* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
Download Rapid displacement of native species by invasive species: effects of
Ecological fitting wikipedia , lookup
Occupancy–abundance relationship wikipedia , lookup
Biodiversity action plan wikipedia , lookup
Latitudinal gradients in species diversity wikipedia , lookup
Reconciliation ecology wikipedia , lookup
Habitat conservation wikipedia , lookup
Island restoration wikipedia , lookup
Introduced species wikipedia , lookup
Biological Conservation 89 (1999) 143±152 Rapid displacement of native species by invasive species: eects of hybridization Gary R. Huxel* Department of Environmental Science and Policy and Institute of Theoretical Dynamics, University of California, Davis, CA 95616, USA Received 23 March 1998; received in revised form 24 October 1998; accepted 4 November 1998 Abstract The introduction of non-native populations can lead to the competitive exclusion (displacement) of native populations. This has been hypothesized to be further exacerbated by the potential of hybridization, which can dilute or genetically assimilate the native genotype leaving no ``pure'' natives. With relatively moderate to high rates of immigration, the loss of the native species can be rapid with or without hybridization. Using single-locus, two-allele models, I ®nd that species replacement can occur very rapidly and the time to displacement decreases rapidly with increasing immigration and selection dierential. Immigration and selection act in two dierent ways: increasing immigration results in displacement by overwhelming the native; whereas increasing the selection dierential in favor of the invader leads to displacement via genetic assimilation. The implications of these results are the need for more empirical studies on the immigration patterns of invasive species and their potential for interbreeding with natives. # 1999 Elsevier Science Ltd. All rights reserved. Keywords: Introgression; Species displacement; Hybridization; Genetic assimilation; Invading species 1. Introduction Hybridization is a common phenomenon in plants, birds, ®sh, and many other taxa (Mayr, 1942; Grant, 1981; Harrison, 1990; Grant and Grant, 1992, 1996; Dowling and DeMarais, 1993; Levin et al., 1996; Rhymer and Simberlo, 1996; Williamson, 1996; Arnold, 1997). Yet in studying the displacement of native species by invading species the potential for interbreeding between an invading species and a native species is often ignored or understated. It has been suggested that interbreeding increases the threat of extinction for a number of species due to hybridization introgression (Levin et al., 1996; Rhymer and Simberlo, 1996). For example, ``pure'' native Pecos pup®sh may no longer exist due to introgression with an introduced bait ®sh, the sheepshead minnow (Echelle and Connor, 1989). Conditions under which hybridization is expected to increase or decrease the rate of species displacement has not been explicitly examined. In cases where related taxa are able to interbreed, introductions may lead to the introduced taxa dying out, coexistence, new hybrid * Tel.: +1-530-752-5162; fax: +1-530-752-3350; e-mail: grhuxel@ ucdavis.edu. taxa or extinction of the native taxa. Although theoretical studies have addressed the introduction of new alleles into populations (i.e. the third phase of Wright's shifting balance theory), few have focused on the displacement of native taxa (alleles). Here, I use single locus, two-allele (one native and one non-native) to examine the in¯uence of interbreeding with invading taxa on native taxa. Historically, geographical isolation has been a major factor in limiting the impacts of hybridization and introgression. Today, however, taxa are being brought into contact with related taxa from which they have been isolated through many anthropogenic pathways (Carlton, 1979, 1989; Carlton and Geller, 1993; Williamson, 1996). For example, the movement of ballast water throughout the world is increasing the opportunities for closely related taxa to interbreed leading to homogenization of near-shore marine communities worldwide and the widespread loss of species (Carlton, 1979, 1989, 1996; Carlton and Geller, 1993). Similar large distance (inter-regional), short time scale immigration occurs with the introduction of nurse stocks of plants and the incidental movement of herbivores and parasitoids (Heywood, 1989). Furthermore, introduction of bird species has been extremely widespread 0006-3207/99/$ - see front matter # 1999 Elsevier Science Ltd. All rights reserved. PII: S0006-3207(98)00153-0 144 G.R. Huxel / Biological Conservation 89 (1999) 143±152 throughout the world (Moulton and Pimm, 1986; Moulton, 1993; Case, 1996). Human-mediated introductions can be frequent and re-occurring, greatly increasing immigration rates (Carlton and Geller, 1993). In the absence of hybridization the interaction between the native and introduced species is essentially a competitive one between species that are similar in many aspects of their ecology and life histories. Assuming a homogeneous environment, without interbreeding, the species with the greater ®tness (e.g. better competitor or higher reproductive rate) will dominate at equilibrium. Switching from discussing individuals, I now focus on interactions between native and non-native genotypes. If immigration rates are very low and given no (or a weak) selection advantage to the invader, then one would expect that Allee or stochastic eects may lead quickly to elimination of the invader as its population density remains low. At very high rates of immigration, Kondrashov (1992) has shown that immigration eects alone in the absence of interbreeding can cause the loss of the native