The Shadow Reaction of Diadema Antillarum Philippi: II. Inhibition by Light

In our previous study of the shadow reaction (Millott & Yoshida, 1960) the relation between the response and the preceding illumination and shadow led to the suggestion that the reaction may be a rebound from inhibition produced by light. Also, that the shorter duration of the reaction, together with the diminished amplitude and frequency of the contractions produced by brief shadows as compared with those produced by longer, is due to the inhibiting effect of the re-admitted light rather than to any effect of the duration of shading on processes initiated in darkness. Similarly, where the light was not cut off completely the lesser effect could be explained as due to the inhibiting action of the light remaining. A shadow would thus lack any intrinsic value as a stimulus, being a mere interruption of the inhibitory influence exercised by light.

Other explanations are possible, but a cardinal feature of this notion is the existence of an inhibitory effect of light. It is the object of this study to show that light exerts such an action.

If light has an inhibiting effect, it should be possible to make it suppress the shadow reaction and, by arranging events so that light is re-admitted after a reaction has been set in train, the inhibition should become clearly evident.

To show that light, and not the shadow, is the factor which moulds the ensuing reaction, the shading may be kept constant and the intensity of the re-admitted light varied, as a result of which the degree of inhibition should vary in the same way.

Again, if light inhibits, it should be possible to make its effect at one point inhibit the response to a shadow cast elsewhere. Experiments based upon those reported in the previous account were designed to demonstrate this in both radial nerve and skin.

The responses of a single spine were examined in preparations of the same kind as those previously employed. The experimental tank and the means of recording spine movement, together with the signals showing the onset of light and shade, the time trace, etc., were the same as before.

To produce timed shadows and light spots in rapid succession, in different places, two light beams were used, obtained from twin lamps in the manner already described, each focused to form a separate light spot. The optical axes were inclined so as to allow the two spots to be superimposed as required and the size and intensity of each was controlled by interposed diaphragms and filters. To facilitate manipulation, the terminal lens was a long focal distance objective, and the lamps with their ancillary lenses, etc., were moved by rack and pinion. Vertically the two moved together, but only one could be moved horizontally, and in a restricted way, to keep it parfocal with its partner. The linear horizontal distance between the light spots was measured by a scale fixed to the moving member.

Projection of the spots was timed so that the first illuminated the preparation for 5 min., after which it was extinguished so as to set up the spine response. This is referred to as spot I. After a definite interval, a second light spot (spot II) was projected on to the same or a different place. Brief intervals were timed by the automatic rotating shutter already described, which was interposed in both beams, its form being so designed as to cut off the beam from one lamp and, after a predetermined interval, to admit light from the other. This interval was most commonly 38 msec., known from the previous study to be adequate to release a response. To avoid the disturbing effect of flicker, the lamps were operated on d.c. while the shutter was in use. Longer intervals were achieved by operating the shutter manually, the timing being only approximate ; the precise timing was recorded on the photographic record and measured on the accompanying time trace.

By means of the records responses could be compared as regards reaction time (latency, as defined in the previous account), amplitude and frequency of the contractions, together with the duration of the reaction, the validity of which as a basis for comparison has already been shown (Millott & Yoshida, 1960). Responses modified by spot II were also compared with unmodified control responses, elicited by using spot I alone.

Because the responses of a particular preparation decline steadily in vigour during the day, frequent controls are necessary to reveal the extent of such decline. As an added check experiments were repeated throughout the day with light spots in the same positions, either alone or with both in sequence. Testing the responsiveness in this way also served to check that the light spots were still projected on to the radial nerve.

When the reaction elicited by a brief shadow is compared with that which follows longer shading there are conspicuous differences (Fig. 1).

The most striking appear in the later part of the reaction, particularly in its duration which is clearly curtailed. Other features are affected; the amplitude of the contractions, their frequency and their uniformity of size is diminished. When the shadows are brief, i.e. less than 40 msec., the reaction time is increased. The effect on all these is graded, increasing as the shadow is made shorter and the period during which light can affect the reaction becomes longer.

That this is due to the light is shown by the effect of the complementary variable, namely intensity. If the initial (field) illumination is kept constant, but fight of different intensity is re-admitted, the effect on the duration of the reaction is as shown in Fig. 2, where the time at which light was re-admitted (abscissae) is plotted against the duration of the reaction (ordinates) expressed as a fraction of the duration of the longest reaction, regarded as unity. The progressively shorter reaction produced by increasing intensities displaces the curves down the ordinate axis.

