A Classification of Leindopterous Larvse. 199 



Modification of the Tubercles, 



The primitive form of tubercle consists of a little chitinous button 

 on the skin bearing a single long hair. It is found in the less spe- 

 cialized families of all the groups discussed above, and exclusively 

 in the Jugat^e and in the Psychida?. When this form is present, 

 there are, in general, no other hairs on the body. 



The Frenatas, exclusive of the Psychidte, may be divided into 

 groups on the character of the tendency to coalescence or separa- 

 tion of tubercles iv and v, as I have remarked above. The first 

 group corresponds to Prof. Comstock's division of " Microfrenatae" 

 and " Generalized Frenulum Conservers," plus one family each from 

 his Zygjeuina and Saturnina.^ The second group corresponds to his 

 "Specialized Frenulum Conservers" and "Frenulum Losers" with 

 the exceptions just noted. If we shift the order of Prof. Comstock's 

 characters, and disregard the two exceptions, we may say that the 

 first group corresponds to the Generalized Frenatse, the second to 

 the Sjyecialized Frenatse. Now, in the lower families of these two 

 groups, we have the primitive form of tubercles, as already stated. 

 In the higher families of both groups we have at first a parallel line 

 of modification, which consists in the tubercles becoming enlarged 

 and many haired. Each hair arises from its own minute tubercle, 

 and the whole are borne upon an enlarged base or wart.^ The 

 arrangement of the warts is, at first, strictly that of the primitive 

 tubercles, and they correspond in number and equality of develop- 

 ment. The next step is a reduction in the number of tubercles, by 

 unequal development and final obliteration of certain ones. In the 

 higher Generalized Frenata?, tubercle iv has disappeared by coales- 

 cence with V, and in the higher Specialized Frenatse it becomes 



1 Prof. Comstock informs me in correspondence that he has somewhat 

 changed his views in regard to these two latter snperfamilies, so that he will 

 place them more in accordance with my own view, derived from studies on 

 the larvse. I omit reference to the Sesiidae, which Prof. Comstock places 

 among the specialized frenulum conservers, as he states in the text of " Evo- 

 lution and Taxonomy," that they probably belong with the Microfrenatje, a 

 conclusion which is abundantly justified by larval characters. 



^2 It is a striking fact that we do not find a series of intergrading forms 

 between the single-haired tubercle and the many-haired wart, though both 

 may occur in diflTerent genera of the same family. This can be understood on 

 the principle of discontinuous variation whicli is insisted upon by Bateson 

 ("Materials for the study of variation"). 



Annals N. Y. Acad. Sci., VIII, May, 1894.— 14 



