Heparinized samples of plasma, cerebrospinal fluid (CSF) and lymph from intestinal, prescapular and popliteal lymph nodes were collected from young lambs in order to characterize and compare the insulin-like growth factor-binding proteins (IGFBPs) in extracellular fluids with those from the circulation. Prior to collection and analysis, the superiority of heparin for plasma preparation was established over EDTA and citrate or the use of serum, according to the extent of IGF-I and IGF-II binding achieved in the high molecular mass IGFBP region in vitro. The IGFBPs were characterized by ligand blotting and competitive binding techniques using radiolabelled IGF-I, IGF-II and the truncated IGF analogue, des(1–3)IGF-I, as well as by ligand blotting of fractions after Superose 6 chromatography of incubations of fluids with labelled factors. This combined analysis demonstrated an IGF-II-specific binding protein at approximately 250 kDa that was present in plasma and each lymph type and presumably represented the soluble type-2 IGF receptor; a complex of 130 kDa containing 52 kDa and 46 kDa binding proteins that was labelled by all three IGF peptides was particularly evident in plasma and intestinal lymph and was probably a complex between IGFBP-3 and the acid-labile subunit; and a group of binding proteins that chromatographed as IGF complexes at approximately 50 kDa. This last group contained IGFBP bands of 52, 46, 35, 28 and 23·5 kDa in plasma and all lymphs as well as an IGF-II-specific band of 22 kDa in prescapular and popliteal lymphs. CSF differed qualitatively from plasma and lymph in that the 52/46 kDa IGFBP bands were undetectable in this fluid; the 35 kDa band was the predominant binding protein, and neither this nor the 28, 23·5 and 22 kDa proteins bound des(1–3)IGF-I to any significant extent. The 52,46 and 28 kDa bands in plasma and lymph did bind this ligand. Immunoblots using antisera against bovine IGFBP-2 showed binding at 35 kDa in all fluids as well as several bands at lower molecular masses. Taken together these results show not only marked differences in the binding protein profiles of sheep plasma, lymph and CSF, but both qualitative and quantitative differences between intestinal, prescapular and popliteal lymphs. We speculate that the differences between lymphs may result from tissuespecific release of binding proteins into extracellular fluid. Journal of Endocrinology (1991) 129, 59–68