Published online by Cambridge University Press: 14 December 2007
The nutrient intakes of mammalian herbivores depend on the amount and the nutrient content of the plant species and plant parts which they eat. We review the merits of oesophageal-fistulated (OF) animals, microhistological procedures, stable C-isotope discrimination and plant cuticular-wax markers as methods for estimating diet composition and intake in both ruminant and non-ruminant herbivores. We also briefly discuss methods based on grazing behaviour measurements or on H2O or Na turnover, and methods for estimating supplement or soil intake. Estimates of intake in ruminants are often based on separate measurements of faecal output and herbage digestibility. We review this approach and emphasize that, under some circumstances, the applicability of in vitro digestibility estimates based on OF extrusa is questionable. We discuss how plant-wax marker patterns can be used to check whether OF and test animals are consuming similar diets, but also emphasize that a major advantage of the use of plant-wax markers is that this approach may obviate altogether the need for OF animals. Estimates of total herbage intake can be partitioned into the intakes coming from different plant species and/or parts, provided diet composition can be measured. Diet composition estimates based on C-isotope discrimination have the major disadvantage that they cannot be taken to species level. By contrast, microhistological methods can identify many plant species in extrusa, digesta or faeces, but often a large proportion of plant fragments remains unidentifiable. Plant-wax hydrocarbons show great promise as markers for estimating diet composition and intake. However, we suggest that to be applicable in complex plant communities there is a need with this method either to recruit a wider range of wax markers (e.g. alcohols, sterols, fatty acids) or to use it in combination with other methods. We suggest that, in turn, this generates an urgent need for research on statistical aspects of the combined use of markers or methods, in relation to the error structures of the data or methods being combined and the standard errors of the resultant estimates of diet composition and intake. We conclude by discussing the extension of intake and/or diet composition measurements to the measurement of nutrient transactions within the gut, particularly in relation to the supply of absorbable nutrients.