Reproductive success in the wild.

Differential reproductive success of residents and immigrants (strays) in a wild population.
    Andrew Hendry, John Wenburg, Eric Volk, and Thomas Quinn.

We demonstrate that sockeye salmon populations can exchange many migrants each generation and yet remain genetically distinct, owing to reduced reproductive success in strays.  We studied a small beach population that receives strays each generation from a much larger river population (both in Lake Washington, Washington).  Site-specific otolith microstructure patterns were used to determine which beach spawners had been born at the beach (residents) and which had been born in the river (strays).  In each of two years, about 1% of the river population strayed to the beach but these strays composed 35 - 44% of the beach spawners.  If strays and residents had similar reproductive success, such levels of gene flow would prevent any neutral genetic divergence of the populations.  However, allelic variation at microsatellite loci revealed that beach residents were distinct from the river population and from river fish that strayed to the beach.  Strays were morphologically similar to river fish but quite different from beach fish, suggesting that local adaptation may play a role in their reduced success at the beach.  Local adaptation of residents and declining success of strays can arise early in a population’s history (the beach site was colonized less than 14 generations ago).

Variation in reproductive success within and among populations of steelhead trout.
    Paul Bentzen, Thomas Quinn, Gregory Mackey, and Todd Seamons.

Variation in reproductive success (RS) among individuals within populations is an important determinant of effective population size (Ne), a critical component of the process of natural selection, and an important management issue in the case where non-native populations are introduced to supplement local native production.  We are studying individual RS in three populations of winter-run steelhead trout (Oncorhynchus mykiss) in Washington State to shed light on all of these issues.  One population is entirely wild, native and non-fished.  The other two populations occur in a second drainage; one is wild and native, and the other is a recently established hatchery population of allopatric origin that has been selected to spawn earlier than wild populations to promote reproductive isolation between native wild and feral hatchery fish.  We are non-lethally sampling pre-spawning adult steelhead in each of these populations, and their progeny at juvenile, smolt and adult life history stages.  In each case we record phenotypic data (length, weight, time of upstream migration of adults) and take fin clips for DNA analysis.  We are determining parentage, and hence realized reproductive success, through analysis of genetic variation in a suite of highly polymorphic microsatellite loci.  Early results of this study will be discussed.

Individual reproductive success in wild Atlantic salmon.
    Dany Garant, Louis Bernatchez, and Julian Dodson.

In nature, organisms are selected to maximise their reproductive success.  For anadromous Atlantic salmon (Salmo salar L.), this may involve a larger body size for both males and females.  Previous studies showed that female salmon most commonly mate with a single anadromous male although multiple-male mating is also known to occur.  Males are known to generaly mate with several partners.  In this study, we assessed individual reproductive success in a population of 76 wild anadromous Atlantic salmon that were introduced to a stretch of the Sainte-Marguerite River in Québec, Canada, that was previously unoccupied by salmon.  We used microsatellite loci to perform parentage analysis based on multi-locus genotype information from all spawners and a subsample of the progeny collected several months following emergence.  Our results showed that the body lengths of both sexes were neither correlated with the number of offspring produced nor with the number of spawning partners.  The number of partners seemed to be the major determinant of the number of offspring produced by an individual.  We also found that the mean and variance in the number of mating partners was as high in females as in males.

Use of microsatellites, behavior, and life history traits to assess reproductive success of pink salmon.
    Bobette Dickerson, Tom Quinn, Paul Bentzenand Mary Willson.

Reproductive success of Pacific salmon, Oncorhynchus spp., is influenced by the complex interactions of many factors.  In the past the different factors contributing to reproductive success have been looked at individually or in a few combinations and reproductive success has usually been estimated through indirect means such as male courtship success or fecundity or even more indirect indices such as body size.  The objective of this study is to look at all of the factors influencing reproductive success on the spawning grounds.  This is being accomplished by assessing the interactions and relative importance of life history traits (adult and juvenile body size, fecundity, egg size, longevity) and behavior (arrival date, social status, spawning site quality) to the reproductive success of pink salmon.  Reproductive success will be defined by number of offspring returning to spawn; parentage will be determined through DNA fingerprinting with microsatellites.  A path model will be designed that will define statistically the interactions among these factors and the magnitude of their influence on reproductive success.  This 4-yr. study will differ significantly from previous research in its emphasis on the interactive importance of these factors and the use of a direct measure of reproductive success.

Whole-population estimates of female and male reproductive success in lacustrine brook trout.
    Paul Blanchfield, Mark Ridgway, and Chris Wilson.

Our understanding of mating systems, in general, has been limited by the inability to observe the interactions of individuals on the breeding grounds. This is particularly true of salmonines, in which population size and breeding location often make acquiring data on individual mating success difficult. Here, we report on measures of female and male reproductive success from an intensively studied population of lacustrine brook trout (Salvelinus fontinalis). Female reproductive success is a function of egg survival, which is dependent upon spawning site quality (i.e. groundwater flow), brood loss due to superimposition by other females and egg cannibalism by peripheral males. For male mating success, we estimated spawning frequency and determined genetic contributions (paternity analyses) based on proximity to the female at the time of spawning. Furthermore, estimates of male reproductive success were contingent upon the success of the female with which they spawned.

Genetic variation of life history traits and outbreeding depression in locally adapted populations of Pacific salmon.
    Bill Smoker, Tony Gharrett, and colleagues.

Two isolated populations of pink salmon (Oncorhynchus gorbuscha) spawn in odd- or even-numbered years in Auke Creek, Alaska, and many other streams bordering the subarctic Pacific Ocean. We have used a series of breeding experiments to analyze variation of life history traits in those populations, and to test for outbreeding depression in F2 hybrids between them. Analysis of between family variation of embryonic traits (studied in the laboratory) and of adult traits (in fish liberated to sea and returning as adults to their natal stream) provides evidence of genetic variation of date of return, body size, fecundity, embryo development time, etc. Genetic variation of survival itself was high particularly when cohort survival at sea was high. Genotype by environment interaction was evident in our analysis of variation of embryo development time under different environmental (temperature) regimes. Outbreeding depression (reduced survival at sea in the second generation) was suggested in hybrids (formed by cryopreservation of semen) between sympatric Auke Creek populations, odd-year and even-year salmon. This evidence supports the notion that differences between local populations of salmon have evolved in response to selection in different local environments and that genetic variation

Effective population size.

Effective number of breeders in Atlantic salmon and variance in reproductive success.
    Nathalie Tessier and Louis Bernatchez.

In this paper, we estimated the effective number of breeders (Nb) in a population of Atlantic salmon from the analysis of linkage disequilibrium observed at microsatellite loci. The analysis was performed on aproximately 100 individuals representing each of six consecutive cohorts. These estimates were compared to the estimated total number of breeders (N) that produced each cohort. The Nb estimates from linkage disequilibrium varied between 94 and 350 (average = 180) whereas the total number of breeders varied between 235 and 976 (average = 562). No significant correlation ( r= 0.370, P = 0.47) was observed between Nb and N. In contrast, a negative correlation was observed between the ratio Nb/N, that varied between 0.50 and 0.14, and N (r = -0.791, P = 0.037). As no obvious differences in sex ratio was observed among the different groups of breeders, these results may indicate that variance in reproductive success increase with the density of breeders during a given reproduction event.

Effect of population size on effective population size in coho salmon.
    Adrian Spidle, Paul Bentzen, and Thomas Quinn.

To study the effect of population size on reproductive success in coho salmon, a small experimental population of coho salmon was established in a spawning channel at Big Beef Creek on the Hood Canal in Washington State for each of 3 successive years.  Tissue samples for DNA fingerprinting and a number of physical measurements were also taken from the adults.  Spawning behavior was observed and courtship and competitive interactions were recorded, resulting in an estimated social
hierarchy for the population for each year.  Fry were captured in the spring following each spawning season, their size and distribution in the spawning channel were recorded and a nondestructive fin clip was taken for DNA fingerprinting.  The fry samples have been genotyped for paternity determination using novel microsatellite loci cloned from coho salmon in our lab as well as a locus cloned from chinook salmon.  The genotypes of the fry will reveal the reproductive success of the individual adults, which may or may not be correlated to the position of the adults in the social hierarchy.

Variability in pink salmon family size: effective population size and a genetic basis for family-specific survival, recruitment, and population increases.
    Harold Geiger, Ivan Wang, William Smoker, Anthony Gharrett, and colleagues.

