The Origin and Development of the Nervous System

It is a fact of considerable theoretical interest not merely to the biologist but to the psychologist as well that the cephalic region of the central nervous system arises from that region of the body which primarily is most intensely active and most completely independent of other parts. In fact the development of the central nervous system and particularly of its cephalic portion represents that physiological and morphological reaction of animal protoplasm which is least complicated by physiological correlation with other parts and most directly related to external exciting factors. The cen- tral nervous system arises from the higher levels of the chief axial gradients in the organism, and these represent

the dominant regions of the physiological gradients which are present before the nervous system appears. I have endeavored to show that these gradients are in the final analysis responses to the differential action of external factors upon living protoplasm, and if we admit this it follows that the localization and the general pattern of the centralized and cephalic nervous system is deter- mined physiologically by a response of protoplasm to environmental action upon it. Moreover, the central nervous system is the chief organ of excitation and trans- mission and the chief organ of physiological integration, because it represents the working out in ontogeny and phylogeny of the primary organismic integrating factors, the excitation- transmission relations.

The special characteristics of nervous structure as distinguished from those of other organs are of course determined by the hereditary constitution of the proto- plasm concerned. As I have already pointed out, the physiological gradients merely provide a framework or plan as a basis for the realization of certain hereditary potentialities, and with evolutionary changes in the hereditary mechanism, this physiological plan, and therefore the localization, structure, and relations of the nervous system, undergo alteration. But there can be no doubt that the earlier stages of the development and differentiation of the nervous system are in large meas- ure independent of other organs, and I have endeavored to show that this independence is due to the fact that the nervous system, more than any other organ, is the expression of the primary physiological factors of organismic pattern and integration.

The bodies of the annelids and arthropods among the invertebrates and the bodies of all vertebrates are made up of segments, metameres, or somites, and the problem of the origin and the evolution of this segmentation or metamerism has ranked as one of the great problems of speculative morphology, while its physiological aspects have received but little attention. Without attempting a survey of the various theories of segmentation it may be noted that they involve two different conceptions of the segment and of the process of segmentation. According to the one conception the process of segmentation is essentially a process of agamic reproduction, the segment being equivalent to a new zooid, but instead of becoming a complete new individual and undergoing separation, it is co-ordinated or inte- grated with the head region and other segments into a new individuality. In short this theory conceives the process of segmentation as a modification of certain processes of agamic reproduction, such as the repeated transverse fissions which occur in certain Turbellaria and give rise to temporary chains of zooids.

The other conception regards segmentation as originating in the reduplication of certain internal organs, and most of the theories based on this conception hold that the process is primarily localized in the mesoderm and that the reduplications in other parts are secondary. It is evident that as a physiological process segmen- tation, as well as many other repetitive developmental processes, must involve a rhythmical repetition in physiological conditions of some sort. Ritter and Johnson (Ritter, 1909, 1919 [II, 95-107]; Johnson, 1910; Ritter and Johnson, 1911) have called attention to the significance of these repetitive processes as evi- dence of law and order in organisms and to their bearing on the problem of integration. I have shown that various processes of agamic reproduction in both plants and animals result from physiological isolation (Child, 1910^, 191 id, 1915^, chap, v; Child and Bellamy, 1919) and have suggested that the physiological conditions determining the reduplications of parts, or meristic series, as Ritter terms them, are essentially similar, except that they concern more or less specialized parts and their repeated development rather than entire organisms. If this suggestion is correct there is little fundamental difference, so far as the physiological con- ditions are concerned, between the two chief theories of segmentation. In the one case the process of develop- ment of a new segment presumably begins in the ecto- derm, since this is primarily the most active layer, and later involves the other layers; in the other case the process begins in the mesoderm and secondarily involves other layers.

The theories of segmentation have been primarily morphological rather than physiological. When we con- sider segmentation as a physiological process, as it of course must be in even' case, it is difficult to find an adequate physiological basis for the theory of the mesodermal origin of segmentation. Even if reduplica- tion of mesodermal organs did occur, it is difficult to understand how or why it should have led to the segmen- tation of ectodermal structures and even of the nervous system. Moreover, many of the lower invertebrates, e.g., various Turbellaria, show reduplication of certain mesodermal and of entodermal parts without any indica- tion of actual segmentation in muscles, nervous system, or ectoderm in general. The theory of mesodermal segmentation regards such forms as the precursors of segmented animals, but what evidence we have seems to me to point to a different conclusion.

