Establishment refers to the apparent persistence of a population and is equivalent to the term naturalization as used for the descendants of plant immigrants (Mack 1997). To become established or persistent, immigrant plants, arthropods, or pathogens (or their descendants) need not be widespread or growing rapidly, but they must withstand challenges to their survival. The challenges include the random events that often prove detrimental to the survival of small populations, including a potentially hostile climate; inadequacies in nutrients, hosts, or mates; and competitors and predators. The vast majority of immigrant populations do not become established. For example, although thousands of nonindigenous phytophagous insects arrive in the United States every year, establishment is a rare occurrence (Carey 1996, Lewis and Kareiva 1993). Only an estimated 10% of all nonindigenous insect species that are introduced into a new range become established. Even when insects are carefully selected for intentional introduction as biological control agents, only about 33% become established (Williamson and Fitter 1996).
How does a species become established in a new geographic range? Studies dealing with that question have focused on the environmental forces that impede establishment while identifying particular circumstances of an introduction, such as the size of the founder population or the species’ traits, that might facilitate its ability to overcome that impediment. For example, it has been recognized for more than 150 years that immigrants into a new range gain a substantial potential advantage by leaving behind their native competitors, natural enemies, or other biotic constraints on growth, development, survival, and reproduction. Release from such constraints might in itself explain the superior individual growth and
large increases in population that have been documented for many species in new ranges (Thebaud and Simberloff 2001). When biological control of an invasive species, which involves the deliberate introduction of parasites and predators of the species from its native range, is successful, it is powerful confirmation of the potential role of biotic constraints in curbing a species’ abundance and distribution.
In this chapter, we first review the stochastic character of forces that determine the establishment or persistence of populations, especially the small populations that typify immigrants. We then illustrate how biologists have attempted to categorize the role of a population’s spatial structure in interacting with the stochastic character. Without attempting to be inclusive, we next proceed to illustrate the breadth of environmental factors, both abiotic and biotic, that form the specific forces that immigrant populations encounter. Finally, we provide illustrations of life-history traits that can influence immigrants’ tolerance of a new range. In compiling this chapter, we were aware that some aspects of the discussion of persistence are equally pertinent to the proliferation and spread of a species, the topic of Chapter 4. As a result, we introduce these shared properties or circumstances here and mention them only briefly in Chapter 4. Although additional observations, new hypotheses, and empirical testing are needed to confidently predict the net effect of these countervailing forces on the outcome of a specific species’ introduction, a body of data that underlie the basis for establishment is slowly emerging.
STOCHASTIC EXTINCTION— THE PERILS OF SMALL POPULATIONS
The defining demographic of an introduced species is typically its small population. Even the planet’s most abundant and widespread invasive species commonly began as “rare” species in their new ranges, compared with the size of their populations in their native ranges. (There are exceptions; see Eckert et al. 1996). Although population extinction can result from deterministic processes— ranging from direct eradication by humans to fire or flooding—a pervasive threat to the persistence of small populations is stochastic extinction. This risk of extinction for small populations transcends the taxonomic groups we examine here. For plants and arthropods, the topic is often examined in terms of the minimal viable population (see Box 3-1); for pathogenic microorganisms, it is referred to as the minimal infective dose. Consequently, our remarks should be interpreted as relevant for all taxonomic groups unless stated otherwise.
Demographic models have relied on such quantities as the mean population growth rate and the variance in growth rate (Leigh 1981, Goodman 1987) and predict that persistence time increases slowly with increasing carrying capacity. As a result, extinction is predicted to be much more likely at lower population numbers, a conclusion that is biologically realistic. To facilitate discussion of the importance of stochastic extinction in the early stages of establishment, the forces
BOX 3-1 Minimal Viable Population
Theoretical development associated with the concept of minimal viable population (MVP) for rare taxa (Soulé 1987, Lande 1988, Menges 1991) has demonstrated the profound demographic implications of low population numbers in the extinction process. Although descriptions of what is implied by MVP have varied (Shaffer 1981), most definitions focus on estimating the minimal population density for a given probability of population persistence over a specified period. In applications of the idea of MVP for rare species, the time frames are often thousands of years (Shaffer 1987). As a result, the recommended population sizes in conservation biology are often large to reduce the likelihood of chance extinction over the long term by various stochastic processes that can differ in their frequency and their potential demographic and evolutionary impacts.
Nonindigenous species are often rare during the initial phases of their colonization. By analogy, an application of the concept of MVP to a newly introduced species might be useful in predicting the likelihood of its establishment. Needed here is knowledge of the acceptable maximal persistence time for the newly introduced species. Precisely because they are at low population densities, new introductions are typically not detected, so it is difficult to know how long they persisted without obvious population growth. However, one would expect that the periods for the establishment phase of currently widespread nonindigenous plants were much less than the 1000-year persistence times sometimes recommended for rare species. Actual periods for the establishment phase in invasive species on which data are available indicate that the initial “lag phase” in the population growth trajectory is often less than 100 years. For example, after initial introduction into northern Australia around 1900, Mimosa pigra populations remained small and confined to environs around Darwin until the 1950s. After that short lag phase, this native of Mexico and South America began to spread rapidly throughout Australia’s Northern Territory; it is now a major threat to wetland areas in Kakadu National Park (Lonsdale et al. 1989, Cousens and Mortimer 1995). The population required to ensure persistence for this much shorter period could be very small and thus much more difficult to detect and eradicate. Difficulties in detection because of low population densities might be exacerbated in invasive species that can reproduce vegetatively or by uniparental forms of reproduction. The theory predicts that the population required for persistence of clonal or parthenogenic species is even smaller than those for species that reproduce sexually (MacArthur and Wilson 1967, Shaffer 1981).
Efforts to predict the likelihood of establishment of a recently introduced species will need to consider how an Allee effect may influence the MVP. Life-history traits related to dispersal, reproduction, host finding, predator defense, or other factors can affect the critical density threshold, below which a population that is subject to inverse density dependence cannot recover. Characteristics of the habitat colonized by the founding population can further interact with an Allee effect; the MVP may be higher or lower, depending on the size of the founding colony and the extent of environmental stochasticity.
of stochasticity that can affect population persistence can be grouped into four main categories: demographics, environment, natural catastrophes, and genetics (Shaffer 1981).
Demographic stochasticity refers to the chance variation in survival and reproductive rates in very small populations (for example, the probability that all individuals in a population will not reproduce in a given period). Although average values for survival or reproduction might remain relatively constant, chance variation can occur among individuals. As a simple example, consider an annually reproducing population with nonoverlapping age classes, whose average probability of survival to reproduction is 0.20. Despite this average survival rate, if the population is extremely small—say, 10 individuals—there is a probability of about 11% in any one year that all members of the population will die before reproducing, in which case the population will go extinct. If the population is somewhat larger–say, 50 individuals–the probability that the population will go extinct during a given year is very low, less than 0.002%. Equally important is that the probability of demographic extinction is associated with just a single “trial” (that is, a single year). As the “demographic dice” are cast each year, the cumulative probability of chance extinction increases as a simple arithmetic product of the probabilities for all consecutive years if the population does not increase. Although the numerical threshold where demographic stochasticity can cause extinction depends on the particular situation, a threshold estimate for dioecious organisms of about 50 individuals is widely cited (Pollard 1966, Keiding 1975, Shaffer 1981). The fate of a population for which this type of stochastic process becomes important is bleak; as noted by Gilpin and Soulé (1986), demographic stochasticity might well be viewed as “the immediate precursor of extinction.”
Environmental stochasticity usually reflects the impact of random variation in the environment as it influences a population. In its simplest form, the demographic effects of environmental perturbations are assumed to be equally distributed across all individuals in a particular age or stage class in a population. For example, an increase in seed predators might reduce average reproductive output by 80% in a plant population; this decrease, because it represents a drop in average reproductive output across all individuals in the reproductive age class, can severely reduce the size of the population, regardless of its initial size. Because of its capacity to adversely affect even large populations, environmental stochasticity is considered an important force in promoting chance extinction (Sykes 1969, Cohen 1979, Leigh 1981, Menges 1990, 1991). In essence, envi-
ronmental stochasticity can reduce a population’s size to a point at which demographic stochasticity becomes important and causes chance extinction. Goodman (1987) has suggested that when environmental stochasticity is present, expected population persistence time increases roughly as a linear function of population size.
Given that the typical effect of increasing environmental stochasticity is to reduce the likelihood of population persistence, there should be selective value in the capacity of an organism to reduce the demographic impact of environmental variation. One potential evolutionary response is the development of adaptive phenotypic plasticity: reducing the variance in survival, growth, and reproduction buffers the adverse impact of environmental variation on population persistence (Caswell 1983). Because of plants’ modular construction, phenotypic plasticity in plant structure has been suggested as an adaptive response to buffer localized gradients in resource availability (Bradshaw 1965, Schlichting 1986). In both plants and animals, physiological plasticity in the form of acclimation to variation in climatic factors, such as temperature, has also been suggested as an adaptive plasticity response (Sultan 1987).
Floods, fires, earthquakes, droughts, ice storms, and the like, which all occur relatively infrequently and at random intervals, are also potent threats to the persistence of populations. In a sense, natural catastrophes constitute a more extreme and less predictable form of environmental stochasticity. Given the sporadic occurrence of natural catastrophes, it is difficult to characterize the disturbance caused by them and thus incorporate them realistically into models of population persistence. Some theoretical treatments of natural catastrophes indicate that there is a “diminishing-returns” effect: the average persistence time of a population increases only in correspondence to the logarithm of its size (Ewens et al. 1987). Thus, greater and greater increases in population size are required to gain the same increase in persistence time. Further work has indicated that the form of the relationship depends on the carrying capacity of the habitat and on the severity and frequency of catastrophes (Lande 1993). As a result, a population might need to be very large (in the thousands to millions) if it is to persist in spite of periodic catastrophes (Shaffer 1987).
