Abstract Gibbons' method for methyl pentoses has been found to be a reliable method for the estimation of salivary fucose concentration. The standard error of measurement is of the order of 1 per cent. A spectrophotometric technique is described for detecting salivas which are so contaminated as to prevent their fucose content being accurately determinable. The salivary fucose content of a particular individual varies considerably from time to time. Blood group substances are known to be present in considerable amounts in saliva, and these substances are known to contain 12 to 20 per cent by weight of fucose. The estimation of salivary fucose content may prove to be a reliable index of blood group substance content.