Molecular Phylogenetics and Evolution 52 (2009) 806–824 Contents lists available at ScienceDirect Molecular Phylogenetics and Evolution journal homepage: www.elsevier.com/locate/ympev Rapid radiation of Impatiens (Balsaminaceae) during Pliocene and Pleistocene: Result of a global climate change Steven B. Janssens a,*, Eric B. Knox b, Suzy Huysmans a, Erik F. Smets a,c, Vincent S.F.T. Merckx a a Laboratory of Plant Systematics, Institute of Botany and Microbiology, K.U.Leuven, Kasteelpark Arenberg 31, P.O. Box 2437, BE-3001 Leuven, Belgium Department of Biology, Indiana University, Jordan Hall 142, Bloomington, IN 47405, USA c National Herbarium of the Netherlands, University Leiden Branch, NL-2300 RA Leiden, The Netherlands b a r t i c l e i n f o Article history: Received 28 November 2008 Revised 3 April 2009 Accepted 14 April 2009 Available online 3 May 2009 Keywords: Global cooling Impatiens Pliocene Pleistocene Rapid radiation a b s t r a c t Impatiens comprises more than 1000 species and is one of the largest genera of flowering plants. The genus has a subcosmopolitan distribution, yet most of its evolutionary history is unknown. Diversification analyses, divergence time estimates and historical biogeography, illustrated that the extant species of Impatiens originated in Southwest China and started to diversify in the Early Miocene. Until the Early Pliocene, the net diversification rate within the genus was fairly slow. Since that time, however, approximately 80% of all Impatiens lineages have originated. This period of rapid diversification coincides with the global cooling of the Earth’s climate and subsequent glacial oscillations. Without this accelerated diversification rate, Impatiens would only have contained 1/5th of its current number of species, thereby indicating the rapid radiation of the genus. Ó 2009 Elsevier Inc. All rights reserved. 1. Introduction Global climate changes have played a crucial role in shaping species evolution either by increased rates of extinction, or by bursts of diversification. Earth’s climate has changed dramatically altered during the past 65 million years and likewise several lines of evidence have been found that correlate these fluctuations with an increase in speciation rate (Bowie et al., 2004; Erkens et al., 2007; Gamble et al., 2008; Hughes and Eastwood, 2006; Jaramillo et al., 2006; Johnson et al., 2006; Merckx et al., 2008; Plana et al., 2004; Richardson et al., 2001). Nevertheless, it is still being debated when the majority of the current angiosperm diversity has originated during the course of evolution (Davies et al., 2004). For a long time, it was considered that the recent climatic oscillations of the Pleistocene and Pliocene were crucial in promoting general diversification among tropical and subtropical species (Aubréville, 1962; Crowe and Crowe, 1982; Grey-Wilson, 1980a; Haffer, 1969; Mayr and O’Hara, 1986; Moritz et al., 2000; Prance, 1973). This hypothesis of recent speciation in the Quaternary is also referred to as the ‘refuge’ hypothesis and assumes allopatric speciation in populations as they became isolated from each other due to glacial aridity and rainforest fragmentation. Despite possible molecular evidence for recent explosive diversification events in some tropical lineages (Harris et al., 2000 on Aframomum; Rich* Corresponding author. Fax: +32 16 32 19 55. E-mail address: Steven.Janssens@bio.kuleuven.be (S.B. Janssens). 1055-7903/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.ympev.2009.04.013 ardson et al., 2001 on Inga), several recent studies suggest that these fluctuations played a much smaller role than previously thought and additionally suggest that the massive divergence of tropical species probably started already in the Tertiary or even earlier (Bowie et al., 2004; Glor et al., 2001; Knapp and Mallet, 2003; Moritz et al., 2000; Pennington et al., 2004; Plana et al., 2004; Zink and Slowinsky, 1995). The Pleistocene refuge hypothesis invokes the occurrence of a heterogeneous topography as cool and dry climate conditions forced rainforest taxa to withdraw into small refugial pockets (Haffer, 1969; Haffer and Prance, 2001; Mayr and O’Hara, 1986; Prance, 1973). In Africa for instance most of the postulated distribution areas of glacial forest refugia are assumed to be associated with more mountainous areas (Maley, 1987; Prance, 1982). In South America, the situation is more complex with several refuge areas being located in lowland Amazonia, yet also some uplifted areas are considered as possible refuges for the neotropical flora (Haffer, 1969, 1982; Prance, 1982). Moreover, endemic taxa are frequently associated with montane regions as geological and altitudinal differences create many habitat types (Grey-Wilson, 1980a; Plana et al., 2004). Investigating tropical or subtropical taxa which are often endemic to lowland or montane rainforests would make it possible to test the influence of glacial oscillations in light of the current species diversity in the tropics. In this aspect, balsams are of particular interest. Impatiens (Balsaminaceae) is among the most species-rich genera of flowering plants with over 1000 species and several newly described species every year (e.g. Janssens S.B. Janssens et al. / Molecular Phylogenetics and Evolution 52 (2009) 806–824 et al., in 2009; Pòcs, 2007). Diversity hotspots for the genus can be observed in tropical Africa, the Himalayan region, Madagascar, South India and Sri Lanka and Southeast Asia (Bhaskar and Razi, 1981; Chen, 2001; Grey-Wilson, 1980a, 1989; Rahelivololona et al., 2003; Shimizu, 1979; Toppin, 1920). Balsams are a characteristic element of tropical and subtropical montane forests above 500–800 m, even sometimes growing as high as 5000 m. Although there are notable exceptions, most species of Impatiens cannot endure persistent drought or extended exposure to direct sunlight (Fischer, 2004). As a result Impatiens species are typically confined to stream margins, waterside boulders, and wet and/or montane forests. For moisture-loving plants like Impatiens, upland rainforests are islands of suitable habitats in a vast landscape that is at least seasonally dry. By migrating along the mountain slopes with its moist habitat, Impatiens is able to minimize the effect of large climatic fluctuations. As a consequence, populations of Impatiens are easily isolated from other closely related lineages, which could result in an increase in the rate of species diversification. The present paper investigates the origin and evolution of the Balsaminaceae with special emphasis on tempo and timing of diversification in the genus Impatiens. We assess the evolutionary relationships between the major clades of Impatiens and date their origin in order to gain insight into causes of speciation. In addition, we examine whether the global cooling at the Miocene–Pliocene boundary and the subsequent climatic fluctuations during Pliocene and Pleistocene might have triggered adaptive radiation or whether the vast number of Impatiens lineages had already been largely established in the Tertiary. Furthermore, we provide additional understanding in the importance of the refuge theory as a possible explanation for the present tropical biodiversity. 2. Materials and methods 2.1. Taxon sampling Species of Impatiens were selected to represent the geographic and taxonomic diversity of the genus. In total 117 Impatiens species were included from which chloroplast (atpB-rbcL) and nuclear (introns 4 and 5 of ImpDEF1 and ImpDEF2) sequences have been produced. Janssens et al. (2006) demonstrated that Impatiens omeiana is sister to the other sampled species of Impatiens. Therefore, this species was chosen as outgroup for phylogenetic analyses of the remaining samples. A list of taxa with authorities, localities, voucher information and GenBank Accession Nos. is listed in Tables 1 and 2. 2.2. Molecular protocols and phylogenetic analysis Total genomic DNA was extracted using a modified version of the hot CTAB protocol (Janssens et al., 2006). Primers and temperature profiles used for the amplification of atpB-rbcL, ImpDEF1 and ImpDEF2 follow Janssens et al. (2006) and Janssens et al. (2007), respectively. Amplification reactions were carried out on a GeneAmp PCR system 9700 (Applied Biosystems). Cycle sequencing reactions were performed as in Janssens et al. (2006). Initial alignment of the DNA sequences was carried out with CLUSTALX, followed by manual adjustment using MacClade 4.05 (Maddison and Maddison, 2002). The chloroplast atpB-rbcL and the nuclear ImpDEF1/ImpDEF2 datasets were analyzed separately and combined using a maximum likelihood (ML) based method with RaxML (Stamatakis et al., 2005). A partition homogeneity test (implemented in PAUP4.0b10a; Swofford, 2002) was used to evaluate whether the data matrices provide different signal in the combined analyses. The most appropriate substitution model for each gene marker was determined using a series of likelihood ratio tests as implemented in ModelTest 3.06 (Posada and Crandall, 1998). 807 ML analyses were carried out on the CIPRES computer cluster (CyberInfrastructure for Phylogenetic RESearch) at the San Diego Supercomputing Center (www.phylo.org). Model parameters were computed with RAxML-VI HPC (Stamatakis, 2006). Two runs were performed, starting from a completely randomized tree, each with 100 inferences on the original alignment. Non-parametric bootstrapping (ML-BS) was carried out by calculating six times 167 replicates on the CIPRES computer cluster. Because a few nodes in the ML phylogram were collapsed and a fully resolved tree was necessary to calculate diversification rates and lineages-through-time plots, we set the branch length of these nodes collapsed to 0.000001, under the assumption that a lack of resolution indicated a rapid radiation (Harris et al., 2000; Hughes and Eastwood, 2006). 2.3. Molecular dating When estimating the age of the most recently diversified lineages in Impatiens, we encountered several problems which made it impossible to immediately calibrate the obtained Impatiens phylogeny. (1) Lack of a suitable fossil record for Impatiens made it impossible to directly calibrate the Impatiens topology. Although there is one fossil pollen record available for the genus (Impatiensiditis brevicolpus – Pliocene; Song et al., 2004) the assignment of this fossil pollen to clades in Impatiens was impossible. (2) Lack of resolution with the chloroplast atpB-rbcL spacer for the most recently diversified lineages. This issue could be solved by using the fastevolving introns 4 and 5 of ImpDEF1 and ImpDEF2, which are very effective to elucidate relationships at inter- and intraspecific levels (Janssens et al., 2007). However, because of their high nucleotide and indel substitution rate, it was impossible to amplify and sequence these loci for other taxa than Impatiens. Due to the high amount of missing data for ImpDEF1 and ImpDEF2, it was impossible to use these loci for the construction of the asterid phylogeny, which was applied estimate the age of the lineages towards Impatiens. As a result, we followed a two-stage strategy to calculate the divergence times for the most diversified branches in Impatiens. By extending the atpB-rbcL dataset of Impatiens with 319 atpB-rbcL accessions of asterid taxa from GenBank, we were able to cover almost all major lineages of this large subclass (Cornales – Cornaceae and Hydrangeaceae; Ericales – Actinidiaceae, Ebenaceae, Ericaceae, Lecythidaceae, Marcgraviaceae, Primulaceae, Sapotaceae, Styracaceae and Tetrameristaceae; Garryales – Garryaceae; Aquifoliales – Aquifoliaceae and Helwingiaceae; Gentianales – Apocynaceae, Gentianaceae, Loganiaceae and Rubiaceae; Lamiales – Gesneriaceae, Globulariaceae, Lamiaceae, Lentibulariacaeae, Orobanchaceae, Plantaginaceae and