Genetic Status and Timing of a Weevil Introduction to Santa Cruz Island, GalápagosHoi-Fei Mok, Courtney C. Stepien, MarySka kaCzMarek, Lázaro roque aLbeLo, and andrea S. Sequeira
From the Department of Biological Sciences, Wellesley College, Wellesley (Mok, Stepien, Kaczmarek, and Sequeira); Department of Terrestrial Invertebrates, Charles Darwin Research Station, Puerto Ayora, Galápagos, Ecuador (Roque Albelo). Hoi-Fei Mok is now at the School of Land and Environment, University of Melbourne, Melbourne, Australia; Courtney C. Stepien is now at the Graduate Program, Committee on Evolutionary Biology, University of Chicago, Chicago, IL; Maryska Kaczmarek is now at the Department of Molecular Genetics and Microbiology, The University of Texas at Austin, Austin, TX; Lázaro Roque Albelo is now at the Invertebrate Science Team, Ecologia Environment, Australia and Curtin Institute for Biodiversity and Climate, Curtin University, Perth, Australia.
Address correspondence to Dr. Andrea S. Sequeira, Department of Biological Sciences, Wellesley College, 106 Central Street, Wellesley, MA 02481, or e-mail: [email protected].
Data deposited at Dryad: http://dx.doi.org/doi:10.5061/dryad.3ks4v
AbstractSuccessful invasive species can overcome or circumvent the potential genetic loss caused by an introduction bottleneck through a rapid population expansion and admixture from multiple introductions. We explore the genetic makeup and the timing of a species introduction to Santa Cruz Island in the Galápagos archipelago. We investigate the presence of processes that can maintain genetic diversity in populations of the broad-nosed weevil Galapaganus howdenae howdenae. Analyses of com-bined genotypes for 8 microsatellite loci showed evidence of past population size reductions through moment and likelihood-based estimators. No evidence of admixture through multiple introductions was found, but substantial current population sizes (N0 298, 95% credible limits 50–2300), genetic diversity comparable with long-established endemics (Mean number of alleles = 3.875), and lack of genetic structure across the introduced range (FST = 0.01359) could suggest that foundations are in place for populations to rapidly recover any loss of genetic variability. The time estimates for the introduction into Santa Cruz support an accidental transfer during the colonization period (1832–1959) predating the spurt in human population growth. Our evaluation of the genetic status of G. h. howdenae suggests potential for population growth in addition to our field observations of a concurrent expansion in range and feeding preferences towards protected areas and endemic host plants.Key Words: bottlenecks, Galapaganus h. howdenae, microsatellite loci, Msvar, older introduction, range expansion
Several genetic events are possible on the introduction of exotic species. Introduced species themselves can undergo population bottlenecks, where a drastic reduction in popula-tion size causes the genetic makeup of the remaining few to become the new gene pool (Nei et al. 1975). If severe and long enough, bottleneck events can result in loss of variants, decreased genetic flexibility, and loss of phenotypes (Nei et al. 1975; Garza and Williamson 2001; Allendorf and Lundquist 2003; Colautti et al. 2005)—all possibly detri-mental to the success of the recently introduced population. However, rapid demographic growth and/or range expan-sions after translocations have made the detection of bottle-necks impossible (Zeisset and Beebee 2003) and caused the
retention of genetic diversity (Zenger et al. 2003). Despite the presumption that bottlenecks are disadvantageous on invasiveness, the length of the population size reduction dur-ing the introduction event and the speed of the recovery after the introduction event can effectively diminish or avoid those detrimental effects. Additionally, one key to invasion success could be the occurrence of multiple introductions that may transform among-population variation in native ranges to within-population variation in introduced areas (Kolbe et al. 2004; Williams et al. 2005). Alternatively, maintaining connec-tivity across the area of introduction could ensure the spread of less frequent allelic variants, which could either arrive or arise at the edges of the introduced range, throughout the
Journal of Heredity Advance Access published January 7, 2014 by guest on January 8, 2014
area of introduction. In summary, the substantial loss of variability that can be associated with an introduction can be mitigated or circumvented altogether through at least 3 processes: introductions from multiple genetically differen-tiated continental sources (Kolbe et al. 2004; Facon et al. 2008), rapid demographic and/or range expansions (Zeisset and Beebee 2003; Zenger et al. 2003), and population con-nectivity through effective dispersal (Mishra et al. 2006). If a population maintains or increases genetic variability after introduction through any or all of these processes, chances of deleterious alleles surfacing are decreased and overall fit-ness could be increased (Lindolm et al. 2006). Still, the under-lying assumption that all types of genetic diversity are equally important for the long-term survival of a population is chal-lenged by the fact that some endangered populations have maintained small census sizes and very low levels of neutral genetic diversity more than hundreds or thousands of years (Johnson et al. 2009) and that low levels of diversity may not be accompanied by obvious reductions in fitness (Brodie 2007). Undeniably, newly introduced populations could remain unaffected by a reduction in census size if they do not experience a corresponding reduction in effective popu-lation size. Nonetheless, monitoring the genetic makeup of introduced populations using neutral markers, such as micro-satellite loci, still remains an informative mean of exploring the underlying genetics of invasiveness (Alda et al. 2013; Frisch et al. 2013; Ghabooli et al. 2013; McMahon et al. 2013; Perdereau et al. 2013; Simpson et al. 2013).
The effects of introduced populations on the genetic makeup of their native counterparts can range from increased vigor to complete loss of fitness in the introduced–native hybrids (Huxel 1999; Fitzpatrick and Shaffer 2007). Additionally, hybridization may stimulate invasiveness in the introduced populations (Ellstrand and Schierenbeck 2000). In contrast, the effects of introduced species on the ecological communities that harbor their native counterparts are harder to unravel (Evans 2004). Species invasions of continental hab-itats are rarely accompanied by the total exclusion of native competitors. Island systems, on the other hand, appear more vulnerable because they seemingly are less saturated or the native species are competitively inferior (Simberloff 1986); this has caused at least one oceanic archipelago to become a geographic hotspot of invasion for certain insect taxa (Beggs et al. 2011). The vulnerability of native biotas to ecological disruption may be especially great on oceanic islands invaded by continental species with unique ecological traits (Wilson and Holway 2010). An important correlate of the vulnerabil-ity to displacement by invading counterparts of insect spe-cies on islands is the provenance of the populations studied, where populations of rare endemics appear to be more vul-nerable to displacement by invasives than introduced species sharing their habitat (Krushelnycky and Gillespie 2010).
Although oceanic islands such as those in the Galápagos archipelago are natural models for studying the process of generation and maintenance of species diversity (Emerson and Kolm 2005), such endemic diversity may become threat-ened by the introduction of exotic species. Since human set-tlement on the Galápagos in 1535, the native biodiversity
has been altered through the addition of alien insect spe-cies (Desender et al. 1992; Klimaszewski and Peck 1998; Desender et al. 2002). Although laws restricting transporta-tion of live animals and plants in 1994 and 1998 have reduced introductions, accidental establishment of insects still occurs through importation of construction materials and food supplies, especially on the large and central islands (Peck et al. 1998; Causton et al. 2006). Furthermore, introduced species from nearby continental areas have been identified as the principal threat to the terrestrial ecosystems of the Galápagos Islands, given that many later become established and widespread (Snell et al. 2002a, 2002b). Since humans first arrived to Galápagos the rate of unintentional insect intro-ductions has been about one new insect species per year; however since 1998 the rate has increased by 59% (Peck et al. 1998; Causton et al. 2006). A large proportion of these intro-duced insects are plant feeders (42%), threatening endemic flora with disease transmission and herbivory (Causton et al. 2006). Evidence of regular ongoing human-assisted intro-ductions suggests that the measures already in place are not sufficient to stem the tide of incoming species (Bataille et al. 2009). Considering the susceptibility of Galápagos biota and the environmental damage possible, identifying potential invasive insects and monitoring their impact on native and endemic species on the Galápagos archipelago remains a rel-evant conservation question.
The geologic age of the extant islands within the Galápagos archipelago sits within the “middle age” range of the life cycle of volcanic islands (Peck 2005); the maximum age of the oldest islands is reported to be between 3 and 4 million years (Bailey 1976; Hickman and Lipps 1985) or up to 6.3 mil-lion years (Geist 1996; White et al. 1993). However, the his-tory of human–nature relationships in Galápagos is relatively short and spans no more than 5 centuries (Gonzalez et al. 2008). The four periods within the historical profile of this interaction determined by Gonzalez and coauthors (2008) pinpoint time-frames where the rate of accidental introduc-tions increased significantly. The four periods are described as extractive exploitation (1535–1832), colonization (1832–1959), wilderness conservation (1959–1998), and conservation–development balance (1998–present)(Gonzalez et al. 2008). Even though the first alien species are thought to have entered Galápagos during the early period of extractive exploitation (Tye et al. 2002), it was the establishment of the first human settlements during the colonization period that brought a dra-matic increase in the numbers of exotic plants and animals. Interestingly, human population numbers increased signifi-cantly later, more markedly between 1950 and 2001.
The weevil genus Galapaganus Lanteri (Coleoptera: Curculionidae) has presented a suitable system to study spe-cies radiations on islands (Lanteri 1992; Sequeira et al., 2000; Sequeira, Lanteri et al., 2008; Sequeira, Sijapati et al., 2008; Sequeira et al., 2012). We will now attempt to use introduced island populations of a continental Galapaganus species to explore the genetic signals of a presumed recent introduc-tion in this ecosystem that is hailed as an example for conser-vation efforts while under increasing pressure from tourism and development. The genus Galapaganus contains a total of
by guest on January 8, 2014http://jhered.oxfordjournals.org/
Mok et al. • Genetics of a Weevil Introduction to Galápagos
3
15 species; 13 within the darwini species group are flightless and fairly heavy bodied. Ten of the species within the dar-wini group are endemic to the islands (Lanteri 1992), whereas the remaining three, together with winged Galapaganus in the femoratus group, reside on the Ecuadorian mainland and Perú (Lanteri 2004). Lanteri (2004) described 2 subspecies for one of the femoratus group species, Galapaganus howdenae howdenae from the lowlands of Guayas, Manabí and Los Rios prov-inces, and G. howdenae pecki from the highlands of Pichincha province in Ecuador.
