FISHES—REGAN. 41 
Tertiary land-bridges to account for all cases of this sort would reduce the oceans to 
a few puddles. 
Much has been made of the distribution of the Galaxiidae and Haplochitonidae, for 
some time regarded as fresh-water fishes found in Southern Australia and Tasmania, 
New Zealand, and the southern part of America. It is now known that Galaaias 
attenuatus, the only species common to all these regions, breeds in the sea. In the 
“ Scotia” report | have shown that these two families are Salmonoids related to the 
Osmeridae, and their marine origin may be regarded as certain. Like the northern 
Salmonoids they are establishing themselves in fresh water, and it is interesting to note 
that Galaxias occurs at the Cape and even in New Caledonia, where, like the Trout of 
Algeria, it remains as the witness of a glacial epoch. 
None of the families of true fresh-water fishes of either South America or Africa 
occurs in Australia, except the Osteoglossidae, a generalised and ancient type. Even in 
this case the relationship is not with America or Africa, but with Asia, Scleropages 
comprising one species from Queensland and New Guinea, and one from Borneo and 
Sumatra. 
Thus neither marine nor fresh-water fishes support the theory that the Antarctic 
continent connected America with Australia during the Tertiary Period. This being 
the case, I have been led to examine somewhat critically the other zoogeographical 
evidence in support of this theory. Some of this is derived from the similarity of 
marine faunas, or from the distribution of fresh-water organisms that may have had a 
marine origin. Of more importance are the land animals, and Dollo, in his monumental 
report on the “ Belgica” fishes, comes to the conclusion, “ C’est (Antarctide Tertiaire 
de M. Osborn—ou une Antarctide analogue, indispensable pour les Marsupiaua et 
Miolania—qui explique le mieux la Biogéographie des Poissons Antaretiques et Sub- 
antarctiques.” Unable to accept this for the Fishes, I have looked into the question of 
the Marsupials and JMiolania. 
MARSUPIALS. 
It has been suggested that Caenolestes and the extinct Patagonian members of the 
Epanorthidae may be related to the Australian group Diprotodontia. This is by no 
means generally accepted, and several authorities believe that the Epanorthidae may 
have been derived from a primitive Didelphoid type which has evolved a diprotodont 
dentition independently. This view is supported by the fact that the Epanorthidae are 
eleutherodactyle, whereas the Australian Diprotodonts are syndactyle.* 
It has recently been shown by Gidley f that the Multituberculates are “ Dipro- 
todonts,” so that Marsupials of this type date back to the Triassic, and even if they be 
monophyletic, their occurrence in Australia and South America loses its significance. 
* Thomas, Proc. Zool. Soc. 1895, p. 870 ; Gregory, Bull. Amer. Mus. xxvut, 1910, p. 211. 
+ Proc. U. 8. Nat. Mus. xxxvi, 1909, p. 611. 
G 
