TIMOTHY H. GOLDSMITH 775 



"relative stimulative efficiency," that is plotted as a function of wave- 

 length in the uppermost graph of Fig. 1. Although this correction 

 is surely in the proper direction, it assumes direct proportionality be- 

 tween phototactic effectiveness and intensity, an assumption which is 

 almost certainly incorrect (46, 7, 28) . A further difficulty with these 

 experiments arises from the fact that the curve is pieced together from 

 experiments done at two different times with two different lamps. The 

 two halves of the curve (filled and open circles) are brought to- 

 gether by making equal the "relative stimulative efficiencies" at 436 

 m/x. If, however, the curves had been made to coincide at 546-550 m|u, 

 or at 575-577 m^ (the other points at which data were obtained with 

 both lamps) , the peak in the blue-green would be more prominent. 



The phototactic response curve measured by Wolken (65) for 

 Drosophila is an action spectrum; the ordinate is the reciprocal of the 

 relative energy necessary for a constant phototactic effect. The curve 

 peaks in the blue-green and starts to rise again at shorter wavelengths. 

 Unfortunately, it does not extend to wavelengths shorter than 420 mfx,, 

 but already by that point it has climbed higher than the maximum 

 at 508 irifji. The data of Wolken thus support Bertholf's conclusion 

 that for Drosophila the most effective wavelengths for phototaxis lie 

 in the near idtraviolet. 



Cameron (7) performed several kinds of experiments with the 

 housefly, Musca. The relative phototactic response to wavelengths of 

 equal energy (filled circles, solid curve) and two minor variations of 

 the method of "relative stimulative efficiency" (open circles, dashed 

 curves) all gave results similar to those of Bertholf for Drosophila. 

 None of these curves is a proper action spectrum, and (as was recog- 

 nized by Cameron) the significance of the quantity "relative stimula- 

 tive efficiency" is particularly uncertain. 



The last four curves in Fig. 1 apply to the honeybee, Apis. Bert- 

 holf (4, 5) determined the "relative stimulative efficiency" of different 

 wavelengths for phototaxis. The curve has a large peak in the near 

 ultraviolet and a smaller peak in the green. However, the pair of 

 objections that were made to Bertholf's experiments with Drosophila 

 apply to these data as well. 



Daumer (9) measured the lowest energies at which bees were at- 

 tracted to monochromatic lights to which they had previously been 

 trained. His is thus an action spectrum for threshold recognition of 

 various wavelengths. Sufficient data to locate the highest maximum 

 are not included in Daumer's paper, but it is dear that the major 

 peak lies in the ultraviolet. 



