New Caledonia is a global biodiversity hotspot. Hypotheses for its biotic richness suggest either that the island is a ‘museum’ for an old Gondwana biota or alternatively it has developed following relatively recent long distance dispersal and in situ radiation. The conifer genus Araucaria (Araucariaceae) comprises 19 species globally with 13 endemic to this island. With a typically Gondwanan distribution, Araucaria is particularly well suited to testing alternative biogeographic hypotheses concerning the origins of New Caledonian biota. We derived phylogenetic estimates using 11 plastid and rDNA ITS2 sequence data for a complete sampling of Araucaria (including multiple accessions of each of the 13 New Caledonian Araucaria species). In addition, we developed a dataset comprising 4 plastid regions for a wider taxon sample to facilitate fossil based molecular dating. Following statistical analyses to identify a credible and internally consistent set of fossil constraints, divergence times estimated using a Bayesian relaxed clock approach were contrasted with geological scenarios to explore the biogeographic history of Araucaria. The phylogenetic data resolve relationships within Araucariaceae and among the main lineages in Araucaria, but provide limited resolution within the monophyletic New Caledonian species group. Divergence time estimates suggest a Late Cretaceous-Cenozoic radiation of extant Araucaria and a Neogene radiation of the New Caledonian lineage. A molecular timescale for the evolution of Araucariaceae supports a relatively recent radiation, and suggests that earlier (pre-Cenozoic) fossil types assigned to Araucaria may have affinities elsewhere in Araucariaceae. While additional data will be required to adequately resolve relationships among the New Caledonian species, their recent origin is consistent with overwater dispersal following Eocene emersion of New Caledonia but is too old to support a single dispersal from Australia to Norfolk Island for the radiation of the Pacific Araucaria sect. Eutacta clade.
Citation: Kranitz ML, Biffin E, Clark A, Hollingsworth ML, Ruhsam M, Gardner MF, et al. (2014) Evolutionary Diversification of New Caledonian Araucaria. PLoS ONE 9(10): e110308. doi:10.1371/journal.pone.0110308
Editor: William J. Etges, University of Arkansas, United States of America
Received: June 12, 2014; Accepted: September 11, 2014; Published: October 23, 2014
Copyright: © 2014 Kranitz et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The authors confirm that all data underlying the findings are fully available without restriction. All relevant data are within the paper and its Supporting Information files.
Funding: This work was supported by funds from the Darwin Trust of Edinburgh and the ANR project BIONEOCAL, the PPF MNHN ‘Structure et évolution des écosystèmes’ led by P. Grandcolas. The work of AJL and EB was supported by the Australian Research Council through a Super Science Fellowship (FS1102 00051) and Discovery Grant (DP0665859) awarded to AJL. The Royal Botanic Garden Edinburgh is supported by the Scottish Government's Rural and Environment Science and Analytical Services Division. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
New Caledonia, in the tropical south-western Pacific, is noted for a remarkably rich biota with outstanding levels of endemism . The island has been considered both a ‘museum’ for phylogenetic relicts and a natural laboratory for the study of island radiations . A long held view is that New Caledonia (as part of Zealandia) was rifted from the eastern Australian margin some 80 million years before present (Ma), carrying a complement of Gondwanan lineages that have since evolved there in relative isolation . Alternatively, geological reconstructions for the region indicate lengthy deep-water marine transgressions (Palaeocene-Eocene) with limited evidence of a continual landmass prior to the emersion of New Caledonia in the Late Eocene-Oligocene (c. 37 Ma) . This implies that dispersal events must be central in the assembly of the modern biota . Phylogenetic and molecular dating methods, within the context of a well-constrained geological scenario, can be used to explore these alternative biogeographic scenarios . With respect to New Caledonia such studies provide strong support for the latter view , .
The conifer genus Araucaria Juss. (Araucariaceae) provides a good model to explore alternative hypotheses regarding the age, origins and diversification of the New Caledonian biota , . Araucaria has been considered an old lineage  and for instance, fossil data from both the Northern and Southern Hemisphere of the early Mesozoic (Jurassic, c. 190 Ma, or earlier) have been taken to indicate that at least some of the extant sections of Araucaria had begun to diversify and were widespread by this time. Thirteen of the 19 recognised Araucaria species are now endemic to New Caledonia, with the other species occurring in eastern Australia and New Guinea (3 spp., 1 shared), Norfolk Island (1 sp.) and southern South America (2 spp.). In light of these data extant Araucaria have been considered a relict of a formerly widespread Gondwanan distribution .
A key question is the biogeographic origin of the New Caledonian Araucaria – an old lineage isolated by vicariance that has diversified in situ or a relatively recent radiation following long distance dispersal? There have been some indirect tests of these hypotheses. Setoguchi et al.  used plastid rbcL sequence data to infer the phylogeny of Araucariaceae and found support for the monophyly of the New Caledonian Araucaria. In this study, however, the morphological species were characterised by very low levels of DNA sequence divergence suggestive of a recent radiation. Gaudeul et al.  also argued for a recent diversification of the New Caledonian Araucaria based on a low ratio of among versus within species genetic divergence using AFLP markers. In the context of these findings it has been proposed that the origin of this lineage may be maximally constrained by the age of Norfolk Island (c. 1000 km to the south of New Caledonia) , a volcanic island that in its present form emerged c. 3.7 Ma. This island is home to the sister species (A. heterophylla (Salisb.) Franco; Norfolk Island Pine) of the New Caledonian Araucaria lineage , . However, the use of island age to date island-endemic lineages can be problematic as these taxa could represent historically widespread lineages that have been isolated by extinction . Limitations of the molecular clock have also been well-characterised  and there may be a poor correlation between levels of DNA sequence divergence and divergence times if rates have been slow relative to related taxa. Furthermore, finding low levels of divergence among molecular markers does not rule out the possibility of an ancient lineage and a recent diversification in the absence of a timescale for the origin of the New Caledonian lineage . Recent analyses of araucarian and conifer divergence times – for the most part suggest a relatively recent radiation of extant Araucariaceae and Araucaria but as yet there has been no formal attempt to date the diversification of Araucaria on New Caledonia.
