Regeneration

It has been shown during the last decade that double embryos may be produced artificially by incomplete separation of the first two blastomeres. Driesch, Loeb, and others have demonstrated that if the first two cells of the egg of the sea-urchin be incompletely separated, each may produce a single embryo and the two remain sticking together. Wilson has shown in amphioxus that the same result occurs if the first two cells are partially separated by shaking. Schultze has shown in the frog that if at the two-cell stage the egg is held in an inverted position, i.e. with the white hemisphere turned upwards, each blastomere gives rise to a whole embryo — the two embryos being united, sometimes in one way, sometimes in another, as shown in Fig. 63. In this case it appears that the results are due to a rotation of the contents of each blastomere, so that like parts of the two blastomeres become separated. In the egg of the sea-urchin, and of amphioxus, gravity does not have a similar action on the egg, but the results seem to be due to a mechanical separation of the blastomeres. These cases of double structures, produced by the segmenting egg,

appear to belong to the same category as those described above for adult forms — especially in those cases where pieces regenerate by morphallaxis. In connection with the production of double structures there should be mentioned a peculiar method of formation of new heads, first discovered by Van Duyne in a planarian. He found that if the FlG. 44. — A. Planaria lugubris, cut in two as far forward as region between eyes, regenerating half-heads. B. Same cut in two at one side of middle line. Smaller piece produced a new head. C. Planaria maculata, split in two. It produced two heads in angle. D. Another, that produced a single head in angle.

animal is cut in two in the middle line, the halves being left united only at the head-end, as shown in Fig. 44, D, C, there may appear one or two new heads in the angle between the halves. I have repeated this experiment with the same result, and have found that it may also occur when only a piece is partially split from the side of the body, as shown in Fig. 44, B. In Van Duyne' s experiment the two new heads do not appear unless the cut extends far forward, but if the division extends into the region between the two eyes there may be formed, as I have found, a single eye on each side that makes a pair with the old eye of that side (Fig. 44, A). It is evident in this case that each head has completed itself on the cut-side, the completion including the eye and the side of the head also with its " ear-lobe." The result, in this case, is the same as though the pieces had been completely cut in two. If the cut does not extend quite so far forward there are usually formed one or two heads near the angle, each with a pair of eyes and a pair of ear-lobes (Fig. 44, C). Sometimes a single head develops in the angle itself (Fig. 44, Z>), and it is difficult to tell whether it belongs to one or to the other side, or whether it is common to both sides. Van Duyne spoke of the double and single head of the latter kind which he obtained as heteromorphic structures in Loeb's use of the term. According to this definition, heteromorphosis is the replacement of an organ by one that is morphologically and physiologically unlike the original one, but this statement has been made to cover a number of different phenomena. The examples of heteromorphosis that Loeb gives by way of illustration of the phenomenon are : the production of a hydranth on the aboral end of tubularia, and the formation of roots in place of a stem in antennularia, etc. The formation of the heads in the angle in planarians does not appear to me to belong in this category. It seems rather that the phenomenon is of the same sort as the formation of a new head at the side of a longitudinal piece, and if so the new heads in the angle are, therefore, in their proper structural position for new heads belonging to the posterior halves.

Even if it should prove true that a single head may develop exactly in the angle itself, and belong to both sides, it can be interpreted by an extension of the same principle.1 The position of this median head turned backward suggests an obvious comparison with the production of the heteromorphic head in Planaria Ingubris, but a closer examination will show, I think, that the two cases are different. The heteromorphic head is produced only when the head is cut off close behind the eyes. If cut off slightly behind this region, a posterior end is generally formed. But in the worms split lengthwise the head in the angle may be formed at a level much farther posteriorly than the eyes. If the split extends into the head, then the eyes that develop are the supplements of those of the old part. Our analysis leads, therefore, to the conclusion that the heads, or parts of heads, in the split worms are not heteromorphic structures but supplementary heads.

1 A parallel case is found when a piece partially split in two at the anterior end (Fig. 24) produces one or two heads on each half, according to the extent of fusion of the new material that goes to form the new head or heads. SELF-DIVISION, as a means of propagation, is of widespread occurrence in the animal kingdom. In some cases the animal simply breaks into pieces and subsequently regeneration takes place in the same way as when the animal is cut into pieces by artificial means. In other cases the parts are gradually separated, and during this time new parts are formed by a process resembling that of regeneration after separation. A few zoologists 'have tried to show how the process cf regeneration before separation has been derived from regeneration following self-division. It is our purpose to examine here the evidence in favor of this hypothesis.

