Evaluation of very long-chain fatty acids and n-alkane epicuticular compounds as markers for estimating diet composition of sheep fed heathland vegetation species

Application of very long-chain fatty acids (VLCFA) as an alternative or a complement to the alkane markers to estimate diet composition of sheep offered six different diets was evaluated. Twenty-four adult crossbred sheep were housed in metabolic crates and fed with diets composed of different proportions of herbaceous (Lolium perenne and Trifolium repens) and woody species (Erica spp., Calluna vulgaris and Ulex gallii). Diet composition was estimated from VLCFA (i.e., C22–C34) and alkane (i.e., C25–C33) concentrations in diet and faeces, by least-squares procedures, considering the plant species effectively offered to the animals (C1) or these plant species plus feeds that were not part of the diets but were available in the field at each experimental period (C2). Prior to this, faecal VLCFA and alkane concentrations were uncorrected for incomplete faecal recovery (FC0), corrected using recovery data for each individual ewe (FC1), corrected using mean recovery rate of the dietary treatment that the animal belonged to (FC2), or corrected using mean recovery rate across all experimental diets (FC3). For all diet components, total VLCFA concentrations were higher (P<0.01) than those observed for alkanes, with a predomination of even-chain (P<0.001) over odd-chain. Faecal recovery was incomplete and tended to increase with carbon-chain length in a curvilinear (P<0.001) fashion for the even-chain VLCFA, and a linear (P<0.001) one for odd-chain VLCFA and alkanes. Diet composition had an effect (P<0.001) on faecal recovery of both markers. Estimates of diet composition obtained using alkanes alone were more accurate (P<0.05) than those obtained with VLCFA alone while their combination resulted in an improvement (P<0.05) of the accuracy. Estimates obtained when considering plant species offered effectively to the animals were more accurate (P<0.001) than those obtained when considering more plant species as possible diet components. In general, use of faecal marker concentrations corrected with FC2 and FC3 provided equally accurate estimates of diet composition, indicating that an overall recovery data can be applied to these markers in grazing studies without losing accuracy in the estimation. Although the use of alkanes resulted in more accurate estimates of diet composition, results obtained in this study showed that VLCFA can be useful markers for studying diet selection of sheep grazing on shrubland vegetation communities. However, a suitable adjustment of their faecal concentrations prior to their application is recommended. It is also concluded that use of VLCFA in combination with alkanes could result in more accurate estimates of diet composition.