Physiological Foundations of Behavior

and of the four regions composing it is fairly constant (Fig. 67). Un- der conditions accelerating metabolism it is longer (Fig. 68) and under depressing conditions it is shorter (Fig. 69). In the reconstitution of pieces of Planaria dorotocephala into new individuals localization of parts can be similarly controlled. At a Vics. 70-75.—Size of head, length of prepharyngeal region and position and length of pharynx in reconstitution of Planaria dorotocephala under different conditions: Fig. 70, outline indicating level of section; Fig. 71, reconstitution under standard ' laboratory conditions; Figs. 72-74, different positions and lengths of regions and organs with different degrees of inhibition; Fig. 75, reconstitution with high metabolic rate at high temperature (from Child, ’15 c).

temperature of 20° in well-aérated water the isolated postpharyngeal region of Planaria— the portion posterior to the transverse line in Figure 70 — gives rise to a new animal like Figure 71, the pharynx being somewhat anterior to the middle, the region anterior to it differentiating as a prepharyngeal region and that posterior remaining practically unchanged. Under depressing conditions the head is smaller, the new prepharyngeal region shorter and the pharynx arises nearer the head, the difference from normal increasing with the degree of depression (Figs. 72, 73). Under a sufficient degree of depression reconstitution is practically limited to development of a minute head, almost no traces of reorganization of the anterior region of the piece into a prepharyngeal region occurring and the pharynx being entirely absent (Fig. 74). Changes in physiological condition in the opposite direction determine a longer prepharyngeal region and localization of the pharynx at a more posterior level (Fig. 75). Similar alterations in localization of the pharynx have been observed in a polyclad according as the cephalic ganglia are present or absent. Among the polyclads pieces do not give rise to new heads except when some portion of the cephalic ganglia, or in some cases, of parts of the nerve cords near them are present. When the pieces remain headless the new pharynx arises nearer the anterior end than when a head is present or develops (Child, ’05).

Another line of evidence appears in the modification of development by means of differential susceptibility. Attention was briefly called to such modifications in connection with the discussion of susceptibility in the preceding chapter, but it is necessary to consider these modifications somewhat more at length at this point in their relation to localization and differentiation of parts. Taking up first the question of obliteration of gradients or axes, it is evident that if the gradient represents the axis, obliteration of the gradient ought to result in obliteration of axial localization and differentiation. The obliteration of a gradient through differential susceptibility results from the fact that the levels of higher metabolic rate in a gradient are more susceptible and are therefore more depressed by inhibiting agents than the levels of lower rate. Consequently the gradient becomes less steep and with a certain degree of differential inhibition its steepness may attain zero, 2. é., it may be obliterated. When a gradient is obliterated in this way in early stages of development, we find that further development shows no indication of the order corresponding to that particular gradient.

The obliteration of the apico-basal gradient in the hydroid larva was mentioned in the preceding chapter (pp. 82, 83, Figs. 45-47). After obliteration the larve are apolar so far as can be determined, both as regards the existence of physiological gradients and the capacity for [4 axiate development and axiate motor behavior. They remain spherical indefinitely, even when returned to water, unless they are sub- Fras. 76-81.—Obliteration of bilaterality in larval development of starfish, Asterias forbesti: Fig. 76, normal larva, oral view; Fig. 77, normal larva, side view; Figs. 78, 79, radially symmetrical, Figs. 80, 81, slightly bilateral modifications of development in KCN, m/200000. The ciliated bands are indicated by shading, the entoderm is drawn in broken lines, the ccelom sacs, if separated from entoderm, in dotted lines.