The central island of Santa Cruz in Galápagos harbors 2 endemic and 1 introduced Galapaganus species (G. conway-ensis and G. ashlocki: endemics; G. h. howdenae: introduced). Galápagos introduced specimens are typically more densely covered in scales and have broader bodies than those found at higher altitudes of mainland Ecuador, fitting the descrip-tion of coastal G. h. howdenae (Lanteri 2004). All collections suggest that to date only G. h. howdenae has been introduced into the focus area of this study—the Agricultural Zone (AZ) of Santa Cruz Island—and that it has not been intro-duced elsewhere. Like its endemic relatives, G. h. howdenae larvae are root feeders, whereas adult weevils feed on leaves of a variety of plant hosts. Unlike its endemic counterparts, G. h. howdenae have well-developed and flight-capable wings, giving this introduced species more efficient dispersal abili-ties. Interestingly, this distinction in ecological traits between the continental and island endemic species could contrib-ute to the vulnerability of the populations of island natives. Galapaganus h. howdenae was first reported on the island of Santa Cruz in 1992 (Lanteri 2004). Given the location of the main ports of entry for goods in Puerto Ayora, a likely intro-duction scenario is that G. h. howdenae could have been intro-duced either once or multiple times together with imported plants, first into lowland areas of Santa Cruz and later either passively transported with those plants or actively dispersed into the AZ where it established populations. We have now repeatedly found it beyond the boundaries of the dis-turbed AZ (in 2006, 2007), feeding on introduced as well as endemic vegetation side by side with G. conwayensis—one of its endemic close relatives. However, our most recent obser-vations (in 2010, 2011) suggest that G. h. howdenae has also further broadened its range into the moist highlands reach-ing the highest points in Santa Cruz. It now shares habitats and plant hosts with a highland specialist and single island endemic (G. ashlocki) (Sequeira et al. 2012). Analysis of mito-chondrial sequences comparing populations of endemic and introduced Galapaganus species indicated that time of estab-lishment and dispersal abilities are important determinants of genetic structure in weevil populations. In particular, the 3 introduced populations included in that study, though vari-able, lacked genetic structure and retained signals of a recent change in population size (Sequeira et al. 2012).
Studying the genetic composition of populations of G. h. howdenae across the introduced range using multiple markers can continue to clarify the invasive potential of this weevil introduction (Allendorf and Lundquist 2003; Kolbe et al. 2004; Colautti et al. 2005). Many island introductions show clear evidence of loss of variability (Puillandre et al.
2008); however, that loss is linked with differing effects on the new arrival’s establishment success and its impact on native species. In particular, the potential impacts of the introduced populations of G. h. howdenae are compounded by Santa Cruz’s central location in the Galápagos archipel-ago (Figure 1). Given the flight capabilities of G. h. howdenae, the risk of it dispersing to the other islands from the central island is increased (Roque Albelo et al. 2006). An assessment of alien insect species in the Galápagos has already recog-nized G. h. howdenae as a potentially moderate invasive species (Causton et al. 2006). As with other Galápagos insect intro-ductions (Dudaniec et al. 2008), analyzing the genetic sta-tus of introduced populations can not only shed light on its invasive potential but also aid identifying subsequent steps to prevent further expansion. Considering the recent observa-tions of expanding range of G. h. howdenae, and the potential damages in consequence, such insight may prove important for protection of the Galápagos’ endemic biodiversity.
This study will analyze the variation found at 8 microsat-ellite loci to assess the current genetic status of populations of G. h. howdenae across the known introduced range in the AZ and National Park (NP) areas of Santa Cruz. A second objective is to explore the origin of that genetic variation by reconstructing the timing and number of introduction events into Santa Cruz.
The predictions associated with the genetic status of the introduced G. h. howdenae populations in Santa Cruz are: 1) If populations have undergone a bottleneck associated with their introduction to Galápagos (or some other event) within 2–4 Ne generations, then we would observe i) Low ratio of observed alleles to allele range (M ratio), ii) Excess heterozy-gosity with respect to the mutation-drift equilibrium, and iii) Large differences between likelihood estimates of ancestral and current population sizes. 2) If populations across the introduced range are maintaining or recovering from the potential depletion of variability caused by a bottleneck, then we would find current amounts of variation comparable with those of closely related long-established endemics. 3) If introduced populations across the AZ and NP areas form one cohesive unit, then we would find no significant genetic structuring as demonstrated by lack of significant F statistic values.
The predictions regarding the number and timing of introduction events are the following: 1) If the current genetic composition in the introduced range were the prod-uct of multiple introductions from already differentiated populations in the source range, then we would detect sig-nals of genetically differentiated clusters where either each individual is itself the product of admixed origins or the populations are composed of individuals that can be traced as deriving from genetically differentiated populations. 2) If the introduction of G. h. howdenae has been recent and aided by humans, then we would observe a reduction in popula-tion size contemporaneous with either the establishment of human settlements during the colonization period of the islands between 1835 and 1959 or concurrent with the rapid growth of human populations (between 1950s and 2001) (Gonzalez et al. 2008).
by guest on January 8, 2014http://jhered.oxfordjournals.org/
Figure 1. Map of the island of Santa Cruz in the Galápagos Archipelago. Outline indicates the Agricultural Zone. Collecting localities are labeled with population codes following Table 1. Inset A: Relative location of Santa Cruz within the archipelago. Inset B: Relative location of the archipelago with respect to western South America.
by guest on January 8, 2014http://jhered.oxfordjournals.org/
Mok et al. • Genetics of a Weevil Introduction to Galápagos
5
Materials and MethodsSampling
We obtained samples of G. h. howdenae weevils from 9 localities across the AZ and NP areas of Santa Cruz in the Galápagos archipelago (Figure 1, Table 1). Adult weevils were collected by beating on introduced plant hosts such as Erytrina, Guava and Aguacate with the exception of those collected in El Chato (Figure 1, Table 1) which were found feeding on Tournefortia rufoscericea, an endemic plant host, alongside endemic congener G. conwayensis. Weevils were pre-served in 100% ethanol until processed for DNA extraction. In 2 cases, specimens from nearby sites or from the same location collected in consecutive years were pooled and ana-lyzed as a single locality (Media Luna and La Caseta, respec-tively, Figure 1, Table 1).
DNA Preparation and Genotyping
We used 3 legs from each specimen to isolate DNA accord-ing to the protocol of Normark (1996) or alternatively used the DNeasy Tissue Kit (Qiagen, Valencia, CA). Two microsatellite libraries were constructed using MseI and AseI restriction enzymes, following the FIASCO (Fast Isolation by AFLP of Sequences Containing repeats) protocol (Zane et al. 2002; Stepien et al. 2010). Optimal PCR reaction and cycling conditions differed across polymorphic microsatellite loci (Stepien et al. 2010). Microsatellite variation was analyzed for 8 polymorphic loci in 106 G. h. howdenae individuals from multiple locations across Santa Cruz. Generally, amplification reactions were performed in a final volume of 10 μl contain-ing 1–2 μl DNA template, 1 μl 10× ThermoPol Reaction Buffer containing 2 mM MgSO4 (New England Biolabs, Ipswich, MA), Taq polymerase (1U) (New England Biolabs), 0–3.0 mM of MgCl2, 1 fluorescently tagged forward primer (5 μM; HEX: hexachloro-fluoresceine, 6-FAM: 6-carboxyflu-orescein, NED and PET: property of Applied Biosystems, now Life Technologies, Foster City, CA) and reverse prim-ers (5 μM), dNTPs (0.8 mM) (Invitrogen, Carlsbad, CA), and high performance liquid chromatography water. Quartets of polymerase chain reaction products were combined with
LIZ600 or LIZ500 size standard (Applied Biosystems) and run on an ABI3100 Genetic Analyzer. Genemapper version 4.3 was used to score positive reactions. In order to ensure accuracy in allele calling, individuals from all localities were randomly selected to be genotyped 2 or 3 times, and allele sizes were verified through blind tests.
Genetic Analyses
Genetic Diversity, Population Differentiation, and Mixed Ancestry
We determined microsatellite loci independence through the linkage disequilibrium test in Genepop 4.0 (Raymond and Rousset 1995). A probability test (G-test) was computed for contingency tables of each loci pair generated using the Markov chain algorithm with the following parameters: 10 000 dememorization steps, 100 batches, and 5000 iterations per batch. Linkage disequilibrium tests in Arlequin version 3.11 (Excoffier et al. 2005) confirmed results from Genepop 4.0.
Microsatellite genetic diversity (Mean number of alleles, A; expected and observed heterozygozity, HE and HO) and deviation from Hardy–Weinberg equilibrium (Fis indices) were determined per locus and per locality area (AZ, NP, and all populations: Table 2) using Genepop 4.0. FIS indices were calculated and tested for significance using exact tests with the Markov chain algorithm with 10 000 dememoriza-tion steps, 20 batches, and 5000 iterations per batch (Haldane 1953; Guo and Thompson 1992) (Table 2). Molecular diver-sity indices obtained in arlequin 3.11 (Excoffier et al. 2005) confirmed results from Genepop 4.0. Statistical differences in mean number of alleles and heterozygosity values between the 2 locality areas were assessed using the Mann–Whitney test (Mann and Whitney 1947).
We searched for signals of genetic structuring through analyses of molecular variance (amova) in Arlequin 3.11(Excoffier et al. 2005). The hierarchical distribution of genetic variance was assessed for all localities as a single group. To evaluate if there was any geographic pattern of genetic dif-ferentiation across the distribution range in Santa Cruz, we grouped localities into AZ and NP. Additionally, we assessed differentiation across pairs of localities computing pairwise FST indices. All analyses consisted of 10 000 permutation steps (Cockerham and Weir 1984; Cockerham and Weir 1987) (Table 3) and were run in arlequin 3.11(Excoffier et al. 2005).
