An evaluation of the potential factors affecting lifetime reproductive success in salmonids

Abstract Lifetime reproductive success (LRS), the number of offspring produced over an organism's lifetime, is a fundamental component of Darwinian fitness. For taxa such as salmonids with multiple species of conservation concern, understanding the factors affecting LRS is critical for the development and implementation of successful conservation management practices. Here, we reviewed the published literature to synthesize factors affecting LRS in salmonids including significant effects of hatchery rearing, life history, and phenotypic variation, and behavioral and spawning interactions. Additionally, we found that LRS is affected by competitive behavior on the spawning grounds, genetic compatibility, local adaptation, and hybridization. Our review of existing literature revealed limitations of LRS studies, and we emphasize the following areas that warrant further attention in future research: (1) expanding the range of studies assessing LRS across different life‐history strategies, specifically accounting for distinct reproductive and migratory phenotypes; (2) broadening the variety of species represented in salmonid fitness studies; (3) constructing multigenerational pedigrees to track long‐term fitness effects; (4) conducting LRS studies that investigate the effects of aquatic stressors, such as anthropogenic effects, pathogens, environmental factors in both freshwater and marine environments, and assessing overall body condition, and (5) utilizing appropriate statistical approaches to determine the factors that explain the greatest variation in fitness and providing information regarding biological significance, power limitations, and potential sources of error in salmonid parentage studies. Overall, this review emphasizes that studies of LRS have profoundly advanced scientific understanding of salmonid fitness, but substantial challenges need to be overcome to assist with long‐term recovery of these keystone species in aquatic ecosystems.


| INTRODUC TI ON
Lifetime reproductive success (LRS) is broadly defined as the number of offspring an individual produces over the span of a lifetime (Clutton-Brock, 1988) and is generally considered a sufficient estimate of fitness in natural populations (Arnold & Wade, 1984;Grafen, 1988). LRS is thought to consist of two main components which intrinsically connect both sexual and natural selection. The first component is the number of offspring produced, with the caveat that distinct evolutionary trade-offs exist between offspring number and quality, and the second component is the number of available reproductive seasons within an individual's lifetime (Barrowclough & Rockwell, 1993;Olofsson et al., 2009;Philippi & Seger, 1989). In order to have evolutionary consequences, individual phenotypic differences in LRS need to be heritable across generations (Shuster & Wade, 2003).
Although LRS is a commonly used estimation of fitness across taxonomic groups, gathering LRS data has presented a challenge in natural populations, particularly for species with external fertilization and no postnatal parental care. A direct count of the number of offspring requires both confidence in sampling and accuracy in parentage assignment once sampled. Furthermore, following an individual throughout their lifespan can be technically challenging, costly, and in many cases, simply not possible. Therefore, constraints on technical sampling of parents in natural environments have led some studies to rely on a correlate or "proxy" measure of fitness, such as morphology, phenology, performance measures, and various life-history traits (some examples would include body mass, seasonal timing, growth rate, and offspring size; Barrowclough & Rockwell, 1993;Franklin & Morrissey, 2017;Kingsolver et al., 2012).
Salmonids represent an ideal study system to estimate pedigree structure, reliably measure LRS, and address evolutionary questions given their unique life-history strategies and population structure, high fecundity, wide variability in life-history traits, and the wealth of genetic information gained from commercial and conservation aquaculture . Additionally, salmonids demonstrate very high individual-and population-level variability in LRS, intense intrasexual competition, and high mortality overall, making them also ideal for addressing questions in sexual selection (Fleming & Reynolds, 2004).
Variation in numerous life-history traits occurs both within and across salmonid species, including both semelparous and iteroparous reproductive strategies (Quinn, 2005; see Box 1 for glossary of terms presented). Semelparity is characterized by a single reproductive event, followed by death, and represents a life-history strategy that can be found across the taxonomic spectrum (Braithwaite & Lee, 1979;Crespi & Teo, 2002;Fritz et al., 1982;Young & Augspurger, 1991). Multiple species of salmonid fishes exhibit a semelparous life history, but it is particularly common in the genus Oncorhynchus, including Chinook (Oncorhynchus tshawytscha), Coho (Oncorhynchus kisutch), Sockeye (Oncorhynchus nerka), Pink (Oncorhynchus gorbuscha), and Chum (Oncorhynchus keta) Salmon (Crespi & Teo, 2002).
Since semelparous species of salmonids reproduce offspring during a single reproductive bout prior to death, LRS can be estimated by evaluating a single breeding season. Alternatively, salmonids exhibiting the potential for iteroparity are characterized by repeat breeding episodes, the number of which depends on numerous factors including sex, size, and anthropogenic impacts (Fleming, 1998). Some examples of iteroparous salmonids include European Grayling (Thymallus thymallus), Lake Whitefish (Coregonus clupeaformis), Rainbow and Steelhead Trout (Oncorhynchus mykiss), Cutthroat Trout (Oncorhynchus clarkii), Lake Charr (Salvelinus namaycush),
Salmonids also display a vast array of other alternative life-history strategies that can impact LRS, such as alternative reproductive phenotypes and alternative migratory tactics. For example, precocial males generally mature (i.e., produce milt) at least a year earlier than average adult males and, depending on the species, can either migrate to the ocean and return early (commonly referred to as "jack" males), partially migrate ("minijacks"), or mature entirely in freshwater ("residents") (Johnson et al., 2012;Mullan et al., 1992;Pearsons et al., 2009;Zimmerman et al., 2003). Precocial maturation has often been used to describe resident males that produce milt as parr, but in this review we use the term "precocial" more generally to refer to males that mature at least a year earlier than average adult males of a given species. This precocial maturation, or "jacking," is in contrast to

Anadromous
Fish born in freshwater, followed by growth and rearing in saltwater, and a return to freshwater to reproduce.

Broodstock
Group of reproductively mature fish used for artificial spawning in the hatchery setting; parents of hatcheryorigin fish.

Hatchery-origin
Fish that were born in the hatchery through artificial spawning.

Integrated Broodstock
Hatchery population created by varying proportions of hatchery-origin and natural-origin fish. Fully integrated hatchery programs use only natural-origin fish in broodstock. Typically observed in supplementation programs.

Iteroparity
A life-history strategy characterized by multiple reproductive events across more than one spawning season.

Lifetime Reproductive Success (LRS)
As defined in this study, the total number of offspring produced over the adult lifetime.

Local Hatchery Stock
Broodstock created from naturally spawning fish inhabiting streams that are the same as release location of artificially spawned juveniles.

Natural-Origin
As defined in this study, fish that were born in the natural environment regardless of the origin of parents or grandparents. Note: some studies reviewed here refer to fish as "wild" despite unknown origin of parents or grandparents, but we use "natural-origin" since "wild" implies lack of any ancestry from hatchery-origin fish which is typically unknown.

Nonlocal Hatchery Stock
Broodstock created from fish inhabiting streams that are different from release location of artificially spawned juveniles.

Relative Reproductive Success (RRS)
The average number of offspring produced by hatcheryorigin fish compared to the average number of offspring produced by natural-origin fish. Can be calculated as relative LRS. Also referred to as "relative fitness".

Resident Fish
Fish that do not migrate to the ocean but remain in freshwater to live and spawn.

Segregated Broodstock
Hatchery population created solely by crosses of hatcheryorigin fish as opposed to natural-origin. Typically observed in traditional hatchery programs.

Semelparity
A life-history strategy characterized by death following spawning in a single reproductive season. larger and older "hooknose" males observed in semelparous Pacific Salmon Oncorhynchus spp. (Allen et al., 2007;Gross, 1985;Quinn & Foote, 1994) and other anadromous salmonids, such as Atlantic Salmon. For example, in Atlantic Salmon, both size and age at maturity can vary widely, with some males maturing as precocial parr in their natal streams and using a sneaker strategy to fertilize eggs compared to a fighter strategy displayed in larger anadromous males (Fleming, 1996;Fleming & Reynolds, 2004). Precocial maturation also occurs in other iteroparous salmonid species, such as Steelhead Trout (Viola & Schuck, 1995;Willson, 1997). Similar to males, there are a few salmonid species that display precocial maturation of females as well and are commonly referred to as "jills" (Willson, 1997). Previous work has suggested that alternative reproductive phenotypes in salmonids are heritable and maintained through the processes of both negative frequency-dependent selection and condition-dependent sexual selection (Berejikian et al., 2010(Berejikian et al., , 2011Christie et al., 2018;DeFilippo et al., 2019;Fleming, 1996;Gross, 1985Gross, , 1996Heath et al., 2002;Reed et al., 2019;Taborsky, 2008;Tentelier et al., 2016). One proposed evolutionary explanation for the presence of mature male parr on the spawning grounds is that they can increase overall genetic diversity and reduce inbreeding since they are unlikely to mate with females from their own cohort (i.e., full-or half-siblings; Perrier et al., 2014).
Alternative migratory tactics also include other forms of lifehistory variation within salmonids that, similar to alternative reproductive behavior, are determined by both environmental influences and genetically based developmental thresholds Kendall et al., 2015) and can occur along a continuum both within and across populations (Jonsson & Jonsson, 1993). Anadromous salmonids spend part of their life in freshwater (postnatal development and spawning) and the other part in the ocean where they develop and grow, while resident fish spend the entirety of their lifecycle in freshwater. While the majority of species within the genus Oncorhynchus display anadromy (ex. Chinook, Chum, Coho, and Pink Salmon), other species such as Coastal Cutthroat Trout, Steelhead Trout, and Sockeye Salmon display both an anadromous and a resident form Pavlov & Savvaitova, 2008;Quinn & Myers, 2004;Trotter, 1989). Similarly, Salmo, Salvelinus, Thymallus, and Coregonus species display a continuum of anadromy and residency (Brenkman & Corbett, 2005;Dodson et al., 2013;Jonsson et al., 1988;Klemetsen et al., 2003;Larsson et al., 2013;Morin et al., 1982;Northcote, 1995;Pavlov & Savvaitova, 2008).
Multiple species of salmonids, including Atlantic Salmon, Steelhead Trout, Chinook Salmon, Chum Salmon, Coho Salmon, and Sockeye Salmon, are considered of high concern for conservation (International Union for Conservation of Nature, IUCN; U.S. Endangered Species Act, ESA; Canadian Species at Risk Act, SARA) owing to numerous anthropogenic impacts such as dam construction, overfishing, climate change, and habitat degradation (Boisclair, 2004;Gustafson et al., 2007). Estimating productivity of conservation species provides a vital component of population viability (Garcia de Leaniz et al., 2007;Waples & Hendry, 2008), with the success of conservation management frequently measured by LRS. Therefore, understanding the vast array of recent literature that has addressed salmonid reproductive ecology can provide conservation managers with a framework for constructing successful management plans and assessing success of existing conservation strategies.
Here, we synthesize peer-reviewed studies from the published literature over the last two decades (since the year 2000) that assessed the factors affecting adult-to-adult LRS (i.e., fitness) across salmonids that display a range of life-history strategies using parentage assignment from adult fish to their adult offspring with a few important caveats. First, our primary focus is on species from the genus Oncorhynchus, as the LRS literature is heavily biased toward these species, particularly due to their conservation status. However, we also provide studies that estimated LRS in the genus Salmo (see Box 2 for a glossary of the primary salmonid species presented in this review). Second, the majority of the studies examined in this review estimated LRS during the span of one breeding season for both semelparous and iteroparous species. However, while most LRS studies in iteroparous species did not account for repeat spawning, the occurrence of repeat spawning is quite low for anadromous salmonids (generally <10% in Steelhead Trout populations; Busby et al., 1996;Christie et al., 2018;Seamons & Quinn, 2010). Therefore, we provide the caveat that although an estimate of offspring number is a true measure of LRS in semelparous species, it may not necessarily accurately reflect LRS in iteroparous species which can spawn in more than one breeding season. Third, while the ultimate goal in salmonid fitness studies is to estimate adult-to-adult LRS for multiple generations, sampling juvenile offspring offers a more tractable alternative for studies that do not have capture locations to sample returning adult anadromous fish. Additionally, adult-to-juvenile LRS estimates can provide larger sample sizes in LRS studies because juveniles are sampled prior to outmigration and therefore are largely absolved from extraneous factors that could impact returning adult sampling and survival, such as straying or prespawn mortality (Keefer & Caudill, 2014). While there have been very few studies that have directly compared the factors affecting both adult-to-juvenile and adult-to-adult LRS within the same study system (Berntson et al., 2011;Ford et al., 2006Ford et al., , 2012Kostow et al., 2003), results suggest that adult-to-juvenile estimates provide comparable results and can be extremely informative for study systems that do not have access to adult-to-adult data (Berntson et al., 2011). Therefore, in addition to adult offspring produced from adult parents (i.e., adult-to-adult; Table 1), we also include results of juvenile offspring produced from adult parents (i.e., adult-to-juvenile; Table 2).

