Standard theory assumes that osteoarthritis is a catabolic disease, characterised by cartilage destruction brought about by absolute or relative overload. This theory fits the observed facts in secondary osteoarthritis, but when it is tested against the idiopathic disease, the disease profile includes many phenomena which are difficult to reconcile with the overload theory. It incorporates, for example, no adequate explanation for the prevalence patterns observed in human and primate epidemiology, the idiosyncratic distribution, the sparing of the ankle and wrist joints, the production of osteoarthritis by immobilisation, the anabolic dominance of early disease, and several pathological changes including the osteophyte, the reduplicated tide-mark, and the extra-articular contracture. The evidence better fits the concept of two different diseases with a possible, but not obligatory, final common pathway. In the case of the idiopathic disease, the unused arc hypothesis proposed 50 years ago by Harrison et al., reinforced with the concepts of synovial stasis and positive feedback, tests better against the evidence than the overload paradigm. These proposed adjuncts to the hypothesis are speculative and as yet untested. The results could be of more than academic interest. If the unused arc hypothesis turns out to be correct, idiopathic osteoarthritis should be preventable and conceivably, in its early stages, arrestable.