Kays,Michael Ronsenzweig,Phil Grime,and Ken Thompsonfor providing biod - PDF document

Kays,Michael Ronsenzweig,Phil Grime,and Ken Thompsonfor providing biodiversity data and valuable insights on thistopic.This article also benefited greatly from the thoughtfulcomments and suggestions ofthree anonymous reviewers. References cited Bellwood DR,Hughes TP.2001.Regional scale assembly rules and biodiversityofcoral reefs.Science 292:1532[BONAP] Biodiversity ofNorth America Program.1999.Synthesis oftheNorth American Flora,Version 1.0 [CD-ROM].Chapel Hill:BONAP,North Carolina Botanical Garden,University ofNorth Carolina.Bryant SJ,Jackson J.1999.Tasmanias Threatened Fauna Handbook:What,Where and How to Protect Tasmanias Threatened Animals.Hobart(Tasmania):Threatened Species Unit,Parks and Wildlife Service.Carlton JT,Geller JB,Reaka-Kudla ML,Norse EA.1999.Historical extinc-tions in the sea.Annual Review ofEcology and Systematics 30:515Chave J,Muller-Landau HC,Levin SA.2002.Comparing classical commu-nity models:Theoretical consequences for patterns ofdiversity.Amer-ican Naturalist 159:1Chilton G.1997.Labrador Duck, Camptorhynchus labradorius. Washington(DC):American OrnithologistsUnion.The Birds ofNorth America,no.Cox CB,Moore PD.2000.Biogeography:An Ecological and Evolutionary Ap-proach.Oxford (United Kingdom):Blackwell Science.Antonio CM,Vitousek PM.1992.Biological invasions by exotic grasses,the grass-fire cycle,and global change.Annual Review ofEcology andSystematics 23:63Davis MA,Pelsor M.2001.Experimental support for a resource-basedmechanistic model ofinvasiblity.Ecology Letters 4:421Davis MA,Grime JP,Thompson K.2000.Fluctuating resources in plant com-munities:A general theory ofinvasibility.Journal ofEcology 88:528Elton CS.1958.The Ecology ofInvasions by Animals and Plants.London:Methuen.Fritts TH,Rodda GH.1998.The role ofintroduced species in the degrada-tion ofisland ecosystems:A case history ofGuam.Annual Review ofEcol-ogy and Systematics 29:113Gill B,Martinson P.1991.New Zealands Extinct Birds.Aukland (NewZealand):Random Century.Golani D.1993.Trophic adaption ofRed Sea fishes to the eastern Mediter-ranean environmentreview and new data.Israel Journal ofZoology39:391Gould SJ.1998.An evolutionary perspective on strengths,fallacies,andconfusions in the concept ofnative plants.Arnoldia 58:11Guiler E.1985.Thylacine:The Tragedy ofthe Tasmanian Tiger.Oxford(United Kingdom):Oxford University Press.Herbold B,Moyle PB.1986.Introduced species and vacant niches.Ameri-can Naturalist 128:751Hubbell SP.2001.The Unified Neutral Theory ofBiodiversity and Biogeography.Princeton (NJ):Princeton University Press.Hughes TP,Bellwood DR,Connolly S.2002.Biodiversity hotspots,centresofendemicity,and the conservation ofcoral reefs.Ecology Letters 5:Huston MA.1994.Biological Diversity:The Coexistence ofSpecies onChanging Landscapes.Cambridge (United Kingdom):Cambridge University Press.Kaufman L.1992.Catastrophic change in species-rich freshwater ecosystems:The lessons ofLake Victoria.BioScience 42:846King C.1984.Immigrant Killers.Oxford:Oxford University Press.King WB.1980.Ecological basis ofextinction in birds.Acta Congressus Internationalis Ornithologici 2:905Loope LL.1999.Hawaii and the Pacific Islands.In US Geological Survey.Status and Trends ofthe Nations Biological Resources.Washington(DC):US Government Printing Office.(6 April 2003; http://biology.usgs.gov/s+t/SNT/ noframe/pi179.htm) Marchetti MP.1999.An experimental study ofcompetition between the na-tive Sacramento perch (Archoplites interruptus) and introduced bluegill (Lepomis macrochirus). Biological Invasions 1:55Miller SE,Eldredge LG.1996.Numbers ofHawaiian species,Suppl.1.BishopMuseum Occasional Papers 45:8Montevecchia WA,Kirk DA.1996.Great Auk, Pinguinus impennis. Washington(DC):American Ornithological Union.Birds ofNorth America,no.Mooney HA,Cleland EE.2001.The evolutionary impact ofinvasive species.Proceedings ofthe National Academy ofSciences 98:5446Moyle PB,Light T.1996.Biological invasions offresh water:Empirical rulesand assembly theory.Biological Conservation 78:149Owenby GB,Morley T.1991.Vascular Plants ofMinnesota:A Checklistand Atlas.Minneapolis (MN):University ofMinnesota Press.Peterson RO.1995.Wolves as interspecific competitors.Pages 315byn LN,Fritts SH,Seip DR,eds.Ecology and Conservation ofWolves ina Changing World.Alberta (Canada):Canadian Circumpolar Institute.Petren K,Case TJ.1996.An experimental demonstration ofexploitation com-petition in an ongoing invasion.Ecology 77:118[PIER] Pacific Island Ecosystems at Risk Project.1999.Plant threats to Pacific ecosystems.(6 April 2003; www.hear.org/pier/threats.htm) Pimm SL,Askins RA.1995.Forest losses predict bird extinctions in easternNorth America.Proceedings ofthe National Academy ofSciences 92:Pimm SL,Moulton MP,Justice J.1994.Bird extinctions in the central Pacific.Philosophical Transactions ofthe Royal Society ofLondon,B 344:Polis GA,Anderson WB,Holt RD.1997.Toward an integration oflandscapeecology and food web ecology:The dynamics ofspatially subsidized foodwebs.Annual Review ofEcology and Systematics 28:289Por FD.1978.Lessepsian Migration.Berlin:Springer-Verlag.Porter SD,Savignano DA.1990.Invasion ofpolygyne fire ants decimates na-tive ants and disrupts arthropod community.Ecology 71:2095Randall JM,Rejmanek M,Hunter JC.1998.Characteristics ofthe exotic floraofCalifornia.Fremontia 26:3Richardson DM,Pysek P,Rejmanek M,Barbour MG,Panetta FD,West CJ.2000.Naturalization and invasion ofalien plants:Concepts and defin-itions.Biodiversity and Distributions 6:93Rosenzweig ML.2001.The four questions:What does the introduction ofexotic species do to diversity? Evolutionary Ecology Research 3:361Ross ST.1991.Mechanisms structuring stream fish assemblies:Are therelessons from introduced species? Environmental Biology ofFishes 30:Sale PF.1977.Maintenance ofhigh diversity in coral reeffish communities.American Naturalist 111:337Sax DF,Gaines SD,Brown JH.2002.Species invasions exceed extinctions onislands world-wide:A comparative study ofplants and birds.AmericanNaturalist 160:766Spanier E,Galil BS.1991.Lessepsian migration:A