The Accurate Measurements of Hydrocarbons in the Atmosphere (AMOHA) project was proposed to evaluate and improve current gas chromatographic methods used across Europe to determine concentrations of C2--C9 nonmethane hydrocarbons (NMHCs) in ambient air. Four international intercomparisons of increasing complexity were carried out over a period of 4 years. These were designed to evaluate the analytical performance of a considerable number of European laboratories making frequent ambient NMHC measurements, to identify the major problems encountered in these analyses, and to recommend solutions for these problems. Calibration gas standards prepared gravimetrically and certified by the National Physical Laboratory (NPL) in the UK were used in all of the intercomparisons to eliminate the variability that would have been introduced by laboratories using their own calibration gas mixtures. The results presented here are for the first three of the AMOHA intercomparisons. A four-component synthetic test gas mixture was used for the first intercomparison, a 30-component synthetic test gas mixture for the second intercomparison, and both a pressurized ambient air sample and a 30-component synthetic test gas mixture for the third intercomparison. The results of the 12--14 participating laboratories showed generally good agreement with the gravimetric concentrations of the synthetic NMHC mixtures, although some considerable differences also occurred. The analytical results on the pressurized air samples were similar, but with a greater scatter of values. Most importantly, there were distinct improvements in the measurement quality achieved in this project when compared to previous European intercomparisons. This emphasized the importance of using high-quality NMHC multicomponent gas calibration standards when making accurate measurements of NMHCs in air. Despite the generally good results, some compounds, such as isobutene, butadiene, methyl pentanes, and trimethyl benzenes, gave considerable analytical problems in some laboratories. The most frequent causes of these problems were found to be interference by coelution with other compounds, breakthrough of C2 compounds during sample preconcentration, and adsorptive losses of C7--C9 compounds in the transfer lines before the analytical instrumentation. Calibration with multicomponent NMHC standards was demonstrated to provide significantly more accurate results than the frequently used method of calibration with just a single hydrocarbon species, such as n-butane.