Oesophageal fistulated cattle can give unreliable estimates of the proportion of legume in the diets of resident animals grazing tropical pastures

Estimates of the diet of resident intact steers grazing four different tropical grass/legume pastures were made by two methods. In the first, oesophageal fistulated (OF) steers were used on three days and the botanical composition of the extrusa determined microscopically. The second method measured the carbon ratio (δ 13 C) of the faeces of the resident steers over three days, and the percentage legume in the diet was calculated after correction for digestibility differences between the grass and legume components. Estimates of legume percentages in the diet by the two methods differed markedly; OF steers selected a diet containing much more legume than did resident steers, the regression of extrusa estimates on δ 13 C having a negative slope. This, and other evidence, clearly shows that estimates using OF steers of the diet of resident cattle grazing tropical grass/legume pastures can be most unreliable. Reasons for this unreliability are briefly discussed.