The hydroxyl radical (OH) plays a key role in the oxidation of trace gases in the troposphere. However, observations of OH and the closely related hydroperoxy radical (HO<sub>2</sub>) have been sparse, especially in the tropics. Based on a low-pressure laser-induced fluorescence technique (FAGE – Fluorescence Assay by Gas Expansion), an instrument has been developed to measure OH and HO<sub>2</sub> aboard the Facility for Airborne Atmospheric Measurement (FAAM) BAe-146 research aircraft. During the African Monsoon Multidisciplinary Analyses (AMMA) campaign, observations of OH and HO<sub>2</sub> (HO<sub>x</sub>) were made in the boundary layer and free troposphere over West Africa on 13 flights during July and August 2006. Mixing ratios of both OH and HO<sub>2</sub> were found to be highly variable, but followed a diurnal cycle: OH varied from 1.3 pptv to below the instrumental limit of detection, with a median mixing ratio of 0.17 pptv. HO<sub>2</sub> varied from 42.7 pptv to below the limit of detection, with a median mixing ratio of 8.0 pptv. A median HO<sub>2</sub>/OH ratio of 95 was observed. Daytime OH observations were compared with the primary production rate of OH from ozone photolysis in the presence of water vapour. Daytime HO<sub>2</sub> observations were generally reproduced by a simple steady-state HO<sub>x</sub> calculation, where HO<sub>x</sub> was assumed to be formed from the primary production of OH and lost through HO<sub>2</sub> self-reaction. Deviations between the observations and this simple model were found to be grouped into a number of specific cases: (a) within cloud, (b) in the presence of high levels of isoprene in the boundary layer and (c) within a biomass burning plume. HO<sub>2</sub> was sampled in and around cloud, with significant short-lived reductions of HO<sub>2</sub> observed. Up to 9 pptv of HO<sub>2</sub> was observed at night, with HO<sub>2</sub> above 6 pptv observed at altitudes above 6 km. In the forested boundary layer, HO<sub>2</sub> was underestimated by a steady state calculation at altitudes below 500 m but overestimated between 500 m and 2 km. In a biomass burning plume, observed HO<sub>2</sub> concentrations were significantly below those calculated.