Resurrected seeds from herbarium specimens reveal rapid evolution of drought resistance in a selfing annual.

PREMISE: Increased aridity and drought associated with climate change are exerting unprecedented selection pressures on plant populations. Whether populations can rapidly adapt, and which life history traits might confer increased fitness under drought, remain outstanding questions. METHODS: We utilized a resurrection ecology approach, leveraging dormant seeds from herbarium collections to assess whether populations of Plantago patagonica from the semi-arid Colorado Plateau have rapidly evolved in response to approximately ten years of intense drought in the region. We quantified multiple traits associated with drought escape and drought resistance and assessed the survival of ancestors and descendants under simulated drought. RESULTS: Descendant populations displayed a significant shift in resource allocation, in which they invested less in reproductive tissues and relatively more in both above- and below-ground vegetative tissues. Plants with greater leaf biomass survived longer under terminal drought; moreover, even after accounting for the effect of increased leaf biomass, descendant seedlings survived drought longer than their ancestors. CONCLUSIONS: Our results document rapid adaptive evolution in response to climate change in a selfing annual and suggest that shifts in tissue allocation strategies may underlie adaptive responses to drought in arid or semi-arid environments. This work also illustrates a novel approach, documenting that under specific circumstances, seeds from herbarium specimens may provide an untapped source of dormant propagules for future resurrection experiments.

Local adaptation of seed and seedling traits along a natural aridity gradient may both predict and constrain adaptive responses to climate change.

PREMISE: Variation in seed and seedling traits underlies how plants interact with their environment during establishment, a crucial life history stage. We quantified genetic-based variation in seed and seedling traits in populations of the annual plant Plantago patagonica across a natural aridity gradient, leveraging natural intraspecific variation to predict how populations might evolve in response to increasing aridity associated with climate change in the Southwestern U.S. METHODS: We quantified seed size, seed size variation, germination timing, and specific leaf area in a greenhouse common garden, and related these traits to the climates of source populations. We then conducted a terminal drought experiment to determine which traits were most predictive of survival under early-season drought. RESULTS: All traits showed evidence of clinal variation-seed size decreased, germination timing accelerated, and specific leaf area increased with increasing aridity. Populations with more variable historical precipitation regimes showed greater variation in seed size, suggestive of past selection shaping a diversified bet-hedging strategy mediated by seed size. Seedling height, achieved via larger seeds or earlier germination, was a significant predictor of survival under drought. CONCLUSIONS: We documented substantial interspecific trait variation as well as clinal variation in several important seed and seedling traits, yet these slopes were often opposite to predictions for how individual traits might confer drought tolerance. This work shows that plant populations may adapt to increasing aridity via correlated trait responses associated with alternative life history strategies, but that trade-offs might constrain adaptive responses in individual traits.

Evolutionary history and ecology shape the diversity and abundance of phytochemical arsenals across monkeyflowers.

We examine the extent to which phylogenetic effects and ecology are associated with macroevolutionary patterns of phytochemical defence production across the Mimulus phylogeny. We grew plants from 21 species representing the five major sections of the Mimulus phylogeny in a common garden to assess how the arsenals (NMDS groupings) and abundances (concentrations) of a phytochemical defence, phenylpropanoid glycosides (PPGs), vary across the phylogeny. Very few PPGs are widespread across the genus, but many are common to multiple sections of the genus. Phytochemical arsenals cluster among sections in an NMDS and are not associated with total concentration of PPGs. There is a strong phylogenetic signal for phytochemical arsenal composition across the Mimulus genus, whereas ecological variables such as growing season length, latitude, and elevation do not significantly influence arsenal. In contrast, there is little phylogenetic signal for total PPG concentration, and this trait is significantly influenced by several ecological factors. Phytochemical arsenals and abundances are influenced by plant life history form. Both phylogenetic effects and ecology are related to phytochemical patterns across species, albeit in different ways. The independence of phytochemical defence concentrations from arsenal compositions indicates that these aspects of defence may continue to evolve independently of one another.

To compete or defend: linking functional trait variation with life-history tradeoffs in a foundation tree species.