In the present study, we develop a phylogenetic estimate for Araucaria using rDNA internal transcribed spacer 2 (ITS2) sequences and a mix of 11 coding and non-coding plastid data. We include a complete sampling of species and several accessions of each of the 13 New Caledonian representatives. Our overall aims are: (1), to clarify relationships among Araucaria, particularly the New Caledonian species; (2), using well-constrained fossil dates and molecular clock methods, to estimate a timescale for Araucaria evolution focussing on the origins of the New Caledonian radiation.
Plant material and DNA extractions
Collection, voucher and permit details of all samples used for the phylogenetic analyses are listed in Table S1. Permits allowed for the collection of samples from threatened species and for collection of samples from within protected areas. CITES permits were obtained for Araucaria araucana. DNA was extracted using the Qiagen DNeasy Mini Plant Kit following the manufacturer's protocol.
PCR amplification and sequencing
DNA sequences were obtained from 11 chloroplast regions (matK, rcbL, rpoB, rpoC1, atpF-atpH, atpH-atpI, rps12-rpl20, trnC-ycf6, trnH-psbA, trnS-trnfm, trnS-trnG) and one nuclear region (ITS2). For a subset of 21 taxa (nine Agathis species, and outgroup taxa including a single representative each of the genera Pinus L. (Pinaceae), Taxus L. (Taxaceae), Cryptomeria D.Don (Cupressaceae), Taxodium Rich. (Taxodiaceae), Lepidothamnus Phil., Halocarpus Quinn, Falcatifolium de Laub., Dacrycarpus de Laub., Phyllocladus Rich. ex Mirb., Podocarpus L'Hér. ex Pers., Retrophyllum C.N.Page and Sundacarpus (J.Buchholz and N.E. Gray) C.N. Page (all Podocarpaceae)) we assembled a data set comprising rbcL, matK and non-coding plastid regions atpF-atpH and rps12-rps20 to supplement the Araucaria data for molecular dating analyses (see below). These regions were chosen because they could be unambiguously aligned in all taxa.
PCR primers and conditions for matK, rcbL, rpoB, rpoC1, atpF-atpH and trnH-psbA are detailed in Appendix 4 of the supplementary material of . For the remaining regions primer sequences were taken from the following publications: atpH-atpI , rps12-rpl20 and trnC-ycf6 , trnS-trnfm , trnS-trnG  and ITS2 .
PCR reactions for atpH-atpI, rps12-rpl20, trnC-ycf6, trnS-trnG and ITS2 were performed in volumes of 25 µl using the following protocol: 1xbuffer (Bioline, London, UK), 0.2 mM dNTPs, 2.5 mM MgCl2, 0.3 µM of each forward and reverse primer, (4% Dimethyl Sulfoxide, DMSO, ITS2 only), 0.04 U BioTaq (Bioline, London, UK) and 1 µl of unquantified DNA. The mixture was then cycled through the profile: 2 min at 95°C, 40 cycles of 30 sec at 95°C, 30 sec at 50°C and 1 min at 72°C, ending with 10 min at 72°C to complete extension and subsequent storage at 4°C.
PCR reactions for trnS-trnfm followed this protocol (25 µL): 1x buffer (Bioline, London, UK), 0.2 mM dNTPs, 2.5 mM MgCl2, 0.3 µM of each forward and reverse primer, 2 M Betaine, 0.03 U BioTaq (Bioline, London, UK) and 2 µl of unquantified DNA. The mixture was then cycled through the following profile: 4 min at 94°C, 30 cycles of 45 sec at 94°C, 45 sec at 62°C and 2 min at 72°C, ending with 10 min at 72°C and subsequent storage at 4°C.
PCR products were cleaned using 2 µl exonuclease I and shrimp alkaline phosphatase (ExoSAP) for 5 µl of product and sequenced in two reactions using each of the two PCR primers, following Big Dye v. 3.1 chemistry (Applied Biosystems, Warrington, UK). Sequences were then assembled and aligned using Sequencher v. 3.7 (GeneCodes Corp., Ann Arbor, Michigan, USA).
Araucaria phylogeny was inferred separately for the plastid (53 taxa), ITS2 (48 taxa) and a combined plastid and ITS2 DNA sequence data set (48 taxa) using both maximum likelihood (ML) and Bayesian inference (BI) optimisation criteria (Table S1; vouchers).
We used GARLI 0.951  for the ML analyses with a general time reversible model of sequence evolution, gamma distributed rate variation and a proportion of invariant sites (GTR+I+Γ). Clade support was assessed using 250 non parametric bootstrap (BS) pseudoreplicates.
Bayesian inference was performed using MrBayes 3.1.2 , . For the plastid and combined data sets we used a mixed model approach with GTR+I+Γ model parameters estimated separately for each locus in the concatenated matrices. For the analysis of the ITS2 sequences a GTR+I+Γ model was used. For each analysis we performed four independent runs of 2×106 generations, sampling topology and parameter values every hundredth generation, with default (flat) priors, a random starting tree and four starting chains (one cold, three heated). Convergence between runs was assessed relative to the variance in parameter estimates (average standard deviation of split frequencies <0.01; Ronquist 2005) and by inspection of parameter estimates in Tracer v1.5 . Tracer was used to estimate the burn in proportion, and 50% majority rule consensus topologies were generated from the pooled post-burn in topologies from the four independent runs.
To estimate topology and divergence times for Araucariaceae, we assembled a data set comprising partial sequences of the plastid genes rbcL, matK and non-coding plastid regions atpH-atpI and rps12-rps20 for 45 taxa). Taxon sampling includes all of the 19 currently recognised species of Araucaria, 11 species of Agathis Salisb., Wollemia (monotypic), and outgroup taxa including Cupressaceae, Taxaceae, Podocarpaceae and Pinaceae (Table S1).
