A suite of 34 disinfection by-products (DBPs), including eight halomethanes, nine haloacetic acids, six haloacetonitriles, six haloaldehydes, four haloketones and the halonitromethane chloropicrin, were monitored in two microfiltration (MF) and reverse osmosis (RO) treatment plants as part of a larger study of chemical removal by MF/RO treatment for water recycling purposes. Both DBP detection frequency and concentration increased during treatment, and this was attributed to a chloramination step used to minimize RO membrane fouling. The degree of DBP formation was particularly related to plant residence time, with DBPs falling into two distinct groups; the first group in which DBP concentration increased with increasing residence time (e.g. chloroform and bromochloroacetaldehyde) and a second group in which increased residence time did not affect the concentration (e.g. dichloroacetic acid and 1,1-dichloropropanone). These results indicate that MF/RO plant design and wastewater quality are both important factors in minimizing DBP formation within MF/RO treatment. RO rejection was influenced by several chemical-specific properties, including pKa, log Kow and DBP class. Rejection of haloacetic acids, present as charged molecules, was consistently better than 90% and did not alter with logKow. For all other DBPs, present as neutral molecules, rejection was much more variable, and decreased with decreasing logKow, although the effect of MW and logKow on rejection could not be separated. The DBP formation described in this study lead to variable estimations of DBP removal by RO and thus it is recommended that DBPs are used as indicators of RO removal efficiency with caution, and only after DBP formation within RO treatment has been studied. (C) 2012 Elsevier B.V. All rights reserved.