continuous biogeo-graphical process.Endeavour 15:102Steadman DW.1995.Prehistoric extinctions ofPacific island birds:Biodi-versity meets zooarchaeology.Science 267:1123Tompkins DM,White A,Boots M.2003.Ecological replacement ofnativered squirrels by invasive greys driven by disease.Ecology Letters 6:189-Thompson JD.1991.The biology ofan invasive plant.BioScience 41:Wilcove DS,Rothstein D,Dubow J,Phillips A,Losos E.1998.Quantifyingthreats to imperiled species in the United States.BioScience 48:607Wilson EO.1992.The Diversity ofLife.Cambridge (MA):Belknap Press.Winemiller KO,Polis GA.1996.Food webs:What do they tell us about theworld? Pages 122 in Polis GA,Winemiller KO,eds.Food Webs:Inte-gration ofPatterns and Dynamics.New York:Chapman and Hall. May 2003 / Vol.53 No.5 BioScience489 Articles 03 May Article Davis 4/23/03 3:22 PM Page 489 spread ofthese species and to eradicate them in regionswhere they are already a threat.Some new species,such as introduced predators on islands,are clearly capable ofexter-minating resident species,and most ecologists would prob-ably regard such extinctions as undesirable and hence supportthe use ofsocietys resources to protect these species.Globalization ofEarths biota will not lead to a world com-posed ofzebra mussels,kudzu,and starlings.One need onlylook to one ofthe most diverse communities in the world,coral reefs,to see the consequence ofregular long-distance dis-persal (figure 4).As a result ofocean currents that regularlytransport immature corals and fish hundreds or even thou-sands ofmiles,reefs exhibit comparatively low rates ofendemism and instead are typically composed offish and coralspecies with large geographic ranges (Hughes et al.2002).Aspredicted by the extended unified neutral theory,reefsthroughout the Indian and Pacific Oceans share a commonfauna (Bellwood and Hughes 2001),a biogeographic pat-tern that contrasts sharply with that exhibited by historicterrestrial communities in the same archipelago.Comprehensive analyses ofcoral andreeffish communities suggest that com-munity composition ofreefs throughoutthe Indo-Pacific region is primarily dis-persal rather than niche driven (cf.Bell-wood and Hughes 2001),consistent withHubbells original theory.Although theextensive dispersal patterns homogenizereeffaunas to some extent,the diversecoral and fish communities are funda-mentally the product ofthese high dis-persal (invasion) rates,and these samehigh dispersal rates currently help to main-tain,not threaten,the high species richnessamong corals and reeffish communities(Hughes et al.2002).Thus,although it istrue that the breakdown ofthe worlddispersal barriers will result in a homog-enization ofEarths biota,homogenizationis not synonymous with low diversity.Inthe future,different regions ofthe worldwill be more similar than they are now.They will also be more diverse.Because conservation priorities andmanagement plans involving introducedspecies are largely based on ecologistssessments ofbiological invasions and theireffects,it is vital that these judgments bescientifically sound,that is,well groundedin theory and robustly supported by fielddata.In addition,given the limited hu-man and financial resources available todeal with the biodiversity impacts ofthelarge number ofintroduced speciesthroughout the world,it is important thatconservationists focus their efforts wisely.The biogeographic analysis and data presented here and else-where (Mooney and Cleland 2001,Rosenzweig 2001,Chaveet al.2002,Sax et al.2002) suggest that ecologists and habi-tat managers striving to protect and maximize biodiversity willbe able to maximize their effectiveness by focusing principallyon intertrophic interactions between introduced species andlong-term residents.Taken together,theory and data sug-gest that,compared to the effects ofintertrophic interac-tions and habitat loss,competition from introduced speciesis not likely to be a common cause ofextinctions oflong-termresident species at global,metacommunity,and even mostcommunity levels. Acknowledgments I am grateful to Jim Space (project director,Pacific IslandEcosystems at Risk),David Bellwood (deputy director,Centre for Coral ReefBiodiversity),John Kartesz (director,Biota ofNorth America Project),Gerald McCormack (di-rector,Cook Islands Natural Heritage Project),Ann Dennis(executive director,CalFlora),Dov Sax,Don Wilson,Roland 488BioScience May 2003 / Vol.53 No.5 Articles Figure 4.A diverse Fiji reefcommunity.Shown are both hard and soft corals (the prominent red soft coral is Dendronephthxa sp.),some crinoids,sponges,and several species offishfairy basslets (Pseudanthias spp.) and damselfish (Chromis sp.) . Because ofthe long-distance dispersal capabilities ofmany reeforganisms,endemism is not common in most reefs.Historically,Pacific island reefcommunities were more homogenous in their species composition than the respective terrestrial island environments.However,although homogeneous,thereefcommunities were very diverse,whereas the terrestrial island communities,which were largely isolated from one another and thereby composed ofmany endemic species,were comparatively species poor.The diverse homogeneity illustrated by coral reefs may aptly characterize the worlds biota in the future.Increasingly connected as a result ofbiological invasions,different regions ofthe world will become more similar in their floras and faunas.At the same time,they will become more diverse,in some cases much more diverse.Photograph: Copyright © Brandon Cole. other compelling reasons to prefer the long-term residents overthe new species,including aesthetic,economic,cultural,andhistorical preferences (Gould 1998).While the arrival ofnewspecies often disrupts ecological and evolutionary relation-ships among the long-term residents,it also creates oppor-tunities for the development ofnew ecological andevolutionary relationships (Mooney and Cleland 2001).It is important to reemphasize that the extended unifiedneutral model proposed here addresses species within the sametrophic level.Without question,introduced species can driveresident species in other trophic