( < 0.5-mm-mesh screen). When the three extracting methods were compared, method I gave higher readings than method II for high-viscosity barleys, and the reverse was true for low-viscosity barleys. The ranking of the barleys remained the same. Method III gave the lowest viscosity values, perhaps due to active endogenous B-glucanases in water extracts of the raw barley. When different barley cultivars were extracted by method I and subjected to different shear rates the viscosity differences were much more pronounced at the lower shear rates, particularly for high-viscosity barleys. Finally, the same seven barleys were compared in a feeding trial with chickens. The improvement in weight gain due to supplementation with a crude enzyme preparation from Trichoderma viride (Cellulase Tv) ranged from 2 to 4l% (P <0 .05). Similar improvements were obtained for feed consumption, feed-to-gain ratio and dry matter retention (P < 0.05). The highest simple correlation coefficients were between the viscosity of barleys that had been extracted by method I, measured at low shear rates, and relative weight gain (r : 0 .97 , week I plus week 2). The correlation coefficients using low shear rate data were slightly higher than those with high shear rate data.These results indicate that most pronounced differences among barleys were obtained with finely ground samples (<0.5 mm), extracted for one hour at 38'C, using a low pH 1.5 buffer (method I) and assayed at low shear rates. Under these conditions the viscosity data reflected the antinutritive properties of barley, which in turn were reduced or abolished by adding a crude enzyme preparation to the diet.