We used the clustering approach implemented in Structure 2.3 (Pritchard et al. 2000) to explore the like-lihood of mixed ancestry in the introduced range of G. h. howdenae in Galápagos. If we were to find evidence of multiple population units within Santa Cruz, that could be interpreted as multiple introductions from differentiated continental sources. The natural logarithm of the probability of the data set was calculated more than 10 trials per treat-ment using a burn-in of 10 000 and 90 000 replicates with either independent or correlated allele frequencies for all trials (Figure 2, Table 4). The admixture and no admixture ancestry models were used in each case with and without the addition of sample group information to aid in the cluster-ing (LocPrior function)(Hubisz et al. 2009). For all ancestry
Table 1 Locality information: population codes, locality details, and locality area designations for introduced Galapaganus h. howdenae populations on Santa Cruz, Galápagos
Population code
Locality (altitude, in meters)
Zone N
SR10 Road to Cerro Croker (249) AZ 15SR07 Finca Steve Devine (351) AZ 15SR11 Above El Chato (422) AZ 4SR22 El Chato (106) AZ 9SR06 2 km north of El Cascajo (240) AZ 15SR08/ SR18 Media Luna (469) NP 13SR09 Miconia Zone (500) NP 6SR12/SR17 La Caseta (613) NP 14SR01 Cerro Croker trail (400–800) NP 15
N = number of individuals analyzed per locality.
by guest on January 8, 2014http://jhered.oxfordjournals.org/
models, values of 1.0 and 10.0 were used for initial and maxi-mum alpha, respectively, for all numbers of clusters (K) set between 1 and 6. The locality groupings of AZ and NP as designated populations were used as locality priors under the LocPrior function, given the subtle genetic structure detected between these 2 units (see amova results below). Following the recommendations for additional prior settings (Pritchard et al., 2000), the probability that an individual is an immigrant to the community or has a recent immigrant ancestor, otherwise known as MIGPRIOR or v, was set at 0.05. To analyze results from multiple runs simultaneously, we summarized the mean proportion of membership across all trials and models either by measuring the deviation from perfectly symmetrical assignments (1/k) for the runs without a LocPrior or by calculating mean and standard deviation of membership in the largest cluster across identical runs for the LocPrior models (Table 4). Additionally, we explored the possibility of multiple clusters both visually, with plots of mean likelihood for increasing numbers of inferred clusters, and analytically, through DeltaK values following the Evanno method (Evanno et al. 2005). Structure results for all 8 sets of runs (Admixture and No Admixture; correlated and inde-pendent allele frequencies; with and without addition of LocPrior) were summarized and analyzed using Structure HarveSter (Earl and Vonholdt 2012) (Supplementary Material Figure S1 and Figure 2).
Demographic History
Moment-based estimators. We used 2 lines of evidence to search for genetic signals of a past reduction in population
size within the G. h. howdenae range in Santa Cruz poten-tially associated with its introduction to the archipelago. The population bottleneck was explored through tests of excess heterozygosity with respect to the mutation–drift equilibrium and of a significantly reduced ratio of the number of micro-satellite alleles to allele size range—known as the M ratio. Bottleneck 1.2.02 (Piry et al. 1999) tested for heterozygo-sity excess with respect to mutation–drift equilibrium charac-teristic of population size reductions using the Sign test and the Wilcoxon signed rank one-tailed test for excess (Table 5). Three mutation models were used: infinite allele model (IAM), stepwise mutation model (SMM), and 2-phase model (TPM) with 1000 replicates for each model and parameters derived from the G. h. howdenae data set implemented in the TPM (36% stepwise mutations, variance = 130). As the vari-ance in repeat number at a locus increases, the frequency or magnitude of multi step mutations also increases (Di Rienzo et al., 1994). Though low variance is recommended for microsatellites (Di Rienzo et al., 1994; Piry et al. 1999), the high percentage of large step mutations in the G. h. howdenae data set prompted use of a higher variance value (Table 6).
In addition to the Sign and Wilcoxon tests, mode shifts for allele distribution were calculated to detect any popula-tion bottleneck–related deviation from the normal allele dis-tribution (with one most common allele size and other less frequent allele sizes that differ by varying number of steps). Because low-frequency alleles are lost during a population bottleneck, this can be detected as a shift in the distribution and frequency of allele sizes (Table 5).
During bottlenecks, the range of possible allele sizes stays relatively constant, whereas the number of alleles decreases, leading to a characteristic reduction in M ratio (Garza and Williamson, 2001; Hawley et al., 2006). The software program M (Garza and Williamson, 2001) was used to calculate the M ratios and significant deviations from equilibrium situations (Table 6). M ratios were calculated for all localities as a single group, as well for the 2 locality areas: AZ and NP. The following parameters were used in the analyses: 64–67% proportion of nonstep mutations, 4.6–4.8 average base size of nonstep muta-tions, assumed marker mutation rate (μ) of 5 × 10−4 (Goldstein and Schlotterer 1999), effective prebottleneck population size of Ne = 2000 (Goldstein and Schlotterer 1999; Hawley et al. 2006), and 1000 simulations (Table 6). A recent study explor-ing the reliability of moment-based estimators—M ratio and heterozygote excess—remarks that, in general, bottleneck tests fail to detect declines when changes in population size
Table 3 Genetic structure across localities
Groups Fixation indices P values
All localities FST = 0.01359 (1.36%) 0.058AZ FST = 0.00060 (0.06%) 0.264NP FST = 0.00834 (0.83%) 0.178By locality area (AZ and NP)
Hierarchical analyses of molecular variance (amova) in introduced Galapaganus populations, all grouped by locality (All Localities), only locali-ties within the AZ, and only localities within the NP and all localities grouped by area AZ and NP.F statistics (FST, FSC, and FCT) among all localities, among localities within areas, and among locality groups.
Table 2 Genetic diversity in Galapaganus howdenae howdenae populations across the introduced range
Indices are calculated across all loci for all localities together and each locality area.N = number of individuals; #A = mean number of alleles across all loci; HO and HE = observed and expected heterozygosity; All loc = FIS values for all localities; AZ = localities in the Agricultural Zone; and NP = localities in the National Park.*P < 0.05.
by guest on January 8, 2014http://jhered.oxfordjournals.org/
Mok et al. • Genetics of a Weevil Introduction to Galápagos
7
had been recent (Peery et al. 2012). Even when using the more realistic 2-phased mutation model Peery and coauthors (2012) suggest, among other things, increasing sample size and num-ber of loci compared to add rigor to the inferences and warn against underestimating the percentage of multi step mutation in the TPM model. In line with those recommendations, in this study we are including weevil samples spanning the entire range of G. h. howdenae in Galápagos and using an elevated per-centage of multi step mutations derived from our data set. The number of polymorphic loci analyzed falls within the range of most studies reviewed by Peery and coauthors (2012).
Likelihood-based methods. We inferred past popula-tion dynamics using a likelihood-based Bayesian method (Beaumont 1999; Storz and Beaumont 2002) implemented in mSvar 0.4 and mSvar 1.3. Simulations have shown that mSvar is efficient at detecting population declines and expansions, provided that the event is neither too weak nor too recent (Girod et al. 2011). Additionally, simulations suggest that the method is robust to moderate departures from a strict SMM
(Girod et al. 2011). The method implemented in mSvar 0.4 estimates the posterior probability distributions of the rate of population size change (r = N0/N1), the time since the population size change scaled by the current population size (tf = ta/N0), and the mutation rate as θ = 2 N0μ, where μ is the locus mutation rate. The population size changes inferred in mSvar 0.4 were calculated under an exponential model (Figure 3). To quantify and date the change in population size we sampled from the posterior distribution of parameters r and tf. We used rectangular prior distributions for log r, log tf and log θ as well as combinations of starting values for each parameter. Given that MSVAR 0.4 does not allow for the direct calculations of the time of population change (ta ), this was calculated from tf after independently determining the current effective population size (N0) using the web-based program ONeSAMP (Tallmon et al. 2008). This method assumes that all loci are neutral and unlinked and is based on simulations of a single closed population. The assumptions detailed above seemed realistic in our study system, given the nature of the markers themselves, the general lack of linkage
Figure 2. Structure v. 2.3 analysis of the G. h. howdenae populations in Santa cruz. (A) Plot of estimated ln likelihood against number of inferred genetic clusters (K) and (B) Delta K values without (i) and with (v) the locality areas as locality priors, under admixture with indepedent allele frequencies.
by guest on January 8, 2014http://jhered.oxfordjournals.org/
detected among the loci studied, and the isolated state of the island localities with respect to the continental range of G. h. howdenae. We provided a broad prior distribution for the values of Ne based on informal estimates of population abundance over multiple field seasons (100 individuals for
the lower bound on the prior and either 1000 or 4000 for the upper bound). Ne values have been reported as typically much lower than census size (Frankham 1995). Empirical studies have reported Ne/N to vary widely between popu-lations depending on a variety of conditions, including sex
Table 5 Evidence for significant population size reduction through significant heterozygote excess
Sign test Wilcoxon test
Mode shiftMutational model P value Mutational model P value
All loc IAM 0.0292* IAM 0.0059* Normal (L shaped)SMM 0.4764 SMM 0.4219TPM 0.1561 TPM 0.0098*
AZ IAM 0.0354* IAM 0.0039* Normal (L shaped)SMM 0.5024 SMM 0.1914TPM 0.0431* TPM 0.0039*
NP IAM 0.0374 * IAM 0.0137* Normal (L shaped)SMM 0.4985 SMM 0.4727TPM 0.1810 TPM 0.0195*
Reported P values for the Sign and Wilcoxon tests (one tailed for excess heterozygosity) for departure from mutation–drift equilibrium and occurrence of allele frequency distribution mode shift across all loci for all localities (All loc), localities in the AZ, and localities in the NP.Mutational models used in Bottleneck: IAM, strict (1-step) SMM, and TPM.
