The Origin and Development of the Nervous System

relation between the directions of growth of the axons which arise in the earlier stages of development and these gradients. Second, from what we know concern- ing other developmental processes, there is good reason to believe that the establishment of regions of rapid growth in the nervous system, as well as in other organs, involves a change in electrical potential in those regions. Admitting this, it is evident that in the course of develop- ment there must be changes, both in the intensity of potential and in the direction of the potential gradients. Third, it is a familiar fact that conduction of the nervous impulse is accompanied by marked change in electrical potential, and as soon as any neuron, neuron group, or axon, or axon bundle becomes functionally active in this way, changes in the potential gradients in its vicinity must appear. In short, the existence of gradients in electric potential with the resulting electric currents and the changes in direction and intensity of these gradients during development and in relation to function must be a factor in determining the physiological con- dition of the cells affected, and so constitute a real basis of fact for attacking, at least tentatively, the problem of neuron development. On the other hand, it is not demonstrated, though many facts make it highly probable, that the neuroblast or the neuron is capable of response in any way to such electrical gradi- ents, or, more strictly speaking, to the electrical cur- rents which arise in connection with them. Moreover, the difficulties which arise in the matter of the opposite orientation of axon and dendrites in certain neurons and the absence of orientation in some or all dendrites in others, remain. But in spite of the lack of expcii-

mental data concerning the neuron itself and in spite of the difficulties, it seems worth while to point out certain of the possibilites along this line of interpretation.1 First, however, it is necessary to examine a hypothesis, recently advanced by Kappers, of neuron polarity and growth orientation in terms of response to electrical factors. In a series of papers Kappers2 has described certain phylogenetic and ontogenetic processes of shifting or migration of neurons in the nervous system, which he has called "neurobio taxis." Neurobiotaxis consists, according to Kappers (1917) in the following phenomena : "The growth of the chief dendrite and eventually the displacement of the cell-body itself takes place in the direction whence the majority of stimuli proceed to the cell. Only between correlated centers does this out- growth or shifting take place." In the growth of the axon also synchronic or successive stimulation plays a part. In attempting to account for the apparently directed outgrowth of dendrites and shifting of the cell body as well as for the localization and directed outgrowth of the axon Kappers has advanced the hypothesis that the electrical field in which a neuron develops is the fac- tor which determines its morphological polarity as regards axon and dendrite. He interprets the growth response of the neuron in the electrical field as a galvanotaxis, the axon originating on the side of the neuroblast toward the positive pole or anode of the field and growing

toward the anode, the dendrites arising on the side toward the negative pole or cathode and growing toward the cathode of this field. Kappers accounts for the pres- ence of the electrical field by the fact that a region of ex- citation becomes electronegative to its surroundings, and assumes that the electrical field to which the neuroblast reacts results from the stimulation of a neuron or group of neurons or axons in the vicinity of the neuroblast in question. The neuroblast itself he regards as sur- rounded by a fluid medium, the lymph, containing electrolytes and therefore conducting the current. According to the hypothesis, the neuroblast reacts to this electric current by localization and growth of axon and chief dendrite at opposite poles of the cell and later by shifting of the cell body toward the negative pole or cathode. The growth of the axon is then "stimuli- fugal," that of the dendrite "stimuli-petal."

As various authors have shown, the usual galvano- tactic response of living organisms is toward the cathode, though this reaction may be reversed in many cases by changes in the surrounding medium, such as increase in concentration of electrolytes. If, however, the polarity of the neuron is a galvanotactic response of the sort Kappers postulates, it shows both positive cathodotaxis and positive anodotaxis. In order to account for this opposite reaction of axon and dendrites and for various other facts, such as the usual condition with one axon and several or many dendrites, the appearance of the axon earlier than the dendrites, etc., Kappers is forced to make a number of special assumptions, all of which have a physicochemical basis, but which seem to me greatly to complicate the hypothesis and limit its

applicability. For example, the localization of the axon outgrowth is regarded as corresponding to the "anodic extension" observed by Verworn and others in Protozoa and certain other forms exposed to an electric current. Assuming that this anodic extension possesses a greater permeability for negative than for positive ions and referring to the evidence indicating that the medium about the neuroblast contains a considerable amount of potassium salts/ Kappers maintains that the anodic extension will take up more negative ions than other parts of the cell and also more negative than positive ions and so will become electronegative and therefore positively anodo tactic, i.e., it will be attracted by the anode and grow toward it.

