Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Factors influencing reproductive success of male sockeye salmon Hoysak, Drew J. 2002

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2002-731804.pdf [ 6.14MB ]
Metadata
JSON: 831-1.0090552.json
JSON-LD: 831-1.0090552-ld.json
RDF/XML (Pretty): 831-1.0090552-rdf.xml
RDF/JSON: 831-1.0090552-rdf.json
Turtle: 831-1.0090552-turtle.txt
N-Triples: 831-1.0090552-rdf-ntriples.txt
Original Record: 831-1.0090552-source.json
Full Text
831-1.0090552-fulltext.txt
Citation
831-1.0090552.ris

Full Text

Factors Influencing Reproductive Success of Male Sockeye Salmon by Drew J. Hoysak B.Sc. (Honours), Carleton University, 1983 M . S c , University of Western Ontario, 1993 A THESIS S U B M I T T E D IN P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E OF Doctor of Philosophy in T H E F A C U L T Y OF G R A D U A T E STUDIES (Department of Zoology) We accept this thesis as conforming to the required standard The University of British Columbia December 2001 © Drew J. Hoysak, 2001 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department The University of British Columbia Vancouver, Canada Date )H 3\<L < ^ Q Q S DE-6 (2/88) Abstract In this thesis I studied factors that influence reproductive success in two age classes of male sockeye salmon (Oncorhynchus nerka). In my first experiment I tested for the presence of frequency-dependent reproductive success of male age classes in single spawning events and examined the influence of the number of males present on frequency-dependence. I placed groups of males in an enclosure with a nesting female, allowed them to spawn, and determined paternity of the resulting offspring. In spawning trials with three males, frequency-dependent reproductive success was apparent. That is, both age classes had greater mean reproductive success when they were in the minority. However, frequency-dependent reproductive success was not present in groups with six males and the two age classes did not differ in mean reproductive success. My second study was a detailed examination of factors that may have influenced individual male reproductive success in the previous experiment. Alpha male positions were usually held by five-year-old (5Y) males characterized by large body size and large secondary sex traits. Among satellite males, three-year-old (3Y) males held positions closer to the female than did 5Y males, and distance from the female was a good predictor of whether or not a male would participate in spawning and of the order in which males began spawning. Overall, alpha males fertilized more eggs than did 3Y satellite males, and 3Y satellite males fertilized more than 5Y satellite males. Behaviour during spawning appeared to influence male success. Path analysis suggested that male success was directly affected by the time spent in spawning posture which, in turn, was influenced by the time that a male began to spawn. I found no evidence of a i i relationship between male fertilization success and distance from the female during spawning. In the next experiment, I compared male age classes in their success in sperm competition. I fertilized eggs in vitro with mixtures of milt from pairs of males that differed in age. The milt mixtures contained equal numbers of sperm from each male. Paternity analysis of the resulting offspring revealed that success in sperm competition did not depend on male age. However, success in these competitive fertilizations did not conform to the fair raffle model of sperm competition. Paternity in most of the egg batches was biased in favour of one male. I did not detect any morphological or genetic correlates of success in sperm competition. In my final set of experiments I used in vitro fertilization to examine some aspects of sockeye salmon fertilization dynamics. Sperm maintained high fertility for 10 s after being exposed to water but fertility dropped sharply after that. The association between eggs and sperm occurs very rapidly. 80% fertilization success was achieved with five s of sperm-egg mixing and c. 25% success was achieved with < 0.5 s of sperm-egg mixing. Sperm from 3Y and 5Y males differed little in motility and did not differ at all in their propensity to fertilize eggs. i i i Table of Contents Abstract n Table of Contents iv List of Tables v i i i List of Figures x n List of Abbreviations x v i i Preface X 1 X Acknowledgements x x 1 General Introduction 1 Alternative Reproductive Phenotypes in Pacific Salmon 4 Fitness of Alternative Reproductive Tactics 5 2 General Methods 8 Study Area and Species 8 Fish Capture and Measurement 8 Enclosure Experiments 9 Fish Selection 9 Treatments 10 Spawning Trial Protocol 10 D N A Analysis 12 Gamete Collection 13 Data Analysis 14 iv 3 Frequency-Dependent Reproductive Success in Individual Spawning Events and the Effect of Operational Sex Ratio 20 Introduction 20 Methods 2 4 Results 25 Discussion 26 4 Factors Influencing Male Fertilization Success in Single Spawning Events 3 9 Introduction 39 Methods 42 Data Analysis 43 Results 45 Behaviour and Male Positions Before Spawning 45 Time to Spawn 47 Behaviour and Positions During Spawning 47 Correlates of Male Fertilization Success 49 Discussion 52 5 The Outcome of Sperm Competition Between Alternative Male Phenotypes 81 Introduction 81 Methods 83 Data Analysis 85 Results 86 Discussion 87 v 6 Fertilization dynamics and a comparison of sperm from alternative male phenotypes 95 Introduction 95 Methods 97 Fish Selection and Gamete Collection 97 General Protocol 98 Experiment 1: Fertilization Rate v. Male Age and Sperm Concen-tration Around the Eggs 99 Experiment 2: Changes in Fertilization Rate With Time After Sperm Activation 99 Experiment 3: Changes in Fertilization Rate With Time After Egg Activation 100 Experiment 4-' Fertilization Rate v. Time of Egg-Sperm Association 101 Results 102 Milt & Sperm Characteristics 102 Experiment 1: Fertilization Rate v. Sperm Concentration & Male Age 102 Experiment 2: Changes in Fertilization Rate With Time After Sperm Activation 103 Experiment 3: Changes in Fertilization Rate With Time After Egg Activation 104 Experiment 4- Fertilization Rate v. Time of Egg-Sperm Association 104 Discussion 104 7 General Discussion 120 References 126 vi Appendix A A Mode l of Fertilization Success in Spawning Events . 147 Appendix B Tables for Chapter 4 152 List of Tables 2.1 Distribution of spawning trials over three years and number of trials for which various types of data were collected 15 2.2 Characteristics of microsatellite loci used for paternity determination. . . 16 2.3 P C R conditions for each locus. For each set of cycles the conditions for denaturing, annealing, and extending are given as time-temperature where time is in seconds and temperature is in degrees Celsius. The number of cycles in the third set of cycles varied among loci and the ending cycle number is denoted as i. At the end of these cycles, an additional extension for 5 min at 72 ° C was added for all loci 17 4.1 Behaviours recorded in spawning trial observations 57 4.2 Spearman rank correlations between behavioural variables in first and last observations. Sample sizes were 20 for first observations and 31 for last observations. None of the correlations was significant using otBonj — 0.008. 58 4.3 A N O V A tables for the effects of morphological variables on satellite male distance from the female. Analyses were done as linear mixed-effect models (Pinheiro & Bates 2000) with trial number included as a random factor. The first model included 64 3Y males in 32 trials. The second model included 41 5Y males in 25 trials. See Table B.3 and text for interpretation of P C I and PC2 59 viii 4.4 A N O V A table for the effects of male age and spawning participation on spermatocrit. The analysis was done as a linear mixed-effect model (Pin-heiro & Bates 2000) with trial number included as a random factor. The analysis included 94 males in 22 trials. The interaction between age and participation was not significant and was removed from the model ( F l i 6 9 = 2.002, P = 0.162) 60 4.5 Spearman rank correlations between the proportion of fry sired by alpha males and frequency of three behaviours in first (N = 16) and last (N = 24) observations. None of the correlations was statistically significant. . . 61 5.1 Percentage of offspring sired by 3Y and 5Y males in competitive fertiliza-tions (N = 30 offspring for each replicate) 91 6.1 Attempted and actual time (from video analysis) of egg exposure to sperm in experiment 4. Sample size was six in all cases 110 6.2 Analysis of Variance of fertilization rate (arcsine square root transformed) in experiment 1. Female was included as a random factor I l l 6.3 Spearman rank correlations of fertilization rates of male pairs in experi-ment 1. Sample size was 11 in all cases 112 6.4 Analysis of variance for fertilization rate (arcsine square root transformed) in experiment 2. Female was included as a random factor 113 6.5 Spearman rank correlation coefficients between fertilization success and two measures of motility in experiment 2. Both male age classes were included so that N — 16 for all tests. A l l f s greater than 0.1 except where noted , 114 B . l Linear regressions of three morphological measurements versus hypural length. A l l variables were log-transformed for the analyses 152 ix B.2 Correlations between morphological variables for 174 males. Residuals were calculated as deviations from the reduced major axis regressions. . . 152 B.3 Loadings from principal components analysis of morphological measure-ments for 174 males. The proportion of variance explained by each com-ponent was 0.508, 0.250, 0.136, and 0.106 respectively 152 B.4 Analysis of variance on the number of aggressive acts of alpha males in first and last observations 153 B.5 Analysis of variance on the number of courtship behaviours (square-root transformed) of alpha males in first and last observations. Removing the interaction term did not influence significance of main effects in either model. 153 B.6 Analysis of variance on the number of satellite intrusions (square-root transformed) in first and last observations. The interaction term was not significant in either model (first: df = l,SS = 0.207, F = 0.342, P = 0.567; last: df = 1, SS = 0.022, F = 0.034, P = 0.854). 154 B.7 A N O V A to test the effects of OSR and age group composition on time to spawn. The interaction between OSR and age composition was not significant and was removed from the model (^1,30 = 1.071, P = 0.309). . 154 B.8 A N O V A to test the effects of OSR and age group composition on female spawning duration. The interaction between OSR and age composition was not significant and was removed from the model (Fi^o = 0.162, P = 0.691) 154 B.9 A N O V A to test the effects of OSR and age group composition on the time from the start of the female gape to the start of her first dig. The interaction between OSR and age composition was not significant and was removed from the model (Fl>25 = 0.022, P = 0.883) 155 B.10 A N O V A table for the effects of male age and spawning participation on testes depletion. The analysis was done as a linear mixed-effect model (Pinheiro & Bates 2000) with trial number included as a random factor. The analysis included 77 males in 19 trials B . l l A N O V A to test for effects of OSR and age group composition on al-pha male fertilization success. Alpha male success was arcsine squareroot transformed for the analysis xi List of Figures 2.1 Schematic drawing of the spawning enclosures from above showing the spawning channel (C), holding pens (H), gravel substrate in each enclosure (G), tarpaulin-covered substrate in each enclosure (T), and shoreline (S). The arrow indicates the direction of water flow 18 2.2 Relationship between spermatocrit and sperm count for 3Y males (o) and 5Y males (x). The slope of the relationship (using an arcsine-square root transformation on spermatocrit) does not differ between the two age classes (F l j 2 8 = 0.451, P = 0.507), nor does the elevation ( F l i 2 9 = 1.615, P = 0.214) 19 3.1 Frequency-dependent mating success in a population with large and small male phenotypes. Arrows indicate mating between a male and female. . . 32 3.2 Mean proportion of fry fertilized per male for 3Y (closed circles) and 5Y (open circles) males in spawning trials with three males. Error bars are ± 1 SE (back-transformed from arcsine transformations). Sample sizes indicate the number of spawning trials on which the calculations were based. 33 3.3 Mean proportion of fry fertilized per male for 3Y (closed circles) and 5Y (open circles) males in spawning trials with six males. Error bars are ± 1 SE (back-transformed from arcsine-squareroot transformations). Sample sizes indicate the number of spawning trials on which the calculations were based 34 xii 3.4 Proportion of males that spawned in spawning trials with three males. Includes all trials in which spawning status is known for all male. Closed circles represent 3Y males and open circles represent 5Y males. Sample sizes indicate the number of spawning trials on which the calculations were based 35 3.5 Proportion of males that spawned in spawning trials with six males. In-cludes all trials in which spawning status is known for all males. Closed circles represent 3Y males and open circles represent 5Y males. Sample sizes indicate the number of spawning trials on which the calculations were based 36 3.6 Mean proportion of fry fertilized per male for 3Y (closed circles) and 5Y (open circles) males in spawning trials with three males, including only males that participated in spawning. Error bars are ± 1 SE (back-transformed from arcsine transformations). Sample sizes indicate the num-ber of spawning trials on which the calculations were based 37 3.7 Mean proportion of fry fertilized per male for 3Y (closed circles) and 5Y (open circles) males that spawned in spawning trials with six males, in-cluding only males that participated in spawning. Error bars are ± 1 SE (back-transformed from arcsine-squareroot transformations). Sample sizes indicate the number of spawning trials on which the calculations were based. 38 4.1 Boxplot of the number of aggressive acts per individual per 5min obser-vation for females (F), alpha males (Alpha), 5Y satellite males (5Y), and 3Y satellite males (3Y). First observations (20 trials) were done at least 2 hr before last observations (31 trials). Data from all four treatments are included 62 xiii 4.2 Mean number (± SE) of aggressive acts initiated by alpha males in first (a) and last observations (b) of trials composed of 1/3 or 2/3 3Y males. Open symbols indicate trials with three males and solid symbols indicate trials with six males. Sample sizes are in brackets 63 4.3 Mean number (± SE back-transformed from square-root transformations) of alpha male courtship behaviours in first (a) and last observations (b) of trials composed of 1/3 or 2/3 3Y males. Open symbols indicate trials with three males and solid symbols indicate trials with six males. Sample sizes are in brackets 64 4.4 Mean number (± SE back-transformed from square-root transformations) of satellite intrusions into the nest in first (a) and last observations (b) of trials composed of 1/3 or 2/3 3Y males. Open symbols indicate trials with three males and solid symbols indicate trials with six males. Sample sizes are in brackets 65 4.5 Frequency distribution of the change in male distance from the female as spawning approached for 3Y satellites (a) and 5Y satellites (b). Positive values indicate that males moved closer to the the female as spawning approached and negative values indicate that males moved away from the female as spawning approached 66 4.6 Mean time to spawn (±SE) for trials composed of 1/3 or 2/3 3Y males. Open symbols indicate trials with three males and solid symbols indicate trials with six males. Sample sizes are in brackets 67 4.7 Relationship between rank order in which males began to gape and their distance from the female shortly before spawning. Symbol size is propor-tional to the number of data points. Includes 75 males in 26 trials 68 xiv 4.8 Relationship between time that males began to gape and their distance from the female in the last observation before spawning. Symbol size is proportional to the number of data points. Includes 68 males in 24 trials. 69 4.9 Duration of female gape in trials where 3Y males were at low and high frequency. Open circles represent trials with three males and closed circles represent trials with six males 70 4.10 Time between start of female gape and start of dig in trials where 3Y males were at low and high frequency. Open circles represent trials with three males and closed circles represent trials with six males 71 4.11 Time that males began gaping relative to the first male. Sample sizes are given in brackets 72 4.12 Comparison of male classes in their gape start times. Time was measured relative to the beginning of the female's gape. Sample sizes (number of trials) are given in brackets. For trials with more than one 3Y or 5Y • satellite, means were used 73 4.13 Male spawning position versus the time that gaping was started relative to the first male within a trial. Symbol size is proportional to the number of data points. Includes 84 males in 27 trials 74 4.14 Gape duration of males. Sample sizes (number of trials) are given in brackets. For trials with more than one 3Y or 5Y satellite, means were used. 75 4.15 Frequency distribution of spawning positions of (a) alpha males (N = 26), (b) 5Y satellite males (N = 11), and (c) 3Y satellite males (N = 47). . . 76 4.16 Mean testes depletion (± SE) for 3Y and 5Y males. Open circles represent males that did not participate in spawning and closed circles represent males that did participate in spawning. Sample sizes are given in brackets. 77 xv 4.17 Boxplots of fertilization success per male (a) and per male that partici-pated in spawning (b) for alpha and satellite males. Number of trials is shown in brackets 78 4.18 Alpha male fertilization success in trials where 3Y males were at low and high frequency. Open circles represent trials with three males and closed circles represent trials with six males 79 4.19 Path diagram to examine the influence of male timing in spawning events on fertilization success. Coefficients are standardized multiple regression coefficients from an analysis with fertilization success (arcsine square-root transformed) as the dependent variable and start time, spawning dura-tion, and trial number included as independent variables. Significant path coefficients are indicated by asterisks 80 5.1 Power of the replicated goodness-of-fit test to reject the null hypothesis of no difference between age classes in fertilization success given that one age class actually fertilizes the proportion indicated (on average) on the x-axis. 92 5.2 Scatterplot of the difference between competing males in proportion of fry sired versus the difference in their genetic variability (as measured by mean d2) 93 5.3 Scatterplot of the difference between competing males in proportion of fry sired versus the difference in their relatedness with the female (as measured by the kinship coefficient) 94 6.1 Relationship between proportion of eggs fertilized and sperm concentra-tion for 3Y (open boxes) and 5Y (shaded boxes) males in experiment 1. Lines are drawn at the medians with boxes encompassing the interquartile ranges. Whiskers represent total range of the data 115 xvi 6.2 Proportion of eggs fertilized in the sperm concentration experiment. Only two replicates are included to illustrate difference among replicates and similarity among males within replicates. Squares represent replicate #3 and circles represent replicate #9. Open symbols and dashed lines repre-sent 3Y males. Closed symbols and solid lines represent 5Y males. . . . . 116 6.3 Relationship between proportion of eggs fertilized and duration of milt exposure to water for 3Y (open boxes) and 5Y (shaded boxes) males in experiment 2 117 6.4 Relationship between proportion of eggs fertilized and duration of egg exposure to water in experiment 3 118 6.5 Relationship between proportion of eggs fertilized and duration of egg exposure to milt in experiment 4 119 A . l Index of frequency-dependence versus a (reduction in success for large males) in the spawning model. Line 1 is for six males present and b = 0.2. Line 2 is for three males present and b = 0.2. Line 3 is for six males present and b = 0.8. Line 4 is for three males present and 6 = 0.8 151 xvii List of Abbreviations 3Y three-year-old 5Y five-year-old &Bonf Bonferroni-adjusted alpha-level A N O V A analysis of variance A R P alternative reproductive phenotype d2 a measure of individual genetic variability ESS evolutionarily stable strategy FDS frequency-dependent selection OSR operational sex ratio rs Spearman rank correlation coefficient W M S T Wilcoxon matched-pairs signed ranks test xviii Preface Chapter 6 has been published as: Hoysak, D. J. & Liley, N . R. 2001. Fertilization dynamics in sockeye salmon and a comparison of sperm from alternative male phenotypes. Journal of Fish Biology 58:1286-1300. xix Acknowledgements I am very grateful to R. Liley for his support and patience throughout my degree. I also thank my committee (M. Healey, D. McPhail, D. Schluter, J. Smith) for valuable input. I. Fleming, S. Hinch and M . Whitlock provided insightful comments. M . Whitlock contributed many commas, and I will be forever grateful. Fisheries and Oceans Canada provided an excellent research site at Fulton River, and the staff (D. Chapman, J. Dickerson, C. Harrison, J. Smith) was very helpful. Assistance in the field was graciously provided by R. Liley, Y . Morbey, and P. Rutherford. J.P. Danko and D. Reinders assisted with field work that is not included here, but is not forgotten. Chapter 6 benefited from discussions with R. Montgomerie and B. Leach and comments from M . Gage. I could not have done the paternity analyses without the help of E. Taylor. L. Barrett-Lennard, P. Tamkee, and all of the TayMacs were also of great help in the lab. P. Rutherford and A. Hoysak provided all the support and inspiration I needed to do this work. D R E W J . H O Y S A K The University of British Columbia December 2001 xx Chapter 1 General Introduction Phenotypic variation is ubiquitous in natural populations. Selection usually acts to reduce variability (or at least its genetic portion), but this is not always the case. Variation in some traits may be adaptive and have a genetic basis and therefore may be maintained by selection. An example that has received considerable attention is that of alternative male reproductive tactics (reviewed in: Arak 1984; Austad 1984; Dominey 1984; Gross 1996). Alternative male reproductive tactics have been documented in a variety of taxa. They range from subtle changes in behaviour by individuals to elaborate morphological and behavioural differences among individuals. For example, in bighorn sheep (Ovis canadensis), some males court and defend females that gather in herds (Hogg 1984). Other males fight these tending males for access to females. Still other males avoid gathering sites and try to block females from moving to these sites so that the males can court females with less competition from other males. Many males use more than one of these tactics within a breeding season. A more extreme male polymorphism occurs in Paracerceis sculpta, a marine isopod crustacean (Shuster 1989). In this species, there are three forms of males that differ both in morphology and reproductive behaviour. The largest males (about 9mm in length) defend harems within the spongocoels of intertidal sponges. They have enlarged uropods and telsons, which are used in courting and grasping females and in fights with other males. The smallest males are only about 3mm in length, lack ornamentation and gain access to females by sneaking into occupied spongocoels. Males in the third class are intermediate in size, morphologically similar to receptive females, and mimic female courtship behaviour. They are thus able to deceive alpha males into letting them into the spongocoel. 