The Origin and Development of the Nervous System

some to constitute a formidable difficulty for this con- ception, but I believe the difficulty is only apparent. It is of course true that differences in the velocity alone of a pure chemical reaction do not give rise to qualitatively different products of reaction. In protoplasm, however, we are concerned, not with a pure chemical reaction, but with an exceedingly complex system, involving both physical and chemical factors. The coefficients of quantitative alteration of the different factors of the protoplasmic system in relation to quantitative external factors, such, for example, as temperature, are of very different magnitudes. The temperature coefficients of most chemical reactions differ widely in magnitude from those of many physical processes, and there is even considerable difference in the temperature coefficients of different chemical reactions. A change in tempera- ture then does not alter the different component factors of the protoplasmic system to the same degree, but alters some much more than others. Consequently such a change, which is nonspecific and quantitative for each factor, may alter the system as a whole in such a way that chemical reactions which could not take place before the change may occur after it, and vice versa. Changes in concentration of reacting substances do in many cases determine different products of reaction, and change in temperature may bring about changes in concentration of substances in protoplasm in various ways, e.g., by alteration in the water content, the dis- persion of the colloids, adsorption, and by change in the rate of chemical reaction itself, which may change the concentration of the products of a given reaction and so determine a second reaction which could not take place

previously, and therefore a new product of reaction. Such changes determine others, and there is not the slightest doubt that a change in protoplasmic condition, originating as a purely quantitative alteration in some factor or factors, may become the starting-point of extensive qualitative changes. When we take into account the material exchanges between protoplasm and its environment, various other possibilities of change in concentration of substances in the protoplasm appear. An increase in rate of oxidation in the protoplasm may oxidize certain substances more rapidly than they enter the protoplasm, and, as they disappear, other substances, previously inactive, may enter reaction and different products be formed. As a matter of fact, even with our present fragmentary knowledge of living protoplasm, almost endless possi- bilities appear for the origin of qualitative differences from primarily quantitative changes in the system. It may even be questioned whether it is possible to bring about a change in protoplasm which remains through- out purely quantitative.

Differentiation in the organism consists, at least very largely, in the appearance in protoplasm of substances not present originally in appreciable quantities, and such substances differ in amount or in kind in different regions or cells. In many such cases a simple difference in relation between the rate of nutritive intake and the rate of oxidation is sufficient to account for the appear- ance of a particular substance in a certain region or at a certain time, or its non-appearance or disappearance in another region or at another time. The accumulation and disappearance of fat and various other so-called

reserves, the loading and discharge of gland cells, are cases in point, and there can be no doubt that many other differentiations arise in essentially the same way. That is, a region in which oxidation is occurring slowly becomes different from a region of rapid oxidation because of this difference in relation between intake and rate of oxidation, which determines that in the one region certain substances shall accumulate as part of the structure, while in the other such substances are more or less completely oxidized as rapidly as they are formed.

The changes concerned in excitation, whether they are primarily electrical, chemical, or physical, may similarly become the starting-point of a series of quali- tative changes, particularly if they persist for a con- siderable time, and it is evident that the different levels of an excitation gradient or a more or less permanent physiological gradient afford numerous possibilities for the origin of qualitative differences at the various levels. Such changes may or may not be readily or completely reversible. The changes which constitute differentia- tion are those which are not readily or completely reversible under the conditions of development.

As soon as these qualitative differences make their appearance the basis for definite and orderly chemical or transportative correlation of organismic magnitude is established, and from this time on such correlation plays an increasingly important role in the further sequence of events. The definiteness, variety, and specificity of chemical correlation must increase as differentiation progresses and mechanisms of communication develop, both in the individual and in the course of evolution,

until in the higher animals and man it becomes almost inconceivably complex. Moreover, from the moment of its appearance as a factor in organismic pattern, chemical correlation constitutes a factor in organismic integration, which develops along different lines from nervous integration and makes possible other kinds of control and relation than those determined through the nervous system. The internal secretions and all so-called hormones of the higher animals and man con- stitute the most advanced stages attained in this development.

