Physiological Foundations of Behavior

This relation exists not only for the main stem, but also to a greater or less degree within each subordinate axis, e. g.,a lateral branch. In - like manner the root tip dominates levels below it within a certain range and lateral root tips dominate their own axes to some degree. This relation of dominance and subordination also appears more or less clearly in various other plant organs or organ complexes. Certain characteristics of this relation in plants are of particular interest here. In the first place the relation is not simply an apparent or formal relation resulting from stereochemical or other structural features of the plant protoplasms, but it is a real physiological control mvolving energy changes. The experimental evidence indicates that dominance is associated in some way with the liberation of energy in living protoplasm and it appears at present to be true that the region or part in which intensity or rate of energy-liberation is highest dominates other regions of lower intensity or rate. In fact the relation of dominance and subordination in the plant is funda-

mentally associated with, and determined by the physiological gradients. The high end of the gradient dominates other levels and in partial gradients, 7. e., isolated pieces, the highest level present dominates lower levels. The experimental demonstration of dominance in plants is most readily accomplished through physiological isolation, 7. e., through the removal of subordinate parts from the influence of a dominant part and observation of their behavior following such removal. The various methods and the results of physiological isolation are considered below (pp. 151-160), but it may be pointed out here that the effects of the trimming and pruning of plants are dependent primarily upon this relation of dominance and subordination in the plant axes.

The existence in postembryonic stages of most axiate animals of a functional relation of dominance and subordination of nervous or transmissive character is a familiar fact. In general, the apical region or head, or more specifically the central nervous system and the sense organs of the apical region or head, dominate other levels to a greater or less degree by means of nerve impulses, and we can also distinguish the relation of dominance and subordination at the various postcephalic levels of the central nervous system. In similar manner the central nervous organs of a particular level of the longitudinal axis constitute to some degree functionally dominant regions of the minor body-axes. In annelids and arthropods, for example, the segmental ganglia, situated in the median ventral region exercise a functional nervous dominance over lateral and dorsal regions of the segment. In vertebrates the median dorsal spinal cord similarly dominates lateral and ventral regions. So far as our knowledge goes, such dominance and subordination are always definitely related to the chief physiological gradients of earlier developmental stages, the high end of a gradient becoming the dominant region of that particular gradient.!

* But in the brains of the higher vertebrates there appears to be considerable modification of the primary axial gradients resulting from the development of numerous “centers” or regions of secondary dominance, the correlation centers. The progressive elaboration of these centers has been reviewed by Herrick (’24, Chaps. X VII—XIX) who expresses the opinion that in their more highly developed forms, and notably in the cerebral cortex, these correlation centers become the regions of highest dominance of the body. Herrick’s account includes a discussion — largely theoretical—-of the probable physiological factors involved in the transfer of the region of highest dominance from the periphery in primitive reflex arcs to the deep-seated cerebral cortex of higher mammals.

Functional dominance and subordination also appear in many _ axiate organs and organ complexes and here, as in the more general relations, the high end of a physiological gradient becomes the dominant region. The ctenophore plate row has been discussed in earlier publications (Child, ’17 ¢, ’21 a, pp. 212-220) and I have called attention to the fact that the relations of the plates in the row constitute an almost diagrammatic example of dominance and subordination

in relation to a physiological gradient. Alvarez and his co-workers ! have demonstrated the existence of the physiological gradient and of a functional relation of dominance and subordination in stomach, small intestine and colon. Hyman (’21) has found that the embryonic vertebrate heart is a physiological gradient and that the sinus region which later becomes the pacemaker represents the high end of this gradient. Gowanloch (’23) has been able to modify the development and even to reverse the direction of beat of the heart in fishes by means of differential susceptibility.

Undoubtedly in other axiate organs, such as the ureter (Satani, 19, Penfield, ’20), the fallopian tubes, the contractile blood vessels of various animals, etc., functional dominance and subordination are similarly related to physiological gradients. In “The Origin and Development of the Nervous System,” Chap. XIII, it was pointed out that the reflex are itself is fundamentally a relation of dominance and subordination, that the development of the reflex arcs in the individual is evidently related in a definite way to the physiological gradients and their developmental modifications, and finally that the gradients as excitation-transmission gradients constitute the general physiological basis out of which the reflex are arises, through functional specialization and morphological differentiation.

