The Origin and Development of the Nervous System

type "neuroid" transmission. While this term is undoubtedly useful, it should not be forgotten that such neuroid transmission represents merely a stage in the development and evolution of the excitation- transmission process and its morphological substratum. Even though a definite, anatomical conducting path is not distinguishable, the direction of transmission of the impulses in ciliated surfaces is more or less definitely determined. Griitzner (1882) and Kraft (1890) observed that the increased ciliary activity following direct mechanical stimulation of a local area of a ciliated epithelium is transmitted very largely in one direction from the stimulated area, the effect being observed as a band of increased activity, becoming somewhat broader with increase in distance from the stimulated area, but remaining definite in direction. Briicke's reimplanta- tion experiment, referred to above, also shows that the direction of transmission is definite. As far as I am aware, no adequate physiological grounds for such definiteness of direction in the absence of a definite anatomical path have been discovered by earlier authors, but it is evident from what has been said in earlier chapters that the physiological gradients afford a basis for definitely directed transmission, even in the absence of localized and directed anatomical paths. We expect transmission to occur chiefly downward in such gradient, and, as a matter of fact, in those cases in which the direction of the physiological gradients is known this is very generally, if not always, the case as regards the usual or normal direction. In many ciliated surfaces the direction of transmission may be reversed by ade- quate stimulation of lower levels of the gradient, but

such reversal means merely that a region of excitation arises in consequence of the stimulation and that the impulses initiated in it give rise to a temporary local gradient which overcompensates and reverses the general gradient for the time being and within a certain distance from the stimulated region, and so makes possible transmission in a direction opposite to the normal direction. The ctenophore plate row, discussed in the following section, affords beautiful examples of transmission and its reversal in relation to physiological gradients.

From the present viewpoint these cases of so-called neuroid transmission on ciliated surfaces appear, so far as the facts are at hand, to be merely cases of general cellular or protoplasmic transmission in relation to a general physiological gradient or to a temporary excita- tion gradient. The suggestion advanced by Parker (1905, p. 419) that neuroid transmission occurs irrevers- ibly in one direction and that mechanical transmission is probably concerned in reversal is quite unneces- sary. The directive factor appears beyond a doubt to be the physiological gradient and reversal to be asso- ciated with local or more or less extended reversal of this gradient by local excitation, or in some cases by developmental changes.

The body of the ctenophore possesses eight rows of motor organs, extending from a region near the aboral or apical end toward the oral end (Fig. 67). Each of the plates in a row is a flat, somewhat paddle-like different intervals, these be- after Chun), ing long, in some species several seconds, when the animals are completely undisturbed and quiescent, and much shorter, often only a small fraction of a second, when the animals are excited. Since the progress of the wave along the row is obviously the visible motor effect of a transmitted impulse of some sort, these plate rows constitute most interesting material for the experimental investigation of a relatively primitive type of transmission, and their behavior under various conditions has been repeatedly studied. Extended discussion of the literature of this subject is impossible at this time, but some of the earlier work has been considered in a recent paper (Child, and in an earlier study by Parker (1905).

Before considering the behavior of the plate row, the question of the nature of the transmission requires some attention. Verworn (1890) believed it to be mechanical and others have supposed that a true nervous transmission is concerned, but Parker (1905) has con- cluded on the basis of experimental evidence that transmission in the plate row cannot be mechanical and that, since definite nervous structures have not been found, it must be neuroid. I agreed with this conclusion, but more recent observations of my own on several species raise the question whether true nervous transmission is not more intimately concerned in at least some portions of the path of the impulse than has been previously believed. About the apical sense organ are four tissue masses corresponding to the four quadrants of the body. In some ctenophores, e.g., Mnemiopsis, Bolina, a single ciliated groove passes orally from each of these four masses and at some dis- tance from the apical end each groove divides into two, and each of the eight grooves thus formed connects with the apical end of a plate row. In other forms, e.g., Pleurobrachia, two ciliated grooves arise separately from each of the four masses and connect with the plate rows without division. When the animals are vitally stained with methylene blue a very definite differential staining occurs. The four tissue masses about the apical sense organ stain rapidly and deeply and appear as aggregations of cells. Each ciliated groove shows a perfectly definite, deeply stained strand, and the origin of this strand from the cell mass, its division into two in those forms in which it does divide and its connection with the plate row can be followed with the greatest