Stochastic forces involving founder events, genetic bottlenecks, and genetic drift play a dominant role in determining the fate of immigrant species because frequent colonizing episodes are a central feature of their population biology. These forces tend to dominate during the early stages of establishment after long-distance dispersal, when populations are often very small. Theoretical studies
demonstrate that frequent bottlenecks and low effective population sizes reduce genetic variation, especially the frequencies of rare alleles (Nei et al. 1975, Sirkkomaa 1983). However, strong directional selection during the early stages of population establishment can also reduce genetic variation, particularly at loci that govern fitness. Hence, a population early in its postimmigration history is especially vulnerable because random genetic events erode the little genetic diversity that introduced populations usually contain. Low population size can also result in increased mating among related individuals; increased inbreeding can result in the expression of recessive deleterious alleles in homozygous form and cause reduced fitness or inbreeding depression in the progeny (Barrett and Husband 1990, Barrett and Kohn 1991).
The input of mutational variance, recombination, and gene flow theoretically can counteract those forces. In addition, many species have become invaders despite low genetic diversity. In these cases, extensive areas of the introduced range often comprise a small number of genotypes (Moran and Marshall 1978, Scribailo et al. 1984, Barrett and Shore 1989, Novak and Mack 1993). Low diversity at the population and regional levels is especially evident in plant species that propagate by asexual reproduction, sometimes called selfing.
However, effects of genetic bottlenecks associated with introduction of small populations can also be counterintuitive and complex, as evidenced by the introduction of the Argentine ant (Linepithema humile) into California. In its native Argentina, intraspecific competition limits colony size and population density, and numerous ant species co-exist with the Argentine ants. Suarez et al. (1999) found that populations of the Argentine ant in California exhibited substantially lower heterozygosity than populations in their native range. Genetic changes in the introduced ant populations were associated with altered behavior that reduced nestmate recognition and hence intraspecific competition and resulted in high population densities, competitive displacement of a majority of the native ant species, and adverse effects on ant predators, such as horned lizards (Suarez et al. 2000). Evidence suggests that the imported fire ant (Solenopsis invicta) similarly experienced a pronounced genetic bottleneck when it was introduced into North America (Ross et al. 1993). Altered ecological characteristics and population genetics of the introduced fire ant populations appear to be associated with changes in the social organization of its colonies (Ross et al. 1996). Such changes include multiple queens and zero relatedness between workers and new queens in the introduced populations, compared with few queens and significant relatedness between queens and workers in the native range. Low relatedness has potential advantages for ants and has been associated with rapid colony growth in other ant species (Cole and Wiernasz 1999).
An Allee effect occurs when low-density populations sustain a zero or even negative rate of increase because of reduced reproduction or survival when con-specifics are insufficiently abundant. Eventually, “undercrowding”—inverse density dependence at low density—drives the population below a critical threshold and extinction occurs (Allee 1931, Courchamp et al. 1999). The dynamics of Allee effects, therefore, can exert a substantial influence on whether a colony of a newly arrived organism is able to persist and become established. One frequent cause of Allee effects is a scarcity of reproductive opportunities at low densities. For example, in some insect populations, difficulty in locating conspecific mates will reduce the likelihood that individuals in the next generation will produce offspring, and in the case of arrhenotokous insects (a population in which unmated females produce only males), will result in a population with a male-biased sex ratio. Ultimately, this form of demographic stochasticity leads to collapse of the population. Other factors may also generate inverse density dependence in low-density populations, including a reduction in the ability of individuals to find or use suitable host plants (Way and Banks 1967), decreased ability to cooperate in defense against predators (Turchin and Kareiva 1993), and genetic inbreeding that leads to decreased fitness (Courchamp et al. 1999, Lamont et al. 1993).
The strength of an Allee effect on the persistence of a population depends on the processes influenced by inverse density dependence. Species in which fitness is enhanced in some way by conspecific facilitation or cooperation may be subject to Allee effects only at very low densities. In contrast, species with obligate sexual reproduction may be more strongly affected by an Allee effect and at a wider range of densities (Courchamp et al. 1999). Species that are subject to a strong Allee effect may also be more vulnerable to extinction due to environmental stochasticity because the population size below which they cannot recover from an unfavorable weather event will be larger than for other species.
Influence of Allee effects on establishment has been addressed by examining the dynamics of populations of nonindigenous species released in biological control programs. Hopper and Roush (1993) found that parasitoid wasp and fly species released for biological control of leaf-feeding insects were subject to an Allee effect when dispersal from low-density populations led to low mating success, which resulted in male-biased sex ratios. Modeling showed that the Allee effect could drive populations to extinction and that it limited establishment of introduced parasitoid species more than the limits imposed by stochastic environmental variation or lack of genetic variation. Grevstad (1999a,b) used simulation modeling to show that if the net reproductive rate of a newly introduced species was even slightly greater than 1.0, demographic stochasticity was unlikely to limit population persistence. Environmental stochasticity, however, interacted with an Allee effect. When an Allee effect was present and environmental conditions were relatively constant, establishment of a species was most likely to be
successful if there was a single introduction of many individuals. In contrast, when environmental stochasticity was great, establishment was more likely to occur if there were many introductions of low numbers. Interactions between an Allee effect and environmental variability in larger colonies were more deleterious than the sum of their independent influences; a few years with bad conditions reduced population density to a level where negative population growth eventually led to extinction (Grevstad, 1999a,b). The need to estimate the influence of Allee effects on any immigrant population depends on the species and the circumstance, thereby complicating the population state in a new range.
Effects of Spatial Structure
Models estimating the persistence of a population typically consider only aggregate population statistics, not the spatial distribution of individuals. As a result, the question arises as to whether the extinction dynamics of spatially structured populations (such as a patchy distribution) might differ from those with little spatial structure (such as an aggregate).
The answer is frequently yes. Effects of environmental stochasticity can be reduced if the species is patchily distributed in such a way that not all members in any age class are affected equally by environmental perturbations. The potential for spatial patchiness to function as a buffer against extinctions caused by environmental stochasticity depends, however, on the degree to which environmental fluctuations are correlated across patches (Gilpin 1987, Stacey and Taper 1992). If environmental conditions are correlated across the region occupied by subpopulations of a newly arrived species, “Moran effect” dynamics could lead to a tension between synchronizing effects of extrinsic environmental stochasticity and desynchronizing effects of nonlinear density dependence (Hudson and Cattadori 1999, Ranta et al. 1999). In the Moran effect, originally proposed as a mechanism to explain synchronized fluctuations in Canadian lynx populations (Moran 1953, Royama 1992), when disjunct populations have the same endogenous structure (such as density dependence), a correlated exogenous, density-independent factor (such as weather) will bring population fluctuations into temporal synchrony (Earn et al. 1998, Heino et al. 1997, Hudson and Cattadori 1999, Williams and Liebhold 1995). Thus, when a species is strongly influenced by Moran effect dynamics, a period of unfavorable environmental conditions will promote extirpation of all colonies. A newly arrived immigrant species will persist, therefore, only if at least some subpopulations experience “good times” while other subpopulations suffer reduced population growth during “bad times”. In contrast, spatially structured populations may be subject to inverse density dependence regardless of environmental conditions if subpopulations are diluted by dispersal. Such a scenario could increase the probability of extinction due to stochasticity or Allee effects (Hopper and Roush 1993, Lewis and Kareiva 1993).
If the initial distribution of the populations of an immigrant species across a spatial hierarchy is known, this information might predict the potential for spread. The concepts developed in conservation biology for rare native species can be applied to understand the dynamics of the small populations that characterize immigration and establishment. In an attempt to understand what it means to “become rare”, Rabinowitz (1981) proposed a classification system that uses three attributes to characterize rarity: local population size, habitat specificity, and geographic range. When expressed as dichotomies, the three attributes result in an eight-cell table (Table 3-1) that describes the spatial characteristics of different forms of rarity. For plant pathogens, habitat would refer to the host infected (for example, a wide habitat would mean a wide host range for the pathogen).
The upper-left cell represents the distribution pattern of a common species, but the remaining seven cells describe distinct categories of spatial patterning by which Rabinowitz classified rare taxa and asked questions about the origins of rarity. That approach has application for nonindigenous species: one can ask how nonindigenous species differ in their potential for becoming common. If we assume that immigrants initially consist of small, sparse populations, we can restrict our attention to the bottom row of the table.
Even casual inspection of the categories of the bottom row suggests that some distribution patterns will be more worrisome than others as signals of a species’ potential spread in a new range. The far-right cell in the bottom row suggests a single introduction. Because of its habitat specificity and small geo
TABLE 3-1 Spatial Characteristics of Different Forms of Species Rarity
Local Population Size
Large Geographic Range Habitat Specificity
Small Geographic Range Habitat Specificity
Large, dominant somewhere
Locally abundant over a large range in several habitats
Locally abundant over a large range in a specific habitats
Locally abundant in several habitat but restricted geographically
Locally abundant in a specific habitat but restricted geographically
Small, sparse everywhere
Consistently sparse over a large range and in several habitats
Consistently sparse in a specific habitat but over a large geographic range
Consistently sparse in several habitats but restricted geographically
Consistently sparse in a specific habitat and geographically restricted
SOURCE: Adapted from Rabinowitz 1981.
graphic range, this type of immigrant is probably most susceptible to stochastic extinction. At the other end of the spectrum is the immigrant described by the far-left cell in the bottom row. This species inhabits a broad geographic range and is found in many habitats (or associated with many host species). Its wide distribution (both geographically and by type of habitat) might reflect multiple introductions. Stochastic extinction is very unlikely for a species with this distribution pattern. Obviously, if populations increase in size (move to the upper row), the species will become common and probably invasive. The next cell to the right represents a scenario that, again because of the broad geographic range, suggests multiple introductions. Because the species is restricted to a particular habitat, the emergence of an invader through this scenario might be predictable, and effective management and containment of the species might be quite feasible.
Rabinowitz (1981) suggested that there might be no rare species in the last category (third cell, second row). There might indeed be no rare native species that are limited geographically and are found at low population density in several habitats. However, this scenario could be common among persistent nonindigenous species. It represents a situation similar to that of the most problematic category (the far-left cell) except that the limited geographic range described for this cell suggests a single introduction. One might suspect that such a species could become established, and even invasive, and is limited only by lack of additional introductions throughout its potential range. Early detection of a nonindigenous species with this distribution should evoke immediate consideration of eradication.
Although those categories are admittedly simplistic, they do provide a general conceptual framework for understanding the various pathways to abundance that a newly established immigrant population might follow.
