To determine the appropriateness of the ferret as a model for human beta-carotene (beta-C) metabolism, we have perfused both 15,15'-beta-[14C]C and unlabeled beta-C through the upper 30-cm portion of the small intestine of ferrets in vivo. The effluents of a mesenteric lymph duct cannulation and a common bile duct cannulation, as well as portal vein blood periodically sampled via an indwelling catheter, were collected. Ten percent (9.5 +/- 0.06%) of the total administered beta-C was taken up by the intestine after a 4-h perfusion. Of the radioactivity taken up, 68.6 +/- 6.5% remained in the intestinal mucosa, 3.2 +/- 0.2% was recovered in the lymph, and 28.2 +/- 6.5% (calculated) was absorbed via the portal system. The total uptake/absorption of beta-C was 12.9 +/- 6.8 nmol.h-1.30 cm intestine-1. Large amounts of unchanged beta-C and relatively small amounts of both beta-apo-12'-carotenal and beta-apo-10'-carotenal were isolated in the intestinal mucosa after a 4-h perfusion with beta-C. Considerable amounts of metabolites more polar than retinol were formed and comprised 35% of the total radioactivity recovered in the intestinal mucosa. Polar metabolites were absorbed mostly into the portal venous system, whereas retinol and retinyl esters were absorbed mainly into the mesenteric lymph. Of the total absorbed radioactivity in lymph, 10 +/- 1.0% appeared as unchanged beta-C, with peak absorption occurring at 3 h after beginning the perfusion.(ABSTRACT TRUNCATED AT 250 WORDS)