The shrinkage of natural soil aggregates that accompanies water withdrawal has been examined. Four main stages operate: (1) structural shrinkage, (2) normal shrinkage, (3) residual shrinkage, (4) no shrinkage: The pF ranges over which these stages operate and the fraction that they constitute of the total shrinkage has been measured for a range of soils. The influence of structure and texture upon shrinkage was assessed. These affect both the shrinkage phase and the range over which it operates. Structural development affects shrinkage markedly, giving a very different shrinkage pattern from that found for puddled soil or remoulded soil blocks. The difference in porosity between a natural aggregate and a puddled block of the same soil corresponding to a definite pF value is suggested, therefore, as a means of assessing the degree of structural development in the soil. It provides a suitable index for assessing the rate of structure improvement under various ameliorative treatments. The influence of cracking in increasing the rate of water entry into fine textured soils is discussed. In the particular cases of irrigated soils, where soil water contents are maintained at relatively high levels (above wilting point), it seems that cracking may not have progressed sufficiently to compensate for the low inherent permeability of such soils. Cultivation of the surface soil to form a mulch before irrigation tends to eliminate any contribution to conduction by the smaller cracks which are present at higher soil water contents. Large increases in permeability, due to the presence of gross soil cracks, are attained only when water contents are reduced to levels which are not practicable in continuous irrigation culture.

Soil water depletion rates in a 115‐cm and a 170‐cm deep profile of Adelanto clay loam were compared with lysimetrically obtained consumptive use rates for periods of many days after measured water applications. When the soil was bare, the depletion rates were always higher than the rate of loss to the atmosphere, and the inferred flux at the 170‐cm depth was as high as 2 mm/day 8 days after irrigation. When the test plot was planted to sorghum (Sorghum vulgare Pers.) an initially strong downward flux at the 170‐cm depth reversed itself after about 10 days and became as high as 4 mm/day, representing upward flow of water from wet soil into the root zone above. The data imply that indiscriminate use of soil water depletion rates as representing consumptive use rates can be highly misleading at any time in an irrigation cycle. Further analysis shows that a rational and satisfactory correction of depletion data is not likely feasible, and, at any rate, unworkable for the condition of the experiment.

A convenient method is described for determining the apparent density of soil aggregates down to 2 mm in diameter. The aggregates are evacuated, wetted with a non-polar liquid (kerosene), drained at a standard tension, and their volume determined by displacement in this liquid. Results obtained have shown that apparent density is measurable to within 0.03 g/c.c. or about 2 per cent. Features of this method and their significance are discussed.

The distribution of irrigation water in an Adelanto clay loam profile was studied in a field plot by simultaneous, periodic observations of the water content and hydraulic head profiles. Successive measurement series were made with the plot bare and covered, bare, and planted to a sorghum crop (Sorghum vulgare Pers.). From the first two, the in situ and dynamic relations of water content to water pressure and to conductivity were obtained. From the cropped field data, the root‐extraction pattern was derived, using the established hydraulic properties of the profile. The data demonstrate the variability within depths and locations of water retention and conduction properties and the consequent problem of calculating fluxes. The mobile character of soil water is also evident, confirming the inadequacy of static concepts of soil water “constants” for a profile. Calculated root‐extraction rates agreed reasonably with independent lysimetric measurements of the water loss from the surface to the atmosphere.

The redistribution of rainfall on silverleaf ironbark (Eucalyptus melanophloia) trees in a semiarid environment at Narayen Research Station has been studied using detailed measurements on four trees.Interception accounts for about 11% of the annual rainfall with only 0.6% as stemflow; canopy storage capacity is estimated to be 2 mm and an estimate of the rain required to initiate stemflow is 7 mm.With the exception of nitrogen, the amounts of major elements returned to the soil surface beneath the tree canopy via throughfall and stemflo w is of the same order as that added by litterfall.