We investigate the multilinear epistatic model under mutation-limited directional selection. We confirm previous results that only directional epistasis, in which genes on average reinforce or diminish each other's effects, contribute to the initial evolution of mutational effects. Thus, either canalization or decanalization can occur under directional selection, depending on whether positive or negative epistasis is prevalent. We then focus on the evolution of the epistatic coefficients themselves. In the absence of higher-order epistasis, positive pairwise epistasis will tend to weaken relative to additive effects, while negative pairwise epistasis will tend to become strengthened. Positive third-order epistasis will counteract these effects, while negative third-order epistasis will reinforce them. More generally, gene interactions of all orders have an inherent tendency for negative changes under directional selection, which can only be modified by higher-order directional epistasis. We identify three types of nonadditive quasi-equilibrium architectures that, although not strictly stable, can be maintained for an extended time: (1) nondirectional epistatic architectures; (2) canalized architectures with strong epistasis; and (3) near-additive architectures in which additive effects keep increasing relative to epistasis.

We propose a model to analyze a quantitative trait under frequency-dependent disruptive selection. Selection on the trait is a combination of stabilizing selection and intraspecific competition, where competition is maximal between individuals with equal phenotypes. In addition, there is a density-dependent component induced by population regulation. The trait is determined additively by a number of biallelic loci, which can have different effects on the trait value. In contrast to most previous models, we assume that the allelic effects at the loci can evolve due to epistatic interactions with the genetic background. Using a modifier approach, we derive analytical results under the assumption of weak selection and constant population size, and we investigate the full model by numerical simulations. We find that frequency-dependent disruptive selection favors the evolution of a highly asymmetric genetic architecture, where most of the genetic variation is concentrated on a small number of loci. We show that the evolution of genetic architecture can be understood in terms of the ecological niches created by competition. The phenotypic distribution of a population with an adapted genetic architecture closely matches this niche structure. Thus, evolution of the genetic architecture seems to be a plausible way for populations to adapt to regimes of frequency-dependent disruptive selection. As such, it should be seen as a potential evolutionary pathway to discrete polymorphisms and as a potential alternative to other evolutionary responses, such as the evolution of sexual dimorphism or assortative mating.

Despite growing interest in the effects of landscape heterogeneity on genetic structuring, few tools are available to incorporate data on landscape composition into population genetic studies. Analyses of isolation by distance have typically either assumed spatial homogeneity for convenience or applied theoretically unjustified distance metrics to compensate for heterogeneity. Here I propose the isolation-by-resistance (IBR) model as an alternative for predicting equilibrium genetic structuring in complex landscapes. The model predicts a positive relationship between genetic differentiation and the resistance distance, a distance metric that exploits precise relationships between random walk times and effective resistances in electronic networks. As a predictor of genetic differentiation, the resistance distance is both more theoretically justified and more robust to spatial heterogeneity than Euclidean or least cost path-based distance measures. Moreover, the metric can be applied with a wide range of data inputs, including coarse-scale range maps, simple maps of habitat and nonhabitat within a species' range, or complex spatial datasets with habitats and barriers of differing qualities. The IBR model thus provides a flexible and efficient tool to account for habitat heterogeneity in studies of isolation by distance, improve understanding of how landscape characteristics affect genetic structuring, and predict genetic and evolutionary consequences of landscape change.

We examined patterns of variation and the extent of local adaptation in the interaction between the highly selfing annual weed Arabidopsis thaliana and its foliar bacterial pathogen Pseudomonas viridiflava by cross-infecting 23 bacterial isolates with 35 plant lines collected from six fallow or cultivated fields in the Midwest, USA. We used two measures of resistance and virulence: bacterial count in the leaf and symptom development four days after infection. We found variation in resistance in A. thaliana and virulence in P. viridiflava, as well as a significant difference in symptoms between two distinct genetic clades within P. viridiflava. We also observed that both resistance and plant development rate varied with field type of origin (cultivated or fallow), possibly through age-related resistance, a developmentally regulated general form of resistance. Finally, we did not observe local adaptation by host or pathogen, rather we found patterns of variation across populations that depended in part on P. viridiflava clade. These data suggest that the interaction between A. thaliana and P. viridiflava varies across space and is mediated by the selection regime of the host populations and differential performance of the P. viridiflava clades. This is one of a very limited number of studies examining a bacterial pathogen of wild plant populations and one of a few studies to examine patterns of variation in a plant-pathogen association that is not a highly specialized gene-for-gene interaction.

