Not only do adult organisms have the power of regeneration, but embryos and larval forms possess the same power, and even portions of the segmenting, and also the unsegmented, egg may be able not only to continue their development, but in many cases to produce whole organisms. Haeckel observed in 1869-1870 that pieces of the ciliated larvÆ of certain medusÆ, and even pieces of the segmented egg, could produce whole organisms. The more recent experiments of PflÜger (’83) and of Roux (’83) on the frog’s egg mark, however, the beginning of a new epoch in embryological study. The explanation of this is to be found, I think, not only in the introduction of experimental methods, but also in the fact that PflÜger and Roux realized the important theoretical questions involved in their results.

PflÜger’s experiments were made by changing the conditions under which the egg develops in order to determine what factors control the development. Since these experiments were made with whole eggs, the problems of regeneration were not directly involved in his results, although his conclusions are of great importance in connection with questions concerning the regeneration of the egg. A part of Roux’s work dealt directly with the development of a new organism from a piece of the egg or of the embryo. Roux’s principal discovery[107] (’88) was that a half-embryo develops from either of the first two blastomeres of the frog’s egg, if the other blastomere has been injured or destroyed, but that subsequently the missing half of the embryo is “post-generated.” Roux was led to this experiment by his discovery that the plane of the first cleavage of the egg corresponds very often to the median plane of the body of the embryo.[108] This relation suggested that there might be some causal connection between the two phenomena in the sense that the first cleavage plane divides the material for the right side of the body from that of the left side. In a descriptive sense this would be, of course, true if the two planes do really correspond, and if there was no later shifting of material across the middle line, but whether the two phenomena are causally connected, or are merely due to a coincidence, could only be determined by further experiment. The observations themselves are not beyond question, for the two planes do not always coincide, and may be even ninety degrees apart. These cases of divergence were thought by Roux to be due to an unobserved shifting of the developing embryo, but it is improbable that all cases can be accounted for in this way.

Roux carried out his experiment by plunging a hot needle into one of the first two blastomeres, so that it is injured to such an extent that its development is prevented. The same needle, without heating again, was used for one or two other eggs, for, if the needle had been so hot in the first instance that both blastomeres had been injured by the heat, this might not happen in the second or the third egg. It was found that amongst the eggs that had been operated upon in this way, some had been so much injured that neither blastomere developed, others had been so little injured that both blastomeres developed, but in the successful operations the uninjured blastomere developed, while the injured one did not. In the last case the uninjured blastomere divided, and produced a large number of cells. A segmentation cavity was present in the upper part of the hemisphere (Fig. 61, A). The injured half remained in contact with the other, completing the sphere, but it did not segment. A half-embryo developed from the uninjured half, as shown in Fig. 61, B, C. This embryo has a half-medullary fold along the side in contact with the injured half. At the anterior end somewhat more than half a head is present, and at the posterior end there is a half-blastopore. The cross-sections[109] (Fig. 61, C), through the embryo, show that beneath the half-medullary fold a rod-like notochord is present, which is made up apparently of fewer cells than the normal notochord, but it has, in cross-section, a round and not a half form. At the side, the mesoderm is present, as in the normal embryo, and it has produced the characteristic mesoblastic somites. An archenteron is formed in the half-embryo, and, since it is smaller than the normal, it may, perhaps, be called a half-archenteron. The embryo is, therefore, in most respects a half-structure. The head is, however, nearly a whole head, but whether this is due to a whole head developing out of material derived entirely from one of the two blastomeres, or whether, as Roux supposes, a portion of the material of the injured blastomere has been worked over, i.e. “post-generated,” remains, I think, an open question.

The results of this experiment seem to confirm Roux’s conjecture that the material of each of the first two blastomeres is of such a sort that it gives rise to half the embryo, and, if so, there would be some probability that there is a causal connection between the first cleavage and the separating out of the parts of the embryo. In fact, Roux drew this conclusion, and even attempted to show how such a qualitative division is brought about. It should not be overlooked, however, that this conclusion goes beyond the legitimate bounds of deduction from the results, since the half-development takes place while the injured half retains its connection with the developing half, the former still remaining alive. On the other hand, the presence of the injured half makes the experiment more suitable to demonstrate that each of the first blastomeres gives rise, under normal circumstances, to half of the embryo. If one half had been removed, we can foresee that its absence might lead to other complications that would affect the result.

The most important outcome of this experiment is, I think, to show that a half-structure may develop by itself, i.e. that there is a certain amount of independent power of development in the parts of the egg.

Roux also tried to show that if, after the second cleavage has been completed, the two blastomeres that lie on opposite sides of the first cleavage plane are killed by a hot needle, the remaining two produce either an anterior or a posterior half of an embryo. An embryo derived from the two “anterior” blastomeres is represented in Fig. 61, D. The anterior half of the body is present. Posteriorly the half-embryo abuts against the injured half. It is possible, I think, that this embryo may represent the anterior half of a whole embryo of half size that has been prevented from closing in posteriorly by the mass of injured material of the undeveloped blastomere. Roux did not determine positively whether the two “posterior” blastomeres could give rise to posterior half-embryos; one embryo in his opinion appeared to bear out this interpretation. This part of Roux’s work is, it seems to me, not so satisfactory as the part dealing with the first two blastomeres, and we may leave it, for the present, out of the discussion, and consider only the result of the first experiment, in which one of the first two blastomeres was injured. Since the problems involved in the two cases are essentially the same, nothing will be lost by dealing with the first case alone.

The uninjured blastomere first gives rise to a half-embryo. After this has been accomplished, other changes take place that “reorganize,” according to Roux, the material of the injured half in such a way that the missing half of the embryo is formed by a process that Roux calls “post-generation.” This process can be studied only by means of sectioning the embryos, and since the eggs may be injured to a varying extent, there must be some uncertainty in making out the sequence of events. It is found that the yolk of the injured blastomere is vacuolated in places, and that the protoplasm in the path of the needle has been killed (Fig. 61, A). Irregular pieces of chromatin are found in the protoplasm, which seem to come from an irregular breaking up of the nucleus.

The changes that lead to the reorganization of the injured half may take place at different times in different eggs. Roux describes three kinds of reorganization phenomena. The first includes the formation of new cells in the injured half. Nuclei, surrounded by finely granular protoplasm, appear in the protoplasm of the injured blastomere. These nuclei arise from two sources: in part from the scattered chromatin of the injured blastomere itself, and in part from nuclei, or from cells without walls that have emigrated from the developing half. Around these nuclei, as centres, the protoplasm (with its contained yolk) of the injured half breaks up into cells. This cellulation of the yolk may take place in different eggs at different times. In some cases it may not have appeared as late as the gastrula stage of the uninjured half; in others, it may take place at the time when the uninjured half is segmenting.[110] The formation of the cells in the injured half begins always near the developing half, and extends thence into the injured parts. The new cells are of different sizes, but are larger than those of the uninjured half.

The cellulation of the yolk takes place only in the least injured parts of the protoplasm. Where the protoplasm and yolk have been much injured, they are changed over by the second method of reorganization. This part of the blastomere is either actually devoured by wandering cells, or is slowly changed under the influence of the neighboring cells, so that it becomes a part of these cells.

The surface of the injured half is covered over by ectoderm that grows directly from the developing half (third method of reorganization),—at least this happens where the protoplasm has been much injured. In other parts of the injured half the new cells that have appeared in this part, and that lie at the surface, become new ectoderm.

Post-generation now begins in the reorganized and cellulated half; the cells become changed over into the different layers and organs that make the new half-embryo. A few hours or a night is sometimes sufficient to change a hemi-embryo into a whole embryo. The new half-medullary fold develops from the new ectoderm to supplement the half already present. The mesoblast appears over the side. Its upper part seems to come from the uninjured mesoderm that has grown over to the other side, but this is added to at the free edge by cells that belong to the newly cellulated part. The new differentiation is, in general, in a dorso-ventral direction. The lacking half of the archenteron arises in connection with the half of the archenteron already present in the hemi-embryo. The yolk cells arrange themselves radially, and a split appears in the post-generated part, extending from the archenteron of the hemi-embryo. The split opens, and the new half-archenteron appears. In general, Roux states, the post-generation of the organs of the injured half proceeds from the already differentiated germ-layers of the hemi-embryo. The post-generation begins where the exposed surfaces of the germ-layers of the hemi-embryo touch the newly cellulated regions of the injured half.

