review article phylogeny and biogeography in solanaceae

REVIEW ARTICLE

Phylogeny and biogeography in Solanaceae,Verbenaceae and Bignoniaceae: a comparisonof continental and intercontinentaldiversification patterns

RICHARD G. OLMSTEAD*

Department of Biology and Burke Museum, University of Washington, Box 355325, Seattle, WA98195, USA

Received 10 January 2012; revised 31 July 2012; accepted for publication 25 August 2012

Recent molecular phylogenetic studies of Solanales and Lamiales show that Solanaceae, Verbenaceae andBignoniaceae all diversified in South America. Estimated dates for the stem lineages of all three families implyorigins in the Late Cretaceous, at which time South America had separated from the united Gondwanan continent.A comparison of clades in each family shows (1) success in most clades at dispersing to, and diversifying in, NorthAmerica and/or the Caribbean, (2) a mix of adaptation to novel ecological zones and niche conservation, (3) limiteddispersal to continents outside of the western hemisphere, and, where this has occurred, (4) no association betweenlong-distance dispersal and fleshy, animal-dispersed fruits. Shared patterns among the three families contribute toa better understanding of the in situ diversification of the South American flora. © 2012 The Linnean Society ofLondon, Botanical Journal of the Linnean Society, 2013, 171, 80–102.

ADDITIONAL KEYWORDS: biogeography – long-distance dispersal – Neotropics – South America.

INTRODUCTION

South America is one of the great crucibles of plantdiversity, with some of the most diverse plant ecosys-tems on earth populated in high proportion by plantgroups that originated and diversified in situ (Gentry,1982; Pennington & Dick, 2004 Pennington et al.,2010; Antonelli & Sanmartin, 2011). That said,perhaps more has been written about the biogeographyof immigrants to South America (Renner, Clausing &Meyer, 2001; Davis et al., 2002; Lavin et al., 2004;Richardson et al., 2004; Bell & Donoghue, 2005; Lavin,Herendeen & Wojciechowski, 2005; Särkinen et al.,2007; McDade, Daniel & Kiel, 2008; Soza & Olmstead,2010a, b), and groups that have radiated recently inSouth America (e.g. Richardson et al., 2001; Andersson

& Antonelli, 2005; Andersson, 2006; Hughes & East-wood, 2006; Erkens, Maas & Couvreur, 2009; O’Learyet al., 2009; Lu-Irving & Olmstead, 2012) than aboutpatterns of diversification of indigenous groups thatoriginated and have a long history in South America. Afew examples are available that seem to fit thesecriteria, including Gesneriaceae subfamily Gesnerio-ideae (Smith et al., 1997; Zimmer et al., 2002; Weber,2004; Perret et al., 2012), Bromeliaceae (Givnish et al.,2004, 2011), a large clade of Cactaceae (Edwards,Nyffeler & Donoghue, 2005) and Asteraceae (Funket al., 2005). A well-documented example is Malpighi-aceae, which has a circumtropical distribution, but hasbeen shown to have diversified initially in SouthAmerica at least 64 Mya (Davis et al., 2002) and hascolonized the Old World as many as six times. Threeother large clades that originated and exhibited theirearly diversification in South America are the subject*E-mail: [email protected]

bs_bs_banner

Botanical Journal of the Linnean Society, 2013, 171, 80–102. With 6 figures

© 2012 The Linnean Society of London, Botanical Journal of the Linnean Society, 2013, 171, 80–10280

of this paper: Solanaceae (Olmstead et al., 2008),Bignoniaceae (Olmstead et al., 2009) and Verbenaceae(Marx et al., 2010).

Clade diversification may entail both the geographi-cal spread from a point of origin and adaptation tonovel environments along the way. The processesinvolved in successful migration and adaptationpresent a continuum from migration and diversifica-tion in closely similar environments, referred to as‘niche conservatism’ (Harvey & Pagel, 1991; Wiens &Donoghue, 2004), or ‘biome conservatism’ when con-sidered on a global scale (Crisp et al., 2009), to adap-tation to rather different environmental settings,referred to as ‘niche evolution’ (Wiens & Donoghue,2004). Niche conservation has been shown to be asso-ciated with recently diversified groups in SouthAmerica (e.g. Inga Mill. – Richardson et al., 2001;Coursetia DC – Lavin et al., 2003; Lavin, 2006; Vale-rianaceae – Bell & Donoghue, 2005; Lupinus L. –Hughes & Eastwood, 2006; Cinchoneae – Andersson& Antonelli, 2005; Calceolaria L. – Andersson, 2006)and may be a major determinant in the patternrecognized by Gentry (1982) that there is a significanttaxonomic division between Andean and Amazonianfloras. At the other end of this continuum, nicheevolution, the adaptive shifts between ecologicalzones, has been important in the adaptation of plantsinto geologically new ecosystems, such as oceanicislands (Carlquist, 1974; Baldwin et al., 1998) and thecerrado of South America (Simon et al., 2009), butmay prove to be a barrier to entry into establishedecosystems with well-developed floras (Antonelliet al., 2009; Antonelli & Sanmartin, 2011). Under-standing the course of diversification of large cladesand the balance between ecological shifts and diver-sification in similar habitats can offer insight into therelative importance of these processes.

As reviewed by Pennington & Dick (2004) and inthe present volume by Christenhusz & Chase (2012),explanations for intercontinental disjunctions involv-ing South American plant groups with distributionson other continents typically involve either relictGondwanan distributions (Gentry, 1982), migrationroutes involving northern hemisphere continentalconnections (Renner et al., 2001; Davis et al., 2002) orover-water, long-distance dispersal (Givnish et al.,2004; Renner, 2004; Särkinen et al., 2007). A continu-ous land connection to North America has existedwith certainty only since the closure of the Isthmus ofPanama (c. 3 Mya), so, for groups that originated inSouth America after the split up of Gondwana, exceptfor very recent migrants to North America, virtuallyall emigration required over-water dispersal (Codyet al., 2010). However, recent evidence (Farris et al.,2011; Montes et al., 2012) suggests that there mayhave been a continuous land connection as early as c.

20 Mya, even if it did not persist. An earlier connec-tion, if confirmed, would require re-evaluation ofclassic explanations for mammal (Simpson, 1980;Marshall et al., 1982) and, more recently, bird (Smith& Klicka, 2010) distributions, which have been linkedto the more recent date.

Over-water, long-distance dispersal requires avector; wind, rafting via ocean currents and transportby birds are the three most commonly postulatedmechanisms (Carlquist, 1967; Renner, 2004; Nathanet al., 2008; Baldwin & Wagner, 2010). A summary ofmechanisms for transport of plants to Hawaii, one ofthe most remote oceanic island archipelagos, whichwas populated (prior to arrival of the first Polyne-sians) entirely by long-distance dispersed propagules,indicates that transport by birds can account for c.75% of all immigrants, with wind and ocean driftaccounting for the remainder (Gillespie et al., 2012).Disjunctions between North and South America havealso been attributed mainly to transport by birds(Carlquist, 1967; Raven, 1972; Wen & Ickert-Bond,2009; Cody et al., 2010). However, Renner (2004)reviewed the putative long-distance dispersed dis-juncts between tropical Africa and South America andconcluded that dispersal by rafting on ocean currentsis likely to have been more common than dispersal bybirds or wind. Whereas wind can be an effectivemechanism for seed dispersal in some species,reviews suggest that it is rarely effective for long-distance, over-water dispersal (Renner, 2004; Nathanet al., 2008; Gillespie et al., 2012). From a local, eco-logical perspective, animal-dispersed fruits/seeds,including fleshy fruits or fruits/seeds with mecha-nisms for sticking to the body of an animal, areconsidered to be more effective for medium- to long-distance seed dispersal than are fruits/seeds lackingany such mechanism (Nathan et al., 2008), althoughsuch fruits/seeds may have an alternative dispersalmechanism, such as hitch-hiking on a rafted sub-strate, and modelling long-distance dispersal isfraught with difficulty (Levin et al., 2003).

South America has a long history of geographicalisolation from other continental land masses and hasa biota that reflects that isolation, exemplified mosticonically by its mammal fauna (Simpson, 1980). Itslink to Gondwana was severed about 100 Mya whenSouth America and Africa separated and began todrift apart (Goldblatt, 1993; Burnham & Graham,1999; Morley, 2003), even though east–west runningarchipelagos (Walvis Ridge/Rio Grande rise andSierra Leone rise) derived from volcanic hotspots onor near the mid-Atlantic ridge may have permittedstepping stone migration perhaps until c. 70 Mya (e.g.Morley, 2003; Pennington & Dick, 2004), when themid-Atlantic ridge moved off the stationary hotspotthat formed the Walvis Ridge/Rio Grande rise

SOLANACEAE, VERBENACEAE AND BIGNONIACEAE 81


(O’Connor & Duncan, 1990). After the opportunitiesfor short-distance, over-water migration with Africawaned in the Cretaceous and early Tertiary, suchopportunities arose again, this time with NorthAmerica, in the mid Tertiary when blocks of theCaribbean Plate (proto-Greater Antilles) emergedbetween South and North America at c. 50–40 Mya,via GAARlandia (Greater Antilles and Aves Ridge) atc. 35–33 Mya (Morley, 2003; Pennington & Dick,2004; Pirie et al., 2006) or via a Central Americanpeninsula that reached close to north-western SouthAmerica at c. 20–15 Mya (Cody et al., 2010; Monteset al., 2012). Finally, a continuous land connectionarose between South and North America when theIsthmus of Panama formed, connecting the two con-tinents at c. 3 Mya (Keigwin, 1982; Coates & Obando,1996; Burnham & Graham, 1999). Unlike the verylimited dispersal capabilities of mammals, however,the potentially much longer history of plant dispersalto and from South America makes understandingthe origin of contemporary floras more complicated(Pennington & Dick, 2004; Renner, 2004; Wen &Ickert-Bond, 2009; Cody et al., 2010).

Phylogenetic studies have been published for threeplant families that are widely distributed throughoutSouth America, Solanaceae (Olmstead et al., 2008),Bignoniaceae (Olmstead et al., 2009) and Verbenaceae(Marx et al., 2010). Each family has substantialalthough much smaller representation in NorthAmerica, and each is also represented on other con-tinents. All three families occur in a broad rangeof ecosystems and exhibit a range of dispersalattributes, including animal and non-animal vectors.The phylogenetic hypotheses for these three majorclades of plants suggest that all three originated anddiversified initially in South America, thus providingnatural replicate ‘experiments’ for investigating plantdiversification in South America and their spread toNorth America and beyond. Whereas well-sampled,dated phylogenies are not yet available for any ofthese three families, higher level dating studies thathave included members of these clades suggest thatthe crown ages of each are younger than the separa-tion of South America from Gondwana (Wikström,Savolainen & Chase, 2001; Bremer, Friis & Bremer,2004).

Here we use these phylogenies to explore commonpatterns of geographical distribution and establishexplanatory hypotheses that can form the basis forfuture studies. Major clades identified in previouslypublished phylogenies for Solanaceae (Olmsteadet al., 2008), Bignoniaceae (Olmstead et al., 2009) andVerbenaceae (Marx et al., 2010) here provide replicatecase studies for a series of coarse-scale macro-evolutionary comparisons, including (1) the abilityto extend distributions throughout the western hemi-

sphere, (2) the importance of niche conservatism andadaptive shifts between ecological zones for explain-ing current distribution patterns, (3) the success oftransoceanic, long-distance dispersal to other conti-nents and, where possible, (4) the association of fleshyfruits (conventionally associated with transoceanic,long-distance dispersal via animal vectors) withdispersal to other continental areas.

MATERIAL AND METHODS

To infer the ancestral geographical area for Solanales,for which there is currently no single densely sampledstudy, a tree was assembled by combining resultsfrom several plastid DNA studies of Asteridae(Bremer et al., 2002; Soltis et al., 2011; N. Refulio &R. Olmstead, unpubl. data), Solanaceae (Olmsteadet al., 2008) and Convolvulaceae (Stefanovic, Krueger& Olmstead, 2002). For Solanaceae, Bignoniaceae andVerbenaceae, phylogenetic trees from previously pub-lished studies were used (Olmstead et al., 2008, 2009,and Marx et al., 2010, respectively).

