The effectiveness of street sweeping as a source control measure for stormwater pollution was tested at a site in Toronto, using three types of sweepers employed by the City. A paired-plot experimental design was employed along an arterial road with a traffic volume of 26,000 vehicles/day. Typically, after several days of dry weather, one roadway plot was swept by the available sweeper (treated) and the following plot was left unswept (control). After sweeping, sediment on the roadway was sampled on both plots; wet samples were collected by washing off one half of each plot, and dry samples were collected by vacuum cleaning the remaining halves of both plots. Differences between swept and unswept plots were assessed by comparing: (a) conventional sediment quality parameters, total residue mass, and particle sizes for dry sediment samples, and (b) toxicity, conventional water quality parameters, and particle sizes in wet samples. Results were highly variable and contained large uncertainties. The greatest environmental benefits of sweeping were the reduction of the total mass of sediment on road surfaces and a reduction in some dissolved metals in the runoff (e.g., Cr and Zn).