species. Alternatively, large ®tness dierentials, alone, lead to the ®xation of a favored genotype, therefore the relatively slow process of selection acting without immigration can be quickened by increasing the ®tness dierential. For example, intermediate genotypes can be made lethal for a population with a large, but ®nite, population size so that the inferior parental genotype is rapidly lost (Kondrashov, 1992). In the presence of hybridization, one would expect costs associated with genetic assimilation and/or bene®ts of infusion of new genes into one or both parental taxa. However, hybrids can be fertile yet reproductively isolated from adults for various reasons including fertility selection (allopolyploidy, translocation heterozygosity, recombination speciation, species-speci®c dierences in mitochrondial DNA) and pre-mating factors (behavioural changes, reproductive phenology, niche separation). Interbreeding may also allow rare genotypes to become established and then increase as backcrosses with the non-native parental taxa or interbreeding between hybrids occurs. This of course, is simple in the single locus models used here, but is more complicated in real situations involving multiloci (see below). Large selection advantages for non-native alleles should result in more rapid displacement of native alleles with hybridization than without. While studies have shown that lowering the ®tness of hybrids can result in displacement (Kondrashov, 1992), the question remains as to whether the rate of species displacement increases or decreases with increased interbreeding between an invading genotype and a native genotype. Genotypic displacement occurs when a native species becomes extinct and is replaced by an invading species. By ®tness, I mean the relative competitive ability expressed as the number of viable ospring produced compared with the number produced by other genotypes. Thus, the interesting question in terms of the conservation of species, and the main question of this paper, is whether hybridization enhances or impedes the rate of species displacement and whether this phenomenon is aected by introgression. I address this question using single locus, two-allele models with varying degrees of interbreeding. Additionally, I examined the in¯uence of varying the ®tness of the heterozygote from underdominance to overdominance in a system that exhibits introgression. 2. Model descriptions and analyses Let us consider two populations with non-overlapping generations. One population is native and has a ®xed allele at a given locus (BB) and the other is nonnative with a dierent ®xed allele at the same locus (AA). Non-native individuals can migrate into the native habitat but the opposite in not allowed. This may represent stocking programs or other human-mediated dispersal so that immigration rates can be much greater than otherwise found in natural systems. The population is assumed to follow Hardy±Weinberg proportions with selection immediately following immigration. This assumes no mutation, large populations and random mating. Furthermore, any competitive dierences are accounted for in ®tness dierentials. In the model, the population size is in®nite and ®xation of AA is said to occur when its frequency is greater than 0.99. Four cases of post mating isolation are modeled: (1) lethal heterozygotes; (2) sterile hybrids; (3) hybrids with backcross sterility and (4) introgression. A ®fth case was examined which also involves introgression, but the ®tness of the heterozygote is varied from underdominance to overdominance. The frequency of AA, pAA, (the introduced genotype) at time (t 1): p0AA pAA 1 ÿ m m; 1 where m is the migration rate. This assumes that migrating individuals each replace one resident individual proportionally (no preference for displacement of any genotype by immigrants). Fitness of the native species was assumed to be 1 in all models for mathematical and interpretive ease. Most theoretical studies have focussed on equilibrium analyses of introduced alleles, but I was most interested in the time to displacement. The time frame of displacement has important conservation consequences due to human-mediated immigration rates and the lack of taxonomic work in many systems (Carleton and Geller, 1993; Geller et al., 1997). Thus I allowed the systems to run assuming a large population of constant size from one generation to the next until displacement G.R. Huxel / Biological Conservation 89 (1999) 143±152 145 occurred or 500 generations had elapsed. I used the range of values for ®tness (!AA Юtness of AA in the ®rst four cases or !AB Юtness of AB in the last case) from 0 to 4.0 with step increases of 0.1 and m from 0 to 1.0 with step increases of 0.1. The initial community is comprised only of natives (pBB 1). 2.1. Lethal heterozygotes I simpli®ed the analyses by only following the introduced species in this case. If pAA is the frequency of the introduced genotype in the generation after selection and if no interbreeding can occur then p0AA !AA p2AA =b!AA p2AA m !AB 1 ÿ pAA 2 ÿ mc 2 where !AA is the ®tness of the ``pure'' invader genotype, pAA is the frequency of the ``pure'' invader genotype, m is the migration rate and !BB is the ®tness of the native genotype. This is the baseline case for the discussion, which is essentially a competitive situation where the species have remained reproductively isolated. Fixation of the invader genotype (pAA 1) always occurred when both immigration and !AA were at moderate levels [Fig. 1(a)]. Fixation also occurred when the invader had a disadvantage in ®tness (!AA < 1). For example, when !AA > 0:20, ®xation occurred whenever immigration was greater than 0.30. Under the condition of no immigration (m 0:0) the