LECTURE I INTRODUCTORY — CO-ORDINATION OF THE SIMPLE REFLEX

Nowhere in physiology does the cell-theory reveal its presence more frequently in the very framework of the argument than at the present time in the study of nervous reactions. The cell-theory at its inception depended for exemplification largely on merely morphological observations; just as these formed originally the almost exclusive texts for the Darwinian doctrine of evolution. But with the progress of natural knowledge, biology has passed beyond the confines of the study of merely visible form, and is turning more and more to the subtler and deeper sciences that are branches of energetics. The cell-theory and the doctrine of evolution find their scope more and more, therefore, in the problems of function, and have become more and more identified with the aims and incorporated among the methods of physiology.

The physiology of nervous reactions can be studied from three main points of view.

In the first place, nerve-cells, like all other cells, lead individual lives, — they breathe, they assimilate, they dispense their own stores of energy, they repair their own substantial waste; each is, in short, a living unit, with its nutrition more or less centred in itself. Here, then, problems of nutrition, regarding each nerve-cell and regarding the nervous system as a whole, arise comparable with those presented by all other living cells. Although no doubt partly special to this specially differentiated form of cell-life, these problems are in general accessible to the same methods as apply to the study of nutrition in other cells and tissues and in the body as a whole. We owe recently to Verworn and his co-workers advances specially valuable in this field.

Secondly, nervous cells present a feature so characteristically developed in them as to be specially theirs. They have in exceptional measure the power to spatially transmit (conduct) states of excitement (nerve-impulses) generated within them. Since this seems the eminent functional feature of nerve-cells wherever they exist, its intimate nature is a problem co-extensive with the existence of nerve-cells, and enters into every question regarding the specific reactions of the nervous system. This field of study may be termed that of nerve-cell conduction.

But a third aspect which nervous reactions offer to the physiologist is the integrative. In the multicellular animal, especially for those higher reactions which constitute its behaviour as a social unit in the natural economy, it is nervous reaction which par excellence integrates it, welds it together from its components, and constitutes it from a mere collection of organs an animal individual. This integrative action in virtue of which the nervous system unifies from separate organs an animal possessing solidarity, an individual, is the problem before us in these lectures. Though much in need of data derived from the two previously mentioned lines of study, it must in the meantime be carried forward of itself and for its own sake.

The integration of the animal organism is obviously not the result solely of any single agency at work within it, but of several. Thus, there is the mechanical combination of the unit cells of the individual into a single mass. This is effected by fibrous stromata, capsules of organs, connective tissue in general, e. g. of the liver, and indeed the fibrous layer of the skin encapsulating the whole body. In muscles this mechanical integration of the organ may arrive at providing a single cord tendon by which the tensile stress of a myriad contractile cells can be additively concentrated upon a single place of application.

Integration also results from chemical agency. Thus, reproductive organs, remote one from another, are given solidarity as a system by communication that is of chemical quality; lactation supervenes post partum in all the mammary glands of a bitch subsequent to thoracic transection of the spinal cord severing all nervous communication between the pectoral and the inguinal mammae (Goltz). In digestive organs we find chemical agency co-ordinating the action of separate glands, and thus contributing to the solidarity of function of the digestive glands as a whole. The products of salivary digestion on reaching the pyloric region of the stomach, and the gastric secretion on reaching the mucosa of the duodenum, make there substances which absorbed duly excite heightened secretion of gastric and of pancreatic juice respectively suited to continue the digestion of the substances initiating the reaction (Bayliss and Starling, Edkins). Again, there is the integrating action effected by the circulation of the blood. The gaseous exchanges at one limited surface of the body are made serviceable for the life of every living unit in the body. By the blood the excess of heat produced in one set of organs is brought to redress the loss of heat in others; and so on.

But the integrative action of the nervous system is different from these, in that its agent is not mere intercellular material, as in connective tissue, nor the transference of material in mass, as by the circulation; it works through living lines of stationary cells along which it despatches waves of physico-chemical disturbance, and these act as releasing forces in distant organs where they finally impinge. Hence it is not surprising that nervous integration has the feature of relatively high speed, a feature peculiarly distinctive of integrative correlation in animals as contrasted with that of plants, the latter having no nervous system in the ordinary sense of the word.

The nervous system is in a certain sense the highest expression of that which French physiologists term the milieu interne. With the transition from the unicellular organism to the multi-cellular a new element enters general physiology. The phenomena of general physiology in the unicellular organism can be divided into two great groups; namely, those occurring within the cell, intracellular, and those occurring at the surface of the cell, in which forces that are associated with surfaces of separation have opportunity for play at the boundary between the organism and its environment. But in the multicellular organism a third great group of phenomena exists in addition; namely, those which are intercellular, occurring in that complex material which the organism deposits in quantity in the intercellular interstices of its mass as a connecting medium between its individual living units.

When the intercellular substance is solid, e. g. in many connective tissues, the physiological agencies for which it affords a field of operation are mechanical rather than chemical. The organism obtains from it scaffolding for supporting its weight, levers for application of its forces, etc., and in this degree the intercellular material performs an integrative function. Where the intercellular material is fluid, as in blood, lymph, and tissue juice, it constitutes a field of operation for agencies chemical rather than mechanical. The intricacy of the chemistry of this milieu interne is shown by nothing better than by the specificity of the precipitins, etc., the intercellular media for each separate animal species yielding its own particular kinds. The cells of a multicellular organism have therefore in addition to an environmental medium in which the organism as a whole is bathed, and to which they react either directly or through the medium of surface cells, an internal medium created by their organism itself, and in many respects specific to itself.

