98 Journal of the Asiatic Society of Bengal. [{N.S., XV, 
comparatively easy matter for the water to find its way into 
the Bay of Bengal by river valleys which were already in 
existence. This supposition seems more likely than that the 
watershed was depressed by local subsidence, or faulting, or 
that the reversed rivers were able by their own force to cut a 
way through the watershed to the sea. 
It is evident that the Godavari and the Kistna have at 
some period cut back through crystalline rocks which may 
have formed an outlying portion of the same watershed, since 
they now rise in the Western Ghats. 
Similarly the Brahmaputra may have cut back through 
it in another place and captured the headwaters of the old 
Siwalik river, either at the time when the reversal of drainage 
took place or even earlier than this. 
€ may suppose that the portion of the Siwalik river 
which drained Kangra and Jammu reached the Arabian sea in 
a somewhat similar fashion. Ever since the origin of the 
Indus its Punjab tributaries, the Jhelam, the Chenab, the 
Ravi, the Beas and the Sutlej, must have been cutting back 
towards the north-east, and by the Pleistocene may have cap- 
tured a considerable portion of the drainage of the Siwalik 
river from off the Aravallis and Rajputana. 
The Pleistocene pebbly- gravels which have been found 
here and there on the margins of, or as islands in, the Gangetic 
alluvium, as for example at Allahabad and Sara, are doubtless 
the first deposits of the new southerly flowing rivers before 
they had regularly established themselves on their downward 
gradient to the Bay of Bengal. Ata later stage this class of 
deposit ceased and was replaced by the modern annual accu- 
mulations of silt. 
tion known or inferred which is at our disposal concerning the 
elevation of the Himalayas. On the other hand, so far at least 
