372 
Psyche 
[Vol. 85 
care of pupae. However, no greater proportion of egg-workers also 
tended for larvae and pupae than chance expectation. That is, indi- 
viduals specializing on eggs were not necessarily those specializing 
on pupae or larvae. The final set of behaviors is a large but 
loosely-connected cluster. Regurgitation, grooming, and antenna- 
tion were closely interconnected (Figure 2); these were also 
peripherally connected with inspecting the nest exterior and moving 
outside, since individuals returning from an outside foray elicited 
interest and grooming from nestmates. Similarly, LN and HM do 
not directly intersect, since they co-occurred no more often than 
chance expectation; these two behaviors were indirectly linked 
through ATW and RW (Figure 2). Behaviors in the roles of social 
interactions, physical maintenance, and provisioning were therefore 
not strongly separated. 
Thus analysis of sets of behaviors displayed over 30-minute peri- 
ods illustrated the expectation that individual workers tended to 
specialize, particularly within the brood care role. An ant grooming 
larvae was more likely to also carry or regurgitate to larvae than 
chance expectation. By a similar token, a worker guarding the nest 
Figure 2. Sets of behaviors that co-occurred within 30-minutes. Two intersecting 
behaviors were observed more often than expected by chance (x“ tests; for all 
intersections P <.05). Behaviors not intersecting occurred as often or less than 
random expectation. 
