The Origin and Development of the Nervous System

The starting-point of Lillie's theory is the generally accepted fact of electrical polarization of the plasma membrane, the existence of which is well established by the work of numerous investigators. This polarization is usually regarded as determined by the fact that the membrane in the unexcited condition is normally per- meable to ions of one sign, and impermeable, or less permeable, to ions of opposite sign. The facts indicate that the unexcited plasma membrane acts as if per- meable or reversible to certain cations and impermeable or less permeable to certain anions which are present in higher concentration in the cell than in the surrounding medium. The evidence indicates further

that the change in permeability associated with stimulation is not a direct effect due to merely physical changes in the proto- plasmic surface layer, but is the consequence of a chemical reaction which alters the character of the surface film and temporarily deprives it of its normal semipermeable and electromotor prop- erties, and that this chemical process may in a highly irritable tissue like nerve be initiated by any slight local decrease in polari- zation, provided t1^ change is sufficiently rapid [R. S. Lillie, 1914, p. 443].

In various other papers (e.g., R. S. Lillie, 1917, pp. 80- 82; 1918, p. 3; 1919, p. 458) Lillie points out that the chemical change involved in the process is a metabolic reaction, probably oxidative in nature. Whether the primary change in the membrane consequent upon depolarization is this chemical reaction or a change in colloid condition, or some other change is at present uncertain, but since the whole series is indissociable and more or less mutually related it makes little differ- ence which we call primary and which secondary. For present purposes the significant fact is that, according to Lillie, excitation involves a chemical change, probably oxidative hi nature, as well as other changes, in the sur- face layer of protoplasm.

FIG. 6. — Diagram indicating the electric currents arising in relation to an excitation gradient; the direction of migration of the positive ions is indicated; for further explanation see text. The process of transmission of these changes from point to point is accomplished by the electrical changes which constitute one feature of the excitation. The excited region (Fig. 6a) becomes electronegative (exter- nally) to an unexcited region (b, c, etc.), in other words a local electric circuit is established, the positive current passing in the cell or tissue from the region of excitation to unexcited regions with a " re turn'1 current in the external medium. In consequence of resistance the strength of this current decreases with increasing dis- tance from the region of excitation a, that is to say, a gradient in electrical potential must arise in relation to a region of excitation. Lillie points out that this local

electric current arising in relation to the excitation of a given region a affords the condition necessary to bring about excitation at b, c, etc., i.e., at all points within a certain distance from a where the strength of the current is sufficient to bring about the changes determining excitation. In short, according to this theory of excita- tion and transmission, the electrical changes involved in the excitation of any point a, however that excitation be produced, determine the transmission of excitation from a to b, c, etc. The important point for present purposes is that in consequence of the increasing resist- ance with increasing distance from the point a, a gra- dient in electrical condition of the plasma membrane results from the excitation at #, the (external) electro- negativity decreasing from b to some point at which the resistance becomes so great that the electrical effect of the excitation at a is zero. If Lillie's conception is correct every local excitation gives rise for the time being to such a gradient.

According to Lillie, then, the electrical change pro- duced by excitation at a becomes the " stimulus" which brings about excitation at b, c, d, etc., at all points, that is, within the limit of its physiological effective- ness. If it is true, as the facts indicate (see pp. 80-81), that in protoplasm in general, where specialized con- ducting paths do not exist, a relation between the energy or the intensity of excitation and the energy or the intensity of stimulus exists, it follows that when excita- tion occurs at a given region a of such protoplasm a gradient in the degree, energy, or intensity of excita- tion, i.e., in its ability to act as stimulus for other points, must result in consequence of the electrical conditions,

this ability decreasing from a through b, c, d, etc., to a point x where the electrical change becomes so weak that it is incapable of acting as a stimulus. Moreover, the current resulting from excitation at a tends at the same time to re-establish the polarization at a and so to reverse the depolarization and restore the original condition of polarization at a. Each new point b, c, d, etc., excited by the current arising from a, becomes the source of a current which, as Lillie has shown, may produce excitation at further points and might also produce re-excitation at a if the refractory period following excitation did not prevent. It is evident that according to this conception the velocity of transmission depends upon the velocity of the local processes in the membrane at each excited point and is entirely independent of the velocity of elec- trical transmission \vhich is practically instantaneous. Moreover, in any case in wrhich excitation varies directly (not necessarily proportionally) with stimulus, a decre- ment must occur in the course of transmission. Only in certain highly specialized protoplasms for which the uall-or-none law'3 holds, i.e., in which any stimulus above the threshold produces maximal excitation, is transmission without a decrement possible.

