Figure 1 and 3, Data collection and sampling design
In the field trial, four large blocks (35 x 40 m) were established randomly in a pasture dominated by Lovegrass. Two of these blocks were fenced to exclude grazing by cattle and limit access by other native and exotic gazers (e.g. kangaroos, wallabies, hares and rabbits). The other two blocks were left open to grazing. In each of the four blocks, forty-eight 5 x 5 m plots were established which included a 2 m buffer around the perimeter of each plot (September 2006). A randomised, split plot design was used with the grazing treatment at the block level and three treatments applied randomly at the plot level, including 16 replicates of each treatment combination in each block. In total, the trial was made up of 192 plots with 8 replicates of 24 treatment combinations (Firn et al. 2010).
At the plot level, three types of treatments were applied in October each year from 2006 to 2008: (1) three management treatments to remove and reduce Lovegrass: herbicide application, mowing and a control, and (2) slow-release fertilizer was applied to half of the plots in a pellet form (N 21.6%, P 1.1%, K 4.1%) at a low application rate of 2 kg/ha.
Firn et al. (2010) contains species abundance results from 2006 to 2008. In December 2009, we measured species abundance, functional traits and soil nutrient levels from just the grazing/grazing exclusion treatment and the fertilizer/no fertilizer treatment combinations. We randomly sampled 16 plots from each of the four treatment combinations (8 plots within each block).
We also randomly established 16 plots in a pasture under the same grazing management located less than 200 metres (separated by a forest) from the field trial where native grasses remain the dominant species (Lovegrass abundance was < 2.5%, Table 1). These plots were established in an area the same size as the field trial blocks (35 x 40 m). We did this to compare the leaf traits of a site where the invader is not abundant and that is still subject to the same disturbance regime of light grazing. Because we only have one large restoration experiment and on native reference our experimental design is pseudo-replicated
The abundances of all plant species were recorded within each of the 80 plots in the central 9m2 section using the point-intercept method (modified from, Everson 1987). A 4 mm dowel was placed vertically on set points along a grid of 100 points. Relative abundance was measured by counting the number of leaves, stems and inflorescences of each species that touched the dowel at each point along the grid. Six soil samples (core radius of 5 cm, 10 cm deep) were collected from each plot at the same time as the botanical surveys. Available soil nitrate and phosphate levels were analysed with colorimetric methods using a SEAL AQ2 (Rayment & Higginson 1992)
We collected five leaves from three individuals for all species recorded in each plot using the standardised protocols detailed by Cornelissen et al. (2003). All leaf samples were collected during the same growing season. The leaves collected from each species within a plot were combined (leaves of most species were small) to be weighed immediately. Leaf samples were then dried in an oven for 72 hours at 60ºC and re-weighed.