Clark & Harris (1985) examined the extent to which sheep, intermittently grazing grass/clover swards, preferentially harvested clover. They concluded that on swards in which the grass and clover were intermingled (mixed swards) there was no evidence of preferential selection of clover. However, preference for clover was reported for swards where the grass and clover were grown in alternating strips of varying widths and the preference declined in swards that were more than 50% clover. Our analysis of the Clark & Harris (1985) data does not support these conclusions. We found strong evidence for preferential selection of clover on the mixed swards and little evidence of decreasing preference on the strip swards. We estimated selection ratios in favour of clover of 2:1 on the mixed swards and 5:1 on the strip swards. The impact of these findings on the authors' conclusions is discussed.