Five fossil dates were used to calibrate molecular evolutionary rates amongst Araucariaceae. These comprise two fossil constraints that have been assigned to the Araucariaceae crown group, two within the Podocarpaceae as well as a constraint upon the age of the Cupressaceae sensu lato (Table 1). The araucarian fossils include the oldest unequivocal macrofossil remains of Agathis from south-eastern Australia of the Late Palaeocene (A. vittata R.S.Hill et al., c. 55 Ma)  and a recently described bract-scale complex from the Lower Jurassic (c. 190 Ma) of eastern North America, which has been proposed as the earliest known occurrence of Araucaria section Eutacta (Link) Endl. . It should be noted that this fossil age is not inconsistent with several other well documented fossil types (e.g. Araucaria mirabilis (Spegazzini) Windhausen, A. sphaerocarpa Carruthers) that have suggested affinities within extant Araucaria . Several extant podocarp lineages have reliable macrofossil records including Prumnopitys Phil. ,  and Acmopyle Pilg. , . We used macrofossil remains from K-T boundary exposures (c. 65 Ma) of New Zealand  to constrain the crown group age for Prumnopitys sensu lato (most recent common ancestor, hereafter MRCA, of Prumnopitys + Sundacarpus; see ) and defined a minimum age of 55 Ma to constrain the Acmopyle stem (MRCA Podocarpus + Retrophyllum + Dacrycarpus + Falcatifolium + Acmopyle, based on A. florinii R.S.Hill and Carpenter). The fossil taxon Austrohamia minuta Escapa, Cúneo and Axsmith, from the Middle-Late Jurassic has been placed within Cupressaceae sensu lato following a cladistic analysis of morphological characters from fossil and living taxa . On this basis we assumed a minimum age of 160 Ma to constrain the Cupressaceae stem (MRCA Cupressaceae s.l. + Taxaceae) (Table 1; Figure 1).
The topology presented is the maximum credibility tree (median node heights) obtained by pooling four separate runs under two constraint scenarios (see Table 1, and text for details). Circles indicate nodes that were included in the assessment of candidate nodal constraints, and the dark shaded circles indicate nodes that were constrained for the BRC analyses (see Table 1 for details). Node bars (nodes with a posterior probability ≥0.75) indicate the 95% HPD of divergence times for the two prior scenarios: left, log normal prior mean = minimum fossil age +50%; right, log normal prior mean = minimum fossil age +10% (NC, New Caledonia; NI, Norfolk Island).
Assessment of fossil constraints.
We adopted the approach of Dornburg et al.  to provide an assessment of the palaeontological data described above. These authors extend the methods of Marshall , which use minimum node age estimates derived from the fossil record and an uncalibrated ultrametric tree to bracket absolute divergence times from molecular phylogenies. Given that for a particular lineage the known fossil record is generally incomplete, Marshall  provides an approach to estimate the coverage of a fossil assigned to that clade (empirical scaling factor, Si, i.e. the proportion of the true duration of the lineage represented by its oldest known fossil) which can be used to estimate the true scaling factor, S, i.e. the actual time of divergence of the root. To incorporate uncertainty in Si estimates, Dornburg et al.  sample a distribution of ultrametric tree topologies and relative branch lengths. Rather than applying an age bracket to the single fossil with the highest coverage , they identify a set of internally consistent fossil constraints based upon the Si estimates and this set of consistent calibrations is incorporated in molecular dating analyses to derive absolute estimates of node heights .
We estimated the posterior distribution of uncalibrated ultrametric trees (i.e. branch lengths in units of substitutions/site) from the molecular data using the Bayesian relaxed clock (BRC) application BEAST (v.1.6.1)  with a GTR+I+Γ model of sequence evolution, a Yule prior on branch rates and an uncorrelated log normal relaxed clock model . For these analyses two independent Markov-chain Monte Carlo runs were performed, each of 4×107 steps (sampling topology and parameter values every 5000 steps). Tracer v1.5  was used to assess convergence between runs and estimate an appropriate burn-in proportion, the mean and 95% highest posterior density (HPD) of parameters sampled from the posterior distribution of the combined runs, and to ensure that the effective sample size was sufficient to provide reasonable estimates of model parameter variance (i.e.>200).
For each fossil constraint (Table 1) Si was estimated using Equation 1 of Marshall  and the distribution of Si values was estimated from 1000 topologies sampled from posterior distribution of uncalibrated ultrametric trees. In order to assess the internal consistency of the fossil constraints, we simultaneously compared the interval encompassing 95% of the values for the fossil with the highest Si estimates (i.e. the highest coverage) against all other Si intervals for each fossil. The consistent fossil set is defined as those with 95% Si intervals that overlap the interval with the highest Si values .
Where possible, we assessed the relative placement of the fossil age within a lineage (i.e. crown or stem node). For three of the constraints, taxon sampling (Acmopyle and Cupressaceae) and/or phylogenetic uncertainty (Prumnopitys s.l.) excluded this possibility, but for the Agathis fossil age, we compared crown and stem node placement (MRCA Agathis + Wollemia) (nodes 4 and 5, respectively; Table 1, Figure 1). For the Araucaria section Eutacta fossil age we considered alternative placements on the section Eutacta crown (node 6), Araucaria crown (node 7), Araucariaceae crown (node 8), and on the Araucariaceae stem (MRCA of Araucariaceae + Podocarpaceae; node 9). This is in the light of questions, on a number of grounds , , ,  of generic and/or sectional affinities of Early Mesozoic araucarians.
We placed a caveat on the selection of the fossil with the highest coverage. Fossil prepollen of Potonieisporites D.C.Bhardwaj, a monosaccate type that was produced by walchian conifers, is known from at least as early as the Langsettian (Upper Carboniferous >313–314 Ma) from Scotland  while macrofossils of the walchian conifer Hermitia Kerp and Clement-Westerhof are known from at least as early as the Asturian ( = Westphalian D, 308–306 Ma) of the Netherlands . Extant conifer families have fossil records extending from the Triassic ,  providing an expectation on the ‘upper’ distribution of reasonable Si values. Estimates that consistently exceeded the age of the earliest known conifers were treated as outliers.
Divergence time estimates.
We used the consistent set of fossils identified following the procedure outlined above to estimate araucarian divergence times with BEAST. For each fossil datum, we defined a log normal prior probability distribution on the age of the fossil constrained node, with a hard minimum age equivalent to the age of the associated fossil. The other parameters of the log normal distribution are the mean and standard deviation. For all analyses, the standard deviation was set to 0.5, and to model taphonomic bias of the fossil record, the log normal prior mean was set at c. 10% and 50% older than each minimum fossil age, in separate sets of analyses (Table 1). At the highest value of the mean, the upper bound of the 95% confidence interval (CI) of the log normal constraint priors included values estimated using Marshall's equation 11  approach for estimating a 95% CI for the true time of origin of a lineage, given its oldest known fossil (with n = 5, i.e. the number of fossils included in the consistent set of fossil calibrations) . For each value of the log normal prior mean, two separate BEAST analyses were performed, with model parameters and settings as outlined above. For each mean value, the two separate runs were combined after excluding the burn in fraction, and topologies and parameter values were summarised on the ‘maximum credibility’ tree using Tree Annotator v. 1.6. . For these analyses we additionally constrained the root age with a log normal prior with a mean of c. 300 Ma and a 95% CI ranging from 260–350 Ma approximating the age of stem group conifers .
