Resprouter fraction in Cape Restionaceae assemblages varies with climate and soil type

1. While fire-induced changes in biodiversity are well documented, less is known about how fire impacts life-history variation and diversity of functional traits that represent distinct strategies for persistence in fire-driven ecosystems. One example is the dichotomy in which 'resprouter' species usually survive fires to produce new growth, while 'reseeder' species perish and re-establish from seed. 2. Variable relative numbers of reseeder and resprouter species in local assemblages of Restionaceae (Poales) of the Cape Floristic Region (CFR), and their high species richness and endemism, make this family representative of the region's functional, taxonomic and ecological diversity. 3. We investigate how the proportion of resprouter species (resprouter fraction) changes among edaphic conditions and along a climate moisture (i.e. aridity) gradient that integrates annual precipitation (range 130-1300 mm) and potential evapotranspiration (range 3-8 mm m(-2) day(-1)). We sampled local assemblages using a stratified-random design that fully spanned the environmental conditions of the CFR. 4. As climate moisture across sites with infertile soils increased, so did resprouter fraction. Resprouters may increasingly dominate reseeders as moisture increases under conditions in which productivity is generally limited. This suggests the explanatory hypothesis that in these conditions, greater moisture accelerates both biomass accumulation (fuel) and fire frequency, providing an advantage to resprouters. However, we found no evidence for increasing resprouter fraction with increasing climate moisture on more fertile soils. 5. This statistical interaction suggests the possibility that moisture affects resprouter fraction dependent on soil fertility, potentially through: (i) impacts on species composition, if soil type acts as an environmental filter or promotes biotic interactions, and/or (ii) impacts on fire dynamics, potentially through fuel load drying rates. Future research should include a comprehensive fire model for the CFR, to enable direct examination of relationships of life-history and functional traits across variation in fire dynamics. Additionally, our results indicate that we need both (i) demographic studies on plant establishment, growth rates and survival, and (ii) field experiments that address how environmental and biotic filters affect species composition across variable fire regimes. These together should facilitate mechanistic interpretation of how fire mediates life-history variation and diversity of functional traits in the CFR