The Origin and Development of the Nervous System

1 If the peripheral processes of the spinal ganglion neuron are axons (see p. 196) these cells constitute an exception to this general rule. certainty. The synapse may be a membrane of this sort as regards excitation, but we know nothing of the mechanism of such irreversibility. Since this theory of the synapse is still in large measure based on assumption rather than on experimental data, it is perhaps allowable to point out the possibility of a different interpretation of some of the characteristics of the reflex arc, ordinarily attributed to the synapse.

The functional polarity of the neuron is evidently quite distinct from its developmental polarity, as it exists during the growth period. It is clear that a relation of functional dominance and subordination exists between neurons, and such a relation may depend in part upon the order in which the neurons of a given path attain the stage of functional conduction of impulses and in part upon the frequency or the quanti- tative characteristics of functional impulses in other neurons in their vicinity. The facts indicate that the neurons of the higher levels of a particular gradient develop more rapidly than those of lower levels, and it is also probable that the nervous impulse represents, at least at first, either a steeper or a higher wave of excita- tion in a neuron of higher than in one of lower levels of a gradient. If differences of this sort exist, they must constitute important factors in determining not only the morphological connections of different neurons but the functional relations between them and particularly such matters as "openness'1 or "resistence" of func- tional paths.

As regards the pecularities of inter-neuronic conduc- tion which are usually attributed to the synapse, the possibility that certain characteristics of the dendrite as a conducting path and not merely the synapse itself, may be concerned deserve attention. The dendrite is much less highly specialized morphologically than the axon and is evidently concerned in nutrition and mainte- nance of the neuron as well as in conduction. We know practically nothing concerning conduction in the ordi- nary dendrite, but it is evident that the assumptions usually made concerning the synapse may be made, perhaps on even better grounds, concerning the dendrite. If the dendrite is less readily excitable than the axon, does not show the " all-or-none ' : type of reaction, and conducts with a steep decrement, it will serve as well as the synapse to account for the characteristics of the reflex arc commonly attributed to the synapse. Assum- ing for the sake of argument that the dendrite possesses these characteristics of the more primitive conduction path, an excitation conducted up the axon, through the cell body, and down the dendrite may be so slow or so weak or both, by the time it reaches the tip of the dendrite that it is unable to excite the axon tip with which the dendrite is in contact. The excitation wave in the axon, being steeper or higher, or both, than that in the dendrite and undergoing little or no decrement in its passage down the axon, is able to excite the dendrite or the cell body, with which the axon is in contact.

As regards delay and summation, it may require a number of impulses from the axon to bring the dendrite to a physiological condition in which it is able to initiate an excitation wave, high enough or intense enough to travel to and through the cell body to the axon. In the conduction of impulses toward the apical end of the ctenophore plate row following mechanical stimulation of a lower level, essentially this change occurs (p. 218). The first impulses travel only a short distance from the point of stimulation, but if the stimulation is continued they travel successively farther and farther, until after a time an impulse is produced which reaches the apical end of the row. In this case the continued stimulation and the passage of impulses upward doubtless affect both the condition of the path and the quantitative characteristics of the later impulses. This is obviously a sort of summation.

Students of the nervous system are agreed that the reflex is the unit reaction (Sherrington) or the physio- logical unit (Parker) upon which nervous function and integration are built up, but they are not in complete agreement as to what constitutes a reflex unit's simplest terms. According to Sherrington (1906), for example, a reflex arc consists fundamentally of receptor, conduc- tor, and effector, while Herrick's conception (Herrick, 1918) and Parker's (1919) also involves the presence of a central organ, the adjuster, and so of a central nervous system.

But wherever we may determine to draw the line between the true reflex and the simpler response it is evident that we have in the reflex arc, first, the initiation of excitation in the receptor by some external agent, second, the transmission of excitation over some path, the conductor, from the point of excitation, usually with more or less central complication, and third, the production by the transmitted excitation of some effect through the agency of a more or less specialized organ, the effector. In the higher animals even the simpler reflexes are rather highly adapted reactions in which the biologically purposive character is more or less evident, and the reflex arc in such cases represents a

considerable degree of physiological integration, since it correlates two more or less different organs, the receptor and effector, or, more strictly speaking, subordi- nates one to the other. Even in the primitive receptor- conductor-effector system of Vorticella within the limits of a single cell and without any visible nervous structure there is a considerable degree of integration and biologi- cal purpose. Every neurologist and physiologist recognizes that between the highly adaptive reflex arcs of the vertebrate with their highly specialized receptors, conductors, and effectors, and the sensitivity and conductivity of living protoplasm in general some sort of relation must exist. The problem of the physiological origin of this adapted unit reaction, the reflex, is a part of the problem of the origin and evolution of the nervous system in its physio- logical aspects and its consideration is the first step toward any adequate conception of the nervous system as a functional system.

