Few events are as important in predicting the future role of a nonindigenous plant, arthropod, or pathogen as its attainment of the population size at which it rapidly adds members and spreads simultaneously into a new range (Elton 1958). In this demographic and geographic transition, the species is no longer likely to go extinct through stochastic forces—losses to environmental stochasticity are all within limits that can be absorbed by the population. The population’s basic reproduction number (Ro) is greater than 1, and its deterministic growth rate (λ) will be increasing from 1; hence, the population doubling time will be decreasing. The population is not simply tolerating its new range; the environment of the new range is clearly well within the species’ ecological amplitude. Furthermore, the totality of its traits contributes to its survival in the new range and to its proliferation. In the terms used in this report, the species has become a biotic invader.
An issue in this report is why some species make a transition from low to high rates of population growth, that is, shift from being simply persistent or surviving to becoming invasive. An explicit explanation is needed, one ideally based on the attributes of the species combined with its circumstances in the new range. From the case histories of individual invasions, it is apparent that the transition in each case has been propelled by a unique and accidental chain of events in which community dynamics and the resources of the environment interact with the biological characteristics of an established species in a way that fosters its proliferation (Crawley 1989a). At each turn, these events enable other traits to reinforce the growth and spread of the population. The nexus of the community, its environment, and the size of the population is crucial, but so little
is known quantitatively about many of these interactions that using them in a predictive manner remains difficult.
Among the factors that influence an immigrant’s transition from persistence to invasion are the same abiotic and biotic forces that are faced by organisms during establishment. For example, the establishment of nonindigenous arthropods has often been attributed to their escape from natural enemies in their native habitat. However, knowing whether resident enemies in a newly colonized habitat will attack a nonindigenous arthropod and limit its demographic growth or spread is important if we wish to predict which immigrants will not only establish, but also proliferate and spread.
Cornell and Hawkins (1995) and Hawkins et al. (1997) compiled life tables for herbivorous insects and examined patterns in mortality caused by natural enemies to determine whether established nonindigenous species sustained lower mortality from resident natural enemies than did natives. In the first report, Cornell and Hawkins (1995) found that the invasion status of the herbivores was only weakly related to the mortality caused by natural enemies. Sources of mortality differed most strongly between early-stage and late-stage herbivore larvae and between insects that feed inside plants (endophytic) and insects that feed externally (exophytic). Hawkins et al. (1997) used slightly different methods to differentiate among causes of mortality in 63 resident and 20 invading species. Nonindigenous insects did not sustain lower overall mortality than natives from resident natural enemies, but larvae and pupae of nonindigenous species experienced more predation mortality and perhaps more pathogen mortality than natives, although differences were not statistically significant. It must be noted, however, that the analysis relied on data collected in life-table studies of economically important insect species; this suggests that the nonindigenous species included in the analysis had successfully made the transition from establishment to proliferation and spread. Knowing whether interactions between nonindigenous insects and native natural enemies are similar earlier in the invasion process will require additional research.
Evidence from existing studies, however, could be useful in developing testable predictions about the effects of native natural enemies on nonindigenous insects that have become established in the United States. For example, we might expect that nonindigenous insects that feed externally on plant foliage, where they are exposed, will be more likely to acquire a complement of native predators than insects with more protected or specialized feeding behaviors. We might also predict that native parasitoids that are habitat-specific, rather than host-specific, could eventually become an important cause of mortality among nonindigenous insects, especially those confined within leaf or phloem tissue. In a review of cases involving insects imported for weed control, endophytic herbivores were the group most likely to experience mortality from native parasitoids (Goeden and Louda 1976). Native endophytic leaf-miners sustained the highest mortality from parasitoids in the analysis by Hawkins et al. (1997). Bright (1996) noted
that the ability of native parasitoids to successfully attack larvae of the larger European pine shoot borer (Tomicus piniperda L.) in pine logs was not surprising, given that the parasitoids are habitat-specific rather than host-specific. Furthermore, Hawkins and Gross (1992) showed that parasitoids reduced the density of nonindigenous herbivorous insects more often if the native habitat of the herbivores supported many parasitoid species than if it supported few such species. Thus, information about the diversity, life history, and importance of natural enemies in the native habitat will be valuable in estimating the potential effects of the resident enemies on the dynamics of newly established nonindigenous insects.
The longer a population is established and is expanding into a new range, the more likely it is to come into contact with a greater diversity of conditions and organisms (Strong et al. 1984). Over the longer term, in addition to the factors discussed in Chapter 3, a species’ invasion will be influenced by factors that may have been less prominent during establishment. This chapter discusses three processes that occur in an invasion: dispersal and spread, competition for resources, and evolution. The few elements of these processes that we know about offer clues to predicting invasiveness, but this information is incomplete, and the gaps in our knowledge suggest potentially productive lines of research.
