Physiological Foundations of Behavior

In the discussion of mosaic development in Chapter X, it was pointed out that development is not always a mosaic of independent ; or self-differentiating parts, as the earlier preformistic theories assumed, but that very different degrees of independence or dependence of parts appear, according to species, stage of development and region of body. According to the theory of qualitative nuclear division which postulated the distribution of different determinants or other hereditary entities to different nuclei in early development, development must always be a mosaic. As the advance of knowledge forced the abandonment of this theory, the theory of “formative substances” definitely localized in the cytoplasm and distributed to different cells took its place. But the attempts to interpret mosaic development in terms of such “formative substances” localized in different regions of the egg was practically abandoned when it was shown that the localization and distribution of the supposed formative substances might be very greatly altered without altering polarity or the course of development and differentiation (F. R. Lillie, 08). These facts made it clear that at least many of the supposed formative substances were first of all the products rather than causes of differentiation and that we must account for their presence rather than use them to account for development.

In general the facts show that the stability of organismic pattern in the egg at the beginning of embryonic development differs widely, not only in different species but as regards different regions of the body. While we know practically nothing as yet concerning the physico-chemical basis of this stability in any particular case, its relations with the physiological gradients are evident (Chap. X). In some protoplasms, e. g., annelids and mollusks, the gradients determine a relatively stable pattern before cleavage begins, while in others, such as the sea urchin egg, the changes brought about by the presence of a gradient or gradients are much less stable and may

even be completely obliterated experimentally (pp. 104-107). In general, also, so-calied self-differentiation and correlative differentiation are expressions of conditions existing at different levels of a physiological gradient. The high end of a gradient, the dominant region, is the most nearly self-differentiating, or independent of any body level (see Chap. X, also Child, ’21 a, Chap. VII) and the degree of dependence increases as we go down the gradient.

According to this view “mosaic’”’ development and “self-differentiation’”’ are merely consequences of stability or fixity of certain features of organismic pattern in eggs and early developmental stages of certain forms. It makes little difference whether we say that the annelid egg has developed a greater degree of specialization than the sea urchin egg along the gradient at the beginning of embryonic development, or that the constitution of the annelid protoplasm determines that the differences at different levels of the gradient when once established are much less readily reversible than in the sea urchin egg.

As long as development proceeds without extensive alteration of gradients and patterns the differentiation of different parts becomes in general increasingly independent or mosaic-like, but in the annelid the apparent mosaic character of early stages concerns the larval rather than the adult organs, and disappears more or less completely with metamorphosis and the development of the trunk region. It has long been recognized by students of evolution that in general characteristics of early phylogenetic origin are less modifiable than those of later origin. Similarly, in the development of the individual, it is evident that under the usual range of conditions, the larger more general features of organismic pattern which appear earlier in development are less modifiable than the details of pattern which appear later, and polarity and symmetry, the most general features of organismic pattern are the most stable of all. If we are able to modify polarity and symmetry we modify the whole pattern of the organism (Chaps. VIII, IX). As regards the individual, these differences in modifiability are expressions of the fact that the primary physiological gradients once established, are the most stable, the least modifiable features of organismic pattern, because such gradients are the record in protoplasm of the most generalized form of reaction which underlies all other organismic behavior. Only in

so far as the primary gradients persist in a protoplasm is any further development of organismic pattern possible. Returning again to Ameba and Fucus (pp. 57-61), we see that in Ameba each pseudopod is for the time being a gradient, or the whole body may become temporarily axiate, but the gradients constituting such patterns are completely reversible and therefore do not become the starting point of differentiation. In Fucus, on the other hand, the gradient determined by the illumination differential persists and constitutes the basis of the whole axiate pattern of the plant.

Going one step further, it is also true in general that the fixity or modifiability of the other components of organismic pattern depends upon their relation to the primary gradients. The more general features of pattern more directly and closely related to the general axial gradients are more stable than the minor features, appearing at later stages and representing more remote relations to the primary gradients. For example, the localization, course of development and general pattern of the central nervous system, all of which are very closely related to the primary gradients (Child, ’21 a) are much more stable features of pattern than the relations between different reflex arcs and neuron paths. Even in man these latter vary widely as regards different individuals and are highly modifiable. Again, the bilateral arrangement of appendages on the body is less readily modifiable than the structure and symmetry of the single appendge.

In animals we also find that the components of pattern which represent higher levels of a gradient are less readily modifiable into those representing low levels than the reverse. In Planaria, for example, it. has never been possible to transform a head into any other region of the body, but any other level may transform into a head. Somewhat similar relations exist in other animals in which isolated pieces from the lower levels give rise to new complete individuals, e. g., hydroids, many flatworms and annelids.