Distribution ranges were assembled from mono-graphs and global treatments for families (e.g.Gentry, 1980, 1992a; Hunziker, 2001; Atkins, 2004;Fischer, Theisen & Lohmann, 2004) and specimendatabases (e.g. GBIF – http://data.gbif.org; TROPI-COS – http://www.tropicos.org). Northern and south-ern distribution limits in the New World weredetermined for each clade (not only those exemplarsincluded in the phylogenetic studies) to the nearest5° latitude (Table 1). Effort was made to excludenon-native ranges, where weedy species may haveexpanded their distributions through anthropogenicmeans.

RESULTS

The individual phylogenetic analyses included nearlycomplete generic level sampling (Solanaceae 91%,Bignoniaceae 85%, Verbenaceae 97%), but samplingat the species level was only representative andranged from 7.5% (Solanaceae) to 12% (Bignoniaceae)and 13% (Verbenaceae). This inevitably means thatthe diversity of large genera is underrepresented.In each family, many of these species-rich groupshave been subject to more fine-grained phylogeneticstudies, but for the purposes here, general patternswere inferred from the available sampling in pub-lished higher-level phylogenetic analyses with addi-tional inferences drawn from species-level studieswhere needed.

Solanales comprise three main clades (Fig. 1), witha small clade composed of Montiniaceae, SphenocleaGaertn. and Hydrolea L. sister to a much larger cladecomprising Convolvulaceae sister to Solanaceae.

82 R. G. OLMSTEAD


Solanaceae include about 100 genera and c. 2500species (Olmstead & Bohs, 2007) and are distributedworldwide, but by far the greatest diversity occursin the Neotropics, especially in South America. Theplants are mostly small trees, shrubs and herbs,and several of the world’s most important non-cerealcrops belong to Solanaceae (e.g. potatoes, tomatoes,chili peppers, eggplants). Approximately half of the

species reside in one genus, Solanum L., which isone of the largest genera of flowering plants andwhich has a nearly cosmopolitan distribution.Solanaceae have evolved a diversity of alkaloids forresistance to herbivory, which have been exploitedfor medicinal (atropine), psychotropic (nicotine) andculinary (capsaicin) purposes, among others (Heiser,1969).

Table 1. Major clades of Solanaceae, Bignoniaceae and Verbenaceae, with presence in South America, North America andoutside of the New World, total number of species in each clade, and latitudinal distribution in the western hemisphereestimated to the nearest 5° (Y = present; N = absent)

CladeS.Am.

N.Am.

OldWorld

No. ofspecies

Latitudinalrange

SolanaceaeSchizanthus Y N N 12 40°S–30°SGoetzeoideae/Metternichia/Tsoala Y Y Y 9 25°S–25°NCestroideae Y Y N 206 35°S–30°NBenthamielleae Y N N 15 55°S–30°SPetunieae Y Y N 145 50°S–30°NSchwenckieae Y Y N 28 30°S–30°NNicotiana Y Y Y 76 50°S–50°NAnthocercideae N N Y 32Lycieae s.l. Y Y Y 229 55°S–40°NHyoscyameae N N Y 43Juanulloeae s.l. Y Y N 50 25°S–25°NMandragora N N Y 2Datureae Y Y N 18 30°S–30°NSalpichroina Y Y N 6 35°S–30°NCuatresia Y Y N 11 15°S–20°NWithaninae Y N Y 42 30°S–15°SPhysalinae/Iochrominae Y Y Y 187 35°S–50°NCapsiceae Y Y Y 231 35°S–35°NSolaneae Y Y Y (~5¥) 1378 50°S–50°N

BignoniaceaeJacarandeae Y Y N 52 30°S–20°NTourrettieae Y N N 4 40°S–10°NTecomeae/Argylia Y Y Y (2¥) 57 45°S–40°NDelostoma Y N N 4 15°S–5°NPalaeotropical clade N N Y 151Tabebuia alliance Y Y N 155 30°S–30°NOroxyleae N N Y 6Catalpeae N Y Y 11 20°N–45°NBignonieae Y Y N 392 35°S–35°N

VerbenaceaePetreeae Y Y N 12 25°S–25°NDuranteae Y Y Y 191 30°S–30°NCasellieae Y Y N 20 35°S–25°NCitharexyleae Y Y N 134 30°S–30°NPriveae Y Y Y 21 35°S–30°NRhaphithamnus Y N N 2 45°S–30°SNeospartoneae Y N N 7 45°S–25°SVerbeneae Y Y Y 177 55°S–50°NLantaneae Y Y Y (3¥) 276 50°S–40°N



Divergence time estimates for the crown cladeSolanaceae range from a minimum of c. 35 Mya(Dillon et al., 2009), c. 40 Mya (Wikström et al., 2001)or c. 51 Mya (Paape et al., 2008) and for stemSolanaceae a minimum of c. 65 Mya (Wikström et al.,2001) or 85 Mya (Bremer et al., 2004). The fact thatthe closest relatives of Solanaceae diversified initiallyon the different continental plates derived from theinitial separation of Gondwana (Fig. 1) suggests thatdiversification of Solanaceae coincided with the isola-tion of South America from the Late Cretaceousthrough the Tertiary.

Phylogenetic analyses of Solanaceae (Olmstead &Palmer, 1992; Olmstead & Sweere, 1994; Olmsteadet al., 1999, 2008) have resolved generic relationshipsand provided the basis for a revised classification forthe family (Olmstead & Bohs, 2007). For the purposeshere, 19 clades, ranging in size from as few as two toas many as c. 1400 species, are recognized (Table 1;Fig. 2). Sixteen of the 19 clades are represented inthe New World, with three clades found only outsidethe New World (Anthocercideae – Australia; Hyo-scyameae – Eurasia; Mandragora – Eurasia). All ofthe 16 remaining clades have South American repre-sentatives; of these, 13 have representatives on theNorth American continent or in the Greater Antillesand seven have Old World representatives. One smallclade, Withaninae, with about nine genera, butonly 42 species, occurs in South America, Africa, theMediterranean basin, East Asia and three widelyspaced oceanic islands or archipelagos (Olmsteadet al., 2008). Since the emphasis on sampling was forgeneric diversity, Solanum, with about half of allspecies in the family, is here very underrepresented.However, numerous detailed phylogenetic studies of

Solanum extend these results by providing additionalphylogenetic and geographical resolution in the genus(e.g. Bohs & Olmstead, 2001; Bohs, 2005; Levin,Myers & Bohs, 2006; Weese & Bohs, 2007). Othergenera that were substantially underrepresented, butwhich also have more detailed phylogenies, includeNicotiana L. (Clarkson et al., 2004), Nolana L. exL.f. (Dillon et al., 2007, 2009), Jaltomata Schldl. (R.Miller et al., 2011), Physalis L. (Whitson & Manos,2005) and Lycium L. (J. Miller, Kamath & Levin,2009; J. Miller et al., 2011).

Bignoniaceae include about 80 genera and c. 800species (Lohmann & Ulloa, 2007; Olmstead et al.,2009). They are primarily trees, shrubs and lianasand are important components of Neotropical commu-nities, with secondary distributions in tropical regionsof Africa and Asia. The large clade, Bignonieae, con-sists largely of lianas, which form an important com-ponent of Neotropical forest canopies. The family alsois abundant in tropical savannas, including thecerrado of South America, but only a few lineageshave established in temperate zones in the northernand southern hemispheres. Species of several genera(e.g. Tabebuia Gomes ex D.C., Jacaranda Juss.,Spathodea P.Beauv.) are widely planted tropical orna-mentals and many species have had indigenous uses(Gentry, 1992b).

Estimated ages for the crown clade Bignoniaceaeare a minimum of c. 40 Mya (Wikström et al., 2001) toc. 45 Mya (Nie et al., 2006), although Lohmann,Winkworth & Bell (2012) provide an age estimate forthe included clade Bignonieae of 40–68 Mya, andminimum ages of the stem clade range from 47 Mya(Wikström et al., 2001) to 68 Mya (Bremer et al.,2004). Bignoniaceae belong in a part of Lamiales

Figure 1. Phylogenetic tree of Solanales (see text for sources used to depict relationships). The three main branches areeach inferred to have their earliest crown divergence in one of the three main Gondwanan continental fragments.

84 R. G. OLMSTEAD


Figure 2. Phylogenetic tree of Solanaceae adapted from Olmstead et al. (2008). Clades in Solanaceae and theirdistributions are indicated. Arrows indicate Old World dispersal events.



where resolution is still lacking (Schäferhoff et al.,2010; Soltis et al., 2011; N. Refulio & R. Olmstead,unpubl. data) and it is still unclear what constitutesthe sister group of Bignoniaceae.

Phylogenetic analyses of Bignoniaceae (Spangler &Olmstead, 1999; Zjhra, Sytsma & Olmstead, 2004;Lohmann, 2006; Grose & Olmstead, 2007a; Olm-stead et al., 2009) have established generic relation-ships and a new classification for Bignoniaceae isemerging (Grose & Olmstead, 2007b; Olmstead et al.,2009; Lohmann, in press). Nine clades are recog-nized for the purposes of the present study, rangingin size from four to nearly 400 species (Table 1;Fig. 3). Seven of the nine clades are represented inthe New World, with two exclusively Palaeotropicalin distribution (Oroxyleae – SE Asia; Palaeotropicalclade – Africa, Asia, Australia). Of the seven cladeswith New World distributions, six are represented inSouth America and, of these, three include repre-sentatives in North America or the Antilles and oneincludes Old World representatives. The seventhclade, Catalpeae, occurs exclusively in the northernhemisphere and includes North American, Antilleanand Asian species. As with Withaninae (Solanaceae),one small clade of Bignoniaceae, tribe Tecomeae (12genera, c. 55 species), is exceptionally widely distrib-uted, occurring in South America, North America,Africa, Central Asia, SE Asia and Australia (Olm-stead et al., 2009). More densely sampled phyloge-netic analyses are available for several of the largerclades, including tribe Bignonieae (Lohmann, 2006),the Tabebuia alliance, including Crescentieae (Grose& Olmstead, 2007a), and tribe Coleeae (Zjhra et al.,2004).

Verbenaceae include about 35 genera and c. 900species (Atkins, 2004; Marx et al., 2010) of trees,shrubs, lianas and herbs. They are particularlyimportant components of arid to semi-arid communi-ties in North and South America, but also are presentin wet and dry tropical forests, high Andean grass-lands and cloud forests. A secondary centre of distri-bution is found in Africa. A few species [e.g. Lantanacamara L., Stachytarpheta jamaicensis (L.) Vahl.]have become widespread weeds in tropical regions,and several genera are grown as ornamentals (e.g.Glandularia J.Gmelin, Aloysia Palau, Lantana L.,Duranta L.).

The crown age of Verbenaceae has been estimatedto be a minimum of c. 40 Mya (Nie et al., 2006),whereas the stem age ranges from a minimum of48 Mya (Wikström et al., 2001) to 62 Mya (Bremeret al., 2004). Verbenaceae belong to the same part ofLamiales as Bignoniaceae, where resolution is stilllacking (Schäferhoff et al., 2010; Soltis et al., 2011;N. Refulio & R. Olmstead, unpubl. data). However,unlike Bignoniaceae, Verbenaceae seem to have a

clear sister group in the small west African genusThomandersia (Wortley, Harris & Scotland, 2007;Soltis et al., 2011; N. Refulio & R. Olmstead, unpubl.data, but see McDade et al., 2012).

Phylogenetic analyses of Verbenaceae (Yuan &Olmstead, 2008a; O’Leary et al., 2009; Marx et al.,2010; Yuan et al., 2010; Lu-Irving & Olmstead, 2012)have resolved generic relationships and a new classi-fication for Verbenaceae has been developed (O’Learyet al., 2009; Marx et al., 2010), with the exception ofgeneric circumscriptions and relationships in tribeLantaneae, which are in progress (P. Lu-Irving &R. Olmstead, unpubl. data). Nine well-supportedclades ranging in size from two to c. 275 species arerecognized for the purposes of the present study(Fig. 4; Table 1). All nine clades exhibit their greatestdiversity in South America, with seven having dis-tributions reaching North America. Four clades(Duranteae, Priveae, Verbeneae, Lantaneae) haveboth New and Old World representatives. Moredetailed phylogenetic studies are available for two ofthe most species-rich groups, Verbeneae (Yuan &Olmstead, 2008a, b; O’Leary et al., 2009) and Lanta-neae (Lu-Irving & Olmstead, 2012).

