Evolution of a hotspot genus: geographic variation in speciation and extinction rates in Banksia(Proteaceae)
© Cardillo and Pratt; licensee BioMed Central Ltd. 2013
Received: 1 May 2013
Accepted: 11 July 2013
Published: 19 August 2013
Hotspots of angiosperm species richness and endemism in Mediterranean-climate regions are among the most striking, but least well-understood, geographic patterns of biodiversity. Recent studies have emphasized the importance of rapid diversification within hotspots, compared to non-hotspot regions, as a major contributor to these patterns. We constructed the first near-complete phylogeny of Banksia (Proteaceae) to test whether diversification rates have differed between lineages confined to the southwest Australian hotspot and those found throughout southern, eastern and northern Australia. We then tested for variation in diversification rates among the bioclimatic zones within the southwest hotspot itself.
Although Banksia species richness in the southwest is ten times that of the rest of the continent, we find little evidence for more rapid diversification in the southwest, although this result is inconclusive. However, we find firmer support for substantial rate variation within the southwest hotspot, with more rapid diversification in the semi-arid heaths and shrublands, compared to the high-rainfall forests. Most of the Banksia diversity of the southwest appears to be generated in the heaths and shrublands, with a high migration rate out of this zone boosting diversity of the adjacent forest zone.
The geographic pattern of diversification in Banksia appears more complex than can be characterized by a simple hotspot vs. non-hotspot comparison, but in general, these findings contrast with the view that the high diversity of Mediterranean hotspots is underpinned by rapid radiations. Steady accumulation of species at unexceptional rates, but over long periods of time, may also have contributed substantially to the great botanical richness of these regions.
Mediterranean-climate regions include some of the world’s great repositories of botanical diversity, with levels of angiosperm species richness and endemism that can rival tropical rainforests . However, these environments typically lack features such as high rainfall or productivity that are often associated with high plant diversity on a global scale. Indeed, some of the most species-rich Mediterranean communities are semi-arid heaths and shrublands found on ancient, nutrient-impoverished soils, such as the fynbos of South Africa or the kwongan of southwest Australia [1–3]. Understanding these apparent outliers on global biodiversity gradients may yield insights into the drivers of angiosperm diversification more generally.
A key step in explaining the high botanical diversity of Mediterranean hotspots is demonstrating whether this diversity has arisen as a result of more rapid diversification in these regions, compared to elsewhere, and if so, whether this is the result of a higher rate of speciation or a lower rate of extinction . If diversification rates are not unusually high in Mediterranean hotspots, then alternative explanations for high diversity must be investigated. For example, high diversity may be the result of a steady accumulation of species over a longer period of time, or of higher ecological limits on diversity, in Mediterranean hotspots. However, explicit comparisons of diversification rates between groups within and outside Mediterranean hotspots are rare. Several recent analyses, using very different approaches, support rapid hotspot diversification [5–7]. On the other hand, another study  found no evidence for rapid hotspot diversification in an analysis of a large, densely-sampled genus (Protea) that spans the boundaries of South Africa’s Cape region.
Hopper  acknowledged that geohistorical events alone could not completely explain the high plant diversity of the SWBP. Cowling  offered a more deterministic hypothesis based on ecological processes. According to Cowling, frequent fires, a prominent feature of many Mediterranean communities, serve to repeatedly fragment species populations. Combined with poor dispersal capacities and a complex mosaic of soil types, this gives rise to numerous edaphically-specialized populations, and high rates of speciation ensue  (see also ). This kind of speciation would have become more frequent in the later Tertiary as the climate became drier and the flora became dominated by flammable scleromorphic species, increasing fire frequency .
