Elton Revisited: a Review of Evidence Linking Diversity and Invasibility Author(S): Jonathan M

Nordic Society Oikos

Elton Revisited: A Review of Evidence Linking Diversity and Invasibility Author(s): Jonathan M. Levine and Carla M. D'Antonio Source: Oikos, Vol. 87, No. 1 (Oct., 1999), pp. 15-26 Published by: Blackwell Publishing on behalf of Nordic Society Oikos Stable URL: http://www.jstor.org/stable/3546992 Accessed: 03/12/2010 14:50

Your use of the JSTOR archive indicates your acceptance of JSTOR's Terms and Conditions of Use, available at http://www.jstor.org/page/info/about/policies/terms.jsp. JSTOR's Terms and Conditions of Use provides, in part, that unless you have obtained prior permission, you may not download an entire issue of a journal or multiple copies of articles, and you may use content in the JSTOR archive only for your personal, non-commercial use.

Please contact the publisher regarding any further use of this work. Publisher contact information may be obtained at http://www.jstor.org/action/showPublisher?publisherCode=black.

Each copy of any part of a JSTOR transmission must contain the same copyright notice that appears on the screen or printed page of such transmission.

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].

Blackwell Publishing and Nordic Society Oikos are collaborating with JSTOR to digitize, preserve and extend access to Oikos.

http://www.jstor.org OIKOS 87: 15-26. Copenhagen1999

Minireviews provides an opportunity to summarize existing knowledge of selected M IN I- ecologicalareas, with specialemphasis on currenttopics whererapid and significant advances are occurring. Reviews should be concise and not too wide-ranging. All key REVIENW referencesshould be cited. A summaryis required.

Elton revisited:a review of evidence linking diversity and invasibility

Jonathan M. Levine and Carla M. D'Antonio

Levine, J. M. and D'Antonio, C. M. 1999. Elton revisited:a review of evidence linkingdiversity and invasibility.- Oikos 87: 15-26.

It is commonlybelieved that diversecommunities better resist invasion by exotic species than do simple communities.We examinedthe history of this notion, and evaluatedtheoretical and empiricalwork linkingdiversity and invasions.We found that much of the historicalwork that has contributedto the perceptionthat diverse communitiesare less invasible, including Elton's observationsand MacArthur's species-packingand diversity-stabilitymodels, is based on controversialpremises. Nevertheless,more recent theoreticalstudies consistentlysupported the predicted negative relationshipbetween diversityand invasibility.The results of empirical studies, however, were decidedlymixed. Constructedcommunity studies directly manipulatingdiversity found both positive and negative effects of diversity on invasibilityin both field and microcosmsettings. Other empirical studies tracking the assemblyof ecologicalcommunities generally suggested that communitiesdecline in invasibilityas speciesaccumulate over time, though the role of diversityitself was often ambiguous.Studies of the spatial correlationbetween diversity and invasion and studiesexperimentally adding invaders to naturalsystems indicated that diverse communitiestend to be more invasible.We arguethat theseresults most likelyreflect environmentalfactors spatially covarying with diversityin naturalcommunities (e.g. resources,disturbance), and not the effects of diversity itself as uncoveredby constructedcommunity studies. Nevertheless,the consistent positive relationship between exotic species abundanceand resident species diversityfound in spatial patternstudies suggests that invadersand residentspecies are more similarthan often believed,and the implicationsof this for theoriesof invasionare discussed.

J. M. Levine and C. M. D'Antonio, Dept of Integrative Biology, Univ. of California, Berkeley, CA 94720, USA ([email protected]).

Concern over the impacts of biodiversity loss on the is "resistance" - the ability of a community to resist functioning of ecosystems has stimulated renewed inter- change in the face of a potentially perturbing force est in the importance of species diversity to community (Connell and Sousa 1984, Pimm 1991). These forces are stability. Stability has several components; among these often abiotic, such as storms, fires, and droughts,

OIKOS 87:1 (1999) though increasingly,the major perturbationsexperi- invasion, and has fueled recent challengesto Elton's enced by natural communitiesare biological in origin observationthat species poor, oceanic islands are par- (Drake et al. 1989). Extremelycommon is the human- ticularlyvulnerable to invasion.Simberloff (1986, 1995) mediated introductionof plant, animal, or pathogen and D'Antonio and Dudley (1995)have concludedthat speciesinto communitieswhere they did not previously while island species are often more susceptibleto the exist. The ease with which these newcomersbecome impactsof invadersthan theircontinental counterparts, establishedmembers of communitiesdefines the "inva- thereare insufficientdata to evaluatethe inherentinva- sibility"of a system.Neither the generalcontrols over sibility of intact island communities. communityinvasibility nor the role of speciesdiversity Classic ecological theory has also strongly con- in influencingsusceptibility to invasion are well under- tributedto the belief that more diversesystems are less stood. Yet, the widespreadoccurrence of invasiveintro- invasible.Much of the researchconducted in the 1960's duced species, their documented impact on native and 1970'swas based on assumptionsthat niche space biological diversity (e.g. Mooney and Drake 1986, in naturalcommunities is limiting and that communi- Drake et al. 1989, Vitouseket al. 1997) and the devel- ties are strongly structuredby competition(May and oping interestin restorationof nativediversity mandate MacArthur1972). Several authors, but most notably a clear understandingof the community-levelprocesses MacArthur(1970, 1972)explored the conditionsunder influencinginvasions. Here, we review empiricaland which an assemblagearrayed along a resource axis theoreticalresearch linking diversity and invasibilityin could be invaded by a new species. In general, this an attempt to provide a critical frameworkfor ap- work suggestedthat the more species occupyinga re- proachingresearch in this area. sourceaxis, the more fully resourceswere monopolized, and the more difficultit was to insert new species(but see Case 1991). These models, however, have been on several the that Historical considerations challenged grounds,including point competitorsalong a niche axis may indirectlyfacilitate It is widelybelieved that high native diversitydecreases one another via competing with shared competitors. the invasibilityof communities(Lodge 1993), an idea When such indirectinteractions are included,the same with perhapsthree origins.The first lies in the natural types of models suggestthat for some combinationsof history observationsand writingsof Elton (1958) and species, diverse communitiesmight be more invasible the second and third emerge from classic ecological (Levine 1976, Lawlor 1979, Stone and Roberts 1991, theory exemplifiedby the work of MacArthur(1955, Levine 1999). In spite of the now widespreadaccep- 1972) and colleagues. tance that indirectinteractions are common and more In his book, The ecologyof invasionsby animalsand generally,that many systemsare loosely structuredand plants, Elton (1958) argued that "the balance of rela- not at equilibrium(Wiens 1977),conclusions from clas- tively simplecommunities of plantsand animalsis more sic niche theorycontinue to stronglyinfluence the study easily upset than that of richer ones; that is more of biologicalinvasions. subject to destructiveoscillations in populations... The third origin of the belief that more diverse and more vulnerableto invasions".He supportedthis systems are less invasible lies in MacArthur's(1955) argumentwith observationsthat "oceanicislands and classic work on the relationshipbetween diversity and crop monoculturesare simple ecosystems that show stability.He suggestedthat systemswith fewerlinkages high vulnerabilityto invasions... and frequent out- were more subject to population fluctuationsif one breaksof populationsubsequently". Elton's notion that linkage was disrupted than were systems with more diverse systems are less invasible(also see Allee et al. linkages,work which spurredan enormousdebate over 1949) has been echoed many times by authors such as the diversity-stabilityrelationship (May 1973, Pimm Moyle (1986) examiningfish invasionsin North Amer- 1991). Though MacArthur(1955) did not evaluatein- ica and Moulton and Pimm (1983) examining bird vasibility,this work has contributedto the perception invasionsof Hawai'i.It is important,however, to clar- that diverse systems are more stable and possibly less ify what Elton (1958) was proposing. invasibleas resistanceto invasionis a commonmeasure As evidenced by the above quotations, Elton was of stability (Elton 1958). It is also importantto note interestedin communitystability, and focused on in- that MacArthur and May's models were based on vaders with particularlylarge impacts. He did not linkages among as well as within trophic levels. Yet clearly distinguishbetween the ability of non-resident very few empiricalstudies have examinedmulti-trophic species to establishin depauperatecommunities versus level interactionswhen invokingthe diversity-resistance the susceptibilityof organismsin those communitiesto linkage. the effects of invaders. Only the former is strictly a In this manuscriptwe examine evidence for a link questionof invasibility.This confusionbetween invasi- betweendiversity and invasibilitywith the goal of clar- bility and the vulnerabilityof communitiesto impact ifying some of the strengthsand weaknessesof current permeatesmuch of the literaturelinking diversityand approaches. Overall, we believe that the search for

