Physiological Foundations of Behavior

the localization, length, steepness and persistence of the gradients. The development of a persistent axial gradient, a physiological polarity and symmetry, even in the absence of a definite nervous _ system, determines a relation of dominance and subordination (Chaps. X, XII), and the organism behaves primarily as an axiate organism, ~ although surface-interior pattern may still play a part even in axiate organisms, in determining behavior. As long as axial gradients persist they constitute a relatively stable basis for behavior and we find that the most firmly fixed elements or types of behavior pattern are very directly associated with such gradients. For example, the precedence of the apical end or head in locomotion is a direct expression of an axial gradient. It has been determined experimentally that when the apico-basal gradient is obliterated through differential susceptibility, e. g., in the sea urchin blastulz (Child, ’16 d), or in hydroid planule (see pp. 83, 106), definite orientation of the body is no longer possible. Under such experimental conditions the fundamental mechanism or pattern on which this highly stable feature of axiate

behavior, viz., precedence of the apical end, depends, has been destroyed. Similarly, the integration or codrdination of bilateral or radial organs is very closely associated with the axial gradients and disappears when they are obliterated. In axiate organisms modifications of behavior occur in general with reference to the axiate pattern, that is, they are modifications of axiate behavior. To take a few simple illustrations, local excitation of some part may affect the speed of definitely directed locomotion, e. g., in a creeping earthworm. Unequal excitation of the two sides of the body may partially or wholly obliterate for the moment the symmetry pattern and induce asymmetr ic behavior as in the case of a tropistic orientation according to Loeb (see pp. 230-231). Again, in the quiescent condition basal or posterior regions of the body may be for the time being more or less physiologically isolated and show a greater or less degree of independence, while excitation of the dominant region may bring them under control. Such relations appear in various flatworms and oligochetes (see pp. 153-156).

With the development of the nervous system in animals the axiate — character of the behavior pattern becomes in general still more stable. The very close physiological relations between the central nervous system and the axial gradients have been considered elsewhere (Child, 21a). Reflex arcs arise with definite, functionally irreversible paths, all related in some definite way to the fundamental axiate pattern and there are many facts which indicate that a physiological gradient is the foundation on which each reflex arc develops (Child, ’21 a, Chap. XIII). Evidently the actual development of a reflex arc is possible only when the gradient is to a certain degree fixed and stable. In fact, the single reflex arc, so far as such a thing exists, is to a high degree fixed and stereotyped in structure and function because it is an expression of the more general and more stable factors in organismic pattern and particularly in the pattern of the nervous system (Herrick, ’24, Chaps. VIII, LS XeVvIbye

In the actual behavior of organisms, however, reflex ares do not function singly and separately, but there is always some degree of integration, and it is in this integration of reflex arcs that the modifiability chiefly appears. From the appearence of the nervous System on through evolution and development the modifiability of behavior depends primarily upon the integration pattern in the nervous system and the changes of which it is capable. The structure of the neuron, its sensitiveness and delicacy of reaction during development to certain factors in its environment, perhaps electrical (Child, ’21 a, Chap. XI), the alterations in physiological state of the synapse and doubtless also of dendrites and other portions of the cell body, and perhaps of axons, provide a basis for almost infinite complexity, variety and modifiability of excito-motor behavior,

Certain features of nervous function and structure are not only of fundamental importance for the modifiability of behavior in the higher animals and man, but are of very great interest, both from the physiological and the historical viewpoint. The invertebrate nervous system, even in its highest development in the insects, does not permit any very high degree of modifiability of behavior, and we must interpret the fixity of behavior in these forms as meaning that, even in their minuter details, nervous structure and functional correlation are determined with a high degree of fixity in individual development in relation to the general pattern of the species, 7. €., in the