It is most difficult to account for these post-generative changes, since the new part has, according to Roux, a double and even a three-fold origin. The pieces of the old nucleus, he admits, may take a part in the formation of the new cells; wandering cells migrate from the yolk mass of the old half into the new, and the cells of the formed germ-layers may be pushed over to the other side. Since a certain share, and perhaps a large share, of the new cells comes from the hemi-embryo, it is clear that, in addition to the power of self-differentiation shown by the uninjured blastomere, we must also ascribe to it certain regenerative powers, at least to the extent that each kind of cell that comes from it can give rise in the injured half to cells like itself, and produce similar structures in the other half.

If then, as Roux supposes, the development of the egg consists in an orderly, qualitative series of changes that lead to the subsequent differentiation, we must also suppose that the new parts are gifted with latent powers by virtue of which they can re-create all parts of the other half. Roux supposes, in fact, that each cell carries with it a sort of reserve-plasm, that is dormant in ordinary development but is awakened when any disturbance of the normal development takes place. Objections have been made to this subsidiary hypothesis, since the addition of this to the original assumption of a series of qualitative changes involves such complications that the view can hardly be considered a probable one. This objection is, I think, not as strong as certain critics believe, since the facts of development show beyond a doubt that although the egg has the power of progressive change it has also, as certain experiments show, the power of reorganization, if the ordinary course of events is interrupted. This admission by no means throws us back upon Roux’s hypothesis, for, as will be shown later, a different conception of the development may better account for both phenomena.

Inasmuch as a good deal of discussion has taken place in regard to the process of post-generation described by Roux, it should be stated that Endres and Walter reËxamined the process, and found, as had Roux, that the reorganizing cells migrate from the uninjured to the injured side, and around them the protoplasm of that side makes new cells. They found that the injured half is directly overgrown by the ectoderm from the developing half. When the material of the injured blastomere is only incompletely reorganized, there is formed, after post-generation, an embryo that has a protrusion of yolk in the dorsal part of the body. When the injured material is completely worked over, a perfectly formed embryo may result. The typical half-embryos that Roux obtained were also obtained by Endres and Walter. They deny that whole embryos develop from one of the first two blastomeres, as Hertwig affirms.

Hertwig repeated Roux’s experiment and obtained results entirely different from those of Roux. He injured one of the first two blastomeres of the frog’s egg with a hot needle, or by means of a galvanic current. Hertwig states that after the operation the egg turns so that the uninjured part lies uppermost. This is owing, he thinks, to the appearance of a blastula or of a gastrula cavity in the developing part. The segmentation cavity is found in many cases surrounded by the cells of the segmenting half (Fig. 62, A), but at other times at the border between the new and the old parts. In still other cases the cavity may lie eccentrically, and in some cases the floor of the cavity may be bounded by the yolk substance of the injured half. An embryo appears on the upper, uninjured part, though it is not, according to Hertwig, a half-embryo, but a whole embryo, or at least one approaching that condition (Fig. 62, D, E, F, G, H). It is shorter than the normal embryo, and its posterior end is incomplete. When these embryos are cut into sections, it is found that the part that has developed corresponds to the dorsal part of a normal embryo, but the ventral part is continuous with the yolk substance of the injured half (Fig. 62, B, C, J, K). Hertwig interprets these embryos as forms in which the yolk portion of the developing half, together with the whole of the injured blastomere, represents a yolk mass that has not yet been enclosed by the margin of the developing part.

In nearly all the embryos that Hertwig has described, the medullary folds appear eccentrically on the developing half (Fig. 62, D, F, K), and in some cases they may lie so far to one side that they are situated almost at the edge; and the less development of one of the folds makes the embryo appear almost like the hemi-embryos obtained by Roux. In fact, one embryo seems to have been a true hemi-embryo.

Hertwig attributes the eccentric position of the embryo to the eccentric position of the blastopore of an earlier stage, but he does not attempt to account for the eccentricity of the latter.

It is significant in this connection to find that Hertwig obtained other embryos that show a condition of “spina bifida.” In these there is an exposure of yolk in the mid-dorsal line between the halves of the medullary folds. Still other embryos in the same series of experiments were only slightly injured, and developed nearly normally. In these cases, Hertwig thinks, the blastomere that was stuck had been only slightly injured, and had partly developed. I have also often observed in this experiment that the injured blastomere may segment and add cells to the developing half, but in such cases the development of the injured half may be less regular than is that of the uninjured half. It seems to me not improbable that in several of the embryos described by Hertwig both blastomeres have taken part in the development. The main points of difference between the results of Roux and of Hertwig cannot, however, be explained in this way, and the explanation is to be found in another direction.

