Dichloroacetonitrile (DCAN) is an important example of a reactive disinfection by-product for which a large body of occurrence data exists. Although it is known to undergo base-catalysed hydrolysis, DCAN9s peculiar dependence on reaction time, chlorine dose and pH has never been fully reconciled with expectations based on its presumed precursor (i.e. amino acid residues). The purpose of this research was to improve existing models for DCAN degradation and to use this information for interpretation of DCAN concentration profiles. Laboratory studies were performed using buffered solutions of DCAN, natural organic matter (NOM) and treated drinking waters, both with and without free residual chlorine. DCAN concentrations were measured as a function of reaction time. Results indicate a decomposition scheme encompassing three pathways of hydrolysis: attack by hydroxide, hypochlorite and water. Any one of the three pathways may predominate in drinking water systems, depending on the pH and chlorine residual. The resulting chemical kinetic model was used to show that the DCAN formed (and subsequently decomposed) was often many times the actual measured DCAN concentration. DCAN formation was found to agree with expectations based on the underlying chemistry of chlorine attack on proteinaceous material.