It is significant to find that the theory of pre-formation of the embryo in the egg, that was so very widely held during the seventeenth and eighteenth centuries, and during the first part of the nineteenth century, was at once applied to the explanation of the regeneration of animals when this process became known. Bonnet in 1745 attempted to explain the newly discovered facts in regard to the regeneration of animals by means of the pre-formation theory. Just as the egg was supposed to enclose a pre-formed germ, so he imagined there lay concealed latent germs in the adult animal. At first Bonnet thought that these germs must be whole germs, like those contained in the germ cells of the reproductive organs, and that only as much of any one developed as was needed to replace the missing part. Later, however, he admitted that the germs might be incomplete germs, which are so located in each region that they represent the parts of the body beyond that region. The purpose of these germs is to replace any accidental injuries to the animal. He pointed out that some animals are more subject to injuries than others, and these animals are he thought especially well supplied with germs. Since in some animals the same part may be replaced several times, Bonnet assumed that on each occasion a new germ is awakened. As many sets of germs are present in these animals as the number of times the animal is liable to be injured in the course of its natural life.

Bonnet found that in lumbriculus a new head and a new tail may appear at almost any level, if the worm is cut in two, and, therefore, he supposed, head germs and tail germs are present throughout the worm. But why, if this is so, should a head germ always develop at the anterior end, and a tail germ at the posterior end of a piece cut from the body? Bonnet’s keen mind saw that it was necessary to make a further assumption. He supposed that the fluids of the body that pass forward carry nourishing substances for the head. When the worm is cut in two these substances are stopped at the anterior cut-surface, and there accumulating act on the latent head germ, and nourishing it, cause it to develop. Correspondingly the nourishing substances for the tail flow backward, and accumulating at the posterior cut-surface awaken a tail germ to activity.

The part of the body in which these nourishing substances are supposed to be produced is not specifically stated, but in one passage Bonnet says that the fluids that flow toward the head are there used up in that organ, and we may infer that he held a similar view for the posterior region. He offers no explanation of the cause of the flow of these substances in a given direction, and in this respect his hypothesis lacks support where it is most needed. In fact, it is no more improbable that a head germ should always develop at the anterior end and a tail germ at the posterior end, than that head-forming substances should flow in one direction and tail-forming in another. It is not that it is worth while to object to Bonnet’s hypothesis on the ground that it does not explain everything, but it is worth while to point out that it gives only the appearance of an explanation, and that it begs the whole question by the assumption of particular nourishing fluids flowing in definite directions. So far as the blood is concerned, we know that the different parts of the body take from it those substances or fluids that they make use of, not that special fluids flow to particular regions. It is probable that Bonnet thought of the blood rather than of any other subtler fluids passing through the tissues; and, if so, there is nothing that we know in regard to the behavior of the blood that lends support to Bonnet’s idea.

Bonnet takes care to state that the pre-formed germs may not appear to us like miniature copies of the part into which they develop, but they are so constructed that, as they absorb nourishment and become larger, they assume a characteristic form.

Weismann, who has also accepted the pre-formation hypothesis to account for the development of the egg, has applied the same conception of pre-formation to the process of regeneration. He believes that partial, latent germs are present in different parts of the body, and that these germs are present especially in animals that are liable to injury and in those parts of the body that are likely to meet with accidents. In these essential respects, Weismann’s idea is the same as Bonnet’s; but in regard to the location of the germs, and their manner of awakening, and as to how the forms, liable to injury, have acquired their power to regenerate, Weismann adopts more modern standards. He believes that the germs are located in the nucleus. Those that bring about the development of the egg are supposed to be different from those that bring about regeneration, because the method of regeneration is generally different from the method of development of the egg.

Regeneration, on Weismann’s view, is brought about by latent cells containing pre-formed germs in the chromosomes of the nucleus. These germs are called the determinants. Since at each level in an animal, or in a part of an animal, regeneration may occur and replace the missing part, it is assumed that the germs are correspondingly different at each level, and represent all the parts that lie distal to that region. Weismann does not suppose that there is a single germ at each level that represents all the distal parts, but that in each layer, or organ, or part there are many cells that contain germs corresponding to the distal regions. The qualities of the latent cells are sorted out by means of the qualitative divisions of the chromatic material of the nucleus. Moreover, since the new part can itself regenerate, the further assumption is made that during regeneration new subsidiary or latent cells are laid down at each level. This is supposed to be brought about by a quantitative division of each germ after it has reached its definitive position in the new part.