Hertwig emphasizes the view that the injured blastomere is not dead, but exerts an influence upon the other half—an influence of the same kind as that which the yolk of a meroblastic egg has on the protoplasmic portion of the egg from which the embryo arises. He ventured to prophesy that if the injured yolk mass could be entirely removed, the uninjured blastomere would produce a normal embryo without defect, and one like the normal embryo in every respect except in size.[111]x

Roux interprets Hertwig’s results as due to the sudden partial post-generation of a part of the injured half of the egg. He thinks that a half-embryo had first developed, and then to this there has been quickly added a part of the missing side. This reply fails, however, to meet Hertwig’s description of the method of development of the embryos. Later work, however, has put us in a position to give a more satisfactory account of the differences between the results of Roux and Hertwig. It seemed to me that the two kinds of embryos might be due to the different positions of the eggs after the operation. It had been shown by Schultze (’94) that if a normal egg in the two-celled stage is turned upside down and held in that position two embryos develop from the egg (Fig. 63, B, C, D). These embryos are united in various ways, and arise presumably one from each of the first two blastomeres. These results have been confirmed by Wetzel, who examined more fully into the early development of the twin embryos. He showed with much probability that the protoplasm rotates in each blastomere, so that in many cases the lighter part flows, or starts to flow, toward the upper hemisphere of the egg. In this way similar protoplasmic regions of the two blastomeres may become separated, and under these circumstances each blastomere gives rise to a whole embryo. A cross-section through one of the segmentation stages of one of these eggs is shown in Fig. 63, A. The smallest cells are found at the outer side of each half, and the two segmentation cavities lie one in the upper region of each hemisphere. Some of the different kinds of embryos that develop from inverted eggs are shown in Fig. 63, B, C, D. They are united in Fig. 63, B, by their ventral surfaces, and in Fig. 63, C, C¹, C², by their dorsal surfaces, and in Fig. 63, D, D¹, at the sides. These differences are probably accounted for by the different ways in which the protoplasm of the first two blastomeres rotated before the egg divided.

A consideration of these results led me to carry out the following experiment on eggs operated upon by Roux’s method. After sticking one of the first two blastomeres, some of the eggs were placed so that the uninjured blastomere kept its normal position, i.e. with the black hemisphere upward. Other eggs were turned, so that more or less of the white hemisphere was upward. From the two kinds of eggs two kinds of embryos were obtained. From those with the black hemisphere upward the embryo was a half-embryo like that described by Roux, while from the eggs with the white hemisphere upward embryos developed that were in many respects whole embryos of half size.[112] The explanation of this difference will be obvious from what has been said. When the black hemisphere is uppermost the contents of the uninjured blastomere remain as in the normal egg, and a half-embryo results. When the white hemisphere is uppermost the contents of the uninjured blastomere rotate, so that it generally shifts its relation to the protoplasm in the other injured half, and a whole embryo develops, as in Schultze’s experiment. In one case I obtained a half-embryo from an inverted egg. The result did not appear to be due to a lack of rotation of the protoplasm, because the medullary folds were white, showing that the protoplasm must have changed its position. The result can possibly be explained as due to the protoplasm rotating in each blastomere along the line between the halves, so that it still retains the same relation as that of the normal two-celled stage.

The whole embryos of half size are generally imperfect in certain respects on account of their union with the other half. They resemble in all important points the embryos described by Hertwig, and I see no grounds for interpreting them as embryos of a meroblastic type, but rather as whole embryos of half size, whose development posteriorly and ventrally has been delayed or interfered with by the presence of the other blastomere.

It has not been possible to separate the first two blastomeres of the frog’s egg, for if one is removed the other collapses. In the salamander, that has a mode of development similar to that of the frog,[113] it has been possible to separate the first two blastomeres. Herlitzka, who carried out this experiment, found that each blastomere gives rise to a perfect, whole embryo of half size. We cannot doubt, I think, that the same power of producing a whole embryo is also present in each of the first two blastomeres of the frog’s egg. When the two remain in contact in their normal relation to each other, each produces only a half; when like regions of the two blastomeres are separated, each produces a whole embryo. Thus we see that whatever the factors may be that determine the development of a single embryo from the egg, still each half, and perhaps each fourth also, has the power of producing a whole embryo.

