Biological Journal of the Linnean Society, 2016, 119, 511–522. With 4 figures. Evolutionary history of Gymnocarpos (Caryophyllaceae) in the arid regions from North Africa to Central Asia SHU-WEN JIA1,2, MING-LI ZHANG1,3*, ECKHARD V. RAAB-STRAUBE4 and MATS THULIN5 Key Laboratory of Biogeography and Bioresource in Arid Land, Xinjiang Institute of Ecology and Geography, Chinese Academy of Sciences, Urumqi, 830011, China 2 Graduate University of the Chinese Academy of Sciences, Beijing, 100049, China 3 Institute of Botany, Chinese Academy of Sciences, Beijing, 100093, China 4 € Berlin, K€ Botanischer Garten und Botanisches Museum Berlin-Dahlem, Freie Universitat oniginLuise-Str. 6–8, 14195, Berlin, Germany 5 € Department of Systematic Biology, EBC, Uppsala University, Norbyvagen 18D, SE-752 36, Uppsala, Sweden Received 4 February 2016; revised 7 April 2016; accepted for publication 7 April 2016 Gymnocarpos has only about ten species distributed in the arid regions of Asia and Africa, but it exhibits a geographical disjunction between eastern Central Asia and western North Africa and Minor Asia. We sampled eight species of the genus and sequenced two chloroplast regions (rps16 and psbB–psbH), and the nuclear rDNA (ITS) to study the phylogeny and biogeography. The results of the phylogenetic analyses corroborated that Gymnocarpos is monophyletic, in the phylogenetic tree two well supported clades are recognized: clade 1 includes Gymnocarpos sclerocephalus and G. decandrus, mainly the North African group, whereas clade 2 comprises the remaining species, mainly in the Southern Arabian Peninsula. Molecular dating analysis revealed that the divergence age of Gymnocarpos was c. 31.33 Mya near the Eocene and Oligocene transition boundary, the initial diversification within Gymnocarpos dated to c. 6.69 Mya in the late Miocene, and the intraspecific diversification mostly occurred during the Quaternary climate oscillations. Ancestral area reconstruction suggested that the Southern Arabian Peninsula was the ancestral area for Gymnocarpos. Our conclusions revealed that the aridification since mid-late Miocene significantly affected the diversification of the genus in these areas. © 2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, 119, 511–522. KEYWORDS: ancestral area – diversification – historical biogeography. INTRODUCTION In recent decades, much evidence from biogeographic studies has shown that the changes of paleoclimate and paleogeography have profound effects on biological evolution, diversification, and speciation (e.g. Barrientos et al., 2014; Chen et al., 2014; Gao, Meng & Zhang, 2014). Uplift of the Qinghai-Tibet Plateau (QTP) and retreat of the Tethys Ocean from Central Asia led to significant changes in atmospheric circulation patterns (Liu & Yin, 2002). These events not only resulted in the aridification in Central Asia, but also affected climatic conditions in North Africa (Liu & Yin, 2002; Micheels, Eronen & Mosbrugger, 2009). *Corresponding author. E-mail: zhangml@ibcas.ac.cn Mid-Miocene climatic and cryospheric changes in the Antarctic likewise had a great influence on the increase of aridification in Africa (Flower & Kennett, 1994). Subsequently, the Pleistocene glacial–interglacial cycles imposed further influence on aridification in both the African and Asian regions (Demenocal, 2004; Ding et al., 2005; Wu et al., 2007). To investigate the relationship between aridification and evolutionary history of plants, we should chose a group which is distribute widely in northern Africa and Central Asia. The genus Gymnocarpos Forssk. is distributed across arid and semi-arid regions of Africa and Asia, from the Cape Verde and Canary Islands to Northwest China and Mongolia, with a centre of diversity in the Southern Arabian Peninsula and the adjacent islands in the Indian © 2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, 119, 511–522 511 Downloaded from https://academic.oup.com/biolinnean/article/119/2/511/2701016 by guest on 28 July 2021 1 512 S.-W. JIA ET AL. MATERIAL AND METHODS T AXON AND MOLECULAR SAMPLING We included eight of the ten species of the genus in our study. Samples of G. przewalskii were collected at five sites during field surveys in Northwest China from June to August 2013, which covered the major distribution areas of the species. Use of a global positioning system (GPS) placed the geographical coordinates of sample sites. Fresh leaves were dried immediately in silica gel. Samples of G. decandrus were obtained from eight collections from LE (Komarov Botanical Institute, Russian Academy of Sciences, St. Petersburg, Russia; one individual), B (Botanic Garden and Botanical Museum Berlin-Dahlem, Free University of Berlin, Berlin, Germany; four individuals) and MA (Herbarium of the Real Jardin Botanico of Madrid, Madrid, Spain; three individuals). Many samples came from G. przewalskii and G. decandrus as both are the previously published species that are representative of the eastern and western area distribution of the genus (Chaudhri, 1968). G. mahranus, G. rotundifolius, G. kuriensis, G. bracteatus and G. sclerocephalus were obtained from UPS (Department of Systematic Biology, Evolutionary Biology Centre, Uppsala, Sweden). Details of the collection are show in Table 1. For divergence time estimates, we selected a broad range of 30 species of Caryophyllaceae as outgroups (see Supporting Information, Table S2) according to the family phylogenetic scheme (Oxelman et al., 2002; Frajman, Eggens & Oxelman, 2009; Greenberg & Donoghue, 2011). Those eight species of 30 outgroups allied to Gymnocarpos were used to reconstruct the phylogenetic tree. D NA EXTRACTION, AMPLIFICATION AND SEQUENCING Total genomic DNA was extracted following a modified CTAB protocol (Doyle & Doyle, 1987). Polymerase chain reaction (PCR) amplifications of two chloroplast fragments used the primers rps16 (Chen et al., 2011) and psbB–psbH (Xu et al., 2000). The nuclear fragment was amplified using primers for ITS (White et al., 1990). The PCR amplification program followed the protocols: 94 °C for 5 min followed by 28 cycles at 94 °C for 30 s, 52 °C, 52 and 55 °C (respectively, for the ITS, rps16 