| THE IMPAC T OF HATCHERY ORI G IN AND C AP TIVE RE ARING ON LIFE TIME REPRODUC TIVE SUCCE SS
Understanding the effects of captive breeding is critical to the maintenance and perseverance of species that are of conservation concern (Williams & Hoffman, 2009). An essential component of numerous captive breeding programs is to minimize genetic adaptation to captivity and to confidently estimate the success of reintroducing captive-born animals back into their species range (Fischer & Lindenmayer, 2000;Frankham, 2008). Due primarily to the conservation status of multiple species of salmonids, the differences between fish born to parents that spawned in nature (i.e., naturalorigin) and those fish born in a hatchery setting from broodstock parents (i.e., hatchery-origin) have been extensively studied and reviewed in other papers (Araki et al., 2008;Christie et al., 2014;Naish et al., 2007). As such, we provide a brief summary of findings on the effects of origin on LRS and we defer to previously published review papers for further details on this topic, but also account for additional papers that have been published more recently than those reviews.
Based on author interpretations across salmonid studies and our examination of estimates of hatchery-origin compared to naturalorigin fish LRS (i.e., relative reproductive success; RRS; relative LRS, or relative fitness), we found trends that hatchery-origin fish consistently demonstrated lower LRS than natural-origin fish across species ( Figure 1; Anderson et al., 2013;Berntson et al., 2011;Evans et al., 2015;Ford et al., 2016;Kostow et al., 2003;McGinnity et al., 2003;McLean et al., 2003McLean et al., , 2004Milot et al., 2013;Neff et al., 2015;O'Sullivan et al., 2020;Sard et al., 2015;Thériault et al., 2011;Williamson et al., 2010) as discussed in previous reviews (Araki et al., 2008;Christie et al., 2014;Naish et al., 2007). However, previous research suggests that relative LRS between hatchery-and naturalorigin fish is dependent on the type of hatchery or supplementation program under examination. Results in Steelhead Trout (Araki et al., , 2009Christie et al., 2012)  Significantly lower LRS for captive-bred compared to wild-bred fish in five out of six spawning cohorts with males and females combined. When sexes were separated, female captive-bred LRS was significantly lower in four spawning cohorts and in two spawning cohorts for males. Annual population productivity was lower in years where captive-bred fish represented a greater proportion of potential spawners Chinook Salmon (Oncorhynchus tshawytscha) Ford et al. (2012) Negative relationship between the LRS of individuals bred in captivity and the LRS of their male offspring in the wild. Did not find same result for female offspring. Includes additional estimates of adult-to-juvenile LRS Hess et al. (2012) No significant differences in LRS of hatchery-origin compared to natural-origin fish, but trends for lower LRS for hatchery-origin males Anderson et al. (2013) Hatchery-origin males demonstrated nonsignificant trends for lower LRS than natural-origin males, but LRS for females varied between years. Larger males and females demonstrated trends toward higher LRS. Fish that arrived early, compared to later, to the spawning grounds had higher LRS, but in some years, LRS was highest at intermediate dates Evans et al. (2015) Sex, release date, and the interaction between the two significantly predicted LRS. The effect of origin was close to significant with hatcheryorigin fish exhibiting lower fitness than natural-origin fish  (2007) Hatchery-origin females demonstrated lower LRS than natural-origin females. Hatchery females that were spawned earlier in the season had significantly higher LRS, but there was no effect on spawning date for hatchery males. No relationship between LRS and female or male length. Araki, Ardren, et al. (2007), ) Evaluated LRS in both a traditional hatchery system with no natural-origin fish incorporated into a nonlocal broodstock and an integrated supplementation program with a designated proportion of natural-origin fish incorporated into broodstock. Hatchery-origin fish (both from a summer and a winter stock) demonstrated lower LRS than natural-origin fish if the fish originated from the traditional hatchery program. No differences were observed in hatchery-origin fish from the supplementation program. Compared the LRS between natural-origin females that crossed with either natural-origin males, traditional hatchery-origin males, or supplementation hatchery-origin males and found no statistically significant effect of male type; trend for crosses involving traditional hatchery-origin males to have lower LRS than crosses involving naturalorigin males. Compared the LRS between crosses involving fish with different levels of captive breeding in the previous generation. Found a decline in LRS for each successive generation reared in captivity; fish with two hatchery-origin parents had significantly lower LRS than a fish with two natural-origin parents; fish with only one hatchery-origin parent had intermediate, but nonsignificant, LRS between crosses involving two hatchery-origin parents and those involving zero hatchery-origin parents. Seamons et al. (2007) Evidence for directional selection for larger body length across sexes and years (although some variation still existed). Evidence for directional selection on early male arrival date, but overall arrival date varied across years for both sexes, providing evidence of stabilizing selection. Berntson et al. (2011) Lower LRS of hatchery-origin females and males. Generalized linear models (GLMs) showed that LRS was significantly affected by return date, length, origin, and the number of same-sex competitors. Larger and earlier-returning fish demonstrated higher LRS, but there was also evidence for stabilizing selection on return date. Included additional estimates of adult-to-juvenile LRS. Christie et al. (2012) For the majority of run years, individuals from larger broodstock families demonstrated lower LRS than those producing fewer offspring (<5 offspring). Hatchery-origin broodstock (1 generation in captivity) outperformed wild-origin broodstock (no generations in captivity). Adaptation to captivity can occur over one generation with a trade-off between high performance in captivity and subsequently low performance in the wild.

Christie et al. (2018)
Estimated LRS from both repeat-spawning and single-spawning hatchery-and natural-origin Steelhead Trout. The majority of repeat-spawning females increased in both length and weight between their first and second spawning. Repeat spawners were more likely natural-origin compared to hatchery-origin. Natural-origin repeat spawners of both sexes had more than 2.5 times the number of offspring than single spawners. First-time repeat spawners had lower LRS than same-age single spawners, but higher LRS than same-age single spawners during the second spawning event, suggesting a trade-off between reproduction and survival. For single-spawning fish, LRS increased with male age and size, but for females, there was evidence for negative frequency-dependent selection on age.
Note: Terms used to describe general findings follow those used in each paper. Within each species, studies are presented first by order of publication date and second by alphabetical order. Abbreviation: LRS, lifetime reproductive success.

