Phylogenetic structure across grazing regimes
Phylogenetic dispersion (mean ± SD (range): 0.44 ± 0.08 (0.24 – 0.60)) varied with year (F3,128 = 3.18 P = 0.03) [plots from 2005 (lsmean ± SE: 0.46 ± 0.01) were more phylogenetically dispersed than plots from 2004 (0.41 ± 0.01)], but was independent of wet/dry (F1,128 = 0.32, P = 0.57), grazing intensity (F2,128 = 0.16, P = 0.85; Figure 4a) and the grazing:year interaction (F6,122 = 1.52, P = 0.18). Controlling for the proportion of species that are graminoids in each plot, the effect of year remained (F2,127 = 3.06, P = 0.03). In contrast to the patterns of unweighted phylogenetic dispersion, the abundance-weighted measure (mean ± SD (range): 0.41 ± 0.26 (0.03 – 0.95)) varied with grazing intensity (F2,128 = 5.72, P = 0.004), and wet/dry (F1,128 = 14.25, P = 0.0002) [wet plots (lsmean ± SE: 0.33 ± 0.04) were more clumped than dry plots (lsmean ± SE: 0.48 ± 0.04)], though not with year (F3,128 = 0.91, P = 0.44) or the grazing:year interaction (F6,122 = 0.32, P = 0.93). After controlling for the proportion of the relative abundance of graminoids, the effect of grazing intensity on abundance-weighted phylogenetic dispersion remained (F2,127 = 9.92, P < 0.0001). Importantly, plots with pulse grazing were more phylogenetically dispersed than plots with both light (P = 0.007) and press grazing (P = 0.02), while there was no difference in dispersion between plots with light and press grazing (P = 0.91; Figure 4b).