The Origin and Development of the Nervous System

associated and localized in a definite way in relation to regions of high activity. In this connection it may also be suggested that certain ganglia outside the central nervous system in other forms arise in connection with secondary regions of high activity. The relation of the central nervous system to the symmetry gradients is definite and clear except in some of the lower invertebrates. In most of the bilateral invertebrates in which longitudinal nerve cords form part of the central nervous system, viz., most flatworms, annelids, and arthropods, the postcephalic longitudinal portions of the central nervous system are ventral, in the vertebrates dorsal, and except in some of the lower invertebrates they arise as a bilateral organ at or near the median line. Various lines of experiment and observation indicate that median and ventral in most bilateral invertebrates and median dorsal in the verte- brates, represents the high region of the symmetry gradients and is, at least during earlier stages of develop- ment, a region of more intense physiological activity than the regions on either side of it (pp. 26, 58).

In some of the lower invertebrates, however, the relation between the longitudinal nerve cords and the axial gradients is not so clear. In Planaria, for example, and in most other Turbellaria the nerve cords lie some distance lateral to the median line on each side (Fig. 19, p. 122), in the cestodes the chief nerve cords are near the lateral margins of the proglottides, and in the nemer- teans also the nerve cords are more lateral than median. Even in Planaria it has not yet been possible to deter- mine the primary symmetry gradients with certainty because of the lack of early embryonic stages, but in all the Turbellaria examined the ventral region is, at least primarily, a region of higher activity than the dorsal, and while the lateral margins of the body may develop high activity in the ectoderm in later stages the growth of new tissue from the cut ends of pieces and various other facts suggest that in the parenchyma the median ventral region is more active than the lateral.

There are some reasons for believing that dorsiven- trality is the first step in the development of bilateral symmetry in some of these forms and that the distinction between median and lateral is of later origin. If this is true, the localization of the nerve cords in these lower vertebrate groups may represent the high ends of the symmetry gradients, i.e., the cord may arise where the high activity of the ventral region begins to decrease laterally and dorsally. Where the whole ventral region is alike and the gradient begins laterally the cords may be lateral, and as the width of the ventral active zone decreases they may arise nearer the median line. This interpretation is offered merely as a suggestion, although there is some evidence in support of it. Until we know more concerning the early embryonic stages in these forms, definite conclusions are not possible. But what- ever the actual conditions in these lower forms may prove to be, it appears to be true in general for the higher invertebrates and the vertebrates that differ- entiation of nervous structure is primarily and pre- dominantly localized at the higher levels of the chief

body gradients. The evolutionary progress of centraliza- tion and cephalization from the primitive more or less diffuse localization of nervous structure in relation to the general body surface is one aspect of the evolution of axiate pattern. At present, however, we can do little more than note the relation between the physiological gradients and nervous localization. How it comes about physio- logically that the higher levels of these gradients give rise primarily and predominantly to nervous structure is a problem in the physiology of localization and differentiation like that pertaining to any other organ. We can see that the origin of the chief aggregation of nervous structure from that region of the embryo which is primarily most active, most sensitive, and physio- logically dominant represents a continuous sequence of physiological processes, but our present knowledge of living protoplasm does not permit us to analyze this sequence into its physicochemical terms.

The apparent migration of the cell bodies of periph- eral receptors from various regions of the body surface toward or into the central nervous aggregation consti- tutes another aspect of the evolutionary process of cen- tralization. Primitively these cells differentiated in the general epidermis (Fig. 18) and gave rise to nerve fibers which grew toward the central nervous system and were afferent in function, but in the vertebrates most of these cells arise in close relation to the central nervous system and send fibers peripherally as well as centrally. The olfactory epithelium alone retains the epidermal