The localization and outgrowth of the chief dendrites toward the cathode are regarded by Kappers as repre- senting the normal or usual cathodo taxis. The dendrites grow out only after the cell contains Nissl granules, which are also present in the dendrites. These granules are supposedly of nuclear origin and therefore acid and electropositive, according to Kappers. He also suggests that the cataphoresis of albumens may play a part in the localization and direction of the outgrowths. Alka- line albumen particles move toward the anode and acid albumens toward the cathode of a current passing through the medium in which they are suspended. A more extended consideration of these and other special assumptions which Kappers makes to account for the facts is impossible and, I believe, unnecessary here, for I shall endeavor to show that they are not essential to an interpretation of the developmental processes con- cerned in terms of reaction to an electric current.

Kappers' hypothesis may be illustrated by the accompanying diagram (Fig. 57). In this diagram a represents the source of stimulation which Kappers postulates and which is electronegative externally; d, ax is a neuroblast surrounded by a fluid medium containing electrolytes. The current produced by a has its nega- tive pole on the surface of a and its positive pole in some more or less distant region, indicated by the sign+. The return current must pass in the opposite direction through body tissues. The neuroblast exposed

FIG. 57. — Diagram illustrating Kappers' conception of the action of the bioelectric current on the neuroblast. For explanation see text. to the current arising from a develops an anodic exten- sion ax, which, because of its greater permeability to anions, becomes electronegative and gives rise to the axon growing toward the positive pole because of the attraction between it and this pole. Later the opposite side of the cell gives rise to the dendritic outgrowth d, which contains the supposedly acid and electropositive Nissl granules and is therefore attracted by the cathode and grows toward it.

While this ingenious hypothesis of Kappers' is of great interest as an attempt to interpret the polarity and external morphology of the neuron in terms of a reaction to an electric current, it seems to me to present certain difficulties, some of which may be briefly pointed out. In the first place, tljis hypothesis does not account for the outgrowth and orientation of axons and dendrites of the neuroblasts which first undergo differentiation in the nervous system, for it assumes the presence of functional neurons as the source of the current. Second, Kappers assumes that the neuroblast is more or less completely suspended in a fluid medium, consisting of lymph, containing electrolytes, while, as a matter of fact, it is in its earlier stages part of a very compact tissue, which may be more or less syncytial in character. At any rate, the cells are apparently in close contact, and tissue continuity exists to a high degree. Later in development of course the neurons do become more or less separated by the substance of the neuroglia, by lymph spaces, and probably by colloid gels, but differ- entiation begins long before these changes have occurred. Third, the hypothesis does not take into account in any way the metabolic activity concerned in the localization and outgrowth of axon and dendrites, but this is con- ceived as almost entirely a matter of electrical attraction between bodies of opposite sign. Actually, however, these processes involve localized metabolic activity, including oxidation and synthesis of new protoplasmic substance. Moreover, if positive cathodotaxis, such as Kappers assumes to underlie the outgrowth of the dendrites, is associated with cathodic stimulation, as many facts indicate, we should expect the side of the neuroblast toward the cathode to be more active than the other, and the dendrites should arise before the axon, but the reverse is the case. Fourth, the " anodic extension" observed by Verworn in certain forms in the

current is by no means certainly an excitation, as Verworn regarded it. As a matter of fact, if the current is long continued or sufficiently strong, death and dis- integration begin in this anodic extension. Kappers' assumption of increased permeability to anions in the anodic extension also seems to involve difficulties. In the current the anions migrate toward the anode, and unless the concentration of electrolytes in the medium surrounding the neuroblast is very much higher than inside the cell, which cannot be the case, the anions must migrate out of the cell toward the anode instead of into it. Consequently the anodic extension ought, according to Kappers' assumptions, to become more electropositive internally and therefore to grow toward the cathode, instead of becoming more electronegative and growing toward the anode, as he maintains. While these and other difficulties make it impossible to accept Kappers' hypothesis as it stands, there can be no ques- tion concerning the significance of the facts to which he calls attention, and concerning the difficulties involved in any attempt to interpret them in any other terms than those of relations of some sort to electrical factors.

