The Odoiiafa or Dragonflies of South Africa. 257 



Tillyard iu a spirited correspondeuce, lie felt the desirability of new 

 observations aud pi-obably of a different orientation. Larval prepara- 

 tions of Caloptery.v and of Ischnura were photographed, compared with 

 various other important photographs furnished by Mr. Tillyard {Lestes, 

 Synlestes, Pseudagrion, Neosticta, etc.), and the following conclusion 

 was arrived at provisionally, and for which Mr. Tillyard and the 

 writer alone have to accept for the present the responsibility: (1) 

 Zygoptera are radically different from Anisoptera, having the radius 

 unbranched in larval and mature wing ; what has been termed Bs is, 

 in fact, a branch of the media. (2) This particular branch of the 

 media (Rs of our terminology) has undergone the following (gradual) 

 changes of position : (a) a very proximal origin in the] quadrangular 

 region (most primitive position — Colopteryr and many others of the 

 Calopterygidae) ; (6) considerably more distal origin, very near tlie 

 nodus (the bulk of Agrionidae) ; (c) still more distal origin, shifted 

 so far distally as to become a secondary branch of the following 

 branch (M2) of the media. In this condition a bridge (a very long 

 one in Lestes, a shorter one in Chlorolestes aud Synlestes) appears 

 between this " Bs " and the main branch of the media (the structure 

 thus resulting in an apparent analogy with the anisopterous condition, 

 though actually different). Under this assumption (unbranched 

 radius, distally migrating branch of media at place of anisopterous 

 [i2s]) most of the difficulties of interpretation vanish for the mature 

 as well as for the larval zygopterous wing, and as far as the writer's 

 observations go, all ontogenetic preparations, even of earlier stages 

 (Calopteryx and Ischnura) , dure fully in favour of this theory and none 

 contrary to it. 



Nevertheless I hesitate to draw conclusions from this view in 

 proposing a new terminology for the Zygopterous wing. Mr. Tillyard 

 has done so, and says : " Ms in Zygoptera for the vein in place of Bs 

 in Anisoptera." Fortunate as this term appears to be, I refrain from 

 accepting it, as long as the question is not fully settled by the 

 discussion that must certainly follow. 



The abdomen is joined to the thorax in a free but rather broad 

 joint. It is composed of ten fully-developed segments of subequal 

 value, although the first and the last are somewhat reduced in size, and 

 the second, eighth and ninth distinguished by their relation to sexual 

 organs. The dorsal sclerites, the tergites, are predominant ; the lateral 

 and even part of the ventral surface belong to the tergite. The 

 ventiul sclerites, the sternites, are, as far as not modified in the 

 genital segments, narrow, flat plates, mostly of indifferent structure 

 and colour, folded under the veutrally produced margins of the 