Hopper’s historical scenario and Cowling’s ecological hypothesis both make the simple prediction that diversification rates have been higher in the SWBP compared to less species-rich regions outside the SWBP. A recent broad-scale, genus-level analysis  provides some support for this prediction, but it has never been thoroughly tested for any single, well-sampled clade. Furthermore, both hypotheses involve complexities that suggest diversification rates may also vary within the SWBP itself. Within the SWBP, plant diversity tends to be lower in the forested high-rainfall zone (HRZ) of the southwest coastal margin, compared to the heaths and shrublands of the transitional rainfall zone (TRZ) and the southeast coastal province (SCP; [2, 10]). Hopper  noted that the TRZ and SCP appear to harbour a greater proportion of recently-evolved endemic plant species, suggesting a pattern of more recent, rapid speciation in these regions compared to the HRZ. On the other hand, the proportion of older, “relict” species appears to be higher in the HRZ, suggesting a lower extinction rate in this region. Hopper [2, 10] suggested that more complex erosional patterns in the TRZ led to recurrent fragmentation and isolation of populations in the late Tertiary and Quaternary, hence more rapid speciation, while the comparative stability of the lateritic soils in the HRZ favoured larger, more persistent populations, and a lower extinction rate. Under Cowling’s hypothesis, the rate of speciation is linked to the frequency and intensity of fire, both of which vary substantially across different ecosystem types in the SWBP . Hence, this hypothesis also predicts variation in diversification rates within the SWBP.
In this study we use the iconic Australian genus Banksia (Proteaceae) as a case study for a comparison of diversification rates (1) between the SWBP and elsewhere in Australia, and (2) between the bioclimatic zones within the SWBP. Based on molecular evidence , a recent taxonomic revision  sunk the former genus Dryandra into Banksia, forming a large new genus of 170 species. The majority of these are endemic to the SWBP, but Banksia have dispersed throughout the southern, eastern and northern parts of Australia (Figure 1). We present the first near-complete species-level molecular phylogeny of Banksia sensu lato. We then use this phylogeny to reconstruct the ancestral range of Banksia, and to compare rates of diversification between regions, using recently-developed methods that allow for reciprocal effects of diversification and shifts in geographic distribution .
Sampling, DNA extraction and sequencing
We used a combination of previously-published sequences for 88 Banksia taxa (including multiple subspecies within some species), and newly-generated sequence data for 110 taxa. Tissue samples were collected from dried specimens held at the Australian National Herbarium, Canberra, and from living plants at the Banksia Farm arboretum, Mt. Barker, Western Australia. We extracted total genomic DNA using a Qiagen DNEasy Plant Mini Kit according to the manufacturer’s protocol (Qiagen Inc., California, USA). We amplified DNA from four noncoding regions of the chloroplast genome; 994bp of the trnL-trnF intergenic spacer, 1113bp of the rpl16 intron, 578bp of the psbA/trnH spacer and 873bp of the trnT-trnL spacer. Sequencing was carried out at Macrogen on a 3730XL DNA sequencer (Macrogen Inc., Seoul, Korea) using the same primers used to amplify the product. Details of primer sequences and GenBank accessions are provided in Additional files 1 and 2.
Phylogeny reconstruction and dating
We concatenated published and newly-generated sequences from the four cpDNA markers for 198 Banksia taxa, together with sequences from GenBank for six outgroup species (Panopsis costaricensis, Musgravea heterophylla, Grevillea renwickiana, Grevillea scortechinii, Macadamia integrifolia and Hakea victoria). Sequences were aligned using automated alignment followed by manual editing in Geneious . The phylogeny was estimated together with divergence times using a Bayesian analysis in BEAST 1.6.2 . Calibrations were based on ages estimated from fossil data ; this reference gives details of the fossils used and the method by which stratigraphic ages were converted to absolute time estimates. We placed lognormal priors on the ages (in millions of years) of the Banksia crown node (mean = 42, sd = 0.4) and the Banksia stem node (mean = 62, sd = 0.4). Because the formerly separate genus Dryandra is now believed to be nested within Banksia, our Banksia stem node corresponds to the node considered to be the stem of the tribe Banksieae by Crisp & Cook , for which they assumed a minimum age of 62Mya. We used lognormal priors for these calibrations because this allows for uncertainty around the age estimates, while still reflecting the belief that these ages represent minimum bounds on the estimated divergence times . For the root node we used a normal prior (mean = 77, sd = 1), based on the estimated divergence time between the lineages leading to Banksia and Hakea / Grevillea in a comprehensive genus-level phylogeny of the Proteaceae . Because this was a secondary calibration that does not represent a minimum bound, we felt that a normally-distributed prior was most appropriate. We used a Yule prior for the speciation process, with an uncorrelated lognormal model of variation in rate of molecular evolution. The function modelTest in the R library phangorn v.1.6.0  was used to select an HKY + gamma substitution model. We ran two separate MCMC runs for >40,000,000 generations, sampling trees every 10,000 generations. We diagnosed sampling adequacy by examining the effective sample sizes (ESS) using Tracer 1.5 , and stopped the runs when the parameter values of the combined runs all had ESS > 300.