But the internal interconnection of the multicellular organism is not restricted to intercellular material. Intercellular material is, after all, no living channel of communication, delicately responsive to living changes though it may be. An actually living internal bond is developed. When the animal body reaches some degree of multicellular complexity, special cells assume the express office of connecting together other cells. Such cells, since their function is to stretch from one cell to another, are usually elongated; they form protoplasmic threads and they interconnect by conducting nervous impulses. And we find this living bond the one employed where, as said above, speed and nicety of time adjustment are required, as in animal movements, and also where nicety of spatial adjustment is essential, as also in animal movements. It is in view of this interconnecting function of the nervous system that that field of study of nervous reactions which was called at the outset the third or integrative, assumes its due importance. The due activity of the interconnection resolves itself into the co-ordination of the parts of the animal mechanism by reflex action.

It is necessary to be clear as to what we understand by the expression “reflex” action.

In plants and animals occur a number of actions the initiation of which is traceable to events in their environment. The event in the environment is some change which acts on the organism as an exciting stimulus. The energy which is imparted to the organism by the stimulus is often far less in quantity than the energy which the organism itself sets free in the movement or other effect which it exhibits in consequence of the application of the stimulus. This excess of energy must be referred to energy potential in the organism itself. The change in the environment evidently acts as a releasing force upon the living machinery of the organism. The source of energy set free is traced to chemical compounds in the organism. These are of high potential value, and in immediate or mediate consequence of the stimulus decompose partly, and so liberate external from internal energy. It is perfectly conceivable, and in many undifferentiated organisms, especially in unicellular, e. g. amoeba, is actually the case, that one and the same living structure not only undergoes this physico-chemical change at the point at which an external agent is applied, but is subject to spread of that change from particle to particle along it, so that there then ensue in it changes of form, movement. In such a case the initial reaction or reception of the stimulus, the spatial transmission or conduction of the reaction, and the motor or other end-effect, are all processes that occur in one and the same living structure. But in many organisms these separable parts of the reaction are exhibited by separate and specific structures. Suppose an animal turn its head in response to a sudden light. Large fields of its body take part in the reaction, but also large fields of it do not. Some of its musculature contracts, particularly certain pieces of its skeletal musculature. The external stimulus is, so to say, led to them by certain nerves in the altered form of a nervous impulse. If the neck nerves are severed the end-effect is cut out of part of the field; and the nerves themselves cannot exhibit movement on application of the stimulus. The optic nerve itself is unable to enter into a heightened phase of its own specific activity on the application of light. Initiation of nervous activity by light is the exclusive (in this instance) function of cells in the retina, retinal receptors. In such cases there exist three separable structures for the three processes — initiation, conduction, and end-effect.

These reactions, in which there follows on an initiating reaction an end-effect reached through the mediation of a conductor, itself incapable either of the end-effect or, under natural conditions, of the inception of the reaction, are “reflexes.” The conductors are nerve. Usually the spaces and times bridged across by the conductors are quite large, and easily capable of measurement. Now there occur cases, especially within the unicellular organism and the unicellular organ, where the spaces and times bridged are minute. In them spread of response may involve “conduction” (Poteriodendron, Vorticella) in some degree specific. Yet to cases where neither histologically nor physiologically a specific conductor can be detected, it seems better not to apply the term “reflex.” It seems better to reserve that expression for reactions employing specifically recognizable nerve-processes and morphologically differentiated nerve-cells; the more so because the process of conduction in nerve is probably a specialized one, in which the qualities of speed and freedom from inertia of reaction have been attained to a degree not reached elsewhere since not elsewhere demanded.

The conception of a reflex therefore embraces that of at least three separable structures, — an effector organ, e. g., gland cells or muscle cells; a conducting nervous path or conductor leading to that organ; and an initiating organ or receptor whence the reaction starts. The conductor consists, in the reactions which we have to study, of at least two nerve-cells, — one connected with the receptor, the other with the effector. For our purpose the receptor is best included as a part of the nervous system, and so it is convenient to speak of the whole chain of structures—receptor, conductor, and effector—as a reflex-arc. All that part of the chain which leads up to but does not include the effector and the nerve-cell attached to this latter, is conveniently distinguished as the afferent-arc.

The reflex-arc is the unit mechanism of the nervous system when that system is regarded in its integrative function. The unit reaction in nervous integration is the reflex, because every reflex is an integrative reaction and no nervous action short of a reflex is a complete act of integration. The nervous synthesis of an individual from what without it were a mere aggregation of commensal organs resolves itself into co-ordination by reflex action. But though the unit reaction in the integration is a reflex, not every reflex is a unit reaction, since some reflexes are compounded of simpler reflexes Co-ordination, therefore, is in part the compounding of reflexes. In this co-ordination there are therefore obviously two grades.

The simple reflex. There is the co-ordination which a reflex action introduces when it makes an effector organ responsive to excitement of a receptor, all other parts of the organism being supposed indifferent to and indifferent for that reaction. In this grade of co-ordination the reflex is taken apart, as if separable from all other reflex actions. This is the simple reflex. A simple reflex is probably a purely abstract conception, because all parts of the nervous system are connected together and no part of it is probably ever capable of reaction without affecting and being affected by various other parts, and it is a system certainly never absolutely at rest. But the simple reflex is a convenient, if not a probable, fiction. Reflexes are of various degrees of complexity, and it is helpful in analyzing complex reflexes to separate from them reflex components which we may consider apart and therefore treat as though they were simple reflexes.

