles. Whether a quantitatively exact equilibrium is reached, cannot yet be stated. In proportion to the length of time which elapsed since the last ovulation, the probability of a new spontaneous rupture, with the subsequent changes in the follicles, becomes greater. At this and the preceding period signs of degeneration are present in the ordinary corpora lutea, which become the more marked the older the corpus luteum; the number of mitoses in lutein cells decreases with advancing age ; they may however still be present in corpora lutea immediately following a new ovulation; the latter however is soon followed by further progressing degeneration of the corpus luteum of the preceding ovulation. If the ovulation that took place twenty to twenty-six days previously was accompanied by pregnancy, no new spontaneous rupture of follicles took place, the proliferation of the lutein cells continued, and degenerative processes in the corpora lutea were retarded.

Approximately twenty-five days after the completion of pregnancy the corpora lutea of pregnancy (second generation) have become small vacuolar bodies with thick vessels and fibrous tissue, while corresponding ordinary corpora lutea have at this time apparently been transformed into yellow bodies.



In five animals in which, in order to prevent pregnancy, both (and in one case one) of the Fallopian tubes had been ligated within twenty-six hours after copulation, and in which at a later operation incisions had been made into the uterus, no new ovulation had taken place at the time of the examination, twentysix to thirty-four days after copulation. The corpora lutea (twenty-six to thirty-four days old) showed very marked retrogression; they were very vacuolar; their size was always diminished especially after thirty-two to thirty-four days, but differed somewhat in individual cases. Some corpora lutea formed small bodies containing very dense fibrous tissue in the center and enclosing in the periphery a relatively small number of very vacuolar cells. Other corpora lutea were still somewhat larger and contained a. few better preserved cells.

Besides the retrogressing vacuolar corpora lutea some atretic yellow bodies could be found in some cases ; they were remnants of corpora lutea at least fifty days old. In two other ovaries a spontaneous ovulation had taken place recently and the condition of the follicles was in accordance with the age of the new corpora lutea. Here also the thirty to thirty-two days old corpora lutea of the preceding ovulation were very vacuolar and contained blood vessels with a thick coat and much dense fibrous tissue.


In these ovaries we found good follicles of small, medium and large size without granulosa degeneration and with mitoses in granulosa cells; other follicles showed various stages of granulosa degeneration and of connective tissue atresia. Mitoses were absent or diminished in number in follicles in which granulosa degeneration existed.

In addition to the ordinary large follicles mature or almost mature follicles were seen in which the cytoplasm of the cells was well developed, and in which the granulosa contained only very


few mitoses which were found especially in the discus proligerus. Some of the nuclei of the granulosa cells appeared somewhat contracted in these follicles, but no marked degeneration of the granulosa cells was found.

The corpora lutea of pregnancy were large, the cytoplasm of the lutein cells stained red yellow with eosin; the cell outlines were quite distinct. The large majority of the lutein cells were compact and did not show vacuoles; the nuclei were vesicular. A few mitoses were found in lutein cells. Only very little connective tissue was present in the center of the corpora lutea. Some of the vessels had thick walls, while other vessels were of a capillary character and had either a wide or narrow lumen. We see therefore that also at later stages of pregnancy the follicles continue to grow and to degenerate, and that even at this period of pregnancy follicles may mature. The lutein cells of the corpora lutea of pregnancy continue to show mitotic nuclear figures and well preserved cytoplasm at a time when, in the ordinary corpora lutea, retrogression is very far advanced.


A large number of ovaries were examined of female guinea pigs which had been kept separated from males for various lengths of time.

One set of guinea pigs was separated from males before sexual maturity had been reached. The ovaries were examined, when the animals were six and twelve months old. In every instance ovulation had taken place repeatedly and we usually found the three generations of corpora lutea which we described in the case of guinea pigs which had copulated, namely relatively young corpora lutea, retrogressing vacuolar corpora lutea and atretic yellow bodies.

In another series guinea pigs were guarded against contact with males after delivery, and were kept separated from males for various periods of time. In this case a spontaneous ovulation took place after delivery, at least in the majority of cases, even


without contact with males, and subsequently further ovulations occurred. Under such conditions the successive ovulations do however not occur in the same intervals in all animals; in some cases a delay in ovulation may take place : this accords well with our previous observations. The conditions of the follicles correspond to the time elapsed since the last ovulation, as indicated by the state of the corpora lutea.

Not in every case however does a spontaneous ovulation take place without contact with male. In several cases neither new nor retrogressing corpora lutea could be found in the ovaries of guinea pigs which, according to their age, ought to have ovulated, but in which no sign of heat had been noticed during an observation extending over a certain period of time. In other guinea pigs which had been in heat recently, but in which copulation had been prevented, no new ovulation corresponding to the period of heat had taken place at the time of examination.


In connection with the cyclic changes in the adult ovary of the guinea pig, just described, we thought it of interest to determine the time at which these cyclic changes set in. For this purpose we studied a series of ovaries at differents stages of the growing guinea pig 1. In the ovaries of a fetus near the time of birth many follicles are present in the cortex. These follicles have not yet a cavity and the largest follicles have a granulosa consisting of three, or four rows of granulosa cells ; in the latter some mitoses can be seen. No distinct differentiation appears in the connective tissue of the different parts of the ovary.

2. In the ovaries of guinea pigs four, five and seven days old we find a cavity in a certain number of the follicles ; no atretic processes have as yet taken place. The theca interna cells are distinguished from the surrounding connective tissue through the


increase in the size of their nuclei. The connective tissue around the medullary canals is relatively dense. In the granulosa, theca interna and in the ordinary connective tissue stroma mitoses are frequent.

3. The ovaries of guinea pigs eighteen days old are larger; the follicles also have increased in size. Small and medium sized and in proportion to the as yet small size of the ovaries, relatively large follicles are present. In some of the follicles degenerative processes appear at this time, but the extent to which such changes have taken place differs in the ovaries of different animals. In the ovaries of some guinea pigs no degeneration of the granulosa has as yet taken place. In the ovaries of another guinea pig a few follicles showed a trace of granulosa degeneration, while in another follicle the granulosa degeneration was pronounced.

In the follicles of some ovaries we find even a beginning ingrowth of connective tissue into the follicular cavity, and in one case a cavity of a follicle was filled with loose connective tissue. The majority of the follicles are in a good condition; their cavity is larger than at the preceding stage and the interna is better developed and consists of more rows of cells. Mitoses are present in the theca interna and in the granulosa. The connective tissue between the follicles is a little more fibrous, and around certain blood and lymph vessels it is somewhat edematous and rarefied.

4. In the ovaries of guinea pigs twenty-eight days old the majority of follicles are in good condition and non-atretic; they are of small and medium, but not yet very large size. In some ovaries hardly any degeneration of follicles is visible ; in others we see some follicles which have not yet attained their full size (corresponding to the as yet small size of the ovaries) , presenting various stages of granulosa degeneration. In some follicles the granulosa has been entirely destroyed and connective tissue begins to grow into the cavity.' In some cases we find quite atretic connective tissue follicles. In some small and medium sized follicles the ova may undergo (probably amitotic) nuclear division and a corresponding segmentation of the cytoplasm, the granulosa being still intact.


In other cases, however, such ova are surrounded by connective tissue.

The connective tissue of the ovaries shows more differentiation at this period and is somewhat more fibrous.

5. In ovaries of guinea pigs one to two months old the size of some of the folhcles, in correspondence with the growth of the ovaries, enlarges. We see various stages of granulosa degeneration and of connective tissue atresia. Granulosa degeneration may take place in medium sized and in large follicles. In some ovaries the large majority of follicles may show either granulosa degeneration or connective tissue atresia. Corpora lutea are not yet visible.

6. Ovaries of guinea pigs three months old: Approximately at this period the ovaries have become mature. We find various stages of developing follicles and occasionally mature follicles ready to rupture. We find the various stages of granulosa degeneration and of connective tissue atresia. We notice a greater differentiation in the structure of the stroma in different parts of the ovary.

Corpora lutea, which occasionally are already in the beginning of degeneration, are present in some ovaries; in other animals ovulation has not yet taken place.

It follows from these observations that degenerative processes in follicles set in approximately fourteen to eighteen days after birth, and ovulation and formation of corpoa lutea appear in guinea pigs two to three and a half months old. The ovaries and follicles must have reached a certain size, before ovulation sets in. The time required for the development of small into large follicles, with subsequent beginning of degenerative processes, is somewhat longer in the young growing animal than in the mature guinea pig, but in both the periods of time are of a similar order (approximately nine and fourteen days respectively).



The principal result of our investigations we can state as follows : In the ovary of the guinea pig (and probably of mammals generally) cyclic changes take place independently of copulation and of pregnancy.

A sexual period (the period between two ovulations) is accompanied by a series of changes in the follicles. As a result of the conditions leading to or accompanying ovulation the granulosa of all large and medium follicles undergoes a very rapid degeneration, which is very marked within an hour or two after ovulation, or perhaps even sooner. In the follicles in which the cavity is as yet very small, the degenerative processes are very slight or absent. These follicles do not seem to perish. These degenerative changes affect equally both ovaries of one animal, even if a rupture of follicles should have taken place in only one of the two ovaries. The local effect of the rupture of the follicle can therefore not be the cause of the follicular degeneration. Within the next few days the small follicles grow and gradually attain a large size. Eight days after ovulation large follicles are again noticeable. As soon as good sized and medium sized follicles have been formed they begin to undergo degenerative processes, the granulosa degenerating and becoming dissolved and connective tissue growing into the follicular cavity. This process ends in an almost complete disappearance of these follicles. In the meantime other follicles grow and, having reached a large size, they also degenerate. Thus after a first stage of general growth, comprising approximately ten days after ovulation, a certain equilibrium is reached in which new follicles are growing to a certain size, and in which other follicles of large or medium size degenerate. Whether certain quantitative differences in the proportion of the number of growing and degenerating follicles exist at different periods of this second part of the sexual cycle, will still have to be determined. This second period of equilibrium begins approximately ten days after the last ovulation, and it lasts until a new ovulation occurs. Gradually a few large follicles undergo still further changes, the cytoplasm of their granulosa cells enlarges, the number of mitoses in these


cells decreases and they become more resistant to those processes which lead to degeneration in other follicles. The follicles in which such changes have taken place are mature and ready to rupture. In the meantime the follicles that ruptured during the preceding ovulation developed into corpora lutea. The latter represent principally the hypertrophic granulosa cells of the ruptured folUcles, which proliferate mitotically. After a certain stage of development has been reached, degenerative processes set in in the corpus luteum, which start in its periphery and proceed to the center. These degenerative processes set in very early, are noticeable eighteen to twenty days and are usually marked twenty to twenty-four days after the preceding ovulation. Throughout this period of beginning degeneration, however, some mitoses are still visible in certain lutein cells. At this period usually a new ovulation takes place. The exact time at which the new ovulation occurs varies however somewhat in different animals, ovulation occurring earlier in some animals than in others. In some cases it can be hastened through certain external factors, especially copulation, but in the large majority of cases it occurs sooner or later even without a preceding copulation.^ After the new ovulation has taken place, the degenerative processes progress in the corpus luteum, although within the first twenty hours after ovulation mitoses may still be found in certain lutein cells. In the following period a considerable shrinking of the corpus luteum takes place ; the connective tissue in the cortex and in the periphery becomes hyaline and forms a relatively prominent part enclosing a small number of very vacuolar cells. Gradually yellow pigment is deposited in these vacuolar cells and thus the corpora lutea become transformed into the atretic yellow bodies. The new ovulation was of course again followed by the typical changes in the follicles.

If the ovulation be followed by pregnancy, the principal changes taking place in the ovaries are on the whole the same. The only

■* Whether or not in the guinea pig ovulation can take place independently of a preceding copulation has been a subject of controversy. Concerning the literature of this question see William H. Kirkham, Biological Bulletin, vol. 18, no. 5, April, 1910.


difference consists in a prolongation of the sexual cycle, which lasts as long as the pregnancy continues. The changes in the follicles are identical with those found in the ordinary sexual period not accompanied by pregnancy.

After copulation the period of growth following the sudden degeneration of the follicles is the same as in the ordinary sexual period, but the period of follicular equilibrium is much prolonged.

During this period of follicular equilibrium certain follicles can not only grow to a considerable size, but may even undergo the additional changes which indicate the maturation of the follicle. A rupture of follicles does not however take place during pregnancy under ordinary circumstances.

The corpus luteum of pregnancy differs from the ordinary corpus luteum mainly in its prolonged duration of growth and of life. At a time when, in the ordinary corpus luteum not accompanied by pregnancy, mitoses have ceased to be present and the retrogressive changes are very marked, mitoses are still seen in the corpus luteum of pregnancy. In the corpus luteum of pregnancy degenerative changes set in before the end of pregnancy has been reached, and they continue after delivery. A short time after delivery a new ovulation usually occurs, even if no copulation had taken place after delivery. The retrogression of the corpora lutea of pregnancy continues, but it requires much more time than the retrogression of an ordinary corpus luteum.

The mechanism that governs the sexual cycle in the ovary can be recognized only incompletely by observation and it has been the subject of an experimental investigation, the results of which we shall report in more detail elsewhere. We may however state that our experiments have shown that through extirpation of the corpora lutea the sexual cycle is shortened. The presence of well functioning corpora lutea inhibits a new ovulation. Pregnancy as such does not prevent ovulation. Ovulation can be made to take place even in pregnancy, if the corpora lutea be extirpated at an early period after copulation. And under such conditions the typical follicular changes follow the ovulation during pregnancy. As soon therefore as degenerative processes have set


in in the corpora lutea, either during pregnancy or outside of pregnancy, a new ovulation can take place. How far the presence of ,the corpora lutea influences the transformation of ordinary large follicles into mature follicles and how far its action merely concerns the rupture of the mature follicles, remains still to be determined.

It follows from our observations that the time of ovulation depends upon at least three different factors: (1) Changes taking place in the ovaries. It is necessary that mature follicles have been produced, before rupture can take place. Our experiments indicate that cuts into an ovary causing an opening of a follicle may possibly lead to the formation of a corpus luteum only at a time when mature follicles are present. A certain time must therefore have elapsed after ovulation before another ovulation can take place. During this period small follicles reach their full size. Thus a minimal time which must elapse between two ovulations is required. (2) The time at which the influence of the corpus luteum preventing ovulation ceases to be exerted. Our observations make it very probable that the retrogressive changes observed in the corpora lutea before ovulation indicate the necessary cessation of functional activity. It is however noteworthy that, notwithstanding such a cessation of activity, mitoses can still be observed in the lutein cells at this period. Whether the corpus luteum acts principally upon the last stage in the development of follicles (maturation) or merely upon the rupture of follicles will still have to be determined with certainly. We recall however the fact that we observed the occurrence of mature follicles during various stages of pregnancy, notwithstanding the existence of corpora lutea. (3) Certain more or less accidental conditions, as for instance copulation. It is probable that other circumstances also may accelerate or retard the rupture of the follicles. Such factors act probably indirectly by causing changes in the circulation in the ovaries. In the guinea pig these are not indispensable, but their place can be taken by other factors; or even the total absence of corpora lutea may in the guinea pig be sufficient to allow a new ovulation.


In the guinea pig ovulation occurs in the large majority of cases without any previous copulation. In many cases however copulation is not without significance even in the guinea pig; it accelerates ovulation. While, after delivery, a spontaneous rupture may take place without copulation, in other cases it does not occur without ovulation. Also in the ordinary period of heat ovulation does not need to take place without copulation. Copulation is therefore not without importance; but in almost all of these cases ovulation is only deferred and sooner or later it will take place without the male. So far as the literature has been accessible to us it appears that the role copulation plays had not been fully appreciated by former investigators. Certain observations which we made indicate that other factors besides a preceding copulation may influence ovulation, and we intend to continue our investigation in this direction.

Our observations enable us to give some data concerning the time relations in the growth of various ovarian structures.

a Follicles. In about six days after ovulation small follicles reach medium size. In approximately eight days large follicles have developed and now degenerative processes set in. Mitotic cell division is most pronounced in the granulosa before degenerative processes have commenced; but mitoses may still be seen, if a slight degree of degeneration exist.

h Ordinary corpora lutea. The development of corpora lutea within the first six days after ovulation has been described in a previous paper. At six days we see for the first time, besides the capillary vessels, blood vessels with walls consisting of two rows of cells penetrating into the corpus luteum; they become somewhat more frequent from the tenth day on. In the meantime mitotic division of lutein cells continues and the increase in these cells causes the central connective tissue to become smaller in amount.

In corpora lutea ten to eleven days old a few vacuolar cells are present in the periphery of the corpus luteum. From ten to fifteen days after ovulation vacuolization is still very slight in peripheral luetein cells. From fifteen to eighteen days more fine or coarse vacuolization may appear. Other lutein cells are still


more solid and mitoses are still present. If no new ovulation have taken place, degeneration becomes more marked after twenty days; twenty-four days after ovulation we noticed a small amount of connective tissue growing into the periphery. At this period the number of mitoses is already diminished. In cases in which, between the eighteenth and twenty-sixth day after ovulation, a new rupture of follicles sets in, the degenerative processes are still more marked; mitoses may still be seen in the course of the first day after rupture of the follicles, but they disappear afterwards and the degenerative processes progress. The vacuolization of the lutein cells increases, the corpora lutea shrink, the connective tissue becomes gradually hyaline and is relatively preponderating in quantity over the lutein cells. About six days after the new ovulation (in approximately twenty-six days old corpora lutea) yellow pigment may be seen for the first time in the vacuolar lutein cells. Eight days after the new ovulation the corpus luteum is much shrunken, and ten to eleven days after the new ovulation corpora lutea approximately thirty-one to thirty-two days old have been reduced to small vacuolar bodies, around which a strong connective tissue capsule may appear. Corpora lutea thirty-three to forty days old (twelve to nineteen days after new ovulation) still represent vacuolar bodies; but now gradually the transformation into a yellow body sets in. Corpora lutea about forty-five days old have the appearance of yellow bodies and they may probably persist as such for a long time, after the third ovulation has taken place. Thus three generations of corpora lutea may be present side by side in the same ovary, c. Corpus luteum of pregnancy. The corpus luteum of pregnancy differs from the ordinary corpus luteum in the longer duration of mitotic division, and the delay in retrogressive changes. Although slight vacuolization may be noticeable at relatively early stages, the corpora lutea of pregnancy are still in a good condition thirtyfive to forty days after ovulation and they may still show mitoses at this period. Towards the latter part of pregnancy however degenerative processes set in, vacuolization and loss in staining power of the nuclei, and other changes, are noticeable. Mitoses could not be seen at this stage, and they appeared to be absent


after delivery had taken place. From ten to twelve days after delivery yellow pigment was seen in a few of the lutein cells in the corpus luteum of the previous pregnancy.

Thirteen to twenty days after delivery the corpus luteum is still much larger than an ordinary corpus luteum at the same period after ovulation, but considerable shrinking has already taken place. Twenty-seven days after delivery the corpus luteum is very small and vacuolar, with much hyaline connective tissue, but has not yet been transformed into a yellow body; but at a later stage, approximately sixty days after delivery (or possibly somewhat earlier) the corpus luteum appears as a yellow body, and as such it may persist for some time.

d. In the developing ovaries degeneration of the granulosa and connective tissue atresia of follicles are found as soon as the follicles have reached a relatively large size; these retrogressive changes first appear in guinea pigs approximately fourteen to eighteen days old, while the first ovulation appears much later, namely two to three and a half months after birth.




From the Zoological Department, Columbia University



Introduction 71

Descriptive 73

1 The second spermatocyte-division in Nezara 73

a The idiochromosomes 73

b The double chromosome 77

2 The first spermatocyte-division 78

3 The growth-period and spermatocyte-prophases 80

4 The diploid chromosome-groups 83

General 84

5 The idiochromosomes 84

a Composition and origin cf the XY-pair 85

b Modifications of the X-element 88

c Sex-limited heredity 94

d Secondary sexual characters 99

6 Modes in which the chromosome-number may change 99

Conclusion 105


In the first of these 'Studies' ('05a) I described the idiochromosomes (X and Y-chromosomes) of Nezara hilaris as being of equal size in the male, and reached the conclusion that in this species no visible dimorphism appears in the spermatid-nuclei. In my third 'Study' ('06), after examination of the female diploid groups, this species was assigned a unique position as the single then



known representative of a type in which a pair of idiochromosomes can be identified in both sexes, but are of equal size in both, and in which, accordingly, no visible sexual differences appear in the diploid nuclei. These conclusions, as is now apparent, were based upon a wrong identification of the idiochromosome-pair, which is not the smallest pair, as I then believed, ^ but one of the largest. When this fact was recognized, the true conditions soon became evident.

I was led to re-examine Nezara hilaris by the fact (very surprising to me) that in Nezara viridula, a southern species closely similar to N. hilaris, the idiochromosomes of the male are extremely unequal in size, and the dimorphism of the spermatidnuclei is correspondingly marked. Upon returning to the study of N. hilaris it soon became manifest that the dimorphism is present in this species also, though in far less conspicuous form. The size-difference between the X- and Y-chromosomes is here often so slight that I did not at first distinguish it from an inconstant fluctuation of size, such as is sometimes seen between the members of the other chromosome-pairs. When, however, the identity of the XY-pair was correctly recognized, its constancy of position and of size in the second division enabled me to make an accurate comparison between it and the other bivalents; and this fully established the constant inequality of its members, which is constantly greater than that now and then seen in other pairs. Both species also exhibit some other very interesting features that I overlooked in my former studies.

Nezara can therefore no longer stand as a representative of the third of the types distinguished in my third 'Study,' but belongs with Euschistus, Lygaeus, etc., in the second type

1 This was in part because in most of the other forms known at the time the idiochromosomes are in fact the smallest, or one of the smallest, pairs. In part, also, I followed Montgomery ('01) who described in this species two small " chromatin nucleoli" in the spermatogonial groups, and believed them to be identical with the chromatic nucleolus of the growth-period. In a later paper ('06) Montgomery states these "chromatin nucleoli" to be "apparently not quite equal in volume," and asserts that I was in error in describing them as equal. In my material they are certainly equal in the great majority of cases. However, this is not the idiochromosome-pair.



Since the two species agree very closely save in respect to the idiochromosomes they may conveniently be considered together. Before describing the divisions, attention may be called to a striking difference between the two species in respect to the size of the cells and karyokinetic figures. As a comparison of the figures will show, the spermatocytes and maturation divisionfigures of N. hilaris are much larger than those of N. viridula. In the spermatogonia this difference is also apparent, though less marked. In the ovaries, strange to say, it cannot certainly be detected, either in the dividing cells or in the nuclei of the folliclecells or of the tip-cells at the upper end of the ovary. It would be interesting to make a more accurate study of these relations; but I will here only state that the differences between the two species seem to arise mainly through greater growth of the spermatocytes in N. hilaris. With this is correlated a greater size of the testis as a whole; but the size of the entire body in this species is but little larger, as far as I have observed, than in N. viridula.

As regards the general features of the divisions, the diploid groups of both sexes uniformly contain fourteen chromosomes, the first spermatocyte-division eight and the second seven, the idiochromosomes being, as is the rule in Hemiptera, separate and univalent in the first division.

1. The second spermatocyte-division

a. The idiochromosomes. Polar views of the second division always show 7 chromosomes which are usually grouped in an irregular ring of six with the seventh near its center (fig. 3 j-m, figs. 14, 15). In both species one chromosome of the outer ring (s) can usually be distinguished as the smallest, though this is not always evident owing to the apparent variations produced by different degrees of elongation. This is the chromosome that I formerly supposed to be the idiochromosome-bivalent, despite its peripheral position, and despite the fact, which I had myself described, that a similar small chromosome, also peripheral in posi JOUBNAL OF MORPHOLOGY, VOL. 22, NO. 1


Figures 1 to 9 are from camera drawings, and are not schematized except that in a few instances the chromosomes have been artificially spread out in a series in order to facilitate comparison. Figs. 2 k-l are somewhat more enlarged than the others. In all the figures d denotes the double chromosome or 'd-chromosome,' s the small chromosome, X the large idiochromosome and Y the small.

Fig. 1 The second spermatocyte-division in Nezara viridula. a-d, metaphases in side view; e-gr, anaphases; h, i, polar views of two sister-groups, middle anaphase, from the same spindle and in the same section.

tion, appears in several other pentatomids (e.g., in Euschistus, Coenus and Mineiis). But Nezara forms no exception to the rule that the central chromosome is the idiochromosome-bivalent. In N. viridula this is immediately apparent in side views (often also in polar views) where the central chromosome is seen to consist of two very unequal components, the smaller being not more than one fourth or one fifth the size of the larger (fig. 1 a-c). In the anaphases these separate and pass to opposite poles, while all the others divide equally (fig. 1 e-g). Polar views of middle or rather late anaphases, when both daughter-groups can be seen superposed in the same section, clearly show the marked difference of the two groups in respect to the idiochromosomes (fig. 1 h-i). All the facts are here so nearly similar to those seen in Euschistus or Lygaeus as to require no further description.


In N. hilaris the conditions differ only in tiiat the two components of the central chromosome are but sUghtly miequal; but in the examination of at least two hundred of these divisions I have never failed to detect the inequaUty. A series of side views are shown in fig. 2 a-i, figs. 16-21, two of which show all the chromosomes. These figures illustrate practically all the variations that have been seen in the idiochromosomes. The most characteristic condition is that seen in 2 a, b, d, in which both idiochromosomes (X and Y) are more or less elongated and united end to end. Less often one of them assumes a more spheroidal form (fig. 2 e, h, i, fig. 17). The size-difference, though always evident, seems to vary slightly (perhaps because one or the other component may be more or less compressed laterally), but is always distinctly greater than that now and then seen in other bivalents.

Fig. 2 j shows a mid-anaphase^ (cf. figs. 21-23) in which the inequality would hardly be noticed without close study and the comparison of other cases. Figs. 2 k and I are similar stages showing all the chromosomes spread out in a series for the sake of comparison. In both, the two idiochromosomes are easily distinguishable,^ and the larger is seen to be 07ie of the three largest chromosomes. Figs. 2 m-n, o-p, q-r and s-t are pairs of sistergroups, in each case from the same spindle in anaphase. All of these are selected from cases in which a distinct size-difference appears between X and Y, but there are also many cases in which this cannot be seen. Such a case was figured in fig. 4 e-f of my first ' Study' the correctness of which is confirmed by re-examination of the original section. This condition is due simply to the fact that the large idiochromosome is more elongated than the small, so that the size-difference cannot be seen in polar view; and for the same reason it is often not evident in polar views of the metaphase.

2 This and the two following figures are a little more enlarged than the others .

' Fig. 2 I is the same group figured in fig. 4 d of my first 'Study,' carefully redrawn and corrected. A comparison of the two drawings will show that in the latter a distinct size-difference between X and Y is actually shown but is minimized by the fact that the former is represented a trifle too small, the latter a little too large. It is now also evident that they are connected by two connecting fibres instead of by one.

Fig. 2 The second spermatocyte-division in Nezara hilaris. a-i, metaphase figures in side view, a and e showing all the chromosomes; j~l, mid-anaphases; in A; and I all the chromosomes are shown artificially spread out in series; m-n, o-p, q-r, s-t, four pairs of sister-groups from late anaphases, in polar view, in each case from the same spindle.

b. The double chromosome. A seco;iid interesting feature of the second division that I formerly overlooked is the presence of a remarkable double chromosome which in the metaphase has exactly the appearance of a butterfly with widespread wings. This chromosome (which may be called the d-chromosome) is shown in profile view in 2 b~e and 1 a-d, 16, 17, 20, 24, 25. This is the only chromosome in the second division that shows any approach to a quadripartite form, audits characters are so marked as to constitute the most striking single feature of the division. As the figures show, it is one of the largest of all the chromosomes. It always has an asymmetrical tetrad shape, giving exactly the appearance of a smaller and a larger dyad in close union; and it always lies in the outer ring, so placed as to undergo an equal division, and with the larger wings of the butterfly turned towards the axis of the spindle. In polar view (3 j-m) the duality is far less apparent and sometimes invisible, even upon careful focussing. In N. viridula the duality is always apparent in side view, but the butterfly shape is usually less evident than in N. hilaris.

In the initial anaphases the (i-chromosome divides symmetrically, drawing apart into two bipartite chromosomes (2 j, k, I g); but this is seldom evident save in profile view. Viewed from the pole the duality does not now ordinarily appear, though it may still sometimes be seen upon careful focussing. In the later anaphases the two components tend to fuse, and often can no longer be distinguished. Not seldom, however, the duality is visible even in the final anaphases; and sometimes this is so conspicuous that the spermatid-group seems at first sight to comprise eight instead of seven separate chromosomes (n, r, s, t).

Since the duality of this chromosome does not certainly appear in the spermatogonial groups or in the first spermatocyte-division, its peculiar form in the second division might be supposed to result from some special mechanical relation to the spindle-fibers in that division. This is, however, excluded by examination of the interkinesis, in which the chromosomes are irregularly scattered. • In these stages, even when the spindle is still very small and the chromosomes lie in a quite irregular group, the butterfly shape is already perfectly evident; and it shows no constancy of


relation to the spindle-axis^ often lying at right angles to the latter. Apparently therefore its duality arises quite independently of the spindle or astral rays, and its constant position in the fully formed spindle is the result of a later adjustment. In this species, as in many others, each chromosome is connected with the pole by a bundle of delicate fibers. In case of the d-chromosome this bundle is very broad, but I cannot be sure that it is double.

At first sight any observer would, I think, take the c?-chromosome to be merely a result of the accidental superposition or close adhesion of two separate dyads of unequal size ; but such an interpretation is inadmissible. When all the chromosomes can be unmistakably seen, the d-chromosome is found to constitute one of the seven separate elements invariably present in this division; and since the diploid number is 14 in both sexes this chromosome must represent one chromosome, not two, of the original spermatogonial groups. It is certain, therefore, that the double appearance does not result from close apposition of two separate chromosomes; it is therefore not a tetrad" in the ordinary sense of the word — i.e., not one that results from the synapsis of two chromosomes that are originally separate in the diploid groups.

2. The first spermatocyte-divisiori

This division requires only brief mention. As stated, it shows eight separate chromosomes, of which the only one that can be positively identified is the Y-chromosome of N. viridula. This chromosome, always immediately recognizable in this species by its small size (3 c, d, f, g, i), figs. 12, 13), is usually central in position, like the m-chromosome of the Coreidae, but this is not invariable. Since it divides equally, and without association with any other chromosome (3 g) it is evident that the two idiochromosomes must be separate and univalent in this division. In N. hilaris (3 a, b, figs. 10, 11) the eight chromosomes usually form an irregular ring, there is no central chromosome, and neither idiochromosome can be certainly recognized. It nevertheless seems a safe inference from what is seen in N. viridula that the two idiochromosomes are here also separate and univalent.

Fig. 3 First and second spermatocyte-divisions in the two species of Nezara. a, 6, first division, hilaris, polar views: c, d, corresponding views of viridula; first division, hilaris, side view showing five of the chromosomes in position and the other three to the right above;/, corresponding view of viridula; g, middle anaphase, viridula, showing division of Y; h, first division metaphase, hilaris, all the chromosomes artificially spread out in series; i, corresponding view of viridula; 2, k, second division metaphase, hilaris, polar views; I, m, corresponding views of viridula.


In this division the d-chromosome can not be identified in either species. Figs. S e, f, h, i, show all the chromosomes of the two species, in each case from a single spindle in side view. Most of them have a simple bipartite form, but in each species two or three of them often appear more or less distinctly quadripartite as is, of course, often the case with the bivalents in this division. In N. hilaris one of the largest chromosomes is usually more elongated than the others, and each half shows a slight transverse constriction. I suspect that this may be the d-chromosome, but cannot establish the identification.

3. The groivth-period and spermatocyte-prophases

These stages fully bear out the conclusions based upon the divisions and establish the identity of the idiochromosome-pair with the chromatic nucleolus of the growth-period. Throughout the growth-period each nucleus contains a single intensely staining spheroidal chromatic nucleolus and in addition a very large, ■clearly defined pale plasmosome, which is sometimes double. Series of drawings of these two bodies (in each case from the same nucleus, and in their relative position) are given in figs. 4 i-l and m-p, from cells of the middle growth-period. They are also shown in figs. 26-29. In these stages no sign of duahty is to be seen in the chromatic nucleolus, even after long extraction or in saffranin preparations. In later stages, as the chromosomes begin to condense, this nucleolus becomes less regular in outline, and gradually assumes a tetrad form, which becomes very clear as the chromosomes assume their final shape. This transformation may be traced without a break, successive stages being often seen within the same cyst. Just before the nuclear wall breaks down this tetrad is still clearly distinguishable from the others by its asymmetrical quadripartite form, as seen in 4 y, z, which show all the chromosomes (in each case from two successive sections). Figs. 4 q-t show four views of this tetrad at this period in N. hilaris, while u-x are corresponding views of N. viridula. These figures (which might be indefinitely multiplied) show the marked differences between the two species in respect to this tetrad, obviously corresponding to that seen between the idiochromosome

Fig. 4 The diploid groups, nucleoli of the growth-period, and late prophasefigures of the two species of Nezara. a, b, spermatogonial groups, hilaris ; c, d, the same, viridula; e,f, ovarian groups, hilaris; g, h, the same, viridula; i-l, chromatic nucleolus and plasmosome from the growth-period, in each case from the same nucleolus in their relative position; m-p, corresponding views, viridula; q-t, the idiochromosome-tetrad (chromatic nucleolus) from prophase nucleoli, hilaris; u-x, corresponding views, viridula; y, late prophase nucleus, showing all the chromosomes, hilaris (combination figure from two sections) ; z, corresponding viridula, three of the chromosomes from adjoining section at the right.


bivalents of the two in the second division. ^ The two species may in fact readily be distinguished by mere inspection of the chromatic nucleolus at this period. Already at this time the two components are here and there seen to be separating, but as a rule they do not finally move apart until the nuclear wall has dissolved. From this time forward they cannot be individually identified with exception of the small idiochromosome of N. viridula, which is obvious at every period.

As far as my material shows, the earlier stages of the idiochromosomes can not be so readily traced in Nezara as in some other species, and the chromatic nucleolus can not actually be followed backward to the spermatogonial telophases — as can be done in such forms as Lygaeus or Oncopeltus, of which a detailed account will be given in a later publication. The prophase -figures, however, decisively establish its identity with an unequal pair of chromosomes that divide separately in the first spermatocytedivision; and in N. viridula, one of these is certainly the small idiochromosome. It may therefore confidently be concluded that the chromatic nucleolus is identical with the idiochromosome-pair, as in so many other cases. Comparison of the divisionfigures proves that this pair can not be identical with the small pair that I formerly supposed to be the idiochromosome-pair; and this small pair is moreover usually recognizable in the prophase groups (s, in 5 y, z) in addition to the unequal pair.

The foregoing facts make it clear that in Nezara the idiochromosomes undergo a process of synapsis at the same time with the other chromosome-pairs, and that their separation before the first division is a secendary process, to be followed by a second conjugation after this division is completed. A similar process often takes place in many other Hemiptera. There are, however, some forms, like Oncopeltus, in which the idiochromosomes are always separate, from the last spermatogonial division through all the succeeding stages up to the end of the first division. In this case, which I shall describe more fully hereafter, there can be no doubt that the conjugation which follows the first division is a primary synapsis, to be immediately followed by a disjunction.

^ CJ. the earlier figures of the corresponding tetrad in Brochymena in my first 'Studv,' fis. 7.


4. The diploid chromosome-groups

In these groups the interest centers again in the identity of the idiochromosomes and the d-chromosome. Of the 14 separate chrosomomes present in the diploid nuclei of both sexes, none shows any constant indication of duality (figs. 4 a-h). The dchromosome can not, therefore, be identified in these stages. Secondly, in both species the diploid groups of the two sexes show the same relation as in other Hemiptera of this type, though this is, of course, more readily seen in N. viridula than in hilaris, owing to the small size of the Y-chromosome. In the spermatogonia! groups of this species (4 c, d) this chromosome is at once recognizable while in the female groups {g, h) it is lacking, its place being taken by one of larger size. In both sexes the small pair (s, s) is also recognizable. In this species, accordingly, the Ychromosome is confined to the male line, and the Y-class of spermatozoa must be male-producing, as in other forms.

In N. hilaris the Y-chromosome can not be identified (4 a, h), but the relation of the spermatozoa to sex-production is shown in another way, though less unmistakably than in N. viridula. As already described, the large idiochromosome or X-chromosome is one of the largest three chromosomes seen in the second division. We should therefore expect to see five largest chromosomes in the male diploid groups. This is clearly apparent in figs. 4 a, h, and is also shown in the corresponding figures of N. viridula (c, d) though not quite so clearly. One of these five in the male should be the X-chromosome; and if the usual relation of the spermatozoa to sex hold true, there should be six largest chromosomes in the diploid groups of the female. This relation actually appears in nearly all cases, and is shown in figs. 4 e, f, g, h, in each of which, again, the small pair (s, s) may be recognized. Though this evidence is in itself less convincing than that afforded by N. viridula (since the relation can not always be made out with certainty) it is fully in harmony with the latter, and sustains the same conclusion.^

^ This relation is shown in my original figures of N. hilaris, though not quite as clearly as in the groups here figured. In my first 'Study' ('05) the five largest chromosomes are very clearly shown in fig. 4 h, and are also evident in 4 q. In the third 'Study' the relation is hardly evident in the male but fairly so in the female (figs. 5 I, m).



5. The idiochromosomes

The case of Nezara shows how readily a morphological dimorphism of the spermatid-nuclei may be overlooked when the X- and Y-chromosomes do not differ markedly in size ; and it emphasizes the necessity for the closest scrutiny of these chromosomes in the study of this general question. In my fourth 'Study' I placed with Nezara hilaris, as a second example of my original 'third type/ the lygaeid species Oncopeltus fasciatus (Dall.), on the strength of Montgomery's account of the conditions in the male ('01, '06) and my own unpublished observations on both sexes. While I have carefully re-examined this case also, I am not yet prepared to express an unqualified opinion in regard to it. Certainly, in very many of the cells of this species, at every period of the spermatogenesis, the idiochromosomes (which are always separate up to the second division) seem to be perfectly equal. A slight inequality may indeed be seen in some cases; but as far as I can yet determine this seems to fall within the range of the sizevariation in other chromosomes and gives no positive ground for the recognition of a morphological dimorphism in the spermatozoa. A similar condition has been described in several other insects, notably in some of the Lepidoptera (Stevens, '06; Dederer, '08; Cook, '10), in the earwig Anisolaba (Randolph, '08) and apparently also in the beetle Hydrophilus according to Arnold ('08). I see no reason to question these observations; but the interpretation to be placed on them is by no means clear at present. The experimental evidence on the Lepidoptera seems to demonstrate that in at least one case^that of Abraxas according to Doncaster and Raynor, — it is the eggs and not the spermatozoa that are sexually dimorphic ; that is, in the terms that I have recently suggested ('10a), in this case it is the female that is sexually 'digametic' whUe the male is 'homogametic' Baltzer's careful work on the seaurchins ('09) clearly demonstrates a cytological sexual dimorphism in the mature eggs of these animals, and shows that the spermnuclei are all alike. In cases, therefore, where no visible dimorphism of the spermatid-nuclei is demonstrable, two possibilities


are to be considered, namely, (1) that it may be the female which (as in sea-urchins) is the digametic sex, and (2) that one sex or the other may still be physiologically digametic even though this condition is not visibly expressed in the chromosomes. The first of these possibilities may readily be tested by cytological examination of the female groups. The second can only be examined by means of experiment, and especially by experiments on sex-limited heredity. It is interesting that the work of Doncaster and Raynor, cited above, and the more recent one of Morgan on Drosophila ('10) have given exactly converse results, the former demonstrating a sexual dimorphism of the eggs, the latter of the spermatozoa. This agrees with the cytological data, as far as they have been worked out. The researches of Stevens ('08, 10), on the Diptera establish the cytological dimorphism of the spermatozoa in these animals, while all observers of the Lepidoptera have thus far failed to find such dimorphism in this group. It thus becomes a very interesting question whether a cytological dimorphism of the mature eggs may be demonstrable in the Lepidoptera; though a failure to find it would in no wise lessen the force of the experimental data. Physiological differences between the chromosomes are of course not necessarily accompanied by corresponding morphological ones — indeed such a correlation is probably exceptional.

(1) (a) Composition and origin of the XY~pair. The facts seen in Nezara again force upon our attention the puzzle of the Y-chromosome or 'small idiochromosome.' It is remarkable that two species so nearly akin as N. hilaris and N. viridula should differ so widely in respect to this chromosome; though this is hardly so surprising as the fact that in Metapodius this chromosome, as I have shown ('09, '10) may actually either be present or absent in different individuals of the same species. These facts show, as I have urged, that although the Y-chromosome shows a constant relation to sex when it is present, it can not be an essential factor in sex-production. As the case now stands this might be taken as a direct piece of evidence against the view that the idiochromosomes are concerned with sex-heredity. Further, as I have pointed out ('10) in Metapodius the introduction of super


numerary Y-chromosomes into the female has no visible effect upon any of the characters of the animal, sexual or otherwise; and this might be urged against the whole conception of qualitative differences among the chromosomes and of their determinative action in development. It is especially in view of these and certain other general questions that I wish to indicate some of the many possibilities that must be taken into account in the consideration of this problem. My discussion is throughout based upon the assumption that the chromosomes do in fact play some definite role in determination, and that they exhibit qualitative differences in this respect. I do not hold that they are the exclusive factors of determination; though it is often convenient, for the sake of brevity, to speak of them as if they were such.

(2) Cytologically considered, the morphological dimorphism of the spermatozoa seems to have arisen by the transformation of what was originally a single pair of chromosomes comparable to the other synaptic pairs. We have at present no information as to whether the members of this pair were equal or unequal in size; but in either case there are grounds for the assumption that its two members differed in some definite way in respect to the quality of the chromatin of which they were composed. This pair, which may be called the priixiitive XY-pair, has undergone many modifications in different species, but without altering its essential relation to sex. In the insects (Hemiptera, Coleoptera, Diptera) its most frequent condition is that of an unequal pair, consisting of a 'large idiochromosome' or 'X-chromosome,' and a small idiochromosome" or ' Y-chromosome,' the latter being confined to the male line, while the former appears in both sexes — single in the male and paired in the female. That all gradations exist between cases where X and Y are very unequal (as in many Coleoptera and Diptera and in some Hemiptera) and those in which they are nearly or quite equal (Mineus, Nezara, Oncopeltus) gives some ground for the conclusion that in the original type the XY-pair was but slightly if at all unequal.

By disappearance of the free Y-member of this pair has arisen the unpaired odd or 'accessory' chromosome, which accordingly


has no synaptic mate. This condition seems to have arisen in more than one way. It is almost certain that in many cases the Y-chromosome has disappeared by a process of gradual and progressive reduction (as indicated by the graded series observed in the Hemiptera (Wilson, '056, '06). In some cases (of which Metapodius is an example) the same result may have been produced suddenly by a failure of the idiochromosomes to separate in the second spermatocyte-division (Wilson, '096). A third possibility, first suggested by Stevens ('06), is that the X-element may have separated from a YY-pair with which it was originally united. This possibility seems to be supported by recent observations on Ascaris megalocephala, where the X-chromosome is sometimes fused with one of the other pairs, sometimes free (Edwards, '10).

(3) We have as yet no positive knowledge as to how the Xniember of the XY-pair originally differed, or now differs, from the Y, or as to how this difference arose— a definite answer to these questions would probably give the solution of the essential problem of sex. There are, however, pretty definite grounds for the hypothesis that the X-member contains a specific ' X-chromatin' that is not present in the Y-member, and that the XY-pair is heterozygous in this respect. If this be so, the primary sexual differentiation is therefore traceable to a condition of plus or minus in this pair, accompanied by a corresponding difference between the nuclear constitution of the two sexes. (Cf. Wilson, '10a.) Further, there is also reason for regarding the heterozygous condition of this pair as due to the presence of the X-chromatin in one member of a pair which is (or originally was) homozygous in respect to its other constituents. The latter may be called collectively the 'Y-chromatin'; and we may, accordingly, think of the XY-pair as being essentially a YY-pair with one member of which the X-chromatin is associated.^ Both the X ^ This suggestion is in principle the same as one earlier made by Stevens ('06, p. 54) that the Y-chromosome represents "some character or characters which are correlated with the sex-character in some species but not in others," with one member of which the X-chromosome is fused; and that "a pair of small chromosomes might be subtracted from the unequal pair, leaving an odd chromosome."


chromatin and the Y may themselves be composite, thus giving the possibility of many secondary modifications. The point of view thus afforded opens many possibilities for an understanding of sex-limited heredity, as indicated beyond.

(6) Modifications of the X-element. This view of the XY-pair is based upon two series of facts, of which the first includes the various modifications of the X-member of the pair seen in different species. It is, perhaps, most directly suggested by a study of the pentatomid species Thyanta custator. ' In this common and widely distributed species I have found two races, which thus far can not be distinguished by competent systematists, ^ but which differ in a remarkable way in respect to both the total number of chromosomes and the XY-pair. In one of these races (which I will call the 'A form'), widely distributed throughout the south and west, the total number in both sexes is 16, and the XY-pair of the male is a typical unequal pair of idiochromosomes, exactly like that seen in many other pentatomids {e.g., Euschistus, Coenus or Banasa). These are shown in fig. 5 a, b, their mode of distribution being the usual one. The second race (the 'B form') is thus far known from only a single locality in northern New Jersey. It differs so remarkably from the A fonn that I could not believe the observations to be trustworthy until repeated study of material, collected in four successive years, established the perfect constancy of the cytological conditions and the apparent external identity of the two forms. In this race the XY-pair is represented by three small chromosomes of equal size, which are always separate in the diploid groups and in the first spermatocyte-di vision (fig. 5i), but in the second division are united to form a linear triad series (5 c, d) . This group so divides that one component passes to one pole and two to the other {oe, h), the

^ I am indebted to Mr. E. P. Van Duzee for a careful study of my whole series of specimens of both races. He could find no constant differential between them. Additional studies of this material are now being made by Mr. H. G. Barber.

Addendum. Since this paper was sent to press Mr. Barber, after prolonged study, has reported his conclusion that the 'A form' is Thyanta custator of Fabricius, while the 'B form' is probably Thyanta calceata of Say, which has long been regarded as a synonym of former species.

Fig. 5 Comparison of the XY-group in various Hemiptera. (a-i are original; the others from Payne.) a, b, Thyanta custator, 'A form,' second division in side view; c, d, corresponding views of the 'B form'; e-h, anaphases of same; i, polar view of first division of same; j, k, metaphase chromosomes, second division, Diplocodus exsanguis; i, similar view of Rocconot^ anrtailicornis; m, similar view of Conorhinus sanguisugus; n, Sinea diadema; o, Prionidus cristatus; p, Gelastocoris oculatus; q, anaphase chromosomes of the same species; r, the XYgroup, from the second division, AchoUamultispinosa; s, diagram, slightly modified from Payne, to show the distribution of the XY-components in the second division of the same species.

latter being usually in close contact and in later anaphases sometimes hardly separable (5^), though now and then all three components are for a time strung separately along the spindle in the early anaphases, so that no doubt of their distinctness can exist (5/). Comparison of the diploid groups of the two sexes shows that those of the male contain but three of these small chromosomes and those of the female four, the total respective numbers being 27 and 28 (instead of 16 in both sexes, as in the A form).

These facts make it perfectly clear that one of the small chromosomes in the male passes to the male-producing pole, and therefore corresponds to the Y-chromosome ; while the other two, taken together, represent the large idiochromosome, or X-chromosome, of the A form — precisely as in the reduvioids the single X-chromosome of Diplocodus is represented by a double element in Fitchia, Rocconota or Conorhinus (Payne). Had we no other evidence on this point we might assume simply that the original X-chromosome has divided into two equivalent X-chromosomes. But there are other facts that give reason for the conclusion that the breaking up of a single X-chromosome into separate components means something more than this. In the B form, as in' Fitchia or Rocconota (fig. 5 I), the X-element consists of two equal components, but in Conorhinus the two components are always of unequal size (5 m). In Prionidus and in Sinea there are three equal components (5 n, o), in Gelastocoris four equal ones (5 p, and in A choUa multispinosa five, of which two are relatively large and equal and three very small (5 r, s). In every case these components, though quite separate in the diploid groups (and usually also in the first spermatocyte-division) act as a unit in the second division, though not fused, and pass together to the female-producing pole (Payne, '09, '10).

In the foregoing examples the X-element is accompanied by a synaptic mate or Y-chromosome. The following are examples of a similar breaking up of the X-element into separate components when such a synaptic mate is missing. In Phylloxera (Morgan) the X-element consists of two unequal components, sometimes separate, sometimes fused together. In Syromastes (Gross,

Wilson) it consists of two unequal components, always separate, in the diploid groups but closely in contact (not fused) in both spermatocyte-divisions. The recent work of Guyer ('10) indicates a similar condition in the X-element of man. In Agalena (Wallace) there are two equal components, always separate. Finally, in Ascaris lumbricoides (Edwards, '10) there are five components, separate, and scattered in the diploid groups but closely associated in the spermatocyte-divisions.

In all these cases the significant fact is that not only the number but also the size-relations of these components are constant ; and in many of these forms this fact may be seen in such multitudes of cells, and with such schematic clearness, as to leave no manner of doubt. It seems impossible to understand this series of phenomena unless we assume that the single X-chromosome is essentially a compound body — i.e., one that consists of different constituents that tend to segregate out into separate chromosomes. We are led to suspect, further, that the composition of the Xelement, even when it is a single chromosome, may differ widely in different species because of its great variations of size as between different species. For instance, in the family of Coreidae it is in some cases very large (Protenor), in others of middle size (Chelinidea, Narnia, Anasa), in others one of the smallest of the chromosomes (Alydus). Similar examples might be given from other groups.

In the case of Thyanta, therefore, it seems a fair assumption that the double X-element of the B form likewise represents at least a partial segregation of the X-chromatin from other constituents ; and the latter may plausibly be regarded as representing the 'Y-chromatin' of the original X-member of the pair. In other words, we may think of the triad element as a YY-pair, one member of which is accompanied by a separate X-chromosome. In accordance with this its formula should be X.Y.Y, while that of the A form is XY.Y; and this may also be extended to other forms of similar type. If this be admissible, the male formula, as regards essential chromatin-content, becomes in general XY.Y and the female XY.XY, both sexes being homozygous for the Y-constituents, while in respect to X the male is heterozygous, the female homozygous. The puzzle of the Y-chromosome would thus be solved; for although a separate Y-chromosome, when present, is confined to the male line, its disappearance only reduces the male from a homozygote to a heterozygote in respect to the Y-chromatin, and the introduction of supernumerary Y-chromosomes into the female (as in Metapodius) brings in no new element.

Fig. 6 Compound groups formed by union of the X-chromosome with other chromosomes in the Orthoptera. (a and b, from Sinety, the others from McClung. ) a, triad group; first division of Leptynia, metaphase; b, division of similar triad in Dixippus; c, triad group formed by union of the l^-chromosome with one of the bivalents, first spermatocyte-prophase, Hesperotettix; d, the same element from a metaphase group; e, the same element in the ensuing interkinesis; /, the compound element of Mermiria, from a first spermatocyte prophase; g, the same element in the metaphase (now, according to McClung, united to a second bivalent to form a pentad) ; h, the same element after its division, in the ensuing telophase.

The same general view as that outlined above is suggested by the constant relation known to exist in some cases between the Xchromosome and a particular pair of the 'ordinary chromosomes.' The first observed case of this was recorded by Sinety ('01) in the phasmid genera Leptynia and Dixippus (fig. 6a,h), where the X-chromosome is always attached to one of the bivalents in the


first spermatocyte-division, and passes with one half of the bivalent to one pole. Since the spermatogonial number in Leptynia (36) is an even one and twice that of the separate chromosomes present in the first spennatocyte-division, it may be inferred that the X-element is already united with one of the ordinary chromosomes in the spermatogonia, though Sinety does not state this. Somewhat later McClung ('05) discovered essentially similar relations in the grasshoppers Hesperotettix and Anabrus (fig. 6, c-e) , and in case of the first named form was able to establish the important fact that it is always the same particular bivalent with which the X-chromosome is thus associated. In respect to the intimacy of this association, a progressive series seems to exist, since in Leptynia it seems to take place in the spermatogonia, in Hesperotettix only in the prophases of the first spermatocytedivision, while in Thyanta the union is only effected after the first division is completed.

Finally, the recent observations of Boring ('09), Boveri ('09) and Edwards ('10) seem to establish the fact that in Ascaris megalocephala the X-element, whether in the diploid groups or in the maturation-divisions, may either appear as a separate chromosome (which has the usual behavior of an accessory chromosome) or may be indistinguishably fused with one of the ordinary chromosomes.

These relations may, of course, be the result of a secondary coupling; and I m^yself formerly so interpreted them ('09c). But in view of what is seen in Thyanta or the reduvioids we may well keep in mind the possibility that they are expressions or remnants of a more primitive association, like that which I have assumed for an original XY-pair. Whatever be their origin, the effect is the same — a definite linking of the X-chromatin with that of one of the other pairs.

Fig. 7 shows, in purely schematic form, the general conception of these relations that has been suggested above, the X-chromatin being everywhere represented in black. A is the primitive XYpair from which all the other types may have been derived. By simple reduction of such a pair arises the ordinary or typical idiochromosome-pair (B) ; and from either A or B may be derived

the other types (C-G)/ or the more compHcated ones shown in fig. 5. I represents the possible mode of separation of the X-element from a YY-pair, as suggested by Stevens; and this may be reaUzed in Ascaris megalocephala (H). J and K are, schemes of the relations seen in Hesperotettix, Anabrus and Mermiria (cf. fig. 6). These may be direct derivatives of a primitive XY-pair, as the diagram suggests, or may be a result

Fig. 7 Diagram illustrating the possible relation of the various types of idiochromosomes to a primitive XY-pair. Explanation in text.

of secondary coupling of X with other elements. In either case X may itself have such a composition as is indicated in F (Protenor).

(c) Sex-limited heredity. (1) The foregoing considerations have an important bearing on the problem of sex-limited heredity, for they give us a very definite view of how such heredity may be effected. It is not my intention to consider this subject in ex

  • These figures are not intended to indicate the precise mode of segregation of

the X- and Y-chromatins of the X-element, but only illustrate possible modes.


tenso; but I wish to indicate some of the possibiUties that have been opened by the cytological results, even at the risk of offering what may be regarded as too speculative a treatment of the matter. It is obvious that atiy recessive mutation should exhibit sex-limited heredity when crossed with the normal or dominant form, if it be due to a factor contained in {or omitted from) the X-element. For instance, in the remarkable Drosophila mutants discovered by Morgan ('10) the experimental data establish the fact that white eye-color (which seems to follow the same type of heredity as color-blindness in man) is linked with a sex-determining factor in such a way that when the white-eyed male is crossed with the normal red-eyed female, the former character is never transmitted from father to son, but through the daughters to some of the grandsons (theoretically to 50 per cent), though the daughters are not themselves white-eyed ; that is, after such an initial cross, white eyes fail to appear in the Fi generation in either sex and in the F2 generation appear only in some of the males. As Morgan points out, this follows as a matter of course if the factor for white eye be identical with, or linked with, a sex-determining factor in respect to which the male is heterozygous or simplex, the female homozygous or duplex. The X-element exactly corresponds in mode of distribution to such a sex-determining factor; for this chromosome, too, is simplex in the male, duplex in the female and its introduction into the egg by the spermatozoon produces the female condition, its absence the male. This chromosome therefore, as I have shown ('06), is never transmitted from father to son, but always from father to daughter. Conversely, the male zygote always receives this chromosome from the mother. So precise is the correspondence of all this with the course of sexlimited heredity of this type that it is difficult to resist the conclusion that we have before us the actual mechanism of such heredity — in other words, that some factor essential for sex is associated in the X-element with one that is responsible for the sex-limited character.

This will be made clearer by the accompanying diagram (fig. 8) where the X-element assumed to be responsible for a recessive sex-limited character is underscored (X) . This character may

Fig. 8 Diagram of the distribution of the X- and Y-elements in successive generations, illustrating sex-limited heredity. The underscored X-element (X) is assumed to bear a factor for a recessive character {e.g., white eye-color), while X represents the normal or dominant character {e.g., red eye-color). Y (being the absence of X) likewise represents the recessive character.

Upon pairing the affected male (XY) with the normal female (XX) there are in the Fj generation but two possible combinations, XX and XY. The affected X-chromosome here passes into the female, and the male is normal; but the female of course likewise shows only the normal (dominant) character. In the following F2 generation (5) there are four possible combinations XX, XX, XY and XY, two of each sex. Though X is present in half of each sex, the character appears only in the males, XY, again because of its recessive nature. By crossing together males of the composition XY and females of composition XX, some of the resulting females will have the composition XX , and the sexlimited character is thus made to appear in the female.

When the female is the heterozygous or digametic sex — as in sea-urchins, in Abraxas, the Plymouth Rock fowls, etc. — exactly the converse assumption has to be made. Here, as Spillman ('08) and Castle ('09) have pointed out, the observed results follow if the sex-limited character {e.g., lacticolor color-pattern in Abraxas) be allelomorphic to, or the synaptic mate of, a sexdetermining factor, X, that is present as a single element in the female but absent in the male. The formulas now become'^ (as Spillman has indicated) XG (9 grossulariata), GG (d^ gross.) XG (9 lacticolor) and GG (cf lact). XG X GG then gives in Fi XG and GG (gross. 9 and d^), G having passed from the female to the male. The following cross, XG X GG gives in F2 the four types XG, XG, GG and GG, — i.e., grossulariata appearing in both sexes but lacticolor only in the female. By crossing XG with GG some of the progeny will have the composition GG (cT lacticolor). The other combinations follow as a matter of course.

This interpretation is in all respects the exact converse of that made in the case of Drosophila, which is also the case with

^ These formulas are in substance the same as those stated by Mr. SpiHman in a private letter to the writer, and are a simplified form of those suggested by Castle ('09). The interpretation thus given seems both the simplest and the most satisfactory from the cytological point of view of all those that have been offered. It obviates the cytological difficulties that I urged ('09) against Castle's formulas, and renders unnecessary the secondary couplings that I suggested. All these ways of formulating the matter conform, of course, to the same principle and differ only in details of statement. Whether the synaptic mate of X is directly comparable to the Y-chromosome of other insects (in which case the female formula becomes XY and the male YY) is an open question.

the experimental results, as Morgan has pointed out. It seems probable that all the observed phenomena may be reduced in principle to one or the other of these schemes, though many modifications or complexities of detail probably exist. A possible basis for many such modifications seems to be provided by the cytological facts already known.

(2) We might assume that in cases of the first type {e.g., Drosophila) both sex and the characters associated with it are determined by the same chromatin; and such a possibility should certainly be borne in mind. But in view of the widely different nature of the characters already known to exhibit sex-limited heredity it seems improbable that we can regard them as all alike due to the same chromatin. In the light of the conclusions that have been indicated in regard to the composition of the X-element, it seems more probable that such characters may be determined by the various other forms of chromatin (' Y-chromatin') associated with the X-chromatin. If these constituents be identical with those contained in the free Y-chromosome (the synaptic mate of X) sex-limited heredity would of course not appear, since the two members of the pair would be homozygous in this respect. It should make its appearance as a result of the dropping out, or other modification, of certain Y-constituents of the X-element, and such a mutation might arise in either sex.

We may perceive here the possibility of understanding many different kinds of sex-limited heredity, perhaps of complex types that have not yet been made known. Such a possibility is suggested, for example, by the remarkable relation discovered by McClung ('05) in Mermiria (fig. 6/-/i, fig. 7 in diagram), where the X-chromosome is in the first spermatocyte-division attached at one end to a linear chain of four other elements to form a pentad complex, to which may be given the formula XA . ABB. This so divides as to separate XA from ABB. The interpretation to be placed upon this is a puzzling question under any view, and apparently must await more extended studies on both sexes, perhaps also on other forms, before it can be fully cleared up. Even here the possibility exists, I think, that the entire complex may have arisen by the differentiation of a single original XY-pair; but this question is clearly not yet ready for discussion. However such associations have arisen, the result is equally applicable to the explanation of sex-limited heredity.

(d) Secondary sexual characters. Castle ('09) has offered the interesting suggestion that the free Y-chromosome may be responsible for the determination of secondary sexual characters in the male. Though I have criticized this view ('09c) I now believe it may be true for certain cases. It is obviously excluded when the Y-chromosome is missing; and since nearly related species — in Metapodius even different individuals of the same species — show the same or similar secondary male characters whether this chromosome be present or absent, it seems probable that these characters are in general determined in some other way. But if, as I have suggested, sex-limited heredity may arise through a modification of the Y-constituents of the X-element, it follows that the YY-pair thereby becomes heterozygous. In such case, the free Y-chromosome, being confined to the male line, should continue to represent characters that are no longer present in the female, and hence would be indistinguishable from secondary male characters otherwise determined. It has further become evident (as is indicated below) that the chromosome-groups are so plastic that their specific composition may vary widely from species to species. It may very well be, therefore, that Castle's suggestion may apply to some forms.

6. Modes in which the chromosorne-number may change

The constant and characteristic duality of the 'd-chromosome' in the second division suggests a series of questions regarding the mode in which the chromosome-number may change that have an important bearing on those already considered. The appearance of this chromosome must suggest to any observer that it is a compound body, consisting of two closely united components that are invariably associated in a definite way ; but it is especially noteworthy that its duality does not certainly appear before the last division.^ This case must be added to the steadily increasing evidence that chromosomes which appear single and homogeneous to the eye may nevertheless be compound bodies. An important part of it is derived from the modifications of the Xelement reviewed above; but the evidence is now being extended to the 'autosomes' or ordinary chromosomes as well. The double chromosome of Nezara suggests either the initial stages of a separation of one chromosome into two or the reverse process — in either case that we have before us one way in which the number, and the composition, of the chromosomes may change from species to species. This is supported by the recent observations of Miss E. N. Browne ('10) on Notonecta. In N. undulata there are always, in addition to a typical unequal XY pair, two small chromosomes that appear in all the divisions as separate elements. In N. irrorata there is always but one such chromosome, the total number in each division being accordingly one less than in N. irrorata. N. insulata presents a condition exactly intermediate, there being sometimes one and sometimes two such small chromosomes. When, however, but one seems to be present, the second may often be seen closely adherent to one of the larger chromosomes; and the latter may positively be identified, by its size, as always the same one. It can hardly be doubted, as the author points out, that we here see three stages in a process by which the chromosome-number is changing, either by the fusion of two originally separate chromosomes, or by the separation of one into two. It is of the utmost importance that this process affects a chromosome that can be positively identified as the same in each case; for this proves that the change is not an indefinite fluctation but the expression of a perfectly orderly process. While there is here (as in the case of the d-chromosome of Nezara) no way of knowing in which direction the change is taking place, either alternative involves the conception that the individual chromosomes may be compound bodies, whether as a result of previous fusion or as possible starting points for a subsequent segregation.

The facts now known indicate at least four possible ways in which the chromosome-number (and in three of these also the composition of the individual chromosomes, may change from species to species.

One is that suggested by the foregoing phenomena, i.e., the gradual fusion of separate chromosomes into one or the reverse process.

A second mode may be the gradual reduction and final disappearance of particular chromosome-pairs, as was suggested by Paulmier ('99), and afterwards by Montgomery and myself, in case of the microchromosomes, or ' m-chromosomes' of the coreid Hemiptera. In respect to the size of these chromosomes, a graded series may be traced from forms in which they are very large (as in Protenor) through those where they are of intermediate size down to cases where they are very small (as in Archimerus) and finally to such a condition as that seen in Pachylis (fig. 9 j-l) where they are almost as minute as centrioles and may almost be regarded as vestigial. Four of these stages are shown in fig. 9. In Protenor {a-c) the wi-chromosomes are so nearly of the same size as the next smallest pair that they often can not be positively identified in the spermatogonia! groups. In Leptoglossus phyllopus {d-f) they are always recognizable, though not much smaller than the next pair. In L. oppositus or L. corculus they are a little smaller. In Anasa (the form in which they were first discovered by Paulmier) they are of middle size {g-i) , representing perhaps a fair average of the group. Several other genera {e.g., Metapodius) show intermediate stages between this condition and that seen in Archimerus (figured in my second 'Study,' and more recently by Morrill) where the ??i-chromosomes are almost as small as in Pachylis. It is most remarkable that throughout this whole series the m-chromosomes exhibit essentially the same behavior (Wilson, '056, '06), usually remaining separate throughout the entire growth-period and only conjugating in the final prophases of the first spermatocyte-division, to form a bivalent which with rare exceptions occupies the center of the metaphase group; in some forms, also (e.g., Protenor, Alydus) they show a marked tendency to condense at a much earlier period than the other chromosomes. The m-chromosomes of Pachylis, excessively minute though they are, exhibit a behavior in all respects as constant and characteristic as even the largest of the series. In the Lygaeidae they seem to be present in some

Fig. 9 Comparison of the m-chromosomes in Hemiptera. (In each horizontal row are shown at the left a spermatogonial group, in the middle a polar view of the first spermatocyte-division, at the right a side-view of the same division.) a-c, Protenor belfragei; d-^, Leptoglossus phyllopus; g-i, Anasa tristis; j-l, Pachylis gigas.


species (Oedancala, t. Montgomery), in others absent (Lygaeus). In the Pyrrhocoridae (Pyrrhocoris, Largus) they are absent as far as known. So characteristic is the behavior of these chromosomes as to leave not the least doubt of their essential identity throughout the whole series ; and this series may be regarded as a progressive one, in one direction or the other, with the same reason as in case of any other graded series of morphological characters. The series thus shown in case of the m-chromosomes is as gradual and complete as in case of the Y-chromosome, and may with the same degree of probability be regarded as a descending one. Thirdly, it is probable that the chromosome-number may change by sudden mutations that produce extensive redistributions of the chromatin-substance without involving any commensurate change in its essential content. Were gradual changes, chromosome by chromosome, the usual mode of modification, we should certainly expect to find such conditions as are seen in Nezara, in Notonecta, or in the Coreidae, more frequently. In some groups, however, we find wide differences of chromosomenumber between species even of the same genus, and even between those that are very nearly related, without any accompanying evidence of a gradual process of transition — for instance, among the aphids and phylloxerans (Stevens, Morgan) or in the heteropterous genera Banasa and Thyanta. (Wilson, '09d.) In Banasa dimidiata the diploid number is 16 in both sexes, in the nearly related B. calva 26. Of the two races of Thyanta custator described above, apparently identical in other visible characters, one has in both sexes the diploid number 16, with a simple X-chromosome, while in the other the diploid number of the male is 27 and that of the female 28, and the X-chromosome consists of two components. It is improbable that the differentiation of these two forms has been accomplished by gradual modifications, chromosome by chromosome. It seems far more likely that the change took place by sudden mutation, involving a redistribution of the nuclear material which changed its form but not in an appreciable degree its substance. In the well known case of Oenothera gigas, derived by sudden mutation from Oe. Lamarckiana, a change by sudden mutation is known to be




All the figures from photographs of sections. Enlargement 1500 diameters.

10, 11 First spermatocyte-division (N. hilaris)

12, 13 The same (N. viridula)

14, 15 Second spermatocyte-division (hilaris)

16-25 Side views of second division (hilaris). The XY-pair shown in 16-23, the

d-chromosome in 16, 17, 20, 24, 25; the small chromosome is evident in

10, 12, 13, 14, 15, 17, 18.

22 Initial separation of X and Y

23 Early anaphase, X and Y separating near the center (hilaris)

26-28 Nuclei from the growth-period, showing chromosome-nucleolus and plas mosome (hilaris) 29 Corresponding stage (viridula)

a fact (Lutz, '07; Gates, '08), though it may be due in this instance to a simple doubling of the whole group. Such cases led me several years ago to the conclusion that the nucleus consists of many different materials that segregate in a particular pattern . . . and that the particular form of segregation may readily change from species to species" (Wilson, '09d, p. 2).

Such changes must involve corresponding ones in the morphological and physiological value of the individual chromosomes; and we must accordingly recognize the probability that these individual values, though constant for the species, may change from species to species as readily as does the number. Despite the conformity to a -general type often exhibited by particular genera or even by higher groups, the individual chromosomes are therefore per se of subordinate significance; and it may often be practically impossible to establish exact homologies between those of different species. How closely this bears on the origin of the diverse conditions seen in the composition of the XY-pair is obvious.

Lastly, it is almost certain that changes of number may sometimes arise as a result of abnormalities in the process of karyokinesis, such as the passage of both daughter-chromosomes, or of both members of a bivalent, to one pole. In Metapodius I found ('096) direct evidence of this in case of the XY-pair itself, and endeavored to trace to this initial cause the remarkable variations of number that occur in this genus. Many other observers have recorded anomalies of this kind, in both plants and animals. It seems entirely possible, as has been suggested by McClung ('05) and by Gates ('08) that definite mutations may be traceable to this cause; though probably such abnormalities may in general be expected to lead to pathological conditions.


Some of the suggestions offered in the foregoing discussion are admittedly of a somewhat speculative character; but they are not, as I venture to think, mere a priori constructions, but are forced upon our attention by the observed facts. The time has come



when we must take into account more fully than has yet been done the new complexities and possibilities that have continually been unfolded as we have made better acquaintance with the chromosomes. In this respect the advance of cytology has quite kept pace with that of the experimental study of heredity; and it has established so close and detailed a parallelism between the two orders of phenomena with which these studies are respectively engaged as to compel our closest attention.

Studies on the chromosomes have steadily accumulated evidence that in the distribution of these bodies we see a mechanism that may be competent to explain some of the most complicated of the phenomena that are being brought to light by the study of heredity. New and direct evidence that the chromosomes are in fact concerned with determination has been produced by recent experimental studies, notably by those of Herbst ('09) and Baltzer ('10) on hybrid sea-urchin eggs. But the interest of the chromosomes for the study of heredity is not lessened, as some writers have seemed to imply, if we take the view — it is in one sense almost self-evident — that they are not the exclusive factors of determination. Through their study we may gain an insight into the operation of heredity that is none the less real if the chromosomes be no more than one necessary link in a complicated chain of factors. From any point of view it is indeed remarkable that so complex a series of phenomena as is displayed, for example, in sex-limited heredity can be shown to run parallel to the distribution of definite structural elements, whose combinations and recombinations can in some measure actually be followed with the microscope. Until a better explanation of this parallelism is forthcoming we may be allowed to hold fast to the hypothesis, directly supported by so many other data, that it is due to a direct causal relation between these structural elements and the process of development.

A second point that may be emphasized is the remarkable constancy of the chromosome-relations in the species, and their no less remarkable plasticity in the higher groups. The scepticism that has been expressed in regard to constancy in the species finds, I think, no real justification in the facts. It is perfectly true that


individual fluctuations occasionally are seen in the number of the chromosomes, in the process of synapsis, in the distribution of the daughter-chromosomes, and in all other cytological phenomena. It is, however, also true that most observers who have made prolonged, detailed and comparative studies of any particular group, have sooner or later reached the conviction that in each species all the essential relations in the distribution of the chromosomes conform with wonderful fidelity to the specific type. So true is this that the species "may often at once be identified by an experienced observer from a single chromosome-group at any stage of the maturation-process. No one, I believe, who has engaged for a series of years in the detailed study of such a group, for instance, as the Hemiptera or the Orthoptera, returning again and again to the scrutiny of the same material, can be shaken in the conviction that the distribution of the chromosomes follows a perfectly definite order, even though disturbances of that order now and then occur. But it is equally important to recognize the fact that this order has undergone many definite modifications of detail from species to species, and that while all cases exhibit certain fundamental common features, we cannot without actual observation predict the particular conditions in any given case. It is now evident that the larger groups vary materially in respect to specific conditions. For instance, in the orthopteran family of Acrididae (McClung) the relations seem to be far more uniform than such a group as the Hemiptera, where great specific diversity is exhibited, the details often changing from species to species in a surprising manner — witness the species of Aphis or Phylloxera (Stevens, Morgan), those of AchoUa (Payne) or of Thyanta (Wilson). In these respects, too, the cytologist finds his experience running parallel to that of the experimenter on heredity; and here, once more, we find it difficult not to believe that both are studying, from different sides, essentially the same problem.

December 13, 1910.



Arnold, G. 1908 The nucleolus and microchromosomes in the spermatogenesis of Hydrophilus piceus. Arch. Zellforsch., vol. 2,

Baltzer 1909 Die Chromosomen von Strongylocentrotus lividus und Echinus microtuberculatus. Arch. f. Zellforsch., Bd. 2.

1910 Ueber die Beziehung zwischen dem Chromatin und der Entwicklung und Vererbungsrichtung bei Echinodermenbastarden. Habilitationsschrift, Wurzburg. Engelmann, Leipzig.

Boring 1909 A small chromosome in Ascaris megalocephala. Arch., f. Zellforsch., vol. 4.

BovERi, Th. 1909 " Geschlechtschromosomen" bei Nematoden. Arch. f. Zellforsch., Bd. 4.

Browne, E. N. 1910 The relation between chromosome-number and species in Notonecta. Biol. Bull., vol. 20,1.

Castle, W. E. 1909 A Mendelian view of sex-heredity. Science, n. s., March 5.

Cook, M. H. 1910 Spermatogenesis in Lepidoptera. Proc. Acad. Nat. Sci., Philadelphia, April.

Dbderer, p. 1908 Spermatogenesis in Phyllosamia. Biol. Bull., vol. 13.

Edwards, C. L. 1910 The idiochromosomesin Ascaris megalocephala and Ascaris lumbricoides. Arch. f. Zellforsch., vol. 5.

Gates, R. R. 1908a The chromosomes of Oenothera. Science, n. s., vol. 27, Aug. 2.

1908b A study of reduction in Oenothera rubrinervis. Bot. Gazette, vol. 46,

1909 The behavior of the chromosomes in Oenothera lata x O. gigas. Ibid., vol. 48.

Gross, J. 1904 Die Spermatogenese von Syromastes marginatus. Zool. Jahrb. Anat. u. Ontog., vol. 20.

GuTER, M. 1910 Accessory chromosomes in man. Biol. Bull., vol. 19.

Herbst, C. 1909 Vererbupgsstudien, VI. Die cytologischen Grundlagen der Verschiebung der Vererbungsrichtung nach der miitterlichen Seite. Arch. Entwicklungsm., Bd., 27.

LuTZ, A. M. 1907 A preliminary note on the chromosomes of Oenothera La. marckiana and one of its mutants. Sci., n. s. 26.

McClung, C. E. 1905 The chromosome complex of orthopteran spermatocytes. Biol. Bull., vol. 9.


Montgomery, T. H. 1901 A study of the chromosomes ofMetazoa. Trans. Am. Phil. Soc, vol. 20.

1906 Chromosomes in the spermatogenesis of the Hemiptera Heteroptera. Trans. Am. Phil. Soc, vol. 21.

Morgan, T. H. 190Ga A biological and cytological study of sex-determination in phylloxerans and aphids. Jour. Exp. Zool., vol. 7,

1910 Sex-limited inheritance in Drosophila. Science, n. s. 32, July 22.

Morrill, C. V. 1910 The chromosomes in the oogenesis, fertilization and cleavage of coreid Hemiptera. Biol. Bull., vol. 19.

Patjlmier, F. C. 1899 The spermatogenesis of Anasa tristis. Jour. Morph., vol. 15, Suppl.

Payne, F. 1909 Some new types of chromosome distribution and their relation to sex. Biol. Bull., vol. 16.

1910 The chromosomes of Acholla multispinosa. Biol. Bull., vol. 18.

Randolph, Harriet. 1908 On the spermatogenesis of the earwig, Anisolaba maritima. Biol. Bull., vol. 15.

Sinety, R. de 1901 Recherches sur la biologie et I'anatomie des phasmes. La Cellule, t. 19.

Spillman, W. J. 1908 Spurious allemorphism. Results of some recent investigations. Am. Naturalist, vol. 42.

Stevens, N. M. 1906 Studies in spermatogenesis, II. A comparative study of the heterochromosomes in certain species of Coleoptera, Hemiptera and Lepidoptera, etc. Carnegie Inst. Pub. 36.

1908 A study of the germ-cells of certain Diptera, etc. Jour. Exp. Zool., 5, 3.

1910 The chromosomes in the germ-cells of Culex. Jour. Exp. Zool., vols.

Wallace, L. B. 1909 The spermatogenesis of Agalena nsevia. Biol. Bull., vol. 17.

Wilson, E. B. 1905a Studieson chromosomes, I. The behavior of the idiochromosomes in Hemiptera. Jour. Exp. Zool., vol. 2.

1905b Studies on chromosomes, II. The paired microchromosomes, idiochromosomes, and heterotropic chromosomes in Hemiptera. Jour. Exp. Zool., vol. 2.

1906 Studies on chromosomes, III. The sexual differences of the chromosomes in Hemiptera. Jour. Exp. Zool., vol. 3.

1909a Studies on chromosomes, IV. The accessory chromosome in Syromastes and Pyrrhocoris. Jour. Exp. Zool., vol. 6.


1909b Studies on chromosomes, V. The chromosomes of Metapodius, etc. Jour. Exp. ZooL, vol. 6.

1909c Secondary chromosome-couplings and the sexual relations in Abraxas. Science, n. s. 29, p. 748.

1909d Differences in the chromosome-groups of closely related species and varieties, etc. Proc. Seventh Internat. Zool. Congress, Aug. 1907.

1910a The chromosomes in relation to the determination of sex. Science Progress, no. 16, April.

1910b Studies on chromosomes, VI. A new type of chromosome-combination in Metapodius. . Jour. Exp. Zool., vol. 9.

1910c Note on the chromosomes of Nezara. Science, n. s. 803, May 20.



From Carnegie Institution of Washington: Station for Experimental Evolution

Dr. C. C. Guthrie ('08) has reported the results of transplanting ovaries from black to white hens and vice versa. A blackplumaged hen furnished by transplantation with 'white' eggs and mated to a white cock gave about equal numbers of white and spotted" chicks. Guthrie thinks that these black spots indicate that the black-plumaged foster-mother infected the engrafted 'white' eggs. So far Guthrie. But a person familiar with the results of hybridizing v/ill appreciate that Guthrie's result is better explained on the assumption that the engrafted ovary was absorbed and that the white sperm fertilized the regenerated 'black' eggs of the black hen. For the white by black cross gives white offspring with black spots in the female chicks only, i.e., half of all, as Guthrie found.