invader genotype is always lost. An interesting dynamic happens at m 0:1 which is below the critical immigration rate for which a species cannot successfully replace another without selection when !AA 1 (Crow et al., 1990; Kondrashov, 1992). As predicted with m at this level, at and below !AA of 2.0 the pAA slowly approaches 0.07 so that ®xation does not occur in 500 generations, but increasing !AA to 2.1 leads to ®xation in just 33 generations. This is typical of the dynamics with no hybridization where there exists a sharp boundary between rapid and no ®xation with 500 generations (eventually however ®xation will always occur with one way migration and !AA > 1). The frequency of AA after 500 generations was always less than 0.25 when ®xation did not occur and was usually less than 0.10 [Fig. 1(b)]. When ®xation occurred it always did so in less than 33 generations and as both !AA and m increased, this time to ®xation was reduced to less than 10 generations. The case of no interbreeding demonstrates that displacement is greatly in¯uenced by immigration and differences in relative ®tness. Time to displacement, when it occurred, was extremely rapid and decreased as m and !AA increased. Immigration rates may be increased by human activities; examples of this are found in stocking programs, in which the introduced species may have Fig. 1. The no interbreeding case; (a) time to extinction and (b) frequency of AA. These patterns hold for the sterile hybrid and backcross sterility cases as well. greater population sizes than the native species and in the absence of introgression may displace the native. This seems to be occurring in the introduction of brook trout (Salvelinus fontinalis) into the habitat of native bull trout (S. con¯uentus) in the Columbia and Klamath River drainages (Leary et al., 1993). 2.2. Hybrid sterility Next I examine the extreme case of hybridization without introgression in which the hybrids are sterile (!AB 0), but are part of the t 1 generation. The equation for the frequency of AA (pAA) is the similar to Eq. (2), however to maintain a constant population size, pAA at time t 1 one must account for the loss of hybrid individuals each generation. Therefore I assumed that the two parental lineages produce sucient numbers of ospring so that the population size is constant from one generation to the next and do so at their respective proportions after selection. For example, if pAA 0:40 and pBB 0:60 (®tness of all genotypes is set 146 G.R. Huxel / Biological Conservation 89 (1999) 143±152 to 1 and m is set to zero for this example), random mating results in pAA 0:16, pAB 0:48, and pBB 0:36. As the heterozygotes are sterile, the frequency of A for contribution to the following generation is 0.32/(0.32+0.72) or 0.307. Thus, the only dierence between this model and the no interbreeding case is that some reproductive eort is spent on the hybrids which do not reproduce. The pattern of the rate of species displacement found in this case was similar to the no interbreeding case (see Fig. 1). The dierence was that the transitions were steeper, that is, when ®xation did not occur, pAA was lower. In many instances the invader genotype did not become established even at high rates of immigration (but low AA ®tness values). The number of generations to species displacement, when it did occur, was the same as the no interbreeding case. The frequency of the native (pBB) only slowly decreased with m and !AA until the steep transition to displacement. The frequency of hybrid (pAB) remained low while AA was low and were lost with the native. The results show that hybridization with sterile hybrids has little eect on species displacement as compared to the no interbreeding case. This result is dependent upon the surviving ``pure'' individuals having great enough fecundity to maintain a consistent population size. If fecundity of these individuals cannot maintain a consistent population size then they are likely to become extinct, which is a real possibility for populations that start with small numbers of individuals. Once the extinction of natives and hybrids has occurred, the invader may then become established. Furthermore, this case demonstrates that even when the invader is at low density, small increases in immigration or ®tness of the invader can lead to rapid displacement. 2.3. Hybridization with backcross sterility In this case the hybrids are fertile yet have become reproductively isolated from the two parental lineages (which can interbreed). This can be modeled by the following: p0AA !AA fag=!AA fag !AB fbg !BB fcg p0AB !AB fbg=!AA fag !AB fbg !BB fcg p0BB !BB fcg=!AA fag !AB fbg !BB fcg a p2AA p2AB =4 1 ÿ m b 2pAA pBB p2AB =2 1 ÿ m c p2BB p2AB =4 1 ÿ m; 3 where a, b and c are the total contributions to the next generation from each genotype for AA, AB and BB, respectively; pAA is the frequency of the ``pure'' invader; pAB is the frequency of the hybrid; and pBB is the fre- quency of the ``pure'' native. Compared to the ®rst two cases, the time to displacement is faster at low immigration and at low !AA values, slower otherwise. From the perspective of invading species with a low immigration rate, this is favorable compared to a case where no interbreeding occurs or the hybrids are sterile. At low invasion numbers, the invader may suer from the Allee eect, which can be molli®ed by interbreeding with the native species. In the one locusÐtwo-allele model, even if introgression does not occur, breeding among hybrids produces a signi®cant number of ``pure'' invaders. In the sterile hybrid model, all reproductive eort with natives is lost, eectively reducing the population size of the invaders. At high immigration rates, sheer numbers alone can overcome