Biological Consequences of Ecosystem Fragmentation: a Review Author(S): Denis A

Biological Consequences of Ecosystem Fragmentation: A Review Author(s): Denis A. Saunders, Richard J. Hobbs and Chris R. Margules Reviewed work(s): Source: Conservation Biology, Vol. 5, No. 1 (Mar., 1991), pp. 18-32 Published by: Wiley-Blackwell for Society for Conservation Biology Stable URL: http://www.jstor.org/stable/2386335 . Accessed: 11/10/2012 13:03

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Wiley-Blackwell and Society for Conservation Biology are collaborating with JSTOR to digitize, preserve and extend access to Conservation Biology.

http://www.jstor.org Review

BiologicalConsequences of Ecosystem Fragmentation:A Review

DENIS A. SAUNDERS CSIRO Division of Wildlifeand Ecology LMB 4 P.O. Midland WesternAustralia 6056, Australia RICHARDJ. HOBBS CSIRO Divisionof Wildlifeand Ecology LMB 4 P.O. Midland WesternAustralia 6056, Australia

CHRIS R. MARGULES CSIRO Divisionof Wildlifeand Ecology P.O. Box 84 Lyneham AustralianCapital Territory2602, Australia

Abstract: Research on fragmentedecosystems has focused Resumen: La investigacionsobre los ecosistemasfragmen- mostlyon thebiogeographic consequences of thecreation of tados se ha enfocado principalmente en las consecuencias habitat "islands" of differentsizes, and has provided littleof biogeograficasde la creacion de "islas" de habitat de difer- practical value to managers.However, ecosystem fragmen- entestamaniosy ha sido de muypoco valorpracticopara los tation causes large changes in thephysical environmentas manejadores del recurso.Como quiera que sea la fragmen- well as biogeographicchanges. Fragmentation generally re- tacion de los ecosistemascausa grandes cambios en el medio sults in a landscape thatconsists of remnantareas of native ambiente fisico asi como en el ambito biogeografico.La vegetationsurrounded by a matrixof agriculturalor other fragmentacionresulta generalmente en terrenosque consis- developed land. As a result,fluxes of radiation; momentum ten de areas remanentesde vegetacion nativa rodeada de (i.e., wind), water,and nutrientsacross the landscape are una matriz de tierrasagricolas u otrasformas de uso de la altered significantly.These in turn can have importantin- tierra Como un resultado de esto, el flujo de la radiacion; fluences on thebiota withinremnant areas, especiallyat or del momentum(ej. el viento),del agua y de los nutrientesa near theedge betweenthe remnant and thesurrounding ma- travesde la tierrason alterados significativamente.Esto en trix. The isolation of remnant areas by clearing also has su turno,puede influenciare la biota dentro de las areas importantconsequences for the biota These consequences remanentes,especialmente en o cerca de los limitesentre los varywith the timesince isolation; distancefrom otherrem- remanentesy la matriz que los rodea El aislamiento de las nants, and degreeof connectivitywith otherremnants. The areas remanentesporla tala tambien tieneimportantes con- influencesof physical and biogeographicchanges are mod- secuenciaspara la biotay estas consecuencias varian con, el ified by the size, shape, and position in the landscape of tiempodesde el momentodel aislamiento, la distancia hasta individual remnants,with larger remnants being less ad- los otros remanentesy el grado de coneccion entreellos. La verselyaffected by thefragmentation process. The dynamics influencia de los cambios fisicos y biogeograficoses modi- ficada por el tamanio,la formay la posicion en el terrenode remanentesindividuales siendo los remanentesgrandes los menos afectados adversamantepor el proceso de fragmen- Paper submittedMay 1990; revisedmanuscript accepted 2, Septem- tacion. La dinamica de las areas remanentesson dirigidas ber 20, 1990.

ConservationBiology Volume 5, No. 1, March 1991 Saunderset al. ConsequencesofFragnentation 19 of remnantareas are predominantlydriven by factors aris- predominantementepor factores que surgen en el terreno ing in the surroundinglandscape. Management of, and re- circundante.El manejoy la investigacionde los ecosistemas search on, fragmented ecosystemsshould be directed at fragmentadosdeberi'a de dirigirsetanto al entendimientoy understandingand controllingthese external influencesas control de estas influencias externascomo a las biotas re- much as at the biota of theremnants themselves. There is a manentesen si. Hay unafuerte necesidadpara el desarrollo strongneed to develop an integratedapproach to landscape de un enfoque integradoen el manejo de tierrasque coloca managementthatplaces conservationreserves in thecontext a las reservaspara la conservacionen el contextodel terreno of the overall landscape. en general.

Introduction the originalvegetation has been removed (Saunders et al. 1987a). Remnantsoften have been called habitatis- Since the developmentof agriculture,the naturalvege- lands and the changes thatresult from the isolationof tationcover of every continentexcept Antarcticahas these islandshave been the subject of considerablede- been extensivelymodified. A cycle of agriculturalde- bate in the literature.This debate has centeredmostly velopmentfollowed by overexploitationof the land has on the equilibriumtheory of islandbiogeography (Mac- been repeated throughoutrecorded history.Forman Arthur& Wilson 1963, 1967) and its applicabilityto (1987) quotes Plato's(ca. 2350 BP) descriptionof how conservation. In particular,the importance of size, an area ofancient Greece was strippedof its soil follow- shape,and designof single reserves and reservesystems, ing clearingand grazing,leaving "the mere skeletonof extinctionand colonizationrates, and species-arearela- theland." Overexploitation and theuse ofinappropriate tionshipshave been much discussed (Wilson & Willis agriculturalpractices have led to desertencroachment, 1975; Diamond 1975, 1976; May 1975; Terborgh1976; as in the Sahel region in NorthAfrica (Le Houerou & Whitcombet al. 1976; Simberloff& Abele 1976a4 1982; Gillet 1986; Ehrlich& Ehrlich1987), and to extensive Pickett& Thompson 1978; Game 1980; Marguleset al. loss of soil, oftenwith disastrousand dramaticconse- 1982; Boecklen & Gotelli 1984). Particularattention quences, as in the Dust Bowl of the UnitedStates (Hud- has been paid to the question of whetherone large re- son 1981). serve could preserve more species than several small The process of land clearingand the consequent en- reservesadding up to the equivalentarea of the larger vironmentaldegradation is continuingrapidly in many reserve(the so-calledSLOSS, or "singlelarge or several regionssuch as SoutheastAsia and SouthAmerica, par- small" debate; Simberloff& Abele 1976b, 1984; Gilpin ticularlyin areas of tropicalrain forest(Myers 1988). & Diamond 1980; Higgs & Usher 1980; Jairvinen1982; Australiaprovides an example of recentagricultural de- Willis 1984). velopment,with vast areas being cleared over the last These questionshave been reviewedelsewhere (Sim- 100 yearsfor cereal croppingand stockgrazing (Saun- berloff1988) and are not examinedin detailhere. While ders & Hobbs 1989; Hobbs & Hopkins 1990; Saunders of theoreticalinterest, most of these issues are of little et al. 1990). In some regions over 93% of the native practicalvalue in managingfragmented systems (Zim- vegetationhas been removedand the agriculturalland merman & Bierragaard1986; Margules 1987; Hobbs that has replaced it is subject to extensivewind and 1988; Margules& Stein 1989). The species-areaequa- water erosion or soil salinization,with consequentpol- tion,for example, may give a managera rough idea of lutionof water suppliesfor drinking and irrigation. how manyspecies will be maintainedon a remnantof a One legacy of the extensiveremoval of nativevege- given area, but will yield absolutelyno informationon tationis thatthe remainingvegetation is usuallyin frag- the practicalissue ofwhich habitatscontribute most to mentedpatches across the landscape.These patches or species richnessor on which species are most likelyto remnantsare situatedin differentpositions in the land- be lostfrom the remnant. Simberloff ( 1986) stated,"It is scape and on differentsoil types,possess differentveg- also sad thatunwarranted focus on the supposed lessons etationtypes, and varyin theirsize, shape, isolation, and of island biogeographytheory has detractedfrom the typeof ownership.Over much of the world,conserva- maintask of refugeplanners, determining what habitats tionof regional biotas depends entirely on theretention are importantand how to maintainthem." Margules and managementof these remnants.Conservation man- (1986), in a discussionabout two conservationevalua- agersare thereforefaced with the dual issuesof whether tion exercises,notes that,"no panel membersin either the remnantshave any practical conservationvalues, exercise consideredthe species-arearelationship or the and ifthey do, of how theymust be managedto retain equilibriumtheory of islandbiogeography in theireval- these values (Saunders et al. 1987a). uations;at least not explicitly."In addition,the debates In thispaper we use the word remnantto defineany about reservedesign and SLOSS are of limitedrelevance patch of nativevegetation around which most or all of because, withvery few exceptions, managers of conser-

