276 H. K. HARTLINE, F. RATLIFF AND W. H. MILLER 



curves but are so weak in the particular case we have computed, that there 

 would be httle chance of detecting them in an effort to decide between the 

 two models for human vision. In Limulus, they must be present and perhaps 

 could be demonstrated. 



These exercises with idealized models are perhaps instructive but they are 

 so speculative at present that we will not give any numerical results here. 



Up to this point our analysis of the inhibitory interaction in the eye of 

 Limulus has been confined to the "steady state", in which light was allowed to 

 shine steadily on the eye for a long enough time to permit sensory adaptation 

 of the receptors to take place and to permit the inhibiting influences to take 

 effect and come to a mutual equilibrium. Whenever the pattern of illumination 

 on the eye is changed, transient changes in receptor activity take place and 

 readjustments of the inhibitory influences follow. These transient changes are 

 no less interesting than the steady-state interaction, and are of equal or greater 

 significance in visual physiology. 



Transient inhibitory effects are demonstrated in an experiment in which 

 the responses of two ommatidia were observed (Fig. 23). To begin with, the 

 receptors were steadily illuminated ; then the hght intensity on one of them 

 was increased for a period of several seconds. An isolated receptor similarly 

 stimulated responds to the increment of intensity, after brief latency, with a 

 sharp peak in frequency; as the receptor adapts the frequency soon subsides 

 to a steady level, higher than the value it had before the increment was 

 applied. When the increment is turned off, there is again a short delay and 

 then a sharp dip in the frequency of discharge which reaches a minimum and 

 then returns to a value close to the initial level (MacNichol and Hartline, 

 1948). The upper curve of Fig. 23 illustrates these phenomena in the experi- 

 ment under consideration; these frequency changes were determined pri- 

 marily by the response of the first receptor to the stimulus increment that was 

 applied to it. The second receptor, whose responses are plotted in the lower 

 curve, was steadily illuminated throughout the entire period. Its frequency 

 changes mirrored those of the first, with maxima and minima inverted with 

 respect to those of the first fiber. Evidently the inhibition exerted by the first 

 receptor varied with its discharge rate and these variations were followed 

 with some fidelity by the frequency of the second. It must now be recognized 

 that the variations in the frequency of the second receptor must also have 

 produced changes in the inhibition it exerted back on the first so that the 

 responses of the two receptors, being mutually interdependent, must have 

 affected one another so as to modify each other's transient responses reci- 

 procally. Since there are time delays in the exertion of inhibitory actions, and 

 since the inhibitory process may itself exhibit transient changes in magnitude, 

 it is evident that the temporal interactions of a group of receptors may become 

 quite complex. 



We have begun the analysis of the temporal aspects of the inhibitory inter- 



