PROCEEDINGS OF THE NATIONAL MUSEUM. 



Egijs. — The eggs are circular and mucli flattened, tbe micropyle in 

 the center of the upper side, ribbed and with faint cross-striae. The 

 characters are the same as in the English species, as described by Dr. 

 T. A. Chapman. So few of our species are known that it is impossible 

 to generalize on them and useless to specially discuss the few cases. 



Larvae. — The larvae of this genus are especially interesting from the 

 wide range of the characters of the hairs. The range covers forms with 

 single setae, with true many-haired warts and those in which this pri- 

 mary coating is almost completely replaced by a secondary one of the 

 hairs growing from the skin. This is as great a range as is shown by 

 any of the families of the whole Bombyces. All the species may be 

 traced back to a ground form, which may be described as a typical 

 Noctuid in which warts have been developed, accompanied by a tend- 

 ency to the production of secondary hairs. The species have made 

 the most of this comprehensive structure. In the single-haired forms 

 degeneration has occurred, as may be seen by the presence of true warts 

 in the earlier stages. The Arctian-like forms have iireserved the warts 

 intact without secondary hairs. Another series has developed the sec- 

 ondary hairs at the expense of everything else. 



The species divide into four well-marked groups. The first com- 

 prises those with abundant secondary hairs. All the American species 

 have a series of dorsal tufts or pencils, variously modified, and in some 

 cases absent in the last stage. The coloration is very diverse. (Group 

 americana.) 



In the second group the warts are small, the secondary hairs weak, 

 but present in some degree. There is a wide range in the coloration of 

 these forms at maturity, where various adaptations and cases of mim- 

 icry occur, though before the last stage very diverse forms are often 

 strikingly alike. (Group lobeliae.) 



The third grgup have warts in the early stages in all the American 

 species so far known, which may degenerate to single hairs in the last 

 stage except in the case of tubercle VI. Secondary hairs are absent. 

 In two cases the hairs are spatulate. This group appears at first 

 sight a compact one, but from the results obtained from the study of 

 the male genitalia of the moths it appears that the single-haired 

 species are really derived from two diftereut sources within the genus. 

 One section (group hamanicUs)hiis the structure of retardata, while the 

 other (group persuasa) seems to have arisen from a point near cuspis 

 and tridens. The structure of Group IV may be derived from Group 

 III by the loss of the secondary hairs, and the only unexpected point 

 that the male genitalia indicate is that the single-haired structure has 

 been developed twice. The loss of secondary hairs would be expected 

 to accompany this change. (Groups ^;ersMasa and hamamelis.) 



The fourth group is more compact. The warts are well formed and 

 bear si)reading tufts of hair, in some cases supplemented by bunches 

 of flue feathery hairs which in luteicoma become markedly developed. 



