A remote sensing system for carbon monoxide and hydrocarbon vehicle emissions, devised by Prof. D. Stedman's group at the University of Denver, is being evaluated at the Ministry of Transportation of Ontario (MTO) in Toronto. The system was first checked to compare its readings with on-board emission measurements of vehicles as they passed a test point. Since then, it has been operated routinely over several months to measure the emissions of employee and MTO vehicles entering the site and records have been maintained for approximately 1000 vehicles that come in regularly. About 40 of these were tested on the MTO chassis dynamometer using the hot start and stabilized portion of the Environmental Protection Agency (EPA)/Transport Canada urban test cycle. The aim was to establish the confidence with which one or a limited number of pass-by measurements can be used to estimate the actual road emissions of vehicles and, in particular, reliably identify vehicles that would fail more stringent tests. Results indicate that an average reading of more than 3.5% carbon monoxide will identify vehicles that would fail the pre-1988 Canadian standard of 25 gm/mi, and also the prevailing Ontario idle emission standard. This group, which represents 10% of the vehicles in the commuting fleet studied, is also responsible for 50% of the carbon monoxide emitted. Hydrocarbon emissions were also examined but the results are more scattered and there is no clear threshold to confidently identify vehicles that would fail the standard.