Gene flow between populations is one of the primary mechanisms of evolution. In plants, it can occur either by the dispersal and establishment of seeds into a local population from outside or by the dispersal of pollen and successful reproduction. Either of these sources of genetic variation must then be followed by successful survival and reproduction of the immigrant or offspring so that the introduced genetic variation continues to contribute to the local population over time. This time-integration of the genetic contribution constitutes the realized gene flow between populations or lineages. The rate of gene flow and constraints on it are crucially important to population genetic structure and phylogenetic lineage divergence as well as patterns of local adaptation and genetic variation.

In this dissertation I study gene flow at three evolutionary scales in the western North American clade of white oaks (genus Quercus, section Quercus s.s., series Dumosae).

First, in chapter one, I use two established common garden plantings of blue oak (Q. douglasii), together with surveys of the provenance field populations that provided acorns for the gardens, to investigate gene flow and local adaptation among populations of a single species. I show that there are both environmental and genetic components to variation in spring phenological timing among these blue oaks (as well as a very small genotype×environment effect). There are significant differences in phenological timing associated with the different provenance sites even for trees growing in a common garden environment, reflecting a genetic component of their phenological variation. This genetic variation is correlated with the climate experienced by trees at the provenance sites. In particular, I identify notable influences of spring maximum temperature and fall and spring precipitation on genetic variation in spring phenology. Additional genetic variation, at the individual level, may be reflected by the phenological variation observed among trees from the same provenance. This individual variation is high relative to the variation that can be associated with provenance sites, suggesting that even though there may be local adaptation for this trait, there is also a great deal of genetic variation within populations. This may be due to high rates of gene flow among populations of blue oak, balanced by a moderate effect of selection on local variation, but may also be due to temporally fluctuating selection within populations together with a more moderate rate of gene flow.

Next, in the second chapter, I focus on gene flow between pairs of oak species occurring within hybrid zones where their ranges overlap. Blue oak is again the central species, hybridizing in the northern part of its range with Oregon white oak (Q. garryana var. garryana) and in the southern part of its range with Tucker's scrub oak (Q. john-tuckeri). My research reveals the evolutionary and biogeographic contexts of these two hybrid zones by measuring and partitioning landscape-scale barriers to gene flow within them. I explain pairwise genetic dissimilarity between individuals as a function of their geographic separation (isolation-by-distance), environmental difference (isolation-by-environment), and phenological asynchrony (isolation-by-time). Even though it is commonly considered a basic control on genetic structure, I do not find evidence for isolation-by-distance in either of the hybrid zones, nor in a third geographic data set consisting of only blue oaks. Instead, I find that genetic dissimilarity can be partially explained as isolation-by-environment, both across the hybrid zones and within blue oaks alone. This signal is especially strongly associated with winter temperatures, and to a lesser extent with summer high temperature and aridity. In addition, in the southern hybrid zone only, between blue oak and Tucker's scrub oak, there is a strong signal of isolation-by-time. Using variation partitioning models to separate the effects of these three isolating factors into their independent and overlapping contributions, I suggest the differences in flowering phenology that contribute to genetic structure in the southern hybrid zone result from both environmental differences and genetic differences. This is much like the overall influences on phenology identified in chapter one, but in this case they represent extrinsic and intrinsic reproductive isolating mechanisms, respectively. This pattern is found only in the southern hybrid zone and not in the northern hybrid zone, which may be a reflection of the closer phylogenetic relationships between the hybridizing species in the southern zone and their biogeographic histories.

In chapter three, I place gene flow in its full phylogenetic context within this clade of oaks. I identify two evolutionary scales of gene flow: contemporary hybridization, as already discussed in chapter two, and ancient hybridization. I use methods drawn from both population genetics and phylogenetics to identify individual samples that show signs of contemporary hybridization. Removing these samples from the tips of a maximum likelihood phylogenetic tree dramatically improves resolution of groups within the clade. Using this well-resolved tree, I then turn to phylogenetic invariants methods (D-statistics, ABBA-BABA) to investigate patterns of gene flow deeper within the phylogeny. This reveals a particularly strong signal of hybridization involving the common ancestor of California scrub oak (Q. berberidifolia) and leather oak (Q. durata) and either the common ancestor of the entire southern California scrub oak clade (including Q. douglasii) or the Quercus garryana clade. There are also potentially signals of additional episodes of hybridization at medium depth within the phylogeny. The implications of hybridization at multiple depths in the phylogeny and lasting impacts of these periods are discussed in relation to evolutionary models of oaks as a whole.