To help assess respiratory health risks from sulfur dioxide (SO2) air pollution, we studied 24 normal, 21 atopic, 16 minimal/mild asthmatic, and 24 moderate/severe, medication-dependent asthmatic subjects classified according to history, lung function, allergy skin tests, serum IgE level, and airway reactivity to methacholine. All were exposed in a chamber (21 degrees C, 50% humidity) to 0.0, 0.2, 0.4, and 0.6 ppm SO2 in random order at 1-wk intervals; then exposures were repeated to test consistency of response. The 1-h exposures included three 10-min exercise periods (ventilation approximately 40 L/min). Physiologic response was measured early (approximately 15 min) and late (approximately 55 min) in exposure. Symptoms were evaluated during exposure and for 1 wk afterward. Normal and most atopic subjects showed little response at these SO2 levels. A few atopic subjects and many asthmatics developed bronchoconstriction and respiratory symptoms, but most were able to maintain their exercise. Effects were not markedly different between early and late measurements, nor between the first and second round of studies; however, late and second-round responses appeared slightly more favorable. No statistically significant effect of SO2 on symptoms was found 1 day or 1 wk after exposure. Minimal/mild asthmatics showed, on the average, slight responses at 0.0 ppm (attributable to exercise) and increasing responses at increasing SO2 concentrations. Moderate/severe asthmatics reacted more at 0.0 ppm, but their increments in response with increasing SO2 concentration were roughly similar to those of minimal/mild asthmatics. Thus, responses to SO2 per se were not strongly dependent on clinical severity of asthma, nor on SO2 exposure history during previous weeks.