392 Journal of Agricultural Research voi. x, no. s 



be added to the surface without any run-off, but in the extent to which 

 that from a light or moderate rain may be able to penetrate so deeply 

 into the soil that it will later be exposed to only the minimum of loss 

 through direct evaporation. 



HISTORICAL REVIEW 



Von Liebenberg (ii, p. 31) and Atterberg (4, p. 113) appear to be the 

 only investigators who have studied the downward movement in soils to 

 which a definite depth of water had been applied and then allowed to 

 penetrate as far as it would. Wollny (17, pp. 274, 288) carried out 

 some experiments where the soils were contained in tubes covered with 

 fine wire gauze and the water added drop by drop just as fast as it was 

 absorbed, thus determining the rate of penetration when an unlimited 

 amount of water is applied as rapidly as possible without actually losing 

 any by run-off. In various experiments the rate of percolation has been 

 studied, using a constant head of water, as i cm. in the work of Von 

 Klenze (9, p. 113) and 4 cm. in that of Kdler (7, p. 41). 



Von Liebenberg (11, p. 31), using 22 surface soils to illustrate a wide 

 range in texture, placed these in glass tubes of a diameter of 1.5 cm., 

 added i inch of water, recorded the penetration after periods of one- 

 fourth, one-half, one and one-half, four hours, one, two, and three days, 

 and finally determined the distribution of moisture at intervals of 5 cm. 

 and at the lower limit of penetration. He reports the mechanical analy- 

 sis, the hygroscopic moisture, and the loss on ignition. The hygroscopic 

 coefficients can not be computed (6, p. 69) from his mechanical analysis, 

 as he separated his soils into only the four fractions, sand above i mm. , 

 sand I to 0.5 mm., sand under 0.5 mm., and "fine earth," without sub- 

 dividing the last into silt and clay. In the case of the hygroscopic 

 moisture the percentages he reports are clearly not those which would 

 have been present had the soils all been in equilibrium with the atmos- 

 phere at the same time ; accordingly they do not permit the computation 

 of the relative hygroscopicities (i, p. 351). In fact, at least part of them 

 seem far drier than would be expected if they had been exposed for a 

 few days to the atmosphere of a storeroom or laboratory; thus, two 

 loess samples contained only 1.81 and 1.22 per cent, respectively, of 

 hygroscopic moisture (11, p. 13). In all his experiments he used the 

 soils in this dry condition. 



Our own experiments were in many respects very similar to those of 

 Von Liebenberg, with whose work we were not acquainted until after 

 we had completed our work. He, however, reports no data on the rela- 

 tive hygroscopicity of the soils he used and employed these in only a 

 very dry form. He concluded that the penetration is dependent upon 

 the "fine earth," it being less and taking place more slowly in soils with 

 larger proportions of clay and organic matter. 



Atterberg (4, pp. 11 3-1 24) determined the rate and distance of pene- 

 tration during i to 6 days of different amounts of water, equivalent to 