1 The existence of multiple tactics within a species is, at first glance, paradoxical. Why doesn't selection eliminate all but the most successful tactic, especially since, in many (perhaps most) examples, the alternatives do not yield equal fitness payoffs (Dominey 1984)? One possibility is that selection favours individuals that are able to adopt suboptimal tactics while waiting for a chance to use the optimal tactic. For example, if the optimal tactic is to defend a territory and display to prospective mates and only older males are able to do so, then it may be worthwhile to adopt a satellite tactic (wait near a displaying male and intercept incoming females) at younger ages (e.g., Howard 1978). Even when such differences in ability to employ the 'optimal' tactic are fixed through an animal's life, selection will favour individuals that obtain whatever fitness gains possible through alternative tactics. For example, in an anthophorid bee, Centris pallida, adult males vary in size because of food supply differences at the larval stage (Alcock et al. 1977). Male larvae that are well nourished become large adults and use the mating tactic of patrolling for and fighting over newly emerged females. Male larvae that are undernourished become small adults, and the best mating tactic they can use is to hover near foraging areas to find females. While reproductive success of small hovering males is lower than that of large patrolling males, the small males do better by hovering than they would by patrolling. Alternative tactics may not always be age- or environment-dependent, as in the above examples. Gadgil (1972) hypothesized that some alternative tactics may be a result of genetic polymorphisms that yield equal lifetime fitness. He argued that if male investment in competition for mates incurs a large cost, it may be beneficial to minimize this investment and use a different reproductive tactic. Two reproductive tactics could coexist in a population if their average fitnesses are negatively frequency-dependent. The result would be a stable equilibrium of the two tactics. The idea that the best tactic an individual can use depends on what tactics the rest of the population 2 is using was subsequently formalized into models of the evolutionarily stable strategy (ESS) using game theory (Maynard Smith & Price 1973). It should be noted that frequency-dependent selection is not the only mechanism that could maintain alternative strategies in a population. Processes such as heterozygote advantage and fluctuating selection (spatial or temporal) may also lead to polymorphisms (Gadgil & Taylor 1975; Rubenstein 1980), but there are no known example of alternative reproductive tactics that are maintained by these mechanisms. The terminology used in game theory can be confusing and there are many inconsistencies in the literature (Dominey 1984), so some clarification will be worthwhile here. A strategy is defined as a genetically-based life history plan (Dominey 1984; Gross 1984). Tactics, on the other hand, are the proximate methods used to achieve the life history plan. Each individual possesses only one strategy, but one or more tactics may be employed to achieve the strategy. Thus, alternative reproductive tactics may represent alternative strategies, or they may be components of a single strategy. A given strategy can only be considered to be evolutionarily stable if there are no other strategies that could produce a higher average fitness. Thus, if more than one strategy can persist in a population, an ESS at the individual level does not exist (these cases are often referred to as mixed ESS's - where the stability occurs at the population level). There are few documented examples of alternative reproductive strategies in nature. The three morphs of the marine isopod, Paracerceis sculpta (see above), have been shown to be genetically based, with the three strategies yielding equal fitnesses (Shuster 1989; Shuster & Wade 1991). The mechanism for maintenance of these strategies has not been investigated. One of the most thoroughly documented cases of alternative male strategies occurs in bluegill sunfish, Lepornis macrochirus (Gross & Charnov 1980; Gross 1982; Dominey 1984; Gross 1991). Males in this species follow one 3 of two life history paths. In the 'parental' path, males mature at seven years of age and build nests to which they attract females. In the 'cuckolder' path, males mature at two years of age and, with the aid of their small size and cryptic colouration, mate by sneaking into the nests of parental males when spawning takes place. When cuckolder males reach four years of age, they change tactics from sneaking to female mimicry. Gross (1991) hypothesized that these two paths represent alternative reproductive strategies and provided evidence that they are maintained in the population by frequency-dependent selection. Alternative Reproductive Phenotypes in Pacific Salmon Alternative male reproductive phenotypes are common in.fish (Taborsky 1994). In many fish species there is a wide range in body size among breeding males and there is corresponding variation in competitive ability (e.g., Foote 1990). Since fertilization is often external, males may not be able monopolize females during spawning. Therefore, small competitively inferior males may be able to fertilize eggs by sneaking in with spawning pairs. Gross (1984, 1985) hypothesized that alternative mating strategies based on variation in age at maturity exist in Pacific salmon (Oncorhynchus spp.: but see Repka & Gross (1995)). Pacific salmon are semelparous and anadromous, breeding in temperate freshwater streams. In breeding populations, there may exist anywhere from one to several age classes, usually with a greater age range in males than in females (Fleming 1998). The youngest males (sometimes referred to as jacks) are small and show relatively little or no development of secondary sexual characteristics. Older males can be much larger than jacks and exhibit elongation of their jaws, enlarged teeth, a dorsal hump, and conspicuous body colouration. Males congregate at females' nests and compete to fertilize eggs. Large (older) males typically fight and display for 4 proximity to a female, with the winner assuming the 'alpha' position beside the female. The smallest (youngest) males rarely fight for access to females, but instead employ a sneaking tactic using crypsis and refuges to remain close to a nesting female. During a spawning event the female and alpha male simultaneously release gametes into the nest. Subordinate males may also rush into the nest and release sperm. The mechanisms maintaining variation in age at maturity of Pacific salmon are not well understood. Age at maturity can have a genetic component (Iwamoto et al. 1984; Hankin et al. 1993; Heath et al. 1994), so different age classes may represent alternative reproductive strategies maintained by negative frequency-dependent selection (Gross 1984, 1985). However game theory models assume direct correspondence between genotype and strategy. This is clearly not the case in salmon as maturity is strongly affected by growth rate, with the fastest growing individuals being the first to mature (Hutchings 1993; Wood & Foote 1996). Therefore, variation in age at maturity in male Pacific salmon may represent variation in tactics within one strategy. Models have been proposed by Hutchings & Myers (1994) and Repka & Gross (1995) that incorporate condition-dependent maturity into game theoretical models (reviewed in Hazel & Smock (2000)). Fitness of Alternative Reproductive Tactics A complete understanding of the origin and maintenance of alternative reproductive tactics requires knowledge of genetic and .developmental mechanisms and the effect of variation in tactics on fitness. The factors involved in determining fitness vary among species, and I concentrate on those that are likely to be important in Pacific salmon. Fitness is a function of survival to maturity and lifetime reproductive success (number of offspring produced that survive to breed themselves). For males, the number of offspring produced has three main components: number of matings, number 5 of offspring sired in each mating, and proportion of offspring that survive. The first two factors can be further broken down. The number of matings may be a function of breeding life span, amount of sperm available, rate of female encounter, and potential to mate with females encountered. The number of young sired per mating may be a function of male behaviour, number and tactics of male competitors, number of sperm released, quality of sperm released, and number of viable ova available for fertilization. My research examines reproductive competition among male sockeye salmon (0. nerka) once they have found a nesting female. The goal of these studies is to increase our understanding of the relationship between male age at maturity and fitness. I examine factors that influence the probability that a male will spawn with a particular female and factors that influence the number of young sired per mating. The factors that I consider include male age, morphology, behaviour before and during spawning, and sperm characteristics. In this thesis I do not test hypotheses or make assumptions concerning the developmental nature of alternative reproductive phenotypes (ARP's) in sockeye salmon. Male age classes may represent alternative strategies (mixed ESS) or tactics (ESS with condition-dependent variation). In either case, it is important to understand how competition among and within age classes influences fitness. In Chapter 3, I consider the of effect spawning group composition on fertilization success of males in two age classes. The experiment reported in this chapter tests for frequency-dependent success of male age classes and the effect of the operational sex ratio (OSR: ratio of the number of sexually active males to sexually active females) on male fertilization success in experimentally staged spawning events. In Chapter 4, I examine individual spawning events in detail. The purpose of this analysis is to test for behavioural and morphological correlates of male success in single spawning events. 6 In Chapter 5, I test for differences in competitive ability of sperm from two male age classes by performing in vitro competitive fertilizations. In addition to age, I test for morphological and genetic correlates of success in sperm competition. Ultimately, reproductive success is a function of a sperm cell fertilizing an egg. In Chapter 6, I examine the dynamics of fertilization in sockeye salmon and compare sperm attributes among male age classes. The results are examined in relation to male competition and mate choice. 7 Chapter 2 General Methods Study Area and Species Research was conducted between early September and early November in each of four spawning seasons (1994-97) at Babine Lake, British Columbia. Fisheries and Oceans Canada maintains two spawning channels on Fulton River which provide a convenient site for experimental studies and an abundance of sockeye salmon (Oncorhynchus nerka). The average escapement for 1990-96 was 494,000 (Wood et al. 1998). Spawning occurs from August to November and eggs hatch in the spring. Almost all (>98%) individuals spend one year in the lake before they smolt and migrate to the ocean (McDonald & Hume 1984). Thus, variation in spawning age results mainly from variation in time spent in the ocean. Age of spawning males ranges from three to five years old and females are either four or five years old. Between 1969 and 1977 the spawning population was comprised of 3.3-50.0% three-year-old males with an average of 22.9% (West 1978). Age composition of the spawning population in the years of this study were within this range. Fish Capture and Measurement I captured all fish with a dip net at the Fulton River counting fence. I used fish that were healthy in appearance with no evidence of prior spawning activity for all experiments. Fish with fin wear, excessive fungus or other damage were not used. For experiments in which I wished to compare male age classes, I selected three-year-old (3Y) and five-year-old (5Y) males based on body size (see below). I anaesthetized fish in 0.05% 2-phenoxyethanol for all body measurements. 8 Hypural length was measured as the distance (nearest 5 mm) along the lateral line from the posterior edge of the eye socket to the hypural bone at the base of the tail. Snout length was measured as the distance (nearest mm) from the posterior edge of the maxilla to the tip of the upper jaw. Hump size was measured as the vertical distance (nearest mm) between the lateral line and edge of body just anterior to the dorsal fin. I measured body weight to the nearest 50 g with a 10 kg spring scale. Enclosure Experiments I built enclosures (Fig. 2.1) in spawning channel #1 of Fulton River to perform spawning experiments (Chapters 3 and 4). Enclosures were constructed of wood-framed panels of plastic-coated fencing with a mesh size of 2.5 cm. Enclosure size was 3 m x 5 m with a water depth of approximately 0.5 m. I built three enclosures end-to-end in a section of the spawning channel that provided easy access and observation. I placed them along the shoreline of the channel so that I did not have to climb over a fence to enter them. Four 1.5 m x 1.5 m holding pens were placed on the upstream end of the enclosures. I attempted to mimic natural spawning habitat within each enclosure by liberally distributing cover objects (rocks and logs). I used plastic tarpaulins to cover gravel in the lower 2 m of each enclosure to ensure that spawning would occur only in the upstream end. This allowed males to have downstream access to nesting females. Fish Selection I captured unovulated females (no eggs expressed with gentle pressure on the abdomen) at the Fulton River counting fence and placed them in a holding pen. They were checked daily for signs of ovulation. When a female ovulated I anaesthetized her and attached a numbered disk tag through the dorsal musculature near the anterior end 9 of the dorsal fin. Females were used in enclosure experiments within three days of ovulation. The range of female hypural lengths was 385-520 mm (x = 456.7, SD = 36.2, N = 39). I also captured males at the Fulton River counting fence. I only used ripe males (milt easily expressed with gentle pressure to the abdomen). I captured both 3Y males and 5Y males, using body size to estimate age (Hanson Sz Smith 1967). I subsequently confirmed male age by examining otolith rings (Clutter Sz Whitesel 1956; K i m Sz Roberson 1968). I tagged each male with two small pairs of disc tags. Tags were coloured either black or white and each male in a trial had a unique combination of colours. Males were used in enclosure experiments on the day after their capture. Treatments I varied two factors in spawning trials resulting in four experimental treatments. First, the number of males used was either three or six. Second, the male age composition was either one-third 3Y males plus two-thirds 5Y males or two-thirds 3Y males plus one-third 5Y males. I distributed treatments as evenly as possible among enclosures and over time (Table 2.1). Female hypural length did not vary among treatments (F 3,3 5 = 0.233, P = 0.873). Spawning Trial Protocol I set up spawning trials in the late afternoon or evening when I placed tagged males into all three enclosures. At approximately 7:00 A . M . the next morning I placed a tagged female in each enclosure. I checked each enclosure every 30 min to keep track of nesting activity. When nesting and courtship were well established, I placed a video camera on a tripod overlooking the nest and began filming. I also continued to monitor the enclosures and perform behavioural observations every 30 min. Spawning occurred 10 between 2.5 hr and 9 hr after females were introduced to the pens (34 trials). In most cases where spawning did not occur in an enclosure by the end of the day, the fish were released. In five trials in which spawning did not occur by the end of the day, I removed the female to a holding pen and re-introduced her on the next day. When a spawning event occurred, I quickly entered the enclosure, covered the nest with plywood to avoid disturbing the eggs and removed all fish. The fish were killed with an overdose of anaesthetic. I performed body measurements on the dead fish and collected otoliths and a small sample of liver tissue (stored in 95% ethanol) for D N A analysis. I also removed testes from the males and measured their wet weight (nearest 0.1 g) on an electronic balance. I placed testes in a -20 ° C freezer until dry weight was determined. I collected eggs with a turkey baster while carefully removing rocks from the nest pocket. While collecting, I used a piece of plywood on the upstream edge of the nest to block the water current. A few eggs may have sunk between rocks deeper than I collected and some may have been lost in the current. The number of eggs lost was probably far too small to influence the results. It is unlikely that lost eggs represented a biased sample with respect to their paternity. After collecting the eggs, I used a shovel to thoroughly mix the gravel around the nest to ensure that no eggs remained and to ; smooth the substrate for the next trial. I placed eggs in perforated plastic tubes (30 cm x 5 cm) and buried them in the spawning channel to incubate. In late December I collected all egg tubes and brought them to U.B .C . to complete incubation. I removed unfertilized and dead eggs the remaining eggs were checked every two days for further mortality and hatching. After the eggs had hatched, I continued incubating the alevins until their yoke sacs were nearly used up. At this point I killed the fry with an overdose of anaesthetic and preserved them in 95% ethanol for D N A analysis. 11 I collected morphological data for fish in all 39 trials in which spawning occurred but pre-spawning behaviour, spawning behaviour, and paternity data were not collected for all trials (Table 2.1). As a result, sample sizes vary among analyses. Paternity data were collected for 32 trials by analyzing variation in microsatellite D N A (see below). To establish paternity, I used between one and six microsatellite loci per trial. For 35 fry from six trials (out of a total of 988 fry from 32 trials) I could not unambiguously establish paternity based on microsatellite variation alone. In these cases, I was able to eliminate potential sires based on clear video evidence that certain males did not participate in spawning. I determined paternity for 30 fry selected randomly from each trial. Some of the fry from two of the trials did not exhibit any maternal alleles. These fry were probably remainders of other clutches that did not get completely removed from the gravel. For each of these two trials I randomly selected an additional 30 fry for D N A analysis. I was able to establish the maternal genotype for the non-focal clutch in both cases and, therefore, I could determine which fry came from the focal clutch. Paternity estimates for these clutches were based on 40 and 48 fry, respectively. D N A Analysis I initially used standard phenol extractions to obtain D N A from both adult liver tissue and fry (Taylor et al. 1996). This method provided very pure samples of D N A but was time consuming. I therefore switched to a lysis buffer extraction method. This method was as reliable as phenol extraction in terms of the readability of autoradiographed polymerase chain reaction (PCR) products. The buffer contained 40 mM Tris pH 8.3, 50 mM K C l , 0.5% Tween 20, and 200 / ig/ml proteinase K . To extract D N A , a 6-10 mg sample of tissue was placed in 200 fA of lysis buffer, incubated at 65 0 C for 2 hr, and 95 ° C for 15 min. The sample was then centrifuged to pellet the 12 debris and the supernatant was used as the source of D N A template for P C R . I determined paternity of fry by examining variation in microsatellite D N A . Individuals were screened for as many loci as was necessary to determine paternity unambiguously. The loci that I used are given in Table 2.2. The loci were amplified using P C R with 10 p\ reaction volumes. Each reaction contained 1 x reaction buffer, 0.4 mM dNTP's, 0.05 pM of 3 2P-endlabelled primer 1, 0.25 pM of unlabelled primer 1, 0.6 pM of unlabelled primer 2, 1.5 m M M g C l 2 , 0.5 units of Taq polymerase, and 1.0 pl of the D N A template. The P C R conditions are described for each locus in Table 2.3. After amplification, each reaction tube received 10 pl of loading buffer and was stored at -20 ° C. Electrophoresis was performed using denaturing polyacrylamide (Long Ranger) gels at 5% concentration. Samples were denatured at 95 ° C for 5-10 minutes, placed on ice, and then loaded on the gel (5 pl per sample). Gels were vacuum-dried and exposed to autoradiographic film for 0.5-4 days. In gels where dams and putative sires were initially genotyped, M13 sequencing ladders were used as references to determine allele sizes. On gels where fry samples were genotyped, the dam and putative sires were included as references so that fry alleles could be compared directly to those of the parents. Gamete Collection To obtain gametes, fish were anaesthetized in 0.05% 2-phenoxyethanol. Before collecting gametes, fish were wiped dry with paper towels to prevent contamination by water, faeces, urine, and mucus. With gentle squeezing of the abdomen, eggs were extruded into clean dry beakers. To collect milt, a male was secured upside-down in a harness and gentle pressure was applied to its abdomen. A clean dry 5 ml syringe was used to collect milt as it emerged from the genital pore. Spermatocrit was measured for each milt sample by filling two capillary tubes with 13 milt and centrifuging them for ten minutes in a microhematocrit centrifuge. The average of these two values gives the proportion of milt occupied by sperm cells which is a strong correlate of number of sperm cells per unit volume of milt (Bouk & Jacobson 1976). I confirmed the correlation between spermatocrit and sperm count for sockeye salmon by performing sperm counts and measuring spermatocrit for 16 3Y males and 16 5Y males in 1997. To perform sperm counts I serially diluted 50 pA milt with river water to achieve a final dilution of 1/2000. A hemacytometer chamber was filled with diluted milt, allowed to settle for three minutes, and the number of cells in 80 squares was counted at a magnification of 100 x. Two milt dilutions were prepared and two separate counts were performed for each dilution. The sperm count for each individual was taken as the mean of these four counts. The relationship between spermatocrit and sperm cell count did not vary with male age (Fig. 2.2). The correlation between sperm count and arcsine-square root transformed spermatocrit for the two age classes combined was 0.856. Data Analysis Statistical analyses were performed using R (Ihaka & Gentleman 1996) and J M P IN® (Sail &; Lehman 1996). I tested all data for normality using qq-plots and Shapiro-Wilks test before proceeding with parametric statistical analyses. I also examined normality of residuals in cases where linear models were applied. Data that could not be normalized with transformations were subjected to non-parametric tests. In cases where I performed multiple statistical tests, I adjusted the a-levels using the sequential Bonferroni procedure (Rice 1989). These are indicated as asonf- A l l tests are 2-tailed unless indicated otherwise. Wilcoxon matched-pairs signed-ranks test is abbreviated as W M S T . A l l t-tests are two-sample tests unless otherwise indicated. 14 -a cj o CD O o C J fe ce a -a M - H o CO CD ft co O • i-H -fl cj fe CO C J -O S ce CO a C J ;>> C J C J S-i -fl - u C J > O ce bJO fl ' f l ! a CO =4-1 O fl O + J =3 X> co Q ' f l •~-> C J ce fcuO fl • t-H fl 1 ft CO .a n I . 2 ft 2 co -fl i C J ;0-05 Oi L O OM Oi >> CJ fl CJ fl cr C J CO O ce oo oo co oo C M o i oo r— L O C M CO CO N C M O C O C M L O L O T P L O L O T P C M C M ? bX) -fl O - ~ O .SP — - f l — ' - f l ... -fl -fl — 1 ' ' - f l - f l CN H CN H CN CN CO C J 13 s C J _> O cci fl X C J co co cj 13 a C J > • —H cj ce ce fl C J co C J a A C J -fl o . 2 ce ce S-l X C J co 13 fl . 2 ce C J ft O C J C J O C J O -f l fl O fl C J >, ce 13 fl C J C J CO C J l-l C J fl - Q C J ft ce o C J '> + J O fl C J C J fe -fl o Ifl fe C J CO o -fl C J l-l C J fe CO ce bJO - u fl ce • T-H fl fe hO ce fl ft ' f l CO fe "O ce C J ft > CO !-l C J C J CO Xi fl O O CO C M C J Xi 15 Table 2.2: Characteristics of microsatellite loci used for paternity determination. Locus Repeat type Alleles Reference Omy77 di- 7 Morris et al. (1996) One^i8 di- 7 Scribner et al. (1996) One/xl4 di- 7 Scribner et al. (1996) OtslOO di-, tetra- 9 Nelson et al. (1998) Otsl03 tetra- 15 Beacham et al. (1998) Otsl07 di-, tetra 6 Beacham et al. (1998) Ssa85 di- 16 O'Reilly et al. (1996) 16 Table 2.3: P C R conditions for each locus. For each set of cycles the conditions for denaturing, annealing, and extending are given as time-temperature where time is in seconds and temperature is in degrees Celsius. The number of cycles in the third set of cycles varied among loci and the ending cycle number is denoted as i. At the end of these cycles, an additional extension for 5 min at 72 0 C was added for all loci. Locus Cycle 1 Cycles 2-6 Cycles 7-i i Omy77 120-95, 60-56, 60-72 60-94, 60-56, 60-72 60-92, 60-56, 60-72 31 One//8 120-95, 60-62, 60-72 60-94, 60-62, 60-72 60-92, 60-60, 60-72 31 One/jl4 120-95, 60-62, 60-72 60-94, 60-62, 60-72 60-92, 60-60, 60-72 31 OtslOO 120-94, 60-60, 20-72 20-94, 20-60, 20-72 20-90, 20-58, 20-72 26 Otsl03 120-94, 60-58, 20-72 20-94, 20-58, 20-72 20-94, 20-56, 20-72 26 Otsl07 120-94, 60-50, 20-72 20-94, 20-50, 20-72 20-94, 20-48, 20-72 26 Ssa85 60-94, 20-58, 20-72 20-94, 20-58, 20-72 20-90, 20-58, 20-72 26 17 1m Figure 2.1: Schematic drawing of the spawning enclosures from above showing the spawning channel (C), holding pens (H), gravel substrate in each enclosure (G), tarpaulin-covered substrate in each enclosure (T), and shoreline (S). The arrow indicates the direction of water flow. 18 55 J 50 J 45 J - 40 J o o 13 § 35 a> a. CO 30 25 20 Sperm count (106 cells per ul milt) Figure 2.2: Relationship between spermatocrit and sperm count for 3Y males (o) and 5Y males (x). The slope of the relationship (using an arcsine-square root transformation on spermatocrit) does not differ between the two age classes ( F i i 2 8 = 0.451, P = 0.507), nor does the elevation ( F i , 2 9 = 1.615, P = 0.214). 