16 OIKOS 87:1 (1999) generality in this area will be improved by careful the direct effects of ecological factors that influence definition of the important questions, attention to diversity is critical to properly interpreting these mechanism,consideration of relevantspatial scales, and studies. recognition that factors controlling native biological diversityalso control invasions. Review of the evidence We located the studies a BIO- Why should diversity affect invasibility? following by conducting SIS searchof articlesdating back to 1985,and examin- The most commonly cited and often implicit mecha- ing referencestherein. nism by whichdiversity confers resistance to invasionis that more diverseassemblages more fully utilize available thus little resource for resources, leaving space Models individuals of new species. This mechanism follows from the "species-packing"or "diffuse competition" All of the mathematicalmodels relatingdiversity and models of MacArthur(1970) described above, and is an invasibilityindicate that increasingdiversity decreases extension of the "empty niche" concept in invasion invasibility(Robinson and Valentine 1979, Post and biology (Herbold and Moyle 1986, Crawley 1987). Pimm 1983, Shigesadaet al. 1984, Drake 1988, 1990, More diversecommunities are also more likely to have Case 1990, 1991, Law and Morton 1996).This general a diverse guild of consumers,and this may limit the result was independentof whether the communities number of prey species that can potentially invade were composed of species competing via interference (Pimm 1991, Naeem and Li 1998).By contrast,Palmer (Shigesada et al. 1984) or exploitation (Case 1990, and Maurer (1997) point out that due to differences 1991), or if these specieswere membersof a single or among plants in canopy height, rooting depth, rhizo- multiple trophic levels (Post and Pimm 1983, Drake sphere microbialcommunities, etc., diverse communi- 1988, 1990). ties contain more microheterogeneity than Robinson and Valentine (1979) and Case (1990, monocultures(Aarssen 1983), and might thereforebe 1991)examined the invasibilityof stablemodel commu- more invasible. nities varyingin diversity.The interactionsand dynam- More diverse communitiesare also more likely to ics of these systems were describedby Lotka-Voltera contain species with particularlystrong impacts on equations with interactionstrength most often drawn invasibility.Whether or not this mechanismis consid- from a uniformdistribution. All inter-and intraspecific ered an effect of diversityper se has been considered pairwiseinteractions were compiled as the elementsof a elsewhere(Huston 1997, Tilman 1997a). Regardless,if communitymatrix, with matricesof varyingsize repre- a key species that is more often found in more diverse sentingcommunities of varyingdiversity. Each commu- assemblagesstrongly deters invasion, diverse communi- nity matrix was analyzed for local stability, and ties will show enhanced resistance.Alternatively, key unstablecommunities were discarded. The locallystable species that are strong facilitatorscan cause diverse communitieswere then "invaded"by a new species, systems to be more invasible. These facilitatorshave which expanded the matrix by one row and one been termed "diversity promoters" by Palmer and column, the elementsof which were the interactionsof Maurer (1997) and may be common in communities the invaderwith each of the residentspecies. Communi- where facilitation is an important structuringforce ties were invasible if the new species persistedin the (Bertnessand Callaway 1994, Callaway1995). In sum, expandedcommunity at equilibrium. while it is commonly believed that diversity should While the models of Case (1990, 1991)and Robinson enhanceresistance to invasion,arguments can be made and Valentine(1979) differedin how the elements of that diversitymay promote invasion. the communitymatrix were assigned,both found that Understandinghow invasibilityvaries with diversity the proportionof invaded communitiesdeclined with in natural systems is complicated by the fact that diversity. Case found that even invaders that were variationin diversityis controlledby, and thus covaries equivalentcompetitors to the residentspecies were of- with disturbance,resource availability, physical stress, ten unsuccessfulin more diversecommunities, a result competitors,consumers, etc., the same factors known attributedto the increasedfrequency of multiple do- to influenceinvasibility (Rejmanek 1989, Huston 1994, mains of attractionin more diversesystems (Gilpin and Robinson et al. 1995,Wiser et al. 1998).These covary- Case 1976). These multiple domains of attraction or ing factors are not intrinsiceffects of diversity,though equilibriaare more familiarin the context of the simple most of the studieswe reviewdo not make this impor- Lotka-Volteratwo speciescompetition model with un- tant distinction.Thus, recognizingthat correlationsbe- stable coexistencewhere the initially more abundant tween diversityand invasibilityin naturalcommunities speciesdominates. Case suggestedthat the importance reflect both the effect of diversityon invasibility,and of these "priorityeffects" increaseswith species diver-

