38 DISTRIBUTION OF SOUTHERN FAUNAS, 



Caledonia and part of Polynesia. Subsidence in the Oligocene 

 and Miocene, followed by a third elevation in the Older Pliocene 

 when New Zealand was connected with the Chatham Is , Auck- 

 land Is., and perhaps others to the south, but did not stretch 

 north into Polynesia. This large island was broken up by sub- 

 sidence during the Newer Pliocene.* 



In 187-i Prof. A. Milne-Edwards presented to the Academy of 

 Sciences, Paris, a report on the fossil birds of the Mascarene 

 Islands showing that they were related to those of New Zealand. 

 As an explanation, he supposed that land communication had 

 formerly existed between these islands and New Zealand, which 

 was also joined to some islands in Polynesia, while it remained 

 separated from Australia. The connection with Polynesia was 

 to explain the occurrence of Rhinochetus in New Caledonia and 

 Didunoulus in Samoa. 



In 1876 Prof. H. N. Moseley supported Sir Jos. Hooker's 

 theory of a former greater extension of land in the Antarctic 

 Ocean t; and in the same year Mr. A. R. Wallace published his 

 " Geographical Disti'ibution of Animals," which treats of the whole 

 question. 



In 1880 Mr. Wallace published "Island Life," in which he 

 proposes the following hypothesis relating to Australia and New 

 Zealand. During the Cretaceous period, and probably throughout 

 a considerable portion of the Tertiary era, S. W. Australia (includ- 

 ing the southern part of S. Australia) was separated from Eastern 

 Australia by a broad sea, which contained some islands in what 

 is now Northern Australia. This western island had received its 

 mammalia at an earlier epoch from Asia, and no mammals existed 

 in Eastern Australia. New Zealand was connected with the 

 northern part of Eastern Australia, the land forming a horse-shoe 

 open towards the Tasman Sea. Probably the Bampton Shoal, west 

 of New Caledonia, and Lord Howe's Island formed the western 

 limits of this land; but it is possible, though hardly probable, that 



* Trans. N.Z. Inst. Vol. v. p. 227, and A.M.N.H. Ser. 4, Vol. xv. p. 25. 

 t Linn Sec. Journ. Botany, Vol. xv. p. 485. 



