A primary goal of evolutionary genetics is to discover and explain the genetic basis of fitness-related traits and how this genetic basis evolves within natural populations. Unprecedented technological advances have fueled the discovery of genetic variants associated with ecologically relevant phenotypes in many different life forms, as well as the ability to scan genomes for deviations from selectively neutral models of evolution. Theoretically, the degree of overlap between lists of genomic regions identified using each approach is related to the genetic architecture of fitness-related traits and the strength and type of natural selection molding variation at these traits within natural populations. Here we address for the first time in a plant the degree of overlap between these lists, using patterns of nucleotide diversity and divergence for .7000 unique amplicons described from the extensive expressed sequence tag libraries generated for loblolly pine (Pinus taeda L.) in combination with the .1000 published genetic associations. We show that loci associated with phenotypic traits are distinct with regard to neutral expectations. Phenotypes measured at the whole plant level (e.g., disease resistance) exhibit an approximately twofold increase in the proportion of adaptive nonsynonymous substitutions over the genome-wide average. As expected for polygenic traits, these signals were apparent only when loci were considered at the level of functional sets. The ramifications of this result are discussed in light of the continued efforts to dissect the genetic basis of quantitative traits.A primary goal of population and quantitative genetics is to understand the genetic architecture of ecologically relevant traits (Stinchcombe and Hoekstra 2008; Barrett and Hoekstra 2011;Neale and Kremer 2011). A primary step on the path to this goal is to link genetic with phenotypic variation, either through linkage mapping of quantitative trait loci using pedigrees or through linkage disequilibrium mapping in natural populations (Lander and Schork 1994), with the latter currently being the most utilized. A multitude of studies ranging across a diverse set of taxa have discovered myriad genotype-phenotype correlations (Hindorff et al. 2009;Ingvarsson and Street 2011;Neale and Kremer 2011). Each discovered variant, however, often explains only a small fraction of the heritable phenotypic variance, thus being consistent with a polygenic model for the genetic architecture of complex traits (Lynch and Walsh 1998). Concomitant with the discovery of these associations are population genomic scans documenting deviations from expectations derived using the neutral theory (Nielsen 2005). Such scans have also discovered large numbers of loci putatively underlying phenotypic traits in many different taxa (e.g., Pollinger et al. 2005;Pritchard et al. 2010;Hufford et al. 2012), but in this case the link between genotype and phenotype is not explicit. A natural question thus arises about how much overlap exists between the lists gen...