Viewed from its physiological side the reflex consists in initiation, transmission, and end-effect of excitation induced by some external agent. From this point of view the relationship is evident between the reflex arc and the protoplasmic excitation gradient, which is the first step in axiate integration. In the primitive proto- plasmic gradient, receptor, conductor, and effector are not differentiated, but any part of the protoplasm or at least of the protoplasmic surface may perform any of these functions. The receptor is merely the point first or most intensely excited by an external stimulus, the conductor is any protoplasm which conducts or trans- mits the excitation, and the effector is any point excited

by the transmitted excitation. In fact each point of the path is successively receptor, conductor, and effector, the effect being excitation, consisting of chemical reaction and other changes, the liberation of energy, and the excitation of a further point. As long as the protoplasm concerned remains undifferentiated, such an excitation gradient can never develop a characteristic feature of the true reflex, viz., the specialized function of the effector, which constitutes a part of the founda- tion for the adaptive character of the reflex, but the general physiological basis on which the development of the reflex may occur is clearly present in such a gradient. In short, according to this viewpoint the physiological basis wrhich underlies and initiates orderly development and morphological and physiological unity and order in general is the same as the basis of nervous integration and of the most complicated nervous acts. The physio- logical axis and the reflex arc have a common origin in the protoplasmic excitation gradient.

For the sake of convenience and clarity it is necessary before proceeding farther to make certain definitions. An excitation gradient is the change resulting from irrita- tion or excitation at some point and its protoplasmic transmission with a decrement (pp. 75-82). In such a gradient each level may be successively receptor, con- ductor, and effector. Through gradual changes in irritability and protoplasmic conditions at different levels an excitation gradient may become a more or less permanent axial gradient or physiological axis of a whole organism or of some part of it. As soon as local speciali- zation and differentiation begin at different levels of a gradient it must be distinguished from a simple gradient

because the specialization may alter the primary rela- tions, e.g., an effector at a certain level, although subordinate to the receptor-conductor portion of the system, may show under excitation a more intense activity than levels of the gradient above it. An excitation system in which local specialization of receptor, conductor, and effector, or of some of them, has begun may then be distinguished physiologically from a simple gradient as an excitation arc. Morphologically as well as physiologically, of course, it may differ from the simple gradient in an indefinite number of ways. Except as regards the highly adaptive character and the central complications in the reflexes of higher animals, such an excitation arc is already similar physiologically to a reflex arc, though it is not necessarily associated with visibly differentiated nervous structure. With the evo- lution of the adaptive character, i.e., the preferential development, persistence, or low "functional resistance " of certain arcs which serve a biological purpose, the arc may be conceived, as Dewey (1893) suggests, as physio- logically a circuit.

It is perhaps necessary to call attention to the dis- tinction between organismic and protoplasmic pattern involved in what has been said concerning adaptation in the reflex arc. Even the simplest organism capable of responding in any way to the action of an external factor has unquestionably undergone adaptation as regards its protoplasmic pattern, but a physiological gradient in it may be a matter of direct response of this protoplasm to external factors. This gradient pattern itself is not adaptive in character, though the protoplasmic pattern on which it is based may be highly adaptive. A

simple excitation gradient, arising in direct response to external excitation, is not as a whole an adaptive pattern. The problem of the evolution of a true adaptive reflex from the protoplasmic excitation gradient is a problem of the specialization and adaptation of function and structure. The establishment of definite axial relations in the organism and so of definite relations to the external world affords a physiological basis for the localization and differentiation of definite receptors, conductors, and effectors, and with the beginning of the specialization of these as distinct organs, the reflex in its essential physiological features is present. The adaptive features of the reflex are chiefly in the central organ, the adjuster, and the relations of the particular arc to it. In this aspect the reflex arc becomes a product of evolution, and its development is a historical process associated with biological purpose. At present we are concerned rather with the discovery of the physiological basis of such adaptive evolution than with the evolution itself.

If the physiological basis of the reflex arc is an excitation gradient of the same sort as the gradients which underlie axiate order and integration, we should expect to find receptor, conductor, and effector associ- ated with each other, not independent in their evolu- tionary history. The relation between receptor and conductor is self-evident and leaves little room for dis- cussion, but as regards the relation of the effector to the rest of the arc the physiological facts are not so clear, and a wide difference of opinion has existed.