Scrophulariaceae; Solanales – Solanaceae; Apiales – Apiaceae and Araliaceae; Asterales – Asteraceae, Campanulaceae and Menyanthaceae; Dipsacales – Adoxaceae, Caprifoliaceae, Dipsacaceae and Valerianaceae). Two accessions from the rosid clade were used as outgroup. This large-scale approach allowed us to integrate multiple fossil calibration points in order to minimize the bias produced by a single calibration point (Merckx et al., 2008). The asterid dataset contains 437 taxa and 3880 characters and was analyzed with RAxML using similar settings as for the combined chloroplast-nuclear dataset of Impatiens. ML bootstrapping of the asterid dataset was also carried out as described above. A v2 likelihood ratio test, used to assess rate heterogeneity among lineages, indicated that the substitution rates in atpB-rbcL are not clocklike. As a result, we applied a relaxed clock model using the Penalized Likelihood method (PL; Sanderson, 2002) as implemented the r8s software package. The rooted ML phylogram was used as input file for the dating analysis. The following calibration points were used for age estimation: (1) the crown group of Diospyros, (Ebenaceae) constrained to a minimum of 54.9 mya corresponding to the Eocene age estimate of the fossil Austrodiospyros cryptostoma (Basinger and Christophel, 1985); (2) minimum age 808 S.B. Janssens et al. / Molecular Phylogenetics and Evolution 52 (2009) 806–824 Table 1 Accession Nos., voucher data and origin of plant material for taxa included in the combined DNA analyses of Impatiens. Taxon Impatiens Impatiens Impatiens Impatiens Impatiens Impatiens Impatiens Impatiens Impatiens Impatiens Impatiens Impatiens Impatiens Impatiens Impatiens Impatiens Impatiens Impatiens Impatiens Impatiens aquatilis Hook.f. arachnoides H.Perrier assurgens Baker f. aurea Muhl. aureliana Hook.f. auricoma Baill. balfourii Hook.f. balsamina L. bequaertii De Wild. begonifolia S. Akiyama and H. Ohba bicaudata H. Perrier bombycina Lobin and E. Fischer briartii De Wild. and Th. Dur. bururiensis Grey-Wilson burtonii Hook.f. var. burtonii campanulata Wight capensis Meerb. catati Baill. cecili N.E.Br. chevalieri Tardieu Impatiens chinensis L. Impatiens chungtienensis Y.L. Chen Impatiens clavicornu Turcz. Impatiens columbaria J.J. Bos Impatiens conchibracteata Y.L. Chen Impatiens confusa Grey-Wilson ssp. longicornu Grey-Wilson. Impatiens congolensis G.M. Schulze and R. Wilczek Impatiens cuspidata Wight and Arn. Impatiens cyanantha Hook.f. Impatiens davidi Franch. Impatiens desmantha Hook.f. Impatiens digitata Warb. ssp. digitata Impatiens eberhardtii Tardieu Impatiens edgeworthii Hook.f. Impatiens Impatiens Impatiens Impatiens Impatiens Impatiens engleri Gilg ssp. engleri eryaleia Launert ssp. eryaleia ethiopica Grey-Wilson faberi Hook.f. fenghwaiana Y.L. Chen fischeri Warb. Impatiens flaccida Arn. Impatiens flanaganae Hemsl. Impatiens forrestii Hook.f. ex W.W. Smith Impatiens furcata H. Perrier Impatiens glandulifera Arn. Impatiens gordoni Horne Impatiens hamata Warb. Impatiens hawkeri W. Bull Impatiens hians Hook.f. Impatiens henslowiana Arn. Impatiens hochstetteri Warb. ssp. hochstetteri Impatiens hochstetteri Warb. ssp. hochstetteri Impatiens hoehnelii T.C.E. Fr. Impatiens hydrogetonoides Launert ssp. hydrogetonoides Impatiens imbecilla Hook.f. Impatiens inaperta H. Perrier Impatiens ioides G.M. Schulze Impatiens irvingii Hook.f. Impatiens kamerunensis Warb. ssp. kamerunensis Impatiens kamerunensis Warb. ssp. obanensis (Keay) Grey-Wilson Impatiens keilii Gilg ssp. keilii Impatiens keilii Gilg ssp. pubescens Grey-Wilson Impatiens kerriae Craib Origin Voucher Accession No. ImpDEF1 Accession No. ImpDEF2 Accession No. atpB-rbcL China, Yunnan Madagascar Zambia North American origin, cult. Holden arboretum China, Yunnan Comoros origin, cult. Bot. Gard. Marburg Himalayan origin, cult. Denver Bot. Gard. Indian origin, cult. Kruidtuin Leuven Congo origin, cult. Bot. Gard. Koblenz Univ. China, Yunnan Madagascan origin, cult. by Ray Morgan, UK African origin, cult. Bot. Gard. Koblenz Univ. Zambia Burundi Uganda South Indian origin, cult. by Ray Morgan, UK North American origin, cult. Holden arboretum Madagascar Zimbabwe Vietnam, Balat Song CNY017 (NEU) Janssens SJ010 (LV) Dessein SD720 (BR) Janssens SJ008 (LV) Yuan CN2k1-56 (NEU) Janssens SJ001 (LV) 33051 (DBG) Janssens SJ003 (LV) Fischer NE1 (NEU) Yuan CN2k1-51 (NEU) Ray Morgan s.n. (LV) Fischer NE3 (NEU) Dessein 1018 (BR) Reekmans 8110 (BR) Knox 2803 (LV) Ray Morgan s.n. (LV) Janssens SJ009 (LV) Janssens SJ011 (LV) Knox 4353 (LV) Song and Puoong 2004–07 (NEU) Yuan CN2k1-49 (NEU) Yuan CN2k2-204 (NEU) Ray Morgan s.n. (LV) FB/S2966 (BR) Hao 427 (NEU) Knox 3239 (LV) – FJ826680 FJ826681 EF133560 EF133561 EF133562 EF133564 EF133563 FJ826682 EF133565 EF133566 FJ826683 FJ826684 FJ826685 EF133567 EF133568 FJ826686 FJ826687 – EF133611 – FJ826735 EF133613 EF133614 EF133615 EF133616 EF133617 FJ826736 EF133619 EF133618 FJ826737 FJ826738 FJ826739 FJ826740 EF133620 EF133621 – FJ826741 FJ826742 DQ147811 FJ826628 FJ826629 DQ147813 DQ147814 DQ147815 DQ147817 DQ147816 FJ826630 DQ147819 FJ826631 FJ826632 FJ826633 FJ826885 DQ147822 DQ147823 FJ826634 FJ826635 FJ826636 EF133569 EF133570 EF133571 EF133572 EF133573 FJ826688 EF133622 EF133623 EF133624 EF133625 EF133626 FJ826743 DQ147825 DQ147826 DQ147827 DQ147828 DQ147829 FJ826637 Fischer NE7 (NEU) Ray Morgan s.n. (LV) Yuan CN2k1-84 (NEU) Yuan CN2k-09 (NEU) Yuan CN2k-30 (NEU) Knox 3653 (LV) Song s.n. (NEU) 89.2005 (UC) EF133574 EF133575 EF133576 EF133578 FJ826689 EF133579 EF133580 EF133627 EF133628 EF133629 EF133630 – FJ826744 EF133632 EF133633 DQ147830 DQ147832 DQ147833 DQ147835 DQ147837 FJ826638 DQ147839 DQ147840 Knox 3627 (LV) Knox 3262 (LV) Jansen 5505 (WAG) Song S007 (NEU) Yuan CN2k-41 (NEU) Menn. and Bar.-Kui. 284 (U) FB/S3925 (BR) 19860179 (E) FJ826690 – – EF133581 – FJ826691 FJ826745 FJ826746 – EF133634 EF133635 – FJ826639 FJ826640 FJ826641 DQ147841 DQ147842 DQ147843 EF133582 – EF133636 FJ826747 DQ147845 DQ147846 Yuan CN2k-79 (NEU) Fischer EF8 (NEU) Janssens SJ002 (LV) EF133583 FJ826692 EF133584 EF133637 FJ826748 EF133638 DQ147847 FJ826642 DQ147848 Kew 2398 Knox 3558 (LV) Janssens SJ006 (LV) Schwerdtfeger 9492a (B) Fischer NE10 (NEU) Knox 2633 (LV) Knox 4355 (LV) Knox 2846 (LV) Dessein SD719 (BR) FJ826693 FJ826694 EF133586 EF133585 FJ826695 FJ826696 FJ826697 FJ826698 FJ826699 FJ826749 FJ826750 EF133640 EF133639 FJ826751 FJ826752 FJ826753 FJ826754 FJ826755 FJ826643 FJ826644 DQ147850 DQ147849 FJ826645 FJ826646 FJ826647 FJ826648 China, Sichuan Madagascar Tanzania Gabon Cameroon Ghana Hao 426 (NEU) JBE2 (NEU) Knox 3571 (LV) Ngok Banak 2022 (WAG) J.J.F.E. de Wilde 8638 (BR) Jongkind 1926 (WAG) EF133587 FJ826700 FJ826701 FJ826702 FJ826703 FJ826704 EF133641 – FJ826756 FJ826757 FJ826758 FJ826759 DQ147851 FJ826649 FJ826650 FJ826651 FJ826653 FJ826652 Tanzania Tanzania Thailand, Qingmai Knox 3575 (LV) Knox 3587 (LV) Chassot 99–238 (NEU) FJ826705 FJ826706 – FJ826760 FJ826761 EF133642 FJ826654 FJ826655 DQ147853 China, Yunnan China, Yunnan South Indian origin, cult. by Ray Morgan, UK African origin, cult. Nat. Bot. Gard. Meise China, Yunnan Tanzania African origin, cult. Bot. Gard. Koblenz Univ. South Indian origin, cult. by Ray Morgan, UK China, Yunnan China, Fujian China, Yunnan Tanzania Vietnam, Anam Himalayan origin, cult. Univ. California Bot. Gard. Berkeley Tanzania Tanzania Ethiopia China, Sichuan China, Guangxi Kenya South Indian origin, cult. Nat. Bot. Gard. Meise South-African origin, cult. Roy. Bot. Gard. Edinburgh China, Yunnan Madagascar Belgium, Leuven (introduced; Himalayan origin) Seychelles Tanzania New Guinean origin West African origin, cult. Bot. Gard. Berlin Sri Lanka origin, cult. Bot. Gard. Koblenz Univ. Kenya South Africa Kenya Zambia 809 S.B. Janssens et al. / Molecular Phylogenetics and Evolution 52 (2009) 806–824 Table 1 (continued) Taxon Origin Voucher Accession No. ImpDEF1 Accession No. ImpDEF2 Accession No. atpB-rbcL Impatiens kilimanjari Oliv. Impatiens latifolia L. Impatiens lukwangulensis Grey-Wilson Impatiens mackeyana Hook.f. ssp. claeri (N.Hallé) Grey-Wilson Impatiens mackeyana Hook.f. ssp. zenkeri (Warb.) Grey-Wilson Impatiens macroptera Hook.f. Impatiens mannii Hook.f. Tanzania South Indian origin, cult. by Ray Morgan, UK Tanzania Gabon origin, cult. Nat Bot. Gard. Koblenz Univ. Knox 3652 (LV) Ray Morgan s.n. (LV) Knox 3570 (LV) Fischer EF5 (NEU) FJ826707 EF133588 EU723716 FJ826708 EF133643 EU723723 FJ826762 FJ826656 DQ147854 EU723709 FJ826657 African origin, cult. Bot. Gard. Koblenz Univ. Fischer EF21 (NEU) FJ826709 – DQ147857 Equatorial Guinea East Congo FJ826710 FJ826711 FJ826763 FJ826764 FJ826658 FJ826659 Impatiens mannii Hook.f. Impatiens mazumbaiensis Grey-Wilson Impatiens mengtseana Hook.f. Impatiens meruensis Gilg ssp. cruciata (T.C.E. Fr.) Grey-Wilson Impatiens meruensis Gilg ssp. meruensis Impatiens mildbraedii Gilg telekii (T. C.E.Fr.) Grey-Wilson Impatiens monticola Hook.f. Impatiens nana Engl. Impatiens napoensis Y.L. Chen Impatiens niamniamensis Gilg Impatiens noli-tangere L. Impatiens omeiana Hook.f. Cameroon Tanzania China, Yunnan Kenya de Wilde 12014 (WAG) Geuven Carvalho 4534 (BR) Merckx s.n. (LV) Knox 3643 (LV) Yuan CN2k1-38 (NEU) Ray Morgan s.n. (LV) FJ826712 EU723717 EF133589 FJ826713 FJ826765 EU723724 EF133644 FJ826766 FJ826660 EU723710 DQ147858 FJ826661 Tanzania Kenya Knox 3328 (LV) Knox 3882 (LV) FJ826714 FJ826715 FJ826767 FJ826768 FJ826662 FJ826663 China, Sichuan Tanzania China, Yunnan African origin, cult. Nat. Bot. Gard. Meise Korea China, Sichuan, cult. Univ. California Bot. Gard. Berkeley Tanzania Hao 425 (NEU) Knox 2589 (LV) Yuan CN2k1-61 (NEU) FB/S2590 (BR) Janssens SJ008 2002.0214 (UC) EF133591 EU723719 FJ826716 FJ826717 – EF133592 EF133646 EU723726 FJ826769 FJ826770 – EF133647 DQ147860 EU723712 DQ147861 DQ147862 FJ82664 DQ147864 Knox 3572 (LV) FJ826718 – FJ826665 Tanzania Gabon Indonesian origin, cult. by Ray Morgan, UK Belgium, Leuven (introduced; Asian origin) Tanzania Bali, Indonesian origin, cult. by Ray Morgan, UK Vietnam, Anam Tanzania Gabon African origin, cult. Roy. Bot. Gard. Edinburgh Tanzania China, Yunnan Tanzania China, Yunnan Tanzania Knox 3552 (LV) Jongkind 5724 (WAG) Ray Morgan s.n. (LV) Janssens SJ004 (LV) Knox 4420 (LV) Ray Morgan s.n. (LV) Song s.n. (NEU) Knox 3553 (LV) de Wilde 10390 (WAG) 19680124 (E) Knox 3229 (LV) Song Y007 (NEU) Knox 2592 (LV) Yuan CN2k1-26 (NEU) Knox 3462 (LV) EU723720 FJ826719 FJ826720 EF133593 FJ826721 EF133594 EF133595 FJ826722 FJ826723 EF133597 FJ826724 EF133598 FJ826725 EF133599 FJ826726 FJ826772 FJ826771 EF133648 FJ826773 EF133649 EF133650 FJ826774 FJ826775 – FJ826776 EF133652 FJ826777 EF133653 – EU723713 FJ826666 FJ826667 DQ147866 FJ826668 DQ147868 DQ147869 FJ826669 FJ826670 DQ147871 FJ826671 DQ147872 FJ826672 DQ147874 FJ826673 Tanzania China, Yunnan Cameroon Himalayan origin, cult. Holden arboretum Tanzania China, Yunnan Tanzania Zimbabwe China, Yunnan African origin, cult. Bot. Gard. Koblenz Univ. Tanzania Kenya Knox 3667 (LV) Yuan CN2k1-44 (NEU) Merckx VM125 (LV) 941314 (DBG) Knox 3551 (LV) Yuan CN2 k-80 (NEU) Knox 3318 (LV) Knox 4324 (LV) Yuan CN2 k-57 (NEU) Fischer NE18 Knox 3576 (LV) Knox 2639 (LV) EU723721 EF133600 FJ826727 EF133601 FJ826728 – FJ826729 EU723722 EF133602 – FJ826730 - EU723727 EF133654 FJ826778 EF133655 FJ826779 FJ826780 FJ826781 EU723728 EF133656 FJ826782 FJ826783 FJ826784 EU723714 DQ147876 FJ826666 DQ147877 FJ826674 – FJ826675 EU723715 DQ147882 DQ147883 FJ826676 FJ826677 China, Yunnan Madagascan origin, cult. Bot. Gard. Univ. Kopenhagen Kenya Tanzania China, Yunnan China, Taiwan African origin, cult. Bot. Gard. Koblenz Univ. South Indian origin, cult. by Ray Morgan, UK Tanzania African origin, cult. Nat. Bot. Gard. Meise China, Yunnan Yuan CN2k1-68 (NEU) Janssens SJ005 (LV) EF133603 EF133604 EF133657 EF133658 DQ147885 DQ147886 Knox 2776 (LV) Knox 3554 (LV) Yuan CN2k2-173 (NEU) Zhengyu Jiang T1 (NEU) Fischer NE20 (NEU) Ray Morgan s.n. (LV) Knox 4117 (LV) S3926 (BR) Yuan CN2k1-55 (NEU) FJ826731 FJ826732 EF133605 EF133606 EF133608 FJ826733 EF133609 FJ826734 FJ826785 FJ826786 EF133659 EF133660 EF133661 EF133663 – EF133664 FJ826787 FJ826678 DQ147887 DQ147888 DQ147889 DQ147891 FJ826679 DQ147892 DQ147894 Impatiens pallide-rosea Gilg ssp. lupangaensis (G.M. Schulze) Grey-Wilson Impatiens pallide-rosea Gilg ssp. pallide-rosea Impatiens palpebrata Hook.f. Impatiens parasitica Bedd. Impatiens parviflora DC. Impatiens percordata Grey-Wilson ssp. percordata Impatiens platypetala Lindl. Impatiens poilanei Tardieu Impatiens pseudohamata Grey-Wilson Impatiens pseudomacroptera Grey-Wilson Impatiens pseudoviola Gilg. Impatiens pseudozombensis Grey-Wilson Impatiens purpurea Hand.