Habitat choice plays a critical role in the processes of host range evolution, specialization, and ecological speciation. Pea aphid, Acyrthosiphon pisum, populations from alfalfa and red clover in eastern North America are known to be genetically differentiated and show genetic preferences for the appropriate host plant. This species feeds on many more hosts, and here we report a study of the genetic variation in host plant preference within and between pea aphid populations collected from eight genera of host plants in southeastern England. Most host-associated populations show a strong, genetically based preference for the host plant from which they were collected. Only in one case (populations from Vicia and Trifolium) was there little difference in the plant preference spectrum between populations. All populations showed a significant secondary preference for the plant on which all the aphid lines were reared: broad bean, Vicia faba, previously suggested to be a “universal host” for pea aphids. Of the total genetic variance in host preference within our sample, 61% could be attributed to preference for the collection host plant and a further 9% to systematic differences in secondary preferences with the residual representing within-population genetic variation between clones. We discuss how a combination of host plant preference and mating on the host plant may promote local adaptation and possibly ecological speciation, and whether a widely accepted host could oppose speciation by mediating gene flow between different populations.

Hybridization is an important factor in the evolution of plants; however, many of the studies that have examined hybrid fitness have been concerned with the study of early generation hybrids. We examined the early- and late-generation fitness consequences of hybridization between two ecotypes of the selfing annual Avena barbata in a greenhouse environment as well as in two natural environments. Fitness of early generation (F₂) hybrids reflects both the action of dominance effects (hybrid vigor) and recombination (hybrid breakdown) and was not significantly different from that of the midparent in any environment. Fitness of later generation (F₆) recombinant inbred lines (RILS) derived from the cross reflect both the loss of early generation heterozygosity as well as disruption of any coadapted gene complexes present in the parents. In all environments, F₆ RILs were on average significantly less fit than the (equally homozygous) midparent, indicating hybrid breakdown through the disruption of epistatic interactions. However, the inbred F₆ were also less fit than the heterozygous F₂, indicating that hybrid vigor also occurs in A. barbata, and counteracts hybrid breakdown in early generation hybrids. Also, although the F₆ generation mean is lower than the midparent mean, there are individual genotypes within the F₆ generation that are capable of outperforming the parental ecotypes in the greenhouse. Fewer hybrid genotypes are capable of outperforming the parental ecotypes in the field. Overall, these experiments demonstrate how a single hybridization event can result in a number of outcomes including hybrid vigor, hybrid breakdown, and transgressive segregation, which interact to determine long-term hybrid fitness.

Reproductive isolation can evolve either as a by-product of divergent selection or through reinforcement. We used the Cape flora of South Africa, known for its high level of pollination specialization, as a model system to test the potential role of shifts in pollination system in the speciation process. Comparative analysis of 41 sister-species pairs (representing Geraniaceae, Iridaceae, and Orchidaceae) for which complete pollinator, edaphic, and distribution data are available showed that for sister species with overlapping distribution ranges, pollination system shifts are significantly associated with edaphic shifts. In contrast, there is no significant association between pollination system shifts and edaphic shifts for allopatric sister species. These results are interpreted as evidence for reinforcement.

