Diversification times and biogeographic patterns in Apiales.
The dicot order Apiales exhibits a worldwide geographic distribution, in which its two most speciose families have a center of diversity either in the north temperate zone (Apiaceae) or in the tropics (Araliaceae). The remaining families of the order are more geographically restricted (e.g., Pittosporaceae, which is largely Australasian, and Myodocarpaceae, which is restricted almost exclusively to New Caledonia) (Table 1). Distributional data from fossils and extant species provide evidence that the species of Apiales have occupied all major phytogeographic regions, but very little is known about the biogeographic relationships among the major apialean clades. Traditional theories suggest an origin of Araliaceae in the Paleotropics during the Cretaceous or earlier, and the derivation of the more temperate Apiaceae from proto-araliaceous stock due to climatic changes during the Late Cretaceous or early Cenozoic (Mathias, 1965; Rodriguez, 1971). This Paleotropical-origin theory is supported by the high diversity of "ancient" apialean taxa in Old-World tropical regions and parallels that of the angiosperms in general (Axelrod, 1952; Shields, 1991; Plunkett et al., 1996a; Plunkett & Lowry, 2001). Australasia has been of major interest in biogeographic studies due to its complicated geologic history and high levels of endemism, especially for plants and invertebrates. Theories have been proposed to explain both the distribution of taxa in this region and the links of Australasian taxa to those of other regions of the world, including both vicariance, resulting from the break-up of Gondwana (e.g., Brundin, 1966; Raven & Axelrod, 1972; Nelson, 1975; Linder & Crisp, 1996), and dispersal of taxa across narrow ocean basins or by means of "island stepping stones" (e.g., Carlquist, 1974, 1981; Diamond, 1984; Takhtajan, 1986; Pole, 1994). Explanations favoring dispersal often invoke the submergence of many landmasses (especially islands) during the Oligocene, necessitating post-Gondwanan colonization and diversification (Pole, 1994, 2001; McPhail, 1997). This theory suggests that the affinity between taxa of post-Gandwanan origin on isolated areas may be best explained by long-distance dispersal. However, many studies of endemic lineages from different Australasian subregions have dated these lineages to ancient Gondwanan ancestors, which appear to have been separated subsequently by vicariance events (e.g., McLoughlin, 2001; Swenson et al., 2001; Stockier et al., 2002; Ladiges et al., 2003). Most of these studies attribute the survival of such lineages to the presence of refugia, in which some Gondwanan species persisted during periods of submergence and climatic fluctuations. Based on the current distributions of its major lineages, we attempt to determine which clades were influenced by vicariance and which were influenced by long-distance-dispersal events. To achieve this, different biogeographic-reconstruction and estimation tools were applied to gene phylogenies, geologic histories, current distributions, and patterns of endemism.
The earliest known angiosperm fossils date from Valanginian-Hauterivian deposits, demonstrating the presence of angiosperms in the Early Cretaceous (141-132 Ma) (Brenner, 1996; Doyle & Endress, 2010). The angiosperm fossil record suggests that these plants underwent a rapid diversification from the Barremian of the Early Cretaceous (c. 115 Ma) through the Late Cretaceous (c. 90 Ma) (Friis et al., 1999; Herendeen et al., 1999; Magallon-Puebla et al., 1999; Bell et al., 2005). Differences in methodology or statistical approach have resulted in several different estimates of the origin of angiosperms, some of which are quite disparate (Sanderson & Doyle, 2001). Molecular clocks calibrated with fossil data estimated the origin of the angiosperms to be as early as the Triassic or Jurassic (> 200 Ma) (Sanderson, 1997; Chaw et al., 2004; Bell et al., 2005). The core eudicot lineage was estimated to have diverged from other angiosperms c. 100-147 Ma (Bremer, 2000; Wikstrom et al., 2001; Chaw et al., 2004; Bell et al., 2005). For Apiales, early work by Bessey (1897) reported that fossil "Umbellales" (reflecting on out-dated circumscription of the order that included not only Umbelliferae and Araliaceae, but also Comaceae) stretched back to the Cretaceous. Bessey also estimated that the number of extant species relative to the total number of angiosperms has decreased since the Late Cretaceous and Eocene. By contrast, the more recent study of Magallon and Sanderson (2001) estimated an increase in the diversification rate for Apiales since its divergence, but this rate may be attributed to recent radiations in the more speciose Apiaceae (~75 % of the 4,898 species counted in the study) compared to the other families of the order, which appear to be older but have fewer extant species (e.g., Myodocarpaceae and Pittosporaceae). Also, the divergence time used for Apiales (45.15 Ma) represents an underestimate of that inferred from fossil records (> 60 Ma; Farabee, 1993). Within suborder Apiineae (Table 1), more than 60 fossil taxa have been discovered for Araliaceae, dating back to the Cretaceous (Europe and North America) and early Cenozoic (Siberia and Australasia). Whereas most fossils referable to Apiaceae are scarce in the Oligocene and Miocene records, they are increasingly represented in the Pleistocene record (e.g., Axelrod, 1952; Mathias, 1965). Considering recent advancements in fossil dating and phylogenetically based classifications, the oldest apialean fossils are those collected in Germany belonging to the Maastrichtian flora of the Cretaceous Period (c. 70 Ma) and placed in the araliad genera Acanthopanax (= Eleutherococcus) (A. friedrichii, A. gigantocarpus, A. mansfeldensis, and A. obliquocostatus) and Aralia (A. antiqua) (Knobloch & Mai, 1986). The age of these fossils was among six reference dates used by Bremer et al. (2004) to calibrate their molecular clock, resulting in a stem age of 113 Ma for the order Apiales and a crown age of 84 Ma. However, the reliability of many of these fossils remains questionable (Martinez-Millan, 2010). Schneider et al. (2004) estimated an age of 37 Ma based on fossil data and 50 to 80 Ma based on dating molecular phylogenies, whereas Wikstrom et al. (2001) used a fixed fossil age of 69 Ma and estimated the age of Apiales at 85-90 Ma, with the Araliaceae clade originating 41-45 Ma. Farabee (1993) relied on pollen fossils to date the history of Araliaceae back to the Paleocene (55-65 Ma), but few other studies have provided estimates for the timing of diversification among families and genera in Apiales. Moreover, none of the aforementioned studies were focused on diversification within Apiales and the representative sampling form this order was very sparse.
The goal of the present study is to test the hypothesis of an Australasian origin of Apiales dating to the Cretaceous, paralleling that of the angiosperms as a whole. To accomplish this objective, we have determined divergence estimates for all major clades of Apiales, and examined patterns of historical biogeography in the order using an expanded version of the plastid data first presented in Nicolas and Plunkett (2009), which included the widest taxonomic sampling from the order to date, representing all major lineages and most genera of Apiales. In this study, we infer biogeographic patterns in the order using information drawn from a combination of sources, including biogeographic reconstructions, estimated times of divergence, published accounts of paleobotanical evidence and geological history, and patterns of phylogenetic relationships (based on this and prior studies). Given the current limitations in biogeographic analyses, we envision this study as an initial attempt to produce a broad framework that will produce hypotheses for future testing using more focused studies within each of the major clades of Apiales.
Materials and Methods
Taxon Sampling, DNA Sequencing, and Alignment
The sampling included 223 terminals representing sequences from the plastid rpll6 intron and the trnD-trnY-trnE-trnTregion including the three intergenic spacers, previously published in Nicolas and Plunkett (2009). To this, we added two key South African species, both from monotypic genera, Marlothiella gummifera (C. Mannheimer 2987, JRAU; genbank accessions KC881526 and KC881532 for rpll6 intron and trnD-trnT respectively) and Phlyctidocarpa flava (C. Mannheimer 2889; JRAU, GB accessions KC881527 and KC881533) in order to improve confidence in relationships in the early diverging lineages of subfamily Apioideae. We also added four outgroup taxa: Lonicera japonica (Plunkett 2255, NY; GB accessions KC881530 and KC881536) and Sambucus canadensis (Plunkett 2256, NY; GB accessions KC881531 and KC881537) from Dipsacales, and Helwingia japonica (Xiang 04C62, NCSC; GB accessions KC881528 and KC881534) and Ilex opaca (Plunkett 2262, NY; GB accessions KC881529 and KC881535) from Aquifoliales. Hence, the final dataset included 229 terminals from 144 genera. All protocols for DNA extraction, PCR amplification, DNA sequencing, and the assembly and alignment of the data matrix followed the methods detailed in Nicolas and Plunkett (2009).
Inferences of Historical Biogeography and Times of Divergence
We selected six fossils attributed to Torricelliaceae, Araliaceae, and Apiaceae to set calibration points. Torricelliaceae offered very reliable macrofossils from fruits of Torricellia bonesii (Manchester) Manchester comb, nov., which are relatively common in deposits of the early Eocene from Europe and North America (Manchester, 1999). More recently, this fossil taxon has been identified from late Paleocene deposits of North America (North Dakota; Manchester et al., 2009), prompting us to calibrate the crown age of Torricelliaceae at 57.25 Ma with 95 % confidence interval of 55.8-58.7 Ma, which spans the late Paleocene Epoch. We used reliable macrofossils of Araliaceae from Dendropanax eocenensis identified from leaf material (Dilcher & Dolph, 1970) and Paleopanax (= Metapanax) identified from fruit material, collected from middle Eocene deposits in North America (Manchester, 1994). Both taxa belong to the Asian Palmate clade of Araliaceae, a group of taxa that appears to have undergone a rapid radiation and probably reticulate evolution, which has hindered the resolution of its phylogenetic relationships in this clade (Wen et al., 2001; Mitchell & Wen, 2004; Plunkett et al., 2004c). Upon examination of published photos of the fossils and considering Dilcher and Dolph's (1970) assessment of a possible relationship of D. eocenensis to Oreopanax, we decided to take a more conservative approach by using the two fossils to calibrate the minimum crown age of the Asian Palmate clade at 42.9 Ma with a 95 % confidence interval spanning the middle Eocene from 37.2 to 48.6 Ma. We also used pollen microfossils assigned to three early diverging lineages in Apioideae related to Steganotaenia and Bupleurum from the lower Eocene, and Heteromorpha from the upper Eocene (Gruas-Cavagnetto & Cerceau-Larrival, 1984). We used the fossils related to Steganotaenia and Bupleurum to calibrate the stem ages for the clades representing Steganotaeneae and Bupleureae (respectively), both at 52.2 Ma with a 95 % confidence interval between 48.6 and 55.8 Ma, which spans the entire lower Eocene. The fossil related to Heteromorpha was used to set the minimum stem age of the Heteromorpheae clade to 35.55 Ma and 95 % confidence interval spanning the middle Eocene from 33.9 to 37.2 Ma. To set a constraint on the root, we had to rely on prior estimates for the ages of Aquifoliales and Dipsacales. Because these estimates have not been consistent across previous studies, we relaxed the stringency for placing an age at the root. In doing so, we considered prior estimates and calibrations (e.g., Bremer, 2000; Wikstrom et al., 2001; Chaw et al., 2004; Bell & Donoghue, 2005) to set the median age of the root to 115 Ma, and relaxed the estimate with 97.5 % confidence intervals between 100 and 130 Ma.
