In an experiment started in June 1976, 14-16-month-old Hereford steers weighing 300 kg were allocated to five treatments. Cattle in four of the treatment groups, viz. high (H), medium (M), low (L) and low-high (L-H) growth rates were held in feedlots and offered whole oats supplemented with linseed meal, hay and minerals to obtain growth rates of 1.0, 0.7, 0.4 and 0.4-1.0 kg day-1 respectively. Steers in the fifth treatment group were grazed on pasture (P) to gain weight at 0.7 kg day-1 and received no supplements. Each treatment group of 10 steers was slaughtered at a mean liveweight of 450 kg. The carcasses of H steers were shorter, had more total fat and a greater depth of fat at the 12-13th rib than those of the L steers (fat depths of H, M and L groups were 10.1, 8.6 and 7.1 mm). The H steers also had a larger eye muscle area, a higher ABCAS (Australian Beef Carcase Appraisal System) score and higher carcass grades than the L steers. The L-H treatment appeared to reduce the fat content of the total side relative to the M treatment but the treatment was confounded with differences in carcass gain and carcass weight. Pasture grazing reduced carcass fat content as compared with grain feeding (15.2 v. 18.6%; P c 0.05), but differences in depth and weight of subcutaneous fat in the carcass side or in individual cuts were not significant. There was no detectable difference in tenderness due to level of feeding or feed type. Separately, the effects of feeding level and grain feeding increased carcass fatness by 16 and 22% respectively; we suggest that these effects are of commercial importance, particularly if the combined effects are additive.