Are species–area relationships from entire archipelagos congruent with those of their constituent

14 Pages·2010·0.37 MB·English

GlobalEcologyandBiogeography,(GlobalEcol.Biogeogr.)(2010)19,527–540
Are species–area relationships from
RESEARCH
PAPER entire archipelagos congruent with
those of their constituent islands?
geb_536527..540
AnaM.C.Santos1,2*,RobertJ.Whittaker3,KostasA.Triantis2,3,
PauloA.V.Borges2,OwenR.Jones1,DonaldL.J.Quicke1,4,5and
JoaquínHortal4
1DepartmentofBiology,ImperialCollege ABSTRACT
London,SilwoodParkCampus,Ascot,
Aim To establish the extent to which archipelagos follow the same species–area
BerkshireSL57PY,UK,2Universidadedos
relationshipastheirconstituentislandsandtoexplorethefactorsthatmayexplain
Açores,Dep.deCiênciasAgrárias–CITA-A
departuresfromtherelationship.
(AzoreanBiodiversityGroup),Terra-Chã,
9700-851AngradoHeroísmo,Portugal, Location Thirty-eightarchipelagosdistributedworldwide.
3BiodiversityResearchGroup,Oxford
UniversityCentrefortheEnvironment,South Methods Weusedninety-sevenpublisheddatasetstocreateislandspecies–area
ParksRoad,OxfordOX13QY,UK,4NERC relationships(ISARs)usingtheArrheniuslogarithmicformof thepowermodel.
CentreforPopulationBiology,Imperial Observedandpredictedspeciesrichnessofanarchipelagoandofeachofitsislands
CollegeLondon,SilwoodParkCampus,Ascot, were used to calculate two indices that determined whether the archipelago fol-
BerkshireSL57PY,UK,5Departmentof lowedtheISAR.Archipelagicresiduals(ArcRes)werecalculatedastheresidualof
Entomology,NaturalHistoryMuseum, thepredictionprovidedbytheISARusingthetotalareaof thearchipelago,stan-
CromwellRoad,LondonSW75BD,UK dardizedbythetotalrichnessobservedinthearchipelago.Wealsotestedwhether
anycharacteristicofthearchipelago(geologicaloriginandisolation)and/ortaxon
accountsforwhetheranarchipelagofitsintotheISARornot.Finally,weexplored
therelationshipbetweenArcResandtwometricsofnestedness.
Results ThearchipelagowasclosetotheISARofitsconstituentislandsinmostof
the cases analysed. Exceptions arose for archipelagos where (i) the slopes of the
ISAR are low,(ii) observed species richness is higher than expected by the ISAR
and/or(iii)distancetothemainlandissmall.Thearchipelago’sgeologicalorigin
wasalsoimportant;ahigherpercentageofoceanicarchipelagosfitintotheirISAR
than continental ones.ArcRes indicated that the ISAR underpredicts archipelagic
richness in the least isolated archipelagos. Different types of taxon showed no
differencesinArcRes.NestednessandArcResappeartoberelated,althoughtheform
oftherelationshipvariesbetweenmetrics.
Main conclusions Archipelagos, as a rule, follow the same ISAR as their con-
stituentislands.Therefore,theycanbeusedasdistinctunitsthemselvesinlarge-
scalebiogeographicalandmacroecologicalstudies.DeparturefromtheISARcanbe
usedasacrudeindicatorofrichness-orderednestedness,responsivetofactorssuch
asisolation,environmentalheterogeneity,numberandageofislands.
*Correspondence:AnaM.C.Santos, Keywords
DepartmentofBiology,ImperialCollege
Archipelagic residuals, archipelagos, geological origin, island biogeography,
London,SilwoodParkCampus,Ascot,
island species–area relationship, isolation, macroecology, nestedness, power
BerkshireSL57PY,UK.
E-mail:[email protected] model,z-values.
©2010BlackwellPublishingLtd DOI:10.1111/j.1466-8238.2010.00536.x
www.blackwellpublishing.com/geb 527
A.M.C.Santosetal.
Schoener,1976;Wright,1983;Adler,1992,1994;Adler&Dudley,
INTRODUCTION
1994; Adler etal., 1995; Biber, 2002; Carvajal & Adler, 2005;
Thespecies–arearelationshipisoneofthemoststudiedpatterns Hamiltonetal.,2009).Infact,inhisdiscussionoftheso-called
inecology,oftenbeingreferredtoasoneofecology’sfewlaws ‘singlelargeorseveralsmall’debateontheimplicationsofisland
(Schoener,1976;Rosenzweig,1995,2003;Lawton,1996,1999). theoryforreservedesign,Rosenzweig(1995,p.382)arguedthat
According to this‘rule’, the number of species increases with ‘thediversityandtheareaofwholearchipelagosfallsinthesame
area,andtherateofincreaseofspeciesrichnessusuallydeclines species–areacurveastheseparateislandsthatconstitutethem’,
asareaincreases.Thereareanumberof classificationsforthe althoughstatingthatthishypothesisdeservesfurtherexamina-
differenttypesof species–arearelationships,dependingonthe tion.However,formaltestsoftheassumptionthatarchipelagos
scaleatwhichtheyareanalysedorwhethertheyaremeasured act as homogeneous entities in biogeographical terms are
from nested areas or not (e.g. Rosenzweig, 2003; Scheiner, lacking.Shouldthisassumptionberejected,eitherlocalecologi-
2003;Grayetal.,2004;Whittaker&Fernández-Palacios,2007; calfactorsortheparticularcharacteristicsofthegroupstudied
Dengler,2009).Rosenzweig(1995)describedthreemainscales (e.g. life-history traits, physiological adaptations) would pre-
(andtypes)ofspecies–arearelationships(four,ifthepointscale, dominateoverclassicalislandbiogeographyprocesses,challeng-
whichdependsonsamplingeffort,isincluded),thatcorrespond ing the universality of regional processes as the main factor
to different spatial/temporal scales. Following Whittaker & shapingthediversityofislandbiotas.
Fernández-Palacios(2007)wemaydescribethemas:(1)archi- Here,weevaluatewhetherentirearchipelagosfollowthesame
pelagic (or island species–area relationship), which is the species–area relationship as that defined by their constituent
species–area relationship within a group of islands; (2) intra- islands.Thatis,weassesstowhatextentthetotalrichnessofthe
provincial(orregionalspeciesaccumulationcurve),whichisa archipelagodepartsfromtheextrapolationof theISARtothe
speciesaccumulationcurvewithinalargecontinentalareaona total area of the islands that compose it. By implication we
regionalscale;and(3)inter-provincial,whichencompassesdif- therefore test the assumption that archipelagos act as single
ferentbioticregions(seealsoRosenzweig,2004;Triantisetal., entities in biogeographical terms, and hence the reliability of
2008a). In the present work we focus on the archipelagic- or using them as single units in large-scale biogeographical and
island-scale species–area relationship, henceforth termed the macroecological studies. We then evaluate our findings with
ISAR. regard to the type of taxon (invertebrates, vertebrates and
TheformandslopeofISARsdependontheparticularpro- plants),thegeologicalorigin(continentaloroceanic)andiso-
cess(es) that dominate(s) the study system (immigration,spe- lation(distancetothemainland)of eacharchipelago,andthe
ciationandextinction)(Rosenzweig,1995;seealsoMacArthur possible biological interpretation of departures of the archi-
& Wilson, 1963, 1967; Triantis etal., 2008a; Whittaker etal., pelagodatapointfromtheISAR,includingitsrelationshipwith
2008;Borges&Hortal,2009).Somearchipelagos(e.g.Hawaii, thenestednessofislandbiotas.
