Research Article

Relaxed Molecular Clock Provides Evidence for Long-Distance Dispersal of Nothofagus (Southern Beech)

  • Michael Knapp,

    Affiliation: Allan Wilson Centre for Molecular Ecology and Evolution, Institute of Molecular BioSciences, Massey University, Palmerston North, New Zealand

  • Karen Stöckler,

    Affiliation: Allan Wilson Centre for Molecular Ecology and Evolution, Institute of Molecular BioSciences, Massey University, Palmerston North, New Zealand

  • David Havell,

    Affiliations: Allan Wilson Centre for Molecular Ecology and Evolution, Institute of Molecular BioSciences, Massey University, Palmerston North, New Zealand, Universal College of Learning, Palmerston North, New Zealand

  • Frédéric Delsuc,

    Affiliations: Allan Wilson Centre for Molecular Ecology and Evolution, Institute of Molecular BioSciences, Massey University, Palmerston North, New Zealand, Institut des Sciences de l'Evolution, Université Montpellier II, France

  • Federico Sebastiani,

    Affiliation: Università degli Studi di Firenze, Dipartimento di Biotecnologie Agrarie Laboratory, Genexpress, Polo Scientifico, Sesto Fiorentino, Italy

  • Peter J Lockhart mail

    To whom correspondence should be addressed. E-mail:

    Affiliation: Allan Wilson Centre for Molecular Ecology and Evolution, Institute of Molecular BioSciences, Massey University, Palmerston North, New Zealand

  • Published: January 04, 2005
  • DOI: 10.1371/journal.pbio.0030014


Nothofagus (southern beech), with an 80-million-year-old fossil record, has become iconic as a plant genus whose ancient Gondwanan relationships reach back into the Cretaceous era. Closely associated with Wegener's theory of “Kontinentaldrift”, Nothofagus has been regarded as the “key genus in plant biogeography”. This paradigm has the New Zealand species as passengers on a Moa's Ark that rafted away from other landmasses following the breakup of Gondwana. An alternative explanation for the current transoceanic distribution of species seems almost inconceivable given that Nothofagus seeds are generally thought to be poorly suited for dispersal across large distances or oceans. Here we test the Moa's Ark hypothesis using relaxed molecular clock methods in the analysis of a 7.2-kb fragment of the chloroplast genome. Our analyses provide the first unequivocal molecular clock evidence that, whilst some Nothofagus transoceanic distributions are consistent with vicariance, trans-Tasman Sea distributions can only be explained by long-distance dispersal. Thus, our analyses support the interpretation of an absence of Lophozonia and Fuscospora pollen types in the New Zealand Cretaceous fossil record as evidence for Tertiary dispersals of Nothofagus to New Zealand. Our findings contradict those from recent cladistic analyses of biogeographic data that have concluded transoceanic Nothofagus distributions can only be explained by vicariance events and subsequent extinction. They indicate that the biogeographic history of Nothofagus is more complex than envisaged under opposing polarised views expressed in the ongoing controversy over the relevance of dispersal and vicariance for explaining plant biodiversity. They provide motivation and justification for developing more complex hypotheses that seek to explain the origins of Southern Hemisphere biota.


An important principle of evolutionary inference is that explanations for the past require an understanding of mechanisms and processes applicable in the present [1]. It is perhaps this sticking point more than any other that has polarised views over the relative importance of vicariance and dispersal for explaining extant plant biodiversity. In 1915, Alfred Wegener put forward a testable hypothesis and mechanism that could explain the transoceanic distribution of animal and plant species. In the 21st century, with many DNA studies now implicating the importance of long-distance dispersal for explaining plant biodiversity [2,3,4,5], it is disconcerting that there is currently a very poor understanding of the mechanisms of transoceanic dispersal (but see [6,7,8,9,10]). Indeed, the inference that the seeds of extant Nothofagus species are not suited for dispersal across large distances has played a major role in motivating the hypothesis that transoceanic distributions of Nothofagus (Figure 1) can only be explained by vicariance [11,12,13,14,15]. This hypothesis posits that following the Cretaceous breakup of Gondwana, Nothofagus rafted and evolved in situ upon different Southern Hemisphere lands. Whilst very attractive, this hypothesis fits somewhat uncomfortably with the findings from analyses of morphological and molecular data. In particular, whilst earlier molecular data have been insufficient for rigorous molecular clock analyses, their interpretation has favoured hypotheses of transoceanic dispersal [16,17,18].


