Drivers of Acacia and Eucalyptus growth rate differ in strength and direction in restoration plantings across Australia.

Functional traits are proxies for a species' ecology and physiology and are often correlated with plant vital rates. As such they have the potential to guide species selection for restoration projects. However, predictive trait-based models often only explain a small proportion of plant performance, suggesting that commonly measured traits do not capture all important ecological differences between species. Some residual variation in vital rates may be evolutionarily conserved and captured using taxonomic groupings alongside common functional traits. We tested this hypothesis using growth rate data for 17,299 trees and shrubs from 80 species of Eucalyptus and 43 species of Acacia, two hyper-diverse and co-occurring genera, collected from 497 neighborhood plots in 137 Australian mixed-species revegetation plantings. We modeled relative growth rates of individual plants as a function of environmental conditions, species-mean functional traits, and neighbor density and diversity, across a moisture availability gradient. We then assessed whether the strength and direction of these relationships differed between the two genera. We found that the inclusion of genus-specific relationships offered a significant but modest improvement to model fit (1.6%-1.7% greater R than simpler models). More importantly, almost all correlates of growth rate differed between Eucalyptus and Acacia in strength, direction, or how they changed along the moisture gradient. These differences mapped onto physiological differences between the genera that were not captured solely by measured functional traits. Our findings suggest taxonomic groupings can capture or mediate variation in plant performance missed by common functional traits. The inclusion of taxonomy can provide a more nuanced understanding of how functional traits interact with abiotic and biotic conditions to drive plant performance, which may be important for constructing trait-based frameworks to improve restoration outcomes.