Are first-generation (F 1 ) hybrids typically intermediate for all traits that differentiate their parents? Or are they similar to one parent for most traits, or even mismatched for divergent traits? Although the phenotype of otherwise viable and fertile hybrids determines their fate, little is known about the general patterns, predictors, and consequences of phenotype expression in hybrids. To address this empirical gap, we compiled data from nearly 200 studies where traits were measured in a common environment for two parent populations and F 1 hybrids. We find that individual traits are typically halfway between the parental midpoint and one parental value (i.e., hybrid trait values are typically 0.25 or 0.75 if parents' values are 0 & 1). When considering pairs of traits together, a hybrid's multivariate phenotype tends to resemble one parent (pairwise parent-bias) about 50 % more than the other while also exhibiting a similar magnitude of trait mismatch due to different traits having dominance in conflicting directions. We detect no phylogenetic signal nor an effect of parental genetic distance on dominance or mismatch. Using data from an experimental field planting of recombinant hybrid sunflowers-where there is among-individual variation in dominance and mismatch due to segregation of divergent alleles-we illustrate that pairwise parent-bias improves fitness while mismatch reduces fitness. Importantly, the effect of mismatch on fitness was stronger than that of pairwise parent-bias. In sum, our study has three major conclusions. First, hybrids between ecologically divergent natural populations are typically not phenotypically intermediate but rather exhibit substantial mismatch while also resembling one parent more than the other. Second, dominance and mismatch are likely determined by population-specific processes rather than general rules. Finally, selection against hybrids likely results from both selection against somewhat intermediate phenotypes and against mismatched trait combinations.

Are first-generation (F 1 ) hybrids typically intermediate for all traits that differentiate their parents? Or are they similar to one parent for most traits, or even mismatched for divergent traits? Although the phenotype of otherwise viable and fertile hybrids determines their fate, little is known about the general patterns, predictors, and consequences of phenotype expression in hybrids. To address this empirical gap, we compiled data from nearly 200 studies where traits were measured in a common environment for two parent populations and F 1 hybrids. We find that individual traits are typically halfway between the parental midpoint and one parental value (i.e., hybrid trait values are typically 0.25 or 0.75 if parents' values are 0 & 1). When considering pairs of traits together, a hybrid's multivariate phenotype tends to resemble one parent (pairwise parent-bias) about 50 % more than the other while also exhibiting a similar magnitude of trait mismatch due to different traits having dominance in conflicting directions. We detect no phylogenetic signal nor an effect of parental genetic distance on dominance or mismatch. Using data from an experimental field planting of recombinant hybrid sunflowers-where there is among-individual variation in dominance and mismatch due to segregation of divergent alleles-we illustrate that pairwise parent-bias improves fitness while mismatch reduces fitness. Importantly, the effect of mismatch on fitness was stronger than that of pairwise parent-bias. In sum, our study has three major conclusions. First, hybrids between ecologically divergent natural populations are typically not phenotypically intermediate but rather exhibit substantial mismatch while also resembling one parent more than the other. Second, dominance and mismatch are likely determined by population-specific processes rather than general rules. Finally, selection against hybrids likely results from both selection against somewhat intermediate phenotypes and against mismatched trait combinations. hybridization | speciation | phenotypic mismatch | opposing dominanceCorrespondence: ken.thompson@zoology.ubc.ca

No abstract

Women in science, technology, engineering, and math are not equally represented across tenure-track career stages, and this extends to grant funding, where women applicants often have lower success rates compared with men. While gender bias in reviewers has been documented, it is currently unknown whether written language in grant applications varies predictably with gender to elicit bias against women. Here we analyse the text of ∼2000 public research summaries from the 2016 Natural Sciences and Engineering Research Council (NSERC) individual Discovery Grant (DG) program. We explore the relationship between language variables, inferred gender and career stage, and funding levels. We also analyse aggregated data from the 2012–2018 NSERC DG competitions to determine whether gender impacted the probability of receiving a grant for early-career researchers. We document a marginally significant gender difference in funding levels for successful grants, with women receiving $1756 less than men, and a large and significant difference in rejection rates among early-career applicants (women: 40.4% rejection; men: 33.0% rejection rate). Language variables had little ability to predict gender or funding level using predictive modelling. Our results indicate that NSERC funding levels and success rates differ between men and women, but we find no evidence that gendered language use affected funding outcomes.

1. Populations are embedded in communities, but despite their potential to affect individual fitness, it is unknown whether and how species interactions evolve in communities. Evolutionary outcomes are likely more complex in natural communities because (a) the evolution of interactions may not be evenly distributed among all community members and (b) coevolution is conditional on the environmental conditions within which interactions are playing out. 2. To test the evolution of interaction strengths in natural communities, we performed two common garden experiments in grassland communities in Northern California. In each garden, we transplanted individuals of four populations (one local, three foreign) of an annual invasive grass Bromus hordeaceus into natural communities, characterized the interaction neighbourhood around each focal individual, and quantified individual fitness. This method allowed us to fit multispecies competition models to fitness data, estimating interaction strengths between focal B. hordeaceus populations and each of seven species that were common in the interaction neighbourhoods, in each garden.3. We found that interaction strengths significantly differed among local and foreign source populations, but the direction and magnitude of evolution differed among common gardens and among neighbour species-in neither garden were interactions experienced more strongly by foreign populations compared to local populations. The fitness of local populations (relative to foreign populations) decreased when neighbours were removed, strongly enough in one garden to cause strong local maladaptation, and the local population did not perform the best in either garden. 4.Synthesis. Together, our results demonstrate how species interactions evolve to determine fitness in ecological communities, providing a richer view of adaptation in natural systems. In our study, this richness included the unique challenges populations face in nature: uneven abundances and a diffuseness of species interactions, nonlinear density effects on fitness, and evidence of (mal)adaptation that is conditional on local conditions. We conclude by hypothesizing the causes and consequences of challenges to adaptation and how they help identify priority areas for the field.