TA B L E 1 (Continued)
TA B L E 2 Summary of studies estimating lifetime reproductive success (adult-to-juvenile) across species

General findings
Atlantic Salmon (Salmo salar) Garant et al. (2001) No relationship between size and LRS for males or females, but there was a significant relationship between number of mates and LRS Landry et al. (2001) There was no evidence for inbreeding avoidance. Mating patterns appeared random overall. Mates chose each other to increase heterozygosity and increase immune defenses of their offspring, as determined by significant differences in the amino-acid composition of the MHC peptide-binding region between partners LRS was affected by body size, male tactic, and the interaction between the two. Larger anadromous males had higher LRS than smaller males, but the opposite effect was observed for precocious males Mobley et al. (2019) Both females and males that return to their natal spawning grounds had higher LRS compared to strays (termed "dispersers"). Higher LRS for males and females with older sea age at maturity Prévost et al. (2020) Number of mates significantly predicted LRS among adults that successfully reproduced in one of the two tributaries examined. Both males and females that mated with two or more partners had higher LRS than individuals that only mated with one partner. Body mass, sex, and arrival date were not significant predictors of LRS Mobley et al. (2020) Negative relationship between LRS and freshwater age for females, but not males. Positive relationship between LRS and sea age for both males and females. More time spent in juvenile freshwater habitat was associated with reduced LRS for females but not males. More time spent in marine habitat was associated with increased LRS for males and females. Overall, younger freshwater age was significantly related to older sea age and subsequent increased LRS for females but not males  Hatchery-origin fish were younger on average and had significantly lower LRS (Continues)

General findings
Evans et al. (2012) Both male courting and dominance were positively associated with LRS. Male body size was negatively associated with LRS, but there was no relationship between body size and LRS for females. LRS was not affected by intersexual aggression or dominance in females. Aggression toward females did not significantly predict LRS Ford et al. (2015) Naturally spawning resident, early-maturing, hatchery-origin males from a captive broodstock program likely represented a large proportion of missing male parents. Anadromous male LRS was significantly higher than LRS of resident males. In general, females spawned with a single anadromous male and multiple resident males, which demonstrated lower per-capita spawning success than the anadromous males Sard et al. (2015) Hatchery-origin males (not females) had lower LRS than natural-origin males, but origin was not significant after accounting for body length.
Overall positive relationship between length and LRS for both sexes.

TA B L E 2 (Continued)
that differences between hatchery-and natural-origin fish in certain systems could, in part, be attributed to hatchery-origin fish originating from nonlocal origin broodstock and maintained as part of a segregated hatchery system. While reduction of fitness has been documented after a single generation of captive rearing (Araki et al., 2008;Milot et al., 2013), genetic divergence of hatchery strains is expected to be much more rapid in segregated programs with multiple generations of spawning hatchery-origin fish compared to integrated programs that incorporate natural-origin fish (Ford et al., 2016;Paquet et al., 2011;Waters et al., 2015). Further, adaptation to captivity can occur in very few generations (Christie et al., 2012), so repeated generations of hatchery rearing would be expected to strengthen domestication selection without input from naturalorigin stocks as predicted by previous models (Baskett & Waples, 2013;Ford, 2002). Similarly, the length of time that hatchery-origin fish are reared in a hatchery setting may negatively affect LRS (Berejikian et al., 2020). Studies in integrated programs have found little to no significant differences in LRS of crosses containing a hatchery-origin parent compared to those containing two naturalorigin parents (Ford et al., 2012;Hess et al., 2012), even after two generations (Janowitz- Koch et al., 2019), providing evidence that there was no reduction in LRS for natural-origin fish that spawn with hatchery-origin fish. However, it also worth noting that in integrated programs that aim to incorporate hatchery-origin fish into the naturally spawning population, RRS estimates may become upwardly biased due to overall reductions in long-term fitness of the whole population as the proportion of hatchery ancestry increases over time (Willoughby & Christie, 2017). Thus, hatchery programs must carefully weigh goals for conservation versus production when considering spawning and rearing protocols that can lead to varying degrees of hatchery ancestry and domestication.
A second pattern that was detected in our review indicated that differences in LRS for hatchery-origin fish are influenced by precocial males (e.g., jacks) that tend to have low LRS. Across salmonid studies that analyzed precocial males separately, the difference between hatchery-versus natural-origin male and hatchery-versus natural-origin female LRS appeared smaller (e.g., Hess et al., 2012;Janowitz-Koch et al., 2019;Thériault et al., 2011;Williamson et al., 2010), and in some cases, female relative LRS was lower than male relative LRS overall (Figure 1; Hess et al., 2012;Janowitz-Koch et al., 2019). These trends could be the result of different selective pressures on males that adopt a sneaker strategy compared to males that compete for access to females (Thériault et al., 2011). These trends also suggest that inclusion of precocial males across salmonid species can affect estimates of relative LRS, and when possible, the addition of a separate analysis for precocial males could help to determine the effect on LRS patterns in a population. Overall, given the wide range of life-history variation across salmonids (e.g., resident/anadromous migration, semelparity/iteroparity, premature/mature migration, age at maturity, age at juvenile emigration), disentangling domestication selection due to captive rearing rather than unintentional artificial selection due to inability to account for complex life-history variation and lack of mate choice (Auld et al., 2019) will provide a clearer picture on the effects of captive breeding.

| Adult migration timing
While animal migration is a well-documented phenomenon across the taxonomic spectrum, an inherent complexity exists in understanding how best to manage species of conservation concern with a geographic range that encompasses multiple habitats (Milner-Gulland et al., 2011  The term "migration timing" is frequently used interchangeably with the terms "run timing," "arrival timing," and "return timing"; however, these terms may describe distinct phenotypes that refer to different time points in the migratory trajectory, including entry into freshwater, a common passage point along the migratory corridor, and/or arrival onto the spawning grounds (Quinn et al., 2016).
Therefore, we use the general term "migration timing" throughout this manuscript to refer to sampling date at a common migration location but provide clarification on specific migratory phenotypes where appropriate.
General trends in the literature suggest fish that migrate earlier to the spawning grounds demonstrate higher LRS compared to those that arrive later (  Morbey & Ydenberg, 2003;Quinn et al., 2009Quinn et al., , 2016. Our review of the literature continues to support this finding across species, where males and females that returned earlier to the spawning grounds generally had positive LRS effects compared to those that arrived later. While there were strong trends for an advantage of early migration, variation still existed in the overall direction and strength of selection on migration timing. For example, Kodama et al. (2012) used selection gradients to estimate the strength and direction of selection across multiple age classes, sexes, and years. Selection differentials measure the mean phenotype both before and after selection, which can be used as a metric of both strength of direct and indirect selection, while selection gradients measure the strength of direct selection on a trait while removing indirect selection of other traits (Arnold & Wade, 1984;Brodie et al., 1995;Falconer & Mackay, 1996;Lande & Arnold, 1983). Kodama et al. (2012) showed that 2-year-old Coho Salmon males (jacks) from one broodyear demonstrated significant evidence for disruptive selection on migration timing (Table 3). However, for 3-year-old males, there was evidence for selection favoring early migration timing, and in females, there was evidence for stabilizing selection in one broodyear (Table 3). Similarly, Ford et al. (2006) estimated selection differentials on migration timing in Coho Salmon and found evidence of stabilizing selection on migration timing for both sexes (Table 3). Another study did not find a significant effect of migration timing in jack Coho Salmon males, which may possibly reflect a lack of selection on migration timing in jacks which tend to employ sneak fertilization attempts, rather than direct competition for ac-

| Effects of age and size
Body size is a ubiquitous trait across the animal kingdom that correlates with numerous physiological, ecological, and life-history processes and is driven by both sexual and natural selection (Andersson, 1994;Berns, 2013;Blackburn & Gaston, 1994   TA B L E 4 (Continued) (McLean et al., 2007), or the experiment was conducted under artificial conditions (Berejikian et al., 2005), for example.