ConservationBiology Volume 5, No. 1, March 1991 20 ConsequencesofFragmentaton Saunderset a). vation reservesare faced with a faitaccompli. Conser- etation, especially where the native vegetation was vation considerationshave rarelybeen taken into ac- dense beforeclearing. Removing native vegetation and count duringthe developmentof areas foragriculture, replacingit with crop species with differingarchitec- mining,forestry, or othersuch uses. Conservationman- ture and phenologyalters the radiationbalance by in- agersmust work with the remnantsleft following these creasingthe solar radiationreaching the groundsurface developmentsand virtuallynever have the opportunity duringthe day,changing the albedo, and increasingre- to design a reserve networkbefore an area is frag- radiationat night.These featuresvary depending on mented.There is an increasingneed to utilize design time of year; ploughing,crop growth,and harvesting .criteriato improve conservationnetworks in already produce an alternationof bare groundand varyingde- fragmentedareas, but thisrequires a clear understand- grees of vegetationcover (Geiger 1965; Milthorpe& ingof the problems created by fragmentationin thefirst Moorby 1974). In cleared areas, in general, daytime place. temperaturesare higherand nighttemperatures lower Hence we believe thatresearch and discussionshould thanin naturallyvegetated areas. This leads to greater focus on practicalissues relatingto the impactof frag- temperatureranges both at the surfaceand in the upper mentationon naturalecosystems and managingrem- layersof the soil, and an increased incidence of frost nants for conservation(Saunders et al. 1987b). We (Geiger 1965). share the fearof Noss and Harris(1986) thatconserva- These changes in the cleared parts of the landscape tionagencies have not realizedthe importantbiological impingeon the remnantnative vegetation, especially at consequences of ecosystem fragmentationand have the edge between the two. Except near the equator,the thereforenot developed policies to manage theirrem- orientationof the edge affectsthe degree to which solar nantsto maintainconservation values. The aim of this radiationincreases within a remnantat differenttimes of paper is to point out the physicaleffects of fragmenta- year (Geiger 1965; Wales 1972; Ranneyet al. 1981). tion, the biological consequences for naturalecosys- Latitudealso affectsradiation input, and at highlatitudes temsof these effects,and the optionsavailable for con- especially,where solar angles are low, a remnantedge servationresearch and management. can receive significantlymore solar radiationthan un- fragmentedareas receive (De Walle & McGuire 1973; Characteristicsof Fragmented Ecosystems Hutchinson& Matt1976,1977). Airtemperatures at the edge ofa forestremnant can be significantlyhigher than Fragmentationof the landscape produces a series of those foundin eitherthe interiorof the remnantor the remnantvegetation patches surroundedby a matrixof surroundingagricultural land (Geiger 1965; Kapos differentvegetation and/or land use. Two primaryef- 1989). fectsof this are an alterationof the microclimatewithin The consequences of increasingsolar radiationat the and surroundingthe remnantand the isolation of each edges of remnantsare not clear. The indicationsare, area fromother remnantpatches in the surrounding however,that a differentsuite of species may come to landscape. Thus, in a fragmentedlandscape there are occupy thisaltered habitat. Lovejoy et al. (1986) report changesin the physicalenvironment as well as biogeo- thatfragmentation in tropicalforests results in the rapid graphicchanges. Most discussionsof habitatfragmenta- growthof vines and other secondary vegetationin a tion have concentratedon the biogeographicaspects, 10-25 m striparound the remnantedge. This may ef- and the physicalchanges have received littleattention. fectivelyseal offthe remnantand maintainan environ- All remnantsare exposed to thesephysical and biogeo- ment within the remnantsimilar to that that existed graphicchanges to a greateror lesser degree,but their priorto fragmentation(F. Crome,personal communica- effectsare modifiedby the size, shape, andposition in tion). This also occurs to some extentin temperatere- the landscape of individualremnants. We examinefirst gions(Gysel 1951; Trimble& Tryon1966; Ranneyet al. the physicaleffects of fragmentationand then discuss 1981). the operationof the modifyingfactors. Shade-tolerantspecies may become restrictedto the interiorparts of the remnant,with differentspecies re- Changesin Microclimate quiringdifferent distances from the edge. The compo- Fragmentationof the landscaperesults in changesin the sitionof remnantedges maybe affectedby edge aspect physical fluxes across the landscape. Alterationsin (Wales 1972; Palik & Murphy1990). Distinctsets of fluxesof radiation,wind, and water can all have impor- "interior"and "edge" species have been recognizedin tanteffects on remnantsof nativevegetation. landscapes thathave been fragmentedfor a long time, forinstance in the easternUnited States (Ranney et al. 1. RADIATIONFLUXES 1981). The energybalance ofa fragmentedlandscape will differ Nutrientcycling processes may be affectedby in- markedlyfrom one witha completecover ofnative veg- creased soil heatingand its effecton soil microorganism

Conservation Biology Volume 5, No. 1, March 1991 Saunderset aW. ConsequencesofFragmentation 21 and invertebratenumbers and activity(Klein 1989; these processes must be significantlymodified in rem- Parker1989), on litterdecomposition, and on soil mois- nantvegetation areas, where thewind profilewill notbe ture retention.Changes in the radiationbalance also in equilibriumwith the remnantvegetation (Jarvis & affectlarger fauna both directlyand indirectlythrough McNaughton1986). Given the fetch requirements,a altering resource availability(via changes in plant woodlandwith 20-m-tall trees would need to be at least growthand phenology).Lovejoy et al. (1986) attribute 2-4 km wide before wind profileswould resemble changesin butterflycommunity composition in tropical those in an unfragmentedsituation. The implicationsof forestfragments partly to increased insolationwithin thisfor plant gas exchange and growthhave not been smallremnants. Increased radiation load and dessication examined,but could be significant. ratesmay lead to reduced foragingopportunities. Saun- A more obvious effectof landscape fragmentationis ders (1982) considered thatelevated temperaturesin thatremnants are subjected to increased exposure to fragmentedlandscapes reduced the foragingtime avail- wind. This may resultin damage to the vegetation,ei- able to adultCarnaby's cockatoos (Calyptorbynchusfu- therthrough direct physical damage (Moen 1974; Grace nereus latirostrus)and contributedto theirlocal ex- 1977), or by increasing evapotranspirationwith re- tinction. Alternatively,some species dependent on duced humidityand increaseddessication (Tranquillini temperaturethresholds may be able to foragefor longer 1979; Lovejoyet al. 1986). Direct physicaldamage can periods.For instance,the stronglydominant ants of the take the form of wind pruning (Caborn 1957) or Iridomyrmexgenus are knownto forageonly when in- windthrowof trees.Distinct edge structureshave been solationand temperaturesare high,and other ant spe- foundto develop at the edges oftree plantations (Fraser cies forageonly when Iridomyrmexis absent (Green- 1972), and thisis likelyto be the case forremnant areas slade & Halliday 1983; Andersen 1987). Increased also (see Ranneyet al. 1981 for a discussion of edge insolation and ambient temperaturescould then in- structures).Trees near the edge of recentlyisolated crease the foragingtime available to Indomyrmexbut remnantsare particularlyat risk of windthrowsince reduce it forsubordinate species. they have maturedwithin a closed canopy and have Alteredtemperature regimes can also have the effect thereforedeveloped in the absence of strongwinds and of destabilizingcompetitive, predator-prey, and para- lack the necessary support mechanismsto deal with siticinteractions. Geiger (1965) quotes the example of such winds. Windthrowof dominanttrees results in the timberpest Ocneria monacha, which lays eggs on changesin the vegetationstructure and increasedavail- treetrunks. Elevated springtime temperatures on trunks abilityof regenerationgaps, allowingrecruitment, par- at the forestedge allow larvae to emerge beforetheir ticularlyof pioneer or light-demandingspecies. In- parasites,which emerge fromthe cooler forestfloor. creased litterfall through tree damage is likelyto alter This givesthe pest a head startand resultsin population soil surface characteristicsand hence the habitat of buildups. ground-dwellingfauna. Similarly, increased exposure to wind may remove loose bark and reduce the substrate availablefor bark-dwelling invertebrates, and hence also 2. WIND reduce theiravailability as a food resource. Increased Withthe removalof surroundingvegetation, the entire wind turbulencedue to clearing has been shown to patternof momentumtransfer over the landscape is al- affectthe breedingsuccess of birdsby creatingdifficul- tered.As air flowsfrom one vegetationtype to another, ties in landingdue to wind shear and vigorouscanopy the wind profileadjusts to the new roughnesscharac- movement(Brett 1989; Revilleet al. 1990). In the case teristics.When air flowingover one vegetationtype of tropicalforests, fragmentation can resultin hot, dry comes to a boundarywith a new vegetationtype, the windsblowing into remnant areas fromthe surrounding upper partof the wind profileinitially retains the char- cleared areas,with the probable resultof increasedtree acteristicsof the previous vegetation type, but thelower mortality(Lovejoy et al. 1986) and preventionof regen- part takes on new characteristicsreflecting the rough- eration of species whose successfulestablishment re- ness of the new vegetationtype. The wind profiledoes quires humid conditions or persistentsoil moisture not fullyequilibrate with the new vegetationfor some (anzen 1986). This mayalso be importantin the regen- distance.Rules ofthumb for wind profilemeasurements eration of species in other areas, such as those with give the minimum"fetch" (i.e., the minimumdistance mediterraneanclimates, where successfulestablishment fromthe vegetationboundary that will ensure thatthe requires adequate soil moisture (e.g., Gordon et al. profilehas takenon the characteristicsof the vegetation 1989; Williams& Hobbs 1989). This effectwill be less- under study) as 100-200 timesthe heightof the vege- ened in cases where theedges ofremnants are sealed off tationunder study(Monteith 1975; Grace 1983; Jones by rapid secondarygrowth. 1983). Turbulenttransfer is importantfor the transport Increasedwind speeds at remnantedges have the sec- of atmosphericgases, and gas fluxesabove vegetation ondaryeffect of increasingthe transferof materialsuch are controlledby these processes.It follows,then, that as dust and seeds in fromthe surroundingmatrix. Gei-