In a second set of experiments, Guthrie found that when a white hen carrying a 'black' ovary was mated to a White Leghorn male, the offspring were either white or black or spotted. Guthrie says : "The black, therefore, must have come through the black ovary." But the student of hybridization on poultry will recognize at once that, if the white-plumaged cock produced only ' white ' germ cells, none of his offspring would be black even if the eggs were 'black.' Hence, the cock must have had 'black' germ cells and, very likely, the hen also, since 'White Leghorn' hens that carry 'black' germ cells are very common and frequently show, in adult life, a pure white plumage.

If two 'White Leghorns' with 'black' germ cells be mated expectation is that in four chicks one shall be black; one spotted, and

' A preliminary paper covering these results was read before the Society for Experimental Biology and Medicine, June 1910.



two white; Guthrie got five chicks, one black, one spotted and three white.

Guthrie found that a black hen containing a 'white' ovary, mated with a black cock gave black-plumaged chicks, of which two out of six had white feet. He concludes that the white condition of the feet must have come from the engrafted eggs of the White Leghorn. In criticism it must be pointed out that the cross, white egg X black sperm, normally gives offspring whose plumage color is white, either pure or with black specks. The fact that all the offspring had black plumage proves that the eggs were the normal ' black ' eggs regenerated by the black hen. The white toes are frequently found in the offspring of two black birds. Thus in my pen 1041 two extracted blacks (Sumatras) mated give ten black chicks in six of which white toes are recorded. The results of this cross of Guthrie's confirm the conclusion that the transplanted ovaries were not functional and that the normal ovaries had regenerated.

To -test the possibility of such regeneration of ovaries I removed the ovaries of some hens in the autumn of 1909 and transplanted into them eggs from dissimilar hens. The operated birds were then mated to cocks resembling the soma of the so-called 'fostermother.' Were there regeneration of the ovary the offspring should be of the straight breed; but if the 'grafts' persisted and became functional the chicks should be hybrids.

Experiments 1 and 2, operations: The protocol of the grafting operations is as follows :

No. 11379, pure-bred Dark Brahma bantam, hatched February, 1909; made to fast two days. On September 29, 1909, injected with 0.005 grain of atropin in 1 cc. of water, etherized in about twelve minutes and opened up between two left intercostals. Large ovary, badly torn in removal, removal tolerably complete. One piece of ovary from no. 11605 fastened by cotton thread to inesentery near attachment of ovary. Sewed up.

No. 11605, hatched March, 1909, from White Leghorn,-Houdan ancestry. Clean-footed, with five toes on each foot, V-comb, modified high nostril, plumage color white (with black recessive). On September 29, 1909, injected 0.005 grain atropin in 1 cc. of


water. Etherized in twenty minutes. Plucked feathers and opened body wall between last two ribs. Large ovary completely removed or nearly so, in three or four pieces. Hemorrhage slight. Stitched in small piece of ovary of no. 11379 to peritoneum near attachment of old ovary. Sewed up. Bird recovered rapidly. Some Dark Brahma in ancestry, but its characters had become eliminated.

Results, Experiment 1. Mated in pen 1027, no. 11605 9 (with engrafted ovary from no. 11379, Dark Brahma) and 11291°", straight Dark Brahma. Table 1 gives the juvenile characteristics of 1, the male; 2, the White Leghorn-Houdan, so-called foster-mother; 3, the hen from which the ovaries were transplanted ; 4, expectation on the hypothesis that the graft succeeded; 5, expectation on the hypothesis that the graft failed and the proper ovary was regenerated ; and 6, the observed characteristics of the young offspring.

An examination of table 1 shows at once that it cannot be true that the engrafted ovary replaced the hen's proper ovary, for if it had, columns six and four should agree. On the contrary, column six agrees essentially with column five and supports the hypothesis that the engrafted eggs did not become functional.

One discordant fact there is, however, namely, the occurrence in column six of three cases of cinnamon offspring. Such offspring are to be expected on the hypothesis that some eggs of the graft became functional. If that hypothesis be true, then the other characters of the same individuals should be like those of the pure Dark Brahma. Of the three the first has extra toes, split comb and a boot of one row; it is no Dark Brahma; the second has extra toes, wide nostril and a two rowed boot; it is not a Dark Brahma; and the third has really black down with some red at the tips, five toes on the right foot, a split comb and one row of feathers on the shank; so it is not a Dark Brahma. These therefore, are not from the engrafted Dark Brahma eggs. They represent cases of imperfect dominance of the black down over cinnamon. The conclusion to be drawn from this experiment is that the engrafted eggs did not mature in the foster-mother.

s ^


white; white + specks (or smoked) 21 black, white below ISl cinnamon above, white below 3J

3 rows, 12; 2 rows, 15; 1 row, 1; row, 1; unrecorded, 3

pea, 16;

split, 23; unrecorded, 3

4 toes, 18; 5 toes, 24

narrow, 12; Intermediate and wide, 12; unrecorded, 18



50 per cent white or white + smoky; 50 per cent black back


3 to rows pea, 50 per cent;

split pea, 50 per


4, 50 per cent

5, 50 per cent narrow, 50 per cent Interm. and wide,

50 per cent



cinnamon; light below

heavy pea

4 narrow, gr. 2

light cinnamon; gray below

heavy, 5 to 7 rows pea


narrow, gr. 2


a 1

white (+ smoky) [black back]


Y [V, 1]

5 [4] wide, grade 7


cinnamon; gray below

heavy, 5 to 7 rows pea

4 narrow, grade 2


3 ,

a ^









Experi7nent No. 2. No. 11379, Dark Brahma with engrafted ovary from no. 11605 (White Leghorn-Houdan) was mated in pen 1050 with 14122^, a single-comb Black Minorca. Table 2 gives the juvenile characters of 1, the male parent; 2, the Dark Brahma, so-called foster-mother; 3, the hen from which the ovaries were transplanted; 4, the expectation of offspring on the assumption that the graft succeeded; and 5, that the graft failed and the proper ovary was regenerated; also, 6, the observed characters in the offspring.

Without exception the characters of the offspring are clearly those of the Dark Brahma X Minorca cross and none of the White Leghorn or Houdan differential characters enter into their composition. The grafted ovary produced no eggs that developed, the extirpated ovary was regenerated.

Experiments 3 and 4, operations. The protocol of the grafting operations is as follows:

No. 11541 9 is a white-plumaged hen derived from a cross between 8681 9 ,aWhiteLeghorn-Minorca-Polishbird, and 7811 cf, a Houdan cross hatched (in pen 905) in February, 1909; fasted two days. On October 2, 1909, injected with 0.005 grain atropin in 1 cc. of water; etherized and opened. Ovary very large, two large pieces (60 per cent) of ovary removed. Strong hemorrhage. Two small pieces of ovary from no. 11383 9, Dark Brahma, sewed with peritoneum close to ovarial artery. Sewed up. Bird slow in recovery.

No. 11383 9, straight Dark Brahma, hatched February, 1909, from mating 907: 7549. On October 2, 1909, injected with atropin, etherized and opened, ovaries small, incompletely removed. Two large pieces of ovary of no. 11541 sewed into peritoneum. Sewed up. Bird recovered rapidly.

Results, experiment 3. No. 11541, White Leghorn-Black Minorca-PoHsh-Houdan hybrid, with engrafted ovary from No. 11383, Dark Brahma, was mated in pen 1027 with 11291 o^, a pure bred Dark Brahma. The results of this mating are given in table 3.

Experiment 4- No. 11383 9, pure-bred Dark Brahma with engrafted ovary from no. 11541 (White Leghorn-Black Minorca




1 1












i ^

m \ "*



1 i

1 1" i i


°? . £ S &a 1 i

If 1 ss ss 1

> O


IM f li


i i

is "^ -i 5

F 1 i i:

1 !

ill i:

a I







S £


S 8

s S

I i ■Is

„• t

S « "^

s fe g

g ft

a -^

S. S

a a ftS

I - « 1 1









1 1




- 1


2. "foster-mother"

« i

. > 1 1 ,

1 1 1 1

i 5

ill " I .

1" 1 1 § ^


i i


' .


5 1



Polish — Houdan hybrid) was mated with a Black Minorca 14122 d". The results of this mating are given in table 4.

Experiment 5. No. 11693 9 , used in this experiment, is a white bird that had ' smoke ' on down when hatched. It is of somewhat complex origin. Its mother was an Fi hybrid between a Black Spanish cock and a White Leghorn; its father had the same elements and also white Silkie in its ancestry. No. 11693 has, consequently, black recessive. It has a single comb, is free of the skin pigment of the Silkie, is clean-shanked and has four toes on the right foot and five on the left.

On September 19, 1909, this pullet (which was hatched March, 1910) was treated with atropin, etherized during half an hour and opened as usual between the last two ribs. All of the ovary, as far as could be seen, was removed. Pieces of ovary from no. 11280 ? (a straight-bred Dark Brahma bantam) were placed in contact with the peritoneum, near the removed ovary, but not stitched in, as the bird showed signs of succumbing. The cut was sewed up and the bird set aside where it lay quiet for half an hour.2 The Dark Brahma from which the ovary (whose eggs measured 0.5 mm. in diameter) was removed died in consequences of hemorrhage.

Later No. 11693 was mated with 11291 o^ (in mating 1027: 11693). He is a straight-bred dark Brahma bantam cock, used also in experiments 1 and 3. The results are shown in table 5.

Experiment 6. No. 11826 9, hatched March, 1909, a pure bred Dark Brahma was opened October 2, 1909, and ovary imperfectly removed. Ovary of no. 12550 (a White Leghorn-Minorca-PolishHoudan hybrid) sewed on to peritoneum at point of removal. The ovary had been kept out of body of hen about ten minutes, but covered and moist.

In the late winter of 1910 no. 11826 9 was mated in pen 1050 with 14122 cf, a single-comb Black Minorca. The results are given in table 6.

See postscript.

All black; white below

intermediate, to 3 rows

pea, modified

low, grade 1 or 2


white, 25 per cent; white, black specks, 25 per cent; black and white 50 per cent



Intermediate grade 5 to 2

I 5




white [black and white]



wide, grade 8

. s

cinnamon above; gray below

heavy, 5 to 7 rows


low, grade 1 or 2



black; white below



narrow, grade 1

i i



i 1







In the six experiments described above there is no evidence that the engrafted ovary ever became functional but all results are in accord with the conclusion that the more or less completely extirpated ovary regenerated and produced an abundance of eggs. With the results the data of Dr. Guthrie's paper are not in disaccord. His data, like ours, furnish no evidence for the survival of the engrafted ovaries, far less of an effect of the soma of the foster-mother on the introduced germ plasm.

Cold Spring Harbor, N. Y. September 26, 1910.


On January 4, 1911, No. 11693 ? was killed and opened on the left side. An ovary of fairly typical size for a hen entering her second year of laying was found. It contained numerous eggs, 4 to 5 mm. in diameter. Slightly ventrad of the main artery of the ovary is an irregular mass 5x4x2 mm. of cheesy consistency, imbedded in and covered by peritoneum. Its general appearance is that of a dried, hardened ovary, with clear traces of follicles. It doubtless represents the engrafted ovary, entirely encysted in the peritoneum.

January 30, 1911.


Davenport, C. B. 1906 Inheritance in poultry. Publication no. 52, Carnegie Institution of Washington.

1910 Inheritance of plumage color in poultry. Proc. Soc. Exper. Biol, and Med., vol. 7, p. 168.

Guthrie C. C. 1908 Further results of transplantation of ovaries in chickens. Jour. Exp. Zool., 5, pp. 563-576.





Indiana University, Bloomington, Indiana


Introductory 124

Material and methods 124

Inbreeding and selection on fertility and vigor 126

1 Introductory 126

2 Sterility 127

a Character of the sterility 127

b Degrees of sterility 128

3 Inbreeding and vigor 131

4 Sterility and selection 134

6 Discussion of results 138

Sex-ratio and selection 141

1 Introductory 141

2 The normal sex-ratio 141

3 Control of sex-ratio by selection 142

a History of strain 206 143

b History of strain 207 147

c Discussion 147

4 Influence of male and female in determining the sex-ratio 148

5 Discussion of results on sex-ratio 151

Summary 153

Literature cited 154




The present report includes the results of two series of experiments on the fruit fly — -Drosophila ampelophila. One set concerns itself primarily with the effects of inbreeding and the other with sex-ratios. The experiments on inbreeding grew out of work I had been carrying on on hybridization. In these hybridization experiments the effects on the developmental processes of hybrids between species too remotely related were especially emphasized. The converse of these experiments was, naturally, to study the effect upon the young between individuals too closely related. Fishes, upon which all my experiments in hybridization were made, do not lend themselves for purposes of inbreeding without elaborate breeding facilities. Mice seemed suitable for this purpose but, both at the outset of these experiments and since, these creatures have proven miserable failures in my hands. Among the insects, I tried the common willow beetle but this proved to throw only one generation annually in this latitude. It was desirable to have an animal with a brief life history, whose food could be easily obtained at all seasons and in which the sexes could be readily distinguished. In these respects the fruit fly is almost ideal. The facts herein considered confine themselves to this species.

The experiments on sex-ratio suggested themselves in connection with the inbreeding experiments and so were carried out along with the latter and after they were completed.


The strain which is mostly under discussion in my inbreeding experiments came from a well-filled female that was taken from the window of my residence in Bloomington. Other strains were started at the onset. Some of these came from the banana bunches at the various groceries and others came from fruit which I had laid out for this purpose. None of these were carried further than two or three generations excepting two, called 6 and 7 in my records. The latter was discontinued after the tenth generation


since it had been from the beginning apparently less prolific. The strain 6 was carried for over seventy-five generations and is the one on which the experiments in inbreeding of this report are based.

For vivaria, tall stender dishes, tumblers, quinine bottles and lamp chimneys were given a trial. They were discarded in favor of 8-dram shell vials. These were compact, so that a large number of matings could be kept in a small space, and they were most convenient in manipulating the pairs during the frequent changes to new cages that was necessary all along. The open end of the shell vial was closed with a plug of absorbent cotton, not too compact, so as to afford some ventilation. The flies are strongly positive to light, so that the vials could be laid with their bottom toward the light and the cotton plug removed with safety for the introduction of food etc. Small trays holding fifteen of these vials were used and in this way the experiments could be readily and compactly stored in the incubator, or they could be packed into a valise to be taken along wherever I went. The food was exclusively well-ripened bananas. To prevent the larvae ;rom pupating in the food, narrow strips of blotter or filter paper were introduced in which they seemed to be especially fond of pupating. It is, of course, apparent that the greatest care had to be taken to avoid contamination from flies without. The stock food had to be scrupulously watched and the instruments kept clean to avoid the introduction of eggs laid on them by extraneous females. The bananas, especially, as they come from the stores, are likely to be infected with eggs and larvae if the skin be in any way bruised or split.

The brothers and sisters were paired off, always within the first ten or twelve hours of their life as imagos. Up to this time mating has not occurred. In fact I have never found a pair that copulated during the first twenty-four hours or, if so, that produced fertile eggs.



That continued inbreeding acts deleteriously on the fertility and vitahty of a race is a beUef so firmly and generally established that it is seldom questioned. This has its origin largely in the common experience of breeders whose observations, unfortunately, are too often unreliable. There are not wanting experiments such as those of Van Guaita ('98) and Bos ('94) and others, scientifically conducted, which bear out this conclusion.

On the other hand, it is refreshing to encounter in the literature such reports as that of Gentry ('05) who believes not only that inbreeding is not necessarily harmful, but also that it maybe beneficial to conserve and intensify the good points in his breed. Gentry's experiments were made on Berkshires. The most prolonged tests of close inbreeding that have been recorded were made by Castle ('06) on the same species with which the present paper deals. He inbred (brothers with sisters) for fifty-nine generations. He concludes that such close inbreeding does not necessarily result in a loss of productiveness and of vigor; at least that inbreeding cannot be regarded as a causal factor. Some of his results so nearly parallel those of the present writer that further reference to his results will be made in the body of the paper.

During the early part of October, 1903, a number of pairs were started breeding. These came from various sources in Bloomington. These different pairs were reared for the most part only a few generations, excepting pair No. 6 which was continued for about four and one-half years. During this time over seventy-five generations were produced. Toward the close of this period no exact count was kept of the generations so that only an approximate figure can be given. Five pairs of brothers and sisters were mated in each generation to insure against accidents that might terminate the strain if but one mating were made.

Along at the fifth and sixth generation it became more and more difficult to keep the strain alive with the five pairs of brothers and sisters that were mated each generation. The failure of an


occasional pair to produce young had hitherto been attributed to accidental conditions of food, etc., but this no longer seemed a satisfactory explanation of all the failures to produce young. This condition, was, therefore looked into more thoroughly. This was done by laying out instead of five pairs a much larger number from thf: offspring of a given productive pair. The greatest care was taken with the food, temperature etc. and it soon developed that a variable per cent of the pairs were sterile. These sterile pairs were to all appearances normal. It was clear now that, while inbreeding had not reduced the general vitality of the strain thus far, there had appeared a high degree of sterility.

2. Sterility

  • a. Character of the sterility. Examination of all the matings

brought out the fact that in all cases eggs were present in large numbers. This seemed to suggest that the difficulty lay in the larvae either failing to emerge from the egg envelope or, succeeding in this, failing to carry themselves through the feeding stage or the transformation.

By a careful search of the food of the sterile pairs, after sufficient time for the larvae to mature had been allowed, it became evident that the difficulty lay at a time earlier than the pupal stage for none of the latter could ever be found. The food supplied these sterile pairs was the same as that of the fertile ones since it could not be foretold which pairs were going to prove infertile. Furthermore, the infertile pairs were usually kept for from twenty to thirty days, the best of food being supplied them from time to time. The same search showed that no larvae were present, at least so far as direct inspection of the food under a dissecting microscope could be depended upon.

It was always possible, of course, that the larvae failed to carry their development very far, and, thus, being small when they first emerge from the egg, might have been overlooked. It became necessary, consequently, to take the eggs as they were laid from time to time and keep them under observation to see whether the larvae ever emerged. This was done by placing a piece of banana


in the vial with a sterile pair and from time to time removing the eggs one by one with the point of a needle and placing them on a piece of moist filter paper in a separate vial. Usually twenty were placed in each vial and some food added for the larvae, should they emerge. Inspection of the eggs after twenty-four, forty-eight and seventy-two hours would readily reveal the number of eggs that had produced larvae. I have laid out thus at a great expense of time literally thousands of eggs from many infertile pairs, in many cases all the eggs that a given pair produced during the first twenty-five days of its life, but I have never seen a single egg that had hatched. Eggs of fertile pairs thus laid out will readily hatch so that all the larvae will have taken to the food twenty-four hours after the eggs are deposited.

Such infertile pairs copulate frequently and it would seem that impregnation should follow. I have never sectioned the eggs to see whether spermatozoa enter the eggs or whether they contain partially developed larvae which fail to hatch. I have, however, been able to determine in this strain which of the sexes is at fault. This was done in the following manner. After a pair by sufficient trial had proven itself infertile, the male was mated to a virgin female of a fresh strain that had not been inbred and possessed a high degree of fertility, and the female was similarly mated with a male, usually one whose fertility had been established. Sixty-four such cases were tried and in no case did the females fail to produce young and in no case did the males produce any although repeated copulations took place. It is evident from the foregoing, that, in this strain, the sterility lies exclusively in the male and that the female has lost, apparently, nothing in fertility. Castle (p. 735) reports, on the other hand, that either sex may be sterile. However, Castle took no account of the eggs and larvae but merely the production of pupae, so that his sterility cannot be with certainty compared to mine. It would seem, however, that in some strains infertility may be strictly confined to the males and in others to both sexes. That sterility is complete for all males, when it occurs, is shown by both our results.

h. Degrees of sterility. The foregoing experiments concerned themselves with such pairs as were completely sterile. Other pairs


of brothers and sisters from the same parents, however, were fertile. Judging from the productiveness of these, there was often a wide divergence. It would seem that, as a result of inbreeding, we had a condition of fertility ranging from absolute infertility to comparatively high fertility among the different pairs of brothers and sisters from any given pair of parents. To test this the following experiment was carried out: About two-hundred eggs from each of fifteen pairs of flies were laid out after the fashion indicated above. Ten of these pairs had been inbred for seventeen generations while five belonged to fresh stock that had not been inbred. Of the ten pairs of the inbred strain, five belonged to a strain which had arrived at a very low degree of fertility, namely only 36 per cent of the forty- two pairs tested were fertile (table 3, seventeenth generation, strain, A) . These five pairs were brothers and sisters to many of the sterile pairs considered in the preceding section.

The other five pairs (of the ten inbred) were from a strain which had been held by selection to a high degree of fertility, namely 97 per cent of the thirty-four pairs tested were fertile. Both of these strains were descended from common great grandparents (table 3, seventeenth generation, strain B).

We have, thus, for comparison three conditions, namely, (1) eggs from a highly infertile inbred strain; (2) eggs from a highly fertile inbred strain; and (3) eggs from a presumably norma strain that had not been inbred. It should be added that the five pairs were taken at random and were not selected. Approximately the first two-hundred eggs of each pair were laid out in batches of about twenty to twenty-five to the vial. The number of eggs that hatched was noted in each case and also the number that emerged as imagos. Table 1 gives the summary of results.

From this table it appears that from the eggs which were taken from the inbred pairs with low fertility practically as large a per cent (97.27) hatched as from the eggs that came from the inbred pairs that showed a high fertility (98.2). The same is true in regard to the number that produced imagoes, 86.8 per cent and 85.1 per cent respectively. The fact clearly brought out here is that when infertility arises in this strain it arises suddenly and




does not present all intergradations. In other words, one does not find that among a large number of brothers and sisters some pairs whose eggs only partially hatch and other pairs that range in this respect, on the one hand, to perfect fertility and, on the other, to complete sterility. The fertility is either completely lost or it is of a high degree. Furthermore, when we compare the inbreds with the normals (not inbred) in regard to the percentage of eggs hatched no essential difference is observable. It would seem, therefore,


Inbred {low fertility)





193 184 160

95.3 94.0 98.0 100.0 100.0



200 188

201 197 198 198 123 123

169 182 180 104














Inbred (high



201 173 204

197 175

198 172 200 193 169

182 156 161 165 145

98.5 99.4 98.0 97.9 96.5
















Normals (not inbred)


215 70 153 224 158 146 223

211 70 152 218 155 127 222

193 48 132 144 144 109 205

98.1 100.0 99.9 97.3 98.1 87.7 99.9






86.2 64.2

e". .:













that the pairs that had not completely lost then- fertility, in so far as hatching their eggs is concerned, had suffered no deterioration whatever as a result of seventeen generations of closest inbreeding.

A fact of further importance brought out by table 1 is that of the percentage of eggs that successfully produced imagos. This does not differ essentially in the two groups of inbreds nor do these differ essentially from the normals. Castle used as his measure 'productiveness,' meaning thereby the number of pupae that were successfully produced. Making allowance for some pupae which do not emerge, the imagos produced in my experiments were an approximation to his 'productiveness.' Inbreeding, consequently, does not affect adversely the productiveness of pairs that show any fertility at all.

Castle found that his strains showed an annual fluctuation in productiveness, the period of least productiveness falling in the late autumn and early winter. My own experiments extended over about four and one half years and, although I have been on the lookout for this, I have never observed it. As Castle himself suggests, this fluctuation was probably a function of the temperature of the room. My flies were kept in a room which varied from 60 to 80 degrees and, when this was not possible, they were placed in an incubator kept at about the same range of temperature. It may also be that the productiveness of his strain ran low at this time of the year because they were placed in new hands at the opening of the college year. My observation has been that it takes some time for a new man to learn all the conditions that make for a favorable rearing of these creatures so that Castle's low productive periods may be merely a measure of the training period of the experimentor.

3. Inbreeding and vigor

At the outset of the experiments it was the expectation of the writer that such rigorous inbreeding would early and violently show itself in the vigor and fertility of the animals. In this, however, he was largely disappointed. In the strain that is here under consideration no untoward results could be detected during the


first five or six generations. As previously stated, up to this time the method consisted in placing pairs of brothers and sisters in each of five vials to insure against mishaps. These mishaps consisted of drying up of the food, attacks of fungus and in some cases the escape of the flies themselves during the process of feeding etc. Those pairs that produced young were regarded as having escaped these various possible mishaps and were taken as indications of the vitality and productiveness of the strain. The expectation at that time was that any deleterious effect of the inbreeding would show itself in the offspring of any of the pairs. Consequently, when a given pair would produce offspring that was numerous, all well formed^ vigorous, and in no apparent way differing from normal offspring, to see whether some slight influence might not be present that could not be detected by ordinarj^ observation a definite measure was taken of (1) their rate of reaction to light and gravity, (2) the total number of eggs produced and (3) the percentage of eggs which hatched and emerged. An attempt was made to determine their length of life but this proved too prolonged to allow one to carry it out together with all the other incidents of the already too laborious experiments.

The reaction of this animal toward light and against gravity is well known. To get a measure of the rate of reaction the animals were made to travel through a glass tube that had been blackened for 16 cm. on the inside. This tube had a light placed at one end and was inclined about twenty-five degrees. From a glass vial the flies were admitted, one at a time, into the tube and the time from the moment of entrance into the blackened portion of the tube to their emergence was recorded. It was found essential that the two batches of flies (inbreds and normals) should be of the same age, be reared under the same conditions and that the temperature of the room be the same for the two batches. The results are as foflows: at a temperatureof 27.2° C. 133 normals took 16 seconds, average, to travel the distance, and 140 inbreds took 15.4 seconds. The two sexes in these two groups were about equal in number. In both groups the males travel the distance on an average in three seconds less time. It is clear from this that the normals and inbreeds are equally responsive to these two



agents and that the latter have not suffered in this regard as a result of inbreeding.

In order to determine the total number of eggs produced it was necessary to isolate the pairs and twice each day pick off all the eggs that had been deposited in and around the food provided. This proved to be a most laborious process, for the eggs are too small to be followed safely with the naked eye and had to be removed individually with the point of a needle. Too much value must not be attached to this measure for the reason that the rate and, therefore, probably the number of eggs deposited seems to depend somewhat, at least, on the condition of the food present, and for the


Strain 6

Number of generations inbred 2

Number of days eggs were counted 27

Total number of eggs laid 433







6 34








Strain 7

Number of generations inbred ....


26 654


33 662




6 23



33 907


Number of days eggs were counted

Total number of eggs laid

28 429

reason that only the eggs deposited during the first twenty-five or thirty days were counted. These creatures live to he very much older. We have kept females alive 153 days, but after the first twenty-five or thirty days the eggs come only in small numbers. Table 2 gives the actual counts of several females of both strains 6 and 7.

We see from the above counts that no material reduction has occurred in egg production during nine and ten generations of inbreeding. Such variations as occur may, of course, represent individual differences in the females.

The data given in table 1 of the relative hatching and emerging qualities of the young of normals and of pairs inbred for seventeen generations shows that there is no difference in this respect.


In so far as the above determination may be taken as a measure of the vitality of this species we are justified in concluding that from six to seventeen generations of inbreeding no appreciable deterioration has resulted. No such exact determinations were made in later generations, and it is possible that eventually the effects of inbreeding would manifest themselves, but my observations during seventy-five or more generations does not lead me to believe this.

Jf.. Sterility and selection

Along at the thirteenth and fourteenth generations the sterility had become very pronounced. Of the offspring of some of the pairs, more than 50 per cent of the males were sterile. On the other hand, while practically all pairs showed at least some degree of sterility this varied very much in the different brothers and sisters of the same brood. That this sterility was a direct physiological result of the inbreeding seemed to me very doubtful. To find the effects of inbreeding showing itself in such a specific way upon the males only, did not, to say the least, meet expectations. Furthermore, sterility was not wholly wanting in forms that had not been inbred.

It was highly desirable to continue the experiments on inbreeding, and yet to keep the strain alive, it was necessary to find some way to eliminate this high degree of sterility. The process that was most effective was selection. By continuing the strain of those pairs whose offspring showed the highest degree of fertility but at the same time continuing the rigorous inbreeding, it was possible almost completely to eliminate the sterility. This at the same time gave one of the severest tests as to whether inbreeding was the responsible factor, for if the sterility could be eliminated by continuing the very process of inbreeding the latter could not well be held to be the cause of it.

This was done as follows: In the fourteenth generation three fertile pairs of brothers and sisters from the same brood were isolated and mated. The offspring of each of these were mated in pairs to determine the degree of sterility. By reference to table

3, it will be seen that the pair marked A produced offspring out of which nine of twelve pairs tested were infertile; pair' 5 produced offspring of which four pairs out of fourteen tested were infertile and pair C threw offspring with five pairs out of fifteen infertile. We have here, then, three pairs showing a wide variation in the degree of fertihty of their offspring. Pair A showed 75 per cent of the pairs infertile and pairs B and C approximately the reverse ratio. In the further progress of the experiment pair C was dis

TABLE 4 Strain A






18 (1)

52 51 52 66


27 37 37 46 19

25 14 15 11 9

51 72 71 80 69

49 28 29 20 31

18 (2)

18 (3)

18 (4) . . .

18 (5) . .

Average for 238 pairs 69 per cent.

Strain B

18(1)... 18(2)... 18(3).^.

18(4)... 18(5)...

100 100 100 68 100

Average for 93 pairs 92.5 per cent.

continued so that only pairs A and B were used. I shall in the further description of the experiment refer to the descendants of A as strain A and of B as strain B.

Before entering upon the experiment of selection it was necessary to ascertain whether, without selection, the descendants of pairs A and B continued to show a low and high fertility respectively. Accordingly, a single one of the fertile pairs of the 15th inbred generation of strain A and B was tested. Reference to the table shows that in strain A 27 pairs or 57 per cent of the forty seven pairs tested were infertile, while in strain 5 none of the thirty


seven pairs tested were infertile. The same process was repeated with a pair of the sixteenth generation of the two strains. Strain A showed twenty-seven or sixty-four per cent of the forty-two pairs tested infertile and strain B one or three per cent of the thirty-six pairs tested.

Up to this point in the experiment only a single pair in each generation was tested as to the fertility of its offspring. It might well be that by chance in each case a pair of low fertility was taken in strain A and a pair of high fertility in strain B. To eliminate this possible error five pairs w^ere taken in each strain and the fertility of their offspring determined. It was further desirable to obtain an estimate of the variability in the fertility of the pairs in the two strains as well as to get a more correct estimate of the average fertility of both. In the diagram these five pairs are designated as 18 (1), 18 (2), etc. Table 4 shows the number of pairs of offspring tested for each pair and the number and percentage of pairs fertile and infertile.

The fertility thus varied in strain A from 51 per cent in 18 (1) to 80 per cent in 18 (4), with an average fertility of 69 per cent. In strain B the fertility was much less variable in the different pairs, the only exceptions being 18 (4), the average fertility being 92.5 per cent.

We now have definitely established two strains, one of low and another of high fertility. The important part to be emphasized here is that this was produced by the process of selection from among the variable offspring of generation fourteen of the inbred strain. To make the experiment more complete it was now necessary to obtain a highly fertile strain out of the one with low fertility. Accordingly strain B was discontinued at this point and attention restricted to strain A. Five pairs, 19 (1), 19 (2), 19 (3), etc., were taken from among the offspring of 18 (4) because this showed the highest percentage of fertility. These were tested in the same way as in the preceding generation. Table 5 gives the details.

By selection it will be seen that the average fertility was raised from 69 per cent in the 18th generation to 75 per cent in the 19th generation. Among the five pairs used one 19 (2) showed an unusually high fertility (96 per cent). This pair was accordingly




19(1). 19(2) 19(3) 19(4) 19(5)


Average fertility of 239 pairs 75 per cent.

taken to select from. Five pairs were again taken as before. The results appear in table 6.

Thus it will be seen that all five pairs showed a uniformly high degree of fertility. The average fertility of all the pairs was raised to 93. 8 per cent.

5. Discussion

From the above series of experiments a number of important facts are birought out. 1. Sterility, as it appeared in the strain under consideration, is strongly transmissible through inheritance. 2. It is readily controlled by selection. 3. Inbreeding is probably not the physiological cause of it.

That this sterility is transmissible cannot be doubted. The faithfulness with which this occurs appears in the strains A and B. Both were derived from a common pair that showed a variability with respect to this character in the three pairs of its offspring

20(1). 20 (2) . 20 (3) . 20(4). 20 (5) .

Average for 211 pairs 93.8 per cent.


tested. One of these possessed a high degree of sterility, while the two other pairs showed a low degree. The descendants of the latter constituting strain 5, retained this low degree of infertility throughout. Similarly the descendants of the former, constituting strain A, retained their high degree of infertility up to the time when selection away from this condition was introduced. In the latter process the transmissibility of the character is again emphatically revealed. In the eighteenth generation, pair 4 showed a lower degree of sterility than any of the remaining four pairs of brothers and sisters. Breeding from this pair at once showed offspring with a decided decrease in sterility, compared with the eighteenth generation, the average of the nineteenth generation being 75 per cent of the pairs fertile as compared to 69 per cent of the latter. Again, in the nineteenth generation, pair 19 (2) showed a much lower degree of infertility than the other pairs. Continuing the strain from this pair, this character is faithfully reproduced in the offspring in that they average fertility of the latter is raised to 93.8 per cent.

It is important to note in this connection that Castle, in his experiments upon Drosophila, found that productiveness (which as previously noted is quite a different thing from the sterility here considered) was similarly transmissible and amenable to selection. Furthermore, Castle's experiments would seem to indicate that this character of productiveness behaves, in inheritance, after the Mendelian fashion, low productiveness acting as the recessive character. We have evidently to do here, both in the productiveness in Castle's experiments and in the sterility in my own, with characters that are germinal for they behave as such. In the strain upon which my experiments Vv^ere made we have the further remarkable condition that the infertility is inherited only by the males.

It is clear that whatever the causal factor or factors to which the sterility may be attributed, it is relatively insignificant compared to the effect of selection upon it. Furthermore, the modification is a germinal one. That inbreeding may be responsible for its prevalance in the strain seems probable, but that it is responsible


for its origin is not believed. We have seen that the general vitality of the strain, as measured by its productiveness and its reaction to light and gravity, did not suffer as a result of seventeen generations of closest inbreeding. Failing in this, it is not probable that its effect would show itelf in so specific a way as the sudden and complete sterility in certain males of the strain. The improbability is further supported by the fact that the inbreeding may be continued unabated if only care be exercised in the selection of the brothers and sisters to be mated, thereby even eliminating practically what sterility may have existed.

It is much more probable that the sterility arose spontaneously in this strain or that it is present to a varying degree in this species. With the character present and highly transmissible and subject to selection it is only necessary to carry on indiscriminate breeding to have the character appear in varying intensities depending upon the chance combinations. The rule of inbreeding would be only to intensify the chance combination of the character and to insure the more or less continued presence in the successive generations.

That this character of sterility is not unique to this inbred strain is evident from its rather frequent presence in pairs not inbred. In my own experience this sterility nearly always showed itself in the males. In one instance I found among a brood, besides a sterile male, two females that failed to deposit eggs although eggs were evidently present in the oviducts. Similarly Castle found in his strain a considerable amount of sterility, and this in some cases among the females. We see, therefore, that sterility is not altogether rare even in broods that were not inbred.

The same facts doubtless hold for the character of productiveness. Castle has shown this to be transmissible and amenable to selection. Inbreeding does not produce it but is instrumental, with indiscriminate mating, in intensifying it, or, if the strain be not eliminated thereby, of preserving it in the strain.


The once rather generally accepted notion that nutrition was an influential factor in the control of sex, based on the experiments of Yung ('85), Born ('81), and others, has given place to the now as commonly accepted idea that sex is determined prior to or at the time of fertilization and is independent of the food. The experimental work of Cuenot ('99) King ('07) and others, and the splendid cytological researches of Wilson and his students are largely responsible for this change of view and have been so frequently reviewed in the various recent discussions of the problem of sex that they need not be further detailed here.

The writer tried some starvation experiments on Drosophila in 1904. During the past year more extensive experiments were carried on under his direction by Mr. Claude D. Holmes, on the effects of starvation during successive generations upon the sexratio. These are published under a separate title ('10). It will suffice in this connection, to state that the results coincide with those of recent workers, namely that nutrition does not affect the sex-ratio.

2. The normal sex-ratio

One fact was very apparent in these earlier tests and in all subsequent experiments, that, under the varying conditions in these creatures were reared, there was the same persistance of the predominance of females over males. Below (table 7) is given the






10506 2161 4048






5534 1166 2105 5461

1:1 113


1:1 171

Tomatoes and grapes

Bananas . . . .

1:1.083 1*1 14





1-1 126


summary of four determinations on a large scale to obtain the normal sex-ratio. The flies were reared in the following manner. Mason jars containing a large quantity of food were exposed to flies in nature. The jars were left open until the larvae began to pupate when all flies were excluded by tying a guaze over the top. As the imagos emerged from time to time they were preserved and the sex-ratios determined. For 26933 individuals, the ratio was one male to 1.126 females.

In regard to these determinations only one question, so far as I can see, can be raised. This is the academic one of the greater mortality of the males during development or, to push the matter back a little further and to make it applicable to recent developments in our idea of sex, the greater mortality of the male determining sex cells. In reference to this it may be pointed out that the developmental conditions were as nearly normal as one can imagine. There was an abundance of food, air, light and moisture, and the larvae pupated in the remnants of the food in much the same manner as one finds them doing in nature. In this connection the experiments of Miss King ('07) on the influence of food on the sex ratio of Bufo are of importance. In this she finds that the mortality among the males is not greater than among the females. From these facts and from the knowledge that has come to me from the extensive rearing of Drosophilas for six years I am convinced that the sex-ratio in this species is not one of equality.

3. Control of sex-ratio by selection

If the sex-ratio of this species, then, is that of 1 male to every 1.126 females, this should be regarded as specific just as any other of the specific characters of the species. It should, therefore, be subject to fluctuations and to control like other specific characters.

Starting with this conception of sex-ratio, I wished to see whether it were possible to control this, within limits, of course, by the process of selection. The results of these experiments I propose to detail below.

To apply the selective process on the sex-ratio, the following simple method was employed. Two pairs were selected from


nature, the one showing a high, the other a low female ratio. These were selected as the parents of the two strains to be developed. From among the offspring of each of these two pairs a number of single matings were made. From among these the pair that showed the most favorable ratio in the desired direction was selected to continue the strain. The same process was repeated as often as desired.

From a number of pairs taken from a banana bunch in Bloomington June 12, 1907, two such pairs were obtained. These two pairs go by the numbers 206 and 207, showing the following ratio :

206— 52 cf: 135 9 or 1:2.59 207—84 d^ : 75 9 or 1 : 0.89

A. Strain 206 {high female ratio). The 206 strain will, for convenience, be called the female strain and the 207 strain the male strain, although, as will appear, the latter never developed into a predominantly male strain. In tables 8 and 9 are given in diagramatic form the results of selection for five generations in the former and six generations in the latter. At the margin the generations are numbered 1, 2, 3 etc., and the sex-ratios are indicated.

The sex-ratio of the eleven pairs of brothers and sisters mated from the first generation of the female strain (206) varied from 1 :93 (76 d'; 71 9) to l:-7.00 (8cf:56 9).

The unusually high female ratio in the latter is probably attributable to the small number of individuals obtained from this pair. Two of the pairs threw a predominance of males (table 8 nos. 4 and 8). With the exception of no. 5, all the remaining pairs threw a high female ratio. The ratio for all the pairs was 1:1.67 (578 &: 969 9). We have here a female ratio very much higher than that characteristic of the species (1:1.14) and yet considerably below that of the parent pair (1 :2.59). This may be regarded as a regression toward the normal ratio. It should be pointed out here that too much emphasis should not be placed upon the exact figures representing the ratios in the different pairs, since the number of individuals at best are rather small. In most cases, however, when the number of offspring obtained is fairly large, the ratio approximates the true one, so that in any given







1-1 02


pair from which a fairly large number of offspring has been obtained shows a high female ratio for instance, this may be taken as a pretty safe indication that the female ratio would be high if all or a much larger number had been obtained.

For the next generation ten pairs were taken from brood 9 with a ratio of 1.1.94. Brood 3, with a ratio of 1 :2.31, would have been a more favorable one to select from, but this is not always possible since the matings must be made before all the offspring have emerged and therefore all the data for the complete ratio is obtained. Onh^ four pairs of this series of matings came through safely, due purely to the lack of time to give them the attention they should have had. The four pairs threw the following sexratios: 1: 2.33; 1:2,29; 1:1.27; 1:1.81. The ratio for the entire brood was 1 :1.82 (215 d" : 391 9 ). This ratio was somew^hat more predominantly female.

Pairs were now selected from the brood 8 with a ratio of 1 :2.29. Of the seven pairs mated the offspring of only four was obtained and the number of young in each case was quite small. The ratio for all the offspring of the generation was 1:2.17 (93 cf to 201 ?). The total number here involved is so small that not too much importance should be attached to the increased female ratio.

For the matings of the next generation there is little doubt that an unfortunate selection was made. The brood from which the matings were taken showed a ratio of 1 :2.46 but this ratio was based on numbers so small (52) that it probably did not represent the true ratio of the pair. This may account for the drop in the ratio for all the broods of the 4th generation to 1:1.36 (354 d' 483 9).

Two sets of matings were now made from as many broods of the fourth generation. One of these series was again taken from the brood showing the most favorable female ratio 1:1.90 (85 cf 162 9), but the other series was taken from a brood showing a relatively low female ratio, 1:1.04 (64 & 67 9). From the former the ratio of five pairs was obtained showing a ratio of 1 :1.39 (372 cT — 518 9) and from the latter the ratio of 7 pairs, showing a ratio of 1 :1.07 (496 c^ : 535 9 ).


b. Strain 207 {low female ratio). From pair 207 with a ratio of 1 :0.89 (84 o" : 75 9 ) it was hoped to develop by selection a strain showing a low female ratio. Seven matings from the first generation produced 536 d" and 579 $, or a ratio of 1:1.08. The range of ratios of the individual pairs was from 1 :1.22 (99 d" : 121 9) to 1:0.86 (79 cT: 68 9). This selection was continued for four generations, the matings being made from broods with a low female ratio. The ratios of all the offspring in the successive generations were 1:1.06 (220 cf : 223 9) 1:1.10 (581 c^ : 640 9); 1: 1.04 (142 o^ : 147 9); 1:1.17 (518 d' : 607 9) for the second, third, fourth and fifth generations respectively (See Table 9) . This low female ratio showed itself rather uniformly in all the individual matings, a notable exception occurring in the fifth genertion (see Table 9, pair 3.) with a ratio of 1:2.53 (45 c^r 144 9). On the other hand no pairs threw a great preponderance of males, the most notable among those from which a large number of progeny was obtained being pair 2 in the third generation in which the ratio was 1:0.87 (115 o-: 101 9). For the sixth generation two sets of matings were made as in the fifth generation of the strain 206. One of these was made from a brood with a ratio of 1:2.53 (45 d" : 144 9) and the other from a brood with a relatively low female ratio, 1:1.36 (72 c?: 98 9). From the former the total progeny of eight matings gave a ratio of 1:1.42 (461 d : 654 9 ) and from the latter the ratio of eleven matings was 1.1.05 (944 d^:997 9).

c. Discussion. It seems from the above experiment that the sexratio in this creature is a strongly transmissible character. Starting with a pair that throws an offspring showing either high or a low female ratio it was possible to maintain, by selection, a strain maintaining the respective ratios. The offspring from a given pair, when mated in pairs, show^ a considerable variation in the sex-ratio of their children. It is thus possible to develop a strain with a low female ratio from one with a high female ratio, or the reverse, as is shown in the fifth and sixth generation of experiment 206 and 207 respectively (tables 8 and 9). The sex-ratio is clearly amenable to selection like any other character.


It is an interesting fact that it is possible to develop a strain with a high female ratio much more easily and pronouncedly than a male strain. I have repeatedly tried to hold the sex-ratio to or below that of unity but without success. Not infrequently pairs will throw a predominance of males but it has not been possible to hold them there. The best I have ever been able to do is to hold it considerably below that of the normal, but never as low as unity. On the other hand, it is relatively easy to select in the direction of females even to the extent of 1 to 2.

It should be observed that in the breeding of these strains the most rigorous inbreeding was practiced. It might, therefore, be that the difficulty of selecting for a low female ratio results from the possibility that inbreeding tends toward the elimination of the males. My extensive experience in inbreeding these creatures, however, does not bear out this explanation. Furthermore, in the sixth generation of the high female strain it, was possible in two generations to reduce this ratio to near unity notwithstanding that the same rigorous inbreeding was continued.

4. Relative influence of male and female in determining the sex-ratio

Having thus produced two strains showing a decided difference in the sex-ratio of their offspring I wished to determine two further points. First, whether the maternal or the paternal elements had an equal share in the control of this ratio, and second, whether this ratio was determined in the process of fertilization. To this end reciprocal crosses were made between the two strains and the proportion of the sexes in the offspring ascertained. Three experiments were performed in the following manner. From among a brood of each of the two strains a large number of individuals were taken. Before sexual maturity a number of males and females were isolated, while the remainder were allowed to reproduce. The latter gave a control for each of the strains. The isolated virgin females of one strain were mated with the males of the other. Each experiment thus consisted of four multiple matings. (1) A number of brothers and sisters belonging to the male strain. This furnished a control for the male strain. (2) A number of brothers and sisters belonging to the female strain. This furnished a control


for the female strain. (3) Females from the male strain mated with males from the male strain, and (4) the reciprocal of '(3)'.

In crossing two strains as in the above experiment three possibilities might obtain. First, that the two sexes have an equal influence in determining the sex-ratio; second, that either sex have a predominant influence and third, that a ratio result unlike that obtaining in either of the parental strains. While the first is probably the expected result, the experiments show in a most decided way that the male has little or no influence in determining the sex-ratio in this species (tables 10, 11 and 12). In most of the cases the ratio of the offspring falls pretty closely around that of the strain from which the females were taken. In two instances the ratios exceeded 100 per cent influence. The remaining ones, with the exception of strain 244 in which the male influence amounted to 35 per cent show the female influence almost near enough to 100 per cent to justify one in regarding the differences merely as fluctuations incident to the small number of individuals involved. The unusually great influence of the male in strain 244 might be accounted for in two ways. First the number of individuals involved in this experiment are relatively small so that the ratios of both the control and the crossed broods are not as reliable as in the other experiments. Secondly, the flies used for this experiment were taken from the earlier generations of the two strains, before, we may believe, any considerable selection had been appUed to fix the character of the respective strains. Indeed, this seems to be borne out in the other experiments.

The materials of the three experiments were not all taken from the same generation but were taken from different generations in the development of the strain. Thus, in experiment 1 the broods were taken from the first generation of strain 206 and 207. In experiment 2 the broods came from the second generation of strain 206 and the third of 207. The third experiment was made from the fourth and fifth generations of strains 206 and 207 respectively. Arranging these experiments in a series, based on the length of time that selection had been practiced on the broods used, we see that the male influence decreases as the selective time increases.




Experiment 1

No. of strain mated |

No. 242 2122 X 2122

No. 245 2122 9 X2149d'

No. 243 2149 X 2149

No. 244 214g9 X212CC?




• 9





Number of individuals

Sex-ratio (actual)

208 1.00

194 0.98

463 1.00 1.00

475 1.03


171 1.00

273 1.60

225 1.00 1.00

311 1.38

Theoretical ratio


Influence of male parents.... Influence of female parents..

7 . 3 per cent 92.7 per cent

35 per cent 65 per cent

TABLE 11 Experiment 2

No. of the strains mated. . |

No. 271 252io X 252io

No. 274 252,0 9 X2558cf

No. 272 255s X 2558

No. 273

2558 9 X252iocf









Number of individuals

332 1.00

545 1.69

589 1.00 1.00

919 1.56 1.365

739 1.00

818 1.106

680 1.00

698 1.026

Theoretical sex-ratio



Influence of male parents... . Influence of female parents. .

22 per 78 per

cent cent


per cent (—13) per cent (1 . 13)


Experiment 3

No. of strain mated [

No. 279 2758 X 2758

No. 281 2758 9 X2787cf

No. 280 No. 282

278r X 2787 i 278; 9 X 2758c:f




9 &

9 ' C


Number of individuals

Sex-ratio (actual)

289 1.00

42-, 1.477

382 1.00 1.00

551 1022 1.50 1.00 1.249:


1044 752

1.021 1.00


825 1.083

Theoretical ratio


Influence of male parents.... Influence of female parents..

per cent 13 per cent 100 per cent 87 per cent






Experiment 1-from broods selected for one/ 212 214' 92.7 i 7.3

generation \ 214 212; 65 [ 35

Experiment 2-from broods selected for 2 and 3/ 252 255' 78 22

generations \ 255 252' 100 I

Experiment 3-from broods selected for 4 and 5/ 275 278 100

generations \ 278 275j 87 13

This fact of the prevaihng or exclusive influence of the female in determining the sex-ratio occurs in some other species of animals. Phylloxerans (Morgan '09) and Dinophilus apatris (Korschelt '82) . On the other hand, Whitney ('09) seems to have shown that in rotifers certain eggs which will produce males if unfertilized are changed to females, if impregnated. In the case of Drosophila, we can not be certain that the sex-ratio is established before fertilization since the experiments do not with certainty entirely exclude the male influence.

5. Discussion of sex-ratio

It is not the intention to enter into an elaborate discussion of the problem of sex control. The literature is certainly already sufficiently burdened with such. The writer wishes merely to point out briefly a few conclusions about sex in this species which his results seem to warrant.

The property of sexuality possessed by this species expresses itself not in the equal production, numerically, of its two states, male and female, but in an unequal production. Studies in normal sex-ratios involving a sufficiently large number of individuals are not numerous. The unequal production of the two sexes in the human species is well established. Montgomery ('08) has given the data of a large number of individuals of Theridium and finds a marked inequality in the sexes. The general assumption seems to be that an equal sex-ratio is the rule. It is not improbable,


however, that, as careful determinations upon different species multiply, the condition of unequal ratios will be found increasingly common. Any theory of sex must take into consideration this normal inequality in the sex-ratios.

Sex-ratio like color, size etc., is a character belonging to a species. Sexuality of course is not, for it is common to all species reproducing by the sexual method. The particular form of sexuality, however, the proportion of the two sexual persons to which it gives expression in the process of differentiation, this is specific. For Drosophila ampelophila, the ratio of one male to 1.126 females is a specific character. This is not a ratio of merely the present generations but has been transmitted from generations remote. It is inherited. It is the expression of the physiological condition to which the species has been developed by its environmental demands.

Like other specific characters this ratio should be subject to modification, but this should not be more easily done or by other methods, in general, than those used in the modification of other characters. From this view point it should not be expected that the sex-ratio in an animal could be materially changed by such agents as food, temperature,etc. A change in the proportion of the sexes involves a much more fundamental modification than simple starvation or the reverse is likely to induce. In regard to other characters, we have long ago ceased to regard them as modifiable by such methods, but in the case of sex, it is only recently that their futility is being entertained. The most potent factor and the one most generally used to modify a character is selection. If the experiments herein recorded prove what they are held to prove, this process of selection is a potent factor in the modification of the sex-ratio also. It would be interesting to try to line this fact up with the chromosomal conception of sex. However, the writer regards this as the task of those who are engaged in these interesting and important investigations.


1. Drosophila ampelophila may be inbred (brothers and sisters) for seventy-five or more generations.

2. Inbreeding in itself is not deleterious to the fertility or vigor of this species.

3. ' Infertility normally occurs to a varying degree among the offspring of any pair. Promiscuous inbreeding among such offspring may perpetuate and even intensify this character. When sterility appeared in the strain experimented with, it was always complete, appeared suddenly and was confined to the male.

4. By the judicious selection of the brothers and sisters to be mated from a brood that shows a high degree of infertility, this infertility can be eliminated by selection although continuing the inbreeding in the closest possible way.

5. There is a wide divergence in the fertility and productiveness among the different pairs taken in nature, but by the proper selection and closest inbreeding these may be readily brought to either a high or low state with respect to these characters.

6. Many generations of closest inbreeding does not necessarily cause any loss in size, perfection of form, rate of reaction to light and gravity, egg production or length of life and sex-ratio.

7. The normal sex-ratio of this species in nature when reared under diverse conditions of food is one male to 1.126 females.

8. Different pairs in nature show a wide divergence in the sexratio of their offspring.

9. When the offspring from a pair with a given ratio are mated in pairs their offspring will show a wide range in the sex-ratio but in the aggregate will tend to reproduce the ratio of the brood to which they belong.

10. Sex-ratio is therefore a character that is strongly transmissible. By the proper selection of pairs tending to throw a high female ratio on the one hand or a low female ratio on the other it is possible to develop strains characterized by high or low female ratios.

11. In this species it is comparatively easy to develop a strain with a female ratio considerably higher than the normal but very



difficult to develop a strain with a female ratio much lower than the normal or even one in which the sexes are equal in number.

12. Sex-ratio is one of the qualities that is, like color, an inherent characteristic of this creature, strongly transmissible and amenable to the process of selection.

13. The female is almost wholly responsible in the transmission of the sex-ratio. For, if females from a strain possessing a high female ratio be mated with males from a strain possessing a low female ratio or vice versa, the offspring will show a sex-ratio which is wholly or very near that of the strains from which the females were taken.

14. Sex is probably very little, if at all, influenced at fertilization in this species, but is probably determined much earlier and by the female, but there seems no reason why this may not be influenced by various factors and in some species at fertilization.


Born, G. 1881 Experimentelle Untersuchungen iiber die Entstehung der Geschlechtsunterschiede. Breslauer iirtzliche Zeitschrift. Bd. 3.

Bos, J. R. 1894 Untersuchungen fiber die Folgen der Zucht in engster Blutsverwandtschaft. Biol. Centralbl. pi. Bd. 14.

Castle, W. E. and others. 1906 The effects of inbreeding, cross-breeding and selecLion upon the fertility and variability of Drosophila. Proc. Amer. Acad. Arts and Sciences, vol. 41 .

CuENOT, L. 1899 Sur la determination du sexe chez les animaux. Bull, scientif. de la France et de la Belgique, t. 32.

Gentry, N. W. 1905 Inbreeding Berkshires. Proc. Amer. Breeders Association vol. 1.

Guaita, G. VON 1898 Versuche mit Kreuzungen von verschiedenen Rassen des Hausmaus. Ber, ub. d. Verhandl. d. Naturforsch. Gesellsch. zu Freiburg, Bd. 10.

Holmes, Claude D. 1910 The effect of starvation for five successive generations on the sex-ratio in Drosophila ampeloph'.la. Indiana University Studies No. 2.

King, Helen D. 1907 Food as a factor in the determination of se> in Amphibians. Biol. Bull., vol. 13.

KoRSCHELT, E. 1882 tJber Bau und Entwickelung des Dinophilus apatris. Zeitschrift f. wiss. Zool. Bd. 37.

Morgan, T. H. 1909 A biological and cytological study of sex determination in Phylloxerans and Aphids. Jour. Exp. Zool. vol. 7.

Whitney, D. D. 1909. Observations on the maturation stages of the parthenogenetic and sexual eggs of Hydatina senta. Jour. Exp. Zool. vol. 6.

Yung, E. 1885 De 1' influence des \ariations du millieu physico-chimique sur le development des animaux. Arch, des Sci. phys. et naturelles, t. 14.




The snail's foot in locomotion is so striking and so easily observed that it has excited the interest of naturalists for a long time and yet a complete solution of even the mechanical problems connected with its action seems not to have been attained. Within recent times a number of investigators have attacked the problem of locomotion in snails, but their efforts have been directed chiefly toward the elucidation of the action of the neuromuscular mechanism rather than toward an understanding of the external mechanical conditions that accompany locomotion. It is the object of this paper to consider, in the light of the more recent investigations and from the standpoint of renewed observation, the external mechanical factors involved in the movements of the gastropod foot.

The observations recorded in this paper were made partl}^ at the Bermuda Biological Laboratory, at the Harvard Zoological Laboratory, and at the Biological Laboratory of the United States Bureau of Fisheries at Woods Hole. I am under obligations to the directors of the laboratories mentioned for the materials and opportunities for carrying on these studies.


When the locomotor movements of the foot in many species of gastropods are compared, a surprising diversity is found. These different types of movement have been well classified by Vies


156 G. H. PARKER

('07) and are apparently characteristic not only for species but for larger groups of gastropods. In the majority of species thus far examined, the pedal waves course forward over the foot, thus agreeing in direction with the animal's locomotion. Vies has appropriately designated this type of movement as the direct type and has given the following gastropods as examples; the pulmonates (including Onchidium), Aplysia, Aeolis, Doris, Haliotis, Trochus, Cyclostoma, and certain small species of Littorina. I can confirm this statement for such of these molluscs as I have examined, namely, many pulmonates, including Onchidium, and I can add to this list Crepidula fornicata. In other gastropods the waves pass over the foot from anterior to posterior and this type has been designated by Vies as retrograde. As examples he has given Acanthochites fascicularis, Littorina littorea, and L. rudis. Besides confirming Vies' observation on Littorina littorea, I can add to this list Dolabrifera virens Verrill, Tectarius nodulosus Gmel., Nerita tessellata Gmel., and Chiton tuberculatus Linn. According to the observations of Jordan ('01, p. 99) Aplysia belongs under this head and not under that of the direct type as given by Vies.

In both chief types of movement several subtypes can be distinguished as determined by the lateral extent of the pedal waves. In some gastropods each wave extends over the functional width of the foot and thus the foot is occupied by only a single series of waves. This subtype has been termed by Vies monotaxic, and is exemplified by the pulmonates and chitons. In addition to these gastropods, Dolabrifera virens also has a monotaxic wave. In other gastropods the foot is functionally or even structurally divided along the median plane and exhibits a double system of waves, one right and the other left. This subtype has been designated ditaxic by Vies and is exemplified by Haliotis, Trochus, and Cyclostoma among the direct types, and by Littorina littorea among the retrograde types. Besides confirming Vies' statement as to Littorina littorea, I can add Tectarius nodulosus and Nerita tessellata as ditaxic gastropods. In Tectarius the waves on the two sides of the foot usually alternate and they are


SO extensive that never more than two waves can be seen on one side of the foot at once. The foot, therefore, moves forward in alternate steps, first on the right side and then on the left, the motion resembUng that of a person in a sack walk. In Nerita the wave begins anteriorly as a single wave whereupon it breaks and passes down the right and left sides of the foot to unite as one wave again at the posterior margin. These two conditions of alternate waves, as in Tectarius, and opposite waves, as in Nerita. will probably be found exemplified in other ditaxic gastropods. In certain small species of Littorina with direct movements, Vies has described four parallel sets of waves, fulfilling the requirements of a tetrataxic subtype. This occurs, according to Vies, only in connection with the retrograde type of movement. I have seen no example of it.

Among those snails that I have examined, one species, Ilyanassa obsoleta (Say), seems to find no place in Vies' classification. This snail is a vigorous, active creeper. Its foot covers a large area compared with the size of its body. Anteriorly the foot is truncated and auriculate; posteriorly it is bluntly rounded. Its ventral surface is whitish, flecked over with irregular grayish splotches. In resting, the snail uses chiefly the posterior part of the foot, the anterior part being sometimes more or less withdrawn into the shell. In locomotion the anterior part seems to be the more active. Notwithstanding the fact that this snail is very easily observed in active creeping and that its foot is marked in a most favorable way for exhibiting wave-like movements, I have never been able to discover any evidence of such movements. When in locoixiotion, the whole foot seems to glide at a uniform rate over the surface of attachment such as that of a glass plate. Only along the anterior edge and over a small portion of the anterior ventral surface of the foot, can slight variations in the rate of movement be discovered and these variations are so local and scattered that they can in no sense be regarded as forming a wave. The movement of the foot of Ilyanassa has a most striking resemblance to that of the foot of a planarian in which cilia may be the chief motor organs, but on testing the foot

158 G. H. PARKER

of Ilyanassa with carmine suspended in seawater. not the least evidence of cilia could be discovered. I therefore believe that Ilyanassa moves by a form of muscular activity that does not appear as pedal waves and it is not improbable that other gastropods will be found that have the same peculiarity. That Vies recognized something of this kind may be inferred from his statement that no changes in color can be seen in the creeping foot of Na-ssa, Buccinum, Aeolis, etc., and that the direction of the waves in these instances can be judged only by the deformations produced at the edge of the foot. As Vies makes no further mention of Nassa in his subsequent account, I suspect that it is more or less like Ilyanassa and is capable of little or no pedal-wave movement. The locomotion of such gastropods I should designate as due to arhythmic pedal movements as contrasted with rhythmic pedal movements, such as have been fully classified by Vies.

It is a significant fact that all gastropods, irrespective of their type of movement (direct or retrograde), are restricted to forward locomotion. None, so far as I am aware, can reverse and move backward as, for instance, an earthworm can. Whatever differences these various types of pedal movements possess, they still lead to but one result, the forward locomotion of the snail.


The snail's foot subserves the double function of attachment and locomotion. As means of attachment snails secrete a bed of mucus, and use the foot as a sucker. Both methods are commonly employed by the same species, but in a given form one method is usually developed much in excess of the other. For instance, in Helix pomatia, Limax maximus, and other allied species, the moist surface of the expanded foot will stick with some tenacity to glass. But if such an animal be allowed to creep its length over a glass surface and thus spread a bed of mucus on which it can rest, it will be found to have multiplied the strength of its attachment many times. The mucus adheres to the glass and the surface of the foot to the mucus very much more power


fully than the foot alone can adhere to the glass. That this attachment is due chiefly to the adhesive properties of the mucus and not to the sucking action of the foot, is seen from the fact that the attachment can be completely accomplished over a minute hole in a plate of glass. When a snail in such a position is seized and drawn off, air is sucked in through the hole in the glass as the middle of the foot rises, showing that under these extreme circumstances, the foot does act as a sucker, but in the ordinary resting state of the snail no such suction is exerted. All snails with which I am acquainted deposit more or less mucus and though this is sometimes so small in amount that it can be demonstrated only by means of powdered carmine, it serves, I believe, in so far as it is present, as a means of attachment. This production of mucus is highly developed in the pulmonates. Its relation to creeping on the surface-film of water, as exhibited by many fresh-water snails, has long been recognized.

In some snails the foot serves as an organ of attachment chiefly through its power of suction. The general surface of the foot is applied closely to the substrate after which the central portion is lifted thus converting the foot into a sucker. This kind of attachment is well exemplified in Patella, Crepidula, etc. Crepidula fornicata can be made to creep over a surface of glass and can move with ease and security over a minute hole in the glass. If, however, the snail is disturbed by being touched several times when its foot is over the hole, it will actually dislodge itself by endeavoring to suck firmly to the glass, for in so doing it will fill to repletion the forming concavity on the underside of its foot by sucking water through the underlying hole. When one contrasts the difficulty with which Crepidula is dislodged from its natural surface of attachment, particularly after it has been induced to exert full suction, with the ease with which it can be made to dislodge itself when over a small hole, the magnitude of its power of suction becomes apparent. The action of the foot of Aplysia as a suction apparatus has already been demonstrated by Jordan ('01). These two methods -of attachment, suction, and adhesion through mucus are the chief means by which snails hold to the surfaces on which they creep.

160 . G. H. PARKER


Locomotion by the gastropod foot, is not dependent upon ciliary action but is a muscular operation as shown by Dubois and Vies ('07) . The precise way in which the movements of locomotion are accomplished can best be made out by examining good examples of direct and retrograde movement. The first is well exemplified in Helix pomatia and Limax maximus; the second in Chiton tuberculatus and Dolabrifera virens.

In an expanded and actively creeping Helix pomatia, the foot may measure as much as seven to eight centimeters in length by two and a half in width. Over this a succession of transverse, dark-brownish waves run from posterior to anterior. At any instant there may be as many as ten or a dozen such waves on the foot. Each wave is separated from its neighbor by a space equal to about three-times its own thickness. The waves travel over the foot in about thirty seconds, or at a rate of a centimeter in seven to eight seconds. These records, taken from a normal individual, agree fairly well with those given by Bohn ('02) and by Biedermann ('05).

As the snail creeps, it spreads from the mucous gland at the anterior edge of its foot a broad path of slime over which it makes its way. An active snail marks its course in this manner by a long track of slime. A somewhat exhausted snail, when placed upon an appropriate substrate, will almost always creep far enough to lay a mucous path that will subtend the whole of its foot, after which it will cease creeping. If it is removed to another position, it will usually repeat this operation, but it will seldom creep farther. This habit is doubtless connected with the effectual attachment of its foot to the substrate.

Locomotion in Helix, like that in other pulmonates (Kiinkel, '03), is apparently inseparable from the wave movement of its foot. When a snail is placed upon a glass plate preparatory to creeping, it lengthens and expands its foot; almost immediately thereafter pedal waves appear and the animal begins to move forward. Such a snail will creep over a perforation in a glass plate


without isuc'king air through the perforation, thus demonstrating that its attachment in locomotion/ as in rest, is due to adhesion and not to suction. In fact in a creeping Helix the foot not only does not suck but actually presses on the substrate. If, as the snail creeps, a bubble of air is introduced under it by a capillary tube or other means, this air will usually escape at the edge of the foot in such a way as to show that it was under considerable pressure. The action of such bubbles demonstrates that the foot as a whole is firmly attached to the mucous substrate, in fact presses against it. Locomotion in Helix pomatia, then, has to overcome under ordinary circumstances only the adhesion of the foot and this is accomplished apparently by the pedal waves. In snails in which the attachment is due to suction as well as to adhesion, locomotion requires that both attractive forces shall have been overcome, but, as suction is muscular, it seems likely that this would be relaxed somewhat, as seems to be the case in Crepidula, before locomotion begins.

How the pedal waves accomplish locomotion is still a disputed question. According to von Uexkiill ('09, p. 181), who has followed Jordan ('01) and Biedermann ('05) in many particulars, each pedal wave is formed by the contraction of the longitudinal muscles of the foot and takes the form of a slight swelling on the underside of the organ. Such a wave, as von Uexkiill rightly remarks, would effect nothing by way of locomotion unless some portion of the foot were fixed. Von Uexkiill ('09, p. 187) believes that the foot is provided with some such mechanical device as the setae of the earthworm, which, resist backward movement while they allow forward motion and that, therefore, the region in front of each wave may be regarded as a fixed region. Hence the contraction waves would always draw that portion of the foot where they temporarily were forward over the substrate toward the fixed point in front and as a result forward locomotion would be accomplished.

Although this explanation is free from mechanical objections, it is doubtful whether it really applies to the case in hand. Von Uexkiill has maintained in support of this view, that a snail can

162 G. H. PARKER

be slipped over a glass plate more easily forward than backward, just as an earthworm can be drawn over an appropriate surface more easily headward than tailward. I must confess that I have not been able to convince myself that there is any difference in this respect in Helix pomatia or Limax maximus; both seem to slip over the glass forward and backward with equal ease.

Moreover, the view advanced by von Uexklill is based upon what I believe to be a somewhat erroneous conception of the pedal wave. Biedermann ('05, p. 11) pointed out that the foot of Helix pomatia has great advantages over that of many other gastropods for studies of this kind because of the numerous small specks contained in its outer layer. These specks can be discerned clearly by means of a hand lens and they give a true picture of the movements of the foot. As watched through a plate of glass over which the anuTial is creeping, they can be seen, as Biedermann has described, to move momentarily forward, then come to rest, and then again to move forward. This is best demonstrated on a sheet of glass on which there are numerous scratches. Such scratches serve as landmarks and by them it can be seen that the minute specks in the foot do remain essentially fixed in position and then momentarily move forward to assume again for a brief period a position of rest. When this motion is examined in relation to the foot as a whole, it is evident that the forward motion takes place in the dark waves and that quiescence is characteristic of the intermediate lighter portions of the foot. Each wave, then, is a pulse of forward motion and the rest of the foot is momentarily quiescent. The area covered by the waves is probably a fourth or a fifth of the total area of the foot. At any moment, therefore, about three-fourths to four-fifths of the surface of the foot is stationary and about one-fourth to one-fifth is moving forward. In other words the snail stands on the greater part of its foot while it moves forward with a much lesser part.

Essentially the same conditions as have been described for Helix pomatia can be demonstrated in Limax maximus. If particles of carmine be driven into the substance of the median, active band of the foot of this slug, they can be seen to exhibit exactly


the same type of movement as has been described for the specks in the foot of HeUx. In Limax the waves, however, are light in color, instead of being dark as in Helix, and their surfaces, .as seen in the air, are marked with fine wrinkles transverse to the longitudinal axis of the animal. These wi'inkles show that the waves are regions of longidudinal contraction, as has been maintained by most recent writers on this subject.

The chief error in most previous accounts of the locomotion of the gastropod foot is found in the physical configuration ascribed to the underside of this organ. Biedermann ('05, pp. 10, 17) states that the waves are convexities on the surface of the foot and that they press more firmly against the substrate than does the rest of the foot. This view was adopted by von Uexkiill ('09, p. 187) in his discussion of gastropod locomotion. In Helix pomatia it is by no means easy to determine whether the waves are convexities or not, for the reason that they are at most only very slightly different in level from the general surface of the foot. On inspecting by reflected light the free ventral surface of a part of a Helix foot over which waves were running, I was unable to tell with certainty whether the surfaces of the waves were convex, concave, or flat. If, however, the creeping foot be closely studied through glass, evidence of a conclusive kind can be found. If, under these circumstances, a very minute air bubble entangled in the mucus under the snail is watched, it will be seen to change its form and position slightly as each wave passes over it. As the wave approaches it, it will elongate slightly on its face next the wave and at times move a little towards the wave, and as the wave leaves it, it will elongate slightly in the opposite direction and at times follow slightly the retreating wave. The motions of the bubble are exactly those that should be expected provided the wave exerted a slight suction in its passage and the reverse of what would occur supposing the wave pressed upon the bubble. The evidence, though slight, is clear and I, therefore, believe that each wave on the underside of the foot of Helix pomatia is a slight concavity.

Although the configuration of the surface of the wave in Helix pomatia could be determined only indirectly, in Limax maximus

164 G. H. PARKER

it can be seen with distinctness. If the anterior part of the foot of this slug be appHed to a glass surface, the pedal waves appear quickly over the whole foot. On inspecting the portion of the foot not yet in contact with the glass, the waves can be identified as dark bands alternating with light areas. On examining from the side the portion of the foot not yet in contact with the glass, it can be clearly seen that the waves are concavities in the foot as compared with the areas between the waves. I am, therefore, entirely convinced that, contrary to the opinion expressed by Biedermann and others, the pedal waves of the gastropods are concavities and not convexities on the foot. In these concavities, which are probably filled with the more fluid portion of the mucus, the foot moves forward, the rest of this organ being temporarily at a standstill.

The mechanical advantage of this arrangement must be obvious. The snail is attached to the substrate chiefly by adhesion to the denser mucus. This attractive force is overcome by drawing certain parts of the foot, the region of the waves, away from the substrate. These parts are then in a position to move with reduced resistance and are momentarily shifted forward while the snail supports itself on the rest of its foot. As this release from adhesion is propagated as a wave over the whole of the foot, this whole organ, together with the rest of the snail, is eventually moved forward. At first thought it might seem that such a wave movement could not produce so uniform a motion as snails show, but it must be remembered that the uniformity of this movement is seen only in parts of the animal some distance from the foot. On the foot itself the operation is alternate movement an4 rest, which becomes more and more continuous motion as points on the body more and more distant from the foot are reached. The locomotion is in many fundamental respects like that of the human being. In our locomotion each foot is alternately at rest and in motion and yet distant parts of our body, like the head, show a motion which in comparison with that of our feet is almost continuously uniform. In fact, a ditaxic gastropod with alternate, direct, single waves on the foot would almost exactly reproduce the method of locomotion found in the human being.


Tectarius, as already noted, practicalh^ fulfills these conditions except that its waves are retrograde. This general theory of the mechanics of gastropod locomotion is an elaboration of the views already set forth by Jordan ('01).

It is not my purpose in this paper to enter into an account of the musculature by which the movements already described are carried out, for I have made no observations on this part of the subject. It is, however, pertinent to show that the elements of motion implied in the preceding description are not inconsistent with the general structure of the snail's foot. The work of Jordan ('01), Biedermann ('05), and others shows conclusively, I believe, that the musculature of the snail's foot works against the elastic-walled, fluid-filled cavities of the animal's interior and that these cavities are often temporarily closed from one another. It is these spaces which, acting collectively as a vacuolated, erectile tissue, give rise to such rigidity as is possessed by the expanded foot of the snail. In this tissue two sets of muscles, longitudinal and dorso-ventral, have been identified. The dorso-ventral muscles lift the foot locally from the substrate. They are imbedded in the vacuolated tissue already mentioned and when they contract, their dorsal ends, being more firmly set than their ventral ones, serve as relatively fixed points and the ventral ones, therefore, move. The mechanical support that these muscles receive comes primarily from the tissue adjacent to their dorsal ends which in turn gets its support from other tissues reaching to the parts of the foot fixed on the substrate in front and behind the region of elevation. The "action of the ventral end lifts the foot locally and overcomes adhesion in the given region. When the muscle relaxes, the portion of the foot that was elevated is returned to its former level chiefly by the elastic action of the vacuolated tissue and the muscle recovers its original length and position. This action of the dorso-ventral muscles takes place in sequence from behind forward and thus a concave wave runs on the surface of the foot from tail to head.

The second element in the pedal wave is the forward movement of that portion of the foot which is temporarily lifted from the substrate. This must be accomplished by the contraction of the

166 G. H. PARKER

longitudinal muscles and can be best pictured by reference to the accompanying diagrams. These diagrams represent steps in the passage of a concave wave over the foot of a snail from an anterior position to a posterior one (left to right in the diagram) whereby the pointlc is temporarily released from full adhesion to the mucous surface, moved forward, and brought to full adhesion again. The point x is supposed to be associated with a particular longitudinal muscle fiber, number 2, through whose action it is moved. In A, this fiber is shown in its relaxed condition with the wave approaching. In B, the wave has released the point x from full

I 2 3

A ' ' — '

B i 1

C ' i

D ' '

K ' ' '—~

?-^- ^-^-^ ' ^<^

adhesion. In C, fiber 2 has contracted and since the posterior end of it is over a released part of the foot and the anterior end over a fixed part, the posterior end with the underlying point x has been moved anteriorly. In D, the fiber remains contracted and the point x has come again to adhere to the substrate. In E, the wave has reached the next longitudinal fibre anterior, number 3, which has contracted and drawn out the relaxing fiber, number 2, to its original length and position in reference to point x. The contraction of each longitudinal fibre then serves two purposes: it moves the foot forward as the releasing wave passes over the region and it extends the relaxing posterior fiber. In this way each


point on the foot is lifted, moved forward, and set down again and thus the foot, and with it the animal as a whole moves forward. From this theoretic consideration, it is evident that the theory of pedal-wave action advanced in the preceding paragraphs is entirely consistent with such an arrangement of muscles as has long been known to occur in the gastropod foot.

Vies ('07) has called attention to the fact that the majority of theories as to the locomotor action of the gastropod foot apply only to the direct type of movement and do not take into account the retrograde type. The theory put forward in this paper is believed to apply equally well to both types. Among retrograde gastropods. Chiton tuberculatus is an excellent example. This mollusc uses its foot as a sucker, but nevertheless can creep with considerable rapidity. It exhibits, as a rule, not more than two waves on the foot at a time; these course posteriorly at the rate of about a centimeter in five seconds. In a Chiton creeping over a glass plate, the wave when viewed from the side can be seen to be an area lifted well off the substrate. This feature is much more conspicuous in Chiton than in any other mollusc that I have examined. As in the pulmonates, the surface of the Chiton foot in direct contact with the substrate is motionless ; that in the wave area moves forward. At any moment about one quarter of the Chiton foot is moving forward while the animal supports itself on the remaining three quarters.

In Dolabrifera the foot is pear-shaped in outline with the rounded end posterior. It is about 8 mm. in length. In creeping, one to two waves can be seen on its surface at once; each wave sweeps the length of the foot in about seven seconds. As in Chiton, the waves can be clearly seen to be areas in which the foot is lifted completely from the substrate to which the rest of the foot is firmly applied. The pedal surface is mottled and in the wave area it can be seen to be moving forward, whereas on the rest of the foot it is motionless. The total wave area is about one-half the total area of the foot.

The conditions in Chiton and in Dolabrifera are essentially similar to those in the pulmonates, except that the pedal waves progress posteriorly instead of anteriorly, i.e., the dorso-ventral

168 G. H. PARKER

muscles contract in sequence from the anterior to the posterior end instead of the reverse and the longitudinal muscles follow the same sequence; otherwise they act as they do in the direct type. It is evident from this brief discussion of the nature of the waves in the retrograde type that the theory developed in connection with the direct type applies perfectly to this second type.

It remains still to point out that what I have called the arhythmic form of pedal locomotion, a form well exemplified in Ilyanassa, may be explained on the same general basis as that which has just been given for the two types of arhyhmic locomotion. If the foot of such a snail as Ilyanassa be thought of as composed of a multitude of small areas, each one of which can be lifted from the substrate, moved forward, and set down again separately, and that this action takes place irregularly and without reference to any sequence, it can easily be seen how the animal could move forward but without the formation of pedal waves. It is my belief that this is the condition in the foot of the arhythmic gastropods, but because of the small size of Ilyanassa, I have not been able to subject this opinion to experimental test.

Before closing this paper, I wish to add a word concerning the very remarkable method of locomotion observed by Carlson ('05) in Helix dupetithouarsi. The movement carried out by this snail is appropriately described as a gallop, both from its rate and configuration. The snail on strong provocation lifts the head and projects it forward, and eventually brings it to the ground, thus initiating a giant wave which proceeds backward over the length of the body. Several such waves may be present at once. Carlson suggests that this movement is only an exaggerated form of the ordinary locomotion, but I am inclined to agree with Jordan ('05, p. 104) that this is probably an entirely different type of locomotion and I suspect that this snail also possesses the typical pedal wave. In fact it seems to me likely that the gallop was, so to speak, superimposed on the pedal wave system and, had the snail when in gallop been examined from below, the pedal waves would have been seen in operation in conjunction with the body waves. I am the more inclined to the view that the gallop is an independent form of locomotion as compared with the pedal


waves, because in the gallop the body waves of this species, as reported by Carlson, were retrograde whereas the pedal waves in all Helices thus far reported are direct.


Ordinary gastropod locomotion is accomplished either without pedal waves (arhythmic) or with pedal waves (rhythmic). In rhythmic locomotion the waves may run from posterior to anterior (direct) or the reverse (retrograde). The foot may exhibit one (monotaxic), two (ditaxic), or four (tetrataxic; series of waves. In the ditaxic foot the waves may be alternate or opposite.

