The intra- and inter-observer variation in the measurement of the penile brachial index (PBI) was prospectively assessed in 30 patients. An appreciable error was discovered that cast doubt on the reliability of the measurements. Up to 20% of patients deemed normal by one observer would be deemed abnormal by a second. The PBI appears to be an inaccurate and poorly reproducible test of penile blood flow which can lead to a false diagnosis and subsequent management of impotent patients. There would appear no justification for the continued use of the penile brachial index in the assessment of erectile dysfunction.