Table 6 Evidence for significant overall population size reduction through significant M ratio with chosen model parameters for Theta, size, and percentage of nonstepwise mutation
Group Theta %Nonstep Size nonstep M P value
All Loc 4 0.67 4.6 0.414061 0.0224*AZ 4 0.64 4.8 0.411157 0.0923NP 4 0.64 4.7 0.410856 0.0951
Ratio calculated for all localities (All Loc), localities in the Agricultural Zone (AZ), and localities in the National Park (NP).*P < 0.05
Table 4 Proportion of membership in clusters inferred from Structure v. 2.3 analyses of the introduced populations of Galapaganus howdenae howdenae
Symmetry of assignments in the No loc prior runs illustrated through the range of sum of squared differences (SSD) between proportion of membership and 1/k scaled by the number of values (N) calculated across 10 runs per set of conditions for different numbers of inferred clusters (k = 2–6).Preponderance of assignments to a major cluster in the loc prior runs illustrated through the range of mean and SD values of proportion of membership across runs for the cluster with the largest (L) and smallest (S) proportion of assignments calculated across 10 runs per set of conditions for different num-bers of inferred clusters (k = 2–6).Each run set was performed with different allele frequency models (independent and correlated), different ancestry models (admixture and no admixture) and with or without the addition of locality prior information to aid in the clustering (Loc prior and No loc prior).
by guest on January 8, 2014http://jhered.oxfordjournals.org/
Mok et al. • Genetics of a Weevil Introduction to Galápagos
9
ratios, variance in family sizes, and fluctuations in population size (Falconer 1989; Mace and Lande 1991). An additional factor in the calculation is the transformation of generations into calendar years; given that there are no ecological studies for G. h. howdenae, we incorporated generation time data from closely related species (Lapointe 2000; Ju et al. 2011) alter-natively using 1 or 0.5 years for all mSvar runs (either 1 or 2 generations per year).
Even though simulation studies have shown that the scaled parameters are more precisely estimated than the natural parameters and advocate drawing conclusions from the scaled rather than the natural parameters (Girod et al. 2011), our interest in estimating the timing of the popu-lation size reduction and presumably the introduction of G. h. howdenae into Galápagos, independently of the uncer-tainties described above, led us to implement the method of Storz and Beaumont (2002) to directly estimate N0, Na (or N1), and tf (Figure 4). Fourteen independent runs were performed with different combination of priors, hyperpri-ors, starting values, and run lengths (Supplementary Material Table 1). Runs with the same priors were repeated 3 times as an initial test for convergence (Goossens et al. 2006; Bourke et al. 2010).
For all mSvar runs (either in version 0.4 or 1.3) we dis-carded the first 10% of the updates. The outputs were ana-lyzed using the Locfit and Coda packages in R (Loader 1996; Plummer et al. 2006; R-Core-Team, 2012). Convergence and stationarity of individual chains were checked visually and using the Geweke (1992) and Heidelberger and Welch (HW) (1983) diagnostic statistics. Results of individual runs were
considered only if all variables (No, Na, tf, and θ) passed at least 1 test (provided that the values for the Geweke statistic Z were not too far beyond the 2.5% and 97.5% quantiles of the null distribution [Normal 0,1] and displayed a P value larger than 0.05 for the HW statistic). All runs were repeated after the exclusion of 10 individuals consistently assigned to smaller secondary clusters in the Structure 2.3 runs.
ResultsSubstantial Genetic Diversity
Linkage disequilibrium tests showed that the majority of microsatellite loci pairs in G. h. howdenae localities (27 out of the 28 pairwise comparisons) do not contain alleles that are inherited together more often than expected by chance (P > 0.05) and that the remaining comparison rejects independ-ence only marginally (P = 0.048).
Genetic diversity within assigned groups of localities (AZ and NP) and within all localities grouped together was assessed through calculation of average number of alleles per locus and expected and observed heterozygosity val-ues. The average expected heterozygote frequency across all localities was 0.5388, and the number of alleles per locus ranged from 2 to 7 (mean = 3.875, Table 2). A slightly larger mean number of alleles across all loci was recorded for the NP area compared with the AZ; however, those differences are not significant (P = 0.8, Table 2). Similarly, neither the
Figure 3. Population size change. Posterior distributions of the demographic parameters on a logarithmic scale with the Beaumont (1999) method. r = N0/N1 represents the ratio of present to past populations size; (tf): (ta/N0) represents the ratio between the time in generations of the population change and the present population size.
Figure 4. Ancestral and present population sizes. Posterior distributions on a logarithmic scale for past (N1) and current (N0) effective population sizes from the Stortz and Beaumont (2002) method. Results are shown using 2 alternative generation times (g). Each line corresponds to a different run.
by guest on January 8, 2014http://jhered.oxfordjournals.org/
mean observed nor the expected heterozygosity values dif-fer significantly across the 2 areas (P = 0.5403, for both HO and HE). As a comparison point, values of mean number of alleles from the introduced localities of G. h. howdenae are in line with those derived from 13 localities of G. conwayen-sis, a long-established endemic in the same island, (A = 3.3, SD = 0.8, range [1.9–4.9], data available from the authors), albeit with a different set of microsatellite loci. This suggests that Galápagos populations of G. h. howdenae bear genetic variability comparable with populations of long-established endemic close relatives, despite their supposedly recent establishment.
Overall, deviation from Hardy–Weinberg equilibrium measured through FIS averaged across all the loci and all localities resulted in significant deficiency of heterozygotes (P < 0.01) (Table 2). Given the presumed recent introduction of G. h. howdenae into Santa Cruz, this is not an unexpected result because evolutionary forces such as genetic drift could be acting within the introduced range (Stepien et al. 2010).
Genetic Homogeneity Across the Introduced Range
Results from the analyses of molecular variance (amova) indicate that the majority of the genetic variation present lies within localities and that there is no marked structuring among collecting sites. FST values calculated among all locali-ties, AZ localities and NP localities, were all nonsignificant (Table 3). In every case, the variance among populations was a very small percentage of the total variance, ranging from 0.06–1.36% (Table 3) and nonsignificant FST values. Moreover, these results are in line with findings derived from mitochondrial DNA sequence variation for a smaller number of G. h. howdenae localities (Sequeira et al. 2012). The lack of significant genetic differentiation among localities is also consistent with expectations considering the flight capability of G. h. howdenae and the potential homogenizing effects of mobility in disturbed environments.
When localities are grouped according to their placement in either the AZ or the NP, a significant P value was reported, even though the variance between these 2 groups was a small percentage of the total variance (Table 3). Additionally, signif-icant pairwise FST values were only found between localities in the AZ and the NP (6 among the 36 pairwise compari-sons performed), not within either of the groups. Pairwise FST values were also significant when we compared clusters including all individuals from each locality area (FST = 0.015). This suggests that despite a general trend indicating a geneti-cally homogenous population across the introduced range of G. h. howdenae in Santa Cruz, there is subtle genetic differen-tiation between the AZ and NP areas.
Contemporary Effective Population Size
Using ONeSAMP, our median estimate of Ne was 184.37 with 95% credible limits (CL) of 139.01–387.52 breed-ing individuals when using 100 and 4000 for the lower and upper priors, respectively. However, in this particular case, ONeSAMP values were not robust to changes in the priors
and produced a lower estimate when using 1000 for the upper prior (median Ne = 60.09 with 95% CL of 51.98–81.78 breeding individuals). We chose to use the lower estimate, Ne = 60.09, for quantifying the timing of population decline (tf) from MSvar 0.4, given that the range of values seemed more in line with our field observations of weevil abundance over multiple field seasons.
No Clear Evidence of Multiple Introductions
Differences between values of ln likelihood of the data (Ln PD) for increasing number of inferred clusters provide limited information about the likeliest value of k and the existence of structure across the introduced range for G. h. howdenae in Galápagos; under some conditions indicating a single clus-ter as the most likely scenario while in others suggesting up to four clusters (Figure 2A, Supplementary Materials Figure S1A). Overall, the differences among the ln likelihood values of runs with increasing k are modest, in some cases as small as 50, and often associated with large variance within each set of runs. Even in the light of the Delta K results suggesting alternatively k = 2 or k = 3 as the uppermost hierarchical level of structure (Figure 2B, Supplementary Materials Figure S1B), we are skeptical to accept this as an indication of real structure (and potentially of mixed origins of samples within the introduced range) because no other indicators support this result. In every case, random clustering models, assum-ing k populations = 2–6, showed symmetrical proportions of membership in said number of clusters (Table 4). In the face of real structure, proportions would be asymmetric and the probability that a random individual had recent ancestry in a particular population would be skewed. The correlated allele frequency model showed slightly more asymmetric mem-bership proportions than the independent allele frequency model (Table 4), but there was still no strong asymmetric assignment for 1 cluster, as demonstrated by an almost negli-gible deviations from complete symmetry in the assignments. Without incorporating geographical information, we find no strong support for the notion that the current populations of G. h. howdenae in Santa Cruz could be the product of multiple introductions from already differentiated source populations.
When incorporating the locality information (LocPrior option) the proportion of membership within 1 cluster was higher for both of the allele frequency models (Table 4). A very large proportion of individuals from both predefined zones (AZ and NP) were assigned to a single mixed cluster (even with higher numbers of assumed clusters); whereas a few, usually less than 10%, were assigned to one of the other clusters (Table 4). The inferred ancestry of individuals assigned to any cluster other than the larger cluster was above 0.6 only when the models incorporated correlated allele fre-quencies (either with or without admixture). The individuals originate from localities within both areas: in most “admix-ture” runs from NP locality SR18 and AZ, SR11 and SR07, whereas most “no-admixture” runs add 1 or 2 individu-als from NP localities SR01 and SR17 and AZ, and SR22. Although simulations indicate that the LocPrior model does not detect structure when there is none (Hubisz et al., 2009),
by guest on January 8, 2014http://jhered.oxfordjournals.org/
Mok et al. • Genetics of a Weevil Introduction to Galápagos
11
the contribution of a second introduction pulse seems very small compared with the one that originated the majority of the introduced G. h. howdenae weevils now distributed across the AZ and NP. An additional indicator of the importance of location information under the LocPrior model is the value of r, which parameterizes the amount of information carried by the locations. High values of r, such as the ones obtained in our analyses of the G. h. howdenae data set (from 0.1 to 16.3), suggest that either there is no population struc-ture or the structure is independent of the localities. Given the mixed composition of the smaller cluster with individuals from both AZ and NP, it becomes clear that those a priori locality areas we established (AZ and NP) bear no biological or historical significance. The current composition across the introduced range could be the combination of descendants from a larger introduction pulse either preceded or followed by a minor introduction event.
