The kinetics of the reaction of dissolved S(IV) with H2O2 was studied in freshly collected precipitation samples in order to examine whether there exist any important influences on this reaction from trace constituents present in natural atmospheric liquid water, that is, rainwater and snowmelt. The kinetic study was carried out on over 300 freshly collected precipitation samples obtained at Brookhaven National Laboratory, Long Island, New York, during the period from October 1983 to September 1985. The second-order rate constant k(2), defined as -d2O2>/dt=k(2)2O2>, which was determined using the precipitation samples covering pH range 3.3--5.6, was found to be linearly dependent on H+ concentration, conforming to results previously obtained with purified water. The observed scatter in rate constant (¿30%, standard deviation) is comparable to interlaboratory scatter observed for purified water; the average rate constant is 15% less than values obtained in our laboratory using purified water at the same pH and ionic strength. The scatter and the systematic bias of the data were examined in terms of experimental techniques employed as well as chemical composition of the samples. Sample scatter is attributed in large part to uncertainty in measured pH. The difference in the average rate constant is small compared to the large scatter exhibited by the data. Causes of the negative bias were not clearly identified, but pH uncertainty, ionic strength, and the presence of formaldehyde were ruled out. The small negative bias observed on the kinetics of this reaction, however, is not expected to exert an appreciable effect on the extent of atmospheric SO2 oxidation, especially under conditions where the reagent concentrations are high, and the reaction tends to be complete within cloud lifetimes. The utilization of rate constants for this reaction, determined using purified water in numerical calculations and modeling, is therefore justified.