Persistence with episodic range expansion from the early Pleistocene: the distribution of genetic variation in the forest tree Corymbia calophylla (Myrtaceae) in south-western Australia

Persistence with episodic range expansion from the early Pleistocene: the distribution of genetic... Abstract Phylogeographical patterns of trees in topographically subdued, unglaciated landscapes are under-reported, and might reflect population persistence and the influences of environment and distance over historical (~2.6 Mya to present) and contemporary (recent generations) timescales. We examined this hypothesis using genetic analyses of four slowly evolving non-coding chloroplast sequences and 16 nuclear microsatellites in the tree Corymbia calophylla from south-western Australia, an area that has been unglaciated since the Permian (c. 300–250 Mya). We found strong population differentiation for chloroplast DNA and low differentiation for nuclear loci, consistent with higher gene flow by pollen than by seed. We identified three divergent chloroplast lineages distributed in central, northern and southern regions, and diversifying from the early (c. 3.028 Mya), mid- (c. 0.793 Mya) and late (c. 0.426 Mya) Pleistocene, respectively. Moderate to high nucleotide diversity with population-specific haplotypes supported long-term persistence, but diversification of lineages provided evidence of unexpected episodic range expansion. We suggest this pattern reflects environmental influences of climatic oscillations during progressive drying of south-western Australia from the early Pleistocene. Significant tests for isolation by environment for nuclear loci also supported an influence of contemporary environmental (aridity) conditions on genetic structure, but isolation by distance (IBD) was greater. Significant chloroplast and nuclear IBD suggested distance was a major influence on gene flow at both timescales. INTRODUCTION Gene flow in plants is influenced by geographical features (e.g. topography, distance) and environmental factors (e.g. habitat heterogeneity, climate, pollen dispersal) and these influences can lead to a variety of patterns in the genetic structures of plant species (Wang & Bradburd, 2014). Changes in geographical and environmental factors over multiple timescales, such as contemporary (several recent generations) or longer-term historical periods including the Quaternary (~2.6 Mya to present), can leave signals in the spatial genetic structure of populations (Hewitt, 2001; Wang & Bradburd, 2014). These may be especially apparent in long-lived, sessile organisms such as trees (Petit & Hampe, 2006; Gugger, Ikegami & Sork, 2013; Sork et al., 2016; Zhang et al., 2016) due to the importance of founding effects and drift in sessile species (McLachlan, Clark & Manos, 2009). Climatic oscillations during the Pleistocene (2.588 Mya to 11.700 kya) are considered to have been major drivers of broad phylogeographical patterns (Hewitt, 2000). Many studies of trees, predominantly in the Northern Hemisphere, reveal the overriding influence of population extinctions due to glaciation, with chloroplast haplotype distributions reflecting colonization patterns through rapid expansion following the Last Glacial Maximum (LGM, c. 21 kya) (reviewed by Hewitt, 2000; Petit et al., 2003). Phylogeographical studies of plant species that have been undertaken in unglaciated landscapes have inferred the influence of climatic oscillations (Byrne & Hines, 2004; Nistelberger et al., 2014), topographic heterogeneity (Soltis et al., 2006; Gugger et al., 2013), habitat heterogeneity (Llorens et al., 2016) and fire ephemeral life history (Bradbury et al., 2016a; Bradbury et al., 2016b) on genetic patterns. In unglaciated landscapes, many species show persistence, rather than contraction and expansion, as the main response to climatic oscillations (Byrne & Hines, 2004; Wheeler & Byrne, 2006; Gugger et al., 2013; Llorens et al., 2016) with phylogeographical patterns characterized by localized distribution of diverse chloroplast haplotypes without signs of expansion. However, signals of spatial expansion have been identified in some species (Soltis et al., 2006; Nevill et al., 2014; Dalmaris et al., 2015) and minor range shifts in others (Gugger et al., 2013). Many trees share life history traits (sessile life-form, size, high reproductive output, leptokurtic pollen dispersal, longevity) that affect the mode and tempo of evolution, and trees typically experience slow speciation and extinction (Petit & Hampe, 2006). Longevity can promote persistence of populations by maintaining effective population size, and the production of genetically diverse offspring over long lifetimes can help prevent extinction of local populations during climatic fluctuations (Petit & Hampe, 2006; Sork et al., 2016). The forest environments in the topographically subdued landscape of south-western Australia provide an excellent opportunity to infer responses to historical and contemporary environmental conditions through their influences on genetic structure. The Southwest Australian Floristic Region (SWAFR) is a geologically stable, highly weathered and relatively flat, low plateau, with occasional emergent granite inselbergs, that has not been glaciated since the Permian (300–250 Mya) (Hopper & Gioia, 2004). Following progressive drying since the Pliocene (5–3 Mya), the SWAFR has a seasonal Mediterranean-type climate with the most mesic conditions (600–1500 mm rainfall per annum) occurring in the south-west corner where there is a strong west to east and south-west to north-east gradient of reducing rainfall and increasing summer evaporation (Bradshaw, 2015). Forests are the dominate vegetation type in the more mesic south-west corner (Hopper & Gioia, 2004). Corymbia calophylla (Lindl.) K.D. Hill & L.A.S. Johnson is a long-lived, forest tree that can reach heights of 40–60m with a continuous distribution in mesic forest and wetter woodland regions receiving 800–1300 mm rainfall per annum in the SWAFR. There are some northern, isolated outlier populations confined to wetter soil patches in the transitional rainfall zone (300–600 mm) and C. calophylla is usually replaced by another forest species, Eucalyptus diversicolor F. Muell., in the areas of highest rainfall (1300–1500 mm) (Churchill, 1968). The distribution of C. calophylla includes the Darling Range but elevations are subdued by global standards (Hopper & Gioia, 2004). Most Corymbia are pollinated by insects and birds and it is likely that both vectors pollinate C. calophylla (O’Brien & Krauss, 2010). Seeds are primarily gravity-dispersed with the most likely potential dispersal rate equivalent to about 1–2 m per year (Booth, 2017). Occasional long-distance (>1 km) seed dispersal events may occur following fire storms or cyclones or by red-tailed black and Baudin’s cockatoos (Calyptorhynchus banksia naso and Calyptorhynchus baudinii, respectively) as C. calophylla seeds are a primary food source for these birds (Cooper et al., 2003). Complementary DNA markers that evolve at different rates are often used to distinguish between genetic divergence patterns at different timescales. In plants, bi-parentally inherited, recombining and rapidly evolving nuclear microsatellite loci are often used to reveal contemporary (several recent generations) processes that are influenced by both demographic changes, and pollen- and seed-mediated gene flow. In contrast, the chloroplast genome is slowly evolving, non-recombining and maternally inherited in most angiosperms and therefore genetic signals reflect responses of genealogical lineages over historical timescales (Schaal et al., 1998). The difference in mutation rates between nuclear and chloroplast markers is of the order of 105 (Graur & Li, 2000; Wang, 2010), establishing the utility of chloroplast DNA (cpDNA) for revealing divergence of deeper genealogical lineages (Schaal et al., 1998). We investigated the evolutionary and contemporary history of C. calophylla across its range in SWAFR given that tree species in unglaciated landscapes are expected to reveal insights into evolutionary processes and their influences on genetic structure over periods extending beyond the LGM (Byrne et al., 2011; Sork et al., 2016). We used a combination of non-coding cpDNA sequences and nuclear microsatellites to investigate spatial genetic structures at historical and contemporary timescales, respectively. We hypothesized that there would be evidence of population persistence in the phylogeographical history of C. calophylla in the mesic forests in south-west Australia, an area that has not been glaciated since the Permian (300–250 Mya). We expected geographical distance and the environment (climate) to have been key factors driving historical and contemporary patterns of genetic structure. MATERIAL AND METHODS Study species Corymbia KD Hill & LAS Johnson is a sister lineage to Eucalyptus L’Hér. and Angophora Cav. (Bayly et al., 2013) and one of three sclerophyll genera commonly known as eucalypts. The forest in which C. calophylla occurs is often separated into Northern and Southern Jarrah Forests on the basis of flowering time in the dominant species Eucalyptus marginata (Bradshaw, 2015). The environment is a fire-prone Mediterranean type and although C. calophylla is killed by severe fires, trees can be long-lived (~400 years) as they re-sprout from epicormic shoots (Wardell-Johnson, 2000) following less severe fires. Sampling and genotyping Leaves were sampled from 24 well-dispersed adult plants in each of 27 C. calophylla populations spanning the north–south distribution of the species’ range (Table 1; Fig. 1A). Genomic DNA was extracted from lysed, freeze-dried leaf material by the method described in Byrne et al. (2016). Table 1. Characteristics of 27 Corymbia calophylla populations from south-western Australia used for genotyping Species/population  Code  Latitude  Longitude  Annual rainfall (mm)  Aridity Index*  Landscape character  Northern Forest region  Mt Lesueur  LES  −30.14147200  115.11430600  600  0.44  Wheatbelt Plateau  Moochamulla  MOO  −30.98341700  115.81213900  600  0.44  Wheatbelt Plateau  Julimar  JUL  −31.40622200  116.35602800  600  0.47  Darling Plateau Uplands  Mt Helena  HEL  −31.86822200  116.22297300  900  0.54  Darling Plateau Uplands  Perry Lakes  LAK  −31.94630600  115.78118300  800  0.51  Swan Coastal Plain  Dale  DAL  −32.10136100  116.18927800  1100  0.57  Darling Plateau Uplands  Serpentine  SER  −32.35269400  116.07647200  1200  0.61  Darling Plateau Uplands  Whittaker  KER  −32.55605600  116.03100000  1100  0.62  Darling Plateau Uplands  Wearne  WEA  −32.52066700  116.49900000  700  0.54  Darling Plateau Uplands  Peel  PEE  −32.68466700  115.74266700  900  0.55  Swan Coastal Plain  Pindalup  PID  −32.80163900  116.27741700  1000  0.59  Darling Plateau Uplands  Saddleback  SAD  −32.99702800  116.48466700  700  0.55  Darling Plateau Uplands  Godfrey  GOD  −33.20758300  116.56197200  700  0.55  Darling Plateau Uplands  Yourdaming  YOU  −33.30244400  116.24094400  900  0.59  Darling Plateau Uplands  Eaton  EAT  −33.31119400  115.76336100  900  0.57  Swan Coastal Plain  Lennard  LEN  −33.32613900  115.96827800  1100  0.63  Darling Plateau Uplands  Southern Forest region  Meelup  MEE  −33.57527800  115.08600000  900  0.58  Swan Coastal Plain  Grimwade  GRI  −33.76127800  115.99991700  900  0.61  Darling Plateau Uplands  Mowen  MOW  −33.90788900  115.54211100  1000  0.63  Darling Plateau Uplands  Bramley  BRA  −33.91638900  115.08327800  1200  0.63  Leeuwin Naturalist Coast  Kingston  KIN  −34.08116700  116.33044400  800  0.61  Darling Plateau Uplands  Milylannup  MIL  −34.19658300  115.66205600  1100  0.66  Darling Plateau Uplands  Carey  CAR  −34.41963900  115.82130600  1200  0.67  Darling Plateau Pemberton Slopes  Muir  MUI  −34.65341700  117.49908300  800  0.63  Darling Plateau Uplands  Boorara  BOO  −34.63894400  116.12380600  1300  0.69  Darling Plateau Pemberton Slopes  Beadmore Rd  BEA  −34.81644400  116.65852800  1200  0.70  Darling Plateau Pemberton Slopes  Denmark  DEN  −34.90816700  117.37077800  1100  0.67  Darling Plateau Uplands  Species/population  Code  Latitude  Longitude  Annual rainfall (mm)  Aridity Index*  Landscape character  Northern Forest region  Mt Lesueur  LES  −30.14147200  115.11430600  600  0.44  Wheatbelt Plateau  Moochamulla  MOO  −30.98341700  115.81213900  600  0.44  Wheatbelt Plateau  Julimar  JUL  −31.40622200  116.35602800  600  0.47  Darling Plateau Uplands  Mt Helena  HEL  −31.86822200  116.22297300  900  0.54  Darling Plateau Uplands  Perry Lakes  LAK  −31.94630600  115.78118300  800  0.51  Swan Coastal Plain  Dale  DAL  −32.10136100  116.18927800  1100  0.57  Darling Plateau Uplands  Serpentine  SER  −32.35269400  116.07647200  1200  0.61  Darling Plateau Uplands  Whittaker  KER  −32.55605600  116.03100000  1100  0.62  Darling Plateau Uplands  Wearne  WEA  −32.52066700  116.49900000  700  0.54  Darling Plateau Uplands  Peel  PEE  −32.68466700  115.74266700  900  0.55  Swan Coastal Plain  Pindalup  PID  −32.80163900  116.27741700  1000  0.59  Darling Plateau Uplands  Saddleback  SAD  −32.99702800  116.48466700  700  0.55  Darling Plateau Uplands  Godfrey  GOD  −33.20758300  116.56197200  700  0.55  Darling Plateau Uplands  Yourdaming  YOU  −33.30244400  116.24094400  900  0.59  Darling Plateau Uplands  Eaton  EAT  −33.31119400  115.76336100  900  0.57  Swan Coastal Plain  Lennard  LEN  −33.32613900  115.96827800  1100  0.63  Darling Plateau Uplands  Southern Forest region  Meelup  MEE  −33.57527800  115.08600000  900  0.58  Swan Coastal Plain  Grimwade  GRI  −33.76127800  115.99991700  900  0.61  Darling Plateau Uplands  Mowen  MOW  −33.90788900  115.54211100  1000  0.63  Darling Plateau Uplands  Bramley  BRA  −33.91638900  115.08327800  1200  0.63  Leeuwin Naturalist Coast  Kingston  KIN  −34.08116700  116.33044400  800  0.61  Darling Plateau Uplands  Milylannup  MIL  −34.19658300  115.66205600  1100  0.66  Darling Plateau Uplands  Carey  CAR  −34.41963900  115.82130600  1200  0.67  Darling Plateau Pemberton Slopes  Muir  MUI  −34.65341700  117.49908300  800  0.63  Darling Plateau Uplands  Boorara  BOO  −34.63894400  116.12380600  1300  0.69  Darling Plateau Pemberton Slopes  Beadmore Rd  BEA  −34.81644400  116.65852800  1200  0.70  Darling Plateau Pemberton Slopes  Denmark  DEN  −34.90816700  117.37077800  1100  0.67  Darling Plateau Uplands  *Rainfall/potential evapotranspiration (see Material and Methods). View Large Figure 1. View largeDownload slide Genetic structure of sampled populations of Corymbia calophylla in south-western Australia, inferred from analysis of cpDNA (psbA-trnH, trnQ-rps16, trnG) haplotypes. (A) Distribution of haplotypes overlaid on geographical map of sampling sites. Pie charts show proportion of individuals with a given haplotype. Colour coding of the haplotypes corresponds to those of the median-joining parsimony haplotype network (Fig. 2). Geographical regions containing the Central, North and South chloroplast lineages are identified by dotted lines. The background blue scale represents annual mean rainfall (mm). (B) Maximum-clade-credibility tree from Bayesian phylogenetic analyses with the outgroup C. gummifera, calibrated using a 0–6 My CI calibration applied to a known root age from eucalypt fossils. Figure 1. View largeDownload slide Genetic structure of sampled populations of Corymbia calophylla in south-western Australia, inferred from analysis of cpDNA (psbA-trnH, trnQ-rps16, trnG) haplotypes. (A) Distribution of haplotypes overlaid on geographical map of sampling sites. Pie charts show proportion of individuals with a given haplotype. Colour coding of the haplotypes corresponds to those of the median-joining parsimony haplotype network (Fig. 2). Geographical regions containing the Central, North and South chloroplast lineages are identified by dotted lines. The background blue scale represents annual mean rainfall (mm). (B) Maximum-clade-credibility tree from Bayesian phylogenetic analyses with the outgroup C. gummifera, calibrated using a 0–6 My CI calibration applied to a known root age from eucalypt fossils. The chloroplast psbA-trnH and trnQ-rps16 intergenic spacer regions and the trnG intron were selected for amplification and sequencing in eight random samples from each of the 27 study populations. Sequences were amplified according to Byrne & Hankinson (2012) and sequenced via Macrogen Inc. (Seoul, South Korea). SEQUENCHER 5.0 (Genecodes Corp., Ann Arbor, MI, USA) was used to edit miscalls, and to align and trim sequences. All three cpDNA regions were concatenated in MESQUITE 3.04 (Maddison & Maddison, 2016) to a total sequence length of 2451 bp. One 20-bp and one 21-bp inversion were uncovered in the psbA-trnH region. Following Whitlock, Hale & Groff (2010), one configuration of inversions was replaced with its reverse-complement and coded as a single transversion. Microsatellite loci were developed from a representative individual of C. calophylla according to Byrne et al. (2016) and screened in eight individuals from six populations (Supporting Information Table S1). Loci were amplified using the Multiplex60 PCR program of the Qiagen Multiplex kit (Hilden, Germany), separated on an Applied Biosystems (Foster City, CA, USA) 3730 capillary sequencer, and 648 individuals (24 per population) were genotyped at 16 nuclear microsatellite loci using GENEMAPPER version 5.0 (Applied Biosystems). Chloroplast diversity and phylogeographical structure analysis Chloroplast haplotypes were identified using DNAsp 5.1.1 (Librado & Rozas, 2009; Supporting Information Table S2). Nucleotide (π) and haplotype (HD) diversity indices were calculated using ARLEQUIN 3.5.2.2 (Excoffier & Lischer, 2010). The number of haplotypes per population (Happop), HD and π were compared between groups using t-tests. To visualize the phylogenetic relationships among all chloroplast haplotypes, a median-joining maximum parsimony (MJMP) network was constructed in NETWORK 5.0 (Bandelt, Forster & Röhl, 1999). Population differentiation estimates (NST, GST) were calculated using PERMUT 2.0 (Pons & Petit, 1996). GST considers only allele frequencies in analysis (unordered analysis) but NST also takes distances between alleles into account (ordered analysis). NST > GST indicates a phylogenetic signal, i.e. that haplotypes within populations are on average more similar to each other than to a random set of haplotypes within the species, suggesting that genetic drift alone cannot explain the pattern and that the mutational process is influencing genetic differentiation among populations. A hierarchical analysis of molecular variance (AMOVA) was conducted in ARLEQUIN to estimate differentiation