The Occurrence of Dinosaurs in Bushmanland. 
271 
rather over 120 feet to the mile, which is a very steep slope for the lov^er 
end of a tributary draining at least 2,000 square miles of country. The 
slope of the valley above the water is also very great for the first 10 miles 
or so. 
The surface of the ground in the valley is a coarse sand or grit derived 
from gneiss, but the narrow slopes between the terraces are made of damp, 
muddy, and calcareous sand, very often seen to be resting on calcareous 
tufa, which frequently crops out over hundreds of square yards. These 
damp belts are more or less covered with vegetation, especially rushes 
and mimosa thorns, while the sand between and above the damp belts is 
almost bare. Water is got on the bush-covered belts at a very shallow 
depth, and often appears at the surface, so there are swampy patches on 
them. The slope of the surface on the belts is great, while the stretches 
of sand between them are very gently inclined. The water of the top belt 
is brak, but quite good for drinking, though it leaves a distinct white 
efflorescence on the ground after evaporation ; the waters below the top 
terrace become more and more salt as one goes down the valley, and at 
the time of my visit (May, 1913) that of the 4th belt and below seemed to 
be too salt to drink. A small stream issued from the 4th belt and made 
its way down to the river, being strengthened by additions from the 5th 
and 6th belts. This stream was finally half an inch deep and a foot wide, 
and it was the only stream seen entering the Orange Eiver on the left 
bank between Eaman's Drift and the mouth, a distance of nearly 200 
miles along the river. 
Outcrops of gneiss on the valley floor are first seen below the 5th belt. 
The calcareous tufa of the belts is a very sandy rock cemented by white 
limestone ; in the larger outcrops where springs have cut channels it is 
seen to be roughly bedded, and it was evidently formed by the cementation 
of the sand of the valley at the places where there has been long-con- 
tinued evaporation of brakwater. A good explanation of the formation of 
the terraces has not been found, but it seems clear that the progressive 
saltness of the water is due to evaporation at the surface in each belt ; the 
water appearing at the 1st belt sustains loss by evaporation, sinks under 
the sand and reappears at the lower belts, each time with a greater pro- 
portion of salts in solution. Whether the position of the belts was 
determined by the shape of the rock floor is uncertain, but wherever the 
water flowing underground maintained a level very near the surface 
during former stages of the valley's history, plant growth must have been 
encouraged, and this growth checked the downward travelling of sand, 
thus tending to form the terraces of low slope between the belts. 
Evaporation of the water brought about deposition of carbonate of lime 
between the sand grains, and such action would obviously have been 
greater in the damp belts than in the terraces between them. Erosion by 