Distributional ranges for each clade in all threefamilies were plotted on north and south latitudinallimits (Fig. 5). Most clades fall within latitudinallimits of 35° north and south of the equator. A handfulof outliers represent mostly very small cladesrestricted to South America (e.g. Schizanthus Ruiz &Pav. – Solanaceae, Tourrettieae – Bignoniaceae andNeospartoneae – Verbenaceae) and one small north-ern hemisphere clade (Catalpeae – Bignoniaceae). Asecond set of outliers represent those clades that haveextended their ranges into higher latitudes; theseclades typically are large (e.g. Solaneae – Solanaceae,Tecomeae – Bignoniaceae, Lantaneae – Verbenaceae)and include Old World representatives.

DISCUSSION

Well-sampled phylogenetic analyses of large familiesincreasingly permit inferences about the geographicalorigins of those clades, and, thus, to importantinsights into the detailed biogeographical history ofthe earth’s flora. Digging deeply into the past becomesincreasingly difficult with age, because very old cladesoften have either broad or relictual distributions thatreflect complicated histories of plate tectonics, migra-tion and extinction.

The difficulty in resolving the branching order ofthe four primary lineages of Lamiidae (Lamiales,Solanales, Gentianales and Boraginaceae: Schäferhoffet al., 2010; Soltis et al., 2011; N. Refulio & R. Olm-stead, unpubl. data) suggests that Solanales andLamiales have approximately similar stem ages,

86 R. G. OLMSTEAD


Figure 3. Phylogenetic tree of Bignoniaceae adapted from Olmstead et al. (2009). Clades in Bignoniaceae and theirdistributions are indicated. Arrows indicate Old World dispersal events; asterisk indicates dispersal back to SouthAmerica.



estimated at a minimum of c. 80 Mya (Wikströmet al., 2001; Magallón & Castillo, 2009) or c. 106 Mya(Bremer et al., 2004). Bignoniaceae and Verbenaceaerepresent two clades nested well within core Lamiales

(Olmstead et al., 2001; Schäferhoff et al., 2010),whereas Solanaceae represent one of three earlydiverging lineages in Solanales (Bremer et al., 2002;N. Refulio & R. Olmstead, unpubl. data).

Figure 4. Phylogenetic tree of Verbenaceae adapted from Marx et al. (2010). Clades in Verbenaceae and their distribu-tions are indicated. Arrows indicate Old World dispersal events.

88 R. G. OLMSTEAD


Lamiales also represent a phylogenetically chal-lenging clade, in which resolving the branching orderhas been difficult (Olmstead et al., 2001; Bremeret al., 2002; Schäferhoff et al., 2010; Soltis et al., 2011;N. Refulio & R. Olmstead, unpubl. data). However,the geographical origins of several large family-levelcrown clades of Lamiales (Oleaceae – Wallander &Albert, 2000; Acanthaceae – McDade et al., 2008;Bignoniaceae – Olmstead et al., 2009; Verbenaceae –Marx et al., 2010) have been determined and cancontribute to our understanding of plant diversifica-tion on a global scale. Two of these clades, Bignon-iaceae and Verbenaceae, have unequivocally SouthAmerican ancestral areas, and have diversifiedthroughout the New World, including North America,and have established toeholds on other continents,too, primarily in the southern hemisphere (Olmsteadet al., 2009; Marx et al., 2010).

Solanales represent a much smaller clade of similarage to Lamiales. However, the relationship among thethree primary clades of this order are well resolvedand robustly supported (Fig. 1), with each of the threecrown clades inferred to have different southernhemisphere ancestral areas, the Montiniaceae/Sphenoclea/Hydrolea clade in Africa, Convolvulaceaemost likely in the continental fragment knownas East Gondwana (SE Asia/India/Madagascar/

Antarctica) and Solanaceae in South America. ThusSolanales may present a classic case of vicariancefollowing Gondwanan breakup, with its descendantson each of the three initial continental fragmentsgiving rise to the three extant clades.

Solanaceae, Bignoniaceae and Verbenaceae thusrepresent replicated comparative datasets for inves-tigating the diversification of plants with origins inSouth America. Divergence time estimates for thecrowns of all three family clades (going as far back asthe common ancestor of all extant representatives, asdistinct from the stem clades, which go back to thedivergence of each lineage from its nearest livingrelatives) are inferred to be in the Late Cretaceous orearly Tertiary, after the separation of South Americafrom Africa (Wikström et al., 2001; Bremer et al.,2004), with Solanaceae inferred to be somewhat olderthan Bignoniaceae or Verbenaceae (Wikström, Savol-ainen & Chase, 2001; Bremer et al., 2004). Thus eachfamily represents a lineage probably derived fromGondwanan ancestors the diversification of which hascoincided with the geological and climatologicalchanges that South America has experienced duringits long existence spanning almost 100 million yearsas an isolated continent (Gentry, 1982; Burnham &Graham, 1999; Antonelli et al., 2009; Antonelli & San-martin, 2011).

Phylogenetic studies of each family form the basisfor revised classifications that recognize well-supported clades as taxa. Although these clades areall of different ages, the absence of dated phylogenetictrees constrains the potential for testing hypotheses.However, these multiple, more or less independentgroups can still be used to explore broad continental-scale geographical patterns of diversification withinand beyond South America. The latitudinal extent towhich multiple clades have been able to spread, theability to spread into North America and to othercontinents and the degree to which niche conserva-tism constrains the habitat diversity occupied bymembers of each clade can be examined for eachclade. These patterns provide more explicit hypoth-eses for testing with more densely sampled, time-calibrated phylogenies.

SOLANACEAE

South American origin and diversificationThe 16 clades of Solanaceae present in the New Worldall have their origins in South America. The onepossible exception is the Goetzeoideae/Metternichia/Tsoala clade, which has one lineage in Brazil, one inthe Greater Antilles and one in Madagascar. However,evidence suggests that Amazonian DuckeodendronKuhlm. is sister to this clade (Santiago-Valentin &Olmstead, 2003; Olmstead et al., 2008), and all other

Southern limit(latitude)

Northern limit(latitude)

10 20 30 40 50 60ºS

60ºN

50

40

30

20

10

30ºN 20 10

10

20

30ºS

Figure 5. Graph depicting the latitudinal limits for the32 New World clades found in Solanaceae (red), Bignon-iaceae (green) and Verbenaceae (blue). The bold dashedline represents the equal northern and southern latitudi-nal limits, with the fine dashed lines representing 5°latitude on either side. Lines at 35° latitude roughlyrepresent the 10 °C minimum average monthly tempera-ture for the coldest month (January in the north; July inthe south).



early diverging lineages of Solanaceae in this part ofthe phylogenetic tree are also South American inorigin, so a South American ancestral area for thisclade is also likely. Most of the early diverging line-ages in Solanaceae are either temperate SouthAmerica [Schizanthus, Benthamielleae, Petunieae,Lycieae sensu lato (s.l.)] or Andean (Cestroideae, Nico-tiana) in origin. Tropical members of these clades andclades with primarily tropical distributions are allrelatively derived in Solanaceae.

Distribution in North America and the CaribbeanOnly three clades (Schizanthus, Benthamielleae,Withaninae) have not reached North America (Fig. 5)and all three are small, with c. 12–15 species each inSouth America (Withaninae are also represented inAfrica and Asia). Schizanthus and Benthamielleaeare found only in far southern South America,whereas Withaninae (Athenaea Sendt. and AurelianaSendt.) are found only in the Atlantic coastal forestsof Brazil.

There is a striking consistency with respect to thenorthern and southern distribution limits for theclades that have successfully reached NorthAmerica. Given that all of the 16 New World cladesapparently originated in South America and south-ern South America in many cases, the same distri-bution limits seem to have been reached in mostclades, regardless of clade age. Of the 13 clades thathave reached North America, ten have north/southlatitudinal limits that are within 5° of each other onboth sides of the equator (Table 1; Fig. 5), and allexcept Lycieae have more or less continuous distri-butions through the Tropics. Lycieae exhibit a dis-junct distribution between the arid regions oftemperate South and North America (a pattern seenagain in Verbenaceae). Exceptions include Phy-saleae, which has diversified in colder regions oftemperate North America, where Physalis extendsto 50°N (c. 15° further north than the clade occursin the south), and Petunieae, which only reaches c.30°N, but in which Fabiana Ruiz & Pav. extends asfar as 50°S. Whereas several clades do extend tolatitudes further than c. 30–35° north or south ofthe equator, a relatively small number of species,perhaps no more than 125, or c. 5% of New Worldspecies occur at those latitudes. To the extent thatlatitude is a placeholder for ensemble environmentalconditions, or something as simple as the latitudinallimits of a frost-free zone (Thompson, Anderson &Bartlein, 2000; Wiens & Donoghue, 2004), thedistribution limits for the clades of Solanaceaesuggest that for most clades, successful migrationin the New World has reached its full poten-tial geographical ranges, within the constraints of

the ecological preferences of the plants in eachclade.

Niche conservatism vs. adaptation to novelecological zones in the diversificationof SolanaceaeSolanaceae can be found in virtually all terrestrialecosystems in South America, evidence of a historythat involved numerous ecological shifts during theirdiversification. Similarly, many of the clades identi-fied here for the purposes of comparison includemembers found in two or more distinct habitats.Most of the larger clades (e.g. Solaneae, Capsiceae,Physalinae/Iochrominae, Petunieae, Nicotiana) havesignificant representation in both xeric and mesicecological zones and broad latitudinal ranges, reach-ing into cool temperate climates, providing furtherevidence that adapting to novel ecological settingshas been an important process in diversification ofSolanaceae. However, at the generic level, mostgroups do exhibit significant levels of niche conserva-tism, and even a couple of the larger clades showconsistent, if contrasting, environmental preferences.The Lycieae/Nolana/Sclerophylax Miers clade isrestricted to arid habitats, with Lycium mostly in cooltemperate habitats, distantly disjunct between Northand South America, Sclerophylax Miers in the aridhigh Andes in northern Argentina, and Nolana occu-pying coastal lomas. Cestreae, by contrast, arerestricted to mostly mesic, tropical settings. Assampling for broad, family-level phylogenetic treat-ments of Solanaceae necessarily under-samplesspecies diversity within genera, the relative impor-tance of niche conservation vs. niche evolution willprobably be underestimated by examining trends ongeneric-level phylogenetic trees. Much more denselysampled phylogenetic trees will be needed to quantifythe balance between these two contrasting modes ofdiversification in Solanaceae.

Transoceanic dispersalTen clades of Solanaceae have successfully colonizedother parts of the world outside of North and SouthAmerica (prior to European exploration), includingthe three clades that do not occur in the New World(Anthocercideae, Hyoscyameae, Mandragora L.), eachof which is inferred to have descended from a NewWorld immigrant. A total of c. 15–17 pre-Columbiandispersal events are needed to account for the presentdistribution, with about five to seven events inSolanum (Bohs, 2005; Olmstead et al., 2008) and twoin Lycium (J. Miller et al., 2009, 2011). In the absenceof a dated tree, the best one can do is infer relativeage of dispersal events on the basis of depth in thetree where the event was inferred to occur (Fig. 2).Many of the postulated dispersal events are clearly

90 R. G. OLMSTEAD


recent in origin (a single species of Physalis,P. alkekengi L., in East Asia, a Hawaiian endemicsubspecies, Lycium carolinianum Walt. subsp. sand-wicense (Gray) C.L.Hitchcock, of a North Americanspecies, and a species of Solanum section Geminata(G.Don) Walp. in Australia, an otherwise entirelyNew World group of c. 140 species), whereas othersrepresent potentially relatively old dispersal events(e.g. Tsoala in Madagascar, Anthocercideae in Aus-tralia, Hyoscyameae in Eurasia).

Withaninae present a microcosm of Solanaceaedispersability. With c. 40 species assigned to ninegenera, their distribution includes Brazil, Africa, thewestern Mediterranean, Central Asia and China, aswell as remote oceanic islands, such as the Canaries,Hawaii and St Helena. The clade probably originatedin tropical South America and dispersed east to Africaor the Mediterranean initially, and then on to EastAsia and Hawaii (Olmstead et al., 2008), although twoAsian genera have not been sampled in previousphylogenetic studies. Despite an origin in SouthAmerica, and an apparent vagility that has permittedcolonization of remote oceanic islands and widelydisjunct localities in Africa and Eurasia, the distri-bution of Withaninae in the New World remainsrestricted to the Atlantic coastal forest of Brazil, incontrast to so many other clades that have colonizedmuch of the New World.