Drosophila melanogaster has colonized temperate habitats on multiple continents over a historical time period, and many traits vary predictably with latitude. Despite considerable attention paid to clinal variation in Drosophila, the mechanisms generating such patterns in nature remain largely unidentified. In D. melanogaster, the expression of reproductive diapause can be induced by exposure to low temperatures and shortened photoperiods. Both diapause expression itself and the underlying genetic variance for diapause expression have widespread impacts on organismal fitness, and diapause incidence exhibits a 60% cline in frequency in the eastern United States. The major aim of this study was to evaluate whether the relative fitness of diapause and nondiapause genotypes varies predictably with environment. In experimental population cages in the laboratory, the frequency of genotypes that express diapause increased over time when flies were exposed to environmental stress, whereas the frequency of nondiapause genotypes increased when flies were cultured under benign control conditions. Other fitness traits correlated with the genetic variance for diapause expression (longevity, mortality rates, stress resistance, lipid content, preadult viability, fecundity profiles, and development time) also diverged between experimental treatments. Similarly, sampling of isofemale lines from natural populations revealed that the frequency of diapause incidence cycled over time in seasonal habitats: diapause expression was at high frequency following the winter season and subsequently declined throughout the summer months. In contrast, diapause expression was low and temporally homogeneous in isofemale line collections from human-associated urban habitats. These data suggest that genetic variation underlying the diapause-nondiapause dichotomy may be actively maintained by selection pressures that vary spatially and temporally in natural populations.

Maintaining an immune system is costly. Resource allocation to immunity should therefore trade off against other fitness components. Numerous studies have found phenotypic trade-offs after immune challenge, but few have investigated genetic correlations between immune components and other traits. Furthermore, empirical evidence for the costs of maintaining an innate immune system in the absence of challenges is rare. We examined responses to artificial selection on phenoloxidase (PO) activity, an important part of the insect innate defense against multicellular pathogens, in yellow dung flies, Scathophaga stercoraria (L.). After 15 generations of successful selection on PO activity, we measured reproductive characters: clutch size, egg hatching rates, adult emergence rates, and adult longevity. We found no evidence for negative genetic correlations between PO activity and reproduction. In fact, flies of lines selected for increased PO activity had larger first clutches, and flies of lines selected for decreased PO activity had smaller ones. However, flies from high-PO lines died earlier than did low-PO flies when no food was available; that is, there is a survival cost of running at high PO levels in the absence of challenge. Variation in resource acquisition or use may lead to positive genetic correlations between PO and fertility and fecundity. The negative correlation between PO and longevity under starvation may indicate that variation for resource acquisition is maintained by a cost of acquisition, based on a genotype-environment interaction.

Adaptive radiations consist of two intertwined processes, diversification of species and diversification of their ecological niches, but it is unclear whether there is a causal link between the processes. In phytophagous insects, ecological diversification mainly involves shifts in host-plant associations and in larval feeding habits (internal or external) on different plant parts, and several observations indicate that speciation is facilitated by host shifts. Data on host use in individual species suggest that internal feeders are less likely to colonize new hosts than external-feeding taxa and, consequently, increases in collective host ranges and species numbers should be slowed down in endophagous lineages. We tested these related hypotheses by using phylogenetic information to reconstruct the evolutionary history of larval resource use in the sawfly subfamily Nematinae, a group of 1000 plus species with a broad range of niches: the subfamily's combined host range includes over 20 plant families, and larvae may feed externally on leaves or needles, or internally, for example, in buds, fruits, leaves, or galls. The results show that: (1) Most internally feeding groups have evolved independently from external-feeding ancestors, but several distinct internal habits have appeared convergently multiple times; (2) Shifts among host taxa are clearly more common than changes in larval habits; (3) The majority of host switches have occurred among phylogenetically close plant groups, but many shifts are manifest among distantly related, ecologically proximate hosts; (4) Although external feeding characteristic of the common ancestor of Nematinae is associated with relatively high rates of host-shifting, internal feeders are very conservative in their host use; (5) In contrast, the effect of endophagy on speciation probabilities is more variable: net speciation rates are lowered in most internal-feeding groups, but a striking exception is found in species that induce galls on Salicaceae. The loose connection between collective host ranges and species diversity provides empirical support for theoretical models suggesting that speciation rates are a function of a complex interplay between “intrinsic” niche width and resource heterogeneity.

