The Origin and Development of the Nervous System

According to the conception of organismic integra- tion advanced in the preceding chapters such integration is, at least in axiate forms, primarily excitatory and transmissive in character, and the role played by chemi- cal or transportative factors, hormones, etc., is secon- dary, though of great importance, particularly in the more complex, highly differentiated forms. If the organism were primarily nothing but an aggregation of different chemical substances with the possibility of chemical or transportative relations be- tween them, it is difficult to understand how it could ever become anything else; in other words, how a nervous system with its excitation-transmission relations could arise in it. But living protoplasm is something more than such an aggregation of chemical substances, because it is irritable, that is, it possesses the capacity for excitation and transmission and therefore, as I have pointed out, the possibilities of excitation-transmission relations between different parts exist in it. Most biologists are agreed that the appearance of a nervous system in protoplasm does not involve the origin of a new functional activity different from the fundamental activities of protoplasm in general. The current view

is rather that nervous function develops in some way from the functions of excitation and transmission which are characteristic of all protoplasm. If this view is correct, some sort of physiological continuity must exist between excitation and transmission in general, and development and function of the nervous system. In forms possessing it the nervous system is not only the chief organ of excitation and transmission but also the chief organ of physiological integration, and the physiological factors concerned in its origin in ontogeny must be factors connected with organismic integration in general. Chemical, transportative, or hormonic re- lations throw no light on the problem, and the only logical procedure is to seek these factors in the general, non-nervous excitation-transmission relations. Admit- ting that the physiological gradients originate in excitation-transmission relations and that they repre- sent fundamental features of organismic pattern and integration, the developmental relations between these gradients and nervous structure and function become highly significant.

In the first place the localization of nervous differ- entiation shows a very definite relation to surface- interior pattern. The primary receptors develop at the surface, and the nervous system as a whole is primarily of ectodermal origin and even in the higher animals arises superficially. Even in the neutral tube of verte- brates we find indications of the influence of the surface- interior factor in the localization and direction of out-growth of the earlier axons. These axons arise on the

FIGS. 13-15. — Axial gradients in development of the central nervous system: Fig. 13, embryo of Plagiostomum girardi, a rhabdocoel turbel- larian; longitudinal section near median plane with anterior end at left; the mass of cells anterior to the pharynx represents the cephalic ganglia and one eye. These organs are advanced in development, but postcephalic regions have scarcely begun to develop (from Bresslau, 1904) ; Figs. 14, 15, surface views of two stages in development of chick, showing progress of development in posterior direction. In Fig. 14 the central nervous system is marked off as a distinct region, the medullary plate, only in the anterior head region. In Fig. 15 the neural tube has developed by closure of the medullary folds to a level posterior to the somites, 5. Still farther posterior is the secondary region of growth (from F. R. Lillie, 1919).

sides of the neuroblasts which face away from the central canal, i.e., the surface originally external, and the first outgrowth of the axons is usually away from the central canal, though this direction may be altered by the axial relations (pp. 185-87). In forms in which axiation is but little developed the nervous system may be a more or less diffuse nerve not developing in relation to the body surface with little centralization, as in various coelenterates.

In animals with more highly developed axiate pattern we find the chief receptors and the chief aggregations of nervous tissue differentiating from the high ends of the chief body gradients. This relation is most conspicu- ous with respect to the major or polar gradient, the apical end, or the head, with its special sense organs and nervous ganglia, arising from the high end of the gradient, but it also appears more or less clearly with respect to the symmetry gradients (see p. 149); the general direction of growth and differentiation of nervous structure is down the gradients. Differentia- tion of nervous structure begins at the anterior end and progresses posteriorly (Figs. 13-17) and growth of nerves is in general from central to peripheral regions, particularly in the higher animals. The direction of early axon outgrowth in the vertebrate neural tube, while showing primarily a more or less definite relation to the surface-interior pattern, as already noted, reacts very early to the axiate pattern and the axons grow in general postero-ventro-laterally, that is, down a resultant of the chief body gradients (pp. 185-87).

