ABSTRACTThe DM digestibility (DMD) and organic matter digestibility (OMD) of 36 tropical grass samples were determined bothin vivoandin vitro. Regressions relating OMD to DMD were calculated.In vivoOMD=1·02in vivoDMD+ 0·1, r=0·99.In vitroOMD=1·05in vitroDMD—4·7, r=0.99. At the same DMD, thein vitromethod predicted OMD 2·8 to 3·4 percentage units lower than thein vivomethod. This difference was caused by the difference in mean apparent digestibility of the ash,in vitro71%,in vivo47%.In vivoDMD was correlated (r=0·92, P<0·01) within vitroDMD and the regression relating these two factors had a residual standard deviation of ±1·42 digestibility units. This deviation was reduced to ±1·9 digestibility units when OMD was determined. For both DMD and OMD significant differences were found between grass species in the relation betweenin vivoandin vitrodigestibility.To overcome biased estimates of OMD it was suggested that conversion equations should be avoided, and that thein vitromethod be standardized with samples of knownin vivoOMD as similar as possible to thos