Estimates of Divergence
Estimation of rate variation was assessed using the software BEAST vl.5.2 (Drummond & Rambaut, 2007). The model is optimized through Bayesian Markov chain Monte Carlo (MCMC) methods without requiring rate autocorrelation or a starting phylogram, and thus provides a better account for phylogenetic uncertainty (Drummond et al., 2006; Rutschmann, 2006). Estimates of divergence time were determined by a data matrix of aligned DNA sequences and reliable fossils to calibrate branches across various lineages. BEAST XML files (available from the authors upon request) were generated using the Beauti vl.5.2 software (available in the BEAST package). jModeltest 0.1.1 (Posada, 2008) was used to estimate the best model of sequence evolution to be implemented in BEAST. We also conducted a likelihood ratio test (LRT) to determine whether the sequences evolve according to a strict or relaxed molecular clock. Consequently, we used the GTR + [GAMMA] +1 as the model of evolution and performed the analyses using relaxed clock models with uncorrelated lognormal (UCLN) distribution, which account for rate variations across branches and do not assume a priori correlation between a lineage and its ancestor (Drummond & Rambaut, 2007). We set the tree prior to the Birth-Death speciation process (Gemhard, 2008) and ran two separate analyses, each for 100 million generations. The two runs were combined using LogCombiner vl .5.2 for a total 200 million generations with sampling of trees every 5,000 generations (for a total of 40,000 trees). We used Tracer v 1.4.1 (Rambaut & Drummond 2007) to estimate the bum-in and assess the effective sample sizes (ESS) for the traces saved in the BEAST log file. ESS values are indicators of the reliability of the results and whether they provide a good representation of the posterior distribution (see Tracer Tutorial at http://beast.bio.ed.ac.uk/Analysing_BEAST_output). The tree file was transferred to TreeAnnotator vl.5.2 (in the BEAST package), where 20 % of the trees were discarded as the bum-in, and a maximum clade credibility tree was estimated from the remaining 32,000 trees. The maximum clade credibility tree was viewed in FigTree vl.3.1 (available from http://tree.bio.ed.ac.uk/software/figtree/).
The process of tree searching in BEAST involves a single chain, which leaves it susceptible to getting stuck on local optima in tree space and thus producing a tree different from the one produced in more robust Bayesian methods (e.g., MrBayes). Therefore, we decided to compare the chronogram produced in BEAST to the trees produced in MrBayes v3.1.2 (Ronquist & Fluelsenbeck, 2003) and to trees estimated by maximum likelihood in GARLI vl.O (Zwickl, 2006) under the GTR + [GAMMA] + 1 model. We analyzed the matrix in MrBayes by setting it to perform two simultaneous runs, each with one hot and three cold chains. Each chain was ran for 5 million generations, after which the standard deviation of split frequencies of different chains was less than 0.01, an indication that the runs have converged on the same topology. The files were checked in Tracer, which estimated the bum-in to be 10 %. Twelve separate maximum likelihood analyses were performed in GARLI and the support values of nodes were estimated with 1,000 bootstrap replicates.
We defined geographic areas based on divisions given by Pielou, (1979; cf. Sclater, 1858 and Wallace, 1876) with consideration to those given by Morrone (2009) and the distribution pattern of major groups in Apiales. As such, six broad geographic characters were defined: Nearctic (Area A), Neotropical (Area B), Palearctic (Area C), Ethiopian (Area D), Oriental (Area E), and Australasian (Area F; includes Oceanian and Antarctic regions) (Fig. 1). We used two different approaches to examine historical biogeographical relationships: an event-based maximum parsimony approach and a parametric maximum likelihood approach. The event-based method of reconstruction of areas was implemented using the DIVA software package (Ronquist, 1996, 1997), where vicariance was set as the default cause of speciation (no cost) while accounting for dispersal and extinction events using a priori cost assignment (one per event). The outgroup was removed and the maximum number of areas at each node was set to two. Due to the large size of our data set and the limitations in the number of terminals and characters accepted in DIVA (see DIVA manual), we constructed two separate DIVA analyses. The first included all taxa in the Apioideae-Saniculoideae clade with 53 terminals. The second included 109 terminals, using a single terminal to represent the Apioideae-Saniculoideae clade and 108 terminals for the remaining clades of Apiales. As recommended by Ronquist (1996), the area given as the origin of Apioideae-Saniculoideae in the first analysis was used to code that terminal in the second analysis.
The second approach to biogeographic reconstruction uses a maximum-likelihood method as implemented in the software Lagrange (Ree & Smith, 2008). Lagrange allows the use of parametric models to test different biogeographic scenarios of dispersal and extinction, and for each clade it provides a probability based on the likelihood that an area is the center of origin of that clade (Ree et al., 2005, Ree & Smith, 2008). For analyses in Lagrange, we used the tree generated in BEAST (after trimming the outgroup) and a matrix representing the distribution of taxa among the six geographic areas mentioned above. We then estimated the most likely origin for each node in the phylogeny under different scenarios of area relationships using 11 geographic ranges, with and without geologic and time constraints (i.e., stratified and non-stratified models; Fig. 1). Python files were generated using the web-based configuration tool developed by the authors of Lagrange (http://www.reelab.net/lagrange) and the files were run in Python 2.5. The phylogenetic placements of two important clades in Apiaceae, Klotzschia and Lichtensteinia have been unstable in prior studies (see Nicolas & Plunkett, 2009. To explore the impact of alternative placements on the biogeographic inferences, we repeated the analyses using these alternative topologies.
The DNA sequence matrix comprised 229 terminals and 3469 aligned characters, 1742 of which were parsimony informative. The likelihood ratio test gave a value for p that approached 0, which indicated that the sequences did not evolve according to a strict molecular clock. After assessing the topologies and support values of trees resulting from different methods of phylogeny estimation, we identified three major sources of incongruence, namely the placements Klotzschia, the Lichetensteinia clade, and Harmsiopanax. Klotzschia was placed either as a lineage that diverged after Azorelloideae and before Hermas, or as sister to the rest of Azorelloideae. The Lichtensteinia clade had three possible placements, either sister to Saniculoideae sensu lato, or as a clade that diverged before or after Saniculoideae sensu lato. Harmsiopanax was either sister to Hydrocotyle-Trachymene or sister to a lineage that diverged after Hydrocotyle-Trachymene (and thus sister to the rest of Araliaceae). However, none of these placements had strong support and the only main difference in the biogeographic inferences resulted from the alternate placements of Klotzschia (discussed below). Results for the estimates of dates of origin generated in BEAST (Fig. 2, Table 2; detailed tree is shown in Appendix SI) and the geographic areas of origin that were reconstructed in DIVA and Lagrange (Table 3; detailed Lagrange output is shown in Appendix S2) are reported together with the Discussion.
Relationships Among the Major Clades of Apiales
The early diverging families Griseliniaceae, Torricelliaceae, and Pennantiaceae form successive sister groups to Apiineae (see Fig. 2, Table 1). The three families differ from Apiineae in several structural features, including the lack of schizogenous secretory canals (see Plunkett & Lowry, 2001) and their pollen exine texturing, which is usually reticulate in most Apiineae but varies in the three remaining families (Karehed, 2003). The morphological evidence supporting the inclusion of these families in Apiales is scant. Several features common in Apiineae are also found in most of these groups, such as ovary-roof nectaries (Erbar & Leins, 2010), a single functional ovule (either per locule or per ovary), drupaceous fruits, and sheathing petiole bases, but there are significant exceptions to each. By contrast, molecular evidence has consistently suggested that these families belong to Apiales as early-diverging lineages.
Within Apiineae, Pittosporaceae have strong support as the sister group to the remaining clades. A distinguishing morphological feature of Pittosporaceae is the presence of superior ovaries with parietal or axile placentation, whereas the other three families of Apiineae have inferior ovaries with an apical placentation. Despite its traditional association with Rosidae, there were early indications of an affinity of Pittosporaceae to Apiales based on chemical data (Hegnauer, 1971; Bohlmann, 1971), anatomical characters (Jurica, 1922; Rodriguez, 1971), and cytology (Darlington & Wylie, 1955; Jay, 1969). The traditional circumscriptions of Araliaceae and Apiaceae and their relationships as a "family pair" (sister families) are not supported. The principle exceptions are the placements of taxa now recognized as Myodocarpaceae and the Hydrocotyle-Trachymene clade of Araliaceae. This picture of relationships helps to explain the pre-cladistic view that the genera of Myodocarpaceae (Myodocarpus and Delarbrea) represent "bridging groups" between Apiaceae and Araliaceae (Plunkett & Lowry, 2001). These plants share the woody habit of most Araliaceae, and Delarbrea also has the drupaceous fruits characteristic of that family, but with thin endocarps. By contrast, Myodocarpus has dry, schizocarpic fruits with free carpophores, reminiscent of many Apioideae. Basic chromosome numbers (x=12) also provide a link between Myodocarpaceae and Araliaceae. However, wood anatomical characters, such as non-septate fibers and thin intervessel pits, suggest a connection to Apiopetalum and Mackinlaya, both formerly placed in Araliaceae, but now placed in Apiaceae subfamily Mackinlayoideae (Plunkett et al., 1996a, b, 1997, 2004a, b, c; Oskolski & Lowry, 2000).
Within Apiaceae, four major clades are evident, the earliest diverging of which is Mackinlayoideae, followed by the Platysace clade. Both groups are largely Australian. The two other clades are the Andean-Subantarctic-Australian Azorelloideae, and the cosmopolitan Apioideae-Saniculoideae. Of uncertain placement here is Klotzschia, a Brazilian genus of three species, which has morphological affinities to Azorelloideae, where it is placed incertae sedis (Nicolas & Plunkett, 2009). This genus was also thought to have affinities to Araliaceae (especially K. glaziovii), mostly due to similarities in pollen morphology (Shoup & Tseng, 1977). Relationships among early diverging lineages in Apioideae have been problematic, and the line between subfamilies Apioideae and Saniculoideae has recently been blurred due to the results outlined by Magee et al. (2010), who suggested the merging of the two subfamilies. Hermas, a genus endemic to the fynbos regions of the South African Cape Province, appears as the earliest diverging lineage of this group. As in Magee et al. (2010), our trees place Marlothiella sister to Choritaenea in the Lichtensteinia clade, and Phlyctidocarpa appears as the earliest diverging lineage in the Steganotaenae-Saniculae clade.
