Methane emissions from a flock of 14, 1-year old sheep grazing on a grass and legume pasture were measured using a micrometeorological mass- balance method and a sulphur hexaflouride (SF6) tracer technique. The former measured the mean emission, over 45 min intervals, from all the sheep within a fenced 24 m x 24 m enclosure, from the enrichment of methane (CH4) in air as it passed over the sheep. The tracer technique measured emissions from a subset of 7 individual animals over 24 h periods from measurements of CH4 and SF6 concentrations in air exhaled by the sheep, and from the known rate of release of SF6 from small permeation tubes placed in the animals' rumens. Both methods gave highly similar results for 4 out of 5 days. When the species composition of dietary intake was steady during the last two days of measurement, the mean emission rate from the mass-balance method was 11.9 ± 1.5 (SEM) g CH4 sheep-1 d-1, while the rate from the tracer technique was 11.7 ± 0.4 (SEM) g CH4 sheep-1 d-1. These rates are for sheep with mean live mass of 27 kg, with a measured dry matter intake of 508 g sheep-1 d-1 and pasture dry matter digestibility of 69.5%. There was close agreement between these measurements and estimates from algorithms used to predict methane emissions from sheep for the Australian National Greenhouse Gas Inventory.