Reports of low-level subtherapeutic concentrations of pharmaceuticals in surface and ground waters have raised questions as to whether such compounds would survive typical drinking water treatment (i.e., coagulation, settling, filtration, and disinfection) and ultimately result in inadvertent human exposure. To examine the impact of these processes, a bench-scale drinking water treatment plant (DWTP) was constructed and used to examine the potential removal of caffeine, trovafloxacin mesylate, estradiol and salicylic acid relative to a conservative bromide (Br) tracer at pH 8. Radiolabeled compounds were used to provide good sensitivity within the small sample volume constraints of the system and to enable a more rigorous mass balance analysis. The bench-scale system was comprised of several unit operations including: coagulation, flocculation, sedimentation, dual-media gravity filtration, granular activated carbon treatment (GAC), and chlorination. Without GAC treatment, steady state analyses indicate removals of 3.4 to 13%, 21 to 31%, 6.9 to 12%, and 31 to 39% (95% confidence interval) for caffeine, trovafloxacin mesylate, estradiol, and salicylic acid, respectively, with most removal associated with flocculation/sedimentation and filtration. Biological degradation was likely the main process contributing to the removal of salicylic acid. The addition of GAC treatment was found to significantly enhance the overall removal of caffeine (>94%), trovafloxacin mesylate (>95%), and estradiol (93–95%), but had limited effectiveness for salicylic acid (39–56%).