These examples serve to illustrate what appears to be a fact of very general, if not of universal significance in the development of organismic pattern, viz., that there is a relation between the stability of the components of organismic pattern and the metabolic rate of the region or developmental stage in which they arise. In general those pattern components determined in the earlier stages of development and at the higher levels of a gradient are more stable than those determined at later stages and lower levels. The rate of metabolism is in general higher in earlier stages of development from the begin-

ning of morphogenesis on, than in later stages (Child, ’15b) and is \ _also higher at higher levels of the gradient than at lower. In many organisms there are regressive stages in development, as in the atrophy and loss of larval organs in nemerteans, echinoderms, polychetes, insects, etc. In such cases previously existing gradients may disappear and be replaced by others and under the altered conditions the stability of pattern of the earlier stages may also disappear and new relations of stability be built up with the new gradients. Evidently the stability of any pattern component, structural or functional, can

_ persist only so long as the more general components of pattern on which its existence depends are stable. In a particular environment a certain organismic pattern may ap- - pear highly stable, but experimental changes in environment may show that even its most general features are readily modifiable by many factors. This is strikingly true in the case of the sea urchin. The normal pluteus presents a very definite, highly constant pattern, but through experimental conditions affecting the height, slope, etc., of the gradients, this pattern is modifiable to such an extent that if we did not know the origin of the forms thus produced, we should never recognize them as belonging to the same species, or in the more extreme cases, to the same class, or perhaps even the same phylum as the normal larvee (Child, ’16 d, also pp. 83, 84 ,106).

If we consider the development of the individual as a dynamic equilibration (Chap. XIII), we see that modifiability of organismic pattern follows in general the same laws as modifiability of pattern in other dynamic systems. In the flowing stream the more general, larger features of pattern are more stable than the details and the morphological features determined at a high rate of flow are more stable than those determined at a low rate of flow. Whatever future investigation may teach us concerning the particular physical and chemical conditions concerned in the development of the organism, its similarity in many respects to non-living dynamic systems ise believe, highly significant.

When we turn to phylogeny we find, in so far as we can analyze the evidence, that the same general laws hold as for the individual. As I have suggested in earlier publications (Child, ’11 f,’15b, pp. 266-270), evolution, as well as individual development, appears to be an equilibration process, that is, the course of evolution has been in general a progressive fixation or stabilization of mechanisms composing organismic pattern. We do not know whether the rate of metabolism has undergone a general decrease during evolution, as it does during in-

dividual development and senescence, but if the “biogenetic law” that ontogeny repeats phylogeny means anything, it must mean that individual development and evolution represent changes in protoplasm which are similar in general character and course, but of very different period. If the individual organism represents the behavior of a particular cell or cell mass in a particular environment, evolution represents the behavior of protoplasm in general in environment in general, or more specifically, the behavior of the primordium or substratum which we call in abstract terms the “ germ plasm,” and on which the pattern of the individual organism is superimposed. There is in fact some ground for the assertion that the pattern of the individual organism represents primarily, or in large measure, the behavior pattern of the cytoplasm, and the evolutionary pattern primarily the behavior pattern of the nucleus. To conceive evolution as behavior and the patterns of the various phyla, classes, etc., as behayior patterns, does not by any means require the adoption of Lamarckian views. The behavior of the “germ plasm” in relation to its envirenment rather than the behavior of each individual organism may be the significant factor in evolution.

At the beginning of this chapter it was noted that to the student of animal behavior modifiability usually means the possibility of alteration of behavior under identical external conditions through changes in physiological condition or state of the organism. Modification in this sense, 7. e., learning by experience, occurs when traces, records, or effects of a, previous reaction are more or less irreversible and persist after the external factor determining them has ceased to act and so become factors in altering the course of subsequent reactions.

The biologist ordinarily thinks of development as something very different from such modification of behavior by experience, but from time to time the idea that the basis of heredity and development is fundamentally similar to memory has been advanced. More than forty years ago Hering (76), in his highly suggestive paper “On Memory as a General Function of Organized Substance”’ advanced this idea in general form. Later Semon (04 and later works) developed the idea in great detail, but in Lamarckian terms. Various other suggestions along the same line have also been made. Manifestly this mnemic conception of heredity is not necessarily Lamarckian. Even if heredity is protoplasmic memory, it is not necessarily memory of

the characteristics acquired by the parent individual or earlier ancestors, but may be merely memory of the reactions of the germ plasm to its environment. The present purpose is to call attention briefly to certain physiological features of individual development as they appear from a non- Lamarckian mnemic viewpoint. If the physiological axial gradient originates primarily as a reaction to an external differential, its persistence after the external differential has ceased to act, makes it a factor in modifying all subsequent reactions of the individual. What else is the gradient determined by light in the Fucus egg than a physiological memory of the differential exposure to light? And is it not true in a general physiological sense that through the experience of differential exposure to light the organism of Fucus learns to behave as an axiate organism? Ameba, on the other hand, may also learn to behave in an axiate manner (Figs. 13, 14), but here the protoplasmic memory is short and the axiate pattern usually persists only a short time.

