The roles of demography and genetics in the early stages of colonization

Abstract

Colonization success increases with the size of the founding group. Both demographic and genetic factors underlie this relationship, yet because genetic diversity normally increases with numbers of individuals, their relative importance remains unclear. Furthermore, their influence may depend on the environment and may change as colonization progresses from establishment through population growth and then dispersal. We tested the roles of genetics, demography and environment in the founding of Tribolium castaneum populations. Using three genetic backgrounds (inbred to outbred), we released individuals of four founding sizes (2–32) into two environments (natal and novel), and measured establishment success, initial population growth and dispersal. Establishment increased with founding size, whereas population growth was shaped by founding size, genetic background and environment. Population growth was depressed by inbreeding at small founding sizes, but growth rates were similar across genetic backgrounds at large founding size, an interaction indicating that the magnitude of the genetic effects depends upon founding population size. Dispersal rates increased with genetic diversity. These results suggest that numbers of individuals may drive initial establishment, but that subsequent population growth and spread, even in the first generation of colonization, can be driven by genetic processes, including both reduced growth owing to inbreeding depression, and increased dispersal with increased genetic diversity.

1. Introduction

Understanding the dynamics governing the founding of new populations is a central focus of evolutionary biology [1–3] and has clear application in conservation, restoration, biological control and invasion biology [4–6]. The most consistent predictor of successful colonization is the number of individuals founding a population [7,8]. Large founding populations can result in high colonization success owing to both demographic and genetic processes. The demographic consequences of having more individuals in a founding population include buffering against demographic and environmental stochasticity and minimizing Allee effects [9,10]. The genetic consequences of a large founding population include reducing the probability of inbreeding and thus also of inbreeding depression and increasing genetic variation with which founders can adapt to the new environment [11]. Clearly, both demographic and genetic processes drive the relationship between number of founders and colonization success, yet because in nature genetic variation typically increases with the number of individuals, the comparative importance of demography and genetics remains unclear [6]. In this study, we experimentally decouple demography from genetics to test the relative contributions of each to colonization success.

The debate over the roles of demography (numbers of individuals in the population) and genetics (diversity of alleles in the population) in the persistence of small populations, such as founding populations, is polarized. In an influential article (cited more than 1000 times), quantitative geneticist Lande [1] argued that in most cases demographic factors will decide the fate of small populations before genetic factors can impact them. However, recent studies emphasize that genetics can play an important role even in the early stages of colonization [12,13]. With evidence mounting that both demography and genetics can play important roles in founding success (table 1), research is needed that investigates their relative importance at different stages of colonization [6,17].

Table 1.

Summary of studies that manipulate demography (D) and genetics (G) of founding populations simultaneously, and that measure establishment success and/or performance either in one or more environments (E). (Asterisks (*) indicate significant effects of individual factors and their interactions, including the level of significance, while ‘n.s.’ indicates non-significant results. *p < 0.05, **p < 0.01, ***p < 0.001.)

G	D	G × D	E	G × E	D × E	species	reference
establishment success
*	*	n.a.				Aquarius najas	Ahlroth et al. [14] (fig. 1)
***	n.s.	n.s.				Arabidopsis thaliana	Crawford & Whitney [15] (table 3a)
n.s.	***	n.s.				Senecio vernalis	Erfmeier et al. [16] (table 1)
*	n.s.	n.s.	***	*	n.s.	Bemisia tabaci	Hufbauer et al. [17] (table 2)
n.s.	***	n.s.	n.s.	n.s.	n.s.	Tribolium castaneum	this study
performance
***	***	n.s.				Raphanus sativus	Elam et al. [18] (table 1)
*	n.s.	n.s.				Arabidopsis thaliana	Crawford & Whitney [15] (table 3)
*	***	n.s.*,b				Lolium multiflorum	Firestone & Jasieniuk [19] (table 1 and 2b)
*	n.s.	n.s.	***	n.s.	n.s.	Bemisia tabaci	Hufbauer et al. [17] (table 2)
***	***	***	***	***	**	Tribolium castaneum	this study

^aPer cent seedling emergence used as a measure of establishment success, and number of fruits as performance measure.

^bSignificance depends on the measure used for population size. The interaction is non-significant when the number of individuals is used but it is significant when above ground biomass is used as a proxy for population size.

