Trait means and reaction norms: the consequences of climate change/invasion interactions at the organism level

RESEARCH ARTICLE

Trait means and reaction norms: the consequencesof climate change/invasion interactionsat the organism level

Charlene Janion • Hans Petter Leinaas • John S. Terblanche •

Steven L. Chown

Received: 20 January 2010 / Accepted: 28 June 2010 / Published online: 15 July 2010� Springer Science+Business Media B.V. 2010

Abstract How the impacts of climate change on biological invasions will play out at the

mechanistic level is not well understood. Two major hypotheses have been proposed:

invasive species have a suite of traits that enhance their performance relative to indigenous

ones over a reasonably wide set of circumstances; invasive species have greater phenotypic

plasticity than their indigenous counterparts and will be better able to retain performance

under altered conditions. Thus, two possibly independent, but complementary mechanistic

perspectives can be adopted: based on trait means and on reaction norms. Here, to dem-

onstrate how this approach might be applied to understand interactions between climate

change and invasion, we investigate variation in the egg development times and their

sensitivity to temperature amongst indigenous and introduced springtail species in a cool

temperate ecosystem (Marion Island, 46�540S 37�540E) that is undergoing significant cli-

mate change. Generalized linear model analyses of the linear part of the development rate

curves revealed significantly higher mean trait values in the invasive species compared to

indigenous species, but no significant interactions were found when comparing the thermal

reaction norms. In addition, the invasive species had a higher hatching success than the

indigenous species at high temperatures. This work demonstrates the value of explicitly

examining variation in trait means and reaction norms among indigenous and invasive

species to understand the mechanistic basis of variable responses to climate change among

these groups.

Electronic supplementary material The online version of this article (doi:10.1007/s10682-010-9405-2)contains supplementary material, which is available to authorized users.

C. Janion (&) � S. L. ChownCentre for Invasion Biology, Department of Botany and Zoology, Stellenbosch University,Private Bag X1, Matieland 7602, South Africae-mail: [email protected]

H. P. LeinaasDepartment of Biology, University of Oslo, Blindern, P.O. Box 1066, 0316 Oslo, Norway

J. S. TerblancheDepartment of Conservation Ecology and Entomology, Stellenbosch University, Private Bag X1,Matieland 7602, South Africa

123

Evol Ecol (2010) 24:1365–1380DOI 10.1007/s10682-010-9405-2

Keywords Biological invasions � Climate change � Development rate �Phenotypic plasticity � Soil arthropod

Introduction

The major drivers of biodiversity change are well known. Habitat alteration, over-

exploitation, biological invasions and co-extinctions were identified as an ‘evil quartet’

nearly two decades ago (Diamond 1989). Together with climate change, their present and

likely future effects on biodiversity and human well-being remain among the most sig-

nificant problems presently facing humanity (Sala et al. 2000; Chown and Gaston 2008).

Although much is known about the mechanistic basis of each of these drivers, how they are

likely to interact has not been comprehensively investigated across the full range of

interactions (Brook et al. 2008). For example, understanding of the nature and outcome of

interactions among habitat alteration and invasion is growing rapidly, and it is clear that the

‘driver-passenger’ relationship is often context dependent (Daehler 2003; Didham et al.

2007). Similarly, appreciation of how habitat alteration, patchiness, climate change and

short-term evolutionary responses act in isolation and in concert to affect species distri-

butions across the landscape is also growing rapidly (Warren et al. 2001; Carroll and Fox

2007; de Mazancourt et al. 2008). By contrast, interactions between climate change and

invasion are poorly understood.

It is widely thought that climate change will exacerbate the extent and impact of

biological invasions, affecting each of the stages in the invasion process (Cannon 1998;

Hobbs and Mooney 2005; Theoharides and Dukes 2007), and, where non-indigenous

species have established, often favouring them over indigenous species (Dukes and

Mooney 1999). Whilst evidence is accumulating to support some of these conjectures (e.g.,

Holzapfel and Vinebrooke 2005; Chown et al. 2007), the number of empirical studies of

interactions between climate change and invasion remains relatively limited (Walther et al.

2009). Indeed, several recent reviews have concluded that the information available is

insufficient for empirical generalization (Ward and Masters 2007; Chown and Gaston

2008; Brook 2008). This is especially true for how such interactions might play out at the

mechanistic level (Agrawal 2001; Stachowicz et al. 2002). Nonetheless, two major

hypotheses have been proposed.

First, invasive species have a suite of traits that enhance their performance relative to

indigenous ones over a reasonably wide set of circumstances. In consequence, climate

change will continue to mean that these species perform better than their indigenous

counterparts. The foundation of this idea stretches back at least to Elton (1958) and has

formed the basis of a large literature attempting to identify the species-level characteristics

that might promote invasion. For example, in plants, invasive species are often charac-

terized by large stature, high seed mass, rapid and profuse seedling emergence, and self-

fertilization (Richardson and Pysek 2006; van Kleunen and Johnson 2007; van Kleunen

et al. 2008, but see also Davis and Shaw 2001; Daehler 2003). In consequence, these

species are able to take advantage of a reasonably broad range of conditions, especially

those that are considered favourable for plant performance generally, such as warm, moist,

nutrient rich circumstances (Stohlgren et al. 2008). This idea has also been extended to

many other groups. For example, invasive invertebrate species are apparently often small-

bodied and have high-growth rates, so promoting their success in novel environments

(Lawton and Brown 1986; Gaston et al. 2001). In birds, behavioural flexibility, broad

1366 Evol Ecol (2010) 24:1365–1380

123

environmental tolerances and wide resource use are all thought to favour invasion (Duncan

et al. 2003). Thus, if invasive species have a suite of characteristics that enable them to

succeed under a reasonably wide range of conditions they should perform better under

warmer conditions too. In ectotherms, the generally positive relationship between growth

rate and temperature (Chown and Nicolson 2004; Ward and Masters 2007) means that this

will be especially likely as long as the temperature increase does not exceed organismal

thermal safety margins (see also Deutsch et al. 2008; Frazier et al. 2008), and mismatches

in phenology do not limit resource availability (Parmesan 2007). However, this hypothesis

provides no grounds for expecting that climate change will increasingly favour invasive

species.

