The diagnostic yield of different electrophysiological criteria was examined to establish whether a peroneal palsy was due to compression of the nerve in the region of the capitulum fibulae. Slowing of sensory conduction along the segment of the nerve across the capitulum fibulae localized the lesion in 64% of 47 consecutive patients with a history indicating or suggesting compression of the nerve in the vicinity of the capitulum fibulae and there were no false positive findings in 18 patients whose peroneal palsy was not due to compression at the capitulum fibulae. In 20% of the patients with slowing along the segment across the capitulum, conduction velocity was normal when measured from the superior retinaculum to the popliteal fossa. Slowing along motor fibres (m. extensor digitorum brevis) localized the site of the lesion in one-third of the patients. Differences in amplitude and in split-up of the sensory responses recorded in the popliteal fossa as compared with those recorded distal to the capitulum fibulae were of limited diagnostic value because of many false positive findings among patients whose peroneal palsy was not due to compression of the nerve at the capitulum fibulae.