-Mazz. Impatiens raphidothrix Warb. Impatiens rectangula Hand.-Mazz. Impatiens rubromaculata Warb. ssp. imagiensis Grey-Wilson Impatiens rubromaculata Warb. ssp. rubromaculata Impatiens rubrostriata Hook.f. Impatiens sakeriana Hook.f. Impatiens scabrida DC. Impatiens serpens Grey-Wilson Impatiens siculifer Hook.f. Impatiens sodenii Engl. Impatiens sylvicola Burtt Davy and Greenway Impatiens taronensis Hand.-Mazz. Impatiens teitensis Grey-Wilson Impatiens thamnoidea G. M. Schulze Impatiens tinctoria A. Rich ssp. elegantissima (Gilg) Grey-Wilson Impatiens trichosepala Y.L. Chen Impatiens tuberosa H. Perrier Impatiens Impatiens Impatiens Impatiens Impatiens Impatiens Impatiens Impatiens Impatiens tweedieae E. A. Bruce ulugurensis Warb. uliginosa Franch. uniflora Hayata usambarensis Grey-Wilson viscida Wight volkensii Warb. walleriana Hook.f. yingjiangensis S.Akiyama and H.Ohba of the crown group of Asclepiadoideae (Apocynaceae) constrained to 38 mya on the basis of the earliest fossil of this group from the Oligocene and Miocene (Muller, 1981); (3) a minimum age constraint of 87 mya to the Cornaceae based on the fossil fruits of Hiro- noia fusiformis that were found in the Coniacian stratigraphic zone (Takahashi et al., 2002; Anderson et al., 2005); (4) minimum age constraint of 42 mya for Sambucus (Adoxaceae) based on the finding of fossil endocarps from the late Eocene to Pliocene found in 810 S.B. Janssens et al. / Molecular Phylogenetics and Evolution 52 (2009) 806–824 Europe (Reid and Chandler, 1926); (5) based on reports of fossil fruits of Fraxinus (Oleaceae), occurring in America from the Eocene onwards, a minimum age for this genus was constrained to 54.9 mya (Call and Dilcher, 1992); (6) minimum age of Weigela (Caprifoliaceae) constrained to 24.6 mya based on reports of distinctive winged seeds from the Miocene and Pliocene (LancuckaSrodoniowa, 1967); (7) the crown node of Solanaceae is constrained to a minimum age of 54.9 mya, corresponding to the earliest Solanaceae fossil known from the Eocene (Collinson et al. (1993)); (8) recent estimation of divergence times based on ndhF and rbcL suggest that Asteraceae have originated in the mid Eocene (Kim et al., 2005). As a result, the crown node of this clade was constrained at a minimum age of 47 mya; (9) the lineage towards Valeriana (Valerianaceae) is constrained at a minimum age of 11.2 mya, corresponding with the report of fossil fruits from the late Miocene and Pliocene in Europe (Bell and Donoghue, 2005); (10) minimum age of Ixora (Rubiaceae) constrained to 5.3 mya based on the fossil collection of I. casei from the Marshall Islands (Leopold, 1969); (11) in addition, we assigned a fixed age of 128 mya to crown node of the asterids, a calibration point estimated by Bremer et al. (2004). In order to determine the impact of the different fossil calibration points on the overall dating estimation, we used the method of ‘fossil cross-validation’ developed by Near and Sanderson (2004). Their approach evaluates the difference between the fossil ages and the estimated molecular ages by re-evaluating the dating analysis with only a single calibration point. By calculating the summed square value (SS) between fossil constraint and molecular estimate, we noticed that calibration points 1 and 2 exhibited the highest scores (Fig. 1). Removal of these calibration points from our dataset resulted in a significant decline of the average squared deviation (s), whereas subsequent removal of other fossil constraints did not alter s (Fig. 2). Consequently, we removed fossil constraints 1 and 2, as they were determined as most inconsistent calibration points for the asterid dataset. Confidence intervals for each node were obtained as described in the r8s manual. PHYLIP (Felsenstein, 1995) was used to generate 100 bootstrapped datasets of the original dataset. For each bootstrapped data matrix, we generated a phylogram with the RAxML procedure. Using the PL method in r8s 1.70, we recalculated divergence times for each of the phylograms using the same calibration points as above. These results were then used to estimate standard deviations for each node. The obtained age estimate for the split between I. omeiana and the remainder of the genus was then applied as fixed calibration point to calculate the divergence events within Impatiens. The boundaries of the confidence interval for this node were taken into account in the secondary calibration approach of Impatiens. Hence we analyzed the Impatiens dataset for an age of 22.5, 22.5 + 5.6 and 22.5 5.6 million years ago. In addition to these indirectly obtained age estimates, we used the following biogeographic age estimates for cross-validation: The age of Mt Kilimanjaro, a recent volcanic mountain dated at 1.1 mya harbouring the endemic species Impatiens digitata ssp. digitata and Impatiens kilimanjari (Knox and Palmer, 1998). To infer age estimates for the most diversified lineages in Impatiens, we used the combined chloroplast and nuclear data matrix. A likelihood ratio test showed that a clocklike evolution for the ImpDEF1/ImpDEF2 dataset also had to be rejected. Consequently, we calculated the divergence times on the combined data using the penalized likelihood method (Sanderson, 2002) onto the rooted ML phylogram. 2.4. Ancestral area distribution In order to reconstruct biogeographic ancestral areas, we used the likelihood reconstruction method (Pagel, 1999; Schluter et al., 1997) implemented in Mesquite v1.12 (Maddison and Maddison, 2008). The geographic distribution of Impatiens has been coded using major geographic units: Africa, Madagascar, Southwest China, the Himalayas, India and Sri Lanka, North America, Southeast Asia. A Markov k-state one-parameter model (Mk1), assuming a single rate of transition between two character states was applied for ML ancestral area reconstruction. In order to determine the best estimate of the reconstructed characters state at each node, a likelihood ratio test was used with a likelihood decision limit of 2.0. With log-likelihoods differing by 2 or more, the lowest negative log-likelihood was regarded as best value, whereas nodes with log-likelihoods differences lower than two were assumed to be ambiguous. 2.5. Speciation rates and timing of diversification in the genus Impatiens One aspect of this study is to examine whether the timing of speciation in Impatiens is correlated with climate changes during the Pliocene–Pleistocene. By generating a lineages-through-time (LTT) plot for the PL chronogram derived from the combined Impatiens dataset, we obtain a schematic visualization of the net diversification rate of the genus Impatiens. The outline of the LTT plot is used to evaluate whether changes in diversification rate have occurred during the history of Impatiens. Only one individual per species was included in the LTT plot as the inclusion of individuals representing additional subspecies would inflate the speciation rate. Present sampling only contains about 11% of all known Impatiens species (113 out of 1000 species). As a result, we tested the consequence of an incomplete taxon sampling on the profile of the empirical LTT plot by generating simulated phylogenies with the software program Phylogen 1.1 (Rambaut, 2002). The following simulation was executed. A total of 1000 replicate phylogenies were generated, containing 1000 extant taxa of which 884 were randomly pruned from each tree. Each simulation was generated under the assumption of a constant birth–death rate (b/d = 0). Branch lengths of the resulting phylograms were rescaled with TreeEdit 1.0 (Rambaut and Charleston, 2002) to 22.5 million years, the estimated crown node age of Impatiens. We subsequently evaluated the empirical LTT curve by comparison with the mean LTT simulation and its 95% confidence interval. In order to assess a quantitative method that could illustrate an overall change in diversification rate within the evolutionary history of Impatiens, the rate of speciation was calculated for each lineage ([ln(N) ln(N0)/T]; Baldwin and Sanderson, 1998). Additionally, we used the constant rate (CR) test of Pybus and Harvey (2000) to investigate a possible increase in overall diversification rate within Impatiens. The CR method as implemented in Genie 3.0 (Pybus and Harvey, 2000) follows an approach of c-statistics in which the relative position of the nodes in the empirical phylogeny is compared to the position of those nodes that are assumed to be under a model of constant diversification. Positive c-values (c > 0) indicate an apparent acceleration of the overall diversification rate, assuming that nodes are closer to the tips than expected under a pure-birth–death model (c = 0). On the contrary, negative c-values (c < 0) display a significant decrease in diversification rate. The null hypothesis of constant birth–death rate can be rejected when c < 1.645 (one-tailed test). Because the c-statistics is biased by extinction and incomplete taxon sampling, we simulated pure-birth topologies (d/b = 0) using Phylogen 1.1 (Rambaut, 2002). Analyses were carried out under the assumption that our Impatiens sampling (n = 116) only represents a small fraction (f) of the real diversity of the genus. We simulated 1000 pure-birth topologies with n/f tips, with the f-scores varying from 0.06 to 1. The resulting trees were then pruned to 117 taxa and for each of the pruned trees the c-statistics was computed using Genie 3.0 811 S.B. Janssens et al. / Molecular Phylogenetics and Evolution 52 (2009) 806–824 Table 2 GenBank accessions of atpB-rbcL spacer sequences of the asterid used in the large age estimation analysis. Species Accession No. Family Order Kirengeshoma palmata Cornus mas Actinidia chinensis Hydrocera triflora Impatiens Diospyros kaki Euclea pseudebenus Lissocarpa guianensis Acrothamnus hookeri Acrothamnus maccraei Acrotriche ramiflora Andersonia sprengelioides Andromeda polifolia Androstoma empetrifolia Archeria comberi Archeria serpyllifolia Astroloma xerophyllum Brachyloma daphnoides Brachyloma preissii Bruckenthalia spicufolia Budawangia gnidioides Chamaedaphne calyculata Coleanthera myrtoides Conostephium preissii Cosmelia rubra Croninia kingiana Cyathodes platystoma Cyathodes pumila Cyathopsis floribunda Daboecia cabtabrica Diplycosia cinnamomifolia Dracophyllum longifolium Epacris obtusifolia Erica hispidula Gaultheria tomentosa Harrimanella hypnoides Kalmia latifolia Leptecophylla tameiameiae Leucopogon milliganii Leucothoe fontanesiana Lissanthe synandra Lyonia lucida Lysine ciliatum Melichrus urceolatus Melichrus urceolatus Monotoca elliptica Oligarrhena micrantha Pentachondra involucrata Pernettya tasmanica Planocarpa petiolaris Prionotes cerinthoides Rhododendron simsii Richea pandanifolia Rupicola sprengelioides Sphenotoma dracophylloides Styphelia tubiflora Trochocarpa laurina Woollsia pungens Zenobia pulverulenta Corytophora alta Eschweilera simiorum Marcgravia maguirei Marcgravia umbellata Norantea guianensis Souroubea sp. Primula cuneifolia Chromolucuma rubriflora Chrysophyllum