The Ne/N ratio in applied conservation.

Recent developments within the IUCN and the Convention on Biological Diversity have affirmed the increasingly key role that effective population size (N e) and the effective size: census size ratio (N e/N) play in applied conservation and management of global biodiversity. This paper reviews and synthesizes information regarding the definition of N e and demographic and genetic methods for estimating effective size, census size, and their ratio. Emphasis is on single-generation estimates of contemporary N e/N, which are the most informative for practical applications. It is crucial to clearly define which individuals are included in the census size (N). Defining N as the number of adults alive at a given time facilitates comparisons across species. For a wide range of applications and experimental designs, inbreeding N e is simpler to calculate and interpret than variance N e. Effects of skewed sex ratio are generally modest, so most reductions to N e/N arise from overdispersed (greater-than-Poisson) variance in offspring number (σk2). Even when fecundity changes with age, overdispersed within-age variance generally contributes most to overall σk2, and both random and deterministic (mediated by selection) factors can be important. Most species are age-structured, so it is important to distinguish between effective size per generation (N e) and the effective number of breeders in one season or year (N b). Both N e and N b are important for applied conservation and management. For iteroparous species, a key metric is variance in lifetime reproductive success (σk•2), which can be affected by a variety of additional factors, including variation in longevity, skip or intermittent breeding, and persistent individual differences in reproductive success. Additional factors that can be important for some species are also discussed, including mating systems, population structure, sex reversal, reproductive compensation, captive propagation, and delayed maturity.

Low-coverage sequencing and Wahlund effect severely bias estimates of inbreeding, heterozygosity and effective population size in North American wolves.

vonHoldt et al. ((2024), Molecular Ecology, 33, e17231) (vH24) used low-coverage (average ~ 7X read depth) restriction site-associated DNA sequence data to estimate individual inbreeding and heterozygosity, and recent effective population size (Ne), in Great Lakes (GL) and Northern Rocky Mountain (RM) wolves. They concluded that RM heterozygosity rapidly declined between 1991 and 2020, and that Ne declined substantially in GL and RM over the last 50 generations. Here, we evaluate the effects of low sequence coverage and sampling strategy on vH24's findings and provide general recommendations for using sequence data to evaluate inbreeding, heterozygosity and Ne. Low-coverage sequencing resulted in downwardly biased estimates of individual inbreeding and heterozygosity, and an erroneous large temporal decline in RM heterozygosity due to declining read depth through time. Additionally, vH24's sampling strategy-which combined individuals from several genetically differentiated populations and across at least eight wolf generations-is expected to have resulted in severe downward bias in estimates of recent Ne for RM. We recommend using high-coverage sequence data ( ≥ $$ \ge $$ 15-20X) to estimate inbreeding and heterozygosity. Carefully filtering individuals, loci and genotypes, and using genotype imputation or likelihoods can help to minimise bias when low-coverage sequence data must be used. For estimation of contemporary Ne, the marginal benefits of using more than 103-104 loci are small, so aggressive filtering of loci with low average read depth potentially can retain most individuals without sacrificing much precision. Individuals are relatively more valuable than loci because analyses of contemporary Ne should focus on roughly single-generation samples from local breeding populations.

Influence of molecular marker type on estimating effective population size and other genetic parameters in a critically endangered parrot.

Genetics is a fast-moving field, and for conservation practitioners or ecologists, it can be bewildering. The choice of marker used in studies is fundamental; in the literature, preference has recently shifted from microsatellites to single nucleotide polymorphism (SNP) loci. Understanding how marker type affects estimates of population genetic parameters is important in the context of conservation, especially because the accuracy of estimates has a bearing on the actions taken to protect threatened species. We compare parameter estimates between seven microsatellites, 3761 SNP loci, and a random subset of 100 SNPs for the exact same 324 individual swift parrots, Lathamus discolor, and also use 457 additional samples from subsequent years to compare SNP estimates. Both marker types estimated a lower H O than H E. We show that microsatellites and SNPs mainly indicate a lack of spatial genetic structure, except when a priori collection locations were used on the SNP data in a discriminant analysis of principal components (DAPC). The 100-SNP subset gave comparable results to when the full dataset was used. Estimates of effective population size (N e) were comparable between markers when the same individuals were considered, but SNPs had narrower confidence intervals. This is reassuring because conservation assessments that rely on population genetic estimates based on a few microsatellites are unlikely to be nullified by the general shift toward SNPs in the literature. However, estimates between markers and datasets varied considerably when only adult samples were considered; hence, including samples of all age groups is recommended to be used when available. The estimated N e was higher for the full SNP dataset (2010-2019) than the smaller comparison data (2010-2015), which might be a better reflection of the species status. The lower precision of microsatellites may not necessarily be a barrier for most conservation applications; however, SNPs will improve confidence limits, which may be useful for practitioners.