Dominance and subordination, however, appear in animals, not only in the special functional relations of postembryonic stages, but also as factors in growth and differentiation of parts, agamic reproduction in nature and the regulatory reconstitution of new individuals from isolated pieces. Some fourteen years ago attention was called to the existence of a relation of dominance and subordination in the hydroid Tubularia (Child, ’07 a, ¢, 712 b) and Rand’s work on hydra (Rand, ’11, 712) led him quite independently to recognize the dominance in organization and development of the apical body-region in that form.

In the reconstitution of isolated pieces of Planaria it has been found that a piece from any level of the body is capable of determining the development of all parts posterior to its own level, even though it remains completely headless. On the other hand, if it remains headless, such a piece never develops any trace of parts anterior to its own level in the body, but if even a rudimentary head develops, then all parts normally characteristic of the regions intermediate between the head and the level from which the part was taken develop.! In fact the study of Planaria has demonstrated beyond question, first, that the head region, or more properly speaking, the high end of the major physiological gradient in the body ‘of Planaria determines the organization and development of the parts and organs of levels posterior to it, so far as these are not already present, and second, that the development of a new head on an isolated piece is not, physiologically speaking, a replacement of missing parts by the piece, but actually occurs in spite of the piece. That is to say, in order that a new head shall develop on a piece, the cells at the anterior cut surface, which give rise to the head must be active enough as compared with the rest of the piece to be physiologically independent of it. If these cells are dominated by other parts of the piece, no head develops. Consequently, as repeated experiment with various agents has shown, the head-frequency (7. e., the frequency of head development in pieces of like size from animals of like size taken from a standardized stock), may be increased by conditions which inhibit the metabolism of other parts of the piece. Similarly head-frequency is decreased when the rate of metabolism of other parts of the piece is increased in relation to that of the head-forming cells.

These experiments have led to the conclusion that the reconstitutional development of such pieces is essentially similar to embryonic development, in that determination and control in the organization of the new individual proceeds from the anterior end posteriorly. The formation of the new head on an isolated piece is not determined by the piece but is a “self-differentiation,” The only effect other parts of the piece can have on head-formation is a negative, inhibitory effect. The head, once established as a region of high metabolic rate, determines the organization and differentiation of levels posterior to it, so far as these are not already present.

It has long been known to students of regulatory development in the lower an‘mals that in isolated pieces of Planaria the missing parts anterior to the level of the piece do not develop in order proceeding anteriorly, with the head appearing last. The fact is that, no matter what the level from which the piece is taken, the new head begins to develop first at the anterior cut surface of the piece and other parts arise either by redifferentiation of the old parts posterior to the cut - surface or by later growth of new tissue posterior to the new head

(Figs. 129-131). This is true not only for Planaria but for many if not for all forms in which the regulatory development of new apical regions or heads occurs on isolated pieces. For example, in the hydroid Tubularia, whatever the level at which the stem is cut, the new hydranth arises at the apical end of the piece (see Figs. 67-69, 91-98). In the case of the earthworm Morgan (’97, ’02) observed that when more than five segments were removed from the anterior end only five segments regenerated. Similarly, in various limicolous oligochetes only a certain number of segments characteristic for each species develops anew at the anterior ends of species when more than that number of segments are removed (Hyman, ’16 a). In the species studied by Hyman the intermediate regions are formed by later redifferentiation of some of the old segments. In pieces from the more posterior levels of the body then the new head is “out of place” since the parts which normally lie between it and the level of the piece on which it develops are not present. If we regard the development of the new head as determined by the piece these facts remain inexplicable, but as soon as we recognize that the new head is, so far as the rest of the piece is concerned, a self-differentiation and that the region from which it arises dominates levels posterior to it because it represents the region of highest metabolic rate in the piece the difficulty disappears. In regulatory, as in embryonic development, the so-called law of antero-posterior development holds, and the apical end or head constitutes the first step in the formation of a new individual.

The chief physiological difference between the reconstitution of a piece and embryonic development is that in the former case the development of a new posterior end is determined by the regions already present in the piece even in the absence of a head, while in embryonic development the whole individual is determined from the apical or anterior end posteriorly. That the apical or cephalic portions of the central nervous system develop independently of other parts is indicated by various facts. In the case of Planaria, Flexner (98) and Keiller (710) showed that the

ganglia of the new head may develop quite apart from and independ- — ently of the cut ends of the nerve cords in the piece, the connections being established later by outgrowth from the new ganglia. The fact determined by Goldfarb, (’09), that in the earthworm a new head and new ganglia will develop at the anterior cut surface even after the ventral cord is removed from several segments posterior to the end, also indicates the independence of the cephalic ganglia. The experiments of Fischel (710, ’14) and Waelsch (’14) indicate that in amphibia and birds the earlier embryonic stages of the central nervous system arise independently of other parts. Moreover, Spemann (’18, ’21) and Mangold (’20) have shown that the region anterior to the dorsal lip of the blastopore in the amphibian embryo, 7. e., the region which plays a large part in the formation of the cephalic nervous system, not only develops independently of other parts after a certain stage, but dominates or determines their development. Spemann, in fact, has called this region the “organizer.”