Kleinenberg (1872), for example, held that the primitive condition with receptor, conductor, and effector within the limits of a single cell is found in the neuro-muscle cell of the Hydra ectoderm (Fig. 68). The outer exposed surface of the cell is the receptor, the elongated contractile base the effector, and the intermediate region the conductor. The effect of excitation in this unicellular arc is motor in character, a simple contraction of the basal effector portion of the cell. Ac- cording to this conception, re- ceptor, conductor, and effector develop in functional rela- tion, first within a single cell, •ag^a^ and only later in the course FiG.68— Anepithelio-muscle of evolution become sepa- cell (" neuro-muscle cell") of rated into different cells. The

Hydra, the contractile basal brothers Hertwig (l878) agree portion indicated by shading. the simultaneous development of nerve and muscle, but their studies on the medusae lead them to believe that receptor, conductor, and effector arise as separate cells (Fig. 69). Claus (1878) and Chun (1880) maintained that the nervous system arises independently of muscles and only secondarily becomes connected with them, while Samassa (1892) and Schaeppi (1964) hold that a nervous system without effectors is difficult to conceive. More recently Parker (1909, 19100, 19106, 1911, 1914, 19170, 1919) concludes from his study of sponges that muscles appear first as independent effectors and in connection with them the first true nervous organs in the form of receptors, sensory cells, develop, a condition very similar

to that in the coelenterates, while the central organ, the adjuster, is of still later origin. With the exception of Kleinenberg's conception of the neuro-muscle cell, these hypotheses and suggestions are concerned more with the origin of the nervous system as a morphological structure than as a physiological entity. A brief con- sideration will show that from the physiological point of view these different views appear rather as special cases of a general physiological process than as mutually exclusive conceptions.

FIG. 69. — Diagram of simple multicellular arcs: s, sensory cell (receptor); n, nerve cell (conductor); m, m, muscle cells (effectors). If the reflex arc develops from the protoplasmic excitation gradient, the development and evolution of the nervous system as a morphological structure must follow in general the lines of the chief or the most persistent excitation gradients established in the pro- toplasm of a particular species. According to the con- ception of physiological axiation developed in the earlier chapters, an excitation gradient may arise in primitive protoplasm at any point of the surface where excitation occurs, and any particular gradient may persist only momentarily or for a short time, as is apparently the case in Amoeba (Hyman, 1917). Under

such conditions the development of a definite, persistent receptor-conductor-effector structure is of course impos- sible. But where the gradients persist, whether from the surface inward or as axial gradients of the body as a whole, they afford a physiological basis for the develop- ment of a structural nervous system. In a multicellular organism, such as Hydra, where the body cells form coherent epithelial layers and only certain portions of the cell surface are directly exposed to the extra-organic environment the conditions exist for the development of an excitation gradient and a physiological axis in each ectoderm cell with the exposed surface as receptor, and this pattern may coexist with a general axiate pattern of the whole body. The visible differentiation of an effector region, e.g., a contractile portion, in such a cell may occur before anything recognizable as nervous structure is visible, for reception and transmission of excitation apparently do not bring about modification of proto- plasmic structure as rapidly or as extensively as the development of contractility. In short, the neuro- muscle cell of Kleinenberg is exactly the sort of system we might expect to find in a cell layer writh definite surface relations to the exterior. It represents simply one possibility in the origin of a receptor-conductor- effector system, and the grounds for doubting its exist- ence are morphological rather than physiological.

Parker's conclusion that the myocytes of sponges are independent effectors is doubtless morphologically correct, but physiologically either some portions of these myocytes themselves function as receptors and con- ductors or else other cells perform these functions. The presence or absence of anything recognizable as nerve cells or fibers is physiologically of secondary importance. As a physiological entity the receptor- conductor-effector system is present in the sponges, either in single cells or in cell series, and if the myocytes react directly, without relation to other cells, they themselves must be physiologically more or less like neuro-muscle cells. Physiologically speaking, a pure effector is impossible, for even the most highly specialized effector must be capable of receiving stimuli from with- out. As a matter of fact every part of the reflex arc like the primitive excitation gradient functions as receptor, conductor, and effector, at least in one direction. Nevertheless the different regions of the arc are in the broad sense predominantly receptors, conductors, and effectors, as their differentiation indicates. In this sense it is undoubtedly true, as Parker maintains, that effectors appear, i.e., become morphologically distin- guishable earlier than nervous structures. The condi- tion in the medusae, as described by the Hertwigs, in which receptor, conductor, and effector are different cells, does not by any means exclude the possibility of the existence of the neuro-muscle cell, i.e., the system in unicellular form, even in the same individual. In Figure 69, for example, where the system as conceived by the Hertwigs is shown, it is certainly possible that the contractile basal portion of the cell m may be activated not only by the receptor-conductor system represented by the other two cells, s and n, but by excitation at its own receptor surface and transmission in its own protoplasm. Such a cell may be a neuro- muscle cell and at the same time its contractile portion may be the effector of one or several other arcs.