among and within populations and regions or chloroplast lineages, with 1000 permutations. Isolation by distance (IBD) effects were assessed by Mantel tests of pairwise population differentiation [FST/(1 − FST)] estimated in ARLEQUIN against the logarithm (log10) of pairwise geographical distances. Molecular dating and phylogeny reconstruction were simultaneously completed via a strict clock Bayesian analysis in BEAST 1.7.5 (Drummond & Rambaut, 2007). The substitution model was set to GTR+I+G as inferred as the best fit to the data by jModelTest (Posada, 2008). Divergence times were estimated under a strict clock model (i.e. with uniform rates across branches). Date estimates were constrained via the inclusion of a root calibration using the estimated time since the most recent common ancestor (TRMCA) of C. calophylla and Corymbia gummifera (Gaertn.) K.D.Hill & L.A.S.Johnson of 3.0 My. This calibration date was based on an unpublished dated version of the Gonzalez-Orozco et al. (2016) eucalypt phylogeny that was calibrated by Andrew Thornhill (pers. comm) using the same eucalypt fossils as previously defined and used by Thornhill & Macphail (2012). In the absence of a 95% confidence interval for the TRMCA, dating was conducted using three different hypothetical confidence intervals applied to the root age calibration, 2—4 My, 1—5 My and 0—6 My, with four independent runs of 10 million generations carried out for each of the three scenarios, sampling every 1000 generations. Convergence was assessed in Tracer 1.6 (Drummond & Rambaut, 2007) and trees were combined using LOGCOMBINER 1.6.2. TREEANNOTATOR 1.6.2 (Drummond & Rambaut, 2007) was used to identify a maximum clade credibility tree. Tests for neutrality and expansion were calculated with Tajima’s D (Tajima, 1989) and Fu’s Fs (Fu, 1997) in ARLEQUIN, and R2 (Ramos-Onsins & Rozas, 2002), and F and D (Fu & Li, 1993) calculated in DNAsp using C. gummifera as an outgroup. To infer spatial and demographic history, we used mismatch distribution analyses in ARLEQUIN to compare observed differences with model expectations assuming spatial and demographic expansion. Goodness-of-fit was tested with Harpending’s raggedness index (HRag) and the sum of squared differences (SSD). Nuclear diversity and differentiation analysis Tests for stutter bands and large allele dropout were conducted using MICROCHECKER 2.2.3 (van Oosterhout et al., 2004). Global tests of population heterozygote deficiency, and tests of linkage disequilibrium among pairs of loci, were performed with GENEPOP 4.2 (Rousset, 2008). The frequency of null alleles was estimated using FREE NA (Chapuis & Estoup, 2007). Mean multilocus genetic diversity parameters per population were estimated using GENALEX 6.501 (Peakall & Smouse, 2012). To identify genetic evidence of recent population bottlenecks, we used the Wilcoxon sign-rank test of excess heterozygosity under a two-phase model (TPM) of 70% stepwise mutation model (SMM) and 30% infinite allele model (IAM) with 30% variance, and mode shifts from an L-shaped distribution with 1000 replicates using BOTTLENECK 1.2.02 (Piry, Luikart & Cornuet, 1999). Values of among-population differentiation (FST) per locus were calculated using FREENA, correcting for null alleles by the excluding null alleles (ENA) method, using 1000 bootstraps to generate 95% confidence intervals (Chapuis & Estoup, 2007). Pairwise genetic differentiation was estimated with FST using GENALEX with statistical testing by random permutations. Genetic variation among populations was partitioned using AMOVA in ARLEQUIN. An indirect measure of gene flow among populations was estimated using the private allele frequency method of (Slatkin, 1985) in GENEPOP. To estimate the relative contribution to total gene flow of seeds versus pollen, we calculated the pollen-to-seed gene flow ratio of Ennos (1994) as [A(1 + FIS) − 2C]/C, where A = [(1/FST) − 1] and C = [(1/GST − 1)]. Values for FIS and FST were obtained from GENEPOP using nuclear microsatellite data, and GST from chloroplast data using PERMUT. IBD and isolation by environment (IBE) were investigated by testing for correlations between log10 pairwise geographical distances and genetic differentiation (FST/(1 − FST) and the normalized differences of the ‘Aridity Index’ among populations, respectively, using a Mantel procedure in GENALEX. The ‘Aridity Index’ (annual mean monthly ratio of precipitation to potential evaporation) for each population was obtained from the Atlas of Living Australia (http://www.ala.org.au/, accessed 19 October 2016) and used as an environmental variable to indicate the local climate of each population. The Aridity Index is a numerical environmental indicator of the degree of dryness of the climate at a given location gridded at 0.01° (~1 km) and is intended to describe the magnitude of potential water limitation and biological stress influencing biotic distribution patterns. To compare the influence of IBD and IBE, we also used Mantel tests under a multiple correlation and regression design to simultaneously evaluate variables, with the proportion of the total explained variance due to geographial distance or aridity estimated following Telles & Diniz-Filho (2005) by removing each variable alternately. To assess genetic relationships among populations, we performed a phenetic analysis based on the CS Chord genetic distance calculated using MSA 4.05 (Dieringer & Schlötterer, 2003) and constructed an unrooted neighbor-joining (NJ) tree using PHYLIP 3.69 (Felsenstein, 1989) with clustering patterns validated with 1000 bootstraps. Patterns of population genetic structure were visualized using a Bayesian model-based approach implemented in STRUCTURE 2.3.4 (Pritchard, Stephens & Donnelly, 2000). STRUCTURE was used to identify the most likely number (K) of genetically homogeneous clusters, and to assign proportionate membership (q) of individuals to clusters. Analyses used no prior knowledge, the admixture ancestry model and a burn-in of 150 000 with 200 000 iterations for Markov chain Monte Carlo parameters. Analyses were done for K = 1–27 and assessed using the similarity coefficient h′ in CLUMPP 1.1.2 (Jakobsson & Rosenberg, 2007). The optimum K of genetically homogeneous clusters was determined from 20 replicate runs for each value of K using the statistic ΔK from the program STRUCTURE HARVESTER 0.6.93 (Earl & von Holdt, 2012). RESULTS Chloroplast diversity and phylogeographical structure Twenty-nine unique haplotypes were found among individuals (Fig. 2; Table S2). Overall nucleotide diversity was moderate (π = 0.086, SD = 0.044, Table 2) and haplotype diversity was high (HD = 0.931, SD = 0.007). Most haplotypes (N = 24, 82.8%; Fig. 1) were found in single populations and within-population haplotype diversity was low (hS = 0.200, SD = 0.045). Unique haplotypes were distributed throughout the species’ range rather than in restricted groups of peripheral populations. Two haplotypes were more widely distributed: H01 occurred in six populations and H02 in five populations. Table 2. Genetic diversity and differentiation indices of chloroplast haplotypes for 27 populations of Corymbia calophylla from south-western Australia   All populations  Central lineage  North lineage  South lineage  Number of populations  N  27  4  3  20  Number of samples  n  216  32  24  160  Number of haplotypes  HN  29  9  4  16  Mean number of haplotypes per population  Happop  1.593 (0.134)  2.250 (0.479)  1.333 (0.333  1.500 (0.136)  Haplotype diversity  HD  0.931 (0.007)  0.859 (0.029)  0.750 (0.040)  0.884 (0.011)  Nucleotide diversity  π  0.086 (0.044)  0.126 (0.065)  0.052 (0.029)  0.029 (0.017)  Within-population haplotype diversity  hs  0.200 (0.045)  0.375 (0.128)  0.179 (0.179)  0.168 (0.048)  Population differentiation (unordered)  GST  0.791 (0.047)  0.625 (0.129)  0.856 (0.180)  0.818 (0.052)  Population differentiation (ordered)  NST  0.836 (0.053)  0.559 (0.165)  0.821 (0.173)  0.833 (0.072)  Phylogeogenetic signal within populations  NST > GST  NS  NS  NS  NS    All populations  Central lineage  North lineage  South lineage  Number of populations  N  27  4  3  20  Number of samples  n  216  32  24  160  Number of haplotypes  HN  29  9  4  16  Mean number of haplotypes per population  Happop  1.593 (0.134)  2.250 (0.479)  1.333 (0.333  1.500 (0.136)  Haplotype diversity  HD  0.931 (0.007)  0.859 (0.029)  0.750 (0.040)  0.884 (0.011)  Nucleotide diversity  π  0.086 (0.044)  0.126 (0.065)  0.052 (0.029)  0.029 (0.017)  Within-population haplotype diversity  hs  0.200 (0.045)  0.375 (0.128)  0.179 (0.179)  0.168 (0.048)  Population differentiation (unordered)  GST  0.791 (0.047)  0.625 (0.129)  0.856 (0.180)  0.818 (0.052)  Population differentiation (ordered)  NST  0.836 (0.053)  0.559 (0.165)  0.821 (0.173)  0.833 (0.072)  Phylogeogenetic signal within populations  NST > GST  NS  NS  NS  NS  Standard deviations in parentheses. NS, non-significant. View Large Figure 2. View largeDownload slide Median-joining maximum parsimony haplotype network of Corymbia calophylla from south-western Australia, based on cpDNA sequences. Groups are enclosed in boxes. Circle size in the network is relative to haplotype frequency and boxes on the network branches represent mutations. Figure 2. View largeDownload slide Median-joining maximum parsimony haplotype network of Corymbia calophylla from south-western Australia, based on cpDNA sequences. Groups are enclosed in boxes. Circle size in the network is relative to haplotype frequency and boxes on the network branches represent mutations. The haplotype network of phylogenetic relationships showed an asymmetrical structure revealing three haplotype groups: a group of divergent haplotypes that were found geographically in populations in the central–west of the species’ range (henceforth Central lineage), a group of moderately divergent haplotypes found in populations in the north (North lineage) and a radiation of weakly diverged haplotypes found in populations in the south (South lineage). Only one population had haplotypes from more than one lineage; WEA contained seven individuals with a South lineage haplotype (H24) and one individual with a Central lineage haplotype (H29). Nucleotide diversity was higher in the Central lineage (π = 0.126, SD = 0.065) compared to the North (π = 0.052, SD = 0.029) or South lineages (π = 0.029, SD = 0.017) but haplotype diversities (HD) were similar (0.859, SD = 0.029; 0.750, SD = 0.040; 0.884, SD = 0.011, respectively). Within-population haplotype diversity was highest in the Central lineage (hs = 0.375, SD = 0.128; North, hs = 0.179, SD = 0.179; South, hs = 0.168, SD = 0.048). Both unordered (GST) and ordered (NST) estimates indicated significant population differentiation (Table 2). However, there was no significant phylogenetic signal (NST > GST) within populations for cpDNA over all populations (GST/NST = 0.704/0.802, P < 0.01) or within any lineage. Non-hierarchical AMOVA analyses over all populations indicated significant differentiation of populations with most of the variance among (83.63%) rather than within (16.37%) populations. There was significant support for IBD (r = 0.192, P < 0.05). Hierarchical analyses of populations partitioned according to the chloroplast lineages suggested by the haplotype network, or the geographical regions they aligned with, showed strong structuring with the majority of variance between lineages (65.59%; Table 3) rather than among populations within lineages (24.13%) or within populations (10.28%). Table 3. Analysis of molecular variance (AMOVA) of Corymbia calophylla for chloroplast haplotype data from 27 populations from Central, North and South chloroplast lineages*, and for nuclear microsatellite data from 27 populations separated into Northern Forest and Southern Forest regions in south-western Australia Source of variation  d.f.  SS  Variance component  Variance (%)  Chloroplast haplotypes  Among lineages  2  473.6  4.919  65.59  Among populations within lineages  24  366.002  1.81  24.13  Within populations  189  145.75  0.771  10.28  Nuclear microsatellites  Among forest regions  1  44.026  0.056  1.48  Among populations within forest regions  25  217.994  0.105  2.76  Within populations    4902.124  3.656  95.76  Source of variation  d.f.  SS  Variance component  Variance (%)  Chloroplast haplotypes  Among lineages  2  473.6  4.919  65.59  Among populations within lineages  24  366.002  1.81  24.13  Within populations  189  145.75  0.771  10.28  Nuclear microsatellites  Among forest regions  1  44.026  0.056  1.48  Among populations within forest regions  25  217.994  0.105  2.76  Within populations    4902.124  3.656  95.76  *Central, North and South lineages were found in geographical Central, North and South regions as described in the Results. View Large Bayesian analyses of the combined chloroplast sequences revealed early divergence of two haplotypes found in the central part of the geographical distribution followed by three chloroplast lineages of sequentially younger ages that correspond to populations (with the exception of WEA) in the geographical centre–west (henceforth Central region), north (North region), and centre–east and south (South region) of the species distribution (Fig. 1B). These lineages and geographical regions aligned with the Central, North and South groups of the haplotype network (Fig. 2), respectively. Trees and dates derived from the 0–6, 1–5, and 2–4 My CI applied to the root age calibration were similar with a small difference in the location of the clade containing Peel population (PEE) in the 2–4 My CI scenario (Supporting Information Fig. S1). Using the 0–6 My CI, the TMRCA of all haplotypes was 3.028 Mya [CI 2.307–3.842 Mya, posterior probability (PP) = 1] in the late Pliocene. Bayesian analyses revealed strong support for divergence of North and South lineages from the Central lineage haplotypes in the early Pleistocene at 1.982 Mya (CI 1.422–2.637 Mya, PP = 1). The Central lineage is characterized by haplotypes found in single populations with deep divergences up to 3.028 Mya. There was strong support for divergence of North and South lineages at 0.793 Mya (CI 0.517–1.129 Mya, PP = 1) and for later divergence of haplotypes within the South lineage from 0.426 Mya (CI 0.256–0.625 Mya; PP = 1). Across all populations, neutrality tests were not significant with the exception of Tajima’s D (Table 4). When chloroplast lineages were analysed separately, neutrality statistics were low or negative and Tajima’s D was significant in the South lineage only. Mismatch analyses indicated spatial expansion for the South and North lineages but not for the Central lineage. Considering the results together, tests for C. calophylla suggest spatial expansion in the South lineage found in the geographically south region. Table 4. Tests for neutrality and for demographic and spatial expansion based on Corymbia calophylla chloroplast haplotypes sampled from 27 populations in south-western Australia Chloroplast lineage  All populations  Central  North  South  Number of populations  27  4  3  20  Number of samples  216  32  24  160  Tajima’s D  −1.498 (P < 0.05)  −0.866 (P > 0.05)  1.622 (P > 0.05)  −1.497 (P < 0.01)  Ramos-Onsins and Rozas R2  0.043 (P = 0.054)  0.093 (P > 0.05)  0.196 (P > 0.05)  0.060 (P > 0.05)  Fu’s FS  0.439 (P > 0.05)  7.408 (P > 0.05)  6.942 (P > 0.05)  −1.598 (P > 0.05)  Fu and Li’s D*  0.693 (P > 0.10)  1.338 (P > 0.10)  1.662 (P < 0.05)  −2.036 (P > 0.05)  Fu and Li’s F*  −0.476 (P > 0.10)  0.615 (P > 0.10)  1.985 (P < 0.05)  −2.179 (P > 0.05)  Demographic expansion (SSD)†  0.016 (P > 0.05)  0.095 (P < 0.01)  0.112 (P < 0.01)  0.016 (P < 0.05)  Demographic expansion (HRag)†  0.018 (P > 0.05)  0.240 (P < 0.01)  0.275 (P < 0.01)  0.043 (P < 0.05)  Spatial expansion (SSD)†  0.016 (P > 0.05)  0.082 (P < 0.05)  0.067 (P > 0.05)  0.014 (P > 0.05)  Spatial expansion (HRag)†  0.018 (P > 0.05)  0.240 (P < 0.05)  0.275 (P > 0.05)  0.043 (P > 0.05)  Chloroplast lineage  All populations  Central  North  South  Number of populations  27  4  3  20  Number of samples  216  32  24  160  Tajima’s D  −1.498 (P < 0.05)  −0.866 (P > 0.05)  1.622 (P > 0.05)  −1.497 (P < 0.01)  Ramos-Onsins and Rozas R2  0.043 (P = 0.054)  0.093 (P > 0.05)  0.196 (P > 0.05)  0.060 (P > 0.05)  Fu’s FS  0.439 (P > 0.05)  7.408 (P > 0.05)  6.942 (P > 0.05)  −1.598 (P > 0.05)  Fu and Li’s D*  0.693 (P > 0.10)  1.338 (P > 0.10)  1.662 (P < 0.05)  −2.036 (P > 0.05)  Fu and Li’s F*  −0.476 (P > 0.10)  0.615 (P > 0.10)  1.985 (P < 0.05)  −2.179 (P > 0.05)  Demographic expansion (SSD)†  0.016 (P > 0.05)  0.095 (P < 0.01)  0.112 (P < 0.01)  0.016 (P < 0.05)  Demographic expansion (HRag)†  0.018 (P > 0.05)  0.240 (P < 0.01)  0.275 (P < 0.01)  0.043 (P < 0.05)  Spatial expansion (SSD)†  0.016 (P > 0.05)  0.082 (P < 0.05)  0.067 (P > 0.05)  0.014 (P > 0.05)  Spatial expansion (HRag)†  0.018 (P > 0.05)  0.240 (P < 0.05)  0.275 (P > 0.05)  0.043 (P > 0.05)  Tajima’s D is expected to be negative when there is an excess of rare, less diverged mutations, which may be due to sudden, recent population growth. R2 is expected to be low when there is an excess or rare, less diverged mutations, which may be due to sudden, recent population growth. FS is expected to be negative when there is an excess of rare mutations. *Corymbia gummifera used as an outgroup. †Demographic and spatial expansion models test for deviation from the distribution expected under expansion and therefore P > 0.05 supports a scenario of expansion. View Large Nuclear diversity and structure No evidence of stutter or large allele dropout was detected. We found 25 (out of a possible 3240) instances of significant composite genotypic disequilibrium within populations. The loci involved were not consistent across populations. We detected 27 frequencies of null alleles significantly greater than 0.05 over all loci and populations (out of a possible 432; Supporting Information Table S3). Comparison of FST (95% CI) estimates showed that null alleles did not cause significant bias (without ENA: 0.032, 0.029–0.037; with ENA: 0.033, 0.029–0.037). Therefore, all loci were included in analyses. A total of 319 alleles were amplified from 16 microsatellite loci among all individuals. Genetic diversity parameters were high (Table 5) and inbreeding coefficients were generally low (F = −0.002 to 0.151), as expected for a mixed mating species (Byrne, 2008). Bottleneck analysis detected significant probabilities of heterozygote excess for six populations from the Northern Forest region (LAK, WEA, PEE, SAD, GOD, YOU; Table 5) and three Southern Forest populations (MEE, BRA, MIL) using the IAM model. Table 5. Nuclear microsatellite diversity of 27 Corymbia