OPINION

Dispersalofaquaticorganismsbywaterbirds:areview ofpast researchand prioritiesforfuturestudies

JORDI FIGUEROLA andANDY J. GREEN

DepartmentofAppliedBiology,Estacio´nBiolo´gicadeDon˜ana,CSIC,Avda.Mar´ıaLuisas/n,Sevilla,Spain

SUMMARY

1. Inlandwetlandsconstituteecological islandsofaquatichabitatoften isolatedbyhuge areas ofnon-suitableterrestrialhabitats.Althoughmost aquaticorganismslackthe capacitytodispersebythemselvestoneighbouringcatchments,many species present widespreaddistributionsconsistentwith frequentdispersalbymigratorywaterbirds.

2. Aliteraturereview indicatesthat bird-mediatedpassivetransportofpropagules of aquaticinvertebratesand plantsisafrequentprocess inthefield,atleastatalocalscale. Bothendozoochory(internaltransport)and ectozoochory(externaltransport)are importantprocesses.

3. Thecharacteristicsofthedispersedand thedisperserspecies that facilitate such transportremainlargely uninvestigated,but asmall propagulesizetends tofavour dispersalbyboth internaland externaltransport.

4. Wereview theinformationcurrentlyavailableontheprocessesofwaterbird-mediated dispersal,establishingthelimits ofcurrentknowledgeand highlightingproblemswith researchmethodsused inpreviousstudies.Wealsoidentifystudiesrequiredinthefuture tofurtherour understandingoftheroleofsuch dispersalinaquaticecology.

Keywords:dispersalcapacity, seed dispersal,eggdispersal,habitatcolonisation,community composition

Introduction

Continentalwetlandsusuallyconsist ofisolatedunits oftemporaryorpermanentlyfloodedareasina‘sea’of habitat not suitable for aquatic organisms. Many aquatic plantsand invertebrateslackthemobility necessary totravel directlyfromonecatchment to another, tocolonise newareas and todisperseto neighbouring water bodies indifferentcatchments. Despite thisapparentisolationoffreshwaterhabitats, many aquatic plant and invertebratespecies have widespreaddistributions (Good, 1953;Raven, 1963; Sculthorpe,1967;BrownGibson,1983;WCMC,1998), spanningseveral continentsinsome cases. Although

some recent genetic studieshave demonstratedthat some‘species’areinfactmorphologically similar complexesofsibling species (HebertWilson, 1994), others have identified widelydistributedspecies (HebertFinston, 1996). However,there arealso numerousaquaticorganisms with very restricted distributions(Frey,1987;Stemberger,1995).Onemajor reason forsuch differencesisthat widespreadspecies have agreater capacity todisperseand canthereby colonise new areas and maintaingene flow between differentareas(Bohonak, 1999;Clobert etal., 2001).

Passive dispersalcapacity isakey trait explaining patternsofdistributionand communitycomposition ofnon-mobileorganisms(Jenkins Buikema, 1998).

For example, Primack Miao (1992) showedthat

Correspondence: J.Figuerola,DepartmentofAppliedBiology, Estacio´nBiolo´gicadeDon˜ana, CSIC,Avda. Mar´ıa Luisa s/n,

E-41013Sevilla,Spain. E-mail:

limiteddispersalcapacity canrestrict thedistribution ofmany terrestrialspeciesofplants.Similarly,despite the high potential for seed dispersal by water

483

currents,thesame occurs inZosteramarina,amarine macrophyte,80% ofseedlingswere foundless than

5maway fromthemotherplant (Orth,Luckenbach& Moore, 1994). Dispersal capacity is also of vital importance in determining responses to climatic changes(Davis, 1989;Graham,TurnerDale,1990). Althoughthereismuch informationavailableabout animal-mediateddispersal in terrestrial ecosystems (seereviewsinSorensen,1986;Traveset, 1998),relat- ivelylittleisknownabouttheprocessesaffectinglong- distancedispersal betweenaquatichabitats,where

waterbirdspresumablyareanimportantvector.

