Metaproterenol (M) and theophylline (T) were compared in an eight-week randomized, double-blind, crossover study in 20 moderately severe asthmatic children. All subjects were clinically maintained on both M and T. T was superior when evaluating cough, attacks, extra medicines, and PEFR values. Diary analysis revealed T better in 11, M in six and M/T equal in three subjects. Side effects were more common with T. M might be an adequate alternative to T in the treatment of asthmatic children requiring daily bronchodilators.