Underground Water in South-Eastern Bechuanaland. 259 



even if they are followed by intervals of lesser rainfall. This has 

 been shown clearly by experiment * as well. 



Along ridges and watersheds very little moisture is able to pene- 

 trate the soil, but in hollows and valleys the conditions are very 

 much more favourable. 



(3) Evaporation and transpiration. — By far the greater loss^ must 

 occur directly from the surface, more especially as the rain falls 

 during the heat of summer. The rate of evaporation depends not 

 so much upon the temperature as upon the humidity of the air. 

 There are no records available for this area, but at Kimberley, 

 where the conditions are very similar, the mean annual humidity 

 has the low value of 55 per cent. 



Not only does evaporation take place directly from the surface 

 after rainfall, but moisture is brought up from below ground by 

 capillary action and so dissipated ; the formation of calcareous tufa 

 by this means has already been noted. 



Some experiments made by Prof. P. H. King f show that the 

 capillary movements of water are considerable even at a distance of 

 4 feet below ground ; hence moisture is readily brought up from such 

 a depth and evaporated. The importance of the rapid descent of the 

 moisture through the soil is thus made apparent. They also indicate 

 what has been proved in other ways, that, in the case of light 

 showers, all the water is brought back to the surface and evaporated 

 before it can get below the critical depth. Another source of loss is 

 the transpiration of moisture by vegetation. Bechuanaland is fairly 

 thickly clothed, and the roots of some of the thorn trees descend to 

 depths of over 80 feet in search of nutriment. The amount of 

 moisture transpired by a tree of average size is estimated at from 

 2 to 2J gallons daily, but I am unable to obtain any figures for the 

 transpiration losses per square mile of wooded country. The 

 removal of water by vegetation must, however, take place on an 

 immense scale annually. 



Proportion of rainfall retained. — From the preceding paragraphs it 

 is clear that this depends upon quite a number of factors, e.g., 

 whether the rainfall is above or below the mean, upon the porosity 

 of the soil and the rock underlying, upon the depth down to 

 decomposed material and thence to unfissured rock, upon the slope 

 of the surface, upon the amount of vegetation it supports, &c. 



I think, though, that if we consider the annual rainfall to be 



* Prestwich, " Water Supply of London," p. 113, 1895. 



f King, "Nineteenth Annual Keport United States Geological Survey," vol. ii., 

 p. 85, &c, 1899. 