ConservationBiology Volume 5, No. 1, March 1991 22 ConsequencesofFragnentation Saunderset al. ger (1965) gives an example where particulatematter naturalwetland now requirepumping systems to retain depositionat the edge of a forestremnant increased by adequate waterlevels (Hutchinson1980; Rowell 1986). 40% over that in the open. Transferof nutrientsby Changes in water fluxes and associated particulate saltationof surfacesoil particlesis also possible, and and nutrientfluxes can have importantinfluences on stronggradients in soil nutrientlevels have been found thebiota of remnants. Altered patterns of erosion lead to at theedges ofremnant areas (Muir 1979; Cale & Hobbs changesin drainagepatterns and the productionof new 1991). Wind can deposit seeds of nonnativespecies substratesfor plant colonization.Of particularimpor- over considerabledistances into remnantareas (Hobbs tance is the depositionof nutrient-richmaterial in run- & Atkins1988). Transferof insectsand disease organ- on areas,which can act as a focusfor invasion by species isms into remnantareas may also be increased. requiringdisturbance and/or nutrient enrichment for successful establishment (Hobbs & Atkins 1988). Changes in surfaceand soil moisturelevels could also 3. WATERFLUX lead to changes in decompositionrates, altered seed- Fragmentationof the landscape resultsin modification bed characteristics,and changes in habitatfor ground- of the local water regimeby alteringthe variouscom- dwellingfauna. ponents of the hydrologicalcycle. Removal of native vegetationchanges the ratesof rainfallinterception and Isolation evapotranspiration,and hence changes soil moisture Landscape fragmentationhas two importantconse- levels (Kapos 1989). The pathwaysby which water pen- quences forthe biota. First,there is a reductionin the etratesthe soil mayalso be markedlyaltered (Bormann total area of habitatavailable, with possible increased & Likens 1979; Nulsen et al. 1986; Peck & Williamson densitiesof surviving fauna in theremnants, and second, 1987; Sharmaet al. 1987; Bell 1988). Replacementof the habitatthat is leftis brokenup into remnantsthat deep-rootedperennial species withherbaceous crop or are isolated to varyingdegrees (Lovejoy et al. 1984, pasturespecies leads to greatlyreduced evapotranspi- 1986; Haila & Hanski 1984; Wilcove et al. 1986). The ration and increased surface-and groundwaterflows. timesince isolation,the distancebetween adjacent rem- The hydrologicalsystem in generalbecomes much less nants,and the degree of connectivitybetween themare bufferedwith more extremerun-off events (Hornbeck all importantdeterminants of the biotic response to 1973; Simons1989). Increasedsurface water flows lead fragmentation. to increasederosion and transportof particulate matter (e.g., Bormannet al. 1974). Topsoil removedfrom high in the catchmentends up as sedimentin the riversys- 1. TIME SINCE ISOLlATON tem. Transportof nutrientsinto streamsalso increases Upon isolation,a remnantis likelyto have more species (Likens et al. 1970). Rises in water tables can bring thanit will be capable of maintaining,and species will storedsalts to the surfaceand cause secondarysalinity, be lost as the changes broughtabout by fragmentation with considerableimpacts on both remnantvegetation take effect(Miller & Harris1977; Miller 1978; Wilcox and the surroundingagricultural matrix (Peck 1978; 1980; Harris1984). This process of "species relaxation" Williamsonet al. 1987). Movementof stored salts,nu- is consideredan inevitableconsequence of area reduc- trients,and pesticides washed fromcleared land can tionand isolation,on the basis of islandbiogeographical have significantimpacts on river systems(Kendrick predictions. However, the various mechanisms by 1976; Karr& Schlosser1978). which local extinctionsoccur will resultfrom the phys- The impactof these changes on an individualremnant ical changes discussed above and resultingchanges in depends greatlyon itsposition in the landscape(Swan- biotic interactions.The rate of species relaxationwill son et al. 1988). Remnantsat the top ofa catchmentcan differamong different taxa. The most rapid extinctions be expected to be relativelylittle affected by changesin are likelyin species thatdepend entirelyon nativeveg- water flux,whereas remnantson midslopesand valley etation,those that require large territories,and those bottomswill be moreaffected. Remnants in run-offareas thatexist at low densities.Dispersal behavior and de- can be expected to experience more erosion, while mographywill determinethe response of individual spe- those in run-onareas will experiencemore soil deposi- cies to fragmentation(Karieva 1987). tion,especially on the up-slopeedge. Populationsthat are too smallto be viable maypersist Furtherimpacts on remnantareas can be expected forlong periods simplybecause of the longevityof in- followingmanagement operations in the surrounding dividuals.For example,in remnantsin the WesternAus- matrixthat alter hydrological processes. Thus irrigation, tralianwheatbelt, female trapdoor spiders Anidiops vil- water storage,or drainagemay affect remnant areas. An losus can live for at least 23 years (Main 1987), and extremeexample of thisis foundin the fensof eastern smallAustralian passerines of about 25 g may live over England,where drainagehas led to peat shrinkageand a 20 years (AustralianBird BandingScheme records). It drop of4 m in land level in 130 years.Remnant areas of may take several hundredyears to lose some species

ConservationBiology Volume 5, No. 1, March 1991 Saunderset al. ConsequencesofFragmentation 23 such as long-livedtrees, especially since adultplants are ment.Organism size is also important,and 100 m over often less sensitiveto changed environmentalcondi- agriculturalfields may be a completebarrier to dispersal tions than seedling and juvenile stages.Alterations in forsmall organisms such as invertebrates(Mader 1984) disturbanceregime in remnantareas mayprevent suc- and some species ofbird (Saunders & de Rebeira 1991). cessfulregeneration (Hobbs 1987; Bond et al. 1988). The persistenceof such species on a remnantwould Presence of a species in a remnantis thusno guarantee then depend entirelyon the retentionof enough suit- of its continuedexistence there;successful reproduc- able habitatto maintainsufficient numbers to withstand tion and recruitmentare required.Managers therefore therisks of extinction (Soule 1987a, Ewens et al. 1987). may need to examine the age structureof species on Some evidence existsthat fragmentation of largepopu- remnantsto identifyvulnerable species to be targeted lations into subpopulationsmay decrease the risk of forspecial management. overallspecies extinctioneven thoughlocal extinctions Time since isolationwill thereforedetermine how far may occur (Higgs 1981; Quinn & Hastings 1987). It down the "relaxationtrack" any given remnant has trav- seems likelythat differentspecies will respond differ- eled. Recentlyisolated remnantscan be expected to entlyto the creationof subpopulationsand thatknowl- continuelosing species; thisprocess may continuefor edge of the details of an organism'sbehavior will be relativelylong periods in the absence of interventive necessaryto predict its response (Karieva 1987; Mer- management(Soule et al. 1988; Saunders1989). Long- riam 1991). isolated remnantscan be expected to have lost a pro- portionof the species originallypresent, and gained an 3. CONNEClvmY additionalcomponent of invadingspecies thatare capable of establishingin the fragmentedsystem. It is thus Associatedwith the effectsof distanceis the degree to wrong to consider only species numbersand not spe- which individualremnants are connected in some way cies compositionwhen discussingspecies diversityin to adjacentareas. The issue of connectivityand the use- remnantareas: species numberscan potentiallyincrease fulnessof corridorsconnecting remnants has received in fragmentedsystems where invasiveand edge species increasingattention in the literature(MacClintock et al. can establish,but the numbers of species originally 1977; Wegner& Merriam1979; Baudry 1984; Forman foundin thearea maycontinue to decline(Verner 1986; & Baudry1984; Merriam1984; Harris1984, 1985; Fah- Murphy1989; Webb 1989; Harris& Scheck 1991). rig & Merriam1985; Noss & Harris1986; Bridgewater It is commonlyassumed thatat some stage the rem- 1987; Simberloff& Cox 1987; Noss 1987; Soule et al. nantwill reequilibratewith the surroundinglandscape. 1988), and was the subject of a recent symposium It is, however,questionable whether a new stableequi- (Hobbs et al. 1990; Saunders& Hobbs 1991a). Corri- libriumwill be reached since the equilibrationprocess dors are generallybelieved to provide benefitssuch as is liable to be disruptedby changingfluxes from the enhanced biotic movement,extra foragingareas, ref- surroundingmatrix, disturbances, and influxof new in- uges duringdisturbances, and enhancementof the aes- vasive species. The finalequilibrium can be likenedto theticappeal of the landscape. In some areas theysig- an idealized endpointthat is neverlikely to be reached, nificantlyadd to the area of native vegetation left in much the same fashionas the climaticclimax is now followingfragmentation. conceptualized in succession theory.Management of Simberloff& Cox (1987) pointedout thatmost of the remnantareas will thusbe an adaptiveprocess directed workon the value of corridorshas not been sufficiently at minimizingpotential future species losses. controlledto demonstratean unequivocal role in in- creasingimmigration and/or decreasing extinctions. An increasingnumber of studies, 2. DISTANCEFROM OTHER REMNANTS however, now indicate thatcorridors are of value formovement, at least fora The abilityof species to colonize a remnantdepends to subset of the biota (papers in Saunders & Hobbs, some extenton the distanceof the remnantfrom other 1991a). Simberloffand Cox (1987) also discussed po- areas of native vegetation,be they other remnantsor tentialdisadvantages of corridors,which include facili- nearbyuncleared areas. Colonizingability is relatedto tationof the spread of disease,pests, and fireand other dispersalmode, with wind-dispersed and vagile species disturbances,increased predation,and high costs of more likely to arrive at isolated remnants.However, maintaininglinear remnants with highedge:area ratios. whethersuch species become successfulcolonists de- The relativemerits of corridorsand theirrequired char- pends on physicaland biotic factorssuch as nutrient acteristics(i.e., width,composition, etc.) will varyfrom availabilityand competitiveinteractions (Vepsalainen & place to place and will depend on the targetspecies. Pisarski1982). Animalspecies may have the physical Detailed predictionsof corridorvalue in reducingiso- abilityto disperselong distances, but lack thebehavioral lation of remnantareas are not possible withoutinfor- repertoireto traversethe matrixsurrounding the rem- mationon the movementof individualspecies across nant;the matrixbecomes an effectivebarrier to move- the landscape. Such informationis, however, difficult

