The Origin and Development of the Nervous System

In the simpler animals we see new polarities arise at cut ends of pieces, e.g., the development of new apical regions and axes from the aboral ends of pieces in various hydroids, the development of heads from posterior cut surfaces in Planaria (see Figs. 30-34, p. 125). In such cases the new axis is always represented by a new gradient and the relation between the new axis and the occurrence of differential exposure is obvious, though whether the wound stimulus, oxygen supply, or some other factor is chiefly concerned in determining the new gradient is not known. Loeb (1892) has maintained that new polarities are determined by gravity in the

regulatory development of the hydroid Antennularia antennina, but Morgan (1901) and Stevens (1902, 1910), while not disputing Loeb's results, showed that other factors besides gravity were concerned in determining polarity. In the development of sponges from dissociated tissue cells described by H. V. Wilson (1907, 1911) the polarity of the new individual is determined by some sort of differential between free and attached surfaces of the cell mass, the osculum developing on the free surface. Similarly in the experiments on obliteration of pre- existing polarity and the establishment of a new polar axis in hydroids (Child, 1915^, pp. 142-46) the apical region of the new axis arises in the region of greatest exposure to the environment, but the particular factors chiefly concerned have not been determined. In both these cases, the sponge and the hydroid, the difference in oxygen supply on the exposed surface of the cell mass and the surface in contact suggests a probable factor in determining the new axis.

In the actinian Harenactis the localization of a region of more rapid and more extensive growth by injury is sufficient under certain conditions to deter- mine the position and development of new polarities (Child, iqiob, 191 5c, Figs. 79-83). The axial grada- tion results in this case from the fact that the activity of the cells is greatest in the middle region of such an area and decreases*toward its borders. In this connec- tion a recent statement of Harrison's concerning the limb rudiment of Amblystoma is of interest. Harrison says: :The limb rudiment may be thus regarded, not as a definite circumscribed area like a stone in a mosaic,

but as a center of differentiation in which the intensity of the process diminishes as the distance from the center increases, until it passes away into an indifferent region. Many other systems, such as the nose, ear, hypophysis, gills, seem to have the same indefinite boundaries which may even overlap one another" (Harrison, 1918, p. 456). Evidently Harrison conceives these primordia as gradi- ents in activity in a more or less specialized cellular region of the embryo. Such gradients differ from the general axial gradients of the body only in that they are determined in some way, presumably by intra- organismic correlative conditions in specialized body regions, and are concerned with particular organ com- plexes instead of with the body as a whole. In still other cases new polarities are apparently determined and localized by slight differences in activity between different cells of a mass. Such differences determine the more or less definite localization of a region of growth in which the activity decreases toward the periphery, and as growth progresses an axial gradient arises. Determi- nation of new polarities in this way apparently occurs in many cases when pieces of naked hydroid stems give rise to multiple stolons, each of which represents a new axis and a new gradient. These multiple polarities have been observed by many investigators, and I have been able to produce them experimentally in hydrozoan planulae by first obliterating the original polarity through differential inhibition. In the origin of adventitious buds from the epidermal cells of the Begonia leaf similar local growth areas with gradients in activity from center toward periphery and from the surface inward are the first indications of the new plant axes (Regel, 1876;

Child, 191 5c, Figs. 38, 39). The localization of the new axis in such cases appears to be largely a matter of slight fortuitous differences in activity in different cells or cell groups, in consequence of which certain cells or groups react more rapidly than others to the experi- mental conditions. As regards the animal egg, the evidence is very incomplete but indicates that in at least many forms polarity is determined during the growth period of each egg by differential exposure. In various hydromedusae, for example, the growing oogonia constitute a columnar epithelium, one end of each cell being separated from the exterior only by a layer of very thin flattened cells, while the opposite end is attached and adjoins the radial canal (Figs. 1-2). When portions of the ovaries are slightly teased to separate the eggs it is found that the free end of the egg, the end nearest the exterior, repre- sents the high region, and the attached end, the low region of a gradient in susceptibility and permanganate reduc- tion (Fig. 3). A similar gradient appears in develop- mental stages, and while the absence of good landmarks makes it impossible to demonstrate that the later gradient is identical with the earlier, there can be little doubt that it is. In the sea urchin the oogonium is attached to the wall of the ovary at one small region of its circumference and here also, as Boveri has shown, the free pole becomes the apical pole and represents the high end of a gradient. In Sternaspis an aberrant annelid, a peduncle containing a vascular loop, develops in connection with each growing oogonium, but my observations indicate that the free pole, not the attached pole of the egg, into which the vascular loop enters,

