60 Flora of Narcondam and Barren Island. 
volcanoes did not carry the chain beyond Barren Island, but Griffith, 
who in passing Narcondam recognised its volcanic natnre, suggested 
to McClelland that here might be seen a northward extension of the 
same chain. McClelland not only adopted the suggestion but sought a 
still further extension to the north, in the mud-volcanoes of Ramri and 
Cheduba, ofE the Arracan coast ;* * * § and other writers, such as Daubeny, 
Scrope, Mrs. Somerville and Mallettf have adopted the same view. 
Ramri and Cheduba lie to the west of a tertiary ridge that composes 
the Tomah of Arracan, which, in the latitude of Ramri, reaches a 
height of 4,000 feet. This range is continued southward into and 
beyond the Andaman group. Thus it passes through Diamond Island 
to the Alguada reef, beyond this, across a channel less than 60 fathoms 
deep, to Preparis, and again across another of 160 fathoms to the Coco 
Group, Great Andaman and Little Andaman. It would appear after 
this to pass to the westward of the Nicobars, though its precise relation¬ 
ship to that group has not yet been made clear ; finally it reappears, not 
in Sumatra, but in a long line of islands—the Nias group—that stretches 
south-eastward along the western coast of Sumatra.J The line of 
volcanic activity to which Barren Island and Narcondam presumably 
belong, lies from Narcondam southwards to the east of this tertiary 
ridge; if, therefore, Ramri and Cheduba belong to the same line, we 
have to believe that, after continuing for the whole length of Sumatra 
and the Andamans parallel to this ridge, the volcanic line at its nor¬ 
thern end, where its activity is weaker than elsewhere, crosses the 
tertiary formations where they have become thicker and stronger. 
This is in itself a proposition, the truth of which is so hard to accept, 
that when Blanford§ suggests that the true northern continuation 
of the Sunda volcanic range is to be found in the extinct Burmese 
volcano of Popah, and the extinct Yunnan one of Han-shuen-shan, we 
realise that he must be right, and are surprised that, after all. Mallet is 
inclined, in a modified sense, to favour the earlier view.|| The volcanoes 
of Ramri are of a different type from those of the Sunda Range; they 
belong to a series of gas vents, all of the same general character, though 
none of them so active as the Ramri ones. The Sitakund in Chittagong, 
* McClelland, Joum. As. Soc. Beng., vii., 77. 
t Mallet does this (Eecords of the Geol. Survey of India, xi., 203) in a different 
sense from the earlier writers; they, owing to a want of definiteness in the accounts 
on which they relied, mistook the “ gas ” volcanoes of the Arracan Coast for true 
“ steam ” volcanoes. 
ij; Kurz : Journ. As. Soc. Beng., xlv., pt. 2, 105. 
§ Manual of the Geology of India, iii., 725. 
II Mallet: Memoirs of the Geol. Survey of India, xxi., 253. 
274 