ease. The four tissue masses are generally regarded as nervous in character and the very definite differential staining of these and of the strands arising from them make it highly probable that the strands are actually nerves. Each strand terminates in an expanded end in contact with the base of the first plate of each row. Methylene blue has been very widely used as a more or less specific stain intra mtam for nervous structures, and the very definite and sharply bounded differential staining of these structures in ctenophores leaves little doubt as regards their nervous nature. Moreover, in forms like Mnemiopsis, in which four strands arise from the apical masses and each strand divides into two and passes to two plate rows, the two plate rows thus " innervated'' always show the same rhythm of beats, while in Pleurobrachia and other forms, in which two strands arise separately from each mass and pass with- out division to the two plate rows of a quadrant, these two rows often show an independent rhythm when the animal is more or less quiescent, but beat synchron- ously in stimulated animals. The cells of the plate row also stain with methylene blue earlier and more deeply than the general body surface, but no structures which can be identified as distinctly nervous appear in them. In the light of all these facts I am inclined to believe that actual nervous connection exists between the plate row and the apical nervous tissue, and that the plate row, as regards its relation to this tissue, represents a motor end organ or effector, but one capable of initiating and transmitting impulses. The paths from the apical nervous tissue are, however, not the only paths of trans- mission, for, as various authors have noted, the plate

row may be affected by direct stimulation of the mouth region or other parts of the body surface, even after the apical nervous tissue and the sense organ have been removed. Evidently also the plate row is not a pure effector, for it is able to initiate and transmit impulses. After section of the plate row at any level, transmission of the impulses from the apical region ceases at the level of section, but the portion of the row oral to the level of section begins, usually almost at once, or within a few moments, to beat with a rhythm of its own and the impulses are transmitted orally from the level of section. Moreover, even a single plate isolated by section from its neighbors may show rhythmic beats.

An investigation of the plate row in the genera, Mnemiopsis (Child, 191 yc), Bolina, Beroe, and Pleuro- brachia (unpublished) by means of the susceptibility method shows that the row is a susceptibility gradient, its apical end being the most susceptible region. When individuals are exposed to the action of inhibiting agents, e.g., cyanide and various anesthetics, the apical end of the row is affected first, and at a certain stage the impulses from the apical organ appear at the apical end of the row but are transmitted only a short distance instead of the whole length of the row, as in normal animals. Under these conditions the lower levels of the row, being less susceptible, develop rhythmic impulses of their own. As I have pointed out (191 yc), they are physiologically isolated. At first there may be only one such rhythm, originating in the region where the apical impulse disappears, but later, as the action of the agent progresses, several different rhythms may appear at different levels of the row, those of the more

oral levels usually being at first more rapid than those of the more apical. In most individuals, particularly those of Pleurobrachia and Beroe, a second region of high susceptibility is present at the extreme oral end of the row, where new plates are developing as the animal grows. This region is less susceptible than the apical end, but more susceptible than the middle regions of the row. As regards its relation to the gradient of the row, this region corresponds to the secondary posterior growing region of many animals (see chap. viii). Since it is normally less completely under the control of the apical region, as various facts indicate, this oral region becomes physiologically isolated and shows independent rhythm at a very early stage of the action of inhibit- ing agents. Complete disappearance of transmitted impulses occurs first at the apical end and progresses orally along the row, except as regards the extreme oral region. Even after transmission is completely abolished in a particular region, independent movements of slight amplitude may continue for some time in single plates. Usually the apical and the extreme oral regions are dead and opaque white in consequence of colloid changes before inhibition in the least susceptible regions is complete. This sequence of events, which occurs in essentially the same way in all the rows at about the same rate, affords a beautiful example of the gradual obliteration of apical dominance and of the progressive physiological isolation of various levels of the gradient, as determined by the differences in susceptibility at different levels.