the Allee eect. 2.4. Introgression This is the standard Hardy±Weinberg scenario with selection giving Eq. (3), but now a p2AA p2AA =4 pAA pAB =2 1 ÿ m m b 2pAA pBB p2 =2 pAA pAB =2 pBB pAB =2 1 ÿ m c p2BB p2AB =4 pBB pAB =2 1 ÿ m Using this model, I performed two analyses. In the ®rst, in this section, I varied !AA as in the earlier two cases and in the second, Section 2.5, I varied !AB to examine the eects of hybrid ®tness. In the former, !AB and !BB were set to 1 and in the latter, !AA and !BB were set to 1. Hybrid vigor can occur whenever hybrids are produced, sterile hybrids may thrive by vegetative growth and reproductively isolated hybrids may also have increased ®tness. For demonstrative purposes, I chose only to model sexual reproduction in Section 2.5. As with all other cases, ®xation occurs over a wide range of m and !AA values [Fig. 2(a)]. However, in this case ®xation does not occur when !AA < 0:50, but does occur when m50:05 and !AA > 0:90. Thus, ®xation can arise at lower levels of immigration, but displacement is restricted to conditions when the relative ®tness of the native is low. Comparing time to ®xation, the introgression model takes longer than the previous three models, but the dierences are relatively small. The frequency of AA is also much greater when ®xation does not occur in this case compared to the others [Fig. 2(b)]. Because of high frequencies of hybrids, pBB was much lower in this case than the previous models [Fig. 2(c)]. As m and !AA values (as well as pAA) increased, pAB became greater than pBB [Fig. 2(d)]. The hybrid increased as m and !AA increased until the native was driven to low frequencies and then both pAB and pBB went extinct. In the case of introgression, the rate of species displacement is faster at low levels of immigration, but G.R. Huxel / Biological Conservation 89 (1999) 143±152 147 Fig. 2. The introgression case; (a) time to extinction; (b) frequency of AA; (c) frequency of AB; and (d) frequency of BB. slower at lower values of relative ®tness of the invader when compared to the no interbreeding case [Figs. 1(a) and 2(a)]. Owning to a lower frequency of the native genotype and increased frequency of the hybrid, the native is more susceptible to demographic or environmental stochasticity as well as genetic assimilation by the invader. The presence of large numbers of hybrids may also arise at points of contact between species at the limits of their ranges creating hybrid zones. As an example, introgression seems to be playing a role in the loss of the ``pure'' native red deer in Scotland (Abernethy, 1994). In this system a hybrid zone was created and it is in disequilibrium so that the zone is shifting, resulting in a decrease of the native red deer range. Displacement occurs at lower values of !AA for immigration rates greater than 0.05 in comparison of the introgression case with the hybridization with backcross sterility case. Similarly, the hybridization with backcross sterility case is intermediate in terms of the frequency of the three genotypes. The native has a lower frequency when it persists compared to all other cases. The other two genotypes have greater frequencies com- pared to the ®rst two cases, but they have lower frequencies compared to the introgression case. 2.5. Heterozygote ®tness At values of !AB < 1 (underdominance), the species displacement of the native is enhanced [Fig. 3(a) and (b)]. However, when !AB > 1 (overdominance), this was greatly reduced and complete displacement did not occur at any level of m when !AB > 2:1. As pAB increased with !AB , pBB was driven to <0.13 at low m, decreasing to pBB < 0:02 at m > 0:40 [Fig. 3(c) and (d)]. The hybrid also became dominant over the ``pure'' invader as pAA decreased and !AB increased at all levels of m [Fig. 3(b)]. These results may be an artifact of the single locus models in which hybrids interbreeding among themselves or with either parental genotype can produce ``pure'' parental genotypes (see below). This may not occur given a large number of interacting loci. The loss of the ``pure'' native due to genetic assimilation can be hastened by hybrid vigor (overdominance) 148 G.R. Huxel / Biological Conservation 89 (1999) 143±152 Fig. 3. The introgression with underdominance and overdominance: (a) time to extinction; (b) frequency of AA; (c) frequency of AB; and (d) frequency of BB. whether introgression occurs or the hybrids are reproductively isolated from the parental lineages. Though complete displacement may be reduced as the relative ®tness of the hybrid increases, the frequency of the native decreases (above a threshold) and approaches zero when the rate of immigration is high [Fig. 3(a)]. As an illustration, the introduction of Spartina alterni¯ora into the British Isles shows aspects of hybrid sterility, hybridization without backcrossing and hybrid vigor. In plants (and other organisms), hybrids may be reproductively isolated from the parental lineages due to polyploidy. When this occurs and the new species shows hybrid vigor, both ``pure'' species may be displaced. This can be seen in the case of S. anglica (Thompson, 1991). The introduction of S. alterni¯ora, a native of the North American Atlantic coast, into the British Isles lead to its hybridizing with S. maritima, a European native, producing S. townsendii which was sterile. Subsequently a chromosome doubling occurred in S. townsendii producing S. anglica, which was fertile but reproductively isolated from the parental lineages. This new species has since displaced both ``pure'' species and the sterile hybrid and become a dominant plant extending much further down into the tidal zone while the original hybrid has not spread (Thompson, 1991). 