Differentiation appears, not only in the development of the mature organism from the egg, but in many animal eggs the appearance of the yolk in the protoplasm con- stitutes a characteristic differentiation. The relation of this differentiation to the physiological gradient in the egg is of interest. Very commonly a gradient in yolk accumulation appears, the apical (" animal") pole of the egg showing little or no yolk and the amount increasing toward the basal (" vegetative ") pole. Yolk is sub- stance which can under certain conditions be oxidized and furnish energy, but it is evidently not oxidized to any great extent during the growth period of the egg in the region where it accumulates, and it has been sug- gested that a reduction is involved in yolk formation. Evidently the different regions of the egg are different in some way, and the gradient in yolk accumulation suggests that the difference is essentially quantitative, not qualitative. Moreover, we find in such an egg a quantitative gradient in physiological activity, and when development begins the region containing least yolk shows the greatest developmental and apparently

the greatest oxidative activity, and is certainly doing the most work. Evidently the localization of the yolk in the egg is determined by the axial gradient and con- stitutes the first differentiation in relation to that gradient. The facts indicate very clearly that the yolk accumulates in the region of least oxidation, even though the substances from which it is formed enter at the opposite pole, as they do in many cases (Fig. 5). Such differentiation along the axis of the egg becomes the basis of the two primary germ layers, the ectoderm arising from the more active, the entoderm from the less active, region. The region of origin of mesoderm differs in different forms, but is often intermediate between ectoderm and entoderm. The gradient hypothesis maintains that differences in behavior of the different egg regions and of the germ layers arising from them result from the fact that they represent different levels of the gradient.

The gradient which determines the differentiation of the oogonial cytoplasm into yolk-bearing and yolkless regions is merely the beginning of a more or less complex series of events, each of which is determined in a definite and orderly way by the preceding. It will be evident at once that as soon as any such differentiation along a gradient occurs chemical or transportative correlation becomes possible, and it is certain that such correlation plays an increasingly important and complex role as development progresses. The accumulation of yolk in one region of the egg becomes itself a basis for trans- portative correlation, for most of this yolk is finally used, not in the region where it accumulated, but in the region which originally contained little or no yolk.

There are of course many cases in which the gradient does not determine in the egg any such gross structural differentiation as the yolk gradient, but such differences are incidental rather than fundamental, and depend upon differences in the specific constitution of the proto- plasm of different eggs, differences in nutritive conditions, etc. But sooner or later the structural differences appear, always in definite and orderly relations to the gradients, and it has been shown that the localization and relative size and proportions of parts along an axis can be altered experimentally by altering the length, steepness, or direction of the gradient.1 I believe that the evidence at hand warrants the statement that it has been experimentally demonstrated that such a gradient in a specific protoplasm constitutes an adequate basis for the initiation and so for the orderly progressive com- plication, with reference to the particular axis which the gradient represents, of the differentiation characteristic of the process of development in that particular proto- plasm. Once again it may be pointed out that the gradi- ent does not determine the character of differentiation and the kind of organs produced at different levels in different organisms. These are determined by the specific inherited constitution of the protoplasm.

Where symmetry gradients as well as polar gradients are present the situation is of course further compli- cated and orderly differences arise in more than one direction. Similarly, as particular organs are localized and develop, new organ gradients or axes may arise at any angle to the original axis in consequence of local conditions, until in the higher animals the axial relations may become almost inconceivably complex. In the mammalian nervous system, for example, the polarities of the different neurons form all possible angles with the primary body axes, yet it seems evident that these polarities arise in response to local conditions.

Turning to another aspect of physiological integra- tion, it is evident that a relation of dominance and subordination is a characteristic feature of the axial gradient. The region of most intense activity, the high end of the gradient, must influence the physiological condition at other levels to a greater degree than these other levels can influence it. It is the region of highest electronegativity and the primary factor in determining the electrical gradient, and therefore must dominate or control other levels to a greater or less extent. When a new excitation occurs at any level of an already established physiological gradient it must, if the electric current is the transmitting factor, as Lillie maintains, be transmitted to a greater distance down the gradient than up, because as it passes to higher levels its own electric current will be more and more completely com- pensated by the current arising at the higher levels. When an axial gradient is already present we may expect to find, at least in the simpler organisms, a difference of this kind in transmission with respect to the two direc- tions, upward and downward. Figure 7 is an attempt to show in a purely diagrammatic way the reason for the disappearance within a comparatively short distance of an excitation transmitted up a physiological gradient.

In this figure xy is the gradient in electronegativity characteristic of the physiological gradient. Let us suppose an excitation to occur at a with an increase in electronegativity ab. It is evident that the current produced by excitation at a will be compensated at d by the current of the physiological gradient, but if points on the gradient above a increase the strength of FIG. 7. — Diagram illustrating transmission up a gradient; for explana- tion see text.

the excitation current the excitation may be trans- mitted to a higher level e and so on. An excitation at the level a which increases the electronegativity to c will produce electrical effects at all levels up to x and if the excitation brought about by this current at levels above a increases this electro- negativity still further, the excitation ac may be trans- mitted all the way to the high end of the gradient and may produce an increase in electronegativity even there. These relations may of course be altered by various factors: for example, higher levels of the physiological gradient may and probably often do have a lower threshold or a higher maximum of excitation than lower levels, consequently the decrement may be different at different levels. In tissues which react according to the uall-or-none" law transmission may occur to an indefinite distance up a gradient, and if the maximal excitation is higher at higher levels the energy or inten- sity of the excitation may perhaps increase as it is trans- mitted upward.