jected to a new differential in environmental conditions sufficient to determine a new gradient, or gradients. When this occurs, axiate development proceeds again, but the new axis or axes may arise In any direction with respect to the original (see pp. 123-125). Obliteration of bilaterality in the starfish larva is readily brought about by the same methods. The normal larva of Asterias is distinctly bilaterally symmetrical in form (Figs. 76, 77), but when development from the eight, or the sixteen cell stage on takes place in KCN, m/200,000 at temperatures ranging from 20° to 24° the larve show various degrees of modification resulting from differential inhibition. A large percentage of these larve are either completely radially symmetrical in structure (Figs. 78, 79) or show some slight indications of bilaterality in the position of the ciliated bands (Figs. 80, 81)

seal (oe ees 8&3 Fie. 83.—Sea urchin larva from 82 which all visible indications of Fic. 82.—Radially symmetrical polarity and symmetry have been frog embryo: a differential inhibiobliterated through differential intion resulting from exposure of unhibition by KCN (many other segmented egg to LiCl m/10.62 agents give same result): (ec), ectofor 76 hours; then 20 hours in derm; (en), entoderm; mesenchyme water; (bp), blastopore lip (from cells scattered between ectoderm

or other organs. Larvee of the sea urchin, Arbacia, in which bilaterality has been completely obliterated, so far as distinguishable structural and physiological differences are concerned, have been described elsewhere (Child, 16d). With the same methods Bellamy has almost or completely obliterated bilaterality in amphibian embryos (Fig. 82). By means of somewhat more extreme inhibiting conditions it is possible to obliterate completely, not only the bilaterality, but the polarity of echinoderm embryos. In the case of the sea, urchin such forms lose their definitely directed movement and remain spherical in form indefinitely. Structurally they consist of a spherical ectodermal vesicle containing a spherical entodermal vesicle, and in the cayity between these a few mesenchyme cells (Fig. 83). Even when they

are returned to water, such Jarvee may live for weeks without further development. Moreover, susceptibility tests show that the polar susceptibility gradient characteristic of normal larve is completely absent. In these cases, then, all the gradients have been obliterated and with their disappearance, the capacity for axiate development and differentiation has also disappeared, and the animal no longer Fias. 84-88.—The head forms appearing in the reconstitution of pieces of Planaria dorotocephala: Fig. 84, normal head; Fig. 85, teratophthalmic head, development of median region inhibited; Fig. 86, various degrees of teratophthalmia; Fig. 87A—D, teratomorphic heads, showing the range of forms. The degree of medio-lateral, differential inhibition is always greater than in teratophthalmia; Fig. 88A—D, anophthalmic heads representing a still greater degree of medio-lateral differential inhibition.

shows even axiate motor behavior, but merely rolls about in one direction or another indefinitely. Dr. J. W. MacArthur permits me to state that he has produced by the same methods apolar starfish larve similar to those of the sea urchin (Fig. 83). With less extreme degrees of differential inhibition acting at early stages of development, it is possible to inhibit more or less completely the development of higher levels of a gradient, while less susceptible levels proceed more or less rapidly. In this way proportions may be

altered in such manner that the parts representing the higher levels of a gradient are reduced relatively to those representing the lower levels, or even completely inhibited in development. Differential inhibitions of this sort in the sea urchin larva were briefly described and figured in the preceding chapter as examples of differential susceptibility (pp.83-94, Figs. 48-52). With the same methods and a wide range of agents, Dr. MacArthur has obtained similar modifications in the starfish and various other echinoderms, and Bellamy (19, ’22) has described numerous cases in the frog which illustrate the same principle.

‘ Some of the most interesting modifications resulting from differential inhibition concern bilateral organs. In most if not all bilateral animals metabolic rate and susceptibility decrease primarily from the median ventral, or the median dorsal, region laterally (see pp. 81, 92). Fie. 89.—Cyclopia in frog, resulting from differential inhibition by treatment with LiCl, m/7 for three hours in early gastrula stage. The single eye is in the median plane beneath the surface and is seen through the body wall (from Bellamy, ’19).