The alignment of the 11 used plastid regions and the nuclear ribosomal ITS2 comprised 9970 bp and 500 bp, respectively, totalling 10470 bp in the combined dataset. For all Araucaria samples (except A.schmidii), Agathis moorei (Lindl.) Mast., Agathis lanceolata Warb., Wollemia nobilis W.G.Jones et al., Prumnopitys ferruginoides (Compton) de Laub. and Acmopyle pancheri (Brongn. and Gris) Pilg., sequencing was successful for all individuals and regions (Table S2). Certain regions could not be obtained for Prumnopitys ferruginoides (rpoB, trnH-psbA, trnS-trnfm), Acmopyle pancheri (trnH-psbA), A. schmidii de Laub. (ITS2), Taxodium distichum (atpF-atpH), Podocarpus gnidioides (rps12-rpl20) as well as Lepidothamnus fonkii and Retrophyllum rospigliosii (both atpF-atpH and rps12-rpl20) (Table S2).
Phylogeny of Araucaria
Phylogenies estimated from the plastid (Figure 2) and ITS2 (Figure 3) data resolved broadly similar groupings including Wollemia + Agathis and Araucaria. Within Araucaria, section Eutacta is resolved as sister to a clade including the sections Araucaria and Bunya Wilde and Eames + Intermedia C.T.White. ITS2 data (Figure 3) resolved A. heterophylla (Salisb.) Franco as sister to the remaining A. section Eutacta clade, but this placement is relatively weakly supported (BS<50%; PP = 0.93). In the combined plastid and ITS2 data, there is robust support for A. cunninghamii Mudie as sister to the Pacific A. section Eutacta (A. heterophylla+New Caledonian species) (Figure 4). There is strong support for the monophyly of New Caledonian A. section Eutacta, but low resolution within this clade. In the combined data, well-supported groupings (here, defined as having a BS≥75% and PP≥0.95) include A. nemorosa de Laub. + A. luxurians (Brongn. and Gris.) de Laub. + A. columnaris (G.Forst.) Hook., and multiple accessions of some species including A. nemorosa, A. humboldtensis J.Buchholz and A. subulata Vieill. In the plastid data set, the three accessions of A. schmidii are also resolved with strong support, but were not included in the combined analyses because ITS2 sequences could not be obtained (Figures 2–4).
Support values, below the branch, are: ML bootstrap/Bayesian posterior probability. Branches marked with an asterisk have a BS of 100% and a PP of 1.0. Detail of the relationships among New Caledonian Araucaria is shown at the bottom left. Branch lengths are proportional to the inferred number of substitutions along that branch.
Support values are as detailed for Figure 2.
Molecular dating of Araucariaceae
Assessment of fossil constraints.
We used an expanded sampling of taxa and four plastid loci to infer araucarian phylogeny using a BRC model (Figure 1). With respect to relationships, these data are largely consistent with the results reported above.
Scaling factor distributions estimated under a BRC approach with unconstrained branch lengths are shown in Figure 5. Of the potential calibrations tested, the highest Si estimates were associated with the ‘Eutacta’ fossil constraint, although assigning an Early Jurassic age to the Araucaria section Eutacta crown, Araucaria crown or Araucariaceae crown (nodes 6, 7 and 8, respectively; Table 1) results in a distribution of Si values that consistently exceeds the age of the earliest known conifer fossils (i.e.> c. 310 Ma), or indeed, the oldest known land plants. Similarly, constraining the Agathis crown node to a Late Paleocene age returned unrealistic Si estimates. After excluding these outliers, we identified the constraint on node 5 as returning the highest empirical scaling factor, and 4 potential clade age calibrations (fossils associated with nodes 1, 2 3 and 9) with Si intervals that substantially overlap with node 5.
The mean (circle) and 95% highest posterior density interval (lines) of Si is shown for each proposed calibration, including: (a) alternative nodal placements of fossil Agathis and Araucaria (vertical boxes). In (a), the Si intervals are contrasted with the estimated age of conifers (c. 320 Ma; horizontal box). In (b), candidate fossils with Si estimates that consistently exceeded the maximum age of conifers have been removed. The grey shading indicates the region of overlap in Si intervals for the five retained calibrations.
Araucarian divergence times.
Five fossil constraints with overlapping Si intervals (Figure 5; Table 1) were used to estimate araucarian divergence times, varying the mean of the log normal calibrations priors to model taphonomic bias. In general, node age estimates substantially overlapped, irrespective of the log normal prior mean but differed substantially in terms of the upper versus lower values encompassed within the estimated 95% HPD. As there is no strong basis to favour a particular distribution of node age priors, we consider the full range of posterior value estimates as equally plausible (Table 2). Our estimates for the Araucariaceae stem (node 9) range from c. 284–202 Ma; the Araucariaceae crown (node 8), c. 138–60 Ma; the MRCA of Agathis + Wollemia (node 5) is estimated to have diversified at c. 91–55 Ma, while the Agathis crown (node 4) ranges from c. 55–12 Ma. The Araucaria crown (node 7) is estimated to have diverged c.103–29 Ma and our analyses support a preponderance of Cenozoic radiations among extant Araucaria. For instance, the MRCA of sections Araucaria, Bunya and Intermedia radiated between c. 58–9 Ma, while the diversification of crown Eutacta (node 6) is estimated at between c. 51–9 Ma. The radiation of the Pacific Eutacta (MRCA A. heterophylla + New Caledonian Eutacta) is Oligocene-Miocene age (c. 33–5 Ma), and the New Caledonian species appear to have diversified from the Miocene-Pliocene (c. 19–3 Ma) (Table 2).