In later papers Roux has stated that he had also, even in his earlier experiments, found other kinds of embryos than the half-embryos that he described. Some of these were whole embryos that had developed from the uninjured blastomere without the injured one taking any part or only a very small share in their formation. He found, he states, all stages between those embryos that had used up all the yolk material of the injured side (though post-generated) and those that had not used any part of it. The latter kind of embryo he does not recognize as a whole embryo of half size in the sense that a single blastomere has developed directly into a smaller whole embryo, but he believes that there must have been formed at first a half-blastula, half-gastrula, half-embryo, and that the last stage completed itself laterally without using any material from the injured half. That the uninjured blastomere may at first segment as a half is not improbable, but that whole embryos are formed only by the formation of new material at the side of a half-embryo is, I think, hardly possible, since the results of Schultze, Wetzel, Hertwig, and myself show that a whole embryo may develop directly out of the material of a single blastomere.

Spemann (1900) has carried out some novel experiments on the eggs of triton, and has shown how in another way double structures may be produced. If a ligature is tied loosely around the egg at the first cleavage exactly along the division plane between the first two blastomeres, it will be found later that the long axis of the single embryo lies, in the great majority of cases, across the ligature, and only in a small percentage of cases does the median plane correspond with that of the ligature, and, therefore, with the first cleavage plane.

If one of the latter eggs is allowed to develop to the blastula stage, and the ligature is then drawn tighter, so that the blastula is completely constricted, an embryo develops from each half.

If one of the former eggs is allowed to develop to a stage when the medullary plate is laid down, but is not yet sharply marked off, and the ligature is then tightened, there will be formed (the plane of constriction being across the medullary plate) from the anterior part a normal head with eyes, nasal pits, ears, and a piece of the notochord, and from the posterior part there will be formed, at its anterior end, another new head just behind the ligature. Ear-vesicles develop in this part at the typical distance from the anterior end. The brain that develops has a typical cervical curvature, and eye evaginations appear at the anterior end. The chorda, that extended at first to the anterior end of this region, is partially absorbed.

If the ligature is drawn tighter at a later stage, when, for instance, the medullary plate is plainly visible but is still wide open, a different result is obtained. The posterior part no longer forms a new head at its anterior end, but develops into those structures that it would form normally. In some cases it was found that the region from which the ear develops had been pinched in two, and in consequence a small vesicle appears in front of the constriction and another behind it.

In those cases in which the ligature lies in the median plane of the embryo, it is found that a double anterior end is produced. As the embryo develops it tends to elongate, and in consequence the material is pushed forward on each side of the ligature. A double head is the result. The extent of the doubling depends on the depth of the constriction between the halves. In the most extreme cases two complete heads are formed with an inner nasal pit, eye, and ear on each head, as well as the normal outer ones. The results show that even such complicated structures as the eyes and ears, etc., may arise from parts of the body where they never appear under normal conditions.

A series of experiments that have been made on the eggs of sea-urchins has led to equally important results. The earliest experiments are those of O. and R. Hertwig, who, in addition to studying the effect of different drugs on the developing egg, found that fragments of the eggs of sea-urchins, obtained by violently shaking the eggs in a small vial, could give rise, if they contained a nucleus, to small whole embryos. Boveri made the important discovery in 1889 that if a non-nucleated piece of the egg of the sea-urchin is entered by a single spermatozoon, the piece develops into a whole embryo of a size corresponding to that of the piece. Fiedler, in 1891, separated the first two blastomeres by means of a knife, and found that the isolated blastomere divides as a half, but he did not succeed in obtaining embryos from the halves. Driesch has made many experiments, beginning in 1891, with the eggs and embryos of the sea-urchin. He separated the first two blastomeres (’91) by means of Hertwig’s method of shaking the eggs, and studied the development of the isolated blastomeres. He found that the cleavage was strictly that of a half, and not like that of a whole egg. The normal egg divides into two, four, and eight equal parts. At the next division, four of the cells divide very unequally, producing four very small cells, the micromeres, at one pole. The four cells of the other hemisphere divide equally (Fig. 64, I). The isolated blastomere divides at first into two equal parts, then again into equal parts. At the next division two of the cells produce micromeres and two divide equally (Fig. 64, G). This is exactly what happens at this division in each half, if the blastomeres are not separated. In later stages a half-sphere is formed that is equivalent to half of the normal sphere (Fig. 64, C). The open side corresponds to the side at which the half would have been united to the other half. Thus up to this point a half-cleavage and a half-blastula have appeared.[114]

In later stages the open half-blastulÆ close in, producing a whole sphere that becomes perfectly symmetrical (Fig. 64, D). A symmetrical gastrula develops (Fig. 64, E) by the invagination of a tube at one pole, and a symmetrical embryo is formed (Fig. 64, F) that resembles the normal embryo except in point of size.

