BY K. J. TILLYAKD. 375 



remnant at the present day, consisting of only three species (*S'. 

 grayi Selys, .S'. braueri Selys, and S. villosa Rambiir; the first 

 two from New Zealand, the last from Chili). The main body of 

 the southern SomatochlorarStoi:\^ may have included a large number 

 of species at a [)eriod when our southern land-areas were much 

 larger than they are now. Before and during Miocene times, 

 these were becoming faced with a harder struggle for existence, 

 mainly due to the shrinkage of their territory; so that a remnant 

 gradually fought their way noi'thwards to the warmer tempex'ate 

 and even to tropical regions. As this proceeded, they became 

 modified, firstly to the intermediate Procordulia-type, and finally 

 to the ffei7iicordulia-tyTpe, which, in its rounded hindwing in both 

 sexes, shows a csenogenetic convergence to the Libellulinoe. 



The most archaic forms are evidently those northern species of 

 Somatochlora whose females still retain a remarkably enlarged 

 terebra. Our three southern species show a definite advance in 

 the reduction of the vulvar lamina to the more usual Libellulid 

 proportions. The gradual rounding of the anal angle of the hind- 

 wing of the male is cleai'ly seen in comparing the three genera, 

 Procordulia showing a " half-way " stage, between Somatochlora 

 and Hemicordulia. Also the second cubital cross- vein of Soma- 

 tochlora is eliminated in the two latter genera. 



Hespekocordulia, nov.gen. (Plate x., fig.l). 



Head and thorax robust. Legs long and slender. Abdomen 

 slender, cylindrical. Triangles of all wings and subtriangle of 

 forewings free. Sectors of arculus separated at bases. Arculus 

 placed between first and seccmd antenodals, but nearer to former 

 than to latter. Triangle of hindwing almost completely recessed 

 to level of arculus. Only one cubital cross-vein in hindwing, and 

 hence no hindwing-" suljtriangle." Post-trigonal space of fore- 

 wing beginning with one roto of cells, but sooner or later giving 

 place to two (variable). Anal loop long and narrow, fairly well 

 defined, apical end not enlarged, but cut straight off by a trans- 

 verse vein; longitudinal bisector fairly well defined. A small 

 cross- vein low down in anal triangle of hindwing. ^j^^TTp j' 



'.y^.^Hesperocorduliaberthoud.,..^. ^^i^/^ 



library"^ 



