4431 LARVAE OF THE TENTHREDINOIDEA—YUASA 125 



with the thoracic legs and homologous with the abdominal appendages of 

 generalized insects. There is no reason for considering the larvapods of 

 the Xyelidae as embryologically and morphologically different from those of 

 Hylotoma, consequently the larvapods of the Xyelidae must be the 

 true appendages of the abdomen; and since the larvae of this family 

 are provided with the maximum number of larvapods, they must be 

 considered as the most generalized of the Tenthredinoidea. The Ten- 

 thredinidae with six to eight pairs of larvapods and certain other mor- 

 phological and biological characters are unquestionably related to the 

 Xyelidae and probably represent a line of evolution from a xylelid-like 

 ancestral stock. Among the Tenthredinoidea with apodous larvae, the 

 Pamphiliidae, with a pair of segmented subanal appendages, is undoubtedly 

 the most generalized of all five families. The origin of the Pamphiliidae 

 is consequently an important question. For the reasons already stated 

 in connection with the larvapods, the progenitor of the Hymenoptera has 

 been considered as possessing a pair of subanal appendages on the caudal 

 segment of the body. In this character as well as in all others the Pamphilii- 

 dae approach most nearly the primitive condition and, except for the 

 absence of larvapods, unquestionably represents the most generalized 

 condition found in the Tenthredinoidea, ovitranking even the Xyelidae. 

 The loss of larvapods in this case is just as difficult to explain as the loss 

 of subanal appendages in the case of the Xyelidae. These structures, the 

 larvapods and subanal appendages, must have been lost during the course 

 of phylogeny since the progenitor undoubtedly possessed both of these 

 structures, and these two families, in spite of their generalized conditions, 

 must represent the end-products of evolution in their particular lines. 

 It is, then, natural and proper to assume that there have taken place two 

 distinct lines of development from the ancestral type of the Hymenoptera. 

 In the one, the specialization consisted in the suppression of the develop- 

 ment of larvapods, as in the Pamphiliidae, and in the other in the suppres- 

 sion of the development of subanal appendages, as in the Xyelidae. These 

 two families, then, represent two independent lines of evolution and are 

 the most generalized families not only of the Tenthredinoidea but of the 

 Hymenoptera. Whether the Xyelidae is more generalized than the 

 Pamphiliidae, or vice versa, must, from the very nture of the case, remain 

 a question till the advancement of our knowledge shall perhaps make the 

 answer possible. There are, however, a few things that should be pointed 

 out regarding this question. If the suppression of the development of 

 larvapods is considered of equal phylogenetic significance with the sup- 

 pression of the development of the subanal appendages, and if the head 

 and the appendages of these two families alone are compared, there is no 

 doubt that the Pamphiliidae are more generalized than the Xyelidae. 

 But since the subanal appendages are true abdominal appendages homo- 



