ON AN INVESTIGATION OF THE BATOKA GORGE. 297 
resolved into two great curves, the first with its convexity toward the 
south, and the second, convex toward the north. 
Except in the upper part of the southerly bend below the Falls, the 
general direction of the numerous tributaries which join the Zambesi 
from the north in the country traversed is from north-west to south-east, 
though the deep ravines into which they plunge on approaching the 
Zambesi often reproduce in miniature the erratic zigzags of the main 
gorge. In its wider aspect this part of the country exhibits a tree-clad 
plateau of basalt sloping gradually toward the south-east, edged on the 
north by broad smooth ‘ bults’ of thick red sand resting on the basalt and 
cJad with larger trees. This plateau is broken by wide shallow valleys of 
low gradient, whose flat bottoms are generally covered with two or three 
feet of blackish loam which supports a luxuriant growth of tall rushy 
grass without trees ; but as they approach the Zambesi these open valleys 
break away into the deep precipitous ravines already described-; and it is 
important to note that the length of these ravines was found to increase 
steadily as we journeyed eastward from Victoria Falls. 
It is clear that the open valleys represent a mature stage of the tribu- 
taries in their relation to the trunk-drainage, similar to that which still 
exists in the tributaries that join the Zambesi above the great Falls, and 
that the caiions in which they now terminate are a measure of the 
erosion which has taken place since the streams were rejuvenated by the 
gradual backward excavation of the Batoka Gorge. Owing to this reju- 
venation a gradually widening belt of the plateau which borders upon the 
main gorge has become exceedingly rent and stony from the washing 
away of the surface soil down the spreading network of gullies ; but the 
level spaces of limited extent that have not yet come under this influence 
possess, where not covered with the surface-sands, a firm red lateritic soil 
derived from the decomposed basalt. 
That the Batoka Gorge itself has been carried gradually backward into 
the heart of the plateau by the erosive agency of the Zambesi in its rapid 
descent from the basaltic upland is further shown by the falling off in the 
angle of slope of its sides as we descend the river. At Victoria Falls, 
where the gorge is freshly cut, its walls are practically vertical ; buta few 
hundred yards below, they are already beginning to show the effect of 
weathering by the slight recession of their crest line and by indications of 
terracing along the planes of stratification. Seven miles farther down, 
this recession and terracing become so pronounced that the average angle 
of slope from base to crest is reduced to 60° or less ; thirty miles below 
the Falls, it is no more than 35° ; and farther eastward, the sides of tho 
gorge have been weathered down into bushy slopes, broken here and there 
by low bars of crag, with an average inclination of 30° or under. 
The basaltic country traversed south of the Zambesi, while presenting 
the same general features as that north of the river, differs from it in some 
respects, owing to the comparatively great antiquity of its low-level drain- 
age. The Matetsi and the Deka run much more obliquely to the general 
course of the Zambesi than do the northern tributaries. Thus the Matetsi 
leaves only a narrow wedge of country to drain directly to the main river, 
and itself draws many streams from the southern side of this wedge. It 
makes its confluence with the Zambesi, as we have already seen, below the 
eastern termination of the great gorge, while the Deka’s confluence lies stil] 
farther eastward ; so that these rivers have had time, not only to carve out 
long deep valleys into the plateau, but to develop wide drainage basins 