Moreover, the excitation arc may arise not only within a single cell or within a small group of adjacent cells, but also in relation to the axiation and integration of the body as a whole. In other words, not merely intra- cellular gradients and local multicellular gradients, but the general axial gradient or gradients of the whole body may constitute the basis of excitation arcs and so of nervous function and structure. My point is then that intracellular arcs, local multicellular arcs, and what we may call general axial arcs may all exist in the same individual, and so far as nervous develop- ment and differentiation is concerned, all be equally primitive. It is not necessary to assume that nervous structure was at first completely diffuse and equally distributed in all parts and that centralization occurred only in the course of evolution. If the physiological gradients in the body constitute the basis of the localiza- tion of nervous structure, we must expect to find more or less nervous centralization in any axiate animal in which nervous structure differentiates. A completely uncentralized system is possible only when axiate pattern as regards the body as a whole is absent, but, as a matter of fact, all animals in which a nervous system differentiates are more or less axiate. Whether the unicellular arcs, the local, or the general axial arcs are the more important in a particular case must depend upon the degree of development of physiological axiation and so of physiological integration in the body as a whole. In the medusa, a motile form, the development of axiation is higher and the general arcs as represented by the marginal sense organs, the ganglionic aggregation, and the nerve net are evidently more highly developed

than in the sessile coelenterates. It may be pointed out, however, that in the sessile coelenterates there is normally a reversal of polarity during development. The planula possesses a well-marked apico-basal gra- dient, but as the period of attachment approaches this gradually disappears, and finally it is the original apical end which becomes attached, and the sessile individual arises as a new axial gradient, essentially a bud, at the original basal end. Moreover, in these sessile forms axiation is never very highly developed. Again in the sponge, which Parker regards as representing the primi- tive condition, the early developmental motile stage possesses a definite physiological axis, i.e., a primitive physiological gradient exists, but this disappears more or less completely during the later development, and the sessile adult exhibits an extremely low degree of physiological axiation and integration, so that practically nothing but local or unicellular arcs exist. But the sessile sponge and the sessile coelenterate develop from motile forms and the almost anaxiate condition oi the adult is not primitive but secondary. Looking at the matter from this point of view, it is scarcely possible to agree with Parker, either that the contractile cells of the sponge are pure effectors or that the conditions in the sponges and sessile coelenterates represent the first step in the evolution of the receptor-conductor-effector system in free-living forms. They appear rather as special cases in which a form originally free-swimming and axiate has become sessile and almost anaxiate. Corresponding to the very low degree of integration in the adult sponge no distinguishable nervous structure is present, and in the unicellular or local arcs only the

effector portion is visibly differentiated. In the actinian some degree of axial integration is undoubtedly present though it may be slight, as Parker's experiments indicate. If the free-swimming sponge larva developed into a free-swimming adult instead of becoming sessile, there is no doubt that the axiation and physiological integra- tion would progress instead of disappearing to a large extent and that one factor of this progressive develop- ment would be some sort of central nervous system. In such a case the local excitation arcs consisting of one or few cells would be more or less completely subor- dinated to the great general arcs of the chief axis. In short, the course of development of the excitation arc and the structures which represent it must be different in a form which remains motile throughout life from that in a form which becomes sessile after an early motile stage. In a free-living form nervous differen- tiation may occur as early as, or even earlier than, the differentiation of effectors, and this condition may be just as primitive or even more so than in the sponge, for since the nervous system is so intimately associated with motility we should scarcely look for the most primitive conditions in forms which are secondarily sessile.

Histologically, the condition in the sponges certainly supports the views of Claus and Chun that muscles and nerves originate independently and of course the embryonic differentiation of muscles in complete inde- pendence of nervous connections in the higher animals may be regarded as pointing to the same conclusion, but the origin and differentiation of muscles as of other organs is unquestionably associated with physiological gradients or excitation arcs, either local or general or both, and excitation and its transmission occur in proto- plasm where no visible nerve structure exists.

