Eighteen purebred steers of three genotypes, Aberdeen Angus (AA), Charolais (CH) and Holstein (HO), were divided within genotype into three groups of six animals and offered one of three different levels of feeding either moderate (M/M) or high (H/H) both for 20 weeks or moderate for the first 10 weeks followed by high for the remaining 10 weeks (M/H). Growth rates during the final 10 weeks of the experimental period differed between dietary regimen (M/M = 0.87; M/H = 1.25; and H/H = 1.02 kg/day; s.e.d. = 0.08; P < 0.001). Over the entire 20 week experimental period animals offered the M/M level of feeding grew more slowly (0.97 kg/day) than those offered the M/H and H/H level of feeding (1.20 kg/day; s.e.d. = 0.06; P < 0.002;. Mean growth rates for CH, HO and AA steers were 1.21, 1.23 and 1.03 kg/day (s.e.d. = 0.06; P < 0.05). The animals were all slaughtered at a fixed age of 18 months, according to the Meat and Livestock Commission Blueprint for beef and, 48 h post mortem, samples of m. longissimus lumborum (LL) and m. vastus lateralis (VL) were removed for analyses. Muscle fibres were classified histochemically, according to their contractile and metabolic properties, and muscle fibre size was measured. Fibre type frequency was calculated and, in LL, the total fibre number of the muscle was estimated. There was little impact of feeding level, or consequentially growth rate, on muscle fibre frequency and size. The effects seen were confined mainly to LL where there were significant differences between the M/M and H/-H groups with respect to fast twitch glycolytic fibres (mean % frequency (M/M = 40.1 and H/H = 44.3; s.e.d. = 1.4; P < 0.01); mean % area (M/M = 51.9 and H/H 56.0; s.e.d. = 1.5; P < 0.05)) and apparent total fibre number (M/M = 35.0; and H/H = 41.9 × 104; s.e.d. = 2.7; P < 0.05) which were greater in H/H than in M/M groups. However, in both LL and VL the predominant differences were related to genotype; in particular, overall fibre size was smallest in CH, while slow oxidative (SO; type I) fibre area was highest in AA. For LL, analysis across all animals showed a positive relationship between SO area, % area, % frequency and overall acceptability of meat at 14 days as evaluated by a trained sensory panel. No such relationship was observed for VL. The data suggest that in this study manipulation of feeding level has only a small impact on muscle fibre characteristics and that the differences between genotype and muscle type may be more important in determining the variability of overall acceptability than growth rate.