The Origin and Development of the Nervous System

The problem of the factors concerned in the origin and development of the dendrites appears much less simple than the problem of the axon, but the facts afford a basis for certain suggestions. In general the dendrites appear much later than the axon, they are evidently less highly specialized structures, their growth is usually less rapid, in some neurons they arise from all parts of the cell surface and some of them may show definitely directed growth, while others belonging to the same neuron grow in various directions, and in many cases only certain dendrites persist, others under- going resorption in the course of development. More- over, the dendrites undoubtedly are of great importance as nutritive organs of the neuron, as well as organs of conduction.

As regards the factors concerned in dendrite develop- ment, it appears necessary to recognize two distinct aspects or phases of the developmental process, the origin and localization of the dendrites on the cell body and the later orientation and definitely directed growth of some or all of them. In connection with dendrite origin two possible factors suggest themselves, viz., physiological isolation and the nutritive relations of the cell with its environment. Physiological isolation may lead to budding on the cell body as well as on the axon. The fact that the dendrites usually arise only when the axon tip has attained a considerable distance from the cell body, or has ceased to grow, favors this possiblity. From this viewpoint the origin of the dendrites is in some degree similar to the origin of the basal stolons in hydroids and perhaps of the root system in plants. But it is also possible that nutritive conditions in the cell body play a part in determining these outgrowths as they apparently do in root formation.

As regards the orientation and definitely directed growth of dendrites, we must apparently look to the electrical factor. By the time the dendrites arise, the developing neurons have become more or less separated from each other by the development of neuroglia and intercellular colloidal substance containing water and salts. The dendrites which show definite orientation grow toward certain axons or axon tracts of other neurons and establish synaptic contacts with them. Assuming that these tracts are regions of high developmental or functional nervous activity, electrical conditions are those of the second type of polarization (Fig. 59) in which the cell n and the electronegative surface a are not in continuity, but separated by a conducting medium. Under these conditions the dendrites must grow toward the electronegative surface of the axon tract. It is quite unnecessary to assume, as Kappers has done, that the dendrites react differently to the electric current from the axon. The physiological effect of polarization is the same in both axon and dendrite, but the two outgrowths develop, so to speak, in the two different halves of the electrical circuit, in which the currents flow in opposite directions.

The apparently selective or specific relations between the axons of certain neurons and the dendrites of others may then be determined by the developmental or func- tional activity of the one and its effect on the direction of growth of the other. Different dendrites of a par- ticular neuron may enter different electrical fields and so establish connections with different neurons, and it is conceivable that some dendrites may remain nutritive organs.

The dendrite is evidently a less highly specialized axis than the axon, and its frequent branching suggests a low degree of dominance during growth. It reacts chiefly with reference to local conditions rather than to the more general gradients, because at the time of its development the simpler general gradients of earlier stages have undergone complication in various ways. The neurons of the spinal ganglia of the vertebrates present certain special problems. When the neural tube arises by the development of the neural folds and the borders of the neural plate (Fig. 62) and the meeting and fusion of these folds (Fig. 63), some of the cells of the dorsal border of the folds, the neural crest, are not incorporated in the neural tube, but migrate ventrally on each side of it (Fig. 64), forming, according to Johnston (1906, p. 50), segmental series of flaps attached to the dorsolateral surface of the cord and hanging down on its lateral surface. Later these flaps may separate entirely from the cord and form an aggregation of cells which constitutes the primordium of the spinal ganglion and lies between the cord and the mesoderm (Fig. 65).

