Florida Spodosols are sandy, inherently low in Fe- and Al-based minerals, and sorb phosphorus (P) poorly. We evaluated runoff and leachate P losses from a typical Florida Spodosol amended with biosolids and triple superphosphate (TSP). Phosphorus losses were evaluated with traditional indoor rainfall simulations but used a double-deck box arrangement that allowed leaching and runoff to be determined simultaneously. Biosolids (Lakeland, OCUD, Milorganite, and Disney) represented contrasting values of total P, percent water-extractable P (PWEP), and percentage of solids. All P sources were surface applied at 224 kg P ha(-1), representing a soil P rate typical of N-based biosolids application. All biosolids-P sources lost less P than TSP, and leachate-P losses generally dominated. For Lakeland-amended soil, bioavailable P (BAP) was mainly lost by runoff (81% of total BAP losses). This behavior was due to surface sealing and drying after application of the slurry (31 g kg(-1) solids) material. For all other P sources, BAP losses in leachate were much greater than in runoff, representing 94% of total BAP losses for TSP, 80% for Milorganite, 72% for Disney, and 69% for OCUD treatments. Phosphorus leaching can be extreme and represents a great concern in many coarse-textured Florida Spodosols and other coastal plain soils with low P-sorption capacities. The PWEP values of P sources were significantly correlated with total P and BAP losses in runoff and leachate. The PWEP of a source can serve as a good indicator of potential P loss when amended to sandy soils with low P-retention capacities.