and psbB–psbH regions) for 30 s, 72 °C for 45 s and a single final step at 72 °C for 1 min. PCR was carried out in a total volume of 30 lL. The PCR mixtures contained 2 lL of 109 PCR reaction buffer, 2 lL of 25 mM MgCl2, 0.9 lL of each primer at 10 ng lL 1, 2 lL of 2.5 mM dNTP solutions in an equimolar ratio, 0.5 lL of Taq DNA-polymerase and 1 lL of genomic DNA at 60 ng lL 1. The PCR products were purified with purification kits (Shanghai SBS, Biotech Ltd, China). Sequences were obtained using ABI PRISM 3700 DNA automatic sequencer (Shanghai Bioengineering, Shanghai, China). To check for consistency, all individuals were sequenced using both forward and reverse primers. Sequences were edited and aligned using the software SeqMan (Lasergene, DNASTAR Inc., Madison, Wisconsin, USA) and CLUSTAL_W (Thompson, Higgins & Gibson, 1994), respectively. Some unphased sequences with double peaks sites were observed in the ITS data, indicating heterozygous genotypes. The unphasable heterozygous sites in the ITS data were coded ambiguously and one-site heterozygotes were directly phased (Huang, Ji & Zhang, 2008; Shi et al., 2013). P HYLOGENETIC ANALYSES AND ESTIMATION OF DIVERGENCE TIME Phylogenetic analyses were performed using Maximum parsimony (MP) and Bayesian inference (BI) methods for ITS, cpDNA, the combined dataset, respectively. The MP analysis was conducted with PAUP* v.4.0b10 (Swofford, 2002), with 1000 replicates of heuristic search using tree-bisection–reconnection (TBR) branch-swapping with ten random sequence additions. BI analysis was conducted using MrBayes version 3.2.3 (Ronquist et al., 2012). The best fit nucleotide substitution models for BI and BEAST analysis were inferred using the © 2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, 119, 511–522 Downloaded from https://academic.oup.com/biolinnean/article/119/2/511/2701016 by guest on 28 July 2021 Ocean (Petrusson & Thulin, 1996). Therefore, it is an ideal model. In the genus, G. decandrus and G. sclerocephalus have wide distributions in North Africa and Asia Minor. All other species such as G. kuriensis, G. bracteatus, G. rotundifolius, G. mahranus, and G. argenteus, are endemic to the Southern Arabian Peninsula and the adjacent islands in the Indian Ocean (Petrusson & Thulin, 1996). G. przewalskii is endemic to Mongolia and Northwest China (Oxelman, Ahlgren & Thulin, 2002). Detailed distribution information on Gymnocarpos is in the Supporting Information (Table S1). Initially, the genus Gymnocarpos consisted of only two species G. decandrus and G. przewalskii (Chaudhri, 1968). Oxelman et al. (2002) recognized ten species in the genus based on both molecular and morphological evidence. The evolutionary history of Gymnocarpos has been poorly understood so far. In this study, we use phylogenetic and biogeographic methods and a wide sampling of the genus to discuss the biogeographical history of the genus, to attempt to clarify its geographic origin and dispersal across Africa and Asia, and to assess the effect of climate aridification on diversification of Gymnocarpos. EVOLUTIONARY HISTORY OF GYMNOCARPOS 513 Table 1. Voucher information for the eight species of Gymnocarpos Source Voucher GenBank accession nos (ITS, rps16, psbB–psbH) G. przewalskii Bunge ex Maxim. /1 G. przewalskii Bunge ex Maxim. /2 G. przewalskii Bunge ex Maxim. /3 G. przewalskii Bunge ex Maxim. /4 G. przewalskii Bunge ex Maxim. /5 Wuqia, Xinjiang, China; 39.68N/75.01E Akesu, Xinjiang, China; 41.55N/81.25E Hami, Xinjiang, China; 42.89N/93.94E Jiayuguan, Gansu, China; 39.79N/98.19E Wulate Rear Banner, Inner Mongolia, China; 41.67N/108.5E Fuerteventura, Canary Islands, Spain Tenerife, Canary Islands, Spain 10 km SE of Erfoud toward Merzouga, Morocco Eastern Antiatlas Mountains, Tagmoute, Morocco envirous.t. Algeria 5 km NE of the Monastery St. Catherine. rocky NE flank of Jabal Berqa facing Wadi Sabayia, Sinai, Egypt; 28.58N/33.98E Wadi Agrav, near the street ca 15,5 km WSW Mitzpe Ramon, Central Negev, Israel; 30.55N/34.65E Perisa, C: Yazd. in saxosis calc. 18 km ENE Chupunum (Chupanan), Iran; 33.55N/54.32E Hadramaut, 25 km W of Al Ridah, Yemen Mahrah, Ras Fartak, Yemen S.W. Jia, 2013045, XJBI KX012966; KX012948; KX012984 S.W. Jia, 2013056, XJBI KX012965; KX012947; KX012983 S.W. Jia, 2013063, XJBI KX012964; KX012946; KX012982 S.W. Jia, 2013068, XJBI KX012963; KX012945; KX012981 S.W. Jia, 2013082, XJBI KX012967; KX012949; KX012985 I. Alvarez et al. 768040, MA J.J. Greuter, 100021572, B C. Aedo. et al. 552332, MA T. Buira & J. Calvo 758240, MA P. Bochantsev, 924, LE P. Hein, 100193002, B KX012974; KX012956; KX012992 KX012969; KX012951; KX012987 KX012971; KX012953; KX012989 M. Ristow, 100355390, B KX012970; KX012952; KX012988 K. H. Rechinger, 100355987, B KX012972; KX012954; KX012990 M. Thulin et al., 9531, UPS M. Thulin et al., 9632, UPS Thulin & Gifri, 8795, UPS Hedberg, 92034, UPS KX012979; KX012961; KX012997 KX012980; KX012962; KX012998 Miller et al., 11384, UPS KX012977; KX012959; KX012995 Thulin, 8330, UPS AJ310973y (ITS) G. decandrus Forssk. /1 G. decandrus Forssk. /2 G. decandrus Forssk. /3 G. decandrus Forssk. /4 G. decandrus Forssk. /5 G. decandrus Forssk. /6 G. decandrus Forssk. /7 G. decandrus Forssk. /8 G. rotundifolius Petruss. & Thulin G. mahranus Petruss. & Thulin G. bracteatus (Balf.f.) Petruss. & Thulin G.sclerocephalus (Decne.) Ahlgren & Thulin G. kuriensis (Radcl.-Sm.) Petruss. & Thulin G. argenteus Petruss. & Thulin Socotra, Muqadrion pass, Yemen 65 km W of Riyadh, Saudi Arabia Abd al-kuri, Jebal Hassala, Yemen Yemen KX012968; KX012950; KX012986 KX012975; KX012957; KX012993 KX012973; KX012955; KX012991 KX012978; KX012960; KX012996 KX012976; KX012958; KX012994 [Corrections added on 10 June 2016, after first online publication: Some of the data in Source and Voucher columns of Table 1 that were previously incorrect have now been corrected]. mrModelTest 2.3 program (Nylander, 2004). GTR + G, GTR + I and GTR + I + G were chosen as the substitution models for ITS, cpDNA and the combined dataset, respectively. The BI analysis ran the chains for 2 000 000 generations and sampling every 1000 generations. We discarded the first 10% © 2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, 119, 511–522 Downloaded from https://academic.oup.com/biolinnean/article/119/2/511/2701016 by guest on 28 July 2021 Taxon/code 514 S.