Finally, the difficulty of conceiving a nervous system without effectors which was pointed out by Samassa and Schaeppi is a real difficulty, but when we remember that primitive protoplasm is effector as well as receptor and conductor and that passage of excitation over any protoplasmic region involves reception, conduction, and effect, this difficulty remains only for the teleological point of view. When we consider the nervous system as an adaptation for biological purposes, its existence without adapted effectors is difficult or impossible to conceive, but when we consider it as the visible expres- sion and record in protoplasm of an excitation gradient or gradients resulting primarily from a differential relation to the external world, the presence or absence of a specialized effector becomes a matter of minor importance. In the first case we are considering the nervous system as a means, and without relation to the end — the effector or the effect produced by it — it becomes inexplicable, while in the second case we are considering it as an effect or result, and as such it follows necessarily from the irritability of protoplasm, the gradients, the capacity for morphological differen- tiation, etc.

The question of the localization of receptor, con- ductor, and effector and of what may be termed the functional direction of the reflex arc with respect to the axial gradients remains to be considered. As regards the coelenterates, the work of Parker on actinia ns and Corymorpha and the work of others on medusae1 indicate that in these forms nervous transmission is highly diffuse and may occur in all directions from most or all regions of the body in the actinians and at least of the subumbrellar surface in the medusae. The experimental data do not indicate any differences in conductivity or range in any direction with respect to the body axis. I believe, however, that crucial experi- ments bearing upon this point have not yet been per- formed. Parker's experiments on the actinians, for example, are not adapted to show whether axial differ- ences in irritability, conductivity, and range are present or not. For this purpose stimuli graded in intensity would be necessary, and aquatic forms are not favorable material for work of that kind. Mayer and Harvey have shown that impulses may apparently continue to travel indefinitely without any indication of decrement in circular strips of the medusa, Cassiopea, from which the sense organs have been removed, but they have not determined whether there is any difference in conduc- tivity in the direction away from and toward the sense organs. There can be no doubt that in the medusae the sense organs are the most irritable regions and therefore in nature usually the initiators of impulses.

In the free-swimming medusae the specialization of both receptors and effectors is far more advanced than in the sessile coelenterates, and the whole receptor- conductor-effector mechanism is much more highly developed on the apical body surface than elsewhere. Other polar specialization is not very great, conse- quently we need not expect to find any very marked difference between general axial arcs and others, except perhaps in the elongated manubria of certain species and in the tentacles, and in these organs we do find evidence of more or less definite axial arcs (see p. 112).

It has been generally maintained that some degree of structural axiation exists in the nervous system of the sea anemones and corals and probably in that of the hydroids. According to this view, there is a more or less distinct aggregation or a greater development of nervous tissue in the apical region in relation to the mouth and tentacles. Wolff's (1903) figures of the general plan of the nervous system of the hydroid and the coral polyp indicate a considerable degree of axiation and centralization, but these figures, while based on the data of observation, are admittedly diagrammatic. If any degree of axial localization of nervous tissue and of receptors does exist in these sessile coelenterates, and the facts permit little doubt on that point, some degree of functional axiation undoubtedly exists in the excita- tion arcs, even though experimental stimulation not quantitatively controlled may show conduction occurring in all directions.

Many facts indicate that in these forms the receptor function is most highly developed in the apical regions, probably in the tentacles and about the mouth, and these regions are normally most exposed to excitation. In the tentacles the excitation arcs are as already noted, predominantly basipetal in both structure and function (pp. 112, 203), and there can be no doubt that in the body as a whole there is some degree of apico- basal integration, with the apical region acting as receptor and dominant. In certain elongated actinians such as Cerianthus and Harenactis, which live in tubes in the sand with the apical end exposed, this is very clearly the case.

The functional direction of the arcs in forms with a synaptic nervous system is generally believed to be determinate and irreversible as regards the arc as a whole, although conduction supposedly occurs in both directions in the axon. The synapse, the region of contact between two neurons, where passage of a cell membrane is involved in conduction of the impulse, is usually regarded as responsible in some way for this functional irreversibility of the arc (see pp. 228-32). The evidence for irreversibility, however, has been obtained chiefly from the higher vertebrates in which it appears to be a characteristic feature of the arc, but it is possible that even in these forms some paths in the central nervous system may be reversible as regards functional direction. If it is true that in the synaptic nervous system the reflex arc is functionally irreversible,, then different or opposed functional directions must mean different paths. Certain conditions in some of the lower bilateral invertebrates are of interest as indicating the relation between functional direction and the chief axial gradient, though regarding the actual paths concerned in anterior and posterior directions we know practically nothing.