The existence of a discrete 'link' peptide in epithelial mucins has been debated for many years. There is evidence that at least some mucins contain a specific 'link' peptide (or glycopeptide) that enhances mucin polymerization by forming disulphide bridges to large mucin glycoprotein subunits. A major difficulty has been to know whether the reported differences in putative 'link' components represent artifacts generated by inter-laboratory differences in technical procedures used in mucin purification. The present paper outlines the results of a collaborative study involving five laboratories and 53 samples of purified gastrointestinal mucins (including salivary, gastric, small-intestinal and colonic mucins) prepared by five techniques from four different animal species. An early step in mucin purification in all cases was the addition of proteinase inhibitors. Representative mucins were analysed for their composition, electrophoretic mobility in SDS/polyacrylamide-gel electrophoresis before and after disulphide-bond reduction, and for their reactivity with monospecific antibodies developed against the 118 kDa putative 'link' glycopeptide isolated from either rat or human small-intestinal mucins. Our results indicate that, despite differences in laboratory techniques, preparative procedures, organs and species, each of the purified mucins contained a 'link' component that was released by disulphide-bond reduction and produced a band on SDS/polyacrylamide-gel electrophoresis at a position of approx. 118 kDa. After electroelution and analyses, the 118 kDa bands from the different mucins were found to have similar amino acid profiles and to contain carbohydrate. It would appear therefore that a 'link' glycopeptide of molecular mass approx. 118 kDa is common to all of the gastrointestinal mucins studied.