1S79.] 17 



similar " mimicry " in C. blattarice. " In repose, with its legs contracted, it affords 

 an exact representation of a small patch of bird-droppings." The larva of S. ruviicis 

 bears also a close likeness to the flowering tops of the Rutnex, among which it feeds. 

 Nevertheless, it appears to be very liable to the attacks of parasites, and indeed of 

 two species of IchneumonidcB. The cocoon of the one is larger (2^ — 2\ lines in 

 length), looser in texture, and attached to the outer Hypera-cocoon by an open mesh- 

 work of threads, and remains always white in colour. On one occasion I extracted 

 a white apodous larva, wliich had just emerged from the body oi & Si/jjera, from the 

 Si/pera-CQCoon, and kept it several days in a watch glass. For four or five days it 

 kept on spinning as if making a cocoon, but without investing itself in the web. On 

 the third day it had deposited a great thickness of silk on the glass, the absence of 

 surrounding points of attachment (normally afforded by the Sypera-cocoon), appar- 

 ently preventing it from forming a cocoon; and had become quite a purple-brown in 

 colour from the shining through of its contents. On the fourth and fifth days it 

 seemed stUl spinning, but languidly ; but on the seventh it was quiet, contracted, 

 and thickened. I did not succeed in obtaining a pupa from it. Kirby and Spence 

 say, " Introduction," vol. iii, p. 225, " It is a general rule, that those larva? which spin 

 cocoons never in ordinary circumstances become pupas without having thus enclosed 

 themselves." Many of the Hi/pera larvae which I bred, however, pupated without 

 forming cocoons, and the same thing happened, less frequently, with Cionus scrophu- 

 laricB, of which I shall have more to say anon. The second parasitic cocoon is much 

 smaller (Ij lines long), elongate, egg-shaped, compact in structure, and lying loose 

 within the Hypera cocoon. When fresh it is white like the larger one, but on the 

 second day begins to acquire a pink or reddish colour at the ends, which gradually 

 deepens and extends, till on the third or fourth day only a narrow equatorial belt 

 of white remains, the rest of the cocoon being of a dusky purple. Similarly belted 

 cocoons are mentioned by K. and S., vol. iii, p. 221. " Sometimes the same cocoon is 

 of two different colours. Those of certain parasites of the tribe of Chalcidites, &c., 

 Latr., the motions of one of which I noticed on a former occasion [vol. ii, p. 296], 

 are alternately banded with black or brown and white, or have only a pale or white 

 belt in the middle, which gives them a singular appearance." But their account of 

 the manner in which this dilference of colour is produced does not seem to apply in 

 the case of the Hypera parasite. " In both cases the difference of colour depends 

 upon the different tints with which the silky gum is imbued in the reservoirs : the 

 fii'st portion of it is white, and with this the larva first sketches the outline of its 

 cocoon, and then thickens the layers of silk considerably, in those parts where the 

 white bands appear: when these are finished, its stock of white sUk is exhausted, and 

 the remainder of the interior of the cocoon is composed of brown sUk." My cocoon, 

 which was made in the end of July or beginning of August, 1877, is now (Apr., 1879) 

 of a brittle, membranous, or gummy texture. The fly has escaped out of a circular 

 opening near one of the ends. I have cut it open and perceive that, while the thickness 

 of the shell is if anything darker at the equatorial region than elsewhere, the outer 

 whitish belt seems quite superficial, and looks almost as if it could be peeled off. 

 Coi-responding with this white external equatorial belt, there is on the inner surface 

 of the shell a dark opaque ingrained belt ; and to the one side of this, and partly 

 overlapping it, a superficial belt of white glistening silk, with a groove in the middle 

 running round the inside of the shell. On the cut edge of the couoon it can be 