Most of the special assumptions as well as most of the difficulties involved in Kappers' hypothesis can, I believe, be avoided by considering the reaction of the neuroblast, not simply as a matter of electrical attrac- tion, but rather as a change in physiological condition, resulting primarily from the electrical polarization of the cell through the action of the current to which it is exposed. When an electric current passes through any conducting medium in which ions are free to move, negative ions (anions) migrate toward the positive pole or anode and positive ions (cations) toward the negative pole or cathode. From this migration of the ion? electrical polarization results. If the current passes through a living cell, it is evident that ions of opposite sign must migrate in opposite directions in the cell, and if the cell possesses a membrane which is not completely permeable to ions, and cells in general do possess such membranes, the ions of opposite sign must accumulate to some extent at opposite poles of the cell inside the membrane, the positive ions on the side toward the cathode, the negative ions on the side toward the anode. The cell is then electrically polarized and this polariza- tion must determine changes in its physiological condition

and activity. As I pointed out in earlier chapters (pp. 70-82), a region of excitation or of rapid oxidation is electronegative externally with respect to unexcited or inactive regions. This means that it is internally electropositive with respect to such regions. Increase in internal electropositivity is then associated with increase in the fundamental physiological activity of protoplasm, though it may also be brought about in other ways as well. Moreover, various lines of evidence and particularly the data concerning electrical stimula- tion indicate that an increase in internal electropositivity of living cells, directly produced by electrical means, generally if not universally determines a condition of excitation or an increase in fundamental physiological activity. It may be that in any stimulation of living protoplasm the primary change is electrical, but whether that is the case or not it is evident that the electrical change is a factor in the change in physiological state. If we admit this, it follows that electrical polarization of a cell may determine changes in its rate of metabolism and in its physiological state in general.1 We may expect to find the region of increased positivity in such a cell showing increased physiological activity or excita- tion, the region of increased negativity showing decreased activity or depression. As Hyman (1918) has suggested, the phenomena of electrotonus in the nerve fiber appar- ently fall under this head. When a constant current is passed through a portion of a nerve the irritability of a

1 If the membrane of the unstimulated cell is normally polarized with internal surface negative and external positive, as commonly assumed, an increase in positivity in any region really constitutes primarily a depolarization of that region and such depolarization is, according to R. S. Lillie (see pp. 72-73), an essential factor in stimula- tion. region about the cathode, the region in which the nerve becomes internally positive and externally negative, is increased (catelectrotonus) , and if the current is strong enough the cathode gives rise to a nerve impulse when the circuit is closed. A region of the nerve about the anode, however, where the nerve becomes externally more positive and internally more negative, shows decreased irritability while the current is passing, but the anode may give rise to a nerve impulse when the circuit is broken, that is, the disappearance of the depression constitutes an excitation. Electrotonus is apparently merely a special case of the physiological effect of electrical polarization in highly specialized tissues.

Electrical conditions as factors in development and differentiation have received comparatively little atten- tion of late years, though they are probably of consider- able importance, particularly in the case of highly susceptible cells, such as the neuroblasts, which in early developmental stages are the most susceptible cells of the organism. Concerning the presence of differences in electrical potential and of electric currents there can be no doubt, for, as I have pointed out (p. 44), they are characteristic features of the physiological gradients from the earliest stages on, and their existence in con- nection with the specialized activities of later stages has been abundantly demonstrated. It is of course not necessary to assume that electrical factors are the only factors concerned in determining the morphological polarity, directions of outgrowths, and connections of the neurons. Various other factors may play a part, but it appears impossible to find any other than the