Second, invasive species have greater phenotypic plasticity than their indigenous

counterparts. In consequence, they will be better able to retain performance under the

altered circumstances precipitated by climate change. This hypothesis is a component of

the idea first proposed by Baker (1965) of a ‘‘general purpose genotype’’, and has garnered

some support both for plants (Daehler 2003; Richardson and Pysek 2006) and animals

(Trussell and Smith 2000; Rosecchi et al. 2001; Duncan et al. 2003; Dzialowski et al. 2003,

but see also Lee et al. 2007). In the context of climate change, Stachowicz et al. (2002)

demonstrated that invasive ascidians had steeper growth rate–temperature relationships

than an indigenous species, and argued that this has favoured the introduced species under

warming, an idea supported by several years of demographic data. By contrast, Chown

et al. (2007) found no difference in the extent of phenotypic plasticity for desiccation

tolerance in indigenous and invasive springtails. However, they showed that the form of

plasticity differed between these two groups. Invasive species are favoured following

exposure to warm conditions, whilst indigenous species perform better following a low

temperature treatment.

Clearly, plasticity may evolve independently of or in concert with the trait mean, and

indeed how this interaction should be viewed and modelled has been the subject of much

discussion (see Scheiner 1993; de Jong 1995, 2005; Via et al. 1995; Ghalambor et al.

2007). Nonetheless, as a matter of convenience the mean and plasticity can be considered

relatively independent (see Ghalambor et al. 2007: 396). Thus, the mechanistic, organis-

mal-level exploration of interactions between climate change and invasion can be

approached from complementary perspectives: the way that differences in mean trait

values among indigenous and invasive species are likely to affect performance or survival

under a new set of conditions, and the way that variation in reaction norms among these

two groups of species may affect fitness under altered conditions (Lee 2002; Lee et al.

2003; Dybdahl and Kane 2005). This does not mean that mean trait values and plasticity

are necessarily independent (see de Jong 1995, 2005; van Kleunen and Fischer 2005;

Ghalambor et al. 2007), but does provide a framework within which the historical per-

spectives of differences in trait values and in plasticity among indigenous and invasive

species can be interpreted.

Here, we adopt this approach to investigate the likely impacts of interactions between

climate change and invasion, mediated through egg development rate and its response to

temperature, in a springtail assemblage from a cool, temperate ecosystem that is under-

going significant climate change (Chown et al. 2007; see Dybdahl and Kane 2005 for a

similar approach). Springtails were chosen as the study organisms because they are a

globally significant group of soil invertebrates that play major roles in ecosystem func-

tioning (Rusek 1998), including in the system we investigate (Gabriel et al. 2001; Hugo

et al. 2004). Egg development rate (or time, given that rate = 1/development time) was

selected because it is an important component of fitness in springtails (as is development

Evol Ecol (2010) 24:1365–1380 1367

123

rate more generally—Sibly and Calow 1986; Roff 2002), and the egg development period

encompasses a substantial portion of the life cycle of an individual springtail (van Straalen

1994; Birkemoe and Leinaas 2000).

Specifically, we first test the hypothesis that over the linear portion of the rate–

temperature curve (this portion of the rate–temperature curves is most often the subject of

investigation—see e.g., Trudgill et al. 2005; de Jong and van der Have 2008), mean egg

development should be faster in the invasive than in the indigenous species. Next we

evaluate the prediction of the reaction norm hypothesis that the invasive species should

have steeper rate–temperature relationships than the indigenous species (which has its basis

not only in the findings of Stachowicz et al. (2002) for ascidians, but also in terms of the

same hypothesis proposed for insects from the same region (Chown et al. 2002)). The

outcomes of these two tests are subsequently explored in the context of what is known

about the population responses of these species to current climates and climate change, to

further knowledge of the likely effects of rapid climate change on this system (see le Roux

and McGeoch 2008).

Methods

Study site and species

All species used in this study were collected from sub-Antarctic Marion Island (46�540S37�540E), which has a cool, wet, windy climate that has shown a substantial change over

the last 50 years, including more than a 1.3�C increase in mean annual temperature

(Chown and Froneman 2008). Sixteen springtail species occur on the island, of which six

are invasive, four endemic and six have a broader, sub-Antarctic distribution (Gabriel et al.

2001; Frenot et al. 2005). The invasive springtails are mostly of European origin (Chown

et al. 2002) and are thought to have been introduced by humans with animal fodder

provided for sheep that were kept on the island in the 1950s (Cooper and Condy 1988).

Thus they have been on the island for at least 40 years (Deharveng 1981). In Europe,

original records of the species are mostly from Scandinavia and northern areas of Germany

(Fjellberg 1998; Potapov 2001). Low mitochondrial cytochrome oxidase I (COI) haplotype

numbers (typically a single haplotype) in the invasive springtail species are indicative of

colonization by a few individuals, whilst the indigenous species are characterized by

considerable haplotype diversity (Myburgh et al. 2007).

For this study, the seven most common Poduromorpha and Entomobryomorpha species

on the island were investigated. Cryptopygus dubius Deharveng (Isotomidae), C. tricuspisEnderlein and C. antarcticus travei Deharveng are indigenous, whilst Ceratophyselladenticulata Bagnall (Hypogastruridae), Parisotoma notabilis Schaeffer (Isotomidae),Isotomurus cf. palustris Muller (Isotomidae) and Pogonognathellus flavescens Tullberg

(Tomoceridae) are invasive (Gabriel et al. 2001). The species also differ in depth distri-

bution within the soil profile: Isotomurus cf. palustris, C. antarcticus travei and Pogon-ognathellus flavescens are epedaphic or soil-surface dwellers, Cryptopygus dubius and

Ceratophysella denticulata hemi-edaphic or litter dwellers, and C. tricuspis and Pariso-toma notabilis euedaphic or soil dwellers (see van Straalen 1994 and Hopkin 1997 for

discussion of depth distributions in springtails).

Several hundred adult specimens of each species were collected in the field (below

100 m above sea level) using an aspirator and placed into 30 ml vials with moist Plaster-

of-Paris substrates and small amounts of detritus for shelter and food, and transported back

1368 Evol Ecol (2010) 24:1365–1380

123

to the laboratory within 5 h of collection. In the laboratory, specimens were kept in 40 ml

containers, with moist Plaster-of-Paris substrates and small amounts of detritus and algae

as food, in incubators (LABCON, Johannesburg, South Africa, accurate to ±1�C) at 5�C

with a 12L: 12D photoperiod. The animals were then transferred to Oslo, Norway, where

most work was undertaken except for a few additional trials with Cryptopygus antarcticustravei and Isotomurus cf. palustris, which were done in Stellenbosch South Africa. In both

places animals were maintained under identical conditions in the same make of incubators

(Sanyo MIR 153, Osaka, Japan, accurate to ±0.5�C). All individuals used for experiments

were acclimated at 15�C (or 10�C in the case of Cryptopygus antarcticus travei) for at least

14 days before the start of the experiment, as egg production was low at 5�C. Distilled

water was added once a week to maintain high humidity and springtails were provided with

excess algae as a food source (replaced once to twice a week to prevent growth of mould).

