198 PHILIP p. CALVERT. 



of living Odonate imagos are universally three-jointed, and while 

 most nymphs agree in this respect, the first and second tarsi of the 

 nymphs of the Gomphinse are never more than two-jointed. The 

 oldest nymphs rarely shov,' any trace of the tooth on the tarsal nails. 



The abdomen is always much shorter and broader than in the 

 imago, and owing to its flattened form, the distinction between dorsal 

 and ventral surfaces is much more evident. The pleura are equally 

 well chitinized as the terga and the sterna ; in the Zygoptera they 

 are infolded, in the Anisoptera exposed to view on the ventral sur- 

 face ; the sterna are rectangular. Along the lateral margins of the 

 segments a spine is frequently developed, the arrangement of which 

 is often of generic character. 



The three tracheal-gills which terminate the abdomen of the Zy- 

 goptera are, in the young nymph, conical, tapering gradually to the 

 apex, longer than the abdomen and sparsely hairy. In Calopteryx 

 they become triangular in cross-section ; in the Agrioninse, each gill 

 becomes a flat, thin plate ; in both cases they enclose tracheal 

 branches. Their relative length decreases. 



In the Anisoptera, where, from the first, the places of the terminal 

 gills are occupied by chitinous, non-respiratory appendages, the mid- 

 dle or dorsal appendage appears to be the shortest even at hatching. 

 The pair of chitinous pieces, described as being constantly in motion 

 in young Anisopterous nymphs, apparently disappears in early nym- 

 phal life. It is perhaps a fact of some significance that the hairs 

 which are found in the youngest nymphs on the thorax, abdomen, 

 legs, terminal gills in the Zygoptera or three terminal appendages in 

 the Anisoptera are early lost. 



The superior terminal appendages of the imago appear in the 

 nymph at about the same time as the wing-covers, as a small conical 

 process on either side of the dorsal gill or dorsal appendage. They 

 increase in length with each moult, but never become as long as the 

 gill or appendage between them. A comparison of recently trans- 

 formed imagos and old nymphs indicates that the anal segment of 

 the imago is formed by the representatives of the three caudal gills 

 (Zygoptera) or appendages (Anisoptera) of the nymph. In the 

 females (imagos) of both of these groups neither dorsal nor lateral 

 pieces are greatly developed, and the anal opening lies between all 

 three. In the males of the Zygoptera the two inferior appendages 

 correspond in position to the two lateral gills of the nymph ; the 

 anal opening lies between them and is bounded above by a plate 