The plate row also shows certain very interesting features as regards reversal of the direction of transmis- sion. Such reversal is most easily brought about in Pleurobrachia. In a normal individual of Pleuro- brachia which has been undisturbed for a time and shows a slow rhythm, reversal is readily produced by direct mechanical stimulation of any level of the row. Such a stimulation usually gives rise to rhythmic impulses which are transmitted in both directions, and the rhythm of the impulse transmitted orally is almost always more rapid than that of the impulse transmitted api- cally. Evidently the region of stimulation is able to initiate impulses with greater frequency down the gradient than in the opposite direction. Moreover, the first impulses transmitted apically usually disappear within a short distance, often only two or three plates from the level of stimulation, but if the stimulation is continued they gradually extend their range and after a few seconds may be transmitted all the way to the apical end of the row. When impulses passing in opposite directions meet, the result differs in different cases. They may obliterate each other completely, so far as further motor effect is concerned, or one of them may obliterate the other and pass on, sometimes with a visibly lower rate of transmission than before, appar- ently because of its decreased intensity. In general the rate of transmission of the impulse passing toward the apical end is lower than that in the opposite direction, or becomes lower as the impulse travels.

In normal individuals of Pleurobrachia reversal of direction of transmission over the whole length of the row sometimes occurs apparently spontaneously for a short time, and it may usually be brought about by direct stimulation of the oral end of the row. In this case also the first impulses are transmitted only a short distance, but the range gradually increases with con- tinued stimulation, until they arrive at the apical end. On cessation of the stimulation the reverse occurs, the normal apical impulses gradually extending their range as the oral impulses decrease theirs, until finally the normal direction is re-established. In the earlier stages of the action of inhibiting agents reversal of direction often occurs in Pleurobrachia without external stimula- tion, beginning at the oral end and often extending over the whole row. In such cases the greater degree of inhibition of the more susceptible apical regions has obliterated or reversed the original gradient to such a degree that transmission toward the apical end becomes the normal direction. Similar reversal of physiological gradients has been observed in many other cases. Some degree of reversal may be produced in Beroe by these methods, but usually only in the oral half of the row, and I have never seen the reversal extend to the apical end. In Mnemiopsis and Bolina reversal has been seen only at the extreme oral end and then only in the presence of inhibiting agents. Apparently the excita- tion waves are not high enough or not steep enough to travel up the gradient in these forms.

In Pleurobrachia and Beroe a difference in rate of transmission appears, not only with difference in direc- tion, but with difference in degree of excitation of the animal. In a strongly stimulated individual the rhythm is much more rapid and the impulses travel more rapidly, in fact, almost twice as rapidly as in a quiescent animal. That the ctenophore plate row represents a physio- logical gradient with a rise at its growing oral end is demonstrated not only by the susceptibility method

but in various other ways. The behavior of the impulses with respect to the two directions not only constitutes additional evidence for the existence of the gradient, but is, I believe, inexplicable on any other basis. In this behavior we find in the first place dominance and subordination in the normal transmission, in which all levels are subordinate to the apical end, although each level is capable of initiating an impulse. Second, by decreasing experimentally apical dominance or by excit- ing subordinate levels to a sufficient degree, physiological isolation may be brought about, and, according to con- ditions, the isolated region either retains the original gradient or develops a new one, as in the case of reversed transmission. In fact, the plate row constitutes an example, almost diagrammatic in its clearness, of the dynamic relations of parts as determined by a physio- logical gradient. Whether the relations depend upon the presence of nervous structure or upon neuroid trans- mission is of minor importance from the physiological viewpoint. In any case the plate row exhibits the most various degrees of integration of the organismic type, and may, I believe, be regarded almost as a model of a physiological gradient.

