We come now to a very vital problem in the physical geography of deserts, namely, 

 what becomes of such rain as does fall. The data we have on this important aspect of 

 arid areas are still very incomplete, nor have they been collected into any general 

 summary so far as I know. We know little as to the relation between evaporation, run- 

 off and absorption in deserts, which is hardly to be wondered at for not only are sta- 

 tions suitable for such observations very rare but the number of factors involved is so 

 large that data can rarely be of general application. It goes without saying that 

 evaporation must be high and that run -off on loose sand must be low or non-existent. 

 It is the remaining proportion which is so important and so difficult to measure. 



We cannot here become involved in the figures available from French engineers in 

 North Africa, American engineers from work in their drier states, and British workers 

 chiefly in India, since they do not mean very much without accessory observations of 

 a very local kind such as the nature of the ground, the rate at which the rain falls, the 

 temperature of air and ground, the wind, and even the time of day. All we can say is 

 that in most deserts there is some portion of the rainfall which succeeds in passing 

 through the upper layers beyond the reach of plants, there to form a water-table which 

 will appear as seepage springs in an oasis or can be tapped by bores. Much of it may 

 remain in partially enclosed areas underground, to become brackish and to cement the 

 sand grains into a calcrete or a silcrete. The term 'fossil water' has been used for 

 such occurrences but it is not a very useful phrase. Obviously there must be critical 

 points as to the amount of rain, its rate of fall, and the other factors mentioned, below 

 which no water is conserved beneath the surface, and this critical point will vary with 

 each desert and again at different places within that desert. 



It seems likely that the most critical factor may well be the rate at which rainfall 

 can seep downwards through the sand and soil, since it is a question of z race between 

 capillary action taking it back towards the surface and gravity leading it downwards, 

 beyond the reach both of plant roots and of the capillary rise, and the issue of the 

 race is largely dependent on the porosity. An accessory factor may be the imprison- 

 ment of air below the sodden layer after a storm; at least, that was my interpretation 

 of an observation we made in the northern, wetter, portion of the Kalahari. We were 

 boring with a hand- sampler, which in this case took us down 20 feet, and it was in 

 the middle of the rainy season. The first 3 feet produced damp, not saturated, sand 

 and then we suddenly ran into 10 feet of perfectly dry sand. At about 13 feet the sand 

 was again damp, and so continued to the limit we reached. From a nearby bore the 

 water-table here was at 40 feet. We took it that the zone of dry sand represented the 

 previous dry season and were surprised that six weeks of a rainy season, which on 

 average should have yielded about 8 inches, had here penetrated to only 3 feet. It was 

 the usual fine yellow Kalahari sand, which by experiment absorbed water at a much 

 greater rate than that. 



An isolated observation of that kind is of little value, but nevertheless one seeks 

 for an explanation, just as one wonders how the reservoir at 40 feet or so can get any 

 significant replenishment each year under such a regime.. 



The hydrology of deserts must sound to the layman like a contradiction in terms, 

 yet it is the study of a desert's water- supply which is the basis of this whole con- 