imperiale Eberhardtia aurata Elaeolum glabrescens Changiostyrax dolichocarpa Pterostryrax psilophyllus Sinojackia sarcocarpa Pelliciera rhizophorea Aucuba japonica EF437413 X83988 FJ866479 DQ147895 DQ147810–DQ147894 X91004 FJ866480 AF421094 AY971370 AY971371 AY636036 AF155843 AF366584 AY372540 AF155840 AY971372 AY372554 AF155859 AY372555 AY520757 AF155852 AF366616 AY372556 AY372557 AF155842 AF208750 AY372560 AY372558 AY636040 AY520758 AF366587 AF155845 AY636042 AY520767 AF366611 AF155833 AB247970 AY372569 AY372575 AF366612 AY636038 AF155836 AF155848 AY372595 AY372595 AY005085 AF155854 AY005087 AF155835 AY372594 AF155838 AM296078 AF155844 AF155851 AF155846 AY372591 AY005092 AF155847 AF366615 AF076769 X91001 DQ147896 DQ147897 DQ147898 DQ147900 AB003575 EF558591 EF558592 EF558594 EF558593 DQ317984 DQ317980 DQ317979 DQ147899 AB087779 Hydrangeaceae Cornaceae Actinidiaceae Balsaminaceae Balsaminaceae Ebenaceae Ebenaceae Ebenaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Ericaceae Lecythidaceae Lecythidaceae Marcgraviaceae Marcgraviaceae Marcgraviaceae Marcgraviaceae Primulaceae Sapotaceae Sapotaceae Sapotaceae Sapotaceae Styracaceae Styracaceae Styracaceae Tetrameristaceae s.l. Garryaceae Cornales Cornales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Ericales Garryales Superorder Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Basal Asterids Euasterids 1 (continued on next page) 812 S.B. Janssens et al. / Molecular Phylogenetics and Evolution 52 (2009) 806–824 Table 2 (continued) Species Accession No. Family Order Superorder Garrya ovata Alstonia scholaris Dischidia astephana Gunessia pepo Hoya kentiana Madangia inflata Marsdenia carvalhoi Micholitzia obcordata Gentiana oreodoxa Mitrasacme pygmaea Mitreola petiolata Adina rubella Alberta minor Argostemma hookeri Asperula cynanchica Augusta longifolia Aulacocalyx jasminiflora Belonophora coriacea Bertiera dewevrei Cephalanthe occidentalis Chiococca alba Chomelia brasialiana Cinchona pubescens Coffea congensis Colletoecema dewevrei Commitheca liebrechtis Coptosapelta diffusa Coryanthe mayumbensis Cosmocalyx spectabilis Coussarea hirticalyx Crossopteryx febrifuga Cruciata glabra Cruckshansia hymencodon Cubanola domingensis Dictyandra arborescens Didymosalpinx lanciloba Diplospora dubia Duperrea pavettifolia Ecpoma hierniana Eosanthe cubensis Erithalis fruticosa Euclinia longiflora Exostema lineatum Faramea porophylla Fernelia buxifolia Gaertnera sp. Galium scabrum Gardenia imperialis Guettarda platypoda Hayataella michelloides Hedyotis littoralis Heinsia crinita Hilllia triflora Hoffmannia refulgens Hydnophytum formicarum Isertia rosea Ixora coccinea Ixora finlaysonia Keetia multiflora Kellogia galioides Ladenbergia sp. Leptactina leopoldi-secundi Luculia pinceana Manettia cordifolia Mitragyne inermis Morelia senegalensis Morinda lucida Mycetia malayana Myrmecoda platytyrea Myrmephytum beccarii Otiophora scabra Paratriaina xerophila Pausinystalia johimbe Pavetta ternifolia Pentas lanceolata Phialanthus stillans AB087788 DQ359161 DQ334576 DQ334570 DQ334564 DQ334583 DQ334563 DQ334601 DQ398621 DQ131694 DQ131696 DQ131698 DQ131699 AJ234032 X81689 DQ131703 DQ131704 DQ131706 DQ131707 DQ131710 DQ131711 DQ131712 AJ233990 AM412393 DQ131713 AJ233999 AJ233987 DQ131715 DQ131716 DQ131717 DQ131719 X81097 AJ234004 DQ131720 DQ131723 DQ131724 DQ131725 DQ131726 DQ131727 DQ131729 DQ131730 DQ131731 DQ131732 AJ234008 DQ131736 AJ234012 X76462 DQ131737 DQ131739 AB247242 AJ234027 DQ131740 AJ233993 X81684 X76480 DQ131743 AM412400 DQ131744 DQ131745 AY570768 DQ131747 DQ131748 DQ131749 AJ234023 DQ131751 DQ131752 DQ131753 AJ234033 AB044151 AB044152 DQ131756 DQ131759 DQ131760 DQ131761 AB247149 DQ131762 Garryaceae Apocynaceae Apocynaceae Apocynaceae Apocynaceae Apocynaceae Apocynaceae Apocynaceae Gentianaceae Loganiaceae Loganiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Garryales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 813 S.B. Janssens et al. / Molecular Phylogenetics and Evolution 52 (2009) 806–824 Table 2 (continued) Species Accession No. Family Order Plocama pendula Pomax umbellata Portlandia grandiflora Pouchetia africana Pseudosabicea floribunda Psilanthus leroyi Psychotria undata Putoria calabrica Remijia macronemia Rothmannia sp. Rubia tinctorum Rudgea bremekampiana Rutidea schlechteri Sabicea diversifolia Sacosperma paniculatum Salzmannia axillaris Sarcocephalus latifolius Schumanniophyton magnificum Serissa foetida Spermacoce assurgens Stelechante cauliflora Stipularia elliptica Tricalysia elliotii Trichostachys perfoliatum Urophyllum glabrum Valantia muralis Vangueria infausta Vangueria infausta Wendlandia sp. Achimenes admirabilis Aeschynanthus longiflorus Alsobia dianthiflora Besleria melancholica Chirita longgangensis Chirita spadiciformis Chrysothemis pulchella Cobanantha calochlamys Codonantha serrulata Columnea sanguinea Cyrtandra cupulata Didissandra frutescens Didymocarpus citrinus Didymocarpus cordatus Emarhendia bettiana Episcia cupreata Epithema taiwanese Fieldia australis Haberlea rhodopensis Henckelia albomarginata Jovellana punctata Koellikeria erinoides Kohleria eriantha Lenbrassia australiana Loxonia hirsutum Mitraria coccinea Monophyllae horsfieldii Nautilocalyx melittifolius Nenatanthus villosus Paliavana sericiflora Paraboea capitata Petrocosmea nervosa Primulina tabacum Ramonda myconi Rhabdothamnops sinensis Rhynchoglossum obliquum Rhytiophyllum tomentosumbesleria Saintpaulia velutina Sarmiaeta scandens Sinningia bulbosa Stauranthera grandiflora Streptocarpus holstii Streptocarpus rexii Titanotrichum oldhamii Trisepalum speciosum Vanhouttea gardneri Whytockia sasakii AJ234035 DQ131767 DQ131768 DQ131770 DQ131771 AM412399 DQ131774 X81672 DQ131775 DQ131776 X76474 DQ131778 DQ131779 DQ131781 DQ131782 DQ131784 DQ131785 DQ131786 AJ234034 X81679 DQ131789 DQ131790 DQ131791 AF446995 DQ131793 X76473 DQ131794 DQ131794 DQ131796 AJ439982 AJ490920 AJ490924 AJ490923 AJ490903 AJ490904 AY423115 AJ490926 AJ439981 AJ490927 AJ490886 U91313 AJ490906 AJ490907 AJ490908 AJ490928 AY423117 AY423112 AJ490909 AJ490910 AY423109 AJ439983 AY423114 AJ490921 AJ490891 AY423113 U91315 AJ439984 AJ439980 AJ439963 AJ490911 AJ490912 AJ490913 AJ490914 AJ490915 AJ490898 AJ490930 AJ490916 AJ490922 AJ439910 AJ490900 AJ490917 AJ490918 AY423111 AJ490919 AJ439974 AY423116 Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Rubiaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gesneriaceae Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Gentianales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Superorder Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 Euasterids 1 (continued on next page) 814 S.B. Janssens et al. / Molecular Phylogenetics and Evolution 52 (2009) 806–824 Table 2 (continued) Species Accession No. Family Order Superorder Globularia salicina Phlomis crinita Utricularia gibba Fraxinus excelsior Plantago media Antirrhinum majus Calceolaria arachnoidea Digitalis purpurea Gratiola aurea Paulownia tomentosa Scrophularia canina Tetranema mexicanum Verbascum speciosum Veronica incana Aureliana fasciculata Capsicum baccatum Datura stramonium Jaltomata auriculata Lycianthe lenta Lycopersicon pimpinellifolium Nicandra physalodes Schizanthus x wisetonensis Solanum shanesii Tubocapsicum anomalum Wythania coagens Chaerophyllum nodosum Myrrhis odorata Oreomyrrhis azorellaceae Osmorhiza longistylus Aralia cachemirica Brassaiopsis shweliensis Eleutherococcus trifoliatus Fatsia japonica Harmsiopanax aculeata Hedera maderensis Heteropanax fragrans Macropanax dispermus Merilliopanax cordifolius Panax assamicus Polyscias javanica Schleffera hypoleucoides Tupidanthus calyptratus Helwingia japonica Ilex discolor Achillea millefolium Bellis perennis Ligularia franchetiana Apetahia margaretae Brighamia insignis Burmeistera crispiloba Centropon gutierrezii Clermontia fauriei Cyanea pilosa Delissea subcordata Isotoma axillaris Lobelia hypoleuca Nymphoides cordata Sclerotheca jayorum Trematolobelia macrostachys Adoxa moschatellina Sambucus racemosa Sinadoxa corydalifolia Tetradoxa omeiensis Viburnum acerifolium Leycesteria formosa Lonicera sempervirens Symphoricarpos orbiculatus Triosteum perfoliatum Triplostegia glandulifera Diervilla sessilifolia Weigela hortensis Dipsacus mitis Knautia macedonica Scabiosa columbaria Heptacodium miconioides Abelia x grandiflora AY818898 AY792758 EF529719 AY911655 AY818903 AJ490883 AY423108 AY818897 EF529725 AY423104 AY423105 AJ490884 AJ490885 AY818908 AF397083 AF397101 AF397076 AF397081 AF397093 AF397079 AJ490882 AY423103 AF397088 AF397082 AF397084 DQ829733 DQ829738 DQ829705 DQ829737 AY753224 AY753233 AY753235 AY163545 AY753251 AY163533 AY753236 AY753241 AY753243 AY753245 AY753252 AY753247 AY753250 X94941 AF471629 EU129084 X91000 AB375447 DQ285286 DQ285256 DQ285281 DQ285282 DQ285259 DQ285291 DQ285264 DQ285283 DQ285266 EF529716 DQ285273 DQ285271 AF446990 AF446988 AF446989 AF446991 AF446987 EU265521 EU265563 AF446994 AF446995 AF447009 AF446997 AF446998 AF447007 AY362501 AF447008 AF446996 AF446999 Globulariaceae Lamiaceae Lentibulariaceae Oleaceae Plantaginaceae Scrophulariaceae Scrophulariaceae Scrophulariaceae Scrophulariaceae Scrophulariaceae Scrophulariaceae Scrophulariaceae Scrophulariaceae Scrophulariaceae Solanaceae Solanaceae Solanaceae Solanaceae Solanaceae Solanaceae Solanaceae Solanaceae Solanaceae Solanaceae Solanaceae Apiaceae Apiaceae Apiaceae Apiaceae Araliaceae Araliaceae Araliaceae Araliaceae Araliaceae Araliaceae Araliaceae Araliaceae Araliaceae Araliaceae Araliaceae Araliaceae Araliaceae Helwingiaceae Aquifoliaceae Asteraceae Asteraceae Asteraceae Campanulaceae Campanulaceae Campanulaceae Campanulaceae Campanulaceae Campanulaceae Campanulaceae Campanulaceae Campanulaceae Menyanthaceae Campanulaceae Campanulaceae Adoxaceae Adoxaceae Adoxaceae Adoxaceae Adoxaceae Caprifoliaceae Caprifoliaceae Caprifoliaceae Caprifoliaceae Caprifoliaceae Diervillaceae Unplaced Dipsacaceae Dipsacaceae Dipsacaceae Unplaced Linnaeaceae Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Lamiales Solanales Solanales Solanales Solanales Solanales Solanales Solanales Solanales Solanales Solanales Solanales Apiales Apiales Apiales Apiales Apiales Apiales Apiales Apiales Apiales Apiales Apiales Apiales Apiales Apiales Apiales Apiales Apiales Aquifoliales Aquifoliales Asterales Asterales Asterales Asterales Asterales Asterales Asterales Asterales Asterales Asterales Asterales Asterales Asterales Asterales Asterales Dipsacales Dipsacales Dipsacales Dipsacales Dipsacales Dipsacales Dipsacales Dipsacales Dipsacales Dipsacales Dipsacales Dipsacales Dipsacales Dipsacales Dipsacales Dipsacales Dipsacales Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 815 S.B. Janssens et al. / Molecular Phylogenetics and Evolution 52 (2009) 806–824 Table 2 (continued) Species Accession No. Family Order Superorder Dipelta yunnanensis Kolkwitzia amabilis Linnaea borealis Zabelia biflora Acanthocalyx albus Cryptothladia chinensis Morina longifolia Centranthus ruber Fedia graciliflora Nardostachys jatamansii Plectritis macrocera Valeriana bractescens Valeriana celtica Valerianella locusta Castanea crenata Vatairea fusca AF447000 AF447002 AF447001 EU265530 AF447003 AF447004 AF447005 AF448572 AF448575 AF447010 AF447015 AF448580 AY362516 AF447014 AB124938 EF466253 Linnaeaceae Linnaeaceae Linnaeaceae Linnaeaceae Morinaceae Morinaceae Morinaceae Valerianaceae Valerianaceae Valerianaceae Valerianaceae Valerianaceae Valerianaceae Valerianaceae Rosids Rosids Dipsacales Dipsacales Dipsacales Dipsacales Dipsacales Dipsacales Dipsacales Dipsacales Dipsacales Dipsacales Dipsacales Dipsacales Dipsacales Dipsacales Rosids Rosids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Euasterids Outgroup Outgroup 3500 3000 2500 SS 2000 1500 1000 500 0 1 2 3 4 5 6 7 8 9 10 Fossil calibration Fig. 1. Histogram of the summed square values (SS) of the deviations between estimated molecular and fossil ages for each calibration point. 