levels to extinction,even quitequickly,as illustrated by the introductions ofpredators andpathogens to islands (figure 3).The expected impacts onbiodiversity ofsuch intertrophic interactions might be ef-fectively addressed using approaches based on food web the-ory and trophic cascades (Winemiller and Polis 1996,Poliset al.1997).In addition,it must be stressed that both the fluctuating re-source availability theory and the unified neutral theory weredeveloped primarily with the ecology ofplant communitiesin mind.Although the predictions presented here seem to holdfor many animal groups as well,it is possible that competi-tion-driven extinctions might be more likely higher on thefood web.For example,some predators (e.g.,wolves) areknown to drive other predators (e.g.,coyotes) out ofan areawith their aggressive behavior (Peterson 1995).Thus,the in-troduction ofa dominant predator into a region,particularlya geographically isolated region such as an island or lake,po-tentially could cause the extinction ofone or more residentpredators through competition.Although the reasons be-hind the extinction two to three thousand years ago ofthe thy- (Thylacinus cynocephalus) in Australia are not fullyknown,it is suspected that competition from the dingo wasa major factor (King C 1984,Guiler 1985).Overhunting,habitat loss,and disease are believed to have driven the lastremaining thylacine populations to extinction in Tasmania inthe early 1900s (Bryant and Jackson 1999).One problem that arises from extending the spatial scaleofHubbells unified neutral theory is that the species neutralityassumed by Hubbell at the local and regional scale is less likelyto hold at the global scale.Hubbell recognized this,pointingout that while species evolving in different regions would haveevolved neutrality (equivalence ofper capita fitness) with re-spect to the other controphic species in their local commu-nities,there would be no reason to expect neutrality amongcontrophics from different regions.Thus,some introducedspecies may enjoy a pronounced competitive advantage intheir new environment.Hubbell (2001) refers to such speciesrule breakers.These are species capable ofsubstantiallyreducing the abundance ofmany long-term resident species.However,the paucity ofdocumented cases ofcompetition-driven extinctions indicates that,even in instances whereconsiderable interspecific variation in competitive ability ex-ists between new species and long-term residents,extinc-tion is seldom the outcome.Furthermore,the introductionofthese rule breakerswould alter the selection regimes inthe host environment,which would be expected to lead,overtime,either to an increase in the relative competitive abilityofthe long-term residents (Hubbell 2001) or,as explainedabove,to a niche shift that would reduce the intensity ofcompetition (Ross 1991,Spanier and Galil 1991,GolaniAlthough introduced species seldom drive resident con-trophics to extinction through competitive interactions,theycan bring with them new pathogens that may cause extinc-tions among the resident species (Loope 1999).Also,intro-duced species may significantly alter disturbance regimes,asin the case offlammable grasses introduced to Hawaii Antonio and Vitousek 1992);such alterations could leadto substantial enough changes in habitat to result in the ex-tinction oflong-term resident and controphic species.Richardson and colleagues (2000) refer to species that canchange the character or condition ofan ecosystem over a sub-stantial area as transformers. Conclusions and recommendations Few new species are returning to the metacommunities fromwhich they came,and thus it is inevitable that communitiesofthe future will consist ofa mixture ofnew and long-termresidents.However,that we cannot turn back the ecologicalclock (except possibly in small restoration efforts requiringperpetual vigilance and intensive management) does notmean that we should cease all efforts to protect long-term res-idents or that we should welcome all introduced species withopen arms.The serious health and economic problems causedby a small percentage ofintroduced species certainly de-mand aggressive action,including efforts to prevent the May 2003 / Vol.53 No.5 BioScience487 Articles Figure 3.The brown tree snake, Boiga irregularis, intro-duced into Guam around 1950,is believed to have causedthe extinction ofmore than 10 bird species,some endemicto Guam,in only a few decades.Biological invasions aremore likely to threaten long-term resident species withextinction through intertrophic interactions,such as pre-dation and pathogenic infection,than through competi-tion.Photograph: Gordon Rodda,US Geological Survey. a contributing factor in some cases (King WB 1980).For example,ifhabitat fragmentation on continents continues,some long-term resident species,particularly those with lim-ited ranges to begin with,may end up persisting in just oneor a few fragments ofhabitats.With populations reduced insize and confined to geographically restricted habitats,thesemuch like species inhabiting oceanic islandswillbe more vulnerable to extinction.That competition-driven extinctions have been uncommon even on oceanic islands sug-gests that factors other than competition are likely to causeextinctions in these fragmented continental habitats as well.Given that small populations are vulnerable to extinctionresulting from a wide variety ofcauses,the fact that a long-term resident species goes extinct in the presence ofan introduced controphic will not be enough alone to justify aclaim ofcompetition as the cause.In most instances where an introduced species has exter-minated a resident,the species interactions have been betweentrophic levels,e.g.,predatorprey,pathogenhost (King C1984,Gill and Martinson 1991,Kaufman 1992,Fritts andRodda 1998,Loope 1999).The fact that competition-drivenextinctions have been so uncommon during the past fewcenturies suggests that,compared to habitat loss and intertrophic interactions,competition is a relatively weakextinction threat.For example,although introduced freshwaterfish have been shown to reduce the population