500
Abstract:FeedingbehaviouroftheMarbledTeal,Marmaronettaangustirostris,Mallard,Anasplatyrhynchos,Garganey,
Anasquerquedula,andFerruginousDuck,Aythyanyroca,intheGöksuDelta,Turkey,wascomparedfrom10Julyto
6August1995.Almostall individualsobservedwerepostbreedingadultsandjuveniles.MarbledTealfedclosesttothe surface(meandepthwithinthewatercolumn8.4cm),chieflybybilldipping(66%)andgleaning(14%).Garganeyfedata meandepthof 9.1 cm, mainlybybilldipping(57%)andneckdipping(35%).Mallardsfedatgreaterdepths(mean31.8cm), mainlybyupending(tipping 46%)andneckdipping(41%).FerruginousDucksfedatthegreatestdepths(mean38.4cm), chieflybydiving(76%).MarbledTealmovedmostfrequentlybetweenfeedingeventsandMallardsmovedleastfrequently. Asinpreviousstudiesofdabblingducks,thelargestspecies(Mallard)upendedmoreandfeddeeperinthewatercolumn. However,Mallardsusedshallowermicrohabitatsthansmallerdabblingducks.Dabblinganddiving duckguilds werenot discerniblein eitherhorizontal(feeding habitat)orvertical(feedingbehaviour)nichedimensions,and theMallardand FerruginousDuckwererelatedinbothdimensions.Nicheoverlapsbetweenspeciespairsalongthetwodimensionswere negativelycorrelated(r —0.71,P 0.12),supportingnichecomplementarity.
Résumé:Lecomportementalimentaireaétécomparechezquatre espècesde canards, laMarmaronettemarbrée, Marmaronettaangustirostris,leCanardcolvert,Anasplaryrhynchos,laSarcelled’été,Anasquerquedula,etleFuligule nyroca,Aythyanyroca,dansledeltadeGöksu,Turquie,du10juilletau6aoüt1995.Laplupartdescanardsobservesdtaient desadultespost-reproducteursetdesjuveniles.CesontlesMarmaronettesmarbréesqui senourrissentleplusensurface (profondeurmoyennede8,4cmdanslacolonne d’eau),surtoutparimmersiondubec(66%) et pargrappillage(14%).Les
Sarcellesd’étdsenourrissentauneprofondeurmoyennede9,1 cm,surtoutparimmersiondubec(57%) etimmersionducou
(35%).LesCanards colvertssenourrissentplusprofondément(31,8cmenmoyenne),surtoutparbasculement(46%)etpar
immersionducou(41%).LesFuligulesnyrocasontlescanardsquisenourrissentauxplusgrandesprofondeurs(38,4cmen moyenne),surtoutparplongée(76%).CesontlesMarrnaronettesmarbréesqui se déplacentleplussouvententrelessessions d’alimentationetlesCanards colvertsquise déplacentlemomssouvent.Commel’ontdémontréd’autresetudessurles canardsbarboteurs,laplusgrandeespèce(leCanardcolvert)aplusrecoursaubasculementet senourritplusprofondément dansIacolonne d’eau.Cependant,leCanardcolvertutiliselesmicrohabitatsd’eaumomsprofondequelescanardsbarboteurs pluspetits.Laguilde descanardsbarboteursetcelledescanardsplongeursne differentpasd’aprèslesdimensionshorizontales(habitatalimentaire)ouverticales(comportementalimentaire)deleursnichesrespectivesetlesCanardscolvertsetlesFuligulesnyrocasont semblablesdanslesdeuxdimensions.Les partiessuperposéesdesnichesdepairesd’espèceslelongdesdeuxdimensionssontencorrelationnegative(r =—0,71,P=0,12),ce quiappuiel’hypothèsedelacomplémentarité desniches.
[Traduitpar la Redaction]
Introduction
Various studies of habitat use by duck communities have shown niche separationbetweenspeciesinhorizontaldimen sions such as wetland size, vegetation characteristics, and water chemistry (Bengston 1971; Weller 1975; White and James1978;Amat 1982;Toft et al. 1982;Andersonand Ohmart1988;Monda andRatti1988;Bergan andSmith1989; Nummi and Pöysä 1993; Nudds et al. 1994). Some studies have addressedvertical partitioninginforagingmethodsand in the depths in the water columnat which feedingoccurs (DanellandSjoberg1982;Thomas1980,1982).However,few
.
A.J.Green.EstaciónBiológicadeDoflana,AvenidaMaria Luisas/n,Pabellóndel Perd,41013Sevilla,Spain(e-mail: ).
studies haveaddressedtheseissues simultaneouslytoidentify therelationshipbetweenhorizontalandvertical partitioningin duckcommunities.Amat(1984a)showed thattheRed-crested Pochard,Netta rufina, andCommonPochard,Aythya ferina, used differenthabitatsbutshared feedingmethodsinwinter, thenshared thesamehabitat butuseddistinct feedingmethods in spring.Poysä (1983a,1983b)found that partitioning of feedingmethods wasmoreimportant indabblingducks, while partitioningoffeedinghabitatswasmore importantindiving ducks. Pöysä (1987) found anonsignificantnegativecorrela tionbetweennichebreadth inthehorizontal(feedingsites)and vertical (feedingmethod)resourceaxesfordabblingducksand Amat(1984b)obtainedasimilarresultfordivingducks.These findingssupport nichecomplementarity(Schoener1974).