final analysis the physiological gradients and their modifications in that particular protoplasm. (Herrick, ’24, Chap. XI). In the invertebrate the axiate pattern appears to be predominant throughout the nervous system and the behavior therefore consists of relatively rigid reflexes and reflex complexes. In the course of vertebrate evolution, however, a new factor in behavior makes its appearance (Herrick, ’24, Chapters XVII-XXI). A portion of the central nervous system becomes, so to speak, emancipated, or physiologically isolated to a greater or less degree, from the relatively rigid determining action of the general axial gradients in the body. This portion is, or becomes the cerebral cortex. In the arrangement of its neurons the cortex shows predominantly a surfaceinterior pattern with only slight and very general axiate features. The consequence of this pattern is that structural and functional relations are physiologically determined to a large extent by local factors which vary from moment to moment. Histological investigation has shown that the morphological connections of one region of the cortex with others are exceedingly complex and not the more or less uniformly directed connections characteristic of axiate pattern. - The cortex seems, in fact, to be to some extent what we may call a superaxiate region. It arises from the higher levels of the chief body gradient, 7. e., broadly speaking, from the dominant region of the individual, and as it is superaxiate in position and structure, so it is superdominant in function. Judging from behavior, the relations of dominance and subordination within the cortex are not definite and fixed, but shift from moment to moment, according to the impulses coming in and the physiological state of the various cells.

According to this view, the cortex really represents a new form or phase of organismic integration. The gradient pattern is fundamentally an autocratic pattern. The dominant region controls other regions but is relatively independent of them. In the higher vertebrates, however, the development of excitation and transmission in the nervous system have made it possible for the lower levels of a eradient to affect the dominant region and the origin and development of the cortex are undoubtedly associated in some way with this situation. The cortex functions as a deliberative assembly, a parliament in a representative form of goverment. That is to say, the efferent impulses finally emerging from it are determined, not simply by its immediate relations to environment, as in the dominant region of a physiological gradient, but rather by the sum total of afferent impulses reaching it from lower centers and through these from all

parts of the body, and by its own physiological state as determined — by past activities and their records in the cells as functional alteration, memory, etc. (cf. Herrick, ’24 at the end of Chap. XVII). We may say then that in the higher vertebrates the gradient pattern of organization is undergoing modification from the autocratic form of dominance or control characteristic of the simple gradient toward the democratic form with representative government, with all its plasticity, modifiability and uncertainty of result in any particular case (see Chap. XVII). In this development of the cortex as a superaxiate pattern with an inconceivable number of individual neuron axes, but only slightly developed general axiation, lies the physiological and morphological foundation of the modifiability of excitomotor behavior in the higher vertebrates and man. The physiological problem of the origin of the cortex has been touched upon elsewhere (Child, ’21 a, pp. 261-7), and since at present nothing more than speculation along these lines is possible, no further discussion is necessary here.

Thus far we have been concerned chiefly with the mechanisms and components of behavior of excito-motor character, rather than with the integrated behavior of the organism as a whole. Any particular type of reaction of an organism to an external factor involves the integration of various mechanisms, and the character of the reaction as a whole depends on the integration pattern in each particular case. The question is, then, to what extent and how is the integration pattern of the various forms or types of behavior modifiable? That this pattern is modifiable through differences in the direct action of external factors is demonstrated by universal everyday experience. The excito-motor pattern determined by the sight of food in a hungry dog is different from that determined by the sight of his master or of a stranger. The behaviorist usually takes such modifiability for granted and concerns himself more particularly with modifiability of the behavior pattern through internal changes without change in the external factor.

When, for example, an individual organism changes its reaction to an external factor acting continuously without change, or when the separate reactions to interrupted or repeated identical actions of an external factor differ, or finally, when different individuals of the same species behave differently with respect to the same external factor, we say that the behavior is modifiable. Such modifiability the physiological state of the organism. The occurrence of modifica-

tion of behavior in this way in man and the higher animals has long been a familiar fact, and the studies of Jennings and others have made it clear that in the simpler forms, even in the protozoa, some degree of such modifiability exists.! At present, however, we are primarily concerned, not with a general discussion, but rather with the question of the relation of this sort of modifiability and of physiological state on which it depends, to organismic pattern in general.

Physiological state.— Physiological state represents the sum total of the physiological factors which determine the excito-motor integration pattern at any particular moment. As Jennings has pointed out, many factors may be concerned in determining physiological state. Among such factors, for example, the persistent effects of previous reactions, such as altered irritability, morphological alteration in pattern, memory, etc., the effect of different excitations occurring simultaneously, metabolic condition as determined by nutrition, respiratory, endocrine and other factors, physiological age and other periodicities, in short all factors, external or internal, which affect the organism.

