Eight Angus steers whose methane emissions had had been found to be higher or lower than predicted when fed a commercial feedlot diet were re-tested on a medium quality forage diet. Methane emissions were within published ranges (136.4 g d⁻¹), but differences in actual vs predicted production between high and low ranked animals were diminished and several animals changed in rankings. This suggests that methane emission characteristics may not persist over time, and that any selection of animals for low methane emission may need to be diet specific.