Inthisstudy Icomparethefeedingbehaviourinlatesum
meroffour species intheGöksu Delta, Turkey:theMarbled Teal, Marinaronettaangustirostris,the Mallard, Anas platy rhynchos,theGarganey,Anasquerquedula,andtheFerruginous
Green501
Duck,Aythya nyroca.Strongpartitioninginhabitatuseoccurs betweenspeciesinthiscommunity(Green1996’,Green1998). HereIconsidertherelationshipbetweenhorizontalpartition inginfeedinghabitatandverticalpartitioning infeedingbe haviour.Iaskhowfeedingdepthisrelatedtothedepthof microhabitatsexploitedbyeachspecies.Iconsiderwhether the dabblinganddivingducksformseparateguildsinthis community andItest fornichecomplementarity.Ialsocon sidertherelationshipbetweenfeedingbehaviourandbodysize indabblingducksinthelightoftheresultsofpreviousstudies suggestingthatlarger speciesfeedlowerinthewatercolumn (Poysa 1983a;Nudds1992).
TheMarbled TealandFenuginousDuckareglobally threatenedspecies(Collaretal.1994;TuckerandHeath1994; Green1996),andthisisthemostdetailedstudytodateoftheir foragingecology(butseeAmatandSonguer1982;Navarro andRobledano1995). TheGarganeyhasbeenstudiedat breedingsites(Poysä1983a,1983b,1986),butasfarasIam aware,thisisthefirstecologicalstudyofpostbreedingmigra tion.TheMarbledTealwaslongconsidereda typicalsurface- feeding duck,but isnowconsideredthelivingmemberofthe Aythyini(pochards)thatbearstheclosestresemblancetothe ancestor ofthistribe(Johnsgard1961; Livezey1996).Inthis studyIconsidertheMarbledTealtobea“dabblingduck,” usingtheterm intheecologicalratherthanthetaxonomic sense.
Studyarea
TheGdksuDeltaislocatedontheTurkishMediterraneancoast(see
DHKD 1992;Green 1996’;for adetaileddescriptionseeGreen
1998).ItisaRamsarsite(Frazier1996),a SpeciallyProtectedArea,
andanImportantBirdArea(MagninandYarar1997),andsupports
internationally important breeding populations of Marbled Teal
(Ca.50pairs) andFerruginousDucks(Ca.30pairs),aswellasabout
50pairsofMallards.Garganeysusethedeltaonpostbreedingpas
sage,beginningtoarriveinearlyJuly,withnumberspeakingat
3000—5000inlateAugust(Green1996’). Breedingandpostbreeding ducksarehighlyconcentrated atLakeAkgol,apermanent,well-. vegetated, freshwater tobrackisheutrophiclake(1400ha)with extensivebedsofPhragmitesaustralis, Typhasp.,Scirpus(=Schoeno plectu.s)litoralis,andPotamogetonpectinatus.
ThisstudywasconductedwhenMallardshadcompletedbreeding andlargenumbersofpostbreedersfromotherareaswerepresent. MarbledTealandFerruginousDuckswereattheendoftheirbreeding season,andsomebroodshadnotyetfledgedwhenthisstudybegan. However,ducklingswereexcludedfromthisstudyandalmostall individualsincludedwerepostbreedingadultsorjuveniles.
Methods
From10Julyto 6August1995inclusive,Akgoland otherwetlands inthedelta(seeGreen19961;fordetailssee Green1998)weresys tematicallysurveyedthree timesforAnatidae.AllareasofLakeAk golaccessiblebycanoeweresurveyedoverseveraldays.Fieldwork wasconductedfromfirstlight(06:45localtime)andwasusually completedby14:00.ThreecompletesurveysofLakeAkgolwere
A.J.Green.1996.ThesummerecologyoftheMarbled Teal Marmaronettaangustirostris,FerruginousDuckAythyanyroca andotherducksin theGöksuDelta,TurkeyinI995. Unpublishedreport,EstaciónBiologicadeDoñana,Seville, Spain.
conductedovertheperiods18—23July,24—28July,and30July—
5August,datesinclusive.
Duckswere observedwithatelescopeduringatotalof150hof
fieldwork.AllmaleMallardsandGarganeyswereineclipseplumage.
Breedingandeclipseplumages couldnotconsistentlybedistin guishedfortheMarbledTealorFerruginousDuck,althoughsome adultsofbothspecieswereclearlymolting.SomeflightlessGar
ganeyswereobserved.Owingtothedistancesandlightconditionsin whichobservationsweremade,itwasrarelypossible todistinguish thesexes,oradultsfromjuveniles,soallobservationswerepooledfor eachspecies.
Thebehaviourofeachduckfeedingatthetimeofinitialobserva tionwasrecordedasoneofthefollowing:dive;upend(tipping);neck dip(partorallofnecksubmerged);headdip(eyesubmerged);billdip (partorallofthebillsubmerged);glean(billheldintheplaneofthe watersurfacewithonlythelowermandiblesubmerged,straining itemsfromthewatersurface).Feedingbirdsweredescribedasmobile orstatic,mobilebirds beingthosemovingduringmore than50%of interdiveintervalsorintervalsbetweenotherforagingevents(upends, dips,etc).