calophylla populations from south-western Australia Population  Code  n  A  Np  Ho  UHe  F  Bottleneck test†  Northern Forest region  Mt Lesueur  LES  24  7.50 (0.84)  3  0.661 (0.052)  0.671 (0.052)  0.000 (0.029)  0.0649  Moochamulla  MOO  24  8.31 (1.08)  2  0.681 (0.049)  0.681 (0.045)  −0.002 (0.026)  0.3161  Julimar  JUL  24  8.88 (1.00)  3  0.648 (0.061)  0.714 (0.048)  0.101 (0.060)  0.0523  Mt Helena  HEL  24  9.13 (0.93)  2  0.675 (0.049)  0.735 (0.032)  0.089 (0.047)  0.1489  Perry Lakes  LAK  24  7.88 (0.96)  2  0.638 (0.057)  0.686 (0.051)  0.068 (0.040)  0.0145*  Dale  DAL  23  8.50 (0.89)  2  0.655 (0.047)  0.699 (0.046)  0.046 (0.042)  0.2809  Serpentine  SER  22  8.44 (0.97)  1  0.605 (0.035)  0.689 (0.039)  0.111 (0.034)*  0.4105  Whittaker  KER  24  9.06 (0.94)  2  0.651 (0.061)  0.717 (0.047)  0.122 (0.065)  0.1261  Wearne  WEA  24  9.00 (0.94)  0  0.708 (0.049)  0.745 (0.038)  0.062 (0.037)  0.0046*  Peel  PEE  24  9.06 (1.22)  4  0.692 (0.046)  0.735 (0.044)  0.047 (0.035)  0.0000*  Pindalup  PID  19  7.94 (0.88)  3  0.635 (0.049)  0.693 (0.053)  0.074 (0.040)  0.2983  Saddleback  SAD  23  8.63 (0.99)  0  0.664 (0.042)  0.729 (0.041)  0.080 (0.034)*  0.0026*  Godfrey  GOD  24  7.94 (0.88)  1  0.614 (0.044)  0.701 (0.043)  0.112 (0.041)*  0.0008*  Yourdaming  YOU  23  8.63 (1.11)  2  0.612 (0.053)  0.713 (0.049)  0.151 (0.038)*  0.0055*  Eaton  EAT  24  8.88 (1.08)  4  0.625 (0.054)  0.715 (0.047)  0.123 (0.037)*  0.1156  Lennard  LEN  24  9.31 (1.23)  2  0.646 (0.063)  0.711 (0.053)  0.096 (0.043)*  0.0523  Southern Forest region  Meelup  MEE  24  8.38 (0.88)  0  0.654 (0.055)  0.707 (0.044)  0.082 (0.039)*  0.0222*  Grimwade  GRI  24  8.31 (0.83)  5  0.628 (0.049)  0.710 (0.047)  0.101 (0.045)*  0.1261  Mowen  MOW  24  9.38 (1.03)  1  0.695 (0.045)  0.719 (0.045)  0.026 (0.026)  0.2477  Bramley  BRA  24  9.13 (0.89)  1  0.740 (0.037)  0.751 (0.036)  0.004 (0.038)  0.0327*  Kingston  KIN  24  8.81 (1.01)  3  0.625 (0.050)  0.704 (0.048)  0.100 (0.046)*  0.2019  Milylannup  MIL  24  9.06 (1.11)  5  0.688 (0.052)  0.722 (0.050)  0.040 (0.034)  0.0327*  Carey  CAR  24  8.94 (0.87)  1  0.698 (0.052)  0.709 (0.045)  0.013 (0.033)  0.3342  Muir  MUI  23  8.69 (1.00)    0.691 (0.054)  0.706 (0.048)  0.018 (0.036)  0.1489  Boorara  BOO  24  8.75 (1.01)  0  0.692 (0.042)  0.722 (0.037)  0.042 (0.027)  0.0795  Beadmore Road  BEA  24  8.75 (0.81)  2  0.678 (0.051)  0.706 (0.041)  0.053 (0.034)  0.2983  Denmark  DEN  23  8.69 (0.95)  1  0.642 (0.054)  0.696 (0.046)  0.074 (0.045)  0.4500  Mean    23  8.66 (0.18)    0.661 (0.010)  0.711 (0.009)  0.068 (0.008)    Population  Code  n  A  Np  Ho  UHe  F  Bottleneck test†  Northern Forest region  Mt Lesueur  LES  24  7.50 (0.84)  3  0.661 (0.052)  0.671 (0.052)  0.000 (0.029)  0.0649  Moochamulla  MOO  24  8.31 (1.08)  2  0.681 (0.049)  0.681 (0.045)  −0.002 (0.026)  0.3161  Julimar  JUL  24  8.88 (1.00)  3  0.648 (0.061)  0.714 (0.048)  0.101 (0.060)  0.0523  Mt Helena  HEL  24  9.13 (0.93)  2  0.675 (0.049)  0.735 (0.032)  0.089 (0.047)  0.1489  Perry Lakes  LAK  24  7.88 (0.96)  2  0.638 (0.057)  0.686 (0.051)  0.068 (0.040)  0.0145*  Dale  DAL  23  8.50 (0.89)  2  0.655 (0.047)  0.699 (0.046)  0.046 (0.042)  0.2809  Serpentine  SER  22  8.44 (0.97)  1  0.605 (0.035)  0.689 (0.039)  0.111 (0.034)*  0.4105  Whittaker  KER  24  9.06 (0.94)  2  0.651 (0.061)  0.717 (0.047)  0.122 (0.065)  0.1261  Wearne  WEA  24  9.00 (0.94)  0  0.708 (0.049)  0.745 (0.038)  0.062 (0.037)  0.0046*  Peel  PEE  24  9.06 (1.22)  4  0.692 (0.046)  0.735 (0.044)  0.047 (0.035)  0.0000*  Pindalup  PID  19  7.94 (0.88)  3  0.635 (0.049)  0.693 (0.053)  0.074 (0.040)  0.2983  Saddleback  SAD  23  8.63 (0.99)  0  0.664 (0.042)  0.729 (0.041)  0.080 (0.034)*  0.0026*  Godfrey  GOD  24  7.94 (0.88)  1  0.614 (0.044)  0.701 (0.043)  0.112 (0.041)*  0.0008*  Yourdaming  YOU  23  8.63 (1.11)  2  0.612 (0.053)  0.713 (0.049)  0.151 (0.038)*  0.0055*  Eaton  EAT  24  8.88 (1.08)  4  0.625 (0.054)  0.715 (0.047)  0.123 (0.037)*  0.1156  Lennard  LEN  24  9.31 (1.23)  2  0.646 (0.063)  0.711 (0.053)  0.096 (0.043)*  0.0523  Southern Forest region  Meelup  MEE  24  8.38 (0.88)  0  0.654 (0.055)  0.707 (0.044)  0.082 (0.039)*  0.0222*  Grimwade  GRI  24  8.31 (0.83)  5  0.628 (0.049)  0.710 (0.047)  0.101 (0.045)*  0.1261  Mowen  MOW  24  9.38 (1.03)  1  0.695 (0.045)  0.719 (0.045)  0.026 (0.026)  0.2477  Bramley  BRA  24  9.13 (0.89)  1  0.740 (0.037)  0.751 (0.036)  0.004 (0.038)  0.0327*  Kingston  KIN  24  8.81 (1.01)  3  0.625 (0.050)  0.704 (0.048)  0.100 (0.046)*  0.2019  Milylannup  MIL  24  9.06 (1.11)  5  0.688 (0.052)  0.722 (0.050)  0.040 (0.034)  0.0327*  Carey  CAR  24  8.94 (0.87)  1  0.698 (0.052)  0.709 (0.045)  0.013 (0.033)  0.3342  Muir  MUI  23  8.69 (1.00)    0.691 (0.054)  0.706 (0.048)  0.018 (0.036)  0.1489  Boorara  BOO  24  8.75 (1.01)  0  0.692 (0.042)  0.722 (0.037)  0.042 (0.027)  0.0795  Beadmore Road  BEA  24  8.75 (0.81)  2  0.678 (0.051)  0.706 (0.041)  0.053 (0.034)  0.2983  Denmark  DEN  23  8.69 (0.95)  1  0.642 (0.054)  0.696 (0.046)  0.074 (0.045)  0.4500  Mean    23  8.66 (0.18)    0.661 (0.010)  0.711 (0.009)  0.068 (0.008)    n, Number of samples; A, mean number of alleles per locus; Np, total number of private alleles within species; Ho, observed heterozygosity; UHe, unbiased expected heterozygosity; F, Wright’s inbreeding coefficient; †, Wilcoxon test, probability of a heterozygote excess under the IAM model. Standard errors in parentheses. *, Significant, P < 0.05. View Large The genetic differentiation of C. calophylla populations was significant but low (FST = 0.033, 95% CI 0.029–0.037) and there were 52 private alleles detected among the total 319 alleles. Mantel tests showed there was support for both IBD (r = 0.535, P < 0.01) and IBE (r = 0.452, P < 0.01). The effect of geographical distance was five times higher than the environmental effect (aridity) when the association between both parameters and genetic distance was partitioned by regression analyses. Hierarchical ANOVA showed a small proportion of variance was due to differences between the Northern and Southern Forest regions (1.48%; Table 3) although differentiation among populations was higher in the Northern Forest (FST = 0.054, SE 0.003) compared to the Southern Forest (FST = 0.038, SE 0.001). The indirect estimate of gene flow among populations (mean Nm) was 3.45. The estimate of pollen to seed gene flow ratio based on comparisons of nuclear and chloroplast differentiation (68:1) inferred substantially higher pollen- than seed-mediated gene flow. Populations of C. calophylla did not form distinct clusters with strong bootstrap support in the consensus NJ tree of CS Chord genetic distance (Fig. 3B). A geographical north/south division was present although support for this was weak (< 50%). The STRUCTURE analysis identified an optimum of two clusters (q1, q2) with a high similarity of runs (h′ = 0.995; Fig. 3C; Supporting Information Fig. S2). Individuals were admixtures of the two clusters with the relative proportions changing in a cline that generally followed the north–south geographical distribution and the north-east to south-west gradation of increasing rainfall and decreasing temperatures. The proportional cluster membership of populations was strongly correlated with the Aridity Index (Population q1, r = −0.771, P < 0.01). Figure 3. View largeDownload slide Genetic structure among sampled populations of Corymbia calophylla in south-western Australia, inferred from nuclear microsatellites. (A) Map showing the locations of sampled populations. Corymbia calophylla populations are shown as pie charts representing the proportion of genetic assignment to genetic clusters in STRUCTURE analyses. (B) Neighbour-joining tree of CS Chord distance. The pies on branches correspond to the proportion of genetic assignment to genetic clusters in STRUCTURE analyses. Support is shown on the branches as the number of bootstraps out of 1000. Values > 500 are shown. (C) Ancestry of individuals inferred from nuclear microsatellites data using STRUCTURE 2.3.4. Each individual is represented as a single line with coloured segments representing the proportion of ancestry from clusters (q). Results are the optimal alignment of 20 replicates. Figure 3. View largeDownload slide Genetic structure among sampled populations of Corymbia calophylla in south-western Australia, inferred from nuclear microsatellites. (A) Map showing the locations of sampled populations. Corymbia calophylla populations are shown as pie charts representing the proportion of genetic assignment to genetic clusters in STRUCTURE analyses. (B) Neighbour-joining tree of CS Chord distance. The pies on branches correspond to the proportion of genetic assignment to genetic clusters in STRUCTURE analyses. Support is shown on the branches as the number of bootstraps out of 1000. Values > 500 are shown. (C) Ancestry of individuals inferred from nuclear microsatellites data using STRUCTURE 2.3.4. Each individual is represented as a single line with coloured segments representing the proportion of ancestry from clusters (q). Results are the optimal alignment of 20 replicates. DISCUSSION Our analysis of genetic patterns in C. calophylla largely confirms our hypothesis that long-lived trees will show population persistence in stable, unglaciated mesic landscapes. However, evidence of unexpected range expansion in episodes during the Quaternary (~2.6 Mya to present) revealed a major role of the environment, through climatic oscillations and progressive drying of the forest habitat, in shaping genetic structure in this forest tree. In the absence of major topographic features within the distribution of the species, geographical distance appears to have been a key determinant of the spatial distribution of genetic variation at both historical and contemporary timescales, as expected for a sessile species in which restricted seed dispersal and localized pollen dispersal limits gene flow. Our results also confirmed our expectation that the environment may have shaped contemporary genetic structure through higher gene flow among plants with similar local climates, although the influence of distance was much greater than that of environment. Long-term persistence with episodic range expansion The phylogenetic structure of C. calophylla revealed evidence of persistence with unexpected range expansion in episodes during the Quaternary (~2.6 Mya to present). In the central and north regions, we found the expected signals of persistence with overall high haplotype and mid- to high nucleotide diversity, and highly divergent, localized haplotypes in populations in the lineages. In contrast, the lineage found in the south region included some widespread, shared, derived haplotypes and lower nucleotide diversity that is a pattern typically seen in range expansions (Hewitt, 2000; Pannell & Dorken, 2006). Neutrality and mismatch statistics also supported spatial expansion in the south, but not in the central or northern regions. The time sequence of diversification of lineages in C. calophylla during the Pleistocene (2.588 Mya to 11.7 kya) suggests range expansion consistent with the southward progress of increasing aridity in the SWAFR from the mid-Pleistocene (Hopper & Gioia, 2004; Byrne et al., 2008) that opened up suitable habitat within the forests. At the broad scale, the three chloroplast lineages (Central, North, South) correspond to divergence in a time sequence (early to mid to late-Pleistocene) over geographic regions (central, north, south). Diversification of the earliest lineage (Central) began in the late Pliocene (c. 3.028 Mya) during the period when progressive drying of mesic environments accelerated (see Byrne et al., 2011, 2008). The later divergence of the North and South lineages (c. 0.793 Mya) corresponds to the mid-Pleistocene transition (0.7–0.8 Mya), when climatic oscillations became more extreme (Zachos et al., 2001), and is consistent with the early (Nistelberger et al., 2014) or mid-Pleistocene (Byrne, Macdonald & Brand, 2003; Byrne & Hines, 2004) lineage divergence observed in other trees and woody shrubs in south-western Australia. However, the later range expansion of the South lineage in C. calophylla beginning around 0.426 Mya was unexpected. It suggests a significant influence of later Pleistocene climatic oscillations of increasing aridity in this more mesic environment, and is consistent with evidence of a significant shift in hydrological state to widespread aridity from approximately 500000 years ago in lakes in western, central and eastern Australia (Zheng et al., 1998). The southern progress of aridity in the SWAFR since the Pliocene (5–3 Mya) led to vegetation changes as denser forest dried and gave way to more open forests (Hopper & Gioia, 2004; Martin, 2006) and this may have provided opportunities for C. calophylla to colonize areas that had previously been dominated by species favouring wetter environments. Major expansions appear to have been rare in the mesic biota (Byrne et al., 2011, but see Nistelberger et al., 2014) although a similar pattern of expansion in the southern forest was seen in Allocasuarina humilis (Otto & A. Dietr) L.A.S. Johnson (Llorens et al., 2016), and the divergence of lineages in Eucalyptus wandoo Blakely from a geographically central refugium that were estimated to have been no later than 0.53 Mya (Dalmaris et al., 2015) may also be associated with expansion into the mesic forests as the environment dried. Influence of distance on genetic structure A signal of IBD for chloroplast markers indicated the influence of geographical distance in shaping spatial genetic structure at the historical scale. A lack of significant phylogeographical signal within populations (NST > GST) over the entire species’ range was consistent with theoretical expectations (Vekemans & Hardy, 2004) of population differentiation being determined by a balance between drift and gene flow rather than by mutation. Strong structuring and differentiation of populations and highly localized distributions of haplotypes is consistent with gene flow via seeds being too low to counter the diversifying effects of mutation and drift (Lowe & Allendorf, 2010). Seed-mediated gene flow is likely to be low in eucalypts because they have very limited seed dispersal capabilities with dispersal rates of 1–2 m per year (Booth 2017). Seed movements tend to take place in a stepwise fashion, taking advantage of disturbance events such as bushfires. Almost all seed dispersal is expected to be localized although occasional long-distance seed movements might be achieved by wind vortices in firestorms or cyclonic winds. Long-distance seed movements make also occur by cockatoos, as these large birds use C. calophylla seed as a food source and are known to carry capsules over some distance (Cooper et al., 2003). It is assumed that uneaten seeds are sometimes dispersed but there is no published data on how far or often this occurs (Booth, 2017). Foraging by cockatoos is more common in the southern forest (as flocks in the northern forests migrate to the coastal plain in summer) and might facilitate maintenance of greater connectivity in these populations. The high pollen to seed gene flow ratio (68:1) and low nuclear differentiation indicated that gene flow via pollen dispersal in C. calophylla has been more effective in maintaining the genetic connectivity of populations. Studies of pollination distances in the eucalypts show that dispersal is leptokurtic but with a fat tail, and is generally localized in continuous forest populations (Barbour, Potts & Vallaincourt, 2005; Jones et al., 2008). However, pollination can be extensive over long distances in more open woodlands or in fragmented landscapes (Byrne et al., 2008; Sampson & Byrne, 2008; Mimura et al., 2009). Pollen-mediated gene flow is therefore likely to be far higher than that achieved by restricted seed dispersal, and both overlapping gene pools and long-distance pollen dispersal are expected to retard differentiation in the nuclear genome. Low differentiation is generally observed in studies of widespread eucalypt species (Byrne, 2008) including the co-distributed forest tree E. marginata (Wheeler, Byrne & McComb 2003). However, although differentiation was low in C. calophylla, the strong signal of IBD in the nuclear genetic structure indicates a spatial component in the distribution of genetic variation and the influence of geographical distance at the contemporary timescale. The distribution of nuclear genetic variation is expected to be related to distance in most plants because pollen dispersal is usually leptokurtic (Vekemans & Hardy, 2004), particularly in a landscape with few topographical barriers. Genetic drift does not appear to have significantly disrupted the pattern of IBD despite evidence in nuclear data of bottlenecks in some populations. Bottlenecks could be the result of demographic changes following recurrent severe fires that have characterized the landscape (Pickett, 1997; Prideaux et al., 2010) but pollen-mediated gene flow has apparently been high enough in C. calophylla to prevent significant drift. We also found evidence that contemporary genetic structure may be shaped by the current environment, measured as local climate in the Aridity Index, although to a lesser extent than by distance (IBD ~ 84.6% vs. IBE ~ 15.5%). IBE can be produced by selection or if gene flow is restricted among plants in habitats with environmental differences through a variety of processes (Orsini, Andrew & Eizaguirre, 2013; Sork et al., 2016). Such processes could include isolation due to differences in flowering time created by environmental variation or changes in pollinator behaviour in different environments (Llorens et al., 2013). Flowering varies within the range of C. calophylla (Churchill, 1968; Cooper et al., 2003) and therefore restricted pollen dispersal among plants that flower concurrently might be expected to contribute to IBE. CONCLUSION In an unglaciated and topographically subdued landscape, the genetic structure of the mesic forest tree C. calophylla reflects persistence over multiple climatic oscillations during the Pleistocene (2.588 Mya to 11.7 kya). Environmental factors have been important in shaping genetic structure particularly on historical timescales. Deep divergence of lineages and unexpected episodic range expansion since the early Pleistocene appear to have been associated with climatic oscillations and progressive drying of the landscape. Geographical distance has been a key influence on both contemporary and historical genetic structure. Our study demonstrates that analysis of patterns of genetic diversity at historical