620 CAUDAL GILLS OF ZYGOPTEHID LARVAE, 



it, as brandies of the tenth-segment nerves of the ahdonien. 

 Prohablv correlated with this would be the regulation of the 

 blood-circulation, by the development of a closed afierent canal, 

 the primary hlood-cmial . As the positions of both these struc- 

 tures are different in Perlaria and Odonata, it is logical to 

 assume that the two Orders became separated at a stage in the 

 evolution of the Insect Larva earlier than is here indicated. 



The third stage would be the development of the main trachea 

 along the interior of the processes, the cerci receiving one each, 

 from the dorsal trunk of its own side, but the median appendix 

 dorsalis receiving two, one from each side. Unless these trachea:' 

 can be shown to exist in some Plectopterous larvae, we must 

 hold that the Odonata became differentiated from their common 

 stock with the Plectoptera befoi-e this outgrowing of the trachea^ 

 took place. 



Finally, as all the appendages of the Odonata, in every part 

 of the body, show a great reduction in the number of joints U.y., 

 the antennae, the tarsal joints, the labial and maxillary palpi), 

 we must assume that the reduction of the caudal processes to a 

 one- or two-jointed form proceeded parallel with these other 

 reductions; in other words, the original Protodonate ancestor 

 alread}' possessed all these reductions, derived from a Palaeo- 

 dictyopterous ancestor. 



Thus we conclude, that the original form of Zygopterid caudal 

 process was a short, one- or two-jointed appendage, in which the 

 nerves and blood-canals were developed as in the caudal pro- 

 cesses of the Plectoptera, and with the addition of a single longi- 

 tudinal trachea in each cercus, and two in the appendix dorsalis. 

 In this organ, neither internal lamina? nor alveolar meshwork 

 were developed. 



Evolution of the Cmidaf Gills within the Suborder Zygoptera. 



We have now to consider the evidence of the Ontogeny, with 

 respect to the further evolution of gill-types within the Suborder, 



As far as I can ascertain, all newly-hatched Zygopterid larvae 

 have the gills filiform, one-jointed, and hairy. We may safely 

 assume that the filiform condition is primitive, since all the evi- 