19 Chapter 3 Frequency-Dependent Reproductive Success in Individual Spawning Events and the Effect of Operational Sex Ratio Introduction One of the most powerful mechanisms for maintaining genetic and phenotypic variation is negative frequency-dependent selection (FDS). Negative FDS occurs when fitness of a genotype decreases as the genotype becomes more common in a population (Ayala & Campbell 1974). For example, predators may preferentially feed on common forms of prey, resulting in higher fitness for rare forms (Allen 1988). Gadgil (1972) hypothesized that negative FDS could also maintain alternative male reproductive phenotypes (ARP's). He proposed that, in a population where all males use a mating behaviour such as territory defense, a single male that adopts an alternative mating behaviour such as sneaking would have very high fitness because he avoids the cost of territoriality but is still able able to mate successfully. Conversely, if sneaking males become very common then males that adopt territorial behaviour may be very successful because of limited competition with other territorial males. Early studies (reviewed in: Arak 1984; Dominey 1984; Gross 1984) attempted to place male ARP ' s in the theoretical framework provided by evolutionary game theory (Maynard Smith 1982). However, very few examples of male ARP ' s fit the assumptions of simple game theory models. That is, ARP ' s do not usually result from genetic polymorphisms but are instead a function of male condition (Dominey 1984; Gross 1996). Nevertheless, recent theoretical studies have shown that FDS can play an 20 important evolutionary role in maintaining variation even when alternative phenotypes do not result from genetic polymorphisms (e.g., Hazel et al. 1990; Repka & Gross 1995; Day & Taylor 1996; Hines & Turelli 1997; Roff 1998). In spite of the potential importance of FDS in the evolution of male ARP 's , only two studies have tested for its occurrence. Gross (1991) used a behavioural measure of reproductive success in naturally spawning bluegill sunfish (Lepomis macrochirus) to compare cuckolder (mature early, mate by sneaking) male and parental (mature late, build nests to attract females) male fitness. He found that success of cuckolders relative to parentals first increased as the frequency of cuckolder males increased (positive FDS) but then decreased at high cuckolder frequency (negative FDS). In a population of side-blotched lizards (Uta stansburiana) with three male reproductive phenotypes, Sinervo & Lively (1996) measured male mating success over six years and found that the morph that was rarest in a given year had the highest mating success. With the dearth of studies of FDS on male ARP 's , there has also been little consideration of the nature of FDS. Male reproductive success is a function of the number of matings and the number of young sired per mating and FDS could act on either (or both) of these components. FDS in the number of matings may be illustrated by a simple hypothetical example of a species with two male size classes. If the size classes use different reproductive tactics such as courting and sneaking, then competition within age classes could result in a single large male and a single small male mating with each female. Imagine a population consisting of six females and twelve males (Fig. 3.1a). Two of the large males may be able to mate twice and two of the small males do not mate at all. The average number of matings is 0.75 for small males and 1.5 for large males. Thus small males are half as successful as large males on average. If the population has instead eight large males and four small males (Fig. 3.1b) then competition among large males could exclude two of them from 21 mating. This would allow two of the small males to mate twice. Therefore the average number of matings would be 1.5 for small males and 0.75 for large males. Comparing these two scenarios, it is apparent that the number of matings is negatively frequency-dependent. That is, the number of matings that a male can expect to achieve depend both on his phenotype and on the phenotypic composition of the population. This is similar to the mechanism thought to produce FDS on sex ratio (Fisher 1930). FDS could also act on fertilization success at individual mating events (Myers 1986). In species with external fertilization, it is possible for more than one male to participate in the fertilization of a female's eggs. When this happens, large males may be able to force their way into advantageous mating positions close to the female. If large males are common, space may be limited and outer males may have low success. A small male, on the other hand, could squeeze between large males and achieve moderate or even high success. However, if small males are common they would also compete for positions resulting in low success for some small males. In this situation, rare large males would be able to force their way into favourable positions and achieve high success. Given this effect of male size on fertilization success, male reproductive success would again be negatively frequency-dependent. Operational sex ratio (OSR: ratio of the number of sexually active males to sexually active females) can also influence the expression and success of ARP ' s . For example, male behaviours such as sneaking and female mimicry may be absent or rare when the OSR is female-biased, but become more common as the OSR becomes more male-biased (Verrell 1983; Kodric-Brown 1988; Quinn et al. 1996; Jirotkul 1999). In the water strider Gerris odontogaster, there is substantial variation among males in the length of their abdominal processes, used to grasp females (Arnqvist 1992). Males with longer processes have an advantage in mating when the OSR is not strongly male-biased because females resist mating. However, as the sex ratio becomes more male-biased, 22 females become less reluctant to mate and the advantage of long abdominal processes decreases. Thus, the success of male ARP ' s may depend on the local OSR. The potential for OSR to influence FDS has not been addressed. However, OSR can be affected by population density (Conner 1989; Noltie 1990) and several studies have examined the effect of population density on alternative phenotypes. Early game theory models assumed large and constant population sizes but this assumption is unlikely to be valid in most cases. Changes in population density may not affect all phenotypes equally and therefore frequency-dependence and stability of ARP ' s may be influenced by density (Cressman & Dash 1987). For example, Barrow's goldeneye (Bucephala islandica) females sometimes lay eggs in nests of other females. This behaviour can be maintained by FDS but only at high population densities (Eadie & Fryxell 1992). Experimental studies of predation have found that predators select disproportionately more rare food when food density is high, but disproportionately more common food when food density is low (Allen et al. 1998; Weale et al. 2000). Thus, an increase in density results in a change in form of selection from negative FDS at low density to positive FDS at high density. The first objective of this study was to test for the presence of frequency-dependent reproductive success of male sockeye salmon (Oncorhynchus nerka) age classes in single spawning events. I staged spawning events with single females and groups of males in enclosures so that male group composition (i.e., number of males in each of two age classes) could be manipulated. If negative frequency-dependent success occurs at this level, the mean success of individuals in an age class (relative to the second age class) should decrease as the frequency of individuals in that age class increases in a spawning event. My second objective was to examine the effect of local OSR on FDS in sockeye salmon spawning events. While other studies have examined the effect of population 23 density on FDS (see above), the literature does not provide a basis for predictions about the effect of OSR. Therefore, I developed a mathematical model to simulate male success in spawning events in which frequency composition of male age classes and OSR vary (Appendix A) . This model formalizes the qualitative description given above of how FDS could occur at single spawning events. It predicts that frequency-dependence will be present at the two levels of OSR that I use in this experiment and that the intensity of FDS will increase with an increase in OSR (i.e., an increase in the number of males present in a spawning group). The mean success of individuals in an age class is a function both of the proportion of eggs fertilized by individuals that participated in spawning and of the proportion of males that participated. Thus my third objective was to examine the influence of spawning participation and fertilization success (of participating males) on frequency-dependent reproductive success. Methods The general protocol for this experiment was to place a female and a group of males in an enclosure, allow them to spawn, and determine paternity of the resulting offspring. The groups consisted of either three or six males with one female (low and high OSR, respectively) with two levels of 3Y male frequency (one-third and two-thirds). For paternity analyses I included eight spawning trials performed at each of the four combinations of OSR and 3Y male frequency. I did not obtain both paternity data and spawning participation data for all trials so sample sizes vary among analyses. Details of the experimental protocol and paternity analysis are given in Chapter 2. To avoid pseudoreplication (Hurlbert 1984) in tests for frequency-dependent fertilization success in the spawning trials, I considered each trial and not each male to be an independent replicate. For each trial I determined the mean proportion of fry 24 sired per male for both the 3Y and 5Y age classes. Relative success of individuals in an age class can be calculated as the mean success of individuals in that age class divided by the mean success of individuals in the other age class (Gross 1991). However, in my experiment there were trials in which the mean success of one age class was zero. Therefore I calculated relative success of individuals in an age class in each trial as the difference between the mean successes of the two age classes. Specifically, I defined relative success in each trial to be arcsine-transformed mean 3Y success minus arcsine-transformed mean 5Y success. Negative frequency-dependent success occurs if relative success of 3Y males (as defined above) decreases as 3Y male frequency increases (FDS can also be tested in terms of changes in relative 5Y male success, but the result would be the same). I used 1-tailed t-tests to test for this effect at both levels of OSR. When variances were significantly different, I used Welch's t-test. Frequency-dependence in the proportion of males spawning in each age class could be tested using log-linear analysis. However, I could not do this for my experiment because the number of males that did not spawn was small (zero or one) in three cells. R e s u l t s At low OSR and low 3Y male frequency, 3Y males sired more offspring on average than 5Y males (Fig. 3.2). When 3Y male frequency was high, 5Y males sired more offspring on average than 3Y males (Fig. 3.2). The difference between transformed means of 3Y and 5Y males thus decreased with increasing 3Y male frequency and this effect was significant (Welch's t-test: t= 1.916, df = 9.351, Pi-taiied = 0.043), indicating that success is negatively frequency-dependent when OSR is low. At high OSR, there was little difference between the 3Y and 5Y age classes in mean success of individuals (Fig. 3.3). The difference between transformed means of 3Y 25 and 5Y males did not decrease with increased 3Y male frequency (t-test: t= —0.962, df = 14, Pi-taiied = 0.824), indicating that success is not negatively frequency-dependent when OSR is high. Some of the variation in mean proportion of offspring sired by individuals in each age class may have resulted from variation in spawning participation while some may have resulted from variation in success of males that did participate. At low OSR, the proportion of 3Y males participating in spawning decreased slightly as the frequency of 3Y males increased while the proportion of 5Y males participating increased substantially (Fig. 3.4). Thus, as with the previous analyses, spawning participation appeared to be negatively frequency-dependent at low OSR. At high OSR, the proportion of both 3Y and 5Y males participating in spawning decreased as the frequency of 3Y males increased (Fig. 3.5). The frequency of 3Y males had little effect on the relative proportions of males spawning in the two age classes, so spawning participation does not appear to be frequency-dependent at high OSR. Another factor that could contribute to frequency-dependent success is the success of males that participated in spawning. The success of males that participated in spawning when OSR was low appeared to be negatively frequency-dependent (Fig. 3.6), but the trend was not statistically significant (t-test: t= 0.942, df = 13, Pi-taiied = 0.182). The success of males that participated in spawning when OSR was high was, again, not frequency-dependent (Fig. 3.7, t-test: t= —0.548, df = 9, Pl-tailed = 0.701). Discussion This is the first systematic experimental study of FDS on male ARP ' s . There is clear evidence of fine-scale negative FDS on male age classes when OSR is low but not when OSR is high. At low OSR, frequency-dependence is mediated by two factors. 26 First, the alpha males (5Y males in all but one trial) were often able to exclude other 5Y males from participating in spawning (Fig. 3.4). Second, 3Y males are rarely excluded from spawning and so competition between 3Y males results in a trend towards frequency-dependent fertilization success among males that participate (Fig. 3.6). When OSR is high, alpha males are less able to exclude other males from spawning. Fertilization success in high OSR spawnings appears to be mediated through scramble competition (as defined by Andersson (1994)). Thus, fertilization success did not differ between male age classes and did not depend on 3Y male frequency. This result is opposite to that predicted by the spawning model in Appendix A , which predicted that the intensity of FDS should increase with OSR. This could be resolved by allowing the model parameters to vary with OSR. The model did not account for changes in the nature of male competition when OSR changes. The effects of male group size and composition may have implications for the distribution of males in a breeding population. At low OSR, a male's success would be maximized by choosing a spawning group in which his age class is rare or absent. When OSR is high, on the other hand, there is no expected payoff from choosing groups based on age composition. Thus, variation in age composition of male groups would be greater at high OSR than at low OSR. A complete understanding of FDS in this population can only be achieved by including all three male age classes and examining competition both for females and for fertilizations. However, the use of only two age classes is a reasonable simplification because many naturally spawning groups contain only two age classes (personal observation). The addition of a third age class may result in complex interactions between OSR and frequency. Quantifying these interactions would require a large number of treatments (Antonovics &; Kareiva 1988). An additional factor in some populations (but not the one used in this experiment) is the presence of kokanee, a 27 non-anadromous form of sockeye salmon. Kokanee males are often included in pre-spawning groups of sockeye salmon (Foote &; Larkin 1988) but the extent to which they are successful in fertilizing eggs and the impact they have on FDS on sockeye males is not known. The study of FDS on the number of matings that males achieve will be much more' logistically difficult than that of FDS on fertilization success in single spawning events. Larger enclosures with several females could be used but this would restrict male movements between spawning events and may result in spurious measurements of reproductive success. Naturally spawning male salmon can make extensive movements over the course of their breeding lives (Healey & Prince 1998;personal observation). Therefore, data on the number of matings would ideally be obtained from whole streams, making experimental manipulation difficult. The group sizes used in this experiment are within the range observed in natural sockeye salmon populations (Hanson &; Smith 1967; Quinn et al. 1996) but there are few published data on the number of males that participate in natural spawning events. The number of males in pre-spawning groups varies with factors such as female size and imminence of spawning (Hanson & Smith 1967; Sargent et al. 1986; Noltie 1990). It is also influenced by the population OSR which can vary between years, within years, and even within days (Quinn et al. 1996). Within-year variation in OSR occurs because of differences between males and females in arrival date (Morbey 2000), lifespan (Hendry et al. 1995; Quinn et al. 1996), and movement (Quinn & Foote 1994; Healey & Prince 1998). The number of males present in pre-spawning groups can also be affected by population density. Noltie (1990) found that pink salmon (O. gorbuscha) females nested far from each other when population density was low, and this allowed large males to monopolize females, leaving many males excluded from spawning groups. In 28 contrast, when population density was high, females nested close together and males were not able to monopolize access. This resulted in larger pre-spawning groups. A decrease in the ability of males to monopolize females as density increases has also been observed in coho salmon (0. kisutch: Fleming & Gross 1994). If spawning group size is related to population density in sockeye salmon, the results of the present study have important implications for sockeye salmon population ecology. In this study I found evidence that selection on male age at maturity varies with the number of males in spawning groups. Increasing numbers of males in spawning groups appears to lead to a breakdown of FDS with no difference in fertilization success between age classes. If mean success of age classes in single spawning events is indicative of lifetime reproductive success, younger age classes would have higher fitness than older age classes (since generation time is an important component of fitness (Roff 1992)). In agreement with this conclusion, the proportion of early-maturing male coho salmon is correlated with population density (Young 1999). This could result in positive feedback, with increasing population density resulting in selection for reduced age at maturity which, in turn, would lead to faster population growth. A positive relationship between population density and frequency of sneaker male phenotypes has also been found in grasshoppers (Ligurotettix coquilletti: Greenfield & Shelly 1985) and seabass (Serranus subligarius: Oliver 1997). In the fungus beetle Bolitotherus cornutus males with long horns have greater insemination rates than males with short horns, but this advantage is more pronounced in low density populations (Conner 1989). Population dynamics can thus interact in a complex way with the phenotypic composition of a population (Parker 1985). This study also has implications for the population genetics of sockeye salmon. If age at maturity has a genetic basis, FDS could result in maintenance of genetic variation. Interestingly, Roff (1998) found that the amount of additive genetic variance 29 maintained by FDS increases with population size. However, his model did not take into account the possibility that FDS may itself be influenced by population size. The potentially high reproductive success of early-maturing males has led to speculation that early maturation is the 'best' evolutionary strategy for Pacific salmon (Gross 1996; Foote et al. 1997). This hypothesis is intriguing, but jacks (early-maturing males) are either rare or absent in most salmon populations (Groot & Margolis 1991). Several factors may explain this paradox. First, population densities may not be high enough for selection to favour early maturation. Population density can be limited by many factors in both freshwater and saltwater life stages. For example, stream habitat and water quality can influence egg survival, food abundance and intraspecific competition can limit population size in rearing lakes; and ocean survival can be limited by fishing pressure, food abundance, and temperature (Burgner 1991). Second, there is evidence that maturation in salmon does not occur until a size threshold is reached (Wood & Foote 1996). If male growth is limited by factors such as food abundance or temperature, then early maturation could be constrained. The evolution of this size threshold has not been well studied. There is evidence to suggest that long migration to spawning streams selects against early maturation (Young 1999) but many Alaskan populations of sockeye salmon that have very short upstream migrations exhibit very low proportions of jacks (Burgner 1991). Finally, selection on age at maturity in females may favour later maturation because fecundity increases with body size (Burgner 1991). Male age at maturity probably has a genetic component that is partially inherited from mothers (Heath et al. 1994; Silverstein & Hershberger 1992). Therefore, selection that favours early maturity in males could be counteracted by selection favouring late maturity in females. The evolution of age at maturity in male sockeye salmon has been influenced by many of factors. In this study I found that competition for fertilizations at individual 30 spawning events can lead to FDS. This may be one of the factors contributing to variation in male age at maturity in sockeye salmon. 31 a cf cf \ / 9 9 9 cf cf 9 9 9 + + + f f fi cf cf cf cf T + + .77 /7 .Tf cf cf cf cf b aid rcid 9 9 9 d dd d '"'U '""^  9 9 9 + + + cf cf + + + 'cf 'V . Figure 3.1: Frequency-dependent mating success in a population with large and small male phenotypes. Arrows indicate mating between a male and female. 32 Figure 3.2: Mean proportion of fry fertilized per male for 3Y (closed circles) and 5Y (open circles) males in spawning trials with three males. Error bars are ± 1 SE (back-transformed from arcsine transformations). Sample sizes indicate the number of spawning trials on which the calculations were based. 33 T3 C o o Q . O c CO CD 0.8 J 0.6 0.4 -J 0.2 J 0.0 2/6 (N=8) 4/6 (N=8) 3Y male frequency Figure 3.3: Mean proportion of fry fertilized per male for 3Y (closed circles) and 5Y (open circles) males in spawning trials with six males. Error bars are ± 1 SE (back-transformed from arcsine-squareroot transformations). Sample sizes indicate the number of spawning trials on which the calculations were based. 34 1/3 (A/=11) 3Y male frequency 2/3 (W=9) Figure 3.4: Proportion of males that spawned in spawning trials with three males. Includes all trials in which spawning status is known for all male. Closed circles represent 3 Y males and open circles represent 5 Y males. Sample sizes indicate the number of spawning trials on which the calculations were based. 35 Figure 3.5: Proportion of males that spawned in spawning trials with six males. Includes all trials in which spawning status is known for all males. Closed circles represent 3Y males and open circles represent 5Y males. Sample sizes indicate the number of spawning trials on which the calculations were based. 36 Figure 3.6: Mean proportion of fry fertilized per male for 3Y (closed circles) and 5Y (open circles) males in spawning trials with three males, including only males that participated in spawning. Error bars are ± 1 SE (back-transformed from arcsine transformations). Sample sizes indicate the number of spawning trials on which the calculations were based. 37 T3 CD C o '•e o a. o i CL C CO cu 0.8 0.6 J 0.4 J 0.2 0.0 2/6 (N=8) 3Y male frequency 4/6 (A/=8) Figure 3.7: Mean proportion of fry fertilized per male for 3Y (closed circles) and 5Y (open circles) males that spawned in spawning trials with six males, including only males that participated in spawning. Error bars are ± 1 SE (back-transformed from arcsine-squareroot transformations). Sample sizes indicate the number of spawning trials on which the calculations were based. 