THE CHRYSALIS. 47 



posite end and the secondary channels which 

 course along the body. 



The abdomen is always of a subconical shape, 

 tapering toward the hinder extremity, and more 

 or less constricted where it joins the thorax. The 

 spiracles remain in the same position as in the 

 caterpillar, excepting on the first segment, which 

 is so far covered by the folded wings that the 

 spiracle is lost, and the segment reduced to a 

 small dorsal plate ; at the extremity of the body, 

 however, an apparently new structure is devel- 

 oped, the cremaster, or anal button, a slender 

 tubercular prolongation at the extremity of the 

 body. Kiley, however, has shown* that it is not 

 a new structure, but that it corresponds to the 

 anal plate (or hindmost segment) of the cater- 

 pillar. Kiinckel indeed contendsf that it repre- 

 sents the soldered anal prolegs of the caterpillar ; 

 but Biley clearly proves that these are repre- 

 sented in the chrysalis by what he calls the sus- 

 tentors, ridges on the ventral surface which ter- 

 minate anteriorly in little knobs or hooks, and 

 play an important part in pupation. The form 

 of the cremaster varies considerably in different 

 groups, but always bears on its tip, or under sur- 

 face, or on both, a crowd of microscopic recurved 

 hooks [Fig. 64], which, in chrysalidisation, are 



* Amer. Entom., iii., 162-167, July, 1880. 



f Comptes Rendus, xci., 395-397, August, 1880. 