and contemporary timescales in unglaciated landscapes can reveal insights into evolutionary processes and influences in forest species over long periods since the early Pleistocene. This knowledge contributes to informed management and climate adaptation strategies in forest trees that are critical for maintenance of forest ecosystems ACKNOWLEDGEMENTS We thank Andrew Thornhill for providing node dates for divergence estimates and advice on BEAST analyses. We thank the anonymous reviewers for their comments on our manuscript. SUPPORTING INFORMATION Additional Supporting Information may be found online in the supporting information tab for this article: Figure S1. Maximum-clade-credibility trees from Bayesian phylogenetic analyses for Corymbia calophylla made with (A) 2–4 My CI and (B) 1–5 My CI calibration applied to a known root age from eucalypt fossils. Figure S2. Delta K values from STRUCTURE HARVESTER 0.6.93 (Earl & von Holdt, 2012) of Bayesian inference of the number of nuclear microsatellite genetic marker clusters using STRUCTURE 2.3 4 (Pritchard et al., 2000) for Corymbia calophylla. Table S1. Primer sequences and characteristics of 16 microsatellite loci isolated from Corymbia calophylla. Allele size ranges and the number of alleles (NA) are based on 648 individuals from 27 natural populations. Table S2. List of GenBank accessions for haplotypes uncovered in 27 Corymbia calophylla populations in south-western Australia via sequencing of psbA-trnH, trnG and trnQ-rps16 chloroplast intergenic spacer regions. Coloured boxes next to haplotypes correspond to colours used in Figures 1 and 2. Table S3. Estimates of microsatellite null allele frequencies made for 27 populations of Corymbia calophylla from south-west Australia. REFERENCES Bandelt HJ, Forster P, Röhl A. 1999. Median-joining networks for inferring intraspecific phylogenies. Molecular Biology and Evolution  16: 37– 48. Google Scholar CrossRef Search ADS PubMed  Barbour RC, Potts BM, Vaillancourt RE. 2005. Pollen dispersal from exotic eucalypt plantations. Conservation Genetics  6: 253– 257. Google Scholar CrossRef Search ADS   Bayly MJ, Rigault P, Spokevicius A, Ladiges PY, Ades PK, Anderson C, Bossinger G, Merchant A, Udovicic F, Woodrow IE, Tibbits J. 2013. Chloroplast genome analysis of Australian eucalypts – Eucalyptus, Corymbia, Angophora, Allosyncarpia and Stockwellia (Myrtaceae). Molecular Phylogenetics and Evolution  69: 704– 716. Google Scholar CrossRef Search ADS PubMed  Booth TH. 2017. Going nowhere fast: a review of seed dispersal in eucalypts. Australian Journal of Botany  65: 401– 410. Google Scholar CrossRef Search ADS   Bradbury D, Tapper SL, Coates D, Hankinson M, McArthur S, Byrn M. 2016a. How does the post-fire facultative seeding strategy impact genetic variation and phylogeographical history? The case of Bossiaea ornata (Fabaceae) in a fire-prone, mediterranean-climate ecosystem. Journal of Biogeography  43: 96– 110. Google Scholar CrossRef Search ADS   Bradbury D, Tapper SL, Coates D, McArthur S, Hankinson M, Byrn M. 2016b. The role of fire and a long-lived soil seed bank in maintaining persistence, genetic diversity and connectivity in a fire-prone landscape. Journal of Biogeography  43: 70– 84. Google Scholar CrossRef Search ADS   Bradshaw FJ. 2015. Reference material for jarrah forest silviculture, Forest Management Series FEM061 . Perth: Department of Parks and Wildlife. Byrne M. 2008. Phylogeny, diversity and evolution of eucalypts. In: Sharma A, Sharma A, eds. Plant genome: biodiversity and evolution. 1E: Phanerogam – angiosperm . Enfield: Science Publishers, 303– 346. Byrne M, Elliott CP, Yates CJ, Coates DJ. 2008. Maintenance of high pollen dispersal in Eucalyptus wandoo, a dominant tree of the fragmented agricultural region in Western Australia. Conservation Genetics  9: 97– 105. Google Scholar CrossRef Search ADS   Byrne M, Steane DA, Joseph L, Yeates DK, Jordan GJ, Crayn D, Aplin K, Cantrill DJ, Cook LG, Crisp MD, Keogh JS, Melville J, Moritz C, Porch N, Kale Sniderman JM, Sunnucks P, Weston PH. 2011. Decline of a biome: Evolution, contraction, fragmentation, extinction and invasion of the Australian mesic zone biota. Journal of Biogeography  38: 1635– 1656. Google Scholar CrossRef Search ADS   Byrne M, Coates DJ, Macdonald BM, van Leeuwen S. 2016. High nuclear genetic differentiation, but low chloroplast diversity in a rare species, Aluta quadrata (Myrtaceae), with a disjunct distribution in the Pilbara, Western Australia. Australian Journal of Botany  64: 687– 695. Google Scholar CrossRef Search ADS   Byrne M, Hankinson M. 2012. Testing the variability of chloroplast sequences for plant phylogeography. Molecular Ecology Resources  60: 569– 574. Byrne M, Hines B. 2004. Phylogeographical analysis of cpDNA variation in Eucalyptus loxophleba (Myrtaceae). Australian Journal of Botany  52: 459– 470. Google Scholar CrossRef Search ADS   Byrne M, Macdonald B, Brand J. 2003. Phylogeography and divergence in the chloroplast genome of Western Australian Sandalwood (Santalum spicatum). Heredity  91: 389– 395. Google Scholar CrossRef Search ADS PubMed  Chapuis MP, Estoup A. 2007. Microsatellite null alleles and estimation of population differentiation. Molecular Biology and Evolution  24: 621– 631. Google Scholar CrossRef Search ADS PubMed  Churchill DM. 1968. The distribution and prehistory of Eucalyptus diversicolor F. Muell., E. marginata Donn ex SM., and E. calophylla R. Br. in relation to rainfall. Australian Journal of Botany  16: 125– 151. Google Scholar CrossRef Search ADS   Cooper CE, Withers PC, Mawson PR, Johnstone R, Kirkby T, Prince J, Bradshaw D, Robertson H. 2003. Characteristics of Marri (Corymbia calophylla) fruits in relation to the foraging behaviour of the Forest Red-tailed Black Cockatoo (Calyptorhynchus banksii naso). Journal of the Royal Society of Western Australia  86: 139– 142. Dalmaris E, Ramalho CE, Poot P, Veneklaas EJ, Byrne M. 2015. A climate change context for the decline of a foundation tree species in south-western Australia: insights from phylogeography and species distribution modelling. Annals of Botany  116: 941– 952. Google Scholar CrossRef Search ADS PubMed  Dieringer D, Schlötterer C. 2003. MICROSATELLITE ANALYSER: a platform independent analysis tool for large microsatellite data sets. Molecular Ecology Notes  3: 167– 169. Google Scholar CrossRef Search ADS   Drummond AJ, Rambaut A. 2007. BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evolutionary Biology  7: 214. Google Scholar CrossRef Search ADS PubMed  Earl D, von Holdt B. 2012. STRUCTURE HARVESTER: a website and program for visualizing STRUCTURE output and implementing the Evanno method. Conservation and Genetic Resources  4: 359– 361. Google Scholar CrossRef Search ADS   Ennos R. 1994. Estimating the relative rates of pollen and seeed migration among plant populations. Heredity  72: 250– 259. Google Scholar CrossRef Search ADS   Excoffier L, Lischer H. 2010. Arlequin suite ver 3.5: a new series of programs to perform population genetic analyses under Linux and Windows. Molecular Ecology Resources  10: 464– 467. Google Scholar CrossRef Search ADS   Felsenstein J. 1989. PHYLIP – Phylogeny Inference Package (Version 3.2). Cladistics  5: 164– 166. Fu YX. 1997. Statistical tests of neutrality of mutations against population growth, hitchhiking and background selection. Genetics  147: 915– 925. Google Scholar PubMed  Fu YX, Li WH. 1993. Statistical tests of neutrality of mutations. Genetics  133: 693– 709. Google Scholar PubMed  Gonzalez-Orozco C, Pollock L, Thornhill A, Mishler B, Knerr N, Laffan S, Miller J, Rosauer D, Faith D, Nipperess D, Kujala H, Linke S, Butt N, Külheim C, Crisp M, Gruber B. 2016. Phylogenetic approaches reveal biodiversity threats under climate change. Climate Change  6: 1110– 1114. Graur D, Li W. 2000. Fundamentals of molecular evolution . Sunderland: Sinauer Associates Inc. Gugger PF, Ikegami M, Sork VL. 2013. Influence of late Quaternary climate change on present patterns of genetic variation in valley oak, Quercus lobata Née. Molecular Ecology  22: 3598– 3612. Google Scholar CrossRef Search ADS PubMed  Hewitt G. 2000. The genetic legacy of the Quaternary ice ages. Nature  405: 907– 913. Google Scholar CrossRef Search ADS PubMed  Hewitt GM. 2001. Speciation, hybrid zones and phylogeography - or seeing genes in space and time. Molecular Ecology  10: 537– 549. Google Scholar CrossRef Search ADS PubMed  Hopper SD, Gioia P. 2004. The Southwest Australian Floristic Region: evolution and conservation of a global hot spot of biodiversity. Annual Review of Ecology, Evolution, and Systematics  35: 623– 650. Google Scholar CrossRef Search ADS   Jakobsson M, Rosenberg NA. 2007. CLUMPP: a cluster matching and permutation program for dealing with label switching and multimodality in analysis of population structure. Bioinformatics  23: 1801– 1806. Google Scholar CrossRef Search ADS PubMed  Jones M, Shepherd M, Henry R, Delves A. 2008. Pollen flow in Eucalyptus grandis determined by paternity analysis using microsatellite markers. Tree Genetics & Genomes  4: 37– 47. Google Scholar CrossRef Search ADS   Librado P, Rozas J. 2009. DnaSP v5: a software for comprehensive analysis of DNA polymorphism data. Bioinformatics  25: 1451– 1452. Google Scholar CrossRef Search ADS PubMed  Llorens TM, Yates CJ, Byrne M, Nistelberger HM. 2013. Complex interactions between remnant shape and the mating system strongly influence reproductive output and progeny performance in fragmented populations of a bird-pollinated shrub. Biological Conservation  164: 129– 139. Google Scholar CrossRef Search ADS   Llorens TM, Tapper SL, Coates DJ, McArthur S, Hankinson M, Byrne M. 2016. Does population distribution matter? Influence of a patchy versus continuous distribution on genetic patterns in a wind-pollinated shrub. Journal of Biogeography  44: 361– 374. Google Scholar CrossRef Search ADS   Lowe WH, Allendorf FW. 2010. What can genetics tell us about population connectivity? Molecular Ecology  19: 3038– 3051. Google Scholar CrossRef Search ADS PubMed  Maddison WP, Maddison DR. 2016. Mesquite: a modular system for evolutionary analysis, Version 3.04. Available at http://mesquiteproject.org, accessed 23 October 2017. Martin HA. 2006. Cenozoic climatic change and the development of the arid vegetation in Australia. Journal of Arid Environments  66: 533– 563. Google Scholar CrossRef Search ADS   McLachlan JS, Clark JS, Manos PS. 2009. Molecular indicators of tree migration capacity under rapid climate change. Ecology  86: 2088– 2098. Google Scholar CrossRef Search ADS   Mimura M, Barbour RC, Potts BM, Vaillancourt RE, Watanabe KN. 2009. Comparison of contemporary mating patterns in continuous and fragmented Eucalyptus globulus native forests. Molecular Ecology  18: 4180– 4192. Google Scholar CrossRef Search ADS PubMed  Nevill PG, Bradbury D, Williams A, Tomlinson S, Krauss SL. 2014. Genetic and palaeo-climatic evidence for widespread persistence of the coastal tree species Eucalyptus gomphocephala (Myrtaceae) during the Last Glacial Maximum. Annals of Botany  113: 55– 67. Google Scholar CrossRef Search ADS PubMed  Nistelberger H, Gibson N, Macdonald B, Tapper SL, Byrne M. 2014. Phylogeographic evidence for two mesic refugia in a biodiversity hotspot. Heredity  113: 454– 463. Google Scholar CrossRef Search ADS PubMed  O’Brien EK, Krauss SL. 2010. Testing the home-site advantage in forest trees on disturbed and undisturbed sites. Restoration Ecology  18: 359– 372. Google Scholar CrossRef Search ADS   van Oosterhout C, Hutchinson WF, Wills DPM, Shipley P. 2004. MICRO-CHECKER: software for identifying and correcting genotyping errors in microsatellite data. Molecular Ecology Notes  4: 535– 538. Google Scholar CrossRef Search ADS   Orsini L, Andrew R, Eizaguirre C. 2013. Evolutionary ecological genomics. Molecular Ecology  22: 527– 531. Google Scholar CrossRef Search ADS PubMed  Pannell JR, Dorken ME. 2006. Colonisations as a common denominator in plant metapopulations and range expansions: effects on genetic diversity and sexual systems. Landscape Ecology  21: 837– 848. Google Scholar CrossRef Search ADS   Peakall R, Smouse PE. 2012. GenALEx 6.5: genetic analysis in excel. Population genetic software for teaching and research-an update. Bioinformatics  28: 2537– 2539. Google Scholar CrossRef Search ADS PubMed  Petit RJ, Aguinagalde I, de Beaulieu JL, Bittkau C, Brewer S, Cheddadi R, Ennos R, Fineschi S, Grivet D, Lascoux M, Mohanty A, Müller-Starck G, Demesure-Musch B, Palmé A, Martín JP, Rendell S, Vendramin GG. 2003. Glacial refugia: hotspots but not melting pots of genetic diversity. Science (New York, N.Y.)  300: 1563– 1565. Google Scholar CrossRef Search ADS PubMed  Petit RJ, Hampe A. 2006. Some evolutionary consequences of being a tree. Annual Review of Ecology, Evolution, and Systematics  37: 187– 214. Google Scholar CrossRef Search ADS   Pickett EJ. 1997. A Late Pleistocene and Holocene vegetation history of three lacustrine sequences from the Swan Coastal Plain . PhD Thesis, The University of Western Australia. Piry S, Luikart G, Cornuet JM. 1999. Computer note. BOTTLENECK: a computer program for detecting recent reductions in the effective size using allele frequency data. Journal of Heredity  90: 502– 503. Google Scholar CrossRef Search ADS   Pons O, Petit RJ. 1996. Measuring and testing genetic differentiation with ordered versus unordered alleles. Genetics  144: 1237– 1245. Google Scholar PubMed  Posada D. 2008. jModelTest: phylogenetic model averaging. Molecular Biology and Evolution  25: 1253– 1256. Google Scholar CrossRef Search ADS PubMed  Prideaux GJ, Gully GA, Couzens AMC, Ayliffe LK, Jankowski NR, Jacobs Z, Roberts RG, Hellstrom JC, Gagan MK, Hatcher LM. 2010. Timing and dynamics of Late Pleistocene mammal extinctions in southwestern Australia. Proceedings of the National Academy of Sciences of the United States of America  107: 22157– 22162. Google Scholar CrossRef Search ADS PubMed  Pritchard JK, Stephens M, Donnelly P. 2000. Inference of population structure using multilocus genotype data. Genetics  155: 945– 959. Google Scholar PubMed  Ramos-Onsins SE, Rozas J. 2002. Statistical properties of new neutrality tests against population growth. Molecular Biology and Evolution  19: 2092– 2100. Google Scholar CrossRef Search ADS PubMed  Rousset F. 1997. Genetic differentiation and estimation of gene flow from F-statistics under isolation by distance. Genetics  145: 1219– 1228. Google Scholar PubMed  Rousset F. 2008. genepop’007: a complete re-implementation of the genepop software for Windows and Linux. Molecular Ecology Resources  8: 103– 106. Google Scholar CrossRef Search ADS PubMed  Sampson JF, Byrne M. 2008. Outcrossing between an agroforestry plantation and remnant native populations of Eucalyptus loxophleba. Molecular Ecology  17: 2769– 2781. Google Scholar CrossRef Search ADS PubMed  Schaal BA, Haywood DA, Olsen KM, Rauscher JT, Smith WA 1998. Phylogeographic studies in plants, problems and prospects. Molecular Ecology  7: 465– 474. Google Scholar CrossRef Search ADS   Slatkin M. 1985. Gene flow in natural populations. Annual Review of Ecology, Evolution, and Systematics  16: 393– 430. Google Scholar CrossRef Search ADS   Soltis DE, Morris AB, McLachlan JS, Manos PS, Soltis PS. 2006. Comparative phylogeography of unglaciated eastern North America. Molecular Ecology  15: 4261– 4293. Google Scholar CrossRef Search ADS PubMed  Sork VL, Gugger PF, Chen JM, Werth S. 2016. Evolutionary lessons from California plant phylogeography. Proceedings of the National Academy of Sciences of the United States of America  113: 8064– 8071. Google Scholar CrossRef Search ADS PubMed  Tajima F. 1989. Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics  123: 585– 595. Google Scholar PubMed  Telles MP, Diniz-Filho JA. 2005. Multiple Mantel tests and isolation-by-distance, taking into account long-term historical divergence. Genetics and Molecular Research  4: 742– 748. Google Scholar PubMed  Thornhill AH, Macphail M. 2012. Fossil myrtaceous pollen as evidence for the evolutionary history of Myrtaceae: a review of fossil Myrtaceidites species. Review of Palaeobotany and Palynology  176– 177: 1– 23. Vekemans X, Hardy OJ. 2004. New insights from fine-scale spatial genetic structure analyses in plant populations. Molecular Ecology  13: 921– 935. Google Scholar CrossRef Search ADS PubMed  Wang IJ. 2010. Recognizing the temporal distinctions between landscape genetics and phylogeography. Molecular Ecology  19: 2605– 2608. Google Scholar CrossRef Search ADS PubMed  Wang IJ, Bradburd GS. 2014. Isolation by environment. Molecular Ecology  23: 5649– 5662. Google Scholar CrossRef Search ADS PubMed  Wardell-Johnson GW. 2000. Responses of forest eucalypts to moderate and high intensity fire in the Tingle Mosaic, south-western Australia: comparisons between locally endemic and regionally distributed species. Austral Ecology  25: 409– 421. Google Scholar CrossRef Search ADS   Wheeler MA, Byrne M. 2006. Congruence between phylogeographic patterns in cpDNA variation in Eucalyptus marginata (Myrtaceae) and geomorphology of the Darling Plateau, south-west of Western Australia. Australian Journal of Botany  54: 17– 26. Google Scholar CrossRef Search ADS   Wheeler MA, Byrne M, McComb JA. 2003. Little genetic differentiation within the dominant forest tree, Eucalyptus marginata (Myrtaceae) of south-western Australia. Silvae Genetica  52: 254– 259. Whitlock BA, Hale AM, Groff PA. 2010. Intraspecific inversions pose a challenge for the trnH-psbA plant DNA barcode. PLoS ONE  5: e11533. Google Scholar CrossRef Search ADS PubMed  Zachos J, Pagani M, Sloan L, Thomas E, Billups K. 2001. Trends, rhythms, and aberrations in global climate 65 Ma to present. Science (New York, N.Y.)  292: 686– 693. Google Scholar CrossRef Search ADS PubMed  Zhang YH, Wang IJ, Comes HP, Peng H, Qiu YX. 2016. Contributions of historical and contemporary geographic and environmental factors to phylogeographic structure in a Tertiary relict species, Emmenopterys henryi (Rubiaceae). Scientific Reports  6: 24041. Google Scholar CrossRef Search ADS PubMed  Zheng H, Wyrwoll K-H, Li Z, Powell CM. 1998. Onset of aridity in southern Western Australia - a preliminary palaeomagnetic appraisal. Global and Planetary Change  18: 175– 187. Google Scholar CrossRef Search ADS   © 2018 The Linnean Society of London, Biological Journal of the Linnean Society http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Biological Journal of the Linnean Society Oxford University Press

Persistence with episodic range expansion from the early Pleistocene: the distribution of genetic variation in the forest tree Corymbia calophylla (Myrtaceae) in south-western Australia

Loading next page...
