ebook img

butterflies – North Africa is the source of ancestral populations of all Pararg PDF

12 Pages·2006·0.15 MB·English
by  
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview butterflies – North Africa is the source of ancestral populations of all Pararg

SystematicEntomology(2006),31,621–632 DOI:10.1111/j.1365-3113.2006.00333.x Speciation in Pararge (Satyrinae: Nymphalidae) butterflies – North Africa is the source of ancestral populations of all Pararge species ELISABET WEINGARTNER, NIKLAS WAHLBERG and SO¨ REN NYLIN DepartmentofZoology,StockholmUniversity,Stockholm,Sweden Abstract. The genus Pararge comprises three species: P. aegeria, distributed in Europe and North Africa; P. xiphia, endemic to Madeira; and P. xiphioides, endemic to the Canary Islands. Two subspecies are recognized in P. aegeria, P. a. tircis and P. a. aegeria, distributed in northern and southern Europe, respectively. In the 1970s, P. aegeria appeared on Madeira. However, despite the status of P. aegeria as a model species in ecological studies, the evolutionary history of Pararge remains unknown. We studied the phylogenetic relationships of the three Pararge species, using the mitochondrial gene cytochrome oxidase subunit I and the nuclear gene wingless to infer modes and times of speciation. On the basis of our analyses, Pararge forms a strongly supported monophyletic group, with the DNA haplotypes of the three species also forming well-supported monophyletic groups. We found that P. xiphia diverged first from the common ancestor a maximum of five million years ago, with P. xiphioides and P. aegeria being sister species that diverged a maximum of three million years ago. The two subspecies, P. a. tircis and P. a. aegeria, were not distinguishable on the basis of DNA haplotypes; instead, our data clearly distinguished between European speci- mens and those from North Africa. Madeiran P. aegeria has North African haplotypes and thus originated from there rather than from Europe. We hypothe- size that the Mediterranean Sea forms a strong barrier to dispersal for Pararge butterflies,andhasdonesoforapproximatelythepastonemillionyears. Introduction (Nylin et al., 1993, 1995; Gotthard et al., 1994, 2000) and thermoregulatory ability in different colour morphs (van Thespeckledwoodbutterfly,Parargeaegeria(Linnaeus,1758), Dyck&Matthysen,1998).However,theevolutionaryhistory isoneofthemostfrequentlyusedbutterflyspeciesinecological of the genus remains unstudied. This paper focuses on the studies.Ithasbeenamodelspeciesinstudiesofbehaviourand speciation processes in Pararge Hu¨bner, 1819 to understand territorialdefence(Davies,1978;Wickman&Wiklund,1983; better the differences found between species and populations Shreeve,1984,1987),lifehistorytraits(Gotthardetal.,1994, and to provide a phylogenetic framework for comparative 2000; Sibly et al., 1997; van Dyck et al., 1997; van Dyck & studies. Wiklund, 2002; Stevens, 2004) and the impact of climate The genus Pararge is considered presently to comprise changeonthespeciesdistributions(Hilletal.,1999).Thetwo three species: P. aegeria, P. xiphia (Fabricius, 1775) and subspecies,P.a.tircis(Godart,1821)andP.a.aegeria,have P. xiphioides Staudinger, 1871. Previously, these were beenusedcommonlyincomparativestudiesoflifehistorytraits considered to form a single polytypic species with four clearly distinguished subspecies, three of which differed in their genitalia (Higgins, 1975). Today, P. xiphia and P. xiphioides have separate species status, but P.a. aegeria Correspondence: Niklas Wahlberg, Department of Zoology, andP.a.tircisaretreatedstillastwoconspecificsubspecies. StockholmUniversity,10691Stockholm,Sweden.E-mail:niklas. ParargeaegeriaisdistributedthroughoutEurope(eastto [email protected] the Urals), Asia Minor and the Middle East (Bozano, #2006TheAuthors Journalcompilation#2006TheRoyalEntomologicalSociety 621 622 E.Weingartneretal. 1999), and from the southern slopes of the High Atlas in Lasiommatamaera(Linnaeus,1758),L.megera(Linnaeus, northern Africa to Scandinavia. Pararge aegeria tircis and 1767), L. petropolitana (Fabricius, 1787) and Lopinga P. a. aegeria are distributed north and south of the Alps, achine (Scopoli, 1763). The origin and voucher identifica- respectively, with a zone of intermediates in France. The tions of the specimens used in this study are detailed in two subspecies are distinguishable clearly on morphology: Appendix 1. Specimens are presently stored at (cid:1)20(cid:2)C at P. a.tircishasayellowgroundcolourandP. a.aegeriahas the Department of Zoology, Stockholm University, anorangegroundcolour.Since1976,P. a.aegeriahasbeen Sweden, and will be deposited at the Swedish Museum of recorded also from Madeira, where it is now well estab- NaturalHistory,Stockholm,Sweden. lished (Owen et al., 1986). Thus, the colonization of Madeira by P. aegeria offers a unique opportunity to study the interactions between the endemic P. xiphia and Molecularmethods the newcomer (Owen et al., 1986; Jones, 1992; Shreeve & Smith,1992;Jonesetal.,1998). We extracted DNA from two legs of the frozen or The other two species, P. xiphia and P. xiphioides, dried butterflies using a QIAGEN (Hilden, Germany) are endemic to Madeira and the Canary Islands, respec- DNeasy Extraction Kit following the manufacturer’s tively. Pararge xiphioides has been recorded from the instructions. CanaryIslandsofGranCanaria,Tenerife,LaGomeraand Wesequencedtwogeneregions:onemitochondrialgene, La Palma. With such species distributions, the speciation cytochrome oxidase subunit I (COI), of 1450bp, and one processinParargeislikelytohavetakenplaceallopatrically. nuclear gene, wingless, of 407bp. COI was sequenced for As species of the genus Pararge frequently are used as all seventy-one individuals sampled. From these seventy- models in ecological studies, it is of great importance to oneindividuals,wechosetwenty-threeindividualsshowing understandtheirevolutionaryhistory.Ithasbeensuggested the most divergent COI sequences for which we then also that Madeira was colonized first by individuals carried by sequenced wingless. Polymerase chain reactions (PCRs) strong air currents or, in the case of P. aegeria, perhaps were performed in a 20ml reaction volume. The protocol deliberately or accidentally introduced by humans (Owen used was: 1 ml DNA, 11.5 ml H O, 2 ml 10(cid:3) buffer, 2 ml 2 etal.,1986).Wehypothesizethatthecommon ancestorof MgCl2(25 mM),1 mlofeachprimer,0.4mldNTP(10 mM) Pararge species colonized both Madeira and the Canary and 0.1ml Taq Gold. The cycling profile was 95(cid:2)C for Islands, followed by speciation due to isolation. However, 5 min, 35 cycles of 94(cid:2)C for 30s, 47(cid:2)C for 30s, 72(cid:2)C wereMadeiraandtheCanaryIslandscolonizedbyindivid- for 1min 30s and 72(cid:2)C for 10min. We used the primers ualsfromNorthAfricaorfromtheIberianPeninsula?The ofBrower&DeSalle(1998)forwinglessandofWahlberg& colonization of Madeira by P. aegeria might offer a suffi- Zimmermann(2000)forCOI. ciently recent event, in which it is still possible to find the The same primers were used in the sequencing reaction. samehaplotypeinpopulationsonthemainland.Thisisone Inaddition,forCOI,theinternalprimer‘Patty’(Wahlberg oftheaimsofthispaper.Inaddition,wewanttodetermine & Nylin, 2003) was used. The sequencing reactions were whether P. aegeria is divided into two genetically differen- performed in a total volume of 20ml: 13.5 ml H O, 1.5 ml 2 tiated subpopulations corresponding to the two currently 10(cid:3)buffer,2mlCEQ2000Mix,2mlprimerand1 mlDNA. recognized subspecies, that is, are P. a. aegeria and Sequencing was undertaken using a Beckman Coulter P. a. tircis separate entities, with different mitochondrial (Bromma, Sweden) CEQ8000 capillary sequencer. All haplotypes,aswecanexpectfrommorphology? sequence chromatograms were checked in the program BIOEDIT(Hall,1999)andalignedbyeye. Materialsandmethods Phylogeneticanalyses Sampling In total, three datasets were analysed: COI and Pararge aegeria was sampled as widely as possible to winglessaloneandbothgenescombined.Forthecombined maximize the number of represented haplotypes analysis,weusedonlythoseindividuals(twenty-threespeci- (Appendix 1). Both subspecies are represented by several mens) for which both genes had been sequenced (see populations: P. a. aegeria from populations in southern Appendix1). Portugal, central Spain, southern France and southern The most parsimonious cladograms were searched for Greece; P. a. tircis from populations in Sweden, the UK from theequally weightedandunordereddatamatrix using andBelgium.Parargexiphioideswassampledfromthreeof a heuristic search algorithm in the program NONA 2.0 the four islands where it is present, but unfortunately we (Goloboff,1998)viatheprogramWINCLADA1.00.08(Nixon, could not obtain specimens from the island of Gran 2002). The heuristic searches were conducted with 200 Canaria. Pararge xiphia was sampled from various local- randomadditionreplicatesusingtreebisection-reconnection ities on Madeira. As outgroup taxa, we chose representa- (TBR)branchswapping,withtentreesheldduringeachstep tives from other genera and species belonging to the same and a final swapping to completion. The robustness of the tribe, Parargini: Kirinia roxelana (Cramer, 1777), phylogenywasevaluatedusingbootstrap(Felsenstein,1985) #2006TheAuthors Journalcompilation#2006TheRoyalEntomologicalSociety,SystematicEntomology,31,621–632 PhylogenyofPararge 623 and Bremer supportvalues (Bremer, 1988, 1994). Bootstrap Dating of the phylogenies was made with the program valueswereobtainedusing1000replicates.Whencalculating R8S (Sanderson, 2004). We used three algorithms to esti- Bremer supports, TREEROT version 2 (Sorensen, 1999) was mate the ages of divergence: the Langley–Fitch method used in conjunction with PAUP 4.0b10 (Swofford, 2001) to (which assumes a global molecular clock), nonparametric defineconstrainttreesforeachnode.Themostparsimonious ratesmoothing(Sanderson,1997)andpenalizedlikelihood trees without a particular node were found in PAUP using a (Sanderson,2002).Forthepenalizedlikelihoodmethod,we heuristic search with random additions of twenty replicates estimated the value of the smoothing parameter using a andonetreeheldduringeachstep.Maxtreeswassetto500 cross-validationprocedure.Inallcases,topreventthealgo- andbranchswappingwasperformedusingTBR.Toevaluate rithms from converging on a local optimum, the searches the contribution of each dataset to the combined result, we werebegunatfivedifferentinitialtimeestimates.Thelocal used partitioned Bremer support values (Baker & DeSalle, stability of the solutions for each estimate was checked by 1997;Gatesyetal.,1999).Pairwisesequencedivergencesand perturbing them and restarting the search five times. We base frequencies were calculated using MEGA version 3 compared our age estimates with those obtained from the (Kumaretal.,2004). now standard constant mutation rate of 2.3% pairwise sequencedivergencepermillionyears(Brower,1994). Errors in age estimates resulting from the stochastic natureofsubstitutionprocesseswereassessedusingaboot- Timingofdivergences strap resampling procedure, with the help of Perl scripts madeavailablebyEriksson(2005).Onehundredbootstrap Fordatingpurposes,themostparsimonioustreefromthe replicates of the combined dataset were constructed using combinedanalysiswasprunedtocontainonerepresentative the SEQBOOT program (Felsenstein, 1993), and branch of each major Pararge lineage (i.e. one P. xiphia, one lengthswerecalculatedusingmaximumlikelihoodandthe P. xiphioides from each island, one Moroccan P. aegeria GTR þG þ I model for each replicate and then input to and one P. aegeria from Europe), together with the out- the R8S program. The divergence ages were estimated for groups. Branch lengths for this tree were estimated from the bootstrapped datasets using nonparametric rate the combined dataset using maximum likelihood and a smoothing and fixing the age of the split between the La GTRþ Gþ Imodel.Branchlengthsandparametervalues Palma and La Gomera lineages of P. xiphioides at two were estimated in PAUP 4.0b10 (Swofford, 2001) with and millionyears.Bootstrapestimatesofthestandarddeviation without a molecular clock enforced. The evolution of the for eachnodewerecalculatedforthe agedistribution esti- sequences in a clocklike manner was tested with the like- matesobtained. lihoodratiotest.Ifthelikelihoodsofthemodelsaresignifi- cantlydifferent,themolecularclockcanberejected.Owing to practical problems with estimating branch lengths at the Results root node (see accompanying instructions of Sanderson, 2004), after the branch lengths had been estimated, two The COI dataset (n¼ 71 individuals) consisted of 1450 outgroup taxa (Kirinia Moore, 1893 and Lopinga Moore, base pairs, of which 235 were parsimony informative. The 1893) were pruned from the tree, leaving Pararge and its base frequencies were: T, 40.9%; C, 15.4%; A, 30.1%; G, putativesistergenusLasiommataWestwood,1841. 13.6%.Intheheuristicsearch,110treesoflength604steps To obtain reliable age estimates for clades, the ages of were found (CI¼0.71, RI ¼0.92). The strict consensus particular nodes need to be fixed or constrained using treewithbootstrapvaluesisshowninFig.1. external information from fossils or geological events. In Thewinglessdataset(n ¼ 23)consistedof407basepairs, thepresentcase,nofossilsofeitherParargeoritsputative ofwhichthirty-threewereparsimonyinformative.Thebase sister genus Lasiommata are known. There are two geo- frequencies were: T, 19.2%; C, 28.4%; A, 22.6%; G, logical events that potentially can be used to calibrate our 29.7%. The analysis found a single most parsimonious age estimates. The age of Madeira, at five million years tree of length seventy steps (CI ¼0.91, RI¼ 0.94). The (Geldmacher & Hoernle, 2000), can be used to place the tree,withbranchlengths,isshowninFig. 2. maximum age of the divergence between the ancestor of Analysisofthecombineddataset(n ¼23)generatedone P.xiphiaandtheothertwoParargespecies.However,this mostparsimonioustree(Fig.3)with647steps(CI ¼0.74, estimateiscomplicatedbythefactthattheislandofPorto RI ¼ 0.83). Of the 1857 characters, 260 were parsimony Santoismucholder(fourteenmillionyears;Geldmacher& informative. Hoernle, 2000) and, although it does not support popula- ThegenusParargeissupportedstronglyasa monophy- tionsofP.xiphianow,itmayhavedonesointhepast.The letic group by both the COI and wingless datasets (Figs 1 second calibration age is a more certain maximum age, and 2). We find that P. aegeria is the sister species to namely the age of the island of La Palma at two million P. xiphioides and that P. xiphia is sister to these two. In years(Ancocheaetal.,1994).Currentmethodsrequirethat theCOIanalyses,theP.xiphioidespopulationsondifferent atleastonenode agebefixed.Wetherefore fixed thesplit islands are monophyletic with high bootstrap support between the La Gomera lineage and the La Palma lineage values. The population from Tenerife is sister to those ofP.xiphioidesattwomillionyears. from La Gomera and La Palma. However, using wingless, #2006TheAuthors Journalcompilation#2006TheRoyalEntomologicalSociety,SystematicEntomology,31,621–632 624 E.Weingartneretal. Kirinia roxelana Lopinga achine Lasiommata maera Lasiommata megera Lasiommata petropolitana Pararge xiphia 24-32 100 Pararge xiphia 5-1 Pararge xiphia 24-29 Pararge xiphia 3-12 Pararge xiphia 5-2 Pararge xiphia 24-28 Pararge xiphia 24-31 Pararge xiphia 24-30 Pararge xiphia 24-27 100 Pararge xiphioides 3-10 Te Pararge xiphioides 28-16 Te 70 100 Pararge xiphioides 28-17 Te Pararge xiphioides 27-1 Lg 98 75 PParaarragreg ex ixpihpihoiiodiedse s2 72-76- 2L pLg 100 Pararge xiphioides 28-19 Lp Pararge xiphioides 28-18 Lp Pararge xiphioides 27-4 Lp Pararge xiphioides 27-5 Lp Pararge xiphioides 27-7 Lp Pararge a. aegeria 6-4 Ma Pararge a. aegeria 27-12 MC Pararge a. aegeria 27-8 MC Pararge a. aegeria 25-7 Ma Pararge a. aegeria 6-5 Ma 99 Pararge a. aegeria 6-6 Ma Pararge a. aegeria 27-13 MC Pararge a. aegeria 27-15 MC 100 Pararge a. aegeria 27-11 MC Pararge a. aegeria 27-9 MC Pararge a. aegeria 6-7 Ma Pararge a. aegeria 26-4 Ma Pararge a. aegeria 26-5 Ma Pararge a. aegeria 27-14 MC Pararge a. aegeria 25-9 Ma Pararge a. aegeria 25-10 Ma Pararge a. aegeria 29-19 Ma Pararge a. aegeria 29-20 Ma Pararge a. aegeria 29-21 Ma Pararge a. aegeria 25-8 Ma 100 Pararge a. aegeria 1-3 FR Pararge a. tircis 10-8 BE Pararge a. tircis 16-3 UK Pararge a. tircis 17-13 SE Pararge a. aegeria 25-27 GR Pararge a. aegeria 1-1 FR Pararge a. aegeria 1-6 FR Pararge a. aegeria 1-4 FR Fig.1. Strict consensus tree of 110 most Pararge a. tircis 16-1 UK parsimonious trees of equal length, 604 Pararge a. tircis 18-1 SE Pararge a. aegeria 24-18 FR steps,basedonthecytochromeoxidasesub- 100 PPaararargrgee a a. .t iarceisg e1r8ia-3 1 S7-E8 FR unitI(COI)dataset(n¼71taxa).Bootstrap Pararge a. aegeria 1-8 FR support is shown above the branches. Pararge a. aegeria 17-9 FR Pararge a. tircis 17-14 SE Branchlengthsareproportionaltothenum- Pararge a. tircis 18-2SE Pararge a. aegeria 25-28 GR ber of changes inferred from parsimony. Pararge a. tircis 16-2 UK Geographic abbreviations: BE, Belgium; Pararge a. aegeria 1-2 FR Pararge a. aegeria 31-5 ES ES, Spain; FR, France; GR, Greece; Lg, Pararge a. aegeria 30-7 GR 1 step Pararge a. aegeria 31-6 ES La Gomera; Lp, La Palma; Ma, Madeira; Pararge a. aegeria 33-22 PT MC, Morocco; SE, Sweden; Te, Tenerife; Pararge a. aegeria 33-23 PT Pararge a. aegeria 33-21 PT UK,UnitedKingdom. only the population from La Palma is resolved as mono- The phylogeny of the combined dataset unsurprisingly phyletic and does not share haplotypes with the other correspondstothetopologies ofCOI andwingless(Fig.3). islands. In contrast, the two subspecies of P. aegeria are The monophyly of all three species receives moderate to notrecoveredasmonophyletic.Instead,accordingtoCOI, strong Bremer support values, although this is mainly due the species is divided into two clades, with continental to the COI partition. The sister group relationship of European specimens forming one group, and the African P. aegeria and P. xiphioides is supported strongly by COI andMadeiranspecimensformingtheother.Theanalysisof datapartition,butweaklybythe winglesspartition.Within winglesssupportstheEuropeanclade,buttheAfricanand P. xiphioides, the populations from different islands receive Madeirangroupisunresolved. highsupportfromtheCOIpartition,butlowornosupport Thegenusitselfreceiveshighbootstrapsupportforboth fromthewinglesspartition.Thehighlysupporteddivisionof genes (Figs 1 and 2), as do P. aegeria and P. xiphia. P. aegeria into European and North Africanþ Madeiran However, the value is less for P. xiphioides, 70% for COI cladesisduealmostentirelytotheCOIpartition. and 87% for wingless. The sister group relationship of PairwisesequencedivergencesareshowninTable1.For P. aegeria and P. xiphioides is strongly supported by COI COI,anaverageofabout1.9%differencebetweenthetwo (bootstrapvalueof99%),butlesssobywingless(bootstrap major P. aegeria lineages is detected. The pairwise COI valueof70%). sequencedivergencewithinthetwolineagesiscloseto0% #2006TheAuthors Journalcompilation#2006TheRoyalEntomologicalSociety,SystematicEntomology,31,621–632 PhylogenyofPararge 625 Kirinia roxelana Lopingaachine Lasiommata maera Lasiommata megera Lasiommata petropolitana 89 Pararge xiphia 5-1 Pararge xiphia 24-29 Pararge xiphioides 3-10 Te Pararge xiphioides 27-1 Lg 93 87 Pararge xiphioides 27-2 Lg Pararge xiphioides 28-16 Te Pararge xiphioides 28-17 Te 80 Pararge xiphioides 27-6 Lp 70 Pararge xiphioides 28-19 Lp Pararge a. aegeria 6-4 Ma Fig.2. Mostparsimonioustreefoundforthe Pararge a. aegeria 27-12 MC winglessdataset(treelength,70steps;n¼23 98 Pararge a. aegeria 27-8 MC taxa). Bootstrap support is shown above the Pararge a. aegeria 25-7 Ma branches. Branch lengths are proportional to Pararge a. aegeria 1-3 FR the number of changes inferred from parsi- mony. Geographic abbreviations: BE, Pararge a.tircis 10-8 BE 62 Belgium; ES, Spain; FR, France; GR, Pararge a.tircis 16-3 UK Greece; Lg, La Gomera; Lp, La Palma; Ma, Madeira; MC, Morocco; SE, Sweden; Te, 1 step Pararge a.tircis 17-13 SE Tenerife;UK,UnitedKingdom. Pararge a. aegeria 25-27 GR (i.e. haplotypes tendto differby onlyonebase pair outof (1994)molecularclockestimateof2.3%pairwisesequence 1450 sequenced). The most common haplotype in Europe divergence per million years to Pararge gives similar ages also is the most widespread, being found in individuals for the younger splits, including the assumed two million collected in southern Portugal, central Spain, southern yearsforthesplitbetweenP.xiphioidesonLaGomeraand France, the UK and Sweden. Most haplotypes in LaPalma(Table2). Morocco and Madeira are identical to each other. The highest intraspecific diversity in COI is found within P.xiphioides,withamaximumof4.4%differencebetween specimens.Haplotypediversitywithinthemajorlineagesis Discussion verypoor(closeto0%),suggestingrepeatedbottlenecksor selective sweeps purging the populations of mitochondrial Speciationinbutterfliesmaybepromotedbyanexpanded haplotypediversity. host plant range (Janz et al., 2001; Weingartner et al., Clockwise evolution of the sequences in the combined 2006).However,speciesofthegenusParargearespecialists datasetwasrejected(P <0.05),andsowebaseourdiscus- of grasses and show no evidence for an expanded host sion on the nonclock results. However, the choice of algo- range.Nevertheless,hostplantusecouldstillbeanimport- rithm did not change the age estimates very much antfactorinthediversificationprocess.Ofthetenendemic (Table2). Calibrating the split between the La Palma and butterflyspeciesoftheCanaryIslands,fivearegrass-feeding La Gomera lineages of P. xiphioides at two million years satyrines (Hipparchia wyssii Christoph, 1889, H. bacchus provides us with a maximum estimate of the ages of the (Higgins,1967),H.gomera(Higgins,1967),H.tilosi (Manil othersplitsinthetree(Fig.4).Theageofthesplitbetween 1984)andP.xiphioides).Oftheothers,oneisastrongflier P.xiphiaandtherestofParargeisestimatedatamaximum (Vanessa vulcanica Godart, 1819; Nymphalidae), one is a of 5.6 million years with a standard deviation of 0.9 hesperid (Thymelicus christi Rebel, 1894), which also feeds (Table2). This age is consistent with the age of Madeira ongrass,twobelongtothefamilyPieridae(Pierischeiranthi and suggests that the ancestor of P. xiphia may have (Hu¨bner, 1808), Gonepteryx cleobule (Hu¨bner 1825)) and reached the island soon after it emerged from the ocean. one is a lycaenid (Cyclyrius webbianus (Brulle´, 1840)). The TheageofthesplitbetweenP.xiphioidesandP.aegeriais dominance of grass feeders on the Canary Islands might estimatedatamaximumof3.1millionyears((cid:4) 0.46),while reflect their good colonization ability. Grass is a resource thesplitbetweenNorthAfricanpopulationsandEuropean availablealmosteverywhere,anditmightbeeasytocolonize populations of P. aegeria may have occurred around 0.87 new habitatseven though the butterflies themselvesare not millionyears((cid:4) 0.24)ago.Interestingly,applyingBrower’s verygooddispersers. #2006TheAuthors Journalcompilation#2006TheRoyalEntomologicalSociety,SystematicEntomology,31,621–632 626 E.Weingartneretal. Kirinia roxelana Lopinga achine Lasiommata maera 100 Lasiommata megera 100 Lasiommata petropolitana Pararge xiphia 27,2 100 Pararge xiphia 58 Pararge xiphioides Te 24,0 100 Pararge xiphioides Te Pararge xiphioides Te 10,3 3,2 99 86 Pararge xiphioides Lg 20,0 100 Pararge xiphioides Lg 4,0 Pararge xiphioides Lp 80 20,2 100 Pararge xiphioides Lp 11,1 Pararge a. aegeria Ma 99 Fig.3. Most parsimonious tree of the Pararge a. aegeria Ma combineddataset[cytochromeoxidasesub- 13,0 unit I (COI)þwingless]. Tree length, 647 100 Pararge a. aegeria MC steps(n¼23).PartitionedBremersupport valuesare shown abovethebranches.The Pararge a.aegeria MC firstvaluereferstothesupportcontributed 17,3 byCOIandthesecondtothesupportvalue 100 Pararge a. aegeria FR of the wingless dataset. Bootstrap support is shown below the branches. Branch Pararge a. aegeria GR lengths have been estimated using maxi- 11,1 mum likelihood (GTRþGþI model). Pararge a. tircis SE 100 Geographic abbreviations: BE, Belgium; Pararge a. tircis UK ES, Spain; FR, France; GR, Greece; Lg, La Gomera; Lp, La Palma; Ma, Madeira; Pararge a. tircis BE MC, Morocco; SE, Sweden; Te, Tenerife; 0.01 UK,UnitedKingdom. In the case of Pararge, it thus appears that speciation Where these ancestors came from is thus of interest. The happened through processes not directly related to host islandofMadeiraisabout850kmsouth-westofPortugal, plants. Quite likely, P. xiphia and P. xiphioides speciated about540kmwestofMoroccoandabout450kmnorthof after their ancestors had colonized their respective islands. theCanaryIslands.TheislandofGranCanaria(theclosest Table1. AverageuncorrectedpairwisedistancesofcytochromeoxidasesubunitI(COI)haplotypesforeachofthemajorlineagesofPararge. P.xiphioides P.xiphioides P.xiphioides P.aegeria P.aegeria Taxon Lasiommata P.xiphia (Te) (Lg) (Lp) (Africa) (Europe) Lasiommata 0.061 0.089 0.098 0.093 0.093 0.095 0.092 P.xiphia 0.001 0.060 0.057 0.056 0.057 0.064 P.xiphioides(Te) 0.002 0.043 0.043 0.047 0.049 P.xiphioides(Lg) 0.001 0.032 0.048 0.045 P.xiphioides(Lp) 0.000 0.047 0.048 P.aegeria(Africa) 0.000 0.019 P.aegeria(Europe) 0.001 Lg,LaGomera;Lp,LaPalma;Te,Tenerife. #2006TheAuthors Journalcompilation#2006TheRoyalEntomologicalSociety,SystematicEntomology,31,621–632 PhylogenyofPararge 627 Table2. EstimatedagesforsplitswithinPararge. Node Constant Clock NPRS PL BootSD Bootrange Root 4.1 9.07 9.96 9.88 1.59 4.26–12.50 Pararge 2.6 5.16 5.63 5.64 0.90 3.02–8.13 P.aegeriaþP.xiphioides 2.1 3.09 3.06 3.08 0.46 2.14–4.56 P.aegeria 0.8 0.94 0.86 0.87 0.24 0.42–1.58 P.xiphioides 1.9 2.45 2.43 2.44 0.26 2.00–3.09 LaPalmaþLaGomeraa 1.4 2.00 2.00 2.00 0.00 2.00 Constant,assumingaconstantmutationrateof1.2%permillionyears(Brower,1994);NPRS,nonparametricratesmoothing;PL,penalized likelihoodwithasmoothingrateof0.0032;BootSD,standarddeviationofbootstrappeddatasets;Bootrange,minimumandmaximumages estimatedfrombootstrappeddatasets. aNodefixedattwomillionyears. islandtotheAfricanmainlandwhichhasParargeonit)is islands of the archipelago, most probably due to the lack about190kmwestoftheAfricancontinent.Thesourceof ofsuitablehabitatsontheseislands. the ancestors of P. xiphia is not clear and may have been One explanation for the short basal branch lengths for eitherEuropeorNorthAfrica. the P. xiphioides haplotype clade (see Fig. 3) could be a Most likely, the Canary Islands were colonized from rapid diversification process on these islands. This would NorthAfrica,fromwhichthenearestislandisonly95km. produce the observed basal branches with few character Although we have yet to sample P. xiphioides from Gran changes, but large genetic distances between islands, indi- Canaria,thepatternofcolonizationseemstobefromolder cating genetic isolation between them (Table1). High to younger islands, which corresponds to an east to west genetic divergences amongst the populations on different direction, with the population on Tenerife sister to the islandsintheCanaryIslands,comparedwiththemainland populationsonLaGomeraandLaPalma.Thisresultagrees populations, are found also in psyllids (Percy, 2003). The with other studies on Canary Islands colonizations estimatedmaximumageofP.xiphioidesis3.1millionyears (Carranza et al., 2000; Emerson et al., 2000a; Rees et al., ((cid:4)0.46)(Fig. 4).Thus,P. xiphioides ismuch younger than 2001). With regard to butterflies, there is one study of the theformationofTenerife(11.6millionyearsago)and,once genusGonepteryxLeach,1815,inwhichtheauthorssuggest it colonized the Canary Islands, it was able to spread that the colonization of the Canary Islands has occurred amongsttheislandsquickly. fromAfrica(Brunton&Hurst,1998).However,itiscrucial Contrary to what might be expected from morphology, toincludethepopulationonGranCanariainordertodraw we have not discovered any indication that P. a. aegeria any conclusions about the colonization pattern of P. andP. a.tircisareseparateevolutionaryentitiesbasedona xiphioides. The species has not been recorded on the two mitochondrial gene and a nuclear gene. Instead, we found islands Lanzarote and Fuerteventura, the easternmost twoseparatecladeswithinP.aegeriacorrespondingtothe Lasiommata maera Lasiommata megera Lasiommata petropolitana Pararge xiphia Pararge aegeria North Africa Pararge aegeria Europe Pararge xiphioides Tenerife Pararge xiphioides La Gomera * Pararge xiphioides La Palma Fig.4. Chronogram for Pararge. The 9 8 7 6 5 4 3 2 1 0 asteriskmarksthenodewhichwasfixedat twomillionyears. Millions of Years Before Present #2006TheAuthors Journalcompilation#2006TheRoyalEntomologicalSociety,SystematicEntomology,31,621–632 628 E.Weingartneretal. European and African þ Madeiran populations. This TheevolutionofParargebutterfliesisthusanexampleof pattern is consistent in both genes. However, the morpho- aseeminglysimplecaseofspeciationthat,infact,showsa logical differences indicate that other genes are differen- rather complicated pattern. In order to understand this tiated between P. a. tircis and southern European pattern, we must rely on the phylogeny. Five observations populations known as P. a. aegeria. The implication of canbemade:(1)thesistergrouptoParargeis(orappears this finding will affect comparative studies of P. a. tircis tobe)thePalaearcticgenusLasiommata(infact,allpoten- and P. a. aegeria in which Madeiran specimens have been tial sister groups to Pararge are Palaearctic; see Bozano, used as representatives of the subspecies P. a. aegeria (e.g. 1999); (2) the two island species of Pararge are not sister Nylin et al., 1994, 1995). It cannot be concluded that life species, suggesting that Madeira and the Canary Islands historydifferencesbetweenthetwosubspeciesareaconse- werecolonizedindependently;(3)thedirectionofcoloniza- quenceofrecentadaptationtothehabitatonMadeira(e.g. tionontheCanaryIslands(fromwesttoeast)accordswith Nylinetal.,1995).Instead,theMadeiranpopulationsmay other studies and suggests that the source of the colonists carry with them adaptations to habitats in North Africa, wasNorthAfrica;(4)therecentcolonizationofMadeiraby whicharebeneficialalsoonMadeira.Itisthuspreliminary P.aegeriatookplacefromNorthAfrica;and(5)thehaplo- to conclude that over 100 generations is enough time to types of P. aegeria form a monophyletic group to the allow life history traits to evolve on Madeira (Nylin et al., exclusion of the two other species and, importantly, the 1995), without a comparative study of populations from haplotypes in North Africaþ Madeira and in Europe are NorthAfricaandMadeira. reciprocallymonophyletic. It is of great interest to determine when the two mito- The foregoing observations suggest strongly the follow- chondrialcladesofP.aegeriadiverged,asthiscanhelpus ing scenario to us. The ancestor of the genus Pararge tounderstandtheprocessthatmighthavepromoteddiver- colonized North Africa from the Palaearctic, perhaps dur- sification. The results from dating should, however, be ingtheMessiniansalinitycrisis(assumingthatP.xiphiais interpreted carefully. When using the age of islands to atmostfivemillionyearsold).Theancestralpopulationsin calibrate nodes, we obtain the maximum age of a clade if Europesubsequentlywentextinct.Whenthisoccurred,we we assume that divergence has taken place on the island. cannot say, but at least prior to the recolonization of Most likely, the ancestral butterflies colonized the islands Europe by P. aegeria some four million years later (evi- later, maybe much later. However, the much used 2.3% denced by no endemic European species or ancient haplo- pairwise sequence divergence per million years (Brower, typesinEuropeanpopulationsofP.aegeria).Theancestral 1994) gives a very similar age (1.4 million years) for the species then colonized Madeira from North Africa soon split between La Gomera and La Palma lineages as our after the island appeared above sea level, and speciated assumed two million years (the age of the island of La subsequently and evolved into what we know currently as Palma),lendingsomecredencetotheageswediscuss. P. xiphia. The ancestral populations in North Africa con- Accordingtoourdating,thesplitbetweentheEuropean tinuedtoevolvethroughtimeuntilasecondcolonizationof P. aegeria and the North African populations occurred at theislandsandspeciationeventtookplace,thistimeinthe about0.87millionyears((cid:4)0.24)ago(Table2,Fig.4).This Canary Islands. The subsequent spread of P. xiphioides fallswithintheQuaternary,whichinEuropewasdominated amongsttheislandsappearstohavebeenrapid.Thecolo- byrepeatedintervalsofglaciationsthataffectedtheevolu- nizationoftheCanaryIslandsapparentlyalsowasasingle tion of European species (Hewitt, 1996). During glacial event, evidenced by the monophyly of current haplotypes periods,northernpopulationswereextinguishedorpushed ofP.xiphioides. furthersouth.Whentheclimatebecamewarmer,thepopu- After the divergence of P. xiphioides from ancestral P. lations expanded their ranges northwards. The aegeriapopulations,theNorthAfricanpopulationscontin- Mediterranean has been a strong barrier to migration ued to develop in isolation until perhaps 0.8 million years (Taberlet et al., 1998). During the Messinian salinity crisis ago, when Europe was colonized. This colonization also (5.57–4.93millionyearsago),whentheMediterraneanwas appears to be a single event, evidenced by the deep diver- desiccated,EuropeandAfricawerejoinedintoasingleland gence between, and reciprocal monophyly of, North mass(DeJong,1998,andreferencescitedtherein),andthis AfricanandEuropeanpopulations.Subsequenttothecolo- possiblyenabledmigrationbetweenthetwocontinents. nizationofEurope,thespecieseventuallyspreadtooccupy The phylogenetic patterns we have uncovered present a its current range. It is quite likely that the European difficult case for investigating the historical biogeography P.aegeriahasspreadandretracteditsrangeseveraltimes, of the group. Three of the four major lineages show auta- astherehavebeenseveralglaciationperiodsduringthepast pomorphic distributions (Madeira, Canary Islands and one million years. Recent observations in the U.K. have NorthAfrica),whereastheonelineagesharingitsdistribu- shown that P. aegeria has the ability to expand its range tion with the sister group (Europe) is in a highly derived rapidly in response to climate changes (Hill et al., 1999). position, suggesting convergence. Conventional methods GiventherelativelydeepdivergencebetweentheEuropean would suggest that Europe is the ancestral range of the andNorthAfricanlineages,itispossiblethattheEuropean genus Pararge, and this is undoubtedly true. However, it P.aegeriasurvivedglacialmaximainrefugiaawayfromthe doesnotshedlightonthesourcefromwhichtheislandsor Iberian peninsula, such as the Balkan peninsula. Further NorthAfricawerecolonized. evidence for the scenario described above comes from the #2006TheAuthors Journalcompilation#2006TheRoyalEntomologicalSociety,SystematicEntomology,31,621–632 PhylogenyofPararge 629 colorationofthewings.Fourtaxonomicentitiesretainthe entities within P. aegeria, one in North Africa (and ancestral orange wing colour (which is also found in Madeira), one in Europe south of the Alps and one in Lasiommata):P.xiphia,P.xiphioides,P. a.aegeria(North EuropenorthoftheAlps. Africa) and P. a. aegeria (Europe). Our phylogenetic There are names available for each of the three entities: hypothesis suggests that the yellow colour of the northern meoneStoll,1780(typelocality:Algeria),aegeriaLinnaeus, P. a. tircis has evolved later, perhaps as an adaptation to 1758 (type locality: South Europe and North Africa) and colderclimates. tircis Godart,1821(typelocality: France).However, using Analternativebiogeographicalhypothesis,andonemore thesenamesiscomplicatedbythefactthatthetypelocality parsimoniousbasedonourphylogenetichypothesis,would ofaegeriaisgivenas‘SouthEuropeandNorthAfrica’and be that Madeira was colonized first from Europe by the notypematerialexists(seeHoney&Scoble,2001).Fixing ancestralPararge.TheCanaryIslandswerecolonizedsub- thenametopopulationsinsouthernEuropewouldrequire sequently from Madeira by the ancestor of P. xiphioides a neotype, and such an act is beyond the scope of this andP.aegeria,andthenNorthAfricawascolonizedfrom paper. On the other hand, the names aegeria and tircis the Canary Islands (before the differentiation of have been used consistently for southern and northern P.xiphioidesontheislands).Finally,Europewascolonized European populations, respectively, and taking the name from North Africa, completing a circle that took perhaps meone for the North African and Madeiran populations five million years. Both hypotheses require that the ances- would not cause much taxonomic instability. We thus tral Pararge went extinct in Europe at some stage before raise the name meone Stoll, 1780 from synonymy to sub- the present. It is impossible to differentiate between these speciesstatus(Parargeaegeriameonestat.rev.)andrestrict two hypotheses based on the evidence we have on hand. its use to populations in North Africa and Madeira. The However,webelievethatindirectevidencefavoursthefirst name Pararge aegeria aegeria refers to populations in hypothesisoverthesecond.Studiesonotherinsectssuggest southern Europe with an orange background colour on thattheislandgroupshavebeencolonizedfromcontinental theforewings,andthenameParargeaegeriatircistopopu- sources (Brunton& Hurst, 1998; Emerson etal., 2000a,b; lationsinnorthernEuropewithalightyellowbackground Reesetal.,2001;Hundsdoerferetal.,2005).Althoughthe colourontheforewings. colonization of Madeira by P. aegeria from North Africa Different genetic lineages in Europe and Africa have cannot be seenas direct evidence of the same routeby the been found in several studies (Burban et al., 1999; Percy, ancestorofP.xiphia,intheabsenceofotherevidence,itis 2003; Hundsdoerfer et al., 2005). Quite clearly, the suggestive. There is some evidence of inter-island group Mediterranean remains a barrier for genetic exchange, colonizations, but so far the direction of colonization has even though the Iberian Peninsula and Morocco are sepa- been inferred to be from the Canary Islands to Madeira ratedbyonly14kmatonepoint(Dobson&Wright,2000). (e.g.Emersonetal.,2000b;Hundsdoerferetal.,2005). This does seem paradoxical, as Pararge clearly crossed Our results have important taxonomic implications for large bodies of water to reach Madeira and the Canary Pararge.Firstofall,P.aegeria,P.xiphiaandP.xiphioides Islands. However, a general pattern of intraspecific diver- clearlyformthreegoodspeciesthatproducemonophyletic gencebetweenthetwocontinentsmayexist,butstudies,as clades with both mitochondrial and nuclear genes. This yet,arefew,especiallyinflyinginsects.Asimilarpatternis implies that the three species have had a relatively long foundinthesatyrineLasiommatamegera(E.Weingartner, period of independent evolution, an implication which is unpublished),aswellasinthenymphalinesMelitaeacinxia corroboratedbyouranalysisoftheagesoftheclades.The (Linnaeus, 1758) and Melitaea phoebe (Denis & second taxonomic implication is within the species Schiffermu¨ller, 1775) (N. Wahlberg, unpublished). P. aegeria, for which there is conflicting information However, in the strong flyer Polygonia c-album (Linnaeus, regarding morphological and geneticdata. On the basis of 1758)(Nymphalidae),deepdivergencebetweenEuropeand morphology, there are two clear subspecies found north Morocco could not be found (E. Weingartner, unpub- andsouthoftheAlps,respectively.Basedonthesequences lished).Moretaxaneedtobestudiedinordertoseeifthe oftwogenes,therearetwoclear lineageswhich arehighly patternisconsistent.Ifthispatternoccursinothertaxa,it diverged from each other, one in North Africa and the would suggest that there has been an event that has facili- other in Europe. DNA taxonomy is enjoying great popu- tated movement between North Africa and Europe at one larityatthemoment(see,forexample,Hebertetal.,2003), point in the recent geological history of the region. What although calls are being made not to discard traditional this event could have been remains to be discovered morphological data (see, for example, Wahlberg et al., throughcomparativestudiesofotherspecieswithdistribu- 2005). In this case, there are two COI haplotype lineages tionsinNorthAfricaandEurope. which have apparently not interbred for a long period of time(perhapsaslongas0.8millionyears,althoughtheyare able to produce fertile offspring in the laboratory; Nylin Acknowledgements etal.,1994),andtwomorphologicallineagesthatdointer- breed but form a narrow hybrid zone and retain their WearegratefultoEnriqueGarcı´o-Barros,KarlGotthard, morphological distinctiveness outside the hybrid zone. Bengt Karlsson, Georg Nygren, Thomas Schmitt, This would suggest that there are in fact three taxonomic Constantı´ Stefanescu, Michel Tarrier and Christer #2006TheAuthors Journalcompilation#2006TheRoyalEntomologicalSociety,SystematicEntomology,31,621–632 630 E.Weingartneretal. Wiklund for providing specimens for this work. We thank Emerson, B.C., Oromi, P. & Hewitt, G.M. (2000a) Colonisation AndrewBrower,IanKitchingandananonymousreviewer and diversification of the species Brachyderes rugatus for constructive comments on a previous version of the (Coleoptera) on the Canary Islands: evidence from mitochon- manuscript. drialDNACOIIgenesequences.Evolution,54,911–923. Emerson, B.C., Oromi, P. & Hewitt, G.M. (2000b) Interpreting colonization of the Calathus (Coleoptera: Carabidae) on the Canary Islands and Madeira through the application of the References parametricbootstrap.Evolution,54,2081–2090. Eriksson,T.(2005) BergianskaTra¨dga˚rdenSoftwareURLhttp:// Ancochea,E.,Hernan,F.,Cendrero,A.,Cantagrel,J.M.,Fuster,J.M., www.bergianska.se/index_kontaktaoss_torsten.html [accessed Ibarrola, E. & Coello, J. (1994) Constructive and destructive on24May2005]. episodes in the building of a young oceanic island, La Palma, Felsenstein, J. (1985) Confidence limits on phylogenies: an CanaryIsland,andgenesisoftheCalderadeTaburiente.Journal approachusingthebootstrap.Evolution,39,783–791. ofVolcanologyandGeothermalResearch,60,243–262. Felsenstein, J. (1993) Phylip (Phylogeny Inference Package), Baker, R.H. & DeSalle, R. (1997) Multiple sources of character Version 3.63. Available from author at URL http://evolution. information and the phylogeny of Hawaiian drosophilids. gs.washington.edu/phylip.html[accessedon12May2005]. SystematicBiology,46,654–673. Gatesy,J.,O’Grady,P.&Baker,R.H.(1999)Corroborationamongdata Bozano,G.C.(1999)SatyridaePartI,SubfamilyElymniinae,Tribe setsinsimultaneousanalysis:hiddensupportforphylogeneticrelation- Lethini.OmnesArtes,Milan. shipsamonghigherlevelartiodactyltaxa.Cladistics,15,271–313. Bremer, K. (1988) The limits of amino acid sequence data in Geldmacher, J. & Hoernle, K. (2000) The 72 Ma geochemical angiospermphylogeneticreconstruction.Evolution,42,795–803. evolutionoftheMadeirahotspot(easternNorthAtlantic):recy- Bremer,K.(1994)Branchsupportandtreestability.Cladistics,10, cling of Paleozoic ((cid:5)500 Ma) oceanic lithosphere. Earth and 295–304. PlanetaryScienceLetters,183,73–92. Brower,A.V.Z.(1994)Rapidmorphologicalradiationandconver- Goloboff, P.A. (1998) NONA, Version 2.0. Published by author. gence among races of the butterfly Heliconius erato inferred URLhttp://www.cladistics.com[accessed2001]. frompatternsofmitochondrialDNAevolution.Proceedingsof Gotthard,K.,Nylin,S.&Wiklund,C.(1994)Adaptivevariationin theNationalAcademyofSciencesUSA,91,6491–6495. growthrate:lifehistory costsand consequences inthe speckled Brower, A.V.Z. & DeSalle, R. (1998) Patterns of mitochondrial woodbutterfly,Parargeaegeria.Oecologia,99,281–289. versus nuclear DNA sequence divergence among nymphalid Gotthard,K.,Nylin,S.&Wiklund,C.(2000)Matingopportunity butterflies: the utility of wingless as a source of characters and the evolution of sex-specific mortality rates in a butterfly. forphylogeneticinference.InsectMolecularBiology,7,73–82. Oecologia,122,36–43. Brunton, C.F.A. & Hurst, G.D.D. (1998) Mitochondrial DNA Hall, T.A. (1999) BioEdit: a user-friendly biological sequence phylogeny of brimstone butterflies (genus Gonepteryx) from alignmenteditorandanalysisprogramforWindows95/98/NT. the Canary Islands and Madeira. Biological Journal of the NucleicAcidsSymposiumSeries,41,95–98. LinneanSociety,63,69–79. Hebert, P.D.N., Cywinska, A., Ball, S.L. & deWaard, J.R. (2003) Burban,C.,Petit,R.J.,Carcreff,E.&Jactel,H.(1999)Rangewide BiologicalidentificationsthroughDNAbarcodes.Proceedingsofthe variation of the maritime pine bast scale Matsucoccus feytaudi RoyalSocietyofLondon,SeriesB,BiologicalSciences,270,313–322. Duc. (Homoptera: Matsucoccidae) in relation to the genetic Hewitt, G.M. (1996) Some genetic consequences of ice ages, and structureofitshost.MolecularEcology,8,1593–1602. theirroleindivergenceandspeciation.BiologicalJournalofthe Carranza, S., Arnold, E.N., Mateo, J.A. & Lo´pez-Jurado, L.F. LinneanSociety,58,247–276. (2000) Long-distance colonisation and radiation in gekkonid Higgins, L.G. (1975) The Classification of European Butterflies. lizards,Tarentola(Reptilia:Gekkonidae),revealedbymitochon- WilliamCollinsSons&Co,London. drial DNA sequences. Proceedings of the Royal Society of Hill, J.K., Thomas, C.D. & Blakeley, D.S. (1999) Evolution of London,SeriesB,BiologicalSciences,267,637–649. flightmorphologyinabutterflythathasrecentlyexpandedits Davies,N.B.(1978)Territorialdefenceinthespeckledwoodbut- geographicrange.Oecologia,121,165–170. terfly (Pararge aegeria): the resident always wins. Animal Honey, M.R. & Scoble, M.J. (2001) Linnaeus’s butterflies Behaviour,26,138–147. (Lepidoptera: Papilionoidea and Hesperoidea). Zoological DeJong,H.(1998)Insearchofhistoricalbiogeographicpatterns JournaloftheLinneanSociety,132,277–399. in the western Mediterranean terrestrial fauna. Biological Hundsdoerfer,A.K.,Kitching,I.J.&Wink,M.(2005)Thephylo- JournaloftheLinneanSociety,65,99–164. genyoftheHyleseuphorbiaecomplex(Lepidoptera:Sphingidae): Dobson,M.&Wright,A.(2000)Faunalrelationshipsandzoogeo- molecular evidence from sequence data and ISSR-PCR finger- graphicalaffinitiesofmammalsinnorth-westAfrica.Journalof prints.OrganismsDiversityandEvolution,5,173–198. Biogeography,27,417–424. Janz,N.,Nyblom,K.&Nylin,S.(2001)Evolutionarydynamicsof van Dyck, H. & Matthysen, E. (1998) Thermoregulatory differ- host-plantspecialization:acasestudyofthetribeNymphalini. ences between phenotypes in the speckled wood butterfly: hot Evolution,55,783–796. perchersandcoldpatrollers?Oecologia,114,326–334. Jones, M.J. (1992) The speckled wood butterflies Pararge xiphia van Dyck, H. & Wiklund, C. (2002) Seasonal butterfly design: and P. aegeria (Satyridae) on Madeira: distribution, territorial morphological plasticity among three developmental pathways behaviour and possible competition. Biological Journal of the relative to sex, flight and thermoregulation. Journal of LinneanSociety,46,77–89. EvolutionaryBiology,15,216–225. Jones,M.J.,Lace,L.A.,Harrison,E.C.&Stevens-Wood,B.(1998) vanDyck,H.,Matthysen,E.&Dhondt,A.A.(1997)Theeffectof Territorial behaviour in the speckled wood butterflies Pararge wingcolouronmalebehaviouralstrategiesinthespeckledwood xiphiaandP.aegeriaofMadeira:amechanismforinterspecific butterfly.AnimalBehaviour,53,39–51. competition.Ecography,21,297–305. #2006TheAuthors Journalcompilation#2006TheRoyalEntomologicalSociety,SystematicEntomology,31,621–632

Description:
Speciation in Pararge (Satyrinae: Nymphalidae) butterflies – North Africa the status of P. aegeria as a model species in ecological studies, the evolutionary history of van Dyck, H. & Wiklund, C. (2002) Seasonal butterfly design:.
See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.