Divergence and Biogeography
Starting at the deepest branches in Apiales (Figs. 2 and 3), our analyses estimate that the divergence of the Pennantiaceae lineage from the rest of the order occurred in the mid-Cretaceous, with a median minimum age of c. 117 Ma and 95 % confidence interval between 104 and 130 [104, 130]. Torricelliaceae diverged later, c. 107 Ma [94, 120], followed by Griseliniaceae, which diverged from the rest of Apiales at c. 103 Ma [90, 116] (Fig. 2; Table 2). Suborder Apiineae (Pittosporaceae, Araliaceae, Myodocarpaceae, and Apiaceae) originated c. 98 Ma [85, 110]. The results from DIVA can be interpreted in two ways for the biogeographic origin of Apiales, either Australasia (F) or Australasia + Oriental (EF), followed by a vicariance event that led to the separation of Torricelliaceae (with an origin in the Oriental region) and diversification in Australasia for the remaining lineages. Similarly, all models applied using Lagrange suggested that the highest relative probability (rp) for the split at the root is EF/F (rp values for the three models ranged between 0.32 and 0.36). The second most-likely scenario was F/F (0.26<rp<0.32), followed by F/E at the stem of Torricelliaceae (0.42<rp <0.56), interpreted as an origin in Australasia for Pennantiaceae and Oriental-Australasia for the rest of the order, which subsequently experienced a split between the Oriental area for Torricelliaceae and Australasia for Griseliniaceae + Apiineae (Fig. 3; Table 3). With a date exceeding 100 Ma, the Oriental area (E) at that time was represented by India (still not separated from Madagascar) and Australasia (F) was represented by Australia, Antarctica, and Zealandia (which includes the current landmasses of New Zealand and New Caledonia); Zealandia eventually broke away from Australia and West Antarctica in the Late Cretaceous, c. 80 Ma (Kula et al., 2007; Trewick et al., 2007). The same scenario was evident in the DIVA reconstruction for the origins at the nodes following Pennantiaceae. Considering the ages of Pennantiaceae, Torricelliaceae, Griseliniaceae, and Apiineae, these four clades appear to represent ancient lineages of Gondwanan origins. A likely conclusion, therefore, is that the ancestors of Apiales inhabited East Gondwana in the Early to Mid-Cretaceous. The divergence of Torricelliaceae from the rest of Apiales may have been correlated with the separation of India and Madagascar from the rest of East Gondwana, which is estimated to have occurred during the Early Cretaceous (Briggs, 2003; Cox & Moore, 2010; and consistent with our results), followed by the drifting of India towards Asia.
The Early-Diverging Families
Pennantiaceae. The link between Pennantiaceae and the rest of Apiales is tenuous, where only ovary position and low carpel number are shared. Despite its placement in plastid phytogenies in this and prior studies (e.g. Karehed, 2003; Chandler & Plunkett, 2004), results from two nuclear datasets of Apiales have placed it among the outgroups (Dipsaclaes and Aquifoliales) rather than as sister to the rest of Apiales (Chandler & Plunkett, 2004; Nicolas, 2009). Regardless, exclusion of Pennantiaceae from the ingroup in the Lagrange analyses had no effect on interpretations in the rest of the tree, and the origin of Torricelliacea + Griseliniaceae + Apiineae remained Oriental + Australasia (EF).
If Pennantiaceae belong to Apiales, they appear to have diverged c. 117 Ma and Lagrange placed the origin of the stem and crown of Pennatiaceae in Australasia. The four extant species of Pennantia are each restricted in distribution to a single landmass (Australia, New Zealand, Norfolk Is., or Three Kings Is.). The age of the lineage makes Australia + Zealandia the likely origin. Evidence from geologists and biologists support the theory that after the separation of Zealandia, most or all of New Zealand underwent submergence periods and reemerged in the early Miocene, which suggests that the age of its flora may be much younger than 100 Ma (Trewick et al., 2007; Landis et al., 2008; but see Jolivet & Verma, 2010 and references therein for a different view). This implies that the current presence of Pennantia in New Zealand is most likely due to dispersal from Australia, which agrees with Mildenhall's (1980) indication that Pennantia appeared in New Zealand in the Quaternary.
Torricelliaceae. Torricelliaceae includes three genera with quite distinct distributions, Melanophylla (7 spp.) in Madagascar, Torricellia (3 spp.) from SW China, the eastern Himalayas (Nepal to Bhutan) and India (Khasia Hills and Assam), and Aralidium (1 sp.) in western Malesia (Fig. 3). The median minimum age of divergence of Torricelliaceae is 107 Ma, which makes the lineage of Gondwanan age. The origin at the stem is most likely due to vicariance events in East Gondwana, associated with the separation of Australia and Africa and the subsequent drifting of India and Madagascar from Africa. The origin of the crown of Torricelliaceae appears to be either Oriental (in this case India; 0.42<rp<0.58) followed by dispersal into the Ethiopian region (Madagascar), or alternatively an origin in India-Madagascar (area DE; 0.17<rp< 0.34), followed by speciation through vicariance (Fig. 3). However, the separation of India and Madagascar occurred in the Late Cretaceous (c. 90 Ma; Ravala & Veeraswamya, 2003), much earlier than the age of divergence of the Malagasyendemic Melanophylla from the rest of Torricelliaceae (c. 20 Ma with an uppermost limit at c. 43 Ma), but dispersal from India (where Torricellia persists to this day) to Madagascar remained possible until the Miocene (Gentry, 1982; Schatz, 1996; Briggs, 2003; Renner, 2004). Based on these data, the diversification of Melanophylla in Madagascar most likely occurred in the late Oligocene-early Miocene as India drifted farther from Madagascar and gene flow between populations in India and Madagascar would have become increasingly unlikely. Routes of dispersal out of India explain present day distributions of Aralidium and Torricellia in Malesia and tropical Asia (see also Schatz, 1996). Fossils from species of Torricelliaceae were located in Europe and North America, as old as the late Paleocene (Manchester, 1999, Manchester et al., 2009), which suggests that the family had dispersed from Asia into Europe and North America during a time when these areas were tropical. Extinction in Europe and North America may be explained by the shift from a tropical to a cooler, more temperate climate, which commenced around the end of the Eocene into the Oligocene for Europe and South America (Manchester, 1999).
Griseliniaceae. The monogeneric Griseliniaceae exhibits a South American-New Zealand disjunction, with a greater number of species in austral South America (5 spp.) than New Zealand (2 spp.). Our BEAST analyses date this lineage to the mid-Cretaceous (c. 103 Ma) and results from both DIVA and Lagrange suggest an origin of Griseliniaceae in Australasia (rp>0.90) (Fig. 3). The oldest fossils attributed to Griselinia are macrofossils (leaf material) from lower Miocene New Zealand (Otago, South Island) (Pole, 2008). Based on pollen records, Mildenhall (1980) predicted that Griselinia appeared in New Zealand in the Miocene, which is supported by the minimum crown age retrieved for the genus (c. 12 Ma). This age also represents the split of the New Zealand taxa from the South American ones. Given that the minimum age of the lineage (c. 103 Ma) far exceeds the time of the initiation of the break-up of Zealandia from Australia (c. 80 Ma), we predict an origin of Griseliniaceae in Australia + Zealandia and suggest that the presence of the only two extant taxa in New Zealand could be explained by one of the three following scenarios. In the first scenario, the high geologic activity and low stability of Zealandia in the Late Cretaceous and early Cenozoic, along with the submergence of New Zealand during the Eocene-Oligocene, suggests that Griselinia may have arrived in New Zealand during the Miocene through long distance dispersal from Australia (where it subsequently went extinct).
Long-distance dispersal of Griselinia is possible because of certain adaptations of its seeds that make them well suited for dispersal by migratory birds (which are known to migrate long distances), including their sensitivity to dryness and their inability to germinate until the fleshy fruit has been removed (Burrows, 1995, 1999; Bryan et al., 2011). Under this scenario, the presence of Griselinia in southern South America was likely caused by long-distance dispersal via stepping stones in the Subantarctic region during the Miocene. This agrees with a prior theory proposed by Dillon and Munoz-Schick (1993) in their revision of Griselinia.
A second scenario suggests that the common ancestor of Griseliniaceae originated in the landmass that included Australia + Zealandia, but diversified only on the part of Zealandia that is now New Zealand, or had a wider distribution and subsequently went extinct in Australia and other parts of Zealandia (e.g., New Caledonia). In this scenario, the two extant species in New Zealand represent relicts of an old lineage that survived the extensive geologic activity, and remained on island refugia of Zealandia that escaped submergence and glaciation. If Griseliniaceae originated in New Zealand as a result of the break up of Zealandia from Australia, then the early lineages of Apiales were shaped by vicariance events involving the break up of Australasia-Africa (Torricelliaceae), followed by the separation of Zealandia (Griseliniaceae) from Australia (suborder Apiineae) in Australasia.
A third scenario would involve assuming an origin of Griseliniaceae in South America, where five species of Griselinia are extant in Chile and Argentina (one of which also extends to southeastern Brazil). Here, diversification into New Zealand would likely have resulted from dispersals across the Subantarctics in the Miocene. Under this scenario, the three oldest lineages of Apiales (Torricelliaceae, Griselineaceae, and Apiineae) would fit the Africa-South America-Australasia breakup model of Cretaceous Gondwana. While agreement of the third scenario with this model of vicariance may have some appeal, it largely ignores the results of our biogeographic analyses (which suggest an Australasian origin), and is based merely on the greater number of extant species of Griselinia in South America.
Our results indicate that suborder Apiineae originated in the mid-Cretaceous (c. 98 Ma) in Australasia. Radiations within a period of 15 Ma led to the origin of the four lineages that are currently recognized as families: Pittosporaceae, Araliaceae, Myodocarpaceae, and Apiaceae (Fig. 2). The Australasian origin was estimated by DIVA and supported by Lagrange (rp>0.95). Among the mid-Cretaceous Australasian landmasses, Australia appears to be the most likely center of origin. Based on the extant distribution and ages of many lineages retrieved in our analyses, it appears that New Caledonia, New Zealand, and New Guinea played important roles in the preservation and diversification of these lineages during subsequent epochs at a time when many taxa went extinct in Australia.
Pittosporaceae. Pittosporaceae appear to have diverged during the Cretaceous with a stem age of c. 98 Ma [85, 110]. The family has its greatest generic diversity in Australia, where nearly all of the nine genera are endemic. The largest genus, Pittosporum, also has centers of diversity in New Caledonia and New Zealand, but it extends to other islands of the Pacific, into Asia, and across the Indian Ocean to Madagascar and eastern Africa. Inferences from DIVA and Lagrange reconstructions suggest an Australasian origin for the group (rp>0.95; Fig. 4). The estimate of the minimum median age of the Pittosporaceae crown group was c. 19 Ma, which indicates that diversification among the major lineages within the family is post-Gondwanan. Apart from the eight genera which diversified in Australia, the most speciose genus, Pittosporum, appears to have dispersed from Australasia (most likely Australia) to Madagascar through the Indian Ocean Basin (JOB), and through the Pacific into Malesia and tropical Asia. Such dispersal would have been facilitated by adaptations to the fruits of many Pittosporaceae (especially Pittosporum), which are often brightly colored and contain sticky seeds, making them attractive to birds. Within Australasia, the genus dispersed from Australia to the Pacific islands and into tropical Asia (see Gemmill et al., 2002; Chandler et al., 2007). These dispersals are very recent, with divergence time estimated at less than 10 Ma.
