274 MESSES. MIAI.L AND HAMMOND ON THE DEVELOPMENT 



become bit by bit convex, not by any g-radual widening of tbe fold, but by a steady 

 extension of tbe convex surface at tbe expense of tbe concave fold (PL XXXI. fig. 33). 

 Tbe process is bard to describe, but may be closely imitated by cutting a bollow india- 

 rubber ball into halves, and everting one of tbe bemispberes with the fingers. During 

 the process of eversion the compound eyes are drawn downwards and backwards, so 

 that they get behind and beneath tbe bases of tbe antennae (PL XXX. fig. 29). The 

 morphologically external surface of tbe eyes, Avhicb was previously turned towards the 

 lumen of the invagination, now looks outwards (PL XXXI. fig. 33) ; tbe optic nerve, 

 which was distributed to the (temporarily) convex and outer surface of the eye, still 

 takes its course to tbe same surface, now become concave and internal ; and tbe walls 

 of tbe head now for tbe first time enclose tbe brain. Tbe inner walls of the paired 

 invaginations, which were very long and enclosed a narrow median space or sinus 

 («', PL XXIX. figs. 18-22), contract more and more, and give rise to the central parts of 

 the head of the fly. 



Sections taken through the pupal antenna indicate the change resulting in tbe 

 formation of the split in that organ to w'hich reference has been naade (page 268). Tbe 

 portion of tbe wall destined to form the hollow of the shaft projects in a marked manner, 

 and is much more delicate than tbe remaining portion (see PL XXX. fig. 30). 



Tbe muscles of tbe larval head, as well as of some other parts of the body w^liicb become 

 completely transformed, disappear by a process which leaves behind a number of more 

 or less empty sheaths. This process reminds us strongly of the bistiolysis which takes 

 place in Muscidte, as well as in many other animals. We have not, however, in spite 

 of many efforts, arrived at such proofs of bistiolysis as can readily be obtained in 

 Muscidse. We have never, for example, seen unmistakable sarcolytes within tbe 

 supposed phagocytes, and cannot tlierefore appeal to any of our preparations as 

 furnishing a demonstration of tbe process of bistiolysis in Chironormis, which is on 

 general grounds highly probable. 



VII. Comparison o/" Chironomus with some other Insects. 



In the larva of Culex, as we find from Dr. C. H. Hurst's partly unpublished descrip- 

 tions and preparations, there are no deep invaginations for the compound eyes or antennge 

 of the imago. The compound eye forms beneath the larval eye-spots, and is at first 

 relatively simple and of few facets. The number increases by the gradual formation 

 of partial and marginal invaginations, each of which forms a new element. The imaginal 

 antenna grows to a much greater length than that of the larval antenna, and its base is 

 accordingly telescoped into tbe bead while the shaft becomes irregularly folded*. Culex, 

 though more modified than Chironotmis in many respects, e. g. in tbe mouth-parts, is 

 relatively primitive with respect to the formation of the imaginal head, and shows a 

 mode of development of the eye and antenna which we may suppose to have characterized 

 a remote and comparatively unspecialized progenitor of Chlrononms. 



* "The Pupal Stage of Culex" (Studies from Biol. Lab. Owens Coll. vol. ii. 1S90). See also Manch. Micr. 

 Soo. 1890. 



