622 CAUDAL GtLI.S Of ZYGOPTERID LAEV.t:, 



ence of tlie huge, cumbersome, Uladcler-like gills of Diph/ebnr and 

 Cora. It seems clear, also, that these organs are primarily an 

 attempt to evolve blood -fj ills, and not tracheal-yills', since it is the 

 blood-space that undergoes an immense increase in volume, — the 

 tracheal supply remaining far inferior, in these gills, to that to 

 which it eventually attains in the Lamellar Gill. 



Although we designate the Saccoid I ype as the most archaic 

 of all, there is one qualification to add. 'Ihese huge bladder- 

 like gills could not be held apart in the watei*. The two lateral 

 ones must always have rested either on the river-bed, a rock, oi" 

 a convenient sten), on which the larva mav have been resting:. 



o 



o 

 a. 



° op m. 



d e f g-^ 



Text-tig. 4"). 

 Evolution of the outward form of the eaudal gill>. a^; shown in T.S. 

 Phyhjgcnetic seiie.s from caudal hlameiit, through simple Saecoid 

 and Tiiquetro-quadrate Types, to Lestid form of Veitical Lamellai 

 Type: a, caudal filament; h, simple saccus; c, Diphlehiu-Hi&^^e of 

 same: d, c, evolution of the Tiiquetro-quadrate Tj-pe (t); /, (/, evo- 

 hition of the Lestid form of Vertical Lamellar Type (;/). 



The median gill must also have rested symmetrically upon the 

 other two. Hence, even in the huge bladder-^ike gills of Di- 

 phlebia,t\\e beyinnhigs of the Triquetro-quadrate Type are evident. 

 That type, as seen in its highest development in the Caloptery- 

 (jinxe, is clearly a specialised development from the older Saccoid 

 Type ( Text-fig. 4 5, rt-e). Thus we may say that a single line of 



