PHOTORECEPTION 193 



surface (Stephens et al., 1953; Baylor and Smith, 1953) does not con- 

 form to a number of experimental facts. For one thing, the necessary 

 properties would be realized only in an optically isotropic medium, 

 and the corneal lenses of Drosophila are anisotropic (Stockhammer, 

 1959). Nor is a mechanism of double refraction present. When the 

 dioptric apparatus is observed through a rotating polaroid no change 

 in the pattern or intensity of transmitted light is observed (Autrum and 

 Stumpf, 1950; de Vries and Kuiper, 1958; Stockhammer, 1959). 

 Furthermore, analyses of the ERG have failed to support the hypo- 

 thesis (Autrum and Stumpf, 1950; Baylor and Kennedy, 1958; de 

 Vries and Kuiper, 1958). 



The on-effect of the ERG is known to depend in its magnitude on 

 the intensity of light. When the action potential resulting from the 

 stimulation of a total of twelve-fifteen ommatidia by polarized light 

 was measured there was no change in magnitude of response as the 

 plane of polarization was rotated. It seems unlikely, therefore, that the 

 eye as a whole acts as an analyser. Similar results obtained by stimula- 

 tion of a single ommatidium also indicated that the single ommati- 

 dium does not act as an analyser. However, when the effects of polar- 

 ized and non-polarized light of equal intensities were tested the 

 polarized light always elicited the greater response (Autrum and 

 Stumpf, 1950). From this it was concluded that each of the retinal 

 cells (eight in the honeybee and seven in Calliphora) is maximally 

 sensitive to light of a definite vibration direction, namely, that per- 

 pendicular to the radial direction of the cell within the ommatidium. 

 With unpolarized light each cell receives less than maximal stimula- 

 tion; with polarized light some receive maximal and some minimal. 

 The net result in the ERG is a response of greater magnitude than to 

 ordinary light because maximal and minimal effects do not cancel out. 

 Thus, the unit of analyses seems to be the single retinal cell, and the 

 central nervous system is presumed to integrate the pattern of light and 

 dark from each ommatidium. 



This hypothesis is supported by the electrophysiological work of 

 Liidtke (1957), Naka and Kuwabara (1959), and Burkhardt and 

 Wendler (1960), as well as by field experiments of von Frisch (1950 a, 

 1950 b). 



Von Frisch (1950 a, 1950 b) provided support for the hypothesis by 

 comparing the actions of bees dancing under a polaroid material with 

 the changes observed in the light-intensity pattern of the sky when 

 viewed through an artificial ommatidium constructed of eight seg- 

 ments of polaroid material each oriented differently. When this model 



