For the hydrolysis of the two glycosidic bonds of fluorescein di-beta-D-galactoside (FDG) by beta-galactosidase from Escherichia coli, small [Hofmann, J. & Sernetz, M. (1983) Anal. Biochem. 131, 180-186] to dramatic [Huang, Z. (1991) Biochemistry 30, 8535-8540] deviations from simple stepwise substrate-intermediate-product kinetics have been reported. Intermediate channelling, a preferred hydrolysis of the intermediate fluorescein mono-beta-D-galactoside (FMG) formed from FDG at the active site and thus in a favourable position for further reaction, has been postulated. As there were reasons to doubt the previous findings and conclusions, the hydrolysis experiments have been repeated at initial FDG concentrations of 7-200 microM, following the concentrations of FDG, FMG and fluorescein with a reliable method, quantitative HPLC, to completion of the reaction. The transient appearance of substantial amounts of the intermediate FMG also in experiments with 200 microM FDG already rules out the existence of the most efficient intermediate channelling deduced by Huang (1991) from measurements of the initially developing fluorescence, incorrectly ascribed to fluorescein. Redetermination of the Michaelis constants for FDG and FMG led to much higher values than those reported previously. Fitting the progress curves by means of nonlinear regression combined with numerical integration of the rate equations resulted in good fits of the normal stepwise substrate-intermediate-product mechanism, without any necessity of assuming a more complex course of the reaction. So one of the rare examples of the hydrolysis of two bonds at a single enzyme-substrate encounter has been invalidated.