ConservationBiology Volume 5, No. 1, March 1991 24 ConsequencesofFragnentation Saunderset al and time-consumingto collect (Saunders & de Rebeira roundingthe remnantbut use the remnantfor shelter. 1991). Nevertheless,while such data are being gath- An example fromWestern Australia is the Wedge-tailed ered, we need to take the approach thatcorridors do Eagle (Aquila audax), which breeds or roosts in rem- have value forbiotic movementand attemptto retaina nantsbut forageson carrionin the surroundingfarm- good corridor networkwherever possible (Harris & land,bringing parts of sheep carcassesback to the roost Scheck 1991; Saunders& Hobbs 1991b). to consumethem. Wool fromcarcasses carries seed that is dropped under the roost tree. Nutrientinput from eagle droppingsand the disturbancecaused by other 4. CHANGESIN THE SURROUNDINGIANDSCAPE scavanginganimals provides a focus for the establish- Removalof the vegetationfrom the area surroundinga mentof nonnativespecies (Saunders,personal observa- remnantleads to the remnantbecoming the onlyarea of tion). Thus, even in the absence of deliberate distur- suitablehabitat remaining for biota displaced by clear- bance withinremnant areas, invasionsmay occur. ing. This may lead to the concentrationof mobile ele- Invasivespecies can have significantimpacts on the mentsof the biota in the remnants(Lovejoy et al. 1986). plantcommunities within remnants; for instance, invad- This concentrationor crowdingeffect may be rapidand ingplant species can significantlyalter the fuel structure resultin supersaturationof the remnantby some spe- and hence fireregime, and can inhibitthe regeneration cies. Crowdingcan alterintra- and interspecificinterac- of native species (Wycherly 1984; Macdonald et al. tions.Competition and predation,for example, can be 1989; Panetta& Hopkins 1991). Nonnativeherbivores, increased,resulting in changesin fecundityand the po- includingstock, can also dramaticallychange vegetation tentialultimate collapse of the population. Resource structureand prevent regeneration.Species that in- availabilityis also affectedby overexploitation;for increase because of landscape modificationcan also have stance,increased herbivory by largeherbivores such as significantimpacts on the restof the biota.For instance, the elephantin Africanreserves can lead to quite dra- in NorthAmerica increased pressurefrom nest preda- maticchanges in habitat(Spinage & Guiness 1971; Laws tors and parasitessuch as the Brown-headedCowbird et al. 1975; Walker 1981). (Molothrus ater) has affectedpasserine bird popula- Supersaturationresults from the influxof species na- tions in fragmentedsystems (Brittingham & Temple tive to the area, but thereare also potentialinfluxes of 1983; Wilcove 1985; Andren& Angelstam1988). Sim- new suitesof species thathave increasedin abundance ilarly,the Galah (Cacatua roseicapilla) has moved into or established.in the surroundinglandscape following all agriculturalareas in Australiaas a resultof the devel- fragmentation.Such species include those that have opmentof cereal croppingand the provisionof water- been introducedin the process of agriculturaldevelop- ingpoints for stock (Saunders et al. 1985). It now roosts ment (pasture and crop plant species and livestock), in remnantwoodland areas,often competing with other other deliberateand accidental introductions,and na- indigenous hole nesters (Saunders & Ingram 1987; tive or migrantspecies thatcan take advantageof the Saunders1990), and can damage tree foliageand bark, new habitatconditions caused by fragmentation. in extremeexamples causing tree mortality. Natural communitiesvary in their susceptibilityto invasion,although there is stilldebate over which char- ModifyingInfluences acteristicsrender one communitymore invasiblethan another(Fox & Fox 1986; MacDonald et al. 1986; Craw- 1. REMNANTSIZE ley 1987; Usher 1988; Rejmanek1989). For vegetation, establishmentof nonnativespecies seems to be en- The smallera remnantis, the greaterthe influenceex- hanced by some formof disturbance,especially if this ternal factors are likely to have. In small remnants, increasesthe availabilityof a limitingresource (Hobbs ecosystemdynamics are probablydriven predominantly 1989; Panetta & Hopkins 1991). Thus the opening of by externalrather than internalforces. Of importance light gaps in dense forestswhere light was limiting here are "edge effects"(Williamson 1975; Janzen1983; could enhanceinvasion, whereas in nutrient-limitedsys- Harris1988; Yahner 1988). Largerremnants have a big- tems,nutrient input significantlyincreases the perfor- ger core area that is unaffectedby the environmental mance ofnonnatives, especially in conjunctionwith soil and biotic changes associated with edges. Here, we re- disturbance(Hobbs & Atkins1988). Invasion may be gardedge effectsas mainlydetrimental; this is opposite restrictedto the edge of remnantsif disturbance factors to the traditionalview thatedges and ecotones are ben- decline with distancefrom edge (Cale & Hobbs 1991; eficialto wildlife(Harris 1988). The differenceis that Panetta & Hopkins 1991), but species with wind- or remnantedges are createdby removalof the surround- animal-dispersedseeds can establishin suitable areas ingvegetation and place the remainderin juxtaposition withina remnant,away fromthe edge. Invadingspecies with a completelyaltered surrounding matrix. can establish,for example, from seeds carriedin by,or Noss and Harris(1986) have pointed out thatwe do deposited in feces of,animals that feed in the area sur- not knowthe minimumcritical size an ecosystemneeds

ConservationBiology Volume 5, No. 1, March 1991 Saunderset al. ConsequencesofFragmentation 25 to be to preserveits characteristicspecies diversityand have gone throughgenetic "bottlenecks"are likelyto species composition(Lovejoy & Oren 1981). In factthe suffera reduction in heterozygositywith consequent "minimumdynamic area" of Pickett and Thompson loss of abilityto adapt to changingconditions. Species (1978) or "the smallestarea witha naturaldisturbance isolatedon remnantsmay go throughsuch geneticbot- regime which maintains internal recolonization tlenecksbecause of smallpopulation sizes, and deliber- sources" would probablyexist only in the largestcon- ate transferof individualsbetween populationsmay be servationparks. required(Boecklen & Bell 1987). However,the general Largerremnants usually contain greater habitat diver- assumptionthat heterozygosity is essentialfor long term sitythan smallerones. A collection of smallerreserves populationviability is stillopen to debate. may,however, cover a greaterarray of habitatsthan a Large reservesmay have some disadvantages;in par- single large one simplybecause a single large reserve ticular,the possibilityof disease spreadingthrough en- will not contain all of the habitatslikely to occur in a tirepopulations on a large reservehas been discussed. region.These argumentshave been discussed in detail However,there is a wide rangeof species' lifehistories elsewhereand will not be pursuedhere. It is important, and distributionpatterns in the biota,and the effectsof however,to recognizethat the process offragmentation reserve size are largely species-specific.Species that is generallynot random(Usher 1987). Land clearance have large area requirementsor thatrequire combina- usually occurs on a selective basis, with the best soil tions of differenthabitats are likelyto surviveonly in types being cleared first.For example, in southwest relativelylarge areas,whereas organismswith small, lo- Western Australia,settlers selectively cleared wood- calized populationsand simplehabitat requirements can lands because theyoccurred on the heavier soils best surviveon smallerremnants. However, all will be af- suited foragriculture. As a result,woodland communi- fected by the disruptionof physical ecosystem pro- ties are now poorly representedin conservationre- cesses that result fromfragmentation and were dis- serves and most woodland remainingon farmsis in a cussed earlier. highlydegraded state (Saunders & Hobbs 1989). Few reservesin the area are large enough to containrepre- 2. SHAPE sentativesamples of all preexistingvegetation types. The shape of a remnantis importantonly for relatively Kitcheneret al. (1980a b, 1982) found,however, that smallareas; there is some size beyondwhich shape does even relativelysmall reserves (i.e., 30 ha) could be rich not reallymatter. However, for small remnants,shape in some groupsof fauna, but whetherthese populations determinesthe perimeter:core(or edge:interior)ratio. are viable in the long termis debatable (see Saunders Long, thin remnantshave proportionallymuch more 1989). edge than square or round remnants(Diamond 1975; Remnantsize determinesthe potentialsize of popu- Wilson& Willis 1975), and are more open to detrimen- lations of componentspecies. Clearly,the numberof tal edge effects.However, some vegetationtypes, such individualsof any particularspecies thata given rem- as riparianstrips, are naturallythin, and corridorsare by nantwill supportdepends on organismsize and require- definitiongenerally long thin remnants(although the ments.The largerthe remnant,the more likelyit is that widerthey are themore useful they may be to aid move- populations will be large and more likely to resist ment of biota; Saunders & Hobbs 199 lb). Long, thin chance extinctions(Gilpin & Soule 1986; Soule 1987a). remnantsmay also, dependingon theirorientation, lie Retainingpopulations in the long term may require along environmentalgradients and thus contain more largepopulation sizes (of the orderof hundredsor very vegetationtypes and habitatsthan a square reserve of much greater;Shaffer 1987), althoughthe actual num- similararea. Linearfeatures are thuspart of the natural bers requiredwill depend on the lifehistory and pop- and fragmentedlandscape, and thereis no point in try- ulationgrowth rate of the species involved.Pimm et al. ingto develop optimaldesign principles that do not take (1988) confirmthat over a fewdecades extinctionrisk theminto account; the most importantquestion is how does decrease with populationsize, but theyfound no to managethe remnants,whatever shape theyare, so as extinctionsamong Britishisland birds numberingover to minimizeexternal effects. 18 pairs.The issue ofminimum population size has been discussed extensively in the literature (see Soule 1987b), but thereis stillno real resolutionas to what 3. POSITION IN LANDSCAPE constitutesa minimumviable population. There has The positionof a remnantin the landscape has impor- been extensivemodeling of the concept,but littleex- tant influenceson all the featuresso far discussed. It perimentalwork. affectsprefragmentation patterns of geomorphology, Largerpopulations tend to have higherlevels of het- soil,and vegetation,and hence determinesthe structure erozygositythan small isolated populations. Current and vegetationcomposition of any givenremnant area. thinkingis thatheterozygosity is beneficial.Species that It also significantlyaffects postfragmentation processes.

ConservationBiology Volume 5, No. 1, March 1991 26 Consequencesof Fragnentation Saunderset al.

