The results of two long-term field experiments and two 1-year experiments are reported. In three of these, severe phosphate deficiency was present initially. At high phosphate supply, the annual total yield was not significantly related to age of pasture over periods of 10–13 years. At intermediate and low supply, yields relative to high phosphate supply increased significantly with time; these increases are believed to demonstrate residual effects of phosphate. Indirect evidence for nitrogen accretion from clover under severely phosphate-deficient conditions is presented. Sward components—in the long term—responded differentially to phosphate supply. With very low phosphate, erodium (Erodium botrys (Cav.) Bertol.) and flatweed (Hypochoeris glabra L.) were dominant; whereas with high phosphate, cape-weed (Cryptostemma calendula (L.) Druce) and ripgut brome grass (Bromus rigidus Roth)—or else barley grass (Hordeum leporinum Link)—were dominant. Subterranean clover (Trifolium subterraneum L.), though present under these extremes, was relatively more plentiful at intermediate levels of supply. However, at "steady state" conditions, the range in clover content was fairly narrow (from c. 20 to 40%). The significance of these findings to a sheep infertility problem ("clover disease") of subterranean clover-dominant pastures is discussed.