The gastropod foot is an organ of attachment through adhesion (mucus) or suction, or both.

The pedal wave is an area of the foot that is lifted off the substrate as compared with the rest of the foot and thereby freed more or less from aahesion. It is also the region of the foot that moves forward, the rest of the foot remaining temporarily stationary. Locomotion is the cumulative result of local forward motion on the part of one section of the foot after another till the whole foot has been moved. The same type of muscular movement as that seen in rhythmic locomotion can be present in a diffuse form (not wave-like) in a gastropod foot and will result in locomotion.

170 G. H. PARKER


BiEDERMANN, W. 1905 Studien zur vergleichenden Ph}^siologie der peristaltischen Bewegungen. II. Die locomotorischen Wellen der Schneckensohle. Arch. f. ges. Physiol., Bd. 107, pp. 1-56, Taf. 1-2.

BoHN, G. 1902 Des ondes musculaires, respiratoires et-locomotrices, chez les Annelides et les Mollusques. Bull. Mus. Hist. Nat., Paris, tome 8,pp. 96-102.

Carlson, A. J. 1905 The physiology of locomotion in gastropods. Biol. Bull. , vol. 8, pp. 85-92.

Dubois, R., et Vles, F. 1907 Locomotion des Gasteropodes. Compt. rend. Acad. Sci., Paris, tome 144, pp. 658-659.

Jordan, H. 1901 Die Physiologic der Locomotion bei Aplysia limacina. Zeit. f. Biol., Bd. 41, pp. 19&-238, Taf. 2.

1905 The physiology of locomotion in gastropods. Biol. Bull., vol. 9, pp. 138-140.

KtJNKEL, K. 1903 Zur Locomotion unserer Nacktschnecken. Zool. Anz., Bd. 26, pp. 560-566.

Uexkull, J. V. 1909 Umwelt und Innenwelt der Tiere. Berlin, 8vo, 261 pp.

Vles, F. 1907 Sur les ondes pedieuses des Mollusques reptateurs. Compt. rend. Acad. Sci., Paris, tome 145, pp. 276-278.



Rull Zoological Laboratory, University of Chicago

Introduction 171

The organism as a physico-chemical system 173

1. The relation between metabolism, and structure 173

2. Physiological correlation and the physiological system or individual . . 179

3. The basis and nature of physiological correlation ISl

The nature of regulation 182

1. Organic or physiological equilibrium and equilibration 182

2. Regulation as equilibration 188

The regulatory processes 199

1. The relation between form regulation and functional regulation 199

2. The inducing conditions and the results 199

3. The provisional classification of the regulatory processes 200

a. The two methods of regulation 200

h. Regulatory compensation 202

c. Regulatory transformation 205

The nature of reconstitution 207

1. Restitution or reconstitution? 207

2. The initiating factor in reconstitution 211

3. The process of equilibration in reconstitution . 212

4. The complexity of reconstitution .215

5. The limits of reconstitution 217

Reproduction in general as a form of reconstitution 218

Conclusion 221

Bibliography 222


Of late years the term 'regulation* has come into such general use and has been applied to so wide a range of organic phenomena, that it seems desirable to attempt a general consideration and analysis of the regulatory processes from the present viewpoint of physiology. Tne biologist who takes the position that there is at the present time, when the investigation and analysis of the physics and chemistry of the organic processes is still only


172 C. M. CHILD

at its beginning, no adequate basis for 'vitalistic' interpretations of regulatory phenomena finds but little satisfaction or enlightenment in Driesch's 'entelechy' or in other assumptions of the neovitalistic school.

In the present state of our knowledge these views are and must remain expressions of personal opinion. Driesch's first two Beweise der Autonomie der Lebensvorgange" (Driesch, '01, '03 etc.), which are based on certain phenomena of form regulation, constitute proofs only when we accept Driesch's premises, and as I have pointed out (Child, '08b) these premises are pure assumptions. Neither Driesch nor anyone else has placed them on a foundation of fact. The existence of the 'harmonious-equipotentiai system,' for example, which is of so great importance to Driesch, is a matter of assumption, not of fact. So far as the systems, which according to Driesch belong in this category, have been thoroughly examined, they have shown themselves to be neither harmonious nor equipotential in Driesch's sense, and to the extent which he has assumed. It is of course easy to assume, as Driesch has done, that the harmony of these systems is due to entelechy and their limitations to physico-chemical factors, but such assumptions, since they are so manifestly invented ad hoc, do not carry conviction to the minds of most biologists, what ever, their effect upon their author.

Much the same is true of other modern vitalistic hypotheses: as expressions of personal opinion, they are of great interest in the histor}^ of scientific thought, but none of them thus far has presented any convincing arguments in its own support.

Furthermore, with the exception of Driesch's analytical consideration of the regulatory phenomena in organisms, most of the recent published works of general character, which concern themselves primarily with the regulations which involve the visible morphological features of the organism, e. g., the books of Morgan ('07), Korschelt ('07) and Przibram ('09), have been devoted chiefly to the descriotive, rather than the analytical and interpretative aspects of the subject.

In view of these facts, an attempt at physiological analysis of the regulatory processes or of some of them can scarcely be re


garded as superfluous. The further my own investigations in this field proceed, the more completely I am convinced that those phenomena, which we are accustomed to call regulations are among the most characteristic, perhaps it is not too much to sa}-, the most characteristic phenomena of life.

The views of different authors concerning the relation between regulatory and ' normal' or ' typical' phenomena are very different. Roux ('95, II, pp. S43-4), for example, makes a sharp distinction between typical and regulatory development, though he admits that the distinction is analytical rather than practical. For Driesch regulation is areturn approach to the normal condition, after this condition has been disturbed by some external factor. Many physiologists, on the other hand, have used the term 'regulation ' in a much broader sense, as applying not only to the extra-normal or extra-typical but at least to many of the most typical phenomena of life.

Because there is no general agreement concerning the real basis and nature of these processes, and because the regulations are of great importance for any interpretation of life, it seems worth while to undertake a brief analysis of them and particularly of the regulations which involve form and structure to a large extent. The present paper is concerned with such an analysis.


1. The relation between jnetdbolism and structure

The structural basis of living organisms consists primarily of colloids. These colloids, together with water, make up the greater portion of what we are accustomed to call protoplasm and in this protoplasm the various reactions and processes which characterize life occur. The universal association of colloids with life suggests that these substances play an important part in some manner in determining some of the characteristic features of life.

Metabolism has been commonly conceived in the past as consisting, on the one hand, of the synthesis of an exceedingly complex

174 C. M. CHILD

and highly labile molecule, and on the other, of the breaking down of this molecule in functional activity. According to this view the colloid structure built up is used in function and must be continually replaced.

During recent years, however, many facts have been discovered which seem to make necessary some modification of this view of the relation between metabolism and colloid structure. In the first place, most proteids, and indeed most organic colloids, are relatively inactive chemically. The common interpretation of this fact has been that death involved, or perhaps consisted in a change from lability to relative stability of the proteid molecule. But the recent work of Fischer and others upon proteids makes it highly probable that the proteid molecule, although very large, is not so complex as has been supposed, but may be polymeric in high degree. Thus the assumption of extreme lability of this molecule in the living organism becomes even more difficult than before.

Work along other lines has demonstrated that the nitrogen metabolism is only a fraction of the total metabolism of the organism and that it does not necessarily increase in proportion to functional activity. Moreover, the nitrogen requirement for maintenance in animals is apparently much smaller than had been supposed. All of these facts seem to indicate that in the living organism, as in vitro, the proteids, or many of them, are relatively inactive chemically; that after they are formed, they are excluded to a large extent from metabolism simply because of their relative inactivity. If these conclusions be correct, the accumulation of proteids in the organism is a process not very different from the deposition of other forms of inactive substance in or about the cell, e. g., chitin, insoluble salts, etc. Indeed it is highly probable that the accumulation of most or all structural substances in the organism is due to the fact that they are relatively inactive under the existing conditions. After they have arisen in metabolism they persist or disappear much more slowly than other substances, simply because under the exi.sting conditions they do not enter chemical reactions as readily as other substances.

But the proteids and other substances deposited in the cell are not absolutely inactive and undoubtedh' do enter metabolism


to some extent at all times. Under certain conditions, however, i. e., in the absence of certain other substances, they, or some of them, may reenter metabolism to a much larger extent and furnish energy. It is a familiar fact that in the absence of nutritive material from without the organism uses up its own substance, ^. e., the relatively inactive substances which under other conditions had accumulated in it. In certain of the lower organisms this process may continue until the organism is reduced to a minute fraction of its original size. These facts do not, however, conflict with the suggestions made above as to the relative inactivity of structural substance, but serve rather to confirm the idea that the accumulation of these substances, when other nutritive material is present, is due to their relative inactivity.

Various authors have attempted to distinguish between a morphological and a functional metabolism, but it is doubtful whether such a distinction is valid, except as an expression of the fact that substances of different degrees of chemical activity arise in the course of metabolism and that certain of the less active substances constitute the structural basis of the organism, while others undergo chemical transformation and elimination.

It is of course not merely the nature of the substances themselves, but the existing conditions as well, which determine the degree of activity or inactivity. Under certain conditions a cell or an organ may accumulate certain substances and so acquire a certain characteristic structure, while under altered conditions these substances may rapidly disappear and others be accumulated. Thus, for example, the oocyte, during its growth period, accumulates yolk, which under the existing conditions is almost wholly inactive chemically and so appears as a structure-building substance. But when fertilization occurs the conditions within the cell are so altered that the accumulated yolk rapidly reenters metabolism and serves as nutritive material. In fact we may say that the egg does not produce yolk because it is to develop into a new organism, but that it develops as it does because it has accumulated yolk. In the periodic changes in the cell connected with growth and division there is also abundant evidence for the occurrence of changes of this character.

176 C. M. CHILD

If we may accept this view of structure in the organism, and all the facts are in its favor, then it is actually very similar in its relation to the energy current to the morphological characteristics of a river system except of course that the latter are mechanically produced. The constructed islands and bars, the depositions of the river, represent those particles or masses which have, under the conditions existing at a given time and place, been left behind by the current. Under certain conditions the river may produce structure of a certain kind at a certain point in its course, while under different conditions this structure may disappear and give place to structure of a different kind.

But the most important fact for present purposes is that in the organisms, as in the river, structure, as soon as it appears, begins to influence the metabolism, the energy current. From this time on the metabolic processes, like the flow of the river, occur in a certain structure and here the mutual interactions begin. Of the character of these interactions in organisms we are only beginning to obtain some vague conceptions, but that they occur, it is impossible to doubt.

Perhaps a few words will not be out of place concerning the bearing of these facts and suggestions upon the theory of * formative substances ' which has played a considerable role in embryological investigation during the last few years. Most of the supporters of this theory have attempted to identify the so-called formative substances with visible granules or other accumulations in the cytoplasm, without considering the fact that the appearance of these substances in visible structural form indicates that they are, at least for the time being, relatively inactive, and that they are first of all products or incidents of metabolism (Child, '06b). Of course some or all of these substances might reenter metabolism under altered conditions and so play a part in determining its character, but the important point is that they are indications of a difference in metabolism already existing in the different regions where they are formed. We might expect that the differences in metabolism, which are certainly more important as formative factors than these accumulations of granules, would persist in the different regions, even if the granules could


be removed. Fortunately the embryologists themselves have now a method of removing these granules from their usual position, i. e., the method of centrifuging the egg, and the results of recent experiments along this line indicate, as was to be expected, that the granules are quite unessential to regional localization and differentiation of the embryonic structures.

It is evident from the above suggestions that our fundamental conceptions of the relation between structure and function in organisms must be intimately connected with our ideas concerning the nature of colloid substances and their significance as a substratum or medium for chemical reactions. Within recent years it has been pointed out repeatedly that these substances afford various means for the partial or total isolation, of different chemical reactions in organisms and that their mere presence may bring about such isolation, e.g., by the formation of semipermeable membranes. Here then we have a physico-chemical basis for localization and differentiation. Moreover, the changes in the physical aggregate condition of colloids, together with the possibility of the simultaneous existence of different phases of high and low water-content, must play a part in determining the degree and place of dissociation of various substances, and therefore in determining the speed of reactions in different regions, as well as the occurrence or non-occurence of certain reactions at particular points. It is then, to say the least, highly probable that the possibilities of localization, physiological specification and the accompanying possibilities of physiological correlation of parts and of regulation are very closely connected with the fact that the formation of colloids is a component of the reaction complex known as metabolism. Moreover, as I have attempted to show in another paper (Child, 'lib), the accumulation of relatively inactive substances, particularly colloids, in the cell is undoubtedly a factor in senescence, in that it constitutes an obstacle to the metabolic interchange and so brings about a decrease in the rate of metabolism.

If the conclusion be correct, that the visible structural elements of the cell are, at least for the time being, relatively inactive chemically, then it follows that these elements do not represent the

178 C. M. CHILD

'living substance' in the stricter sense, but are really the least 'alive' of any part. The reactions which furnish the energy of life undoubtedly occur, at least in large measure, in the more fluid parts of the cell, the parts which present the least characteristic structure. The so-called living substance is actually then, so far as it presents a visible structure, chiefly a substratum or m€dium in which the reactions occur, and is itself the product of past reactions. That these structural elements, as they accumulate, must modify the rate and character of the reactions to an increasing extent, cannot be doubted. The advancing specialization of metabolism in different organs and cells is probably closely connected with the fact that these parts produce different structural elements, either in consequence of an original specification or in consequence of different correlative or external conditions which induce specification.

If we accept this view of the relation between function and structure in the organism, we must give up the idea of a definite 'living' substance in the chemical sense, and the basis of life becomes, not a specific substance, but a series of reactions in a field or medium of a certain complex constitution, which is itself the product of past reactions. We can agree with Driesch ('01, p. 140), as regards the absence of a specific living substance, though we cannot follow him in his further conclusions along this line. The life process has become individualized, not because of entelechy, but because it forms its own field or medium of action, as the river forms its channel, particularly in the later stages of its course, where deposition exceeds erosion. To put the matter briefly, life as we know it consists not in metabolism alone nor in a specific substance or structure alone, but in the physiological correlation of processes in a structural medium or substratum of a certain constitution, which makes possible localization and correlation of processes.

It is then the existing relation between the processes and the structural substratum, the mutual interaction and dependence of both, that forms the basis for the phenomena of regulation or equilibration which occur in the organism.


2. Physiological correlation and the physiological system or individual

Organisms in general appear in the form of more or less sharply defined physiological systems or individuals and in the more complex organisms we can distinguish systems or individuals of various kind and degree. What is the basis of this unity?

The experimental investigation of orgaaisms has led those who are not yet ready to accept vitalistic hypotheses to the conclusion that two factors are chiefly involved in the formation of a living system or individual, viz. constitution and physiological correlation. In its grosser aspects the first of these is the morphological, the second the physiological factor. Most of us believe, however, that the morphological features of organisms are essentially visible expressions of dynamic processes past or present and that sooner or later we must interpret constitution in dynamic terms. The factor of physiological correlation in the organism is essentially the problem of physiology, for in the final analysis function is impossible without such correlation.

Wherever in the universe unity can be recognized, there some sort and some degree of correlation must exist, either conceptually or as a datum of nature, between the elements which compose the unit, and vice versa, wherever correlation between conceptual or phenomenal elements is recognized or established, there a unity of some sort and some degree exists. On the other hand, the character of the unity is determined by the nature or constitution, however this may have arisen, of its elements.

There is at present no adequate ground for believing that organisms differ from other phenomena in these respects. We cannot conceive an organic individual without correlation of some sort between the parts which compose it, nor can we conceive it without elements or parts of a certain more or less characteristic constitution.

The development of morphology and its separation from other fields of biology during the latter half of the nineteetith century has led, particularly in the field of zoology, to the consideration of the problem of constitution apart from that of correlation. But the

180 C. M. CHILD

introduction of the experimental method into zoology has alreadydemonstrated the limited scope and value of pure morphology for the interpretation of life. In the organism as we find it, the two factors, constitution and correlation are mutually determining. We cannot alter the constitution without altering the correlation of parts, neither can we alter correlation without changing constitution to a greater or less extent. The cases of so-called selfdifferentiation constitute no real exception to this statement,^ which has the value and significance of a law of nature. Morphology and physiology are inseparable except analytically and their artificial separation can lead only to the formulation of many pseudo -problems and to uncertain or false conclusions and hypotheses.

In so far as the organism is a physiological, i. e., a physicochemical individual or unity, in so far must physiological correlation exist between its parts in the form of actual physical and chemical processes, conditions and substances. Until it is proven by the profoundest investigation and the strictest analysis that physiological correlation does not suffice to account for the organic individual, there is no need of turning to the vitalistic hypotheses for an interpretation.

Indeed our knowledge of physiological correlation is in its earliest stages. One need only refer to the work on the conduction of stimuli in plants and through protoplasm in general and to the investigations of recent years on the thyreoid, the adrenals, the reproductive organs, the pancreas, etc., as organs of chemical correlation and to the work on hormones, to become aware of the advances in knowledge along this line within the last few years. At present we are willing to believe, in fact we find it difficult not to believe, that every metabolically active organ in the body is an organ of chemical correlation. And we also know that many

1 As Roux ('95, p. 822 etc.) has pointed out development depends primarily upon correlation and absolute self-differentiation cannot occur. In cases where parts differentiate in a relatively high degree of independence from each other, we must believe, and in some cases, e. g., the nemertean egg, we know that this condition is preceded by, and is the result of an earlier condition in which the parts are in much closer correlation.


of these correlative factors show a high degree of specificity. Moreover, the chemical factors are by no means the only factors in correlation: mechanical and other physical factors also play a part. And finally, the experimental investigations themselves have demonstrated the importance of physiological correlation in morphogenesis.

In short, there is at present every reason to believe that the existence and continuity in time and space of organic individuality are essentially dependent upon physiological correlation, i.e., upon processes and conditions which are accessible to scientific investigation and analysis.

3. The basis and nature of physiological correlation

That Dhysiological correlation is in general dependent upon the physical and chemical processes and conditions in the various parts which make up the individual cannot be doubted. These in turn are deDendent upon the constitution of the parts, which itself depends in part upon preexisting correlation and to a greater or less extent upon conditions and processes in the extra-individual environment. At every step in our consideration we recognize the mutual interdependence of constitution and correlation.

But if we consider the organic individual only as it exists at the present time, then we may say that the existing physiological correlation between parts is dependent upon the conditions and processes in the parts, however these may have been brought about.

In general we can recognize at present three main groups of correlative factors : first, mechanical or mass correlation (Roux, '95, II, p. 240), which results merely from the existence of mass without respect to constitution; second, substantial or material correlation, which consists in the actual transference or transportation of substance possessing a certain physical or chemical constitution, e. g., chemical correlation; and third, dynamic correlation, of which the essential feature is the transmission of energy rather than the actual transportation of material over any appreciable distance. None of these forms of correlation can be sharply

182 C. M. CHILD

separated from the others in the final analysis, but in its extreme forms each type is readily distinguishable.

The physiological correlative effect of a part upon others is then the result of all that that part is and has been in the past, of its physical and chemical constitution, its position, its relation to external factors and of the changes which are occurring in it. It is apparent that there exists in physiological correlation the possibility of an almost infinite variety and specificity. Driesch ('09) has recently maintained that the specificity of the 'Restitutionsreiz' together with the specificity of the reaction to it constitute' an 'Individualitat der Zuordnung' which is inexplicable on a physico-chemical basis and which therefore constitutes a new and independent 'proof of the 'Autonomic der Lebensvorgange.' Comment seems scarcely necessary. One sees here merely an assertion, a jump at conclusions, but no proof, where proof of the most convincing character is absolutely essential. If vitalism can present no more convincing arguments than this its future prospects in science are not bright.

THE NATURE OF REGULATION 1. Organic or physiological equilibrium and equilibration

One of the most characteristic features of organisms is, as Roux ('95, I, pp. 145, 154, 392, etc.) has said, their continued existence as individuals, their 'Dauerfahigkeit' amid changing internal and external conditions. On the other hand this 'Dauerfahigkeit' is only relative, not absolute, i. e., it is limited. The organism is constantly changing, and- so far as our knowledge goes, never twice the same, yet the continuity of individuality is obvious.

Nevertheless the continuity of the existence of individuality must not be emphasized to the exclusion of the fact that under certain conditions this individuality may disappear, at least in the simpler organisms, and be replaced by other individualities in larger or smaller number. Certain factors concerned in this physiological disintegration will be discussed below, but for the present we are concerned with the individual, the system as we see it


in the organism or part which constitutes a unity cHstinct to a certain extent from others.

When we investigate the processes in the organism, we find that they are very intimately connected with one another: a change in one conditions changes in others. Moreover, and this is an important point, the physiological specification of different parts is not in most cases absolute. In the highest, most complex forms absolute specification is doubtless approximated more or less closely in certain organs, but in general we find that the processes in different parts of the organism are not fixed in character. The characteristic series of reactions in a part does not represent the only possible series, but rather the particular series determined by a particular complex of conditions. A certain process may occur at one time in a certain part, at another time in others. In short there is more or less possibility of substitution among the different parts.

Let us suppose, for example, that a certain correlative factor x originates in a certain part. Under certain conditions this factor may influence various other parts, a b c — n,of which one, a, let us say, reacts with greater speed or intensit}^ than others. The reaction of this part may itself produce new correlative factors and so alter conditions in the others that their reactions are changed. But if we suppose that the receptivity of the part a to the correlative factor X is decreased, or that the part a is itself removed or rendered incapable of reaction, then the reactions of a b — n or of some of them are not altered or prevented by the effect of the reaction of a, and these parts may take the place of a in the system, though perhaps at first reacting more slowly or less intensely than a, until their constitution has become altered by repeated or continued reactions.

Through such a series of reactions the individuality of the organism is maintained, or restored, even though it may have lost a part. We see exactly such reactions in various organisms, and we can devise physico-chemical systems which show similar correlation between parts and a similar method of maintenance or restoration of something approaching the preexisting condition. In the system which we devise such processes are processes of equili

184 C. M. CHILD

bration. We find that the system is capable within certain limits of attaining or approaching a condition of equilibrium after a disturbance of a previously existing equilibrium.

And again in the law of mass action and the general principles of chemical equilibrium, together with what we know of katalysis, we have the possibility of accounting for a great variety of processes of equilibration in the organism. Until we have exhausted these and other physico-chemical possibilities and found them inadequate, we have no adequate reason for oelieving that organic individuality and its maintenance are anything unique.

The fact that physiological correlation exists between different parts of an organism must necessarily determine a certain relation, a certain proportionality in the activities of the different parts. It is this relation, this proportion in activity determined by correlation which constitutes what we call organic or physiological equilibrium in the organism. This equilibrium is dynamic, not static, it is an equilibrium of processes, not of masses and it must be dependent either upon physiological correlation, or upon something else which controls the supply of energy to this or that part in very much the way in which the man in charge controls the workings of a complex machine, e.g., a steam-shovel, turning the steam into this or that cylinder as required for the harmonious working of the whole. Driesch's entelechy is comparable to the man in charge of the engine.

But it is not the mere existence of an organic equilibrium which constitutes the real problem; it is the apparent power of adjustment, of equilibration, the harmony of action of the parts, as in the engine, which has been regarded as the strongest argument for vitalism. How, the vitalist asks, is it conceivable that a machine with such capacity of adjustment, of equilibration as the organism, which can even repair itself, can be constructed and continue to exist and work unless there is something comparable to the man in charge concerned in these processes.

As a matter of fact this question is based on a wrong conception of the organism. The organism as we see it, i. e., morphologically, is not the machine whose action constitutes life, but rather simply a part of the products of that machine, which accumulate during


its action and as they accumulate, alter and determine the character of its activity. In other words, as the products are formed they become a part of the machine. Starting with the egg, the organism is not, as Driesch asserts that it must be according to the 'machine theory,' a machine developed /or function (Driesch, '05, p. 790), but rather a machine developed by function. The result at any stage represents morphologically the products of a preexisting machine and physiologically the action of the machine as altered from the preceding stage by the products of its own activity. Each stage of development is the result of the machine plus the product of the preceding stage. Our experiments have shown that physiological correlation, not predetermined harmony is the basis of development, and that where a predetermined harmony appears to exist it is certainly in some cases, probably in others, the result of an earlier condition of correlation.

On the basis of this conception of the organism it is inconceivable that processes of adjustment of the parts to each other, i. e., processes of equilibration, should not occur, both in development in nature and under experimental conditions. The parts are what they are, not simply because of their original constitution, but because they have been acting in correlation with each other. From the moment the organic machine began to work in the first organism 'adjustment' of the parts to each other began and it has continued ever since. Could we but read it completely, every part is a record, an epitome, more or less complete according to circumstances, of what has been going on, not merely in itself, but in the whole organism. Moreover, in different parts this record is written in different characters, in different languages, according to the constitution of the part.

The distinction which Roux makes between the formative and the functional periods of development (Roux, '95, II, p. 281), is, according to this view, not fundamental in character. The formative period is functional and the functional period is formative. But this distinction is based upon the fact that at a certain more or less sharply defined stage of development the accumulated products of the activity of the machine begin to play a more or less definite role in its further physical and chemical activity. The adult

186 C. M. CHILD

organism is not then to be compared with a machine constructed of certain definite Darts, which have been put together in some way, and which, after completed construction, begin to function. It is much more nearly comparable to a river, which molds its banks and bottom, forming here a bar, there an islaad, here a bay, there a point of land, but still flowing on, though its course, its speed, its depth, the character of the substances which it carries in suspension and in solution all are altered by the structural conditions which it has built up by its own past activity. In such a system a wide range of equilibration exists and we see both the adjustment of function to form and of form to function. The relation between structure and function in the organism is similar in character to the relation between the river as an energetic process and its banks and channel. From the moment that the river began to flow it began to produce structural configurations in its environment, the products of its activit}^ accumulated in certain places and modified its flow% but just so long as the flow continues the process of equilibration goes on. If we consider merely a certain region of the river with the water containing certain substances in suspension and in solution entering at one end, depositing some of these substances and taking up others as conditions determine in the course of its passage, and finally passing out at the other end bearing certain substances more or less different from those which it brought in, the analogy becomes even more complete. In fact this region of the river, together with its bed, shows a real, though chiefly a mechanical rather than a chemical metabolism.

I believe that this comparison between a river with its channel and the organism is far more than a fanciful analogy. Theindividdual organism is merely a section from that current of energy which constitutes the essence of life, and in the individual we see the mutual correlation and interaction between the current and the conditions under which it finds itself, between the energetic process

- Rignano ('07) has referred briefly to this analogy between the river and the organism, using the case of a river equilibrating itself in connection with the piers of a bridge to illustrate the process of equilibration in organisms. See also Delage, L'Heredite, etc., 1903.


and the structural features which its activity has produced. As the banks and the channel are 'adjusted' to the activity of the current, and the current to the morphological characteristics of the banks and bed, so, and in no otherwise are structure and function in the organism, correlated with each other. It is absolutely inconceivable that 'adjustment,' equilibration should not occur. So long as the current flows, equilibration must take place in one way or another.

The organism has often been compared to a flame. Roux particularly has carried out this comparison in detail (Roux, '05, p. 109, et seq.). Although this analogy contains much that is valuable and on the chemical side is much closer than that of the river, yet on the other hand the morphological features of the river are more nearly comparable to those of the organism, In their localization, their often complex structure and their modifying effect upon the activity of the current. For these reasons I have chosen the river rather than the flame as a physico-chemical system with which the organism may be compared.

When we take the view of the organism suggested above, I believe that Driesch's first two proofs of the autonomy of vital processes (Driesch, '03, p. 74, etc. Cf . also p. 197 below) appear in their proper light. They apply only to the morphological conception of the organism as a machine constructed for function, i. e., to the banks and channel without the river. In the organism the current is working from the beginning, the organism is functioning in one way or another, and the real machine is the process, the function, plus the existing structure which past processes have produced, just as in the case of the river the real machine is the current plus the banks and channel. The process of development in the organism is comparable, not to the digging of a channel into which, after its completion, the water is turned, but to the formation of a channel with certain characteristics determined by a variety of conditions, by the activity of the current itself. From the moment the current begins to flow, structure and function become mutually interdependent and mutually determining, but there can be no river -structure without the current. Machines like the steam engine, constructed by man and considered without


188 . C. M. CHILD

their motive power, are comparable rather to the dead than to the living organism. They are merely the conditions under which the energy acts, but the living organism consists from the beginning of these conditions plus the energy. Development is not comparable to the construction of such a machine by man, but rather to its action after the steam is turned on. Every steam engine possesses a certain power of equilibration dependent upon its constitution, and the only reason its powers in this direction are so narrowly limited is because the energy current and the structure have not been working together from the beginning.

The only possible basis for a scientific, as opposed to a philosophical vitalistic hypothesis is the proof that the energy of organic life is something essentially different from the energj^ of the physico-chemical world. When the vitalists shall succeed in proving this or even in making it probable, then their views will be given more general consideration. But even the most extreme among this school at the present day do not attempt such proof. If we admit that the energy of the organism is not different from that in the physico-chemical world, then I believe we are forced to regard the organism as a physico-chemical system, for as I have shown above, physico-chemical systems exist in which the relation between structure and function, between the conditions of action and the energy itself, are of the same character as in the organism itself and give rise to a power of equilibration of the same character.

2. Regulation as equilibration

From what has been said it will be at once apparent that the processes which we commonly call regulatory are processes of equilibration in the organism (Holmes, '04, '07, Child, '06, '08a). They enable the organism to persist and to maintain its individuality under changing conditions, although it cannot be supposed that the condition of dynamic equilibrium is the same for different conditions, and indeed we have evidence that it is not. But within certain Umits, and for certain factors, the organism is capable of a greater or less degree of equilibration, when a change in external conditions occurs.


The question at once arises as to whether all processes of equilibration are to be regarded as regulations, or only certain of them. By zoologists the term 'regulation' has been applied mostly to processes occurring under experimental conditions outside the usual range of conditions in nature and the regulations of form and structure have been the chief, though not the only objects of investigation. Jennings ('06) has used the term with reference to phenomena of behavior which are characteristic features of life and not of abnormal or pathological conditions. Among the physiologists also we find the term often used as referring to various changes in metabohsm and reactions of different kinds in response to conditions to which every individual is subjected repeatedly.

If we define regulation as a return or approach to a condition of dynamic equilil^rium in a living organism after a previously existing condition has been disturbed by some external factor (Child , '06) , we shall include all the above phenomena as well as many others. According to this definition, the simplest reflex as well as the restoration of a missing part is a regulation, the simplest correlative compensation in metabolism, as well as the development of a whole from an isolated blastomere of an egg.

Moreover, when a complex part of an organism undergoes an equilibrating change in reaction in response to altered correlation with another part or other parts, a regulation occurs as truly as when the whole organism responds to some change in conditions outside of it. In short, regulations are equilibrating reactions to changes external to the reacting system, whether this system be a part or a whole of an organism.

And finally, regulation is not limited to the return or approach to the preexisting condition, but may be an approach to a condition very different from that, i.e., the organism or the part may become something more or less widely different from what it was originally. In every case of regeneration of lost parts some of the cells become something different from what they were before the part was removed, and their change is a reaction to altered conditions and specifically to altered correlation.

But that the regulatory process is always and necessarily of advantage to the organism does not follow from the definition.

190 C. M. CHILD

So far as it enables the organism to persist, it may be of advantage, and I see no escape from Roux's argmnent (Roux, '95, 1, p. 145, 154, etc.), that systems possessing such reactions will persist longer than others. But not all regulatory processes are of advantage to the organism and many of them, e. g., the so-called axial hetermorphoses, lead to its destruction or its disruption, but they are no less regulations because of this result.

According to Driesch ('01, p. 92), Regulation ist ein am lebenden Organismus geschender Vorgang oder die Anderung eines solchen Vorgangs, durch welchen oder durch welche eine irgendwie gesetzte Storung seines vorher bestandenen 'normalen' Zustands ganz oder teilweise, direkt oder indirekt, kompensirt und so der 'normale' Zustand oder wenigstens eine Anniiherung an ihn wieder herbeigefiihrt wird."

If we accept this definition, then the processes which do not constitute a return or approach to the previously existing 'normal' condition are not regulations. This normal condition is nothing but the condition which corresponds to a certain complex of external factors or to changes within certain limits. Under changed conditions a new equilibrium, not the old, is established. In short, if we accept such a definition, we not only exclude many processes which are as truly regulatory as any, but we are forced to assume the existence of an entelechy or other similar principle to account for the 'normal' condition and its maintenance.

Regulatory processes are determined in character and direction by the nature of the organism, on the one hand, and the nature and amount of the external change, on the other. Under the given conditions, the organism or part is capable of doing only the one thing; under other conditions, or with a different constitution, the regulation may occur in a different manner and may often lead to a different result. In Planaria, for example, the course and result of regulation differ according to the size of the piece, the region of the body from which it is taken, the temper iture, the nutritive conditions and other factors. To say that the pieces always produce a whole under all these conditions means but little, for the wholes which they produce are not alike. In plants the character of the external change often plays a very large part in deter


mining the character of the regulatory processes. In many cases, however, in both plants and animals, the action of the external factor is so indeterminate, or the external conditions are so complex that equilibration may occur in various ways under what seem to be, but are not actually similar conditions. Thus, as Jennings has pointed out, in the regulation of behavior the disturbance, the stimulus, may merely bring about reactions of an indeterminate character, which sooner or later, in one way or another lead to equilibration. Evidently then the relation between the character of the external change and the character of the regulatory process differs very widely in different cases.

The initiating factor in regulation is the external change, the disturbance of the preexisting condition. This change brings about changes within the organism or the part and these in turn lead to changes in the correlative factors, and so to equilibration or to disruption and death, in case the external change is such that equilibration of the system as a whole is impossible. But so long as the energetic processes of life continue in the system, equilibration of some sort must occur. To return to the analogy of the river, so long as the water flows, equilibration of some sort occurs, whatever the changes and whatever the obstacles. The river may alter its course, it may transform its banks and its channel so that they bear little or no resemblance to those existing before the change, it may divide into a number of streams, each of which pursues its own course, according to the conditions under which it finds itself, and builds up its own structural characteristics. In all cases, however, unless the conditions are such as to stop the flow of the water, equilibration takes place in some manner.

The range of regulatory capacity in the organism represents then merely the range of possibilities within which the flow of the current of energy which constitutes metabolism and which is the essential feature of life, is possible. Within these limits it is absolutely inconceivable that regulation or equilibration should not occur. The nature of the process depends upon the nature of the orgaaism and the conditions which it meets.

Equally inconceivable is the occurrence of regulation as a process of life under conditions which stop the metabolic current.

192 C. M. CHILD

Organisms which meet such conditions are simply eliminated. The survival and elimination of organisms is determined primarily, not by their morphological characteristics, but by their capacity for regulation of one kind or another under the conditions in which they find themselves. To attempt to understand the course of evolution from morphological characteristics alone can only lead to confusion and failure. Only a knowledge of the nature of the metabolic current in organisms and the possibilities of its equilibration under different conditions can lead to a theory of evolution and heredity which will stand.

For example, the evolution of animals and plants, like fevery other evolution, is based primarily upon differences in the metabolic processes. These undoubtedly originated as regulations and as soon as they had arisen, gave different possibilities ot further regulation : in the course of the realization of these different possibilities in accordance with the conditions of existence, animals and plants with their different morphological characteristics have arisen. In each case the visible structure represents merely a partial record of the realized possibilities. All the structural 'adaptations' in both animals and plants are based upon the processes of equilibration of the energy current and must sooner or later be expressed in terms of this current and its environment. They are not the primary and essential features of the organisms, they give us merely an outline, a diagram of the most characteristic activities of the energy current. As the banks and channel of the river, even after the water has ceased to flow, enable us to gain some conception, though a very incomplete one, of what the river has done in the past, so the structure of the organism is merely a rough sketch of what the current of life has done in the "way of deposition, arrangement and removal of materials along its course. Many of the past activities of the current are not distinguishable in the structure because their effects were slight or transitory, or because they have been masked or altered by later activity of a different character. As the river in some process of equilibration, e. g., in a flood, a period of increased energy, may sweep away many of the records of its previous activity, so the


organism, in a period of increased metabolism may remove the structural evidences of past metabolism.

In the present state of our knowledge we should think it absurd to attempt to account for the configuration of the banks and bed of the river without taking into account the action of the current. It would remain a miracle, which we could ascribe to the caprice or other quality of a personal creator, or to some other mysterious natural force. In the same way, when we attempt to interpret the structure of organisms without direct reference at every step to the current of energy of which the structure is evidence, we must neces-' sarily go astray or end in confusion or in the most bizarre hypotheses. We can do as Driesch has done and shift the burden to the shoulders of entelechy, to which we can ascribe such qualities as may please us. Or we can speak of biophores and determinants, pangens, or whatever we please to call them, or we may pin our faith to the visible chromosomes, but these are nothing but creators of a type which appeals to certain minds.

On the other hand, when we take as our starting point the process of metaboHsm, we are proceeding as the physiographer has learned to proceed in his study of rivers. As we learn how metabolism produces structure we shall be able more and more completely to interpret the nature and the past history of the organism from its structure, but at every step we must return to the process, the current, in order to understand, and we can never hope to understand all through structure, simply because structure is an incomplete record. Life is first of all an energy process, a flowing current. All that is relatively stable, all that persists as visible form and structure, represents merely some past action of the current occurring under certain conditions. Almost sixty years ago Huxley said concerning the cells: They are no more the producers of the vital phenomena than the shells scattered along the sea-beach are the instruments by which the gravitative force of the moon acts upon the ocean. Like these, the cells mark only where the vital' tides have been, and how they have acted. "-^ And even yet the truth of these words is not recognized as it should be by biologists.

  • British and Foreign Medico-chirurgical Review, vol. 12, p. 314, Oct. 1853.

Cited from Whitman, the inadequacy of the cell-theory, Jour. Morph., vol. 8, 1893.

194 C. M. CHILD

Only when we take into consideration the motive power and the method of its action under the given conditions, can we hope really to advance in our knowledge of how things come to be as they are in the organism, or to determine and predict what they shall be. Man has attained his present position by acquiring knowledge and control of energy in nature. Can he hope to advance in his insight into the problems and his control of the processes of life in any other way?

According to this point of view, life, like every other continuous energetic process, is essentially a series of equilibrations, of regulations. When regulation shall cease, evolution and li'e will also cease. The power of regulation in organisms is nothing unique, but is something which they possess in common with all energetic processes in nature,which continue for any appreciable time. In fact, strictly speaking, all energetic processes in nature are equilibrations.

As was suggested above, the range of regulatory capacity in organisms is undoubtedly due in large measure to the fact that the process of metabolism produces certain colloid substances, among which the proteids and lipoids are the most characteristic. With the first proteid synthesis under certain conditions in nature the processes of regulation of the type which we find in organisms began. Perhaps we may say that life began here also. The reaction which was concerned in the first synthesis must of course have preceded the completed synthesis, but as water apart from the channel which it forms for itself in its environment is not a river, so a given chemical reaction, or a series of reactions, apart from the conditions which it produces where it takes place, is not life. We may say if we please that life began as a chemical reaction, but we must recognize the fact that the occurrence of that reaction produced certain characteristic conditions, which played a part in determining the course and character of further reactions : in short, the reaction determined the existence of structure and the mutual interrelations between structure and function: and finally, with the existence of structure of colloid nature, the possibility of regulation of the organic type also appeared, and regulation began.


My purpose in laying special emphasis upon the point that regulation is an essential characteristic of life and that life must cease when regulation ceases, is merely to show that the extreme forms of regulation, which occur under experimental or accidental conditions, are in no way different from the processes of life apart from experimental or accidental interference. The capacity for regulation is not something secondary or something acquired in the course of evolution, but it is as inseparable from life itself as the power of equilibration from the flow of the river. Not only life but the universe is an unceasing series of regulations. Every experimental investigation performed with living organisms is, so far as it does not lead to the death of the organism, an investigation of organic regulation, and death itself is an equilibration, though of another type.

To set the regulations ofT as a special category of phenomena, occurring only in organisms and of secondary or incidental significance in these, must of necessity lead to conclusions of the same character and value as those which would be reached by one who should attempt to investigate the phenomena of equilibration in the river, without considering the flow or the resistance of its banks and bed. Such a one would doubtless marvel at the wonderful harmony of action displayed by the simultaneous disappearance of a part of the bank and the encroachment of the water upon it, or by the appearance of an island and the division of the river into two channels. He would doubtless call attention to the remarkable fact that both the channel and the river were narrow and deep at some points and broad and shallow at others. He might wonder why stones moved along where the bed was steep and only fine particles where it was nearly horizontal. If he were of an investigating turn of mind, he might throw stones into the river and observe the consequences, or he might dig a ditch and turn part of the water into it. Thus he would observe further remarkable harmonies of action. If he were inclined to look for causes, he would probably conclude that the complex of phenomena was determined and controlled by some mysterious being or principle, which, judging from his own ability to bring about harmony of action between different things in his world, he would

196 C. M. CHILD

conceive to be more or less like himself, though greater, more perfect and more powerful. Doubtless also he would give it a name.

But when once the idea of the flow of the river as a motive power has entered his mind, his whole attitude toward what he has seen is altered. He sees that it is the current which carries away from the bank or stones and mud along the channel. On the other hand he sees that the banks confine the river, that the island, which it has formed divides it, that it accommodates its form to the ditch which he has dug and at the same time begins to change it. He begins to realize that the remarkable harmonies which he has observed are the result, on the one hand, of the flow of the river, i. e., in i further analysis, of the characteristics of water, and on the other, of the nature of its banks and bed. He will also realize in time, that just as long as the flow continues these harmonies of action will continue to occur. Then he may begin to investigate the characteristics of currents and of water in general, and later we find him devising water-wheels, dams, pumps etc., i. e., bringing about the most various harmonies of action between the flow of water and other phenomena.

His conception of what he saw was at first more or less similar to that of the vitalist concerning organisms and all his investigation could only end in speculation, which did not advance his real knowledge. But when he once began to reahze the action of the current as an energetic and a constructive process, then he saw that the harmonies of action were only apparent, not real, because he was dealing with mutually dependent phenomena rather than with those which were independent and predetermined.

Driesch, for example has maintained in criticism of some of my own earlier statements, that development is for function (Driesch, '95, p. 790) and the same view is apparent in his repeated comparison of the organism to a machine constructed by man. This is as if our hypothetical man should maintain that because he could dig a ditch and turn water into it, therefore the channel of the river must have been constructed by some 'entelechy,' or other principle for the water, and then the water turned in. And more specifically, Driesch's 'proofs' of the autonomy of vital proc


esses, ^ which are based on the phenomena of regulation are not proofs at all, because the 'machine' which he has in mind is comparable to the dead, rather than to the living organism, to the river frozen solid, rather than to the river flowing. Tf we could separate a portion of this frozen river with its channel from the rest it would of course remain what it was, i. e., a part, so long as it remained frozen. But if we divert any sufficient quantity of water from the flowing river it is capable of forming a whole which shows all the essential characteristics of the original river, though not identical with it. In short each flowing river, with its banks and bed is a 'machine' according to Driesch's definition, eine typische chemisch-physikalische Spezifitatskombination" (Driesch, '01, p. 187), and it may become whole when parts are taken from it or when their relative position is changed; moreover, when it is divided, each part may form a whole essentially similar in its processes and structure to the original whole. The existence of such a 'machine' is therefore a sufficient refutation of these 'proofs' of Driesch's. So long as the current flows such regulatory processes are not only possible but necessary, when the conditions arise. Neither the organism nor the river 'remain w^iole' when parts are taken from them, but they become new wholes, which under similar conditions, may become more or less like the original whole (Child, '08b), but which under other conditions, may be different.

Driesch's error is two-fold : although his general definition of a 'machine' is sufficiently broad, his argument in the 'proofs' is based only on a certain type of machine, viz., that constructed by man for function, a type which is wholly passive during its con

  • A brief statement of the first two ' proofs' is as follows :

"Erstens: Eine Maschine bleibt nicht dieselbe, wenn man ihr beliebige Telle nimmt oder ihre Telle beliebig verlagert; deshalb kann das sich auf Basis harmonisch-aquipotentieller Systeme abspielende Formbildungsgeschehen kein maschinelles chemisch-physikalisches Geschehen sein.

"Zweitens: Eine nach den drei Dimensionen typisch spezifisch verschiedene Maschine bleibt nicht ganz, wenn sie geteilt wird, deshalb liegt der Geneseaquipotentieller Systeme mit komplexen Potenzen im Bereiche des Formbildungsgeschehen kein maschinelles chemisch-physikalisches Geschehen zu Grunde." Driesch, '03, p. 74.)

198 C. M. CHILD

struction, rather than on a type in which, as in the organism, the structure at each stage is determined by the function and the structure in the preceding stage. The organism is comparable not to the constructed 'machine' alone but to the machine plus the constructing activity, and since Driesch has confined his argument to the type of machine constructed by man for a definite purpose, he is very naturally and logically led to the assumption of a constructor. His 'proofs' are equivalent to the argument that because a ditch built by man for a particular purpose and possessing a specific structure but containing no water does not remain whole or the same when we take away parts of its banks or bottom, therefore the river, as we see it in nature cannot be a physicochemical system.

Similarly in his consideration of the organism he has failed to take account of the constructive activity of the continuous flow of energy in a given environment. The organism is, he says, constructed for function. His position is identical with that of our hypothetical man who concluded that the channel of the river must have been constructed for the water, and like him, Driesch has given his imagined constructor a name, or rather has adopted an old one for it, viz., entelechy. Most of us have concluded from our observations and experiments that the channel of the river is constructed by the activity of the current and we have some rather conclusive evidence upon that point. Before he can hope to see his views accepted, our man must actually prove or make it at least probable that this is not so. The burden of proof lies wholly upon him. And similarly, until Driesch can make it at least probable that the organism is constructed for and not by function, instead of merely assuming this to be the fact, he cannot expect to find wide acceptance for his views. Nowhere in Driesch's work do we find any convincing evidence upon this point: Driesch has simply chosen to assume that it is so. I am of course aware that Driesch regards entelechy as in constant connection with physicochemical factors and as working with these as means. But I see no reason why, if we postulate an entelechy for the organism, we should not at least be consistent and postulate another for the



1. The relation between form regulation and functional regulation

A distinction between regulations of form and regulations of function has \'ery commonly been made. As might be expected from his conception of the organism, Driesch ('01) has attempted to draw the line very sharply. But if we adopt the point of view suggested above the distinction becomes apparent rather than real. First of all every regulation in organisms is primarily an energetic process and secondly, it occurs in a certain structure and must affect that structure to a greater or less extent. On the other hand, every change in structure must lead to a regulation of function. Structural and functional regulation are in fact inseparable in organisms. If we go further and interpret structure in terms of the constructive energy, we may say that all regulations are essentially functional, i. e., energetic.

It is sufhciently evident from what has been said, that the flow of energy in the organism is essential for regulation, and that the structure must play a part in determining its character. It should be possible, therefore, to interpret the regulatory processes in terms of the energy current, i. e., metabolism, and the preexisting structure. To refer again to the analogy of the river, both the channel and the current are involved to a greater or less extent in each equilibration in the system. The distinctioa between form regulation and functional regulation is then in part conventional and connected with the separation of morphobgy and physiology from each other, and in part a matter of convenience, since some regulatory processes involve the visible structure to a much greater extent than others. As in the classification of other natural phenomena, we separate for convenience of thought or reference a graded series into a number of (in this case two) different classes.

2. The inducing conditions and the results

As already noted, the first factor in regulation is a change of some sort in the external conditions affecting the system. In the

200 C. M. CHILD

case of a part of an organism this change may be a change in physiological correlation resulting from changes in other parts, however produced. This change, external to the system concerned, produces an internal change of some sort in some part or parts, and this in turn alters the physiological correlation between the components of the system affected. So long as life continues, these correlative changes must result in equilibration in one way or another. The processes of equilibration may be very different in different cases : they may bring about sooner or later a return or approximation to the preexisting condition — according to Driesch, this alone constitutes regulation and only when the preexisting condition was the 'normal' condition. On the other hand, the correlative changes may result in the establishment of, or approach to a condition of equilibrium more or less widely different from the preexisting, and I believe that most, if not all regulations which we usually regard as an approach or return to the preexisting condition actually represent an approach to a new equilibrium, often only slightlj^ different from the old ; it seems at least doubtful whether the organism ever really returns to a preexisting condition in the strict sense.

The new ecjuilibrium may diffei- (quantitatively or qualitatively from the old, or it may even result in the separation of the system into a larger or smaller number of systems, more or less completely isolated from each other. In all these cases, as the rate or character of the metabolic processes are changed, changes in structure as well as in function occur to a greater or less degree. The following suggestions foi- a classification of the regulatory processes are based primarily upon the metabolic processes concerned.

3. A provisional classification of the regulatory processes

a. The two methods of regulation. It is evident that any really analytical classification which is based upon the conception of regulation suggested above must take account, not merely of the visible features, but of the character of the different energetic processes, since regulatioji is, according to this view, essentially a complex of energetic processes in a substratum of a certain


constitution. Such a classification must also be available for both the so-called functional and form regulations, since every regulation probably involves both to some extent.

Driesch's classification of the regulations (Driesch, '01, p. 95 et seq.) is based upon a conception of regulation so widely different from the one developed in this paper that it does not assist us in distinguishing the processes involved. From Driesch's point of view, the physico-chemical processes in regulation are to a large extent of secondary importance and therefore cannot serve as a basis for classification.

At present, however, we are practically unable to attain the proper basis for classification, since our knowledge of the processes involved is incomplete. Nevertheless we can distinguish with more or less certainty the resemblances and differences between different equilibria, and the following suggestions are based upon the character of the equilibria.

We may distinguish two chief type of regulatory processes, first, quantitative equilibrations or compem^alions, and second, qualitative changes in equilibrium or frans/orma/zons. In compensation the rate or intensity of the processes, their continuation in time or their extension in space are concerned: in the transformations their character as energetic processes, i. e., the nature of the chemical reactions and the physical changes. In the compensation the system remains much like that previously existhig, as regards its character and the processes of equilibration are quantitative. In the transformations a new system, qualitatively different from that previously existing, arises as the result of equilibration.

Most regulations, as they occur in nature and experiment, involve both compensations and transformation in various degrees. This is especially true of the regulatory processes which follow the removal of a part. Here some parts of the organism undergo transformation in consequence of altered correlation, while compensatory processes of various kinds are evident, both in the increase in size of the new part and often also in a decrease in old parts.

Moreover our use of these terms will depend upon the particular processes to which we have reference in a given case. The process of compensatory growth, for example, is highly complex in charac

202 C. M. CHILD

ter aad coasists in a variety of both compensatory and transformatory processes, but when an increment in all of these processes occurs, the change is quantitative in character and when it constitutes a process or part of a process of equihbration we are justified in calling it a compensation.

It is of course evident that a classification of regulatory processes must finally become identical with the classification of processes occurring in the organism, for, as I have pointed out above, the regulations are not a peculiar form of organic activity; they represent merely the equilibrations resulting from the existence of physiological correlation between parts. But we shall probably always have occasion to refer to the organism, the system, as a whole undergoing equilibration or, relatively speaking, in equilibrium, consequeatly some means of distinguishing between the different methods of equilibration is useful. This is the chief significance which any classification of the regulations can possess.

h. Regulatory compensation. Several different types of compensation can be distinguished, though they do not of course in most cases exist in nature or even in experiment apart from other processes. The following divisions under this head are suggested :

Incremental compensation: The system shows an increment as compared with that previously existing.

Decremental compensation: the system shows a decrement as compared with that previously existing.

Reversional compensation: an increment or a decremeat in some part of the system, indaced by some external factor is correlatively more or less completely elimiaated and the system approaches its previous condition.

Alterative compensation: an increment or decrement in one part produces change in the opposite direction in another or in others, so that the proportional relations in the system differ from those previously existing.

The first step in all compensations is of course a change in some part {a) induced by some factor external to the system. What particular form of compensation shall occur depends upon the degree of the change in the part and upon the character of the correlation existing between it and other parts (6, c, d-7i). If for


example the part a dominates the other parts, or if the change in a is so great that the correlative factors resulting from it become dominant, then an increment in a may bring about an incremental compensation, a decrement in a a decremental compensation. On the other hand, if the parts b, c, d-n, or certain of them, dominate a, then they may inhibit or reverse the incremental or decremental change in a and reversional compensation results. A.nd finally, alterative compensations occur whenever changes in one part induce correlatively changes in the opposite direction in others.

An incremental compensation occurs when increased metabohsm and growth follow the ingestion of food, a decremental compensation, when decreased activity of a sense organ or a muscle induces a correlative decrease in activity and perhaps atrophy in parts with which it is connected, e. g., the center in the case of the sense organ, the tendon, or even the bone in the case of a muscle. In various temperature regulations in warm-blooded animals we have reversional compensations, and finally, in many cases of regeneration and probably also often in normal development, alterative compensations occur, e. g., when increased growth of one part retards correlatively the growth of another or perhaps induces reduction in it.

The chemical substances which arise in the course of metabolism in certain parts very often produce compensations of various kinds in other parts. A good example is the correlative effect of increase in the carbon dioxide in the blood through the nervous system upon the rate of respiration. The recent work of Bayliss and Starling and others on 'hormones' gives us some insight into various other cases of compensation and other regulatory reactions ; the distribution of nutritive substances in the starving animal and under various other conditions also constitutes compensations of various kinds; the correlative changes in so-called functional structure are in many cases very characteristic compensations.

Incidentally it may, be pointed out that the view of the relation between metabolism and structure suggested above affords a basis for interpretation to a certain extent of the processes of functional hypertrophy and atrophy from disuse.


204 C. M. CHILD

We have seen that structure-formation of some sort, i. e., the accumulation of relatively inactive substances in or about the cell is a characteristic feature of metabolism. The close relation between the syntheses and the oxidation processes has been pointed out repeatedly by Loeb as well as by others. The structural substances, when once formed, play only a relatively small part in further metabolism, provided other more active substances, i. e., nutritive materials, are at hand, and provided the general character of metabolism is not changed. Any condition, e. g., the 'functional stimulus' which leads to increased metabolic activity of the particular kind which constitutes what we call the special function of the cell or part leads, when nutritive material is present, to increased accumulation of the inactive substances and hypertrophy is the result. On the other hand, in the absence of the functional stimulus, or when its frequency or intensity is decreased, the use of nutritive material aad the accumulation of structural substance do not occur or are less rapid, and the result is that below a certain level of functional activity the gradual breaking down of the accumulated substance, which is not immediately connected with the special functional activity of the part, exceeds the constructive processes and decrease in size and atrophy occur. The constructive processes continue only, or very largely, in connection with the functional stimulus and, for the addition of new structure nutritive material must be taken in from without, but this functional activity does not under these conditions, increase proportionally the rate of reentrance of the structural substances into metabolism; in fact, if other more active substances are present in sufficient quantity, the structural substances may be spared to a large extent.

Hypertrophy and atrophy are then the result of two different kinds of processes, the one connected with the specialized function of the part in its relation to other parts, the other to a considerable degree independent of this except in starving animals. In its 'functional activity' the part builds structure, but does not destroy it to so great an extent. The destructive process is largely independent of function and goes on more or less continuously. Whether hypertrophy or atrophy shall occur in a given case depends merely


on whether the one or the other of these processes is the more rapid. Hypertrophy is then in no sense a 'regeneration in excess' ; it is merely a direct result of increased metabolic reactions of the kind which constitute or accompany the 'function' of the part concerned. Nevertheless, it is without doubt a compensation. The occurrence of one series of metabolic reactions determines the occurrence of another series according to chemical laws; the one series furnishes energy, the other forms relatively inactive substances, which persist as structure. Closely related to functional hypertrophy is the growth in size of regenerating parts after their formation: here the 'functional stimulus' is the quantitative factor in the correlative influences from other parts. This factor induces a certain rate or frequency of reaction in the small new part, which leads to rapid accumulation of material, i. e., to hypertrophy (Child, '06a, p. 407). But as the structural substance accumulates, the structure itself constitutes an obstacle to metabolism (Child, '116) the rate of hypertrophy decreases and finally equilibrium is attained.

c. Regulatory transformation. The character of many metabolic reactions is more or less definitely known, but the exact relation of the reactions to the production of a particular kind of visible structure is a much more difficult matter to determine. The visible characteristics of organic structure are by no means adequate criteria of the character of the processes involved in its formation. We are not always justified in concluding from the differences in the visible appearance of structures that the processes concerned in their formation are actually different in nature. Great differences in appearance may arise in the same colloid substance in consequence of differences in aggregate condition or phase. But when we find substances of different constitution in different cells or parts, it is evident that processes of different character were concerned in their formation. Consequently we can often determine that a transformation has occurred by the change in the character of the structural substance. Many features of correlative differentiation, whether in ontogeny in nature or under experimental conditions are undoubtedly transformations, e. g., the formation of a bud from a differentiated cell in

206 C. M. CHILD

the plant, in consequence of the removal of other vegetative tips, the formation of a hydranth from cells of the stem of Tubularia, etc.

The difficulty lies in distinguishing qualitative from quantitative regulations. In living organisms the two are evidently very closely associated, and probably in every regulation which we can observe directly both are concerned. And in the final analysis the question of the relation between quality and quantity in general is involved, though this is scarcely a biological problem.

As regards the further classification of the regulatory transformations, I think that at present the most satisfactory basis for classification is the comparison of the new system with that existing before regulation. The following division of this group of regulations is therefore suggested:

Progressive transformation: the regulatory formation of a system possessing a greater degree of complexity, more varied localization of structure and function and consequently more varied correlation than the system existing before regulation.

Regressive transformation: the regulatory formation of a system of simpler character than the preexisting.

Transgressive transformation: the regulatory formation of a system which cannot be distinguished as more or less complex, but merely as different from the preexisting.

Suh a classification serves merely to suggest the various possibilities. As our knowledge of the processes concerned in the changes of the organic system increases, the basis of classification will change. Without doubt many progressive transformations occur in normal development. The adult organism is certainly a more complex and qualitatively different system from the blastula, and we know that correlative factors have been concerned in the changes in many parts. A regressive transformation occurs when a part undergoes dedifferentiation in consequence of altered correlation, as in various cases where cells which give rise to new parts first lose their old differentiation.

The group of transgressive transformations possesses little more than a conventional significance, since it is based upon difference from the normal which is essentially merely the usual


Very probably various 'sports' and mutations can be placed under this head, perhaps also certain of the neoplasms, more specifically the malignant tumors, though this is by no means certain. But whatever the categories which we may establish for the different regulatory processes, the important point is that we should at least make the attempt to find a physico-chemical basis for our analysis. If we do this we cannot separate structure and function since both are merely different aspects of the same process-complex and are dependent upon and determine each other. Doubtless we shall still find it convenient to speak of form regulation as distinguished from functional regulation, but we must remember that the distinction is not a real one and that every regulation in the organism is undoubtedly a regulation of both form and function, of both structure and reactions. Furthermore, we must regard our experiments on regulation as means of analyzing the factors of the process. With the proper care in experiment we can do much toward determining the nature and action of various correlative factors in regulation, and every step in this direction is a step in advance in our knowledge of the system which constitutes the organism,


1. Restitution or reconstitution f

When an organism 'restores' a missing part or in general when a part of an organism forms a whole, the process seems at first glance to be so obviously a restoration in which something removed is replaced, that the term 'restitution' has found very general favor. Although I have used this term to a large extent, it has always seemed inadequate, for the reason that the process is not simply one of restoration but something more. There is no case of so-called restitution known in which the changes following the removal of a part are limited to the formation of a new similar part. In every case changes of one kind or another, quantitative or qualitative, or both, occur in other parts, sometimes limited chiefly to parts adjoining the part removed, sometimes extending throughout the system. The removal of a part

208 C. M. CHILD

brings about not simply its restitution, but an equilibration extending more or less widely, and strictly speaking, probably throughout the system. The system reconstitutes itself and the whole formed is different from the original quantitatively or qualitatively^ The whole process is a complex of equilibrations, of compensations and transformations, resulting in that which we call a whole, but no two wholes are alike.

If for example, we consider this process in a piece of the Planarian body, we find that it differs in rate and character according to size of the piece, region of the body from which it is taken, temperature and other factors which influence metabolism. The animal formed as the result resembles the original in its general shape and activity, but it is far from being identical with it. It is usually smaller than the original, the pharynx may be in quite different position in the body, and the arrangement, number and form of the intestinal branches differs more or less widely, according to conditions. Moreover, under various conditions, various degrees and kinds of incompleteness appear. Some pieces develop only a single eye, or the eyes are partially fused or otherwise different from those in the original animal, some pieces develop no pharynx and no posterior end, others no head or an imperfect' one, some develop the postpharyngeal intestinal branches more rapidly and more completely, others the prepharyngeal branches, some produce a larger, others a smaller head, some show more 'regeneration,' others more 'redifferentiation' and so on. If we place the different sorts of wholes under closely similar conditions and give them food they become more or less like each other because these conditions bring about further regulations but these regulations do not properly belong to the regulatory process which resulted from the isolation of the part, but are independent of it and are such as were occurring in the original animal during its life. It is probably not too much to say that no two pieces of the Planarian body attain the same condition in the process of regulation. When we say that because some or all of them produce wholes they are all potentially alike, we are simplj- assuming that all wholes must be alike, which is obviously untrue. As each piece is different at the start from the others, so it attains a di^erent


result. Driesch's assumption of 'equipotentiality' of different parts is shown by the facts themselves to be incorrect, and I believe that his 'harmonious-equipotential systems' do not exist in nature, as systems capable of development but only as abstractions of the human mind.

For these reasons the term 'restitution' seems to me to carry with it implications which, when we analyze them, we cannot, in the light of the facts, accept. The piece does most certainly not restore what it lost : it reconstitutes itself into something more or less widely different from that of which it formed a part, and this something often possesses visible structural characteristics which we have come to regard as characteristic of a whole. Seeing these resemblances, we abstract from the differences and say that it is the same as that of which it formed a part.

Closer examination shows us that even visibly it is not the same, but different; moreover, visible characteristics are not the sole criterion of resemblance and difference in organisms. The processes occurring are just as characteristic as the visible structure. I have shown elsewhere (Child, 'lib) that in Planaria the process of form regulation results in a rejuvenation, the pieces after undergoing regulation are physiologically younger than the animals from which they were taken, and the degree of rejuvenation varies with the degree of reconstitutional change. Manifestly then these pieces are not the same after regulation as the wholes of which they formed parts. To say that they are is simply to deny the facts as they stand before us.

In many cases, moreover, the pieces do not produce anything that can be called a whole. Pieces of Planaria may produce double heads or double tails, tailless heads or headless forms. For those who with Driesch regard the formation of a 'whole' as the uniform result of so-called restitution, these cases are difficult to interpret. The process of regulation has apparently followed the wrong track, it has gone astray and so failed of the correct result. But when we take the position that the part, when isolated, undergoes a reconstitution, which differs in its results according to the existing conditions, internal and external, we see that these 'abnormalities' differ from 'normal' results simply because different

210 C. M. CHILD

conditions existed at the beginning or elsewhere in the course of the regulation, and in many cases we can determine what those conditions are. For example, pieces of Planaria which give rise to 'wholes' at a certain temperature and under certain other conditions, can be made to produce headless forms or tailless heads, according to the region of the body from which they are taken, by subjecting them to lower temperatures, by starving the animals before beginning the experiment, by placing the pieces in dilute alcohol or ether, etc. Nothing has gone astray in t'lese cases, there is no error, the same laws have been followed as when 'wholes' are produced, different conditions simply lead to different results. Elsewhere (Child, '10c) I have attempted to analyze some of the conditions which bring about so-called heteromorphosis in Tubularia and other forms, and have shown that they are similar in character to those which bring about asexual reproduction in nature.

There are many cases in which the occurrence of reconstitution as opposed to restitution is so obvious that there can be no questioning it. A piece from the body of Hydra, for example, does not restore the missing parts, but reconstitutes itself into an organism, smaller, simpler, possessing fewer tentacles and undoubtedly physiologically younger than the original animal. In Clavellina also, as Driesch himself has shown (Driesch, '02), the isolated branchial region or a part of it does not replace the missing parts, but undergoes a process of reconstitution. Tn these cases the physiological effect upon the parts remaining of the removal of certain parts is so great that these parts do not retain their original structure, and a dedifferentiation and redifferentiatioQ occurs. But the effects so apparent in these cases are simply more extreme than in cases where only a small part is removed. Przibram ('07) has called attention to a number of cases which show verj" clearly that the removal of a part results, not in the restoration of the original, but in the establishment of a new equilibrium, differing more or less widely from that.

It is obvious that the process of reconstitution is an equilibration and just as obvious that it leads to different results under different internal and external conditions. As different rivers differ


from each other, so may the results of reconstitution, even in different pieces of the same individual, differ from each other. Moreover, as a certain amount of water does not, except under certain conditions, form a river, so does a piece of an organism reconstitute itself to a whole only under certain conditions.

2. The initiating factor in reconstitution

As Driesch has pointed out (Driesch, '01), it is evident that reconstitution occurs as the result of the absence CNichtmehrvorhandensein') of something. What is this something? Is it the structure, the form, or is it activity? In plants it is possible to bring about reconstitution, i. e., the formation of new buds, roots, etc., by inhibiting the metabohc activity of the existing growing regions without their removal, e. g., by enclosing them in plaster or in an atmosphere of hydrogen, or even by applying anesthetics locally between them and the regions concerned. In another paper I have considered numerous cases of this sort and have discussed their significance at length (Child, 'Ua). These facts show very clearly that it is not the form or structure which is involved but a process, an activity, whose effect is transmitted in some way from one part to another. If we stop the metabolism of the one part for a time the effect on the other is the same as if the first part had been removed. These facts alone should be sufficient to prevent us from regarding form regulation as a process distinct from functional regulation. It is the absence of the effect of certain processes in a certain part or in certain parts, which initiates reconstitution. In short, it is the absence or decrease below a certain point of certain physiological correlative factors which were previously present, that initiates reconstitution. In the absence or decreased effectiveness of these correlative factors, the remaining parts, still being subjected to other correlative factors, which may themselves gradually change in consequence of the removal or decreased activity of the part, react in a manner different from their previous reactions, simply because they are under different physiological conditions. Reconstitution is then initiated by a change in physiological correlation. Recently Driesch ('09) has

212 C. M. CHILD

expressed himself in somewhat similar terms, but he finds nevertheless, as already noted above, that the 'Individuahtat der Zuordnung' between agent and effect cannot be accounted for on a physico-chemical basis and therefore regards it as a new 'proof of the autonomy of vital processes.

3. The process of equilibration in reconstitution

The process of equilibration in reconstitution differs in many details in different cases, but it possesses certain more or less characteristic features, and it is desired to call attention briefly to some of these. The change in ph37siological correlation is the internal factor which has disturbed the preexisting condition, whatever that may have been. This change may or may not lead to equilibration of the living organism. If the change be great, if the other parts possess but little capacity for altering their reactions, it may lead to death. On the other hand, it may lead to reconstitution in various ways according to conditions.

Let us consider first the case where a part is removed and is formed again without any great changes in other parts, e. g., the 'regeneration' of the posterior end of Planaria.

In the absence of the correlative factors which originated in the part removed (a), certain regions (6) of the remaining parts {b,c

d n), which were before prevented by these correlative factors

from reacting as their own constitution and the correlative factors from other parts would determine, now begin to react in this manner. In the region adjoining the part removed, i. e., in the cells b,

the correlative factors originating in the parts cd n are more or

less siniilar in their effect to those which affected the part removed (a). So far as they are and remain similar, and so far as the constitution of b permits, this region will be forced by the correlative factors to react more or less in the manner of a, which is no longer present, and b will replace a more or less completely and more or less rapidly, according to conditions in the particular case^ If

' The formation of a new head in Planaria or a new hydranth in Tubularia is a somewhat different process from the formation of the proximal or posterior end. In these forms the anterior or distal region is physiologically dominant over parts


the cells of the region b adjoining the part removed, be chiefly affected correlatively by the removal of a or if they react much more rapidly than others further away, the process of formation of a will be a 'regeneration.' But if parts further away from a are also affected to a considerable extent by the change and are capable of reacting as rapidly or almost as rapidly as b, then they may also take part in the process of replacing a, which then takes on more or less completely the character of a 'redifferentiation.' In some cases we can determine experimentally whether a part shall be formed chiefly by regeneration of redifferentiation. In Planaria, for example, the amount of regeneration, as opposed to redifferentiation, in the formation of a new posterior end increases with increasing distance of the cut surface from the old posterior end. The farther the level of the cut from the old posterior end, the more completely is the development of the new part confined to cells near the cut, and vice versa. The cells near the cut are those which are most affected by the removal of the part a. Even when this part does not develop anew, they react by healing the wound. That is to say, they change their reactions most rapidly of all cells, they lose their old specification, they become capable of a more rapid metabolism (Child, 'lib) and being subjected to the correlative factors of the parts cd n, they begin to develop into something more or less like a in advance of other parts. The processes in these cells establish certain correlative factors which determine that the cells farther away from the cut shall remain as they are or take other forms of reaction.

But if we decrease the rate of metabolism in Planaria by extreme starvation or by the use of anesthetics, then parts which under the

posterior or proximal to it and controls their development directly or indirectly. Briefly stated, the regulatory formation of a dominant part is a reconstitution resulting primarily from isolation, while for the formation of a subordinate part correlation with other parts of the original system is necessary. For example, a piece of the tubularian stem may reconstitute itself into a hydranth without any other parts (Child, '07a, b, c), or a piece of Planaria into a head without other parts but a tubularian stem or stolon or a planarian tail is never formed except in connection with a more distal or anterior region of the original organism. The question of the dominance and subordination of parts and its significance will be discussed more fully elsewhere.

214 C. M. CHILD

usual conditions are formed chiefly by regeneration, e. g., the head, may be formed largely by redifferentiation (Child, '10a). T.iis means simply that the cells near the cut do not react so rapidly s under the usual conditions, so that other cells further away have ime to change their reactions and take part ia the process, while ordinarily they would be prevented from doing this by the correlative factors arising from the activity' in the cells near the cut. These instances are merely special cases under the general rule that the less rapidly the missing part is replaced, the more extensive are the changes in the remaining parts, so far as their constitution permits change.

Regeneration and redifferentiation in their extreme forms represent the extreme terms of a graded series, of which all terms are essentially of the same physiological character, i. e., all consist in a change in reaction in consequence of a change in physiological correlation. The designations 'regeneration' and 'redifferentiation' serve merely as convenient descriptions of the visible phenomena.

Where all the cells of the remaining parts are so sensitive to the absence of the correlative factors originating in the part removed that they cannot maintain themselves after its removal, the old structure of all the remaining parts may disappear to a greater or less extent, i. e., a 'dedifferentiation' occurs, as, for example in the isolated pieces of the branchial region of Clavellina. In the different cells of the mass the metabolic processes become less specified and in this respect it approaches the 'embryonic' condition, and the correlative factors in the mass approach those existing in the embryo. But during this process some of the cells have been subjected to correlative factors more or less similar to those to which the part removed was subjected at some stage of development, consequently these become in some degree the physiological representatives of that part. In short the system becomes physiologically a whole, but in consequence of the rapid dediffereatiation of the old parts it corresponds to a whole in a relatively early stage of development. From this condition renewed differentiation as a whole results necessarily from continued metaboUsm,


i. e., continued life. The new whole is, however, different from the old in size, physiological conditions, number and proportion of various parts, etc.

In the process of regeneration in the stricter sense the new part is usually at first smaU and increases rapidly in size. I believe that this growth in size is essentially similar to the functional hypertrophy of organs. The part which was removed possesses a certain size in relation to other parts, because its size was determined chiefly by correlative factors. Just so far as the new developing part is subjected to similar correlative factors, it will tend to attain the same size as the part removed. Consequently it does not always attain the same size with respect to other parts. In Planaria the relative size of the new head differs according to the region of the body from which the piece is taken, to the nutritive condition and various other factors. The process of reconstitution ceases when a certain stage, differing under different conditions, is attained. This stage represents an equilibrium of physiological correlation, i. e., of interaction between the parts; it is primarily a dynamic equilibrium, a proportionalit}' of processes, not of form. We can alter this condition of equilibrium experimentally by food, by starvation, by temperature, and in short by all factors which affect the processes.

In various papers Driesch has distinguished a number of different forms of reconstitution (restitution). His distmctions are based primarily upon differences in the visible phenomena of development or dedifferentiation and for him the chief interest lies in the recognition of the different forms, rather than in the attempt to determine how they differ from each other physiologically, since from his point of view the physiological factors are in many cases only 'means' which the enfelechy employs. It is impossible to consider here these various forms of reconstitution, and since my point of view is so widely different from that of Driesch such a consideration would show merely that his basis of distinction could not be accepted for purposes of a physiological analysis. While it is convenient to distinguish different forms or methods of reconstitution, I believe that it is much more important to resolve the phenomena into processes.

216 C. M. CHILD

Ij.. The complexity of reconstitution

The process of reconstitution is not a simple process which cannot be analyzed, but rather an exceedingly complex one; in fact its complexity is of the same order and character as that of development. It consists of a series of compensations and transformations in different parts of the system. It is only when we take into account the complexity of physiological correlation between parts and the almost infinite possibility of change and variety in this correlation that we have any hope of gaining an insight into the complex series of events. The specificity of correlation and reaction does not, as Driesch apparently beheves' (Driesch, '09) constitute a physico-chemically insoluble problem except when we follow Driesch in ignoring the energy current as an equilibrating factor in the organism and as the efficient factor in construction of the visible and tangible characteristics. The energy current performs its work under specific conditions in each case and leads to a specific result. As soon as a specific condition arises in any part of the system, from whatever cause, it determines other specific conditions in at least certain other parts. From the experiments on Planaria it is perfectly apparent that the cells at every level of the body posterior to the ganglia, are capable under certain conditions of developing into a head, but under the usual conditions they are prevented from doing this because the correlative factors arising from the presence, ^. e., the activities, of a head determine their activities in another direction. As soon as the old head is eliminated from the system, those cells, which in consequence of their past correlation are most similar to it, begin at once to form a new head, provided the piece is not too small and as soon as this occurs it determines correlatively a variety of reactions in other cells. The same may be said of the reconstistitution of any part. The place where a particular part shall arise is determined by constitutional and correlative factors in the existing system — so far of course as external factors are not concerned — and as soon as one such place is determined, it determines others and so on. 'Ihus any case of reconstitution consists of a series of regulations, each of which determines others. This


is shown very clearly by the fact that isolation of a part from certain others during the course of reconstitution may alter the course of the process in the part, according to the degree and character of the isolation, i. e., according to the correlative factors which are eliminated. By means of experiments of this kind it is possible even now to determine the action of various correlative factors in different stages of the process of reconstitution.

5. The limits of reconstitution

In every case the reconstitutional processes are limited and determined by existing conditions as the river is limited and defined by its banks and channel which its own activity has constructed in the environment through which it flows. It is not true without qualification that any part of certain organisms is capable of giving rise to any part. Driesch's often repeated statement to this effect requires modification and limitation. The part is at most only provisionally capable of giving rise to any part; in other words, only when it constitutes a component of a system possessing certain characteristics, i. e., only when it is subjected to or isolated from certain correlative factors. This is apparent from every recorded series of observations on reconstitution, except perhaps the most superficial. In some cases the system may consist of a single cell, in others of a large number of pells, but the fact remains the same. The power of reconstitution is limited, not unlimited. As I hope to show elsewhere for Planaria, and as I have shown for Tubularia (Child, '07a, '07b, '07c), the investigation of these limitations is of the greatest importance in throwing light upon the nature of the reconstitutional processes. When we find that the removal of a certain part, or even a certain amount of material, determines a different result from the removal of another part or a larger or smaller amount of material, ■ we are forced to the conclusion that the part or the material removed has some connection with the character, place or other factors in the result, and furthermore, when w^e find that inhibition of the metabolic processes or certain of them in the part or the material is as effective in certain

218 C. M. CHILD

respects as the removal of the part or the material, we have attained a basis for investigation and analysis which is proof against such assumptions as those which Driesch has made, e. g., concerning the nature of the 'harmonious-equipotential system.' For Driesch the limitations of the reconstitutional processes appear to be of secondary importance, but I believe that any one who will investigate and analyze these limitations at all thoroughly will find that they are not only essential features of the regulatory processes, but that they afford us one of the best means of gaining some insight into their nature. As water does Qot constitute a river, except under certain limiting conditions, so certain substances or processes do not constitute an organism or even life except under certain limitations. The water contains the potentialities for giving rise to any kiad of a river, as well as other specific 'machines,' but none of these exist until the specific hmitations are present. The case is essentially similar as regards the organism. The specific 'machine' exists only so far as the limitations exist. And the investigation of the limitations of reconstitution affords at present one of the best methods, if not the best, for demonstrating this to be a fact.


In another paper (Child, '11a) I have discussed at length the significance of physiological isolation of parts as a factor in reproduction. I have shown that certain degrees and kinds of physiological isolation of parts may arise as the result, first, of an increase in size; second, of decrease in correlative control or physiological dominance of a part in consequence of decreased activity in it; third, of decreased conductivity^ or transmissibility of correlative processes, agents or conditions; fourth, of decreased receptivity, sensitiveness or irritability of certain parts to the correlative factors originating in other parts. Furthermore, we know from experiment, as I have shown, that in a considerable number of cases physiological isolation of parts serves as well as physical isolation by section to bring about reconstitution; and if it were possible to perform the experiment, it is practically certain that we should find the same to be true for many other cases.


In the paper just referred to I have also attempted to show that at least a great variety of natural and experimental forms of reproduction, reduplication of parts, etc. are essentially processes of reconstitution following physiological isolation of parts. The chief difference between them and the cases of reconstitution following experimental section is, first, that the isolation of the part or parts is brought about within the organism physiologic ally and not by the crude method of cutting the organism into pieces; and second, that this isolation is usually partial at first and differs in degree and kind in different cases.

And finally, I have called attention to certain evidence in support of the view that the formation of sex cells and the development of organisms from them are processes not fundamentally different from other forms of reproduction, i. e., that the sex cells are first physiologically parts of the organism like other organs, and that the development of a new organism from them is initiated by changes similar in character to those which occur in other parts capable of reconstitution, when they are physiologically or physically isolated (Child, '10b, '10c).