OIKOS 87:1 (1999) 17 sity,,and therebydecreases community vulnerability to In each of the models of diversityand invasion,the competitorsentering in low numbers,such as invaders. interactionsof the invaderwere similarto those of the The assembly models of Post and Pimm (1983), residentspecies. In the assemblymodels, the invaders Drake (1988, 1990), and Law and Morton (1996) are were drawnfrom the same speciespool as the residents, also based on Lotka-Volteraequations, but they dif- while Robinson and Valentine(1979) and Case (1991) fered from those of Case (1990) and Robinson and drew invaderand residentspecies interactions from the Valentine(1979) in how the communitieswere "assem- same distribution.Though Case (1991) showed that bled". These models began with a finite species pool increasingthe invader'sgrowth rate had little effect on with all pairwise interactionsdrawn from a specified model results, what happenswhen invadersare better range of values. Specieswere then randomlydrawn to competitorsor predatorsthan all of the native species? invade initiallydepauperate communities. Invasion was Severalinvestigators have observedthat some success- successfulif the invaderincreased when rare,and main- ful invaders,particularly ones with large impacts,tend tained positive equilibrialabundance as inferredfrom to have qualitativelydifferent traits than the resident the communitymatrix expanded to includethat species species(Vitousek 1990, Chapin et al. 1994).We hypoth- (for a mathematicalcritique, see Case 1991, Law and esize that wheninvaders are superiorcompetitors or use Morton 1996). This process was iterated hundredsof differentresources than the nativespecies, the effectsof times, with each successful invasion yielding a new diversityon invasibilityare likely to be weak. locally stable equilibrium.Law and Morton (1996) Another important assumption of most of these providedan alternative"permanence" criteria for suc- models is that a communityis invasibleif the invader cessful invasion that allows for coexistenceor invader persistsat equilibrium.Crawley (1987) points out that successeven when speciesdo not tend to an equilibrium understandingwhat factors allow the equilibriumper- point. sistence of invaders, does not in itself explain why Regardlessof the invasioncriteria, all of these mod- communitiesare invasible in the first place (see also els indicatedthat the rate of successfulinvasions de- Rejmanek1989, Law and Morton 1996).Furthermore, clinedwith time, a resultthat has been inferredby some invaders that fail to persist in communitiesat equi- to mean that diversityenhances resistance. Case (1991), libriummay still stronglyimpact these systems.One of however,notes that due to the finite species pool, the the outcomesof Case's(1990, 1991)models are what he degree to which this result reflectsthe effects of diver- terms"indirect failures" - situationswhere the invader sity versusthe exhaustionof good invadersearly in the populationinitially grows, but in the processinfluences assembly process is unclear. Furthermore,Law and other speciessuch that its own populationgoes extinct. Morton (1996) showed that certain combinations of Case (1991) notes that though the frequencyof these speciesbetter resist invasion than othersof equal diver- indirectfailures is relativelysmall, their frequencyin- sity. These invasion resistant and presumablymore creases with diversity. Further exploring non-equi- connectedcommunities develop as the requisitemem- libriumoutcomes of invasionis an interestingarea for bers are introduced,and thus invasion resistancemay futurework. develop over time independentof communitydiversity. Lastly, in all of the models, prior to invasion, the Law and Morton (1996)also exploredhow the diver- communitieshave reached equilibrium,a state rarely sity of the speciespool influencescommunity invasibil- reachedby naturalsystems (Connell and Sousa 1984). ity and found that communitiesassembled from richer This mismatchbetween the models and field systemsis pools were more resistantto invasion. This result was furthercomplicated by the fact that communitydiver- not an effect of local diversitysince all communities, sity is often dynamic and the outcome of non-equi- regardlessof the pool size, tended towardsfive species. libriumprocesses. How diversityrelates to invasibility Rather, Law and Morton argued that communities in successionalor unsaturatedsystems is an interesting derived from richer pools have been tested by more questionripe for explorationin models. kinds of species and are thus less likely to be invaded by a new species than are communitiesof the same but assembledfrom a smaller diversity pool. studies There are severalfeatures of all of these models that Spatial pattern we believequalify their generality. First, the predictions Spatial pattern studies correlate the abundanceof a may be most appropriatefor relativelysmall spatial single or multiple invaders with the diversity of the scales. Lotka-Volteramodels assumewell-mixed popu- surroundingassemblage. These studies are not mecha- lations, such that all species interact with a potential nistic in that correlationsbetween diversityand inva- invader.This does not invalidateLotka-Voltera predic- sion reflectboth the independenteffects of diversityas tions, but ratherrestricts their applicabilityto scales at well as the effectsof processescovarying with diversity. whichall constituentspecies do interact- the neighbor- As we describebelow, it is these latter factors which hood scale (Tilman 1994, Naeem et al. 1998). Yet, as tend to drive broad scale field patterns.Nevertheless, we will show, empiricalwork at this scale is rare. these studies address the question, "Are naturallydi-

18 OIKOS 87:1 (1999) verse assemblagesmore or less invadedthan their spe- ity. For example,Timmins and Williams(1991) found cies-poor counterparts?".Interestingly, most of the that area could explain the positive correlation be- studies we found suggest that more diverse assem- tween native plant and weed diversity among forest blages are more invaded, contrary to Elton's (1958) and scrub reserves in New Zealand. By contrast, observations. MacDonald et al. (1986, 1989) analyzedinvasion pat- Note from the question above that spatial pattern terns among nature reserves in South Africa, and studies do not explicitly measure invasibility,but in- found that while more diverse reserveswere more in- use invader abundanceor as a stead diversity proxy vaded (R2 = 0.44, p < 0.001), reservesize was not cor- for to invasion. These measuresnot susceptibility only relatedwith the numberof exotic species.The positive reflect and features intrinsic to the com- invasibility correlation between native and exotic diversity was but also externalfactors, such as variationin munity, attributedto the greater environmentalheterogeneity propagule supply. In systems such as rivers, where in diversereserves (also see Higgins et al. 1999). propagulesupply covaries with residentspecies diver- Many studies have used equal size areas to explore sity (Nilsson et al. 1993), correlationsbetween invader of and invasion, and most indicated abundance and diversity may be driven entirely by patterns diversity correlationsbetween these factors. invader input. A second caveat about spatial pattern positive Examining invasion in communitiesof the studies is that they correlate invasion with current patterns multipleplant diversity, which is not diversity at the time of inva- central United States, Stohlgren et al. (1998, 1999), sion (but see Wiser et al. 1998). Given sufficienttime, showed that more diverse communities contained exotic species may reduce the diversityof invaded as- greater numbers of exotic species than their species- semblages,so the resultsof some of these studiesmay poor counterparts,though within communities,at the better reflect the impact of invaders on native diver- scale of 1 m2, results were mixed (see below). Focus- sity ratherthan vice versa. ing on ripariansystems, Planty-Tabacchiet al. (1996) Species lists from well defined areas, such as re- showed that sites with high native plant diversity serves, provide interesting data sets to explore pat- along the McKenzie River in Washington and the terns of diversity and invasion. Fox and Fox (1986) Adour River in France containedmore exotic species showed that the percentageof alien plant species is than sites with lower native diversity.They attributed negatively correlated with native plant diversity in this result to riverinefactors covaryingwith diversity. Australianheathland and shrublandreserves. In con- Pickard (1984) found that floristicallydiverse sites in et al. showed the trast, Knops (1995) that percentage the Lord Howe Islands had more exotic species than of alien was in more diverse plant species greater did species-poorsites, a result he attributedto the soil communitiesof coastal, central California. However, nutrientstatus of diversesites. Rejmanek (1989), Brown (1989), and others have In a of invasion Wiser et out that the of that are rigorousanalysis patterns pointed percentage species al. (1998) found that the occurrenceof the exotic forb alien is a biased measure of invasibilitybecause even if invasibilityis independentof diversity,diverse systems will contain a lower percentageof invaderssim- 50 because of the number of native in ply high species 45 - R2 = 0.77 * the denominator(similar problems emerge in a study g* P =0.002 by Bridgewaterand Backshall1981). For the study by S.4040 -> Knops et al., this means that the strengthof the posi- .~ 35 tive correlationwas under-estimated.For the study by Fox and Fox (1986), we recalculatedthe relationships. I 30- Using the numberof invadersrather than the percent- o 25 - as age the dependent variable, we found a positive - correlationbetween native richnessand exotic richness 20 for shrublandsites (Fig. 1) and a non-significantrela- Z 15 - tionship for the heathlandsites (R2 = 0.002, = 0.80). p 10 .... Similarly,Kruger et al. (1989) showed that for Med- 0 20 40 60 80 100 120 140 160 iterraneanplant systems in California, more diverse native communities also contained more non-native Numberof NativeSpecies species. Fig. 1. Reanalysisof the relationshipbetween diversity and Area can often be a confoundingvariable in these invasionfor shrublandsites presentedby Fox and Fox (1986). correlational studies between diversity and invasion See text for biases in the originalpresentation of these data. because areas are to contain both more Data were obtained by digitizing Fig. 5 in Fox and Fox larger likely We back-calculatedthe numberof exotic exotics and more (1986). species per natives (Case and Bolger 1991), site from the numberof native speciesand the percentageof leading to positive correlationsunrelated to invasibil- speciesthat were exotic.

