The glycosyl-enzyme intermediate in lysozyme action has long been considered to be an oxocarbonium ion, although precedent from other glycosidases and theoretical considerations suggest it should be a covalent enzyme-substrate adduct. The mutation of threonine 26 to glutamic acid in the active site cleft of phage T4 lysozyme (T4L) produced an enzyme that cleaved the cell wall of Escherichia coli but left the product covalently bound to the enzyme. The crystalline complex was nonisomorphous with wild-type T4L, and analysis of its structure showed a covalent linkage between the product and the newly introduced glutamic acid 26. The covalently linked sugar ring was substantially distorted, suggesting that distortion of the substrate toward the transition state is important for catalysis, as originally proposed by Phillips. It is also postulated that the adduct formed by the mutant is an intermediate, consistent with a double displacement mechanism of action in which the glycosidic linkage is cleaved with retention of configuration as originally proposed by Koshland. The peptide part of the cell wall fragment displays extensive hydrogen-bonding interactions with the carboxyl-terminal domain of the enzyme, consistent with previous studies of mutations in T4L.