ForeachindividualMallard,MarbledTeal,andFerruginousDuck seenfeeding,thefollowingdetailsofitspositionatthetimeofinitial observationwererecorded:distancetothenearestemergentvegeta tion(estimatedvisually inmetres);species ofemergentvegetation; presenceorabsenceofsubmergedmacrophytesatthe surfacelayer. Suspectedrepeatedobservationsofthe sameindividualonthe same daywerediscarded.Garganeyswereobservedinlargeflocksofupto severalhundred, andtheabovedatawere recordedforarandom sampleoffeedingindividuals.
Adipstickwasusedtomeasurewater depthatlocationswhere Marbled TealandFerruginousDuckswerefeeding.Althoughsome speciesofAythyinidonotalwaysfeedonthebottomwhendiving (TomeandWrubleski 1988),diving depthforfeedingwasassumed tobewaterdepth,sinceobservationsofbillcontentsandbehaviour uponsurfacing,combinedwithgrabandnetsampling,suggestedthat alldivingbirdswerefeedingonbenthicseedsorCharavulgaris, whichwasconfinedtothebottomofthewatercolumn.Feedingdepth forgleaningwasconsideredtobe1cmandforotherfeedingmethods wasestimatedusingmorphologicalmeasurementsfrommuseum specimens(Appendix)followingPoysa(l983a): forupendthiswas
3/5xbodylength+necklength;forneckdipitwasnecklength;for headdipitwasskulllength;andforbilldipitwasbilllength.
Nicheorganizationforfeedingbirdswascomparedforthehori zontal(habitatuse)andvertical(feedingbehaviour)axes.Percent nicheoverlapbetweenresourcecategories(Krebs1989)wasusedto calculateoverlapforeachspeciespair forthefollowinghabitat-use variables:nearest emergentvegetationtype;distancetonearestemer gentvegetation;presenceorabsenceofP.pectinatusinthesurface layer(Table1).Sincethesethreevariableswerenotindependent, mean nicheoverlapwascalculatedasanoverallmeasureofhonzontal partitioning(seePoysa1983b;MondaandRatti1988).Likewise, mean nicheoverlapforfeedingbehaviourwascalculatedusingfeed ingmethodanddepthdata(seeFigs.I and2).Clusteranalysiswas usedtoproducedendrogramsbasedoninterspecificdistancesinthe horizontalandverticaldimensions(UPGMAmethod;Krebs1989).
Results
Feedingbehaviour
Strong differencesinfeedingmethodwereobserved between duck species (Fig.1).MarbledTealfedmainlybymobilebill dippingandweretheonlyspeciesobservedgleaningandthe onlydabblingduckobserveddiving.Divetimesforjuvenile MarbledTealwererecordedon30July,whentheywerediving toadepthof49cm (n=17,range2—7s,4.9±1.5s(mean±
Table1.Percentage offeedingduck observationsbycategoryamongthreehabitat variablesinLakeAkgol.
Mallard(n=ll0) / Garganey
(n=70) / MarbledTeal
(n=24) / Ferruginous
Duck(n=41)
Nearestemergent
vegetation
Scirpuslitoralis / 38.2 / 33.9 / 70.8 / 45.0
Phragmitesoustralis / 26.4 / 8.1 / 25.0 / 35.0
Typhasp. / 35.5 / 58.1 / 4.2 / 20.0
Distancetoemergent
vegetation(m)
0 / 3.6 / 0 / 8.3 / 020.9 / 6.5 / 0 / 27.5
2—5 / 43.6 / 9.7 / 58.3 / 42.5
6—10 / 21.8 / 12.9 / 16.7 / 5.0
11—50 / 10.0 / 61.3 / 16.7 / 25.0
>50
P.pectinatusatsurface?
09.700
Yes / 95.4 / 94.3 / 75.0 / 75.6No / 4.6 / 5.7 / 25.0 / 24.4
Table2.Feeding-nicheoverlaps betweenduckspeciesfor horizontal(above thediagonal)andvertical(beneaththediagonal) resourceaxes.
neck-dippingoneswereprobably feedingatthelakebottom (Green,seefootnote1;Green1998). Lake depth atfeeding sitesusedbyMarbledTeal(n=16,range31—66cm,47.9±
10.3cm)andFerruginousDucks(n=9,range30—51cm,
MallardGarganey
Marbled
teal
Ferruginous
Duck
46.3 ±7.0cm)withinLakeAkgolwereverysimilar(two- tailedtwo-samplettest,t=0.44,22df,P=0.66).Maximum
MarbledTeal
Ferruginous
Duck
0.140.67
0.290.22
0.30
0.79
MarbledTealmovedmoreoftenbetween foragingevents thantheotherspecies did(Fig.1).Excluding unknowns,
92.0%offeedingMarbledTeal(n 50),45.5%ofFerruginous
Note:Horizontaloverlapsaremeansofthreedimensions(nearest emergentvegetation;distancetoemergentvegetation;presenceorabsence ofsubmergedvegetation).Verticaloverlapsaremeansoftwodimensions (feedingmethod;feedingdepth).
SD)).FerruginousDucksfedmainlybydiving,whilstMal lardsfedchieflybystaticupending(tipping)andneckdipping (Fig.1).Garganeysfedlargely bystaticbillandneckdipping (Fig. 1).Differencesinfeedingmethodamongspecieswere
highlysignificant(x2=429, 15df,P0.0001),even when
Ferruginous Ducks were excluded (x2=193,10df,P<
0.0001).
