- 
; 
CaRPENTER—The Geographical Distribution of Dragonflies. 455 
into the southern Holarctic sub-regions, having nine Mediter- 
ranean species. Anisogomphus, with far fewer species, has a much 
more discontinuous range in the tropics, and invades only the 
south-eastern Holarctic districts. Diastatomma and Cordulegaster 
may perhaps be reckoned as characteristically Holarctic genera ; 
but in both cases the Canadian sub-region is hardly richer in 
species than the Sonoran. Both genera occur in most of the 
European and Asiatic sub-regions of the Holarctic. The large 
genus Aischna or Gomphus shows a somewhat similar range, but the 
great majority of its species are about equally divided between 
the Canadian sub-region and the Sonoran. A fair number of 
species occur in the European and Asiatic districts of the Hol- 
arctic, stragglers ranging as far south as Abyssinia in the Ethi- 
opian, and Assam and Madras in the Oriental Region. The range 
of Hagenius is very remarkable; one species is Neotropical and 
Sonoran, one Japanese, and one North Indian; but it should be 
mentioned that the two latter are doubtfully referable to the 
genus. Ictinus is widely spread over the Oriental and Ethiopian 
Regions, and has five species in Madagascar; as noted above, it 
invades the Manchurian sub-region of the Holarctic, aud there is 
also one species in Australia. Anotogaster, with two species in 
northern India and one in Japan, is hardly more Oriental than 
Holarctic. 
Striking facts in the distribution of Gomphine dragonflies are 
their absence from the Austro-Malayan and Polynesian sub- 
regions, and their scarcity in Madagascar and Ceylon. ‘The 
affinity of the southern Neotropical with the Australian and New 
Zealand gomphines has already been mentioned. The genera 
in question all belong to the division Cordulegasterina, which is 
absent not only from Austro-Malayan, but also from the whole 
Neotropical Region except Chili. Such isolation of animal groups 
at the extremities of the southern regions has often been invoked 
to support the theory of an ancient Antarctic continent. In the 
present instance, however, we fortunately have some evidence 
from fossils by which to cheek our speculations. The Gom- 
phinae are believed to be one of the most primitive of the odonate 
sub-families, and remains of dragonflies referable to them have 
been discovered in rocks as old as the Lias. Quite a number of 
SCIEN. PROC. R.D.S., VOL. VIII.) PART Ve 2K 