Embryonic development rates

Previous studies have demonstrated low survival when springtail individuals are kept

separately (e.g., Birkemoe and Leinaas 2000). Thus, 10–20 adults from the stock cultures

were randomly placed in culture boxes (diameter = 3.4 cm, depth = 3 cm) lined with

Plaster-of-Paris mixed with charcoal powder (ratio of 9:1). For optimum egg production,

and to reduce maternal effects, culture boxes were kept at 15�C for one generation, or, for

Cryptopygus antarcticus travei only, at 10�C because mortality in this species was high at

15�C. Egg batches (Cfive eggs) were then collected daily from culture boxes of each

species (except P. flavescens which laid single eggs) and assigned to a particular tem-

perature treatment (see below). For each species and each treatment, eggs were distributed

at random among at least four culture vials to avoid container effects. The vials were then

placed into the incubators (Sanyo MIR) at a set temperature and 12L:12D light regime.

Eggs were inspected daily until hatching, and at this time containers were randomized

among shelves to avoid shelf effects. Egg development times (determined as the number of

days between egg laying and hatching (Birkemoe and Leinaas 2000)) were obtained at

c. 5�C intervals from 5 to 28.8�C to provide complete performance curves (see Angilletta

2006; Deere and Chown 2006). Due to technical difficulties (incubator failure) no egg

development time measurements were obtained for P. flavescens above 25�C. Cryptopygusantarcticus travei eggs failed to hatch at 20�C and above.

Hatching success is often traded off against egg development rate (Stearns 1992;

Stillwell and Fox 2005; Geister et al. 2009). This means that for a given species, a high egg

development rate may not necessarily also mean high survival under those conditions,

resulting in low fitness at high temperatures. Therefore, we also examined hatching success

at 15�C and at 25�C for all species except C. antarcticus travei which had zero hatching

success at 20 and 25�C. Specifically, we determined whether hatching success declined in

the indigenous vs. invasive species (irrespective of soil depth category) at the higher

temperatures. Hatching success was recorded as the number of eggs that hatched from the

total number of eggs in the batch. This value was converted to a percentage value (number

of hatched eggs/total number eggs in batch 9 100).

Analyses

First, whilst recognizing that rate–temperature relationships typically have a non-linear

form (Stearns 1992; Angilletta 2006), and that a variety of analytical approaches are

available to deal with this non-linearity (e.g. David et al. 1997; Izem and Kingsolver 2005)

Evol Ecol (2010) 24:1365–1380 1369

123

only the linear portion of the reaction norm (5–20�C in this case, hereafter rate–temper-

ature or R–T curve) was analyzed. Such analyses are typical of investigations of rate–

temperature relationships (see Trudgill et al. 2005; Ikemoto 2005; de Jong and van der

Have 2008). Moreover, the aim here was to consider the sensitivity to development over

the linear part of the temperature range as the reaction norm, and the mean across these

temperatures as the mean trait value.

For the analysis of mean trait values and their reaction norms over the linear portion of

the R–T curve, general linear models, implemented in Statistica v.8 (Statsoft, Tulsa, OK,

USA), were used. In all cases, mean egg development rate per egg batch was used as the

independent data because individual eggs within batches are unlikely to represent inde-

pendent data (with the exception of Pogonognathellus flavescens which laid eggs singly).

The first hypothesis tested was that mean log10 egg development rate, holding temperature

constant, differs among the invasive and indigenous species and among the soil depth

categories. Thus, the model included temperature as a covariate to estimate the least-

squares mean trait value. However, because the rate–temperature relationship might differ

among species, a general linear model specifically testing for whether interactions are

present between temperature and status (indigenous vs. invasive), and between temperature

and depth category was initially implemented (see Quinn and Keough 2002: 349, the test

was implemented in Statistica v. 8). No slope heterogeneity was found, and in effect this

formed the test of whether the linear reaction norms differed among the species.

Because each status by depth group was typically only represented by a single species

(except for the ‘epedaphic invasive’ group which included Isotomurus cf. palustris and

Pogonognathellus flavescens), the power of the above tests to assess the extent to which

any differences among the status and depth groups might be considered general, could be

thought of as limited. Therefore, further to examine the extent to which these differences

might be biologically significant we also undertook a randomization analysis where

development rate data were randomized among the same number of status and depth

groups. Here we used a resample-without-replacement protocol for both status and depth

categorical variables (see Parr et al. 2005). Then, we calculated mean log10 egg devel-

opment rate for 100 random re-samples (run in Microsoft Excel) and computed their 95%

confidence limits. This allowed us to undertake a comparison of randomized depth/inva-

sive status samples relative to observed egg development rate for each category. We

reasoned that if particular status and depth group means were different to a random

selection of the data, additional confidence could be placed in their biological significance.

Because phylogenetic relatedness may have played a major role in influencing the

patterns in trait variation found among these species (given that the data set for the

indigenous species included only the genus Cryptopygus) we undertook a second set of

analyses using phylogenetic generalized least squares (PGLS). The PGLS and generalized

estimating equation (GEE) analyses were implemented in the APE module (Paradis et al.

2004) of the R language open-source software package (following methods outlined in

Halsey et al. 2006). The species level phylogeny was constructed using mitochondrial COI

sequence data (Supplementary material S3). The tree concurred with previously published

analyses (Stevens et al. 2006), with the exception of Cryptopygus tricuspis clustering with

Ceratophysella denticulata and Pogonognathellus flavescens, as opposed to with other

Cryptopygus species (C. antarcticus travei and C. dubius). No reason exists to suppose that

Cryptopygus is paraphyletic (based on limited 28S data and on Stevens et al. 2006). Rather

the outcome was likely due to hemiplasy (Avise and Robinson 2008). Therefore, to

determine branch lengths, the evolutionary tree was constrained to group C. tricuspiswithin the other Cryptopygus species. Branch lengths were determined in PAUP*

1370 Evol Ecol (2010) 24:1365–1380

123

(Swofford 2001) using maximum likelihood analyses and plotted on the tree. Thereafter,

PGLS, using these estimated branch lengths in units of mutational change, was used to

assess the strength of the effect of the hypothetical phylogeny on the dependent variable by

estimating lambda (k). k is an estimate of the phylogenetic covariance in the data and lies

between 0 and 1; k = 1 indicates a strong influence of phylogeny while k = 0 indicates

little influence of phylogeny (see Halsey et al. 2006). In the GEE models significance

levels were calculated from phylogenetically adjusted degrees of freedom. The PGLS and

GEE analyses were used to investigate variation among the least squares means obtained

from the General Linear Model.

