1953] 
Wilson — Caste Determination 
17 
from unfertilized eggs, because they are laid prematurely 
and at least some time before the first meiotic division. 
But whatever the explanation, the important point con- 
sidered here is that workers can develop from eggs of either 
of the two different sizes, whereas queens develop from at 
least the smaller of the two sizes. While it is true that the 
size of the Oecophijlla egg may not absolutely determine 
yolk content, yet it seems very probable that the differences 
in yolk content between two such markedly distinct egg 
sizes would greatly outweigh the very slight differences 
that must be assumed to determine caste in the worker- 
laid egg if Flanders’ hypothesis is correct. 
The preponderance of available evidence seems to in- 
dicate that the caste of female individuals is determined 
in the larval period, without regard to the original condi- 
tion of the egg. The work of Wesson (1940), Goetsch 
(1937) , and many others in the past half century has shown 
that larval feeding plays a major role; an excellent review 
of this work is presented by Light (1942-43) . Gregg (1942) 
has shown that the major-minor ratio in colonies of Pheid- 
ole morrisi Forel affects significantly the ratio of these 
castes appearing in the brood, and he has offered as the 
simplest explanation that this condition originates through 
trophic influences, possibly ectohormonal, on the develop- 
ing larvae. Recently, Brian (1951) has outlined his pre- 
liminary results from current work on caste determination 
in Myrmica ruhra L. The threshold for queen-worker 
divergence is reported to be in the larval period; in order 
to attain queenness an overwintering larva must reach by 
a certain time a weight of about 6.5 milligrams. It must 
then race to reach another threshold in order to become a 
fully developed queen; if it fails, it ends its development 
abnormally small or as an intercaste. Ledoux (op. cit.) 
presents convincing evidence that in Oecophylla longinoda 
the threshold for queen-worker divergence is in very early 
larval life, while the threshold for major-minor divergence 
is at some time in the second larval stadium. In the work of 
both Brian and Ledoux it is especially noteworthy that 
larval nutrition and larval size are apparently the major 
factors involved. 