38 Chapter 4 Factors Influencing Male Fertilization Success in Single Spawning Events Introduction When a male encounters a sexually receptive female, the probability that he will mate with her may depend on his success in competition with other males. In many animals male competitive ability is strongly related to body size (Wilson 1975). This is the case in most salmon species, in which males form dominance hierarchies downstream from nesting females. In these aggregations dominant males maintain positions closer to the female than subordinate males (Hanson & Smith 1967; Chebanov 1980; Maekawa 1983; Keenleyside & Dupuis 1988; Jarvi 1990; Kitano et al 1994). However, male positions may not be entirely a result of aggressive competition between males. Very small males may obtain positions close to a female because their colouration and small size make them cryptic (Gross 1985; Healey &; Prince 1998) or unlikely to elicit aggression from other males (Koseki & Maekawa 2000). Another factor that could influence male positions is female behaviour. Although female salmon are less aggressive than males and tend to direct more aggression towards females than males, females are more aggressive to small males than large males (Hanson & Smith 1967;but see Healey & Prince (1998)). Close proximity to the female has usually been assumed to allow dominant males closer and quicker access to the female during oviposition and, consequently, higher reproductive success. One advantage of being close to the female is that the male may be exposed to reproductive pheromones that influence the male's hormones and increase the number of sperm available for ejaculation (Liley & Kroon 1995). Gross 39 (1985) and Fleming & Gross (1992, 1994) used distance from the female at an unspecified time before spawning as a measure of reproductive success for male coho salmon (Oncorhynchus kisutch). This was based on an apparent correlation between male proximity to the female and order of male entry into the nest during spawning (see Gross (1985)) and Schroder's (1982) study of captive chum salmon (0. keta) in which male dominance was apparently correlated with the proportion of eggs fertilized. However, there are assumptions involved in using distance from the female as a measure of male reproductive success in salmon. First, sperm competition among males is assumed to be random. This assumption will be addressed below, but it should be noted that Schroder's (1982) study included a narrow size range of males, so his results cannot necessarily be extrapolated to compare success of males from different age classes (whose sperm output and spawning behaviour may differ). Second, group composition and relative positions are assumed to remain constant between the observation and time of spawning. Foote et al. (1997) found that when a three-year-old (3Y) male sockeye salmon (0. nerka) competed with an older male in spawning, the older male always held a closer position to the female before spawning, but during spawning the 3Y male usually rushed in on the side of the female opposite to that of the older male. Therefore, the relationship between pairing success (proximity to female prior to spawning) and reproductive success of male salmon is not clear. At the time of spawning, male behavioural tactics may play an important role in determining reproductive success. In animals with external fertilization, males that release sperm in closer proximity to the eggs have a better chance of fertilizing eggs than males that release sperm from farther away (e.g. horseshoe crabs (Brockmann et al. 1994; Brockman et al. 2000); marine invertebrates (Yund & McCartney 1994)). Timing of sperm release is also likely to be important for two reasons. First, fertilization takes place very rapidly after gametes come in contact (Ginzburg 1972;Chapter 6). If a 40 male releases sperm too long after the female releases eggs, he risks losing fertilizations to males that released sperm sooner. Second, salmon sperm is short-lived (motility ceases in 30-60 s (Ginzburg 1972;Chapter 6). Thus if a male releases gametes before the female, his sperm cells may not be active long enough to fertilize any eggs. The influence of timing and location of sperm release on male fertilization success have not been examined in detail in any fish species, but three studies provide anecdotal information. Schroder (1982) videotaped six captive spawning groups consisting of one female and three male chum salmon each and determined paternity of the resulting offspring. He found that alpha males sired an average of 72% of young, first satellites sired 26%, and second satellites sired 3%. Analyses of spawning positions and behaviour were not performed, but Schroder concluded that "the closer and more attuned a male was to a female, the more eggs he was able to fertilize" (p. 284). Maekawa &; Onozato (1986) performed a similar experiment with Miyabe charr (Salvelinus malma miyabei), except that only two males were present with each female (a large male and a small sneaker male). They found that sneaker males succeeded in fertilizing eggs both when positioned beside the female and when the alpha male was between the sneaker and the female. Foote et al. (1997) made a similar observation in spawnings involving sockeye salmon. Three-year-old male fertilization success was not obviously different between spawnings in which the 3Y male was beside the female and those in which the alpha male was between the 3Y male and the female. Thus, the effect of spawning position on reproductive success remains unclear, particularly when males of different phenotypes are involved. The purpose of this study was to examine competition among sockeye salmon males that had the opportunity to spawn with a nesting female in an enclosure. I analyzed competition between two male age classes as well as competition within age classes. Specifically, I tested how male age, morphology, and behaviour influenced male 41 proximity to the female prior to spawning, probability of participating in spawning, and behaviour during spawning. I also tested for correlates of male fertilization success including age, sperm concentration (number of sperm per unit volume of milt), an index of potential ejaculate size, behaviour and position prior to spawning, and behaviour and position during spawning. Methods The general protocol for this experiment was to place a female and a group of males in an enclosure, allow them to spawn, and determine paternity of the resulting offspring. Details of the protocol are given in Chapter 2. Once the female was settled on a site and began digging a nest, I performed a 5 min behavioural observation every 30 min when possible (observations were sometimes pre-empted by more pressing tasks). Data were recorded on audio tape. The behaviours were: digging, satellite intrusion, aggression, and courtship (Table 4.1). These classifications are similar to those used in other studies of salmonids (e.g., Hanson & Smith 1967; Tautz & Groot 1975; Matsushima et al. 1989; Healey & Prince 1998; Quinn 1999). The identity of individuals involved in each behaviour was also recorded. At the end of each observation I noted the location where each male spent most of its time. Distance from the female was estimated to the nearest 0.25 m for males less than 1 m from the female and to the nearest 0.5 m for males that were 1 m or greater from the female. Normally one male spent most of his time at the female's side and exhibited frequent courtship behaviours. This individual was designated as the alpha male. To examine changes in behaviour as spawning approached, I selected two 5 min observation periods designated as first and last. The last observation was the last observation before spawning occurred. The first observation was the first behavioural observation after the female had settled on a site and begun digging a nest. For the 42 analyses, I included a trial's last observation only if it was done less than one hour before spawning. I included a trial's first observation only if it was done at least two hours before the last observation. A given trial may have data for the last observation only, both first and last observations, or neither. Videotapes were examined to analyze spawning events. A spawning event consisted of a female and one or more males curving their bodies into the nest pocket, quivering, widely opening their jaws (gaping), and releasing gametes. Females normally begin to cover the eggs as soon as they stop gaping but I interrupted them after one or two digs. By playing tapes one frame at a time I noted the following times for each fish that participated in spawning: start of gape, end of gape, and start of first female dig. I also identified all males that participated and noted the positions in which they spawned. Males that spawned beside the female were assigned a spawning position of one. Males that spawned with other males between themselves and the female were assigned a spawning position of one plus the number of males between them and the female. Some males' spawning positions changed over the course of a spawning event, for example when another male inserted itself between the focal male and the female. In these cases the male was assigned the mean of the two spawning positions that it occupied over the course of spawning. Data Analysis Complete data were not collected for all trials (see Chapter 2). Therefore, sample size varies among analyses. For several of the ANOVA's in this chapter, I present the data in figures. In these cases, and those dealing with the morphological principal components analysis (see below), A N O V A tables are included in Appendix B. The data in this study consisted of a number of independent spawning trials each of which included several males. Males within trials are not independent because males influence each other. To avoid artificially inflating the degrees of freedom for the 43 analyses (Hurlbert 1984), I used trials (and not individual males) as replicates wherever possible. I examined the effect of four morphological traits on male status. The first was hypural length which is a measure of body size. The others were body weight, hump size, and upper jaw length. Rather than using these three variables directly in subsequent analyses, I adjusted them for body size by using residuals from regressions of each log-transformed variable versus log-transformed hypural length (Table B . l ) . Some of the resulting variables were correlated (Table B.2) so I performed a principal components analysis to create independent variables (Table B.3). The first two components accounted for over 75% of the variance and I used these in subsequent analyses. The first component is an index of body condition (body weight adjusted for body size) and relative size of secondary sexual traits and the second component is an index of body size. When more than one male spawns with a female, a male's fertilization success may be influenced by the amount of sperm he releases relative to other males (Parker 1990a). I did not measure ejaculate size but I could compare testes weight after spawning to testes weights of pre-spawning males. To do this I measured dry testes weight and hypural length of 56 males at the Fulton River counting fence in the early part of the 1996 and 1997 spawning seasons. None of these males were spermiating. The log-log regression of testes dry weight versus hypural length was statistically significant (r2 = 0.796, FXM = 211.0, P < 0.001). I calculated testes depletion for males in the spawning trials as their observed testes dry weight minus their expected testes dry weight (for their hypural length) from this regression line. Fleming & Gross (1992) found that testes depletion was correlated with the number of times that male coho salmon participated in spawning. 44 Results Behaviour and Male Positions Before Spawning Alpha males were usually established as soon as the female settled on a nest site. There were only three trials in which the alpha males were supplanted between the first and last behavioural observations (out of 29 trials for which I had data from both first and last observations). The alpha male in the last behavioural observation before spawning was a 5Y male in all but one trial. I examined morphological correlates of male status using a multiple logistic regression of body size (PCI from the principal components analysis) and relative size of secondary sexual traits (PC2) versus male status (alpha or satellite) for 83 5Y males in 38 trials (trial number was included as a random factor). The probability of becoming an alpha male increased with body size (z = 3.941, P < 0.001) and with relative size of secondary sexual traits (z = 3.892, P < 0.001). Most aggression was initiated by alpha males and their aggression rate was higher just before spawning than it was in the first observation (Fig. 4.1; W M S T : V = 21, N = 19, P = 0.005). Frequency of alpha male aggression in the first observation was influenced significantly by an interaction between the operational sex ratio (OSR) and age class frequency composition of male groups (Table B.4; Fig. 4.2a). In first observations there was more alpha male aggression in six-male groups with two-thirds 3Y males than in the other three OSR-frequency combinations (Tukey test: P's < 0.025). Frequency of alpha male aggression in the last observation before spawning did not vary with OSR or the frequency composition of age classes (Table B.4; Fig,4.2b). To test if alpha males were more aggressive to 3Y or 5Y males, I considered only trials in which alpha males had one 3Y satellite and one 5Y satellite to contend with. In these trials, the frequency of alpha male aggression towards 3Y satellites did not 45 differ from the frequency of alpha male aggression towards 5Y satellites in either first (WMST: V = 21.5, N = 8, P = 0.236) or last (WMST: V = 26, iV = 11, P = 0.563) observations. The frequency of courtship behaviour exhibited by alpha males increased from the first to last observation in 16 of 19 trials, and this change was significant (WMST: V = 2, N = 19, P < 0.001). Courtship frequency did not depend on OSR or the frequency composition of age classes in either first or last observations (Table B.5; Fig. 4.3). The frequency of satellite male intrusions into nests did not change between first and last observations (WMST: V = 73.5, N = 19, P = 0.905). In first observations the mean intrusion rate was greater in six-male than in three-male groups, but it was not affected significantly by the age composition of the group (Table B.6; Fig. 4.4a). In last observations the effect of OSR was not significant, but the intrusion rate tended to be greater in groups comprised of two-thirds 3Y males than in groups comprised of one-third 3Y males (Table B.6; Fig. 4.4b). Rates.of alpha male aggression, alpha male courtship, and satellite male intrusions were not correlated significantly in either first or last observations, but the magnitude of correlations was stronger in first observations (Table 4.2). Distances of satellite males from the female did not change as spawning approached. Distance from the female in the first observation was not different from that in the last observation for either 3Y satellite males (Fig. 4.5a; W M S T for 46 males in 24 trials: V = 270, P = 0.665) or 5Y satellite males (Fig. 4.5b; W M S T for 30 males in 19 trials: V = 115, P = 0.720). In the last observation before spawning, the mean distance of 5Y males (within trials) from the female was less than that of 3Y males (paired t-test: mean difference = -0.40 m, t = -2.338, df = 31, P = 0.026). However, most alpha males 46 were 5Y, and their distance from the female was 0 m. When alpha males were excluded from the analysis, 3Y satellite males were significantly closer to the female prior to spawning than were 5Y satellite males (paired i-test: mean difference = 0.57 m, t = 2.727, df = 24, P = 0.012). Distance from the female was not related to body size or other morphological measurements for either 3Y satellite males or 5Y satellite males (Table 4.3). Time to Spawn For the trials in which spawning occurred on the first day (N = 34), the time between introduction of the female to the enclosure and spawning tended to be greater in groups with six males than in groups with three males and was greater in groups with 1/3 3Y males than in groups with 2/3 3Y males (Table B.7; Fig. 4.6). Of the five trials in which spawning did not occur on the first day, three were six-male groups, two were three-male groups, one had 2/3 3Y males and four had 1/3 3Y males. Behaviour and Positions During Spawning Male distance from the female in the last observation before spawning was a significant predictor of whether or not he would participate in spawning (logistic regression including 126 males in 30 trials with trial number included as a random factor: z = —4.505, P < 0.001, model r 2 = 0.544). This distance was also correlated with the rank order in which males began to gape (Fig. 4.7: rs = 0.553, N = 26, P < 0.001) and with time that males began to gape relative to the start of the female's gape (Fig. 4.8: rs = 0.415, N = 24, P < 0.05). Females always began gaping before males and the range in duration of female gapes was 6.0-12.9 s. This duration was not correlated significantly with female size (r = 0.134, N = 24, P = 0.533) but was greater in trials with low 3Y male frequency than in trials with high 3Y male frequency (Table B.8; Fig. 4.9). The range in times 47 between the start of the female gape to the start of her first dig was 10.3-19.0 s. This time was not correlated significantly with female size (r = 0.226, iV = 29, P — 0.238) and was not influenced significantly by OSR or by 3Y male frequency (Table B.9; Fig. 4.10). The first male to start gaping began 0.3-3.5 s after the female (median = 1.5 s, x = 1.5 s). The range in times between the start of the first male's gape and the start of subsequent males' gapes was 0-7.7 s (Fig. 4.11). Alpha males usually began gaping before other males (Fig. 4.12). The alpha male was the first male to begin gaping in 21 of 27 trials. In two of these 21 trials, another male started gaping at the same time as the alpha male. In five trials the alpha male was the second male to start gaping and in one trial the alpha male was the third male to start gaping. The mean gape start time of 5Y males (relative to the female start time) was less than that of 3Y males within the same trials (Fig. 4.12; paired i-test: mean difference = —1.4 s, t — —4.969, df = 24, P < 0.001). However, when alpha males were excluded from the analysis, the difference was not significant (Fig. 4.12; paired t-test: mean difference = —0.4 s, t = —0.544, df = 7, P = 0.604). Start time of male gaping was correlated significantly with spawning position (Fig. 4.13; rs = 0.434, N = 27, P < 0.05) such that males that spawned close to the female began to gape sooner than males that spawned farther from the female. The range in duration of male gapes was 0.9-11.0 s. Mean (within-trial) gape duration of 5Y males was greater than mean (within-trial) gape duration of 3Y males (Fig. 4.14; paired t-test: mean difference = 1.5 s, t = 4.972, df = 24, P < 0.001). However, if alpha males were excluded from this comparison, then 3Y and 5Y males did not differ significantly in gape duration (Fig. 4.14; paired t-test: mean difference = 0.1 s, t = 0.293, df = 7,P = 0.778). 48 The two male age classes did not differ significantly in spawning position. There were 25 trials in which at least one male in each age class participated in spawning and spawning positions were known. In 19 of these trials, at least one 3Y male spawned beside the female and in 18 trials at least one 5Y male spawned beside the female (x2 = 0.000, df = 1, P = 1.000). Mean spawning position of 3Y males within a trial did not differ significantly from mean spawning position of 5Y males within the same trial (WMST: V = 136, N = 25, P = 0.484). There were nine trials in which spawning positions were known and at least one 3Y satellite and at least one 5Y satellite participated in spawning along with the alpha male. In four of these trials the alpha male spawned beside the female, in one trial a 5Y satellite spawned beside the female, and in six trials a 3Y satellite spawned beside the female. While 3Y satellites tended to be more likely to spawn beside the female, the difference among the three male classes was not statistically significant (Fisher exact test: P = 0.076). Correlates of Male Fertilization Success Testes depletion was significantly greater in males that participated in spawning than in males that did not participate, and this effect tended to be stronger in 3Y males than in 5Y males (Table B.10; Fig. 4.16). To examine the effect of gape duration on testes depletion, I selected pairs of males within trials'and examined the difference in their gape duration versus the difference in their testes depletion. A n increase between males in gape duration was not correlated with an increase in testes depletion (for 17 pairs of males in 14 trials: rs = —0.549, Pi-taiied — 0.988). I performed a similar analysis to examine the relationship between fertilization success and testes depletion. An increase between males in testes depletion was not correlated significantly with an increase in fertilization success (for 21 pairs of males in 16 trials: rs = 0.269, Pl-tailed = 0-119). Sperm concentration (proportion of milt volume occupied by sperm cells) was 49 greater in 3Y male milt (x = 42.8%, SD = 7.5, N = 53) than in 5Y male milt (x = 38.3%, SD = 7.3, N = 57), but milt of males that participated in spawning did not differ in sperm concentration from milt of males that did not spawn (Table 4.4). For 23 male pairs in 17 trials, difference in sperm concentration was not significantly correlated with difference in fertilization success (rs = —0.185, P — 0.397). Alpha male fertilization success was greater than mean (within-trial) 5Y satellite male success (Fig. 4.17a; W M S T : N = 22, P = 0.002, aBonf = 0.017) but did not differ from mean (within-trial) 3Y satellite male success (Fig. 4.17a; W M S T : 7Y = 30, P = 0.080, ason} — 0.05). Mean 3Y satellite male success was significantly greater than mean 5Y satellite male success (Fig. 4.17a; W M S T : N = 24,P = 0.010, aBonf = 0.025). If only males that participated in spawning were included in this analysis alpha, 5Y satellite, and 3Y satellite males did not differ in fertilization success (Fig. 4.17b; alpha vs 5Y: N = 10, P = 0.044; alpha vs 3Y: N = 27, P = 0.106; 3Y vs 5Y: N = 9, P = 0.944; aBonf = 0.017). Alpha male fertilization success was not correlated significantly with frequency of courtship, aggression, or intrusion by satellite males (Table 4.5). There was a significant interaction between OSR and age group composition in their effects on alpha male success (Table B . l l ; Fig. 4.18). In three-male groups alpha males had greater success when there was one 5Y and two 3Y males present, but in six-male groups alpha males had greater success when 5Y males outnumbered 3Y males. However, the mean success of alpha males did not differ significantly between any pair of treatments (Tukey test: P 's> 0.2). Female size was not correlated significantly with alpha male fertilization success (r = 0.262, N = 30, P = 0.162). To examine the relationship between satellite male behaviour prior to spawning and spawning success, I calculated for each trial the mean number of intrusions by satellite males that participated in spawning and the mean number of intrusions by 50 those that did not spawn. Satellite males that spawned had significantly greater mean intrusion rates than those that did not spawn (WMST for first observations: V = 6, N = 10, P = 0.032; last observations: V = 0, iV = 18, P < 0.001). To examine the relationship between satellite male intrusion rate and fertilization success of males that participated in spawning, I paired satellite males within trials and tested for a correlation between the difference in their fertilization success and the difference in their intrusion rates. For trials in which three satellite males participated in spawning, I randomly selected two males of the same age. For trials in which four males participated in spawning I randomly selected two pairs of males. This resulted in 19 pairs of satellite males from 16 trials. Fourteen of these pairs were made up of 3Y males, two were made up of 5Y males, and three consisted of a 3Y and a 5Y male. The difference between fertilization success in these male pairs was not correlated with the difference in their intrusion rate either in the first observation (r s = —0.078, N = 11, P = 0.818) or the last observation (rs = 0.039, N = 19, P = 0.871). Fig. 4.19 shows "a path diagram to examine the effects of gape start time and spawning duration on male fertilization success. Success was not influenced directly by start time, but it increased with the duration of a male's gape which was influenced by starting time. However, most of the variation in male fertilization success arose from other sources (including random variation). There was also substantial variation in gape duration that could not be explained by starting time. Spawning position was not included in the path analysis. It could have a direct effect on fertilization success and could itself be influenced by gape start time. However, I measured spawning position as a rank which does not conform to the assumptions of this type of analysis. To examine the effect of spawning position, I compared the fertilization success of pairs of males within trials. I paired males so that the difference in spawning position was as large as possible and males were the same 51 age if possible. This included 25 pairs of males from 18 trials. Within these pairs, the male that spawned closer to the female did not have significantly higher fertilization success than the male that spawned farther from the female (paired t-test: mean difference in success= 9.1%, t = 1.055, P\-taUed — 0.151). Discussion The status and behaviour of male sockeye salmon in a spawning group are related to age and morphology, as in other salmonids (Maekawa & Onozato 1986; Fleming 1996; Healey & Prince 1998; Koseki & Maekawa 2000). In the present study 5Y males competed for dominance and alpha male status was usually attained by males with large body size and relatively large secondary sex traits. 5Y males that do not achieve alpha male status may participate in spawning as satellite males, but often do not participate (Chapter 3). Quinn & Foote (1994) also found an effect of body size and relative hump size on status of sockeye salmon spawning in an Alaskan lake. Large body size is an advantage in aggressive interactions which sometimes involve physical contact. Contests between competing males frequently include a series of displays in which humps and snouts figure prominently (Hanson & Smith 1967). Males with large humps and snouts may be perceived as superior competitors during these displays. In addition, females may show a preference for large body size and large secondary sex traits (Jarvi 1990). 3Y males usually behaved as satellite males. They exhibited little aggression but maintained positions nearer to the female than 5Y satellite males. Unlike other studies, I did not find a relationship between satellite male (within age-class) positions and body size (Maekawa 1983; Koseki & Maekawa 2000). In my study, male positions were measured only to the nearest 0.5 m, and this may not have afforded enough precision to detect such a relationship. However, 3Y males interacted very little in this study, and it is possible that males settled in the first unoccupied suitable positions that they found 52 without contesting positions of other 3Y males. Alpha male aggression followed an interesting pattern in this study. In most trials, alpha males exhibited a higher frequency of aggressive behaviour as spawning approached. However, in trials with four 3Y and two 5Y males, the frequency of alpha male aggression was high from the beginning (Fig. 4.2). These trials also had high rates of satellite intrusion during first observations (Fig. 4.4). When the number of males and the frequency of 3Y males are high, alpha males apparently have much less control over their social group and, as a result, have lower fertilization success than alpha males in other types of groups. This may explain why fertilization success of male age classes is not frequency-dependent in six-male groups (Chapter 3). The single most important factor influencing a male's fertilization success in a given spawning is whether he participates or not. Nevertheless, 5Y males in this study often did not participate and, when they did participate, their fertilization success was very low unless they were able to attain alpha male status. 