THE ISLANDS OF THE BAY OF BENGAL. 135 



It seems to me impossible to avoid the conclusion that the 

 ornis of these islands has' altogether a very far stronger affinity 

 with that of the Indian region, than with those of either 

 the Indo-Burmese, Indo-Malayan, (which inosculates with the 

 former north and south of Moulmein) or Archi pelagian. Yet 

 this involves very great difficulties ; for, in the first place, these 

 islands seem to form parts of a chain of mountains, once con- 

 tinuous from Arakan to Acheen ; and in the second place if we 

 take Port Blair as a centre, we shall find that its average 

 distance in all directions, north and east, from Tenasserim 

 (where the Indo-Malayan fauna preponderates), and north of 

 this from the Indo-Burmese sub-region, is only about 350 

 miles, while per contra its distance, from the nearest points 

 of the Indian sub-region, all round from Calcutta to Madras, 

 is 900 miles. 



That so many of the most characteristic birds of the Ara- 

 kan hills, especially amongst the Rasores, should be entirell 

 wanting in these islands, we may partly account for by 

 supposing that the chain of mountains never was continuous, 

 and that the same agency that raised the Arakan hills only 

 raised portions of their continuation, between these hills and 

 Sumatra, above the sea level, and that therefore these islands 

 never were directly connected either with Acheen or Pegu. 

 This might, of course, be true ; but looking to the fringing reefs 

 of the Andamans and Cocos, and to many other indications, I 

 confess that I should be disposed to believe that these islands 

 belonged to an area of subsidence, and that the}' are even now 

 sinking. But grant that the other is the correct view, that these 

 groups first made their appearance as, and have ever since remain- 

 ed, islands, detached alike from Arakan and Sumatra, even then 

 it is inconceivable how the great bulk of the work of coloniza- 

 tion should have gone on from a region 900 miles distant, while 

 so little should have been done from others separated by little 

 more than one-third of that distance.* 



The prevalent winds of these seas will in no ways explain 

 or help to explain the difficulty, on the contrary they altogether 

 intensify it, since their influence would be distinctly more 

 favorable to emigration from south-west, east, and north-east 



* Indeed I do not feel sure that some colonization has not gone on from 

 the Nicobars to the Indo-Malayan region instead of the converse. Palceornis 

 appears to me an eminently Indian genus ; and it is not impossible that 

 both Luciani and longicaudatus are modified forms of erythrogenys, ov 

 affinis, themselves modifications of one of the existing Indian races, or com- 

 mon offshoots with them from some antecedant Indian one. 