DISPERSAL AND SPREAD
Once an immigrant population has arrived, it will become a successful invader only if the population is able to increase in abundance and spread from its point of entry. Spread of a nonindigenous species in a new range is determined by several components, including the number of propagules available for dispersal, the opportunities for and mechanisms of dispersal, the communities into which the organism spreads, and the availability of suitable hosts, nutrients, and other resources in the habitat. The difficulty of predicting the rate of population growth and spread is often compounded by interactions between life-history traits of the organism and characteristics of the environment.
Expansion of an immigrant population consists of three steps: an initial establishment phase with little or no expansion, an expansion phase, and a saturation phase (Shigesada and Kawasaki 1997). Shigesada and Kawasaki (1997) describe three situations that can initiate the beginning of the expansion phase. In some species, expansion begins only after the territory occupied by the initial invading population becomes filled. When the population is subject to an Allee effect, the time for the initial population density to reach this level will be increased (Lewis and Kareiva 1993). In other species, expansion begins only after the occurrence of a favorable mutation in the colonizing population. In other words, expansion is triggered by a genetic adjustment to the new habitat that leads to higher fecundity, survival, improved dispersal ability, or some other significant trait (Bazzaz 1986). In the third scenario, a few individuals that were initially released into a single area disperse rapidly. Range expansion becomes
evident only when reproduction enables these populations to increase above a detection threshold.
The rate at which expansion proceeds can be assessed by evaluating the relationship between range distance and time. According to Shigesada and Kawasaki (1997), the shape of the curve relating range distance and time can assume three general forms. The range of a population with a type 1 curve expands linearly with time. Offspring typically settle in the neighborhood of the parent population, and dispersal occurs primarily over short distances. Shigesada and Kawasaki (1997) suggest that natural expansion of gypsy moth populations, which occurs when recently hatched larvae balloon on the wind, may exemplify a type 1 curve.
In contrast, species with a type 2 or type 3 curve are able to expand their range not only by random movement into surrounding adjacent areas, but also by long-distance dispersal, a process referred to as stratified diffusion (Andow et al. 1993, Hengeveld 1989). These populations will demonstrate an accelerating range-versus-time curve. A population with a type 2 curve initially expands slowly, then linearly at a higher rate. Short-distance migration expands the occupied area from its periphery; long-distance migrants generate new satellite colonies that expand in isolation for a short period until they coalesce with the parent population. In a population with a type 3 curve, the rate of spread increases linearly with time, and the curve assumes a convex shape. In this scenario, long-distance migrants establish satellite colonies; in the long run, the range of these colonies expands independently of the parent population and other colonies. Hengeveld (1989), however, noted that an accelerating range-versus-time curve could also reflect genetic adjustments early in the expansion phase.
In some plant-pathogen populations, movement of propagules away from the source population follows a dispersal gradient that reflects the population’s net reproductive rate and the dispersal distances of the propagules. When spores settle rapidly from the air, the distribution of propagules will be exponential and the expansion rate will be roughly linear with time. In such a population, the radius of the focus increases linearly with time (Shigesada and Kawasaki 1997, van den Bosch et al. 1990), and new foci developing ahead of the wave front will be subsumed by the front (for example, a type 1 curve). However, if the distribution of propagules is not exponentially bounded, propagules are disseminated further from the focus. If new foci are not subsumed within the much larger foci of the invasion, the average population expansion rate increases with distance from the original focus (Shaw 1995). Growing evidence indicates that the latter type of distribution occurs commonly in many plant-pathogen species (Ferrandino 1993), and this type of distribution apparently operates among different taxonomic groups. Establishment of multiple foci accelerates the spread of a population. Moody and Mack (1988) showed that spatially growing foci expanded in area as a strict function of their circumference-to-diameter ratio. Thus an area could be occupied by an invading species much sooner if the species is distributed as an
array of widely separated foci than if the composite area occupied by foci is expanding from a much larger, single focus, unless they are expanding at a markedly lower rate (Moody and Mack 1988).
The relationship between the number of foci and the dispersal distribution provides one explanation of the change in the rate of expansion of a nonindigenous species from the so-called lag phase to the log phase. Initially, the immigrants and their immediate descendants are confined to a small area, a “beachhead” that consists of a single focus in the new range. Propagules of this population may be carried well beyond the beachhead, and some fraction is able to establish new foci. These foci are a threshold or minimum that precipitate the log phase of range expansion; they are both expanding and creating new foci, and range expansion accelerates markedly. The emergence of the pronounced range expansion depends, of course, on multiple demographic and biotic factors (such as competition, predation, parasitism, and resources), not the least of which is sufficient distance between nascent foci for them to add individuals to the unoccupied range. Convergence among foci would create the potential for intraspecific competition, which could partly explain why a population expanding spatially from a single large focus occupies space so much more slowly; many offspring land in an area that is already occupied, as opposed to being carried far from the population boundary. The transition from a species that is persistent or merely surviving in a new range to an invasive species might simply represent the time and opportunity for a nonindigenous species to be transported to some minimal number of isolated locales where new foci establish and expand. The minimum is probably unique to each species and perhaps even to each combination of locales at which a given species arrives.