Galápagos,Azores)aresufficientlyisolated,inspaceortime,to
host a distinctive (‘disharmonic’) species pool, often drawn
METHODS
from more than one source region but with many shared
elements (e.g. species or lineages) among the islands of the Information on the species richness on islands was compiled
archipelago. The small number of colonization events, a from several sources for 97 archipelago/taxon combinations,
characteristic of isolated archipelagos, creates homogeneity in pertainingto38islandgroupsdistributedworldwide(Appen-
thespeciescolonizingtheseislands,whichcouldimplythatthe dix1).Ourdataincludearchipelagosofoceanic(i.e.bothtrue
processesestablishingislandspeciesrichnesswouldlargelybea oceanic islands and continental fragments sensu Whittaker &
property of the archipelago rather than of each constituent Fernández-Palacios,2007,followingWallace,1902),mixed,and
islandonitsown.Consequentlyeacharchipelago–oratleast continental origin and of varying size and degree of isolation
eachremotearchipelago–mayberegardedasauniqueentity fromtheclosestmainland,andcomprisedataonseveralgroups
similar to a province (Triantis etal., 2008a), regardless of the ofvertebrates,invertebratesandplants.Weexcludedintroduced
particularitiesofeachisland(seediscussioninWhittakeretal., speciesandsubspecifictaxafromalldatasets.Totalspeciesrich-
2008).The homogeneity in the processes that build up island ness for each archipelago was obtained by pooling the species
biotas would only be broken in cases where local (i.e.within- listsoftheirconstituentislands.Weusetheterm‘ISAR’torefer
island) idiosyncratic processes are predominant or the archi- to the species–area relationship constructed from the islands
pelago is composed by different subsets of islands that draw that constitute the archipelago (following Whittaker &
theircomponentsfromsignificantlydifferentspeciespools. Fernández-Palacios, 2007). We also use‘archipelagic point’ to
Regionalfactorsactingonthewholeofthearchipelago(such refertothetotalareaandrichnessofthecorrespondingisland
asarchipelagoisolation,age,originoftheislands)aregenerally group(archipelago).
thoughttohaveaconsistenteffectonthelocalpatternsofdiver- Foreacharchipelago/taxoncombination,anISARregression
sityattheislandlevel.Therefore,archipelagosareusuallycon- model, with observed species richness (S ) as the response
obs
sideredtobehomogeneousentities,anditisthusnotsurprising variableandislandarea(A)asthepredictor,wasconstructedon
that many authors have used complete archipelagos as single a log–log scale.This follows the same approach as Arrhenius’
data points in their analyses (e.g. Wilson, 1961; Scott, 1972; (1921) power model log S = c + b ¥ log A, where c is the
obs
528 GlobalEcologyandBiogeography,19,527–540,©2010BlackwellPublishingLtd
Archipelagospecies–arearelationship
interceptandbtheslope.Intheparticularcaseof theCanary that particular observation [i.e. if the residual was 1 and the
Islands, ISARs were constructed for both (1) all seven main predictedlog(speciesrichness)foragivenisland’sareawas3,
islands,and(2)allislandswiththeexceptionofFuerteventura thenPropMaxReswouldbe0.33].WeusedPropMaxResasanaid
andLanzarote,becausethesetwoislandsareknowntodeviate to examine the amount of disparity between the observed
fromthegeneralISARoftheirarchipelago,beingenvironmen- whole-archipelago species richness (SA ) and its predicted
obs
tallydifferentfromtheyoungerislandslocatedtothewestand species richness (SA ). If SA was within the bounds of
pred obs
lacking mesic upland habitats (see discussion in Whittaker & SA (cid:2)(SA ¥PropMaxRes),thenweassumethatwearenot
pred pred
Fernández-Palacios,2007,andWhittakeretal.,2008). able to reject the hypothesis that the archipelago follows the
Wewereunabletofindpublishedprotocolsforevaluatingthe ISAR.Conversely,ifSA wasoutsidethesebounds,weassume
obs
congruenceofarchipelagosandtheirISARs.Logically,thearchi- thatthishypothesiscanberejectedandthearchipelagospecies
pelagicpointintermsofspeciesrichnessmustfallsomewhere richnessisdeemedtoviolatetheISAR.Weconductedanalogous
betweenthespeciesrichnessofthemaximumindividualisland analysesusingthemedianoftheresiduals(MedRes)asamore
richnessandthesumofthespeciesrichnessofallislandsinthe conservative criterion, determining whether the archipelago
archipelago,andarchipelagicareaissimplythesumofthearea fallsintotheintervaldefinedbySA (cid:2)(SA ¥PropMedRes)
pred pred
of all constituent islands. Therefore, the archipelagic point ornot.Wherethearchipelagohadanevennumberofislandsthe
cannot be considered independent of the ISAR, and so we PropMedReswascalculatedusingthemedianvaluesoftheabso-
cannotformallytestcongruenceusingstandardregressiontech- luteresidualsandofalltheSI .
pred
niques.We therefore generated two simple indices to evaluate The above procedure can be illustrated using the vascular
thedepartureof thearchipelagicpointfromtherichnesspre- plantsof CapeVerdeasanexample(Table1).Forthisdataset,
dictedbyextrapolatingtheISARtothetotalareaforthewhole theISARequation(onalog–logscale)islogS=1.385+0.291¥
archipelago(seebelow).Theseindicesarebasedonanexami- logA;SA andSI wereobtainedbyapplyingthisequation
pred pred
nationofthevariationaroundthepredictedISARbutwithout to the logarithm of the total land area of the archipelago
any statistical probability being attached to them. However, (4076km2)andtothelogarithmofeachislandarea,respectively
havingcalculatedtheindicesoffit,wedomakeuseofinferential (Table1). The maximum residual was 0.157 (Santo Antão
statistical tests to evaluate the strength of the relationships island)andPropMaxReswas0.074,whichwasobtainedbycal-
betweenthesemetricsandthepropertiesofthearchipelagoand culatingtheproportionofMaxResoveritscorrespondingSI
pred
thetypeoftaxonstudied. (2.127).ThevalueobtainedforSA ¥PropMaxReswas2.436¥
pred
Foreacharchipelago/taxoncombination,weestimatedarchi- 0.074 = 0.179, so the interval defined by SA (cid:2) (SA ¥
pred pred
pelagic(SA )andconstituentislands’(SI )speciesrichness PropMaxRes)was2.436(cid:2)0.179.Sincethelogarithmofthetotal
pred pred
fromtheISARregressionmodel.ValuesofSI weresimplythe richness of the archipelago (2.441) lies inside this interval,we
pred
fittedvaluesoftheregressionmodel,whileSA wasestimated cannot reject that the archipelago is following the ISAR. The
pred
fromthemodelusingthetotalland-surfaceareaof thearchi- sameprocedurewasfollowedusingthemedianresidual(0.087)
pelagoasthepredictor.Wethencalculatedtheresidualsofthe insteadofMaxRes.
regression model (i.e. observed species richness minus SI ) WeusedtheratiobetweenMaxResandMedResasameasure
pred
and identified their maximum absolute value (MaxRes). We of the dispersion of the most distant island points within the
expressed MaxRes as a proportion (PropMaxRes) of SI for archipelago for a preliminary evaluation of whether such dis-
pred
Table1 Valuesusedtocalculatethe
intervalthatdelimitswhetheran Islands Sobs A logSobs logA SIpred SApred AbsResid
archipelagicpointiscongruentwithits
BoaVista 126 634.1 2.100 2.802 2.201 0.101
islandspecies–arearelationship(ISAR)
Brava 100 66.6 2.000 1.823 1.916 0.084
ornot(formoredetailsseetext).