This definitely directed growth of nerves in relation to the gradients is not limited to the higher animals. A FIGS. 16, 17. — Transverse sections of chick embryo at different levels, to indicate developmental gradients: Fig. 16 indicates the condi- tion of the posterior half or more of the stage of Fig. 14 and of the extreme posterior region of the stage of Fig. 15; Fig. 17 indicates the condition at the level of the sixth somite of Fig. 15; the region of most intense developmental activity is primarily median dorsal, and develop- ment progresses in general laterally and ventrally (from Child, 1915^).

striking example of such growth of nerve fibers in an axiate organ is found in the tentacles of sea anemones. Groselj (1909) showed that the fibers arising from the sensory cells in actinian tentacles grow predominantly in the basipetal direction (Fig. 18), though in some cells the basal outgrowth divides and gives rise to fibers growing in both directions. Parker (1919, p. 132) concludes that the physiological polarity of such tentacles is dependent

on this definitely di- rected growth of the fibers. Undoubtedly this conclusion is cor- rect as regards the difference in conduc- tivity in the two di- rections in the fully developed tentacle, but Parker apparently fails to recognize the fact that the directed growth of the nerve fibers is not a mat- ter of independent choice, but must itself be determined by physiological conditions. From the time of their appearance as buds, these tentacles show a definite physiological gra- dient, with differences in electrical potential, suscepti- bility, reducing power, etc., and the directed growth of the nerve fibers is unquestionably determined by this pre-existing gradient.

While the evidence is by no means complete, all that we have indicates very clearly that the differentia- Fig. 18. — Sensory cells of the sea- anemone tentacle and their outgrowths; the arrow points toward the base of the tentacle (from Groselj, 1909). tion of nervous structure is in general localized, first, with reference to the physiological surface-interior rela- tions, and, second, with reference to the axiate relations as represented by the primary or chief body gradients. The earliest appearance and the chief aggregation of nervous structure is localized at or near the high ends of the gradients present in the organism concerned, and the general direction of nervous differentiation and of nerve outgrowth during the earlier stages of develop- ment is down the chief body gradients. Later of course the situation becomes always more and more compli- cated by the localization of new regions of high activity and new gradients, which may determine locally the direction of nervous differentiation and growth, but the sequence of events is orderly and definite.

The very early differentiation of nervous structure is another striking feature of ontogeny. In the coelen- terates nervous differentiation is so slight that it is difficult to determine when it begins, but in other animals we find that the central nervous system becomes recog- nizable at least as a region of localized growth at the anterior end soon after the establishment of the primary germ layers and before other definitive organs make their appearance (Figs. 13-17).

These various facts appear to me to indicate a very close ontogenetic relation between the primary physio- logical gradients and nervous differentiation, and they afford a physiological basis for various features of nervous development which have heretofore remained obscure. If the primary physiological relations in the pattern of the organism are chemical or transportative and excitation and transmission relations arise only secondarily the localization of the nervous system with respect to the gradients, its early development, and the fact that it becomes the chief integrating organ all remain remarkable and completely inexplicable facts of ontogeny, but if it is true that organismic pattern originates in excitation-transmission relations and that the axial gradients are protoplasmic records of such relations, all these facts appear in a new light. The localization, the early differentiation, and the function of the nervous system, which represents the specialized mechanism of excitation and transmission in the organ- ism, appear as the natural and necessary consequence of the existence and physiological nature of the gradients. Moreover, from this point of view nervous structure appears as the primary or fundamental developmental reaction of the more active body regions. The dis- tinction between ectoderm and entoderm precedes its appearance, but the entoderm represents physiologically the lower levels of the chief axial gradient and the nervous system is the first definite and definitive organ to differentiate from the higher levels of the gradient or gradients.

The older conception of ontogeny as a process of construction of a machine which, after construction is completed, begins to function appears less and less satis- factory as our knowledge advances. Living protoplasm is functioning at all times and development is a process of functional construction, that is, beginning with a given structure and function, the continuance of func- tion modifies the structural substratum, and this in turn modifies further function and so on. If the effect of functional activity on the structural substratum is ever completely reversible it is only or chiefly in the fully developed organism where a dynamic equilibrium between structure and function is attained or approached. Development represents the progress from a disturbed protoplasmic or cellular equilibrium to an organismic equilibrium, or, as in physiologically isolated regions or physically isolated pieces, the progress from a disturbed organismic equilibrium to a new equilibrium, and the metabolic or physiological gradients constitute the framework or ground plan on which this progress takes place.

