FORMATION OF ASSOCIATIONS IN THE MAY-FLY 19 



different from those of H. inter punctata. However, she obta ned 

 such results with only a single specimen and although attempts 

 were made to investigate the learning capacity of another larva 

 of the same family, she was unable to get any further evidence 

 of the formation of associations. " Seit einigen Wochen befindet 

 sich wieder eine Libellenlarve in meinem Aquarium, doch habe 

 ich bei ihr noch nicht viel Erfolg mit meinen Fiitterungsver- 

 suchen gehabt." 



CONCLUSION 



In conclusion it might be said that we have three types of 

 experiments with large numbers of H. inter punctata nymphs 

 which clearly show the power of formation of associations 

 by three distinct processes. In case of the investigations on 

 the reactions to objects it was found that the insect grad- 

 ually increased the distance at which it responded positively 

 to objects in an evenly lighted environment. In another set 

 of experiments it gradually inhibited its usual negative response 

 to light and proportionately reinforced its reaction to an object 

 against the rays of light. It is also important to note the fact 

 that this inhibition in phototactic response was only tempo- 

 rary; it occurred only when the object was present, for wdienever 

 the object was removed the nymph would manifest its usual 

 dislike for the bright light. In case of the experimental work 

 on the manifestations of fear through previous experience we 

 have results which can hardly be interpreted in any other way 

 except as a clear example of the associations of my movements 

 with pain, for why should the disturbed specimens show such 

 obvious signs of agitation when the undisturbed group evinced 

 no indications of fear under otherwise practically identical con- 

 ditions. Lastly, we have results which seem to indicate quite 

 conclusively that the insects associated my presence with food. 

 These results, again, become more convincing when compared 

 with the unvarying individual records of the group not so 

 trained. The foregoing conclusions are not based on the results 

 tabulated in this paper alone, but on the results of a large 

 number of similar experiments performed with many large 

 sets of specimens and continued for two years during the var- 

 ious seasons. 



My thanks are due to Professor S. J. Holmes for his kind 

 suggestions and criticisms in preparing this paper. 