The phylogeny of Solanum is under active investi-gation (Bohs, 2005; Levin et al., 2006; Weese & Bohs,2007; Stern, Agra & Bohs, 2011; L. Bohs, pers. comm.),but it is too early to delineate all postulated dispersalevents in that group confidently. The dispersal eventout of the New World that has resulted in the greatestradiation in the Old World has occurred in subgenusLeptostemonum (Dunal) Bitter, the spiny solanums,which are worldwide in distribution and include aclade of c. 175 species widely distributed in the OldWorld. However, in addition to this major diversifica-tion in Leptostemonum following dispersal to the OldWorld, the relatively ancient clade comprising thenon-spiny African species, the Normania Lowe groupfrom the Mediterranean and the Canary Islands (assegregate genera until Bohs & Olmstead, 2001) andthe Australian ‘kangaroo apples’ (Symon, 1994),Solanum section Archaesolanum Marzell (Weese &Bohs, 2007; Poczai, Hyvönen & Symon, 2011) alsorepresents a significant Old World lineage. The latterclade appears to have involved initial radiation inAfrica, followed by dispersal east to Australia, whereasthe direction of dispersal is uncertain in section Lep-tostemonum, where resolution remains poor. Addi-tional dispersal events in Solanum are likely to berelatively recent and involve small groups in thedulcamaroid (Eurasia) and Morelloid (Europe, Africaand Asia) clades, or individual species (S. aculeatissi-

mum Jacq. in Africa; S. lasiocarpum Dunal andS. repandum G.Forst. in the South Pacific; S. spiraleRoxb. in Australia/SE Asia), which in some cases maybe early anthropogenic dispersal events. Thus, thereappear to have been at least four and possibly as manyas seven or eight pre-Columbian long-distance disper-sal events in Solanum.

Using simple parsimony to determine the directionof dispersal events overstates the confidence withwhich we can discuss long-distance dispersal historyin any clade, and Solanaceae are no exception.However, based on current evidence, long-distancedispersal in Solanaceae appears to have been both tothe east and west (Fig. 6). At least four events arepostulated to have been to the east from SouthAmerica to Africa or Madagascar: (1) Tsoala to Mada-gascar; (2) Lycium to Africa (and on to Asia andAustralia); (3) Withaninae to Africa (and on to Asiaand Hawaii); and (4) the Solanum non-spiny African/Normania/Archaesolanum to Africa (and on to Asiaand Australia). At least five events seem to have beento the west: (1) Nicotiana subgenus Suaveolentes toAustralia; (2) the ancestor of Anthocercideae to Aus-tralia; (3) Lycium carolinianum to Hawaii (where it isrecognized as L. carolinianum var. sandwicense); (4)Physalis alkekengi to East Asia; and (5) a clade withinLycianthes (Dunal) Hassler to SE Asia. Hyoscyameae,Mandragora and at least three of the clades inSolanum have distributions that extend acrossEurasia or Africa/Asia/Australia, such that no simpleinferences are possible as to the directionality of thedispersal events responsible for them. The recentindividual species dispersals in Solanum, if theyare not of anthropogenic origin, have gone in bothdirections.

Association of seed dispersal withlong-distance dispersalThere is a fundamental split in Solanaceae betweenfleshy, animal-dispersed fruits and dry dehiscent cap-sular fruits with no evident means of seed dispersal.Along with a few other characteristics, this traitdefines the two subfamilies Solanoideae (fleshy fruits)and Cestroideae (dry fruits) in traditional classifica-tions (although Cestrum L., itself, has berries). Phy-logenetic studies of Solanaceae (e.g. Olmstead &Palmer, 1992; Olmstead et al., 2008) have shown thatCestroideae form a paraphyletic grade from whichSolanoideae are derived. Each of the 19 clades con-sidered here can be characterized by either fleshy ordry dehiscent fruits, with only minor exceptions. Atfirst glance, it appears that there is a close associationbetween long-distance, intercontinental dispersal andfleshy fruits, with 12 of the c. 15 events (80%) occur-ring in fleshy-fruited clades. However, approximately87% of species in Solanaceae belong to fleshy-fruited



lineages, so the distribution of long-distance dispersalamong clades of Solanaceae does not seem to reflect abias towards fleshy-fruited clades.

BIGNONIACEAE

South American origin and diversificationOf the seven New World clades in Bignoniaceae, fourcan confidently be inferred to be of South Americanorigin (Tourrettieae, Tecomeae/Argylia D.Don, Delos-toma D.Don, Bignonieae). These clades represent

three of the earliest diverging lineages of Bignon-iaceae, with Bignonieae being more derived. Jacaran-deae, sister to the rest of the family, are widespreadin lowland tropical wet and dry forest and savannasthroughout the Neotropics. Their phylogeny is poorlyresolved (R. Farias & L. Freitas, unpubl. data), andincludes two widespread South American species, theranges of which extend into southern CentralAmerica, and one clade of six Antillean species.Although the phylogenetic analysis does not permitan unambiguous ancestral area, the Antillean species

Figure 6. Direction of transoceanic migration events in the history of Solanaceae (red), Bignoniaceae (green) andVerbenaceae (blue) divided into those inferred to originate in North America and South America. Only events wheredirection can be inferred are indicated; direction for six to eight additional events in Solanaceae and one in Bignoniaceaecannot be inferred.

92 R. G. OLMSTEAD


belong to the putatively derived Jacaranda Juss. sub-genus Jacaranda (= Monolobos DC.), characterized bythe derived condition of having one anther sacmissing; all species of subgenus Dilobos DC. arerestricted to South America. Therefore, the geographi-cal origin of Jacarandeae was most likely in SouthAmerica. Argylia (sister to Tecomeae), Tourrettieaeand Delostoma are all endemic Andean groups.Tecomeae (exclusive of Argylia) are of New Worldorigin, although most of the species diversity is inAustralasia, Africa and the Himalayas. The first twodiverging branches of Tecomeae are a temperatenorthern hemisphere taxon (Campsis Lour.), with onespecies in North America and one in east Asia, and aprimarily Andean taxon (Tecoma Juss.; one wide-spread species extends into Central America), withthe remainder of the clade in the Old World (with oneexception, see below), thus the geographical originseems uncertain. However, with modest support forthe southern Andean taxon Argylia as the sister toTecomeae and the evidence that all the other basalbranches of Bignoniaceae (Jacarandeae, Tourrettieaeand the isolated Delostoma) are South American inorigin would also place the origin of Tecomeae inSouth America.

The remaining clade found in South America is theTabebuia alliance, which comprises the large taxonTabebuia s.l., now split into three genera, Handroan-thus Mattos, Roseodendron Miranda and Tabebuia(Grose & Olmstead, 2007b), Crescentieae and a fewother small Neotropical genera. Crescentieae arenested within this clade and are Central American inorigin. There is also a Caribbean radiation in Tabe-buia sensu stricto (s.s.), but much of the remainingdiversity, including two small lineages that emerge ator near the base of the clade (Sparattosperma Mart.ex Meisn. and the Cybistax Mart. ex Meisn./Godmania Hemsl./Zeyheria Mart. clade) are SouthAmerican. The sister to this clade is a strictly Pal-aeotropical clade, but poor resolution in this partof the phylogenetic trees for Bignoniaceae meansthat the sister to the inclusive clade remainsunknown. Thus, there remains uncertainty about theancestral area of this clade, but it is probably alsoSouth American.

Distribution in North America and the CaribbeanOf the six clades of Bignoniaceae with putative SouthAmerican ancestral areas, four also include elementsin North America. The exceptions are two smallAndean clades, Tourrettieae, which includes three orfour species of vines found in the Andes from Chile toColombia, and Delostoma, with four species of trees inthe north-central Andes. As with Solanaceae, theclades that have reached North America seem to havesimilar northern and southern latitudinal limits

(Table 1; Fig. 5). The two largest such clades, Bigno-nieae and the Tabebuia alliance, each have nearlyidentical northern and southern limits (35°S/N and30°S/N, respectively), whereas the Tecomeae/Argyliaclade has northern and southern limits of 45°S/40°Nand Jacarandeae extends c. 10° further south of theequator than north (30°S/20°N). Bignoniaceae areapparently the most constrained of these three fami-lies to warm tropical regions, with perhaps no morethan seven or eight species (c. 1% of New Worlddiversity) occurring further than c. 35° north or southof the equator. This pattern is consistent with thetropical niche conservation model of Wiens & Dono-ghue (2004). Catalpeae are the only clade recognizedfor the purposes of this study in any of the threefamilies that are restricted to the northern hemi-sphere; their distribution includes the deserts ofsouth-west North America, the Greater Antilles andmesic forests in eastern North America and east Asia.Lack of resolution prevents determining what theclosest relatives of Catalpeae are, with Oroxyleaefrom SE Asia as a possible candidate (Fig. 3) andBignonieae and the Tabebuia alliance/Palaeotropicalclade as other possibilities. Thus, it is uncertain atthis point whether it is derived directly from SouthAmerican ancestors or not.

The pattern observed in Bignoniaceae suggests thatrelatively old dispersals to North America haveoccurred, with many recent, post-Panama land bridgedispersals. Relatively ancient dispersals include Cat-alpeae, Campsis, one lineage of Jacaranda in theAntilles (R. Farias & L. Freitas, unpubl. data) andperhaps three lineages of the Tabebuia alliance(Ekmanianthe Urb. and a large clade of Tabebuia s.s.in the Antilles and Crescentieae in Central America).

Putative post-isthmian dispersals include one wide-spread Tecoma sp., two Jacaranda spp., eight speciesof Tabebuia s.l. and several widespread species ofBignonieae spanning both South and Central America(Gentry, 1982, 1992a). Also, three or four species ofCentral American Crescentieae have ranges thatextend into northern South America, perhaps theresult of mammal migrations made possible by theclosure of the Isthmus of Panama, which assistedwith dispersal of their large, fleshy, mammal-dispersed fruits (Janzen & Martin, 1982; Grose &Olmstead, 2007a).

Niche conservatism vs. adaptation to novelecological zones in the diversificationof BignoniaceaeMuch of the early diversification in Bignoniaceaeoccurred in Andean South America, perhaps in theseasonally dry habitats where Tourrettieae, Argylia,Tecoma and Delostoma all occur today (Gentry,1992a). Even though much of the Andes represent a



relatively recent physiographic province in SouthAmerica (Antonelli et al., 2009; Antonelli & Sanmar-tin, 2011), there is both phylogenetic and fossil evi-dence for older plant diversification in the Andes,especially for Andean dry forest clades (Penningtonet al., 2004, 2010; Särkinen et al., 2012). That beingsaid, Jacarandeae, which comprise the sister to therest of Bignoniaceae, are largely absent from theAndes, where only two species occur, and are found ina variety of low-elevation ecosystems, including wetAtlantic coastal and Amazonian forests and dry orfire-prone ecosystems such as cerrado and the edaphi-cally dry coastal restingas (Gentry, 1992a). However,limited resolution and support in the phylogenetictrees for Jacarandeae (R. Farias & L. Freitas, unpubl.data) currently provide few insights into the histori-cal biogeography of this clade. The Tabebuia allianceis also found in diverse ecosystems, including wetand dry forests, and savannas, including the cerrado.Bignonieae are common elements of the liana flora inwet tropical forests, including Amazonian and Atlan-tic coastal forests, and probably originated in thoseecosystems, but have successfully colonized the dry,fire-prone cerrado of Brazil, most probably once, butperhaps multiple times (Lohmann et al., 2012). Thusthe history of diversification of Bignoniaceae appar-ently represents a complex of radiations withinbiomes (e.g. Andean seasonally dry forests, wet tropi-cal forests, cerrado), punctuated by ecological shiftsinto new habitats, including several lineages in thecerrado (Jacaranda, Bignonieae, Tabebuia) thatmostly probably represent niche shifts from moremesic habitats (Simon et al., 2009). Despite theapparent ability of several clades of Bignoniaceae toshift ecological habitats, very few lineages, perhapsfive, representing a handful of species in Tecomeae/Argylia, one in Tourrettieae, one in Bignonieae and c.10 in Catalpeae occur outside the tropics or subtrop-ics, where cold tolerance is thought to present one ofthe more difficult adaptive barriers to cross (Wiens &Donoghue, 2004; Donoghue, 2008).