 
/lp/ou_press/persistence-with-episodic-range-expansion-from-the-early-pleistocene-PHn0pDNWtp
Publisher
The Linnean Society of London
Copyright
© 2018 The Linnean Society of London, Biological Journal of the Linnean Society
ISSN
0024-4066
eISSN
1095-8312
D.O.I.
10.1093/biolinnean/blx168
Publisher site
See Article on Publisher Site

Abstract

Abstract Phylogeographical patterns of trees in topographically subdued, unglaciated landscapes are under-reported, and might reflect population persistence and the influences of environment and distance over historical (~2.6 Mya to present) and contemporary (recent generations) timescales. We examined this hypothesis using genetic analyses of four slowly evolving non-coding chloroplast sequences and 16 nuclear microsatellites in the tree Corymbia calophylla from south-western Australia, an area that has been unglaciated since the Permian (c. 300–250 Mya). We found strong population differentiation for chloroplast DNA and low differentiation for nuclear loci, consistent with higher gene flow by pollen than by seed. We identified three divergent chloroplast lineages distributed in central, northern and southern regions, and diversifying from the early (c. 3.028 Mya), mid- (c. 0.793 Mya) and late (c. 0.426 Mya) Pleistocene, respectively. Moderate to high nucleotide diversity with population-specific haplotypes supported long-term persistence, but diversification of lineages provided evidence of unexpected episodic range expansion. We suggest this pattern reflects environmental influences of climatic oscillations during progressive drying of south-western Australia from the early Pleistocene. Significant tests for isolation by environment for nuclear loci also supported an influence of contemporary environmental (aridity) conditions on genetic structure, but isolation by distance (IBD) was greater. Significant chloroplast and nuclear IBD suggested distance was a major influence on gene flow at both timescales. INTRODUCTION Gene flow in plants is influenced by geographical features (e.g. topography, distance) and environmental factors (e.g. habitat heterogeneity, climate, pollen dispersal) and these influences can lead to a variety of patterns in the genetic structures of plant species (Wang & Bradburd, 2014). Changes in geographical and environmental factors over multiple timescales, such as contemporary (several recent generations) or longer-term historical periods including the Quaternary (~2.6 Mya to present), can leave signals in the spatial genetic structure of populations (Hewitt, 2001; Wang & Bradburd, 2014). These may be especially apparent in long-lived, sessile organisms such as trees (Petit & Hampe, 2006; Gugger, Ikegami & Sork, 2013; Sork et al., 2016; Zhang et al., 2016) due to the importance of founding effects and drift in sessile species (McLachlan, Clark & Manos, 2009). Climatic oscillations during the Pleistocene (2.588 Mya to 11.700 kya) are considered to have been major drivers of broad phylogeographical patterns (Hewitt, 2000). Many studies of trees, predominantly in the Northern Hemisphere, reveal the overriding influence of population extinctions due to glaciation, with chloroplast haplotype distributions reflecting colonization patterns through rapid expansion following the Last Glacial Maximum (LGM, c. 21 kya) (reviewed by Hewitt, 2000; Petit et al., 2003). Phylogeographical studies of plant species that have been undertaken in unglaciated landscapes have inferred the influence of climatic oscillations (Byrne & Hines, 2004; Nistelberger et al., 2014), topographic heterogeneity (Soltis et al., 2006; Gugger et al., 2013), habitat heterogeneity (Llorens et al., 2016) and fire ephemeral life history (Bradbury et al., 2016a; Bradbury et al., 2016b) on genetic patterns. In unglaciated landscapes, many species show persistence, rather than contraction and expansion, as the main response to climatic oscillations (Byrne & Hines, 2004; Wheeler & Byrne, 2006; Gugger et al., 2013; Llorens et al., 2016) with phylogeographical patterns characterized by localized distribution of diverse chloroplast haplotypes without signs of expansion. However, signals of spatial expansion have been identified in some species (Soltis et al., 2006; Nevill et al., 2014; Dalmaris et al., 2015) and minor range shifts in others (Gugger et al., 2013). Many trees share life history traits (sessile life-form, size, high reproductive output, leptokurtic pollen dispersal, longevity) that affect the mode and tempo of evolution, and trees typically experience slow speciation and extinction (Petit & Hampe, 2006). Longevity can promote persistence of populations by maintaining effective population size, and the production of genetically diverse offspring over long lifetimes can help prevent extinction of local populations during climatic fluctuations (Petit & Hampe, 2006; Sork et al., 2016). The forest environments in the topographically subdued landscape of south-western Australia provide an excellent opportunity to infer responses to historical and contemporary environmental conditions through their influences on genetic structure. The Southwest Australian Floristic Region (SWAFR) is a geologically stable, highly weathered and relatively flat, low plateau, with occasional emergent granite inselbergs, that has not been glaciated since the Permian (300–250 Mya) (Hopper & Gioia, 2004). Following progressive drying since the Pliocene (5–3 Mya), the SWAFR has a seasonal Mediterranean-type climate with the most mesic conditions (600–1500 mm rainfall per annum) occurring in the south-west corner where there is a strong west to east and south-west to north-east gradient of reducing rainfall and increasing summer evaporation (Bradshaw, 2015). Forests are the dominate vegetation type in the more mesic south-west corner (Hopper & Gioia, 2004). Corymbia calophylla (Lindl.) K.D. Hill & L.A.S. Johnson is a long-lived, forest tree that can reach heights of 40–60m with a continuous distribution in mesic forest and wetter woodland regions receiving 800–1300 mm rainfall per annum in the SWAFR. There are some northern, isolated outlier populations confined to wetter soil patches in the transitional rainfall zone (300–600 mm) and C. calophylla is usually replaced by another forest species, Eucalyptus diversicolor F. Muell., in the areas of highest rainfall (1300–1500 mm) (Churchill, 1968). The distribution of C. calophylla includes the Darling Range but elevations are subdued by global standards (Hopper & Gioia, 2004). Most Corymbia are pollinated by insects and birds and it is likely that both vectors pollinate C. calophylla (O’Brien & Krauss, 2010). Seeds are primarily gravity-dispersed with the most likely potential dispersal rate equivalent to about 1–2 m per year (Booth, 2017). Occasional long-distance (>1 km) seed dispersal events may occur following fire storms or cyclones or by red-tailed black and Baudin’s cockatoos (Calyptorhynchus banksia naso and Calyptorhynchus baudinii, respectively) as C. calophylla seeds are a primary food source for these birds (Cooper et al., 2003). Complementary DNA markers that evolve at different rates are often used to distinguish between genetic divergence patterns at different timescales. In plants, bi-parentally inherited, recombining and rapidly evolving nuclear microsatellite loci are often used to reveal contemporary (several recent generations) processes that are influenced by both demographic changes, and pollen- and seed-mediated gene flow. In contrast, the chloroplast genome is slowly evolving, non-recombining and maternally inherited in most angiosperms and therefore genetic signals reflect responses of genealogical lineages over historical timescales (Schaal et al., 1998). The difference in mutation rates between nuclear and chloroplast markers is of the order of 105 (Graur & Li, 2000; Wang, 2010), establishing the utility of chloroplast DNA (cpDNA) for revealing divergence of deeper genealogical lineages (Schaal et al., 1998). We investigated the evolutionary and contemporary history of C. calophylla across its range in SWAFR given that tree species in unglaciated landscapes are expected to reveal insights into evolutionary processes and their influences on genetic structure over periods extending beyond the LGM (Byrne et al., 2011; Sork et al., 2016). We used a combination of non-coding cpDNA sequences and nuclear microsatellites to investigate spatial genetic structures at historical and contemporary timescales, respectively. We hypothesized that there would be evidence of population persistence in the phylogeographical history of C. calophylla in the mesic forests in south-west Australia, an area that has not been glaciated since the Permian (300–250 Mya). We expected geographical distance and the environment (climate) to have been key factors driving historical and contemporary patterns of genetic structure. MATERIAL AND METHODS Study species Corymbia KD Hill & LAS Johnson is a sister lineage to Eucalyptus L’Hér. and Angophora Cav. (Bayly et al., 2013) and one of three sclerophyll genera commonly known as eucalypts. The forest in which C. calophylla occurs is often separated into Northern and Southern Jarrah Forests on the basis of flowering time in the dominant species Eucalyptus marginata (Bradshaw, 2015). The environment is a fire-prone Mediterranean type and although C. calophylla is killed by severe fires, trees can be long-lived (~400 years) as they re-sprout from epicormic shoots (Wardell-Johnson, 2000) following less severe fires. Sampling and genotyping Leaves were sampled from 24 well-dispersed adult plants in each of 27 C. calophylla populations spanning the north–south distribution of the species’ range (Table 1; Fig. 1A). Genomic DNA was extracted from lysed, freeze-dried leaf material by the method described in Byrne et al. (2016). Table 1. Characteristics of 27 Corymbia calophylla populations from south-western Australia used for genotyping Species/population  Code  Latitude  Longitude  Annual rainfall (mm)  Aridity Index*  Landscape character  Northern Forest region  Mt Lesueur  LES  −30.14147200  115.11430600  600  0.44  Wheatbelt Plateau  Moochamulla  MOO  −30.98341700  115.81213900  600  0.44  Wheatbelt Plateau  Julimar  JUL  −31.40622200  116.35602800  600  0.47  Darling Plateau Uplands  Mt Helena  HEL  −31.86822200  116.22297300  900  0.54  Darling Plateau Uplands  Perry Lakes  LAK  −31.94630600  115.78118300  800  0.51  Swan Coastal Plain  Dale  DAL  −32.10136100  116.18927800  1100  0.57  Darling Plateau Uplands  Serpentine  SER  −32.35269400  116.07647200  1200  0.61  Darling Plateau Uplands  Whittaker  KER  −32.55605600  116.03100000  1100  0.62  Darling Plateau Uplands  Wearne  WEA  −32.52066700  116.49900000  700  0.54  Darling Plateau Uplands  Peel  PEE  −32.68466700  115.74266700  900  0.55  Swan Coastal Plain  Pindalup  PID  −32.80163900  116.27741700  1000  0.59  Darling Plateau Uplands  Saddleback  SAD  −32.99702800  116.48466700  700  0.55  Darling Plateau Uplands  Godfrey  GOD  −33.20758300  116.56197200  700  0.55  Darling Plateau Uplands  Yourdaming  YOU  −33.30244400  116.24094400  900  0.59  Darling Plateau Uplands  Eaton  EAT  −33.31119400  115.76336100  900  0.57  Swan Coastal Plain  Lennard  LEN  −33.32613900  115.96827800  1100  0.63  Darling Plateau Uplands  Southern Forest region  Meelup  MEE  −33.57527800  115.08600000  900  0.58  Swan Coastal Plain  Grimwade  GRI  −33.76127800  115.99991700  900  0.61  Darling Plateau Uplands  Mowen  MOW  −33.90788900  115.54211100  1000  0.63  Darling Plateau Uplands  Bramley  BRA  −33.91638900  115.08327800  1200  0.63  Leeuwin Naturalist Coast  Kingston  KIN  −34.08116700  116.33044400  800  0.61  Darling Plateau Uplands  Milylannup  MIL  −34.19658300  115.66205600  1100  0.66  Darling Plateau Uplands  Carey  CAR  −34.41963900  115.82130600  1200  0.67  Darling Plateau Pemberton Slopes  Muir  MUI  −34.65341700  117.49908300  800  0.63  Darling Plateau Uplands  Boorara  BOO  −34.63894400  116.12380600  1300  0.69  Darling Plateau Pemberton Slopes  Beadmore Rd  BEA  −34.81644400  116.65852800  1200  0.70  Darling Plateau Pemberton Slopes  Denmark  DEN  −34.90816700  117.37077800  1100  0.67  Darling Plateau Uplands  Species/population  Code  Latitude  Longitude  Annual rainfall (mm)  Aridity Index*  Landscape character  Northern Forest region  Mt Lesueur  LES  −30.14147200  115.11430600  600  0.44  Wheatbelt Plateau  Moochamulla  MOO  −30.98341700  115.81213900  600  0.44  Wheatbelt Plateau  Julimar  JUL  −31.40622200  116.35602800  600  0.47  Darling Plateau Uplands  Mt Helena  HEL  −31.86822200  116.22297300  900  0.54  Darling Plateau Uplands  Perry Lakes  LAK  −31.94630600  115.78118300  800  0.51  Swan Coastal Plain  Dale  DAL  −32.10136100  116.18927800  1100  0.57  Darling Plateau Uplands  Serpentine  SER  −32.35269400  116.07647200  1200  0.61  Darling Plateau Uplands  Whittaker  KER  −32.55605600  116.03100000  1100  0.62  Darling Plateau Uplands  Wearne  WEA  −32.52066700  116.49900000  700  0.54  Darling Plateau Uplands  Peel  PEE  −32.68466700  115.74266700  900  0.55  Swan Coastal Plain  Pindalup  PID  −32.80163900  116.27741700  1000  0.59  Darling Plateau Uplands  Saddleback  SAD  −32.99702800  116.48466700  700  0.55  Darling Plateau Uplands  Godfrey  GOD  −33.20758300  116.56197200  700  0.55  Darling Plateau Uplands  Yourdaming  YOU  −33.30244400  116.24094400  900  0.59  Darling Plateau Uplands  Eaton  EAT  −33.31119400  115.76336100  900  0.57  Swan Coastal Plain  Lennard  LEN  −33.32613900  115.96827800  1100  0.63  Darling Plateau Uplands  Southern Forest region  Meelup  MEE  −33.57527800  115.08600000  900  0.58  Swan Coastal Plain  Grimwade  GRI  −33.76127800  115.99991700  900  0.61  Darling Plateau Uplands  Mowen  MOW  −33.90788900  115.54211100  1000  0.63  Darling Plateau Uplands  Bramley  BRA  −33.91638900  115.08327800  1200  0.63  Leeuwin Naturalist Coast  Kingston  KIN  −34.08116700  116.33044400  800  0.61  Darling Plateau Uplands  Milylannup  MIL  −34.19658300  115.66205600  1100  0.66  Darling Plateau Uplands  Carey  CAR  −34.41963900  115.82130600  1200  0.67  Darling Plateau Pemberton Slopes  Muir  MUI  −34.65341700  117.49908300  800  0.63  Darling Plateau Uplands  Boorara  BOO  −34.63894400  116.12380600  1300  0.69  Darling Plateau Pemberton Slopes  Beadmore Rd  BEA  −34.81644400  116.65852800  1200  0.70  Darling Plateau Pemberton Slopes  Denmark  DEN  −34.90816700  117.37077800  1100  0.67  Darling Plateau Uplands  *Rainfall/potential evapotranspiration (see Material and Methods). View Large Figure 1. View largeDownload slide Genetic structure of sampled populations of Corymbia calophylla in south-western Australia, inferred from analysis of cpDNA (psbA-trnH, trnQ-rps16, trnG) haplotypes. (A) Distribution of haplotypes overlaid on geographical map of sampling sites. Pie charts show proportion of individuals with a given haplotype. Colour coding of the haplotypes corresponds to those of the median-joining parsimony haplotype network (Fig. 2). Geographical regions containing the Central, North and South chloroplast lineages are identified by dotted lines. The background blue scale represents annual mean rainfall (mm). (B) Maximum-clade-credibility tree from Bayesian phylogenetic analyses with the outgroup C. gummifera, calibrated using a 0–6 My CI calibration applied to a known root age from eucalypt fossils. Figure 1. View largeDownload slide Genetic structure of sampled populations of Corymbia calophylla in south-western Australia, inferred from analysis of cpDNA (psbA-trnH, trnQ-rps16, trnG) haplotypes. (A) Distribution of haplotypes overlaid on geographical map of sampling sites. Pie charts show proportion of individuals with a given haplotype. Colour coding of the haplotypes corresponds to those of the median-joining parsimony haplotype network (Fig. 2). Geographical regions containing the Central, North and South chloroplast lineages are identified by dotted lines. The background blue scale represents annual mean rainfall (mm). (B) Maximum-clade-credibility tree from Bayesian phylogenetic analyses with the outgroup C. gummifera, calibrated using a 0–6 My CI calibration applied to a known root age from eucalypt fossils. The chloroplast psbA-trnH and trnQ-rps16 intergenic spacer regions and the trnG intron were selected for amplification and sequencing in eight random samples from each of the 27 study populations. Sequences were amplified according to Byrne & Hankinson (2012) and sequenced via Macrogen Inc. (Seoul, South Korea). SEQUENCHER 5.0 (Genecodes Corp., Ann Arbor, MI, USA) was used to edit miscalls, and to align and trim sequences. All three cpDNA regions were concatenated in MESQUITE 3.04 (Maddison & Maddison, 2016) to a total sequence length of 2451 bp. One 20-bp and one 21-bp inversion were uncovered in the psbA-trnH region. Following Whitlock, Hale & Groff (2010), one configuration of inversions was replaced with its reverse-complement and coded as a single transversion. Microsatellite loci were developed from a representative individual of C. calophylla according to Byrne et al. (2016) and screened in eight individuals from six populations (Supporting Information Table S1). Loci were amplified using the Multiplex60 PCR program of the Qiagen Multiplex kit (Hilden, Germany), separated on an Applied Biosystems (Foster City, CA, USA) 3730 capillary sequencer, and 648 individuals (24 per population) were genotyped at 16 nuclear microsatellite loci using GENEMAPPER version 5.0 (Applied Biosystems). Chloroplast diversity and phylogeographical structure analysis Chloroplast haplotypes were identified using DNAsp 5.1.1 (Librado & Rozas, 2009; Supporting Information Table S2). Nucleotide (π) and haplotype (HD) diversity indices were calculated using ARLEQUIN 3.5.2.2 (Excoffier & Lischer, 2010). The number of haplotypes per population (Happop), HD and π were compared between groups using t-tests. To visualize the phylogenetic relationships among all chloroplast haplotypes, a median-joining maximum parsimony (MJMP) network was constructed in NETWORK 5.0 (Bandelt, Forster & Röhl, 1999). Population differentiation estimates (NST, GST) were calculated using PERMUT 2.0 (Pons & Petit, 1996). GST considers only allele frequencies in analysis (unordered analysis) but NST also takes distances between alleles into account (ordered analysis). NST > GST indicates a phylogenetic signal, i.e. that haplotypes within populations are on average more similar to each other than to a random set of haplotypes within the species, suggesting that genetic drift alone cannot explain the pattern and that the mutational process is influencing genetic differentiation among populations. A hierarchical analysis