The question might be raised whether the assump- tions concerning the semi-permeability and electric polarization of the plasma membrane and its depolari- zation in excitation are the Only possible basis for the electrical phenomena associated with excitation. It has been pointed out by Lillie and others that acid production and consequent increase in hydrogen ion concentration at points of excitation will not account for the electrical changes, since the currents produced in this way are opposed in direction to those which are actually observed. It is very probable, as Lillie sug- gests, that the membrane changes may be the primary factors in highly irritable tissues like nerve, where the processes of excitation and transmission and recovery are extremely rapid, but it is possible that in excitation in general the changes in the membrane may be incidental rather than primary. In an oxidation-reduction cell the electric currents produced show the same relation to the oxidation end of the cell as do the bioelectric cur- rents to the region of excitation. That is, the region where oxidation is going on in the oxidation-reduction cell is externally electronegative to the region where reduction is occurring. I have repeatedly called atten- tion to the apparent impossibility of dissociating the process of excitation in protoplasm in general from an acceleration of the fundamental dynamic activities of protoplasm, particularly those concerned with the liberation of energy. If current conceptions are correct, the chief source of energy in living protoplasm is oxida- tion. Moreover, it is apparently true that a region of rapid oxidation in protoplasm gives rise to electrical conditions similar to those which we observe in relation to the more highly specialized forms of excitation. While it is by no means necessary, nor at present justifi- able, to maintain that oxidation is the primary factor in excitation, it apparently is true, so far as electrical phenomena are concerned, that a region of more rapid oxidation in protoplasm behaves with respect to

regions of less rapid oxidation like a region of excitation with respect to less excited or unexcited regions, i.e., it becomes electronegative and a current similar to the excitation current is set up between it and other less rapidly oxidizing regions. If this current constitutes the exciting factor for adjoining regions and so the primary factor in transmission, then a region of rapid oxidation, however determined, may initiate a process of transmission which induces excitation, involving increase in the rate of oxidation at points within the range of effectiveness of the current, and so may give rise to an oxidation gradient. Doubtless even in such a case membrane changes are involved, but my point is merely that they need not necessarily be regarded as the most significant changes in the complex process. In some of his later papers Lillie calls attention to the probable significance of oxidation-reduction in the more primitive excitation-transmission relations.

In such a process as this the velocity of transmission may be very low as compared with nerve, so low indeed that it might in many cases scarcely be termed trans- mission in the ordinary sense. Moreover, in such a process a decrement is to be expected, since a relation unquestionably exists between the electrical condition and the rate of oxidation, and this being the case a gradient must result. In fact, in the extreme case such a gradient may arise without any wavelike transmission of excitation, such as occurs in the nerve, electrical transmission only being involved and determining a gradation in rate of oxidation and associated condi- tions from the original region of high rate to a certain limit of distance.

Moreover, if the original region of high rate persists the electrical conditions and so the gradient as a whole must persist and become static instead of advancing in the protoplasm. The length of such a gradient will depend, first, upon the potential arising at the original region of rapid oxidation and, second, upon the con- ductivity of the protoplasm. In that they represent a gradation from a region of high to one of low physio- logical activity, such gradients are comparable to the ascending phase or front of an excitation wave in a nerve fiber, viz., that portion of the wave from the region where excitation is just beginning to the region of maxi- mal excitation. But they differ from the ascending phase of the nervous impulse in that they do not advance, though they may undergo change in length with change in the rate of reactions in the original region or at other levels.

It seems probable that the most primitive excitation- transmission phenomena in protoplasm possess more nearly the characteristics or such oxidative or metabolic gradients than those of the highly specialized sorts of excitation and transmission processes, such as the nerve impulse. Apparently all gradations exist between the two extremes, and it is at least possible that the more rapidly arising and more rapidly reversible excitation- transmission processes, even in protoplasm in general, are chiefly or wholly plasma-membrane processes, while the slower, more persistent, and less readily revers- ible processes approach more closely the oxidation gradients and are not necessarily limited to cell mem- branes, but may occur in relation to limiting surfaces or phase boundaries more or less distant from the surface. But since protoplasm is a system in which a considerable