| Resident versus anadromous
There have been a limited number of studies comparing LRS differences between resident and anadromous life-history forms within a population, which are inherently correlated with body size. However, studies noted sex-specific effects of residency versus anadromy on LRS that provide insight to the maintenance of both life-history types. One of the most intriguing studies comparing LRS between anadromous versus resident life-history types in Brook Charr found that anadromous females had higher LRS than resident females, with results driven by the larger size of anadromous compared to resident females (Thériault et al., 2007). The authors posited that the tactic of residency may be beneficial and therefore persist, where small tributary streams are easily accessible to smaller residents but could exclude larger anadromous females (Thériault et al., 2007). In the same study, there were no observed differences in LRS between resident and anadromous males, a potential reflection of opportunistic behavior, such as sneaking by smaller resident males rather than fighting (Thériault et al., 2007). Other studies in Steelhead Trout revealed only a very small number of offspring were assigned to resident parents overall (Berntson et al., 2011), which reflected both logistical constraints regarding sampling resident fish (resulting in a large fraction of missing resident parents) and the overall low assignment success of hatchery-origin resident fish (Berntson et al., 2011).

| Local adaptation
Numerous salmonid species demonstrate high fidelity to natal spawning sites Quinn, 1993). Higher fitness in the local habitat (i.e., natal spawning sites compared to foreign sites in salmonids) can affect spatial distribution and genetic diversity and can promote reproductive isolation and is, therefore, a central theme in animal conservation (Fraser et al., 2011;Kawecki & Ebert, 2004;Savolainen et al., 2013;Taylor, 1991). However, although numerous studies have measured survival differences between salmonids that stray from their natal spawning sites and those that return, very few studies have actually estimated differences in LRS, a key component of adaptive variation and evolution (Fraser et al., 2011;Garcia de Leaniz et al., 2007). Strays among distinct lineages or geographically distant populations are expected to be relatively rare and have low LRS (Hess et al., 2014;Keefer & Caudill, 2014;Quinn, 1993).
However, straying among streams within the same drainage is common (Ford et al., 2015;Keefer & Caudill, 2014) and colonization of newly available habitat (Anderson et al., 2013) may be beneficial for preventing extirpation of local populations (Hill et al., 2002).
Further, different ecotypes that occur within the same system may be under selective pressure at temporal or fine geographic scales resulting in outbreeding depression (Gharrett et al., 1999) and introgression (Hess et al., 2011) when ecotypes are interbred artificially, but little is known about LRS among salmonid ecotypes. A study of Sockeye Salmon ecotypes in southwest Alaska found that dispersers from beach habitat to stream habitat (i.e., immigrants) had significantly lower LRS than fish spawning in their natal stream or fish spawning in another stream (Peterson et al., 2014). The authors provided potential mechanisms shaping differences in LRS in this study, including morphological maladaptation and reduced survival of hybrid offspring. Similarly, Atlantic Salmon that were native to spawning grounds demonstrated higher LRS compared to those dispersing from neighboring areas (Mobley et al., 2019). On a broad scale, these results suggest that differences in LRS between immigrant and philopatric fish can serve as a barrier to reduce gene flow between populations and thus further reinforce local adaptation.
However, the maintenance of gene flow, even at low levels observed in the Peterson et al. (2014) study, can still promote adaptive genetic diversity within populations, a key component of salmonid evolution, persistence, and conservation policy decisions (Waples, 1991).

| Spawning behavior
Animal behavior is a central theme in ecology and evolution that spans across species, involving both intrasexual and intersexual interactions (Alcock, 2001;Dugatkin, 2020). Numerous behavioral factors, such as dominance, courtship, and density, can affect LRS in salmonids. For example, Berntson et al. (2011) used the number of same-sex competitors on the spawning grounds of Steelhead Trout as a proxy for competition. Berntson et al. (2011) found that the number of same-sex competitors negatively affected LRS and that females, in particular, were negatively affected by a greater female density on the spawning grounds. Previous research has suggested that breeding density on the spawning grounds is the primary driver of female competition in salmonids with females competing for nest sites, displacing other females, and disturbing nests (i.e., "redds"; Fleming van den Berghe & Gross, 1989). However, while some studies have demonstrated a relationship between spawner density and redd superimposition (Beard Jr & Carline, 1991;Fukushima et al., 1998), others have not (Essington et al., 1998;Gortázar et al., 2012;Peterson et al., 2020).
Aggressive or dominant mating behavior among males on the spawning grounds has also been examined across a limited number of studies, with a general positive relationship of LRS with increasing aggressiveness. Aggressive behaviors in males that have been tied to higher LRS include dominance in courting behaviors (Dickerson et al., 2005;Evans et al., 2012), attack frequency to competitors on spawning grounds (Schroder et al., 2010), and frequency of courting attempts with females (Evans et al., 2012;Schroder et al., 2010).
However, effects of aggressive male behavior may diminish later into the spawning season, potentially as a result of a shift in the operational sex ratio that could cause a reduction in the ability of larger males to maintain access to spawning females through a decrease in physical condition over time, the overall completion of female egg deposition, and an increase in the arrival of new spawners (Dickerson et al., 2005;Quinn et al., 1996). A positive relationship between the number of mates and LRS (Dannewitz et al., 2004;Garant et al., 2001;Haddeland et al., 2015;Prévost et al., 2020;Richard et al., 2013) and a negative relationship between the number of days spent on the spawning grounds and LRS (Schroder et al., 2010) have also been documented.  (Bernatchez & Landry, 2003;Milinski, 2006;Tregenza & Wedell, 2000). In salmonids, general trends in the literature suggest that there are higher levels of population differentiation at MHC genes compared to neutral genes (reviewed in Bernatchez & Landry, 2003 Salmon (Garner et al., 2009) and Atlantic Salmon (Evans et al., 2011;Landry et al., 2001)