electrical factor which is adequate as the primary factor in these processes. The point may be emphasized, however, that according to the conception advanced here the electrical factor is effective, not directly as electrical but only through the changes in the physio- logical state brought about by it and the localization of those changes in relation to the electric currents. The direction of electrical polarization in a given cell depends on the position of the poles of the polarizing current, and it must be pointed out that with respect to any particular source of current two opposite possi- bilities exist, depending on the relation of the cell concerned to the source. The situation as regards bio- electric currents is not fundamentally different from that in the galvanic cell. We call one pole of the cell positive, the other negative, but it is a familiar fact that the pole which is negative when the two poles are connected by a conducting medium outside the cell is positive with respect to the other pole inside the cell. We commonly say that the current flows in opposite directions inside and outside the galvanic cell. Similarly in the case of bioelectric phenomena, if a particular region of an organism is electronegative with respect to another, when the two regions are connected by a con- ducting medium, e.g., a wire, outside the protoplasm, it must be electropositive internally, that is, through the protoplasm, with respect to the other region. Here, as in the galvanic cell, the directions of the current inside and outside the tissue are opposite. As regards the electric polarization of a cell by such a current, it is evident that two opposed possibilities exist, according as the cell lies in one or the other half of the circuit.

The following diagrams will assist in making the point clear. Let ab (Fig. 58) represent a mass of con- tinuous, closely coherent tissue, the cells of which are FIGS. 58, 59. — Diagrams illustrating electrical polarization of a cell by bioelectric currents: Fig. 58, polarization of first type; Fig. 59, polarization of second type; further explanation in text. in contact, such, for example, as the embryonic neural tube. In consequence of differences in physiological condition the region a is electropositive internally to the region b and a current passes through the tissue between

the two. Positive ions will then migrate toward b, negative ions toward a. In a given cell, n, of the mass this migration will occur if the current passes through it, and the ions will accumulate at two opposite regions of the cell, the side toward b becoming more positive inter- nally, the side toward a more negative. For convenience in the following discussion this is called the first type of polarization. We turn now to the other alternative, Figure 59, in which the cell n is not in direct tissue-continuity with the source of current a but suspended in a conducting medium. The cell or tissue a is electropositive inter- nally, because of its physiological condition and is therefore electronegative externally, i.e., with respect to the medium in which the cell n lies. Consequently in this medium positive ions will migrate toward the cathode on the surface of a, negative ions in the opposite direction. The cell n will therefore undergo electrical polarization with the positive pole toward a, the negative in the opposite direction. This may be called for our purposes the second type of polarization. Comparing the two diagrams, Figures 58 and 59, it is evident at once that although the relations as regards space and direction between the region of high positivity a and the cell n are the same in both cases, the polarization of n is opposite. If the region of increased positivity in the polarized cell becomes more active physiologically and growth occurs in it in consequence of the polariza- tion, the outgrowth in the first case (Fig. 58) will be away from a and toward b, and in the second case (Fig. 59) toward a. In the following sections a few suggestions are advanced concerning the occurrence of

these two types of electrical polarization in the cells of the developing nervous system and their role in the differentiation, direction of growth, and connections of the neurons. In the localization and development of at least the earlier axons in the vertebrate central nervous system the conditions represented in Figure 58 seem to be more or less exactly realized. The embryonic neural tube constitutes a compact tissue and is the region of highest internal positivity in the embryo. Moreover, the region of highest internal positivity in this tissue is probably primarily that adjoining the original external surface (see p. 26), which becomes the surface of the central canal after closure of the tube (see Figs. 17, 62, 63). Secondarily, the edges of the neural folds which become the dorsal regions of the neural tube are, to judge from their susceptibility, the most active regions of the tube preceding the closure and therefore must represent regions of high positivity internally, as compared to more ventral regions. We may expect then to find that in the transverse section of the neural tube the electropositive regions of the neuroblasts as determined by electrical polarization, will appear primarily on the sides of the cells away from the central canal and from the dorsal region, or in the resultant of these two gradi- ents. Assuming that the region of highest positivity in each cell becomes the region of most rapid growth, extensions or outgrowths of the cells will appear on the sides of the cells away from the central canal, or the dorsal region, or in the resultant of these two directions and will grow laterally, or ventrally, or ventrolaterally

in the neural tube. Figures 60 and 61, cross-sections of the cord of the chick and of the fish in early stages of neuron development, show very clearly the localization and directions of growth of the early axons and indicate their very definite relation to the central canal in the cells nearest it and to the dorsiventral direction in others. In short, the neuroblasts are apparently reacting to the physiological gradients, and the only factor in these gradients which we can at present conceive as deter- mining polarity and direction of outgrowth is the elec- trical factor, acting through the polarization of the cells.

