A simple and nearly direct procedure has been developed for the determination of As in beers, based on hydride generation and atomic fluorescence detection. Using 2 mol l 21 HCl with a carrier flow of 6 ml min 21 , 3% m/v NaBH 4 with a flow of 1.3 ml min 21 , a reaction coil of 200 cm and an argon flow of 400 ml min 21 , total arsenic in beer could be determined in samples diluted 1 : 10 with a final concentration of 2 mol l 21 HCl, 1% m/v KI and 0.08% v/v of an antifoam agent. For the estimation of As III and As V concentrations in beer samples, the difference between the analytical sensitivities of the fluorescence signals obtained for As hydride, without and with previous treatment of samples with KI, can be used, based on a simple proportional equations approach. The methodology developed provides a limit of detection of 39 ng l 21 As, a dynamic range between 0 and 1 ng ml 21 , a variation coefficient between 0.1 and 4.9% (1.01-12.89 ng ml 21 As) and recovery values of 95¡2% for As III and 96¡2% for As V . Data found for the direct determination of total As in market samples varied from 1.01 to 12.89 ng ml 21 , being in relatively good agreement with those obtained after a previous sample ashing.