In the simple reflex there is exhibited the first grade of co-ordination. But it is obvious that if the integration of the animal mechanism is due to co-ordination by reflex action, reflex actions must themselves be co-ordinated one with another; for co-ordination by reflex action there must be co-ordination of reflex actions. This latter is the second grade of co-ordination. The outcome of the normal reflex action of the organism is an orderly coadjustment and sequence of reactions. This is very patently expressed by the skeletal musculature. The co-ordination involves orderly coadjustment of a number of simple reflexes occurring simultaneously, i. e. a reflex pattern, figure, or “complication,” if one may warp a psychological term for this use; orderly succession involves due supercession of one reflex by another, or of one group of reflexes by another group, i. e. orderly change from one reflex pattern or figure to another. For this succession to occur in an orderly manner no component of the previous reflex may remain which would be out of harmony with the new reflex that sets in. When the change from one reflex to another occurs it is therefore usually a far-reaching change spread over a wide range of nervous arcs.

This compounding of reflexes with orderliness of coadjustment and of sequence constitutes co-ordination, and want of it inco-ordination. We may therefore in regard to co-ordination distinguish co-ordination of reflexes simultaneously proceeding, and co-ordination of reflexes successively proceeding. The main secret of nervous co-ordination lies evidently in the compounding of reflexes.

Co-ordination in the simple reflex. It is best to clear the way toward the more complex problems of co-ordination by considering as an earlier step that which was termed above, the first grade of co-ordination, or that of the simple reflex. From the point of view of its office as integrator of the animal mechanism, the whole function of the nervous system can be summed up in the one word, conduction. In the simple reflex the evidence of co-ordination is that the outcome of the reflex as expressed by the activity induced in the effector organ is a response appropriate to the stimulus imparted to the receptor. This due propriety of end-effect is largely traceable to the action of the conductor mediating between receptor and affector. Knowledge of the features of this “conduction” is therefore a prime object of study in this connection.

But we have first to remember that in dealing with reflexes even experimentally we very usually deal with them as reactions for which the reflex-arc as a whole and without any separation into constituent parts is laid under contribution. The reflex-arc thus taken includes the receptor. It is assuredly as truly a functional part of the arc as any other. But, for analysis of the arc’s conduction, it is obvious that by including the receptor we are including a structure which, as its name implies, adaptation has specialized for excitation of a kind different from that obtaining for all the rest of the arc. It is therefore advantageous, as we have to include the receptor in the reflex-arc, to consider what characters its inclusion probably grafts upon the functioning of the arc.

Marshall Hall^{28 *} drew attention to the greater ease with which reflexes can be elicited from receptive surfaces than from afferent nerve-trunks themselves; and this has often been confirmed (Eckhard, Biedermann). Steinach¹⁹⁰ has measured the lowering of the threshold value of stimulation when in the frog a reflex is elicited by a mechanical stimulus applied to skin instead of to cutaneous afferent nerve. The lowering is considerable. There are numerous instances in which particular reflexes can be elicited from the receptive surface by particular stimuli only. Goltz⁴⁵ endeavoured in vain to evoke the reflex croak of the female frog by applying to the skin electrical stimuli. Mechanical stimuli of non-nocuous kind were the only stimuli that proved effective. From the afferent nerve itself by direct stimulation the reflex could but rarely be elicited at all. Later Goltz’s pupil Gergens⁶⁸ succeeded in provoking the reflex by applying to the skin a mild discharge from an influence machine.

A remarkable reflex³⁰⁴ is obtainable from the planta of the hind foot in the “spinal” dog. The movement provoked is a brief strong extension at knee, hip, and ankle. This is the “extensor-thrust.” It seems obtainable only by a particular kind of mechanical stimulation. I have never succeeded in eliciting it by any form of electrical stimulation, nor by any stimulation applied directly to an afferent nerve-trunk.

Again, a very characteristic reflex in the cat is the pinna-reflex.³⁰⁴ If the tip of the pinna be squeezed, or tickled, or in some cases even touched, the pinna itself is crumpled so that its free end is turned backward, as in Darwin’s⁵⁰ picture of a cat prepared to attack. The afferent nerve of this reflex appears to be in part at least not the cranial fifth nerve, but the foremost cervical. The reflex emerges very early from the shock of de-cerebration and is submerged very late in chloroform narcosis. This reflex, easily elicitable as it is by various mechanical stimuli to the skin, I have never succeeded in provoking by any form of electrical stimulation.

The same sort of difference, though less marked in degree, is exhibited by the scratch-reflex.^{87, 126, 251, 252, 300} This reflex is one in which various forms of innocuous mechanical stimulation (rubbing, tickling, tapping) applied to the skin of the back behind the shoulder evoke a rhythmic flexion (scratching movement) of the hind limb, the foot being brought toward the seat of stimulation. This reflex in the spinal dog, although usually elicitable, varies much under various circumstances in its degree of elicitability. When easily elicitable it can be evoked by various forms of electrical stimulation as well as by mechanical; but when not easily elicitable electrical stimuli altogether fail, while rubbing and other suitable mechanical stimuli still evoke it, though not so readily or vigorously as usual.

