The EEGs of 90 aged psychiatric patients were subjected to a frequency analysis. In 48 cases a manual analysis was performed on the occipital tracing; in 52 cases an electronic analysis was undertaken of the occipital, temporal and frontal tracings; both methods were used in 10 cases. Half of the patients were diagnosed as having chronic brain syndrome, the other half as having a “functional” disorder. The following results were obtained: 1. 1. Brain syndrome patients had significantly lower peak frequencies than functional cases in the occipital and temporal areas. Peak frequency in the frontal area did not differentiate groups. 2. 2. Frequency spectra based on per cent-time measurements (manual analysis) and on per cent pen deflection (electronic analysis) were reliably different for the two diagnostic categories. Brain syndrome cases had more delta activity; functional patients showed more alpha and beta activity. 3. 3. Frequency spectra based on absolute voltage (electronic analysis) did not discriminate groups well. Only occipital alpha and beta values gave significant differences, with the brain syndrome patients having less voltage. 4. 4. The relative measure (per cent pen deflection) was more often in agreement with clinical EEG ratings than was the absolute measure (voltage). 5. 5. Total voltage for all frequencies showed little or no relation to psychiatric diagnosis. 6. 6. High intercorrelations were obtained between delta, alpha and beta voltage in the functional group, indicating the existence of a general factor of voltage level. Brain syndrome cases gave significantly lower correlations for delta activity, suggesting a greater independence of abnormal slow waves. These results offer substantial confirmation of our previously reported findings on clinical EEG ratings in the same group of patients. Discussion centered about a comparison of manual and electronic analysis and the merit of absolute and relative measures. It was concluded that the manual frequency analysis, which does not involve amplitude measurements, compares favorably with the electronic tronic technique, which is partially dependent on amplitude. It was further suggested that electronic measurements expressed in relative terms are more meaningful than those expressed in absolute terms, probably because they are not influenced by variations in voltage level.