Although chemical deterrents to herbivory often exact costs in terms of plant growth, the manner in which those costs arise, and their physiological relationship to other functional traits, remain unclear. In the absence of appreciable herbivory, we examined interrelationships among chemical defense levels and other foliar functional traits (e.g., light-saturated photosynthesis, specific leaf area, nitrogen concentration) as co-determinants of tree growth and, by extension, competitive ability in high-density populations comprising 16 genotypes of Populus tremuloides. Across genotypes, concentrations of chemical defenses were not significantly related to other leaf functional traits, but levels of the salicinoid phenolic glycosides (SPGs) salicin, salicortin and tremulacin were each negatively correlated with relative mass growth (RMG) of aboveground woody tissue (P ≤ 0.001). RMG, in turn, underpinned 77% of the genotypic variation in relative height growth (our index of competitive ability). RMG was also positively related to light-saturated photosynthesis (P ≤ 0.001), which, together with the three SPGs, explained 86% of genotypic RMG variation (P ≤ 0.001). Moreover, results of a carbon balance simulation indicated that costs of resource allocation to SPGs, reaching nearly a third of annual crown photosynthesis, were likely mediated by substantial metabolic turnover, particularly for salicin. The lack of discernible links between foliar defense allocation and other (measured) functional traits, and the illustrated potential of metabolic turnover to reconcile influences of SPG allocation on RMG, shed additional light on fundamental physiological mechanisms underlying evolutionary tradeoffs between chemical defense investment and competitive ability in a foundation tree species.

The Genetic Architecture of Plant Defense Trade-offs in a Common Monkeyflower.

Determining how adaptive combinations of traits arose requires understanding the prevalence and scope of genetic constraints. Frequently observed phenotypic correlations between plant growth, defenses, and/or reproductive timing have led researchers to suggest that pleiotropy or strong genetic linkage between variants affecting independent traits is pervasive. Alternatively, these correlations could arise via independent mutations in different genes for each trait and extensive correlational selection. Here we evaluate these alternatives by conducting a quantitative trait loci (QTL) mapping experiment involving a cross between 2 populations of common monkeyflower (Mimulus guttatus) that differ in growth rate as well as total concentration and arsenal composition of plant defense compounds, phenylpropanoid glycosides (PPGs). We find no evidence that pleiotropy underlies correlations between defense and growth rate. However, there is a strong genetic correlation between levels of total PPGs and flowering time that is largely attributable to a single shared QTL. While this result suggests a role for pleiotropy/close linkage, several other QTLs also contribute to variation in total PPGs. Additionally, divergent PPG arsenals are influenced by a number of smaller-effect QTLs that each underlie variation in 1 or 2 PPGs. This result indicates that chemical defense arsenals can be finely adapted to biotic environments despite sharing a common biochemical precursor. Together, our results show correlations between defense and life-history traits are influenced by pleiotropy or genetic linkage, but genetic constraints may have limited impact on future evolutionary responses, as a substantial proportion of variation in each trait is controlled by independent loci.

Linking plant genes to insect communities: Identifying the genetic bases of plant traits and community composition.

Community genetics aims to understand the effects of intraspecific genetic variation on community composition and diversity, thereby connecting community ecology with evolutionary biology. Thus far, research has shown that plant genetics can underlie variation in the composition of associated communities (e.g., insects, lichen and endophytes), and those communities can therefore be considered as extended phenotypes. This work, however, has been conducted primarily at the plant genotype level and has not identified the key underlying genes. To address this gap, we used genome-wide association mapping with a population of 445 aspen (Populus tremuloides) genets to identify the genes governing variation in plant traits (defence chemistry, bud phenology, leaf morphology, growth) and insect community composition. We found 49 significant SNP associations in 13 Populus genes that are correlated with chemical defence compounds and insect community traits. Most notably, we identified an early nodulin-like protein that was associated with insect community diversity and the abundance of interacting foundation species (ants and aphids). These findings support the concept that particular plant traits are the mechanistic link between plant genes and the composition of associated insect communities. In putting the "genes" into "genes to ecosystems ecology", this work enhances understanding of the molecular genetic mechanisms that underlie plant-insect associations and the consequences thereof for the structure of ecological communities.

Optimal defense theory explains deviations from latitudinal herbivory defense hypothesis.