It is informative to compare these results with those obtained from experiments of the same kind where the intensity of the field illumination was also varied so as to make it the same as that of the re-admitted light (Fig. 3). Here there is no such displacement ; the values for duration, which increase as the re-admission of light is delayed, fall on a common curve. The increased inhibition is here counterbalanced by the action of the stronger light before the shadow.

In these experiments both light spots with a constant diameter (1·0 mm.) were moved along the radial nerve both orally and aborally, over a range of up to 6·0 mm. ; but since this was measured on a flat scale, the true distance was slightly greater because of the curvature of the preparation.

Typical results are shown in Fig. 4, where a standard system of notation is used ; position 0 is approximately midway between the ambitus and the periproct and the degree of linear separation from this is indicated in units of 1·0 mm. on the flat scale, positive values being oral, negative aboral.

The effect of varying the position of spot II with respect to spot I is very clear. First, when the two spots are separated the inhibitory effect of spot II still appears and is exerted over distances as great as 6·0 mm. Inhibition is maximal when the two spots coincide (Fig. 4, position 0, where the reaction is completely suppressed) or are near together, and decreases progressively as they are separated. Such a gradient appeared wherever the nerve was shaded and it affected all the aspects of the reaction we have examined, though most clearly the duration and frequency. The considerable distance over which inhibition is exerted eliminated the possibility of the effect being the result of stray light.

These effects cannot be explained as due simply to differences in sensitivity between various regions of the radial nerve, because control experiments in which spot I alone was extinguished at different points on the radial nerve revealed no comparable differences.

The slope of the inhibitory gradient varied in different preparations. Most often it appeared as shown in Fig. 5, with a decidedly steeper slope when spot II is moved to the aboral side; less frequently, the decline was symmetrical and least often, the steeper slope occurred on the oral side. In some cases the effect was reversed in that, after inhibition from the second light spot had declined so as to be imperceptible, shifting it still further away resulted in a progressive increase in the frequency of the later contractions and in the duration of the reaction after illumination. This potentiating effect was seen only in some of the instances where spot II was aboral with respect to spot I and where the inhibitory gradient had a steeper slope aborally.

Neither the position of spot I on the nerve nor the position of both light spots with respect to the spine made any difference; the same type of gradient, with its varying slopes and sometimes with potentiation, could appear in any position.

Repetition of the type of experiment just described with variation of the intensity of the second light spot yields results of the type shown in Fig. 5.

They confirm what has been said above, the inhibiting effect of spot II increasing with its intensity. Since the shadows employed were brief an effect on the earlier part of the response (e.g. the reaction time) also appears, though it is less than that on the frequency or duration.

The results are informative in showing not only that the gradient is preserved but also that its form is reproduced at each intensity, so that the curves showing the effect of varying position are by and large shifted so as to form a parallel series (Fig. 5 C, D).

This was determined by projecting spots of three sizes (0·3,1·1 and 2·5 mm. in diameter) on to the radial nerve at five positions with respect to spot I, which was kept constant in size (1·1 mm. in diameter).

In each position inhibition increased as the area illuminated by spot II was increased. Here the effect appeared on the earlier as well as on the later parts of a reaction, so that the reaction time lengthened and the size of the initial (as well as of the later) contractions, their frequency and the duration of the reaction were reduced.

However, the size of the area illuminated by spot II affects the gradient of inhibition. With small and medium-sized spots it is of the type already described. With large spots, though the degree of inhibition is greater, the gradient is abolished and the reactions are always feeble, short and have a suspiciously long reaction time.

The explanation may be found in the controls (E 60 and 62) where projection of spot II by itself produces an ‘on’ response of the type already described (Millott & Yoshida, 1959), being brief and feeble with a long reaction time. This suggests that the responses following illumination of large areas may be ‘on’ responses and so it is necessary to be cautious in interpreting the effect of area.

The remarkably long reaction time raises a point of interest, for it is longer than when a reaction is elicited by the light of spot II alone; this means that the ‘on’ response has been affected, perhaps by the preceding shadow response.

Information concerning the pathways of the inhibitory influence and the site of its interplay with the shadow reflex would clearly be useful. It was obtained from simple nerve-cutting experiments. Preparations of the same type were used and the experiments were repeated using light spots 0·8–1·0 mm. in diameter, after making certain cuts in the radial nerve and its branches. Several difficulties were experienced. The branch nerves are translucent and because of their looseness are sometimes difficult to cut successfully. Cuts were made by fine needles or by glass knives, under a low-power binocular microscope, before the preparation was used. After use, it was fixed and the position and completeness of the cuts were checked again and their distances from the light spots, spine, etc., were measured. Where there was doubt, the experimental results were rejected.