-W. JIA ET AL. A NCESTRAL AREA ESTIMATES The ancestral areas were reconstructed using Statistical Dispersal-Vicariance Analysis (S-DIVA; Yu, Harris & He, 2010), maximum likelihood – based DEC model (Lagrange; Ree & Smith, 2008;.), and the Bayesian Binary Method (BBM; Ronquist & Huelsenbeck, 2003) as implemented in RASP (Reconstruct Ancestral State in Phylogenies) 3.1 (Yu et al., 2015). According to geological conditions or species occurrence, eight distributional areas of the Gymnocarpos and its relatives were delimited: A, North America; B, Europe (mainly Spain); C, North Africa (including Canary Islands, Libya, Tunisia, Algeria, Morocco, Western Sahara, Mauritania, Chad); D, Horn of Africa (including Eritrea, Somalia); E, Northern Arabian Peninsula (including Israel, Lebanon, Sinai, Jordan, Iraq, Syria, Saudi Arabia, Kuwait, Bahrain, Qatar); F, Southern Arabian Peninsula (including Yemen, Oman, Socotra, Abd alKuri); G, Iran plateau & Central Asia (including Iran, Afghanistan, Pakistan); H, Mongolia & Northwest China (details of the setting are in Supporting Information, Table S3 and Fig. S1). For the endemic species, we defined distribution ranges based on species occurrence. For the widely distributed species, G. decandrus and G. sclerocephalus, and the relatives of Gymnocarpos, we defined distribution ranges based on Global Biodiversity Information Facility (GBIF, http://www.gbif.org/, accessed 19 January 2016). The maximum number of areas for each node was set as 2, 3, 4, or 5 for comparison, because most species are not observed in more than five areas in our study. We loaded previously produced 10001 BEAST trees and discard the first 9001 trees. The last 1000 BEAST trees were used in this analysis. The condensed tree was also estimated in BEAST. Outgroups of distant relatives were excluded in calculations of ancestral distributional area by RASP 3.1 (Yu et al., 2015). The S-DIVA ran using default settings. The BBM ran using model F81 according to Akaike Information Criterion (AIC: F81 = 33 024.1016, JC = 33 246.4531). Due to only a few of the study of the genus, the dispersal rates of Gymnocarpos are very difficult to estimate. Thus, we set all dispersal events equally likely in DEC analyela & Rouhan, 2016). sis (Gaudeul, V RESULTS P HYLOGENETIC ANALYSES AND DIVERGENCE TIME ESTIMATES The total aligned matrix consisted of 2120 characters. The aligned positions of psbB–psbH, rps16 and ITS datasets were 666, 848 and 606 base pairs, respectively. The combined matrix of the three markers consisted of 155 polymorphic informative sites, including 122 nucleotide substitutions and 33 indels (Supporting Information, Tables S4–S6). All the variations were considered in the present study, including substitutions and indels. In our study, many outgroups were used, the positions of gaps of Gymnocarpos were inconsistent with outgroups. Thus, we did not treated indel as a single mutation event. Phylogenetic trees resulting from MP and BI analyses were generally similar (Fig. 1, Supporting Information, Figs S2, S3). The Gymnocarpos group was corroborated as monophyletic with high bootstrap (BP) support and posterior probability (PP) © 2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, 119, 511–522 Downloaded from https://academic.oup.com/biolinnean/article/119/2/511/2701016 by guest on 28 July 2021 as burn-in. The remaining trees were used to construct a Bayesian consensus tree. Divergence time was estimated using a Bayesian approach of BEAST version 1.8.1 (Drummond & Rambaut, 2007), as implemented in CIPRES Science Gateway V. 3.3 (Miller, Pfeiffer & Schwartz, 2010). The likelihood ratio test performed in PAUP* 4b10 (Swofford, 2002) rejected the hypothesis of a strict molecular clock model (ln-likelihood without clock: – 15 538, ln-likelihood with clock: – 16 939, v2 value = 2802, 47 d.f. = 82.720, P < 0.001). Thus, the Uncorrelated Lognormal relaxed clock was employed for clock models. Because the Yule process (pure birth) and the birth–death process (considered the extinction issue) are commonly used for modelling speciation, the two processes were employed to run separately. We calibrated two nodes using one macrofossil and one published divergence time of Caryophyllaceae. First, the previous report of Kool (2012) provided evidence that Caryophyllaceae originated in the Paleocene, around 60 Mya. Therefore, for the place of Caryophyllaceae origin, we used a normal prior distribution with mean 60 Mya and standard deviation 3 Mya. This time range (95% HPD: 52–64 Mya) corresponds to the Paleocene period. Second, we used the inflorescence fossil (Caryophylloflora paleogenica; Jordan & Macphail, 2003) as a fixed age for the Alsinoideae/Caryophylloideae node. We set a lognormal prior distribution with offset 34 Mya and standard deviation 1 Mya for the node according to Frajman et al. (2009). The MCMC search was run for 20 000 000 generations and sampled every 2000 generations. Four independent Markov chains were used in this process. A maximum clade credibility (MCC) tree was generated in TreeAnnotator (Drummond & Rambaut, 2007) using the product method, after which we discarded the first 10% as burn-in. The convergence of MCMC and adequate effective sample sizes (ESSs > 200) were checked by Tracer version 1.6 (Rambaut et al., 2014). EVOLUTIONARY HISTORY OF GYMNOCARPOS 515 Clade2 (BP = 100, PP = 1.00), and was related to Paronychia. Two clades within the genus were recognized, and each was well supported (clade 1, BP = 99.69, PP = 1.00; clade 2, BP = 99.6, PP = 1.00). Samples of G. sclerocephalus and G. decandrus formed a clade and the two species are distributed in western parts of the geographic range, whereas the other clade includes species distributed in the Southern Arabian Peninsula and the western part of China. Results of divergence time analysis from birth– death and Yule priors showed no significant differences. The details are shown in Figure 2 and Supporting Information (Fig. S4). Because the Yule model