1864.] 555 



does not vary in the same species in the G-all-gnats of the Willow, but 

 on the other hand it does not diiFer in different species, except of course 

 in its proportional size ; so that it is of no service hei'e towards distin- 

 guishing species. 2nd. The larva varies very considerably in its colora- 

 tion, and becomes more deeply colored after it has reached maturity 

 and formed its cocoon, as it approaches the period when it tranforms 

 into the pupa state; but the " breast-bone" (See Dipt. N. A. p. 182,) 

 is tolerably constant in the same species. Unfortunately, however, this 

 last character does not differ materially, there being only two distinct 

 forms, the Y-shaped, varying in the same species by a considerable cur- 

 tailment or prolongation of the lower (or posterior) arm of the Y and 

 the clove-shaped breast-bone. I thought at first that this curtailment 

 or prolongation might be due to the greater or less degree in which the 

 joint bearing the breast-bone is overlapped by the following joint; but 

 on carefully examining at the same hour 26 larvae belonging to 5 dif- 

 ferent but closely allied species, and distending the fore part of their 

 bodies by pressure so as to obviate any such overlapping, I satisfied 

 myself that it was a honn fide variation, and that it occurs in at least 4 

 out of the 5 species. The comparative length and breadth of the larva 

 is somewhat inconstant; for although C. s. siliqua n. sp. (?) is always, 

 so far as I have observed, (9 specimens) elongate, yet others (e. g. C. 

 s. brassicoides n. sp. and C s. strobiloides n. sp.) are generally short, 

 but occasionally as elongate as C. s. siliqua. ord. The pupa varies very 

 considerably in coloration, becoming much darker before it transforms. 

 On the other hand the coloration of the empty pupal integument is 

 very constant, and presents a few very remarkable differences in differ- 

 ent species, even when the pupae themselves are undistinguishable in 

 their coloration. The characters drawn from the structure of the horns 

 at the base of the antennae and the thoracic bristle (ibid. p. 185-6) are 

 invariable, but do not differ much in different species, no less than five 

 species being precisely identical in these respects. It is remarkable that 

 I could not find in any species the bristle behind the base of the an- 

 tenna, which is stated by Osten Sacken to occur in most Cecidomyide 

 pupse. Ath. The structure, shape and comparative dimensions of the 

 cocoon, which the larva of most species constructs in the autumn, and 

 in which it lies through the winter, not transforming into the pupa till 

 a few weeks before it is ready to assume the imago state in the follow- 



