An experiment was carried out to evaluate the use of alkanes for estimating diet composition of goats and sheep offered three different dietary treatments. Twelve animals as two groups of 4 crossbred goats (G1, 24 kg live weight; G2, 22 kg) and 4 crossbred sheep (S, 26 kg live weight), were housed in metabolism pens. Animals were offered daily a total of 1 kg DM/100 kg live weight. G1 received 70% ryegrass (Lolium perenne) and 30% gorse (Ulex gallii), G2 received 70% ryegrass and 30% heather (Erica sp.) and S group ate 100% ryegrass. Diet composition was estimated from the alkane concentrations (using all alkanes from C23 to C36 or only odd-chain alkanes C27, C29, C31 and C33) in diet and faeces (with or without correction for incomplete faecal recoveries) using least-squares procedures.
Dietary treatment and animal species significantly affected alkane faecal recoveries, except for C24 and C36. When applying the faecal recovery corrections, there were no significant differences between measured proportions of dietary components and those estimated using all alkanes or odd-chain alkanes. In contrast, the proportions calculated without faecal recovery correction differed significantly (P<0.05) from the actual proportions and over-estimated the amount in the diet of those plant components with higher concentrations of long-chain alkanes (Erica sp. and Lolium perenne). The results indicate that alkanes are useful markers to estimate diet composition, however, it was observed that animal species and diet composition influenced the faecal recovery of alkanes. This suggests that the use of the alkane methodology for estimating the diet selection of grazing animals should be preceded by a calculation of the actual alkane faecal recoveries for each experimental condition.