The present experiment was designed to elucidate the mechanism of the methionine-sparing effect of excess branched-chain amino acids (BCAA) reported in the previous paper (Langer & Fuller, 2000). Twelve growing gilts (30–35 kg) were prepared with arterial catheters. After recovery, they received for 7 d a semipurified diet with a balanced amino acid pattern. On the 7th day blood samples were taken before (16 h postabsorptive) and after the morning meal (4 h postprandial). The animals were then divided into three groups and received for a further 7 d a methionine-limiting diet (80 % of requirement) (1) without any amino acid excess; (2) with excess leucine (50 % over requirement); or (3) with excesses of all three BCAA (leucine, isoleucine, valine, each 50 % over the requirement). On the 7th day blood samples were taken as in the first period, after which the animals were killed and liver and muscle samples taken. Plasma amino acid and branched-chain keto acid (BCKA) concentrations in the blood and branched-chain keto-acid dehydrogenase (BCKDH; EC 126.96.36.199) activity in liver and muscle homogenates were determined. Compared with those on the balanced diet, pigs fed on methionine-limiting diets had significantly lower (P < 0·05) plasma methionine concentrations in the postprandial but not in the postabsorptive state. There was no effect of either leucine or a mixture of all three BCAA fed in excess on plasma methionine concentrations. Excess dietary leucine reduced (P < 0·05) the plasma concentrations of isoleucine and valine in both the postprandial and postabsorptive states. Plasma concentrations of the BCKA reflected the changes in the corresponding amino acids. Basal BCKDH activity in the liver and total BCKDH activity in the biceps femoris muscle were significantly (P < 0·05) increased by excesses of leucine or all BCAA.