For instance,there is an importantdistinction between thisis to attemptto maintainrepresentative examples of remnantsthat are predominantlyrun-on versus those each ecosystemor communitytype present before frag- thatare predominantlyrun-off. This influencesnot only mentation.To do this,we need to know the distribu- the hydrologicalregime of the remnant,but also the tionsof species and communitiesand then select areas movementof soil, nutrients,and seeds into and out of thatrepresent them. In general,there are two possible the remnantarea. scenarios.In the first,we have a systemthat is about to be fragmentedand we have to design the ideal set of Lessons forManagement reservesfor the area. In the second,we have an already- fragmentedsystem that we need to make the most of. Managementof fragmentedecosystems has two basic Most theoriesin conservationbiology, including virtu- components:(1) managementof the naturalsystem, or allyall the discussionsof island biogeography in relation the internaldynamics of remnantareas, and (2) Man- to reserves,have dealtwith the firstscenario, whereas it agementof the externalinfluences on the naturalsys- is the second scenario thatwe mostfrequently have to tem.For largeremnant areas, the emphasis should be on confront.Here we presenta seriesof guidelines for man- managingthe internaldynamics, including, for instance, agementin thissituation. the disturbanceregime and populationdynamics of key 1. The initialstep must be to determinethe minimum organisms.For smallremnants, on the otherhand, man- subsetof the existingremnants that are requiredto rep- agementshould be directedprimarily at controllingthe resent the diversityof a given region (Margules et al. externalinfluences. Janzen (1983, 1986), however,has 1988; Margules& Stein 1989). This requires thatwe pointed out thatexternal influences can be important have some knowledgeof the distributionof species or whateverthe remnantsize. ecosystemtypes. Clearly, it would be desirableto have Since most impactson remnantareas originatefrom all existingremnants available for thispurpose, but in the surroundinglandscape, there is clearly a need to many cases this is not achievable, and there must be departfrom the traditionalnotions of reservemanage- prioritiesfor reserve retention or acquisition. ment, and look instead toward integratedlandscape 2. The systemmust then be managedto maintainthe management.It will become increasinglydifficult to diversityof species or ecosystems. The question of maintainremnants of nativevegetation if the manage- whethermanagement should be for individualspecies mentpractices in the surroundingmatrix have contin- or whole ecosystemsis largelyirrelevant, because indi- uous adverseimpacts on them.Traditional reserve man- vidual species require functioningecosystems to sur- agementstops at the reserveboundary; fluxes of water, vive. Managementguidelines will be area-specific,but particulates,and organismsdo not. Placingthe conser- the need to manage externalinfluences is universal. vation reserves firmlywithin the context of the sur- 3. Prioritiesfor managementmust be established. roundinglandscape and attemptingto develop comple- Clearlythere are manyproblems to be tackled,and usu- mentarymanagement strategies seems to be the only ally there are limitedresources available for the job. way to ensure the long termviability of remnantareas There musttherefore be a clear priorityranking to en- (Hobbs & Saunders1991). This has importantimplica- sure that resources are deployed optimally.Problems tionsfor land managerssince it involvesa radicallynew thatare likelyto disruptecosystem processes and hence way ofviewing management and requiresthat neighbor- threatenthe viabilityof a remnantarea should be given ing land uses, and hence neighboringlandowners, inter- highpriority for treatment. act in a positiveway. This is difficult,but notimpossible, 4. Continuous managementis needed to maintain and thereare encouragingexamples of attemptsat this remnantareas in theircurrent state, due to the constant type of integratedmanagement (e.g., FitzgeraldBio- pressureof altered internaldynamics and externalin- sphere Project 1989; Bradby1991). fluences.Here again, the allocation of scarce manage- The landscapeapproach to managementis also essen- ment resources must be considered. Effortshould go tialsince severalremnant areas takentogether may rep- into maintainingsome remnant areas in as near a resent a systemover which componentsof the biota "natural"state as possible,but it will not be feasibleto travelto meet habitatand food requirements.The loss do this for all remnants.There is a strongcase to be of a singlecomponent of such a networkcould severely made for lettingsome areas degrade so that they be- affectthe capabilityof the remainingremnants to carry come less naturalbut are easier to manage and stillre- out the same functions,if for instance a particularspe- tainsome conservationvalue (Bridgewater1990). This cies or habitatwas lost.Such a networkconsists not only is not as radical as it may sound, since the process is of the designatedreserves, but also otherremnant areas ongoing in manyremnant areas anyway.Once we ac- and linkagesbetween them. cept thatmany remnants now contain"synthetic" com- The goal of conservationmanagement usually is to munitiesthat are not likelyever to returnto theirpris- maintainspecies diversity,and the methodof achieving tine state,management priorities become easier to set.

ConservationBiology Volume 5, No. 1, March 1991 Saunderset al. ConsequencesofFragmentation 27

ResearchRequirements fragmentation.Emphasis also has been on biogeographic explanationsfor the patternsof species loss afterfrag- Research to date on fragmentedecosystems has pro- mentation,whereas a whole suite ofphysical and biotic vided fewanswers on the issues ofpractical importance parametersare significantlyaltered in the fragmented to management.It is just as importantto set priorities systemand have significantimpacts on remnantbiota. In forresearch as formanagement, and in the same way, particular,the switch from predominantlyinternally research costs must be taken into account since re- drivento predominantlyexternally driven dynamics is a sources forresearch are also limited.For instance,is it key factorin the fragmentedsystem. Management and betterto concentrateon single-speciesstudies that can researchshould focus on thisfactor. There is a pressing produce resultswith direct practical application but are need for an integratedapproach that treatsthe land- verycostly and time-consuming,or shouldwe concen- scape as a whole insteadof as a collection of separate trate on the community/ecosystemapproach that is biotic and legal entities. cheaper but mayyield more equivocal results?Clearly, balanced use of both approaches is needed. We have Acknowledgments identifieda number of priorityareas that require research effort(see also Soule & Kohm 1989). We are grateful to Francis Crome, Harry Recher, 1. A majorpriority is to understandthe effectsof ex- Michael Soule, and two anonymousreferees for their ternalfactors. Comparisons of pre- and postfragmenta- constructivecomments on the manuscript. tion systemswill be particularlyuseful (Lovejoy et al. 1986; Margules 1987). Effectsof changes in radiation LiteratureCited and water fluxes are particularlyimportant, as is the biotic invasionprocess. Andersen,A. N. 1987. Ant communityorganisation and envi- 2. Changesin internalprocesses since fragmentation ronmentalassessment. Pages 43-53 in J.D. Majer,editor. The role of invertebratesin conservationand biological survey. also require furtherinvestigation. In particular,the in- Departmentof Conservationand Land Management,Perth, teractionbetween internaland externalprocesses is Australia. likelyto be of criticalimportance. 3. Isolationfactors need to be betterunderstood. In Andren,H., and P. Angelstam.1988. Elevatedpredation rates particular,rates of genetic change in isolated popula- as an edge effectin habitatislands: experimentalevidence. Ecology 69:544-547. tions require study,as does the question of whether reductionin genetic variabilityis important.We also Baudry,J. 1984. Effectsof landscape structureon biological requiremore and betterdata on the role of corridorsin communities:the case ofhedgerow network landscapes. Pages allowingbiotic (and hence genetic) movementin frag- 55-65 in J. Brandtand P. Agger,editors. Proceedings of the mentedlandscapes. firstinternational seminar on methodologyin landscape ecological researchand planning.Volume 1. 4. While theoreticalstudies have theirplace, thereis InternationalAssoci- ation for Landscape Ecology, Rosskilde UniversityCenter, an urgentneed forfield experimentation in both man- Rosskilde,Denmark. agementand restoration.While such experimentsare costly to set up, it is possible to make use of many Bell, A. 1988. Trees,water and salt-a finebalance. Ecos 58: situationsthat offer ready-made experiments. Our un- 2-8. derstandingof succession has benefitedgreatly from the Boecklen,W. J., and C. W. Bell. studyof abandoned old fields,and in the same way,we 1987. Consequences of faunal collapse and genetic driftfor the design of naturereserves. can use currentor past managementactivities as large- Pages 141-149 in D. A. Saunders,G. W. Arnold,A. A. Burbidge, scale experiments;there is plentyof experimentalma- and A.J. M. Hopkins,editors. Nature conservation: the role of terialaround (McNab 1983; Hopkins& Saunders1987; remnantsof nativevegetation. Surrey Beatty and Sons, Chip- Pimm1986;Jordan et al. 1987; Hobbs & Hopkins1990). ping Norton,Australia. Research has much to gain froma close liaison with Boecklen,W.J., and N.J. Gottelli. Island if 1984. biogeographic management,especially managementoperations actu- theoryand conservationpractice: species-area or specious- allycan be carriedout as designedexperiments and the area relationships?Biological Conservation29:63-80. resultssuitably monitored. Bond,W. J., J. Midgley, and J. Vlok. 1988. When is an islandnot an island?Insular effects and theircauses in fynbosshrublands. Oecologia (Berlin) 77:515-521. Conclusions Bormann,F. H., and G. E. Likens.1979. Patternand process in a forestedecosystem. Springer, New York. Emphasisin the literaturehas been on the design of naturereserves, but we are usuallytoo late to do any- Bormann,F. H., G. E. Likens,R. S. Siccama,R. S. Pierce,and J. S. thingexcept tryto maintainthe remnantsleft following Eaton. 1974. The export of nutrientsand recoveryof stable

ConservationBiology Volume 5, No. 1, March 1991 28 ConsequencesofFragnentation Saunderset al. conditionsfollowing deforestation at HubbardBrook. Ecolog- Forman, R. T. T., and J. Baudry. 1984. Hedgerows and ical Monographs44:255-277. hedgerow networks in landscape ecology. Environmental Management8:495-510. Bradby,K 1991. A data bank is never enough-the local approach to landcare.Pages 377-385 in D. A. Saundersand R.J. Fox, M. D., and B.J. Fox. 1986. The susceptibilityof natural Hobbs,editors. Nature conservation: the role ofcorridors. Sur- communitiesto invasion.Pages 57-66 in R. H. Grovesand J.J. rey Beattyand Sons, ChippingNorton, Australia. In press. Burdon,editors. Ecology of biologicalinvasions: an Australian perspective.Australian Academy of Science, Canberra,Austra- Brett,D. 1989. Sea birdsin the trees.Ecos 61:4-8. lia.

