Abstract. Current photochemical models developed to simulate the atmospheric degradation of aromatic hydrocarbons tend to underestimate OH radical concentrations. In order to analyse OH budgets, we performed experiments with benzene, toluene, p-xylene and 1,3,5-trimethylbenzene in the atmosphere simulation chamber SAPHIR. Experiments were conducted under low-NO conditions (typically 0.1-0.2 ppb) and high-NO conditions (typically 7-8 ppb), and starting concentrations of 6-250 ppb of aromatics, dependent on OH rate constants. For the OH budget analysis a steady-state approach was applied in which OH production and destruction rates (P OH and D OH ) have to be equal. The P OH were determined from measurements of HO 2 , NO, HONO, and O 3 concentrations, considering OH formation by photolysis and recycling from HO 2 . The D OH were calculated from measurements of the OH concentrations and total OH reactivities. The OH budgets were determined from D OH /P OH ratios. The accuracy and reproducibility of the approach were assessed in several experiments using CO as a reference compound where an average ratio D OH /P OH = 1.13 ± 0.19 was obtained. In experiments with aromatics, these ratios ranged within 1.1-1.6 under low-NO conditions and 0.9-1.2 under high-NO conditions. The results indicate that OH budgets during photo-oxidation experiments with aromatics are balanced within experimental accuracies. Inclusion of a further, recently proposed OH production via HO 2 + RO 2 reactions led to improvements under low-NO conditions but the differences were small and insignificant within the experimental errors.