The purpose of this study was to identify the level of house dust mites (HDMs) and their allergens in mattresses, not selected on their owners atopic status, and to find associated factors. Dust was collected from 68 mattresses. The recruitment was population-based and conducted during the screening phase of a HDM intervention study. The visited persons declared to have had a "cold" bedroom the previous winter. HDMs were counted and dust was analysed by ELISA for Der 1 (= Der f 1+ Der p 1+ Der m 1). Multiple regression analysis was carried out to find housing conditions associated with high HDM levels. Type of housing, mattress age and self-assessed winter bedroom-temperature explained 47% of Der 1. Median concentrations were 3.77 microg Der 1/g and 1 HDM/0.1 g dust. Both immunochemically and microscopically Dermatophagoides farinae was dominant; D. pteronyssinus less frequent but important; and D. microceras insignificant. In 62% of these suburban homes the mattress dust exceeded 2 microg Der 1/g; and measurement of both Der f 1 and Der p 1 was necessary and sufficient to evaluate HDM allergen exposure. The association with a high HDM level was highest and most consistent for one-family houses.