origin. It is evident that in the more primitive con- ditions the localization of these cells is determined primarily by the surface-interior, rather than the axiate pattern. The whole ectoderm as that part of the organisms most directly related to the external world may be regarded as potentially nervous. It is obvious, however, that in the course of evolution the axiate pattern has become predominant over the general surface-interior pattern. The whole hereditary mechan- ism concerned in reproduction provides in more and more definite ways as evolution proceeds for the estab- lishment of such axiate pattern in the egg or for its persistence through the reproductive process. At the same time the surface-interior pattern, while still evident in many ways in the higher forms, becomes less and less important as a primary factor in differentiation. Doubtless the phylogenetic development of protective coverings of the egg and early embryonic stages is concerned in these changes, the greater degree of pro- tection of the body surface by egg coverings, etc., from the direct action of external factors in the early develop- mental stages of higher forms, determining that the general surface-interior gradients shall be less definite and not so far above the general level of intrinsic activity. With this phylogenetic change, the evolution of axiation, the relation of nervous differentiation to the regions of most intense metabolism in the embryo appears in the shifting of the region of differentiation of the receptor cells from the general body surface toward the high regions of the axiate pattern and they, like the motor neurons, send fibers peripherally, i.e., down the chief axial gradients.

From this point of view this process of shifting in position is physiologically similar to certain features of the process of cephalization, e.g., the phylogenetic migration of segmental ganglia toward, and their aggre- gation in, more anterior regions. The localization of nervous structure in any particular organism is pri- marily an expression of the localization of the regions of higher rate of the fundamental dynamic activities in that organism and the progressive growth and differen- tiation of nervous structure is primarily from the regions of more intense to those of less intense activity. The statement that the whole ectoderm is potentially nervous, a statement which has been repeatedly made, means in terms of this conception that wherever the fundamental physiological activity of the ectoderm cells is sufficiently intense there nervous structure will arise, provided the protoplasm is capable of recording this activity in per- manent structural differentiation. The evolutionary progress of nervous centralization and cephalization signifies that, as the axiate pattern becomes more and more predominant over the primitive surface-interior pattern, it plays a larger part hi determining the localiza- tion of the physiological conditions which underlie the development of nervous structure. And since these conditions are evidently associated with the higher levels of metabolic or oxidative activity in protoplasm the development of nervous structure becomes more and more definitely localized in the high regions of the axial

The phylogenetic changes in localization and aggre- gation of nervous structure are of course not independent of the changes in other parts of the organisms. We see clearly enough that the organism has evolved as a whole, not as a mosaic of independent parts, but the changes in the nervous system are among the most fundamental and significant changes which have occurred. As long as we regard the nervous system as representing physiologically a new factor superadded at some period in evolution to the primitive organismic system, the phylogeny as well as the ontogeny of the nervous system and its relation to other parts remains a complete mystery, and we can merely record and not interpret the facts. But phylogeny as well as ontogeny must sooner or later be interpreted into terms of develop- mental physiology and potentialities of the hereditary mechanism of the various protoplasms. The conception which I have attempted to develop in this book provides at least a working hypothesis for further attack upon the problems of developmental physiology, not only as they concern the individual but also in their phylo- genetic aspects. It must not be forgotten, however, that this hypothesis is purely physiological and is based primarily on physiological evidence. It concerns the actual physiological conditions under which potentiali- ties of the hereditary mechanism or action system are realized. The problem of the hereditary mechanism remains, although I believe that increasing knowledge concerning the physiological conditions under which hereditary potentialities are realized must teach us something concerning the physiological basis of these potentialities.

It has been shown in preceding chapters that a definite relation exists, as regards localization differ- entiation, and the course and degree of centralization and cephalization, between the nervous system as a whole and the physiological pattern of the organism. Moreover, it has been pointed out that the problem - of the nature of this relation is first of all a problem of developmental physiology and that it must be solved in terms of the physiology of organismic pattern. And finally, an attempt has been made on the basis of a physiological conception of organismic pattern to throw some light on this problem.

