Adsorption/thermal desorption with multisorbent air-sampling cartridges was developed for the determination of 87 method analytes including halogenated alkanes, halogenated alkenes, ethers, alcohols, nitriles, esters, ketones, aromatics, a disulfide, and a furan. The volatilities of the compounds ranged from that of dichlorofluoromethane (CFC12) to that of 1,2,3- trichlorobenzene. The eight most volatile compounds were determined using a 1.5-L air sample and a sample cartridge containing 50 mg of Carbotrap B and 280 mg of Carboxen 1000; the remaining 79 compounds were determined using a 5-L air sample and a cartridge containing 180 mg of Carbotrap B and 70 mg of Carboxen 1000. Analysis and detection were by gas chromatography/mass spectrometry. The minimum detectable level (MDL) concentration values ranged from 0.01 parts per billion by volume (ppbv) for chlorobenzene to 0.4 ppbv for bromomethane; most of the MDL values were in the range 0.02-0.06 ppbv. No breakthrough was detected with the prescribed sample volumes. Analyte stability on the cartridges was very good. Excellent recoveries were obtained with independent check standards. Travel spike recoveries ranged from 90 to 110% for 72 of the 87 compounds. The recoveries were less than 70% for bromomethane and chloroethene and for a few compounds such as methyl acetate that are subject to losses by hydrolysis; the lowest travel spike recovery was obtained for bromomethane (62%). Blank values for all compounds were either below detection or very low. Ambient atmospheric sampling was conducted in New Jersey from April to December, 1997. Three sites characterized by low, moderate, and high densities of urbanization/traffic were sampled. The median detected concentrations of the compounds were either similar at all three sites (as with the chlorofluorocarbon compounds) or increased with the density of urbanization/traffic (as with dichloromethane, MTBE, benzene, and toluene). For toluene, the median detected concentrations were 0.23, 0.42, and 0.70 ppbv at the three sites. Analytical precision was measured using duplicate sampling. As expected, the precision deteriorated with decreasing concentration. At concentrations greater than 0.2 ppbv, most duplicates differed by less than 20%; below the MDL values, the differences between the duplicates were larger, but they were still typically less than 40%.