The P Index for Virginia uses coefficients describing the risk of P runoff from pastures for different nutrient sources, but these coefficients have not been verified on Virginian soils. An experiment was conducted to evaluate the relative risk of P losses from pasture after P application of manures or commercial fertilizer. Four sources of manure (dairy slurry, swine slurry, beef solids, and poultry litter) and triple superphosphate were spring applied to pasture plots at an estimated crop P removal of 27.4 kg ha(-1) and compared with a no-P control treatment. Three pairs of rainfall simulations (65-70 mm h(-1)) were conducted (June, August, and October). Each month, the second rainfall simulation, conducted 24 h after the first, was to evaluate effects on saturated soils. The control plots had lower total P (TP; 0.125 mg L-1) and lower dissolved reactive P (DRP; 0.067 mg L-1) concentrations than all other treatments (0.189-0.241 mg TP L-1 and 0.127-0.174 mg DRP L-1) in runoff during the first rain simulation 40 days after amendment. The control generated the lowest (0.117 mg L-1) and triple superphosphate the greatest (0.241 mg L-1) TP concentrations during the second simulation 24 h later (saturated ground). Slurry manures led to greater TP runoff concentrations than solid manures during the second simulation in both June and August. Furthermore, TP and DRP concentrations dropped throughout the 30-min runoff period for all P treatments. Cropland-derived coefficients in the Virginia P Index are generally appropriate for pasture in soils typical of the Shenandoah Valley, but reduction in their magnitude and a differentiation between liquid and solid manures may be warranted. In more complex process-based models, inclusion of season and rainfall duration may be appropriate, to reflect the greater runoff risk during longer rain events and from wet fields early after manure application.