The efficacy of metaproterenol (orciprenaline) and theophylline given orally at currently recommended doses was examined in 34 children with chronic asthma using a randomized double-blind cross-over evaluation of four weeks' duration for each active regimen. No serious adverse effects were seen with either medication, but tremor occurred more frequently with metaproterenol (P less than 0.01). No significant differences were observed in the frequency of nausea, vomiting, headache, or insomnia (P greater than 0.05). Symptoms of wheezing, coughing, exercise intolerance, and interference with sleep were more frequently associated with the oral metaproterenol regimen; completely asymptomatic days occurred 50% more frequently in association with theophylline therapy (P less than 0.01). Mean peak flows, performed twice daily during each of the four-week study periods, were 86 and 92% of predicted for metaproterenol and theophylline, respectively (P less than 0.05). Pulmonary function decreased significantly less with theophylline than with metaproterenol among those who completed six minutes of treadmill exercise during both regimens (P less than 0.05). Corticosteroids, used for acute symptoms that failed to respond to the addition of inhaled metaproterenol, were required in four patients during both regimens, in ten patients only during the metaproterenol regimen, and in one patient only during the theophylline regimen (P less than 0.02). Thus, theophylline therapy was associated with fewer adverse effects, fewer symptoms of asthma, better pulmonary function, better exercise tolerance, and less requirements for corticosteroids than was treatment with metaproterenol.