The evidence bearing upon the first point is briefly as follows; first, the sex cells always arise in, or attain by migration particular regions of the body in a particular organism, therefore, their physiological correlation with other parts cannot be purely nutritive in character, for if it were, there is no reason why they should not take the most various positions in the same species. Second, they undergo characteristic differentiations during the life of the individual, as do other organs and these differentiations begin at a certain stage of development of the organism, i. e., at or near the end of the period of vegetative growth. This cannot be accounted for by quantitative differences in nutrition at different stages, because the growth of the primitive germ cells in earlier stages often requires very large amounts of nutritive material. If this development is predetermined, then physiological correlation between the germ cells and other parts must have existed at some earlier stage, or else we are forced to a hypothesis of preestablished harmony, which amounts to some form of vitalism.

Moreover, in organisms, which show both asexual and sexual

220 C. M. CHILD

reproduction in the same individual, the asexual reproduction occurs earlier in the life cycle than the sexual, and in organisms which produce naturally both parthenogenetic and non-parthenogenetic eggs, the parthenogenetic eggs appear earlier than the non-parthenogenetic. In both these cases the earlier product is usually capable of a greater degree of regulation when it is isolated from the parent body, than the later; in other words, both the non-parthenogenetic egg and the sperm appear from their behavior to be more Highly specified or differentiated than the asexual or parthenogenetic reproductive elements. There is then considerable evidence in support of the view that the history of the germ cells, like that of other organs, is in part the result of physiological correlation.

As regards the 'stimulus to development,' I have shown by experiment (Child, 'lib) that in Planaria the process of reconstitutution after physical isolation as well as extreme starvation followed by feeding, accomplish rejuvenation and that it is highly probable that various other factors bring about similar changes. And finally, I have considered the facts which indicate that the process of fertilization and the conditions inducing artificial parthenogenesis produce changes in the egg similar in character to the rejuvenation occurring in the other cases.

From this point of view, the stimulus to development of the egg is essentially a process or the beginning of a process of reconstitution and so is similar in its physiological effect to the factors initiating the various processes of asexual reproduction. Experimental reconstitution following section is then merely a special case of reproduction occurring under certain conditions, or we may say just as correctly that each form of reproduction in nature or experiment is a special case of reconstitution occurring under certain special conditions.

If this view be correct, then the fundamental problems of development and heredity are before us in every case where a physically or physiologically isolated part of an organism produces a new organism, just as truly as they are in sexual reproduction. In fact sexual reproduction constitutes the most complex case of all, but I am convinced that a recognition of its essential similarity to the


processes following experimental section and the physiological solation of parts is of the greatest significance for our conception and solution of the problems of inheritance and development.


It is sufficiently evident from what has been said that I consider the phenomena of regulation in organisms as congtitating the essential chacteristic of life as a coatinuing process. I agree with Jennings that the problem of regulation is the fundamental problem of life. All of our experimental investigations on living organisms are directly concerned with the problem of regulation in one way or another. Tn fact there are only two possible methods of investigating and analyzing the phenomena of life : one is concerned with regulation in the living organism, the other with the observation and analysis of the results of stopping the life-processes at this or that particular point, under these or those particular conditions. In the one case we observe and control the process in its action, in the other we seek to determine the effects of its past action. As we can watch the river at work and investigate the processes of equilibration resulting from alteration of its flow in one way or another, so we can investigate the living organism. And as we can stop the flow of the stream or divert it into other channels and determine something of what it has done along its course up to a certain time by examination of its channel, so from the dead organism, we can determine something of its past activity.

But the conclusions drawn from the examination of the channel of the 'dead' river are only fragmentary at best. Only by observing and controlling the river in action is it possible to acquire any adequate conception of what it really is. And so, I believe, with regard to the organism : the living organism will teach us more than the dead one though we must work with both. And when we work with the living organism we come at once face to face with the problem of equilibration, of regulation. And finally, I believe that the further our knowledge of the processes of equilibration, in the organism advances, the greater will be the difficulty of finding an adequate foundation in biolog}'^ for vitalistic or dualistic hypotheses.




Child, C. M. 1906a Contributions toward a theory of regulation. I. The significance of the different methods of regulation in Turbellaria. Arch, f. Entwickelungsmech. Bd. 20, H. 3,

19066 Some considerations regarding so-called formative substances. Biol. Bull. vol. 11, no. 4 .

1907a An analysis of form regulation in Tubularia. I. Stolon formation and polarity. Arch, f . Entwickelungsmech. Bd. 23, H. 3. 1907b An analysis etc. IV. Regional and polar differences in the time of hydranth-formation as a special case of regulation in a complex system. Arch. f. Entwickelungsmech. Bd. 24, H. 1. 1907c An analysis etc. V. Regulation in short pieces. Arch f. Entwickelungsmech. Bd. 24. H. 2.

1908a The physiological basis of restitution of lost parts. Jour. Exp. Zool., vol. 5, no. 4.

1908b Driesch's harmonic-equipotential systems in form regulation. Biol. Centralbl. Bd. 28, nos. 18 und 19.

1910a Analysis of form regulation with the aid of anesthetics. Biol. Bull., vol. 18, no. 4.

1911b A study of senescence and rejuvenescence, based on experiments with Planaria. Arch. f. Entwickelungsmech. Bd. 31, H. 4. 1911a Die physiologische Isolation von Teilen des Organismus, Vortr. u. Aufs. ti. Entwickelungsmech. H. 11.

Driesch, H. 1901 Die organischen Regulationen. Leipzig.

1902 tJber ein neues harmonisch-aquipotentielles System und iiber solche Systeme iiberhaupt. Studien uber das Regulationsvermogen der Organismen. 6. Die Restitutionen der Clavellina lepadiformis. Arch, f. Entwickelungsmech. Bd. 14, H. 1 u. 2.

1903 Die 'Seele' als elementarer Naturfaktor. Leipzig.

1905 Die Entwickelungsphysiologie von 1902-1905. Ergebnisse der Anat u. Entwickelungsgesch. Bd. 14 (1904).

1908 The science and philosophy of the organism. Vol.1. London.

1909 Der Restitutionsreiz. Vortr. u. Aufs. ii. Entwickelungsmech. H vii.

Holmes, S. J. 1904 The problem of form regulation. Arch. f. Entwickelungsmech. Bd. 17, H. 2 u. 3.

1907 Regeneration as functional adjustment. Jour. Exp. Zool. vol. 4, no. 3.

Jennings, H. S. 1906 Behavior of the lower organisms. New York.

KoRSCHELT, E. 1907 Regeneration und Transplantation. Jena.

Morgan, T. H. 1907 Regeneration. Ubersetzt von M. Moszkowski. Leipzig.

Przibram, H. 07 Equilibrium of animal form. Jour. Exp. Zool., vol. 5, no. 2. 1909. Experimental-Zoologie. 2. Regeneration. Leipzig u. Wien.

RiGNANO, E. 1907. Die funktionelle Anpassung und Paulys psychophysische Teleologie. Riv. di Scienza, vol. 2.

Roux, W. 1895. Gesammelte Abhandlungen fiber Entwickelungsmechanik der Organismen. Bd. 1, u. 11. Leipzig.



Fruyn the Sheffield Biological Laboratory, Yale University


Leeuwenhoek in 1677 described^ some little animals longer than an oval" which he had discovered two years previously, and there is some reason to believe that this is the first published record of an organism belonging to the genus Paramaecium. The name Paramaecium, however, was first employed by HilP to designate certain small organisms which were more or less oblong, in contrast to others which were round or decidedly vermiform, and either the present species aurelia or caudatum is probably the animal which he designated as 'Paramaecium species 3.'

Although Hill was the first to attempt to apply scientific names to microscopic animals, it remained for O. F. Miiller to give a general classification of these forms, and to apply the Linnean nomenclature. He began this work on the infusoria as a section of a treatise entitled, Vermium terrestrium et fluviatilium Historia, which appeared in two volumes in 1773. Unfortunately he did not live to see the publication of his special work, Animalcula Infusoria fluviatilia et marina, 1786, which was edited by his friend, O. Fabricius. Miiller described a Paramaecium and applied the specific name aurelia in the former of these works. In the latter work he described and figured^ Paramaecium aurelia

^Philosophical transactions, London, 11, 133. 1677. ^History of animals, 3, 1751. 'Plate 12, figs. 1-14.




together with several other forms, which at the present time are assigned to other genera. Miiller's description is as follows*: —

Paramaecium. Vermis inconspicuus, simplex, pellucidus, membranaceus, oblongus.

Paramaecium aurelia. Paramaecium compressum, versus antica plicatum, postice acutum.

Thus the organism described by Baker^ as Animalcules in pepper water, first sort/' by Joblot^ as Chausson, by Ellis^ as Volvox terebella, etc., received the name which, in spite of various vicissitudes, has come down to the present time.

The next great student of the lower organisms, C. G. Ehrenberg, in the first two of his treatises,* described several species of Paramaecium, and one of these is Paramaecium aurelia. In his third treatise^ he described still another species which he named Paramaecium caudatum.^" Five years later, in 1838, Ehrenberg brought out his monumental monograph. Die Infusionsthierchen als vollkommene Organismen, and in this work he described these two species as follows :^^

Paramecium Aurelia, Pantoffelthierchen.

P. corpore cylindrico, subclavato, antica parte pauUo tenuiore, plica longitudinali obliqua in os multum recedens exeunte, utrinque obtuso.

Paramecium caudatum, geschwanztes Pantoffelthierchen.

P. corpore fusiformi, antica parte obtusiore, postica magis attenuata.

Thus Ehrenberg described, on the basis of shape and size, the two common forms of colorless paramaecia which appear in

  • Page 86.
  • The microscope made easy, London, 1742. 3rd ed., 1744, p. 72, PI. 7, fig. 1.

Observat. fait, avec le microscope, Paris, 1754.

Observations on a particular manner of increase in the Animalcula of vegetable infusions, etc. Phil. Trans., London, 1769.

^Abhandl. der Akademie d. Wissensch. zu Berlin, 1830, 1831.

"Ibid, 1833.

lOEhrenberg notes that Herrmann (Naturforscher, 1784) applied the name caudatum to a form which was probably a species of Amphileptus; also Schrank (Fauna boica, 1803) used the same name.

iiPp. 350-352. PI. 39, figs. 6, 7.


modern systematic works as P. aurelia O. F. M. and P. caudatum Ehrbg.

Dujardin, in 1841, in his treatise on the Infusoria '^ recognized but two species of Paramaecium as follows :

Paramecie Aurelie. — Paramecium aurelia.

Corps ovale oblong, arrondi ou obtus aux deux extremit^s, plus large en arriere. — Long de 0, 18 a 0, 25.

Paramecie a queue. — Paramecium caudatum.

Corps fusiforme, obtus ou arrondi en avant, aminci en arriere.— Long de 0, 22.

His figures of the two species show clearly the characteristic form which he considered diagnostic.

Various investigators, including Stein, and Claparede and Lachmann, questioned the justification of considering these two forms as distinct species, basing their opinions, as had Ehrenberg and Dujardin, solely on external characters, and they united these two forms under one species, and applied Miiller's original name, P. aurelia. This union of aurelia and caudatum into one species was accepted by all the subsequent students of Paramaecium, e.g., Balbiani, Blitschli, Engelmann, Gruber and Kolliker and consequently all the early literature on the conjugation of this infusorian, refers to the organism as P. aurelia, although it had but a single micronucleus.

Maupas, in 1883, in his studies on the ciliates,i^ wrote:- —

Tous les auteurs jusqu'ici ont decrit Paramecium aurelia comme ne poss^dant jamais qu'un nucleole d'assez grande taille et mesurant de 0mm,005 h. 0mm,008. C'est en effet la forme que Von rencontre la plus frequemment. Mais j'ai observe aussi de nombreux individus pourvus de deux nucl^oles plus petit s et de structure diff^rente de la piecedente. lis etaient de forme spherique et composes d'un corpuscule central opaque vivement color^ par les teintures et ne mesurant que 0mm,003; enveloppe d'une couche corticale mesurant en diametre 0mm,005, claire et ne se colorant pas.

i^Histoire naturelle des Zoophytes. Infusoires, etc. Paris, 1841. Pp. 481-483, PI. 8, figs. 5, 6, 7.

"Contributions a I'etude morphologique et anatomique des Infusoires cilies, Arch, de zool. exp. et gen., (2), I, 1883, p. 660.


Thus Maupas tacitly accepted the current view that there was one large species of Paramaecium, but observed, for the first time, that certain paramaecia have a different nuclear apparatus from that previously described. This author, however, in 1888, stated that in his earlier work he, as all his immediate predecessors, had confused two species, and he wrote^^ as follows :

Ces deux formes de micronucleus constituent le caractere distinctif le plus important entre les deux esp^ces de Parameeies. La premiere forme appartient toujours et uniquement au P. caudatum, la seconde, egalement toujours et uniquement, au P. aurelia.

Pour Ehrenberg et Dujardin, P. caudatum se distingue par un corps allonge, fusiforme, obtus en avant, aminci en arriere: P. aurelia par un corps plus large, presque ovale, obtus aux deux extremites. Ces differences de contour general, tout en ^tant reelles, ne sont pas absolument rigoureuses; car, si on ne trouve jamais de Paramecie k un seul micronucleus affectant la formed trapue obtuse, il n'est pas tr^s rare d'en rencontrer a deux micronucleus, ayant pris la forme allongee k queue. Dans ce dernier cas, il est impossible de savoir a quelle esp^ce on a affaire, sans une preparation permettant de voir les micronucleus. Ce charact^re distinctif, bas^ sur le contour general, n'a done qu'une valeur relative. II est cependant bon d'en tenir compte; car lorsqu'on s'est exerce k bien distinguer les deux especes, il suffit presque toujours et trompe rarement.

Le P. caudatum paralt avoir une taille un peu plus grande que celle du P. aurelia. Ainsi, j'ai mesur^ des premiers depuis 120 jusqu'a 325 /x, tandis que les seconds ont varie seulement entre 70 et 290 /i. En outre, P. caudatum se conjugue avec une taille variant entre 125 k 220 n, et P. aurelia entre 75 a 145 fi. Pendant la conjugaison, le deroulement rubanaire, preparant la fragmentation du nucleus, s'effectue chez le P. aurelia, des le stade D, tandis que chez le P. caudatum il ne commence que vers le milieu du stade G. Chez cette derniere espece, le nucleus mixte de copulation donne naissance finalem'ent a huit corpuscules, chez P. aurelia il n'en produit que quatre; il en resulte que chez celle-ci I'^tat normal se trouve r^tabh des la premiere bipartition qui suit la conjugaison, et chez P. caudatum seulement apres la seconde.

Toutes ces differences anatomiques et physiologiques me paraissent plus que suffisantes pour justifier la distinction des deux especes. II

"Sur la multiplication des Infusoires cilies, Arch, de zool. exp. et gen., (2), 4, 1888, pp. 231-235.


est fort possible que Claparede et Lachmann aieiit eu raison, en considerant la forme caudatwn comme plus typique que la forme aurelia. Si, en effet, on examine avec soin les dessins de O. — F. Miiller, on penche k croire que le vieux micrographe a vu et figure la premiere seulement. En se conformant strictement au principe de la loi de priorite, ce serait done le nom aurelia, donne par Miiller, qui devrait etre conserve a la forme fuselee. Mais, d'un autre cote, Ehrenberg et Dujardin ont distingue ce type et I'ont decomme caudatum. Si nous lui conservons la vieille denomination aurelia, il devient impossible de transmettre le qualificatif caudatum a la forme qui, le plus souvent, est obtuse k ses deux extremites. II faudrait alors creer un nouveau nom. Je crois plus simple de conserver les denominations d'Ehrenberg.

Since 1889, when Maupas^^ and Hertwig^, in studies on conjugation added further evidence for the distinction of the two forms, they have been generally accepted as 'good' species. Calkins, however, again raised the question in 1906: I personally believe that the slight differences that distinguish the two types of Paramecium are not of specific value, and hold that P. caudatum should be regarded as a mere variant of P. aurelia. "^^ He based this view chiefly on the following observations. One of a pair of ex-conjugants of P. caudatum, which he was studying by his well-known accurate culture methods, reorganized as P. caudatum and the other as P. aurelia, i.e., the latter had two small micrpnuclei, instead of one, and remained in this condition for about forty-five generations in pedigree culture, and then reverted to the caudatum type with one large micronucleus. While the aurelia phase existed, the rate of«division was comparatively slow, and when the caudatum phase was reassumed the rate of division immediately increased considerably. Calkins also considered the relative size of the two forms, and the conjugation phenomena as described by Maupas and Hertwig, and concluded that these are not of such a character as to warrant their being considered diagnostic.

i*Le rajeunissement karyogamique chez les cilies, Arch, de zool. exp. et gen., (2), 7, 1889.

i^Ueber die Konjugation der Infusorien, Abh. kgl. bayr. Akad. d. Wiss. Miinchen, 2, CI. 17, 1889.

"Paramecium aurelia and Paramecium caudatum. Studies by the pupils of W. T. Sedgwick, 1906.


Jennings, in his studies on heredity in Paramaecium/^ showed that he could readily isolate a considerable number of pure lines from a wild culture, and that these pure lines breed true, i.e., there exist inherent hereditary differences in size, persisting when all other conditions remain the same. These different lines fall usually into two main groups, one group having a mean length greater than 170^, and the other having a mean length less than 140ai. But he was able finally to isolate a line intermediate in size, and thus to bridge over the gap. As Jennings points out, even if it were not possible to isolate a strain of intermediate size between the two large groups, this would not give a basis for distinguishing two species. However, he states: I may be permitted to add to the precise data thus far given a personal impression or surmise. Though, as I -have shown, intermediate lines occur, I believe that it will be found that most Paramecia can be placed in one of the two groups that we have called ' caudatum' and 'aurelia'. In other words, if my impression is correct, most lines will have a mean length either below 145 microns or above 170 microns; rarely will lines be found whose mean falls between these values. Such at least has been my experience in a large amount of work. Furthermore, I am inclined to believe that those belonging to the smaller group (mean length below 145 microns) will be found to have as a rule two micronuclei; those belonging to the large group but one micronucleus. This matter is worthy of special examination."

Jennings and Hargitt in 1909 made this examination and in a preliminary communicatioif stated^* that two sets of races could be distinguished, one set having two micronuclei, the other but one. The races with two micronuclei were all smaller than those with one. The larger races together thus correspond with what had before been described as P. caudatum, the smaller races with P. aureha. The two differ also in the size, position and

i^Heredity, variation and evolution in Protozoa. II. Heredity and variation of size and form in Paramecium, with studies of growth, environmental action and selection, Proc. Amer. Philosophical Society, 47, no. 190, 1908.

Characteristics of the diverse races of Paramecium, Proc. Amer. Soc. Zoologists, 1909 meeting, in Science, March 25, 1910.


staining relations of the micronuclei, in ways that correspond to the descriptions of Hertwig and Maupas. But in rare cases specimens of the caudatum races have two micronuclei, those of aurelia races but one, thus confirming the observation of Calkins on this point."

In accordance with the conclusions of Calkins, I have used the specific name aurelia to include both the aurelia and caudaturn forms; but my extended study of Paramaecia cultures has demonstrated that these two forms are remarkably constant, and I am inclined to the view that they are distinct species, in the sense in which this term is generally used in biological work. The data on which I base this conclusion are chiefly as follows: the pedigree culture of P. aurelia which I have had under daily observation for (so far) more than three and one half years, during which time more than 2100 generations have been attained, has bred practically true to the aurelia type as described by Maupas in the passage quoted. The pedigree culture of P. caudatum which I have carried for nearly seven months, and which has attained more then 300 generations up to the present time, has bred practically true to the caudatum type as described by that author.

The pedigree culture of P. aurelia was started on May 1, 1907, with a 'wild' individual which was found in a laboratory aquarium, and was carried on at Williams College during May and June, 1907; at the Woods Hole Marine Biological Laboratory during parts of the summers of 1907 through 1910; and at Yale University during the academic years from 1907 to the present time, November 30, 1910. The pedigree culture of P. caudatum was started on May 14, 1910, with a 'wild' individual collected from a pond at New Haven, Conn., and was carried on at Yale University except for a period of a few weeks in the summer when it was taken to the Woods Hole Laboratory.

The original specimen of each culture was placed in about five drops of culture fluid on a glass slide having a central ground concavity, and when the animal had divided twice, producing four individuals, each of these was isolated on a separate slide to form the four lines of the respective cultures. The pedigree cultures have been maintained by the isolation of a specimen from


each of these lines practically every day up to the present time, thus precluding the possibihty of conjugation taking place between sister cells. The number of divisions of each line has been recorded daily at the time of isolation and the average rate of these four lines has been again averaged for ten-day periods (cf. fig. 1). The culture medium has consisted of materials collected practically at random from laboratory aquaria, hay infusions, ponds, etc. The infusions were thoroughly boiled to prevent the contamination of the pure lines of the pedigree cultures by 'wild' individuals. Permanent preparations have been preserved from time to time for the study of the cytological changes during the life history.

In the light of this experience with cultures I shall consider each of the characters emphasized by Maupas.

Shape. The general shape of the aurelia and caudatum forms is, in nearly all specimens, quite distinctive; aurelia is slightly more broad at the posterior than at the anterior end, while caudatum, as the name implies, is quite pointed at the posterior end as compared with the anterior end. The posterior end, in the specimens in my pure culture, is markedly pointed, and being free from endoplasmic inclusions, appears transparent and clearly delineated even under a lens with a magnification of ten diameters. I have been accustomed to allow stock material from my pedigree aurelia culture to multiply in large flasks of hay infusion, for various experiments on conjugation, etc. Frequently I have used this material for ni}^ elementary class in biology and I have found that even the novice has called attention to the fact that the shape of the ends was reversed as compared with the figure of caudatum in the text-book. McClendon, however, stated that in his study of aurelia and caudatum he found no characters of outward form" which were diagnostic.

Changes in the vitality of my pedigree lines never have been very marked, and consequently I have not had organisms, in the direct lines of my pedigree cultures, representing physiological extremes to compare. Numerous experiments, however, have been made with 'stock' material left over after the daily isolations of the pure lines, which have clearly shown that, for


example, even when the aurelia and caudatum cultures are subjected to unfavorable environmental conditions, as, for example, scarcity of food, the very great majority of individuals retain the shape which is characteristic of the race.

Size. As has frequently been pointed out, size alone is an entirely inadequate character on which to base species. It is significant, however, I believe, that during the long life of my pure strains, I have never observed the relative size of the individuals of the aurelia and caudatum forms, when bred under identical conditions, to change greatly during any single period. Experiments have shown that even when the two forms have been bred under diverse conditions, for example, aurelia in a medium rich in food and caudatum in a medium with a very small amount of bacterial growth, the size of the caudatum form always has remained sufficiently great to render it distinguishable from the aurelia form. On the basis of size alone, then, it has been possible, with great accuracy, to separate the two forms when mingled together. It is probable, of course, that I began my pedigree cultures with very typical specimens of the aurelia^" and caudatum groups as described by Jennings. If such be the case, then my cultures add considerable evidence in favor of the different strains which Jennings has isolated. It appears to me, however, that what that author has done for Paramaecium, can probably be done for many closely related species of infusoria, and the very fact that he did find it difficult to secure an intermediate race between the aurelia an-d the caudatum groups is a strong point in favor of the. distinctness of the forms.

Vitality. It has been customary to regard the rate of reproduction of infusoria in culture as a just criterion of vitality. Maupas wrote i^^ "Cette faculte de reproduction (aurelia) resemble beaucoup a celle de la precedente espece (caudatum)." My cultures completely corroborate this statement, for during the six and one half months of the life of the caudatum culture, 324 generations have been attained, while during the same period,

^"For further details of the culture see: L. L. Woodruff, Two thousand generations of Paramaecium; Archiv fiir Protistenkunde, 21, 3, 1911. '^'Sur la multiplication des Infusoires cilies, loc. cit., p. 234.



under identical conditions, the aurelia culture has advanced from the 1785th generation to the 2117th generation, or 332 generations. This gives a difference of only eight generations in the rate of reproduction of the two forms during seven months (cf. fig. 1). These cultures obviously do not support the statement, frequently made, that aurelia is a weaker form than caudatum.

Maupas remarked that P. aurelia was one of the most common infusoria, and Jennings found that a typical wild culture could





Fig. 1 Diagram showing the comparative rate of division of the pedigree cultures of Paramaecium aurelia and Paramaecium caudatum, when bred under identical conditions, from May 14, 1910, to November 30, 1910. During this period P. aurelia (designated by continuous line) advanced from 1785 to 2117 generations, while P. caudatum (designated by broken line) advanced from 1 to 324 generations. The rate of division is averaged for ten-day periods. The ordinates represent the average daily rate of division of the four lines of the cultures.

be resolved into caudatum and aurelia groups. It has been my experience that it is as easy to procure one form as the other in the wild state. Certainly my aurelia culture, which theoretically would provide individuals to the number represented by 2 to the 2117th power, gives more evidence of vitality and reproductive power than has been demonstrated for any other animal.

Conjugation. I have no data in regard to the conjugation of either of these forms, for, so far, in all experiments with stock


material left over after the daily isolations from my pure lines, I have failed to observe a single syzj-^gy, either between aurelia lines or caudatmn lines, or between aurelia and caudatum lines. Jenning's" experiments on conjugation in Paramaecium bring out data which add further evidence that in certain strains at least a predisposition to conjugation does not exist. Maupas wrote: C'est bien certainement une des especes (aurelia) qui se recontrent les plus frequemment a I'etat conjugu^."

Maupas, as we have seen, pointed out a difference in the nuclear phenomena during conjugation which he held to be of diagnostic value, and Hertwig apparently showed that aurelia has two micronuclei at the reorganization after conjugation. Calkins, on the other hand, has shown that P. caudatum, in one case, reorganized with two micronuclei and later reverted to the uninucleate type. Such a case can readily be considered a ' sport ' which has arisen possibly by the persistence of the stage with two micronuclei immediately following the separation of the conjugants, or by the precocious division of a single micronucleus previous to the first regular vegetative division after conjugation. Although, as Calkins stated also, forty-five generations is a long time for an abnormality, if it be such, to persist; nevertheless, I believe it is very significant that, whereas during the presence of two micronuclei the division rate averaged only 0.8 of a division per day, after the loss of one of the micronuclei the division rate increased to the remarkable rate of 2.2 divisions per day, on the average for a period af four months. It is also of interest that the other exconjugant which reorganized 'normally' as caudatum failed to live.

So far as I am aware, the following statement^'* by Simpson is the only record of a possible case of conjugation between aurelia and caudatum: Out of twenty-one attempts I had but two partial successes. Conjugation took place on two slides: the period was normal. After separation each of the ex-conjugates divided once : on the third day they died off. In anticipation of something

"What induces conjugation in Paramecium? Jour. Exp. Zool. 9, 2, 1910. ^^Observations on binary fission in the life-history of Cihata, Proc. Royal Soc. Edinburgh, 1901, pp. 407-408.


of this sort from analogy in higher forms, I intended to let the two pairs run their natural course, foregoing the desire to examine their nuclear condition. In view, therefore, of the incompleteness of the experiment, it is perhaps unwarrantable to draw any results regarding hybridization and infertility, or even the 'fixity of species' so far down in the animal scale." Simpson gives no data to prove that these were actually syzygies between the two forms, but if they were, it is obvious that they were not fertile. Jennings and Hargitt stated that they had been unable to induce the two forms to conjugate.

In view of the fact that, for example, Maupas studied conjugation of both P. aurelia and P. caudatum, and Hertwig studied conjugation of P. aurelia, and also that Jennings observed conjugation in both his aurelia races and in his caudatum races, it is clear that aurelia forms conjugate and caudatum forms conjugate, but there is no positive evidence that conjugation takes place between individuals of aurelia and caudatum.

Macronucleus. The normal macronucleus of aurelia was described by Hertwig and Maupas and that of caudatum agrees very closely. It is an ellipsoidal body with a smooth contour, except for a slight depression, in which the micronucleus is usually located. But the form of the macronucleus of both aurelia and caudatum frequently departs very greatly from the 'normal' condition. It is not unusual to find paramaecia of my aurelia cultures with the macronucleus resolved into several parts. These parts apparently may be gathered together into a typical nucleus for division, or the cytoplasm and micronuclei may divide, the macronuclear fragments which are in the posterior part forming the macronucleus of one daughter cell and those in the anterior part forming the macronucleus of the other daughter cell. I shall reserve the full discussion of these interesting changes for a special paper. It is - important to emphasize the fact that these are not pathological conditions, since the general vitality, as indicated by the rate of division, is not appreciably affected.


Calkins,-^ however, found nuclear fragmentation in degenerating individuals of caudatum, Wallengren-^ and Kasanzeff^^^ showed that various changes including fragmentation of the macronucleus occur when paramaecia are starved, and Popoff-' described a large increase in size and fragmentation of the macronucleus in degenerating caudatum which paralleled the conditions observed in specimens ripe for conjugation. He also obtained similar changes by subjecting the animals to various reagents. ^^ Mitrophanow^* emphasized the fact that the structure of the macronucleus varied considerably under the influence of diverse conditions, and he described fragmentation and figured spherical pieces which very closely resembled micronuclei.

It is evident, then, from my cultures that the macronucleus of both aurelia and caudatum is subject to great morphological variation without appreciably affecting the rate of reproduction, i.e., it is entirely normal. It is also apparent from the work of the other authors cited that degeneration changes become manifest in the fragmentation of the macronucleus. Consequently the macronucleus presents no character which is of permanent diagnostic value.

Micronucleus. Maupas, as we have seen, regarded the micronucleus as the chief distinguishing character of aurelia and caudatum, and my cultures substantiate his view. Fixed and stained individuals show that the micronuclei of the aurelia culture for over two thousand generations have conformed in a remarkable degree to the aurelia type as described by the French investigator, and the micronuclei of the caudatum culture have conformed to his caudatum type.

2*Studies on the life history of Protozoa. IV. Death of the A Series, Jour. Exp. Zool., 1, 3, 1904.

-^Inanitionserscheinungen der Zelle, Zeit. f. allg. Physiologie, I, 1, 1901.

-"Experimentelle Untersuchungen ueber Paramecium caudatum. Inaug.' — Diss., Zurich, 1901.

-^Depression der Protozoenzelle und der Geschlechtszellen der Metazoen, Archiv fur Protistenkunde, R. Hertwig Festband, 1907.

2«Experimentelle Zellstudien III. Ueber einige XJrsachen der physiologischen Depression der Zelle. Archiv fiir Zellforschung. 4, 1909.

-"L'appareil nucleaire des Param6cies, Arch. Zool. Exp. et Gen., (4), I, 1903.


It is not only the presence of two micronuclei, but their peculiar morphology, as emphasized by Maupas, which is characteristic of the aurelia type. I have found one individual of the aurelia culture with three micronuclei, and a few specimens in which I have been unable, in total mounts, to distinguish a single micronucleus or more than one micronucleus. ,But when only one micronucleus could be seen it has been of the aurelia type, and other individuals of the culture at the same period have had the two characteristic micronuclei. I have observed a variation in the number of micronuclei in various pedigree cultures of hypotrichs,^" Popoff has found reduplication in Stylonychia mytilus and Paramaecium caudatum during degeneration, and Kasanzeff has observed the same in starved P. caudatum. Thus, while my cultures of Paramaecium and various hypotrichous species substantiate Wallengren's and Calkins' statement that the micronuclei are the most stable elements in the cell, and the last to be visibly affected by environmental changes, nevertheless it is apparent that they are subject to variations under certain unknown conditions. Temporary variation, therefore, cannot be considered as having weight in determining species. The essential fact is, however, that throughout the existence of my aurelia and caudatum cultures, the morphology of the micronuclei has conformed to Maupas' description for the respective species. It must be borne in mind also that P. caudatum has been the subject of more extended study by exact culture methods than any protozoon except P. aurelia, and in all these long pedigree cultures it has bred true to the caudatum type, at least with respect to the single micronucleus. Calkins, for example, in his important investigations on the life history of this form, carried three distinct cultures, by the aid of artificial stimuli during periods of physiological depression, through 379, 570, and 742 generations respectively. McClendon, also, studied mass cultures of Paramaecium for considerable periods and stated that he never found individuals with different numbers of micronuclei in the same culture. "^

^"An experimental study on the life history of hypotrichous Infusoria, Jour. Exp. Zool., 2, 4, 1905. "Protozoan studies, Jour. Exp. Zool.. 6, 2, 1909.


Briefly stated, I am convinced from my study of paramaecia that —

1 . A very great maj ority of individuals of aurelia and caudatum can be distinguished on the basis of shape alone;

2. A very great majority of individuals of aurelia and caudatum can be distinguished on the basis of size alone;

3. The power of reproduction, or general vitality, of aurelia and caudatum is practically identical;

4. The macronucleus of aurelia and caudatum is subject to such great variation that it affords no diagnostic feature;

5. The micronuclear apparatus of aurelia and caudatum affords crucial diagnostic characters.

I have summarized the various characters of the two forms as they have shown themselves in my long pedigree cultures, and it is evident that they have conformed practically identically to the Maupasian types — such variations as have appeared not being so great as have been observed to occur in undisputed species, or as one would expect to find when the intimate relation of the unicellular organism to the environment is considered. Therefore, I believe, that since one of the crucial tests of a species is its ability to breed true to its type indefinitely, aurelia and caudatum have adequately met this test during more generations than any other animal under observation, and accordingly Paramaecium aurelia O. F. M. and Paramaecium caudatum Ehrbg. should be regarded as distinct species. ^-^ ^^

  • 2In this paper I have followed the spelling of the name of the genus as given

by its founder, except in direct quotations from other authors.

'*I have the satisfaction to note that my conclusions are in accord with the final results published by Jennings and Hargitt in the last number of this journal, which was received when this paper was in press. Hargitt says, "There is cytological warrant for distinguishing caudatum races from aurelia races, and it seems probable that it will continue to be convenient to distinguishthese as two species."





From the Zoological Laboratory, Johns Hopkins University



The present paper is a contribution to our knowledge of the means that lead to the fertilization of the egg. It is part of the history of the sperm outside the body of the animal.

Sexual reproduction in most complex animals involves the transfer of the sperm from one animal to another, before the eggs can be fertilized.

Among animals the various methods by which the sperm is transferred may be grouped under the three heads, diffuse, direct, indirect. By diffuse sperm transfer we mean the discharge of the sperm into the water, where it may meet the eggs outside of the female, as in certain coelenterates, echinoderms and annelids, or may be drawn into the body of the female, as in certain lamellibranchs. By direct sperm transfer we mean the method found in the majority of complex animals, in which there is more or less direct application of the terminal parts of the passages leading the sperm to the exterior to the passages leading from the exterior direct to the eggs. In this group there is commonly a true copulation.

By indirect sperm transfer we mean those peculiar complex methods of getting the sperm from the testis to the eggs that are found in a few cases amongst the great groups of animals, as in




earthworms, spiders, some cephalopods and leeches. The essence of these cases of indirect sperm transfer Hes in the fact that while the sperm is transferred by organs, and not by floating, yet these organs either do not put the sperm into the egg passages, or else if they do they are not organs directly concerned with the discharge of sperm, or both may be true. In indirect sperm transfer there is no true copulation, or intromission, but at most conjugation or clasping.

The three methods are not always sharply separable, and may be regarded as only convenient groupings of physiological processes that occur here ^nd^here among animals without reference to their systematic positions. The diffuse method is obviously the one open to the most hazard in the sperm and eggs meeting; the direct method by intromission the best assured method; the indirect method most perculiar and needing special explanation in each case.

In the crayfishes and lobsters most interesting cases of indirect sperm transfer occur, and it is the purpose of the present paper to describe the organs of the males that are used to transfer the sperm to the receiving organs that have been described in previous communications (1, 2, 3). In these animals the male transfers the sperm to the outside surface of the female where it remains till the eggs are laid, when fertilization takes place outside the body. In the American lobster and the sixtj^ and more species of crayfishes of the genus Cambarus that are found in all but the most western parts of North America, the sperm on the shell of the female is stored in a special pocket, or receptacle, but in the other genera of crayfishes, all the world over, there is no such receptacle and the sperm is believed to be distributed over the shell of the female in separate spermatophores. While the sperm pocket has been described (1,2, 3) the organs of the male that fill the pocket have had only such consideration as was necessary for the systematist, who found them to be of the greatest value in distinguishing species and in forming subgenera.

In the present paper the anatomy of the male organs is examined and their use as organs of sperm- transfer is explained.



While the sexual habits of all the species of Cambarus, agree in the main, the species afFmis has been more studied, and as in describing the female organs concerned in sperm transfer we first considered this species, we will also give chief attention to the male organs of sperm transfer in this species.

As elsewhere described (4, 5) conjugation is here a long series of activities of the male accomplishing the accurate adjustment of the essential transfer organs of the male to the receptacle of the female. The receptacle of the female is a single pouch in the shell, but the transfer organs of the male are three pairs of outgrowths. On each side of the body there is a papilla, or special termination of the sperm duct, and two limbs, those of the first and second segments of the abdomen, which we may call the stylets.

To introduce the sperm into the receptacle the papilla must be adjusted both to the first and to the second stylet and both sides of the body play a necessary part in the process of sperm transfer.


One external instrument concerned in the process of sperm transfer is the modified end of the sperm duct that emerges from the base of the fifth leg, on each side of the animal. These organs are the papillae.

Since systematic work has been done largely upon preserved specimens it is not so generally known that the sperm duct ends in life, in a soft, turgid protuberance, which may be so collapsed after death as to leave only the rounded hole in the firmer shell as the apparent ending of the sperm duct. These papillae lie concealed by the stylets, at rest, but on raising the stylets the papillae are seen as conspicuous, clear, tubes about 3 mm. long and 1^ mm. wide, jutting out from the base of each fifth leg.

At the time of conjugation the papillae are also concealed from view since the necks of the first and the second stylets form a nicely adjusted frame about the papillae and this frame is fitted in between the bases of the fifth legs. Certain in and out move


ments of these bases seem to adjust the papillae so that they fit accurately into the orifices of the first stylets (figs. 30, 31) and by that means the sperm discharged from each papilla is passed into the cavity of the stylet.

The papilla (P. fig. 1) is on the under side of the large first segment of the leg and projects downward and toward the median plane; but its tip turns away from the middle line of the body. The papilla is a cone with bent apex. It is translucent and distended with colorless blood. When directly injured, or upon lessening of the blood pressure from injury elsewhere, the papilla collapses, being but a thin uncalcified protrusion of the skin, kept turgid, or erected, by blood pressure. Within the papilla one can see a large central tube passing toward the tip and also chalky white masses suspended between the central tube and the thin outer wal!s.

On the shell at the base of the papilla there is, anteriorly, a single row of very long setae (fig.l ) that form a sort of protective screen over the anterior face of the papilla.

Sections show that the papilla is a continuation of the deferent duct, blood cavity and skin, so constructed that the bent, conical apex, with its soft walls can be adjusted to the hard opening of the stylet so as to fit hermetically, as a tense rubber bag might. Moreover the bent tip can be opened to discharge the sperm, when special muscles remove the obstructing valve that holds the tube closed.

A lengthwise section through this delicate papilla (fig. 2) shows that the central tube is a direct continuation of the deferent duct that leads the sperm from the testes to the tip of the papilla. Between this duct and the outer cuticle there is a large space fall of blood, traversed by little connective tissue and in it are the white bodies just mentioned, now seen to be small tubular glands, opening into the central duct. The central duct presents two strikingly distinct parts; the one continued from within the leg has the thick muscular wall and peculiar secreting lining of the deferent ducts, the other is lined by the thin cuticle inflected at the orifice at the tip of the papilla, and lacks muscle. In place of muscle the wall has only epidermis, which extends irregular!}'



Fig. 1 Posterior face of the left fifth leg of a living male 95 mm. long to show the translucent papilla. (P.) 2. Qq.

Fig. 2 Longitudinal section through the papilla and the base of the fifth leg, showing the orifice of the sperm duct, the valve, the muscles and the glands. 2. 90 mm. A.


into the blood space as the tubular glands alluded to above. The orifice at the tip is small and is not closed by any muscle, but apparently by blood pressure only. The part of the tube lined by cuticle has its lumen much reduced by a valve, or great longitudinal ridge, which extends out as far as the abrupt bend at the orifice. In a cross section (fig. 3) this ridge is well seen, as is also the fact that some muscle fibres run into it and that the glands are chiefly on the side opposite the ridge. The ridge appears to act like a valve to hold this part of the tube closed, while contractions of the muscle would tend to open the tube wide and let the sperm pass to the orifice, which would then be forced open by the internal pressure of the sperm squeezed by the muscles of the wall all along the length of the duct, or some extent of it at all events.

The upper part of the duct, as seen in the cross section fig. 4, has its thick muscles arranged chiefly in transverse fibres and is lined by an epithelium that evidently in large measure breaks down to furnish a great mass of secretion about the sperms. It is probably this secretion that envelops the sperm in the form of macaroni-like tubes, when they pass out in a slow stream.


The most complex of the organs concerned in sperm transfer are the modified limbs of the abdomen which we will call the stylets. In the male the sixth pair of abdominal appendages form the large side parts of the tail fan while the third to the fifth inclusive are the simple and apparently rather useless swimmerets. The first and second pairs are specially constructed to serve as transfer organs for the sperm.

These appendages of the first and second somites are much stouter and longer than the following swimmerets and have a very firm attachment to the abdominal sterna. The calcified ridge across the middle of the sterna is much more developed in the first and second somites, and where the appendages are fastened it rises up as a decided elevation which remains as a stump when the appendage is cut off. On the second somite these stumps are far apart, (some 10 mm. in a male of 100 mm.) while on the first



Ex. m. ',

Fig. 3 Cross section of the sperm duct and valve along the line 3 of fig. 2, showing the duct closed by valve ridge. 2. A.

Fig. 4 Cross section of the sperm duct along the line of 4- of fig. 1, showing the muscular wall and the lining epithelium disintegrating in secretion. 2. A.

Fig. 5 Extreme tip of right first stylet, showing the groove bottom coming to the surface, posterior face. 2. A. Ex. m.— the external mass. M.m.— the internal mass. IS' — the level of the section, fig. 13.

Fig. 6 Diagram of stylet as in plate i, fig. i, to show location of glands in the interior, and the location of the sections, 7 to 13, shown in figs. 7 to 13.


somite they are in contact at the median Une of the abdomen. The eUiptical transversely elongated stumps of the first appendages are 5 mm. long and those of the second about 3 mm.

Commonly these appendages are carried forward horizontally under the thorax between the thoracic legs in a deep depression of the thoracic sterna. The first pair lie close side by side with their median faces in contact. The second pair lie over and largely conceal the first, since their form enables them to come to the middle of the body beneath the first pair in spite of the fact that their bases are attached to the sterna, so far from the middle line.

In a dead male one may move the appendages upon their attached bases as follows:

The first may be moved upon its base from the horizontal up toward the vertical only about 45°. The membrane on the anterior face of the joint at the base of the appendage is stretched to its limit when the appendage is pulled up a little beyond sixty degrees, so that this appendage is never vertical and cannot swing back and forth through a wide arc as do the ordinary swimmerets. The distance traversed by its tip is some two cm. The appendage may also be rotated a very little at its base and moved from side to side a little so that ts tip travels some 5 or 6 mm.

The apex of the second may be drawn from the horizontal up a little beyond the vertical; but neither the basal protopodite nor the endopodite travels more than 90°. They are set together at a large angle, so that while the main length of the appendage is horizontal the basal part never is, and when the base goes back some 90 degrees the horizontal part is swung past the vertical line. The tip traverses some 3 cm. The base may be rotated a little and moved from side to side so that the apex travels 6 or 7 mm.


The first abdominal appendage of the male is a very stiff calcified mass of the general shape of an awl, some 17 mm. long, but having two tips. There is a groove along more than half its length and the base is articulated to the ventral shell of the anima' so


that the appendage has very Httle mobihty back and forth through some 45°. The normal position of the stylet is pointed forward under the thorax, where it lies horizontally in a deep groove, but in use it is dropped down and backward toward a vertical position. It has an anterior face, which is usually carried as the dorsal side, a posterior face which is usually the ventral aspect, and an outer and an inner or median face.

The general appearance of the stylet is seen in the photographs, figs. I, II, III, IV, which represent respectively the posterior, median (or rather median and posterior somewhat diagonally), the anterior and the outer faces of the same left stylet. Fig. i, the posterior face, is the view got by looking at the under side of the crayfish, after lifting up the second stylet, which lies over the first and largely conceals it.

The first pair of stylets do not spring from the sternal surface far apart as is the case with the common, unmodified swimmeret, but they arise very close together; in fact the median faces, (fig. II,) of the two come into contact so that these two appendages really form one mass. If looked at from the dorsal side, the two are seen to lie in contact at the base and all along the distal half, leaving between the constricted parts of the two a square opening that is occupied, in rest, by part of the second stylet.

In describing the stylet we will distinguish the base, the neck, and the scroll or spiral that contains the groove. The scroll ends in two tips, the more slender, side outgrowth, or spatula, and the real end bearing the groove, the canula

The base is some 6 mm. wide and long and only 2 thick, being flattened from before back. The posterior or ventral face of the base, fig. I, presents a wide groove bounded on the median side by a rounded knob and on the outer side by a long ridge which, as it passes on to the neck, bears a tuft of long, finely plumose setae, that are seen again in profile in figs, ii, iv. In this deep groove the second stylet lies when not in use, so that the two appendages are firmly packed together under the thorax of the male.

The part of the base joined to the sternum of the animal is an oblique eUiptical area, around the edge of which the hard shell gives place to the soft articular membrane that makes it possible


to cut the whole appendage away from the sternum. In this membrane there is an articular, whitish plate that is seen in figs. Ill and IV. The whole base is pyramidal and except the posterior all its faces (figs, ii, iii, iv) are convex and rounded.

The neck is the narrowest part, before the sudden enlargement of the spiral part; it is the smallest of the three regions; and is best seen in figs, ii, iii, iv. The neck passes gradually into the base and ends abruptly at the spiral. It is some 3 mm. long and 2 wide and thick. It has an angle along the ventral face that continues the ridge of the base up to the outer part of the spiral.

The spiral or scroll may be likened to a long triangular plate with its edges rolled in together so as to leave a groove between them, but it is a plate some 8 mm. long, with the edges greatly thickened, so that the resulting mass is apparently solid. The groove begins on the median side, fig. ii, and passes in a sinuous course to the ventral side and along this diagonally to the very tip. The apparent bifid nature of the stylet is due to an outgrowth from the median part, quite separate from the real end of the organ, in which the groove is continued through its entire length. We have then to describe a sinuous groove and its two boundaries, which we will call the median mass and the external mass ; and also the two tips. The external mass, seen from the ventral side on the right of fig. i, shows a proximal part about 2 mm. long and 1 mm. wide, bearing a marked ridge parallel to its sides and continued up from the neck. And then it suddenly turns at a large angle and becomes a rounded and gradually tapering terminal part, something less than ^ nam. wide at first, and 6 mm. long. This passes behind the slender protuberance of the median mass to end as a flattened, horny tip together with the like ending of the median mass. In other words both external and median masses unite as the horny tip that we will call the canula. The sudden change in direction of the mass is accompanied by a like change in the groove whose edge it forms; this change of the groove we will call the angle of the groove.

Seen from the outer face, fig. iv, the external mass is widely swollen proximally, some 2^ mm. deep, and gradually narrows into the distal part. The round canula is bent somewhat, ventrally.


On the dorsal face, fig. iii, the external mass is confluent with the median mass, without boundary line. Thus the distinction between the two masses is useful chiefly on the anterior face where they form the two sides of the groove.

In fig. Ill, the long triangular region running from the notch that marks off the neck from the spiral region and ending distally in the rounded and pointed canula, is to be regarded as made up chiefly of the median mass, but the depressed part along the left edge is part of the external mass.

On the median face, fig. ii, (which is unfortunately turned so that part of the posterior face shows) the external mass shows only its prc-ximal end along the side of the diagonal groove, and into this groove the external mass here sends a narrow horny shelf, dimly seen as light in fig. ii. The external mass has an angular projection, or lip, at the very beginning of the groove which will be described in connection with the orifice of the groove. At the tip, part of the external mass is seen making the lower part of the canula, to the right, that is, the curved strip of external mass seen is flat and on a lower level than the median mass.

On the ventral face, fig. i, the median mass looks like a long rounded white bone that begins suddenly without apparent connection with the neck and, after running nearly straight for some 6 mm., turns externally across the external mass as a flat, curved process that we will call the spatula. Beyond the spatula, w^hich stands out freely as the second tip of the appendage, the median mass continues as the narrow median edge of the canula. From the external view, fig. iv, the visible part of the median mass, the spatula is back of the external mass.

In the dorsal view, fig. iii, the main part of the spiral region is median mass, forming a long triangle, beginning at a deep notch near the neck and extending in the foreground as the vis ble part of the canula and back of that as the spatula. At the notch may be seen part of the lip on the external mass.

The median face best shows the median mass, but, fig. ii, being not an exactly median view, does not do justice to it. In reality this face is markedly flat where it comes against the like face of of the other stylet of the pair. This flat face is a long ellipse, 2


mm. wide and 5 long, and is smooth except for a roughened area near its proximal end where there is a long tuft of finely plumose setae which bend abruptly downward, that is, posteriorly, as if an adjustment to the fact that they are pressed in between the two stylets. These setae are so long as to be visible from all points of view, cf. figs, i, ii, iii, iv.

The groove itself is seen only from the median and ventral views. It is some 7 mm. long and begins as the orifice on the median face where it meets the ventral, fig. ii. The orifice is a conical opening bounded by that depression of the neck that makes the notch so conspicuous in fig. iii, by the rounded origin of the median mass, fig. ii, and by the overhanging lip of the external mass. It is of such shape that the tip of the spout, fig. 1, can fit into it. The groove leads from the orifice obliquely outward and distally between the external and median masses some 3 mm. and then turns to make a rounded angle, fig. i, toward the median line some 3 mm. more. In this part of its course it is soon concealed behind the median mass that is rising to form the base of the spatula, but it still exists there and emerging again runs the entire length of the spatula as a very narrow slit with horny edges. The groove is thus a long double curve, bending abruptly outward, then forward and slightly inward and finally outward again, as seen from the ventral side. But it also bends in the vertical plane, passing downward, then forward and upward and finally a little downward at the tip. While the walls of the groove seem to be merely hard rounded bone there projects into the groove from the side a narrow shelf of horn that springs from the external mass only. This will be seen in sections.

The spatula is a flat flagellum-like process some 2 mm. long, | wide and perhaps | to iV thick. It is curved and pointed as seen in the figures. It springs from the median mass where this suddenly narrows to help form the canula, fig. ii. In life the spatula is milky white and pliable, not bony, more like leather. At its base it passes suddenly into the bony walls of the median mass and there can be bent as if in a socket. After drying it looks more like a thin chitinous membrane over a dried contents. It is somewhat concave at the base on the dorsal face. With methyl


green the horny tip of canula and the shelf in the groove stand out clearly as distinct from the substance of the spatula.

The canula is some 3 mm. long and at base f mm. wide and thick. It is a long cone, flattened somewhat from before back, bent upward dorsall3\ and ending in a rather sudden point that bends outward from the median side. The canula is made up of both external and internal masses. Most of the length of the canula is clear, yellow, horny matter, but at the base this is continued as the white calcified material of the rest of the stylet. The bone of the external mass stops rather suddenly, while that of the median mass is continued in the midst of the horny cap as a central area, as seen from the median view. An enlarged view of the tip of the canula, fig. 5, shows that both external and internal masses make about the same amount of the canula, since the groove continues sinuously almost to the exact tip of the organ, but yet there is a greater prolongation of the external mass to form a short ungrooved apex. This sketch is from a canula of the opposite side of the body from that in fig. i. The two canular tips flare away from one another.

The groove may be said to begin and to end on the median face and to be shoved away from it through most of its course by the ridge that we have called the median mass (fig. i.).


When the stylet is macerated some days the entire contents may be drawn out of the hard shell; such a cast of the shell has its general long conical form with a short conical tip that came out of the canula and a short flat plate that came out of the spatula. It is made of connective tissue and blood covered with epidermis with some red pigment cells and shows at the base some muscles and at the middle some glands.

The muscles, as made out by dissection of fresh and preserved crayfishes, are weak and run from the base of the stjdet into the adjacent ridge of the sternum upon which the stylets articulate. There is a wide thick fan of muscle that passes from the bony articular plate of the anterior face of the stylet, fig. iii. When


this is pulled the stylet is raised dorsally into its position of rest. Since it lowers the organ into the groove on the thorax it may be called the depressor, though it really swings the appendage forward.

This depressor muscle is lodged in the protruding ridge of the sternum from which the stylets spring, and its fibres are made fast to the posterior wall of this ridge. There is also a smaller muscle attached to the base of the stylet at its external edge which would seem to antagonize the other and to tend to swing the stylet backward, that is, to raise it up from its horizontal position of rest into the erect position of use; it may be called the erector muscle.

The internal anatomy of the stylet as well as the character and mode of use of the groove, were made clear from sections.

The diagram fig. 6 shows the ventral view of a left stylet as if transparent, the extent of the glandular area being shaded; the glands occur in both external and internal masses, but not in the base of the stylet, and the}' extend from the neck to near the origin of the spatula, filling most of the cavity of the region in which they occur.

The sections, (figs. 7 to 13, inclusive), were taken across the stylet along the planes indicated by the like numerals in fig. 6.

The transverse section, (fig. 7) shows in black the exceedingly thick shell with the depression on one side that forms part of the orifice of the groove, overhung by the solid lip. Through the thickness of the shell that forms this part of the orifice are seen many fine tubes, passing from the internal glands to discharge on the surface. The interior of the stylet is a delicate mass of connective tissue, chiefly blood sinuses, crossed by few strands of tissue, and bounded by the thin epidermis against the shell.. Scattered all through this are the tubular glands that bend and are cut at various angles. These glands ultimately discharge by the numerous fine ducts that penetrate the shell. In this section the sharp angle above is the ridge {R) seen in figs. 6 and I passing along the external mass. The angle to the right is the line between the ventral and external faces of the external mass.



Sections 8 and 9 show the orifice passing into the groove; they are cut obUquely transverse and, in addition to the section of the first stylet, show also the section of the second stjdet as it lies locked in the first. Disregarding for the present all but the lower part of the sections we see that the stylet has widened out from the constricted neck into a wide flattened mass sub-divisible with reference to the groove into the external and median masses. In

Fig. 7 Section across the stylet, in the region of the neck, just below the orifice, on the level 7 of fig. 6. 72— the sharp ridge on the external mass, fig. 6 and i. 2. 90 mm. A.

Fig. 8 Cross section of the same at the level 8, showing the groove above the orifice filled by the head of the accessory stylet, which is the separate. mass lying to the left and above. 2. 90 mm. .4.

fig. 8 the orifice is so overhung by the lip as to be in section a Cshaped bay, embracing the head and neck of part of the second stylet. Here again the shell is remarkably thick, but is penetrated by the ducts of the glands discharging on the surface that lines the orifice. In fig. 8 the lower straight side to the left is the flat face that is normally applied against its fellow on the outer side of the body. Above is the angle {R) that represents the ridge of



the external mass, just as in fig. 7. In the interior some of the glands are very large. The section distal to this, (fig, 9) shows the bottom of the groove receded from the surface and constricted from the rest bj^ the continuation of the lip so that it forms a rather elliptical hole with only a very narrow slit opening into the deep groove that is seen from the surface. This surface groove is bounded on the left by the greatly thickened shell substance of the median mass and on the right by the thick shell of the external


Fig. 9 Cross section of the same at the level 9, showing also the grasping second stylet, above, and its wedge, to the left, where it is entering into the groove of the spiral. 2. 90 mm. A.

Fig. 10 Cross section of the same at the level 10, showing the bottom of the groove cut off by the shelf from the external mass. 2. 90 mm. A.

Fig. 11 Cross section of the same, at the level 11, showing, above, the base of the spatula. 2. 90 mm. A.


mass. The cavity within the shell of the external mass is reduced to a narrow space and the glands have become few.

Further along the stylet, (fig. 10) the groove has passed from opening to the left (fig. 8), through the position shown in fig. 9, to open more toward the right. The groove is a deep and narrow one. Into it still open some few gland ducts from the remaining glands of the median mass. As before the side walls of the groove are made of very thick shell. The most unexpected fact is that the bottom of the groove is shut off as a very minute hole overhung by the continuation of the lip, which is now a horny shelf passing all along the groove, near its bottom, and so nearly meeting the opposite side as to practically shut off the bottom of the groove as a special tube. This figure shows the form of the stylet at the level, 10, of fig. 6. The flat side to the left is the flat face of the median mass, while the rounded edges of the groove are the two narrow parts of the external and median masses seen from the ventral side in figs, i and 6, just proximal to the base of the spatula.

A section through the base of the spatula, (fig. 11) shows the groove above overhung by the rising spatula that conceals it from surface view, (figs, i and 6) but still allows access to the groove from the right, in under the spatula base. The external mass {Ex. m.) is now the greater, but it contains no glands, while the median mass is reduced to a nearly solid shell prolonged as the slightly hollow spatula. The tube at the bottom of the groove is still there, overhung by the little chitinous shelf.

Near the apex of the organ, (fig. 12) the groove is again open above, as we have passed beyond the base of the spatula, only the tip of which is cut, lying well over to the right. This figure being magnified twice the diameter of the preceding figures, shows plainly the shelf that cuts off the bottom of the groove. The median mass is a narrow and nearly solid shell that forms the left wall of the straight, deep groove. The external mass is the main part of the section and contains much very watery connective tissue, covered with epidermis. In this section, the calcified part of the shell is represented in black, as in the other sections, while the chitinous or horny parts are dotted. From the surface this region of the canula looks to be only chitin. Farther on the



calcified part of the shell fades away and only pure chitinous matter is left, so that a section at the very tip of the canula, (fig. 13) is only chitin. This view is enlarged four times as much as the preceding one and shows the disappearance of the superficial part of the groove though the bottom, which is now close to the surface, is still overhung by the shelf from the external mass. That is the tube at the bottom of the groove can now discharge by a slit to the surface at the tip of the canula; see fig. 5, where the surface slit of the groove is represented by the black line and the bottom of the groove, or the tube, is represented by the dotted line, which comes finally to the surface at the tip as seen in the section across the level 13. As fig. 5 is of *a right stylet and the section 13 from a left stylet it shows the parts reversed; the main bulk of the section is really of the external mass, as in fig. 12.

The specialization of the bottom of the groove had not been expected till sections revealed it and suggested some special use. Sections of stylets taken when being used in conjugation soon showed that the tube at the bottom of the groove is the channel for the transfer of sperm. Along this minute tube all the sperm passes from the papilla to the sperm pocket of the female. A section across the stylet where the median surface bears a tuft of setae, between the levels of 8 and 9 of fig. 6, when sufficiently enlarged, shows that the sperm is contained inside the tube of the groove, as in fig. 14. This shows only the part of the shell about the tube, with the sharp edge of the shelf above, jutting out to almost meet the wall of the median mass (see fig. 9). The cavity of the tube is full of a secretion containing at its centre a pearshaped mass of the peculiar sperm of the crayfishes. As was shown (6) these sperm do not assume the star shape they have in books as long as they are in the male and not even when in the sperm receptacle of the female when normally protected from the water, and in this section, where they are seen in transit, they are still spherical, clear bodies with the peculiar bowl-shaped central part that, as represented in the sketch, might be thought a central nucleus. All along the groove above the orifice there is thus a strand of sperm surrounded by a paste-like white mass that fits tightly into the tube.

Fig. 12 Cross section of the same near tip, at level 12, showing the spatula cut off to the right. 2. A.

Fig. 13 Cross section of the very tip of the stylet, at the level 13, of figs. 5 and 6, showing the groove coming to the surface of the horny canula. 2. D.

Fig. 14 Enlarged view of section across the tubule, between the levels 8 and 9 of fig. 6 showing the sperm cells enveloped in a secretion and shut in by the shelf above. 2. D.

Fig. 15 Enlarged view of section of tubule and bottom of groove, about the level 10 of fig. 6 showing the fewer sperm and little secretion in the tubule, surrounded by a thick horny layer. The calcified skeleton is represented as black. g. D.



That this mass is run in under pressure seems indicated by the way it tends to flow out at the narrow sUt leading up from the tube into the groove and by the form of the sperm mass that tends hkewise to copy the shape of the cavity that is filled, being pointed toward the slit (fig. 14) . In successive sections this sperm mass is found all along the length of the groove, always in the bottom of the tube only, while the enveloping secretion for the most part disappears. Thus in the fig. 15 from the level 10, where there are still some secretion tubes coming through the heavy shell of the median mass, (fig. 10) there are a dozen or so sperm enclosed in the minute tube together with very little secretion and the sperm seem to come into contact with the shell. At this level, however, the thick and well-calcified shell (fig. 10) is covered by a thick layer of horny substance that ntiakes the shelf and continues on up the face of the external mass bounding the groove, (fig. 15). The discharge of the milk-white sperm from the tip of the canula, (figs. 5, 6 and 13) was seen in some males separated from females in conjugation.

The anatomy of the stylet thus shows it to be a more refined and specialized tool for sperm transfer than had been expected. It is essentially a very fine tube receiving sperm at its larger base and discharging it at its attenuated tip; but it has walls that give it great strength and rigidity while allowing the tip some elasticity. Moreover the receiving part of the tube is provided with glands of problematic value.

In looking for further light upon the nature of this sperm transfer organ we turn to its development in the individual.


We find that in Cambarus affinis the first and second larval stages are externally alike, in both sexes, while the third shows the male openings on the fifth legs, or the female on the third legs. In the first stage, there are no abdominal appendages on the first somite and but a crowding of epidermal nuclei under the shell where the appendage will be. In the second stage, these appendages are slight papillae. These indifferent stages are fol


lowed by the third, in which the external openings are differentiated but the appendages of the first somite are still simple papillae, alike in both sexes, unless they be longer in the male.

In the fourth stage, which is about 11 mm. long, the pleopods of the female still are simple papillae but little longer than in the third stage, while in the male they are long, simple spines, pointing toward one another and but slightly forward, as indicated in fig. 10 p. 127, Andrews, Ontogeny of Annulus, Biol. Bull., 1906.

The ventral face of the left spine or slightly specialized first pleopod, of a male 11 mm. long, is seen in fig. 16, magnified 430 in the camera sketch. This is from a larva killed July 1st, from late spring hatching. The organ is like a club ; it is very simple, nearly cylindrical and very blunt. It is not jointed, although there is a faint groove marking off the base from what will be the neck and spiral.

On the base there is a slight ridge with depressions on the median side of it. Internally there are two muscles from the base into the sternum of the abdomen. The distal part of the appendage is slightly grooved along its ventral face, thus marking off an external from a median mass. In cross section, fig. 17, the shell is not very thick and beneath it is a well formed epidermis with large nuclei, from which connective tissue strands traverse the large blood space in which blood corpuscles float. This section shows the groove on the lower side. The appendage is articulated to a slightly elevated stump on the sternum that holds one of the articular muscles and part of the other and ends in an elliptical orifice into which the base of the stylet fits. This articulation is so oblique that the stylet lies down and cross-wise towards its fellow and is but little elevated or directed forward.

In the male of this stage, the openings on the fifth legs are short slits, not a third of the width of the above simple stylet, and to each slit there leads a strand of nuclei that represents the efferent duct.

In males of 15 to 18 mm., in the fifth stage, the stylet (fig. 18) is about 1 mm. long and is somewhat more specialized. The base is set off from the terminal part by a more pronounced fur



Fig. 16 Posterior face of left first stylet of male 11 mm. long. Enlarged 21S diameters.

Fig. 17 Cross section of the same stylet. 2. D.

Fig. 18 Posterior face of left stylet of male 18 mm. long. Length of stylet 1 mm. f. A.

Fig. 19 Section of stylet of a male 12 or 15 mm. long. 2. D.


row, but there is no movable joint. The organ is more pointed and the groove is very deep from the rising up of its sides. Thus in section fig. 19, the narrow median mass, {M.m) to the left, rises high up beyond the groove and the groove itself is a narrow space between the wide external and the narrow median masses. In the surface view, (fig. 18) the bottom of the groove is indicated by the broken line; it is already twisted so that the groove looks towards the median side along its proximal part and then for a short distance toward the observer, that is toward the ventral side, and finally at the tip toward the median side again. Where the groove is open ventrally the median mass is rising up as a protuberance that will form the spatula. As yet the canula is only the spoon-shaped end of the organ.

In a male 22 to 21' mm. long and probably in the sixth stage, (fig. 20) killed October 4th, we find the same stage as in other males of this size killed in July, this being an exceptional male that failed to grow as the average do to be nearly two inches long in October. Here the spatula is quite evident as a blunt rounded finger-like elevation that crosses over the groove. As shown by the dotted line the bottom of the groove is to the right of its mouth along the proximal part of its course and to the left along under the base of the spatula; that is, the sinuousness of the groove is exaggerated by the fact that the sides not only rise up but grow over the groove, the external mass overhanging toward the median line proximally and the median mass growing over away from the median line, distally. The base of the stylet now bears a few short acicular setae and is provided with three muscles at its attachment to the sternal elevation upon which it stands. By this time the stylets point forward under the thorax. The canula is now a short rounded blunt termination of the stylet in which the groove is no longer widely open but reduced to a slit by the upgrowth of its walls.

In an autumnal male 38 mm. long, (fig. 21) the stylet has become much longer and more modeled but still shows the stiff joint between the base and the partly-formed neck. The few setae extend along the ridge of the base on to the proximal part of the external mass. The median mass sticks out abruptly at



Fig. 20 Posterior face of left stylet of male 22 mm. long in October. S. A.

Fig. 21 Posterior face of left stylet of male 38 mm. long in October. Enlarged 25 diameters. 2. 90 mm. A.

Fig. 22 Anterior face of left accessory stylet, somewhat turned to show part of the external face: a view between vii and viii. 2. ao- On the left an enlarged sectional view of the cup at the end of the radius and the wedge cut off.


the notch, or orifice, and bears a tuft of short setae. The spatula is long, flat and pointed. The canula is bluntly pointed and turned outward.

Later when the animal is 64 mm. long, the false joint of the stylet has disappeared and the tips become more sharp and long. Even before this size the males are known to conjugate, when about two inches long.

We thus find that the complex stylet of the adult starts from a slender papilla that becomes slightly flattened and grooved so as to form a very clumsy spoon with its depression rather more median than ventral. Then the sides of this groove grow up and make the groove into a cleft, which opens as before toward the median face proximally and distally; but along the middle of its course is forced to open ventrally and even externally by the overhanging growth of the median mass. The organ might be imitated by taking a long strip of clay with a slight length-wise groove on it and rolling the sides up over the groove, the median side tending to roll over outside the other. How the shelf from the external mass first grows out over the groove to cut off its inner part as a tube was not made out, but it is evidently a secondary specialization of the shell made by some special activity of the epidermis in a line near the bottom of the groove after the groove has becoriie deep.


The accessory stylets (figs, v-viii) are evidently specializations of the common type of abdominal appendages, (fig. 26). They are elevated only when in use in conjugation; and at rest are carried forward under the thorax, horizontally, where they rest upon the first stylets and are closely packed in with them inside the special sternal groove of the male thorax.

Figs. V, VI, VII, VIII, represent the left second stylet as seen from the ventral or posterior, the median, the anterior or dorsal, and the exterior faces, respectively. Like the unmodified pleopods this has a basal protopodite, an exopodite, an endopodite. The exopodite is a slender offset with setae, while the endopodite is


the complex large part of the appendage that bears a terminal fiabelum and the remarkable side protuberance, found on no other limbs, which may be called the triangle.

Describing the entire stylet from the base outward, we see that the protopodite is chiefly a very strong flattened bony mass extending diagonally inward so that while the endopodite and exopodite are about parallel to the median line of the animal the protopodite forms an angle of 45° with it. This makes it possible for the endopodites of the two stylets to come together at the median line and for the endopodite of each side to lie upon the groove of the base of the first stylet, like a lance in its rest, although the bases of the two second stylets are fastened to the sternum of the second abdominal somite some distance from the median line. The protopodite is not entirely one-jointed but at its base is a soft membrane where it is joined to the sternum and in this are two large calcified plates, (figs, v and vi) besides two minute ones, (fig. vii) all of which together make a narrow basal section of the protopodite. Dissection shows there are muscles passing from this base of the protopodite into the sternum that may depress and elevate the appendage.

The protopodite is some 6 mm. long, 2 wide and 1^ thick. The exopodite is a slender filament some 9 mm. long and ^ mm. thick; a slightly flattened tapering cylinder set with long setae on External and median face. The setae are really plumose and together form a sparse brush. The exopodite is obscurely divided into some twenty segments. The basal 2 mm. is partly calcified, the rest membranous. It articulates freely with the outer distal corner of the protopodite so that it may be moved from the position of rest parallel to the endopodite, outward through 90° and swung back and forth some 45°. The tip of the exopodite often lies dorsally within the cavity or hollow of the triangle, and may have some use as a cleaning brush.

The endopodite is the stout calcified mass, roughly cylindrical but flattened from before back, some 9 mm. long on the median (fig. vi) and 7 mm. on the external face (fig. viii), and bearing at its distal end a flagellum on the external side and the flat triangle on the internal side. This bony mass is set on the protopodite


by a very stiff oblique joint at about 45° and allows of very little lateral and rotative motion. It may be forced outward and inward through but few degrees, its tip traveHng only 4 mm. It may be twisted so that the triangle, from being almost concealed dorsal to the end of the bony mass (fig. v), may be turned outward a few degrees toward a horizontal position and present more of its median face, somewhat as in fig. vi. The movement is comparable to that of a stiff arm that should allow only a little sidewise movement and a verj- little twisting at the elbow with the end result that the triangle, or hand, at the end, accomplishes a little adjustment to the orifice of the first stylet. This is done as if by supination, though done by the above twist at the elbow.

The flagellum is the real termination of the endopodite; it is some 3 mm. long, 1 mm. wide and rapidly tapering, also flattened, being a long triangular terminal tip to the essentially flat endopodite. By the presence of white lateral areas in the otherwise membranous flagellum, it is obscurely divided into 9 or 12 joints. At the tip and along the sides it bears long plumose setae that are often sparse or worn off along the outer side. The flagellum springs from a socket in the bony shell of the wide end of the endopodite. The external angle of the edge of this socket, figs, v and VIII, forms a hard protuberance at the end of a bony ridge (the Guide). The setae along the flagellum as well as those along the exopodite do not stand out horizontally, right and left, but slant ventrally, or posteriorly, (fig. viii).

The most novel and characteristic part of the second appendage of male crayfishes is the lateral outgrowth which we will call the triangle. It is a form of the Decapod appendix masculina of Boas. The triangle stands up dorsally so that at rest, it, with its fellow of the other side of the body, fits into the squarish cavity left between the two necks of the first stylets. It is not well seen normally from the ventral view, (fig. v) but it may be pulled outwards through 90° and then looks as in the median view (fig. vi). It is a flat triangular outgrowth, partly calcified and partly membranous. The edges are calcified and the centre membranous, so that the whole suggests a bent arm or wing with skin stretched across it. Each long side of the triangle is about 3 and


the shorter base about 2 mm. The bony rhns of the triangle as seen in fig. vi may be called the humerus and the radio-ulna.

The distal free part of the apparatus, ,(figs. vi, vii) is a trihedral mass set with long plumose setae and might be likened to a sort of hand at the end of the fore-arm. We will call it the wedge from its appearance and use as seen in sections (fig. 9).

The humerus articulates at each end; proximally loosely with the side of the exodopodite mass, (fig. vi) ; distally at the elbow, firmly with the other firm edge of the triangle, the radio-ulna. On the external or concave face of the triangle, (fig. viii) the humerus is not as well separable from the membranous part of the triangle, and between its proximal end and the bone of the main mass of the endopodite there is more or less expanse of membrane. On this outer face, (fig. viii) we find that all the concave aspect of the triangle is membranous.

The humerus is wide and smooth and flat on the inner face, fig. VI, but on the outer face forms only a narrow edge to the membrane, fig. VIII,

The soft hollow face of the triangle in life is swollen with contained liquid. The soft area is not only the outer face of the triangular protuberance but also half of the dorsal face of the distal part of the main trunk of the endopodite.

The whole darkened area of fig. viii might be compared to the soft inside of the palm of a hand and it is this which comes against the neck of the first stylet, in conjugation.

While the humerus is wider toward the base and slender at the elbow end, the radio-ulna is the reverse; that is, it begins narrow at the elbow and widens to the hand or terminal part. . The radio-ulna is a thick plate-like mass that is not in the same plane as the humerus, but about 45° with it, so that it has the appearance of a scroll rolling in over the depressed membranous outer face of the triangle, (figs, vii, viii). The radius part is the free rounded edge, (fig. viii) and this ends abruptly opposite the base of the hand, which is back of it in the figure, while the ulna plate runs on continuously in the background of this figure and passes imperceptibly into the hand, or wedge, (fig. vii).

The radius stands free, away from the membrane, as a rounded


bony ridge much thicker than the ulna plate from which it is faintly marked off by a suture. Thus in sections (fig. 8), the radius looks like a head on a slender neck. The abrupt termination of the radius is very like the elbow end of the human radius, a shallow cup. The actual cup is made by clear horny matter of considerable thickness and is prolonged as a horny sharp ridge all along the radial edge of the pyramidal wedge. The head of the radius stands out as wider than the neck (fig. vii).

The ulna is but a vaguely defined thick area of the general shell and it continues as the hand or wedge, which is, next to the head of the radius, the most peculiar part of the triangle. This wedge is a hard horny pyramid of three faces. One is rounded and setose, two flat, meeting at a sharp edge, (see small sketch, fig. 22). Its exposed rounded face (figs, vi, vii) is set with a dense brush of plumose setae. The external or ulnar face (fig. vii) is smooth bone, bearing setae along its right edge and ending, to the left, in the sharp horny ridge that runs up from the head of the radius and is shown as a dark shade in fig. v. The concealed innermost face is bony and contains orange pigment; along its left edge it bears setae (fig. viii), and its right edge is the sharp horny membrane that runs up from the head of the radius. In the union of this face with the soft membrane of the concavity of the triangle there is a bony articular plate.

The photographs do not represent one feature of the triangle and that is the small tuft of some five or six, or so, very wiry bent plumose setae that spring from the elbow of the triangle and, for the most part, curve so as to lie down close to the soft membrane. These setae are roughly shown in fig. 22 at the elbow. This also gives in the side sketch, an end view of the head of the radius as seen when the base of the wedge was cut off and the stump of the ulna and free end of the radius viewed from the face where the wedge had been. This is intended to show the head of the radius as a rounded saucer with flat bottom, not deep, but with flaring and rounded sides that form a rim thicker than the neck of the radius below. The cut off setae in this figure are the bases of those on the union of ulna and wedge, just above the level of the line 23 in the main fig. 22.



Dissections and sections showed the presence of the same general structures as in the case of the first stylet, with the important difference that the special glands of the tube of the first stylet are absent and on the other hand the intrinsic muscles that are absent in the first are well developed in the second stylet. The muscles are arranged as in the younger stages (figs. 27, 28). Besides the three muscles at the base that pass into the sternum of the second abdominal somite a very short distance there are long muscular strands within the stylet itself.

The protopodite springs from a considerable elevation of the sternum and in the adult two muscles were found within this elevated articular region. Pulling one tended to depress the stylet into its position of rest while the smaller muscle was thought to be probably concerned with the erection of the stylet. Pulling all the basal muscles made the stylet not only lie down but also move toward the median line, which would enable it to fit in nicely with the first stylet. Some of these extrinsic muscles extend a distance into the protopodite itself, to be attached to the shell. There are also long strands arising from the shell of the protopodite and running to the exopodite and the endopodite. Those of the exopodite seem associated with the basal muscles, so that pulling the muscle in the sternum made slight twitching movements of the exopodite, simulating those seen during conjugation, which may thus be caused by contractions of the muscles that hold the entire appendage in position. Pulling the muscles that are in the distal part of the protopodite made both exopodite and endopodite move dorsally and also away from the median plane.

The muscles that move the exopodite are better developed than those of the endopodite. Within the exopodite there is a long intrinsic muscle that would seem fit to bend the slender filament slightly. Inside the endopodite, beside the slight muscles of the base concerned with the movement upon the protopodite, there are in the adult two slight threads that represent the muscle seen in early stages (figs. 27, 28) passing from the terminal flagellum down



into the region whence springs the triangle. These muscles are seen in the sections of the triangle (figs. 23, 24) as two black dots. These sections, with those in figs. 8, 9, show the anatomy of the triangle. Fig. 23 is a section along the Une 23 of fig. 22. The great thickness of the calcified shell is shown by the black mass. The membranous parts are shown by the thin black, as to the left in fig. 24. The cavity within is blood space traversed by connective tissue strands and faced by epidermis against the shell. In

Fig. 23 Cross section of the triangle on the level 23 of fig. 22. Fig. 24 Cross section of the same about the level 24 of fig. 22.

90 mm. A. 90 mm. A.

fig. 23 the triangle and the distal part of the protopodite are cut across with the hollow face to the left. The dense shell mass to the left above is the guide ridge, (fig. viii) which somewhat overhangs the cavity of the triangle and bears on its median face some setae, (fig. vi), which are connected at the root with the epidermis by the long canals of which one is seen to the left (fig. 23) penetrating the shell. Opposite this on the median face of the endopodite there are also a few setae which do not appear in the


photo^ruph (lig. v) but present one of their canals in the shell of lig. 23, to the right. In contrast to the excessive thickness of the shell of this main stem of the endopodite, the triangle, as represented by the lower part of this section, is relatively thin shelled. The radius is th(^ thick knob in the lower left corner. The shell to the I'ight is the ulna, the thick mass against the concavity of the triangle is in reality more membranous than calcified, but as yet thick. But further toward the elbow (fig. 24) along the hne 24, (fig. 22), the corresponding region is a thin membrane reaching from tlu^ n(^ck of the radius across to the thick guide ridge. In reality the elbow stands out more as in fig. viii so that the width of the section 24 is nmch greater than fig. 28. Fig. 24 shows clearly, on the riglit, the hinge-like line of demai-cation between the outstanding triangle and the main stem of the endopodite, being in fact cut at the edge of the proximal articulation of the humerus (fig. vi), wliere there is a sudden change in level in passing from the humerus to the main stem. In sections 23 and 24, the small black dots above within the connective tissue, are the muscles that iim up int-o the flagellum, much as hi fig. 28.

Section 8 shows the radius standing out from the flat triangle with the thick mass of the humerus above in the figure, wliile fig. 9 shows the thick end of the endopodite above and in the groove of the first stylet the cut off wedge, as will be described below in considering the :idjustm(Mits of the first and second stylets during conjugation.


