While invasion success is usually associated with the biological fitness of the invader and environmental similarity between the area of origin and the invaded range, some of the most notorious aquatic invasions involve species with traits that for millions of years were a burden, rather than an advantage, for their survival. These odd characteristics became major assets after man started reshaping the surface of the earth, facilitating their spread. Invasion risk-assessment models, chiefly based on environmental match parameters, are unlikely to predict the dispersion of these (and probably many other) species, whose invasive nature involves subtle and intricate mechanisms that operate at levels normally ignored by (and often unknown to) the models. Much of the literature on introduced species is focused on demonstrating their negative impacts on the ecosystems invaded. While the fact that invasive organisms can, and very often do, have enormous negative impacts, is beyond doubt, and all efforts possible for keeping biological invasions at bay should be made, once the introduction happens and the eradication of the invader is unfeasible, research efforts should be centered on objective analyses of how the invader interacts with the new ecosystem, untainted by efforts to forcibly demonstrate its negative impact.
Although the birth of invasion biology is usually associated with the publication of the book, The Ecology of Invasions by Animals and Plants (Elton, 1958), it was only 30 years later when introduced species started attracting the attention of scientists and the media, spurring an exponential growth in the literature on the subject (Lockwood et al., 2007; Ricciardi and MacIsaac, 2008). The importance of this field also fostered the creation of journals and the publication of books specifically dedicated to biological introductions (Driesche and Driesche, 2000; Mooney and Hobbs, 2000; Pimentel, 2002; Ruiz and Carlton, 2003; Sax et al., 2005; Cadotte et al., 2006; Lockwood et al., 2007; Davis, 2009; Rilov and Crooks, 2009; Thompson, 2014), establishing a new area of ecology – “Invasion Ecology.” The new paradigms involved promoted new hypotheses and concepts (e.g. the enemy-release hypothesis, propagule pressure, invasional meltdown: (Simberloff and Von Holle, 1999; Keane and Crawley, 2002; Lockwood et al., 2005), which became the leading subject of many surveys. However, attempts at integrating them into a distinct and cohesive body of knowledge, and especially their application to the prediction of future introductions and their probable impact, have had limited success (Davis, 2009).
While scientific work strives to derive general rules from observations based on limited sets of phenomena, and this inductive approach is the foundation on which much of what we know about our world was built, some areas of knowledge are intrinsically more amenable to this process than others. Unfortunately, “Invasion Ecology” is not among the former.
In this work, I try to offer new insights into two aspects of the ecology of introduced species. The first section is related to notions of the variables that govern the process of invasibility. The second section deals with the attitude of scientists when dealing with the purported impacts of biological invasions. The leading theme that emerges from these considerations is that, with the exception of some obvious mechanisms, general rules applicable to all introduced species seem far from “general,” suggesting that species- and environment-specific surveys cannot yet be adequately encompassed by many of the tenets of “Invasion Ecology.” This review relies heavily on what we know about freshwater introduced mussels, in particular the Mussel, Limnoperna fortunei; its significance beyond this scope may therefore be limited.
Invasibility and species/environment-specific traits
Limnoperna fortunei is a small (∼3 cm) freshwater Mussel, indigenous to the Pearl River basin in southern China. During the last few centuries, it spread throughout Indochina, presumably aided by human migrations (Morton and Dinesen, 2010). Its geographic range has been restricted to southern China for millions of years, and to southeastern Asia for centuries. Around 1960 it started spreading northwards and eastwards in China and Korea (Morton, 1975; Xu, 2015), and in 1990 it was recorded in Japan (Ito, 2015) and in Argentina (Pastorino et al., 1993). In South America its upstream expansion along the Paraná-Paraguay Rivers was impressively fast. By 2016, it was present in most of the Río de la Plata watershed, in several endorheic and coastal basins (Mar Chiquita, Guaíba, Patos-Mirim, Tramandaí; Oliveira et al., 2015), and in the São Francisco river basin (Barbosa et al., 2016), >4,000 km from its first sighting, covering Argentina, Bolivia, Brazil, Paraguay, and Uruguay. In most of the areas colonized, Limnoperna swiftly reached very high abundances, often exceeding 200,000 ind. m−2 (Correa et al., 2015).