degree of correlation and mutual dependence exists among the essential conditions and reactions, change in any essential factor cannot proceed very far without involving changes in other factors. Undoubtedly, therefore, oxidation is not independent of the condition of membranes or phase boundaries and, on the other hand, the condition of these is not independent of oxidation. In other words the most primitive and the most highly specialized forms of excitation and trans- mission in protoplasms probably involve much the same complex of factors, though the specialization of different protoplasmic mechanisms may determine various modifications of the general process with different factors predominant in different mechanisms. It is possible, for example, that in nervous excitation and conduction oxidation may play a role of minor importance, and it is even conceivable that in certain protoplasmic mechan- isms some other reaction than oxidation may be involved. Again, in some mechanisms the plasma membrane and its changes may be of fundamental importance, while in others this factor may perhaps be secondary and inci- dental to others. Unquestionably the protoplasmic system affords a basis for the specialization of various mechanisms of excitation and transmission, but it does not follow that the primitive process is exactly like any one of the more highly specialized processes. The general problem has apparently been somewhat obscured by the fact that a large part of the work on excitation and transmission has been concerned with the process as it occurs in the muscle and nerve of the higher animals, and this work has constituted the chief basis for gen- eralization. It is commonly held, for example, that

excitation is transmitted as a wavelike change and that it is completely reversible. This is undoubtedly true in large measure for certain protoplasmic mechanisms, but it does not follow that it is true for all or under all con- ditions. In fact the existence and persistence of the metabolic and physiological gradients in protoplasms without specialized mechanisms indicates, if it does not prove, that neither wavelike transmission of the excita- tion process nor complete reversibility is a universal characteristic of the excitation-transmission relation. As I have pointed out, the excitation itself does not necessarily travel in these gradients, but the electro- motor changes determine different degrees of excitation at different distances from the point of original excita- tion. Moreover, in most protoplasms these gradients are not rapidly reversible but persist after the factors inducing them have ceased to act, and become factors in determining the developmental order and relation of parts. They are in a sense fixed or static excitation gradients, and the electrical changes rather than the excitation itself are transmitted. The various lines of evidence which have been considered in preceding chap- ters seem to me to indicate that the primitive excitation- transmission relation in protoplasm approaches more or less closely this condition and that oxidation is an important factor in it. With the specialization of protoplasm and the appearance of one mechanism or another changes occur. Perhaps the most significant of these changes is the change in the relation between the exciting factor and the excitation produced. In the more primitive mechanisms, as pointed out above, the energy, the intensity, or, in lack of definite knowledge

we may say, the degree of excitation, is certainly under some if not all conditions a function of the energy, the intensity, or the effectiveness of the exciting factor, and the same protoplasm is capable of different degrees of excitation. But with the specialization of mechanisms of excitation and transmission, the relation apparently becomes more and more like that of the spark and the powder, and such specialization reaches its final terms in the more highly differentiated muscle and nerve which apparently react according to the ;'all-or-none': law, i.e., maximal excitation results from any stimulus above the threshold. Under these conditions trans- mission of the excitation as a wave to an indefinite distance without decrement is possible, provided the maximal excitation at successive points of the path does not decrease, in other words, provided the path is not already a physiological gradient.

In conditions intermediate between the two extremes transmission may occur to a greater or less distance beyond the limit of the gradient resulting from the original excitation, but with a decrement. If, for example, the excitation induced at a point b (Fig. 6) by the electrical changes resulting from excitation at a gives rise to sufficient electric current to increase the strength of the current produced at a the limit of effect- iveness will be not x but a more distant point y. In this case then the excitation at b will become a factor in determining excitation at a point beyond the range of the current produced at a and the length of the excitation- transmission gradient will be increased. Similarly, exci- tation at c may be sufficient to determine excitation at z and so on. In such a case transmission is not limited by

the conditions arising at the point a of original excitation, but points 6, c, etc., become factors in extending the range. Unless the degree of excitation at the various points b, c, d, etc., within the range of the current pro- duced at a is the same and therefore independent of the strength of the current at these points, the excitation will undergo a decrement in the course of transmission and will finally reach a limit, and the steepness of the decrement will differ according to the relation between stimulus and excitation in the protoplasm concerned. The facts indicate that between the primitive gradient in which transmission depends entirely upon conditions at the point of original excitation and the condition of transmission without decrement various intermediate conditions exist, even in the same organism at different stages of development and perhaps in some cases with different degrees of excitation (see p. 225).

