In eight anaesthetized post-absorptive dogs we measured the concentration and specific radioactivity of phenylalanine and leucine in arterial and femoral-venous plasma, together with hindlimb flow during a continuous infusion of L-[ring-2,6-3H]phenylalanine and [1-14C]leucine. The femoral-venous plasma concentration was greater than arterial for both phenylalanine and leucine (P less than 0.05 for each). Despite net amino acid release there was a significant removal of both labelled phenylalanine and labelled leucine. Consequently, a significant dilution of specific radioactivity was observed between artery and vein for both radio-tracers. The uptake of leucine from the arterial circulation by the hindlimb exceeded by 2.6-fold that of phenylalanine; the measured molar ratio of leucine to phenylalanine in hindlimb muscle protein averaged 2.4 +/- 0.1. Since phenylalanine is neither synthesized nor degraded by muscle tissue, the measured removal of tracer and the dilution of tracer specific radioactivity across the hindlimb can be used to estimate rates of phenylalanine incorporation into, and release from, tissue protein. The estimated rate of protein synthesis by hindlimb averaged 644 +/- 250 nmol of phenylalanine/min. This was exceeded by the rate of tissue protein degradation (987 +/- 285 nmol of phenylalanine/min). The present results demonstrate that the dilution of the specific radioactivity of labelled phenylalanine can be readily measured across dog hindlimb. This measurement, coupled with an estimate of tissue blood flow, can provide a readily measured, non-destructive, method for estimation of protein turnover in specific muscle beds in vivo. Measurements can be made repeatedly over time in a single experiment, allowing the study of factors which regulate protein turnover. The method developed here in dogs can be readily extended to clinical studies.