Temperate grassy woodlands are endangered ecosystems in Australia, and many degraded remnants are in desperate need of understorey restoration. This experiment compared the effects of soil disturbance, weed control and mulch treatments on establishment of the original dominant grass, Themeda triandra Forssk., in a degraded white box (Eucalyptus albens Benth.) woodland at Cowra in central New South Wales (NSW). Awned Themeda seeds were surface-sown into replicated plots treated as follows: soil scalping, soil disturbance (by ripping), herbicide (simazine) application and retention of natural mulch.

Scalping combined with soil disturbance best promoted Themeda establishment (47.8% after 40 days and 28% after 518 days), and also reduced broadleaf-herb densities. By contrast, scalping without soil disturbance had the worst effect on Themeda establishment (5.2% after 40 days and 4.5% after 518 days). Disturbance significantly enhanced Themeda establishment and decreased the density of annual grasses and the basal cover of non-Themeda species. By contrast, the retention of 500–800 kg of natural surface mulch had no apparent effect on Themeda establishment. Contrary to expectations, simazine reduced the density and basal cover of all species, including Themeda, which is normally resistant to this herbicide. All combinations of the mulched, disturbed and herbicide treatments (i.e. all treatments except scalping) gave similar results, ranging from 10.7 to 22.0 Themeda plants m^–2 after 518 days. These results suggest that Themeda stands can be established in degraded box woodlands by using awned seed materials, with minimal seedbed preparation and simple sowing techniques. Further studies are required to determine whether established swards can resist weed invasion in the absence of ongoing weed management, and whether establishment success varies with soil conditions and landscape position.