Transoceanic dispersalFive to six long-distance dispersal events need to beinvoked to account for the Old World distribution ofBignoniaceae (Fig. 6). Two clades of Bignoniaceae areOld World in distribution, Oroxyleae, with fourgenera and six species restricted to SE Asia, and alarger Palaeotropical clade distributed in Africa, Aus-tralasia and Madagascar. As noted above, poor reso-lution renders the origin of Oroxyleae uncertain,although it undoubtedly represents an independentdispersal event from the New World. The Palaeotropi-cal clade is sister to the New World Tabebuia alliance,with which it is part of a poorly resolved section of thetree also including Oroxyleae and New World Bigno-

nieae and Catalpeae. Its origin is most likely NewWorld and South American. Resolution at the base ofthe Palaeotropical clade is poor and sampling of Asiantaxa, in particular, is limited, so it is uncertainwhether dispersal originally went east or west fromSouth America.

The two clades that originated first in the NewWorld and now have both New and Old Worldmembers each have unique biogeographical histories.Crown Catalpeae originated in North America, withChilopsis D.Don (and perhaps Astianthus D.Don) insouth-western USA and arid regions of CentralAmerica, and Macrocatalpa (Griseb.) Britton in theAntilles (Li, 2008; Olmstead et al., 2009) as succes-sively closer sister groups to Catalpa Scop., whichoccurs in eastern North America and East Asia. TheAsian species of Catalpa represent two lineages (Li,2008), suggesting either two dispersal events to EastAsia, or one dispersal to Asia followed by dispersalback to North America.

Tecomeae are similar to Withaninae (Solanaceae),in that a small clade of c. 12 genera and 50 specieshas dispersed to all habitable continents exceptEurope (South and North America, Africa, HimalayanAsia, SE Asia, Australia). In the New World, theTecomeae/Argylia clade is represented by Argylia andCampsidium Seem. in the southern Andes, Tecoma inthe central/northern Andes (with one widespreadspecies extending into Central America), and Campsisin eastern North America. Two dispersal events toAsia are inferred to have given rise to the Old Worldtaxa. One dispersal to Asia occurred in Campsis, inwhich C. grandiflora (Thunb.) K.Schum. is sister toC. radicans (L.) Seem. of eastern North America. Thesecond dispersal was most likely in the southernhemisphere and gave rise to the clade that includesthe Himalayan Incarvillea Juss. and a radiation ofabout five genera in SE Asia and Australia. Onelineage of this clade has reached Africa [includingTecomaria (Endl.) Spach, incorrectly included inTecoma by some authors, e.g. Gentry (1992a)].Another lineage is postulated to have dispersed backacross the South Pacific from Australia to Chile whereit is represented by the monotypic Campsidium.Campsidium represents the only hypothesized disper-sal back from the Old to the New World in Bignon-iaceae, Solanaceae or Verbenaceae.

Association of seed dispersal with long-distancedispersalDispersal in Bignoniaceae is typically facilitated bydry dehiscent fruits with winged seeds, which arereadily dispersed by wind, and which are found inmost members of the clade. However, three lineagesin the family have animal-dispersed, indehiscentfruits. None of these lineages originated in South

94 R. G. OLMSTEAD


America (Crescentieae in Central America, Coleeae inMadagascar and Kigelia DC. in Africa), and thus theyall represent lineages with a history of long-distancedispersal. However, seed dispersal of these taxa isbelieved to be facilitated by mammals (Janzen &Martin, 1982; Zjhra et al., 2004; Grose & Olmstead,2007a) and mammals are rarely thought to be vectorsof long-distance seed dispersal. The phylogeny ofBignoniaceae suggests that each case could representmore recent in situ evolution of animal-dispersedfruits, after the lineage to which each group belongswas established in the continent in which they occurtoday. Thus, all long-distance dispersal to the OldWorld is inferred to have been by plants with seedspresumed to be wind dispersed.

VERBENACEAE

South American origin and diversificationEvery clade of Verbenaceae recognized by Marxet al. (2010) originated in South America and hasits greatest species diversity in South Americatoday. The initial diversification is inferred to haveoccurred in the tropics, where Petreeae are part ofthe diverse Amazonian liana flora. The next threediverging clades, Duranteae, Casselieae and Cith-arexyleae, also occur primarily in tropical habitatsin wet tropical forests and Andean cloud forests.Citharexylum L. exhibits a pattern of diversificationfirst in lowland wet tropical forests, and sub-sequently radiating in Andean montane forests(C. argutedentatum Moldenke and C. ilicifoliumKunth in Fig. 4). This pattern appears to berepeated in Duranta L. (V. Thode & R. Olmstead,unpubl. data). However, Verbenaceae exhibit theirgreatest radiation in arid, temperate South America,presumably as the Andean uplift contributedto those environments in what is now Argentina(Blisniuk et al., 2004; Barreda & Palazzesi, 2007;Graham, 2010, 2011). Priveae, Neospartoneae, Ver-beneae and Lantaneae are all found today orinitially diversified primarily in such habitats.

Distribution in North America and the CaribbeanWhereas all of the clades of Verbenaceae identifiedhere originated in South America, only two areendemic to that continent, Neospartoneae, with threegenera and seven species, and Rhaphithamnus Miers,with two species (one in the Patagonian Andes andone endemic to the Juan Fernández islands). Theother seven clades all exhibit more or less continuousdistributions between South and North America. Asin Solanaceae and Bignoniaceae, most of these cladesreach similar latitudinal limits in the northern andsouthern hemispheres (Table 1; Fig. 5). Only two ofthese clades exhibit more than a 5° latitudinal differ-

ence between the northern and southern limits. Inthe case of Lantaneae, the 10° latitude greaterdistribution in the southern hemisphere can beaccounted for by one species, Acantholippia seriphio-ides (Gray) Moldenke, which is the only species in theclade to occur south of 40°S. In total, approximately65 species (c. 7.5% of New World diversity) occurfurther than c. 35° north or south of the equator.Glandularia, Verbena L. (Verbeneae) and Aloysia(Lantaneae) all exhibit amphitropical disjunctions ofseveral thousand kilometres between arid North andSouth America, a pattern observed in numerous unre-lated groups, suggesting long-distance dispersal(Lewis & Oliver, 1961; Solbrig, 1972; Yuan & Olm-stead, 2008a; P. Lu-Irving & R. G. Olmstead, unpub-lished), although there is evidence for Verbenasuggesting that a dispersal route along the Andes,where seasonal dry forests have existed for a longtime in a series of discontinuous patches (Pennington,Lavin & Oliveira-Filho, 2009), may be responsible(Marx et al., 2010; V. Thode & R. Olmstead, unpubl.data).

Niche conservatism vs. adaptation to novelecological zones in the diversificationof VerbenaceaeAs with Solanaceae and Bignoniaceae, diversificationin Verbenaceae represents a complex of ecologicalshifts between ecosystems and radiations within eco-systems. Verbenaceae most likely originated in wettropical forests, but following a successful shift to aridecosystems in temperate South America (possiblycoincident with the expansion of such habitats duringuplift of the Andes), diversified significantly inthose environments. Phylogenetic niche conservationapparently has dominated diversification followingthis shift to arid ecosystems. This can be seen in tworather different manifestations. In one, amphitropicaldisjunctions are found between arid zones of temper-ate and subtropical North and South America inVerbeneae and Lantaneae. The fact that these cladeshave hardly colonized geographically adjacent, moremesic habitats, yet have succeeded through dispersalin colonizing geographically distant, but ecologicallycomparable arid habitats in both hemispheres is con-sistent with Donoghue’s (2008) ‘easier to move than toevolve’ hypothesis.

In another instance, the Lantana/Lippia L. complexhas successfully colonized the relatively young, fire-adapted savannas of central Brazil and easternBolivia, where they have diversified extensively in thecerrado (Lu-Irving & Olmstead, 2012). In contrast,Stachytarpheta Vahl. has its roots in the wet tropicsduring the early diversification of Verbenaceae, buthas also diversified extensively following one or moreniche shifts into the Cerrado, where it co-occurs with



the Lippia/Lantana complex. The xeric vs. mesicdivide seems to present a strong adaptive barrier inVerbenaceae, with the repeated occurrence of phylo-genetically niche-conserved dry or wet-adapted line-ages across the family. For example, the centralAndes (northern Bolivia to central Ecuador) are hometo groups of Verbenaceae that apparently arose inarid regions of temperate South America (Verbeneae,Lantaneae), and which also occupy seasonally dryhabitats at mid to high elevations, and to groups thatarose in wet tropics (Duranta, Citharexylum), andwhich occur in wet Andean cloud forests today. Thishas been observed in many other Neotropical groupsthat are diverse in arid habitats (reviewed by Pen-nington et al., 2004, 2009; Lavin, 2006).

Transoceanic dispersalFour clades of Verbenaceae have representativesoutside the New World and six long-distance dispersalevents are required to account for them (Fig. 6). Mostof these Old World occurrences represent individualspecies or small clades within genera with primarilyNew World distributions and most likely representrecent dispersal events. Verbena officinalis L. occursin Eurasia, but is the only member of Verbena foundoutside the New World. It is phylogenetically nestedin a derived group of North American Verbena spp.and represents a relatively recent dispersal event toEurope (Marx et al., 2010). The other dispersal eventsall follow a pattern of South America to Africa;Lantana, Lippia and Priva Adans. all have a smallnumber of species in Africa. Chascanum E.Meyer isfound only in the Old World (Africa, Arabia, Indiansubcontinent), but is a close relative of South Ameri-can Bouchea Cham. and Stachytarpheta. Coelocar-pum Balf.f., a Madagascan genus of five species, issister to the rest of Lantaneae, and thus represents arelatively older dispersal event than those withinLantana and Lippia, although Lantaneae themselvesare a relatively recent group within Verbenaceae. Allsix inferred dispersal events appear to have beenfrom west to east.

Association of seed dispersal withlong-distance dispersalDry fruits are inferred to be ancestral in Verbenaceae(O’Leary et al., in press), but fleshy, putativelyanimal-dispersed fruits have arisen in several clades(Duranteae, Citharexyleae, Neospartoneae, Rhaphi-thamnus, Lantaneae), including multiple times in thespecies-rich Lantaneae, in which Lantana, tradition-ally defined by the presence of fleshy fruits, ispolyphyletic (Lu-Irving & Olmstead, 2012). Fleshyfruits are found in groups inhabiting both wet and dryhabitats and represent about 33% of all species ofVerbenaceae. Despite the relatively frequent occur-

rence of fleshy, animal-dispersed fruits in Verben-aceae, only one of the long-distance, over-waterdispersal events appears to coincide with a fleshy-fruited disperser (Lantana in Africa). Plants of Coe-locarpum in Madagascar also have fleshy fruits, butthe lineage is inferred to have arisen from dry-fruitedancestors (Lu-Irving & Olmstead, 2012), so it is notpossible to infer whether the dispersing colonizershad fleshy or dry fruits. Similarly, the lineages withamphitropical desert distributions are all dry-fruitedgroups (Glandularia, Verbena, Aloysia). Thus, to aneven greater extent than in Solanaceae, fleshy-fruitedlineages are under-represented relative to dry-fruitedones among long-distance dispersers.

CONCLUSIONS

Solanaceae, Bignoniaceae and Verbenaceae representthree ecologically and floristically important SouthAmerican plant families, the origin and diversifica-tion of which occurred primarily in situ. By breakingdown each of the three families into constituentclades, 37 in total, and looking at their geographicaldistributions in the context of the phylogeny of eachfamily, patterns emerge that contribute to a greaterunderstanding of the historical development of SouthAmerican floristic diversity.

The patterns observed suggest that a large majorityof clades in each family, regardless of age, have suc-ceeded in colonizing North America, which had nocontiguous land connection throughout most of thehistory of the South American continent since itssplit from Gondwana c. 100 Mya. Few apparent con-straints on migration/dispersal seem to exist forplants in these groups in the western hemisphere(Cody et al., 2010). Clade limits within the New Worldseem to be primarily ecological, rather than strictlygeographical, suggesting that adaptive barriers thatlimited geographical spread and constrain mesic/xerictransitions have been important factors shaping thediversification of these three families, in line withrecent finding for other groups (Lavin et al., 2004;Donoghue, 2008). Inability to evolve cold toleranceseems to restrict most clades to within a range of c.30–35°N/S latitude (Fig. 5), roughly coinciding withthe 10 °C average temperature for the coldest month,a placeholder for the frost-free zone on most conti-nents (Thompson et al., 2000). Only an estimated 5, 1and 7.5% of New World Solanaceae, Bignoniaceae andVerbenaceae, respectively, occur beyond c. 35°N/S,thus fitting the tropical niche conservation model(Wiens & Donoghue, 2004; Donoghue, 2008). Thoseclades that have succeeded in reaching high latitudestend to be species-rich (although most of the diversityis still tropical) and also to have colonized the OldWorld much more often than the clades restricted to

96 R. G. OLMSTEAD


lower latitudes. Bromeliaceae provide an interestingcontrast. They are even more highly constrained tofrost-free latitudes, where they are largely restrictedto mesic ecosystems, with rare shifts to arid habitats;following such shifts, the resulting clades remainconstrained to arid ecosystems (Givnish et al., 2011).

