Amelioration of subsoil acidity using gypsum (CaSO4·2H2O) or lime (CaCO3) was studied on sandy textured soils with low water holding capacity in a low rainfall environment. Field trials were established in 1989 at two sites on yellow sandplain soils to investigate whether different rates, sources and combinations of gypsum and lime application could be used to increase wheat and lupin yields. Gypsum increased wheat yields by up to 45% in the first two growing seasons whereas lime increased wheat yields by up to only 15% in the second season. The highest yields were generally recorded when gypsum and lime were applied together. The response of wheat to the various treatments varied both regionally and temporally and it is suggested that the inherent soil solution composition affected the magnitude and rapidity of wheat responses to gypsum. The rate of gypsum application affected the longevity of the wheat responses, with a low application rate (1 t ha-1) increasing yields for only one season. No differences in wheat yields were recorded between different sources of gypsum or application rates higher than 3 t ha-1. In contrast to wheat, lupin yields were substantially lower on gypsum-treated plots. The yield decline did not appear to be related to any simple nutritional factor and the gypsum effect was generally minimized when lime was added with the gypsum. The results indicated that lower rates of gypsum than used in previous subsoil amelioration studies were suitable for increasing wheat yields on sandy soils in low rainfall environments, and that gypsum should not be used if lupins are to be grown within at least 2 years of its application to the soil surface.