A study of the forms that propagate by means of self-division shows that the process is present in many groups of the animal kingdom. In the unicellular forms this method is universally present ; and in the multicellular forms the division of the individual cells is looked upon as a process similar to the method of propagation in the protozoa. The sponges do not multiply by self-division. In the coelenterates, on the other hand, we find this mode of propagation present in most forms. Hydra appears rarely, if at all, to divide by a cross-division, and, although one or two cases of longitudinal division have been described, it is not improbable that they have been started by the accidental splitting of the oral end. The hydromedusae, Stomobrachium mirabile, Phialidium variabilc, Gastroblasta Raffcelei, are known to increase by division.1 Several actinians and many corals divide longitudinally, while the scyphistoma of the scyphomedusae produce free-swimming ephyras by cross-divisions of the fixed strobila stage. The ctenophors do not divide.

It is known that several fresh-water planarians propagate by division, the tail-end breaking off in the region behind the old pharynx. In one form,2 and possibly in others, regeneration may begin before the separation takes place. Many of the rhabdocoelous planarians increase by cross-division — the separation taking place more nearly in the middle of the body. In these forms the parts develop new organs more or less completely before they separate. In the trematodes self-division does not take place. The division of the body of the tapeworm into proglottids may represent a process of self-division, but the proglottids do not regenerate after separation.

The nemertians break up readily into pieces, if roughly treated or if the conditions of life are unfavorable, but this can scarcely be spoken of as a process of voluntary self-division. Regeneration takes place in some species, but imperfectly or not at all in others. In the group of annelids we find many cases of self-division, especially in marine polychaetes and in fresh-water oligochaetes. One of the most interesting forms, belonging to the first group, is the palolo worm in which the swimming headless form, that is set free by division, serves to distribute the sexual products. Subsequently it appears that the piece dies without regenerating a new head. If we examine more in detail some of the cases of self-division in annelids, we find the following interesting facts. In nereis the posterior region of the body undergoes great changes of structure, the new worm being known under a different name, viz. heteronereis. In this part of the worm, eggs (or sperm) are produced, but it does not separate from the anterior end as a distinct individual. In the family of scyllids the changes that take place in the posterior or sexual end of the body are often accompanied by non-sexual modes of fission. In some species the changes that take place are like those in nereis, and no separation occurs ; in other species the sexual region becomes separated from the anterior or non-sexual regions. In scyllis a new head develops, after separation, on the sexual or posterior piece. A new tail is also regenerated by the non-sexual or anterior piece, and as many new segments are formed as are lost. The new posterior region may again produce sexual cells, and again separate. In autolytus a new head develops on the posterior piece before it separates. A region of proliferation is also found at the posterior end of the anterior part. In some species new individuals develop in this zone of proliferation, and a chain of as many as sixteen worms may be present before the one first formed drops off. A still more complicated process is found in myriana. The region just in front of the anus elongates, and gives rise to a large number of segments.

These form a new individual with the head at the anterior end. Then another series cf segments is proliferated at the posterior end of the old, or anterior worm, and just in front of the first-formed individual. This region also makes a new individual. The process continuing, a chain of individuals is produced, with the oldest individual at the posterior end and the youngest at the anterior end of the series. Each individual grows larger, and produces

more segments at its posterior end. Reproductive organs appear in each individual, and when the germ-cells are mature the chain breaks up. None of the earthworms propagate by self-division, although occasionally, under unfavorable conditions, pieces may pinch off at the posterior end.1 Lumbriculus, on the other hand, propagates by selfdivision, although it has been disputed whether the division takes place without the intervention of an external injury or disturbance of some sort, or whether the division may take place entirely from internal causes, that is, spontaneously. Von Wagner has shown that at certain seasons lumbriculus breaks up much more readily than at other times, which may only mean that it is more sensitive to stimuli at one time than at another.

The pieces into which lumbriculus breaks up regenerate after separation. In another form, Ctenodrilus monostylos, division takes place first in the middle of the body behind a cross-septum. Each half may again divide in the same way, and the same process may be repeated again and again until some of the pieces are reduced to a single segment. A new anterior and posterior end may then develop on each piece. In Ctenodrilus pardalis each segment of the middle region of the body constricts from the one in front and from the one behind, and each produces a new head at its anterior end and an anal opening at its posterior end. The worm then breaks up into a number of separate worms. In this series, self-division of the individual is not associated with the development of sexual forms, but seems to be a purely non-sexual method of reproduction. In the leeches selfdivision does not occur, and no cases are known in the mollusks.

