OCCURRENCE OF SMALL DESERT TRACTS. 167 
for several months of the year.* During the rains the desert 
tracts drain off rapidly to the rivers, as indicated by the 
narrow effluent ravines choked with sand (plate I., fig. 2). 
The deserts tend to assume the form of more or less ellip- 
tical depressions bounded by scarped faces, which are most 
conspicuous along the borders furthest from the main drainage 
line of the district. 
Approaching the largest and most typical desert tract (a) 
from the south one passes through dense forest of the dry-zone 
flora springing from a brick-red soil, which evidently makes a 
very fertile ground. The forest ends abruptly on the edge of 
an escarpment, the upper fifteen feet or so of which is a red 
loamy deposit, while the lower half is composed of sedimentary 
deposits (plate II., fig. 1). The barren tract is scarcely 
more than a quarter of a mile wide, but extends laterally for a 
mile orso. On the northern side of the desert the jungle starts 
again, and nearly half a mile beyond this point lies the Modera- 
gam-aar (Uppu-aar). 
The sedimentary rocks which form the floor of the desert 
tract reach down to the coast and extend scme miles inland. 
They vary in composition from compact limestone to arena- 
ceous and calcareous beds, in which limonite concretions are 
common. The compact limestone does not occur, except as 
pebbles and isolated fragments, in any of the barren tracts. 
During the early part of the monsoons strong winds blow 
across the country, catching up sand denuded from the rocks 
of the desert floor and carrying it along in scurries and whirl- 
pools. These effect much wear and tear of softer parts of 
exposed rocks, so that the harder bands and cencretions come, 
in time, to stand out in bold relief. As may well be imagined, 
the resulting surface is rough in the extreme and very tiresome 
to walk over. It recalls in some measure the less sandy parts 
of the great African deserts, which are locally such “ bad 

* At the time of my visit—February and March—the Pomparippu 
river was chest deep at the ford, while the Moderagam-aar was rapidly 
drying up. The Kall-aar was completely dry to its mouth. The 
country is almost uninhabited, and information concerning it is difficult 
to obtain, but I gathered from some fishermen at the coast that the 
Kall-aar never contains much water now-a-days. 