In the polychaete annelids, so far as they have been studied in this way, the earliest stages of development show a single apico-basal gradient in susceptibility, but at a stage much earlier than the beginnings of segmen- tation in the mesoderm an ectodermal region of high susceptibility, indicating high physiological activity, appears in the posterior part of the developing larvae (Fig. 37) and from this region the ectoderm of the trunk arises (Child, 191 7 d). The appearance of this region resembles physiologically the appearance of a second zooid in the posterior region of Planaria (Child, 191 id, 19136), and it seems probable that it is determined by some degree of physiological isolation in much the same way as in Planaria, except, of course, that it is much earlier in development. This region is already estab- lished as a growing region when the mesoderm consists of only a few cells and the formation of successive seg- ments from it is evidently a repeated reproductive process.

We may conceive the process somewhat as follows: the physiological isolation of this posterior region of the developing organism results first in increased cell division and growth, i.e., it becomes or remains less specialized instead of contin- uing to differentiate, like other parts. This is the same pro- cess that we find in other cases of physiological isolation. Second, in this mass of em- bryonic cells the development of a new individual begins, but in this case the individual becomes a segment instead of a complete organism. Its failure to develop a head may be due either to incomplete isolation from the head region or to the presence of some degree of specialization in the cells concerned which makes them incapable of giving rise to a head.

the lower invertebrates, e.g., su^epp£bility. the flatworms, the first limi- tation in the power of regeneration of pieces of the body with advancing specialization is the loss of capacity to give rise to a new head. In the triclad turbellarian, Planaria, for example, pieces above a certain length from any level of the body are able to regenerate heads Fig. 37- — Early larval stage of Chaetopterus, a polychaete annelid, showing the primary apical and the secondary poste- rior regions of high susceptibil- ity as they appear when the cells begin to swell and separate in various agents; at this stage the mouth is developing and

as well as other parts and become complete new indi- viduals (Figs. 19-22). In the polyclad Turbellaria, however, pieces from levels posterior to the head are unable to develop new heads, though they regenerate all other parts. In certain other forms, e.g., the triclad, Dendrocoelum (F. R. Lillie, 1901) and various oligo- chaetes (Hyman, 1916), the capacity to regenerate a head exists in the more anterior body regions and dis- appears at a certain level. In one of the rhabdocoeles, Stenostomum, the capacity to regenerate a head exists in the younger asexual stages but disappears when the animals become sexually mature. Moreover, in embry- onic development the head arises earlier than other parts, i.e., when the cells are physiologically younger and their rate of physiological activity higher, and it is localized in the most active region of the organism. And finally, the earlier stages of head development are more susceptible to the action of inhibiting agents than those of other parts of the body. All these facts indicate that a higher rate of metabolic activity in the cells con- cerned is necessary for the development of a head than for that of other parts, i.e., in order to produce a head the cells must be physiologically younger, or must arise from a higher level of the axial gradient and therefore undergo more complete dedifferentiation and attain a higher rate of activity, than is necessary for the develop- ment of other parts.

Returning to the process of segmentation it appears then not improbable that the cells of the posterior growing region in annelids have already become special- ized to such an extent that they are capable of giving rise only to incomplete headless individuals. In any case the segment resembles a headless individual of Planaria (see Figs, n, 12). The process of growth and differentiation is repeated for each segment, but instead of becoming free and developing heads, like the zooids of Planaria, the segments are integrated into a new whole, probably chiefly through the development and increase in conductivity in the nervous system.

The early appearance of a posterior growing region involving primarily the ectoderm and giving rise to the segments has been demonstrated not only in the poly- chaetes but in the oligochaetes (Hyman, 1916), various Crustacea and fishes, and Bellamy (1919) has shown that it appears in the frog at an early stage (Figs. 38, 39) and that through the action of inhibiting agents its localization can be altered so that it arises nearer the apical pole of the egg than normally, much as the anterior end of a new zooid can under inhibiting condi- tions be localized nearer the head in Planaria and various other forms (Child, 191 id). In unsegmented animals a definite posterior growing region of this sort is usually not present, except in cases of agamic repro- duction, though even in these forms more or less growth may occur at the posterior end as they increase in length. These facts show at least that the posterior growing region is determined in the ectoderm earlier than in the mesoderm and, I believe, increase the probability that segmentation as a physiological process begins in the most active layer, the ectoderm, and that it is physi- ologically a process resembling various processes of agamic reproduction.

