Figure 1.
Relationships between ant species richness and energy and climate variables within the FB and SKB.
(A) productive energy, (B) mean annual precipitation, (C) mean temperature. Triangles and circles refer to data from FB (n = 29) and SKB (n = 5) sites respectively. NDVI is taken as a surrogate of productive energy. All measures are averaged over several years to account for natural variability (see Appendix S3 for methods). Ant species richness was ln(x+1)-transformed for all analyses and regression line and confidence intervals were backtransformed. The regression line and 95% confidence intervals for NDVI are shown. The relationships between ant species richness and precipitation or temperature were not significant. Regression for NDVI: ln(species richness+1) = 2.4325983+0.00035032780 * NDVI−0.000000035595287 * NDVI2, r2 = 0.22, F2,31 = 4.50, P = 0.0192.
Figure 2.
Comparison of ant species richness in the FB and SKB with that of sites worldwide.
The relationships between global ant species richness and (A) productive energy, (B) mean annual precipitation, and (C) mean temperature are shown with the data from the FB and SKB superimposed. NDVI is used as a surrogate of productive energy. Clear circles represent global data extracted from the literature (n = 331) while triangles and filled circles refer to our own data from FB (n = 29) and SKB (n = 5) sites respectively. Ant species richness was ln(x+1)-transformed for all analyses. The backtransformed regression lines and 95% confidence intervals for precipitation and temperature are shown. The confidence interval for high values of precipitation is very large and thus not fully shown. The regression for NDVI was not significant for the global dataset. Regressions are: mean annual precipitation: ln(species richness+1) = 2.9100926−0.00056787062 * precipitation+0.00000038199406 * precipitation2, r2 = 0.86, F2,328 = 15.51, P<0.0001; temperature: ln(species richness+1) = 0.48074860+0.22574050 * temperature−0.0046141963 * temperature2, r2 = 0.30, F2,328 = 68.55, P<0.0001.
Table 1.
General linear model (type 3) showing the effects of hemisphere and latitudea on ant species richnessb (Akaike weight: 6.77 * 10−13c).
Table 2.
General linear model (type 3) showing the effects of hemisphere, habitat typea, and temperature on ant species richnessb (Akaike weight: 0.656241c).
Figure 3.
Comparison of ant species richness in the FB and SKB with that of similar habitats.
Data for other shrublands and arid shrublands were extracted from the literature and comprise a variety of different vegetation types from desert shrublands to thickets. The other southern hemisphere shrublands and arid shrublands are other South African vegetation types, such as Nama-karoo, and Australian and South American shrublands and arid shrublands. Error bars show 95% confidence intervals. The confidence interval for other southern hemisphere arid shrublands is wide and thus not shown fully, as only the fact that the SKB falls within it is of interest. FB Fynbos biome (n = 29), SKB Succulent karoo biome (5), SS other Southern hemisphere shrublands (6), SAS other Southern hemisphere arid shrublands (3), NS Northern hemisphere shrublands (33), NAS Northern hemisphere arid shrublands (16).