Weismann’s general attitude toward the problem of regeneration is summed up in the following statements: “It may, I believe, be deduced with certainty from those phenomena of regeneration with which we are acquainted, that the capacity for regeneration is not a primary quality of the organism, but that it is a phenomenon of adaptation.” Again, “Hence there is no such thing as a general power of regeneration; in each kind of animal this power is graduated according to the need of regeneration in the part under consideration.” “We are, therefore, led to infer that the general capacity of all parts for regeneration may have been acquired by selection in the lower and simpler forms, and that it has slowly decreased in the course of phylogeny in correspondence with the increase in complexity of organization, but that it may, on the other hand, be increased by special selective processes in each stage of its degeneration in the case of certain parts which are physiologically important and at the same time frequently exposed to loss.”

The evidence brought forward in the preceding pages leads, I think, to precisely the opposite conclusions, and, in certain cases at least, it has been shown that there can be no relation between the power of regeneration and the extent of exposure of a part to injury or to loss. It is unnecessary to enter here further into this question, since it has been discussed already in Chapter V.

Weismann’s statement that the power of regeneration has decreased “in correspondence with the increase in the complexity of the part” cannot by any means be entirely accepted. If the complexity of a part is of such a kind that the part cannot sustain itself independently until regeneration has taken place, or if the exposed surface of the wound is such that it cannot be closed over, or if the new part cannot be properly nourished, or if the tissues have changed in such a way that their cells can no longer multiply, then the statement is, to a certain extent, true. On the other hand, when we find that one of the most complicated organs of the body, the eye, can regenerate in the salamander, if only a piece of the optic cup is left attached to the nerve, we may well doubt if there is any such direct and general connection between regeneration and complexity as Weismann maintains.

Weismann’s so-called “mechanism” of qualitative nuclear division is the basis of his conception of pre-formation. We are, I think, at present in a position to reject not only this conception, since it finds no support either in observation or experiment, but also his view that regeneration is brought about by latent cells; for it has been shown in a large number of cases that the new cells come directly from the old, differentiated ones. In a previous chapter it has been pointed out that Weismann’s idea that regeneration has been acquired by a process of natural selection, and is under the influence of this supposed agent, is in direct contradiction to a number of known facts. Under these circumstances we are warranted, I think, in concluding that the entire Weismannian point of view is wrong.

The process of regeneration has been often compared to the process by which a broken crystal completes itself. Herbert Spencer, in particular, has elaborated this idea. In his book on the Principles of Biology, he says: “What must we say of the ability an organism has to recomplete itself when one of its parts is cut off? Is it of the same order as the ability of an injured crystal to recomplete itself? In either case new matter is so deposited as to restore the original outline. And if, in the case of a crystal, we say that the whole aggregate exerts over its parts a force which constrains the newly integrated molecules to take a certain definite form, we seem obliged, in the case of the organism, to assume an analogous force.” Spencer has called attention to a superficial resemblance between the renewal of a part of a crystal and the regeneration of an animal, and without further inquiry into the profound differences between the processes, assumes that “analogous forces” are at work. Now that we know something more of both processes, we find so much that is totally different, that there remains no basis for Spencer’s conclusion, namely, that analogous forces must be present. Furthermore, Spencer’s statement that the whole crystal aggregate exerts over its parts a force of some kind is diametrically opposed to our idea as to the method of “growth” of a crystal in a saturated solution. The new material is added always at the surface of the crystal, and the growth of each point is self-determining. There is no central force that controls the deposition of new material in the different regions. Rauber’s work on the so-called regeneration of the crystal has given us a clearer conception of how the process is brought about. He has shown that when a piece is broken from a crystal, and the crystal suspended in a saturated solution of the same substance, it becomes larger by the deposition of new material over all its surfaces. The addition of new material may be more rapid over the cut-surface than elsewhere, but it must not be supposed that the more rapid “growth” takes place in order to complete the form of the crystal, for the growth over the cut-surface follows precisely the same laws that regulate the “growth” over all the other surfaces, that is taking place at the same time. In this respect we find an essential difference between the regeneration of a crystal and that of an animal, since in the latter the growth takes place only over the cut-surface; and, in forms that regenerate by proliferation, at the expense of the old material, so that the old material is correspondingly diminished as the new part grows larger. Regeneration may even take place in an animal deprived of all food, and also in one that is starving to death and diminishing in size. In those forms that regenerate by a change in shape of the entire piece into that characteristic of the typical form, the process bears not even the remotest resemblance to the process in the crystal. It is so obvious from every point of view that the comparison is entirely a superficial one, that it seems useless to point out further differences between the two processes.

