449] LARVAE OF THE TENTHREDINOIDEA—YUASA 131 



of the family. There are, however, certain considerations which suggest 

 a possible relationship between this family and the Siricidae. Morpho- 

 logically the oryssid larvae are more closely related to the apodous boring 

 larvae of the Cephidae, Xiphydriidae, and Siricidae than to the polypodous 

 free-living larvae of the Xyelidae and Tenthredinidae. The siricid larvae 

 are more closely related to those of the Oryssidae than to those of the 

 Cephidae and Xiphydriidae. The parasitic habit also suggests a closer re- 

 lation to the wood-boring larvae since it is more plausible to imagine the 

 possibility of a wood-boring larvae becoming parasitic on other wood- 

 boring insect larvae under some unknown but not entirely inconceivable 

 circumstances than to imagine the development of a parasitic habit de novo 

 in free-living leaf-feeders. Since the oryssid larvae are parasitic on the 

 larvae of Buprestis inhabiting plants which are also infested by the larvae 

 of the Siricidae, if any transformation of habit of the larvae has taken 

 place, it is more natural to expect the larvae of the Siricidae or some 

 siricid-like insect to become parasitic than any other larvae. The recent 

 investigation by Baumberger (1919) on the role of microorganisms in 

 the physiology of insect nutrition offers a valuable suggestion in regard to 

 the possibility of radical changes in food habits. For these reasons it is 

 considered reasonable to ascribe a common progenitor to the Siricidae 

 and Oryssidae, at least for the time being. It may be added that it is 

 not entirely unreasonable to assume an independent line of evolution 

 for the Oryssidae apart from all other Tenthredinoidea and consider 

 this family as having no close relation to any of the modern families of the 

 Tenthredinoidea. In that case, the Oryssidae must have arisen from the 

 ancestral stock before the Pamphiliidae and Xyelidae had their origin. 

 There is, however, no clear evidence in support of such relation and 

 since the relation is reasonably explained by associating the Oryssidae 

 with the Siricidae, the former is considered the most highly specialized 

 family of the Tenthredinoidea with a common origin with the ancestor of 

 the Siricidae. 



The conclusions on the systematic position and relationship of the 

 different families of the Tenthredinoidea based exclusively on larval 

 characters and derived entirely independent of the opinions of the spe- 

 cialists who have paid more attention to the adults are of necessity not 

 the final words on the subject. The true significance of such conclusions 

 lies in their complemental and collateral value. It is interesting on this 

 account to compare the writer's opinion with the conclusions of the 

 modern authorities on this group of the Hymenoptera. 



The relationship suggested here supports in its essential points the 

 three more important systems of classification proposed by Konow (1905), 

 MacGillivray (1906), and Rohwer (1911). MacGillivray considered the 

 Xyelidae the most primitive because of the venational character but 



