Manitoba has a universally accessible health-care system that records physician contacts and hospitalizations in such a way that they can be ascribed to individuals. We examined the prevalence of physician-diagnosed asthma, bronchitis, and airways obstruction (total respiratory morbidity [TRM]) in Winnipeg in 1988 and 1992, using place of residence to divide people into quintiles according to average family income. Physician office visits, hospitalizations, and consultation referrals were each examined. Three age groups: 0 to 14 yr, 15 to 34 yr, and > or = 35 yr were studied. The prevalence of TRM was greater in low- than in high-income quintiles. Asthma prevalence was unrelated to income in the younger age groups; in the older group asthma was more common in low-income groups, but was less strongly related to income than was TRM. Asthma prevalence increased over the years studied, but the increase was not related to income level. There was some evidence of income-related diagnostic bias in that low-income patients were more likely to be labeled with a related diagnosis in addition to asthma than were high-income patients. Low-income patients had more physician contacts than did high-income patients. In terms of physician office visits, care continuity did not differ among income quintiles. Low-income quintiles had more hospitalizations than did high-income quintiles, and differences were larger than could be accounted for by diagnostic bias; asthma was probably more severe in low-income quintiles. High-income quintiles had more consultation referrals than did low-income quintiles.