Colonization can be divided into three important stages: the founding event, establishment and spread. Evidence to date suggests that demography is most critical in the first generation of colonization, with larger groups of founders being better buffered against demographic and environmental stochasticity, and thereby enabling some founding individuals to survive and reproduce [20,21]. The establishment phase, on the other hand, is sensitive to the genetic composition of the founders because genetic variation can dictate population growth rates with higher variation leading to higher population growth [22]. It can also strengthen a population's ability to respond to selection [23,24]. In combination, higher growth and more rapid adaptation strongly shape establishment. Geographical spread of populations following establishment is probably driven by a combination of demographic, genetic and environmental factors. Dispersal usually increases with population density as a result of competition [25–27]. On the other hand, individual responses to intraspecific competition [28] and dispersal tendencies both [24,29] have strong genetic components. Dispersal tendencies can increase in populations with related individuals as a strategy to avoid kin competition [30] or inbreeding [31]. However, inbreeding can reduce dispersal rates by reducing fitness and imposing energetic constraints on inbred individuals [32,33]. In addition, the propensity to disperse can increase in fluctuating environments and in lower quality habitats as competition for limited resources increases [34–36].

The degree of compatibility between the founding individuals and their new environment is a critical factor determining colonization success. Suboptimal environments result in lower growth rates (demonstrating demographic effects) and inbreeding depression (demonstrating genetic effects). Inbreeding depression can manifest to varying degrees according to the environment [37,38], often being more severe in stressful environments [39,40]. Inbreeding depression may increase in stressful environments because the expression of genetic load (deleterious alleles) can change under stress [41,42]. For example, the dominance of deleterious alleles as well as selection against deleterious recessive alleles was shown to increase with stress, resulting in higher mortality of genotypes homozygous at deleterious loci [41]. The differential expression of deleterious alleles can also lower fitness of individuals by increasing their overall susceptibility to environmental stress and by hindering evolutionary responses, such as adaptive phenotypic plasticity, which could provide means for short-term survival in the face of sudden environmental change [42,43].

Designing experiments to disentangle the relative contributions of demography and genetics to colonization is difficult because these two factors covary. Larger populations typically harbour more genetic variation simply because they have more individuals and small populations harbour less genetic variation because they have fewer individuals. A growing number of studies are now controlling for genetic variation by simultaneously manipulating the number of founders and the genetic background of founding individuals. Unfortunately, these studies have not yet led to a consensus. They have found that demography [16], genetics [15,17] or both demography and genetics [14] can be responsible for initial establishment success (table 1). Subsequent performance of newly established populations depend either on the genetic background of founders [15,17] or on demography and genetics together [18,19].

We used a factorial experimental design to explicitly test the interactions between demography, genetics and environment during colonization. Using a biological model system, red flour beetles (Tribolium castaneum: Coleoptera, Tenebrionidae), we created replicate beetle lineages varying in their genetic background from inbred to outbred, which were then released at four different founding population sizes into experimental arenas in two different environments (natal/benign or novel/stressful). We divided the initial stages of colonization into three discrete, biologically relevant stages: (i) establishment success, defined to occur when offspring of the founding individuals survive to adulthood, (ii) initial population growth, which reveals differences in the vigor of colonizing populations, and (iii) dispersal beyond the location of the initial introduction, which gives insight into the long-term trajectory of the new populations.

We anticipated that propagule size would largely determine establishment success and that the negative effects of low genetic diversity could become also apparent in the novel stressful environment even at this early stage. We expected genetic diversity and the quality of the environment both to influence initial population growth and spread.

2. Material and methods

(a). Rearing

The experiment took place in 4 × 4 × 6 cm plastic boxes (hereafter ‘patches’) with 14 g of medium at 31°C and an average of 54% humidity. Rearing procedures were designed to mimic a seasonally breeding organism with discrete generations and an adult dispersal phase [44,45]. This life history is widespread among animals and plants. Populations were founded by placing adults in a patch with fresh medium to mate and lay eggs for 24 h. At the end of this period, adults were removed and discarded, and offspring were allowed to develop undisturbed into adults. These adults were censused after five weeks and used to evaluate dispersal.

(b). Experimental design

Our treatments, described in detail below, included three genetic backgrounds (inbred, low diversity to outbred and high diversity), four founding population sizes (2–32) and two environments (benign and stressful), applied in a 3 × 4 × 2 factorial experimental design.

