84 INTRODUCTION. 



in our estimation, than those afforded by the butter- 

 fly; but generally these characters confirm each 

 other, in other words, two caterpillars presenting 

 the same characters produce flies which likewise 

 partake of the same generic characters. It will no 

 doubt be objected to us, that it is illogical to deduce 

 characters except from the animals we are attempt- 

 ing to classify. We reply, that by following any other 

 plan we must despair of attaining to a natural 

 method. Besides, it is not necessary to be acquaint- 

 ed with the caterpillars of every species ; it is suf- 

 ficient to study a caterpillar and chrysalis of Vanessa 

 or Pieris to have an exact idea of those of the two 

 genera. Even though it should be very difficult to 

 verify the characters taken from the larva, this is 

 no reason why we should abstain from employing 

 them. It is not merely from the fruit that botanists 

 obtain their characters, but likewise from the flower, 

 and even from the first developement of the vege- 

 table embryo. The flower is to the plant what the 

 caterpillar is to the lepidopteron, and the different 

 modes of metamorphosis have as much value as 

 those of inflorescence. 



" It appears to us that Latreille's three divisions, 

 taken from Linnaeus — Diurnal, Crepuscular, and 

 Nocturnal — are too inaccurate to be retained, espe- 

 cially the crepuscular section. The denomination 

 Diurnal not only applies to all the known kinds of 

 day butterflies, but also to an almost infinite num- 

 ber of others forming a part of the two other divi- 

 sions, such as Macroglossa, Zygenides, Castniarice, 