is a special case of the birth–death process where the rate for extinction is equal to zero, the birth–death results are more realistic than those from Yule. The following information will show the results of birth–death analysis. The estimated divergence (stem) age of Gymnocarpos was c. 31.33 Mya (95% HPD: 19.94–45.42); the crown age was c. 6.69 Mya (95% HPD: 3.97–10.48); and the crown ages of clade 1 and clade 2 were c. 2.73 Mya (95% HPD: 1.52–4.43) and 2.86 Mya (95% HPD: 1.66– 4.38), respectively. The diversification crown age of G. przewalskii was 1.16 Mya (95% HPD: 0.55–2.01) and that of G. decandrus was 1.29 Mya (95% HPD: 0.7–2.15). A NCESTRAL AREA ESTIMATES Results from ancestral area reconstruction using SDIVA, BBM and DEC are shown in Figures 3, 4, and Supporting Information (Figs S5–S7). The results of the S-DIVA analysis using different maximum area numbers were almost identical (Supporting Information, Fig. S5). The results of BBM using different maximum area numbers analysis were also almost identical (Supporting Information, Fig. S6). DEC results using different maximum area numbers were significantly different and less informative than those of S-DIVA and BBM, because DEC inferred numerous ancestral areas or null distribution in some notes (Supporting Information, Fig. S7). Therefore, we discuss results based on S-DIVA and BMM analyses. The results of the two methods were similar except for the order of dispersal events. The ancestral area was estimated as F by both S-DIVA and BBM (Fig. 3). As F (Southern Arabian Peninsula) always occurs as an ancestral area, it was identified as the most likely ancestral area. The multiple © 2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, 119, 511–522 Downloaded from https://academic.oup.com/biolinnean/article/119/2/511/2701016 by guest on 28 July 2021 Clade1 Figure 1. Phylogenetic tree based on ITS, rps16 and psbB–psbH sequences. Numbers above branches are posterior probability support values from Bayesian inference, and below branches are bootstrap support values from maximum parsimony analysis. 516 S.-W. JIA ET AL. Figure 2. Divergence time estimates of Gymnocarpos based on the three plastid and nuclear regions inferred using the birth–death process. Blue bars represent 95% higher posterior densities (95% HPD) of divergence time. Red lines represent the nodes of calibration. dispersal events were recognized, from Southern Arabian Peninsula (F) to all other distribution areas. The order of dispersal events for S-DIVA was from F to H and from F to C, D, E, G. The order of dispersal events for BBM is from F to H; from F to CF, and from CF to D, E, G. DISCUSSION P HYLOGENETIC RELATIONSHIPS OF GYMNOCARPOS In terms of traditional taxonomy, there were formerly two species in Gymnocarpos (Chaudhri, 1968), i.e. western G. decandrus, and eastern G. przewalskii, both form a distribution disjunction. Petrusson & Thulin (1996) resurrected Gymnocarpos and included eight species in the genus. The molecular phylogeny indicated that the genus Sclerocephalus was related to G. decandrus and nested in Gymnocarpos and was highly supported, so that Sclerocephalus was included within Gymnocarpos (Oxelman et al., 2002) and therefore ten species are recognized in the genus so far. Our molecular phylogenetic analysis (Fig. 1) confirmed the previous phylogeny and taxonomic treatment, namely, Gymnocarpos is monophyletic, Gymnocarpos together with Paronychia and Herniaria, forms a well supported monophyletic clade, and the subfamily Paronychioideae should be included in Caryophyllaceae (Bittrich, 1993; Oxelman et al., 2002; Greenberg & Donoghue, 2011; Kool, 2012). Within Gymnocarpos, the two formed clades were consistent with the distribution pattern: clade 1 was comprised of two widely distributed species in the western part, G. decandrus and G. sclerocephalus, whereas clade 2 was comprised of G. przewalskii in eastern China and Mongolia, as well as species in Southern Arabian Peninsula, G. mahranus, G. decandrus, G. argenteus, G. bracteatus, and G. rotundifolius. G. przewalskii is shown to be closer © 2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, 119, 511–522 Downloaded from https://academic.oup.com/biolinnean/article/119/2/511/2701016 by guest on 28 July 2021 Mya EVOLUTIONARY HISTORY OF GYMNOCARPOS BBM 517 S-DIVA LEGEND LEGEND * * (F) Gymnocarpos kuriensis CDEFG 100 (F) Gymnocarpos bracteatus (F) Gymnocarpos kuriensis CDEFG 100 CEFG CEFG (F) Gymnocarpos argenteus 97 CEG F (F) Gymnocarpos mahranus (F) Gymnocarpos mahranus 100 CF 100 FH 61 61 (F) Gymnocarpos rotundifolius H F (H) Gymnocarpos przewalskii4 100 62 62 Dispersal 100 (H) Gymnocarpos przewalskii1 100 100 (H) Gymnocarpos przewalskii5 (H) Gymnocarpos przewalskii5 Dispersal (CEFG) Gymnocarpos decandrus2 (CEFG) Gymnocarpos decandrus2 100 100 100 (CEFG) Gymnocarpos decandrus1 100 100 (CEFG) Gymnocarpos decandrus6 (CEFG) Gymnocarpos decandrus6 (CEFG) Gymnocarpos decandrus5 (CEFG) Gymnocarpos decandrus5 (CEFG) Gymnocarpos decandrus7 100 (CEFG) Gymnocarpos decandrus7 100 (CEFG) Gymnocarpos decandrus1 (CEFG) Gymnocarpos decandrus8 (CEFG) Gymnocarpos decandrus8 100 (H) Gymnocarpos przewalskii1 100 100 (CEFG) Gymnocarpos decandrus3 (CEFG) Gymnocarpos decandrus3 100 100 100 100 (CEFG) Gymnocarpos decandrus4 (B) Paronychia suffruticosa 68 (CEFG) Gymnocarpos decandrus4 (CDEFG) Gymnocarpos sclerocephalus (CDEFG) Gymnocarpos sclerocephalus (B) Paronychia suffruticosa 68 (A) Paronychia americana (A) Paronychia americana 100 100 100 100 (A) Paronychia fastigiata (A) Paronychia fastigiata (CEG) Herniaria hemistemon (CEG) Herniaria hemistemon Figure 3. Ancestral area reconstruction of Gymnocarpos conducted by RASP. The maximum number of areas is two. Arrows represent the direction of dispersal events. Figure 4. Several dispersal events are illustrated. Blue dotted lines represent dispersal events of BBM. Red dotted lines represent dispersal events of S-DAVA. The direction of the arrow represents the direction of dispersal event. to Southern Arabian Peninsula species than those in North Africa and SW Asia (G. decandrus and G. sclerocephalus). This is also supported by a typical morphological character, being