Evidence of a Past Reduction in Population Size
Tests in Bottleneck 1.2.02 for excess heterozygosity with respect to the mutation–drift equilibrium gave differing results, depending on the test and underlying mutational model employed. For the IAM, a significant excess het-erozygosity was found with both the Sign and Wilcoxon one-tailed test for excess for all the locality configurations tested (Table 5). Neither test found significant values under the stepwise mutational model (SMM). Under the TPM, only the Wilcoxon test found significantly excess heterozygosity across all localities, localities in the AZ and NP, though the Sign test under TPM was significant for the AZ. SMM and TPM are generally more appropriate than IAM for micro-satellite loci with TPM being the more suitable of the 2 (Di Rienzo et al. 1994; Luikart et al. 1998; Piry et al. 1999). Bottleneck 1.2.02’s dependency on mutational model for detection of excess heterozygosity should be taken into con-sideration. However, given that the Wilcoxon test is more powerful than the other tests with a small number of loci (<20), the significant P values found under TPM across all the populations using the Wilcoxon one-tailed test for excess should not be disregarded. Thus, the excess heterozygosity could support the hypothesis of G. h. howdenae populations having undergone a reduction in population size within 2–4 Ne generations (Luikart et al. 1998; Piry et al. 1999). The mode shift test did not find significant deviations from the normal L-shaped allele frequency distribution in the analy-ses for all localities, AZ and NP, which is contrary to our expectations.
The significant M ratio found for all individuals as a single group and its implication for population reduction is congru-ent with the excess heterozygosity results (Table 6). Although the populations in the AZ and NP analyzed separately did not show significantly reduced M ratios, all localities grouped together showed a significant reduction (Table 6). Thus, both moment-based estimator tests support the hypothesis that G. h. howdenae populations as a whole have undergone a pop-ulation bottleneck potentially due to a significant reduction in population size during their introduction to Santa Cruz.
Likelihood-based estimators also support the proposed sce-nario of a population size reduction in the past (Figures 3 and 4). Using the method of Beaumont (1999), we found a clear signal of past population decline (Figure 3). Under a model of exponential change, the mode for the rate of size change was estimated as log −3.62 (with a 95% highest probability density interval, −4.41; −2.93). These estimates correspond to declines larger than 1000-fold in effective population size. Using the method of Stortz and Beaumont (2002) the signal for a past reduction in population size is also quite clear under multiple starting conditions (Figure 4). Interestingly, despite the sharp decline, likelihood estimates of current population size indicate substantial populations (Mode N0 298 individuals, 95% HPD 50–2300)—larger than anticipated from field observations.
Simulation studies have warned that even the full-likeli-hood Bayesian methods used here to detect population size changes (Storz and Beaumont 2002) could be affected by the presence of population structure in the sample and/or by the lack of coverage in the sampling scheme (Chikhi et al. 2010). However, we contend that the FST values reported here are mostly nonsignificant or suggest very subtle structuring across the introduced range and that our sampling covers the entirety of the known distribution of G. h. howdenae in Santa Cruz Island, making any population size changes detected less likely to result from spurious signals.
Timing of the Population Size Change
Likelihood-based estimators provide differing assessments of the timing of the reduction in population size, depending on the value used in the calculations to incorporate the period between reproductive cycles (generation time) and whether we estimate the scaled parameters through mSvar 0.4 or the natu-ral parameters through mSvar 1.3. When using 1 year as the generation time, mSvar 1.3 estimates the mode of the density distribution of the timing of population size change as 90 years ago, whereas the mode is around 119 years ago when incor-porating the shorter generation time in the initial conditions (t = 0.5). The upper and lower limits of the 95% HPD interval of the timing of the population size change under both sets of conditions suggest a decline between 50 and 250 years ago. Admittedly, this provides a wide time frame for the population size change and presumably the introduction of G. h. howdenae to Santa Cruz. The window for this change broadens further and encompasses an earlier time frame when deriving the timing from the scaled parameters (ta = tf × N0). Despite the narrower range obtained for the scaled time parameter itself (90% HPD interval log tf 0.48–0.82), the combination of the uncertainties derived from 1) the time interval between reproductive cycles included in the calculation to transform generations into years since the population size change (1 vs. 2 generations per year), 2) the appropriate Ne/N ratio used for this particular situation (0.8, 0.5, or 0.3), and 3) the 95% CL of the Ne values produced by ONeSAMP generate a broader and less informative time estimate for the introduction (107–800 years ago). In an attempt to narrow the timing of the major introduction of G. h. howde-nae, we excluded the individuals that were consistently assigned to a smaller, distinct genetic cluster in the Structure v. 2.3 runs
by guest on January 8, 2014http://jhered.oxfordjournals.org/
(a total of 10 individuals) that presumably could represent a separate smaller introduction pulse. With the culled data set, the estimates of the timing of the population size change and, pre-sumably, the predicted time frame for the larger introduction to Santa Cruz Island narrow somewhat. mSvar 1.3 estimates the mode of the timing of introduction to 65 years ago (95% HPD interval 35–90) using 1 generation per year in the calculations. Interpreting the results derived from both the culled and com-plete data sets under either set of run conditions, the reduction in population size could be interpreted as a signature of an acci-dental introduction in historical times either contemporaneous or at the end of Gonzalez’s et al. (2008) colonization period (1862–1959) together with the establishment, rather than the later growth, of human settlements.
DiscussionRange Expansion and Connectivity Maintaining Genetic Diversity From a Single Introduction Pulse
One of the possible mechanisms proposed to mitigate the allele loss that can be brought about by a population bot-tleneck is a rapid increase in population size once the new population has been established (Zeisset and Beebee 2003; Zenger et al. 2003). Similarly, to the situation posed by a natu-ral colonization, the reduction in population size will con-tribute more substantially to the loss of genetic variability of the newly founded population if the bottleneck is sustained and even repeated (Clegg et al. 2002). Microsatellite genetic variation in European rabbit populations (Oryctolagus cunicu-lus) in Australia showed that a rapid population expansion at the time of establishment severely diminished the founder effects normally associated with introduction (Zenger et al. 2003). Lack of reduced genetic variance was observed in the introduced rabbit populations, whereas historical records of the rabbit expansion across Australia gave sufficient support for population growth as a mechanism for genetic diversity retention (Zenger et al. 2003). Similarly, a study of the well-documented marsh frog introduction in Britain revealed that the rapid reproductive success of Rana ridibunda resulted in effective colonization with little bottleneck effect on genetic diversity (Zeisset and Beebee 2003). Indeed, the authors concluded that significant bottlenecking was undetectable despite the small number of founding individuals.
Similarly, G. h. howdenae populations in Galápagos, though still displaying a signature of population size reduction, may not have detectable signals of long-term genetic loss. Measures of microsatellite genetic diversity in introduced populations are in line with that of populations of long-established endemic close relatives in the same island (in both number of alleles and heterozygosity), even though the most valuable contrast with G. h. howdenae populations in the native rage is not available. Additionally, mitochondrial DNA varia-tion found in some G. h. howdenae populations was comparable with that of younger endemics close relatives (Sequeira et al. 2012). As for the processes underlying the retention of diver-sity, one possible explanation is that the observed expansion in range might also be accompanied by an increase in census
numbers after a lag phase following the initial colonization (Sakai et al. 2001; Holt et al. 2005). During the last 2 collect-ing trips to the Galápagos in 2010 and 2011, G. h. howdenae individuals were found farther away from the AZ on the west-ward side of Santa Cruz as well as at high altitudes in the NP. In the context of the flight capability of G. h. howdenae, it is reasonable to predict that the range expansion will continue across the island, possibly instigating a rapid revival of genetic diversity.
In addition to an expansion in range, the lack of strong genetic structuring found across G. h. howdenae localities sup-ports the notion that gene flow could be a mechanism for homogenizing the genetic diversity present. Though sub-tle but still significant structuring was detected between the AZ and NP localities, the amount of variance between the grouped populations was too small to ascribe any biological significance to those groupings. Mishra and coauthors (2006) suggested that frequent gene flow contributed to the high levels of genetic diversity found within the localities of the fungal pathogen Fusarium pseudograminearum. In the case of the fungus, the novel genotypes resulting from regular gene flow and random mating were suggested to contribute to range expansion and heightened disease severity (Mishra et al. 2006). In G. h. howdenae populations, lack of genetic struc-turing suggests there is regular gene flow between relatively distant localities, which would in turn increase the available mating pool for individuals and could lead to eventual evo-lution of new genetic combinations. With observations sup-porting range expansion in G. h. howdenae, there is potential to quickly overcome any allele loss caused by the population bot-tleneck and potentially adapt to new territory in Santa Cruz. As noted by Keller and Taylor (2008), although it is tempting to ascribe phenotypic divergence to adaptive evolution, evolu-tion of phenotypic traits during dispersal, establishment, and range expansion of introduced populations may reflect either adaptive evolution or neutral phenotypic changes.
When disentangling the invasion history of island or conti-nental introductions, multiple introduction pulses and admix-ture have been associated with maintenance of genetic diversity (Kelager et al. 2013) and genetically diverse populations (Riva Rossi et al., 2012)—the establishment of distinct strains (Lee and Akimoto 2013), rapid evolution (Konečný et al. 2013), and increase of invasive potential (Kolbe et al. 2004; Kolbe et al. 2007; Roman and Darling 2007; Kolbe et al. 2008; Riva Rossi et al. 2012). Alternatively, admixture in the native range, long before the introduction occurred, has been ascribed as the underlying process generating exceptionally high genetic diversity in other recently introduced island populations (Tonione et al. 2011). However, it has also been suggested that the importance of admixture to invasion success has been overemphasized, because introductions are capable of being successful both in the pres-ence and absence of admixture (Chapple et al. 2013).
