Positive Interactions of Nonindigenous Species: Invasional Meltdown?
<ul><li><p>Biological Invasions 1: 2132, 1999. 1999 Kluwer Academic Publishers. Printed in the Netherlands.</p><p>Positive interactions of nonindigenous species: invasional meltdown?</p><p>Daniel Simberloff & Betsy Von HolleDepartment of Ecology and Evolutionary Biology, University of Tennessee, Knoxville, TN 37996-1610, USA;Author for correspondence (e-mail: firstname.lastname@example.org; fax: C1-423-974-3067)</p><p>Received 4 June 1998; accepted in revised form 22 February 1999</p><p>Key words: biotic resistance, dispersal agent, facilitation, habitat modification, indirect effects, mutualism,pollination, synergism</p><p>Abstract</p><p>Study of interactions between pairs or larger groups of nonindigenous species has been subordinated in the literatureto study of interactions between nonindigenous and native species. To the extent that interactions among introducedspecies are depicted at all, the emphasis has been on negative interactions, primarily resource competition andinterference. However, a literature search reveals that introduced species frequently interact with one another and thatfacilitative interactions are at least as common as detrimental ones. The population significance of these interactionshas rarely been determined, but a great variety of types of direct and indirect interactions among individuals ofdifferent nonindigenous species is observed, and many are plausibly believed to have consequences at the populationlevel. In particular, mutualisms between plants and the animals that disperse and/or pollinate them and modificationof habitat by both animals and plants seem common and often important in facilitating invasions. There is littleevidence that interference among introduced species at levels currently observed significantly impedes furtherinvasions, and synergistic interactions among invaders may well lead to accelerated impacts on native ecosystems an invasional meltdown process.</p><p>Introduction</p><p>The concept of environmental resistance was intro-duced by Chapman (1931) to describe the forces, pri-marily biotic, that hinder the establishment of speciesin a new location. The emphasis (e.g., Elton 1958;Udvardy 1969) has been on the biological aspects ofthis resistance the complex of native predators, par-asites, pathogens, and competitors, as well as previ-ously introduced species, that oppose a newly arrived,nonindigenous propagule. Thus, the notion is perhapsmore accurately described as biotic resistance, andit has dominated thinking on why some introducedspecies survive and spread while others die out orpersist tenuously and perhaps temporarily (Simberloff1986). For example, two of the commonest generaliza-tions claimed for introduction success that islands aremore easily invaded than mainland, and that disturbedhabitats are more readily invaded than pristine ones are both usually interpreted as at least partly due to dif-ferences in species richness of the recipient community</p><p>(Simberloff 1986 and references therein). Islands usu-ally have fewer species than mainland, and disturbedareas have fewer species than undisturbed ones. Simi-larly, occasional claims have been voiced that speciesintroduced earlier to a site have excluded one or moreof those introduced later. For example, in the biolog-ical control literature, a debate raged about whethersome introductions of natural enemies of Homoptera,Coleoptera, and Lepidoptera failed because the specieswere excluded by others previously introduced for thesame purpose (references in Simberloff and Boecklen1988). For the introduced Hawaiian avifauna, Moultonand Pimm (1983) argued that species introduced laterhad a higher probability of going extinct preciselybecause of competition with species introduced earlier.</p><p>The view that biotic resistance determines invasionsuccess or failure fits well with the dynamic equilib-rium model of island biogeography (MacArthur andWilson 1963, 1967), which became enormously popu-lar for some 20 years after its publication (Simberloff1974; Williamson 1989). The model posits that the</p></li><li><p>22</p><p>arrival of a new species on a real or habitat island will becompensated for by the extinction of a species alreadypresent, so that the number of species remains constant.The model refers only to species richness, not speciesidentities, though implicitly the notion in the originalpapers and most successors was that extinction wouldlikely occur in a species taxonomically related to theinvader. This view was consistent with the dominantinterpretation that the main force causing the extinc-tion was competition, especially diffuse competition(Simberloff 1981). Some of the papers invoking bioticresistance to explain the failure of biological controlintroductions (e.g., Tallamy 1983) specifically relatethe failure to the island biogeographic model, as doesthe study of the introduced Hawaiian birds (Moultonand Pimm 1983). Additionally, the original and almostall subsequent depictions of the extinction curve inthe equilibrium theory were concave upward thatis, the per-species rate of extinction rose faster thanthe number of species did. This approximately expo-nential shape was partly attributed to the increasingprobability of interference among species with a resul-tant accelerating detrimental effect (MacArthur andWilson 1967, p. 22).