Cation-exchange chromatography effectively concentrates the cell growth activity present in whey and we have used this process as a basis to characterise further the growth factors present in bovine milk. Under neutral conditions, total bioactivity in the growth factor-enriched cation-exchange fraction chromatographed with an apparent molecular mass of 80-100 kDa. In contrast, acid gel-filtration chromatography resolved two peaks of cell growth activity. A peak at 15-25 kDa contained the bulk of growth activity for Balb/c 3T3 fibroblasts while bioactivity for L6 myoblasts and skin fibroblasts eluted with a molecular mass of 6 kDa. A peak of inhibitory activity for Mv1Lu and MDCK cells also eluted at 15-25 kDa. Both IGF-I and IGF-II were purified from fractions that eluted at 6 kDa, although the IGF peptides alone did not account for the total bioactivity recovered. Platelet- derived growth factor (PDGF), identified by radioreceptor assay, eluted at a slightly higher molecular mas than the peak of growth activity for Balb/c 3T3 cells, and an anti-PDGF antibody was without effect on the growth of Balb/c 3T3 cells in response to the whey-derived factors. Further purification of the inhibitory activity for epithelial cells yielded a sequence for transforming growth factor β (TGF-β), and all inhibitory activity for MvILu cells was immuno-neutralised by an antibody against TGF-β. In contrast, this antibody decreased the growth of Balb/c 3T3 fibroblasts in the whey-derived extract by only 10%. Finally, a cocktail of recombinant growth factors containing IGF-I, IGF-II. PDGF, TGF-β and fibroblast growth factor 2 stimulated growth of Balb/c 3T3 cells to a level equivalent to only 51% of that observed in the milk-derived growth factor preparation. We conclude that: (i) cell growth activity recovered from bovine whey is resent in acid- labile high molecular weight complexes; (ii) all cell growth inhibitory activity for epithelial cells can be accounted for by TGF-β; (iii) IGF-I and IGF-II co-elute with the major peak of activity for L6 myoblasts and skin fibroblasts, although the IGF peptides alone do not explain the growth of these cells in the whey-derived extract; and (iv) neither PDGF nor TGF-β account for the 15-25 kDa peak of Balb/c 3T3 growth activity. These data suggest the presence of additional mitogenic factors in bovine milk.