DISCUSSION. 421 



parts of South Australia, Victoria, and Tasmania, so that the 

 proportion is about the same for this side of the barrier also. 



In the case of Bass Straits, of the twenty-four species found 

 in Tasmania, twenty-three occur also in Southern Victoria. It 

 is curious, however, to notice that several common South Vic- 

 torian species, including the large strong-flying Hetnianax jmpu- 

 enses, have not succeeded in establishing themselves in Tasmania. 



In the case of New South Wales and Chili, there are, as might 

 be expected, no species common to both; and only one pair of 

 closely allied species. 



When the 43 closely allied pairs chosen are classified, it is 

 found that they fall easily into three main groups, which may be 

 termed respectively Inlying Pairs, Overlapping Pairs, and Isolated 

 Pairs*: — 



{I) Inlying Pairs. — Those in which the distribution of one 

 species is a small region entirely surrounded by a larger region, 

 over which the commoner and more widely distributed second 

 species ranges. There are ten cases of this out of the 43. 



(2) Overlapping Pairs. — Those in which the distribution of 

 one species coincides with that of the other for a greater or less 

 portion of the total area. There are 24 cases of this. 



(3) Isolated Pairs. — Those in which the two regions of distri- 

 bution are quite distinct. There are nine cases of this, viz., the 

 eight "geminate" pairs, and the isolated Lestes aridus Tillyard, 

 which has been paired with L analis Ramb., a species that 

 surrounds it on all sides but does not reach the same region. 



Assuming that, in the case of the Odonata, geographical 

 isolation has played a distinct part in the formation of new 

 species; and, admitting also, the strong probability of " muta- 

 tion " or the sudden arising of new forms, a simple explanation 

 can be offered not only for the occurrence of these tiiree forms, 

 but for their relative frequency: — 



(A). In fig.l of Plate x., let the large, lightly-shaded area repre- 

 sent the region of distribution of a species, X. Suppose a muta- 



* Three exceedingly apt terms suggested by Mr. A. H. 8. Lucas. 