A question germane to this is the oft-debated sensitivity of various internal organs. Direct stimulation of various afferent nerves of the visceral system is itself well known to yield reflexes on blood-pressure, etc. But in regard to the sensitivity of the organs themselves we have, on the one hand, the passage of bilestones, renal calculi, etc., accompanied by intense sensations, and on the other hand the insensitivity of these ducts and various allied visceral parts as noted by Haller⁶ and observed by surgeons working under circumstances favorable for examining the question. The stimulation which excites pain in these internal organs is usually of mechanical kind, e. g. calculus, and the surgeon’s knife and needle provide mechanical stimuli, and Haller and his co-workers in their research employed multiform stimuli, many of them mechanical in quality. But though mechanical, the latter are remote in quality from the former; the former are distensile. The action of a calculus can be imitated by injecting fluid of itself innocuous. Marked reflex effects can then be excited¹⁹⁴ from the very organs (Fig. 1), the cutting and wounding of which remains without effect. For Haller’s and the surgical experience to be harmonized with the medical evidence from calculi, etc., all that is necessary is that the mechanical stimulation be adequate, and to be adequate it must be of a certain kind. Thus we see that when the mechanical stimulation employed resembles that occurring in the natural accidents that concern medicine, the experimental results fall into line with those observed at the bedside.

Therefore we may infer provisionally—for the facts justify only a guarded judgment—that the part played by the receptor in the reflex-arc is in the main what from other evidence it is inferred to be in the case of the receptors as sense-organs; namely, a mechanism more or less attuned to respond specially to a certain one or ones of the agencies that act as stimuli to the body. We may suppose this special attuning acts as does specialization in so many cases, namely by rendering more apt for a certain kind of stimulus and at the same time less apt for stimuli of other kinds. The main function of the receptor is therefore²⁰⁵ to lower the threshold of excitability of the arc for one kind of stimulus, and to heighten it for all others. This is quite comparable with the low threshold for touch-sensation under mechanical stimulation applied to a hair (v. Frey)¹⁷² contrasted with the high threshold under electrical stimulation of the skin (v. Frey). Adaptation has evolved a mechanism for which one kind of stimulus is the appropriate, that is, the adequate stimulus: other stimuli than the adequate not being what the adaptation fitted the mechanism for, are at a disadvantage. Electrical stimuli are in most cases far the most convenient to use for experimental work, because of their easy control, especially in regard to intensity and time. But electrical stimuli not being of common occurrence in nature, there has been no chance for adaptation to evolve in the organism receptors appropriate for such stimuli. Therefore we may say that electricity never constitutes the adequate stimulus for any receptor, since it is always an artificial form of stimulus, and every adequate stimulus must obviously be a natural form of stimulation. It is therefore rather a matter for surprise that electrical stimuli applied to receptor organs are as efficient excitors of reflexes as they in fact prove to be. It is particularly in regard to a class of reflexes whose receptive cells seem attuned specially to react to nocuous agents, agents that threaten to do local damage, that electrical stimuli are found to be excellently effective. But the conditions of adaptation to stimuli appear here peculiar; and there will be better opportunity of considering them later.

We infer, therefore, that the main contribution made to the mechanism of the reflex-arc by that part of it which constitutes the receptor is selective excitability. It thus contributes to co-ordination, for it renders its arc prone to reply to certain stimuli, while other arcs not having that kind of receptor do not reply, and it renders its arc unlikely to reply to certain other stimuli to which other arcs are likely to respond. It will thus, while providing increase of responsiveness on the part of the organism to the environment, tend to prevent confusion of reactions (inco-ordination) by limiting to particular stimuli a particular reaction.

On the whole, we may regard the receptor as being concerned with the mode of excitation rather than with the features of conduction of the reflex-arc, and may now return to that conduction, which itself has important co-ordinative characters.

Nervous conduction has been studied chiefly in nerve-trunks. Conduction in reflexes is of course for its spatially greater part conduction along nerve-trunks, yet reflex conduction in toto differs widely from nerve-trunk conduction.

Salient among the characteristic differences between conduction in nerve-trunks and in reflex-arcs respectively are the following:

Conduction in reflex-arcs exhibits (1) slower speed as measured by the latent period between application of stimulus and appearance of end-effect, this difference being greater for weak stimuli than for strong; (2) less close correspondence between the moment of cessation of stimulus and the moment of cessation of end-effect, i. e., there is a marked “after-discharge;” (3) less close correspondence between rhythm of stimulus and rhythm of end-effect; (4) less close correspondence between the grading of intensity of the stimulus and the grading of intensity of the end-effect; (5) considerable resistance to passage of a single nerve-impulse, but a resistance easily forced by a succession of impulses (temporal summation); (6) irreversibility of direction instead of reversibility as in nerve-trunks; (7) fatigability in contrast with the comparative unfatigability of nerve-trunks; (8) much greater variability of the threshold value of stimulus than in nerve-trunks; (9) refractory period, “bahnung,” inhibition, and shock, in degrees unknown for nerve-trunks; (10) much greater dependence on blood-circulation, oxygen (Verworn, Winterstein, v. Baeyer, etc.); (11) much greater susceptibility to various drugs — anaesthetics.

These differences between conduction in reflex-arcs and nerve-trunks respectively appear referable to that part of the arc which lies in gray matter. The constituents of gray matter over and above those which exist also in nerve-trunks are the nerve-cell bodies (perikarya),¹⁷⁰ the fine nerve-cell branches (dendritic and axonic nerve-fibres), and neuroglia.