Araliaceae. The analyses show that Araliaceae had a median minimum stem age c. 95 Ma [83, 107], and like the suborder as a whole, originated in Australasia (rp>0.95) (Fig. 4). A thorough understanding of the biogeography of Araliaceae continues to be hindered by the many unresolved or poorly supported relationships remaining in this family. For that reason, we limit our focus on the major, well supported clades to examine trends of diversification within such clades. One such example is the Hydrocotyle-Trachymene group, which belongs to a lineage that appears to have diverged from the rest of Araliaceae c. 65 Ma and has an Australasian origin (rp>0.95). The Hydrocotyle and Trachymene subclades subsequently diverged from one another c. 51 Ma. Compared to most Araliaceae, Hydocotyle and Trachymene mark a shift from woodiness to herbaceousness along with a preference for temperate habitats, especially in Western Australia. Despite their prior placement in Apiaceae subfamily Hydrocotyloideae, the affinity of these genera to Araliaceae has been noted as far back as Seemann (1863), who concluded that the valvate petal aestivation of Hydrocotyle warranted its transfer to Araliaceae. Seemann also noted similarities to Horsfieldia Blume ex DC. (= Harmsiopanax), which appears close to Hydrocotyle and Trachymene in some trees (Nicolas & Plunkett, 2009). Other morphological similarities also link the Hydrocotyle-Trachymene group to members of Araliaceae, including sclerified endocarps and laterally-compressed, bicarpellate fruits.
Apart from Hydrocotyle-Trachymene and Harmsiopanax, Afro-Madagascan Schefflera is the earliest diverging clade among the rest of Araliaceae, dating back to the early Eocene with a stem age of c. 48 Ma [34, 64], Our results indicate that the presence of this clade in Africa and Madagascar is likely due to long-distance dispersal from Australia (p>0.70), across the western Indian Ocean Basin (Fig. 4). Dispersal events across the IOB have been reported in many plant groups and at different ages (including in other groups of Araliaceae, such as Polyscias s. late, see Plunkett et al, 2004b), and may have led to secondary dispersals into Africa from Madagascar (Schatz, 1996; Sanmartin & Ronquist, 2004).
The remaining groups of Araliaceae have a stem age of c. 38 Ma (95 % confidence interval between 29 and 47 Ma), with the Raukaua group (includes Raukaua, Cephalaralia, Motherwellia, true Schefflera, Cheirodendmn) diverging first and having a minimum crown age of c. 23 Ma. This group has an Australasian origin (rp>0.99), where it is currently found in Australia, New Zealand, New Caledonia, and the Pacific Islands (Fig. 4). The Asian Palmate clade diverged c. 34 Ma and the origin of its crown group is most likely the Palearctic region (Area C) (0.52<rp<0.69). Extant taxa in this clade exhibit a disjunct distribution in eastern Asia and South America. However, the presence and timing of fossils from the late Eocene of North America (Dilcher & Dolph, 1970, Manchester, 1994) suggest that North American representatives reached that region through Europe (Fig. 4). Extinction in both Europe and North America most likely occurred as these continents cooled in the Miocene, which agrees with suggestions by some authors of a boreotropical origin for the Asian-Neotropical disjunctions in this clade (e.g., Plunkett et al., 2004c).
A third major group of Araliaceae includes the Polyscias and Pseudopanax clades. The BEAST estimate for the crown age of this group at c. 17 Ma [9, 25] and the Lagrange analysis suggested an origin of the crown group in Australasia (0.89<rp<0.95). While the Pseudopanax clade is restricted to the Pacific region, the Polyscias clade has a Paleotropical distribution that may have originated in Australia, from which it dispersed multiple times both eastward into the Pacific and westward into the IOB (Fig. 4; see also Plunkett et al., 2001, 2004b; Plunkett & Lowry, 2010).
The resolution of phylogenetic relationships in Araliaceae remains a work in progress. Conclusions regarding the biogeographic events at the stem of these groups are hampered by the lack of support at many nodes and the topological incongruence resulting from analyses of different sources of data. However, with the available data, our analyses show support for an origin of the family in Australasia, followed by major dispersal events to Africa and Asia.
Myodocarpaceae. Myodocapaceae represent an ancient lineage in Apiineae, and the BEAST analysis suggests that it diverged from the rest of the suborder in the Late Cretaceous, c. 91 Ma [80, 103], The Lagrange and DIVA analyses suggest that the only two genera in this family, Delarbrea and Myodocarpus, originated in Australasia (rp> 0.90), where all extant species are found (Fig. 5a, b). In fact, 15 of the 17 species in the family are entirely restricted to New Caledonia, excepting only two species of Delarbrea (D. michieana, which is endemic to Queensland, and D. paradoxa, which ranges from New Caledonia to other nearby Pacific islands). The estimated time of origin for Myodocarpaceae suggests that the family may have originated before the split of New Caledonia from Australia, which is estimated to have occurred as part of the break up of Zealandia from Australia in the Late Cretaceous (65-85 Ma; see McLoughlin, 2001 and references therein). Conflicting theories regarding the age of the New Caledonian flora and the persistence of refugia during times of submergence have been suggested (see Lowry, 1998; Ladiges & Cantrill, 2007; Grandcolas et al., 2008; Heads, 2008, 2010). This presents two possible hypotheses regarding the origin of Myodocarpaceae: (1) that New Caledonia experienced complete submergence, in which case this lineage most likely dispersed to New Caledonia after its re-emergence, followed by rapid diversification on the island but extinctions in Australia and other nearby regions during Eocene climatic changes, or (2) that there was incomplete submergence of New Caledonia, allowing the persistence of emergent refugia, in which case Myodocarpaceae most likely originated as a result of the separation and isolation of New Caledonia. In the latter case, the current distribution of two species of Delarbrea in Australia and the Pacific Islands most likely resulted from post-Oligocene dispersal. The median minimum crown age of the family is estimated to be c. 25 Ma and thus dispersal was possible through connections that persisted during the Miocene between New Caledonia and Australia through the Sunda arc (Barlow, 1981). Such dispersal events may be mediated by birds, as the fleshy fruits of Delarbrea are readily dispersed by medium- to large-size birds (Lowry, 1986), whereas the dry, winged fruits of the endemic Myodocarpus are strictly anemochorous. One key factor for the diversification of Myodocarpus on New Caledonia may have been its adaptability to serpentine soils. Eight of its ten species grow exclusively on serpentine-like ultramafic soils, an adaptation that may have contributed to species endemism and increased competitive success (Lowry, 1991; Mayer & Soltis, 1994; Morat, 1993; De Kok, 2002).
Apiaceae. The origin of Apiaceae in Australasia (area F) is supported by both DIVA and Lagrange (0.73<rp<0.84) (Fig. 5a, b), and our BEAST analyses date this origin to the Late Cretaceous/early Eocene (c. 87 Ma). The four main clades correspond to origins in three distinct areas: two clades in Australasia (Mackinlayoideae and Platysace), a third in South America (Azorelloideae), and the fourth in Africa (Apioideae-Saniculoideae). The estimated divergence times of these four clades (<90 Ma) is too recent to correspond to the break-up of Australasia, South America, and Africa landmasses, but connections among these three areas were still possible during the times estimated herein, and the diversification may be attributed to the break up of Antarctica. Connections between Australasia and South America were present through West Antarctica until the Eocene, but the connection to Australia was affected by the separation of Zealandia and the submergence of the South Tasman Rise c. 64-90 Ma (Gaina et al., 1998; McLoughlin, 2001 and references therein). Connections between Australasia and Africa were affected by the geology of East Antarctica, with Africa breaking apart from Australia-East Antarctica more than 100 Ma. This exceeds the age of the oldest clade of African origin in Apiaceae (c. 71 Ma, uppermost limit of 80 Ma), but studies have speculated that connections between Africa and East Antarctica remained possible through the Kerguelen Plateau up to c. 80 Ma (see Schwarz et al., 2006 and references therein). South America and Africa remained in close proximity until the Late Creatceous (80-90 Ma; Gentry, 1982). However, more recent dispersals between these two continents remained possible during the estimated time of origin for Apioideae-Saniculoideae in Africa (c. 71 Ma) via land connections across the Rio-Grande Rise and the Walvis Ridge, especially from Brazil (where Klotzschia is extant) to southern Africa (where Hermas is extant) (see Morley, 2003 and references therein).
The node representing the divergence of Platysace from the rest of Apiaceae had different possible reconstructions in DIVA, depending on the placement of Klotzschia. The issue was not resolved in Lagrange because different models failed to provide a consensus regarding vicariance, dispersal, or a combination of both as the explanation for the divergence of the three clades. Two alternative scenarios may explain the diversification of Apiaceae out of Australasia. In the first scenario, in which Klotzschia is sister to the rest of Azorelloideae, the ancestral area given by DIVA was either South America-Australasia (BF) or Ethiopian-Australasia (DF), and the same two reconstructions (B/F or D/F) were almost equally likely in the Lagrange results (Fig. 5a). This implies that dispersal occurred from Australasia, through West Antarctica, to South America (where Azorelloideae originated) and from Australasia, through East Antarctica across a Kerguelen land bridge, to Africa (where Apioideae-Saniculoideae originated). In the second scenario, in which Klotzschia diverged after Azorelloideae, the ancestral area given by DIVA was BF at the Platysace-Azorelloideae node (compared to B/F or F/F in Lagrange) and BD at the Klotzschia-Hermas node (compared with D/B in Lagrange) (Fig. 5b). This scenario implies dispersal from Australia to South America (most likely through Antarctica), then from South America to Africa (most likely through the Rio-Grande Rise/Walvis Ridge connection). Such exchanges were dramatically reduced after the breakup of Antarctica from Australia, the submergence of many landmasses in the region, and the Antarctic glaciation in the late Paleocene-early Eocene (Lawver & Gahagan, 2003), which led to the disjunctions and the isolation of these plant groups in their respective areas. The following sections provide greater detail for the diversification of the major clades of Apiaceae.
