342 
F. Kingdon Ward. 
which extend to well beyond the limit of trees. In this case the 
trees end sooner in the valley-bottom than they do on the enclosing 
mountain slopes, due to the concentration of the wind along the 
valley-bottom as the mountains rise higher, since the enclosing ridges 
usually rise more rapidly than does the floor of the valley (PI. 7, Fig. 3.) 
The shrub vegetation on the other hand, usually represented at 
these altitudes by dwarf Snlix, extends further up the valley-bottom 
than it does up the mountain slopes (where the place of Snlix is 
taken by Rhododendron), for the reason that the water-supply con¬ 
centrates in the valley-bottom. 
Wind and water are indeed the two continuously warring factors 
which determine the formation and its limit, the selection of a 
certain type of vegetation being dependent on its invulnerability to 
the ravages of one or lack of the other: when at last they combine 
forces they quickly overwhelm all forms of vegetation. 
To sum up, the tree-limit varies between approximately 14,000 
and 15,500 feet, with local peculiarities of distribution dependent 
on topographical details; it is determined by the level at which 
absorption of water, more and more retarded by removal of the 
protecting snow-blanket and consequent lowering of the soil-tem¬ 
perature, is finally overbalanced by the rate of transpiration, more 
and more accelerated by a prevailing wind. 
In north-facing valleys pre-eminently, the forest-belt, which 
here begins with deciduous-leaved trees and includes Larix , up to 
the limit of trees, passes rather abruptly into a scarcely developed 
belt of dwarf Rhododendron immediately giving place to alpine 
grassland ; in south-facing valleys, the forest-belt, which towards 
the tree-limit comprises only evergreen Conifers, passes gradually 
into a well-marked shrub vegetation, which in turn gives place to 
dwarf Rhododendron and alpine grassland. 
Valleys facing in intermediate directions are dependent for 
their characteristic vegetation not so much on the presence or 
absence of the prevailing wind as on the fact of its blowing up or 
down the valley, a circumstance varying with the topographical 
features of the surrounding country. 
While valleys well exhibit the effects of wind and water both in 
the composition of their flora and in the different altitudes at which 
a particular formation is curtailed, such differences are, as already 
indicated, still more clearly marked in the case of mountain ridges ; 
and here the problem is simplified by the appearance of an entirely 
new formation, namely desert, 