Avoiding Stochasticity: Multiple Introductions and Population Size
Aside from the obvious inability of a population to survive under a climatic regimen that is well beyond its tolerance, stochasticity is the greatest threat to small populations. And as stated above, small numbers typify most unintentional introductions of nonindigenous species. But if the founder population is large enough or replenished with additional propagules, it can withstand stochastic forces—provided it can tolerate the overall character of the new environment.
Empirical studies of the relationship between the probability of establishment and the size of the founding population stem from evaluations of efforts to establish insects for biological control. In general, establishment is predicted to increase with population size (Grevstad 1999a). However, it is also possible that persistence is independent of population size if density-independent factors— such as weather, habitat conditions and the size of the habitat patch—are the main determinants of persistence or if the population’s numerical increase allows initially small populations to escape rapidly the risk of extinction.
Bierne (1975), in a widely cited retrospective analysis of the role of initial colony size in establishment, reviewed Canadian biological control programs. He found that if fewer than 5000 individuals were released, about 9% of the species became established—a percentage that is similar to the 10% estimated for accidental introductions (Williamson 1996). If at least 30,000 individuals were released, 79% of the species became established. The average number of individuals per release site also seemed to be important; establishment rates increased from 15% to 65% of species if at least 800 individuals were released at a single location. Bierne’s (1975) approach has limitations: the role of the number of individuals released in each case might be confounded by the traits of particular species; and these traits also contribute to establishment, such as a high reproductive rate or abundant distribution in their native range (Crawley 1986, 1989b). Grevstad (1999a) took an experimental approach in assessing relationships between population size and persistence. She followed the fate of 92 experimental releases of two chrysomelid beetles for three full generations and found that the probability of establishment increased over the range of initial beetle density (20, 60, 180, and 540 beetles). Population growth rates varied among environments but were positively related to release size. In a second experiment, in 20 releases of single gravid females, only one female founded a population that persisted for the duration of the 3-year study.
Although Levine and D’Antonio (1999) contend that any community, given enough propagules, can be invaded, nonindigenous insects are usually introduced accidentally and presumably arrive in low numbers in most cases. Small populations have a greater random chance of extinction than large populations and are more vulnerable to inbreeding depression, so the descendants of founders might also need propitious conditions to survive. Liebhold et al. (1995) and MacArthur and Wilson (1967) suggest that the probability of establishment can be simply described as a continuous function of initial population size. Translating these conclusions into practical terms is currently difficult because the small immigrant population that is detected may be only one of a large group of small populations that arrived at the same time. Although many may soon go extinct, some may survive unless they are deliberately eradicated.
Not only a large number of immigrants per introduction, but also frequent introductions will mitigate stochastic processes and increase the likelihood of establishment. Crawley (1986) found that the probability of establishment increased with both the sizes and the number of population releases. Repeated introductions increase the chance that an immigrant species will encounter a combination of resources, scarcity of competitors, and low density of predators or pathogens that will permit establishment. That expectation appears to be borne out in the release of insects for biological control. In the Canadian biological control programs, nonindigenous species that were released in at least 10 episodes established in 70% of the cases; those released in fewer than 10 episodes established in only 10% of the cases (Bierne 1975).
Simulation analyses using matrix models can examine how such factors as environmental stochasticity can interact with demographic characteristics, such as deterministic growth rate (λ), to influence the persistence of small populations. Using stage-structured matrix models, Menges (1990, 1991, 1992) explored how the persistence of plant populations is affected by environmental stochasticity and how the effects of environmental stochasticity are modulated by variation in rates of increase. Using life-table data on several plant species, Menges (1992) found that, except for very small populations with deterministic growth rates near 1.0, demographic stochasticity was much less important than environmental stochasticity in causing chance extinction. Moderate environmental stochasticity was found to cause extinction even in populations with positive deterministic growth rates. By systematically varying environmental stochasticity in a series of simulations, he demonstrated that populations with deterministic growth rates near 1.0 are much more likely than populations with higher growth rates to go extinct when exposed to moderate environmental stochasticity. Populations with deterministic growth rates greater than 1.2 were affected only by extreme environmental stochasticity. The potential importance of that final result to invasion biology is that if the deterministic growth rate of a newly established population of invaders is much greater than 1.0 (say, 1.5 or higher) the likelihood of chance extinction is much reduced.
THE ENVIRONMENTAL CONTEXT
Against the continuous backdrop of stochastic forces that affect a fledgling population are the environmental characteristics of the new range. Abiotic factors–such as climate and landscape–will determine whether a new range is at least minimally habitable by an immigrant. Biotic factors–including the availability of hosts or pollinators and the presence of competitors, predators, and plant-pathogen antagonists–challenge the biological tolerance and competitive abilities of the newcomer. The following discussion is by necessity abridged and serves only to illustrate the breadth of environmental factors that affect establishment.
The likelihood of establishment will be affected by the general climatic match between the donor habitat and the new habitat of the immigrant species. Moreover, the geographic distribution and range of climatic conditions known to have been suitable for the immigrant species in its native range or in previously invaded regions provide some indication of potentially suitable habitats. Prediction of establishment and invasiveness based on climate-matching between original and potential ranges of nonindigenous species is a subject of active research (Kriticos and Randall 2001). This correspondence is related to the potential
inability of a nonindigenous species to tolerate weather and climatic conditions in a new range, a common cause of founder-population extinction (Crawley 1986, Leigh 1981, Levins and Heatwole 1973).
In addition, the latitudinal range can affect the ability of many species to become synchronized rapidly with a new habitat. Insects in a new range must initiate and terminate diapause at appropriate times to cope with climatic extremes. If we consider insects in the Northern Hemisphere, we might expect that insects from northern latitudes would be more likely to become established in southern latitudes of the Northern Hemisphere than the reverse. The higher probability of establishment in southern latitudes occurs because the consequences of failing to enter diapause soon enough in autumn or breaking diapause too early in spring are presumably more detrimental than entering diapause too early or breaking it relatively late. Niemelä and Mattson (1996) noted that deciduous forests in Europe span a higher latitudinal range (43-60°N) than deciduous forests in the United States and Canada (30-48°N); this might partially explain the invasion of many European forest insect species in North America.
Similarly, establishment of univoltine insects (those with only one generation each year) would be unlikely if an insect moved between the Northern Hemisphere and the Southern Hemisphere. Such a species would be disadvantaged because dormant eggs or pupal stages would encounter antipodal seasons in the new habitat (Simberloff 1986, Ridley et al. 2000).
For introduced forest pathogens, temperature and moisture appear to be critical. The influence of climate on the incidence of plant disease incidence has been studied for those indigenous and nonindigenous microorganisms that exist in the United States and infect agricultural crops (Coakley 1988). Numerous climatic factors, particularly moisture and temperature, dictate where a microbe survives and whether it is able to infect a suitable host. Predictive models, based in part on climatic factors, have become increasingly useful in estimating disease severity and showing patterns of potential pathogen distribution on crops, for example, with Peronospora tabacina, the tobacco blue mold (Main et al. 1998).
Damage by the introduced European larch canker organism (Lachnellula willkommii) appears restricted to the coasts in eastern Canada and Maine (Ostaff 1985), where fog, more rainfall, less snow, and higher mean monthly winter temperatures than inland sites are prerequisites for disease development. With Scleroderris canker, another disease of conifers, the introduced European race of the causal fungus, Ascocalyx abietina, predominantly infects seedlings and lower branches of some pine species. As with the North American race, a period of low temperatures and snow appears necessary for the disease to develop (Marosy et al. 1989). As a result, this pathogen has not spread substantially beyond the sites where it was introduced on infected nursery stock in Canada and the United States (Laflamme et al. 1998). Cronartium ribicola, the introduced pathogen responsible for white pine blister rust, has been the object of one of the more detailed evaluations of microclimatic factors that constrain infections. Higher
infection rates occur where air drainage influences spore dispersal and provides the specific microclimatic conditions required for infection (Van Arsdal 1967, Dahir and Cummings Carlson 2001).
To understand how variation in the physical environment might affect immigrant populations, it is critical to relate the environment to demographic characteristics. For example, even if the average climatic conditions at a site favor population growth and persistence in most years, variation around the average growth rate caused by an occasional year of unfavorable weather can drive the population to extinction. Weather, the combined random variation in a host of climatic variables, is often the primary source of the environmental stochasticity incorporated into population viability analysis. Extreme weather events are one aspect of this stochasticity. Thus, a species from a native range without extreme weather events could conceivably be at heightened risk in a potential new range simply because of the occurrence of this expression of stochasticity.
Although the importance of human-generated disturbance in promoting the spread of invaders is well established (Harper 1965, Mack 1989), the role of disturbance in affecting the early establishment of invasive species is uncertain. Disturbance can reduce an immigrant population to such a low density that demographic stochasticity could cause extinction. It can also adversely affect population persistence if it increases environmental stochasticity or takes the form of natural or human-mediated catastrophes.
However, if disturbance effectively reduces environmental stochasticity or increases resource availability or both, it can promote the persistence of the initial colonists. A common example of this type of enhancement of population persistence by disturbance is cultivation. Plant cultivation typically reduces seasonal and year-to-year variation in a wide array of environmental factors, both abiotic and biotic. In addition to reducing variation in demographic characteristics, the higher level of resource availability typical in cultivated systems increases average population growth and further decreases the impact of environmental stochasticity on population persistence. By reducing environmental stochasticity and eliminating native competitors, cultivation can be a potent force in promoting the persistence of nonindigenous populations, thereby eventually allowing them to become established and even invasive (Mack 2000).
Native organisms in the new range of a recently introduced species can be essential for its survival as hosts, mutualists, and vectors, or they can threaten its survival as predators, grazers, competitors, parasites, and pathogen antagonists. The consequences of these agents might not be expressed on entry of the immigrants into the new range. Strong et al. (1984) showed that the likelihood that a nonindigenous species will encounter such new biotic constraints depends on the length of time in the new range, the size of the new range, and the availability of
nonindigenous enemies. That an immigrant species will eventually acquire new biotic constraints seems inevitable, although the constraints might not arrive until long after the nonindigenous species has become persistent or even invasive (Strong et al. 1984).