large and scarious or small and with stipule-like bracts among the endemic species of Southern Arabia Peninsula and G. przewalskii (Oxelman et al., 2002), whereas G. sclerocephalus and G. decandrus with leaf-like bracts. D ISPERSAL EVENTS IN GYMNOCARPOS Our ancestral areas reconstruction showed that the Southern Arabian Peninsula (including the adjacent islands in the Indian Ocean) was the ancestral area of Gymnocarpos (see Fig. 4). Subsequently, dispersal events in clade 2, indicated that G. przewalskii is a dispersal event from the Southern Arabian Peninsula, this is different from the pattern from eastern © 2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, 119, 511–522 Downloaded from https://academic.oup.com/biolinnean/article/119/2/511/2701016 by guest on 28 July 2021 H (H) Gymnocarpos przewalskii3 (H) Gymnocarpos przewalskii2 100 (H) Gymnocarpos przewalskii2 100 (F) Gymnocarpos rotundifolius (H) Gymnocarpos przewalskii4 100 (H) Gymnocarpos przewalskii3 FH 100 (F) Gymnocarpos bracteatus (F) Gymnocarpos argenteus 97 518 S.-W. JIA ET AL. C LIMATE CHANGES TRIGGERED THE DIVERSIFICATION HISTORY OF GYMNOCARPOS The estimated divergence age of Gymnocarpos was c. 31.33 Mya, following the global abrupt shift to glacial conditions near the Eocene and Oligocene transition (EOT) boundary c. 34 Mya (Crowley & North, 1988; Coxall et al., 2005; Liu et al., 2009). Numerous studies on biostratigraphy and paleontology evidenced that a dramatic variety of composition and evolution of taxa correlated with the abrupt change of climate (e.g. Dockery, 1986; Cavagnetto & Anadon, 1996; Haasl & Hansen, 1996; Ivany, Patterson & Lohmann, 2000; Pan et al., 2006; Seiffert, 2007). Notable species turnover and decline of palms (Arecaceae) occurred at the Eocene and Oligocene boundary 33.9 Mya in Africa (Pan et al., 2006). The vegetation of Ebro Basin in northeastern Spain changed from mangrove swamps to sclerophyllous forests in order to adapt to the climate change during that period (Cavagnetto & Anadon, 1996). Therefore, the divergence of Gymnocarpos was also likely to be affected by climate changes at EOT. Diversification of Gymnocarpos was at c. 6.69 Mya in late Miocene, corresponding to global cooling and aridization c. 7 Mya (Quade, Cerling & Bowman, 1989; Willis & McElwain, 2002). The uplift of the Himalayas and the QTP significantly affected the climate of Asia, and enhanced the aridity of the Asian interior about 8–9 Mya (An et al., 2001). North Africa covered with an open grassland and the Sahara desert appeared firstly also was at c. 7 Mya (Fortelius et al., 2002; Liu et al., 2009; Micheels et al., 2009), where previously it had been dominated by tropical trees in early Miocene. The vegetation changes were driven by climatic oscillations during the Miocene (Linares, 2011; Chen et al., 2014). Our evidence offers a case of Arabian Peninsula or northern Africa origin and driven by climatic cooling and aridification during the mid-late Miocene. From Figure 2, the diversification within the genus occurred during the Pliocene and Pleistocene, corresponding to the Quaternary glaciations, which have profoundly affected the distribution and genetic variation of many organisms throughout the Northern Hemisphere (Hewitt, 2000, 2004). In Northwest China, aridification has presumably begun at early Miocene, drastically increased aridity and expansion of deserts occurred during the Pleistocene (Yang et al., 2011). For example, the several deserts greatly enlarged during the Pleistocene (Fang et al., 2002; Sun, 2002; Zhang & Men, 2002; Li et al., 2014). The genetic structures, population isolation and diversification of many plants in those regions were affected significantly by severe aridity (e.g. Su, Zhang & Cohen, 2012; Meng & Zhang, 2013; Xie & Zhang, 2013). From our dating of G. przewalskii to after 1.16 Mya, we can infer that the aridification prompted the formation of intraspecific genetic variations during the Quaternary glaciations. This finding is consistent with previous phylogeographical studies of G. przewalskii (Ma & Zhang, 2012; Ma, Zhang & Sanderson, 2012). Quaternary glaciations also had a certain effect on the Mediterranean region (Bertoldi, Rio & Thunell, 1989). The diversification of G. decandrus in North Africa occurred starting from 1.29 Mya. This result not only corresponds with increasingly dramatic climate oscillations (Hewitt, 2000), but is also linked closely to the changes of sea surface temperatures in the tropical Atlantic Ocean and hence large-scale vegetation changes in the mid-Pleistocene (Schefuß et al., 2003). In the Arabian Peninsula, the increasing aridification began in the late Tertiary, resulting in the shrinkage of a once continuous distributional ranges of species (Meister et al., 2005). Moreover, the numerous climate oscillations between humid and arid period, which were caused by glacial–interglacial episodes during the Pleistocene and the early Holocene, shaped the present distribution of plants in the Arabian Peninsula (Bray & Stokes, 2004; Meister et al., 2005). Therefore, we infer that climate changes since the Pliocene has had a profound influence on the diversification of the Gymnocarpos species. CONCLUSIONS Gymnocarpos exhibits an interesting geographical disjunction, western North Africa, Southern Arabian Peninsula, and eastern Northwest China and Mongolia. In this study, we provide an initial © 2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, 119, 511–522 Downloaded from https://academic.oup.com/biolinnean/article/119/2/511/2701016 by guest on 28 July 2021 QTP to western Mediterranean, the Arabian Peninsula and Africa, such as Hippopha€ e rhamnoides (Jia et al., 2012) and Isodon (Yu et al., 2014). In clade 1, the S-DIVA result illustrated that all dispersals were from the Arabian Peninsula to different directions in the Miocene–Pliocene, whereas BBM result revealed that dispersal events occurred from the Southern Arabian Peninsula to North Africa, then returned to Northern Arabian Peninsula and Iran. Even though previous studies have indicated that the Sahara desert might be a barrier of gene flow between northern and southern African c. 7 Mya (Besnard, Rubio de Casas & Vargas, 2007), some biological cases have also evidenced that organisms can migrate in the Sahara desert (Drake et al., 2011), thus our dispersals between North Africa and Arabia Peninsula are possible and justifiable. EVOLUTIONARY HISTORY OF GYMNOCARPOS understanding of the origin and evolutionary history of this disjunction. The ancestral area is estimated as the Southern Arabian Peninsula. The aridification events since the mid–late Miocene are presumed to have significantly influenced the diversification of the genus. Diversification within North Africa, Northwest China, and the Southern Arabian Peninsula is hypothesized to be recent, since the Pliocene, and most likely driven by the Quaternary glacial oscillation, Sahara Desert effects in North Africa, QTP uplift, and desert expansion in Northwest China. Thanks to Dr Gabriele Dr€oge at the DNA bank (http://www.bgbm.org/en/dna-bank) of Botanischer Garten and Botanisches Museum Berlin-Dahlem, Freie Universit€ at Berlin (Germany) for her kind donation of DNA material, to Prof. Mats Thulin at UPP herbarium of Department of Systematic Biology, Evolutionary Biology Centre, Uppsala University (Sweden) and Dr Abelardo Aparicio at the herbarium of the Real Jardin Botanico of Madrid (Spain) for sending leaves. Thanks to Dr Stewart C. Sanderson at the Shrub Sciences Laboratory, USDA, Utah, USA, for his English improvement to the manuscript. This study is financially supported by China Natural Key Basic Research Programme (2014CB954201), and the Xinjiang Institute of Ecology and Geography, Chinese Academy of Sciences. Finally, thanks to the nine anonymous reviewers for helpful comments on this paper. REFERENCES An ZH, John EK, Warren LP, Stephen CP. 2001. Evolution of Asian monsoons and phased uplift of the HimalayaTibetan plateau since Late Miocene times. Nature 411: 62– 66. Barrientos R, Kvist L, Barbosa A, Valera F, Khoury F, Varela S, Moreno E. 2014. Refugia, colonization and diversification of an arid – adapted bird: coincident patterns between genetic data and ecological niche modelling. Molecular Ecology 23: 390–407. Bertoldi R, Rio D, Thunell R. 1989. Pliocene-pleistocene vegetational and climatic evolution of the south-central mediterranean. Palaeogeography, Palaeoclimatology, Palaeoecology 72: 263–275. Besnard G, Rubio de Casas R, Vargas P. 2007. Plastid and nuclear DNA polymorphism reveals historical processes of isolation and reticulation in the olive tree complex (Olea europaea). Journal of Biogeography 34: 736–752. Bittrich V. 1993. Caryophyllaceae. In: Kubitzki K, Rohwer JG, Bittrich V, eds. The families and genera of vascular plants II. Flowering plants-dicotyledons. Berlin: Springer, 206–236. Bray HE, Stokes S. 2004. Temporal patterns of arid-humid transitions in the south-eastern Arabian Peninsula based on optical dating. Geomorphology 59: 271–280. Cavagnetto C, Anad on P. 1996. Preliminary palynological data on floristic and climatic changes during the Middle Eocene-Early Oligocene of the eastern Ebro Basin, northeast Spain. Review of Palaeobotany and Palynology 92: 281–305. Chaudhri MN. 1968. Revision of the Paronychiinae. Mededelingen van het Botanisch Museum en Herbarium van de Rijksuniversiteit te Utrecht 1: 3–440. Chen RF, Zhang Z, Tang Z, Yu MD, Xu L, Wang XL. 2011. Morus ITS, trnL-F, rps16 sequence and phylogenetic analysis of mulberry resources. Scientia Agricultura Sinica 44: 1553–1561. Chen C, Qi ZC, Xu XH, Comes HP, Koch MA, Jin XJ, Fu CX, Qiu YX. 2014. Understanding the formation of Mediterranean–African–Asian disjunctions: evidence for Miocene climate-driven vicariance and recent long-distance dispersal in the Tertiary relict Smilax aspera (Smilacaceae). New Phytologist 204: 243–255. Coxall HK, Wilson PA, P€ alike H, Lear CH, Backman J. 2005. Rapid stepwise onset of Antarctic glaciation and deeper calcite compensation in the Pacific Ocean. Nature 433: 53–57. Crowley TJ, North GR. 1988. Abrupt climate change and extinction events in earth history. Science 240: 996–1002. Demenocal PB. 2004. African climate change and faunal evolution during the Pliocene-Pleistocene. Earth and Planetary Science Letters 220: 3–24. Ding ZL, Derbyshire E, Yang SL, Sun JM, Liu TS. 2005. Stepwise expansion of desert environment across northern China in the past 3.5 Ma and implications for monsoon evolution. Earth and Planetary Science Letters 237: 45–55. Dockery DT III. 1986. Punctuated succession of Paleogene mollusks in the northern Gulf Coastal Plain. Palaios 1: 582–589. Doyle J, Doyle J. 1987. A rapid DNA isolation procedure for small quantities of fresh leaf tissue. Phytochemical Bulletin 19: 11–15. Drake NA, Blench RM, Armitage SJ, Bristow SC, White HK. 2011. Ancient watercourses and biogeography of the Sahara explain the peopling of the desert. Proceedings of the National Academy of Sciences 108: 458–462. Drummond AJ, Rambaut A. 2007. BEAST: Bayesian evolutionary analysis by sampling trees. BMC Evolutionary Biology 7: 214. Fang X, Shi Z, Yang S, Yan MD, Li JJ, Jing PA. 2002. Loess in the Tian Shan and its implications for the development of the Gurbantunggut Desert and drying of northern Xinjiang. Chinese Science Bulletin 47: 1381–1387. Flower BP, Kennett JP. 1994. The middle Miocene climatic transition: East Antarctic ice sheet development, deep ocean circulation and global carbon cycling. Palaeogeography, Palaeoclimatology, Palaeoecology 108: 537–555. Fortelius M, Eronen J, Jernvall J, Liu L, Pushkina D, Rinne J, Tesakov J, Vislobokova I, Zhang Z, Zhou L. © 2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, 119, 511–522 Downloaded from https://academic.oup.com/biolinnean/article/119/2/511/2701016 by guest on 28 July 2021 ACKNOWLEDGEMENTS 519 520 S.-W. JIA ET AL. long-term climatic change and glacial refugia. Journal of Biogeography 38: 619–630. Liu X, Yin Z-Y. 2002. Sensitivity of East Asian monsoon climate to the uplift of the Tibetan Plateau. Palaeogeography, Palaeoclimatology, Palaeoecology 183: 223–245. Liu Z, Pagani M, Zinniker D, DeConto R, Huber M, Brinkhuis H, Shah SR, Leckie RM, Pearson A. 2009. Global cooling during the Eocene-Oligocene climate transition. Science 323: 1187–1190. Ma SM, Zhang ML. 2012. Phylogeography and conservation genetics of the relic Gymnocarpos przewalskii (Caryophyllaceae) restricted to northwestern China. Conservation Genetics 13: 1531–1541. Ma SM, Zhang ML, Sanderson SC. 2012. Phylogeography of the rare Gymnocarpos przewalskii (Caryophyllaceae): indications of multiple glacial refugia in north-western China. Australian Journal of Botany 60: 20–31. Meister J, Hubaishan MA, Kilian N, Oberprieler C. 2005. Chloroplast DNA variation in the shrub Justicia areysiana (Acanthaceae) endemic to the monsoon affected coastal mountains of the Southern Arabian Peninsula. Botanical Journal of the Linnean Society 148: 437–444. Meng HH, Zhang ML. 2013. Diversification of plant species in arid Northwest China: species-level phylogeographical history of Lagochilus Bunge ex Bentham (Lamiaceae). Molecular Phylogenetics and Evolution 68: 398–409. Micheels A, Eronen J, Mosbrugger V. 2009. The Late Miocene climate response to a modern Sahara desert. Global and Planetary Change 67: 193–204. Miller MA, Pfeiffer W, Schwartz T. 2010. Creating the CIPRES Science Gateway for inference of large phylogenetic trees. Gateway Computing Environments Workshop 2010 (GCE), , 1–8. Nylander JAA. 2004. MrModeltest V. 2.2. Uppsala, Sweden: Uppsala University, Department of Systematic Zoology. Oxelman B, Ahlgren B, Thulin M. 2002. Circumscription and phylogenetic relationships of Gymnocarpos (Caryophyllaceae-Paronychioideae). Edinburgh Journal of Botany 59: 221–237. Pan AD, Jacobs BF, Dransfield J, Baker WJ. 2006. The fossil history of palms (Arecaceae) in Africa and new records from the Late Oligocene (28–27 Mya) of north-western Ethiopia. Botanical Journal of the Linnean Society 151: 69–81. Petrusson L, Thulin M. 1996. Taxonomy and biogeography of Gymnocarpos (Caryophyllaceae). Edinburgh Journal of Botany 53: 1–26. Quade J, Cerling TE, Bowman JR. 1989. Development of Asian monsoon revealed by marked ecological shift during the latest Miocene in northern Pakistan. Nature 342: 163– 166. Rambaut A, Suchard MA, Xie D, Drummond AJ. 2014. Tracer v1.6. Available at: http://beast.bio.ed.ac.uk/Tracer Ree RH, Smith SA. 2008. Maximum likelihood inference of geographic range evolution by dispersal, local extinction, and cladogenesis. Systematic Biology 57: 4–14. © 2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, 119, 511–522 Downloaded from https://academic.oup.com/biolinnean/article/119/2/511/2701016 by guest on 28 July 2021 2002. Fossil mammals resolve regional patterns of Eurasian climate change over 20 million years. Evolutionary Ecology Research 4: 1005–1016. Frajman B, Eggens F, Oxelman B. 2009. Hybrid origins and homoploid reticulate evolution within Heliosperma (Sileneae, Caryophyllaceae) – a multigene phylogenetic approach with relative dating. Systematic Biology 58: 328– 345. Gao XY, Meng HH, Zhang ML. 2014. Diversification and vicariance of desert plants: evidence inferred from chloroplast DNA sequence variation of Lagochilus ilicifolius (Lamiaceae). Biochemical Systematics and Ecology 55: 93– 100. Gaudeul M, V ela E, Rouhan G. 2016. Eastward colonization of the Mediterranean Basin by two geographically structured clades: the case of Odontites Ludw. (Orobanchaceae). Molecular Phylogenetics and Evolution 96: 140– 149. Greenberg AK, Donoghue MJ. 2011. Molecular systematics and character evolution in Caryophyllaceae. Taxon 60: 1637–1652. Haasl DM, Hansen TA. 1996. Timing of latest Eocene molluscan extinction patterns in Mississippi. Palaios 11: 487– 494. Hewitt G. 2000. The genetic legacy of the Quaternary ice ages. Nature 405: 907–913. Hewitt G. 2004. Genetic consequences of climatic oscillations in the Quaternary. Philosophical Transactions of the Royal Society of London. Series B: Biological Sciences 359: 183– 195. Huang ZS, Ji YJ, Zhang DX. 2008. Haplotype reconstruction for scnp DNA: a consensus vote approach with extensive sequence data from populations of the migratory locust (Locusta migratoria). Molecular Ecology 17: 1930–1947. Ivany LC, Patterson WP, Lohmann KC. 2000. Cooler winters as a possible cause of mass extinctions at the Eocene/ Oligocene boundary. Nature 407: 887–890. Jia DR, Abbott RJ, Liu TL, Mao KS, Bartish IV, Liu JQ. 2012. Out of the Qinghai-Tibet Plateau: evidence for the origin and dispersal of Eurasian temperate plants from a phylogeographic study of Hippopha€ e rhamnoides (Elaeagnaceae). New Phytologist 194: 1123–1133. Jordan GJ, Macphail MK. 2003. A middle-late Eocene inflorescence of Caryophyllaceae from Tasmania, Australia. American Journal of Botany 90: 761–768. Kool A. 2012. Desert plants and deserted islands: systematics and ethnobotany in Caryophyllaceae. PhD Thesis, Uppsala University. Li G, Jin M, Wen L, Zhao H, Madsen D, Liu XK, Wu D, Chen FH. 2014. Quartz and K-feldspar optical dating chronology of eolian sand and lacustrine sequence from the southern Ulan Buh Desert, NW China: implications for reconstructing late Pleistocene environmental evolution. Palaeogeography, Palaeoclimatology, Palaeoecology 393: 111–121. Linares JC. 2011. Biogeography and evolution of Abies (Pinaceae) in the Mediterranean basin: the roles of EVOLUTIONARY HISTORY OF GYMNOCARPOS phylogenetics. PCR Protocols: A Guide to Methods and Applications 18: 315–322. Willis KJ, McElwain JC. 2002. The evolution of plants. Oxford: Oxford University Press. Wu F, Fang X, Ma Y, Herrmannd M, Mosbrugger V, An Z, Miao Y. 2007. Plio-Quaternary stepwise drying of Asia: evidence from a 3-Ma pollen record from the Chinese Loess Plateau. Earth and Planetary Science Letters 257: 160–169. Xie KQ, Zhang ML. 2013. The effect of Quaternary climatic oscillations on Ribes meyeri (Saxifragaceae) in northwestern China. Biochemical Systematics and Ecology 50: 39–47. Xu D, Abe J, Sakai M, Kanazawa A, Shimamoto Y. 2000. Sequence variation of non-coding regions of chloroplast DNA of soybean and related wild species and its implications for the evolution of different chloroplast haplotypes. Theoretical and Applied Genetics 101: 724–732. Yang X, Scuderi L, Paillou P, Liu Z, Li H, Ren X. 2011. Quaternary environmental changes in the drylands of China – a critical review. Quaternary Science Reviews 30: 