For analyses of hatching success variation among the indigenous and invasive species at

the higher temperatures (15 and 25�C), a generalised linear model analysis, assuming a

binomial distribution and a logit link function, was undertaken in SAS (V 9.1) to assess

effects of temperature variation, status (indigenous vs. invasive species) and interactions

between temperature and species.

Owing to the substantial emphasis that has been placed recently on identifying how the

temperature optimum (Topt), rate at the temperature optimum (umax) from performance

curves might differ among species, environments and treatments (e.g., Gilchrist 1996;

Kingsolver and Huey 1998; Angilletta et al. 2002, 2003; Deere and Chown 2006; Deutsch

et al. 2008) these traits were estimated for each species. This was done using a modification

of the method proposed by Angilletta (2006). For each species, the mean rate per batch at

each temperature was calculated and a curve was fitted to these data using TableCurve 2D

(SYSTAT Inc, 2002, San Jose, California, USA). The curve with the largest coefficient of

determination was selected, irrespective of the number of terms, because the aim was to

find the best-fitting curve so that Topt and umax could be determined from the fitted curve (in

the latter case using Gilchrist’s (1996) approach with 0.1�C temperature increments).

Typically, the best fits were provided by the Exponentially Modified Gaussian, and Half

Gaussian Modified Gaussian curves (Supplementary material S1). In the case of Crypto-pygus antarcticus travei, individual level data were used because mean data were available

only at four temperatures so precluding use of the peaked curve equations of TableCurve

2D. In this case a LogNormal4 curve provided the best fit (Supplementary material S1).

Values for Topt and umax were compared among the indigenous and invasive groups using

a single classification ANOVA.

Results

Temperature had the expected effect on development rate (Fig. 1; Table 1), with maximum

development temperatures found at 27.5�C in the invasive species Isotomurus cf. palustris,

Ceratophysella denticulata and Parisotoma notabilis, and typically at lower temperatures

for the indigenous ones (see Appendix S2 for mean data values). In the generalized linear

model analyses, the homogeneity of slopes model revealed that depth and status had a

significant effect on the mean egg development rate, but no slope heterogeneity was

identified among the groups (Table 2). Means and confidence intervals for each of the

groups indicated that invasive species typically had faster mean egg development rates

than the indigenous species except in the hemi-edaphic depth group (Fig. 2). Because the

slopes of the rate–temperature relationships did not differ significantly among the groups

(including no 3-way interaction), it is clear that the reaction norms, or form and extent of

plasticity, do not differ among them.

Evol Ecol (2010) 24:1365–1380 1371

123

Calculation of lambda suggested limited influence of phylogeny on the least-squares

(LS)-adjusted mean development rate (k = 0) (Appendix S3, S4). Thus, phylogenetic-

adjusted results were excluded from further interpretation.

Mean egg development rates for the invasive species were significantly and substan-

tially faster than those obtained by randomization of the data (Fig. 3), suggesting that these

species have faster rates than expected on average. By contrast, no differences were found

for the indigenous groups except for the epedaphic group, which showed much slower

development rates than those obtained from a random selection of the data.

Fig. 1 Reaction norms for egg development rates for seven species of Collembola from Marion Island(mean ± 95% CI at each temperature). The species marked with an asterisk are indigenous to the sub-Antarctic

Table 1 Summary statistics of the log10 of egg development rate–temperature relationships, using batchmeans of egg development rate

Species Slope ± SE Intercept ± SE F df P R2 SE

Isotomuruscf. palustris

0.0588 ± 0.0012 -1.9704 ± 0.0159 2447.57 1, 34 0.0001 0.9863 0.0431

Pogonognathellusflavescens

0.0542 ± 0.0023 -2.0828 ± 0.0337 568.62 1, 9 0.0001 0.9844 0.0352

Parisotomanotabilis

0.0530 ± 0.0016 -1.9651 ± 0.0219 995.63 1, 24 0.0001 0.9765 0.0446

Ceratophyselladenticulata

0.0510 ± 0.0020 -1.9408 ± 0.0292 635.31 1, 20 0.0001 0.9695 0.0504

Cryptopygusdubius*

0.0458 ± 0.0025 -1.8901 ± 0.0369 330.81 1, 43 0.0001 0.885 0.0795

C. tricuspis* 0.0503 ± 0.0199 -1.9821 ± 0.0199 1352.51 1, 37 0.0001 0.9734 0.0465

C. antarcticustravei*

0.0561 ± 0.0087 -2.3544 ± 0.1096 41.90 1, 3 0.0075 0.9332 0.0775

The species marked with an asterisk are indigenous to the sub-Antarctic

1372 Evol Ecol (2010) 24:1365–1380

123

The curve fitting approach indicated that whilst the indigenous species tended to

have lower Topt values than the invasive species, the difference was not significant

(F(1,5) = 3.63, P = 0.11). Likewise, umax did not differ significantly among these groups

(F(1,5) = 4.45, P = 0.09) (Table 3). However, hatching success was significantly lower at

25�C in the two indigenous species by comparison with the invasive species (Table 4;

Fig. 4a, Appendix S5). These differences remained significant irrespective of whether the

species were considered separately or grouped by invasive status (Table 4; Fig. 4b).