Mackinlayoideae and Platysace
Mackinlayoideae are another ancient lineage in Apiales, and the earliest diverging lineage of Apiaceae, originating in Australasia (Fig. 5a, b) and dating to c. 87 Ma [76, 99]. The divergence of Actinotus and Apiopetalum from the remaining taxa occurred c. 66 Ma, followed by Mackinlaya, which separated from the mostly herbaceous taxa in the clade c. 57 Ma. DIVA and Lagrange reconstructions suggested an Australasian origin for each of the three main clades (rp>0.95). Australasian groups from this clade include Actinotus (Australia), Apiopetalum (New Caledonia), and Mackinlaya (Australia to Malesia). The mackinlayoid genera Xanthosia, Chlaenosciadium, Pentapeltis, and Schoenolaena diversified entirely within Australia and represent another shift to herbaciousness originating in this region, parallelling that of Hydmcotyle-Trachymene in Araliaceae. The node ages for the mackinlayoid clades do not indicate rapid radiation events (node ages c. 47, 32, and 20 Ma). Dispersal events to Mesoamerica (Micmpleura) and Africa (Centella) must be invoked to explain the distribution of these taxa, which appear to have diverged after Schoenolaena, c. 20 Ma. This suggests a Miocene dispersal into South Africa and South America. Use of the alternative models supplied in the Lagrange analyses resulted in different scenarios for the dispersal out of Australia: dispersal to Africa (Centella) and thence to South America, or vice-versa, or dispersal to both South America and Africa followed by isolation in these respective areas. In each of these scenarios, a likely explanation for the distribution of these mackinlayoids is relatively recent dispersal through Subantarctic stepping stones.
The Platysace clade (Platysace + Homalosciadium) diverged from the rest of Apiaceae c. 84 Ma [73, 95]. Support for this clade is not very strong, however, and plastid data place it as a separate lineage from Mackinlayoideae, sister to the rest of Apiaceae (i.e., Azorelloideae + Saniculoideae + Apioideae). The Platysace clade is restricted to Australia. A shift from wet tropical climates to more arid temperate climates is correlated with Platysace as well as with the clade of herbaceous Mackinlayoideae (Chlaenosciadium, Xanthosia, Schoenolaena, and Pentapeltis), both of which exhibit high levels of endemism in temperate southwestern Western Australia. Platysace appears to be the youngest of the major Australasian lineages of Apiales. The remaining lineages are more recently derived, but originated either in South America (Azorelloideae) or Africa (Apioideae-Saniculoideae).
Azorelloideae and Klotzschia
The split between Azorelloideae and Apioideae-Saniculoideae dates back to c. 76 Ma [66, 86]. The placement of Klotzschia remains unsettled, but if we assume it belongs to Azorelloideae, its divergence from the rest of that group dates back to c. 74 Ma. This subfamily appears to have had a South American origin (rp>0.70), where its major clades all still exhibit high diversification, especially in the Andes. The radiation of its three major clades (Azorella, Asteriscium, and Bowlesia) is estimated at c. 66 Ma, and their median crown ages that range between 48 and 56 Ma. These diversification times represent periods of gradual upliftings and other geological changes in the Andes, especially in the central and southern parts of that range (Gentry, 1982; Megard, 1984). The diversification of these groups may reflect adaptations either to open areas at high elevations (e.g., Azorella), to shady, humid areas of high elevations (e.g., Bowlesia), or to deserts (e.g., Eremocharis). Many azorelloids are also characterized by a sufffutescent and/or cushion habit, which may have been an adaptation to harsh weather conditions such as severe cold (e.g., Azorella) or aridity (e.g., Eremocharis). Remarkably, each of these three major clades exhibits one or more independent dispersal events out of southern South America, including dispersals to Australia (Diplaspis in the Azorella clade, Oschatzia in the Asteriscium clade, and Dichosciadium in the Bowlesia clade), the Subantarctic Islands (Stilbocarpa and Azorella selago), and New Zealand (Schizeilema) (see Nicolas & Plunkett, 2012).
The minimum age of Apioideae is estimated at c. 71 Ma [62, 80], with an origin in the Ethiopian region (rp>0.63) (Fig. 5a, b). The South Africa endemic genus Hermas is the earliest diverging lineage. Although the placement of the African Lichtensteina clade remains tenuous, alternative placements of this clade in the phylogeny had no impact on the biogeographic analyses. Such congruence is to be expected because the clades diverging before and after the Lichtensteina clade are all of African origin. The separation between Saniculoideae and Apioideae most likely occurred c. 66 Ma [58, 74] and both subfamilies exhibit an African origin (rp>0.95). In Saniculoideae, the earliest diversification out of Africa is estimated to be c. 30 Ma, for the clade that unites Eryngium + Sanicula + Astrantia and related genera. These diversifications occurred in the Palearctic (North Africa, Europe, and Asia) as Africa and Europe were coming into close proximity during the Oligocene. Within the African region, South Africa appears to be the center of origin for Apioideae-Saniculoideae, which is reflected in the distribution and endemicity of the oldest lineages, especially in the South African Cape region (e.g., Hermas, Steganotaenia, Arctopus, Heteromorpha, Anginon, Lichtensteinia, Marlothiella, Annesorhiza). Many plants of these lineages are woody, which may suggest that woodiness was a characteristic of the ancestor of Apioideae, with another major shift to herbaceousness in clades that diverged after Bupleurum (c. 43 Ma). In our sampling of Apioideae, the earliest diverging clades not found in Africa include the Central American Neogoezia, a representative of tribe Oenantheae (c. 43 Ma), and the New Zealand-Australian Anisotome group, which represents tribe Aciphylleae (c. 38 Ma compared to c. 35 Ma by Spalik et al., 2010).
Beyond these early-diverging lineages, our sampling from Apioideae is very limited and was not a major focus of this study. The apioids experienced a massive diversification both within and outside of Africa, especially in the temperate regions. This is evidenced by the large number of its species in Africa, the Mediterranean, and throughout Eurasia. Recently, Spalik et al. (2010) addressed the amphitropic amphiantarctic disjunctions with much denser sampling in Apioideae, but more detailed sampling and statistical analyses are necessary to explain the complex diversification routes of the very speciose Apioideae-Saniculoideae groups out of (and sometimes back to) Africa.
Summary and Hypotheses of Biogeographic History
Based on the analyses presented here, Apiales appears to have originated in Australasia, with Australia as the likely center of origin. Vicariance due to the Gondwanan break-up shaped the distribution and diversification of the early-diverging families in Asia-Africa and Australasia. Suborder Apiineae originated in Australasia in the Middle to Late Cretaceous. Collectively, Araliaceae, Myodocarpaceae, Mackinlayoideae, and Pittosporaceae represent more than 200 extant endemic species in Australia. New Caledonia likewise has more than 130 endemic species from the same four groups (Morat, 1993), and many of these represent early-diverging lineages in their respective clades (i.e., paleo-endemics). New Caledonia separated from the Australian continent in the Late Cretaceous, but maintained links to Australia through the Miocene. Between the Cretaceous and the Oligocene, major geological events helped to shape the distribution and diversification of the flora on the island (see Lowry, 1998; Murienne et al., 2005). The exceptionally high levels of species diversity and endemism in New Caledonia may be explained by the persistence of refugia that helped to preserve "relictual lineages" during the aridification of Australia and the submergence of other island archipelagoes during the Oligocene (Chazeau, 1993; Jaffre, 1993; Lowry, 1998). Considering the geologic evidence and the age of Apiineae, Australia appears to be its most likely center of origin, with New Caledonia serving as a refugium where some ancient relicts survived and then later diversified, leading to a proliferation of neo-endemic species. The island's flora shares the most connections to that of Australia (Morat, 1993; Morat et al., 2001), but has undergone less alteration due to fewer climatic changes (Stevenson & Hope, 2005). Inhabitation by humans was also delayed in New Caledonia relative to other landmasses (c. 3,000 years, compared to Australia, which was inhabited by humans since c. 50,000 years), saving the island's natural habitats from detrimental anthropogenic effects (the "blitzkrieg" hypothesis) and fires that annihilated many forests elsewhere (Miller et al., 2005; Trueman et al., 2005).
Dispersals to Madagascar (through the Indian Ocean), to Asia (through Malesia), to North America (through Asia and Europe), and into South America (either through Antarctica or Asia) provide possible explanations for the geographic histories of Pittosporaceae and Araliaceae to these continents. An Australasian (especially New Caledonian and Australian) distribution is retained in the three clades diverging next in the evolutionary history of Apiineae (i.e., Myodocarpaceae, Mackinlayoideae, and the Platysace clade).
Apiaceae probably originated in Australia, where Mackinlayoideae and Platysace are well represented. Origins out of Australia can be explained through vicariance events that led to the separation of Antarctica from other land masses, which commenced in the Late Cretaceous. This would help to explain the diversifications in the Austral realm (Morrone, 2009) of southwestern Australia (Platysace), southern South America (Azorelloideae) and South Africa (early diverging lineages of Apioideae). An alternative explanation would be dispersal from Australia to South America and thence to Africa, followed by isolation and diversifications occurring in parallel in each of these regions. Africa and South America started to separate more than 120 Ma, but contact was possible until c. 80 Ma (Raven & Axelrod, 1974). Plant dispersals between Africa and South America have been recorded as late as the Eocene (Sanmartin & Ronquist, 2004), which is later than the age of divergence of the Azorelloideae and Apioideae-Saniculoideae clades. Diversification in South America was followed by rapid diversification throughout the Andes, and dispersal to the Sub-Antarctic Islands, New Zealand and Australia. The Cape region of southern Africa was the likely center for the third radiation, in Apiaceae, from where subsequent diversifications occurred throughout Africa and then to the North temperate regions.
We have also presented for the first time (and with a nearly comprehensive sampling of genera from throughout the order) a detailed estimate of divergence times for all the major clades of Apiales. Nevertheless, it will be necessary to test these estimates using both additional markers and calibration points. While the interpretation of biogeographic patterns at this level of breadth and depth is complicated, our study provides an historical-biogeographic framework for all the main lineages of Apiales that can be tested in future studies based on detailed work with additional markers (such as nuclear and mitochondrial sequences) and a focus within individual lineages of the order.
DOI 10.1007/s 12229-014-9132-4
Acknowledgments The authors thank the following people and institutions for providing assistance in obtaining plant samples: P. P. Lowry II, G. T. Chandler, P. Goldblatt, P. B. Phillipson, [dagger] L. Constance, J.-P. Reduron, J. Wen, R. J. Bayer, C. Gemmill, A. D. Mitchell, B.-E. van Wyk, P. M. Tilney, A. R. Magee, P. C. Zietsman, Q.-Y. Xiang, G. E. Schatz, D. A. Neill, W. Takeuchi, G. Keppel, B. Gray, R. Jensen, L. W. Cayzer, I. R. H. Telford, L. Hufford, M. E. Mort, D. M. E. Ware, P. Fiaschi, D. Lorence, D. K. Harder, M. O. Dillon, L. A. Johnson, and the Missouri Botanical Garden (MO), Museum National d'Histoire Naturelle (P), United States National Herbarium (US), University of Waikato (WAIK), Australian National Herbarium (CANB), Royal Botanic Gardens Kew (K), New York Botanical Garden (NY), Huntington Botanical Garden (HNT), National Tropical Botanical Garden (PTBG), University of California Botanical Garden (UC), Bloemfontein Museum (BLFU), Parc Zoologique et Botanique de la Ville de Mulhouse, Bogor Botanical Garden, South Pacific Regional Herbarium (SUVA), Universidade de Sao Paulo (SPF), CSIRO-Atherton, and Washington State University (WS). Assistance was also provided by the Integrated Life Sciences Program of Virginia Commonwealth University. Support for field and laboratory work was provided by the National Science Foundation (DEB 0949819 and 0613728/0943958) and the National Geographic Society (CRE 8355-07).