One of the tenets of Invasion Ecology is that organisms better adapted for survival (e.g. abundant and widely distributed in their original range, with wide environmental tolerance, short generation times, rapid growth, high reproductive capacity, broad diet, high phenotypic plasticity) are more likely to become invasive (Ricciardi and Rasmussen, 1998; Richards et al., 2005; but see also Hayes and Barry, 2008). The expansion of Limnoperna suggests that this principle, which might seem quite obvious at first sight, is actually not as straightforward as it appears. Limnoperna has a short life-cycle, early sexual maturation, fast growth, presumably high fecundity, and ample environmental tolerance (Boltovskoy and Correa, 2015), traits which are shared by most animals that succeeded in colonizing fresh waters. In waters which are warm year round, its reproductive period extends for up to 9–10 months (Boltovskoy et al., 2009b), but if the temperature drops significantly in the winter, like in Korea and Japan, it is restricted to 1–2 months (Boltovskoy et al., 2015).
On the other hand, Limnoperna depicts two traits that are clearly undesirable in fresh waters: planktonic larvae and the possession of a byssus, which involves the need of hard substrate. Free-swimming planktonic larvae, typical for most marine benthic animals (Thorson, 1961; Pechenik, 1999), are a rarity among freshwater benthic invertebrates (Moss, 1988), which depict various strategies to avoid this stage because of its vulnerability to expatriation into the sea (Lopez, 1988). Among freshwater bivalves, planktonic larvae, probably a relic of their recent marine origin, are restricted to very few species; suggestively, three of these are conspicuous invaders worldwide: Dreissena rostriformis bugensis, D. polymorpha and L. fortunei.
The other trait that does not seem to be advantageous for colonizing freshwater habitats is the attached mode of life, derived from the possession of a byssus and the consequent requirement for hard substrate (although colonization of soft bottom areas has occasionally been noticed for L. fortunei and D. polymorpha, and is quite common for D. rostriformis bugensis, (Dermott and Munawar, 1993; Beekey et al., 2004; Karatayev et al., 2014; Nalepa et al., 2014; Pollick, 2014; Correa et al., 2015). In freshwater habitats, hard substrata are normally restricted to high-slope streams and the coastal fringe of lentic and lotic waterbodies, the rest of the bottom-surface being dominated by loose sediments. Limitations imposed by the scarcity of hard substrata are vividly illustrated by the fact that only ∼1% of North- and South American freshwater bivalves produce a byssus, whereas 99% live buried in the mud (McMahon and Bogan, 2001; Pereira et al., 2014).
The conflict between the presumed adaptive ubiquity of these species and their actual dispersion rates raises the question whether their invasion success is due to intrinsic traits associated with invasiveness, to environmental similarities between the areas of origin and those invaded, or to other factors unassociated with the above. As mentioned above, evidence on the adaptive ubiquity is mixed. For Dreissena spp., adaptive ubiquity is probably lower than for Limnoperna because it is significantly less tolerant to limiting conditions, in particular pollution, dissolved oxygen and calcium concentrations (Karatayev et al., 2007a). Assessments of environmental similarity are more complicated because the range of conditions of the waterbodies invaded by these Mussels varies widely. For example, while ecological conditions in southern Chinese lowland rivers are probably similar to many of those of the Río de la Plata watershed in South America, invaded by Limnoperna, in Korea this Mussel has succeeded in establishing permanent populations in waterbodies that freeze in the winter.
We speculate that neither ecological ubiquity, nor environmental similarity were key factors for the invasion success of these Mussels, which is largely due to the reshaping of the earth's surface by man. The major changes brought about by man that have been instrumental for the expansion of these species are (1) the increase in inter- and intra-basin connectivity through the building of canals for navigation and water distribution, (2) intra-basin and sea-borne inter-basin shipping, (3) overland intra-, inter-basin and upstream transport, (4) the construction of reservoirs and (5) man-made constructions in lakes, reservoirs and along rivers increasing the availability of hard substrata (Karatayev et al., 2007b; Boltovskoy, 2015a).
