Haloacetic acids (HAAs) and trihalomethanes (THMs) were generated in bench-scale chlorination experiments using treated waters (prior to final chlorination) of the three major drinking water utilities of the Quebec City area. The purpose was to investigate the formation and occurrence of these chlorination by-products (CBPs) on a seasonal basis. Data for HAAs, THMs and other physico-chemical parameters were produced through a six-month sampling program with variable conditions of water quality, water temperature, applied chlorine dose and reaction time. In waters from the three utilities, chloroform (THM specie), dichloroacetic and trichloroacetic acid (HAA species) were the most prevalent compounds due to the low concentrations of bromide in the utilities' raw waters. Significant differences in CBP occurrence were noted between the three utilities' chlorinated waters, mainly due to the type of disinfectant applied to raw water. The use of pre-ozonation, as opposed to pre-chlorination (or direct chlorination) in one of the utilities appears to be the major factor contributing to that utility's potential for compliance with current THM and future HAA standards. Seasonal variations in THMs and HAAs were mainly associated with variations in organic precursors and to changes in water temperature (two parameters which vary widely on a seasonal basis in surface waters of southern Quebec), with CBP occurrence at its highest in spring. Statistical correlations between HAAs and THMs were moderate and only temperature appeared to affect the preponderance of one CBP or the other. Finally, a regression analysis was carried out aimed at associating each CBP to water quality and the experimental parameters. Thanks to their predictive ability, multivariate models seem to be the tools with the best potential for decision-making purposes.