The phylogenetic position, historical phylogeography, and population genetics of foxtail pine (Pinus balfouriana Grev. and Balf.)

The origin and maintenance of genetic variation in natural populations is a function of multiple processes acting across several levels of biological organization. Using a multidisciplinary approach, I develop and test a phylogeographic explanation for patterns of genetic variation found at four chloroplast, four mitochondrial, and 10 nuclear gene regions for foxtail pine ( Pinus balfouriana Grev. & Balf.). This species is a haploxylon, five needle pine placed within subsection Balfourianae. The distribution of this California endemic pine is characterized by a 500 km range disjunction separating populations located in the Klamath Mountains from those located in the southern Sierra Nevada. Paleofloristic data suggest that the Holocene Xerotherm approximately 10,000 years ago produced this range disjunction. Phylogenetic analyses indicate that the divergence of subsection Balfourianae occurred in the late Cretaceous or early Tertiary. The distribution of Miocene fossils suggests that the ancestral foxtail-bristlecone pine complex intermingled with the members of the Arcto-Tertiary Geoflora to form conifer forests of unparalleled diversity. Similarly structured forests are currently found in the Klamath Mountains of northern California. Using measures of stand structure and demographic data, I show that the persistence and population dynamics of extant populations of foxtail pine in the Klamath Mountains is a function of habitat heterogeneity at multiple spatial scales. Further work with coalescent-based population genetic models illustrates patterns of population subdivision that are consistent with long-term isolation (Pleistocene), population bottlenecks, and source-sink dynamics. Divergence times among populations in the Klamath Mountains (150,000-350,000 years ago) pre-dated those observed in the southern Sierra Nevada (25,000-30,000 years ago). Populations in both regions experienced glaciation-induced bottlenecks followed by limited levels of gene flow that differed between dispersal vectors (i.e., seed versus pollen). At the regional level, the divergence time between populations located in the Klamath Mountains and those located in the southern Sierra Nevada was correlated with the Sherwin glacial maximum (∼1,000,000 years ago). Moreover, those regional populations are connected only through rare, long-distance pollen dispersal. A phylogeographic explanation is generated from the aforementioned results, subsequently tested using parametric bootstrapping, and compared to literature-based phylogeographic hypotheses.