198 



A Classification of Lepidopterous Larvae. 



are all present, and the upper and lower of the posterior annulet 

 the two substigmatal tubercles may represent the substigmatal 

 ridge of the Tenthredinidte. The others have no honiologue in 

 these Hymen optera. 



In the FrenatiB we have the same arrangement still further 

 reduced. But here appears an important difiference to which I 

 have alluded above. In the Psychidai the three tubercles are 

 retained on the middle annulet, while both are lost on the poste- 

 rior one (see Fig. 3); the substigmatal tubercles are retained and 

 approximated ; the anterior one of the four on base of the leg seems 

 to have been moved up, forming tubercle vi, which is thus anterior. 

 In the other type, which includes all the other families of the 



Fig. 3. 



Fis:. 4. 



Fig. .5. 



Fig. 3. — Segment of an undeseribed Psychid larva from New Mexico. 



Fig. 4. — Segment of Margarodia quadrislir/initlis (from figure in Insect Life, vol. I, p. 24, 

 with addition of setse). 

 Fig. 5. — Arctian type of tubercles (diagram). 



Frenatse, the middle tubercle of the three on the middle annulet is 

 lost, but the upper on the posterior annulet is retained; the two 

 below the spiracle are retained also, as in the Psychidae ; but they 

 are either approximated (see Fig. 4) or separated (see Fig. 5), giving 

 us characters for the next subdivision. Of the four tubercles above 

 the base of leg, the p)Osterior one is moved up, forming tubercle vi. 

 These two types must represent dichotomous lines of descent. If 

 this is true, the Psychidse should be separated from the other 

 FrenatEe as a great group, equal in rank to all that follow.^ 



1 This can be done, in most cases, also by characters of venation ; viz., the 

 permanent cross vein in the Psychidse between veins viii and ix on primaries, 

 which exists also in the Hepialidse, and to which more classificatory import- 

 ance can be attached than has been done bv Prof. Comstock. 



