1981] 
Carlin — Polymorphism in Orectognathus 
237 
estimated total repertory size for medias was twenty-eight, the 95% 
confidence interval (27,33); for majors, twenty-seven, with a confi- 
dence interval of (24,37). 
Minor and media workers engaged in the same tasks with 
essentially similar frequencies, while majors, with a smaller reper- 
tory, also performed certain acts with quite different frequencies. 
Self-grooming was the commonest act in all castes. Allogrooming 
and regurgitation occurred freely among all castes, with a tendency 
among minors and medias to interact with their own class. After 
self-grooming, brood care and foraging were the most frequently 
performed acts in the minor and media repertories. An ant was 
scored as “foraging” any time it left the nest tube - an act that does 
not necessarily signify hunting for food. Though majors did “forage” 
by this definition, they captured almost no prey and returned none 
to the nest. “Processing”, in which workers tore at, dismembered 
and occasionally stung prey that had been brought inside the tube, 
was rarely performed by majors, despite the seeming usefulness of 
their heavy mandibles for such a task. 
The province of the majors was “guarding”: walking to the tube 
mouth and facing outward without setting foot on the container 
floor; after self-grooming, it was their most frequent act. A guarding 
ant might station itself at the opening for less than a minute or up to 
half an hour. That this is in fact a defensive behavior will be shown 
below. Minors and medias also guarded in large numbers, but less 
frequently than they foraged or attended brood. 
Nest maintenance was undertaken almost exclusively by the small 
size classes. Carrying refuse down the tube, to be dropped inside or 
just outside the entrance, was defined as “in-tube refuse removal”, 
while carrying trash out to corner refuse piles on the container floor 
(to which dead ants were also brought) was defined as “out-of-tube 
refuse removal.” “Manipulation of nest material”, that is, of the 
fibers of the cotton plug, may not be an actual maintenane behavior 
used in natural colony sites (under stones, in rotting wood); 
similarly, ants may lick the tube wall only to drink condensation on 
the glass, and not exhibit any such behavior in the wild. 
The division of labor among minor and media size classes, and 
the role of the major caste, were better elucidated by constructing 
polyethism curves, depicting the percent contributions of each caste 
to the total colony performance of behaviors (figs. 3 and 4). For 
simplicity, certain behavioral categories from the ethogram were 
