A routinely employed and specific affinity chromatography purification method was used to isolate cellobiohydrolase II (CBHII) from a Trichoderma reesei culture supernatant. Two different identically purified batches of CBHII were found to have distinctly different properties in barley beta-glucan hydrolysis. The abilities of the two preparations to produce small oligosaccharides and reduce the viscosity of beta-glucan were substantially different, but no contamination or heterogeneity was detected in sodium dodecyl sulfate-electrophoresis used to assess the protein purity. Furthermore, the specific activities of the preparations on cellotetraose substrate were similar. Careful analysis with specific chromogenic oligosaccharide substrates showed that the considerable variation in beta-glucan hydrolysis was caused by a minor endoglucanase contamination consisting of only 0.4% of the total protein. The contamination is possibly caused by nonspecific interactions between the proteins and chromatographic column material or by unfavorable protein-protein interactions during the chromatophoric separation. The data demonstrate clearly the caution needed in the interpretation of hydrolysis studies on polymeric substrates with cellulolytic enzyme preparations.