Measurements of submicron particles by Fourier transform infrared spectroscopy in 14 campaigns in North America, Asia, South America, and Europe were used to identify characteristic organic functional group compositions of fuel combustion, terrestrial vegetation, and ocean bubble bursting sources, each of which often accounts for more than a third of organic mass (OM), and some of which is secondary organic aerosol (SOA) from gas-phase precursors. The majority of the OM consists of alkane, carboxylic acid, hydroxyl, and carbonyl groups. The organic functional groups formed from combustion and vegetation emissions are similar to the secondary products identified in chamber studies. The near absence of carbonyl groups in the observed SOA associated with combustion is consistent with alkane rather than aromatic precursors, and the absence of organonitrate groups can be explained by their hydrolysis in humid ambient conditions. The remote forest observations have ratios of carboxylic acid, organic hydroxyl, and nonacid carbonyl groups similar to those observed for isoprene and monoterpene chamber studies, but in biogenic aerosols transported downwind of urban areas the formation of esters replaces the acid and hydroxyl groups and leaves only nonacid carbonyl groups. The carbonyl groups in SOA associated with vegetation emissions provides striking evidence for the mechanism of esterification as the pathway for possible oligomerization reactions in the atmosphere. Forest fires include biogenic emissions that produce SOA with organic components similar to isoprene and monoterpene chamber studies, also resulting in nonacid carbonyl groups in SOA.atmospheric aerosol | organic particles | smog chamber aerosol | alkane oxidation products | photochemical reactions R ecent literature reviews have highlighted significant advances in identifying compounds formed as "secondary" organic aerosol (SOA) in a wide range of laboratory-simulated atmospheric conditions, and field studies have identified several individual products from chamber studies (1-3). Paulot et al. have used global modeling to show that extrapolating these laboratory results for modeled global oxidant distributions helps explain biogenic SOA (4). However, the observations needed to confirm the proposed sources of organic aerosols (OAs) in global modelsnamely, quantitative field measurements of the proposed products to compare with the model predictions-remain elusive (5). Without such confirmation, it is difficult to identify which of the mechanisms proposed to explain controlled laboratory studies of simple volatile organic compound (VOC)-oxidant systems would satisfactorily capture those aspects of the chemistry that determine SOA formation in the more complex atmosphere.There has been significant progress in quantifying organic mass (OM, the particle-phase components of OA) from the fragments produced by online mass spectrometry of particles, with recent promulgation of techniques to quantify the resulting mixtures by two to four fragment-based positive matri...