Although the erythrocyte sedimentation rate (ESR) is frequently used and well known to both doctors and patients, the rationale behind its clinical use has been little explored. In this study we evaluated 559 consultations in general practice when an ESR was performed. It was most often done in response to musculoskeletal aches and pains, followed by respiratory tract problems, a general feeling of illness, and abdominal complaints. In half the cases it was used for diagnostic purposes, and in a third for monitoring progression of disease or treatment. In these cases the general practitioners found the test useful. Fourteen per cent were taken as a screening procedure, and were afterwards evaluated as being of little or no clinical value. In 60% of the consultations the test exerted its influence mainly by supporting or reinforcing the doctor's clinical opinion. In 11% the results were unexpected, and forced the doctor to reconsider. In 22% the results were felt to be of little or no clinical consequence. The ESR is perceived as very useful by the doctors when a specific diagnosis is considered. The test in itself is rather unspecific. Its major impact lies in its ability to help reinforce or lessen diagnostic probabilities. It is imperative, though, for the doctor to have a pretest hypothesis against which the test result can be evaluated. An ESR taken without an aim, or when somatic disease is unlikely, is of little or no clinical value.