Neuroglia exists in white matter as well as in gray, and there is no good ground for attributing the above characteristics of conduction in reflex-arcs to that part of the arcs which consists of white matter. It is improbable, therefore, on that ground that the features of the conduction are due to neuroglia. Indeed there is no good evidence that neuroglia is concerned directly in nervous conduction at all. As to perikarya (nerve-cell bodies) the experiment of Bethe¹⁷⁸ on the motor perikarya of the ganglion of the second antenna of Carcinus, and the experiments of Steinach¹⁹⁰ on the perikarya of the spinal-root ganglion, also the observation by Langley²²⁰ that nicotin has little effect when applied to the spinal-root ganglion, though breaking conduction in sympathetic ganglia, all indicate more or less directly that it is not to the perikarya that the characteristic features of reflex-arc conduction are referable. Similarly, the experiments of Exner,⁷¹ and of Moore and Reynolds,¹⁹³ detecting no delay in transmission through the spinal-root ganglion, — though observations by Wundt⁶⁹ and by Gad and Joseph¹²¹ had a different result,—withdraw from the perikaryon the responsibility for another feature characteristic of reflex-arc conduction. Again, histological observations by Cajal, van Gehuchten, and others, indicate that in various cases the line of conduction may run not through the perikaryon at all, but direct from dendrite stem to axone.

As to the nerve-cell branches (dendrites, axones, and axone-collaterals) which are so prominent as histological characters of gray matter, they are in many cases perfectly continuous with nerve-fibres outside whose conductive features are known by study of nerve-trunks, and they also are themselves nerve-fibres, though smaller in calibre than those outside. It seems therefore scarcely justifiable to suppose that conduction along nerve-fibres assumes in the gray matter characters so widely different from those it possesses elsewhere as to account for the dissimilarity between reflex-arc conduction and nerve-trunk conduction respectively.

In this difficulty there rises forcibly to mind that not the least fruitful of the facts which the cell-theory rests upon and brings together is the existence at the confines of the cells composing the organism of “surfaces of separation” between the adjacent cells. In certain syncytial cases such surfaces are not apparent, but with most of the cells in the organism their existence is undisputed, and they play an important rôle in a great number of physiological processes. Now in addition to the structural elements of gray matter specified above, there is one other which certainly in many cases exists. The gray matter is the field of nexus between neurone and neurone. Except in sympathetic ganglia, the place of nexus between neurone and neurone lies nowhere else than in gray matter. We know of no reflex-arc composed of one single neurone only. In other words, every reflex-arc must contain a nexus between one neurone and another. The reflex-arc must, therefore, on the cell-theory, be expected to include not only intracellular conduction, but intercellular conduction. But on the current view of the structure of the nerve-fibres of nerve-trunks the conduction observed in nerve-trunks is entirely and only intracellular conduction. Perhaps, therefore, the difference between reflex-arc conduction and nerve-trunk conduction is related to an additional element in the former, namely, intercellular conduction. If there exists any surface or separation at the nexus between neurone and neurone, much of what is characteristic of the conduction exhibited by the reflex-arc might be more easily explicable. At the nexus between cells if there be not actual confluence, there must be a surface of separation. At the nexus between efferent neurone and the muscle-cell, electrical organ, etc., which it innervates, it is generally admitted that there is not actual confluence of the two cells together, but that a surface separates them; and a surface of separation is physically a membrane. As regards a number of the features enumerated above as distinguishing reflex-arc conduction from nerve-trunk conduction, there is evidence that similar features, though not usually in such marked extent, characterize conduction from efferent nerve-fibre to efferent organ, e. g., in nerve-muscle preparation, in nerve-electric-organ preparation, etc. Here change in character of conduction is not due to perikarya (nerve-cell bodies), for such are not present. The change may well be referable to the surface of separation admittedly existent between efferent neurone and effector cell.

If the conductive element of the neurone be fluid, and if at the nexus between neurone and neurone there does not exist actual confluence of the conductive part of one cell with the conductive part of the other, e. g. if there is not actual continuity of physical phase between them, there must be a surface of separation. Even should a membrane visible to the microscope not appear, the mere fact of non-confluence of the one with the other implies the existence of a surface of separation. Such a surface might restrain diffusion, bank up osmotic pressure, restrict the movement of ions, accumulate electric changes, support a double electric layer, alter in shape and surface-tension with changes in difference of potential, alter in difference of potential with changes in surface-tension or in shape, or intervene as a membrane between dilute solutions of electrolytes of different concentration or colloidal suspensions with different sign of charge. It would be a mechanism where nervous conduction, especially if predominantly physical in nature, might have grafted upon it characters just such as those differentiating reflex-arc conduction from nerve-trunk conduction. For instance, change from reversibility of direction of conduction to irreversibility might be referable to the membrane possessing irreciprocal permeability. It would be natural to find in the arc, each time it passed through gray matter, the additive introduction of features of reaction such as characterize a neurone-threshold (Goldscheider).¹⁸⁷ The conception of the nervous impulse as a physical process (du Bois Reymond) rather than a chemical, gains rather than loses plausibility from physical chemistry. The injury-current of nerve seems comparable in mode of production (J. S. Macdonald)²³⁴ with the current of a “concentration cell,” a mode of energy akin to the expansion of a gas and physical, rather than chemical, ‘volume-energy.’ Against the likelihood of nervous conduction being pre-eminently a chemical rather than a physical process must be reckoned, as Macdonald well urges, its speed of propagation, its brevity of time-relations, its freedom from perceptible temperature change, its facile excitation by mechanical means, its facilitation by cold, etc. If it is a physical process the intercalation of a transverse surface of separation or membrane into the conductor must modify the conduction, and it would do so with results just such as we find differentiating reflex-arc conduction from nerve-trunk conduction.

