Few methods exist for estimating quantitatively the diet composition of free-ranging herbivores. The current study examined whether long-chain fatty alcohols (alcohols) or long-chain fatty acids (acids) could be used along with n-alkanes to allow reliable diet composition estimates to be made in herbivores consuming complex diets. 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 experiment. Concentrations of cuticular wax n-alkanes, alcohols and acids were estimated in samples of individual plant species, and the faeces from animals that consumed mixtures of these species. These concentrations were then used to calculate the dietary proportions of each species by a least-squares optimization procedure. To explore the differences between the estimation methods (individual markers and their combinations), the mean squares of errors (EMS) between the actual and predicted proportions of plant species were calculated.

In three out of the four mixtures, alcohols had the lowest discrepancies (lowest EMS values), followed by n-alkanes and then acids. Acids yielded the lowest discrepancy in one mixture and the highest in the others.

It is concluded that, for this particular set of mixtures, alcohols had great potential to estimate composition of complex diets. However, the estimation using acids was less good and n-alkanes were of intermediate potential. Estimation from the combination of the three marker classes was always better than using the poorest individual marker.