Practical application of the linkage disequilibrium method for estimating contemporary effective population size: A review.

The method to estimate contemporary effective population size (Ne ) based on patterns of linkage disequilibrium (LD) at unlinked loci has been widely applied to natural and managed populations. The underlying model makes many simplifying assumptions, most of which have been evaluated in numerous studies published over the last two decades. Here, these performance evaluations are reviewed and summarized, with a focus on information that facilitates practical application to real populations in nature. Potential sources of bias that are discussed include calculation of r2 (a measure of LD), adjustments for sampling error, physical linkage, age structure, migration and spatial structure, mutation and selection, mating systems, changes in abundance, rare alleles, missing data, genotyping errors, data filtering choices and methods for combining multiple Ne estimates. Factors that affect precision are reviewed, including pseudoreplication that limits the information gained from large genomics datasets, constraints imposed by small samples of individuals, and the challenges in obtaining robust estimates for large populations. Topics that merit further research include the potential to weight r2 values by allele frequency, lump samples of individuals, use genotypic likelihoods rather than called genotypes, prune large LD values and apply the method to species practising partial monogamy.

Partitioning variance in reproductive success, within years and across lifetimes.

Variance in reproductive success (sk2, with k = number of offspring) plays a large role in determining the rate of genetic drift and the scope within which selection acts. Various frameworks have been proposed to parse factors that contribute to sk2, but none has focused on age-specific values of ϕ=sk2/k¯, which indicate the degree to which reproductive skew is overdispersed (compared to the random Poisson expectation) among individuals of the same age and sex. Instead, within-age effects are generally lumped with residual variance and treated as "noise." Here, an ANOVA sums-of-squares framework is used to partition variance in annual and lifetime reproductive success into between-group and within-group components. For annual reproduction, the between-age effect depends on age-specific fecundity (b x), but relatively few empirical data are available on the within-age effect, which depends on ϕ x. By defining groups by age-at-death rather than age, the same ANOVA framework can be used to partition variance in lifetime reproductive success (LRS) into between-group and within-group components. Analytical methods are used to develop null-model expectations for random contributions to within-group and between-group components. For analysis of LRS, random variation in longevity appears as part of the between-group variance, and effects (if any) of skip breeding and persistent individual differences contribute to the within-group variance. Simulations are used to show that the methods for variance partitioning are asymptotically unbiased. Practical application is illustrated with empirical data for annual reproduction in American black bears and lifetime reproduction in Dutch great tits. Results show that overdispersed within-age variance (1) dominates annual sk2 in both male and female black bears, (2) is the primary factor that reduces annual effective size to a fraction of the number of adults, and (3) represents most of the opportunity for selection. In contrast, about a quarter of the variance in LRS in great tits can be attributed to random variation in longevity, and most of the rest is due to modest differences in fecundity with age estimated for a single cohort of females. R code is provided that reads generic input files for annual and lifetime reproductive success and allows users to conduct variance partitioning with their own data.

Robustness of Hill's overlapping-generation method for calculating Ne to extreme patterns of reproductive success.

For species with overlapping generations, the most widely used method to calculate effective population size (Ne) is Hill's, the key parameter for which is lifetime variance in offspring number ([Formula: see text]). Hill's model assumes a stable age structure and constant abundance, and sensitivity to those assumptions has been evaluated previously. Here I evaluate the robustness of Hill's model to extreme patterns of reproductive success, whose effects have not been previously examined: (1) very strong reproductive skew; (2) strong temporal autocorrelations in individual reproductive success; and (3) strong covariance of individual reproduction and survival. Genetic drift (loss of heterozygosity and increase in allele frequency variance) was simulated in age-structured populations using methods that generated no autocorrelations or covariances (Model NoCor); or created strong positive (Model Positive) or strong negative (Model Negative) temporal autocorrelations in reproduction and covariances between reproduction and survival. Compared to Model NoCor, the other models led to greatly elevated or reduced [Formula: see text], and hence greatly reduced or elevated Ne, respectively. A new index is introduced (ρα,α+), which is the correlation between (1) the number of offspring produced by each individual at the age at maturity (α), and (2) the total number of offspring produced during the rest of their lifetimes. Mean ρα,α+ was ≈0 under Model NoCor, strongly positive under Model Positive, and strongly negative under Model Negative. Even under the most extreme reproductive scenarios in Models Positive and Negative, when [Formula: see text] was calculated from the realized population pedigree and used to calculate Ne in Hill's model, the result accurately predicted the rate of genetic drift in simulated populations. These results held for scenarios where age-specific reproductive skew was random (variance ≈ mean) and highly overdispersed (variance up to 20 times the mean). Collectively, these results are good news for researchers as they demonstrate the robustness of Hill's model even in extreme reproductive scenarios.

