38 Pehr Ohson-Sefier. 
tation, is so effective that except for the counteracting influence of 
surface-tension very little water would be retained in the sand. 
The permeability of sand is variable according to the texture of the 
soil, and it increases as the particles increase in size. Lysimetrical 
measurements given by Wiley 1 show that in ten hours’ time under a 
water-pressure of 10 cm., 0-151 liters will pass through a layer of sand 
10 cm. thick of 0-01—0-71 mm. grain-diameter, while with a water- 
pressure of 100 cm., 0-596 liters will percolate through the same sand 
10 cm. deep in the same time. With a sand of a coarseness of 
0-071—0-114 mm. the corresponding amounts of water filtrated are 
2-194 and 8-530 liters. 
Comparing sand with soil of finer texture we find that the 
greater permeability of the former depends on the fact that the 
interspaces between the soil-particles are so large that they are 
not filled with the fine sediment which is brought down from the 
surface with the sinking water, while in the case of clay its almost 
complete imperviousness is due to the inability of the soil-air to 
escape upwards through the super-saturated surface. So long as this 
air cannot escape it prevents the water from entering the soil-spaces 
occupied by it (King). In fine sand the water therefore does not 
sink so rapidly as in coarser, and it can be observed after a heavy 
rain that on those places on a dune where the finer material is 
accumulated, the water has not penetrated so deep as where coarser 
grains compose the bulk of the sand. The air has been confined 
under the water, and impedes percolation until an escape is affected 
either upward or laterally. This circumstance makes it difficult to 
use cylinders or vessels with closed sides for experiments on the 
downward movement of water, and by allowing the air to escape 
laterally the source of error will always be considerable. The 
values obtained by such experiments are therefore relative, and cannot 
be compared unless the experiments are conducted under conditions 
as nearly similar as possible. I have made a number of attempts 
to determine the relative rate of percolation in sands of different 
coarseness, but the results have not satisfied me as being a true 
index of the power of filtration, because of the difficulties in 
establishing conditions approximately similar to those in nature. I 
do not therefore feel justified in mentioning the results obtained, 
and I must add that in the case of all the data I have seen in the 
literature, those quoted above included, we have to accept them with 
1 Principles and practice of agricultural analysis, 1895, Vol. I., 
p. 163. 