3. Discussion The results suggest that displacement of native taxa by non-native taxa can occur very rapidly [e.g. less than ®ve generations; see Fig. 2(a)]. Given the number of cryptic species and the high rates of human-mediated species introductions, extinctions may be on the increase from the eects of hybridization alone. The results of the models are dependent on a number of factors implicitly assumed in the model. These include: (i) that the ®tness of AA and AB can have a large range of values; (ii) migration rates can be extremely high; (iii) a single locus can infer large relative ®tness dierences; and (iv) mating is random. Below I brie¯y discuss each of these. G.R. Huxel / Biological Conservation 89 (1999) 143±152 149 3.1. Fitness of AA and AB 3.2. Migration The range of values of ®tness for AA used in this study may seem fairly extreme, however in the context of an invading species these values may not be extreme. Invading species that are transported across large barriers may have the advantage of escaping from predators, parasites and competitors (Elton, 1958). Essentially species can land in ``enemy free space''. This may explain some of the diculty in developing general theories concerning which species will successfully invade which community (RejmaÂnek, 1995; Reichard and Hamilton, 1997). Competition, predation and parasitism (or a combination of some or all) may drive a species into a narrow niche in its native range, once released from this the species may become more common in its new habitat. Additionally, a species that is rare in its native habitat may become more common with increasing amount of favorable habitat in its new range. New habitat can include prey species with which they have no previous evolutionary contact so that no defenses are encountered. Conversely, a species may ®nd increased competition, predation and parasitism and less favorable habitat in the new region lowering its ®tness. One of the potential consequences of hybridization is that the hybrid may be able to invade territory that its parental species cannot, thus the total population size may be increased due to new resource availability. Spartina anglica's occupation of lower regions of the intertidal zone is an example that is consistent with this scenario (Thompson, 1991). A possible limitation to the model is that it does not allow for hybrids to reproduce vegetatively. Sterile hybrids may expand their range and spread via vegetative growth, but there is no evidence for this. As to the relative ®tness of hybrids, there are only a limited number of documented cases of overdominance. Endler (1986) lists only six examples. However, this may be due to the lack of solid data testing for this phenomenon, but with improving genetic techniques this will become easier to test. Blossey and NoÈtzold (1995) suggested that the evolution of increased competitive ability can result from shifts in resource allocation patterns due to selective forces in the new habitat. This may lead to rapid changes in relative ®tness of the invader and any hybrids. In a test of this hypothesis, Blossey and Kamil (1996) found that introduced Lythrum salicaria (purple loosestrife) grew better in identical conditions than it did in its native range. Similarly, Strong (pers. comm.) has found that introduced Spartina alterni¯ora populations in Washington State grows rapidly, but has lost many of its defenses to herbivores. Levin et al. (1996) further suggested that hybrids might perform best under conditions not favored by either parent. Similarly, Floate et al. (1993) found that herbivory has greatest in hybrid zones relative to ``pure'' zones. The movement of coastal marine species via ship fouling, incidental passengers and ballast water has been occurring for many of years, with most the movement in terms of ballast water occurring recently (Carlton, 1979, 1989; Carlton and Geller, 1993). With many species in the marine environment yet undescribed, the impact of species displacement may be highly signi®cant. Many marine taxa are widely distributed and geographically distinct. Terrestrial systems have also been subjected to introduction rates that are occurring at three or more orders of magnitude greater than the major paleontological invasions such as the Great American Interchange (Williamson, 1996). In addition to human-mediated transportation, changing global climate can also signi®cantly increase migration rates. For marine systems, near-shore temperature is a major limiting factor in species distributions; increased temperatures in these systems will facilitate dispersal. For example, the interaction between Mytilus galloprovinicialis and M. trossolus may be further complicated by warming of northern Paci®c waters allowing M. galloprovinicialis to spread northward (Suchanek et al., 1998). Migration and ®tness may also be altered by dierences in gamete production and fertility. The invading species may genetically ``swamp'' the native by producing large numbers of gametes (i.e. pollen or sperm by males) and increased fertility in the expanded range. Many rare plant populations contain few individuals (Levin et al., 1996). Low population size results in increased susceptibility for rare species to genetic assimilation and species displacement due to interbreeding. This is an important factor of animal species as well. For example, introgression has been cited as a substantial contributing factor for three (the Tecopa pup®sh, Cyprinodon nevadensis; the Amista gambusia, Gambusia amistadensis; and the longjaw cisco, Coregonus alpenae) of the 13 species that have become extinct since the creation of the Endangered Species Act (Rhymer and Simberlo, 1996). An increasing source of migration of large numbers of