Major pharmacokinetic processes influencing polychlorinated biphenyl (PCB) accumulation in mammals include uptake, biotransformation, respiration, and excretion. We characterized some of the factors underlying PCB accumulation/loss by evaluating PCB concentrations and patterns in pre- and posthibernation grizzly bears (Ursus arctos horribilis) and their prey. The PCB congeners with vicinal meta- and para-chlorine unsubstituted hydrogen positions consistently showed loss both before and during hibernation, supporting the idea of a dominant role for biotransformation. Retention of all other studied congeners relative to that of PCB 194 varied widely (from <1 to 100%) and was highly correlated with log octanol--water partition coefficient (p < 0.0001). A lack of loss for most of these other congeners during hibernation supports the notion that excretion (e.g., fecal or urinary) or lack of uptake during the feeding season underlies their lack of accumulation, because hibernating bears do not eat or excrete. We estimate that grizzly bears retain less than 10% of total PCBs taken up from their diet. Our results suggest that for grizzly bears, depuration of PCBs via biotransformation is important (explaining ∼40% of loss), but that nonbiotransformation processes, such as excretion, may be more important (explaining ∼60% of loss). These findings, together with the approximately 91% loss of the persistent PCB 153 congener relative to PCB 194 in grizzly bears, raise important questions about how one defines persistence of PCBs in wildlife and may have bearing on the interpretation of food-web biomagnification studies.