Several studies have used molecular data to infer Araucariaceae phylogeny , , , –, . These are generally consistent regarding relationships among genera and among the major lineages within Araucaria. While the New Caledonian species of Araucaria section Eutacta consistently form a clade, establishing a robust understanding of interspecific relationships has proven difficult , , . AFLP markers resolved three main genetic groups including a small leaved, large leaved and a coastal species group . Escapa and Catalano  presented an analysis of a combined morphological and molecular data set, and found reasonable agreement with these interspecific groupings but their hypothesis differed from Gaudeul et al.  in the resolution of group inter-relationships. In the present study, employing a relatively large amount of sequence data (Table S1) did return some well supported resolutions among the New Caledonian Araucaria including the coastal group (A. nemorosa, A. luxurians, A. columnaris) as noted above. However, additional lines of evidence are clearly required to more fully resolve relationships and to test the monophyletic status of some of the species (Figures 3 and 4). Low levels of genetic divergence in previous studies and ours argue for a relatively recent and rapid radiation of the New Caledonian species , , .
Assessment of fossil constraints
Molecular dating has become a routine approach in molecular phylogenetic studies, and is central to evolutionary interpretation . While there have been several advances in molecular clock methods, a key issue is the placement of fossil constraints, which, regardless of the clock methods employed, can strongly influence branch lengths. Several recent studies have used cross-validation approaches to identify sets of internally consistent constraints , , .
The fossil record of Araucariaceae extends to the Triassic, although in many cases the affinities of pre-Jurassic remains are doubtful . Nevertheless, several fossil types from the Early Mesozoic have been assigned to extant araucarian lineages , , . In the light of our analyses using unconstrained branch lengths, we find little support for this approach: assigning an Early Jurassic age to the Araucaria section Eutacta crown, Araucaria crown or Araucariaceae crown (nodes 6, 7 and 8, respectively; Table 1) results in a distribution of Si values that are unrealistic in the context of the evolutionary time scale of land plants . We interpret these values as reflecting incorrect phylogenetic placement of the ‘Eutacta’ fossil constraint which would bias age estimates to be too old if retained in molecular clock analyses. Indeed, Biffin et al.  found that minimally constraining Araucaria to a Middle Cretaceous age resulted in implausibly high levels of molecular rate variation between stem and crown Araucariaceae or alternatively, unrealistically old estimates for the divergence of extant conifers. In the present study, assuming an Early Jurassic age for stem Araucariaceae (node 9) returned a distribution of Si estimates that substantially overlapped with the intervals for four other fossils, and in light of the conifer fossil record, provides a credible estimate of the age of the root  (Figures 1 and 5; Table 1).
If the Araucaria crown group radiated not more than c. 100 Ma (Figure 1; Table 2) this casts doubt on the status of several relatively ancient fossil types with putative affinities to extant taxa (e.g. Axsmith et al., 2008). However, aspects of this record have previously been questioned on a number of grounds. Firstly, several authors have noted that the extant diversity of Araucariaceae almost certainly underestimates past diversity and morphological variation within the family , , , . In this context, araucarian fossils have been described that possess unique combinations of character states but these overlap with extant variation, e.g. Yezonia Stopes and Fujii , making the placement of fragmentary fossil material problematic. Secondly, in the light of the discovery of Wollemia and its phylogenetic placement in the family, it has been suggested that important characters used to diagnose Araucaria and its extant sections may be plesiomorphic for the family , , . Thirdly, the low levels of molecular sequence divergence among extant Araucaria are difficult to reconcile with a Jurassic fossil record for these lineages . Taken together, it has been suggested that the pre-Cretaceous fossil record of putative Araucaria may entirely represent stem Araucariaceae, or (at best) stem lineages of the extant genera , . However, the molecular dates presented here are not inconsistent with a hypothesis that crown Araucaria had evolved by the late Early Cretaceous as indicated by fossil taxa assigned to living sections, e.g. A. grandiflora Feruglio and A. otwayensis Cantrill, estimated at c. 105–95 Ma from Patagonia and southern Australia and referred to sections Araucaria and Eutacta, respectively , ,  (Table 2, Figure 1).
The fossil record for the crown Agathis clade (i.e. Agathis+Wollemia) extends from the Late Cretaceous in reasonable agreement with our molecular age estimates for the corresponding nodes (Figure 1, Table 2). While there are no formal descriptions from the macrofossil record, pollen that is almost indistinguishable from that of extant Wollemia nobilis (Dilwynites W.K.Harris sp.) is first recorded in Turonian (c. 93–89 Ma) strata of Australia and the Maastrichtian (c. 70–65 Ma) of New Zealand, representing the oldest probable record of that genus  (but see ). In a recent review of the macrofossil record of Agathis it was concluded that the oldest unequivocal remains for that genus are from the Late Palaeocene (c. 55–50 Ma) of Southern Australia . Putative Agathis macrofossils have been recorded from the Cretaceous but lack organic preservation, e.g. A. victoriensis Cantrill of the Lower Cretaceous, c. 110 Ma, of Southern Australia  or include characters that are atypical of modern Agathis ,  and therefore remain equivocal . Using a cladistic analysis of morphological characters, the fossil taxa Wairarapaia mildenhallii Cantrill and Raine, and Emwadea macrocarpa Dettmann, Clifford and Peters, were resolved as stem group lineages of the Agathis-Wollemia clade  and date from the Early Cretaceous of New Zealand and Australia, respectively , .
The present distribution of Araucaria, which is Gondwanan in character, is often interpreted as relictual, reflecting regional (Northern Hemisphere) extinction, localised (Southern Hemisphere) contraction and Gondwanan tectonic events , . Our molecular dating scenario is consistent with aspects of this view. For instance, there is evidence for the presence of Araucaria in Europe until c. 66 Ma, which has been proposed as the youngest reliable record of the genus for the Northern Hemisphere . In light of our findings, this date does not reject placement of these fossils within crown Araucaria (Figure 1) although in contrast to some interpretations of the Northern Hemisphere fossil record, our dated phylogeny suggests that the crown groups of the extant sections are derived entirely from Cenozoic Southern Hemisphere radiations. The regional (Northern Hemisphere) extinction of Araucariaceae was probably multicausal . It has been linked to palaeoclimatic and vegetation change  as well as to short term environmental perturbations at the Cretaceous-Palaeogene (K-Pg) boundary (c. 65 Ma) following the asteroid impact at Yucatan (see also  where the virtual disappearance of Araucaria at the K-Pg boundary of New Zealand is noted). A recent study documents the survival of a seed fern lineage (Komlopteris Barbacka) into the Eocene of Tasmania . These authors hypothesise that a delayed radiation of the angiosperms at high latitudes, and the remoteness of southern Gondwana from the Yucatan impact site may have afforded refuge to Komlopteris and other ‘archaic’ gymnosperm lineages into the Cenozoic . Given this hypothesis, Araucaria, too might have persisted in isolated Gondwanan refuges following the K-Pg boundary event and radiated from these during the Palaeocene . In this case, one implication is that the pre-Cretaceous Northern Hemisphere fossil record of Araucariaceae may have little direct bearing on the evolution of extant Araucaria.