Figure 1. Southern Hemisphere Maps and Present-Day Nothofagus Distribution

(A) Transoceanic distribution of Nothofagus subspecies Lophozonia and Fuscospora and South American species N. nitida (subgenus Nothofagus). Map adapted from Swenson et al. [43]. ASE, Australia; NCA, New Caledonia; NGU, New Guinea; NZE, New Zealand; SAM, South America; TAS, Tasmania.

(B) Relationship of Australia, New Zealand, and South America 65 Myr and 35 Myr before present, reconstructed from (link “Plate Tectonic Reconstructions”).


Based on the sequence of Gondwana breakup, a hypothesis of vicariance most parsimoniously predicts that Australian Nothofagus species should be most closely related to South American species. This follows since South America and Australia were connected via Antarctica until approximately 35 million years (Myr) ago (Figure 1). In contrast, New Zealand is thought to have separated from Australia 80 Myr ago [19,20]. Thus to explain the close relationship between Australian and New Zealand species by vicariance, it is necessary to argue that extinction of Australian and/or closely related South American species has occurred [12]. Whilst this explanation is ad hoc, the fossil record does provide evidence for numerous Nothofagus extinctions in Australia, South America, and New Zealand [21,22,23].

However, the fossil record has also been interpreted as indicating multiple events of transoceanic dispersal of Nothofagus from Australia to New Zealand. Whilst the extinct “ancestral” Nothofagus pollen type occurred in New Zealand prior to the breakup of Gondwana, Fuscospora pollen first appeared in New Zealand during the Palaeocene (65 Myr ago) and Lophozonia pollen first appeared during the late Eocene (50 Myr ago; [24]). Sixty-five Myr ago the Tasman Sea had already reached its present-day size [19,20]. Hence it is possible that extant New Zealand Nothofagus subgenera did not have the opportunity to reach New Zealand via overland migration. Hill [25] has also described the species Nothofagus cethanica, which first appeared in Oligocene macrofossils from Tasmania. This species shares unique features with extant N. fusca and N. truncata from New Zealand and may share a sister relationship with these species explained by trans-Tasman Sea dispersal [26].

A contribution to the debate over the relative importance of vicariance and dispersal can be made by estimating the divergence times of extant species. However, DNA sequences of insufficient length have prevented robust molecular clock analyses from being undertaken. For this reason, we report the sequencing of a 7.2-kb chloroplast genome fragment covering the gene regions (trnL–trnF and atpB–psaI; see Table 1 for accession numbers) for 11 species of three Nothofagus subgenera (Lophozonia, Fuscospora, and Nothofagus). Our aim has been to date divergence of extant species in the subgenera Lophozonia and Fuscospora. We have carried out relaxed molecular clock analyses using the methods of Sanderson [27,28] and Thorne et al. [29]. Our findings are that, whilst vicariance is likely to explain some transoceanic relationships amongst Nothofagus species, phylogenetic relationships between trans-Tasman species in both Lophozonia and Fuscospora can only be explained by mid- to late-Tertiary transoceanic dispersal.


Table 1. Origin of Nothofagus Samples and Sequence Accession Numbers



Figure 2 shows an optimal maximum-likelihood reconstruction of phylogenetic relationships for chloroplast DNA sequences (7.2-kb comprising the atpB–psaI region and the trnL–trnF region; 7,269 nucleotide sites) for Nothofagus (subgenera or pollen groups (a) Lophozonia, (b) Fuscospora, and (c) Nothofagus) and outgroup Castanea sativa (not shown). In a sensitivity analysis of 60 substitution models, the tree shown in Figure 2 was always recovered with very little difference in branch lengths regardless of the substitution model used. Of all substitution models evaluated, K81uf+G was identified as the best fitting one for the data based on hierarchical likelihood ratio tests and the Akaike Information Criterion. This substitution model and also the F84+ Γ8 model were used for further analyses. The latter was included because the Bayesian relaxed molecular clock (BRMC) approach as implemented in the program MULTIDIVTIME (see Materials and Methods) only allows the use of the JC and the F84 models. Thus analysis with the F84+ Γ8 model allowed a comparison of date estimates to be obtained using different relaxed molecular clock methods. All nodes of the optimal ML tree recovered in the sensitivity analysis received nonparametric bootstrap support greater than 97%, with the only exception being the grouping of N. cunninghamii with N. moorei, which received 72% support.