Consequently an inhibiting agent inhibits the median region more than the lateral, and bilateral organs may be approximated to the median line. The different forms of head in the regulatory development of Planaria dorotocephala (Figs. 84-88) can all be experimentally determined in this way and represent different degrees of differential inhibition in the medio-lateral axis (Child, ’16 b, ’20 a, ’21 ¢; Buchanan, 22; Sivickis, ’23). Under proper conditions forms representing differential acclimation and differential recovery can also be obtained (Child, ’21c). Similar medio-lateral differential modifications in the larval development of the sea urchin have already been noted (Chap. VII, Figs. 51,52). Cyclopia in fishes (Stockard, ’07, ’09, ’10, 53 WA ER 1 and forms intermediate between this condition and normal are likewise different degrees of differential inhibition in the medio-lateral

gradient. Bellamy has produced similar medio-lateral differential inhibitions in Amphibia (Fig. 89), The changes in localization and differentiation in differential acclimation and recovery and in differential acceleration are, as already noted, in the opposite direction from those characteristic of differential inhibition. Differential acclimation in the sea urchin was shown in Figures 53-56 with Figures 48, 49 representing normal larve for comparison. Bellamy’s work on the amphibia, unpublished work by MacArthur on several species of echinoderms and my data on the sea urchin and Planaria (Child, ’16 d, ’21 c) show that differential acclimation and recovery determine changes in localization and differentiation in the same direction as regards any particular axis in all these forms and without specific relation to any particular agent.

The fact may be emphasized in passing that in these experiments, as elsewhere, polarity and symmetry appear as non-specific or quantitative regional differentials in a specific protoplasm. The production of anaxiate or modified forms does not require any particular agent or action, but polarity and symmetry can both be obliterated or modified by a large number of widely different agents and conditions. The obliteration or modification of the axes depends simply upon the differential susceptibility of different levels of the gradient which constitutes physiologically the axis.!

The data on modification of development through differential susceptibility leave no room for doubt that differentiation is very closely associated with, and dependent upon, the physiological gradients. When the gradients are altered, localization and differentiation are altered, and when a gradient is obliterated, differentiation does not occur in that particular axis. Moreover, these changes are determined by changes which are primarily non-specific or quantitative. We do not have to use particular agents to obliterate specific primordia, but by obliterating the gradient at a sufficiently early stage, we can prevent a primordium from developing and attaining a stage of specific differentiation. Similarly, by altering length and slope of the gradients we can alter localization of organs along the axes represented by the gradients. And finally, we shall

1 Recently Dr. M. A. Hinrichs (’23 a) and I have observed independently that the same developmental modifications, resulting from differential susceptibility at different levels of the gradients, can bé obtained in the sea urchin with ultraviolet radiation and with sunlight after photochemical sensitization as with other agents. With proper procedure either differential inhibitions like Figures 50-52 or differential recoveries like Figures 53-56 can be obtained. No attempt has been made thus far to obtain differential acclimation with continuous exposure, but there is little doubt that with a certain range of intensity it will

see in the following chapter that the appearance and establishment of new gradients determines axial differentiation in new directions. The data on developmental modification through differential susceptibility not only do not require the postulation of symmetrical or asymmetrical molecules (Harrison, ’21) or a space lattice (Przibram, ’21), but they are difficult to interpret in such terms, for it is highly improbable that agents of widely dissimilar constitution acting upon protoplasm in widely different ways, can all produce the same stereochemical changes, nevertheless, with proper concentration, period of exposure, etc., they do produce similar developmental modifications, not only in a single species, but in widely different species.

The problem of the origin and development of the nervous system as it appears in the light of our present knowledge concerning the physiological gradients has been considered at length elsewhere (Child, ’21 a). Here it is possible only to refer briefly to a few points concerning localization and differentiation. In axiate animals the chief aggregation of nervous tissue is localized in the apical region or head, the region which primarily possesses a higher metabolic rate than any other part. Similarly the postcephalic portions of the central nervous system arise from the regions of highest rate in the symmetry gradients (see Child, ’21 a, Chap. VI) and the progress of nervous differentiation in the posterior direction is an expression of the antero-posterior gradient.

