Power spectral analysis of the heart rate was compared with standard tests of autonomic function performed on 15 insulin-dependent diabetic patients with symptoms of autonomic dysfunction. The high-frequency power, low-frequency power, and total power of the heart rate power spectrum measured in the supine position--possible measures of parasympathetic nervous system function--were significant predictors of the expiratory-inspiratory ratio, the beat-to-beat heart rate difference on deep respiration, the standard deviation of the resting heart rate, the 30:15 ratio, and the Valsalva ratio. The change in low-frequency power on moving from the supine to upright position, a possible measure of sympathetic nervous system function, was a modest predictor of the systolic and diastolic blood pressure fall in response to postural change, but a poor predictor of the blood pressure response to isometric exercise and a cold stimulus. This study supports the role of heart rate power spectral analysis as a clinical test of autonomic function in patients with known or suspected autonomic failure.