We used T. castaneum populations from three sources (see the electronic supplementary material) to create all experimental lineages with three genetic backgrounds. The base populations were reared in panmictic colonies of at least 500 individuals (hereafter, standard lineages). A single generation of full-sibling mating was used to create populations that had reduced genetic diversity and were also more inbred (hereafter, inbred). Reciprocally crossing beetles from the three different populations (each done both male to female and female to male) was used to create populations that had higher genetic diversity (hereafter, outbred). Thus, for each genetic background (inbred, standard, outbred) there were three replicate, genetically distinct lineages.

We founded populations with 2, 4, 12 or 32 individuals. The low end of this range was chosen as the size that maximizes the role of demographic stochasticity, including stochasticity in sex ratio. The upper end was chosen as a population size in which stochasticity in sex ratio is low, and, at least in the benign environment, competition for resources should be low.

The 12 types of founders (three genetic backgrounds by four population sizes) were released into two environments: the natal environment, which was a benign environment consisting of standard rearing medium (95% wheat flour (enriched), 5% brewer's yeast by weight), and a novel and stressful environment, which had inferior nutritional quality, made up of 99% wheat flour (non-enriched) and 1% brewer's yeast.

We counted emerging adults in the first generation following population founding. We defined the presence of at least two offspring as establishment success. Initial population growth rate was calculated by dividing the number of offspring in generation one by the founding population size. Dispersal was measured in linear arrays following Melbourne & Hastings [44] on a subset of experimental populations by choosing randomly two replicates of founding size 32 of each population type, breeding level and environment. Arrays were comprised five patches held together by stout rubber bands, connected via 2 mm holes drilled in the sides and aligned to allow movement. All emerging adults were placed in the first patch of dispersal arrays following census and allowed to move across the arrays for 24 h. At the end of this period, the number of beetles in each patch was recorded.

To evaluate whether the lineages of the three genetic backgrounds differed in genetic diversity, we genotyped 19–23 individuals at 12 microsatellite loci from each of the nine lineages used (three initial populations × three breeding levels) (for details of molecular analyses and their results, see the electronic supplementary material and tables S1–S2).

(c). Data analyses

We evaluated the influence of genetic background, founding size, environment and their interactions on establishment success (binary response) using a generalized linear mixed model (SAS 9.3, PROC GLIMMIX) in which temporal block, colony identity, and interactions between block and genetic background, and block and colony identity were included as random effects (see the electronic supplementary material for more details). All but one population founded with 12 individuals and all populations founded with 32 individuals established, and thus models including those two founding sizes could not converge as there is no uncertainty to estimate (the separation problem, [46]). We thus modelled establishment success only for founding sizes 2 and 4. We treated founding size as a categorical effect, as that produced models with lower Akaike information criterion values. We also evaluated whether or not establishment rates differed from that expected given stochasticity in sex ratio. Fifty per cent of populations initiated by two founders can be expected to include a male and female, and therefore possibly establish, whereas 87.5% of populations initiated by four founders can be expected to include both sexes. Thus, we compared establishment rates of populations from different genetic backgrounds against those expectations. We examined population growth and dispersal rates with linear mixed models (PROC MIXED). The model of population growth included the same effects as that for establishment success. Population growth rates were log transformed to meet the assumption of constant variance. Mean and maximum dispersal distances were evaluated for the largest founding size (32) only using linear mixed models (PROC MIXED). The models included the same random effects as previous analyses, and genetic background, environment, their interaction and population size at time of dispersal as fixed effects.

3. Results

(a). Genetic diversity

Our goal to create populations with low, intermediate and higher genetic diversities was met. Inbreeding reduced diversity, and outbreeding increased diversity relative to the colonies from which they were created (table 2).

Table 2.

Genetic diversity measures for experimental lineages of T. castaneum over 12 microsatellite loci. (Mean values are shown for the three inbred, standard and outbred populations. Number of individuals (N), number of alleles (N_a), number of effective alleles (N_e), observed heterozygosity (H_o), and unbiased expected heterozygosity (uHe) as calculated in GenAlEx 6.5.)

breeding levels	N	Na	Ne	Ho	uHe
inbred	65	2.00	1.57	0.244	0.279
standard	66	3.14	2.15	0.382	0.426
outbred	65	4.11	2.64	0.458	0.569

