10 Journal New York Entomological Society. [vol. vr. 



ENDROMIDID/E. 



The subprimary tubercles of the larva (3 and 5) in stage I are 

 wanting on the thoracic segments (I to III). On the abdominal seg- 

 ments the subprimary tubercle 6 is also wanting. In the " Saturnii- 

 den," page 6, 1 figure abdominal segment, but the bristles marked " 6 " 

 are too low down for this and evidently belong to 7. I indicate this 

 doubt in the text (p. 5). On the same segments the tubercles 5 and 4. 

 are separate. There appears to me no insuperable objection to the 

 view, that Endromis represents an isolated form of the Hawk Moth 

 stem, separating soon after this stem had emerged from the Tineid 

 trunk. The retention of the intercostal vein is then an independent sur- 

 vival of a character shared by both when the Endromid branch made its 

 separate way. Dyar has enabled us to show that the type of the En- 

 dromid larva is opposed to the Saturnian and the latter to the Sphing- 

 oid type. The neuration excuses me in considering Endromis as an aber- 

 rant Sphingoid type. It is not improbable that the larva of an existing 

 generalized Sphinx might throw some light on the matter through a 

 comparison of parallel stages. When we turn from the larva to the 

 pupa, we find that the segments of the abdomen are capable of move- 

 ment in Endromis and by their aid the pupa is forced out of the cocoon, 

 before exclusion, as in Anthrocera, Cossus and the Tineides gen- 

 erally. Preparations before me of Endromis and Anthrocera, hatched in 

 my breeding cages, show a striking similarity in this habit. The Sphing- 

 idge seem to have the habit also, inasmuch as the naked pupa is stated 

 to wriggle its way to the surface of the ground to allow the escape of the 

 moth. Such species as transform on the surface within a slight cocoon 

 have not, to my knowledge, been observed as to this point. This habit 

 indicates a direct connection of the Sphingides with the Tineides. 

 The links between Endromis and Sphinx appear to have dropped out ;. 

 also those by which we might more surely trace the relationship between, 

 adult forms of the Sphingides and Tineides. Nevertheless, I call atten- 

 tion to the fact that the Anthroceridse represent a Tineid branch possi- 

 bly related to the stem which threw off the existing Sphingides. 



Leaving these characters, we will consider the neuration. And first 

 the shape of the wings is modified, and this probably in accordance 

 with the method of flight. Disposed as I am to consider the Sphingides- 

 and Saturniades as parallel groups, each specialized in a different way, 

 and the Saturniades unquestionably the more highly so, I would com- 

 pare the Endromididse with the Saturniadse and the Sphingidae with the 



