The homes of 120 patients with asthma, 57 in Vancouver and 63 in Winnipeg, were studied. The characteristics of the homes were assessed by a questionnaire. Dust samples were collected and the indoor relative humidity was measured four times during the year covering all four seasons in both cities. Mite allergen levels were determined using monoclonal antibodies against Der p I and Der f I by the ELISA method. The mean levels of both mite allergens in mattress and floor samples in the homes in Vancouver and in Winnipeg were relatively low for all seasons. Mite allergen levels were found to be associated with city, season and individual home differences. They were significantly higher in Vancouver than in Winnipeg. Der p I and Der f I in mattress samples in both cities and Der f I in floor samples in Vancouver, varied by season. The indoor relative humidity level in the homes in Vancouver were also significantly higher than those in Winnipeg. There was, however, no significant association between the levels of indoor relative humidity and the levels of mite allergens after adjusting for variations in city, season and individual home. Although individual home differences were highly associated with mite allergen levels, only a few home characteristics were found to be related to mite allergen levels such as the type and the age of the home, the type of heating, the use of feather pillows and the number of occupants in the homes. Whether low levels of mite allergens are partially responsible for the relatively low prevalence of childhood asthma in Canada remains to be investigated.