The current conception of the cellular constitution of the nervous system is that, except in its most primi- tive forms, it is made up of distinct cells, the neurons, which usually give rise to outgrowths of two sorts, the axons or neurites and the dendrites (Figs. 45-53). In such a nervous system functional continuity from cell to cell is established during the course of development. It has been demonstrated in many cases and is believed by many to hold for all that this continuity consists in a contact relation between axon and dendrite, axon and cell body or dendrite and cell body of different cells with persistence of the cell membrane between the parts in contact. This region of contact between two cells is

termed the synapse, and a nervous system consisting of distinct neurons connected by means of synapses is known as a synaptic nervous system. In some of the lower invertebrates and in certain organs of the higher animals, however, the nervous system consists of a net- work or reticulum of branching cells forming a syncy- tium, i.e., direct protoplasmic continuity apparently exists from cell to cell through the branches, without the intervention of the membrane characteristic of the synaptic type of connection. The cellular elements of such a nervous network have been called by Parker protoneurons (Fig. 66, p. 203). The protoneurons usu- ally show little or no definite morphological axiation or polarity, but the forms of neuron most characteristic of both invertebrates and vertebrates show very definite axiate patterns, though a wide range of variation in form occurs.

The problem of the polarity and morphological pat- tern of the neuron is of interest from various points of view : First, as regards its origin, the question whether it is inherent in the neuroblast or is determined in relation to correlative factors is of the same importance as in other cases of axiate pattern. Second, if this polarity is of correlative origin, the question of the nature of the factors concerned is of fundamental significance, since the neuron patterns are essential factors in determining nervous structure and inter- relations. Third, the question of the relation between the visible morphological polarity and the functional polarity of the neuron as an element of the nervous system is an important aspect of the problem of nervous function.

The problem of the nature and origin of neuron polarity has received comparatively little attention from neurologists, and in view of the fact that experimental data are lacking, any consideration of the problem at this time will doubtless appear to some as a completely futile undertaking. Neuron polarity and pattern are, however, essential features of nervous pattern, and it is scarcely possible in a discussion of the origin of nervous pattern to ignore the problem of neuron pattern. In its most highly developed form this pattern represents a specialized type of intracellular axiate pattern which appears in a very definite and orderly manner, though with extreme variation and complication as regards axial direction and spatial relations. We must at least attempt to discover whether there is any basis for believing that this pattern is similar in nature and origin to other types of axiate pattern. Any exhaustive consideration is of course impossible, but it is desired to call attention to certain facts which appear significant and to suggest certain interpretations which may possibly point the way to an experimental attack on the problem.

Our knowledge of the morphology of the neuron is based upon the extensive investigation of Golgi, Ramon y Cajal, Ehrlich, Retzius, Bethe, and many others, and an extensive literature has grown up, from which is evident notonly the bewildering variety of neuron pattern as regards detail but also certain general characteristic features. As reference to any of the general works on the histology of the nervous system will show, the following FIGS. 45-47. — Fig. 45, motor cell of ventral horn of spinal cord from human embryo 30 cm. long (from Barker, 1901, after von Lenhossek); Fig. 46, pyramidal neuron of cerebral cortex of rabbit (from Cajal, 1909); Fig. 47, neuron with short, branching axon, from cerebral cortex (from Cajal, 1909).

brief survey serves merely to call attention to some of the more common and characteristic aspects of neuron pattern. The most characteristic feature of the differentiation of the neuron from the neuroblast is the appearance of outgrowths of the cell body which often develop an extremely complex system of branches, resembling the branching of the multiaxial plants. In most vertebrate neurons the two kinds of outgrowths, axon and dendrites, are clearly distinguishable. The axon is slender and of uniform diameter over all or most of its length and possesses a characteristic structure differing from that of other parts of the cell. The axon may extend for a long distance from the cell body with no, or very few, branches, as in many motor cells (Fig. 45), or it may give rise to lateral branches, the collaterals, in the immediate vicinity of the cell body (Fig. 46). In other neurons it may break up into branches and terminate within a short distance from the cell (Fig. 47). It may terminate in a complex system of branches, the terminal arborization, near or far away from the cell body, or it may end with little or no branching. The relations of axons to dendrites of other neurons also differ widely in different cases. In some the axon or some of its branches form extensive spirals about the dendrite while in others the approach is more or less direct. Usually a neuron gives rise to only one axon, though dioxonic and even polyoxonic neurons occur, and usually also the axon is