Not only the primary axial gradients, but their later alterations and the new gradients of parts, organs or cells, may also be regarded as protoplasmic memories of er-vironmental differentials of some sort. Viewed in this way the whole course of development is a process of physiological learning, beginning with the simple experience of differential exposure to an external factor, and undergoing one modification after another, as new experiences in the life of the organism or of its parts in relation to each other occur. Memory and learning in the narrower, psychological sense represent that part of the general developmental learning process which concerns the minute pattern of certain regions of the nervous system in advanced stages of development, particularly in the higher animals. There is no evidence of any fundamental physiological difference between the general protoplasmic memory as expressed in physiological gradients and their effects and the higher forms of memory characteristic of the central nervous system.

To sum up: the development of the individual may be regarded as the expression of a general protoplasmic memory, and experiment shows us that developmental behavior is modifiable by experience. Such modification or learning is going on at all times from the differential exposure which determines the primary axial gradient, to the end of life. To the objection that interpretation of a series of phenomena such as development in terms of a less known series serves no useful purpose, it may be replied that, although we know more in cer-

tain respects about individual development than about memory in the strict sense, yet in certain other respects we know more about memory than we do about development. To interpret development in terms of memory and learning is simply to identify certain characteristics of development with certain characteristics of memory and learning, and such identification is one step in a process of synthesis. If we accept the viewpoint of preceding chapters that organismic pattern is a behavior pattern from the beginning, it is evident that the behavior of an organism in the narrower sense, as comprising fundamentally the excito-motor and closely related reactions, constitutes the most highly integrated, the most highly regulatory and in general the most highly reversible expressions of the pattern in relation to environmental factors. Some fifteen years ago Jennings (06) suggested that the laws of this behavior are the same as those governing the other aspects of life, and the present book is in large measure an attempt to show that the experimental investigation of recent years supports and confirms this conclusion. !

All processes of equilibration or regulation are expressions of the integrative aspects of pattern, that is, of the correlative factors which make the organism physiologically a whole. But in excito-motor behavior the integrative aspects appear more clearly than in any other activity of life. Nowhere is the wholeness of the organism so conspicuous as in this field. The relative fixity or modifiability of such behavior must therefore depend in larger measure upon the fixity or modifiability of the integrative mechanisms of organismic pattern than upon any other feature of pattern. Moreover, the dynamic integrative or correlative mechanisms are much more important in this behavior than the transportative mechanisms (see Chap. V). Of the dynamic mechanisms, those of excitation and transmission are the most effective means of integration.

Modification of excito-motor behavior must mean change of some sort, either temporary or permanent, in the integration pattern and particularly in the integration of excitation and transmission. Such change may be determined in various ways, not only by preceding 1 See also Herrick’s discussion of modifiability of behavior from the neurological viewpoint; Herrick, ’24, Chaps. DID. DLO. excitation-transmission patterns, but also by change in the chemjcal or transportative factors of correlation, e. g., in amount or character of the internal secretion of the thyroid, the gonads, ete., by the complex changes associated with the satisfaction of hunger by — food, or of thirst by water, by fatigue, bacterial intoxication and many other conditions. Even in such cases, however, the mechanisms of excitation and transmission are involved, for it is the effect of the chemical changes on these mechanisms which brings about the modification of excito-motor behavior. Frequently also, the chemical change is the result of a preceding modification of the excitation-transmission pattern. For example, the sight or smell of food determines a modification of behavior, viz., feeding, and fatigue is the result of other preceding modifications, usually in large part excito-motor. With the progress of differentiation and the increase in complexity of the organism, both in individual development and in evolution, the possibilities of physiological correlation increase and with such increase the possibilities of modification of correlative factors become greater, even though the general morphological and functional pattern of the organism is relatively very stable. Because of the physiological character and role of excitation and transmission in organisms, the possibilities of modification are greater for these than for any other correlative factors and, the higher the development of the nervous system, the greater these possibilities, From these facts it is evident that at least certain features of organismic behavior pattern may become more varied, more plastic and more modifiable, both in the course of individual development and of evolution, even though the general structural and functional pattern becomes at the same time more and more stable.

