THE EUMENIDAE 309 



channel were four pockets, all alike and all containing cater- 

 pillars of the species used exclusively by O. geminus. (See 

 figure in lower right-hand corner of fig. 57). There were 

 fifty-two caterpillars in all, or an average of thirteen to a 

 cell. They were writhing briskly and seemed very much 

 alive, but they were unable to walk. The larvae were small 

 but apparently healthy. Another of the sealed burrows near 

 by had precisely the same external appearance, but had 

 only one cell below ground (upper right-hand figure). This 

 contained eleven caterpillars of the same species of Loxo- 

 stege as the others, and the egg was hanging from the wall 

 by a short but strong silken thread. 



One certain incomplete turret surmounting an open hole 

 was watched for three days, but it neither increased nor de- 

 creased in height, and since we saw no signs of activity 

 about it, we dug it up. The main channel went down as 

 usual for an inch, where it branched and terminated in 

 two pockets (see fig. 57, lower left figure). One of these 

 contained a large O. geminus larva and was well sealed 

 with a plug of mud in the mouth of the chamber, while the 

 second was open and contained only four caterpillars, and 

 apparently was unfinished. Probably the mother wasp had 

 met with an accident which had prevented her from finish- 

 ing the nest, but the incomplete nest gives us the evidence 

 that the turret is left standing until the pockets below are 

 all finished. 



Hungerford and Williams have some evidence that the 

 cells are often used more than once as brood chambers. 

 They also illustrate many galleries and cells to one nest, 

 apparently made by one mother. The most complex nest 

 in our experience is that shown in figure 57, in which the 

 cells diverge from the central gallery in a shower. 



When we opened one new nest, we were surprised to find 

 in its single cell one lone caterpillar and the mother wasp 