Fogo 154 474.8 2.188 2.677 2.164 0.023
Maio 117 279.0 2.068 2.446 2.097 0.029
Sal 92 221.5 1.964 2.345 2.068 0.104
SantaLucia 61 36.7 1.785 1.565 1.841 0.055
Santiago 183 991.0 2.262 2.996 2.228 0.034
SantoAntão 192 787.3 2.283 2.896 2.127 0.157
SãoNicolau 144 352.2 2.158 2.547 2.257 0.099
SãoVincente 146 232.8 2.164 2.367 2.074 0.090
CapeVerde 276 4076.0 2.441 3.610 – 2.436 –
S istheobservedspeciesrichnessofthevascularplantsofCapeVerde,Aistheisland/archipelago
obs
areaexpressedinkm2,logS isthelogarithmofS ,logAisthelogarithmofA,SI isthepredicted
obs obs pred
valueforthespeciesrichnessofeachisland,SA isthepredictedvalueforthespeciesrichnessof
pred
thearchipelago,andAbsResidistheabsolutevalueoftheresidualsobtainedbysubtractingSI from
pred
logS .
obs
GlobalEcologyandBiogeography,19,527–540,©2010BlackwellPublishingLtd 529
A.M.C.Santosetal.
persion would affect the results of our analyses. The cases in Table2 Resultsofthenestednessanalysesforseveralarthropod
agreementwiththeISARshowedsimilarMaxRes/MedResratios groupsintheAzoresandCanaryIslands(datafromBorgesetal.,
tothosefallingoutsidethisrelationship,forboththemaximum 2005,andIzquierdoetal.,2004,respectively).
andmedianresidualcriteria(notshown).Thedegreeofdisper-
Archipelago n Taxon ArchRes T NODF
sion of the distant island points was, however, higher for the
cases meeting only the MaxRes criterion than for the cases
Azores 9 Allarthropods -0.701¥10-2 19.38 31.53
meetingbothcriteria,asexpected.Wethereforeassumethatthe
9 Arachnids -6.585¥10-2 20.87 31.04
degree of dispersion of the island points does not affect the
9 Coleoptera -3.653¥10-2 18.29 23.35
probabilityofrejectingthecongruenceofanarchipelagicpoint 9 Lepidoptera 2.376¥10-2 28.20 34.55
withitsISAR.Rather,itonlyaffectstheprobabilityofmeeting
CanaryIslands 7 Arachnids* 6.791¥10-2 33.62 14.49
just the less restrictive or both criteria. Therefore, we expect
5 Arachnids† 1.522¥10-2 34.72 16.92
differencesinthecriterionmettobemainlydrivenbythedegree 7 Coleoptera* 8.468¥10-2 40.83 23.05
of dispersion of the islands, and not by factors causing the 5 Coleoptera† 2.405¥10-2 40.12 26.20
archipelagicpointtodepartfromtheISAR.Itfollowsthatboth 7 Lepidoptera* -6.093¥10-2 20.64 26.45
criteria are equally reliable in terms of identifying whether an 5 Lepidoptera† -7.394¥10-2 15.57 30.98
archipelagofollowstheISARofitsconstituentislandsornot.
Todetermineifanyarchipelagoand/ortaxoncharacteristics *ReferstoallCanaryIslands(seetext).
†ReferstoallCanaryIslandsexceptFuerteventuraandLanzarote(see
accountforthefitofanarchipelagotoitsISAR,weclassifiedthe
text).
datasets according to: (1) the kind of taxon they belong to
n,numberofislands.
(invertebrates,vertebratesorplants),and(2)theoriginof the
ArchResisthearchipelagicresidual(theresidualofthepredictionpro-
archipelago (continental, mixed or oceanic). In addition, we videdbytheislandspecies–arearelationship(ISAR)usingthetotalarea
measured (3) the isolation of the archipelago as the smallest ofthearchipelago,dividedbythetotalrichnessobservedinthearchi-
distancebetweenanyoftheislandsandthenearestmainland. pelago),TistheoriginalTemperaturemeasureofnestedness(Atmar&
To obtain a measure of how much the archipelago departs Patterson,1993),andNODFisthenestednessmetricbasedonoverlap
anddecreasingfillproposedbyAlmeida-Netoetal.(2008).
fromtheISAR,andallowanexplorationofpotentialcausesof
AnalyseswerenotperformedforallarthropodsintheCanaryIslands
deviation,wecalculatedthearchipelagicresidual(ArcRes)asthe
due to the limit of 3000 lines (i.e. species) of the program used to
residualofthepredictionprovidedbyISARusingthetotalarea
computethenestednessmeasure(aninhado;Guimarães&Guimarães,
of the archipelago. To enable comparisons between different 2006).
archipelagos,westandardizedthisresidualbydividingitbythe
total richness observed in the archipelago. Using the above
exampleofthevascularplantsfromCapeVerde,ArcReswould pared these measures and ArcRes by simple correlations and
be calculated as (log SA - SA )/log SA ,that is (2.441 - visualexamination.Nestednessmeasureswerecalculatedusing
obs pred obs
2.436)/2.441=0.002.WeusedArcResasaresponsevariablein aninhado(Guimarães&Guimarães,2006).Allotherstatistical
regression models designed to explore the potential causes of analyseswereperformedusingStatistica6.1(StatSoft,2003).
deviationfromtheISARinthedatasetsthatyieldedsignificant
regressions(P<0.05).Potentialexplanatoryvariablesincluded
RESULTS
taxontype,geologicaloriginofthearchipelagoandisolation,as
above. Seventy-two (74%) out of the 97 ISARs examined had slopes
Finally,weexploredwhetherthereisarelationshipbetween significantlydifferentfromzero(Appendix1).Mostofthenon-
the magnitude of ArcRes and the degree of nestedness of the significantISARscamefromtheCanaryIslands(18outof25),
island biotas within the archipelago. Detailed data on species butmostofthesebecamestatisticallysignificantafterexcluding
composition per island were not available for many of the the two more xeric and older islands, Fuerteventura and
datasetsusedfortheformeranalyses.Duetothis,andtoavoid Lanzarote(seeexampleinFig.1a,b).Allsubsequentresultsare
problemsrelatedtounevensamplingeffort,weonlyanalyseda basedonthesignificantISARsonly.Inthesearchipelago/taxon
reducednumberofarthropodgroupsintwoarchipelagosweare combinations, slopes (i.e. z-values) ranged from 0.08 to 0.94,
morefamiliarwith:theAzoresandtheCanaryIslands(Table2). withthelowerandupperquartilesbeing0.22and0.52,respec-
These datasets,however,present an ample variation in ArcRes tively;themedianwas0.33;andtheoverallmeanwas0.38.