The primary germ layers represent the primary regional differentiation with respect to the gradient pattern. They result from the differences in rate of metabolism at different levels of the gradient and con- stitute the basis for a chemical or transportative relation which arises sooner or later, when ectoderm grows at the expense of the yolk in the entoderm, or of food digested by the entoderm. This transportative relation between ectoderm and entoderm is, however, secondary and the consequence of the differences in rate of activity between them. The nervous system, on the other hand, repre- sents the primary differentiation with respect to the functions of excitation and transmission of the gradients, and as the gradients represent the primary integrating factor in the development of the organism, the nervous system as the differentiated mechanism of this primary integrating function necessarily becomes the chief organ of integration.

According to this conception the nervous system is the physiological and morphological expression of the excitation-transmission relations, first with respect to the primary or chief physiological gradients, and later with respect to the progressive developmental complica- tions as they arise. The absence of a nervous system in plants apparently depends upon the inability of plant protoplasm to synthesize molecules sufficiently stable under the exist- ing physiological conditions to constitute the basis of a permanent morphological structure of nervous character, that is, possessing a stable protoplasmic structural sub- stratum and at the same time a mechanism permitting intense dynamic activity. In general, the capacity of plant protoplasm for morphological differentiation is slight as compared with that of animal protoplasm, and plant structure consists very largely of cellulose. Con- sequently physiological integration in plants retains more or less its primitive character, based on a gradient in physiological activity, physiological dominance of the apical region, the growing tip, is not very effective, and repeated physiological isolation and reproduction, as evidenced by the development and arrangement of multiple axes in the form of branches and various other parts, is a characteristic feature of the life of most plants.

This conception of the ontogenetic origin of the nervous system is purely physiological and involves no assumptions concerning inheritance or evolution. The gradients do not determine the nature of nervous struc- ture but merely provide the physiological basis for the realization of certain of the hereditary potentialities of the protoplasm concerned and the problem of the evolutionary origin of those potentialities remains as before. We are primarily concerned with the develop- mental physiology of the nervous system and the problem of the phylogenetic origin, and evolution of nervous structure is involved only in so far as it can be analyzed in terms of developmental physiology.

Attention has already been called (p. 97) to the fact that the establishment of a physiological gradient deter- mines the existence of a relation of dominance and subordination in which the high end of the gradient, because of its greater irritability and activity, is physio- logically dominant over all levels of the gradient within a certain range of distance, which varies with the activity of the dominant region and the conductivity of the protoplasm. Similarly, any level of the gradient is to some extent dominant over more basal or posterior levels and subordinate to more apical or anterior levels. This relation of dominance and subordination appears most clearly in the simpler axiate organisms, both plants and animals, and its significance in ontogeny and in the regulatory development of isolated parts as well as in function has been considered at length elsewhere (Child, igioa, igncy d, 1915^, chaps, iv, v).

This relation may be stated not only in terms of dominance and subordination, but in terms of inde- pendence and dependence. So far as one region or level of the axis is dominant over another, it is independent of the other. In other words, in so far as a more irritable or more active region A becomes a factor in determining through transmission of excitation the physiological con- dition of another less irritable or less active region B, in so far is A independent of B. Evidently then the high end of a gradient must be, at least to a large extent, independent of other levels, and in bilateral forms not only must the anterior region be largely independent of more posterior levels, but the median ventral (inverte- brates) or the median dorsal (vertebrates), so far as it represents the high end of the symmetry gradients, must be relatively independent of more lateral regions.

Such a relation, if it actually exists, must appear in the order of development and the functional relations of parts along an axis. The so-called law of anteroposterior development is merely a statement of this relation for the longitudinal axis. Development and differentiation begin anteriorly or apically and progress posteriorly or basally because the anterior or apical region, being more intensely active than other regions, is so far as they are concerned self-determining, at least in large measure, and each body level is also to some extent self-determining with respect to more basal or more posterior levels and at the same time to some extent determined by more apical or anterior levels (Figs. 8-12). Similarly the median region is to some extent self-determining with respect to the lateral, and more lateral levels are to some extent determined by more median. In plants the growing tip arises first in the development of an indi- vidual axis, whether stem or root, and is manifestly self-determined with respect to other parts, since it may arise in their complete absence, and other parts are sub- ordinate to and determined by it.

