Dissolved organic matter (BOM) from three low-hardness surface waters was isolated and concentrated using a reverse osmosis (RO) membrane system. The efficacy of the RO isolation method and its impact on the subsequent reactivity between DOM and chlorine were examined. DOM mass balances (quantified as dissolved organic carbon) ranged from 96.1 to 102.1 % for the three waters tested, and DOM mass recoveries of 93.9 to 98.2% indicated successful isolation, minimal fractionation, and negligible loss of organic matter. RO isolates were diluted using distilled and deionized water in the laboratory to reconstitute the source waters. Both source water (collected at the time of isolation) and reconstituted source water samples were chlorinated. Formation of several disinfection by-products (DBPs: e.g., THMs. HAA(9), HANs, and HKs) were measured. For all waters tested, DBP formation of source and corresponding reconstituted source water agreed within 95% confidence intervals. Therefore, RO isolation had no impact on the DOM reactivity of the three low-hardness surface water tested in this study. In addition, the degree of bromine substitution. as expressed by the bromine incorporation factor, was calculated. Comparison of bromine incorporation factors for sourer and reconstituted source waters further indicated that, as with the total DBP formations, bromine speciation and the relative occurrence of individual species in THMs and HAA(9) did not change as a result of the isolation. Overall, in terms of DBP formation, RO isolation appears to maintain the integrity and reactivity of DOM. (C) 2001 Elsevier Science Ltd. All rights reserved.