Evolution

Populations may be rescued from extinction via sufficiently rapid adaptive evolution. Evolution is faster with stronger selection but this may come with a demographic cost, creating opposing effects on evolutionary rescue. The outcome of this trade-off determines the optimal strategy for avoiding herbicide/drug resistance evolution. Here we examine the effect of stronger selection, and the associated demographic change, across four models of evolutionary rescue. We find that stronger selection cannot facilitate rescue in two quite different population genetic models unless it increases the absolute fitness of the rescue homozygote. Similarly, in a quantitative genetic model of rescue, where stronger selection leaves maximum fitness unchanged, stronger selection cannot facilitate rescue despite elevating the rate of evolution. We also explore a quantitative genetic model of rescue in a gradually changing environment, generalizing the finding that an intermediate selection strength maximizes survival at steady-state to a wider class of fitness functions.

Adaptive radiation results in part from ecological opportunity in a new environment, but it is unclear how pre-existing constraints in the founding population may limit this process. Genetic lines of least resistance, and by proxy morphological variance, are known to limit adaptive divergence, but ecological variance is rarely investigated. Here we test whether ecological or morphological lines of least resistance in generalist populations may have constrained the directions of species divergence in two independent Caribbean adaptive radiations of Cyprinodon pupfishes. We find almost universal congruence between the major multivariate dimensions of intraspecific craniofacial and dietary variance within generalist populations and the major axes of interspecific divergence within each adaptive radiation. This is surprising given that we document unique trophic specialists within each radiation, including a bivalve-specialist, zooplanktivore, molluscivore/ostracod-specialist, and scale-eating specialist, while nearly all generalist populations were observed to feed rarely on these same resources. We conclude that pre-existing genetic constraints within each founding generalist population, resulting in dimensions of greater ecological and morphological variance, may partially constrain and predict the directions of species divergence and dietary specialization during adaptive radiation. We also provide a new framework for examining ecological lines of least resistance.

Hybridization between distinct populations injects genetic variation, which can bring fitness benefits. However, these benefits often appear as F1 heterosis, and might not persist into later generations; especially since, as emphasized by classical theories, heterosis can be caused in several different ways. Here, we study the long-term outcomes of hybridization, using a model that allows us to tune several properties of the genetic variation, including the strength and architecture of heterosis, thereby unifying the classical theories. Results suggest that long-term outcomes depend mainly on the variance in epistasis, which determines the ruggedness of the fitness landscape, but without affecting the heterosis. Together, results suggest that the study of heterosis may tell us relatively little about the long-term outcomes of hybridization, and that hybridization might bring benefits more often than has been assumed.

Coevolutionary negative frequency dependent selection has been hypothesized to maintain genetic variation in host and parasites. Despite the extensive literature pertaining to host-parasite coevolution, the effect of matching-alleles (MAM) coevolution on the maintenance of genetic variation has not been explicitly modelled in a finite population. The dynamics of the MAM in an infinite population, in fact, suggests that genetic variation in these coevolving populations behaves neutrally. We find that while this is largely true in finite populations two additional phenomena arise. The first of these effects is that of coevolutionary natural selection on stochastic perturbations in host and pathogen allele frequencies. While this may increase or decrease genetic variation, depending on the parameter conditions, the net effect is small relative to that of the second phenomena. Following fixation in the pathogen, the MAM becomes one of directional selection, which in turn rapidly erodes genetic variation in the host. Hence, rather than maintain it, we find that, on average, matching-alleles coevolution depletes genetic variation.