In the simpler organisms the range of changes in physiological state is relatively narrow, because the instability, 7. e., the high degree of reversibility of the protoplasmic substratum permits only a relatively slight chemical and morphological differentiation representing the more general and constant factors in developmental behavior. Both transportative and dynamic correlation must remain simple in such organisms, and the transitory excito-motor reactions leave little or no persistent effect or record upon the protoplasm, or one which is soon obliterated by the reversal of the changes determining it. Memory in such organisms is, in short, a generalized protoplasmic function and but slightly developed. The excito-motor behavior of these organisms must of necessity be relatively simple and stereotyped in character, though, as will appear below, a certain degree of modifiability of behavior in relation to physiological state is possible, even in such forms.

As chemical and structural differentiation become more stable and general organismic pattern therefore becomes more complex, the possibilities of modification of physiological state increase. Particularly with respect to the nervous system is this the case. The mor- 1See for example Jennings, ’06, particularly Chaps. X-—XII, XVI-XVIII, and numerous references there given; also later works on animal behavior. phological pattern of the nervous system, from the general anatomical features down to the hypothetical structural substratum of memory, constitutes the most stable and least reversible structure in the body. At the same time, the large number of neurons, the branching of dentrite and axon, making possible great complexity of relation between neurons, the relations of nervous organs to environment and to other organs, the complexity of the processes of excitation and transmission, their sensitiveness to mechanical, thermal, chemical and electrical factors, all these, as well as chemical correlative factors, hormones, products of metabolism of the organism, of bacteria, etc., play parts in determining that physiological state in the nervous system is more labile than in any other organ. For similar reasons physiological state in the cerebral cortex is more labile than in any other part of the nervous system. :

Summing up, we may say that physiological state at any given moment represents the process or stage of regulation or equilibration or “adjustment of internal relations to external relations” which is occurring at that particular moment. It is, in fact, nothing else than life in a particular individual at a particular time. If this is true, excitomotor behavior, as the most highly integrated expression of organismic pattern, is the most adequate and complete expression of the process of living in a protoplasm. As regards this point, this general physiological conception leads us to the same conclusions as our observations of, and relations with our fellow-men and animals.

Trial and error behavior. — Modifiability of physiological state is an essential factor in trial and error behavior. It is the modification of physiological state that determines the successive changes in reaction from which selection is made. According to Jennings, the changes in physiological state result from the continuance of the disturbance and must go on until some reaction determined by them brings relief, or until the disturbing factor ceases to act. The trial form of behavior occurs when the pattern of the organism does not include any behavior mechanism which gives directly an effective reaction in the Special case in question. As expressions of modifications of physiological state the successive changes in reaction really represent changes in the integration pattern and the successful trial constitutes the effective integration. It is often possible to observe the progressive development of the new integration. In the anaxiate Ameba, for example, the reaction to an external factor may consist first in the formation of several pseudopodia, often in various directions, but with continuation of the external action the animal may develop

temporarily a more or less clearly marked axiate pattern which has a definite relation to the external factor (Figs. 12-15); see also Jennings, 04, ’06, Chap. I). This relation is such that the Ameba finally moves more or less directly away from — or toward — the source of disturbance. In such cases the change in physiological state is chiefly or wholly the progressive integration toward an axiate pattern which is more or less definitely referable to the localization or direction of action of the external factor.

In axiate forms the axiate pattern is of course already present, but adjustment to the special case may be necessary. In the case of Paramecium the axiate pattern does not enable the animal to orient itself directly and at once with reference to the external factor acting upon it (Jennings, 06, Chaps. III-VI). It can only repeat the characteristic avoiding reaction, consisting of backward movement, for a short time, followed by forward movement at an angle to the original direction (Fig. 146), until an adjustment of the pattern to the particular case occurs, 7. e., the reacticn is effective, or until fatigue, death, cessation of the exciting action or some other change terminates the reaction.