45 40 35 30 s 25 20 15 10 5 0 1 2 3 4 6 5 7 Fossils removed 8 9 10 Fig. 2. Plot illustrating the effect of the removal of fossil calibration points on the average squared deviation (s). (Pybus and Rambaut, 2002). In addition, we calculated the 95% confidence interval and the mean c-value for each of the simulated datasets. 3. Results 3.1. Sequence characteristics and phylogenetic results Sequence characteristics for all datasets are summarized in Table 3. Ambiguously aligned nucleotides were removed from both chloroplast and nuclear data matrices. Although some loci have not been amplified from a few species, their absence probably did not 2 2 2 2 2 2 2 2 2 2 2 2 2 2 alter the analyses when coded as missing data in the combined data matrix. A recent study demonstrated that highly incomplete taxa can be accurately placed in phylogenies as long as many characters were sampled throughout (Wiens, 2005). The ML analysis of the enlarged chloroplast Impatiens-asterids dataset yielded a well-supported phylogenetic hypothesis in which the different orders are well defined and the relationship between the orders largely corroborates the study of Bremer et al. (2002). Nevertheless, we observed some discrepancies between their phylogeny and the present study, yet all of which are weakly supported in both studies. Whereas our results reveal a sistergroup relationship between Solanales and Gentianales (ML-BS: 56), and Asterales and Dipsacales (ML-BS: 73), the study of Bremer et al. (2002) shows a closer affinity between Solanales and Lamiales (maximum parsimony-BS: <50) on the one hand, and an unresolved polytomy with Dipsacales, Asterales and Apiales on the other. The phylogeny of Bremer et al. (2004) however, which is based on the molecular results of Bremer et al. (2002) indicates a sistergroup relationship between Dipsacales and Asterales. Within the order Ericales, the balsaminoid clade is sister to the other Ericalean families (ML-BS: 100; e.g. Bremer et al., 2002; Schönenberger et al., 2005). In contrast to several molecular studies on Ericales and asterids, in which the mutual relationships between Balsaminaceae, Tetrameristaceae s.l. and Marcgraviaceae remained unresolved (APG, 2003; Savolainen et al., 2000; Schönenberger et al., 2005), the present study reveals a well supported sistergroup relationship between Tetrameristaceae s.l. (Pelliciera) and the Balsaminaceae–Marcgraviaceae clade (ML-BS: 98). Although many of the early diversified nodes in Impatiens are well resolved, the most recent nodes within the genus show limited resolution (Fig. 3). However, the enlarged asterid dataset was put together to resolve the basal relationships within the balsaminoid clade, which are subsequently used in the time divergence estimation procedure. ML analysis of the nuclear ImpDEF1/ImpDEF2 data matrix of Impatiens yielded highly congruent trees, supported by moderate bootstrap support (data not shown). Topologies based on ImpDEF1/ImpDEF2 are slightly better resolved than those based on plastid atpB-rbcL sequences. The separately analyzed plastid and nuclear datasets of Impatiens corroborated the major clades recognized by Janssens et al. (2006) with no incongruent clades found between the two major data sets. Furthermore, the partition homogeneity test was not significant (P > 0.05), showing that chloroplast and nuclear partitions of the combined data were not in conflict. For the combined data matrix, the ML search yielded a well-resolved topology with strong support for most of the clades (Fig. 4). In comparison to the separate ImpDEF1/ImpDEF2 and atpBrbcL analyses, the combined phylogeny is better resolved and support values are generally higher. 816 S.B. Janssens et al. / Molecular Phylogenetics and Evolution 52 (2009) 806–824 Table 3 Sequence characteristics of the nuclear and chloroplast loci and the combined matrix. Analyzed markers Analyzed characters Variable characters Parsimony informative characters Asterid dataset Impatiens dataset atpB-rbcL ImpDEF1/ImpDEF2 atpB-rbcL Combined 3822 908 687 5637 804 468 989 248 133 6626 1052 601 100 3 Cornales Impatiens Balsaminaceae 100 Ericales 100 Balsaminoids 98 100 100 11 Marcgraviaceae Tetrameristaceae s.l. 75 1 Lamiales 89 5 97 7 100 Solanales 100 10 56 Gentianales 96 100 99 2 100 Garryales Aquifoliales 100 95 73 Asterales 8 93 95 Apiales 9 96 Dipsacales 6 68 140 120 100 4 80 60 Rosids 40 20 0 Fig. 3. ML phylogram of the asterid atpB-rbcL dataset. The considered calibration points (1–11) are indicated on the tree. ML bootstrap support values for the most basal nodes until the nodes at order level are highlighted on the branches. Support values at lower taxonomical level are not shown. S.B. Janssens et al. / Molecular Phylogenetics and Evolution 52 (2009) 806–824 3.2. Divergence time estimates Using the ‘fossil cross-validation’ method of Near and Sanderson (2004) we identified the Ebenaceae fossil A. cryptostoma and 817 the earliest known Asclepioid fossil as most inconsistent calibration points in our analysis. The node to which A. cryptostoma was assigned has been dated 11 million years younger than assumed by fossil data, whereas the crown node of the Asclepioideae was Fig. 4. ML phylogram based on the combined chloroplast and nuclear dataset of Impatiens. Asterisks indicate the presence of collapsed branches in strict consensus tree. Numbers on the branch represent ML bootstrap support values. Clades M (Madagascar), SEA1 (Southeast Asia), SEA2 (Southeast Asia), SI1 (South India), SI2 (South India), NA (NorthAmerica), A1 (Africa), A2 (Africa) and A3 (Africa) are indicated with an arrow. Blue bars indicate age intervals. 818 S.B. Janssens et al. / Molecular Phylogenetics and Evolution 52 (2009) 806–824 Table 4 Estimated ages for the crown and stem groups of the asterid orders. Order Penalized likelihood age estimates (Mya) Bremer et al. (2004) age estimates (Mya) Stem Crown Stem Crown Cornales Ericales Garryales Lamiales Gentianales Solanales Aquifoliales Apiales Dipsacales Asterales 128* 123 ± 10.5 112 ± 9.3 104 ± 8.2 101 ± 7.9 101 ± 11.8 113 ± 9.8 104 ± 11.2 105 ± 14.2 104 ± 12.1 104 ± 13.1 117 ± 9.2 20 ± 8.6 95 ± 11.9 79 ± 10.2 58 ± 9.1** 62 ± 11.9*** 87 ± 14.1 99 ± 15.4 94 ± 11.2 128* 127 114 106 108 106 121 113 111 112 112 114 – 97 78 100 113 84 101 93 * ** The crown node of the asterid was fixed at 128 Mya. The Convolvulaceae were not included in our dating analysis, which explains the young age estimate for the crown group of the Solanales. The Stemonuraceae and Cardiopteridaceae were not included in our dating analysis, which explains the young age estimate for the crown group of the Aquifoliales. *** estimated approximately 15 million years younger than suggested by fossil data. However, when including both fossil calibration points in the overall dating analysis, the dating estimates of the Impatiens nodes were hardly altered. Nevertheless, both calibration points were removed from the final dating analysis. Applying the divergence times of the selected asterid fossils and the previously estimated age of the asterid crown group by Bremer et al. (2004), we dated the split between Marcgraviaceae and Balsaminaceae at 48.2 mya (SD = ±9.3 mya). The divergence between Impatiens and Hydrocera is estimated at 30.7 mya (SD = ±8.6 mya), whereas the crown group of Impatiens is estimated at 22.5 mya (SD = ±5.6 mya). The age estimations of the asterid orders with their standard deviations are listed in Table 4. In order to infer the divergence times for the remaining lineages in Impatiens, we used the estimated crown age of Impatiens (22.5 my) as secondary calibration points for the ML phylogram of the combined nuclear-plastid dataset (Fig. 4). In addition, geological evidence was used to cross-validate the asterid-based divergence times. 3.3. Ancestral area distribution Maximum likelihood-based ancestral area reconstruction illustrates that the vast majority of the current diversification of Impatiens has originated in Asia (Fig. 5 and Table 5). Southwest China was unambiguously reconstructed as the ancestral area for the earliest-diverging lineages. Although likelihood ratio values were in general significant for the Impatiens topology, we noticed that the root node, which connects Impatiens to its sistergroup Hydrocera, has an insignificant likelihood ratio value. Apparently, this insignificance occurs rather frequently in likelihood-based ancestral state reconstruction analyses as the amount of uncertainty related with reconstruction increases in time (Schluter et al., 1997). Biogeographic reconstructions indicate that Impatiens dispersed into South India via two independent colonization events from Southwest China (clades SI1 and SI2), whereas the African continent was colonized from Southwest China in three independent dispersal events (clades A1, A2 and A3). Similarly, our reconstructions show that only one speciation event from Southwest China accounts for the diversity of Impatiens in North America (clade NA). Furthermore, the Southeast Asian (clades SEA1 and SEA2) and Himalayan species originated from Southwest China. In contrast, the Malagasy clade (clade M) is derived from a single colonization event from Africa (Fig. 5 and Table 5). 3.4. Speciation rates and timing of diversification in Impatiens The slope of the LTT plot obtained from the PL tree for Impatiens is slightly convex until 4.5 million years ago and then becomes steeper (Fig. 6). Simulated LTT curves for Impatiens diversity indicate a hypothesized steep slope near the root of the genus that gradually decreases towards the tips. Interestingly, between 15 million years ago and present-day, the empirical LTT plot is situated outside the 95% significance level of the simulated LTT plots. In order to statistically test whether this observation truly indicates a late increase in net diversification rate, we evaluated our results with c-statistics (Fig. 7). For our chronogram, we computed a negative c-value (c = 6.27), indicating that the process of lineage accumulation in Impatiens did not remain constant in time. Simulations carried out to consider the effect of incomplete sampling (113 species sampled from 1000 extant species), indicated that under a true value of 0, Impatiens would only contain 207 species with the 95% interval, ranging between 120 and 427 species to obtain a c-value as extreme as 6.27. As a result, we are able to reject the hypothesis that the obtained c-value was the result of a poor sampling density. Similar results are obtained when the diversification rate was calculated for every possible node and its accompanying lineage (Fig. 8). Between 22.5 and 5 mya, computed speciation rates are estimated between 0.03 and 0.56 species per million years. During the next 5 my, rates of species diversification rates fluctuate between 0.07 and 3.8 species per million years. 