With no evidence for mixed ancestry from substantial popu-lation admixture, there is no support for the idea that G. h. howde-nae populations can rebuild genetic variability through 2 or more equally large introduction pulses. Rather, the contribution of admixture to the current genetic diversity levels seems mod-est at best, and the existing diversity needs to be attributed to in
by guest on January 8, 2014http://jhered.oxfordjournals.org/
Mok et al. • Genetics of a Weevil Introduction to Galápagos
13
situ demographic processes such as a rapid range expansion and growth after the original introduction bottleneck.
Bottleneck of G. h. howdenae During Introduction
Microsatellite markers and moment-based estimators have been successfully used to explore the introduction history of species with diverse life histories and dispersal strategies such as Chinese mitten crabs (Eriocheir sinensis) (Herborg et al. 2007) and the common ragweed (Ambrosia artemisiifolia) in Europe (Genton et al. 2005) and invasive house finches (Carpodacusm mexicanus) in North America (Hawley et al. 2006). The sen-sitivity of microsatellite markers to reductions in popula-tion size up to 500 generations in the past provides a broad time frame for the introductions to be studied (Garza and Williamson 2001). Results from moment-based estimators in G. h. howdenae support the hypothesis that introduced wee-vil populations on Santa Cruz have suffered a reduction in population size in the past. These findings are consistent with theoretical and empirical studies of recently introduced popu-lations, where allelic richness was reduced to a greater extent than heterozygosity, disturbing the mutation–drift equilibrium (Cornuet and Luikart 1996; Allendorf and Lundquist 2003).
Although microsatellite markers continue to be widely employed to study the introduction history of a wide array of organisms (Garcia et al. 2012; Frisch et al. 2013; Ghabooli et al. 2013; McMahon et al. 2013; Perdereau et al. 2013; Simpson et al. 2013), more recently, moment-based esti-mators have been under scrutiny and their effectiveness in detecting population declines has been questioned (Peery et al. 2012). Bayesian coalescent methods have been adopted to infer demographic histories (Jones et al. 2013; Paz-Vinas et al. 2013), including the detection of severe population bot-tlenecks (Bourke et al. 2010). Several recent studies remark that in order to avoid spurious signals of population decline, details of the dispersal strategies and population structure of the species under study cannot be overlooked (Girod et al. 2011; Paz-Vinas et al.2013). The results of the likelihood-based estimators in island populations of G. h. howdenae sup-port the notion of a reduction in population size possibly linked to introduction. This result has 2 implications. The first is that in G. h. howdenae the recovery from the genetic depletion that should accompany the introduction can be accomplished by mechanisms other than admixture and independent from multiple introductions. The second implication is that along with the substantial current population sizes estimated (even though significantly smaller than the estimated pre-introduc-tion ancestral sizes of putatively continental populations) we have also observed a concurrent expansion in range that now places endemic plants at an increased risk of herbivory.
An Older Introduction to Galápagos
The timing of other human-mediated plant and animal intro-ductions to Galápagos (either confirmed or suspected) spans all of Gonzalez’s (2008) time periods from when humans first arrived in the archipelago, all the way to ongoing introduc-tions (Bataille et al. 2009; Jager et al. 2009; Caetano et al. 2012). Thorough reviews focusing on insect fauna in Galápagos
show that accidental insect introductions, among those bee-tles, have never been uncommon (Roque-Albelo and Causton 1999; Causton et al. 2000; Causton et al. 2006); humans have accidentally carried more than 22% of the presently known total beetle fauna (Peck 2005). The first records of an intro-duced beetle species date back to 1835: Dermestes maculatus was presumably accidentally transported into the islands with stored products. The most probable mode of accidental intro-duction proposed for G. h. howdenae is accompanying plant products (Peck 2005). Our most conservative estimate of the timing of the introduction of G. h. howdenae to Santa Cruz predates the first collecting record in 1992 by at least four dec-ades whereas our most extreme ones by almost a century. The time estimates for the introduction of G. h. howdenae into Santa Cruz supports our predictions of an accidental transfer dur-ing the colonization period (1832–1959). By most estimates, the introduction of G. h. howdenae into the AZ predates the spurt in human population growth that took place between the 1950s and 2001. The timing of arrival of an introduced species can feature in the estimation of its invasive potential, given that it can inform the speed of dispersal and naturaliza-tion of the new arrival. Following scoring system of Causton et al. (2006) for predicting invasiveness in the archipelago, an older introduction would render a lower invasive ranking.
Conservation Implications
Fitness loss associated with introductions and population bot-tlenecks may occur due to diminished genetic variation and exposure of detrimental alleles (Colautti et al. 2005). Sufficient genetic variability and its presumed links to fitness are neces-sary to permit adaptation and expansion of an introduced species in the introduced range (Genton et al. 2005). Despite the intermediate invasive ranking attached, G. h. howdenae has successfully adapted to at least 2 different environments in Santa Cruz, and its flight capability suggests that it could also spread further across the island and into other islands. Given the results of this study, we must incorporate to our current knowledge the combined effect of substantial popu-lation sizes and expansions in introduced range and feeding preferences. Because this introduced weevil’s herbivorous preferences include several endemic host plants (Scalesia and Tournefortia), the ecological impact would not only be limited to damage to endemic plants, but the forced resource shar-ing between introduced and endemic Galapaganus weevils, specifically those with specialized highland habitats and with restricted populations. Even though not of the highest con-servation priority, the situation merits consideration when bearing in mind the central geographical location and logisti-cal relevance of Santa Cruz within the Galápagos archipelago. Results from this study as well as of others evaluating trans-port of alien insect species within the archipelago (Roque Albelo et al., 2006) can inform the design of measures that attempt to control accidental interisland transfer.
It is unclear what effect, if any, interactions between endemic and introduced Galapaganus species are having on the genetic pool of both species. Given that the introduced and endemic Galapaganus now share endemic plant hosts
by guest on January 8, 2014http://jhered.oxfordjournals.org/
throughout part of the introduced range, future behavioral and ecological studies should focus on the details of their interactions and explore the possibility of interspecies hybrid-ization. Furthermore, little is known about the host prefer-ences and genetic status of source populations within the native range in Ecuador, and samples from those locations are currently unavailable. Direct comparisons between source and introduced populations of G. h. howdenae may clarify whether the genetic patterns observed in island populations are only a function of the introduction history of these localities and whether they are related to invasive potential of G. h. howdenae . If genetic variability of the introduced populations proves to be higher than in the native range, this could indicate that introduced G. h. howdenae may be predisposed for success-ful establishment and expansion beyond their present range. Thorough sampling of the native range would aid in identify-ing the precise geographic source of this island introduction.
Supplementary MaterialSupplementary material can be found at http://www.jhered.oxfordjournals.org/.
FundingNational Science Foundation (0817978 to A.S.S.); the Brachman Hofmann Funds (to A.S.S); the Howard Hughes Medical Institute (52006325 to C.C.S and A.S.S. through Wellesley College).
AcknowledgmentsScientific research permits for the collecting trips were obtained from the Parque Nacional Galápagos through the Department of Terrestrial Invertebrates at the Charles Darwin Research Station (Santa Cruz). We gratefully acknowledge A. Lanteri, S. Cárdenas, L. Cruz (and the crew from “El Pirata”), H. Herrera, P. Lincagno, A. Mieles, A. M. Ortega, J. Rosado, and station volunteers for invaluable assistance in the field. Field logistical sup-port was provided by S. Cisneros, P. Couenberg, and R. Pépolas, Division of Visiting Scientists at the Charles Darwin Research Station.
ReferencesAlda F, Ruiz-Lopez MJ, Garcia FJ, Gompper ME, Eggert LS, Garcia JT. 2013. Genetic evidence for multiple introduction events of raccoons (Procyon lotor) in Spain. Biol Invasions. 15:687–698.
Allendorf F, Lundquist L. 2003. Introduction: population biology, evolution, and control of invasive species. Conserv Biol. 17:24–30.
Bailey K. 1976. Potassium-argon ages from the Galápagos Islands. Science. 192:465–467.
Bataille A, Cunningham AA, Cedeno V, Cruz M, Eastwood G, Fonseca DM, Causton CE, Azuero R, Loayza J, Martinez JD, et al. 2009. Evidence for reg-ular ongoing introductions of mosquito disease vectors into the Galápagos Islands. P Roy Soc B-Biol Sci. 276:3769–3775.
Beaumont MA. 1999. Detecting population expansion and decline using microsatellites. Genetics. 153:2013–2029.
Beggs JR, Brockerhoff EG, Corley JC, Kenis M, Masciocchi M, Muller F, Rome Q, Villemant C. 2011. Ecological effects and management of invasive alien Vespidae. Biocontrol. 56:505–526.
Bourke BP, Frantz AC, Lavers CP, Davison A, Dawson DA, Burke TA. 2010. Genetic signatures of population change in the British golden eagle (Aquila chrysaetos). Conserv Genet. 11:1837–1846.
Brodie ED. 2007. Population size is not genetic quality. Anim Conserv. 10:288–290.
Caetano S, Currat M, Pennington RT, Prado D, Excoffier L, Naciri Y. 2012. Recent colonization of the Galápagos by the tree Geoffroea spinosa Jacq. (Leguminosae). Mol Ecol. 21:2743–2760.
Causton CE, Peck SB, Sinclair BJ, Roque-Albelo L, Hodgson CJ, Landry B. 2006. Alien insects: Threats and implications for conservation of Galápagos Islands. Ann Entomol Soc Am. 99:121–143.
Causton CE, Zapata CE, Roque-Albelo L. 2000. Alien arthropod species deterred from establishing in the Galapagos, but how many are entering undetected? Noticias de Galapagos. 61:10–13.
Chapple DG, Miller KA, Kraus F, Thompson MB. 2013. Divergent introduc-tion histories among invasive populations of the delicate skink (Lampropholis delicata): has the importance of genetic admixture in the success of biological invasions been overemphasized? Divers Distrib. 19:134–146.
Chikhi L, Sousa VC, Luisi P, Goossens B, Beaumont MA. 2010. The con-founding effects of population structure, genetic diversity and the sampling scheme on the detection and quantification of population size changes. Genetics. 186:983–995.