of molecular variance (AMOVA) was conducted in ARLEQUIN to estimate differentiation among and within populations and regions or chloroplast lineages, with 1000 permutations. Isolation by distance (IBD) effects were assessed by Mantel tests of pairwise population differentiation [FST/(1 − FST)] estimated in ARLEQUIN against the logarithm (log10) of pairwise geographical distances. Molecular dating and phylogeny reconstruction were simultaneously completed via a strict clock Bayesian analysis in BEAST 1.7.5 (Drummond & Rambaut, 2007). The substitution model was set to GTR+I+G as inferred as the best fit to the data by jModelTest (Posada, 2008). Divergence times were estimated under a strict clock model (i.e. with uniform rates across branches). Date estimates were constrained via the inclusion of a root calibration using the estimated time since the most recent common ancestor (TRMCA) of C. calophylla and Corymbia gummifera (Gaertn.) K.D.Hill & L.A.S.Johnson of 3.0 My. This calibration date was based on an unpublished dated version of the Gonzalez-Orozco et al. (2016) eucalypt phylogeny that was calibrated by Andrew Thornhill (pers. comm) using the same eucalypt fossils as previously defined and used by Thornhill & Macphail (2012). In the absence of a 95% confidence interval for the TRMCA, dating was conducted using three different hypothetical confidence intervals applied to the root age calibration, 2—4 My, 1—5 My and 0—6 My, with four independent runs of 10 million generations carried out for each of the three scenarios, sampling every 1000 generations. Convergence was assessed in Tracer 1.6 (Drummond & Rambaut, 2007) and trees were combined using LOGCOMBINER 1.6.2. TREEANNOTATOR 1.6.2 (Drummond & Rambaut, 2007) was used to identify a maximum clade credibility tree. Tests for neutrality and expansion were calculated with Tajima’s D (Tajima, 1989) and Fu’s Fs (Fu, 1997) in ARLEQUIN, and R2 (Ramos-Onsins & Rozas, 2002), and F and D (Fu & Li, 1993) calculated in DNAsp using C. gummifera as an outgroup. To infer spatial and demographic history, we used mismatch distribution analyses in ARLEQUIN to compare observed differences with model expectations assuming spatial and demographic expansion. Goodness-of-fit was tested with Harpending’s raggedness index (HRag) and the sum of squared differences (SSD). Nuclear diversity and differentiation analysis Tests for stutter bands and large allele dropout were conducted using MICROCHECKER 2.2.3 (van Oosterhout et al., 2004). Global tests of population heterozygote deficiency, and tests of linkage disequilibrium among pairs of loci, were performed with GENEPOP 4.2 (Rousset, 2008). The frequency of null alleles was estimated using FREE NA (Chapuis & Estoup, 2007). Mean multilocus genetic diversity parameters per population were estimated using GENALEX 6.501 (Peakall & Smouse, 2012). To identify genetic evidence of recent population bottlenecks, we used the Wilcoxon sign-rank test of excess heterozygosity under a two-phase model (TPM) of 70% stepwise mutation model (SMM) and 30% infinite allele model (IAM) with 30% variance, and mode shifts from an L-shaped distribution with 1000 replicates using BOTTLENECK 1.2.02 (Piry, Luikart & Cornuet, 1999). Values of among-population differentiation (FST) per locus were calculated using FREENA, correcting for null alleles by the excluding null alleles (ENA) method, using 1000 bootstraps to generate 95% confidence intervals (Chapuis & Estoup, 2007). Pairwise genetic differentiation was estimated with FST using GENALEX with statistical testing by random permutations. Genetic variation among populations was partitioned using AMOVA in ARLEQUIN. An indirect measure of gene flow among populations was estimated using the private allele frequency method of (Slatkin, 1985) in GENEPOP. To estimate the relative contribution to total gene flow of seeds versus pollen, we calculated the pollen-to-seed gene flow ratio of Ennos (1994) as [A(1 + FIS) − 2C]/C, where A = [(1/FST) − 1] and C = [(1/GST − 1)]. Values for FIS and FST were obtained from GENEPOP using nuclear microsatellite data, and GST from chloroplast data using PERMUT. IBD and isolation by environment (IBE) were investigated by testing for correlations between log10 pairwise geographical distances and genetic differentiation (FST/(1 − FST) and the normalized differences of the ‘Aridity Index’ among populations, respectively, using a Mantel procedure in GENALEX. The ‘Aridity Index’ (annual mean monthly ratio of precipitation to potential evaporation) for each population was obtained from the Atlas of Living Australia (http://www.ala.org.au/, accessed 19 October 2016) and used as an environmental variable to indicate the local climate of each population. The Aridity Index is a numerical environmental indicator of the degree of dryness of the climate at a given location gridded at 0.01° (~1 km) and is intended to describe the magnitude of potential water limitation and biological stress influencing biotic distribution patterns. To compare the influence of IBD and IBE, we also used Mantel tests under a multiple correlation and regression design to simultaneously evaluate variables, with the proportion of the total explained variance due to geographial distance or aridity estimated following Telles & Diniz-Filho (2005) by removing each variable alternately. To assess genetic relationships among populations, we performed a phenetic analysis based on the CS Chord genetic distance calculated using MSA 4.05 (Dieringer & Schlötterer, 2003) and constructed an unrooted neighbor-joining (NJ) tree using PHYLIP 3.69 (Felsenstein, 1989) with clustering patterns validated with 1000 bootstraps. Patterns of population genetic structure were visualized using a Bayesian model-based approach implemented in STRUCTURE 2.3.4 (Pritchard, Stephens & Donnelly, 2000). STRUCTURE was used to identify the most likely number (K) of genetically homogeneous clusters, and to assign proportionate membership (q) of individuals to clusters. Analyses used no prior knowledge, the admixture ancestry model and a burn-in of 150 000 with 200 000 iterations for Markov chain Monte Carlo parameters. Analyses were done for K = 1–27 and assessed using the similarity coefficient h′ in CLUMPP 1.1.2 (Jakobsson & Rosenberg, 2007). The optimum K of genetically homogeneous clusters was determined from 20 replicate runs for each value of K using the statistic ΔK from the program STRUCTURE HARVESTER 0.6.93 (Earl & von Holdt, 2012). RESULTS Chloroplast diversity and phylogeographical structure Twenty-nine unique haplotypes were found among individuals (Fig. 2; Table S2). Overall nucleotide diversity was moderate (π = 0.086, SD = 0.044, Table 2) and haplotype diversity was high (HD = 0.931, SD = 0.007). Most haplotypes (N = 24, 82.8%; Fig. 1) were found in single populations and within-population haplotype diversity was low (hS = 0.200, SD = 0.045). Unique haplotypes were distributed throughout the species’ range rather than in restricted groups of peripheral populations. Two haplotypes were more widely distributed: H01 occurred in six populations and H02 in five populations. Table 2. Genetic diversity and differentiation indices of chloroplast haplotypes for 27 populations of Corymbia calophylla from south-western Australia   All populations  Central lineage  North lineage  South lineage  Number of populations  N  27  4  3  20  Number of samples  n  216  32  24  160  Number of haplotypes  HN  29  9  4  16  Mean number of haplotypes per population  Happop  1.593 (0.134)  2.250 (0.479)  1.333 (0.333  1.500 (0.136)  Haplotype diversity  HD  0.931 (0.007)  0.859 (0.029)  0.750 (0.040)  0.884 (0.011)  Nucleotide diversity  π  0.086 (0.044)  0.126 (0.065)  0.052 (0.029)  0.029 (0.017)  Within-population haplotype diversity  hs  0.200 (0.045)  0.375 (0.128)  0.179 (0.179)  0.168 (0.048)  Population differentiation (unordered)  GST  0.791 (0.047)  0.625 (0.129)  0.856 (0.180)  0.818 (0.052)  Population differentiation (ordered)  NST  0.836 (0.053)  0.559 (0.165)  0.821 (0.173)  0.833 (0.072)  Phylogeogenetic signal within populations  NST > GST  NS  NS  NS  NS    All populations  Central lineage  North lineage  South lineage  Number of populations  N  27  4  3  20  Number of samples  n  216  32  24  160  Number of haplotypes  HN  29  9  4  16  Mean number of haplotypes per population  Happop  1.593 (0.134)  2.250 (0.479)  1.333 (0.333  1.500 (0.136)  Haplotype diversity  HD  0.931 (0.007)  0.859 (0.029)  0.750 (0.040)  0.884 (0.011)  Nucleotide diversity  π  0.086 (0.044)  0.126 (0.065)  0.052 (0.029)  0.029 (0.017)  Within-population haplotype diversity  hs  0.200 (0.045)  0.375 (0.128)  0.179 (0.179)  0.168 (0.048)  Population differentiation (unordered)  GST  0.791 (0.047)  0.625 (0.129)  0.856 (0.180)  0.818 (0.052)  Population differentiation (ordered)  NST  0.836 (0.053)  0.559 (0.165)  0.821 (0.173)  0.833 (0.072)  Phylogeogenetic signal within populations  NST > GST  NS  NS  NS  NS  Standard deviations in parentheses. NS, non-significant. View Large Figure 2. View largeDownload slide Median-joining maximum parsimony haplotype network of Corymbia calophylla from south-western Australia, based on cpDNA sequences. Groups are enclosed in boxes. Circle size in the network is relative to haplotype frequency and boxes on the network branches represent mutations. Figure 2. View largeDownload slide Median-joining maximum parsimony haplotype network of Corymbia calophylla from south-western Australia, based on cpDNA sequences. Groups are enclosed in boxes. Circle size in the network is relative to haplotype frequency and boxes on the network branches represent mutations. The haplotype network of phylogenetic relationships showed an asymmetrical structure revealing three haplotype groups: a group of divergent haplotypes that were found geographically in populations in the central–west of the species’ range (henceforth Central lineage), a group of moderately divergent haplotypes found in populations in the north (North lineage) and a radiation of weakly diverged haplotypes found in populations in the south (South lineage). Only one population had haplotypes from more than one lineage; WEA contained seven individuals with a South lineage haplotype (H24) and one individual with a Central lineage haplotype (H29). Nucleotide diversity was higher in the Central lineage (π = 0.126, SD = 0.065) compared to the North (π = 0.052, SD = 0.029) or South lineages (π = 0.029, SD = 0.017) but haplotype diversities (HD) were similar (0.859, SD = 0.029; 0.750, SD = 0.040; 0.884, SD = 0.011, respectively). Within-population haplotype diversity was highest in the Central lineage (hs = 0.375, SD = 0.128; North, hs = 0.179, SD = 0.179; South, hs = 0.168, SD = 0.048). Both unordered (GST) and ordered (NST) estimates indicated significant population differentiation (Table 2). However, there was no significant phylogenetic signal (NST > GST) within populations for cpDNA over all populations (GST/NST = 0.704/0.802, P < 0.01) or within any lineage. Non-hierarchical AMOVA analyses over all populations indicated significant differentiation of populations with most of the variance among (83.63%) rather than within (16.37%) populations. There was significant support for IBD (r = 0.192, P < 0.05). Hierarchical analyses of populations partitioned according to the chloroplast lineages suggested by the haplotype network, or the geographical regions they aligned with, showed strong structuring with the majority of variance between lineages (65.59%; Table 3) rather than among populations within lineages (24.13%) or within populations (10.28%). Table 3. Analysis of molecular variance (AMOVA) of Corymbia calophylla for chloroplast haplotype data from 27 populations from Central, North and South chloroplast lineages*, and for nuclear microsatellite data from 27 populations separated into Northern Forest and Southern Forest regions in south-western Australia Source of variation  d.f.  SS  Variance component  Variance (%)  Chloroplast haplotypes  Among lineages  2  473.6  4.919  65.59  Among populations within lineages  24  366.002  1.81  24.13  Within populations  189  145.75  0.771  10.28  Nuclear microsatellites  Among forest regions  1  44.026  0.056  1.48  Among populations within forest regions  25  217.994  0.105  2.76  Within populations    4902.124  3.656  95.76  Source of variation  d.f.  SS  Variance component  Variance (%)  Chloroplast haplotypes  Among lineages  2  473.6  4.919  65.59  Among populations within lineages  24  366.002  1.81  24.13  Within populations  189  145.75  0.771  10.28  Nuclear microsatellites  Among forest regions  1  44.026  0.056  1.48  Among populations within forest regions  25  217.994  0.105  2.76  Within populations    4902.124  3.656  95.76  *Central, North and South lineages were found in geographical Central, North and South regions as described in the Results. View Large Bayesian analyses of the combined chloroplast sequences revealed early divergence of two haplotypes found in the central part of the geographical distribution followed by three chloroplast lineages of sequentially younger ages that correspond to populations (with the exception of WEA) in the geographical centre–west (henceforth Central region), north (North region), and centre–east and south (South region) of the species distribution (Fig. 1B). These lineages and geographical regions aligned with the Central, North and South groups of the haplotype network (Fig. 2), respectively. Trees and dates derived from the 0–6, 1–5, and 2–4 My CI applied to the root age calibration were similar with a small difference in the location of the clade containing Peel population (PEE) in the 2–4 My CI scenario (Supporting Information Fig. S1). Using the 0–6 My CI, the TMRCA of all haplotypes was 3.028 Mya [CI 2.307–3.842 Mya, posterior probability (PP) = 1] in the late Pliocene. Bayesian analyses revealed strong support for divergence of North and South lineages from the Central lineage haplotypes in the early Pleistocene at 1.982 Mya (CI 1.422–2.637 Mya, PP = 1). The Central lineage is characterized by haplotypes found in single populations with deep divergences up to 3.028 Mya. There was strong support for divergence of North and South lineages at 0.793 Mya (CI 0.517–1.129 Mya, PP = 1) and for later divergence of haplotypes within the South lineage from 0.426 Mya (CI 0.256–0.625 Mya; PP = 1). Across all populations, neutrality tests were not significant with the exception of Tajima’s D (Table 4). When chloroplast lineages were analysed separately, neutrality statistics were low or negative and Tajima’s D was significant in the South lineage only. Mismatch analyses indicated spatial expansion for the South and North lineages but not for the Central lineage. Considering the results together, tests for C. calophylla suggest spatial expansion in the South lineage found in the geographically south region. Table 4. Tests for neutrality and for demographic and spatial expansion based on Corymbia calophylla chloroplast haplotypes sampled from 27 populations in south-western Australia Chloroplast lineage  All populations  Central  North  South  Number of populations  27  4  3  20  Number of samples  216  32  24  160  Tajima’s D  −1.498 (P < 0.05)  −0.866 (P > 0.05)  1.622 (P > 0.05)  −1.497 (P < 0.01)  Ramos-Onsins and Rozas R2  0.043 (P = 0.054)  0.093 (P > 0.05)  0.196 (P > 0.05)  0.060 (P > 0.05)  Fu’s FS  0.439 (P > 0.05)  7.408 (P > 0.05)  6.942 (P > 0.05)  −1.598 (P > 0.05)  Fu and Li’s D*  0.693 (P > 0.10)  1.338 (P > 0.10)  1.662 (P < 0.05)  −2.036 (P > 0.05)  Fu and Li’s F*  −0.476 (P > 0.10)  0.615 (P > 0.10)  1.985 (P < 0.05)  −2.179 (P > 0.05)  Demographic expansion (SSD)†  0.016 (P > 0.05)  0.095 (P < 0.01)  0.112 (P < 0.01)  0.016 (P < 0.05)  Demographic expansion (HRag)†  0.018 (P > 0.05)  0.240 (P < 0.01)  0.275 (P < 0.01)  0.043 (P < 0.05)  Spatial expansion (SSD)†  0.016 (P > 0.05)  0.082 (P < 0.05)  0.067 (P > 0.05)  0.014 (P > 0.05)  Spatial expansion (HRag)†  0.018 (P > 0.05)  0.240 (P < 0.05)  0.275 (P > 0.05)  0.043 (P > 0.05)  Chloroplast lineage  All populations  Central  North  South  Number of populations  27  4  3  20  Number of samples  216  32  24  160  Tajima’s D  −1.498 (P < 0.05)  −0.866 (P > 0.05)  1.622 (P > 0.05)  −1.497 (P < 0.01)  Ramos-Onsins and Rozas R2  0.043 (P = 0.054)  0.093 (P > 0.05)  0.196 (P > 0.05)  0.060 (P > 0.05)  Fu’s FS  0.439 (P > 0.05)  7.408 (P > 0.05)  6.942 (P > 0.05)  −1.598 (P > 0.05)  Fu and Li’s D*  0.693 (P > 0.10)  1.338 (P > 0.10)  1.662 (P < 0.05)  −2.036 (P > 0.05)  Fu and Li’s F*  −0.476 (P > 0.10)  0.615 (P > 0.10)  1.985 (P < 0.05)  −2.179 (P > 0.05)  Demographic expansion (SSD)†  0.016 (P > 0.05)  0.095 (P < 0.01)  0.112 (P < 0.01)  0.016 (P < 0.05)  Demographic expansion (HRag)†  0.018 (P > 0.05)  0.240 (P < 0.01)  0.275 (P < 0.01)  0.043 (P < 0.05)  Spatial expansion (SSD)†  0.016 (P > 0.05)  0.082 (P < 0.05)  0.067 (P > 0.05)  0.014 (P > 0.05)  Spatial expansion (HRag)†  0.018 (P > 0.05)  0.240 (P < 0.05)  0.275 (P > 0.05)  0.043 (P > 0.05)  Tajima’s D is expected to be negative when there is an excess of rare, less diverged mutations, which may be due to sudden, recent population growth. R2 is expected to be low when there is an excess or rare, less diverged mutations, which may be due to sudden, recent population growth. FS is expected to be negative when there is an excess of rare mutations. *Corymbia gummifera used as an outgroup. †Demographic and spatial expansion models test for deviation from the distribution expected under expansion and therefore P > 0.05 supports a scenario of expansion. View Large Nuclear diversity and structure No evidence of stutter or large allele dropout was detected. We found 25 (out of a possible 3240) instances of significant composite genotypic disequilibrium within populations. The loci involved were not consistent across populations. We detected 27 frequencies of null alleles significantly greater than 0.05 over all loci and populations (out of a possible 432; Supporting Information Table S3). Comparison of FST (95% CI) estimates showed that null alleles did not cause significant bias (without ENA: 0.032, 0.029–0.037; with ENA: 0.033, 0.029–0.037). Therefore, all loci were included in analyses. A total of 319 alleles were amplified from 16 microsatellite loci among all individuals. Genetic diversity parameters were high (Table 5) and inbreeding coefficients were generally low (F = −0.002 to 0.151), as expected for a mixed mating species (Byrne, 2008). Bottleneck analysis detected significant probabilities of heterozygote excess for six populations from the Northern Forest region (LAK, WEA, PEE, SAD, GOD, YOU; Table 5) and three Southern Forest