568 THE EYE IN EVOLUTION 



indeed he does see. For this reason Konig gave up the method of 

 introspection entirely and trusted only to colour-matches in his 

 investigation of colour-blindness ; only if every colour in the spectrum 

 could be matched by a mixture of a given pair of colours should the 

 subject be considered a dichromatic colour-blind. There are occasions, 

 however, when the observation of the behaviour of animals in their 

 natural surroundings can yield satisfying results. We have already 

 noted many instances of such cecological research, for example, in the 

 study of the conduct of different Arthropods in their orientation towards 

 light ; a particularly good example is von Frisch's experiments on 

 bees, or the means of orientation employed by birds in navigation. 

 These methods, however, valuable as they are, are applicable only to 

 certain restricted types of complex behaviour of a nature such that 

 other incidental variables can be neglected. 



Two more generally applicable methods of research are available. 

 In the objective methods of approach a measurable physical phenomenon 

 presumably determined by a specific stimulus is observed — a contrac- 

 tion of the pupil to light, for instance, or an electroretinographic 

 response — and it is assumed that this reaction bears a relatively 

 constant relation to events on the sensory level. If a response of this 

 nature follows stimulation by one band of wave-lengths of light and 

 not by another, for example, it is probable that the first gives rise to a 

 sensation and the second does not. A further analysis is possible by the 

 study of reflex responses. If an animal exhibits characteristic reflex 

 reactions to varying stimuli it is reasonable to suppose that these 

 affect it in different and specific ways. The optomotor reaction 

 illustrates this. If an animal, be it insect or vertebrate, is faced with 

 a revolving striped drum and reacts to the succession of stimuli thus 

 presented to it by compensatory movements of its eyes or its body, we 

 can assume that the alternating stimuli have a different effectivity ; or 

 if an animal salivates when presented with one stimulus associated 

 by training with food and not with another, the deduction seems 

 inescapable that a discrimination is made between the two stimuli ; 

 but whether the differentiation remains on the reflex level or is 

 appreciated as a sensation is sometimes problematical. 



The elicitation of such reflex responses, however, although 

 suggestive, gives us little idea of the conscious appreciation of sensations 

 and their effectiveness in determining conduct. A more satisfying 

 approach is the subjective rnethod of the study of what appears to be 

 conscious behaviour. The simplest technique in this respect is a study 

 of " preference " : if an animal goes towards light and avoids darkness, 

 or vice versa, it evidently can distinguish between them ; a similar 

 argument applies to a fish which swims towards a red rather than a 

 green li^ht. How far this conduct implies that the discrimination is 



