318 WASP STUDIES AFIELD 



eter at the top. The hole goes straight down for one-quarter 

 to one-half an inch and then widens to form a spherical 

 chamber a half -inch in diameter (see figs. 60, 61 ) . From the 

 bottom of this the gallery continues, the same diameter as 

 before, for a short distance and expands into a second cham- 

 ber similar to the one above. The lower cell is filled with 

 caterpillars, the egg is left swinging by a delicate thread 

 from the wall of the cell, and the mouth of the compartment 

 is sealed up with a plug of mud ; then the upper cell is simi- 

 larly filled and sealed with a mud plug in the neck of the cell. 

 Finally a closure of mud is inserted in the top of the hole 

 flush with the surface, sometimes artfully constructed so 

 as to be indistinguishable from the surrounding earth, but 

 more often there is the saucer-like depression, caused by bit- 

 ing out earth to fill the hole, which Iseley has so well de- 

 scribed. But right here is a very interesting and curious 

 point : the entire gallery from the top of the upper cell to 

 the surface plug is not filled in solid for its entire length, as 

 are the holes of other burrowing wasps, although often it 

 is a short channel (fig. 61) ; but it has only the firm plug 

 at the top and bottom and an air-chamber between. The 

 purpose of this arrangement leads one into pretty specula- 

 tion. Does the air-chamber in the channel help to maintain 

 a more uniform temperature in the cells beneath? The 

 depth is sometimes so slight that the difference could not 

 amount to much at any event. 



Iseley, too, finds that the nests are vertical with one or 

 two cells. He describes one exception, however, which is 

 very surprising and interesting: a colony of eight nests 

 built in the face of a vertical clay bank, with from three 

 to seven cells to each burrow. We have never yet found 

 more than two cells except in one nest which was appar- 

 ently abnormal, which had an extra (small) cell between the 

 two normal ones (figure 62) ; neither have we found these 



