DISCUSSION. 419 



T. crinita Distant, described from Australia, without exact 

 locality; Gippsland, Melbourne, and Kunell East, are the only 

 known localities for it — a wide range along the southern portion 

 of Southern Victoria; Kershaw says it is found, like the 

 Tasmanian species, in fern-tree gullies. 



Mr. Tillyard — Australian dragonflies(ODONATA) seem to offer 

 a very good test-case, for several reasons. (1) The group is a 

 small one, so that careful records of localities, sufficient to indicate 

 the range of distribution of species, can be made and analysed 

 with great exactitude. (2) There is very little variation in the 

 species of the group. (3) Hybridisation between closely allied 

 species is impossible, owing to the remarkable divergence in 

 formation between the genital appendages of both sexes in closely 

 allied species. A male of one species cannot pair with a female 

 of another closely allied species. In many cases the two species 

 may be almost exactly alike in colouring and may occur together 

 on the same river or lagoon, yet crossing cannot possibly take 

 place. This is a most important point, as it simplifies matters 

 very greatly. 



An analysis of the 170 species comprising the Australian 

 Odonate i&unsi gives the following results : — (1) Nearly 50 species 

 belong to monotypic genera; these are of no value in this dis- 

 cussion. (2) Between fifty and sixty have their distribution 

 too imperfectly known to be of any real value. (3) Out of 65 

 remaining species, belonging to 18 genera, all possible pairs 

 which are clearly very closely allied* have been chosen. A few 

 pairs, not really closely related, but still placed in the same 

 genus in the present classification, have also been omitted. 

 The total is forty-thiee pairs. Of these only eight are geminate, 

 in the sense of being closely allied species separated by a definite 

 barrier. They are as follows : — 



* It is necessary to bear in mind the operation of the " personal equa- 

 tion " in this choice. By including the few pairs which have been rejected 

 as being, in the speaker's opinion, not sufficiently closely allied, the number 

 of geminate species is not only not increased, but their proportion to the 

 whole is diminished. 