PflÜger (’83) has given, in brief outline, an hypothesis to account for the process of regeneration. He states that since there is always replaced exactly what is lost, the new part cannot arise from a pre-existing whole germ. If, for instance, the leg of a salamander is cut off at any level, as much comes back as is removed. The assumption of a leg germ is insufficient to account for the fact that only as much comes back as is lost, and not always a whole leg. PflÜger, therefore, offers another hypothesis. He assumes that food material is taken up at the wounded surface and organized into the substance of the new part. The new material is laid down at the surface of the old material, and is then organized into the kind of tissue that lay just beyond that region in the whole limb. Upon this first layer a new layer is deposited that is organized into the next part of the limb, and so on, until the whole missing part is replaced. PflÜger does not give any idea of how the new material is deposited at the cut-surface, but from what we know of the histology of the process we must suppose, if we should adopt PflÜger’s interpretation, that new cells are produced by the old ones, and that these new cells form the successive layers out of which the new limb is produced. PflÜger speaks of an arranging molecular force, which we can only suppose, in the light of what has just been said, to act from cell to cell through the continuous protoplasm. PflÜger also pointed out that in certain cases the organization can take place only in a certain direction, that is, in some forms regeneration can take place from one side of a cut-surface, but not from the other. He interprets this as due to a polarization of the protoplasm, one surface having peculiarities that are absent in the other.

There are certain objections to PflÜger’s hypothesis that suggest themselves. In the first place the new part does not, in many cases, replace all that has been removed, and hence it is difficult to see how the building up in the way PflÜger supposes, could take place. In these cases the new material forms only the distal end of the part removed, and the relation of the old to the new part is of secondary importance. Again, in cases of heteromorphosis, as when a tail develops on an anterior cut-surface of a piece of an earthworm, the result must be due to quite different factors from those suggested by PflÜger. The results are, in fact, the reverse of what the hypothesis demands. Furthermore, when the entire piece is transformed into a whole new organism, there is very little in the process to suggest a change like that postulated by PflÜger. On the other hand there cannot be much doubt that the old part may have some influence, and in certain cases a very important influence on the new part, but whether this is a purely molecular influence is open to doubt. In whatever way this influence may act, it is only one of a number of factors that take a share in the result. The amount of new material, that is formed before the organization of the new part begins, seems to be also a factor; and the one that determines how much of the missing part can be replaced, and this in turn seems to be connected with the lowest organization size that can be produced. The distal end of the new part forms always the distal end of the organ that is to be produced. If enough new material has developed (before the organization of the new part takes place) to produce all of the missing part, the latter is formed, but if the material is insufficient to produce the whole structure, then as much of the distal end as possible is formed. In some cases, as in the planarians, the missing intermediate regions may subsequently develop behind the distal part that is first produced.