</p><p>However, a diametrically opposite conception ofinvasion success is also possible, one in which non-indigenous species, instead of interfering with oneanother, facilitate each others establishment and/orcontinued existence. Thus, for example, Crosby (1986)depicts the colonization of the Americas, Australia,New Zealand, and the Canary Islands by European ani-mals, plants, and pathogens as a mutualistic process.The European species, including European humans,are seen as biological allies that together constituteda synergistic juggernaut crushing native peoples andtheir ecosystems because the European species hadcoevolved with humans and with one another. Forinstance, European weeds, having coevolved with pigs,sheep, and cattle, were adapted to their activities, whileNorth American plants were devastated by them.</p><p>Similarly, many classic examples of the ravages ofintroduced species include facets that entail facilita-tion, either one-way or two-way, between differentintroduced species. Elton (1958) describes how theArgentine ant, Linepithema humile, tends the Asian redscale insect, Aonidiella aurantii, in California citrusorchards. Ant removal demonstrated that scale densi-ties were several times higher on trees with ants thanon trees without them; ants remove some scale nat-ural enemies. The population impact of the scale onthe ant is not described, but it surely is not negative.</p><p>This example of interaction between a neotropical antand an Asian scale in California suggests that facili-tation among nonindigenous species need not be gen-erated by a coevolutionary history. Also, the fact thatsome successful biological control programs entail newassociations between parasitoid and host or predatorand prey (Hokkanen and Pimentel 1984, 1989) indi-cates that at least some introduced species can bene-fit from interactions with others even if they are notcoevolved.</p><p>In short, it is possible to imagine an invasion modelvery different from the dominant scenario of bioticresistance. At the most basic level, positive interactionsamong invaders may, for at least some of them, enhancethe probability of survival and/or increase populationsize. In such instances, there may or may not be syn-ergy that is, a greater impact of a group of invaderson the recipient community than would have been pre-dicted by the summed impacts of the individual species.Howarth (1985) foresaw this possibility:</p><p>Often two or more harmful alien species may act inconsort so that their joint impact is more severe thanthat of the several species acting separately. Even anotherwise innocuous or seemingly beneficial alienmay, in fact, act in consort with other aliens with aconsequent synergistic effect, causing great harm tothe native biota. (p. 163)</p><p>We suggest the term invasional meltdown for the pro-cess by which a group of nonindigenous species facili-tate one anothers invasion in various ways, increasingthe likelihood of survival and/or of ecological impact,and possibly the magnitude of impact. Thus, thereis an accelerating accumulation of introduced speciesand effects rather than a deceleration as envisioned inthe biotic resistance model. The analogy to mutationalmeltdown (Lynch et al. 1995) is evident. Our purpose inthis paper is to determine the frequency of invasionalmeltdown and the variety of processes that can con-tribute to it. Is it a rare phenomenon generated by a fewsorts of interactions, or is it occurring all around us?</p><p>Methods</p><p>To assess the frequency of facilitative interactions thatis, enhanced survival and/or population size amongnonindigenous species, we used a data base, compiledby Ingrid Parker, consisting of all papers between theyears 1993 and 1997 in the Biosis data base that hadthe key words species AND inva# OR introduced OR</p></li><li><p>23</p><p>alien OR exotic OR non-native OR non-indigenous.These keywords did not specifically address impact oreffect of introduced species, and the majority of papersdid not. The data base excluded any reports of biologi-cal control agents on their target organisms but includedthose of biological control agents on non-targets. Therewere over 5,000 papers. We determined the journalsrepresented by the greatest number of papers for oneyear in this data base. Three journals (Biodiversity andConservation, Biological Conservation, and Ecology)each had approximately twice as many papers as didany other journal. In addition to these three journals,we choose four of the twelve journals that fell in thenext frequency category (Conservation Biology, Jour-nal of Animal Ecology, Journal of Applied Ecology, andNatural Areas Journal) because it was clear from thetitles in the data base that these journals tended to havearticles that dealt with effects of introduced species inaddition to just presence and absence data. For theseseven journals, we examined all 254 articles in the database to determine the extent and nature of facilitativeinteractions among two or more introduced species.