5Y males that do not attain this status may be better off to abandon the current spawning opportunity and search for other nesting females. The temporal and spatial dynamics of spawning events have important implications for male fertilization success. When a female releases eggs into the water, males compete to fertilize them. The fertilization process in most fish is unlike that in more familiar taxa such as mammals and sea urchins. In fish, fertilization is achieved after a sperm cell enters a small opening in the egg membrane called the micropyle (Ginzburg 1972). There are few studies on the speed with which fertilization occurs (Chapter 6), but the sooner a male begins to release sperm, the sooner his sperm can locate and enter micropyles. As eggs continue to be released it may be advantageous for a male to maintain a steady flow of milt, effectively 'placing' sperm on each egg as it is released. It may also be advantageous for a male to release his sperm as close as 53 possible to the eggs so that his sperm can find and enter egg micropyles before sperm of his competitors. I did not find a direct relationship between the start time of a male's gape and fertilization success, but early start times allowed males to maintain sperm release for a longer time and this increased fertilization success. It is possible that starting time does influence male success directly but that the effect is too small to have been detected in this study. It is also possible that the onset of gaping is not a precise measure of the beginning of sperm release. It is surprising that I did not detect an effect of spawning position on male fertilization success, especially since spawning position is correlated with starting time and spawning duration and I did not control for these effects in the analysis. However, my measure of spawning position is a rank with limited variability. The relevant measure of spawning position is the location of the male's vent relative to that of the female and other males. This depends not only on male body position (as measured in this study) but also on the ability of the male to maneuver his vent close to that of the female. In this study the temporal and spatial variables (start time, duration, and position) that could influence fertilization success were correlated. Path analysis (Fig. 4.19) was useful in examining the effects of some of these variables but only provides correlational evidence (Grace Sz Pugesek 1998; Gregory Sz Skebo 1998). An experimental approach will be necessary to fully resolve the relationships between these variables and fertilization success. In addition, male success may be influenced by differences among males in factors such as rate of sperm release, sperm swimming speed, longevity, micropyle-finding ability, and compatibility with eggs of the given female (Chapters 5 and 6). There are at least three mechanisms through which females could influence fertilization success of competing males. First, females may attack males to exclude 54 them from spawning or to force them to increase their pre-spawning distance from the female. Females did show greater aggression to 3Y males than 5Y males, but this did not appear to be effective since 3Y males maintained positions closer to the females than 5Y satellite males and were more likely to participate in spawning than 5Y satellite males. Second, previous studies found that females delayed spawning when paired with small males (Foote 1989; van den Berghe et al. 1989; Blanchfield & Ridgway 1999; Berejikian et al. 2000; de Gaudemar et al. 2000), which may be an indirect form of mate choice (delaying spawning could give time for other males to displace the current one). However, in the present study females spawned sooner when 3Y males were common than when 3Y males were rare. Third, females could manipulate the number of eggs they lay. In the present study, females had shorter spawning durations when 3Y males were common than when 3Y males were rare, and this could mean that they laid fewer eggs. Females have been shown vary clutch size in response to male size in other species (Cote & Hunte 1989; Berejikian et al. 2000), but these studies involved females mating with single males. In the present study, it is also possible that a high frequency of 3Y males resulted in disruption of 'normal' spawning behaviour. For example, 3Y males that rush into the nest after spawning begins could jostle the female and cause her to stop spawning. Finally, females could bias paternity by directing eggs towards the vent of a preferred male. There is no evidence for this mechanism, but little is known about the spatial and temporal dynamics of gamete release. The experimental methodology of this study did not completely mimic natural spawning salmon since males in enclosures cannot leave the immediate area, other males cannot intrude, and individuals cannot interact with neighbouring groups. However, the behaviour of fish in these spawning trials was not qualitatively different from wild spawning fish (personal observation). The use of enclosures is well suited to address questions about factors influencing male status, which males participate in 55 spawning, and what influences fertilization success of males that do participate. Questions about larger scale processes such as lifetime reproductive success will require a different approach. 56 4.1: Behaviours recorded in spawning trial observations. Behaviour Description Dig turn on side and flip tail Aggression includes chase, threaten, ram, bite, and display Courtship includes quiver (male moves alongside the female and vibrates his body) and crossover (male rapidly moves over the female from one side to the other) Intrusion satellite enters nest within two body widths of the female 57 Table 4.2: Spearman rank correlations between behavioural variables in first and last observations. Sample sizes were 20 for first observations and 31 for last observations. None of the correlations was significant using asonf = 0.008. Observation Courtship Aggression First Aggression -0.243 Intrusions -0.235 0.457 Last Aggression 0.105 Intrusions 0.070 0.040 58 Table 4.3: A N O V A tables for the effects of morphological variables on satellite male distance from the female. Analyses were done as linear mixed-effect models (Pinheiro & Bates 2000) with trial number included as a random factor. The first model included 64 3Y males in 32 trials. The second model included 41 5Y males in 25 trials. See Table B.3 and text for interpretation of P C I and PC2. Source df F P 3Y males P C I PC2 1,30 1,30 0.029 0.427 0.867 0.519 5Y males P C I PC2 1,14 1,14 0.082 0.013 0.779 0.912 59 Table 4.4: A N O V A table for the effects of male age and spawning participation on spermatocrit. The analysis was done as a linear mixed-effect model (Pinheiro & Bates 2000) with trial number included as a random factor. The analysis included 94 males in 22 trials. The interaction between age and participation was not significant and was removed from the model ( F 1 ] 6 9 = 2.002, P = 0.162). Source df F P Male age 1,69 12.330 0.001 Participation 1,69 0.080 0.779 60 Table 4.5: Spearman rank correlations between the proportion of fry sired by alpha males and frequency of three behaviours in first (N — 16) and last (N = 24) observations. None of the correlations was statistically significant. First Last Courtship 0.418 0.193 Aggression -0.083 0.168 Intrusions -0.325 -0.169 61 o CD > 'to to CD 1_ O) O) CC CD JD E 3 15 J 10 J 5 -J 0 J i 1 r F Alpha 5Y First observation Last observation Figure 4.1: Boxplot of the number of aggressive acts per individual per 5min observation for females (F), alpha males (Alpha), 5Y satellite males (5Y), and 3Y satellite males (3Y). First observations (20 trials) were done at least 2 hr before last observations (31 trials). Data from all four treatments are included. 62 c o w tn co \ O) O) (0 CO E CO .n < 15 10 J 5 J 0 J 2/3 1/3 3Y male frequency Figure 4.2: Mean number (± SE) of aggressive acts initiated by alpha males in first (a) and last observations (b) of trials composed of 1/3 or 2/3 3Y males. Open symbols indicate trials with three males and solid symbols indicate trials with six males. Sample sizes are in brackets. 63 Q. CO 20 15 J c 10 O O 5 J 2/3 1/3 3Y male frequency Figure 4.3: Mean number (± SE back-transformed from square-root transformations) of alpha male courtship behaviours in first (a) and last observations (b) of trials composed of 1/3 or 2/3 3Y males. Open symbols indicate trials with three males and solid symbols indicate trials with six males. Sample sizes are in brackets. 64 15 J 10 « 5 J 0 J 2/3 1/3 3Y male frequency Figure 4.4: Mean number (± SE back-transformed from square-root transformations) of satellite intrusions into the nest in first (a) and last observations (b) of trials composed of 1/3 or 2/3 3Y males. Open symbols indicate trials with three males and solid symbols indicate trials with six males. Sample sizes are in brackets. 65 CO O CO e 30 . 20 10 J CD E 1 5 3 10 T - 2 - 1 0 1 Change in distance from female Figure 4.5: Frequency distribution of the change in male distance from the female as spawning approached for 3Y satellites (a) and 5Y satellites (b). Positive values indicate that males moved closer to the the female as spawning approached and negative values indicate that males moved away from the female as spawning approached. 66 1/3 2/3 3Y male frequency Figure 4.6: Mean time to spawn (±SE) for trials composed of 1/3 or 2/3 3Y males. Open symbols indicate trials with three males and solid symbols indicate trials with six males. Sample sizes are in brackets. 67 4.0 3.5 3.0 J c CO k_ •c S 2.5 CO d) Q . CO C5 2.0 1.5 1.0 -J Distance from female (m) Figure 4.7: Relationship between rank order in which males began to gape and their distance from the female shortly before spawning. Symbol size is proportional to the number of data points. Includes 75 males in 26 trials. 68 CD E •c CO co CD Q . CO O 10 J 8 J 6 J 2 J 0.0 1.0 2.0 Distance from female (m) 3.0 Figure 4.8: Relationship between time that males began to gape and their distance from the female in the last observation before spawning. Symbol size is proportional to the number of data points. Includes 68 males in 24 trials. 69 o 3 Q cu CL CO (5 0) CO E CD 13 J 12 J S 11 A 10 J 8 -J 6 J 1/3 2/3 3Y Male Frequency Figure 4.9: Duration of female gape in trials where 3Y males were at low and high frequency. Open circles represent trials with three males and closed circles represent trials with six males. 70 20 -O 18 - O 16 _ i O • t • • o o • 14 _ o • 1 12 _ 0 o • o 10 - o • 1/3 2/3 3Y Male Frequency Figure 4.10: Time between start of female gape and start of dig in trials where 3Y males were at low and high frequency. Open circles represent trials with three males and closed circles represent trials with six males. 71 8 J 6 J Q CO E V/J I £ 4 CD % CD E 2 J 0 J Male number Figure 4.11: Time that males began gaping relative to the first male. Sample sizes are given in brackets. 72 6 J CD E •4—' •c CO CD Q. CO o 0 J Alpha 5Y Satellite 3Y Satellite Figure 4.12: Comparison of male classes in their gape start times. Time was measured relative to the beginning of the female's gape. Sample sizes (number of trials) are given in brackets. For trials with more than one 3Y or 5Y satellite, means were used. 73 c o O CL 3 J o o o o o o CO CL CO 2 J o o o a n D oQO 0 0 ° o o o o o o o 1 — | O O rt CDD O O OO O • " O O i 1 r 0 2 4 8 Gape start time (s) Figure 4.13: Male spawning position versus the time that gaping was started relative to the first male within a trial. Symbol size is proportional to the number of data points. Includes 84 males in 27 trials. 74 c o 3 T3 CD Q . CO O 12 10 J 8 J 4 J Alpha 5Y Satellite 3Y Satellite Figure 4.14: Gape duration of males. Sample sizes (number of trials) are given brackets. For trials with more than one 3Y or 5Y satellite, means were used. 75 0.8 . 0.6 . 0.2 0.0 0.8 0.6 . 0.4 J Q) CO E O c o o 0 2 -I CL o 0.8 0.6 0.4 0.2 0.0 1.5 2.5 Spawning position 3.5 Figure 4.15: Frequency distribution of spawning positions of (a) alpha males (N (b) 5Y satellite males (N — 11), and (c) 3Y satellite males (N = 47). 76 0.00 J a> 'CD CO & to CD -0.05 r -o . ioJ CO -g 'co CD DC -0.15 J -0.20 J 3Y males 5Y males Figure 4.16: Mean testes depletion (± SE) for 3Y and 5Y males. Open circles represent males that did not participate in spawning and closed circles represent males that did participate in spawning. Sample sizes are given in brackets. 77 CD i _ co & • "o c o •c o Q . o 1 -\ 0.75 J 0.5 ol 0.25 J Alpha 5Y sat. 3Y sat. Alpha 5Y sat. 3Y sat. Figure 4.17: Boxplots of fertilization success per male (a) and per male that participated in spawning (b) for alpha and satellite males. Number of trials is shown in brackets. 78 TJ CD i _ 'co o c o t o CL O 1.0 J 0.8 J 0.6 J 0.4 0.2 J 0.0 J 1/3 2/3 3Y Male Frequency Figure 4.18: Alpha male fertilization success in trials where 3Y males were at low and high frequency. Open circles represent trials with three males and closed circles represent trials with six males. 79 -0.678 Unknown 0.470 t Spawning Duration -0.124 0.534 Fertilization Success A 0.798 Unknown Figure 4.19: Path diagram to examine the influence of male timing in spawning events on fertilization success. Coefficients are standardized multiple regression coefficients from an analysis with fertilization success (arcsine square-root transformed) as the dependent variable and start time, spawning duration, and trial number included as independent variables. Significant path coefficients are indicated by asterisks. 80 Chapter 5 The Outcome of Sperm Competition Between Alternative Male Phenotypes Introduction There is evidence from many taxa that sperm competition imposes selection for increased numbers of sperm (Birkhead & M0ller 1998). In species where females commonly mate with more than one male, males have larger testes and larger ejaculates than males of species in which females mate only once. A similar form of selection may act within species. For example, males that use mating behaviour such as sneaking have a high probability of undergoing sperm competition because they always sneak in the presence of a male that uses behaviour such as courtship. Courting males, on the other hand, may have a lower probability of undergoing sperm competition because, if sneaker males are rare, then some courting males will mate without competition from sneakers (Parker 1990b). A relationship between male phenotype and sperm production has been established. In many fish species small sneaker males have relatively larger testes and relatively greater numbers of strippable sperm (a measure of potential ejaculate size) than large courting males (Taborsky 1994; Ruchon et al. 1995; Gage et al. 1995; Foote et al. 1997; Scaggiante et al. 1999). Similarly, sneaking males in the dung beetle Onthophagus binodis have relatively larger testes and ejaculate volume than guarding males (Simmons et al. 1999). The selective effect of sperm competition is usually considered in terms of the number of sperm produced. However, this effect can be expressed more generally as an increased investment in sperm. Investment may be expressed in greater numbers of sperm but it may also involve production of higher 'quality' sperm (e.g., size, longevity, 81 speed, etc.). Differences in sperm traits between male phenotypes have been found in Atlantic salmon (Salmo salar), in which sperm from parr (small, early-maturing sneaker males) have greater motility than sperm from anadromous (large, late-maturing courting males) males, although sperm length does not differ between the two phenotypes (Gage et al. 1995). In the dung beetle Onthophagus binodis, sneaking males have longer sperm than guarding males (Simmons et al. 1999). One of the problems with examining sperm characteristics is that, for most species, little is known about the relationship between sperm traits and the ability of sperm cells to compete for fertilizations. A n alternative method for comparing sperm is to examine directly the fertilization success of sperm from different males in a competitive environment. For example, females can be inseminated with a mixture of semen from two males and paternity of the resulting offspring determined (e.g., Dziuk 1996). One drawback of this approach is that it is difficult to distinguish the effect of sperm quality from that of female choice (Birkhead 1998). This problem may be reduced in animals with external fertilization since sperm do not have to negotiate a female reproductive tract which may impose selection (but see Palumbi (1999)). Salmonid fishes are ideal for studying competition among sperm from different males because a) they undergo external fertilization which makes in vitro experiments easy to perform, and b) most species have alternative male phenotypes. Male phenotypes have not been compared for competitive sperm quality but, in semi-natural matings, Foote et al. (1997) found that sneaker and courting phenotypes of sockeye salmon (Oncorhynchus nerka) did not differ in fertilization success even though the sneaker males have fewer sperm available. While factors such as timing and position of male spawning may have played a role in male success, this result is consistent with the hypothesis that sneaker males have sperm that are competitively superior to sperm from courting males. 82 There have been no experimental studies that compare alternative male phenotypes in the ability of their sperm to compete for fertilizations. The purpose of this study was to test for differences in competitive ability of sperm from alternative male phenotypes (two male age classes) of sockeye salmon using in vitro fertilizations. As I did not detect a difference in sperm competition success between male phenotypes, I also tested for morphological and genetic correlates of success in sperm competition. Methods This study was conducted during the spawning season of 1995. In this experiment, 3Y males ranged in hypural length from 280 mm to 305 mm, 5Y males ranged from 485 mm to 530 mm and females ranged from 410 mm to 475 mm. Methods for selecting fish and collecting gametes are described in Chapter 2. The basic protocol for this experiment was to mix milt from two males, fertilize a batch of eggs with this mixture, and determine paternity of the fertilized eggs. I used spermatocrit as a measure of sperm abundance in the milt (Chapter 2) so that I could add an equal number of sperm from each male. Each replicate in this experiment consisted of three in vitro fertilizations. I fertilized one batch of approximately 150 eggs (range: 105 to 199) with milt from a 3Y male. A second batch of approximately 150 eggs (range: 115 to 186) was fertilized with milt from a 5Y male. I alternated the order of these fertilizations between replicates. The volume of milt used was that which yielded 0.5 ml of packed sperm cells (as calculated from spermatocrit). These single-male fertilizations allowed me to assess the combined effects of fertilization success and survival of fertilized eggs for each male. I fertilized a third batch of approximately 300 eggs (range: 176 to 377) with a mixture of milt from the two males. This mixture consisted of 0.5 ml sperm from each male and was stirred thoroughly before being added to eggs. I performed a total of 10 replicates 83 using different individuals in each. I completed all fertilizations in a replicate within 45 min of starting to collect gametes. They were performed by simultaneously adding milt and river water to the eggs. The volume of water used was 250 ml for the single male fertilizations and 500 ml for the competitive fertilization. The gametes were swirled for several seconds and allowed to stand for three minutes. The eggs were then rinsed thoroughly and placed in perforated plastic tubes. The tubes were buried in gravel in a spawning channel and left to incubate. After the fertilizations, parental fish were given a lethal dose of anaesthetic and body measurements taken. Size-adjusted morphological measurements were calculated as the residual of the given measurement from a reduced major axis regression of logio (trait size) versus logi 0(body size) for a sample of 36 males. A sample of liver tissue was stored in 95% ethanol for D N A analysis. Approximately two months after the fertilizations, I removed all tubes from the spawning channel and placed them in a laboratory incubator. At this time, undeveloped and damaged eggs were removed and counted. I checked the surviving eggs every two to three days for further mortality and allowed them to hatch. When the offspring reached the free-swimming fry stage I gave them a lethal dose of anaesthetic and preserved them in 95% ethanol. I determined paternity of competitive fertilizations by examining variation in microsatellite D N A for 30 randomly selected fry in each replicate. Parents were typed at seven microsatellite loci. The method of D N A analysis and the loci used are described in Chapter 2. In eight replicates there was sufficient variability at single loci to establish paternity unambiguously. In one replicate a second locus was necessary and in another replicate three loci were required to unambiguously determine paternity for all fry. 84 Data Analysis To test for differences between 3Y males and 5Y males in competitive fertilization success, I used a replicated goodness-of-fit test (Sokal Sz Rohlf 1981). In this test the null hypothesis is that 3Y and 5Y males do not differ in competitive fertilization success so that each age class fertilizes 50% of the eggs on average. If this hypothesis is false, the probability of rejecting it (power) depends on the true mean proportions fertilized by the age classes. For example, if one age class fertilizes 90% of the eggs on average while the other age class sires 10%, then the power of the test is likely to be much higher than if one age class sires 55% on average and the other age class sires 45%. To examine the power of this test I performed Monte Carlo simulations of my experiment. This involved setting the mean proportion fertilized by one of the age classes to a particular value and generating random fertilization data for 10 replicates of 30 eggs using the binomial distribution. I repeated this 100,000 times, performing the replicated goodness-of-fit test each time. The proportion of these simulations that resulted in rejection of the null hypothesis represents the power of the test accurate to two decimal places (Thomas Sz Juanes 1996). I performed these simulations using values for the mean proportion fertilized between 0.5 and 0.7 in increments of 0.01. When the mean proportion sired was set to values greater than 0.62, some of the simulations produced results in which one male fertilized all of the eggs in a replicate. When this occurred, the test statistic could not be calculated (it is based on the log of the number of eggs fertilized), so I changed zero values to 0.00001. The results of these simulations are shown in Fig. 5.1. The power of the replicated goodness-of-fit test to reject the null hypothesis was greater than 0.8 when the mean proportion fertilized by one age class was 0.58 or higher. Microsatellite data were used to test for genetic effects on male success. The genetic measures I used were male relatedness to the female (Queller &; Goodnight 85 1989), male heterozygosity, and mean d2 (a measure of individual genetic variability: see Coulson et al. 1998). I predicted that the difference in paternity between two males would be negatively correlated with the difference in their relatedness with the female (i.e., males closely related to the female are less successful than distantly related males) and positively correlated with the differences in their genetic variability (i.e., males with high genetic variability are more successful than males with low variability). Results An average of 84.3% of eggs were fertilized and survived to hatch (median: 88.5; range: 51.8 - 98.9) in the single-male and competitive fertilizations combined. For nine replicates I was able to measure survival of eggs in single-male fertilizations (fry for one of the single-male fertilizations in the tenth replicate were not recovered). In single-male fertilizations, survival of clutches fertilized by 3Y males did not differ from survival of clutches sired by 5Y males (WMST: P = 0.570). There was a significant correlation between fry survival in 3Y male fertilizations and fry survival in 5Y male fertilizations (rs = 0.533, Pi-taiied = 0.016). Difference in fry survival between single-male fertilizations within replicates was not correlated significantly with difference in paternity in competitive fertilizations (rs = 0.368, Pi-taiied = 0.149). In most replicates, paternity of competitive fertilizations was biased toward one male (Table 5.1). A replicated goodness-of-fit test revealed that 3Y and 5Y males did not differ in competitive fertilization success (pooled G = 1.334, df = 1, P = 0.248). However, there was significant deviation from the distribution expected if males had an equal probability of fertilizing eggs (total G — 32.116, df — 10, P = 0.0001), and there was significant heterogeneity among replicates (heterogeneity G = 30.782, df = 9, P = 0.0001). Order of male capture did not influence paternity (replicated goodness-of-fit test: pooled G = 2.617, df = 1, P = 0.106). 86 Success in sperm competition was not related to male morphology. Difference in paternity between males was not correlated with difference in residual body weight (rs = 0.059, Pi-taiied = 0.434), residual hump size (r s = 0.195, Pi-taiied = 0.229), or residual snout length (r s = 0.189, Pi-taiied = 0.236). I did not detect any effects of male genotype on success in sperm competition. Difference in paternity was not significantly correlated with difference in male heterozygosity (r s = 0.182, P\-taiied = 0.308) or difference in mean d2 (rs = —0.500, Pi-taiied = 0.930). I also failed to detect a negative correlation between difference in paternity and difference in relatedness to females (rs = 0.433, Pi-taiied = 0.895). The relationships of mean d? and relatedness with success in sperm competition, while not significant, were in the opposite directions to those predicted (fig. 5.2, fig. 5.3). Discussion Sperm from alternative phenotypes of sockeye salmon did not differ in competitive success in this study. This result is consistent with observations that sperm from 3Y and 5Y males are equally capable of fertilizing eggs in non-competitive fertilizations (this chapter; Chapter 6). Thus, in this species, there is no evidence that differences in sperm competition risk (as defined by Parker et al. (1996)) between phenotypes has produced differences in competitive ability of sperm. In this study I did not examine eggs immediately after insemination and so my measure of fertilization success includes the combined effects of fertilization success and survival of embryos. However, embryo survival was not likely to have had a strong effect on my measure of male fertilization success. In addition to the fact that differences in paternity were not correlated with differences in survival, survival rates were high in most replicates. For example, replicates 2 and 3 had large differences in paternity (Table 5.1), but survival of both clutches was greater than 98%. Thus the 87 differences in paternity in these replicates could not have resulted from differences in embryo survival. The volume of milt used in these fertilizations was sufficient to ensure 100% fertilization (Chapter 6). The results of this study leave unresolved the mechanism through which 3Y males obtain high success in semi-natural competitive matings (Foote et al. 1997). One possible solution is that the assumption that ejaculate size differs between male age classes is incorrect. Ejaculate size has not been measured in this species, and it is possible that measures of the number of sperm available for ejaculation are not correlated with ejaculate size. A n alternative explanation of Foote et al.'s results is that 3Y males may be able to use their small size to their advantage during spawning and obtain closer access to eggs being released by the female. A factor that may have influenced both the results of the present study and those of Foote et al. (1997) is the social environment of the males. Male fish that are exposed to ovulated females undergo behavioural and physiological changes (referred to as the priming response) which include an increase in the volume of milt available for ejaculation (Liley et al. 1993; Liley & Kroon 1995), an increase in sperm motility (Miura et al. 1992; DeFraipont & Sorensen 1993), and, in the case of goldfish (Carassius auratus), an increase in competitive ability of sperm (Zheng et al. 1997). When two males compete for access to a reproductive female, the priming response is stronger in the dominant male (Liley et al. 1993; Liley &; Kroon 1995), but priming responses in alternative male phenotypes have not been examined. In the present study males did not undergo priming, and it remains to be seen if the priming response differs among male phenotypes and if success in sperm competition is influenced by priming. The possibility remains that female choice influences the outcome of sperm competition in this species. In sea urchins, eggs discriminate among sperm on the basis of sperm genotype at a surface protein locus (Palumbi 1999). Similarly, when female 88 decorated field crickets (Gryllodes supplicans) are mated both to a sibling and an unrelated male, sperm from the unrelated male fertilize more eggs than do sperm from the sibling male (Stockley 1999). However, the process of fertilization in salmon is very different from that in sea urchins and insects. In the latter taxa, sperm cells can attach anywhere on the egg and undergo an acrosome reaction. In teleost fish, fertilization is achieved when a sperm cell swims into a small opening (micropyle) in the egg membrane and there is no acrosome reaction. Observations of this process indicate that the first sperm to enter the micropyle fertilizes the egg (Yanagimachi et al. 1992). Thus, if eggs discriminate among sperm, the mechanism would likely involve an influence of eggs either on the ability of sperm to locate the micropyle or on the swimming speed of sperm. Two studies in other fish species found that a male's success in competitive in vitro fertilizations does vary across females (Gharrett & Shirley 1985; Rakitin et al. 1999), but males in a third study had consistent success across females (Gile & Ferguson 1995). A mechanism for female choice of sperm has not been demonstrated in teleost fish but ovarian fluid is known to influence sperm motility (Litvak & Trippel 1998), and if this effect varies among males then females could influence male fertilization success. In rainbow trout (0. mykiss) male fertility in single-male fertilizations is correlated with the proportion of sperm possessing a specific antigen (Trummel et al. 1992), but neither the mechanism of this effect nor the effect of the antigen on competitive fertility are known. I did not detect any phenotypic correlates of success in sperm competition, but I did find substantial differences between males. It should be noted that without replicated tests of individual males I cannot conclude that males differ in competitive success but rather that the particular milt samples that I collected showed significant variation in success. Nevertheless, the outcome of sperm competition evidently is not a 'fair raffle' in which success is determined by the number of sperm (Parker 1990a). 89 More research is needed to find determinants of success in sperm competition, both in vitro and in natural matings. Sneaker males in sockeye salmon clearly achieve high reproductive success through some mechanism other than high sperm quality. Little is known about the dynamics of gamete release (e.g., relative positions of milt and egg release, timing of gamete release, rate of gamete release and total ejaculate volume) in salmonid fishes and how these dynamics influence male reproductive success. In addition, the relationships I observed between genetic variables and success in sperm competition are intriguing and worthy of further research. 90 Table 5.1: Percentage of offspring sired by 3Y and 5Y males in competitive fertilizations (N = 30 offspring for each replicate). Replicate 3Y male 5Y male Binomial P 1 53.3 46.7 0.135 2 70.0 30.0 0.013 3 70.0 30.0 0.013 4 43.3 56.7 0.112 5 36.7 63.3 0.051 6 63.3 36.7 0.051 7 63.3 36.7 0.051 8 20.0 80.0 0.001 9 66.7 33.3 0.028 10 46.7 53.3 0.135 total 53.3 46.7 91 CD o Q. 1.0 0.8 J 0.6 0.4 J 0.2 J Mean proportion fertilized Figure 5.1: Power of the replicated goodness-of-fit test to reject the null hypothesis of no difference between age classes in fertilization success given that one age class actually fertilizes the proportion indicated (on average) on the x-axis. 92 CD O c CD CD T J 2" c 1— CD •S Q. 0.6 J 0.4 0.2 J 0.0 J -0.2 Mean d difference Figure 5.2: Scatterplot of the difference between competing males in proportion of fry sired versus the difference in their genetic variability (as measured by mean d2). 93 - 0 . 4 -0.2 0.0 0.2 Kinship difference Figure 5.3: Scatterplot of the difference between competing males in proportion of fry sired versus the difference in their relatedness with the female (as measured by the kinship coefficient). 94 Chapter 6 Fertilization dynamics and a comparison of sperm from alternative male phenotypes Introduction Gamete biology exhibits considerable variation in fish and the processes leading up to fertilization are often very different from those in well studied taxa such as sea urchins and mammals (Gilkey 1981). For example, in Salmonids fertilization is external, eggs are very large relative to sperm, and eggs are fertilized after a sperm cell enters a single opening (the micropyle) in the egg membrane which is barely wide enough to allow entry of one sperm cell (Ginsburg 1963; Kobayashi & Yamamoto 1981; Yanagimachi et al. 1992). Thus, in contrast to sea urchins and mammals in which thousands of sperm may associate with an egg before one of them achieves fertilization, in Salmonids the first sperm cell to enter the micropyle fertilizes the egg (Kobayashi & Yamamoto 1981; Yanagimachi et al. 1992). Another notable characteristic of Salmonid reproduction is the limited lifespan of the gametes. Sperm are motionless until they come into contact with water and only remain motile for c. 30 s (Billard et al. 1986; Billard 1992), a fact that prompted Huxley (1930) to conclude that trout (Salmo trutta) sperm are maladapted for freshwater life. When eggs are exposed to water, they remain fertile for c. 40 s before osmotic swelling blocks the micropyle and prevents fertilization from occurring (Ginsburg 1963; Billard 1992). The short time span of gamete viability has implications for both male and female reproductive success. For both sexes, synchronizing gamete release may be crucial for maximizing the proportion of eggs that is fertilized. Males face an additional challenge 95 in that several competitors may attempt to fertilize eggs at the same time. Therefore, male success may depend on the timing and position of sperm release relative to those of other males, ability of the sperm to compete with those from other males, and on the number of sperm released. Thus, both male and female reproductive success are closely bound to fertilization dynamics. The influence of fertilization dynamics on male reproductive success is of particular interest in species in which males exhibit alternative reproductive tactics (Taborsky 1998). In these species, it has been hypothesized that sperm characteristics may vary among male phenotypes because of variation in selective pressures (Gage et al. 1995). Males that use parasitic reproductive tactics (e.g., sneaking) are expected to invest more in sperm than males that use non-parasitic tactics (e.g., courtship) because parasitic males are more likely to experience direct competition than are non-parasitic males (i.e. parasitic males have a greater risk of sperm competition) and the average number of competitors is likely to be higher for parasitic males than for non-parasitic males (i.e. parasitic males experience higher intensity of sperm competition). This difference in selective pressure can result in differences in the number and quality of sperm produced. For example, mature Atlantic salmon (Salmo salar) parr have relatively larger testes and relatively greater strippable sperm numbers than do anadromous males (Gage et al. 1995). Atlantic salmon parr have also been found to have greater sperm motility than anadromous males (Daye & Glebe 1984; Gage et al. 1995). In this study I examined some aspects of sockeye salmon (Oncorhynchus nerka) fertilization dynamics in vitro. First, I compared sperm characteristics and fertilization performance over a range of sperm concentrations in alternative male phenotypes. Gamete longevity was documented also. This was achieved by exposing eggs or sperm to fresh water for varying lengths of time before mixing them with gametes from the opposite sex. Finally, I examined the relationship between duration of gamete 96 association and fertilization success to gain understanding about how quickly fertilization occurs. Methods Fish Selection and Gamete Collection A l l fish selected for this experiment had free-flowing gametes (easily collected with gentle pressure to the abdomen). Five-year-old males (hereafter referred to as 5Y males) were used in all experiments and ranged in hypural length from 470 mm to 585 mm. For some experiments smaller (younger) males were also used (hereafter referred to as 3Y males). These males ranged in hypural length from 270 mm to 435 mm and were mostly 3-year-olds but may have included some 4-year-olds. Females were either 4-or 5-years-old (hypural length ranged from 425 mm to 535 mm). The methods for collecting gametes are described in Chapter 2. The time of gamete collection was noted for each fish and gametes were stored on ice until experiments were performed. Sperm motility was measured in two of the experiments, immediately after milt collection. A microhematocrit capillary tube was used to place a small drop (c. 1 (A) of milt in a 1.5 ml microcentrifuge tube. Then a syringe was used to add 1 ml river water to the drop of milt to initiate motility. The water was added forcefully to ensure thorough mixing of milt and water. A timer was started at the moment that water was added. 50 /A of the milt-water mixture was placed immediately on a depression slide resting on the stage of a pre-focused microscope (100x magnification). Ten seconds after sperm activation, the proportion of motile sperm was assessed using a 6-point scale similar to DeFraipont & Sorensen (1993): no motility, very low (1-10%), moderately low (10-30%), medium (30-75%), moderately high (75-95%), and very high (95-100%). When the proportion of sperm exhibiting motility declined below 5%, the 97 time was noted as a measure of the duration of motility. Both measures of sperm motility were performed two to three times for each male. General Protocol Each replicate in these experiments consisted of a number of fertilization treatments (for example, using different volumes of milt) and, in some cases, milt from two different males (one 3Y and one 5Y). To avoid potential deterioration of gametes over the time it took to complete a replicate, fertilizations were performed as quickly as possible. The order of fertilization treatments was randomized within replicates but was the same for both males. This ensured that any effects of the order of treatments would be the same for both males. The order of males for two-male experiments was the same as their order of capture. No individuals were used in more than one replicate. The procedures for fertilizing eggs varied between experiments and are described for each experiment below. In replicates for experiments 1, 2 and 4 the time between collection of the first fish and completion of the last fertilization ranged from 30 to 64 min. In experiment 3, milt samples were pooled and used for five replicates. The time between collection of the first male and completion of the last fertilization ranged from 55 to 174 min for this experiment. After fertilization, eggs were rinsed thoroughly and placed in Heath tray incubators. Incubating eggs were checked approximately every second day and eggs that had turned cloudy were removed. After one to three weeks, eggs were removed from the incubators and soaked in 5% acetic acid for 15 min. This turned embryos white, making them clearly visible to the naked eye. Eggs without visible embryos were assumed to be unfertilized. The total number of fertilized and unfertilized embryos was counted for each batch of eggs. 98 Experiment 1: Fertilization Rate v. Male Age and Sperm Concentration Around the Eggs The goals of this experiment were to determine the minimum concentration of sperm required for subsequent experiments and to compare fertilization success of sperm from 3Y and 5Y males over a range of sperm concentrations. Each replicate involved fertilizing batches of eggs from one female at four levels of sperm concentration from each of two males. A total of 11 replicates was performed. One 3Y and one 5Y male were captured (alternating the order of capture of the two age classes between replicates), measured, and stripped for milt. For each male the volume of milt required to yield 5, 15, 50, and 100 p\ of packed sperm was calculated using the measurement of spermatocrit. A female was then captured and eight egg batches of 20 ml each (c. 150 eggs per batch) were measured. Fertilizations were performed one at a time, as quickly as possible. For each fertilization, eggs were kept to one side of the beaker and milt was pipetted onto a dry spot on the other side of the beaker. 500 ml of river water was added quickly and the beaker was swirled for 5 s then set aside for 3 min. Thorough mixing of gametes occurred very rapidly with this method, regardless of the milt volume used. The eggs were rinsed and placed in an incubator. Experiment 2: Changes in Fertilization Rate With Time After Sperm Activation In this experiment sperm longevity was examined, specifically, the relationship between fertilization success and time since activation of sperm for both 3Y and 5Y males. The relationship between sperm motility and fertilization success was examined also. The basic protocol was to activate sperm by adding water to milt samples and then to add this milt-water mixture to eggs after varying time periods (0, 5, 10, 20, 40, and 80 s). Sperm motility was assessed also for all males in this experiment. Nine replicates were performed, each of which involved using milt from one 5Y male and one 99 3Y male to fertilize batches of eggs from one female. The volume of milt used in each fertilization was that which yielded 30 p\ packed sperm. This volume was chosen because it corresponded approximately to the minimum sperm concentration required to fertilize 100% of the eggs (as determined in the first experiment). In treatments with a delay between sperm activation and exposure to eggs, milt was pipetted onto the bottom of a beaker, and 20 ml of eggs was measured into a second beaker. Then 500 ml of river water was poured into the sperm beaker. When the water hit the bottom of the beaker, a timer was started. After the appropriate time period had passed, the water-milt mixture was poured quickly into the beaker containing eggs. For the treatment with no delay between sperm activation and exposure to eggs, eggs were placed on the bottom and to one side of a beaker, milt was placed on the other side, and 500 ml of river water was added. A l l egg-milt mixtures were allowed to stand for three minutes before rinsing the eggs and placing them in an incubator. Experiment 3: Changes in Fertilization Rate With Time After Egg Activation The purpose of this experiment was to examine the relationship between fertilization rate and time since immersion of eggs in water. The basic protocol was to add water to eggs and then add sperm after varying time periods. Ten replicates were performed with treatments of 0, 5, 10, 20, 40, 80, and 160 s. The first five replicates were performed using a mixture of sperm from six 5Y males and the next five replicates using a mixture of milt from six different 5Y males. A different female was used in each replicate. To perform a fertilization, 20 ml eggs was measured into a clean dry beaker. Then 250 ml river water was added and a timer was started when the water made contact with the eggs. After the appropriate time period, a volume of milt containing 30 pl packed sperm and 250 ml river water was added simultaneously to the egg-water 100 mixture. The purpose of adding the second batch of water was to ensure mixture of eggs, sperm, and water. The egg-sperm mixture was allowed to stand for 3 min before rinsing the eggs and placing them in an incubator. Experiment 4-' Fertilization Rate v. Time of Egg-Sperm Association The purpose of this experiment was to determine how quickly eggs were fertilized once they were exposed to sperm. However, in practice it is difficult to determine the exact moment when fertilization occurs. Instead, I measured the time it took for sperm to become so closely associated with eggs so that fertilization was inevitable. The basic protocol was to add water to eggs and sperm and rinse the eggs after varying periods. Sperm motility was assessed for all males in this experiment. Nine replicates were performed using milt collected from 5Y males only. The time periods for which I exposed eggs to sperm were <1, 1, 2, 5, 10, and 180 s. In the treatment of <1 s, eggs were dipped in the milt-water mixture and removed as quickly as possible. This experiment was carried out by two people so that timing could be controlled precisely. 20 ml of eggs was measured into a basket made of 3 mm plastic netting. The eggs rested in a single layer on the bottom of the basket so that rinsing would affect all eggs equally. 30 /A of sperm was measured onto the bottom of a beaker and 500 ml river water was added quickly. The basket of eggs was then dipped into the sperm-water mixture. The time delay between adding water to sperm and dipping eggs was 1-3 s. After the appropriate period of time the egg basket was pulled out of the sperm-water mixture and immediately plunged into flowing river water to rinse off sperm. A metronome was used to aid in controlling egg-dipping time. Six replicates of this experiment were recorded on videotape to determine the actual timing to the nearest 1/60 s. The actual dipping times were very consistent and close to the attempted times (Table 6.1). There was also a brief time lag between pulling eggs out of the sperm-water mixture and plunging them into flowing water to be 101 rinsed. This time ranged from 0.50 to 1.17 s (x = 0.61, SD = 0.12). Results Milt & Sperm Characteristics Milt of 3Y males contained more sperm cells per unit volume than did milt of 5Y males. For the first two experiments combined, spermatocrit of 3Y males ranged from 27.75% to 49.0% (x = 39.1%, median = 39.5%, N = 20) and spermatocrit of 5Y males ranged from 22% to 47.5% (x = 32.5%, median = 31.0%, N = 20). The difference between 3Y and 5Y males (paired within replicates to minimize variation from sources such as time, date and year) in spermatocrit was significant (WMST: Pi-taiied = 0.002). Sperm motility was compared between 3Y and 5Y males in the second experiment. The ranked proportion of initially motile sperm ranged from two to five for both 3Y males (x = 3.8, median = 3.5) and 5Y males (x = 3.4, median = 4.0). The difference between age classes was not significant (WMST: Pi-taiied = 0.311, N = 9). Duration of motility ranged from 22 to 27 s for 3Y males (x = 24.2 s, median = 24.0 s) and from 21.5 to 24.7 for 5Y males (x = 23.1, median = 24.5 s). Duration of motility was significantly greater in 3Y males than in 5Y males (WMST: Pi-taiied = 0.035, N = 9). The two measures of sperm motility were not correlated significantly (data pooled for all males in the second and fourth experiments: rs = 0.204, P = 0.298, N = 27). Experiment 1: Fertilization Rate v. Sperm Concentration & Male Age The range of sperm volumes used in this experiment resulted in obvious differences in opacity of the water. When 5 [A packed sperm were added to 500 ml water there was no noticeable cloudiness. At the other extreme, the addition of 100 fA of packed sperm resulted in very cloudy water. Even the lowest sperm concentration sometimes yielded high fertilization rates. 102 When 5 iA packed sperm volume per 500 ml water was used to fertilize eggs, fertilization rates were highly variable, ranging from 8.3% to 97.5% (Fig. 6.1). Sperm concentration had a strong effect on fertilization rate (Fig. 6.1, Table 6.2), and the concentration required to achieve fertilization rates consistently >80% was between 15 fA and 50 /A packed sperm per 500 ml water (Fig. 6.1). Male age did not affect fertilization rate, but there was significant variation among females (Table 6.2; Fig. 6.2), indicating that the quality of eggs varied among females. This conclusion is supported by strong correlations between fertilization rates of pairs of males tested with the same female (Table 6.3). This correlation existed at all but the highest sperm concentration, where there was not enough variation in fertilization rate for a correlation to be detected. Experiment 2: Changes in Fertilization Rate With Time After Sperm Activation In one replicate of this experiment eggs were damaged, and fertilization success was very low. This replicate was not included in the analyses. Sperm maintained a high capacity for fertilization for 10 s after exposure to water but fertilization rate declined sharply after that (Fig. 6.3). 3Y and 5Y males did not differ in fertilization success at any time, but there were significant differences among females, again suggesting differences among females in egg quality (Table 6.4). Correlations between fertilization success and both measures of sperm motility were tested (Table 6.5). Motility rank showed a moderate correlation with fertilization success when there was no time delay, but this effect was not significant with a Bonferroni-adjusted ct-level of 0.004. Motility rank was not correlated significantly with fertilization success when sperm were activated before being added to eggs. Duration of sperm motility was not correlated significantly with fertilization success for any of the treatments. 