12 



HOW WE SEE 



Purkinje shift would not occur there. 

 This is essentially correct. Walters and 

 Wright have shown that relative luminosity 

 curves measured in the fovea change 

 slightly as the luminance is changed from 

 full dayhght to full twilight conditions. 

 The change is very small, however, as 



The magnitude of the Purkinje shift at 3° 

 was not quite as great, and did not proceed 

 at as rapid a pace, as at 10°. 



Absolute Sensitivity of the Cones and Rods. 

 The photopic and scotopic relative lumi- 

 nosity curves shown in Fig. 5 are frequently 

 misinterpreted in another important respect : 



< > => 



Q Q > 



< "5 



> UJ (t 



»- cr < 



< I - 



u 

 cr 



2 



X 

 < 



o 



400 



500 600 



WAVELENGTH IN Tn>i 



700 



Fig. 7. Relative amounts of radiant flux required to stimulate the rods and cones, 

 replotted from the data shown in Fig. 5. 



These curves are 



compared with the full range of the Purkinje 

 shift which occurs in the peripheral regions 

 of the eye. Because of the varying densit}^ 

 of the rods in different retinal locations, 

 (see Fig. 2), it might also be anticipated 

 that the Purkinje shift would occur at 

 different rates in different parts of the eye. 

 This expectation is also in accord with the 

 experimental facts. Walters and Wright 

 studied the retina 3° and 10° from the fovea. 



it is not generally appreciated that the 

 curves shown there are relative curves; each 

 is drawn relative to its own maximum. 

 This misunderstanding leads to the frequent 

 mis-statement that the rods are not as 

 sensitive to red as are the cones. Several 

 lines of evidence, in particular recent 

 experiments by Chapanis (7, 8), and Wald 

 (91), show conclusively that the rods are, 

 in fact, just as sensitive to red as are the 