Because natural selection must optimise multiple traits at once, previous work has suggested that phenotypic dimensionality can substantially worsen the equilibrium fitness defect of a population relative to the phenotypic optimum. However, it remains unclear how conclusions drawn from classical theoretical phenotype-fitness maps extend to models grounded in explicit biological mechanisms. Here we introduce weakest-link epistasis (WLE), a framework in which fitness is determined by the least-fit phenotypic component, an extreme form of diminishing returns epistasis. We show that in this framework, increasing dimensionality amplifies the load in a manner comparable to, but surprisingly not more than, Fishers geometric model (FGM). Building on this similarity, we demonstrate why genetic load is often invariant across different rules for combining trait-specific fitness components into an overall organismal fitness. We explore these ideas by considering the family of models where the organismal fitness is determined based on the{ell} p-norm of the vector of trait-specific fitness defects, a framework that includes both FGM and WLE, but also captures a continuum of genetic architectures, ranging from generalist to specialist regimes. Altogether, our approach proposes a new perspective on the geometry of adaptive landscapes, and may help provide quantitative insight into the cost of complexity.

Theory predicts that hybrid incompatibilities accumulate faster than linearly with genetic divergence, a phenomenon known as the snowball effect. While this prediction is mathematically robust under simplifying assumptions, accumulating evidence suggests that the structure of gene interaction networks can alter both the rate and organization of incompatibility evolution. Here, we extend classic DMI models with a network approach, equating the assumptions of the Orr model with a complete graph of gene interactions. We simulate the evolution of hybrid incompatibilities under different gene interaction networks and evaluate the effects of network density, topology, and substitution model. We find that network density strongly governs the rate of DMI accumulation, particularly under models permitting multiple substitutions per locus, while network topology shapes the agglomeration of incompatibilities into large, connected clusters. Substitution rate heterogeneity, especially when anti-correlated with node degree, further suppresses both accumulation and clustering. These results highlight that while the snowball effect remains qualitatively valid, the structure and evolution of the incompatibility network exhibit nontrivial departures from previous expectations, with implications for observable quantities in empirical systems. Our findings underscore the importance of incorporating genomic architecture and network constraints into models of speciation.

Mutation rate plays an important role in adaptive evolution due to its effect on the rate of appearance of both beneficial and deleterious mutations and is therefore subject to second-order selection. The mutation rate varies between and within species and populations, increases under some stresses, and can be modified by mutator and anti-mutator alleles. It may also vary among genetically identical individuals: empirical evidence from bacteria suggests that the mutation rate can be affected by translation errors and expression noise in various proteins. Importantly, this non-genetic variation may be heritable via a transgenerational epigenetic mode of inheritance, giving rise to mutator phenotype switching that is independent from mutator alleles. Here we investigate mathematically how the rate of adaptive evolution on rugged, complex fitness landscapes is affected by the rate of mutation rate phenotype switching. Motivated by recent experimental results of mutation rate variation, we model an asexual population with two mutation rate phenotypes, non-mutator and mutator. An offspring may switch from its parental phenotype to the other phenotype. Thus, the mutation rate can be interpreted as a genetically inherited trait when the switching rate is low, as an epigenetically inherited trait when the switching rate is intermediate, or as a randomly determined trait when the switching rate is high. We find that intermediate switching rates maximize the rate of adaptation on rugged fitness landscapes. This is because an intermediate switching rate can maintain within the same individuals both a mutator phenotype and pre-existing mutations, a combination that facilitates the crossing of fitness valleys. Further, intermediate switching rates allow the population to quickly revert to a low mutation rate after adaptation is achieved, avoiding the accumulation of deleterious mutations linked to mutator alleles. Our results rationalize recently observed noise in the expression of proteins that affect the mutation rate and suggest that non-genetic inheritance of this phenotype may facilitate evolutionary adaptive processes.

