We have measured the hydrolyses of α- and β-cellobiosyl fluorides by the Cel6A [cellobiohydrolase II (CBHII)] enzymes of Humicola insolens and Trichoderma reesei, which have essentially identical crystal structures [Varrot, Hastrup, Schülein and Davies (1999) Biochem. J. 337, 297-304]. The β-fluoride is hydrolysed according to Michaelis-Menten kinetics by both enzymes. When the ~ 2.0% of β-fluoride which is an inevitable contaminant in all preparations of the α-fluoride is hydrolysed by Cel7A (CBHI) of T. reesei before initial-rate measurements are made, both Cel6A enzymes show a sigmoidal dependence of rate on substrate concentration, as well as activation by cellobiose. These kinetics are consistent with the classic Hehre resynthesis-hydrolysis mechanism for glycosidase-catalysed hydrolysis of the ‘wrong’ glycosyl fluoride for both enzymes. The Michaelis-Menten kinetics of α-cellobiosyl fluoride hydrolysis by the T. reesei enzyme, and its inhibition by cellobiose, previously reported [Konstantinidis, Marsden and Sinnott (1993) Biochem. J. 291, 883-888] are withdrawn. 1H NMR monitoring of the hydrolysis of α-cellobiosyl fluoride by both enzymes reveals that in neither case is α-cellobiosyl fluoride released into solution in detectable quantities, but instead it appears to be hydrolysed in the enzyme active site as soon as it is formed.