The effects of the density of lupin ( Lupinus angustifolius L. cv. Illyarrie) taproots on a following wheat crop ( Tritium aestivum L. cv. Gutha) were investigated on a compacted, earthy sand soil near Geraldton, Western Australia. In 1985, plots were sown to lupins at densities ranging from 35 to 220 plants m-2. Because peak lupin biomass varied by less than 25%, and 100 kg N ha-1 of mineral fertiliser was supplied to the wheat, the effects of variation in residual N from the lupins were considered to be minimal. In 1986, all plots were split for shallow or deep tillage and wheat sown. There were no effects of tillage or lupin density on stored soil water at sowing. Growth and yield of wheat on the shallow tilled plots increased linearly with lupin plant density in the previous year, to equal or better wheat yields on the deep tilled plots, which were unaffected by lupin density. At normal farm lupin plant densities, the biological plough effect was estimated to improve wheat yields by 100 kg ha-1, substantially less than the benefits from nitrogen fixation and breaks in disease cycles. Nevertheless, it still suggests that lupin stand densities should be maintained or increased in crop rotations on compacted sands.