The purpose of this study was to identify whether the central extracellular fluid volume status following and pypo- hypervolaemia can be measured by the initial distribution volume of glucose or by the extravascular lung water. These two estimates were compared with the initial distribution volume of sucrose which has been used as an indicator for the measurement of the extracellular fluid volume. The above three estimates were determined by the administration of glucose, chilled saline and sucrose solutions, before and after haemorrhage (30 mL kg−1), and subsequent fluid load (lactated Ringer's solution 90 mL kg−1). The distribution volumes of glucose and sucrose decreased after haemorrhage and increased after fluid load compared with normovolaemic values, and a linear correlation was obtained between these two distribution volumes (r=0.93, P<0.001, n=36). How ever, the extravascular lung water remained statistically unchanged throughout the procedure, despite a weak linear correlation with the sucrose distribution volume ( r=0.38, n=33, P<0.05). These results indicate that the initial distribution volume of glucose is more useful as an indicator of the central extracellular fluid volume status than the extravascular lung water.