Many waterbirdsundertakelong migratory jour- neys from their breedingareas at extremelatitudes towardsmore temperateareas duringthewinter(del Hoyo, ElliottSargatal, 1992).Waterbirdshave long been consideredamajor disperserofaquaticorgan- isms, transportingplant and invertebratepropagules intheir guts (endozoochoryorinternaldispersal)or attached totheirbodies (ectozoochoryorexternal dispersal) (Darwin, 1859; Ridley, 1930). However, mostevidence forsuch dispersalbywaterbirdsis anecdotaland detailedquantitativeinvestigations on thepatternsofsuch transportareveryscarce(seeour review below). Recent indirectevidencesupporting the role of long-distance dispersal by waterbirds comes from studiesofthe genetic populationstruc- ture ofdifferentaquaticspecies. Mostofthese studies have concludedthat geographicaldistanceisusually unrelatedtothegenetic distancebetweenpopulations (HebertFinston, 1996; Hollingsworth,Preston& Gornall, 1996; Vanoverbeke & DeMeester, 1997; Mader, van Vierssen Schwenk, 1998; Freeland, Noble Okamura,2000).Oneexceptionisastudyof thefennelpondweedPotamogetonpectinatusinEurope (Mader etal.,1998).However,inthis casetheassoci- ation betweengenetic and geographicdistanceswas onlysignificantforpondsnotvisited byswans during migration(consideredthemain potentialdisperserof P.pectinatusinthe studyarea). Thegenetic distribu- tion of Daphnia laevisreflects the major waterfowl (ducks, geese and swans)flyways inNorthAmerica, with higher similarity between populations in a North-South than anEast-West direction(Taylor, Finston Hebert, 1998).Similarly, the genetic distri- butionofthe bryozoanCristatellamucedoinnorthern Europefollows the major waterfowlflyway in this region (Freelandetal.,2000). However, thelackof associationbetweengenetic and geographicdistances

in aquaticorganismsdoes not constitutedirect evidence supportingbird-mediateddispersal,but ratherindicatesthat populationsare not at equilib- rium, orthatcolonisation events areindependentof geographicdistance.Thismaybebecause dispersalis more frequentinthedirectionofbird migration or betweenmajor stopoverareas, but alternatively may be because loss of alleles as a result of drift is not balancedbytheir replacementbygene flowbetween populationsinaregion (McCauley,1993).

Inthis paperwe review the researchconductedto date onwaterbird-mediateddispersal ofaquatic organisms,highlighting the series of processes involved and assessing thelikelyfrequency ofsuch dispersalinthefield.Weconcentrateonquantitative and semiquantitative studiesratherthan anecdotal observations.Weidentify pitfalls and limitationsof existing studies and methodologies and highlight some ofthe key questionsthat remainunanswered in this field. Finally, we identifysome prioritiesfor futureresearch ontheprocessesinvolvedinthe dispersal of aquatic organisms by birds. In this review, wedonotdealwith thequestionofhow well apropaguleislikely tocompetetoestablishitselfin anaquaticcommunityonceithasbeendispersedbya waterbird(e.g.questionssuch aswhetherthat prop- agule will bewell adaptedtothe new environment, whether itcancompetewith conspecifics already present,or whetherthe arrival of a new genotype wouldbedetectablegiven thesizeofexisting seed or propagulebanks inthewetland).

Internaltransportstep 1:dowaterbirds consumepropagules?

Tofacilitate internaldispersal,waterbirdsfirstneedto ingest propagulesofaquaticorganisms.Such inges- tion may be voluntarywhen the propagulesthem- selves constitutethe foragingresourceexploitedby theanimals,orinvoluntary when propagulesare ingestedwith leaves, invertebratesorother preferred fooditems.Therelative frequencyofdispersalmaybe higherfollowinginvoluntarydigestion,asseed-eating birds can destroymost seeds ingested,whereas ingestionofseeds byconsumersofleaves orinsects may have less impact on their viability. Thus, in a given bird community,some species may belegitim- ate seed dispersers and others seed predators (Jordano, 1992;Traveset, 1998).

©2002Blackwell ScienceLtd,FreshwaterBiology,47,483–494

Diet studies of migratory waterfowl (Anatidae) have shown thattheyconsumelargequantitiesof propagulesofaquaticorganismsfrom awide variety ofspecies. Thisisespeciallythecaseforduck species (see,e.g.CrampSimmons,1977;Thomas, 1982;Batt etal.,1992).Thefilter-feedingmethodtypicallyused byducks isrelativelyunselective,and they typically ingest variousclasses offooditems atagiven time.It isthus difficult aprioritoassess whetherseeds and other propagulesareingestedvoluntarilyornot.