Our phylogenies included multiple subspecies or varieties for some species. To avoid biasing our analyses by the variable representation of intraspecific taxa, we pruned the trees to the species level before all analyses. Each time a new tree was called from the posterior distribution, we randomly sampled one tip from each species represented by multiple tips.
Assigning species to geographic regions
Geographic range maps were constructed for each species by obtaining all Banksia and Dryandra collection records from the Australian Virtual Herbarium (http://chah.gov.au/avh/), then removing records that were clearly outside the natural distribution limits of each species (e.g. those from botanic gardens in capital cities) by checking each map against published maps in atlases of Banksia and Dryandra species. We first classified each species as belonging to the SWBP or not. This was relatively straightforward as distributions of southwestern Australian Banksia are clearly separated from the rest of Australia by arid barriers to the north and east, and no species has a distribution than spans these barriers. The distributions of a few species are slightly outside the formal boundary of the SWBP, but still clearly separated from the eastern and northern Australian species – these species were assigned to the SWBP. We classified each SWBP species as belonging to the HRZ, the TRZ/SCP, or both. This was done by overlaying each species distribution onto a map of the rainfall zone boundaries and counting the number of records within each zone. If >90% of a species’ records fell in one zone, the species was assigned to this zone; otherwise the zone was recorded as “both”. All mapping was done using the maptools, rgdal, and raster libraries in R. The numbers of species of each geographic state that were present / not present in the phylogeny are as follows. non-SWBP: 15/1; SWBP: 143/12; HRZ: 11/0; TRZ+SCP: 78/9; both HRZ and TRZ+SCP: 54/3.
We then reconstructed the ancestral geographic range of all nodes in the Banksia tree as a binary variable (SWBP/non-SWBP) by maximum likelihood, using the Mk2 model in the R library diversitree.
Geographic variation in diversification rates
To compare rates of diversification among regions, and rates of migration across regions, we used two related methods. First, we tested whether diversification varies between SWBP and non-SWBP species using BiSSE , implemented in diversitree. The BiSSE method implements maximum likelihood estimates of six parameters under a constant-rates birth-death process: speciation and extinction rates under each of the two states of a binary character (λ0, λ1, μ0, μ1), rates of transition from state zero to state one (m01), and rates of transition from state one to state zero (m10). In our analyses, we treated SWBP and non-SWBP species as displaying state zero and state one, respectively. Second, we tested whether diversification rates vary among the rainfall zones within the SWBP, using GeoSSE . GeoSSE differs from BiSSE by allowing for reciprocal effects of geographic distribution and diversification rates; for example, a species distributed across two geographic regions may speciate to produce two daughter species each with a distribution restricted to a single region. In addition to the six BiSSE parameters, therefore, GeoSSE estimates a seventh parameter for between-region speciation. For this test we compared species belonging to the HRZ with those belonging to the TRZ/SCP. Because the GeoSSE model requires that all three geographic states (region A, region B, and both) be occupied by at least some species, we could not apply GeoSSE to the SWBP/non-SWBP comparison, since no species are distributed across the SWBP boundary. For this comparison we used BiSSE, which does not have this restriction. For both BiSSE and GeoSSE, we specified the proportion of species in each geographic zone that are not present in the phylogeny.
To test for variation in diversification and migration rates between SWBP and non-SWBP species, we began by fitting a BiSSE model in which all six parameters were allowed to vary (full model). We then fitted three constrained models in which (1) speciation and extinction rates were held equal across regions, and migration from and to the SWBP were held equal (equal-rates model); (2) speciation and migration were held equal but extinction allowed to vary (equal-speciation model); and (3) extinction and migration were held equal but speciation allowed to vary (equal-extinction model). We did not fit a model that explicitly tests the influence of varying migration rates while holding speciation and extinction constant, because migration rates into and out of the SWBP are likely to have been too low to have any important influence on differences in species richness. We compared the fit of the four models to the data using AIC. All tests were repeated for 250 phylogenies sampled evenly from the Bayesian posterior distribution; we felt that this number was sufficient to adequately represent the variation among trees within the posterior set.
