Evolution of Edaphic Ecology in Ceanothus (Rhamnaceae)

Edaphic factors---those pertaining to the substrate or soil---are thought to play an important role in the diversification of plants. However, little is known about the evolutionary origin of most edaphically specialized plant species, preventing inference of potential mechanisms by which substrate properties may influence speciation. The North American genus Ceanothus (Rhamnaceae) contains eight edaphic endemics, species that are strictly associated with specific soil types. The three components of my dissertation research aim to improve understanding of how edaphic ecology has evolved during the diversification of Ceanothus. First, I use DNA sequence data from the low-copy nuclear gene nitrate reductase to reconstruct a phylogeny of Ceanothus. This research indicates that diversification of the two Ceanothus subgenera ( Cerastes and Ceanothus) is centered in the California Floristic Province (CFP) of western North America and is characterized by shallow divergence and phylogenetic relationships defined predominantly by geography. Dating exercises indicate that the two subgenera diverged from one another about 12 million years ago (Ma), with subsequent independent diversification starting around 6 Ma. NIA gene trees also revealed major lineages within Ceanothus not suggested by previous work, including a northern CFP clade comprising 15 of the 24 Cerastes species, a southern CFP clade associated with five southern California Cerastes, and a small clade containing members of subgenus Ceanothus that possess thorn-tipped twigs. Edaphic-endemic taxa are not phylogenetically clustered, suggesting that the origin of edaphic endemism has occurred on multiple occasions. Nevertheless, the large northern CFP clade of Cerastes contains five of the edaphic endemics.;The second chapter of my dissertation uses soil chemistry data in combination with a more detailed examination of genetic variation in nitrate reductase to elucidate the evolution of C. roderickii, a specialist on soils derived from an outcrop of gabbro rock in western El Dorado County, California. This research indicates that C. roderickii is very closely related to C. cuneatus, a geographically widespread soil generalist that occurs peripatrically with C. roderickii in western El Dorado County. Though not conclusive, the phylogenetic data are consistent with a progenitor-derivative relationship between C. cuneatus and C. roderickii. Soil chemistry analyses show that C. roderickii is specialized on a nutrient-deficient form of gabbro-derived soil, while peripatric populations of C. cuneatus are on a previously uncharacterized nutrient-rich form of gabbro-derived soil.;The third chapter of my dissertation focuses yet more closely on C. roderickii and C. cuneatus, using biosystematics, experimental growth trials, and amplified fragment length polymorphism (AFLP) data to learn how edaphic conditions may contribute to ecological and genetic differentiation between the species. My research indicates that prezygotic barriers to gene flow between the species are limited, and that the species are divergently specialized to their substrates. AFLP data confirm a close relationship between the species, but also demonstrate pervasive population genetic differentiation, even in areas of very close peripatry; putative admixture between the species is rare and restricted mostly to populations of seeds rather than adults. Paired with evidence for decreased fitness of hybrids on either parental substrate, the data on genetic admixture suggest that genetic and ecological differentiation of the species may be driven, at least in part, by selection against hybrid progeny in either parental environment.