Peripheral parts may influence the development of the central nervous system in later stages, at least in the higher animals as Shorey (09) has shown, and the incomplete or retarded differentiation or small size of certain nervous regions or centers in the absence, incomplete development, or lack of functional connection of peripheral parts normally associated with them is a general rule in the higher animals. Levi has found a correlation between the size of nerve cells and the size of the animal, 7. e., supposedly the size of the area innervated by each cell, and Terni has demonstrated this correlation experimentally.!’ Nevertheless, all the evidence at hand indicates that in the lower animals and in the earlier stages of the higher animals the central nervous system develops independently, or almost independently of other parts, and the apical or cephalic portion of it independently of other levels of the nervous system itself (see, also, Child, ’21 a, pp. 125-128).

It was noted above (pp. 101-110) that the localization of organs along an axis may be experimentally altered by altering the length or steepness of the gradient. At present it seems possible to interpret these facts only in terms of dominance and subordination in relation to a physiological gradient. The facts, considered from the physiological viewpoint, permit little doubt concerning the significance of dominance and subordination in both embryonic and regulatory development and in both animals and plants. There is, in fact, good ground for believing that the re-

~ lation of dominance and subordination is a necessary expression and consequence of the existence of a physiological gradient. Moreover, a physiological continuity apparently exists between the relation of dominance and subordination in development and that which appears in later stages as a characteristic feature of nervous function. Thus we reach once more the conclusion that between the simple physiolog- _ ical gradient, the localization and differentiation of organs and the complexities of nervous control a physiological continuity exists.

We must of course expect to find that dominance and independence go hand in hand, 7. e., that so far as one region or level of a gradient or physiological axis dominates another, it is independent of that other. This fact is strikingly illustrated by normal embryonic development, in many agamic reproductive processes and in the regulatory development of isolated pieces. In embryonic development, for example, the organs representing the high ends of the axial gradients develop before other parts (Child, ’15 c, pp. 67-72). In processes of budding in plants, the apical end, the growing tip of each particular axis, is the first part of that axis to be established (Child, ’15e, pp. 83-87), and in budding and fission in animals, as in embryonic development, the region representing the high end of a particular axis or gradient develops first. Again in the regulatory development of short isolated pieces of the lower animals such as hydroids, flatworms, etc., only the parts representing the apical or anterior end of the body may appear (Figs. 93-101, also Child, ’15 c, pp. 96-101). In other words, a short piece isolated from any level of the body may transform completely into an apical structure or head, other parts being completely absent. No other region or level of the body has ever been seen to originate in this way in regulatory development, 7. e., in complete independence of other parts. The development of organs representing lower levels of a gradient is always determined by the presence of higher levels. Relative independence of course exists in the gradients of minor axes. The development of the median ventral trunk region in most, if not all, bilateral vertebrates and of the median dorsal region in vertebrates, is relatively independent of the development of lateral and dorsal, and lateral and ventral regions respectively.

In axiate organs as well we find a similar independence in development of the regions representing the high ends of the gradients. Functionally also a given level of a gradient is relatively independent. of lower levels, and the apical or anterior end is therefore the most independent region of all. In the simpler animals this functional independence in relation to the gradient appears most clearly. It has been observed in many invertebrates that regions of the body apical or anterior to a region of excitation are but little affected by it, while regions basal or posterior to it are much affected. If an annelid or a turbellarian be cut in two, the piece anterior to the cut is but little affected and soon recovers its normal behavior, while the levels posterior to the cut are much affected (Norman, ’00). In the higher animals and even in man the functional independence of higher levels still persists to some extent, but the autocratic or oligarchic pattern of relations of the simpler animal and plant axes has undergone modification in the direction of democracy and representative government in the course of evolution (Child, ’21 ¢, pp. 248-267).