Life history, morphology, and behavioral traits related to dispersal of a newly established species obviously play important roles in determining the rate of range expansion (Hastings 1996), and knowledge of such characteristics would be useful in predicting the likelihood of invasion. The process of dispersal facilitates cued responses to environmental variables or “bet-hedging” by species in unpredictable or harsh environments (Levin 1989). Dispersal can reduce dependence of populations on localized climatic variables, enabling the presence or abundance of an invader at a specific location to be influenced by spatially disparate populations (Davis et al. 1998). Source-sink dynamics enable invader populations to be sustained in less than optimal conditions through frequent or continued immigrant dispersal from source populations (Davis et al. 1998, Pulliam 1998). Emigration and immigration among colonies can counteract environmental stochasticity, however, only if the individual colonies are exposed to independent environmental events and if they are close enough for dispersal to occur (Lewis and Kareiva 1993, Stacy and Taper 1992).
Populations of numerous plants, insects and plant pathogens include forms or behaviors that lead to both short- and long-distance dispersal, consistent with the type 2 or 3 curves described by Shigesada and Kawasaki (1997). In some
moth and locust species, for example, long-distance dispersers become more common as population density increases. Long-distance dispersal can also occur when plants’ propagules are carried by insects, wind, or water or by birds or mammals. In addition, our ability to detect a newly established population is typically limited when the density of a population is very low. The population must exceed a threshold density before it can be detected; this threshold will depend on traits or behavior of the organism, including the extent of the damage it causes.
Dispersal information about a species is often anecdotal and usually lacks quantification (Ridley 1930). Comparisons of biotic and abiotic dispersal among woody invaders and noninvaders suggested that the dispersal mechanisms of a species were not predictive of its likelihood of becoming invasive (Reichard and Hamilton 1997). Data from mark-recapture studies of 12 plant-feeding insects indicated that variation in diffusion was considerable among ecologically similar species and even within the same species (Kareiva 1983). In addition, inadvertent transportation of a nonindigenous organism by humans can establish new foci at substantially greater distances than would occur by natural dispersal mechanisms of the species. Such transportation has been shown to have substantially increased the spread rate of such species as gypsy moth (Liebhold et al. 1992) and cereal leaf beetle (Andow et al. 1993). Differences in dispersal, however, have been recognized among plant community types. Wind dispersal, for example, is more common among arid treeless ecosystems, and bird dispersal is more common in forest systems. Special attention should be given to the detection of newly established species with large, fleshy fruits in habitats that support an array of frugivorous birds. Such birds have contributed substantially to the spread of naturalized species, such as Clidemia hirta (Koster’s curse) and Hedychium gardnerianum (kahili ginger), in Hawaii (Cuddihy and Stone 1990).
The economic and ecological importance of invasive species has given rise to numerous models that seek to describe how invading species spread and increase in abundance (Liebhold et al. 1992). In a simple diffusion model, the range of an immigrant expands solely by diffusion without population growth. In other words, the population expands in a concentric manner around the origin, and the density of the population decreases rapidly away from the origin.
Logistic expansion occurs when a population spreads by growth alone without diffusion (Skellam 1951). In that case, the rate of spread will depend on the reproductive rate of the organism and the degree of competition it encounters. If competition does not occur, the relationship between rate, distance, and time approaches a Malthusian curve. When competition increases with density, however, a logistic curve results because growth slows as population density increases. The spread rate becomes asymptotic as the population approaches the carrying capacity of the habitat.
Perhaps the most widely used model to describe the spread of an invader is a growth-diffusion model that links an exponential population growth term with a
diffusion model (Fisher 1937, Liebhold et al. 1992, Skellam 1951). Range expands because of the combined effects of diffusion and growth, and the change in population density over time is approximated by an asymptotic rate of spread. The advancing front of the population will appear to advance at a constant velocity and shape–a wave front. As described by Shigesada and Kawasaki (1997), such a scenario occurs when the density of the initial colonists invading the origin decreases rapidly because of diffusion and there is little competition. As the population gradually recovers through reproduction, competition effects start to be manifested, so the density in the occupied range approaches the carrying capacity of the habitat. The range front forms a sigmoidal pattern and spreads at a constant speed–a traveling frontal wave. At the edge of the expanding population, competition is low, so expansion by diffusion predominates and density-dependent effects on growth rate are insignificant. Despite its relative simplicity, the growth-diffusion model has been used successfully to simulate the spread of a variety of organisms (Hastings 1996, Levin 1989, Liebhold et al. 1992, Long 1977, Willamson and Brown 1986).
