1. An experiment was conducted with dairy cows over two grazing seasons to investigate the feasibility of fixed rotational grazing. Each year a comparison was made between treatment A, which consisted of four equal paddocks grazed for 1 week and rested for 3, and treatment B, which contained five paddocks grazed for 1 week and rested for 4. The influence of the length of the rest periods on output per hectare and per cow was measured. Within each paddock approximately one-seventh of the area was allocated daily to each treatment group, there were no back fences and the paddocks were not topped or harrowed. Stocking rates were identical on both treatments at all times at 6·2 cows per ha throughout 1964; 7·1 cows per ha for weeks 1·9 inclusive and 6·2 cows per ha for weeks 10·20 inclusive in 1968. The fertilizer nitrogen applied (282 kg/ha) was equivalent to a daily rate of 2·02 kg/ha. The feed intake of each cow was measured on three occasions during 1967 and four in 1968.

2. Differences between treatments in outputs per ha were small and on average about 13 000 kg milk and 890 cow grazing days were produced per ha during the 20-week experimental period.

3. Differences between treatments A and B in herbage intake, milk production, milk composition and live-weight change were not statistically significant.

4. When 3·6 kg of a supplementary concentrate was given to some cows, positive responses in milk production, milk composition and live-weight change were found, but the value of the extra milk produced was less than the cost of the concentrate.