Between the individual development of the first and the second stylets there is this imi)ortant diffeivnce that while the first never at any time looks like one of the ordinary pleoiK)ds but is of late appearance and is also a dwarfed, specialized, or reduced appendage from the first, the second appendage is present as soon as the others are and is at first like the ordinary appendage and becomes specialized by the addition of an outgrowth and not by the loss of parts.


The plcopods of the second, third, fourth and fiftli somites of both males and females are represented at the time of hatching and all alike have the appearance seen in i\g. 2() whi(;h is magnified 75 diameters and represents the anterior face of the third left pleopod of a male 18 mm., in July, when in the fifth larval stage. The pleopod is flat and translucent; the endopodite (En.) is longer than the exopodite {Ex.) and both are fringed by long setae that are really plumes, though not so figured. Both endopodite and exopodite are obscurely joint(Hl and the proto]x)dite has a short annular segment as well as a long main segment. Through the thin shell may be seen the muscles, represented by the dotted lines. At the base are three large and one minute muscles; two of the main three are posterior and one anterior, and apparently the movement of the entire appendage would be a more powerful backward swing and weak forward recovery, as in swimming. Within the main segment of the protopodite are three long muscles that would seem to aid in bending the appcnulage at its base, while distally there are two muscles which both go to the exopodite to move it. The endopodite is left with only intrinsic muscles to move it at its base and with a long branched muscle that can act only to bend the endopodite itself. Th(^ exopodite has also intrinsic muscles at its base as well as i\w musck^s of the protopodite to move it. There is likewise a long branched muscle to bend the exopodite.

In the early stages the second appendage of the male is quite like this third pleopod, but in a male of 21 mm. (probably in the same larval stage as the male having the third appendage shown in fig. 26) we find th(^ pleopod of the second somite modified as in fig. 27, that is, there has been added to it the excrescence seen on the median side of the endopodite. This is to become the triangle or appendix mascuUna of the adult.

The first discovered trace of this outgrowth was seen in a larva of the fourth stage, 11 mm. long, in July. This first beginning of the triangle is the slight elevation {x) seen in fig. 25, on the side of the endopodite. This figure represents only that part of the endopodite which is not well jointed and forms a sort of base beyond which is the more flabelliform distal part, (fig. 26). It

jotniNAi, OK Moupiioi.ofiy, vol,. 22, NO. 2

will be noted that the row of plumes on the right, or median side of the endopodite (fig. 25), is interrupted distally so that there is a blank space where one would expect one or two setae, and in this space there protrudes to the right a rounded elevation. The position of this slight elevation with reference to the muscles leaves

Fig. 25 Posterior view of basal part of the endopodite of the accessory stylet of a male 11 mm. long. Enlarged 215 diameters.

no doubt that it is the same thing as the larger elevation of the next larval stage (fig. 27) . In the preparation the epidermis, not here shown, grew out to form this elevation as a hollow outgrowth, leaving no question as to the possible artificial nature of the bulging of the cuticle shown in fig. 25.



In the fifth larval stage (fig. 27) the protopodite has become wider and stouter and the basal part of the endopodite is much expanded distally where the protuberance arises from it. The

26 27

Fig. 26 Anterior face of third left pleopod of a male 18 mm. long. 2. D. Fig. 27 Anterior faoe of left accessory stylet of male 21 mm. long. Enlarged 75 diameters.

result is that the exopodite begins to take on that relative insignificance in size, characteristic in the adult accessory stylet.


The new growth on the median side of the basal part of the endopodite (fig. 27), is a sort of knob set on a neck and indined at about 45° to the axis of the endopodite. Its form is not spherical but rather more that of a short cylinder on a slightly shorter neck. The long axis of the cylinder and of the neck is at an angle of 45 degrees to the side of the endopodite. Not only this protruding knob must be reckoned as part of the future triangle but also the neighboring widened area of the endopodite which is depressed as indicated in the shadow in fig. 27 and which will be the depressed anterior face of the future triangle. In fact this depression is accentuated by the position of the knob, which not only stands out as represented in the figure but also rises up toward the observer; that is, anteriorly away from the general plane of the endopodite. The base of the flabeUiform distal part of the endopodite is continued on to the external distal corner of the basal region of the endopodite as a ridge standing up above the depressed area, and forpiing what will be the guide ridge of the perfected organ.

In a small male, 38 mm. long, in October, the second pleopod had advanced to the state of perfection shown in fig. 28, which is an external view of a left accessory stylet, which was about three times as long as the one shown in fig. 27. The muscles in the protopodite remain as before, though not so well seen from this point of view, and the same is true of the endopodite and the exopodite. The protopodite and the exopodite have grown so large and massive that the slender exopodite is much subordinated. The great increase in the basal part of the endopodite, along with the enlargement and specialization of the triangle, leaves the plumose terminal part of the endopodite as a slender palp-like remnant of the original end of the endopodite. The triangle is now so much longer at its free edge than at its attached part that it has the adult triangular form when seen from the median face ; or more explicitly, the obliquely set cylindrical knob of fig. 27 has grown so much longer at its free edge than at its attachment that the length between its ends about equals the distance of the proximal end or elbow from the main mass of the endopodite, which



29 28

Fis. 28 External face of left accessory stylet of male 38 mm. long in October, enlarged 25 diameters.

Fig. 29 External view of the united first and second stylets of the left side of an adult male, 110 mm. long. 2. 90 mm. Oq.

Fig. 30 View of the median face of the same. 2. 90 mm. Cq.

Infigs. 29, 30, 31: I| = 1st stylet II = 2nd stylet 5 = crossed fifth leg seen in section

5' = other fifth leg not crossed. C =canula Sp = spatula.


gives the wide scalene triangle as seen from the median side (fig. vi). The proximal elongation of the cylinder makes the elbow of the triangle, while the distal elongation has made the pyramid or wedge that runs up toward the flagellum of the endopodite. As yet no setae were seen on the wedge. The triangle, however, is not merely a flat plate that grows out diagonally, but from the first it is thick through in the anterior-posterior direction, thus producing the cylindrical edge seen in fig. 27, where the thick edge is restricted and marked off by a less thick neck; moreover the thickening of the cylinder is toward the anterior face. By the stage shown in fig. 28 there is great thickening toward the external face. Moreover the external free edge of this thickened cylinder is now itself thickened as a ridge hanging out from the ventral rim over the depressed area as indicated by the broken line in fig. 28. This rounded thick edge is the future radius. (Compare figs. 28 and viii.) From this state it is an easy transition to the more sculptured form of the appendage seen in adults.

The second pleopods of the male thus owe their special structure to a gradual emphasis of the endopodite and protopodite with the addition of an outgrowth peculiar to these appendages, the triangle. The triangle at first is a mere blister on the median side of the endopodite but soon becomes an oblique plate that is surmounted by a thickening. The plate grows anteriorly and the thickening of its free edge becomes longer than the base of the plate, with a resulting triangular form as seen from the median face. The thick ridge grows out externally and this extension itself acquires a thickened rim, posteriorly, which is the radius.

The triangle is thus a triangle only as seen from the median face of the pleopod, in its entirety the triangle is a curved object like a half open hand, and as such is capable of being applied to the rounded surface of the first stylet. It is made of a cylinder obliquely set along the edge of a plate and curving over it, like fingers over the palm. A slip of paper if cut of angular form and bent twice at right angles may be made to represent the stylet.



The way in which the various parts of the stylets are used in the process of conjugation and sperm transfer has been found out partly by direct observation, partly by experiment, and partly by more indirect inferences that still leave some questions unanswered.

The phenomena of conjugation in general have been described elsewhere (5) and we will here consider chiefly the use of the stylets. There is a stage in the early part of conjugation, where the male has seized the female and clasped all her claws, when he rises up away from her sufficiently to allow the pleopods to swing back and forth. In this swinging the long stiff stylets and accessory stylets take part and then are soon locked together, after which the stylets are held by the crossed fifth leg so that henceforth they make a rigid mass which cannot be folded down against the thorax again by any pressure until that fifth leg is removed. The process of locking together of the stylets is as follows :

The swinging of the pleopods is caused by their basal muscles ; and likewise the muscles in the bases of the stylets move them slightly backward, or erect them, and forward, or depress them. While both first and second generally move together and right and left alike, they have been seen to move independently. By a special movement of the second stylets they are clasped against the first in such a way that the triangle is applied to the neck of the first stylet. By arching the abdomen, cat-like, the second stylet is drawn up dorsally along the first, and then, by partial relaxation of the arch of abdomen, the second is shoved distally, along the first, while held tight against it; the result is that the wedge glides along in the groove of the stylet and the radius enters into the inner tubule through the flaring orifice and is shoved in so far that it remains fast. In sections (fig. 8) it is seen that radius fits into the groove as in a socket and, all the walls being thick and solid, the radius cannot be forced out again without running it back along the orifice. The fact is that the locking is very firm and when one tries to pull the second stylet backward the first is dragged with it and only by pulling the second dorsally toward


the base of the first can one separate the two, as by that means the radius is brought to the orifice out of which it readily passes.

When the two stylets have been erected by their own erector muscles and locked together by their muscular movements which lead to this mechanical fastening of the edge of the triangle within the groove, they form one organ, physiologically, which is to transfer the sperm without any further muscular activity within it.

The appearance of the two locked organs is indicated in the somewhat diagrammatic sketches 29, 30. In 29 the external view of the left stylets and part of the fifth thoracic legs is shown. The second stylet, to the right of the figure shows the solid tip region of the endopodite applied closely against the most protuberant part of the posterior face of the spiral of the first stylet, while the terminal flabellum runs along parallel to the canula and spatuula. In fact .the tip of the bony endopodite seems to overlap the contours of the spiral and this is due to the soft nature of the depressed region of the median face of the end of the endopodite as is seen in fig. viii. The guide ridge is the part seen external to the spiral in fig. 29, while the soft surface is squeezed against the rounded face of the spiral and the triangle is applied close against the median face of the spiral so that it can be seen only from the median view.

Turning to the median view we see, (fig. 30) the triangle lying over the neck and extending out along the groove. The elbow of the triangle lies over the orifice. The radial edge of the triangle conforms with the obliquity of the groove since both the wedge and the radius are firmly inserted in the groove.

Figures 29 and 30, show the supporting fifth leg in section, as a rounded cross-hatched area. It will prevent the locked stylets from being shoved forward, or closed up against the sternum anteriorly. It is also obvious that the movement backward toward a vertical position will be hindered, not only by the inclination and rigidity of the basal joint of the first stylet, but by a like joining of the base of the second stylet, since one cannot move back without the other, for the radius and wedge will go no further


toward tip of groove. The second forms a mechanical brace tending to hold the first from going backward.

In order to separate the two the second must move toward the animal and glide along the first till free from it. And this motion is actually seen. The locking is not always done without trial and may be broken and renewed during conjugation, so that we often see two positions of the stylets, that of perfect locking,

Fig. 31 Same view when the accessory is drawn back into position of recession showing the papilla at the mouth of the groove.

as in figs. 29 and 30 when the triangle is most advanced toward the tip of the spiral, and a preliminary and alternate position of recession when the triangle is applied against the base of the first style proximal to the orifice. This position of recession is shown in fig. 31. The triangle goes as far toward the basal end of the first stylet as possible, till stopped by the knob on the base (fig. ii). In this recession the orifice with the papilla meeting it, is exposed and the ventral lip is seen.

It should be borne in mind that the back and forth play of the triangle on the first stylet is limited not only by the knob basally and the narrowness of the groove that prevents the radius from going into it dorsally beyond the position of figure 30, but it is limited laterally by the fact that the triangles of the two sides


of the body are in contact and are held together by being placed in the squarish hole between the necks of the first two stylets.

The two triangles play back and forth like two hands with bent fingers, back to back, in a narrow space between the first stylets and, like hands, each runs its palm or soft flat surface along the median constricted part of the first stylet and the firm guiding ridge — its thumb, as it were — along the external face of the stylet (fig. 29). In one case from 3 to 4 seconds were taken to glide the triangles back from the normal position to the recession (fig. 31) ; there they remained four or five seconds and advanced strongly in two seconds. Another recession took 12 seconds, but the advance occupied 2 seconds.

If we imagine figure v apphed to i, vi to ii and vii to iii, VIII to IV, we will appreciate how nicely all the surfaces adjust themselves. The oblique ridge of the external mass of fig. i is overlaid by the soft depressed area, (figs, viii, 22) so that the thumb-like guide shows external to the ridge as in fig. 29.

In life the two sets of appendages, right and left, are so closely applied together that the median face of neither can be seen, directly, without mutilation experiments on one side, but the presence of the guide ridge along the external face of the spiral (fig. 29) enables one to judge where the triangle must be at any stage of advance or recession, a matter of importance in deciding as to its use in sperm transfer.

That an application of the second, or accessory stylet, to the first is necessary for the completion of normal conjugation and the filling of the sperm pocket by transferred sperm, was determined not only by the above facts of structure and use but by the following experiments. The instincts of the male are so strong that, when in the process of conjugation the second stylet on one side was cut off, there was no immediate visible effect, except the escape of some blood from the stump of the appendage. And when on the next day all the stylets, both first and second, were cut off, the male seized and turned a female and carried the conjugation as far as possible in the absence of the organs of transfer. The instincts thus go on without the means of carrying them to completion.


It was then easy to get males to begin conjugation when the accessory stylets had been removed from both sides. Three such males made conjugation experiments with several females, successively, but in no case was there an evidence that the annulus had been filled by these mutilated males, through in one case the union lasted for eight and one-half hours. In these attempted conjugations it was not evident how the absence of the second stylet prevented perfect sperm transfer. In one case the male let fall three or four sperm masses, or pseudo-spermatophores, about 1 mm. long on the telson of the female but it was not determined how this happened. Apparently this was from failure to have a close union at the orifice, which would lead one to think the failure due to absence of the triangle that normally holds the papilla tight to the orifice. But the failure may have been due to the absence of piston like movements of the radius. More experiments should show what the uses of the different parts of the triangle really are.


The adjustment of the papillae, whose anatomy has been described, to the stylets must now be considered in order to appreciate the final use of the stylet.

As seen in fig. 1, the papilla juts out toward the median plane so far that it can be placed across the narrowest part of the first stylet where the notch is (fig. in); that is across the dorsal face of the first stylet. But its tip turns abruptly inward far enough to reach along the median face (fig. ii) as far as the orifice, into which its tip fits. In figs. 30, 31, this position of the papilla is crudely represented; in reality the tensely swollen translucent spout is very nicely applied to the rounded faces of the entrance .to the groove. The papilla is seen in this position when the triangle is receded (fig. 31) and in the advance of the triangle its tip becomes concealed, but it doubtless remains as before.

Returning to the actions of the combined stylets which embrace the papillae we note certain 'tamping' movements. Besides the advance and recession of the second stylet along the first, the first and second together when locked, are seen to execute quick jerks


that carry the tips of the first back and forth a part of a millimeter only. \Yhen the tips of the stylets have gained entrance into the annulus, these thrusts may serve to introduce the tip farther into its cavity. As in the movements of recession the force here must be exerted by the muscles of the abdomen, as the stylets themselves have no telescopic power; and actual twitching of the anterior part of the abdomen were seen.


In normal conjugation nothing is seen of the sperm so that its transfer from the deferent duct to the cavity of the annulus is a matter of inference. The papilla is applied to the orifice of the tube of the first stylet so that it may discharge into it and sections show the tube full of sperm, (figs. 14, 15) ; moreover in some abnormal cases the sperm is seen to issue from the tip of the canula into the water, and, as the tip of the canula is normally inside the sperm pocket, it is evident that the sperm must pass along the stylet from the papilla. The force that propels the sperm is no doubt muscular contraction, but it is not clear at first what muscles are concerned; there are none within the first stylet which acts merely as a passive tube.

From such figures as 2, it is evident that the deferent duct has powerful transverse muscles that could squeeze out the sperm with force and this seems the main if not only motive force to carry thie sperm through the papilla and all along the tube of the stylet into the annulus.

The force necessary to propel the liquid sperm through a tube that is only some 20 to 40 ^ in diameter (figs. 13, 15) is great and attempts to force india ink through the tubule of the stylet with a small hypodermic syringe failed. When the specially ground canula was inserted into the orifice, while the radius was engaged in the tubule, no ink could be forced out of the tip of the stylet. It was inferred that the radius blocked the way, as it fits in so as to nearly occlude the lumen (fig. 8) , but the same failure was met with when the triangle was removed from the stylet, but then the ink jetted out along the proximal part of the groove where the


triangle had been. Apparently the wedge of the triangle is well fitted to hold the liquid in the tubule since it fills up the groove external to it (fig. 9), where the sides of the groove are not as close together as they are distally (fig. 10), which is beyond the wedge. When the ink had been introduced into the tubule and not forced out of the tip of the stylet the triangle was applied to the stjdet and the radius worked back and forth like a piston in the tubule with the result that some of the ink issued from the tip' of the canula of the stylet.

This suggested that the radius might act like a piston in normal sperm transfer and thus propel the sperm from the papilla along the tubule to the annulus. We also saw that when a pair was separated in conjugation the sperm that issued from the tip of the canula of the stylet was mixed with bubbles of air when held out of the water, which suggested some action at the base of the tubule (at the orifice) to draw the air into the tubule. However, this might be movements of the triangle or simply failure of the triangle to hold a tight joint around the tip of the papilla and orifice, for thus air could be drawn in by the stream of sperm advancing, driven by pressure of the muscles of the deferent duct. When the radius was inserted into the orifice and shoved along in the tubule, sperm was forced out of the tip of the canula, which seemed to demonstrate the ability of the radius to act as a propelling piston.

We failed to detect any such piston motions during conjugation, but they would be of very slight extent and not readilj'^ observed. The movements of advance and recession described above are of a much grosser magnitude than the piston movements that might be supposed to take place. The movements 31, 30 are only for getting right adjustment of the enveloping triangle over the papilla tip and the entrance of the radius into the tubule so that the hand-like triangle may make such tight binding of the papilla to the orifice that no sperm escapes or comes into contact with the water. Yet the piston may then presumably be in position to advance or recede a little. When we thrust the triangle strongly so far along the stylet that the elbow was at the orifice, (fig. 30), the triangle tended to spring slowly back out of the groove


till only half of the length of the radius remained in the groove, owing apparently to the elastic side walls of the groove shoving against the wedge (figs. 9, 10) as these walls are the closer together toward the tip of the stylet.

By this mechanical means the piston might tend to recede, while the movements of the muscles of the abdomen might make the entire second stylet advance enough to shove the piston along the groove again. We can easily pump the radius back and forth in the groove by moving the whole second stylet. The muscles of the abdomen make the slight twitching back and forth jerks of both first and second stylets above mentioned as tamping movements. Now after the first stylet, with the second locked to it, is introduced into the cavity of theannulus as far as possible, these movements of tamping, if they be continued, could not advance the first stylet but may push the second further along the first and so cause the piston to act on the sperm. The dish-like head of the end of the radius (fig. 22) receives explanation upon the assumption that it is useful in shoving the sperm along in the tubule, in fact, the solid bone-like piston with horny cupped tip provided with elastic flaring edge seems a remarkably well made apparatus for pushing liquid along in a tube that it fits so well.

Some such piston movements might be expected from the statements of Coste (C. R. 46, 1858), that Gerbe in his laboratory saw the male Astacus apply the foliacious part of the second stylet to the first stylet and by reiterated back and forth motions during the passage of sperm, keep as he thought, the trough of the first stylet free from sperm that might harden there else. Schillinger, states that the second stylet is used to push the spermatophores out of the first stylet.'

The groove and its concealed inner part that forms the tubule are of course open to the water and if the sperm is to pass free from contact with the water to the cavity of the annulus the assumed piston movements of the radius may serve to clean out the tubule and fill it with harmless secretions. The source of such secretions may be surmised to be the glands in the tip of the pap ' As reported by Ortmann in Bronn's Klassen und Ordnungen.


ilia (fig. 2) or those along the tubule itself (fig. 8). Possibly this preparatory action of the radius is all that it has to fulfill and that the pressure of the muscle of the efferent duct is all sufficient to cause the sperm to run through the length of the stylet. In connection with this question we have to bear in mind that the sperm is in some way freed from its envelope of secretion made in the efferent duct before it is laid away inside the sperm pocket where it exists pure (1).

This separation of sperm from enveloping secretion takes place in the tubule of the stylet. In the proximal part of the tubule the secretion of the deferent duct (fig. 2), is still all around the strand of sperm (fig. 14), but distally the sperm is almost pure inside the tubule (fig. 15).

We found also that in one case a male, fallen on the side while still holding a female, had the stylets only partly erected so that they were free in the water and from the tip of each canula a very fine stream of sperm, finer than the tip of the spatula, issued slowly and coiled up in a small mass. From one canula the sperm then slowly sank in ten minutes down in the still water as a fine thread with a coil at the tip. Another male showed faint sperm jelly on the tip of the flagellum of the endopodite of the second stylet and this was pure sperm becoming modified by the water; there was no secretion.

There are however besides these escapes of pure sperm, escapes of sperm inside of secretions that resemble spermatophores. In a male, in which the triangle was in the position of recession, (fig 31), there were such white sperm threads, | to 1 mm. long, about the orifice of the groove. The pseudo-spermatophores that in abnormal or interrupted conjugations were sometimes seen were soft, paste-like tubes containing a central mass of sperm. The short pieces of tube stick by their ends to the inside of a pipette used to pick them up and to the shell of the crayfish on which they fall.

The wall of these tubes is a very thin layer of secretion which is vesiculate and stringy like dough and can be drawn out into clear threads with minute droplets along them. These would seem to be not normal spermatophores, which in Astacus have


thick walls, but only rods of sperm enveloped in some slight secretion from the deferent duct, or possibly that of the papilla or of the glands of the spiral. The thin walls of these tubes break open, hernia like, and sperms ooze out.

The separation of the sperm from the secretion of the deferent duct may be due merely to the diminution in diameter of the tubule; the pressure of the duct would drive the central part of the current faster than the envelope and thus the central sperm might flow out of the very narrow tip of the canula and leave the envelope of secretion behind in the wider parts of the tubule. Finally, when enough sperm had passed along to fill the annulus, the enveloping secretion might be forced out and this would make that wax-like mass that fills the external parts of the annulus and projects in excess from its mouth as the so-called sperm plug. Possibly again the piston movements of the radius might come into play to clean out the secretion from the stylet tubule and ram it into the annulus. In fact in the last stages of conjugation of one pair slow and repeated movements of advance and recession of the triangle were seen which may be interpreted as concerned with plug making.

The use of the glands of the spiral is not known. Possibly their secretion cleanses the surfaces to be used in sperm transfer and aids in keeping water from the sperm. Possibly the secretion may help the enveloping secretion of the deferent duct to adhere to the walls of the tubule of the spiral and thus hold it back till the sperm has passed on into the annulus.


The striking fact of the exact duplication of both first and second stylets right and left suggests questions as to the use of right and left in conjugation. Are both sides used at each conjugation?

Again the remarkable dimorphism of the females of C. affinis and of probably all other species of that genus, which expresses itself in the occurrence of females with the vestibule of the sperm pocket opening a little to the right of the middle line and of females with the pocket opening to the left, so that the symmetry


of the two i.s reversed, raises the question as to whether the males are adjusted, in habit, to these two kinds of females, so as to use the right set of sperm transfer organs for a left-handed female and vice versa, or not.

The crucial experiments to determine whether males actually use the one stylet for right and the other for left-handed females have not yet been made. However, some facts and considerations make it improbable that a male is obliged to do so and indicate that a male may adjust his stylets so as to use either right or left on any form of female annulus, leaving the question still open as to what is the normal habit of the males with reference to the two forms of females.

In the first place we found that though the two first stylets seem to be in the annulus they are never both firmly inserted. One is fixed firmly by its tip while the other may be drawn away by a pair of forceps. Moreover the one that is inserted has its tip some f to If mm. in advance of the other and its base is locked against the base of the other, diagonally, the abdomen being advanced more on one side than the other.

Observations showed that not only were there cases of the right stylet in the left annulus but of right stylet in the right annulus and of left stylet in left annulus and of left stylet in right annulus. Whether in these cases the sperm was actually transferred was, unfortunately, not made out. It is possible that a male may insert one stylet and afterwards the other till finally the actual sperm transfer takes place with some more definite reference to the symmetry of the annulus than the above observations would indicate.

That there is any alteration in the advance of the stylets was not made out, but there is often an alteration in the use of the fifth leg, right and left. At any one time many males will be found with the left and others with the right leg crossed, but continuous observations show that the male will change from right to left in difficult cases especially, till a better adjustment is obtained.

It was at first thought that there was a relation between the fifth leg and the advance and use of the first stylet so that these were on the same side, that is, the stylet being advanced by the



use of the leg of that side of the body, but cases were recorded in which the advanced stylet was on the opposite side from the crossed leg. Males crossed the right leg with either right or left stylet advanced and males crossed the left leg with either right . or left stylet advanced. Here again there is the possible objection that the condition observed was not permanent or the one employed in actual sperm transfer. More minute observation of several normal cases are necessary.

One good case seems, however, rather conclusive. In this a male, in November, crossed the left fifth and advanced the left stylet, but after an hour of attempts to enter the annulus, crossed the right fifth and five hours later the right stylet was one mm. in advance of the other and the female had a sperm plug in a right annulus. Here the leg and stylet used did coincide, but the annulus was not the one to be expected.

Again in some conjugates killed by boiling while united it was found that in one a right stylet was advanced to a right annulus and in others a left stylet to a left and to a right annulus.

As far as the evidence goes it gives the impression that the male is free to use either right or left stylet with either right or left fifth legs till successful in getting some one tip of the stylets into the vestibule of the annulus, which may be a right or a left one, indifferently. Yet future observations may show that the lines of least resistance are for the male to use the left stylet for the right-handed female, and the reverse, and that this actually takes place, in nature as the normal, though we doubt if it be at all necessary. Observations show that both papillae are ready to discharge sperm at the same time and it should be determined by experiment whether the male uses both right and left sets of sperm transfer organs, alternately, at each conjugation or not.

When the first and the second stylets were cut off from one side of some sixteen males and, either at once or some weeks after, these males were given females, the unexpected result followed that in spite of many repeated attempts, one lasting nine hours, the numerous conjugations of these unilaterally mutilated males did not result in any clear cases of successful sperm transfer. In


many cases the annuli of the females were artificially cleared out so that any new plugs would have been seen.

Among these cases there were males that alternately used the fifth left and right legs in crossing, though some had only the left series of stylets and others the right; the leg being crossed on the side where there was no stylet and on the side where there was a stylet. And these same cases were attempting conjugation with females that were of both kinds, right and' left forms, so that there was no agreement between the kind of annulus and the fifth leg used.

In only one case was there any sperm seen and this was seen twice in successive conjugations of the same male that seems to have been peculiar. This sperm lay in pseudo spermatophores, 8 mm. long, upon the telson of the female under the left stylet, and probably escaped from some imperfection of the closure of the triangle.

While it was not found out why there was this apparent inability to complete sperm transfer while the stylets of one side were missing it is thought that this is not due to the need of using sperm from both sides of the body at each conjugation but rather to the mechanical factor that the two sets of stylets are always applied to one another so firmly as to hold the tips of the stylets at the annulus, so that when one is absent the tip of the remaining one lacking the usual support cannot be readily brought to the middle line of the body. Moreover it is possible that the triangle will not be well applied to the orifice unless the fellow triangle be there to shove against it, as both are packed in side by side between the necks of the first stylets.


Though the sperm of the crayfish, Cambarus affinis, is injured by exposure to water, it is transferred from the male to the female under water and stored up in an external pouch.

The part played by the female in this insurance against injury in transit has been elsewhere described.


The present paper describes only those organs of the male that are combined to form a safe conduit for the sperm from the male to the receptacle on the female.

The actual sperm transit apparatus of the male consists of three organs on each side of the body. The anatomy and use of these three organs are here described in detail.

The 'papilla' or end of the deferent duct is provided with glands and a valve. It is distended by blood and applied to fit accurately to the beginning of a tube.

This tube is the innermost part of the groove of the first stylet, or limb of the abdomen, and hitherto its existence and use has not been described.

The first abdominal limb is, in action, a duct leading the sperm uninterruptedly from the deferent duct into the receptacle of the female. It contains large glands of problematical use, and relies for mechanical support upon the habit of the male in using the second abdominal limb as well as one of the fifth thoracic limbs to insure the entrance of the first stylet into the receptacle of the female.

The second stylet is accessory to the first in applying its handlike outgrowth over the papilla and insuring a tight joint. It also gives mechanical support to the first stylet. How much it may also serve as a piston for cleaning the tube or even for aiding in sperm transfer is left undecided.

The ontogeny of the first stylet shows that it begins after the other abdominal limbs and is from the first a simple unbranched outgrowth which becomes a tube by the depression of its central and elevation of its lateral parts to form a deep groove, the bottom of which is ultimately isolated by a shelf.

The morphology of the organ, based upon its use, anatomy and development, gives the basis for its utilization in defining species and subgenera. The tip or canula that is inserted into the receptacle to discharge sperm is the real tip of the organ and all other tips are to be referred to lateral outgrowths from one or the other side of the original groove.

The ontogeny of the second stylet shows that in the first larva it is just like the following abdominal limbs; but its subsequent


fate is to add on a lateral outgrowth {appendix masculina) wliich becomes the useful part of this organ when acting as a necessary part of the sperm transit apparatus.

The duplication of all three organs, right and left, seems necessary in, as far as removal of one set leads to the lack of necessary mechanical support for the perfect functioning of the opposite set.

The evidence is against the conclusion that the right and left openings of different receptacles upon different females are necessarily met by the males employing the stylets of one side rather than an other. In each case the male may by trial obtain the entrance of some one of the two stylets into the receptacle of the female.

The extreme solidity of the shell of the stylets is to be correlated with the amount of force exerted by the male in making a water tight passage for the sperm from the deferent duct into the receptacle of the female.

While all six organs are necessary for sperm transfer, most of them may be removed without preventing the males from carrying out many of the stages of conjugation that would normally lead up to sperm transfer.

Many of the peculiarities of the form and structure of the transfer organs are demonstrated to be of use, or even necessary.

The accurate interadjustment of the six organs is necessary for the perpetuation of the species.

It is difficult to believe that in the evolution of Cambarus the increasing perfection of these organs could have been decisive in eliminating the less perfect organs. Astacus survives with more simple organs and the majority of genera of crayfish. have no stylets at all. The perfection of the organs, characteristic of Cambarus may have been brought about from laws of change that it will require much experimentation to discover.




1 Andrews, K. A. U)()B The luumluw vcntralis. Proc. Boston 8oc. Nat. Hist.

vol. 32.

2 1908 The aiinulus of a Mexican crayfish. Biol. Bull. vol. 14.

3 1908 The sperm receptacle of the crayfishes Cambaius cubensis and C. para

doxus. Proc. Wash. Acad. Sci. vol. 10.

i 1904 Breeding habits of crayfish. Am. Nat. vol. 38.

o 1910 Conjugation in the crayfish C'ambarus^affinis. Jour. Exp. Zoo!, vol. 9.

6 1904 Crayfish spermatozoa, .\natoin. .\nz. vol. 2h.

PLATES 1, 2. 3, 4


I. Photograph taken with a magnification of about ten diameters, of the posterior face of the first stylet of the left side.

II. Photograph of the same, taken from the median side, but diagonally, so that the posterior side is also shown in part.

III. Photograph of the same from the anterior face. I\'. Photograph of the same from the external face.

V. Photograph taken enlarged about ten diameters, of the second, or accessory stylet, of the left side of adult male. Posterior face.

VI. The same from the median face. V'll. The same from the anterior face. VIII. The same from tiie externa! face.



Department of Philosophy, University of Maiiie

The influence of the male upon the time of oviposition is a matter in regard to which pigeons differ from some other birds, notably the domestic fowl. With regard to the fowl I have consulted a number of poultry keepers and experts, chiefly Dr. Raymond Pearl and Dr. Frank M. Surface, of the Maine Agricultural Experiment Station, where the most extensive studies of the egglaying of fowls have been, and are being carried on. Dr. Pearl and Dr. Surface tell me that the domestic hen, and also the hen of the wild Gallus bankiva so far as can be ascertained, commence their spring laying at an approximately fixed date which can neither be deferred by withholding the cock nor advanced by giving the cock before the usual time.

Pigeons differ widely from poultry in this respect. If, from the winter season onward, an old female piegon be kept unmated and isolated, she refrains from egg-laying, in evident distress for want of a mate, until the breeding season is far advanced; at length she does begin to lay, but her laying without a mate manifestly partakes of the abnormal. And a virgin pigeon, if kept isolated from other pigeons, may postpone her laying for a still longer period. On the other hand, a female pigeon, young or old, will lay very early in the season if she be early mated. Moreover, there is a pretty definite interval between the first copulation and the laying of the first egg, namely six or seven days; if the egg be delayed much beyond this time, the fact indicates some indisposition on the part of the female. And as the pair rear brood after brood throughout the season, this time-relation between copulation and egg-laying is regularly repeated.



The utility of this time adjustment in pigeons seems obvious. The male pigeon takes his turn daily in the duty of incubation: hence the female must not lay the eggs before he is ready to sit. This aspect of the matter, which has to do with pigeon sociology, has already been treated elsewhere (Craig '08) and will be discussed more fully in a book dealing with pigeon behavior. The present paper is to show, not why the male should determine the time of oviposition, but how he does determine it.

The thesis of the present paper is, that the influence of the male in inducing oviposition is a psychological influence; that the stimulus to oviposition is not the introduction of sperm, for the male can cause the female to laj^ even though he does not copulate with her. This is easily proven by an experiment, which requires only pigeons, patience, and time, and I shall now recount seven repetitions of such experiment, the first two being accidental cases, the other five being trials designed and carried out on purpose to test the thesis.

Case 1 (1903). In the spring of 1903 I brought together a virgin female dove (individual female no. 7, the species in all these trials being the blonde ring-dove, Turtur risorius) and a young inexperienced male, intending simply that they should mate in the normal manner. The young male played up to the female, but due to his inexperience and to other causes which need not be discussed here, his mating behavior was imperfect and he did not copulate with her. Nevertheless, in due time (six days) she laid an egg, and a second egg, as usual, forty hours later. This was the first intimation to me that a male bird can stimulate the female to lay, without copulating with her. Such an explanation seemed so absurd at that time that I dismissed it with the assumption that the birds must have copulated unobserved, and I did not even test the eggs to see if they were fertile. Looking back on that case now, however, and considering the observed behavior of that male, I feel reasonably certain that he did not fertilize the eggs but simply stimulated oviposition through the psychic (neural) channels.

Case 2 (1904). A female dove (no. 5) had been kept alone ever since her mate had died in November, 1903, and as time wore on


she showed mtense anxiety to mate. She being a very tame bird, I had often caught and held her gently, but she did not like to be held, so one day in early March I tried tickling her head and pulling the feathers about her neck somewhat as a courting male would do it, and, finding that the poor lonely bird received these attentions with intense pleasure and became still more tame, I continued to preen her neck daily. She now acted toward the hand as if it were a mate, went through a nesting performance in her seed dish, there being no nest in her cage, and to my astonishment laid her eggs in due season. The first egg was laid March 11 and the second March 13. There is no doubt in my mind that the caressing of this bird's head and neck brought on oviposition. I once tried to repeat the experiment with another female dove, but she would not accept the touch of the hand as the former dove had done. Yet there is other evidence indicating that, with a specially tamed bird, this experiment, inducing oviposition by the hand, could be successfully repeated.

This case called to mind that of 1903, and suggested an experiment to determine definitely whether the male dove can stimulate the female to lay, without actual copulation. Opportunity to try this experiment was not found till 1907 and following years, when it was planned as follows.

Method of the regular trials

The experiment requires an unmated female dove that is not laying eggs, preferably a young dove that has never laid. It is best tried early in the season (e.g., in February), especially if an old dove be used, for, as said above, if the female is kept too long without a mate she may lay without one. Side by side with this female, in a separate cage, is placed an unmated male, and the two are given several days to become acquainted. When they act toward one another like mate and mate, the doors separating them are opened and they are allowed to come together for a time, under constant supervision. When they attempt to copulate, a slender rod which can be thrust between the bars of the cage is used to keep them apart. Such attempts are made many times in a day,


mostly in the afternoon, and are continued for several days in succession; hence it is best that the experimenter should be able to devote some hours a day for several days in succession to a single pair or at most two pairs of birds. Whenever the birds are not under surveillance they are shut apart, each in his or her own cage. But they should be allowed to come together daily until the egg is laid.

A factor which caused difficulty in one of my trials was the nest. In cases 1, 2, 3 and 6, the bird laid without any nest at all (except that in case 6 a nest was given just a few hours before the egg was deposited). But in case 4 {q.v.) the female refused to lay without a nest : it was then necessary to remove the male and make the trial again, first giving the female a nest, and waiting long enough to prove that the nest alone would not cause her to lay.

Results of the regular trials

Case 3 (1907). Female dove, no. 20. This bird had been bought recently from a dealer, and it was not known whether she had laid earlier in the season. But she was kept isolated for some time, during which she showed no inclination to lay. She was then given a male in the manner indicated. No nest given.

June 9. Male allowed in cage of female, and plays up to her.

June 15. First egg.

June 17. Second egg. (The second egg was of no special interest. After the first egg was laid, I generally left the doors open, allowing the pair to come together without surveillance.)

Case 4- (1908). Female, the same. She had not laid since the close of last season. No nest given.

February 4. Male allowed to enter.

The female was unresponsive and showed by her behavior that this time she was holding back for want of a nest. This deficiency was supplied in the following manner (vide ut supra.)

February 8. Male taken away to another building.

March 10. Nest put in cage. Female paid practically no attention to it. Many days were allowed to pass, in order to make sure that the nest alone would not stimulate the female to lay.


March 21. Male (after short period in sight of female, that they might become re-acquainted) allowed to enter.

March 27. Egg laid.

Case 5 (1910). Female, the same as in cases 3 and 4. She has laid no eggs since last season (1909.)

January 20. I begin to allow male in cage, at same time putting nest in.

January 29. Egg laid.

Case 6 (1908). Female, no. 19. Virgin, has never laid. No nest given. In this case, the date on which the female was first given the requisite stimulus cannot be stated so definitely as in the other cases.

July 12. Male,- in his cage, placed close to cage of female. Cooing commences. Female so excited that she several times assumes, and maintains in extreme degree, the copulation posture.

July 14. Male allowed into cage of female, but he fights her, so that it is necessary to remove him (otherwise the female might be painfully injured), and to allow the pair a few days more of preliminary acquaintanceship.

July 18. Male allowed to begin his series of daily visits.

July 22. Egg laid.

Case 7 (1910). Female no. 19, the same as in case 6. She has laid no eggs since last summer (1909.)

For several days before contact with the male, a nest was kept in her cage; but she paid no attention to it, showing that the nest alone would not stimulate her to lay.

January 20. Male allowed to enter.

January 26. Egg laid.


1. In six cases, stimulation of a female dove by a male, without copulation, was followed by oviposition; and in one other instance (case 2), stimulation by the hand of man in imitation of a male dove was followed by oviposition

2. In six of the seven cases (being all except case 3, in which the previous history was unknown), it was known that the female



had laid no eggs previously during the current year. In two of these six cases the dove was a virgin aad had never laid.

3. It is true that the female may, if left without a mate, begin to lay late in the season. Hence it might be suspected that the sequence of stimulation and egg-laying in the seven cases was mere coincidence. But this is precluded, first of course by the fact that coincidences are not known to happen seven times in succession, and further by the following considerations.

4. In some of the trials it was proven that the female when stimulated by the male laid much earlier in the season than she did when not so stimulated. This is shown in the following table.

Female, no. 20.

1908. (Case 4), stimulated by male, laid March 27.

1909. (Control), without male, began to lay May 13.

1910. (Case 5), stimulated by male, laid January 29. Female, no. 19.

1909. (Control), without male, began to lay April 26.

1910. (Case 7), stimulated by male, laid January 26.

5. The interval between the first stimulation by the male, and the laying of the first egg, was as follows:

Case 1. 6 days.

Case 2. (Male not used.)

Case 3. 7 days.

Case 4. 6 days.

Case 5. 9 days.

Case 6. 4 to 10 days, depending on what is regarded as the first stimulation in this case.

Case 7. 6 days.

The average and the variation of these intervals tally closely with the average and the variation of the interval in normal breeding, between the first copulation and the laying of the first egg.

6. There were no exceptions. Ovoposition never failed to follow within nine days after the first contact with the male. (The onlj^ partial failure was that of the first trial in case 4, which was due to faulty experimental conditions.)



These facts make it certain that the male dove can stimulate the female to lay, without copulating with her.

Harper ('04) mentioned the fact that ovulation in the pigeon does not take place until after the bird is mated, but he was in doubt as to how far the influence of mating was a 'mental' one and how far it was a matter of the introduction of sperm. The present paper goes to show that the stimulus to the whole process of egg development and laying is a psychic (neural) stimulus, not dependent upon the introduction of sperm.


Craig, Wallace 1908 The voices of pigeons regarded as a means of social control. Am. Jour. Sociol., vol. 14, pp. 86-100.

Harper, Eugene Howard 1904 The fertilization and early development of the pigeon's egg. Am. Jour. Anat., vol. 3, pp. 349-386.



As a zoologist, reared among what are now rapidly coming to be regarded as antiquated ideals, I confess to a feeling of great diffidence in addressing an audience so thoroughly versed in the very latest as well as the very oldest biological facts, methods and hypotheses. I feel, indeed, like some village potter who is bringing to the market of the metropolis a pitiable sample of his craft, a pot of some old-fashioned design, possibly with a concealed crack which may prevent it from ringing true. Although in what I have to say, I shall strenuously endeavor to be modern, I can only beg you, if I fail to come within hailing distance of the advance guard of present day zoologists, to remember that the range of adaptability in all organisms, even in zoologists, is very limited.

Under the circumstances, my only hope lies in appealing to our permanent common biological interests and these, I take it, must always center in the organism. But the point of view from which we study this most extraordinary of nature's manifestations, is continually shifting. Twenty years ago we were captivated by the morphology of the organism, now its behavior occupies the foreground of our attention. Once we thought we were seriously studying biology when we were scrutinizing paraffine sections of animals and plants or dried specimens mounted on pins or pressed between layers of blotting paper; now we are sure that we were studying merely the exuviae of organisms, the effete residua of the life-process. If the neovitalistic school has done nothing else, it has jolted us out of this delusion which was gradually taking possession of our faculties. It is certain that whatever changes may overtake biology in the future, we must henceforth grapple

1 A lecture prepared for delivery at the Marine Biological Laboratory, Woods Hole, Mass., August 2, 1910.



with the organism as a dynamic agencj' acting in a very complex and unstable environment. In using the term organism, therefore, I shall drop the adjective ' living,' since I do not regard pickled animals or dried plants as organisms.

As I wish to describe a peculiar type of organism, I may be asked, before proceeding, to state more concisely what I mean by an organism. It is obvious that no adequate definition can be given, because the organism is neither a thing nor a concept, but a continual flux or process, and hence forever changing and never completed. As good a formal definition as I can frame is the following: An organism is a complex, definitely coordinated and therefore individualized system of activities, which are primarily directed to obtaining and assimilating substances from an environment, to producing other similar systems, known as offspring, and to protecting the system itself and usually also its offspring from disturbances emanating from the environment. The three fundamental activities enumerated in this definition, namely nutrition, reproduction and protection seem to have their inception in what we know, from exclusively subjective experience, as feelings of hunger, affection and fear respectively.

Biologists long ago constructed an elaborate hierarchy of organisms. Those of a speculative turn of mind, like Spencer and Weismann, postulated the existence of very simple organisms, the physiological units, or biophores, which, though invisible, were nevertheless conceived as combining the fundamental activites above enumerated. These biophores were supposed to form by aggregation the cells, which may exist as independent organisms in the Protozoa and Protophyta or unite with other cells to form more complex aggregates, for which Haeckel's term 'persons' may be adopted. The person may be merely a cell-aggregate or consist of complexes of such aggregates as the metameres of the higher animals, for the separate metameres, according to a very generally accepted theory, are supposed to be more or less modified or highly specialized persons. Somewhat similar conditions are supposed to obtain in the composition of the vascular plants. The integration both of the metameric and non-metameric Metazoa may proceed still further, the simple persons combining to


form colonies in which the persons are primarily nutritive and acquire fixed and definite spatial relations to one another, whereas the more specialized animals, like the social insects, may constitute families of mobile persons with reproduction as the 'Leitmotiv' of their consociation. In man we have families associating to form still more complex aggregates, the true societies. Other comprehensive organisms are the coenobioses, or more or less definite consociations of animals and plants of different species, which the ecologists are endeavoring to analyze. Finally we have philosophers, like Fechner, stepping in with the assertion, that the earth as a whole is merely a great organism, that the planetary systems in turn are colonies of earths and suns and that the universe itself is to be regarded as one stupendous organism. Thus starting with the biophore as the smallest and ending with the universe as the most comprehensive we have a sufficiently magnificent hierarchy of organisms to satisfy even the most zealous panpsychist. As biologists we may, for present purposes, lop off and discard the ends of this series of organisms, the biophores as being purely hypothetical and the cosmos as involving too many ultrabiological assumptions. We then have left the following series: first, the Protozoon or Protophyte, second the simple or non-metameric person, third the metameric person, fourth the colony of the nutritive type, fifth the family, or colony of the reproductive type, sixth the coenobiose, and seventh the true, or human societ}'. Closer inspection shows that these are sufficiently heterogeneous when compared with one another and with the personal organism, which is the prototype of the series, but I believe, nevertheless that all of them are real organisms and not merely conceptual constructions or analogies. One of them, the insect colony, has interested me exceedingly, and as I have repeatedly found its treatment as an organism to yield fruitful results in my studies, I have acquired the conviction that our biological theories must remain inadequate so long as we confine ourselves to the study of the cells and persons and leave the psychologists, sociologists and metaphysicians to deal with the more complex organisms. Indeed our failure to cooperate with these investigators in the study of animal and plant societies has blinded us to many aspects of the cellular and personal activities with which we are constantly dealing. This failure, moreover, is largely responsible for our fear of the psychological and the metaphysical, a fear which becomes the more ludicrous from the fact that even our so-called 'exact' sciences smell to heaven with the rankest kind of materialistic metaphysics.

Leaving these generalities for the present, permit me to present the evidence for the contention that the animal colony is a true -organism and not merely the analogue of the person. To make this evidence as concrete as possible I shall take the ant-colonj^ as a paradigm and ask you to accept my statement that the colonies of the termites, social bees and wasps, which the limited time at my disposal does not permit to consider, will be found to offer the same and in some cases even more satisfactory data. I select the ant-colony not only because I am more familiar with its activities, but because it is much more interesting than that of the polyps, more typical and less specialized than that of the honey bee, less generalized than that of the wasps and bumble-bees, and has been much more thoroughly investigated than the colonies of the stingless bees and the termites.

The most general organismal character of the ant-colony is its individuality. Like the cell or the person, it behaves as a unitary whole, maintaining its identity in space, resisting dissolution and, as a general rule, any fusion with other colonies of the same or alien species. This resistance is very strongly manifested in the fierce defensive and offensive cooperation of the colonial personnel. Moreover, every ant-colony has its own peculiar idiosyncrasies of composition and behavior. This is most clearly seen in the character of the nest, which bears about the same relation to the colony that the shell bears to the individual Foraminifer or mollusc. The nest is a unitary structure, built on a definite but plastic design and through the cooperation of a number of persons. It not only reflects the idiosyncrasies of these persons individually and as a whole, but it often has a most interesting adaptive growth and orientation which may be regarded as a kind of tropism. In many species the nest mounds, which are used as incubators of the brood and as sun-parlors for the adult ants, are constructed in


such a manner as to utilize the solar radiation to the utmost. In the Alps and Rocky Mountains we find the nests oriented in such a manner that the portions in which the brood is reared face south or east, and as time goes on the nests often grow slowly in these directions, like plants turning to the light, so that they become greatly elongated. This orientation is, in fact, so constant in some species that the Swiss mountaineers, when lost in a fog, can use it as a compass.

Every complete ant-colony, moreover, has a definite stature which depends, of course, on the number of its component persons. And this stature, like that of personal organisms, varies greatlj^ with the species and is not determined exclusively by the amount of food, but also by the queen mother's fertility, which is constitutional. Certain ants live in affluence but are nevertheless unable to form colonies of more than fifty or a hundred individuals, while others, under the same conditions, have a personnel of thousands or tens of thousands.

One of the most general structural pecuharities of the person is the duality of its composition as expressed in the germ-plasm on the one hand and the soma on the other, and the same is true of the ant-colony, in which the mother queen and the virgin males and females represent the germ-plasm, or, more accurately speaking, the ' Keimbahn,' while the normally sterile females, or workers and soldiers, in all their developmental stages, represent the soma. In discussing the question of the inheritance or non-inheritance of acquired characters the Neodarwinians trace all the congenital modifications of the worker and soldier phases to the queen, just as in the personal organism all the congenital somatic characters are traced to the germ-plasm of the egg. Since the homologue of the reproductive organ of the ant-colony consists of the virgin males and females, and since the males mature earlier than the females, the colony may be regarded as a protandric hermaphrodite. Some colonies, however — and this is probably characteristic of certain species — produce only males or females and are therefore in a sense gonochoristic, or dioecious. And this protandric hermaphroditism and gonochorism, like the corresponding conditions in persons, may be interpreted as a device for, or, at any rate, as an aid, in insuring cross-fertilization. The fecundated queen of the ant-colony represents the first link in the 'Keimbahn' and therefore corresponds to the fertilized egg of the personal organism. She produces both the worker personnel and the virgin males and females, just as the fertilized egg produces both the soma and the germ-cells. The colonial soma, moreover, may be differentiated as the result of a physiological division of labor into two distinct castes, comprising the workers in which the nutritive and nidificational activities predominate, and the soldiers, which are primarily protective. Here, too, the resemblance to the differentiation of the personal soma into entodermal and ectodermal tissues can hardly be overlooked.

The structure of the ant-colony thus appears to be very simple as compared with that of its component persons. The question naturally arises as to the particular type of unicellular or personal organism which it most resembles. Undoubtedly, if we could see it acting in its entirety, the ant-colony would resemble a gigantic foraminiferous Rhizopod, in which the nest would represent the shell, the queen the nucleus, the mass of ants the Plasmodium and the files of workers, which are continually going in and out of the nest, the pseudopodia.

The ant-colony, of course, like the person, has both an ontogenetic and a phylogenetic development; the former open to observation, the latter inferred from the ontogeny, a comparison of the various species of ants with one another and with allied Hymenopterous insects, and from the paleontological record. The fecundated queen, as I have stated, represents the fertilized egg which produces the colonial organism, but she is a winged and possibly conscious egg, capable not only of actively disseminating the species, like the minute eggs of many marine animals, but of selecting the site for the future colony. After finding this site she discards her wings and henceforth becomes sedentary like the wingless workers which she will produce. The whole colony rests satisfied with the nesting site selected by its queen if the environmental conditions remain relatively constant. If these become unfavorable, however, the colony will move as a whole to a new site. In most species such movements are rather limited, but the nomadic driver and legionary ants are almost continually moving from place to place and must cover a considerable territory during the year. After the queen has selected the nesting site, she immures herself in some earthen or vegetable cavity, laj^s a number of eggs, supplying them with yolk derived by metabolism from her fat-body and now useless wing-muscles, and feeds the hatching larvae on her salivary secretion, which, though highly nutritious, is, nevertheless, very limited in quantitj^ so that the offspring when mature are dwarfed and very few in number. They are in fact, workers of the smallest and feeblest caste; but they set to work enlarging the nest, break through the soil or plant tissues, construct an entrance on the surface and seek food for themselves and their famished mother. This food enables her to replenish her fat-body and to produce more eggs. Her expansive instincts and activities now contract, so to speak, and becoine reduced henceforth to a perpetual routine of assimilation, metabolism and oviposition. She produces brood after brood during her long life which may extend over a period of ten to thirteen years. Her workers assume the duties of foraging, of feeding the larvae and one another, and of completing the nest. Their size and polymorphism increase with successive broods, till the soldier forms, if these are characteristic of the species, make their appearance. Then the individuals which correspond to the reproductive cells of the personal organism, namely, the virgin males and females develop, and the colonial organism may be said to have reached maturity. Like the personal organism, it may persist for thirty or forty years or,, perhaps, even longer without much growth of its soma, since the workers and soldiers of which this consists are exposed to many vicissitudes and live only from three to four years and probably, as a rule, for a much shorter period. If the queen grow too old or die the colonj^, as a rule, dwindles and eventually perishes unless her place is taken by one or more of her fertile daughters.

This is the ontogenetic history of most ant-colonies. It is so similar to the phylogenetic history derived from the sources mentioned above that we have no hesitation in affirming that it conforms in the most striking manner to the biogenetic law. The very ancient behavior of the soUtary female Hymenopteron is still reproduced during the incipient stage of colony formation, just as the unicellular phase of the Metazoon is represented bj^ the egg. A further correspondence of the ontogeny and phylogeny is indicated by the fact that the most archaic and primitive of living ants form small colonies of monomorphic workers closely resembling the queen, whereas the more recent and most highly specialized ants produce large colonies of workers not only verj^ unlike the queen but unhke one another.

In order to complete the foregoing account it will be necessary to consider some interesting modifications of the usual method of colony formation and growth, especially as these modifications furnish additional and striking evidence in favor of the contention that the ant-colony is a true organism. In many species, after the colony has reached maturity and especially if the food-supply continue to be abundant, several of the virgin females may be fecundated in the nest, lose their wings and remain as members of the colony. This may, indeed, contain half a dozen and in extreme cases as many as forty or fifty or even more fertile queens. But often the growth of the colonial organism becomes excessive through an increase in the worker personnel and passes over into a form of colonial reproduction, when the young fertilized queens, each accompanied by a band of workers, start new nests in the vicinity of the parental formicary. In this manner a very large and complex colony may arise and extend over many adjacent nests. For some time the new settlements may remain in communication with the home-nest through files of workers, but eventually the daughter settlements may become detached and form independent colonies. The resemblance of this method of reproduction, which is essentially the same as the^swarming in the honey-bee, to the asexual reproduction of many unicellular and multicellular organisms by a process of budding, is too obvious to need further comment.

The important role of nutrition in the development of the colony will be clear from the foregoing remarks. It becomes even more striking in the methods adopted by the queens of certain parasitic species in starting their colonies. Some European


observers and myself have found a number of queen-ants that are unable to found colonies without the aid of workers of allied species. These queens may be separated into four groups, as follows:

1. The queen which enters a colony of an alien species and decapitates its queen or is the occasion of her being killed off by her own workers. The intrusive queen is then adopted by the workers and a compound colonial organism arises, consisting of the germ-plasm of one species and the soma of another. The queen proceeds to lay eggs, which are reared by the alien workers, thus relieving her of all the labor and exhaustion endured by the independent typical ant-queen during the early stages of colony formation. Pari passu with the development of the worker offspring of the intrusive queen, the worker nurses grow old and die, so that the colony eventually comes to consist of only one species, the soma of the host being replaced bj^ that of the parasite. This method of colony formation, first observed among our American ants and later among certain European and North African species, I have called temporary social parasitism. Now many of the species, which behave in this manner, have extremely small queens, or queens provided with a peculiar pilosity or sculpture that tend to endear them to the workers of the alien colonies which they invade. If we regard the large fertilized queens of ordinary ants, which are supplied with a voluminous fat-body and wing-musculature, as representing eggs provided with a great amount of yolk, and the diminutive queens of the temporary social parasites as the equivalents of alecithal eggs, we have another striking resemblance between the personal and colonial organisms, for the large queens, like the yolk-laden eggs of many vertebrates, are produced in small numbers but are able to generate the colonial soma independently, whereas the small queens, which are produced in great numbers, in order that some of them may survive the vicissitudes of a parasitic life, correspond to the small yolk-less eggs of many parasites, which have to be deposited in plant or animal tissues in order that the imperfect young on hatching may be surrounded by an abundance of food.

2. The queen of the blood-red slave-maker (Formica sanguinea) adopts a different method. She enters the colony of an


allied species, snatches up the worker brood and kills any of the workers or queens that endeavor to dispute her possessions. The ants hatch with a sense of affiliation with their foster mother and proceed to rear her eggs and larvae as soon as they appear. Here, too, the colony is formed by a mixture of two species, but the workers produced by the intrusive queen inherit her predatory instincts and therefore become slave-makers. They keep on kidnapping worker larvae and pupse from the nests of the alien species, carry them home, and eat some of them but permit many to mature, so that the mixed character of the colony is maintained. This, however, is not invariably the case, for old and vigorous sanguinea colonies may cease to make slave-raids and the slaves may die off and leave a pure colony of the predatory species. The advantages of this method of colony formation arfe obvious, for the colonial soma, being composed of two species, grows more rapidly and is much more efficient as a nutritive and protective support to the colonial germ-plasm, which is restricted to the predatory species.

3. The colony-founding queen of the amazon ants of the genus Polyergus resorts to a modification of the method adopted by sanguinea, as has been shown by Emery's recent observations. She enters the colony of an alien species, perforates its queen's head with her sickle-shaped mandibles and permits herself to be adopted by the workers. She pays no attention to the brood but begins to lay eggs, the larvae from which are carefully reared by the workers. The Polyergus offspring inherit the pugnacity of their mother, but, like the sanguinea workers, have the ability to kidnap the brood of other ants. They are, in fact, slave-makers of a very deft and ferocious type. Like their mother, however, the}' are unable to excavate the nest, to care for their own young or to take food except from the mouths of the workers that hatch from the kidnapped larvae and pupae. The mixture of the two species is therefore obligatory, and the slave personnel, which represents the nutritive and nest-building portions of the colonial soma, has to be maintained throughout the life of the colony.

4. Certain feeble queen ants belonging to a few aberrant genera (Anergates, Wheeleriella) invade populous nests of an alien species and are adopted in the place of their queens, which are


destroyed by their own workers. The parasites then proceed to lay eggs but these give rise only to males and females as the worker caste is entirely suppressed. The colony retains a mixed character, the parasitic species usurping the functions of the germplasm, while the host is purely somatic. As there are no means of prolonging the lives of the host-workers and as they do not reproduce, the whole colony is short-lived and the maturation of the parasitic sexual individuals has to be accelerated so that it will fall within the brief life-time of the worker hosts. This condition I have called permanent social parasitism.

These four peculiar types of colony-formation all lead to the formation of compoand colonial organisms, comparable to certain compound personal organisms which, with few exceptions, can be produced only by artificial means. In temporary social parasitism the colonial egg can develop its soma only when grafted on to the soma of another species. This soma eventually perishes and the colony then assumes a normal complexion. This condition reminds us of certain tropical plants, like the species of Clusia and Ficus, which develop as epiphytes on other trees but after killing their hosts take root in the soil and thenceforth grow as independent organisms. The slave-makers of the sanguinea or facultative type are also unable to develop the soma except when grafted on to the soma of another species, but in this case the cooperation of both somas in nourishing and protecting the germplasm is maintained for a much longer period. This kind of colony may be compared with a graft made by uniting the longitudinal half of one plant with that of another so that both take nourishment through their roots. To make the resemblance more complete one of the grafted halves would have to be pruned in such a manner as to prevent flowering. In the amazons or obligatory slave-makers and the permanent social parasites the alien soma alone has a nutritive function, so that the conditions are like those in ordinary vegetable grafts, in which the stock retains the roots and the scion produces the flowers and fruit.

I have dwelt on the various methods of colony formation not only because they give us an insight into colonial reproduction, but because they throw light on the colonial organism from the


standpoint of parasitology. That the four types of queens and their offspring are directly comparable with entoparasitic persons is not so remarkable as the fact that in ants the host and parasite form a mixed organism which could only be obtained with persons by jumbling together the component cells of host and parasite like two kinds of peas shaken in a bottle. Notwithstanding this mixture the parasitic colony not only retains its identity and the anticipatory character of its behavior but castrates the host colony and constrains its soma either to cooperate in many of its activities or to specialize as a purely nutritive or nest-building auxiliary. The host is thus reduced to the status of a nourishing or protective organ of the parasite. This behavior has many striking analogies among persons. Giard long ago called attention to the fact that when the cirriped Sacculina settles under the abdomen of a male crab and sends its rootlike haustoria into the tissues of its host, the latter undergoes castration, and its narrow abdomen expands to form a protection for the soft-bodied parasite. In other words, the parasite acts as if it were a mass of crabs' eggs and the male crab behaves as if it had changed its sex and develops an abdomen of the female type.

Not only are there ants, like those already considered, that may be regarded as colonial entoparasites, but there are also a number of species that may be called colonial ectoparasites. These form the so-called 'compound nests,' in which two or more species live amicably side by side, or may even mingle freely with one another, but rear their broods in separate nests, thus indicating in the clearest manner the integrity of the colonial organism. This is also shown by the vast number of myrmecophilous insects, which are, of course, ento- or ectoparasitic persons, and behave towards the ant colony as if it were a rather incoherent and therefore more vulnerable, or exploitable personal organism.

Finally we come to what the neovitalists regard as the most striking autonomic manifestations of the organism, namely the regulations and restitutions, and face the question as to whether these, too, have their counterpart in the colonial organism. I believe that the following facts compel us to answer this ques


tion in the affinnative. If the worker personnel be removed from a young ant-colony, leaving only the fertile queen, we find that this insect, if provided with a sufficiently voluminous fat-body, will set to work and rear another brood, or, in other words, regenerate the missing soma. And, of course, any portion of the worker or sexual personnel, that is removed from a vigorous colony will be readily replaced by development of a corresponding portion of the brood. On the other hand, if the queen alone be removed, one of the workers will often develop its ovaries and take on the egg-laying function of the queen. In ants such substitution queens, or gynaecoid workers are not fertilized and are therefore unable to assume their mother's worker- and queen-producing functions. The termites, however, show a remarkable provision for restituting both of the fertile parents of the colony from the so-called complemental males and females. In ants we have a production of fertile from normally infertile individuals, but the incompleteness of the result does not disprove the existence of a pronounced restitutional tendency.

Very striking examples of this tendency are exhibited when colonies are injured by parasitic myrmecophiles. I shall consider only the case of the peculiar beetle Lomechusa strumosa, which breeds in colonies of the blood-red slave-maker (Formica sanguinea) . Though the beetle and its larvse are treated with great affection, the latter devour the ant larvse in great numbers, so that little of the brood survives during the esirlj sumAier months when the colony is producing its greatest annual increment to the worker personnel. The ants seem to perceive this defect and endeavor to remedy it by converting all the surviving queen larvse into workers. But as these larvse have passed the stage in their development when such an operation can be successful, the result is the production of a lot of pseudogynes, or abortive creatures structurally intermediate between the workers and queens and therefore useless in either capacity. It is instructive to compare this case with the regeneration of the lens from the iris in the Amphibian eye. In his recent analysis of the stimuli of restitution in personal organisms Driesch reaches the conclusion that "the specificity of what is taken away certainly forms part of the



stimulus we are searching for, and it does so by being communicated in some way by something that has relations to many, if not all, parts of the organism and not only to the neighboring ones." He also says that "each part of the organism assigns its specific share to an unknown something and that this something is altered as soon as a part is removed or absolutely stopped in its functional life, and that the specific alteration of the something is our stimulus of restitutions." These quotations and Driesch's further discussion of the problem are even clearer in their application to the colonial than to the personal organism, for in the former it is much easier to see how each individual insect "can do more than one thing in the service of restitution" than it is to understand how each cell of the person can do more than one thing in restoring a lost organ.

I fear that I may have wearied you with this long attempt to prove that the ant-colony is a true organism, especially as this statement must seem to some of you to be too trite for discussion, but when an author like Driesch writes a large work in two volumes on the "Philosophy of the Organism" and ignores the colonial organisms altogether, an old-fashioned zoologist may perhaps be pardoned for calling attention to a well-founded, though somewhat thread-bare, biological conception.

If it be granted that the ant-colony and those of the other social insects are organisms, we are still confronted with the formidable question as to what regulates the anticipatory cooperation, or synergy of the colonial personnel and determines its unitary and individualized course. The resemblance of the ant- or bee-colony to the human state long ago suggested a naive reply to this question. Aristotle naturally supposed the colonial activities to be directed and regulated by a ^aaCkebs or riyeixuv, because these personages managed affairs in the Greek states. After the sex of the fertile individual had been discovered by Swammerdam, the word 'queen' was naturally substituted for (SaatXevs or 'king/ and as queens in human states do not necessarily govern and are often rather anabolic, sedentary and prolific persons and the objects of much flattering attention, the term is not altogether inapt when applied to the fertile females of insect colonies. It


has been retained although everybody knows that these colonies represent a form of society very different from our own, a kind of communistic anarchy, in which there is neither guide, overseer nor ruler," as Solomon correctly observed. In this respect too, the colony is essentially the same as the personal organism, at least in the opinion of those who do not feel compelled to assume the existence of a 'soul' in the scholastic sense. For it is clear, that to primitive thinkers the soul was supposed to bear the same relation to the person as the iSaatXevs to the insect colony and the king to the human state. This supposition is still held though in a more subtle form, by writers of the present day. Some of these, like Maeterlinck, clothe the postulated controlling agency in a mystical or poetic garb and call it the 'spirit of the hive.' The following passage from the Belgian poet's charming account of the honey-bee will serve to illustrate this method of meeting the problem:

What is this 'spirit of the hive' — where does it reside? It is not like the special instinct that teaches the bird to construct its well planned nest, and then seek other skies when the day for migration returns. Nor is it a kind of mechanical habit of the race, or blind craving for life, that will fling the bees upon any wild hazard the moment an unforeseen event shall derange the accustomed order of phenomena. On the contrary, be the event never so masterful, the 'spirit of the hive' still will follow it, step by step, like an alert and quickwitted slave, who is able to derive advantage even from his master's most dangerous orders.

It disposes pitilessly of the wealth and the happiness, the Hberty and life, of all this winged people; and yet with discretion, as though governed itself by some great duty. It regulates day by day the number of births, and contrives that these shall strictly accord with the number of flowers that brighten the country-side. It decrees the queen's deposition or warns her that she must depart; it compels her to bring her own rivals into the world, and rears them royally, protecting them from their mother's political hatred. So, too, in accordance with the generosity of the flowers, the age of the spring, and the probable dangers of the nuptial flight will it permit or forbid the first-born of the virgin princesses to slay in their cradles her younger sisters, who are singing the song of the queens. At other times, when the season wanes, and flowery hours grow shorter, it will command the workers themselves to slaughter the whole imperial


brood, that the era of revolutions may close, and work become the sole object of all. The 'spirit of the hive' is prudent and thrifty, but by no means parsimonious. And thus, aware, it would seem, that nature's laws are somewhat wild and extravagant in all that pertains to love, it tolerates, during summer days of abundance, the embarrassing presence in the hive of three or four hundred males, from whose ranks the queen about to be born shall select her lover; three or four hundred foolish, clumsy, useless, noisy creatures, who are pretentious, gluttonous, dirty, coarse, totally and scandalously idle, insatiable, and enormous.

But after the queen's impregnation, when flowers begin to close sooner and open later, the spirit one morning will coldly decree the simultaneous and general massacre of every male. It regulates the worker;^' labours with due regard to their age; it allots their task to the nurses who tend the nymphs and the larvae, the ladies of honour who wait on the queen and never allow her out of their sight ; the house-bees who air, refresh, or heat the hive by fanning their wings, and hasten the evaporation of the honey that may be too highly charged with water ; the architects, masons, wax- workers, and sculptors who form the chain and construct the combs ; the foragers who sally forth to the flowers in search of the nectar that turns into honey, of the pollen that feeds the nymphs and the larvae, the propohs that welds and strengthens the buildings of the city, or the water and salt required by the youth of the nation. Its orders have gone to the chemists who ensure the preservation of the honey by letting a drop of formic acid fall in from the end of their sting; to the capsule makers who seal down the cells when the treasure is ripe, to the sweepers who maintain public places and streets most irreproachably clean, to the bearers whose duty it is to remove the corpses ; and to the amazons of the guard who keep watch on the threshold by night and by day, question comers and goers, recognize the novices who return from their very first flight, scare away vagabonds, marauders and loiterers, expel all intruders, attack redoubtable foes in a body, and, if need be, barricade the entrance.

Finally, it is the spirit of the hive that fixes the hour of the great annual sacrifice to the genius of the race: the hour, that is, of the swarm; when we find a whole people, who have attained the topmost pinnacle of prosperity and power, suddenly abandoning to the generation to come their wealth and their palaces, their homes and the fruits of their labour; themselves content to encounter the hardships and perils of a new and distant country. This act, be it conscious or not, undoubtedly passes the fimits of human morahty. Its result will sometimes be ruin, but poverty


always; and the thrice-happy city is s'cattered abroad in obedience to a law superior to its own happiness. Where has this law been decreed which, as we soon shall find, is by no means as blind and inevitable as one might believe? Where, in what assembly, what council, what intellectual amd moral sphere, does this spirit reside to whom all must submit, itself being vassal to an heroic duty, to an intelligence whose eyes are persistently fixed on the future?

It comes to pass with the bees as with most of the things in this world; we remark some few of their habits; we say they do this, they work in such and such fashion, their queens are born thus, their workers are virgin, the}^ swarm at a certain time. And then we imagine we know them, and ask nothing more. We watch them hasten from flower to flower, we see the constant agitation within the hive; their life seems very simple to us, and bounded, like every life, by the instinctive cares of reproduction and nourishment. But let the eye draw near, and endeavour to see ; and at once the least phenomenon of all becomes overpoweringly complex; we are confronted by the enigma of intellect, of destiny, will, aim, means, causes; the incomprehensible organization of the most insignificant act of life.

Other authors like Driesch, give the postulated controlling agency the sharper outlines of a would-be scientific but in reality metaphysical entity and call it the 'entelechy.' It is true that the entelechy is deduced by Driesch from the autonomic peculiarities of the personal organism, but as the colony has all the essential attributes of the organism, he would undoubtedly assign it an entelechy, which according to the definition would have to be nonspacial, but working into space, nonspsychic, but conceivable only after analogy with the psychic, and non-energetic, but nevertheless capable of determining the specificity of the colonial activities through releasing and distributing energy.

I confess that I find the entelechy quite as useless an aid in unravelling the complex activities of the ant-colony as others have found it in analyzing the personal organism. This angel-child, entelechy, comes, to be sure, of most distinguished antecedents, having been mothered by the Platonic idea, fathered by the Kantian Ding-an-sich, suckled at the breast of the scholastic forma substantialis and christened, from a strong family likeness, after old Aristotle's darhng evTeXexeta, but nevertheless, I believe that


we ought not to let it play about in our laboratories, not because it would occupy any space or interfere with our apparatus, but because it might distract us from the serious work in hand. I am quite willing to see it spanked and sent back to the metaphysical house-hold.

But, speaking seriously, it seems to me that if the organism be inexplicable on purely biological grounds, we should do better to resort to psychological agencies like consciousness and the will. These have at least the value which attaches to the most immediate experience. And even the subconscious and the superconscious are more serviceable as explanations than such anaemic metaphysical abstractions as the entelechy. Of course, psychic vitalism is one of Driesch's pet aversions and he will have none of it, because he is a solipsist, but the fact that he is compelled to operate with a 'psychoid' and with an entelechy conceivable only jper analogiam with the psychic, shows the inconsistency of his position.

Before we can adopt any ultrabiological agencies, however, except in a tentative and provisional manner, an old and very knotty problem will have to be more thoroughly elucidated. I refer to the problem of the correlation and cooperation of parts. If the cell is a colony of lower physiological units, or biophores, as some cytologists believe, we must face the fact that all organisms are colonical or social and that one of the fundamental tendencies of life is sociogenic. Every organism manifests a strong predelection for seeking out other organisms and either assimilating them or cooperating with them to form a more comprehensive and efficient individual. Whether, with the mechanists, we attribute this tendency to chemotropism or cytotropism, or with the psychic neovitalists, interpret it as conscious and voluntary, we certainly cannot afford to ignore the facts. The study of the ontogeny of the person, i.e., the person in the process of making, in the hands of recent experimentalists, has thrown a flood of light on the peculiarities of organization, but the animal and plant colony are in certain respects more accessible to observation and experiment, because the component individuals bear such loose spacial relations to one another. Then too, the much simpler and more primi


tive organismal type of the colony, as compared with that of the person, should enable us to follow the process of consociation and the resulting physiological division of labor more successfully. In the problem, as thus conceived, we must include, not only the true colony and society, and the innumerable cases of symbiosis, parasitism and coenobiosis, but also the consociation and mutual modification of hereditary tendencies in parthenogenetic and biparental plants and animals, since in all of these phenomena our attention is arrested not so much by the struggle for existence, which used to be painted in such lurid colors, as by the ability of the organism to temporize and compromise with other organisms, to inhibit certain activities of the aequipotential unit in the interests of the unit itself and of other organisms ; in a word, to secure survival through a kind of egoistic altruism. ^

2 Since this paragraph was written I have found that several recent authors have given more explicit expression to a very similar conception to the role of cooperation and struggle in the development of organisms. Especially worthy of mention in this connection are Kammerer (Allgemeine Symbiose und Kampf urns Dasein als gleichberechtigte Triebkrafte der Evolution. Arch. f. Rass. u. Ges.-Biol.6, 1909, pp. 585-608), Schiefferdecker (Symbiose. Sitzb. niederrhein. Ges. f. Natur. u. Heilk. zu Bonn, 13, Juni, 1904, 11 pp.), Bolsche (Daseinskampf und gegeuseitige Hilfe in der Entwicklung. Kosmos, 6, '1909); and Kropotkin (Mutual aid, a factor of evolution, London, 1902).




From the University of Maine, Orono, Maine


This account, which deals with some of the sexual activities of the squid, is based upon observation made on specimens kept in glass sided aquaria at the Marine Biological Laboratory, Woods Hole, Mass. Specimens caught in the fish traps of the immediate vicinity may, by careful handling, be kept in aquaria in fairly good condition for a number of days. Such specimens occasionally copulate and eggs are sometimes laid.

There are two methods of copulation. By one method the spermatophores ejaculate their contents so the sperm reservoirs thrown from them are attached in a special depression on the inner side of the outer buccal membrane opposite the junction of the two ventral arms (figs. 8 and 10). They then slowly emit sperm, which are carried to and stored in, a special sperm receptacle that opens near this depression and is imbedded in the tissue of the outer buccal membrane (figs. 10 and 11). In this receptacle the sperm are mixed with a secretion and are not active. How long the sperm may be retained in the receptacle is not known, but there is some reason to think that they may be retained for at least some weeks. Females with eggs that can be fertilized may be found during the four months, June to late September, that I have worked at Woods Hole. Without exception every adult female that had not spawned had the sperm receptacle filled more or less completely with sperm, although in many cases the



eggs were far from mature. This, together with the dormant condition of the sperm in the receptacle, and the fact that they seem to be poured out only during egg laying, point to a possible long retention. It is certain that the same female may have sperm reservoirs attached near this receptacle a number of times after it has been filled, and it is possible that the same sperm do not continue long in the receptacle. There seems, however, to be no evidence that they are discharged except during the period of egg laying.

The other method of copulation results in fastening the sperm reservoirs of the ejaculated spermatophores near the end of the oviduct (fig. 8, s) usually directly on its walls but sometimes on the mantle, gill or visceral mass. There is no special receptacle for the sperm from these sperm reservoirs. They escape into the water, becoming active as they escape, and pass out with the water through the funnel. The escape of the sperm is rather rapid but there are vast numbers in each reservoir, from which they are constantly poured like smoke from a chimney until the reservoir is empty. It is not known how long it takes to empty a reservoir but by keeping reservoirs from spermatophores that ejaculated in dishes of sea-water, and by examining reservoirs normally attached to the oviducts and buccal membranes of females, it seems probable that the sperm do not all escape for two or more days.

In aquaria I have seen rather more cases of copulation where the spermatophores are inserted into the mantle chamber than where the sperm reservoirs are attached to the buccal membrane. This may be because of the limited quarters in aquaria. In the larger floating tanks, in which specimens are sometimes kept before they are brought into the laboratory, the buccal membrane copulation seems proportionally more common than in aquaria, but even here the mantle chamber copulation seems to be rather more frequent.

The same individuals may copulate several times in the course of a few hours. In general the male is aggressive. The female may attempt to escape or she may be quite passive. Spermatophores seem to be inserted in the mantle chambers of only those


females that are nearly ready to deposit their eggs. In the large number of trials made it was found that the eggs of these individuals were so nearly mature they could be artificially fertilized. Females that are nearly ready to deposit eggs have the nidamental glands considerably swollen and the accessory nidamental glands are highly colored with bright red. Wherever the spermatophores were inserted in the mantle chamber these glands were in this condition.

Before copulation both female and male are usually especially active and may be known as sexually excited animals by their peculiar movements. The female in swimming seems to be nervous or excited. She throws short but rapid puffs of water from the funnel, moves the tail fin very rapidly and, leaving the arms quite limp, spreads them apart and frequently throws them to one side. This gives the arms a jerky or trembling motion not shown in ordinary swimming. Except during the most rapid movements of the female, the male solemnly swims by her side, an inch or two away, but parallel, and with his head in the same direction. He frequently manipulates his arms, spreading them apart, commonly with the two dorsal arms elevated nearly or quite to a perpendicular position, and the third arms spread far to the sides (fig. 3). This position is not infrequently accompanied by localized activity of chromatophores. A spot may appear near the base of each third arm and a smaller spot on each second arm a little further from its base. These spots do not remain continuously while the male is in this attitude but suddenly appear with each increase of activity on the part of either the male or female. Occasionally blushing is quite general over the head and anterior end of the body and sometimes includes the whole body but the bodies of both animals generally remain colorless except for the special spots mentioned on the male. The attitude of the male, with elevated and spread arms, is not continuous but is assumed every few minutes, or in some cases seconds, and the arms may be brought into the usual position of a swimming animal for periods of many minutes.

Males do not all respond equally to the presence of sexually active females. Not uncommonly one male in an aquarium containing


several males will follow the females around by the hour while the other males remain entirely inattentive. Usually when a male begins to show sexual activity he will follow a single female although other females that show similar activities are present in the aquarium. Occasionally he may change to another individual but he nearly always returns after a few minutes to the one to which he has been paying chief attention.

