BY R. J. TILLYARD, 335 



In studying the remaining thirteen species, we can at once 

 pick out a homogeneous group of four (group 2 above-mentioned) 

 characterised by the short anal appendages of the males, the 

 absolute loss of the ovipositor in the females, and the generally 

 rather shorter and less constricted abdomen. These are S hrevi- 

 styla, S. virgula, S. yittlata, and S. nigra. I regard these as the 

 most advanced members of the group Synthemina. Their vena- 

 tion is more open and .less inclined to variation than that of the 

 other species, and in the complete loss of the ovipositor, they have 

 at last reached an invariant stage. These four species are also 

 very much more closely allied to one another than any other two 

 species outside them. I therefore propose to place them in a new 

 genus, Metathemis, of which the type will be S. guttata Selys. 



There now remain nine species (including the impei'fect female- 

 type, S. miranda Selys, which is retained next to S. regina, to 

 which it is closely allied in venation), in which the superior 

 appendages of the male are always long atid wavy, the mernbra- 

 nule always present, the front large, and the abdomen long, 

 pinched at 3-4, and (except in the smallest species, S. cyanitincta) 

 somewhat narrowed again at 8 or 9. Of these, the female of S. 

 primigenia possesses an enormously long ovipositor; the ovipositors 

 of S. eustalacta, S. tasmanica, S. leachii, S. macrostigma are large 

 and conspicuous; that of S. regina (closely allied in other respects 

 to S. eustalacta) is very small and those of *S'. claviculata and ^S'. 

 cyanitincta almost obsolete. The two latter are otherwise not by 

 an}- means closely allied, and are from widely different localities. 

 Of all the species it may be said that, in contrast to the four 

 species above separated out, they all possess at least a rudiment of 

 an ovipositor. I propose to retain these in the genus Synthemis 

 (restricted), the tyjie of the restricted genus, of course, remaining 

 S. eustalacta Bnvxa., the original type of the whole genus proposed 

 by de Selys. 



I cannot see any advantage in suggesting further subdivision. 

 The nine species just considered form, without doubt, a less 

 homogeneous division than the other two, but they are in many 

 ways very closely allied, and are certainly the most archaic 

 members of the group still extant. 



