Forty-five plots (3 by 2 m) of perennial ryegrass-white clover pasture were randomly allocated within 3 sites to 5 winter defoliation interval treatments to determine the effect of site on best defoliation interval as measured by winter and spring herbage accumulation. Defoliation intervals were 14, 28, 42, 56, and 84 days. The defoliation intervals resulting in maximum total winter dry matter (DM) accumulation at sites 1 and 3 were 42, 56 and 84 days. Available feed accumulated at the end of winter was highest for the 84-day defoliation interval at both sites (2.0 and 1.4 t DM/ha, respectively). At site 2, maximum total winter DM accumulation was achieved for defoliation intervals of 28, 42 and 56 days. Available feed accumulation at the end of winter was 1.2 t DM/ha for the 42-day interval and did not increase thereafter. The difference in response was associated with pasture composition. The results of the experiment showed that there is scope for flexibility in winter grazing management to suit livestock requirements without compromising total winter dry matter accumulation.