Evolutionary biologists have long been intrigued by the factors that sustain genetic and phenotypic variation within and among natural populations. Polymorphisms underlying components of floral display - such as floral color, size, and shape - are uniquely of interest since variation in these traits impact pollinator attraction, rates of visitation, and pollinator efficiency, which ultimately influence patterns of plant reproduction and therefore fitness. We leverage existing floral variation present within and among populations of Hesperis matronalis (Dames Rocket) to disentangle the relative influence of natural selection and genetic drift in shaping floral trait variation. We employ a multi-tiered approach: we determine if variation in floral traits (color, size, and petal shape) is influenced by geography or environmental variables such as temperature and precipitation, we evaluate whether selection underlies trait variation by comparing phenotypic divergence (PST) with neutral genetic structure (FST), and we perform a Lande-Arnold selection analysis to explore the relationship between fitness and floral trait variation within natural populations. We find that selection underlies the divergence of floral color, floral size, and petal width among H. matronalis populations, with PST > FST for each trait. We find no indication, however, that variation in size in this species is influenced by the environment, but some evidence that variation in floral color and petal shape may be influenced by temperature. Finally, selection analyses of contemporary populations indicate divergent selection affecting combinations of color, petal shape, and plant size. These results suggest that the variation in floral shape in this species may be maintained due to environmental pressures, whereas floral color is influenced by pollinator visibility and the presence of different pollinator groups.

The divergence of reproductive traits frequently underpins the evolution of reproductive isolation. One of the most enduring puzzles on this subject concerns the variability in egg coloration among species of tinamou (Tinamidae), endemic to neotropics. Here we investigated the hypothesis that tinamou egg coloration is a mating signal and its diversification was driven by reinforcement. For most tinamou species, the male guards the nest that is sequentially visited and laid eggs in by multiple females. The colorations of the existing eggs in the nest could signal mate quality and species identities to the upcoming females, preventing costly hybridization, thus were selected to diverge among species (Mating Signal Character Displacement Hypothesis). If so, two predictions should follow: (1) egg colors should coevolve with known mating signals as the tinamou lineages diverged; (2) species that partition similar ecoregions should display different egg colors. The tinamou songs are important mating signals and are highly divergent among species. We found that the egg luminance was significantly associated with the first principal component of the song variables, which supports prediction (1). In addition, we found support for (2): tinamou species that co-partition ecoregions tend to display different egg colors, controlling for song variation. Egg color and songs could be multimodal mating signals that are divergently selected as different tinamou species diverged. Mating signal evolution could be opportunistic and even exploit post-mating trait as premating signals that undergo character displacement at sympatry.

Changes in the structure and relative size of the brain are thought to be key transformations in the origins and continued evolution of birds, reflecting innovations and diversity of neurosensory and cognitive capabilities. However, these neuro-anatomical and functional changes do not occur in isolation, being accompanied by a host of other derived morphological characteristics associated with the evolution of flight. In the avian head alone, these include the evolution of a toothless beak, increase in relative eye size, and reduction and restructuring of jaw muscles. Several hypothesized developmental trade-offs have been proposed to explain the interrelationships among the hard and soft tissues of the head. How these developmental patterns translate into evolutionary trade-offs in other cranial traits is poorly understood, despite brain shape evolution being well documented in birds. Here, we use two-block partial least squares analyses and Ornstein-Uhlenbeck models of adaptive trait evolution to explore the phenotypic evolution of hard and soft cranial tissues and test hypotheses of correlated trait evolution. In pairwise analyses we found that all traits (endocast shape, neurocranium shape, rostrum shape, jaw muscle shape and residual orbit diameter) are significantly correlated. We found strongest support for a modular hypothesis of trait evolution across the whole head, with the rostrum and jaw muscles forming one module and brain, neurocranium, and eye forming the other. Within modules, traits are tightly integrated, but integration is relaxed between modules, allowing them to develop and evolve with a degree of independence. Together, these results highlight the integrated nature of the avian head and reveal that rather than being driven overwhelmingly by selection on a single trait, the shape of the avian head is a result of multiple interactions among hard and soft tissue traits.