Null Models for the Opportunity for Selection.

AbstractCrow's "opportunity for selection" (I=variance in relative fitness) is an important albeit controversial eco-evolutionary concept, particularly regarding the most appropriate null model(s). Here, we treat this topic in a comprehensive way by considering opportunities for both fertility selection (If) and viability selection (Im) for discrete generations, both seasonal and lifetime reproductive success in age-structured species, and experimental designs that include either a full or partial life cycle, with complete enumeration or random subsampling. For each scenario, a null model that includes random demographic stochasticity can be constructed that follows Crow's initial formulation that I=If+Im. The two components of I are qualitatively different. Whereas an adjusted If (ΔIf) can be computed that accounts for random demographic stochasticity in offspring number, Im cannot be similarly adjusted in the absence of data on phenotypic traits under viability selection. Including as potential parents some individuals that die before reproductive age produces an overall zero-inflated Poisson null model. It is always important to remember that (1) Crow's I represents only the opportunity for selection and not selection itself and (2) the species' biology can lead to random stochasticity in offspring number that is either overdispersed or underdispersed compared with the Poisson (Wright-Fisher) expectation.

Effective population size in a partially clonal plant is not predicted by the number of genetic individuals.

Estimating effective population size (N e) is important for theoretical and practical applications in evolutionary biology and conservation. Nevertheless, estimates of N e in organisms with complex life-history traits remain scarce because of the challenges associated with estimation methods. Partially clonal plants capable of both vegetative (clonal) growth and sexual reproduction are a common group of organisms for which the discrepancy between the apparent number of individuals (ramets) and the number of genetic individuals (genets) can be striking, and it is unclear how this discrepancy relates to N e. In this study, we analysed two populations of the orchid Cypripedium calceolus to understand how the rate of clonal versus sexual reproduction affected N e. We genotyped >1000 ramets at microsatellite and SNP loci, and estimated contemporary N e with the linkage disequilibrium method, starting from the theoretical expectation that variance in reproductive success among individuals caused by clonal reproduction and by constraints on sexual reproduction would lower N e. We considered factors potentially affecting our estimates, including different marker types and sampling strategies, and the influence of pseudoreplication in genomic data sets on N e confidence intervals. The magnitude of N e/N ramets and N e/N genets ratios we provide may be used as reference points for other species with similar life-history traits. Our findings demonstrate that N e in partially clonal plants cannot be predicted based on the number of genets generated by sexual reproduction, because demographic changes over time can strongly influence N e. This is especially relevant in species of conservation concern in which population declines may not be detected by only ascertaining the number of genets.

The opportunity for selection: A slippery concept in ecology and evolution.

Natural selection can only occur if individuals differ in fitness. For this reason, the variance in relative fitness has been equated with the 'opportunity for selection' ( I ), which has a long, albeit somewhat controversial, history. In this paper we discuss the use/misuse of I and related metrics in evolutionary ecology. The opportunity is only realised if some fraction of I is caused by trait variation. Thus, I > 0 does not imply that selection is occurring, as sometimes erroneously assumed, because all fitness variation could be independent of phenotype. The selection intensity on any given trait cannot exceed I , but this upper limit will never be reached because (a) stochastic factors always affect fitness, and (b) there might be multiple traits under selection. The expected magnitude of the stochastic component of I is negatively correlated with mean fitness. Uncertainty in realised I is also larger when mean fitness or population/sample size are low. Variation in I across time or space thus can be dominated (or solely driven) by variation in the strength of demographic stochasticity. We illustrate these points using simulations and empirical data from a population study on great tits Parus major. Our analysis shows that the scope for fecundity selection in the great tits is substantially higher when using annual number of recruits as the fitness measure, rather than fledglings or eggs, even after adjusting for the dependence of I on mean fitness. This suggests nonrandom survival of juveniles across families between life stages. Indeed, previous work on this population has shown that offspring recruitment is often nonrandom with respect to clutch size and laying date of parents, for example. We conclude that one cannot make direct inferences about selection based on fitness data alone. However, examining variation in ∆ I F (the opportunity for fecundity selection adjusted for mean fitness) across life stages, years or environments can offer clues as to when/where fecundity selection might be strongest, which can be useful for research planning and experimental design.