OIKOS 87:1 (1999) 19 Hieraciumlepidulum in New Zealand was positively Invader addition studies correlatedwith richnessacross 250 species permanent Invaderaddition studies differ from the studies above Theirdata includedthree time over23 plots. points years in that is measured a fixed and measuresof a wide of other site conditions only invasibility by adding variety numberof to that influenceinvasion. propagules assemblagesvarying naturally might Using multiple logistic in the knows the out the factors diversity.Thus, investigator diversity regression,they partitioned covarying of the at the time of invasion. This elimi- with and still found a correla- assemblage speciesdiversity, positive nates the that the invader has somehow tion between and invasion.In anotherinvesti- possibility diversity influenced of native In addition, an invasion patterns diversity. by gationtracking throughtime, Holway (1998) the numberof added, invasibil- found that the of 20 ant controlling propagules spread Argentine populations reflectsfeatures intrinsic to the community,and not in Californiawas not correlatedwith nativeant richness. ity differential Nevertheless,like spatial Severalstudies have found that invadedcommunities propagulesupply. studies, diversity is not manipulated,so the are less diverse than their uninvaded In pattern counterparts. results are still driven by both diversity and factors each however,the invaderswere so abundantthat case, covarying with diversity. Nonetheless, invasibility is the resultsmost reflectedthe of the invader likely impact actuallymeasured, though how this is done can influ- on native diversity. Woods (1993) used the negative ence the conclusions. correlationbetween exotic coverand native honeysuckle Invasion is a probabilisticprocess (Crawley 1987, richnessin EasternUS foreststo that argue honeysuckle 1989). Diversity does not deterministicallyrender a native negativelyimpacted diversity.Similarly, Morgan communitycompletely resistant or vulnerableto inva- found a weak correlationbetween native (1998) negative sion, but rather,influences the probabilitythat a given richnessand exotic richness = < in an (R2 0.16,p 0.001) propagulewill be successful.One approachfor assess- Australian with low and grassland.By comparingplots ing invasibilityis to score communitiesas invaded or exotic he thatthis correlation reflected high cover, argued not invadedafter addinga given numberof propagules the of the exotic on native richness. impact species (e.g. Robinsonand Dickerson1984, McGrady-Steedet et al. that 1 m2 Similarly,Stohlgren (1999)reported plots al. 1997, Tilman 1997b).However, with this approach, in invaded communitiesof the placed more heavily the magnitudeof any effects on invasibilitydepend on betweennative CentralUS showednegative correlations the number of propagules added. Even if diversity and exotic diversity,while those in less invaded areas influences the probability of successful invasion, if showed positive relationships.Pysek and Pysek (1995) enough propagulesare added, all communitiesmay be reportedthat habitatsin the CzechRepublic invaded by invasible.The sensitivityof invadersuccess to propag- the tall forb Heracleum mantegazzianum had lower ule input is well known from the purposefulintroduc- diversityand richnessthan uninvadedplots. They re- tion of biocontrol agents (e.g. Crawley 1987, Hopper ported that H. mantegazzianum occupied between 50% and Roush 1993) and birds (Duncan 1997). Alterna- and 100%cover in the study area and replacesnative tively, by examining the proportion of successful vegetation,providing the simplestexplanation for the propagulesas a function of diversity(e.g. Robinson et lower diversityof invadedareas. al. 1995),studies can estimatethe per capitaprobability Several investigators have explored biogeographic of success.This probabilityis independentof the num- scalepatterns of diversityand invasion.Case and Bolger ber of addedpropagules as long as propaguledensity is (1991)found a weak negativecorrelation between native low enough to prevent interactions among the and exotic reptile diversityon islands throughoutthe colonists. world.They noted, though,that the degreeto whichthis A secondproblem with assessinginvasibility is deter- reflectsthe numberof native species per se versus the mining the point at which an invader is considered evolutionaryisolation of species poor islands remains "successful".We find the definitionof Crawley(1987) difficultto untangle.Weaker trends for birdsand mam- useful:an invaderis successfulif it increaseswhen rare. mals are presentedby Brown (1989) and Case (1996). Using this definition,invader addition studies evaluat- Rejmanek(1996) found that for vascular plants, the ing sexualor asexualpopulation growth make the most tropicsare less invadedthan the temperatezone, though convincingcase for examininginvasibility. he concluded that this correlation was unrelated to Robinson et al. (1995) added seeds of California diversity.In anotheranalysis, Rejmanek (1996) reported poppy (Eschscholziacalifornica) to grasslandplots that a positivecorrelation between the numberof nativeand varied naturallyin diversityand found that more di- exotic plants per log area on islands between50?S and verse plots were more invasible by this poppy than 50?N. He arguedthat native diversityseemed to be an species-poorplots. Specifically,they found that more indicator of favorable conditions. These results are poppiesgerminated, flowered, and producedfruit in the supportedby Lonsdale's(1997) study of global patterns more diverseplots. This resultcan be partly attributed plant invasions,which showedthat worldwide,commu- to the fact that Bromusdiandrus, a strongsuppressor of nitiesricher in nativeplant species are also moreinvaded. poppies and other plants was more abundantin low