The importance of Gondwanan tectonic vicariance versus long distance dispersal in the assembly of the Southern Hemisphere flora has been widely debated and generally, the latter appears to be prominent . Dispersal is implicated by clade origination times (i.e. stem group age estimates) that are younger than the timing of geological events . In the context of our data, vicariance is not rejected for the divergence of the clade that includes Araucaria section Araucaria (southern South America) and sections Bunya + Intermedia (Australia and New Guinea, respectively). Southern South America and Australia remained connected via Antarctica until the Neogene ,  and potentially harboured a widespread ancestor of these lineages .
The biogeography of New Caledonia has been considered as an example where geological and biological evidence are at odds , , , . New Caledonia supports an apparently ancient biota  despite the lack of firm geological evidence for a continuously emergent land surface following the Late Cretaceous-Palaeogene rifting of New Caledonia (as part of the continental fragment, Zealandia) from East Gondwana , . The re-emergence of New Caledonia c. 37 Ma places an upper limit on the age of the island's biota. However, the radiation of Araucaria section Eutacta on New Caledonia may be younger still if the emersion of Norfolk Island (c. 3.7 Ma) is assumed to constrain the possible age of Pacific Araucaria section Eutacta  (Figures 1 and 4). In the present study, we find that the relevant divergence (c. 33–5 Ma) is too old to support this hypothesis and implicitly rejects a single long distance dispersal event from Australia to Norfolk Island for the MRCA of Pacific Araucaria section Eutacta. Alternative plausible scenarios for the origin of A. heterophylla involve a dispersal event from Australia to New Caledonia and a subsequent dispersal from either New Caledonia or Australia. In either case, the second dispersal event requires an extinction event post-dating the emersion of Norfolk Island involving the A. heterophylla lineage in New Caledonia, or the Pacific Araucaria section Eutacta lineage in Australia. Given the uncertainty in divergence time estimates for the Eutacta crown (51-9 Ma), the origin of New Caledonian Araucaria could be explained by a single dispersal or alternatively by short range dispersals among ephemeral landmasses such as those identified or are believed to have existed on the South Norfolk Rise ,  and in the Greater New Caledonia region , . In either case, our data do reject a ‘museum’ type hypothesis for New Caledonian Araucaria and implicate at least some over water dispersal, given that the New Caledonian endemic clade has arisen entirely after the estimated timing of the Palaeocene-Eocene marine transgressions and has radiated in the Neogene (c. 19-3 Ma; Table 2, Figure 1).
Accession details of individuals used in the phylogenetic analyses. Country abbreviations in the ‘Location’ column are AU…Australia, AR…Argentina, CL…Chile, NC…New Caledonia. Superscripts denote cultivated material: 1 Royal Botanic Garden Edinburgh (UK), 2 Mount Lofty Botanic Garden (AU), 3 Koishikawa Botanic Garden (Japan), 4 Adelaide Botanic Garden (AU). Voucher numbers refer to herbarium specimens deposited in Edinburgh (E), University of Adelaide (ADU) or Allan Herbarium, Christchurch (CHR).
Genbank accession numbers of all samples used for the phylogenetic analyses.
We are grateful to IRD Nouméa – in particular G. Dagostini, T. Jaffré, J. Munzinger and F. Rigault – for logistical support in the field, and to the Province Sud and Province Nord authorities of New Caledonia for granting collection permits. We thank the Adelaide Botanic Garden of South Australia for access to living collections. We also thank Bob Hill, Richard Winkworth and an anonymous reviewer for feedback on earlier versions of the manuscript. The Royal Botanic Garden Edinburgh is supported by the Scottish Government's Rural and Environment Science and Analytical Services Division.
Conceived and designed the experiments: PMH RAE MK. Performed the experiments: MK AC MLH EB. Analyzed the data: MK EB PMH. Contributed reagents/materials/analysis tools: MK AC MLH RAE MFG MG PT RRM AJL. Wrote the paper: EB PMH MR RRM.
- 1. Myers N, Mittermeier RA, Mittermeier CG, da Fonseca GAB, Kent J (2000) Biodiversity hotspots for conservation priorities. Nature 403: 853–858.
- 2. Grandcolas P, Murienne J, Robillard T, Desutter-Grandcolas L, Jourdan H, et al. (2008) New Caledonia: A very old Darwinian island? Philos Trans R Soc Lond B Biol Sci 363: 3309–3317.
- 3. Morat P (1993) Our knowledge of the flora of New Caledonia: endemism and diversity in relation to vegetation types and substrates. Biodiv Lett 1: 72–81.
- 4. Pelletier B, editor (2006) Geology of the New Caledonia region and its implications for the study of the New Caledonian biodiversity. Nouméa, New Caledonia: Documents Scientifiques et Techniques, Vol. II 7, 2nd ed. IRD.
- 5. Murienne J (2009) Testing biodiversity hypotheses in New Caledonia using phylogenetics. J Biogeogr 36: 1433–1434.
- 6. Pillon Y (2012) Time and tempo of diversification in the flora of New Caledonia. Bot J Linn Soc 170: 288–298.
- 7. Gaudeul M, Rouhan G, Gardner MF, Hollingsworth PM (2012) AFLP markers provide insights into the evolutionary relationships and diversification of New Caledonian Araucaria species (Araucariaceae). Am J Bot 99: 68–81.
- 8. Setoguchi H, Asakawa Osawa T, Pintaud J-C, Jaffré T, Veillon J-M (1998) Phylogenetic relationships within Araucariaceae based on rbcL gene sequences. Am J Bot 85: 1507–1516.
