The aim of this experiment was to investigate whether different diets at weaning determine the microbial populations established in the rumen, together with its methanogenic capacity, and whether these differences are consistent over a longer time period. Twenty ewes with single lambs were used in two periods. Period I: 10 lambs had access only to grass hay whilst with the dam and for 8 weeks after weaning (group H). The other 10 lambs had free access to concentrate and grass hay whilst with the dam and were fed with a 60 : 40 mix of concentrate and grass hay for 8 weeks after weaning (group C). Eight weeks after weaning, methane emissions were measured in polycarbonate chambers over a 3-day period. After methane emission measurements, five lambs from each group were randomly selected and slaughtered and samples of rumen content collected for measuring rumen fermentation parameters and for microbial enumeration (total and cellulolytic bacteria and methanogenic archeaea) by most probable number. Period II: the remaining 10 lambs were grouped together and fed the same diet (grass and concentrate) for 4 months. After this period, all animals were fed concentrate and grass hay (60 : 40) for 2 weeks and introduced in to the chambers to measure methane emissions over a 3-day period. After measurements, they were slaughtered and rumen samples collected and analysed as in Period I. Lambs from group H produced more (P = 0.04) methane than group C lambs (26.0 v. 22.5 L/kg DM intake) in Period I. Group H lambs also had less total bacteria (10.2 × 1010 v. 61.6 × 1010 cells; P = 0.284) but more cellulolytic bacteria (40.6 × 109 v. 10.0 × 109 cells; P = 0.098) and methanogenic archaea (37.1 × 109 v. 19.0 × 109 cells; P = 0.113) than group C lambs in Period I. The acetate to propionate ratio tended to be higher (P = 0.089) in group H lambs than in group C lambs (3.00 v. 2.35). In Period II, methane produced was not different (P > 0.05) between the groups (26.6. v. 25.7 L/kg DM intake by group C and H lambs, respectively). Microbial numbers and fermentation parameters were also similar in samples collected from both experimental groups in Period II. Our results show that the differences observed as a result of providing different diets at weaning disappear in the long-term. It may be appropriate to study a wider range of dietary treatments to better understand the factors determining the microbial populations establishing in the rumen.