1892.] 
119 
at least of the Nicobar Islands. Beyond the Nicobar group this system 
is continued southwards, not as is commonly said, (and perhaps generally 
supposed), into Sumati’a, hut into the Nias, a chain of islands lying along 
the west coast of Sumatra. 
The two volcanic islands under discussion—as well as a third peak. 
Flat Rock, which reaches the surface but no more at a point some 50 
miles south of Barren Island—are separated from the Andaman chain 
proper by a strip of sea 90 miles wide and over 1,000 fathoms in depth. 
In all probability therefore we see in these islands a northward con¬ 
tinuation of the chain of volcanoes that stretches upwards from Flores and 
Sambawa though the whole length of Java and of Sumatra to Barren 
Island and Narcondam and perhaps even beyond them to the extinct 
volcanic peaks of Popah in Upper Burma and of Han-Shuen-Shan in 
Western Yunnan. But even if this be true there is little doubt that 
both Barren Island and Narcondam conform essentially to the class of 
“ Oceanic ” islands and never have possessed any previous land-connection 
with neighbouring islands either to the south or to the west, or with con¬ 
tinental Asia to the North-east. The present physical conditions in 
Narcondam appear moreover to be very ancient; there is no trace of a 
crater at the top of its peak* which rises 2,330 feet above the level of 
the Andaman Sea, and the whole island is clad with a dense jungle 
much richer in species than the forest on Barren Island is. But though 
the present biological features of Barren Island are of much more 
modern aspect, is it not necessary to consider that island as really less 
ancient than Narcondam. The topography of its outer cone, combined 
with the historical fact of recent activity on the part of the volcano, 
points to the possibility of some catastrophe similar to that which 
devastated Krakatau having once happened in Barren Island. And if 
this has been the case it would follow that the island must have re¬ 
quired, even if previously covered with vegetation, to be stocked de novo 
with vegetable and animal life. Still, granting that the present fauna 
and flora of Barren Island are of more recent introduction than those 
of Narcondam, the fact remains that we must look upon every species 
present, even in the island with the older biological features, as an 
immigrant one. 
Dealing first with the birds since these may include both voluntary 
and involuntary immigrants, we And that of the eighteen species enum¬ 
erated, the Sea-Eagle, all the sea- and shore-fowl, the three Fruit-Pi¬ 
geons and the large Corby, making altogether ten species, or 55 per 
* The appearance of the hill indicates however that there never has been a 
crater. 
245 
