F. Kingdon Ward—The Land of Deep Corrosions. 215 
then any cessation or diminution of that pressure might allow the 
overlaying strata to sag back. The result would be that on the 
steeper slopes—for the slope from west to east is manifestly not 
uniform—the massive anticlines, dragging on the synclines, would 
cause the latter to crack, and rifts such as the parallel rivers now 
occupy would take the place of simple anticlines, accounting at the 
same time for the differences of level at which the rivers flow. The 
rifts once formed would be deepened by erosion, the valley walls 
being protected by the aridity enjoyed, which is itself increased 
automatically by the up-valley winds. It is, however, evident that 
these aggravating factors alone are not sufficient to account for the 
rifts in the first instance, since the Salween, south of latitude 28° in 
a monsoon region, flows through a very narrow valley, the rocks in the 
bed of which in all respects resemble those further north, and in 
places through gorges modified by the heavy summer rainfall; still 
more does the upper ’Nmai-hka (or Taron as it is now called) flow in 
a gorge, though the region is drenched with rain almost all the year 
round, 
Or, we may account for the rifts, and perhaps also for the marked 
differences of level, not by any lateral movement of the Himalayan 
axis, but by the elevation of that range itself, setting up a great 
tension strain at right angles in the adjacent crust, causing a sequence 
of parallel rifts to appear during the second phase referred to above, 
followed by an irruption of molten rock along the lines of weakness. - 
Lines of volcanic activity can be followed southwards from Batang 
through Yunnan (‘I‘engyueh volcano), Burma (Popa), the Arrakan 
Hills, and the Andamans (Barren Island), to the Indies, where the 
voleanic forces are at present concentrated, having forged southwards 
like a line of fire. In the case of a tension strain like that, the 
region most affected would be that nearest the seat of force, that is 
in the west nearest the Himalayas, where the broadest and deepest 
valleys would be formed; the least affected would be those furthest 
away, in the east, which would consequently be narrowest and 
highest. But while the rift formation may be due to this cause, 
and not to any pushing of the crust up a gently inclined plane with 
subsequent sagging back, it does not affect the synclinal origin of 
the great breach in the Sino-Himalayan axis, and the successive 
elevation of the parallel divides, the most easterly being those 
first elevated. This I now regard as proved from botanical and 
geological evidence. 
It will be noticed that there is a tendency for the river valleys not 
only to lie at higher elevations, but also to become narrower, so far 
as is consistent with their different powers of erosion, as we go 
eastwards. Thus the Brahmapootra valley, once the Himalayan 
axis is cut across, is broader and flatter than that of the Mali, and the 
Mali than that of the ’Nmai-hka. There is probably little to choose 
between the gorges of the upper ’Nmai-hka (Taron), the Salween, 
and the Mekong, as regards breadth,’ but the Yangtze valley is 
rather broader, the river itself being quite twice as broad as the 
Mekong. If no disturbing factors intervene the western rivers will 
eventually tap the eastern, owing to the cutting back of the tributaries 