When we take into account all the facts it seems necessary to conclude, as I have already suggested, that excitation in a broad sense is essentially nothing else than an acceleration of the fundamental activities of life, and particularly or primarily those concerned with energy liberation. Excitability or irritability is a fundamental characteristic of living protoplasm and life is apparently made up of a series of excitations and their effects. The processes may differ as the mechanisms differ, but the fact remains that excitation means a rise in the level of living. The physiological or metabolic gradients which, as I have endeavored to show, constitute the funda- mental features of axiate pattern, are essentially gradients

in the rate of living, possess all the characteristics of excitation gradients, and are determined by differential exposure to the action of external factors. I believe we are amply justified in the light of all the facts in regarding them as the most general and primitive expres- sion of the excitation- transmission relation in protoplasm. The fact that these gradients are usually more or less persistent and not readily or rapidly reversible requires some consideration. From our knowledge of momen- tary excitations and the highly specialized processes in nerve and muscle we are accustomed to think of excita- tion as completely reversible. Excitation may be completely reversible in the fully developed nerve fiber, but certainly in many parts of the nervous system it is not completely reversible, as memory and the possi- bility of learning show clearly enough. Moreover, it is evidently not completely reversible in muscle, since frequent excitation, if not exhausting, leads to growth, i.e., functional hypertrophy, and after a certain stage of development excitation is necessary for the continued existence of the muscle. Whether we invoke a mys- terious " over-compensation " or some other hypothetical process to account for these facts, they show that while muscular contraction is completely reversible the pro- cess of excitation as a whole is not.

The fully developed organism represents an approach to dynamic equilibrium. The protoplasmic record of the events of development is about as complete as is possible for the protoplasm concerned. The various mechanisms are all developed and the further changes in them are largely temporary and reversible except so far as the gradual changes which constitute senescence are concerned. That is to say, the fully developed organism represents a condition in which most of the irreversible or less readily reversible changes possible in a particular protoplasm under the conditions of life and development have already occurred. It is, in fact, the occurrence of these changes which constitutes development. Under altered conditions reversal, or more properly as regards the process as a whole, regres- sion may occur, at least in the simpler organisms with less stable structural features, and the protoplasm may return to or approach an embryonic or unspecialized condition and undergo development anew (Child, 191 56). The earlier stages of development, however, do not and cannot be expected to show the high degree of reversibility of processes that we find in the fully developed organism. Embryonic protoplasm is proto- plasm from which the developmental record has been, so to speak, wholly or in large measure erased, and development is the recording process.

I have tried to show that the physiological gradient is the first step in the developmental record. If oxida- tion and synthesis are associated, as many authorities now believe, the higher rate of oxidation at the higher levels of the gradient must be a factor in determining a higher rate of synthesis, and more rapid growth may occur at these levels than elsewhere, as is often the case. Moreover, different molecules may be synthesized, or different products of synthesis may be relatively stable and remain in the protoplasm, at the different levels and so differentiation may begin. The chief point for present purposes is that if the conditions giving rise to the gradient persist long enough, whether this be hours,

days, or weeks, the gradient usually becomes to some extent more or less permanently recorded in the proto- plasmic substratum. Concerning the exact nature of this record we can only say that it consists primarily in the factors determining the rate of fundamental proto- plasmic activity, for we see that the differences in rate persist. The persistence of the gradient as a gradient must depend upon the fact that the changes in the protoplasmic substratum go on more rapidly at the higher levels than elsewhere. In short, I believe that the determination by an external factor of a more or less persistent gradient in physiological condition is the most general case of incomplete reversibility of excitation and also of so-called functional adaptation, that is, the process of the more or less persistent alteration of protoplasmic structure and function, by the dynamic activity occurring in it.

Concerning the existence of the axial gradients and their significance in development there can be no doubt. These are demonstrated facts, however we may inter- pret them, and they are the earliest indications of an order or pattern on a larger scale than that of proto- plasm. Organismic pattern is, as already noted, a pattern of regions or masses of protoplasm or cells and of a higher order of magnitude than protoplasmic pattern; physiological polarity and symmetry appear in protoplasms of very different specific constitution and many different lines of evidence agree in indicat- ing that they are primarily quantitative gradations in physiological condition, and that they can be