3219–3233. Yu Y, Harris A, He X. 2010. S-DIVA (statistical dispersalvicariance analysis): a tool for inferring biogeographic histories. Molecular Phylogenetics and Evolution 56: 848–850. Yu XQ, Maki M, Drew BT, Paton AJ, Li HW, Zhao JL, Conran JG, Li J. 2014. Phylogeny and historical biogeography of Isodon (Lamiaceae): rapid radiation in south-west China and Miocene overland dispersal into Africa. Molecular Phylogenetics and Evolution 77: 183–194. Yu Y, Harris AJ, Blair C, He X. 2015. RASP (reconstruct ancestral state in phylogenies): a tool for historical biogeography. Molecular Phylogenetics and Evolution 87: 46–49. Zhang HY, Men GF. 2002. Stratigraphic subdivision and climatic change of the Quaternary of the center Taklimakan Desert. Xinjiang Geology 20: 257–261. SUPPORTING INFORMATION Additional Supporting Information may be found online in the supporting information tab for this article: Figure S1. Distribution of Gymnocarpos species and outgroups used for ancestral area reconstruction, data from our field investigations in China, published literature, Global Biodiversity Information Facility and herbaria (LE, B, MA, UPS). Eight distribution areas are defined: A, North America; B, Europe (mainly Spain); C, North Africa; D, Horn of Africa; E, Northern Arabian Peninsula; F, Southern Arabian Peninsula; G, Iran plateau and Central Asia; H, Mongolia and Northwest China. Figure S2. Phylogenetic tree based on rps16 and psbB–psbH sequences. Numbers above branches are posterior probability support values from Bayesian inference, and below branches are bootstrap support values from maximum parsimony analysis. Figure S3. Phylogenetic tree based on ITS sequences. Numbers above branches are posterior probability support values from Bayesian inference, and below branches are bootstrap support values from maximum parsimony analysis. Figure S4. Divergence time estimates of Gymnocarpos based on the three plastid and nuclear fragments inferred using Yule process. Blue bars represent 95% higher posterior densities (95% HPD) of divergence time. Red lines represent the nodes of calibration. © 2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, 119, 511–522 Downloaded from https://academic.oup.com/biolinnean/article/119/2/511/2701016 by guest on 28 July 2021 Ronquist F, Huelsenbeck JP. 2003. MrBayes3: Bayesian phylogenetic inference undermixed models. Bioinformatics 19: 1572–1574. Ronquist F, Teslenko M, van der Mark P, Ayres DL, Darling A, H€ ohna S, Larget B, Liu L, Suchard MA, Huelsenbeck JP. 2012. MrBayes 3.2: efficient Bayesian phylogenetic inference and model choice across a large model space. Systematic Biology 61: 539–542. Schefuß E, Schouten S, Jansen JHF, Damste JSS. 2003. African vegetation controlled by tropical sea surface temperatures in the mid-Pleistocene period. Nature 422: 418–421. Seiffert ER. 2007. Evolution and extinction of Afro-Arabian primates near the Eocene-Oligocene boundary. Folia Primatologica 78: 314–327. Shi CM, Ji YJ, Liu L, Wang L, Zhang DX. 2013. Impact of climate changes from Middle Miocene onwards on evolutionary diversification in Eurasia: insights from the mesobuthid scorpions. Molecular Ecology 22: 1700–1716. Su ZH, Zhang ML, Cohen JI. 2012. Phylogeographic and demographic effects of Quaternary climate oscillations in Hexinia polydichotoma (Asteraceae) in Tarim Basin and adjacent areas. Plant Systematics and Evolution 298: 1767–1776. Sun JM. 2002. Source regions and formation of the loess sediments on the high mountain regions of northwestern China. Quaternary Research 58: 341–351. Swofford DL. 2002. PAUP*: phylogenetic analysis using parsimony (*and other methods). Version 4.0b10. Sunderland, MA: Sinauer Associates. Thompson JD, Higgins DG, Gibson TJ. 1994. CLUSTAL W: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, positionspecific gap penalties and weight matrix choice. Nucleic Acids Research 22: 4673–4680. White TJ, Bruns T, Lee S, Taylor J. 1990. Amplification and direct sequencing of fungal ribosomal RNA genes for 521 522 S.-W. JIA ET AL. © 2016 The Linnean Society of London, Biological Journal of the Linnean Society, 2016, 119, 511–522 Downloaded from https://academic.oup.com/biolinnean/article/119/2/511/2701016 by guest on 28 July 2021 Figure S5. Ancestral areas reconstruction of the Gymnocarpos using Statistical Dispersal-Vicariance Analysis (S-DIVA) with maximum number of areas from 2 to 5. The four analyses produced similar results. A, North America; B, Europe; C, North Africa; D, Horn of Africa; E, Northern Arabian Peninsula; F, Southern Arabian Peninsula; G, Iran plateau and Central Asia; H, Mongolia and Northwest China. Figure S6. Ancestral areas reconstruction of the Gymnocarpos using Bayesian Binary Method (BBM) with maximum number of areas from 2 to 5. The four analyses produced similar results. A, North America; B, Europe; C, North Africa; D, Horn of Africa; E, Northern Arabian Peninsula; F, Southern Arabian Peninsula; G, Iran plateau and Central Asia; H, Mongolia and Northwest China. Figure S7. Ancestral areas reconstruction of the Gymnocarpos using the maximum likelihood-based DEC model with maximum number of areas from 2 to 5. The four analyses produced different results. A, North America; B, Europe (mainly Spain); C, North Africa; D, Horn of Africa; E, Northern Arabian Peninsula; F, Southern Arabian Peninsula; G, Iran plateau and Central Asia; H, Mongolia and Northwest China. Table S1. The geographical distribution of species of Gymnocarpos. For the endemic species, the distribution ranges were defined based on species occurrence. For the widely distributed species, G. decandrus and G. sclerocephalus, distribution ranges were defined based on Global Biodiversity Information Facility (GBIF; http:// www.gbif.org/, accessed 19 January 2016). Table S2. GenBank accession numbers of the sequences of outgroups included in this study. Table S3. Distribution area set for RASP analysis. Table S4. DNA sequence polymorphisms detected in the rps16 region of the Gymnocarpos. Table S5. DNA sequence polymorphisms detected in the psbB–psbH region of the Gymnocarpos. Table S6. DNA sequence polymorphisms detected in the ITS region of the Gymnocarpos.