Aim The East Asia endemic Taiwania cryptomerioides Hayata is an iconic and relictual monotypic conifer whose main extant populations are now restricted to the Yunnan-Myanmar border, northern Vietnam and Taiwan. It has also been reported from several localities in Guizhou, Hubei and Fujian Provinces, China. Its fossil record indicates that, while it was more widely distributed in the Northern Hemisphere and grew under a range of different ecological conditions, it has remained almost unchanged in its morphology for over 100 Myr. We investigate whether these remaining extant, disjunct populations have diverged genetically; when such a divergence may have occurred; and which, if any, of the extant populations exhibit refugial characteristics.Location East Asia.Methods Sequences of five chloroplast DNA markers (petG-trnP, trnH-psbA, trnV-trnM, trnC-ycf6 and trnL-trnF) from all extant populations of T. cryptomerioides were analysed to reveal their phylogeography. Molecular clock models with fossil calibrations were used to estimate divergence times between extant populations.Results Extremely low nucleotide diversity was found in the overall population (p = 0.00077) with only nine haplotypes distinguished. The mainland Asia populations share one major ancestral haplotype. The insular populations in Taiwan all possess a unique haplotype with at least an eight-mutational-step difference to the mainland Asia haplotype. Molecular clock estimations demonstrated that the mean divergence time between the predominant insular population haplotype and the mainland Asia haplotype occurred at c. 3.23-3.41 Ma, followed by a split into Vietnamese and Yunnan-Myanmar populations (c. 1.0-1.39 Ma).Main conclusions Strong genetic differentiation exists between insular (Taiwan) and mainland Asia populations. The split between insular and mainland haplotypes can be dated back to the end of the Pliocene. The Yunnan-Myanmar border area, northern Vietnam and Taiwan are identified here as potential refugia for T. cryptomerioides. Other populations in mainland China are unlikely to be the result of historical fragmentation and their origins require further investigation.

No abstract

Phylogeographic studies have revealed spatial genetic structure and inferred geographical processes that may have generated genetic diversity and divergence. These study results have implications not only on the processes that generate intraspecific and interspecific diversity but also on the essential integrals for defining evolutionary entities (e.g., species). However, the resulting phylogeographic inferences might be impacted by the sampling design, i.e., the number of individuals per population and the number of geographic populations studied. The effect of sampling bias on phylogeographic inferences remains poorly explored. With a comprehensive sampling design (including 186 samples from 56 localities), we studied the phylogeographic history of a Taiwanese endemic damselfly, Psolodesmus mandarinus, with a specific focus on testing the impact of the sampling design on phylogeographic inference. We found a significant difference in the genetic structure of eastern and western populations separated by the Central Mountain Range (CMR) of Taiwan. However, isolation by the CMR did not lead to reciprocally monophyletic geographic populations. We further showed that, when only a subset of individuals was randomly included in the study, monophyletic geographic populations were obtained. Furthermore, historical demographic expansion could become undetectable when only a subset of samples was used in the analyses. Our results demonstrate the impact of sampling design on phylogeographic inferences. Future studies need to be cautious when inferring the effect of isolation by a physical barrier.