Adaptive evolution requires both raw genetic material and an accessible path of high fitness from one fitness peak to another. In this study, we used an introgression line (IL) population to map quantitative trait loci (QTL) for leaf traits thought to be associated with adaptation to precipitation in wild tomatoes (Solanum sect. Lycopersicon; Solanaceae). A QTL sign test showed that several traits likely evolved under directional natural selection. Leaf traits correlated across species do not share a common genetic basis, consistent with a scenario in which selection maintains trait covariation unconstrained by pleiotropy or linkage disequilibrium. Two large effect QTL for stomatal distribution colocalized with key genes in the stomatal development pathway, suggesting promising candidates for the molecular bases of adaptation in these species. Furthermore, macroevolutionary transitions between vastly different stomatal distributions may not be constrained when such large-effect mutations are available. Finally, genetic correlations between stomatal traits measured in this study and data on carbon isotope discrimination from the same ILs support a functional hypothesis that the distribution of stomata affects the resistance to CO 2 diffusion inside the leaf, a trait implicated in climatic adaptation in wild tomatoes. Along with evidence from previous comparative and experimental studies, this analysis indicates that leaf traits are an important component of climatic niche adaptation in wild tomatoes and demonstrates that some trait transitions between species could have involved few, large-effect genetic changes, allowing rapid responses to new environmental conditions. T HE course of phenotypic evolution depends upon both the genetic basis of functionally important traits and the mechanistic relationship between traits. All else being equal, traits governed by mutations of large phenotypic effect with low pleiotropy can respond to selection more readily than traits that are complex, highly pleiotropic, and/or strongly correlated with other traits. Hence, the genetic basis of trait differences and covariation influences which phenotypic axes are most likely to respond to selection in new environmental conditions; understanding this underlying architecture therefore provides insight into the genetic mechanisms constraining or facilitating phenotypic evolution (Lee et al. 2014). It can also indicate which evolutionary forces are responsible for trait differences within and between species. When adaptation to a complex ecological niche involves many traits, for example, genetic analysis can determine whether correlations between traits are caused by a shared genetic basis (pleiotropy) or whether natural selection favors trait combinations because they function well together (coselection). The magnitude and direction of allelic effects between phenotypically divergent genotypes can also be used to infer whether this divergence is consistent with directional natural selection (Orr 1998b;Rieseberg et al. 2002)...