In the echinoderms several forms reproduce by voluntary self-division. In the brittle-stars some forms divide by the disk separating into two parts, one having two and the other three of the old arms. Each piece of the disk then regenerates the missing part of the disk as well as the additional arms. In the starfishes the arms may be thrown off if injured, and, while in certain forms the lost arm does not regenerate a new disk, yet, according to several writers, it may in other species regenerate a new animal. Dalyell observed a process of self-division in a holothurian, each part producing a new individual, and more recent observers have confirmed this discovery.

No cases of self-division are known in the groups of myriapods, insects, crustaceans, spiders, polyzoans, brachiopods, enteropneusta, or vertebrates. Before discussing the general problems connected with the preceding cases, I should like to point out that it is certainly a striking fact that in all, or nearly all, of these cases of self-division, the separation takes place in the shortest axis, without regard to the structure of the animal. A law similar to that enunciated in connection with the division of the cell seems to hold for the organism as a whole : namely, division takes place, as a rule, in the shortest diameter of the form. The protozoa are, in a sense, excluded, since being unicellular forms they come under the rule for the division of the cell. In the ccelenterates we find the actinians and corals, that have short, cylindrical bodies, dividing from the oral to the aboral end, while the longer scyphistoma divides transversely. The flat, bell-shaped medusa, gastroblasta, divides in an oral-aboral plane. The flat-worms and annelids divide transversely, and, therefore, in the plane of least resistance. The most important illustrations of this principle are furnished by the echinoderms. Those brittle-stars that divide through the disk do so in the shortest direction, that is, from the oral to the aboral side, whilst the holothurians that are long, cylindrical forms divide across the body and, therefore, in a structural plane at right angles to that of the brittle-stars. It may be claimed that in all these cases the plane of division is that in which the animal is most likely to be broken in two by external agents, but this is, I think, only a coincidence, and the result is really due to internal conditions. The division is brought about in most cases, and perhaps in all, by the contraction of the muscles ; and the arrangement of the muscles in connection with the form of the body is the real cause of the phenomenon.

Returning to the general question of the occurrence of the process of division in the different groups, we find that in nearly all of them in which self-division occurs it is found in a number of different forms in the same group. The process seems to be characteristic of whole groups rather than of species, and so far as evidence of this sort has any value it points to the conclusion that the process is not necessarily a special case of adaptation to the surroundings, because the species that divide may live under very diverse conditions.

A further examination of the facts throws a certain amount of light on the relation between the processes of self-division and of regeneration. The following questions may serve to guide us in our examination : — (i) Is regeneration found only in those groups in which self-division takes place as a means of propagation ; or, conversely, does self-division only occur in those groups that have the power of regeneration ? (ii) Is regeneration confined, in the groups that make use of selfdivision as a means of propagation, to those regions of the body where the self-division takes place ?

(iii) Is regeneration as extensive in the groups that do not propagate by self-division as in those that do ? (iv) Can we account, in any way, for the presence of self-division in certain groups, and for its absence in others ? (v) What relation exists between the forms that prepare for subsequent self-division and those that do not ? The first question is easily answered. Regeneration is also found in nearly all the other groups that do not propagate by self-division, — as, for instance, the mollusks, vertebrates, etc. The second half of the question may also be answered. All the groups that propagate by self-division have also the power of regeneration.1

In answer to the second question there is ample evidence showing that regeneration is by no means confined to those regions of the body in which the self-division occurs. In answer to the third question, it may be stated that although, in the groups that propagate by self-division, regeneration may be present in nearly all parts of the body, the same phenomenon occurs in other groups that do not propagate by division. The fourth question offers many difficulties, and our answer will depend largely upon what we mean by " accounting for" the process in certain groups. If the question is interpreted to ask, Why does an animal divide ? no answer can be given. If it is meant to ask, Can we see how the process would be difficult, or even impossible, in certain groups and not in others ? then an approximate answer may be given, or at least an hypothesis formed. In the first place, the power of regeneration must be present in the region at which the selfdivision takes place in order that the result may lead to the formation of new individuals, or else be acquired in that region along with the acquirement of the means for division. A leech is not much more complicated than a marine annelid, yet it has little or no power of regeneration ; hence, perhaps, propagation by division could not be acquired by the leeches until they had first acquired the power to regenerate. In the second place, in certain forms a separation of the body into two parts would lead to the death of one or of both parts, owing to the dependence of the different regions upon each other. In forms like the vertebrates, insects, Crustacea, etc., we can readily see why this would be the case. Hence propagation by means of self-division could not be acquired, since the division itself would lead to the destruction of the organism. In the third place, the structure of the body may be such that the process of self-division would be mechanically impossible. A hard outer coat, like that of the sea-urchin, combined with a weak development of the mus-