Bridgewater,P. B. 1987. Connectivity:an Australianperspec- Fraser,A. I. 1972. The effectof climatefactors on the devel- tive.Pages 195-200 in D. A. Saunders,G. W. Arnold,A. A. Bur- opmentof plantation forest structure. Review Papers in Forest bidge, and A.J. M. Hopkins,editors. Nature conservation: the Meteorology1972:109-125. role of remnantsof nativevegetation. Surrey Beatty and Sons, ChippingNorton, Australia. Game, M. 1980. Best shape for nature reserves. Nature 287:630-632. Bridgewater,P. B. 1990. The role of syntheticvegetation in presentand futurelandscapes of Australia.Proceedings, Eco- Geiger,R. 1965. The climate near the ground.Harvard Uni- logical Societyof Australia16:129-134. versityPress, Cambridge, Massachusetts.

Brittingham,M. C., and S.A. Temple. 1983. Have cowbirds Gilpin,M. E., and J.M. Diamond. 1980. Subdivisionof nature caused forestsongbirds to decline?BioScience 33:31-35. reserves and the maintenanceof species diversity.Nature 285:567-568. Caborn,J. M. 1957. Shelterbeltsand microclimate.Forestry CommissionBulletin 29. Her Majesty'sStationary Office, Ed- Gilpin,M. E., and M. E. Soul& 1986. Minimumviable popula- inburgh,Scotland. tions:processes of species extinctions.Pages 19-34 in M. E. Soule, editor. Conservationbiology. The science of scarcity Cale, P., and R.J.Hobbs. 1991. Conditionof roadside vegeta- and diversity.Sinauer Associates, Sunderland, Massachusetts. tionin relationto nutrientstatus. Pages 353-362 in D. A. Saun- ders and R.J.Hobbs, editors.Nature conservation the role of Gordon,D. R.,J. M. Welker,J. W. Menke,and KJ. Rice. 1989. corridors.Surrey Beatty and Sons,Chipping Norton, Australia. Competitionfor soil waterbetween annual plants and blue oak (Quercus douglasii) seedlings. Oecologia (Berlin) 79:533- Crawley,M.J. 1987. Whatmakes a communityinvasible? Pages 541. 429-453 in M.J.Crawley, P.J. Edwards, and A.J.Gray, editors. Plant Colonization, succession and stability.Blackwell Scientific Grace,J. 1977. response to wind. Academic Press,Lon- Publications,Oxford, England. don, England. Grace,J. 1983. Plant-atmosphererelationships. Chapman and De Walle,D. R.,and S. G. McGuire.1973. Albedo variationsof Hall, London,England. an oak forest in Pennsylvania.Agricultural Meteorology 11:107-113. Greenslade,P.J. M., and R. B. Halliday. 1983. Colony disper- sion and relationshipsof meat antsIridomyrmex purpureus Diamond,J. M. 1975. The island dilemma:lessons of modern and allies in an aridlocality in SouthAustralia. Insectes Sociaux biogeographicstudies for the design of naturereserves. Bio- 30:82-99. logical Conservation7:129-146. Gysel, L. W. 1951. Borders and openings of beech-maple Diamond,J. M. 1976. Island biogeographyand conservation: woodlands in southernMichigan. Journal of Forestry49:13- Science strategyand limitations. 193:1027-1029. 19.

Ehrlich,A. H., and P. R. Ehrlich.1987. Earth.Franklin Watts, Haila, Y., and I. Hanski. 1984. Methodologyfor studyingthe New York. effectof fragmentationon land birds.Annals Zoologica Fen- nica 21:393-397. Ewens,W. J., P.J. Brockwell, J. M. Gani,and S. I. Resnick.1987. Minimumviable populationsize in the presence of catastro- Harris,L. D. 1984. The fragmentedforest. Island biogeographic phes. Pages 59-68 in M. E. Soule,editor. Viable populations for theoryand the preservationof biotic diversity.University of conservation.Cambridge UniversityPress, Cambridge,En- Chicago Press,Chicago, Illinois. gland. Harris,L. D. 1985. Conservationcorridors: a highwaysystem Fahrig,L., and G. Merriam.1985. Habitatpatch connectivity for wildlife.ENFO Report 85-5. EnvironmentalInformation and populationsurvival. Ecology 66:1762-1768. Centerof the FloridaConservation Foundation, Winter Park, Florida. FitzgeraldBiosphere Project. 1989. The bush comes to the city.Fitzgerald Biosphere Project, Perth, Australia. Harris,L. D. 1988. Edge effectsand conservationof biotic diversity.Conservation Biology 2:330-332. Forman,R. T. T. 1987. The ethics of isolation,the spread of disturbance,and landscape ecology. Pages 213-229 in M. G. Harris,L. D., and J. Scheck. 1991. From implicationsto appli- Turner, editor. Landscape heterogeneityand disturbance. cations:the dispersalcorridor principle applied to conserva- Springer,New York. tion of biological diversity.Pages 189-220 in D. A. Saunders

ConservationBiology Volume 5, No. 1, March 1991 Saunderset al. ConsequencesofFragmentation 29 and R.J. Hobbs, editors.Nature conservation the role of cor- Janzen,D. H. 1983. No park is an island:increase in interfer- ridors.Surrey Beatty and Sons, ChippingNorton, Australia. ence fromoutside as park size decreases. Oikos 41:402-410.

Higgs,A.J. 1981. Island biogeographictheory and naturere- Janzen,D. H. 1986. The eternalexternal threat. Pages 286-303 serve design.Journal of Biogeography8:117-124. in M. E. Soule, editor. Conservationbiology. The science of scarcityand diversity.Sinauer Associates, Sunderland, Massa- Higgs,A.J., and M. B. Usher. 1980. Should naturereserves be chusetts. large or small?Nature 285:568. Jarvinen,0. 1982. Conservationof endangeredplant popula- Hobbs, R.J.1987. Disturbanceregimes in remnantsof natural tions:single large or several small reserves?Oikos 38:301-7. vegetation.Pages 233-240 in D. A. Saunders,G. W. Arnold, A.A. Burbidge,and A.J. M. Hopkins,editors. Nature conserva- Jarvis,P. G., and K G. McNaughton.1986. Stomatalcontrol of tion:the role of remnantsof nativevegetation. Surrey Beatty transpiration:scaling up fromleaf to region.Advances in Eco- and Sons, ChippingNorton, Australia. logical Research15:1-4.

Hobbs,R. J. 1988. Whatis ecological theoryand is it ofany use Jones,H. G. 1983. Plants and microclimate.Cambridge Uni- to managers?Pages 15-27 in D. A. Saunders and A.A. Bur- versityPress, Cambridge, England. bidge, editors.Ecological theoryand biological management of ecosystems.CALM Occasional Paper 1/88.Department of Jordan,W. R.,III, M. E. Gilpin,andJ. D. Aber. 1987. Restoration Conservationand Land Management,Perth, Australia. ecology: ecological restorationas a technique for basic research. Pages 3-21 in W. R. JordanIII, M. E. Gilpin,and J.D. Hobbs, R.J. 1989. The natureand effectsof disturbancerela- Aber, editors. Restorationecology: a syntheticapproach to tive to invasions.Pages 389-405 in J.A. Drake,H. A. Mooney, ecological research.Cambridge University Press, Cambridge, F. di Castri,et al., editors.Biological invasions.A global per- England. spective.John Wiley & Sons,New York. Kapos,V. 1989. Effectsof isolation on thewater status of forest Hobbs,R. J., and L. Atkins.1988. The effectof disturbanceand patches in the BrazilianAmazon. Journal of Tropical Ecology nutrientaddition on native and introducedannuals in the 5:173-185. WesternAustralian wheatbelt. Australian Journal of Ecology 13:171-179. Karieva,P. 1987. Habitat fragmentationand the stabilityof predator-preyinteractions. Nature 326:388-90. Hobbs, R.J.,and A.J.M. Hopkins.1990. Fromfrontier to fragments:European impact on Australia'svegetation. Proceed- Karr,J. R., and I.J. Schlosser. 1978. Water resources and the ings,Ecological Societyof Australia16:93-114. land-waterinterface. Science 201:229-34.

Hobbs, R.J., and D. A. Saunders. 1991. Reintegratingfrag- Kendrick,G. W. 1976. The Avon:faunal and othernotes on a mented landscapes: a proposed frameworkfor the Western dyingriver in south-westernAustralia. West AustralianNatu- Australianwheatbelt. Proc. V Aust.Soil Cons. Conf.In press. ralist13:97-114. Kitchener,D. J.,A. Chapman,J. Dell, B. G. Muir,and M. Palmer. Hobbs, R.J.,D. A. Saunders,and B. M. T. Hussey. 1990. Nature 1980a. Lizardspecies assemblageand reservesize and struc- conservation:the role of corridors.Ambio 19:94-95. turein the WesternAustralian wheatbelt-some implications forconservation. Biological Conservation17:25-62. Hopkins,A.J. M., and D. A. Saunders.1987. Ecological studies as the basis formanagement. Pages in D. A. 15-28 Saunders, Kitchener,D. J.,A. Chapman,B. G. Muir,and M. Palmer.1980b. G. W. A. A. and A. M. editors. Arnold, Burbidge, J. Hopkins, Na- The conservationvalue formammals of reservesin the West- ture conservation:the role of remnantsof nativevegetation. ern Australianwheatbelt. Biological Conservation 18:179- SurreyBeatty and Sons, ChippingNorton, Australia. 207.

Hornbeck,J.W. 1973. Stormflow from hardwood-forested and Kitchener,D.J., J. Dell, B. G. Muir,and M. Palmer.1982. Birds cleared watershedsin New Hampshire.Water ResourcesRe- in WesternAustralian wheatbelt reserves-implicationsfor search 9:346-354. conservation.Biological Conservation22:127-163.

Hudson,N. 1981. Soil conservation.Cornell University Press, Klein,B. C. 1989. Effectsof forestfragmentation on dung and Ithaca,New York. carrion beetle communitiesin Central Amazonia. Ecology 70:1715-1725. Hutchinson,B. A., and D. R. Matt.1976. Beam enrichmentof diffuseradiation in a deciduous forest.Agricultural Meteorol- Laws, R. M., I. S. C. Parker,and R. C. B. Johnstone.1975. Ele- ogy 17:93-1 10. phantsand theirhabitat. Clarendon Press, Oxford, England.