103 Experiment 3: Changes in Fertilization Rate With Time After Egg Activation Eggs retained high fertility after submersion in water for up to 20 s but, beyond this, fertilization success was greatly reduced (Fig. 6.4). However, even after 180 s of exposure to water, some eggs were fertilized in all but one replicate. Experiment 4'- Fertilization Rate v. Time of Egg-Sperm Association Irreversible egg-sperm association can occur very rapidly. Even when eggs were exposed to sperm for <1 s, a small to moderate proportion was always fertilized in this experiment (Fig. 6.5). Consistently high fertilization success was achieved after 5-10 s of exposure. Multiple comparisons between successive time exposures revealed significant increases in fertilization success between the 2 and 5 s treatments and between the 5 and 10 s treatments (Scheffe's F-test, P's < 0.05). Neither sperm motility rank nor duration of motility were correlated significantly with fertilization success for any of the treatments (Spearman rank correlations, P's > 0.05). Discussion Several factors may influence the proportion of eggs fertilized in a given spawning event. The first of these is sperm concentration around the eggs. This is influenced chiefly by the number of sperm released by the male(s). Salmon milt contains an extremely large number of sperm per unit volume of milt (of the order of 107 sperm cells per pl milt (Bouk & Jacobson 1976;Chapter 2). Ejaculate volume has not been measured in sockeye salmon, but during spawning events in the wild there is a visible cloud of milt which persists around the eggs until the female buries them (personal observation). This suggests that sperm concentration around the eggs is in the range of the higher concentrations used in this experiment (50-100 pl packed sperm per 500 ml 104 water). Furthermore, strippable milt volume, even in small males, is typically much greater than the volumes used in this experiment (Foote et al. 1997;personal observation). Foote et al. (1997) found a minimum of 400 pl milt in prespawning males, and, in the presence of nesting females, this volume would be expected to increase by a factor of about four (Olsen & Liley 1993; Liley & Kroon 1995). This can be compared to the present study, in which the average milt volume required to yield 30 pl packed sperm was 99 pA. This represents only a small fraction of milt available to spawning males. Therefore, it is unlikely that fertilization success in the wild is limited by the number of sperm present. Sperm concentration around the eggs and therefore fertilization success is reduced by water turbulence (Petersen et al. 1992), but the structure of salmon nests limits this by creating an eddy in the nest in which water flow is restricted (Foerster 1968). Sperm concentration is also likely to be reduced at the periphery of a gamete cloud, but this again may be affected by the water flow. Fertilization success in the wild is thought to be very high, with estimates for sockeye salmon being >97% (Mathisen 1962). Longevity of eggs and sperm following exposure to fresh water is unlikely to limit fertilization success because the two sexes release gametes simultaneously and spawning events usually last less than 15 s (Mathisen 1962;Chapter 4). There was little or no decrease in fertilization success over this period of time. This is consistent with other reports of salmonid gamete longevity (reviewed in Billard 1992). The duration of spawning events may also influence fertilization success if fertilization is not instantaneous and the process of egg burial washes sperm away from the eggs. The present study suggests that 5-10 s of egg-sperm contact is sufficient to ensure high fertilization success. While most spawning events last at least this long, females sometimes begin to bury eggs <5 s after the start of a spawning event (Mathisen 1962;Chapter 4). While increasing the time of egg-sperm contact may 105 increase fertilization success, it may also increase the probability that eggs are washed away by the current or eaten by predators. Another factor that could influence fertilization success is sperm quality. However, poor sperm quality does not seem to be a significant factor in this population as no case has been observed in which low fertilization success could be attributed to low male fertility. Studies of other species have reported variation in male fertility (e.g., Aas et al. 1991; Billard 1992), but many of the studies of male fertility have been done with hatchery-raised fish and it is possible that this environment does not favour maximum male fertility. Also, there is a methodological problem associated with performing in vitro fertilizations in that some milt samples may be contaminated with urine which can adversely affect sperm function (Dreanno et al. 1998). Male reproductive success is more likely to be affected by competition with other males than by incomplete fertilization (Warner et al. 1995). As a result, species with high risk of sperm competition (probability that more than one male will participate in spawning) have relatively larger testes and greater sperm output than species with low risk of sperm competition (Stockley et al. 1997). This effect can also be seen within species such that phenotypes with a high risk of sperm competition have relatively large testes and milt volume (Gage et al. 1995; Taborsky 1998). This has been confirmed in sockeye salmon (Foote et al. 1997) in which 5Y males are more likely than 3Y males to spawn alone with a female and therefore experience a lower risk of sperm competition, on average, than 3Y males (personal observation). The number of a given male's sperm (relative to that of other males) that surround an egg is influenced not only by the volume of milt released. First, spawning occurs in three-dimensional space, and a male's position relative to that of other males will affect the relative proximity of his sperm to the eggs. Second, spawning occurs over time, and the sooner a male begins to spawn, the sooner and the more of his sperm are available 106 to compete against those of other males. The importance of timing is influenced by how quickly eggs are fertilized. While the speed of fertilization was not measured directly and the efficiency of the egg-rinsing technique was not tested, the results correspond closely to observations of Iwamatsu et al. (1991) on fertilization in medaka [Oryzias latipes). They videotaped gametic interactions and found that most micropyles were occupied by sperm in 1-6 s at high sperm concentration and this time increased as sperm concentration decreased. Therefore, males that start late in a competitive spawning event may be severely limited in the number of eggs they can fertilize. Future models of sperm competition may need to incorporate both the continuous time scale over which gametes are released (Ball & Parker 1997) and variation in the speed with which sperm attach to eggs, enter the micropyle, and achieve fertilization. It has been hypothesized that a high risk of sperm competition imposes selection not only for large numbers of sperm, but for high quality sperm as well (Gage et al. 1995; Stockley et al. 1997). In support of this hypothesis, Gage et al. (1995) found that Atlantic salmon parr had a higher percentage of motile sperm and greater sperm longevity than did anadromous males. There was a difference between male age classes of sockeye salmon in sperm longevity but not in the proportion of sperm that were motile. Given the speed with which fertilization occurs, the slight difference in longevity is unlikely to influence success in sperm competition. Furthermore, there was no difference in fertilizing ability between age classes of male sockeye salmon. Another important test of the sperm quality hypothesis will be to examine success of alternative phenotypes in competitive fertilizations. The relationship between sperm motility and fertilizing ability remains equivocal. A positive correlation has been found in hatchery-raised rainbow trout (0. mykiss) (Moccia & Munkittrick 1987; Ciereszko & Dabrowski 1994; Lahnsteiner et al. 1998)), but not in Atlantic salmon (Aas et al. 1991) or Atlantic cod (Gadus morhua) (Trippel 107 & Neilson 1992). There was no relationship between sperm motility and fertilization success in my study of sockeye salmon. While my experiment was not powerful enough to detect a weak or moderate correlation, I agree with Litvak & Trippel (1998) that in vitro methods for measuring sperm motility may not reflect what happens in the presence of eggs and ovarian fluid. In addition, it is not clear which measures of motility are relevant to fertilizing ability. Until a link between sperm motility and fertilizing ability is clearly established (e.g., Levitan 2000), the only reliable measure of a male's sperm quality is its ability to fertilize eggs. The present study was not initially intended to examine variation in egg quality among females. While only females that were in good condition and had not spawned previously were selected, I found considerable variation in egg quality. This agrees with other studies (Craik &; Harvey 1984; Nagler et al. 2000). Some of the variation in this study may have resulted from variation in ovulation dates, since egg quality decreases with time after ovulation (Craik & Harvey 1984; Liley & Rouger 1990; Flett et al. 1996; de Gaudemar & Beall 1998). Other factors such as health and nutrition may also influence fertility and survival of eggs. Regardless of the source, this variation has important implications for male mate choice. In addition to the well-established benefit of choosing females with greater numbers of eggs (e.g., Grant et al. 1995), males may benefit from choosing females on the basis of egg quality. The combined effects of intrinsic egg quality, timing of nest initiation relative to ovulation, and the passage of time with successive nests are likely to create substantial variation in egg quality among nesting wild females. While it may not be possible for males to distinguish females based on egg quality directly, females in the wild undergo obvious physical deterioration and males could choose females on this basis. The breeding system of salmonids is likely to be very dynamic, with female quality continuously changing as both the current nesting bout and the breeding season progress. 108 In conclusion, this study demonstrates that the proportion of a female's clutch that gets fertilized is not affected by her mate's age and is not likely to be limited by the number and longevity of sperm released. However, the brief time window during which eggs remain available to be fertilized make position and timing critical factors in determining male success in sperm competition. Finally, male choice of females based on egg quality may be more important than previously recognized. Given that the present experiments were performed in vitro, further studies on naturally spawning fish are necessary to confirm these conclusions. 109 Table 6.1: Attempted and actual time (from video analysis) of egg exposure to sperm in experiment 4. Sample size was six in all cases. Attempted Actual time (s) time (s) Range Mean SD <1 0.30 - 0.40 0.36 0.03 1 0.83 - 1.17 1.05 0.12 2 1.95 - 2.08 2.04 0.06 5 4.92 - 5.10 5.02 0.07 10 9.97- 10.17 10.04 0.07 180 179.45 - 180.28 180.07 0.31 110 Table 6.2: Analysis of Variance of fertilization rate (arcsine square root transformed) experiment 1. Female was included as a random factor. Source df SS F P Female 10 3.383 12.4449 <0.001 Male age 1 0.009 0.313 0.578 Sperm concentration 3 3.777 46.322 <0.001 Male age x sperm cone. 3 0.039 0.481 0.696 Error 70 1.90 111 Table 6.3: Spearman rank correlations of fertilization rates of male pairs in experiment 1. Sample size was 11 in all cases. Sperm concentration (fA packed sperm per 500 ml water) r P 5 0.727 0.022* 15 0.788 0.013* 50 0.743 0.019* 100 0.453 0.152 * Still significant with a-levels adjusted using the sequential Bonferroni procedure (Rice 1989) 112 Table 6.4: Analysis of variance for fertilization rate (arcsine square root transformed) experiment 2. Female was included as a random factor. Source df SS F P Female 7 0.52 3.57 0.002 Male age 1 0.0004 0.02 0.886 Time after activation 5 31.38 302.24 <0.0001 Male age x time 5 0.02 0.18 0.971 Error 77 33.52 113 Table 6.5: Spearman rank correlation coefficients between fertilization success and two measures of motility in experiment 2. Both male age classes were included so that N — 16 for all tests. A l l P's greater than 0.1 except where noted. Time after activation (s) Correlation between fertilization rate and: Motility rank Motility duration 0 0.575* -0.050 5 0.306 0.368 10 0.125 0.350 20 0.265 -0.089 40 -0.145 -0.221 80 0.250 0.076 *P = 0.026, a B o n f = 0.004 114 TJ CD N •E 0) c g o CL o 5 15 50 100 Packed sperm concentration (ui per 500ml water) Figure 6.1: Relationship between proportion of eggs fertilized and sperm concentration for 3Y (open boxes) and 5Y (shaded boxes) males in experiment 1. Lines are drawn at the medians with boxes encompassing the interquartile ranges. Whiskers represent total range of the data. 115 5 15 50 100 Packed sperm concentration (u.l per 500ml water) Figure 6.2: Proportion of eggs fertilized in the sperm concentration experiment. Only two replicates are included to illustrate difference among replicates and similarity among males within replicates. Squares represent replicate #3 and circles represent replicate #9. Open symbols and dashed lines represent 3Y males. Closed symbols and solid lines represent 5Y males. 116 1.0 J 0.8 = 0.6 J tr CD C o o CL O 0.4 0.2 J 0.0 J Time after sperm activation (s) Figure 6.3: Relationship between proportion of eggs fertilized and duration of milt exposure to water for 3Y (open boxes) and 5Y (shaded boxes) males in experiment 2. 117 c o '•c o Q . O 1.0 J 0.8 J = 0.6 •c cu 0.4 0.2 J 0.0 J r 20 40 80 Time after egg activation (s) Figure 6.4: Relationship between proportion of eggs fertilized and duration of egg exposure to water in experiment 3. 118 1.0 J 0.8 T 3 CD N P 0.6 J CD c o '•c o Q_ o 0.4 0.2 J Sperm-egg exposure time (s) Figure 6.5: Relationship between proportion of eggs fertilized and duration of egg exposure to milt in experiment 4. 119 Chapter 7 General Discussion Many animal species exhibit striking variation in male reproductive behaviour, morphology, and life history. Such is the case for most Pacific salmon species which are semelparous and in which males vary in age at maturity (female age at maturity also varies, but to a lesser extent (Groot & Margolis 1991; Fleming 1998)). Variation in male age at maturity is associated with variation in body size, colouration, size of secondary sexual traits, and behaviour. In this thesis I have examined several aspects of sockeye salmon (Oncorhynchus nerka) reproduction that may contribute to variation in reproductive success among males and, ultimately, to variation in male age at maturity. A field experiment dealt with competition among males for mating opportunities and fertilization success in single spawning events. A set of in vitro fertilization experiments dealt with competition among males' sperm to fertilize eggs. In my first experiment I found that mean fertilization success of male age classes can be negatively frequency-dependent at single spawning events (Chapter 3). An individual is likely to have higher fertilization success in a spawning event where there are relatively few other males of the same age than when that age class is in the majority. This provides a mechanism through which a mixed ESS may be maintained (Maynard Smith 1982). That is, different ages at maturity could represent genetically distinct life history strategies with equal fitness. However, while there is evidence that age at maturity may have a genetic component (Iwamoto et al. 1984; Hankin et al. 1993; Heath et al. 1994), maturation occurs when individuals reach a size threshold (Wood & Foote 1996). Thus, the age at which maturation occurs is likely to be influenced strongly by environmental conditions that affect growth rate. This does not diminish the importance of frequency-dependent selection (FDS) in systems without 120 genetic polymorphisms. When male alternative reproductive phenotypes (ARP's) are condition-dependent, FDS can still maintain genetic variation (Roff 1996, 1998), and it influences the point at which males 'switch' between tactics, thus affecting the phenotypic makeup of a population (Hutchings & Myers 1994; Repka & Gross 1995; Roff 1996, 1998). The presence of frequency-dependent reproductive success in my experiments was dependent on the number of males present or the operational sex ratio (OSR). Frequency-dependence occurred when three males had the opportunity to spawn with a nesting female but not when six males were present. This could lead to instability in selection on age at maturity since both population size and OSR fluctuate in natural salmon populations (Groot & Margolis 1991; Quinn et al. 1996). However, my study only considered FDS at the level of individual spawning events. To fully understand the effect of population age structure on individual reproductive success it will be necessary to consider FDS at the population level. In Chapter 4 I examined factors influencing male fertilization success in spawning events. Several other studies have examined fertilization success of salmon (various species) spawning in enclosures or semi-natural channels (e.g., Maekawa & Onozato 1986; Hutchings & Myers 1988; Jordan & Youngson 1992; Moran et al. 1996; Thomaz et al. 1997; Mj0lner0d et al. 1998; Thompson et al. 1998). Behavioural data were not collected in all of these studies, the sample sizes are generally small, and they vary in the number and size of males used. However, the importance of male dominance in fertilization success is a general pattern that emerges (but see Foote et al. (1997) for a notable exception). My study identified two important sources of variation in male success: probability of mating and proportion of eggs fertilized. 5Y males with small bodies and small secondary sex traits were less likely to participate in spawning than large males with large secondary sex traits. A small-bodied 5Y male that encounters a 121 female with a larger alpha male may choose to spawn with that female as a satellite or abandon the opportunity and search for other females. Because small-bodied 5Y males are still large relative to the whole male population, a small 5Y male may be able to find a female with which it can achieve alpha status and obtain a high reproductive payoff. For males that do not abandon the current reproductive opportunity and participate in spawning, fertilization success is partly determined by the following sequence: body size ,. . . + —»• dist. from female —¥ spawn start time —> spawn duration —>• success behaviour Once the process of spawning (gamete release) begins, competition between individual males shifts to competition between their sperm. The outcome of this competition may be partly a function of the number of sperm released by competing males (Parker 1982). However, variation in competitive ability of male gametes and female choice may also contribute to the outcome of sperm competition (e.g., Gile & Ferguson 1995; Bishop et al. 1996; Dziuk 1996; Olsson et al. 1997; Delph & Havens 1998; Radwan 1998; Birkhead et al. 1999; Palumbi 1999). In Chapter 5 I showed that success in sperm competition between sockeye salmon males is also influenced by factors other than the number of sperm. In this in vitro fertilization experiment, pairs of males contributed equal numbers of sperm that were thoroughly mixed before being added to eggs so that eggs had an equal probability of first coming in contact with sperm from either male. While age classes did not differ in competitive fertilization success, there was significant variation among males. The sources of differences in sperm competitive ability remain unknown, but these differences could contribute to variation in fertilization success in spawning events as seen in Chapter 4. However, it will be necessary to test if differences in competitive success between males are consistent across females (Gharrett & Shirley 1985; Gile & Ferguson 1995; Clark et al. 1999). It will also be necessary to test if differences in success between males are 122 transitive. That is, if male A beats male B and male B beats male C, does male A beat male C? This is not the case in Drosophila (Clark et al. 2000). In addition, to extend our understanding of the link between success in sperm competition and success in natural spawning events, further experiments are required to determine the relationship between number of sperm released and male success, the relationship between the timing of sperm release (i.e., when a male begins spawning relative to other males) and male success, and the mechanisms underlying, differences in sperm competitive ability. Success in sperm competition may be influenced by characteristics of the gametes (Birkhead et al. 1999) and by the spatial and temporal dynamics of fertilization (Yund & McCartney 1994). In Chapter 6 I performed in vitro fertilization experiments to examine these factors in sockeye salmon. The results emphasized the importance of timing in sockeye salmon fertilization. Sperm maintained high fertility only for about 10 s and the association between eggs and sperm occurred rapidly. This suggests that the sooner males begin to spawn and the closer they release their sperm to the eggs, the higher will be their fertilization success relative to other males. However, spawning position was not correlated with fertilization success in my experimental spawning trials, and path analysis indicated that starting time did not have a direct effect (Chapter 4). The spawning trial analysis may not have been powerful enough to detect the effects of these variables or they may have been overwhelmed by other factors. A drawback of both fertilization experiments is that males were not exposed to mating stimuli before their milt was collected. When male salmonids interact with nesting females they undergo hormonal changes and the volume of milt in their sperm ducts increases (Liley et al. 1993; Rouger & Liley 1993; Cardwell et al. 1996). Exposing male goldfish (Carassius auratus) to a female pheromone results in increased sperm motility and increased success in competitive fertilizations compared to males that are not exposed (Zheng et al. 1997). Thus, the dynamics of fertilization and outcome of 123 sperm competition may be influenced by the social environment of the males involved. Another factor that may influence competition between male classes is health of the fish. In my experiments, fish were free of external damage caused by agents such as parasites, fungus, and predators. However, some fish that arrive at the spawning grounds are afflicted with health problems and these may differ between age classes. For example, nematodes are common parasites in sockeye salmon, and the degree of infection (relative to body size) increases with sea age (Berg et al. 1995). The risk of predation by bears is also greater for larger (older) fish than for smaller (younger) fish (Ruggerone et al. 2000). However, comparisons of age classes in wounding from unsuccessful predation attempts have not been performed. Overall, the results of my research suggest that 5Y males do not have an advantage over 3Y males when competing to fertilize a given female's eggs. This is true in spite of large differences between the two male age classes in body size and secondary sex traits. One reason for this similarity in success is that each age class has the highest success when it is rare. The lack of difference originates with competition for positions around a nesting female, with one 5Y male usually winning the alpha position, 3Y males obtaining intermediate positions, and 5Y satellite males being furthest from the female. Position prior to spawning influences the starting time of sperm release (relative to other males), the duration of sperm release and, ultimately, fertilization success. Sperm from the two age classes are equally capable of fertilizing eggs and show no difference in competitive ability. Thus, competitive equivalence of 3Y and 5Y males extends from behavioural competition between males when they encounter a nesting female right through to the competition between their sperm to fertilize eggs. These aspects of male competition represent important components of male fitness, and my results help to explain why we see such a great degree of variation in male age at maturity. In fact, my results raise the question of why early maturing males are rare in most populations. 124 Future studies need to examine the number of matings, characteristics of females mated with (number of eggs, stage in spawning cycle, etc.), and offspring survival for a complete picture of reproductive success in different age classes of male sockeye salmon. 125 References Aas, G. H. , Refstie, T. k Gjerde, B. 1991. Evaluation of milt quality of Atlantic salmon. Aquaculture 95:125-132. Alcock, J. , Jones, C. E. & Buchmann, S. L. 1977. Male mating strategies in the bee Centris pallida Fox (Anthophoridae: Hymenoptera). American Naturalist 111:145-155. Allen, J. A . 1988. Frequency-dependent selection by predators. Philosophical Transactions of the Royal Society of London B 319:485-503. Allen, J. A . , Raison, H. E. & Weale, M . E. 1998. The influence of density on frequency-dependent selection by wild birds feeding on artificial prey. Proceedings of the Royal Society of London B 265:1031-1035. Andersson, M . 1994. Sexual selection. Princeton University Press, Princeton, N.J . Antonovics, J. & Kareiva, P. 1988. Frequency-dependent selection and competition: empirical approaches. Philosophical Transactions of the Royal Society of London B 319:601-613. Arak, A . 1984. Sneaky breeders. In Producers and scroungers. Strategies of exploitation and parasitism, edited by Barnard, C. J., Chapman and Hall, New York, pp. 154-194. Arnqvist, G. 1992. The effects of operational sex ratio on the relative mating success of extreme male phenotypes in the water strider Gerris odontogaster (Zett. Heteroptera; Gerridae). Animal Behaviour 43:681-683. Austad, S. N . 1984. A classification of alternative reproductive behaviors and methods for field-testing ESS models. American Zoologist 24:309-319. 126 Ayala, F. J. & Campbell, C. A . 1974. Frequency-dependent selection. Annual Review of Ecology and Systematics 5:115-138. Ball, M . A. &; Parker, G. A . 1997. Sperm competition games: inter- and intra -species results of a continuous external fertilization model. Journal of Theoretical Biology 186:459-466. Beacham, T. D., Margolis, L. & Nelson, R. J. 1998. A comparison of methods of stock identification for sockeye salmon (Oncorhynchus nerka) in Barkley Sound, British Columbia. North Pacific Anadromous Fish Commission Bulletin 1:227-239. Berejikian, B. A. , Tezak, E. P. & LaRae, A . L. 2000. Female mate choice and spawning behaviour of chinook salmon under experimental conditions. Journal of Fish Biology 57:647-661. Berg, O. K . , Foote, C. J. &; Quinn, T. P. 1995. Fish age, nematode (Philonema oncorhynchi) infection, and development of sexual dimorphism by the adult male sockeye salmon, Oncorhynchus nerka, in western Alaska. Canadian Journal of Zoology 73:1999-2004. Billard, R. 1992. Reproduction in rainbow trout: sex differentiation, dynamics of gametogenesis, biology and preservation of gametes. Aquaculture 100:263-298. Billard, R., Christen, R., Cosson, M . P., Gatty, J. L. , Letellier, L . , Renard, P. & Saad, A. 1986. Biology of the gametes of some teleost species. Fish Physiology and Biochemistry 2:115-120. Birkhead, T. R. 1998. Cryptic female choice: criteria for establishing female sperm choice. Evolution 52:1212-1218. Birkhead, T. R., Martinez, J. G. & Froman, D. P. 1999. Sperm mobility determines the 127 outcome of sperm competition in the domestic fowl. Proceedings of the Royal Society of London B 266:1759-1764. Birkhead, T. R. k M0ller, A . P. (editors). 1998. Sperm competition and sexual selection. Academic Press, New York. Bishop, J. D. D., Jones, C. S. k Noble, L. R. 1996. Female control of paternity in the internally fertilizing compound ascidian Diplosoma listerianum. II. Investigation of male mating success using R A P D markers. Proceedings of the Royal Society of London B 263:401-407. Blanchfield, P. J. & Ridgway, M . S. 1999. The cost of peripheral males in a brook trout mating system. Animal Behaviour 57:537-544. Bouk, G. R. k Jacobson, J. 1976. Estimation of salmonid sperm concentration by microhematocrit technique. Transactions of the American Fisheries Society 105:534-535. Brockman, H. J., Nguyen, C. k Potts, W. 2000. Paternity in horseshoe crabs when spawning in multiple-male groups. Animal Behaviour 60:837-849. Brockmann, H. J., Colson, T. k Potts, W. 1994. Sperm competition in horseshoe crabs (Limulus polyphemus). Behavioral Ecology and Sociobiology 35:153-160. Burgner, R. L. 1991. Life history of sockeye salmon (Oncorhynchus nerka). In Pacific salmon life histories, edited by Groot, C. k Margolis, L. , U B C Press, Vancouver, pp. 1-117. Cardwell, J. R., Sorensen, P. W., Kraak, G. J. V . D. k Liley, N . R. 1996. Effect of dominance status on sex hormone levels in laboratory and wild-spawning male trout. General and Comparative Endocrinology 101:333-341. 128 Chebanov, N . A . 1980. Spawning behavior of the pink salmon, Oncorhynchus gorbuscha. Journal of Ichthyology 20:64-73. Ciereszko, A . & Dabrowski, K . 1994. Relationship between biochemical constituents of fish semen and fertility: The effect of short term storage. Fish Physiology and Biochemistry 12:357-367. Clark, A . G., Begun, D. J. k Prout, T. 1999. Female x male interactions in Drosophila sperm competition. Science 283:217-220. Clark, A . G., Dermitzakis, E. T. k Civetta, A . 