The disposition of the cuts is shown in Fig. 6.

Cutting across the radial nerve between the position of the two light spots (Fig. 6 B) does not abolish inhibition. Thus spot II may be sited in positions corresponding to either 2 or 4 in Fig. 6B when spot I is in a position corresponding to 3. This means that some, at least, of the interaction must occur outside the central nervous system. Such was the case with light spots as far apart as 4 mm.

On the other hand, cutting the lateral branches at positions c and d in Fig. 6 C, so as to remove the direct pathways to the periphery from portions of the radial nerve as long as 3·5 mm., left the inhibitory gradients in that area unimpaired. Increasing the distance between the two light spots projected on to this region still reduced the degree of inhibition, a fact which holds, whatever the position of the spine with respect to the denervated area. This suggests interaction within the radial nerve, rather than at the effector; the interaction would be mediated by distinct peripheral pathways of excitation and inhibition, since the mechanism responsible for the gradient is preserved.

Experiments in which the branch nerves on either side of the radial nerve for a distance of 3·5 mm., as well as the radial nerve itself, have been cut across (Fig. 6 D) show that shading the radial nerve at position 3, just aboral to the cut across the main nerve (e), still elicits a response. This means that nerve pathways for the shading reflex can travel along about 3·5 mm. of the nerve cord before emerging to the periphery.

Similarly, projecting light spots between positions 2 and 3 (see Fig. 6D) in the same type of preparation shows that they can interact when 2·5 mm. apart. Further, since inhibition can be elicited by a spot just aboral to e, with the spine at either position 2 or 4, the inhibiting pathways traverse the radial nerve for at least this distance.

The effects on the reaction time are particularly interesting (Table 2). Sectioning the lateral branches (Fig. 6C) increases the reaction time of the response which follows shading the cord in the same region (pieces III and IV, Table 2); moreover, the effect is greater if the shadow falls in the middle of the denervated area than at the ends, and is greater still if the radial nerve is then cut across, particularly when the shaded area is nearer the cut end (Fig. 6D ; piece IV, Table 2). Although possible effects of denervation on the speed or responsiveness of the effector should not be overlooked, it is difficult to avoid the suggestion that the nerve cutting brings into action pathways that are more devious than usual.

Another effect of transecting the radial nerve and its branches is to make the shadow response more readily inhibited. Thus although the response to extinction of one spot projected on to the transected nerve cord (e.g. at position 3 in Fig. 6B), may remain the same, the inhibitory effect of a second spot projected on to a region of the cord, separated from the first by a cut (e.g. at either position 2 or 4 in Fig. 6B), is increased, so that the reaction time is sometimes strikingly greater and the response diminished to a greater degree by light spots of the usual intensity.

The pattern of inhibition is sometimes altered by cutting across the radial nerve, so that light spots projected near to the transected region of the nerve exert a marked inhibition, whereas before cutting they produce only a slight increase in reaction time. Again, the inhibitory gradient may be altered so that the point of maximum inhibition shifts from the point where the two light spots coincide to a neighbouring position.

Finally, transection sometimes augments the later part of the response which follows the extinction of a single light spot (Fig. 7), so that the amplitude, frequency and regularity of the beats is increased, as is also the duration of the reaction. This suggests the existence of a tonic inhibitory effect in the nerve cord.

In all, there is bewildering variety of effects, indicating a complex pattern of nervous interaction that is altered by transection. Numerous alternative pathways for the response and its inhibition seem to exist and nervous interaction can occur both centrally and peripherally. To what extent these alternative pathways play a part in normal responses remains unknown.

Preparations exactly the same as those previously described were used and the same technique was employed, except that here the preparation was held with the spine directed upwards, the two light spots being projected on to the outside (upper) surface.

Typical results from preparations reproduced in Fig. 8 show that the response to shading is inhibited by light spots projected on to most areas of the test, both ambulacral and interambulacral.

The inhibition is manifest in the same way as when the light spots were projected on to the radial nerves. Moreover, it forms meridional gradients, maximum inhibition being produced when the two spots fall on the same position, the effect declining as one spot is moved in the direction of the oral or aboral pole. The region over which the two spots interact may extend along a meridian for slightly more than 6 mm.

The slope of the gradient varies as in the radial nerve, not only in different preparations but also according to whether spot I is projected on to positions oral or aboral to the ambitus. It may be steeper orally or aborally and there is no consistency in the variation. Here it is pertinent to recall the irregularity of the outside surface due to the many spine bases, etc., which may to some extent account for the inconsistency.