The latitudinal herbivory defense hypothesis (LHDH) postulates that the prevalence of species interactions, including herbivory, is greater at lower latitudes, leading to selection for increased levels of plant defense. While latitudinal defense clines may be caused by spatial variation in herbivore pressure, optimal defense theory predicts that clines could also be caused by ecogeographic variation in the cost of defense. For instance, allocation of resources to defense may not increase plant fitness when growing seasons are short and plants must reproduce quickly. Here we use a common garden experiment to survey genetic variation for constitutive and induced phenylpropanoid glycoside (PPG) concentrations across 35 Mimulus guttatus populations over a ~13° latitudinal transect. Our sampling regime is unique among studies of the LHDH in that it allows us to disentangle the effects of growing season length from those of latitude, temperature, and elevation. For five of the seven PPGs surveyed, we find associations between latitude and plant defense that are robust to population structure. However, contrary to the LHDH, only two PPGs were found at higher levels in low latitude populations, and total PPG concentrations were higher at higher latitudes. PPG levels are strongly correlated with growing season length, with higher levels of PPGs in plants from areas with longer growing seasons. Further, flowering time is positively correlated with the concentration of nearly all PPGs, suggesting that there may be a strong trade-off between development time and defense production. Our results reveal that ecogeographic patterns in plant defense may reflect variation in the cost of producing defense compounds in addition to variation in herbivore pressure. Thus, the biogeographic pattern predicted by the LHDH may not be accurate because the underlying factors driving variation in defense, in this case, growing season length, are not always associated with latitude in the same manner. Given these results, we conclude that LHDH cannot be interpreted without considering life history, and we recommend that future work on the LHDH move beyond solely testing the core LHDH prediction and place greater emphasis on isolating agents of selection that generate spatial variation in defense and herbivore pressure.

Foliar bacterial communities of trembling aspen in a common garden.

Microbial associations with plants are widely distributed and are structured by a number of biotic and physical factors. Among biotic factors, the host plant genotype may be integral to these plant-microbe interactions. Trees in the genus Populus have become models for studies in scaling effects of host plant genetics and in plant-microbe interactions. Using 454 pyrosequencing of the 16S rRNA gene, we assessed the foliar bacterial community of 7 genotypes of mature trembling aspen trees (Populus tremuloides Michx.) grown in a common garden. Trees were selected based on prior analyses showing clonal variation in their concentration of chemicals conferring resistance against insect herbivores. At broad taxonomic designations, the bacterial community of trembling aspen was similar across all plant genotypes. At a finer taxonomic scale, the foliage of these trees varied in their community composition, but there was no distinct pattern to colonization or abundance related to plant genotype. The most abundant operational taxonomic units (OTUs) were classified as Ralstonia, Bradyrhizobium, Pseudomonas, and Brucella. These OTUs varied across the common garden, but there was no significant effect of host plant genotype or spatial position on the abundance of these members. Our results suggest that aspen genotype is less important in the structuring of its foliar bacterial communities than are other, poorly understood processes.

Rapid phytochemical analysis of birch (Betula) and poplar (Populus) foliage by near-infrared reflectance spectroscopy.

Poplar (Populus) and birch (Betula) species are widely distributed throughout the northern hemisphere, where they are foundation species in forest ecosystems and serve as important sources of pulpwood. The ecology of these species is strongly linked to their foliar chemistry, creating demand for a rapid, inexpensive method to analyze phytochemistry. Our study demonstrates the feasibility of using near-infrared reflectance spectroscopy (NIRS) as an inexpensive, high-throughput tool for determining primary (e.g., nitrogen, sugars, starch) and secondary (e.g., tannins, phenolic glycosides) foliar chemistry of Populus and Betula species, and identifies conditions necessary for obtaining reliable quantitative data. We developed calibrations with high predictive power (residual predictive deviations ≤ 7.4) by relating phytochemical concentrations determined with classical analytical methods (e.g., spectrophotometric assays, liquid chromatography) to NIR spectra, using modified partial least squares regression. We determine that NIRS, although less sensitive and precise than classical methods for some compounds, provides useful predictions in a much faster, less expensive manner than do classical methods.

Patterns of phytochemical variation in Mimulus guttatus (yellow monkeyflower).

