1. Herbage intake and diet composition of sheep were measured on two swards containing mainly Lolium perenne, Festuca rubra and Trifolium repens, on an organic soil, maintained at sward surface heights of approximately 3 and 5 cm. Small areas within the swards were artificially contaminated by injecting Cs-134 into the topsoil. The Cs-134 concentrations of each diet component were obtained by sampling the vegetation within the contaminated areas. The Cs-134 intake by the sheep was estimated under the assumption that the complete pasture had been contaminated. 2. The concentrations of Cs-134 in the diet components increased in the order: dead matter, grasses, T. repens, Cerastium fontanum. The Cs-134 intake by sheep was higher in summer than in spring and lowest in autumn. Seasonal differences in the Cs-134 intake resulted from changes in the Cs-134 concentrations of plant species, changes in the botanical composition of the sward and changes in herbage intake. There was a strong indication that the Cs-134 intake was lower on the shorter sward as a result of lower Cs-134 concentrations in the vegetation and a lower herbage intake. 3. The botanical composition of sward and diet was compared using similarity coefficients and principal coordinate analysis. The sheep's diet closely reflected the botanical composition of the 'top layer' of the sward, i.e. the top 2 cm of the short sward and the top 3 cm of the tall sward. Diets differed slightly between seasons due to changes in sward composition. Grasses made up 70-90% of the diet and accounted for most of the Cs-134 intake in spring and autumn. In summer, C. fontanum had Cs-134 concentrations which were up to nine times higher than in other species and it contributed up to 35% of the Cs-134 intake, despite low proportions in the diet. 4. The cumulative amount of Cs-134 ingested by the sheep between May and September through intake of herbage was equivalent to 40-70% of the Cs-134 taken up into the above-ground vegetation and 2.2-2.7% of the Cs-134 injected into the soil. It was estimated that most of the Cs-134 ingested would be returned to the pasture via excreta and only 0.10-0.16% of the total Cs-134 would be removed with the sheep if they were taken off the pasture at the end of September.