The results thus obtained do not, of course, apply directly to the same soils undisturbed in place in the field; where, moreover, the air is confined by the wetting of the surface and thus directly opposes penetration of the water. Still, they doubtless give a correct idea of their _relative_ permeability for water when in the tilled condition. The water level was automatically maintained at the depth of half an inch above the surface of the soil columns. Pore-spaces given are calculated from volume-weight and specific gravity. This diagram shows plainly that there is no direct relation between the total pore-space in a soil and the facility of water-penetration. The highest pore-space, in the fine-grained alluvial loam, allows more rapid percolation than the heavy clay or adobe soil, but is greatly exceeded by the coarser sandy soil. In all it is very apparent that the downward movement slows down as the water descends, doubtless because the great friction in a longer column gradually diminishes the effect of hydrostatic pressure. It may be presumed that at a certain distance from the surface the downward movement becomes practically uniform, and independent of the pressure from above. [Illustration: FIG. 41.—Diagram showing differences in rates of percolation through different soils.] _Summary._—Two salient points are revealed by even a cursory inspection of the preceding diagram, viz.: 1. The downward percolation is most rapid in the same soils in which the capillary ascent is quickest, that is, in the coarse, sandy soil. 2. The rapidity of percolation decreases materially as the wetted soil column increases in length. The first point is readily foreseen and needs no comment. As regards the second, it results from the fact that as the wetted column lengthens the frictional resistance increasingly counteracts the effects of the hydrostatic pressure from above, until the water’s descent becomes but little more rapid than would be its lateral diffusion, or its ascent at the end of a similar column supplied by capillary rise from below. In both cases the frictional resistance has so far counteracted the effect of gravity that the capillary coefficients of the soil-material become the controlling factors of the water movement. _Influence of Variety of Grain-sizes._—King (Physics of Agriculture, pp. 159, 160), compared the rapidity of the percolation of water through definitely graded pure sands on the one hand, and a sandy loam and a clay soil on the other. The materials were arranged in 8-foot columns fully saturated with water at the outset, and then allowed to drain freely. The following abridged table shows the tenor of his results: