Figure 2.

(a) Proportional change in species richness between 2003/2004 and 2007/2008 was compared using Mann–Whitney–Wilcoxon comparison of means (one-tailed), for the 12 mistletoe-free sites (grey), 11 control sites (black) and 17 treatment sites (white). Treatment sites lost significantly more species than control sites, in terms of both total species (p = 0.004) and woodland-dependent species (p = 0.002), lending support to prediction 4. (b) The effect of mistletoe removal on sensitivity to drought was tested by comparing the proportional change in residents after mistletoe removal. Treatment sites lost significantly more residents than control sites (p = 0.008 for total residents; p = 0.002 for woodland-dependent residents), providing support for prediction 5.