30 CONTRIBUTIONS TO THE ORNITHOLOGY OF INDIA. 



It must be understood that in " the islands of the Bay of 

 Bengal," I only include Preparis, the Cocos, the Andamans, 

 the Nicobars, Barren Island, and Narcondam. I do not 

 include the islands of the Mergui Archipelago, which I con- 

 sider may be treated as an integral part of the Tenasserim 

 Provinces, which we are now working, and in regard to w T hich 

 I shall hope to have something to say next year. 



The accompanying sketch map will show, sufficiently clearly 

 for our purposes, the position of these islands. It will be seen 

 that, neglecting Narcondam and Barren Island, the whole of the 

 rest of the islands constitute a broad, irregular, curved chain 

 (many of the links of which, however, are broken and missing) 

 connecting Cape Negrais, the south-western point of Pegu, 

 with Acheen Head, the north-westerly point of Sumatra. 



Close to Cape Negrais terminate the Arrakan Hills, part 

 of a series of ranges which run down from the Eastern 

 Himalayas dividing Assam, Cachar, Chittagong, and Arakan 

 from Independent Burmah and Pegu. 



Just south of Acheen Head Ave have the Golden Mountain 

 and other hills, and looking at the map it seems difficult to 

 avoid the conclusion that the whole of the chain of islands 

 above referred to. is nothing but a continuation southwards of 

 these Arakan Hills, of which, owing to a general subsidence of 

 the tract of country they traverse, only the more elevated por- 

 tions now remain above the sea level. 



It cannot, however, I fear, be asserted, that this very simple 

 and obvious explanation of the origin of these islands derives 

 prima facie much support from a consideration of their fauna ; 

 and if they ever were in uninterrupted connection with the 

 Arakan Hills, it must, apparently, have been at an immensely 

 distant period, since not only are almost all the most character- 

 istic species of the Arakan Hills, as we now find them, absent 

 from these islands, but these latter exhibit a great number of 

 distinct and peculiar forms, constituting where the ornis is 

 concerned, if we except the cosmopolite waders and swimmers, 

 considerably more than one-third of the whole number known. 



As for Narcondam and Barren Island these lie altogether 

 outside of the main chain. Both are entirely volcanic, and the 

 latter a still smoking volcano, and it is curious that in the 

 valley of the Irawaddy, hot springs and other evidences of 

 volcanic action occur in the same relative position to the 

 Arakan Hills that these two islands occupy in respect to the 

 Andamans. There seems little doubt that both these islands 

 belong to the same great line of volcanic disturbance that 