Sachs has advocated a view which has many points of similarity to that of Bonnet, although, in reality, it is not a theory of pre-formation at all, but one of pure epigenesis. His idea rests on the view that the form of a plant, or of an animal, is the expression of the kind of material of which it is composed. Any change in its material leads to a corresponding change in the form of the new parts. Sachs holds that the idea of many morphologists, that there is for each organism a specific form that tends to express itself, and which controls the development of the organism, is a metaphysical idea that has no ground in science. For instance, Sachs thinks that the flower buds of a plant develop, not because of some innate, mystical force that causes the plant to complete its typical form, but because some substance is made in the leaves which, being carried into the growing region, becomes there a part of the material of that region, and from this new material a flower is formed. Simple and clear as this hypothesis appears to be at first sight, it will be found on more careful examination that it fails to account for some of the most characteristic phenomena of development and of regeneration. It may be granted at the outset that the presence of certain substances may undoubtedly influence the kind of growth of a new part; but, on the other hand, one of the most characteristic things of the organism is that it asserts its specific nature within quite a wide range of change, and, on the whole, resists the influence of other kinds of substances than those connected with its ordinary life. While Sachs looked no farther than the material substratum, and supposed that any change in this altered the form, there is, at present, sufficient evidence to show that it is the structure of the material that determines the most important changes that take place in it. This means, if we attempt to divest the statement of its somewhat metaphysical appearance, that the material of the organism is not simply a mixture of different kinds of materials, but a special kind of substance that has a definite structure of its own. This structure may, of course, be changed, but only by the addition of materials that the structure can take up as a part of itself. If the material does not become a part of the structure or organization, it is without effect on the form.[130] My meaning can, perhaps, best be illustrated by the method of regeneration of the tail of the fish from an oblique cut-surface. The growth of the new part is not determined by the kind or by the amount of the new material that is brought to the growing part, for, if it were, the new part would grow at an equal rate at every point; but the growth of the new part is regulated by the form of the tail of each particular kind of fish. The structure of the new part controls the growth of the material of the new part, and not the reverse. The only interpretation that can be given to this result is, I believe, that the new material assumes a definite structure, or what we may call an organization, and the subsequent changes are controlled by the kind of structure that is present; and since this structure has, as a whole, a definite form, we can state that the form controls the material, although the substitution of the word “form” for that of “the structure of the new material” may give the statement an unfortunate, metaphysical appearance.

In order to explain the regeneration of a piece of a plant, Sachs supposes that two substances are produced by the plant,—one a stem-(or leaf-) forming substance and the other a root-forming substance. If either of these substances combines with the protoplasm of any part, a stem or a root is produced from that part. When a piece of the stem is cut from a plant, these two substances accumulate, one at the distal end and the other at the proximal end of the piece, and their presence in these regions determines that new shoots develop at or near the apex, and new roots at the base. Sachs tried to show that the direction of the flow of these two substances is determined by the action of gravity,—the lighter substance flowing to the higher parts, and the heavier to the lower parts. We find here reproduced Bonnet’s idea of specific substances flowing in definite directions; but Sachs goes farther, and gives an explanation of the cause of the different directions taken by the two kinds of substances, viz. that it is due to the action of gravity. VÖchting has shown, as we have seen, after a thorough examination of the method of development of pieces of plants, that Sachs’s hypothesis fails to account for the results; and he shows also that an internal factor, which he calls the polarization, has the most important influence on the regeneration.

It is not difficult to show that there are many other cases to which the stuff hypothesis does not apply. If, as Bonnet attempted to show, the regeneration is due to different stuffs, there is no explanation to account for the flow in animals of head-forming stuffs forward and tail-forming stuffs backward. In animals that regenerate laterally as well as anteriorly and posteriorly, we should be obliged to assume side-forming stuffs as well as head-forming and tail-forming stuffs; and since the kind of structures that are produced at the side are different at each level, we should be obliged to assume that there are many kinds of lateral stuffs. If regeneration can take place in a dorsal and in a ventral direction, as, for example, when the dorsal and the anal fins of teleostean fishes regenerate, there must also be stuffs to account for their development. When regeneration takes place from an oblique surface, it must be supposed that two or more of these kinds of stuff are brought into action. The regeneration of just as much of the limb of the salamander as is cut off also offers difficulties for Sachs’s view. If we assume a leg-forming substance, it fails to account for the difference in the result at each level. If we assume that different substances come into play according to the amount of the leg that has been cut off, the hypothesis becomes as complicated as the facts that it pretends to explain.

