LOCAL ADAPTATION WHEN COMPETITION DEPENDS ON PHENOTYPIC SIMILARITY

Recent work incorporating demographic–genetic interactions indicates the importance of population size, gene flow, and selection in influencing local adaptation. This work typically assumes that density‐dependent survival affects individuals equally, but individuals in natural population rarely compete equally. Among‐individual differences in resource use generate stronger competition between more similar phenotypes (frequency‐dependent competition) but it remains unclear how this additional form of selection changes the interactions between population size, gene flow, and local stabilizing selection. Here, we integrate migration–selection dynamics with frequency‐dependent competition. We developed a coupled demographic‐quantitative genetic model consisting of two patches connected by dispersal and subject to local stabilizing selection and competition. Our model shows that frequency‐dependent competition slightly increases local adaptation, greatly increases genetic variance within patches, and reduces the amount that migration depresses population size, despite the increased genetic variance load. The effects of frequency‐dependence depend on the strength of divergent selection, trait heritability, and when mortality occurs in the life cycle in relation to migration and reproduction. Essentially, frequency‐dependent competition reduces the density‐dependent interactions between migrants and residents, the extent to which depends on how different and common immigrants are compared to residents. Our results add new dynamics that illustrate how competition can alter the effects of gene flow and divergent selection on local adaptation and population carrying capacities.     

Fitness declines towards range limits and local adaptation to climate affect dispersal evolution during climate‐induced range shifts

Dispersal ability will largely determine whether species track their climatic niches during climate change, a process especially important for populations at contracting (low‐latitude/low‐elevation) range limits that otherwise risk extinction. We investigate whether dispersal evolution at contracting range limits is facilitated by two processes that potentially enable edge populations to experience and adjust to the effects of climate deterioration before they cause extinction: (i) climate‐induced fitness declines towards range limits and (ii) local adaptation to a shifting climate gradient. We simulate a species distributed continuously along a temperature gradient using a spatially explicit, individual‐based model. We compare range‐wide dispersal evolution during climate stability vs. directional climate change, with uniform fitness vs. fitness that declines towards range limits (RLs), and for a single climate genotype vs. multiple genotypes locally adapted to temperature. During climate stability, dispersal decreased towards RLs when fitness was uniform, but increased when fitness declined towards RLs, due to highly dispersive genotypes maintaining sink populations at RLs, increased kin selection in smaller populations, and an emergent fitness asymmetry that favoured dispersal in low‐quality habitat. However, this initial dispersal advantage at low‐fitness RLs did not facilitate climate tracking, as it was outweighed by an increased probability of extinction. Locally adapted genotypes benefited from staying close to their climate optima; this selected against dispersal under stable climates but for increased dispersal throughout shifting ranges, compared to cases without local adaptation. Dispersal increased at expanding RLs in most scenarios, but only increased at the range centre and contracting RLs given local adaptation to climate.     

Among‐year variation in selection during early life stages and the genetic basis of fitness in Arabidopsis thaliana

Incomplete information regarding both selection regimes and the genetic basis of fitness limits our understanding of adaptive evolution. Among‐year variation in the genetic basis of fitness is rarely quantified, and estimates of selection are typically based on single components of fitness, thus potentially missing conflicting selection acting during other life‐history stages. Here, we examined among‐year variation in selection on a key life‐history trait and the genetic basis of fitness covering the whole life cycle in the annual plant Arabidopsis thaliana. We planted freshly matured seeds of >200 recombinant inbred lines (RILs) derived from a cross between two locally adapted populations (Italy and Sweden), and both parental genotypes at the native site of the Swedish population in three consecutive years. We quantified selection against the nonlocal Italian genotype, mapped quantitative trait loci (QTL) for fitness and its components, and quantified selection on timing of germination during different life stages. In all 3 years, the local Swedish genotype outperformed the nonlocal Italian genotype. However, both the contribution of early life stages to relative fitness, and the effects of fitness QTL varied among years. Timing of germination was under conflicting selection through seedling establishment vs. adult survival and fecundity, and both the direction and magnitude of net selection varied among years. Our results demonstrate that selection during early life stages and the genetic basis of fitness can vary markedly among years, emphasizing the need for multiyear studies considering the whole life cycle for a full understanding of natural selection and mechanisms maintaining local adaptation.     

