Phylogeography of the rare, perennial, Californian tarweed Carlquistia [Raillardiopsis] muirii was studied to assess whether range disjunctions in the southern Sierra Nevada and between the Sierra Nevada and South Coast Ranges (Ventana Double Cone, Monterey Co.) are best explained by long-distance dispersal or vicariance. Occurrence of C. muirii, a granite endemic, on Ventana Double Cone in the Santa Lucia Range, ca. 240 km from the nearest Sierran population, is intriguing in part because granitic exposures are extremely limited between the southern Sierra/Tehachapi region and the northern Santa Lucia Range/Gabilan Range and unexposed granitics are covered by sediments of at least Miocene age. Vicariance could explain most of the Sierra/Santa Lucia disjunction, even without assuming past occurrence of C. muirii off of granitics, if the species existed ca. 5 Ma, when the Santa Lucia/Gabilan region was juxtaposed with the Sierra/Tehachapi region. Any populations then extending from the Sierra/Tehachapi onto Santa Lucia/Gabilan granitics could have been transported in situ 240 km northwesterly by tectonic slippage of the Salinian Block along the San Andreas Fault. Phylogenetic analysis of 18S-26S rDNA ETS and ITS variation from populations throughout the distribution of C. muirii yielded a robust, well-resolved tree that is congruent with the area cladogram for the southern Sierra Nevada and South Coast Ranges and with the hypothesis that vicariance explains distributional patterns in C. muirii. However, a calibrated, rate-constant rDNA tree for perennials of the "Madia lineage" yields an estimated time of divergence between Sierran and Santa Lucian populations of only ca. 1 Ma, in conflict with a vicariance hypothesis. Patterns of ancient, long-distance dispersal mimicking vicariance may explain at least part of the fascinating phylogeography of C. muirii.

Key words: Carlquistia muirii, Compositae, dispersal, phylogeography, vicariance