The effects of residual phosphate (P) fertilizer were monitored for six seasons on Mediterranean grassland. The phosphate fertilizer was originally applied annually at three rates (0, 25 and 60 kg P2O5/ha) for 7 years (1984–1990) to phosphate-deficient grassland at Tel Hadya, northern Syria. The herbage biomass productivity, botanical composition and the seed bank in the soil were monitored for six seasons (1991/92–1996/97). The experiment was grazed at two annual stocking rates (1·1 sheep/ha (low) and 2·3 sheep/ha (high)). The experimental site was typical of native grassland within the cereal zone of west Asia, where cropping is not possible because of shallow, stony soils and steep slopes.

Available soil phosphorus in May 1991 was 6·5, 20·8 and 40·1 mg P/kg under the 0, 25 and 60 kg P2O5/ha treatments and 6·6, 13·4 and 14·8 mg P/kg respectively, in May 1997. Yields of both legumes and total herbage (legume+grass+other species) were significantly influenced by the residual phosphate. Legume yields were between 6- and 7-times the control yield in the first two years of the study. This decreased with time but was still in the range of 1·5 to 1·9 times the yield of the control in 1997, six years later. Total herbage yield was consistently higher on the plots previously fertilized with P, which ranged between 1·5- and 2·5-times the control. Both legume seed and grass seeds were significantly larger with residual P, which ranged between 5·4- and 2·0-times the control for the legume and 2·5- and 1·4-times for the grasses. All these factors have practical implications for the use of P fertilizer on grassland which could help control soil erosion and improve livestock production on marginal lands on which farming communities largely depend.