That the central nervous system of segmented animals shows more or less definite indications of being itself segmented is a fact of common knowledge. In the less highly specialized segmented invertebrates a bilateral pair of ganglionic enlargements of the ventral nerve cord, or a single ganglionic mass with more or less FIG. 38. — Unsegmented egg of frog, viewed from apical pole, show- ing susceptibility gradient to temperature of o° C. The stippled areas represent areas of disintegration, the cross-marks the apical pole, the dotted circle the boundary between pigmented and unpigmented regions, and the dotted crescent the outline of the gray crescent, as they appear on the lower hemisphere (from Bellamy, 1919).

distinct double bilateral structure, appears, but the vari- ous modifications of segmental structure in the different groups are accompanied by modifications of ganglionic arrangement. This is particularly the case as regards the head region which, in the higher invertebrates, apparently consists of a number of more or less com- pletely fused segments, and in the vertebrates of a still larger number. Even in the oligochaete annelids Hyman (1916) has found indications in the behavior in

FIG. 39. — Stages of disintegration of frog embryo during late cleavage, in side view; the secondary posterior region of growth is already appearing as an area of high susceptibility in the equatorial region (from Bellamy, 1919). regulatory development of a number of anterior seg- ments, characteristic for each species, that they con- stitute a more or less definite head region. The problem of the segmentation of the vertebrate head has been widely discussed and an extensive litera- ture exists which cannot be considered here. The

segmentation of the cephalic, and in fact of the whole central nervous system of the vertebrate, appears in the adult in the arrangement of cranial and spinal nerves and spinal ganglia and to some extent in certain parts in the arrangement of centers or nuclei of the nerves. Concerning the embryonic development of this segmen- tation and its relation to the segmentation of other parts there has been much difference of opinion, centering chiefly about the question whether the segmentation of the nervous system is primary or secondary, i.e., whether segmentation is primarily an ectodermal or general reproductive process or a mesodermal process. Like the evidence on segmentation in general the evidence con- cerning the segmentation of the vertebrate nervous system seems to me to support the former alternative. The discovery by Locy (1895) of distinct neuromeres in very early stages of the embryonic development of vertebrates, and the later work along this line by Hill (1900) indicate very clearly the occurrence of some sort of repetitive process in the primordium of the central nervous system as it arises (Figs. 40-44). Neal (1898, 1918) among others has criticized this evidence and has attempted to show that the neuromeres of early stages do not correspond to the final segmentation, but in spite of these and other criticisms the physiological significance of Locy's neuromeres appears beyond ques- tion as indicating a rhythmical repetitive process of some sort along the anteroposterior axis in very early stages.

The neuromeres of the head region appear of course in that portion of the embryo developing from the primary region of growth which has its anterior end at FIGS. 40-44. — Embryonic segmentation of the central nervous sys- tem of vertebrates: Figs. 40, 41, Squalus acanthias, dogfish, surface view of young embryo and side view of nervous system of later stage (from Locy, 1895); Figs. 42-44, chick; Fig. 42, dorsal view of embryo with one somite, twenty-one hours old, neural segments traceable, not only across wide neural groove but some distance laterally into adjacent tissue; Fig. 43, dorsal view of anterior portion of neural tube of embryo twenty-two hours old; Fig. 44, right surface view of same neural tube as Fig. 43, showing that segments encircle neural axis (from Hill, 1900).

or near the apical pole of the egg. That is to say, they are wholly anterior to the secondary region of growth from which the postcephalic segments arise, and it has been suggested above that the appearance of this secondary growing region is the result of physiological isolation. The presence of the cephalic neuromeres indicates the occurrence of a rhythmic, repetitive process in the nervous primordium even in this anterior region. If we regard the cephalic neiiromeres as true segments, we account in evolutionary terms for their appearance by assuming that they represent the remain- ing traces of an ancestral segmented condition of the cephalic region. So far as I am aware, however, the problem of the physiological process involved has not been considered. Obviously it is a true repetitive or reproductive process in the cephalic primordium. There is little prospect of obtaining actual experimental knowledge concerning this process until physiological technique is much more highly developed than at present, but it may be suggested that the increasing physiological activity and its progress in the posterior direction in the embryonic primordium are the chief factors in the repetitive process of neuromere forma- tion. As each level of the primordium attains a certain degree of activity it becomes to some slight extent physiologically isolated from more anterior levels and some slight amount of localized growth occurs, but the degree of isolation is so slight that it does not proceed very far.