A special case, to which the stuff hypothesis has been applied by Loeb and by Driesch, is that of tubularia, although the latter writer has used the hypothesis only to a limited extent as involving quantitative rather than qualitative results. There is present in the hydranth and stem of tubularia a red pigment in the form of granules in the endodermal cells. There is more of the red pigment in the stem near the hydranth than elsewhere. If a piece of the stem is cut off, it closes its cut-ends, and a circulation of fluid begins in the central cavity. In this fluid globules now appear that contain the red-pigment granules. The globules appear to be free endodermal cells, or parts of such, that have been set free in the central cavity. In the course of twenty-four hours the new hydranth begins to appear near one end of the stem, and in this region of the stem a much larger number of granules appear. A little later all the red granules disappear from the circulation.

Driesch has supposed that the red granules of the circulation become a part of the wall of the new hydranth. The disappearance of the red granules at this time from the circulation would seem to give color to this view. But, on the other hand, I have found evidence showing that this interpretation is incorrect. In the first place, the granules that disappear from the circulation can be found lying in a ball within the digestive tract of the newly formed hydranth; hence their disappearance can be accounted for, and we find that they are not, or at least in large part are not, absorbed into the forming hydranth.[131] In the second place, there is a great increase in the number of endodermal cells in the region in which the hydranth is about to appear, and the thickening that results takes place some time before the granules begin to disappear from the circulation. The new granules appear in the new endodermal cells, and are presumably formed by them. Again, the hydranth, that develops later at the distal end, appears when there are no granules in the circulating fluid, and yet the hydranth may contain as much red pigment as does the proximal one. Lastly, the development of very short pieces shows that at the time of the formation of the new hydranth there is an enormous increase in the number of red granules in the piece, for there are many more of them contained in the new hydranth than were present in the entire piece at the time of its removal.

Loeb has not referred to the red granules in the circulating fluid, but simply to the red pigment which is present in the walls of the piece. This is supposed to move forward into the hydranth region, and call forth the development of a new hydranth. A study of the number of the granules in the stem gives no support to this idea, and the method of formation of single and of double hydranths in short pieces shows that the increase in the number of granules in the hydranth-forming region is not due to migration, but to local formation.

That specific substances may have an influence on the growth of certain parts cannot be denied, but it appears that in general they play a very secondary rÔle as compared with other factors that determine the form of the organism or the development of a part. VÖchting’s beautiful experiments (’86) on tuberous plants show that the presence of an excessive amount of food substances in the plant, brought about by the artificial removal of the natural storehouses for such material, may act on certain parts, such as the axial buds, or on the stem, and cause them to produce structures that they do not produce under ordinary circumstances. The axial buds become swollen and produce tuber-like bodies above ground, especially if the parts are enclosed so as to be in the dark, since the light retards the growth of tubers of all sorts. But it should not be overlooked that these buds and stems are structurally the same things as the tuberiferous stolons that have been removed, and hence the excess of material is stored up in them in the same way as it is under normal circumstances in the underground stems or stolons. The reaction is one normal to the plant, although it usually takes place in a different part.

The preceding hypotheses that have been advanced to account for the phenomena of regeneration, draw attention to some of the most fundamental problems of regeneration and, even in those cases in which the hypotheses have not given a satisfactory solution of the problems, some of them have served the good purpose, both of directing attention to important questions and of leading biologists to make experiments to test the new points of view. We should not underrate their value, even if they have sometimes failed to give a solution of problems, for they have been useful if only in eliminating certain possibilities, and this simplifies all future work. So long as an hypothesis is of a sort that it is within the range of observational and experimental test, it may be of service, even if it prove erroneous; for our advance through the tangled thread of phenomena is not only assisted by advances in the right direction, but all possibilities must be tested before we can be certain that we have discovered the whole truth. The value of a scientific hypothesis depends, it seems to me, first, on the possibility of testing it by direct observation, or by experiment; second, on whether it leads to advance; and, lastly, on its elimination of certain possibilities.

The experiments described in Chapters II, III, IV, have shown that there are many resemblances between the phenomena of growth and of regeneration. It has been pointed out that when it could be shown that certain external agents have a determining influence upon growth, these same agents have a similar effect upon regeneration. This also holds apparently for internal factors, although it is much more difficult to demonstrate that this is true. The presence of an abundance of food material in the tissues hastens regeneration in the same way that growth is more rapid in a well-fed organism. Food may, however, be looked upon rather as an external factor than as an internal one. An excellent example of an internal factor is found in the interrelations of the parts to each other. This is shown in the development of a piece of a plant in which the apical buds develop faster than the proximal ones, and it appears that, in some way, the development of the latter are held in check by the development of the apical ones. Another case is found in the development of the bilobed tail of certain fish in which particular regions are held in check, while others grow at the maximum rate.

