Downloaded from rstb.royalsocietypublishing.org on May 20, 2010 Plant speciation in continental island floras as exemplified by Nigella in the Aegean Archipelago Hans Peter Comes, Andreas Tribsch and Christiane Bittkau Phil. Trans. R. Soc. B 2008 363, 3083-3096 doi: 10.1098/rstb.2008.0063 References This article cites 99 articles, 17 of which can be accessed free http://rstb.royalsocietypublishing.org/content/363/1506/3083.full.html#ref-list-1 Article cited in: http://rstb.royalsocietypublishing.org/content/363/1506/3083.full.html#related-urls Rapid response Respond to this article http://rstb.royalsocietypublishing.org/letters/submit/royptb;363/1506/3083 Email alerting service Receive free email alerts when new articles cite this article - sign up in the box at the top right-hand corner of the article or click here To subscribe to Phil. Trans. R. Soc. B go to: http://rstb.royalsocietypublishing.org/subscriptions This journal is © 2008 The Royal Society Downloaded from rstb.royalsocietypublishing.org on May 20, 2010 Phil.Trans.R.Soc.B(2008)363,3083–3096 doi:10.1098/rstb.2008.0063 Publishedonline25June2008 Plant speciation in continental island floras as exemplified by Nigella in the Aegean Archipelago Hans Peter Comes1,*, Andreas Tribsch1 and Christiane Bittkau2 1Fachbereich fu¨rOrganismische Biologie, Paris-Lodron-Universita¨tSalzburg, 5020Salzburg,Austria 2Institut fu¨rSpezielle Botanik undBotanischer Garten, Johannes Gutenberg-Universita¨tMainz,55099 Mainz, Germany Continental shelf island systems, createdby rising sea levels, provideapremier setting for studying the effects of geographical isolation on non-adaptive radiation and allopatric speciation brought about by genetic drift. The Aegean Archipelago forms a highly fragmented complex of mostly continental shelf islands that have become disconnected from each other and the mainland in relatively recent geological times (ca !5.2Ma). These ecologically fairly homogenous islands thus provide a suitable biogeographic context for assessing the relative influences of past range fragmentation, colonization, gene flow and drift on taxon diversification. Indeed, recent molecular biogeographic studies on the Aegean Nigella arvensis complex, combining phylogenetic, phylogeographicandpopulationlevelapproaches,exemplifytheimportanceofallopatryandgenetic drift coupled with restricted gene flow in driving plant speciation in this continental archipelago at differenttemporalandspatialscales.Whiletherecent(LatePleistocene)radiationofAegeanNigella, as well as possible instances of incipient speciation (in the Cyclades), is shown to be strongly conditionedby(palaeo)geographicfactors(includingchangesinsealevel),shiftsinbreedingsystem (selfing) and associated isolating mechanisms have also contributed to this radiation. By contrast, founder event speciation has probably played only a minor role, perhaps reflecting a migratory situation typical for continental archipelagos characterized by niche pre-emption because of a long establishedresidentflora.Overall,surveysofneutralmolecularmarkersinAegeanNigellahavesofar revealedpopulationgeneticprocessesthatconformremarkablywelltopredictionsraisedbygenetic drifttheory.Thechallengeisnowtogainmoredirectinsightsintotherelativeimportanceoftherole of genetic drift, as opposed to natural selection, in the phenotypic and reproductive divergence among these Aegean plant species. Keywords:Aegean palaeogeography;allopatric speciation; continental shelf islands; genetic drift; Nigella;non-adaptive radiation 1.INTRODUCTION see Linder 2008). In most cases, such radiations have Thetermspeciesradiation,i.e.thedivergentevolution been attributed to ecological opportunities afforded of a relatively large, monophyletic group of species by the emergence of new habitats and the absence within a relatively short time (Stanley 1979; Schluter of competition on recently formed oceanic islands 2000),ismostfrequentlyusedincombinationwiththe and continental mountain ranges, rather than to the epithet ‘adaptive’ (Givnish 1997), which in turn evolution of key morphological or physiological inno- implies ecological specialization and the probable vations (but see Klaket al.2003;Kayet al.2005). co-occurrence of closely related species (MacArthur By contrast, only a very few studies have explicitly 1972; Schluter 2000; Savolainen & Forest 2005). In addressed the question of whether species prolifera- recent years, molecular phylogenetic data have pro- tion might occur by ‘non-adaptive radiation’ (sensu vided a number of spectacular examples of explosive Cain1944;Givnish 1997;Savolainen &Forest2005), adaptive species radiations for plants in the recent i.e. without appreciable ecological divergence and evolutionarypast(LateTertiary/Quaternary),oneither evolution of corresponding adaptations, and simply as a consequence of geographical isolation and allopa- oceanic islands (Baldwin & Sanderson 1998; Bo¨hle tric divergence among sibling species maintaining et al. 1996; see also Whittaker & Ferna´ndez-Palacios similarecologicalnichesoverevolutionarytimescales. 