Discussion

Invasive alien species are thought to differ from their indigenous counterparts by virtue of

differences in trait means and/or the extent of their phenotypic plasticity. The former idea

has proven controversial with some studies demonstrating significant and sometimes

Table 2 Outcome of the generallinear model investigating theinteraction between status, depthdistribution and temperature

Effect Sums of squares df F P

Status 0.0379 1 8.13 0.004898

Depth 0.08881 2 9.52 0.000120

Temperature 5.40866 1 1159.54 0.000001

Status 9 depth 0.08056 2 8.64 0.000267

Status 9 temperature 0.0035 1 0.75 0.387570

Depth 9 temperature 0.01144 2 1.23 0.295990

Status 9 depth 9temperature

0.00952 2 1.02 0.362622

Error 0.80229 172

Indigenous Invasive

Fig. 2 The effect of depth distribution, and whether the species are invasive or indigenous, on mean (±95%CI) egg development rate, when temperature is taken into account using a General Linear Model (seeTable 2)

Evol Ecol (2010) 24:1365–1380 1373

123

substantial differences among these two groups of species in trait means and other

investigations finding no differences at all (see Daehler 2003; Duncan et al. 2003;

Richardson and Pysek 2006; Ward and Masters 2007; Walther et al. 2009 for recent

-1.7

-1.6

-1.5

-1.4

-1.3

-1.2

Eue Hemi Epe Eue Hemi Epe

Ind Ind Ind Inv Inv Inv

Group

Mea

n e

gg

dev

elo

pm

ent

rate

(lo

g10

)

RandomObserved

Fig. 3 Summary statistics (means ± 95% CLs) for estimated (observed) egg development rates for eachdepth by status category relative to data randomised for depth and status across all species and categories.Status (ind indigenous or inv invasive) and soil depth category (euedaphic, hemi-edaphic or epedaphic) wererandomised without replacement 100 times

Table 3 Values for the optimumtemperature (Topt) and eggdevelopment rate at Topt (umax)for each species

Species marked with anasterisk are indigenousto the sub-Antarctic

Species Topt Umax

Parisotoma notabilis 27.0 0.151

Ceratophysella denticulata 27.8 0.169

Pogonognathellus flavescens 24.6 0.134

Isotomurus cf. palustris 27.3 0.192

Cryptopygus dubius* 24.8 0.119

C. tricuspis* 23.9 0.139

C. antarcticus travei* 13.8 0.037

Table 4 The results from ageneralised linear model assum-ing a binomial distribution and alogit link function, showing theeffects of temperature (Temp) onthe hatching success between(a) species and (b) status(indigenous or invasive)

Effect df v2 P

(a)

Species 5 681.55 \0.0001

Temp 1 58.97 \0.0001

Species 9 temp 4 17.54 0.0015

(b)

Status 1 0.01 0.9143

Temp 2 28.45 \0.0001

Status 9 temp 1 19.44 \0.0001

1374 Evol Ecol (2010) 24:1365–1380

123

reviews). In this case, substantial differences in trait means across the linear part of the

rate–temperature curve (while taking temperature into account) were found among the

indigenous and invasive species, with the invasive species having, on average, faster egg

development rates between 5 and 20�C than the indigenous species. For umax, the indig-

enous species mean (0.0985 or 10 days for development at Topt) was only 61% of that

found for the invasive species (0.162 or 6 days for development at Topt), but because of the

relatively large variance around the means, and the small numbers of species involved, the

difference was not significant. Likewise, Topt did not differ among the two groups. In

consequence, the trait means differ among the indigenous and invasive species over the

40

50

60

70

80

90

100

110

120

130

Indigenous Invasive

15 25

a

b

Fig. 4 The effect of temperature on hatching success at 15 and 25�C with a species and b status(indigenous or invasive). The species marked with an asterisk are indigenous to the sub-Antarctic. Errorbars show 95% CIs

Evol Ecol (2010) 24:1365–1380 1375

123

5–20�C part of the reaction norm, and possibly at its optimum. Importantly, however,

hatching success was lower in the indigenous species at 25�C (zero in C. antarcticus above

20�C) than in the invasive species, although such differences were not apparent at 15�C.

Thus, the invasive species appear to be at an advantage in terms of egg development rate

and hatching success at high temperature compared with the indigenous species in most

groups, except perhaps for the hemi-edaphic species where the difference was not

significant.

Given the fact that each status x depth category was typically only represented by a

single species (a consequence of the nature of the springtail assemblages at the site we

investigated—see Gabriel et al. 2001), it might be argued that the differences among the

groups are simply an artefact of the species investigated. The randomization procedure

suggested that, perhaps with the exception of the slow developing C. antarcticus, the

indigenous species are no different to what might be expected for a random assortment of

egg development rates. However, the invasive species always had faster rates than might

be expected from a random re-allocation of the data. Thus, further support is provided for

the idea that the invasive species typically have faster development rates than the indig-

enous ones. The limited influence of phylogeny (k = 0) also showed that the result is not

confounded by a strong phylogenetic signal in the data. In consequence, at least for this

assemblage of springtails, it does appear that the invasive species typically have much

faster development rates than the indigenous ones, and a lower susceptibility to hatching

failure at higher temperatures.

This outcome of a significant difference in trait means is in keeping with what has been

found for adult lower and upper lethal temperatures for a similar group of species from

Marion Island (including many of those studied here), where the indigenous species tended

to have lower freezing points (=lower lethal temperature in this case) and lower, high

temperature tolerances than the invasive species (Slabber et al. 2007). By contrast, duration

of survival of dry conditions (75% RH) did not differ among the indigenous and invasive

species for the adults, although acclimation to high temperature certainly increased sur-

vival time in the invasive species, whilst often having the converse effect on the indigenous

ones (Chown et al. 2007). In consequence, it appears that conclusions about whole

organismal responses cannot simply be generalized. Rather, the responses likely vary from

trait to trait, and in concert will determine the performance outcome for the organism

concerned (see also van Kleunen and Johnson 2007). Nonetheless, overall it appears that

warmer and drier conditions will certainly benefit invasive springtail species on Marion

Island relative to their indigenous counterparts.

Across the 5–20�C part of the reaction norm, the slopes of the R–T relationships did not

differ among the species. This outcome is in keeping with that found for other traits in the

same species, notably survival time under desiccating conditions and lower and upper

critical thermal limits (Chown et al. 2007; Slabber et al. 2007), and with other investigations

where no differences in plasticity have been found (see Daehler 2003; Lee et al. 2007).

However, it differs from a range of studies where significant and often substantial differ-

ences in phenotypic plasticity among indigenous and invasive species within the same taxa

have been found (e.g., Trussell and Smith 2000; Rosecchi et al. 2001; Stachowicz et al.

2002; Duncan et al. 2003; Dzialowski et al. 2003; Richardson and Pysek 2006).