4. Discussion 4.1. Divergence time estimates in asterids Table 4 lists the stem and crown group estimates for all asterid orders. Although the obtained divergence estimates generally corroborate the results of Bremer et al. (2004), the majority of the stem group estimates are slightly younger, whereas most of the crown group estimates are a little older. The largest differences in age estimate between our study and the study of Bremer et al. (2004) were observed for the stem node age of the orders in the euasterids 2 clade and the crown group age of the orders Solanales, Cornales and Aquifoliales. These larger dissimilarities are possibly due to the use of a single gene marker (atpB-rbcL) that was not applied in the multigene dating analysis on the asterids by Bremer et al. (2004). In addition, we were unable to sample all the earliest-diverging lineages of each order, resulting in much younger crown node estimates for some of the above-mentioned lineages. The use of a single gene for generating age estimates also caused S.B. Janssens et al. / Molecular Phylogenetics and Evolution 52 (2009) 806–824 819 Fig. 5. ML topology illustrating the reconstruction of ancestral area distribution of Impatiens ancestors. Pie charts at each node indicate proportional likelihoods of the most common recent ancestor for each clade. Clades M (Madagascar), SEA1 (Southeast Asia), SEA2 (Southeast Asia), SI1 (South India), SI2 (South India), NA (North America), A1 (Africa), A2 (Africa) and A3 (Africa) are indicated with an arrow. the bootstrap confidence intervals to be fairly large, indicating a certain inaccuracy of the atpB-rbcL dataset. However, the overall similarity in ordinal age estimates in Bremer et al. (2004) and the present study suggests that our chronogram presents a plausible scenario for other nodes among the asterids (e.g. Impatiens). Despite the slightly younger age estimate for most of the stem nodes of the asterid orders, the crown group of the Ericales, to 820 S.B. Janssens et al. / Molecular Phylogenetics and Evolution 52 (2009) 806–824 Table 5 Proportional likelihoods values for reconstruction of ancestral areas. Values in bold indicate best estimates. Node numbers correspond to those shown in Fig. 5. NODE Africa Madagascar Himalaya Southwest China North America Southeast Asia South India-Sri Lanka Inferred ancestral area 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 6.5 5.6 6.1 99.1 99.0 0.8 0.4 0.1 <0.05 0.1 <0.05 1.9 <0.05 98.2 <0.05 0.9 <0.05 <0.05 99.1 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 0.8 99.0 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 1.3 2.5 3.7 3.7 2.6 93.3 94.2 93.7 0.8 <0.05 <0.05 99.1 98.4 0.8 98.5 0.8 95.2 0.8 1.6 <0.05 99.0 98.4 1.7 0.8 99.1 98.6 97.4 96.2 96.2 97.3 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 1.5 98.2 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 0.1 0.1 <0.05 <0.05 <0.05 0.2 1.3 <0.05 <0.05 99.0 2.8 99.1 <0.05 0.9 <0.05 <0.05 <0.05 <0.05 0.8 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 0.1 <0.05 99.1 1.2 <0.05 <0.05 <0.05 <0.05 99.0 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 <0.05 Southwest China Southwest China Southwest China Africa Africa Madagascar Southwest China Southwest China South India-Sri Lanka Southwest China Southeast Asia Southwest China Southeast Asia Africa South India-Sri Lanka Southwest China Southwest China North America Africa Southwest China Southwest China Southwest China Southwest China Southwest China Southwest China which Impatiens belongs, was dated three million years older (117 ± 10.5 my) than calculated by Bremer et al. (2004). In addition, the study of Bremer et al. (2004) estimated the crown group of the balsaminoid clade at approximately 64 million years, a date that corresponds with our results (58.9 my). The crown group age of the balsaminoids is also slightly older than the first known fossil of Pelliciera rhizophorae from the Late Eocene (54.9 mya; Graham, 1975; Rull, 1999). We decided not to apply this age estimate in the asterid dating analysis but to use the fossil pollen age of Pelliciera to cross-validate our obtained results. By applying the fossil age of P. rhizophorae to the node directly leading towards Balsaminaceae and Marcgraviaceae, the estimated age of Impatiens would have been strongly influenced by this one calibration point, 6 4 3 2 Number of reconstructed lineages (ln) 5 1 0 20 15 10 5 0 Million years before present Miocene Pliocene Pleistocene whereas all other fossil calibration points would have had no value in dating Impatiens. 4.2. Origin and evolution of the Balsaminaceae Only recently, molecular evidence has demonstrated that Balsaminaceae are closely related to Marcgraviaceae, both being part of the balsaminoid clade in the Ericales (Bremer et al., 2002; Schönenberger et al., 2005). Despite their close affinity, these families have very different distribution areas. The Balsaminaceae have a worldwide distribution except for Australia and South America, whereas the Marcgraviaceae are restricted to tropical forests of South and Central America (Dressler, 2004; Fischer, 2004). It seems plausible that their most recent common ancestor was present by the start of the Tertiary (Grey-Wilson, 1980a). Although the split between Marcgraviaceae and Balsaminaceae is dated in this study in the Middle Eocene (48.1 mya), the ancestral stock of both families already existed during the Middle Paleocene (58.9 mya). This Proportion (f) of Impatiens sampled 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1.1 0 -2 -4 -6 Average γ -value=-6.27 -8 -10 -12 -14 Fig. 6. Semilogarithmic lineage-through-time (LTT) plot for the Impatiens clade sister to I. omeiana are indicated by a black line. LTT plot from simulated phylogenies illustrating the effect of an incomplete sampling is shown by a grey line. The thin grey lines indicate the 95% interval while the thick grey line designates a mean value. Fig. 7. Relationship between c and the proportion of extant taxa in Impatiens (f). The thin grey lines indicate the 95% interval while the thick black line appoints the mean c-value. The shaded area represents the range of f-values consistent with the observed c at the 95% level. 821 S.B. Janssens et al. / Molecular Phylogenetics and Evolution 52 (2009) 806–824 3.5 1 3 2 2.5 3 2 4 1.5 5 1 6 0.5 Diversification rates 0 18 O (ppt) 4 0 20 15 10 Million years before present Miocene 5 0 Pliocene Pleistocene Fig. 8. Diversification rates calculated for the genus Impatiens (black line). Superimposed (grey line) is a time-averaged record of sea surface temperatures (Zachos et al., 2001). confirms the hypothesis of Grey-Wilson (1980a) suggesting an origin for Balsaminaceae ca. 50 million years ago. It lasted another 18 million years before the lineage leading to the extant species of Impatiens diverged from the lineage leading to the extant species of Hydrocera (30.7 mya, Late Eocene). Despite this early separation, the extant species of Impatiens have a coalescence point in the Early Miocene (22.5 mya; Fig. 3). It is unclear whether this phenomenon of no net diversification during a period of 8 million years is due to low rates of speciation or high rates of extinction because no fossil data are available for this group. A similar question applies to the sister group of Impatiens, the monospecific Hydrocera triflora, which has a restricted Indo-Malaysian distribution (Grey-Wilson, 1980b). Since the origin of this lineage in the Late Eocene, climatic conditions have changed significantly, resulting in several periods of drought and huge fluctuations in sea level (Berggren et al., 1995; Haq et al., 1987; Riggs, 1984). Due to its restriction to swampy lowland habitats (not occurring above 100 m altitude) and a less efficient dispersal mechanism than its sistergroup Impatiens, other Hydrocera lineages could have gone extinct during that time. Alternatively, speciation could be slowed down by other causes (e.g. fluctuating sea levels), resulting in only very low diversification. This remarkable situation in which lack of diversification occurred in the sister lineage of a very species-rich genus is analogous with the monospecific genus Hillebrandia and its species-rich sister genus Begonia (ca. 1400 species, Clement et al. 2004). 4.3. Diversification patterns in Impatiens: the effect of a global climate change Despite our ability to reconstruct the basal phylogenetic relationships within Impatiens using only the chloroplast atpB-rbcL spacer data, most of the recently diversified lineages remained unresolved. By combining the rapidly evolving ImpDEF1 and ImpDEF2 markers with the chloroplast atpB-rbcL spacer, we were able to resolve many of the previously unresolved relationships in Impatiens. However, due to rapid evolution of the ImpDEF genes, we were obliged to use an approach of secondary calibration to assess the evolutionary history of the rapid diversification of Impatiens. Several studies indicate that secondary calibration can cause considerable problems in molecular dating (Graur and Mar- tin, 2004; Hedges and Kumar, 2003, 2004). Nevertheless, we considered this a valid approach for our study because we used the same dating protocol, a similar phylogeny reconstruction method and complementary taxon sampling for both dating analyses. Furthermore, in order to reduce the problem of secondary calibration, we used additional geologic evidence together with the calibration points based on direct and indirect fossil data. The present study indicates that diversification in Impatiens increased since the last 4.6–5 million years (Fig. 8). This period of exceptional diversification, during which 84% of all Impatiens lineages were established, can be linked with the global cooling that occurred during the Pliocene and Pleistocene (Fig. 6 In addition, when comparing the proportion of 207 Impatiens lineages (b/ d = 0) equalling the calculated c-value of 6.27 with the approximately 1000 currently described taxa, it is clear that the climatic conditions correlate well with the rapid diversification within the genus. This can also be seen in the fluctuating species diversification rates of the last 4.6 million years (between 0.09 and 4 species per million year; Fig. 8), which are probably related to the glacial– interglacial cycles that were initiated during this period (Zachos et al., 2001). Our study demonstrates that the current species richness of Impatiens is the result of a sudden diversification boost and did not originate via gradual accumulation of species over a long geological period. In addition the massive species diversity in Impatiens occurred through changing climatic conditions in the Pliocene and Pleistocene, and was not the outcome of an enhanced speciation event in the Tertiary or earlier. Consequently, the ‘‘refuge” hypothesis, in which rainforest species populations became isolated from each other during times of cool and dry climate conditions, may be a plausible model to explain the enormous diversity in Impatiens. The plants are characterized by explosively dehiscent capsules with seeds travelling for the most part less than a meter. In addition, most balsams are montane plants, often restricted to a small area, sometimes no more than a single mountain peak or a mountain range. The periodicity of glacial cycles during the Pleistocene could have resulted in several alternating episodes of contraction and expansion of the montane and lowland rainforests with vegetation belts slightly shifting down and upslope during these varying climatic conditions (Kebede et al. 2007). As a result, Impatiens would have been forced to migrate along with the rainforest belts in order to retain its required habitat type. With 822 S.B. Janssens et al. / Molecular Phylogenetics and Evolution 52 (2009) 806–824 its area of distribution being extremely fragmented during certain climatic episodes, many different Impatiens populations were probably isolated for several thousands of years. The continuous cycle of dispersal subsequent to fragmentation and isolation that has been induced by climatic fluctuation during the Pliocene and Pleistocene almost certainly contributed to the rapid radiation of Impatiens. 