individuals is the introduction of biological control agents and genetically engineered genes (Howarth, 1991; Klinger and Ellstrand, 1994; Simberlo and Stiling, 1996). Crop plants are a large source of engineered genes and the spread of these genes into wild populations has been documented in several cases (Klinger and Ellstrand, 1994; Bergelson et al., 1998). Bergelson et al. (1998) recently demonstrated that engineered genes in crop plants may have greater risks of spread than wildtype genes in weedy forms. Additionally, many biological control agents have very narrow range of hosts, however, hybrids between these non-native species and natives may allow for a broadening of hosts or host 150 G.R. Huxel / Biological Conservation 89 (1999) 143±152 switching by the hybrids. This potential hybridization is usually not considered when selecting for biological control agents. One the above mentioned species that became extinct after the creation of the Endangered Species Act, Gambusia amistadensis, was in part due to two conspeci®cs, G. anis and G. holbrooki, introduced for biological control of mosquitos. Additionally, other endangered or vunerable Gambusia species hybridize with the mosquito®sh, G. anis (Minckley et al., 1991). 3.3. Number of loci In many cases of hybridization, the species may not be sister species so that segregation between many loci can frequently occur during recombination. Segregation can in¯uence interactions between genes and lead to hybrids being phenotypically closer to one of the parents. In a study of gene interactions in hybrid speciation, Rieseberg et al. (1996) found that selection to a large extent has governed past hybrid species formations. They also stated that nonrandom rates of introgression and signi®cant associations among unlinked genetic markers of three synthesized hybrid lineages implied that interactions between coadapted parental speices' genes constrain the genomic composition of hybrid species. In many cases, reduced hybrid fertility or sterility is due to unfavorable genomic interactions. However, successful origin of new species via hybridization implies that this is not always the case. The selection process will weed out unfavorable combinations while favorable ones are preserved creating new genotypes (or possibly new species) or introducing new alleles into one or both of the parental species. In a theoretical study of relative roles of migration and selection in the introduction of new alleles, Kondrashov (1992) compared a single locus system with a model using an in®nite number of loci. He found that both migration and selection were important factors and that the coexistence of genotypes is more likely to occur with in®nite loci. Yet in both model systems, displacement occurs even in the extreme case of lethal hybrids. While Kondrashov (1992) examined models of extreme numbers of loci, either one or in®nite, most systems will involve an intermediate number of loci. Crow et al. (1990) found in models of two to nine loci that with interbreeding, ®xation of the invading genotype could occur at high rates of two-way migration (the more conservative case compared to one-way migration) and moderate rates of ®tness of the invader. The critical migration rate for the ®xation of the non-native genotype for ! 1 increases to 0.5 in the in®nite loci case, up from 0.18 in the single locus model. However, for presence of the non-native alleles, the migration rate could be several orders of magnitude lower (0.005 in the case of a triple dominate genotype in a three loci model). Furthermore, introgression may lead to the development of a hybrid swarm consisting of a large number of hybrid types due to large numbers of loci. The introgression between the native Pecos pup®sh (Cyprinodon pecosensis) and the introduced sheepheads minnow (C. variegatus) generated a hybrid swarm. C. variegatus was introduced, by sport ®shers, as a bait®sh. Within 5 years of introduction of the sheepheads minnow, the native had been excluded from approximately half its original range and all remaining individuals were hybrids (Echelle and Connor, 1989). Thus, multiple loci seem to enhance the number of hybrid types and genetic mixing and have sped the diminution of the ``pure'' native types. Also, large numbers of loci essentially reduce the probability of a ``pure'' individual of either parental lineage from being produced by two hybrid individuals, an event that is common in the single-locus models. Displacement holds, in model systems, for the in®nite loci case and may exaggerate the in¯uence of introgression due to the large number of potential hybrid combinations. Thus, one would expect that genetic assimilation would be more common with increased, but ®nite, numbers of loci, preserving some of the natives' alleles in the hybrids (however unfavorable genomic interactions will limit the number of hybrid types; Rieseberg et al., 1996). One then can assume that the selection dierences among genotypes are dependent upon a relatively small number of loci. Given even a small number of loci, a greater number of hybrid types will be produced than in the single locus model and the likelihood that one will be able to outcompete both the native and the non-native is increased. Without introgression these hybrids may quickly form a new species as the hybrids are reproductively isolated, but with introgression this may be slowed as backcrosses with parental lineages occurs. 