A primary driver of the wheat yield gap in Australia and globally is the supply of nitrogen (N) and options to increase N use efficiency (NUE) are fundamental to closure of the yield gap. Co-application of N with phosphorus (P) is suggested as an avenue to increase fertiliser NUE, and inputs of N and P fertiliser are key variable costs in low rainfall cereal crops. Within field variability in the response to nutrients due to soil and season offers a further opportunity to refine inputs for increased efficiency. The response of wheat to N fertiliser input (0, 10, 20, 40 and 80 kg N ha(-1)) under four levels of P fertiliser (0, 5, 10 and 20 kg P ha(-1)) was measured on three key low rainfall cropping soils (dune, mid-slope and swale) across a dune-swale system in a low rainfall semi-arid environment in South Australia, for three successive cropping seasons. Wheat on sandy soils produced significant and linear yield and protein responses across all three seasons, while wheat on a clay loam only produced a yield response in a high rainfall season. Responses to P fertiliser were measured on the sandy soils but more variable in nature and a consistent effect of increased P nutrition leading to increased NUE was not measured.