sizes oflong-term resident species through competition,these interactionsseldom cause extinctions (Herbold and Moyle 1986,Ross1991).Ifcompetition is a relatively weak extinction threat,extinctions caused by competition should take longer thanthose caused by predation and habitat loss.Ifthis is the case,then the reason so few competition-driven extinctions havebeen documented may be that not enough time has passedfor competitive exclusion to occur.This would suggest thatwe might expect more competition-driven extinctions in thefuture.However,the increased time frame needed for theseextinctions also provides more time for other factors to disrupt the competitive asymmetry between the new andlong-term resident species,thereby reducing the likelihood thatsuch extinctions would ever occur.These factors would includeevents and processes that would reduce the abundance ofthenew species,such as disturbances,disease,environmentalfluctuations,or even a new introduced species.To illustrate,Marchetti (1999) concluded that although the Sacramentoperch (Archoplites interruptus) is threatened by the aggressivedominance ofan introduced bluegill (Lepomis macrochirus), competitive exclusion ofthe perch may never occur becauseoffluctuating environmental conditions.A longer time frame also means that the resident speciesmay have time to adapt to the new competition pressure intheir environment and thereby reduce the intensity ofcom-petition to a level that permits coexistence.For example,asdescribed by Mooney and Cleland (2001),the introductionofmore than 250 new fish species into the Mediterranean Seafollowing the completion ofthe Suez Canal has resulted in onlya single extinction (Por 1978).This has been attributed to theability ofthe long-term residents to respond to the compet-itive interactions with the Red Sea species by adjusting theirforaging depths.This niche adjustment enabled the long-term residents,which prefer to feed in the warmer surface waters ofthe Mediterranean (Spanier and Galil 1991,Golani1993),to accommodate the introductions.Similar niche ad-justments in response to introduced species have been doc-umented for freshwater fish (Ross 1991).For these reasons,it seems unlikely that competition willemerge as a primary,or even frequent,cause offuture ex-tinctions.It is significant that interspecific variation in com-petitive abilities is not incorporated into the extended unifiedneutral model presented here.The fact that biodiversity datasets from around the world are consistent with the predictionsofsuch a theory,and at multiple spatial scales,is an additionalargument against the notion that competition from intro-duced species constitutes a major extinction threat for long-term residents.Thus,the observed increases in local andregional biodiversity resulting from species introductionsare not likely to be a temporary phenomenon,but representreal increases that,in the absence offuture habitat loss,willlikely persist indefinitely into the future (Rosenzweig 2001). Biodiversity at the meta-metacommunity (global) level Biodiversity is declining at the global level (Wilson 1992).Thisis indisputable.The extended unified neutral model pre-sented here predicts that increased dispersal among meta-communities will be responsible for some ofthe speciesextinctions at the global level.A theorem ofHubbellwhich is based on a biotically saturated landscape,is thatany increase in the population size ofone species must be ac-companied by a corresponding decrease in the collectivenumber ofindividuals ofall other species in the community.Thus,any new species that become dominant and wide-spread in communities would be expected to depress thepopulations ofresident species,making already rare and lo-cal species particularly vulnerable to extinction from sto-chastic processes (i.e.,ecological drift).The result oftheseextinctions would be that total meta-metacommunity bio-diversity (i.e.,global biodiversity) would decline,even thoughlocal communities and metacommunities may be experi-encing a net increase in diversity.Nevertheless,it is habitat loss,and not the globalization ofEarths biota,that principallythreatens a substantial and permanent reduction in global bio-diversity (Rosenzweig 2001). Caveats and exceptions Even ifnew species seldom drive long-term residents to com-plete extinction through competition,they can sharply reducethe numbers ofonce abundant long-term residents (Porterand Savignano 1990,Petren and Case 1996,Carlton et al.1999).However,it does not necessarily follow that changes inthe dominanceabundance distribution ofa community,oreven local extirpations oflong-term residents,constitute anecological or evolutionary catastrophe,although there may be 486BioScience May 2003 / Vol.53 No.5 Articles the community and metacommunity levels.However,theincrease at the community level would be expected to begreater in the Type 4 dispersal system,because communitieswould regularly receive new species from other communitieswithin their metacommunity as well as from other meta-communities.Owing to the high rates ofdispersal among bothlocal communities and metacommunities,endemism at bothlevels would be expected to be the lowest among the four dis-persal systems.Ofcourse,some long-term resident species have gone extinct on both island and continental landmasses duringthe past several hundred years.Has the number ofnewspecies actually exceeded the number ofextinctions,therebysupporting the models prediction that biodiversity shouldbe increasing at the community and metacommunity levels in a Type 4 dispersal system? In the vast majority ofcases,the answer is yes (Mooney and Cleland 2001,Rosen-zweig 2001,Sax et al.2002).Where good data sets exist,regional biodiversity levels have almost always increasedduring the past few centuries.For example,the introductionand naturalization ofnew vascular plant species into the stateofMinnesota (from outside the Midwest) during the pastseveral hundred years have increased the stateplant diversity by almost 25% (Owenby and Morley 1991).Naturalized species have increased the diversity ofthe Cal-ifornia