High values of the ratio of the variance of family size to the mean family size -- sometimes called the index of variability --  may be telling about the extent to which a population has "prepared" for change in a random environment.  For low breeding population sizes, this ratio indexes the amount of genetic variability a population can hold.  High values of this index of variability imply a lower value of the effective population size, as a relatively small number of families contribute disproportionately  to the next generation.  To examine how family-specific survival and recruitment in pink salmon alter the effective population size, we first assumed the size of each pink salmon family is a function of three family-specific factors.  For the ith family, (1) let rI denote an underlying recruitment to the marine lifestage, (2) let sI denote an underlying survival in the marine lifestage, and (3) let kI  denote the random realization of actual family size (i.e., E(kI ) = rI  sI ).  Thinking of family size this way led to an approximation to the index of variability that involves the mathematical expectation of family size together with the coefficients of variation of rI  and sI.  Using genetic breeding studies, we test the hypothesis that variation in rI  and sI have a genetic basis.  We also estimate these coefficients of variation for rI  and sI in controlled pink salmon populations, and speculate about values in wild salmon populations.

Assessing reproductive success.

Factors influencing the power of multi-locus genotypic information for quantifying reproductive success at the population level.
    Louis Bernatchezand P. Duchesne.

Multi-locus genotypic information, particularly derived from microsatellite analysis, is increasingly used for parentage inference and quantification of reproductive success. The factors potentially affecting assignment success are well known. However, the quantification of their effects on the predicted power of markers and their impact on assignment success have been little explored, particularly in situations where the parent-progeny relationships is unknown for both sexes. In this paper, we will use a simulation approach to explore the effects of the following factors on the proportion of resolved parentage assignment : the number and quality of markers, defined both from the number of alleles and their distribution, the number of potential parents in the population, and the rate and distribution of typing errors. We will then compare how assignment success obtained from empirical data fits the predictions made from the simulated approach.

Sampling requirements in a maximum-likelihood approach to parentage determination and estimation of differential reproductive success--more is better, to a point.
    Paul Moran, Linda Park, Perry Thornton, and Robin Waples.

A critical information gap in Pacific salmon conservation has been the paucity of data regarding relative reproductive success of hatchery and wild fish spawning naturally.  The use of multilocus microsatellite genotypes in natural populations to infer parentage provides a fundamentally new approach to research of this kind.  Computer simulations were used with microsatellite data for a Snake River steelhead population to estimate; 1) the confidence of parentage assignment, and 2) sample size of juveniles required to detect changes in proportional representation of wild and hatchery fish (relative to the numbers of adults passed over a weir).  Several interesting points emerged from this exercise.  If genotyping error rates are low, and all or nearly all potential parents have been sampled, then reasonable numbers of microsatellite markers (18, in this case) provide considerable power to assign parentage, even when the number of potential parents is large (600-900).  It was noted, however, that genotyping error rates combine with missing parents (i.e., not included in the sample) to substantially reduce successful parentage determination.  In this system, with a large proportion of hatchery fish (reducing sampling variance), we expect reasonable power to detect differences as small as 1% with juvenile samples of less than 200.

The effective use of genetic markers.
    Bryan Neff, Joe Repka, and Mart Gross.

Although genetic markers are available for salmonids, there is a lack of theoretical knowledge about the best means for using these markers to measure reproductive success (and parentage, kin and social structure).  We now present several models that were developed by our lab for measuring reproductive success in the wild.  We have identified two general situations: 'single-sex' competition, such as multiple male salmon spawning in a single female's nest; and 'two-sex' competition, such as fry in a stream that have been parented by multiple fathers and mothers.  When all the candidate parents have in fact been sampled, and therefore the 'breeding population' is known, we show how to use 'fractional allocation' methods to identify the cumulative probability of parentage per individual.  When the entire breeding population has not been sampled, which is the typical case in salmon research, we show how population allele frequencies can be used to calculate parentage.  In both cases we show how to use the genetic information in the situations of single-sex and in two-sex competition.  Finally, we present new models to determine how many offspring and loci to sample, thereby optimizing research effort and statistical power.  This research is supported by the Natural Sciences and Engineering Research Council of Canada, division of Ecology and Evolution.

Parentage analysis from progeny arrays: the benefits and pitfalls of linked loci and clustered mutations.
    Adam Jones and John Avise.