Asaresult ofthedifferencesinfeedingmethod,eachspe cieswas feedingatdifferentdepthswithinthewatercolumn (Fig.2).MarbledTealwereingestingfoodattheshallowest depths (range1—49cm,8.4±12.1cm),followedbyGar ganeys (range4—18cm,9.1±6.7cm)andthenMallards (range6—41 cm,31.8±10.3cm).FerruginousDuckswere feedingatthegreatestdepthsofall(range4—50cm,38.4±
17.6cm).Differencesinfeedingdepthamongspecieswere
highlysignificant(Kruskal—Wallis test, H=190,3df,P
0.0001).
WhereasMarbledTealandGarganeys inLakeAkgolfed mainly inthetophalfofthewatercolumn,amongstbedsof P.pectinatusthatfilledthesurfacelayer(Table1),Ferrugi nousDucksfedmainlyatthelakebottom.Mallardwereinter mediate, but most upending birds and many of the
Ducks(n=44),16.0%ofGarganeys(n=81),and only5.3%
ofMallards(n=113)weremobile.Thesedifferencesamong specieswerehighlysignificant(x2= 138,3df,P0.0001),
andremainedsimilarwhentheanalysiswasconfinedto those behaviours(bill,head,andneckdipping)exhibitedbyallfour species(90.5%ofMarbledTeal(n=42),53.8% ofFerrugi nousDucks(n=13),16.0%ofGarganeys(n=81),and 7.8% ofMallards(n=64)weremobile;x2=96,3df,P0.0001).
Nichesegregation
Therewasanonsignificantnegativecorrelationbetweenniche
overlapsshownbetweenspeciespairsalongthevertical(feed
ingbehaviour)andhorizontal(feedinghabitat)resourceaxes (Table2;r=—0.705,P=0.12). This correlationisstrongly affectedbythe MarbledTeal— Garganeypair,whichhadthe mostsimilarverticalnichesandtheleastsimilarhorizontal niches(Table 2).Clusteranalysisshowedthatalongtheverti calresourceaxis,theMarbled TealandGarganeywere very similar,whilsttheMallardandFerruginousDuckwererelated toeachother,butwithmuchlesssimilarity (Fig.3a).Along thehorizontalresource axis,theMallardandFerruginous Duckwere againgroupedtogether,butwith muchgreater similarity(Fig.3b).TheGarganeywasisolatedfromtheother three species inhorizontalniche space (Fig.3b).
Discussion
Markeddifferencesinfeedingmethodwerefoundbetweenthe
reen503
rig.1.ProportionsofMallards(n=134)(a),Gaganeys(n=Si)(b),MarbledTeal (n=58)(c),andFerruginousDucks(n=62)(d)using lifferentfeedingmethodsintheGöksuDelta.SeeMethodsforafulldescriptionoffeedingmethods.Birds“notrecorded”arethoseforwhich hedistinctionbetween“mobile”and“static”feedingwas notmade. DataexcludeducklingsbutincludeoneadultMarbled Teal(billdip)and neadultFerruginousDuck(headdip)accompanyingbroods.
(a)
U,
I
a0)
0
C
0
0
0
0.
(c)
0
0
0
Ca) U a)
(b)
U,
0
a,
a,
-a
a,
0
Ia-,
0
0Static
Mobile
NotRecorded
(d)
Cl,
0
a,
0)
a,
a, C) a,
Glean Bill HeadNeckUpendDiveGlean Bill HeadNeckUpendDive dip dipdipdipdip dip
©1998NRCCanada
Fig.2.FeedingdepthsusedbyduckspeciesintheGöksuDelta.SeeFig.1forsamplesizes.These figuresrefertothedepthwithinthewater columnatwhichingestionoccurred,whichwasoftenmuchlessthenthetotaldepthofthewatercolumn (seeMethods).
0.
Percentage of Birds
CMailard
0Garganey
JFerruginousDuck
•MarbledTeal
fourduckspeciesstudied,resultinginclearverticalseparation offeedingactivityinthewatercolumn.Thiscomplements differences inhabitatuseanddiet(Green,seefootnote 1; Green1998).Marbled Tealfedclosesttothewatersurface, whichconfirmsthatthisspeciesisadabblingduckinecologi calterms, despiteitstaxonomicpositionintheAythyini (Livezey 1996).NotsurprisinglyforanAythyaspecies,Per ruginousDucksfedatthegreatestdepths,althoughtheiraver age feedingdepthwasonly21%morethanthatofMallards. Indeed, MallardsshowedgreateroverlapwithFerruginous Ducksin thedepthrangesutilized(23%accordingto Fig.2) thanwith theothertwo dabblingduckspecies(14%with the MarbledTealand9%with theGarganey).Similarly,theMar bledTealandGarganeyshowed muchmoreoverlapwitheach other(67%)thanwitheitheroftheothertwo species.
