PHOTORECEPTION 183 



americana and Blaberus craniifer show a peak at 500 my. (Goldsmith 

 and Ruck, 1958). 



One of the more interesting aspects of all of these descriptions of 

 spectral sensitivity is the great stimulating effectiveness of the near 

 ultra-violet. Hess (1920) had questioned the sensitivity of the insect 

 retina to ultra-violet and suggested the possibility of stimulation 

 actually occurring as a result of fluorescence set up in various tissues of 

 the eye. Fluorescence may occur (Walther and Dodt, 1959), but 

 Merker (1929) and Lutz and Grisewood (1934) showed that crushed 

 eyes of Drosophila did not fluoresce in ultra-violet of the wavelength 

 to which flies respond. They showed, furthermore, that the cornea 

 of Apis and Sarcophaga can transmit in the 253-m[JL band of the 

 spectrum. Additional reasons for rejecting Hess's hypothesis are 

 mentioned by Walther and Dodt (1959) and discussed by Goldsmith 

 (1961). 



Taken at face value, the spectral sensitivity data indicate what may 

 constitute brilliance for an insect, and a number of studies bear this 

 out wholly or in part (MoUer-Racke, 1952, with Dytiscus, but compare 

 Schone, 1953; Rokohl, 1942; Llidtke, 1953, and Resch, 1954, with 

 Notonecta). Demonstration of true colour vision has been technically 

 more difficult. There is now evidence for colour vision in thirty-three 

 genera of insects in six orders. The evidence has been derived princi- 

 pally from training experiments, optomotor responses, and electro- 

 physiological analyses. 



Von Frisch (1914) trained honeybees to collect food from a dish 

 placed on a card of a particular colour. This card was then placed in 

 one square of a checkerboard of greys (white to black) of different 

 intensities. The position of the coloured card was changed constantly. 

 Under these conditions the bees could always pick out a blue or yellow 

 card from the checkerboard of greys. Thus, it was demonstrated that 

 bees conditioned to blue confused blue-violet and purple, while those 

 conditioned to yellow confused yellow, orange, and yellow-green. 

 None could distinguish red. Later tests of the papers used by von 

 Frisch showed that some of the blues and greens reflected ultra- 

 violet, whereas some of the yellows and greens reflected red and blue 

 (Lutz, 1924). Similar experiments employing spectral colours pro- 

 jected on a white background demonstrated that bees could dis- 

 tinguish four regions of the spectrum: yellow-green and orange 

 (510-650 m|j.), blue-green (480-500 ma), blue and violet (400-480 m\L\ 

 and the near ultra-violet (300-400 m\3) (Kiihn and Pohl, 1921 ; Kiihn, 

 1927). These results were confirmed by the training experiments of 



N 



