SU MMARYBased on the knowledge that alpaca (Lama pacos) have a lower fractional outflow rate of feed particles (particulate FOR) from their forestomach than sheep (San Martin 1987), the current study measured methane (CH 4 ) production and other digestion parameters in these species in three successive experiments (1, 2 and 3): Experiment 1, lucerne hay fed indoors ; Experiment 2, grazed on perennial ryegrass/white clover pasture (PRG/WC) ; and Experiment 3, grazed on birdsfoot trefoil (Lotus corniculatus) pasture (Lotus). Six male alpaca and six castrated Romney sheep were simultaneously and successively fed on the forages either ad libitum or at generous herbage allowances (grazing). CH 4 production (g/day) (using the sulphur hexafluoride tracer technique), voluntary feed intake (VFI), diet quality, and protozoa counts and volatile fatty acid concentrations in samples of forestomach contents were determined. In addition, feed digestibility, energy and nitrogen (N) balances and microbial N supply from the forestomach (using purine derivatives excretion) were measured in Experiment 1.Diets selected by alpaca were of lower quality than those selected by sheep, and the voluntary gross energy intakes (GEI, MJ) per kg of liveweight 0 . 75 were consistently lower (P<0 . 001) for the alpaca than for the sheep (0 . 74 v. 1 . 36, 0 . 61 v. 1 . 32 and 0 . 77 v. 2 . 53 on lucerne hay, PRG/WC and Lotus, respectively). Alpaca and sheep did not differ (P>0 . 05) in their CH 4 yields ( % GEI) when fed on lucerne hay (5 . 1 v. 4 . 7), but alpaca had a higher CH 4 yield when fed on PRG/WC (9 . 4 v. 7 . 5, P<0 . 05) and Lotus (6 . 4 v. 2 . 7, P<0 . 001). When grazing on Lotus, the sheep had very high protozoa counts in their forestomach contents, compared with those with the other forages and those in the alpaca. On lucerne hay and Lotus, but not on PRG/WC, the alpaca had higher (P<0 . 01) acetate/propionate ratio in their forestomach fluid than sheep. When fed on lucerne hay, alpaca and sheep did not differ (P>0 . 05) in diet N partition or microbial N yield, but alpaca had higher (P<0 . 05) neutral detergent fibre digestibility (0 . 478 v. 0 . 461) and lower (P<0 . 01) urinary energy losses (5 . 2 v. 5 . 8 % GEI) than sheep. It is suggested that differences between these species in forestomach particulate FOR might have been the underlying physiological mechanism responsible for the differences in CH 4 yield, although the between-species differences in VFI and diet quality also had a major effect on it.