Male nutrient provisioning is widespread in insects. Females of some species use male-derived nutrients for increased longevity and reproductive output. Despite much research into the consequences of paternal nutrient investment for male and female fitness, the heritability, and therefore the potential of this trait to respond to selection, has rarely been examined. Males of several butterfly species provide the female with nutrients in the spermatophore at mating. Females of the green-veined white butterfly Pieris napi (Lepidoptera: Pieridae) use male donations both for developing eggs (resulting in higher lifetime fecundity of multiply mated females), but also for their somatic maintenance (increasing longevity). Using half-sib, father-son regression and full-sib analyses, I showed that paternal nutrient investment is heritable, both in terms of the absolute but also the relative size of the spermatophore (controlling for body size). Male size and spermatophore size were also genetically correlated. Furthermore, a separate study showed male genotype had a significant effect on female longevity and lifetime fecundity. In contrast, male genotype had no influence on the immediate egg-laying rate of females following mating, suggesting limited scope for male manipulation of immediate female oviposition. These results indicate that females may derive both direct (increased lifetime fecundity and longevity) and indirect (sons with greater reproductive success) fitness benefits from paternal nutrient donations in this species.

A unique reproductive system has previously been described in Wasmannia auropunctata, a widespread invasive ant species, where males are produced clonally, female queens are parthenogens, and female workers are produced sexually. However, these findings were mostly based on samples originating from only a limited part of the native range of the species in South America. We used microsatellite markers to uncover the reproductive modes displayed by a large number of nests collected in various invasive W. auropunctata populations introduced 40 years ago into New Caledonia, where the species now forms a single 450-km-long supercolony. Although the main reproduction system in New Caledonia remained clonality for both male and female reproductives, we found evidence of rare sexual reproduction events that led to the production of both new queen and male clonal lineages. All clonal lineages observed in New Caledonia potentially derived from sexual reproduction, recombination, and mutation events from a single female and a single male genotype. Hence, the male and female gene pools are not strictly separated in New Caledonia and the two sexes do not follow independent evolutionary trajectories. Our results also suggest genetic determination for both parthenogenesis and caste. We discuss the evolutionary implications of the emergence of sex in the clonal reproduction system of introduced populations of W. auropunctata.

The complex interactions between genetic diversity and evolution have important implications in many biological areas including conservation, speciation, and mate choice. A common way to study these interactions is to look at heterozygosity-fitness correlations (HFCs). Until recently, HFCs based on noncoding markers were believed to result primarily from global inbreeding effects. However, accumulating theoretical and empirical evidence shows that HFCs may often result from genes being linked to the markers used (local effect). Moreover, local effect HFCs could differ from global inbreeding effects in their direction and occurrence. Consequently, the investigation of the structure and consequences of local HFCs is emerging as a new important goal in evolutionary biology. In this study of a wild threespine stickleback (Gasterosteus aculeatus) population, we first tested the presence of significant positive or negative local effects of heterozygosity at 30 microsatellites loci on five fitness components: survival, mating success, territoriality, length, and body condition. Then, we evaluated the direction and shape of total impact of local HFCs, and estimated the magnitude of the impacts on fitness using regression coefficients and selection differentials. We found that multilocus heterozygosity was not a reliable estimator of individual inbreeding coefficient, which supported the relevance of single-locus based analyses. Highly significant and temporally stable local HFCs were observed. These were mainly positive, but negative effects of heterozygosity were also found. Strong and opposite effects of heterozygosity are probably present in many populations, but may be blurred in HFC analyses looking for global effects only. In this population, both negative and positive HFCs are apparently driving mate preference by females, which is likely to contribute to the maintenance of both additive and nonadditive genetic variance.