(b). Establishment

The only factor that significantly influenced establishment success was the size of the founding population (F_1,1374 = 189.30, p < 0.0001). There was not a significant effect of environmental quality or genetic background of founders (F_1,1374 = 0.47, p = 0.4921; F_2,8 = 1.30, p = 0.3237, respectively), and none of the interactions were significant. Establishment was lowest at the smallest founding size (2) with an average 43% success rate, which increased to about 83% at n = 4 (figure 1). All but one of those founded with 12 individuals and all populations founded with 32 individuals established. When comparing establishment against expected proportions for sexual organisms, the inbred populations founded with two individuals had substantially lower establishment rates than expected (mean 36.3% with an upper 95% confidence interval (CI) of 45.5% versus the expected 50%). The standard populations founded with four individuals had somewhat lower establishment rates than expected (mean 80.3% with an upper 95% CI of 87.0% versus the expected 87.5%).

Figure 1.

Establishment success of T. castaneum populations founded at four different population sizes (2, 4, 12 or 32 individuals) and with three different genetic backgrounds (inbred, standard and outbred). Bars indicate 95% confidence intervals (CIs) for founding sizes of 2 and 4. Raw values of 1 or close to it are graphed for populations founded with 12 and 32 individuals, as there is essentially no uncertainty in those estimates (see text for statistical details). Overlapping data points on the x-axis are offset to improve clarity.

(c). Population growth

Population growth decreased with increasing founding population size (F_3,1467 = 104.06, p < 0.0001; figure 2a), exhibiting classic negative density dependence [47]. Population growth was also reduced by inbreeding (genetic background F_2,8 = 30.55, p = 0.0002), indicating inbreeding depression, and was lower in the novel environment relative to the natal environment (F_1,1467 = 159.54, p < 0.0001; figure 2b). A significant interaction between genetic background and founding size (F_6,1467 = 4.27, p = 0.0003) was owing to genetic background having a larger effect at small founding size than at large founding size (figure 2a): inbred populations exhibited the lowest growth rate at small founding size, but less negative density dependence than standard and outbred populations, and thus population growth rates at large founding sizes were comparable across genetic lineages. Similarly, a significant interaction between genetic background and environment (F_2,1467 = 7.74, p = 0.0005) was owing to genetic background having more potent effects in the natal environment than the novel environment (figure 2b). Interestingly, though the standard and outbred lineages did not differ significantly in their performance (figure 2a,b), the means for outbred lineages were consistently slightly lower than those for standard lineages, suggesting weak outbreeding depression with respect to population growth. An interaction between founding size and environment was statistically significant (F_3,1467 = 5.25, p = 0.0013) but does not appear biologically interesting, as differences in how strongly the environment affects growth rates of populations with different numbers of founders were minor (electronic supplementary material, figure S1).

Figure 2.

Initial growth rates (N₁/N₀) of T. castaneum populations (a) founded at four different population sizes (2, 4, 12 or 32 individuals) and with three different genetic backgrounds (inbred, standard and outbred), (b) in a novel (stressful) versus a natal (benign) environment. Bars indicate 95% confidence intervals (CIs). Note that statistical tests were based on log-transformed data but back-transformed values are shown on the graphs. Overlapping data points on the x-axes are offset to improve clarity.

(d). Dispersal

Dispersal through patch space increased with genetic diversity in both environments. This pattern was most pronounced for maximum dispersal (F_2,8 = 9.16, p = 0.0085; figure 3). Mean dispersal showed the same pattern, however, suggesting that the results are not driven only by a few strong dispersers moving further, but may reflect population level processes (F_2,8 = 4.34, p = 0.0529; electronic supplementary material, figure S2). Dispersal was greater in the novel environment than the natal environment, but that was significant only for maximum dispersal rates (F_1,144 = 5.76, p = 0.0177; figure 3).

Figure 3.

Dispersal rates (maximum number of patches moved) of T. castaneum populations founded with 32 individuals and with three different genetic backgrounds (inbred, standard, and outbred) in a novel (stressful) versus a natal (benign) environment. Bars indicate 95% confidence intervals (CIs). Overlapping data points on the x-axis are offset to improve clarity.

4. Discussion

Both demography and genetics are important in explaining why larger founding populations are more successful, but their relative effects are unknown, largely because experimental research is lacking [6,14]. We experimentally manipulated the number of founders, their genetic background and the quality of the founding environment in a full-factorial experimental design to examine how these factors and their interactions influence the initial stages of colonization. We discuss observed patterns during three stages: (i) establishment success, (ii) initial population growth, and (iii) subsequent dispersal.