The search for general patterns in the production and allocation of plant defense traits will be facilitated by characterizing multivariate suites of defense, as well as by studying additional plant taxa, particularly those with available genomic resources. Here, we investigated patterns of genetic variation in phytochemical defenses (phenylpropanoid glycosides, PPGs) in Mimulus guttatus (yellow monkeyflower). We grew plants derived from several natural populations, consisting of multiple full-sibling families within each population, in a common greenhouse environment. We found substantial variation in the constitutive multivariate PPG phenotype and in constitutive levels of individual phytochemicals within plants (among leaves of different ages), within populations (among full-sibling families), and among populations. Populations consisting of annual plants generally, but not always, had lower concentrations of phytochemicals than did populations of perennial plants. Populations differed in their plastic response to artificial herbivory, both in the overall multivariate PPG phenotype and in the individual phytochemicals. The relationship between phytochemistry and another defense trait, trichomes, differed among populations. Finally, we demonstrated that one of the PPGs, verbascoside, acts as a feeding stimulant rather than a feeding deterrent for a specialist herbivore of M. guttatus, the buckeye caterpillar (Junonia coenia Nymphalidae). Given its available genetic resources, numerous, easily accessible natural populations, and patterns of genetic variation highlighted in this research, M. guttatus provides an ideal model system in which to test ecological and evolutionary theories of plant-herbivore interactions.

Life history and chemical defense interact to drive patterns of adaptation in an annual monkeyflower.

Although chemical defenses and herbivore pressure are widely established as key targets and agents of selection, their roles in local adaptation and determining potential evolutionary responses to changing climates are often neglected. Here, we explore fitness differences between 11 rangewide M. guttatus populations in a field common garden experiment and assess the agents and targets of selection driving relative fitness patterns. We use piecewise structural equation models to disentangle associations between chemical defenses, (phenylpropanoid glycosides; PPGs), and life history traits with herbivory and fitness. While the historical environment of populations is not predictive of fitness differences between populations, >90% of variation in fitness can be predicted by the flowering time and foliar PPG defense arsenal of a population. Piecewise structural equation models indicate that life history traits, particularly earlier flowering time, are strongly and directly linked to fitness. However, herbivory, particularly fruit predation, is also an important agent of selection that creates indirect links between fitness and both chemical defenses and life history traits. Our results emphasize the multivariate nature of the agents and targets of selections in producing adaptation and suggest that future responses to selection must navigate a complex fitness landscape.

Relative importance of genetic, ontogenetic, induction, and seasonal variation in producing a multivariate defense phenotype in a foundation tree species.

Plant adaptations for defense against herbivory vary both among species and among genotypes. Moreover, numerous forms of within-plant variation in defense, including ontogeny, induction, and seasonal gradients, allow plants to avoid expending resources on defense when herbivores are absent. We used an 18-year-old cottonwood common garden composed of Populus fremontii, Populus angustifolia, and their naturally occurring F(1) hybrids (collectively referred to as "cross types") to quantify and compare the relative influences of three hierarchical levels of variation (between cross types, among genotypes, and within individual genotypes) on univariate and multivariate phytochemical defense traits. Within genotypes, we evaluated ontogeny, induction (following cottonwood leaf beetle herbivory), and seasonal variation. We compared the effect sizes of each of these sources of variation on the plant defense phenotype. Three major patterns emerged. First, we observed significant differences in concentrations of defense phytochemicals among cross types, and/or among genotypes within cross types. Second, we found significant genetic variation for within-plant differences in phytochemical defenses: (a) based on ontogeny, levels of constitutive phenolic glycosides were nearly three times greater in the mature zone than in the juvenile zone within one cottonwood cross type, but did not significantly differ within another cross type; (b) induced levels of condensed tannins increased up to 65 % following herbivore damage within one cottonwood cross type, but were not significantly altered in another cross type; and (c) concentrations of condensed tannins tended to increase across the season, but did not do so across all cross types. Third, our estimates of effect size demonstrate that the magnitude of within-plant variation in a phytochemical defense can rival the magnitude of differences in defense among genotypes and/or cross types. We conclude that, in cottonwood and likely other plant species, multiple forms of within-individual variation have the potential to substantially influence ecological and evolutionary processes.

Climate and the biotic community structure plant resistance across biogeographic groups of yellow monkeyflower.