Figure 2. ML Tree Indicating Evolutionary Relationships for Nothofagus Species Based on the atpB–psaI and trnL–trnF Region of the Chloroplast Genome (7,269 bp)

Divergence dates (in Myr) were obtained with an F84+ Γ8 substitution model using the BRMC approach of Thorne et al. [29]. For the dates indicated, the age of the root node and that of F/N1 were constrained to 70–80 Myr; L2 was also constrained in accordance with fossil data [26] at 20 Myr. Violet numbers show bootstrap values. The pollen grains represent the first appearance of the respective pollen type in the New Zealand fossil record. Plio, Pliocene; Oligo, Oligocene; Palaeo, Palaeocene; Ma, Maastrichian; Campan, Campanian. L1–L4, Lophozonia 1–4; F1–F2, Fuscospora 1–2; F/N1, Fuscospora/Nothofagus 1.


Divergence times for the nodes in this tree (Figure 2) were estimated using the penalized likelihood (PL) method [27] and BRMC method [29,30,31]. For these analyses, a period of 70–80 Myr was used to calibrate the divergence between the three fossil pollen groups representing subgenera Lophozonia, Nothofagus, and Fuscospora. These four pollen groups all first appeared in the fossil record approximately 75 Myr ago [32]. A second constraint of a minimum of 20 Myr for the divergence of N. cunninghamii and N. moorei was also used. This constraint was based on observations reported by Hill [26] that 20-Myr-old fossils intermediate between N. moorei and N. cunninghamii were recorded from Tasmania and that fossils closely resembling N. moorei were also present at that time. The inferred ages for the remaining nodes of the tree, obtained under the F84+ Γ8 model of substitution are given in Table 2 and graphically illustrated on Figure 2. The variance on these estimates was low and the values were little influenced by the choice of substitution model (Table 3). The robustness of the estimates to calibration error was tested by constraining the divergence of Australian and New Zealand sister taxa to 65 Myr (the time before present when the Tasman Sea reached its present position; thus this date provided us with a lower bound for divergence times of trans-Tasman Nothofagus disjunctions that might be explained by vicariance). Constraining these two nodes in this way produced unrealistic age estimates for all basal nodes. For example, using the BRMC method, which additionally required a prior expectation to be specified for the age of the root node (which we specified at 75 Myr—the time of appearance of all four extant pollen types), we estimated a more likely age for the root node at 191 Myr. For the PL approach, which does not require specification of a prior, we estimated the age of the root node at 634 Myr. Other basal nodes in both the Fuscospora and Lophozonia lineages were also much older than reasonably expected (see Table 2).


Table 2. Estimated Divergence Dates and Standard Deviations (in Brackets) of Different Nothofagus Clades


Table 3. Variation of Estimated Divergence Times (in Myr) under 60 Symmetrical Models of DNA Substitution



Our findings from molecular clock analyses using five independent calibrations (for four nodes), suggest that the sister relationships of the Australasian (Australia and New Zealand) species within both Lophozonia and Fuscospora lineages are too young to be explained by continental drift (as indicated by the inferred ages of nodes F1 and L3). Transoceanic dispersal appears the most likely explanation for the trans-Tasman sister relationships indicated in Figures 1 and 2. In contrast, the age inferred for node F2, using both relaxed clock methods is compatible with a hypothesis of continental drift as an explanation for the sister relationship between South American and Australasian Fuscospora lineages. The age for node L4, which separates Australasian and South American Lophozonia, may also be consistent with vicariance. The BRMC method dates it at 34 Myr before present. However, the PL method estimates this node to be only 25 Myr old, an age too recent to be consistent with vicariance. Thus we regard our results for node L4 as equivocal. Nevertheless, southern beeches are likely to have been present in Antarctica 25 myr ago [33], and thus long-distance dispersal across the young southern ocean between South America and Australia via Antarctica may be conceivable.