Many organisms undergo ontogeny, whereby individuals change in state (e.g. in size, morphology, or condition) as they age. Understanding the evolution of traits influencing ontogeny is challenging because their fitness effects unfold across an individuals lifetime and may differ between classes such as sexes. Here, we analyse selection on non-plastic traits (e.g., fixed resource allocation strategies) that determine the development of dynamical states throughout life (e.g., body size), with consequences for fecundity and survival in age- and class-structured populations. Using invasion analysis, we derive expressions for directional and quadratic selection that decompose into age- and class-specific components. This allows us to identify convergence stable trait values, assess whether they are uninvadable or potentially experience evolutionary branching, and pinpoint the age and class pathways through which correlational and disruptive selection act. Applying our results to a model of growth under size-mediated sexual selection, we show how selection on growth rates distributes across ages and sexes, and how its relative strength depends on genetic correlations, individual ploidy and life-history. We also show how sex-specific developmental trade-offs and constraints can generate disruptive selection on male growth and favour the evolution of alternative male life histories. More broadly, our results highlight how adaptation is mediated by the interaction of development and demography, and provide tools to investigate how conflicts across ages and classes may influence senescence, sexual dimorphism, and the diversification of ontogenetic strategies. Author SummaryHow does natural selection shape traits that influence how organisms develop through their lifetime? This question is difficult because the effects of such traits unfold dynamically as individuals change in states such as size or condition, which in turn influence survival and reproduction throughout life. We formulate a mathematical model that follows these causal links and characterises the gradual evolution of non-plastic traits affecting development. Our results apply to populations with different classes, such as males and females, that follow distinct developmental and demographic paths. Our model reveals when selection is stabilizing, holding populations fixed for an optimal trait value, and when it is disruptive, favouring trait diversification and the emergence of alternative life-history strategies. As an illustration, we analyse the evolution of sex-specific growth under size-based sexual selection. We show how the evolution of male and female growth rates depends on genetics, such as individual ploidy or genetic correlation between traits, but also on life history, which determines the relative strength of selection acting on trait expression at different ages. We further show how developmental trade-offs and demographic differences can generate disruptive selection on male growth, leading to the coexistence of fast-growing short-lived males and slow-growing long-lived ones. More broadly, our results link development, demography, and adaptation, and show how their interplay shapes biological diversity both within and between populations.

1Phenotypic evolution in sympatric species can be strongly impacted by species interactions, either mutualistic or antagonistic, which may favour local phenotypic divergence or convergence. Interspecific sexual interactions between sympatric species has been shown to favour phenotypic divergence of traits used as sexual cues for example. Those traits may also be involved in local adaptation or in other types of species interactions resulting in complex evolution of traits shared by sympatric species. Here we focus on mimicry and study how reproductive interference may impair phenotypic convergence between species with various levels of defences. We use a deterministic model assuming two sympatric species and where individuals can display two different warning colour patterns. This eco-evolutionary model explores how ecological interactions shape phenotypic evolution within sympatric species. We investigate the effect of (1) the opposing density-dependent selections exerted on colour patterns by predation and reproductive behaviour, and (2) the impact of relative species and phenotype abundances on the fitness costs faced by each individual depending on their species and phenotype. Our model shows that reproductive interference may limit the convergent effect of mimetic interactions and may promote phenotypic divergence between Mullerian mimics. The divergent and convergent evolution of traits also strongly depends on the relative species and phenotype abundances and levels of trophic competition, highlighting how the eco-evolutionary feedbacks between phenotypic evolution and species abundances may result in strikingly different evolutionary routes.

Evolutionary biologists have long been interested in understanding the mechanisms underlying Haldanes rule. The explanatory theories of dominance and faster-X, which are based on recessive alleles being expressed in the heterogametic sex, have been proposed as common mechanisms. These mechanisms predict that greater hemizygosity leads to both faster evolution and greater expression of intrinsic postzygotic isolation. Under these mechanisms, haplodiploids should evolve and express intrinsic postzygotic isolation faster than diploids because the entire genome is analogous to a sex chromosome. Here, we measure sterility and inviability in hybrids between Neodiprion pinetum and N. lecontei, a pair of haplodiplopids that differ morphologically, behaviorally, and genetically. We compare the observed isolation to that expected from published estimates of isolation in diploids at comparable levels of genetic divergence. We find that both male and female hybrids are viable and fertile, which is less isolation than expected. We then discuss several potential explanations for this surprising lack of isolation, including alternative mechanisms for Haldanes rule and a frequently overlooked quirk of haplodiploid genetics that may slow the emergence of complete intrinsic postzygotic isolation in hybrid males. Finally, we describe how haplodiploids, an underutilized resource, can be used to differentiate between mechanisms of Haldanes rule.