According to this point of view the origin of the secondary posterior growing region in the embryos of segmented animals and the development of the trunk segments from it are physiologically merely a con- tinuation in later developmental stages of a process essentially similar to that which in the very early stages gives rise to the cephalic neuromeres. Moreover, the very early appearance of indications of segmentation in the primordium of the nervous system, which is ecto- dermal, favors the conclusion that segmentation is primarily an ectodermal rather than a mesodermal process and this conclusion again is in accord with the fact that the ectoderm is physiologically the most active body-layer in the earlier embryonic stages and the layer in which the earliest differentiations appear. The mesodermal theory of segmentation is based pri- marily on morphological speculation rather than on physiological experiment, and it seems to me to offer little prospect of interpretation or synthesis of the various facts as they now stand, and particularly of the very early participation of the ectoderm and its earliest definitive differentiation the nervous primor- dium in the repetitive process. The facts apparently indicate that segmentation begins in the ectoderm and that the processes in this the most active and most advanced body layer affect conditions in other layers. The segmental ganglion is then the primary definitive organ of the segment as the cephalic ganglion is of the individual.

The terms " centralization'2 and " central nervous system" possess of course a morphological and a physio- logical significance. Morphologically, centralization is the process, both ontogenetic and phylogenetic, of localization and aggregation of the chief mass of nervous tissue in a particular region of the body. Speaking physiologically, centralization represents integration, and the central nervous system is really the integrative nervous system, as Sherrington (1906) has so beauti- fully demonstrated. The term "cephalization" refers to one particular aspect of centralization, viz., the con- dition or process of integration with respect to the longitudinal axis. The point to which I wish to call particular attention is that these terms refer not merely to a morphological localization or aggregation, but to a physiological condition and process, and also to an evolutionary process. Morphological centralization and cephalization in any particular case are at best only the indication or record in structure of the degree of integra- tion attained. Moreover, a high degree of nervous centralization or cephalization may exist in compara- tively simple forms with relatively slight degree of integration as compared with the higher animals, e.g., ascidians, rhabdocoele and polyclad Turbellaria in which the central nervous system is entirely in the head.

The aggregation of the whole central nervous system in a single ganglion or pair of ganglia, as in these forms, does not of course mean that these animals have attained a higher degree of integration than the vertebrates, but merely that the degree of integration is higher than would be the case in similar forms with a more diffuse system. In the adult ascidian the function of the ganglion as an integrating organ is certainly not highly developed, but the ascidian tadpole represents a rather high degree of centralization and the high degree of morphological centralization in the adult appears to be rather the consequence of the earlier course of develop- ment than directly associated with conditions in the adult. The degree of integration in the rhabdocoeles and polyclads is unquestionably much higher than in the triclad Planaria, where the central nervous system includes not merely the cephalic ganglia but the longi- tudinal cords and apparently some of their branches (Fig. 19, p. 1 22) . The degree of integration corresponding to a particular degree of centralization or cephalization in a particular group depends upon the hereditary poten- tialities of the protoplasm concerned as regards the development of nervous structure and function.

According to current views, the nervous system is primitively diffuse and uncentralized and centralization appears only in the course of evolution. I have endeav- ored to show that the localization and differentiation of nervous structure, and the development of nervous func- tion constitute only a part of a developmental process, that they must depend upon physiological conditions already present in the developing organism before the nervous system as a specialized organ appears, and that the physiological gradients as the earliest distin- guishable features of organismic pattern are fundamental factors in the process which determines the localization and development of a nervous system.

From this viewpoint the nervous system appears as a morphological and physiological expression of the gradients in the developing organism, and both its early differentiation and its function as an organ of excitation and transmission indicate its intimate relation to these fundamental factors of organismic pattern. If we accept this view, we can expect to find an uncen- tralized nervous system only in animals which are completely anaxiate and which possess only a surface- interior pattern, and it is difficult to conceive how such a system could contribute in any way to organismic integration. All receptors of such a system would be equally susceptible to excitation, conduction would be diffuse and equal in all directions, and the resulting integration would be temporary, and probably in many cases only partial as in Amoeba, and would soon give place to another integration with another receptor as the dominant region. As a matter of fact, some degree of organismic and even of axiate pattern and integration already exists before the nervous system appears (p. 242) and the development of the nervous system merely con- tinues the process. If strictly an axiate animals exist they do not give rise to nervous systems. All animals in which any visible differentiation of nervous structure appears are to some extent axiate before nervous differ- entiation occurs, though in the sessile coelenterates,

and particularly in the sponges, such axiation is not very highly developed. We must expect therefore to find at least some degree of structural or functional axiation, even in the most primitive types of nervous system existent. Even in the sessile coelenterates we do find various indications of nervous axiation and integration (pp. 243-44) and in . the medusae nervous axiation is much more marked. In radial forms, e.g., coelenterates and echinoderms, the different radii are more or less equivalent. There is no indication of any persistent dominance of any particular radial axis either in the medu- sae or in most of the sessile coelenterates. In the medusa, Cassiopea, that particular marginal sense organ which at any given time gives rise to impulses most rapidly, i.e., which has the highest rhythm and is therefore pre- sumably the most active, is for the time being dominant, and similar radial relations apparently exist in hydroids and most actinians. In some actinians, e.g., the Cerian- thidae, there are, however, some indications of a differ- ential or preferential relation among the radii in certain reactions, and in the echinoderms we find this preferential relation or functional dominance of certain radii, in all degrees of development even involving marked structural bilaterality of the body wall in certain groups.

