GENETIC VARIATION AND DIFFERENTIATION AMONG SOUTHERN RED CEDAR (JUNIPERUS VIRGINIANA VAR. SILICICOLA) POPULATIONS ACROSS ITS GEOGRAPHICAL RANGE

Premise of research. Juniperus virginiana is the most widespread native conifer in the eastern United States. There are two disjoint varieties of J. virginiana, the eastern red cedar (var. virginiana) and the southern red cedar (var. silicicola). While the eastern red cedar occurs in the eastern United States, the southern red cedar primarily occurs along the coast marshes and swamps from North Carolina south to Florida and west to southeast Texas. Methodology. We studied the genetic variation and differentiation of the southern red cedar in Florida, Georgia, and North Carolina using eight microsatellite markers. Pivotal results. The genetic diversity of the southern red cedar was similar to that of other species of Juniperus. The mean number of alleles per locus was highest in the west coast region of Florida and the lowest in North Carolina. Analysis of molecular variance and Wright's F coefficients revealed significant differentiation among regions. Principal coordinate analysis and phylogenetic analysis revealed three distinct groups among the six regions. Mantel test showed significant correlations between geographic distance between regions and their genetic differentiation, suggesting that limited gene flow determines the genetic structure. In addition, a population tree suggested that coastal regions are central regions while North Carolina and inland Florida are marginal. Our findings also showed that a Bayesian model best describes the genetic structure of a southern red cedar with four clusters (K = 4). The model supports isolation by distance, but our results also indicate that environmental factors explain differences observed in inland regions. Conclusions. We found significant genetic diversity and differentiation among southern red cedar regions. Genetic differentiation can be explained, at least in part, by limitation in gene flow because of isolation by distance and isolation by environment. Our results support the predictions of the central-marginal model.