178 DWIGHT E. MINNICH 



in the hope of counteractmg the possible effect of the death 

 feint. I am certain, however, that with all these precautions, 

 the death feint together with the effect produced by holding the 

 animal accounted for many failures to respond to chemical 

 stimulation. This opinion was strengthened by the fact that 

 several animals which had failed to respond during a trial did 

 so immediately upon being released from the holder. 



Since the data to be presented in this paper have been obtained 

 solely through a study of the proboscis response and the con- 

 ditions which effect it. a more detailed account of this reaction 

 is necessary at this point. In a few instances the proboscis 

 was completely uncoiled only to be recoiled shortly afterwards. 

 As a rule, however, once the organ was extended, the butterfly 

 actively explored the substrate with it, and upon contact with 

 the stimulating substance began to feed. The animal was 

 generally removed as soon as the proboscis was completely 

 extended, in order that feeding might not intei'fere with the 

 sensitivity to food substances. So strong was the response^ 

 however, that frequently, as the animal was being carried back 

 to the cage, the proboscis would remain extended and continue 

 to probe the empty air in a ^'igorous mamier. 



In many cases the extension of the proboscis was not complete. 

 The degree of these partial extensions varied from slight, and 

 occasionally, barely visible jerks to almost complete extensions. 

 These were easily observed in the great majority of cases, and 

 indeed it was only in a few trials of a very few indi\'iduals that 

 the barely visible jerks were noted. Frequently the proboscis 

 was partially uncoiled and, like the hair-spring of a watch, was 

 kept springing back and forth for some seconds. Occasionally, 

 this culminated in a complete extension, but more often it was 

 followed by a complete recoiling of the organ. 



The duration of each trial was one minute, unless the proboscis 

 was completely extended before that interval had elapsed. In 

 the latter case, as stated above, the animal was immediately 

 removed to prevent feeding. If there occurred any ^'isible 

 movement of the proboscis whatever during the trial, it was 

 counted as a response. If, on the contrary, there was no visible 



