The purpose of this study was to test the validity of a multiple frequency bioimpedance spectroscopy (BIS) technique that estimates extracellular fluid volume (ECV), intracellular fluid volume (ICV), and total body water (TBW). Thirteen healthy males (mean +/- SD: age, 23 +/- 3 yr; body mass, 80.6 +/- 14.7 kg) had their TBW and ECV measured by ingesting dilution tracers (7.27 g deuterium oxide, 1.70 g sodium bromide; blood samples at 0 and 4 h). ICV was calculated as TBW minus ECV. Impedance was measured (50-500 kHz) at rest, on a nonconducting surface, with a BIS analyzer. Electrode placement, posture, exercise, food/fluid intake, and ambient temperature were controlled. Dilution measures (TBW, 51.00 +/- 9.30; ECV, 19.88 +/- 3.14; ICV, 31.12 +/- 6.80 L) and BIS volumes (TBW, 50.03 +/- 7.67; ECV, 20.95 +/- 3.33; ICV, 29.04 +/- 4.51 L) were significantly different for ECV (P < 0.01) and ICV (P < 0.05); some individual differences were large. The correlation coefficients of dilution versus BIS volumes (r = 0.93 to 0.96) were significant at P < 0.0001; SEEs were: TBW, 2.23 L; ECV, 1.26 L; and ICV, 1.71 L. We concluded that BIS is valid for between-subject comparisons of body fluid compartments, is appropriate in clinical settings where change in ECV/ICV ratio is important, and should be used by comparing the required level of accuracy to the inherent technique error/variance.