It has already been shown that certain axiate organs behave in many respects much like the ctenophore plate row and probably further physiological analysis will show similar conditions in many other axiate organs. The vertebrate heart is a case in point. In its earlier embryological stages the heart is a tubular structure and Hyman has found that in both the fish and the chick (unpublished) the heart of the earlier stages shows a distinct susceptibility gradient, the end which is to become the sinus end being the high end of the gradient. This end becomes functionally the dominant region, the pacemaker, and the only conceivable basis for this dominance is the fact that it is originally the high end of the gradient. Its dominance depends essentially upon the fact that it is able to give rise to impulses with a more rapid rhythm than other levels under the same general conditions, and this ability is unquestionably associated with its position in the gradient. It makes little difference for this conception whether we accept the myogenic or the neurogenic theory of the heart beat.

If the sinus region is cooled or otherwise inhibited, the beat may be initiated by lower levels of the gradient, but the rhythm is slower. Any level of the heart is able to initiate a beat when sufficiently isolated from the sinus region, either by inhibition of the latter, by block- ing the passage of the beat, or by sufficient stimulation of the level concerned. Even reversal in direction of transmission of the beat is possible to a greater or less degree under conditions which inhibit the sinus region and excite the arterial end, that is, which tend to reverse the gradient. In all these respects the heart as a con- ducting path and as a gradient resembles the plate row. The invertebrate heart differs more or less in form in different groups, but is in general tubular, and, though it has not yet been demonstrated to be a physiological gradient, it behaves in general like one.

The ascidian heart shows periodic reversal in the direction of transmission of the beat, the reversal occurring after a few or a considerable number of beats, according to species and conditions. The heart is tubular but evidently is not a fixed physiological gradient. I have already suggested elsewhere (191 yc) that the reversal of the direction of beat is determined by fatigue of the temporarily dominant region with consequent decrease of its dominance. This decrease in dominance permits the other end of the tube, which is recovering or has recovered from fatigue, to initiate a beat and become dominant. The situation is practically identical with the spontaneous reversal which occurs in Pleuro- brachia in the early stages of inhibition (p. 219) in consequence of the decrease in dominance of the more susceptible apical region and the physiological isolation of the oral end of the row.

Alvarez and his co-workers (p. 48) have demonstrated that the stomach, the small intestine, and the colon of mammals and of at least some other vertebrates represent physiological or metabolic gradients. Here, as in the other cases described, the relations of dominance and subordination exist and determine the normal direction of peristalsis, and physiological isolation and independent function of parts as well as reversal in direction of per- istalsis may be brought about in essentially the same ways as in the plate row. Alvarez' evidence for the existence of the gradient rests on many different lines of experimental investigation and is now very complete. Moreover, he has pointed out that recognition of the presence of the gradient in the alimentary tract of the higher vertebrates and man is not only of theoretical interest but of great practical significance in medicine.

There is considerable evidence from its functional behavior that the ureter is also a physiological gradient. Portions of the circulatory system of various animals, which transmit active contraction waves in a particular direction, also probably arise as physiological gradients. In fact we shall probably find that all organs which are axiate as regards their special function as well as organ- isms and organs which are structurally axiate represent physiological gradients. The gradient is, in short, as I have attempted to show, the physiological basis and starting-point of the organismic type of relation, and structural and functional relations are merely different aspects of the general pattern.

The "all-or-none" principle or so-called law main- tains that intensity or degree of excitation is independent of the strength or intensity of the stimulus for all stimuli above the threshold, the minimum stimulus which brings about reaction. In other words, in tissues which show the uall-or-none': type of reaction, all stimuli above the threshold produce maximal excitation. Lucas (1917) has concluded from his and Adrian's investiga- tions that the nerve fiber, the axon, shows this type of reaction in the vertebrates.