What Is Ne, Anyway?

Few doubt that effective population size (Ne) is one of the most important parameters in evolutionary biology, but how many can say they really understand the concept? Ne is the evolutionary analog of the number of individuals (or adults) in the population, N. Whereas ecological consequences of population size depend on N, evolutionary consequences (rates of loss of genetic diversity and increase in inbreeding; relative effectiveness of selection) depend on Ne. Formal definitions typically relate effective size to a key population genetic parameter, such as loss of heterozygosity or variance in allele frequency. However, for practical application to real populations, it is more useful to define Ne in terms of 3 demographic parameters: number of potential parents (adult N), and mean and variance in offspring number. Defined this way, Ne determines the rate of random genetic drift across the entire genome in the offspring generation. Other evolutionary forces (mutation, migration, selection)-together with factors such as variation in recombination rate-can also affect genetic variation, and this leads to heterogeneity across the genome in observed rates of genetic change. For some, it has been convenient to interpret this heterogeneity in terms of heterogeneity in Ne, but unfortunately, this has muddled the concepts of genetic drift and effective population size. A commonly repeated misconception is that Ne is the number of parents that actually contribute genes to the next generation (NP). In reality, NP can be smaller or larger than Ne, and the NP/Ne ratio depends on the sex ratio, the mean and variance in offspring number, and whether inbreeding or variance Ne is of interest.

Implications of Large-Effect Loci for Conservation: A Review and Case Study with Pacific Salmon.

The increasing feasibility of assembling large genomic datasets for non-model species presents both opportunities and challenges for applied conservation and management. A popular theme in recent studies is the search for large-effect loci that explain substantial portions of phenotypic variance for a key trait(s). If such loci can be linked to adaptations, 2 important questions arise: 1) Should information from these loci be used to reconfigure conservation units (CUs), even if this conflicts with overall patterns of genetic differentiation? 2) How should this information be used in viability assessments of populations and larger CUs? In this review, we address these questions in the context of recent studies of Chinook salmon and steelhead (anadromous form of rainbow trout) that show strong associations between adult migration timing and specific alleles in one small genomic region. Based on the polygenic paradigm (most traits are controlled by many genes of small effect) and genetic data available at the time showing that early-migrating populations are most closely related to nearby late-migrating populations, adult migration differences in Pacific salmon and steelhead were considered to reflect diversity within CUs rather than separate CUs. Recent data, however, suggest that specific alleles are required for early migration, and that these alleles are lost in populations where conditions do not support early-migrating phenotypes. Contrasting determinations under the US Endangered Species Act and the State of California's equivalent legislation illustrate the complexities of incorporating genomics data into CU configuration decisions. Regardless how CUs are defined, viability assessments should consider that 1) early-migrating phenotypes experience disproportionate risks across large geographic areas, so it becomes important to identify early-migrating populations that can serve as reliable sources for these valuable genetic resources; and 2) genetic architecture, especially the existence of large-effect loci, can affect evolutionary potential and adaptability.

Pseudoreplication in genomic-scale data sets.

In genomic-scale data sets, loci are closely packed within chromosomes and hence provide correlated information. Averaging across loci as if they were independent creates pseudoreplication, which reduces the effective degrees of freedom (df') compared to the nominal degrees of freedom, df. This issue has been known for some time, but consequences have not been systematically quantified across the entire genome. Here, we measured pseudoreplication (quantified by the ratio df'/df) for a common metric of genetic differentiation (FST ) and a common measure of linkage disequilibrium between pairs of loci (r2 ). Based on data simulated using models (SLiM and msprime) that allow efficient forward-in-time and coalescent simulations while precisely controlling population pedigrees, we estimated df' and df'/df by measuring the rate of decline in the variance of mean FST and mean r2 as more loci were used. For both indices, df' increases with Ne and genome size, as expected. However, even for large Ne and large genomes, df' for mean r2 plateaus after a few thousand loci, and a variance components analysis indicates that the limiting factor is uncertainty associated with sampling individuals rather than genes. Pseudoreplication is less extreme for FST , but df'/df ≤0.01 can occur in data sets using tens of thousands of loci. Commonly-used block-jackknife methods consistently overestimated var (FST ), producing very conservative confidence intervals. Predicting df' based on our modelling results as a function of Ne , L, S, and genome size provides a robust way to quantify precision associated with genomic-scale data sets.