Both Dreissenids and Limnoperna have been restricted to small southern Eurasian basins for millions of years, and the start of their spread outside of their indigenous ranges is clearly associated with human activities. Dreissena polymorpha evolved around the Black and Caspian seas, and started expanding towards western Europe since the late 1700s through man-made canals for connecting the Black and Baltic Sea basins (Karatayev et al., 2007b). The expansion of D. rostriformis bugensis was somewhat slower and started later, but is also clearly associated with human activities, which were instrumental for the presence of both species in North America since the 1980s (Karatayev et al., 2007b). The indigenous range of Limnoperna is quite more remote than those of the Dreissenids, and man-made canals connecting its indigenous and invasive areas are so far restricted to China (Morton, 1975; Xu, 2015; Zhan et al., 2015). Like the Dreissenids, its introduction to the Americas is associated with sea-borne transportation, most probably in the ballast water of transoceanic vessels. While Dreissena spp., first introduced in the Great Lakes, started their spread downstream the Mississippi River (Benson, 2014), Limnoperna expanded upstream the Paraná and Uruguay Rivers. Interestingly, a rough comparison of their rates of geographic expansion yields similar numbers: around 400 km y−1 (Karatayev et al., 2015). While fast downstream spread was obviously aided by Dreissena's planktonic larvae, unaided upstream expansion of freshwater bivalves is only 0.1–10 km y−1 (Kappes and Haase, 2012), which indicates that the attachment of adult Mussels to northbound cargo vessels along the Paraná-Paraguay waterway was the most likely mechanism by which they reached the upper sections of this system so fast (Boltovskoy et al., 2006). Thus, while increasing inter-basin connectivity may favor the spread of most aquatic species (Minchin and Gollasch, 2002), upstream shipping selectively promotes the expansion of the few that can attach to watercraft hulls, while downstream dispersal selectively favors benthic organisms that produce planktonic larvae.
As opposed to the majority of freshwater organisms, some molluscs can spend extended periods of time outside of the water. Bivalves, in particular, endure desiccation by tightly closing their shells, decreasing metabolic rates, and retarding the loss of water. D. polymorpha and L. fortunei can survive in air for up to ∼10 days (Ricciardi et al., 1995; Montalto, 2015); this resilience, and the ability to attach to hard substrata, allows them to be transported overland between waterbodies on the hulls or entangled macrophytes of trailered leisure watercraft (Padilla et al., 1996; Johnson et al., 2001; Collas et al., 2017). Incidentally, this mechanism is most probably responsible for the colonization of the inland Argentine Mar Chiquita watershed by Limnoperna (Boltovskoy et al., 2006).
Construction of reservoirs is yet another major feature which has facilitated spread and enhanced survival of these invasive Mussels, especially in short, non-navigable rivers. Reservoirs act as refugia and seeding spots in water systems where otherwise the extinction of benthic species with a planktonic larval stage is facilitated by downstream flushing. Coastal and floating man-made structures are also likely important seeding spots and stepping stones for the Mussels in lowland rivers where natural hard substrata are scarce.
These considerations raise the question whether there are significant rules applicable to biological invasions in general, or if, as noticed by Davis (2009), invasibility should be interpreted only in the context of a particular species and a particular environment. The above review suggests that the combination of factors that made these freshwater Mussels so successful invaders has little in common with those that facilitated the invasion of the comb jelly Mnemiopsis in the Back Sea (Kideys, 2002), or the marine alga Undaria in Patagonia (Irigoyen et al., 2010), or the carp and the sparrow worldwide. Admittedly, some overarching themes do exist, such as environmental match, or propagule pressure, but even these widely accepted and rather obvious principles have pitfalls when inspected from the perspective of different invaders. For example, several models have been proposed and applied for the prediction of the introduction and spread of invasive species, most of which use environmental matching as input variables (GARP, Stockwell, 1999; MaxEnt, Phillips et al., 2006). These models have been used for predicting the worldwide spread of Limnoperna (Kluza and Mc Nyset, 2005; Campos et al., 2014), concluding that it could colonize areas along the western coasts of South America (among others). Although these waterbodies share many physical and chemical properties with those presently occupied by the Mussel, and even assuming a high propagule presure (from ballast water discharges, for example), the spread of Limnoperna into these waterbodies is improbable because areas to the west of the Andes mountain chain have short, high-slope, fast-running rivers which are unlikely to host permanent, self-sustaining populations. Again, predictions of spread of different potentially invasive species must take into account quite different environmental variables in order to produce meaningful results. For Limnoperna, calcium, pH, and, to a lesser degree, water temperature, seem to be marginally significant, but river length and navigability, the presence of lakes and reservoirs, and that of freshwater ports that operate with transoceanic traffic, are likely very important. Sediment load, probably of minor importance for infaunal and deposit-feeding species, is a major limiting factor for suspension feeders, including Dreissena and Limnoperna (Lei et al., 1996; Tokumon et al., 2015).
All invasives are harmful. Or perhaps not all and not always?
An interesting debate appeared in Nature a few years ago. In June 2011 Mark Davis published an article entitled “Don't judge species on their origins” (Davis et al., 2011), where he advocated for a more balanced and objective outlook on the impacts of introduced species. A few months later, a reply was published in the same journal where 141 scientists endorsed Daniel Simberloff's response rejecting Davis' stand, essentially arguing that introduced species bring about more harm than benefit, and should therefore be treated differently (Simberloff and signatories, 2011; see also Richardson and Ricciardi, 2013). In the section that follows I will attempt to summarize our current knowledge of the impacts of Limnoperna in South America in the light of this provocative, and still ongoing, discussion.
