We assessed total body water (TBW) and extracellular water (ECW) in thirty-four non-ascitic cirrhotics and twenty healthy controls by *H,O and Br dilution. In the same subjects, bioelectric impedance (BI) was recorded at multiple frequencies. Body hydration was similar for controls (mean 55.6 (SD 67)), lesssevere cirrhotics (Child-Pugh classification A; CPA; n 21, mean 56.2 (SD 6.2)) and moderately-severe cirrhotics (Child-Pugh classification B; CPB; n 13, mean 57.2 (SD 5.4)). However, intracellular water standardized per litre TBW was significantly higher in CPB subjects (mean 27.0 (SD 7-5); P c 0.01) compared with CPA (mean 21.3 (SD 10.6)) and control subjects (mean 18.0 (SD 9%)). Published formulas for predicting TBW and ECW from BI at multiple frequencies were applied to the cirrhotics. These formulas gave accurate predictions of TBW and ECW, although standard errors of estimates were higher for CPB subjects (TBW < 2 5 and ECW < 2-1 1) than those for CPA (TBW < 2-0 and ECW < 1.8 I) and control (TBW 1.4 and ECW 0.9 1) subjects.Liver cirrhosis: Total body water: Extracellular water: Bioelectric impedance Assessment of body water compartments is of great clinical interest. Body hydration and extracellular water (ECW) : intracellular water (ICW) ratio are known to change with nutritional status and under a variety of disease conditions. Malnourished patients often exhibit an increase in ECW relative to total body water (TBW; Rose 1994b). Nonetheless, obese subjects often show an increase in ECW relative to TBW, the pathophysiological basis of which is currently under investigation (Waki et al. 1991 ;Battistini et al. 1995). In addition to under-and overweight states, it has been shown that an increase in ECW relative to TBW may occur also in highly-trained athletes (Battistini et al. 1994). Thus, body-water distribution appears to be significantly linked to nutritional status in both physiological and clinical conditions.Measurements of TBW and ECW involve the use of invasive and time-consuming tracerdilution techniques and, therefore, are not suitable for use in clinical practice. Multifrequency bioelectric impedance analysis (MFBIA) is a novel technique for the assessment of TBW and ECW (van Marken Lichtenbelt et al. 1994). MFBIA has many advantages over dilution methods in that it is safe, inexpensive, portable and requires minimal training of the operator. Thus, MFBIA has a great potential to be employed as a 'bedside' technique for assessing body water volumes in ill subjects.