380 PROCEEDINGS OF THE AMERICAN ACADEMY. 



animal with eyes comparatively so well developed we might expect 

 more evidence of image-formation to be apparent in its reactions to the 

 lights of different areas. An inspection of Figure 13, however, shows 

 that the " mode " for the negative responses of the cockroach to unilat- 

 eral light lies only 20° to 4(J° from the indift'erent position, in which 

 respect it agrees more closely with Oniscus (compare Figure 12) and 

 differs markedly from the Tenebrio larva (Figure 11), although its eyes 

 are undoubtedly vastly better adapted to ordinary vision than those of 

 the mealworm. This means that the cockroach is less responsive to 

 directive light than the mealworm, or at any rate that its responses are 

 less definite and constant. Another factor which must be taken into 

 consideration is the greater rapidity with which the former animal 

 travels, so that in a given distance the light acts upon it during a much 

 shorter interval than is the case with an animal which moves more 

 slowly. Under these circumstances the records would tend to be less 

 divergent ; and undoubtedly this was an important factor in the reac- 

 tions to both lights, where, it will be observed, the number of records 

 in the class was very large (Figure 13). The same factor would have 

 a tendency to mask any differences there might be in the reactions to 

 the two lights. This fault was partially corrected in the experiments 

 by taking the records for the cockroach when it crossed the outer circle 

 (15 cm. radius), whereas the records on Tenebrio were taken when it 

 reached the innermost circle (5 cm. from the starting point). 



The conclusions to be drawn from the experiments on the cockroach 

 may, then, be stated briefly as follows : Periplaneta americana reacts 

 negatively to directive light as used in these experiments in an excess 

 of about 50 per cent (42 per cent to 55 per cent) of its responses. It 

 possesses relatively large eyes, which, one would suppose from their 

 structure, were capable of much better image-formation than those of any 

 of the other forms so far employed ; but the results of the experiments 

 fail to confirm this point. The explanation is probably to be looked for 

 in the fact that on account of the influence of other factors the reactions 

 to light are masked, and probably to a certain extent inhibited. 



6. Moui'ning-cloak Butterfu ( Vanessa antiopa Linn.). 



The work of Parker ( : 03) on Vanessa naturally suggested that animal 

 as a favorable subject for these experiments. The following conclu- 

 sions reached by Parker (: 03, p. 4G7) are of importance in the present 

 connection : 



"3. V. antiopa creeps and flies toward a source of light, that is, it 

 is positively phototropic in its locomotor responses. 



