CH. Ill] FRESH-WATER FISH. 171 



between Patagonia and New Zealand than between either 

 of these countries and the Cape of Good Hope. These 

 however are by no means the only facts which seem to 

 point to some read} 7 method of transit between these 

 different regions for purely terrestrial or fresh-water 

 animals. Many writers have with justice emphasised the 

 extremely significant fact that the fresh- water fish Ga- 

 laxias attenuates occurs in Tasmania, New Zealand, the 

 Falkland Islands and the southern extremity of the South 

 American continent. This is perhaps the only case of a 

 species of fish with so wide a range in the antarctic hemi- 

 sphere ; but the two families of Salmon oid fishes Ga- 

 laxiidcv and Hoplockitonidce are restricted to the same 

 countries with the addition of southern Australia. A 

 species of Galaxias has recently been described from 

 India ; but Mr Blanford states that the determination is 

 doubtful. Now a fresh-water fish, even more if possible 

 than the purely terrestrial OligocJiceta, needs a continuity 

 of free land surface for its migrations 1 . And the fishes of 

 these two families form a considerable proportion of the 

 entire fish fauna of the countries where they occur. It 



/ 



is well known, of course, that the ova of many fish, par- 

 ticularly of Salmonoids, can be safely transferred in ice 

 and will hatch out at the other end : rivers in New 

 Zealand have been stocked in this very way. So that 

 there is no antecedent improbability in the iceberg theory 

 of the migration of these fishes. But in this case it is 



1 The possibility however must be borne in mind that they have been 

 independently evolved from marine ancestors. This can hardly perhaps 

 apply to Galaxias attenuates, to the same species. 



