Abstract
The ability to predict the growth of 124 white clover cultivars in a grass sward under sheep grazing from performance in a spaced plant block was examined. Over a 3-year period the extent of variation in cultivar yield in grass swards explained by growth as spaced plants decreased from 45% in Year 1 to 10% in Year 3. Spaced plants gave a better prediction of performance of large-leaved than smallleaved cultivars. However, lines which had highest scores (i.e., when in the top 10) when grown as spaced plants in Years 2 and 3, were generally the best when grown in a grass sward in Years 2 and 3. This would suggest that selection of genotypes from the best lines on the basis of spaced plant data would provide material of reliable agronomic potential when grown in grass swards that are to be grazed.