Digestibility was estimated in vitro from a total of 167 oesophageal fistula samples collected over a period of 16 to 24 months from adult sheep grazing eight Phalaris tuberosa-Trifolium repens pastures stocked continuously at rates of from 2·5 to 37·1 sheep per ha. Corresponding estimates of the nitrogen content of the faeces and of the fistula samples, the live weight and faecal output of the sheep, and of the herbage availability of the pasture were obtained.
The observations were stratified on the bases of stocking rate, digestibility, month of the year, herbage availability or organicmatter intake. Relationships between digestibility estimated by the fistula technique and faecal nitrogen content (model B) and between digestibility and faecal nitrogen content, nitrogen content of the dietary dry and organic matter, and faecal output per unit live weight (model A) were calculated for each stratum.
There were significant differences between relationships derived for the various stocking rates, levels of digestibility, times of year, levels of intake and levels of herbage availability; predicted values of digestibility for given values of the independent variables differed by as much as 20 units of digestibility. Model A was substantially more precise than model B.
Differences between estimates of digestibility, derived by the fistula technique and by a general faecal nitrogen—digestibility relationship, established by Lambourne and Reardon (1963a) were calculated. The differences were related to digestibility and to herbage availability.
The implications of these results are discussed and it is concluded that faecal nitrogen content is not a satisfactory index of the digestibility of the diet selected by grazing sheep.