Molecular and morphological support for a Florida origin of the Cuban oak


Paul F. Gugger, Department of Ecology and Evolutionary Biology, University of California, 4145 Terasaki Life Sciences Building, 610 Charles E. Young Drive E, Los Angeles, CA 90095-7239, USA.


Aim  The origins of the Cuban biota are of long-standing interest in biogeography, and the source of a small live oak (Quercus series Virentes) population on Cuba remains unresolved. Based on morphological evidence, previous authors have hypothesized a Florida origin from either Q. geminata or Q. virginiana or both; a Mexican origin from Q. oleoides; or a hybrid origin from both sources. We use molecular data and taxonomically informative leaf morphology to identify the source species and timing of colonization.

Location  Cuba, Central America, Mexico and the south-eastern United States.

Methods  We collected representative samples of Cuban oaks and each putative source species and genotyped each sample at 12 nuclear microsatellites and two chloroplast DNA sequences. We estimated population structure using a Bayesian clustering analysis and F-statistics, pairwise migration rates among taxa, and divergence time using an isolation-with-migration model. We measured seven leaf traits and conducted an analysis of similarity (ANOSIM) to determine which putative source species was most similar to Cuban oaks.

Results  Cuban oak contains one chloroplast DNA haplotype, which is common in southern Florida. Bayesian clustering analysis of microsatellites revealed that the Cuban oak forms a distinct and pure population cluster, and F-statistics showed that Cuban oaks are differentiated least from Q. virginiana and most from Q. geminata. Migration rates were highest out of Cuba to Q. oleoides. Molecular diversity, the ratio of allelic richness to allele size range, and effective population size of the Cuban oak were relatively low, suggesting a founder effect. Divergence time estimates fell entirely within the Pleistocene (628–6 ka), considering a range of mutation rates and generation times. Cuban oaks were morphologically most similar to Q. virginiana and least similar to Q. geminata.

Main conclusions  Molecular and morphological data support a Pleistocene dispersal of Q. virginiana from Florida to Cuba, followed by isolation and divergence, then limited dispersal and introgression from Cuba to Q. oleoides in Central America. Birds could have dispersed acorns to Cuba during a glacial period when sea levels were low. These results highlight the varied origin of the Cuban biota and the possible role of Pleistocene glaciations in the establishment of temperate taxa in the tropics.