Nervous differentiation is not coincident with the beginnings of axiation in animals but represents a later stage in the development of axiation, consequently the nervous system shows some degree of axiate pattern even in its most primitive forms; in other words, wherever differentiation progresses far enough to give rise to any distinguishable nervous structure some degree of struc- tural or functional axiation or centralization is probably always present. The physiological basis of nervous cen- tralization exists as soon as axiate pattern is established, that is, more or less definite axiate excitation-transmission relations are determined before the appearance of special- ized nervous structure, and when such structure appears its localization and general pattern are determined by these still more fundamental features of pattern.

Starting from the slight degree of centralization or axiation of the nervous system which is present in its most primitive clearly recognizable forms, we find many indications of progressive centralization and cephaliza- tion, which is centralization with respect to the longi- tudinal axis. In most of the great phyla or in some of their groups the evolutionary process appears more or less clearly. Even within the triclad Turbellaria differ- ent degrees of cephalization appear and, as already noted (p. 142), the rhabdocoeles and polyclads show a higher degree of cephalization than the triclads. In the mollusks again conditions range from the "stepladder" nervous system of the amphineura through the gang- lionic system of pelecypods and gastropods to the highly cephalized system of the cephalopods. The annelids, on the other hand, show but little progressive change in these respects, although certain data of regeneration indicate that a number of anterior segments definite for the species, but differing in different species, are more or less integrated with the prostomium to form a head region (Hyman, 1916). In various groups of the Crustacea, in the insects and to some extent in the

arachnids the process of cephalization appears in one form or another. To what extent these processes occur in the echinoderms is uncertain, but the holothurians probably represent a somewhat more advanced cephali- zation than the other groups. And finally the progress of nervous cephalization from the lower to the higher vertebrates is perhaps the most significant feature of vertebrate evolution. The evolutionary progress of nervous cephalization is accompanied by various other changes in form, proportion and relations of parts, par- ticularly in the segmented animals, in which the fusion of segments at the anterior end of the body, the decrease in total number of segments, and the increasingly intimate physiological relations of the segments in general are characteristic features.

Concerning the physiological basis of these changes we know as yet practically nothing, but it is evident that the changes represent modifications of the axiate pattern with respect to the longitudinal axis, and if the gradient is the physiological basis of axiate pattern, these evolutionary changes must be interpreted physio- logically in terms of gradients. Undoubtedly the form or curve of a physiological gradient, its slope or steep- ness, is different in different kinds of protoplasm, and undoubtedly also the capacity of protoplasm to develop a high conductivity has increased in the course of evo- lution. The experimental data at hand do not afford a basis for exact quantitative measurement or comparison of the gradient curves in different organisms, but they do indicate that the gradients undergo changes in the course of evolution as well as in ontogeny. The evolu- tionary change means merely that as the constitution

and hereditary potentialities of protoplasm are altered in the course of evolution, however we suppose such alteration to occur, the form, slope, or curve which a physiological gradient will take and the rate or intensity of physiological activity at its high end may also undergo change. In other words, the reaction of the protoplasm to the conditions giving rise to a gradient, or in the presence of a persistent or inherited gradient, must undergo modification in the course of evolution, like other protoplasmic reactions. In whatever terms we may prefer to interpret the problem of cephalization with respect to inheritance and evolution, I believe the physiological basis of the differences in different forms will be found to consist in the differences in the charac- ter of the physiological gradients in the different pro- toplasms.

The localization of the "central" nervous system in more or less distinct regions of the body in the pelecy- pods and gastropod mollusks indicates that the physio- logical conditions concerned in the localization of nervous structure in these forms are somewhat different from those in most other invertebrates. As regards the gradients in development some data have been obtained for certain gastropods, and these indicate that the primary polar gradient of early embryonic stages is more or less completely obliterated in larval stages and that in place of it several regions of high physiological activity arise, the foot, the visceral mass, the mantle folds, and the branchial region. It may be suggested that the peculiar ganglionic form of the nervous system in these groups is associated with these physiological facts, the differentiation of nervous structure being