(a). Establishment

Initial establishment success depended strongly on the size of the founding group (figure 1). This is consistent with both reviews [7,8] and empirical studies [20,48] showing that colonization success increases with the number of founders. The overwhelming importance of number of founders held even under stressful novel environmental conditions when genetic variation might be expected to play a larger role. A detailed look at effect of genetic background on establishment showed that inbred populations at the smallest founding size (2) established 14% less often than expected, given natural stochasticity in sex ratio. Curiously, standard populations showed slightly lower than expected establishment at the founding size of 4, but the 95% CI of that estimate almost included the expectation. These results indicate that establishment success is mostly driven by variation in sex ratio at small founding sizes but that beyond the demographic processes the negative effects of inbreeding can further reduce establishment.

Of the five studies that manipulate demography and genetics (table 1), two found that demography alone explains initial establishment, two others found that genetics (either alone or in combination with environment) explains initial establishment, and one found both to be important. With these contrasting results, drawing general conclusions about the effects of demography and genetics on establishment success is not yet feasible. Clearly, both can play a role but what determines which will be important under various introduction scenarios remains an open question.

The quality of the environment may be key in answering that question. Hufbauer et al. [17] found that genetic background had no influence in establishment on a natal host (environment), but played a key role in establishment on a novel host, where outbred founders were much more successful than inbred ones. In our study, the environment did not influence establishment success, however, our environment did not represent a fundamentally novel resource as in the case above [17]. The data available lead to the prediction that the effects of low genetic diversity and inbreeding will be more pronounced in an environment not encountered before, owing at least in part to the well-established pattern that inbreeding depression is often environment dependent [39,40,49]. This prediction suggests that intentional and unintentional introductions may differ in their sensitivity to inbreeding. Deliberate introductions (e.g. reintroductions, restoration and biological control) are unlikely to be released into a completely novel environment, perhaps making them less susceptible to the effects of inbreeding depression in the initial establishment phase, whereas unintentional introductions (exotic species) may indeed find themselves in an environment drastically different from their native habitats, where genetic background could be critical.

(b). Population growth

Our results show that both demography and genetics can influence population growth rates in newly founded populations, as well as the environment where the introduction takes place. Population growth rate did not increase with founder size, but decreased. Thus, after initial establishment, we observed classic negative density dependence. The two other studies that found significant demographic effects (table 1) investigated self-incompatible plant species, and their results highlight the importance of positive density-dependent processes, such as Allee effects [18,19]. In the case of our model system, the mechanism for negative density dependence is probably increased egg cannibalism with increasing population densities [47]. Interestingly, negative density dependence was more pronounced in more diverse founders, than inbred founders (figure 2a), supporting previous findings that cannibalism is reduced among close relatives [50]. Such potential benefits of low diversity are also seen in the reduced aggression of genetically depauperate invasive ants [51].

Inbred populations had lower growth rates than standard and outbred populations at the smaller founding sizes (2, 4 or 12 founders) but not at the highest founding size (32) (figure 2a). This interaction is important as it highlights that genetic effects can depend on the number of individuals in the population and that the negative impacts of even strong inbreeding, such as shown here, may not become apparent until populations become very small. Because our study focuses on only the first generation, increased importance of inbreeding at smaller founding sizes cannot be owing to additional build-up of inbreeding in smaller populations over time, but owing to the immediate effects of being inbred relative to being outbred.

Growth rates of standard and outbred populations were lower in the novel relative to the natal environment. However, we did not see the commonly observed pattern of increased inbreeding depression in the low-quality environment [40] as inbred populations had similarly low growth rates in both environments (figure 2b).

In our study, the outbred lineages with the highest genetic diversity did not perform better than the standard lineages with respect to population growth, and in fact showed weak signs of outbreeding depression as their growth rates were consistently (though not significantly) slightly lower than those of the standard lineages (figure 2a,b). Outbreeding depression is usually attributed to the break-up of co-adapted gene complexes upon mating between divergent populations and may manifest in the F₁ [52] or more likely in later generations [53,54]. However, the effects of outbreeding may differ among traits and during different life stages [52,55], among genotypes and generations [56,57], thus its impact on colonization is hard to predict. Indeed, in a study using the same model organism we use, outbreeding reduced extinction risk and enhanced adaptation to a novel resource over multiple generations [55].