Driesch has also found that a number of twin embryos arise from the shaken eggs. They arise from eggs whose blastomeres have been disturbed or shifted, so that each produces a small whole embryo, the two embryos being united to each other in various ways.

In a second paper, published in the following year, Driesch extended his experiments, and attempted to discover how far the “independence” of the blastomeres extends; i.e. he tried to find out if all the blastomeres resulting from the cleavage are alike. When one of the first four cells is separated from its fellows by shaking, it continues to divide, in most cases as a quarter, and produces later a small spherical blastula. Many of these blastulÆ, although apparently healthy, never develop further, although they may remain alive for several days. In one experiment only eight out of twenty-six reached the pluteus stage, with a typical digestive tract and skeleton.

From these experiments Driesch drew the important conclusion that the cleavage cells or blastomeres of the sea-urchin’s egg are equivalent, in the sense that if they were interchanged a normal embryo would still result. A somewhat similar view is expressed in the dictum that the position of a blastomere in its relation to the others determines what part it will produce, if its position is changed it gives rise to another part, etc.,—or, expressed more concisely, the prospective value of a blastomere is a function of its position.[115] Driesch extended these experiments further in 1893. His aim was to separate different groups of cells at the sixteen-cell stage in order to see whether the cells around the micromere pole (or “animal pole”) if separated from those of the opposite (or “vegetative pole”) could produce a whole embryo, etc. Eggs whose membranes had been removed by shaking immediately after fertilization were allowed to develop normally to the sixteen-cell stage and were then shaken into pieces. Amongst the groups of cells that were present those that contained the micromeres were picked out. It was found that they give rise to whole embryos. In order to obtain cells that belong to the vegetative hemisphere, the blastomeres were shaken apart at the eight-cell stage, and those groups of cells that in later divisions did not produce micromeres were isolated. From these also whole embryos develop. The results show that the cells of both hemispheres are able to produce whole embryos, and that at the sixteen-cell stage the different parts of the egg are still capable of producing all parts of the embryo. It is important to observe that the results of the experiment do not show that if the normal development goes on undisturbed any part of the egg may become any part of the embryo, for it is highly probable that a definite region of the egg may always produce a definite part of the embryo. The results do show, however, that, even if this is true, any cell has the power of producing any or all parts of the embryo if the normal conditions are changed.

In connection with these experiments Driesch discussed the factors that determine the axial relations of the embryo. If all the cells have the power of producing all parts, what determines in the normal development, and also in the development of a part of the whole, the axial relations of the embryo? Driesch assumed that the egg has a polar structure, and that the same polarity is found in all parts of the protoplasm. Around this primary axis all the parts are alike or isotropous.[116] The origin of the mesenchyme and the position of the archenteron, that develop at one pole, are determined by the polarity of the protoplasm. The plane of bilateral symmetry may appear in any one of all the possible radial planes around the primary axis. The selection of a particular one is due to some accidental difference in the structure of the protoplasm, or to some external factor. In later papers Driesch modified this view, and assumed that along with the primary polarity a bilateral structure also exists in the protoplasm.

Wilson (’93) studied the development of isolated blastomeres of amphioxus, and found that it agreed in all essential respects with the mode of development of the blastomeres of the sea-urchin. The isolated blastomeres of the two-cell and four-cell stages produce whole embryos, but the blastomeres of the eight-cell stage develop only as far as the blastula. The blastomeres segment, after separation, in most cases not as a part, but as a whole egg would divide, although the cleavage of the one-eighth blastomere only approaches that of the entire egg, but is never identical with it. Incompletely separated blastomeres give rise to twins, triplets, etc. Wilson agreed with the Hertwig-Driesch conception of the value of the early blastomeres, and accepted the view that the fate of each is a function of its position, and that at first they are qualitatively alike. During the early cleavage he supposed that a change takes place that is slight at the two-cell stage, greater at the four-cell stage, and in the eight-cell stage the differentiation has gone so far that the blastomere can no longer return to the condition of the ovum. “The ontogeny assumes more and more the character of a mosaic work as it goes forward.”

