Low Genetic Diversity in the Rare Madeiran Endemic Armeria maderensis (Plumbaginaceae)

Abstract

The genetic diversity and possible geographic structure of the Madeiran endemic Armeria maderensis have been assessed with AFLP. Its scarce distribution (less than 3km between the two most distant localities) and restricted
habitat (vertical pastures on the highest elevations of Madeira), at least in part due to grazing by goats, suggest an assessment
of its conservation status. Diversity estimates obtained for A. maderensis were evaluated through comparison with reference values of AFLP diversity for outcrossing plants and, in order to correct
for phylogenetic constraints, with a widespread congener analyzed with the same AFLP markers. Our results reveal that low
levels of genetic diversity and a weak intraspecific genetic structure underlie the restricted distribution of A. maderensis.

Full text

1
2
3
4
5
6
7
8
9
10
11
LowGeneticDiversityintheRareMadeiranEndemicArmeriamaderensis
(Plumbaginaceae)


RosalíaPiñeiro1,JavierFuertesAguilar1,MiguelMenezesdeSequeira2andGonzaloNietoFeliner1

1RealJardínBotánicodeMadrid,CSIC,PlazadeMurillo2,28014Madrid,Spain.
2DepartamentodeBiología,CentrodeEstudosdaMacaronésia,UniversidadedaMadeira,Campusda
Penteada,9000Funchal,Portugal

Correspondingauthor:
12
13
14
15
RosalíaPiñeiro*:E‐mail:pineiro@rjb.csic.es;Phone:+34914203017;Fax:+34
914200157


1

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
LowgeneticdiversityintherareMadeiranendemicArmeriamaderensis
(Plumbaginaceae)

RosalíaPiñeiro1,JavierFuertesAguilar1,MiguelMenezesdeSequeira2andGonzaloNietoFeliner1

AbstractThegeneticdiversityandpossiblegeographicstructureoftheMadeiran
endemicArmeriamaderensishavebeenassessedwithAFLP.Itsscarcedistribution(less
than3kmbetweenthetwomostdistantlocalities)andrestrictedhabitat(vertical
pasturesonthehighestelevationsofMadeira),atleastinpartduetograzingbygoats,
recommendanevaluationofitsconservationstatus.Diversityestimatesobtainedfor
A.maderensiswereevaluatedthroughcomparisonwithreferencevaluesofAFLP
diversityforoutcrossingplantsand,inordertocorrectforphylogeneticconstraints,
withawidespreadcongeneranalyzedwiththesameAFLPmarkers.Ourresultsreveal
thatlowlevelsofgeneticdiversityandaweakintraspecificgeneticstructureunderlie
therestricteddistributionofA.maderensis.

KeywordsAFLP,Bayesianclusteringanalyses,Conservationbiology,Islandendemic

Runningtitle:LowgeneticdiversityinArmeriamaderensis
2

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Introduction

Propensitytoextinctionofislandplantshasbeenwelldocumented.The
reviewsbyPrimack(1998)andWCMC(1992)indicatethatmostspeciesgoneextinct
between1600andthe1990soccurredonislands(Frankhametal.2002).This
susceptibilityseemstobeduetotheparticulardemographicconditionsenduredon
islands,ofteninvolvingsmalleffectivepopulationsizes,isolationandlowmigration
rates,colonizationthroughfoundingeventsand/orhabitatdestructionbyhumansor
introducedanimals(Coblentz1978;RiesebergandSwensen1996;Frankhametal.
2002;Holmgren2002;CowieandHolland2006;CuevasandLeQuesne2006).
Animportantmatterwhenassessingtheconservationstatusofvulnerable
islandplantsistotestwhethertheraredistributionofthespeciesisparalleledbylow
levelsofgeneticdiversity(Frankham1997).Inordertoproperlyassesstheamountof
geneticdiversity,thespeciesunderstudyneedstobecomparedwithvaluesof
reference.However,interspecificcomparisonsmightbebiasedbydifferencesini)the
biologyofthespeciescompared,specificallytheparticularphylogeneticpositionorlife
historytraits,aswellasinii)theanalyticalproceduresacrossstudies,i.e.,sampling
strategy,typeofgeneticmarker,labprotocolanddiversityparametersused(Loveless
andHamrick1984;Felsenstein1985;HamrickandGodt1989;Stoneburneretal.1991;
GitzendannerandSoltis2000;Culleyetal.2002;Nybom2004;Petitetal.2005).
Inplants,severalapproacheshavebeenproposedtoobtainreliable
comparisonsofgeneticdiversityacrossspecies.Intheclassicapproach,theestimate
foraparticularspeciesiscomparedtoaveragediversityvalues,accountingforthetype
ofmarker,thelifehistorytraitscategoryandthepopulationgeneticparameterused
(HamrickandGodt1989:proteinmarkers;Nybom2004:RAPD,AFLPorSSR).However,
thismethoddoesnotcontrolforphylogeneticconstraintsand,therefore,whena
phylogenyisavailable,otherapproacheshavebeenapplied.Thefirstoneisthe
comparisonwiththemostcloselyrelatedwidespreadcongener,assuggestedby
GitzendannerandSoltis(2000),whofoundempiricalsimilarityofpopulationgenetic
parameterswithinrelatedtaxa.Thisapproachhasbeenusedtostudyalargenumber
ofplantendemics,includingtaxarestrictedtoislands(YoungandBrown1996;
Frankham1997;Soltisetal.1997;DoddandHelenurm2002;Ellisetal.2006;Moreira
3

daSilvaetal2007).Anotherpossibilityistoperformphylogeneticallyindependent
contrasts(PIC;Felsenstein1985).Aguinagaldeetal.(2005)usedthelattermethodto
assesschloroplast‐DNA‐basedG
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
sttrendsinEuropeantemperatetreesandshrubsand
foundlittleassociationwithlifehistorytraitswhenphylogeneticpositionwas
accountedfor,thuschallengingtheconclusionsoftheclassicmeta‐analysesbyHamrick
andGodt(1989)andNybom(2004),Petitetal.(2005).
OurstudyfocusesontheislandendemicplantArmeriamaderensisLowe,which
hasanextremelyrestricteddistributioninrockyareasexposedtohumidwindsinthe
mountainsofMadeira(Vieira1992;PressandShort1994;JardimandFrancisco2000).
Madeira,avolcanicislandoriginatedwithintheAfricanplateintheTertiary
(approximately5.3MYA;Geldmacheretal.2000),comprises1226describedvascular
plantspecies,ofwhich234areMacaronesianendemicsand157oftheseendemicsare
exclusivetotheisland(Vieira1992;PressandShort1994;JardimandFrancisco2000).
Armeriamaderensisoccursonlyabove1600m,i.e.,strictlyinthesupratemperatebelt,
spanningasmallareabetweenthehighestpeaks,Ruivo(1862m)andAreeiro(1818m)
(Mesquitaetal.2004).ThisareaispartoftheNaturalParkofMadeira,ageologicaland
highaltitudevegetationreserve,alsopartofaSpecialProtectedArea(SPA)includedin
theNatura2000NetworkasaCommunityImportantSite(CIS).Climaxvegetationin
thisbeltincludesseveralrupicolouscommunities,whichmaydevelopwheretreecover
andgrazingareabsent.OneofthemistheArmeriomaderensis‐Parafestucetum
albidae,dominatedbyArmeriamaderensisLowe,Deschampsiamaderensis(Hack.&
Bornm.)Buschm.,Parafestucaalbida(Lowe)Alexeev,Anthoxanthummaderensis
Teppner,andAnthyllislemmanianaLowe.Moderategrazingofthiscommunityseems
toresultinthereplacementbytheassociationViolarivianae‐agrostietumcastellanae,
dominatedbynon‐endemics(Costaetal.2004;M.Silva,D.Menezes,S.King,E.
MenezesdeSequeira,M.MenezesdeSequeira,unpubl.).Inlocationsexposedtoheavy
grazingandsoilerosion,evenpoorerannualcommunitiesareestablished(Leontodo
longirrostris‐Ornithopetumperpusilli).
Grazingseemstobeanimportantfactordeterminingtherestricteddistribution
ofthespecies.In2003,theregionalgovernmentcompletedtheprogramofremovalof
goatsintheMadeiranMountainsinitiatedin1994intheframeoftheEUHabitats
Directive92/43/EEC.Fromthatmomenton,Armeriamaderensishasexperienceda
4

