[ 169] 



plus ag^es, en croissance active, le kyste, d^truit ant^rieurement, n'abrite 

 plus que la region postdrieure; le calice chitineux devenu plus dpais emboite 

 etroitement I'extr^mit^ postanale du parasite, allong^e en un court siphon 

 respiratoire bisegment^. ... La structure histologique de ces organes permet en 

 effet d'affirmer leur nature trach^enne. A la base, les cellules hypodermiques 

 sont abondamment d^veloppees et en plusieurs couches. La s^cr^tion chiti- 

 neuse ne forme plus, par suite, un simple filament spiral, mais une couche 

 continue, 6paisse et noire, de chitine: c'est cette region qui constitue propre- 

 ment le calice. Anterieurement, I'epaisseur de la parol kystale s'att6nue, 

 comme par^ etirement de la formation precedente, jusqu'a se reduire k une 

 mince couche chitineuse incolore ofi Ton ne distingue plus que quelques ilots 

 de cellules hypodermiques, les debris de mues s'ajoutent k I'ensemble." 



In Bucentes geniculata we have found structural relations between host 

 and parasite similar to those cited above (PI. XIV, figs. 1, 3, 4, 5, 6, 7 and 8). 



The larva of B. geniculata lives in the body cavity of Tipula paludosa. 

 We have not, so far, found it free in this situation, but always attached to 

 one of the main tracheal trunks of it.s host by means of a chitinous sheath-like 

 structure similar to that described by the observers quoted. 



At its junction with the trachea of the host, and for a short distance along 

 its length, this sheath is thicker and of a dark brown colour. Beyond this, it 

 is membranous in appearance and completely encloses the parasite. Usually, 

 however, and more particularly in the older larvae, the head end of the parasite 

 is extruded from the sheath freely into the body cavity of the host. 



Kecalling Pantel's description and figures of the genesis of the funnel, in 

 which he shows that the hooks at the posterior end of the parasite are utilised 

 to perforate the skin of the host, it seems probable to us that the attachment 

 between parasite and the tracheal system is established in a similar manner 

 in Bucentes geniculata. The body spines of the first-stage larva appear well 

 suited for such a purpose. At first the hind end of the parasite is adherent to 

 the host at the place where the inflammatory reaction is developed, and its skin 

 becomes incorporated with the funnel. On moulting taking place, the larva 

 leaves this portion of its cuticle and retracts away, leaving a clear space in 

 this area. The relation thus set up between parasite and host involves the 

 perforation of the host's tracheal system and establishes a common respiratory 

 system for both. Within the sheath there are always to be found the mouth- 

 parts and spiracles of the previous moults. In this way we have obtained 

 from third-stage larvae two sets of moulted mouth-parts and both anterior 

 and posterior spiracles. 



Description of Larva. 



First-stage Larva. As in almost all Cyclorraphous Diptera, the larva 

 passes through three stages separated by two moults. The first-stage larvae 

 obtained measured about 1-5 mm. Thirteen segments could be distinguished. 

 The head, as is usual in Cyclorraphous Diptera, is divided anteriorly by 