In the process of differentiation each of these cells which becomes a neuron gives rise first to an outgrowth which extends dorsomedially and enters the cord, and somewhat later to a second outgrowth, which grows in the opposite direction (Fig. 60). Both of these out- growths are structurally similar to the axons of other neurons and both may develop medullation. Function- ally they constitute the afferent or sensory paths, and the direction of functional conduction in the peripheral outgrowth is toward the cell body instead of away from it, as in other axons. The question which of these out- growths is axon and which dendrite, or whether both

FIGS. 62-65. — Developmental stages of neural tube and spinal ganglia in salamander, Amblystoma punctatum: Fig. 62, one side of neural plate, showing early stage of neural fold; Fig. 63, neural tube just after closure, cells of neural crest indicated by shading; Fig. 64, later stage, showing beginning migration of spinal ganglion cells; Fig. 65, still later stage; cells of neural crest have migrated lateroventrally to form spinal ganglion (from Johnston, 1906).

are axons, as they appear to be, is answered differently by different authorities. . Some, taking into account the fact that the centrally directed outgrowth arises first and conducts away from the cell body, while the peripheral outgrowth appears later and conducts toward the cell body, regard the central outgrowth as axon, the periph- eral as dendrite. When the peripheral process is separated from the cell body, chromatolysis occurs in the latter, as in other neurons after separation of the axon from the cell body, and this fact has led some to regard the peripheral process as the axon. Still others conclude from the structural characteristics that both processes are axons.

As regards the developmental factors concerned in this process of differentiation, three possibilities suggest themselves. The first is that the cells become electrically polarized while they are still continuous with the cord as the flaps mentioned above. Such polarization is of the first type (Fig. 58), and the region of high positivity determined by it must be ventrolateral. If this is the case, we should expect these cells to give rise first to axons growing ventrolaterally, but the ventrolateral migration of the cells may be a reaction to this polariza- tion and may delay the axon outgrowth. As these cells lose their continuity with the cord, its electronegative external surface must tend to produce a second polari- zation in them, with the region of high positivity toward the negative surface, i.e., the dorsolateral regions of the cord. This is a polarization of the second type (Fig. 59) , and it may also be a factor in retarding the outgrowth of the peripheral process because it tends to obliterate and reverse the primary polarization. If the effects of the

first polarization are not immediately reversible, these cells may develop two growing regions at opposite poles and become strictly biaxial. The second possibility is that the polarization giving rise to the dorsomedially directed outgrowth is the first effective polarization and that the peripheral outgrowth is determined only after the central process attains con- tinuity with the cord. If this is the case, the migration of the cells before the outgrowths appear must be regarded as due to mechanical or other factors and not to polarization.

The third possibility is that both processes may represent reactions to polarization of the second type, the central process with respect to the most negative region of the cord, the peripheral with respect to the less strongly negative surfaces of peripheral parts, e.g., perhaps primarily the myomeres. Which of these possibilities or whether any of them represent the actual course of events, it remains for the future to determine. They are not entirely mutually exclusive and each of them may be in part realized at some stage. In any case, however, it is evident that as soon as the centrally directed process attains con- tinuity with the cord, the direction of growth of the peripheral process is determined in the same way and in the same general direction as that of the axons grow- ing out from the ventrolateral regions of the cord

As regards the differentiation of both outgrowths as axons, various facts indicate that axon and dendrites do not represent reactions of different kind but rather of different degree, the more rapidly growing, more active, and probably more highly polarized region becoming an axon, the less active and less highly polarized, or perhaps not primarily polarized, a dendrite. Both outgrowths being highly polarized and active in this case, both develop as axons and the functional polarity of these cells, i.e., the direction in which they conduct, is determined, not by the fact that one out- growth represents a dendrite, the other an axon, but by their relations to other parts. The reason for the occurrence in the cell body of chromatolysis following the section of the peripheral and not that of the central outgrowth, is not at present clear. It may be that the peripheral outgrowth makes greater demands on the cell body in the matter of nutrition and maintenance than the central, and therefore the effect on the cell following section of it is greater than that following section of the central outgrowth. The later change of the spinal ganglion neurons from the " bipolar'1 to the umonopolar': type by the approach and union of the bases of the two outgrowths is perhaps the result of "neurobio taxis," which is considered below.