It is a curious fact that while we can cite several kinds of external influences that affect the development and the regeneration of organisms, the only internal factors that have been discovered are the so-called polarity and this interrelation of the parts. Perhaps there should also be added the specific nature of certain parts, limiting the possibilities of new growth in these parts, and the presence of the nucleus as necessary for the growth and regeneration of the organism.

If it be admitted that the same factors that affect the growth also affect in the same way the regeneration, we have made a distinct advance. It is, moreover, not difficult for us to understand how this is possible. If we consider first those cases in which growth takes place at one or more points at which the cells are undifferentiated, and compare this condition with that in regenerating animals that produce new tissue by proliferation, we can picture to ourselves that the same factors would act on the undifferentiated tissue in the same way in both cases. This does not explain what causes the organism to produce the new cells that appear over an exposed surface, and we must search for other factors to account for the out-wandering of cells, and for the local multiplication of the cells at the cut-end. We find a parallel to those cases in which the growth of an organism takes place throughout the whole body, in those animals in which the regeneration also takes place in the old part. This comparison should not, however, be pushed too far, since, in some forms, as, for example, a salamander, the growth of the animal takes place throughout the body, while regeneration takes place by the proliferation of new material. The difference in the regenerative process in a salamander and in a form like hydra is not due so much to the inability of the old cells of the salamander to increase in number as compared with those of hydra, but rather, it appears, to a certain rigidity or stiffness of the body of the salamander that prevents the rearrangement of the parts; and the recompletion of the form takes place in the direction of least resistance, i.e. at the open or cut-end of the body.

Regeneration by means of morphallaxis takes place only in those forms in which the body is not made up of a series of separated parts. This kind of regeneration occurs in those organisms in which the normal growth consists only in the enlargement of a system of organs already present. A piece of an animal of this sort usually contains the elements of each kind of organ, and from these the new parts are produced, both by proliferation at the cut-ends and by the enlargement of the parts that are present in the piece. In forms with separate segments we find, in some cases, resemblances between normal growth and regeneration, as shown, for example, in the earthworm. There is present in the young worm a region in front of the last segment, or, rather, a part of this segment, from which new segments are formed. In the regeneration of the posterior end a knob of new tissue is formed, and out of this a few segments develop, the last one having a growing region similar to that in the young worm. The subsequent stages in the regeneration involve the formation of new segments from the last one, as in the young worm. There is no such growing zone at the anterior end of the young worm, and none is formed in the regeneration of an anterior end, so that only the segments that are first laid down in the new part are present in the new anterior end.

An interesting comparison may be made between the phenomenon of growth and that of contraction and expansion of the protoplasm. The bending of heliotropic organisms toward or away from the light, and the similar bending of negatively stereotropic forms away from contact with a solid body, are supposed to be phenomena of growth, and resemble in many ways the phenomenon of contraction. In a plant that bends toward the light, it is found that the most obvious change involves the amount of water on the two sides of the stem, and this is most probably connected with a fundamental structural change in the protoplasm, that is too subtile for further analysis. In the regeneration of some forms it is found that they respond in the same way to light. While it cannot be demonstrated that these phenomena really depend on processes of contraction and of expansion, the results are nevertheless suggestive from this point of view. Furthermore, I think, one cannot study the regeneration of such forms as planarians, hydras, stentors, etc., without being struck by the apparent resemblance of the change in form that they undergo to a process of expansion. The idea of the expansion of a viscid body carries with it, of course, the idea of tension within the parts, and the return to the former condition is brought about by a release from the tension and a return to a more stable condition. If by the intercalation of new material the extended condition is fixed, a new state of equilibrium will be established.

