The aim of this study was to evaluate the sulphur hexafluoride (SF 6) tracer technique for methane (CH4) emission measurement in sheep. Ten cryptorchid Romney sheep were involved in two indoor trials (T1 and T2), where daily CH4 emissions were individually measured both by the SF6 tracer ('tracer CH 4') and by the indirect calorimetry chamber ('chamber CH 4') techniques while fed on lucerne hay at 1.2 times maintenance requirements. Separate sets of permeation tubes with pre-calibrated permeation rates ('pre-calibrated PRs') were used in the two trials (for tracer CH 4) and at the time of T1 and T2 these tubes had been deployed in the rumen for 250 and 30 days, respectively. The tracer CH4 measurements were carried out for 2 (T1) and 5 (T 2) days in digestibility crates housed within a building (T 1) or a well-ventilated covered yard (T2). Sheep were transferred to calorimetry chambers for 3 days acclimatisation, followed by measurement of CH4 emission for 7 (T1) and 3 (T 2) days. In T1 samples from the chamber, outflow and inflow (collected over ∼22 h) were analysed for CH4 and SF 6 concentrations using the tracer protocol. Thus, PRs of SF 6 at the time of the trials ('calculated PRs') could be inferred and the corresponding CH4 emissions are then calculated using either the pre-calibrated PR or calculated PR. Permeation tubes were recovered at the end of the animal trials and their 'post-recovery PR' determined. In trial T 1, the tracer CH4 estimates (based on the pre-calibrated PR) were much higher and more variable than the chamber CH4 values. In this trial, the calculated PR and the post-recovery PR were similar from each other but smaller than the pre-calibrated PR, and when the calculated PR was used in place of the pre-calibrated PR the CH4 emission estimates agreed well with the chamber CH4 values. This suggested that the discrepancy was due to a declining PR during the long deployment time of the tubes in T1, an observation reported elsewhere. When the long intra-ruminal deployment was avoided in T2, good agreement between the techniques for CH4 emission measurement was observed. © 2008 The Animal Consortium.