Both the Roux-Weismann theory of qualitative nuclear division and the theory of the segregation of formative substances in the cytoplasm of the egg involve the assumption that different cells or cell groups of the embryo go through the developmental processes independently of each other. According to these conceptions, the developing organism is a mosaic of these independent parts, which enter into relation with each other only in later stages. Omitting discussion of the theories, the experimental data show that in some animals development even from the beginning, actually appears to be of this character. In such cases we find that, at least within certain limits, parts of the egg, isolated cells or cell groups of early stages, or organ primordia continue their development with little or no change, 7. ¢., as if the other parts were present.! In some forms this apparent independence of parts already exists to a greater or less extent in the egg at the beginning of development, while in others cells or parts isolated in early stages react to the altered conditions by change in the course of development, but in later stages such reaction becomes more and more narrowly limited. Evidently then development of apparently mosaic character appears to very different degrees and in different stages in different forms. At the

one extreme are forms such as the annelids, mollusks and ctenophores, in which at least certain regions are so fixedly predetermined at the beginning of cleavage that isolation from other parts does not alter their behavior to any marked degree. At the other extreme are eggs like those of the sea urchin, the medusa, ete., in which isolation of parts is followed by extensive regulatory change. In such forms - whole embryos may develop from single isolated blastomeres or other parts of embryos, but with certain limitations as regards size of pieces, regions of embryo represented and stage of development, which differ in different species. Even in such cases a gradual determination or fixation of differences in different regions or cells occurs in the course of development, so that in general as development progresses the regulatory plasticity of earlier stages disappears or becomes more and more limited. In some forms, however, in which metamorphosis involves extensive development of new parts, the larval stages preceding metamorphosis may show a high degree of mosaic behavior, while later stages have a high regulatory capacity. Such relations appear in the polychete annelids, the mosaic character of early development having to do with larval organs.

How are we to account for these differences? In the more extreme mosaic type of development the parts appear to develop, or, more strictly speaking, are able to develop up to a certain point independently of each other and in the regulatory types, physiological correlation between the parts appears to be an essential factor, since the course of development is altered by isolation. A very simple interpretation of this series of facts is possible in terms of physiological gradients. In the more extreme mosaic types of development, e. g., annelids and mollusks, specialization or differentiation at different levels of the gradient has occurred to such an extent that isolation at a certain stage does not alter the course of events during a certain period of development following that stage. It is of interest to note further that in the annelid and mollusk this specialization is apparently established first in the upper levels of the primary gradient, 7. e., the regions which give rise to the anterior regions of the larval body, and particularly the ectodermal regions. In fact, the lower levels of the gradient, which give rise to the ectoderm, mesoderm and entoderm of the postcephalic regions are not so highly specialized. In the annelid these regions remain capable of growth and formation of new segments up to late developmental stages, or throughout life, and even in the adult isolated pieces of postcephalic regions are capable of more or less extensive regulatory development,

often of development of complete new individuals. If specialization and differentiation have any relation to the physiological gradients, we should expect them to appear first or to be more stable at the higher than at the lower levels of a gradient because of the differences in metabolic rate at the different levels. Whenever a sufficient degree of such specialization exists, development will appear to a greater or less extent as a mosaic of independent processes in different cells or regions, 7. ¢., the regulatory capacity of isolated parts will be more or less limited. In annelid and mollusk eggs and in some other forms, in which the upper levels of the polar gradient have already attained such specialization at the beginning of cleavage, early development appears to be largely a mosaic of independent processes. In the egg of the jellyfish, on the other hand, none of the levels of the polar gradient has become fixedly specialized at the beginning of development, consequently in this form early development appears as a correlative process, 7. @., extensive regulation follows isolation of parts. In fact it is only in advanced stages of development of the hydroid that any marked degree of fixity of differentiation takes place and then it is limited to the extreme apical end of the gradient, 7. e., to the hydranth, and even this is capable of a considerable degree of regulation in many hydroid species.

If what has been said in preceding sections concerning the dominance and independence of the high end of the polar gradient and of the apical end or head developing from it is correct, it is evident that the early development of this region must always be of the mosaic type, so far as its relations to other parts are concerned. That is to say, its specialization is independent of that of lower levels of the gradient. According to this view this is the only primary or fundamental mosaic feature of development. Apparent independence of other parts does not mean primary or even actual independence, but merely a certain degree of specialization or differentiation resulting from physiological correlation in earlier stages. Such specialization may limit the regulatory capacity of the parts and make it possible for them to continue their development with little or no change for a time after isolation.