The robustness of our phylogenetic inferences has been investigated by varying the substitution model (60 symmetric models were used), estimating the variance of age estimates, and evaluating the influence of calibrations on divergence times. With the exception of the root node, the PL method consistently gave more recent age estimates than did the BRMC method. Both methods showed sensitivity to the number of calibration points used, a finding consistent with recent observations on the performance of relaxed molecular clock methods [34]. In general, the date estimates produced by the BRMC approach were more consistent with the fossil record [26]. A relevant question is whether or not additional calibration points could make date estimates older and thus change our conclusion of trans-Tasman dispersal. We suggest that this may be unlikely, given the observation that constraining a minimum age for trans-Tasman sister species to 65 Myr leads to greatly inflated and unrealistic age estimates for all basal nodes. Hence to explain this finding we would need to invoke a further hypothesis of a dramatic and independent slowing in the rate of evolution in Lophozonia, Fuscospora, and Nothofagus lineages.

Thus the hypothesis that present-day distribution patterns of Nothofagus can be explained by continental drift following the breakup of Gondwana and subsequent extinction of some species [24] can be rejected on the basis of the divergence dates that we have estimated. These dates also indicate that present-day Nothofagus species in New Zealand are not the direct descendants of the Fuscospora and Lophozonia southern beeches that reached New Zealand in the Palaeocene and Eocene eras, respectively [24]. This finding highlights the caution that needs to be taken when interpreting fossil evidence for the apparent first appearance of extant taxa. Fossils that identify specific evolutionary lineages may not necessarily indicate the origins for extant taxa or suggest a continuous presence for these taxa. Similar concerns follow from the findings of molecular analyses for Ascarina and Laurelia in New Zealand [2,4].

The strength of our molecular analyses highlights the importance of future research into potential mechanisms of long-distance dispersal, and in particular reinvestigation of the transoceanic dispersal properties of Nothofagus seeds. For the reasons that we outline in our introduction, it seems likely that only once the mechanisms of long-distance dispersal are understood will hypotheses based on DNA divergence time estimates be truly convincing. DNA sequence analyses have also suggested that long-distance dispersal and continental drift are both important for explaining distributions of the conifer Agathis (Araucariaceae) in the South Pacific [35]. Although the molecular evidence for Agathis is not as strong as it is for Nothofagus, the findings from the molecular studies on these genera highlight the importance of considering more complex hypotheses of relationship in debates concerning the relative importance of dispersal and vicariance.

Materials and Methods

Sequence data.

Chloroplast DNA sequences (7.2 kb comprising the atpB–psaI region and the trnL–trnF region) were determined for each of 11 accessions of Nothofagus (subgenera or pollen groups Lophozonia, Fuscospora, and Nothofagus) sampled in South America, Australia, and New Zealand (see Table 1). These genome regions were also determined for C. sativa (an outgroup taxon from Fagaceae) and aligned using progressive multiple-sequence alignment: ClustalX version 1.81 [36]. This resulted in an unambiguous alignment of 7,269 nucleotide sites. Data are missing for approximately 250 bp of the atpB gene and atpBrbcL intergene region of Nothofagus.

Tree building.

Phylogenetic analyses were conducted with PAUP* version 4.0b10 [37] under the ML criterion. A model sensitivity test was conducted, investigating a range of 60 symmetrical models of DNA substitution corresponding to the 56 implemented in MODELTEST version 3.06 [38] (​t.html) plus F84, F84+I, F84+Γ8, and F84+I+Γ8. ML parameters of these models were estimated by PAUP* following the approach used in MODELTEST. These parameters were then used to conduct 60 individual ML heuristic searches in PAUP* with tree bisection-reconnection branch swapping and a neighbour-joining starting tree. ML bootstrap proportions were obtained after 100 replications, using the same search strategy and ML parameters as for the analysis of the original dataset.

Molecular dating: The PL method.

