Functional trait patterns: investigating variation-covariation relationships and the importance of intraspecific variability along distinct vegetation types

Plants adjust to abiotic conditions by changing their anatomical, morphological, and physiological traits. Traits can vary independently or in an integrated manner, known as trait variation and covariation respectively. It has been hypothesized that a trade-off would emerge along a gradient of abiotic constraints in which trait variation would be favored under resource-rich conditions while covariation under resource-limited ones. Although many studies have provided empirical support for this trade-off, a consensus has not yet emerged, due to a lack of support in some cases. This study investigated variation and covariation in three leaf and four wood traits of 74 woody species from a rainforest, a semideciduous forest, and a Restinga heath vegetation in the Atlantic Forest, which are subjected to different water-related constraints. We asked: Is there a variation-covariation trade-off within and across vegetation types? How does incorporating intraspecific variability change the magnitude and pattern of trait covariation? We found a variation-covariation trade-off and a positive relationship both within and across vegetation types. Wood variation was higher and covariation was lower in the rainforest, likely due to the greater water availability. Conversely, wood trait covariation was higher and variation was lower in the Restinga and seasonal semideciduous forest. Differences between vegetation types are likely related to the species’ strategies to prevent hydraulic failure, particularly for Restinga species that adjust their wood density and xylem vessel density in a coordinated manner. Accounting for intraspecific trait variability increased covariation across all vegetation types, particularly in the Restinga. This highlights the loss of functional information when analyses are based solely on species’ mean trait values. Our results also contribute to this discussion by providing evidence that the trade-off or a positive relationship between trait variation and covariation may be context-dependent.