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
spectacularincreaseofitspopulations,colonizingmoreaccessiblealpinepastures
(Violarivianae‐agrostietumcastellanaeandLeontodolongirrostris‐Ornithopetum
perpusilli)inadditiontotheinaccessiblerockcrevicesthatweretheexclusivehabitat
before.Moreover,ageneralincreaseoffloristicdiversityinMadeiranMountaiscould
beobserved(M.Silva,D.Menezes,S.King,E.MenezesdeSequeira,M.Menezesde
Sequeira,unpubl.).
LikeotherMacaronesianendemics(reviewedinJuanetal.2000;Emerson2002;
Carineetal.2004),A.maderensisappearstohaveitsclosestrelativesintheWestern
Mediterranean.NuclearribosomalITSphylogeniesareconsistentwiththisgeneral
trendandrevealasubstantialdivergence(NietoFelineretal.2001;FuertesAguilarand
NietoFeliner2003).Unluckily,accurateestimationofisolationtimeandaswellas
identificationofitssisterspeciesareprecludedbythelowlevelresolutionoftheITS
tree,duetothefrequentinterspecifichybridizationinnaturalpopulationsofArmeria
andthelikelyrecentdiversificationofthegenus(GutiérrezLarenaetal.2002;Fuertes
AguilarandNietoFeliner2003;Tauleigne‐GomesandLefèbvre2005).
ThisstudyusesAFLPdatatoassessthegeneticstructureanddiversitylevelsof
Armeriamaderensis,tointerpretitsevolutionaryhistoryandtoevaluateits
conservationperspectives.Oursamplingwasperformedin2003.Therefore,itprovides
anaccuratedescriptionofthegeneticstructureofthespeciesjustbeforegoatswere
definitivelyremoved,afteralmost600yearsofgrazing(Sousa2003).
Inordertocorrectforphylogeneticbias,followingGitzendannerandSoltis(2000),
wehaveevaluatedthelevelsofgeneticdiversityinA.maderensisincomparisonwitha
previousAFLPstudyonthewesternMediterraneanwidespreadcongenerA.pungens
(Piñeiroetal.2007).ArmeriapungensisdisjunctlydistributedintheAtlanticandthe
Mediterranean.Ithasamainareaalonga500kmstripeinsouthernAtlanticIberia,
fromthemouthoftheTagusRivertotheGibraltarStrait.Italsooccursontwodisjunct
archipelagos:intheAtlantic,intheCíesislands(offshoreGaliciancoast,northern
Spain),andintheMediterraneaninsouthernCorsicaandnorthernSardinia.Previous
morphologicalandmoleculardatarevealthatA.pungenspresentsadiverseancestral
lineageoccurringintheAtlanticcoastsofSWPortugal(Piñeiroetal.2007;Piñeiroet
al.,unpubl.),whileintrogressionevents(Piñeiroetal.,unpubl.)mayexplainthe
relativelyhighdiversityintheCíesislands.TheremainingpopulationsintheSouthern
5

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
partoftheAtlanticrange(GulfofCadiz)aswellasthedisjunctMediterranean
populationsinCorsicaandSardiniaaretheresultofrecentexpansionsfromthe
originalPortuguesearea.Colonizationeventsmeantthelossofgeneticdiversity,
especiallyintheGulfofCadiz,probablyduetothedifferentenvironmentalconditions
inthisarea.

Specifically,inthisstudythefollowingquestionsareinvestigated:i)Istherestricted
distributionofA.maderensisparalleledbyalowgeneticdiversity?ii)Isgenetic
variationgeographicallystructuredinA.maderensis?iii)Whichmightbethefactors
influencingcurrentgeneticstructureandlevelsofdiversity?iv)Whatarethe
implicationsofthecurrentgeneticstructureanddiversityfortheconservationofthe
species?

MaterialandMethods

StudySystem

Armeriamaderensisistheonlyendemicrepresentativeofthegenusin
Macaronesia.Morphologically,themostuniquefeatureofthisspecieswithinthe
genusisthelackofimbricationofinvolucrebracts,whicharefew,narrowand
seeminglyarrangedinasinglerowsothatitcanbehardlysaidthattheyconstitutean
involucre.Theinsertionoftheflowerpedicelonthecalyxisfrequentlymoretruncate
andthuslessspurredthaninothercongeners.Anotherapparentfeatureisthepatent
arrangementoftheinnerflowerineachspikeletwithintheglomerule.These
morphologicalcharactersareusefultoidentifyA.maderensisbuttotheextentthat
theyarenotsharedwithotherspecies,theydonothelpinfindingitsclosestrelatives.

SamplingStrategy,DNAisolation,AFLPprotocol

Thesamplingwasperformedin2003inthesmalldistributionareabetween
peaksRuivoandAreeirobeforegoatswereremovedfromtheisland(Table1,Fig.1).
Toourknowledge,nopreviousfieldstudieshadattemptedtolocalizeorquantifythe
6