</p><p>The articles fell into four categories. In those denotedC=C, individuals of two or more nonindigenousspecies each benefited from the presence of the other(s).The C=0 category described situations in which indi-viduals of one species benefited from the presence ofthe other, while the second species was not knownto affect individuals of the first. The category C=included invasions in which individuals of one non-indigenous species benefit from the presence of a sec-ond species, while the individuals of the second werenegatively affected by those of the first. Finally, to rep-resent a sort of interaction envisioned as frequent inthe governing paradigm, competition or other formsof mutual detriment between individuals of pairs ofspecies, we used the category =. In the greatmajority of tabulated studies, the population impactof one species on another was not demonstrated. Forexample, in cases recorded as C=C, the activities ofindividuals of each species were shown to benefit indi-viduals of the other, but there was generally no evi-dence on the effect of these activities on the populationof either species, though often an effect was reason-ably inferred. Thus, the quantitative evidence for facil-itation and even more so for synergy is usuallyabsent or weak. For example, grazing nonindigenousmammals may disperse the seeds of a nonindigenousplant, and may even aid their germination, but repli-cated, quantitative, probably experimental study wouldbe required to show that the population of either species</p><p>is therefore more likely to survive or be greater than itwould have been without the other species. Of course,the same caveat applies to arguments that two non-indigenous species negatively affect one another.</p><p>In addition to the classification of articles from thedata base, we sought examples from the literature(including the data base, but also other journal articles,books, and gray literature sources) to depict the rangeof ways in which nonindigenous species can facilitateone another.</p><p>Results</p><p>The numbers of different types of interactions aredepicted in Table 1. Of the 254 articles reviewed, 30recorded at least one interaction between two intro-duced species, and one recorded a great number: anintroduced phytophage eating many introduced plantspecies. The majority of perceived introductions (156)are of the latter sort at least at the individual level,one species benefits and the other is harmed (C=), aswhen one species eats another. However, it is notewor-thy that almost as manyC=C cases were adduced (10),in which two species facilitate one another, as =cases (12) that accord with the governing paradigmof mutual interference or competition. In addition, 12instances were recorded in which individuals of oneintroduced species benefit, while individuals of theother are unaffected (C=0). No case of amensalism(=0) was found. Table 1 also lists the more specificnature of the interactions of the four types. With respectto the preponderance of C= interactions entailingplants and phytophagous insects, it should be notedthat 128 of these were reported in a single study. Noother study reported even ten interactions, and mostreported just one.Table 1. Numbers of different types of interactions betweenintroduced species cited in 254 articles in seven journals duringa five-year period (see text).</p><p>Interaction Number Nature of interactiontype</p><p>C=C 10 Disturbance D 6, indirect effects D 3,pollination D 1</p><p>C=0 12 Disturbance D 9, commensalism D 1,host/parasite and similarinteractions D 2</p><p>C= 156 Predator/prey D 23, phytophagousinsect/plant D 131, other D 2</p><p>= 12 Competition D 12</p></li><li><p>24</p><p>The data base plus the literature search turned up agreat variety of types of interactions in which intro-duced species facilitate one another, at least at the indi-vidual level. Although there are many idiosyncraticvarieties of facilitative interactions, most examples canbe broadly classified as follows:</p><p>Animals pollinating and dispersing plants</p><p>There is some evidence that introduced plants may alterpollination regimes for native plants (e.g., Butz Huryn1997), but the introduction of non-native pollinatorsdoes not have such a big effect. Nor is it clear that anewly introduced pollinator will automatically enhancethe reproduction of introduced plants it can pollinate.Although the introduced honey bee is a major polli-nator of such weeds as yellow star thistle (Centaureasolstitialis) (Barthell et al. 1994) and purple loosestrife(Lythrum salicaria) (Mal et al. 1992) in North Americaand barberry shrub (Berberis darwinii) in New Zealand(Butz Huryn 1997), these weeds would all be pollinatedby other insects in the absence of honey bees. Whetheror not the pollination regime is altered by the introduc-tion of honey bees and whether altered seed set andgene flo...</p></li></ul>