A multitude of biotic factors can prevent establishment of immigrant insects and pathogens in a new range. The immigrants must be able to locate suitable host plants or come into contact with a suitable host through the deposition of spores and eggs or other dispersal units, survive interactions with newly encountered enemies, and compete with native species. A new nonindigenous insect or pathogen is often not detected for years or even decades (Carey 1996), and the difficulty of early detection makes it difficult to identify factors that were conducive to establishment. General patterns can be recognized, however, and might be useful in efforts to predict future invasions.
Host Availability and Distribution
The establishment of nonindigenous pathogens and arthropods is subject to the availability of a suitable host. Some plant pathogens, such as the cereal rust fungi, are extreme specialists; they infect only specific plant varieties. Others— for example, such root pathogens as Pythium species, will infect a wide taxonomic range of hosts. Although the wide host compatibility of some pathogens increases the probability of their finding a suitable host, many pathogens with narrow host ranges (such as Sporisorium sorghi, the sorghum smut fungus) have become established and even invasive in the United States. For example, plant pathogens that have invaded the temperate forests of North America cause epidemics on a few closely related species. Generalist pathogens are rare in forest ecosystems; one example is Erwinia amylovara, the fire blight pathogen (Vanneste 2000).
Not surprisingly, North American forests have been invaded by pathogens from areas with similar hosts and climates. But the long-term isolation of the North American forest flora from floras that are the source of pathogens might predispose the forests to invasion because susceptible genotypes could have evolved in the absence of pathogens that would have constrained their survival (von Broembsen, 1989).
Pathogens of minor significance in one location can cause epidemics when transferred elsewhere. For example, races of the stem rust pathogen Puccinia graminis were blown from moderately resistant wheat varieties in Australia to highly susceptible ones in New Zealand (Watson 1970). A most noteworthy example in the United States is the recent identification of the fungus Phytophthora ramorum, which has been associated with the disease, sudden oak death. Its hosts in California include oak and numerous other plant species. Although the origin of the fungus is not known, it was probably introduced and is known to cause a leaf spot and dieback of rhododendron in Europe (California Oak Mortality Task Force 2001).
There is a close genetic relationship between pathogens and their hosts. Many plant pathogens contain avirulence genes that trigger defense responses in plants that have corresponding resistance genes. In the absence of one or both corresponding genes, the pathogen escapes recognition and is able to replicate to high levels (Keen 1990). Information about the regional distribution of pathogen populations and the frequency of different avirulence genes in those populations can be important predictors of whether a host plant species will be available (susceptible) or not (resistant). If the virulence or avirulence of a potential invading pathogen is known, the vulnerability of a plant species can be predicted rather accurately (Mekwatanakarn et al. 2000).
The likelihood of establishment of nonindigenous, phytophagous insects depends critically on the availability of host plants that foster larval development and other life stages. Crawley (1986) analyzed establishment of insects released for the biological control of nonindigenous plants and concluded that the most common reason for their failure to establish was insect-plant incompatibility. Similarly, Barbosa and Schaefer (1997) suggested that abundance of invading herbivorous insects depended primarily on the availability and quality of host plants.
The trans-Atlantic establishment of insects is facilitated by the high degree of similarity between North American and European hosts in structural, biochemical, and spectral properties (Fraser and Lawton 1994, Jones and Lawton 1991, Niemelä and Mattson 1996). That conclusion is based on the assumption that immigrant species are conservative and are less likely to attack plants that are taxonomically or chemically distinct from their native host plants. Therefore, the degree of taxonomic difference between host plants in the native range of a nonindigenous insect and potential host plants in a new habitat could be a useful predictive tool (Strong et al. 1984, Niemelä and Mattson 1996). Most European forest insects that have become established in the United States attack the same tree genera in the United States that they attack in Europe (Mattson et al. 1994). Niemelä and Mattson (1996) found that North American trees that supported the highest numbers of European invasive insects were well represented at the genus or family level in Europe; trees peculiar to North America (such as Robinia pseudo-acacia) support few European insects.
The fact that more European insects have invaded North America than the reverse could also reflect the diverse flora of North America that includes many European plants and the relatively high rate of plant extinctions in Europe. The total vascular plant flora of North America is composed of roughly 18,000 species, including 16 genera and 97 species of gymnosperms and 143 genera and 503 species of angiosperms. In comparison, there are about 12,000 vascular plant species in Europe, including eight genera and 30 species of gymnosperms and 78 genera and 256 species of angiosperms (Niemelä and Mattson 1996 and references therein). At least 20 tree genera are extinct in Europe but are still found in North America. The high numbers of congeneric and confamilial plant species in
North America presumably made it relatively easy for immigrant European species to find suitable hosts on this continent. Niemelä and Mattson (1996) have speculated that insects that relied exclusively on plant hosts that are now extinct are themselves extinct if they could not attack a confamilial relative.
Diet breadth of a nonindigenous insect is important in assessing host availability. Nonindigenous insects that are capable of feeding on a wide variety of hosts often can forage on novel hosts in their new range. Therefore, insects that are able to use many species as hosts in their native range could be less reliant on locating plants in a new range that are similar to their native diet; this lack of specificity could facilitate establishment. Simberloff (1989) pointed out, however, that the assumption that broad diet breadth confers ecological versatility and hence higher success rates for nonindigenous insects of this type is not always reflected in empirical data. For example, various insects that became established in biological control programs were often intentionally chosen for their high degree of host specificity.
Niemelä and Mattson (1996) classified tree-feeding insects of European origin that are found in North America as monophagous, oligophagous, or polyphagous, depending on whether they fed on one genus, more than one genus in a single family, or more than one family of plants, respectively. Of the nearly 400 European invaders in North America, 68% were found to be monophagous or oligophagous, and the remaining 32% were polyphagous (Mattson et al. 1994, Niemelä and Mattson 1996). It is not clear, however, whether this pattern indicates that specialized feeders are more likely to become established than generalists or that there simply were more opportunities for invasion of specialized feeders. A random sample of the British insect fauna found that 75% of plant-feeding insects were monophagous or oligophagous, and 25% were polyphagous (Bernays and Chapman 1994). If the British fauna are representative of broad patterns across Europe, invasion by polyphagous insects may be about as likely as for monophagous species.
The relation of diet breadth to host availability may have implications for nonindigenous herbivorous insects originating in other regions, such as Asia. Insects with specialized diets may be less likely to become established in a new area than insects with broad diets because they are less likely to encounter acceptable hosts. The contrast that emerges from a comparison of diet breadth of forest insects originating in Europe and Asia appears to support this hypothesis. Mattson et al. (1994) compiled host-plant and origin data for 266 forest insect species established in North America that are native to Europe and 68 species that are native to Asia. More than 50% of the Asian species were polyphagous, but, as cited above, only 32% of the European species were polyphagous. The higher degree of polyphagy characteristic of the Asian invaders may reflect lower overall floristic similarity between trees in Asia and North America than between trees in Europe and North America. This observation depends heavily, however, on the regions being compared: temperate China’s mesophytic forests share a
striking similarity to tree genera in North America (Axelrod et al. 1998 and references therein). Although data on other groups of nonindigenous insects should be examined, the pattern suggests that polyphagy is most likely to favor establishment when the flora in a new range differs substantially from the flora in the immigrant insect’s native range.
Spatial distribution and abundance of potential hosts affect the ability of a nonindigenous insect to locate suitable hosts and the probability that a pathogen will come into contact with a susceptible host (MacArthur and Wilson 1967, Shigesada and Kawasaki 1997). A host with a patchy or fragmented distribution might be less likely to be colonized by a nonindigenous species, or its distribution could reduce the movement among local populations of a nonindigenous species and thereby lead to a greater chance of extinction than a host that is abundant or continuously distributed. A patchy host population can also reduce the rate of increase of pathogen or insect populations (Hughes et al. 1997). For example, one factor that might partially account for the order-of-magnitude higher number of European forest insects established in North America than the reverse is the lower total area and less continuous distribution of European forests than of forests in North America (Niemelä and Mattson 1996). In Britain, Fraser and Lawton (1994) reported that more than 2% of the European moths that originally fed on angiosperms have begun feeding on conifers. They suggested that a continent-wide conifer afforestation program in Europe has been a major factor in this host-range extension.
Abundance and distribution of host plants may also be related to how Allee effects influence a low-density population of a newly introduced species. If individuals of a founding population must disperse widely to find suitable hosts, founding populations may be “diluted” to even lower densities. Inverse density dependence arising from reduced reproductive success or poor survival at low densities may lead eventually to extinction (Courchamp et al. 1999). Life-history traits, such as aggregation pheromones or the ability to detect host volatiles, should enhance the ability of a newly introduced species to effectively locate a suitable host. Presumably, such traits should increase the probability of establishment, although empirical tests of this assumption are lacking.
Hosts must be available to immigrant arthropods and pathogens temporally, as well as spatially. In other words, a nonindigenous arthropod must rapidly establish phenological synchrony with its potential host plants in a new range (Quiring 1992, Wood et al. 1990). Establishment of an immigrant arthropod or pathogen might be facilitated by host plants that are distributed across a wide range of latitudes or elevations and by multiple introductions of the arthropod or pathogen during the year. Insects that use endophytic cues, such as oviposition within plant tissues, might more readily synchronize with the phenology of their hosts (Wood et al. 1990). Synchronization with hosts enables insects to take advantage of ephemeral peaks in nutrient availability or foliage quality and could
also promote temporal synchronization of mate-finding or dispersal (Wood et al. 1990).
Typically, nonindigenous plants, insects, and pathogens that are accidentally introduced into a new range arrive without the complement of natural enemies— such as predators, parasitoids, parasites, antagonists, and pathogens—with which they interacted in their native range. Whether native species can expand their diet to include the immigrants and whether the newly encountered enemies will prevent establishment of the nonindigenous species are important questions.
Establishment will be less likely if the intrinsic rate of increase (r) of the immigrant is small, if the immigrant is particularly vulnerable to attack by resident enemies, or if populations of resident enemies are large (Lawton and Brown 1986). The likelihood that native enemies will exclude a nonindigenous species presumably depends on the numbers and densities of the potential predator and pathogen species, their feeding preferences, and the total number of native prey species that are of higher preference than the immigrants (Crawley 1986). All together, the more similar an immigrant species is to a native species, the more vulnerable it will be to native enemies, because the rate of attack should be high (Lawton 1990).