QTL mapping of freezing tolerance: links to fitness and adaptive trade‐offs

Local adaptation, defined as higher fitness of local vs. nonlocal genotypes, is commonly identified in reciprocal transplant experiments. Reciprocally adapted populations display fitness trade‐offs across environments, but little is known about the traits and genes underlying fitness trade‐offs in reciprocally adapted populations. We investigated the genetic basis and adaptive significance of freezing tolerance using locally adapted populations of Arabidopsis thaliana from Italy and Sweden. Previous reciprocal transplant studies of these populations indicated that subfreezing temperature is a major selective agent in Sweden. We used quantitative trait locus (QTL) mapping to identify the contribution of freezing tolerance to previously demonstrated local adaptation and genetic trade‐offs. First, we compared the genomic locations of freezing tolerance QTL to those for previously published QTL for survival in Sweden, and overall fitness in the field. Then, we estimated the contributions to survival and fitness across both field sites of genotypes at locally adaptive freezing tolerance QTL. In growth chamber studies, we found seven QTL for freezing tolerance, and the Swedish genotype increased freezing tolerance for five of these QTL. Three of these colocalized with locally adaptive survival QTL in Sweden and with trade‐off QTL for overall fitness. Two freezing tolerance QTL contribute to genetic trade‐offs across environments for both survival and overall fitness. A major regulator of freezing tolerance, CBF2, is implicated as a candidate gene for one of the trade‐off freezing tolerance QTL. Our study provides some of the first evidence of a trait and gene that mediate a fitness trade‐off in nature.     

Cooperation‐mediated plasticity in dispersal and colonization

Kin selection theory predicts that costly cooperative behaviors evolve most readily when directed toward kin. Dispersal plays a controversial role in the evolution of cooperation: dispersal decreases local population relatedness and thus opposes the evolution of cooperation, but limited dispersal increases kin competition and can negate the benefits of cooperation. Theoretical work has suggested that plasticity of dispersal, where individuals can adjust their dispersal decisions according to the social context, might help resolve this paradox and promote the evolution of cooperation. Here, we experimentally tested the hypothesis that conditional dispersal decisions are mediated by a cooperative strategy: we quantified the density‐dependent dispersal decisions and subsequent colonization efficiency from single cells or groups of cells among six genetic strains of the unicellular Tetrahymena thermophila that differ in their aggregation level (high, medium, and low), a behavior associated with cooperation strategy. We found that the plastic reaction norms of dispersal rate relative to density differed according to aggregation level: highly aggregative genotypes showed negative density‐dependent dispersal, whereas low‐aggregation genotypes showed maximum dispersal rates at intermediate density, and medium‐aggregation genotypes showed density‐independent dispersal with intermediate dispersal rate. Dispersers from highly aggregative genotypes had specialized long‐distance dispersal phenotypes, contrary to low‐aggregation genotypes; medium‐aggregation genotypes showing intermediate dispersal phenotype. Moreover, highly aggregation genotypes showed evidence for beneficial kin‐cooperation during dispersal. Our experimental results should help to resolve the evolutionary conflict between cooperation and dispersal: cooperative individuals are expected to avoid kin‐competition by dispersing long distances, but maintain the benefits of cooperation by dispersing in small groups.     

Parasite‐mediated selection in a natural metapopulation of Daphnia magna

Parasite‐mediated selection varying across time and space in metapopulations is expected to result in host local adaptation and the maintenance of genetic diversity in disease‐related traits. However, nonadaptive processes like migration and extinction‐(re)colonization dynamics might interfere with adaptive evolution. Understanding how adaptive and nonadaptive processes interact to shape genetic variability in life‐history and disease‐related traits can provide important insights into their evolution in subdivided populations. Here we investigate signatures of spatially fluctuating, parasite‐mediated selection in a natural metapopulation of Daphnia magna. Host genotypes from infected and uninfected populations were genotyped at microsatellite markers, and phenotyped for life‐history and disease traits in common garden experiments. Combining phenotypic and genotypic data a Q_ST–F_ST‐like analysis was conducted to test for signatures of parasite mediated selection. We observed high variation within and among populations for phenotypic traits, but neither an indication of host local adaptation nor a cost of resistance. Infected populations have a higher gene diversity (Hs) than uninfected populations and Hs is strongly positively correlated with fitness. These results suggest a strong parasite effect on reducing population level inbreeding. We discuss how stochastic processes related to frequent extinction‐(re)colonization dynamics as well as host and parasite migration impede the evolution of resistance in the infected populations. We suggest that the genetic and phenotypic patterns of variation are a product of dynamic changes in the host gene pool caused by the interaction of colonization bottlenecks, inbreeding, immigration, hybrid vigor, rare host genotype advantage and parasitism. Our study highlights the effect of the parasite in ameliorating the negative fitness consequences caused by the high drift load in this metapopulation.     

