This study examines short-term relationships between doctors' house calls and urban air pollution in Greater Paris for the period 1991-1995. Poisson regressions using nonparametric smoothing functions controlled for time trend, seasonal patterns, pollen counts, influenza epidemics, and weather. The relationship between asthma visits and air pollution was stronger for children. A relative risk (RRP95/P5) of 1.32 [95% confidence interval (CI) = 1.17-1.47)] was observed for an increase from the 5th to the 95th percentile (7-51 micrograms/m3) in daily concentrations of black smoke (BS). The risks for 24-hr sulfur dioxide and nitrogen dioxide levels were in the same range. Cardiovascular conditions, considered globally, showed weaker associations than angina pectoris/myocardial infarction, for which RRP95/P5 was 1.63 (95% CI = 1.10-2.41) in relation to ozone ambient levels. Eye conditions were exclusively related to ozone (RRP95/P5 = 1.17, 95% CI 1.02-1.33). Asthma visits and ozone showed an interaction with minimum temperature: an effect was observed only at 10 degrees C or higher. In two-pollutant models including BS with, successively, SO2, NO2, and O3, only BS and O3 effects remained stable. Along with mortality and hospital admissions, house call activity data, available on a regular basis, may be a sensitive indicator for monitoring health effects related to air pollution.