A WEEK WITH A MINING EUMENID. 159 



reached the point where they were bent abruptly to form a support. 

 A series of sub-parallel, semi-elliptical wires were constructed 

 across this fan-like expanse of wires (Fig. 2). 



A number of these interference mazes were used. They were 

 all similar, but no two were identical (Figs. 2 and 3 are from the 

 same maze and Figs. 4 and 5 are from another). The only points 

 invariably adhered to were two : first, in the immediate vicinity of 

 the nest the wires were so arranged as to make it difficult for the 

 wasp to get a flying start with her pellet of mud; second, the 

 meshes were made sufficiently open for the wasp to fly through. 

 Two mazes were used at a time. Each experiment extended over 

 a period of from one to two hours. Three days were devoted to 

 these experiments. The wasps, which invariably spent the night 

 in the nests, became active between 7:30 and 8:00 A.M. and con- 

 tinued to work until about sundown. Each day the experiments 

 were begun as soon as the wasps became active. On the first day 

 they were continued until they became inactive ; on the two other 

 days I stopped work between four and five in the afternoon. Since 

 two mazes were used at a time, this gave time enough for fifteen 

 experiments. 



The results of these experiments were conclusive and positive 

 and sufficiently uniform to be epitomized in a single paragraph. 

 In each case the wasp flew away with a few pellets (usually less 

 than six) and then began to walk. In the midst of this uniformity 

 there was individuality. The majority of the wasps walked from 

 the nest, deposited the pellet (Fig. 2, /), and then walked back to 

 the nest ; some walked from the nest, deposited the pellet, then flew 

 out of the maze and back to the nest. One walked to and from 

 the pellet pile most of the time, but occasionally it would fly and 

 scatter a pellet. Another would walk from nest to pellet pile and 

 back until it came time to get some more water, then she would 

 fly with the pellet out of the maze, drop it, and then go for water. 

 Thus we have conclusive experimental proof that the walk-and- 

 deposit-the-pellet behavior of these wasps has been derived from 

 the fly-and-scatter-the-pellet behavior in response to an environ- 

 ment which made flying with a load unpleasant. 



Occasionally, in arranging the maze over a certain nest, the 



