Print PageFunctional traits are valuable for understanding how plants respond to their environment. Some traits may be evolutionarily conserved, such as wood density (WD), while others may be more plastic, such as specific leaf area (SLA). Many traits may be correlated with each other, and thus exhibit coordinated responses. Given that both WD and SLA represent structural traits, one might expect the two traits to be correlated across species. Conversely, given that WD is associated with long-lived tissues, while SLA is associated with short-lived tissues, one may expect these two traits to be uncorrelated due to their different “functional time-scales.” To evaluate WD-SLA correlations, we classified species into different resource tolerance, wood type, and leaf habit categories to determine if the relationship between WD and SLA differed between these groupings. We obtained species-specific estimates of these traits for 305 US tree species by implementing a hierarchical Bayesian meta-analysis of literature-derived WD and SLA data. This allowed us to estimate trait values for data-poor species, to partition the sources of variation affecting both traits, and to evaluate potential correlations between these two traits.
Results/Conclusions
Across all 305 species, there was no correlation between WD and SLA. Using shade-tolerance to group species, the highest correlation (r = 0.31) occurred for shade-intolerant species, whereas the lowest (r = 0.08) occurred for intermediate shade-tolerant species. Species-specific resource tolerances, wood type, and leaf habit explained 35% and 24% of the variation in WD and SLA, respectively. While the predicted species-specific WD and SLA estimates from this model were correlated (r = 0.50), reflecting potential constraints imposed by the categorical covariates, the residual errors in WD and SLA, after having accounted for the effects of the covariates, were uncorrelated (r = -0.07). Since SLA and WD have contrasting evolutionary strategies, absence of correlation between the two traits is not surprising. In support of this, we found that, for SLA, the study-to-study variability was up to 67 times greater than the variation explained by different taxonomic levels. Conversely, taxonomic identity explained about twice the variation in WD compared to the study effects. Thus, SLA appears to be strongly influenced by environmental heterogeneity (plastic), and WD exhibits a strong evolutionary signal. Our study highlights the importance of considering evolutionary and environmental influences for predicting plant functional traits and their coordinated behavior.