Coots (Fulica sp.), flamingoes (Phoenicopterussp.) and other groupsofwaterbirdsalsoconsumeimport- ant quantities of seeds and other propagules (e.g. Cramp Simmons,1977). Waders(Charadridaeand Scolopacidae)are a particularlyinteresting group because oftheirlong-distancemigrationsandcapacity forlong-distance dispersal.Many species consume invertebratesthat maycontain resting eggs(CrampSimmons,1983),and some species often ingest seeds ofsaltmarshplants(Pe´rez-Hurtado,Goss-Custard& Garcia, 1997).

Dietstudiesofwaterbirdsgenerally have anaut- ecological perspectivefocussingonthe nutritionalor habitatrequirementsof the waterbirdspecies and, while giving information about theingestionof propagules,theydonotconsiderwhetherornotsuch propagulessurvivedigestion.Moststudiesarecarried out onoesophagusorgizzardcontents,but even diet studiesviafaecalanalysesdonotconsiderhow many propagules survive ingestion intact (e.g. Pe´rez- Hurtadoetal.,1997;Green Selva2000).

Internaltransportstep 2:can propagules survivedigestionbywaterbirds?

Various experimentson the survivalofinvertebrate and plant propagulesafter digestionby waterbirds have been carriedout by feeding alarge number of propagules to captive-reared ducks or waders (Table 1).Droppingswere generallycollected fora variablenumberof hoursafter consumptionof the seeds(usually24horless,butProctor, 1968,included droppingscollected atleast 10days after thefeeding of the propagules),and the intact propaguleswere thenculturedinthelaboratorytoassesstheirviability. Theseexperimentshave shownthat arange ofseeds, oospores,phytoplanktonsporesand crustaceaneggs cansurvivedigestion (seeTable1).Formostofthe species tested, some propagulessurviveddigestion,

exceptforphytoplanktonwhere, amongstsevenalgae species tested, only the sporesofone diatomgermi- natedafter digestion(Atkinson,1970,1971).

Someofthese studiescompared thesurvivalof differentspecies after digestiontoinvestigatewhich propagulecharacteristicsfavourinternaltransportby birds. Thesestudies have producedcontradictory results.DeVlaming Proctor (1968) tested the resistanceofseeds of23wetland plant species to digestionbymallardAnasplatyrhynchosand killdeer Charadriusvociferus.Theyconcludedthat resistanceto digestionishigherforsmall seeds with hardcovers and is also affected by the characteristics of the disperser species (gizzard structure and retention time ofseeds inthe gut were proposedasthe most importantcharacteristics). Inasimilar studywith mallard,Holt (1999)concludedthat resistance to digestionwas unrelatedtosizeand lignin content in seeds ofeightwetland plant species. However,she reportedthat the proportionofnon-destroyedseeds germinating afteringestionbybirds waspositively associated with the size and lignin content of the seeds.

The study of the propagule characteristicsthat facilitate survival afterdigestionisimportantinorder tounderstandthe factors that affect the colonisation and dispersal capacitiesof differentaquaticorgan- isms.However,variousproblemslimittheusefulness oftheexperimental set-upsused sofartoaddressthis topic.First,thenumbersofbirds used inthesestudies were very low(Table1)and sometimesnotstatedby theauthors,makingithard tointerprettheresults obtainedgiven thehigh intraspecificvariabilityinthe structureand size ofthe digestiveorgansofwater- birds (Kehoe, AnkneyAlisauskas,1988).Secondly, variousfactors canaffectthestructure ofdigestive organsofcaptive waterbirds,includingtime incap- tivity(ClenchMathias, 1995) and diet(Al-Dabbagh, JiadWaheed,1987).Thismakes resultsdifficult to comparebetween studies.Althoughwecanbecon- fident of resultsobtainedby feeding differentseed types simultaneouslytothesame individualbird, the percentageofseeds survivingdigestion reportedin different studiesarenotdirectlycomparable,as differencesin seed viability betweenstudies could bebecause ofeither thecharacteristicsoftheseeds or thecharacteristicsofthebirds used intheexperiment (even when either dispersedor disperserorganisms belong tothesame species).