Differences in transmission up and down an axial gradient appear in many functional relations. It has been noted by various observers that when flatworms or annelids are cut in two, at least in the more anterior body regions, the portion anterior to the cut is much less affected than the posterior portion (Norman, 1900). Coelenterate tentacles show distinct physiological gra- dients, and conductivity down the gradient, that is, basipetally in the tentacle, is much better than in the opposite direction. Various other similar cases might be mentioned.

This relation of dominance and subordination, how- ever, is not simply a functional relation in fully developed organisms, but is concerned in development and is quite independent of the presence of a nervous system. The existence of such a relation in plants with the growing tip or, in certain cases, some other region of high activity as the dominant region has long been known. This relation in the higher plants has been interpreted by Sachs, Loeb, and others in terms of

transportation of formative or inhibiting substances, but even in these plants it can be blocked by a zone of low temperature about the stem between dominant and subordinate parts, which merely decreases the metabolic activity of the cooled region but does not stop trans- portation (Child and Bellamy, 1919). Moreover, the relation of dominance and subordination exists in algae in which a transportative mechanism is not, so far as is known, developed.

A similar relation exists in animals, the apical or anterior region being primarily the dominant region in the polar axis, and it has been possible to show experi- mentally that this relation plays an important part in the localization, proportions, and relations of parts.1 In Planaria, for example, a piece taken from any level of the body is capable of giving rise to all parts posterior to its own level, even if it does not give rise to a head, but it never produces any part anterior to its own level unless some degree of head development occurs first (Figs. 8-12). Moreover, the distance from the head at which certain organs arise in regeneration can be altered experimentally by altering the rate of physiological activity in the regenerating head region (Child, 191 ic). In short, any given level of the body dominates more posterior levels to a certain degree, and is dominated by more anterior levels.

Although the relation of physiological dominance and subordination originates in the physiological gradient, 1 In 1907 I called attention to the existence of such a relation in the hydroid, Tubularia (Child, 19070, c, 19126), and Rand's work on Hydra led him to recognize its existence in that form (Rand, 1911, 1912). Since then further data bearing on the question of dominance and its relation to physiological condition have accumulated (Child, 191 ic, 1914^, e, 1915^, chap, iv, 19166).

FIGS. 8-12. — Reorganization in pieces from different body levels, a and b, in relation to presence or absence of a head; Planar la doroio- ccphala: Fig. 8, diagrammatic outline, indicating levels from which pieces are taken; Fig. 9, normal animal developing from either a or b when head arises; Fig. 10, anophthalmic form with rudimentary head, but developing all parts of body, whether taken from a or b; Fig. n, headless form from region a, develops all parts posterior, but none anterior to region a, from which it was taken; Fig. 12, headless form from region b, develops all parts posterior, but none anterior to its own level (from Child, 191 5c).

it is not necessary, for the persistence of the relation, that the original gradient shall persist unchanged. After definite anatomical relations of parts and definite conduction paths have developed, the dominant region may still maintain its dominance, even though the metabolism or other physiological change in it is less in amount than in the parts which it dominates. The initiation of a nervous impulse in a receptor of the head and its conduction through the nervous system to an effector certainly does not involve a very large amount of metabolic, electrical, or other physiological change, but by means of such impulses the nervous system maintains its dominance, even though the physiological change brought about in the effector by the impulses may be far greater in amount than that in the nervous system. The physiological gradient represents merely the first step in the determination and establishment of the relation of dominance and subordination.

Since this relation of dominance and subordination is a feature of the physiological gradient and since the length of the gradient is limited, the range within which such dominance is effective must be limited, though variable, according to physiological condition. In terms of Lillie's hypothesis the range of dominance must be determined primarily by the range of effectiveness of the electrical gradient which is a feature of the physio- logical gradient. This limitation of the range of physiological dominance makes possible the occurrence of physiological isolation, i.e., of the more or less com- plete isolation of a part of the body of the organism

from the controlling or dominating action of the high end of the gradient. Such physiological isolation is a familiar fact in plants and I have shown that it also occurs in animals and that it determines at least many processes of agamic reproduction.1 Such physiological isolation may occur in four ways: first, by increase in the size of the body beyond the limit of the axial gradi- ent, the region most distant from the high end of the gradient may be isolated; second, decrease in the activity of the dominant region may decrease the length of the gradient and so determine isolation of the most distant regions without actual increase in size of the body; third, transmission between the dominant region and another part may be blocked by one means or another and parts beyond the block may thus be isolated; fourth, a subordinate part may be directly excited to such a degree that its own activity and electric potential equal that of the dominant part and the current produced by it compensates the current produced by the dominant region, consequently the part concerned is physiologically isolated.

