An inventory of 162 hay samples from Austrian permanent grasslands was used to obtain information about the precision of the in vitro NDF degradability (NDFd) measured by the Daisy fermenter and its accuracy to predict in situ NDFd. The within forage standard error of the in vitro NDFd triplicate, obtained in five consecutive incubations, was equal to 2.8%, while the effect of the four jar positions in the fermenter was not significant. The cutting frequency had a great impact on the in situ effective NDFd of hays, which ranged (P<0.01) from values of 32.9, 43.1 and 48.3% in hays obtained from 2, 3 and 4 cuts/season, respectively. The regression analysis between the in vitro and in situ NDFd values (measured at 48h and effective, k=3%/h) allowed to obtain medium degrees of correlation (r2 = 0.69 - 0.71; P<0.01) and low levels of accuracy (RSE = 4.0 -4.6 %).
