184 THE PHYSIOLOGY OF INSECT SENSES 



Kuwabara (1957) and of Daumer (1956, 1958), who also demon- 

 strated that bees can even discriminate between wavelengths within 

 these bands, but with less precision. 



Colour vision can also be investigated by colour-matching tests. 

 Daumer (1956, 1958) applied these with outstanding success to honey- 

 bees by training them to come to a particular colour. The training 

 colour was then paired with a colour produced by mixing wave- 

 lengths. Various mixtures were tested until one was found which was 

 as acceptable to the bees as the training colour. For example, bees can 

 distinguish between white light containing ultra-violet and that lack- 

 ing ultra-violet. White light with ultra-violet ('bee white') can be 

 matched by a mixture of 15 per cent ultra-violet (360 mjx) and 85 per 

 cent blue-green (490 mii). The complementariness of ultra-violet and 

 blue-green had already been suggested by Hertz (1939), and Use (1937) 

 had shown that purple adjacent to white sets up a green colour con- 

 trast in the white. Similarly, *bee purple' can be matched by mixtures 

 of yellow and ultra-violet. Goldsmith (1961) has summarized Daum- 

 er's results in a preliminary chromaticity diagram which is explained 

 in Fig. 90. 



From the results of all of his matching experiments, Daumer con- 

 cluded that the eye of the honeybee possesses a receptor maximally 

 sensitive to ultra-violet, one for blue-violet, and one for green-yellow. 

 The existence of two of these has been confirmed by the electrophysio- 

 logical analyses of Goldsmith (1958 a, 1958 b, 1959, 1961). The wave 

 form of the ERG in the early studies with grasshoppers and Dytiscus 

 had shown no change as the wavelength of the stimulus was changed, 

 provided the intensities were matched (Crescitelli and Jahn, 1939; 

 Jahn, 1946; Jahn and Wulff, 1948). On the other hand, as Walther 

 (1958 a) found in the cockroach eye, there are some areas of the eye 

 where differences in the form of the ERG arising from differences in 

 the colour of the stimulus cannot be matched by adjusting intensity. 

 Goldsmith and Ruck (1958) also found that the ERG of the ocellus of 

 the bee had at low stimulus intensities a different contour for ultra- 

 violet than for green. 



By adapting with lights of different wavelengths it is possible to 

 demonstrate the existence of different receptors (Hamilton, 1922; 

 Fingerman and Brown, 1953). The dark-adapted eye of the worker 

 honeybee shows a peak of sensitivity in the green (535 m^t) and a 

 smaller peak in the ultra-violet (Goldsmith, 1960). If the eye is 

 adapted with red or yellow light the peak at 535 miJ. diminishes and the 

 peak at 340-345 m[j. increases (Fig. 91). The complementary experi- 



