LAUSR

Premise of research. Phylogenetic comparative methods provide a powerful approach for exploring the macroevolution of plant functional traits. Such approaches can uncover trait-trait correlations through evolutionary time as well as provide evidence of the role of traits in adaptation across environmental gradients. For continuous traits, most phylogenetic comparative approaches to date employ a single trait value per species—often a mean of sampled individuals—or alternatively incorporate intraspecific variation as a distribution around such a mean. It has been known for quite some time that many of the most physiologically and ecologically important plant traits are actually highly plastic, changing dynamically across a growing season, with whole-plant development or in response to environmental conditions. Here we demonstrate one possible approach to assessing the evolution of such dynamic traits: the use of function-valued phylogenetic comparative methods. Methodology. Leaf traits were sampled across 25 taxa in the genus Cornus at six time points throughout the growing season in a common-garden context, followed by contrasting sets of alternative analyses to demonstrate the consequences of researcher decisions on study conclusions. Pivotal results. The vast majority of assessed traits exhibit substantial seasonal shifts. These shifts cause traditional macroevolutionary correlations assessed at different sampling dates to yield conflicting results. Function-valued approaches indicate that seasonal shifts in many traits are evolutionarily correlated, with implications for the origin of trait-trait trade-offs. Seasonal trait plasticity is also evolutionarily correlated with native habitat environmental gradients across Cornus. Conclusions. Because a very large number of plant functional traits are not fixed but vary dynamically over time or with environmental conditions, stronger insights into the evolution of plant functional traits can emerge when this dynamism is explicitly incorporated into phylogenetic comparative approaches. We encourage the adoption of such approaches, as well as the development of better tools for doing so.