Phosphorus in runoff from fields where poultry litter is surface-applied is an environmental concern. We investigated the effect of adding phytase and reducing supplemental P in poultry diets and composting poultry manures, with and without Fe and Al amendments, on P in manures, composts, and runoff. We used four diets: normal (no phytase) with 0.4% supplemental P, normal + phytase, phytase + 0.3% P, and phytase + 0.2% P. Adding phytase and decreasing supplemental P in diets reduced total P but increased water-extractable P in manure. Compared with manures, composting reduced both total P, due to dilution of manure with woodchips and straw, and water-extractable P, but beyond a dilution effect so that the ratio of water-extractable P to total P was less in compost than manure. Adding Fe and Al during composting did not consistently change total P or water-extractable P. Manures and composts were surface-applied to soil boxes at a rate of 50 kg total P ha(-1) and subjected to simulated rainfall, with runoff collected for 30 min. For manures, phytase and decreased P in diets had no significant effect on total P or molybdate-reactive P loads (kg ha(-1)) in runoff. Composting reduced total P and molybdate-reactive P loads in runoff, and adding Fe and Al to compost reduced total P but not molybdate-reactive P loads in runoff. Molybdate-reactive P in runoff (mg box(-1)) was well correlated to water-extractable P applied to boxes (mg box(-1)) in manures and composts. Therefore, the final environmental impact of dietary phytase will depend on the management of poultry diets, manure, and farm-scale P balances.