Most considerations of protein digestion assume that amino acids are made available to the body tissues as the free form but recent reports suggest that a substantial portion may be absorbed from the stomach region of ruminants as small peptides (1). However, the quantitative relevance of this absorption to amino acid supply to tissues remains unclear. We have previously indicated that in early lactation at least part of the amino acid supply to the lactating mammary gland may be met from blood derived peptides or small proteins (2) but at present there is no direct evidence to suggest that peptides can contribute amino acids to the gland for milk protein synthesis. However, it has been demonstrated in dairy animals that the uptake of certain amino acids across the mammary gland is insufficient to account for their output in milk protein (J.A. Metcalf, unpublished observation) and the possible utilisation of amino acids in ‘non-free’ form must be considered. The present study involves the use of a dual-labelled tracer approach to evaluate the ability of the mammary gland to utilise amino acids in peptide-bound form for milk protein synthesis. The technique involves infusion into the external pudic artery (EPA) supplying one half of die gland of a dipeptide XY where Y is a [13C]-labelled amino acid, coupled with a simultaneous (jugular) infusion of the amino acid Y but with a deuterium label. The jugular infusion allows a correction for recycled amino acid generated by whole animal (i.e. non-mammary) hydrolysis of the infused peptide. In theory if the half of the gland receiving direct infusion of the dipeptide utilises peptide-bound Y for milk protein synthesis then the ratio of [13C] : deuterium should be greater in casein secreted from that half of the gland compared with the other (control) side of the gland.