180 THE PHYSIOLOGY OF INSECT SENSES 



light of variable intensity was substituted for the coloured light, and 

 the intensity was found which, when tested against the white stand- 

 ard, gave the same ratio of attractiveness as the test colour. Despite 

 these attempts at control, Bertholf did not have control over the 

 energy of his stimuU (see Weiss, 1943 a, 1943 b, and Goldsmith, 1961, 

 for a full discussion). The results of his experiments showed a peak 

 sensitivity in the ultra-violet (365 mjjt) and a slight indication of sensi- 

 tivity in the blue-green (487 my.) for Drosophila and in the yellow-green 

 (553 mji.) for Apis. Sander (1933) adjusted his lights so that the in- 

 tensities of the white standard and the test colours were equal. He then 

 used the number of insects {Apis) attracted to each as an index of rela- 

 tive effectiveness and found peaks in the yellow-green (570 mji.) and 

 in the blue (470 m^), but none in the ultra-violet. Cameron (1938) 

 employed all of the aforementioned methods in tests with Musca and 

 found a maximum in the ultra-violet. Weiss and his co-workers ex- 

 posed approximately 1 5,000 insects representing forty species to ten 

 wavelengths of light of equal intensities. The composite group be- 

 haviour pattern consisted of a peak at 492 m[ji and a maximum at 

 365 ra\i. Wolken determined for Drosophila the relative energy re- 

 quired to produce a constant phototactic response and found a peak 

 in the blue-green and a maximum commencing towards the ultra- 

 violet. Heintz measured, as a function of intensity, the number of bees 

 that crawled per unit time over a slit in a box illuminated by different 

 wavelengths and found a small peak in the green and the maximum 

 in the ultra-violet. He suggested that Sander's failure to detect a peak 

 in the ultra-violet stemmed from the very high light intensities em- 

 ployed. Since these intensities elicited maximum responses in the blue- 

 green, no higher response could possibly have been obtained in the 

 ultra-violet. Goldsmith (1960), perturbed by Sander's failure to find a 

 maximum for the bee in the ultra-violet, tested the phottoactic effective- 

 ness of green (546 mfx) and ultra-violet (365 m[j.) lights of equal in- 

 tensities and did find a peak at 365 mt^. In short, whether the curves 

 describing the relative effectiveness of different wavelengths in eliciting 

 phototactic responses are true action spectra (relative energy for a 

 constant effect) or not, all show that the most effective part of the 

 spectrum for Musca, Apis, and Drosophila is the near ultra-violet and, 

 to a lesser degree, the blue-green. This finding has been confirmed in 

 whole or in part for so many species by Weiss and his associates that it 

 is probably of general occurrence. 



The danger, of course, in interpreting these behavioural tests with- 

 out caution as indicative of spectral sensitivity is that the response of an 