Kappers and others have shown that in many neurons and neuron groups the cell body undergoes an apparent migration in the ' direction of growth of the chief dendrites, i.e., toward the region with which the neuron or group is functionally correlated. So far as this migration is an ontogenetic, not a phylogenetic, process, it is doubtless determined by the factors which direct the growth of the dendrites and probably results from some degree of polarization of the cell body. On

its more positive side, at this stage the side of the chief dendrites, the cell body probably adds to its sub- stance by growth, while on the axon side it furnishes material for maintenance and further growth of the axon and therefore appears to migrate toward the region determining the polarization. Attention has already been called to the fact that in earlier stages, preceding or during axon origin, the cell body often appears to migrate in the direction of axon outgrowth (p. 187). The two reactions are apparently not differ- ent in kind, but represent rather the same reaction under different conditions.

Certain neurons, occurring particularly among the lower vertebrates in certain regions of the nervous system and as a characteristic feature of the invertebrate nervous system, possess a single main outgrowth or extension of the cell body, which for some distance remains dendritic in structure and usually gives rise to dendritic branches, but sooner or later decreases in diameter and becomes an axon (Figs. 51, 52, 53). The primary outgrowth of such cells apparently corre- sponds, as regards its origin to the axon, although it is structurally dendritic and only gradually reaches the degree of specialization represented by axon structure. In fact, in some cells the outgrowth ends in dendritic branches and never gives rise to an axon. The occurrence of such neurons indicates clearly enough that dendrite structure and axon structure do not mean opposite reactions but rather different degrees of the same reaction. Either the degree of

polarization and of the resulting growth activity is not sufficient to give rise at first to axon structure but gradually becomes so, or else these are cells which react to polarization more slowly than the ordi- nary type of neuron. Bethe has shown that in cer- tain crustacean neurons of this general type (Fig. 52) the physical isolation of dendritic and axonic portions of the cell from the nucleate portions by section of the basal unbranched region of the outgrowths does not interfere with conduction from the dendrites to the axon. Evidently the nucleate portion of the cell has little or no direct relation to the nervous function but is concerned primarily with maintenance. This type of neuron represents then a gradual and partial speciali- zation of the cell body, presumably in the region of high positivity determined by the polarization of the cell.

These cells (Fig. 66), characteristic not only of the simplest nervous systems in the lower invertebrates but of certain organs in the higher animals, even the vertebrates, apparently have attained little or no definite developmental or morphological polarity. Their branching form and their apparent continuity with other cells may be due, at least in part, as Parker has suggested, to the fact that they represent the persistence of an embryonic syncytium. The absence of a definite structural polarity in these cells may be accounted for either by the indefiniteness and inconstancy of the physiological gradients in the organisms or organs con- cerned and the consequent indefiniteness and incon- stancy of polarization or by the inability of the cells to

react to such polarization as does occur. Both these factors may be concerned in many cases. Even in the lower invertebrates, however, the reticular nervous structure without definite morpho- logical polarity occurs chiefly in those portions of the nervous system which consti- tute the connecting links be- tween the peripheral receptors and the muscles or other effectors. The peripheral re- ceptors, which are in direct connection with the external surface of the body and there- fore with the physiological factors concerned in surface- interior pattern, do possess a more or less definite polarity, and may give rise to an out- growth which shows definite growth direction with refer- ence to an axial gradient. The receptor cells of the tenta- cles of certain sea anemones, for example, are not only elongated vertically to the surface, but give rise at their inner ends to nerve fibers which grow chiefly down the tentacle, i.e., down the gradient (Fig. 18, p. 112). A more or less definite polarity of this sort is a general characteristic of the peripheral receptors, even in the simplest forms in which we can distinguish definite nerve cells. Even the epithelial muscle cell of Hydra (Fig. 68, p. 238) wrhich, as I have suggested in another chapter (p. 240) , is probably to some extent a receptor and conductor as well as a muscle cell, shows a definite polarity with respect to the body surface. Evidently

FIG. 66. — A cell of the nerv- ous reticulum of a coelenterate; branches of such cells are ap- parently protoplasmically con- tinuous with those of other cells of the reticulum. polarity may arise in the cells of these simple forms in relation to the surface-interior pattern and in some cases in relation to the axiate pattern. There can be little doubt that in these forms the surface-interior pattern is more definite and constant than the axiate pat- tern and that the physiological factors concerned in it are therefore more effective in determining, at least locally, a polarity and pattern in individual cells, than the general axiate pattern of the animal. If electrical polarization is concerned in determining such pattern, it is evidently of the first type, since there is tissue- continuity, and the basal outgrowths of such cells correspond physiologically, as regards origin and direc- tion of growth, to the axons of the more highly special- ized neurons.