values (compared with the variation found in all studied We could not reject the hypothesis that the archipelago
datasets)andincludecasesthatbothenterandfailtoenterinthe speciesrichnessfollowstheISARin63cases(88%)whenusing
medianandmaximumresidualcriteria(seeAppendix1).Given the maximum residual criterion (Appendix1; Fig.1b,f,g,h).
currentdebateonthemostappropriatemeasureofnestedness TheninecaseswherethearchipelagodidnotfollowtheISAR
(e.g.Almeida-Netoetal.,2007;Ulrichetal.,2009),wecalculated according to this criterion had significantly lower slopes
twodifferentmeasures:thenestednessmetricbasedonoverlap (median=0.21,lowerandupperquartiles=0.18and0.28)than
and decreasing fill (NODF;Almeida-Netoetal.,2008),as rec- thosethatdidfit(median=0.40,lowerandupperquartiles=
ommendedbyUlrichetal.(2009);andtheoriginalTemperature 0.27 and 0.53) (Mann–Whitney U = 156.5,Z = 2.162,n = 9,
1
(T)measureproposedbyAtmar&Patterson(1993).Wecom- n =63,P<0.05).Usingthemoreconservativemedianresidual
2
530 GlobalEcologyandBiogeography,19,527–540,©2010BlackwellPublishingLtd
Archipelagospecies–arearelationship
criterion,only45outof72caseswerecongruentwiththeirISAR in such cases, the ISAR tends to under-predict archipelagic
(63%)(Appendix1;seeexamplesinFig.1b,g,h).Again,mostof speciesrichness,especiallyintheleastisolatedislands.However,
thecasesinwhichthehypothesisof thearchipelagofollowing inthecaseoftheinvertebratesfromoceanicislandstheopposite
the ISAR was rejected showed lower slopes (median = 0.27, relationship appears to occur: the least isolated archipelagos
lower and upper quartiles = 0.19 and 0.33) than those where typicallyhadnegativeArcResvalueswhilethemostdistantones
suchahypothesiswasnotrejected(median=0.40,lowerand exhibitedpositiveresiduals.Foralltheotherarchipelago/taxon
upperquartiles=0.29and0.58)(Mann–WhitneyU=301.5,Z= combinations the few data points available were much more
3.559,n =27,n =45,P<0.001). scatteredandthereforeshowednoevidenttrend.
1 2
When the maximum residual interval was used,vertebrates Thedegreeofnestednessofislandbiotasshowedaclearrela-
hadmorearchipelagosfollowingtheirISARthaninvertebrates tionshipwithArcRes:themorenestedadataset,thelower(and
or plants (vertebrates 92%, invertebrates 88%, plants 78%; negative)thearchipelagicresidual(Table2,Fig.3).Inthecases
Appendix1). However, these differences were not significant analysed this pattern is more evident using the Temperature
(c2=1.06,P=0.59).Furthermore,whenconsideringtheinter- index(SpearmanR=0.83,n=10,t[n-2]=4.21,P<0.01)than
valdefinedbythemedianresidual,theproportionofcasesthat for NODF (Spearman R = -0.48, n = 10, t [n - 2] = -1.54,
werecongruentwiththeISARwasalmostthesameforeachone P=0.16).
of the three groups (vertebrates and invertebrates 62%, and
plants67%;c2=0.08,P=0.96).Inanycase,differencesseemed
DISCUSSION
to be stronger between archipelagos than between taxa.Some
evidenceofthiscamefromthearchipelagosforwhichwehave Itisoftencommentedthatcommunityecologyhasfewgeneral
dataondifferenttaxonomicgroups;whereasinseveraltaxaof rules or laws (Lawton, 1996, 1999, 2000; but see Gaston &
theCanaryIslandsthearchipelagicpointfelloutsideitsISAR, Blackburn,2000;Simberloff,2004;Ricklefs,2008).Oneof the
thiswasnotthecasefortheAzoresorCapeVerde(Appendix1). fewexceptionsisheldtobethespecies–arearelationship,which
OceanicarchipelagoswerecongruentwiththeirISARmore iswidelyapplicableforallscalesandtypesoforganisms,from
oftenthancontinentalones,accordingtotheMaxRescriterion bacteria to vertebrates (Rosenzweig, 1995). The species–area
(95%,vs.76%;c2 = 5.43,P < 0.05).This difference was more relationship has commonly been described for discrete geo-
pronounced when the MedRes criterion was used; while most graphic units, such as islands within an archipelago (Rosenz-
oceanic archipelagos remain within their ISAR (75%), only weig,1995;Whittaker&Fernández-Palacios,2007).However,it
41% of the continental archipelagos showed the same result isnotunusualtolumptogetherdataontheconstituentislands
(c2=7.98,P<0.05).Mixedarchipelagoswerenotconsideredin of archipelagos by adding island areas and combining island
thesecomparisonsbecauseofthesmallnumberofcases(n=3). speciesliststoobtaintheoverallareaandrichnessofthearchi-
Thedistancefromthearchipelagotothemainlandalsohada pelagos(e.g.Wilson,1961;Scott,1972;Wright,1983;Adleretal.,
significanteffectontheprobabilityofrejectingthehypothesis, 1995; Carvajal & Adler, 2005; among others). Although this
since more isolated archipelagos were more congruent with allowstheuseofthesedatainlarge-scaleanalyses,italsoimplies
their ISARs than less isolated ones (Mann–Whitney U = 159, that the archipelagos as a whole follow the same relationship
Z = -2.12, n = 9, n = 63, P < 0.05 for the MaxRes; Mann– withareathattheirconstituentislandsdo,anassumptionthat
1 2
WhitneyU=413,Z=-2.26,n =27,n =45,P<0.05forthe hasneverbeenevaluatedbefore.