20 OIKOS 87:1 (1999) diversity plots. In a similar study, Peart and Foin which calls to question the diversity effect (but"see (1985) examined the invasibility of patches within a Moulton 1993). In a similaranalysis of bird invasions coastal Californiagrassland by addingthe seeds of five into New Zealand,Duncan (1997)was unableto isolate grasses to plots varying in composition.Though their an effect of diversity. study did not focus on diversity,Peart and Foin found that plots dominated (near 100%cover) by a single productiveperennial grass were more difficultto invade Constructed community studies than were mixed species plots of annual grasses and forbs. More generally,low diversitypatches may often Studies that experimentallyconstruct communities, di- be the least invasible when the low diversity results rectly manipulatingdiversity are the only ones that from competitivedominance by one or a few species. experimentallyisolate the independenteffects of diver- Tilman (1997b) added seeds of 54 native species to sity on invasibility.In this light, they are the most Minnesota grasslandplots varying naturallyin diver- comparableto the modeling work. In addition, they sity. In contrast to Robinson et al. (1995) and Peart have the potentialto isolate the mechanismsunderlying and Foin (1985), he found that more diverseplots were the effects of diversity,though this has not yet been less invasible.The mechanism,however, remained un- done. clear. The effects of diversityon invasibilitywere not Knops et al. (1997) examinedthe number of inva- mediated by effects on nitrogen, since as Tilman sions into Minnesotagrassland plots whereplant diver- pointed out, plots with the less nutrientswere more sity had been experimentallymanipulated (see Tilman invaded. et al. 1996).Their results indicated that invasibilitywas a negativebut asymptoticfunction of diversity,providing some of the first empiricalevidence that diversity may deter invasion in a natural system (Naeem et al. 1998).Palmer and Maurer(1997) examinedweed inva- Assembly studies sion into plots plantedwith one or five crop plants. As Assemblystudies assess patternsof diversityand inva- in Tilman et al. (1996), the one-speciestreatment in- sion over time, asking, "Do communitiesbecome more cluded monoculturesof all species in the pool. They difficultto invade as time progresseswith the accumu- found that more diverseplots were significantlymore lation of species?".Invasibility is measuredas the num- invaded,though the mechanismsremained unclear. ber of new species establishing over a given time Due to the relativeease of manipulatingcommunities interval. If diverse communitiesbetter resist invasion, of microorganisms,these systems have proven useful the numberof new speciesin a communityshould level for examiningthe relationshipbetween diversity and off with time. However,as with the assemblymodels, a invasibility.Much like the models, these studiescreate declining number of invasions over time may also communitiesof varyingdiversity and then subjectthem reflect the formation of more connected communities to invasion. In the studies we describe below, the (Law and Morton 1996) or the exhaustion of good identityof the speciesin each replicatecommunity was invadersearly in the assemblyprocess, selectivelyleav- not randomlyassigned, and prior to invasion,the com- ing poorer invadersto establishlater on (Case 1991). munitieswere given time to equilibrate,during which Due to the difficultyof obtainingfrequent censuses species were lost. Thus it is difficultto determinethe of a developingcommunity and the long-termdata sets extent to which these studies reflectjust the indepen- required,most assemblystudies are conductedin mi- dent effects of diversity.Nevertheless, since the initial crocosms. These have been reviewed elsewhere by numberof speciesin each replicatewas determinedby Drake et al. (1996), the generalresult being that com- the investigator,diversity is still unlikely to be highly munitiesdecline in invasibilityover time. The degreeto correlatedwith other environmentalfactors. In Mc- which this depends on diversityis less clear. Moulton Grady-Steedet al. (1997), roughly 17 microorganism and Pimm (1983) took advantage of historical bird communitiescontaining up to 17 species, and three counts to examinethe successof avian introductionsto trophic levels were invaded with a facultativehetero- Hawai'i.They found that on certainislands, the success troph. Of the five successfullyinvaded communities, all rate of introductionsdecreased over time as the com- contained fewer than six species. Thus, their work munities accumulatedmore introduced species. They generally supports the hypothesis that diversity en- arguedthat this resultedfrom increasedcompetition in hances resistanceto invasion,though they arguedthat the more diversecommunities. Simberloff and Boecklen ratherthan a species-packingmechanism, the effect of (1991), however,showed that nearly all of the success- diversitywas attributableto particularkey speciesthat ful invadersin Moulton and Pimm's study were suc- were more abundantin the diversecommunities. cessful when introduced throughout the Hawaiian In a similartype of experiment,Robinson and Dick- islands, regardlessof island diversityand unsuccessful erson (1984) used threedifferent invaders to examineif species tended to be unsuccessfuleverywhere, a result diversity enhanced ecological resistance.The first in-

OIKOS 87:1 (1999) 21 100 exotic invaders, there is considerableopportunity to explore the diversity/invasibilityrelationship with an 80 , 80- experimentalapproach. Some revegetationstudies have exploredthe effects of using various native seed mixes on weed invasion (e.g. Tyser et al. 1998) but have not .a 60 specifically addressed the importance of species diversity. E 40- / o ^^^ o 20 Discussion O , I , , Currentmodels evaluatingthe effects of diversityon 1 2 3 4 5 6 7 invasibilityunilaterally support the commonbelief that Species Richness of the Community speciesdiversity enhances resistance to invasion.Those studies most similarto the the con- Fig. 2. Reanalysisof microcosmresults from Robinson and empirical models, Dickerson(1984). We used simplelinear regression to examine structedcommunity studies, only sometimessupported the relationshipbetween community diversity and the percent- this notion, suggestingthat diversitydoes not consis- age of communitiessuccessfully invaded three weeks after enhance resistancein natural communities.The introductionof (0) Staurastrum(R2 = 0.91,p < 0.05) and (@) tently Dictyosphaerium. (R2 = 0.97, p < 0.05). In order for these re- most strikingfinding of this reviewwas that the major- sults to be comparableto those of McGrady-Steedet al. ity of the spatial pattern and invaderaddition studies (1997),we selectedweek 3 data from Robinsonand Dickerson indicatedthat naturallydiverse assemblages tended to (1984) since week 2 data were not availablefor both invaders (McGrady-Steedet al. presentedweek 2 data). Nevertheless,a be more invasiblethan their species-poorcounterparts, positive relationship between diversity and invasibility contraryto Elton's (1958) observations.This findingis emergedregardless of when invasion was assessed.Diversity particularlyimportant from a conservationstandpoint levels with four or fewer replicateswere excludeddue to the since it that our most diversehabitats are the inherentinaccuracy of percentagescalculated from smallnum- suggests bers (all percentsin the figure were based on at least nine most prone to invasion (Stohlgrenet al. 1999). replicates). Why might diverseassemblages be the most invaded? Nearly all of the spatialpattern studies cited favorable vader, though capable of growing in isolation, failed to biotic or abiotic factors covaryingwith diversitythat invade any community, regardless of diversity. For the were presumed to underlie the positive correlations other two invaders, Robinson and Dickerson (1984) between diversityand invasibility,including moisture, concluded that persistencetime was independentof nutrients,habitat heterogeneity,physical factors, and diversity. This conclusion was based on a contingency competitors(e.g. Pickard1984, MacDonald et al. 1989, table analysis where the rows were progressively in- Robinson et al. 1995, Rejmanek1996, Stohlgrenet al. creasing levels of diversity, and the columns were pro- 1999).In general,unless native and exotic speciesdiffer gressively increasing invader persistence times. The markedly,the factorscontrolling native diversity should simple G-test used by the authors treated each cell as similarly control invaders, and thus the number of independent, and was inappropriately weak considering native speciessimply becomes an indicatorof biotic or that both the rows and columns were ordered (Feinberg abiotic conditions favorable to invasion (Rejmanek 1980). We reanalyzed their data using doubly ordered 1989;but see McIntyreand Lavorel 1994). contingency table analysis, and found that diversity Because of the importance of factors consistently significantly influenced invader persistence time for covaryingwith diversity,the observationthat naturally both species (linear by linear association tests, p < diverse communities are more invasible cannot be 0.05), but in contrast to their expectations, invaders translated to mean that diversity promotes invasion persisted longer in more diverse communities. We also (e.g. Stohlgren et al. 1999). While we believe that used simple linear regression to examine the relation- positive effects of diversity on invasibility are very ship between community diversity and the percentage much under-appreciated,the small scale at which the of communities successfully invaded three weeks after mechanismsunderlying the effects of diversityoperate introduction (Fig. 2). These results also indicated a precludesthem from drivingcommunity or landscape- significant positive relationship between diversity and level patterns.Each of the proposedmechanisms, such invasibility. as species-packingor the effects of key species,operate In spite of the possibilities provided by studies that on the scale of species interactions,or the neighbor- directly manipulate diversity, very little of such work is hood (Naeem et al. 1998; but see Law and Morton currently published. We believe that particularly in light 1996).Yet, the scale of nearlyall of the spatialpattern of the large number of habitat restoration projects and invader addition studies was significantlygreater currently taking place and management concerns over than the neighborhood,with most examining entire