populations (MEE, BRA, MIL) using the IAM model. Table 5. Nuclear microsatellite diversity of 27 Corymbia calophylla populations from south-western Australia Population  Code  n  A  Np  Ho  UHe  F  Bottleneck test†  Northern Forest region  Mt Lesueur  LES  24  7.50 (0.84)  3  0.661 (0.052)  0.671 (0.052)  0.000 (0.029)  0.0649  Moochamulla  MOO  24  8.31 (1.08)  2  0.681 (0.049)  0.681 (0.045)  −0.002 (0.026)  0.3161  Julimar  JUL  24  8.88 (1.00)  3  0.648 (0.061)  0.714 (0.048)  0.101 (0.060)  0.0523  Mt Helena  HEL  24  9.13 (0.93)  2  0.675 (0.049)  0.735 (0.032)  0.089 (0.047)  0.1489  Perry Lakes  LAK  24  7.88 (0.96)  2  0.638 (0.057)  0.686 (0.051)  0.068 (0.040)  0.0145*  Dale  DAL  23  8.50 (0.89)  2  0.655 (0.047)  0.699 (0.046)  0.046 (0.042)  0.2809  Serpentine  SER  22  8.44 (0.97)  1  0.605 (0.035)  0.689 (0.039)  0.111 (0.034)*  0.4105  Whittaker  KER  24  9.06 (0.94)  2  0.651 (0.061)  0.717 (0.047)  0.122 (0.065)  0.1261  Wearne  WEA  24  9.00 (0.94)  0  0.708 (0.049)  0.745 (0.038)  0.062 (0.037)  0.0046*  Peel  PEE  24  9.06 (1.22)  4  0.692 (0.046)  0.735 (0.044)  0.047 (0.035)  0.0000*  Pindalup  PID  19  7.94 (0.88)  3  0.635 (0.049)  0.693 (0.053)  0.074 (0.040)  0.2983  Saddleback  SAD  23  8.63 (0.99)  0  0.664 (0.042)  0.729 (0.041)  0.080 (0.034)*  0.0026*  Godfrey  GOD  24  7.94 (0.88)  1  0.614 (0.044)  0.701 (0.043)  0.112 (0.041)*  0.0008*  Yourdaming  YOU  23  8.63 (1.11)  2  0.612 (0.053)  0.713 (0.049)  0.151 (0.038)*  0.0055*  Eaton  EAT  24  8.88 (1.08)  4  0.625 (0.054)  0.715 (0.047)  0.123 (0.037)*  0.1156  Lennard  LEN  24  9.31 (1.23)  2  0.646 (0.063)  0.711 (0.053)  0.096 (0.043)*  0.0523  Southern Forest region  Meelup  MEE  24  8.38 (0.88)  0  0.654 (0.055)  0.707 (0.044)  0.082 (0.039)*  0.0222*  Grimwade  GRI  24  8.31 (0.83)  5  0.628 (0.049)  0.710 (0.047)  0.101 (0.045)*  0.1261  Mowen  MOW  24  9.38 (1.03)  1  0.695 (0.045)  0.719 (0.045)  0.026 (0.026)  0.2477  Bramley  BRA  24  9.13 (0.89)  1  0.740 (0.037)  0.751 (0.036)  0.004 (0.038)  0.0327*  Kingston  KIN  24  8.81 (1.01)  3  0.625 (0.050)  0.704 (0.048)  0.100 (0.046)*  0.2019  Milylannup  MIL  24  9.06 (1.11)  5  0.688 (0.052)  0.722 (0.050)  0.040 (0.034)  0.0327*  Carey  CAR  24  8.94 (0.87)  1  0.698 (0.052)  0.709 (0.045)  0.013 (0.033)  0.3342  Muir  MUI  23  8.69 (1.00)    0.691 (0.054)  0.706 (0.048)  0.018 (0.036)  0.1489  Boorara  BOO  24  8.75 (1.01)  0  0.692 (0.042)  0.722 (0.037)  0.042 (0.027)  0.0795  Beadmore Road  BEA  24  8.75 (0.81)  2  0.678 (0.051)  0.706 (0.041)  0.053 (0.034)  0.2983  Denmark  DEN  23  8.69 (0.95)  1  0.642 (0.054)  0.696 (0.046)  0.074 (0.045)  0.4500  Mean    23  8.66 (0.18)    0.661 (0.010)  0.711 (0.009)  0.068 (0.008)    Population  Code  n  A  Np  Ho  UHe  F  Bottleneck test†  Northern Forest region  Mt Lesueur  LES  24  7.50 (0.84)  3  0.661 (0.052)  0.671 (0.052)  0.000 (0.029)  0.0649  Moochamulla  MOO  24  8.31 (1.08)  2  0.681 (0.049)  0.681 (0.045)  −0.002 (0.026)  0.3161  Julimar  JUL  24  8.88 (1.00)  3  0.648 (0.061)  0.714 (0.048)  0.101 (0.060)  0.0523  Mt Helena  HEL  24  9.13 (0.93)  2  0.675 (0.049)  0.735 (0.032)  0.089 (0.047)  0.1489  Perry Lakes  LAK  24  7.88 (0.96)  2  0.638 (0.057)  0.686 (0.051)  0.068 (0.040)  0.0145*  Dale  DAL  23  8.50 (0.89)  2  0.655 (0.047)  0.699 (0.046)  0.046 (0.042)  0.2809  Serpentine  SER  22  8.44 (0.97)  1  0.605 (0.035)  0.689 (0.039)  0.111 (0.034)*  0.4105  Whittaker  KER  24  9.06 (0.94)  2  0.651 (0.061)  0.717 (0.047)  0.122 (0.065)  0.1261  Wearne  WEA  24  9.00 (0.94)  0  0.708 (0.049)  0.745 (0.038)  0.062 (0.037)  0.0046*  Peel  PEE  24  9.06 (1.22)  4  0.692 (0.046)  0.735 (0.044)  0.047 (0.035)  0.0000*  Pindalup  PID  19  7.94 (0.88)  3  0.635 (0.049)  0.693 (0.053)  0.074 (0.040)  0.2983  Saddleback  SAD  23  8.63 (0.99)  0  0.664 (0.042)  0.729 (0.041)  0.080 (0.034)*  0.0026*  Godfrey  GOD  24  7.94 (0.88)  1  0.614 (0.044)  0.701 (0.043)  0.112 (0.041)*  0.0008*  Yourdaming  YOU  23  8.63 (1.11)  2  0.612 (0.053)  0.713 (0.049)  0.151 (0.038)*  0.0055*  Eaton  EAT  24  8.88 (1.08)  4  0.625 (0.054)  0.715 (0.047)  0.123 (0.037)*  0.1156  Lennard  LEN  24  9.31 (1.23)  2  0.646 (0.063)  0.711 (0.053)  0.096 (0.043)*  0.0523  Southern Forest region  Meelup  MEE  24  8.38 (0.88)  0  0.654 (0.055)  0.707 (0.044)  0.082 (0.039)*  0.0222*  Grimwade  GRI  24  8.31 (0.83)  5  0.628 (0.049)  0.710 (0.047)  0.101 (0.045)*  0.1261  Mowen  MOW  24  9.38 (1.03)  1  0.695 (0.045)  0.719 (0.045)  0.026 (0.026)  0.2477  Bramley  BRA  24  9.13 (0.89)  1  0.740 (0.037)  0.751 (0.036)  0.004 (0.038)  0.0327*  Kingston  KIN  24  8.81 (1.01)  3  0.625 (0.050)  0.704 (0.048)  0.100 (0.046)*  0.2019  Milylannup  MIL  24  9.06 (1.11)  5  0.688 (0.052)  0.722 (0.050)  0.040 (0.034)  0.0327*  Carey  CAR  24  8.94 (0.87)  1  0.698 (0.052)  0.709 (0.045)  0.013 (0.033)  0.3342  Muir  MUI  23  8.69 (1.00)    0.691 (0.054)  0.706 (0.048)  0.018 (0.036)  0.1489  Boorara  BOO  24  8.75 (1.01)  0  0.692 (0.042)  0.722 (0.037)  0.042 (0.027)  0.0795  Beadmore Road  BEA  24  8.75 (0.81)  2  0.678 (0.051)  0.706 (0.041)  0.053 (0.034)  0.2983  Denmark  DEN  23  8.69 (0.95)  1  0.642 (0.054)  0.696 (0.046)  0.074 (0.045)  0.4500  Mean    23  8.66 (0.18)    0.661 (0.010)  0.711 (0.009)  0.068 (0.008)    n, Number of samples; A, mean number of alleles per locus; Np, total number of private alleles within species; Ho, observed heterozygosity; UHe, unbiased expected heterozygosity; F, Wright’s inbreeding coefficient; †, Wilcoxon test, probability of a heterozygote excess under the IAM model. Standard errors in parentheses. *, Significant, P < 0.05. View Large The genetic differentiation of C. calophylla populations was significant but low (FST = 0.033, 95% CI 0.029–0.037) and there were 52 private alleles detected among the total 319 alleles. Mantel tests showed there was support for both IBD (r = 0.535, P < 0.01) and IBE (r = 0.452, P < 0.01). The effect of geographical distance was five times higher than the environmental effect (aridity) when the association between both parameters and genetic distance was partitioned by regression analyses. Hierarchical ANOVA showed a small proportion of variance was due to differences between the Northern and Southern Forest regions (1.48%; Table 3) although differentiation among populations was higher in the Northern Forest (FST = 0.054, SE 0.003) compared to the Southern Forest (FST = 0.038, SE 0.001). The indirect estimate of gene flow among populations (mean Nm) was 3.45. The estimate of pollen to seed gene flow ratio based on comparisons of nuclear and chloroplast differentiation (68:1) inferred substantially higher pollen- than seed-mediated gene flow. Populations of C. calophylla did not form distinct clusters with strong bootstrap support in the consensus NJ tree of CS Chord genetic distance (Fig. 3B). A geographical north/south division was present although support for this was weak (< 50%). The STRUCTURE analysis identified an optimum of two clusters (q1, q2) with a high similarity of runs (h′ = 0.995; Fig. 3C; Supporting Information Fig. S2). Individuals were admixtures of the two clusters with the relative proportions changing in a cline that generally followed the north–south geographical distribution and the north-east to south-west gradation of increasing rainfall and decreasing temperatures. The proportional cluster membership of populations was strongly correlated with the Aridity Index (Population q1, r = −0.771, P < 0.01). Figure 3. View largeDownload slide Genetic structure among sampled populations of Corymbia calophylla in south-western Australia, inferred from nuclear microsatellites. (A) Map showing the locations of sampled populations. Corymbia calophylla populations are shown as pie charts representing the proportion of genetic assignment to genetic clusters in STRUCTURE analyses. (B) Neighbour-joining tree of CS Chord distance. The pies on branches correspond to the proportion of genetic assignment to genetic clusters in STRUCTURE analyses. Support is shown on the branches as the number of bootstraps out of 1000. Values > 500 are shown. (C) Ancestry of individuals inferred from nuclear microsatellites data using STRUCTURE 2.3.4. Each individual is represented as a single line with coloured segments representing the proportion of ancestry from clusters (q). Results are the optimal alignment of 20 replicates. Figure 3. View largeDownload slide Genetic structure among sampled populations of Corymbia calophylla in south-western Australia, inferred from nuclear microsatellites. (A) Map showing the locations of sampled populations. Corymbia calophylla populations are shown as pie charts representing the proportion of genetic assignment to genetic clusters in STRUCTURE analyses. (B) Neighbour-joining tree of CS Chord distance. The pies on branches correspond to the proportion of genetic assignment to genetic clusters in STRUCTURE analyses. Support is shown on the branches as the number of bootstraps out of 1000. Values > 500 are shown. (C) Ancestry of individuals inferred from nuclear microsatellites data using STRUCTURE 2.3.4. Each individual is represented as a single line with coloured segments representing the proportion of ancestry from clusters (q). Results are the optimal alignment of 20 replicates. DISCUSSION Our analysis of genetic patterns in C. calophylla largely confirms our hypothesis that long-lived trees will show population persistence in stable, unglaciated mesic landscapes. However, evidence of unexpected range expansion in episodes during the Quaternary (~2.6 Mya to present) revealed a major role of the environment, through climatic oscillations and progressive drying of the forest habitat, in shaping genetic structure in this forest tree. In the absence of major topographic features within the distribution of the species, geographical distance appears to have been a key determinant of the spatial distribution of genetic variation at both historical and contemporary timescales, as expected for a sessile species in which restricted seed dispersal and localized pollen dispersal limits gene flow. Our results also confirmed our expectation that the environment may have shaped contemporary genetic structure through higher gene flow among plants with similar local climates, although the influence of distance was much greater than that of environment. Long-term persistence with episodic range expansion The phylogenetic structure of C. calophylla revealed evidence of persistence with unexpected range expansion in episodes during the Quaternary (~2.6 Mya to present). In the central and north regions, we found the expected signals of persistence with overall high haplotype and mid- to high nucleotide diversity, and highly divergent, localized haplotypes in populations in the lineages. In contrast, the lineage found in the south region included some widespread, shared, derived haplotypes and lower nucleotide diversity that is a pattern typically seen in range expansions (Hewitt, 2000; Pannell & Dorken, 2006). Neutrality and mismatch statistics also supported spatial expansion in the south, but not in the central or northern regions. The time sequence of diversification of lineages in C. calophylla during the Pleistocene (2.588 Mya to 11.7 kya) suggests range expansion consistent with the southward progress of increasing aridity in the SWAFR from the mid-Pleistocene (Hopper & Gioia, 2004; Byrne et al., 2008) that opened up suitable habitat within the forests. At the broad scale, the three chloroplast lineages (Central, North, South) correspond to divergence in a time sequence (early to mid to late-Pleistocene) over geographic regions (central, north, south). Diversification of the earliest lineage (Central) began in the late Pliocene (c. 3.028 Mya) during the period when progressive drying of mesic environments accelerated (see Byrne et al., 2011, 2008). The later divergence of the North and South lineages (c. 0.793 Mya) corresponds to the mid-Pleistocene transition (0.7–0.8 Mya), when climatic oscillations became more extreme (Zachos et al., 2001), and is consistent with the early (Nistelberger et al., 2014) or mid-Pleistocene (Byrne, Macdonald & Brand, 2003; Byrne & Hines, 2004) lineage divergence observed in other trees and woody shrubs in south-western Australia. However, the later range expansion of the South lineage in C. calophylla beginning around 0.426 Mya was unexpected. It suggests a significant influence of later Pleistocene climatic oscillations of increasing aridity in this more mesic environment, and is consistent with evidence of a significant shift in hydrological state to widespread aridity from approximately 500000 years ago in lakes in western, central and eastern Australia (Zheng et al., 1998). The southern progress of aridity in the SWAFR since the Pliocene (5–3 Mya) led to vegetation changes as denser forest dried and gave way to more open forests (Hopper & Gioia, 2004; Martin, 2006) and this may have provided opportunities for C. calophylla to colonize areas that had previously been dominated by species favouring wetter environments. Major expansions appear to have been rare in the mesic biota (Byrne et al., 2011, but see Nistelberger et al., 2014) although a similar pattern of expansion in the southern forest was seen in Allocasuarina humilis (Otto & A. Dietr) L.A.S. Johnson (Llorens et al., 2016), and the divergence of lineages in Eucalyptus wandoo Blakely from a geographically central refugium that were estimated to have been no later than 0.53 Mya (Dalmaris et al., 2015) may also be associated with expansion into the mesic forests as the environment dried. Influence of distance on genetic structure A signal of IBD for chloroplast markers indicated the influence of geographical distance in shaping spatial genetic structure at the historical scale. A lack of significant phylogeographical signal within populations (NST > GST) over the entire species’ range was consistent with theoretical expectations (Vekemans & Hardy, 2004) of population differentiation being determined by a balance between drift and gene flow rather than by mutation. Strong structuring and differentiation of populations and highly localized distributions of haplotypes is consistent with gene flow via seeds being too low to counter the diversifying effects of mutation and drift (Lowe & Allendorf, 2010). Seed-mediated gene flow is likely to be low in eucalypts because they have very limited seed dispersal capabilities with dispersal rates of 1–2 m per year (Booth 2017). Seed movements tend to take place in a stepwise fashion, taking advantage of disturbance events such as bushfires. Almost all seed dispersal is expected to be localized although occasional long-distance seed movements might be achieved by wind vortices in firestorms or cyclonic winds. Long-distance seed movements make also occur by cockatoos, as these large birds use C. calophylla seed as a food source and are known to carry capsules over some distance (Cooper et al., 2003). It is assumed that uneaten seeds are sometimes dispersed but there is no published data on how far or often this occurs (Booth, 2017). Foraging by cockatoos is more common in the southern forest (as flocks in the northern forests migrate to the coastal plain in summer) and might facilitate maintenance of greater connectivity in these populations. The high pollen to seed gene flow ratio (68:1) and low nuclear differentiation indicated that gene flow via pollen dispersal in C. calophylla has been more effective in maintaining the genetic connectivity of populations. Studies of pollination distances in the eucalypts show that dispersal is leptokurtic but with a fat tail, and is generally localized in continuous forest populations (Barbour, Potts & Vallaincourt, 2005; Jones et al., 2008). However, pollination can be extensive over long distances in more open woodlands or in fragmented landscapes (Byrne et al., 2008; Sampson & Byrne, 2008; Mimura et al., 2009). Pollen-mediated gene flow is therefore likely to be far higher than that achieved by restricted seed dispersal, and both overlapping gene pools and long-distance pollen dispersal are expected to retard differentiation in the nuclear genome. Low differentiation is generally observed in studies of widespread eucalypt species (Byrne, 2008) including the co-distributed forest tree E. marginata (Wheeler, Byrne & McComb 2003). However, although differentiation was low in C. calophylla, the strong signal of IBD in the nuclear genetic structure indicates a spatial component in the distribution of genetic variation and the influence of geographical distance at the contemporary timescale. The distribution of nuclear genetic variation is expected to be related to distance in most plants because pollen dispersal is usually leptokurtic (Vekemans & Hardy, 2004), particularly in a landscape with few topographical barriers. Genetic drift does not appear to have significantly disrupted the pattern of IBD despite evidence in nuclear data of bottlenecks in some populations. Bottlenecks could be the result of demographic changes following recurrent severe fires that have characterized the landscape (Pickett, 1997; Prideaux et al., 2010) but pollen-mediated gene flow has apparently been high enough in C. calophylla to prevent significant drift. We also found evidence that contemporary genetic structure may be shaped by the current environment, measured as local climate in the Aridity Index, although to a lesser extent than by distance (IBD ~ 84.6% vs. IBE ~ 15.5%). IBE can be produced by selection or if gene flow is restricted among plants in habitats with environmental differences through a variety of processes (Orsini, Andrew & Eizaguirre, 2013; Sork et al., 2016). Such processes could include isolation due to differences in flowering time created by environmental variation or changes in pollinator behaviour in different environments (Llorens et al., 2013). Flowering varies within the range of C. calophylla (Churchill, 1968; Cooper et al., 2003) and therefore restricted pollen dispersal among plants that flower concurrently might be expected to contribute to IBE. CONCLUSION In an unglaciated and topographically subdued landscape, the genetic structure of the mesic forest tree C. calophylla reflects persistence over multiple climatic oscillations during the Pleistocene (2.588 Mya to 11.7 kya). Environmental factors have been important in shaping genetic structure particularly on historical timescales. Deep divergence of lineages and unexpected episodic range expansion since the early Pleistocene appear to have been associated with climatic oscillations and progressive drying of the landscape. Geographical distance has been a key influence on both contemporary