Limnoperna and Dreissena are known as “ecosystem engineers,” meaning that their presence has wide-ranging influences, affecting the water column communities (abio- and bioseston, including the plankton, and fishes), and the benthos (Karatayev et al., 2002). Since its introduction in Japan and South America, several aspects of the interactions between Limnoperna and other organisms have been investigated. Some of the most salient results are outlined below.
Like all filter-feeding Mussels, Limnoperna filters out organic particles from the water column and produces mucus-embedded feces and pseudofeces that are deposited on the bottom. Studies carried out in an Argentine reservoir since 1996, which was invaded by Limnoperna around 2001, show that, after having been colonized by the Mussel, particulate organic matter in the water column dropped from 4 g m−3 (in 1996–2003) to 2.4 g m−3 (2004–2008), penetration of 1% of incident light increased from 7 to 10 m, phytoplanktonic production decreased from 33 to 17 mg C m−3 h, and the reservoir was populated by submerged macrophytes (Elodea callitrichoides) and waterfowl (Fulica spp.) (Boltovskoy et al., 2009a).
Several surveys in the lower delta of the Paraná River comparing natural and artificial substrata and soft-bottom areas influenced by the presence of Limnoperna with those barren of the Mussel invariably show that the former host higher invertebrate numbers, biomass and diversity (Sardiña et al., 2011; Sylvester and Sardiña, 2015; but see also Linares et al., 2017).
Since its establishment in the Río de la Plata watershed, veligers of Limnoperna became a common (and often dominant) component of the diet of local fish larvae. At least 18 larval fish species were observed to feed on the veligers, and laboratory studies showed that veliger-rich diets promote faster growth than diets restricted to the indigenous prey (rotifers, cladocerans and copepods) (Paolucci and Thuesen, 2015).
As of 2015, >50 adult fishes were observed to consume adult Limnoperna, in some cases representing up to 100% of the diet. Some of these species are major components of the diet of larger, piscivorous species, which suggests that the availability of this new prey item may cascade up the trophic chain favoring these larger predatory species as well (Cataldo, 2015). A conspicuous increase in Argentine freshwater fish landings, around 1995, has been tentatively associated with the introduction of Limnoperna in the Río de la Plata basin (in 1990) (Boltovskoy et al., 2006).
While the long-term balance of these changes is unpredictable, so far most of them can hardly be qualified as negative. On the other hand, 400 L mesocosm experiments carried out in a large South American reservoir of the Uruguay River (Salto Grande), clearly showed that the presence of Limnoperna strongly enhances the densities of the cyanobacteria Microcystis spp. (Cataldo et al., 2012), often toxigenic and a pervasive problem of lentic tropical to temperate waterbodies worldwide (Huisman et al., 2005). Obviously, this is a clearly negative impact.
The assessment of negative vs. positive effects is complicated by the fact that most of the impacts of a species on the community are rarely unidirectional, vary geographically, seasonally, inter-annually, and with the time elapsed after introduction. Dreissena and Limnoperna provide vivid examples of impacts that may operate in opposite directions. For example, grazing of phytoplankton tends to decrease algal densities, but the concomitant increase in water transparency and the supply of nutrients can enhance phytoplankton densities (Figure 1). Predation on zooplankton can decrease densities of these animals, especially rotifers and cladocerans (Rojas Molina et al., 2015), but the facilitation of macrophyte growth due to a deeper photic layer, less competition for nutrients with the phytoplankton, and enhanced nutrient regeneration, provides more shelter for the zooplankton from its predators. The supply of organic-rich deposits from the Mussels' feces and pseudofeces improves the feeding quality of the sediments, but their decomposition lowers dissolved oxygen levels in the near-bottom layers (Figure 1).
Interestingly, the intricacy and ambiguity of the environmental effects of Limnoperna are not restricted to impacts on other communities, but may also significantly affect the Mussel itself. Nutrient recycling by the Mussel and modification of P:N ratios, associated with production of chemical cues, have been shown to boost the growth of the toxic Cyanobacteria Microcystis and enhance the proportion of its colonial forms over isolated cells. Paradoxically, Microcystis' toxin (microcystin L-R) is highly toxic to Limnoperna's larvae, effectively suppressing recruitment at times when reproduction is highest in waterbodies where Microcystis does not bloom (Boltovskoy et al., 2013; see Figure 1).