3.4. Mating While my models assume random mating, non-random mating can greatly in¯uence the results. In the case of the Pecos pup®sh, female mate preference for males of other species may have played a role in the rapidity of the spread of exotic genotype and the loss of the ``pure'' native (Echelle and Connor, 1989). In the case of the introduction of brook trout into the Columbia and Klamath Rivers, the lack of introgression created an advantage for the more numerous introduced species (Leary et al., 1993). However for species invading a new habitat, mating is likely to occur with any potential mates. Species that are most at risk to hybridization are ones that have external fertilization (e.g. as broadcast spawning marine invertebrates and outcrossing plants). Wind pollined taxa such as oaks have high rates of introgression. Hybridization is more likely in groups G.R. Huxel / Biological Conservation 89 (1999) 143±152 that have extended breeding cycles (Rhymer and Simberlo, 1996). Associated with random mating is the assumption of large population size that does not vary. This has consequences, as discussed above, in the sterile hybrid model and for newly evolved species and/or hybrids that have slightly dierent ecological requirements. In the latter, the carrying capacity of the system may be increased. Further, habitat segregation and selection have been found to limit random mating in studies of two or three patches (Holt, 1987; Holt and Gaines, 1992) and in metapopulations (Levins, 1969; Hedrick and Gilpin, 1997; Peacock and Smith, 1997). 4. Conclusions Hybridization with introgression can exacerbate species displacement rates under the assumptions of the models presented here. Hybridization with sterile ospring seems to have little eect on displacement. Introgression enhances displacement at low immigration rates, but impedes it when the native has a large selective advantage and at higher rates of immigration. However, because of increased frequency of hybrids with introgression, the native is still heavily impacted and its likelihood of extinction is greatly increased. Varying the ®tness of the hybrid demonstrated that at low !AB , displacement is enhanced and at greater values of !AB , displacement is impeded. But again, because the increased frequency of hybrids, the native species is susceptible to genetic assimilation or extinction. This is especially important because the taxa that are most threatened by introgression and hybridization have low population sizes and restricted ranges (Levin et al., 1996). This includes the potential loss of genetic integrity of native populations through selective stocking of non-native individuals (LargiadeÁr and Scholl, 1996). As Kondrashov (1992) pointed out, both immigration and selection are important in species displacement. For invasive species, both can play important roles when immigration occurs from long distances at frequent intervals leaving ``enemies'' behind (or new niche space is available) and maintained at a high rate of immigration. Because of rapidity of species displacement under conditions of high ! and m, limiting opportunity for species invasions is critical. The loss of natve species due to a closely related introduced species may not just be a simple ecological or evolutionary trade-o. Further extinctions due to unrelated species result in a greater loss of genetic information as some genetic information of the native is assimilated into the invader. In the models presented here, the invader always becomes established and increases in frequency when immigration is greater than zero. Given the rate of species movement by human activity one can expect immigra- 151 tion rates to be increasing. Albuquerque et al. (1996) found that species that are sympatric may have greater barriers to interbreeding than species (or populations) that have evolved allopatically. These results suggest that more empirical studies need to be carried out to examine the genetic structure of populations that are being invaded and determine whether hybridization and introgression are possible. Additionally, the information on the relative ®tness provides better predictive power. This information will allow for predictions of whether displacement is likely. However, due to the rapidity of displacement under many conditions of the models, it is best to prevent biological invasions from occurring when possible. Acknowledgements I would like to thank Dan Simberlo, Bernd Blossey, Susan Harrison, Alan Hastings, Kevin McCann, Donald Strong, Alisa Swann, David Cutler, Amy Wolf and two anonymous reviewers for suggestions, comments and references. Support for this work was provided by a NSF Research Training Grant (BIR-960226) and a NSF Population Biology Grant (DEB-9629236). References Abernethy, K., 1994. The establishment of a hybrid zone between red and sika deer (genus Cervus). Molecular Ecology 3, 551±562. Albuquerque, G.S., Tauber, C.A., Tauber, M.J., 1996. Postmating reproductive isolation between Chrysopa quadripunctata and Chrysopa slossonae: mechanisms and geographic variation. Evolution 50, 1598±1606. Arnold, M.L., 1997. Natural hybridization and evolution. Oxford University Press, New York. Bergelson, J., Purrington, C.B., Wichmann, G., 1998. Promiscuity in transgenic plants. Nature 395, 25. Blossey, B., Kamil, J., 1996. What determines increased competitive ability of invasive non-indigenous plants? In: Moran, V.C., Homan, J.H. (Eds.) Proceedings of the IX International Symposium on Biological Control of Weeds, 19±26 January 1996. Stellenbosch, South Africa, pp. 3±9. Blossey, B., NoÈtzold, R., 1995. Evolution of increased competitive ability in invasive nonindigenous species: a hypothesis. J. of Ecol. 83, 887±889. Carlton, J.T., 1979. Introduced invertebrates of San Francisco Bay. In: Conomos, T.J. (Ed.), San Francisco Bay: an urbanized estuary. California Academy of Science, San Francisco, pp. 427±442. Carlton, J.T., 1989. Man's role in changing the face of the ocean: biological