| Hybridization with non-native species
Hybridization and introgression with non-native species of fish with distinct life histories can also affect LRS. In general, hybridization plays an important role in evolution either constraining the evolution of new species or in propelling diversification (Arnold, 1997;Mayr, 1963) and can have important implications for setting conservation policies (Allendorf et al., 2001). For example, in one study, the LRS of native Westslope Cutthroat Trout significantly declined for both males and females with increasing admixture from nonnative Rainbow Trout (Muhlfeld et al., 2009). Likewise, hybrid offspring from introduced Brook Trout and native White-spotted Charr showed significantly lower LRS compared to their parental species (Fukui et al., 2018).
While there are few studies directly evaluating LRS related to local adaptation in salmonids, results support expectations that native fish that are adapted to local environments have higher fitness than strays or non-native species. However, this is a rich area in need of further study to better estimate fitness related to local adaptation. Only a limited number of studies in salmonids have examined the differences in LRS between resident and anadromous fish (Berntson et al., 2011;Christie et al., 2011;Ford et al., 2015;Thériault et al., 2007) due to sampling challenges in experimental design. In species with both anadromous and resident forms, sampling efforts have been directed almost exclusively on returning anadromous fish due to low densities and practical challenges of sampling residents, resulting in a large fraction of "missing" resident parents (Araki, Ardren, et al., 2007;Christie et al., 2011;Seamons et al., 2004b). One potential method to circumvent incomplete sampling of resident fish that has recently received attention is to trace anadromous offspring back to their grandparents, essentially filling in the blanks of a pedigree that has complete sampling of anadromous, but not resident, fish (Christie et al., 2011;Ford et al., 2015;Sard et al., 2016).

| Accounting for various life-history strategies
There have also been very few studies that have examined fitness differences between semelparous and iteroparous life-history types that occur within the same species (Christie et al., 2018;Seamons & Quinn, 2010). In particular, salmonids that display iteroparity (i.e., repeat spawners) are greatly under-represented in LRS studies. Since very few studies have been able to account for repeat spawners, it is possible that iteroparous individuals that either minimized energy investment and subsequent reproductive output in the first breeding season and/or failed to reproduce during the second breeding season display disproportionally low LRS compared to the semelparous individuals in the same population, potentially lowering overall population-level trends of LRS (Seamons & Quinn, 2010). Two studies in Steelhead Trout showed that LRS of iteroparous fish averaged more than twice the LRS of those spawning one time, suggesting that iteroparous fish can increase population abundance overall (Christie et al., 2018;Seamons & Quinn, 2010).
Despite logistical challenges of assessing LRS in iteroparous species, increasing the incidence of iteroparous fish within a population has become a tool to increase overall genetic variability and population abundance, particularly for declining populations of Steelhead Trout (Copeland et al., 2019;Hatch et al., 2013;Narum et al., 2008).
Therefore, it is important and necessary for the breadth of LRS studies to continue to expand across a wide range of life-history types, especially as these studies help to inform broad-scale management and conservation decisions for iteroparous species.
Life-history traits such as adult migration timing and duration spent in freshwater versus ocean are often variable within salmonid populations (Quinn et al., 2016) but have also largely been unaccounted for in studies of LRS with only a few recent exceptions (Ford et al., 2016;Janowitz-Koch et al., 2019;Mobley et al., 2020).
Given that genes of major effect have been shown to drive phenotypic variation for these traits across species of salmonids (Barson et al., 2015;Prince et al., 2017;Waters et al., 2021)

Pathogens
Salmoninae: -salmon/trout -charr -lenok that are heterozygous for genes of major effect relative to alternative homozygous fish for phenotypic traits such as adult migration timing where one of the life-history types is under high conservation concern (i.e., spring-run Chinook Salmon; Thompson et al., 2019). In general, we expect that future studies of LRS will need thorough study designs that account for life-history types and traits.

| Expanding studies across salmonid species
It is also important to point out that the majority of LRS studies have

| Estimating reproductive success across multiple generations
An important area of research that is beginning to receive more attention in salmonid LRS studies is estimating fitness across multiple generations. However, only a small number of studies have utilized grandparentage in estimating salmonid LRS using either genetic exclusion methods (Christie et al., 2011;Ford et al., 2015;Sard et al., 2016), maximum likelihood approaches (Letcher & King, 2001), or in instances where long-term pedigree data are available, directly tracing pedigrees over consecutive generations (Janowitz-Koch et al., 2019). Grandparentage has the potential to provide estimates of long-term fitness effects, particularly in hatchery-versus natural-origin RRS studies, and further could help to determine levels of introgression between hatchery and native fish. For example, Janowitz- Koch et al. (2019) provided evidence that LRS did not significantly decline for natural-origin fish that spawned with hatcheryorigin fish, even after the offspring of these crosses were tracked for two generations. Although very useful and informative, constructing multigenerational pedigrees presents logistical challenges, with some programs not able to sample fish for multiple generations, particularly for longer lived salmonids. In addition to practical challenges, grandparentage assignment requires large and diverse genetic marker panels (Letcher & King, 2001), which have not yet been developed for many salmonid species. With the further development of grandparentage software programs (e.g., Huisman, 2017), the long-term fitness effects of factors such as captive breeding and anthropogenic impacts for salmonids will be further explored and potentially integrated into management practices.

