The energy expenditure of freely grazing cattle was investigated in the National Institute of Agricultural Technology and Agricultural Science College of Balcarce, Argentina (37° 45'S, 58° 18'W), by the CO2 entry rate technique. Two experiments were carried out in the autumn in March 1994 (experiment 1) and in April 1995 (experiment 2) with animals prepared with catheters in the parotid gland (collection of saliva) and into the peritonea for infusion of a solution of 14C. Six Angus steers (259 (s.e. 11) kg) were used in experiment 1 and seven (298 (s.e. 36) kg) in experiment 2. In experiment 1 animals grazed ryegrass pastures for 5·5 h in two periods of 1·5 h in the morning and of 4 h in the afternoon and in experiment 2 animals grazed oat pasture in one period of 1 h in the morning. Twenty hours before and during the experiments a solution ofNaH14CO3 was infused at a rate of 9·4 (experiment 1) and 8·1 (experiment 2) μiCi/h for 48 h with portable peristaltic pumps carried by each animal. Saliva samples were collected at least after the first 20 h of infusion. The first sample was collected in the corral just before grazing and two samples were collected in each grazing period. Also, in experiment 1 three samples were taken during resting (noon, 1 h after grazing and the next morning). In addition, bite frequency, pasture availability, plant height, in vitro digestibility and crude protein were measured. Carbon dioxide production was calculated as the ratio between the rate of infusion of 14C (μCi/h) and the specific activity of CO2 (μCi/l CO2) in saliva samples. Bite frequency was 59 and 28 bites per min on the respective pasture of ryegrass (148 g dry matter (DM) per m2 and 10·5 cm height) and oat (228 g DM per m2 and 27 cm height). Energy expenditure (EE, kJ/h per kg M0·75) in corrals was 14·9 (experiment 1) and 14·3 (experiment 2), increasing to 22·6 (proportionately 0·52) when grazing at 59 bites per min and to 16·6 (0·16) when grazing was at 28 bites per min. One hour after grazing at 59 bites per min (experiment 1) the EE was as high as during grazing, and in the next morning (after 5·5 h) of grazing remained at 19·7 kJ/h per kg M0·75 (0·32). No differences in energy expenditure were found between periods of grazing in experiment 1. It was concluded that the increase in energy expenditure of cattle due to the activity of grazing depends on the rate of biting. Grazing for 10 h at a moderate rate may boost EE proportionately by only 0·06, however grazing at the highest rates could easily add proportionately 0·20.