flora by 20% (Randall et al.1998).Considering allofNorth America,naturalized species have increased thecontinents vascular plant diversity by more than 20%(BONAP 1999).Introductions ofnew species have had aneven larger impact on the biodiversity ofislands.For example,the naturalization ofintroduced vascular plants onPacific islands has doubled the species richness ofthe floraon most ofthese islands (Sax et al.2002).As is the case with plants,the most dramatic increases inlocal and regional animal biodiversity caused by species introductions occur on islands,the fauna ofwhich havesometimes lacked entire classes ofanimals.For example,Hawaii,which had no terrestrial amphibian or reptile speciesand only one species ofmammal (a bat) before the arrival ofhumans,now has a terrestrial faunaincluding 4 amphibian,21 reptile,and 44 mammal species (Miller andEldredge 1996).The primary cause ofextinctionsat the local and metacommunity levels in most areas ofthe world ishabitat loss (Mooney and Cleland2001,Rosenzweig 2001).However,in most cases,this loss in habitat(which translates to a decline in JM Hubbells original theory),has beenmore than compensated by the sub-stantial increase in V, the metacom-munity species proliferation rate,resulting from the high invasion rates (i) among metacommunities.Thus,except in areas where natural environments have all but dis-appeared through habitat destruction,the fundamental bio-diversity number,  (defined using V), and overall speciesrichness have been increasing in most regions ofthe world. Are competition-driven extinctions likely to increase in the future? Recent past extinctions oflong-term residents can seldom beattributed to competition from new species (Mooney and Cle-land 2001,Sax et al.2002).The six bird species that have goneextinct in North America during the past 150 years (passen-ger pigeon,great auk,Labrador duck,Carolina parakeet,Bachmans warbler,and ivory-billed woodpecker) succumbedto overhunting,habitat loss,or both (Pimm and Askins 1995,Montevecchia and Kirk 1996,Chilton 1997).In the Cook Is-lands,no plant extinctions can be definitely attributed tocompetition from introduced species.Likewise,none oftheextinctions ofbird or land snails in the Cook Islands werecaused by competition from introduced species;instead,these extinctions resulted from overharvesting and habitat alteration (in the case ofthe birds) and from predation by anintroduced ant species (in the case ofthe land snails) (Gerald McCormack,Cook Islands Natural Heritage Project,personal communication,2002).Warren B.King (1980) listscompetition as a contributing factor in only 11% ofthe ex-tinctions ofbirds from oceanic islands since 1600,whereas pre-dation and habitat loss are listed as causes in 70% and 30%ofthe cases,respectively.Tompkins and colleagues (2003) haveargued that the decline in the abundance ofthe European redsquirrel is due less to competition from the introduced NorthAmerican gray squirrel,once thought to be the primarycause for the red squirrels decline,than to disease from theparapoxvirus carried by the gray squirrel.As already men-tioned,no introduced plant species are known to have causedany North American long-term resident plant species to goextinct.Extinctions often result from the combined effects ofmultiple processes,and thus,although competition may notbe the primary cause ofmost extinctions,it may prove to be May 2003 / Vol.53 No.5 BioScience485 Articles Table 2.Origin of410 plant species listed by the Pacific Island Ecosystems at Risk(PIER) project as invasive,or potentially invasive,in Indo-Pacific island ecosystems. Number of invasive orPercentage of total Area of originpotentially invasive species(410 species) North America,Central America,South America,and Caribbean19347.1Eurasia52 12.7Africa,Madagascar50 12.2Australia,New Zealand215.1Multiple continents17 4.1Indo-Pacific islands,including Java,New Guinea,Indonesia,and Philippines77 18.8 Note: Species listed as originating both in the Indo-Pacific islands and in one or more continents werelisted in the Indo-Pacific islands column. individual islands comprised a large number ofendemics(Loope 1999).Although individual island diversity was low,the total number ofendemics from the many Pacific islandsmeant that the archipelago itselfwas quite species rich.Forexample,it is estimated that before human colonization,thePacific islands (including New Guinea and New Zealand)were home to between 15% and 20% ofthe worlds birdspecies (Pimm et al.1994,Steadman 1995),despite repre-senting less than 1% ofthe worldIn a Type 2 dispersal system,dispersal is not highly limitedwithin tier 2 (among local communities) but is highly lim-ited within tier 3 (among metacommunities).As described byHubbell,frequent dispersal among local communities willdrive up and sustain diversity at the local level.However,thissame dispersal pattern will suppress metapopulation biodi-versity (tier 2) compared to a Type 1 dispersal system becauseofthe extinctions ofsome rare species within the respectivelocal communities (tier 1).In the Type 2 system,endemismwould be expected to be common at the metacommunity,butnot at the local community,level.An example ofa Type 2 dispersal system would be conti-nental biotas up to approximately a thousand years ago,be-fore humans began to travel regularly between continents,transporting other species with them.Because dispersalamong continents more than a thousand years ago normallywould have been exceedingly rare,endemism at the continentallevel would be expected to be common,as is known to be thecase in most plant and animal groups (Cox and Moore 2000).Although endemism would be expected to be common at themetacommunity level in a Type 2 system,the frequent dis-persals among local communities would homogenize thecomposition ofthese communities.Thus,endemism at thelocal level would be expected to be lower than in a Type 1 dis-persal system.This is borne out by the ranges ofmost conti-nental species,including vascular plants,birds,and mammals,whose ranges typically extend broadly east to west providedsuitable habitat is present.This