Salivas from 35 type Le(a-b-x-) subjects were tested for Le^a, Le^b, Le^x, A, B and H substances by both manual and automated quantitative hemagglutination inhibition methods. For comparison, similar data were gathered from 226 Lewis positive adults. Among the 13 Le(a-b-x-) nonsecretors, 7 had small quantities of salivary Le^a activity and 11 had low concentrations of salivary Lex by manual methods. None showed any salivary Le^b activity. By the automated method, low concentrations of Le^a activity were demonstrable in 12. Among the 22 Le(a-b-x-) secretors, 6 had minute quantities of salivary Lewis activity, mainly Le^b. The Le^b activity could not represent cross reactivity with H substance because most of those subjects with the highest H activities had no Le^b activity. Our quantitative studies show, when salivary Lewis activity is present in individuals whose red cells era type Le(a-b-x-), that the activity is much lower in comparison with individuals whose red cells are Lewis positive. Thus, based on quantitative salivary studies, it is possible to distinguish between Lewis positive (Le/) and Lewis negative (le/le) individuals. These observations suggest that the le gene is not a completely inert allele but that it governs the biosynthesis of serologically active substances. The le gene can be regarded as an inefficient Le which produces exactly the same products in a qualitative way as the Le gene but quantitatively far less. Another possibility is, as recently suggested, that the le gene actually governs the biosynthesis of distinct products Le^c and Le^d and that salivary Le^a and Le^b activities among Lewis negative subjects represent cross reactivity with these products.