BY R J. TILLYARD. 445 



forms of Nannodythemis. I also obtained the information that 

 Brauer's species was, as I had expected from its locality, most 

 certainly the smaller form. 



I now had materials for a short paper on the genus; but there 

 was a further surprise in store for me. After making careful 

 descriptions of the two species, I paid a visit to Wentworth 

 Falls, hoping to get a better series of N. australis than I had at 

 the time. There I took, in February of this 3 7 'ear, two distinct 

 species of this genus, one of which was certainly N. australis, but 

 the other quite distinct from it and from the western form. Of 

 the new form I was unfortunately only able to take four males 

 and two females, but these, with my long series of the other two 

 species, are sufficient to determine accurately the existence of 

 three distinct but closely allied species of the genus. 



Brauer distinguished his genus Nannodythemis from Nanno- 

 phya Rambur* chiefly by the fact that the triangle of the hind- 

 wing is normal {i.e., three-sided) in Nannophya while in Nanno- 

 dythemis it is abnormal {i.e., quadrilateral). Now in the two 

 new species before me, which are evidently so closely allied to 

 Brauer's N. australis that there can be no doubt as to their being 

 congeneric, we find the following remarkable fact : — in the 

 western form, the triangle of the hind wing in both sexes is 

 normal; in the form from Wentworth Falls the males have 

 normal triangles, while the two females I possess have an abnormal 

 triangle in the hindwings. The former should then be placed in 

 Nannophya, together with the male of the latter; while the 

 female of the latter is a true Nannodythemis ! The solution of 

 this difficulty is an obvious one. Brauer, in creating the genus 

 Nannodythemis, chose in defining it a variable character, which, 

 far from being of true generic value, is not even of specific value. 

 I can even find in my series of N. australis several individual 

 specimens which possess a normal or nearly normal triangle in 

 the hindwings, either on one side only or on both; while in my 

 series of the western form careful examination reveals the 



* Rambur, Ins. Nevr. p.27(1842). 



