Effects of Castration in hi sects 39 1 



suggested. The ontogenetic explanation suggested by Brues is 

 one of these, implying that a red stage precedes the brown or 

 black of the mature form of dark species like P. metricus. This 

 is borne out by the development of the color pattern in such species. 

 On this view st\lopization inhibits color development in an 

 ontogenetically and presumably' therefore in what corresponds to 

 a phylogenetically earlier stage. A second explanation is, how- 

 ever, suggested by Miss Enteman's studies. These tend to show 

 that the dark-colored races or species of Polistes are due to cold 

 and moisture, the lighter yellow and red forms to heat and aridity. 

 This seems to be clearly indicated in the distribution of the spe- 

 cies, e.g., in such extreme forms as the yellow P. texanus and the 

 black canadensis. It is possible, therefore, that the erythrism of 

 stvlopized P. metricus, which in normal coloration is closely 

 related to P. canadensis, is due to withdrawal of water from the 

 tissues by the developing parasites. This does not contradict the 

 ontogenetic and phylogenetic explanations but supplements them, 

 if we suppose that the primitive yellow or red color cannot pass 

 on to the piceous or black stage unless the tissues contam a suffi- 

 cient amount of water. Miss Enteman has shown that the piceous 

 or black color is in the form of pigment granules in the chitinous 

 cuticle of the wasp's integument, whereas the yellow is deposited 

 in the hypodermis. Erythrism is probably due, therefore, to a 

 diminution in the cuticular pigment which permits the yellow hy- 

 podermal pigment to shine through. As both kinds of pigment 

 are the result of metabolism in the pupa, we can see how a disturb- 

 ance of metabolism either through withdrawal of water by the para- 

 sites or through other causes might lead to the deposition of a 

 smaller amount of the black pigment and hence to erythrism. 



It is more difficult to account for the absence of all modifications 

 of the secondary sexual characters in stylopized Polistes, when 

 such modifications are so evident in Andrena. We may, perhaps, 

 account for this difference on one of the following hypotheses: 



I. As will be shown in the sequel, complete extirpation of the 

 gonads in young larval insects, has produced in the few species 

 on which it has been performed, no appreciable effects on the de- 

 velopment of the secondary sexual characters. This indicates that 