More-sophisticated models with additional variables may be needed when an organism is transported by wind or water or tends to orient to a stimulus. For example, in a convection model, a velocity term is applied to the reaction-diffusion equation, which may better describe an insect, plant, or plant-pathogen species that is transported by wind (Lewis and Kareiva 1993). Insects may select their flight direction or respond to host or conspecific volatiles, whereas the spread of seed or pollen may be affected such variables as settling velocity, height of release, wind speed, and turbulence (Okubo and Levin 1989). Spread of populations can be linear or streamlike if dispersing propagules are channeled by topographic or climatic factors (Carey 1996).
If the population is influenced by an Allee effect, success or failure of an invasion may be determined before the asymptotic form of expansion is achieved (Lewis and Kareiva 1993). Presence of an Allee effect or an increase in the diffusion coefficient will require the initial “beachhead” to be larger if the population is to avoid extinction. Even if an immigrant population initially arrives in a new area at a density above its critical Allee effect threshold, an Allee effect will increase the period before expansion begins (Veit and Lewis 1996) and can substantially reduce the traveling-wave speed, resulting in a lower asymptotic rate of spread. Presence of an Allee effect will affect the curvature of the boundary between invaded and uninvaded regions and affect the rate of spread. When “fingers” of an uninvaded area protrude into the invaded area, the population will spread faster than when occupied territory curves sharply into unoccupied territory (Lewis and Kareiva 1993).
Environmental heterogeneity—for example, the patchiness of the environment—can also influence the rate of spread. Diffusion coefficients may vary widely, depending on the degree of environmental, habitat, or resource heterogeneity and how the nonindigenous species responds to the heterogeneity. When
the habitat is not uniform, the expansion rate of a population will be affected by the spatial distribution of favorable and unfavorable patches (Shigesada and Kawasaki 1997). Environmental variables interact with biological traits of an organism, and this increases the difficulty of predicting invasion rates. For example, the spread of Scotch broom (Cytisus scoparius) from multiple foci was lower in urban areas than in prairies, primarily because survival and reproductive rates were lower in the less favorable urban habitat (Parker 2000).
Predicting the potential of populations of nonindigenous plants and plant pests to grow and spread will depend in part on what can be determined about the dispersal mechanisms and circumstances of organisms in any range, whether native or new. Studies are needed to quantify the role of the various factors— such as wind, birds, natural migrations, and human-mediated actions—that contribute to long-distance transport of individual species. Continued research into and documentation of the spatial and temporal aspects of known invasions will be necessary if we are to improve our understanding of patterns of range expansion and of the mechanisms by which invasions progress.
COMPETITION FOR RESOURCES
Community Diversity and Resources
Community diversity is viewed as the factor that most influences a community’s susceptibility to invasion by nonindigenous species. Different theories (but few experimental studies) support opposing views of the role of diversity in invasion (Levine and D’Antonio 1999, Dukes 2001, Naeem et al. 2000, Stohlgren et al. 1999). Levine and D’Antonio (1999) argue that natural plant communities with high biodiversity tend to be more prone to invasion than their species-poor counterparts because the factors that control native diversity also control diversity among invaders. A supplemental argument is that the “microheterogeneity” of diverse communities provides differences in the spatial context (such as canopy height and rooting depth) that in turn provide more opportunities for immigrant species than are found in monocultures.
Those arguments are in contrast with the concept of “limited resources space”, which suggests that the more species that occupy an area, the more fully resources are used and the more difficult it is for a new species to become established (Tilman 1987, 1988 and references therein, Knops et al. 1999). In addition, there are many cases in which nonindigenous species have become established in species-poor communities.
Indirect interactions among species can also facilitate invasiveness, regardless of diversity. Some species in a community can deter invasion; others can facilitate it (Palmer and Maurer 1997). Interspecific competition initially appeared to be responsible for the declining numbers of a native leafhopper after establishment by a nonindigenous leafhopper (Settle and Wilson 1990). Field and caging
experiments showed that neither species was competitively superior; mortality caused by interspecific competition was equal to mortality caused by intraspecific competition. The two species did, however, share a parasitoid enemy, and the native leafhopper experienced higher attack rates by this parasitoid than did the nonindigenous leafhopper. Differential parasitism apparently shifted the competitive balance so that the native species was at a disadvantage compared with the nonindigenous species. Other studies have similarly shown that the establishment of a nonindigenous insect can be affected by complex interactions involving both potential competitors and resident natural enemies (Davis et al. 1998, Muller and Godfray 1997). Discussions about diversity and competition within communities often refer to the likelihood of invasion as depending on the availability of “vacant niches” for colonization. Observations that some island communities with simplified structures or impoverished floras are particularly susceptible to invasion have been used to support the possibility that they contain vacant niches (Elton 1958). As noted in Chapter 3, other explanations are possible, including the greater numbers of mainland species and the greater opportunities for them to be introduced to islands (Simberloff 1995).
