Soil water balance and plant data from a time of sowing trial, and estimates of transpiration efficiency, were used to assess the importance of soil cover, provided by developing canopies, on the water use and its partitioning between soil evaporation and transpiration. As reported in an earlier paper, time of sowing strongly affected the timing and rate of canopy development, and had little effect within years on total water use, which was 221 mm in 1982 and 185 mm in 1983. Time of sowing also modified the pattern of water use, and this was reflected in substantial effects on the partitioning of water use, between soil evaporation and transpiration, and in turn on yield and water use efficiency. Estimated water use by soil evaporation ranged from 100 to 125 mm in 1982 and from 75 to 115 mm in 1983. The majority of this, an average of 80 mm, in 1982 and 75 mm, in 1983, occurred during the winter months, June to August, and varied mildly with soil cover. Soil cover had its greatest effect on water use through transpiration during spring, when temperatures were rising rapidly. Transpiration varied between sowing dates by 20 mm in 1982, and by 40 mm in 1983. Measures to improve water use efficiency should aim to reduce soil evaporation during winter both directly, by increasing soil cover, for example, by mulches or earlier-developing canopies, and indirectly by increasing infiltration. In spring, measures to improve water use efficiency should aim at reducing transpiration by minimising canopy development to what is required by the crop to maximise harvest index.