and historical genetic structure. Our study demonstrates that analysis of patterns of genetic diversity at historical and contemporary timescales in unglaciated landscapes can reveal insights into evolutionary processes and influences in forest species over long periods since the early Pleistocene. This knowledge contributes to informed management and climate adaptation strategies in forest trees that are critical for maintenance of forest ecosystems ACKNOWLEDGEMENTS We thank Andrew Thornhill for providing node dates for divergence estimates and advice on BEAST analyses. We thank the anonymous reviewers for their comments on our manuscript. SUPPORTING INFORMATION Additional Supporting Information may be found online in the supporting information tab for this article: Figure S1. Maximum-clade-credibility trees from Bayesian phylogenetic analyses for Corymbia calophylla made with (A) 2–4 My CI and (B) 1–5 My CI calibration applied to a known root age from eucalypt fossils. Figure S2. Delta K values from STRUCTURE HARVESTER 0.6.93 (Earl & von Holdt, 2012) of Bayesian inference of the number of nuclear microsatellite genetic marker clusters using STRUCTURE 2.3 4 (Pritchard et al., 2000) for Corymbia calophylla. Table S1. Primer sequences and characteristics of 16 microsatellite loci isolated from Corymbia calophylla. Allele size ranges and the number of alleles (NA) are based on 648 individuals from 27 natural populations. Table S2. List of GenBank accessions for haplotypes uncovered in 27 Corymbia calophylla populations in south-western Australia via sequencing of psbA-trnH, trnG and trnQ-rps16 chloroplast intergenic spacer regions. Coloured boxes next to haplotypes correspond to colours used in Figures 1 and 2. Table S3. Estimates of microsatellite null allele frequencies made for 27 populations of Corymbia calophylla from south-west Australia. REFERENCES Bandelt HJ, Forster P, Röhl A. 1999. Median-joining networks for inferring intraspecific phylogenies. Molecular Biology and Evolution  16: 37– 48. Google Scholar CrossRef Search ADS PubMed  Barbour RC, Potts BM, Vaillancourt RE. 2005. Pollen dispersal from exotic eucalypt plantations. Conservation Genetics  6: 253– 257. Google Scholar CrossRef Search ADS   Bayly MJ, Rigault P, Spokevicius A, Ladiges PY, Ades PK, Anderson C, Bossinger G, Merchant A, Udovicic F, Woodrow IE, Tibbits J. 2013. Chloroplast genome analysis of Australian eucalypts – Eucalyptus, Corymbia, Angophora, Allosyncarpia and Stockwellia (Myrtaceae). Molecular Phylogenetics and Evolution  69: 704– 716. Google Scholar CrossRef Search ADS PubMed  Booth TH. 2017. Going nowhere fast: a review of seed dispersal in eucalypts. Australian Journal of Botany  65: 401– 410. Google Scholar CrossRef Search ADS   Bradbury D, Tapper SL, Coates D, Hankinson M, McArthur S, Byrn M. 2016a. How does the post-fire facultative seeding strategy impact genetic variation and phylogeographical history? The case of Bossiaea ornata (Fabaceae) in a fire-prone, mediterranean-climate ecosystem. Journal of Biogeography  43: 96– 110. Google Scholar CrossRef Search ADS   Bradbury D, Tapper SL, Coates D, McArthur S, Hankinson M, Byrn M. 2016b. The role of fire and a long-lived soil seed bank in maintaining persistence, genetic diversity and connectivity in a fire-prone landscape. Journal of Biogeography  43: 70– 84. Google Scholar CrossRef Search ADS   Bradshaw FJ. 2015. Reference material for jarrah forest silviculture, Forest Management Series FEM061 . Perth: Department of Parks and Wildlife. Byrne M. 2008. Phylogeny, diversity and evolution of eucalypts. In: Sharma A, Sharma A, eds. Plant genome: biodiversity and evolution. 1E: Phanerogam – angiosperm . Enfield: Science Publishers, 303– 346. Byrne M, Elliott CP, Yates CJ, Coates DJ. 2008. Maintenance of high pollen dispersal in Eucalyptus wandoo, a dominant tree of the fragmented agricultural region in Western Australia. Conservation Genetics  9: 97– 105. Google Scholar CrossRef Search ADS   Byrne M, Steane DA, Joseph L, Yeates DK, Jordan GJ, Crayn D, Aplin K, Cantrill DJ, Cook LG, Crisp MD, Keogh JS, Melville J, Moritz C, Porch N, Kale Sniderman JM, Sunnucks P, Weston PH. 2011. Decline of a biome: Evolution, contraction, fragmentation, extinction and invasion of the Australian mesic zone biota. Journal of Biogeography  38: 1635– 1656. Google Scholar CrossRef Search ADS   Byrne M, Coates DJ, Macdonald BM, van Leeuwen S. 2016. High nuclear genetic differentiation, but low chloroplast diversity in a rare species, Aluta quadrata (Myrtaceae), with a disjunct distribution in the Pilbara, Western Australia. Australian Journal of Botany  64: 687– 695. Google Scholar CrossRef Search ADS   Byrne M, Hankinson M. 2012. Testing the variability of chloroplast sequences for plant phylogeography. Molecular Ecology Resources  60: 569– 574. Byrne M, Hines B. 2004. Phylogeographical analysis of cpDNA variation in Eucalyptus loxophleba (Myrtaceae). Australian Journal of Botany  52: 459– 470. Google Scholar CrossRef Search ADS   Byrne M, Macdonald B, Brand J. 2003. Phylogeography and divergence in the chloroplast genome of Western Australian Sandalwood (Santalum spicatum). Heredity  91: 389– 395. Google Scholar CrossRef Search ADS PubMed  Chapuis MP, Estoup A. 2007. Microsatellite null alleles and estimation of population differentiation. Molecular Biology and Evolution  24: 621– 631. Google Scholar CrossRef Search ADS PubMed  Churchill DM. 1968. The distribution and prehistory of Eucalyptus diversicolor F. Muell., E. marginata Donn ex SM., and E. calophylla R. Br. in relation to rainfall. Australian Journal of Botany  16: 125– 151. Google Scholar CrossRef Search ADS   Cooper CE, Withers PC, Mawson PR, Johnstone R, Kirkby T, Prince J, Bradshaw D, Robertson H. 2003. Characteristics of Marri (Corymbia calophylla) fruits in relation to the foraging behaviour of the Forest Red-tailed Black Cockatoo (Calyptorhynchus banksii naso). Journal of the Royal Society of Western Australia  86: 139– 142. Dalmaris E, Ramalho CE, Poot P, Veneklaas EJ, Byrne M. 2015. A climate change context for the decline of a foundation tree species in south-western Australia: insights from phylogeography and species distribution modelling. Annals of Botany  116: 941– 952. Google Scholar CrossRef Search ADS PubMed  Dieringer D, Schlötterer C. 2003. MICROSATELLITE ANALYSER: a platform independent analysis tool for large microsatellite data sets. Molecular Ecology Notes  3: 167– 169. Google Scholar CrossRef Search ADS   Drummond AJ, Rambaut A. 2007. BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evolutionary Biology  7: 214. Google Scholar CrossRef Search ADS PubMed  Earl D, von Holdt B. 2012. STRUCTURE HARVESTER: a website and program for visualizing STRUCTURE output and implementing the Evanno method. Conservation and Genetic Resources  4: 359– 361. Google Scholar CrossRef Search ADS   Ennos R. 1994. Estimating the relative rates of pollen and seeed migration among plant populations. Heredity  72: 250– 259. Google Scholar CrossRef Search ADS   Excoffier L, Lischer H. 2010. Arlequin suite ver 3.5: a new series of programs to perform population genetic analyses under Linux and Windows. Molecular Ecology Resources  10: 464– 467. Google Scholar CrossRef Search ADS   Felsenstein J. 1989. PHYLIP – Phylogeny Inference Package (Version 3.2). Cladistics  5: 164– 166. Fu YX. 1997. Statistical tests of neutrality of mutations against population growth, hitchhiking and background selection. Genetics  147: 915– 925. Google Scholar PubMed  Fu YX, Li WH. 1993. Statistical tests of neutrality of mutations. Genetics  133: 693– 709. Google Scholar PubMed  Gonzalez-Orozco C, Pollock L, Thornhill A, Mishler B, Knerr N, Laffan S, Miller J, Rosauer D, Faith D, Nipperess D, Kujala H, Linke S, Butt N, Külheim C, Crisp M, Gruber B. 2016. Phylogenetic approaches reveal biodiversity threats under climate change. Climate Change  6: 1110– 1114. Graur D, Li W. 2000. Fundamentals of molecular evolution . Sunderland: Sinauer Associates Inc. Gugger PF, Ikegami M, Sork VL. 2013. Influence of late Quaternary climate change on present patterns of genetic variation in valley oak, Quercus lobata Née. Molecular Ecology  22: 3598– 3612. Google Scholar CrossRef Search ADS PubMed  Hewitt G. 2000. The genetic legacy of the Quaternary ice ages. Nature  405: 907– 913. Google Scholar CrossRef Search ADS PubMed  Hewitt GM. 2001. Speciation, hybrid zones and phylogeography - or seeing genes in space and time. Molecular Ecology  10: 537– 549. Google Scholar CrossRef Search ADS PubMed  Hopper SD, Gioia P. 2004. The Southwest Australian Floristic Region: evolution and conservation of a global hot spot of biodiversity. Annual Review of Ecology, Evolution, and Systematics  35: 623– 650. Google Scholar CrossRef Search ADS   Jakobsson M, Rosenberg NA. 2007. CLUMPP: a cluster matching and permutation program for dealing with label switching and multimodality in analysis of population structure. Bioinformatics  23: 1801– 1806. Google Scholar CrossRef Search ADS PubMed  Jones M, Shepherd M, Henry R, Delves A. 2008. Pollen flow in Eucalyptus grandis determined by paternity analysis using microsatellite markers. Tree Genetics & Genomes  4: 37– 47. Google Scholar CrossRef Search ADS   Librado P, Rozas J. 2009. DnaSP v5: a software for comprehensive analysis of DNA polymorphism data. Bioinformatics  25: 1451– 1452. Google Scholar CrossRef Search ADS PubMed  Llorens TM, Yates CJ, Byrne M, Nistelberger HM. 2013. Complex interactions between remnant shape and the mating system strongly influence reproductive output and progeny performance in fragmented populations of a bird-pollinated shrub. Biological Conservation  164: 129– 139. Google Scholar CrossRef Search ADS   Llorens TM, Tapper SL, Coates DJ, McArthur S, Hankinson M, Byrne M. 2016. Does population distribution matter? Influence of a patchy versus continuous distribution on genetic patterns in a wind-pollinated shrub. Journal of Biogeography  44: 361– 374. Google Scholar CrossRef Search ADS   Lowe WH, Allendorf FW. 2010. What can genetics tell us about population connectivity? Molecular Ecology  19: 3038– 3051. Google Scholar CrossRef Search ADS PubMed  Maddison WP, Maddison DR. 2016. Mesquite: a modular system for evolutionary analysis, Version 3.04. Available at http://mesquiteproject.org, accessed 23 October 2017. Martin HA. 2006. Cenozoic climatic change and the development of the arid vegetation in Australia. Journal of Arid Environments  66: 533– 563. Google Scholar CrossRef Search ADS   McLachlan JS, Clark JS, Manos PS. 2009. Molecular indicators of tree migration capacity under rapid climate change. Ecology  86: 2088– 2098. Google Scholar CrossRef Search ADS   Mimura M, Barbour RC, Potts BM, Vaillancourt RE, Watanabe KN. 2009. Comparison of contemporary mating patterns in continuous and fragmented Eucalyptus globulus native forests. Molecular Ecology  18: 4180– 4192. Google Scholar CrossRef Search ADS PubMed  Nevill PG, Bradbury D, Williams A, Tomlinson S, Krauss SL. 2014. Genetic and palaeo-climatic evidence for widespread persistence of the coastal tree species Eucalyptus gomphocephala (Myrtaceae) during the Last Glacial Maximum. Annals of Botany  113: 55– 67. Google Scholar CrossRef Search ADS PubMed  Nistelberger H, Gibson N, Macdonald B, Tapper SL, Byrne M. 2014. Phylogeographic evidence for two mesic refugia in a biodiversity hotspot. Heredity  113: 454– 463. Google Scholar CrossRef Search ADS PubMed  O’Brien EK, Krauss SL. 2010. Testing the home-site advantage in forest trees on disturbed and undisturbed sites. Restoration Ecology  18: 359– 372. Google Scholar CrossRef Search ADS   van Oosterhout C, Hutchinson WF, Wills DPM, Shipley P. 2004. MICRO-CHECKER: software for identifying and correcting genotyping errors in microsatellite data. Molecular Ecology Notes  4: 535– 538. Google Scholar CrossRef Search ADS   Orsini L, Andrew R, Eizaguirre C. 2013. Evolutionary ecological genomics. Molecular Ecology  22: 527– 531. Google Scholar CrossRef Search ADS PubMed  Pannell JR, Dorken ME. 2006. Colonisations as a common denominator in plant metapopulations and range expansions: effects on genetic diversity and sexual systems. Landscape Ecology  21: 837– 848. Google Scholar CrossRef Search ADS   Peakall R, Smouse PE. 2012. GenALEx 6.5: genetic analysis in excel. Population genetic software for teaching and research-an update. Bioinformatics  28: 2537– 2539. Google Scholar CrossRef Search ADS PubMed  Petit RJ, Aguinagalde I, de Beaulieu JL, Bittkau C, Brewer S, Cheddadi R, Ennos R, Fineschi S, Grivet D, Lascoux M, Mohanty A, Müller-Starck G, Demesure-Musch B, Palmé A, Martín JP, Rendell S, Vendramin GG. 2003. Glacial refugia: hotspots but not melting pots of genetic diversity. Science (New York, N.Y.)  300: 1563– 1565. Google Scholar CrossRef Search ADS PubMed  Petit RJ, Hampe A. 2006. Some evolutionary consequences of being a tree. Annual Review of Ecology, Evolution, and Systematics  37: 187– 214. Google Scholar CrossRef Search ADS   Pickett EJ. 1997. A Late Pleistocene and Holocene vegetation history of three lacustrine sequences from the Swan Coastal Plain . PhD Thesis, The University of Western Australia. Piry S, Luikart G, Cornuet JM. 1999. Computer note. BOTTLENECK: a computer program for detecting recent reductions in the effective size using allele frequency data. Journal of Heredity  90: 502– 503. Google Scholar CrossRef Search ADS   Pons O, Petit RJ. 1996. Measuring and testing genetic differentiation with ordered versus unordered alleles. Genetics  144: 1237– 1245. Google Scholar PubMed  Posada D. 2008. jModelTest: phylogenetic model averaging. Molecular Biology and Evolution  25: 1253– 1256. Google Scholar CrossRef Search ADS PubMed  Prideaux GJ, Gully GA, Couzens AMC, Ayliffe LK, Jankowski NR, Jacobs Z, Roberts RG, Hellstrom JC, Gagan MK, Hatcher LM. 2010. Timing and dynamics of Late Pleistocene mammal extinctions in southwestern Australia. Proceedings of the National Academy of Sciences of the United States of America  107: 22157– 22162. Google Scholar CrossRef Search ADS PubMed  Pritchard JK, Stephens M, Donnelly P. 2000. Inference of population structure using multilocus genotype data. Genetics  155: 945– 959. Google Scholar PubMed  Ramos-Onsins SE, Rozas J. 2002. Statistical properties of new neutrality tests against population growth. Molecular Biology and Evolution  19: 2092– 2100. Google Scholar CrossRef Search ADS PubMed  Rousset F. 1997. Genetic differentiation and estimation of gene flow from F-statistics under isolation by distance. Genetics  145: 1219– 1228. Google Scholar PubMed  Rousset F. 2008. genepop’007: a complete re-implementation of the genepop software for Windows and Linux. Molecular Ecology Resources  8: 103– 106. Google Scholar CrossRef Search ADS PubMed  Sampson JF, Byrne M. 2008. Outcrossing between an agroforestry plantation and remnant native populations of Eucalyptus loxophleba. Molecular Ecology  17: 2769– 2781. Google Scholar CrossRef Search ADS PubMed  Schaal BA, Haywood DA, Olsen KM, Rauscher JT, Smith WA 1998. Phylogeographic studies in plants, problems and prospects. Molecular Ecology  7: 465– 474. Google Scholar CrossRef Search ADS   Slatkin M. 1985. Gene flow in natural populations. Annual Review of Ecology, Evolution, and Systematics  16: 393– 430. Google Scholar CrossRef Search ADS   Soltis DE, Morris AB, McLachlan JS, Manos PS, Soltis PS. 2006. Comparative phylogeography of unglaciated eastern North America. Molecular Ecology  15: 4261– 4293. Google Scholar CrossRef Search ADS PubMed  Sork VL, Gugger PF, Chen JM, Werth S. 2016. Evolutionary lessons from California plant phylogeography. Proceedings of the National Academy of Sciences of the United States of America  113: 8064– 8071. Google Scholar CrossRef Search ADS PubMed  Tajima F. 1989. Statistical method for testing the neutral mutation hypothesis by DNA polymorphism. Genetics  123: 585– 595. Google Scholar PubMed  Telles MP, Diniz-Filho JA. 2005. Multiple Mantel tests and isolation-by-distance, taking into account long-term historical divergence. Genetics and Molecular Research  4: 742– 748. Google Scholar PubMed  Thornhill AH, Macphail M. 2012. Fossil myrtaceous pollen as evidence for the evolutionary history of Myrtaceae: a review of fossil Myrtaceidites species. Review of Palaeobotany and Palynology  176– 177: 1– 23. Vekemans X, Hardy OJ. 2004. New insights from fine-scale spatial genetic structure analyses in plant populations. Molecular Ecology  13: 921– 935. Google Scholar CrossRef Search ADS PubMed  Wang IJ. 2010. Recognizing the temporal distinctions between landscape genetics and phylogeography. Molecular Ecology  19: 2605– 2608. Google Scholar CrossRef Search ADS PubMed  Wang IJ, Bradburd GS. 2014. Isolation by environment. Molecular Ecology  23: 5649– 5662. Google Scholar CrossRef Search ADS PubMed  Wardell-Johnson GW. 2000. Responses of forest eucalypts to moderate and high intensity fire in the Tingle Mosaic, south-western Australia: comparisons between locally endemic and regionally distributed species. Austral Ecology  25: 409– 421. Google Scholar CrossRef Search ADS   Wheeler MA, Byrne M. 2006. Congruence between phylogeographic patterns in cpDNA variation in Eucalyptus marginata (Myrtaceae) and geomorphology of the Darling Plateau, south-west of Western Australia. Australian Journal of Botany  54: 17– 26. Google Scholar CrossRef Search ADS   Wheeler MA, Byrne M, McComb JA. 2003. Little genetic differentiation within the dominant forest tree, Eucalyptus marginata (Myrtaceae) of south-western Australia. Silvae Genetica  52: 254– 259. Whitlock BA, Hale AM, Groff PA. 2010. Intraspecific inversions pose a challenge for the trnH-psbA plant DNA barcode. PLoS ONE  5: e11533. Google Scholar CrossRef Search ADS PubMed  Zachos J, Pagani M, Sloan L, Thomas E, Billups K. 2001. Trends, rhythms, and aberrations in global climate 65 Ma to present. Science (New York, N.Y.)  292: 686– 693. Google Scholar CrossRef Search ADS PubMed  Zhang YH, Wang IJ, Comes HP, Peng H, Qiu YX. 2016. Contributions of historical and contemporary geographic and environmental factors to phylogeographic structure in a Tertiary relict species, Emmenopterys henryi (Rubiaceae). Scientific Reports  6: 24041. Google Scholar CrossRef Search ADS PubMed  Zheng H, Wyrwoll K-H, Li Z, Powell CM. 1998. Onset of aridity in southern Western Australia - a preliminary palaeomagnetic appraisal. Global and Planetary Change  18: 175– 187. Google Scholar CrossRef Search ADS   © 2018 The Linnean Society of London, Biological Journal of the Linnean Society

Journal

Biological Journal of the Linnean SocietyOxford University Press

Published: Mar 1, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off