1 The proglottids of the cestodes seem to be an exception, hut they are little more than sacs filled with embryos at the time of their separation. How far regeneration may take place in the scolex, or young proglottids, is not known, but it is not improbable that some of the abnormal forms that have been described may be due to regeneration. culature of the body, would effectively prevent the self-division of the animal. The fifth question has many sides. It involves us on the one hand in a historical question of the origin of self-division, and on the other hand in a discussion of the stimulus that brings about, not only the division, but the changes that precede the division in those cases in which the new part develops before division takes place.

Several zoologists have held that the process of self-division followed by regeneration has been the starting-point for the process of propagation preceded by regeneration. Von Kennel, for instance, maintains that self-division in some of the annelids has arisen in this way. He says : " We recognize everywhere in the animal kingdom the power of organisms to replace lost parts, and we call this regeneration. It may be developed in very different degrees in animals, and, as a rule, only those parts of the body have the power of regeneration that still possess the organs that are essential for independent existence. The higher the organization of the animal, so much the less is its power of regeneration, perhaps, because the division of labor of the different organs has gone so far that extensive injuries cannot be repaired. . . . There is no doubt that this power is adaptive, in a high degree, to preserve the species under unfavorable conditions, so that they are much better off in the battle for existence than are the animals that live under the same conditions but have not the power of regeneration. . . .

The power of regeneration that gives the animal a better chance in the battle for existence and, therefore, makes more certain the continuance and the distribution of the species will be, as is well known from numerous observations, in a high degree inherited, indeed even increased so that its descendants will possess that power in a higher degree than their forefathers ; and, in consequence, a much smaller stimulus (motive) suffices, than at first, to bring about the division of the parts." After showing, according to the usual formula, that the process of regeneration is useful, and, therefore, would come under the guidance of natural selection, von Kennel proceeds to show how the result is connected with an external stimulus! He asks: "Can accidental injuries account for the result (viz. for the division in lumbriculus, planarians, and starfish), since how few starfish are there with regenerating arms in comparison with the enormous number of uninjured individuals? Should we not rather look for the external stimuli that have initiated the process of self-division ? " " Animals that have developed the power of regeneration by a long process of inheritance will have acquired along with this the property of easier reaction to all external adverse conditions. In a sense the sensitiveness for such stimuli is sharpened, and the animal responds at once by breaking up. In the same way the

ear of a good musician becomes more sensitive through practice. If we think of the same stimulus as regularly recurring, and as always answered in the same way, then we may look upon it as a normal condition of the life of the animal and its response as also a normal process in the animal. If, for instance, the breaking into pieces of lumbriculus is a consequence of the approach of cold weather or of other external conditions, then the organization of this animal must react by breaking up in consequence of its adaptation to the conditions acquired through heredity. The self-division becomes a normal process under normally recurring conditions. If the organism has been accustomed to respond through numerous generations, and, therefore, its sensitiveness has become highly developed, it will be seen that it may be influenced by the slightest change in the unfavorable conditions, and although, at first, the change may not be sufficiently strong to cause the animal to divide, yet the introductory changes leading to the division may be started, which will in turn make the division, when it occurs, easier and the animal that possesses this responsiveness more likely to survive. This would be the case if a slow process of constriction took place, so that, at the time of separation, no wounds of any size are -formed." " By a further transfer of the phenomenon, a partial, or even a complete, regeneration may set in before division takes place." " We find changes like this in the series of forms, Lumbriculus, Ctenodrilus monostylos, Ctenodrilus pardalis, Nats, Clicstogaster. It appears in a high degree probable that the series has originated in the way described.

Perhaps zoologists will find after some thousands of years that lumbriculus propagates as does nais at present." In this way von Kennel tries to show how the process of regeneration, that takes place before division, has been evolved from a simple process of breaking up in response to unfavorable conditions. The imaginary process touches on debatable ground, to say the least, at every turn, and until some of the principles involved have been put on a safer basis, it would be unprofitable to discuss the argument at any length.

We should never lose sight of the fact that the arranging of a series like that beginning with lumbriculus and ending with chaetogaster is a purely arbitrary process and does not rest on any historical knowledge of how the different methods originated or how they stand related, and no one really supposes, of course, that these forms have descended from each other but at most that the more complicated processes may have been at first like those shown in other forms. Even this involves assumptions that are far from being established, and it seems folly to pile up assumption on top of assumption in order to build what is little more than a castle in the air.

