The hypothesis that manual work and exposure to vibration are antecedents to the development of osteoarthrosis was assessed employing a cross sectional study design. The frequency of osteoarthrosis in the acromioclavicular joint was studied in three groups of workers in the construction industry. Two groups were manual workers (54 bricklayers and 55 rock blasters); the third group consisted of 98 foremen. The radiographic appearance of the right and left acromioclavicular joints was classified into one of five grades of osteoarthrosis. A protocol was developed to assess exposure on the basis of job title, years of manual work, total weight lifted during working life, and total hours of exposure to vibrating tools. Odds ratios for job titles (manual worker v foreman) and for years of manual work as indicators of exposure were of similar magnitude of around 2.5. Construction workers who had lifted more than 709 tonnes had an increased risk of developing severe osteoarthrosis of the right acromioclavicular joint, odds ratio: 2.62 (95% confidence interval (95% CI), 1.13-6.06). The odds ratio for the left side was 7.67 (95% CI, 2.76-21.34). In the analysis of vibration exposure, workers who had been highly exposed to vibration had an odds ratio of 1.99 (95% CI, 1.00-3.92) on the right side and 2.20 (95% CI, 1.07-4.56) on the left. This effect almost disappeared after simultaneous adjustment for manual work. Occupational and ergonomic factors, such as the sum of lifted tonnes during working life, job title, and the sum of years of manual work seem to be risk factors for osteoarthrosis of the acromioclavicular joint, whereas vibration alone was a weaker risk factor.