We are studying Dirca palustris L. to address the paucity of data on the phylogeography of shrubs in forests of eastern North America. An interesting case study, D. palustris is distributed broadly but is relatively uncommon; is without obvious vectors for long-distance dispersal of seeds; and has an obscure historical distribution because, as an insect-pollinated shrub, the taxon is absent from the pollen record of the late Pleistocene and Holocene. Concern about how diverse taxa of plants respond to changing climates, the scarcity of information about postglacial range expansion of insect-pollinated shrubs, and the paradox that contemporary distributions of many taxa imply rates of migration greater than expected based on present-day observations (Reid’s paradox of plant migration) motivated us to conduct this work. Our goals were to survey the diversity of the species, elucidate the locations of glacial refugia, and identify routes of migration and barriers to dispersal. We sequenced the psbD-trnT(GGU), trnL-trnF, and rpl32-trnL(UAG) noncoding chloroplast regions of three plants from each of 92 populations across the range of D. palustris, as well from populations of three other species of Dirca. A concatenated sequence of 3236 bases was attained, from which haplotype diversity was assessed.
Results/Conclusions
Twenty-four haplotypes were identified, including private haplotypes in 14 populations. Putative regions of glacial refugia include the Cumberland Plateau (Kentucky/Tennessee), the Ozark Mountains (Arkansas/Missouri), the southern Atlantic Seaboard (Carolinas), and the Gulf Coast (Mississippi/Florida). These regions have high diversity of haplotypes and/or multiple private haplotypes. Haplotypes at northern latitudes appear derived from northern refugia (e.g., Cumberland Plateau or Ozarks), and two haplotypes farther south have not dispersed northward. The Atlantic Seaboard and Northeast have a distinct collection of haplotypes, suggestive of dispersal limitation across the Appalachian Mountains. The impact of the Mississippi River is less clear, although one common haplotype is largely restricted to the west of this river. Several phenotypically distinct populations in the Southeast are united by a chloroplast mutation, indicating they are of a common maternal lineage. Hypotheses will be tested formally by using coalescent simulations with sequences from a total of five gene regions, and we will compare our results to those obtained from ecological niche modeling based on distribution data. Our results should inform decision-making in conservation biology and provide additional information for researchers pursuing modeling of the influence of climate change on plant distributions, with additional insights addressing Reid’s paradox of plant migration.