Axelrod, A. J. 1952. A theory of angiosperm evolution. Evolution 6: 29-60.
Barlow, B. A. 1981. The Australian flora: its origin and evolution. In: A. S. George (ed). Flora of Australia, 1: 25-76. Australian Government Publishing Service, Canberra.
Bell, C. D., D. E. Soltis & P. S. Soltis. 2005. The age of the angiosperms: a molecular timescale without a clock. Evolution 59: 1245-1258.
--& M. J. Donoghue. 2005. Dating the diversification of Dipsacales: comparing models, genes, and evolutionary implications. American Journal of Botany 92: 284-296.
Bessey, C. E. 1897. Phytogeny and taxonomy of the angiosperms. Botanical Gazette 24: 145-178.
Bohlmann, F. 1971. Acetylenic compounds in the Umbelliferae. Botanical Journal of the Linnean Society 64(suppl): 279-291.
Bremer, K. 2000. Early Cretaceous lineages of monocots flowering plants. Proceedings of the National Academy of Sciences, USA 97: 4707-4711.
--, E. Friis & B. Bremen 2004. Molecular phylogenetic dating of asterid flowering plants shows early Cretaceous diversification. Systematic Biology 53: 496-505.
Brenner, G. 1996. Evidence for the earliest stage of angiosperm pollen evolution: a paleoequatorial section from Israel. Pp 91-115. In: D. W. Taylor & L. J. Hickey (eds). Flowering plant origin, evolution, and phylogeny. Chapman and Hall, New York.
Briggs, J. C. 2003. The biogeographic and tectonic history of India. Journal of Biogeography 30: 381-388.
Brundin, L. Z. 1966. Transantartic relationships and their significance, as evidenced by chitonomid midges with a monograph of the subfamilies Podonominae and Aphroteninae and the austral Heptagyiae. Kungliga Svenska Vetenskapsakadamiens Handlingar 11: 1-174.
Bryan, C. L., B. D. Clarkson & M. J. Clearwater 2011. Biological flora of New Zealand 12: Griselinia lucida, puka, akapuka, akakopuka, shining broadleaf. New Zealand Journal of Botany 49: 461-79.
Burrows, C. 1995. Germination behaviour of seeds of the New Zealand species Fuchsia excorticata, Griselinia liltoralis, Macmpiper excelsum, and Melicytus ramijlorus. New Zealand Journal of Botany 33: 131-140.
--1999. Germination behaviour of seeds of the New Zealand woody species Beilschmiedia tawa, Dysoxylum spectabile, Griselinia lucida, and Weinmannia racemosa. New Zealand Journal of Botany 37: 95-105.
Carlquist, S. 1974. Island biology. Columbia Univ, Press, New York.
--1981. Chance dispersal. American Scientist 69: 509-516.
Chandler, G. T. & G. M. Plunkett 2004. Evolution in Apiales: nuclear and chloroplast markers together in (almost) perfect harmony. Botanical Journal of the Linnean Society 144: 123-147.
--, --, S. M. Pinney, L. Cayzer & C. Gemmill. 2007. Molecular and morphological agreement in Pittosporaceae: phylogenetic analysis using nuclear ITS and plastid trnL-trnF sequence data. Australian Systematic Botany 20: 390-401.
Chaw, S.-M., C.-. C. Chang, H.-L. Chen & W.-. H. Li. 2004. Dating the monocot-dicot divergence and the origin of core eudicots using whole chloroplast genomes. Journal of Molecular Evolution 58: 424-441.
Chazeau, J. 1993. Research on New Caledonian terrestrial fauna: achievements and prospects. Biodiversity Letters 1: 123-129.
Cox, C. B. & P. D. Moore. 2010. Biogeography: An ecological and evolutionary approach. Wiley, Hoboken.
Darlington, C. D. & A. P. Wylie. 1955. Chromosome Atlas of flowering plants. George Allen & Unwin, London.
De Kok, R. 2002. Are plant adaptations to growing on serpentine soil rare or common? A few case studies from New Caledonia. Adansonia 24: 229-238.
Diamond, J. M. 1984. "Normal" Extinctions of Isolated Populations. Pp 191-246. In: M. Nitecki (ed). Extinctions. University of Chicago Press, Chicago.
Dilcher, D. L. & G. E. Dolph. 1970. Fossil leaves of Dendmpanax from Eocene sediments of southeastern North America. American Journal of Botany 57: 153-160.
Dillon, M. O. & M. Munoz-Schick. 1993. A revision of the dioecious genus Griselinia (Griseliniaceae), including a new species from the coastal Atacama Desert of northern Chile. Brittonia 45: 261-274.
Doyle, J. A. & E K. Endress. 2010. Integrating Early Cretaceous fossils into the phylogeny of living angiosperms: Magnoliidae and eudicots. Journal of Systematics and Evolution 48: 1-35.
Drummond, A. J., S. Y. W. Ho, M. J. Phillips & A. Rambaut. 2006. Relaxed phylogenetics and dating with confidence. PLoS Biology 4: e88.
--& A. Rambaut. 2007. BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evolutionary Biology 7: 214.
Erbar, C. & P. Leins. 2010. Nectaries in Apiales and related groups. Plant Diversity and Evolution 128: 269-295.
Farabee, M. J. 1993. Morphology of triprojectate fossil pollen: form and distribution in space and time. The Botanical Review 59: 211-249.
Friis, E. M., K. R. Pedersen & R R. Crane. 1999. Early angiosperm diversification: the diversity of pollen associated with angiosperm reproductive structures in early Cretaceous floras from Portugal. Annals of the Missouri Botanical Garden 86: 259-296.
Gaina, C., D. R. Muller, J.-Y. Royet; J. Stock, J. Hardebeck & P. Symonds. 1998. The tectonic history of the Tasman Sea: a puzzle with 13 pieces. Journal of Geophysical Research 103: 12, 413-12, 433.
Gemmill, C. E. C., G. J. Allan, W L. Wagner & E. A. Zimmei: 2002. Evolution of insular Pacific Pittosporum (Pittosporaceae): origin of the Hawaiian radiation. Molecular phylogenetics and evolution 22:31-42.
Gentry, A. H. 1982. Neotropical floristic diversity: phytogeographic connections between Central and South America, Pleistocene climatic fluctuation, or an accident of the Andean orogeny. Annals of the Missouri Botanical Garden 69: 557-593.
Gernhard, T. 2008. The conditioned reconstructed process. Journal of Theoretical Biology 253: 769-778.
Grandcolas, P., J. Murienne, T. Robillard, L. Desutter-Grandcolas, H. Jourdan, E. Guilbert & L. Deharveng. 2008. New Caledonia: a very old Darwinian island? Philosophical Transactions of the Royal Society B: Biological Sciences 363: 3309-3317.
Gruas-Cavagnetto, C. & M.-T. Cerceau-Larrival. 1984. Apport des pollens fossiles d'ombelliferes a la connaissance paleoecologique et paleoclimatique de l'Eocene ftanqais. Review of Palaeobotany and Palynology 40: 317-345.
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. Journal of Biogeography 35: 2153-2175.
--2010. Biogeographical affinities of the New Caledonian biota: a puzzle with 24 pieces. Journal of Biogeography 37: 1179-1201.
Hegnauer, R. 1971. Chemical patterns and relationships of Umbelliferae. Pages 267-277 in: V. H. Heywood (ed.). The biology and chemistry of the Umbelliferae. Botanical Journal of the Linnean Society Supplement Vol 64.
Herendeen, P. S., S. Magallon-Puebla, R. Lupia, P. R. Crane & J. Kobylinska. 1999. A preliminary conspectus of the Allon flora from the Late Cretaceous (late Santonian) of central Georgia, U.S.A. Annals of the Missouri Botanical Garden 86: 407-471.
Jaffre, T. 1993. Relation between ecological diversity and floristic diversity in New Caledonia. Biodiversity Letters 1: 82-91.
Jay, M. 1969. Chemotaxonomic researches on vascular plants XIX. Flavonoid distribution in the Pittosporaceae. Botanical Journal of the Linnean Society 62: 423-429.
Jolivet, R & K. K. Verma. 2010. Good morning Gondwana. Annales de la Societe entomologique de France (N.S.): International. Journal of Entomology 46: 53-61.
Jurica, H. S. 1922. A morphological study of the Umbelliferae. Botanical Gazette 74: 292-307.
Karehed, J. 2003. The family Pennantiaceae and its relationships to Apiales. Botanical Journal of the Linnean Society 141: 1-24.
Knobloch, E. & D. H. Mai. 1986. Monographie der Fruchte und Samen in der Kreide von Mitteleuropa. Rozpravy ustredniho ustavu geologickenho, Praha 47: 1-219.
Kula, J., A. Tulloch, T. L. Spell & M. L. Wells. 2007. Two-stage rifting of Zealandia-Australia-Antarctica: Evidence from [sup.40]Ar/[sup.39]Ar thermochronometiy of the Sisters Shear Zone; Stewart Island, New Zealand. Geology 35: 411-414.
Ladiges, P. Y., E Udovicic & G. Nelson. 2003. Australian biogeographical connections and the phylogeny of large genera in the plant family Myrtaceae. Journal of Biogeography 30: 989-998.
--& D. Cantrill. 2007. New Caledonia-Australian connections: biogeographic patterns and geology. Australian Sysematic Botany 20: 383-389.
Landis, C. A., H. J. Campbell, J. G. Begg, D. C. Mildenhall, A. M. Paterson & S. A. Trewick. 2008. The Waipounamu erosion surface: questioning the antiquity of the New Zealand land surface and terrestrial fauna and flora. Geological Magazine 145: 173-197.
Lawver, L. A. & L. M. Gahagan. 2003. Evolution of Cenozoic seaways in the circum-Antarctic region. Palaeogeography, Palaeoclimatology, Palaeoecology 198: 11-37.
Linder, H. P. & M. D. Crisp. 1996. Nothofagus and Pacific biogeography. Cladistics 11: 5-32.
Lowry, P. P. 1. I. 1986. A systematic study of Delarbrea (Araliaceae). Allertonia 4: 169-201.
Lowry, P. R II. 1991. Evolutionary patterns in the flora and vegetation of New Caledonia. Pages 373-379. in The unity of evolutionary biology. Proceedings of the Fourth International Congress of Systematic and Evolutionary Biology. Dioscorides Press, Portland, Oregon.
Lowry, P. P., II. 1998. Diversity, endemism, and extinction in the flora of New Caledonia: a review. Pp 181-206. In: C-. I. Peng & P. P. Lowry II (eds). Rare, threatened, and endangered floras of the Pacific Rim. Institute of Botany, Academica Sinica, Monogr, Taipei. Ser. No. 16.
