About 80 per cent of the population secretes soluble blood group substances with A, B, and O(H) specificities corresponding to each individual's blood type; the remaining 20 per cent, the "nonsecretors," do not.' Most nonsecretors, however, do secrete large amounts of Lea-specific blood group substance, a glycoprotein similar in composition to the A, B, and O(H) substances of secretors.2 For the purposes of this communication, the most important difference between the substances is that the antigenic determinants of A, B, and O(H) substances contain L-fucose linked a-(1 -) 2) to D-galactosyl residues, while they are absent in Lea substance.In a scheme for the biosynthesis of blood group substances, Watkins and Morgan proposed that differences among individuals resulted from the genetically determined presence or absence of specific transglycosylases.3' 4 In their hypothesis, the product of the "H gene" is an enzyme that catalyzes the addition of L-fucose to D-galactosyl residues to form a-(1 --2) linkages and takes part in the synthesis of all blood group substances containing this particular grouping. The absence of the enzyme in the secretory organs of nonsecretors results in their inability to synthesize glycoproteins with A, B, or O(H) structures. This scheme was recently supported by the finding that the presence of 2'-fucosyllactose in human milk is directly correlated with the secretor status of the donor, and is found in the milk of secretors only.5 The structure of this trisaccharide as well as the structures of other sugars found in human milk are shown in Table 1. 2'-Fucosyllactose itself is synthesized by transfer of L-fucose from GDP-L-fucose to the galactosyl residue of lactose.8 The simultaneous absence in nonsecretors of glycoproteins containing O-a-L-fucopyranosyl-(1 -l 2)-O-3-D-galactopyranosyl structures and 2'-fucosyllactose, which has the same structure, suggests that the fucosyltransferase responsible for the synthesis of the trisaccharide is also a key factor in the synthesis of soluble blood group substances and that its presence or absence determines secretor status. It was possible to test this hypothesis directly because fucosyltransferases occur in human milk in soluble form, as demonstrated by the experiments reported in this communication.9In the present study, milk samples from secretors and nonsecretors were assayed for their ability to transfer L-fucose from GDP-L-fucose to the 2-position of galactosyl residues. As discussed below, this transfer is catalyzed by milk from secretors but not by milk from nonsecretors. Thus, as originally hypothesized by Watkins and 1\forgan,3' 4 secretor status in humans appears to be determined by the presence or absence of a specific enzyme which can be described as a GDP-t-fucose: f3-D-galactosylsaccharide a-2-L-fucosyltransferase.