ENVIRONMENT AFFECTS OFFSPRING 483 



have more precise data as to the time when the sex of the tadpole 

 is unmistakably distinguishable. 



When a crowd of caterpillars are under-fed, there is an un- 

 usually large proportion of males (Landois, Treat, Gentry, and 

 others). But as it was shown long ago that the sex is determined 

 in the larva before it leaves the egg, the starving experiments 

 were irrelevant. They only show that there may be great 

 differences in the rate of juvenile mortality in the two sexes. 

 Thus Prof. Poulton points out in regard to the poplar hawk- 

 moth (Smerinthus populi), for instance, that the female cater- 

 pillars, being larger, require more food, and will therefore die first 

 when supplies are scarce. 



Nor is there agreement among the results of experiment. 

 Kellogg and Bell found that the sex of the silkworm is not 

 appreciably affected by the nutrition of the parents or even 

 grandparents. Cuenot found that the proportion of the sexes 

 in blow-flies, where its visible determination is later than in 

 butterflies, was not affected by what the larvae ate, or by what 

 their parents ate. 



What then is our conclusion in regard to the first theory ? 

 It must be admitted that there is no cogent evidence to show that 

 environmental influences operating on a developing organism may 

 decide what its sex is to be. Yet we should be slow to assert 

 that this is impossible. Consider, for instance, Nussbaum's 

 elaborate experiments on Hydra grisea, which he subjected to 

 varying nutritive conditions. In this species there are both 

 hermaphrodite and dioecious forms. Nussbaum found that the 

 optimum nutritive conditions resulted in predominance of female 

 polyps, and that groups wholly male could be produced by rela- 

 tive starving. From these experiments it seems that in Hydra 

 the nutrition of the body determines the production of ovary 

 or testis. 



There are analogous experiments in regard to some plants. 

 Prantl found that spores of the Royal Fern (Osmunda) and of 