Magallon-Puebla, S., P. R. Crane & P. S. Herendeen. 1999. Phylogenetic pattern, diversity, and diversification of eudicots. Annals of the Missouri Botanical Garden 86: 297-372.
Magallon, S. & M. J. Sanderson. 2001. Absolute diversification rates in angiosperm clades. Evolution 55: 1762-1780.
Magee, A. R., C. I. Calvino, M. Liu, S. R. Downie, P. M. Tilney & B.-E. van Wyk. 2010. New tribal delimitations for the early diverging lineages of Apiaceae subfamily Apioidae. Taxon 59: 567-580.
Manchester, S. R. 1994. Fruits and seeds of the middle Eocene Nut Beds Flora, Clamo Formation, Oregon. Paleontographica Americana 58: 38-39.
--1999. Biogeographical relationships of North American Tertiary floras. Annals of the Missouri Botanical Garden 86: 472-522.
--, Z.-D. Chen, A.-M. Lu & K. Uemura. 2009. Eastern Asian endemic seed plant genera and their paleogeographic history throughout the Northern Hemisphere. Journal of Systematics and Evolution 47: 1-42.
Martinez-Millan, M. 2010. Fossil records and age of the Asteridae. The Botanical Review 76: 83-135.
Mathias, M. E. 1965. Distribution patterns of certain Umbelliferae. Annals of the Missouri Botanical Garden 52: 387-398.
Mayer, M. S. & R S. Soltis. 1994. The evolution of serpentine endemics: a cpDNA phylogeny of Streptanthus glandulosus complex (Cruciferae). Systematic Botany 19: 557-574.
McLoughlin, S. 2001. The breakup history of Gondwana and its impact on pre-Cenozoic floristic provincialism. Australian Journal of Botany 49: 271-300.
McPhail, M. K. 1997. The New Zealand flora: entirely long-distance dispersal? Journal of Biogeography 24: 113-117.
Megard, E 1984. The Andean orogenic period and its major structures in central and northern Peru. Journal of the Geological Society Geological Society of London 141: 893-900.
Mildenhall, D. C. 1980. New Zealand Late Cretaceous and Cenozoic plant biogeography: a contribution. Palaeogeography, Paleoclimatology, Palaeoecology 31: 197-233.
Miller, G. H., M. L. Fogei, J. W. Magee, M. K. Gagan, S. J. Clarke & B. J. Johnson. 2005. Ecosystem collapse in Pleistocene Australia and a human role in megafaunal extinction. Science 309: 287-290.
Mitchell, A. & J. Wen. 2004. Phylogenetic utility and evidence for multiple copies of granule-bound starch synthase I (GBSSI) in the Araliaceae. Taxon 53: 29-41.
Morat, P. H. 1993. Our knowledge of the flora of New Caledonia: endemism and diversity in relation to vegetation types and substrates. Biodiversity Letters 1: 72-81.
Morat, P., T. Jaffre & J. M. Veillon. 2001. The flora of New Caledonia's Calcareous substrate: tolerance or dependence. Adansonia 23: 1-19.
Morley, R. J. 2003. Interplate dispersal paths for megathermal angiosperms. Perspectives in Plant Ecology, Evolution and Systematics 6: 5-20.
Morrone, J. J. 2009. Evolutionary biogeography: An integrative approach with case studies. Columbia University Press, New York.
Murienne, J., P. Grandcolas, M. D. Piulachs, X. Belles, C. D'Haese, E Legendre, R. Peilens & E. Guilbert. 2005. Evolution on a shaky piece of Gondwana: is local endemism recent in New Caledonia. Cladistics 21: 2-7.
Nelson, G. 1975. Historical biogeography: an alternative formalization. Systematic Zoology 24: 555-558.
Nicolas, A. N. 2009. Understanding evolutionary relationships within the angiosperm order Apiales based on analyses of organeliar DNA sequences and nuclear gene duplications. Ph.D. Dissertation. Virginia Commonwealth University, Virginia.
--& G. M. Plunkett. 2009. The demise of subfamily Hydrocotyloideae (Apiaceae) and the realignment of its genera across the whole order Apiales. Molecular Phylogenetics and Evolution 53: 134-151.
-- & --. 2012. Untangling Azorella, Laretia, and Mulinum (Apiaceae, Azorelloideae): insights from phylogenetics and biogeography. Taxon 61: 824-840.
Oskolski, A. A. & R P. Lowry II. 2000. Wood anatomy of Mackinlaya and Apiopelalum (Araliaceae) and its systematic implications. Annals of the Missouri Botanical Garden 87: 171-182.
Pielou, E. C. 1979. Biogeography. Wiley, New York.
Plunkett, G. M., D. E. Soltis & P. S. Soltis. 1996a. Higher level relationships of Apiales (Apiaceae and Araliaceae) based on phylogenetic analysis of rbcL sequences. American Journal of Botany 83: 499-515.
--, -- & --. 1996b. Evolutionary patterns in Apiaceae: inferences based on matK sequence data. Systematic Botany 21: 477-495.
--, -- & --. 1997. Clarification of the relationship between Apiaceae and Araliaceae based on matK and rbcL sequence data. American Journal of Botany 84: 565-580.
--& P. P. Lowry II. 2001. Relationships among 'ancient araliads' and their significance for the systematics of Apiales. Molecular Phylogenetics and Evolution 19: 259-276.
--, P. P. Lowry II, & M. K. Burke. 2001. The phylogenetic status of Polyscias (Araliaceae) based on nuclear ITS sequence data. Annals of the Missouri Botanical Garden 88: 213-230.
--, G. T. Chandler, R P. Lowry II, S. M. Pinney, & T. S. Sprenkle. 2004a. Recent advances in understanding Apiales and a revised classification. South African Journal of Botany 70: 371-381.
--, P. P. Lowry II, & N. V Mi. 2004b. Phylogenetic relationships among Polyscias (Araliaceae) and close relatives from the Indian Ocean basin. International Journal of Plant Sciences 165: 861-873.
--, J. Wen, & P. P. Lowry II. 2004c. Infrafamilial relationships in Araliaceae: insights from plastid (trnL-trnF) and nuclear (ITS) sequence data. Plant Systematics and Evolution 245: 1-39.
--& P. P. Lowry II. 2010. Paraphyly and polyphyly in Polyscias sensu lato: molecular evidence and the case for recircumscribing the "pinnate genera" of Araliaceae. Plant Diversity and Evolution 128: 23-54.
Pole, M. S. 1994. The New Zealand flora--entirely long-distance dispersal? Journal of Biogeography 21: 625-635.
--2001. Can long-distance dispersal be inferred from the New Zealand plant fossil record? Australian Journal of Botany 49: 357-366.
Pole, M. 2008. Dispersed leaf cuticle from the early Miocene of southern New Zealand. Palaeontologia Electrionica 11.3(15A): 1-117.
Posada, D. 2008. jModelTest: phylogenetic model averaging. Molecular Biology and Evolution 25: 1253-1256.
Rambaut, A. & A. J. Drummond. 2007. Tracer v. 1.4: MCMC trace analyses tool. Available from: http://beast.bio.ed.ac.ukyTracer
Ravala, U. & K. Veeraswamya. 2003. India-Madagascar separation: breakup along a pre-existing mobile belt and chipping of the craton. Gondwana Research 6: 467-485.
Raven, P. H. & D. I. Axelrod. 1972. Plate tectonics and Australasian paleobiogeography. Science 176: 1379-1386.
-- & --. 1974. Angiosperm biogeography and past continental movements. Annals of the Missouri Botanical Garden 61: 539-673.
Ree, R. H., B. R. Moore, C. O. Webb & M. J. Donoghue. 2005. A likelihood framework for inferring the evolution of geographic range on phylogenetic trees. Evolution 59: 2299-2311.
--& S. A. Smith. 2008. Maximum likelihood inference of geographic range evolution by dispersal, local extinction, and cladogenesis. Systematic Biology 57: 4-14.
Rennet; S. S. 2004. Multiple Miocene Melastomataceae dispersal between Madagascar, Africa, and India. Philosophical Transactions of the Royal Society B: Biological Sciences 359: 1485-1494.
Rodriguez, R. L. 1971. The relationships of the Umbellales. Pages 63-91 in: V. H. Heywood (ed.), The biology and chemistry of the Umbelliferae. Botanical Journal of the Linnean Society Supplement Vol 64.
Ronquist, F. 1996. DIVA version 1.1. Computer program and manual available by anonymous FTP from Uppsala University (ftp.uu.se or ftp.systbot.uu.se).
--1997. Dispersal-vicariance analysis: a new approach to the quantification of historical biogeography. Systematic Biology 46: 195-203.
--& J. P. Huelsenbeck. 2003. MRBAYES 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19: 1572-1574.
Rutschmann, F. 2006. Molecular dating of phylogenetic trees: a brief review of current methods that estimate divergence times. Diversity and Distributions 12: 35-48.
Sanderson, M. J. 1997. A nonparametric approach to estimating divergence times in the absence of rate constancy. Molecular Biology and Evolution 14: 1218-1231.
--& J. A. Doyle. 2001. Sources of error and confidence intervals in estimating the age of angiosperms from rbcL and 18S rDNA data. American Journal of Botany 88: 1499-1516.
Sanmartin, I. & F. Ronquist. 2004. Southern Hemisphere biogeography inferred by event-based models: plant versus animal patterns. Systematic Biology 53: 216-243.
Schatz, G. E. 1996. Malagasy/Indo-Australo-Malesian phytogeographic connections. Pp 73-83. In: W. R. Lourenco (ed). Biogeography of Madagascar. Editions ORSTOM, Paris.
Schneider, J., E. Schuettpelz, K. M. Pryer, R. Cranfill, S. Magallon & R. Lupia. 2004. Ferns diversified in the shadow of angiosperms. Nature 428: 553-557.
Schwarz, M. P., S. Fullec S. M. Tierney & S. J. B. Cooper. 2006. Molecular phylogenetics of the exoneurine allodapine bees reveal an ancient and puzzling dispersal from Africa to Australia. Systematic Biology 55: 31-45.
Sclatec P. L. 1858. On the general geographical distribution of the members of the class Aves. Zoological Journal of the Linnean Society 2: 130-145.
Seemann, B. 1863. On the position of the genera Hydrocotyle, Opa, Commia, and Blastus in the natural system. The Journal of Botany, British and Foreign 1: 278-282.
Shields, O. 1991. Pacific biogeography and rapid earth expansion. Journal of Biogeography 18: 583-585.
Shoup, J. R. & C. C. Tseng. 1977. Pollen of Klotzschia (Umbelliferae): a possible link to Araliaceae. American Journal of Botany 64: 461-463.
Spalik, K., M. Piwczynski, C. A. Danderson, R. Kurzyna-Mlynik, T. S. Bone & S. R. Downie. 2010. Amphitropic amphiantarctic disjunctions in Apiaceae subfamily Apioideae. Journal of Biogeography 37: 1977-1994.