The niche concept in general and the vacant-niche idea in particular are based largely on classical ecological theory that emphasizes the importance of interspecific competition in structuring communities. The validity of the niche concept in explaining community structure has been questioned, in part because the paradigm of competition as the primary organizing force in communities has been challenged (Simberloff, Connor, Strong, Wiens, etc). Crawley (1987) suggests that the confirmation that a niche is vacant is difficult without knowledge of the other factors that can render those resources unexploitable.
It would be more fruitful to focus on resource availability and how variation in limiting resources may foster or hinder the invasion process (Davis et al. 2000). Invasions of communities probably include both cases in which resources are usurped by an invader at the expense of natives and cases in which unused or underused resources are commandeered. Centaurea solstitialis (yellowstar thistle), for example, is able thrive with minimal water in the annual grasslands in California’s Central Valley (Gerlach et al. 1998). Invasive nitrogen-fixing species have proliferated in habitats where there apparently were few native nitrogen-fixers or where native nitrogen-fixers operated weakly (Vitousek and Walker 1989).
A potentially productive approach to predicting invaders is examination of the functional groups in communities, regardless of the taxonomic groups involved. Community ecologists have long noted that some species appear to play similar roles in a community, for example, tree species that occupy different levels in the forest canopy at maturity, and diverse cryptogams that collectively form a crust-like layer in grasslands that are dominated by perennial grasses (Evans and Johansen 1999). They have also noted that some communities lack occupants in apparent roles. In New Zealand, for example, some of the Nothofagus forests in
mesic environments have extraordinarily low species richness (Wiser et al. 1998); the forest understory can consist of fewer than 12 species. It is unlikely that the native resident species collectively play all the roles (and to the same degree) of light and nutrient capture as are played by the scores of species in Northern Hemisphere forests in similar physical environments. The testable question that emerges is whether the low species richness of the native New Zealand forests has led to their being invaded more than, for example, forests with similar physical environments in North America or Europe. The many examples of nonindigenous species that have become invasive with little or no continued assistance by humans suggest that such opportunities do exist.
Disturbance and Resources
Disturbance can alter the availability and use of resources. In general, ecological communities modified by human or exogenous natural events, such as hurricanes and cyclones, to a condition that is not otherwise common are considered more susceptible to invasion. The almost ubiquitous occurrence of ruderal (usually nonindigenous) species along roadways, footpaths, and pastures—sites of routine disturbance—is forceful testimony to this principle. In addition, disturbances that remove biomass and recycle limited resources create opportunities for colonization and occupancy by nonindigenous species, such as ragwort Senecio jacobaea (McEvoy et al. 1993, McEvoy and Coombs 1999) in Oregon. So strong is the link between disturbance and the proliferation of agricultural invaders that it becomes necessary to separate these cases categorically from those in which disturbance is minor or even nil. The degree to which plant invasions in natural communities have been not only sparked but also sustained by continual disturbance deserves further investigation (Mack 1989).
Competitive Traits and Resources
Once established, some invaders can usurp resources in a new range, use them more efficiently than the natives, or even alter the resources themselves. In many invasive species, these competitive traits are revealed as the organism begins to proliferate. Evidence of interspecific competition among newly established insects or pathogens is scarce, but competition often plays a substantial role in plant invasions. In many plant communities, competition for light is so intense that a nonindigenous species’ ability to invade could be related directly to its ability to capture light that could otherwise be used by native species. So the “upmanship” or growth in stature that has arisen through natural selection since plants emerged on land is illustrated in some plant invasions. In its simplest terms, nonindigenous plants that can overtop their neighbors can replace them in the canopy, especially if they reduce light below the light-compensation point for any stage in the natives’ life cycle. In some cases, climbing nonindigenous vines
have reached the top level of the native tree canopy and altered light to the detriment of all species below them. Their influence, although often limited to the edge of the forest, can nevertheless severely affect species diversity locally. A classic example in the United States is Pueraria lobata (kudzu), but it is not unique, and apparently other naturalized climbing species are becoming invasive, such as Paederia foetida (skunk vine) in Florida (Schmitz et al. 1997) and Bryonia alba in the Pacific Northwest (Novak and Mack 1995). Thus, any nonindigenous species that can form a denser shade in a community than is formed under the native species could potentially become invasive. Competition for light is so intense among many plants in a wide array of community types in the United States (such as most forests, most perennial grasslands, and many wetlands communities) that special attention should be paid to the light-sequestering ability of any plant immigrants.