Amongstpreviousstudiesofdabblingducks,largerspecies
(e.g.,Mallard,NorthernPintail,Anasacute)fedlowerinthe watercolumnthansmallerspecies(e.g.,GarganeyandGreen-
wingedTeal, Anascrecca), partlybecause theyshowed a greatertendency toupend(Danelland Sjoberg1982;Thomas
1980,1982;POysa1983a,1983b;Nudds1992).Theresultsof thepresentstudy,whichincludedthepreviouslyunstudied, relativelysmallMarbledTeal,confirmthispattern.Ittherefore seemstobeafairlyuniversal rulethatlargerdabblingducks upendmore,andPUysd(1983a)proposedthatthisisbecause itismore energeticallyefficientforthemtodoso,sincethey canremainsubmerged longer.Heproposedthatsmallerspe ciesare selectedtofeedmainlyatthe surface,partlybecause theirlowerenergyrequirementsareeasiertomeetwithout upending.Alternatively,Iproposethatthisupending rulemay beexplained bydiminishingbenefits ofupendingforsmaller species,owingtothesmallerincreaseinaccessibledepths(see Fig.8inThomas1982andFig.IinPoysa1983a).
Thedifferencesinfeedingmethodobservedinthisstudy
wereconnectedwithdifferencesinfoodselection,althoughall fourspeciesfedmainlywithindensebedsofP.pectinatus (Table1).GleaningMarbledTealwere feedingondense, floatingconcentrationsofrelativelysmallS. litoralisseeds, whilstthosefeeding inotherwaysweretakingmainlyinsects. MallardswerefeedingmainlyonthebasalpartsofP. pecti natusandoncharophytes(Charavulgaris)(Green,see foot note1).Garganeyswereprobablyconsumingmainlytheupper parts of P.pectinatus, whereas Ferruginous Ducks were probablydivingtofeedon theseedsofP.pectinatus,Potamo getonpanormitanus,and S.litoralisabundantinthebenthos, andtakingP.pectinatusseedswhenfeedingbyothermethods (CrampandSimmons1977;Green,seefootnote 1;Kear and Wildfowland WetlandsTrust1998).
Owingtotheirphysicaladvantagewhenfeedingatgreater depths,largerdabblingduckspeciesmayoftenexploitdeeper habitats(e.g.,EulissandHarris1987).However, Poysa (1983a) foundonlyaweak,nonsignificantpositivecorrelation betweennecklengthandthemeandepthoffeedinghabitat, andWhiteandJames(1978)foundthatMottled Ducks, Anasfulvigula, usedshallower habitatsthansixsmallerAnas species. IntheGoksuDelta,therewasanegativerelationship betweenfeedingdepthanddepthofthehabitatused,since MallardsmademostuseofshallowerpartsofLakeAkgOland MarbledTealmademostuseofdeeperparts(Green,seefoot note1;Green1998).Theresultsofthecurrentstudyshowthat therewasnodifferenceinlakedepthbetween areasselected forforagingbyMarbledTealandFerruginousDucks,inspite ofthemajordifferenceinfeedingdepth.Thiswaslargelya consequenceofthedensebedsofP.pectinatusin thesurface layerof relativelydeeppartsofthelake,providingplantand invertebrate foodforducksclosetothewatersurface.Ina lake where submerged vegetation wasonly available to wardsthebottomofthewatercolumn,Poysa(l983a)found
Fig.3.Dendrogramsshowingsimilarity(Euclideandistances)betweenduckspeciesinfeedingbehaviour(verticalpartitioning)(a)and feedinghabitat(horizontalpartitioning)(b)intheGöksuDelta.FD,FerruginousDuck;Gar,Garganey;Mal,Mallard;MT,MarbledTeal.
(a)
1.2
1.1
1-
O 0.8-
a,
0)
a,.-
C
J 0.6 -
0.5-
0.4
(b)
0.55-
0.5-
0
0.45-
)0.4-
Ca
.5 035-
GarMTEDMal
thatMallardsfedindeeperpartsthanGarganeys.Similarly,in sitesinSpainwheresubmergedvegetationwasrareandducks foragedin thesediments,MarbledTealfedinshallower areas than largerduckssuchasMallards(A.J.Green,unpublished data).
Unfortunately,datacollectedinthisstudyaretoolimited toreflectthevariationinfeedingbehaviourshownbyducks inresponseto differencesinseason,sex,age,depthfluctua tion,andeventimeofday(Danell and Sjoberg1982;Thomas
1982;Amat1984a;Jacobsen 1991;Battetal.1992).Some of theapparentdifferencesamongspeciescouldhavebeenpartly
theresultofdifferencesin therelativeproportionofsexesand
age-classes,althoughthesame istrueofthemajorityofcom parativestudiesofduckecology.DatafromSpainandMo roccoconfirmthatMarbledTealfeedpredominantlyatornear thesurfacethroughouttheannualcycle,andthatupendingand divingarerarelyobserved(NavarroandRobledano1995;A.J. Green,unpublisheddata).Inthecurrentstudy,divingwasre corded innewlyfledgedjuvenileMarbledTealbutnotin adults.Divingwasalsoobservedinclassha(LarsonandTaber
1980)ducklings(Green,seefootnote1).Although these observationsareofnote,giventhe classificationofMarbled TealwithintheAythyini,divingisoccasionallyobservedina broadrangeofAnatini(KearandJohnsgard1968;Brodsky andWeatherhead1985).Inprevious studies,Ferruginous Ducksfedmainlyby diving,butupending, headdipping,or billdippingwasfrequentlyobserved(CrampandSimmons
1977;AmatandSoriguer1982;Dvoraketal.19962).Pöysä
(1983a,1983b)foundthatfromMaytoAugust,Garganeysfed mainlybygleaning(39%),bill dipping(38%),andneck dipping(17%).
