Reliability of the clinical and electromyographic examination of tendon reflexes.
 The reliability of clinical examination of the tendon reflexes was examined by studying inter-observer agreement.
 Twenty patients were examined by three neurologists.
 The briskness of the tendon reflexes in arms and legs was scored on a nine-point scale.
 In 28% of the 160 examined reflexes the observations disagreed 2 scale units or more.
 Disagreement on the presence of asymmetry occurred in 45% of the 80 reflex pairs.
 In 15% one observer judged a reflex pair to be symmetrical while another observer found asymmetry of at least 2 scale units.
 In a second experiment clinical observation of apparently asymmetrical quadriceps reflexes was compared with measurement by surface electromyography.
 A significant, semi-logarithmic relationship was found between clinical scores and measured reflex amplitudes.
 Measured reflex asymmetry always agreed with clinical asymmetry, and the magnitudes of right-left amplitude differences were correlated with the magnitude of clinically observed asymmetry.
 The bedside examination of tendon reflexes is subject to considerable inter-observer disagreement.
