ELECTRICAL PHENOMENA IN VISION G03 



off- and on-responses. It has been found to vary from 0.001 to 10,000 

 in the cat. There is some evidence for a relation between the off/on 

 ratio and wave-length sensitivity. Granit believes that the on-off ele- 

 ments are most hkely to be the basis for color discrimination in the cat. 



3. "Dominator" elements: In the dark-adapted cat the spectral sensi- 

 tivity measured for a retinal element often resembles that of visual-purple 

 absorption. Such an element is known as a '"'scotopic dominator." Cer- 

 tain fish have a visual-violet system, and in them the scotopic dominator 

 element is found to have a spectral-sensitivity curve with the higher wave- 

 length maximum characteristic of that substance. After light adap- 

 tation most elements in an eye having cone vision become "photopic 

 dominators" with a maximum sensitivity shifted to a higher wave length 

 (Purkinje shift). Fish that have a visual-violet system shift to a pho- 

 topic dominator at a still higher wave length. No scotopic dominator 

 was found in the cone eye of the snake, and no photopic one was found in 

 the rat or guinea pig, whose receptors are mostly rods. 



4. "Modulator" elements: Some light-adapted elements of the frog, 

 rat, guinea pig, and snake showed narrower spectral-sensitivity curves 

 than those described as dominators. These elements were individually 

 tuned to respond to particular wave lengths of light and were named 

 "modulators" on the assumption that they were the basis for qualitative 

 discrimination of color. In the cat it was necessary to use an indirect 

 method to find such modulators, since in only a single experiment was 

 a modulator ever found by the direct method in this eye. The indirect 

 procedure used in finding modulators in the cat was, first, to light-adapt 

 the eye selectively by the use of red, green, or blue light obtained through 

 Ilford spectral filters. Then the spectral sensitivity function was deter- 

 mined for any given retinal element. If this function resembled that of 

 the scotopic dominator, Granit concluded that only receptors containing 

 visual purple were connected to the fiber being studied. If, on the con- 

 trary, the spectral-sensitivity curve departed from that of the scotopic 

 dominator, the conclusion was drawn that cone receptor elements were 

 present along with rod receptors in the region served by the fiber in 

 question. This situation is shown in Fig. 13-14. Here it is shown that 

 a colored adapting light has changed the spectral sensitivity function of 

 one particular element from that shown in curve P to that in curve u. 

 Curve p is the result that might have been expected for light-adapted 

 rods alone. The fact that u shows relatively less diminution in the region 

 near 550 m/x is interpreted to mean that one or more cone elements are 

 present whose maximum sensitivity lies in this region. These cone ele- 

 ments are assumed to be modulators whose response curves are obtained 

 by the procedure of subtracting curve p from curve u. Figure 13-15 

 shows some examples of the resulting modulator curves for the cat. 

 These are averages obtained in the course of several experiments. Granit 