As might be expected from its position in the gradients, the nervous ‘system becomes morphologically distinguishable earlier than other definitive organs. Evidently then, its development, at least in its earlier stages, is independent of other organs which arise later. In short, the central nervous system arises from the dominant regions ,of the chief axial gradients and becomes, with its receptors, the organ par excellence of functional dominance or control in the behavior of the individual organism. As I have tried to make clear, the facts seem to me to indicate that the nervous system represents the morphological and physiological development and complication of the excitation-transmission relation which in its simplest form originates from the non-specific differences in physiological condition at different levels of a physiological gradient. If this is true, it follows, since the physiological gradient originates as a protoplasmic re-

~ action to local or differential action of an external factor, that, physiologically speaking, the nervous system itself originates in protoplasmic behavior. The nervous system shows relations not only to the axiate, but to the surface-interior pattern. Sensory cells very generally are definitely arranged with reference to a surface. The epithelio-muscle cell of hydra (Fig. 90) probably represents a simple excito-motor ap- -paratus originating in relation to surfaceinterior pattern (Child, ’21 a, Chap. XIII), the receptor or sensory end being at the surface, the effector or motor portion in the elongated contractile base, and the conductor between the two.

If this conception of the réle of the zs es physiological gradients in localization and Fic, 90—-Hpitheliommuscle differentiation of the nervous system is cell of hydra. correct, they must play a part not merely in determining the general region from which nervous structure develops, but also in determining the polarity of individual neuroblasts, the localization and direction of outgrowth of axons, etc. In Chapters X and XI of “The Origin and _ Development of the Nervous System’ this question was discussed and a modified form of Kappers’ (’17 and references there given) hypothesis of determination of neuroblast polarity and outgrowth by bioelectric currents was advanced. According to this hypothesis the electric currents are primarily associated with the general axial eradients and so determine neuron polarity in relation to these gradients.

The suggestion advanced as to the manner in which the physiological polarity of the neuroblast is electrically determined is, briefly stated, that the neuroblasts are electrically polarized by the bioelectric currents to which they are exposed and that such polarization determines physiological polarity. The higher levels of a physiological gradient are electro-negative galvanometrically to lower levels. According to current conceptions the higher levels must therefore be internally electro-positive to lower levels. Consequently when a cell becomes polarized by exposure to an electric current the region or pole which becomes internally electro-positive to other parts will become the high end of a physiological gradient, a region’ of outgrowth and of dominance.

In the discussion of this hypothesis, it was assumed that the membranes of the neuroblasts are impermeable at least to positive ions to such an extent that these ions accumulate within the cell on the side toward the bioelectric cathode and that this polarization determines the polarity of the neuron. This suggestion was advanced because it seemed to fit the facts in the case better than any other. It may, however, be pointed out that it represents only one of two possibilities of electrolytic determination of physiological polarity. If the cell membranes are highly permeable to both positive and negative ions, the positive ions will leave the cell on the cathodic and enter on the anodic side, while the negative ions will pass out on the anodic and enter on the cathodic side. Under such conditions the polarization of the cell will be opposite to that which occurs in cells with relatively impermeable membranes and the internally positive region of the cell will be determined on the side toward the anode.

According to this conception physiological polarity may be determined by an electric current in two opposite ways, depending upon conditions of permeability to positive and negative ions. As already pointed out, this hypothesis of electric determination of physiological and particularly of neuron polarity is nothing more than an attempt to call attention to possibilities along this line.+ The general physiological gradients are present in axiate animals and, the evidence indicates that bioelectric currents are characteristic features of them, but extensive experimental investigation is necessary in order to determine whether these currents affect the neuroblasts in the manner suggested or in some other way, or whether they affect them at all. Asa beginning along this line it may be noted that Ingvar (’20) has recorded the determination of direction of outgrowth of cells in tissue cultures, and Lund (21 a) has determined polarity in pieces of hydroid stems by electric current, the apical ends being determined on the side toward the anode.