1 2
MedRes). We have shown that archipelagos do often follow the same
InFig.2weshowtheArcResof thedatasetswithsignificant ISAR as their constituent islands.In the majority of the cases
ISARsagainsttheirrespectivedistancestothemainland,indi- studied,thearchipelagicpointiscongruentwithitsISAR,which
catingalsothetypeoftaxonandgeologicaloriginofthearchi- begsthequestionofwhywholearchipelagosshouldfollowthe
pelago. For continental islands ArcRes values showed a samespecies–arearelationshipastheirconstituentislands?The
marginallynon-significantnegativerelationshipwithisolation answer must be related to the endogenous dynamics of the
(SpearmanR=-0.33,P=0.08),whereasforoceanicislandsthis system (speciation, immigration and extinction), which are
relationship was significant and positive (Spearman R = 0.39, determinedbyanumberofregionalfactorssuchastotalareaof
P < 0.05). ArcRes of continental archipelagos and vertebrates thearchipelago,numberofislands,environmentalheterogene-
showedhighervariationthaninothercategories(SD=0.14and ity,isolationandgeologicalage.Thesefactorsactaslocalfilters
0.16,respectively,versus0.12forplants,0.09forinvertebrates, for the regional processes acting over the species pool of the
0.08formixedarchipelagosand0.07foroceanicarchipelagos). archipelago(seeRicklefs,2007,2008).Giventheimportanceof
In fact, there were significant differences between the ArcRes regional processes common to all their constituent islands,
valuesasafunctionofthegeologicaloriginofthearchipelagos many archipelagos can be considered to behave as a coherent
(i.e. ArcRes values were bigger for continental archipelagos; entity for the different processes establishing species diversity.
Mann–WhitneyU=344,Z=2.87,P<0.01;mixedarchipelagos Interestingly,asaconsequenceof sourcepooleffectsthereisa
werenotconsideredforthisanalysis)butnotbetweendifferent closeassociationbetweenspecies–areaandlocal–regionalrich-
taxa[Kruskal–WallisH(2d.f.,n=72)=0.093;P=0.95].Itis ness relationships (see Srivastava, 1999; He etal., 2005), so
noteworthy that invertebrates from continental islands had speciesrichnessonindividualislandsistosomeextentareflec-
morepositiveArcResvaluesthannegativeones,indicatingthat, tionofthespeciespoolofthearchipelago.
GlobalEcologyandBiogeography,19,527–540,©2010BlackwellPublishingLtd 531
A.M.C.Santosetal.
4.2 4.4
(a) (b)
4.0 4.2
3.8
4.0
s
s
e 3.6
hn 3.8
Ric 3.4
3.6
g 3.2
o
l 3.4
3.0
2.8 3.2
2.6 3.0
2.5 3.0 3.5 4.0 4.5 2.5 3.0 3.5 4.0 4.5
3.6 3.6
(c) (d)
3.3
3.4
ss 3.0
e
n 3.2
h
c 2.7
Ri
g 3.0
o 2.4
l
2.8
2.1
2.6 1.8
2.5 3.0 3.5 4.0 4.5 2.5 3.0 3.5 4.0 4.5
2.8 2.8
(e) (f)
2.4
2.4
2.0
ss 2.0
e
n 1.6
h
Ric 1.2 1.6
g
o 1.2
l 0.8
0.4 0.8
0.0 0.4
-3 -2 -1 0 1 2 3 4 1.5 2.0 2.5 3.0 3.5 4.0 4.5
2.8 3.2
(g) (h)
2.4
2.8
2.0
s
es 2.4
n 1.6
h
c
Ri 1.2 2.0
g
o
l 0.8 1.6
0.4
1.2
0.0
-1 0 1 2 3 4 0 1 2 3 4
log Area (km2) log Area (km2)
Figure1 Relationshipbetweenspeciesrichnessandareaforseveralarchipelago/taxoncombinations.Individualislandsarerepresentedby
blackcirclesandthearchipelagosbygreytriangles.Theislandspecies–arearelationship(ISAR)predictedbytheregressionfunctionis
shownasacontinuouslineineachcase.Theintervalsdefinedbythemaximumandmedianresidualscriteria(seetext)arerepresentedby
thedashedandthedottedlines,respectively:(a)CanaryIslands–arthropods(allislands);(b)CanaryIslands–arthropods(without
FuerteventuraandLanzarote);(c)CanaryIslands–Coleoptera(withoutFuerteventuraandLanzarote);(d)CanaryIslands–Lepidoptera
(withoutFuerteventuraandLanzarote);(e)AegeanIslands–landsnails;(f)Fiji–ants;(g)ChatmanIslands–birds;(h)CookIslands–
plants.SeedatasourcesandregressionstatisticsinAppendix1.
532 GlobalEcologyandBiogeography,19,527–540,©2010BlackwellPublishingLtd
Archipelagospecies–arearelationship
0.5 a)
0.4 0.10
0.08
0.3
0.06
0.2
0.04
Res 0.1
Arc 0.02
0.0 es
R 0.00
c
r
-0.1 A
-0.02
-0.2
-0.04
-0.3 -0.06
0 1 5 50 500 5000
log Distance (km) -0.08
Figure2 Distributionofthearchipelagicresidual(ArcRes)for -0.10
15 20 25 30 35 40 45
the72significantarchipelago/taxoncombinations,accordingto
thelogarithmofthedistancetothemainland(inkm).ArcReswas T
calculatedastheresidualofthepredictionprovidedbytheisland
b)
species–arearelationship(ISAR)usingthetotalareaofthe
archipelago,standardizedbythetotalrichnessobservedinthe 0.10
archipelago.Thesymbolshaderepresentsthetypeoftaxon
0.08
(black,vertebrates;grey,invertebrates;white,plants),andtheir
shapesrepresentthetypeofarchipelago(circles,continental; 0.06
triangles,oceanic;squares,mixedarchipelagos).Thehorizontal
0.04
lineservesonlyasaguidelinetoArcRes=0.
0.02
s
e
R 0.00
Giventhatareaisoneofthebestmacroecologicaldescriptors rc
A
ofislandspeciesrichness(see,e.g.,Rosenzweig,1995;Whittaker -0.02
& Fernández-Palacios, 2007; Triantis etal., 2008a), the island
-0.04
group as a whole should also be expected to follow its endo-
-0.06
genousspecies–areadynamics.Wethusarguethatifthearchi-
pelago is sufficiently isolated for all the islands to have an -0.08
equivalent species pool,and the geological characteristics and
-0.10
evolutionaryprocessesarecoherentfromislandtoisland,itis 10 15 20 25 30 35 40
unsurprising that the biota of the whole of the archipelago NODF
wouldfollowthesamerelationshipwithareaastheoneoper-
Figure3 Relationshipbetweenthearchipelagicresidual
atingwithintheconstituentislands.Inotherwords,theaccu-
(ArchRes)andtwonestednessmeasures:(a)Temperature(T),the
mulation of new species with additional islands will show a
originalmeasureproposedbyAtmar&Patterson(1993),and(b)
consistentrelationshipwiththeirarea.Itfollowsfromourratio-
theoverlapanddecreasingfillmetric(NODF)proposedby
nalethatthedepartureofthearchipelagofromtheISARwillbe
Almeida-Netoetal.(2008).NotethatthehighertheNODF,the
relatedtosomeextenttothedegreeofnestednessoftheisland
morenestedthearchipelago,whichistheconverseoftheordering
biotas(seealsoWrightetal.,1998). oftheTmetric.Datacorrespondtoseveralarthropodgroupsof
The additional analyses carried out on several arthropod theAzoresandtheCanaryIslands(seeTable2).
