THE CANADIAN ENTOMOLOGIST. 3 



posed by Prof. Comstock, and the division into superfamilies which I have 

 suggested and which Mr. Grote has adopted with improved nomencla- 

 ture*, let us see where the families of " Bombyces " fall. 



From the Jugat/e, we find the Hepialidte only, the most highly 

 specialized Jugates in respect to the abortion of the mouth parts. From 

 the Fkenat^ as follows : — 



Stiperfamily Tineides. — The Eucleidas, Megalopygid^e, Anthroceridae 

 and Pyromorphida? from the apex of development along the main stem ; 

 the Psychidae, Lacosomidse and Heterogynidis, side branches, but all 

 specialized (the much specialized Sesiidas went with the Sphingidse), 

 and finally the Cossidse, a low type, but of large size. 



Superfamily Agrotides. — All the families, except those called Zygje- 

 nidse, the Agrotidfe and Geometridie, the two latter (with the exception of 

 the Notodontid?e) the lowest types in the superfamily, 



Superfamily Bombycides.—T\\t whole group. 



Superfamily Spliingides. — None, this group being recognized as 

 distinct, although the Sesiids and Thyridpe were associated with it. 



Superfamily Papilionides. — None. 



Thus it will be seen that the Bombyces consisted of the higher types 

 in all lines of development, regardless of relationship. If we imagine the 

 genealogical tree of Lepidoptera as growing upright from the ground, the 

 several branches and twigs representing the families and being of length 

 proportional to their degree of specialization, the old classification would 

 be represented by horizontal planes. The uppermost would cut off the 

 very summit of the tree, the Papilionides ; the next would take the next 

 succeeding top branches, perhaps the Sphingides, and the tip of a side 

 branch from the Tineid trunk, say the Sesiidae. The next cut might give 

 the old Zygsenidfe, consisting of some families from the Agrotid and 

 Tineid trunks, and the fourth cut is our Bombyces, taking branches of all 

 the trunks that are approximately equal in degree of specialization. The 

 base of the tree would comprise the rest of our old familiar families, the 

 Noctuida), Micros, etc. 



It is the aim of more recent work to follow the lines of genealogy, a 

 classification cutting our imaginary tree in vertical planes, including in 

 each group all families related to each other in the same line of descent, 

 regardless of degree of specialization. 



*Syst. Lep, Hildesii^. 1895. 