Hosts can use avoidance (e.g., behavior) to reduce their contact rates with pathogens; after contact, they can use resistance (e.g., immunity) to reduce the establishment and proliferation of an infection. Because both defenses preserve host fitness and reduce pathogen fitness, we expect that their epidemiological and evolutionary effects will be inter-dependent. This study used a two-locus model to understand the evolution of allelic associations (i.e., linkage disequilibrium) between genes determining levels of avoidance and resistance in the presence of an infectious disease or parasite. We found that polymorphism in both avoidance and resistance was possible, but only for a limited range of parameter values of avoidance and resistance; at equilibrium, avoidance and resistance alleles were negatively associated in these polymorphic populations. However, most commonly, whichever defense was more effective and less costly went to fixation. This result suggests that avoidance and resistance may be more likely to covary negatively across than within populations.

Hybrid zones have been viewed as an opportunity to see speciation in action. When hybrid zones are replicated, it is assumed that if the same genetic incompatibilities are maintaining reproductive isolation across all instances of secondary contact, those incompatibilities should be identifiable by consistent patterns in the genome. In contrast, changes in allele frequencies due to genetic drift should be idiosyncratic for each hybrid zone. To test this assumption, we simulated 20 replicates of each of 12 hybrid zone scenarios with varied genetic incompatibilities, rates of migration, selection and different starting population size ratios of parental species. We found remarkable variability in the outcomes of hybridization in replicate hybrid zones, particularly with Bateson-Dobzhansky-Muller incompatibilities and strong selection. We found substantial differences among replicates in the overall genomic composition of individuals, including admixture proportions, inter-specific ancestry complement, and number of ancestry junctions. Additionally, we found substantial variation in genomic clines among replicates at focal loci, regardless of locus-specific selection. We conclude that processes other than selection are responsible for some consistent outcomes of hybridization, whereas selection on incompatibilities can lead to genomically widespread and highly variable outcomes. We highlight the challenge of mapping between pattern and process in hybrid zones and call attention to how selection against incompatibilities will commonly lead to variable outcomes. We hope that this study informs future research on replicate hybrid zones and encourages further development of statistical techniques, theoretical models, and exploration of additional axes of variation to understand reproductive isolation.

A wide variety of animals engage in same-sex sexual behavior (SSB). Some instances of SSB appear to arise from individuals mating indiscriminately without regard to the sex of their partners. Theory suggests that indiscriminate mating strategies can be selected if sexes are not completely distinguishable--that is, without strong signals of sexual identity. We model the coevolution of sexual signaling and discrimination using the adaptive dynamics approach to long-term trait evolution. Our model relaxes two major assumptions from past phenomenological theory. First, we consider a mechanistic mating process with tunable encounter rates and mating costs. Second, we vary reproductive investment roles to allow for better correspondence with empirical examples of SSB. We confirm that coevolution of sexual signaling and discrimination can lead to two distinct equilibria: one with no sexual signals and indiscriminate mating; the other with perfect signaling and exclusively different-sex sexual behavior. Selection for indiscriminate mating is strongest at very low or very high encounter rates. Mating costs lead to selection for reduced overall mating rate, but this can lead to either more or less SSB, depending on how this reduction is achieved. Overall, our model highlights the importance of ecologically relevant parameters for the evolution of SSB.

