THE EVOLUTION OF INSTINCT 129 
to degenerate into an almost shapeless mass devoted entirely 
to feeding and the production of eggs. If now the develop- 
ment of some of the females could be checked, by insuffi- 
cient nutrition at an earlier period of metamorphosis we 
should have a caste of sterile forms more highly organized in 
structure and of more varied instincts than the fertile females. 
If would certainly be an error to conclude that the superiority 
of the sterile individuals was due to their having independ- 
ently acquired, since the differentiation of the caste, charac- 
ters which were not a part of the legacy of their parents. 
This is, I think, a fair statement of the neo-Lamarckian 
side of the question. Granting that the theory is otherwise 
acceptable the Lamarckian doctrine is still open to serious ob- 
jections when applied to the instincts of neuter insects. 
It would compel us to assume that all the adaptive struc- 
tures and instincts of the worker bee were but the survival of 
a more primitive period of social evolution before the complete 
separation of the fertile and sterile castes, when presumably 
they might have been evolved through the inherited effects 
of habit. And not only should we have to assume that there 
had been no improvement in the non-reproductive activities 
since the queen had ceased performing them, but the theory 
would lead us to expect a certain amount of deterioration 
through disuse. It certainly cannot be supposed that 
during a period long enough for the queen to lose her wax 
glands, pollen basket, and all her instincts for gathering 
honey, making comb, and caring for the young, the structures 
and instincts of the workers should have suffered no degenera- 
tion. Instead of this there has undoubtedly been a decided 
improvement. Nowhere in social bees do we find so high a 
degree of structural adaptation, such varied activity, such 
perfect workmanship in the construction of comb with its 
varied cells for queen, drone and worker, and such complex 
9 