Table1Summaryofexperimentscarriedout totesttheviability ofpropagulesafter digestionbywaterbirds.Thenumberofindividualbirds used intheexperiments isreportedinbrackets

PropagulespeciesDisperserspeciesSummaryofmain resultsReference

Eggsoftwo gastropodsAnasplatyrhynchos(5)Asmall proportionoftheeggsoftwo species ofMalone (1965b)

Charadriusvociferus(10)gastropodaremainedviable after digestion

Artemia(Crustacea)Anasplatyrhynchos(5–20)Somesurvivedpassage,eggsenteringinthececaareMalone (1965a)

retainedformuch longer than expectedfrom intestinelength

ArtemiaTadornatadorna(2)Hatchingefficiency ofcystswas unchangedbypassageMacDonald(1980)

Phoenicopterusruber(2)throughPh.Ruber,and reducedbyT.tadorna

Twelve genera ofcrustaceansAnasplatyrhynchos(34)Somespecies hatchedafter digestion,others did notProctor (1964)

Artemia,other invertebrates,algaeCharadriusvociferusSurviveddigestionbyCh.vociferusbetter than byProctor etal.,(1967)

and Chara(Characeae)AnasplatyrhynchosA.platyrhynchos

Artemiaand other invertebrates,algaeAnasplatyrhynchosAllsurviveddigestionProctor Malone (1965)

and Chara

Nitella(Characeae)Duck (1)Somegerminatedafter digestionImahori(1954) Riellaamericana(Riellaceae) Anasplatyrhynchos(3) Germinationand development‘similar’ toundigestedspores Proctor (1961) Sixspecies ofChara Anasplatyrhynchos(3) Somegerminatedafter digestion Proctor (1962) Asterionella(Diatomea) Anasplatyrhynchos(4) None germinatedafter digestion Atkinson(1970) Fivediatomea,oneCyanophyta Anasplatyrhynchos(4) Only onediatomeaspecies germinatedafter digestion Atkinson(1971) Potamogetonnatans(Potamogetonaceae) Anasplatyrhynchos 60%germinatedafter digestion Guppy(1906) Scirpuspaludosus(Cyperaceae) Anasplatyrhynchos(1) Somegerminatedafter digestion Low(1937)

Seedsof23wetlandplant speciesCharadriusvociferus(‡5)*Smallseeds and those with ahard cover have higherDeVlaming Proctor (1968)

Anasplatyrhynchos(‡4)*resistancetodigestion.Retentiontime longer in

Ch.vociferusthan inA.platyrhynchos

Seedsofeight wetlandplantsspeciesAnasplatyrhynchos(4–16)Large seeds and those with ahardcover germinatebetter afterHolt (1999)

digestion,noeffectsofthese variablesonresistancetodigestion

Najasmarina(Najadaceae)Anasplatyrhynchos(4)Thirty per centsurviveddigestion,more than 30%germinatedAgami Waisel (1986) Seedsof13plant species Twelve species ofbirds (2–5) Retentiontimes longer inwadersthat inducks (up to10days) Proctor (1968)

Seedoffiveplant speciesThree species ofducks (5)Viableseeds recoveredforfour oftheseed speciesPowersetal.(1978)

*Indicates that differenttypes ofpropaguleswere tested indifferentindividualbirds.

There arealsostatisticalproblemswith those studies that take each seed asthe unit for analyseswithout accountingfortheeffectsofindividualducks (thecase forallstudies citedinTable1).Thisleads topseu- doreplication,asthe resistanceand viability ofseeds digested bythesame duck cannot beconsidered statistically independent observations.Thiscanbe solved byincludingarandom factorcontrollingfor individual-relatedeffectsonseed digestionand viab- ilityintheanalyses(seeBenningtonThayne, 1994, formore informationonrandomfactors).

It is also difficult to compare differentstudies because themethods used tofeedpropagulesto experimentalbirds differ widelybetweenand within studies (force-feeding,mixed with leaves, grain or inside gelatinouspills, see, e.g. Proctor Malone,

1965).Thetype offood ingestedwith thepropagules canstronglyinfluence therateofpassagethroughthe digestivetract (Malone, 1965a).Furthermore, the conditionsoptimalforgermination differbetween propagule species. AsProctor (1964)recognisedfor crustaceaneggs, the apparentdifferencesinviability ofdifferentpropagulespecies afterdigestion may merely reflectinterspecificdifferencesinrequirements for germination or hatching. This problem can be solved bytheuse ofcontrol (undigested)propagules todeterminethesuitabilityofthegerminationproce- duresused (see,e.g.Agami Waisel, 1986),and by makingstatisticalcomparisonsofgerminationratesof digestedand undigestedpropagules. Todate, few studieshave includedsuch analyses,and none have separated theeffectsofdigestiononthespeed of propagulegerminationfrom its effects on long-term viability (see Traveset, 1998, for a review of this probleminstudies ofterrestrialorganisms).For example,Agami Waisel(1986)foundthat digestion by mallardsacceleratedgerminationofholly leaved naiadNajasmarinaseeds, butdid notassess itseffects onlong-termviability.

