Management of beef cows grazing extensively grazed semi-natural pastures in temperate regions in late autumn can require supplements to be offered. The effects of supplementation with soya bean meal on the diet selected by Charolais cows and on their subsequent performance were examined for an 8-week period in late autumn in 2 years. Three groups of eight cows were compared: non-supplemented dry cows (D), non-supplemented (L) and supplemented (LS) lactating cows. The amount of soya bean meal supplement offered per cow was 250 g d(-1) in year 1 and 800 g d(-1) in year 2. Dietary choices were measured by direct observations and herbage intake was estimated in year 2. Supplementation affected neither diet selection (L: 0.42 vs. LS: 0.43 for the proportion of bites on green patches in year 1; 0.24 vs. 0.22 in year 2) nor daily organic matter intake of herbage (L: 15.6 vs. LS: 15.9 kg d(-1)), which may have resulted from an adequate crude protein concentration of herbage. The higher total dry matter intake by cows offered the supplement reduced losses in live weight (L: -1212 vs. LS: -828 g d(-1); P < 0.01) rather than increased milk production (L: 5.1 vs. LS: 5.0 kg d(-1)). This may be linked to the low milk yield potential of the Charolais cows. The use of lactating cows rather than dry cows for pasture management in late autumn would increase the utilization of herbage but a reduction in liveweight losses of cows by supplementation is unlikely to be economic.