The dendrites, on the other hand, are structurally very similar to the cytoplasm of the cell body and are FIGS. 48-51. — Fig. 48, Purkinje cell from human cerebellum (from Stohr); Fig. 49, fusiform cell from anterior corpus quadrigcminum of rabbit (from Cajal, 1909); Fig. 50, monopolar neuron from mesen- cephalic nucleus of fifth cranial nerve (from Cajal, 1909); Fig. 51, neuron of smaller gustatory type from carp (from Herrick, 1905). obviously extensions of it. As the term "dendrite" implies, they usually branch repeatedly, giving rise to treelike complexes, and in general the diameter of the dendrites decreases from the bases of the main trunks to the ultimate branches (Figs. 45-53). Moreover, the neuron usually gives rise to several or many dendrites at different points of its surface. In some neurons, e.g., the pyramidal cells of the cerebral cortex, a chief dendrite is distinguishable, arising from the opposite end of the cell and growing in the opposite direction from the axon (Fig. 46), and some neurons may show only a single dendrite trunk. In still other types of cells the den- drites may arise in two chief groups, or from two main trunks at opposite ends of the cell (Fig. 49), or several groups or trunks may develop, and in some cases the axon takes its origin from some region of a dendritic outgrowth. Again, numerous dendrites arise in some neurons from all regions of the cell surface and extend and branch in all directions (Fig. 45). At the other extreme are neurons without dendrites, for example, peripheral sensory neurons and some central neurons (Fig. 50) . Among the lower vertebrates and the inverte- brates a common type of invertebrate neuron pattern is that with a single outgrowth from which both dendritic branches and an axon arise (Figs. 51, 52, 53). Anaxonic neurons also occur frequently among the invertebrates. As regards its developmental history, neuron pattern shows still further complications. The final relations of axon and dendrites, particularly of the latter, are often quite different from their relations in earlier stages, and it is only by following the course of development that we can hope to gain any light concerning the factors

involved. In the pyramidal cell of the cerebral cortex (Fig, 46) the course of development according to Cajal (Fig. 54, A-E) is as follows: the primary outgrowth, the axon arises on that side of the cell away from the external surface of the cortex and its growth is directed away from the surface (A). The chief dendrite appears at the opposite end of the cell and its growth is directed toward the surface (B, C); as the axon elongates, collateral branches arise from it and grow more or less vertically to its direction (D); and still later dendrites arise from various parts of the cell surface and grow in various directions (£). Taking the cell at the stage B or C it appears that axon and dendrite are localized

at opposite poles of the cell, and it might be supposed that some factor in the cell environment to which the two outgrowths react in opposite ways, e.g., the one positively, the other negatively, determines the direction of their growth. Later, however, the axon gives rise to collaterals which grow in a direction more or less vertical to that of the primary axon (D), and finally the cell gives rise to dendrites which grow in various direc- tions (£). The apparently simple orientation and opposed directions of growth of the earlier stages are much complicated later and no simple interpretation appears possible. Taking all the facts into account it is impossible to doubt that the pattern of such a neuron arises in some sort of relation to its environment, but it is evident that the environmental factors, or the reactions of the cell or its parts to the environment, or both environmental and intracellular factors, undergo alteration and complication during development.

Again, in the fully developed Purkinje cell of the human cerebellum (Fig. 48) the axon arises from the side of the cell away from the external surface of the cerebellar cortex and passes inward, i.e., away from the surface, but usually not vertically to it. The collateral branches of the axon form more or less obtuse angles with the axon direction. The dendrites apparently arise at or near the opposite pole of the cell from the axon and their more basal portions extend parallel to the surface, while their branches show an orientation apparently with respect to the surface, which becomes more and more definite with approach to it.

In the new-born child, however, the Purkinje cell has according to Cajal the form of Figure 55. The chief dendrites arise opposite the axon, but numerous other dendrites are present on all parts of the cell surface and extend in all directions. Evidently the situation at this stage is very different from that in the mature neuron. The primary dendrites are more or less definitely oriented, but many others exist which show no uniform orientation. These latter are evidently temporary and are resorbed in later stages. Cajal distinguishes a period of outgrowth and one of regulation and order- ing of the dendrites. Here, as in the pyramidal cell, very considerable changes occur in the course of devel- opment, both in localization and orientation of the dendrites.

