Positron emission tomography with 11C-2-deoxyglucose was used to determine the test-retest variability of regional cerebral glucose metabolism in 22 young normal right-handed men scanned twice in a 24-h period under baseline (resting) conditions. To assess the effects of scan order and time of day on variability, 12 subjects were scanned in the morning and afternoon of the same day (a.m.-p.m.) and 10 in the reverse order (p.m.-a.m.) with a night in between. The effect of anxiety on metabolism was also assessed. Seventy-three percent of the total subject group showed changes in whole brain metabolism from the first to the second measurement of 10% or less, with comparable changes in various cortical and subcortical regions. When a scaling factor was used to equate the whole brain metabolism in the two scans for each individual, the resulting average regional changes for each group were no more than 1%. This suggests that the proportion of the whole brain metabolism utilized regionally is stable in a group of subjects over time. Both groups of subjects had lower morning than afternoon metabolism, but the differences were slight in the p.m.-a.m. group. One measure of anxiety (pulse at run 1) was correlated with run 1 metabolism and with the percentage of change from run 1 to run 2. No significant run 2 correlations were observed. This is the first study to measure test-retest variability in cerebral glucose metabolism in a large sample of young normal subjects. It demonstrates that the deoxyglucose method yields low intrasubject variability and high stability over a 24-h period.