Total intakes (chaff+SFM) were estimated using either the C31/C32 or the C32/C33 alkane pairs, based on
the whole-diet and faecal concentrations of these alkanes and the C32 release rate from the alkane-CRD
(Mayes et al. 1986, Dove and Mayes 1996). Supplement intakes were estimated either directly (from the
concentrations of C38 alkane in labelled-SFM and faeces), or indirectly by estimating the proportions of
chaff and SFM in the total intake, using the computer package EatWhat (Dove and Moore 1995). This uses
a least-squares procedure to find the combination of the alkane patterns of diet components (in this case,
chaff and SFM) which best matches the pattern of alkane concentrations observed in faeces. The latter
calculations were performed with and without the inclusion of the C28 alkane added to the beeswax. However,
the use of C28 in these calculations did not improve the estimates and will not be discussed further.
Calculations of the proportion of supplement in the consumed diet required the adjustment of faecal alkane
concentrations for incomplete faecal recovery of alkanes; the required estimates of recovery were obtained
from the total faecal collection made with the crate sheep. For the reasons discussed elsewhere (Mayes et al.
1986, Dove and Mayes 1996), alkane-based estimates of total intake do not require such corrections.