534 J. w. BEWS. 



"oases" where the grasses are not dominant. On the buttress- 

 slopes the soil, which is the direct product of the underlying- 

 lavas, is a comparatively thin layer, rarely more than a foot 

 in depth. The slope is often very steep, yet the soil does not 

 get washed away, though it often slips, leaving lenticular 

 bare patches on the hillside, or a large boulder from the cliffs 

 may become loosened and tear away a longitudinal strip of 

 soil right down the mountain-side. On the whole, however, 

 one is surprised at the fact that soil accumulates at all, and 

 that the whole slope is not washed bare, as would happen in 

 temperate or sub-arctic regions. There is no heavy covering 

 of snow to cause landslips when it melts, and the heavy 

 thunderstorms affect mostly the mountain - top, the water 

 being rapidly drained off, through the numerous ravines, 

 without disturbing the intervening buttress -slopes. The 

 tussock grasses themselves of course help to bind the soil. 



Tussock veld may also be observed invading the river-beds. 

 In the upper parts of the river-bed, where the slope is steep, 

 it has to compete with scrub ; but lower down, as the valley's 

 widen and the river-bed is flatter, soil accumulates, and 

 usually the veld gains the upper hand, probably because the 

 climatic factors, especially the valley frosts, are unfavourable 

 to tree growth. The pioneer grasses are usually the wiry 

 species of Aristida and Koeleria cristata. These grow 

 in isolated tufts among the boulders, and fix a certain amount 

 ot" soil. This gradually accumulates until the boulders get 

 almost buried, and then taller and coarser species invade, 

 e.g. species of Andropogon. Still lower down, the rivers 

 form alluvial plains, and the veld is of the Low Veld type, 

 described in former papers, with Anthistiria dominant. 



(2) Associated Veld Plants and Tkansitional Types. 



In previous papers the writer has given somewhat extensive 

 lists of plants which occur associated with the grasses in the 

 veld. At the same time, it was remarked that, in typical 

 veld consisting of pure Anthistiria association, they were 



