In popular belief patients with chronic arthritis take alcohol for its analgesic effect. To test this we studied by validated questionnaire the past and present alcohol consumption of 103 patients with primary osteoarthritis of the hip (OA), 95 patients with rheumatoid arthritis (RA), and 90 orthopaedic non-arthritic controls. OA men were most likely and RA men least likely to have been heavy drinkers at any time of their lives. Mean red corpuscular volume (MCV), gamma-glutamyltransferase (GGT), and serum uric acid (SUA) levels did not correlate with reported alcohol consumption. Two of 93 OA femoral heads examined had avascular change; both were from heavy drinkers. The abstemiousness of RA men compared with their OA counterparts was due to a striking increase in joint pain after drinking alcohol (p = 0.004), fear of adverse drug reactions with alcohol, and a widespread belief not expressed by OA men that 'alcohol and arthritis do not mix'.