In a population sample health survey, body mass, previous trauma and physical stress were studied for associations with coxarthrosis. A sample of 8,000 persons representative of the Finnish population aged 30 or over was invited for examination, and 90 percent participated. On the basis of a standardized clinical examination, a physician diagnosed coxarthrosis in 6 percent of the women and 4 percent of the men. The prevalence rose with age. In persons with a past traumatic lower-limb injury, the odds ratio of unilateral coxarthrosis was 2.1 and of bilateral coxarthrosis 1.5, as adjusted for sex, age and other determinants using logistic regression. The sum index reflecting self-reported features of physical stress in present or previous occupations was directly proportional to the prevalence of coxarthrosis. Body mass index (kg/m2) was closely associated with bilateral coxarthrosis; the adjusted odds ratio (95 percent confidence intervals) for indices > 35, compared to those < 25, was 2.8 (1.4-5.7). In terms of the population attributable fraction, prior trauma, physical stress and body mass were estimated to explain 59 percent of the prevalence of coxarthrosis. The potential for primary prevention may be great, but longitudinal population studies are necessary to elucidate causal significance of the risk factors.