This and the preceding chapter constitute merely a brief consideration of some of the more important aspects of neuton development. The present chapter is nothing more than a tentative development of a conception of neuron polarity and external morphology on the basis of physiological factors concerned in organismic pattern. The factors themselves exist, and I have attempted to show that they afford a basis for interpretation of many facts of neuron development in relatively simple terms. That neuron pattern is and must be related physio- logically to organismic pattern is, I think, beyond the possibility of doubt, and if the conception of organismic pattern developed in earlier chapters has any foundation in fact, it must afford a basis for attack on the neuron problem. This chapter is concerned with showing that it does afford such a basis.

No one is more keenly aware than myself of the hypothetical character of the discussion and of the criticisms which may be brought against it. It is evi- dent, however, that a consideration of the origin and development of the nervous system, such as has been attempted' in this book, cannot stop short before the problem of neuron pattern and refuse to consider it. Neuron pattern is a feature, both of nervous and of organ- ismic pattern. Moreover, the neuron itself develops an axiate pattern, and the resemblance of this to certain types of axiate organism, suggests the possibility of physiological resemblance. I have merely suggested an interpretation of neuron pattern in terms of organismic pattern.1

1 Since this and the preceding chapter were written, I have been informed through the kindness of Dr. R. G. Harrison, under whose direction the work was done, that Dr. Ingvar has succeeded in determin- ing the direction of outgrowth of neuroblasts in tissue culture by means of electrical currents. A report of Ingvar's work is now in press and soon to appear in Proceedings of the Society for Experimental Biology and Medicine. This work represents, so far as I know, the first step in experimental determination of polarity in neuroblasts and, so far as it goes, constitutes an experimental basis for a hypothesis of electrical determination, but much further work is of course necessary to deter- mine, whether or how far the particular hypothesis advanced in these chapters will account for the facts.

In chapter iv the part played by excitation and trans- mission in originating and establishing the physiological gradients was discussed. In chapters vi-xi it was shown that the physiological gradients afford a general physiological basis for the localization, the course of differentiation and the general plan of structure of the nervous system as a whole, as well as for tentative con- clusions concerning the origin and development of neuron pattern. Since there can be no doubt concern- ing the close relation between nervous structure and nervous function, i.e., the conduction of impulses, and since the facts indicate that the physiological gradients originate in excitation and its protoplasmic transmis- sion, there is every reason to believe that these gradients, once established, must play a part in the development of the more or less specialized forms of transmission, which are usually called conduction. The gradients provide, first of all, a mechanism for giving conduction a more or less definite, or at least a preferential, direction and thereby for determining in some degree the physio- logical relations to each other of various parts. The pur- pose of the present chapter is to call attention to certain facts in connection with the transmission or conduction of impulses in paths which do not show any distinguish-

able nervous structure and to consider certain aspects of the problem of conduction in nervous tissue. It was pointed out in chapter iv that protoplasmic transmission in its most primitive form is probably little or nothing more than the gradient in physio- logical condition associated with the electrical gradient which results from a local excitation. With the develop- ment of protoplasmic excitability and conductivity, the transmission of what we call impulses, which consist of more or less definite waves of excitation, becomes increasingly possible. And, finally, the specialization of more or less definite paths affords the basis for a more definite conduction, as distinguished from the general protoplasmic transmission.

