240 
Psyche 
[Vol. 90 
df = 38 1 , p <0.00 1 ; small compared to large nest, t = 8.55, df = 542, 
p <0.001). There were no differences in egg development times 
between the medium and large nests (t = 0.57, df=839, n.s.). Aver- 
age larva development times varied from 13 to 18 days depending on 
nest (Table 1). The small nest and the large nest did not differ in 
average larva development times (t = 0.68, df=343 n.s.). However 
they both had longer larva development times than did the medium 
nest (small compared to medium, t = 4.84, df=322, p <0.001; 
medium compared to large, t = 8.89, df=611, p <0.001). Pupa 
development times averaged 13 days on all nests, and there were no 
significant differences among the nests. 
Development times of eggs, larvae and pupae did not vary signifi- 
cantly from one region of the nest to another on the small and 
medium nests (Tables 2-4, Fig. 2). In the large nest both eggs and 
larvae developed most quickly in region 1 (the oldest part of the 
nest), and most slowly in the edge regions 3 and 4 (Tables 2-4, Fig. 
2). In no case did region of nest explain more than 10% of the 
variance in development time. 
Date did not have a consistent effect on development times. Eggs 
developed more slowly towards the end of the season in the large 
nest, and more quickly towards the end of the season in the small 
nest (Table 5). There was no change in egg development time with 
date in the medium nest. Larva development time increased with 
date in both the large and the medium nest, and decreased with date 
in the small nest (Table 5). Development times of pupae did not 
change with date in the large nest, but decreased with date in the 
medium and small nests (Table 5). Date explained 35% to 51% of 
the variance in larva development times depending on nest. It 
explained smaller percentages of the variance in egg and pupa 
development times except on the small nest (Tables 2-4). 
Interaction between date and region of nest was examined using a 
2-way ANOVA (Tables 2-4). There was a significant (p <0.05) 
interaction between date and region of nest for egg and larva devel- 
opment times in the large nest that explained 7% and 5% of the 
variance respectively (Tables 2-4). This interaction thus explains a 
trivial amount of the variance in development compared to that 
explained by date. 
The effect of numbers of females tending the nest was found to 
be quite variable. Looking only at the medium and large nests, it 