Clegg SM, Degnan SM, Kikkawa J, Moritz C, Estoup A, Owens IP. 2002. Genetic consequences of sequential founder events by an island-colonizing bird. Proc Natl Acad Sci U S A. 99:8127–8132.
Cockerham C, Weir B. 1984. Estimating F-statistics for the analysis of popu-lation structure. Evolution. 38:1358–1370.
Cockerham CC, Weir BS. 1987. Correlations, descent measures: drift with migration and mutation. Proc Natl Acad Sci U S A. 84:8512–8514.
Colautti RI, Manca M, Viljanen M, Ketelaars HA, Bürgi H, Macisaac HJ, Heath DD. 2005. Invasion genetics of the Eurasian spiny waterflea: evidence for bottlenecks and gene flow using microsatellites. Mol Ecol. 14:1869–1879.
Cornuet JM, Luikart G. 1996. Description and power analysis of two tests for detecting recent population bottlenecks from allele frequency data. Genetics. 144:2001–2014.
Desender K, Baert L, Maelfait JP. 1992. Distribution and speciation of car-abid beetles in the Galápagos Archipelago (Ecuador). Bull Inst R Sci Nat Belg Entomol. 62:57–65.
Desender K, Casale A, Baert L, Maelfait J-P, Verdyck P. 2002. Calleida migra-toria Casale, new species (Coleoptera: Carabidae), a newly introduced ground beetle in the Galápagos Islands, Ecuador. Coleopt Bull. 56:71–78.
Di Rienzo A, Peterson AC, Garza JC, Valdes AM, Slatkin M, Freimer NB. 1994. Mutational processes of simple-sequence repeat loci in human popu-lations. Proc Natl Acad Sci U S A. 91:3166–3170.
Dudaniec RY, Gardner MG, Donnellan S, Kleindorfer S. 2008. Genetic vari-ation in the invasive avian parasite, Philornis downsi (Diptera, Muscidae) on the Galápagos archipelago. BMC Ecol. 8:13.
Earl DA, Vonholdt BM. 2012. STRUCTURE HARVESTER: a website and program for visualizing STRUCTURE output and implementing the Evanno method. Conserv Genet Resour. 4:359–361.
Ellstrand NC, Schierenbeck KA. 2000. Hybridization as a stimulus for the evolution of invasiveness in plants? Proc Natl Acad Sci U S A. 97:7043–7050.
Emerson BC, Kolm N. 2005. Species diversity can drive speciation. Nature. 434:1015–1017.
Evanno G, Regnaut S, Goudet J. 2005. Detecting the number of clusters of individuals using the software STRUCTURE: a simulation study. Mol Ecol. 14:2611–2620.
Evans EW. 2004. Habitat displacement of North American ladybirds by an introduced species. Ecology. 85:637–647.
by guest on January 8, 2014http://jhered.oxfordjournals.org/
Mok et al. • Genetics of a Weevil Introduction to Galápagos
15
Excoffier L, Laval G, Schneider S. 2005. Arlequin (version 3.0): an integrated software package for population genetics data analysis. Evol Bioinform Online. 1:47–50.Facon B, Pointier JP, Jarne P, Sarda V, David P. 2008. High genetic variance in life-history strategies within invasive populations by way of multiple intro-ductions. Curr Biol. 18:363–367.Falconer DS. 1989. Introduction to quantitative genetics. New York: Longman.Fitzpatrick BM, Shaffer HB. 2007. Hybrid vigor between native and intro-duced salamanders raises new challenges for conservation. Proc Natl Acad Sci U S A. 104:15793–15798.Frankham R. 1995. Conservation genetics. Annu Rev Genet. 29:305–327.Frisch D, Havel JE, Weider LJ. 2013. The invasion history of the exotic freshwater zooplankter Daphnia lumholtzi (Cladocera, Crustacea) in North America: a genetic analysis. Biol Invasions. 15:817–828.Garcia JT, Garcia FJ, Alda F, Gonzalez JL, Aramburu MJ, Cortes Y, Prieto B, Pliego B, Perez M, Herrera J, et al. 2012. Recent invasion and status of the raccoon (Procyon lotor) in Spain. Biol Invasions. 14:1305–1310.Garza JC, Williamson EG. 2001. Detection of reduction in population size using data from microsatellite loci. Mol Ecol. 10:305–318.Geist D. 1996. On the emergence and submergence of the Galápagos Islands. Not Galapagos. 56:5–9.Genton BJ, Shykoff JA, Giraud T. 2005. High genetic diversity in French invasive populations of common ragweed, Ambrosia artemisiifolia, as a result of multiple sources of introduction. Mol Ecol. 14:4275–4285.Geweke J. 1992. Evaluating the accuracy of sampling-based approaches to calculating posterior moments. In: Bernardo JM, Berger JO, Dawiv AP, Smith AFM, editors. Bayesian statistics, Vol. 4. Oxford: Clarendon Press. p. 169–192.Ghabooli S, Zhan AB, Sardina P, Paolucci E, Sylvester F, Perepelizin PV, Briski E, Cristescu ME, MacIsaac HJ. 2013. Genetic diversity in introduced golden mussel populations corresponds to vector activity. PLoS ONE. Advance Access published March 22, 2013, doi:10.1371/journal.pone.0059328Girod C, Vitalis R, Leblois R, Fréville H. 2011. Inferring population decline and expansion from microsatellite data: a simulation-based evaluation of the Msvar method. Genetics. 188:165–179.Goldstein DB, Schlotterer C. 1999. Microsatellites: evolution and applica-tions. Oxford: Oxford University Press.Gonzalez JA, Montes C, Rodriguez J, Tapia W. 2008. Rethinking the Galápagos Islands as a complex social-ecological system: Implications for conservation and management. Ecol Soc. 13:1–26.Goossens B, Chikhi L, Ancrenaz M, Lackman-Ancrenaz I, Andau P, Bruford MW. 2006. Genetic signature of anthropogenic population collapse in orang-utans. PLoS Biol. 4:e25.Guo SW, Thompson EA. 1992. Performing the exact test of Hardy-Weinberg proportion for multiple alleles. Biometrics. 48:361–372.Haldane J. 1953. An exact test for randomness of mating. Symp. Soc. Exp. Biol. 7:274–289.Hawley DM, Hanley D, Dhondt AA, Lovette IJ. 2006. Molecular evidence for a founder effect in invasive house finch (Carpodacus mexicanus) popula-tions experiencing an emergent disease epidemic. Mol Ecol. 15:263–275.Herborg LM, Weetman D, van Oosterhout C, Hänfling B. 2007. Genetic pop-ulation structure and contemporary dispersal patterns of a recent European invader, the Chinese mitten crab, Eriocheir sinensis. Mol Ecol. 16:231–242.Hickman CS, Lipps JH. 1985. Geologic youth of Galápagos islands confirmed by marine stratigraphy and paleontology. Science. 227:1578–1580.Holt RD, Keitt TH, Lewis MA, Maurer BA, Taper ML. 2005. Theoretical models of species’ borders: single species approaches. Oikos. 108:18–27.Hubisz MJ, Falush D, Stephens M, Pritchard JK. 2009. Inferring weak popu-lation structure with the assistance of sample group information. Mol Ecol Resources. 9:1322–1332.Huxel GR. 1999. Rapid displacement of native species by invasive species: Effects of hybridization. Biol Conserv. 89):143–152.
Jager H, Kowarik I, Tye A. 2009. Destruction without extinction: long-term impacts of an invasive tree species on Galápagos highland vegetation. J Ecol. 97:1252–1263.
Johnson JA, Tingay RE, Culver M, Hailer F, Clarke ML, Mindell DP. 2009. Long-term survival despite low genetic diversity in the critically endangered Madagascar fish-eagle. Mol Ecol. 18:54–63.
Jones FA, Ceron-Souza I, Hardesty BD, Dick CW. 2013. Genetic evidence of Quaternary demographic changes in four rain forest tree species sampled across the Isthmus of Panama. J Biogeogr. 40:720–731.
Ju RT, Wang F, Wan FH, Li B. 2011. Effect of host plants on develop-ment and reproduction of Rhynchophorus ferrugineus (Olivier) (Coleoptera: Curculionidae). J Pest Sci. 84:33–39.
Kelager A, Pedersen JS, Bruun HH. 2013. Multiple introductions and no loss of genetic diversity: invasion history of Japanese Rose, Rosa rugosa, in Europe. Biol Invasions. 15:1125–1141.
Keller SR, Taylor DR. 2008. History, chance and adaptation during biologi-cal invasion: separating stochastic phenotypic evolution from response to selection. Ecol Lett. 11:852–866.
Klimaszewski J, Peck SB. 1998. A review of Aleocharine Rove Beetles from the Galápagos Islands, Ecuador (Coleoptera: Staphylinidae, Aleocharinae). Rev Suisse Zool. 105:221–260.
Kolbe JJ, Glor RE, Schettino LR, Lara AC, Larson A, Losos JB. 2007. Multiple sources, admixture, and genetic variation in introduced Anolis lizard populations. Conserv Biol. 21:1612–1625.
Kolbe JJ, Glor RE, Rodríguez Schettino L, Lara AC, Larson A, Losos JB. 2004. Genetic variation increases during biological invasion by a Cuban liz-ard. Nature. 431:177–181.
Kolbe JJ, Larson A, Losos JB, de Queiroz K. 2008. Admixture determines genetic diversity and population differentiation in the biological invasion of a lizard species. Biol Lett. 4:434–437.
Konečný A, Estoup A, Duplantier JM, Bryja J, Bâ K, Galan M, Tatard C, Cosson JF. 2013. Invasion genetics of the introduced black rat (Rattus rattus) in Senegal, West Africa. Mol Ecol. 22:286–300.
Krushelnycky PD, Gillespie RG. 2010. Correlates of vulnerability among arthropod species threatened by invasive ants. Biodivers Conserv. 19:1971–1988.
Lanteri AA. 1992. Systematics cladistics and biogeography of a new weevil genus Galapaganus (Coleoptera Curculionidae) from the Galápagos Islands and coasts of Ecuador and Perú. Trans Am Entomol Soc (Phila). 118:227–267.
