Polyelectrolytes when used as coagulants or coagulant aids in water treatment can react with oxidants used for disinfection, the by-products formed being analogous to those obtained from the reaction of chlorine with natural organic matter. A review of currently available literature shows that, with the concentrations of the commonly used cationic polymers and chlorine normally met in water treatment, the formation of trihalomethanes is minimal, and the polymer is not the principal precursor, When unsatisfactorily high levels of by-products are formed on polymer usage, other chemicals present in the commercial product have usually been identified as the precursors. The most potent of these is residual acrylamide monomer. AS long as strict checks are maintained on the amount of this monomer present, trihalomethanes are formed in insignificant amounts. However, the other by-products that can be produced may warrant further exploration, such as the total halogenated organic species obtained from polyamines (192 mu g l(-1)). For poly(diallyldimethylammonium chloride) the formation is much lower (12 mu g l(-1)). Nitrosodimethylamine has been reported as another serious contaminant, formed to a limited extent from normal operations with cationic polymers, but at levels of concern when other nitrogen compounds are present. More data are needed on cationic polyacrylamides, which appear to have been somewhat neglected, especially the possible contribution from the cationic acrylate monomer when it is present. Anionic and non-ionic polyacrylamides do not form significant levels of trihalomethanes on reaction with chlorine either. These polymers also have contributions equivalent to the amount produced by the residual monomer present. The choice of the most appropriate treatment polymer should hence always be on the basis of the absence of low molecular weight material as well as on process performance. Strict regulatory control on monomer and impurity content hence has to be maintained.