groups of the Canary Islands and the Azores show that the
magnitudeofthedepartureofthearchipelagofromtheISARis
related to nestedness.The more nested the biota of the archi- etal.,2008;Ulrichetal.,2009).Inotherwords,whileTempera-
pelago, the lower the archipelagic residual, a trend that is ture values reflect how widespread species are distributed in
especiallyevidentwhenusingtheTemperaturemetric.Thesen- progressively less rich islands, NODF also takes into account
sitivitytothechoiceofnestednessmetric[Temperaturemetric whetherpoorislandshostrarespeciesornot.Theserarespecies
ofAtmar&Patterson(1993)versusNODFmetricofAlmeida- arespeciespresentinjustoneorafewislands,soNODFwould
Neto etal. (2008)] might reflect the nature of the nestedness beexpectedtobemoresensitivewhenthepatternsofnestedness
patterns. While traditional ‘gap-counting’ nestedness metrics within the archipelago are driven by the numbers of single
such as Temperature are biased towards the loss of species island endemics (SIEs). Within this framework, the less tight
among islands,NODF also accounts for the degree of coinci- relationship between the archipelagic residual and NODF
denceofspeciespresencesinthepoorersites(seeAlmeida-Neto compared with its relationship with Temperature allows us to
GlobalEcologyandBiogeography,19,527–540,©2010BlackwellPublishingLtd 533
A.M.C.Santosetal.
postulate that the number of SIEs per se does not necessarily differingcharacteristics,theprocessesbuildingupislandbiotas
have an effect on the departure of the archipelago from the mightvaryamongsttheconstituentislands.Thiswillhappenin
ISAR.Infact,manycaseswithdisproportionatelyhighnumbers archipelagoswhere:(1)theproximitytothesource(s)ofcolo-
of SIEs (and therefore highly non-nested biotas) fall into the nizersallowsinter-islandvariationincolonizationratesand/or
confidence intervals we used in this work [e.g.Hawaiian land the arrival probability of particular species or lineages (thus
snails(Cowie,1995)andlobeliads(Givnishetal.,2009)]. different sets of widespread species will be found in different
The residual variation of the archipelagic data point seems clustersofislands);(2)oneorsomeoftheislandsshowhigher
thustoreflectaparticularaspectofnestedness:theabsenceof speciation rates (e.g. because they are significantly larger; see
species present in the richer islands in progressively poorer Losos&Schluter,2000);andalso(3)someislandssufferanoma-
islands (i.e. richness-ordered nestedness sensu Whittaker & louspulsesofextinction(e.g.islandsterilizationprocesses;see
Fernández-Palacios,2007).DeparturesfromtheISARarethus Whittaker & Fernández-Palacios, 2007). All these cases will
expectedinsystemsthatareeitherhighlynestedornotnestedat produceanomalouspatternsofspeciesaccumulationwitharea,
all(Fig.4),independentlyofthenumberofSIEs.Inthecaseof andthedepartureoftheoverallrichnessofthearchipelagofrom
highly nested systems, the richness of the archipelago will theISARofitsconstituentislands.
scarcelyexceedthatofthelargestislandandwillbefairlyinsen- Thisraisesamoregeneralpoint:almostallsituationswhere
sitivetothenumberandareaofsmallerislands.Thus,thepre- thearchipelagicpointdeviatessignificantlyfromtheISARcame
dictednumberofspeciesforthetotalareaofthearchipelagowill fromcaseswhereboththeslopeoftheISARandthedegreeof
behigherthantheobservedspeciesrichness(e.g.Lepidoptera isolationwerelow,andthetotalspeciesrichnesswashigherthan
from the Canary Islands; Fig.1d). Conversely, in highly non- expectedbytheISAR.GiventhattheISARslopesoftheseexcep-
nested systems there is a high rate of accumulation of new tionalcasesarelow,itisnotsurprisingthatthetotalrichnessof
species with the addition of each island. Hence, the observed the archipelago falls above the relationship observed for the
archipelagic species richness (a cumulative total) should be islands,asseenfortheColeopteraoftheCanaryIslandsorthe
higher than that predicted by the non-cumulative series pro- land snails of the Aegean Islands (Fig.1c,e, respectively). The
videdbytheISAR(e.g.ColeopteraoftheCanaryIslandsorland reasonsforsuchanon-nestedpatterncouldbe:(1)aparticu-
snails of the Aegean Islands; Fig.1c,e). However, as discussed larlyhighspeciationrateand/orparticularpatternsinclustersof
above,high overall numbers of SIE in the archipelago are not islands(i.e.limitationstodispersalandavailabilityofecological
enoughtocausesignificantdeparturesfromtheISAR(e.g.the spacethatallowahighernumberofspeciationeventsandthus
highly species rich land snails and lobeliads of Hawaii men- thegenerationoflargenumbersofspeciesoverthewholearchi-
tioned above). Rather, departures will appear when the pro- pelago);(2)aheterogeneousgeologicalhistoryamongtheindi-
cessesleadingtotheappearanceofhighnumbersofSIEdiffer vidual islands, enough to differentiate sets of islands with
withinthesetofislandscollatedtogetherforananalysis.Where particular dynamics (e.g. the Canary Islands; see Whittaker
anarchipelagoiscomposedofdifferentgroupsofislandswith etal.,2008);or(3)multiplesourcesforthearrivalofnewspecies
(which is the case for the Aegean Islands). Conversely, highly
nestedsystemsmightarisefromseveralcircumstances,includ-
ing:(4)alowrateofcolonizationfromthecontinent,whichwill
resultinasmallnumberoflineagesinhabitingthearchipelago,
Non–nested
andthuslowercompositionalreplacement;and(5)highratesof
ss Prediction
e dispersalbetweenislands,whichwillalsoresultinlowcompo-
n Nested
h sitional replacement from island to island. Canarian Lepi-
c
Ri dopteraprovidedtheonlyexampleinwhichtheoverallrichness
og of the archipelagic point falls significantly below the ISAR
l
accordingtoourmaximumresidualcriterion,indicatingalow
log Total Area
degree of dissimilarity for this group (see Fig.1d). Butterflies
and moths are in general very good dispersers (e.g. Borges &
log Area Hortal, 2009), which results in low compositional differences
between island faunas (our reason 5).Hence,the total species
Figure4 Hypotheticalrelationshipsbetweendifferentdegreesof richnessofthearchipelagodoesnotincreasesignificantlywith
nestednessandthedepartureofthearchipelagopointfromthe theadditionofnewislands.
islandspecies–arearelationship(ISAR),inalog–logspecies–area
Our analytical approach presents several limitations that
plot.Here,nestednessisaby-productoftherateofaccumulation
reflecttheexploratorynatureofourstudy.Forexample,many
ofnewspeciestothetotallistofthearchipelagowiththeaddition
of the cases we studied pertain to just five island groups
ofnewislands(i.e.richness-orderednestednesssensuWhittaker&
(Aegean Islands, Azores, Canary Islands, Cape Verde and
Fernández-Palacios,2007).Whereasystemtendstowardsperfect
Hawaii) for which relatively comprehensive data are available.