Lanteri AA. 2004. New taxonomic and biogeographic information on Galapaganus femoratus species group (Coleoptera: Entiminae). Trans Am Entomol Soc (Phila). 130:177–192.
Lapointe SL. 2000. Thermal requirements for development of Diaprepes abbreviatus (Coleoptera: Curculionidae). Environ Entomol. 29:150–156.
Lee W, Akimoto S. 2013. Recent multiple introductions of the gall-forming aphid Pemphigus bursarius into Japanese islands. Entomol Sci. 16:248–251.
Lindolm A, Breden F, Alexander H, Chan W, Thakurta S, Brooks R. 2006. Invasion success and genetic diversity of introduced populations of guppies Poecilia reticulata in Australia. Mol Ecol. 14:3671–3682.
Loader CR. 1996. Local likelihood density estimation. Ann Stat. 24:1602–1618.
Luikart G, Sherwin WB, Steele BM, Allendorf FW. 1998. Usefulness of molecular markers for detecting population bottlenecks via monitoring genetic change. Mol Ecol. 7:963–974.
Mace GM, Lande R. 1991. Assessing extinction threats - toward a reevalua-tion of IUCN threatened species categories. Conserv Biol. 5:148–157.
Mann HB, Whitney DR. 1947. On a test of whether one of two random variables is stochastically larger than the other. Ann Math Stat. 18:50–60.
McMahon CR, Isagi Y, Kaneko S, Bowman DMJS, Brook BW, Bradshaw CJA. 2013. Genetic structure of introduced swamp buffalo subpopulations in tropical Australia. Austral Ecol. 38:46–56.
by guest on January 8, 2014http://jhered.oxfordjournals.org/
Mishra PK, Tewari JP, Clear RM, Turkington TK. 2006. Genetic diversity and recombination within populations of Fusarium pseudograminearum from western Canada. Int Microbiol. 9:65–68.
Nei M, Maruyama T, Chakraborty R. 1975. The bottleneck effect and genetic variability in populations. Evolution. 29:1–10.
Paz-Vinas I, Quemere E, Chikhi L, Loot G, Blanchet S. 2013. The demo-graphic history of populations experiencing asymmetric gene flow: combin-ing simulated and empirical data. Mol Ecol. 22:3279–3291.
Peck SB. 2005. The beetles of the Galápagos Islands, Ecuador: evolution, ecology, and diversity (Insecta: Coleoptera). Ontario, Canada: NRC Research Press.
Peck SB, Heraty J, Landry B, Sinclair BJ. 1998. Introduced insect fauna of an oceanic archipelago: The Galápagos Islands, Ecuador. American Entomologist. 44:218–237.
Peery MZ, Kirby R, Reid BN, Stoelting R, Doucet-Bëer E, Robinson S, Vásquez-Carrillo C, Pauli JN, Palsbøll PJ. 2012. Reliability of genetic bottleneck tests for detecting recent population declines. Mol Ecol. 21:3403–3418.
Perdereau E, Bagnères AG, Bankhead-Dronnet S, Dupont S, Zimmermann M, Vargo EL, Dedeine F. 2013. Global genetic analysis reveals the putative native source of the invasive termite, Reticulitermes flavipes, in France. Mol Ecol. 22:1105–1119.
Piry S, Luikart G, Cornuet J. 1999. BOTTLENECK: a computer program for detecting recent reductions in the effective population size using allele frequency data. J Hered. 90:500–503.
Plummer M, Best N, Cowles K, Vines K. 2006. CODA: convergence diag-nosis and output analysis for MCMC. R News. 6:7–11.
Pritchard JK, Stephens M, Donnelly P. 2000. Inference of population struc-ture using multilocus genotype data. Genetics. 155:945–959.
Puillandre N, Dupas S, Dangles O, Zeddam JL, Capdevielle-Dulac C, Barbin K, Torres-Leguizamon M, Silvain JF. 2008. Genetic bottleneck in invasive species: the potato tuber moth adds to the list. Biol Invasions. 10:319–333.
R-Core-Team. 2012. A language and environment for statistical comput-ing [Internet]. Vienna, Austria: R Foundation for Statistical Computing. Available from: http://www.R-project.org/.
Raymond M, Rousset F. 1995. GENEPOP (Version 1 2): Population genet-ics software for exact tests and ecumenicism. J Hered. 86:248–249.
Roman J, Darling JD. 2007. Paradox lost: genetic diversity and the success of aquatic invasions. Trends in Ecology and Evolution. 22:454–464.
Roque Albelo L, Berg M, Galarza M. 2006. “Polizontes pelogrosos”, disper-sión de insectos entre las islas Galápagos en barcos de turismo. Puero Ayora, Galapagos: Charles Darwin Research Station.
Roque-Albelo L, Causton C. 1999. El Nino and introduced insects in the Galapagos Islands: Different dispersal strategies, similar effects. Noticias de Galapagos. 60:30–36.
Riva Rossi CM, Pascual MA, Aedo Marchant E, Basso N, Ciancio JE, Mezga B, Fernández DA, Ernst-Elizalde B. 2012. The invasion of Patagonia by Chinook salmon (Oncorhynchus tshawytscha): inferences from mitochondrial DNA patterns. Genetica. 140:439–453.
Sakai AK, Allendorf FW, Holt JS, Lodge DM, Molofsky J, With KA, Baughman S, Cabin RJ, Cohen JE, Ellstrand NC, et al. 2001. The population biology of invasive species. Annu Rev Ecol Syst. 32:305–332.
Sequeira AS, Lanteri AA, Albelo LR, Bhattacharya S, Sijapati M. 2008. Colonization history, ecological shifts and diversification in the evolution of endemic Galápagos weevils. Mol Ecol. 17:1089–1107.
Sequeira AS, Lanteri AA, Scataglini MA, Confalonieri VA, Farrell BD. 2000. Are flightless Galapaganus weevils older than the Galápagos Islands they inhabit? Heredity (Edinb). 85:20–29.
Sequeira AS, Sijapati M, Lanteri AA, Roque Albelo L. 2008. Nuclear and mitochondrial sequences confirm complex colonization patterns and clear species boundaries for flightless weevils in the Galápagos archipelago. Philos Trans R Soc Lond B Biol Sci. 363:3439–3451.
Sequeira AS, Stepien CC, Sijapati M, Roque Albelo L. 2012. Comparative genetic structure and demographic history in endemic galapagos weevils. J Hered. 103:206–220.
Simberloff DS. 1986. Introduced insects: a biogeographic and systematic perspective. In: Mooney HA, Drake JA, editors. Ecology of biological inva-sions of North America and Hawaii. Berlin, Germany: Springer. p. 3–26.
Simpson S, Blampied N, Peniche G, Dozières A, Blackett T, Coleman S, Cornish N, Groombridge JJ. 2013. Genetic structure of introduced popu-lations: 120-year-old DNA footprint of historic introduction in an insular small mammal population. Ecol Evol. 3:614–628.
Snell HL, Powell G, Tye A, Bensted-Smith R, Bustamante RH, Branch GM. 2002a. Approach to predicting the future of Galápagos Biodiversity. In: Bensted-Smith R, editor. A biodiversity vision for the Galápagos Islands. Galápagos, Ecuador: Charles Darwin Foundation and World Wildlife Fund. p. 6–11.
Snell HL, Tye A, Causton CE, Bensted-Smith R. 2002b. The status of and threats to terrestrial biodiversity. In: Bensted-Smith R, editor. A biodiver-sity vision for the Galápagos Islands. Galápagos, Ecuador: Charles Darwin Foundation and World Wildlife Fund. p. 30–47
Stepien CC, Kaczmarek M, Mok H, Stuckert A, Chen P, Downer SJ, Sequeira AS. 2010. Isolation and characterization of microsatellite loci for the introduced broad-nosed weevil Galapaganus howdenae howdenae in the Galapágos archipelago. Molecular Ecology Resources. 10:576–579.
Storz JF, Beaumont MA. 2002. Testing for genetic evidence of population expansion and contraction: an empirical analysis of microsatellite DNA vari-ation using a hierarchical Bayesian model. Evolution. 56:154–166.
Tallmon DA, Koyuk A, Luikart G, Beaumont MA. 2008. ONeSAMP: a program to estimate effective population size using approximate Bayesian computation. Molecular Ecology Resources. 8:299–301.
Tonione MA, Reeder N, Moritz CC. 2011. High genetic diversity despite the potential for stepping-stone colonizations in an invasive species of gecko on Moorea, French Polynesia. PLoS One. 6:e26874.
Tye A, Soria MC, Gardener MR. 2002. A strategy for Galápagos weeds. In: Clout CRVaMN, editor. Turning the tide: the eradication of invasive spe-cies. Proceedings of the International Conference on Eradication of Island Invasives (IUCN Species Survival Commission Occasional Paper). Gland, Switzerland: IUCN. p. 336–341.
White WM, McBirney AR, Duncan RA. 1993. Petrology and geochemis-try of the Galápagos Islands: portrait of a pathological mantle plume. J Geophys Res. 98:19533–19563.
Williams DA, Overholt WA, Cuda JP, Hughes CR. 2005. Chloroplast and microsatellite DNA diversities reveal the introduction history of Brazilian peppertree (Schinus terebinthifolius) in Florida. Mol Ecol. 14:3643–3656.
Wilson EE, Holway DA. 2010. Multiple mechanisms underlie displacement of solitary Hawaiian Hymenoptera by an invasive social wasp. Ecology. 91:3294–3302.
Zane L, Bargelloni L, Patarnello T. 2002. Strategies for microsatellite isola-tion: a review. Mol Ecol. 11:1–16.
Zeisset I, Beebee TJ. 2003. Population genetics of a successful invader: the marsh frog Rana ridibunda in Britain. Mol Ecol. 12:639–646.
Zenger KR, Richardson BJ, Vachot-Griffin AM. 2003. A rapid population expansion retains genetic diversity within European rabbits in Australia. Mol Ecol. 12:789–794.
Received July 17, 2013; First decision October 16, 2013; Accepted December 3, 2013
Corresponding Editor: Christopher Smith
by guest on January 8, 2014http://jhered.oxfordjournals.org/