Both Lucas and R. S. Lillie conclude further that if a conducting tissue, e.g., the nerve fiber, reacts according to the "all-or-none': principle, it must necessarily be capable of conducting to an indefinite distance without decrement an excitation once initiated. This conclu- sion is correct, provided the maximal excitation is the same in degree or intensity at all points of the path, hi other words, if all points of the path are in the same physiological condition as regards their capacity for exci- tation. If, however, the conduction path is a prede- termined gradient the maximal excitation may differ quantitatively at different levels of the gradient, that is, a gradient in the quantitative characteristics of maximal excitation may occur along the path. Moreover, we might expect to find a rise in threshold with progress down the gradient. Such a path, even though the reaction is strictly "all or none" at all levels, may con- duct with a decrement. If it is long enough, a limit may even be reached, beyond which conduction cannot proceed because the stimulus resulting from maximal excitation at this point is so weak that it falls below the threshold for points beyond. The "all-or-none" type of reaction does not then exclude the possibility of con- duction with a decrement and of a limited range or distance of conduction. In such a path, however, the steepness of the decrement and the position of the limit are determined by the path, not by the quantitative characteristics of the original stimulus.

As pointed out in chapter iv, the primitive forms of transmission do not show the uall-or-none': reaction, but quantitatively different stimuli determine quanti- tatively different reactions between certain limits. A slight stimulus may produce a slight response only in the region immediately about the point of stimulation, a stronger stimulus a stronger response, which is trans- mitted to a greater distance, and so on up to the maxi- mum. Even in more or less specialized conduction paths like the ctenophore plate row such differences appear and are different in degree according to the direction of conduction in the gradient. Even in the

nervous reticulum such differences have been observed in various cases. In the coelenterate tentacle, for example, impulses pass more readily and to greater distances basipetally than in the opposite direction. But the observations of Mayer (1906, 1908, 1916) and Harvey (1912) on conduction in the medusa, Cassiopea, indicate that in this form the nerve net may conduct both with and without a decrement. Slight stimuli may produce reactions which disappear within a short distance, but the conduction of the "trapped'1' excita- tion waves in circular strips of the medusa body may continue around the strip for days, or until it dies, without appreciable decrement.

It is evident from these facts that in Cassiopea some submaximal excitations are conducted with a decrement, while maximal excitations, or those above a certain quantitative level, are conducted without appreciable decrement, at least in the direction around the body. Apparently in such cases as this the weaker excitations do not set in action the whole of the excitatory mechan- ism, perhaps do not bring about certain chemical reactions or other changes, and the effects of such excitations as stimuli on following points of the path also give partial excitations weaker than the original. Maximal or sufficiently strong excitation, however, sets the whole mechanism in action and brings about the same result at following points of the path; consequently no decrement appears. Whether such differences in conduction are characteristic of the reticular type of nervous system is not yet known.

It is generally assumed that the axon of the higher animals conducts without a decrement, but it has already been pointed out that if the nerve fiber is a physiological gradient when the impulse passes, a decrement may occur, even though the reaction be "all or none." During its growth period the axon possesses a very active growing tip, and for the time being a gradient probably extends a greater or less distance from this tip. When the axon reaches its end organ the tip as a region of special activity disappears, and the gradient associated with the tip probably disappears soon after. Exactly such changes as these occur in many plant axes of limited growth. A gradient present during the growth period in relation to a growing tip may disappear completely after the tip ceases its activity at the end of the growth period, and in some cases new gradients may arise. In general the axon grows down a general or local gradient, and after its growth ceases it lies in the electrical field of the gradient which deter- mined its growth, provided this gradient persists. It seems probable that under such conditions the different levels, particularly of the longer axons, must often be in different physiological condition. Moreover, the relation of the axon to the cell body, whether it be a matter of transportative or of transmissive correla- tion, may determine differences of condition at different levels of the axon.