Students of behavior, particularly of animal behavior, have usually drawn a more or less sharp distinction between fixed and modifiable forms or patterns of behavior in the individual organism. This distinction is useful but somewhat arbitrary, since all degrees of fixity and modifiability of behavior exist and these characteristics also depend upon various conditions, The fixed type is often said to be innate or inherited. This does not mean that the behavior pattern itself persists through reproduction, but merely that it represents a potentiality of the protoplasmic constitution which under normal conditions (see Chap. XIII) js always realized as a feature of the organismic pattern. All the modifications of behavior in the indi-

- vidual are just as truly potentialities of the protoplasmic action system, 7. e., Just as truly hereditary, as the fixed behavior patterns, but they are not realized in all individuals of the species. For the realization of a particular modification, the coincidence of a particular physiological state and a particular complex of environmental factors is necessary, and the degree of specificity of the physiological state and of the environmental complex differs widely, both as regards different organisms and different reactions in the same organism. In Amaba and Paramecium, for example, certain modifi-

_ cations of behavior occur whenever an exciting factor acts for a certain length of time. Similarly, in man certain modifications of behavior appear under almost any sufficiently intense or long continued excitation. In Ameba and Paramecium, however, the modifications of behavior are very few and simple and general in character, 7. e., essentially the same for a wide range of environmental complexes, while in man modifiability is practically unlimited. Some modifications of behavior in man are features of daily life, but others may be realized only once in a lifetime or only once in many generations.

In the preceding chapter it was noted that fixity and modifiability of the developmental and morphological aspects of organismic pattern depend upon their relation to the fundamental factors of the pattern. The more general, more primitive features of pattern, those which are more directly related to the primary gradients, are the more fixed and stable, while the details, the features of later origin, are more modifiable. If it is true that excito-motor behavior is merely the most conspicuous and highly integrated expression of organismic pattern and that the same laws hold for it as for the other features of this pattern, we must expect to find a similar relation between fixity and modifiability of excito-motor mechanisms and components and the fundamental or primary factors of pattern. That is to say, the mechanisms of the more firmly fixed or stable behavior patterns must be more closely or directly associated with the primary factors of organismic pattern, the physiological gradients, than the mechanisms of the more highly modifiable patterns.

In one sense the mechanisms of excito-motor behavior might be said to comprise all the mechanisms of the body, but the special mechanism of such behavior is transmissive correlation. In the development and evolution of the organism, we see transmissive correlation first becoming more or less fixed as physiological gradients, then undergoing further structural fixation with the origin of the nervous system. In the higher animals and man the individual nervous mechanisms, the neurons, the reflex arcs, possess a high degree of fixity both morphologically and physiologically, and modifiability is a matter of the integration or linking up of particular mechanisms into a particular excitation-transmission pattern. The complexity of nervous structure, the specialization of excitatory and transmis- Sive processes, the significance of physiological condition of the synapse or of other parts of the neuron in determining the course of transmission, all make possible an increase in the modifiability of excito-motor patterns, with an increasing morphological fixity of the mechanisms concerned (Herrick, ’24, Chaps. XVII, XTX) in fact the excito-motor patterns of the higher animals and man are physiologically less dependent upon the fundamental factors of organismic pattern than any other feature of the individual. That this is true will appear more clearly in the following sections,

The mechanisms of excitation and transmission, with which we are primarily concerned in the behavior of the organism as a whole, appear in various forms. The simplest and most primitive is apparently the excitation-transmission gradient in protoplasm (see Chap. XI). This in its simplest form is temporary and evanescent: it may arise at any point in relation to local or differential excitation and may soon disappear and be replaced by another. In an organism with gradient patterns of this temporary and shifting character the gradients themselves, the fundamental factors of axiate pattern, are directly concerned in the modifications of behavior.

In the case of Ameba, for example, motor behavior is a direct expression of physiological gradients. Each pseudopod is for the time being such a gradient (Hyman, ’17) and in those cases in which the whole body is temporarily a single gradient, the Ameba is temporarily an axiate animal (pp. 57-59). We do not hesitate to call these different patterns of Ameba behavior patterns, but the gradients of the pseudopods or of the whole body differ from the physiological axes of axiate animals only in their instability. The motor behavior pattern of Ameba is in fact a gradient pattern and all modifications of motor behavior, e. g., localization and development or reduction and disappearance of single pseudopods, as well as the initiation, change in direction, or cessation of directed locomotion of the whole animal represent modifications of the gradient pattern.

Actually Ameba constitutes a direct demonstration of the fundamen- _ tal identity of axiate pattern and behavior pattern. Axiate pattern in Ameba is a behavior pattern in the strict sense, arising in response to the differential action of external factors, and rapidly reversible. Conversely the behavior of Ameba always involves the develop- - ment, persistence or disappearance of axiate pattern, and the only 4 modifications of behavior possible are made up of modifications in