Adequate phosphorus (P) nutrition underpins cereal production in Australia and soil tests are commonly used to adjust fertiliser rates. The critical soil test values (i.e. required to achieve 90% of maximum yield) used for fertiliser recommendations have been derived from historic fertiliser trials sown mainly in May and June, with sowing date not considered in the interpretation of the critical values. However, the availability of long-season wheat cultivars has meant that crops can now be sown earlier. Experiments were conducted to investigate the effect of sowing time on optimum P rates for yield. Bread wheat (Triticum aestivum L.) cvv. Mace and Trojan were grown at five P rates at three sites in the Mid-North and Yorke Peninsula of South Australia, on alkaline calcareous soils, in each of 2 years. Both varieties showed similar responses to P. Crops sown in late April–early May often had a low optimum P rate, and there was a general trend for the optimum P rate to increase with later sowing. However, early sowing reduced the optimum P rate only when April rainfall was high. Grain yield responded positively to increases in vegetative growth from P fertiliser, and the optimum P rate for vegetative growth and grain yield were correlated. Early sowing increased grain P concentrations, and P removal rates were similar to, or greater than, those of later sown crops. Sowing in April–early May can reduce the P requirements of wheat, but there is an increased likelihood of a negative P balance; therefore, monitoring of soil P reserves will be required. A relationship between the optimum P rate for yield and the partial nutrient balance may provide a benchmark to help manage P.