There isevidencethat waterbirdspecies candiffer widelyintheir potentialforinternaldispersal(seeDe Vlaming & Proctor, 1968),butthevariablesexplaining such differences remainuninvestigated.Gizzard structure (assuggested byProctor, Malone De Vlaming, 1967), the nature of grit in the gizzard (Mateo, Guitart Green, 2000) and intestineand caeca length (see Malone, 1965a) are variablesthat seem likely to influence survivalofpropagulesand their retentiontime (see below) and that shouldbe

consideredinfuture studies.Fibrous dietsinmallards lead to heaviergizzards(Miller, 1975;Kehoe etal.,

1988)more likely tocrush propagules,and gizzards become heavierinperiodsofthe annualcyclewhen waterbirddiets are more plant-based(see review in Piersma, Koolhaas Dekinga, 1993). Herbivorous and omnivorouswaterbirds have heaviergizzards than carnivorous species (Kehoe & Ankney, 1985; Barnes Thomas, 1987), and such differencesare likely to influence the proportion of propagules passingundigestedthroughthe birds. The ingestion ofArtemiaeggswithgrit(sand) reducedtheirsurvival ofdigestion(MacDonald,1980).

Internaltransportstep 3:dopropagulesstay long enoughin awaterbirdtobedispersed over along distance?

Themaximumdispersaldistance ofapropaguleby internaltransportislimitedbyhow long the propa- guleremainsinthegutofawaterbird(retentiontime), and how farthe bird iscapableofflying duringthat period. There is little information about retention timesofpropagulesinwaterbirds,althoughstudiesof terrestrialbirds suggestthat retentiontime shouldbe longer forsmaller propagules(Levey Grajal, 1991; Izhaki, Korine Arad, 1995).Agami Waisel (1986) reported retention times of 10–12h for N. marina seeds inmallards,whilst SwansonBartonek (1970) foundthat some Scirpusseeds canberetainedinthe guts of blue-wingedteal Anasdiscorsfor over 72h, althoughtheirviability wasnottested. Malone(1965a) investigatedhow foods consumedwith the propa- gules influence retentiontime inMallards.Retention time changed significantlywith the food provided with eggsofthecrustacean Artemiasalina,with minimum and maximum retention time of viable eggsbeingtwiceas longwhenfedwithcorn(mean maximumretentionofviable propagulesofover 5h) thanwhen fedwithanon-fibrousaquaticplant (Elodea canadensis).

Ducks flyatspeedsof60–78kmh–1,and wadersat

speeds of 48–60kmh–1 (Welham, 1994). Green- wingedtealAnascreccahave been knowntomigrate over 1200km in less than 24h (P. Clausen,pers. comm.).Duringmigration,Pintail A. acutaoftenmove about 1000km between satellite locations at 72h intervals (Miller etal., 2000). Thus, it is clear that propagules can be dispersed long distances by

internal transport, although the frequency of such long-distancedispersalevents remainsunknown.

Externaltransport:can propagulesadhere towaterbirds?