Microsatellite-based analyses of paternity and maternity can be extremely useful in efforts to assess reproductive success in natural populations.  Studies of fish often involve entire arrays of progeny for which certain a priori relationships are known.  Simultaneous consideration of the entire progeny array often facilitates parentage assignment, but it also may reveal unexpected complications.  In studies of parentage in pipefishes and seahorses, we encountered two phenomena that may be important in parentage analysis: linked loci and clustered mutations.  For hypervariable microsatellites, linkage is easily detected from large progeny arrays but is difficult to detect at the population level.  Linked loci generally should be avoided in parentage analysis, but under certain conditions the use of linked loci can actually lead to an increase in exclusionary power.  A second source of concern arises from clustered mutations.  Mutant alleles arising early in meiosis may be passed to numerous progeny, clouding the reconstruction of parental genotypes and possibly affecting the interpretation of inferences based on population genetics theory.

Hatchery-wild interactions.

Reproductive Success of Hatchery and Wild Coho Salmon.
    Mart Gross, Bryan Neff, and Ian Fleming.

 Although salmon supplementation and conservation programs often use hatchery fish, there is a lack of empirical knowledge about their behaviour, ecology and reproductive success in the wild.  We now present the results of several experiments in which we studied their behaviour and ecology and quantified their reproductive success.  Both wild and hatchery coho salmon were allowed to freely breed within a spawning channel in the wild.  The behaviors and interactions of the fish were recorded and after all spawning had been completed we collected the alevins from the nests.  Using microsatellite genetic markers, we determined the parentage, including maternity and paternity, of the fish.  Several important relationships emerged, including that between male position in the mating hierarchy and paternity, between male size and reproductive success, between stock type (hatchery or wild) and paternity, and between mating partner and success.  Overall, hatchery males attended fewer mating hierarchies, obtained lower paternity within a position, and made up only about a third of the male contribution to the next generation.  Hatchery females were also significantly less successful than wild females.  Hatchery fish were therefore relatively maladapted and decreased the wild population's effective population size.  Finally, our measures of reproductive success within hierarchies may be widely applicable to studies of salmon in the field.  This research is supported by NSERC and DFO of Canada, and NINA of Norway.

Natural reproductive success of hatchery and wild steelhead in the Kalama River, Washington.
    Patrick Hulett, Cameron Sharpe, and Chris Wagemann.

Allozyme genetic marking approaches were used in two long-term studies to estimate the reproductive success of non-locally derived stocks of hatchery summer and hatchery winter steelhead spawning naturally in the Kalama River. Results of the winter-run study corroborate those previously published from the summer-run study. Reproductive success (offspring produced per spawner) of the hatchery steelhead was substantially lower than that of the wild fish. Also, the disparity in reproductive success was increasingly pronounced at successive (subyearling, smolt, and adult) life history stages of the offspring.  These results are believed to reflect genetic differences between the wild and non-local hatchery stocks. Because their natural spawning poses genetic and ecological risks to wild steelhead, the non-local hatchery adults are no longer permitted access to principal wild Kalama steelhead spawning areas. Moreover, new research has been initiated to assess the wild stock conservation merits of using locally derived wild broodstock as a source for hatchery steelhead production. Specifically, the reproductive success of hatchery-reared steelhead spawned from wild Kalama summer-run broodstock will be compared to that of their wild-reared counterparts by relating microsatellite DNA profiles of naturally produced offspring to those of their prospective hatchery and wild parents.

Reproductive success of captively-reared, naturally spawning coho salmon.
    Linda Park, Jeff Hard, Barry Berejikian, Skip Tezak, and Eric LaHood.

One captive broodstock strategy being used in the management of Pacific salmon  involves capturing juveniles, rearing them to maturity in captivity, and  releasing them to spawn naturally with returning wild adults; however, captively-reared fish typically do not exhibit many of the physical  characteristics of sexually mature adults.  Since salmon compete for mates in a courtship ritual, it is not clear whether such fish would successfully mate with wild fish.  In one experiment, captively-reared adult coho salmon were placed in experimental channels with an equal number of wild-caught adults.  The fish were allowed to court and spawn undisturbed.  Microsatellite genotyping of the resulting progeny was used to evaluate reproductive success of the two groups.  We were able to assign over 98% of the progeny examined to single pair matings.  The variance in reproductive success among individuals was quite high.  Because all potentially spawning parents were known and their number was limited (20 per channel), the number of loci needed to determine individual pedigrees was minimal (from 3-5); however, we also genotyped additional individuals from the adult source populations and performed simulations to examine how many loci might be necessary to perform such an analysis with a larger population.