22 OIKOS 87:1 (1999) communitiesor landscapes.We believe that indepen- The biotic resistance paradigm dent of whetherthe neighborhoodeffects of diversity Much of the researchand the on invasion are positive or negative these effects are hypothesesconcerning effects of on invasion have been likely to be swampedby factors covaryingwith diver- diversity strongly driven the focus on "biotic a focus that sity at broader spatial scales. This hypothesis is by resistance", is evidentin the seminalwork of Elton Basedon analogous to the observationthat competitorswhich (1958). the resultsof this we believe that the are negativelyassociated at small scales are often pos- review, resolving and more itively associated at broader scales due to similar diversity-invasibilityrelationship generally, the of invasions that habitat requirements(Wiens 1986). enhancing predictability requires Positive correlationsbetween native and exotic di- we not only uncoverthe communitymembers that repel but also those that facilitateinvasion. A new versity suggest that if local scale diversity enhances invaders, focus on that invasionsis to resistanceto invasion, this effect is weak relative to species promote beginning Richardsonet al. that mutual- other factors influencinginvasions in these systems. emerge. (1999) suggest isms between introduced and native Quantitativelyevaluating the importanceof diversity plants organisms are relative to disturbance, nutrients, consumers, and (pollinators, dispersers,fungi, N-fixing bacteria) common and are in other factors limiting invasions has rarely been ac- quite possiblyvery important pro- invasion.Palmer and Maurer and Wiser complished, though is an important goal for future moting (1997) work in this area. Such information is essential to et al. (1998) discuss "invasionpromoters", whose in- creased abundancein areas of make accuratelypredicting the consequencesof species loss greaterdiversity these areas to invasion. for communityinvasibility. particularlysusceptible Indeed, a surprisingfraction of the empiricalstudies suggest that in some systemsdiversity might promoteinvasion (Robinsonand Dickerson1984 (reanalyzed),Robinson et al. 1995,Palmer and Maurer1997, Wiser et al. 1998). Are native and exotic species all that different? Even more generally,facilitations, both directand indirect and A common perceptionin the study of biologicalinva- (Bertness Callaway 1994, Callaway 1995, Levine are as sions is that exotic species somehow differ from the 1999) increasinglyrecognized important forces in and the residents of the communitiesthey invade. However, structuring communities, implications of this for the should the results of spatial pattern and invader addition diversity-invasibilityrelationship not be studies suggest that exotic and native species respond ignored. to the same factors. The notion that major differences separateinvader and resident species is based largely on of what makes a invader classic theories good Conclusions (Baker 1965) and the invaderimpact literature.While it does seem true that invaders with large impacts Ecologistshave long been searchingfor a generalthe- differ considerably from native species (Vitousek ory of the factors controllinginvasions. Based on the 1990, Chapin et al. 1994), the traits behind large im- results of this review, and more specifically,the sug- pacts need not be the same as those allowing the gestion that exotic and resident species show similar invader to enter in the first place. Moreover, while responsesto the environment,we believe that our un- the exact percentageis debated,most invadersdo not derstandingof the ecologicalcontrols over native spe- have major impacts (Simberloff 1981, Herbold and cies diversity may provide our best theory of the Moyle 1986, Moyle and Light 1996). controls over invasibility.It is no coincidencethat the That invaders and resident species must be some- dominantfactors known to influencediversity such as what similar is also an implicit assumptionof argu- competition, disturbance, resource availability, and ments regarding species-packing: the resource propagule supply (see Huston 1994 for review) are requirementsof the invader and the resident species the same factors known to influenceinvasions. must overlap for the species-packingmechanism to Consideringthe controversialfoundations and weak operate. If an invader requires a portion of the re- empiricalsupport for a negative relationshipbetween source space outside that of the residentspecies, then diversityand invasibility,we believethat thereis much residentspecies diversity may be inconsequentialto its room for innovativemodels and carefulempirical stud- invasion. Thus, if invaders with large impacts are ies that go beyond currentapproaches. More specifi- truly differentfrom the residentsof the communities cally, we urge investigatorsto explore the mechanisms they invade, an interestingproblem arises:the factors underlyingthe effects of diversityon invasibility.Em- controllinginvasion may differ for species with large pirical studies rigorously evaluating mechanismsare and small impacts. We note that few diversity-in- conspicuouslyabsent from the literature,yet a mecha- vasibility investigationshave focused on high-impact nistic understandingof the diversity-invasibilityrela- species. tionship is essential for accurately predicting the