becomes the apical pole (Fig. 4) . In these cases it is not the pole through which nutrition enters, but the un- attached pole, which in the medusa is more exposed to external factors and in the sea urchin and Sternapis to the fluids of the ovary which becomes the apical pole. It seems probable that a differential in oxygen supply FIGS. 1-3. — Ovarian eggs of jellyfish, Phialidium gregarium; diagrammatic: Fig. i, early growth stages; Fig. 2, later stages; in these stages one surface (the lower in the figure) is separated from the exterior only by thin epithelium and opposite surface is attached and adjoins radial canal; Fig. 3, eggs separated from ovary by teasing, showing region of attachment; shading indicates susceptibility gradient as indi- cated by disintegration and by permanganate.

and perhaps also in CO2 concentration are chiefly con- cerned in determining the polarity and the gradient which represents it in these cases. In the eggs of at least some higher animals, where oxy- gen as well as nutritive substances reach the egg chiefly or wholly through the blood, the polarity is apparently determined by relation to the blood supply. Bellamy (1919) has shown that in the frog's egg polarity appar- ently develops in definite relation to the vascular

supply, the apical pigmented pole arising on the arterial, the basal unpigmented pole on the venous side (Fig. 5). In these cases the oxygen supply or the conditions determining rate of respiratory exchange are apparently FIG. 5. — Semi-diagrammatic view of ovarian egg of frog, showing distribution of arteries and veins in relation to pigmentation and polar- ity: a, artery, v, vein, p, pedicle (from Bellamy, 1919). the chief factor in determining polarity, the region of most rapid exchange becoming the apical pole.

Symmetry in animals, like polarity, is primarily indicated by gradations in physiological condition, and although a particular kind of protoplasm may give rise normally to a radial or a bilateral animal, experiment shows that for certain forms, at least, the normal sym- metry of pattern is not inherent and unchangeable. In the radial anemone Harenactis, for example, bilateral tentacle groupings may arise under certain experi- mental conditions (Child, 1909), and it is possible through differential susceptibility to obliterate bilater- ality and produce radial larval forms in the sea urchin (Child, 1916^) and also in the starfish (unpublished). Moreover, in pieces of Planaria, under experimental conditions which practically obliterate the polar gradi- ent the symmetry gradient may become the polar axis of the new individual (Child, 1915^ pp. 163-65).

A characteristic feature of radial symmetry is the repetition about a center of parts, each of which is usually axiate and often bilateral in pattern. Such repetition must be largely a matter of the space rela- tions of specialized growth centers. Each growth cen- ter involves or dominates a certain area, and only at a certain distance from it can another similar center arise. Thus the number of such growth centers arising on a given circumference depends on the area dominated by each center, which varies with physiological condition, and on the size of the circumference, and can be altered experimentally in many cases. Moreover, as growth in size of the circumference occurs and the distance between the repetitive parts increases, we often find additional new parts arising, e.g., mesenteries and tentacles in many actinians, etc. Such processes are physiologically similar to many forms of agamic repro- duction, being essentially reproductions of specialized parts instead of new wholes, resulting from physiological isolation (Child, 1915^, chaps, iv, v).

