BOTANY OF THE CEYLON PATANAS. 815 
rain, and would thus be enabled to descend the eastern slopes ; 
but it would undoubtedly at length reach a limit beyond which 
the rainy portion of the period of the S.W. monsoon would 
be too short for the maintenance of a forest of the western 
type. 
On the eastern boundary of the patana-region, at elevations 
not exceeding 4000 feet, there 1s a savannah-forest of which a 
description has already been given (vide p. 306). This forest 
flourishes in a climate which is almost identical with that enjoyed 
by the drier parts of the Uva patana-distriet. The peculiar 
feature of this savaunah-forest, as has been pointed out, is that 
there is a total absence of shrubby undergrowth, and in place of 
it a growth of coarse grasses. We have only to imagine a suffi- 
cient depth of soil on the patanas, and it is then easy to see that 
such a forest might have flourished all over the now barren 
grassy plains of Uva. This is supported by the fact that in some 
of the driest parts of the Uva patanas, where the conditions 
have been such as to allow of the accumulation of soil, a con- 
siderable growth of small trees, such as Dodonea viscosa, 
Eugenia sp, Mussenda frondosa, Osyris arborea, Flacourtia sp., 
Psidium Guajava, is found, This is strikingly the case in pro- 
tected hollows in various localities, and also in open places at 
Bandarawela, Wilson’s Bungalow, and elsewhere, where aban- 
doned termite-heaps provide accumulations of soil to which is 
confined almost all the shrubby vegetation. It seems that the 
termite-earths resisted the N.E. monsoon rains long enough 
to allow of the establishment of such vegetation upon them as 
now protects them from being washed away by the rains. 
It seems, then, a justifiable conclusion that the absence of 
soil on the greater part of the Uva patana-area below 4500 feet 
is of itself a sufficient reason for the absence of trees and tall 
shrubs; and, given a sufficient depth of soil, there can be little 
doubt that the whole of the Uva slopes below 4500 feet would 
bear a savannah-forest identieal with that which now flourishes 
in the Park-country to the east. If this be granted, it is not 
difficult to account for the disappearance of the forest and the 
soil which supported it, assuming only that the modern system 
of periodically firing the grasses has been practised for centuries 
by the Singhalese graziers—an assumption whieh is undoubtedly 
a safe one, although, owing to the peculiar conditions of the case, 
it can be supported by no direct evidence. 
LINN. JOURN.—BOTANY, VOL, XXXIV. z 
