Phylogenetic Divergence and Ecophysiological Variation in the Disjunct Kalmia buxifolia (Sand-myrtle, Ericaceae)

Abstract

Abstract—Kalmia buxifolia (sand-myrtle, Ericaceae) is disjunctly distributed across the high-elevation rock outcrops of the southern Appalachians, upper monadnocks and pine savannas of the Carolina Piedmont and Coastal Plain, and the New Jersey Pine Barrens. Here, we sampled plants from each region and reconstructed the phylogeographic history of K. buxifolia to test a rock-outcrop Pleistocene refugium hypothesis, estimate the potential direction(s) and timing of migration, and date divergence from its alpine sister species, K. procumbens. We also assess whether isolation in these different environments has led to variation in intrinsic water-use efficiency. Dating analysis challenges the current hypothesis that rock-outcrop species are relics of Pleistocene refugia (< 18,000 ybp), placing the divergence of K. buxifolia and K. procumbens much earlier, in the late-Miocene (9.40 Ma). Chloroplast haplotype analysis indicates four potential refugial sites, with the most ancient on Mount LeConte in the Great Smoky Mountains, and point to an Appalachian corridor as the likely Pine Barrens colonization route. The sister species divergence time and population level divergences within K. buxifolia generally coincide with major climatic shifts from the late-Miocene to mid-Pleistocene. Results from carbon isotope discrimination indicate that plant water-use varies geographically within K. buxifolia, as does leaf morphology, although it is unclear whether this variation is due to genetic adaptation or phenotypic plasticity. These patterns of phylogenetic divergence and resulting ecophysiological diversity within K. buxifolia are significant for clarifying long-held questions about the biogeographic history and trait differentiation within this species. Further, our results suggest that high-elevation rock outcrop communities may have been inhabitated by northern-affinity species for much longer than previously assumed, and that subsequent population disjunction and isolation may have resulted in ecophysiological differentiation in these communities.