570 THE EYE IN EVOLUTION 



the main chamber, is offered the choice of two stimuli ; these, for example, may 

 be light stimuli made up of two illuminated milk -glass panels set into hinged 

 doors and lit from behind so that they can be suitably varied in intensity, hue 

 or saturation. Either of these the animal can open to receive a reward (food) 

 or punishment (an electric shock). Trained initially to go towards one (the 

 positive) of two well -differentiated alternative stimuli and to avoid the other, 

 the negative stimulus is approximated progressively to the first until the limit 

 of discrimination is reached. Thi'oughout the exjaeriment the relative positions 

 (right or left) of the two stimuli are randomly alternated, while other stimuli 

 (olfactory, etc.) are eliminated as by j^lacing similar food in each box, that in 

 the negative box being inaccessible. Such training techniques, of course, are 

 laborious, several hundred " runs " being usually required in each experiment ; 

 moreover, they are time-consuming for much cannot be accomplished at one 

 session lest fatigue be induced or interest lost ; and they are restricted to species 

 which are relatively intelligent and docile, for a stupid or an untrainable animal 

 or one that gets cross or sulks is useless. 



It is also to be remembered that any response of this nature made 

 by an animal depends upon complex factors ; few stimuli are in fact 

 simple, most involve more than one receptor, and all responses are 

 complicated by mutual excitations and inhibitions, for the animal 

 reacts not to one stimulus alone (such as food) but to a complex 

 situation wherein each stimulus must be differentiated against a 

 changing background and varies with past experience and its present 

 psychological state. Even in the most adequately controlled experi- 

 ments in the laboratory an ideal environment can rarely be realized. 

 The very fact of the artificial isolation of the stimulus is outside the 

 animal's natural experience and thereby something important in the 

 experiment is lost. It follows that the results of such analyses can be 

 accepted only with reservation ; indeed, any claim that a scientifically 

 exact appreciation of the physiology or psychology of any animal can 

 be based on conditioning experiments is illusory. 



Within these limits, the method undoubtedly produces results in 

 terms of sensational responses of greater reliability than any other 

 and forms the best means of analysing the nature of the sensation 

 concerned. Considering these difficulties, however, as well as the varia- 

 tion in psychology between different members of the same species and 

 the probable differences in apperception and interpretation between any 

 species and our own, it is not surprising that the results thus obtained 

 have often been inconsistent. 



THE LOWER LNVERTEBBATES 



PROTOZOA. We have already seen that Protozoa exhibit fixed 

 reactions to a variety of " sensory " stimuli — light, heat, gravity, 

 contact, electrical shock — the only observable response being a tropism. 

 We ha^ also seen that there is no observable difference in behaviour 