As to the existence or the non-existence of a surface of separation or membrane between neurone and neurone, that is a structural question on which histology might be competent to give valuable information. In certain cases, especially in Invertebrata, observation (Apathy, Bethe, etc.) indicates that many nerve-cells are actually continuous one with another. It is noteworthy that in several of these cases the irreversibility of direction of conduction which is characteristic of spinal reflex-arcs is not demonstrable; thus the nerve-net in some cases, e. g. Medusa, exhibits reversible conduction (Romanes, Nagel, Bethe, and others). But in the neurone-chains of the gray-centred system of vertebrates histology on the whole furnishes evidence that a surface of separation does exist between neurone and neurone. And the evidence of Wallerian secondary degeneration is clear in showing that that process observes strictly a boundary between neurone and neurone and does not transgress it. It seems therefore likely that the nexus between neurone and neurone in the reflex-arc, at least in the spinal arc of the vertebrate, involves a surface of separation between neurone and neurone; and this as a transverse membrane across the conductor must be an important element in intercellular conduction. The characters distinguishing reflex-arc conduction from nerve-trunk conduction may therefore be largely due to intercellular barriers, delicate transverse membranes, in the former.

In view, therefore, of the probable importance physiologically of this mode of nexus between neurone and neurone it is convenient to have a term for it. The term introduced has been synapse.¹⁷⁰

The differences between nerve-trunk conduction and reflex-arc conduction are so great as to require for their exhibition no very minute determination of the characters of either; but we may with advantage follow these differences somewhat further. In doing so we may take the reflexes of the hind limb of the spinal dog as a field of exemplification.

Reflex latency. A dissimilarity between nerve-trunk conduction and reflex-arc conduction which has often been stressed is the slowness of the latter as measured by the latent interval between application of stimulus and appearance of end-effect. In nerve-trunks the interval between the moment of stimulation and the appearance of response (electrical) at any distant point is strictly proportional to the distance of that point from the seat of stimulation. There is in the nerve-trunk no measurable delay or latent interval for the response at the seat of excitation. The latent time for nerve-trunk response is therefore entirely a propagation time. The speed of propagation in frog’s nerve at 15° C. is about 3 cm. per sigma (σ = .001 second). We may compare with this the latent period of the flexion-reflex of the “spinal” dog’s hind leg. The movement of this reflex is a flexion at knee, hip, and ankle. It is easily and regularly evoked by nocuous or electrical stimuli applied to the skin of the limb or to any afferent nerve of the limb. For measurements of the reflex latency I have stimulated with break or make shocks of regular but varied frequency. Assuming that in warm-blooded nerves the conduction is the same (Helmholtz found it faster) as in the frog, and that the length of the reflex-arc of the dog’s knee is two thirds of a metre, and assuming that we may add 5 σ for mechanical latency of the flexor contraction of the limb, we should have about 27 σ as the latent time for the flexion-reflex, supposing its conduction proceeded as does nerve-trunk conduction. But, as a fact, a period double that is common enough for this reflex under ordinary moderate intensities of stimulation.

But with intenser stimuli the latent period of this reflex is much less. A period of 30 σ from commencement of stimulus to commencement of mechanical response is not then uncommon. I have met, at shortest, with 22 σ. There is here little difference between speed of reflex conduction and speed of nerve-trunk conduction. Similarly François Franck¹¹¹ has recorded latent periods for reflex action differing little from those of simple nerve-trunk conduction. Thus, 17 σ were obtained for a reflex contraction of the crossed gastrocnemius evoked by stimulation of the afferent root of the first lumbar nerve. These short latencies Franck obtained with strong stimuli.

It would seem, therefore, that the more intense the stimulation the more the conduction along the reflex-arc comes to resemble in speed the conduction along simple nerve-trunks.

It is with mild stimuli that the difference in speed between reflex conduction and nerve-trunk conduction becomes most obvious. The latent period for the flexion-reflex, then, lies usually between 60 σ and 120 σ. I have met with it as long as 200 σ. There is no good evidence that the speed of propagation in nerve-trunk conduction is in response to weak stimuli appreciably slower than to strong. This slackening of propagation speed under weak stimuli (Fig. 2) is, I would urge, a more significant difference between reflex-conduction and nerve-trunk conduction than is the mere greater slowness of the former than of the latter. Another difference between the two in regard to conduction-speed is that in the various cerebrospinal nerve-trunks of the same animal species the conduction-speed appears to be practically the same. But reflex conduction-speed as measured by the latent period differs greatly in the various type-reflexes of even one and the same limb. The latent time of the scratch-reflex is, on the average, much longer than that of the flexion-reflex or extensor-thrust, although the spatial distance of the nerve-fibre conduction is not greater. The latency of the former usually in my experience lies between 140 σ for intenser stimulation and 500 σ for weaker, and I have seen it extend to 2440 σ and even to 3540 σ. So that although a weakly provoked flexion-reflex may have a lengthier latency than a strongly provoked scratch-reflex, the latency of the scratch-reflex is nevertheless on the average very characteristically longer than that of the flexion-reflex. Now there is no evidence that this is referable to a difference in the conduction rate along the nerve-trunks of the two reflexes; indeed, the efferent nerve-trunks for the two reflexes are the same.