2007) or continents (Richardson et al. 2001; Klak Such an allopatric model of non-adaptive species etal.2003;Kayetal.2005;Hughes&Eastwood2006; radiationessentiallyimpliesthatmutationandrandom genetic drift, rather than habitat-mediated selection, *Authorforcorrespondence([email protected]). will be the primary factors causing divergence of populations occupying ecologically similar habitats. Onecontributionof12toaThemeIssue‘Speciationinplantsand animals:patternandprocess’. The best-known examples include species-rich taxa of 3083 Thisjournalisq2008TheRoyalSociety Downloaded from rstb.royalsocietypublishing.org on May 20, 2010 3084 H. P.Comeset al. AegeanNigella speciation N. a. arvensis N. a. glauca N. a. aristata N. d. barbro N. icarica N. d. jenny N. d. degenii N. a. brevifolia N. d. minor N. stricta N. a. brevifolia N. carpatha km N. a. brevifolia N. doerfleri 0 50 100 Figure1.Approximatedistributionrangesofthesixspecies(12taxa)oftheAegeanNigellaarvensiscomplex(afterStrid2002; see also table 1). Solid versus dashed lines delineate outcrossing versus selfing taxa. Note that the range of N. arvensis ssp. brevifoliaonCretecomprisesonlyfourknownpopulations.N.a.,N.arvensis;N.d.,N.degenii. land snails on either Crete (Albinaria: Gittenberger Ferna´ndez-Palacios 2007). However, based on their 1991; Giokas 2000) or the Madeiran island of Porto thorough reassessment of the speciation literature, Santo (Helicidae: Cameron et al. 1996), as well as Coyne& Orr(2004,pp.383–410)recently concluded particular plant lineages in the Aegean Archipelago thatfirmempiricalevidenceforaroleofgeneticdriftin (e.g. Snogerup 1967; Strid 1970; Stork 1972; see §2). speciation is rare, and they also raised the same In addition, Savolainen & Forest (2005) claimed objection about drift-induced peripatric or founder- that the term non-adaptive radiation might also be effect speciation. Also, mathematical models indicate an appropriate way of describing the diversification that conditions for speciation via drift are fairly of the temperate legume genus Astragalus (see also restrictive (Barton 1996; Coyne et al. 1997; Turelli Sanderson & Wojciechowski 1996; Sanderson 1998). et al. 2001; Coyne & Orr 2004). For example, while Similarly, Whittaker & Ferna´ndez-Palacios (2007) showing that geographical isolation can lead to hypothesized that a non-adaptive component might complete reproductive isolation (Nei et al. 1983), beresponsiblefortheradiationofseveralMacaronesian such models have also revealed that even a small plantgenera(i.e.Cheirolophus,Helianthemum,Limonium) amount of gene flow can retard differentiation by drift that exploit similar habitats in different islands. (Coyne & Orr 2004, p. 87). Moreover, genetic drift is Althoughasystematicsurveyoftheirprevalencewould generally thought to cause speciation and quantitative be valuable, it is perhaps no coincidence that most phenotypic evolution more slowly than does either cases proposed for non-adaptive radiation are from directionalselectionaloneorincombinationwithdrift island systems. In such geographical settings, random (Coyne et al. 1997). On the other hand, there is good genetic drift (whether in established populations or evidence for a role of genetic drift in chromosomal during founder events) has long been implicated as speciation, particularly in plants (Rieseberg 2001). contributingtospeciationviz.reproductiveisolationand Again, especially within plants, a probable important phenotypic evolution (e.g. Mayr 1954; Carson 1975; mechanism favouring speciation by drift is selfing, Templeton1980;seealsoGrant1998). which has recently been invoked in a study of arctic Although patterns of molecular variation in island Draba species as providing an efficient non-adaptive populations often do correspond to what would be expected if the loci were affected by genetic drift (e.g. means for the accumulation of intraspecific hybrid Wendel & Percival 1990; Barrett 1996; Berry 1996; incompatibilities (Grundt et al. 2006). Another well- Stuessy et al. 2006), the contribution of genetic drift known, but fairly unusual, example for speciation by to island speciation remains the subject of keen drift involves shell coil reversals in land snails debate (e.g. Barton 1996; Grant 1998; Whittaker & (Gittenberger 1988; Ueshima &Asami 2003). Phil.Trans.R.Soc.B(2008) Downloaded from rstb.royalsocietypublishing.org on May 20, 2010 AegeanNigella speciation H. P.Comeset al. 3085 Table1.Geographicaldistribution,habitatpreference,floweringperiodandbreedingsystemofthetaxaoftheNigellaarvensis complex (after Strid1970and modified accordingto Bittkau& Comesinpress).Note that mosttaxaoccur atlowaltitudes (approx.0–500ma.s.l.)exceptforNigellaarvensisssp.glauca(0–1000ma.s.l.)andNigellaicarica(500–1000ma.s.l.). geographical flowering breeding Nigellataxon distribution habitat period system N.arvensisL. ssp.arvensis NGreece/EuropeandNAfrica cultivated/abandonedfields VI–IX outcrosser ssp.aristata(Sibth. SandCGreekmainland,Euboea phrygana,abandonedfields,roadsides mid-V–VII outcrosser andSm.)Nym. ssp.glauca(Boiss.) EAegeanislands,WTurkey, phrygana,abandonedfields,olive VI–VIII outcrosser Terracc. Thrakia groves ssp.brevifoliaStrid Elafonisos(southofPeloponnisos) phrygana,abandonedfields,roadsides VI–VIII outcrosser WCretea,Rhodes N.carpathaStrid SKarpathos,Kasos phrygana mid-V–mid-VII outcrosser N.degeniiVierh. ssp.degenii CandSCycladesb phrygana,cultivatedfields,seashores VI–VII outcrosser ssp.barbroStrid NWCyclades(Andros,Tinos, phrygana,roadsides,seashores mid-V–mid-VII outcrosser Mikonos) ssp.jennyStrid NWCyclades(Siros) phrygana,abandonedfields end-V–mid-VI outcrosser ssp.minorStrid SCyclades(Pakhia,southof phrygana end-V–VI outcrosser Anafi) N.icaricaStrid Ikaria phrygana VI–VII outcrosser N.doerfleriVierh. Andikithira,Cyclades,Crete phrygana end-IV–mid-V selfer N.strictaStrid Kithira,SWCrete stablesanddunes mid-IV–mid-V selfer aKnownfromonlyfourlocationsonCrete(Strid1970;Z.Kypriotakis2002,personalcommunication). bKnownfromthefollowingislands:Amorgos,Andiparos,Folegandros,Kimolos,Milos,Naxos,Paros,Santorin,Sifnos,Sikinos. There can be no doubt that natural (and sexual) islands(Sumatra,Java,Borneo),andtheNorthPacific selection plays a much larger role in speciation and Vancouver and Queen Charlotte islands). Such island phenotypicdiversificationthandoesdrift(e.g.Macnair& systemshavemanyoftheadvantagesofoceanicislands Gardner 1998; Rieseberg et al. 2002; Coyne & Orr (e.g. Hawaii, Gala´pagos, Canaries) in that they allow 2004;Waser&Campbell2004;Johnson2006;Noor& insights into colonization and restricted gene flow. As Feder2006;seeLexer&Widmer2008).Nonetheless, an important difference, however, continental islands it would seem to be premature to close the book on alsoprovideapremiersettingforstudyingtheeffectsof speciation by drift, no less than dismissing its role in past range fragmentation via geologically dated sea phenotypic evolution at all. Rather, considering both barriers, and thus for testing genetic drift-models of theoryandempiricalevidence,apotentialroleofnon- allopatric speciation and non-adaptive radiation adaptive radiation by drift will still apply to all species (Wright 1940;Mayr1954; Schluter 2000). occurring in effectively isolated (allopatric) popu- In fact, the most suggestive evidence for non- lations, especially if these are permanently or tempor- adaptiveradiationcomesfromplantgroupsinhabiting arily small (Wright 1948; Slatkin 1985). Yet, despite the ecologically fairly homogenous continental island allopatric speciation being widely considered as the system of the Aegean Archipelago (Runemark 1969; most common geographical mode (Barraclough & Strid 1970; Barrett 1996; Levin 2000). This claim, Vogler 2000; Turelli et al. 2001; Coyne & Orr 2004), specifically, relates to the biosystematic work of Strid and notwithstanding an expanding literature on the (1970) conducted on the so-called Nigella arvensis joint roles of allopatry and ecology in speciation complex (Ranunculaceae, tribe Nigelleae), which (Schluter 2000; Ogden & Thorpe 2002; Wiens 2004; contains six species (12taxa) that are mostly allo-/or Kozak & Wiens 2006; Thorpe et al. 2008), the parapatrically distributed on numerous islands and relationship between geographical isolation, drift, adjacent mainland areas of Greece and western phenotypic differentiation and reproductive isolation Anatolia (figure 1, table 1). Based on crossing experi- remains poorly understood in ‘non-ecological specia- ments, aswellas morphological and palaeogeographic tion’ (sensu Schluter 2000). evidence, Strid (1970) argued that differentiation in the N. arvensis complex has been determined princi- 2.THE AEGEAN ARCHIPELAGO ASA pally by two factors: (i) range fragmentation triggered LABORATORY OFALLOPATRICSPECIATION by Plio-/Pleistocene changes in sea level; and AND NON-ADAPTIVE RADIATION (ii) subsequent evolution of the fragmented ancestral Perhaps the most suitable natural laboratories for stockviageneticdriftviz.non-adaptiveradiation.Strid studying the effects of geographical isolation on (1970) further hypothesized that genetic drift in the allopatric speciation are ‘continental shelf islands’ complexislargelybroughtaboutbydrasticpopulation (sensu Whittaker & Ferna´ndez-Palacios 2007) that bottlenecks in established populations because of have become disconnected from each other and/or (seasonal) fluctuations in population size rather than the mainland in relatively recent geological times (e.g. by repeated interisland (or mainland-island) founding mostAegeanislands,theSouth-EastAsianSundashelf colonizations. His contemporaries working on other Phil.Trans.R.Soc.B(2008) Downloaded from rstb.royalsocietypublishing.org on May 20, 2010 3086 H. P.Comeset al. AegeanNigella speciation endemic components of the Aegean flora favoured high degree of human-mediated secondary sympatry similar evolutionary scenarios (e.g. Snogerup 1967; of plant populations, and thus may create an obstacle Stork1972;Bentzer1973;VonBothmer1974,1987). to the interpretation of extant biogeographic and Here, we will synthesize data from our recent and molecular patterns in the Aegean region. In the ongoing molecular biogeographic studies of the N. arvensis complex, however, there is only limited N. arvensis complex using phylogenetic, phylo- sympatry (except for a pair of species in the Cyclades; geographic and population level approaches in order see §4), as would be predicted under a scenario of todissecttheevolutionaryhistoryofthisplantgroupat non-adaptiveradiation. increasingly smaller temporal and spatial scales. We will first introduce the Aegean region, its palaeo- (b)Aegean palaeogeography geographic history and our study group, and then Paleogeographic/climatic changes affecting the entire giveaconciseoverviewofthemainresults,withafocus MediterraneanbasinduringtheLateTertiary/Quatern- on the times and orders of evolutionary divergence aryhavelikelyexertedamajorinfluenceonthepatterns within the complex, and the population genetic ofgeographicalisolationandallopatricspeciationinthe processes underlying its supposedly non-adaptive N. arvensis complex. Of particular relevance is the fact radiation. In essence, our data indicate that allopatry thatseveral palaeogeographicevents providemaximum (often but not exclusively vicariant) and genetic drift timesofisolationbetweentheAegeanislandsfromeach (coupled with restricted gene exchange) are the other andthemainland. Inthisrespect, the ‘Messinian dominant evolutionary processes driving population salinitycrisis’(ca6.0–5.3Ma;Duggenetal.2003)clearly differentiation and speciation in Aegean Nigella. sets an upper time boundary for the last, large-scale Because of lack of comparable data, however, we are reconnection of the entire region. During this time presently far from being able to evaluate whether this period,theMediterraneancompletelydriedupasaresult scenario also holds for other plant radiations in the oftheclosingoftheStraitofGibraltar,andtheAegean Aegean Archipelago or similarly complex continental islands then became mountains in a steppe or desert island systems. Moreover, as a critical limitation, our (Blondel & Aronson 1999). At some 5.2G0.1Ma, the surveys of neutral molecular polymorphisms alone Strait of Gibraltar reopened and the basin was refilled providelittledirect insight intothepotentialevolution- fromtheAtlanticOceanwithin1000years(Duggenetal. ary processes (drift versus selection) and genetic 2003).Asaresult,oneofthefirstAegeanseabarriersto changes causing phenotypic differentiation and repro- become permanently established was the East Aegean ductive isolation in our study group. We conclude, Sea, separating the modern Cyclades from the east therefore, with some current and possible future Aegeanislands(ca4.5Ma;Chatzimanolisetal.2003). approachesfor tacklingthese more difficult issues. The sequence of Plio-/Pleistocene island formation is particularly well documented for ‘Hellenic Arc’ islands. As another consequence of post-Messinian 3.THE STUDY AREA AND ITS flooding, Crete became isolated, and it has remained PALAEOGEOGRAPHIC HISTORY isolated since; Kithira, however, was completely (a) The Aegean Archipelago submerged and did not re-emerge until the late The Aegean Archipelago (figure 1) is located between Pliocene (ca 3.0Ma; Meulenkamp et al. 1972). At the Greek peninsula in the west and the Turkish coast about the same time, Karpathos became permanently of Asia Minor in the east.Approximately 611kmlong disconnected from Rhodes-Anatolia (Daams & Van and 299km wide, the total area of the Aegean Sea is der Weerd 1980), while the permanent separation of some 214000km2, whereby the total area of the RhodesfromAnatoliaoccurredatthePlio-/Pleistocene multitude of Aegean islands and islets (O100) is boundary(ca 2.4Ma; Kuss1975). approximately 9012km2 (Strid 1997, 2002). Geo- For much of the remainder of the archipelago, the graphically,theseislands can be roughly arranged into Quaternarysea-levelrecordcanbeusedtoapproximate fourgroups:(i)EuboeaandtheSporades,agrouplying timesofislandseparation.Duringtheglacialperiodsof off mainland Greece; (ii) the east Aegean islands, theRiss(250–125ka)andWu¨rm(75–10ka),sealevels includingIkariaandSamos,lyingofftheTurkishcoast; were,respectively,200and120mlowerthanatpresent (iii) the Cyclades, including 16 major islands such as (Sfenthourakis 1996; Perissoratis & Conispoliatis Andros, Tinos, Mikonos, Siros, Paros, Naxos and the 2003). Thus, because of shallow sea barriers, Euboea volcanic island of Santorin/Thera (from north to andseveraleastAegeanislands(e.g.Samos) havebeen south); and (iv) the southern Aegean (‘Hellenic Arc’) separated from the nearby mainland only after the last islands, including Kithira, Crete, Karpathos, Kasos glacial maximum (LGM; 23–18ka). The seafloor and Rhodes, which mark the southern boundary of between Samos and Ikaria, however, is much deeper the archipelago. (albeit less than 200m), so these neighbouring islands Ecologically, the commonest habitat type in these were already separated by the sea transgression islandsisatypicaldryMediterraneanmaquis-phrygana following the Riss glacial (Beerli et al. 1996). community of evergreen, mostly degraded (under) AsregardstheCyclades, most oftheseislands once shrub vegetation (Blondel & Aronson 1999), which is formed a landmass that remained connected to main- also one of the most important habitats for the land Greece/Euboea until the end of the Pliocene N. arvensis complex. Concomitantly, this vegetation (ca 2Ma; Fattorini 2002; Chatzimanolis et al. 2003). type reflects the influence of humans on all aspects of During the glacial/interglacial periods of the Pleisto- the environment that dates back to more than 5000 cene, most central islands were alternately joined and years BP (Broodbank 2000). This may have led to a disentangled, with the most recent fragmentation Phil.Trans.R.Soc.B(2008) Downloaded from rstb.royalsocietypublishing.org on May 20, 2010 AegeanNigella speciation H. P.Comeset al. 3087 dating back to late-glacial/Early Holocene times Inaccordancewithoneofthemajorcriteriafornon- (Lambeck 1996). Notably, several Cycladic islands adaptive radiation, the habitats occupied by most taxa further south (e.g. Milos, Santorin/Thera, Anafi) have of the N. arvensis complex vary little among islands or been formed by volcanic activity that started in the the mainland (table 1). At least from a macro-habitat Pliocene, and most of them were temporarily con- perspective, these taxa are ecologically unspecialized nected to nearby islands during some period of their withpopulationsoccurringinawidevarietyofmoreor existence (Sfenthourakis 1996). As will be seen in §7, less disturbed habitats (0–1000m a.s.l.) such as stony the Cyclades’ intricate configuration and post-glacial seashores, phrygana communities, abandoned/culti- palaeogeography offers an outstanding context in vatedfieldsorroadsides.Incontrasttotheoutcrossers, which to explore the times and orders of recent however, the two selfers tend to occupy or tolerate population splitting in a Nigella species endemic to morearidhabitats(Strid1969,1970),includingstable these islands. sand dunes (N. stricta) or phrygana sites on small, dry and isolated islands (N.doerfleri). 4.THE MODEL SYSTEM: THE NIGELLAARVENSIS COMPLEX 5.SPECIES-LEVEL PHYLOGENETICSAND (a) Breeding system and experimental PHYLOGEOGRAPHY crossing barriers (a)Phylogenetics Nigella is a genus of approximately 23 annual, diploid A recent molecular-phylogenetic study (Bittkau & (2nZ12) and self-compatible species of mainly Med- Comesinpress),withacompletespeciesrepresentation iterranean/Southwest Asian distribution (Strid 2002). ofNigellas.lat.(ZNigellaanditssistergenusGaridella) The six species (12taxa) of the N. arvensis complex and based on internal transcribed spacer (ITS) (figure 1, table 1) are mainly distinguished by flower sequencesofnuclearribosomalDNA,revealedthatthe andfruitcharactersand,toalesserextent,byvegetative N.arvensiscomplextrulyqualifiesasa‘radiation’.The characters,alwaysincombinationwiththeirgeographi- clock-calibrated ITS phylogeny generated showed caldistribution(Strid1970).Fourspecies(N.arvensis, that all the species of the complex began diverging Nigella carpatha, Nigella degenii, Nigella icarica) are rapidly from a single ancestor in the Late Pleistocene, summer-flowering and predominantly outbreeding, between ca 0.78 (G0.39) and 0.16 (G0.08)Ma, whereas self-pollination is the normal condition in the whereas the origin of this ancestor could only roughly spring-flowering, smaller-sized Nigella doerfleri and beapproximatedtobetween6.2and1.3Ma(Messinian Nigella stricta, because their styles regularly twist to mid-Pleistocene). In addition, both log lineages- around the dehiscing anthers (Strid 1969). Attempts through-time(LTT)plots(Harveyetal.1994;Neeetal. to produce fertile hybrids between N. doerfleri and 1994) and likelihood survival analyses (Paradis 1998) N. stricta, or between the selfers and outcrossers, have indicated that the proliferation of Nigella species in the been unsuccessful, whereas species of the latter group Aegean region caused a significant departure from a are interfertile, with pollen-fertility values in hybrids stochastic, speciation–extinction process of diversifica- ranging between 30 and100% (Strid 1970). tionduringtheevolutionofNigellas.lat.Otherwise,the genuswouldhavefolloweda‘constant-ratesbirth–death’ (b)Geographicaldistributionand habitat (CR-BD)model(Harveyetal.1994;Barraclough&Nee The10taxaoftheoutcrossingspeciesexhibitastriking 2001), a scenario under which speciation rate and allo- or parapatric pattern of geographical distribution extinctionrateareassumedtobeconstantacrosslineages (figure 1, table 1). The four outcrossing subspecies of withinaphylogenyandovertime(thenullhypothesis). N.arvensisaredistributedintheformofaringaround Finally,wecouldshowthattheobservedupturncurvein the Cyclades: ssp. aristata occurs in southern Greece, theLTT-plottowardsthepresent,mainlyassociatedwith ssp.arvensisinnorthernGreece,ssp.glaucaintheeast theAegeanN.arvensiscomplex,wasmuchmorerecent Aegean islands and Anatolia, and ssp. brevifolia is thanwouldhavebeenexpectedundertheCR-BDmodel disjunctlydistributedonthe‘HellenicArc’(Elafonisos, (Bittkau&Comesinpress).Thisfindingindicatesthat Crete, Rhodes). The three remaining outcrossing the recent radiation of the Aegean group truly reflects species are island endemics: N. carpatha is found on an increase in the rate of speciation and not decreased Karpathos (and nearby Kasos), N. icarica on Ikaria, extinction(Barraclough&Nee2001).Thus,themajor and N. degenii is widespread throughout the Cyclades, hypothesis emerging from this molecular-phylogenetic where four allopatric subspecies are recognized which study is that the accelerated rate of speciation in are endemic to island groups (spp. barbro, degenii) or the Aegean N. arvensis complex is possibly related single islands (spp. jenny, minor). Like the outcrossing to increased opportunities for allopatric speciation taxa,theselfingspeciesarealsonotsympatricwithone afforded by the (palaeo)geographic complexity of the another: N. stricta is restricted to Kithira and south- archipelago, combined with Late Pleistocene changes western Crete, whereas N. doerfleri is more wide- inclimateandsealevel(Bittkau&Comesinpress). spread (Andikithira, central/eastern Crete, Cyclades, numerous small islets). The broad range overlap (b) Phylogeography between N. doerfleri and N. degenii in the Cyclades is For plant groups inhabiting continental island exceptional for the entire complex (figure 1). It thus systems, the smaller effective population size of their appears that only species of the complex that differ in chloroplast (cp)DNA, compared with the nuclear breeding system (and associated traits) are able to live genome(Birkyetal.1989),shouldbeparticularlyuseful insympatryonthesameislands. for detecting signatures of past range fragmentation Phil.Trans.R.Soc.B(2008) Downloaded from rstb.royalsocietypublishing.org on May 20, 2010 3088 H. P.Comeset al. AegeanNigella speciation The main conclusion drawn from the phylogeo- N. icarica N. arvensis graphic cpDNA study is that various regional sets of populations of the Aegean N. arvensis alliance are N. arvensis geographically isolated and probably on independent (ssp. brevifolia, evolutionary trajectories. That said, none of its four Rhodes) constituent species is fixed for a particular cpDNA haplotype, or uniquely characterized by a haplotype clade. The same applies to the two selfers N. doerfleri and N. stricta, which are also nearly fixed for an ancestral haplotype (A) present in all species of the alliance except N. icarica (C. Bittkau & H. P. Comes 2004, unpublished data). In general, such cpDNA haplotype sharing across species boundaries could N. degenii N. doerfleri indicate either the retention of ancestral haplotypes N. stricta or cytoplasmic gene flow/hybridization following species divergence. Figure 2. Unrooted neighbour-joining phenogram based on To distinguish between these two possibilities, we Nei & Li’s (1979) genetic distances among 106 multilocus have undertaken a preliminary survey of amplified AFLPphenotypesobservedin106plantsfrom42populations fragmentlengthpolymorphisms(AFLPs)inallspecies of the Nigella arvensis complex, and representing all species of the complex (except N. carpatha), taking advantage andsubspeciesexceptN.carpatha,N.arvensisssp.arvensisand of the fact that such fast evolving markers are N. degenii ssp. minor. A list containing detailed geographical predominantly of nuclear origin and also potentially information on the location of all collection sites and the suitable for resolving ‘shallow phylogenies’ (Meudt & numberofindividualssurveyedpertaxon/siteisavailableupon Clarke 2007). Genetic distance analysis of individual request. Solid (dashed) branches indicate bootstrap support valuesabove(below)70%basedon10000replicates.AFLP AFLP phenotypes results in an unrooted network profiles were generated from total genomic DNA samples (figure 2) in which almost all species of the complex (Bittkau & Comes 2005) using the AFLP methodology and surveyedareseparatedinclusters,albeitofunresolved scoring procedure described in Kropf et al. (2006) with two relationships. While a more extensive AFLP survey EcoRI–MseI primer combinations (E38/M53; E37/M56) that ofapproximately500individualswiththree(insteadof resultedinatotalof264polymorphicfragments. two) primer combinations is currently underway to obtainbetter-resolvedspeciesandpopulationrelation- (becauseofrisingseas)anditsconsequences,including ships (U. Jaros, A. Tribsch & H. P. Comes 2008, restricted dispersal, drift and population bottlenecks/ unpublished data), two important insights are already founderevents.InitialevidencethatpresentAegeansea evident from figure 2. First, most nominal species of barriers keep Nigella populations isolated emerged thecomplexaregeneticallydistinctentities,suggesting from a survey of polymerase chain reaction-restriction that erratic instances of cpDNA haplotype sharing fragment length polymorphism haplotype variation of across species boundaries reflect incomplete lineage cpDNA(whichismaternallyinheritedinNigella)inthe sortingratherthanongoing,interspecifichybridization. four outcrossing and cross-compatible species of the Second,N.arvensisssp.brevifoliasampledfromRhodes complex (Bittkau & Comes 2005), henceforth called formsaseparateclusterrelativetoconspecificmaterial the ‘N. arvensis alliance’. Based on the geographical ofmainlandoriginsurveyed(spp.aristata/glauca).This distributions of major haplotype clades, as inferred pronounced intraspecific divergence of N. arvensis, from genealogical (nested clade) reconstructions, this which is also supported by cpDNA data (Bittkau & studyidentifiedtheEasternand,toalesserextent,the Comes 2005), once more indicates a complete lack of Western/Southern Aegean Seas as important isolating gene flowacross the East Aegean Seaviz.theStrait of barrierstoNigelladispersal,ashadlongbeensuspected Marmaris (approx. 18km), which separates Rhodes for other members of the Aegean flora (Rechinger from the nearby Anatolian coast. It is noteworthy, 1950; Runemark 1969, 1980; Greuter 1979; Strid however, that lack of gene exchange between Nigella 1996). Notably, there was a sharp break in cpDNA populations from either Rhodes (N. arvensis ssp. haplotypefrequenciesobservedacrosstheEastAegean brevifolia) or the Cyclades (N. degenii) and Anatolia Sea,i.e.atthe‘Rechinger’sline’(sensuStrid1996),the (N.arvensisssp.glauca)mayalsobeexplained,atleast phytogeographic borderline between Europe and Asia in part, by reciprocal cross-incompatibility or reduced Minor. For the first time this testified molecularly to hybrid fertility, respectively, as indicated by experi- the long-held notion that the East Aegean Sea poses a mentalcrosses(Strid1970,p.117).Thus,evenifseed strong physical barrier to dispersal in many plants materialfromRhodesortheCycladesmanagestocross (Greuter1979;Runemark1980;Strid1996),aswellas the East Aegean Sea,the probabilityof gene exchange for various animals (e.g. Giokas 2000; Fattorini 2002; withresidentAnatolianpopulationsisprobablysmall. Poulakakis et al. 2003, 2005). Furthermore, the restricted distribution of a large number of derived cpDNA haplotypes or clades, on both mainland areas 6.BIOGEOGRAPHY OFAEGEAN NIGELLA andparticularislands,identifiedtheN.arvensisalliance SPECIATION as a plant group with low seed dispersal ability and, Knowingaboutthestrengthofageographicalbarrierin consequently, high susceptibility to genetic drift currentlyrestrictinggeneflowdoes nottellusanything (Bittkau & Comes2005). aboutitsimportanceinspeciation(Coyne&Orr2004). Phil.Trans.R.Soc.B(2008) Downloaded from rstb.royalsocietypublishing.org on May 20, 2010 AegeanNigella speciation H. P.Comeset al. 3089 Hence, a crucial question is whether the establishment oflocaladaptationinresponsetolowwateravailability, of Aegean sea barriers actually caused allopatric which thus would render the two selfers unlikely vicariant speciation within the N. arvensis complex, candidates for non-adaptive speciation. Given their i.e. via range fragmentation of a formerly widespread genetic distinctness (figure 2), it also seems unlikely (pan-Aegean)ancestorofMessiniantomid-Pleistocene that selfing evolved under sympatric conditions as a origin (Bittkau & Comes 2005). However, considering consequenceofreinforcementtoreducegeneflowfrom the Late Pleistocenediversification of the complex (see nearby outcrossing populations not adapted to arid §5a) and in the light of our current understanding of habitats. Rather, we presume that selfing and associ- post-Messinian sea-level changes in the Aegean (see ated isolating mechanisms (phenological, ecological, §3b), there are only two probable instances of sea genetic) evolved more or less simultaneously as barrier-inducedvicariantspeciation,i.e.(i)theoriginof by-products of geographical isolation (Coyne & Orr N. degenii following the separation of the Cycladic 1999). Experimental crossings and, ideally, sequence landmass (peninsula) from mainland Greece/Euboea analysesofrelevantgenes(Noor&Feder2006)willbe (ca 2Ma); and (ii) the probably more recent origin of neededtodeterminethetemporalorderinwhichthese N. icarica following a post-glacial rise in sea level that multiple isolating mechanisms evolved between out- separated Ikaria from Samos/Anatolia (ca 125–75ka). crossersandselfers.Importantly,whilemostspeciesof By contrast, for the remaining island species the N. arvensis complex have allopatric distributions, (N. doerfleri, N. stricta, N. carpatha), as well as the precludingdirectobservationofreproductiveisolation, enigmaticN.arvensisssp.brevifolia,thereislittlereason N. doerfleri and N. degenii coexist in the Cyclades to suspect that rising seaswere required for their origin. (figure 1) without hybridizing (Strid 1970). Evidently, This is because these taxa mainly (albeit not it is here, in an area of possibly secondary range exclusively) occur on a group of southern ‘Hellenic overlap, that these multiple isolating factors act Arc’ islands that became permanently isolated well together to prevent gene flow between these two before the Late Pleistocene (i.e. Crete, ca 5.5Ma; species, allowingtheircoexistence. Kithira,Karpathos/Kasos,ca3Ma;Rhodes,ca2.4Ma; To summarize so far, the above data indicate that see § 3b). In support of this hypothesis, the cpDNA plant speciation in the Aegean Archipelago cannot data indicate a peripatric origin of N. carpatha through simply be viewed as mainly resulting from vicariant relatively recent over-sea dispersal from Anatolia (or speciationviahistoricalrangefragmentationbecauseof Ikaria) to Karpathos (Bittkau & Comes 2005). post-Messinian (Plio-/Pleistocene) rises in sea level. However, the spatial-temporal origins of N. doerfleri, Thiswastheviewcommonlyheldbyearlierbiosystem- N. stricta and also N. arvensis ssp. brevifolia are not yet atistsworkingonallopatricplant‘neo-endemics’ofthe fullyunderstood.Nonetheless,ourcurrentcpDNAand Aegean region (e.g. Snogerup 1967; Stork 1972; AFLP data (Bittkau & Comes 2005; figure 2) show no Bentzer 1973; Von Bothmer 1974, 1987). In contrast, evidence of a founder origin of these island taxa from themoleculardatapresentedheresuggestaremarkably the mainland. Rather, they may constitute remnants of recent (Late Pleistocene) radiation of the Aegean the formerly widespread ancestral stock from which N. arvensis complex that resulted from a combination they originated in situ within islands long in existence, of vicariant, (infrequent) peripatric and intra-island and subsequently dispersed to others to attain their speciation, with the latter process probably facilitated presently disjunct distributions (figure 1). orevenaccomplishedbyachangeinpollinationsystem At this juncture it is worth emphasizing that the (selfing).Thatspeciationviafoundereventshadonlya change fromoutcrossingtopredominant selfing inthe minor role in the group’s history perhaps comes as a ancestor(s)ofN.doerfleriandN.strictaprobablyhada surprise, but apart from taxon-specific attributes (e.g. major effect on the origin of these species. Although dispersal ability, genetic incompatibilities) this could ‘selfing’ not necessarily constitutes a reproductive also reflect a migratory situation more typical for isolating-barrier causing speciation (Coyne & Orr continental (compared with oceanic) island systems, 2004), it nonetheless may contribute to pre-mating wheremostsuccessfulintroductionsofplantdiaspores barriers between diverging populations and accelerate are generally thought to fail because of niche pre- the development of intrinsic post-zygotic isolation emption of a resident flora that has long been well (Fishman & Stratton 2004; Grundt et al. 2006). In adapted to existing habitat conditions (Runemark fact, the reproductive failure of hybrids in crosses 1969;Silvertown2004;Silvertownet al.2005). involvingthesetwoselfers(see§4a)mayreflectgenicor More generally speaking, and to compound this chromosomal incompatibilities that evolved as aresult complexity, it should be recalled that although plant of reductions in effective population size (because of radiations in continental archipelagos are liable to selfing) and the fixation of alternate, negatively be partly conditioned by (palaeo)environmental interacting alleles or chromosomal segments in separ- (extrinsic) factors, such as changes in sea level and ated populations, often referred to as Dobzhansky– geographical configurations, they may also be pro- Muller incompatibilities (Orr & Turelli 2001). More- foundly influenced by taxon-specific (intrinsic) attri- over,bothspeciesshowotherpossibleeffectsofselfing, butes that are thought to influence rates of including reductions in overall habit and flower size, diversification, including types of pollination, advanced flowering, as well as successful interisland mechanisms of dispersal, generation time, population colonization and production of populations in novel, size or hybridization/polyploidization (Coyne & Orr particularly arid habitats at which the outcrossers are 2004; Ricklefs 2007). Future evaluation of Aegean absent (see §3b). Hence, the evolution of selfing may plant radiations will thus require not only good have also led to reproductive isolation as a by-product phylogenetic and phylogeographic data for the Phil.Trans.R.Soc.B(2008) Downloaded from rstb.royalsocietypublishing.org on May 20, 2010 3090 H. P.Comeset al. AegeanNigella speciation (a) (b) 20km (c) (d) Andros Tinos Mikonos Siros Paros/Naxos Santorin Figure3.ReconstructionofthetopographyoftheinsularlandmassofCycladiaattheLGM(ca23–18ka;(a)18ka,(b)14ka, (c) 12.5ka, (d) 10ka), and the degrees by which it drowned because of rising sea levels towards the end of the Pleistocene (modifiedfromLambeck(1996)asillustratedinBroodbank(2000)).NamesofislandsfromwhereNigelladegeniipopulations weresampledforAFLPanalysis(seefigure4,table2)areunderlined. existence of sister species and accurate molecular landmass of approximately 6000km2, named clocks to broaden our understanding of the biogeo- ‘Cycladia’ (with several satellite islands), in the place graphy of plant speciation in the Aegean Archipelago, ofthemodernCyclades.Moreover,refinedcalibrations but also similar molecular investigations across ofeustaticchangescombinedwithmodellingofcrustal evolutionarily independent lineages to separate the reboundenabledLambeck(1996)togenerateremark- effects of historical versus taxon-specific factors on ably precise reconstructions of the topography of diversification. Finally, further detailed intraspecific Cycladiaandthedegreesbywhichitdrownedtowards investigations of the population genetic structure of the end of the Pleistocene because of rising seas (ca Aegean plant species together with experimental 14–10ka;figure3).BythestartoftheHolocene,there studies are needed to test the strong hypothesis of was a substantial tract of land still present between Strid (1970) and others (Snogerup 1967; Runemark ‘Greater Paros’ (i.e. the current islands of Paros, 1969, 1980; Stork 1972; Bentzer 1973; VonBothmer Antiparos, Despotikon) and Naxos, but this land 1974,1987)thatallopatricdivergenceviageneticdriftis bridge also rapidly vanished by ca 9ka (Van Andel & theruleratherthantheexceptionforplantevolutionin Shackleton1982). theAegeanArchipelago. Our cpDNA population-level survey of Cycladic N.degenii(Bittkau&Comes2005),althoughsuffering from low levels of haplotype variation, has revealed 7.INCIPIENT SPECIATION IN THE CYCLADES high levels of genetic differentiation (F ) both ST An ideal Aegean plant species with which to test the between sampled islands connected during the LGM impact of recent population fragmentation on the (i.e. Tinos, Mikonos, Siros, Paros, Naxos), as well potential for incipient allopatric speciation and as between populations at the species-wide range genetic/phenotypic divergence via drift is the Cycladic (F Z0.782 versus 0.840). Both results indicate that ST endemic N. degenii. The Cycladic region de facto seed flow (NmZ0.139 versus 0.095) is not strong providesanexcellentsettingforstudyingthehistoryof enough to prevent N. degenii populations from populationsplittinganditsevolutionaryconsequences, diverging purely by drift (i.e. Nm/1; Wright 1931; giventhatsea-levelchangesintheareasincetheLGM Slatkin 1985). To further investigate the species’ (ca !23–18ka) are well documented (Van Andel & history of population splitting and divergence in the Shackleton 1982; Lambeck 1996; Perissoratis & Cyclades, we have undertaken a survey of AFLP Conispoliatis 2003). Starting with the LGM, there variation in 110 individuals, representing six Cycladia wasamajordropinsealevelthatcreatedalargeinsular populations (ssp. barbro: Tinos, Mikonos; ssp. jenny: Phil.Trans.R.Soc.B(2008) Downloaded from rstb.royalsocietypublishing.org on May 20, 2010 AegeanNigella speciation H. P.Comeset al. 3091 Table2.LocationofsitesusedinthepresentAFLPstudyforNigelladegeniidistributedacrosstheCyclades,andsummaryof AFLPvariationforatotalofsevenpopulations.VariationisdescribedbytheproportionofAFLPmarkers(‘loci’)polymorphic (PLP)andNei’s(1973)expectedheterozygosityH (or‘genediversity’,withF assumedtobezero)calculatedbytheprogram e IS FDIST(Beaumont&Nichols1996). N.degeniitaxon localitya latitude(N) longitude(E) PLP H n e ssp.barbro Tinos,EofAgiosFokasbeach 3783105400 2581301100 0.342 0.110 14 Mikonos,Panormosbeach 3782802700 2582104200 0.361 0.109 18 ssp.jenny Siros,nearKokkinabeach(Finikas) 3782304400 2485201600 0.390 0.152 19 ssp.degenii Paros1,SofCapeKorakas 3780805200 2581302000 0.417 0.128 15 Paros2,SofLefkes 3780300800 2581204500 0.426 0.155 18 Naxos,WofHimaros 3780303600 2582703900 0.333 0.110 13 Santorin,SofFira 3682404000 2582601500 0.491 0.173 13 aAllcollectionsbyC.Bittkau.Seefigure4(inset)forgeographicallocation. Naxos Tinos Mikonos Siros Naxos Paros Paros1 nii Santorin ge e d p. s Paros2 s 62 68 Santorin y 96 Siros nn e p. j s s 59 Mikonos o r b r a b Tinos p. s s 100 57 Figure 4. Neighbour-joining analysis of 110 multilocus AFLP phenotypes observedin 110 plants from sevenpopulations of Nigella degenii from the Cyclades (inset) based on Nei & Li’s (1979) genetic distances (see table 2 for locality information). Intermingled individuals from populations Paros1, Paros2 and Naxos are highlighted (dark, medium and light grey, respectively).Bootstrapvalues(O50%)basedon10000replicatesareshownabovebranches.AFLPprofilesweregenerated fromtotalgenomicDNAsamples(Bittkau&Comes2005)accordingtoScho¨nswetteretal.(2002)withslightmodifications, and using two EcoRI–MseI primer combinations (ECACA/MCCAGC; ECACC/MCCAGA) that resulted in a total of 108 polymorphicfragments. Siros; ssp. degenii: Paros1/2, Naxos), plus one ssp. recognizing five population clusters corresponding to degeniipopulationfromthevolcanicislandofSantorin/ Tinos, Mikonos Siros,Santorin and Paros (1/2)/Naxos Thera, which has been isolated from Cycladia at least (figure5).Moreover,therewasnosignificantassociation since the last interglacial (Snogerup 1967; Greuter between genetic and geographical distances of these 1979; Sfenthourakis 1996; see table 2 and inset of populations(rZ0.249,Manteltest,pZ0.177).Thislack figure4forsamplingsites).Geneticdistanceanalysisof ofisolation-by-distanceinconjunctionwithpronounced these individuals results in an unrooted neighbour- population subdivision (F Z0.290; figure 5) once ST joiningphenogram(figure4)inwhichmostindividuals moresuggeststhatgeneticdrifthasbeenofmuchgreater cluster according to their populations except for historical importance in these Cycladic populations those from Paros (1/2) and Naxos, which cannot be compared with recurrent gene flow (Hutchison & clearly distinguished. Bayesian analysis of population Templeton1999). structure as implemented in the program Structure Despite low internal bootstrap support, the tree v. 2.1 (Pritchard et al. 2000) confirms this pattern by topologyshowninfigure4isstrikinglycongruentwith Phil.Trans.R.Soc.B(2008)