Forty-two samples (seven from each grass) and six composite samples of Merker grass (Pennisetum purpureum), Guinea grass hybrid (Panicum maximum), and Buffel grass (Cenchrus ciliaris) were used in this study to evaluate the artificial rumen, or in vitro cellulose digestion, technique (1). These forage samples of known nutritive value (4), as determined in conventional digestion trials with cattle (3), were harvested at 49 to 55 and at 63 to 69 days of age at the Gurabo Substation farm. No relationship was found between ad libitum intake (R.I. or V.I.) and 12- and 24-hour I.V.C.D. Within-trial variability due to lag-time differences could probably account for this. I.V.C.D. at 36 hours may be used as the best estimate of forage digestibility in terms of D.D.M. (r = 0.95), D.E. (r = 0.94), and T.D.N, (r = 0.94). Forty-eight-hour I.V.C.D. values could also be used as a predictor of forage digestibility in terms of D.D.M. (r = 0.87) and N.V.I, (r = 0.85). Nevertheless, the highest correlations were obtained with in vitro values at 36 hours. There was a tendency for correlation coefficients to increase up to 36-hour digestion periods with increasing fermentation rates. From the data in the present study, the regression equations to predict in vitro D.D.M. (Y), in vitro D.E. (Y), and in vitro T.D.N. (Y) from 36-hour I.V.C.D. (X) were Y = 12.84 + 1.23 X, Y = 1.801 X - 11.76, and Y = 17.79 + 0.906 X, respectively. Maximum cellulose digestion was apparently obtained at 60 hours for four of the six forages studied. When two inoculum donor animals were fed the same forage, no marked differences in 60- and 72-hour I.V.C.D. values were observed.