OIKOS 87:1 (1999) 23 assemblages or circumstances where diversity matters D'Antonio, C. M. and Dudley, T. L. 1995. Biologicalinva- - most. Furthermore, what mechanistic evidence we do sions as agents of change on islands versusmainlands. In: P. L. L. and H. have little for the often-cited Vitousek, M., Loope, Adsersen, (eds), provides support species Islands: biological diversity and ecosystem function. packing mechanism and biotic resistance paradigm. Springer-Verlag,Berlin, pp. 103-121. Moving beyond this paradigm is critical to better un- Drake, J. A. 1988. Models of communityassembly and the the and structureof ecological landscapes.- In: Hallam, T. G., derstanding diversity-invasibility relationship L. J. and S. A. Mathematical more the of Gross, Levin, (eds), ecology: generally, enhancing predictability biologi- proceedingsof the autumncourse research seminars, Inter- cal invasions. nationalCentre for TheoreticalPhysics, Miramare-Trieste, Italy, November 24-December 12, 1986. World Press, 585-605. Acknowledgements- Commentsby Mary Power, Will Sat- Singapore,pp. terthwaiteand Sousa the Drake,J. A. 1990.The mechanicsof communityassembly and Wayne greatlyimproved manuscript. succession.- J. Theor. Biol. 147: 213-233. JML acknowledgessupport from a NSF graduateresearch and dissertation and the Drake, J. A., Mooney, H. A., DiCastri,F. et al. (eds) 1989. fellowship improvement grant, - Williamand Flora Hewlett Foundation.CMD acknowledges Biologicalinvasions: a global perspective. John Wiley, supportfrom the MillerInst. for Basic Researchin Science. Chichester. Drake, J. A., Huxel, G. R. and Hewitt, C. L. 1996. Micro- cosms as models for generatingand testing community theory. - Ecology 77: 670-677. Duncan,R. P. 1997.The role of competitionand introduction References effort in the success of passeriformbirds introducedto New Zealand.- Am. Nat. 149: 903-915. Aarssen,L. W. 1983. Ecologicalcombining ability and com- Elton, C. S. 1958. The ecology of invasionsby animalsand petitivecombining ability in plants:toward a generalevo- plants. - Methuen,London. lutionarytheory of coexistencein systemsof competition. Feinberg,S. E. 1980. The analysisof cross-classifiedcategori- - Am. Nat. 122: 707-731. cal data. - MIT Press, Cambridge,MA. Allee, W. C., Emerson,A. E., Park, O. et al. 1949.Principles Fox, M. D. and Fox, B. J. 1986.The susceptibilityof natural of animalecology. - Saunders,Philadelphia. communitiesto invasion.- In: Groves,R. H. and Burdon, Baker, H. G. 1965. Characteristicsand modes of origin of J. J. (eds), Ecology of biologicalinvasions: an Australian weeds. - In: Baker,H. G. and Stebbins,G. L. (eds), The perspective.Australian Academy of Science,Canberra, pp. genetics of colonizing species. Academic Press, London, 57-66. pp. 147-172. Gilpin, M. E. and Case, T. J. 1976. Multiple domains of Bertness,M. D. and Callaway,R. M. 1994. Positive interac- attractionin competitioncommunities. - Nature261: 40- tions in communities.- TrendsEcol. Evol. 9: 191-193. 42. Bridgewater,P. B. and Backshall,D. J. 1981. Dynamics of Herbold, B. and Moyle, P. B. 1986. Introducedspecies and some WesternAustralian ligneous formations with special vacant niches. - Am. Nat. 128:751-760. - referenceto the invasionof exotic species. Vegetatio46: Higgins, S. I., Richardson, D. M., Cowling, R. M. and 141-148. Trinder-Smith,T. H. 1999. Predictingthe landscape-scale Brown,J. H. 1989. Patterns,modes, and extents of invasions - distributionof alien plants and their threatto plant diver- by vertebrates. In: Drake, J. A., Mooney, H. A., Di- - Conserv.Biol. 13: 303-313. F. et al. invasions:a sity. Castri, (eds), Biological global per- Holway, D. A. 1998. Factors governingrate of invasions:a spective.John Wiley, Chichester,pp. 85-109. - - naturalexperiment using Argentineants. Oecologia115: Callaway,R. M. 1995. Positiveinteractions among plants. 206-212. Bot. Rev. 61: 306-349. Hopper,K. R. and Roush, R. T. 1993.Mate finding, dispersal, Case,T. J. 1990.Invasion resistance arises in stronglyinteract- - number released, and the success of biological control ing species-richmodel competitioncommunities. Proc. introductions.- Ecol. Entomol. 18: 321-331. Natl. Acad. Sci. USA 87: 9610-9614. T. J. 1991. Invasion and Huston, M. A. 1994. Biologicaldiversity: the coexistenceof Case, resistance,species build-up on - Univ. Press, communitycollapse in metapopulationmodels with inter- species changinglandscapes. Cambridge - Biol. J. Linn. Soc. 42: 239-266. Cambridge. speciescompetition. M. A. 1997. Hidden treatmentsin Case, T. J. 1996. Global patternsin the establishmentand Huston, ecologicalexperi- distributionof exotic birds. - Biol. Conserv.78: 69-96. ments:re-evaluating the ecosystemfunction of biodiversity. - 110:449-460. Case, T. J. and D. T. 1991. The role of introduced Oecologia Bolger, J. M. J. R. and A. C. 1995. speciesin shapingthe distributionand abundanceof island Knops, H., Griffin, Royalty, - Evol. Ecol. 5: 272-290. Introducedand native plants of the Hastings reservation reptiles. - Chapin, F. S. III, Reynolds, H. L., D'Antonio, C. M. and centralcoastal California:a comparison. Biol. Conserv. Eckhart, V. M. 1994. The functional role of species in 71: 115-123. terrestrialecosystems. - In: Walker, B. and Steffen, W. Knops, J. M. H., Tilman,D., Naeem, S. and Howe, K. M. International 1997. Biodiversityand plant invasions in experimental (eds), Geosphere-BiosphereProgramme - Book Series,Vol. 2. Global Changeand TerrestrialEcosys- grasslandplots. Bull. Ecol. Soc. Am. 78: 125. tems. CambridgeUniv. Press, Cambridge,pp. 403-428. Kruger,F. J., Breytenbach,G. J., MacDonald,I. A. W. and Connell,J. H. and Sousa,W. P. 1984.On the evidenceneeded Richardson,D. M. 1989. The characteristicsof invaded to judge ecologicalstability or persistence.- Am. Nat. 121: Mediterranean-climateregions. - In: Drake, J. A., 789-824. Mooney, H. A., DiCastri,F. et al. (eds), Biologicalinva- Crawley,M. J. 1987. What makes a communityinvasible? - sions: a global perspective.John Wiley, Chichester,pp. In: Crawley,M. J. and Edwards,P. J. (eds), Colonization, 181-213. succession,and stability.Blackwell Scientific, Oxford, pp. Law, R. and Morton,R. D. 1996.Permanence and the assem- 429-451. bly of ecologicalcommunities. - Ecology 77: 762-775. Crawley,M. J. 1989. Chance and timing in biologicalinva- Lawlor, L. R. 1979. Direct and indirecteffects of n-species sions. - In: Drake, J. A., Mooney, H. A., DiCastri,F. et competition.- Oecologia43: 355-364. al. (eds), Biological invasions:a global perspective.John Levine,S. H. 1976. Competitiveinteractions in ecosystems.- Wiley, Chichester,pp. 407-423. Am. Nat. 110: 903-910.