Bilaterality is inherited through many processes of agamic reproduction, e.g., in flatworms and annelids, and may of course also be inherited in many eggs. On the other hand, it may conceivably be determined in some eggs by ovarian conditions, by conditions connected with maturation or with fertilization or perhaps even by conditions arising later in development. For example, the median ventral trunk region in the polychaete annelids arises from the fusion or concrescence of the growing borders of the somatic plate (Child, • 191 yd), and in the vertebrates from the concrescence of the growing border of the blastoderm or the lips of the blastopore. The higher frequency of cell division and the smaller size of cells in the lateral border of the somatic plate of various annelids suggest that it is a region of high physiological activity, and in certain cases I have found that it is more susceptible to lethal concentrations than other parts of the somatic plate. Hyman has observed that in certain of the fishes the border of the blastoderm becomes a region of high susceptibility as it begins to form the germ ring and to grow down over the yolk (unpublished) . As the embryo is formed, this active border undergoes concrescence to form the median dorsal region. In the frog (Bellamy, 1919) after gastrulation the lips of the blastopore, which are overgrowing the yolk, constitute a region of high susceptibility to lethal concentrations, most marked in the dorsal region. These lips form the median dorsal region of the embryonic body. These facts suggest that the mediolateral gradients in the annelids and in at least the lower vertebrates may be associated to some extent with the differences in physiological con-

dition between the more exposed and probably more stimulated growing borders of the somatic plate, the blastoderm or the blastopore lips and other adjoin- ing regions. Moreover, many bilateral forms develop characteristic asymmetries during the course of devel- opment, e.g., the asymmetry of gasteropod mollusks and the visceral asymmetry of vertebrates, or a well- developed bilateral symmetry may give rise in metamor- phosis to radial-bilateral pattern of very different kind, as in certain echinoderm groups. Our knowledge con- cerning the physiological aspects of the origin of sym- metry in animals is still very fragmentary, but the earliest indications of the presence of a particular sym- metry pattern are gradients in physiological condition, which, so far as the evidence goes, are similar to the polar gradients, and in plants we see the different symmetry patterns arising through differential exposure to the action of external factors. In the light of all the facts we are justified in concluding that even though a particular symmetry pattern may persist through repro- duction, i.e., be inherited in a particular case, sym- metry like polarity must in the final analysis arise through differential exposure to the action of external factors.

Even if we grant that the differential exposure of the egg of the medusa, the sea urchin, Sternaspis, etc. (see p. 54), determines the polarity and that symmetry may also be determined by relation to environment, it is evident that, except in some of the simpler organisms, the differential exposure of the egg is not fortuitous, but is determined by the hereditary mechanism of the organism. The epithelial arrangement of eggs in the medusa gonad, the position of the sea urchin egg in the ovary, the development of the peduncle and the vascular loop in SternaspiSj and the circulatory pattern in the chorion of the frog's egg are all features of the hereditary mechanism of the organism concerned. Even in such cases, then, as well as in cases where the axiate pattern persists through reproduction, the hereditary mechanism is concerned in the origin of the axiate pattern of the new individual. Moreover, even in the case of Fucus, where the determination of polarity by the direction of incident light is apparently wholly fortuitous, the hereditary mechanism, as expressed in the constitution and pattern of the egg protoplasm, determines the occurrence and the nature of the reaction to incident light. In other words, the hereditary constitution of the Fucus egg, as well as the direction of incident light is a factor in the determination of the polarity. There is, in short, no conflict between this physiological con- ception of the origin of organismic or of axiate pattern and conceptions of heredity. The origin and develop- ment of organismic pattern in nature is simply the realization of certain hereditary potentialities of a particular protoplasm in a particular environmental complex, which may itself be determined in large measure, or wholly, by the hereditary mechanism of the protoplasm concerned.

In the light of the conclusions reached concerning the nature and origin of axiate pattern, the question of cell pattern, touched upon above (p. 23) requires some further consideration. It was suggested that the cell is primarily a surface-interior pattern resulting from exposure of the surface of a mass of protoplasm to the action of external factors. Such an exposure is a differential exposure as regards surface and interior. Both the respiratory exchange and excitation can occur only through the surface; therefore differences must arise between surface and interior, and a more or less definite gradient in such conditions from the surface inward must result. As different organs are localized at different levels of an axial gradient, so the localization and differentiation of the nucleus in the first instance may have resulted from the conditions in the interior of the protoplasmic mass. In fact it is difficult to see how the nucleus as a definite organ could have arisen otherwise. The differences between nucleus and cyto- plasm as regards acidity and electric potential, as well as the behavior of nuclei in such specialized cells as spermatozoa, w^here cytoplasm is practically absent, all suggest that the nucleus is fundamentally an internal cell organ, and if the origin of cell pattern has any relation to environmental factors, the differentiation of the nucleus must have been determined originally by conditions in the interior of a protoplasmic mass. The fact that the nucleus persists from one cell generation to another means merely that the pattern once estab- lished is persistent or inherited, although it is difficult to determine to what extent the persistence of the surface-interior conditions is concerned in the per- sistence of pattern.