From a climate change perspective the present data, in the context of what is known

about means and reaction norms of other traits in these species (Chown et al. 2007; Slabber

et al. 2007), suggest that the invasive species should be predominant in low elevation warm

areas of Marion Island, and should continue to be favoured by ongoing warming and

drying. This is indeed what has been found to date. Ecological surveys of the island have

1376 Evol Ecol (2010) 24:1365–1380

123

found exceptionally high densities and dominance of the invasive species at low elevations

across the island, but an absence of these species at higher, cold, dry elevations (Gabriel

et al. 2001). Moreover, field experiments that have simulated predicted warming and

drying on the island (see le Roux and McGeoch 2008) resulted in substantial declines in the

abundances of the indigenous species and little change in the abundance of an invasive

species (Chown et al. 2007). The outcomes of this study also support previous predictions,

made on the grounds of qualitative comparisons of insect life history data, that invasive

species will be favoured by climate change at the island given their short development

times (Chown et al. 2002). However, they contradict the idea that steeper rate–temperature

relationships in invasive than in indigenous species will further benefit the former under

warmer conditions (Chown et al. 2002). Nonetheless, further evidence that climate change

is benefiting the invasive springtails would be provided by demonstrations of an altitudinal

range expansion. This work is presently underway.

In conclusion, the present study has demonstrated that mean egg development rate and

hatching success at high temperature, but not the thermal reaction norms of development

rate, differ consistently between indigenous and invasive springtails on Marion Island, and

that these differences are likely to have substantial implications for assemblage structure as

the climate continues to change (see also Janion et al. 2009). In so doing it demonstrates

how, by considering mean and plasticity of traits separately (see Ghalambor et al. 2007),

insight into the mechanistic basis of interactions among climate change and invasion may

be obtained. It also provides further evidence that the interactions among climate change

and invasion are likely to be synergistic (see also Stachowicz et al. 2002) and that such

interactions need urgent further investigation (Brook 2008). Whether our results for

springtails can be generalized more broadly to this group remains to be determined through

further study of assemblages of indigenous and invasive species elsewhere. Nonetheless, it

supports a growing body of work showing that the traits of indigenous and invasive species

differ considerably, probably owing to the characteristics that ensure survival and estab-

lishment in the first place (Blackburn et al. 2009).

Acknowledgments We thank Erika Nortje, Heidi Sjursen Konestabo and various members of the MarionIsland relief teams for assistance in the field. Bettine Jansen van Vuuren and Angela McGaughran assistedwith the COI sequence data and phylogenetic analysis. Janne Bengtsson, Melodie McGeoch and twoanonymous referees provided useful comments on a previous version of the ms. The South African NationalAntarctic Programme provided logistic support. This work was funded partly by a SA-Norway scienceliaison grant awarded jointly to HPL and SLC. The work forms a contribution to the SCAR EBAProgramme.

References

Agrawal AA (2001) Phenotypic plasticity in the interactions and evolution of species. Science 294:321–326.doi:10.1126/science.1060701

Angilletta MJ (2006) Estimating and comparing thermal performance curves. J Thermal Biol 31:541–545.doi:10.1016/j.jtherbio.2006.06.002

Angilletta MJ, Niewiarowski PH, Navas CA (2002) The evolution of thermal physiology in ectotherms.J Thermal Biol 27:249–268. doi:10.1016/S0306-4565(01)00094-8

Angilletta MJ, Wilson RS, Navas CA et al (2003) Tradeoffs and the evolution of thermal reaction norms.Trends Ecol Evol 18:234–240. doi:10.1016/S0169-5347(03)00087-9

Avise JC, Robinson TJ (2008) Hemiplasy: a new term in the lexicon of phylogenetics. Syst Biol57:503–507. doi:10.1080/10635150802164587

Evol Ecol (2010) 24:1365–1380 1377

123

Baker HG (1965) Characteristics and modes of origin of weeds. In: Baker HG, Stebbins GL (eds) Thegenetics of colonizing species. Academic Press, New York, pp 147–168

Birkemoe T, Leinaas HP (2000) Effects of temperature on the development of an Arctic Collembola(Hypogastrura tullbergi). Funct Ecol 14:693–700. doi:10.1046/j.1365-2435.2000.00478.x

Blackburn TM, Lockwood JL, Cassey P (2009) Avian invasions. The ecology and evolution of exotic birds.Oxford University Press, Oxford

Brook BW (2008) Synergies between climate change, extinctions and invasive invertebrates. Wildl Res35:249–252. doi:10.1071/WR07116

Brook BW, Sodhi NS, Bradshaw CJA (2008) Synergies among extinction drivers and global change. TrendsEcol Evol 23:453–460. doi:10.1016/j.tree.2008.03.011

Cannon RJC (1998) The implications of predicted climate change for insect pests in the UK, with emphasison non-indigenous species. Global Change Biol 4:785–796. doi:10.1046/j.1365-2486.1998.00190.x

Carroll SP, Fox CW (2007) Dissecting the evolutionary impacts of plant invasions: bugs and beetles asnative guides. Global Change Biol 13:1644–1657. doi:10.1111/j.1365-2486.2007.01403.x

Chown SL, Froneman PW (2008) The Prince Edward Islands. Land–sea interactions in a changing climate.African Sun Media, Stellenbosch

Chown SL, Gaston KJ (2008) Macrophysiology for a changing world. Proc R Soc B 275:1469–1478. doi:10.1098/rspb.2008.0137

Chown SL, Nicolson SW (2004) Insect physiological ecology. Mechanisms and patterns. Oxford UniversityPress, Oxford

Chown SL, McGeoch MA, Marshall DJ (2002) Diversity and conservation of invertebrates on the sub-Antarctic Prince Edward Islands. Afr Entomol 10:67–82

Chown SL, Slabber S, McGeoch MA et al (2007) Phenotypic plasticity mediates climate change responsesamong invasive and indigenous arthropods. Proc R Soc B 274:2661–2667. doi:10.1098/rspb.2007.0772

Cooper J, Condy PR (1988) Environmental conservation at the sub-Antarctic Prince Edward Islands: areview and recommendations. Environ Conserv 15:317–326

Daehler CC (2003) Performance comparisons of co-occurring native and alien invasive plants: implicationsfor conservation and restoration. Ann Rev Ecol Evol Syst 34:183–211. doi:10.1146/annurev.ecolsys.34.011802.132403

David JR, Gibert P, Gravot E et al (1997) Phenotypic plasticity and developmental temperature in Dro-sophila: analysis and significance of reaction norms of morphometrical traits. J Thermal Biol22:441–451. doi:10.1016/S0306-4565(97)00063-6