A few males have been observed that were so sexually excited they followed individuals around quite indiscriminately. Under such conditions I have upon three occasions seen a male catch another male and insert spermatophores into his mantle chamber. Two of the three instances were between the same individuals, the second performance being only a few minutes after the first. In each of these cases the male seized made great efforts to get away and finally to get hold of the male that was holding him but was unsuccessful. Upon killing the male that received the spermatophores, sperm reservoirs were found attached to the base of the left gill and to the adjacent visceral mass. Such exceptionally active males may copulate repeatedly with a single female. In a few cases this has been carried so far that the female has actually been killed. Even after the female has beco-me entirely inactive and apparently dead the male may copulate with her several times. In one case, a male that had been several days without food, after copulating with a weakened female, retained his hold and killed her by eating a considerable hole through the mantle.

The male always uses the same arm for transferring the spermatophores. This arm, the left ventral, is not greatly modified, but a short distance from its tip some of the suckers, especially those in the row farthest from the midline of the body, and a ridge between the rows of suckers show modification (fig. 4, h). The peduncles of a dozen or more of the suckers of the outer row are considerably elongated and the sucking discs of a few, (six or eight) are greatly reduced in size or entirely absent. In both directions from these, the discs become increasingly normal until no modification is apparent. The suckers of the row toward the midline of the body are somewhat modified, the peduncles being somewhat shorter than those of the other suckers in the row, and the suck


ing discs somewhat smaller, but in none of the suckers of this row are the sucking discs entirely absent. A glandular plaited ridge extends lengthwise between the suckers of this region and gives off branches that join each of the peduncles. This ridge is highest and broadest opposite the suckers that are most modified and gradually disappears as the suckers become normal. At its highest point it hag about the same elevation as the shortest modified suckers, which are adjacent. Sections of the modified portion of the arm show that the ridge and suckers mentioned are covered by a thick columnar epithelium that stains deeply. Many of these epithelial cells are filled with large rounded granules that stain with eosin. The cells that cover other portions of the arm are flattened or cubical, do not stain very deeply, and do not contain granules. It seems probable that the cells of the hectocotylized region secrete a substance that aids the arm in holding the spermatophores. The modified suckers probably make the bending and grasping necessary for the transfer of the spermatophores more easily accomplished.

The positions of the animals during copulation are rather hard to determine as the whole process generally does not occupy more than ten seconds and during this time the animals are usually swimming and the arms are changing positions, but by carefully focusing attention during different acts upon first one arm and then another, the positions and movements have been determined with some accuracy I think. Fig. 1 represents the positions of the animals while the arm of the male that bears the spermatophores is inserted into the mantle chamber of the female. This figure is the result of my conception of positions after having carefully observed copulation more than twenty times. Since drawing the figure many other observations have been made and the positions always seem to be essentially as given.

The male usually grasps the female while both are sw^imming. Occasonally the female maybe resting on the bottom in the characteristic attitude, with the tips of the arms and the posterior end of the body touching and the head and funnel region somewhat elevated. If not swimming, she usually, when grasped, starts to swim, but in a few cases that I have observed she made no effort


and left the bottom only as she was lifted or turned by the male. In every case the male attached from the left side of the female. He frequently swims close to her and brushes the tips of his arms along her head and mantle. Just before attaching, if both are swimming, he sinks slightly beneath her and grasps her body with his arms so that his right arms are all on the right side of her body and his left arms are all on her left side. The body of the male is seldom exactly ventral to the female but usually slightly toward the left side. Attachment is evidently made as nearly as possible in the required position but when the female darts ahead, as she frequently does, the male is likely to attach too far posteriorly . In such cases he does not let go his hold but crawls rapidly forward, arm over arm, until the right position is attained. Naturally the positions of the individual arms differ somewhat but in general the arrangement is reasonably well shown in fig. 1.

For about a second after his position is attained the arms seem busy in making firm attachments, then with a rapid sweep his left ventral arm is passed by the end of his funnel and is immediately inserted into the mantle chamber along the left side of her neck, near the funnel. During the act both animals are usually quite without color and the inserted arm of the male may be seen fairly distinctly inside the mantle chamber.

The movement of the arm past the funnel is rapid and only once have I actually seen the grasping of the spermatophores and their transference to the mantle chamber. In this case while watching squid in an aquarium that was placed so the squid were between me and a window, a male grasped a female that was resting on the bottom. The female, contrary to the usual custom, did not move. As the male had attached far back on the body, opportunity was given me to get into position for observation before the male could crawl forward. As the female made no attempt to get free, the male seemed far more deliberate than usual. Just before the arm was passed by the end of the funnel, the penis could be seen protruding into it. A number of spermatophores appeared in the opening of the funnel and were grasped by bending the tip of the arm around them. With a rapid sweep of the arm they were immediatelv inserted into the mantle chamber of the female


where they were held about five or six seconds. The arm was then withdrawn and in about five or six seconds more the empty cases of the spermatophores passed out of the funnel of the female with a respiratory jet of water. These spermatophore cases were pretty closely attached to each other by having the tubes of their ejaculatory apparatus twisted together. They were recovered and found to be 41 in number. To the cluster were attached five sperm reservoirs. Examination of the female later showed that most of the other reservoirs were attached near the end of the oviduct. While the number of spermatophores used in an act of copulation varies greatly, the observations that have been made, indicate that this may be a little, but not much above the average.

The animals nearly always separate almost immediately after the arm is withdrawn. Beside the male which started to eat the female, a very few individuals have remained attached for from some seconds to nearly a minute after the arm has been withdrawn.

After copulation the female frequently seems considerably fatigued and may settle to the bottom and rest some minutes before becoming active again. I am rather inclined to think that this is due to her struggles, for when the female remained quiet, the apparent fa,tigue did not seem so marked. The male does not seem greatly affected, but is likely to continue to be very active for some time.

The copulation that leads to the filling of the sperm receptacle on the buccal membrane does not seem to be preceded by special movements. Although I have observed it several times the absence of preparatory movement has left me rather unprepared for the observations that must necessarily be made so quickly, for in this, as in the other form of copulation, the animals are seldom in contact more than ten seconds. In the cases I have observed my attention has been attracted by the sudden dart of one squid, the male, from one end of the aquarium directly at another, the iemale. Before the dart the squid face each other, and are separated by thirty centimeters or more. The movement was always exceedingly rapid and was probably due in each case to the


expulsion of a single jet of water. The male seemed to reach the female before she had time to move much, although she has given me the impression of attempting to dodge as if frightened. The two animals become attached head to head with their arms intermingled, each grasping the other (fig. 2). Then as in the other method, the male sweeps his left ventral arm past the end of the funnel and grasps the bundle of spermatophores. These are immediately thrust between the ventral arms of the female and held there for a few seconds. The animals then separate and examination has shown fresh sperm reservoirs attached to the receiving depression on the buccal membrane of the female. The empty cases of the ejaculated spermatophores may be held between the arms several minutes but they are finally dropped. Here, as in the other method of copulation, only the sperm reservoirs are retained for any length of time.

The spermatophores begin to ejaculate immediately after leaving the penis and the whole process is completed in a very few seconds. Pulling the filament attached to the ejaculatory end of a spermatophore is all that is needed to start its ejaculation. As the ejaculatory end of the spermatophore leaves the penis last and, as the spermatophores in the penis and the spermatophoric sac are imbedded in a viscid secretion, there is every reason to believe that the pull given the spermatophores by the arm with which they are grasped, when this arm starts to transfer them from the penis to the mantle chamber or to the buccal membrane, is sufficient to start ejaculation. The arm carries the spermatophores into the position necessary for the attachment of the sperm reservoirs while they are ejaculating and holds them there until the ejaculation is complete and the reservoirs are attached.

The structure of the spermatophores and the mechanics of ejaculation which lead to the attachment of the reservoirs will be treated in another paper. It should, however, be understood that the spermatophores are never attached as such, but they ejaculate and the sperm reservoirs are attached. As the reservoirs are attached by cement carried inside the spermatophores and liberated by the ejaculation, they may be stuck anywhere.

The sperm slowly escape from these reservoirs and may then


become free in the water, as when they are attached in the mantle chamber, or may be stored in a special receptacle, as when they are attached in the special depression on the outer buccal membrane. They are mixed with a viscid secretion in the reservoir and probably also before entering the reservoir, although I am not certain about the latter. The epithelium of the region is abundantly supplied with goblet cells which very possibly supp'y secretion for this purpose.

The depression in which the sperm reservoirs are mostly attached is supplied with a deeply staining columnar epithelium which is covered by a mass of rather hard material, evidently secreted by these cells, that shows distinct markings parallel with the surface of the epithelium (figs. 11 and 12). These markings seem to indicate that the material is secreted intermittently and thus is formed in layers. This material forms a suitable place for attachment of sperm reservoirs and probably serves no other purpose. Reservoirs are sometimes attached to other portions of the buccal membranes or to the tentacles but they are far more abundant in the depression than anywhere else. The sperm that escape from the reservoirs that are not attached in this depression probably do not find their way into the sperm receptacle.

The sperm receptacle has the shape of a compound alveolar gland (fig. 11). It is imbedded in the outer buccal membrane and opens on the inner surface of this membrane at a point opposite the junction of the two ventral arms. Simple cubical epithelium lines the deeper alveoli of the receptacle, and cubical epithelium with many goblet cells the portion nearer the opening. Some, but not many, cilia have been seen on these cells. The killing fluids used may not have preserved them, for the tails of sperm in the reservoirs are not often individually visible in the sections. With the exception of the tails of the sperm and the possible cilia on the cells the material gives evidence of good preservation. A layer of muscle fibers surrounds the receptacle as a whole and bundles of fibers run between and around the individual alveoli.

It was not determined whether the sperm are active in the interval between their discharge from the reservoirs and their en JOURXAL OP MORPHOLOGY, VOL. 22, NO. 2


trance into the receptacle or not. That they are not active while stored in the receptacle is shown by opening filled receptacles on dry slides. The sperm are invariably quiet, but immediately become active when sea-water is added. In specimens killed soon after copulation, sections show the sperm entering the receptacle in narrow streams and not spread out as one might expect them to be if the sperm were active (fig. 11). It was not possible to remove all the sea-water from living specimens in which the receptacles were being filled without causing disturbances in the vicinity of the reservoirs and that made it impossible to determine the normal condition of the sperm in transit from one to the other. In the sections that show sperm entering the reservoir the tails all point in the same direction, as would be the case if they were not swimming actively but were being moved by an outside force. The heads go first and the tails all trail behind. Swimming sperm usually move in all directions but there may be some directive cause that would account for their positions even if they are stored through their own activities.

As previously stated, a female that is nearly ready to deposit her eggs can be told by her peculiar nervous movements and the way she manipulates her arms. Frequently the borders of the accessory nidamental glands, which are very red at this time, may be seen through the semi-transparent mantle and thus form a further indication that the eggs are nearly ready to be deposited. The nidamental and oviducal glands of such an animal are always somewhat, and frequently greatly, enlarged. Immediately after the eggs have been deposited these glands, while still large, are soft and flabby.

As is well known the squid deposits her eggs imbedded in strings of a jelly-like substance which vary in size with the size of the animal depositing them but which probably average about 8 mm. in diameter and 90 mm. long. The jelly consists of an inner mass that surrounds the eggs and a thick, rather tough but still jellylike sheath that forms the outer covering. The inner jelly is secreted inside the oviduct by the oviducal glands. The outer jelly is secreted by the nidamental glands and is apparently moulded into shape as it passes through the funnel. The accessory nidamental


glands, which he j List in front of the anterior ends of the nidamental glands and open by wide openings near the narrow openings of these glands, are very active during this period and secrete a viscid material. What the special function of this secretion is has not been determined. It would seem from position and activity that the secretions from both sets of glands must be mixed as the}' are poured out.

Until recently eggs have not commonly been deposited in aquaria at Woods Hole. This maybe due to the way the animals have been handled. Squid will not stand rough handling, either in capture or transportation, and live well in aquaria afterward. When captured in fish traps, quickly and carefully transferred to live cars where they are supplied with an abundance of water, and transported to the aquaria with as little excitement and as good water as possible, they may be kept several days in pretty good condition, but they wear themselves out by constantly bumping against the walls of the aquaria and are not vigorous many days. During each of the months I have worked at Woods Hole, June to late September, specimens have been obtained that have deposited eggs in aquaria. During the first three months specimens ready to deposit eggs are rather easy to get. In September only a small proportion of those captured still contained eggs.

Eggs are somewhat more frequently deposited in aquaria at night than during the day, but this may be due to the frequent if not nearly continuous disturbance to which they are subjected during the day in a laboratory where many people are working. The usual number of strings deposited by a female in what would seem to be a continuous laying period ranged from one to six. These strings were commonly delivered from fifteen to forty minutes apart, the time between any two strings being quite variable in an individual. One specimen, however, deposited twentythree strings in an hour and thirty-five minutes. These were laid during a comparatively dark day when the laboratory was quiet. Possibly the small number deposited by other females was due to disturbance.

The end of the egg string begins to protrude from the end of the funnel while the female rests upon the bottom in the attitude


habitually assumed by resting squid (fig. 5). When from one to two centimeters of the egg string protrudes from the funnel, the female leaves the bottom and begins to swim slowly backward. This swimming is apparently due both to movements of the tail fin and to small jets of water forced from the funnel along the sides of the egg string. The jets of water cause the egg string to be protruded gradually. The protruding end is now caught by the ends of the two dorsal arms, which are bent ventrally between the other arms for this purpose (fig. 6), and as the string is ejected from the funnel, it is drawn between the circlet of arms. It usually takes from half a minute to a minute for the egg string to pass through the funnel and to disappear between the arms. It is then held between the arms about two minutes or sometimes a little longer. While the string is held between the arms it is completely enclosed by them and their free ends keep twisting around each other. In this position they form a cone with the apex at the ends of the arms (fig. 7). At other times the arms are held so they form a dorso-ventrally flattened expansion that serves somewhat as a rudder or anterior fin. The arms while enclosing the eggs are never entirely still but move slightly upon each other and are probably busy in moving the string about. While the string is thus held the animal slowly swims back and forth, never rapidly but continuously.

Toward the end of the period during which the string of eggs is held, the animal shows an increasing tendency to turn the body into a nearly perpendicular position to bring and keep the tips of the arms in contact with the bottom (left animal in fig. 9). With the arms held quite rigid and the tail fin moving rapidly she goes bounding along on the tips of her arms, dorsal side foremost, with a movement somewhat similar to the bounces that may be obtained by pushing a lead pencil, held by one extremity and slightly inclined from the perpendicular, over a table. This action is generally repeated several times. She occasionally catches hold of objects with her suckers, finally catches some object firmly, draws down into close contact with it for two or three seconds (right animal in fig. 9) and, when she releases her hold, leaves the string of eggs fastened to the object she had laid hold of. At


this time the jelly of the string is soft and sticky. It hardens quite rapidly and soon will not stick to objects, but at this time it adheres readily. The position of the string when taken between the arms indicates that the string is finally stuck by the end that first leaves the funnel.

After sticking a string of eggs the female rests upon the bottom some minutes before another string makes its appearance. She usually selects some protruding object like a stone, shell, or water pipe upon which to stick the egg strings. Having stuck one string she usually, but not always, returns to the same place to stick later strings. If strings are present when a female begins to deposit she usually attaches to these strings, or to nearby objects. This no doubt accounts for the very large clusters, with strings containing eggs in various stages of development, that are sometimes found. Upon several occasions clusters in fishtraps a ad live-cars have been found that would not go into an ordinary ten-quart pail. Such clusters are of course formed by many females.

It is evident that the eggs may be fertilized in the oviduct, in the mantle chamber, or between the arms. Examination of the contents of the oviduct have in no case given evidence of sperm. Eggs taken from the oviduct may easily be fertilized by placing them in sea-water containing sperm, but in no case did eggs taken from the oviduct show evidence of fertilization although many sperm reservoirs that were giving off active sperm were attached to the walls of the oviduct and to surrounding organs.

There can be no doubt, however, that eggs may be and are fertihzed in the mantle chamber and also between the arms. That the eggs may be fertilized in the mantle chamber is indicated by reason rather than by obervation. When sperm reservoirs are attached in the mantle chamber the sperm are constantly liberated in the water in this chamber as long as the supply lasts. The eggs upon leaving the oviduct also pass into the mantle chamber and, as before stated, when eggs and sperm are mixed in seawater, fertilization results.

That fertilization may be delayed until the egg string is formed and held between the arms is indicated by observations made on


the female already mentioned that deposited twenty-three strings. She was in a rather large aquarium with a number of other squid. Copulation had occurred several times but this particular squid, . which had been under observation some hours, had not been seen to copulate. Dissection later showed that there were no sperm reservoirs attached in her mantle chamber. Because of disturbance she upon six occasions failed to get the egg string between her arms. When she reached for the string with her dorsal arms she was each time disturbed so she dropped the string and ejected it directly into the water. Four of these strings were recovered as quickly as possible after they were dropped, and placed in dishes of fresh sea-water where the proportion of fertilized eggs could be determined. From 40 to 50 per cent of the eggs in the strings developed. More than 99 per cent of the eggs m strings that had been held between the arms and then placed in similar dishes developed. As already mentioned there had been copulation among other squid in the aquarium and as the reservoirs were attached in the mantle chambers there must have been many free sperm in the water of the aquarium. It seems probable that enough of these sperm reached the strings that were dropped, before they could be removed from the aquarium, to fertilize a portion of the eggs. Microscopic examination of these strings immediately after they were dropped revealed very few sperm, but the strings that were held between the arms were swarming with them. Sperm were able to penetrate and move actively about in the soft jelly of a recently formed string, but the jelly soon hardened so fresh sperm brought in contact with it were not able to work their way in.

A "curious bit of habit reflex was exhibited by this squid each time she dropped a string of eggs. Immediately after the disturbance she took the attitude she would normally have taken had the egg string been successfully lodged between the arms. The arms were held in the form of a cone, the tips were twisted together and she passed on through each of the succeeding phases even to drawing down tight against an object as if to attach the egg string that had never been between the arms. After this she rested until the next string was formed, but she never interrupted the orderly


sequence of her activities because she had accidentally lost a string of eggs.

The methods of copulation of cephalopods have attracted the attention of observers from very early times but the act of copulation has not been actually seen for many species and where observations have been made they have for the most part been incomplete. Aristotle makes several statements regarding the breeding habits of cephalopods. It is quite possible that he saw something of the act of copulation for some species, but his statements are hard to follow and are evidently inaccurate. The most important statements are here quoted to show the curious medley of facts and fiction. In chapter 5, book 5, he says:

1 . All the malacia, as the polypus, sepia and teuthis, approach each other in the same manner, for they are united mouth to mouth: the tentacula of one sex being adapted to those of the other; for when the polypus has fixed the part called the head upon the ground, it extends its tentacula which the other adapts to the expansion of its tentacula, and they make their acetabula answer together. And some persons say that the male has an organ like a penis in that one of its tentacula which contains the two largest acetabula. This organ is sinewy, as far as the middle of the tentaculum, and they say it is all inserted into the nostril of the female.

2. The sepia and loligo swim about coiled together in this way, and with their mouths and tentacula united, they swim in contrary directions to each other. They adapt the organ called the nostril of the male to the similar organ in the female; and the one swims forwards, and the other backwards. The ova of the female are produced in the part called the physeter, by means of which some persons say that they copulate.

Again in chapter 10, book 5, he says:

1. The malacia breed in the spring, and first of all the marine sepia, though this one breeds at all seasons. It produces its ova in fifteen days. When the ova are extruded, the male follows, and ejects his ink upon them when they become hard. They go about in pairs. The male is more variegated than the female, and blacker on the back. The sexes of the polypus unite in the winter, the young are produced in the spring, when these creatures conceal themselves for two months. It produces an ovum like long hair, similar to the fruit of the white poplar. The fecund


ity of this animal is very great, for a great mumber of young are produced from its ova. The male differs from the female in having a longer head, and the part of the tentaculum which the fishermen call the penis is white. It incubates upon the ova it produces, so that it becomes out of condition, and is not sought after at this season.

Part of these statements, such as The sepia and loligo swim about coiled together in this way, and with their mouths and tentacula united, they swim in contrary directions to each other" would seem to be based upon such observations as could be made from above but the further statement that they adapt their nostrils (funnels) together, probably indicates the ease with which observation and supposition can be mixed. It is not necessary further to analyze Aristotle's statements. No doubt much was based upon fishermen's stories but he evidently did study the anatomy and habits of these animals and recognized the probability that one of the arms of the male is used in copulation.

While the modified arm of the male thus early received attention, the true hectocotylus that separates entirely from the male and attaches itself in the mantle chamber of the female escaped notice for many centuries. To quote from the Cambridge Natural History :

The typical hectocotylus seems to have entirely escaped notice until early in the present (last) century, when both Delle Chiaje and Cuvier described it, as detected within the female, as a parasite, the latter under the name of Hectocotylus octopodis. KoUiker, in 1845-49 regarded the Hectocotylus of Tre?noctopus as the entire male animal, and went so far as to discern in it an intestine, heart, and reproductive system. It was not until 1851 that the investigation of V^rany and Filippi confirmed a suggestion of Dujardin, while H. Miiller in 1853 completed the discovery by describing the entire male as Argonauta.

While nearly all male cephalopods show some modification of one or more arms, the only ones that have been reported with detachable arms are Argonauta, Ocythoe, and Tremoctopus.

Extended studies have been made on the modification of the arms of cephalopods, and there have been a few observations upon


the functional activities of these arms, but most of the observations have consisted in finding sperm reservoirs recently attached to various portions of females.

In 1869 Lafont described copulation in Sepia. A translation of that portion that deals with the act itself is as follows:

In copulation the male and female precipitate themselves upon one another, hold together by their arms which are twined together, and remain thus, mouth to mouth, for a variable time, which may last for two or three minutes. This act is followed in the female by a state of very marked general prostration, while in the case of the male the general excitation is greatly prolonged and for a considerable time it keeps the splendid appearance these animals show as the result of the accomplishment of the function of reproduction.

He supposed that while the animals were attached by their arms, head to head, the male ejected a packet of spermatophores, which ejaculated while in his mantle chamber and the sperm reservoirs were then thrown from the funnel of the male into the branchial chamber of the female with the current of water entering her branchial chamber.

Sepia, like Loligo, has a receptacle for the storage of spermatozoa in the buccal membrane, and the position observed by Lafont of animals attached head to head was doubtless a true position of copulation, but it seems probable that the spermatophores were not disposed of in the way suggested, but were transferred to the buccal membrane by one of the arms of the male. Lafont found sperm reservoirs attached in the mantle chamber of the females near the mouths of the oviducts, so it seems probable that in this form, as in Loligo pealii, both methods of copulation occur.

Racovitza (1894, a) observed copulation inSepiola. The male seized the female, turned it over and inserted his first pair of arms into the mantle chamber. Copulation lasted eight minutes during which the female struggled to free herself. He speaks of the spermatophores being fixed on the folds of a large pocket situated on the left side of the pallial cavity of the female. These ejaculate and the freed reservoirs deliver their sperm into the pocket


which m turn ejects them (from his description I take it they are not stored up in this pocket as in the receptacle on the buccal membrane of a squid) into the pallial cavity where they are supposed to meet the eggs as they are laid.

The most complete account of copulation that I have seen for anycephalopod was given by Racovitza in 1894 (b) for Octopus vulgaris. He observed copulation in an aquarium and gives a figure showing the positions of the animals. The copulation differs markedly from that of Loligo, as might be expected, for Octopus has a hectocotylized arm that is much more differentiated than that of Loligo. The animals were some distance apart in the aquarium. The male reached over with the hectocotyhzed arm, which for this species is the third on the right side and, after caressing the female with its tip, introduced its end into her mantle chamber by the side of the funnel. Here it remained for something more than an hour. During this time the female remained quiet, except for certain spasmodic movements, while the male showed only slight movements of the hectocotylized arm which were supposed to be associated with the movements of spermatophores down the longitudinal groove of this arm. Although it was not possible actually to see the spermatophores in transit, examination of the female after copulation showed numbers of the sperm reservoirs, derived from the ejaculated spermatophores, within the oviducts.

Evidently there are at least three methods of copulation practiced by cephalopods. A method of caducous hectocotylism in which the charged hectocotyl is liberated in the mantle chamber of the female; a method in which the arm does not liberate any special portion but is so modified that it can transfer spermatophores by a mechanism within itself to the region of the oviduct of the female; and finally a shght modification of the arm that simply enables it to grasp the spermatophores which are then transferred directly to the female by moving the arm. Where the latter method is employed there may be two kinds of copulation, as in Loligo pealii.

Racovitza, (1894, c) in commenting on the copulation of Rossia believes that, although special receptacles are found outside the


mantle chamber of this species, they cannot be considered as normally functional. He seems led to this conclusion by finding sperm reservoirs attached to various portions of the bodies of the animals as well as in the immediate neighborhood of the mouths of the oviducts. It would seem more likely in the light of the observations here recorded for Loligo, that a copulation that leads to the filling of these receptacles is normal and that the sperm so stored may be used in fertilizing the eggs.

It is certainly hard to conceive by what steps a complicated method of transferring sperm that has led to the formation of a hectocotylized arm and complicated spermatophores might be perfected. The modification of different arms for copulation by different cephalopods further increases the difficulty in understanding the history of hectocotylism as a whole.

While evidence that bears directly upon the history of the hectocotylism seems to be lacking, such complications are so frequently considered to be impossible to explain by known evolutionary factors that it may be well at least to consider the great difficulties presented by such structures. It must not be supposed that in so doing I put myself in the position of defending a thesis. This would be too much like the methods employed by many of the Greek philosophers who needed little or no basis of fact upon which to build. My only reason for considering the matter here is to show that, with all the difficulties, the condition of hectocotylism among modern cephalopods cannot be considered beyond the possible range of evolutionary factors.

Among the Dibranchiata the arms that show hectocotylism are the first, the third and the fourth on both sides of the body. Sometimes more than a single one is affected. In such cases the modified arms may be symmetrically placed on the two sides of the body, or they may be adjacent arms on the same side of the body. Steenstrup attempted to base the classification of cephalopods upon their hectocotylized arms but Brock and Hoyle have shown that forms whose general body structure would seem to indicate relationship, do not always have homologous arms modified.

While the arm is usually constantly on one side for all members of a genus, unless both sides are modified as not infrequently hap


pens, a genus whose general body structure indicates near relationship may have the similar arm of the other side modified. The position of the arm on the right or left side of the body is not generally considered very significant. The somewhat frequent occurrence of genera showing hectocotylism of arms symmetrically placed on the two sides of the body may indicate a primitive paired condition that has been replaced among the majority of existing cephalopod genera, by specializing on one side and dropping out on the other. Whether this accounts for the condition or not, the shifting of a modification from one side of the body to the other, sometimes involving modifications of other body structures and sometimes apparently not, is not uncommon among animals, and even if not easily explained, evidently has no very great phylogenetic significance. Shifting in series is not so common and when we find in the same family, genera with the fourth and others with the first arm hectocotjdized it becomes difficult to imagine ancestral conditions that made this posisble.

Wherever known, male cephalopods use one or more of the arms to transfer sperm to the female. Copulation has not been described for many of the species but the presence of more or less modified arms in more than half the recogaized families may be taken as an indication that either these animals or their ancestors used their arms in copulation.

If the spadix of Nautilus is used in copulation we have a possible indication that a number of arms may have been employed in the transfer of sperm by primitive cephalopods. It is of course possible that all the arms were used for this purpose and that the present diversity can be accounted for by the specialization of one or the other of the arms involved in this primitive condition. This, however, does not seem reasonable when the diversity within the limits of a single family is considered.

The arm that is used, and the way in which it is used, is associated with the character of the spermatophores and the position of their final discharge. The Octopoda show the greatest structural modification in their hectocotylized arms. While two of the families of this group give no evidence of hectocotylism, none of the genera of the remaining families are known to be free from it.


and wherever found it is always the third arm that is involved. Sometimes this arm is on the right and sometimes it is on the left side. In three genera it is known to be caducous and in a fourth (Alloposus) it is supposed to be. In the remaining genera in which the hectocotylized arm has been studied, the modifications, while not resulting in the actual separation of the arms, are of an extensive nature. In Octopus, for instance, they involve not only changes. in size, form, and the condition of suckers, but a special groove is present through which the spe matophores are supposed to be carried from the base, presumably from the penis to the tip. The tip in turn is modified so it is supposed to function in placing the spermatophores in position for ejaculation.

The Decapoda do not show such extensively modified hectocotylized arms. The changes are here chiefly confined to some of the suckers and their immediate vicinity. In Loligo this modification apparently serves to aid the arm in grasping the spermatophores, which are then transferred by the movement of the arm. While the actual grasping of the spermatophores has not been previously observed, there can be little doubt that other forms of the Decapoda use the arms in a similar manner. Where copulation has been observed the movements of the arms indicate that they are used in the transfer, and the positions of the sperm reservoirs that have been found attached to the females indicate that some arm must have functioned in getting them into position. As there is no special transferring mechanism, this must have been accomplished by the free movements of the arms.

Where structural modification is shght and the placing o the spermatophores is due to dexterity, there is less difficulty in understanding how the function may be shifted from one arm to another in response to changes in the position of the attachment of the reservoirs on the female, than would be the case were great structural changes involved. It would be much more difficult to understand how there could be a shifting in series of arms as highly modified as those of the Octopoda, where only the modified arm could possibly perform the function.

It mast not be understood that habit formation requiring such dexterity is considered easier to originate than modification in


structure that will perform similar acts. When, however, the habit and dexterity have been acquired, it is not inconceivable that they might be shifted to another closely similar appendage if the position of this appendage becomes more suitable for the purpose. The modification is so slight in the arms of most of the Decapoda, and the modification varies so greatly in the different genera, that it may have been functionally acquired in each case. So far as can be seen it would be mechanically quite possible for a squid to use an unmodified arm, instead of the one that shows the modification, for the transfer of the spermatophores. The spermatophores might not be so tightly or compactly held but the normal suckers would hardly seem to interfere greatly in the performance of the function.

There is still another question involved. Is there any genetic relation between these two methods of transfer and if there be, which, if either, most probably came first?

A special method of copulation that requires the use of arms and complicated spermatophores is not found among animals often enough to make it at all probable that it has arisen in this group more than once, so we can hardly doubt that the two methods are genetically related.

At first sight the squid's method of grasping the spermatophores and transferring them directly might be considered the simpler process, but there is some reason to doubt that this method was at the beginning of the series. While it would be hazardous to say that Octopoda were the ancestors of Decapoda, there is much reason to believe that the ancestors of the latter lived upon the bottom and were far less active than the modern animals. Such animals would not seem to be so well adapted for the transfer of spermatophores by dexterous movements as the more active, freeswimming forms. It is at least certainly true among modern cephalopods that those that show great modifications in the structure of the hectocotylized arms are found entirely among the less active bottom forms. If the method of transferring sperm by means of the arms originated before the Decapoda became free-swimming animals, and this seems the only explanation of its prevalence


among both Decapoda and Octopoda in modern times, it would seem that structural modification probably came early.

Possibly this modification was based upon the use of one or more arms as guides for the transfer of the sperm. It is possible that having first used the arms as guides, structural modifications and dexterous movements were developed as divergent methods. If the two methods form a linear series, there is some reason to think structural modifications came first. It would seem much easier to explain modifications that lead to the change in the structure of appendages for the transfer of spermatozoa, as the grooved hectocotyhzed arm of Octopus or the modified abdominal appendages of certain Crustacea, than to explain a sudden change that would result in a practically unmodified arm functioning by grasping spermatophores of a very specialized kind, transferring them quickly and accurately to the required position and holding them there until they have had ample time to ejaculate and fix their contents. It seems more reasonable to suppose that an arm modified as a machine to perform this process, with its tip serving to place the spermatophores in position, might in time acquire the necessary dexterity and then lose the modifications previously acquired, than to look at this as the beginning of the series. Again we find that in such cases as the squid, where the arm is little modified but very dexterous, there is a special receptacle at some distance from the opening of the oviduct that is norma ly filled with sperm during the breeding season. This would certainly seem to be a comparatively recently acquired receptacle, so the copulation leading to its being filled would also be considered comparatively recent. That this receptacle is concerned in the fertilization of the eggs is shown by observations made while the eggs were being laid.

With no personal knowledge of the breeding habits of other cephalopods than the squid, it would seem more reasonable to consider the method of using the detachable hectocotyl of such forms as Tremoctopus as one extreme, the method used by Loligo in grasping spermatophores and transferring them directly as another extreme and the condition shown by Octopus as the modern greatly specialized product of a modification such as early cephal


opods probably developed. This would mean that the detachable hectocotyl is an extreme specialization in structure and that the modification shown by the squid represents possibly a degeneration in structure but a remarkable specialization in habit.

Why a form should have two methods of copulation is not at all clear. Certainly the introduction of the spermatophores into the mantle chamber to a position near the oviduct is to be considered more primitive than their being placed in a position to fill a receptacle outside of the mantle chamber, but why mantle chamber copulation should be retained after the receptacle has been perfected is not clear. That mantle chamber copulation is not absolutely necessary for the fertilization of the eggs I think is proved by my observations; that it is common is certain. That the sperm receptacle is an improvement over the free attachment of the sperm reservoirs in the mantle chamber is evident from the longer possible retention of the sperm in the receptacle. In a limited period after the sperm reservoirs are freed from the spermatophores, as when deposited in the mantle chamber, the sperm all escape and are wasted unless oviposition takes place in the meantime.


Squid have two methods of copulation. By one method sperm reservoirs are attached in the mantle chamber on or near the oviduct and immediately begin to discharge their contents freely in the water. By the other method sperm reservoirs are attached to the outer buccal membrane and the sperm become stored in a special receptacle in the membrane, which is placed opposite the junction of the two ventral arms and opens on its inner surface.

The left ventral arm of the male is always used in transferring the spermatophores, which are grasped by the arm and transferred by its free movement. Ejaculation of the spermatophores is evidently started by the pull given their filaments when the arm starts to transfer them from the penis to the mantle chamber or buccal membrane. The transfer requires rapidity and dexterity and the spermatophores are held in position until ejaculation is complete and their sperm reservoirs are fastened. As many as forty spermatophores may be transferred at a time.


The egg strings are composed of two kinds of jelly. One kind is supplied by the oviducal gland and the other by the nidamental and probably accessory nidamental glands. The string is apparently molded into shape by passing through the funnel. The jelly is at first soft and sticky but soon becomes tough and loses most of its stickiness.

From the funnel the egg string is drawn between the circlet of arms, where it is held two or more minutes. In sticking the string the female grasps some object with her arms and draws down tight so the string is evidently crowded against it. When she releases her hold the string is left sticking to the object.

Fertilization evidently does not take place inside the oviduct. It doubtless may take place in the mantle chamber when sperm reservoirs are present there, and is known to take place while the egg string is held between the arms. The sperm are hberated from the receptacle while the eggs are between the arms.

Notwithstanding complications, the conditions of hectocotylism shown by cephalopods need not be considered beyond the influence of factors of evolution.


The cephalopod literature is very extensive. Only those papers directly referred to are here given.

Aristotle History of animals. Trans, by Richard Cresswell. 1891.

Brock, J. 1882 Anat. u. Syst. d. Cephalopoden. Z. f. wiss. Zool. 36. 1884 Mannchen d. Sepioloidea lineata. Z. f. wiss. Zool. 40.

HoYLE, W. E. 1907 Presidential address of Zoological Section. Rept. Brit. Ass. Adv. Sci.

Lafont, M. a. 1869 Observations sur la fecundation des Mollusques Cephalopods der Golfe de Gascogne. Ann. des Sci. Nat. (5) 11.

Racovitza, Emile. G. 1894a Sur I'accouplement des quelques Cephalopods Sepiola rondeletii (Leach), Rossia macrosoma (d. Ch.) et Octopus vulgaris (Lam.). Comp. Rend. I'Acad. des Sci. 118.

1894b Notes de Biologic. I. Accouplement et Fecondation chez rOctopus vulgaris Lam. Arch. d. Zool. Exper. et Gen. (3) 2. 1894c Notes de Biologie. III. Moeurs et Reproduction de la Rossia macrosima (D. Ch.). Arch. d. Zool. Exper. et Gen. (3) 2.

Steenstrup, J. J. S. 1856-57 Hectocotyl. hos Octopodstsegterne. Vid. Selsk. Skr. (5) 4, Translated Ann. N. H. (2) 20. 1881 Sepiadarium og Idiosepius. Vid. Selsk. Skr. (6) L. 1887 Nota; Teuthologica? 7. Overs. Vid. Selsk. Forh.




All of the figures that represent the attitudes of squid were drawn from memory after repeated observations. While each figure is thus really a composite, and must represent impressions received rather than the actual positions of particular individuals, much care has been given to the preparation of the figures and it is believed that the general attitudes are reasonably well represented. Sexually mature squid are usually as much as 15 cm. and may exceed 40 cm. in length.


hni, inner buccal membrane n, nidamental gland

bmo, outer buccal membrane na, accessory nidamental gland

d, depression in which sperm reservoirs o, oviduct

are attached r, rectum

g, gill s, sperm reservoirs (ejaculated from sper h, modified (hectocotylized) portion of matophores)

arm sr, sperm receptacle

j, jaws sro, opening of sperm receptacle



1 Copulating squid showing the positions taken by the animals when the spermatophores are inserted into the mantle chamber. The figure shows the animals during the period the arm of the male is inserted in the mantle chamber of the female. Drawn from memory after many observations.

2 Copulating squid showing the positions of the animals when the spei'matophores are placed so that their reservoirs become attached to the outer buccal membrane. The figure .shows the male in the act of grasping the spermatophores with the tip of his arm as they are ejected through the funnel. Drawn from memory after many observations.

3 A common attitude of a sexually excited male. The arms are not kept rigidly in a set position, but are frequently spread as shown in the illustration and held thus for from a few seconds to a minute or more at a time. The drawing is based upon sketches made of active animals.

• 4 Photograph of the two ventral arms of a male squid, showing the slight modification (h) consisting of enlarged peduncles, reduced sucking discs and a ridge between the suckers, toward the tip of the left arm. The wrinkles on the arms are due to shrinkage. A bit of the outer buccal membrane shows between the arms. The arms from which the photograph was made are 9^ cm. long.



5 A female at rest with the egg string beginning to protrude. Drawn from memorj' and hurried sketches after many observations.

6 A female after she starts to swim, reaching for the egg string with her dorsal arms. With these arms she draws the string between the circle of arms as it is ejected from the funnel. Drawn from memory after many observations.

7 A swimming female, showing the positions of the arms while they surround the egg string. They are held in this position, with the tips somewhat twisted together, for two or three minutes. While the arms closely surround the egg string they show slight individual movements that may be of service in moving the egg string so sperm will be more evenly distributed over it. Drawn from memoiy after many observations.

8 A female squid with the mantle cut and spread o])eii and the arms separated to show the position of attached sperm reservoirs (s) on the oviduct (o) and the sperm receptacle (sr) in the outer buccal membrane.



lill.MAX A. DHKW

mi'HOI.OCY, V<



9 The specimen on the left 8ide shows a female in the position she assumes as she bounces over the bottom on the tips of her arms just previous to selecting a place for sticking the egg string. The specimen on the right side shows the position of a female during the act of sticking an egg string to a rock. Only a few seconds are required to stick the string. The positions of the animals are drawn from memory after many observations.



10 Jaws and buccal moinbrane of a female squid, with the out er luemliranc [b/no) pulled ventrally to expose the sperm receptacle (the opening of which is shown at .sTo) and the adjacent depression (d). Several sperm reservoirs (.s), ejaculated from spermatophores, arc shown attached in the depression. Magnified about 7 diameters.

11 Section of the outer buccal membrane taken through the sperm receptacle (.s/-). This was taken from an animal shortly after the sperm reservoirs (.s) had been attached and shows sjierm in transit from reservoirs to receptacle. Magnified about 22 diameters.

12 Section through the e])ithelium and secretion lining the depression in which sperm reservoirs are attached. Magnified about 300 diameters.

13 Section through an alveolus of a sperm receptacle. The clear spaces in the epithelivmi are goblet cells. Traces of the Hagella on the sperm and possibly cilia on some of the epithelial cells were visible but they were not definite enough to be put in the drawing. Magnified about 300 tliameters.






From the Hull Zoological Laboratory, University of Chicago



In many animals one of the immediate effects of fertilization is to cause the egg to throw off a membrane, which is therefore known as the fertilization membrane. This is the case for instance in the eggs of echinids and nematodes. In other cases, where a definite vitelline membrane exists prior to fertihzation, cortical changes occur in the egg immediately after insemination, leading to the formation of a space, the so-called perivitelline space, between the protoplasm of the egg and the vitelline membrane. This is the case for instance in the eggs of at least many annehds, molluscs and vertebrates, and it is unquestionably a more common phenomenon than the formation of a fertilization membrane. There can be little doubt that these apparently different phenomena are simply varying expressions of a change in the cortex of the egg, which is of the same nature in all cases. Loeb's studies ('09) have thrown much light on the nature of these cortical changes. In the case of the egg of Nereis they are relatively obvious in their character and readily followed.

The ovocyte of Nereis is somewhat flattened in a polar direction, measuring about 87.5 x 100^; it is girdled by a double equatorial zone of large oil drops. The large germinal vesicle is central and somewhat elongated in a polar direction.

In his description of the unfertilized egg, Wilson ('92) distinguished two membranes: a delicate outer vitelline membrane,



and a subjacent membrane or layer, about 6-7^ in thickness, which he called the zona radiata. As will appear from the sequel however, the latter is not a membrane in the usual meaning of the word, but a cortical, coarselj alveolar layer of the egg. Jt is transparent and somewhat granular, and the granules tend to be arranged in radiating lines. There is no perivitelline space in the unfertilized egg.

In sections of unfertilized ovocytes fixed in Flemming's fluid, the zona radiata is seen to be a coarsely alveolar layer with homogenous alveolar contents (fig. 1). The walls of the alveoli are continuous internally with the potoplasm of the egg, and unite externally to form a protoplasmic la^'er applied to the vitelline membrane. The alveoli are closed externally (figs. 1 and 2). The zona radiata is in fact a coarse emulsion or foamstructure.

Unfertilized eggs of Nereis are entirely devoid of jelly and they lie in immediate contact in the sea-water. If India ink be ground up in the water, the particles come in contact with the vitelline membrane. ILsich fertilized egg, on the other hand, is surrounded by a thick layer of colorless transparent jelly; If many eggs are contained in the dish, fusion of the contiguous gelatinous membranes binds the eggs into a mass; the cortical layer (zona radiata) is absent in fertilized eggs, and there is a narrow perivitelline space between the vitelline membrane and the surface of the egg (fig. 3).

The jelly is formed by the extrusion, or diffusion, of the alveolar contents of the cortical layer through the vitelline membrane ; the egg of Nereis, in fact, secretes its own jelly, as may be readily demonstrated in life by inseminating under the microscope with excess of sperm. If the sperm be added to closely placed eggs and a cover glass applied so as to force the eggs into a single layer, aiid the preparation examined with no loss of time, the spermatozoa will be seen in large numbers in contact with the vitelhne membrane. In one or two minutes the spermatozoa are moved away from the surface of the membrane by some invisible repelling substance, and, if the eggs be numerous, the spermatozoa unite in three to five minutes to form lines that bound hexa



gonal areas with the eggs in the centers of the hexagons (fig. A). The substance that sweeps the spermatozoa away from the surface of the eggs is the jelly. Synchronously with its formation, the alveoli of the cortical layer are emptied and the alveolar walls now appear as delicate lines crossing a wide perivitelline space^ (fig. B).

However, not all of the spermatozoa are thus carried out by the secreted jelly, but in the case of each egg a single spermatozoon remains attached to the vitelline membrane. This is very prettily demonstrated if the eggs are under some pressure, so that

Fig. A. Diagram of fertilization with excess of sperm. The outflow of jelly from the eggs has carried the supernumerary spermatozoa away from the surface of the eggs (see text). In the case of each egg the single effective spermatozoon remains attached. From a sketch of the living object.

the spermatozoa are prevented from reaching the eggs above or below. In this case one can discover the single spermatozoon attached to the vitelline membrane in practically every egg (fig. A). All stages of the disappearance of the cortical layer may be readily and rapidly observed. The alveolar walls, however,

1 Wilson ('92) states that "from twenty to thirty minutes after fertilization the striae of the zona suddenly become indistinct and in the course of two or three minutes the zona itself entirely disappears, leaving only the outer membrane." But inasmuch as he was under the impression that the unfertilized eggs possess a transparent, thick, gelatinous envelope like the fertilized ones, he failed to observe the interesting phenomenon of formation of the jelly described here.



remain as delicate strands of protoplasm uniting the vitelline membrane to the surface of the egg. The jelly, therefore, represents the alveolar contents only of the cortical layer of the unfertihzed egg, and the perivitelline space is nothing but the contracted alveoli of the cortical layer filled with fluid. The perivitelline space must, therefore, be regarded as intraovular with a delicate external cytoplasmic wall lining the vitelline membrane; this we may distinguish as the plasma membrane; it is comparable in some respects to the ferflization membrane of sea-urchins.

Unfertilized eggs allowed to stand in the sea-water form no jelly and retain the cortical layer; the germinal vesicle remains intact; but, if they be strongly centrifuged, the jelly forms, the perivitelline space arises, the germinal vesicle breaks down and both polar bodies are formed; but parthenogenetic development, usually at least, does not take place. Similarly, the addition of a fairly strong solution of potassium chloride to the sea-water causes formation of the jelly and the perivitelhne space while the eggs are still in the solution; maturation takes place after transfer to sea-water but cleavage does not occur (in my experiments; cf . Fischer) , though some differentiation may take place without cleavage. It would appear, then, that conditions that so alter the permeability of the plasma and the vitelline membranes as to permit the outflow of the alveolar contents of the cortical layer initiate development, but that the normal continuation of development is dependent on other factors.

In the normal fertilization of Nereis the stimulus of the spermatozoon causes the formation of the jelly and the perivitelline space iong before it has penetrated the membrane; in fact penetration does not take place until 40 to 50 minutes after insemination. However, mere contact of the spermatozoon with the membrane is apparently not sufficient; but actual attachment of at least a single spermatozoon is required ; this is shown by the fact that the effective spermatozoon is not carried away from the membrane with the unsuccessful ones by the outflow of jelly. Yet the effect of the locafized stimulus of the attached spermatozoon is practically instantaneously effective over the entire


extent of the membrane; it is more like an electrical discharge or some other physical disturbance than a chemical effect.

J. Loeb ('09) has formed the hypothesis that the cortical layer of the egg, especially of sea-urchins, is an emulsion which is rendered stable by a third substance consisting of lipoids, especially cholesterin. The emulsion becomes unstable on solution of the lipoids; this enables the albuminous drops, which he conceives to form one phase of the emulsion, to take up water; hence the layer liquefies and the perivitelline space arises; the fertilization membrane is thus formed. Hence, according to Loeb, the action of lipoid-dissolving substances is to cause the formation of the fertilization membrane. Without committing ourselves to these specific views of the nature of the cortical emulsion, which Loeb himself does not hold very strongly, we may admit that Loeb's hypothesis, that the formation of the fertilization membrane is due to the breaking down of a cortical emulsion, fits the case of Nereis very well. If we go further, however, we must note an important lack of agreement with Loeb's hypothesis. As Loeb himself points out, the theory implies that the membrane of the egg is permeable for sea-water and crystalloid substances, and on the other hand impermeable for colloids; in Nereis the contents of the cortical alveoli are unquestionably colloid, as Loeb's hypothesis requires, but it is perfectly certain that they diffuse through the membrane to form the external jelly ; at the same time, unquestionably, sea-water enters to take the space previously occupied by the colloid. The membrane is therefore permeable for both crystalloids and colloids at this time. I have not, however, investigated farther the properties of the egg membranes and must leave this problem to those who are better qualified as physiologists to make such a study.

It would appear that the presence of this colloid substance in the cortex is an inhibition to the maturation of the egg, because as soon as it is removed, maturation processes are set in motion and both polar bodies are formed. In what manner it inhibits is of course problematical. In the egg of Ascaris megalocephala there is a similar excretion of a cortical colloid which forms, in this case, the thick resistant perivitelline membrane. The ap


pearance of the fertilization membrane of echinids might be similarly due to excretion of a cortical colloid which is removed by diffusion and hence is not detected. It is a problem worthy of careful investigation whether the loss of cortical colloids is not the first step in fertilization geuerally.


Two functions of the spermatozoon in fertilization may be sharply distinguished. The first is the initiation of the development and the second is the transfer of paternal qualities to the fertilized ovum (heredity from male parent). The first function alone is under consideration in these experiments.

We have seen that in Nereis the immediate effect of attachment of the spermatozoon is essentially the same as a mechanical shock (centrifuging), or a chemical stimulus (KCl); that is, it causes the breaking down of the cortical emulsion and consequent formation of the gelatinous envelope of the egg. But apparently the resemblance extends no farther, for in the case of mechanical or chemical stimulation the impulse to development is lost or greatly weakened after maturation has occurred; and the eggs do not segment. On the other hand the normally fertilized egg does not stop after maturation, but proceeds with its development in a normal fashion. Now the cause of this difference might be either: (a) because the stimulus of the spermatozoon is qualitatively different from, or stronger than, mechanical or chemical stimulation, or (b) because the fertihzing action of the spermatozoon is not completed with the cortical changes but continues after its entrance into the egg. If the first alternative were correct, then the elimination of the spermatozoon after membrane formation should not prevent the normal cleavage and development of ova which had once been stimulated by it; but if the second alternative were correct and the sperm nucleus were prevented from entering the egg after it had induced membrane-formation, then such ova should proceed no further in their development than those mechanically or chemically stimulated.


I have been able to perform this experiment on the eggs of Nereis and have found that eggs in which the spermatozoon is removed after the cortical changes have occurred proceed but little farther in their development than eggs mechanically or chemically stimulated, and they do not undergo segmentation. Fertilization is therefore still incomplete after the formation of the fertilization membrane.

It will be seen that if the results above indicated be demonstrated, the process of fertilization is obviously something more than a beginning of cytolysis or a mere alteration of permeability of the peripheral cell membrane. It would appear to be a progressive change, starting at the periphery and gradually involving the more central portions of the cell. We would, at least, have to distinguish two stages in the fertihzing action of the spermatozoon, one before and the other after penetration.

I shall consider first the evidence for the statement that in the egg of Nereis elimination of the spermatozoon after membraneformation leaves the process of fertilization incomplete. In the second place I shall note the respects in which the completely fertilized egg differs from the partially fertilized egg, and finally, shall consider the bearing of the facts on the theory of fertilization. Inasmuch as it will be necessary to make frequent comparisons with the normal fertihzation, a brief account of the salient features of this process will be given first.

A. Salient features of the normal fertilization

The egg of Nereis is difficult to fix in a thoroughly satisfactory fashion; owing, no doubt, to the presence of the large oil-drops and yolk-granules, uneven fixation with shrinkage is hard to avoid. The eggs appear likewise difficult of penetration, owing probably to the rather viscid jelly from which they cannot be separated; this also makes any considerable number of eggs very bulky and the killing fluid is apt to be much diluted if used in ordinary amounts. After considerable experimenting with picric acid, corrosive subhmate and osmic acid fixing fluids, I finally found one which gives practically perfect results in all


stages of maturation and fertilization. This is Meves' modification of Flemming's fluid made as follows: chromic acid, 0.5 per cent, 15 cc; osmic acid, 2 per cent, 3.5 cc; glacial acetic acid 3 drops. The eggs were left in the fluid from thirty to fortyfive minutes. Fixation in this fluid causes no shrinkage, the oil is so changed that it is not dissolved out by subsequent imbedding in parafline; the sections stain beautifully m iron haematoxylin, and certain substances are clearly differentiated which can be detected only with the greatest difficulty after fixation in any other fluid tried.

a. The penetration of the spermatozoon. It was noted in the first part of tjiis paper that a single spermatozoon becomes attached to the egg-membrane immediately after insemination, and that the breaking down of the cortical layer, secretion of the jelly and formation of the perivitelline space follow immediately, though the actual penetration of the spermatozoon is delayed forty or fifty minutes.

The act of penetration involves no motile activity on the part of the spermatozoon ; after the latter has become attached to the vitelline membrane all movement of the spermatozoon ceases and it remains absolutely immotile throughout the forty or fifty minutes that elapse before it is taken into the egg. The events of this period as seen in the living egg are as follows :

1. The jelly is formed by outflow of the alveolar contents of the cortical layer-as already noted; although a large amount of jelly is formed in two or three minutes, yet the process lasts ten or fifteen minutes before the deeper alveoli are emptied. There is then a very wide perivitelline space crossed by the alveolar walls which are attached to the plasma membrane, presenting a very striking appearance (fig. B).

2. The protoplasm of the egg immediately beneath the attached spermatozoon then forms a rounded elevation, the entrance cone, which gradually moves across the perivitelline space and comes in contact, and fuses, with the membrane (fig. B, a). This condition is usually fully attained about fifteen to seventeen minutes after insemination.

3. The entrance cone then gradually retracts, drawing the membrane down to form a depression in which the spermatozoon is included. At this stage one may easily imagine that the sper

Fig. B. History of the fertilization-cone as been in the living egg. Four camera drawings of the same egg: —

a Seventeen minutes after insemination,

b Nineteen minutes after insemination,

c Twenty-two minutes after insemination,

d Twenty-four minutes after insemination. The fertilization-cone is shown at the height of its development in a, its gradual recession and the simultaneous formation of a depression in the membrane is shown in b, c and d.

matozoon has been taken into the egg, as it is apt to be concealed in the depression of the membrane; but this is not the case. The stage of best development of the depression, corresponding to


the complete retraction of the entrance cone, is about twenty-two to twenty -five minutes after fertihzation (fig. B, b, c, d).

4. The perivitelhne space then narrows around the entire egg, and the depression of the membrane in which the spermatozoon is seated disappears; in consequence, the spermatozoon again becomes prominent externally.

5. It remains prominent for ten or fifteen minutes (about forty to fifty minutes after insemination), and then disappears rather abruptly within the egg as though some resistance had given away. Its penetration coincides with the late anaphase of the first maturation division; in a few cases it may be a little earlier or a little later.

The egg is changing form at this time and in consequnce the perivitelline space is often widened locally, especially in the animal hemisphere; if this happen in the region of penetration, which may be any part of the egg, strong cytoplasmic strands are drawn out between the membrane of the egg and the point of penetratio.a, showing that the egg membrane and the cytoplasm are actually fused here.

To repeat and extend the observations on the living egg several series of eggs were preserved every five minutes from the time of insemination in Meves' fluid. The study of the sections confirmed and extended the above observation on the living egg as follows :

Ten minutes after insemination the entrance cone is quite well formed and the spermatozoon is clearly seen outside, separated from the entrance cone only by the thickness of the vitelline membrane with which it is in contact.

Fifteen minutes after insemination essentially the same condition persists. The entrance cone is homogeneous, shading off into the surrounding yolk-filled protoplasm. It stains very dark in iron haematoxylin. The head of the spermatozoon appears exactly as before.

Twenty minutes after insemination the entrance cone has flattened out, but the spermatozoon is stiH external to the membrane. The substance of the entrance cone is, however, as readily recognized as when it projected above the surface of the egg.


About thirty-seven minutes after insemination (metaphase of first maturation division) the sperm is still readily found on the exterior of the vitelline membrane external to the substance of the entrance cone which is now lens shaped. The substance of the entrance cone is homogeneous and it stains less than before ; it is sharply marked off from the unaltered egg cytoplasm by a layer of small basophile granules. In the center of its external face is a sharply differentiated granule which stains intensely black in iron haematoxylin, and which is connected to the sperm head by a fine thread passing through the vitelline membrane; penetration has already begun.

Forty-three minutes after insemination (late metaphase of the first maturation division) the entrance cone sinks into the egg-cytoplasm, and the head of the spermatozoon begins to be drawn within the egg in the form of a thick thread, less than onethird the diameter of the sperm head, however. The sperm nucleus is being drawn through the small perforation in the vitelline membrane.

Forty-eight minutes after insemination (stages of anaphase of the first maturation division), nearly all of the sperm head is drawn into the egg in the form of a thick thread several times longer than the original sperm head. Before the head is entirely within the egg its inner end begins to swell and becomes vesicular. The entire entrance cone penetrates the egg-protoplasm always retaining its original connection with the apex of the spermatozoon, so that the original orientation of the sperm is preserved and may be readily recognized after penetration.

Fifty-four minutes after insemination (telophase of the first maturation division), the entire head of the spermatozoon is within the egg. The tail and middle piece usually remain without.

As I intend to publish a separate account of the interesting details of penetration of the spermatozoon, and as the later stages do not concern the present problem, I shall simply say, therefore, that as the united sperm-head and entrance cone penetrate farther into the egg cytoplasm, they rotate in such a way that the entrance cone which was originally in advance of the sperm nucleus comes to lie behind it. During the rotation the sperm


aster arises from the pole of the sperm nucleus opposite the entrance cone, thus in the position of the original middle piece.

Morgan has recently ('10), with entire justice as it appears to me, taken a stand against the current view that penetration of the sperm is due to mechanical boring into the egg. He believes that the presence of the sperm calls forth a reaction on the part of the egg that leads to the absorption of the former. There can be no question that the latter conception fits the case of Nereis much better than the former. In the first place the spermatozoon is absolutely motionless after its attachment to the membrane ; in the next place the formation of the entrance cone shows a verj^ decided reaction on the part of the egg to the presence of the spermatozoon; in the third place the retraction of the spermatozoon into a depression of the membrane is due to the retraction of the entrance cone; and finally, as I shall show in a subsequent cytological study, the inclusion of the spermatozoon within the egg appears to be due to activity of the substance of the entrance cone, and not to active penetration by the spermatozoon. The spermatozoon does not penetrate the egg, it is drawn in or engulfed.

6. The later history of the sperm nucleus. The sperm amphiaster is visible in the preparations all through the period of the second maturation division (fig 4). After the formation of the second polar body the sperm-nucleus begins to enlarge and the amphiaster gradually wanes, but it may be recognized up to the time of contact of the germ nuclei. The centrosomes of the first cleavage spindle then begin to appear. Whether or not they are continuous with those of the sperm amphiaster is a question which I shall take up in the next study of this series. The cleavage asters rapidly become very large and conspicuous (figs. 5 and 6). During the growth of the germ nuclei a considerable number of large granules staining strongly in iron haematoxylin appear in their immediate vicinity.

The main point of these observations on the normal fertilization, both in the living eggs and also in section, is to demonstrate for elucidation of the experiments following: (1) That membrane formation precedes penetration of the spermatozoon by a long time. (2) That the spermatozoon does not penetrate


the vitelline membrane and enter the egg until at least forty to fifty minutes after insemination, although its attachment to the membrane takes place immediately. (3) That the presence of the sperm-nucleus is readily demonstrable in all stages after penetration.

B. Removal of the spermatozoon after membrane formation

In the summer of 1909 I was studying the effects of centrifuging on the egg of Nereis with the aim of getting more data on the problem of polarity and the theory of formative stuffs. It soon became apparent that the effects of centrifuging varied with the stage of development, and so several series of experiments were made in which the eggs were centrifuged at regular intervals from before fertilization up to the time of the first cleavage.

The effects of centrifuging may be divided into three categories: (1) A certain proportion of centrifuged eggs develop approximately normally, the percentage varying greatly with the stage of centrifuging. (2) A certain proportion of eggs, varying at different stages, segment more or less abnormally, sometimes extremely so {e.g. meroblastic), and produce embryos with more or less pronounced abnormalities. (8) At certain stages of centrifuging a variable proportion of eggs fails to carry out even the first cleavage. The investigation of the causes of such failure to segment revealed the fact that it was owing to the removal of the spermatozoon after membrane-formation. It is the evidence for this statement that is now under consideration.

The results with reference to failure to segment were, in general, as follows :

1. If unfertilized eggs were centrifuged and then fertilized, all segmented, and a large percentage tended to develop quite normally.

2. A disturbing factor comes in shortly after insemination, owing to the fact that when the jelly is first secreted by the eggs it is so viscid that the eggs stick together in the bottom of the centrifuge in a mass which cannot be separated into its constitu

ent eggs. The extreme viscidity of the jelly gradually disappears, and after about twenty minutes from insemination, the eggs no longer mat together. It is therefore difficult to investigate the effects of centrifuging on the developmental capacity of the eggs during the first ten or fifteen minutes after insemination. However, when the viscid stage begins to pass away and eggs can be separated from the mass for examination, the majority are found to undergo segmentation, as many as 98 percent in one case (experimeat 2, 1910) twenty-one minutes after insemination.

Fig. C. The effects of centrifuging on the power to segment in Nereis. The abscissae represent minutes from the time of i semination, the ordinates the percentage of eggs dividing. At position a penetration of the spermatozoon is just completed in most of the eggs. At position b the first polar body is extruded. Data from experiment 2, 1910.

3. For about the next thirty minutes (twenty-one to fiftythree minutes after insemination) centrifuging tends to inhibit the cleavage of a certain proportion of the eggs which gradually increases up to about thirty-seven minutes after insemination and then decreases again. For instance, in experiment 2 of 1910, of the eggs centrifuged twenty- one minutes after insemination 98 per cent segmented; twenty-six minutes after insemination 36 per cent segmented; thirty-two minutes, 33 per cent; thirty-seven minutes, 21 per cent segmented; forty-three minutes, 26 per cent segmented; forty-eight minutes, 52 per cent segmented; fifty-three minutes, 75 per cent segmented; fifty



eight minutes, 90 per cent segmented; sixty-three minutes, 90+ per cent segmented; sixty-nine minutes, 95+ per cent segmented; control eggs, 99 per cent segmented. (See Fig. C.)

4. From this time on nearly all of the eggs segment after centrifuging until, during the anaphase and telophase of the first cleavage spindle, centrifuging again tends to inhibit cleavage.

The following table (Experiment 29) gives a fairly typical series of results. There were twelve such experimental series in all, more or less complete, giving concordant results except that in some at the period corresponding to 29D, 90 to 98 per cent were so affected that they failed to segment. On either side of this critical period there is a decreasing susceptibility to such injury by centrifuging.

Experiment 29 September 8, 1909



29 A.

29B. 29C.

Control (not centri fuged) Before insemination

20 minutes 30 minutes

29D i 41 minutes

29E. 29F. 29G. 29H. 291. 29 J. 29K.

51 minutes 66 minutes 79 minutes 95 minutes 114 minutes 121 minutes 127 minutes

29L 149 minutes


100 per cent

100 per cent practically Majority Majority 20 to 30 per cent 70 to 80 per cent 90 to 95 per cent 100 per cent 100 per cent 100 per cent Some unsegraented Less than majority and these irregular Most segmented further

Eggs matted loosely

Centrifuged during process of Ist cleavage

Centrifuged in 2-cell stage

Two major processes are going on in the egg at this time, viz. : maturation and fertilization. The injury is not primarily to the process of maturation, for the eggs that do not segment form the polar bodies; nor is it probable that there is a general



systemic injury to the egg protoplasm at this time not received at other times, when the fact that maturation continues and polarity is preserved in these eggs, is considered. It would, therefore, appear probable that the injury is to the process of fertilization itself, and this conjecture is completely confirmed by cytological study. The most conclusive experiment in this respect is no. 27, the details of which are as follows:

The eggs were fertilized at 9:28 a.m., September 4, 1909. Some of these were kept for control and all segmented normally. The remainder were centrifuged about 60 x 120 revolutions at a radius of 6 cm. in one minute, at the following times : 27A at 9 :58 A.M. ; 27B at 10 :03 ; 27C at 10 :08 ; 27D at 10:12; 27E at 10:16. About 20 per cent of 27A segmented, 5 to 10 per cent of 27B, 20 per cent of 27C, 50 to 60 per cent of 27D, and 75 to 90 per cent of 27E. Samples of the control and of each of 27A, 27B, 27C, 27D and 27E were preserved at 10:31 and 10:43 a.m.


Eggs fertilized at 9:28 a.m., Sept. 4, 1909









27 Con

Xot centrifuged

(1) 10:311 A.M.

All segmented


(2) 10:45 §A.M.

27A . . .

60 X 120 rev. in

(1) 10:30 A.M.

About 20 per cent

1 min. 9:58 a.m.

30 min.

(2) 10:431 A.M.



60 X 120 rev. in

(1) 10:30J A.M.

About 5-10 per

1 min. 10:03 a.m.

35 min.

(2) 10:431 A.M.

cent segmented


60 X 120 rev. in

(1) 10:30iA.M.

About 20 per cent

1 min. 10:08 A.M.

40 min.

(2) 10:44 A.M.



60 X 120 rev. in

(1) 10:31 A.M.

About 50-60 per

1 min. 10:12 a.m.

44 min.

(2) 10:441 A.M.

cent segmented


60 X 120 rev. in

(1) 10:31 A.M.

About 75-90 per

1 min. 10:16 a.m.

48 min.

(2) 10:45 A.M.

cent segmented

^Since the above was written this experiment has been repeated (Exp. 2, '10), with the added precaution of preserving a sample of the normal eggs corresponding to each stage of centrifuging, in order to make certain of the stages of fertilization in each case. The results completely confirm those of experiment 27, and the sperm was found to be external in the most critical stage (37 minutes after insemination in this experiment; see page 371).


Cytological studj' of the twelve lots of preserved eggs showed stages ranging from the metaphase of the second maturation spindle to the prophase of the first cleavage, the earlier stages of course being found in lot 1 in each case.

In the control lots it was easy to demonstrate the sperm nucleus at all stages to the formation of the first cleavage spindle. The sperm nucleus is rendered particularly conspicuous during the second maturation division by the large amphiaster that accompanies it (fig. 4), both lying in the yolk-free protoplasm. After the formation of the second polar body the sperm amphiaster gradually fades, but the sperm nucleus can be recognized by its position and by the remnants of radiations up to the time of union of the two germ nuclei; and in the later stages its presence may be inferred by the degree of development of the cleavage amphiaster and the number of chromosomes. There is, therefore, no time from the beginning of the second maturation division up to the formation of the first cleavage spindle when the presence of the sperm nucleus cannot be readily demonstrated.

In the study of the serial sections of the control eggs I found no egg in which, all sections being present, the sperm nucleus could not be demonstrated. In the serial sections of 27A, the sperm nucleus could be recognized in only about 37 per cent of the eggs; in 27B in only 10 per cent to 20 per cent; in 27C in about 25 per cent; in 27D in about 53 per cent; in 27E in about 76 per cent. The stages of maturation of lots A to E corresponded very closely with the stages of maturation of the control eggs killed at the same time.

It is a relatively simple matter to demonstrate the presence of the sperm nucleus, for a single positive observation suffices; but, to be sure of the absence of a sperm nucleus from any particular egg, it is necessary to examine practically every section of the egg, and the absence of two consecutive sections is sufficient reason for excluding an egg from the count. This may be one reason why the number of eggs in the different lots shown to contain sperm nuclei tends to be somewhat larger than the estimate of the number of eggs that segmented. Another reason



probably is that a sperm nucleus may persist to a certain extent even if injured and unable to produce the full fertilizing effect and cause cleavage.

A third reason for discrepancy in the results is that abnormalities of maturation may be produced by centrifuging which render the determination of the sperm nucleus more difficult than usual. It frequently happens that, as the first maturation spindle is driven from the surface by the centrifugal force, it divides before it reaches the surface again, producing two maturation nuclei within the egg. The two second maturation spindles may then unite to form a tetraster, one pole of which approaches the surface and a single polar body is formed, leaving three nuclei within the eggs (fig. 7). Such eggs were readily recognized by the absence of the first polar globule, and by the presence of the extra nuclei. But it was sometimes difficult to determine in certain stages whether there were only three nuclei, the sperm nucleus being absent, or four, the sperm nucleus being present.

A fourth reason for a certain discrepancy between the estimate of the number of eggs that segmented and the number determined to have sperm nuclei might be that at the time the experiment was made the importance of exact determination of the number of eggs that segmented was not realized, and the determination was made only roughly. Putting the results side by side we have :



27 Control


About 20 per cent 5 to 10 per cent About 20 per cent About 50-60 per cent About 75-90 per cent



About 37 per cent About 10-20 per cent About 25 per cent About 53 per cent About 76 per cent





Considering the various sources of error, the agreement is very close except in 27A. But in this case we do not have to


explain why eggs in which the sperm nucleus was absent segmented but on the contrary, why certain eggs that possessed the sperm nucleus failed to segment, which is a very different thing. There is no evidence that any egg in which the sperm nucleus was absent succeeded in dividing.

The general conclusion that removal of the spermatozoon at the times noted in the experiments involves incomplete fertilization, is sufficiently demonstrated by the results.

Let us call the stage at which the spermatozoon is eliminated in the greatest proportion of eggs, the critical period. The exact number of minutes from the time of mixing eggs and sperm to this stage varies of course through the season, owing to the variations of temperature. Moreover, it is not exactly determined in all experiments, for in some the stages of centrifuging fall on either side of it. This being understood, we may note that in eight experiments the critical period occurred at from 25 minutes to 40 minutes after fertilization. This is quite a wide variation, but when the time is represented as a fraction of the entire period between fertilization and the first cleavage, it is found that in all cases the period up to the critical period is between 27 and 33 per cent of the total time up to the first cleavage. It is obviously a corresponding stage in all cases, for the observed differences fall within the chances of error, viz : that the critical period is hit exactly in only very few experiments, and that the time of beginning of the first cleavage must be stated rather arbitrarily on account of the variation in rate of individual eggs.

The critical period occurs shortly before the penetration of the spermatozoon into the egg. We noted in the first part of this paper that the penetration of the spermatozoon is extremely gradual; my observations on this point, both from the study of the living egg and also of sections, show that it requires forty to fifty minutes for the head of the spermatozoon to disappear through the membrane.

As the most critical period comes in the great majority of experiments from thirty-five to forty minutes after insemination, it is obvious that the spermatozoon is in some way prevented



from entering the egg. The explanation is comparatively simple; the spermatozoon is imbedded in the jelly by which the egg is surrounded. When the jelly is first formed it is very viscid, and adheres to the eggs during centrif uging so that they mat together in the centrifuge. However, this stage passes and the result of centrif uging is then to separate the jelly from the eggs. In many cases the jelly carries off the attached spermatozoon with it. After penetration this can of course no longer happen. The curve of variation of the per cent of eggs centrifuged before penetration that fail to segment is due to the following factors: (a) At about twenty-five minutes the sperm head is sunk in a deep depression of the membrane and hence is less likely to be torn away by the jelly; (fe) the change in consistency of the jelly presumably extends from without inwards; hence at first the innermost layer in which the spermatozoon is imbedded tends for a time (also presumably) to remain with the egg; (c) the time of penetration varies somewhat in any lot of eggs. These facts together would explain why the percentage of eggs that fail to segment after centrifuging rises to a maximum and sinks again to a minimum in the successive stages of centrifuging.

The fact that centrifuging inhibits cleavage in a small per cent of the eggs from fifty to fifty-five minutes after insemination, leads me to suspect that in some cases the sperm may be destroyed after its penetration into the egg. In experiment 2, 1910, for instance, cleavage was inhibited in 25 per cent of the eggs centrifuged fifty-three minutes after f ertihzation ; in the control eggs killed at the time of centrifuging, penetration of the sperm is completed in the great majority of eggs, though it is found external still in a very few ; it is difficult to estimate the per cent of the latter, but the impression is that it is less than 25 per cent. However, it is impossible to confirm this, and I mention the matter here to call attention to the error in my first paper read before the Central Branch of the American Zoological Society (Abstract in Science, vol. 18, p. 36, May, 1910), in which I stated that the destruction of the sperm nucleus followed penetration. A renewed study of the penetration has proved that this is not the case, usually at least.


We are not, of course, free to infer that fertilization is complete as the stimulus to development immediately after penetration of the spermatozoon. The experiments prove directly that in the egg of Nereis the stimulus of the spermatozoon as the impulse to development consists of two distinct parts: (1) an external stimulus that causes membrane formation and which is sufficient of itself to induce the maturation ; (2) an internal stimulus dependent on penetration of the spermatozoon. How long after penetration fertilization is still incomplete cannot be decided on the basis of these experiments.

In concluding this section, we may note that the cause of failure to segment following centrifuging during the anaphase of the first cleavage is an entirely different one. The cause in this case is the breaking up of the karyokinetic figure and function, and dispersing the chromosomes. This is readily demonstrated in sections. In the case of eggs centrifuged at the ' critical stage', the sections show that the maturation spindle receives no injury from centrifuging, but appears coherent and normal in the sections. The sections of eggs centrifuged during the anaphases of the first cleavage show the chromosomes dispersed through the cytoplasm and the cleavage spindle no longer coherent, but broken up. In the first case the cause of failure to segment is elimination of the sperm-nucleus, as shown by study of series 27. This cannot be the cause in the second case, and a sufficient explanation of the failure to segment is found in the destruction of the karyokinetic figure.

C. Comparison of co??ipletely and partially fertilized eggs in later

We have noted so far that definite proportions of eggs centrifuged at definite periods in the process of fertilization fail to develop a sperm-nucleus, and that similar proportions of the same lots of eggs when left to develop fail to undergo segmentation. The facts (1) that all the control eggs of the same lot segment, and (2) that the centrifuged eggs that fail to segment, nevertheless had formed the fertilization membrane and under


gone maturation, prove that the unsegmented eggs had received at least the first stimulus of fertilization. It was also shown that the critical period for suppressing segmentation by centrifuging occurs at a time shortly before entrance of the spermatozoon, and that it is due to prevention of penetration. The partially fertilized eggs, therefore, resemble the normal ones in the fact that membrane formation and the first stimulus to development are called forth by action of the spermatozoon, and they differ from the normally fertilized eggs in that the internal egg protoplasm has not received the direct stimulus of the spermatozoon. A cy tological examination of such eggs could not fail to be of interest and might give some clue to the internal function of the spermatozoon in fertilization.

Both polar bodies form regularly in such eggs as already noted, and the egg-nucleus (female pronucleus) arises and attains the same size as in normally fertihzed eggs. The chromosomes of the first cleavage spindle then form in the usual fashion and at the usual time, accompanied by disappearance of the nuclear membrane. But, whereas, in the presence of a sperm nucleus, cytoplasmic asters accompany these processes and a spindle rapidly arises during the prophases of the first cleavage, in the absence of the sperm nucleus there is absolutely no sign of c\^tasters or evidence of spindle formation. The chromosomes lie naked in the cytoplasm surrounded by a clear area (fig. 7).

Each chromosome then splits longitudinally in the usual fashion, but the halves do not separate. At the time of the telophase of the normal first cleavage there is a tendency to scattering and breaking up of the chromosomes. When the normal eggs have reached the two and four-celled stages, the scattering and breaking up of the chromosomes have progressed much farther in the unsegmented eggs, and in the course of two or three hours there remains no differentiated nucleus or chromosomes, but only numerous chromatic granules scattered throughout the cytoplasm.

The behavior of the partially fertilized eggs may be compared on the one hand with that of normally fertilized eggs and on the other with that of eggs caused to mature by centrifuging. As com


pared with the former, the chief difference observable by cytological methods is the entire absence of the achromatic part of the karyokinetic figure. The differences in later stages may be conceived as secondary effects of this defect or of the conditions determining such defect. When eggs are caused to mature by centrifuging the process begins as in normally fertilized eggs by the breaking down of the cortical layer and formation of the jelly; the germinal vesicle ruptures and the two maturation divisions follow. After the completion of the maturation the chromosomal vesicles of the egg nucleus usually fail to unite perfectly, and in a httle while they separate and scatter in the cytoplasm without formation of chromosomes, so that each egg appears to possess a considerable number of small nuclei. In a few cases the first indications of chromosome formation may be observed in the vesicles shortly after maturation but not later. Subsequently the chromosomal vesicles appear to dissolve in the cytoplasm liberating small chromatic nucleoli.

The partial stimulus of the spermatozoon is thus somewhat more effective than the mechanical shock of centrifuging, though both produced the same initial changes, apparently equally well. This may possibly be due to entrance of a small amount of matter from the spermatozoon; for at the critical period the perforatorium of the sperm has penetrated the membrane and is imbedded in the entrance cone.

D. General discussion

The general conclusion that the function of the spermatozoon in the stimulus to development involves at least two factors has already been clearly stated by Boveri ('07) and Loeb ('096) : According to Loeb, one factor is the cytolysis of the very thin cortical layer of egg"; but while this stimulates development, the latter is often abnormal and therefore usually comes to a halt. A second process is necessary to ensure more normal and lasting development (Loeb '096). Apparently Loeb is not very clear concerning the nature of the second factor, but is inclined to regard it as inhibiting the cytolysis which he conceives to be


begun as the first factor of the developmental stimulus. This conclusion was formed as a result of two kinds of experiments: In the first Loeb found that the best results in artificial parthenogenesis are obtained, in the egg of a Californian sea-urchin, by a double treatment : first using a cytolytic agent and then following it by treatment with hypertonic sea-water, or by inhibiting oxidation for a while. In the second class of experiments Loeb and Elder found that mere membrane formation might be induced in sea-urchin eggs by external contact of starfish spermatozoa, but farther development did not take place except in the relatively few cases in which the spermatozoon entered the egg (Loeb '09b, p. 249), or unless the eggs were treated after membrane formation with hypertonic sea-water. Although this experiment is complicated by the hybridizing, yet it demonstrates the same distinction between external and internal functions of the spermatozoon in fertilization that I have shown for Nereis by a different method.

Artificial parthenogenesis may be induced in the sea-urchin egg without membrane formation and this fact appears to me to indicate that the internal function of the spermatozoon is probably at least as fundamental as the external function (membrane formation), though, as Loeb points out, development without membrane formation takes place in a less normal fashion than after membrane formation. But inasmuch as we may have membrane-formation without development, and development without membrane formation, it would seem premature at least to consider membrane formation as the chief function of the spermatozoon in fertilization.

The experiments described in this paper show that m Nereis fertilization by the spermatozoon is incomplete after the formation of the membrane. The question then arises, when is the function of the spermatozoon in fertilization completed? Ziegler's and Wilson's experiments show that it is incomplete even some time after entrance of the spermatozoon. Ziegler's experiments ('98) consisted "in so constricting the egg of the sea-urchin after penetration of the spermatozoon that the one part contains the sperm nucleus, the other part the female sex-nucleus. The


part that contains the sperm nucleus undergoes cleavage and develops farther; in the other part the female sex-nucleus undergoes remarkable transformations, dissolving and reappearing, a process which is repeated several times." In spite, therefore, of the presence of the sperm-nucleus in a constricted part of the same egg, the part containing the egg nucleus was not fully fertilized. It made abortive attempts at division, but the karyokinetic figure was too feeble to carry the process through.