Some invasive plants have better-developed root systems (denser, more finely divided, or deeper) or more-efficient root systems than do natives. Tamarix ramossisima, which has a high rate of transpiration, has roots that extend far into the soil profile. This species’ roots withdraw water from lower levels than do the roots of the native species (Horton et al. 2001a,b). Nonindigenous plants that produce more-abundant nectar or have more-abundant fruit than native species outcompete natives for visits from pollinators, and this results in reduced seed production among the natives (Sallabanks 1993, Brown and Mitchell 2001).
In addition to vying for resources that native plants require, invasive plants can alter the environment. The introduction of salt on the soil surface in the case of Tamarix or of nitrogen in the soil in the case of root-modulating plants can alter species composition and even succession (Vitousek 1986). An increase in the production of combustible fuel is perhaps the best-known environmental alteration caused by some invasive species. Some nonindigenous plants facilitate fire or alter the frequency or intensity of fires, whether by rapidly increasing fuel load or through the plant’s chemical composition. For instance, in California chaparral, introduced grasses have increased the fire frequency markedly; this prevented native grasses from maintaining a sufficiently large seed bank and led to a reduction in the native grasses (Zedler et al. 1983, Jackson and Roy 1989). Attention in the United States has centered on Bromus tectorum, a long-term invader in the intermountain West that has devastatingly altered the fire regime and in turn the composition of plant communities in a vast area. The steppe communities in which B. tectorum is dominant have always burned, but the invasion of B. tectorum has increased the frequency of fire and probably its severity. The result has been major and probably permanent alteration of the species composition in these communities (Daubenmire 1970). High concentrations of resins and other highly combustible compounds and litter in far greater quantities than is cast by natives in a community are not in themselves characteristics of invasive species. But the entry of species with those traits or abilities should be a source of concern. If for unrelated reasons these species became
invasive, they would alter the environment substantially (D’Antonio and Vitousek 1992).
Invasive species can also alter ecosystems through sediment deposition when they grow in areas of water flow where plants have not been found historically. Such alteration has occurred in the Bristol Channel in Great Britain. Spartina anglica, the invasive product of hybridization between North American and European species and subsequent chromosome doubling, is causing accretion of soil in the natural mudflat at rates of 8-10 cm/year (Ranwell 1964).
Genetic diversity can be important to a species’ invasion in a new range. Much information has accumulated on patterns of genetic diversity among plant invaders in recent years, allowing some generalizations to be made. The two most obvious are that the genetic consequences of invasion vary widely among taxonomic groups, depending on their reproductive systems and life histories, and that many nonindigenous species have produced invasions despite their low genetic diversity.
As discussed in Chapter 3, many plant invaders have low diversity at marker loci, such as allozymes (reviewed in Barrett and Shore 1989), but this lack of diversity does not necessarily indicate that they lack genetic variation at all loci, particularly those governing quantitative traits related to fitness (Lewontin 1984, Brown and Burdon 1987). Population differentiation within an invading population after range expansion should be expected, even in species with relatively low variation at neutral marker loci. Studies of genetic differentiation in agricultural weeds or ruderals have often revealed evidence of race formation despite allozyme uniformity (reviewed in Barrett 1988, Warwick 1990). Mutation causes that discrepancy. Genetic variation in polygenic life-history traits is maintained, even in selfing populations, because polygenic mutation rates are considerably higher than those for loci coding for single proteins (Lande 1977). The tempo of differentiation varies considerably among species, depending on the balance between sexual and asexual reproduction and on life-history characteristics.
A source of variation that can favor the evolutionary diversification of invading species is the mixing of genes from multiple founders (Novak and Mack 1995). If there are introductions of a species from different portions of its native range, these will inevitably be genetically differentiated from one another because geographic variation is nearly universal. During range expansion, genotypes derived from separate introductions can cross with one another and give rise to highly variable offspring in the second and later generations. Many of the products of such crosses are likely to be maladapted to local conditions because of outbreeding depression (Templeton 1986), but through selection some individuals with appropriate gene combinations probably will arise from these “genetic soups”. In that way, novel phenotypes that are locally adapted to conditions in
BOX 4-1 Sleeper Species
Sleeper species are those whose populations appear to remain in a quiescent phase for long periods before they begin to proliferate (Groves 1999). The current range expansion and proliferation of rocket (Hesperis matronalis) might qualify it for inclusion in this category. It was introduced early in the 19th century in the United States. It apparently became naturalized soon after arrival, but its numbers did not increase noticeably. In the last 10 years in eastern Washington, what had once been small isolated populations of the species are clearly becoming more abundant, and the individual populations are growing. It is premature to declare this once-inconspicuous species (R.N. Mack, personal observation) an invader, but its range expansion and increase in abundance could well lead to its playing that role.