Hutchinson,B. A., and D. R. Matt. 1977. The distributionof Le Houerou,H. N., and H. Gillet.1986. Desertizationin African solar radiationwithin a deciduous forest.Ecological Mono- arid lands. Pages 44-461 in M. E. Soule, editor.Conservation graphs47:185-207. biology.The science of scarcityand diversity.Sinauer Associ- ates,Sunderland, Massachusetts. Hutchinson,J. N. 1980. The recordof peat wastagein the East Anglianfenlands at Holme Post, 1848-1978 AD. Journalof Likens,G. E., F. H. Bormann,N. M. Johnson,D. W. Fisher,and Ecology 68:229 -49. R. S. Pierce. 1970. Effectsof forest cutting and herbicidetreat-

ConservationBiology Volume 5, No. 1, March 1991 30 ConsequencesofFragnentation Saunderset al. ment on nutrientbudgets in the Hubbard Brook watershed- Margules,C. R.,A.J. Higgs, and R. W. Rafe. 1982. Modem bio- ecosystem.Ecological Monographs40:23-47. geographictheory: are there any lessons for nature reserve design?Biological Conservation24:115-128. Lovejoy,T. E., R. 0. Bierregaard,K S. Brown,L. H. Emmons, and M. E. Van der Voort. 1984. Ecosystemdecay of Amazon Margules,C. R.,A. 0. Nicholls,and R. L. Pressey.1988. Select- forestfragments. Pages 295-325 in M. H. Niteki,editor. Ex- ing networksof reservesto maximisebiological diversity.Bi- tinctions.University of Chicago Press,Chicago, Illinois. ological Conservation43:63-76.

Lovejoy,T. E., B. 0. Bierregaard,Jr., A. B. Rylands,et al. 1986. Margules,C. R.,andJ. L. Stein.1989. Patternsand distributions Edge and other effectsof isolation on Amazon forestfrag- of species and the selection of nature reserves:an example ments.Pages 257-285 in M. E. Soule, editor.Conservation bi- fromEucalyptus forestsin south-easternNew South Wales. ology. The science of scarcityand diversity.Sinauer Associ- BiologicalConservation 50:219-238. ates,Sunderland, Massachusetts. May,R. M. 1975. Island biogeographyand the design of wild- lifepreserves. Nature 245:177-178. Lovejoy,T. E.,and D. C. Oren. 1981. The minimumcritical size of ecosystems.Pages 7-12 in R. L. Burgessand D. M. Sharpe, Merriam,G. 1984. Connectivity:a fundamentalecological editors.Forest island dynamics in man-dominatedlandscapes. characteristicof landscape pattern. Pages 5-15 inJ. Brandt and Springer,New York. P. Agger,editors. Proceedings of the FirstInternational Semi- nar on Methodologyin Landscape Ecological Research and MacArthur,R. H., and E. 0. Wilson. 1963. An equilibriumthe- Planning.International Association for Landscape Ecology, ory of insularbiogeography. Evolution 17:373-387. RosskildeUniversity Center, Rosskilde, Denmark.

MacArthur,R. H., and E. 0. Wilson. 1967. The theoryof island Merriam,G. 1991. Corridorsfor connectivity: animal popula- biogeography.Princeton University Press, Princeton, NewJer- tions in heterogeneousenvironments. Pages 133-142 D. A. sey. Saunders and R.J. Hobbs, editors. Nature conservation:the role of corridors.Surrey Beatty and Sons, ChippingNorton, MacClintock,L., R. F. Whitcomb,and B. L. Whitcomb.1977. Australia. Islandbiogeography and the "habitatislands" of easternforest II. Evidence for the value of corridorsand minimizationof Miller,R. I. 1978. Applyingisland biogeographictheory to an isolation in preservationof biotic diversity.American Birds East Africanreserve. Environmental Conservation 5:191-195. 31:6-16. Miller,R. I., and L. D. Harris.1977. Isolationand extirpationsin wildlife Macdonald,I. A. W., L. L. Loope, M. B. Usher,and 0. Hamann. reserves.Biological Conservation12:311-315. 1989. Wildlifeconservation and the invasion of nature re- Milthorpe,F. L.,andJ. Moorby.1974. An introductionto crop servesby introducedspecies: a globalperspective. Pages 215- physiology.Cambridge University Press, Cambridge, England. 255 in J.A. Drake, H. A. Mooney,F. di Castri,et al., editors. Biologicalinvasions. A global perspective.John Wiley & Sons, Moen, A. N. 1974. Turbulence and the visualizationof wind New York. flow.Ecology 55:1420-1424.

Macdonald,I. A. W.,F. J. Powrie, and W. R. Siegfried.1986. The Monteith,J. L. 1975. Vegetationand the atmosphere.Volume differentialinvasion of southernAfrica's biomes and ecosys- 1. Principles.Academic Press,London, England. tems by alien plants and animals.Pages 209-225 in I. A. W. Macdonald,F. J. Kruger, and A. A. Ferrar,editors. The ecology Muir,B. G. 1979. Observationsof wind-blown superphosphate and managementof biological invasionsin SouthernAfrica. in nativevegetation. West AustralianNaturalist 14:128-130. OxfordUniversity Press, Cape Town, SouthAfrica. Murphy,D. D. 1989. Conservationand confusion:wrong spe- McNab,J. 1983. Wildlifemanagement as scientificexperimen- cies, wrong scale, wrong conclusions. ConservationBiology tation.The WildlifeSociety Bulletin 11:397-401. 3:82-84.

Mader,H.-J. 1984. Animalhabitat isolation by roads and agri- Myers,N. 1988. Tropicalforests and theirspecies: going,going culturalfields. Biological Conservation29:81-96. ... ? Pages 28-35 in E. 0. Wilson and F. M. Peter,editors. Bio- diversity.National Academy Press, Washington, D.C. Main,B. Y. 1987. Persistenceof invertebratesin small areas: case studies of trapdoorspiders in WesternAustralia. Pages Noss, R. F. 1987. Corridorsin real landscapes: a replyto Sim- 29-39 in D.A. Saunders,G. W. Arnold,A.A. Burbidge,and berloffand Cox. ConservationBiology 1:159-164. A.J. M. Hopkins,editors. Nature conservation: the role ofrem- Noss,R. F.,and L. D. Harris.1986. Nodes, networksand MUMs: nantsof nativevegetation. Surrey Beatty and Sons, Chipping preservingdiversity at all scales. Norton,Australia. EnvironmentalManagement 10:299-309.

Margules,C. R. 1986. Conservationevaluation in practice. Nulsen,R. A.,K J.Bligh, I. N. Baxter,E. J.Solin, and D. H. Imrie. Pages 297-314 in M. B. Usher,editor. Wildlife conservation 1986. The fate of rainfallin a mallee and heath vegetated evaluation.Chapman and Hall, London,England. catchmentin southernWestern Australia. Australian Journal of Ecology 11:361-371. Margules,C. R. 1987. The Wog Wog habitatpatch experiment: Background,objectives, experimental design and samplestrat- Palik,B.J., and P. G. Murphy.1990. Disturbanceversus edge egy.Technical Memorandum 85/18. CSIRO, Divisionof Water effectsin sugar-maple/beechforest fragments. Forest Ecology and Land Resources,Canberra, Australia. and Management32:187 -202.

ConservationBiology Volume 5, No. 1, March 1991 Saunderset al. ConsequencesofFragmentation 31

Panetta,D., and A.J.M. Hopkins. 1991. Weeds in corridors: Saunders,G. W. Arnold,A. A. Burbidge,and A.J. M. Hopkins, invasionand management.Pages 341-351 in D.A. Saunders editors.Nature conservation:the role of remnantsof native and R.J. Hobbs, editors.Nature conservation the role of cor- vegetation.Surrey Beatty and Sons, ChippingNorton, Austra- ridors.Surrey Beatty and Sons, ChippingNorton, Australia. lia.