It has been already pointed out that in a piece of a plant suspended in a moist atmosphere, the apical buds are those that first develop, and also grow faster than the others. The buds situated nearer the base may not even begin to develop, although they are at first as favorably situated, so far as external circumstances are concerned, as the uppermost ones. The roots appear first over the basal end, and those nearer the base grow faster than do those nearer the apex. There cannot be much doubt that the suppression of the basal buds and of the more apical roots is connected with the development of the apical buds and of the basal roots. This can be shown by cutting a piece in two, when some of the basal buds will grow into shoots and the apically situated root-buds, that are now on the base of one piece, will begin to grow. It seems to me this relation can be at least more fully grasped, if we look upon it as connected with some condition of tension in the living part. The tension can be thought of as existing throughout the softer, more plastic parts. As long as the apical bud is present at the end of a stem or branch, or even near the apex, it exerts on the parts lying proximal to it a pull, or tension, that holds the development of these parts in check; but if the apical bud is removed the tension is relaxed, and the chance for another bud developing is given.

It may be asked, how can it be explained that only the more apically situated buds of a piece develop, rather than the basal ones, since with the removal of the piece from the plant the tension has been removed also. The only answer that can be made, so far as I can see, would be that from the apex of the plant to its base the tension is graded, being least at the apex and increasing as we pass to the base. Those buds will first develop that are in the region of least tension, and their development will hold in check the other buds by increasing or reËstablishing the tension on the lower parts of the piece. A new system is then established, like that in the normal plant.

There are certain experiments with hydra that can, perhaps, be brought under the same point of view. When two long posterior pieces are united by their anterior cut-surfaces, each piece regenerates a circle of tentacles near the region of union, and each may produce a new head; or only one head, common to both pieces, develops at the side. Each piece has retained its individuality, which may be interpreted to mean that each piece has retained its original condition of tension. If, however, after a union of this kind one piece is cut off, as soon as the two have well united, near the place of union, so that it is relatively small as compared with the other component, it may produce a head at its exposed basal end, and neither heads nor tentacles may develop at the place of union of the pieces.

It is probable, in this case, that the larger component has acted on the smaller one, so that its polarity is changed and becomes like that of the larger component. It is possible, I think, to interpret this result in terms of our tension hypothesis. The condition of tension in the larger piece has overcome that of the smaller piece, so that the latter comes to have the same orientation that the larger piece has; and the development of a head at the free end then takes place. The development of this head holds in check the development of a head at the anterior end of the larger piece in the region of union of the pieces. When two pieces of hydra are united by unlike poles, i.e. so that they have the same orientation, it is found that if the pieces are not too long, a head develops at the free end and none in the region of grafting. The result is similar to that in plants; the development of the head at the free end suppressing any tendency that may exist to produce a new head by the posterior piece at the place of union. If the pieces united in this way are very long, a head develops at the apical end, and, in some cases, also near the line of union. This may be due to the pieces being so different at the place of union, that a head develops below this region before the unification of the two pieces is brought about, or because the formation of the head at the free end is relatively so far removed from the place of union of the pieces, that it does not influence the development of a head in this region.

These cases of grafting also illustrate another point of some interest. They show that the development of a head at the anterior end of a piece is not the result of the injury from the cutting or due to the action of some external condition on the free end, for the regeneration may take place when two anterior ends have been perfectly united to each other. The result can only be explained as the outcome of some internal factor such as polarity.

These examples have been chosen from hydra rather than from tubularia, in which somewhat similar phenomena have been observed, because in hydra the development of heteromorphic structures is of rare occurrence, while in tubularia external influence often calls forth a heteromorphic development. There cannot be much doubt, however, that in tubularia the same kind of internal factors are also at work.

A more striking illustration of the possible influence of tension of the parts is shown by an experiment with planarians. If the head of a planarian is cut off and the posterior piece is split partially in two along the middle line, as shown in Fig. 31, A, and then one of the halves is cut off just anterior to the end of the longitudinal cut, the result is as follows: A new head develops at the anterior end of the long half (Fig. 31, B), but no head develops on the posterior cut-surface, provided this part has reunited along the middle line with the long half, and a line of new tissue connects the anterior cut-surface of the long half and the more posterior cut-surface of the shorter half. At least this happens if the piece is not split too far posteriorly, i.e. through the region of the pharynx. If this is done, a new head may develop from the posterior cut-surface. In another way the development of the more posterior head can be brought about. If the shorter side-piece is kept from fusing with the longer side-piece in the middle line, it will invariably produce a new head (Fig. 31, C). The lack of development of the posterior head, when the two cross-cut surfaces are united by a connecting part of new material, can, it seems to me, be best explained by the influence of the developing anterior head, or of the new side on the posterior new tissue, and this influence can, I think, be better appreciated if we suppose some sort of tension to be the influence at work.