The effects of the mating system on the evolution of migration in a spatially heterogeneous population

Summary Verbal explanations for the evolution of migration and dispersal often invoke inbreeding depression as an important force. Experimental work on plant populations indicates that while inbreeding depression may favor increased migration rates, adaptation to local environments may reduce the advantage to migrants. We formalize and test this hypothesis using a two-locus genetic model that incorporates lowered fitness in offspring produced by self-fertilization, and habitat differentiation. We also use the model to address questions about the general theory of genetic modifiers and the modifier reduction principle. We find that even under conditions when migration would increase the mean fitness of a population, migration may not be favored. This result is due to the associations that develop between genotypes at a locus subject to overdominant selection and at a neutral locus controlling the migration rate. Thus, it appears that, in this model, the forces of local adaptation, which favor a reduction in the migration rate, overwhelm those of inbreeding depression, which may favor dispersal.     

Nonrandom larval dispersal can steepen marine clines

Sharp and stable clinal variation is enigmatic when found in species with high gene flow. Classical population genetic models treat gene flow as a random homogenizing force countering local adaptation across habitat discontinuities. Under this view, dispersal over large spatial scales will lower the effectiveness of adaptation by natural selection at finer spatial scales. Thus, random gene flow will create a shallow phenotypic cline across an ecotone in response to a steep selection gradient. In sedentary marine species that disperse primarily as larvae, nonrandom dispersal patterns are expected due to coastal hydrodynamics. Surprisingly sharp phenotypic and genotypic clines have been documented in marine species with high gene flow. We are interested in the extent to which nonrandom dispersal could accentuate such clines. We model a linear species range in which populations have stable and uniform densities along a selection gradient; in contrast to random dispersal, convergent advection of larvae can amplify phenotypic differentiation if coupled with a semipermeable dispersal barrier in the convergence zone. The migration load caused by directional dispersal pushes the phenotypic mean away from the local trait optimum in downstream populations, that is, near the convergence zone. A dispersal barrier is possible as a result of colliding currents if the water and larvae are mostly displaced offshore, away from suitable settlement habitat. Disjunctions in a quantitative trait were enlarged in the convergence zone by faster current flows or a more complete dispersal barrier. With advection of larvae per generation one-third as far as the average dispersal distance by diffusion, convergence on a dispersal barrier with 40% permeability generated a trait disjunction across the convergence zone of two phenotypic standard deviations. Without directional dispersal, similar clines also developed across a habitat gap, where population density was low, or across dispersal barriers with less than 1% permeability. These findings suggest that the types of hydrographic phenomena often associated with marine transition zones can strongly affect the balance between gene flow and selection and generate surprisingly steep clines given the large-scale gene flow expected from larvae.     

Biased dispersal of Metrioptera bicolor,a wing dimorphic bush‐cricket

In the highly fragmented landscape of central Europe, dispersal is of particular importance as it determines the long‐term survival of animal populations. Dispersal not only secures the recolonization of patches where populations went extinct, it may also rescue small populations and thus prevent local extinction events. As dispersal involves different individual fitness costs, the decision to disperse should not be random but context‐dependent and often will be biased toward a certain group of individuals (e.g., sex‐ and wing morph‐biased dispersal). Although biased dispersal has far‐reaching consequences for animal populations, immediate studies of sex‐ and wing morph‐biased dispersal in orthopterans are very rare. Here, we used a combined approach of morphological and genetic analyses to investigate biased dispersal of Metrioptera bicolor, a wing dimorphic bush‐cricket. Our results clearly show wing morph‐biased dispersal for both sexes of M. bicolor. In addition, we found sex‐biased dispersal for macropterous individuals, but not for micropters. Both, morphological and genetic data, favor macropterous males as dispersal unit of this bush‐cricket species. To get an idea of the flight ability of M. bicolor, we compared our morphological data with that of Locusta migratoria and Schistocerca gregaria, which are very good flyers. Based on our morphological data, we suggest a good flight ability for macropters of M. bicolor, although flying individuals of this species are seldom observed.     