In several groups of animals a process of budding takes place that presents certain features not unlike those of self-division. It is difficult, in fact, to draw a sharp line between budding and selfdivision, and although several writers have attempted to make a distinction between the two processes, it cannot be said that their definitions have been entirely successful. It is possible to make a distinction in certain cases that may be adopted as typical, but the same differences may not suffice in other cases. For instance, the development of a new individual at the side of the body of hydra is a typical example of budding, while the breaking up of lumbriculus or of a planarian into pieces that form new individuals is a typical example of division. In a general way the difference in the two processes involves the idea that a bud begins as a small part of the parent animal, and increases in size until it attains a typical form. It may remain permanently connected with the parent, or be separated off. By division we mean the breaking up of an organism into two or more pieces that become new individuals, the sum-total of the products of the division representing the original organism. Von Kennel first sharply formulated this distinction, and it has been also supported by von Wagner, who has attempted to make the distinction a hard and fast one ; 1 but as von Bock has pointed out, there are forms like pyrosoma and salpa in which the non-sexual method of propagation partakes of both peculiarities, and in Syl/is ramosa the individuals appear to bud from the sides, while in other annelids a process of division takes place. Von Bock assumes, therefore, as more probable, that budding and self-division are only different phenomena of the same fundamental process.

It might be better, I think, to go even further in order to clear this statement from a possible historical implication, and state only that the two processes involve some of the same factors. Budding occurs in several groups of the animal kingdom. There are numerous cases in the protozoa, such, for instance, as that in noctiluca. In the sponges buds are formed that go to build up a colony in most instances. In the coelenterates cases of lateral budding are found in nearly all the main groups, and in one and the same individual, as in the scyphistoma of aurelia, in fact both budding and division occur. In the polyzoa, in the ascidians, and in cephalodiscus lateral budding takes place. In the rhabdocoel turbellarians, and in some of the annelids, we find chains of new individuals produced by a process that is often spoken of as budding. It is convenient, however, to distinguish these cases of axial budding from those of lateral

budding; for, while they both involve an increase in the products over that of the original animal, the axial relations in lateral buds are established in a new plane, while in axial budding the main axis of the new animal is a part of that of the old, and this difference may involve different factors. The process of budding does not occur in the insects, spiders, crustaceans, mollusks, ctenophores, brachiopods, nematodes, vertebrates, or in several other smaller groups.

This examination shows that there are groups in which both processes take place, viz. ccelenterates, planarians, annelids ; and others in which budding alone takes place, viz. ascidians, polyzoa, cephalodiscus; and one group at least in which division, but not budding, takes place, the echinoderms. It is obvious that from the occurrence of the process of budding in the animal kingdom we cannot infer anything as to its relation to division or to regeneration.

It has been pointed out that in the flowering plants, in which the growth takes place by means of buds, the power of terminal regeneration is very slightly developed, and its absence is sometimes accounted for on the ground that the new growth takes place by means of the development of lateral buds. If by this statement it is meant that buds being present the suppression of regeneration in other regions may occur, then there may be a certain amount of truth in the statement. If, however, it is intended to mean that because a plant has acquired the power of reproducing new parts by means of buds it has, therefore, lost the power to regenerate in other ways (or has never developed the power to regenerate), then the argument is, I think, fallacious ; for we find even in flowering plants that the new buds sometimes arise from the new part, or callus, that forms over the cut-end, and this process resembles a real regenerative process. We also find that hydroids that produce lateral buds also regenerate freely from an exposed end. But we are at present so much in the dark in regard to the causes that bring about budding in organisms that a discussion of the possibilities would hardly be profitable.

The process of autotomy differs only in degree from the process of self-division. In autotomy the part thrown off does not produce a new animal. The breaking off of the tail of the lizard at the base, if the outer part is injured, is an example of a typical process of autotomy. The throwing off of the crab's leg, if the leg is injured, is also another typical case of autotomy. There is a definite breakingjoint at the base of the crab's leg at which the separation always takes place (Fig. 45, A i-i). The breaking-joint is in the middle of the second segment from the base of the leg, where there is found,

on the outside of the leg, a ring-like groove that marks the place of rupture. A comparison of the legs of the crab with the walking legs of the crayfish, or of the lobster, shows that the groove in the crab's legs corresponds to a joint in the legs of the two other forms. In the crayfish and lobster the walking legs generally break off at this same level, although by no means as easily or with as much certainty as in the crab. The first pair of legs of the crayfish and