- 9. Kershaw P, Wagstaff B (2001) The southern conifer family Araucariaceae: History, status, and value for paleoenvironmental reconstruction. Annu Rev Ecol Syst 32: 397–414.
- 10. Kunzmann L (2007) Araucariaceae (Pinopsida): Aspects in palaeobiogeography and palaeobiodiversity in the Mesozoic. Zool Anz 246: 257–277.
- 11. Escapa IH, Catalano SA (2013) Phylogenetic analysis of Araucariaceae: integrating molecules, morphology, and fossils. Int J Plant Sci 174: 1153–1170.
- 12. Heads M (2010) Biogeographical affinities of the New Caledonian biota: a puzzle with 24 pieces. J Biogeogr 37: 1179–1201.
- 13. Bromham L, Penny D (2003) The modern molecular clock. Nat Rev Genet 4: 216–224.
- 14. Ladiges PY, Cantrill D (2007) New Caledonia-Australian connections: biogeographic patterns and geology. Aust Syst Bot 20: 383–389.
- 15. Biffin E, Hill RS, Lowe AJ (2010) Did Kauri (Agathis: Araucariaceae) really survive the Oligocene drowning of New Zealand? Syst Biol 59: 594–602.
- 16. Crisp MD, Trewick SA, Cook LG (2011) Hypothesis testing in biogeography. Trends Ecol Evol 26: 66–72.
- 17. Knapp M, Mudaliar R, Havell D, Wagstaff SJ, Lockhart PJ (2007) The drowning of New Zealand and the problem of Agathis. Syst Biol 56: 862–870.
- 18. Leslie AB, Beaulieu JM, Rai HS, Crane PR, Donoghue MJ, et al. (2012) Hemisphere-scale differences in conifer evolutionary dynamics. Proc Natl Acad Sci USA 109: 16217–16221.
- 19. Liu N, Zhu Y, Wei ZX, Chen J, Wang QB, et al. (2009) Phylogenetic relationships and divergence times of the family Araucariaceae based on the DNA sequences of eight genes. Chin Sci Bull 54: 2648–2655.
- 20. Hollingsworth ML, Clark AA, Forrest LL, Richardson J, Pennington RT, et al. (2009) Selecting barcoding loci for plants: evaluation of seven candidate loci with species-level sampling in three divergent groups of land plants. Mol Ecol Res 9: 439–457.
- 21. Grivet D, Heinze B, Vendramin GG, Petit RJ (2001) Genome walking with consensus primers: application to the large single copy region of chloroplast DNA. Mol Ecol Notes 1: 345–349.
- 22. Shaw J, Lickey EB, Beck JT, Farmer SB, Liu W, et al. (2005) The tortoise and the hare II: relative utility of 21 noncoding chloroplast DNA sequences for phylogenetic analysis. Am J Bot 92: 142–166.
- 23. Demesure B, Sodzi N, Petit RJ (1995) A set of universal primers for amplification of polymorphic noncoding regions of mitochondrial and chloroplast DNA in plants. Mol Ecol 4: 129–131.
- 24. Hamilton MB (1999) Four primer pairs for the amplification of chloroplast intergenic regions with intraspecific variation. Mol Ecol 8: 521–523.
- 25. Chen S, Yao H, Han J, Liu C, Song J, et al. (2010) Validation of the ITS2 Region as a novel DNA barcode for identifying medicinal plant species. PloS ONE 5: e8613.
- 26. Zwickl D (2006) Genetic algorithm approaches for the phylogenetic analysis of large biological sequence datasets under the maximum likelihood criterion: PhD dissertation, The University of Texas, Austin.
- 27. Huelsenbeck JP, Ronquist F (2001) MrBAYES: Bayesian inference of phylogenetic trees. Bioinformatics 17: 754–755.
- 28. Ronquist F, Huelsenbeck JP (2003) MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19: 1572–1574.
- 29. Rambaut A, Drummond A (2007) Tracer Version 1.5. http://tree.bio.ed.ac.uk/software/tracer/
- 30. Hill RS, Lewis T, Carpenter RJ, Whang SS (2008) Agathis (Araucariaceae) macrofossils from Cainozoic sediments in south-eastern Australia. Aust Syst Bot 21: 162–177.
- 31. Axsmith BJ, Escapa IH, Huber P (2008) An araucarian conifer bract-scale complex from the lower Jurassic of Massachusetts: implications for estimating phylogenetic and stratigraphic congruence in the Araucariaceae. Palaeontologia Electronica 11: 13A.
- 32. Stockey RA (1982) The Araucariaceae - an evolutionary perspective. Rev Palaeobot Palynol 37: 133–154.
- 33. Hill RS, Brodribb TJ (1999) Turner Review No. 2 - Southern conifers in time and space. Aust J Bot 47: 639–696.
- 34. Pole M, Vajda V (2009) A new terrestrial Cretaceous-Paleogene site in New Zealand-turnover in macroflora confirmed by palynology. Cretac Res 30: 917–938.
- 35. Hill RS, Carpenter RJ (1991) Evolution of Acmopyle and Dacrycarpus Podocarpaceae foliage as inferred from macrofossils in south-eastern Australia. Aust Syst Bot 4: 449–480.
- 36. Sinclair WT, Mill RR, Gardner MF, Woltz P, Jaffré T, et al. (2002) Evolutionary relationships of the New Caledonian heterotrophic conifer, Parasitaxus usta (Podocarpaceae), inferred from chloroplast trnL-F intron/spacer and nuclear rDNA ITS2 sequences. Plant Syst Evol 233: 79–104.
- 37. Escapa I, Cuneo R, Axsmith B (2008) A new genus of the Cupressaceae (sensu lato) from the Jurassic of Patagonia: Implications for conifer megasporangiate cone homologies. Rev Palaeobot Palynol 151: 110–122.
- 38. Dornburg A, Beaulieu JM, Oliver JC, Near TJ (2011) Integrating fossil preservation biases in the selection of calibrations for molecular divergence time estimation. Syst Biol 60: 519–527.
- 39. Marshall CR (2008) A simple method for bracketing absolute divergence times on molecular phylogenies using multiple fossil calibration points. Am Nat 171: 726–742.
- 40. Drummond AJ, Rambaut A (2007) BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evol Biol 7: 214.
