60 



DWIGHT E. MINNTCH 



evaluations of groups of responses may be compared as well as 

 the mere numbers of responses. 



Fig. 1 Diagrams showing the proboscis in the unextended condition and in 

 various stages of extension: 1, unextended; 2 and 3, partial extensions which 

 would be weighted at 0. 5; 4, partial extension which would be weighted at 1; 5, 

 complete extension, which would be weighted at 1. 



In the present experiments every trial was immediately pre- 

 ceded by a preHminary trial held under the same conditions. 

 The purpose of the preliminary trial was fourfold : 1) to overcome 