Evenwhenusingthesamebasicfeedingmethod,duckspe ciesdifferinthedurationoffeedingeventsandspeedofmove ment(Eadieetal.1979).Thefailuretorecordthespeedof movementoffeedingbirdsinthisstudyunderemphasizesthe
2 M.Dvorak,E.Nemeth,S.Tabbich,M.Rössler,andK.Busse.
1996.Bestand,Okologie undHabitatwahischilfbewohnender
VogelArteninderNaturzonedesNationalparks
Neusiedlersee-Seewinkel.Unpublishedreport,BirdLifeAustria,
Vienna.
©1998NRCCanada
extenttowhichMarbledTealweremoreactivefeedersthan MallardsorGarganeys.Whereasthelattertwospecieswere almostcompletelystationarybetweenforagingevents,Mar bledTealtypicallyswamalmostcontinuouslywiththeneck outstretchedandtheheadturneddownatanangleofCa.45°, occasionallydippingtheirbillorheadrapidlyintothewater totakeinvertebratesnearthesurface(Green,seefootnote1).
WhereasPoysa(1983b)foundaclearseparationbetween divingduckanddabblingduckguildsinbothfeedinghabitat andfeedingmethods,thiswasnotthecaseinthecurrentstudy. TheFerruginousDuckwasnotecologicallyisolatedfromdab blingducksineitherhabitatuseorfeedingbehaviour,andwas relatedto theMallardalongbothaxes ofnichespace(Fig.3). Furtherworkisrequiredtoestablishifthesimilaritybetween thesetwo species is confined to the postbreeding period (e.g., owingtosimilarrequirementsduringmoult).Norwas thereanyseparationbetweenAnatiniandAythyinispeciesin thisstudy.Similarly,MondaandRatti(1988)foundthatRing- neckedDuck,Aythya collaris,broodsshowedgreaterniche overlapwithdabblingducksthanwithotherdivingducks.This shows theimportanceofstudyingthetwogroupssimultane ously andbringsintoquestiontheassumptionoftenmade (e.g.,Nuddsetal.1981;Nudds1983; NuddsandWickett
1994)thatdabblinganddivingducksformtwoneatguildsand thatinterspecificinteractionswithinthemaremuchmoreim portantthanthosebetweenthem.
Whenmore thanoneresource dimensionisconsidered, competition theorypredictsthatspeciessimilaralongonedi mensionshoulddifferfromeachotheralonganotherdimen sion(nichecomplementarity;Schoener1974;Nudds1992). Pdysa(193b)foundanonsignificantnegativecorrelationbe tweennicheoverlapsalongthehorizontal(feedinghabitat)and vertical(feeding methods)axesinacommunityofducks, grebes,and coots.Likewise,inthepresent study,therewas a negativecorrelationbetweennicheoverlapsalongthehorizon taland verticalaxes,providingTypeIIevidenceforcompeti tionbetweenduckspecies(seeNudds1992).Anotherformof nichecomplementarity canoccurifspeciesfeeding insimilar ways differindietaryspecialization.Differencesinlocomo tionwhenfeedingareconnectedwithdifferencesinselection ofpreyitems(Eadieetal.1979;NuddsandBowlby1984),and therewasanonsignificantnegativecorrelation(r —0.278,P=
0.59)betweennicheoverlapinfeedingmethod/depth(from
Table2)andoverlapinfeedingmobility(from Fig.1).The MarbledTealandGarganeyweremostsimilarinfeeding methodanddepth, yet theMarbledTealwasmuchmoremo bile(Fig. 1).
Thisstudy addstoagrowingbodyofevidencethatinter specificinteractionshaveanimportantinfluenceonnicheor ganizationinduckcommunities(Nudds1992).Theevidence itprovides fornichecomplementarityisnotstatisticallysig nificant,butthisisalmostinevitableinacommunityoffour species.Lnanycase,instudieswhereanalysesofoverlaps amongspeciespairsaresignificant(e.g.,Eadieetal.1979; NuddsandWickett1994),thevalidityofthesestatisticsis questionable, sincethesamedataforeachspeciesarerepre sentedrepeatedly,causingnon-independence.Morestudiesof differentduckcommunitiesarerequiredtoassesswhether nichecomplementarityalongthehorizontalandverticalaxes isageneralrule.As far asIamaware, thisisthefirststudyof anessentiallypostbreedingcommunity.Garganeyswereonly
presentonpostbreedingmigration,butinsuchrelativelyhigh numbersthattheymaypotentiallyimposenicheshifts onthe speciesbreedingin theGdksuDelta(seePdysä1986).Further researchisrequiredtocomparenicheorganization during breeding andpostbreedingperiodsinthis andotherMediter raneanduck communities.