Overall, our results strengthen the emerging pattern that genetic effects can dramatically shape population performance following founding. This has now been observed in each of the five experiments manipulating demography and genetics in both laboratory and field contexts (table 1) and is also supported by a recent meta-analysis of studies that manipulate genetic background but hold numbers of founders constant [12,13]. Multiple mechanisms underlie the role of genetic background, including the negative effects of reduced genetic diversity and the positive effects of increased genetic diversity [17]. Populations with low genetic diversity commonly show reduced fitness as a result of inbreeding depression or owing to increased genetic load [58]. This is the case in our study, as well as others like ours (table 1) [17,18,19].

The positive effects of increased variation in colonizing populations include more rapid adaptation to novel conditions as well as longer term population persistence and stability [55,59,60]. In the short term, rapid adaptation can be facilitated by sampling effects—simply the presence of an appropriate genotype for the environment, whereas in the long term, admixture will lead to a diverse array of genotypes on which selection can act [61,62]. Having diverse genotypes in a population could also reduce competition through niche partitioning [63,64], and interactions among genotypes may facilitate the success of others [65,66]. Outbreeding can also have immediate fitness effects by masking genetic load, which can result in hybrid vigor in the first generation that can lead to increased population growth, and thus establishment success [11].

(c). Dispersal

The quality of the environment and genetics affected dispersal tendencies of populations founded with 32 individuals. Dispersal was overall higher in the low-quality novel environment (figure 3). This is not surprising as increased dispersal in response to shortage of resources is common [35,36,67]. Genetic diversity had a strong effect on dispersal with inbred populations moving the shortest and outbred populations the longest distances in both the natal and novel environments (figure 3). These results are in line with studies which show that inbreeding can reduce dispersal propensity by reducing fitness of individuals and thus rendering them physically less able to spread [32,33]. It is interesting that standard populations dispersed less than outbred ones, even though they had similar population densities (figure 2a,b). In our study species, and other organisms, where dispersal occurs regardless of environmental conditions as a means of colonizing new habitats, individuals dispersing the furthest are assumed to be more fit and have characteristics that increase chances of successful colonization, such as faster development or higher fecundity [68,69]. Thus, outbred populations may possess traits (possibly faster development) that make them better dispersers but which were not measured in this study. Alternatively, the intensity of competition, and thus the tendency to leave a habitat, may change with the extent of relatedness among individuals. However, data are scarce and there is no consensus whether there is generally more or less competition between kin or unrelated individuals ([70] and references therein).

Variation in dispersal tendencies of founding populations can have various effects on colonization success. High rates of dispersal may prevent establishment by acting as a drain on a founding group [71]. The critical propagule size leading to establishment may be higher for species or populations with higher dispersal rates [15]. On the other hand, successful dispersal to unoccupied habitats may lead to the establishment of several small populations, spreading the risk of overall extinction [72]. In our experimental system, with clear evidence of negative density dependence, dispersal may often be beneficial, thus here we have evidence of a direct benefit of outbreeding and not only a negative effect of inbreeding.

5. Conclusion

Numbers of individuals and the genetic variation they harbour covary, making it difficult to determine the underlying causes of the consistently positive effect of the number of founding individuals on the founding of new populations. We evaluated founding through three sequential phases: initial establishment, population growth and dispersal. We found that the initial establishment success of populations is mainly driven by the size of the founding group but the negative effects of inbreeding can have negative impacts on the smallest populations even at this early stage. Demographic and genetic processes along with the quality of the environment can jointly affect population growth in the first generation with inbreeding and stressful environments reducing the fitness of founders. However, the negative effects of inbreeding were only apparent in small populations. While during establishment and population growth, negative effects of low genetic diversity were evident, in the dispersal stage positive effects of increased genetic diversity were revealed.

The emerging evidence shows that colonization success improves with number of founders not just owing to demographic processes, but also owing to the genetic background of the individuals in the founding group. Both demography and genetics play a role in initial establishment success, and it is not yet clear which will be stronger in different contexts. By contrast, while both mechanisms can play a role in initial population growth, the effect of genetic background is consistent and strong at that stage, while the effect of founding group size is more variable. The field now needs research that follows populations beyond the first generation to gain more insight into the long-term outcomes of colonization events, and how the interactions among demographic, genetic and environmental factors play out.

Data accessibility

Data for this study are available in Dryad.

Funding statement

Funding for this research was provided by the US National Science Foundation (DEB-0949619), and additional support came from the United States Department of Agriculture via the Colorado Agricultural Experiment Station.