It is a familiar fact that the excitation of living proto- plasm depends not merely upon the amount of change in action of an external factor, but also upon the rate of change. In fact the rate of change is, in most cases, if not always, more important than the amount. Similarly, the rate of change at which the excitatory process takes place at a given point is a factor of fundamental import tance in determining its effectiveness as a means of excitation of other points or regions, and so of trans- mission. A slow excitatory process is likely to be rela- tively ineffective, a rapid one much more effective. The development of excitability and conductivity in protoplasm very evidently consists in large measure of an increase in rate of the excitatory process and of the consequent increase in its effectiveness in exciting other points. Such development is a process not only of

ontogeny but of evolution, and as such is of course associated with changes in the hereditary mechanism of excitation, however we may conceive such changes to be brought about. The ascending phase, the front of the excitation wave or impulse, is itself essentially a physiological gradient which, instead of being stationary, advances over the protoplasm. The descending phase of the wave represents the fact that this gradient is readily and rapidly reversible, at least to a large extent. The more rapid the change from the unexcited condition to the completely excited condition, the steeper the gra- dient, and the steeper it is the more effective it becomes in inducing excitation at other points. With the increase in conductivity, i.e., the increase in the effective- ness of the excitation at any point in bringing about excitation at another point, the decrement becomes less and less steep, until in the vertebrate nerve fiber a con- dition is attained in which the "all-or-none law" holds, and conduction supposedly occurs without a decrement (see pp. 8 1, 223). In such protoplasm a certain minimal degree of stimulation sets off, so to speak, the whole mechanism of excitation and a maximal excitation results automatically, just as a tremendous explosion may result from the reaction initiated by a minute spark. Concerning the actual physicochemical basis of these changes in the character of excitation in protoplasm we know little, but concerning the occurrence of the change there can be no doubt.

In this development of excitation rhythmicity fre- quently appears, that is, instead of the continuous existence of a state of excitation and its transmission or conduction definite waves or impulses appear, alter- nating with periods of rest. In one form or another rhythmicity in excitation is of very wide occurrence in organisms and ranges from such phenomena as the rhythmic impulses to ciliary movement to the heart beat. At this time it is impossible to do more than note the occurrence of rhythmicity.

While the development and evolution of the processes of excitation and conduction involve changes in the hereditary mechanism of protoplasm, there are certain features of the conduction pattern which are clearly related to the physiological gradients and certain others which may be interpreted in terms of such relation. Some of these relations have already been mentioned, particularly in chapters iv and v, but certain facts and questions remain to be considered.

It is a familiar fact that in ciliated surfaces, whether composed of single cells or of many cells, forming an epithelium, the beat of successive cilia occurs in a definite sequence of such a character as to give the appearance of successive waves of bending and recovery of the cilia, passing over the surface. Such ciliated surfaces cover large parts or, in many cases, all of the external surfaces of many of the lower aquatic inverte- brates and of early developmental • stages of numerous animals, both invertebrates and vertebrates, and are also found on many internal surfaces in various forms, ranging from the sponges to man. Each beat of a cilium is obviously due to an exciting impulse of some

kind, and some sort of transmission must be concerned in the progress of the wave of beats over the surface. These ciliated surfaces occur in many animals and in stages of development where no distinguishable nervous structure exists, and even in cases in which a well- developed nervous system is present it has not been generally demonstrated that ciliary movement is under nervous control, although Putter (1903) states that in certain animals such control does occur. Verworn (1890) maintained that transmission in such cases is purely mechanical, each cilium in its movement touch- ing the next and so exciting it. Kraft (1890) and others have shown, however, that Verworn's view is incorrect, since transmission of the impulse may occur through a region in which the actual movement of the cilia has been completely inhibited by cold and after passing through such a region is still able to excite the cilia beyond. Briicke (1916) also found that when pieces of ciliated epithelium were cut out and reim- planted in the same spot in reversed position with respect to the direction of transmission, they showed a direction of transmission opposite to that in other regions of the epithelium. Such pieces evidently retain the original direction of transmission quite independ- ently of any mechanical action of the cilia of adjoining regions. Mechanical transmission may perhaps occur in some cases of ciliary activity, but it is certainly not the fundamental factor in the passage of impulses over ciliated surfaces. The impulse is evidently an excita- tion of some kind and its transmission a protoplasmic process, but no definite conducting path is distinguish- able. Parker (1905) has called transmission of this