Solanaceae and Verbenaceae also include notablearid temperate radiations (e.g. Lycium in Solanaceae;Aloysia, Glandularia and Verbena in Verbenaceae)that occupy distantly disjunct distributions betweendeserts of North and South America. Amphitropicaldisjuncts of this kind are relatively common betweenarid habitats of North and South America in coastal,Mediterranean ecosystems in California/Chile and ininterior arid habitats in the south-west USA andnorthern Mexico/central and northern Argentina(Raven, 1972; Wen & Ickert-Bond, 2009). Coastal,Mediterranean climate disjuncts appear to be pre-dominantly the result of north to south migrationevents (Wen & Ickert-Bond, 2009), whereas the inte-rior desert disjuncts exhibit a mix of both south tonorth events (e.g. Larrea Cav. – Lia et al., 2001;Hoffmannseggia Cav. – Simpson, Tate & Weeks,2005), and north to south events (e.g. Tiquilia Pers. –Moore, Tye & Jansen, 2006; Astragalus L. – Scherson,Vidal & Sanderson, 2008). The four lineages inSolanaceae and Verbenaceae in which this pattern isfound all represent South America to North Americadispersals (Yuan & Olmstead, 2008a; J. Miller et al.,2011; P. Lu-Irving & R. Olmstead, unpubl. data).

Again, these patterns provide good evidence forlarge-scale phylogenetic niche or biome conservatismsensu Donoghue (2008). Despite the prevalence ofphylogenetic niche conservatism, major habitat shiftsare also apparent within each family both at theorigin of clades and within clades, resulting inpresent distributions for each family that include allmajor biomes in South America. Indeed, niche shiftswithin clades may account for some impressive recentradiations, e.g. the diversification of Stachytarpheta(Duranteae: Verbenaceae) and Bignonieae (Bignon-iaceae) in the cerrado vegetation of Brazil. Althoughevolution of the Cerrado has been shown to haveinvolved recruitment of lineages from wet tropicalforest, dry tropical forest and subtropical pampaancestors (Simon et al., 2009), the diversification ofthe Lantana/Lippia complex in the cerrado representsa diversification derived from arid-adapted ancestorsin temperate South America (Lu-Irving & Olmstead,2012). Solanum seems to have adapted to virtually allhabitable Neotropical ecosystems, even though itsuffers the same limited ability to colonize cold tem-perate regions as observed in all three families. Thefact that habitat outliers within clades are often theexception suggests that niche conservation prevailswithin most clades in all three families, but niche

evolution is important in the history of each family, asit is likely to be in any large clade of ancient origin(e.g. Fabaceae – Schrire, Lavin & Lewis, 2005) andeven limited niche shifting can lead, with time, tosuccessful establishment in many biomes.

Despite the near universal success of multiple inde-pendent clades within each family to colonize NorthAmerica, in many cases most likely preceding theclosing of the Isthmus of Panama, long-distance dis-persal to colonize Old World continents apparentlyhas been much more limited, and diversification fol-lowing those events typically has been modest (theseare, by definition, relatively recent events). The spinysolanums (Solanum section Leptostemonum) havebeen the most successful diversification outside theNew World in any of these three families, with c. 175species in the Old World. Elsewhere in Solanaceae,Hyoscyameae have about 40 species in Eurasia,Lycium has about 35 Old World species distributed inAfrica, Asia and Australia, and Anthocercideae andNicotiana have about 30 and 16 species in Australia,respectively. In Bignoniaceae, the Palaeotropicalclade has also diversified with c. 150 species, andTecomeae in Australasia and the Himalaya with c. 40species. In Verbenaceae, Chascanum (Duranteae)includes about 25 species in Africa, Arabia and theIndian subcontinent, whereas Lantana and Lippia(Lantaneae) each have c. 15 species in Africa result-ing from independent dispersal events (Lu-Irving &Olmstead, 2012).

The patterns exhibited by these three families, ofprimary radiations within South America, and limitedcolonization of the Old World continents by cladesthat have not themselves radiated extensively, havesimilarities and one stark contrast with Asteraceae,another large clade that is thought to have originatedand initially diversified in South America. In similarways to the families described here, many of the earlydiverging lineages of Asteraceae also expanded tocolonize North America (Funk et al., 2005). However,the vast majority of Asteraceae are derived primarilyfrom a transatlantic colonization event that led to anexplosive radiation in Africa, with subsequent spreadand radiations on other continents, especially NorthAmerica, from which many lineages eventually rec-olonized South America (Funk et al., 2005). Aster-aceae were tremendously successful at colonizingtemperate biomes, unlike any of the three familiesconsidered here. Another large family, Bromeliaceae,in contrast, exhibited relatively limited colonization ofNorth America and only a single transatlantic disper-sal event (Givnish et al., 2011).

It is not possible with the present level of taxonsampling in the phylogenetic analyses for these threefamilies to infer all instances of dispersal from South toNorth America or all biome shifts and, without time-



calibrated trees, it is impossible to rigorously testhypotheses of dispersal routes and timing (e.g. over-land through the Isthmus of Panama vs. an over-waterdispersal route). Work is planned or in progress ondating phylogenies of Solanaceae (S. Knapp, pers.comm.) and Bignoniaceae (L. Lohmann, pers. comm.)and is also anticipated in Verbenaceae. Dated phylog-enies for these clades will permit more rigorous testingof explicit biogeographical hypotheses of the sortdescribed here on the basis of undated phylogenies andspecies distribution data. Enhanced phylogenies usingsupermatrix and supertree approaches to increasetaxon sampling will also permit much finer-scale eco-logical assessment of niche conservation vs. ecologicalshifts in the history of diversification of these threeiconic South American plant families.

ACKNOWLEDGEMENTS

Thanks to the many students and collaborators (inparticular L. Bohs, L. Lohmann, S. Grose, N. O’Leary,M. Múlgura) who participated in the phylogeneticstudies of Bignoniaceae, Solanaceae and Verbenaceae,and to the many herbaria, Botanical Gardens andindividuals who provided plant material for thosestudies. Thanks to A. Antonelli and T. Pennington forinviting me to speak in the symposium at IBC 2011,to C. Stromberg and A. Graham for responses toqueries, and to P. Lu-Irving, L. Lohmann, L. Bohs,Colin Hughes, John Klicka and two anonymousreviewers for discussions on these issues and/orcomments on the manuscript. Funding for thesestudies came from numerous sources, including NSFgrants DEB0309065, DEB0309065, DEB0542493,DEB0710026 and EF-0431184 to R.G.O.

REFERENCES

Andersson L, Antonelli A. 2005. Phylogeny of the tribeCinchoneae (Rubiaceae), its position in Cinchonioideae, anddescription of a new genus, Ciliosemina. Taxon 54: 17–28.

Andersson S. 2006. On the phylogeny of the genus Calceo-laria (Calceolariaceae) as inferred from ITS and plastidmatK sequences. Taxon 55: 125–137.

Antonelli A, Nylander JAA, Persson C, Sanmartin I.2009. Tracing the impact of the Andean uplift on Neotropi-cal plant evolution. Proceedings of the National Academy ofSciences of the United States of America 106: 9749–9754.

Antonelli A, Sanmartin I. 2011. Why are there so manyplant species in the Neotropics? Taxon 60: 403–414.

Atkins S. 2004. Verbenaceae. In: Kadereit JW, ed. The fami-lies and genera of flowering plants, Vol. 7. Berlin: Springer,449–468.

Baldwin BG, Crawford DJ, Francisco-Ortega J, KimS-C, Sang T, Stuessy TF. 1998. Molecular phylogenetic

insights on the origin and evolution of oceanic island plants.In: Soltis PS, Soltis DE, Doyle JJ, eds. Molecular systemat-ics of plants II: DNA sequencing. Boston: Kluwer, 410–441.

Baldwin BG, Wagner WL. 2010. Hawaiian angiospermradiations of North American origin. Annals of Botany 105:849–879.

Barreda V, Palazzesi L. 2007. Patagonian vegetationturnovers during the Paleogene–Early Neogene: origin ofarid-adapted floras. The Botanical Review 73: 31–50.

Bell CD, Donoghue MJ. 2005. Phylogeny and biogeographyof Valerianaceae (Dipsacales) with special reference tothe South American valerians. Organisms, Diversity andEvolution 5: 147–159.

Blisniuk PM, Stern LA, Chamberlain CP, Idleman B,Zeitler PK. 2004. Climatic and ecologic changes duringMiocene surface uplift in the southern Patagonian Andes.Earth and Planetary Science Letters 230: 125–142.

Bohs L. 2005. Major clades in Solanum based on ndhFsequence data. In: Keating RC, Hollowell VC, Croat TB, eds.A festchrift for William G. D’Arcy: the legacy of a taxonomist,Vol. 104. Monographs in systematic botany. St. Louis: Mis-souri Botanical Garden Press, 27–49.

Bohs L, Olmstead RG. 2001. A reassessment of Normaniaand Triguera (Solanaceae). Plant Systematics and Evolution228: 33–48.

Bremer B, Bremer K, Heidari N, Erixon P, OlmsteadRG, Anderberg AA, Kallersjo M, Barkhordarian E.2002. Phylogenetics of asterids based on 3 coding and 3non-coding chloroplast DNA markers and the utilityof non-coding DNA at higher taxonomic levels. MolecularPhylogenetics and Evolution 24: 274–301.

Bremer K, Friis EM, Bremer B. 2004. Molecular phyloge-netic dating of asterid flowering plants shows early Creta-ceous diversification. Systematic Biology 53: 496–505.

Burnham RJ, Graham A. 1999. The history of Neotropicalvegetation: new developments and status. Annals of theMissouri Botanical Garden 86: 546–589.

Carlquist S. 1967. The biota of long-distance dispersal. V.Plant dispersal to Pacific Islands. Bulletin of the TorreyBotanical Club 94: 129–162.

Carlquist S. 1974. Island biology. New York: ColumbiaUniversity Press.

Christenhusz MJM, Chase MW. 2012. Biogeographical pat-terns of plants in the Neotropics – dispersal rather than platetectonics is most explanatory. Botanical Journal of theLinnean Society 171: doi: 10.1111/j.1095-8339.2012.01301.x

Clarkson JJ, Knapp S, Aoki S, Garcia VF, Olmstead RG,Chase MW. 2004. Phylogenetic relationships in Nicotiana(Solanaceae) inferred from multiple plastid DNA regions.Molecular Phylogenetics and Evolution 33: 75–90.

Coates AG, Obando JA. 1996. The geologic evolution of theCentral American isthmus. In: Jackson JBC, Budd AF,Coates AG, eds. Evolution and environment in tropicalAmerica. Chicago: University of Chicago Press, 21–56.

Cody S, Richardson JE, Rull V, Ellis C, Pennington RT.2010. The great American biotic interchange revisited.Ecography 33: 326–332.

98 R. G. OLMSTEAD


Crisp MD, Arroyo MTK, Cook LG, Gandolfo MA, JordanGJ, McGlone MS, Weston PH, Westoby M, Wilf P,Linder HP. 2009. Phylogenetic biome conservatism on aglobal scale. Nature 458: 754–756.

Davis CC, Bell CD, Mathews S, Donoghue MJ. 2002.Laurasian migration explains Gondwanan disjunctions: evi-dence from Malpighiaceae. Proceedings of the NationalAcademy of Sciences of the United States of America 99:6833–6837.

Dillon MO, Tu T, Quipuscoa Silvestre V, Wen J. 2009.Biogeographic diversification in Nolana (Solanaceae), aubiquitous member of the Atacama and Peruvian Desertsalong the western coast of South America. Journal ofSystematics and Evolution 47: 457–476.