The evidence, as far as it goes, indicates that these gradients do exist in some axons, at least. Tashiro (1914, 1915^, b) has found a gradient in CO2 production in the resting nerve, decreasing in the direction of func- tional conduction, and in certain of the nerves investi- gated by Tashiro I have found a susceptibility gradient in the same direction, as indicated by the structural death changes (Child, 1914^). Moreover, Adrian's data on the frog's sciatic indicate that more peripheral levels of the nerve are slightly less susceptible to the action of anesthetics (Lucas, 1917, chap, iv), though the author attributes the difference to defects in the appara- tus. The frog's sciatic is unfavorable material for the detection of a gradient because the length of nerve available is not great. If a gradient is present in the axon it certainly cannot be steep, and in many cases it may be appreciable only in long nerves. As the matter stands at present, the experimental evidence indicates that gradients do exist, at least in some axons, and I have shown that the "all-or-none" type of reaction is not in the least incompatible with the existence of these gradients.

Even if we admit that physiological gradients are present in axons after the growth period is over, the limit of conduction is certainly not reached nor even approached in the higher animals within the lengths of axons attained in the body. Possibly in some of the elongated lower vertebrates and invertebrates the axon length may approach the limit, but on this point we have no data. If it exists then in the higher forms, the axon gradient is important, not as determining functional relations, but rather as a record, an effect of the physio- logical gradients of the body.

When the vertebrate nerve is cooled or partially anesthetized, a relatively steep decrement in conduction appears, as many investigators have shown. Appar- ently under these conditions the more susceptible parts of the excitatory mechanism have been put out of action, so that the excitation which occurs is only partial and its effect as a stimulus on following points produces an even weaker excitation. A relatively steep decrement therefore appears in the cooled or anesthetized region, but if this is not so long that the excitation becomes completely ineffective the impulse regains its normal characteristics when it emerges from the experimentally altered region.

These and various other characteristics of nervous conduction suggest that the evolution of the excitatory process has consisted in part in the modification of the primitive mechanism in the direction of high intensity, but such modification evidently also involves high susceptibility. Inhibiting agents and conditions first eliminate, reversibly or irreversibly, these highly sus- ceptible features of the mechanism, but affect the more primitive factors more slowly. At certain stages of such action, therefore, highly specialized paths, such as the axon, appear to lose some of their specialization, and their behavior resembles that of more primitive paths.

In its functional relations as a part of the nervous system the neuron conducts impulses from the tips of the dendrites to the tips of the axons, but if an axon is artificially stimulated at some point, the electrical change, the negative variation, which accompanies the passage of a nervous impulse, runs in both directions from the point of stimulation. If this negative variation is a characteristic feature of the nervous impulse, as there is every reason to believe, the axon is evidently capable of conducting impulses in both directions, but, so far as we know, the impulse which passes toward the

cell body in an axon never affects another neuron.1 Con- ceivably such an impulse may be unable for some reason to pass the cell body, or it may be unable to pass through the dendrite, or finally it may be unable to pass the synapse. Current views regard the synapse as chiefly responsible for this functional irreversibility of direction of conduction from one neuron to another. Moreover, blocking, delay and the phenomena of sum- mation of impulses in the reflex arc are ordinarily attributed to the synapse. According to Lucas (1917) the junction between motor nerve and muscle is a region in which the nervous impulse apparently under- goes a decrement, and he suggests that conditions at this junction may throw some light on the problem of the synapse. To account for the delay and for summa- tion, it is commonly assumed that the synapse differs in some way from other parts of the nervous system as regards the excitation process in it. Either it is not readily excitable or the intensity or height of the excita- tion wave in it is less than in other parts. Consequently it constitutes to a greater or less extent an obstacle in the way of the impulse, but continued or repeated excitation may alter its condition so that the impulse is able to pass. Passage, however, is apparently in one direction only, from axon to dendrite, and some further assumption is necessary to account for the irreversibility of direction. Membranes, more or less irreversible as regards direction with respect to passage of chemical substances, do occur in other organs, but the physico- chemical basis of this characteristic is not known with