24 OIKOS 87:1 (1999) Levine,J. M. 1999. Indirectfacilitation: evidence and predic- Peart,D. R. and Foin, T. C. 1985.Analysis and predictionof tions from a ripariancommunity. - Ecology 80: 1762- populationand communitychange: a grasslandcase study. 1769. - In: White, J. (ed.), The populationstructure of vegeta- Lodge, D. M. 1993. Biologicalinvasions: lessons for ecology. tion. Junk, Dordecht,pp. 313-339. - TrendsEcol. Evol. 8: 133-137. Pickard, J. 1984. Exotic plant distributionon Lord Howe Lonsdale,M. W. 1997. Global patternsof plant invasions.- Island: distributionin space and time, 1853-1981. - J. Bull. Ecol. Soc. Am. 78: 21. Biogeogr.11: 181-208. MacArthur,R. H. 1955. Fluctuationsof animal populations Pimm,S. L. 1991.The balanceof nature?- Univ. of Chicago and a measure of community stability. - Ecology 36: Press, Chicago. 533-536. Planty-Tabacchi,A., Tabacchi,E., Naiman, R. J. et al. 1996. MacArthur,R. H. 1970. Species-packingand competitive Invasibilityof speciesrich communities in riparianzones. - equilibriumfor many species. - Theor. Popul. Biol. 1: Conserv.Biol. 10: 598-607. 1-11. Post, W. M. and Pimm,S. L. 1983.Community assembly and MacArthur,R. H. 1972.Geographical ecology: patterns in the food web stability.- Math. Biosci. 64: 169-192. distributionof species. - Harperand Row, New York. Pysek, P. and Pysek, A. 1995. Invasionby Heracleumman- MacDonald,I. A. W., Powrie,F. J. and Siegfried,W. R. 1986. tegazzianumin differenthabitats in the Czech Republic.- The differentialinvasion of South Africa's biomes and J. Veg. Sci. 6: 711-718. ecosystemsby alien plantsand animals.- In: MacDonald, Rejmanek,M. 1989. Invasibilityof plant communities.- In: I. A. W., Kruger,F. J. and FerrarA. A. (eds),The ecology Drake, J. A., Mooney, H. A., DiCastri, F. et al. (eds), and management of biological invasions in southern Biological invasions: a global perspective.John Wiley, Africa. OxfordUniv. Press, Cape Town, pp. 209-225. Chichester,pp. 369-388. MacDonald, I. A.W., Loope, L. L., Usher, M. B. and Rejmanek,M. 1996. Speciesrichness and resistanceto inva- Hamann,0. 1989. Wildlifeconservation and the invasion sions. - In: Orians,G., Dirzo, R. and Cushman,H. (eds), of naturereserves by introducedspecies: a global perspec- Biodiversityand ecosystem processesin tropical forests. tive. - In: Drake,J. A., Mooney,H. A., DiCastri,F. et al. Springer-Verlag,Berlin, pp. 153-172. (eds), Biologicalinvasions: a global perspective.John Wi- Richardson,D. M., Allsopp,N., D'Antonio,C. M. et al. 1999. ley, Chichester,pp. 215-255. Plant invasions - the role of mutalisms.- Biol. Rev. (in May, R. M. 1973. Stabilityand complexityin model ecosys- press). tems. - PrincetonUniv. Press, Princeton,NJ. Robinson, G. R., Quinn, J. F. and Stanton, M. L. 1995. May, R. M. and MacArthur,R. H. 1972. Niche overlapas a Invasibilityof experimentalhabitat islands in a California functionof environmentalvariability. - Proc. Nat. Acad. winterannual grassland. - Ecology 76: 786-794. Sci. USA 69: 1109-1113. Robinson,J. V. and Valentine,W. D. 1979. The conceptsof McGrady-Steed,J., Harris, P. M. and Morin, P. J. 1997. elasticity,invulnerability, and invadibility.- J. Theor.Biol. Biodiversityregulates ecosystem predictability. - Nature 81: 91-104. 390: 162-165. Robinson,J. V. and Dickerson,J. 1984.Testing the invulnera- McIntyre,S. and Lavorel,S. 1994. Predictingthe richnessof bility of laboratory island communitiesto invasion. - native, rare, and exotic plants in responseto habitat and Oecologia61: 169-174. disturbancevariables across a variegated landscape. - Shigesada,N., Kawasaki, K. and Teramoto, E. 1984. The Conserv.Biol. 8: 521-531. effects of interferencecompetition on stability,structure, Mooney, H. A. and Drake, J. A. (eds) 1986. Ecology of and invasionof a multispeciessystem. - J. Math. Biol. 21: biological invasions of North America and Hawaii. - 97-113. Springer-Verlag,New York. Simberloff,D. 1995. Why do introducedspecies appear to Morgan,J. W. 1998.Patterns of invasionof an urbanremnant devastateislands more than mainlandareas? - PacificSci. of a species-richgrassland in southeasternAustralia by 49: 87-97. non-nativeplant species. - J. Veg. Sci. 9: 181-190. Simberloff,D. 1986. Introducedinsects: a biogeographicand Moulton, M. P. 1993. The all-or-nonepattern in introduced systematicperspective. - In: Mooney,H. A. and Drake,J. Hawaiianpasseriforms the role of competitionsustained. - A. (eds), Ecologyof biologicalinvasions of North America Am. Nat. 141: 105-119. and Hawaii. Springer-Verlag,New York, pp. 3-26. Moulton,M. P. and Pimm,S. L. 1983.The introducedHawai- Simberloff,D. 1981.Community effects of introducedspecies. ian avifauna:biogeographic evidence for competition.- - In: Nitecki, T. H. (ed.), Biotic crises in ecologicaland Am. Nat. 121: 669-690. evolutionarytime. AcademicPress, New York, pp. 53-81. Moyle, P. B. 1986. Fish introductionsinto North America: Simberloff,D. and Boecklen,W. 1991. Patternsof extinction patternsand ecologicalimpact. - In: Mooney, H. A. and in the introducedHawaiian avifauna: a reexaminationof Drake, J. A. (eds), Ecology of biological invasions of the role of competition.- Am. Nat. 138: 300-327. North Americaand Hawaii. Springer-Verlag,New York, Stohlgren,T. J., Bull, K. A., Otsuka,Y. et al. 1998.Riparian pp. 27-43. zones as havens for exotic plant species in the central Moyle, P. B. and Light, T. 1996.Fish invasionsin California: grasslands.- Plant Ecol. 138: 113-125. do abioticfactors determine success? - Ecology77: 1666- Stohlgren,T. J., Binkley,D., Chong,G. W. et al. 1999.Exotic 1670. plant speciesinvade hot spots of native plant diversity.- Naeem, S. and Li, S. 1998. Consumerspecies richnessand Ecol. Monogr.69: 25-46. autotrophicbiomass. - Ecology 79: 2603-2615. Stone, L. and Roberts,A. 1991. Conditionsfor a speciesto Naeem, S., Knops, J., Tilman, G. D. et al. 1998. Resident gain advantagefrom the presenceof competitors.- Ecol- plant diversityand resistanceto biologicalinvasion at the ogy 72: 1964-1972. level of the plantneighborhood. - Abstractsof the Annual Tilman, D. 1994. Competitionand biodiversityin spatially Meeting of the EcologicalSociety of America,Baltimore, structuredhabitats. - Ecology 75: 2-16. p. 99. Tilman,D. 1997a.Distinguishing between the effectsof species Nilsson, C., Nilsson, E., Johansson,M. E. et al. 1993. Pro- diversityand speciescomposition. - Oikos 80: 185. cesses structuringriparian vegetation. - Curr.Topics Bot. Tilman,D. 1997b.Community invasibility, recruitment limita- Res. 1: 419-431. tion, and grasslandbiodiversity. - Ecology 78: 81-92. Palmer, M. W. and Maurer,T. 1997. Does diversitybeget Tilman,D., Wedin,D. and Knops, J. 1996. Productivityand diversity?A case study of crops and weeds. - J. Veg. Sci. sustainability influenced by biodiversity in grassland 8: 235-240. ecosystems.- Nature 379: 718-720.

OIKOS 87:1 (1999) 25 Timmins,S. M. and Williams,P. A. 1991. Weed numbersin Wiens,J. A. 1977.On competitionand variableenvironments. New Zealand'sforest and scrubreserves. - N. Z. J. Ecol. - Am. Sci. 65: 590-597. 15: 153-162. Wiens, J. A. 1986. Spatial scale and temporalvariation in Tyser,R. W., Asebrook,J. M., Potter,R. W. and Kurth,L. L. shrubsteppebirds. - In: Diamond,J. and Case, T. J. (eds), 1998. Roadside revegetationin Glacier National Park, Communityecology. Harper and Row, New York, pp. U.S.A.: effects of herbicide and seeding treatments. - 154-172. RestorationEcol. 6: 197-206. Wiser, S. K., Allen, R. B., Clinton, P. W. and Platt, K. Vitousek, P. M. 1990. Biological invasions and ecosystem H. 1998. Communitystructure and forest invasion by processes:towards an integration,of population biology an exotic herb over 23 years. - Ecology 79: 2071- and ecosystemstudies. - Oikos 57: 7-13. 2081. Vitousek,P. M., D'Antonio, C. M., Loope, L. L. et al. 1997. Woods, K. D. 1993. Effects on invasionby Loniceratatarica Introducedspecies: a significantcomponent of human- L. on herbsand tree seedlingsin four New Englandforests. caused global change. - N. Z. J. Ecol. 21: 1-16. - Am. Midi. Nat. 130: 62-74.

26 OIKOS 87:1 (1999)