The causes of the emergence of such nonindigenous species into a more prominent role may be related to intrinsic biological attributes or happenstance (Crawley 1989). Such cases could represent the emergence of a new array of genotypes in the nonindigenous populations through postimmigration selection; this proposal is attractive, but evidence among higher plants and animals has been elusive. The proliferation of sleeper species could simply represent the end of a protracted lag phase in the new range–a phase that has no definable limits. For example, Opuntia aurantiaca resided in low numbers in South Africa for over 80 years before it rapidly rose in prominence and abundance (Moran and Zimmerman 1991). Whether the increase was due to the emergence of new genotypes or to the attainment of the threshold number of foci for range expansion, or both, is unknown. Such range expansions would be more apparent if tied to a rapid and potentially radical change in the environment, particularly one fostered by disturbance. Probably many species have undergone such range expansion as some aspect of the new range environment was altered and the change favored their proliferation.
In contrast, the appearance of sleeper species can reflect our inability to detect new or recently established populations until some chance event occurs. Populations of Asian long-horned beetle and pine shoot beetle, for example, were established for at least 5 or 10 years before they were detected. They came to the attention of entomologists and regulatory officials once specimens were sent to an appropriate expert, but not necessarily because of increased population density or spread. In such situations, the points of entry and establishment and the initial rate of spread by founding populations can be difficult to identify; these limitations increase the difficulty of developing predictions.
Finally, when a nonindigenous species is mistaken for a congener, its spread goes unnoticed. In a park in Halifax, Nova Scotia, the brown spruce long-horned beetle (Tetropium fuscum) was collected in low numbers as early as 1990, but it was thought to be a closely related native species, T. cinnamopterum. It was not recognized as an interloper until numerous trees were dying in 2000, at which point beetles were collected from at least 23 other locations in the Halifax area and correctly identified. Such situations can be especially problematic when they involve groups of arthropods or plant pathogens whose systematics are not well understood.
the new range can potentially evolve. This scenario seems likely in the history of the plant invaders Echium plantagineum (Patterson’s curse) in Australia and Lythrum salicaria (purple loosestrife) in North America; in both cases, there is evidence that multiple introductions from Europe resulted in considerable genetic diversity in introduced populations (Thomson et al. 1987, Burdon and Brown 1986). Such diversity is likely to favor rapid evolution of local races, despite recent efforts at biological control.
The idea that the mixing of distinct genetic lineages through multiple introductions can give rise to novel phenotypes is supported by evidence from hybridization studies in several flowering plant groups. For example, morphological and genetic evidence demonstrates that the recently derived allotetraploid ruderals Tragopogon mirus and T. miscellus originated in the U.S. Pacific Northwest through hybridization between allopatric species from Europe that were introduced in historical times (reviewed in Novak et al. 1991). Similarly, molecular evidence indicates that the common cattail Typha glauca, an important component of wetlands surrounding the Great Lakes, is a stabilized F1 hybrid between the native T. latifolia and introduced European T. angustifolia (Kuehn et al. 1999). Not all hybridization events result in the origin of stable F1 hybrids. Recent molecular studies of putative hybrids between the native Californian cord grass (Spartina foliosa) and S. alternifolia, an introduced species from the East Coast of North America, demonstrated extensive introgression between the two parental species in the San Francisco Bay (Ayres et al. 1999). Abbot (1992) discusses other examples of new plant invaders that arose from hybridization. Rhymer and Simberloff (1996) review cases in which such hybridization and introgression of plants and animals pose ecological and genetic threats to the survival of native species.
The capacity of nonindigenous species to evolve after their arrival in a new range complicates our ability to predict their postarrival behavior. Novel genotypes emerge through strong selection, hybridization, and the sharing of genes among members of the same species drawn from different parts of the native range. Selection in the new range sorts among these new genetic products, and the result can be organisms more adapted than their ancestors to the new range. As cited above, there are cases in the United States of new species, created through allotetraploidy, whose existence, to say nothing about their role in the new range, could not have been predicted. Those results have implications for current quarantine practices. With few exceptions, nonindigenous species already found in the United States are not barred from further entry. That practice has the unintended result of potentially allowing the introduction of a species’ genotypes that were previously unrepresented in the United States. As has already occurred repeatedly, the continuous potential exists for the eventual assembly of new, invasive genotypes from among an array of genotypes each of which by itself is innocuous or at least only has the ability to become naturalized, not invasive. The
degree to which this phenomenon has occurred and can occur, particularly among species with broad native ranges, should be the subject of active research.