2000. Nontransitivity of sperm precedence in Drosophila. Evolution 54:1030-1035. Clutter, R. I. k Whitesel, L. E. 1956. Collection and interpretation of sockeye salmon scales. Technical report, International Pacific Salmon Fisheries Commission Bulletin. Conner, J. 1989. Density-dependent sexual selection in the fungus beetle, Bolitotherus cornutus. Evolution 43:1378-1386. Cote, I. M . k Hunte, W. 1989. Male and female choice in the redlip blenny: why bigger is better. Animal Behaviour 38:78-88. Coulson, T. N . , Pemberton, J. M . , Albon, S. D., Beaumont, M . , Marshall, T. C , Slate, J., Guinness, F. E. k Clutton-Brock, T. H. 1998. Microsatellites reveal heterosis in red deer. Proceedings of the Royal Society of London B 265:489-495. Craik, J. C. A . k Harvey, S. M . 1984. Egg quality in rainbow trout: the relation between egg viability, selected aspects of egg composition, and time of stripping. Aquaculture 40:115-134. Cressman, R. k Dash, A . T. 1987. Density dependence and evolutionary stable strategies. Journal of Theoretical Biology 126:393-406. 129 Day, T. Sz Taylor, P. D. 1996. Evolutionarily stable versus fitness maximizing life histories under frequency-dependent selection. Proceedings of the Royal Society of London B 263:333-338. Daye, P. G. Sz Glebe, B. D. 1984. Fertilization success and sperm motility of Atlantic salmon (Salmo salar L.) in acidified water. Aquaculture 43:307-312. de Gaudemar, B. Sz Beall, E. 1998. Effects of overripening on spawning behaviour and reproductive success of Atlantic salmon females spawning in a controlled flow channel. Journal of Fish Biology 53:434-446. de Gaudemar, B. , Bonzom, J. M . Sz Beall, E . 2000. Effects of courtship and relative mate size on sexual motivation in Atlantic salmon. Journal of Fish Biology 57:502-515. DeFraipont, M . Sz Sorensen, P. W. 1993. Exposure to the phermone 17a,20/?-dyhydroxy-4-pregnen-3-one enhances the behavioural spawning success, sperm production and sperm motility of male goldfish. Animal Behaviour 46:245-256. Delph, L. F. Sz Havens, K . 1998. Pollen competition in flowering plants. In Sperm competition and sexual selection, edited by Birkhead, T. R. Sz M0ller, A . P., Academic Press, New York, pp. 147-173. Dominey, W. J . 1984. Alternative mating tactics and evolutionarily stable strategies. American Zoologist 24:385-396. Dreanno, C , Suquet, M . , Desbruyeres, E., Cosson, J., Delliou, H. L. Sz Billard, R. 1998. Effect of urine on semen quality in turbot (Psetta maxima). Aquaculture 169:247-262. Dziuk, P. J. 1996. Factors that influence the proportion of offspring sired by a male following heterospermic insemination. Animal Reproduction Science 43:65-88. 130 Eadie, J. M . & Fryxell, J. M . 1992. Density dependence, frequency dependence, and alternative nesting strategies in goldeneyes. American Naturalist 140:621-641. Fisher, R. A . 1930. The genetical theory of natural selection. Clarendon Press, Oxford. Fleming, I. A . 1996. Reproductive strategies of atlantic salmon: ecology and evolution. Reviews in Fish Biology and Fisheries 6:379-416. Fleming, I. A . 1998. Pattern and variability in the breeding system of Atlantic salmon (Salmo salar), with comparisons to other salmonids. Canadian Journal of Fisheries and Aquatic Science 55(Suppl. l):59-76. Fleming, I. A . & Gross, M . R. 1992. Reproductive behavior of hatchery and wild coho salmon (Oncorhynchus kisutch): does it differ? Aquaculture 103:101-121. Fleming, I. A . & Gross, M . R. 1994. Breeding competition in a Pacific salmon (coho: Oncorhynchus kisutch): measures of natural and sexual selection. Evolution 48:637-657. Flett, P. A. , Munkittrick, K . R. &; Van Der Kraak, G. 1996. Overripening as the cause of low survival to hatch in Lake Erie coho salmon (Oncorhynchus kisutch) embryos. Canadian Journal of Zoology 74:851-857. Foerster, R. E. 1968. The sockeye salmon, Oncorhynchus nerka. Fisheries Research Board of Canada Bulletin 162:1-422. Foote, C. J. 1989. Female mate preference in pacific salmon. Animal Behaviour 38:721-723. Foote, C. J . 1990. A n experimental comparison of male and female spawning territoriality in a Pacific salmon. Behaviour 115:283-314. 131 Foote, C. J., Brown, G. S. & Wood, C. C. 1997. Spawning success of males using alternative mating tactics in sockeye salmon, Oncorhynchus nerka. Canadian Journal of Fisheries and Aquatic Science 54:1785-1795. Foote, C. J. & Larkin, P. A . 1988. The role of male choice in the assortative mating of anadromous and non-anadromous sockeye salmon (Oncorhynchus nerka). Behaviour 106:43-62. Gadgil, M . 1972. Male dimorphism as a consequence of sexual selection. American Naturalist 106:574-580. Gadgil, M . & Taylor, C. E. 1975. Plausible models of sexual selection and polymorphism. American Naturalist 109:470-472. Gage, M . J. G., Stockley, P. & Parker, G. A . 1995. Effects of alternative male mating strategies on characteristics of sperm production in the Atlantic salmon (Salmo salar): theoretical and empirical investigations. Philosophical Transactions of the Royal Society of London B 350:391-399. Gharrett, A . J. & Shirley, S. M . 1985. A genetic examination of spawning methodology in a salmon hatchery. Aquaculture 47:245-256. Gile, S. R. k, Ferguson, M . M . 1995. Factors affecting male potency in pooled gamete crosses of rainbow trout, Oncorhynchus mykiss. Environmental Biology of Fishes 42:267-275. Gilkey, J. C. 1981. Mechanisms of fertilization in fishes. American Zoologist 21:359-375. Ginsburg, A . S. 1963. Sperm-egg association and its relationship to the activation of the egg in salmonid fishes. Journal of Embryology and Experimental Morpholology 11:13-33. 132 Ginzburg, A . S. 1972. Fertilization in fishes and the problem of polyspermy. Isreal Program for Scientific Translations, Jerusalem. Grace, J. B. & Pugesek, B. H. 1998. On the use of path analysis and related procedures for the investigation of ecological problems. American Naturalist 152:151-159. Grant, J. W. A. , Casey, P. C , Bryant, M . J. & Shahsavarani, A . 1995. Mate choice by male Japanese medaka (Pisces, Oryziidae). Animal Behaviour 50:1425-1428. Greenfield, M . D. & Shelly, T. E. 1985. Alternative mating strategies in a desert grasshopper: evidence of density-dependence. Animal Behaviour 33:1192-1210. Gregory, P. T. & Skebo, K . M . 1998. Trade-offs between reproductive traits and the influence of food intake during pregnency in the garter snake, Thamnophis elegans. American Naturalist 151:477-486. Groot, C. & Margolis, L. (editors). 1991. Pacific salmon life histories. U B C Press, Vancouver. Gross, M . R. 1982. Sneakers, satellites and parentals: polymorphic mating strategies in North American sunfishes. Zeitschrift fur Tierpsychologie 60:1-26. Gross, M . R. 1984. Sunfish, salmon and the evolution of alternative reproductive strategies and tactics in fishes. In Fish reproduction: strategies and tactics, edited by Potts, G. W. & Wootton, R. J., Academic Press, London, pp. 55-75. Gross, M . R. 1985. Disruptive selection for alternative life histories in salmon. Nature 313:47-48. Gross, M . R. 1991. Evolution of alternative reproductive strategies: frequency-dependent sexual selection in male bluegill sunfish. Philosophical Transactions of the Royal Society of London B 332:59-66. 133 Gross, M . R. 1996. Alternative reproductive strategies and tactics: diversity within sexes. Trends in Ecology and Evolution 11:92-98. Gross, M . R. & Charnov, E. L. 1980. Alternative male life histories in bluegill sunfish. Proceedings of the National Academy of Science 77:6937-6940. Hankin, D. G., Nicholas, J. W. & Downey, T. W. 1993. Evidence for inheritance of age of maturity in chinook salmon (Oncorhynchus tshawytscha). Canadian Journal of Fisheries and Aquatic Science 50:347-358. Hanson, A . J. & Smith, H. D. 1967. Mate selection in a population of sockeye salmon (Oncorhynchus nerka) of mixed age-groups. Journal of the Fisheries Research Board of Canada 24:1955-1977. Hazel, W. & Smock, R. 2000. Inheritance in the conditional strategy revisited. Journal of Theoretical Biology 204:307-309. Hazel, W. N . , Smock, R. & Johnson, M . D. 1990. A polygenic model for the evolution and maintenance of conditional strategies. Proceedings of the Royal Society of London B 242:181-187. Healey, M . C. & Prince, A . 1998. Alternative tactics in the breeding behaviour of male coho salmon. Behaviour 135:1099-1124. Heath, D. D., Devlin, R. H. , Heath, J. W. k Iwama, G. K . 1994. Genetic, environmental and interaction effects on the incidence of jacking in Oncorhynchus tshawytscha (chinook salmon). Heredity 72:146-154. Hendry, A . P., Leonetti, F. E. & Quinn, T. P. 1995. Spatial and temporal isolating mechanisms: the formation of discrete breeding aggregations of sockeye salmon (Oncorhynchus nerka). Canadian Journal of Zoology 73:339-352. 134 Hines, W. G. S. & Turelli, M . 1997. Multilocus evolutionarily stable strategy models: additive effects. Journal of Theoretical Biology 187:379-388. Hogg, J. T. 1984. Mating in bighorn sheep: multiple creative male strategies. Science 225:526-529. Howard, R. D. 1978. The evolution of mating strategies in bullfrogs, Rana catesbeiana. Evolution 32:850-871. Hoysak, D. J. & Liley, N . R. 2001. Fertilization dynamics in sockeye salmon and a comparison of sperm from alternative male phenotypes. Journal of Fish Biology 58:1286-1300. Hurlbert, S. H. 1984. Pseudoreplication and the design of ecological field experiments. Ecological Monographs 54:187-211. Hutchings, J. A . 1993. Adaptive life histories affected by age-specific survival and growth rate. Ecology 74:673-684. Hutchings, J. A . &; Myers, R. A . 1988. Mating success of alternative maturation phenotypes in male Atlantic salmon, Salmo salar. Oecologia 75:169-174. Hutchings, J. A . & Myers, R. A . 1994. The evolution of alternative mating strategies in variable environments. Evolutionary Ecology 8:256-268. Huxley, J. S. 1930. The maladaptation of trout spermatozoa to fresh water. Nature 125:494. Ihaka, R. & Gentleman, R. 1996. R: a language for data analysis and graphics. Journal of Computational and Graphical Statistics 5:299-314. 135 Iwamatsu, T., Onitake, K . , Yoshimoto, Y . & Hiramoto, Y . 1991. Time sequence of early event in fertilization in the medaka egg. Development, Growth and Differentiation 33:479-490. Iwamoto, R. N . , Alexander, B. A . & Hershberger, W. K . 1984. Genotypic and environmental effects on the incidence of sexual precocity in coho salmon (Oncorhynchus kisutch). Aquaculture 43:105-121. Jarvi, T. 1990. The effects of male dominance, secondary sexual characteristics and female mate choice on the mating success of male Atlantic salmon Salmo salar. Ethology 84:123-132. Jirotkul, M . 1999. Operational sex ratio influences female preference and male-male competition in guppies. Animal Behaviour 58:287-294. Jordan, W. C. & Youngson, A . F. 1992. The use of genetic marking to assess the reproductive success of mature male Atlantic salmon parr (Salmo salar, L.) under natural spawning conditions. Journal of Fish Biology 41:613-618. Keenleyside, M . H. A . & Dupuis, H. M . C. 1988. Courtship and spawning competition in pink salmon (Oncorhynchus gorbuscha). Canadian Journal of Zoology 66:262-265. Kim, W. S. & Roberson, K . 1968. On the use of otoliths of sockeye salmon for age determination. University of Washington Publications in Fisheries New Series 3:151-168. Kitano, S., Maekawa, K . , Nakano, S. k, Fausch, K . D. 1994. Spawning behavior of bull trout in the Upper Flathead drainage, Montana, with special reference to hybridization with brook trout. Transactions of the American Fisheries Society 123:988-992. 136 Kobayashi, W. k Yamamoto, T. S. 1981. Fine structure of the micropylar apparatus of the chum salmon egg, with a discussion of the mechanism for blocking polyspermy. Journal of Experimental Zoology 217:265-275. Kodric-Brown, A . 1988. Effect of sex-ratio manipulation on territoriality and spawning success of male pupfish, Cyprinodon pecosensis. Animal Behaviour 36:1136-1144. Koseki, Y . k Maekawa, K . 2000. Sexual selection on mature male parr of masu salmon (Oncorhynchus masou): does sneaking behavior favor small body size and less-developed sexual characters? Behavioral Ecology and Sociobiology 48:211-217. Lahnsteiner, F., Berger, B. , Weismann, T. k Patzner, R. A . 1998. Determination of semen quality of the rainbow trout, Oncorhynchus mykiss, by sperm motility, seminal plasma parameters, and spermatozoal metabolism. Aquaculture 163:163-181. Levitan, D. R. 2000. Sperm velocity and longevity trade off each other and influence fertilization in the sea urchin Lytechinus variegatus. Proceedings of the Royal Society of London B 267:531-534. Liley, N . R. k Kroon, F. J. 1995. Male dominance, plasma hormone concentrations, and availability of milt in male rainbow trout (Oncorhynchus mykiss). Canadian Journal of Zoology 73:826-836. Liley, N . R., Olsen, K . H. , Foote, C. J. k Van Der Kraak, G. J. 1993. Endocrine changes associated with spawning behavior in male kokanee salmon (Oncorhynchus nerka) and the effects of anosmia. Hormones and Behavior 27:470-487. Liley, N . R. k Rouger, Y . 1990. Plasma levels of gonadotropin and 17a:,20/3 -dihydroxy-4-pregnen-3-one in relation to spawning behaviour of rainbow trout, Oncorhynchus mykiss (Walbaum). Journal of Fish Biology 37:699-711. 137 Litvak, M . K . & Trippel, E . A . 1998. Sperm motility patterns of Atlantic cod (Gadus morhua) in relation to salinity: effects of ovarian fluid and egg presence. Canadian Journal of Fisheries and Aquatic Science 55:1871-1877. Maekawa, K . 1983. Streaking behaviour of mature male parrs of the Miyabe charr, Salvelinus malma miyabei, during spawning. Japanese Journal of Ichthyology 30:227-234. Maekawa, K . & Onozato, H. 1986. Reproductive tactics and fertilization success of mature male Miyabe charr, Salvelinus malma miyabei. Environmental Biology of Fishes 15:119-129. Mathisen, O. A . 1962. The effect of altered sex ratios on the spawning of red salmon. In Studies of Alaska Red Salmon, edited by Koo, T. S. Y . , University of Washington Press, Seattle, pp. 137-245. Matsushima, T., Takei, K . , Kitamura, S., Kusunoki, M . , Satou, M . , Okumoto, N . &; Ueda, K . 1989. Rhythmic electromyographic activities of trunk muscles characterize the sexual behavior in the Hime salmon (landlocked sockeye salmon, Oncorhynchus nerka). Journal of Comparative Physiology A 165:293-314. Maynard Smith, J. 1982. Evolution and the theory of games. Cambridge University Press, Cambridge. Maynard Smith, J. & Price, G. R. 1973. The logic of animal conflict. Nature 246:15-18. McDonald, J . & Hume, J . M . 1984. Babine Lake sockeye salmon (Oncorhynchus nerka) enhancement program: testing some major assumptions. Canadian Journal of Fisheries and Aquatic Science 41:70-92. Miura, T., Yamauchi, K . , Takahashi, H. & Nagahama, Y . 1992. The role of hormones in 138 the aquisition of sperm motility in salmonid fish. Journal of Experimental Zoology 261:359-363. Mj0lner0d, I. B. , Fleming, I. A . , Refseth, U . H. k Hindar, K . 1998. Mate and sperm competition during multiple-male spawnings of Atlantic salmon. Canadian Journal of Zoology 76:70-75. Moccia, R. D. k Munkittrick, K . R. 1987. Relationship between the fertilization of rainbow trout (Salmo gairdneri) eggs and the motility of spermatozoa. Theriogenology 27:679-688. Moran, P., Pendas, A . M . , Beall, E . k Garcia-Vazquez, E. 1996. Genetic assessment of the reproductive success of Atlantic salmon precocious parr by means of V N T R loci. Heredity 77:655-660. Morbey, Y . 2000. Protandry in Pacific salmon. Canadian Journal of Fisheries and Aquatic Science 57:1252-1257. Morris, D. B. , Richard, K . R. k Wright, J. M . 1996. Microsatellites from rainbow trout (Oncorhynchus mykiss) and their use for genetic study of salmonids. Canadian Journal of Fisheries and Aquatic Science 53:120-126. Myers, R. A . 1986. Game theory and the evolution of Atlantic salmon (Salmo salar) age at maturation. Canadian Special Publications in Fisheries and Aquatic Science 89:53-61. Nagler, J. J. , Parsons, J. E. k Cloud, J. G. 2000. Single pair mating indicates maternal effects on embryo survival in rainbow trout, Oncorhynchus mykiss. Aquaculture 184:177-183. Nelson, R. J., Beacham, T. D. k Small, M . P. 1998. Microsatellite analysis of the population structure of a Vancouver Island sockeye salmon (Oncorhynchus nerka) 139 stock complex using nondenaturing gel electrophoresis. Molecular Marine Biology & Biotechnology 7:312-319. Noltie, D. B. 1990. Intrapopulation variation in the breeding of male pink salmon (Oncorhynchus gorbuscha) from a Lake Superior tributary. Canadian Journal of Fisheries and Aquatic Science 47:174-179. Oliver, A . S. 1997. Size and density dependent mating tactics in the simultaneously hermaphroditic seabass Serranus subligarius (Cope, 1870). Behaviour 134:563-594. Olsen, K . H. &; Liley, N . R. 1993. The significance of olfaction and social cues in milt availability, sexual hormone status, and spawning behavior of male rainbow trout (Oncorhynchus mykiss). General and Comparative Endocrinology 89:107-118. Olsson, M . , Shine, R. Sz Madsen, T. 1997. Sperm choice by females. Trends in Ecology and Evolution 12:445. O'Reilly, P. T., Hamilton, L. C , McConnell, S. K . Sz Wright, J . M . 1996. Rapid analysis of genetic variation in Atlantic salmon (Salmo salar) by P C R multiplexing of dinucleotide and tetranucleotide microsatellites. Canadian Journal of Fisheries and Aquatic Science 53:2292-2298. Palumbi, S. R. 1999. A l l males are not equal: fertility differences depend on gamete recognition polymorphisms in sea urchins. Proceedings of the National Academy of Science 96:12632-12637. Parker, G. A . 1982. Why are there so many tiny sperm? Sperm competition and the maintenance of two sexes. Journal of Theoretical Biology 96:281-294. Parker, G. A . 1985. Population consequences of evolutionarily stable strategies. In Behavioural ecology. Ecological consequences of adaptive behaviour, edited by Sibly, R. M . Sz Smith, R. H. , Blackwell Scientific Publications, Oxford, pp. 33-58. 140 Parker, G. A . 1990a. Sperm competition games: raffles and roles. Proceedings of the Royal Society of London B 242:120-126. Parker, G. A . 1990b. Sperm competition games: sneaks and extra-pair copulations. Proceedings of the Royal Society of London B 242:127-133. Parker, G. A . , Ball , M . A. , Stockley, P. & Gage, M . J. G. 1996. Sperm comeptition games: individual assessment of sperm competition intensity by group spawners. Proceedings of the Royal Society of London B 263:1291-1297. Petersen, C. W., Warner, R. R., Cohen, S., Hess, H. C. & Sewell, A . T. 1992. Variable pelagic fertilization success: implications for mate choice and spatial patterns of mating. Ecology 73:391-401. Pinheiro, J. C. & Bates, D. M . 2000. Mixed-Effects Models in S and S-PLUS. Springer, New York. Queller, D. C. & Goodnight, K . F. 1989. Estimating relatedness using genetic markers. Evolution 43:258-275. Quinn, T. P. 1999. Variation in Pacific salmon reproductive behaviour associated with species, sex and levels of competition. Behaviour 136:179-204. Quinn, T. P., Adkinson, M . D. & Ward, M . B. 1996. Behavioral tactics of male sockeye salmon (Oncorhynchus nerka) under varying operational sex ratios. Ethology 102:304-322. Quinn, T. P. & Foote, C. J. 1994. The effects of body size and sexual dimorphism on the reproductive behaviour of sockeye salmon, Oncorhynchus nerka. Animal Behaviour 48:751-761. 141 Radwan, J. 1998. Heritability of sperm competition success in the bulb mite, Rhizoglyphus robini. Journal of Evolutionary Biolgy 11:321-327. Rakitin, A. , Ferguson, M . M . & Trippel, E. A . 1999. Sperm competition and fertilization success in Atlantic cod (Gadus morhua): effect of sire size and condition factor on gamete quality. Canadian Journal of Fisheries and Aquatic Science 56:2315-2323. Repka, J . & Gross, M . R. 1995. The evolutionarily stable strategy under individual condition and tactic frequency. Journal of Theoretical Biology 176:27-31. Rice, W. R. 1989. Analyzing tables of statistical tests. Evolution 43:223-223. Roff, D. A . 1992. The evolution of life histories. Chapman and Hall, London. Roff, D . A . 1996. The evolution of threshold traits in animals. Quarterly Review of Biology 71:3-35. Roff, D. A . 1998. The maintenance of phenotypic and genetic variation in threshold traits by frequency-dependent selection. Journal of Evolutionary Biolgy 11:513-529. Rouger, Y . & Liley, N . R. 1993. Effect of social environment on plasma hormones and availability of milt in spawning male rainbow trout (Oncorhynchus mykiss Walbaum). Canadian Journal of Zoology 71:280-285. Rubenstein, D. I. 1980. On the evolution of alternative mating strategies. In Limits to action. The allocation of individual behavior, edited by Staddon, J . E . R., Academic Press, New York, pp. 65-100. Ruchon, F., Laugier, T. k. Quignard, J. P. 1995. Alternative male reproductive strategies in the peacock blenny. Journal of Fish Biology 47:826-840. 142 Ruggerone, G. T., Hanson, R. & Rogers, D. E. 2000. Selective predation by brown bears (Ursus arctos) foraging on spawning sockeye salmon (Oncorhynchus nerka). Canadian Journal of Zoology 78:974-981. Sail, J. & Lehman, A . 1996. JMP® start statistics. Duxbury Press, Toronto. Sargent, R. C , Gross, M . R. & van den Berghe, E. P. 1986. Male mate choice in fishes. Animal Behaviour 34:545-550. Scaggiante, M . , Mazzoldi, C , Petersen, C. W. & Rasotto, M . B . 1999. Sperm competition and mode of fertilization in the grass goby Zosterisessor ophiocephalus (Teleostei: Gobiidae). Journal of Experimental Zoology 283:81-90. Schroder, S. L. 1982. The influence of intrasexual competition on the distribution of chum salmon in an experimental stream. In Proceedings of the salmon and trout migratory symposium, edited by Brannon, E. L. & Salo, E. O., University of Washington Press, Seattle, Washington, pp. 275-285. Scribner, K . T., Gust, J. R. & Fields, R. L. 1996. Isolation and characterization of novel salmon microsatellite loci: cross-species amplification and population genetic applications. Canadian Journal of Fisheries and Aquatic Science 53:833-841. Shuster, S. M . 1989. Male alternative reproductive strategies in a marine isopod crustacean (Paracerceis sculpta): the use of genetic markers to measure differences in fertilization success among a-, /3-, and 7-males. Evolution 43:1683-1698. Shuster, S. M . & Wade, M . J. 1991. Equal mating success among male reproductive strategies in a marine isopod. Nature 350:608-610. Silverstein, J. T. & Hershberger, W. K . 1992. Precocious maturation in coho salmon (Oncorhynchus kisutch): estimation of the additive genetic variance. Journal of Heredity 83:282-286. 143 Simmons, L. W., Tomkins, J. L. &; Hunt, J. 1999. Sperm competition games played by dimorphic male beetles. Proceedings of the Royal Society of London B 266:145-150. Sinervo, B. & Lively, C. M . 1996. The rock-paper-scissors game and the evolution of alternative male strategies. Nature 380:240-243. Sokal, R. R. & Rohlf, F. J. 1981. Biometry. W. H. Freeman and Company, New York. Stockley, P. 1999. Sperm selection and genetic incompatibility: does relatedness of mates affect male success in sperm competition. Proceedings of the Royal Society of London B 66:1663-1669. Stockley, P., Gage, M . J. G., Parker, G. A . & M0ller, A . P. 1997. Sperm competition in fishes: the evolution of testis size and ejaculate characteristics. American Naturalist 149:933-954. Taborsky, M . 1994. Sneakers, satellites, and helpers: parasitic and cooperative behavior in fish reproduction. Advances in the Study of Behavior 23:1-100. Taborsky, M . 1998. Sperm competition in fish: 'bourgeois' males and parasitic spawning. Trends in Ecology and Evolution 13:222-227. Tautz, A . F. &, Groot, C. 1975. Spawning behavior of chum salmon (Oncorhynchus keta) and rainbow trout (Salmo gairdneri). Journal of the Fisheries Research Board of Canada 32:633-642. Taylor, E. B. , Foote, C. J. & Wood, C. C. 1996. Molecular genetic evidence for parallel life -history evolution within a pacific salmon (sockeye salmon and kokanee, Oncorhynchus nerka). Evolution 50:401-416. Thomas, L. & Juanes, F. 1996. The importance of statistical power analysis: an example from animal behaviour. Animal Behaviour 52:856-859. 144 Thomaz, D., Beall, E . Sz Burke, T. 1997. Alternative reproductive tactics in Atlantic salmon: factors affecting mature parr success. Proceedings of the Royal Society of London B 264:219-226. Thompson, C. E. , Poole, W. R., Matthews, M . A. Sz Ferguson, A . 1998. Comparison, using minisatellite D N A profiling, of secondary male contribution in the fertilisation of wild and ranched Atlantic salmon (Salmo salar) ova. Canadian Journal of Fisheries and Aquatic Science 55:2011-2018. Trippel, E . A . Sz Neilson, J. D. 1992. Fertility and sperm quality of virgin and repeat-spawning Atlantic cod (Gadus morhua) and associated hatching success. Canadian Journal of Fisheries and Aquatic Science 49:2118-2127. Trummel, D. E. , Fulcher, K . D., Beck, J. C. Sz Cloud, J. G. 1992. Fertility of rainbow trout males relative to differences in the proportion of sperm that binds to a specific antibody. Aquaculture 104:175-182. van den Berghe, E. P., Wernerus, F. Sz Warner, R. R. 1989. Female choice and the mating cost of peripheral males. Animal Behaviour 38:875-884. Verrell, P. A . 1983. The influence of the ambient sex ratio and intermale competition on the sexual behavior of the red-spotted newt, Notophthalmus viridescens (Amphibia: Urodela: Salamandridae). Behavioral Ecology and Sociobiology 13:307-313. Warner, R. R., Wernerus, F., Lejeune, P. Sz van den Berghe, E. 1995. Dynamics of female choice for parental care in a fish species where care is facultative. Behavioral Ecology 6:73-81. Weale, M . E. , Whitwell, D., Raison, H . E. , Raymond, D. L. Sz Allan, J . A . 2000. The influence of density of frequency-dependent food selection: a comparison of four experiments with wild birds. Oecologia 124:391-395. 145 West, C. J. 1978. A review of the Babine Lake development project 1961-1977. Fisheries and Marine Service Technical Report Series 812:105 pp. Wilson, E. 0 . 1975. Sociobiology. Belknap Press of Harvard University Press, Cambridge, Mass. Wood, C. C. & Foote, C. J. 1996. Evidence for sympatric genetic divergence of anadromous and nonandromous morphs of sockeye salmon (Oncorhynchus nerka). Evolution 50:1265-1279. Wood, C. C , