CONCLUSION. 233 



once of the regular metliod or she refuses to proceed. Again, 

 we see Occyheliis scratching open her nest while on the wing and 

 ent(!ring at once with the fly held tightly in her legs. Each 

 way is characteristic of the species and wonld be an important 

 part of any definition of the animal based upon its habits. 



The general style of the nest depends upon instinct. Tri/- 

 poxylon uses hollow passages in trees, posts, straws, or brick 

 walls; Diodontiis americamts, a member of the same family, 

 always burrows in the ground, as do Bemheoe, Ammophila, and 

 Sphex. In the case of Trypoo'i/Ion the passage may be ready 

 for use or may require more or less preparation; the instinctive 

 part is the impulse that impels the insect to use a certain kind 

 of habitation. Any one familiar with T. rubroc'mctum would 

 never look for their nests in standing stems or under stones; to 

 use Mr. Morgan's test, he would be willing to bet on the general 

 style of the dwelling place. All of these acts are similarly per- 

 formed by individuals of the same sex and race, not in circum- 

 stantial detail but quite in the same way in a broad sense. Var- 

 iation is always present but the tendency to depart from a cer- 

 tain type is not excessive. In the drawing of the nest of 

 Cerceris nigrcscens the burrow is seen to be tortuous, this style 

 of work being common to many species in the genus and very 

 characteristic. ISTo SpJiex nor AmmopliiJa constructs any such 

 tunnel. The adherence of all the members of a species to a cer- 

 tain style of architecture, is, then, due to instinct. 



The spinning of the cocoon, in those species in which the larva 

 is protected in this manner, and its shape, are instinctive. "We 

 find that closely allied species in the same genus make very dif- 

 ferent cocoons as is seen in T. ruhrocinctum and T. hidentatum. 

 Some wasps never cover themselves with a cocoon, as in the Au- 

 stralian species Alastor eriiirgiis and Ahispa splendida* It 

 is a well known fact that silk w^orms sometimes omit the spin- 

 ning of a cocoon; but this does not affect the argimient since 



*Nests and Habits of Australian Vespidae and Larridae, by Walter 

 "W. Froggatt, Proceedings Linnean Society of N. S. W., Vol. IX., 

 Series 2nd. 