Acknowledgements
Thisprojectwasfinancedbyapostdoctoralgrantfromthe SpanishMinistryofEducationandScienceundertheprogram “Estancias TemporalesdeCientIficosyTecnólogosExtran jerosenEspafla”;theConsellerladeMedioAmbiente,Junta deAndalucIa;theMinistryofAgriculture,NatureManage mentandFisheries,theNetherlands;andtheRoyalSocietyfor theProtectionofBirds.Thefieldworkwas conductedaspart oftheNetherlands’contributiontotheEuropeanCommunity fundedGdksuDeltaproject(TurkishSocietyfortheProtection ofNature,DHKD)onintegratedmanagement,with theper missionoftheAuthorityfortheProtection ofSpecialAreas. TheDHKDhelpedto organize theproject. TheWildfowland WetlandsTrustprovidedessentialequipment.Specialthanks areextendedtoGuneinAydemir, VincentvandenBerk, KeremBoyla,HasanGunen,BorjaHeredia,FernandoHiraldo, SühendanKarauz,RosarioPintos,LaurenceRose, Gurdoar Sarigul,JaninevanVessem,andMuratYarar. Helpfulcom mentsonearlierdraftsofthemanuscriptwereprovidedby Juan A.Amat,Vincent vanden Berk,JoanCaries Dolz, Jean-François Giroux,TommyMichot,HannuPoysa, andan anonymousreviewer.HannuPdysaprovidedthe morpho metricmeasurementsforMallardsandGarganeys.Datafor MarbledTealandFerruginousDucksweretakenfromthecol lectionoftheEstaciónBiológicadeDoñana.
References
Amat,J.A.1982.ThenestingbiologyofducksintheMarismasofthe
Guadalquivir, south-westernSpain.Wildfowl,33:94—104.
Amat,J.A.I984a.Ecologicalsegregationbetween Red-crestedPo chardNetta rufinaandPochardAythyaferinainafluctuating en vironment.Ardea,72:229—233.
Amat,J.A.1984b.Interaccionesentrelospatosbuceadoresenuna
lagunameridionalespaflola.DoflanaActaVertebrata,11:
105—123.
Amat,J,A.,andSoriguer,R.C.1982. Datossobreseleccióndehabitat yecologIaalimenticiadelPorrónPardo(Aythyanyroca).Donana ActaVertebr.9:388—394.
Anderson,B.H.,and Ohmart, R.D.1988.Structureofthewinterduck communityonthelowerColoradoRiver:patternsandprocesses. InWaterfowlinwinter.EditedbyMW.Weller.Universityof MinnesotaPress,Minneapolis.pp.191—236.
Bait, B.D.J., Afton, A.D., Anderson, MG., Ankney, CD., Johnson,D.H.,Kadlec,J.A.,andKrapu,G.L.(Editors).1992.Ecol ogyandmanagementofbreedingwaterfowl.UniversityofMinnesota Press,MinneapolisandLondon.
Bengston,S.A.1971.HabitatselectionofduckbroodsinLakeMy
vatnArea,North-EastIceland.OrnisScand.2:17—26.
Bergan,J.F.,andSmith,L.M.1989.Differentialhabitatusebydiving duckswinteringinSouthCarolina.J.Wildl.Manage.53:1117—1126. Brodsky,L.M.,andWeatherhead,P.J.1985.Divingbywintering BlackDucks:anassessmentofatypical foraging.Wildfowl,36:
72—76.
Collar,N.J.,Crosby, M.J.,andStattersfield,A.J.1994.Birdsto
watch 2.Theworldlistofthreatenedbirds.BirdLifeConservation
Ser.No.4.BirdLifeInternational,Cambridge,England.
Cramp,S.,andSimmons,K.E.L.1977.Handbookofthebirdsof Europe,MiddleEastandNorthAfrica.Vol.1.OxfordUniversity Press,Oxford.
Danefi,K.,andSjoberg,K.1982.Seasonalanddielchangesinthe feedingbehaviour ofsomedabblingduckspeciesonabreeding lakeinnorthernSweden.OrnisScand.13:129—134.
DHKD.1992.TowardsintegratedmanagementintheGdksuDelta,a
ProtectedSpecialAreainTurkey,feasibilityreport,September
1992.TurkishSocietyfortheProtection ofNature(DHKD),
Istanbul.
Eadie,J.McA.,Nudds,T.D.,andAnkney,CD. 1979.Quantifying interspecificvariationinforagingbehaviorofsyntopicAnas(Ana tidae).Can.J.Zool.57:412—415.
Euliss,N.H.,andHarris,SW. 1987.Feedingecologyofnorthern pintails andgreen-wingedtealwinteringinCalifornia.J.Wildi. Manage.51: 724—732.
Frazier,S.1996.Anoverviewoftheworld’sRamsar sites.Wetlands InternationalPubl.No.39.WetlandsInternational,Slimbridge,U.K. Green,A.J.1993.ThestatusandconservationoftheMarbledTealMar maronettaangustirostris.IWRBSpec.PubI.No.23,International
WaterfowlandWetlandsResearchBureau,Slimbridge,U.K.
Green,A,I.1996.Analyses ofgloballythreatenedAnatidaeinrela tiontothreats,distribution,migrationpatternsandhabitatuse. Conserv.Biol.10:1435—1445.
Green,A.J.1998.Habitat selectionbytheMarbledTeal (Marmaro nettaangustirostris),FerruginousDuck(Aythyanyroca)andother ducksintheGöksuDelta,Turkeyinlatesummer.Rev.Ecol.Terre Vie.Inpress.
Jacobsen, O.W.1991.FeedingbehaviourofbreedingWigeonAnas penelopeinrelation toseasonal emergenceandswarmingbehav jourofchironomids. Ardea,79:409—418.
Johnsgard,P.A.1961.ThesystematicpositionoftheMarbledTeal.
Bull.Br.Ornithol.Club,81:37—43.
Kear,J.,andJohnsgard,P.A.1968.Foragingdivesby surface- feedingducks.WilsonBull.80:231.
