Alternate 2·4 m wide strips of a mixture of birdsfoot trefoil (Lotus corniculatus L.) with white clover (Trifolium repens L.) (BW), and of red clover (Trifolium pratense L.) (RC) were offered in the proportions (by area) of 0·80[ratio ]0·20, 0·67[ratio ]0·33, 0·33[ratio ]0·67 and 0·20[ratio ]0·80 to yearling heifers in groups of three for periods of 3 days over four replicates in time, balanced for effects of previous treatments. Observations on the distribution of grazing activity and biting rate were made over 3 h periods each evening, and on the morning of Day 2. Biting rates were consistently higher on BW than RC (52·3 v. 46·3±0·59 bites/min, P<0·001). Initially the animals showed a strong tendency to concentrate grazing on the sward of smaller proportional area, but the selection coefficient (log θ) changed from positive to negative with time as herbage on minority strips was depleted. These results are discussed in the context of concepts of partial grazing preference.