Certain observations indicate, however, that even in Paramecium a modification of the avoiding reaction through change in physiological state by repetition may occur under certain conditions, 2. @., Paramecium may learn by experience (Stevenson Smith, ’08; Day and Bentley, ’11). According to these authors a Paramecium confined in a capillary tube so narrow that the characteristic spiral course cannot be followed in swimming, advances in the tube with a rotational movement about the long axis and when the end of the tube is reached, reverses the direction of this movement several times. Thus far the behavior is of the usual sort, altered only by the limited

diameter of the tube. But it is not effective in bringing about adjustment and sooner or later the animal turns itself about in the tube by a new sort of reaction, consisting of a bending of the oral end, followed by sudden jerking movements (Fig. 147). Since this sort of reaction is unknown in unconfined locomotion, it is evident that the Fic. 147.—Modification of behavior by experience. q of Paramecium in eapillary tube (from The flatworm, Planaria doroto-

200 Sanit, 08), cephala, in reacting to a chemical factor, e. g., extractives diffusing from food, usually lifts its head and moves it from side to side before orientation of the direction of locomotion with reference to the diffusion gradient occurs. And even when it has begun to advance, it may alter direction more than once before the reaction becomes effective, that is, before the adjustment, the orientation, is accomplished. In trial reactions in the higher animals and man the progressive integration with reference to the particular case is often very clearly seen. The reaction of the decapitated frog to a drop of acid placed on one thigh, when the other leg is held, is a trial reaction experimentally induced. That is, the effective mechanism is not absent but is prevented from acting and with its failure the behavior becomes essentially a matter of trial with a progressive integration of the reflex ares of the body with special reference to the local excitation. Evidently the excitation, at first local in the nervous system, spreads, involving one region after another until the whole body is more or less integrated, whether effectively or not. In trial behavior in man the progressive development of the new integration appears even more clearly. In attempts to learn a new complex coérdination, e. g., riding a bicycle, skating, the first failures lead to greater efforts, the use of other muscles, changes in posture, etc., and gradually the new behavior pattern develops out of the complex.

As Jennings has shown, such new integration patterns may often persist for a time even in unicellular animals, but usually disappear rapidly in these and in the simpler multicellular animals. The “ physiological memory” of these organisms is rudimentary, the axial and symmetry gradients being the chief records and even these are not very stable. Apparently the possibilities of development of persistent records in the protoplasm reach or approach their limits with the differentiation in relation to these gradients, andthe effects of new differentials and integrations are either rapidly reversible, or, if the

action of the external factor is sufficient in degree or duration, they may obliterate the old gradients and determine new ones. In such forms the ability to learn, to acquire habits, goes but little beyond the most general habit of all, 7. e., the axial gradient, and more or less differentiation in relation to it. With the increase in stability of protoplasmic records of behavior which is an essential feature of evolution, the possibility of persistence of the new effective integration in a trial reaction increases,

“4. e., the ability to learn and acquire new habits increases. The gradual recording in the protoplasm of the new integration corresponds to Jennings’ law of resolution of physiological states: The resolution of one physiological state into another becomes easter and more rapid after it has taken place a number of times (Jennings, 06, p. 291). The simpler organisms are what they are because the protoplasmic records of behavior reactions and integrations are to a large extent readily reversible and the more persistent features of such records are of very general character and, so far as organismic integration is concerned, rudimentary. In such forms excito-motor behavior must consist primarily of few and simple stereotyped components and the trial form of reaction must be predominant, with rudimentary and short-lived memory. Modifiability of pattern, whether morphological or excito-motor, by the persistence of protoplasmic records, 7. e., by learning, must therefore also be slight and rudimentary.

Tropisms. — The tropism in Loeb’s sense (pp. 230-231) postulates the presence of an effective integration pattern, a pattern which gives the effective reaction directly. According to Jennings’ conception of a tropism (Jennings, 06, Chap. XIV, ’08, ’09, 710), such a pattern is not necessarily predetermined or inherited, but may have developed in the individual as the result of suecessful trial behavior and the tropism may represent the working of the new integration thus developed. Undoubtedly some reaction patterns which approach more or less closely the form of a tropism are characteristic features of the normal organismic pattern of the species and definitely related to the axial gradients, and are therefore what we commonly distinguish as inherited from the patterns acquired by the individual during its life. Other patterns of this sort are doubtless developed in the individual on a trial foundation. The important point for present purposes is that the tropism requires the existence of a behavior mechanism or pattern already adjusted to the particular case. The orientation of the body of a swimming crustacean

or a flying insect with reference to light requires a complex polarbilateral reflex, or excito-motor mechanism of some sort. Granting this, we must conclude that tropisms are not the most general and primitive form of excito-motor behavior, or else we must follow Loeb in assuming that such a mechanism exists in all motile organisms, but the evidence is strongly against such an assumption. The tropism in motile organisms obviously postulates a relatively high degree of organismic integration and the evidence does not indicate that such a degree of integration exists in the simpler organisms.