4.4. Biogeographic patterns in Impatiens Biogeographic reconstruction indicates that the center of origin for Impatiens is Southwest China, from where the genus subsequently dispersed to Africa, India, the Himalayas, Southeast Asia, Central Asia and North America. These results are consistent with the fact that China contains 222 species or roughly 1/4th of all currently described Impatiens species, thereby largely surpassing South India and Sri Lanka, the Himalayas, Africa (Grey-Wilson, 2008). Of these Chinese species, the majority is endemic to the provinces of Sichuan and Yunnan in the Southwest of China. In addition, the only fossil bearing a relation to extant Impatiens was found in the East China Sea (I. brevicolpus – Santan Formation, Pliocene; Song et al., 2004). In the past, several morphological similarities have been observed in species endemic to Africa and South India, suggesting a close affinity between these taxa and even a possible migration route connecting these two areas (Grey-Wilson, 1980a). Despite the resemblance between various African and South Indian balsams, the present study illustrates that the South Indian species originated from two independent dispersal events (clades SI1 and SI2), with Southwest China reconstructed as the area of origin for each clade. However, one of the two South Indian lineages (clade SI2) is closely related to one of the African subclades (clade A3), hence explaining the overall similarities between taxa from two clades. This African-affiliated lineage colonized South India between the Late Pliocene and the Early Pleistocene. In contrast, the other South Indian lineage (clade SI1) previously dispersed in the Late Miocene. Based on the close resemblance of the North American lineage to some Chinese species, Grey-Wilson (1980a) suggested that the genus arrived only recently in the New World. Our data corroborate his interpretation, as the colonization of North America is estimated to have occurred in the Pleistocene between 1.32 and 1.27 million years ago (clade NA). Interestingly, this time estimate corresponds with a glacial maximum that lasted from 1.35 until 1.25 mya (Berger et al., 1999; Bintanja and Van de Wal, 2008; Raymo 1997). The North American clade is sister to Impatiens nolitangere, one of the few Impatiens species that is not endemic to a restricted geographic area but is a widespread species in temperate Eurasia, including Southwest China (Chen et al., in press). North American may have been colonized via the Bering land bridge in association with glacial–interglacial stages during the Pleistocene. Because some Chinese species are distributed as far as the Himalayan chain, it has been considered that there is a strong biogeographic correlation between these two distribution areas (Grey-Wilson, 2008). Present data show that all species endemic to the Himalayas have a sister species which is distributed or even endemic to China. As many other hotspots for Impatiens, also the Himalayan region is characterized by a considerable number of species (ca. 120 ssp.). All these Himalayan endemics appear to find their origin in (Southwest) China via several separate dispersal events, which occurred as early as the Late Miocene and not via a single colonization event as for example with the Malagasy or North American Impatiens species. The sister species to the New Guinean endemic, Impatiens hawkeri, is located in the Southeast Asian Archipelago. In addition, the closest relative of these two species is distributed in South China and the northern provinces of India, Myanmar and Vietnam close to the Chinese border. As a result, the most recent common ancestor of the New Guinean species dispersed from South China towards the Southeast Asian Archipelago from where New Guinea was eventually reached (clade SEA1). According to our results, the lineage towards I. hawkeri recently crossed the Wallace’s Line from Sunda to Sahul Shelves, thereby arriving on New Guinea during the Early Pleistocene. This recent age estimate is in concordance with the difficulties that were encountered to delimit species of the existing lineages on the island (Grey-Wilson, 1980b). As a result, Grey-Wilson lumped all previously described Papuasian Impatiens species in a single hugely variable species, which is subdivided in 15 groups. A population genetic approach will be required to further investigate the complex evolution of this group. Ancestral area reconstruction clearly indicates that Impatiens colonized Africa on three separate occasions. Previous studies demonstrated that Africa was colonized more than once, but the exact number of dispersal events could not be determined (Janssens et al., 2006; Yuan et al., 2004). The most recent common ancestor of the three African clades is located in Southwest China. Although our initial work (Janssens et al., 2006) left open the possibility that India may have been colonized from Africa, this expanded study suggests that African Impatiens never recolonized Asia. The only colonization events of Impatiens from Africa were to the offshore islands of the Seychelles and Madagascar (and then to the Comoros) in the Indian Ocean. Although only 4% (7/ ± 150) of the Malagasy Impatiens species were sampled for this study, all species are closely related, raising the possibility that most or all of this diversity was derived from a single colonization event (clade M). The initial colonization of Africa by Impatiens apparently occurred relatively early, as the oldest group arrived during the Late Miocene (clade A1). All species present in this old African clade are restricted to East and South Africa and characterized by disjunct distribution areas. Considering the older age of this lineage, the disjunct distribution areas of its representatives and the low rate of diversification, this clade may have experienced extinction in the intervening areas. The second dispersal of Impatiens to Africa occurred between the Late Miocene and Early Pliocene (clade A2). Representatives of this clade are endemic to West Africa. The third colonization of Africa occurred during the Late Miocene and Early Pliocene (clade A3). This lineage quickly diversified after reaching Africa, resulting in the largest diversification of Impatiens in Africa. Most Impatiens species in this clade are restricted to the montane forests and highlands of Central and East Africa. Additionally, only a few species are present in both Central and West Africa. Most species that occur in both Central and West Africa have a disjunct distribution, only occurring in mountainous regions of West Africa, and the eastern part of Congo and neighbouring mountainous regions, yet absent in the lowland forests of central Congo. One exception is Impatiens niamniamensis, which has a continuous distribution from Central to West Africa throughout the Congo River basin. Because the present Congolian lowland forests have been considerably affected by periods of glacial aridity, we assume that this species recolonized the Congo basin only recently. Species with disjunct distribution areas in Central and West Africa may have also been affected by ice age aridity. For example Impatiens mannii is inferred to have originated during the Pleistocene in the volcanic regions of Central Africa and may have migrated to West Africa, in which case it would have been more widespread throughout the tropical rainforests of central Africa. Climatic fluctuations related to glacial events reduced the central African rainforests (Coetzee, 1993; Plana, 2004). Only the more mountainous populations of I. mannii may have survived, having failed to recolonize the lowlands like I. niamniamensis. Alternatively, I. mannii may have reached West Africa via long- S.B. Janssens et al. / Molecular Phylogenetics and Evolution 52 (2009) 806–824 distance dispersal. In either case, the morphological similarity between both western and central African populations of I. mannii indicates a recent isolation. Acknowledgments We thank the Tanzanian Commission for Science and Technology, the University of Dar es Salaam, the Ugandan National Research Council, Makerere University, the National Museums of Kenya, the National Herbarium of Zimbabwe and the University of the Witwatersrand for permission and collaboration while conducting fieldwork in the respective countries. We are also grateful to the National Botanic Garden of Belgium (BR), National Herbarium of the Netherlands, Leiden University branch (L), Utrecht University branch (U) and Wageningen University branch (WAG), Botanischer Garten Berlin-Dahlem (B), University of California Botanical Garden at Berkeley (UC), Botanical Garden of Marburg (MB), Botanical Garden of the University of Copenhagen (C), Royal Botanic Garden Edinburgh (E), South China Botanical Garden (IBSC), Denver Botanic Gardens (DBG), Holden Arboretum, YongMing Yuan, Yi Song, Ray Morgan, Steven Dessein and Anke Geeraerts for providing plant material. Tom Van England is acknowledged for his help with the statistical methods used in this paper. This study was financially supported by research grants of the K.U.Leuven (OT/05/35) and the Fund for Scientific ResearchFlanders (FWO Belgium) (G.0104.01). Steven Janssens holds a PhD research grant from FWO. References Angiosperm Phylogeny Group, 2003. An update of the Angiosperm Phylogeny Group classification for the orders and families of flowering plants: APG II. Bot. J. Linn. Soc. 141, 399–436. Anderson, C.L., Bremer, K., Friis, E.M., 2005. Dating phylogenetically basal eudicots using rbcL sequences and multiple reference points. Am. J. Bot. 92, 1737–1747. Aubréville, A., 1962. Savanisation tropicale et glaciations Quaternaires. Adansonia 2, 16–84. Basinger, J.F., Christophel, D.C., 1985. Fossil flowers and leaves of the Ebenaceae from the Eocene of southern Australia. Can. J. Bot. 63, 1825–1843. Baldwin, B.G., Sanderson, M.J., 1998. Age and rate of diversification of the Hawaiian silversword alliance (Compositae). Proc. Natl. Acad. Sci. USA 95, 9402–9406. Bell, C.D., Donoghue, M.J., 2005. Dating the Dipsacales: comparing models, genes, and evolutionary implications. Am. J. Bot. 92, 284–296. Berger, A., Li, X.S., Loutre, M.F., 1999. Modeling northern hemisphere ice volume over the last 3 Ma. Quat. Sci. Rev. 18, 1–11. Berggren, W.A., Kent, D.V., Swisher, I.C.C., Aubry, M.-P., 1995. A revised Cenozoic geochronology and chronostratigraphy. In: Berggren, W.A., Kent, D.V., Aubry, M.-P., Hardenbol, J. (Eds.), Geochronology, Time Scales and Global Stratigraphic Correlation. Society for Economic Paleontology and Mineralogy, Tulsa, Oklahoma, pp. 129–212. Bhaskar, V., Razi, B.A., 1981. Peninsular Indian Impatiens L. (Balsaminaceae). Bull. Bot. Surv. India. 23, 191–196. Bintanja, R., van de Wal, R.S.W., 2008. North American ice-sheet dynamics and the onset of 100,000-year glacial cycles. Nature 454, 869–872. Bremer, B., Bremer, K., Heidari, N., Erixon, P., Olmstead, R.G., Anderberg, A.A., Källersjö, M., Barkhordarian, E., 2002. Phylogenetics of asterids based on 3 coding and 3 non-coding DNA at higher taxonomic levels. Mol. Phylogenet. Evol. 24, 274–301. Bremer, K., Friis, E.M., Bremer, B., 2004. Molecular phylogenetic dating of asterid flowering plants shows Early Cretaceous diversification. Syst. Biol. 53, 496–505. Bowie, R.C.K., Fjeldså, J., Hackett, S.J., Crowe, T.M., 2004. Molecular evolution in space and though time: mtDNA phylogeography of the Olive Sunbird (Nectarinia olivacea/obscura) throughout continental Africa. Mol. Phylogenet. Evol. 33, 56–76. Call, V.B., Dilcher, D.L., 1992. Investigations of angiosperms from the Eocene of southeastern North America: Samaras of Fraxinus wilcoxiana. Rev. Paleobot. Palynol. 74, 249–266. Chen, Y.-L., 2001. Balsaminaceae. In: Chen, Y.-L. (Ed.), Flora Reipublicae Popularis Sinicae, Tomus 47. Science Press, Beijing, pp. 1–243. Chen,Y., Akiyama, S. Ohba, H., in press. Flora of China: Balsaminaceae. Clement, W.L., Tebbitt, M.C., Forrest, L.L., Blair, J.E., Brouillet, L., Eriksson, T., Swensen, S.M., 2004. Phylogenetic position and biogeography of Hillebrandia sandwicensis (Begoniaceae): a rare. Hawaiian relict. Am. J. Bot. 91, 905–917. Coetzee, J.A., 1993. African flora since the terminal Jurassic. In: Goldblatt, P. (Ed.), Biological Relationships Between Africa and South America. Yale University Press, New Haven, pp. 37–61. 