invasions and implications for conservation of nearshore environments. Cons. Biol. 3, 265±273. Carlton, J.T., 1996. Biological invasions and crytogenic species. Ecology 77, 1653±1655. Carlton, J.T., Geller, J.B., 1993. Ecological roulette: the global transport of non-indigenous marine organisms. Science 261, 78±82. Case, T.J., 1996. Global patterns in the establishment and distribution of exotic birds. Conservation Biology 78, 69±96. Crow, J.F., Engels, W.R., Denniston, C., 1990. Phase three of Wright's shifting-balance theory. Evolution 44, 233±247. Dowling, T.E., DeMarais, B.D., 1993. Evolutionary signi®cance of introgressive hybridization in cyprinid ®shes. Nature 362, 444±446. Echelle, A.A., Connor, P.J., 1989. Rapid, geographically extensive genetic introgression after secondary contact between two pup®sh species (Cyprinodon, Cyprinodontidae). Evolution 43, 717±727. 152 G.R. Huxel / Biological Conservation 89 (1999) 143±152 Elton, C.S., 1958. The ecology of invasions by animals and plants. Methuen, London. Endler, J.A., 1986. Natural selection in the wild. Princeton University Press, Princeton, NJ. Floate, K.D., Kearsley, M.J.C., Whitham, T.G., 1993. Elevated herbivory in plant hybrid zones: Chrysomela con¯uens, Populus and phenological sinks. Ecology 74, 2056±2065. Geller, J.B., Walton, E.D., Grosholz, E.D., Ruiz, G.M., 1997. Cryptic invasions of the crab Carcinus detected by molecular phylogeny. Molec. Ecol. 6, 901±906. Grant, P.R., Grant, B.R., 1992. Hybridization of bird species. Science 256, 193±197. Grant, P.R., Grant, B.R., 1996. Speciation and hybridization in island birds. Phil. Trans. R. Soc. Lond. B 351, 765±772. Grant, V., 1981. Plant Speciation, 2nd ed. Columbia University Press, New York. Harrison, R.G., 1990. Hybrid zones, windows on evolutionary processes. In: Futuyma, D., Antonovics, J. (Eds.), Oxford Surveys in Evolutionary Biology, Vol. 7. Oxford University Press, Oxford, pp. 69±128. Hedrick, P.W., Gilpin, M.E., 1996. Genetic eective size of a metapopulation. In: Hanski, I.A., Gilpin, M.E. (Eds.), Metapopulation Biology: Ecology, Genetics and Evolution. Academic Press, San Diego, CA, pp. 166±182. Heywood, V.H., 1989. Patterns, extents and modes of invasions by terrestrial plants. In: Drake, J.A., Mooney, H.A., di Castri, F., Groves, R.H., Kruger, F.J., RejmaÁnek, M., Williamson, M. (Eds.), Biological invasions, a global perspective. John Wiley & Sons, Chichester, UK, pp. 31±60. Holt, R.D., 1987. Population dynamics and evolutionary processes: the manifold roles of habitat selection. Evol. Ecol. 1, 331±347. Holt, R.D., Gaines, M.S., 1992. Analysis of adaptation in heterogeneous landscapes: implications for the evolution of fundamental niches. Evol. Ecol. 6, 433±447. Howarth, F.G., 1991. Environmental impacts of classical biological control. Annu. Rev. Entomol. 36, 485±509. Klinger, T., Ellstrand, N.C., 1994. Engineered genes in wild populations: ®tness of weed-crop hybrids of Raphanus sativus. Ecological Applications 4, 117±120. Kondrashov, A.S., 1992. The third phase of Wright's shifting-balance: a simple analysis of the extreme case. Evolution 46, 1972±1975. LargiadeÁr, C.R., Scholl, A., 1996. Genetic introgression between native and introduced brown trout Salmo trutta L. populations in the RhoÃne River Basin. Molec. Ecol. 5, 417±426. Leary, R.F., Allendorf, F.W., Forbes, S.H., 1993. Conservation genetics of bull trout in the Columbia and Klamath River drainages. Cons. Biol. 7, 856±865. Levin, D.A., Francisco-Ortega, J., Jansen, R.K., 1996. Hybridization and the extinction of rate plant species. Cons. Biol. 10, 10±16. Levins, R.R., 1969. Some demographic and genetic consequences of environmental heterogeneity for biological control. Bull. Entomol. Soc. Am. 15, 237±240. Mayr, E., 1942. Systematics and the origin of species Columbia University Press, New York. Minckley, W.L., Mee, G.K., Soltz, D.L., 1991. Conservation and management of short-lived ®shes: the Cyprinodontoids. In: Minckley, W.L., Deacon, J.E. (Eds.), Battle Against Extinction. University of Arizona Press, Tucson, AZ, pp. 247±282. Moulton, M.P., 1993. The all-or-none pattern in introduced Hawaiian passiformes: the role of competition sustained. American Naturalist 141, 105±119. Moulton, M.P., Pimm, S.L., 1986. Species introductions to Hawaii. In: Mooney, H.A., Drake, J.A. (Eds.), Ecology of Biological Invasions of North America and Hawaii. Springer±Verlag, New York, pp. 231±249. Peacock, M.M., Smith, A.T., 1997. Nonrandom mating in pikas Ochotona princeps: evidence for inbreeding between individuals of intermediate relatedness. Molec. Ecol. 6, 801±811. Reichard, S.H., Hamilton, C.W., 1997. Predicting invasions of woody plants introduced into North America. Cons. Biol. 11, 193±203. RejmaÂnek, M., 1995. What makes a species invasive? In: PysÏ ek, K., Prach, M., RejmaÂnek, M., Wade, M. (Eds.), Plant InvasionsÐ General Aspects and Special Problems. SPB Academic Publishing, Amsterdam, pp. 3±13. Rhymer, J.M., Simberlo, D., 1996. Extinction by hybridization and introgression. Ann. Rev. Ecol. and Syst. 27, 83±109. Rieseberg, L.H., Sinervo, B., Linder, C.R., Ungerer, M.C., Arias, D.M., 1996. Role of gene interactions in hybrid speciation: evidence from ancient and experimental hybrids. Science 272, 741±745. Simberlo, D., Stiling, P., 1996. How risky is biological control? Ecology 77, 1965±1974. Suchanek, T.H., Geller, J.B., Kreiser, B.R., Mitton, J.B., 1998. Zoogeographic distributions of the sibling species Mytilus galloprovinicialis and M. trossolus (Bivalvia: Mytilidae) and their hybrids in the North Paci®c. Biol. Bull. 193, 187±194. Thompson, J.D., 1991. The biology of an invasive plant. BioScience 41, 393±401. Williamson, M., 1996. Biological Invasions. Chapman & Hall, London.