To test for variation in diversification and migration rates among rainfall zones within the SWBP, we used GeoSSE rather than BiSSE. In this set of tests we were explicitly interested in the role of migration rates between rainfall zones, so we tested the following five models: (1) all six parameters allowed to vary (full model); (2) all parameters held equal (equal-rates model); (3) speciation held equal but extinction and migration allowed to vary (equal-speciation model); (4) extinction held equal but speciation and migration allowed to vary (equal-extinction model); (5) migration held equal but speciation and extinction allowed to vary (equal-migration model). As above, we compared models using AIC and repeated tests on the same sample of 250 phylogenies as the BiSSE analysis.
Banksia phylogeny and ancestral region
Temporal patterns of diversification
Geographic patterns of diversification
Although the nearest extant relatives of Banksia (Austromuellera and Musgravea) are confined to the rainforests of northeastern Queensland, at the opposite end of the Australian continent, our reconstruction of ancestral regions places the crown node of Banksia in the SWBP with a high probability. It seems likely, therefore, that Banksia began diversifying within the SWBP, and that the extant non-SWBP species of Banksia are derived from two independent dispersal events out of the southwest and throughout northern, southern and eastern Australia. One important issue relevant to understanding the high floristic diversity and endemism of the SWBP hotspot is the role played by the Nullarbor Plain in driving divergences between eastern and western species. The Nullarbor Plain is an arid limestone plateau that developed around 13–14 Myr ago, potentially acting as both an aridity barrier and an edaphic barrier to the east–west dispersal of plant species . Crisp & Cook  found that divergences between eastern and western Australian clades within a number of plant genera seemed to cluster around 13–14 Myr ago, leading them to conclude that the Nullarbor Plain played a significant role in the isolation of the SWBP flora. Their reconstructed divergence time for the smaller of the two eastern Banksia clades was 16.3 Myr (CI = 4.3 – 31.3Myr), based on the split between B. benthamiana and B. serrata. However, our phylogeny reveals several SWBP species more closely-related to B. serrata than B. benthamiana, giving a median estimate for the divergence of this clade from its SWBP sister clade of only 8.1 Myr (95% HPD = 3.2 – 13.3 Myr), post-dating the formation of the Nullarbor Plain. This suggests that while the aridity and edaphic barrier formed by the Nullarbor Plain may have restricted the frequency of east–west dispersal of Banksia, it is unlikely to have been a hard barrier to dispersal.
Our BiSSE analysis found no evidence that the two dispersal events out of the southwest corresponded either with a reduced speciation rate or with an increased extinction rate. One possible explanation for this is a simple lack of statistical power. A recent simulation study  showed that the BiSSE method has relatively low power to detect rate differences among states of a binary trait in phylogenies with fewer than 300 tips and a high degree of asymmetry in tip states. Our Banksia phylogenies have 158 tips of which 15 (9.5%) have state zero (i.e. non-SWBP). The results of our BiSSE analysis must therefore be regarded as inconclusive, and treated cautiously. Indeed, the position of the larger non-SWBP clade on the plot of clade age x log(species richness) (Figure 4) hints at the possibility that migration out of the SWBP was accompanied by reduced diversification rates, although again, this result is inconclusive.