Divergence dates were obtained using the PL method of Sanderson [27] as implemented in the program r8s, version 1.60 [28] ( with the TN algorithm. The outgroup was excluded using the “prune” command. The degree of autocorrelation within lineages was estimated using cross-validation as suggested by Sanderson [27], and the correcting smoothing parameter λ defined accordingly. Divergence dates were estimated on the 60 ML phylograms recovered in the phylogenetic model sensitivity analysis. Ages for each node across the 60-ML trees were summarized using the “profile” command. Confidence limits on dating estimates were computed by using nonparametric bootstrapping of the original dataset as suggested by Sanderson and Doyle [39]. The program SEQBOOT of the PHYLIP 3.6 package [40] was used to generate 100 bootstrap resampled datasets of 7,269 sites in length. ML branch lengths of the optimal topology were then estimated under the F84+ Γ8 model for each of the bootstrap resampled datasets using PAUP*. Divergence estimates were then calculated for each of the 100 bootstrap replicates using r8s to obtain standard deviations on each node by the “profile” command and the settings described above.

Molecular dating: The BRMC method.

The BRMC approach was applied using the program MULTIDIVTIME as implemented in the Thornian Time Traveller (T3) package [41]. First, the program BASEML of the PAML package version 3.13 [42] (​ml.html) was used to estimate the ML parameters of the F84+ Γ8 substitution model, using the ML topology previously identified. Second, the program ESTBNEW ( was used to estimate branch lengths of the ML topology and the corresponding variance–covariance matrix. Finally, the program MULTIDIVTIME was used to run a Markov chain Monte Carlo for estimating mean posterior divergence times on nodes with associated standard deviations from the variance–covariance matrix produced by ESTBNEW. The Markov chain was sampled 10,000 times every 100 cycles after a burn-in stage of 100,000 cycles. We used a 75-Myr (SD = 37.5 Myr) prior [32] for the expected number of time units between tip and root and a prior of 200 Myr for the highest possible number of time units between tip and root. Other priors for gamma distribution of the rate at root node and the Brownian motion constant describing the rate variation (i.e., the degree of rate autocorrelation along the descending branches of the tree) were derived from the median branch length. As for the PL method, the outgroup was not included in this analysis.

Supporting Information

Accession Numbers

The GenBank ( accession numbers for the sequences discussed in this paper are given in Table 1.


We thank Bruce Christie, Peter Seemann, Andrew Rozefelds, and Martin Roberts for help with obtaining tissue samples, and also Trish McLenachan, Tim White, and Michael Heads for valuable discussions. Leon Perrie, Heidi Meudt, Robert Hill, Matt McGlone, Peter Raven and an anonymous reviewer provided helpful and critical comments on our manuscript that were much appreciated. We would like to thank the New Zealand Marsden Fund and the Alexander von Humboldt Foundation for funding our study. MK is supported by doctoral scholarships from Massey University and the German Academic Exchange Service.

Author Contributions

MK, KS, DH, FD, and PJL conceived and designed the experiments. MK and KS performed the experiments. MK, KS, FD, and PJL analyzed the data. DH and FS contributed reagents/materials/analysis tools. MK, DH, and PJL wrote the paper.