Dillon MO, Tu T, Soejima A, Yi T, Nie Z, Tye A, Wen J.2007. Phylogeny of Nolana (Nolaneae, Solanoideae,Solanaceae) as inferred from granule-bound starch synthaseI (GBSSI) sequences. Taxon 56: 1000–1011.

Donoghue MJ. 2008. A phylogenetic perspective on the dis-tribution of plant diversity. Proceedings of the NationalAcademy of Sciences of the United States of America 105:11549–11555.

Edwards EJ, Nyffeler R, Donoghue MJ. 2005. Basalcactus phylogeny: implications of Pereskia (Cactaceae) para-phyly for the transition to the cactus life form. AmericanJournal of Botany 92: 1177–1188.

Erkens RHJ, Maas JW, Couvreur TLP. 2009. From Africavia Europe to South America: migrational route of a species-rich genus of Neotropical lowland rain forest trees (Guatte-ria, Annonaceae). Journal of Biogeography 36: 2338–2352.

Farris DW, Jaramillo CA, Bayona GA, Restrepo-MorenoSA, Montes C, Cardona A, Mora A, Speakman RJ,Glasscock MD, Reiners P, Valencia V. 2011. Fracturingof the Panamanian isthmus during initial collision withSouth America. Geology 39: 1007–1010.

Fischer E, Theisen I, Lohmann LG. 2004. Bignoniaceae.In: Kadereit JW, ed. The families and genera of floweringplants, Vol. 7. Berlin: Springer, 9–38.

Funk VA, Bayer RJ, Keeley S, Chan R, Watson L,Gemeinholzer B, Schilling E, Panero JL, Baldwin BG,Garcia-Jacas N, Susanna A, Jansen RK. 2005. Every-where by Antarctica: using a supertree to understand thediversity and distribution of the Compositae. BiologiskeSkrifter 55: 343–373.

Gentry AH. 1980. Bignoniaceae part I – Tribes Crescentieaeand Tourretieae. Flora Neotropica Monograph 25: 1–131.

Gentry AH. 1982. Neotropical floristic diversity: phytogeo-graphical connections between Central and South America,Pleistocene climatic fluctuations, or an accident of theAndean orogeny? Annals of the Missouri Botanical Garden69: 557–593.

Gentry AH. 1992a. Bignoniaceae – part II (Tribe Tecomeae).Flora Neotropica Monograph 25: 1–370.

Gentry AH. 1992b. A synopsis of Bignoniaceae ethnobotanyand economic botany. Annals of the Missouri BotanicalGarden 79: 53–64.

Gillespie RG, Baldwin BG, Waters JM, Fraser CI,Nikula R, Roderick GK. 2012. Long-distance dispersal: a

framework for hypothesis testing. Trends in Ecology andEvolution 27: 47–56.

Givnish TJ, Barfuss MHJ, Van Ee B, Riina R, Schulte K,Horres R, Gonsiska PA, Jabaily RS, Crayn DM, SmithAC, Winter K, Brown GK, Evans TM, Holst BK, LutherH, Till W, Zizka G, Berry PE, Sytsma KJ. 2011. Phyl-ogeny, adaptive radiation, and historical biogeography inBromeliaceae: insights from an eight-locus plastid phylog-eny. American Journal of Botany 98: 872–895.

Givnish TJ, Millam KC, Evans TM, Hall JC, Pires JC,Berry PE, Sytsma KJ. 2004. Ancient vicariance orrecent long-distance dispersal? Inferences about phylogenyand South American-African disjunctions in Rapateaceaeand Bromeliaceae based on ndhF sequence data.International Journal of Plant Sciences 165 (Suppl): S35–S54.

Goldblatt P. 1993. Biological relationships between Africaand South America. New Haven: Yale University Press.

Graham A. 2010. Late Cretaceous and Cenozoic history ofLatin American vegetation and terrestrial environments. St.Louis: Missouri Botanic Garden Press.

Graham A. 2011. The age and diversification of terrestrialNew World ecosystems through Cretaceous and Cenezoictime. American Journal of Botany 98: 336–351.

Grose SO, Olmstead RG. 2007a. Evolution of a charismaticNeotropical clade: molecular phylogeny of Tabebuia s.l.,Crescentieae, and allied genera (Bignoniaceae). SystematicBotany 32: 650–659.

Grose SO, Olmstead RG. 2007b. Taxonomic revisions in thepolyphyletic genus Tabebuia s.l. (Bignoniaceae). SystematicBotany 32: 660–670.

Harvey PH, Pagel MD. 1991. The comparative method inevolutionary biology. New York: Oxford University Press.

Heiser CB. 1969. Nightshades, the paradoxical plants. SanFrancisco: Freeman and Company.

Hughes C, Eastwood R. 2006. Island radiation on acontinental scale: exceptional rates of plant diversificationafter uplift of the Andes. Proceedings of the NationalAcademy of Sciences of the United States of America 103:10334–10339.

Hunziker AT. 2001. Genera solanacearum: the genera ofSolanaceae illustrated, arranged according to a new system.Ruggell: A.R.G. Gantner Verlag.

Janzen DH, Martin PS. 1982. Neotropical anachronisms –the fruits the gomphotheres ate. Science 215: 19–27.

Keigwin L. 1982. Isotopic paleoceanography of the Caribbeanand East Pacific: role of Panama uplift in Late Neogenetime. Science 217: 350–353.

Lavin M. 2006. Floristic and geographical stability of discon-tinuous seasonally dry tropical forests explains patterns ofplant phylogeny and endemism. In: Pennington RT, RatterJA, Lewis GP, eds. Neotropical savannas and seasonally dryforests: plant biodiversity, biogeography and conservation.Boca Raton: CRC Press, 433–437.

Lavin M, Herendeen PS, Wojciechowski MF. 2005. Evo-lutionary rates analysis of Leguminosae implicates a rapiddiversification of lineages during the Tertiary. SystematicBiology 54: 575–594.



Lavin M, Schrire B, Lewis G, Pennington T, Delgado-Salinas A, Thulin M, Hughes C, Beyra Matos A,Wojciechoswki MF. 2004. Metacommunity process ratherthan continental tectonic history better explains geographi-cally structured phylogenies in legumes. PhilosophicalTransactions of the Royal Society London (Biological Sci-ences) 359: 1509–1522.

Lavin M, Wojciechowski MF, Gasson P, Hughes C,Wheeler E. 2003. Phylogeny of robinioid legumes(Fabaceae) revisited: Coursetia and Gliricidia recircum-scribed, and a biogeographical appraisal of the Caribbeanendemics. Systematic Botany 28: 387–409.

Levin RA, Myers NR, Bohs L. 2006. Phylogenetic relation-ships among the ‘spiny solanums’ (Solanum subgenus Lep-tostemonum, Solanaceae). American Journal of Botany 93:157–169.

Levin SA, Muller-Landau HC, Nathan R, Chave J. 2003.The ecology and evolution of seed dispersal: a theoreticalperspective. Annual Review of Ecology, Evolution, and Sys-tematics 34: 575–604.

Lewis WH, Oliver RL. 1961. Cytogeography and phylogenyof the North American species of Verbena. American Journalof Botany 48: 638–643.

Li J. 2008. Phylogeny of Catalpa (Bignoniaceae) inferred fromsequences of chloroplast ndhF and nuclear ribosomal DNA.Journal of Systematics and Evolution 46: 341–348.

Lia VV, Confalonieri VA, Comas CO, Hunziker JH. 2001.Molecular phylogeny of Larrea and its allies (Zygophyl-laceae): reticulate evolution and the probable time of creo-sote bush arrival to North America. Molecular Phylogeneticsand Evolution 21: 309–320.

Lohmann LG. 2006. Untangling the phylogeny of Neotropi-cal lianas (Bignonieae, Bignoniaceae). American Journal ofBotany 93: 304–318.

Lohmann LG. In press. A new generic classification ofBignonieae (Bignoniaceae) based on molecular phylogeneticdata and morphological synapomorphies. Annals of the Mis-souri Botanical Garden.

Lohmann LG, Ulloa CU. 2007. Bignoniaceae in iPlantsprototype checklist. [online]. Available at: http://www.iplants.org [accessed 15 December 2011].

Lohmann LG, Bell CD, Calió MF, Winkworth RC. 2012.Pattern and timing of biogeographic history in the Neotropi-cal tribe Bignonieae (Bignoniaceae). Botanical Journal ofthe Linnean Society 171: doi: 10.1111/j.1095-8339.2012.01311.x.

Lu-Irving P, Olmstead RG. 2012. Investigating theevolution of Lantaneae (Verbenaceae) using multiple loci.Botanical Journal of the Linnean Society 171: doi: 10.1111/j.1095-8339.2012.01305.x.

Magallón S, Castillo A. 2009. Angiosperm diversifica-tion through time. American Journal of Botany 96: 349–365.

Marshall LG, Webb DD, Sepkoski JJ, Raup DM. 1982.Mammalian evolution and the Great American Interchange.Science 215: 1351–1357.

Marx HE, O’Leary N, Yuan Y-W, Lu-Irving P, Tank DC,Múlgura M, Olmstead RG. 2010. A molecular phylogeny

and classification of Verbenaceae. American Journal ofBotany 97: 1647–1663.

McDade LA, Daniel TF, Kiel C. 2008. Toward a compre-hensive understanding of phylogenetic relationships amonglineages of Acanthaceae s.l. (Lamiales). American Journal ofBotany 95: 1136–1152.

McDade LA, Daniel TF, Kiel CA, Borg AJ. 2012.Phylogenetic placement, delineation, and relationshipsamong genera of the enigmatic Nelsonioideae (Lamiales:Acanthaceae). Taxon 61: 637–651.

Miller JS, Kamath A, Damashek J, Levin RA. 2011. Outof America to Africa or Asia: inference of dispersal historiesusing nuclear and plastid DNA and the S-RNase self-incompatibility locus. Molecular Biology and Evolution 28:793–801.

Miller JS, Kamath A, Levin RA. 2009. Do multiple tortoisesequal a hare? The utility of nine noncoding plastid regionsfor species-level phylogenetics in tribe Lycieae (Solanaceae).Systematic Botany 34: 796–804.

Miller RJ, Mione T, Phan H-L, Olmstead RG. 2011. Colorby numbers: nuclear waxy gene phylogeny of Jaltomata(Solanaceae), the sister genus to Solanum, supports threeclades differing in fruit color. Systematic Botany 36: 153–162.

Montes C, Cardona A, McFadden R, Morón SE, Silva CA,Restrepo-Moreno RDA, Hoyos N, Wilson J, Farris D,Bayona GA, Jaramillo CA, Valencia V, Bryan J, FloresJA. 2012. Evidence for middle Eocene and younger landemergence in central Panama: implications for Isthmusclosure. Geological Society of America Bulletin 124: 780–799.

Moore MJ, Tye A, Jansen RK. 2006. Patterns of long-distance dispersal in Tiquilia subg. Tiquilia (Boraginaceae):implications for the origins of amphitropical disjuncts andGalápagos islands endemics. American Journal of Botany93: 1163–1177.

Morley R. 2003. Interplate dispersal paths for megathermalangiosperms. Perspectives in Plant Ecology, Evolution andSystematics 6: 5–20.

Nathan R, Schurr FM, Spiegel O, Steinitz O, Trakhten-brot A, Tsoar A. 2008. Mechanisms of long-distance seeddispersal. Trends in Ecology and Evolution 23: 638–647.

Nie Z-L, Sun H, Beardsley PM, Olmstead RG, Wen J.2006. Evolution of biogeographic disjunction betweeneastern Asia and eastern North America in Phryma (Phry-maceae). American Journal of Botany 93: 1343–1356.

O’Connor JM, Duncan RA. 1990. Evolution of the WalvisRidge-Rio Grande Rise Hot Spot System: implications forAfrican and South American plate motions over plumes.Journal of Geophysical Research 95: 17475–17502.

O’Leary N, Calviño CI, Martinez S, Olmstead RG,Múlgura ME. In press. Evolution of morphological traitsin Verbenaceae. American Journal of Botany.

O’Leary N, Yuan Y-W, Chemisquy A, Olmstead RG. 2009.Reassignment of species of paraphyletic Junellia s.l. tothe new genus Mulguraea (Verbenaceae) and new circum-scription of genus Junellia: molecular and morphologicalcongruence. Systematic Botany 34: 777–786.

100 R. G. OLMSTEAD


Olmstead RG, Bohs L. 2007. A summary of molecular sys-tematic research in Solanaceae: 1982–2006. In: SpoonerDM, Bohs L, Govannoni J, Olmstead RG, Shibata D, eds.Solanaceae VI: genomics meets biodiversity. Proceedings ofthe Sixth International Solanaceae Conference. Acta Horti-culturae 745, Leuven: International Society for Horticul-tural Science, 255–268.

Olmstead RG, Bohs L, Abdel Migid H, Santiago-ValentinE, Garcia VF, Collier SM. 2008. A molecular phylogeny ofthe Solanaceae. Taxon 57: 1159–1181.

Olmstead RG, Palmer JD. 1992. A chloroplast DNA phyl-ogeny of the Solanaceae: subfamilial relationships and char-acter evolution. Annals of the Missouri Botanical Garden79: 346–360.

Olmstead RG, dePamphilis CW, Wolfe AD, Young ND,Elisens WJ, Reeves PA. 2001. Disintegration of the Scro-phulariaceae. American Journal of Botany 88: 348–361.

Olmstead RG, Sweere JA. 1994. Combining data in phylo-genetic systematics: an empirical approach using threemolecular data sets in the Solanaceae. Systematic Biology43: 467–481.

Olmstead RG, Sweere JA, Spangler RE, Bohs L, PalmerJD. 1999. Phylogeny and provisional classification of theSolanaceae based on chloroplast DNA. In: Nee M, SymonDE, Jessup JP, Hawkes JG, eds. Solanaceae IV, Advances inbiology and utilization. Kew: Royal Botanic Gardens, 111–137.

Olmstead RG, Zjhra ML, Lohmann LG, Grose SO, EckertAJ. 2009. A molecular phylogeny and classification of Bigno-niaceae. American Journal of Botany 96: 1731–1743.

Paape T, Igic B, Smith SD, Olmstead RG, Bohs L, KohnJR. 2008. A 15-Myr-old genetic bottleneck. MolecularBiology and Evolution 25: 655–663.

Pennington RT, Dick CW. 2004. The role of immigrants inthe assembly of the South American rainforest tree flora.Philosophical Transactions of the Royal Society London(Biological Sciences) 359: 1611–1622.

Pennington RT, Lavin M, Oliveira-Filho A. 2009. Woodyplant diversity, evolution, and ecology in the tropics: per-spectives from seasonally dry tropical forests. AnnualReview of Ecology, Evolution, and Systematics 40: 437–457.

Pennington RT, Lavin M, Prado DE, Pendry CA, PellSK, Butterworth C. 2004. Historical climate change andspeciation: neotropical seasonally dry forest plants showpatterns of both Tertiary and Quaternary diversification.Philosophical Transactions of the Royal Society London(Biological Sciences) 359: 515–537.

Pennington RT, Lavin M, Särkinen T, Lewis GP, Klit-gaard BB, Hughes CE. 2010. Contrasting plant diversifi-cation histories within the Andean biodiversity hotspot.Proceedings of the National Academy of Sciences of theUnited States of America 107: 13783–13787.

Perret M, Chautems A, Araujo AO, Salamin N. 2012.Temporal and spatial origin of the Gesneriaceae in the NewWorld inferred from plastid DNA sequences. BotanicalJournal of the Linnean Society 171: doi: 10.1111/j.1095-8339.2012.01303.x

Pirie MD, Chatrou LW, Mols JB, Erkens RHJ, OosterhofJ. 2006. ‘Andean-centered’ genera in the short-branch cladeof Annonaceae: testing biogeographical hypotheses usingphylogeny reconstruction and molecular dating. Journal ofBiogeography 33: 31–46.

Poczai P, Hyvönen J, Symon DE. 2011. Phylogeny ofkangaroo apples (Solanum subg. Archaesolanum,Solanaceae). Molecular Biology Reports 38: 5243–5259.

Raven P. 1972. Plant species disjunctions: a summary.Annals of the Missouri Botanical Garden 59: 234–246.

Renner SS. 2004. Plant dispersal across the tropical Atlanticby wind and sea currents. International Journal of PlantSciences 165 (Suppl): S23–S33.

Renner SS, Clausing G, Meyer K. 2001. Historical bioge-ography of Melastomataceae: the roles of Tertiary migrationand long-distance dispersal. American Journal of Botany 88:1290–1300.

Richardson JE, Chatrou LW, Mols JB, Erkens RHJ,Pirie MD. 2004. Historical biogeography of two cosmopoli-tan families of flowering plants: Annonaceae and Rham-naceae. Philosophical Transactions of the Royal SocietyLondon (Biological Sciences) 359: 1495–1508.

Richardson JE, Pennington RT, Pennington TD, Holl-ingsworth PM. 2001. Rapid diversification of a species-richgenus of Neotropical rain forest trees. Science 293: 2242–2245.

Santiago-Valentin E, Olmstead RG. 2003. Phylogenetics ofthe Antillean Goetzeoideae (Solanaceae) and their relation-ships within the Solanaceae based on chloroplast and ITSDNA sequence data. Systematic Botany 28: 452–460.

Särkinen TE, Newman MF, Maas PJM, Maas H, PoulsenAD, Harris DJ, Richardson JE, Clark A, Hollings-worth M, Pennington RT. 2007. Recent oceanic long-distance dispersal and divergence in the amphi-Atlanticrain forest genus Renealmia L.f. (Zingiberaceae). MolecularPhylogenetics and Evolution 44: 968–980.

Särkinen TE, Pennington RT, Lavin M, Simon MF,Hughes CE. 2012. Evolutionary islands in the Andes: per-sistence and isolation explain high endemism in Andeandry tropical forests. Journal of Biogeography 39: 884–990.

Schäferhoff B, Fleishmann A, Fischer E, Albach DC,Borsch T, Heubl G, Müller KF. 2010. Towards resolvingLamiales relationships: insights from rapidly evolving chlo-roplast sequences. BMC Evolutionary Biology 10: 352. doi:10.1186/1471.2148-10-352.

Scherson RA, Vidal R, Sanderson MJ. 2008. Phylogeny,biogeography, and rates of diversification of New WorldAstragalus (Leguminosae) with an emphasis on SouthAmerican radiations. American Journal of Botany 95: 1030–1039.

Schrire BD, Lavin M, Lewis GP. 2005. Global distributionpatterns of the Leguminosae: insights from recent phylog-enies. Biologiske Skrifter 55: 375–422.

Simon MF, Grether G, de Queiroz LP, Skema C, Pen-nington RT, Hughes CE. 2009. Recent assembly of thecerrado, a Neotropical plant diversity hotspot, by in situevolution of adaptations to fire. Proceedings of the National



Academy of Sciences of the United States of America 106:20359–20364.

Simpson BB, Tate JA, Weeks A. 2005. The biogeography ofHoffmannseggia (Leguminosae, Caesalpinoideae, Caes-alpinieae): a tale of many travels. Journal of Biogeography32: 15–27.

Simpson GG. 1980. Splendid isolation: the curious history ofSouth American mammals. New Haven: Yale UniversityPress.

Smith BT, Klicka J. 2010. The profound influence of theLate-Pliocene Panamanian uplift on the exchange, diversi-fication, and distribution of New World Birds. Ecography 33:333–342.

Smith JF, Wolfram JC, Brown KD, Carroll CL, DentonDS. 1997. Tribal relationships in the Gesneriaceae:evidence from DNA sequences of the chloroplast gene ndhF.Annals of the Missouri Botanical Garden 84: 50–66.

Solbrig OT. 1972. The floristic disjunctions between the‘Monte’ in Argentina and the ‘Sonora Desert’ in Mexico andthe United States. Annals of the Missouri Botanical Garden59: 218–223.

Soltis DE, Smith SA, Cellinese N, Wurdack KJ, Tank DC,Brockington SF, Refulio-Rodriguez NF, Walker JB,Moore MJ, Carlsward BS, Bell CD, Latvis M, CrawleyS, Black C, Diouf D, Xi Z, Gitzendanner MA, SystmaKJ, Qiu Y-L, Hilu KW, Manchester SR, Davis CC,Sanderson MJ, Olmstead RG, Judd WS, DonoghueMJ, Soltis PS. 2011. Angiosperm phylogeny: 17-genes, 640taxa. American Journal of Botany 98: 704–730.

Soza VL, Olmstead RG. 2010a. Molecular systematics oftribe Rubieae (Rubiaceae): evolution of major clades, devel-opment of leaf-like whorls, and biogeography. Taxon 59:755–771.

Soza VL, Olmstead RG. 2010b. Evolution of breedingsystems and fruits in New World Galium and relatives(Rubiaceae). American Journal of Botany 97: 1630–1646.

Spangler RE, Olmstead RG. 1999. Phylogenetic analysis ofthe Bignoniaceae based on the cpDNA gene sequences rbcLand ndhF. Annals of the Missouri Botanical Garden 86:33–46.

Stefanovic S, Krueger LE, Olmstead RG. 2002.Monophyly of the Convolvulaceae and circumscriptionof their major lineages based on DNA sequences of multiplechloroplast loci. American Journal of Botany 89: 1510–1522.

Stern S, Agra M, Bohs L. 2011. Molecular delimitation ofclades within New World species of the ‘spiny solanums’(Solanum subg. Leptostemonum). Taxon 60: 1429–14441.

Symon DE. 1994. Kangaroo apples. Adelaide: D. Symon.Thompson RS, Anderson KH, Bartlein PJ. 2000. Atlas of

relations between climatic parameters and distributions of

important trees and shrubs in North America. Washington:United States Geological Survey Professional Paper 1650-A.

Wallander E, Albert VA. 2000. Phylogeny and classificationof Oleaceae based on rps16 and trnL-F sequence data.American Journal of Botany 87: 1827–1841.

Weber A. 2004. Gesneriaceae. In: Kadereit JW, ed. The fami-lies and genera of flowering plants. Berlin: Springer, 7:63–158.

Weese TL, Bohs L. 2007. A three gene phylogeny ofthe genus Solanum (Solanaceae). Systematic Botany 32:445–463.

Wen J, Ickert-Bond SM. 2009. Evolution of the Madrean–Tethyan disjunctions and the North and South Americanamphitropical disjunctions in plants. Journal of Systematicsand Evolution 47: 331–348.

Whitson M, Manos PS. 2005. Untangling Physalis(Solanaceae) from the physaloids: a two-gene phylogeny ofthe Physalinae. Systematic Botany 30: 216–230.

Wiens JJ, Donoghue MJ. 2004. Historical biogeography,ecology and species richness. Trends in Ecology and Evolu-tion 19: 639–644.

Wikström N, Savolainen V, Chase MW. 2001. Evolution ofthe angiosperms: calibrating the family tree. Proceedings ofthe Royal Society of London. Series B, Biological Sciences268: 2211–2220.

Wortley AH, Harris GJ, Scotland RW. 2007. On the tax-onomy and phylogenetic position of Thomandersia. System-atic Botany 32: 415–444.

Yuan Y-W, Liu C, Marx HE, Olmstead RG. 2010. Anempirical demonstration of using PPR (pentatricopeptiderepeat) genes as phylogenetic tools: phylogeny of Verben-aceae and the Verbena complex. Molecular Phylogeneticsand Evolution 54: 23–35.

Yuan Y-W, Olmstead RG. 2008a. A species-level phyloge-netic study of the Verbena complex (Verbenaceae) indicatestwo independent intergeneric chloroplast transfers. Molecu-lar Phylogenetics and Evolution 48: 23–33.

Yuan Y-W, Olmstead RG. 2008b. Evolution and phylogeneticutility of the PHOT gene duplicates in a recently diversifiedgroup: dramatic intron size variation and footprint of ances-tral recombination. American Journal of Botany 95: 1166–1176.

Zimmer EA, Roalson EH, Skog LE, Boggan JK, IdnurmA. 2002. Phylogenetic relationships in the Gesnerioideae(Gesneriaceae) based on nrDNA ITS and cpDNA trnL-Fand trnE-T spacer region sequences. American Journal ofBotany 89: 296–311.

Zjhra ML, Sytsma KJ, Olmstead RG. 2004. Delimitation oftribe Coleeae and implications for fruit evolution in Bigno-niaceae inferred from a chloroplast DNA phylogeny. PlantSystematics and Evolution 245: 55–67.

102 R. G. OLMSTEAD


review article phylogeny and biogeography in solanaceae

Documents