contrasts with the ranges ofmost long-term terrestrial residents ofthe South Pacificarchipelago,whose home ranges,until the arrival ofhumans,were often restricted to one island or to a few nearby islandsbecause ofsevere dispersal limitation,even though suitablehabitat existed on other islands in the archipelago (LoopeIn a Type 3 dispersal system,dispersal is highly limitedwithin tier 2 but is not highly limited within tier 3.In this sit-uation,species would disperse regularly between metacom-munities but seldom between local communities.Given theinflux ofnew species into both local communities and meta-communities,biodiversity would be expected to increase atboth levels.At the same time,the global exchange ofspecieswould drive down endemism rates at both the local and themetacommunity level.This is probably an unlikely combi-nation ofdispersal patterns,although it seems to describe theSouth Pacific archipelago since the time the first European ex-plorers reached the islands in the 16th century.An analysis of410 plant species that are listed by the Pacific Islands Ecosys-tems at Risk project as invasive or potentially invasive in dif-ferent Pacific islands (PIER 1999) reveals that 81% ofthespecies originated in continental floras,while only 19% orig-inated from the floras ofother Indo-Pacific islands (table 2).These data show that species introduced to the Pacific islandsduring the past 400 years principally involved introductionsfrom outside the archipelago rather than from other islands.Captain Bligh,the notorious English sea captain,is respon-sible for dispersing at least two species between metacom-munities.He is credited with introducing papaya (Caricapapaya, native to Central America) to the Cook Islands in 1792and then introducing breadfruit (Artocarpus altilis, native toTahiti) to Jamaica and the West Indies in 1793 on his returntrip.In a Type 4 dispersal system,dispersal would not be highlylimited within either tier 2 or tier 3.This probably describes,to varying degrees,most regions in the world today.As in aType 3 dispersal system,biodiversity should increase at both 484BioScience May 2003 / Vol.53 No.5 Articles Table 1.Four dispersal systems involving communities and metacommunities and the biodiversity patterns predicted by theextended unified neutral theory. Dispersal system typePredicted biodiversity patterns Type 1Dispersal highly limitedHigh degree of endemism at both the local and metacommunity levels; local among local communitiescommunities exhibit low diversity; metacommunities exhibit high diversityType 2Dispersal not highly limited High degree of endemism at the metacommunity level but not at the local community among local communities; level; high diversity at the local level; diversity at the metacommunity level lower thandispersal highly limited in a Type 1 systemType 3Dispersal highly limited amongLow degree of endemism at both the local and metacommunity levels; higher diversity local communities; dispersalat both local and metacommunity levels than in a Type 1 systemType 4Dispersal not highly limitedLowest degree of endemism at both local and metacommunity levels of the fouramong either local communities dispersal systems; high diversity at metacommunity level; highest diversity at theor metacommunitieslocal community level of the four dispersal systems primarily from plant communities,although Hubbell alsouses a number ofanimal examples to support his unifiedneutral theory. Extension of the unified neutral theory to incorporate global biological invasions Although Hubbell (2001) mentioned introduced speciesin the context ofthe unified neutral theory,he did not tryto incorporate the phenomenon ofglobal biological in-vasions into his two-tiered model.Owing to the structuraland logical similarities ofthe two theories,one can explicitlyincorporate biological invasions at the global level intoHubbells neutral model by adding a third tier,creating ahierarchical system comprising local communities andmetacommunities (the two tiers ofHubbellone meta-metacommunity,the entire Earth.In this sim-ple extension ofthe unified model,the metacommunities(e.g.,continents and large island groups) can be treated aslocal communities within the meta-metacommunity,andglobal invasionscan be treated as dispersals amongmetacommunities.The introduction ofa new species from outside a meta-community is equivalent to an increase in the speciationrate within the metacommunity.(Ifthese introductionsare rare events and typically involve only one or a very fewnumber ofindividuals,the introductions are equivalent towhat Hubbell describes as speciation occurring by point mu-tation mode.Ifspecies introductions occur frequently or ifintroductions typically involve large numbers ofindividuals,the introductions are equivalent to fission mode speciations.)In accordance with the fluctuating resource availability the-ory,once an introduced species has entered a local commu-nity through an invasion window (i.e.,by usurping availablespace and resources before one ofthe residents could do so),it is now a player in the game just as much as any ofthe long-term residents.From that point on,the new species will sim-ply be part ofthe local ecological system,being defined by thesize (total number ofindividuals) ofthe metacommunity, JM; by the speciation rate, v; and by the dispersal rate within themetacommunity, m. None ofthese defining factors arechanged by the introduction ofthe new species.To explicitly incorporate global invasions into the neutraltheory,one could redefine v (defined as speciation rate withina metacommunity in Hubbells original model) as the speciesproliferation rate (V), which equals the sum ofthe speciationrate occurring within the metacommunity, v, gration or invasion rate, i, ofnew species into the meta-community from other metacommunities.