Relative Precision of the Sibship and LD Methods for Estimating Effective Population Size With Genomics-Scale Datasets.

Computer simulations were used to compare relative precision of 2 widely used single-sample methods for estimating effective population size (Ne)-the sibship method and the linkage disequilibrium (LD) method. Emphasis is on performance when thousands of gene loci are used, which now can easily be achieved even for nonmodel species. Results show that unless Ne is very small, if at least 500-2000 diallelic loci are used, precision of the LD method is higher than the maximum possible precision for the sibship method, which occurs when all sibling relationships have been correctly identified. Results also show that when precision is high for both methods, their estimates of Ne are highly and positively correlated, which limits additional gains in precision that might be obtained by combining information from the 2 estimators.

Big Data in Conservation Genomics: Boosting Skills, Hedging Bets, and Staying Current in the Field.

A current challenge in the fields of evolutionary, ecological, and conservation genomics is balancing production of large-scale datasets with additional training often required to handle such datasets. Thus, there is an increasing need for conservation geneticists to continually learn and train to stay up-to-date through avenues such as symposia, meetings, and workshops. The ConGen meeting is a near-annual workshop that strives to guide participants in understanding population genetics principles, study design, data processing, analysis, interpretation, and applications to real-world conservation issues. Each year of ConGen gathers a diverse set of instructors, students, and resulting lectures, hands-on sessions, and discussions. Here, we summarize key lessons learned from the 2019 meeting and more recent updates to the field with a focus on big data in conservation genomics. First, we highlight classical and contemporary issues in study design that are especially relevant to working with big datasets, including the intricacies of data filtering. We next emphasize the importance of building analytical skills and simulating data, and how these skills have applications within and outside of conservation genetics careers. We also highlight recent technological advances and novel applications to conservation of wild populations. Finally, we provide data and recommendations to support ongoing efforts by ConGen organizers and instructors-and beyond-to increase participation of underrepresented minorities in conservation and eco-evolutionary sciences. The future success of conservation genetics requires both continual training in handling big data and a diverse group of people and approaches to tackle key issues, including the global biodiversity-loss crisis.

Detecting population declines via monitoring the effective number of breeders (Nb ).

Estimating the effective population size and effective number of breeders per year (Nb ) can facilitate early detection of population declines. We used computer simulations to quantify bias and precision of the one-sample LDNe estimator of Nb in age-structured populations using a range of published species life history types, sample sizes, and DNA markers. Nb estimates were biased by ~5%-10% when using SNPs or microsatellites in species ranging from fishes to mosquitoes, frogs, and seaweed. The bias (high or low) was similar for different life history types within a species suggesting that life history variation in populations will not influence Nb estimation. Precision was higher for 100 SNPs (H ≈ 0.30) than for 15 microsatellites (H ≈ 0.70). Confidence intervals (CIs) were occasionally too narrow, and biased high when Nb was small (Nb  < 50); however, the magnitude of bias would unlikely influence management decisions. The CIs (from LDNe) were sufficiently narrow to achieve high statistical power (≥0.80) to reject the null hypothesis that Nb  = 50 when the true Nb  = 30 and when sampling 50 individuals and 200 SNPs. Similarly, CIs were sufficiently narrow to reject Nb  = 500 when the true Nb  = 400 and when sampling 200 individuals and 5,000 loci. Finally, we present a linear regression method that provides high power to detect a decline in Nb when sampling at least five consecutive cohorts. This study provides guidelines and tools to simulate and estimate Nb for age structured populations (https://github.com/popgengui/agestrucnb/), which should help biologists develop sensitive monitoring programmes for early detection of changes in Nb and population declines.

An estimator of the Opportunity for Selection that is independent of mean fitness.

Variation among individuals in number of offspring (fitness, k) sets an upper limit to the evolutionary response to selection. This constraint is quantified by Crow's Opportunity for Selection (I), which is the variance in relative fitness (I = σ2k /(uk )2 ). Crow's I has been widely used but remains controversial because it depends on mean offspring number in a sample ( k¯ ). Here, I used a generalized Wright-Fisher model that allows for unequal probabilities of producing offspring to evaluate behavior of Crow's I and related indices under a wide range of sampling scenarios. Analytical and numerical results are congruent and show that rescaling the sample variance (s2k ) to its expected value at a fixed k¯2 removes dependence of I on mean offspring number, but the result still depends on choice of k¯2 . A new index is introduced, ΔI = Π- E(Îdrift ) = Π- 1/ k¯ , which makes Î independent of sample k¯ without the need for variance rescaling. ΔI has a straightforward interpretation as the component of variance in relative fitness that exceeds that expected under a null model of random reproductive success. ΔI can be used to directly compare estimates of the Opportunity for Selection for samples from different studies, different sexes, and different life stages.