Davis MB, Shaw RG (2001) Range shifts and adaptive responses to quaternary climate change. Science292:673–679. doi:10.1126/science.292.5517.673

de Jong G (1995) Phenotypic plasticity as a product of selection in a variable environment. Am Nat145:493–512. doi:10.1086/285752

de Jong G (2005) Evolution of phenotypic plasticity: patterns of plasticity and the emergence of ecotypes.New Phytol 166:101–118

de Jong G, van der Have TM (2008) Temperature dependence of development rate, growth rate and size:from biophysics to adaptation. In: Whitman DW, Ananthakrishnan TN (eds) Phenotypic plasticity ofinsects: mechanisms and consequence. Science, Enfield, pp 461–526

de Mazancourt C, Johnson E, Barraclough TG (2008) Biodiversity inhibits species’ evolutionary responsesto changing environments. Ecol Lett 11:380–388. doi:10.1111/j.1461-0248.2008.01152.x

Deere JA, Chown SL (2006) Testing the beneficial acclimation hypothesis and its alternatives for locomotorperformance. Am Nat 168:630–644. doi:10.1086/508026

Deharveng L (1981) Collemboles des iles subantarctiques de l’Ocean Indien Mission J. Trave 1972–1973.Comite National Francaise des Recherches Antarctiques 48:33–108

Deutsch CA, Tewksbury JJ, Huey RB et al (2008) Impacts of climate warming on terrestrial ectothermsacross latitude. Proc Natl Acad Sci USA 105:6668–6672. doi:10.1073/pnas.0709472105

Diamond JM (1989) Overview of recent extinctions. In: Western D, Pearl MC (eds) Conservation for thetwenty-first century. Oxford University Press, Oxford, pp 37–41

Didham RK, Tylianakis JM, Gemmell NJ et al (2007) Interactive effects of habitat modification and speciesinvasion on native species decline. Trends Ecol Evol 22:489–496. doi:10.1016/j.tree.2007.07.001

Dukes J, Mooney HA (1999) Does global change increase the success of biological invaders? Trends EcolEvol 14:135–139. doi:10.1016/S0169-5347(98)01554-7

Duncan RP, Blackburn TM, Sol D (2003) The ecology of bird introductions. Annu Rev Ecol Evol Syst34:71–98. doi:10.1146/annurev.ecolsys.34.011802.132353

Dybdahl MF, Kane SL (2005) Adaptation vs. phenotypic plasticity in the success of a clonal invader.Ecology 86:1592–1601. doi:10.1890/04-0898

1378 Evol Ecol (2010) 24:1365–1380

123

Dzialowski AR, Lennon JT, O’Brien WJ et al (2003) Predator-induced phenotypic plasticity in the exoticcladoceran Daphnia lumholtzi. Freshw Biol 48:1593–1602. doi:10.1046/j.1365-2427.2003.01111.x

Elton CS (1958) The ecology of invasions by animals and plants. Methuen, LondonFjellberg A (1998) Fauna Entomologica Scandinavica Volume 35. The Collembola of Fennoscandia and

Denmark. Part I: Poduromorpha. Brill, LeidenFrazier MR, Huey RB, Berrigan D (2008) Thermodynamics constrains the evolution of insect population

growth rates: ‘‘warmer is better’’. Am Nat 168:512–520. doi:10.1086/285797Frenot Y, Chown SL, Whinam J et al (2005) Biological invasions in the Antarctic: extent, impacts and

implications. Biol Rev 80:45–72. doi:10.1017/S1464793104006542Gabriel AGA, Chown SL, Barendse J et al (2001) Biological invasions on Southern Ocean islands: the

Collembola of Marion Island as a test of generalities. Ecography 24:421–430. doi:10.1111/j.1600-0587.2001.tb00477.x

Gaston KJ, Chown SL, Mercer RD (2001) The animal species-body size distribution of Marion Island. ProcNatl Acad Sci USA 98:14493–14496. doi:10.1073/pnas.251332098

Geister TL, Lorenz MW, Hoffmann KH et al (2009) Energetics of embryonic development: effects oftemperature in egg and hatchling composition in a butterfly. J Comp Physiol B 179:87–98. doi:10.1007/s00360-008-0293-5

Ghalambor CK, McKay JK, Carroll SP et al (2007) Adaptive versus non-adaptive phenotypic plasticity andthe potential for contemporary adaptation in new environments. Funct Ecol 21:394–407. doi:10.1111/j.1365-2435.2007.01283.x

Gilchrist GW (1996) A quantitative genetic analysis of thermal sensitivity in the locomotor performancecurve of Aphidius ervi. Evolution 50:1560–1572

Halsey LG, Butler PJ, Blackburn TM (2006) A phylogenetic analysis of the allometry of diving. Am Nat167:276–287. doi:10.1086/499439

Hobbs RJ, Mooney HA (2005) Invasive species in a changing world: the interactions between global changeand invasives. In: Mooney HJ, Mack RN, McNeely JA, Neville LE, Schei PJ, Waage JK (eds) Invasivealien species. A new synthesis. Island Press, Washington, pp 310–331

Holzapfel AM, Vinebrooke RD (2005) Environmental warming increases invasion potential of alpine lakecommunities by imported species. Global Change Biol 11:2009–2015. doi:10.1111/j.1365-2486.2005.001057.x

Hopkin S (1997) Biology of the springtails. Insecta: Collembola. Oxford University Press, OxfordHugo EA, McGeoch MA, Marshall DJ et al (2004) Fine scale variation in microarthropod communities

inhabiting the keystone species Azorella selago on Marion Island. Polar Biol 27:446–473. doi:10.1007/s00300-004-0614-4

Ikemoto T (2005) Intrinsic optimum temperature for development of insects and mites. Environ Entomol34:1377–1387. doi:10.1603/0046-225X-34.6.1377

Izem R, Kingsolver JG (2005) Variation in continuous reaction norms: quantifying directions of biologicalinterest. Am Nat 166:277–289. doi:10.1086/431314

Janion C, Worland MR, Chown SL (2009) Assemblage level variation in lower lethal temperature: the roleof invasive species on sub-Antarctic Marion Island. Physiol Entomol 34:284–291. doi:10.1111/j.1365-3032.2009.00689.x

Kingsolver JG, Huey RB (1998) Evolutionary analyses of morphological and physiological plasticity inthermally variable environments. Am Zool 38:545–560. doi:10.1093/icb/38.3.545

Lawton JH, Brown KC (1986) The population and community ecology of invading insects. Phil Trans R SocB 314:607–617

le Roux PC, McGeoch MA (2008) Changes in climate extremes, variability and signature on sub-AntarcticMarion Island. Clim Change 86:309–329. doi:10.1007/s10584-007-9259-y

