Declining grassland productivity is a major concern in southern temperate Australia. Continuous grazing is thought to be a primary contributor to this decline, which is associated with the loss of perennial grasses. Landholders are evaluating grazing management strategies that might curb the loss of perennials and increase long-term productivity. This study reports on a comparison between continuous grazing and time-control grazing with sheep and cattle using a paired-paddock design at 5 locations in south-eastern Australia (lat. 30–42°S) over 6 years (1994–99). Pasture herbage mass, grassland species composition and basal cover of perennial grasses were assessed at 6-monthly intervals. Species abundance data were analysed by ANOVA, ordination (multi-dimensional scaling) and splining procedures to assess comparative trends between the 2 management treatments at each site. Species were categorised into major functional groups for analysis. Over all 5 sites there were few consistent differences between management treatments (continuous grazing v. time-control grazing). Basal cover was greater on the time-control grazing management compared with continuous grazing for most of the experimental period at 3 sites, but the initial values were also greater, resulting in a non-significant management × time interaction. Based on this study, we conclude that there was no apparent medium-term benefit of a multi-paddock rotational (time-control grazing) grazing system over continuous grazing for encouraging and maintaining a favourable botanical composition. The benefits for land managers from employing systems such as time-control grazing may accrue through other mechanisms. The study also highlights some of the difficulties with conducting on-farm paired-paddock research.