We are accustomed to say that the reflex arc represents the functional unit in the nervous system. The conception of origin of the nervous system under consideration here amounts essentially to saying that the physiological gradient is the simplest, the most generalized, and the primary reflex arc in the individual and so constitutes the physiological basis for the structural and functional development of all other arcs. It is perhaps necessary to point out that this conception of ner-

1 Recently Kappers (’22) has criticized my hypothesis at length, but since it was intended as nothing more than a suggestion and since experimental data on which to base definite conclusions are lacking, discussion of these criticisms can be little more than an academic exercise and is quite unnecessary here, vous origins does not in any way conflict with data or theories of heredity. Heredity is concerned with potentialities of development and differentiation, but here we are concerned with the realiza- _ tion of hereditary potentialities in the individual. Hereditary potentialities of nervous development and differentiation exist in most animal protoplasms, but for the development of an individual nervous system certain physiological conditions are necessary, and I have endeavored to show that those conditions arise primarily in certain non-specific relations between a living protoplasm and environmental factors. When we say that the physiological gradient is the basis of the reflex arc we do not in any sense deny that each particular kind of protoplasm possesses hereditary potentialities which determine the characteristics that distinguish structurally and functionally the reflex ares and the nervous systems of one species from that of another. We mean merely that for the realization in an individual of these hereditary potentialities of a specific protoplasm a physiological gradient is a primary and necessary factor.

There can be no doubt that.a quantitative physiological gradient provides various possibilities for localization and differentiation and the experimental evidence indicates that changes in the gradient determine changes in localization and differentiation and that obliteration of the gradient obliterates, so far as can be determined, the axis which the gradient represents. If the physiological gradients were products or results of some underlying molecular or other structural polarity or symmetry we should not expect to find that obliteration of a gradient through differential susceptibility would obliterate the axis which it represents. As a matter of fact, however, we find that when a gradient is obliterated in early developmental stages localization and differentiation along that axis cease and we are not able to show that the axis is still present. These experimental data obtained with widely different animals are highly significant as indicating that the gradient is itself the physiological basis of the axis, rather than a consequence or incident of some sort of molecular or other protoplasmic structure.

It was shown in the preceding chapter that the gradients and axes can be obliterated in organisms through the differential susceptibility of different levels. The question of their origin now arises. If polarity and symmetry are inherent in the molecular structure of protoplasm, as many zodélogists still maintain, changes can be brought about only by the modification of that structure and we have to assume that the conditions which brought about the change have done it by altering molecular structure. Leaving out of account the fact that if polarity and symmetry in general are matters of molecular structure, we ought to find some indication of optical axes in protoplasm, this hypothesis involves us in other difficulties, some of which are pointed out below.

If, however, the physiological gradient itself is the basis of polarity and symmetry all the assumptions of molecular structure and change, in favor of which there is no actual evidence, become entirely unnecessary. From this viewpoint polarity and symmetry are primarily dynamic in character, and molar or regional, rather than molecular in magnitude. Admitting this, it evidently ought to be possible not only to obliterate old polarities and symmetries, but to determine new axes experimentally, by exposing the protoplasm to a differential in action of some external factor which will determine in one way or another a physiological differential, 7. e., a physiological gradient. Moreover, since the physiological gradients are apparently primarily quantitative we do not need to employ specific ‘factors to determine them. Any factor that will determine a more or less persistent quantitative differential in the protoplasm ought to be adequate.

As a matter of fact we can determine physiological gradients and axes experimentally by differential exposure in many animal and plant protoplasms and in many other cases the data of observation indicate that in nature physiological axes very often arise in the same way, though it is evident that in some cases they persist from earlier individual or cell generations. Light has long been recognized by the botanists as a factor of great importance in determining both polarity and symmetry in plants through its differential action on different parts or regions. The determination of polarity in the egg of Fucus through the action of light has already been noted (pp. 58-61, Figs. 16-18). The action of light in determining polarity is much the same in the spore of Equisetum (Stahl, 1883). Polarity may also be determined by light in the alga, Bryopsis (Winkler, 1900 a). In certain liverworts light determines dorsoventrality and in various alge the thallus is radially symmetrical when illumination is equal on all sides and bilateral when it is unequal or unilateral. According to Jenkinson (09) light may play some part in determining bilaterality in the frog’s egg. It is important to note that in all these cases the molar regional differential in light action appears to be the important factor rather than any stereochemical or other orienting action upon the proto-plasmic molecules. Exact knowledge concerning the nature of light action is, however, lacking. The fact that a certain range of wave length toward the blue-violet end of the spectrum is significant in determination of polarity in Fucus (Hurd, 719, ’20) suggests a chemical effect.