149 



not act as such an efficient protection to the soil as the denser- 

 closer climax types ; (2) by laying- bare the surface of the soil 

 at the time when the first heavy rains fall. 



The effects of stock grazing- are even more marked, and 

 these are largely increased, if the stock are allowed to roam 

 about at will over uufenced land. The surface layers of the 

 soil are hardened and water does not soak in so readily. 

 Cattle and sheep tracks are formed along which the water runs 

 off, and these gradually deepen, and ultimately form dongas. 

 These drain the rest of the soil, leading to a permanent lower- 

 ing of the water table and a drying out. Pronounced changes 

 in the veld result usually, once more, in the direction of estab- 

 lishing more primitive xerophytic open types of grassland. 



The obvious remedy is to remove, as far as possible, the 

 causes. Burning should be avoided as far as possible, and 

 the stock should be prevented from tramping over the veld to* 

 an unnecessary extent. Under present conditions, it is not 

 possible altogether to remove the causes mentioned, and our 

 study of plant succession throws light upon other remedial 

 measures. There are certain plants which are peculiarly 

 adapted to fixing and binding the soil in incipient dongas. 

 Most of those, which are described in the previous section of 

 this work as belonging to the hydrosere, are of this nature. 

 Some introduced species, e.g., Paspalum dilatatuni, will be 

 found very useful. The stoloniferous species of Sedges, and 

 the spreading rooted or stoloniferous grasses are particularly 

 useful. Everything should be done to encourage their growth, 

 and above all things the vegetation in growing dongas should 

 never be burnt. By the time the donga has deepened into a 

 fairly large ravine trees and shrubs have usually taken pos- 

 session — a list of the most common species is given above — 

 and denudation has proceeded too far to be checked by grasses 

 and sedges. It may then be caused to silt up by placing 

 obstructions at intervals across it. The plant succession then 

 goes through the various stages of the hydrosere as outlined 

 above, back ultimately to the climax stage of veld. Examples 

 of dongas, which have thus in the past become filled up with 

 alluvial gravels through tree growth blocking the mouth, may 

 be found in Natal, and often in the centre of them denudation 

 has commenced afresh, new dongas cutting their way down 

 through the gravel. 



The controlling of the flow of water in various ways helps- 

 to prevent dongas from forming, but This aspect of the subject 

 is hardly a botanical question. 



