Herbage intake estimation, using the n-alkane technique, requires the determination of the ratio of ‘dosed’ to ‘natural’ n-alkane concentrations in faeces. The present study was designed to evaluate the accuracy and precision of intake estimates based on the analysis of pooled samples of total daily faecal output and those derived from rectal grab samples.
Wether lambs maintained in individual metabolism crates, offered fresh herbage ad libitumand dosed daily with 120 mg of n-dotriacontane (C32), were used. Actual intake was recorded daily and samples of total faecal output were retained either for individual analysis or to produce 6-day composites by pooling representative amounts of material from each day's faecal output. Different pooling strategies were evaluated: a constant volume of fresh faecal material from each day or a constant weight (or constant volume) of dried and ground material from each day's faecal output. In addition, daily rectal grab samples, obtained at the same time each day, were pooled using a constant daily volume of fresh faeces. Herbage intake was estimated using the ratios of faecal concentrations of n-hentriacontane (C31) or n-tritriacontane (C33) relative to that for C32.
Similar patterns of effects were observed for intake estimates derived from the C31: C32 and C33:C32 n-alkane ratios for the different faecal sample types. Comparisons based on intake estimates derived from the use of the C33: C32 ratio showed that there were no significant differences between actual intake and estimates based either on the mean of daily samples or on the different composites of total daily faecal output. Intake estimates based on the analysis of composites of total daily faecal output were consistently lower (c. 3%) than those based on the analysis of daily faecal samples. A difference of this magnitude is expected, based on the algebraic relationship between the mean of a ratio of two variables (e.g. faecal concentrations of C33 and C32) and the ratio of the corresponding means. Intake estimates derived from the use of rectal grab samples, taken 3 h after the daily dose was given, yielded a proportionate overestimation of actual intake of 0·06 (P < 0·10). No significant differences were detected in the precision of intake estimates derived either from individual daily samples, composites of total daily faecal output, or composites of rectal grab samples, although the latter yielded the lowest precision.
It is concluded that faecal samples can be pooled on a gravimetric or a volumetric basis over several days with negligible loss in precision of intake estimates. Such pooling greatly reduces the workload and costs involved in using the n-alkane technique. Sampling and analytical errors associated with t e technique were found to be very small, and the processing of faecal samples in duplicate would result in a negligible increase in the precision of intake estimates. The need for further evaluation of rectal grab sampling under grazing conditions is highlighted.