Loeb (’94) showed that if the eggs of the sea-urchin are placed in sea water, diluted by distilled water, the egg swells and bursts its membrane, so that a part of its protoplasm protrudes. Into this protrusion some of the first-formed nuclei pass, and from both the part remaining in the egg membrane, as well as from the protruding part, an embryo is produced, the two embryos often sticking together. In several cases two to eight separate groups of blastomeres are formed from one egg and develop into whole embryos.[117]

The question of the number of cells which are produced by the one-half and one-fourth embryos had not up to this time been determined. Until this was known it could not be stated whether the smaller embryos were miniature copies of the normal embryos in all respects, or whether they assumed the typical form with fewer cells. I found (’95) that the blastula from one of the first two blastomeres contains half the number of cells produced by the whole embryo, and that in the later stages also it contains only about half the normal number. The one-fourth blastomere produces only a fourth of the whole number of cells, and yet can develop with this number, in many cases, into a whole embryo. The one-eighth blastomere produces one-eighth the normal number of cells. In most cases I found that these one-eighth blastomeres do not produce embryos, but occasionally they produce a gastrula, and probably a young pluteus stage.

The development of nucleated fragments of the egg was also studied in order to find out if they too produce a smaller number of cells than does the whole egg, and a number in proportion to their size. The problem is different in this case, because the nucleus has not divided before the piece is separated, and the results ought to show whether there is a prescribed number of divisions for the egg nucleus, or whether the number of times it divides is regulated by the amount of the protoplasm. It was found that the number of cells produced by each fragment is in proportion to the size of the piece. This shows that the division of the nucleus is brought to an end when the protoplasm has become subdivided to a certain point.

A further examination of the number of cells that are invaginated in these smaller “partial” larvÆ to produce the archenteron seemed to show that they often use relatively more than their proportionate number. The normal blastula of SphÆrechinus granularis contains about five hundred cells and turns in fifty cells, or one-tenth the total number. The one-half and one-fourth embryos, and some of the small embryos from the egg fragments, seemed to invaginate more than one-tenth of their total number of cells.

Driesch (1900) reËxamined this point, and found that the embryos from isolated blastomeres may use the proportionate number of cells. I have made a new study of the problem on a larger scale and have found that my earlier statement, as well as that of Driesch, is substantially correct, and that the difference that we found is due to the time at which the embryos gastrulate. Thus the one-half embryos and even the one-fourth embryos, that gastrulate as soon as (or only a little later than) the normal, whole embryos, turn into the archenteron about one-half and one-fourth the number of cells invaginated in the whole embryo; but those partial embryos that gastrulate later (as most of them do) turn into the archenteron more than a half or a fourth of the number of cells turned in at first by the whole embryo. This difference between the early and the retarded partial embryos is in large part due to a slow increase of cells that takes place during the delay in development.

Driesch (’95) found that pieces of the blastula wall of the sea-urchin, if large enough, can also produce a gastrula and embryo. I found that the number of cells in these pieces does not increase appreciably after they are cut off (if the operation has been carried out at the end of the cleavage period), and that the new embryo is organized out of the cells present at the time of removal of the piece from the wall. There is, therefore, in this case no chance for “post-generation” by means of new cells produced at the side, which Roux has supposed to take place in the frog embryo.

The development of pieces of the blastula wall, if they are not too small, also shows that the lack of power to develop, found in some of the one-fourth and in many of the one-eighth blastulÆ, is not the result of any special differentiation that they have undergone during the cleavage period, but is due to their size.

A recent series of experiments by Driesch (1900) on the development of isolated blastomeres of the sea-urchin’s egg has given more exact data in regard to their limit of power to produce embryos, and has shown the possibilities in these respects of different parts of the egg. By means of a method discovered by Herbst (1900) it is possible to obtain isolated blastomeres more readily than by the somewhat crude shaking process. If the eggs, after fertilization and after the removal of the membrane by shaking, are placed in an artificial sea water, from which all calcium salts have been left out, the eggs divide normally, but the blastomeres are not held firmly together, and readily fall apart if the egg is disturbed. By means of a fine pipette any desired blastomere or group of blastomeres can be picked out. If these are returned to sea water they continue to develop.

