BY R. J. Tir-LYAKD. 6*29 



this primary blood-canal existed in the earliest Odoimta. When 

 the caudal process swelled out into the 8accoid Type of Gill, a 

 great increase must have taken place in the size of the hyeuiocoele 

 within it; and this increase must have aftected the afferent 

 canal, as well as the rest of the blood-space. At this stage, the 

 efferent or return-current must have proceeded sluggishly back- 

 along the undifferentiated portion of the hiemocfele. The sub- 

 sequent development of the inffvital laminca and the alveolar 

 meshwork, within this undifferentiated portion, closed-off the 

 secondary or efferent canal, which we have Sf^en, from the onto- 

 genetic evidence afforded by an early instar of Neosticta, is 

 laiger and less definitely developed than the older primary or 

 afferent canal The final result is that the original ha^mocoele 

 of the gill becomes divided up into two distinct canals, sur- 

 rounded by a mass of alveolar tissue, with zones of internal 

 lamina?. 



{2) The Xerronn System : —'Vha original Odonata must have 

 possessed caudal processes with a nerve-supply similar to that 

 found in the Plectoptera, viz., two main nerves in each cercus 

 (both on the outer side), and four in tiie appendix dorsalis (two 

 dorsal and two ventral). This remains the same throughout all 

 subsequent developments of gill-types, except only in the case of 

 the Lestid form of the V^ertical Lamellar Oill-'l'ype, in which 

 the median gill has its nerve-supply reduced to two only, by the 

 suppression of the two ventral nerves. 



(3) 7'he Tracheal Sy intern . —It seems clear that the original 

 caudal processes could not have accommodated more longitudinal 

 tracheie than one each in the cerci (on the outer side), and a 

 single pair in the appendix dorsalis. As this is the arrangement 

 also in the filiform gills of the second larval instar (Text-fig. 34), 

 we may take it as the primitive condition in the Zygoptera. 

 When the Saccoid Type developed, each of these main trachea' 

 gave off a l)ranch from the base; so that there were then two 

 main tracheae in each lateral gill, and four in the median. One 

 of the two tracheae in the lateral gills frequently divided a 

 second time; so that the S^accoid Type, as it stands to-day, may 

 show either two or three main tracheae in the lateral gills. 