Lee CE (2002) Evolutionary genetics of invasive species. Trends Ecol Evol 17:386–391. doi:10.1016/S0169-5347(01)02405-3

Lee CE, Remfert JL, Gelembiuk GW (2003) Evolution of physiological tolerance and performance duringfreshwater invasions. Integr Comp Biol 43:439–449. doi:10.1093/icb/43.3.439

Lee C, Remfert J, Chang Y-M (2007) Response to selection and evolvability of invasive populations.Genetica 129:179–192. doi:10.1016/S0169-5347(01)02405-3

Myburgh M, Chown SL, Daniels SR et al (2007) Population structure, propagule pressure, and conservationbiogeography in the sub-Antarctic: lessons from indigenous and invasive springtails. Divers Distr13:143–154. doi:10.1111/j.1472-4642.2007.00319.x

Paradis E, Claude J, Strimmer K (2004) APE: analysis of phylogenetics and evolution in R language.Bioinformatics 20:289–290

Parmesan C (2007) Influences of species, latitudes and methodologies on estimates of phenological responseto global warming. Global Change Biol 13:1860–1872. doi:10.1111/j.1365-2486.2007.01404.x

Evol Ecol (2010) 24:1365–1380 1379

123

Parr CL, Sinclair BJ, Andersen AN et al (2005) Constraint and competition in assemblages: a cross-continental and modeling approach for ants. Am Nat 165:481–494

Potapov M (2001) Synopses on Palaearctic Collembola, volume 3, Isotomidae. In: Dunger W (ed) Staa-tliches Museum fur Naturkunde, Gorlitz

Quinn GP, Keough MJ (2002) Experimental design and data analysis for biologists. Cambridge UniversityPress, Cambridge

Richardson DM, Pysek P (2006) Plant invasions: merging the concepts of species invasiveness and com-munity invasibility. Progress Phys Geogr 30:409–431. doi:10.1191/0309133306pp490pr2006

Roff DA (2002) Life history evolution. Sinauer Associates, SunderlandRosecchi E, Thomas F, Crivelli AJ (2001) Can life-history traits predict the fate of introduced species? A

case study on two cyprinid fish in southern France. Freshw Biol 46:845–853. doi:10.1046/j.1365-2427.2001.00715.x

Rusek J (1998) Biodiversity of Collembola and their functional role in the ecosystem. Biodiv Conserv7:1207–1219

Sala OE, Chapin FS, Armesto JJ et al (2000) Global biodiversity scenarios for the year 2100. Science287:1770–1774. doi:10.1126/science.287.5459.1770

Scheiner SM (1993) Genetics and evolution of phenotypic plasticity. Ann Rev Ecol Syst 24:35–68. doi:10.1146/annurev.es.24.110193.000343

Sibly RM, Calow P (1986) Physiological ecology of animals. An evolutionary approach. Blackwell Sci-entific Publications, Oxford

Slabber S, Worland MR, Leinaas HP et al (2007) Acclimation effects on thermal tolerances of springtailsfrom sub-Antarctic Marion Island: indigenous and invasive species. J Insect Physiol 53:113–125. doi:10.1016/j.jinsphys.2006.10.010

Stachowicz JJ, Terwin JR, Whitlatch RB et al (2002) Linking climate change and biological invasions:ocean warming facilitates nonindigenous species invasions. Proc Natl Acad Sci USA 99:15497–15500.doi:10.1073/pnas.242437499

Stearns SC (1992) The evolution of life histories. Oxford University Press, OxfordStevens M, Greenslade P, Hogg ID et al (2006) Southern Hemisphere springtails: could any have survived

glaciation of Antarctica? Mol Biol Evol 23:574–882. doi:10.1093/molbev/msj073Stillwell RC, Fox CW (2005) Complex patterns of phenotypic plasticity: interactive effects of temperature

during rearing and oviposition. Ecology 86:924–934. doi:10.1890/04-0547Stohlgren TJ, Barnett DT, Jarnevich CS et al (2008) The myth of plant species saturation. Ecol Lett

11:313–322. doi:10.1111/j.1461-0248.2008.01153.xSwofford DL (2001) PAUP*: Phylogenetic Analysis Using Parsimony (*and other methods). Sinauer

Associates, SunderlandTheoharides KA, Dukes JS (2007) Plant invasion across space and time: factors affecting nonindigenous

species success during four stages of invasion. New Phytol 176:256–273. doi:10.1111/j.1469-8137.2007.02207.x

Trudgill DL, Honek A, van Straalen NM (2005) Thermal time: concepts and utility. Ann Appl Biol146:1–14. doi:10.1111/j.1744-7348.2005.04088.x

Trussell GC, Smith LD (2000) Induced defenses in response to an invading crab predator: an explanation ofhistorical and geographic phenotypic change. Proc Natl Acad Sci USA 97:2123–2127. doi:10.1073/pnas.040423397

van Kleunen M, Fischer M (2005) Constraints on the evolution of adaptive phenotypic plasticity in plants.New Phytol 166:49–60. doi:10.1111/j.1469-8137.2004.01296.x

van Kleunen M, Johnson SD (2007) South African Iridaceae with rapid and profuse seedling emergence aremore likely to become naturalized in other regions. J Ecol 95:674–681. doi:10.1111/j.1365-2745.2007.01250.x

van Kleunen M, Manning JC, Pasqualetto V et al (2008) Phylogenetically independent associations betweenautonomous self fertilization and plant invasiveness. Am Nat 171:195–201. doi:10.1086/525057

van Straalen NM (1994) Adaptive significance of temperature responses in Collembola. Acta Zool Fennica195:135–142

Via S, Gomulkiewicz R, DeJong G et al (1995) Adaptive phenotypic plasticity: consensus and controversy.Trends Ecol Evol 10:212–217. doi:10.1016/S0169-5347(00)89061-8

Walther G-R, Roques A, Hulme PE et al (2009) Alien species in a warmer world: risks and opportunities.Trends Ecol Evol 24:686–693. doi:10.1016/j.tree.2009.06.008

Ward NL, Masters GJ (2007) Linking climate change and species invasion: an illustration using insectherbivores. Global Change Biol 13:1605–1615. doi:10.1111/j.1365-2486.2007.01399.x

Warren MS, Hill JK, Thomas JA et al (2001) Rapid responses of British butterflies to opposing forces ofclimate and habitat change. Nature 414:65–69. doi:10.1038/35102054

1380 Evol Ecol (2010) 24:1365–1380

123