WITH AN APPENDIX AN EXAMINATION OF A SEQUENCE OF RANDOMIZED ERRORS BY THE MONTE CARLO METHOD

The water content of the human body averages 60 per cent of its total weight. Techniques for measurement and criteria for accuracy have been developed for total body water but there are no generally accepted standards for extracellular fluid volume (EFV). It is our purpose in this paper to describe our exploration of the erratic varia-tions of extracellular fluid volume in normal sub-jects as measured by sucrose infusions and to dis-cuss wide fluctuations in what had been thought hitherto to be a "steady state." It is believed that most methods of estimating EFV using nonelectrolyte labels achieve rapid equilibration in, and measure mainly, interstitial and plasma water, and that the substances used diffuse incompletely into the cerebrospinal fluid, joint spaces, bone and tendon water, or into glandular lumina, or gastrointestinal tract water (2). This incomplete mixing of test substance in these spaces should not be expected to intro-duce large errors into this estimation since plasma and interstitial fluid must comprise the major por-tion of EFV. Walser, Seldin and Grollman (3) have proposed the term "functional EFV " for the volume thus measured. However, the accuracy of EFV estimates has been further questioned with reports that even large nonelectrolyte mole-cules, such as inulin and sucrose, may leave the EFV and penetrate cell membranes (4, 5). Most methods for measuring extracellular flui