Cyclic parthenogenesis is a form of occasional sexual reproduction. Here we compare the evolvability of cyclic parthenogens and obligate sexuals in adaptive scenarios that entail the crossing of a fitness valley. With individual-based models, we find that cyclic parthenogens tend to cross fitness valleys faster, although that advantage is dramatically reduced by strong genetic correlations affecting multivariate phenotypes. We also find that, when epistasis occurs, cyclical parthenogenesis reduces the evolvability of mutational variances and covariances. In our analysis, such reductions are overwhelmed by the increase in evolvability due to the periodic storage and release of cryptic genetic diversity. Nonetheless, the relative recalcitrance against selection on mutational variances and covariances in cyclic parthenogens could be an evolutionary cost that helps explain the paradoxical rarity of cyclic parthenogenesis.

Inclusive fitness theory has transformed the study of adaptive evolution since 1964, contributing to significant empirical findings. However, its status as a theory has been challenged by the proposals of several alternative frameworks. Those challenges have been countered by analyses that use the Price equation and the regression method. The Price equation is a universal description of evolutionary change, and the partitioning of the Price equation using the regression method immediately yields Hamiltons rule, which embodies the main tenets of inclusive fitness. Hamiltons rule captures the intensity and direction of selection acting on social behaviour and its underlying causal structure. Recent work, however, has suggested that there is an anomaly in this approach: in some cases, the regression method fails to estimate the correct values of the variables in Hamiltons rule and the causal structure of the behaviour. Here, I address this apparent anomaly. I argue that the failure of the simple regression method occurs because social players vary in baseline fecundity. I reformulate the Price equation and regression method to recover Hamiltons rule and I show that the method correctly estimates its key variables. I show that games where baseline fecundity varies among individuals represent a more general set of games that unfold in class-structured populations. This framework supports the robustness and validity of inclusive fitness.

Phenotypes are partly shaped by the environment, which can impact both short-term adaptation and long-term evolution. In dioecious species, the two sexes may exhibit different degrees of phenotypic plasticity and theoretical models indicate that such differences may confer an adaptive advantage when the population is subject to directional selection, either because of a systematically varying environment or of a load of deleterious mutations. The effect stems from the fundamental asymmetry between the two sexes: female fertility is more limited than male fertility. Whether this asymmetry is sufficient for sexual dimorphism in phenotypic plasticity to evolve is, however, not obvious. Here, we show that even in conditions where it provides an adaptive advantage, dimorphic phenotypic plasticity may be evolutionarily unstable due to sexual selection. This is the case, in particular, for panmictic populations where mating partners are formed at random. However, we show that the effects of sexual selection can be counteracted when mating occurs within groups of related individuals. Under this condition, sexual dimorphism in phenotypic plasticity can not only evolve but offset the twofold cost of males. We demonstrate these points with a simple mathematical model through a combination of analytical and numerical results.

Phenotypic trade-offs, predicted to occur due to resource constraints, are not commonly observed. This discrepancy is often explained by the Y-model, which shows that greater variation in resource acquisition (CVA) than variation in allocation strategy (CVX) masks among-individual phenotypic trade-offs. However, the Y-model is a heuristic rather than a quantitative tool with testable predictions. Additionally, the mechanisms modulating trade-offs in the model remain unclear. Here, we simulate different parameters of the Y-model to understand their influence. We find that CVA/(CVA + CVX) accurately predicts the direction and strength of phenotypic correlations. Contrary to common assumptions, the mean resource acquired by a population does not influence trade-offs; instead, the mean allocation strategy of a population does. Importantly, within-individual dependence of allocation on acquisition exacerbates the influence of mean allocation. This causes greater sensitivity of phenotypic correlations to changes in CVA or CVX, compared to independent resource acquisition and allocation strategies. We generate novel, testable hypotheses about trade-offs in the context of dietary restriction, plasticity, polymorphism, and pace-of-life, and validate our model with empirical data to demonstrate its robustness. By systematically partitioning the influence of means, variances, and covariance, in the acquisition-allocation Y-model, we provide a simple, generalisable, quantitative synthesis for understanding phenotypic trade-offs.