Experimentsonthepotentialforexternaltransportof propagulesbyadhesiontothefeathers,billsorlegsof waterbirdsare even scarcer than for internaltrans- port. Anumberofanecdotalobservationsoforgan- isms adheringto plumagehave been reported (Maguire,1959; Maguire,1963; Swanson,1984). Althoughanumberofexperimentstodeterminethe characteristics ofseeds thatfacilitate transportby adhesion havebeenperformed forterrestrialorgan- isms(Sorensen, 1986; KiviniemiTelenius, 1998), experimentalevidenceforaquaticorganismsisscarce. The earliest studywas performedbyDarwin(1859), who removedthelegofadeadwaterbirdand placed itinatank with pondsnails. Thesnails crawledonto thebirdfootandmany stayedtherewhen heremoved the foot from the water and waved it around to simulateflight. Many years later, Segerstra˚le (1954) establishedthat theamphipodGammaruslacustriscan adhereto the plumageofducks for up to 2h, and Cercopagis(Cladocera)alsofouled totheplumageofa deadtrial duck (Makarewiczetal.,2001).Infreshwa- ter environments, exposureto desiccation during transport is likely to limit the potential of many aquaticorganismstodispersebyadhesion.Asyetno study hasaddressed how desiccationaffectsthe viability ofdifferentkinds ofpropagules. Many seeds andresting eggsarelikelytobehighly resistant to desiccation(e.g. Bilton etal.,2001).However,Frisch (2001)showed thatthelarval resting stages oftwo cyclopoidcopepodspecies die within24h, even at

100%relative humidity,suggesting that desiccation limits thedistanceofdispersalofthelarvae by waterbirds.There isanecdotalevidencethat smaller propagulesand those with hook-likestructuresare more likelytobeattached towaterbirdplumage. Statoblastsofbryozoansthat have hooks have often beenobserved onmoultedduck feathers(OkamuraHatton-Ellis,1995), and Vivian-SmithStiles(1994) described the characteristics of the seeds found adheredto waterfowl.Seeds and resting eggs that floatareobviouslymore likelytobecome attachedto plumage.However,we are not aware ofany experi- ments testing how differentpropagulemorphologies

influence their potentialtoremainattachedtowater- birds.

With whatfrequencyarepropagulesdispersed bywaterbirdsinthe field?

The studies reviewed above show that waterbirds have thepotentialtotransportpropagules,butdonot provideus with informationabout the frequency of waterbird-mediateddispersalinthefield.Waterbirds are not only potentiallyimportantasagents oflong- distancedispersaland ofcolonisationofnewhabitats, but alsoasagents ofdispersalatalocalscale. There areveryfewfieldstudiesofthefrequencyoftransport (summarisedinTable2), buttheyareenoughtoshow thatwaterbird-mediated transportoccurs atahigh frequencyinthefieldbothbyinternal and external transport, specially forplant seeds. Forexample, Powers,Noble &Chabreck(1978) reported 17species ofnon-digestedangiospermseeds intheintestinesof

51hunted ducks, and Vivian-SmithStiles(1994) reportedangiospermseeds attachedto the feathers and feetof28outof36brant geeseBrantaberniclaand ducks.

There arenostudiestodate that combineinforma- tion on the quantities of propagules in faeces or attachedexternallytowaterbirdswith precise infor- mationonthe movementsofthe birds; the distances overwhich thepropagules were dispersed inthese studiesarethus open toconjecture.

Prioritiesforfutureresearch

Our review of publishedstudiesdemonstratesthat many organisms canpotentiallybedispersedby waterbirds,and that these processesare likely tobe frequentenoughtohave amajorimpact onmetapop- ulationdynamics and geneflowinmany aquatic organisms, at least at a local scale. The relative importance ofdifferentvectors (waterbirds and other animals;thisstudy,Lowcock Murphy,1990;wind, BrendonckRiddoch,1999;rain, Jenkins Under- wood, 1998;and man, Reise,Gollasch Wolff,1999) fordispersalofdifferentpropaguletypes atdifferent scales needs tobeaddressed.

Furtherresearch isrequiredfrom arange of disciplines to advance our understanding of how bird-mediateddispersal canshape thestructureof aquatic communities. There are various questions

relatingboth tothe characteristicsofthe propagules thataredispersed(size,shape,hardness,resistanceto desiccation)and tothecharacteristicsofthedisperser and their effectsonthepotentialfordispersal.

Arange offocussed experimentsareneededto establishhow the variouscharacteristicsofdifferent propagulesinfluence their potentialtobetransported internally(quantifyingdigestability, retentiontime, etc.)orexternally(quantifyingtime spent attachedto plumage,resistance todesiccation,etc.).Anumberof differentgenetic methodstoestimatethe tails ofthe propagules’dispersal distancecurves areunderactive development(Cain,Milligan Strand, 2000),opening promisingopportunitiestotest the relationbetween propagulecharacteristicsand dispersalcapacity, and toinvestigatethefrequencyoflong-distancedispersal phenomenainthefield.