From its very definite and direct relations to polar-symmetrical pattern in general we should expect the tropism to be a relatively stable or fixed excito-motor pattern, particularly in those cases in which it is a normal characteristic of the species and not acquired by the individual. So far as such tropisms actually exist, their stability is evident. Modification in minor respects, e. g., in latent period, speed of reaction, etc., may result from alterations in physiological state, but the tropism as a whole remains a relatively fixed pattern. The reversal of direction of a tropic reaction does not represent a modification in the sense of a new integration, but may result from the change of excitation to inhibition, or vice versa, with quantitative change in the external factor, or in the critical point between excitation and inhibition in the physiological mechanism. The tropism as Loeb conceives it, affords no room for memory or learning. The mechanism must be stable, essentially unalterable in the individual and polar-symmetrical in character. If pure tropisms in this strict sense exist, they must obviously be very direct expressions of the fundamental features of organismic pattern, the physiological gradients, and are modifiable in any fundamental way only as these are so modifiable.

Instinctive behavior. — The instinctive forms of excito-motor behavior likewise represent highly stable, in common biological parlance, hereditary patterns (Herrick, '24, Chap. XI). These behavior patterns are among those determined in the individual by the hereditary protoplasmic constitution and the normal or standard range of environmental factors (pp. 211, 222). ‘They represent complex integrations of significance in the life of the individual and so apparently adaptive. In general, however, this instinctive pattern remains merely a potentiality in the behavior of the individual until a particular physiological state arises, either by the attainment of a certain stage of development or by a particular configuration of

environmental factors, and determines its realization as an actual ~ reaction pattern. In most, if not in all cases, the environmental configuration plays some part in initiating the actual integration of the instinct pattern into a reaction. The sensory impression aroused in the well-fed squirrel by the nut apparently sets going the storage instinct and it may be that other factors effective at a certain season of the year start an instinctive search for nuts. The sight, sound or odor of the opposite sex initiates the integration of the sex instinct pattern and so on. In short, the component mechanisms of instinctive behavior are normal features of organismic pattern, but there is no reason to believe that their integration into a particular excito-motor behavior pattern occurs autonomously or spontaneously. Apparently, like other behavior patterns, it requires for realization the action of an environmental factor—in the case of instinct usually a more or less specific complex. As in the case of other features of organismic pattern, even the primary axial gradients or a plasma membrane, the potentiality of the pattern is given in the hereditary constitution of the protoplasm, but the action of some external factor or complex is necessary for its realization as an actual pattern in an individual organism.

Intelligent behavior. — The instinct, as a relatively stable pattern, is unquestionably represented by a more or less definite, relatively stable structural basis of some sort in the nervous system, but the integration of this pattern is to some extent modifiable, chiefly through conscious intelligence. The degree of modifiability of instinct varies in general with the development and dominance of intelligent behavior in the individual. In the insects excito-motor behavior is, so far as we can judge, very largely instinctive and based on the mechanisms of a polar-bilateral nervous system. Since the polar-bilateral pattern is highly stable in these forms, even those minuter details of it which represent the mechanisms of instinct are relatively fixed. In the higher vertebrates, on the other hand, the development of the cerebral cortex represents a considerable degree of physiological isolation from, or obliteration of, the general axial gradients in a part of the nervous system (pp. 257-258). In consequence of this unique position of the cortex, it constitutes a mechanism for highly plastic and modifiable excito-motor behavior patterns. The absence of a fixed dominance in any one region provides a basis for patterns of the deliberative, intelligent type. And finally, the development of the cortex from the anterior region of the nervous system makes it a region of superdominance. Even many