CH. I.] METAMORPHOSES OF INSECTS. 29 



be considered as feathers, from being affixed to the 

 wings by minute quills ; but Reaumur considers 

 them rather as scales, from being composed of 

 small membranous plates, having nothing in com- 

 mon with feathers. The number of these scales 

 upon the wings of a single butterfly is really won- 

 derful, Leeuwenhoeck having observed more than 

 400,000 upon the wings of the silkworm moth. 

 Their forms are likewise extremely numerous in the 

 different species. Indeed, a recent author has sug- 

 gested, in the llth number of the Magazine of Nat- 

 ural History, that they would afford very satisfac- 

 tory marks to distinguish the various species of Lepi- 

 doptera. From examinations, however, which we 

 have made for the purpose of verifying this remark, 

 we are convinced that the adoption of such a char- 

 acter would be impracticable, not only from the dif- 

 ficulty attending its examination, for no one would 

 be willing to spoil a fine specimen by rubbing the 

 down off its wings, but also from the circumstance 

 that, upon the different parts of the same wing of 

 a butterfly, several forms of scales are to be found. 

 These scales are arranged upon the wing in trans- 

 verse lines, the extremities of the scales of one 

 row resting upon and lying flat on the base of the 

 succeeding one, like the tiles upon the roof of a 

 house. According to Messrs. Kirby and Spence, 

 there appears to be a double layer of scales on both 

 sides of the wings, the under layer generally coi> 

 sisting of white ones. 



C2 