The likelihood that predators or other enemies in the new range will prevent establishment of a nonindigenous insect could also be influenced by the structure of the predatory “guild” in the community. Pimm (1989) suggested that when predation exerts a strong influence on prey species, an immigrant species would have more difficulty becoming established in a community with a single predator than one with several predatory species. He based that suggestion on observations of a close correlation between predator and prey species in communities with various numbers of species (Jeffries and Lawton 1984) and the assumption that the correlation reflects the influence of predators on numbers of prey, as well as the reverse.
Evidence of the ability of indigenous species to prevent establishment of nonindigenous insects in its new range is scarce, and the topic deserves much further research. Much of the evidence stems from postrelease evaluations of herbivorous insects introduced for biological control. Goeden and Louda (1976) assessed impacts of resident predators, parasitoids, and pathogens on nonindigenous herbivorous insects introduced for biological control in 23 projects. Effects of resident natural enemies varied widely; there was no discernable impact on the introduced species in five to nine projects, there was some adverse effect on the introduced species in 12 to 17 projects, and establishment was prevented in two projects.
In postrelease evaluations of biological control releases of insects, impacts of predators were generally greater than impacts of parasitoids or pathogens. Goeden
and Louda (1976) concluded that invertebrate predation, mostly by insects, sometimes had substantial impacts on the fate of introductions, whereas effects of vertebrate predators were generally minor and localized. Indigenous predators that prevented establishment or had a measurable impact on the introduced species were usually polyphagous. Effects of predators on introduced herbivorous species did not seem to be affected by whether the herbivore was an endophagous or ectophagous feeder, whether its mouthparts were haustellate or mandibulate, or whether the herbivore was univoltine or multivoltine. In one notable case, native insect predators prevented establishment of cinnabar moth (Tyria jacobaeae) when it was introduced into Australia to control ragwort (Senecio jacobaea). Insect and bird predation also prevented its establishment in New Zealand (Goeden and Louda 1976, Lawton and Brown 1986). Goeden and Louda (1976) found no examples of deliberate insect introductions that failed to establish because of indigenous parasitism or pathogens, although these enemies sometimes limited the growth of immigrant populations after establishment. Most parasitoids that had a discernable effect on a nonindigenous species transferred from a native host that was related to the target host at least at the family level. Endophagous herbivores, primarily mandibulate species, were most often attacked by resident parasitoids. Samways (1979) reported that when a sphingid moth entered an experimental plot of cassava in Brazil, resident egg and larval parasitoids prevented any moth larvae from completing development.
If resident enemies are more likely to prevent establishment of a nonindigenous species that is similar to their native prey, the converse prediction is that establishment should be more likely for species that are different from native prey species (Lawton and Brown 1986). Lawton and Brown (1986) suggested that the rapid invasion of Britain by the cynipid Knopper gall wasp exemplifies this prediction. That the Knopper gall wasp apparently has no morphologically similar native counterparts on British oaks appears to account for the inability of indigenous natural enemies of gall wasps to locate or attack it. Similarly, adelgid insects, a group of sap-feeding insects that infest conifers, have no known parasitoids in North America. That could account at least partially for the successful invasion of nonindigenous species, such as hemlock woolly adelgid and balsam woolly adelgid. The degree of taxonomic relatedness could also be important in explaining the ability of nonindigenous plants to become established in new ranges where they lack native relatives (Mack 1996b).
The limited evidence available suggests that the potential influence of resident enemies on newly arrived populations of nonindigenous insects is difficult to predict. Lawton and Brown (1986) noted that Holt’s (1977, 1984) models to predict the outcome of an introduction require knowledge of the abundance of the various potential enemies of the immigrant, estimates of their attack rates, and the rate of increase of the immigrant. Adequate information for predicting how natural enemies will influence an introduction is not likely to be available before or just after the immigration. They also pointed out that it is generally much easier to
understand retrospectively the circumstances that led to an invasion than to predict the outcome of any particular case–a conclusion with which the committee concurs and sees as applicable to the whole issue of establishment.
Nonindigenous pathogens can also face “enemies”. Populations of native leaf epiphytes or endophytes or natural communities of soil microorganisms that suppress native pathogens might also serve as antagonists to the establishment and spread of nonindigenous pathogens (Cook 1993). These microbial antagonists range from bacteria to mycoparasites. A unique group of hypoviruses has enabled substantial biological control of chestnut blight in Europe and Michigan (MacDonald and Fulbright, 1991). When virulent strains of Cryphonectria parasitica are infected with the hypoviruses, their ability to infect chestnut and reproduce is typically reduced; they become hypovirulent. The hypoviruses appear to have their origin in eastern Asia with their fungal host (Peever et al., 1998). Hypoviruses or their genetically modified variants eventually may prove useful as biological control agents for chestnut blight within chestnut’s native North American range (Dawe and Nuss 2001). Similarly, mycoviruses (called “d-factor”) that correlate with the presence of multiple dsRNA segments in the Dutch elm disease fungus, Ophiostoma novo-ulmi may attenuate the pathogenicity of the fungus. However, their overall effect on the dynamics of Dutch elm disease is unclear, as is their potential use as biological control agents (Brasier 1990).
Predators of seed and vegetative tissue cause immediate death of a plant immigrant. Seed predators can be particularly effective because plant immigrants are most likely to arrive as seeds, and of course seed production is likely to be crucial to the immigrants’ persistence in the new range. Ants, for example, have apparently blocked the establishment of some nonindigenous tree species in the Caribbean region (Little and Wadsworth 1964). But the list of seed predators is by no means restricted to ants or even insects; rodents can be such voracious seed predators of Cakile maritima that it has been prevented from expanding its range locally in California. Predation of nonindigenous plants, apart from seeds, has commonly involved seedlings. Again, the plants are characteristically small and unable to withstand even a single attack (Mack 1996a).
The action of grazers–organisms that remove plant material in one or more nonlethal events–can be cumulative to the point at which the plant dies outright or dies from the infections that grazing can facilitate. Some termites, which as a group characteristically attack only dead wood, can attack living wood. Their attacks have been so severe for some nonindigenous plants as to thwart their establishment. Eucalypts, which are native to Australia and New Guinea, have been prevented from establishing in some locales in Brazil and West Africa by chronic termite grazing. And grazing can contribute indirectly to extirpation by reducing the ability of immigrants to survive competition or parasitism (Mack 1996a).
Perhaps the most frequent biotic constraint imposed on plants in a new range has been attack by pathogens. There are spectacular examples in which a nonindigenous plant species has been destroyed by an indigenous parasite. In such cases, establishment or naturalization is out of the question. The fate of cacao in West Africa is illustrative. Cacao (Theobroma cacao) is native to the Amazon Basin. It is successfully cultivated in West Africa but only if scrupulously protected from the cacao swollen-shoot virus (CSSV), an indigenous virus in West Africa that attacks native West African relatives of cacao. A native scale insect serves as the vector of the virus. Cacao, which has no natural resistance in cacao to CCSV, can be cultivated in West Africa only as long as movement of the scale insect from tree to tree is diligently prevented and a quarantine and destruction protocol for infected trees is rigidly enforced (Jeger and Thresh 1993). A further example of host range extension of an endemic pathogen to a nonindigenous plant species is now occurring in the United States. Multiflora rose (Rosa multiflora) was advancing unchecked in its range expansion across the United States from east to west until it came into contact with native roses infected by rose rosette disease (RRD). This endemic disease is caused by a yet uncharacterized agent that is transmitted by a native mite (Phyllocoptes fructiphylus). Because RRD is highly pathogenic to multiflora rose, the incidence and impact of multiflora rose is diminishing in many midwestern and eastern states (Epstein and Hill 1999).
Those examples involve spectacular action by resident pathogens, but we do not know how often such phenomena occur. Some generalist parasites thwart establishment of nonindigenous species. Texas root rot fungus, Phymatrotrichum omnivorum, is a generalist soil parasite that has remained indigenous in the U.S. Southwest and adjacent Mexico. It attacks at least 2000 plant taxa in more than 40 families and is highly virulent in many of these hosts. The parasite attacks so many nonindigenous woody ornamentals that it is a major deterrent to the introduction of woody horticultural species in the Southwest. Some species, such as Ulmus americana, will not persist in the range of Texas root rot, because of their vulnerability to it (Mack 2002).
Hypothetically, establishment of nonindigenous insects could be affected by interference or exploitation competition with resident species. In interference competition, one species reduces the fitness of another through an action, such as fighting or allelopathy, that is not directly related to resource availability or abundance. Exploitation competition occurs when the rate of resource availability or supply determines the rate of change among populations of different species. Resource availability to the immigrants will depend on the standing crop of the resource (which is a function of the feeding behavior of resident competitors), the productivity of the resource, and the rate at which the resource is removed by the
resident competitor (Crawley 1986). Immigrants should have a greater probability of establishment if the rate of resource availability is adequate or the number of competitors is low enough to permit the immigrant population to be maintained. Crawley (1986) proposed that when insect guilds are structured by interference competition, invasion by a larger or fiercer species is more likely. Conversely, a small species could more likely become established within a guild structured by exploitation interference in which it can reduce the resource supply rate to the point where larger, resident species can no longer be supported.
Although those predictions are intuitively appealing, it is difficult to find situations where competition with native residents has directly affected establishment of nonindigenous insects. Levins and Heatwole (1973) introduced a Drosophila species, an ant, a snail, frogs, and lizards to a small Puerto Rican island. The Drosophila species, the frogs, and the snail went extinct rapidly because of severe weather. The introduced ant and lizards survived for a while but eventually went extinct. The authors attributed extinction of the ant to competition with aggressive native fire ants and extinction of the lizards to competition and predation by resident species.
Introductions of herbivorous insects for the biological control of nonindigenous plants often provide little information, because interspecific competition is, by careful prerelease evaluation, minimal in such situations. Some have noted that the populations of entomophagous species for biological control introduced earliest are more likely to establish than populations introduced later (Tallamy 1983, Ehler and Hall 1982). In some cases, that outcome could reflect competitive exclusion of later introductions by previously established species. A more probable explanation, however, is that the order of species’ introductions is determined by the expectations of those importing the species; that is, biological control practitioners attempt to establish the species that are most likely to be rapidly successful (Keller 1984, Simberloff 1989).