| Exploring the effects of environmental factors and anthropogenic impacts
Numerous studies reviewed here reported large differences in LRS across years (e.g., Anderson et al., 2013;Dickerson et al., 2005;Janowitz-Koch et al., 2019;Kodama et al., 2012;Sard et al., 2015;Seamons et al., 2004aSeamons et al., , 2007Thériault et al., 2011). There are numerous factors that could explain these interannual differences including ocean conditions (e.g., upwelling, sea surface temperature), freshwater environment at natal sites and through migratory corridors (e.g., water temperature, precipitation rates, changes in water velocity), and dam passage effects (both downstream and upstream) that are likely shaping these differences in fitness and survival (National Research Council, 2004). Water temperature and water flow, in particular, can strongly affect salmonid development, growth, and survival across all stages of the life cycle (Jonsson & Jonsson, 2009;Pankhurst & Munday, 2011). For example, elevated levels of stress and mortality in early-entry late-run Sockeye Salmon are directly linked to warm river temperatures that are at, or near, thermal maxima of the species (Hinch et al., 2012). During ocean phases of the life cycle, environmental conditions such as upwelling of cold water can directly affect food chain structure and subsequent growth and survival in salmonids (Bi et al., 2011;Black et al., 2011;Emmett et al., 2006;Fisher & Pearcy, 1988;Peterson & Schwing, 2003;Scheuerell & Williams, 2005). Other ocean-related factors such as Pacific Decadal Oscillation can affect salmonid ocean abundance and subsequent returns to freshwater (Hare et al., 1999;Mantua & Hare, 2002;Mantua et al., 1997;Peterson et al., 2010).
Environmental variables can also interact with both demographic variables and phenotypic traits, such as density, size, or origin, to predict survival in salmonids (Bowerman et al., 2021;Crozier et al., 2008;Zabel et al., 2006). While we recognize the need for these types of environmental data in studies predicting fitness, we also understand the potential difficulty in obtaining these metrics during field sampling. One potential source for obtaining environmental data is through pre-existing databases, such as those maintained by federal agencies (e.g., Huang et al., 2017;Isaak et al., 2017) or spatial interpolated climate data across the earth collected from satellites (e.g., Fick & Hijmans, 2017).
Direct anthropogenic impacts are another type of aquatic stressor that can shape patterns in LRS. For example, trace heavy metals, pesticides, and herbicides could all affect survival and fitness in salmonids (Milner et al., 2003;Wedemeyer et al., 1980). The impact of recreational fishing techniques should also be expected to affect LRS, yet studies in the literature are lacking. One example of a recreational fishing technique is "catch and release" (CR), a technique that involves releasing live fish back into the water after catching them. The use of CR, and the subsequent likelihood of mortality, remains controversial from a cultural, ethical, and biological perspective (Arlinghaus et al., 2007). Air exposure during CR has been shown to negatively affect LRS; however, the relationship between LRS and air exposure may be dependent on other factors such as water temperature and size of the fish (Richard et al., 2013). Other studies have shown no effect of air temperature on LRS (Roth et al., 2019); the differences in results potentially attributed to differences in sample sizes. Therefore, the impacts of human activities, such as fishing, on LRS remains an area of research that is largely untested.

| The effect of general body condition and pathogens on reproductive success
Numerous infectious diseases and parasites have been extensively documented in salmonids (Bakke & Harris, 1998). However, there have been very few studies measuring the impact of disease on reproductive success, particularly in natural salmonid populations, despite the fact that heritable variation has been documented in these factors and they can directly impact fitness-related traits, such as growth, feeding behavior, swimming, and osmoregulation (Garcia de Leaniz et al., 2007;Mendel et al., 2018;Miller et al., 2014;Yáñez et al., 2014). While lethal sampling has been employed in numerous studies assessing pathogens in salmonids, other studies have employed nonlethal methods (e.g., Elliott et al., 2015;Fernández-Alacid et al., 2018;Kittilsen et al., 2009;Rees et al., 2015). The use of noninvasive and nonlethal sampling across salmonid studies remains a priority for state, federal, and tribal fisheries agencies, and much progress has been made on developing these sampling techniques (Coble et al., 2019;Lawrence et al., 2020;Teffer & Miller, 2019). Therefore, an assessment of the effect of pathogens on reproductive success in salmonids is a tractable area of research and remains an area of high need. Similarly, the effect of general physiological condition on reproductive success, particularly an overall assessment of energy reserves, is another area of research that is lacking and recent advances in technology have allowed for noninvasive sampling of body condition in salmonids (Hanson et al., 2010). By including factors such as pathogen load and general body condition, a clearer understanding of the aquatic stressors affecting fitness in salmonids will more fully develop. Of particular importance would be determining whether these factors interact with other factors to predict LRS, such as origin or environmental conditions, which could help within the broader context of successful conservation management.

| Explaining variation in results across studies and within studies
While the review presented here demonstrates general patterns and trends of factors affecting LRS, there is a substantial amount of variability in results across studies. One obvious potential source of variability is the vast array of factors, or lack thereof, used to predict LRS across studies. For example, as discussed previously, while many studies estimating differential LRS of hatchery-and naturalorigin fish accounted for body size and/or age (Anderson et al., 2013;Berntson et al., 2011;Ford et al., 2012Ford et al., , 2016Janowitz-Koch et al., 2019;Sard et al., 2015;Thériault et al., 2011;Williamson et al., 2010), several others did not, which could greatly impact the overall results of these types of studies. Several studies provided evidence that hatchery-origin fish in some years were smaller than naturalorigin, which may, in part, help to explain the differences in LRS that have been observed between hatchery-and natural-origin fish.
For male salmonids, while some studies have found that hatcheryorigin precocial males may not experience the same fitness declines as hatchery-origin males of older age classes (Garant et al., 2003;Thériault et al., 2011), other studies show opposite effects, with hatchery-origin precocial males exhibiting significantly lower LRS than natural-origin precocial males (Hess et al., 2012;Janowitz-Koch et al., 2019). These results provide a clear example that age and/or size can interact with origin to predict fitness, and both should be accounted for in these types of studies whenever possible. However, it is also worth noting that including additional factors in models may not necessarily explain additional variability in salmonid LRS studies.
One method to determine assessment of model fit in a study is estimating the proportion of variation in the response variable that can be explained by the model. However, out of the numerous studies that we reviewed here, only a small number of studies provided estimates on proportion of variability in LRS explained by models.
For example, the total variance in LRS models (as reported by the coefficient of determination estimates, R 2 ) explained by body size was the only trait that was consistently reported for a few studies, yet exhibited marked variation ranging from R 2 < 2%  to 46% (Thériault et al., 2007) for female Steelhead Trout and female Brook Charr, respectively. For males, variation explained by body size was also variable, ranging from <3% in Steelhead Trout  to 26% in Chinook Salmon (Schroder et al., 2010) with results for Chum Salmon males somewhere in between (approximately 13%-19%; Berejikian et al., 2009). Similarly, the proportion of variation on LRS explained by migration timing was only provided in a single study . Although major conclusions cannot be drawn across the limited number of studies provided here, the percent of variation in traits explaining LRS models can provide important statistical and biological information within studies that could perhaps be generalizable to similar study systems.
It is also possible that general differences in model results both within studies (across years) and between studies could be the result of populations experiencing different types and varying strength of selection, sometimes experiencing weak selection and residing in a stable optimum, other times facing strong directional selection and shifting optima, for example. Selection fluctuations can be caused by numerous factors in salmonids, including environmental variables, changing frequencies of life-history forms, sex ratio, competition, and overall spawner density, to name a few (Anderson et al., 2010;Christie et al., 2018;Dickerson et al., 2005;Ford et al., 2008;Kodama et al., 2012;Seamons et al., 2007). These factors are likely to shift within populations across years, which could explain the variability in results seen across years within the same study. Similarly, these factors would be expected to be different across populations with varying geographical landscapes and different genetic and historical backgrounds. Therefore, differences across studies in traits predictive of fitness or a lack of significance in certain traits that should be expected to predict fitness, such as body size, could also be the result of differences in temporal or spatial selective pressures.