The developmental stages of the axon of the Purkinje cell also show extensive changes. In the new-born infant the axon shows numerous, much-branched collaterals near the cell body as well as a number of longer and less-branched collaterals. In later stages extensive resorption of the shorter, highly branched collaterals occurs, and in the adult only the longer, less- branched outgrowths remain. This neuron evidentally gives rise to many outgrowths which are only temporary and play no part in the final morphological and functional relations of the neuron. Apparently a selection among the outgrowths takes place during development and presumably on a functional basis, that is, those out- growths which play an essential part in the functional relations of later stages persist, while those which are not thus functionally active disappear, like larval organs and other temporary structures in the development of organisms. It is evident, however, that the appearance of these outgrowths must be determined by certain

physiological conditions and the persistence of some and the resorption of others by certain other conditions. In many other neurons dendrites arise from all parts of the cell surface and grow in all directions without orientation at any stage (Fig. 45). In some cases in which the axon of the mature neuron arises from a dendritic outgrowth, this outgrowth is secondary and the axon developed originally directly from the cell body. The spinal ganglion cell in early stages shows two outgrowths localized at opposite poles (see Fig. 60, p. 1 86), but later the bases of the two outgrowths approach each other and unite, and the cell gives rise to a single secondary outgrowth from the end of which the two primary outgrowths extend in opposite directions. In some spinal ganglion cells in mature condition this secondary outgrowth is not single in its origin, but the cell gives rise to two or more outgrowths which branch and anastomose inside the connective tissue capsule which surrounds the cell, but finally unite to form the single outgrowth and in other cells several regions of branching and anastomosis appear in the course of this outgrowth.

Frequently axons or dendrites on reaching a certain point in their course show a sudden and definite change in direction as if they had come under the influence of some environmental factor acting in a direction different from that to which they had previously been reacting. For example, the longer dendrites of the pyramidal cells of the cerebral cortex grow vertically toward the surface, but as they approach it they branch and the branches become parallel to the surface (Fig. 46). In the mature Purkinje cell of man the chief dendrites

extend more or less parallel to the surface, while the secondary branches are in general vertical to it (Fig. 48) . Axons in many cases show similar changes in direction, sometimes very abrupt. Again, the dendrites of certain cells may be at first unoriented but later, as the axons of certain other neurons grow into their vicinity, they may become definitely oriented with respect to these and grow toward them. Other variations and complications in the form, arrangement, and developmental changes of the out- growths of the neuron might be described almost indefinitely, but those mentioned suffice to call atten- tion to the bewildering complexity of the facts and the difficulties apparently involved in any attempt at physio- logical analysis of neuron development. In addition to this problem of external morphology and development the neuron presents various other physiological problems, e.g., the origin and function of the internal structural features, the Nissl substance and the neurofibrils, the physiological relations of different portions of the neuron, both as regards nutrition and maintenance, as indicated by the data of degeneration and regeneration, and as regards functional conduction of impulses. For the present, however, we are primarily concerned with the question of physiological polarity or axiation and the development of the external morphology.

Neurons are very commonly distinguished as unipolar, bipolar, or multipolar, according to the number of out- growths, though the terms "uniaxial," "biaxial," and "multiaxial'1 are perhaps preferable. There are of course in general two types of axial outgrowths, the axons and dendrites, which differ in time of appearance, degree of differentiation, and in part at least in function. While we have no demonstrative experimental evidence for the conclusion, there is every reason to believe that the neuron outgrowths represent physiological gradients

during the period of their development. The axon outgrowth certainly possesses a growing tip which shows evidences dur- ing the growth period of greater physiologi- cal activity than other parts of the outgrowth (Fig. 56). While the dendrites do not usu- ally show so definite a growing tip as the axon, their growth form, order of branch- ing, and changes in direction of growth show very clearly that growth is primarily terminal in each dendritic axis, although secondary growth in thick- ness may occur in other parts. In short, all the facts at hand indicate that both axons and dendrites are primarily physiological gradients like other physiological axes in organisms.