We use a combination of microsatellite marker analysis and mate-choice behavior experiments to assess patterns of reproductive isolation of the túngara frog Physalaemus pustulosus along a 550-km transect of 25 populations in Costa Rica and Panama. Earlier studies using allozymes and mitochondrial DNA defined two genetic groups of túngara frogs, one ranging from Mexico to northern Costa Rica (northern group), the second ranging from Panama to northern South America (southern group). Our more fine-scale survey also shows that the northern and southern túngara frogs are genetically different and geographically separated by a gap in the distribution in central Pacific Costa Rica. Genetic differences among populations are highly correlated with geographic distances. Temporal call parameters differed among populations as well as between genetic groups. Differences in calls were explained better by geographic distance than by genetic distance. Phonotaxis experiments showed that females preferred calls of males from their own populations over calls of males from other populations in about two-thirds to three-fourths of the contrasts tested. In mating experiments, females and males from the same group and females from the north with males from the south produced nests and tadpoles. In contrast, females from the south did not produce nests or tadpoles with males from the north. Thus, northern and southern túngara frogs have diverged both genetically and bioacoustically. There is evidence for some prezygotic isolation due to differences in mate recognition and fertilization success, but such isolation is hardly complete. Our results support the general observation that significant differences in sexual signals are often not correlated with strong genetic differentiation.

Carotenoid-based colors are thought to play an important signaling role in many animal taxa. However, little is known about evolutionary changes in carotenoid coloration, especially among closely related species. We used a phylogenetic perspective to examine carotenoid color changes within New World orioles (genus Icterus). Oriole color was quantitatively measured using reflectance spectrometry. We found continuous variation from short- to long-wavelength carotenoid colors in extant orioles—perceived by humans as ranging from yellow to scarlet—suggesting that these carotenoid-based colors have evolved as a continuous character. Ancestral state reconstruction suggests that short- and long-wavelength carotenoid colors have evolved independently multiple times, likely from a middle-wavelength ancestor. Although color showed considerable lability, we found a significant amount of phylogenetic signal across the entire genus. This implies that while labile, the colors of closely related taxa tended to resemble each other more than would be expected due to chance. To our knowledge, this is the first study to use quantitative character states derived from reflectance spectra in ancestral state reconstruction. Reflectance spectra provide an unbiased quantitative description of color that allowed us to detect subtle changes among closely related taxa. Using these quantitative methods to score and reconstruct color changes among closely related taxa provides a better understanding of how elaborate animal colors evolve.

Sibling competition was proposed as an important selective agent in the evolution of growth and development. Brood parasitism by the brown-headed cowbird (Molothrus ater) intensifies sibling competition in the nests of its hosts by increasing host chick mortality and exposing them to a genetically unrelated nestmate. Intranest sibling competition for resources supplied by parents is size dependent. Thus, it should select for high development rates and short nestling periods, which would alleviate negative impacts of brood parasitic chicks on host young. I tested these predictions on 134 North American passerines by comparative analyses. After controlling for covariates and phylogeny, I showed that high parasitism rate was associated with higher nestling growth rate, lower mass at fledging, and shorter nestling periods. These effects were most pronounced in species in which sibling competition is most intense (i.e., weighing over about 30 g). When species were categorized as nonhosts versus old hosts (parasitized for thousands of years) versus new hosts (parasitized the last 100–200 years), there was a clear effect of this parasitism category on growth strategies. Nestling growth rate was the most evolutionarily flexible trait, followed by mass at fledging and nestling period duration. Adjustments during incubation (incubation period length, egg volume) were less pronounced and generally disappeared after controlling for phylogeny. I show that sibling competition caused by brood parasites can have strong effects on the evolution of host growth strategies and that the evolution of developmental traits can take place very rapidly. Human alteration of habitats causing spread of brood parasites to new areas thus cascades into affecting the evolution of life-history traits in host species.