| Biological significance and effect sizes
Although we provide compelling evidence that numerous demographic, phenotypic, phenological, environmental, and behavioral factors predict fitness, we rarely uncovered a discussion of biological significance in literature reviews. While statistical significance is an important part of drawing experimental conclusions, it is only one component of biological significance. Biological significance is broadly defined as a biological effect, or the size of a biological effect, that is biologically meaningful based on expert opinion and can have important implications in real-world applications, potentially impacting decisions regarding conservation and management policies (EFSA Scientific Committee, 2011). The size of a biological effect (i.e., effect size) that would be considered relevant or meaningful should be defined a priori through methods such as power analyses (Martinez-Abrain, 2008;Steidl et al., 1997;Taylor & Gerrodette, 1993). The issue of power has been addressed in other review papers in salmonids, particularly in regard to limitations in power in hatchery-versus natural-origin LRS studies (Araki et al., 2008;Christie et al., 2014). However, across the studies reviewed here, only a limited number provided either a priori or retrospective power analyses to determine the minimum effect size that would be detectable under varying degrees of power (Araki, Ardren, et al., 2007;Berejikian et al., 2009;Hess et al., 2012;Mobley et al., 2020;Thériault et al., 2011).
Confidence intervals are often favored over power analyses to convey information on the range of effect sizes that are supported by the data (Colegrave & Ruxton, 2003;Lovell, 2013); yet again, these were not consistently presented across the studies reviewed here (e.g., Anderson et al., 2013;Christie et al., 2018;Janowitz-Koch et al., 2019;Roth et al., 2019). Ideally, studies assessing fitness in salmonids would demonstrate both statistical and biological significance. However, other scenarios typically unfold, such as statistical significance without biological significance or biological significance without statistical significance. Both scenarios provide valuable information to researchers, each lending information regarding study limitations and directions for future research, such as sample sizes that are too small to detect effects or variability in traits that is too wide to be explained by a single variable of interest (Lovell, 2013; Martinez-Abrain, 2008). Therefore, demonstrating power to detect effects is an extremely useful tool, particularly in conservation management scenarios where sample sizes tend to be small and accepting a false null hypothesis based on a single α cutoff value could have substantial consequences (Taylor & Gerrodette, 1993).

| Limitations in pedigree reconstruction
The ability to sample all potential parents in any given year remains a challenge in multiple salmonid populations namely due to logistical and technical constraints, sometimes in conjunction with conservation limitations. The decreased ability to reconstruct pedigrees due to incomplete sampling may be particularly problematic for datasets with an insufficient number of genetic markers as investigated previously (Aykanat et al., 2014;Harrison et al., 2013). Additionally, Araki and Blouin (2005) demonstrated that an increase in incomplete sampling can increase the overall proportion of incorrectly assigned parents. Numerous parentage programs account for the proportion of candidate parents sampled in a study, with some parental reconstruction methods more sensitive to this variable compared to others (Jones et al., 2010). Similarly, relatedness between potential parents can affect parental assignment success, with a reduction in the overall assignment success of offspring whose parents are more related and less heterozygous overall Olsen et al., 2001). Therefore, to increase the validity of results in salmonid LRS studies, it remains particularly important to choose appropriate parental reconstruction methods, provide estimates of power limitations and potential sources of error, such as genotyping or measurement error, determine the appropriate thresholds for accepting assignments, and maximize genomic marker panels whenever possible (see e.g., Flanagan & Jones, 2019 which reviews next-generation genotyping approaches in parentage analysis).

| Applications for conservation management
Overall, estimating productivity through LRS-based studies remains a vital component of protecting and maintaining declining salmonid populations. General declines in salmonids are namely thought to be the result of anthropogenic impacts, including but not limited to habitat loss, overharvest, unintended effects of hatcheries, and hydropower dams, each with potential to propel rapid evolutionary shifts (McClure et al., 2003;Stockwell et al., 2003;Waples & Hendry, 2008). In this review, we uncovered extensive variation in salmonid LRS and in the factors affecting LRS, with a heavy emphasis on factors related to hatchery rearing. We see, however, that salmonid life history adds further complexity to patterns in LRS and can interact with other variables to predict LRS. Density and environmental factors, for example, are innately related to anthropogenic impacts, and numerous studies presented in this review suggest that these factors may directly affect LRS or can interact with life history to shape patterns in LRS. Therefore, while the complexity in life-history forms and variation in LRS make salmonids ideal study species to address evolutionary questions, this can further complicate conservation questions. As such, the inclusion of the vast array of life-history strategies, environmental drivers, and both phenological and phenotypic traits in studies estimating salmonid LRS remains extremely important in synthesizing both species-specific and population-specific conservation decisions.

| CON CLUS IONS
In summary, studies of LRS have provided critical insights toward understanding fitness advantages within salmonid populations and have expanded our general knowledge of salmonid mating behavior and reproductive strategies. Further, these studies have provided a basis for ongoing conservation and management decisions for salmonid species that have vital ecological, economic, and cultural roles throughout their geographic ranges. Future studies should aim to (1) continue to expand the breadth of studies assessing LRS across different life-history strategies, specifically accounting for different reproductive and migratory phenotypes, (2) broaden the array of species represented in salmonid fitness studies, (3) construct multigenerational pedigrees to measure long-term fitness effects, (4) expand on LRS studies that investigate the effects of largely untested traits, such as aquatic stressors including, but not limited to, environmental factors, pathogens, and general body condition, and (5) utilize appropriate statistical approaches to determine the factors that explain the greatest variation in fitness models and provide details on biological significance, power limitations, and potential sources of error whenever possible. These challenging studies have profoundly advanced scientific understanding that will continue to assist with long-term perseverance of these keystone species in aquatic ecosystems.

ACK N OWLED G EM ENTS
We would like to thank the anonymous reviewers for providing valuable, thoughtful, and insightful feedback on our manuscript. We would also like to acknowledge Bonneville Power Administration for funding this research.

DATA AVA I L A B I L I T Y S TAT E M E N T
Data sharing is not applicable to this article as no new data were created or analyzed in this study.