  1. 1. Lyell C (1970) Principles of geology: Being an attempt to explain the former changes of the earth's surface by reference to causes now in operation. 3 v. New York: S-H Service Agency.
  2. 2. Renner SS, Foreman DB, Murray D (2000) Timing transantarctic disjunctions in the Atherospermataceae (Laurales): Evidence from coding and noncoding chloroplast sequences. Syst Biol 49: 579–591.
  3. 3. Winkworth RC, Wagstaff SJ, Glenny D, Lockhart PJ (2002) Plant dispersal N.E.W.S from New Zealand. Trends Ecol Evol 17: 514–520.
  4. 4. Zhang L-B, Renner S (2003) The deepest splits in Chloranthaceae as resolved by chloroplast sequences. Int J Plant Sci 164: 383–392.
  5. 5. Givnish TJ, Millam KC, Evans TM, Hall JC, Pires JC, et al. (2004) Ancient vicariance or recent long-distance dispersal? Inferences about phylogeny and South American–African disjunctions in Rapateaceae and Bromeliaceae based on ndhF sequence data. Int J Plant Sci 165: S35–S54.
  6. 6. Close RC, Moar NT, Tomlinson AI, Lowe AD (1978) Aerial dispersal of biological material from Australia to New Zealand. Int J Biometeorol 22: 1–19.
  7. 7. Carlquist SJ (1996) Plant dispersal and the origin of Pacific island floras. In: Keast A, Miller SE, editors. The origin and evolution of Pacific island biotas, New Guinea to Eastern Polynesia: Patterns and processes. Amsterdam: Academic. pp. 153–164.
  8. 8. Hurr K, Lockhart PJ, Heenan PB, Penny D (1999) Dispersal of the Edwardsia section of Sophora (Leguminosae) around the southern oceans: Molecular evidence. J Biogeogr 26: 565–577.
  9. 9. Muñoz J, Felicísimo ÁM, Cabezas F, Burgaz AR, Martínez I (2004) Wind as a long-distance dispersal vehicle in the Southern Hemisphere. Science 304: 1144–1147.
  10. 10. Renner S (2004) Plant dispersal across the tropical Atlantic by wind and sea currents. Int J Plant Sci 165: S23–S33.
  11. 11. Preest DS (1963) A note on the dispersal characteristics of the seed of the New Zealand podocarps and beeches, and their biogeographical significance. In: Gressit JL, editor. Pacific Basin biogeography, a symposium. Honolulu: Bishop Museum Press. pp. 415–424.
  12. 12. Linder HP, Crisp MD (1995) Nothofagus and Pacific biogeography. Cladistics 11: 5–32.
  13. 13. Craw RC, Gehan JR, Heads MJ (1999) Panbiogeography: Tracking the history of life. Oxford: Oxford University Press.
  14. 14. McCarthy D (2003) The trans-Pacific zipper effect: Disjunct sister taxa and matching geological outlines that link the Pacific margins. J Biogeogr 30: 1545–1561.
  15. 15. Sanmartín I, Ronquist F (2004) Southern Hemisphere biogeography inferred by event-based models: Plant versus animal patterns. Syst Biol 53: 216–243.
  16. 16. Martin PG, Dowd JM (1993) Using sequences of rbcL to study phylogeny and biogeography of Nothofagus species. Aust J Bot 6: 441–447.
  17. 17. Hill RS, Jordan GJ (1993) The evolutionary history of Nothofagus (Nothofagaceae). Aust J Bot 6: 111–126.
  18. 18. Manos P (1997) Systematics of Nothofagus (Nothofagaceae) based on rDNA spacer sequences (ITS): Taxonomic congruence with morphology and plastid sequences. Am J Bot 84: 1137–1150.
  19. 19. Wilford GE, Brown PJ (1994) Maps of late Mesozoic–Cenozoic Gondwana breakup: Some palaeogeographical implications. In: Hill RS, editor. History of the Australian vegetation: Cretaceous to recent. Cambridge: Cambridge University Press. pp. 5–13.
  20. 20. McLoughlin S (2001) The breakup history of Gondwana and its impact on pre Cenozoic floristic provincialism. Aust J Bot 49: 271–300.
  21. 21. Hill RS, Jordan GJ, Macphail MK (1996) History and paleoecology of Australian Nothofagus forests. In: Veblen TT, Hill RS, Read J, editors. The ecology and biogeography of Nothofagus forests. New Haven (Connecticut): Yale University Press. pp. 182–199.
  22. 22. Markgraf V, Romero E, Villagrán C (1996) History and palaeoecology of South American Nothofagus forests. In: Veblen TT, Hill RS, Read J, editors. The ecology and biogeography of Nothofagus forests. New Haven (Connecticut): Yale University Press. pp. 354–386.