How and whether competitive abilities influence establishment of nonindigenous insects have attracted much speculation, particularly with respect to asymmetry in the establishment of insects between two regions or countries. Vermeij (1991) proposed that the prevalence of asymmetrical exchanges of species between regions could be related to differences in the competitive abilities of immigrants originating in different donor ranges. European insects, for example, have disproportionately invaded other regions (Crosby 1986, di Castri 1989, Niemelä and Mattson 1996, Simberloff 1989). Niemelä and Mattson (1996) speculated that climatic and anthropogenic disturbances in Europe shaped selection for suites of traits likely to enhance the survival of insect species in the fragmented and impoverished European forests. A lower ratio of phytophagous insect species to plant species compared with the ratio on other continents might have also intensified interspecific competition in Europe. European plant-feeders that arrive in a new range, therefore, might be inherently strong competitors. That hypothesis would obviously be difficult to test empirically, and the explana-
tion is almost certainly confounded by differences in the number of opportunities for transport (Simberloff 1989). However, the much larger number of European forest insects established in North America than the reverse—despite historical movement of plant material, lumber, and other products from North America to Europe (Niemelä and Mattson 1996)—is consistent with such a pattern. Simberloff (1989) addressed the general assumption that islands are more frequently invaded by mainland species than the reverse because mainland species are superior competitors. He pointed out that the available data, which are primarily from agricultural systems, do not support that assumption. The relatively frequent invasions of islands by mainland species could instead reflect the greater abundance of mainland species or the greater frequency of opportunities for invasion of islands by mainland species. Moreover, although competition had been documented for some guilds of insects, such as ants (Suarez et al. 1999) and pine phloem-feeders (Light et al. 1983, Poland and Borden 1998), evidence of interspecific competition among foliage-feeding and sap-feeding insects is scarce (Denno et al. 1995, Strong et al. 1984).
Plant competitors can locally extirpate an immigrant population of plants. The evidence has been indirect–the abandonment of a commercial planting because competition by native plants was so severe and pervasive that cultivation alone was insufficient to foster plant establishment. Several native vines apparently prevent the establishment of introduced tree species in the Solomon Islands (Neil 1984). Competition for light probably produces many more cases of such biotic constraint among nonindigenous species, although no specific search for examples has yet been attempted. Closed-canopy forest communities in the United States, dominated by angiosperms or conifers, have much lower numbers of naturalized species than the same sites once the canopy is removed. For example, although the naturalized and adventive flora of New England probably exceeds 800 species (Seymour 1982), few (for example, Ailanthus altissima) are naturalized in New England forests. The rest occur only in undisturbed open communities and in forest habitats in which the canopy has been removed, such as sites logged or long held in cultivation. Although shade would not provide a strong barrier to nonindigenous species that are shade-tolerant, shade-tolerant species either have been introduced infrequently or, more likely, other physical or biotic factors have so far constrained their establishment. Nevertheless, nonindigenous shade-tolerant species constitute a functional category of species whose careful monitoring appears warranted upon their introduction in the United States.
Another difficulty in addressing how competition affects establishment arises because we cannot necessarily distinguish between effects of natural enemies and of competitors when a nonindigenous species arrives. An immigrant species and a native prey species can exhibit “apparent competition” through a shared enemy— a concept similar to the competition for enemy-free space described by Jeffries and Lawton (1984). The outcome of such competition can depend on r, the intrinsic rate of increase of the immigrant, and the rate at which the nonindigenous
species is attacked by resident enemies (Holt 1977, Lawton and Brown 1986). Establishment should be more likely when r is high, but interactions between the diet breadth of the native predators and the specificity of their foraging niche make it difficult to test this hypothesis (Lawton and Brown 1986).
Vectors and Mutualists
Some pathogens are transported by insect vectors to potential hosts (Nault 1977, Harris and Maramorosch 1980, Tolin 1991) and a dramatic increase in the incidence of a disease, whether caused by a native or nonindigenous pathogen, can indicate the arrival of a nonindigenous vector. For example, Xylella fastidiosa, the bacterial agent of Pierce’s disease, has been known to occur in vineyards in California since the 1800s and is transmitted by leafhoppers, but the recent arrival of the glassy-winged sharpshooter (Homalodisca coagulata), which feeds on several other important hosts (including almond, citrus, alfalfa, and oleanders), has dramatically increased the potential threat of Pierce’s disease to the agricultural industry in California (Purcell 2000). The abundance, spatial distribution, and temporal availability of the vector can, therefore, affect pathogen establishment.
Nonindigenous plants are autotrophs, but their ability to establish can depend on the presence of another species, a mutualist, in the new range. Mutualists are pairs of species for which association brings mutual benefit. In the context of nonindigenous species’ entry and survival, only one or neither species may be native to the new range (Simberloff and Von Holle 1999, Richardson et al. 2000). The scenarios by which this dependence can occur are well known and begin with the absence of an obligate pollinator. For example, nonindigenous Ficus species, each of which requires a single species of wasp for pollination, were not deemed threats to establish in the United States until the obligate pollinators for three of the species were detected in southern Florida in the 1990s. Seedlings have since germinated. Clearly, the absence of the obligate pollinator was the only constraint to the establishment and spread of these species in southern Florida (McKey and Kaufmann 1991).
As with pollination, the degree of restriction to one or a few hosts and fungal mutualists varies widely. Some fungi appear capable of infecting many hosts or forming symbiotic relations with many hosts; some hosts appear able to form a mutualism with a wide array of fungi. Alternatively, some species are much more restricted; pines are perhaps the most well-known examples. The genus does not have native members in Australia, South America, or all but a northern strip in Africa (Critchfield and Little 1966), but all three continents contain climates that can support pines. Once soils on those continents were inoculated with appropriate mycorrhizal fungi, pine establishment was ensured. In fact, pines have become invasive locally on all three continents (Richardson et al. 2000, Richardson and Higgins 1998).
The spread of pines in the Southern Hemisphere is germane to the assessment of the causes of persistence because nonindigenous pines have rarely become established in the United States. Naturalized pines are largely restricted to Pinus sylvestris in New York State and Pinus nigra (Leege and Murphy 2001) on dune sets around Lake Michigan. The current restriction of pines is curious for at least four reasons. First, the conterminous United States has about 36 native pines that occur in a wide range of physical habitats (Critchfield and Little 1966), from some of the most arid habitats in which trees occur to sites with abundant moisture and from sites with air temperatures that routinely exceed 35°C to the upper timberline in the Appalachian, Cascade-Sierra Nevada, and Rocky Mountain ranges. Second, a long history of deliberate introduction of pines continues. Pinus sylvestris, P. mugo, and P. nigra are common horticultural species. Third, with so many native pines that all maintain associations with one or more native mycorrhizal fungi, it would be surprising if host extensions among the fungi to the introduced pines had not occurred. Finally, as with the inadvertent introduction of the requisite fungi in the Southern Hemisphere with potted pine seedlings from the native range, it would be surprising if similar introductions had not occurred in the long history of pine cultivation in the United States. The further naturalization and even invasion by foreign pines in the United States remain possibilities that should be experimentally evaluated.
LIFE HISTORY TRAITS
Thus far in this chapter, two major categories of factors that influence the process of establishment have been presented: stochasticity and environmental (abiotic and biotic) forces. However, no factor related to predicting the establishment of nonindigenous species has been pursued more assiduously or longer than a link between the life-history traits of species and their ability to become established in a new range. The reason is obvious: life-history traits are related directly to species growth, reproduction, and survival. The value of any such links is appealing and deserving of much further investigation. Having failed, however, to find broad taxonomic agreement between life-history traits and species’ performance in new ranges, we outline here the traits that appear to play a role in influencing establishment for plants, pathogens, and insects.
Although plants collectively display much diversity in reproductive systems, two fundamental dichotomies are apparent. First, reproduction can be sexual or asexual; and second, sexual reproduction can involve a single parent (uniparental) or two parents (biparental). Biparental sexual reproduction, often termed out-
crossing, is the predominant form of reproduction in animals. It is also predominant among plants because of self-incompatibility mechanisms that promote crosses between individual genotypes and because of structural limitations, such as the separation of anther maturation and stigma receptivity in time (dichogamy) and space (herkogamy). Among dioecious species (in which separate plants are male or female) and those with self-incompatibility, pollen must be transferred from plant to plant if fertilization is to be achieved. Pollen flow might not occur in the absence of a suitable pollinator or if an individual is growing far from potential mates (Willson 1983). Dioecy does not, however, appear to be a major limitation for species establishment; there are dioecious invasive plant species, such as Rumex acetosella, Ailanthus altissima, and Ilex aquifolium.
Uniparental sexual reproduction arises from self-fertilization and is facilitated by hermaphrodite sex expression. As long as the flowers are self-compatible, plants with perfect flowers might have the ideal mating system for establishment in a new range. Species that usually display some form of uniparental reproduction also have an advantage as founders because the lack of recombination with other plants preserves multilocus genotypes related to increased fitness.
In contrast, monoecious species, which have separate male and female flowers on the same plant, might face the same limitation in pollination faced by dioecious species. Nevertheless, this limitation has clearly been overcome in some species, such as several pines (Richardson and Higgins 1998). Many perennial species that have been introduced into North America have the capacity for asexual reproduction by apomixis. Apomixis, which includes agamospermy and the clonal or vegetative regeneration of plant parts, allows isolated individuals to establish new populations and produce plants that are presumably adapted to the current environment. Agamospermy allows a plant to produce viable seed, often without the presence of any male gametes; this is an advantage of an isolated individual. In a comparative study of woody plant invaders and noninvaders, agamospermy was found to be slightly correlated with species that had become invasive (Reichard 2001).
Many plants have the ability to regenerate from a stem or root fragment or to resprout from a cut stem (Bell 1991). Vegetative reproduction of this type allows a population to increase rapidly and to regenerate quickly after a trauma. If the fragments are dispersed, as during a flood, distributed populations can be created. There are many examples of invaders for which most or all reproduction in the new range is the result of asexual reproduction. Clonal propagation is especially prevalent among invasive aquatic plants (reviewed in Barrett 1989, Barrett et al. 1993). In some cases, the failure to reproduce sexually occurs because of genetic sterility (as in Salvinia molesta) or the absence of mating types required for sexual reproduction (as in Elodea canadensis); alternatively, restrictions on sexual reproduction might arise because of unfavorable environmental conditions in the introduced range (as in Eichhornia crassipes).