Wilson observed ('03) that if the fertihzed eggs of Cerebratulus be cut in two shortly after the penetration of the spermatozoon only a single fragment develops even though the fragments be refertilized immediately after the operation. In such cases it is almost invariably the nucleated fragments that develop, but in a very few cases I have observed that the enucleated fragment develops, ' while the nucleated one forms the polar bodies, but proceeds no further." By the nucleated fragment in this case Wilson means the fragment containing the maturation spindle. Farther on he adds The few cases in which the enucleated fragment of the bisected fertihzed egg develops are doubtless those in which the plane of section separates the sperm-nucleus from the egg-nucleus." This is indeed the only possible explanation. In such cases the fragment containing the egg-nucleus is only partially fertilized.

Boveri has also observed that if freshly fertilized sea-urchin eggs be broken into fragments by shaking, certain of the fragments contain the egg nucleus alone. If such fragments are not subsequently entered by a spermatozoon, the nucleus enlarges, dissolves and reappears again; but they do not segment ('96). Later he observed that such pieces may divide at least twice ('02).

These experiments demonstrate that fertilization is still partial even some time after entrance of the spermatozoon into the egg; but they do not show at what stage it is complete. Boveri's very interesting observations on 'partial fertilization' in the sea-urchin egg ('88 and '90) carry the solution of the problem a step farther (cf. also Teichmann '02). In the experiments which furnished the material for his observations both eggs and sperm were weakened, the former by standing for fourteen hours


in sea-water and the latter by treatment with KOH prolonged to a stage in which only a small percentage of the spermatozoa continued to move. Under these conditions in a large number of eggs the sperm aster separated from the sperm nucleus, which was usually left on one side, and proceeded alone to conjugate with the egg-nucleus. Thereupon the cleavage spindle formed with the egg-nucleus alone, and segmentation of the egg ensued. In the four-cell stage usually, but sometimes in the two or eightcell stage, the sperm nucleus united with one of the segmentation nuclei. Boveri concludes from this and other results that the fertihzing action of the spermatozoon consists in the introduction of a centrosome into the egg.^ When this has united with the egg nucleus, with or without participation of the sperm nucleus, fertilization would be complete; or with the sperm nucleus alone in merogony it likewise completes fertilization.

Boveri has used the term 'partial fertilization' for the phenomenon just described, although he admits that it is a misnomer. It is unfortunate that such a significance should have come to be attached to the expression, because, as has been shown, my own results and those of Ziegler, Wilson and Boveri himself prove that partial fertilization in the literal sense really occurs. The various stages of partial fertilization as shown by the results in the literature on the subject are:

1. External contact alone by the spermatozoon producing, a. Formation of the fertilization membrane, (Loeb and Elder

for sea-urchins, Lillie for Nereis).

h. Maturation and formation of the chromosomes from the egg nucleus without spindle, (Lillie for Nereis).

c. Maturation and cleavage to stereoblastula, (Bataillon: hybrid union of eggs of Pelodytes punctatus and Bufo calamita with sperm of Triton alpestris).

2. If the spermatozoon be removed shortlj after entrance, a. Maturation alone may result, (Wilson on Cerebratulus) .

^Herbst ('07 p. 202 and '09 p. 277) interprets Boveri's 'partial fertilization' as a combination of parthenogenesis and fertilization. Such an interpretation does not, however, explain Boveri's account of the behavior of the sperm-centrosomes.


b. An abortive karyokinetic figure may form with the egg nucleus alone, (Ziegler and Boveri on sea-urchins).

c. In some cases at least two cleavages may result, (Boveri on sea-urchins). Boveri's so-called 'partial fertilization, is really full fertilization, so it does not come in this series.

The difference in reaction of the egg-cytoplasm to its own nucleus and to the introduced part of the spermatozoon can be explained on only one of two grounds; either in the general sense of Boveri's theory on purely morphological grounds, or on the ground of a chemical difference, presumably of sexual origin, between the egg on the one hand and the sperm on the other. The latter form of interpretation seems to me preferable because it is a physiological interpretation which takes cognizance of the sexual factor in fertilization,

Boveri's theory of fertilization rests on the identification of the centrosomes of the sperm aster with a definite formed element (centrosome) introduced into the egg by the spermatozoon; but it has never been demonstrated in any case in all the literature on the subject of fertilization that the centrosomes of the cleavage spindle, or indeed of the sperm aster, are derived from any definite formed element of the spermatozoon. Boveri himself admits this in his sixth cell study ('07, page 266) ; and so long as definite proof of the continuity of the so-called spermcentrosomes from the spermatid up to the formation of the first cleavage spindle is lacking, all of Boveri's observations are open to another interpretation than the one he has given; to the interpretation, namely, that the sperm asters represent a reaction of the egg cytoplasm to a male element, or at least a foreign element, represented for the most part by the sperm nucleus.

The biological analysis of fertilization seems to me to rest now upon the problem of the origin of the sperm aster in the egg. More crucial evidence is needed on this point, and I do not believe that any refinement of cytological technique will give the result. Experimental evidence is needed; either, as Boveri ('88) suggested, the introduction of a non-nucleated spermatozoon in the egg to prove whether or not the sperm asters would arise without the nucleus and fertilize the egg, or the introduction of only the


anterior part of the sperm into the egg to prove whether or not the sperm nucleus without the centrosome, which is contained in the middle piece of the spermatozoon, would cause the production of asters in the egg-cytoplasm.

As Boveri, among others, has pointed out, there is not onlyone, but several stages of inhibition in the history of the egg. This may be illustrated by noting the stages at which in the eggs of various animals the need for fertilization arises. In some eggs it is before the rupture of the germinal vesicle {e.g., Nereis), in others at the time of the mesophase of the first maturation spindle {e.g., Chaetopterus and Cerebratulus), in others again after the formation of the first polar body {e.g., Amphioxus, amphibians), in others again, not until after the formation of both polar bodies {e.g., sea-urchin). There is no doubt that the last stage of inhibition is the most difficult one to overcome, both because many eggs pass by the earlier stages without apparent specific stimuli and also because it is possible to cause eggs that normally stop at the first or second stage of inhibition to pass on to the last stage by stimuli that are ineffective when this stage is reached, {e.g., Nereis as noted in this paper and Chaetopterus as noted in various earlier papers).

The nature of the inhibition that causes the need for fertilization is a most fundamental problem. Is it the same in all these cases, i.e., a gradually increasing inhibition that may be effective before maturation, but in some cases not until maturation has progressed a certain distance, or even until it is complete?'* If it is the same cause at all these stages then it is certain that the need for fertilization is not due to any defect of the egg-centrosomes, for the pause takes place in Chaetopterus (e.g.) while the egg-centrosomes are at the very height of their activity. If,

^Bataillon ('10) holds the view that the nature of the inhibition is the same whether the arrest is at the stage of the resting nucleus or in the height of karyokinesis. He holds the view that the inhibition is due to accumulation of excretory products and that the stimulus to development is essentially a process of elimination. Bataillon's paper was received after my own was completely written. His interesting results will be considered more fully in my next paper. In Part I of the present paper (last paragraph) I have presented a view similar in some respects to Bataillon's.


therefore, we are to hold to the theory of Boveri in its Hteral sense, we must beheve that there are different kinds of inhibition. However, it is, I beheve, simpler and more logical to hold that the inhibition differs only in intensity at these various stages; and this point of view seems to be supported by the fact that the same stimulus which at a lower intensity will cause only maturation to take place in Chaetopterus, at a higher intensity will cause differentiation also to proceed, though in this case without cleavage. Boveri ('07), however, holds that there are different kinds of inhibition, that the postulated degeneracy of the eggcentrosomes after maturation is in a sense the more primitive, and that other kinds have been secondarily acquired, a point of view that gives a more or less definitely teleological aspect to the question.

From a physiological point of view we might inquire, what are the conditions that cause the postulated sudden degeneration of the egg-centrosomes? Such a condition if found, would be nearer the fundamental cause of inhibition of the egg and it might turn out to be the same cause that conditions in so many cases an earlier arrest of activities in the egg.

The experiments on artificial parthenogenesis are sometimes regarded as involving the entire problem of fertilization. But if it be true, as many believe, that biological fertilization, (if I may be pardoned such an expression) is fundamentally a sexual reaction, then the physico-chemical analysis of fertilization must compass the entire problem of sex, which is much wider than the problem of parthenogenesis. The physico-chemical analysis of fertilization has dealt, up to the present exclusively, with the latter problem, and for this reason the earlier title of such studies 'artificial parthenogenesis', seems to me much more fitting than 'chemical fertilization' which is sometimes loosely used. From the zoological point of view, at least, parthenogenesis and fertilization are not interchangeable functions. There is a factor present in fertilization which is absent in parthenogenesis, and the latter is never the exclusive mode of reproduction among animals. The biological analysis of fertilization therefore involves problems that do not occur in the physico-chemical analysis of parthenogene>-is.



Bataillon, E. 1910 Le probleme de la fecondation circonscrit par I'impregnation sans amphixie et la parthenogenese traumatique. Arch, de Zool. exp. et gen. 5 ser. Tome 6.

BovERi, Th. 1888 Ueber partielle Befruchtung. Sitz'b. d. Ges. fiir Morph. u. Phys. in Munchen. Bd. 4, H. 2.

1890 Zellenstudien. Heft. 3. Ueber das Verhalten der chromatischen Kernsubstanz bei der Bildung der Richtungskorper und bei der Befruchtung. Jena. pp. 32 ff.

1896 Zur Physiologic der Kern und Zelltheilung. Sitz'ber. d. Phjs.Med. Ges. zu Wurzburg, (cited from Teichmann — unfortunately the original paper was inaccessible to me).

1902 Das Problem der Befruchtung. Jena, G. Fischer.

1907 Zellen-Studien. Heft 6. Die Entwicklung dispermer Seeigeleier. Ein Beitrag zur Befruchtungslehre und zur Theorie des Kerns. Jena. Gustav Fischer.

Fischer, Martin H. 1903 Artificial parthenogenesis in Nereis. Am. Jour. Physiol, vol. 9, pp. 100-109.

Herbst, Curt 1907 Vererbungsstudien V. Auf der Suche nach der Ursache der grosseren oder geringeren Ahnlichkeit der N^chkommen mit einem der beiden Eltern. Arch. f. Entw'mech. Bd. 24, p. 185.

1909 Vererbungsstudien VI. Die cytologische Grundlagen der Verschiebung der Vererbungsrichtung nach der miitterlichen Seite. I Mittheilung. Arch f. Entw'mech. Bd. 27, p. 266.

LoEB, Jacques 1909a Ueber das Wesen der formativen Reizung. Berlin, Julius Springer.

1909b Die chemische Entwicklungserregung des tierischen Eies. Berlin, Julius Springer.

Morgan, T. H. 1910 Cross and self-fertilization in Ciona intestinalis. Archiv f. Entw'mech. der Organismen. Bd. 30, ii, Theil.

Teichmann, Ernst 1902 Ueber Furchung befruchteter Seeigeleier ohneBeteiligung des Spermakerns — Jen. Zeitschr. f. Naturw. N. F. Bd. 30, p!105.

Wilson, E. B. 1892 The cell-lineage of Nereis — A contribution to the cytogeny of the annelid body. Jour. Morph. vol. 6.

1903 Experiments on cleavage and localization in the Nemertine egg. Arch. f. Entw'mech. vol. 16, pp. 417-418.

ZiEGLER, H. E. 1898 Experimentelle Studien tiber die Zelltheilung, II. Arch. f. Entw. mech. vi.



1 Axial section of an unfertilizod normal ovocyte of Nereis, fixed in Flemming's fluid, weaker solution. In this fixing fluid the yolk granules swpll and tend to run together. The oil drops are dissolved out in the preparation and are represented by emptj^ spaces, v. in. vitelline membrane, c.l. cortical layer, from which the jelly is formed.

2 Section of an egg of Nereis, fixed in Aleves' fluid five minutes after insemination. The Cortical layer is already somewhat reduced in thickness. The yolk granules are not swollen. The oil drops are not dissolved out. The section is approximately horizontal, c.l., remains of cortical layer; v.m., vitelline membrane.

3 Section of an egg of Nereis, fixed in Meves' fluid fifteen minutes after insemination. The cortical layer has entirely disappeared, and the perivitelline space is formed. The germinal vesicle is breaking down and the first maturation spindle is forming, p.v. perivitelline space, v.m., vitelline membrane.

4 Section of an egg of Nereis, fixed in Meves' fluid fifty-seven minutes after insemination. Only one centrosome of the sperm amphiaster is shown.

5 Section through the first cleavage-spindle of Nereis, normal, one hour and twenty-seven minutes after insemination.

6 Tetrapolar second maturation spindle of Nereis. See text for description (p. 378). Three egg nuclei are formed in such a case.

7 Egg nucleus of egg of Nereis in which the spermatozoon was removed by centrifuging. The chromosomes of the first cleavage are formed, but there are no asters. Cf. fig. 5.

All figures drawn with the camera lucida with Zeiss comp. oc. 6 and 2 mm. horn, oil im. obj.





From the Laboratory of the Marine Biological Association of San Diego



Purpose of the research 396

1. Special 396

2. General 396

Brief description of the species 398

Measurements of the zooids of the wheels and of a portion of the chain not yet

transformed into wheels 399

Treatment of the quantitative data 406

Attempt to connect the formation of wheels with morphological, physiological,

and mechanical phenomena presented by the animals 414

1. Segmentation of the stolon, and the deploying point 414

2. Position and relation of the zooids in the chain from the deploying point '

to the twist 414

a Shifting of the zooids 414

b Peduncles and foot-pieces 417

c Emergence of the chain to the outside world 417

d Twist in the chain 418

e Reduction of the foot-pieces, first break in the chain, and formation

of the first wheel 418

3. Comparison of the rate of growth of the chain as a whole with the rate of

other animals 419

Discussion of the observations from the causal standpoint 420

1. Cause of the twist * 420

2. Unequal growth of zooids and foot-pieces as a factor in the breaking up of

the chain 422

3. Impossibility that the character of the blood supply to the zooids can be

the cause of the size scheme within the wheels 427

4. Unlikelihood that the wheel arrangement of the zooids in Cyclosalpa has,

as believed by Brooks, anything to do with the position of the first four blastozooids of Pyrosoma 429




The larger significance of such studies 431

1. Supplementing biological with quantitative observations 431

2. Natural periodicity in organisms and exacter methods in biological re search 432

3. The inadequacy of treating periodicity, generally, as an aspect of fluc tuating variation 440

Bibliography 444


1. Special

One of us (Johnson, '10) has shown that the individuals of the blocks into which the chains of blastozooids of Salpa fusiformisruncinata, S. cylindrica, and S. zonaria-cordiformis become differentiated, fall into size schemes, or systems.

The question naturally arises, how general is this phenomenon among salpae? The possibility that the wheel grouping in Cyciosalpa corresponds to the block grouping in Salpa proper, occurs to one rather readily in spite of the conspicuous differences between the two. If this conjecture be right, we should expect to find a size scheme of zooids in the wheels of Cyclosalpa similar to that in the blocks of Salpa. That such a scheme exists in the wheels even more pronouncedly than in the blocks, the sequel will show.

2. General

This much more evidence is consequently adduced favorable to the idea of correspondence between the wheels and the blocks. But what do we mean by correspondence? In a general sense the blocks and wheels undoubtedly correspond: both are groups of shnilar organisms similarly located with reference to the parent zooid. This much of correspondence is recognizable to cursory inspection. Does the discovery of a similar size scheme among the zooids in the groups in the two species advance our interpretation of these organisms much if at all ? Does it amount to anything more than a recognition of one more resemblance? According to the meaning that interpretation' and 'resemblance' have in most later biological writing, we must probably say no. We


must hold that unless the new correspondence includes somewhere what we hold as one or more 'causal factors' not much has been accomplished. If, for example, we extend the inquiry to the question of the dependence of the size scheme of the zooids upon growth and other internal factors on the one hand, and upon environmental factors on the other; and if here, too, we find further correspondence, our belief in the essential identity as we might say, unless standing for extreme exactness of expression, would be reached.

We shall see that the size scheme of the zooids in the wheels is almost certainly foreshadowed before the wheels themselves are. If this be so, then the resemblance between the Cyclosalpa chain and the Salpa chain is considerably closer before than after the wheels appear in the former. But it is difficult, if not impossible, to attribute the block production in the Salpa chain entirely to other than inherent factors, of which growth seems to be the most immediate. So far, therefore, as we can rely upon our evidence for the marking off of the Cyclosalpa chain into groups before the wheels are formed we seem to have placed the notion of correspondence between the wheels and the blocks on firm ground.

Evidence still more convincing perhaps, that the wheels and blocks correspond in a strict biological sense, in the sense that both are expressions of periodicity in growth, is found in the fact that growth and development are observed to be periodic in so wide a range of living beings. The growth of plants for example, appears to be nearly if not quite always of that nature. Finally, belief in the correspondence would, so far as we can see, reach high water mark, should it be finally made very probable that not only growth and development but all strictly biological processes whatever, are periodic. We are undoubtedly a long way from this last conception. Certain it is, though, that we now have sufficient facts to make the hypothesis of periodicity as warrantable as its opposite, namely, that certain phenomena are continuous in the sense of not being automatically interruptive and group-wise. We may now proceed to the handling of our data.



The species Cyclosalpa affinis (Chamisso) was taken in abundance at La Jolla from May to November, 1909, during which time most of the observational portion of this research was made. The longest chains and the largest wheels were brought in during the earlier part of this period, while the later catches yielded many specimens of the solitary form with short chains, and many medium sized single wheels.

Although the salpae do not survive for more than a day or two in the ordinary aquarium, the material has been sufficiently abundant to admit of considerable work, with the living specimens.

The two generations of this species differ markedly, as do those of all members of the genus. The intestine of the solitary form (fig. 11) is straight, extending nearly the full length of the animal, the anal opening being just back of the ganglion. The intestine of the aggregate generation, on the contrary, projects from the ventral side of the creature as a large, almost circular loop, (fig. 12) the anus being only a little to the left of the oesophageal mouth.

The solitarj^ form has eight body muscle bands, according to our system of enumeration, while the aggregate generation has five on the dorsal and six on the ventral side. The hypophysis (%p.), endostyle (end.), and gill (gi.) present much the same appearance in both forms. The orifices are also similar except that the , solitary form possesses short, tail-like appendages one on each side of the atrial orifice. Our records show the maximum length of specimens of the solitary generation to be 15 cm. and of the aggregate 8 cm. In both generations the test is thin, soft, and highly transparent, without special thickenings. In the aggregate generation, projecting from the ventral side, is the broad, thin peduncle (ped.) by which the zooids are united to form the wheel, and within the pharyngeal cavity on the right side, two thirds of the way back, is the embryo. In the young sohtary individual, the eleoblast, near the heart, and the remnant of the placenta, about one-third of the way back from the oral orifice, are both opaque, nearly spherical bodies and are very prominent.


The stolon originates just above the heart and extends straight forward along the median line. The zooids, as in other salpa chains, are first in single file, but at a certain point, the deploying point, shift to double file. The deploying point in this species occurs close to the anterior end of the heart, 3 or 4 mm. from the root of the stolon. At a point about two-thirds of the way between the heart and the branchial orifice, the chain bends downward and passes to the outside world through an opening in the test. Immediately outside the opening, the chain doubles back under the parent and turns over so that it appears to be greatly twisted at this point. Before the twist, the zooids are arranged in two nearly parallel rows along the common stolonic blood vessel. After the twist, they are arranged in wheels which are connected tangentially and contain six to sixteen zooids each. These wheels show a gradual increase in size toward the distal end (fig. 11).




Serial measurements were made of the zooids of a number of chains with and without wheels as well as of the zooids of separate wheels. Lengths only were taken.

The measurements of the wheels were made with dividers and the results are given in millimeters. Those of the unbroken chains were made with the micrometer eyepiece used in the Zeiss binocular microscope. A unit in the tables represents 0.1 nam. of actual length.

The zooids of Chain I were separated from the chain for measurement, but the others were measured while still on the chain. It appeared that the latter is the more accurate method, since separating the zooids, besides being a tedious process, is apt to distort and mutilate them. In these measurements, the posterior extremity was taken at the atrial orifice. The intestine was not included, as a slight difference in its inclination would make a difference in the apparent length.

In table 1 are given the lengths of the zooids of the unbroken portion of Chains I-VII. Table 2 gives the same data for the


Length measurements of the zooids of wheels of chains VII and VIII Horizontal lines indicate end of wheels. Double lines indicate one wheel lost. Unit — 1 nam.




CHAIN VIII— continued

CHAIN vin— continued














7.7 7.7

7.6 7.8 8.2 8.1

6.7 6.9 7.1

6.8 7.0 6.7






13.1 12.9

12.0 12.2









18.6 19.8 20.7




13.0 13.9 14.0

12.9 13.0 13.4



7.8 8.2

7.4 7.4

6.8 6.8




































21.9 20.5




8.8 8.7




i 51







19.8 21.1 22.9





7.8 7.6

14.9 15.2

12.4 14.0

































! 55



















10.3 10.0 9.3

9.2 9.3 10.6




t 59


14.8 14.0









8.4 9.1

15.4 15.5

21.2 23.7







































10.8 10.1

11.4 10.9



i 63




15.2 14.0

15.4 14.3 13.6





24.0 23.9



9.8 9.6 10.2

8.9 8.9




9.7 11.0

10.6 10.7

16.0 16.3

23.0 25.1






















19 6






11.8 11.5 11.1

10.6 11.2 11.2


9.2 9.1













18.4 18.0 15.7

17.7 17.1 14.7 14.7




26.4 25.5 22.4



9.6 10.9 11.8 12.1 12.3 12.3 11.6



10.2 9.9 10.5 10.8 10.7 11.5



11.7 11.1 12.8 13.0 12.9 12.2 12.5

11.3 11.4 12.4 11.8 11.8 11.7

16.2 18.8 19.4 19.7 19.5





17.3 19.4 19.4 18.3 18.4 16.1 19.1 20.0



12.0 12.7 13.4 13.6

11.5 11.4 11.4 11.6

18.6 19.3 20.5 22.2



wheel portions of Chains VII and VIII. In all, the unbroken portions of seven chains were measured. The number of zooids used depended upon the miminum size measurable. In four cases, 90 zooids were taken but in the other three chains only 80, 70 and 52 zooids respectively were large enough to be measured accurately. [.The measurements of all series are given but only two are plotted, the right-hand series of Chain II in fig. 1, and the right-hand

S 3 13 17 Bl ZS t9 J3 J? 41 'ti 4a S3 37 tl 6} 69 73 77 61 63 89

Fig. 1 Plot of length measurements of the zooids, peduncles, and foot-pieces of chain II, right series. Vertical distances represent length. Horizontal distances represent position in the chain.

series of Chain VII in fig. 2. In the latter figure, 1-90 are the zooids of the unbroken part of the chain while 91-108 are wheel zooids, the divisions between wheels being indicated by vertical dotted fines. Table 3 gives the measurements of zooids of several short chains of wheels which furnish figures for comparing graphs of wheels of various sizes. In each series of results here given except length of Chain I, two measurements were taken and these



Length measurements of zooids of various small groups of toheels. Unit — 1mm.

Group A

Group B

Group C

Group D














10.9 11.7 12.1 12.1 12.6 13.2 14.7 15.3 15.6 16.3 14.8

10.7 10.8 11.1 11.4 13 2


18.4 19.1 19.0 20.5 17.9

18.5 18.6 19.1 19.. 6 20.7 19.9

16.7 18.0 19.1

19.7 20.1 21.6 21.1

20.8 27.9

21.5 29.4 21.2 1 29.2

21.9 i 29.8

22.6 ' 30.1 21.9 29.3











13.5 13.6 14.5 15.7 15.8 15.1



19.6 21.1 21.6 22.0 22.5 22.2

19.8 19.9 20.2 22.1 22.4 23.0 21.8 21.8 23.5 24.2 24.3 23.1

22.0 23.4 24.2 24.2 23.7 24.2 23.0

28.9 30.9 30.8 40 30.4 30.3



22.3 22.6 22.6 22.5 21.0









16.6 17.0 16.9 17.1 18.5 18.1 12.7 18.5 19.4 19.4 18.9 15.1

15.1 17.1 16.7 17.9 18.5 18.9 15.9


13 '. ...

22.3 23.3 22.3 24.6 24.1 23.0

19.9 22.3 24.0 25.5 23 5




24.0 24.1 25.0 25.2 25.7 25. S



25.0 26.5 27.6

28.4 27.8 27.8 23.8



18.8 18.9 18.6 19.8

22.8 24.6 24.9 25.8 25.7 24.9

22.3 24.0 26.1 25.4 25.7 25.7 24.5


24.0 26.6 27.3 28.5 28.4 28.0 25.9






26 ...




Gboup E Group F

Group G

Group H













23.4 24.6 25.2 24.7 25.1 23.6

23.0 22.3 22.9 24.8 23.8 23.3 25.7 24.4 23.9

26.1 , 24.8 1 23.9 26.5 25.0 24.7


21.9 21.6 22.3 22.4

21.0 22.2 22.6 24.2 22.6

25.4 25.1 25.0 24.9 23.9

23 1








24 8

25.3 26.3

25.1 25.0


21.3 22.0 22.5 24.3 25.3 25.7

23.2 25.2 25.9 26.8 26.8 26.1




24.9 26.8 27.6 27.2

27.7 27.2

24.1 26.9



24 2




27.2 1 27.2 25.9 27.5 ' 28.7 25.1 27.7 28.6 ! 26.0 27.2 28.4 26.8




27.4 28.8 28.7 29.7 29.2 27.6

27.5 28.3

28.4 27.2

27.9 27.2 27.2 29.2 29.8 29.9 30.3






30.0 29.8 29.2 30.3


26.6 29.9 29.8 29.3 30.4



were averaged for the final result. Where the first and second measurement differed by more than 10 per cent, a third measurement was taken and the three figures averaged.

Fig. 2 Plot of the length measurements of the zooids of chain VII, right side, including wheels.


At first glance, one sees a resemblance between the curves for the wheels of Cyclosalpa affinis and the blocks of Salpa fusiformisruncinata. In both cases, the end zooids are smaller than those nearest them, the maximum values lying somewhere between, usually nearer the distal end.

Fig. 3 Mean curves for wheels of various sizes. Vertical distances represent length of zooids. Horizontal distances represent position in the wheel.


Some variation in the graphs of wheels of different sizes was noted, and to make sure of its general trend, the data for all the wheels were considered. The wheels were first grouped according to size, Group A included wheels whose zooids averaged 5-10 mm. in length; Group B, 10-15 mm.; and so on. In Group A were ten wheels. Not only does the number of zooids in a wheel vary, but the number in one-half of a wheel is not always the same as in the other half. For this reason the ten wheels were regarded as twenty half wheels.

Among these twenty half wheels of Group A were three containing four zooids; one with five zooids; nine with six zooids; and seven with seven zooids. The corresponding values of the three fourzooid half wheels were averaged, the three first zooids together, the three second zooids, the three third, and the three last zooids. The result was a typical curve for a four- zooid half-wheel whose zooids have an average length of 6-10 mm. The five, six, and seven-zooid half-wheels were averaged in the same way. Similar computations were made for the other four groups and the results plotted. The graphs were smoothed and those for each size were averaged in order to get the typical curve for that size. These curves (fig. 3) show that the size differences between the zooids of a half-wheel greatly increase as the zooids grow and that the typical form already noted becomes increasingly evident.

Passing now to the unbroken portion of the chain, we find that the zooids increase in length very slowly at first and more rapidly later; also that though the curve is fairly smooth at first, it becomes quite irregular toward the end. Upon closer examination of fig. 2 and the graphs of other chains, we surmise that these irregularities are the forerunners of the groups making up the wheels; in other words that the periodicity shown so plainly in the wheel part of the chain extends back into the unbroken part. Were this found to be true, the fact could hardly be ignored in considering the problem of the break-up of the chain and the production of wheels.

In order to test the conjecture more critically we submitted the measurements to Mr. George F. McEwen, the mathematical expert of the Marine Biological Station of San Diego for examina


tion. Out of this examination has come the graphs shown in figs. 4, 5, 6, 7, and 8.

A curve was computed to fit the graph (fig. 2), as nearly as possible. From the equation of this smooth curve we get a 'calculated value' for each zooid; that is the length of each zooid, if the series were as smooth as our calculated curve. We next subtract the observed length of each zooid from the calculated length, and get a series of values, some plus and some minus according as the irregular graph went below or above the smooth curve. When we plot these plus and minus values above and below a horizontal line we have the graph fig. 7. It shows that the values follow the curve fairly well at first and then vary more and more; in other words, that we have a periodic curve of increasing amplitude. ^

'Mr. McEwen gives the following summary of the method used: The sizes for each of the points corresponding to the numbers 45, 50, etc., to 90 were taken as the ordinates of a curve whose abscissae were 1, 2, etc., to 10. It was assumed that the above curve corresponded to an equation of the form

y = a+bxi+cxi

and the most probable values of the coefficients a, b, and c were computed according to the method of least squares. By substituting {2x — 8) for Xi in the above equation, the equation

y = a + b (2x — 8) + c (2x — 8)'

was obtained in. which, if y^ of the number of the point is substituted, will equal the computed value of the corresponding size. (This equation was used to calculate the corresponding values of y, which were used in connection with the observed values for computing the algebraic sum of the residuals and the probable error, for the purpose of determining if the equation was a proper expression for measured values of y.)

It was assumed that this equation, determined from the 10 points was very nearly the same as if it had been computed from the 45 actual points, and therefore represented the relation between the number and the average size of all the points. This assumption was verified in one case by including all the points and comparing with the result when only 10 points were used.

The observed values of y were subtracted from the corresponding computed values and these differences were plotted as ordinates against the numbers as abscissae, thus giving a representation of the deviation of the observed values from those given by the equation. These deviations are due to errors in the measurements, and to the fact that the assumed equation was not a true expression for the relation. As the error in measurement was =*= 0.1, it is evident that the deviations are due mainly to the latter fact.

The periodic character of these curves shows that the true law is a periodic fluctuation of increasing amplitude about a mean value increasing in a regular manner with the number of the point.

Fig. 4 Plot of differences for chain IV, right side.

Fig. 5 Plot of differences for chain VI, right and left sides.

Chain IV, whose plot of differences is shown in fig. 4, is one of the smaller chains and in it one would expect to find the grouping less evident than in the larger chains. However it can be plainly seen even here. The right and left sides of Chain VI are shown in fig. 5. With Chains IV and VI, the differences were figured only for the zooids 45-90. In figs 6 and 7, the two sides of Chain VII

Fig. 6 Plot of differences for chain VII, left side

are given entire, the curves and the differences being figured separately for the two parts of the chain, since it can be fitted better when but half is considered at one time. The complete series being given, one can more readily see how the amplitude of the waves increases toward the end.

It will be remembered that in computing the differences, the observed values were subtracted from the calculated values. Hence upward curves in fig. 2 appear as downward curves in fig. 7, To make the comparison with the wheel graphs easier the signs

were reversed for fig. 8 so that values greater than the corresponding ones in the fitted curve he above the x axis while smaller ones lie below. Vertical dotted lines have also been drawn to indicate a possible grouping of the zooids. Irregularities appear, it is true, but since irregularities often appear in the wheels also it is to be expected here. Moreover, with such small values the chances for error are so great that one would expect considerable variation.

Fig. 7 Plot of differences for chain VII, right side

What we get then from these plots of differences is the probable fact that the unbroken paH of the chain really shows a periodicity or incipient grouping closely resembling that of the wheeled portion of the chain, the groups including four to eight zooids each, which is the number found in the completed wheels.

The plots of differences brought to light an aspect of the matter which had not been anticipated, namely, the existence of another toave with a longer period, shown in all the curves. The plots in



figs. 4 and 5 have been smoothed by averagmg for every ninth point in order to make this curve more plain. The curves are not just the same for the two sides of Chain VI, the difference probably being due to a difference in the way in which the computed curve fits in the two cases.

With what, if any, other biological phenomena in the species this newly discovered periodicity is connected we do not know.

7zoo/ets. : 7zooids \6 zooids. \S zooidj. \S zooids. • S zooids\ AS zooids. \ 4zooids

■<6-S2. \53-J9. \60-eS. \66-70. '7/-76. .77-81. \]«-fftf. '^7-90

Fig. 8 Plot of differences for chain VII, right side Inverted and possible groups indicated

Of its existence however, there seems to be no doubt; and it is certainly interesting to recognize that we have here an instance, by no means uncommon in organic phenomena, of waves, so to speak, of one size riding upon those of another size. It is highly desirable to take these cases in hand with a view to finding their connection with other phenomena.



1 . Segmentation of stolon and the deploying point

To find other factors entering into the wheel production, a study of the structure of the chain was made. The portion of the chain in which the zooids are in single file is of the same general form as that of other species. The incipient zooids, marked off by the infolding ectoderm, have their aboral ends uppermost, and the dorsal side of each against the ventral side of its neighbor, the dorsal sides being towards the base of the stolon. The blood supply passes out through one-half of the large axial blood vessel and back through the other half. The segmentation of the stolon in some cases extends to the root of the stolon, in others not quite so far. A possible significance of this variation will be pointed out in another connection. Where segmentation extends to the root of the stolon, the more proximal segmentation lines are very irregular and this fact may be of considerable interest in a way we shall not stop to consider here. Judging by some chains, one would say that the segmentation begins at the sides of the stolon, but others lead us to suppose the beginning is above and below (along the genital rod and the neural tube) while in still others it seems to be equally advanced in all parts of the circumference. The latter condition is probably the usual one.

2. Position and relation of the zooids in the chain from the deploying point to the twist

a. Shifting of the zooids. What we call the deploying point, is the point, or region where the zooids, by moving alternately to the right and left shift from single to double file. While the zooid is moving out, it also moves upward and begins to turn, so that its dorsal side faces out instead of toward the base of the stolon. These changes .take place gradually. The oral ends shove out and begin to turn and before the turn is complete, the aboral ends


shove out and turn. The sketch of the deploy mg point will make this clearer. Figs. 14, 15 and 16 are dorsal, lateral and ventral, views of the deploying point of one chain. These drawings were outlined with the aid of the camera and much care was taken to make them accurate. Calling the zooid whose oral end has just begun to shift, no. 1, the aboral ends of nos. 8 and 9 (numbering on one side only) are beginning to do the same. No. 25 (not shown in the figure) seems to have reached the final position with the rearrangement of the internal organs complete. We find now that the right sides of the right-hand zooids (considering those to be right-hand zooids that correspond to the right-hand side of the parent) and the left sides of the left-hand zooids are toward the base of the stolon. This statement applies to the chain before it emerges from the parent. The orientation of the older, extruded part of the chain is given later.

All of the observations on the early growth and differentiation of the chain agree with those made by Brooks for C. pinnata with the exception of the orientation. Brooks ('93 p. 79) says of the single file zooids:

At this stage each Salpa is bilaterally symmetrical, and its plane of symmetry is the same as that of the stolon, while its long axis is at right angles to that of the stolon, which becomes converted into a single row of Salpae, so placed that the dorsal surfaces of all of them are toward the base of the stolon, their ventral surfaces towards its tip, their right and left sides on its right and left respectively, their oral ends at its top or neural side, and their aboral ends at its bottom or genital side.

Again in his description of the double row he says :

The single row of Salpae becomes converted into a double row, which consists of a series of right-handed Salpae and a series of left-handed ones, placed with .... the left sides of those on the right and the right sides of those on left towards the base of the stolon.

-The loop-like structures seen at the oral extremity of the zooids in figs. 14 and 15 might easily be mistaken for the intestine. They are not this structure, but indicate very nearly where the oral orifice will appear.


This, as will be seen by comparing it with our description, is the opposite of the condition found inC. aifinis, since Brooks places the oral ends of the zooids uppermost while we find the aboral ends up. This mistake was probably due to lack of sufficient material for the study. He says (p. 87):

In all my preserved specimens the tip of the stolon had been so much flattened by contact with the side of the bottle, in transportation, that I have not been able to study in detail the way in which this wheel-like arrangement is acquired, and the subject should receive the attention of those who are able to stud}^ living specimens.

It is a point upon which one could easily go astray if hampered by a lack of material.

As the changes in internal organization seem to correspond with those of C. pinnata, and as Brooks'description is so clear and complete, we need not go into the subject, but refer to his account (Brooks, '93 pp. 80-106.)

When the zooid has moved into its secondary position it lies upon the stolonic blood vessel rather than to the side or around it. With this change, two small vessels develop for each zooid, one leading to it from each half of the stolonic vessel (fig. 24, ihv.). The blood flows along one-half of the main vessel (say the upper half) out through the upper small vessels to each zooid and returns by the way of the lower set of small vessels to the lower half of the main vessel where it joins the inflowing current. These currents are reversed with the reversal of the blood current in the parent. The zooids now increase in size very rapidly, lengthening out more above the upper level of the vessel than below it, so that at the twist the oral ends extend but a little way below the vessel, while the aboral ends extend far above it. As a result,the aboral ends of the zooids of opposite rows come in closer contact than do the oral ends. Since the zooids of the two rows are arranged alternately, each zooid will lie against two of the opposite row. As growth continues and the zooids, through their increased size, move outward as well as upward, they are forced farther apart, but the connection is retained through peduncles which now develop.


b. The peduncles and foot-pieces. These structures play so important a role in the production of the wheels that they must be described in some detail. Almost all the figures show the peduncle in one stage or another. Fig. 12, best gives its relation to the full grown zooid, showing that it is a thin flap or sheet extending out from the ventral median line. The diagram (fig. 17) shows that the peduncles of the series are parallel throughout the first part of the chain, and that each by means of its 'foot-piece' {fp.) is in contact with four others, its two neighbors in each row. These foot-pieces are also well shown in the right-hand portion of figs. 19 and 22. As the zooids grow and extend out farther from the blood vessel, the peduncles lengthen, and the foot-pieces grow longer as the zooids grow wider, at least until the region of the twist is reached.

Along with the great increase in the size of the zooids and the development of the peduncles comes a change in the circulatory system. The two individual blood vessels coalesce to form one vessel with two channels (fig. 24, ihv.). The blood current has the same course as before except that the incoming and outgoing currents of each zooid pass through one vessel. Observation of the blood currents in the living animals made the task of working out the circulation much easier and more certain than it would have been if confined to preserved specimens. The cross section (fig. 25) shows well the relation of the zooids to the blood vessel, the foot-pieces joining the zooids above the vessel, (fp.) and the individual vessels leading from the zooids to the two parts of the large vessel (ibv.)

c. The emergence of the chain to the outside world. Through the first part of its course, the chain is enclosed within a definite tube in the test just below the endostyle, this tube ending at a point just posterior to the placental vessel and anterior to the first body muscle band. This first muscle bends posteriorly here so that its insertion is along the lower part of the tube opening. The placenta usually disappears before the chain reaches this point. There seems to be more or less of a cavity left in the test where the placenta was, and the chain, as it reaches this point, no longer being held in its horizontal position by the tube, following the


line of least resistance, turns down into the cavity, and by the rapid growth of the zooids, soon breaks through the thin wall to the outside, the tip bending downward.

d. The twist in the chain. The general character of this part of the chain naay be seen from figs. 11 and 13, while the peduncles and blood vessels of the region are shown in the diagrams figs. 17 and 1,8. Before the twist, we have within the parent, a straight double row of zooids with oral ends down. After it, the chain is turned back under the parent, and the zooids are again found with oral ends down. Until the zooids break through the test to the outside the chain has not begun to twist, the zooids still lying symmetrically along both sides of the median line. In fig. 13 thirty-six zooids are outside and the twist is just complete. The presence of two rows of zooids in the chain makes the turn appear more complicated than it really is. The chain simply doubles back under and then turns over, this turn being almost invariably to the left. This leaves the zooids with aboral ends again uppermost, but the row that was before on the left is now on the right side of the parent.

e. Reduction of foot-pieces, first break in the chai7i,and formation of the first wheel. The first visible intimation of the break-up of the chain comes in the peduncles and foot-pieces. The footpieces (fig. 17) gradually grow longer toward the distal end of the chain, coming to their maximum length a little before the end of the unbroken part is reached. After the maximum they shrink (fig. 22). The decrease is much more rapid than the increase, there being only about sixteen to twenty-four zooids in the diminishing series. Fig. 19 shows that the first group consists of nine zooids whose peduncles have broken loose from the rest. The foot-pieces have shrunk still more and the distal ends of the peduncles have been drawn closer together. But while the foot-pieces, by which the zooids are held in the axial line of the stolon, become successively and rapidly smaller just before the beginning of the break in the chain, the zooids themselves are becoming constantly larger. A consequent crowding of the zooids results. This brings about a pushing of the bodies of the zooids forward in the chain beyond the foot-pieces. The strain to which


the series of adhering foot-pieces is thus subjected results inevitably in a pulling apart of the foot-pieces somewhere. As a matter of fact the break produces groups and not single pairs. These groups then promptly shape themselves into the wheels.

3. Comparison of the rate of growth of the chain as a whole with the rate of growth of other animals. As a matter somewhat to one side of the main problem, we have thought it worth while to compare the rate of growth of the chain as a whole with what is known of the growth of other organisms. This was done by the method employed by Minot ('91) in his study- of the rate of growth of guinea pigs; namely, by finding the per cent of increase throughout the chain. The values of the corresponding zooids of Chains I, II, III, VI, and VII were averaged. (Chains IV and V being so much shorter were omitted.) The per cent of gain of the secoiwl over the first, third over the second, etc., was then computed and the values plotted. The result is a very ragged line showing a gradual increase through two-thirds of its length and a more sudden drop at the end. To get a graph whose course was more evident, the increment was computed again, this time taking the series in groups of five. The first value here is the per cent of increment of the second five over the first five, etc. (table 4, fig. 9). The gradual increase, maximum toward the end, and rapid decline is here plainly shown in spite of the limited data.

This result seems strikingly different from that for the guinea pig and other animals of higher order, where the per cent of increment is a diminishing one from birth on. The difference may, however, be more apparent than real since, to make the comparison more correct, it would seem that stages in the mammalian development preceding birth would have to be used.

The drop in rate of increase, when the wheel part of the chain is reached, may be significant for the comparison, but we do not consider our observations carried far enough into the life of the chain as a whole to warrant any speculation based upon them. A study of the growth of still younger and still older, larger zooids will have to be made to meet the requirements here.

Fig. 9 Percentage of increment throughout chain


1. Cause of the twist

Although salpae do not move through the water very rapidly, still there is enough motion to make the end of the chain double back imder the parent, as soon as it projects into the water. The reason for its turning over is less evident. The zooids begin to pulsate some time before the twist is reached and, having been with aboral extremities uppermost in the original or normal position, we may suppose they tend to assume the same position again when the normal state of things is interfered with by the bending back of the chain. Observation of the living animals shows that the chains of wheels and the separate wheels (at least the smaller ones) usually move along with aboral extremities uppermost.

It is easier to say that zooids ' tend to assume the normal position' than to show the cause of this tendency. It may be that the specific gravity of the oral ends is greater, or that the pulsation may have something to do with it, or there may be some tropism



TABLE 4 Per cent of increment throughout the chains




1- 6

6-10 11-15 16-20 21-25 26-30 31-35 36-40 41^5 46-50 51-55 56-60 61-65 66-70. 71-75 76-80 81-85.

9.2 9.9 10.3 9.3 11.7 13.3 13.4 12.6 15.1 14.9 17.9 16.7 17.4 14.2 15.1 16.7 14.5

Wheel portion


involved; but we have no observations under this head. As the chain Hes, with aboral ends of the zooids uppermost, the propulsion of the zooids drives them away from the ventral side of the parent, while if they were with oral ends up the pulsations would drive the oral ends up against the ventral side of the parent. In one specimen in which the, chain had not yet emerged, the end of the chain was turning to get around the placental blood vessel. Had our observations been limited to this one instance, we might conclude that the twist is initiated in this way. But after looking over a large amount of material and finding that the placenta usually disappears before the chain reaches that point, it is evident that this in no way accounts for the twist.

2. Unequal growth of zooids arid foot-pieces as a factor in the breaking up of the chain

We seem to have found a cause sufficient for the present research, for the break, in some way, of the chain of zooids. This is, as already pointed out, the unequal growth of the bodies and footpieces of the zooids. The first question that arises when we attempt to push the analysis farther is, why is the break into groups rather than into single pairs of zooids? Nothing in the differential growth recognized appears to bear upon this question. So far as that is concerned we should suppose the zooids would be picked off one by one, or at most in single pairs.

Just how constant these groups are, may be seen from the frequency polygon (fig. 10). We see that of ninety-two half wheels seventy-three contained six or seven zooids each, while only two contained eight, and five contained four zooids. This constancy is to be expected when we regard the breaking apart as a growth phenomenon depending upon constant causes rather than upon chance.

In some waj^ the wheel phenomenon is clearly dependent to a large extent on the strength of the adherence among the footpieces, which are but parts of the central ends of the peduncles. As may be seen by fig. 18, the radial blood vessels, the other main connection of the zooids, break apart early in the life of the











































































No. of zoo/e/a... Ma/flV/iee/. Whole hf heel.

Fig. 10 Frequency polygon showing the number of zooids in the wheels

wheel, so that in the fully grown wheels the zooids are held together almost entirely by the peduncles. The observed facts certainly suggest group adherence among the foot-pieces themselves. Can direct evidence of any such thing be obtained?

Having proved the existence of a pronounced size grouping of the zooids in the wheels it naturally occurred to us that there may be something of the same sort in the peduncles and foot-pieces. We consequently made a considerable number of measurements on these structures. Some of the numbers are given in table 5, and in fig. 1; the graphs of peduncle lengths (dotted hne) and footpiece lengths (lower continuous line) are presented. It is doubtful if these show anything. We have not assumed that they do. The difficulties in the way of making the measurements are too great for the methods employed. It should, however, be borne distinctly in mind that these negative results prove no more than the insufficiency of the measurements. The fact that the footpieces do cling to one another in groups, and that the zooids to to which they belong are demonstrably different in size, appears to make it probable, a priori, that the adhesive power of the footpieces is of the gradational, or periodic sort, in spite of our failure to find it. The suggestion is that the graded size of the zooids is reflected in the adhesive power of the foot-pieces. Could this




si U






I0»00i00000i0000000i00 0000»0000



i <N CC .-I 1 '



00 "3 00 .-i^oJ

cooo 00 t~-t>

00 o



<x)oot^!£!t^Mc<ic<^(^^l-Hco■*■^co■<4^co(^^coeoc<^Meoc<^<^^'-^c^^.—^,— (.-HC],— looocio



conjecture be proved true, an exceedingly important biological point would have been made.

And now as to the evidence that a periodicity corresponding to the future wheels does exist in the chain before its break-up. In discussing the results of our treatment of the data pertaining to the unbroken part of the chain, we said the curves, as shown in fig. 2, for example, 'probably' show a periodicity. We permitted ourselves to doubt to this extent, in the interest of conservatism. We wish now to sum up the evidence for periodicity. Its strength lies in the fact that it is cumulative rather than in the sufficiency of any one piece.

In the first place, does not the undoubted fact of periodicity in the wheels themselves, and the groups that immediately precede them, make the presence of periodicity in the rest of the unbroken part of the chain probable a priori? It would seem so. In the second place mathematical treatment of the quantitative data makes it almost certain that a periodicity corresponding to theory actually does exist. Third and finally the probable extension of the periods far back into the young part of the chain, leads us to suspect that this fact is connected with another observation of quite a different order, an observation, that is, which strongly indicates that the periodicity is really established at least as early as the segmentation of the stolon itself.

One of us has shown that in Salpa fusiformis-runcinata the very early segmented part of the stolon may be interrupted by an unsegmented part (Johnson, '10, p. 154 and fig. 8). While such interruptions have not been observed in Cyclosalpa affinis attention was called, when speaking of the first stages in the segmentation of the stolon, to the fact that in some cases the segmentation reaches to the very root of the stolon, while in others a stretch of unsegmented stolon exists. May not this difference indicate a periodicity in the segmentation corresponding to the periodicity in growth that we have found?

The reader may think that the grouping, as shown in the plots of differences, is too variable and indefinite to warrant the conclusions we have drawn. True, the groups here are not as regular as the wheel graphs shown at the end of the curve (fig. 2) , but though


the small groups appear to be more irregular on account of their riding on the secondary waves, they are of the same sort. It must be remembered, too, that the values are very small and the chances of error are large. In fact, such a uniformity of result throughout all the graphs examined, in spite of small values and difficulty of measurement, is very convincing.

The transformation of the groups of zooids into wheels is easily understood : The moment the break occurs so that the pressure of the zooids upon one another in the group can exert its effect backward as well as forward, the hindmost pair swings in toward the axial line, each of the other pairs up to the transverse middle line of the group following in its proportional amount. Since by this time the foot-pieces have wholly or almost wholly disappeared and the central ends of the peduncles have become closely appressed, the swing of the zooids disposes the peduncles in the form of the spokes of a wheel, the hub being represented by a small elliptical space. The course of things here described is illustrated in fig. 17. That the pressure tending to force the mid-zooids of the groups outward is considerable is obvious from the zig-zag form into which the axial vessel is thrown, due to the pull on the radial vessels, as seen in the second group of fig. 18. The disappearance of the axial vessel in the older wheels may be supposed to be partly due to the same cause, although probably the vessel is actually in course of degeneration.

3. Impossibility that the character of the blood supply to the zooids can be the cause of the size schemes within the wheels

No study involving the growth of the zooids could be complete without attention having been given to so fundamental a matter as that of the blood supply. For example, the question naturally arises, does not the break-up of the chain into groups so affect the common blood vessel of the stolon that the zooids do not share alike in nutriment received, and is not this inequality responsible for the disparity in size among the zooids?

The changes in the circulatory system are best shown by the diagram fig. 18. At the end of the continuous part of the chain,



the individual blood vessels are arranged at regular intervals along the large vessel. The arrangement is the same for the first wheel, but with the second or third wheel the axial vessel begins to shrink. As the vessel remains in connection with the individual lateral vessels, while growing smaller, it comes to have a zig-zag course, due to the opposite but alternate pulls upon it by the growing zooids. The shrinkage of the vessels goes on so rapidlj^ and to such an extent that in the next wheel the vascular connection between the central zooids is lost. The portion of the main vessel which joins two wheels together persists for some time. In fact this and the transparent cellulose envelope which forms around the wheels, filling in the spaces between the zooids, are all that hold the chains together and very long chains of wheels are sometimes found. The small remnants of the individual vessels gradually disappear. Though these vessels end blindly, the blood may still be seen in them for some time, flowing out one side and back the other. After the disappearance of the main vessel at the center of the wheels, short circuits are maintained between the zooids connected at any point. Thus in fig. 18, zooids 1, 2, 12, 6^, and 7^ have a circuit of their own. Thus it would seem that if any of the zooids of the wheels have an advantage over others the end ones would be favored as against the middle ones, but the middle ones are on the whole larger. Hence inequality in blood, supply seems to be excluded from being a determining factor in the size relations observed.

If there be any communication between the zooids of the unbroken chain or of the wheels, other than by the circulatory system just described, it must be through the peduncles. The vessels in the peduncles are irregularly arranged but they are distinctly larger toward the edges and reach part way into some of of the papillae. They are easily followed in the living specimens. To test the question of blood communication between zooids, injections were made. Methylene blue in sea water was used, which could plainly be seen in the transparent peduncles and in the bodies of the salpae. The first attempt was on two wheels whose stage of development was the same as the third and fourth in fig. 18. The needle was inserted in the stolonic vessel half way


between the two wheels. The color shot out through the small vessels to the peduncles of the zooids still remaining in contact. It went throughout the vessels of the peduncles of the zooids but stopped cleanly at the edge of the peduncle. No zooids whose connection with the main vessel had been lost showed any touch of the color. However, the wheel was again examined about fifteen minutes later. The stained zooids had died and dropped away from the wheel, the peduncle dropping away with the zooid. A slight stain was found around the papillae of the peduncles of one or two of the other zooids where they had come in contact with the stained ones. We conclude that there is no direct vascular connection here but that there is possibly some interchange by absorption through the thin ectoderm. Another injection was made in the peduncle of one of the zooids in a wheel. The color flowed throughout the peduncle and into the zooid but did not enter other zooids of the wheel. We therefore seem driven to conclude that the hlood supply is not a factor in the size differentiation of the zooids of a wheel.

4. Unlikelihood that the wheel arrangement of the zooids in Cyclosalpa has, as believed by Brooks, anything to do with the position of the four first blastozooids of Pyrosoma

Brooks was firmly convinced that the radial, or wheel arrangement of the asexually produced zooids in Cyclosalpa is homologous with the radial disposition of the first four blastozooids of Pyrosoma. This he regarded as one of the strongest evidences of the close relationship between the two genera. Thus he says (Brooks '93, p. 133) :

The opinion that Salpa and Pyrosoma are closely related does not however, rest upon superficial resemblances, but upon their fundamental identity of structure, although one of the details, the resemblance in their asexual multiplication, is so complete as to be almost enough in itself to establish their affinity.

The same view he expresses with only a little less assurance in several other connections. We had no thought, in entering upon


the present study, of considering this point, nor do we propose now to go into it extensively. However, our results on the growth and mechanical factors involved in producing the wheels of Cyclosalpa seem to have so much bearing on the question, that we can hardly pass it by without notice. The resemblance between such a figure of the Cyclosalpa wheel as, for example, that given by Brooks ('93, pi. 1, fig. 2), and reproduced by Delage and H^rouard (p. 203, fig. 151) and a figure of an early Pyrosoma colony like 15, (pi. 31), by Huxley ('59) is considerable and not unnaturally suggests true heredity kinship. The moment, however, one comes to look into the details of how each group comes about ontogenetically rather than phylogenetically, he finds them so different that his imagination is balked at an attempt to interpret them as both referable to a common hereditary operation. In the first place Brooks seems never to have observed the fact that the Cyclosalpa wheel is at the outset bilateral. None of his published figures give any intimation of this, nor does he refer to it in his text. For instance, the two figures, 8 and 9, pi. 2, of his latest publication (Brooks, '08) represent wheels of C. fioridana, and C. pinnata as though they were perfect — as though the zooids were disposed in exactly the same way throughout the circuit. We would not, of course, assert that he did not draw just what he saw in these two instances, especially since we have had no chance to examine the wheels of C. fioridana, and have seen but a single one of C. pinnata. In the one specimen of C. pinnata which we have, attentive study finds that two zooids on opposite sides of the circuit have slightly different positions from the others. These probably indicate where the axis of the chain lay; but the departure from perfect regularity is so slight and of such a character that it might be easily overlooked had one not discovered, by studying the formation of the wheels, what their real nature is. In C. afiinis the bilaterality of the wheels is probably never wholly obliterated.

The first four ascidiozooids in Pyrosoma, on the contrary, stand in single file as do the Salpa zooids before the deploying point is reached and the radial order is taken on by the swinging around of the file so that number four comes to be adjacent to


number one. Further there is no opportunity in the Pyrosoma g^roup for the differential mechanical action caused in Cyclosalpa by the growth and crowding of the zooids while the foot-pieces diminish in size. Neither is it possible seemingly, for the periodic phenomenon to play any such part in the arrangement of the Pyrosoma zooids as it appears to in Cyclosalpa.


1. Supple7nenting biological with quantitative observations

We venture to call attention to the way in which morphological and physiological observations and considerations join hands with quantitative observations in this research. Numerous structural details in the adult individuals of both sexual and asexual generations, in the chain of zooids as a whole, in the individual wheels and the individual zooids composing the wheels, and in the unbroken part of the chain both as a whole and as to its individual elements had to be attended to. On the functional side not only growth in several of its aspects, but the mode of swimming, certain facts pertaining to the circulation of the blood, and some points about nutrition have come in for consideration.

All this sort of thing is so familiar to modern biologists as to need no special mention. Not so with what we have done in a quantitative way. It seems to us that in this we have entered a region of research that biologists will be compelled to regard vastly more seriously in the future than they have in the past or do now. The case in hand furnishes a rather striking illustration of what the quantitative method can do. It can enable us to see facts we cannot see otherwise. It amounts to a great increase in the power of our eyes just as does the microscope. This statement is to be taken literally, not figuratively. One may easily imagine a magnifying instrument that would so enlarge the wheels as to make visible the size differences between the zooids. It would seem that this is what the application of mathematics in physical science very frequently does. We should never have suspected from ordinary examination size differences


of a systematic character among the zooids of the chains. It was only from certain biological considerations combined with aid from this instrument, that the existence of the system was made certain. And it should be specially noted how our results would have been affected by failure to recognize this fact. The breaking up of the chain in some way, and the production of wheels from the breaking, could have been inferred from the unequal growth of the bodies and foot-pieces of the zooids; but why the breaking should be into groups rather than into single pairs would have remained with no definite answer but for the discovery of the periodicity in growth in the unbroken as well as in the broken part of the chain.

2. Natural periodicity in organisms and exacter methods of research

But promptly comes the question from some of the foremost biologists, What of it? What particular good is there in knowing that growth is periodic so long as we have no explanation of why it is so? Our real interest, they say, is in the causes not the mere facts of organic phenomena. This objection displays, in our opinion, one of the most pervasive and fundamental weaknesses in the biological philosophy of the day. Looked at critically, it is found to mean that facts of nature, in order to be interesting and deemed really worth while, must be prejudged; that an explanation of them must be ready at hand before they are observed in order that they may be attractive. The issue must be looked squarely in the face. It is in fact the old, old issue between the inductive and the deductive methods of interpreting nature; between observation and reason going hand in hand, and the power of reason alone; between the a posteriori and a priori modes of reasoning. The objection carries the implication that great numbers of facts of nature can be explained without having been themselves examined; that the unobserved causes of many observable effects may be sufficiently inferred from observations on other effects than the particular ones under consideration. In a word the meaning is implied if not expressed, that some time nature may


be fully known without having been fully studied. This conception of nature and the knowledge of nature is always and everywhere the begetter of dogmatic assertion on the part of leaders, of subserviency to authority on the part of followers, and of idolatry to certain facts and neglect of others by everybody. This is not the place to go into the logic, or rather, the epistemology, of biology. The case under treatment does, however, justify us in a few observations and reflections on procedure in research.

Why is it that the biological sciences are designated as observational and descriptive, to distinguish them from the physical sciences which are called quantitative and exact? Surely no present-day student of nature would contend that living objects are qualitative alone and so must be dealt with in terms of quality, while non-living objects are quantitative and are to be dealt with in terms of quantity ! There is surely no structural part or activity of any organism that does not exist in some quantity or other, and hence is not susceptible of being measured in some way. Contrarywise, there is surely no inorganic body or substance that has not qualities of some sort by which it is described and defined. Yet why is it that in spite of the brave effort made by a few distinguished men of science during the last half century to introduce conceptions of quantity and the methods of mathematics into biology, these efforts have met with only limited success at best, and are ignored in practice and frowned upon in theory by many of the foremost bilogists? Only a few months ago a distinguished investigator declared in the presence of the senior author of this paper that the quantitative method in biology is dead, and this student suiting practice to theory, though working in fields where quantitative conceptions and exact determinations are particularly important, rarely attempts to measure in any rigorous way the biological phenomena with which he deals. Attention cannot be called too strongly to the extent to which much of what is esteemed the very highest type of recent biological work has laid stress on accurate quantitative determination of certain environmental factors of organisms, but has ignored almost wholly quantitative determinations of the vital phenomena themselves. There can be no question about the


importance of exactness in the determination of external factors. So far these methods are admirable; but, it appears to us, it must be recognized that when exactness has gone thus far it has gone at best not more than half the way. Nothing less than equal exactness all along the line will do to fulfil the highest demands of physical science.

Let one recall the degree of refinement with which physicists and chemists are measuring the phenomena with which they deal: the wave lengths and angles of refraction of light; the quantity of heat generated in chemical reactions; diffusion rates of gases and liquids; atomic weights and combining ratios, and innumerable other things. Then let him compare these with the ridiculously crude quantitative determinations made in nearly all departments of biology. A few aspects of physiology, as for instance, the temperature of the human body; and a number of phases of the psychology of higher animals — reaction times, for example — have been brought under mensurational treatment comparable with the standards of exactness long demanded in physics. But the vast fields of morphology, of general physiology, of individual and race growth and decline, of propagation, of variation, of automatic and responsive action, etc., have hardly been touched quantitatively as physics and chemistry would understand this term. As yet we in biology have hardly heard of anything corresponding to physical constants, units of measurement, coefficients of change, etc. Yet will any one, fully alive to the spirit of modern physical science, venture to maintain that inorganic phenomena are so utterly different from organic, that conceptions and practices so enormously fruitful in the one realm are wholly inapplicable in the other?

It is a significant fact that many biologists, the most ardent in defence of the so-called mechanistic or materialistic view of living things, are farthest away from, even most hostile to, the very methods for biology proper that have so largely made the physical sciences what they are. One looks in vain through numbers of technical writings by biologists of this school for anything like exact, comprehensive accounts, either qualitative or quantitative of organsims or parts of organisms, or even functions


of organisms, dealt with. Yet how these writings bristle with such expressions as 'differs considerably/ 'constant results.' 'as a rule/ 'very similar/ 'normal segmentation/ 'normal nuclear spindle/ 'normal blastulae/ 'normal animal/ 'practically identical/ 'essential features,' 'increases in exact proportion,' and so on!

Two rejoinders are frequently made to this demand for carrying more exact methods into biology. One is on the purely theoretical ground that it is not necessary; that 'mere quantity' is of no great moment in life phenomena; that slight differences are of the purely 'fluctuating' or individual sort, so have no large significance. To answer this objection in full would take up much farther into philosophical discussion than we can go here, but it may be the more warrantably passed by because the attitude of mind that makes it is seen to be obviously hostile to the whole trend and spirit of physical science. If the history of progress in science can be relied upon to furnish any clue as to how progress is to be continued in the future, the man of science, who holds a general view of nature that makes many facts insignificant and negligible, is bound to come to grief sooner or later.

The other objection is more practically justifiable. It is that the phenomena of living beings are so complex and subtle, and that animals, especially, are so sensitive to changes in external conditions as to make it impossible to apply to them in more than a very limited way, the exacter methods of the physical laboratory. Our answer to this is two-fold. In the first place, we are persuaded that exact methods could be applied far more widely than they are, and they undoubtedly would be, did our general conceptions call for such applications. The other answer is that if it be true, as it well may be, that many life processes are too subtle and involved to submit to measurement on an exact and large scale, then the only course open for the interpretation of such processes is to iritroduce no considerations that involve the conception of accurately measured quantity. The extent to which this principle, seemingly so obvious and unescapable, has been violated in much biological theory during the last quarter century or more, is seen to be remarkable once


one comes to think about the matter. For example, reflect on the extent to which theories of development and heredity have made use of the notion of equation and reduction nuclear divisions of the germ cells ; yet who has determined in any rigid quantitative way the elements that enter into the hypothetical equalities and inequalities? How familiar is the textbook statement that the chromatin of the male fertilization nucleus is 'exactly equal' to that of the female nucleus with which it fuses! But on what sort of determinations does this assertion rest? On scarcely another thread of evidence than that they ' look equal !' And here we come upon the almost incredible naivete with which biologists in most things eminently sound, have gone down before this fallacy! Only a short time ago while discussing this point with a number of biologists, one of them, a man of excellent standing and great carefulness in nearly all scientific matters, replied to my strictures, "if chromosomes look equal why are they not equal?" The words were hardly off the man's tongue when he saw what a remarkable statement he had made. The incident illustrates the straits to which one may blindly go in following a theory.

We conclude this topic with a quotation from John Tyndall. In his well-known address on the "Scientific Use of the Imagination," he says:

Let me say here that many of our physiological observers appear to form a very inadequate estimate of the distance which separates the microscopic from the molecular limit, and that, as a consequence, they sometimes employ a phraseology calculated to mislead. When, for example, the contents of a cell are described as perfectly homogeneous or as absolutely structureless, because the microscope fails to discover any structure ; or when two structures are pronounced to be without difference, because the microscope can discover none, then, I think the microscope begins to plaj' a mischievous part.

In view of the vast amount of evidence now before us from so many aspects of biolog;^^, that vital processes are periodic in their most fundamental manifestations, it appears unwarrantable to assume without proof that any whatever are not so.- But see what periodicity means ; it means that the phenomena are increas


ing and decreasing ; that they have phases ; that the time element being considered, they change in value from moment to moment. How then can we treat any particular phase, or stage of such phenomena so as to meet the demands of rigorous science without considering each phase in relation to the other phases? So far as they are treated without such reference the procedure would seem to be of the nature of 'random observations' — of the 'grabsample' kind — that always, whether in common life, business, or science finally proves to be inadequate if not disastrous. Astronomy, physics, chemistry, and in general geology, have passed quite out of this portion of their careers.

Taking it as established that biology is allied in essential nature with these older, less complex sciences, does it not seem inevitable that it too must move on and leave its cruder, haphazard methods behind? Does it not look as though this very fact of periodicity, this gradual come-and-go of things in the operations of organisms is to be one of the chief if not the chief way out? To press the inquiry a little closer, does it not look as though the wide prevalence of repetitive parts in reproduction and growth, which though like one another still differ from one another by some regular quantity, is to be one of the most important, though only one, of these exits?

It appears to us that cell division, for example, including the division of all cell parts subject to this process will have to be looked at sooner or later from this standpoint. Take the Foraminifera, for instance, unicellular organisms (according to the current interpretation) the bodies of great numbers of which become divided into many sections called nodes and chambers. In the great majority of species, as a glance at figures enables one to see, these divisions fall into quantitatively differentiated series. To make the point more cogent we introduce figures of two species Reophax membranaceus Brady (fig. 21) and Peneroplis arietinus, Batsch. sp. (fig. 20). Now let one compare these organisms with the salpa chain, the one, for example, represented in fig. 18, and catechise himself something like this : surely there is some resemblance between these objects. Both are composed of a considerable number of sections rather regular in form and much like one an


other, though obviously differing from one another in size. Both objects are hving, and both have come to be what we see them by a process of organic growth. Can we properly ignore these similarities in our efforts to interpret the organism, because on the whole the differences between them are more numerous and conspicuous than are the resemblances? Is it not at least possible that by turning to these few correspondences seriously they may serve as the starting point for the discovery of still others, and finally result in the detection of laws of organic growth and functioning that would greatly broaden our conceptions of, and hold upon, life phenomena?

One reason for selecting the Foraminif era as a group with v.'hich to make the comparison is the fact that the comparison of these organsims with higher ones in somewhat the same way has been made by several other zoologists. For instance, Schaudin ('95) speaks of the production and breaking off of parts in Calcituba polymorpha Roboz. as having "eine gewisse Ahnlichkeit mit der Strobilation."

But the most interesting comparison from our standpoint, of Foraminifera with other organisms was made by L. F. de Pourtales in 1850. At the meeting that year of the American Association for the Advancement of Science Professor L. Agassiz presented a short communication from this young zoologist in which Agassiz said:

Mr. Pourtales has, for the first time, pointed out a direct, well sustained analogy, which is to be found in the order of succession of the cells in foraminiferae of the genera Textularia, Candima, Biloculina, Triloculina, and Quinqueloculina. This succession agrees fully with the succession of leaves in plants — so fully that it can be expressed by the same fractions with which botanists are now in the habit of expressing phyllotaxis in the vegetable kingdom. This is, therefore, an important additional hnk in the investigation of the plan which regulates the normal position of parts in organized beings — a link which may lead to include into one universal formula the rhythmic movements which preside over the development of all finite l)eings. (Pourtales, '50, p. 89.)

This communication appealed strongly to at least one of those who heard it. At the next meeting of the association the presi


dent, Professor A. D. Baehe, said in what we should now call his presidential address :

The germ of two most important discoveries in natural history was contained in papers by two of our youngest members. [The first is omitted as not relevant.] The contents of the other were thus expressed : 'The order of succession of parts in foraminiferae is identical with the successive development of leaves in plants, and can be expressed by the same formulae.' Such discoveries, just warm from the study, it may be, as in these cases, forced to light by the occasion of our meetings, are among our greatest triumphs in the way of advancement. (Proc.,vol. 4, p. 159.)

We find no evidence that these ideas of Pourtales have been carried farther either by him or by any one else, though our examination of the literature with reference to the point has been far from exhaustive. D'Orbigny, twenty-five years before, had done much work on the fundamental types of growth in the foraminiferae, though we find no reference to his having compared the arrangements here found, with phyllotaxy in plants.

Our object in calling attention to this matter is, in the first place, to show that we are not quite alone in thinking such comparisons are profitable; and in the second place, to call attention to the possibility that exact studies in the quantitative relationship existing among the members of a repetitive series as well as upon the ordinal arrangement of these members, may be profitable. But should it be found that such studies are significant when prosecuted on unicellular organisms in which the segmentation does not go to complete severance of the pieces, it would seem to follow that they should also be significant when made on species in which the severance is complete, and then to all cell division whatever.

This, of course, brings us immediately to the cyclical phenomena in the propagation of the Infusoria that has received so much attention in recent years, particularly at the hands of Maupas, Calkins, Jennings, Woodruff, and others. Concerning these researches we do no more now than remark that if the general conceptions on which we are going are sound, the phenomena of


protozoon division, and of all cell division will have to be examined much more systematically and vastly more exactly, quantitatively, than they yet have been.

3. The inadequacy of treating periodicity generally, as an aspect of fluctuating variation

Here seems to be the place to point out how much more objective, more workable, more important 'periodicity' is in our conception than it is as usually conceived by biologists.

We compare our ideas on the subject with those held by only one other investigator, Hugo de Vries has dealt with certain aspects of periodicity exhibited by plants, quite at length and in several of his works. He states the general facts with clearness. (De Vries, '05, p. 721) :

This law of periodicity involves the general principle that every axis, as a rule, increases in strength when growing, but sooner or later reaches a maximum and may afterwards decrease. This periodic augmentation and declination is often boldly manifest, though in other cases it may be hidden by the effect of alternate influences. Pinnate leaves generally, have their lower blades smaller than their upper ones, the tallest being seen sometimes near the apex and sometimes at a distance from it.

There can be no doubt that the phenomena we are dealing with in Salpa and calling periodicity resemble closely those in plants thus described. The question, are they 'exactly the same' phenomena, we do not raise. Rather, we ask, in view of the closeness of resemblance ought they or ought they not to be looked at from much the same standpoint? The truth is de Vries has regarded the phenomena in plants very differently from what we have in Salpa, and his standpoint is surely inadequate for the facts we are dealing with. This dependency on local nutrition," saj^s de Vries, "leads to the general law of periodicity, which, broadly speaking, governs the occurrence of the fluctuating deviations of the organs" (p. 721). Again (de Vries, '01, vol. 1, p. 638) under the section, "Die Periodicitat semilatenter Eigenschaften," we read:


Ueber die grossere oder geringere Haufigkeit des Sichtbarwerdens ssemilatenter Eigenschaften entscheidet nicht nur die augenblickliche Lebenslage, d.h. die ausseren Einfliisse wahrend der empJSndlichen Periode der Entwickeluiig. Fast ebenso gross ist die Bedeutung der individuellen Kraft des jungen Pflanzentheiles, diese aber ist das Ergebniss der Wirkung der ausseren Factoren in den vorhergehenden Zeitabschnitten, theils nach Wochen und Monaten, theils nach Jahren gerechnet. . . . Diese Erscheimmg tritt am deutlichsten zii Tage in der Periodicitat der Anomalien aiif der Pflanze.

Again, pushing the matter a step farther, and in a somewhat different direction:

From a broad point of view, fluctuating variability falls under two heads. They obey quite the same laws and are therefore easily confused, but with respect to questions of heredity they should be carefully separated.^ They are designated by the terms individual, and partial fluctuation. Individual variability indicates the differences between individuals, while partial variability is hmited to the deviations shown by the parts of one organism from the average stature." ('05, p. 717). . . . . The individual differences seem to be due, at least in a very great measure, to such apparent trifles. (As differences in soil, moisture, light, etc.). On the other hand partial differences are often manifestly due to similar causes. . . . The development of the leaves is dependent on their position, whether inserted on strong or weak branches, exposed to more or less light, or nourished by strong or weak roots (p. 721). Then follows the quotation already given, viz., "This dependency upon local nutrition, etc."

De Vries' standpoint seems clear: Periodicity in plants is a special form of the more general phenomenon of fluctuating variation which in turn is due to 'ausseren Factoren.' The quantitative differences that manifest themselves in the periods may be lumped together and treated according to the law of probability as first applied to organic beings by Quetelet. After illustrating the application of the method of statistics, the author says: It should be repeated once more that the empirical result is

' It would be very interesting to have deVries follow up this point critically and impartially.


quite the same for individual, and for partial fluctuations" (p. 732). And: In the present state of our knowledge the fluctuation-curves do not contribute in any large measure to an elucidation of the causes." (p. 734.)

And so we come to the real issue. Certainly, as de Vries says, the differences called partial 7nay he treated en masse, so to speak. For example, we might pick to pieces ten wheels of the same dimensions of the Cyclosalpa chain, mix the zooids indiscriminately in a dish, then measure them and plot the results. The curve would be the same — the normal probability curve — but would give us no clue to the way the zooids are disposed as to size in the individual wheels. In that case the treatment would not, it is true, contribute in any large measure to an elucidation of the causes." But in our case we have seen that no evidence can be found tending to show that the size scheme as it actually does occur in the wheels is dependent on external factors. All the evidence is to the effect that it is due to the growth process itself independently of any correspondingly differentiating external conditions. In other words, the periodicity in growth occurs under external conditions, that so far as the evidence goes, are not correspondingly periodic. Viewed in this light, can we still say the curves teach us "measureably little about the cause of the phenomena under consideration?" It seems to us not. Truly they do not furnish us 'a complete explanation' of the phenomena. They do, however, tell us, seemingly, this much: That the cause is in the nature of the growth process itself; that the growth goes that way.

If now it should turn out as suggested that not only the length of the zooids falls into a size scheme, but that many of the other morphological dimensions, and functional capacities fall into similar schemes, then the instructiveness of the curves would, for us at least, be very great touching the causes of the phenomena.

Whatever view may be held as to the relation of the periodicity in plants to that in the Salpa chain, it will we believe be allowed that the general question is one of many sides and great possible importance to biological theory. We have not pretended to do more than call attention to it here.


We conclude with an acknowledgment of our indebtedness to the work of several other biologists who have entered by one or another gate the course upon which we find ourselves. Of these perhaps the first to be mentioned is Julius Sachs whose idea of the grand period of growth in plants must, it seems to us, expand and pla}^ a much larger role in biological theory in the future than it has in the past. After Sachs, chronologically, the various investigations by C. S.Minot on the rate of growth in aminals has largely influenced our observations and thinking. Another research, that by T. Tammes entitled "Die Periodicitat morphologischer Erscheinungen bei den Pflanzen,"has had considerable to do with shaping our ideas on the strictly biological side. But by far the most important as opening up the way to the quantitative work has been Raymond Pearl's Variation and Differentiation in Ceratophyllum." Although Pearl's quantitative data in this research are entirely enumerative rather than mensural; and although his aims and results are in several rather important particulars different from ours, his fundamental problem really gave us our starting point.




Brooks, W. K. 1893 The genus Salpa. Memoirs of the Biological Laboratory of the Johns Hopkins University, vol. 2.

1908 The pelagic Tunicata of the Gulf Stream. In Publication 102, Carnegie Institution of Washington, pp. 73-94.

Brady, Henry B. 1884 The Foraminifera. The voyage of H. M. S. Challenger. Zoology, vol. 9, and plates.

Delage, Yves, etHerouard, Edgard. 1898 Traite de Zoologie concrete. Tome 8, Les Procordes.

Huxley, T. H. 1859 On the anatomy and development of Pyrosoma. Trans. Linn. Soc. 23 (1862) p. 193-250.

Johnson, Myrtle Elizabeth 1910 A quantitative study of the development of the chain in Salpa fusiformis-runcinata. Univ. of Calif. Publications, Zoology, vol. 6, no. 7, pp. 145-176.

MiNOT, C. S. 1891 Senescence and rejuvenation. First paper: On the weight of guinea pigs. Journal of Physiology, vol. 12, pp. 97-153. (Also numerous other writings by Professor Minot.)

Pearl, Raymond (assisted by Olive M. Pepper and Florence J. Hagle) 1907 Variation and differentiation in Ceratophyllum. Publication no. 58, Carnegie Institution of Washington.

DE PouRTALES, L. F. 1850 On the order of succession of parts in Foraminiferae. Proc. of the American Assoc, for the Advancement of Science. Third Meeting , vol. 3, p. 89. Reference to this by Prof. A. D. Bache, A. A. S., Fourth meeting. Proceedings, vol. 4, p. 159.

Sachs, Julius 1873 Lehrbuch der Botanik, Aufi. 3. (The grand period of growth is dealt with by the author in various other publications.)

SCHAUDiN, F. 1895 Untersuchungen an Foraminiferen. I. Calcituba polymorpha Roboz. Zeitsch. fiir wiss. Zoologie, 59, 2 pp. 191-232.

Tammes, T. 1903 Die Periodicitat morphologischer Erscheinungen bei den Pflanzen. Verhand. Kon. Akad. Wetensch. Amsterdam. Tweede Sectie, Deel. 9, no. 5.

Vries, Hugo de 1901-1903 Die Mutationstheorie. 1905 Species and varieties.



atr., atrial orifice

hi., heart

emh., embryo

i.h.v., individual blood v

end., endostyle

int., intestine

/.p., foot-piece

oes., oesophagus

g., ganglion

or., oral orifice

gi., gill

ped., peduncle

gon., gonad

ph., pharynx

st.b.v., stolonic blood vessel




11 Cyclosalpa affinis Chamisso, solitary generation with chain of five wheels. Natural size.







\!^m ..^..^aiaJHBBI^^^^^^M






12 Cyclosalpa affinis, aggregate generation. X 1|.

13 Cyclosalpa affinis, solitary generation, witli young chain of zooids emerging. X 2.



1- niu



■^to ^^^











r \




/!/• \

\ Br



'\^ •'














Deploying point of chain of Cyelosalpa affinis.

14 Dorsal view.

15 Side view, left side.

16 Ventral view.

Zooids on the right side are nuinbereil 1', 2', 3', etc.; those on the left, 1,2, 3, etc. A given zooid has the same number in all three views.






(4 13 a II 10 4 6 H 10 9 8 "r




17-19 Chain of Cyclosalpa affinis

17 Ventral view of chain, showing unbroken part and four wheels. Somewhat diagramatic but drawn to scale. Natural size.

18 Dorsal view of same.

19 Peduncles of distal part of unbroken chain and of first two wheels.

20 Peneroplis arietinus Batsch, sp. Longitudinal section through the shell. Taken from Brady, Foraminifera, Challenger Expedition, vol. 9, plate 13, fig. 22.