- 41. Drummond AJ, Ho SYW, Phillips MJ, Rambaut A (2006) Relaxed phylogenetics and dating with confidence. PLoS Biol 4: 699–710.
- 42. Cantrill DJ, Raine JI (2006) Wairarapaia mildenhallii gen. et sp nov., a new Araucarian cone related to Wollemia from the Cretaceous (Albian-Cenomanian) of New Zealand. Int J Plant Sci 167: 1259–1269.
- 43. McLean D, Owens B, Pendleton JL, Bodman D (2013) Pennsylvanian (Namurian and Westphalian) miospore assemblages from the west coast of Scotland. Rev Palaeobot Palynol 190: 1–14.
- 44. van Waveren IM, Abbink OA, van Hoof TB, van Konijnenburg-van Cittert JHA (2008) Revision of the Late Carboniferous megaflora from the De Lutte-06 well (Twente, the Netherlands), and its stratigraphical implications. Geol Mijnbouw-N J G 87: 339–352.
- 45. Axsmith BJ, Taylor TN, Taylor EL (1998) Anatomically preserved leaves of the conifer Notophytum krauselii (Podocarpaceae) from the Triassic of Antarctica. Am J Bot 85: 704–713.
- 46. Miller CN (1999) Implications of fossil conifers for the phylogenetic relationships of living families. Bot Rev 65: 239–277.
- 47. Stockler K, Daniel IL, Lockhart PJ (2002) New Zealand Kauri (Agathis australis (D. Don) Lindl., Araucariaceae) survives Oligocene drowning. Syst Biol 51: 827–832.
- 48. Ruane S, Pyron RA, Burbrink FT (2011) Phylogenetic relationships of the Cretaceous frog Beelzebufo from Madagascar and the placement of fossil constraints based on temporal and phylogenetic evidence. J Evol Biol 24: 274–285.
- 49. Rutschmann F, Eriksson T, Abu Salim K, Conti E (2007) Assessing calibration uncertainty in molecular dating: the assignment of fossils to alternative calibration points. Syst Biol 56: 591–608.
- 50. Kenrick P, Crane PR (1997) The origin and early evolution of plants on land. Nature 389: 33–39.
- 51. Chambers TC, Drinnan AN, McLoughlin S (1998) Some morphological features of Wollemi pine (Wollemia nobilis: Araucariaceae) and their comparison to Cretaceous plant fossils. Int J Plant Sci 159: 160–171.
- 52. Ohsawa T, Nishida H, Nishida M (1995) Yezonia a new section of Araucaria (Araucariaceae) based on permineralized vegetative and reproductive-organs of A. vulgaris comb. nov. from the Upper Cretaceous of Hokkaido, Japan. J Plant Res 108: 25–39.
- 53. Cantrill DJ (1992) Araucarian foliage from the Lower Cretaceous of southern Victoria, Australia. Int J Plant Sci 153: 622–645.
- 54. Del Fueyo GM, Archangelsky A (2002) Araucaria grandifolia Feruglio from the Lower Cretaceous of Patagonia, Argentina. Cretac Res 23: 265–277.
- 55. Macphail M, Carpenter RJ (2013) New potential nearest living relatives for Araucariaceae producing fossil Wollemi Pine-type pollen (Dilwynites granulatus W.K. Harris, 1965). Alcheringa 38: 135–139.
- 56. Lee DE, Bannister JM, Lindqvist JK (2007) Late Oligocene-Early Miocene leaf macrofossils confirm a long history of Agathis in New Zealand. N Z J Bot 45: 565–578.
- 57. Pole M (2008) The record of Araucariaceae macrofossils in New Zealand. Alcheringa 32: 405–426.
- 58. Dettmann ME, Clifford HT, Peters M (2012) Emwadea microcarpa gen. et sp nov.-anatomically preserved araucarian seed cones from the Winton Formation (late Albian), western Queensland, Australia. Alcheringa 36: 217–237.
- 59. van der Ham RWJM, Jagt JWM, Renkens S, van Konijnenburg-van Cittert JHA (2010) Seed-cone scales from the upper Maastrichtian document the last occurrence in Europe of the Southern Hemisphere conifer family Araucariaceae. Palaeogeogr Palaeoclimatol Palaeoecol 291: 469–473.
- 60. McLoughlin S, Carpenter RJ, Jordan GJ, Hill RS (2008) Seed ferns survived the end-Cretaceous mass extinction in Tasmania. Am J Bot 95: 465–471.
- 61. Sanmartin I, Ronquist F (2004) Southern Hemisphere biogeography inferred by event-based models: Plant versus animal patterns. Syst Biol 53: 216–243.
- 62. Lawver LA, Gahagan LM (2003) Evolution of Cenozoic seaways in the circum-Antarctic region. Palaeogeogr Palaeoclimatol Palaeoecol 198: 11–37.
- 63. Veevers JJ, Powell CM, Roots SR (1991) Review of seafloor spreading around Australia. I. Synthesis of the patterns of spreading. Aust J Earth Sci 38: 373–389.
- 64. Francis J, Ashworth A, Cantrill D, Crame J, Howe J, et al.. (2008) 100 million years of Antarctic climate evolution: Evidence from fossil plants. In: Cooper A, Barrett P, Stagg H, Storey B, Stump E, Wise W, editors. Antarctica: A Keystone in a Changing World. Proceedings of the 10th International Symposium on Antarctic Earth Sciences. Washington, DC: The National Academies Press. pp. 19–27.
- 65. Hall R (2001) Cenozoic reconstructions of SE Asia and the SW Pacific: changing patterns of land and sea. In: Metcalfe J, Smith M, Morwood I, Davidson A, editors. Faunal and Floral Migrations and Evolution in SE Asia-Australasia. Lisse: Swets and Zeitlinger. pp. 35–56.
- 66. Heads M (2008) Panbiogeography of New Caledonia, south-west Pacific: basal angiosperms on basement terranes, ultramafic endemics inherited from volcanic island arcs and old taxa endemic to young islands. J Biogeogr 35: 2153–2175.
- 67. Meffre S, Crawford A, Quilty P (2006) Arc-continent collision forming a large island between New Caledonia and New Zealand in the Oligocene. ASEG Extended Abstracts, Australian Earth Sciences Convention 2006 Melbourne 2006: 1–3.