In the usual case of infiltration into a soil, the flow resistance of the expelled air is negligible compared with the resistance of the water to flow. If the air escape route deeper into the medium is blocked by an impermeable barrier, however, the air will be trapped and its pressure will increase as the water fills the pores. This back‐pressure of the air will decrease the infiltration rate, and if there is a horizontal, truly impermeable barrier at some depth the interface may eventually stop moving. In the analytic approach to the problem it is assumed that the medium is saturated to the wetting front and that all capillary forces are acting at the water‐air interface. Experiments with vertical columns of glass beads and fine sands generally verified the analysis, but it was observed that for mediums with particles larger than 0.30 mm diameter the air escaped upward into the atmosphere through the wetted medium and the interface kept moving downward.

The performance of the Green‐Ampt model for infiltration problems where the depth of water above the ground surface is varying with time is investigated. In order to yield infiltration rates that agree with those predicted by the Richards equation for flow in a homogeneous, nondeforming, and nonhysteretic soil the effective suction head parameter in the Green‐Ampt model must be considered a function of time, surface water depth, initial moisture content, and soil type. However, when a constant value of the effective suction head is assumed, the response of the Green‐Ampt model to variations in surface water depth is qualitatively equivalent to the response of the Richards model. The effectiveness of a number of definitions of the suction head parameter that have been proposed in the literature is investigated. While the differences in predicted infiltration rates and cumulative infiltration obtained by using the different definitions are, in general, of marginal significance, the best choice of the value of the effective suction head depends upon the particular problem to which the model is being applied.

A solution to the two‐dimensional diffusion equation for unsteady and unsaturated flow in soils has been obtained. The physical problem is that of flow from a cylindrical source of finite radius and infinite length. Neglecting gravity, the diffusion equation is reduced to a simple form which can be transformed to a nonlinear ordinary differential equation. This transformation is achieved after substituting an empirical relation for the diffusivity of the soil. This ordinary differential equation is then solved by numerical techniques. Two soils, Yolo light clay and Pachappa loam, have been used for examples. Infiltration rates have been computed in each case; these are found to be almost constant for this problem. A computer model in extended Algol language has been developed.

If the resistance to flow through a granular medium is to be measured in the laboratory with a permeameter, care must be exercised in interpreting the results so as to account for wall effect, A cursory analysis presented in this paper indicates that in general the ratio of permeameter diameter to mean particle diameter should exceed about 40 to eliminate any sizeable effect of the container wall. For ratios less than 40 the Reynolds number of the flow has great influence on wall effect.

Effects of fluctuations in surface water depth on infiltration rates into initially unsaturated soils were investigated by numerically solving the Richards equation. The numerical model was verified through comparison with published solutions of Philip and with the results of laboratory experiments on a well‐graded crushed dune sand. It was found that infiltration rates may incease with time in response to rapid rates of increase in water depth. Conditions under which this will occur are identified.