Density‐dependent fitness,not dispersal movements,drives temporal variation in spatial genetic structure in dark‐eyed juncos (Junco hyemalis)

Some studies have found that dispersal rates and distances increase with density, indicating that density‐dependent dispersal likely affects spatial genetic structure. In an 11‐year mark–recapture study on a passerine, the dark‐eyed junco, we tested whether density affected dispersal distance and/or fine‐scale spatial genetic structure. Contrary to expectations, we found no effect of predispersal density on dispersal distance or the proportion of locally produced juveniles returning to the population from which they hatched. However, even though density did not affect dispersal distance or natal return rates, we found that density still did affect spatial genetic structure. We found significant positive spatial genetic structure at low densities of (postdispersal) adults but not at high densities. In years with high postdispersal (adult) densities that also had high predispersal (juvenile) densities in the previous year, we found negative spatial genetic structure, indicating high levels of dispersal. We found that density also affected fitness of recruits, and fitness of immigrants, potentially linking these population parameters with the spatial genetic structure detected. Immigrants and recruits rarely nested in low postdispersal density years. In contrast, in years with high postdispersal density, recruits were common and immigrants had equal success to local birds, so novel genotypes diluted the gene pool and effectively eliminated positive spatial genetic structure. In relation to fine‐scale spatial genetic structure, fitness of immigrants and new recruits is poorly understood compared to dispersal movements, but we conclude that it can have implications for the spatial distribution of genotypes in populations.     

Landscape genetics of an early successional specialist in a disturbance‐prone environment

Species that specialize in disturbed habitats may have considerably different dispersal strategies than those adapted to more stable environments. However, little is known of the dispersal patterns and population structure of such species. This information is important for conservation because many postfire specialists are at risk from anthropogenic changes to natural disturbance regimes. We used microsatellite markers to assess the effect of landscape variation and recent disturbance history on dispersal by a small mammal species that occupies the early seral stage of vegetation regeneration in burnt environments. We predicted that a postfire specialist would be able to disperse over multiple habitat types (generalist) and not exhibit sex‐biased dispersal, as such strategies should enable effective colonization of spatially and temporally variable habitat. We found significant differentiation between sites that fitted an isolation‐by‐distance pattern and spatial autocorrelation of multilocus genotypes to a distance of 2–3 km. There was no consistent genetic evidence for sex‐biased dispersal. We tested the influence of different habitat‐ and fire‐specific landscape resistance scenarios on genetic distance between individuals and found a significant effect of fire. Our genetic data supported recently burned vegetation having greater conductance for gene flow than unburnt habitat, but variation in habitat quality between vegetation types and occupied patches had no effect on gene flow. Postfire specialists must evolve an effective dispersal ability to move over distances that would ensure access to early successional stage vegetation. Natural disturbance and natural heterogeneity may therefore not influence population genetic structure as negatively as expected.     

Divergent selection on locally adapted major histocompatibility complex immune genes experimentally proven in the field

Although crucial for the understanding of adaptive evolution, genetically resolved examples of local adaptation are rare. To maximize survival and reproduction in their local environment, hosts should resist their local parasites and pathogens. The major histocompatibility complex (MHC) with its key function in parasite resistance represents an ideal candidate to investigate parasite-mediated local adaptation. Using replicated field mesocosms, stocked with second-generation lab-bred three-spined stickleback hybrids of a lake and a river population, we show local adaptation of MHC genotypes to population-specific parasites, independently of the genetic background. Increased allele divergence of lake MHC genotypes allows lake fish to fight the broad range of lake parasites, whereas more specific river genotypes confer selective advantages against the less diverse river parasites. Hybrids with local MHC genotype gained more body weight and thus higher fitness than those with foreign MHC in either habitat, suggesting the evolutionary significance of locally adapted MHC genotypes.     

Fitness-dependent dispersal in metapopulations and its consequences for persistence and synchrony

1. We present a novel metapopulation model where dispersal is fitness dependent: the strength of migration from a site is dependent on the expected reproductive fitness of individuals there. Furthermore, individuals continue to migrate until they reach a suitable habitat where their expected fitness is above a threshold value.
2. Fitness-dependent dispersal has a very strong stabilizing effect on population dynamics, even when the intrinsic dynamics of populations in the absence of dispersal exhibit complex high-amplitude oscillations. This stabilizing effect is much stronger than that of the density-independent dispersal normally considered in metapopulation models.
3. Even when fitness-dependent dispersal does not stabilize the dynamics in a formal sense, it generally leads to simplification, with complex or even chaotic fluctuations being reduced to simple cycles.
4. This form of dispersal also has a strong tendency to synchronize local population dynamics across the spatial extent of the metapopulation.
5. These conclusions are robust to the addition of strong stochasticity in the migration threshold.     

