V 
The Occurrence of Dinosaurs in Buskmanland. 269 
necessary for the reduction of any given tract of hilly country to a 
uniform surface through the action of wind and rain under semi-arid 
conditions. Figures have been obtained representing the rate of denuda- 
tion in certain drainage basins, and an average rate of lowering of the 
land surface can be got from them ; but the conditions vary so greatly 
that even for the present time the value of such an average is quite 
uncertain. These estimates are based on the area of the basins and the 
volume and contents of the outflow, and they are the only measure of 
denudation yet attempted, but they are not applicable to a region like 
Bushmanland, where there is no outflow. 
From the general surface features of the Kalahari region and the 
nature of the superficial deposits, Dr. Passarge t argued that the whole 
region owes its characters primarily to the prevalence of desert conditions 
during post-Karroo times, though wetter climates are supposed to have 
intervened. The evidence of fossils has hitherto been wanting, but the 
general truth of Passarge' s views has not been questioned. The discovery 
at Kangnas afi'ords some confirmation of those views and also grounds for 
expectation that abundant evidence bearing on the problem awaits dis- 
covery ;■ though the facts at present do not support the hypothesis of 
markedly varying climates. 
Bushmanland is really the southern end of the Kalahari region 
separated from the main portion by the valley of the Orange Eiver, which 
is fed by a country with much higher rainfall far to the east, and it is very 
probable that under present conditions the river always loses more water 
by evaporation than it receives between the Hartebeest Eiver and the 
Fish Eiver, a distance of 340 miles ; usually, in fact, that statement must 
also be true of the much longer stretch from the confluence of the Vaal 
and Orange down to the sea. 
The fact that some unknown thickness of Karroo beds was removed 
by denudation over part of Bushmanland and that valleys were cut 
in the underlying gneiss before the dinosaur bones were buried proves 
that a wetter climate prevailed there before the valleys began to be 
filled up. 
The Kangnas valley is a part of the river system which enters the 
Orange Eiver at Henkries. The last 8 miles of the valley lie between 
steep, bare, rocky hills which rise abruptly from the surface, and small 
kopjes appear above the surface of the wide valley. From the relation 
of these hills and kopjes to the wide but sloping valley floor, it is obvious 
that bed-rock lies far below the present surface. (See Plate XXXVI.) The 
* See the table of results for nine rivers in Chaniberlin and Salisbury's " Geo- 
logy," vol. i., p. 101, and the discussion in Penck, " Morphologie der Erdoberflache," 
379-385. 
t See Passarge, "Die Kalahari," 1904, ch. xxxv. 
