Comparison of in vivo organic matter digestion of native Australian shrubs by sheep to in vitro and in sacco predictions

There is increasing interest in selection of genotypes of Australian perennial shrubs such as saltbush (Atriplex nummularia, Atriplex amnicola and Atriplex semibaccata), orange wattle (Acacia saligna), small-leaved bluebush (Maireana brevifolia) and rhagodia (Rhagodia preissii) for extensive grazing systems with sheep or cattle. A major limitation to cultivar development is that determining the in vivo organic matter digestibility of forage is expensive, time consuming and requires substantial amounts of biomass. A number of in vitro and in sacco techniques are available to predict in vivo digestibility of grasses and legumes however none of these prediction methods have been calibrated to in vivo data for Australian native shrubs. The aim of this study was to determine the in vivo digestibility of native shrub forage and compare these data to predictions using in vitro and in sacco methodologies. The hypothesis tested was that there is a linear relationship between in vivo organic matter digestibility (OMD) of Australian perennial shrub forage and a number of commonly used prediction methods. Of all the methods used, in vitro gas production from microbial fermentation showed the best relationship with in vivo OMD (r2 = 0.904). This method appears to be suitable for broad screening and ranking of genotypes. The in vitro pepsin-cellulase technique did not provide a good first estimate of in vivo OMD across all plant species but may be suitable to rank genotypes if data are calibrated with internal standards to manage intrinsic assay variation (step 1) followed by a correction to account for the high salt content of some plants (step 2) and a final linear correction to account for systematic overestimation of OMD of native shrubs (step 3). Further testing and refining of this third calibration step is required. The prediction of in vivo OMD using a 72 h in sacco digestion was adequate for the saltbush species but not for the orange wattle and small-leaved bluebush accessions. Prediction of in vivo OMD using the DaisyII rumen fluid digestion system was not satisfactory and there was not a statistically significant relationship between the acid detergent fibre, neutral detergent fibre or the acid detergent lignin content of the shrubs and in vivo OMD. The differences between the various laboratory-based methods to predict in vivo OMD are discussed and we hypothesise that the presence of plant secondary compounds may interfere with rumen microbial fermentation.