In vitro studies and a lactation trial were conducted to investigate the effects of fibrolytic enzyme mixtures at different inclusion amounts. Seven enzymes in amounts designed to mimic addition of 1, 5, 15, or 30 g/d to dairy diets were incubated in vitro with either soybean hulls or alfalfa for 24 or 48 h. Enzyme treatments generally increased in vitro dry matter disappearance (IVDMD), but not volatile fatty acid production. For some enzyme mixtures, lesser amounts of enzymes led to greater increases in IVDMD, whereas for others there were no differences among the amounts tested. The enzyme mixture with the most cellulase activity was the most effective enzyme in improving IVDMD. In additional in vitro experiments, the same enzymes were used at an amount of 5 g/d, as well as at other amounts that showed promising responses in the first trial. Preincubation of substrates with enzymes before fermentation also was tested. Alfalfa, soybean hulls, corn silage, and corn gluten feed were used as substrates. Preincubation of the substrate with enzymes for 18 h before in vitro fermentation improved IVDMD. The effect on substrate solubilization of incubating substrates with the enzymes but without rumen fluid was also studied. Addition of enzymes to substrates without subsequent fermentation did not solubilize significant amounts of dry matter, indicating that the positive effect of preincubation cannot be attributed directly to hydrolysis of substrates before the in vitro fermentation with ruminal microbes. The fibrolytic enzyme that appeared most promising in vitro did not affect lactational performance when fed to dairy cows.