21 Reophax membranaceus H. B. Brad}'. Taken from monograph of the Foraminifera of the North Pacific Ocean, Cushman, 1910, U. S. Nat. Museum Bulletin 71, p. 90, fig. 126.

22-25 Chain of Cyclosalpa affinis.

22 Enlarged view of the distal foot-pieces of the unbroken part of the chain.

23 Enlarged view of the foot-pieces of the first wheel.

24 Diagramatic representation of three stages in the development of the circulatory system of the chain.

25 Cross section through chain.







FrDiit Ihe Lnhoralories of Zoology and Experimental Therapeutics, rniversity of Chicago


Introduction -455

A method of measuring the rate of growth of rapidly growing ova 457

The rate of growth of the ovum of the common fowl 458

1. Large ova, more than 6.0 mm. in diameter 458

2. Small ova, less than 6.0 mm. in diameter 459

The thickness of the strata of white and yellow yolk in the egg of the common

fowl 461

The coincicience of the amount of yolk deposited in a day, with the amount

of yolk contained in a layer of white and yellow yolk 462

Yolk stratification in eggs of other animals as seen in the light of its causation

in birds 462

On the chemistry of white and yellow yolk 467

On the mechanism of yolk formation and de-formation 470

1. The part played by the reversible action of enzymes 471

2. The role of the partition coefficient of the elements of yolk 475

3. These two factors and the histological data 477

Summary 482

Literature cited 485


Very many thousands of pages have been written concerning yolk — its presence, formation, varieties and distribution in eggs. Indeed, the task of recording such a series of facts has been repeated on nearly every egg that has come under the closer observation of the biologist; while some eggs, notably those of the frog and the fowl, -have submitted their yolks to the observation and description of dozens of different investigators. Not JOURNAL OF MORPHOLOGY, VOI,. 22, NO. 2



withstanding this great amount of study and description, the literature fails to give satisfactory answer to any of the following questions: (1) Precisely how and where does yolk originate? (2) Why, or how is it that there are two kinds of yolk, (a) smaller spherules (often with enclosures) of white yolk, and (b) larger spherules of finely granular (often pigmented) yellow yolk? and, what is the relation between these? (3) What is the meaning of the stratified condition of the yolk of some eggs, eggs in which layers of white yolk alternate with layers of yellow yolk? (4) What are the chief chemical differences between these two kinds of yolk?

Thinking that we are now able positively to answer questions 3 and 4, and that these solutions bring some light upon the first and second questions, we submit the following data and considerations. These are presented with a minimum of reference to the enormous literature; otherwise this communication must have been increased to several times its actual size.

In carrying out this work, and now in the presentation of it, the author would say that he has not forgotten that 'yolk' is 'non-living substance' and therefore from a certain standpoint has but a minor interest to biologists. But, standpoints change. Until Johannes Miiller declared, and Van Beneden clinched the point, that the yolk of eggs is not living matter, and that it contrasts absolutely with the other part of the egg — the protoplasm — yolk had an all-absorbing interest to naturalists as a substance per se. In the years that followed, yolk has been studied largely with a view to cataloguing its diverse occurrences, forms, origins, distribution, tingibility, etc.; its interest to most students has flagged; though its often overweening bulk in the most studied of all cells has frequently won for it unwilling and tedious description. Perhaps one day we shall have a new standpoint. At any rate, we are only now beginning to realize that, though yolk is non-living substance, it is nevertheless organized substance and a very refined product of the vital laboratory; that it is a product laid down in the meshes of protoplasmic elements; and that the very act of its laying down is a signal of important metabolic states and capabilities of these living elements. More of the im


port of the relation between this 'organized' and this 'living' material we shall know later; in the meantime, each bit of information is doubly welcome because it concerns the most interesting form of protoplasm — the egg — at what is probably its most interesting period. Perhaps, then, the substance that has seemed to have but blundered in where it could blind us most, may itself prove to be a mirror for many a secret that we have elsewhere sought in vain.


The present studies began with an attempt to learn the cause of the stratified condition of the yolk of the hen's egg. It was suggested to me by results of an earlier study ('08) that the alternate layers of white and yellow yolk in the egg may be the result of the daily rhythm of nutrition — connected with high and low blood-pressure — which I had discovered in birds, and which I had shown to be the cause of the alternate fault-bars and fundamental bars of birds' feathers; it being there found that the daily variation in nutritive conditions in birds is sufficient to produce structurally perfect, and structurally imperfect parts in their rapidly growing feather germs. To test the suggestion, then, one might need only to learn the rate of growth of a bird's egg. What is the rate of growth in the eggs of the common fowl? This had not been determined, and no way of determining it was known.

It occurred to me that Sudan III might be used for this purpose. Knowing that Sudan was not destroyed in passing through the intestinal wall, (Daddi) that it circulated tied to the fatty acids of the food, and that the fatty acids of the food were laid down unchanged in the egg (Henriques and Hansen, '03), I inferred that Sudan given with fatty food would be laid down in the egg.

Moreover, it seemed possible by regulating the dosage and using proper intervals between feedings, to get laying hens to put this bright pigment down as definite bands in their growing ova, and thus enable one to determine the rate of growth.

The first experiment was as successful as the last. When such Sudan-containing eggs* were hard-boiled and sectioned under


water it was easy to measure the distance between the innermost borders of two such rings of Sudan, and thus to identify this amount of growth with the time which was known to have intervened between the two feedings.

Having thus discovered a method^ (described in detail by me elsewhere, '10) of measuring the rate and time of growth of ova, many data were collected on this point ; the distance between the normal strata (layers of white and yellow yolk) of the egg was carefully measured; later the problems and considerations growing out of the results were further followed up. We give here the following short statement of the observations and conclusions:

The radius of the hen's egg increases during the last few days of its growth by about 2.0 mm. per twenty-four hours. The thickness of a layer of white yolk and a layer of yellow yolk taken together is usually about 2.0 mm. Our conclusion is that in the fowl a layer of white and another of yellow yolk are laid down each twenty-four hours. Other facts at hand indicate that the yellow yolk is laid down under the best nutritive conditions, while the white yolk is a sort of growth-mark left by poorer nutritive conditions.

THE RATE OF GROWTH OF THE OVUM OF THE CO-MMON FOWL 1. Ova of more than 6.0 mm. in diameter

Table 1, section A, and plates 1 and 2 have been prepared to show the rate of growth of the larger ova as this is indicated by the Sudan method. The reader is referred to the table and plates in order to learn the kind of evidence on which the first conclusion is based. The amount of this evidence could be increased several times. It will be seen that the radius of the larger ova contained in the ovary of a fowl may increase by rather more than 2. mm. during twenty-four hours; also that this rate of growth is quite variable and may often fall to one-half the above amount.

Other data in our possession show that this rate not only varies for the eggs of different ovaries, but for different eggs of the same

' First announcement of the method, and of some of the present results, Riddle ('07).



ovary which were grown at different periods. It has been shown moreover by the Sudan-method that the rate of growth may be reduced not only to one-half that given above, but to absolute zero; this, however, is only a confirmation of what inference has long declared must be so, since ova may even decrease in size while in the ovary, i. e., they may be resorbed.


Showing under section A the rate of growth of hens' eggs as this was measured in central and peripheral parts of the yolk by means of Sudan. The numbers in the first column refer to the number which this egg bears in the plate. In section B are recorded measurements of the thickness of pairs of white and yellow yolk strata in central and peripheral regions of the ovum as this could be seen with unaided eye or with addition of iodine solution. The seven measurements here chosen arbitrarily from nearly forty in the records, are consecutive measurements of eggs from different hens.












5 PI.



1^ PI. 1







2.16 2.16






1.54 1.54










2.5 2.5






1.4 1.4






2.0 2.0





1.47 1.47


Average .




1.85 1.81

2. Ova of less that 6.0 mm. in diameter

It hafi not been possible to obtain a deposit of Sudan in eggs smaller than 6.0 mm. in diameter. This failure is explained by the fact that these ova are growing very slowly, as compared with the more advanced ova, and the intake of the stained food is here not rapid enough to give a perceptible effect. We shall see, moreover, that this white yolk — for ova of this size are com


posed entirely of white yolk — is much poorer in fat than is yellow yolk. Since fat is the only food that can carry Sudan this is another reason for the failure of Sudan to appear in them. The Sudan method is therefore not available for the determination of the rate of growth in these eggs.

One bit of evidence of another sort concerning this rate of growth was obtained and may be recorded. In fig. 3, pi. 2, is shown the striated appearance which the peripheral white yolk of one of the small eggs showed after having lain in a quantity of Mann's formalin-alcohol mixture for a few weeks. Here the noteworthy facts are, that a striation exists, and that the lamellae are not thicker than 0.25 mm. Whether these lamellae are made up of still smaller strata which really represent days of growth I am quite unable to say. I doubt somewhat that the radius of these small eggs is increased by as much as 0.25 mm. in twentyfour hours; any^^ay these strata offer some evidence — ^in the light of what we know of succeeding yolk strata — that these small eggs do not grow faster than 0.25 mm. per day.

One must ask what is the meaning of the extraordinary difference in growth-rate of eggs under, and over, 6.0 mm. in diameter? Wliat old mechanism is inhibited or what new one brought into action, that accounts for this procession of cells — each with months of slow and constant growth behind it — coming to a point from which each jumps in a day from its accustomed rate of increase, to a rate that is probably from eight to twenty times higher? Do the folHcular cells now become more permeaable than formerly to the ingredients of yolk? Is the increased vascularity of the follicular envelopes, that certainly occurs at this time, a cause or a result of the new activity? To these questions there comes no answer. But to us there are few events in the history of the primary oocyte of the fowl more interesting than this one. All the more interesting it is, too, because of its glaring apparent teleology. Here is an ovum within five to eight days of extrusion^ a ad containing less than the hundredth part

^ It is true, however, that if the yolk grow less rapidly than normally the egg remains longer in the follicle; showing that the time of ovulation is not controlled by heredity but is governed quite completely by conditions.


of the yolk necessary to make it capable of producing an animal. Nevertheless five to eight days suffice to supply the missing ninety-nine parts.


The measurement of the thickness of a layer of yolk offers some difficulties and- can rarely be done directly on a single layer; the reasons being that one stratum merges very gradually into another and that the strata are often very indistinct. More frequently, though by no means in every egg, a series of wellmarked layers can be found and a measurement made over all; the number of strata — or rather of pairs of strata — may be easily counted. When the total measurement is divided by this number one obtains the thickness of a combined layer of white and yellow yolk.

The result of eight such measurements is recorded in section B of table 1. These are typical of nearly forty reliable measurements, and indicate a thickness of about 2.0 mm. (1.4 — 2.5) for a layer of white and yellow yolk combined.

The layers of yolk can sometimes be seen in the fresh eggs, proving that they are not artifacts; but for the purpose of measurement it is usually best to hard-boil them, and section (under water) from one side until the exact center of the egg is reached. Sometimes it will be found advantageous to put the egg thus prepared in weak iodine solution for a time. This treatment seems occasionally, though not always, to strengthen the contrast between the layers of white yolk and those of the yellow variety.

For reasons stated above it is impossible satisfactorily to measure the thickness of a layer of white yolk. It can be said with confidence, however, that this so-called layer has but a fraction of the thickness of the adjoined yellow layer. Perhaps one errs but little in saying that the former usually has from one-fourth to one-eighth the thickness of the latter.





A comparison of the two sections of table 1 shows quite convincingly, I think, that the figures, which in the one column indicate the amount of a day's growth, are of the same order of magnitude as those which in the other column indicate the thickness of a stratum of yolk. This fact, and another one, namely, that we know thai there exists in birds a daily nutrition rhythm capable of producing daily growth-marks in their rapidly growing feathers, convince us that a layer of white yolk and another of yellow yolk is laid down during each twenty -four hours.

The well-developed appearance of the yellow yolk, its large yolk-spherules audits much greater thickness than that of the white layer, all indicate, moreover, that this layer, like the broad fundamental bar of the feather, is grown under the best nutritive conditions; while the narrow layer of white yolk with its small spherules gives indication that it, like the fault-bar of the feather, is grown under poor nutritive conditions.

Since I have shown that the poor nutritive conditions which produce the fault-bar occur in the later hours of the night — 1 :00-5:00 A.M. — / consider it as practically certain that the white yolk of the ovum is produced at the same time, and that the yellow yolk is produced during all other hours of the day.

The layer of white yolk of the hen's egg is then a growth-mark left at the ever-changing boundary of the ovum; it represents the results of yolk formation under sub-optimal conditions. It is indeed incomplete, unfinished yolk, as is apparently indicated by the histological data already known, and by the chemical evidence which I shall present in another section.


With the story of the white and yellow yolk of the bird in mind it becomes most instructive to reexamine many of the peculiar types of yolk distribution which from time to time have been re


ported and figured by embryologists and cytologists; for now we can feel fairly sure that wherever we meet alternate layers of white and yellow yolk, such layers indicate just so many alternations of better and poorer nutritive conditions during the time these layers were being formed. The better and poorer nutritive conditions doubtless applying to the organism as a whole. ^

A zonal arrangement of yolk similar to that of the bird has been reported in at least four other groups, viz., turtles, lizards, skates, and m3rxinoids. Some yolk patterns are known which are not distinctly zonal but intermediate to it and the type of yolk arrangement which is usual in small eggs; these help to bring all yolk distribution under a single principle or set of principles.

In order to avoid much tedious description in the text, and also to present more clearly and accurately this part of the subject, I have prepared plate 3, which is to a large extent a reproduction of figures which are not new. To what is shown in the plate, and in the explanation which accompanies it, I here add the following:

In all ripe ova, as in all the growth stages during which yolk is being deposited in the ovum, a layer of yolk composed of very small spherules (white yolk) is to be found at the extreme periphery of the egg. If larger yolk spherules (yellow yolk) also occur, they occupy more central portions of the egg. There is, moreover, scarcely an exception to the rule that the germinal vesicle or egg-pronucleus is immediately surrounded by similar small spherules and not by large ones.

It seems also to be very generally true that in those ova in which considerable yolk is developed, and in which the germinal vesicle makes its way from the center to the periphery of the egg (or remains near one side of the cell) it leaves in its wake a cyhnder of white yolk to which in some cases has been given the name of Pander's nucleus.

All of these features are shown in eggs of such widely separated forms as the skate (fig. 6) the amphibian (fig. 5), the lizard (fig.

^ On the other hand, some eggs, e. g., those of the salmon, may undergo their chief growth at the expense of the somatic tissues and while no food whatever is being ingested. The conditions here, however, are essentially constant and therefore produce no stratification of the yolk.


8), the birds and at least in some mammals (fig. 3). These are the forms, too, which — with the exception of the mammal — ^in addition show a stratification of the main body of the yolk. Two other forms are known, the turtle and the cyclostome (Bdellostoma) in which the stratification and other features occur, as in the above mentioned eggs, except that no Pander's nucleus has been found.

How may we explain at one and the same time the essential similarity of the yolk distribution in eggs of widely separated forms, and the often essential dissimilarity of its distribution in the eggs of closely related species? There seems now no doubt that all can be accounted for when one knows two things : first, the length of the growth period; and, second the chief fluctuations in the nutrition of the animal during the growth period of the eggs.

Most ova have no stratification, then, because the yolk is grown in a short season — -the animal not being subjected to such severe alternations as winter and summer, while the process is going on; or, because the eggs remain very small and develop little yolk'; or, again, because some ova have the extraordinary capacity of growing at the expense of somatic tissues. In such cases fluctuations in the nutrition of the animal are of little moment to the egg; the latter being able to feed well at the expense of the organism as long as it continues to live.

When stratification is present, however, I believe this to be a positive declaration that nutritive fluctuations did occur in the organism, and the number of the strata to be a reliable index to the number of such fluctuations. The presence of yolk stratification in the eggs of an animal then is an invitation to the naturalist and physiologist to look for important nutritional variations in that animal.

Thus far definite causal and time relations between such stratification and nutritional fluctuation has been determined only for the bird. What this time period is in BdeUostoma we can now only conjecture; but the fact that in a mature specimen eggs of a wide range of size exist possibly argues that these eggs are several years in forming. The further fact, that the animals lose much blood and become much weakened at each yearly spawning


period, is significant in that here may be found the means of a nutritional depression which produces a layer of white yolk in all of the remaining eggs of the ovary. If this be the true explanation one can readily understand the lack of stratification in the eggs of the related Petromyzon (fig. 2) since this form spawns but once in a lifetime.

In the skate the main growth period of the oocyte is probably completed in less than a year. The nine or ten pairs of strata figured by Riickert (fig. 6) are probably produced at the rate of about one per month. Whether this refers merely to the number of times the animal has fed during this time, or otherwise, nothing seems to be known.

The amphibian egg has a short growth period, and derives its growth material too from substances stored in the body, and is thus independent of external food supply. Doubtless these facts — together with its usually moderate size — will account for the actual configuration of its yolk.

The eggs of two reptiles — turtle and lizard — show very evident, but dissimilar, yolk strata. What the time, or the nature of the nutritive fluctuations are, that may produce these strata in Lacerta, I can make no suggestion.

In the egg of the tortoise Munson ('04) seems not to have identified (fig. 1) the so-called inner and outer cytocoel as layers of white yolk. A study of my own preparations, however, convinces me that such is their nature and the term cytocoel therefore is unnecessary. The turtle's egg has then alternate layers of white and yellow yolk somewhat comparable to those of the bird.

1 have found indications of four pairs of such zones in some eggs; or rather, by comparing the strata of different eggs from the same animal I have found such indications. But I am not now sure that four such pairs exist, nor that only four exist. Certainly several very thin strata can sometimes be found within

2 mm. of the periphery of some ova.

One wonders much whether the well-marked innermost layers of the turtle's egg can be the indications of years of growth. Agassiz ('57) showed that these ova undergo their greatest growth in four interrupted stages extending over four years. Our predic


tion is that further examination of these yolks, by proper methods for differentiating the strata, will show four pairs of white and yellow zones, to correspond to four yearly periods; each year supplying a period of growth and of rest, or at least of more rapid, and of less rapid growth.

Of the mammal's egg shown in fig. 3 it can be said that the several conditions of its growth seem to be closely similar to those of the amphibian egg which it so much resembles. To be sure, this egg may not, like the amphibian, develop at the expense of substances stored in the body; but, so few eggs are here developing at one time that an adequate food supply is always assured.

We believe then that these data practically give answer to the very important question which has been so well put by Riickert ('99, p. 585):

Diese Uebereinstimmung des Selachier— speciell des Torpedo-Eies, mit dem Vogelei ist, wenn der Vergleich sich zunachst auch nur fur die grobere Structur durchfiihren lasst, immerhin eine auffallende Thatsache.

Es wiirde die Mtihe wert sein, bei einer erneuten Untersuchung der ohnedies seit vielen Jahren vernachlassigten Dotterentwickelung nach Anhaltspunkten zu suchen, ob die Aehnlichkeit nur dadurch hervorgerufen wird, dass die beiderlei Eier unter gleichen Bedingungen sich entwickeln, oder ob es sich um einen durch Vererbung auf das Vogelei iibertragenen Vorgang handelt; mit einem Wort, ob eine Analogie oder Homologie vorliegt. Im letzteren Falle wiirde sich der Schluss Ziehen lassen, dass das meroblastische Ei des Vogels resp. der Sauropsiden ein primar meroblastisches ist wie das Selachierei und das Saugetierei kein tertiar sondern ein sekundar holoblastisches wie das Amphibienei.

The similarity noted above of the amphibian and marsupial eggs is another case in point. My results indicate that the likeness of yolk distribution in these two eggs, and in those of selachian and bird cited by Riickert, does not rest on heredity in any narrow sense of the word, hut on the fact thai they develop under like conditions.


The conception of white yolk which arose from the preceding work was that such yolk is a halted, or intermediate stage, in the


development of yellow yolk. This same conception had been urged on histological grounds by several workers, though opposed by others. The chemistry of the two substances was then appealed to for further evidence of a sort which it alone could give.

An examination of the rather abundant literature on the chemistry of yolk showed that it contained none of the data which our problem required. Analyses of yellow yolk have indeed been made by Prout, by Gobley and bj^ Parke; but it was believed that the extraction methods of their time did not effect a complete separation of the fat from the other constituents of the yolk. These determinations have therefore been made anew. That such was really necessary may be indicated b}^ the fact that Parke ('67) extracted only 66.7 per cent of fat and phosphatids, whereas my analyses always yielded more than 70 per cent of these constituents. It was also imperative of course that results of analyses which were to be compared should be obtained by identical methods. Apparently no analysis of white yolk had been made, so that this had to be done.

Since, moreover, the metabolism of yellow yolk includes not only its formation but also its de-formation into absorbable constituents, it was considered necessary to take account of yolk in a late stage of such modification. Such yolk is met with in two rather different situations: Normally, the whole yolk of the egg (yellow yolk) is subjected during the incubation period to the digestive, i.e., disintegrative action of the embryonic tissues — entoderm and yolk sac. Under such modifying action does yellow yolk become more like white yolk, or does it become less like it? A similar digestive action occasionally overtakes an ovum in situ, i.e., while still in the ovary and surrounded by follicular cells. These are the so-called 'resorbed ova.' How does the yolk of such an ovum in an advanced stage of resorption compare with the yellow yolk which it was before the beginning of resorption? Has it become more like, or less like white yolk?

The complete results of my analyses with a consideration of their points of chemical interest, and an account of the preparation of materials, and of methods used, will be pubhshed elsewhere.


I may say here that the fat and phosphatid extractions were made with the methods recently discussed and described by Prof. Waldemar Koch, in whose laboratory these analyses have been made. At this time it seems most desirable to present only the amount and sort of data which is necessary to give a clear picture of the major differences between the two forms of yolk under consideration, and to answer the two questions just stated above.











63.2 '



49.2 45.7 40.7 36.8

20.9 15.3 15.9 11.1

0.6 2.0 2.4 3.4






38.7 43.5

1 = analysis of fresh egg-j-olk (yellow yolk) (17.670 gr.)

2 = analysis of a resorbed ovum (1.834 gr.)

3 = Average of three analyses of contents of (9) yolk sacs (18 da. inc.), (78.821 gr-)

4 = analysis of white yolk (6.019 gr.)

Table 2 has been so arranged as quickly and accurately to tell the story. Nos. 1, 2 and 4 are single and quite typical analyses. The several analyses of the yolk sac contents varied considerably, and therefore an average of three separate analyses of yolksacs of eighteen days incubation is here given in preference to a single analysis. The white yolk was taken from a great number of eggs under 6.0 mm. in diameter, the yolk being removed without carrying over any traces of the enveloping membranes.

The quantitative differences in each of these chief components of white and yellow yolk are remarkable. Quite as striking and conclusive, too, are the numbers which show that ivhen yellow yolk is subjected to digestive action, in either of the two situations named, each and every component approaches more nearly to the quantity characteristic ofiohite yolk.

It cannot be said, however, that these data conclusivelj^ answer the question we have raised as to whether white yolk is an inter


mediate stage in the formation and indeed of the de-formation (digestion), of yellow yolk; although they do strongly support that view. There seems to be an alternative, namely, that the figures under nos. 2 and 3 approach the composition of white yolk more and more, only because the amount of that sort of yolk originally present in the egg is not diminishing, or is diminishing but slowly, whereas the yellow yolk is here being digested very rapidly. For, it must be remembered that, although we are considering a mature hen's egg as our type of yellow yolk, it still contains white yolk in quantities not easy to estimate; though we are accustomed to think of this amount as small, probably between 5 and 15 per cent of the total.

Parallel to the chemical data are the histological conclusions that it is always white yolk and never yellow yolk that is found applied to a surface into which yolk is being ingested. This is true for the germinal disc of pre-embryonic stages, and for the advancing entoderm and yolk-sac of the embryo (Balfour, Agassiz). Virchow ('91, p. 105) however, questions the correctness of this statement. It is certainly almost always true for the nucleus, or germinal vesicle of the primary oocyte, a seemingly significant fact upon which I shall publish observations elsewhere. Our chemical data themselves show, however, that the alternative cannot be true unless there is several times as much white yolk in an egg as we have reason to believe exists there. In any event the certain and interesting fact remains that when the yolk complex of the hen's egg is subjected to digestive and absorptive processes, the fat and phosphatids digest and disappear much more rapidly than does the protein.


Having presented data to answer questions three and four of the introductory statement, we may now consider the first and second questions in the light of these results, and with the help of other facts. Precisel}^ how and where does yolk originate? Why or how is it that there are two forms of yolk; or, what is the relation between these?


I purpose to preface this inquiry with a statement of my two main conclusions, or theses. (1.) The formation and the de-formation of yolk are one and the same subject. The processes of building are also the processes of tearing-down; only an equilibrium changes. These two sister-subjects have, however, long paraded as independents. The formation of yolk has been considered a subject the investigation of which was connected with a wide variety of study such as the migration of fully-formed yolk granules from follicular cells into the ovum; the origin of yolk granules from migrated particles of the chromatin, or the nucleolus; or again their formation by the yolk nucleus, or by mitochondria, etc. On the other hand, when the other phase of yolk metabolism — its de-formation — was concerned, observers have been pretty generally satisfied to speak only of 'a digestion and ingestion of yolk.'

(2.) Given a region into which the elements of yolk — ivith their vast amount of potential energy — can go and can exist without undergoing oxidation, and yolk {or some of its elements) will there be increased or decreased in amount subject to an equilibrium which is a function of two 'factors; (a) the reversible action of enzymes and, (b) the partition coefficient of the elements of yolk. We do not state that all desirable proof of this thesis is at hand, but we do insist that a very considerable body of evidence supports it. Having been led to the formulation of this view, and to the acceptance of it to the fullest extent ourselves, we shall here outline the evidence which we believe will likewise commend it to others.

It is not necessary to discuss separately what we have called theses one and two. Both rest upon the question of the presence, the effectiveness, and the modus operandi of the two factors which we have proposed as the immediate agentsof yolk transformations ; whether such transformations be of growth or of 'digestion,' whether they be progressive or regressive in character. The discussion therefore hangs upon these factors and we shall consider them separately.

Before proceeding in this direction, however, it is well to be reminded that these theses are the physiological and explanatory counterpart of an histological dictum which in certain of its aspects has been for many years ably maintained by several noted


histologists ; but which has apparently not gained universal acceptance : A spherule of yellow yolk may arise from a spherule of white yolk; in the normal destruction and utilisation of the yellow spherule, a white spherule may be again produced.

1. The part played by the reversible action of enzymes

Kastle and Loevenhart ('00) proved the reversibility of the action of lipase — the enzyme concerned in the analysis and synthesis of fat ; and we have seen that fat is the chief constituent of yolk. Wohlgemuth ('05) demonstrated the presence of lipase in the yolk of the fowl's egg. It was shown by Henriques and Hansen ('03) that the fatty acids of the food, i.e., of foreign fat, are laid down as such in the hen's egg. Since we know that this fat did not originate within the egg; and, since we are assured that fat as such does not pass through living cells, but that it is previously split into alcohol and constituent fatty acids, we must believe that the foreign fat found by Henriques and Hansen was synthesized within the egg cell; or, that it was synthesized in the neighboring follicular cells and thrown from their inner margins into the egg. This last alternative is not true as will be pointed out later.

Thus we come by means of the above series of facts directly to the proof of the existence within the fowl's egg of the synthesis — one side of the enzyme action — of the most voluminous constituent of the yolk.

Has the existence of the splitting action of lipase in the egg also been demonstrated? I believe it has practically been so demstrated by Liebermann's ('88) determination that only the merest traces of free fatty acids are present in the fresh egg, whereas large amounts are present at seven and fourteen days of incubation. The existence of a splitting activity of lipase in the hen's egg is moreover a matter that probably no one will question. From these facts then I think it must be said that the reversible action of lipase within the hen's egg has been indirectly demonstrated.

In fact, one familiar with the picture presented by the deposit and absorption of the yolk of eggs, can but wonder that this picture has not been before specifically pointed out as an example^



a typical example — of the reversibility of lipase effecting speedy and rhythmic transformations. The example too, becomes of considerable zoological interest, since certainly nowhere else does this simple physiological principle have such a relation to interesting features of morphology as just here. For, not only does it in these cases often completely change the features of the egg-cell, but it results in a condition (telolecithal) which later gives direction to a host of events of early development — cleavage, gastrulation, etc. — which proceed from the egg.

When we have spoken above of proof of the synthetic and of the analytic action of lipase we mean, of course, proof that each of these reactions may predominate in the egg. The burden of our whole statement is that both sorts of reaction are going on simultaneously (since the reaction is a reversible one), but that the conditions in the egg are, as a matter of observed fact, shown to be such that during the growth period the synthetic reaction normally exceeds the analytic; and that during incubation the reverse is true.

We say nothing in this connection of the origin and disintegration of the proteins of the egg. This group does not furnish, at present, examples specially proved for conditions in the egg, as do the fats. The reversibility of proteolytic enzyme action has however been demonstrated.

With yolk-forming enzymes (lipase, etc.) accelerating a series of reversible reactions in an egg-cell in which traces of yolk have been deposited, what are the factors which favor each side of the reaction, and thus induce either an increase in the amount of yolk, or a decrease in the traces that already exist? We believe that for the ovum of the fowl which we shall more specifically consider, some of the factors effective elsewhere may be ignored.^ The daily temperature fluctuations, for example, are relatively slight, etc. There seems good reason to believe that the amount and pro ■* There are several other factors or conditions which possibly, even probably, play parts in the storage of facts, i. e., building of yolk, in the egg; most of these however are factors supplementary to those of distribution coefficient and enzyme reversibility; though some are not. Some such factors known to be effective in fat-storage elsewhere are: (1) quantity of lipase (Kastle and Loevenhart) ; (2) different species (?) of lipase (Hanriot) ; (3) alkalinity


portion of the reacting substances present is here, as elsewhere under these conditions, the factor that determines whether the amount of yolk shall from time to time increase, remain constant, or diminish. What then are the conditions in the fowl that would tend to modify the amount of these reacting substances in the egg?

In answer to this we revert to the facts forecasted at the beginning of this paper in regard to our own earlier demonstration ('08) of a daily rhythm of better and poorer nutrition in birds; which rhythms coincide with periods of higher and lower blood-pressure. It was there made certain that very rapidly growing organs (feather-germs) were usually unable to pass over the period of the nightly (1:00-5:00 a.m.) reduction in blood-pressure without showing defects; which defects were proved to be due to insufficient nutrition.

Now we think there is no doubt that these facts lead to an answer to the above question. The egg (like the feather germs) doubtless derives fewer nutritive particles from the blood at this time than during the rest of the day. Possibly, even probably, the low blood-pressure induces at this time feeble but effective currents of fluid from these cells towards the blood and lymph; for it is probable that under low-blood-pressure the volume of the blood tends to increase at the expense of the fluid of the tissues. At any rate it seems certain that at this time the intake of the food substances from the blood is reduced, with the result that the equilibrium of the reaction is shifted. Thus the morphological picture becomes changed. Now growth will proceed more slowly. It is now that the granules must remain small, and poor in fat. It is now that some of the larger yolk spherules (yellow yolk) may possibly suffer reduction to smaller spherules (white yolk) ;

(Hanriotj ; '4) presence of other bodies eg. lecithin (Hewlett); (5) reducing conditions, i. e., conditions favorable to the formation of fat from carbohydrate and protein by reduction. A further reason for only a mere mention of these factors here is that the data for the egg are at present too meagre.

The factors which have to do with the formation and storage of the protein constituents of yolk, and of their union with lipoids and fats to form yolk, are less known than those factors which involve fat metabolism only; therefore the latter only are treated here. Macallum ('91) has some interesting statements on related subjects, and further points out that similar processes occur ill the formation of yolk and in the production of pancreatic zymogen.



the former being robbed more rapidly of their fat than of their protein. Now a layer of white yolk is produced in the egg.

In fig. A is shown a diagrammatic representation of how these fluctuations in the quantity of food-products of a fatty nature in the blood-stream would effect changes in size in oil drops, if these latter were separated from the blood by thin and semi-permeable membranes — the conditions existing at the surface of an egg. Section A represents growing conditions — predominance of fat synthesis due to rapid ingress of the constitutents of fat. Section B


Text Figure A=1

Idealized representation of the relations of the periphery of a mature sauropidan egg to the blood and lymph. Follicular cells not shown; these considered pervious as vitelline membrane, or by their intercellular spaces offering free access of lymph to that membrane. A = optimum growth conditions. B = metabolism of an oil drop in equilibrium. C = impoverished blood bearing away elements of yolk, with extension of white yolk area at the expense of yellow yolk. b. cap. = blood capillar}'; v.m. = vitelline membrane; iv.y. = granules of white yolk; y.y. = granules of yellow yolk. See text .

stationary conditions; as much of fatty ingredients is being given off into the blood, as is being taken from it. Section C droplets reduced in size as a result of continued contact with a blood stream poor in fat.


2. The role of the partition coefficient of the elements of yolk

Of less importance than the reversible action of enzymes, but following upon it, is the distribution between the yolk'" and the blood of the soluble substances concerned in these reactions according to their relative solubilities in these two solvents. There can be no doubt that this distribution, or partition coefficient is a factor in determining the amount of soluble substance which comes from the blood, lymph, follicular cells, or vitelline membrane, to the periphery of the yolk, and vice versa. Such is a physical necessity. The constitutents of fat for example, quite certainly enter the egg in soluble form and must there be subject to the laws of solubility.

The chief thing incumbent upon us in this connection is to point out how this partition coefficient may act selectively in modifying the amount of the reacting substances in the egg; i. e., how this principle may contrive at one time to increase, and at another to decrease the quantity of yolk contained in the egg. Remembering that it is the amount of reacting substances present that decides whether yolk formation or yolk de-formation may occur, the answer can perhaps be more easily given in reference to fig. 1. Let this figure now represent the periphery of the ovum of a turtle, in contact with the lymph and blood streams. During the summer, when the constituents of fat are probably most abundant in the blood, some of these must, on account of their solubility, pass into the egg and there later be built into yolk; their former places being continually taken by new particles from the constant supply of the blood stream. Under these conditions yolk spherules grow, as is represented by section A. In winter, conditions become as in C. The turtle cannot now digest food (Riddle, '09). Its heart-beat and other activities, hoVever, require food for their continuance, and the blood becomes depleted of food. The reversible action of its yolk enzymes is not likewise suppressed, but these now as before set free soluble yolk

^ The word 'yolk' is here made to stand for the whole body of the egg cell. Perhaps egg-protoplasm, follicular cell, vitelline membrane, and yolk, should all be mentioned instead of 'yolk' alone.


constituents. For such to be set free now, however, is to leave the egg entirely; for now the distribution coefficient of each of such substances brings a portion of it into the blood or lymph; and here it is not allowed to accumulate — to saturate this solution and then cease to act, — but is taken up by other organs ; while the blood thus freed from traces of it continues to pick up more of such particles as it passes the ovum. Because of this principle then an ovum may not be able to hold all the yolk that it has once acquired. Apparently we can explain the broad zones of white yolk in the turtles in this way, and the known facts seem to require the mechanisms we have described.

Of course we do not mean to infer that no other factor than the two we are describing have to do with certain aspects of yolk metabolism. For example, these two m^y have little causative influence in deciding that very important matter as to when the rapid growth of the hen's yolk is to begin. Here lie mysteries perhaps of the folhcular cells, or something else, perhaps more distant from the point of actual yolk formation. We are dealing only with the immediate mechanism of yolk formation and deformation.

The possible role of lecithin in increasing the solubility of fatty acids and soap in the follicular cells and in the yolk is an attractive subject. Moore and Parker ('01) have shown how enormously the solubilities of these substances are increased by the addition of small amounts of lecithin and bile salts. I have ascertained the presence of lecithin in the follicular membranes, but as yet have not enough analyses for comparison to draw conclusions. I have determined also, as is indicated in Table II, that the lecithin content of the white yolk — i. e., the layer just beneath follicular membrane, and usually between it and the yellow yolk — is smaller in amount than that of the yellow yolk.

As a concluding word on the role of the partition coefficient we record our belief that it alone accounts for the presence of the yolk coloring matters — vitello-lutein and vitello-rubin — in the yellow yolk, and not in the white. These are lipochrome pigments, soluble only in fat and fat solvents, and are abundant in the large yolk spherules, probably because, as we have shown by compara


tive analyses, these spherules abound in fat. Of similar intei-est is the discovery of Miescher ('97) that at the time of the development of the eggs of the salmon the blood of thfese animals is unusually rich in lecithin, fat and globulins.

3. These two factors and the histological data

One hardl}^ has a right to mention the words 'histology of yolk' without entering upon the consideration of an enormous literature. Since my own contribution is not primarily of histological nature, and for reasons stated at the outset, I refrain from doing so, although by my results I am seeking to put some rather new and additional interpretations upon histological conclusions, and to answer some questions in which histological, and to a less extent microchemical methods have before been largely used.

The view that intermediate forms of spherules exist, connecting white yolk with the yellow yolk spheres, has been maintained by Riickert, Sarasin, Disse, Kolliker and others. The region under the germinal disc of avian, reptilian and selachian ova have furnished the most and the clearest pictures of the transition forms. Previous authors have, however, generally considered only the formation of the yellow from the white spheres during growth, and have not considered the reverse of this process as it occurs during the destruction of the yolk. The engulfing of whole granules of (white) yolk by the entodermal cells has been recorded by His ('00) and others. This I would observe is, if true, not a real contradiction of my thesis, since these granules doubtless later undergo the ordinary processes of digestion in the entodermal cell. Similarly I would note that the presence of yolk granules in follicular cells — demonstrated by many observers — only illustrates the mechanism we have described at work in another cell; the classic example of this sort of formation being the fat globule in the cell of the intestinal mucosa. On the other hand the finding of such granules in a follicular cell is no guarantee whatever that the granule is thrown as such from that cell into the egg. The granule may here, as in the mucosa cell, again undergo digestion and pass from it in solution.


As regards yolk formation in insects, conditions are peculiar; the nurse cell seems here largely to carry out the work of yolk formation; while certainly the de-formation process normally is carried out by the egg only.

All observers agree that the outermost layer of yolk in any egg or growing oocyte consists of finely granular yolk. If this were otherwise our general theory of yolk formation would be untenable. Sarasin ('83) was led to the odd idea that the zones of the Lacerta egg were developed outermost first, and the central ones last. I think our demonstration of the nature of this growth in the bird, and the considerations that have followed, will convince that Sarasin's view is untenable.

Yolk spherules have been seen to grow after the egg leaves the ovary by Agassiz ('59), Van der Stricht ('07) and others. This growth is quite surely due to the spherule taking up by osmosis water-particles from the albumen or other fluid encountered by the egg; such growth is not of the nature we have described, though neither of the above mentioned authors has made the distinction.

In regard to the conclusion of many cytologists that yolk arises from the egg-nucleus, and of still others that it arises from the follicular cell nuclei, or from these cells in toto, I may append the following to show that we can exclude all of these as inadequate in the case of the yellow yolk of the fowl's egg. I have calculated that during the last day that a hen's ovum remains in the ovary it may deposit more than 5000 cubic mm. of yolk! Evidently too much work for an egg-nucleus. Again, since the radius of such an egg is increased by 2.0 mm. per day, this means that if yolk formation be a function of the follicular cell, each such cell must here produce daily a column of yolk 2.0 mm. long and of the diameter of the cell ; that is to say each such cell must form more than 50 times its volume of yolk per day, or more than twice its volume per hour ! Evidently too much vicarious labor for a cell.

It appears then that an exclusive origin of yolk from the nucleus, or within the follicular cells is impossible in the birds. The quantities of yolk laid down daily are amounts compatible with substances undergoing physical translocation by osmosis, solution.


etc., but not compatible with the probable rate of organic synthesis in the restricted regions of either the nucleus or follicular cell. It is of course necessary for all of the material entering into yolkformation to pass through or between the follicular cells; but each particle of this material may have, by undergoing the synthesis in situ in the egg, twenty-four hours or longer to accomplish this; whereas we have seen that if it originated within the follicle each cell would there have to organize completely its own volume of yolk material and empty itself of this more or less solid material at least once in each twenty or thirty minutes of the day.

Since such theories of yolk formation as have been proposed are now shown to be inadequate in a case where a test can be applied, and since it seems clear that the mechanism of yolk building which we have here outlined and described is necessarily present wherever and whenever yolk is formed, there is at present no valid reason for believing that any dissimilar method of yolk formation exists.

In a certain sense, no general theory of yolk formation has as yet been stated. That is to say, no outline of the processes involved in yolk-building and of the conditions affecting these processes has been attempted, and bur own effort leaves at least important chemical phases of the problem quite untouched. Previous efforts have been largely devoted to features of the histogenesis of yolk granules, and to the identification of some cell organ as the directive agent of yolk formation. Thus such cell structures as centrosome, nucleus, chromatin, nucleolus, mitochondria, yolk-nucleus, etc., have each been several times proposed as the seat or source of yolk. Whilst for some eggs, particularly those in which all of the yolk plainly could not have so circumscribed an origin, the seat or source of such yolk was centered upon a similar structure of the follicular cell; yolk particles have sometimes been described as arising in such cell and later making their way through the follicle cell membrane, vitelline, or other egg membranes, into the periphery of the egg. But theory usually has extended only to the matter of the source of yolk, to the relation between the white and yellow granules, or to the designation of one or another cell-organ as the directive agent of yolk forma


tion. There has been no theory to cover the long series of points involved, some of which are the following: WTiat are the conditions which permit yolk to form? In what situations and about what structures does it form (this point much studied and discussed)? What are the processes involved, — what is the mechanism of yolk formation? How are the different forms of yolk genetically and chemically related? How account for the variable amount, distribution, and stratification of yolk?

The statements concerning cell-organs as directive agents of yolk-building have often been quite misleading. This could hardly be otherwise since we have had here attempts to 'explain' a process, not in terms of other processes, but in terms of structure — an error not uncommon even in modern biology. One gets the idea from some descriptions of yolk formation that the nucleus is the absolute, immediate and ultimate source of yolk; and this in spite of the fact that yolk is never present within the nucleus, but onh' outside of it. Just how a vanishingly small fragment of chromatin, thrust from nucleus into cytoplasm — i.e., into an environment so new as to imperil its own existence, — may guide and direct the very rapid production of a thousand times its own volume of yolk (a new and very different substance from itself) we have not been told. Much apparently has been left to the imagination of the reader who is evidently expected to bridge for himself the gap that exists between the chromatin particle in situ and the yolk building process in operatio. But, the high regard which some adherents of this theory have for the kingly chromatin evidently persuades them that chromatin particles — which certainly are thrown from the nucleus into the cytoplasm, and about which traces of yolk certainly are sometimes found — comprise material of such superior quality that the base and foreign matter which meets their Midas-like touch must turn at once into golden yolk! By other workers mitochondria, and still other structures, have been similarly endowed with what would seem to be wonderful and transforming power. The writer would not undervalue the great amount of very valuable work that has led to the determination of the cell-elements about which yolk forms. But it seems to him that much less valuable than this painstaking work


are the inferences that have too often accompanied it to the effect that the structure about which yolk is found to form, is itself the active agent in the yolk formation. From the facts already brought forward we see that whatever the out-wandering chromatin particle — the invisible id or biophor — may be able to accomphsh in directing the course of differentiation in the highly complex Uving cytoplasm, the building of a single inert yolk granule by a plainly visible amount of chromatin is a task which clearly quite surpasses it! At any rate a task which it does not accomplish.

Yolk formation as it is indicated by the facts presented in this paper may be connectedly outlined as follows: Yolk will be formed (1) when conditions are such in the egg, follicular cell, food-supply, or organism that excess of food may enter the egg; but (2) in those regions only where some excess of food fat (and protein) can exist without undergoing oxidation; (3) the maintenance of such excess of food is dependent upon the amount of food, or upon marked fluctuations in the amount of food outside the egg, and (4) upon the distribution coefficient of the elements of yolk in the substance inside and immediately outside of the egg, and doubtless by other undetermined conditions within the egg; (5) the actual and active processes of yolk increase or decrease are essentially identified with the partially known synthetic and analytic, — i.e., reversible-action of the enzymes which act upon the constituents of yolk; (6) in the first stages of the growth of a (white yolk) yolk spherule the proportion of fats and phosphatids in its composition is small; (7) in later and more complete (yellow yolk) stages the proportion of these constituents is large.

In my opinion what we now most need to know is how those conditions arise which permit yolk-building to begin. We need further knowledge on*points (1) and (2) of the above. That is, we need to know why an unusual amount of food enters the egg at this particular time in its history. At present we do not know whether such cause lies inside or outside the egg. Again, what is the source of those reduction centers where foods which yield energy so easily as do the fats may not undergo oxidation but be built unchanged into yolk? It is possible, of course that nucleus,


chromatin, mitochondria, or centrosome, etc., of the egg, maj later be shown to have special causal significance with regard to such changes in amount of food-intake, or with the production of reducing centers, which we now recognize as basic and unknown features of the conditions which primarily initiate yolk formation. If so, then such cell-structure will have been shown to bear indirect causal relation to a result which it was formerly credited with 'causing' directly ; the test of this lies with the future. But, some at least, of the more direct and immediate features of yolkbuilding are quite certainly those which have been described in these pages.


1. A method of measuring the rate of growth of large, rapidly growing ova has been found. It consists in feeding, at known intervals, the fat stain Sudan III to animals developing such ova.

2. Ova of the common fowl smaller than 6 mm. in diameter grow extremely slowly as compared with ova of larger size.

3. The time interval between the beginning of rapid growth of the 6 mm. egg, and the breaking of the egg from the ovarian follicle (ovulation) is normally between five and eight days. In most cases it is either six or seven davs.

4. The radii of ova which are larger than 6 mm. usually increase nearly 2 mm. during each twenty-four hours.

5. The thickness of a layer of white yolk together with an adjacent layer of yellow yolk is nearly 2 mm.

6. A pair of such yolk layers is therefore produced during each twenty-four hours.

7. We conclude that the layer of white yolk in the hen's egg is laid down during poorer nutritive conditions obtaining in the later hours of the night (1-5 a.m.) and that the yellow yolk is deposited during the better growth conditions of the rest of the day.

8. Reasons are found for beheving that white yolk wherever found is but a stage in the formation, or the de-formation, of yellow yolk. That it remains as the final form of yolk, only where it is slowly grown or is halted by sub-optimal growth conditions.


Yellow yolk, on the other hand, probably indicates, wherever it is found in ova, rapid growth under better nutritive conditions.

9. The presence of alternating layers or zones of yolk (Schichtung) in the ova of some animals thus receives an explanation. A period of poor nutrition corresponds to each of such zones of white yolk; a period of better nutrition to each layer of yellow yolk.

10. The time of formation of a pair of such zones is known in the birds to be one day; in the turtle and myxinoid perhaps a year; in the skate possibly nearly a month; in the lizard this is quite unknown.

11. This 'Schichtung' of the yolk, and other peculiarities of yolk distribution, have produced great similarity in the gross morphology of eggs of widely separated forms, e. g., selachian and bird ; amphibian and marsupial. We can be confident that such similarities do not depend upon heredity in a strict sense, but upon the fact that these eggs have developed under like conditions.

12. The gross chemical composition of white and yellow yolk, and of yellow yolk undergoing de-formation or digestion (a) by the embryo and (6) by the follicular cells, have been determined and comparisons made.

13. White yolk contains much more water, proteid, and extractives, and much less fat and phosphatid than does yellow yolk.

14. When yellow yolk is digested, in either of the two situations named, its constituent parts are not digested, utilized, and absorbed at a uniform rate; but in such a way that the composition of what remains approaches the gross normal composition of white yolk. In such digestion fat and phosphatid are broken down more rapidly than is protein.

15. The immediate mechanism of yolk formation and of yolkde-formation are the same. Chiefly involved are two factors — not previously applied here — which we recognize as (a) the reversible action of enzymes, and {h) the partition coefficients of the several constituents of yolk.

16. The presence of the native lipochrome coloring matter — vitello-lutein — in the large spherules of yellow yolk only, is


probably due to the fact that these spherules contain much fat, and the lipochrome pigment is soluble in fat and fat solvents only.

17. The origin of the yolk of the fowl's egg from the nucleus of this cell, or from the nuclei of the follicular cells, is shown to be impossible. It is not probable that the essential features of yolk synthesis in any egg resides in either of these alleged sources.

18. An attempt is made to outline the processes involved in yolk formation.


Agassiz, L. and Clark, H. J. 1857 Contributions to the natural history of the United States, vol. 2.

Caldwell, W. H. 1887 On the embryology of monotremata and nmrsupialia. Phil. Trans. Roy. Soc, vol. 178.

Dean, B. 1899 On the embryology of Bdellostoma stouti: Festschrift f. V. Kuppfer.

Henriques, V and Hansen, C. 1903 Ueber den Uebergang des Nahrungsfettes in das Hiihnerei, und uber die Fettsaure des Lecithins. Skand. Arch, f. Physiologic, vol. 14.

Herfort, K. V. 1900 Der Reifung und Befruchtung des Eies Petromyzon fiuviatilis. Arch. f. Anat. u. Entwick., vol. 57.

His, W. 1900 Lecithoblast und Angioblast der Wirbeltiere. Histogenetische Studien. Abhdl. der math-phys. Klasse d. konigl. sachs. Gesellsch. d. Wiss. Leipzig.

Kastle J. H. and Loevenhart, A. S. 1900 On lipase, the fat-splitting enzyme, and the reversibility of its action. Amer. Chem. Jour., vol. 24.

Liebermann, L. 1888 Embryochemische Untersuchungen. Pflugers Archiv, vol. 43.

Miescher, F. 1897 Histochemische, physiologische Arbeiten. vol. 1, Leipzig.

Moore, B and Parker, W. H. 1901 On the functions of bile as a solvent. Proc. Roy. Soc. Lond., vol. 68.

Munson, J. p. 1904 Researches on the oogenesis of the tortoise, Clemmys marmorata. Amer. Jour. Anat., vol. 3.

Parke, J. L. 1867 Ueber die chemische Constitution des Eidotters. Med. -chem. Untersuchungen f. Hoppe-Seyler, heft. 2.


Riddle, O. 1907 The rate of growth of the egg-yolk of the chick, and the significance of white and yellow yolk in vertebrate ova. Paper before Amer. Soc. Zool., Chicago. Abstract in Science N. S. vol. 27, 1908, p. 945.

1908 The genesis of fault-bars in feathers and the cause of alternation of light and dark fundamental bars. Biol. Bull., vol. 14.

1909 The rate of digestion in cold-blooded vertebrates: the influence of season and temperature. Amer. Jour. Physiol., vol. 24.

1910 Studies with Sudan III in metabolism and inheritance. Jour. Exp. Zool., vol. 8.

RtJcKERT, J. 1899 Die erste Entwickelung des Eies der Elasmobranchier. Festsch. f. v. Kuppfer.

Sarasin, C. F. 1883 Reifung und Furchung der Reptilieneier. Arb. aus d. zool. Inst. Wiirzburg, vol. 6.

Sarasin, P. und C. F. 1887 Zur Entwick. und Anat. d. Ichthyophis glutinosa. Ergeb. naturw. Forsch. auf Ceylon, vol. 2, Wiesbaden.

Van der Stricht, O. 1907 La vitellogenese et la deutoplasma de I'oeuf de chauvesouris. Comptes rendus de 1' Assoc. Anat. Lille.

Virchow, H. 1891 Der Dottersack des Hiihnes. Festsch. R. Virchow., vol. 1.

Wohlgemuth, J. 1905 Ueber den Sitz der Ferment in HUhrerei. Zeitsch. f. physiol. Chem., vol. 44.



All figures natural size. 1-6 and la-6a represent series of eggs grown simultaneously in two Sudan-fed hens. About 20 milligrams of Sudan fed to each hen at 2 P.M., January 27, and at 10 a.m., January 30 (sixty-eight hours).

The bird bearing series la-6a killed February 2, 10 a.m. (70 hours after last Sudan began to be deposited in yolk).

1 Egg laid January 27, with pear-shaped, more solid 'waxy' interior; also two prominent circles of 'modified yolk' near periphery.

2 Egg laid January 29. The outer border line here represents Sudan deposited from feeding of January 27. This layer was 1 mm. in thickness. The two circles of 'modified yolk' showing here as in figs. 1, 3 and 4. The size of each of the yolks at the time of the modification is indicated by these circles.

3 Egg laid January 31; see above.

4 Egg laid February 2. Two layers of Sudan. The time between Sudan feedings was sixty-eight hours; the amount of yolk deposited in this egg during that time was 6.2 mm. = 2. 2. mm. in twenty-four hours. Section nearly in plane of germ.

5 Egg laid February 4. The two layers of Sudan here as in fig. 4, were 6.2 mm. apart = growth of 2.2 mm. in twenty-four hours. Section at right angles to plane of germ.

6 Egg laid February 7. Shows spreading, or dilation, effects in Sudan layers. Apparently the 'spreading' is mostly outwards, though this figure well represents neither the position nor the condition of each layer. This effect noted in eggs that have remained long in ovary, or, as in this case, in laboratory at high temperatures.

la Egg laid January 27. To unaided eye the outer 10 mm. of one side of this egg showed very plainly six pairs of yolk layers = 1.67 mm. each. Interior contained somewhat solid, waxy body 15 X 10 mm.

2a Egg laid January 29. Seven very distinct pairs of layers of white and yellow yolk. The white yolk represented by dotted lines; the yellow yolk by the spaces between these. Probably another layer central to those indicated in figure.

3a Egg 40 X 27 X 28 taken from oviduct (see first statement above). Inner borders of two Sudan layers are 4 mm. apart = 1.41 mm. growth per twentyfour hours. Section through plane of germ.

4a Egg 31 X 26 X 28 from ovary (February 2). Distance between inner borders ofSudanlines = 3.6mm. = 1.3 mm. in twenty-four hours. Section nearly in plane of germ. Figs. 3a and 4a show the spreading or diffusion of Sudan in the region of the germ. Distance between inner border of outer Sudan layer and periphery = 5. mm. ; this growth of seventy hours = growth of 1.7 mm. in twenty-four hours.

5a Egg 24 X 21 X 20, from ovary (February 2). The first feeding of Sudan (January 27) left but faint traces of the dye in this small egg (see small crescent). By mistake the inner border of the thick layer of Sudan of this figure was placed 7 mm. from periphery instead of 5 as it actually was. This 5 mm. = growth of seventy hours = 1.7 mm. growth in twenty-four hours.

6a Egg 15 X 16 X 19, from ovary (February 2). Smallest egg in which trace of Sudan was present; this somewhat diffuse and indicated by dotted circle. From inner border of circle to periphery = 4.8 mm. the growth of seventy hours = 1.64 mm. growth in twenty-four hours.

All eggs are drawn as perfect circles, although, as is indicated above, the boiled yolks quite constantly show three unequal diameters.













Figures all drawn twice natural size, and then reduced one-sixth.

1 Section at right angles to germ of an egg( 32 X 30 X 28) that showed with schematic clearness its rate of growth. The Sudan deposited in the first (inner) broken line was fed forty-eight hours before a following feeding. Thereafter the feedings of the dye were made at thirty-six hour intervals. A growth of 1.8 mm. per twenty-four hours is indicated for the three intervals of thirty-six hours.

2 Through gei'm of egg, 33 X 29 X 27. Shows well the manner of deposit of dye (and therefore of yolk) in immediate region of germ. The fan-shaped figure of dilute Sudan lying deeply beneath the germ is perhaps however not a correct picture of the original position of the dye, but a 'spreading' effect. The first two feedings of Sudan were here twenty-four hours apart; the next forty-eight hours and the last, twenty-four hours later. A growth of 1.5 mm. per twenty-four hours is here indicated for time between first and last feedings.

3 A series of small ova (4-7.5 mm. in diameter) from the ovary of a laying hen which had been once fed about 25 milligrams of Sudan and killed on following day. Only three of these ova (6.5 to 7.5 mm.) showed any trace of the dye. In the drawing ofthe egg of 7.5mm. the position of the layer of dye was placed two mm. too near center of egg. The number along side each ovum indicates its actual diameter in millimeters. The egg of 7 mm. was of special interest. After lying in a quantity of Mann's balanced formalin-alcohol solution for a few weeks the striated appearance of the outer portion of its white yolk was visible with binocular. The lower right hand figure is an attempt to represent the structure of the outer 1.75 mm. of the 7 mm. egg. Eight striae could be here distinguished. Apparently therefore the striae have a thickness of about 0.25 mm. The stain was found to be confined to the large, yellow yolk granules.

4 Egg, 34 X 29 X 31, through plane of germ. Central heavy irregular lines are from Sudan feeding three days apart. Bird was laying at long intervals and next feeding delayed ten days and following this two feedings two days apart. This egg therefore at least seventeen days in developing. But its abnormality is evidenced by the crumpled appearance of the innermost layers of Sudan (the corners of innermost layer are too sharp in drawing) and by a curious depression of the germ. Another peculiarity of this egg was its presentation of a brightly Sudancolored vegetative pole (the stratum of dye here near surface), and a normally colored animal pole (except for the small depressed region of germ).

5 Later egg, sister to no. 4. Much more rapid growth than in 4. First two feedings thirty-six hours apart; others twenty-four. A growth of 1.3 mm. per twentyfour hours is here indicated.






»? >




yolk gran. stain


•white yolk




The distribution of white and yellow yolk in the ova of vertebrata with special reference to zonal or lamellar formation.

1 Egg of tortoise, Clemmys (from Munson, '04), c.c. = cytocenter (centrosphere); y.c. = inner cytocoel of Munson, and what I should call inner or first layer of white yolk; i.y.l. = inner yolk layer of Munson, and inner or first yellow yolk layer I have called it: = outer cytocoel or second layer of white yolk; 0. y. I. = outer yolk layer, or second layer of yellow yolk; s. c. I. = subcuticular layer.

2 Mature egg of Petromyzon (from Herfort, '00). Very small granules in the external ooplasm, gradually merging into the large granules and large vacuoles of internal ooplasm.

3 Nearly grown egg of Phascolarctus (marsupial) (from Caldwell, '87). The darker crescentic body is coarsely granular yellow yolk; the clear area around the nucleus, which is also continued around the periphery of the entire egg is of finely granular white yolk.

4 One end of large (3 cm. long) egg of Bdellostoma (cyclostome) to show stratification of its yolk (from Dean, '99). The fine curved lines represent points richest in minute yolk-spherules (white yolk).

5 Mature egg of Ichthyophis glutinosa (amphibia) (from the Sarasins, '87) 6X9 mm. with central 'latebra' of white yolk; this connects above with the germinal vesicle, forming a nucleus of Pander beneath the latter.

6 Mature egg of Torpedo (from Riickert '99) in meridional section. The lenslike germ above. A central 'latebra' without stratification (Riickert says this is composed of dark, not light, substance). The dark layers are composed of loosely bound, but larger yolk platelets (white yolk?) ; the wider lighter strata of more closely packed but somewhat larger platelets (yellow yolk?).

7 Hen's egg photographed to show something of the concentric deposition of Sudan III. Dark lines = Sudan; these bright orange-red in original. The appearance here is very similar to the always less evident stratification of white and yellow yolk; the^narrow lines of Sudan in the photograph simulating the faint and narrow lines of the white yolk.

8 Part of immature egg of Lacerta (from Sarasin, '83) showing well-marked layers of white and yellow yolk (I infer that the dark lines represent white yolk) ; about one-fourth of egg is here shown. The germinal vesicle lies just outside the figure, above and to the right; all layers are seen to converge toward it, and to become gradually modified in its vicinity.




mm ,('■1(1 '^

/ ■ \






From the Zoological Laboratory, Syracuse University



Introduction 494

Material and methods 495

Observations 497

A Pennaria 497

1 Cleavage 499

2 Nuclear aspects 499

3 Amitosis 501

B Hydractinia echinata Flem 502

1 Cleavage 503

2 Ectosarcal features 508

3 Early embryo, moi'ula 508

4 Organization of the embryo 509

5 Entoderm formation 509

6 The larva, planula 510

C Clava leptostyla Ag 511

1 Maturation 511

2 Nuclear behavior 512

3 The chromatin 513

4 Nucleolar behavior 516

5 Later development 517

6 The morula 518

7 The germ layers 518

a. Ectoderm 519

b. Entoderm 520

Review and discussion 523

1 Origin and growth of germ-cells 524

2 Doctrines of homology 529

a The germ layers 531

b The planula 531

c The morula 532

d The blastocoel 534

e Cleavage homology 535

3 Amitosis 537

Summary 541

Bibliography 542





In the course of investigations carried on by the writer during several years, certain facts have come to light which seem to have important bearings upon several problems of general ontogeny. In various papers phases of these have been suggested, but only incidentally has any attempt been made to discuss their significance or their probable correlations as developmental phenomena. With further investigations still additional facts have been observed, and similar investigations by others have tended to convince me of their importance in a still larger degree. When the honor to cooperate in the preparation of this memorial volume was submitted, it seemed that no more appropriate subject came within the scope of the writer's researches than that involved or implied in the above caption.

My introduction to coelenterate morphology began many years ago with the problem of the origin of sex-cells, a subject at that time brilliantly exploited by Weismann, whose Entstehung der Sexualzellen bei den Hydromedusen, Zugleich ein Beitrag zur Kentniss des Baues und der Lebenserscheinungen" ('83), has long been a recognized classic in its line. It was ably supplemented by the hardly less brilliant researches of Metschnikoff ('86), " Embryologische Studien an Medusen. Ein Beitrag zur Genealogie der primitiv Organe."

The first contribution to the subject by the writer was a very brief and tentative paper before Section F, of the American Association for the Advancement of Science, in 1889. It was adversely commented upon by one who had accepted without question the then prevalent dogma that Hydrozoa were distinguished from all other Cnidaria by the origin of the sex-cells exclusively from the ectoderm. Under this adverse criticism no further utterance was made on the subject for several years, though there was no lapse of interest or investigation.

In the meantime, an observer here and there had dared to question the conclusiveness of the earlier dogma. Little by little facts were accumulating which cast further doubts upon the matter, and even compelled the conclusion that Weismann 's


fundamental contention was inconclusive. Results to be cited from various sources will tend to show that the early attempt to formulate a general theory of embryogeny on the basis of the origin of sex-cells was no less defective and inadequate than it was hasty. For some time past phases of my researches have forced the impression, which has deepened as the investigations have extended, that not a few of the earlier views as to coelenterate ontogeny were seriously defective, or absolutely in error at many points. Certain of these T have taken occasion to point out from time to time, as occasion arose. The purpose of the present paper is two-fold: First, to submit accounts of the development of several species of Hydromedusae which have been under investigation for some time; and secondly, to point out certain errors as to the ontogeny of the groups which, from various reasons, had become associated therewith.


1. Material. The material upon which the results herein described are based (with the exception of that of Pennaria australis, for which I am indebted to Mr. Edgar J. Bradley, of Australia, to whom my thanks are hereby acknowledged) was collected by the writer at various times within the past two years, and chiefly in the immediate vicinity of Woods Hole, though some of that of Clava was collected at Harpswell, Maine. It is a pleasure to express my thanks to the directors of these laboratories for various courtesies.

Attention will be given primarily to two species of Pennaria, and to a single species each of Clava and Hydractinia. Other species will be given attention in relation to the several problems with which the paper has to do.

2. Fixation. In my earlier work great difficulty was encountered in reference to killing and fixing reagents. For killing my first lots of eggs of Pennaria picro-nitric and picro-sulphuric solutions, then much in vogue, were used; but to my sorrow these were found to be almost worthless. This was particularly the case with picro-sulphuric. Almost the whole of one summer's


collection was absolutely worthless by reason of the almost exclusive use of this reagent.

Hermann's and Flemming's solutions afforded fairly good fixation, but subsequent staining was very difficult. Perenyi's solution was absolutely worthless with both Eudendrium and Pennaria material and has since been discarded. The only solution which gave reasonably good and fairly constant results was a strong solution of corrosive sublimate to which had been added 5 per cent of glacial acetic acid.

In later work I made use of various solutions of formaldehyde, but with only fair results. A 10 per cent solution in sea-water gave a good general fixation for immediate use. Combination with corrosive did not seem materially to better it. There was found also to be great variability in different species as to this matter. This was particularly apparent in eggs heavily yolk laden as compared with those in which yolk was lacking, or present in only small quantities. There was also great difference in later differentiating other cytoplasmic elements. For example, in the peculiar proteid granules present in eggs of Clava the first, and only satisfactory reagent was picro-acetic acid (p. 217, Biol. Bull, vol. 10, '06).

In 1906 my attention was directed to Bouin's picro-aceticformol. It was thoroughlj^ tested upon eggs of Pennaria and Hydractinia, and was found to be far superior to any thus far employed. I have since used Zenker's fluid with good results in fixation of eggs of several species. It is worth while to emphasize the importance of this feature of fixation, especially as it relates to coelenterate material. I have called attention to this in several previous papers, but it is absolutely imperative in order to warrant trustworthy results that particular attention be given to this matter.

3. Imbedding. In another respect I have learned to my cost the importance of prompt working up of coelenterate material after fixation. Attention was directed to this point in my paper on Pennaria ('04b, p. 455). This precaution has been abundantly confirmed by later experience, and I take occasion here to emphasize its importance once more. The value of this has been


vouched for by Smallwood ('09). My present method in this particular is to imbed the material in paraffine as early as possible after reasonable time has been given for proper hardening and dehydration. This imbedded preservation may apparently be indefinitely prolonged without detriment. But in my experience it is impossible to preserve material of this group for any considerable period in alcohol without having it suffer considerable deterioration. This is particularly the case with those cytologic factors of mitosis and allied features so important in modern problems of embryology.

4. Staining. This, like the matter of killing and preservation is one of much importance and of varying grades of difficulty, as it related to the problem under review. As in the preceding, I had long since called attention to the extreme difficulty in the staining reactions of coelenterate material. This was most marked, in my experience, in the eggs of Eudendrium and Pennaria. Others have also found similar difficulties with this phase of technique. G. T. Hargitt ('09, p. 163) has recently devoted some attention to the subject, and my own results have been confirmed by those described in his paper.

Difficulties experienced in my earlier work in Pennaria, and the later work on Clava, were such as to leave doubt, particularly in relation to the phenomena of maturation, leading me to conclusions, tentatively adduced, which subsequent work has not confirmed, as shown by G. T. Hargitt (op. cit.) and Smallwood ('09), and by facts herein described.


A . Pennaria

Except for additional facts which have come to light in relation to a species of Pennaria, the development of which has been hitherto unknown, no particular attention would be given to the subject in this connection. Since the issue of my detailed paper on the early development of Pennaria tiarella ('04), repeated observations on the living eggs have confirmed my previous


results in every detail, so far as the general facts are concerned. I think it may now be regarded as beyond doubt or cavil that these results, anomalous as they may appear, are absolutely normal and conclusive. Furthermore, when analogous cases to which I had directed attention, and others to be cited in a later connection, are taken into consideration, it seems rather strange that "early cleavage differing widely from what we have come to think as typical" should be given as adequate grounds for a reexamination of the case! However, when it is recalled that, with certain investigators, it is more important to reduce vital phenomena to a set of formulae, or to corral all development within a common law than to recognize facts as they are, the wonder is less strange than it might at first seem! But additional facts are now availaible from a most unexpected source, and of such character as to remove Siuy further grounds for question or doubt. Somewhat over a year ago I had the good fortune to receive from Mr. Edgar J. Bradley, of Adelaide, Australia, a collection of hydroids, and along with them several colonies of Pennaria australis Bale, together with the medusse and eggs, which had been taken in tow-nets just at the height of the breeding season. The only feature of regret as to the eggs is that they had not been preserved in other than weak formalin, in order to have made them available for cytological study. But, as it is, they show in surface study the external aspects of developmental behavior to such perfection as to leave little to be desired. Figures 5 to 8 are sketches of a few of these stages, which speak for themselves. As will be seen at a glance, they duplicate in a most striking way similar stages in the development of Pennaria tiarella. If one were to pass under review separate series of eggs of the two species, without pains to have critically determined them in advance, it would be practicalh' impossible to say which belonged to the one species and which to the other. There are the same ectosarcal features, — papillae, bridges, strands, etc., in both; the same bizarre, amoeboid characters, the same anomalous phases of cleavage, 'every egg a law unto itself, and finallj^ the same end resultant, a normal embryo. Later phases of development of the Australian species were not present, hence further compari


son was impossible, though there is no reason for doubt as to its subsequent similitude and results. A comparison with figures 1 to 4, of Pennaria tiarella, will make this more evident.

The fact that these eggs had been taken with the tow-net in the open harbor, and had been preserved shortly after in formalin, leaves no grounds for serious question as to their normal condition, and confirms completely the results of my own precautions ('04b, p. 474), to guard against possible effects of artificial conditions of the laboratorj^ These additional facts, together with others of like character which have since come to our knowledge, especially those described by Brooks and Rittenhouse ('07) must suffice once for all to establish the perfectly natural and normal phenomena of extremely erratic and indeterminate modes of cleavage and consequent organogeny.

1. Cleavage. There is nothing new to add concerning the cleavage features of the eggs of Pennaria tiarella. Concerning this feature in Pennaria australis little attempt will be made to give detailed descriptions. The figures cited will afford all that is necessar}'^ as to the general surface aspects. As already stated, there is such essential conformity in every respect to the corresponding stages in Pennaria tiarella that there seems small occasion to do more than refer to the figures and descriptions of the former paper ('04b). While the fixation does not give material fit for cytologic details, it is fairly good for general comparisons. Eggs carefully stained and cleared show fairlj^ well the general internal conditions, and here, as in the surface features, there is essential likeness to corresponding stages in Pennaria tiarella.

2. Nuclear aspects. Brief reference may be made to a fesv points under this head.

Fragmentation. In several of my earlier papers ('04b, pp. 460-1), attention was called to certain nuclear phenomena of a rather peculiar character. Among these was what seems to be a rather promiscuous dissolution, or disintegration of the nucleus and the dispersion of the greater portion of it into the cj^toplasm. To designate this process I used the term fragmentation, long previously employed to designate phases associated with direct nuclear division, and apparently first employed bj^ Van Beneden (Wilson, the Cell. p. 64).


In recent papers both Smallwood ('09) and G. T. Hargitt ('09, pp. 197-8), have expressed doubt as to the process in the eggs of Pennaria, the latter stating that "no sign of its fragmentation has ever been seen." But in the following sentence he adds, the supposed disappearance of the germinative vesicle at this time, I believe to be due simply to the usual dissolution of the nuclear membrane and the mingling of karyoplasm with cytoplasm."

Smallwood expresses similar doubt, saying:

If by fragmentation of the nucleus is meant that the entire nucleus disappears and its contents disperse throughout the cytoplasm, then I find no evidence of such a process in these hydroids. But what shall be said of the chromatin changes before maturation in Hydractinia and in Pennaria after maturation, where large quantities of chromatin migrate into the cytoplasm? (Op. cit., p. 228.)

It was chiefly in this latter sense that I had used the term, and observations of Coe, Lillie, and others were cited in support of facts found in Pennaria. It may also be admitted that there seemed to be evidences of the entire dispersal of nuclear substance through the cytoplasm and their subsequent reorganization into new nuclei. ('06, p. 227, etc.). Further reference to this will be made in another section.

Contention for fragmentation was based almost wholly on chromatin behavior. The facts which I urged in this connection were those involving, first, the enormous dissipation of chromatin and its absorption by the cytoplasm, during the phases of maturation; and secondly, the achromophilous condition of the chromatin at a slightly earlier time. These facts have not been disputed. Whether my inferences or interpretations are valid is quite another matter. As to that upon which I have laid most stress, viz., the disintegration and dispersal of a preponderating portion of the chromatin, certainly not less than 90 per cent in many cases, and that it has little or no subsequent function as chromatin, — I am still firmly convinced of its validity and of the vast significance it involves as to chromosome theory.

Concerning the achromophilous condition above referred to, I have little to add to my previous accounts. G. T. Hargitt


('09, p. 165), whose detailed experiments on differential staining have surpassed my own, was perplexed as to this condition. At the end of the growth period, the nuclear reticulum shows so little affinity for basic stains that there appears to be, so far as this test shows, no chromatin present in the entire nucleus. I can suggest no explanation for this peculiar condition of the chromatin at this period, but it is normal and characteristic of this stage." I am now convinced that there is here a chief ground for my failure to distinguish certain phases of maturation, and my subsequent error in the assumption of their possible suppression or modification in certain cases.

3. Amitosis. Concerning a further problem, that of amitosis, I am in doubt so far as Pennaria is concerned, even as at the time of my previous work. My chief grounds for this view are the facts first cited, and those of the multivesiculate aspects of the nucleus during cell proliferation. And here again Smallwood and G. T. Hargitt ('09), and later Beckwith ('09), all confirm my basis of facts. They find in these vesiculate nuclear conditions essentially the same results which are described in my accounts. Without exception their interpretations differ from mine. To them these facts are believed to be obscure phases, chiefly telophases, of mitosis. While I freely admit the force of their contentions, there are still good reasons for maintaining the plausibility of my own views and interpretations. This is especially the case concerning Eudendrium. Here there seemed to be clear and positive examples of amitosis, as shown in fig. 23, a and h, plate 15 ('04a). It may not be amiss to state here that all these examples of amitosis occurred in association with those 'nuclear nests' so intimately involved in the syncytial phase of development concerned in entoderm formation. The conditions are somewhat different in Pennaria, yet sufficiently similar to lead one to anticipate similar processes, and these appeared probable in the vesiculate 'nuclear nests' mentioned above. But in no case were there found the specific and positive examples figured in the case of Eudendrium. The same must also be said of Clava. But further discussion of this will be reserved for a later section.


B. Hydractinia echinata Flem.

During the summer of 1907 I was fortunate in securing large numbers of this hydroid in the height of its breeding season, and took occasion to study the development and life history of the species. Some account of the life history has already been given ('08) which obviates any call for emphasis here on this point. The early development was studied from living material during two summers and at the same time material was carefully preserved for cytological study. This latter was turned over to my colleague, Dr. Smallwood, and his results have already appeared ('09). It only remains for me to submit such accounts of my observations as seem important in order to afford a more or less complete and connected description of phases of development, especially when correlated with Smallwood's account referred to above.

There are numerous points of difference between my observations and those of Bunting ('94), some of which may be due to the fact that her studies were restricted to material obtained from the small colonies living upon shells occupied by hermit crabs, while my material was derived chiefly from colonies of enormous size, obtained from piles of docks or similar habitat, but with comparison from the former sorts. As pointed out in the paper referred to above ('08, p. 98), there is no adequate reason for regarding these hydroids as other than a single species, hence any differences to be cited must be incidental rather than fundamental.

One of the first points of difference to be noted is concerning the time at which the liberation of sexual products takes place. According to Bunting this is between the hours of 9:30 and 10:30 p. M. That it occurs during the night I have repeatedly demonstrated. Further, that it may occur in certain cases about the time stated by Miss Bunting, I have also found true. But that it may also occur at a much later hour, and also at varying hours, I have also found to be the case. Some of the best cleavage series obtained, especiallj' for the very early stages, were in the mornings from seven to nine o'clock. That is to say, the eggs had been deposited some time after midnight, and at the hours


named were in early stages (two- to eight-cell) of cleavage. This would seem strongly to indicate their deposition at perhaps five or six o'clock in the morning or thereabout, as recorded in my notes of July 11th and 12th. In other cases development had reached the morula stage at nine a. m., which would lead to the conclusion that liberation of sexual products had occurred about midnight. While it is true that in many hydroids the liberation of eggs and sperm occurs at a fairly constant time, yet there are others in which this is not the case, and in which such ripening and discharge is a more or less continuous process during the breeding period.

The character of the egg is much like that of Pennaria, though it is much smaller. Both are alike in general texture of protoplasm, contain yolk, and similar inclusions. There is present a pigment similar to that in the eggs of Pennaria, though less marked in color. Like those of the latter, the eggs are devoid of a definite membrane. They are rather heavy and sink promptly when set free. By reason of this it was practicable to suspend colonies in shallow vessels within wire baskets under docks in freely circulating water and with little liability of their being lost. This was a matter of some importance; for, despite the best precautions, these hydroids soon deteriorate in vitality under the artificial conditions of the laboratory, while by suspending them in open waters about the docks they thrive almost as if in the natural habitat.

jf. Cleavage. So far as I am aware the only definite work on cleavage of Hydractinia is that of Bunting ('94). In this paper we have a characteristically symmetrical portrayal of the process. In general surface aspects it is represented as almost mathematical in its regularity and symmetry.

That the earlier cleavage phases in perhaps a majority of the eggs conform to this in greater or less degree is probably true. But that it represents with any degree of accuracy the average behavior of this phenomenon as a whole none who had carefully followed it could for a moment admit. It has been difficult to conceive how, except by a selective process, any such account could have been formulated. It is quite easy to see that by directing


attention only to eggs which exhibited the regulation aspects of cleavage, and disregarding, as abnormal, those of differing aspects, just such an account might easily have been made up; and this in all probability may have been the method followed.

It is not strange that under prevailing conceptions as to formulated 'laws of cleavage' this method might naturally have been adopted. In the case of Pennaria the present writer deliberately disregarded an entire batch of eggs which were so erratic in behavior as to suggest the probability of pathological conditions. But, by whatever method one may explain the miatter, certain it is that there is a measure of irregularity in a large proportion of the eggs of Hydractinia, especially after the third or fourth cleavage furrows, which at once takes them out of the usual category of geometrical order or s>^llmetry and puts them, if not in the Pennaria class of chaotic irregularity, at least consigns them to the category of the indeterminate and unsymmetrical.

However, it is not my purpose, in thus discrediting an account which gives so inadequate and misleading an impression, to goto a similar extreme in the other direction and convey the impression of predominantly erratic cleavage. On the contrary, let it be noted that in perhaps a majority of the eggs of Hydractinia echinata the cleavage, while seldom exhibiting an approach to geometric order or synrnietrj^, is yet more or less regular and orderly. In such cases cleavage begins, as usual, at the animal pole, cutting vertically downward, and generally divides the egg into symmetrical halves, which adhere to each other by a narrow band, or connective of cytoplasm at the lower pole. The second cleavage likewise may begin at the upper pole and at right angles to the previous division, or may begin at the center and work outward, thus dividing each half into symmetrical fourths, giving a fairly typical four-cell stage. The third cleavage, which is usually equatorial, often begins at the center and extends toward the periphery, a process more or less common in eggs of hydroids. The subsequent phases may continue more or less orderl}^ as in earlier stages, but often grow increasing!}^ irregular and independent, though resulting in a sjTametrical embryo. On the other hand, figs. 14 to 22, which are camera sketches of living eggs,


show liow strikingly irregular and unsymmetrical cleavage may be in eggs of a given lot, developing under identical conditions. But in