336 MODERN CLASSIFICATION OF INSECTS. 
(Turin Trans. vol. vi.; and see Blyth, in Field Nat. p. 470.); and 
Captain Lyon, in his Travels in Mexico (vel. i. p.70.), mentions having 
observed butterflies assembled in vast numbers, settled in patches 
several square yards in extent. M. Donzel has published a curious 
memoir upon the flight of butterflies whilst coupling (Ann. Soc. Ent. 
de France, 1837, p.'77.), showing that whilst the males of Pontia 
Brassice, &c., Colias and Polyommatus, support the females, it is the 
latter which support their partners in the genera Thais, Thecla, 
Argynnis, Melitaea, Hipparchia, and Pieris (P. Crategi, formed by 
Donzel into the genus Leuconea). 
In respect to their transformations, these insects exhibit three prin- 
cipal variations, the chrysalides being either suspended, girt, or enve- 
loped in a cocoon. In the first of these, the caterpillar, when full 
grown, spins a small web upon the under side of a leaf or twig, and, 
holding this by its hind pair of anal feet, it drops its head down, when 
the skin bursting, the head of the chrysalis appears. By degrees the 
caterpillar skin is pushed upwards, until it is reduced to a minute 
shrivelled mass, which, however, still serves as a support to the 
chrysalis, by the pliability of the terminal segments of the body. 
Still retaining this situation, it contrives to disengage the extremity 
of the body through the slit side, to extend it upward, and to fasten 
it to the little cone of silk by means of various minute hooks. 
After it has accomplished this delicate task, it gets rid of the old 
exuvia by giving the body a jerking kind of a twirl, by which the 
slough is detached, and then falls. In the girt chrysalides, the cater- 
pillars not only attach themselves by the tails, but also, by throwing 
the head from side to side, spin a number of threads across the middle 
of the body. The skin of the caterpillar then bursts, and is sloughed 
off; the contractions of the body forcing it beneath the girth. The 
number of the species which enclose themselves in cocoons is but 
small. Some of these spin a slight silken case, whilst others make 
for themselves a more substantial covering, by fastening several leaves 
together with threads, after the manner of the Tortricidae. In the 
two former variations the chrysalis is always angulated, but in the 
latter it is conical, like that of the moths. 
It is to be observed that those chrysalides which are simply sus- 
pended almost always produce tetrapod butterflies ; a curious fact, 
because, if any butterflies required more than the rest a perfect pair 
of fore legs, we should have thought it would have been those which 