(In a similar ex-tension ofHubbells model,Chave and colleagues (2002)combined immigration and speciation rates in the same wayand referred to the collective term as the speciation rate.) Thus,Hubbells fundamental biodiversity number,  , originally de- JMv, would now be defined as 2 JMV, where V = v + i. Owing to the ongoing and rapid globalization ofEarthbiota,the invasion rates experienced by most metacommu-nities will normally be quite high relative to their speciationrate.As a result,under todays conditions (where i È v), Hubbells fundamental biodiversity number,  , would belargely defined by a metacommunitys size and invasion rate.This means that biological invasions occurring on a globalthat is,dispersal taking place among Hubbellcommunities (continents and island groups)have signifi-cantly increased the value of  , and,as described below,would be expected to significantly alter the biodiversity andbiogeographic patterns at the local,metacommunity,andglobal scales. Biodiversity at the community and metacommunity levels In a three-tiered system,dispersal limitation can be appor-tioned among the three levels in four different ways (table 1).In a Type 1 dispersal system,highly limited dispersal amonglocal communities and metacommunities would result in ahigh degree ofendemism at both the local and regional lev-els,because new species appear at both levels primarilythrough speciation events rather than immigration.Themetacommunity,being the sum ofmany endemic-dominated communities,would be very diverse.However,individual communities would be expected to exhibit low levels ofbiodiversity,because their species pools depend almost entirely on speciation events for recruitment ofnewspecies.A good example ofa Type 1 dispersal system wouldbe the terrestrial communities ofSouth Pacific islands beforehuman colonization.During this time,there was very little dis-persal among the islands or between the Pacific archipelagoand metacommunities (continents) in other parts oftheworld.The result was that the species-poor communities of May 2003 / Vol.53 No.5 BioScience483 Articles Figure 2.Steady-state dominancediversity distributions for simu-lated metacommunities and local communities experiencing lowand high dispersal rates (m) among the local communities (m =0.005 and m = 0.5,respectively).The graph at left shows that meta-community diversity declines with increasing dispersal rates,whilethe graph on the right shows that diversity at the local communitylevel increases with increasing dispersal rates.In the simulationshown,metacommunity size (JM) = 7056,the fundamental biodi-versity number (  ) = 10,and the size oflocal communities is 16 individuals.Redrawn from Hubbell (2001). Species rank in abundance Species rank in abundance Metacommunity Local community Percentage relative abundance Percentage relative abundance m = 0.005 m = 0.5 m = 0.5 m = 0.005 The fluctuating resource availability theory The fluctuating resource availability theory ofinvasibilityholds that a mechanistic relationship exists between invasi-bility and resource availability and that changes in invasibil-ity often result from changes in the competition intensity fromresidents,which in turn result from fluctuations in resourceavailability (figure 1).This theory is supported by field datashowing that short-term increases in resource availability inplant communities can temporarily reduce or suspend com-petition from resident vegetation,thereby increasing the invasibility ofthe environment (Davis and Pelsor 2001).Thetheory assumes that whatever individual arrives first at apool ofunused resources is able to co-opt them.Whether thatindividual represents a species new to the region or one thathas resided in the community for a long period oftime,it nowhas the opportunity to become established in the community.Under the fluctuating resource availability theory,commu-nities are equal opportunity employers,not discriminatingamong species based on their geography oforigin.An im-portant component ofthis theory is that it does not requirenew species to be ecologically different from resident speciesin order to successfully colonize a new environment,as hasbeen argued in the past (Thompson 1991).In this way,the theory is similar to that presented by Moyle and Light (1996),who also emphasized the importance offluctuating envi-ronmental conditions on invasion success,and to lottery-based models ofcommunity assembly (Sale 1977) and othernonequilibrium models that emphasize the role played by dis-turbances and fluctuating resources in forestalling competi-tive exclusion (Huston 1994). The unified neutral theory of biodiversity and biogeography The unified neutral theory ofbiodiversity and biogeographyestablished that biodiversity patterns at multiple scales can beaccurately predicted using a model based on random butlimited migration,random speciation,and random fluctua-tions in species abundances (the latter a process Hubbell re-ferred to as ecological drift).In this model,individuals ofall species within a community trophic level are consideredto be per capita ecological equivalents,meaning they exhibitthe same probabilities ofgiving birth,dying,and migrating.In Hubbells model,individual organisms interact in a zero-sum game in a biotically saturated landscape,meaning that,although disturbance may create small patches ofresourcesfor a briefperiod,no significant amount ofspace or other lim-iting resource goes unused for long.Using a two-tiered model consisting ofcommunities withina metacommunity,Hubbell showed how changes in disper-sal rates among local communities would be expected to af-fect biodiversity patterns at both levels.He concluded that highrates ofdispersal among local communities should increasethe diversity at the local level while reducing the diversity atthe metacommunity level as abundant and widespread speciesdrive some rare and local species extinct.Conversely,Hubbellargued that although low dispersal rates result in low diver-sity in local communities,the infrequent introduction ofnew species to these communities allows many rare local en-demic species to survive,thereby sustaining a high diversityat the metacommunity level (figure 2).In the development ofhis theory,Hubbell defined a damental biodiversity number,  , described by Hubbell as adimensionless,fundamental quantity that appears pervasively[in the unified neutral theory] at