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
populationsofA.maderensis.Ninesites(nr.1−9)werefound,themostdistantones
beinglessthan3kmapart.Wesucceededatcollectingfruitsfromeightofthesesites
(Table1).Anadditionalinaccessiblesite,harbouringafewindividualsonavertical
north‐facingwall(closetositenr.8,onthepathtoTorrinhaspeak),wasalsodetected.
Sites2,4,5and6correspondedtoconsistentpopulations,whereassites1,3,7
and8consistedofveryfewscatteredindividuals(Table1,Fig.1).Theremoteand
inaccessiblestatusofpopulationslimitedthenumberofindividualscollectedateach
site(e.g.populationnr.4hadtobesampledbyequippedclimbers).Intotal,wewere
abletosampleripefruitsfrom44separatemotherplants.
Threereplicatespermotherplantweregerminatedafteracoldtreatmentof
onemonth.SeedlingswerecultivatedinthegreenhouseattheBotanicalGardenof
MadridbetweenOctober2003andJune2005inordertoprovidefreshleavesforDNA
isolation.Thisisalimitationofourstudy,becausewehaveactuallyassessedthe
geneticdiversityinapotentialfuturegeneration.
DNAwasextractedwithPlantDNeasyMinikit(Qiagen).TheAFLPprotocolwas
performedaccordingtoGaudeuletal.(2000)forEcoRI/MseIenzymecombinationand
Piñeiroetal.(2007)forKpnI/MseI.Thefollowingthreeprimercombinationswere
used:(6‐FAM)EcoRI+acc/MseI+cacc,(6‐FAM)EcoRI+acg/MseI+ctac,(6‐FAM)KpnI+
atc/MseI+cag.Protocolsandselectiveprimerswerethesameusedinthe
phylogeographicstudyofthecongenerA.pungens(Piñeiroetal.2007).Thefactthat
AFLPmarkershavebeenobtainedusingtwodifferentenzymecombinationsincreased
thegenomicregionsrepresentedinthefingerprints(Vosetal1995;Ulrichetal.1999).
Intotal,31individualsweresuccessfullygenotyped.Theremainingindividuals
collecteddidnotsurviveinthegreenhouseorfailedforamplificationwithatleastone
primercombinations.Samplingofmorethanoneindividualfromthesamemother
plantwasavoided.Thesampling,althoughsmall,mightappropriatelyrepresentthe
geneticvariationofA.maderensisgivenitsextremelyrestricteddistribution.Voucher
specimensarekeptatMA.
Areproducibilitytestwasperformedforeachprimerpairbyre‐extractingDNA
andrepeatingthewholeprocedure(7,6and5individualswerere‐amplifiedusingthe
threeprimercombinationsgivenabove,respectively).Theerrorratewascalculatedas
thetotalnumberoflocidifferencesrelativetothetotalnumberoflocicomparisons
7

1
2
3
4
5
6
7
andsubsequentlyaveragedoverthethreecombinations.Tofurtherassessand
eventuallyrefinethequalityoftheAFLPdata,potentiallynon‐homologousbandswere
checkedfollowingfourdifferentcriteria:i)slightsizedifferencesamongputative
homologousbandsacrossindividuals;ii)lowintensitybands;iii)changingintensityof
onebandacrosssamples;andiv)bandsofhigh(upper10%and20%)orsmall(lower
10%and20%)molecularweight(Bagleyetal.2001;Boninetal.2004).Onceidentified
thebandsfallingintoanyofthosecategories,theerrorrateimprovementwas
calculatedafterremovingbandsfromeachcategory(Piñeiroetal.2007).8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31

GeneticStructureofArmeriamaderensis

BayesianclusteringanalyseswereperformedwithSTRUCTURE2.1(Pritchardet
al.2000;Falushetal.2007)andBAPS3.2(Coranderetal.2006).WithSTRUCTUREwe
usedrunlengthsof106followingaburninof105.Longburninperiodswerenecessary
toreachconvergenceofthealphastatisticbeforetheendoftheburninphase.Each
phenotypewascodedbyasinglealleleandamissingdatumaccordingtothe
indicationsinthemanual(1‐missingfordominantmarkersand2‐missingforrecessive).
Weselectedthemodelofcorrelatedallelefrequencies,appropriateincaseswhere
weakgeneticstructureisexpected,andtestedboththeno‐admixture(tenrepetitions
ateachK)andtheadmixture(fiverepetitions)ancestrymodels.SimulationsfromK=1
toK=8wererun,i.e.themaximumnumberofgeneticgroupstestedequalledthe
numberofgeographicalsites.Thenumberofgeneticgroups(K)inourdatasetwas
inferredtakingintoaccounttheestimatedposteriorlogprobabilityofthedata,LK),as
wellasthestabilityoftheassignmentpatternsofindividualsintoKgroupsacross
repetitions.BAPSsimulationswererunfromK=1toK=8asthemaximumnumberof
groups,withfourreplicatesateachK.
Dicesimilarityamongindividualswascalculatedandvisualizedwithaprincipal
coordinatesanalysis(PCoA;NTSYSpc2.1;Rohlf1998).Inaddition,Jaccardandsimple
matchingcoefficientswerecalculated.Aminimumspanningtree(MST)basedonDice
wasimposedonthePCoAtodetectlocaldistortions.NeiandLidistance(1979)was
alsocalculatedwithPAUP4.0b10(Swofford2002).CorrelationbetweenNeiandLi
8

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
distancesandgeographicdistanceswascalculatedbyperformingaManteltestwith
NTSYSpc2.1.Significancewastestedbyrandomization(1000permutations).

GeneticDiversityofArmeriamaderensis

Threegeneticdiversityestimateswerecomputedatthetotalspeciesleveland
atthewithin‐populationslevel:i)allelicrichness,fromthepercentageofpolymorphic
loci,P(POPGENE3.2;YehandBoyle1997);ii)genediversityofNei(1973),H
(POPGENE3.2);andiii)allelesimilarityusingShannonindex(1948),Sh(POPGENE3.2).
TocalculatetheglobalgeneticdiversitywithinA.maderensis,allgenotyped
individualswerepooled.Incontrast,calculationsonaper‐populationbasiswere
challengedbythefactthatonlythreeoutofeightcollectedsitesarewell‐separated
populations(pops.nr.2,5and6).Therefore,within‐populationgeneticdiversitywas
reportedastheaverageoftheestimatesforthesethreesites.
Standarddeviationsofdiversityestimatesarereported.Monomorphicand
polymorphiclociwereincludedinthecalculations.Nei’s&Shannon’smeasureswere
calculatedforeachlocusandaveragedoverloci.Sincetheconceptofheterozygosity
cannotbeappliedtodominantmarkers,averageNei’sgenediversity(H)issimplya
measureofgeneticvariation.WealsocalculatedNei’sdiversityestimateusingLynch&
Milligan’smethod(1994)implementedwithTFPGA(Miller1997),whichattemptsto
correctthebiasgeneratedbydominantmarkersbypruningfrequentlociforthe
estimationofallelefrequencies.