determined experimentally in many protoplasms by dif- ferential exposure of the cell or cell mass to the action of an external factor. In some protoplasms the gradients are only transient. In Amoeba, for example, the ectoplasm of the pseudopod is a gradient for the time being, but here the changes which give rise to the gradient are readily reversible and no permanent axiation develops (Hyman, 1917). In most protoplasms, however, the changes are less readily reversible, and a definite axiation results which may persist even through reproductive processes. The development of the axial gradient is merely one feature of development in axiate forms, and it represents the first step in the localization of physiological differences of organismic magnitude in the protoplasm concerned. The course of further development and differentiation is determined by the specific hereditary mechanism of each particular proto- plasm and the gradient is merely the framework, the ground plan on which the complex superstructure arises. We have seen (p. 33) that in the simpler forms and earlier developmental stages the axial gradient is very commonly found only in the superficial regions of the body. In most Protozoa and differentiated plant cells this superficial region is apparently denser and more permanent in position and structure than other regions of the cytoplasm, and if the gradient is an excitation gradient it may be expected to appear primarily or chiefly in relation to limiting surfaces, and its persistence is possible only in those regions in which the proto- plasmic substratum possesses a certain degree of sta- bility. But, admitting that excitation is transmitted along limiting surfaces, such surfaces are present, not

only at the cell surface, but may be more or less con- tinuous in the interior according to the phase relations. There is no reason why excitation should not be trans- mitted along these internal limiting surfaces, so far as they are continuous, and in this way penetrate the protoplasm to a greater or less depth. The internal limiting surfaces are doubtless not identical with the external, and conditions in them may be less favorable, but there are no grounds for believing that internal transmission is impossible. If it is possible, the axial gradients may exist not only at the surface but to a greater or less depth in protoplasm, according to con- ditions, and may in some cases involve the whole cell, as is apparently the case in some eggs. This suggestion also bears upon the question of cell pattern and surface- interior pattern in general (p. 60). I believe, in short, that the interpretation of axiate pattern in terms of excitation and transmission and the more or less per- sistent developmental protoplasmic changes which result from it is not only the simplest interpretation possible on the basis of our present knowledge but the interpretation which accounts most satisfactorily for all the known facts.

It is impossible to review the literature of the sub- ject here, but it is of interest to note that many authors have suggested that polarity is a gradient of some sort. Morgan (1904, 1906, 1907), for example, has at various times advanced the idea of polarity as a gradation or stratification of materials. Boveri (1910) maintained that the polarity of the Ascaris egg must be a gradation of some sort, and numerous authors have described gradations in size or form of parts, rate of development

or regeneration, etc., along the chief body axis. In fact, evidence of one kind or another for the existence of gradients of some sort as characteristic features of axiate organisms appears in the literature of many different and widely separated fields of investigation, both physiological and morphological. This evidence is perhaps the more valuable because the authors were in most cases stating or describing facts rather than attempting to formulate theories. In the conception of the excitation gradient and its fixation in protoplasm I have merely attempted to provide a general physio - logical basis for the facts as they stand.

It is evident that a simple excitation- transmission relation constitutes, at least momentarily, a physiological integration, in which the point of primary excitation becomes the dominant or controlling factor, since it initiates the transmitted change and determines what shall occur at the various points along the path of trans- mission. Moreover, this relation is the physiological basis of nervous function and integration, and this in turn represents the highest degree of physiological integration which has been attained in the course of evolution. It was pointed out above (p. 17) that the excitation- transmission relation is the most general and most primitive relation of organismic as distinguished from purely protoplasmic character, and I have endeavored to show that the axial gradients originate as excitation gra- dients. From this viewpoint the excitation-transmission relation appears as the basis both of the beginnings and of the most complex and highly developed features of organismic integration, viz., nervous integration. Such a conception has at least the advantage of placing the problem of organismic pattern and integration before us in definite and consistent physiological terms.

The role of the physiological gradient as an integrat- ing factor follows necessarily from the fact that it is a gradient. Once established it brings different regions within its range into a definite physiological relation. The differences in electric potential which have been found to exist between different levels afford one means of actual correlation, and, if Lillie's theory is correct, the resulting electric currents must be the primary factors in determining and maintaining the conditions at different levels. Conceived in these terms, organismic integration is primarily electrical, that is to say, the factor actually concerned in the first instance in bringing different proto- plasmic regions into physiological relation with each other is the electric current. But this is only the first step in organismic integration. As soon as differences in protoplasmic constitution appear at different levels of a gradient, chemical or transportative correlation begins to play a part in integration, and still later various mechanisms of correlation and integration develop, e.g., the various types of vascular system, the supporting and other mechanical tissues, the organs of secretion and excretion, and in most animals the nervous system as the organ of integration par excellence.

Granting that the physiological gradient is the basis of axiate pattern, it is evident that it must constitute the starting-point of the orderly, progressive complica- tion in differentiation and physiological correlation which we call development. In some way qualitative differences must arise in the protoplasm at different levels of the gradient which was essentially quantitative in its origin. The origin of differentiation from differ- ences which are primarily quantitative may appear to