Conservation and Management of Salmon in the Age of Genomics.

Salmon were among the first nonmodel species for which systematic population genetic studies of natural populations were conducted, often to support management and conservation. The genomics revolution has improved our understanding of the evolutionary ecology of salmon in two major ways: (a) Large increases in the numbers of genetic markers (from dozens to 104-106) provide greater power for traditional analyses, such as the delineation of population structure, hybridization, and population assignment, and (b) qualitatively new insights that were not possible with traditional genetic methods can be achieved by leveraging detailed information about the structure and function of the genome. Studies of the first type have been more common to date, largely because it has taken time for the necessary tools to be developed to fully understand the complex salmon genome. We expect that the next decade will witness many new studies that take full advantage of salmonid genomic resources.

Author Correction: Gene-associated markers provide tools for tackling illegal fishing and false eco-certification.

An amendment to this paper has been published and can be accessed via a link at the top of the paper.

The evolution of microendemism in a reef fish (Hypoplectrus maya).

Marine species tend to have extensive distributions, which are commonly attributed to the dispersal potential provided by planktonic larvae and the rarity of absolute barriers to dispersal in the ocean. Under this paradigm, the occurrence of marine microendemism without geographic isolation in species with planktonic larvae poses a dilemma. The recently described Maya hamlet (Hypoplectrus maya, Serranidae) is exactly such a case, being endemic to a 50-km segment of the Mesoamerican Barrier Reef System (MBRS). We use whole-genome analysis to infer the demographic history of the Maya hamlet and contrast it with the sympatric and pan-Caribbean black (H. nigricans), barred (H. puella) and butter (H. unicolor) hamlets, as well as the allopatric but phenotypically similar blue hamlet (H. gemma). We show that H. maya is indeed a distinct evolutionary lineage, with genomic signatures of inbreeding and a unique demographic history of continuous decrease in effective population size since it diverged from congeners just ~3,000 generations ago. We suggest that this case of microendemism may be driven by the combination of a narrow ecological niche and restrictive oceanographic conditions in the southern MBRS, which is consistent with the occurrence of an unusually high number of marine microendemics in this region. The restricted distribution of the Maya hamlet, its decline in both census and effective population sizes, and the degradation of its habitat place it at risk of extinction. We conclude that the evolution of marine microendemism can be a fast and dynamic process, with extinction possibly occurring before speciation is complete.

Life history and temporal variability of escape events interactively determine the fitness consequences of aquaculture escapees on wild populations.

Domesticated individuals are likely to be maladaptive in the wild due to adaptation to captivity. Escaped aquaculture fish can cause unintended fitness and demographic consequences for their wild conspecifics through interbreeding and competition. Escape events from different sources exhibit great heterogeneity in their frequencies and magnitudes, ranging from rare but large spillover during a storm, to continuous low-level leakage caused by operational errors. The timescale of escape events determines the distribution of gene flow from aquaculture to the wild. The evolutionary consequences of this variation in timescale will depend on the degree of generation overlap and the focal species' life history attributes, especially those under selection in aquaculture (e.g., growth rate, which can influence additional demographically important traits such as age at maturity). To evaluate the effects of variable escape both within and across generations, we construct an age-structured model of coupled genetic and demographic dynamics and parameterize it for species with contrasting life history characteristics (Salmo salar and Gadus morhua). Our results are consistent with earlier discrete-generation models that constant, low-level spillover can have a greater impact than rare, large pulses of leakage, even after accounting for the averaging effects of overlapping generations. The age-structured model also allows detailed evaluation of the role of different life history traits, which reveals that species with longer generation times might experience greater fitness consequences of aquaculture spillover but are less sensitive to variability in spillover. Additionally, environment-induced earlier maturity of escapees can increase the fitness effects on wild fish, especially those with shorter generation times. Our results suggest that effective management to minimize the unintended fitness consequences of aquaculture releases might require extensive monitoring efforts on constant, low-level spillover and assessment of the focal species' life history characteristics.