Viewed from this standpoint, cell pattern originates in the differential between surface and interior in general and axiate pattern in differentials between different parts of the surface of the protoplasmic or cell mass concerned. As regards the axiate pattern, the evidence indicates that the differential is primarily quantitative and involves differences in the rate or degree of funda- mental protoplasmic activity, but as regards cell pattern we have at present no means of determining whether the differential was primarily quantitative, though vari- ous lines of evidence point in that direction.

The presence of an axiate pattern does not neces- sarily imply the disappearance of a general surface- interior pattern, either in the cell or the multicellular organism. All organisms show some kind of surface- interior pattern, at least in the superficial regions of the body, and all the facts indicate that in the final analysis such pattern arises through exposure of the surface. The passage of cells to the interior of the embryo by gastrulation is of course a feature of axiate pattern, but conditions in the interior are undoubtedly factors in determining the further differentiation of such cells into the organs of entoderm and mesoderm. Only in some of the simpler animals does the general surface-interior differentiation arise in situ. Forma- tion of entoderm by delamination in all cells of a blastula, for example, appears to be a case in point, and in the Protozoa definite morphological differentiation occurs primarily and chiefly in the ectoplasm. But in the ectoderm of multicellular animals we find numerous evidences of surface-interior pattern ranging from the basal muscular extensions of ectoderm cells in Hydra

(Fig. 68, p. 238) to various complicated sensory struc- tures and the early differentiation of the neural tube in the higher animals, and the relation of such features of pattern to exposure of the surface appears obvious. Here as in the case of the axiate pattern, however, the surface-interior relation represents merely the physio- logical conditions under which. the potentialities of the hereditary mechanism of the protoplasm are realized. The fact that the embryo possesses a surface determines certain relations in the protoplasm to this surface and the specific constitution of the protoplasm determines the kind of reaction which occurs, the sort of specializa- tion which develops. If an axiate pattern is also pres- ent, physiological conditions are provided for differences in reaction in relation to the surface at different regions or levels of the body. This of course does not mean that exposure to a special external stimulus is necessary for the development of a particular superficial organ. Light, for example, is not necessary for the differentiation in the individual of a photoreceptor or an eye. The surface-interior relation merely determines that the physiological conditions under which the hereditary potentialities of a given protoplasm to produce a photo- receptor are realized, are conditions which arise in development at or near the surface of the developing organism.

The differential exposure of the cell or cell mass to the action of environmental factors is only the first step in the establishment of the gradient. Admitting that this differential exposure determines a higher rate of activity in some region, we may expect, since living protoplasm is irritable and since increased activity in one region serves to some extent to excite adjoining regions, that transmission from the region of increased activity will occur. In the absence of highly specialized conducting paths, we find that protoplasmic excitation apparently undergoes a decrement in intensity or effectiveness with increasing distance from the point of origin, so that an excitation gradient results. Many facts, some of which are considered in the following chapter, indicate that the physiological relation resulting from the differential exposure of a cell or cell mass is primarily a relation of excitation and transmission and that the resulting gradient is essentially the fixation or establishment of an excitation-transmission gradient in the protoplasm through the modification of the proto- plasmic subtratum by the persistence or repetition of the differential exposure. From this viewpoint the gradient represents the most primitive sort of excitation- transmission relation and its effects upon protoplasm.

