Evaluation of assays for phenolic compounds on the basis of in vitro gas production by rumen micro-organisms

12 Bolivian fodder tree samples were selected in order to achieve a range of phenolic compositions as gauged by analyses for total phenols (TP) by the Prussian flue method, protein precipitation activity (PPA) by radial diffusion and condensed tannin (CT) by an acid butanol method. Aqueous acetone extracts of phenols from these samples were prepared and added, with and without the phenol binding agent polyvinylpyrrolidone (PVP), to an in vitro gas production system which used rumen micro-organisms to ferment glucose, Differences in the cumulative volume of gas produced with and without PVP were taken as measures of the effects of phenols on fermentation. Two measures of inhibition were used, the maximum differences achieved (generally after 9 or 12 h. incubation) termed Difference T max and the difference when gas production had virtually ceased (after 166 h incubation) termed Difference T166, Linear regression analysis was used to test the hypothesis that the assays correlated with the inhibition of fermentation. Significant (P < 0.01) correlations were found between Difference T max, Difference T166 and TP (R(2) = 0.82 and 0.83 respectively) and PPA (R(2) = 0.63 and 0.74 respectively). There was no significant (P > 0.5) correlation between CT and either measure of inhibition. It was concluded that TP and, less accurately, PPA were useful indicators of the degree of inhibition of rumen micro-organisms by phenolic compounds in tree leaves. The acid butanol assay for condensed tannins was not an indicator of rumen micro-organism inhibition when used for a range of tree species.