all spatio-temporal scales.The variable  is calculated as 2 JMv, where JM = the size ofthemetacommunity and v = the speciation rate within the meta-community.Species richness and relative abundance pat-terns in local communities and in metacommunities arebelieved to be a function of  and dispersal rates, m, local communities (Hubbell 2001). Commonalities of the two theories A fundamental similarity ofboth theories is that they are dis-persal rather than niche based.Neither theory requires thatnew species possess traits different from those ofthe residentspecies to colonize and become established in a community.Both theories emphasize spatiotemporal variability in habi-tat characteristics (e.g.,availability ofresources).Both theo-ries emphasize the importance ofstochastic events andprocesses,particularly the coincident synchronization ofdis-persal episodes with the availability ofunderutilized resource-rich patches.Both theories focus on diversity issues withinindividual trophic levels,and both were developed using data 482BioScience ¥ May 2003 / Vol.53 No.5 Articles Figure 1.The theory offluctuating resource availabilityholds that a communityÕs susceptibility to invasion increases as resource availability (the difference between gross resource supply and resource uptake) increases.Resource availability can increase because ofa pulse in resource supply (A  B),a decline in resource  C),or both (A  D).In the plot,resourceavailability,and hence invasibility,increases as the trajectory moves farther to the right or farther belowthe supply/uptake isocline (where resource uptake =gross resource supply).Image modified from Davis and Gross resource supply Resource supply-uptake isocline Resource uptake May 2003 / Vol.53 No.5¥ BioScience481 Articles T he globalization of EarthÕs biota is transforming local and regional floras and faunas.From the smallest,most remote islands to the largest continents,the intentionalor accidental introduction ofnew species is altering the com-position and ecology oflong-established biological commu-nities (Wilson 1992).As evidenced by the recent introductionsofthe Asian longhorned beetle (Anoplophora glabripennis) the West Nile virus into the United States,some species introductions can inflict economic harm and threaten humanhealth.In addition,it is widely reported that introducedspecies are threatening many resident species with extinction(Elton 1958,Wilcove et al.1998).There have been numerous well-documented extinctionsoflong-term resident species caused by the introduction ofpredators and pathogens,particularly in spatially restrictedenvironments such as islands and lakes (King C 1984,Gill andMartinson 1991,Kaufman 1992,Fritts and Rodda 1998,Loope 1999).However,there are surprisingly few instances inwhich extinctions ofresident species can be attributed tocompetition from new species,either on continents or islands(Mooney and Cleland 2001,Sax et al.2002).For example,more than 4000 plant species introduced into North Amer-ica north ofMexico during the past 400 years are naturalized(established to various degrees),and these new species nowrepresent nearly 20% ofthe continentÕs vascular plant species.Yet there is no evidence that even a single long-term residentspecies has been driven to extinction,or even extirpatedwithin a single US state,because ofcompetition from an introduced plant species (John T.Kartesz,Biota ofNorthAmerica Program,University ofNorth Carolina,personalcommunication,2002).The paucity ofdocumented extinctions caused by com-petition from new species could mean that extinctions ofthis type simply take longer to occur.Or it could mean that,compared to intertrophic interactions and habitat loss,com-petition from introduced species is not likely to be a commoncause ofextinctions oflong-term resident species.The likelythreat ofintroduced species to resident controphics (speciesin the same trophic level) can be assessed with the help ofex-isting biodiversity and extinction data sets and oftwo recenttheories,one focusing on biological invasions and the otherfocusing on geographic patterns ofbiodiversity.The firsttheory,the fluctuating resource availability theory (Davis etal.2000),was developed to account for changes in the inva-sibility ofcommunities (susceptibility to being colonized bynew species).The second theory,the unified neutral theoryofbiodiversity and biogeography (Hubbell 2001),was pro-posed to account for patterns ofbiodiversity at the commu-nity and metacommunity levels. Mark A.Davis (davis@macalester.edu) is a professor in the Department ofBiology at Macalester College,1600 Grand Ave.,St.Paul,MN 55105.© 2003American Institute ofBiological Sciences. Species Threaten Biodiversity? MARK A. DAVIS The introduction ofnew predators and pathogens has caused numerous well-documented extinctions oflong-term resident species,particularly inspatially restricted environments such as islands and lakes.However,there are surprisingly few instances in which extinctions ofresident speciescan be attributed to competition from new species.This suggests either that competition-driven extinctions take longer to occur than those causedby predation or that biological invasions are much more likely to threaten species through intertrophic than through intratrophic interactions.Thelikely threat ofintroduced species to resident controphics (species in the same trophic level) can be assessed with the help ofexisting biodiversityand extinction data sets and oftwo recent theories: (1) the fluctuating resource availability hypothesis,developed to account fvasibility ofcommunities,and (2) the unified neutral theory,proposed to account for patterns ofbiodiversity at the community and metacommu-nity levels.Taken together,theory and data suggest that,compared to intertrophic interactions and habitat loss,competition from introducedspecies is not likely to be a common cause ofextinctions oflong-term resident species at global,metacommunity,and even most community levels.Keywords: biological invasions,unified neutral theory,fluctuating resource availability hypothesis,extinction threats,competiti