It is perhaps necessary to point out that the con- ception of axiate pattern as primarily a quantitative gradation in physiological condition is nothing more than an .attempt to interpret certain aspects of the physiology of development. The axial gradients do not create anything, they are not the " cause" of growth or differentiation in the organism, they do not determine what organs shall develop in a particular protoplasm. Granting their existence, and granting the significance which I have assigned to them, the gradients represent merely certain physiological conditions, under which the hereditary mechanism of a protoplasm gives rise

to the order or pattern which we call axiate or axiate- symmetrical.1 The egg in most cases requires fertilization or the action of some other factor external to it to initiate development, but the specific hereditary constitution of the egg protoplasm with its potentialities of develop- ment is present, whether fertilization or initiation of development by other means occurs or not. Physio- logically speaking, the spermatozoon, or some other factor, merely sets the mechanism in motion or gives it the necessary speed, and development proceeds. Simi- larly, the gradient in the egg, whether it persists from earlier cell generations, or arises anew in the egg through differential exposure, is merely a physiological condition which determines that the hereditary mechanism shall give rise to a particular order or pattern. Alteration of the gradient relations alters the pattern, though the hereditary mechanism remains the same. The gradient is then nothing more than one of the physiological con- ditions under which the development of axiate organ- isms occurs and the surf ace -interior relation, whether a gradient or not, is merely another even more general physiological condition of organismic development.

The idea of a quantitative gradient in physiological condition as the condition initiating axiate development 1 In his recently published Vertebrate Zoology (1920), Newman has employed the gradient conception in connection with various problems of vertebrate evolution. I have found it desirable thus far to consider the gradient primarily as a physiological factor in individual develop- ment rather than as itself a factor in evolution, but even if the direct transfer of the conception to the evolutionary field is justifiable, some of the applications which this author has made seem to me to be at least open to question. In the present state of our knowledge, how- ever, differences of opinion concerning the scope and limitations of this, as of most other biological conceptions, are to be expected.

and of its origin in the final analysis through the action of an external factor does not require or depend upon any particular assumptions or theories concerning inheritance or evolution. Some of the critics of the conception have regarded it as Lamarckian, but it is not, though of course it might be used readily, if desired, in a Lamarckian way. Strictly speaking, it has nothing directly to do with either inheritance or evolution, except in so far as it maintains that axiation or polarity, and symmetry are not inherent properties or charac- teristics of protoplasm. It is fundamentally a physio- logical conception formulated on the basis of many different lines of observational and experimental evi- dence, and while its formulation in the present state of our knowledge is necessarily incomplete and perhaps vague, I believe a fair consideration of the evidence now available justifies the conclusions advanced.

The determination of a region of greater activity in a cell or cell mass by differential exposure is merely the first step in the establishment of a physiological gradient. Living protoplasm is irritable or excitable, and excitation in a broad sense is conceived to consist in an acceleration of the fundamental activities of pro- toplasm or, more particularly, those concerned with energy liberation. A region, the activity of which has been increased by the action of some external factor, whether such factor be oxygen or a stimulus in the stricter sense, is a region of excitation as compared with other regions. Protoplasm also possesses the capacity to transmit excitation to a greater or less degree or, more strictly speaking, to transmit some sort of dynamic change resulting from excitation at one point, which induces excitation in adjoining regions. These may in turn excite still other regions, and so the process of excitation spreads, irradiates, or is transmitted. In the absence of highly specialized conducting paths in pro- toplasm we find, however, that this capacity for trans- mission is apparently limited by distance. In other words, the process of excitation undergoes a decrement in intensity or effectiveness with increasing distance from the point of origin and at a greater or less distance

disappears or becomes ineffective. There is a large body of evidence for the existence of such a decrement in plant tissues and in the simpler animals.1 It is a familiar fact of observation in the simpler organisms that the effects of slight stimuli are effective only over short distances, while those of more intense stimuli are effective over greater distances, perhaps throughout the whole organism. If the energy concerned in the excita- tion process is derived from the original excitation, and transmission is merely the transmission of this energy without any further energy liberation at the various points of the transmission path, a decrement is readily accounted for as due to the work done in the course of transmission. If, on the other hand, energy is liberated at each point affected by the transmitted change, a decrement and limited range of effectiveness is possible only in protoplasm in which a relation between the energy or the intensity of stimulus and the energy or the intensity of excitation exists, in other words only when the same protoplasm is capable of different degrees of excitation and when the energy or the intensity of stimulus decreases in the course of transmission. In the transmission of excitation in the less highly specialized protoplasms such a relation evi- dently does exist. It is only in tissues for which the :<all-or-none" law holds good, that is, in which maximal excitation occurs with any stimulus above the threshold^ that transmission of excitation without decrement is

possible, but it may be pointed out that even in such tissue a transmission decrement may occur, if the trans- mission path is of such a kind that the maximal excita- tion decreases along it. Certain kinds of muscle tissue and medullated nerve have been found to react according to the "all-or-none" law, although even the medullated nerve, the most highly specialized of conducting tissues, shows a decrement when cooled, partially asphyxiated or partially anesthetized.1 But we are concerned rather with general protoplasmic excitation and transmission than with the highly specialized conductive processes, and concerning the existence, at least very generally, of a decrement in such transmission there seems to be no doubt.

