We studied seven analytical methods of estimating the plasma volume from the decay curves of indocyanine green. Fifteen volunteers received 1.0 mg.kg-1 of the dye by intravenous injection and the plasma concentration was measured continuously using spectrophotometry. Plasma volumes were calculated using three single-regression methods (1-a, 1-b, 1-c) and four biexponential regression methods (2-a, 2-b, 2-c, 2-d). The means (SD) of 1-a, 1-b and 1-c were 39 (5.0), 44 (5.7) and 54 (11.5) ml.kg-1, respectively, and these were significantly different from each other (p < 0.05). The values for methods 2-b, 2-c and 2-d were similar to each other: 39 (4.6), 40 (4.1) and 40 (4.0) ml.kg-1, respectively. These required more than 3 min circulation or mixing time. When the time allowed for mixing was less than 3 min (method 2-a) the plasma volume was underestimated. We conclude that plasma volume estimation using indocyanine green and spectrophotometry is most accurate when the mixing time is adequate (3-5 min) and the decay curves are analysed using biexponential regression.