Parker,C. A. 1989. Soil biota and plantsin the rehabilitationof Saunders,D. A.,G. W. Arnold,A. A. Burbidge,and A.J.M. Hop- degradedagricultural soils. Pages 423-438 in J.D. Majer,ed- kins,editors. 1987b. Natureconservation: the role ofremnants itor.Animals in primarysuccession. The role of faunain re- ofnative vegetation. Surrey Beatty and Sons,Chipping Norton, claimed lands. CambridgeUniversity Press, Cambridge,En- Australia. gland. Saunders,D., and R. Hobbs. 1989. Corridorsfor conservation. Peck,A.J. 1978. Salinisationof non-irrigatedsoils and associ- New Scientist1642:63-68. ated streams:a review. AustralianJournal of Soil Research 16:157-168. Saunders,D. A., and R.J. Hobbs, editors. 199 la. Nature conservation:the role of corridors.Surrey Beatty and Sons,Chip- Peck,A.J., and D. R. Williamson.1987. Effectsof forestclear- ping Norton,Australia. In press. ing on groundwater.Journal of Hydrology94:47-66. Saunders,D. A., and R.J.Hobbs. 1991b. The role of corridors Pickett,S. T. A.,andJ. N. Thompson.1978. Patchdynamics and in natureconservation: what do we know and where do we the size ofnature reserves. Biological Conservation 13:27-37. go? Pages 421-427 in D. A. Saundersand R.J.Hobbs, editors. Natureconservation: the role of corridors.Surrey Beatty and Pimm,S. A. 1986. Communitystability and structure.Pages Sons,Chipping Norton, Australia. 309-329 in M. E. Soule, editor.Conservation biology. The science of scarcityand diversity.Sinauer Associates, Sunderland, Saunders,D. A.,A. J. M. Hopkins,and R. A. How, editors.1990. Massachusetts. Australianecosystems: 200 years of utilisation,degradation and reconstruction.Proceedings of the Ecological Societyof Pimm,S. A., H. L. Jones,and J. Diamond. 1988: On the riskof Australia16. extinction.American Naturalist 132:757-785. Saunders,D. A., and J.A. Ingram.1987. Factorsaffecting sur- Quinn,J. F., and A. Hastings.1987. Extinctionin subdivided vival of breedingpopulations of Carnaby'scockatoo Calypto- habitats.Conservation Biology 1:198-208. rhynchusfunereus latirostrisin remnantsof native vegetation. Pages 249-258 in D. A. Saunders,G. W. Arnold,A. A. Ranney,J. W., M. C. Bruner,and J.B. Levenson.1981. The im- Burbidge,and A.J. M. Hopkins,editors. Nature conservation: portance of edge in the structureand dynamicsof forestis- the role of remnantsof native vegetation.Surrey Beatty and lands. Pages 67-95 in R. L. Burgessand D. M. Sharpe,editors. Sons, ChippingNorton, Australia. Forest island dynamics in man-dominated landscapes. Springer,New York. Saunders,D. A.,and C. P. de Rebeira.1991. Values ofcorridors to avian populationsin a fragmentedlandscape. Pages 221- Rejm'anek,M. 1989. Invasibilityof plant communities.Pages 244 in D. A. Saundersand R.J. Hobbs, editors.Nature conser- 369-388 inJ.A. Drake,H. A. Mooney,F. di Castri,et al. editors. vationthe Role ofCorridors. Surrey Beatty and Sons,Chipping Biologicalinvasions. A global perspective.John Wiley & Sons, Norton,Australia. New York. Saunders,D. A.,I. Rowley,and G. T. Smith.1985. The effectsof Reville,B. J., J. D. Tranter,and H. D. Yorkston.1990. Impactof clearingfor agriculture on the distributionof cockatoos in the forestclearing on the endangeredseabird, Sula abbotti. Bio- south west of WesternAustralia. Pages 309-321 in A. Keast, logical Conservation51:23-38. H. F. Recher,H. Ford,and D. Saunders,editors. Birds of euca- lyptforests and woodlands: ecology, conservation,manage- Rowell, T. A. 1986. The historyof drainage at Wicken Fen, ment.Surrey Beatty and Sons, ChippingNorton, Australia. Cambridgeshire,England, and its relevance to conservation. Biological Conservation35:111-142. Shaffer,M. 1987. Minimumviable populations:coping with uncertainty.Pages 69-86 in M. E. Soule, editor.Viable popu- Saunders,D. A. 1982. The breedingbehaviour and biologyof lations for conservation.Cambridge UniversityPress, Cam- the short-billedform of the White-tailedBlack Cockatoo Ca- bridge,England. lyptorhynchusfunereus. Ibis 124:422-455. Sharma,M. L., R.J. W. Barron,and D. R. Williamson.1987. Soil Saunders,D. A. 1989. Changesin the avifaunaof a region,dis- water dynamicsof lateriticcatchments as affectedby forest trictand remnantas a resultof fragmentationof nativevege- clearingfor pasture. Journal of Hydrology94:109-127. tation:the wheatbeltof WesternAustralia. A case study.Bio- logical Conservation50:99-135. Simberloff,D. 1986. Design of naturereserves. Pages 315-37 in M. B. Usher,editor. Wildlife conservation evaluation. Chap- Saunders,D. A. 1990. Problemsof survivalin an extensively man and Hall, London,England. cultivatedlandscape: the case of Carnaby'sCockatoo Calyp- torhynchusfunereus latirostris. Biological Conservation Simberloff,D. 1988. The contributionof population and com- 54:111-124. munitybiology to conservationscience. Annual Review of Ecology and Systematics19:473-511. Saunders,D. A.,G. W. Arnold,A. A. Burbidge,and A.J. M. Hop- kins. 1987a.The role of remnantsof nativevegetation in na- Simberloff,D. S., and L. G. Abele. 1976a. Island biogeography ture conservation:future directions. Pages 387-392 in D. A. and conservation:strategy and limitations.Science 193:1032.

ConservationBiology Volume 5, No. 1, March 1991 32 Consequencesof Fragnentation Saunderset al.

Simberloff,D. S., and L. G. Abele. 1976b. Island biogeography 2000: modelinghabitat relationships of terrestrialvertebrates. theoryand conservationpractice. Science 191:285-286. Universityof WisconsinPress, Madison, Wisconsin.

Simberloff,D., and L. G. Abele. 1982. Refugedesign and island Wales,B. A. 1972. Vegetationanalysis of north and southedges biogeographictheory: effects of fragmentation. American Nat- in a mature oak-hickoryforest. Ecological Monographs uralist120:41-50. 42:451-471.

Simberloff,D., and L. G. Abele. 1984. Conservationand obfus- Walker,B. H. 1981. Is succession a viable concept in African cations:subdivision of reserves.Oikos 42:399-401. savannah ecosystems?Pages 431-447 in D. C. West, H. H. Shugart,and D. B. Botkin,editors. Forest succession: concepts Simberloff,D., and J. Cox. 1987. Consequences and costs of and application.Springer, New York. conservationcorridors. Conservation Biology 1:63-71. Webb, N. R. 1989. Studies on the invertebratefauna of frag- Simons,P. 1989. Nobody loves a canal with no water. New mented heathiandin Dorset, U.K, and the implicationsfor Scientist1685:30-34. conservation.Biological Conservation47:153-165. Wegner,J. F., and G. Merriam.1979. Movementsby birdsand Soule, M. E. 1987a. Introduction.Pages 1-10 in M. E. Soule, smallmammals between a wood and adjoining editor.Viable populationsfor conservation.Cambridge Uni- farmlandhab- itats.Journal of AppliedEcology 16:349-357. versityPress, Cambridge, England. Whitcomb,R. F., J.F. Lynch,P. A. Opler, and C. S. Chandler. conserva- Soule, M. E., editor. 1987b. Viable populationsfor 1976. Islandbiogeography and conservation:strategy and lim- tion. CambridgeUniversity Press, Cambridge, England. itation.Science 193:1030-1032.

Soule,M. E.,D. T. Bolger,A. C. Alberts,J. Wright, M. Sorice,and Wilcove, D. S. 1985. Nest predationin foresttracts and the S. Hill. 1988. Reconstructeddynamics of rapid extinctionsof decline of migratorysongbirds. Ecology 66:1211-1214. chaparral-requiringbirds in urban habitatislands. Conserva- tion Biology2:75-92. Wilcove,D. S., C. H. McLellan,and A. P. Dobson. 1986. Habitat fragmentationin the temperatezone. Pages 273-256 in M. E. Soule,M. E., and K A. Kohm.1989. Researchpriorities for con- Soule, editor. Conservationbiology. The science of scarcity servationbiology. Island Press,Washington, D.C. and diversity.Sinauer Associates, Sunderland, Massachusetts.

Spinage,C. A., and F. E. Guiness. 1971. Tree survivalin the Wilcox, B. A. 1980. Insular ecology and conservation.Pages absence of elephantsin the AkageraNational Park, Rwanda. 95-117 in M. E. Soule and B. A. Wilcox, editors.Conservation Journalof AppliedEcology 8:723-728. Biologyan Evolutionary-EcologicalPerspective. Sinauer, Sun- derland,Massachusetts. Swanson,F. J., T. K Kratz,N. Caine, and R. G. Woodmansee. 1988. Landformeffects on ecosystempatterns and processes. Williams,K, and R.J.Hobbs. 1989. Controlof shrubestablish- BioScience 38:92-98. mentby springtimesoil water availabilityin an annual grass- land. Oecologia (Berlin) 81:62--66. Terborgh,J. 1976. Island biogeographyand conservation: strategyand limitations.Science 193:1029-1030. Williamson,D. R., R. A. Stokes,and J.K Ruprecht.1987. Re- sponse of inputand outputof water and chlorideto clearing Tranquillini,W. 1979. Physiologicalecology of the alpine tim- foragriculture. Journal of Hydrology94:1-28. berline;tree existence at highaltitudes with special reference to the EuropeanAlps. Springer,New York. Williamson,M. 1975. The design of naturepreserves. Nature (London) 256:519. Trimble,G. R., and E. H. Tryon. 1966. Crown encroachment into openingscut in Appalachianhardwood stands. Journal of Willis,E. 0. 1984. Conservation,subdivision of reservesand Forestry64:104-108. the anti-dismembermenthypothesis. Oikos 42:396-398. Wilson,E. O., and E. 0. Willis. 1975. Applied biogeography. Usher,M. B. 1987. Effectsof fragmentationon communities Pages 522-34 in M. L. Cody and J.M. Diamond,editors. Ecol- and populations:a reviewwith applications to wildlifeconser- ogy and evolutionof communities.Belknap Press, Cambridge, vation.Pages 103-121 in D.A. Saunders,G. W. Arnold,A.A. Massachusetts. Burbidge,and A.J. M. Hopkins,editors. Nature conservation: the role of remnantsof nativevegetation. Surrey Beatty and Wycherly,P. 1984. People, fire and weeds: can the vicious Sons,Chipping Norton, Australia. spiralbe broken?Pages 11-17 in S. A. Moore,editor. The managementof smallbush areas in the Perthmetropolitan region. Usher,M. B. 1988. Biological invasionsof naturereserves: a Departmentof Fisheriesand Wildlife,Perth, Australia. searchfor generalisations. Biological Conservation 44:119-135. Yahner, R. H. 1988. Changes in wildlife communitiesnear Vepsalailnen,K, and B. Pisarski.1982. Assemblyof island ant edges. ConservationBiology 2:333 - 339. communities.Annals Zoologica Fennici19:327-335. Zimmerman,B. L., and R. 0. Bierragaard.1986. Relevance of Verner,J. 1986. Predictingeffects of habitatpatchiness and the equilibriumtheory of island biogeographyand species fragmentation-theresearcher's viewpoint. Pages 327-329 in area relationsto conservationwith a case fromAnazonia. J. Verner,M.L. Morrison,and C.J. Ralph, editors. Wildlife Journalof Biogeography13:133 -143.

ConservationBiology Volume 5, No. 1, March 1991