The relationship between metallothionein levels and concentrations of several metals and radionuclides was examined in liver tissues of raccoons (Procyon lotor, n = 47) from the Department of Energy's Savannah River Site in South Carolina to determine the applicability of metallothioneins as an initial screening device for exposure assessment in free-living mammals and environmental monitoring. Using a fluorescent marker and a cell sorter to measure metallothionein, a significant positive correlation was found across animals between levels of metallothioneins and concentrations of selenium (Pearson's r = .30), mercury (Pearson's r = .3 1), and copper (Pearson's r = .30) in liver tissue. Arsenic, cobalt, silver, thallium, and tin were below detection limits in most or all liver samples. Other metals, including cadmium, chromium, radiocesium (137-Cs), copper, lead, manganese, strontium, and vanadium, showed only weak and nonsignificant correlations with metallothionein. Concentrations of mercury were correlated with concentrations of selenium (Pearson's r = .73), manganese (Pearson's r = .56), and strontium (Pearson's r = .57). In an a posteriori test, there was a still unexplained positive correlation between mercury (Pearson r = .56), selenium (Pearson r = .54), and radiocesium (Pearson's r = .38) concentrations and background cellular autofluorescence, and a negative correlation of strontium with the latter (Kendall tau = -.38). Background cellular autofluorescence may represent a generalized cellular stress response, or a yet unidentified biomarker. To better understand which metals contribute to the induction of metallothionein, principle component analysis (PCA) was performed. The first three principle components explained 78% of the variance, with highest loadings being from mercury and radiocesium. Metallothionein levels did not correlate well with the principal components from the metals and radiocesium, while autofluorescent background levels tended to correlate better.