The apparent independence and capacity for " self- differentiation'1 of different levels or regions of the body in certain animals, e.g., annelids and mollusks, may seem to be completely irreconcilable with the conclusion advanced above, but what the data on " self- differentiation'3 and "mosaic development'3 actually demonstrate is not the absence of relation between parts but merely the fact that protoplasmic specializa- tion has advanced so far that the regulatory reactions to isolation of parts are limited, and the behavior of the part after isolation is therefore much the same as before. The evidence indicates that conditions of this sort are due simply to early cytoplasmic specialization rather than to the fundamentally " mosaic" character of ontogeny.

In normal ontogeny the whole body is potentially present, and objection to the conclusions stated above may be raised on the ground that, even where differen- tiation is not perceptible, it may still be sufficient at lower levels of the axis to influence the development of higher levels. But further evidence in support of these conclusions concerning independence and dependence, some of it demonstrative, is obtained from the regula- tory development of isolated pieces from many of the lower animals. In the first place it has long been known that in the development of pieces isolated from the body of various sp ecies of hydroids, flatworms, and annelids, the apical or anterior region of the pieces gives rise directly to an apical structure or head, irrespective of

the level of the body from which the piece was taken. In Planaria, for example, pieces taken from regions near the original head and those from more posterior levels (Fig. 19, a and b) behave in essentially the same way in this respect. Although the anterior end of piece b represents a body level which is normally separated from the head by the whole pre-pharyngeal and pharyngeal region, a new head develops directly from the anterior end of b as well as from that of a and the body regions corresponding to those between the head and the level of b in the original animal arise later by the reorganization of the tissues of b posterior to the new head (Fig. 22). Moreover, this reorganization never takes place unless a head does develop or begin to develop. In the earthworm and other annelids which are not too highly specialized to give rise to new heads from postcephalic body levels the developmental process is the same in these respects. In fact, in all cases where isolated pieces give rise to new individuals the develop- ment of the new apical end or head is the first new differentiation at the anterior end, irrespective of the body level represented by the piece. That is to say, this process of so-called regulatory development is essentially similar to embryonic development in that differentiation of the new individual begins at the apical or anterior end and proceeds basipetally or posteriorly. In a piece like b (Figs. 19-22) the new head is "out of place" when it arises, i.e., it adjoins parts which in the normal animal are far removed from the head. The normal relations are re-established by the redifferentia- tion into the various intermediate regions of parts of the piece posterior to the new head. The significance

FIGS. 19-22. — Planaria dorotocephala: reconstitution of pieces from different levels: Fig. 19, diagrammatic, showing general plan of central nervous system and alimentary tract, a and b, two regions from which pieces are taken; Fig. 20, early stage of reconstitution in both a and b; Fig. 21, later stage from a; Fig. 22, later stage from b (Figs. 19-21 from Child, of these facts has been discussed elsewhere and it has been shown experimentally that the development of the regions normally intermediate between the head and the level represented by the piece is determined by the presence of the developing head (Child, 19116, 1915^, pp. 102-17). Similarly, the development of a new posterior end at the posterior end of the isolated piece is determined by the regions anterior to it and cannot occur independently of them, though it may occur in the absence of a head (Figs. 8-12). The development of the new head, however, actually occurs in spite of the rest of the piece and may under certain conditions be more or less completely inhibited by the rest of the piece (Child, 1914^, e, 19166, 1920).

The most striking evidence for the physiological independence of the apical end or head is found in the regulatory development of very short pieces. Such pieces give rise to apical or cephalic structures alone, all other parts of the body being entirely absent. Short pieces of Tubularia stem (Figs. 23-29) from any level may give rise to single or biaxial whole hydranths or partial hydranths depending on its length, and short pieces of the planarian body may give rise to single or biaxial heads (Figs. 30-34). The frequency of biaxial structures in these short pieces is due to the fact that the short piece, representing a very small fraction of the polar axial gradient, is almost apolar, and in many cases the exposure and the presence of the wound at the two ends is sufficient to determine two opposed gradients and so a biaxial development. These apical and cephalic structures which develop in the complete absence of other parts may be complete and normal in all respects

so far as can be determined. The single or biaxial planarian heads may contain fully developed cephalic FIGS. 23-29. — Tubularia: Fig. 23, an unbranched individual; Figs. 24-29, single and biaxial apical structures which develop from short pieces of the stem (from Child, ganglia. In such cases there can be no doubt that these apical or cephalic structures have arisen in com- plete independence of other parts of the body. Thus far no other region of the body has ever been seen to develop from pieces in this way. All other levels of the body along the polar axis develop only in relation to more apical or more anterior regions. In Planaria, for example, we never find small pieces transforming completely into a new posterior end, although pieces may