When a region of relatively high activity or of excita- tion is established by the action of an external factor in a protoplasm which is not specialized with reference to transmission, transmission from this region must occur to some extent, and from our knowledge of pro- toplasmic transmission in general we must expect that a decrement in intensity or in effectiveness of the trans- mitted change will occur. In other words, excitation and transmission in protoplasm in general usually give rise to an excitation gradient. The evidence at hand concerning the existence and origin of the axial gradients

1 As regards nervous conduction see Biedermann (1903 and refer- ences), Boruttau (1901), Ducceschi (1901), Fischer (1911), Lodholz (1913), Lucas (1917, particularly chaps, ii-v and references). The question of the normal occurrence of a decrement in certain regions of the nervous system is also considered by Lucas. Tashiro has observed a gradient in CO2 production in certain nerves, the rate decreasing in the direction of functional conduction (Tashiro, 1914, 19150, £, 1917, particularly chap. iv). I have found that susceptibility in certain nerves decreases in the direction of functional conduction (Child,

in organisms, which was briefly discussed in the preced- ing chapters, indicates that these gradients originate as such excitation gradients and become more or less per- manently fixed or established in the protoplasm through the persistence for a certain length of time of the dif- ferential exposure, the region of activity or excitation determined by it and the resulting excitation gradient. As a basis for further consideration of the problem some discussion of excitation and transmission in general is necessary. Various theories of excitation and transmission have been advanced in the past, but most of them are either highly speculative or too exclu- sively concerned with the process as it occurs in nerve.1 These theories differ both as regards the cause of the electromotor phenomena characteristic of excitation and transmission and their role in the process of trans- mission. According to some authorities the electrical changes are merely an incidental feature of the process, a result or a by-product, so to speak, while others hold that they are the chief factors in transmission. Again, according to some, the electrical changes have a purely physical basis, while others believe that chemical re- actions are concerned in the excitation process and in bringing about the electrical changes. Some twenty years ago Waller (1897, 1903) maintained that excita- tion is associated with metabolic activity and more recently Tashiro (1917 and earlier papers) has shown

that excitation, even in the nerve fiber, is accompanied by increase in CO2 production and holds that irritability and excitation in general are associated with oxidative metabolism. Bayliss and Lucas have criticized these conclusions on the ground that Hill (1912) has been unable to dis- cover any appreciable heat production associated with the transmission of the nervous impulse, and have pointed out the possibility that the C02 produced by nerves may be dissolved in the tissues, and that its increase on electrical stimulation may be due to the rise in temperature occasioned by such stimulation. The demonstration by Tashiro and by Riggs (1919) that chemical stimulation, as well as electrical stimula- tion, is accompanied by increase in CO2 appears to an^ver this objection. But whatever the situation may prove to be in the highly specialized excitatory process in the nerves of higher animals, there seems to be no doubt that protoplasmic excitation in general involves metabolic and particularly oxidative reactions. Excita- tion in its most primitive form apparently involves in one way or another an acceleration in the fundamental activities of living protoplasm, particularly those con- cerned in the liberation of energy.

In a series of papers R. S. Lillie1 has developed a general theory of excitation and transmission based on extensive experimental evidence and in terms of current physicochemical conceptions. Since this theory is con- cerned to some extent with the process of excitation and transmission in general and not merely with the process as it occurs in the highly specialized tissues, it will serve as a point of departure for interpretation of the facts in the following discussion. It may be pointed out, how- ever, that the conception of the gradients and their significance in the organism do not depend upon this or any other particular theory of excitation but upon the observed facts.