According to this interpretation, the region of origin of the axon in the neuroblast represents the high end of a physiological gradient, electrically determined in each cell, by the electric currents of its environment. The direction of axon growth may likewise be regarded as electrically determined but as a process of electrical and consequent physiological polarization rather than through a simple electrical attraction. Harrison (1908, 1910) has shown that the growing tip of the axon is amoeboid and sends out pseudopodia, not merely in the direction of growth, but in various other directions (Fig. 56), and various axon tips in Figure 60 show similar conditions. The direction of pseudopod exten- sion in the growing tip is evidently not determined by the electric gradient, but in general the pseudopodia which extend in the direction of the gradient will tend to persist, extend, and grow, because they, and particularly their tips, represent the regions of highest positivity.

An apparent migration of the cell body in the direc- tion of axon outgrowth, i.e., toward the cathode of the current, is a feature of these early stages, at least in the development of the spinal cord. This is probably, like the axon outgrowth, a result of the electrical polariza- tion of the cell, the side toward the cathode being more positive, more active physiologically, and growing more rapidly than, or even at the expense of, the other side. Apparent migrations resulting from such differences in relation to an axial gradient are of frequent occurrence in organisms. One of the most striking examples is found in the stolons of various hydroids. In the absence of food the tip of the stolon, which represents the high end of the axial gradient, grows at the expense of lower levels, and the stolon apparently creeps forward and may even become separated from the parent organ- ism through the resorption of its more basal regions as nutrition for the growing tip. Such free stolons appar- ently wander about over the surface to which they are attached, always growing at the tip and undergoing resorption at the base, until they finally use up their own substance and disappear, or else, under certain con- ditions, give rise to a new hydranth and begin to feed. The apparent migrations of the neuroblasts and later of the neurons (p. 200) are perhaps due to such growth on the electropositive and resorption on the electro- negative side of the cell.

The direction of growth of the axons is not neces- sarily altered by their emergence from the neural tube. Since they are protoplasmically continuous with the cell bodies in the tube, the electrical polarization of their tips may still remain the same. Consequently the general direction of growth of these axons will be down a resultant of the chief physiological gradients of the developing body. As the distance between the axon tip and the regions of high positivity in the chief body gradients increases, the positivity of the tip, and therefore its rate of growth, may decrease. In the connection of the axon with its peripheral organ other than electrical factors may per- haps be concerned. Chemical factors, for example, may play a part in some cases. It may be pointed out, however, that if the peripheral organ represents a region of relatively intense growth or functional activity, and therefore of high positivity internally, as the axon comes into its vicinity, it may bring about electrical polariza- tion of the second type in the axon tip and so determine growth of the axon toward itself. It appears probable that the branching of peripheral nerves and the periph- eral distribution and connections of axons with their end organs may be determined, in part at least, in this way through the developmental activity of par- ticular organs, e.g., particular muscles, at the time when particular axon tips grow into their electrical field.

The growth inward of the axons of peripheral receptors in the invertebrates and in the olfactory primodium of the vertebrates is apparently a reaction of the first type (Fig. 58), like that of the axons in the neural tube, that is, it represents a growth directed by electrical polarization away from the region of high positivity just beneath the surface of the body. As the axons of peripheral receptors approach the central nervous system a polarization of the second type may arise with respect to the negative surface of the central nervous system and so determine their growth toward it. Inside the central nervous system the factors