Kear,J.,andWildfowlandWetlandsTrust(Editors).1998.Bird families oftheworki:ducks,geese,swansandscreamers.Oxford UniversityPress,Oxford.Inpress.
Krebs,C.J.1989.Ecologicalmethodology.HarperandRow,New
York.
Larson,J.S.,andTaber,R.D.1980.Criteriaofsexandage.InWildlife
managementtechniquesmanual.4thed.EditedbyS.D.Schemnitz.
TheWildlifeSociety,Washington,D.C.pp.143—202.
Livezey,B.C.1996.Aphylogeneticanalysisofmodernpochards
(Anatidae:Aythyini).Auk,113:74—93.
Magnin,G.,andYarar,M.1997.ImportantbirdareasinTurkey.
DHKD(TurkishSocietyfortheProtectionof Nature),Istanbul.
Monda,M.J.,andRatti,J.T.1988.Nicheoverlapandhabitat useby sympatncduckbroodsineasternWashington.I.WildI.Manage.
52:95—103.
Navarro,ID,,andRobledano,F.(Coordinators).1995.LaCerceta
PardillaMarmaronettaangustirostrisenEspana.ICONA-MAPA
ColecciónTécnica,InstitutoNacionalparaIaCouservacidndela
Naturaleza,Madrid.
Nudds,T.D.1983.Variationinrichness,evennessanddiversityin divinganddabblingduckguildsinprairiepotholehabitats.Can. J. Zool.61:1547—1550.
Nudds,T.D.1992.Patternsinbreedingwaterfowlcommunities.In Ecologyandmanagementofbreedingwaterfowl.Edited by B.D.J. Batt, A.D. Afton, M.G. Anderson, CD. Ankney, D.H.Johnson, l.A.Kadlec,andG.L.Krapu.UniversityofMinne sota Press,MinneapolisandLondon.pp.540—567.
Nudds,T.D.,andBowlby,J.N.1984.Predator-preysizerelationships inNorthAmericandabblingducks.Can.J.Zool.62:2002—2008. Nudds,T.D.,andWickett,R.G.1994.Bodysizeandseasonalcoex istenceof NorthAmericandabblingducks.Can.I. Zool.72:
779—782.
Nudds,T.D.,Abraham,K.F.,Ankney,CD., andTebbel,P.D.1981.
Aresizegapsindabbling-andwading-birdarraysreal?Am.Nat.
118:549—553.
Nudds,T.D.,Sjoberg,K.,andLundberg,P.1994.Ecomorphological relationshipsbetweenPalearcticdabblingducksonBalticcoastal wetlandsandacomparisonwith theNearctic.Oikos, 69:295—303. Nummi,P.,andPoysh,H.1993.Habitatassociationsofducksduring
phasesofthebreedingseason.Ecography,16:319—328.
Pdysa,H.1983a.Morphology-mediatednicheorganizationinaguild ofdabblingducks.OrnisScand.14:317-326.
Poysä,H.I983b.Resourceutilizationpatternandguildstructureina waterfowlcommunity.Oikos,40:295—307.
Pdysa,H.1986.Foragingnicheshiftsinmultispeciesdabblingduck (Anusspp.)feedinggroups:harmfulandbeneficialinteractions betweenspecies.OrnisScand.17:333—346.
Pdysh,H.1987.Ecologyofforagingbehaviourindabblingducks (Anasspp.). University of Joensuu Publications in Sciences No.10,Universityof Joensuu,Joensuu,Finland.
Schoener,T.W.1974.Resourcepartitioninginecologicalcommuni
ties.Science(Washington,D.C.),185:27—39.
Thomas,G.J.1980.TheecologyofbreedingwaterfowlattheOuse
Washes,England.Wildfowl,31:73—88.
Thomas,G.J.1982.Autumnandwinterfeedingecologyofwaterfowlat theOuseWashes,England.].Zool.(1965—1984),197:131—172.
Toft,C.A., Trauger,D.L,,andMurdy,H.W. 1982.Testsforspecies
interactions: breedingphenologyand habitatuseinsubarctic ducks.Am.Nat.120:586—613.
Tome,M.W.,andWrubleski,D.A.1988.Underwater foragingbe
haviorofcanvasbacks,lesserscaups,andruddyducks.Condor,
90:168—172.
Tucker,G.M.,andHeath,M.F.(Editors).1994.BirdsinEurope:their conservationstatus.BirdLife ConservationSer.No.3,BirdLife International,Cambridge,U.K.
Weller,M.W.1975. HabitatselectionbywaterfowlofArgentineIsla
Grande.WilsonBull.87:83—90.
White,D.H.,andJames,D.1978. Differentialuseoffreshwateren
vironmentsbywinteringwaterfowlofcoastalTexas.WilsonBull.
90:99—111.
Appendix
TableAl.Meanmorphologicalmeasurements(cm)usedtocalculatefeedingdepths. ValuesfortheMallardandGarganeyarethoseusedbyPdysh(1983a).
Mallard(n 57) / Garganey
(n=21) / MarbledTeal
(n=8) / Ferruginous
Duck(11 8)
Bodylength / 19.5 / 15.5 / 16.0 / 16.0
Necklength / 29.1 / 18.2 / 19.8 / 19.2
Skulllength / 11.5 / 7.9 / 9.0 / 8.6
Billlength / 5.7 / 3.8 / 4.5 / 3.9
i1998NRCCanada