62 THE EYE IN EVOLUTION 



the Amphipod, Talitrus saltator, for example, reacts in this way to 

 the moon (Pardi and Papi, 1953). Until recently most writers agreed 

 that this reaction was the essential factor in the orientation of insects 

 out-of-doors. This is probably the case when fog or cirrostratus turns 

 the sun into a small light source, but the lack of consistency in the 

 behaviour of insects in natural conditions when the sun is bright does 

 not substantiate that this is the main or even an effective mechanism, 

 and the experiments of Wellington (1955) would seem to indicate that 

 solar heat and the response to the plane of polarization of light^ are the 

 essential factors in determming their conduct in these circumstances. 



The accuracy of the response of the light -compass reaction in 

 insects is made possible by the structure of the compound eyes, for 



Fig. 38. — Menotaxis. 



The insect moves so that its course makes a constant angle (a) with rays 

 of Hght issuing from a source ; it therefore approaches the source along a 

 logarithmic spiral (after von Buddenbrock). 



they orientate themselves in such a way that the sun's rays stimulate 

 one or at most a few ommatidia all the time.^ The high degree of 

 accuracy thus obtained may be gathered from the fact that insects 

 sometimes correct their angle of orientation if the light merely passes 

 from one ommatidium to its neighbour (von Buddenbrock and Schulz, 

 1933). When the guiding light is sufficiently far away this type of 

 response is effective in orientating the insect in a straight line, but if 

 the stimulus is close an entirely different result is seen. If the insect 

 were to pursue a straight path, the incidence of such a light on the 

 retina would constantly change ; and if the angle of incidence is to be 

 kept constant, the insect must perforce turn along a logarithmic spiral 

 which ends in the light itself (Fig. 38) (von Buddenbrock, 1937). Cater- 

 pillars crav ] to a light in this type of sj^iral path (Ludwig, 1933-34) 

 and it is for liis reason that the moth, applying a mechanism adapted 



1 p. 73. * p. 174. 



