Savanna ant species richness is maintained along a bioclimatic gradient of increasing latitude and decreasing rainfall in northern Australia

AIM: Using a standardized sampling protocol along a 600‐km transect in northern Australia, we tested whether ant diversity within a single biome, tropical savanna, decreases with increasing latitude (as a surrogate of temperature) and decreasing rainfall, as is expected for biodiversity in general. LOCATION: Northern Australia. METHODS: Ants were sampled using pitfall traps on three occasions at 1‐ha sand, loam and clay sites at each of five locations along the Northern Australian Tropical Transect (NATT), from 12°50′ S (1400 mm mean annual rainfall) to 17°21′ S (650 mm). RESULTS: We recorded a total of 246 species from 37 genera. Mean observed species richness pooled across sampling periods was similar at sand (85.4) and loam (82.2) sites, but was less than half this at clay sites (40.0). Ant communities were also compositionally distinct on clay soils compared with sands and loams. Individual genera showed variable diversity patterns, ranging from a linear increase to a linear decrease in species richness along the NATT. However, total species richness was relatively uniform along the gradient. Patterns of ant species turnover were consistent with previously recognized biogeographical boundaries, with a primary disjunction between the arid and monsoonal zones in the south, and a secondary disjunction between the semi‐arid and mesic zones in the north. MAIN CONCLUSIONS: Patterns of ant diversity in Australian savannas do not conform to global patterns of biodiversity declines with increasing latitude and decreasing rainfall. We believe this is due to a lack of significant temperature change across the latitudinal gradient, and, in particular, to the fauna's evolutionary history in association with aridification, which makes it unusually resilient to increasing aridity. The diversity of other important faunal groups such as termites and lizards is also exceptionally high in arid Australia and is likewise not closely linked to rainfall in Australian savannas. We predict that these taxa are far more sensitive to increasing aridity in savannas elsewhere in the world, and especially in the Neotropics, where savannas have an evolutionary association with humid rain forest rather than desert.