determining their distribution must be similar to those acting in the case of other axons. In many cases, how- ever, the direction of growth of the axons of peripheral receptors is altered almost immediately after their origin by the presence of axial gradients, which are more effective than the surface-interior gradients in deter- mining polarization and directed response. Even in the slightly differentiated sensory cells of the actinian tentacle the outgrowth apparently reacts almost immedi- ately after its origin to the axial gradient of the tentacle and grows in most cases down this gradient (Fig. 18). Axons may apparently grow up as well as down a general body gradient. According to the polarization hypothesis, such growth may result from polarization of the first type with respect to a region of high positivity in continuity with the axon, which gives rise to a current in the opposite direction from that of the general gradient and sufficiently strong to overcompensate it. In such a case the axon is in reality not growing up a gradient but down the local gradient to which it is reacting. Undoubtedly in the course of development of the central nervous system such local gradients fre- quently arise. For example, the approach of an axon or a group of axons to other cells less advanced in development or the localized activity associated with the development of a body segment may determine such polarization and direction of growth. An axon growing in a local gradient against a general axial gradient can of course grow only so far as its polariza- tion is maintained by the local gradient. Such axons are therefore likely to be short, while those growing down the general gradients are much longer. A very

interesting case in the development of the Amphibia is considered in another chapter (pp. 251-57). The aggregation of growing axons into nerve trunks is apparently also readily accounted for in terms of electrical polarization and growth. If the axon is internally positive, either on account of growth activity or later because of the passage of nervous impulses, it is negative externally and must therefore increase the positivity in that region of a later outgrowth which is toward its surface (second type of polarization) . Conse- quently the later axons will tend to grow along the sur- face of those which have arisen earlier, except where other regions of organs give rise to a sufficiently strong electrical field to alter these relations.

In the origin of the branches known as collaterals from an axon local electrical fields more or less at right angles to the axon direction may play a part in localizing regions of high positivity and growth at various points along the axon and so determining the occurrence of budding. The axiate pattern of the axon suggests, however, that internal physiological factors may also play a part in this branching. The growing axon unquestionably represents, for the time being, a physio- logical axis with the most active region at its tip, and some degree of dominance and subordination probably exists between the tip and other regions within a certain distance from it. I have repeatedly called attention to the occurrence of physiological isolation in relation to such gradients and to its significance for the processes of budding and branching in multiaxial plants and animals (see chap. v). The axon with its branches bears a close resemblance to the growth forms of various

plants and colonial animals with a main axis and lateral branches, and since it has been shown that physiological isolation is the primary factor in the development of such growth forms, quite independently of the specific constitution of the protoplasm, it is very probably concerned in the origin of branches from the axon. As the length of the axon increases, the effect of the tip on the more basal regions must decrease, until these regions are more or less physiologically isolated from it. Under these conditions, and perhaps also in relation to local electrical fields about the axon, budding may occur. Physiological isolation results not only through increase in distance from the dominant region, but also through decrease in activity of this region. In many central neurons growth of the axon ceases a short distance from the cell body and collaterals. In such neurons the appearance of the collaterals is probably associated with the cessation of growth in the main axon. Moreover, conditions such as this within the axon may render it more susceptible to the local electrical fields.

Kappers has called attention to the fact that the direction of growth of collaterals is usually at first more or less perpendicular to the axon from which they arise and accounts for this fact in terms of his hypothesis. From my point of view this direction of growth may be accounted for as follows: as soon as the growing tip of a collateral is established, its most positive region will be determined farthest away from the axon from which it arises, since this is more or less positive internally, either because growth is still going on, or because of the passage of impulses. Consequently the collateral will tend to grow at right angles to the axon for a certain

distance, or until it enters some other electrical field capable of altering its direction. The development of a terminal -arborization on an axon may result in some cases from its approach toward an externally negative region of large size. Under such conditions the growing tip will tend to spread out and give rise to various outgrowths, all growing more or less toward the negative surface. Again, arborization may perhaps result from the cessation of the action of the previously directing field, so that growth becomes indefinite in direction or responds to various local conditions, perhaps determined by adjoin- ing neurons.

If these suggestions are correct, the axon behaves during its growth period in many respects like other physiological axes, and its origin, like that of other axes, is determined by a quantitative differential in the action of a factor external to it. The appearance or non- appearance of medullation on the axon is of course a matter of differentiation and therefore in part deter- mined by the hereditary constitution of the protoplasm, though within the same individual differences in rate of metabolism in the developing axon are perhaps con- cerned in determining whether medullation shall appear or not.