272 



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



_ 



c 7i 



5 O- rT 



5 10 15 20 25 



Tiber- A 



Frequency (impulses per sec.) 



Fig. 21.Thedependenceof the magnitude of inhibition on distance. The inhibition 

 exerted by a small group of receptors (A) on two other receptors (B and C) is plotted 

 as ordinate. As abscissa is plotted the concurrent frequency of the discharge of im- 

 pulses of one of the receptors in the group A. The geometrical configuration of the 

 pattern of illumination on the eye is shown in the insert. The locations of the 

 facets of the receptors whose discharges were recorded are indicated by the 

 symbol © . The receptor A was at the center of a group of six or seven receptors 

 illuminated by a spot of light 1 mm in diameter. The illumination on B and C was 

 provided by spots of light 0-2 mm in diameter and of fixed intensity. The effects 

 of the group A on B and C were determined separately. (From Ratliff and 



Hartline, 1959). 



region of the eye. Sometimes a near neighbor of an ommatidium will exert 

 a much weaker influence on it than will its more distant neighbors. Further- 

 more, the correlation between the threshold and the coefficients of the in- 

 hibitory action is not perfect. We have even seen a few cases in which it is 

 the reverse of that shown in Fig. 21. We have already noted that the strength 

 of the inhibitory influences between two receptor units is not necessarily 

 equal for the two directions of action, as it was, approximately, for the pair 

 represented in Fig. 12. Sometimes, but rarely, the inequality is considerable: 

 an ommatidium may effect a particular neighbor quite strongly but be only 

 slightly affected by it. However, when we measure the interactions of moder- 

 ately sized, compact groups of receptors we find greater regularity and the 

 rule we have stated is followed in the manner illustrated in Fig. 21. 



It is evident that any law relating the strength of the mutual inhibitory 

 interaction of any two individual ommatidia in the retinal mosaic to the 

 distance by which they are separated must take a statistical form. At the 

 present we do not have enough data to formulate such a law quantitatively. 

 Nevertheless, the broad rule we have stated is significant. The fact that the 

 inhibitory parameters are strongly dependent, albeit statistically, on retinal 

 separation, introduces a topographic factor into the inhibitory interaction 

 which gives it special significance in pattern vision. 



We have alluded to the possible role of retinal inhibition in explaining the 