Stevenson, J. & G. Hope. 2005. A comparison of late Quaternary forest changes in New Caledonia and northeastern Australia. Quaternary Research 64: 372-383.
Stockler, K., I. L. Daniel & P. J. Lockhart. 2002. New Zealand kauri (Agathis australis (D.Don) Lindl., Araucariaceae) survives Oligocene drowning. Systematic Biology 51: 827-832.
Swenson, U., A. Backlund, S. McLoughlin & R. S. Hill. 2001. Nothofagus biogeography revisited with special emphasis on the enigmatic distribution of subgenus Brassospora in New Caledonia. Cladistics 17: 28-47.
Takhtajan, A. 1986. Floristic regions of the world. T.J. Crovello (translator). University of California Press, Berkeley.
Thewick, S. A., A. M. Paterson & H. J. Campbell. 2007. Hello New Zealand. Journal of Biogeography 34: 1-6.
Trueman, C. N. G., J. H. Field, J. Dortch, B. Charles & S. Wroe. 2005. Prolonged coexistence of humans and megafauna in Pleistocene Australia. PNAS 102: 8381-8385.
Wallace, A. R. 1876. The geographic distribution of animals, 2 vols. Harper, New York.
Wen, J., G. M. Plunkett, A. D. Mitchell & S. J. Wagstaff. 2001. The evolution of Araliaceae: a phylogenetic analysis based on ITS sequences of nuclear ribosomal DNA. Systematic Botany 26: 144-167.
Wikstrom, N., V Savolainen & M. W. Chase. 2001. Evolution of the angiosperms: calibrating the family tree. Proceedings of the Royal Society B: Biological Sciences 268: 2211-2220.
Zwickl, D. J. 2006. Genetic algorithm approaches for the phylogenetic analysis of large biological sequence datasets under the maximum likelihood criterion. Ph.D. Dissertation, The University of Texas at Austin, Texas.
Antoine N. Nicolas (1,2) * Gregory M. Plunkett (1)
(1) Cullman Program for Molecular Systematics, The New York Botanical Garden, 2900 Southern Boulevard, Bronx, NY 10458-5126, USA
(2) Author for Correspondence; e-mail: email@example.com
Published online: 26 February 2014
Electronic supplementary material The online version of this article (doi: 10.1007/sl 2229-014-9132-4) contains supplementary material, which is available to authorized users.
Table 1 Major lineages of Apiales, their relative sizes (numbers of genera and species), and native geographic distributions provided as letter codes (following Fig. 1), followed by a more detailed description of the lineages' geographies Genera Species Early diverging families Pennantiaceae 1 4 Torricelliaceae 3 9 Griseliniaceae 1 7 Suborder Apiineae Pirtosporaceae 9 250 Araliaceae 37 1900 Hydrocotyle & 2 236 Trachymene Harmsiopanax 1 3 African Schefflera 1 45 Raukaua et al. 5 23 Aralia et al. 2 92 Asian Palmate clade 1 >250 Cussonia et al. 2 21 Polyscias 4 314 Pseudopanax Myodocarpaceae 2 17 Myodocarpus 1 10 Delarbrea 1 7 Apiaceae >450 >3800 Mackinlayoideae 11 c. 100 Platysace 1 27 Azorelloideae 21 126 Bowlesia clade 5 25 Diposis 1 3 Asteriscium clade 6 32 Spananthe 1 1 Azorella clade 8 65 Klotzschia 1 3 Apioideae + >425 > 3500 Saniculoideae Hermas 1 9 Saniculoideae 11 340 Lichtensteinia clade 3 9 Annesorhiza et al. 4 34 Hetemmorpha et al. 4 24 Bupleurum 1 c. 190 Rest of Apioideae >400 c. 2900 Geographic distribution Code Details of distribution Early diverging families Pennantiaceae CEF Australasia (Australia & New Zealand) Torricelliaceae CDE Ethiopian (Madagascar), Oriental (Malesia), Pale-arctic (central Asia) Griseliniaceae BF Australasia (New Zealand), Neotropical (Andean South America) Suborder Apiineae Pirtosporaceae EF Australasia (Australia, New Zealand, Oceania), Oriental, Ethiopian (Africa, Madagascar) Araliaceae ABCDEF Cosmopolitan (esp. tropics) Hydrocotyle & ABCDEF Cosmopolitan (esp. Andean South Trachymene America and Australia) Harmsiopanax F Australasia (eastern Malesia) African Schefflera D Ethiopian (Africa, Madagascar, Mascarenes, Seychelles) Raukaua et al. BF Australasia (Australia, New Zealand, Oceania), Neotropical (temperate South America) Aralia et al. BCE Nearctic (North America), Neotropical (Central and northern South America), Oriental (Asia) Asian Palmate clade EF Oriental (eastern and southeastern Asia), Australasian (Malesia, Australia, Oceania) Cussonia et al. D Ethiopian (southern Africa, Comoro Is., SW Arabia) Polyscias DF Ethiopian (Africa, Madagascar); Pseudopanax Australasian (Oceania, New Zealand) Myodocarpaceae D Australasia (Oceania, esp. New Caledonia) Myodocarpus D Australasia (New Caledonia) Delarbrea D Australasia (New Caledonia, Australia, Oceania, Malesia) Apiaceae ABCDEF Cosmopolitan (esp. temperate regions) Mackinlayoideae DBF Neotropical (Central and South America), Ethiopian (Africa), Australasia (Australia, New Zealand, Oceania, Malesia) Platysace F Australasia (Australia) Azorelloideae BDEF Neotropical (Central and South America), Ethiopian (Canary Is, Morocco, Somalia), Oriental (S China), Australasia (Australia, New Zealand) Bowlesia clade BDF Neotropical (Central and South America), Ethiopian (Canary Is, Morocco, Somalia), Australasia (Australia) Diposis B Neotropical (South America) Asteriscium clade BF Neotropical (South America), Australasia (Australia) Spananthe B Neotropical (Central and South America) Azorella clade BEF Neotropical (Central and South America), Oriental (S China), Australasia (Australia, New Zealand) Klotzschia Neotropical (Central and South America), Australasia (Oceania) Apioideae + ABCDEF Cosmopolitan Saniculoideae Hermas D Ethiopian (southern Africa) Saniculoideae ABCDEF Nearctic (temperate North America), Neotropical (Central and South America), Palearctic (Europe, SW and E Asia), Ethiopian (esp. southern Africa), Oriental (SE Asia), Australasia (Australia, New Zealand, Oceania) Lichtensteinia clade D Ethiopian (southern Africa) Annesorhiza et al. D Ethiopian (southern Africa, Canary Is., Morocco) Hetemmorpha et al. D Ethiopian (southern Africa, northeastern Africa, Madagascar, Yemen) Bupleurum ACD Nearctic (North America), Palearctic (Europe, Asia, northern Africa), Ethiopian (southern Africa) Rest of Apioideae ABCDEF Cosmopolitan Table 2 Timing of the diversification events of the major clades of Apiales as estimated with BEAST using the uncorrelated lognormal relaxed molecular clock (UCLD) model Node PP MMA [with 95 % HPD] Pennantiaceae 1 117.04 [103.73,129.97] Torricelliaceae 1 106.86 [94.02,119.83] Griseliniaceae 1 103.06 [90.20,115.84] Pittosporaceae 1 97.73 [85.15,110.20] Araliaceae 1 94.76 [82.56,106.80] Hydrocotyle + Trachmene 1 65.03 [48.70,83.19] African Schefflera 1 48.09 [34.32,64.33] Raukaua et al. 1 37.65 [28.58,46.93] Asian Palmate 1 33.71 [24.88,39.36] Polyscias + Pseudopanax 1 16.76 [9.49, 25.14] Myodocarpaceae 1 91.33 [78.63,103.22] Myodocarpus from Delarbrea 1 25.36 [7.96,47.35] Mackinlayoideae 0.97 87.36 [75.99,98.90] Apiopetaulm + Actinotus 1 66.28 [50.16,83.55] Mackinlaya 1 57.49 [41.98,74.27] Rest of Mackinlayoideae 1 47.28 [32.80; 62.68] Platysace Clade 0.84 83.57 [72.94; 94.79] Azorelloideae 1 75.78 [65.75,85.60] Hermas 0.79 70.59 [61.98,80.04] Saniculoideae 1 65.78 [58.21,74.31] Lichtensteinia + Choritaenia clade 0.61 63.66 [56.08,71.61] Rest of Apioideae 1 58.32 [52.25,65.27] MMA mean minimum ages; HPD highest posterior density intervals Table 3 Areas of origin for the major clades of Apiales as estimated by DIVA and Lagrange (under three different models, each with probabilities for alternative reconstructions). The complete output resulting from the Lagrange analyses is provided as Appendix S2 Node DIVA Lagrange Model 1 Model 2 Model 3 Pennantiaceae F, EF EF|F; 0.34 EF|F; 0.36 EF|F; 0.32 F|F; 0.32 F|F; 0.26 F|F; 0.26 DF|F; 0.08 EF|E; 0.07 DF|F; 0.10 Totricelliaceae EF F|E; 0.42 F|E; 0.56 F|E; 0.52 F|F; 0.31 F|F; 0.18 F|D; 0.17 F|D; 0.10 F|D; 0.10 F|F; 0.16 Griseliniaceae F F|F; 0.94 F|F; 0.93 F|F; 0.92 Pittosporaceae F F|F; 0.98 F|F; 0.98 F|F; 0.97 Araliaceae F F|F; 0.99 F]F; 0.98 F|F; 0.97 Myodocarpaceae F F|F; 0.97 F|F; 0.95 F|F; 0.94 Mackinlayoideae F F|F; 0.84 F|F; 0.78 F|F; 0.73 Plalysace BF B|F; 0.37 F|F; 0.47 F|F; 0.48 F|F; 0.26 B|F; 0.27 B|F; 0.32 D|F; 0.14 D|F; 0.11 B|BF; 0.07 B|BF; 0.12 B|BF; 0.07 D|F; 0.05 Azorelloideae B B|B; 0.37 B|B; 0.29 B|B; 0.32 BD|B; 0.33 BD|B; 0.24 BD|B; 0.21 B|BF; 0.09 B|BF; 0.13 B|BF; 0.14 Klotzschia BD D|B; 0.42 D|B; 0.31 D|B; 0.33 B|B; 0.22 B|B; 0.21 B|B; 0.22 D|BD; 0.11 D|D; 0.13 D|BD; 0.11 Hermas D D|D; 0.69 D|D; 0.64 D|D; 0.63
|Printer friendly Cite/link Email Feedback|
|Author:||Nicolas, Antoine N.; Plunkett, Gregory M.|
|Publication:||The Botanical Review|
|Date:||Mar 1, 2014|
|Previous Article:||Stigma diversity in tropical legumes with considerations on stigma classification.|
|Next Article:||Epidermal patterning and silica phytoliths in grasses: an evolutionary history.|