Exercise is a very good and dangerous thing. The harms of exercise provide a workload of sports medicine for the young and orthopaedic surgery for the old. But there are also many specific conditions where appropriate doses of this hazardous intervention may help—fibromyalgia, for instance. A Canadian meta-analysis of 6 studies (J Rheumatol 2008;35:1130–44) provides moderate-quality evidence that aerobic-only exercise training to levels recommended by the American College of Sports Medicine has positive effects on global well-being and physical function, and possibly on pain and tender points. In early rheumatoid arthritis, a randomised trial from Sweden (Arthritis Rheum 2008;59:325–31) found that coaching patients to maintain healthy physical activity has benefits above standard rehabilitation, both in muscle strength and perceived health status; but, the investigators add that “the mechanisms remain unclear, as self-reported physical activity at healthy level[s] did not change.” Trials of exercise are certainly fraught with difficulty, but an Australian team must be congratulated on devising a programme of sham exercise to administer to …