Characterizing correlates of phytochemical resistance trait variation across a landscape can provide insight into the ecological factors that have shaped the evolution of resistance arsenals. Using field-collected data and a greenhouse common garden experiment, we assessed the relative influences of abiotic and biotic drivers of genetic-based defense trait variation across 41 yellow monkeyflower populations from western and eastern North America and the United Kingdom. Populations experience different climates, herbivore communities, and neighboring vegetative communities, and have distinct phytochemical resistance arsenals. Similarities in climate as well as herbivore and vegetative communities decline with increasing physical distance separating populations, and phytochemical resistance arsenal composition shows a similarly decreasing trend. Of the abiotic and biotic factors examined, temperature and the neighboring vegetation community had the strongest relative effects on resistance arsenal differentiation, whereas herbivore community composition and precipitation have relatively small effects. Rather than simply controlling for geographic proximity, we jointly assessed the relative strengths of both geographic and ecological variables on phytochemical arsenal compositional dissimilarity. Overall, our results illustrate how abiotic conditions and biotic interactions shape plant defense traits in natural populations.

The genetics of phenotypic plasticity in plant defense: trichome production in Mimulus guttatus.

Insect herbivory is a major driving force of plant evolution. Phenotypic plasticity and developmental variation provide a means for plants to cope with variable herbivory. We characterized the genetics of developmental variation and phenotypic plasticity in trichome density, a putative defensive trait of Mimulus guttatus (yellow monkeyflower). Our results are evaluated in relation to the optimal defense theory, which provides testable predictions for plastic and developmental patterns in defense traits. We found that both developmental stage and simulated insect damage affected trichome production, but in different ways. Plants were more likely to produce at least some trichomes on later leaves than on earlier leaves, regardless of damage. Damage did not affect the average probability of producing trichomes, but it did increase the density of hairs on trichome-positive plants. We mapped trichome quantitative trait loci (QTL) by selectively genotyping a large panel of recombinant inbred lines derived from two highly divergent populations. Several highly pleiotropic QTL influenced multiple aspects of the trichome phenotype (constitutive, developmental, and/or plastic responses). Only one of the QTL influenced trichome induction following damage. In a result that is consistent with a central prediction of optimal defense theory, the high allele at this location was from the ancestral population with low constitutive trichome production.

Separating ontogenetic and environmental determination of resistance to herbivory in cottonwood.

We used narrowleaf cottonwood, Populus angustifolia, and the gall-forming aphid, Pemphigus betae, to determine the extent to which ontogenetic variation in resistance to herbivory is due to endogenous, stable genetic influences. In a three-year common garden trial using ramets propagated from the top, middle, and bottom of mature trees, we found that the resistance of trees to aphids was significantly higher in top vs. bottom source ramets, supporting the hypothesis of a stable, genetically programmed component to aphid resistance. The magnitude of ontogenetically based variation in resistance within an individual tree is comparable to the genetic variation in resistance among narrowleaf cottonwood genotypes or populations found in other studies. These ontogenetic-based findings have the potential to alter ecological interactions and evolutionary trajectories of plant-herbivore interactions.

Mechanisms of a locally adaptive shift in allocation among growth, reproduction, and herbivore resistance in Mimulus guttatus.

Environmental gradients can drive adaptive evolutionary shifts in plant resource allocation among growth, reproduction, and herbivore resistance. However, few studies have attempted to connect these adaptations to underlying physiological and genetic mechanisms. Here, we evaluate potential mechanisms responsible for a coordinated locally adaptive shift between growth, reproduction, and herbivore defense in the yellow monkeyflower, Mimulus guttatus. Through manipulative laboratory experiments, we found that gibberellin (GA) growth hormones may play a role in the developmental divergence between perennial and annual ecotypes of M. guttatus. Further, we detected an interaction between a locally adaptive chromosomal inversion, DIV1, and GA addition. This finding is consistent with the inversion contributing to the evolutionary divergence between inland annual and coastal perennial ecotypes by reducing GA biosynthesis/activity in perennials. Finally, we found evidence that the DIV1 inversion is partially responsible for a coordinated shift in the divergence of growth, reproduction, and herbivore resistance traits between coastal perennial and inland annual M. guttatus. The inversion has already been established to have a substantial impact on the life-history shift between long-term growth and rapid reproduction. Here, we demonstrate that the DIV1 inversion also has sizable impacts on both the total abundance and composition of phytochemical compounds involved in herbivore resistance.

The genetic correlation between flower size and water use efficiency in monkeyflowers.

