OBJECTIVE: The literature on the bioavailability of selenium (Se) from meats, especially beef, is meager, and that which existed when this research began suggested that Se was not highly bioavailable. In addition, much of the analytical values for Se in beef predated the Food and Drug Administration's 1973 approval of Se as an additive to feeds and mineral premixes of livestock. DESIGN: One hundred and thirty-six weanling female Fischer 344 rats were divided into two dietary groups: the selenium deficient group in which animals were fed a torula yeast (TY) basal diet which contained 0.008 mg/kg Se and the control group in which animals were fed the TY diet to which was added 0.10 mg/kg Se as sodium selenite. RESULTS: After 6 weeks of dietary treatment liver glutathione peroxidase (GSHPx) activity had fallen in the Se-deficient rats to 2.4% of that of control rats. At this time (week 6) rats from the Se-deficient TY diet were refed diets containing 0.10 mg/kg Se as selenite, selenate, raw or cooked ground beef that had been freeze-dried. During the Se-repletion period rats were sacrificed at weeks 1, 3, 5 and 8. Liver GSHPx activity and total Se levels in liver and muscle tissue were the criteria of Se bioavailability. After 8 weeks of Se resupplementation the recovery of liver GSHPx activity compared to the control animals (set at 100%) were selenite (98%, p > 0.05), selenate (117%, p < 0.05), raw beef (127%, p < 0.05) and cooked ground beef (139%, p < 0.05). Total Se in both liver and muscle tissue reflected the liver GSHPx activity with the total Se concentration in tissues being highest for cooked beef. CONCLUSION: The data suggest that bioavailability of Se from ground beef is greater than that from either selenite or selenate.