The relation between reproductive systems and a population’s ability to become established appears linked to the association between self-fertilization and colony establishment (Marshall and Brown 1981, Barrett 1982, Gray 1986, Brown and Burdon 1987). Single self-compatible individuals, with the capacity for autonomous self-pollination (“selfers”), are capable of forming established populations through self-fertilization, whereas self-incompatible or unisexual individuals require the simultaneous arrival of mating partners and pollen vectors (in animal-pollinated species) for reproduction. That simple idea, known as Baker’s law (Baker 1955, Pannel and Barrett 1998), states that self-compatibility is favored among immigrants after long dispersal. It has also led to the prediction that annuals and nonindigenous ruderals (plants that commonly occupy rubbish piles and areas that are frequently disturbed), which depend on recurring dispersal and establishment, are more likely to be selfers than obligate outcrossers. Broad surveys generally support the association between self-fertilization and ruderals (Mulligan and Findlay 1970, Price and Jain 1981), although this pattern is less evident among perennial ruderals, many of which invest heavily in clonal offspring and are also outcrossing (Marshall and Brown 1981, Crawley 1987). Among flowering plants, increased longevity is generally associated with decreased selfing: annuals display the highest selfing rates, followed by herbaceous perennials. Woody perennials are predominantly outcrossing; few woody species are reported to have high levels of selfing (Barrett et al. 1996).
Surveys of plant invaders show that those which rely on sexual reproduction are not all selfers; this indicates that some outcrossers overcome this constraint during establishment. Pannel and Barrett (1998) evaluated the benefits of reproductive assurance in selfers compared with outcrossers in the context of colony formation in a metapopulation. Their results suggest that an optimal mating system for a sexual invader should include the ability to alter selfing rates according to local environmental and demographic conditions. When populations are small or individuals are at low density during the early phases of establishment in a new range, plants should be selfers to maximize fertility, thus increasing population growth rates. However, when populations become large and pollinators or mates are not limiting, outcrossing and its attendant genetic effects will be more beneficial.
Flowering and Fruiting Periods
A long flowering time ensures that a plant’s flowers are receptive when pollinators are available. If a plant has a protracted season of flowering, the probability of fertilization is increased. Annual ruderals in Great Britain (Perrins et al. 1992a) and invasive woody plants in North America (Reichard 1994) have been shown to have a long flowering period. The latter study also showed that a long flowering time correlates highly with the length of the fruiting period. Similar to the case with the length of the flowering period, a long fruiting period
can provide a greater opportunity for seed dispersal. So far, however, these traits have been of little use in predicting which species will be invaders across the broad taxonomic groups of potential immigrant species.
The juvenile period is the time from seed germination to the onset of flowering. A short juvenile period could allow a population to increase rapidly while decreasing the probability of the population’s detection by predators, foragers, and pathogens before its sexual reproduction. Annual plant species, by definition, have a short juvenile period; ruderal or weedy annuals have a shorter juvenile period than nonweedy species (Perrins et al. 1992b), as do invasive woody perennials (Reichard 1994). Length of the juvenile period can be difficult to determine accurately for woody species, for which several years can elapse before onset of reproduction. Length of the juvenile period has been used repeatedly to predict persistence of nonindigenous species (Rejmanek and Richardson 1996, Reichard and Hamilton 1997, Pheloung et al. 1999).
Species with high seed production on an annual or cyclical basis are more likely to become established if their seeds are readily dispersed (Juenger and Bergelson 2000), because the odds are greater that some fraction of the seeds will reach sites suitable for germination. But incorporating this trait into any prediction of establishment is difficult because the level of seed production can be difficult to quantify for newly introduced species, especially woody plants.
Given the influence of environmental stochasticity on establishment, there should be strong selection for life-history traits that reduce its impact on a population. Seed banks buffer against the wide swings in the size of the vegetative population that result from strong year-to-year variation in resource availability (Cohen 1979, Venable and Lawlor 1980, Brown and Venable 1986, Phillipi 1993). Germination cuing, in which environmental signals correlated by habitat quality are sensed by seeds and trigger germination or induce dormancy, is a potentially important form of adaptive dormancy in plants. Given the potential for seed dormancy and germination cuing to reduce the demographic impacts of environmental stochasticity, it is not surprising that both dormancy and germination cuing are widespread in many agricultural (and predominantly nonindigenous) weeds (Cousens and Mortimer 1995).
The ability to use light efficiently may enhance a plant’s ability to live in areas with extensive canopies. Consequently, the ability to establish may be related to this trait. The extent to which shade tolerance is an attribute shared widely among naturalized plants is unknown but deserves systematic survey. Baruch et al. (2000) examined 10 physiological and morphological plant traits of four invasive members of the Melastomataceae (two herbs, a shrub, and a tree) in Hawaii and found that the invasive species were better suited to capturing and using light than a large group of natives.
Life-history traits important for the establishment of plant pathogens include reproductive strategies and genetic variability related to fitness, virulence, and host compatibility.
As with plants, pathogens use many reproductive strategies. Some pathogens (such as viruses and some fungi) reproduce only in the presence of their hosts, whereas others (for example, many fungi) are facultative saprophytes and do not require the plant host for reproduction. Some fungi reproduce only sexually, whereas many pathogens, such as viruses and some fungi, reproduce only asexually. Some pathogens can complete several generations in a single year, whereas others require several years to complete a single generation.
Asexually reproducing pathogens are thought to establish most easily (Agrios 1988). Another major characteristic of invasive pathogens is a high rate of survival when the plant host is not present (for example, in the winter for pathogens that infect the leaves of annual plants and deciduous perennials) or when the physical environment is totally unfavorable. Survival can occur in a dormant state (for example, in overwintering spores), in a saprophytic condition, or as infections in alternative hosts (Agrios 1988).
The most successful pathogens display a short time between one infection cycle and the next, have a high rate of production of infectious units (spores, bacterial cells, nematode cysts or eggs, or viruses), and have a long infectious period–the time that infectious units are produced or plants are contagious (Campbell and Madden 1990). Races of Puccinia helianthi, the sunflower rust pathogen, illustrate this point. The rust’s superior colonizers have higher spore germinability, more rapid spore germination, more rapid formation of appressoria (spore-producing structures), and higher spore production than other genotypes or races (Prudhomme and Sackston 1990).
Genetic Variability in Fitness and Virulence
The characteristics described above might increase the potential for establishment, provided that the pathogens are carried to a genetically compatible host. Although it could be predicted that successfully established pathogen populations would have a wide range of virulence genes (Lawrence and Burdon 1989), this is not routinely the case, particularly in an agricultural context. Pathogen populations typically have the smallest number of virulence genes needed for survival, presumably because carrying unneeded virulence genes imposes a fitness penalty on the pathogen. Common, widespread pathotypes of Pyricularia grisea, the fungal cause of rice blast, always have fewer virulence genes than rare pathotypes–an observation that is consistent with the theory of a fitness disadvantage of accumulated virulence genes (Mekwatanakarn et al. 2000).
It is, however, the continuous generation of novel pathogenic variation that enables pathogen populations to overcome resistance and find susceptible hosts (Mekwatanakarn et al. 2000). A large local effect on the diversity of races and complexity of virulence occur in a pathogen population as it changes in response to host resistance (Andrivon and de Vallavielle-Pope 1995). The importance of variability in virulence for establishment tends to be greater among pathogen populations with narrow host ranges. High complexity for virulence is often found in clonally (asexually) reproducing pathogens, such as Pyricularia grisea tritici, that have long faced race-specific host-resistance genes (Marshall 1989; 1993). Conversely, populations of fungi with frequent sexual reproduction, such as the powdery mildews, often have fewer races but a greater diversity of virulence phenotypes (Roelfs and Groth 1980, Groth and Roelfs 1982).
Because of the local effect of selection by host resistance on the diversity and complexity of the pathogen population, immigrations by new genotypes are often detected on the basis of abrupt changes in pathogenicity or mating type. That was the case in Europe with Phytophthora infestans, the late blight disease pathogen (Fry et al. 1992). Some migrations are cryptic, especially if the virulence pattern of a pathogen population has not been monitored. However, information about the genetic diversity in pathogen populations can also help to reveal new introductions. For example, Sphaeropsis sapinea is an asexually reproducing fungus on Pinus spp. in South Africa, so it could be predicted that its population should consist of clonal lineages. But surveys revealed populations of high genetic diversity, a situation consistent with the occurrence of multiple introductions from different sources over a long period (Smith et al. 2000). Finally, in some wild host-pathogen combinations, extinction and recolonization occur routinely, and these outcomes suggest that migration and gene flow are important contributions to the genotype diversity of the pathogen (Burdon et al. 1995).
Rate of Population Growth
The net effect of whether a pathogen will come into contact with a susceptible host and the chance that such a contact will result in infection are summarized by the intrinsic rate of increase (r) or, more mechanistically, the basic reproduction number (R0), sometimes called the progeny-parent ratio (Madden et al. 2000, Swinton and Gilligan 1996). A pathogen will be established only if R0 is greater than a threshold, which is typically 1.0 (higher for small founder populations subject to the effects of stochasticity). An adequate way of determining a threshold of R0 for persistence over more than one growing season, which incorporates survival and within-season dynamics, has not yet emerged, because of the difficulty of incorporating temporal discontinuities (periods without a susceptible crop) into the predictions (Gubins and Gilligan 1997, Madden and van den Bosch 2000).
Insects and other arthropods usually arrive in a new range in small numbers and must reproduce and increase in density rapidly if they are to become established. Parthenogenesis and other forms of uniparental reproduction, such as mother-son or sibling mating, can facilitate survival of low-density populations or populations surviving in small refugia. Parthenogenesis results in a relatively high ratio of reproductive potential for each unit of resource, enabling nonindigenous organisms to exploit a resource rapidly when ephemerally favorable conditions arise (Niemelä and Mattson 1996). The search for mates can impede establishment if the loss of individuals dispersing to search for mates exceeds the rate of population growth (Lewis and Kareiva 1993). Parthenogenic forms of reproduction can reduce or even eliminate the need to locate mates.