The speed of travel of nervous impulses along nerve-trunks is fairly known. On the not improbable assumption that their velocity along the myelinate fibres of the white tracts of the central nervous system is about the same as along the myelinate fibres of nerve-trunks, the latent period of reflex-actions of moderate intensity is obviously greater than can be accounted for by travel along such conductors of the same length as the reflex-arc itself. The delay in speed occurs whenever the impulses pass through gray matter. This has been especially clearly shown by Exner.¹⁴⁹ The delay in the gray matter may conceivably be due to slower conduction in the minute, branched, and more diffuse conducting elements — perikarya, dendrites, arborizations, etc. — found there; or it may be referable to a fresh kind of transmission coming in there, a process of transmission different in nature to conduction along nerve-fibres. The neurone itself is visibly a continuum from end to end, but continuity, as said above, fails to be demonstrable where neurone meets neurone — at the synapse. There a different kind of transmission may occur. The delay in the gray matter may be referable, therefore, to the transmission at the synapse.

And if the delay occur at the synapse, the possibility suggests itself that the time consumed in the latent period may be spent mainly in establishing active connection along the nervous-arc, which connection once established, the conduction in the arc then proceeds perhaps as speedily as does conduction in a simple nerve-trunk. The latent time would then be comparable with time spent in closing a key to complete an electric circuit or in setting points at a railway junction. The key once closed, the points once set, the transmission is as expeditious there as elsewhere. Measurements of reflex times deal customarily, so far as I am aware, with the latent time of reflexes initiated in arcs not in action at the moment when the exciting stimulus is applied to their afferent end. How the latent time is spent can receive some light from observation on the latent time of an increase of action in an arc already active in the same direction as the incremental action.

To examine this the flexion-reflex was excited by a sub-maximal stimulus, and after its appearance the intensity of the exciting stimulus was abruptly increased by short-circuiting a definite resistance from the primary circuit. The stimulus was a series of break shocks of regular interval given by a key rotating at constant speed in the primary circuit of the inductorium. An electromagnet marked the interruptions of the primary current; the electromagnet was arranged to show by more ample excursion of its armature the point of time from which onward the primary current was increased. The shocks were applied by a needle-point (kathode) to the skin of a digit: the other electrode, large and diffuse, was wrapped round a forefoot, i. e., headward of the spinal transection. In these experiments the earlier reflex elicited may be termed the initial reflex. Its sudden increase on sudden intensification of the stimulus may be termed the incremental reflex. The latent times of the initial reflex and incremental reflex, when compared, showed almost always that the latency of the latter was rather the shorter. But the difference often was not great (Figs. 3 and 4). The average for 30 of the initial reflexes was 48 σ, and for the 30 corresponding incremental reflexes was 38 σ. This difference seems too small to support the supposition that the latent time of the initial reflex is chiefly consumed by “setting” the synapse, which, once set, conducts in much the same way as regards speed of transmission as does the rest of the arc. It might be that the incremental reaction involved the “setting” of other additional synapses. But such an explanation demands that none of the augmentation occur through the synapses already in action, for the latent time is measured to the first beginning of the steplike incremental ascent of the curve.

In the incremental stage of the reaction the reflex is usually a relatively intense one. Now the length of reflex latency is caeteris paribus inversely as the intensity of the reflex. If the reflex as produced in two stages be compared with the reflex produced by delivery of the stimulus in its full strength at the outset, the latent time of this latter is found shorter than that of either the initial or incremental reaction of the other reflex (Fig. 4). The latent time under the same external stimulus is thus less in circumstances that ex hypothesi involve building a bridge and then sending impulses across it, than when the bridge already having been built the impulses have merely to pass. This argues against an amoeboid movement of the protoplasm of the cell being the step which determines its conductive communication with the next (Demoor, Cajal, Renaut, Monti, Duval, Lugaro). It also seems conclusive against any major portion of the latent period being consumed at the synapse in a process which sets the synapse ready to conduct, — a process of preparation for transmission as distinguished from a process of transmission. It argues that the delay is inherent in the process of transmission itself, and that therefore the actual nervous transmission at these points has, when the stimuli are weak, a different order of speed to that in nerve-fibres. The shorter latent time of the reflex induced by the stimulus delivered in full at the outset is in harmony with the reflex, being under that mode of excitation rather more intense (e. g., of greater amplitude) than when excited as an increment to a foregoing reflex of less strength. The shorter latent times given by intenser stimuli seem readily explicable by the minimal quantity of transmitted influence necessary to give detectible effect, being necessarily earlier reached with copious transmission than with weaker transmission.

The observations indicate, therefore, that the latent time belongs to some process which is the same in nature, both in initiating a reflex from a resting arc and in increasing a reflex through an arc already in submaximal activity, — and probably therefore in maintaining a reflex in unaltered continuance in an arc. It argues that any “setting” process in the nerve-centre, if it occur at all, is negligible in regard to the time it consumes. It suggests that even while at rest the reflex apparatus is just as prepared for immediately transmitting impulses as when actually engaged in reflex activity in the very direction the new impulses would require. It therefore suggests the greater need for active inhibition in the co-ordination of activity of arcs which have a final path in common and yet use that path to different effects. If resting paths all lie open for conduction, prevention of confusion must depend not on the path excited being the only one open for conduction, but on its excitation being accompanied by inhibition of others that, did they enter into action, would detrimentally confuse the issue of events.

Reflex after-discharge. Another characteristic difference between conduction in nerve-trunks and in reflex-arcs is the less close correspondence in the latter between moment of cessation of stimulus and moment of cessation of end-effect. The reflex-arc shows marked “after-discharge”; the nerve-trunk does not. Tetanic contraction of the knee-flexor muscles of the dog induced by brief faradization of the motor-nerve usually ceases within 150 σ of the cessation of the stimulation of the nerve, if crude condition of fatigue, etc., be avoided. The contraction of those same muscles, when induced reflexly by a similar brief stimulation, often persists for 5000 σ after cessation of the stimulus (Fig. 5).

Time in seconds below. Clonus is seen in the after-discharge. The measure of intensity of stimulus is given from the units of the Kronecker coil; the measure of the reflex is from the area included between myograph curve and base line. Reflex IV not in the figure.

We must subtract from the period of after-discharge a period equal to the latent time. But usually the latent time is quite insignificant in length as compared with the after-discharge. The after-discharge in flexion-reflex VI (Fig. 6) was more than five hundred times longer than the latent time.

The after-discharge increases with increase of intensity of the stimulus, not only absolutely, but relatively to the whole reflex — when the stimulation is not long lasting. Taking the flexion-reflex for example, the increase of after-discharge with increasing intensity of stimulation is more marked than the increase of contraction-height (Fig. 7). With stimulation lasting not more than 1000 σ the maximum amplitude of the reflex arrives, as the intensity of the stimulus is increased, later and later, so that with weaker stimuli it falls within the excitation period, but with stronger stimuli it is reached only after the application of the stimulus has ceased. Very marked in the after-discharge of this, the flexion-reflex, is a clonus (Figs. 5, 6) with a rate in my records varying between 7.5 and 12 per second. Undulations of similar kinds are clear in the reflex movement even from the outset in weak intensities of the reflex.

Figure 6.— Stimulus for each reflex was 72 break shocks at the rate of 40 per second. This stimulus is registered above by electromagnet in the primary. Abscissae on the myograph curve show its delivery and cessation on that curve. The curve marked V is the actual record of its observation; the other curves are tracings of the records obtained in the consecutive series of which curve V was the fifth member; the record is thus condensed and contrasted.

The clonus of the after-discharge is well seen. Time below in seconds.

Under short-lasting application of rather weak serial stimuli at slow frequency of repetition (break shocks at 20 per second), increase of the number of stimuli, without alteration of their intensity or rate, in other words mere prolongation of the stimulation, increases the after-discharge (Fig. 8). The reflex induced by nine stimuli has an after-discharge three times as great as the reflex from three similar stimuli. The maximum amplitude in this case may remain practically the same, the later stimuli simply prolonging the maximum without increasing it. They prolong it for a far greater time than their own delivery prolongs the stimulation time.

In the scratch-reflex, likewise, intensity increases the after-discharge (Fig. 9). Its after-discharge is rhythmic, a clonus like the rest of the reflex, with the slight lengthening of duration and sequence of the terminal beats that is characteristic of this reflex. The after-discharge from the scratch-reflex is not usually so prolonged as that from a flexion-reflex produced by a stimulus of like length and intensity (Fig. 10). Six to nine beats usually complete the after-discharge.

In the spinal dog there is a reflex of the hind limb in which a movement of extension at knee, ankle, and hip is caused by stimulation of the skin of the contralateral hind limb — the “crossed extension-reflex.” When this reflex is provoked with more than a certain intensity its after-discharge becomes a feature of extraordinary prominence, both as regards degree of contraction and duration. This after-discharge may then be more intense than any other part of the reflex, and may persist, gradually declining for 10 or 15 seconds (Fig. 27). Wundt⁶⁹ and Biedermann²⁰⁴ have noted that in the cooled frog the duration of the reflex after-discharge is prolonged.

Figure 7.— Elicited by 10 break shocks at rate of 20 per second, i.e., stimulation lasting .5″. Intensity of stimuli increased by bringing secondary coil toward primary.

The intensity of stimulus is given in units of the Kronecker inductorium. The amount of reflex is measured by the area between the myograph curve and the base line.

Time in seconds above lowest record.

There is no feature of the conduction of a reflex-arc which distinguishes its mechanism more universally from that of a mere nerve-fibre tract or trunk than lengthy after-discharge. Richet⁹⁸ has paradoxically applied to this feature the old adage modified:”Sublata causa, non tollnan effectus.” The after-discharge can, however, be cut short sharply by “inhibition”; it seems also to remanifest itself sometimes after a passing interruption by inhibition.

The long latency and the marked after-discharge of reflex-conduction easily explain a phenomenon often met when studying reflexes provoked by stimuli that are brief, especially if they

be also weak. The stimulus, though it may last for a good many sigmata, is over and past ere the reflex-response appears (Fig. 9). That response, when it appears, may nevertheless endure for 1000 σ or more. There is nothing closely comparable with this in the conduction of nerve-trunks.

The after-discharge of a reflex may be considered analogous to a positive after-image left by a visual stimulus. The analogy is suggestive in connection with others to be drawn between spinal and visual phenomena.

Conduction along reflex-arcs presents in contrast to that along nerve-trunks characters that may be figuratively described as indicating inertia and momentum. It is as though in the case of a weight to be pulled from a position of rest the tractive force were applied through a rigid rod in nerve-trunk conduction, but through a relatively yielding elastic band in reflex-arc conduction. But there are other differences between the two forms of conduction which this simple simile does not figure. We have to enter on such at our next meeting.