Male competition and breeding success in captively reared and wild coho salmon.
    Barry Berejikian, Skip Tezak, Linda Park, Steve Schroder, and Edward Beall.

 In the Pacific Northwest, releasing captively reared adult salmon (Oncorhynchus spp.) for natural spawning is an evolving strategy for the recovery of imperiled populations.  The ability of captively reared fish to spawn naturally may be compromised by their artificial rearing environments, which differ markedly from those experienced by wild fish.  In this study, wild coho salmon (O. kisutch) males dominated access to spawning females in 11 of 14 independent trials.  In two cases where satellite males (both captively reared) were observed participating in spawning, DNA fingerprinting results determined that they did not sire any of the progeny.  When spawning occurred at night and was not observable, DNA results confirmed continuation of behaviour-based hierarchies determined before nightfall.  Aggression data collected during the first hour of competition indicated that dominance was established soon after the males were introduced into a common arena containing a sexually active female.  We hypothesize that  status signaling and decisions by subordinate males to avoid direct competition may have minimized conflict.  The competitive inferiority of captively reared coho salmon in this and a previous study probably reflects deficiencies in culture environments which fail to produce appropriate body coloration, body shape, and perhaps alter natural behavioral development.

Reproductive strategies and tactics.

The determinants of mating success in male rainbow trout.
    Robin Liley.

Four males of different sizes were allowed to compete for spawning with single females in laboratory stream tanks.  A videotaped record allowed identification of the males participating in spawning and an estimate of the time to ejaculation by competing males.  Eggs harvested from the gravel were raised in an incubator, following which the paternity of fry was determined using microsatellite DNA procedures.  The behavioral success of competing males is compared with their fertilization success.  Of particular interest is the time to ejaculation and the time to the onset of polyspermy block (as esimated in in vitro trials).

Selection against late breeding and small egg size in Atlantic salmon resulting from maternal effects on juvenile fitness.
    Sigurd Einumand Ian Fleming.

Individual variation in maternal traits potentially translates into differences in reproductive success through effects on offspring performance, and this can have consequences on various organisational scales. Timing of breeding and offspring size are two such traits, both exhibiting pronounced variation within populations, and both likely being shaped by several selective agents. However, little is known about how these agents interact in the wild. We therefore undertook a field experiment, while controlling for genetic variation by manipulations, to examine how offspring fitness during the first summer is affected by breeding time and offspring size using Atlantic salmon (Salmo salar).  The first 17 days post-emergence from nests was identified as a critical period for selection, with univariate selection differentials demonstrating significant directional selection against late and small emerging juveniles. Standardised selection gradients indicated stronger selection on date of emergence than on body size at emergence. Date of emergence also influenced subsequent body size, and this is suggested to have fitness consequences during later selective periods. We conclude that timing of emergence and offspring fitness are causally related in Atlantic salmon. The observed effect of body size at emergence on survival, combined with other data, suggest that egg size per se may also affect fitness.

Reproductive success in Atlantic salmon (Salmo salar): effects of parr size and intensity of mate competition.
    Matthew Jones and Jeffrey Hutchings.

We tested the hypothesis that as the intensity of competition among parr increases, any advantage of body size to fertilization success diminishes.  We designed an experiment involving 13 anadromous single-pair crosses with varying numbers of parr (3 replicates of 3 parr and 2 replicates of 5, 7, 10, and 20 parr) of different sizes (range 81-120 mm).  Unintentional fish movement resulted in four crosses as designed (5, 10, 10, and 20 parr), plus three other crosses having parr:female:anadromous male ratios of 5:1:0, 7:2:1, and 23:2:1.  We determined individual paternity for 55-60 offspring from 2-3 nests from each cross using 3-4 hyper-variable microsatellite loci.  We found no relationship between parr size and individual reproductive success at any level of competition when anadromous males were involved.  However, in the cross with no anadromous male, the influence of body size was highly significant.  Variation in individual and total parr reproductive success among nests within crosses was high. We found a negative relationship between total parr reproductive success and
intensity of anadromous male competition.  The evolutionary significance of these results are discussed.

