SUMMARYThe main types of underlying mechanism in Raynaud's phenomenon were tested separately in thirty‐six patients. Within each of three groups of patients who were clinically classified according to strict definitions as having primary Raynaud's disease, systemic sclerosis or rheumatoid arthritis, there was pathophysiological heterogeneity. It is therefore suggested that in studies of treatment, at least when it is aimed primarily at one pathophysiological mechanism, the patients should be selected according to the mechanism rather than what is commonly done now, according to the associated disorde