Measurements of plasma volume with 125I human serum albumin, extracellular water (ECW) with 82Br-minus, and total body water (TBW) with 3-H2O were made on 16 postoperative patients, 15 depleted patients, and three control subjects. Intracellular water (ICW) was calculated as the difference between TBW and ECW. The observed findings for the series as a whole showed no change in blood volume, an increase of 3.7 I in ECW, and a decrease of 1.5 I in ICW as compared to predicted values based on current weight. Compared to predicted values based on normal (pre-illness) weight, the observed body weight decreased 16% and ICW decreased 22%; this suggests that body weight consistently under-estimates the extent of nutritional depletion. Severe depletion seen in postoperative patients indicates that frequently earlier use of total parenteral nutrition would be beneficial. ICW was found to be the most reliable single index of moderate or severe nutritional depletion; errors in estimating normal values interfere with its use in mild depletion. The ratio of ECW:TBW best reflects distortion of body water composition; it is largely independent of weight, and is a more sensitive index than absolute values of ICW. Repeat measurements of body composition were made on 9 patients given total parenteral nutrition for an average period of 18 days. On the average, there was an ECW decrease of 1.8 I, a body weight increase of 2 kg, and an ICW increase of 3.2 I after parenteral nutrition. The increased ICW represents the repletion of half of the average initial deficit of 6.1 The initial ratio of ECW:TBW of 0.58 was reduced to 0.50, returning it about three-quarters of the way to the expected normal value of 0.48. Intravenous administration of 5% glucose as sole source of calories may be a factor in distortion of body water compartments.