It is impossible to proceed very far in the examination of nervous tissue prepared by the Golgi method, or any FIG. 56. — Two views, taken fifty min- utes apart, of same nerve fiber, growing from group of embryonic spinal cord cells into the lymph of tissue culture (from Harrison, 1908). other which shows the external morphological neuron pattern, without being impressed by the resemblance between the usual types of neuron pattern and those of the multiaxial plants and multiaxial animals, such as the hydroids. Many neuron patterns, for example, are much like tree patterns, and different modifications of neuron pattern are at least as clearly distinguishable as the forms of different species of trees. Resemblances of this sort do not necessarily signify identity or even similarity of constitution in the individuals or proto- plasms concerned, but I believe they do at least suggest similarity in axiate pattern, i.e., in the relations of physiological gradients in the different protoplasms. It is difficult, in fact, not to believe that similarities so striking as those between neuron pattern and tree pattern, for example, mean something physiologically as regards the relations between the constituent axes and their integration into the whole individual. The hereditary constitution of the protoplasms of neurons and of trees is doubtless very different, and the environ- mental factors to which the plant is reacting are obviously in large part or wholly different from those to which the neuron protoplasm reacts. Notwithstanding these differences the physiological relations of the axial gradi- ents in the two types of individual may be more or less similar and may determine similar types of growth pattern.

The axon is a highly specialized outgrowth whose chief function is evidently conduction. It probably has little or no direct nutritive relation with its environment except possibly in earlier stages of its development. Not only its growth but its maintenance depends upon its connection with the cell body, as the experiments on degeneration and regeneration have demonstrated. There can be little doubt, however, that the dendrites are important nutritive organs as well as conductors of impulses. In all probability they are primarily nutritive and continue this function after they become conductors. The axon arises at a definite region of the cell and usually grows in a definite direction for at least some distance, but the dendrites very commonly, though not in all neurons, arise on various regions of the cell surface, often on all sides of it and grow in all possible directions, at least during earlier stages. In some neurons, how- ever, such as the pyramidal cell of the cortex (Fig. 46), a primary dendrite develops on the opposite side of the cell from the axon and grows at first or continuously in a definite direction (Fig. 54). In such cases other later dendrites may show other and various directions of growth.

So far as we are able to conceive the conditions in the developing nervous system, there are apparently only two sorts of conditions in the environment of the developing neurons which can possibly determine the localization or the direction of growth of either axon or dendrites. The one of these is chemical, the other electrical. The attempt to interpret the localization and direc- tions of outgrowth of axon and dendrites as a response of some sort, whether chemotactic in the strict sense or not, to chemical substances in the environment of the cells meets with various difficulties. In the first place,

the presence of the chemical substances to which reaction is supposed to occur and the proper concentration gradi- ents must be assumed. Moreover, in a neuron like the pyramidal cell (Figs. 46, 54) the reaction of axon and primary dendrite, if to the same substance, must be opposite in character, or they must be reacting to different substances. The absence of definite orienta- tion of the dendrites of some neurons in earlier stages, and its appearance later, and the sudden changes in direction of growth of both axon and dendrites make it necessary to assume that the reaction of the outgrowths changes, or that the concentration gradients of the sub- stances change their direction, or that new substances appear with concentration gradients in different direc- tions from those of the original substances. Again the establishment of functional connections between differ- ent neurons and between nerves and muscles and other peripheral organs necessitates various other assumptions. While directed response to chemical factors may play some part in neuron development, any attempt to interpret the process entirely in terms of such response is a pure speculation with many difficulties in the way. It may perhaps appear at first glance that the situa- tion is not much better as regards the interpretation in terms of electrical response, and so far as definite experimental data on the neuron itself are concerned, this is the case. Certain facts are at hand, however, which seem to me to be highly significant in this connection. In the first place, the existence of the axial physiological gradients and of gradients in electrical potential as one of their characteristics has been demon- strated beyond question, and we find a very definite