Research isrequiredonhow thediet and structure ofdigestiveorgansofwaterbirdsaffects the propor- tionofpropagulessurvivingdigestion, and their retentiontime. The retentiontime of propagulesin thedigestivesystem determinestherange ofdistances potentiallytravelledbythepropagules viainternal transport (Proctor, 1968). However, retention time may also have profoundeffects on the viability of digested propagules; forexample, thepropagules retainedlongest may have lower survival.Although some authorshave suggestedthat longer retention timesarebecause ofvariabilityinthepassagethrough theintestinesand that these may have littleeffecton theviability ofpropagules(Clench Mathias,1992), others have statedthat variationin retentiontimes was because of variationin the time spent in the gizzard, which islikelytoleadtoastrong effectof retention time upon viability (Proctor etal., 1967). Thus there isaneed fordetailedexperimentsonthe effectsofretentiontimeuponfuture viability of propagules.

Some species ofpasserinebirds seem toinitiate a periodoffasting prior tothestartofamigratoryflight (Fransson,1998).PiersmaGill(1998)alsoreporteda reductioninthesizeofthe digestiveorgansofknots Calidriscanutusduringmigration(seealsoPiersmaLindstro¨m,1997).Ifbirds emptytheir digestivetracts prior tomigration,thiscould reducethepotentialfor long-distancedispersalduringmigration.However, it seems unlikelythat birds cancompletelyemptytheir digestiveorgans, includingthe long caeca character- istic of waterfowl (Clench Mathias, 1992, 1995).

Furthermore,evidencefromterrestrialplantssuggests that even low-frequency, long-distancedispersal events canhave profound effectsonpatternsof colonisationand distribution(Cain,DammanMuir,

1998), especiallyontheexpected speed ofrange expansion(HigginsRichardson,1999).

Noinformationissofaravailableonthecharacter- istics of propagule shadows(microscalepatterns in the dispersalmovementsofpropagulesvia birds) in aquaticenvironments,oronthecharacteristicsofthe locations where propagules areredistributedby waterbirdsinrelationtothegerminationand growth requirementsoftheorganisms.Sofaritisunknownif dispersalbywaterbirdsisdirectional,towardsgood (or bad) places for the growthofthe organisms,or random(seeWenny Levey,1998,foranexampleof directionaldispersalbybellbirds).

Satellite trackingis providingmuch new data on the long- and short-distance movements of water- birds, althoughresearchtodatehassofarfocussedon largeswan and goosespecies that arelikelytobeless important for dispersal of aquaticorganismsthan smaller and more abundantducks and waders(but see Miller etal., 2000). The recoveries of ducks a couple ofdays after markingcan also provideinter- esting informationonthepotentialrange ofdispersal providedbywaterfowl.Existing data aresufficient to show thatlong-distance movementsbymigratory waterbirdsare not entirelyconfinedto autumnand springmigration. Many waterfowlspecies undergo long-distancemovementstomoulting sitesafter breedinghas been completed(Alerstam,1990).After arrivingatwintering sites,waterfowl and other waterbirds oftenundergofurthermid-winter move- ments, e.g.relatedtospells ofharshweather(Ridgill

Fox,1990)ortochangesinwater levelinresponse to heavy rainfall (Simmons, Barnard Jamieson,

1998). There generally appears to be a constant

‘turnover’ of individual waterbirds on wintering areas, and the only detailedstudywe are aware of foundthatthemajorityofgreen-wingedtealsA.crecca winteringinthe Camarguestayedinthe studyarea forlessthan 10days (Pradel etal.,1997).

Many waterbird species breeding atnon-extreme latitudesarenotstrictlymigratorybutrathernomadic, makinglong-distancemovementsatany time ofthe yearinresponse tospatial fluctuations inwetland availability(delHoyo etal.,1992;KingsfordPorter,

1993).Thefrequentdailymovementsofallwaterbirds

from foragingtoresting areas atany time oftheyear and theexploratory movements ofjuveniles can potentiallyresult inthedispersalofplant orinverteb- rate species or genotypes over larger distances, through successive short-distance dispersal steps. Suchbirdmovements atalocalscaletend tooccur withoutanyparticulardirection (asopposedtothe directionalityofmigration),and canbevery frequent duringthe periodsofmaximumpropaguleproduc- tivity (Wilkinson,1997).