ComparisonwiththeWidespreadCongenerArmeriapungens

ForthewesternMediterraneanwidespreadcongenerA.pungens,thesame
threegeneticdiversityestimateswerecalculatedasdescribedabove.Thecomparison
ofthegeneticdiversityvaluesofA.maderensiswithA.pungens,inadditiontothetype
ofmarker,lifehistorytraitscategoryandpopulationgeneticparameters,correctsfor
phylogeneticbiasandhomogenizestheAFLPprotocolinbothspecies,includingthe
selectiveprimersused.
9

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
Still,theAFLPstudyonA.pungensinvolved221individualsfrom23welldefined
populationsspanningarangeofhundredsofkilometres,whichimpliesanimportant
differencewiththesamplingstrategyandspatialscaleofthecurrentassessmentofA.
maderensis(Piñeiroetal.2007).Inordertoaccountforsuchdifferences,aper‐
populationcomparisonofA.pungensandA.maderensiswouldbedesirable.However,
thescattereddistributionofA.maderensisindividualschallengesthereliabilityof
within‐populationestimates.Wehavethusmadeacomparisononthebasisofthe
geneticlineagesdetectedwithinA.maderensisusingBayesianandgeneticdistance
methodswiththosepreviouslyfoundinthewidespreadA.pungens(Piñeiroetal.
2007):Cíesislands,Portugal,Vicente‐Bordeira,GulfofCadizandCorsica‐Sardinia(hear
calledI,II,III,IV,VI,respectively).Anadditionallineage(lineageV,Camarinal
population)wasnotincluded,givenitslowsamplesizeandhybridorigin(Piñeiroetal.
2007;Piñeiroetal.,unpubl.).
Forthecomparisonacrossintraspecificgeneticlineages,independent
presence/absencematriceswereeditedfromthecompleteAFLPmatricesforA.
maderensis(presentstudy)andforeachlineageofA.pungens(fromtheoriginaldata
inPiñeiroetal.2007).Foreachgeneticlineage,twomatriceswereedited,oneatthe
populationandanotheratthetotalgeneticlineagelevel,anddiversityparameters
wererecalculatedfromthem.Accordingtothelowestsamplesizes,thenumberof
individualswasstandardizedton=6atthepopulationlevelandton=15atthetotal
geneticlineagelevel.Thiswasachievedbyrandomexclusionofindividuals.Lociabsent
inallindividualsofonelineagewereremovedtoavoidunderestimationofthegenetic
variability.

ComparisonwithReferenceValuesofAFLPDiversityinPlants

Nei’sunbiasedgenediversity(1978)wasalsocalculatedforA.maderensiswith
TFPGA(Miller1997)atthespecieslevel(Ht)andwithinpopulations(Hs).Thisestimate
waschosensinceitwastheoneusedbyNybom(2004)forcomparisonofwithin‐
populationparametersacrossstudiesbasedondominantmarkers.Itcorrectsfor
differentsamplingsizesbymultiplyingtheindexper2n/2n‐1,wherenisthesample
size(Nei1987:equation8.4).Thebiasiseffectivelyreducedforsamplingsizes<50and
10

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
largenumberoflociavailable,whichfitsoursamplingstrategy.FollowingNybom’s
approach,unbiasedNei’sdiversityisherereportedonlyforpolymorphicbands.
ThecomparisonofNei’sunbiasedgenediversityvaluesforA.maderensiswith
thereferencevaluesreportedinNybom(2004)basedondominantmarkersfor
perennialandoutcrossingplantsaccountsforthetypeofmarker,lifehistorytraitsand
populationgeneticsparameters.

Results

AFLPProfiles

Byrecalculatingtheerrorrateafterremovingeachofthecategoriesof
potentiallynotreproduciblebands,thosebandschangingintensityacrosssamples
wereobservedtobetheleastreliable(Table2).Discardingthemmeantadecreaseof
theerrorratebelow5%(3.1%).Accordingly,67unreliablebandswerediscarded.An
additionalunreliableband,nonreproducibleinmostcomparisons,waseliminated.
Subsequently,16individualsthatwerenotamplifiedforoneoftheprimer
combinationswereremoved,whichmeantthelossof32bands.Afinaldatasetof31
individualsand90markerswasretainedforanalysis.Markersspannedfrom56bpto
447bpandonly58(64.44%)werepolymorphic.Noidenticalphenotypesamong
individualsweredetected.

GeneticStructure

BAPSandSTRUCTUREinferredasingleBayesianclustercomprisingallsampled
individualsofA.maderensis,revealingtheweakgeneticstructureintheAFLPdata(Fig.
2a).BAPSfoundtheoptimalpartitionatK=1(resultsnotshown).IntheSTRUCTURE
analysis,individualswereevenlyassignedtotheKgroupsinsimulationsfromK=2to
K=8forbothadmixtureandno‐admixturemodels(Fig.2a).AsstatedintheSTRUCTURE
manual,thissituation,whentheproportionofthesampleassignedtoeachpopulation
isroughlysymmetric,isindicativeofnopopulationstructure.Consistently,L(K)didnot
showamaximumvalueatanyspecificK.Instead,runsatK=1,K=2,K=6,K=7andK=8
11

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
yieldedverysimilarposteriorprobabilitiesbetweenL(K)=‐589.3andL(K)=‐586.7(Fig.
2b).
Incontrast,thePCoArepresentingDicesimilaritiesamongindividuals(Fig.3)
showedsomedegreeofgeneticstructureatlocalscales,sincegenotypesspecificto
differentpopulationscouldbedistinguished.Nonetheless,theabsenceof
discontinuitiesinthePCoAscatterplotconfirmedthelowlevelofgeneticdivergence
andweakstructuringoftheoverallgeneticvariationwithinA.maderensis,aspointed
byBayesianclusteringanalyses.Jaccardandsimplematchingcoefficientsgavealmost
identicalresultsasDice(r=0.99withJaccardandr=0.98withsimplematching).The
Manteltest(r=‐0.01608;PrandomZ<observedZ=0.6134)corroboratedthelackof
overalllinearcorrelationbetweengeneticandgeographicdistances.

GeneticDiversity

WhenthetotaldiversityestimatesofA.maderensisarecomparedwitheachof
thegeneticlineagesofthewidespreadcongenerA.pungens,diversityofA.maderensis
resultedtobelowerthanthatinthelineagewiththelowestdiversityofA.pungens
(lineageIV;Table3).Thispatternholdsforthepercentageofpolymorphisms,Nei´s
genediversity(Ht;Nei1973)andShannonindex(Sh).TheremaininglineagesofA.
pungensshowedsignificantlyhigherdiversitylevelsforallthreediversityparameters,
especiallylineagesI,IIandIII,followedbyintermediatelevelsofdiversityinlineageVI.
Ineverycase,correctionofNei’sindexfordominantmarkersusingestimationofallele
frequenciesbyLynchandMilligan’s(1994)(resultsnotshown)methodgavealmost
identicalresultsasthosebasedonthesquaredrootofthefrequencyoftherecessive
allele(Nei1973).
TheconclusionofalowerdiversityofA.maderensisascomparedtolineages
withinA.pungens,drawnfromthetotaldiversitymeasures,holdformeasuresona
per‐populationbasis,asshowninTable4.Forcomparison,weestimatedthediversity
levelsofthedifferentlineagesofA.pungensincludingalsothosebandsthatwere
absentinallindividualsofonelineage.Inthiscase,thediversitylevelsofthedifferent
lineagesofA.pungensslightlydecreasedbutwerestillhigherthaninA.maderensis
(resultsnotshown).
12

Finally,theunbiasedwithin‐populationgenediversityofNei(1978)forA.
maderensisaveragedoverpopulations2,5and6wasH
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
s=0.19.Poolingallgenotyped
individuals,itwasHt=0.14.

Discussion

IsArmeriamaderensisGeneticallyDepauperate?

Inherreviewofpopulationgeneticsstudiesbasedondominantmarkers,
Nybom(2004)reportedanaveragewithin‐populationunbiasedgenediversityofNei
(1978)ofHs=0.27foroutcrossingplantsandofHs=0.25forperennials.Thelowerand
upperlimitsofAFLPdiversityinplantswereHs=0.15andHs=0.31,respectively(average
Hs=0.23).Inthiscontext,A.maderensis,withHs=0.19atthepopulationleveland
Ht=0.14atthetotalspecieslevel,exhibitssignificantlylowlevelsofgeneticdiversity.
Nonetheless,thiscomparisonshouldbetakenwithcautionforthreereasons.First,
althoughlifehistorytraitsareconsidered,phylogeneticconstraintsarenotcontrolled
for.Second,Nybom´sestimatesaregiveninaper‐populationbasis,whereasthisis
difficulttoobtainforA.maderensisgivenitsparticulargeographicaldistribution.Third,
Nybom´scalculationswerebasedonpolymorphicloci,butconsideringonly
polymorphicmarkersinaplantlikeA.maderensis,withahighpercentageof
monomorphicloci(52.23%,seeTable3),probablyleadstoanoverestimationofthe
geneticdiversity.
ThecomparisonwiththewidespreadcongenerA.pungensismorereliable
becauseitiscorrectedforbothbiological(lifehistorytraitsandphylogeny)and
methodologicalconstraintsandbecausethephylogeographichistoryofA.pungensis
wellknownbasedonmorphologicalandmoleculardata(Piñeiroetal.2007;Piñeiroet
al.,unpubl.).OurgeneticdiversityestimatesforA.maderensisareshowntobe,bothat
thetotalgeneticlineagelevel(Table3)andatthepopulationlevel(Table4),even
lowerthanfortheextremelyimpoverishedlineageIVofA.pungensthathasrecently
colonizedtheGulfofCadiz.Still,thesecomparisonswithA.pungensmustbealso
consideredcautiously.Ontheonehand,thecomparisononthebasisofgenetic
lineagesdoesnotfullyaccountforthedifferentspatialscalesofA.pungensandA.
13

maderensis,stilllargerinA.pungenslineages(maximaldistanceswithinlineagesfrom
21kmto413km)thatinA.maderensis(maximaldistanceabout3km).Ontheother
hand,theaveragewithin‐populationdiversityforA.maderensisisonlybasedon
calculationsforthreesites.Anotherpossibility,istoconsiderthefactthatdiversity
estimatesforthewholerangeofA.maderensis(H
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
t=0.08,Sh=0.14,n=15;Table3)are
belowthelowestwithin‐populationestimatesofA.pungens(Hs=0.10−0.19,
Sh=0.15−0.27,n=6;Table4).Thisassessment,atcomparablespatialscales,definitively
confirmsthereducedlevelofgeneticdiversitywithinA.pungensbeyondany
reasonabledoubt.
Besidestheinformationprovidedbythepopulationgeneticestimates,the
infraspecificgeneticstructurewithinA.maderensisisweakandshowntoberestricted
tolocalscales.

ThreatsforArmeriamaderensisandFutureProspects

WeconsideredthepossibilitythatthereducedgeneticdiversityinA.
maderensisisactuallyreflectingashifttoautogamyduringestablishmentafterlong‐
distancedispersal.Thebreakdownofself‐incompatibilityisobservedinArmeriain
circumpolarareas.Ithasbeenhypothesizedthatthismechanismfavours
establishmentafterlongdistancedispersalorinareaspoorinpollinators(Baker1966).
However,twodifferentpollen‐stigmamorphsthatareassociatedtotheincompatibility
systeminArmeriahavebeendetectedinA.maderensisplants(NietoFeliner,unpubl.),
suggestingthatincompatibilityhasbeenmaintainedaftercolonisation.Mechanisms
favouringoutcrossinghavebeenreportedforotherMacaronesianspecies(Francisco‐
Ortegaetal.2000).
Therefore,oncediscardedtheinfluenceofthebreedingsystem,thegenetic
impoverishmentofA.maderensisseemstoresultfromhistoricalevents.Thesignificant
increaseofthedistributionofA.maderensisintoaccessiblehorizontalpasturessince
theremovalofgoatsin2003,pointsatgrazingasoneofthefactorsresponsibleforthe
reducedgeneticdiversityofthespecies.IsolatedevolutioninMadeiraprecludinggene
flow,long‐termsmallpopulationsizes(Frankhametal.2002),foundereffectsduring
thecolonizationofMadeirabymainlandancestors(Winkworthetal.1999;Charbonnel
14

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
etal.2002;CowieandHolland2006)ordestructiveeffectofvolcaniceruptionsmight
alsobeconsidered.
Althoughnofragmentationoftherangehastakenplace,theobserved
levelofgeneticdiversitymayincreasetheriskofinbreedingdepression,
whichcouldreducesurvivalandfecundityintheshortterm.However,to
assessaccuratelytheamountofinbreeding,co‐dominantmarkerswouldbe
required.Long‐termadaptationtoenvironmentalchangesmightbealso
compromized(Frankhametal.2002;PetersonandMcCracken2005).To
preventtheserisks,werecommendmonitoringthepopulationsofA.
maderensisinthefollowingyears,tosurveyitsrecoveryintheabsenceof
goats.FuturegeneticassessmentsofA.maderensismaybeperformedand
thepresentgeneticstudyusedasareferenceofthediversitylevels
immediatelybeforeeradicationofgrazing.Iftheevolutionofthepopulations
wasobservedtobenonappropriate,takingintoaccounttheobservedlackof
geneticstructure,areinforcementofpopulationscouldbedesignedinorder
toreducethedangerofoutbreedingdepression.Germinationratesofseeds
areusuallyhighinArmeria(WoodellandDale1993),andaremaximizedwhen
seedsfollowacoldtreatment(personalobservation).
AsidefromthepracticalapplicationofourgeneticstudytotheconservationofA.
maderensis,itprovidestheopportunitytoaddressinthenearfuturethetheoretical
questionofhowthegeneticvariationparallelstherecoveryinnumberofindividuals
afterapopulationbottleneck.Thismightcomplementtheavailablereportsofincrease
ofthecoverofvegetationorspecificrichnessfollowingtheremovalofherbivores
(Mueller‐DomboisandSpatz1975;Coblentz1977;Schofield1989;LorvelecandPascal
2005).

AcknowledgementsWethankAndreaCostaforprovidingplantmaterialandfield
support,andtheNaturalParkofMadeiraforcollectingpermitsandofferofequipped
climberstoassistfieldwork.PilarCatalán,JuliCuajapéandJoséMaríaIriondomade
valuablecommentsonthemanuscript.ThisworkwassupportedbytheSpanish
DirecciónGeneraldeEnseñanzaSuperioreInvestigaciónCientífica(grantBOS2001–
1839).
15

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
Thisresearchcomplieswiththecurrentlawsofthecountriesinwhichitwas
performed.

References

AguinagaldeI,HampeA,MohantyA,MartinJP,DuminilJ,PetitRJ(2005)Effectsoflife‐
historytraitsandspeciesdistributionongeneticstructureatmaternally
inheritedmarkersinEuropeantreesandshrubs.JBiogeogr32:329–339
BagleyMJ,AndersonSL,MayB(2001)Choiceofmethodologyforassessinggenetic
impactsofenvironmentalstressors:PolymorphismandreproducibilityofRAPD
andAFLPfingerprints.Ecotoxicology10:239–244
BakerHG(1966)Theevolution,functioningandbreakdownofheteromorphic
incompatibilitysystems,I.ThePlumbaginaceae.Evolution20:349–368
BoninA,BellemainE,BronkenEidesenP,PompanonF,BrochmannC,TaberletP(2004)
Howtotrackandassessgenotypingerrorsinpopulationgeneticsstudies.Molec
Ecol13:3261–3273
CarineMA,RussellSJ,Santos‐GuerraA,Francisco‐OrtegaJ(2004)Relationshipsofthe
MacaronesianandMediterraneanfloras:molecularevidenceformultiple
colonizationsintoMacaronesiaandback‐colonizationofthecontinentin
Convolvulus(Convolvulaceae).AmerJBot91:1070–1085
CharbonnelN,AngersB,RasatavonjizayR,BrémondP,DebainC,JarneP(2002)The
influenceofmatingsystem,demography,parasitesandcolonizationonthe
populationstructureofBiomphalariapfeifferiinMadagascar.MolecEcol
11:2213–2228
CoblentzBE(1977)SomerangerelationshipsofferalgoatsonSantaCatalinaIsland,
California.JRangeManagem30:415–419
CoblentzBE(1978)Theeffectsofferalgoats(Caprahircus)onislandecosystems.Biol
Conservation13:279–286
CoranderJ,MarttinenP,MäntyniemS(2006)Bayesianidentificationofstockmixtures
frommolecularmarkerdata.FishBull104:550–558
16

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
CostaJC,CapeloJ,JardimR,SequeiraM(2004)Catálogosintaxonómicoeflorísticodas
comunidadesvegetaisdaMadeiraePortoSanto.InCapeloJ(ed)Apaisagem
vegetaldaIlhadaMadeira.Quercetea6:61–183
CowieRH,HollandBS(2006)Dispersalisfundamentaltobiogeographyandthe
evolutionofbiodiversityonoceanicislands.JBiogeogr33:193–198
CulleyTM,WallaceLE,Gengler‐NowakKM,CrawfordDJ(2002)Acomparisonoftwo
methodsofcalculatingGST,ageneticmeasureofpopulationdifferentiation.
AmerJBot89:460–465
CuevasJG,LeQuesneC(2006)Lowvegetationrecoveryaftershort‐termcattle
exclusiononRobinsonCrusoeIsland.PlEcol183:105–124
DoddSC,HelenurmK(2002)GeneticdiversityinDelphiniumvariegatum
(Ranunculaceae)acomparisonoftwoinsularendemicsubspeciesandtheir
widespreadmainlandrelative.AmerJBot89:613–622
EllisJR,PashleyCH,BurkeJM,McCauleyDE(2006)Highgeneticdiversityinarareand
endangeredsunflowerascomparedtoacommoncongener.MolecEcol
15:2345–2355
EmersonBC(2002)Evolutiononoceanicislands:molecularphylogeneticapproachesto
understandingpatternandprocess.MolecEcol11:951–966
FalushD,StephensM,PritchardJK(2003)Inferenceofpopulationstructureusing
multilocusgenotypedata:Linkedlociandcorrelatedallelefrequencies.
Genetics164:1567–1587
22 FalushD,StephensM,PritchardJ.K(2007)Inferenceofpopulationstructureusing
23
24
25
26
27
28
29
30
31
32
multilocusgenotypedata:Dominantmarkersandnullalleles.MolecEcolNotes
7:574–578
FelsensteinJ(1985)Phylogenyandthecomparativemethod.AmerNaturalist125:1–15
Francisco‐OrtegaJ,Santos‐GuerraA,KimS‐C,CrawfordDJ(2000)Plantgenetic
diversityintheCanaryIslands:aconservationperspective.AmerJBot87:909–
919
FrankhamR(1997)Doislandpopulationshavelessgeneticvariationthanmainland
populations?Heredity78:311–327
FrankhamJ,BallouJD,BriscoeDA(2002)Introductiontoconservationgenetics.
CambridgeUniversityPress,Cambridge
17

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
FuertesAguilarJ,NietoFelinerG(2003)Additivepolymorphismsandreticulationinan
ITSphylogenyofthrifts(Armeria,Plumbaginaceae).MolecPhylogenEvol
28:430–447
GaudeulM,TaberletP,Till‐BottraudI(2000)Geneticdiversityinanendangeredalpine
plant,EryngiumalpinumL.(Apiaceae),inferredfromamplifiedfragmentlength
polymorphismmarkers.MolecEcol9:1625–1637
GeldmacherJ,vandenBogaardP,HoernleK,SchminckeH(2000)The40Ar/39Arage
datingoftheMadeiraArchipelagoandhotspottrack(easternNorthAtlantic).
GeochemGeophysGeosyst1:1008,doi:10.1029/1999GC000018
GitzendannerMA,SoltisPS(2000)Patternsofgeneticvariationinrareandwidespread
plantcongeners.AmerJBot87:783–792
GutiérrezLarenaB,FuertesAguilarJ,NietoFelinerG(2002)Glacial‐inducedaltitudinal
migrationsinArmeria(Plumbaginaceae)inferredfrompatternsofchloroplast
DNAhaplotypesharing.MolecEcol11:1965–1974
HamrickJ,GodtM(1989)Allozymediversityinplantspecies.InBrownA,CleggM,
KahlerA,WeirB(eds)Plantpopulationgenetics,breedingandgenetic
resources.Sinauer,SunderlandMA,pp43–63
HolmgrenM(2002)Exoticherbivoresasdriversofplantinvasionandswitchto
ecosystemalternativestates.BiolInvasions4:25–33
HolsingerKE,LewisPO,DeyDK(2002)ABayesianapproachtoinferringpopulation
structurefromdominantmarkers.MolecEcol11:1157–1164
HolsingerKE,LewisPO(2005)Hickory:Apackageforanalysisofpopulationgenetic
datav1.0.4.Storrs,UniversityofConnecticut,CT
JardimR,FranciscoD(2000)FloraendémicadaMadeira.MúchiaPublicaçôes,Funchal
JuanC,EmersonBC,OromíP,HewittGM(2000)Colonizationanddiversification:
towardsaphylogeographicsynthesisfortheCanaryIslands.TrendsEcolEvol
15:104–108
LorvelecO,PascalM(2005)Frenchattemptstoeradicatenon‐indigenousmammals
andtheirconsequencesfornativebiota.BiolInvasions7:135–140
LynchM,MilliganB(1994)AnalysisofpopulationgeneticstructurewithRAPDmarkers.
MolecEcol3:91–99
18

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
MesquitaS,CapeloJ,SousaJ(2004)BioclimatologiadaIlhadaMadeira:abordagem
numérica.InCapeloJ(ed)ApaisagemvegetaldaIlhadaMadeira.Quercetea
6:61–183
MillerMP(1997)Toolsforpopulationgeneticanalyses(TFPGA).AWindowsprogram
fortheanalysisoftheallozymesandmoleculargeneticdata.Departmentof
BiologicalScience,NorthernArizonaUniversity,Flagstaff
MoreiradaSilvaR,WilsonFernandesG,BernadeteLovatoM(2007)Geneticvariation
intwoChamaecristaspecies(Leguminosae),oneendangeredandnarrowly
distributedandanotherwidespreadintheSerradoEspinhaço,Brazil.CanadJ
Bot85:629–636
Mueller‐DomboisD,SpatzG(1975)Theinfluenceofferalgoatsonlowlandvegetation
inHawaiiVolcanoesNationalPark.Phytocoenologica3:1–29
NeiM(1973)Analysisofgenediversityinsubdividedpopulations.ProcNatlAcadUSA
70:3321–3323
NeiM(1978)Estimationofaverageheterozygosityandgeneticdistancefromasmall
numberofindividuals.Genetics89:583–590
NeiM(1987)Molecularevolutionarygenetics.ColumbiaUniversityPress,NewYork
NeiM,LiWH(1979)Mathematicalmodelforstudyinggeneticvariationintermsof
restrictionendonucleases.ProcNatlAcadUSA76:5269–5273
NietoFelinerG,FuertesAguilarJ,RossellóJA(2001)Canextensivereticulationand
concertedevolutionresultinacladisticallystructuredmoleculardataset?
Cladistics17:301–312
NybomH(2004)ComparisonofdifferentnuclearDNAmarkersforestimating
intraspecificgeneticdiversityinplants.MolecEcol13:1143–1155
PetersonPM,McCrackenCL(2005)Geneticconsequencesofreduceddiversity:
heterozytosityloss,inbreedingdepressionandeffectivepopulationsize.In
KrupnickGA,KressWJ(eds)Plantconservation.Anaturalhistoryapproach.The
UniversityofChicagoPress,Chicago
PetitRJ,DuminilJ,FineschiS,HampeA,SalviniD,VendraminGG(2005)Comparative
organizationofchloroplast,mitochondrialandnucleardiversityinplant
populations.MolecEcol14:689–701
19

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
PiñeiroR,FuertesAguilarJ,DraperMuntD,NietoFelinerG(2007)Ecologymatters:
Atlantic‐Mediterraneandisjunctionofthesand‐duneshrubArmeriapungens
(Plumbaginaceae).MolecEcol16:2155–2171
PressJR,ShortMJ(1994)FloraofMadeira.TheNaturalHistoryMuseum,London
PrimackRB(1998)Essentialsofconservationbiology.SinauerAssociates,Sunderland
PritchardJK,StephensM,DonnellyP(2000)Inferenceofpopulationstructureusing
multilocusgenotypedata.Genetics155:945–959
RiesebergLH,SwensenS(1996)Conservationgeneticsofendangeredislandplants.In
AviseJ,HamrickJL(eds)Conservationgenetics.Casehistoriesfromnature.
Chapman&Hall,NewYork,pp305–327
RohlfFJ(1998)NTSYS‐PCnumericaltaxonomyandmultivariateanalysissystem.Exeter
Publications,Setauket
SchofieldEK(1989)Effectsofintroducedplantsandanimalsonislandvegetation:
examplesfromGalápagosArchipelago.ConservationBiol3:227–239
ShannonCE(1948)Amathematicaltheoryofcommunication.BellSystTechJ27:379–
423,623–656(BellSystemTechnicalJournal)
SoltisPS,SoltisDE,NorvellTL(1997)Geneticdiversityinrareandwidespreadspecies
ofLomatium(Apiaceae).Madroño44:59–73
SousaP(2003)50anosaservirafloresta.GovernoRegionaldaMadeira.Secretaria
RegionaldoAmbienteeRecursosNaturais,DirecçãoRegionaldeFlorestas,
Funchal
SwoffordD(2002)Phylogeneticanalysisusingparsimony(andothermethods),Version
4.SinauerAssociates,Sunderland
Tauleigne‐GomesC,LefèbvreC(2005)Naturalhybridizationbetweentwocoastal
endemicspeciesofArmeria(Plumbaginaceae)fromPortugal.1.Populationalin
situinvestigations.PlSystEvol250:215–230
UlrichG,MuellerL,WolfenbargerLR(1999)AFLPgenotypingandfingerprinting.Trends
EcolEvol14:389–394
VieiraR(1992)FloradaMadeira:ointeressedasplantasendémicasmacaronésicas.
ServiçoNacionaldeParques,ReservaseConservaçâodaNatureza,Lisboa
WCMC–WorldConservationMonitoringCentre(1992)Globalbiodiversity:statusof
theearth'slivingresources,ChapmanandHall,London
20

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
VosP,HogersR,BleekerM,ReijansM,vandeLeeT,HornesM,FritersA,PotJ,
PalemanJ,KuiperM,ZabeauM(1995)AFLP:anewtechniqueforDNA
fingerprinting.NuclAcidsRes23:4407–4414
WinkworthRC,RobertsonAW,EhrendorferF,LockhartPJ(1999)Theimportanceof
dispersalandrecentspeciationinthefloraofNewZealand.JBiogeogr
26:1323–1325
WolfeAD,ListonA(1998)Contributionsofthepolymerasechainreactiontoplant
systematics.InSoltisD.E,SoltisP.S,DoyleJ.J(eds)Molecularsystematicsof
plantsII.DNAsequencing.KluwerAcademicPublishers,NewYork,pp43–86
WoodellSRJ,DaleA(1993)Armeriamaritima(Mill.)Willd.(StaticearmeriaL.;S.
maritimaMill.).JEcol81:573–588
YehFC,BoyleTJB(1997)Populationgeneticanalysisofcodominantanddominant
markersandquantitativetraits.BelgJBot129:157
YoungAB,BrownAHD(1996)Comparativepopulationgeneticstructureofthe
rarewoodlandshrubDaviesiasuaveolensanditscommoncongenerD.
mimosoides.ConservationBiol10:1220–1228


21

1
2 Table1DetailsofsamplesitesofArmeriamaderensisfortheAFLPstudy(allinMadeira;seealsoFig.1)

Site
nr.
Site,habitat(altitude)Latitude

Longituden
collected/
genotyped
1PicoAreeiro,stonypasturesinthesummit(1779m)

32º44’16.6’’N

16º55’42.5’’W2/2
2PicoAreeiro,rockcrevices(1668m)32º44’21.5’’N

16º55’55.3’’W

12/6
3FromPicoAreeirotoPicoRuivo,rockcrevices(1753m)32º44’23.7’’N

16º55’58.8’’W

1/1
4FromPicoAreeirotoPicoRuivo,MangaGrande,rockcrevices(ca.1650m)32º44’12.7’’N

16º56’03.3’’W

5/4
5FromPicoAreeirotoPicoRuivo,pr.PicodoGato,rockcrevices(1779m)32º44’35.5’’N

16º56’17.5’’W

12/8
6Pr.PicoAreeiro,westernslopeofPicodoGato,rockcrevices(1599m)32º44’39.3’’N

16º56’13.4’’W

10/8
7FromPicoRuivotoPicodasTorrinhas,stonyopenshrubsandpastures(1739m)32º45’42.2’’N

