Fecal coliforms are often used as indicators of fecal contamination of surface water from surface-applied manure. Manure constituents other than fecal coliforms may serve as natural tracers of fecal contamination provided that rates of their release from manure to runoff are similar to fecal coliform release rates. Objectives of the present work were (a) to compare kinetics of chloride (Cl), fecal coliforms (FC), organic carbon (OC), and water-soluble phosphorus (P) release from dissolving manure applied on a hillslopes with different soil texture and surface cover under simulated rainfall; (b) to assess applicability of three most recent models to describe the kinetics of Cl, FC, OC and P release. Manure was applied at 50 cm x 30 cm miniplots with 20% slope. Both bare and vegetated miniplots were set at sandy loam and clay loam soil outside the area of homogeneous rainfall coverage to provide variations in rainfall intensity. Rainfall was simulated for one hour, and runoff was collected from gutters at the edge of miniplots with five-min interval. Concentrations Cl, FC, OC and P were measured in runoff. Exponential one-parametric model, a power-law two-parametric Vadas-Kleinman-Sharpley model, and a power-rational two-parametric Bradford-Shijven model were fitted to the data. Kinetics of FC release from manure was similar to the release kinetics of other manure constituents. Fecal coliforms tended to be released with the same rate or slower than P and OC at vegetated plots. The two-parametric models had better accuracy in fitting the release kinetics of compared with one-parametric exponential release model. Two parameters of the Vadas-Kleinman-Sharpley were strongly correlated. The Bradford-Shijven model had uncorrelated parameters values of which reflected differences in irrigation rate and vegetation. Two-parametric models appear to be preferable to describe release of manure constituents when manure is deposited in vegetated areas.