A modification of an established technique for quantifying metallothionein (MT) in mammals was used to evaluate the toxicological importance of MT as a biological indicator of stress from chronic cadmium toxicity in brook trout Salvelinus fontinalis. In a 30-d study, fish mortality was significantly increased but growth was not altered by exposure to 3.6 μg Cd/L or more. After chronic exposure to cadmium, both mortality and whole-body residues showed a dose-response relation over the exposure range of 3.6 to 60.6 μg Cd/L. Concentrations of MT were increased significantly in all exposures that resulted in significant mortality; however, they showed no dose-response relation to cadmium exposure, nor were they correlated with mortality or whole-body cadmium residues. Consequently, measurement of MT alone was not a useful indicator of cadmium toxicity in brook trout. A better biological indicator was the amount of free cadmium in liver tissue. Free cadmium (the difference between the amount of cadmium in the supernatant resulting from centrifugation at 100,000 x gravity and the amount bound to MT in an unsaturated condition) showed a dose-dependent increase with increasing cadmium exposure and was highly correlated with mortality and whole-body residues. Free cadmium concentrations were significantly elevated in all exposures to 3.6 μg Cd/L or greater. The presence of free cadmium in tissues of brook trout from all exposures suggests that MT was not saturated with cadmium before the appearance of pathological effects and thus conflicts with the “spillover” hypothesis. As judged by the results of our toxicity studies, the spillover hypothesis should be redefined as a continuum of toxic responses to varying balances between the relative abundance of metals present and their respective binding affinities for MT.