156 TRANSFORMATIONS OF INSECTS. 



of incubation ; but in some species closely allied to the silkworm 

 two or three broods are obtained every year, and the embryonic 

 state is then very short. 



How different to this metamorphosis is that of the Scsia 

 apiformis, the caterpillar of which lives for two years, and then 

 changes into a chrysalis which has some power of progression ; and 

 how different is the physical condition of the perfect Sesia to the 

 slow-moving Bovibyx. The goat moth larva, like its fellow tree- 

 dweller the Sesia, has a long life, and is metamorphosed into 

 a more or less active chrysalis, and these long-living caterpillars 

 change their skins during growth, but no essential difference is 

 noticed as regards their ornamentation during the successive 

 moults. All the supposed causes of the metamorphosis into the 

 chrysalis state act over and over again upon these long-liv^ed 

 caterpillars, and cold winter, genial spring, and hot summer, 

 succeed each other without the change taking place. 



The hybernation of such caterpillars as those of CEiiccira 

 pillcyiana and Lasiocavipa qiicrcus, which go into hiding imme- 

 diately after hatching, proves that there is something more 

 required to produce the metamorphosis besides want of food and 

 heat. The hybernation of well-fed caterpillars, like those of the 

 AntJu'occra already mentioned, must be considered with these 

 proofs that metamorphosis is not a condition to be determined 

 by simple physical influences, but that it often refers back to 

 ancestral peculiarities. The cocoon making and burying in the 

 height of the warm weather, when placed in relation with the 

 occurrence of two or more broods during the season, in some 

 closely allied species, the last being hatched late, are incompre- 

 hensible on the simple explanation that the silken and stony 

 shelters are absolutely required for the preservation of the insect 

 during its longer or shorter inclusion. The case of the Reapers 

 — which undergo metamorphosis both in summer and in autumn — 

 is complicated by the fact that the last brood, if it is very late, 

 hybernate as caterpillars, and not in their admirably protected 

 natural shelters. Here is a direct proof that something more 

 than cold and absence of food is required to determine the 

 first metamorphosis of the Lcpidoptcra, and that it cannot take 

 place until the development of the caterpillar has advanced to 