Similar inhibitory gradients appear running parallel to the equator and these are particularly interesting since they extend over ambulacra and interambulacra, but the sensitivity, unlike that of the meridional gradients and radial nerves, is not uniform. Thus when the inhibitory effect of light spots projected on to these areas is determined and compared, it is seen that they form a gradient as follows :

ambulacral margin → ambulacral centre → interambulacrum.

This corresponds exactly with that of the reaction time (Millott & Yoshida, 1960) and the sensitivity to shading, previously reported (Millott, 1954); indeed the ambulacral margins are so clearly the most sensitive that, even though the two light spots may coincide at the ambulacral centre, inhibition is still less than when spot II falls at the margin.

The present study shows clearly that light inhibits the spine reflex and strengthens the suggestion that the shadow reaction may be the result of release from inhibition.

The other suggestion, that the effect of the shadow intensity was really that of intensity of light remaining, is also substantiated by the experiments in which light readmitted after shading is shown to inhibit the reaction with an efficacy roughly proportional to its intensity. There are thus good grounds for believing that any light remaining after shading would continue to inhibit the reaction that ensues, in the ways that the experiments described above have revealed.

The use of separate light spots has revealed more of the inhibitory mechanism involved. The effect of light spots of differing size shows that spatial summation of inhibition can occur over considerable areas. Interaction has been shown to occur, so that the shadow reaction is moulded by events in separated receptive regions, both in the skin and central nervous system.

The interaction can be complex and a variety of inhibitory patterns, some of which proved unpredictable, have appeared. There are also indications that interplay with the ‘on’ response may sometimes be involved. The work may well have been sufficient to reveal only a little of the complexity, and much more is required before a beginning can be made towards resolving it. Thus we have as yet paid no attention to the possible effects of sensory adaptation which affect interaction in the vision of vertebrates (Barlow, Fitzhugh & Kuffler, 1957).

The importance of nervous interplay in vision involving complex photoreceptors has been emphasized by Granit (1933, 1955) in the case of vertebrates, and by Hartline in invertebrates. Its importance in the dermal light sense of echinoids has now been revealed and it is pertinent to recall the earlier experiments on shadow reactions in Balanus performed by von Buddenbrock (1930), who realized the importance of illumination and showed the inhibitory effect of light, postulating the existence of interaction at nerve centres.

Clear evidence of spatial and temporal interaction in Diadema enables us to introduce the ‘receptor field’ concept into the dermal light sense. We regard the field as the area over which interaction can occur. Although our experiments have been inadequate to reveal the pattern or even the extent of such areas, the existence of gradients showing a gradual change from inhibiting to potentiating effects recalls the complex retinal organization revealed in vertebrates by Kuffler (1953), Barlow et al. (1957) and Wiesel & Brown (1958). Thus in some respects the dermal fight sense of Diadema shows a complexity which parallels that of elaborate photoreceptors. The dermal light sense may thus prove to be of much more complex organization than is commonly suspected.

There is a continually emerging parallel between the shadow response in Diadema and the ‘off’ effect in certain elaborate photoreceptors. Though the planes of analysis differ greatly, the present study extends this parallel. Responses to changes in light intensity are regarded as both prominent (Hartline, 1938 a) and distinctive (Ratliff & Mueller, 1957) features of the vertebrate eye, though ‘off’ effects have been described in the eye of Pecten (Hartline, 1938b). Such responses are widely believed to be the result of interplay between excitation and inhibition in ganglionic layers associated with complex eyes. Inhibition has now been shown to play its part in the shadow reaction of Diadema in which the receptive surface is extensive and diffuse.

It is noteworthy that in the field (or gradient) the inhibitory effect is greatest between regions that are closest together, and this could have an effect, similar to that mentioned by Hartline (1958) in connexion with complex photoreceptors, of emphasizing the contrast generated by patterns of light and shade falling on the skin of Diadema. This would be particularly effective if the pattern were continually changed by a moving shadow. Such an organization would signal very effectively the position of a moving (and possibly harmful) object in the vicinity, calling forth an effective response in the poisonous spines.

We are very grateful to the Zoological Society of London, especially to Dr H. G. Vevers, for much help. Our thanks are due to Mr S. E. White of the College Workshops and to Drs Brindley, Denton, and Pirenne, who criticized the manuscript. The work was accomplished while one of us (M.Y.) held a Research Fellowship awarded by Bedford College.