Within a group of organisms, some morphologies are more readily generated than others due to internal developmental constraints. Such constraints can channel evolutionary changes into directions corresponding to the greatest intraspecific variation. Long term evolutionary outputs, however, depend on the stability of these intraspecific patterns of variation over time and from the interplay between internal constraints and selective regimes. To address these questions, the relationship between the structure of phenotypic variance covariance matrices and direction of morphological evolution was investigated using teeth of fossil rodents. One lineage considered here leads to Stephanomys, a highly specialized genus characterized by a dental pattern supposedly favoring grass eating. Stephanomys evolved in the context of directional selection related to the climatic trend of global cooling causing an increasing proportion of grasslands in southwestern Europe. The initial divergence (up to ∼6.5 mya) was channeled along the direction of greatest intraspecific variation, whereas after 6.5 mya, morphological evolution departed from the direction favored by internal constraints. This departure from the “lines of least resistance” was likely the consequence of an environmental degradation causing a selective gradient strong enough to overwhelm the constraints to phenotypic evolution. However, in a context of stabilizing selection, these constraints actually channel evolution, as exemplified by the lineage of Apodemus. This lineage retained a primitive diet and dental pattern over the last 10 myr. Limited morphological changes occurred nevertheless in accordance with the main patterns of intraspecific variation. The importance of these lines of least resistance directing long-term morphological evolution may explain parallel evolution of some dental patterns in murine evolution.

The species and races of the shrews of the Sorex araneus group exhibit a broad range of chromosomal polymorphisms. European taxa of this group are parapatric and form contact or hybrid zones that span an extraordinary variety of situations, ranging from absolute genetic isolation to almost free gene flow. This variety seems to depend for a large part on the chromosome composition of populations, which are primarily differentiated by various Robertsonian fusions of a subset of acrocentric chromosomes. Previous studies suggested that chromosomal rearrangements play a causative role in the speciation process. In such models, gene flow should be more restricted for markers on chromosomes involved in rearrangements than on chromosomes common in both parent species. In the present study, we address the possibility of such differential gene flow in the context of two genetically very similar but karyotypically different hybrid zones between species of the S. araneus group using microsatellite loci mapped to the chromosome arm level. Interspecific genetic structure across rearranged chromosomes was in general larger than across common chromosomes. However, the difference between the two classes of chromosomes was only significant in the hybrid zone where the complexity of hybrids is expected to be larger. These differences did not distinguish populations within species. Therefore, the rearranged chromosomes appear to affect the reproductive barrier between karyotypic species, although the strength of this effect depends on the complexity of the hybrids produced.

The island rule is the phenomenon of the miniaturization of large animals and the gigantism of small animals on islands, with mammals providing the classic case studies. Several explanations for this pattern have been suggested, and departures from the predictions of this rule are common among mammals of differing body size, trophic habits, and phylogenetic affinities. Here we offer a new explanation for the evolution of body size of large insular mammals, using evidence from both living and fossil island faunal assemblages. We demonstrate that the extent of dwarfism in ungulates depends on the existence of competitors and, to a lesser extent, on the presence of predators. In contrast, competition and predation have little or no effect on insular carnivore body size, which is influenced by the nature of the resource base. We suggest dwarfism in large herbivores is an outcome of the fitness increase resulting from the acceleration of reproduction in low-mortality environments. Carnivore size is dependent on the abundance and size of their prey. Size evolution of large mammals in different trophic levels has different underlying mechanisms, resulting in different patterns. Absolute body size may be only an indirect predictor of size evolution, with ecological interactions playing a major role.

Studies of character evolution often assume that a phylogeny's shape is determined independently of the characters, which then evolve as mere passengers along the tree's branches. However, if the characters help shape the tree, but this is not considered, biased inferences can result. Simulations of asymmetrical speciation (i.e., one character state conferring a higher rate of speciation than another) result in data that are interpreted to show a higher rate of change toward the diversification-enhancing state, even though the rates to and from this state were in fact equal. Conversely, simulations of asymmetrical character change yield data that could be misinterpreted as showing asymmetrical rates of speciation. Studies of biased diversification and biased character change need to be unified by joint models and estimation methods, although how successfully the two processes can be teased apart remains to be seen.