The expansion of Banksia out of the southwest opened up vast new areas of the Australian continent in which the genus could diversify. Regardless of whether diversification rates remained steady or were reduced following this expansion, it appears that the expansion out of the southwest did not provide significant new ecological opportunities for diversification. The difference in mean geographic range size between SWBP and non-SWBP Banksia species supports this view. In the SWBP, geographically fine-scale speciation has produced numerous narrowly-distributed species, whereas the non-SWBP clades are characterized by fewer, more widely-distributed species. Hence, although non-SWBP clades occupy a collective area many times larger than the SWBP, species density is far lower. In this respect, the situation for Banksia is intriguingly similar to that of another hotspot genus, the African Protea. Protea is a large genus of Proteaceae with the great majority of its species endemic to the Cape biodiversity hotspot, but with a small number species, derived from only a single migration event out of the Cape, distributed across a large area of Africa . Like Banksia, the non-hotspot species of Protea tend to occupy far larger geographic ranges than the hotspot species, and Valente et al. found no evidence for a difference in diversification rates between Cape and non-Cape clades (although the above caveat about the statistical power of BiSSE also applies to their analysis). The similarity of patterns between Banksia and Protea suggests a possible common mechanism driving diversification in the two genera. One possibility is that strong habitat affinities (niche conservatism) and poor dispersal capabilities limited migration out of the hotspot regions. Successful migration may have been achieved only rarely, perhaps by chance, or perhaps by unusually good dispersers which then gave rise to clades of descendants that inherited the traits associated with high dispersal ability. Unfortunately, this hypothesis will be difficult to test in Banksia until good data on dispersal ability (e.g. [28, 29] become available for many species. Even then, with only two comparisons it would be difficult to support this hypothesis to the exclusion of possible confounding factors.
While we found little firm evidence that diversification rates differ between SWBP and non-SWBP clades of Banksia, the GeoSSE analysis revealed differences in rates among the rainfall zones within the SWBP itself. In the transitional rainfall zone (TRZ) and south coast province (SCP), speciation rates are higher, and extinction rates lower, than in the high rainfall zone (HRZ). In addition, rates of migration from the TRZ/SCP to the HRZ were higher than in the opposite direction. These results point to a scenario where the sandplain heaths and shrublands of the TRZ/SCP act as a “biodiversity pump”, responsible for generating much of the Banksia diversity of the SWBP, with diversity of the adjacent coastal forests of the HRZ elevated by migration out of the TRZ/SCP. This scenario is at least partly consistent with Hopper’s geohistorical hypothesis [2, 10]. Under Hopper’s hypothesis, high speciation rates in the TRZ/SCP resulted from increasing landscape complexity in the later part of the Tertiary Period, following a long period of relative stability. The resulting fragmentation of populations may have led to widespread allopatric speciation and the formation of numerous narrowly-distributed endemic species. However, this scenario does not account for the reduced extinction rates in the TRZ/SCP, compared to the HRZ, that the GeoSSE analysis indicates. Indeed, Hopper suggested that the relative stability of HRZ landscapes should have led to lower extinction rates there, and the accumulation of “relict” species. One possibility is that regional-scale variation in fire frequency and intensity have played an important role in regulating extinction rates. If fires were more frequent and intense in the densely-vegetated sclerophyll forests, compared to the more sparsely-vegetated heaths and shrublands, extinction resulting from population fragmentation may have been more frequent in the HRZ. There is some evidence that the forests have been more fire-prone than the heaths and shrublands over the comparatively recent period of human occupation of southwestern Australia , but there are little data with which to infer geographic variation in fire patterns across the region over the deeper timescales relevant to macroevolution .
Mediterranean-climate hotspots of botanical diversity are among the most striking, but least well-understood, of the world’s large-scale biodiversity patterns. We have found little evidence that the diversity of Banksia, one of the largest genera in the southwest Australian hotspot, is the result of anything other than a steady accumulation of species over a long period of time. This contrasts markedly with the view that the great diversity of Mediterranean hotspots is underpinned by rapid radiations of plant genera [5, 7]. There are, of course, other genera (such as Hakea) that appear to have radiated much more rapidly than Banksia, and also have their greatest diversity in the SWBP. However, there can be no basis for assuming that diversification in Hakea and other genera has been more rapid in the hotspot region, until comparisons of diversification rates across the hotspot boundaries, similar to those we have presented here, are made. In order to understand the relative contribution of rapid radiations versus steady diversification to the botanical diversity of Mediterranean hotspots, therefore, it will be necessary to produce and analyse densely-sampled phylogenies of multiple hotspot genera.
Availability of supporting data
The data sets supporting this article are included within the article and its additional files.
The full posterior set of Banksia phylogenies from the BEAST analysis will be deposited in the Data Dryad repository (http://datadryad.org/) upon acceptance of the manuscript for publication.
