On some of the lighter textured soils in the wheatbelt of central-western New South Wales near Dubbo, soil acidity is a major problem, and lucerne (Medicago sativa) often establishes and grows poorly. We selected a site with a surface soil pHCa of 4.4 and an exchangeable aluminium of 0.4 cmol(+)/kg, which was also acidic down the soil profile. Experimental plots of 4 application rates of lime (nil, 1, 2 and 3 t/ha) in 4 replications were established. The site was limed in 1990 and lucerne sown in May 1991. Over the next 6 years the trial was periodically grazed with sheep, and lucerne regrowth and stand density were monitored. In October 1997, the lucerne was removed and 3 crops of varying acid tolerance (wheat, barley and canola) were sown as split plots in both 1998 and 1999. Lucerne density was higher in the limed plots compared with the unlimed treatment, and this difference persisted for 6 years. Dry matter production of lucerne was increased by lime applied at rates up to 2 t/ha. All 3 crops sown after the lucerne phase responded to lime applied 8 and/or 9 years earlier. The responses were attributed to the strong residual effect of the lime in the 0'10 cm soil layer, to smaller improvements in the 10'20 cm zone (possibly due to the movement of lime down the soil profile over the 7 years before the date of measurement) and to carry over effects of nitrogen fixation by the lucerne into the cropping phase. The protein content of the wheat grain was increased concurrently with grain yield due to the previous liming and resultant legume nitrogen effects. The results support the application of lime to improve the productivity of lucerne and subsequent crops, even when the soil is acidic to depths below the cultivation layer.