823 Collinson, M.E., Boulter, M.C., Holmes, P.L., 1993. Magnoliophyta (‘Angiospermae’). In: Benton, M.J. (Ed.), The Fossil Record, vol. 2. Chapman and Hall, London, pp. 809–841. Crowe, T.M., Crowe, A.A., 1982. Patterns of distribution, diversity and endemism in Afrotropical birds. J. Zool. 198, 417–442. Davies, T.J., Barraclough, T.G., Chase, M.W., Soltis, P.S., Soltis, D.S., Savolainen, V., 2004. Darwin’s abominable mystery: insights from a supertree of the angiosperms. Proc. Natl. Acad. Sci. USA 101, 1904–1909. Dressler, S., 2004. Balsaminaceae. In: Kubitzki, K. (Ed.), Families and Genera of Vascular Plants, vol. 6. Springer-Verlag, Berlin, Heidelberg, pp. 258–265. Erkens, R.H.J., Chatrou, L.W., Maas, J.W., van der Niet, T., Savolainen, V., 2007. A rapid diversification of rainforest trees (Guatteria; Annonaceae) following dispersal from Central into South America. Mol. Phylogenet. Evol. 44, 399–411. Felsenstein, J., 1995. PHYLIP: Phylogeny Inference Package. Version 3.57c. Department of Genetics, University of Washington, Seattle, Washington. Fischer, E., 2004. Balsaminaceae. In: Kubitzki, K. (Ed.), Families and Genera of Vascular Plants, vol. 6. Springer-Verlag, Berlin, Heidelberg, pp. 20–25. Gamble, T., Bauer, A.M., Greenbaum, E., Jackman, T.R., 2008. Evidence for Gondwanan vicariance in an ancient clade of gecko lizards. J. Biogeogr. 35, 88–104. Glor, R.E., Flecker, A.S., Benard, M.F., Power, A.G., 2001. Lizard diversity and agricultural disturbance in a Caribbean forest landscape. Biodiv. Conserv. 10, 711–723. Graham, A., 1975. Late Cenozoic evolution of tropical lowland vegetation in Veracruz, Mexico. Evolution 29, 723–735. Graur, D., Martin, W., 2004. Reading the entrails of chickens: molecular timescales of evolution and the illusion of precision. Trends Genet. 20, 80–86. Grey-Wilson, C., 1980a. Impatiens of Africa. Balkema, Rotterdam. Grey-Wilson, C., 1980b. Hydrocera triflora, its floral morphology and relationship with Impatiens. Studies in Balsaminaceae: V. Kew Bull. 35, 213–219. Grey-Wilson, C., 1989. A revision of Sumatran Impatiens. Studies in Balsaminaceae: VIII. Kew Bull. 44, 107–113. Grey-Wilson, C., 2008. Impatiens forrestii: Balsaminaceae. Curtis’s Bot. Mag. 25, 139– 145. Haffer, J., 1969. Speciation in Amazonian forest birds. Science 165, 131–137. Haffer, J., 1982. General aspects of the refuge theory. In: Prance, G.T. (Ed.), Biological Diversification in the Tropics. Columbia University Press, pp. 6–24. Haffer, J., Prance, G.T., 2001. Climatic forcing of evolution in Amazonia during the Cenozoic: on the refuge theory of biotic differentiation. Amazoniana 16, 579– 607. Haq, B.U., Hardenbol, J., Vail, P.R., 1987. Chronology of fluctuating sea levels since the Triassic. Science 235, 1156–1167. Harris, D.J., Poulsen, A.D., Frimodt-Möller, C., Preston, J., Cronk, Q.C.B., 2000. Rapid radiation in Aframomum (Zingiberaceae): evidence from nuclear ribosomal DNA internal transcribed spacer (ITS) sequences. Edinburgh J. Bot. 57, 377–395. Hedges, S.B., Kumar, S., 2003. Genomic clocks and evolutionary timescales. Trends Genet. 19, 200–206. Hedges, S.B., Kumar, S., 2004. Precision of molecular time estimates. Trends Genet. 20, 242–247. Hughes, C., Eastwood, R., 2006. Island radiation on a continental scale: exceptional rates of plant diversification after uplift of the Andes. Proc. Natl. Acad. Sci. USA 103, 10334–10339. Janssens, S.B., Knox, E.B., Dessein, S., Smets. E.F., 2009. Impatiens msisimwanensis (Balsaminaceae): Description, pollen morphology and phylogenetic positioning of a new East African species. S. Afr. J. Bot. 75, 104-109. Janssens, S., Geuten, K., Yuan, Y.-M., Song, Y., Küpfer, P., Smets, E., 2006. Phylogenetics of Impatiens and Hydrocera (Balsaminaceae) using chloroplast atpB-rbcL spacer sequences. Syst. Bot. 31, 171–180. Janssens, S., Geuten, K., Viaene, T., Yuan, Y.-M., Song, Y., Smets, E., 2007. Phylogenetic utility of the AP3/DEF K-domain and its molecular evolution in Impatiens (Balsaminaceae). Mol. Phylogenet. Evol. 43, 225–239. Jaramillo, C., Rueda, M.J., Mora, G., 2006. Cenozoic plant diversity in the Neotropics. Science 311, 1893–1896. Johnson, W.E., Eizirik, E., Pecon-Slattery, J., Murphy, W.J., Antunes, A., Teeling, E., O’Brien, S.J., 2006. The late Miocene radiation of modern Felidae: a genetic assessment. Science 311, 73–77. Kebede, M., Ehrich, D., Taberlet, P., Nemomissa, S., Brochmann, C., 2007. Phylogeography and conservation genetics of a giant lobelia (Lobelia giberroa) in Ethiopian and Tropical East African mountains. Mol. Ecol. 16, 1233–1243. Kim, K.-J., Choi, K.S., Jansen, R.K., 2005. Two chloroplast DNA inversions originated simultaneously during the early evolution of the sunflower family (Asteraceae). Mol. Biol. Evol. 22, 1783–1792. Knapp, S., Mallet, J., 2003. Refuting Refugia? Science 300, 71–72. Knox, E.B., Palmer, J.D., 1998. Chloroplast DNA evidence on the origin and radiation of the giant lobelias in eastern Africa. Syst. Bot. 23, 109–149. Lancucka-Srodoniowa, M., 1967. Two new genera: Hemiptelea Planch. and Weigela Thurnb. in the younger Tertiary of Poland. Acta Palaeobot. 8, 1–17. Leopold, E.B., 1969. Miocene pollen and spore flora of Eniwetok, Marshall Islands. US Geol. Surv. Profess. Pap. 260, 1183–1185. Maddison, W.P., Maddison, D.R., 2008. Mesquite: a modular system for evolutionary analysis. Version 2.5. Available from: <http://mesquiteproject.org>. Maddison, D., Maddison, W., 2002. MacClade. Sinauer Associates, Sunderland, Massachusetts. Maley, J., 1987. Fragmentation de la fôret dense humide africaine et extension des biotopes montagnards au Quaternaire recent: nouvelles données polliniques et 824 S.B. Janssens et al. / Molecular Phylogenetics and Evolution 52 (2009) 806–824 chronologiques. Implications paleoclimatiques et biogeographiques. In: Coetzee, J.A. (Ed.), Palaeoecology of Africa and the Surrounding Islands, vol. 18. Balkema, Rotterdam, pp. 307–334. Mayr, E., O’Hara, R.J., 1986. The biogeographic evidence supporting the Pleistocene forest refuge hypothesis. Evolution 40, 55–67. Merckx, V., Chatrou, L.W., Lemaire, B., Sainge, M.N., Huysmans, S., Smets, E.F., 2008. Diversification of myco-heterotrophic angiosperms: Evidence from Burmanniaceae. BMC Evol. Biol. 8, 178–194. Moritz, C., Patton, J.L., Schneider, C.J., Smith, T.B., 2000. Diversification of rainforest faunas: an integrated molecular approach. Ann. Rev. Ecol. Syst. 31, 533–563. Muller, J., 1981. Fossil pollen records of extant angiosperms. Bot. Rev. 47, 1–142. Near, T.J., Sanderson, M.B., 2004. Assessing the quality of molecular divergence time estimates by fossil calibrations and fossil-based model selection. Philos. Trans. R. Soc. Ser. B 359, 1477–1483. Pagel, M., 1999. The maximum likelihood approach to reconstructing ancestral character states of discrete characters on phylogenies. Syst. Biol. 48, 612–622. Pennington, R.T., Lavin, M., Prado, D.E., Pendry, C.A., Pell, S.K., Butterworth, C.A., 2004. Historical climate change and speciation: neotropical seasonally dry forest plants show patterns of both Tertiary and Quaternary diversification. Philos. Trans. R. Soc. Ser. B 359, 515–537. Plana, V., Gascoigne, A., Forrest, L.L., Harris, D., Pennington, R.T., 2004. Pleistocene and pre-Pleistocene Begonia speciation in Africa. Mol. Phylogenet. Evol. 31, 449– 461. Plana, V., 2004. Mechanisms and tempo of evolution in the African Guineo– Congolian rainforest. Philos. Trans. R. Soc. Lond. B. 359, 1585–1594. Pòcs, T., 2007. A new species of Impatiens (Balsaminaceae) from the Nguru Mountains (Tanzania). Acta Bot. Hung. 49, 377–383. Prance, G.T., 1973. Phytogeographic support for the theory of Pleistocene forest refuges in the Amazon Basin, based upon evidence from distribution pattern in Caryocaraceae, Chrysobalanaceae, Dichapetalaceae and Lecythidaceae. Acta Amazon. 3, 5–28. Prance, G.T., 1982. A review of the phytogeographic evidences for Pleistocene climate changes in the Neotropics. Ann. Mo. Bot. Gard. 69, 594–624. Posada, D., Crandall, K.A., 1998. Modeltest, testing the model of DNA substitution. Bioinformatics 14, 817–818. Pybus, O.G., Rambaut, A., 2002. GENIE: estimating demographic history from molecular phylogenies. Bioinformatics 18, 1404–1405. Pybus, O.G., Harvey, P.H., 2000. Testing macro-evolutionary models using incomplete molecular phylogenies. Philos. Trans. R. Soc. Lond. B 267, 2267– 2272. Rahelivololona, M.E., Fischer, E., Wohlhauser, S., 2003. Balsaminaceae. In: Goodman, S.M., Benstead, J.P. (Eds.), The Natural History of Madagascar. The University of Chicago Press, Chicago and London, pp. 402–409. Rambaut, A., 2002. PHYLOGEN: Phylogenetic Tree Simulator Package. University of Oxford, Oxford. Rambaut, A., Charleston., M., 2002. TreeEdit: Phylogenetic Tree Editor v. 1.0 alpha 10. Available from: <http://evolve.zoo.ox.ac.uk/software/TreeEdit>. Raymo, M.E., 1997. The timing of major climate terminations. Paleoceanography 12, 577–585. Reid, E.M., Chandler, M.E.J., 1926. In: Catalogue of Cainozoic plants in the Department of Geology, vol. 1. The Bembridge Flora. Br. Mus. Nat. Hist., London, p. 206. Richardson, J.E., Pennington, R.T., Pennington, T.D., Hollingsworth, P.M., 2001. Rapid diversification of a species-rich genus of neotropical rain forest trees. Science 293, 2242–2245. Riggs, S.R., 1984. Paleoceanographic model of Neogene phosphorite deposition, US Atlantic Continental Margin. Science 223, 123–131. Rull, V., 1999. Palaeofloristic and palaeovegetational changes across the Paleocene– Eocene boundary in northern South America. Rev. Palaeobot. Palynol. 107, 83– 95. Sanderson, M.J., 2002. Estimating absolute rates of molecular evolution and divergence times: a penalized likelihood approach. Mol. Biol. Evol. 19, 101–109. Savolainen, V., Chase, M.W., Hoot, S.B., Morton, C.M., Soltis, D.E., Bayer, C., Fay, M.F., De Bruijn, A., Sullivan, S., Qiu, Y.-L., 2000. Phylogenetics of flowering plants based upon a combined analysis of plastid atpB and rbcL gene sequences. Syst. Biol. 49, 306–362. Schluter, D., Price, T., Mooers, A.O., Ludwig, D., 1997. Likelihood of ancestral states in adaptive radiation. Evolution 51, 1699–1711. Schönenberger, J., Anderberg, A.A., Sytsma, K.J., 2005. Molecular phylogenetics and patterns of floral evolution in the Ericales. Int. J. Plant Sci. 166, 265–288. Shimizu, T., 1979. The lectures on the flora of Thailand. Acta Phytotax. Geobot. 30, 180–187. Song, Z.-C., Wang, W.-M., Huang, F., 2004. Fossil pollen records of extant angiosperms in China. Bot. Rev. 70, 425–458. Stamatakis, A., 2006. RAxML-VI-HPC: maximum likelihood-based phylogenetic analyses with thousands of taxa and mixed models. Bioinformatics 22, 2688– 2690. Stamatakis, A., Ludwig, T., Meier, H., 2005. Raxml-III: a fast program for maximum likelihood-based inference of large phylogenetic trees. Bioinformatics 21, 456– 463. Swofford, D.L., 2002. PAUP. Phylogenetic Analysis Using Parsimony (and other Methods). Version 4. Sinauer Associates, Sunderland, Massachusetts. Takahashi, M., Crane, P.R., Manchester, S.R., 2002. Hironoia fusiformis gen. et sp. nov.; a cornalean fruit from the Kamikitaba locality (Upper Cretaceous, Lower Coniacian) in northeastern Japan. J. Plant Res. 115, 463–473. Toppin, S.M., 1920. Notes on the balsams of Chitral and the Kachin Hills. Bull. Misc. Inf. Kew 34, 5–366. Wiens, J.J., 2005. Missing data and the design of phylogenetic analyses. J. Biomed. Inform. 39, 34–42. Yuan, Y.-M., Song, Y., Geuten, K., Rahelivololona, E., Wohlhauser, S., Fischer, E., Smets, E., Küpfer, P., 2004. Phylogeny and biogeography of Balsaminaceae inferred from ITS sequence data. Taxon 53, 391–403. Zachos, J., Pagani, M., Sloan, L., Thomas, E., Billups, K., 2001. Trends, rhythms, and aberrations in global climate 65 Ma to present. Science 292, 686–693. Zink, R.M., Slowinsky, J.B., 1995. Evidence from molecular systematics for decreased avian diversification in the Pleistocene epoch. Proc. Natl. Acad. Sci. USA 92, 5832–5835.