Relative fitness of Atlantic salmon anadromous males and precociously mature male parr in a small stream.
    Edward Beall, Jacques Dumas, and Eva Garcia-Vazquez.

Seven Atlantic salmon anadromous males of different sizes and six females, tagged with radiotransmitters, were released in a tributary of the Nivelle River (southwest France) in a section inaccessible to migratory fish but which held a population of salmon parr at a density of 6 individuals/100m2. Tagged salmon were radiotracked during the whole spawning period and the distribution of redds was precisely determined. One month after completion of spawning, all the eggs were recovered from the redds and survival estimated. To determine reproductive success of the different males present, paternity analyses with microsatellite probes were performed on eggs sampled in the different egg pockets. Long term studies of the Nivelle salmon population allowed us to determine the rates of adult and parr survival. It was thus possible to compare the relative fitness of the two male alternative phenotypes (anadromous vs. precocious males). Results clearly indicated that sneaking precocious male parr could achieve a higher fitness than their adult counterparts. This confirmed the hypothesis that the coexistence of the two alternative phenotypes, commonly observed in Atlantic salmon, is dependant from a conditional strategy.

Is there an equivalence between energy and reproductive success in Atlantic salmon?
    Ian Fleming, Bror Jonsson, Kjetil Hindar, and Ole Berg.

Reproductive success of organisms is thought to be strongly influenced by their available surplus energy and it has been argued that energy content may be a good surrogate measure of fitness. We evaluated this in Atlantic salmon by examining the association between individual energy content and reproductive success measured during a series of breeding experiments. The experiments were conducted in semi-natural streams, where breeding behaviour and mating were monitored 24 h/day using live and video observations, and reproductive success quantified (number of eyed embryos parented). The results are also contrasted with measures of energy use during spawning in the wild. Female reproductive success proved to be strongly related to estimated energy content at the start of the breeding season and was a function of female fecundity and her ability to have her offspring survive egg incubation. Male reproductive success was also positively related to estimated energy content, however, the relation was considerably weaker than that observed in females. This appeared to hold true for both anadromous and parr males, indicating that the determinants of male success were multifaceted reflecting the diversity of reproductive tactics available to males and that energy content alone may provide a less than satisfactory measure of fitness.

Within-basin life history divergence based on alternate reproductive strategies in steelhead: genetic and environmental considerations.
    Jennifer Nielsen.

Genetic differentiation between two reproductive ecotypes summer- and winter-run anadromous steelhead found in the Middle Fork Eel River (MFER) in northern California was tested using 16 microsatellite loci.  Twelve of these loci showed significant differences in allelic frequency between the two MFER ecotypes (Fisher's exact p <0.025). Fisher's combined test for independence also supported significant genetic separation between these sympatric steelhead populations (p <0.001).  Analysis of molecular variance (AMOVA) indicated that only 1% of the overall microsatellite allelic variation could
be attributed to differences in steelhead run-timing in this basin.  Variation found among individuals within the two runs equaled 18.2%. Analyses showed less genetic distance between the two reproductive ecotypes than in comparisons made with geographically proximate coastal winter-run steelhead populations. Despite estimates of significant gene-flow, molecular divergence between the two reproductive ecotypes based microsatellite diversity was estimated to be 16,000 - 28,000 years.

Frequency-dependent reproductive success of alternative life history tactics in male sockeye salmon.
    Drew Hoysakand Robin Liley.

In this study we tested for the presence of frequency-dependent male reproductive success in individual clutches of sockeye salmon.  We allowed females to spawn with groups of males and then we determined paternity of the resulting offspring.  We manipulated not only the frequency of two male age classes in these spawnings, but also the density of males.  This allowed us to test for an effect of density on frequency-dependence.  We performed eight replicates for each of four treatments, including two levels of male density and two levels of frequency of two male age classes.  Reproductive success of the two male age classes showed strong negative frequency-dependence when male density was low.  However, at high male density reproductive success was not frequency-dependent and the two male age classes had nearly equal success.  Thus the nature of male competition appears to vary with male density.

Last updated March 5, 1999.