Driesch found that the one-half and one-fourth blastomeres develop into proportionate gastrulÆ and larvÆ; that the one-eighth blastomeres, both of the animal and the vegetative hemispheres, sometimes produce gastrulÆ, and even the beginning of the larval stage with the rudiments of a skeleton. There are certain differences between the one-eighth larvÆ that come from the animal hemisphere and those from the vegetative half. More of the one-eighth blastomeres from the animal part of the egg die than from the opposite part, but of those that remain alive a larger percentage reach the gastrula stage than in the case of those from the vegetative pole; their protoplasm moreover is not so clear as is that of the larvÆ from the other hemisphere. These “animal pole” blastomeres develop faster than those of the other sort. The gastrulÆ from the one-eighth blastomeres of the vegetative hemisphere do not die so often after separation, the protoplasm of the larvÆ is clearer, and they often produce long-lived blastulÆ with long cilia. The blastulÆ often develop into gastrulÆ without mesenchyme. These results show that although whole larvÆ may be produced from the one-eighth blastomeres of both hemispheres, yet there are certain characteristics that may be referred with great probability to differences that are present in the protoplasm of the two hemispheres of the egg. The differences are not in all cases sufficient to interfere with the production of all the characteristic structures of the embryo, yet traces of the origin of the larvÆ can be found in their structure. It is probable that the so-called animal (or micromere) pole corresponds to that part of the egg from which the archenteron is produced. Hence the one-eighth blastulÆ from this hemisphere gastrulate sooner and in proportionately larger numbers than do those from the opposite hemisphere. The vegetative hemisphere would correspond to that part of the egg from which the wall of the normal gastrula is derived, and this may account for the clearer protoplasm of these embryos, their inability in many cases to gastrulate, their larger cilia, and the absence of mesenchyme in some of them. Driesch finds that the number of cells that go into the mesenchyme of the partial larvÆ is in proportion to the total number, and that the number of cells in the archenteron is probably also proportionate.[118]

The smallest blastomeres that produce gastrula are the one-sixteenth products. Out of a total of 139 cases only 31 produced true gastrulÆ, 5 produced gastrulÆ with evaginated archenteron, and 103 remained blastulÆ with long cilia. The one-thirty-second blastomeres were not observed to gastrulate.

Driesch (’95) has also made a study of the potentialities of the blastula and gastrula stages of sphÆrechinus, echinus, and asterias. If a blastula is cut in half before the mesenchyme cells are produced, both pieces produce gastrulÆ and larvÆ. Since some of the pieces probably come from the animal hemisphere, and others from the vegetative hemisphere, it follows that all parts of the blastula possess the power of producing whole embryos, and in this respect the potentialities are the same as for the blastomeres. If the experiment is made at a stage just before the archenteron has begun to develop (Fig. 65, A), the results may be different. A half that contains the region from which the archenteron is about to develop will produce a gastrula and a larva (Fig. 65, A, lower row to right of A). A half that contains only the opposite regions of the egg (Fig. 65, A, upper row) may in some cases gastrulate,[119] often abnormally, but as many as half of the pieces do not gastrulate. They may remain alive for a week or more, and even produce a typical ciliated ring with a mouth in the centre, but do not form a new archenteron. These important results show that after the formation of the mesenchyme and archenteron at one pole, the other cells of the blastula wall are no longer able to carry out a process that the same cells were able to carry out at a slightly younger stage, but whether this loss of power is connected with the previous formation of the archenteron, or due to some other change which has by this time taken place in the cells, cannot be determined from the experiment. It is also important to note that these small ectodermal blastulÆ can still develop whole, typical, ectodermal organs, the ciliated ring and the mouth, and that the former especially has the characteristic structure of the whole normal ring.

Fig. 65.—A. Blastula of sea-urchin beginning to gastrulate. Cut in half as indicated by line. Two rows of figures to right show development of upper and lower halves. B. Later gastrula cut in half. Two rows of figures to right show later development. C. End of gastrulation process. Embryo cut in half. Two rows of figures to right show later stages of each half. D. Formation of endodermal pouches from inner end of archenteron. Embryo cut in two. Two rows of figures to right show later stages.