Hybrid fitness, adaptation and evolutionary diversification: lessons learned from Louisiana Irises

Estimates of hybrid fitness have been used as either a platform for testing the potential role of natural hybridization in the evolution of species and species complexes or, alternatively, as a rationale for dismissing hybridization events as being of any evolutionary significance. From the time of Darwin's publication of The Origin, through the neo-Darwinian synthesis, to the present day, the observation of variability in hybrid fitness has remained a challenge for some models of speciation. Yet, Darwin and others have reported the elevated fitness of hybrid genotypes under certain environmental conditions. In modern scientific terminology, this observation reflects the fact that hybrid genotypes can demonstrate genotype × environment interactions. In the current review, we illustrate the development of one plant species complex, namely the Louisiana Irises, into a 'model system' for investigating hybrid fitness and the role of genetic exchange in adaptive evolution and diversification. In particular, we will argue that a multitude of approaches, involving both experimental and natural environments, and incorporating both manipulative analyses and surveys of natural populations, are necessary to adequately test for the evolutionary significance of introgressive hybridization. An appreciation of the variability of hybrid fitness leads to the conclusion that certain genetic signatures reflect adaptive evolution. Furthermore, tests of the frequency of allopatric versus sympatric/parapatric divergence (that is, divergence with ongoing gene flow) support hybrid genotypes as a mechanism of evolutionary diversification in numerous species complexes.     

Modeling a version of the good-genes hypothesis: female choice of locally adapted males

In addition to other potential causes, immigration into locally adapted populations has been suggested to maintain the genetic variance in fitness that is necessary for the good-genes hypothesis. Using population-genetic simulations, the present contribution shows that co-occurring local adaptation and migration can maintain genetic variance in fitness. In combination with an effect of local adaptation on condition and condition-dependent sexual signaling, such a scenario therefore enables the evolution and maintenance of female choice for locally adapted males. The simulations show that this mechanism can also work when choice is costly, and that the potential benefit is similar to that in other good-genes mechanisms. As a consequence of female choice in favor of locally adapted males, differentiation between populations can be expected to increase due to the decreased effective gene flow between populations. Based on such effects, choice of locally adapted males has the potential to play an important role in speciation and adaptive radiation.     

Is local adaptation in Mimulus guttatus caused by trade‐offs at individual loci?

Local adaptation is considered to be the result of fitness trade‐offs for particular phenotypes across different habitats. However, it is unclear whether such phenotypic trade‐offs exist at the level of individual genetic loci. Local adaptation could arise from trade‐offs of alternative alleles at individual loci or by complementary sets of loci with different fitness effects of alleles in one habitat but selective neutrality in the alternative habitat. To evaluate the genome‐wide basis of local adaptation, we performed a field‐based quantitative trait locus (QTL) mapping experiment on recombinant inbred lines (RILs) created from coastal perennial and inland annual races of the yellow monkeyflower (Mimulus guttatus) grown reciprocally in native parental habitats. Overall, we detected 19 QTLs affecting one or more of 16 traits measured in two environments, most of small effect. We identified 15 additional QTL effects at two previously identified candidate QTLs [DIV ERGENCE (DIV)]. Significant QTL by environment interactions were detected at the DIV loci, which was largely attributable to genotypic differences at a single field site. We found no detectable evidence for trade‐offs for any one component of fitness, although DIV2 showed a trade‐off involving different fitness traits between sites, suggesting that local adaptation is largely controlled by non‐overlapping loci. This is surprising for an outcrosser, implying that reduced gene flow prevents the evolution of individuals adapted to multiple environments. We also determined that native genotypes were not uniformly adaptive, possibly reflecting fixed mutational load in one of the populations.     

Coherent synthesis of genomic associations with phenotypes and home environments

Local adaptation is often studied via (i) multiple common garden experiments comparing performance of genotypes in different environments and (ii) sequencing genotypes from multiple locations and characterizing geographic patterns in allele frequency. Both approaches aim to characterize the same pattern (local adaptation), yet the complementary information from each has not yet been coherently integrated. Here, we develop a genome‐wide association model of genotype interactions with continuous environmental gradients (G × E), that is reaction norms. We present an approach to impute relative fitness, allowing us to coherently synthesize evidence from common garden and genome–environment associations. Our approach identifies loci exhibiting environmental clines where alleles are associated with higher fitness in home environments. Simulations show our approach can increase power to detect loci causing local adaptation. In a case study on Arabidopsis thaliana, most identified SNPs exhibited home allele advantage and fitness trade‐offs along climate gradients, suggesting selective gradients can maintain allelic clines. SNPs exhibiting G × E associations with fitness were enriched in genic regions, putative partial selective sweeps and associations with an adaptive phenotype (flowering time plasticity). We discuss extensions for situations where only adaptive phenotypes other than fitness are available. Many types of data may point towards the loci underlying G × E and local adaptation; coherent models of diverse data provide a principled basis for synthesis.     