16º56’41.2’’W

1/1
8FromPicoRuivotoPicodasTorrinhas,rockycrevicesinwalls(1741m)32º45’42.2’’N

16º57’02.9’’W

1/1
3
4 Collectors:A.Costa;G.NietoFeliner;J.FuertesAguilar,M.MenezesdeSequeira,andR.Lansac.Date:July2003

30

1
2
3
3
4
5
6
7
8
9
10
Table2ErrorrateintheAFLPdatasetofArmeriamaderensis,calculatedasthetotal
numberoflocidifferencesrelativetothetotalnumberoflocicomparisons,and
subsequentlyaveragedoverthethreeprimercombinations
Nr.bands
retained
%error
rate
Allbandsincluded1907.1
Typeofbandsremoved:
Differentsizeamongsamples 1476.4
Lowintensity1446.6
Changingintensityamongsamples1233.1
Upper10%molecularweight1726.9
Lower10%molecularweight 1726.7
Upper10%molecularweight
+Lower10%molecularweight
1546.4
Upper20%molecularweight1547.1
Lower20%molecularweight 1546.1

Errorrateswerecalculatedforthewholedatasetandafterremovingpotentiallynon‐
homologousbandsaccordingtofourdifferentcriteria:i)slightsizedifferencesamong
putativehomologousbandsacrossindividuals;ii)lowintensitybands;iii)changing
intensityofonebandacrosssamples;andiv)bandsofhighorsmallmolecularweight
(Bagleyetal.2001;Boninetal.2004;seethetext)


Upper20%molecularweight
+Lower20%molecularweight
1325.6
Nr.bands
retained
%error
rate
Allbandsincluded1907.1
Typeofbandsremoved:
Differentsizeamongsamples 1476.4
Lowintensity1446.6
Changingintensityamongsamples1233.1
Upper10%molecularweight1726.9
Lower10%molecularweight 1726.7
Upper10%molecularweight
+Lower10%molecularweight
1546.4
Upper20%molecularweight1547.1
Lower20%molecularweight 1546.1
Upper20%molecularweight
+Lower20%molecularweight
1325.6
22

1
2
3
Table3ComparisonofthetotalAFLPdiversityinArmeriamaderensiswithfivelineages
detectedwithinthewidespreadcongenerArmeriapungensaccordingtoBayesian
clusteringmethods(recalculatedfromPiñeiroetal.2007)

n=15


P

Ht

Sh

A.maderensis


47.77

0.08(0.14)

0.14(0.20)
A.pungens
lineageI
(mainland+island)

73.91

0.21(0.18)

0.33(0.26)
A.pungens
lineageII
(mainland)

81.81

0.21(0.17)

0.33(0.23)
A.pungens
lineageIII
(mainland)

83.9

0.24(0.18)


0.37(0.24)
A.pungens
lineageIV
(mainland)

51.85

0.14(0.18)

0.22(0.26)
A.pungens
lineageVI
(Island)

77.44

0.19(0.18)

0.30(0.25)
4
5
6
7
8
9
10

P
−
percentageofpolymorphicloci;Ht
−
Nei’sgenediversity(1973)atthe
species/lineagelevel;Sh
−
allelesimilarity,Shannonindex(1948).Samplesizewas
standardizedto15individualsaccordingtothesizeofthesmallestlineage.Standard
deviationisreportedinbrackets.ThelineageofA.pungenswiththelowestdiversity
levelishighlighted.

23

1
2
3
4
5
6

Table4Within‐populationAFLPdiversityofA.maderensisandthefivelineages
detectedwithinthewidespreadcongenerArmeriapungensaccordingtoBayesian
clusteringmethods(recalculatedfromPiñeiroetal.2007)



n=6


P

Hs

Sh
A.maderensis  
Pop.2 17.780.06(0.14)0.09(0.20)
Pop.524.440.07(0.14)0.11(0.20)
Pop.627.780.06(0.11)0.12(0.19)
Mean22.330.060.11
A.pungenslineageI 
Pop.1 43.470.14(0.19)0.21(0.28)
Pop.250.430.18(0.26)0.20(0.29)
Mean46.950.160.20
A.pungenslineageII 
Pop.3 51.510.18(0.20)0.27(0.29)
Pop.4 35.610.11(0.17)0.17(0.25)
Pop.5 43.180.13(0.18)0.21(0.26)
Mean43.430.140.22
A.pungenslineageIII 
Pop.6 50.000.18(0.20)0.26(0.29)
Pop.7 54.240.19(0.20)0.28(0.29)
Mean52.120.190.27
A.pungenslineageIV  
Pop.8 26.670.09(0.17)0.14(0.25)
Pop.9 38.520.14(0.19)0.20(0.28)
Pop.10 23.700.09(0.17)0.13(0.24)
Pop.11 20.000.07(0.16)0.11(0.23)
Pop.12 34.070.11(0.18)0.17(0.25)
Pop.13 25.180.09(0.18)0.14(0.25)
Mean28.020.100.15
A.pungenslineageVI 
Pop.15 50.000.16(0.19)0.25(0.27)
24

Pop.16 29.270.11(0.18)0.16(0.26)
Pop.17 39.020.13(0.18)0.19(0.26)
Pop.18 28.660.10(0.18)0.15(0.26)
Pop.19 42.070.15(0.19)0.22(0.28)
Pop.20 42.680.15(0.19)0.22(0.28)
Pop.21 35.370.13(0.19)0.19(0.27)
Pop.22 35.980.13(0.19)0.20(0.28)
Pop.23 31.710.11(0.17)0.16(0.25)
Mean37.200.130.19
1
2
3
4
5

P
−
percentageofpolymorphicloci;Hs
−
Nei’sgenediversity(1973)atthepopulation
level;Sh
−
allelesimilarity,Shannonindex(1948).Populationsamplesizeswere
standardizedtosixindividuals.Standarddeviationisreportedinbrackets.Thelineage
ofA.pungenswiththelowestdiversitylevelishighlighted
25

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19

FIGURECAPTIONS
Fig.1LocationofsampledsitesofArmeriamaderensisfortheAFLPstudy,numbered
asinTable1

Fig.2AFLPstructureofArmeriamaderensisestimatedwithBayesianclusteringusing
STRUCTURE.aAssignmentof31individualsintoKgroups.Everyindividualis
representedbyaverticalbardividedinstripesofdifferentcolourcorrespondingtothe
estimatedassignmentprobabilitiestoeachgroup.Sampledsitesarenumberedasin
Table1.TenreplicatesateachKproducednearlyidenticalassignmentpatterns;the
highestprobabilityrunsateachKarerepresentedhere.bLogprobabilityofthedata
asafunctionofKaveragedover10STRUCTURErunsfromK=1toK=8.

Fig.3Principalcoordinatesanalysis(PCoA)representingDicesimilaritiesbetweenAFLP
phenotypesofArmeriamadarensis.Plotof31phenotypeseachcorrespondingtoa
singleindividualagainstthefirstandsecondprincipalaxesisshown,withaminimum
spanningtree(MST)superimposed.Thepercentageofvarianceaccountedforbyeach
axisisindicated.

26

1 
2
3
4
5
6
7





27

1
2
3



4
28