Phosphorus applied in excess of crop needs results in soil P buildup and increases the risk of P loss to water bodies through surface or subsurface flow. A field study based on large plots (30 by 6 m) and natural precipitation was conducted in NW Iowa from 2007 until 2012. The soil was Primghar (Aquic Hapludolls) with 2.7 to 3.9% slope. Four sets of treatments were applied to corn-soybean (CS) rotations harvested for grain (each crop was planted each year in the experimental area) and one set was applied to continuous corn (CC) with grain and partial stover baling. For CS the treatments were the combinations of no-tillage (NT), chisel-plow/disk tillage (CHD), broadcast triple superphosphate fertilizer (TSP), and P based liquid swine manure (SMP) injected with both tillage systems. Phosphorus was applied once before corn to maintain initially Optimum soil-test P for the 2-yr CS rotation (112 kg P ha^-1 every 2 years). The CC was managed with CHD, N-based liquid swine manure, and no P fertilizer. The corn of CS received an equivalent N fertilizer rate. Treatments were replicated three times. A tipping bucket technique and electronic devices were used to monitor and collect surface runoff. Runoff was analyzed for dissolved reactive P (DRP, bioavailable P by the Fe-oxide paper method (BAP), total P (TP), and sediment. The last year, mean soil Bray-1 P (15-cm depth) was 24 mg kg^-1, 25 mg kg^-1, and 38 mg kg^-1 for CS with tillage, CS with NT, and CC. There were no consistent soil P and runoff P differences between the P sources for CS, but losses were much higher with tillage and in the corn years. In the corn years, soil and TP losses were 4.59 Mg soil ha^-1 yr^-1 and 4.36 kg P ha^-1 yr^-1 for CWD; and 1.27 Mg soil ha^-1 yr^-1 and 3.59 kg P ha^-1 yr^-1 for NT. In the soybean years, soil and TP losses were 0.78 Mg soil ha^-1 yr^-1 and 1.45 kg P ha^-1 yr^-1 for CHD and 0.30 Mg soil ha^-1 yr^-1 and 0.72 kg P ha^-1 yr^-1 for NT. This was explained by less residue cover after soybean and P applied before corn. Soil and TP losses for CC managed with N-based manure, tillage, and grain plus stover harvest were 1.9 Mg soil ha^-1 yr^-1 and 2.77 kg ha^-1 yr^-1. The DRP and BAP losses were much smaller than TP loss. The proportion of DRP and BAP of the TP for CS was higher with NT than with CHD (63% and 35% higher). No-till management significantly reduced soil and TP loss, but increased the proportion of runoff P most readily-available for algae growth.