Phalaris tiller and plant population characteristics were monitored in sown Australian phalaris–subterranean clover pastures over 3 seasons (1999–2001) to determine the impact of fertiliser and grazing method on phalaris persistence in south-western Victoria. Four grazing systems were tested: set-stocked, low phosphorus (P) fertiliser input (SS low P); set-stocked, high P fertiliser input (SS high P); simple rotation, high P (RG 4 paddock); and intensive rotation, high P (RG intensive). Within each year there was no significant difference in tiller density (tillers/m²) between the grazing systems. Phalaris tiller density declined (P<0.05) on all treatments from June 1999 to June 2001. There was a significant effect of grazing method on tiller size (mg/tiller); tillers growing under rotational grazing systems were much larger than those in set-stocked systems. There was some evidence of tiller size density compensation across the grazing management treatments; however, the slope of the trade-off between tiller size and density was not as steep as those reported for other species. In general, the phalaris tiller populations under rotationally grazed treatments were characterised by lower density per m² of phalaris clump and larger size, compared with the set-stocked treatments. Both the phalaris tiller and clump density declined at a similar rate on all treatments during this experiment, suggesting that there was some limitation to phalaris persistence irrespective of grazing system. Measured leaf appearance intervals (days/leaf) indicated that a lack of tillering sites was not a contributing factor in the observed tiller density decline. It is likely that the combination of high grazing pressure, below average rainfall and subsoil acidity contributed to the observed phalaris population decline.