The intake and the digestibility of herbage, and the effect of level of supplementary concentrate food, were measured in late spring and late summer in two studies, each with eight multiparous, pregnant sows. In the spring study sows were nose rung to prevent rooting but only four nose-rung sows were available for the summer study; the remaining four were unrung. In both cases, sows had access to a plentiful (>2·5 t organic matter (OM) per ha) ryegrass/clover sward in a paddock of 1922 m2. After a week of adaptation to the herbage in the experimental paddock, sows were offered 1·5 or 3·0 kg/day concentrate for consecutive 2-week periods in a change-over experimental design with four sows on each treatment in each period. Samples of herbage were also taken to measure the sward density, chemical composition and n-alkane content. Herbage intake and digestibility estimates were calculated using the n-alkanes technique, with the marker dosed on small food pellets. In the spring study, the herbage intake ranged from 0·9 to 1·8 kg OM per day in the first period (herbage neutral-detergent fibre(NDF) content 439 (s.e.41·6) g/kg OM) and 0·2 to 1·4 kg in the second (475 (s.e.29·3) g NDF per kg OM). The intake was affected by the level of concentrate only in the second period. The results obtained during the summer study with the rung and unrung sows showed an intake between 0·9 and 2·4 kg OM per day in the first period (524 (s.e. 16·0) g NDF per kg OM) and between 1·3 and 4·8 kg in the second (526 (s.e. 21·8) g NDF per kg OM). A high intake estimation for certain unrung individuals appeared to reflect their frenzied feeding behaviour and possible loss of some marker pellets. There were no differences between level of concentrate treatments in either period. Digestibility of the diet was affected by the intake of herbage and the level of fibre consumed (P < 0·01). These herbage intakes equated to proportionately 0·50 (s.e. 0·05) and 0·66 (s.e. 0·1) (or 0·49 (s.e. 0·07) excluding problem sows) of the maintenance energy requirement in each season. Faeces of unrung sows indicated a high ingestion of soil or stones by some individuals: one of the sows produced a bulked faeces sample containing 450 g/kg fresh weight of stones, whilst another sow had a faecal ash content of 937 g/kg DM. The results indicate that the intake of nutrients from herbage by grazing sows is highly variable between seasons and individuals.