Intraspecific variability and reaction norms of forest understorey plant species traits

1. Trait-based models of ecological communities typically assume intraspecific variation in functional traits is not important, although such variation can change species trait rankings along gradients in resources and environmental conditions, and thus influence community structure and function. 2. We examined the degree of intraspecific relative to interspecific variation, and reaction norms of 11 functional traits for 57 forest understorey plant species, including: intrinsic water-use efficiency (iWUE), Delta N-15, five leaf traits, two stem traits and two root traits along gradients in light, nitrogen, moisture and understorey cover. 3. Our results indicate that interspecific trait variation exceeded intraspecific variation by at least 50% for most, but not all traits. Intraspecific variation in Delta N-15, iWUE, leaf nitrogen content (LNC) and root traits was high (47-70%) compared with most leaf traits and stem traits (13-38%). 4. Delta N-15 varied primarily along gradients in abiotic conditions, while light and understorey cover were relatively less important. Intrinsic water-use efficiency was related primarily to light transmission, reflecting increases in photosynthesis relative to stomatal conductance. Leaf traits varied mainly as a function of light availability, with some reaction norms depending on understorey cover. Plant height increased with understorey cover, while stem-specific density was related primarily to light. Resources, environmental conditions and understorey cover did not contribute strongly to the observed variation in root traits. 5. Gradients in resources, environmental conditions and competition all appear to control intraspecific variability in most traits to some extent. However, our results suggest that species cross-over (i.e. trait rank reversals) along the gradients measured here are generally not a concern. 6. Intraspecific variability in understorey plant species traits can be considerable. However, trait data collected under a narrow range of environmental conditions appears sufficient to establish species rankings and scale between community and ecosystem levels using trait-based models. Investigators may therefore focus on obtaining a sufficient sample size within a single set of conditions rather than characterizing trait variation across entire gradients to optimize sampling efforts.