We are grateful to Kevin Collins for permission to collect tissue samples from the Banksia Farm, and to Margaret Pieroni for collecting and posting herbarium specimens. Thanks to Geoff Kay and Lindell Bromham for assisting with field collection of samples, and to Terena Lally for access to specimens at the Australian National Herbarium. The work was funded by an Australian Research Council Discovery Grant and QEII Fellowship to M. Cardillo (DP0879971), with additional funding provided by the Research School of Biology at the Australian National University.
- Cowling RM, Rundel PW, Lamont BB, Arroyo MK, Arianoutsou M: Plant diversity in Mediterranean-climate regions. Trends in Ecology & Evolution. 1996, 11 (9): 362-366. 10.1016/0169-5347(96)10044-6.View ArticleGoogle Scholar
- Hopper SD, Gioia P: The southwest Australian floristic region: evolution and conservation of a global hot spot of biodiversity. Annual Review of Ecology and Systematics. 2004, 35: 623-650. 10.1146/annurev.ecolsys.35.112202.130201.View ArticleGoogle Scholar
- Linder HP: The radiation of the Cape flora, southern Africa. Biological Reviews. 2003, 78 (4): 597-638. 10.1017/S1464793103006171.PubMedView ArticleGoogle Scholar
- Verboom GA, Archibald JK, Bakker FT, Bellstedt DU, Conrad F, Dreyer LL, Forest F, Galley C, Goldblatt P, Henning JF, et al: Origin and diversification of the greater cape flora: ancient species repository, hot-bed of recent radiation, or both?. Molecular Phylogenetics and Evolution. 2009, 51 (1): 44-53. 10.1016/j.ympev.2008.01.037.PubMedView ArticleGoogle Scholar
- Latimer AM, Silander JA, Cowling RM: Neutral ecological theory reveals isolation and rapid speciation in a biodiversity hot spot. Science. 2005, 309 (5741): 1722-1725. 10.1126/science.1115576.PubMedView ArticleGoogle Scholar
- Linder HP: Plant species radiations: where, when, why?. Philosophical Transactions of the Royal Society B-Biological Sciences. 2008, 363 (1506): 3097-3105. 10.1098/rstb.2008.0075.PubMed CentralView ArticleGoogle Scholar
- Sauquet H, Weston PH, Anderson CL, Barker NP, Cantrill DJ, Mast AR, Savolainen V: Contrasted patterns of hyperdiversification in Mediterranean hotspots. Proceedings of the National Academy of Sciences of the United States of America. 2009, 106 (1): 221-225. 10.1073/pnas.0805607106.PubMed CentralPubMedView ArticleGoogle Scholar
- Valente LM, Reeves G, Schnitzler J, Mason IP, Fay MF, Rebelo TG, Chase MW, Barraclough TG: Diversification of the African genus Protea (Proteaceae) in the Cape biodiversity hotspot and beyond: equal rates in different biomes. Evolution. 2010, 64 (3): 745-760. 10.1111/j.1558-5646.2009.00856.x.PubMedView ArticleGoogle Scholar
- Verboom GA, Dreyer LL, Savolainen V: Understanding the origins and evolution of the world’s biodiversity hotspots: The biota of the African ‘Cape Floristic Region’ as a case study. Molecular Phylogenetics and Evolution. 2009, 51 (1): 1-4. 10.1016/j.ympev.2009.03.001.PubMedView ArticleGoogle Scholar
- Hopper SD: Biogeographical aspects of speciation in the southwest Australian flora. Annual Review of Ecology and Systematics. 1979, 10: 399-422. 10.1146/annurev.es.10.110179.002151.View ArticleGoogle Scholar
- Cowling RM: Fire and its role in coexistence and speciation in gondwanan shrublands. South African Journal of Science. 1987, 83 (2): 106-112.Google Scholar
- Hopkins AJM, Griffin EA: Floristic patterns. Kwongan: Plant Life of the Sandplain. Edited by: Pate JS, Beard JS. 1984, Nedlands: University of Western Australia Press, 69-83.Google Scholar