FIGS. 30-36. — Partial structures from short pieces of Planaria: Figs. 30-33, monaxial and biaxial heads from P. dorotocephala; Fig. 34, biaxial heads from P. simplicissima; Fig. 35, biaxial tail from P. doroto- cephala; Fig. 36, biaxial tail from P. simplicissima (Figs. 30-33, 35, from Child, give rise to new posterior ends on one (Figs. 11-12) or on both cut surfaces (Figs. 35-36) in the absence of a head. All the data both from normal ontogeny and from experimental or regulatory reproduction indicate that the apical region or head, at least as regards the earlier

developmental stages, is independent of other parts. In all except the simplest animals the cephalic central nervous system is the primary organ of the head region. If the head region is self-determining with respect to other parts, certainly the same must be true in large measure for the cephalic nervous system. In this connection it is of interest to note that Flexner (1898) and Keiller (1910) have shown that in the regenerating head of Planaria the ganglia may develop quite independently of the cut ends of the nerve cords in the piece and the connections are established later by outgrowth from the new ganglia rather than from the nerve cords. The development of the connections between the new ganglia and the old cords in certain abnormal types of head in Planaria also points to the same conclusion (Child and McKie, 1911). Moreover, Goldfarb (1909) removed the nerve cord from several segments at the anterior cut end of pieces of earthworm and found that the head and the ganglia developed nor- mally in such pieces, and the nerve cords grew poste- riorly from the new head into the old parts, although some regeneration in the anterior direction from the end of the old cord occurred. In this case it is demon- strated beyond question that the ganglia in the regener- ating head develop independently of other parts of the nervous system in the piece.

There seems then to be no escape from the conclusion that, at least in the lower animals and as regards its earlier stages, the apical or cephalic portion of the cen- tral system develops independently of subapical or postcephalic regions. Its physiological dominance and its independence are simply two aspects of the same physiological situation and both are consequences of the fact that this portion of the central nervous system develops from the high end of the primary gradient, i.e., from the region of the body in which activity is most intense during the earlier stages.

The experiments of Waelsch (1914) and Fischel (1914) on birds and Amphibia indicate that in these forms also the earlier embryonic stages of the central nervous system arise independently of other parts. Peripheral parts undoubtedly do exercise some determining influence on the later differentiations in the central nervous system of vertebrates, as is shown by the experimental work of Shorey (1909) on birds and Amphibia and numerous data on the retardation or inhibition of development of central organs in the absence of peripheral function in man and mammals. This, however, is unquestionably a secondary condition and associated with the decrease in "steepness" or decrement in the chief axial gradients with the advance of development. In earlier stages, and in many of the lower organisms throughout life, the difference in irritability or metabolic condition is so great that transmission is chiefly or wholly in one direc- tion, i.e., down the primary or chief gradients because most excitations cannot run very far up the gradients without disappearing as excitations in the higher metabolic level (see pp. 97-99). With the ontogenetic and phylogenetic increase in conductivity along the gradients and the consequent decrease in decrement, excitation may be transmitted to greater and greater distance up a gradient, and peripheral parts or lower levels of a gradient become increasingly important factors in determining differentiation and physiological

condition at higher levels. In other words, the primi- tive type of axiate physiological integration is similar to the autocracy in social integration, but in the devel- opment and evolution of the organism, as in social development and evolution, there is a gradual approach toward democracy. While the evidence is less extensive, the data of normal development and of experiment indicate that the ontogenetic relation of the postcephalic regions of the central nervous system to regions lateral and dorsal (invertebrates) or ventral (vertebrates) to them is similar to the relation of the cephalic to the post- cephalic regions. These postcephalic regions of the central nervous system apparently arise from the high ends of the symmetry gradients (Figs. 13-17), and so far as these regions are dominant they must be primarily independent of the parts which they dominate.