nestedness,thearchipelagorichnesswillbethesameasthatofthe
largestisland(oronlyalittlehigher),whileforcompletely Asaconsequence,althoughtheresultsfromthesearchipelagos
unnestedsystemsthearchipelagorichnessisthesumofthe reassure us that our conclusions are reliable in broad qualita-
richnessofeachconstituentisland. tive terms, we cannot be sure that these results will also turn
534 GlobalEcologyandBiogeography,19,527–540,©2010BlackwellPublishingLtd
Archipelagospecies–arearelationship
outtobereliableinquantitativeterms.Giventherelationship 21496/2005),K.A.T.by a Marie Curie Intra-European Fellow-
between the degree of departure of the ISAR and nestedness ship Program (project ‘SPAR’, 041095) and the Academic
foundhere,werecommendthatfurtherresearchonthistopic VisitorsProgramof theNERCCentreforPopulationBiology,
shouldrelyonthedevelopmentofnullmodelsoftherelation- P.A.V.B.byCITA-A,andJ.H.bytheUKNaturalEnvironment
ship between area and species assembly in a presence/absence ResearchCouncil.
matrix. The assumptions and development of these models
are currently under debate, and several null hypotheses for
REFERENCES
randomassemblyhavebeenproposedsofar(see,e.g.,Wright
etal., 1998; Rodríguez-Gironés & Santamaría, 2006; Almeida- Adler,G.H.(1992)EndemisminbirdsoftropicalPacificislands.
Neto etal., 2007, 2008; Ulrich etal., 2009). In fact, these null EvolutionaryEcology,6,296–306.
hypothesescorrespondtodifferentaspectsofnestedness:rich- Adler,G.H.(1994)Avifaunaldiversityandendemismontropi-
nessorderedorareaordered(seealsoWhittaker&Fernández- cal Indian Ocean islands. Journal of Biogeography, 21, 85–
Palacios, 2007; Ulrich etal., 2009). We anticipate that 95.
departures from the null expectations of these models would Adler, G.H. & Dudley, R. (1994) Butterfly biogeography and
dependonthedisparityinthepresenceof widespreadspecies endemismontropicalPacificislands.BiologicalJournalofthe
amongislands,ratherthanonthenumberofrarespecies(e.g. LinneanSociety,51,151–162.
SIE; see above). Therefore, we hypothesize that these depar- Adler, G.H., Austin, C.C. & Dudley, R. (1995) Dispersal and
tureswouldprovideinsightintotheprocessesdeterminingthe speciationofskinksamongarchipelagosinthetropicalPacific
assembly of island biotas within the archipelago, such as Ocean.EvolutionaryEcology,9,529–541.
varyinggeologicalhistories,islandisolationorhabitatdiversity Almeida-Neto, M., Guimarães, P.R., Jr & Lewinsohn, T.M.
(see,e.g.,Roughgarden,1989;Lomolino&Davis,1997;Hortal (2007) On nestedness analyses: rethinking matrix tempera-
etal., 2009), but they will be relatively independent of differ- tureandanti-nestedness.Oikos,116,716–722.
ences in the evolutionary processes among islands. Almeida-Neto, M., Guimarães, P., Guimarães, P.R., Jr, Loyola,
Tosummarize,wehaveshownthatarchipelagosusuallyfollow R.D.&Ulrich,W.(2008)Aconsistentmetricfornestedness
thesamespecies–arearelationshipastheirconstituentislands.A analysisinecologicalsystems:reconcilingconceptandmea-
straightforward implication of our results is that archipelagos surement.Oikos,117,1227–1239.
can, in most cases, be considered as distinct entities. Hence, Arechavaleta, M., Zurita, N., Marrero, M.C. & Martin, J.L.
researcherswouldbejustifiedinlumpingspecieslistsfromtheir (2005) Lista preliminar de especies silvestres de Cabo Verde
constituent islands when conducting biogeographical and/or (hongos, plantas y animales terrestres). Consejería de Medio
macroecologicalstudies.Itisalsoimportanttonotethatmostof AmbienteyOrdenaciónTerritorial,GobiernodeCanarias,La
thearchipelagosstudiedareinfact‘SLOSSneutral’(sensuRosen- Laguna,Tenerife.
zweig,2004),andthusthatatthislargescalewhetherconserva- Arrhenius, O. (1921) Species and area. Journal of Ecology, 9,
tioneffortsaredevotedtoasinglelargeislandortoseveralsmall 95–99.
onesmaybeof limitedrelevance(seealsoRosenzweig,2004). Atmar,W.&Patterson,B.D.(1993)Themeasureoforderand
Importantly, the degree of departure from the ISAR (i.e. the disorderinthedistributionofspeciesinfragmentedhabitat.
archipelagicresidual)isrelatedtoaparticularaspectofnested- Oecologia,96,373–382.
ness,thelossofspeciespresentontherichestislandsfromthe Azeria,E.T.(2004)Terrestrialbirdcommunitypatternsonthe
poorer ones, and can therefore be used as a crude index of corallineislandsoftheDahlakArchipelago,RedSea,Eritrea.
richness-orderednestednesswhendetailedislandchecklistsare GlobalEcologyandBiogeography,13,177–187.
lacking.Giventhatareaisjustoneofanumberoffactorsdeter- Beck,J.&Kitching,I.J.(2004–2008)TheSphingidaeofSoutheast
miningthespeciesrichnessonislands,withinthisframeworkthe Asia(incl.NewGuinea,BismarckandSolomonIslands),version
departureofsomearchipelagosfromtheirISARwouldbecaused 1.5. Available at: http://www.sphin-sea.unibas.ch/ (accessed
by other factors affecting the assembly of island faunas, and 15June2008).
thereforenestednesspatterns(seeWrightetal.,1998).Further Bennett, D.J. (2008) The ophionine wasps of Hawaii
studiesarerequiredtounderstandthecomplexitiesoftheinflu- (Hymenoptera: Ichneumonidae). Journal of Hymenoptera
enceofthesefactorsonthedegreeofdepartureofthearchipe- Research,17,1–43.
lagicpointfromtheISARandtheexactnatureoftherelationship Biber,E.(2002)Patternsof endemicextinctionsamongisland
betweensuchdepartureandnestedness,andalsotoestablishthe birdspecies.Ecography,25,661–676.
patternofdepartureofarchipelagicdatapointsfromtheircon- Blackburn,T.M.,Cassey,P.,Duncan,R.P.,Evans,K.L.&Gaston,
stituent ISARs for other types of insular system, including K.J.(2004)Avianextinctionsandmammalianintroductions
anthropogenichabitatislandsinfragmentedlandscapes. onoceanicislands.Science,305,1955–1958.
Borges, P.A.V. & Hortal, J. (2009) Time, area and isolation:
factors driving the diversification of Azorean arthropods.
ACKNOWLEDGEMENTS
JournalofBiogeography,36,178–191.
Wethankthreeanonymousrefereesfortheircriticalcomments. Borges,P.A.V.,Cunha,R.,Gabriel,R.,Martins,A.F.,Silva,L.&
A.M.C.S.wassupportedbyaPortugueseFCTgrant(SFRH/BD/ Vieira,V.(2005)Alistoftheterrestrialfauna(Gastropodaand
GlobalEcologyandBiogeography,19,527–540,©2010BlackwellPublishingLtd 535
A.M.C.Santosetal.