Another example may be cited that shows even more clearly that the internal factor regulating the growth in the new part is probably some sort of tension. I refer to the development of the tail of fundulus from an oblique cut, or of the bilobed tail of stenopus from a cross cut. The assumption of the typical form that leads to the holding in check of the growth in certain regions, as compared with others, can be best understood, I think, as due to some sort of tension established in the different parts, that regulates the growth in those regions.

It is evident that whatever factor will serve to explain the preceding cases must also be expected to apply to the development of the whole embryo from parts of the egg or blastula, if the position that I have taken is correct, namely, that these phenomena belong to the same general group. Does the tension hypothesis make clearer the development of a whole embryo from a part of an egg? This means, can we think of the readjustment that takes place as due to the establishment of a characteristic equilibrium that expresses itself in the tensions of the different regions? There is, so far as I can see, no difficulty in supposing that the organization is at bottom a system of this kind; indeed, it seems to me that from this point of view we can get a better appreciation of the organization and of the series of changes that take place in it during development. The example that Driesch has chosen as a typical one of vitalistic action, namely, the proportionate development of a part of the archenteron of the half-embryo, seems to me to be likewise a case to which we can apply the tension view.

In these, as well as in all other cases, we must think of the tensions as existing, not only in one direction, but in the three dimensions of space, and of all combinations of these. The material in which the tensions exist must be thought of as labile, so that a change in one region involves a rearrangement in many cases of the entire system. The new rearrangement appears to take place on the foundations of the old system.

It may appear that this idea of a system of tensions is too vague, that it fails to point out how the reorganization takes place, and that it gives not much more than the facts do themselves. There is a certain amount of truth in these objections which I fully appreciate, but something further can be said on these points. The view is vague in so far as we cannot picture to ourselves in a mechanical way just how such a system could bring about the suppression of growth in one region and allow the maximum amount in another region. But this is asking too much, since the hypothesis can only claim, at present, to furnish a means by which we can at least imagine what sort of a process is involved, and cannot give the details of the process itself. It can be shown experimentally that if the phenomenon is one of tension certain results should follow that are observed to take place, as when by keeping the shorter half of the planarian from reuniting to the larger half, or by breaking the union if it has been formed, a head develops also at the posterior cross-cut. In the second place, although we cannot understand how the rearrangement of the tensions in a piece takes place, yet from a causal point of view we can see how a change in one region of a labile system may produce, by means of a change of tension, a complete rearrangement of the parts throughout. It can even be claimed for the tension hypothesis that it at least becomes easier for us to see how such a change could take place, because it represents the organization as the expression of a system under tension, and hence, if the material is sufficiently flexible, a readjustment will probably take place when the system is changed in any region. It enables us to see how the organization of the egg may be divided by every cell division, and yet after the reunion of the cells the original equilibrium be established. We may perhaps claim, therefore, that in these respects the hypothesis does give us something more than do the facts; and, inasmuch as it brings a large number of phenomena under a common point of view, the idea may be worth further consideration.

In conclusion, I may add that the hypothesis is, I hope, also a legitimate one, in the sense that being within reach of an experimental proof or disproof, it may serve at least as a working hypothesis. Perhaps more fundamental than the idea that a system of tensions exists throughout the organization is the conception that the organization is itself a system of interrelated parts, and not a homogeneous substance or a mass composed of a large number of repeated parts, or rather, despite the presence of smaller, repeated units, the organization is not the result of their interaction, but of their regular arrangement as parts of a whole structure. If, then, this interrelation of the different parts of the structure can be looked upon as the result of a system of tensions, we can at least form a better idea as to how a piece of a whole can readjust itself into a new whole of smaller size. And it is this possibility of rearrangement or regulation that is one of the most characteristic properties of living things.