  23. 23. McGlone MS, Mildenhall DC, Pole MS (1996) History and palaeoecology of New Zealand Nothofagus forests. In: Veblen TT, Hill RS, Read J, editors. The ecology and biogeography of Nothofagus forests. New Haven (Connecticut): Yale University Press. pp. 83–130.
  24. 24. Hill RS (2001) Biogeography, evolution and palaeoecology of Nothofagus (Nothofagaceae): The contribution of the fossil record. Aust J Bot 49: 321–332.
  25. 25. Hill RS (1984) Tertiary Nothofagus macrofossils from Cethana, Tasmania. Alcheringia 8: 81–86.
  26. 26. Hill RS (1991) Tertiary Nothofagus (Fagaceae) macrofossils from Tasmania and Antarctica and their bearing on the evolution of the genus. Bot J Linn Soc 105: 73–112.
  27. 27. Sanderson MJ (2002) Estimating absolute rates of molecular evolution and divergence times: A penalized likelihood approach. Mol Biol Evol 19: 101–109.
  28. 28. Sanderson MJ (2003) r8s: Inferring absolute rates of molecular evolution and divergence dates in the absence of a molecular clock. Bioinformatics 19: 301–302.
  29. 29. Thorne JL, Kishino H, Painter IS (1998) Estimating the rate of evolution of the rate of molecular evolution. Mol Biol Evol 15: 1647–1657.
  30. 30. Kishino H, Thorne JL, Bruno WJ (2001) Performance of a divergence time estimation method under a probabilistic model of rate evolution. Mol Biol Evol 18: 352–361.
  31. 31. Thorne JL, Kishino H (2002) Divergence time and evolutionary rate estimation with multilocus data. Syst Biol 51: 689–702.
  32. 32. Dettmann ME, Pocknall D, Romero EJ, Zamboloa M (1990) Nothofagidites Erdman ex Potonie 1960: A catalogue of species with notes on palaeogeographic distribution of Nothofagus Bl. (southern beech). NZ Geol Surv Palaeontol Bull 60: 1–79.
  33. 33. Hill RS, Truswell EM (1993) Nothofagus fossils in the Sirius Group, Transantarctic Mountains: Leaves and pollen and their climatic implications. In: Kennett JP, Warnke DA, editors. Antarctic research series, Volume 60. The Antarctic palaeoenvironment: A perspective on global change. Washington, DC: American Geophysical Union. pp. 67–73.
  34. 34. Pérez-Losada M, Høeg JT, Crandall KA (2004) Unraveling the evolutionary radiation of the thoracian barnacles using molecular and morphological evidence: A comparison of several divergence time estimation approaches. Syst Biol 53: 244–264.
  35. 35. Stöckler K, Daniel IL, Lockhart PJ (2002) New Zealand Kauri (Agathis australis (D.Don) Lindl., Araucariaceae) survives Oligocene drowning. Syst Biol 51: 827–832.
  36. 36. Thompson JD, Gibson TJ, Plewniak F, Jeanmougin F, Higgins DG (1997) The ClustalX windows interface: Flexible strategies for multiple sequence alignment aided by quality analysis tools. Nucleic Acids Res 24: 4876–4882 (Program available: Accessed 16 November 2004.).
  37. 37. Swofford DL (2003) PAUP*. Phylogenetic analysis using parsimony (*and other methods), version 4 [computer program]. Sunderland (Massachusetts): Sinauer.
  38. 38. Posada D, Crandall K (1998) MODELTEST: Testing the model of DNA substitution. Bioinformatics 14: 817–818.
  39. 39. Sanderson MJ, Doyle JA (2001) Sources of error and confidence intervals in estimating the age of angiosperms from rbcL and 18S rDNA data. Am J Bot 88: 1499–1516.
  40. 40. Felsenstein J (2002) PHYLIP (PHYLogeny inference package), version 3.6 [computer program]. Available:​/phylip.html. Accessed 18 November 2004.
  41. 41. Yang Z (2003) T3 “Thornian time traveller” [computer program]. Available: Accessed 16 November 2004.
  42. 42. Yang Z (1997) PAML: A program package for phylogenetic analysis by maximum likelihood. Comput Appl Biosci 13: 555–556.
  43. 43. Swenson U, Backlund A, McLoughlin S, Hill RS (2001) Nothofagus biogeography revisited with special emphasis on the enigmatic distribution of subgenus Brassospora in New Caledonia. Cladistics 17: 28–47.