- Hassell C, Dodson J: The fire history of south-west Australia prior to European settlement in 1826–1829. Fire in ecosystems of south-west Western Australia: Impacts and management. Edited by: Abbott I, Burrows N. 2003, Leiden: Backhuys Publishers, 71-85.Google Scholar
- Mast AR, Jones EH, Havery SP: An assessment of old and new DNA sequence evidence for the paraphyly of Banksia with respect to Dryandra (Proteaceae). Australian Systematic Botany. 2005, 18 (1): 75-88. 10.1071/SB04015.View ArticleGoogle Scholar
- Mast AR, Thiele K: The transfer of Dryandra R.Br. to Banksia L.f. (Proteaceae). Australian Systematic Botany. 2007, 20 (1): 63-71.View ArticleGoogle Scholar
- Goldberg EE, Lancaster LT, Ree RH: Phylogenetic inference of reciprocal effects between geographic range evolution and diversification. Systematic Biology. 2011, 60 (4): 451-465. 10.1093/sysbio/syr046.PubMedView ArticleGoogle Scholar
- Drummond AJ, Ashton B, Buxton S, Cheung M, Cooper A, Duran C, Field M, Heled J, Kearse M, Markowitz S, et al: Geneious. 2012, http://www.geneious.com/,Google Scholar
- Drummond AJ, Rambaut A: BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evolutionary Biology. 2007, 7: 214-10.1186/1471-2148-7-214.PubMed CentralPubMedView ArticleGoogle Scholar
- Crisp MD, Cook LG: A congruent molecular signature of vicariance across multiple plant lineages. Molecular Phylogenetics and Evolution. 2007, 43 (3): 1106-1117. 10.1016/j.ympev.2007.02.030.PubMedView ArticleGoogle Scholar
- Ho SYW, Phillips MJ: Accounting for calibration uncertainty in phylogenetic estimation of evolutionary divergence times. Systematic Biology. 2009, 58: 367-380. 10.1093/sysbio/syp035.PubMedView ArticleGoogle Scholar
- Schliep KP: phangorn: phylogenetic analysis in R. Bioinformatics. 2011, 27: 592-593. 10.1093/bioinformatics/btq706.PubMed CentralPubMedView ArticleGoogle Scholar
- Rambaut A, Drummond AJ: Tracer. 2012, v1.5: http://beast.bio.ed.ac.uk/Tracer,Google Scholar
- Collins K, Collins K, George A: Banksias. 2008, Melbourne: Bloomings BooksGoogle Scholar
- Cavanagh T, Pieroni M: The Dryandras. 2006, Hawthorn, Victoria: Australian Plants Society Inc.Google Scholar
- Maddison WP, Midford PE, Otto SP: Estimating a binary character's effect on speciation and extinction. Systematic Biology. 2007, 56 (5): 701-710. 10.1080/10635150701607033.PubMedView ArticleGoogle Scholar
- Pybus OG, Harvey PH: Testing macro-evolutionary models using incomplete molecular phylogenies. Proceedings of the Royal Society of London Series B-Biological Sciences. 2000, 267 (1459): 2267-2272. 10.1098/rspb.2000.1278.View ArticleGoogle Scholar
- Davis M, Midford P, Maddison W: Exploring power and parameter estimation of the BiSSE method for analyzing species diversification. BMC Evolutionary Biology. 2013, 13 (1): 38-10.1186/1471-2148-13-38.PubMed CentralPubMedView ArticleGoogle Scholar
- Merwin L, He TH, Lamont BB, Enright NJ, Krauss SL: Low rate of between-population seed dispersal restricts genetic connectivity and metapopulation dynamics in a clonal shrub. PloS One. 2012, 7 (11): e50974-10.1371/journal.pone.0050974.PubMed CentralPubMedView ArticleGoogle Scholar
- Robledo-Arnuncio JJ: Joint estimation of contemporary seed and pollen dispersal rates among plant populations. Molecular Ecology Resources. 2012, 12 (2): 299-311. 10.1111/j.1755-0998.2011.03092.x.PubMedView ArticleGoogle Scholar
- Burrows N, Abbott I: Fire in south-west Western Australia: synthesis of current knowledge, management implications and new research directions. Fire in ecosystems of south-west Western Australia: Impacts and management. Edited by: Abbott I, Burrows N. 2003, Leiden: Backhuys Publishers, 437-452.Google Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.