As part of the Auto/Oil Air Quality Improvement Research Program (AQIRP), a fleet of 299 1983-1993 “real world” light duty vehicles and trucks were acquired from inspection and maintenance (I/M) lanes and tested at prevailing ambient temperatures for their hot soak emissions for the first hour after the engine was turned off. When found, high-emitters were repaired and retested to quantify the effectiveness of the repairs. Also, I/M pressure-purge tests were performed to determine whether such tests could properly identify high-emitting vehicles.Measured hot soak emissions ranged from less than 0.1g HC to as high as 49g HC. Twenty percent of the vehicles tested accounted for nearly 80 percent of the total hot soak emissions, with no single common hardware component identified as the primary cause.Of the 299 vehicles tested, 46 produced as-received emissions in excess of 2g HC; roughly 20 percent of these high-emitting vehicles were found to have either been tampered with or malmaintained. Repairs on 41 of the high-emitting vehicles reduced their emissions by 83 percent.Only 35 of the test vehicles failed the inspection/maintenance (I/M) pressure-purge test and the majority of this group had hot soaks in excess of 2 grams. The remaining 264 vehicles passing the pressure-purge test were not all low-emitters; they produced over half the excess hot soak emissions identified in the pilot study with 16 observed at 2g or greater.Information of the type generated by this study can be used in improving the predictive capability of existing VOC emission inventory models and may aid in the formulation of strategies to reduce hot soak emissions from the existing vehicle fleet.