Uniparental reproduction can facilitate the survival of a small population, but inbreeding depression can eventually become problematic in such populations, depending on the rate of mutation or the rate at which new, unrelated individuals join the population. However, populations of parthenogenic insects that are highly inbred might exhibit little inbreeding depression in fitness if most individuals with deleterious recessive alleles are lost. In addition, parthenogenesis is often linked with polyploidy and high heterozygosity, facilitated by apomixis. These traits presumably confer broad ecological tolerances for new and varying environments (Bullini and Nascetti 1990, Craig and Mopper 1993, Niemelä and Mattson 1996).
Parthenogenesis does appear to be frequently associated with establishment of nonindigenous insects. A large fraction of the nonindigenous invertebrates that became established in Hawaii are parthenogenic or hermaphroditic (Howarth
1985 cited in Simberloff 1989). In Europe, parthenogenic polyploid species of weevils have a much greater range than ancestral species that retain bisexual, diploid reproduction (Suomalainen et al. 1976), and several of these polyploid species have become established in North America. Parthenogenesis is more common among European than North American phytophagous insects and might partly explain the asymmetrical proportion of European invaders around the world. For example, roughly 40% of nonindigenous tree-feeding insects exhibit some form of parthenogenesis, compared with an estimated 11% of native tree-feeding insects in North America (Niemelä and Mattson 1996).
Several parthenogenic insect taxa have been particularly invasive in North America. At least 45 species of Coccoidea scales are established on North American woody plants (Mattson et al. 1994, Niemelä and Mattson 1996); these scale insects are characterized by several types of parthenogenesis and collectively display the most diverse chromosome system of any animal group (Kosztarab 1987). All 60 of the nonindigenous sawfly species and all 23 aphid species established on North American trees and shrubs are parthenogenic (Smith 1993, Niemelä and Mattson 1996). At least 65% of the 33 nonindigenous species of bark beetles (in the family Scolytidae) are facultatively parthenogenic (Atkinson et al. 1990). Establishment of parasitic hymenopterans introduced for biological control is probably favored by reduced inbreeding depression arising from haplodiploidy and the ability to adjust sex ratio according to population density, host condition, or other factors (Simberloff 1989).
Although parthenogenesis is associated with some insect invaders, this trait is not common to all invasive insect taxa. Niemelä and Mattson (1996) compiled data for eight dominant taxa of nonindigenous insect herbivores of woody plants established in North America. Two insect families—including leaf hoppers (Cicadellidae), plant bugs (Miridae), and one moth family (Tortricidae)—have no parthenogenic species. Thus, parthenogenesis or other forms of uniparental reproduction may contribute to the establishment of some taxa but not all.
Rate of Population Growth
The likelihood of establishment of nonindigenous insects has long been assumed to be related to the intrinsic rate of population growth, r. It has been argued that the relative amplitude (coefficient of variation) of a population’s fluctuations is the most important variable affecting the average lifetime of that population (Leigh 1981) and that a high r can reduce the chance of extinction in a founding population (Lawton and Brown 1986). In a review of biological control projects Crawley (1986) found that arthropods with high fecundity, short generation time, or female-biased sex ratios were more likely to establish than comparable arthropods with lower population growth rates. Pimm (1989) noted that across animal taxa, r is inversely correlated with individual longevity. He found that the combination of small body and high r was advantageous for the
establishment of a founder population except at very low population densities (such as six pairs or fewer), when long-lived species would be expected to have a lower extinction rate. In two other studies, insects with small bodies and high r were found to be more likely to establish than insects with large bodies (Crawley 1986, Lawton and Brown 1986).
Limitations in prediction arise, however, if estimates of r are used to predict the likelihood of establishment for a given arthropod species. One problem is that establishment can be linked to other life-history strategies. For example, in Crawley’s (1986) review of biological control introductions, insects with long-lived adult stages were found more likely to establish. Adult longevity presumably enabled oviposition to occur over a protracted period, increasing the probability that the nonindigenous arthropod would encounter suitable conditions for establishment. In addition, r is often considered on a relative or qualitative basis, and it is not clear how large it would need to be to enhance the probability of establishment. Furthermore, it is difficult to disassociate r from other traits, such as reproduction strategy, dispersal, and interactions with predators or other taxa in a new habitat.
The numerous factors identified in this chapter form a basis for predicting the establishment of a nonindigenous plant or plant pest. The degree of uncertainty in our ability to measure these factors depends on whether the identity of the immigrant is known, whether important information about its life history is available, and whether the circumstances of its introduction have been accurately assessed.
The likelihood of establishment of nonindigenous plants and plant pests depends in part on the number of organisms that are introduced and the frequency of the introductions. Nonindigenous plants and plant pests typically arrive in small numbers and are vulnerable to demographic, environmental, and other stochastic forces that drive small populations to extinction.
The chance of extinction due to demographic stochasticity is a function of the number of immigrants, their reproductive rate, and, if sexually reproducing, their success in finding mates. Populations of plants and arthropods of fewer than 50 individuals are highly vulnerable to extinction.
Genetic bottlenecks, small size of the founder population, and strong directional selection on immigration can reduce the probability of establishment. Inbreeding in low-density populations can reduce the fitness of progeny. There are cases, however, where reduced heterozygosity resulting from a genetic bottleneck has enhanced the success of a nonindigenous species. Effects of such
genetic changes can profoundly alter behavior and social organization and alter ecological interactions between nonindigenous and native species.
Environmental stochasticity can reduce populations to the level at which demographic stochasticity becomes important. Models derived for plants suggest that when deterministic growth rates of newly arrived species exceed 1.2, effects of environmental stochasticity will be reduced and extinction will occur only as a result of extreme environmental events.
Weather, the random expression of the amplitude of climate, is an important source of adverse environmental stochasticity.
Natural catastrophes—such as fires, floods, and earthquakes—are difficult to predict but can cause the extirpation of populations of less than several thousand individuals.
An Allee effect arising from low reproductive success or survival in low-density populations can strongly influence the ability of a newly arrived, nonindigenous species to persist. Whether a founder population persists or is driven to extinction can depend on the strength of this inverse density dependence, the population process that is subject to the Allee effect, and interactions with environmental stochasticity.
Phenotypic plasticity, including acclimation, may buffer populations from environmental stochasticity.
Human-generated disturbance can reduce populations to the point where demographic stochasticity causes extinction. But cultivation, whether deliberate or inadvertent, can promote persistence of nonindigenous plant pests by increasing resource availability and decreasing environmental stochasticity.
Spatial distribution of newly arrived populations can affect the influence of stochastic forces. Small populations that are restricted to a specific habitat or host are more susceptible to extinction from stochastic forces than populations distributed across a large geographic area or populations that occupy multiple habitats or infest several hosts.
Demographic, environmental, and other stochastic forces can be overcome by repeated introductions of a species that increase its population number, by introductions that spatially distribute the population, and by the life-history traits (such as diapause and dormancy) that minimize the consequences of stochasticity.
Estimates (based on interception or other data) of the size of an immigrant population, the frequency of introduction, and opportunities for it to be introduced in multiple locations could be useful in determining the likelihood that a nonindigenous population will become established.
The geographic distribution and range of climates known to be suitable for the immigrant species in its native range or in previously invaded regions provide some indication of new habitats in which the immigrant population could
persist. However, commodities that may harbor pests are often distributed from their initial entry point to other areas in the United States, which together encompass broad climatic variation (from arctic to generally tropical climates).
Newly arrived organisms must be in seasonal synchrony with conditions in the new range. Nonindigenous organisms must enter and break diapause or dormancy at appropriate times. Among arthropod introductions in the Northern Hemisphere, species from northern latitudes that are introduced into southern latitudes may be more likely to become established than vice versa.
Availability of suitable host plants is a critical factor in the establishment of nonindigenous arthropods and pathogens. Taxonomic similarity (at the genus or even family level) between host plants of insects or pathogens in the native and new ranges increases likelihood of establishment, although this relationship requires further testing.
Broad diet breadth may enhance the likelihood of a phytophagous arthropod’s establishment, especially if the flora in the new habitat is phylogenetically distant from flora in the native habitat. Most nonindigenous insects known to be established in North America, however, have specialized diets.
Host plants must be temporally and spatially available to newly arriving nonindigenous insects and pathogens for their establishment in the new range.
A vector is necessary for the establishment of some pathogens; in these cases, abundance, spatial distribution, and temporal availability of the vector will affect establishment.
Natural Enemies and Competitors
The presence of competitors and natural enemies in the new range of a nonindigenous plant or plant pest may prevent its establishment. Effects of these biotic forces on new species are difficult to predict, however, and often require detailed ecological information that is rarely available.
Immigrant arthropods or pathogens without a morphological counterpart or close familial relative among the native species could have an advantage in establishment if native natural enemies do not attack them.
Native insect predators are more likely than native parasitoids to prevent establishment of new arrivals of herbivorous insects. Effects of entomopathogens on new arrivals of insects are largely unknown.
There is little evidence that competition with native species has prevented establishment of new arrivals of insects or pathogens, although this could theoretically occur.
Competition, especially for light, can substantially affect the establishment of nonindigenous plants.
Selected life-history traits are frequently associated with persistent nonindigenous species and may be useful in predicting or assessing the likelihood of establishment of a given species. However, many exceptions occur for any given trait, and most can be evaluated only subjectively or qualitatively. Quantitative comparisons between the traits of species that have become established and the traits of species that have failed to establish are rare, especially for arthropods and pathogens. Nevertheless, there are traits that appear to enhance establishment, and these require much further study.
The ability to change from outcrossing to selfing in response to local environmental conditions could optimize the opportunity for establishment. Species with high phenotypic plasticity among many ecologically important traits (for example, the traits collectively considered in connection with phenology) also have an advantage in a new environment. Possession of a resistant dormant phase, particularly a resistant seed bank, appears important, as do alternative forms of asexual and sexual reproduction, rapid growth, and high fecundity.
Nonindigenous plant pathogens with genetic variability in traits associated with reproduction have a higher probability of establishment. Traits of plant pathogens that appear to enhance establishment include a short infection cycle, high productivity of infectious units, and a long infectious period.
High intrinsic rate of increase, uniparental reproduction, and a dormant or resilient life stage that permits surviving temporally unfavorable conditions characterize many insect invaders. Other life-history strategies, such as long-lived adult stages, are common among established nonindigenous insects.