The effect of frequency‐dependent selection on resistance and tolerance to herbivory

Negative frequency‐dependent selection (FDS), where rare genotypes are favoured by selection, is commonly invoked as a mechanism explaining the maintenance of genetic variation in plant defences. However, empirical tests of FDS in plant–herbivore interactions are lacking. We evaluated whether the oviposition preference of the specialist herbivore Lema daturaphila is a mechanism through which this herbivore can exert FDS on its host plant Datura stramonium. The frequency of contrasting resistance–tolerance strategies was manipulated within experimental plots, and the plants were exposed to a similar initial density of their natural herbivore. Herbivore oviposition preference and final density, as well as plant damage and seed production, were estimated. Overall, we found that the high‐resistant–low‐tolerant genotypes produced four times more seeds when common than when rare, whereas the high‐tolerant–low‐resistant genotypes achieved twice its fitness when rare than when common. This pattern was the result of differential oviposition preferences. In addition, when the high‐resistant–low‐tolerant genotypes were common, there was a three‐fold decreased in herbivore final density which led to a decrease in damage level by 10%. Thus, in our experiment positive FDS seems to favour resistance over tolerance. We discuss how this result would change if the extent of herbivore local adaptation and damage modify the pattern of positive FDS acting on resistance and the optimal allocation to tolerance.     

Artificial selection for increased dispersal results in lower fitness

Dispersal often covaries with other traits, and this covariation was shown to have a genetic basis. Here, we wanted to explore to what extent genetic constraints and correlational selection can explain patterns of covariation between dispersal and key life‐history traits—lifespan and reproduction. A prediction from the fitness‐associated dispersal hypothesis was that lower genetic quality is associated with higher dispersal propensity as driven by the benefits of genetic mixing. We wanted to contrast it with a prediction from a different model that individuals putting more emphasis on current rather than future reproduction disperse more, as they are expected to be more risk‐prone and exploratory. However, if dispersal has inherent costs, this will also result in a negative genetic correlation between higher rates of dispersal and some aspects of performance. To explore this issue, we used the dioecious nematode Caenorhabditis remanei and selected for increased and decreased dispersal propensity for 10 generations, followed by five generations of relaxed selection. Dispersal propensity responded to selection, and females from high‐dispersal lines dispersed more than females from low‐dispersal lines. Females selected for increased dispersal propensity produced fewer offspring and were more likely to die from matricide, which is associated with a low physiological condition in Caenorhabditis nematodes. There was no evidence for differences in age‐specific reproductive effort between high‐ and low‐dispersal females. Rather, reproductive output of high‐dispersal females was consistently reduced. We argue that our data provide support for the fitness‐associated dispersal hypothesis.     

Metapopulation dynamics of the bog fritillary butterfly: experimental changes in habitat quality induced negative density‐dependent dispersal

Habitat quality and habitat geometry are two crucial factors driving metapopulation dynamics. However, their intricacy has prevented so far a reliable test of their relative impact on local population dynamics and persistence. Here we report on a long‐term study in which we manipulated habitat quality within a butterfly metapopulation, whereas habitat geometry was kept constant. The treatment consisted in lowering the quality of certain habitat patches while others were kept untreated, using the same spatial design over years. The effect of the treatment on metapopulation dynamics was assessed by comparing residence probability and dispersal rates within the same habitat network on 11 and 6 independent butterfly generations before and after treatment, respectively. Results showed that the experimental decrease in habitat quality generated significantly higher emigration rates from treated patches. This increase was associated with a significant decrease in dispersal rates out of untreated patches, and a significant higher residence probability in these patches. The direct relation between lower habitat quality and higher dispersal propensity in treated patches was expected. However, the lower dispersal from untreated patches after treatment was opposite to the expectation of positive density dependent dispersal generally observed in butterflies. Such negative density‐dependent dispersal would allow a rapid fine‐tuning of dispersal rates to changes in habitat quality, particularly when the spatial autocorrelation of the environmental is low. Accordingly, dispersal would promote an ideal free distribution of individuals in the landscape according to their fitness expectation.