TRANSFORMATIONS OF THYSANURIFORM LARV^. 73 



have a very close resemblance to each other, he conse- 

 quently concludes, with every show of reason, that they 

 will produce pupae and butterflies equally near, and 

 equally resembling each other. He accordingly feeds 

 them, and watches their transformation with great anx- 

 iety. The first, we will suppose, which changes into 

 chrysalis, is that of Podalirius Pompilius* , or the Ja- 

 vanese swallow-tail : he sees with surprise, that the 

 chrysalis, instead of having its head suspended down- 

 wards, as in all the true thysanuriform types, has it in 

 an erect position; while the butterfly, into which it is 

 subsequently transformed, has six perfect legs, and turns 

 out to be of the iuliform, or pre-eminent, type. Sadly 

 perplexed at this, he traces the progress of another ; this 

 he conceives, from the length of its horns, to be more 

 typical, and he consequently expects that it will pro- 

 duce him a perfect thysanuriform butterfly. But he is 

 again baffled; it becomes transformed into the purple 

 emperor, — a butterfly which belongs to the raptorial, or 

 scolopendriform, division. Of two green caterpillars, 

 which, from their similarity, seemed to belong to species 

 of the same genus, one only becomes a true thysanuri- 

 form butterfly, the other changing to one of the scolo- 

 pendriform types. Finally, he gets from a fifih, which 

 closely resembles the last two, a decided species of Iles- 

 peria,—a butterfly totally different from all those which 

 his collection of thysanuriform larva;had produced. How, 

 then, it may be asked, can we maintain that each of the 

 great divisions of butterflies has a peculiarly shaped ca- 

 terpillar; when we see, as in the foregoing instance, those 

 which are caUed thysanuriform are scattered in every one 

 of these divisions ? This question brings us to the point 

 we are to explain, — namely, the principle of their va- 

 riation. It is, theoretically, as follows : — Every natural 

 group of butterflies, either in their caterpillar or perfect 

 state, contains representations of the primary types of 

 larvtc, modified, however, in such a manner, as to indi- 

 cate the real type to which they actually belong. Sup- 



• Zool lU. ii. pi. 105, 