The result of physiological isolation of a part in the simpler organisms is the same as the result of physical isolation. In all forms in which dedifferentiation 1 The literature of both botany and zoology contains many data on physiological isolation of parts, but they have been described under various other terms. The outgrowth in plants, in consequence of inhibition of the growing tip, of buds previously inhibited by the tip is a familiar example of physiological isolation. Similar processes occur in the hydroids and other colonial animals. The occurrence of fission and budding and various other forms of agamic reproduction in many different animals also results from physiological isolation. See Child (igoyc, 19100, 1911^, e, 1915^, chap, v, 1917^, d), Child and Bellamy (1919), Bellamy (1919), Hyman (1916).

occurs readily the physiologically isolated part loses to a greater or less extent its differentiation as a part and returns to or approaches an embryonic condition, and if it retains the original axial gradients, or new gradients arise in it in consequence of local conditions, it develops, like any other embryonic protoplasm of the species, into a new individual. In this way the occurrence of budding, fission, and the spatial arrangements of differ- ent axes in multiaxial plants and colonial animals, such as the hydroids, are determined (Child, 19196). More- over, the evidence indicates that repetition of parts in series, e.g., segmentation (Child, 191 yd; chap, viii, below), and at least many other repetitive develop- mental phenomena are essentially similar in character. In short, any active region, whether it be the apical end of a whole organism or a special organ, such as a hydroid tentacle, dominates to some extent a certain region about it so that only outside of this region and beyond the range of this dominance is a repetition of the indi- vidual part or organ possible.

According to this conception, then, the gradients established in protoplasm may determine not only the course of differentiation along an axis but in the simpler organisms the limit of size of the individual, the occur- rence and order of various agamic reproductive processes, and the orderly repetitive arrangement of parts. There is much evidence to indicate that the range of dominance increases greatly during individual development, par- ticularly in animals in which a nervous system with a much higher conductivity than protoplasm in general develops. Physiological isolation of parts may occur in embryonic stages when the organism is very small, but

in later stages, after the nervous system develops, the parts may be reintegrated into a new whole, as is appar- ently the case with annelid segments (Child, 191 7 d). It is probable, however, that in the higher animals the size of the adult is determined rather by the cessation of growth with increasing differentiation than by the range of dominance, for the conductivity of the highly specialized nerve fiber is so high that the range of domi- nance is certainly much greater than the length of any animal body, and if medullated nerves actually do con- duct without decrement, it is indefinite. But even if physiological isolation of parts of the body should occur in the more advanced developmental stages of the higher animals, it would doubtless not result in agamic reproduction of new individuals, because the capacity of the cells for dedifferentiation is limited. Certain aspects of the tumor problem, particularly with respect to the embryomas and the malignant tumor forms with indefinite growth, are of interest in this connection but cannot be considered at this time.

This brief survey of some of the more important features of physiological integration and isolation will serve to show in a general way how a quantitative physiological gradient involving differences in the rate of fundamental reactions and in protoplasmic conditions associated with them may, on the one hand, initiate localization and differentiation of parts and organs and on the other constitute the basis of a relation of domi- nance and subordination within the individual. Such a gradient, however, is merely the initiator and ordering

factor in the series of events which constitute morpho- logical and physiological development. From the first moment when qualitative differences arise at different levels of the gradients, chemical or transportation corre- lation becomes possible and from this time on it becomes increasingly complex and important as a factor in determining the course of events. The earlier trans- missive relations between certain parts may later be supplemented or perhaps even more or less completely replaced in some cases by transportative relations. Whether or not the gradient persists as such throughout the life of the individual is a question of detail to be determined in each particular case. As a matter of fact, in the simpler organisms it does often persist, at least in the superficial regions and body wall, but it does not follow from its presence in these parts that the internal organs will show the same relations. The changes which occur between the comparatively simple physiological gradient or gradients in the egg and the situation in the adult must be determined, like any other aspect of development, by investigation in each particular case. In these changes the hereditary mech- anisms of the different protoplasms are concerned. Even if the gradients in two different protoplasms were exactly similar in slope and quantitative relation the course and results of development would undoubtedly differ in the two cases.