Arthropoda)andflorafromtheAzores.DirecçãoRegionalde Kohn,D.D.&Walsh,D.M.(1994)Plantspeciesrichness–the
AmbienteedoMardosAçores,Horta,Faial. effectof islandsizeandhabitatdiversity.Journalof Ecology,
Carvajal,A.&Adler,G.H.(2005)Biogeographyofmammalson 82,367–377.
tropical Pacific islands. Journal of Biogeography, 32, 1561– Kotze, D.J., Niemelä, J. & Nieminen, M. (2000) Colonization
1569. successof carabidbeetlesonBalticislands.JournalofBioge-
Conroy,C.J.,Demboski,J.R. & Cook,J.A. (1999) Mammalian ography,27,807–819.
biogeographyoftheAlexanderArchipelagoofAlaska:anorth Kryštufek,B.&Kletecˇki,E.(2007)Biogeographyof smallter-
temperatenestedfauna.JournalofBiogeography,26,343–352. restrial vertebrates on the Adriatic landbridge islands. Folia
Cowie,R.H.(1995)Variationinspeciesdiversityandshellshape Zoologica,56,225–234.
inHawaiianlandsnails:insituspeciationandecologicalrela- Lawton,J.H.(1996)Patternsinecology.Oikos,75,145–147.
tionships.Evolution,49,1191–1202. Lawton,J.H.(1999)Aretheregenerallawsinecology?Oikos,84,
Dapporto,L.&Cini,A.(2007)FaunalpatternsinTuscanarchi- 177–192.
pelagobutterflies:thedominantinfluenceisrecentgeography Lawton, J.H. (2000) Community ecology in a changing world.
not paleogeography. European Journal of Entomology, 104, EcologyInstitute,Oldendorf/Luhe,Germany.
497–503. Lomolino, M.V. & Davis, R. (1997) Biogeographic scale and
Dapporto,L. & Dennis,R.L.H. (2008) Species richness,rarity biodiversity of mountain forest mammals of western North
and endemicity on Italian offshore islands: complementary America.GlobalEcologyandBiogeographyLetters,6,57–76.
signals from island-focused and species-focused analyses. Losos, J.B. & Schluter, D. (2000) Analysis of an evolutionary
JournalofBiogeography,35,664–674. species–arearelationship.Nature,408,847–850.
Dengler, J. (2009) Which function describes the species–area MacArthur,R.H.&Wilson,E.O.(1963)Anequilibriumtheory
relationshipbest?Areviewandempiricalevaluation.Journal ofinsularzoogeography.Evolution,17,373–387.
ofBiogeography,36,728–744. MacArthur, R.H. & Wilson, E.O. (1967) The theory of island
Fattorini, S. (2002) Biogeography of the tenebrionid beetles biogeography.PrincetonUniversityPress,Princeton,NJ.
(Coleoptera,Tenebrionidae)ontheAegeanIslands(Greece). McCormack,G.(2007)CookIslandsbiodiversitydatabase.Cook
JournalofBiogeography,29,49–67. IslandsNaturalHeritageTrust,Rarotonga.Availableat:http://
Gaston, K.J. & Blackburn, T.M. (2000) Pattern and process in cookislands.bishopmuseum.org (accessed 17 November
macroecology.BlackwellScience,Oxford. 2007).
Givnish, T.J., Millam, K.C., Mast, A.R., Paterson, T.B., Theim, Millien-Parra,V. & Jaeger,J.-J. (1999) Island biogeography of
T.J.,Hipp,A.L.,Henss,J.M.,Smith,J.F.,Wood,K.R.&Sytsma, theJapaneseterrestrialmammalassemblages:anexampleofa
K.J. (2009) Origin, adaptive radiation and diversification relictfauna.JournalofBiogeography,26,959–972.
of the Hawaiian lobeliads (Asterales: Campanulaceae). Mylonas,M.A.(1982)Thezoogeographyandecologyoftheter-
Proceedings of the Royal Society B: Biological Sciences, 276, restrialmolluscsofCyclades[inGreek].PhDThesis,University
407–416. ofAthens,Athens,Greece.
Gray,J.S.,Ugland,K.I.&Lambshead,J.(2004)Onspeciesaccu- Niemelä,J.(1988)Habitatoccupancyofcarabidbeetlesonsmall
mulationandspecies–areacurves.GlobalEcologyandBioge- islands and the adjacent Åland mainland, SW Finland.
ography,13,567–568. AnnalesZoologiciFennici,25,121–131.
Guimarães,J.P.R.&Guimarães,P.(2006)Improvingtheanaly- Niemelä,J.,Haila,Y.,Ranta,E.,Tiainen,J.,Vepsäläinen,K.&Ås,
ses of nestedness for large sets of matrices. Environmental S.(1987)Distributionofcarabidbeetlesinfourborealarchi-
ModellingandSoftware,21,1512–1513. pelagoes.AnnalesZoologiciFennici,24,89–100.
Hamilton, A.M., Hartman, J.H. & Austin, C.C. (2009) Island Nishida, G.M. (2002) Hawaiian terrestrial arthropod checklist.
areaandspeciesdiversityinthesouthwestPacificOcean:isthe Availableat:http://www2.bishopmuseum.org/HBS/checklist/
lizardfaunaofVanuatudepauperate?Ecography,32,247–258. query.asp?grp(accessed22February2007).
He, F.L., Gaston, K.J., Connor, E.F. & Srivastava, D.S. (2005) Pandža,M.&Stancˇic´,Z.(2004)Secondcontributiontotheflora
Thelocal–regionalrelationship:immigration,extinction,and of the Kornati islands (Croatia). Natura Croatica, 13, 47–
scale.Ecology,86,360–365. 61.
Holyoak,G.A.,Turk,S.M.&Cameron,R.A.D.(2005)Landmol- Pandža,M.,Franjic´,J. & Škvorc,Z. (2002) The flora of some
luscsontheIslesof Scilly:speciesrichnessandcomposition uninhabitedŠibenikarchipelagoislands(Dalmatia,Croatia).
relatedtoarea.JournalofConchology,38,549–560. NaturaCroatica,11,367–385.
Hortal,J.,Triantis,K.A.,Meiri,S.,Thébault,E.&Sfenthourakis, Panitsa,M.&Tzanoudakis,D.(1998)Contributiontothestudy
S.(2009)Islandspeciesrichnessincreaseswithhabitatdiver- oftheGreekflora:floraandvegetationoftheEAegeanislands
sity.TheAmericanNaturalist,173,E205–E217. AgathonisiandPharmakonisi.Wildenowia,28,95–116.
Izquierdo,I.,Martín,J.L.,Zurita,N.&Arechavaleta,M.(2004) Ricklefs,R.E.(2007)Historyanddiversity:explorationsatthe
ListadeespeciessilvestresdeCanarias(hongos,plantasyani- intersectionofecologyandevolution.TheAmericanNatural-
malesterrestres),2ndedn.ConsejeríadePolíticaTerritorialy ist,170,S56–S70.
Medio Ambiente del Gobierno de Canarias, La Laguna, Ricklefs,R.E. (2008) Disintegration of the ecological commu-
Tenerife. nity.TheAmericanNaturalist,172,741–750.
536 GlobalEcologyandBiogeography,19,527–540,©2010BlackwellPublishingLtd