We have investigated the impact of recognized biogeographic barriers on genetic differentiation of grey box (Eucalyptus moluccana), a common and widespread tree species of the family Myrtaceae in eastern Australian woodlands, and its previously proposed four subspecies moluccana, pedicellata, queenslandica and crassifolia. A range of phylogeographic analyses were conducted to examine the population genetic differentiation and subspecies genetic structure in E. moluccana in relation to biogeographic barriers. Slow evolving markers uncovering long term processes (chloroplast DNA) were used to generate a haplotype network and infer phylogeographic barriers. Additionally, fast evolving, hypervariable markers (microsatellites) were used to estimate demographic processes and genetic structure among five geographic regions (29 populations) across the entire distribution of E. moluccana. Morphological features of seedlings, such as leaf and stem traits were assessed to evaluate population clusters and test differentiation of the putative subspecies.

Haplotype network analysis revealed twenty chloroplast haplotypes with a main haplotype in a central position shared by individuals belonging to the regions containing the four putative subspecies. Microsatellite analysis detected genetic structure between Queensland (QLD) and New South Wales (NSW) populations consistent with the McPherson Range barrier, an east-west spur of the Great Dividing Range. Substructure was detected within QLD and NSW in line with other barriers in eastern Australia. The morphological analyses supported differentiation between QLD and NSW populations, with no difference within QLD, yet some differentiation within NSW populations.

Our molecular and morphological analyses provide evidence that several geographic barriers in eastern Australia, including the Burdekin Gap and the McPherson Range have contributed to the genetic structure of E. moluccana. Genetic differentiation among E. moluccana populations supports the recognition of some but not all the four previously proposed subspecies, with crassifolia being the most differentiated.

This dataset contains psbA-trnH sequences of the chloroplast of several eucalypts. The intergenic region psbA-trnH is among the most variable regions of the chloroplast in E. moluccana (Flores-Renteria et al. 2017). Sequences for this chloroplast intergenic region were obtained using the methods of Flores-Renteria et al. (2017). Additionally, sequences from a subset of seedlings already available were included (accession numbers KY596186 – KY596665). DNA of the chloroplast sequences were aligned using the multiple progressive alignment procedure of MUSCLE (Edgar, 2004), and manual corrections were done around indels or microsatellites. Minimum spanning haplotype networks (Bandelt et al., 1999) were constructed using the program PopART (Leigh and Bryant, 2015) to better visualize non-bifurcating relationships (multifurcations and reticulations) in chloroplast haplotypes (Posada and Crandall, 2001). DNAsp software was used to create the haplotype data file (Librado and Rozas, 2009). PopART software was used to infer the haplotype network by region (http://popart.otago.ac.nz).

A second file contains the dataset of nuclear microsatellites. Ten microsatellite loci of E. moluccana were amplified in a multiplex design. Each specific forward primer was linked with one of the 5′ universal primer sequence tails and fluorescence-labeled (Table S2, Flores-Renteria et al. 2011, 2013). PCR reactions were carried out using QIAGEN Multiplex PCR Kit following manufacturer’s instructions, and amplicons were sized as in Flores-Renteria and Krohn (2013).