30 DONALDSON BODINE. 



lies which " are supposed to retain more nearly than any other 

 Frenatae the form of the primitive Frenatse, those that were the first 

 to appear on earth," and the Specialized Frenatse, including those 

 " that depart more widely from the primitive type of Lei)idoptera, 

 being more highly modified for special conditions of existence." The 

 first group is a small one and comprises only five families, none of 

 which are large. 



The Generalized Frenat^. 



I believe the evidences of the antennae of this group agree with 

 those of the wings as given by Professor Comstock. The five fami- 

 lies represent lower branches from the primitive stem, and so have a 

 bond of connection in that none of them have departed very widely 

 from the primitive type. While the evidence is not conclusive 

 enough to show that any two of the families are genetically related, 

 except through the common stem, there are some indications in the 

 structure of the antennae that the Megalopygidse and Eucleidse are 

 more closely connected than any other two families ; while the Py- 

 romorphidie seem to represent a perfectly distinct line of descent. 

 It is evident that if each of these families represents a distinct line 

 of development, the grou]) as a whole will be a homogeneous one 

 only in the sense of containing those forms which are very gener- 

 alized, and that is the character of the group as given by Px'ofessor 

 Comstock. That there is a wide difierence between the degree of 

 specialization attained by the members of these families and those 

 of any others is quite certain. I do not think any one familiar with 

 various forms of antennae would fail to pick out those belonging to 

 this group. The extent to which the scape and pedicel are clothed 

 with scales ; the method of insertion of the scales more or less ob- 

 liquely to the surface, and the irregularity in their arrangement; 

 the scarcity of pits in most forms and the simplicity of their struc- 

 ture ; the very slight extent of the ventral expansion of the seg- 

 ments ; all are characters which, though perhaps not so constant in 

 all cases as to permit of exact definition of hmits, will yet enable 

 one to distinguish these families without much difficulty. 



The Megalopygid.e. — The members of this family have very 

 generalized antennae. The segments are short and numerous. Me- 

 <l<(loj)ij(i<i crixpata has sixty-one composing the clavola. Tlie ])ecti- 

 nations arise from the extreme ventral aspect of the shaft and the 

 bases of each pair are closely apposed. They are long and well pro- 



