Leaf shape and size variation in bur oaks: an empirical study and simulation of sampling strategies

Premise: Leaf shape and size figure strongly in plants' adaptation to their environments. Among trees, oaks are notoriously variable in leaf morphology. Our study examines the degree to which within-tree, among-tree, and among-site variation contribute to latitudinal variation in leaf shape and size of bur oak (Quercus macrocarpa: Fagaceae), one of North America's most geographically widespread oak species.

Methods: Samples were collected from four sites each at northern, central, and southern latitudes of the bur oak range. Ten leaf size traits were measured, and variance in these traits and eight ratios based on these traits was partitioned into tree and population components. Population means were regressed on latitude. We then parameterized a series of leaf collection simulations using empirical covariance among leaves on trees and trees at sites. We used the simulations to assess the efficiency of different collecting strategies for estimating among-population differences in leaf shape and size.

Results: Leaf size was highly responsive to latitude. Site contributed more than tree to total variation in leaf shape and size. Simulations suggest that power to detect among-site variance in leaf shape and size increases with either more leaves per tree (10-11 leaves from each of 5 trees) or more trees per site (5 leaves from each of 10+ trees).

Conclusions: Our study demonstrates the utility of simulating sampling and controlling for variance in sampling for leaf morphology, whether the questions being addressed are ecological, evolutionary, or taxonomic. Simulation code is provided as an R package (traitsPopSim) to help researchers plan morphological sampling strategies.