The potential use of some plant-wax compounds as faecal markers to measure the botanical composition of herbivore diets
This thesis examined whether long-chain fatty alcohols or long-chain fatty acids could be used with n-alkanes to allow reliable diet composition estimates to be made in herbivores consuming complex diets. Experiment 1 was conducted in Glensaugh (Aberdeenshire) in summer 2001. Twelve Scottish Blackface wether sheep housed in metabolism crates were fed four different mixtures of three plant species (three animals per mixture) for a seven-period trial. Levels of cuticular wax n-alkanes, long-chain fatty alcohols and long-chain fatty acids were estimated in samples of individual plant species and the faeces from animals that consumed mixtures of these species. The original n-alkane method of Mayes et al. (1986) was modified to allow separate hydrocarbon (n-alkanes, n-alkenes and branched-chain alkanes), alcohol (free+esterified) and acid (free+esterified) fractions to be obtained from a single sample. The fraction containing alcohols and sterols was eluted from the silica gel column after removal of the hydrocarbons. Acids were extracted from the aqueous phase of saponification products after removal of hydrocarbons, alcohols and sterols, purified through silica gel columns and were converted into their methyl esters (FAMES) prior to analysis on a GC. Prior tests had been carried out to evaluate the reproducibility of the results obtained from the analytical method developed for quantifying alcohols and acids. Levels of these markers in samples obtained from Experiment 1 were used to calculate the dietary proportions of each species by a least-squares optimization procedure. To explore the differences between the different marker methods, the man squares of errors (EMS) between the actual and calculated dietary proportions of plant species were determined. In three out of the four mixtures, long-chain fatty alcohols were found to have the lowest discrepancies (lowest EMS values), followed by n-alkanes. The highest discrepancy was observed for long-chain fatty acids. Long-chain fatty acids showed the lowest discrepancy in one mixture and the highest in the others. It is concluded that, for this particular set of mixtures, long-chain fatty alcohols have promising potential to estimate composition of complex diets. However, the estimation from long-chain fatty acids was less good and n-alkanes were intermediate. Estimation from the combination of these marker classes was always better than the poorest individual marker.