STATES OF INSECTS. 269 



the hoar. The Ephemerae observed by Reaumur appear 

 at no other time than between eight and ten o'clock in 

 the evening ; and so unalterably is their exclusion fixed, 

 that neither cold nor rain can retard it. Between these 

 hours, in the evenings on which they appear, you may 

 see them fill the air, but an hour before or after, you will 

 in vain look for one a . So also the silkworm-moth and 

 the hawkmoth of the evening primrose [Sphinx CEno- 

 therce) constantly break forth from the pupa at sunrise : 

 and the hawkmoth of the lime [Smerinthus Tilics) as cer- 

 tainly at noon b . Schroeter states, that of sixteen speci- 

 mens of the death's-head-hawkmoth [S. Atropos) which 

 he bred, every one was disclosed between four and seven 

 o'clock in the afternoon c . 



Before I conclude this head, I must observe, that after 

 a caterpillar or gnat has spun its cocoon, it sometimes 

 remains for a considerable period before it incloses itself 

 in the pupa-case, and casts off the form of a larva. Thus 

 the little parasite [Ichneumon glomeratus L.) that destroys 

 the caterpillar of the common cabbage-butterfly, remains 

 a larva in its cocoon for many months, but it becomes a 

 perfect insect a few days after it has put on its pupa- 

 rium d ; and the caterpillars of the great goat-moth ( Cos- 

 sus ligniperda), if they spin their cocoon in the autumn, 

 remain in it through the winter in the larva stale; 

 whereas, if they inclose themselves in the month of June, 

 they assume the pupa, so as to appear as flies in three or 

 four weeks e . It is not therefore easy to state precisely 



a Reaum. vi. 486. b Brahm. 423. 421. 



c Naturf. xxi. 75. A Reaum. ii. 423. 



* De Geer ii. 370. It is not certain, however, that De Geer did 

 not, in this instance, mistake the winter habitation of a larva for a 



