Differences between domestic herbivores species in alkane faecal recoveries and the accuracy of subsequent estimates of diet composition

This study aimed to evaluate effects of animal species on faecal recovery of n-alkane markers. In addition, impacts of using different sets of alkane faecal recoveries on diet composition estimates was assessed. Six mature crossbred mares (383±59kg live weight (LW)), six adult cows (500±72kg LW) of Asturiana de los Valles breed, eight Cashmere does (29±5kg LW) and eight crossbreed ewes (30±5kg LW) were used. Animals were housed in individual stalls and randomly assigned to one of two diets composed of herbage (D1) or herbage plus heather (D2) (3 animals per diet in the equine and cattle groups and 4 animals per diet in the goat and sheep groups). Diet composition was estimated from alkane concentrations (i.e., C₂₅-C₃₃) in diets and faeces by least-squares procedures. Alkane faecal recovery (AFR) was incomplete and increased (P<0.001) linearly with carbon chain length in all animal species. Animal species had an effect on faecal recovery of all alkanes, with equines having the highest recovery and goats the lowest. Differences in AFR between equines and ruminant species were higher for the shorter chain alkanes and tended to decrease with increased of carbon chain length. Within ruminant species, cattle had, in general, the highest faecal recovery for all alkanes indicating a lower level of disappearance of these markers in the digestive tract of cattle. The AFR set used in diet composition calculations affected (P<0.05) the estimated proportions of all plant species in both diets. Utilization of uncorrected faecal alkane concentrations (AFR0) provided the poorest estimates of diet composition, and use of mean faecal recovery data of the dietary treatment (AFR1) yielded the most accurate. Results show that application of a general set of alkane faecal recoveries (AFR2) results in estimates of diet composition which distinguish less well between diet components with similar alkane profiles.