Hybridization of related plant species originating in different areas can influence gene flow in their pathogens. That can have substantial evolutionary consequences, such as influencing host-pathogen interactions with respect to resistance and virulence structure. For ecotypes of the native Australian flax, Linum marginale, which is subject to attack by the rust pathogen Melampsora lini, analysis of the resistance structure of ecotypically different plant populations revealed that plants of one ecotype were generally susceptible to pathogen isolates taken from all sites but that plant hybrids exhibited resistance similar to that of plants from the more resistant population. Similarly, the virulence structure of rust isolates collected from the hybrid plant population was more similar to those isolates taken from the more resistant population than from the susceptible ecotype. In addition, plants from the more susceptible population had substantially higher survivorship than the resistant plants, regardless of where the plants were grown (in the susceptible, resistant, or hybrid zones). Those results suggest that the likelihood of differential gene flow and survival of resistant or susceptible plants of different ecotypes at least partially explains the maintenance of a relatively narrow hybrid zone (Carlsson-Graner et al. 1999).
Given the array of new ranges that introduced species might enter, it is worth considering briefly how ecological context can influence the speed of evolution. The rapidity with which a population can respond to selection will depend on the amount of additive genetic variation for fitness-related traits and the strength and direction of selection. Modern agricultural practices—in which the grower makes every effort to eradicate unwanted plants, often including nonindigenous plants— probably constitute some of the most intense selection pressures that introduced species encounter. Much of this selection can be quite consistent in direction, especially where monocultures are grown from year to year. There is good evidence that under these conditions such weedy (and usually nonindigenous) plant populations have responded through the evolution of locally adapted races that are specifically adapted to particular crops (Baker 1974, Barrett 1988). Indeed, many agricultural races of weeds (agroecotypes) are so specialized to the crop environment that they are rarely encountered outside the agroecosystems to which they have become adapted (Barrett 1983).
Among hosts and pathogens that have undergone coevolution, the pathogens may have an evolutionary advantage. In those cases, pathogen populations could be locally adapted, having higher mean fitness on sympatric than on allopatric hosts. Simple frequency-dependent selection models predict complex patterns of pathogen performance on sympatric and allopatric populations. With local extinction, recolonization, and gene flow in metapopulations, variable selection pressure and stochasticity could obscure local processes or change the extent to which local adaptation occurs. Alternatively, gene flow could introduce adaptive
variation, and differential migration rates could modify the asymmetry of host and pathogen evolutionary rates (Koltz and Shykoff 1998).
Selection pressures encountered during the invasion of natural habitats are likely to be much more complicated than those which occur in agroecosystems. In particular, selection is likely to be less predictable and perhaps less directional. Diversifying selection would then be expected to maintain variation in populations, and the rapid evolution of specialized races of limited ecological tolerance would be less likely at least over short periods. Predicting the course of evolution in such biotically complex situations will be more difficult than in agricultural habitats, where growers strive for environmental homogeneity through the production of a uniform set of growing conditions for the crop.
Some persistent nonindigenous species undergo a transition and become invasive. However, compared with the factors that govern arrival and persistence, there is considerable uncertainty regarding the biotic and abiotic factors that produce an invading population.
The transition from established (or persistent) to invasive does not have a single explanation, but appears to be caused by the interaction of chance events in the new range, the biological attributes of the species, and the recipient community’s composition.
The history of a species’ invasion in another location or country is often a useful predictor of its behavior in a new area.
Many of the characteristics that determine establishment also are important in determining an invading population. For instance, a high reproductive or growth capacity and the availability of suitable habitats, resources, or hosts are important. It is not always known why a species with a moderate rate of population growth makes the transition to a very high rate of growth.
Dispersal is clearly important in facilitating the transition to an invasion. Relevant dispersal components include the number of propagules available for movement, the opportunity for and mechanisms of dispersal, the distances that propagules are transported, and the spatial pattern of the dispersed propagules.
Although some invasions advance as a continuous wave front, most advance spatially by establishment of some minimal number of widely separated foci. Establishment of multiple foci appears to be very important in triggering an invasion, regardless of the taxonomic group.
Habitats and communities differ in their vulnerability to the entry of potentially invasive species. For example, arid or otherwise treeless communities appear more vulnerable to airborne plant dispersal than communities with a continuous forest canopy. In contrast, forest communities might be more vulnerable to animal-dispersed seeds and fleshy fruits. Sometimes, nonindigenous species
become invasive through their ability to tap resources unused or underused by native species.
Biotic agents, including competitors and mutualists, play a role in the transition of an established species to a proliferating and spreading species. However, there appear to be no consistent relationships across groups (plants, arthropods, and plant pathogens).
The roles of biodiversity and habitat disturbances in influencing species invasions are hotly contested. At best, conclusions depend on the invasive group.
The genetic consequences of invasions vary widely among taxonomic groups. High genetic diversity of an established species is not a requirement for its transition to an invader. However, multiple introductions of a species into a new range often facilitate the emergence of new genotypes, some of which will have higher fitness than their parents. The result is an increased probability of yielding an invading population.
Agricultural practices associated with a crop tend to provide intense and highly directional selection of invasive species which results in locally adapted races limited only by the area of the crop.