The number of replicates per treatment was derived from a power calculation (for a power of 0.8), as the number of replicates required to identify a significant difference (P = 0.05) within any one period in CH4/milk yield (g/kg; this being identified one of the most important variables within the study), between treatments