As of 2016, ∼700 publications have been produced dealing specifically with L. fortunei, its biology, ecology, and its effects on the environments invaded. Unfortunately, much of this literature is based on the premise that Limnoperna's impact on the ecosystem is negative, and investigations strive to demonstrate this initial hypothesis, often based on isolated anecdotal observations and little or no supporting evidence. In many cases conclusions are based on the reported effects of Dreissena in the northern hemisphere, assuming that similarities between these species guarantee that their impacts are identical. While this is true for some of the impacts, species- and ecosystem-specific differences determine some very dissimilar effects. For example, in North America enhancement of Microcystis spp. by D. polymorpha is restricted to lakes with low to moderate total phosphorus concentrations (<25 µg total P l−1) (Knoll et al., 2008), whereas in South America strong enhancements have been observed at P levels above 100 µg l−1 (Cataldo et al., 2012). Although field studies in Japan and in Argentina showed that >90% of the Mussel's production is consumed by predators (presumably mainly fishes) (Sylvester et al., 2007; Nakano et al., 2010), in their strive to blame the Mussel for presumed or proven impacts literature on Limnoperna is plagued with remarks that its invasive nature in South America is due to the absence of local predators (Bujes et al., 2007; Terra et al., 2007). Furthermore, the introduction of Limnoperna, which is restricted to fresh and intermittently brackish waters, has been included in the array of impacts associated with the depletion of native exploited species, like Mytilus edulis platensis (the Blue Mussel), which is strictly marine (Defeo et al., 2013).
Literature on Limnoperna soared from ∼0.4 publications per year before 1990, ∼9 per year in 1991–2000, and 39 in 2001–2016, in part due to its geographic spread, and largely because it became a “hot” topic since it started interfering with the operation of power plants (an undeniable negative effect), and its colonies became a salient feature of coastal freshwater areas. This boom spurred a wealth of new data, much of it focused on detecting purportedly negative impacts, often basing their conclusions on the described impacts of D. polymorpha on European and North American waterbodies, rather than on their own observations. Some were cautious in noting that these effects were merely a possibility, but they had a snowball effect whereby subsequent literature used them as proof of the putative impacts (Boltovskoy, 2015b). As noticed by Byers et al. (2002, “… because positive results are more likely to be submitted and published, the invasion literature may be biased toward demonstrating that nonindigenous species have large ecological impacts,” and the same bias seems to permeate grant submissions, conference reports, newspaper and magazine articles, consultant reports, thesis dissertations, book chapters, web pages, etc. where the importance of impacts is exaggerated in the hopes of getting funding or recognition.
Objective studies have shown that introduced consumers are more likely to have greater impacts on the resources available than indigenous ones (Paolucci et al., 2013). However, the invasive vs. indigenous dichotomy should not be equated to harmful vs. non-harmful before proven otherwise. Furthermore, even when the impact of the invasive species is objectively shown to be greater than that of the indigenous one, it does not necessarily imply that the impact is negative. Spartina alterniflora, the Smooth Cordgrass, which dominates much of the coastal marshes of the Atlantic coasts of the Americas, has been the subject of hundreds of surveys concluding that the plant supports rich and diverse communities, and that these systems should be protected from destruction by human activities, in particular grazing by cattle. Recently, Bortolus et al. (2015) suggested that Spartina is not indigenous, but was introduced before the 1800s, and that what are now extensive S. alterniflora marshes were probably bare mudflats in the not so distant past. Interestingly, while the paper does not analyze whether the changes brought about by the introduction of the smooth cordgrass have been positive or negative, they state that these changes were “catastrophic,” altering the pristine state of nature.
Literature on species invasions has provided overwhelming evidence of the harm that biological introductions can bring about, which leaves little doubt that all efforts possible for keeping biological invasions at bay should be made. However, if our efforts fail and a new species makes its way, and especially if it manages to spread and establish beyond reasonable possibility of eradication, investigation of its interactions with the ecosystem invaded should neither be forcefully oriented at finding its negative impacts, nor, of course, at demonstrating that it is innocuous or beneficial (Boltovskoy, 2015b). The “invasive species approach” tends to lump all introduced species in the same bag thus focusing the assessment of their interactions on the traits that they share with other invasives (usually perceived as negative), rather than on those that characterize them best.
Comments by two anonymous reviewers on an earlier version are greatly appreciated.
Funding for this work was provided by the Consejo Nacional de Invstigaciones Científicas y Técnicas (CONICET, Argentina) and by the Agencia Nacional de Promoción Científica y Tecnológica (ANPCyT, Argentina).