Even in the case of the apical end or head, its independence of other parts does not result from inherent qualities of its cells but merely from the fact that it represents the high*end of a gradient. When the gradient is obliterated, it fails to develop (pp. 104-107) and when new gradients are experimentally localized new apical ends or heads appear. We must conclude then that so far as any determination of the apical end or head by other parts is concerned, development is always a mosaic, at least in its earliest stages, but that the apparent independence of other parts is always secondary and the consequence of their specialization or differentiation as particular levels of a gradient. The fact that some subordinate parts are able to continue their development with little change for a time after isolation does not necessarily mean that they were actually independent of other parts before isolation. It means rather that their ability to react to the altered conditions by alteration in the course of development is more or less limited in range, or that reaction occurs slowly.

The idea of a functional pacemaker, that is a region or part determining or controlling the rate of certain activities in other parts, has long been familiar to physiologists. Physiological investigation of the vertebrate heart and its rhythmic activities has demonstrated beyond a question that a certain region at or near the sinus end of the heart normally controls the rate of beat. If this region is inhibited in its activity, e. g., by cold, or is removed, the function of pacemaker may be taken up by the region adjoining, and by inhibition of this region the function of pacemaker may again be transferred to the adjoining intact region and so on. Similar functional pacemakers exist at the upper end of the stomach, the small intestine and the colon and at the upper end of the ureter, and, as in the heart, the function of pacemaker may be transferred successively to lower levels by inhibition or removal of the normal pacemaker.’ In the case of the heart, the alimentary tract and the ureter, regions or levels normally subordinate may also become pacemakers for a part or even for the whole organ if sufficiently excited, even though the normal pacemaker be still present and functioning.

In the plate row of the ctenophore the apical region is normally the pacemaker (Child, ’17¢, ’2la, pp. 212-220), but when the apical region is inhibited or more basal levels are physically isolated from it, the most apical level present or intact becomes the pacemaker and any level of the plate row may be made a pacemaker by sufficient stimulation. Undoubtedly a functional pacemaker exists in many 1 For the heart see Eyster and Meek, ’21, also Gowanloch, ’23; for the alimentary tract see papers by Alvarez and co-authors in bibliography and for ureter, Penfield, ’20.

other organs, e. g., the elongated tubular hearts and contractile blood vessels of various invertebrates, the Fallopian tubes, the vas deferens. In the case of the heart, the alimentary tract, and the ctenophore plate row it has been shown experimentally that the pacemaker represents the high end of a physiological or metabolic gradient characteristic of the organ concerned. When we turn to the development of axiate organisms, we find that the axes are represented primarily by physiological gradients similar, so far as we can determine, to those characteristic of the heart, the alimentary tract and the ctenophore plate row. Moreover, we find that in such a gradient the high end is dominant and controls or determines conditions at lower levels within a certain range of distance. In fact, the high end of such a gradient is essentially a development pacemaker. To take the case of the plant, for example, the growing tip which is the high end of the chief axial gradient controls the development of buds and branches at lower levels. If this tip is removed or inhibited the growing tip of the next lower bud or branch takes up the function of pacemaker, etc. Except that in the heart, alimentary tract and ctenophore plate row a rhythmic excitation is the conspicuous feature, the relations in these gradients of specialized organs and in the axial gradients of axiate plants and animals are very similar. Moreover, the facts indicate that dominance and subordination in development, as well as in the rhythm of heart beat and peristalsis, is fundamentally a relation based on excitation and transmission. The high end of an axial gradient dominates lower levels, because it differs from them as a more strongly excited region differs from one less strongly excited.

Taking the facts as they stand, they point very definitely to two conclusions. First a pacemaker, whether functional or developmental, represents primarily the high end of a physiological gradient, or a region so strongly excited as to alter a preéxisting gradient or determine a new one. Second, wherever a physiological gradient is established a relation of dominance and subordination, of pacemaker and follower, must exist to some degree and within a certain range. And not only do the facts point us to these conclusions, but current theories of heredity make it necessary to discover or postulate some sort of controlling and ordering factor to account for the orderly, definite and correlated character of individual development. If every cell inherits the whole germ plasm, as Morgan maintains, there are only two ways of accounting for the orderly, definite and relatively constant course and character of specialization and differentia-

_ tion of cells and cell groups in development. Either it must depend upon a relation of physiological dominance and subordination, of pacemaker and follower, which is determined in the final analysis by differences in behavior or reaction of different cells or regions to a differential in environment, or else it is the expression of the control and ordering of physico-chemical factors by a metaphysical factor, which we may call entelechy or something else, as we choose. There is, I believe, no escape from this alternative. The phenomena of individual development are inconceivable without a controlling and ordering factor, a pacemaker of some sort, either physiological or metaphysical, and the evidence points very clearly to the existence of a physiological pacemaker in the high end of the physiological gradient.