QUESTION: Does water loss during drought stress represent an important physiological constraint on the evolution of flower size? ORGANISM: A genetically diverse population of Mimulus guttatus (yellow monkeyflower) originally sampled from an alpine meadow in Oregon, USA. METHODS: We grew plants of three different genotypic classes (small, medium, and large flowered) under both well-watered and drought-stress conditions and measured water use efficiency using stable carbon isotopes. RESULTS: There was no difference in water use efficiency among flower size genotypes under well-watered conditions, but the water use efficiency of small-flowered plants was substantially lower than that of medium or large genotypes under drought stress. Whether this paradoxical result is a direct effect of flower size or an indirect (i.e. pleiotropic) effect, the presence of a genetic correlation between floral and physiological traits indicates that selection of one does impact the other.

Genotypic variation in plant traits shapes herbivorous insect and ant communities on a foundation tree species.

Community genetics aims to understand the effects of intraspecific genetic variation on community composition and diversity, thereby connecting community ecology with evolutionary biology. Multiple studies have shown that different plant genotypes harbor different communities of associated organisms, such as insects. Yet, the mechanistic links that tie insect community composition to plant genetics are still not well understood. To shed light on these relationships, we explored variation in both plant traits (e.g., growth, phenology, defense) and herbivorous insect and ant communities on 328 replicated aspen (Populus tremuloides) genets grown in a common garden. We measured traits and visually surveyed insect communities annually in 2014 and 2015. We found that insect communities overall exhibited low heritability and were shaped primarily by relationships among key insects (i.e., aphids, ants, and free-feeders). Several tree traits affected insect communities and the presence/absence of species and functional groups. Of these traits, tree size and foliar phenology were the most important. Larger trees had denser (i.e., number of insects per unit tree size) and more diverse insect communities, while timing of bud break and bud set differentially influenced particular species and insect groups, especially leaf modifying insects. These findings will inform future research directed toward identification of plant genes and genetic regions that underlie the structure of associated insect communities.

Mating system and the evolution of quantitative traits: an experimental study of Mimulus guttatus.

The mating system of a population profoundly influences its evolution. Inbreeding alters the balance of evolutionary forces that determine the amount of genetic variation within a population. It redistributes that variation among individuals, altering heritabilities and genetic correlations. Inbreeding even changes the basic relationships between these genetic statistics and response to selection. If populations differing only in mating system are exposed to the same selection pressures, will they respond in qualitatively different ways? Here, we address this question by imposing selection on an index of two negatively correlated traits (flower size and development rate) within experimental populations that reproduce entirely by outcrossing, entirely by self-fertilizing, or by a mixture of outcrossing and selfing. Entirely selfing populations responded mainly by evolving larger flowers whereas outcrossing populations also evolved more rapid development. Divergence occurred despite an equivalent selection regime and no direct effect of mating system on fitness. The study provides an experimental demonstration of how the interaction of selection, genetic drift, and mating system can produce dramatic short-term changes in trait means, variances, and covariances.

Epigenetic Inheritance across the Landscape.

The study of epigenomic variation at the landscape-level in plants may add important insight to studies of adaptive variation. A major goal of landscape genomic studies is to identify genomic regions contributing to adaptive variation across the landscape. Heritable variation in epigenetic marks, resulting in transgenerational plasticity, can influence fitness-related traits. Epigenetic marks are influenced by the genome, the environment, and their interaction, and can be inherited independently of the genome. Thus, epigenomic variation likely influences the heritability of many adaptive traits, but the extent of this influence remains largely unknown. Here, we summarize the relevance of epigenetic inheritance to ecological and evolutionary processes, and review the literature on landscape-level patterns of epigenetic variation. Landscape-level patterns of epigenomic variation in plants generally show greater levels of isolation by distance and isolation by environment then is found for the genome, but the causes of these patterns are not yet clear. Linkage between the environment and epigenomic variation has been clearly shown within a single generation, but demonstrating transgenerational inheritance requires more complex breeding and/or experimental designs. Transgenerational epigenetic variation may alter the interpretation of landscape genomic studies that rely upon phenotypic analyses, but should have less influence on landscape genomic approaches that rely upon outlier analyses or genome-environment associations. We suggest that multi-generation common garden experiments conducted across multiple environments will allow researchers to understand which parts of the epigenome are inherited, as well as to parse out the relative contribution of heritable epigenetic variation to the phenotype.