The extent of variation in tissue density and hepatic nonheme iron concentration has been examined at autopsy in 21 adult livers. Samples were taken from each liver at inferior and superior sites in the midaxillary, anterior axillary, and midclavicular lines. Histologic examination showed diffuse metastatic carcinoma, cirrhosis, fibrosis, necrosis, steatosis, or congestion in 19 livers; two livers were normal. Density was determined by saline displacement of 0.5- to 1.0-g specimens. Nonheme iron concentration was measured at each site in samples of the size obtained by wedge (0.5 to 1.0 g) and percutaneous needle (0.005 to 0.010 g) biopsy using specially developed chemical assays. Density was uniform within each liver. Despite the inclusion of diseased tissues, the variation in density among the 21 livers was small (coefficient of variation, 1.25%). The mean (+/- 1 SD) hepatic density was 1.051 +/- 0.013 g/mL (range, 1.017 to 1.077 g/mL). Within each liver, the nonheme iron also was uniformly distributed among the six sites. Chemical measurements of nonheme iron concentration were not significantly different in samples of the size obtained by wedge or percutaneous liver biopsy. All the hepatic nonheme iron determinations were below the upper 95% confidence limit of concentrations in adult males (480 micrograms/g). In the absence of focal lesions, the uniformity in hepatic density and nonheme iron distribution supports the assumption of several clinical methods for measuring liver storage iron (wedge and needle biopsy, determination of hepatic magnetic susceptibility, computed tomography, and magnetic resonance imaging) that one sample of liver tissue is representative of the whole organ.