Generality in ecology: testing North American grassland rules in South African savannas

Ecology has emerged as a global science, and there is a pressing need to identify ecological rules-general principles that will improve its predictive capability for scientists and its usefulness for managers and policy makers. Ideally, the generality and limits of these ecological rules should be assessed using extensive, coordinated experiments that ensure consistency in design and comparability of data. To improve the design of these large-scale efforts, existing data should be used to test prospective ecological rules and to identify their limits and contingencies. As an example of this approach, we describe prospective rules for grassland responses to fire and rainfall gradients, identified from long-term studies of North American grasslands and tested with existing data from long-term experiments in South African savanna grasslands. Analyses indicated consistent effects of fire on the abundance of the dominant (grasses) and subdominant (forbs) flora on both continents, but no common response of grass or forb abundance across a rainfall gradient. Such analyses can inform future research designs to refine and more explicitly test ecological rules.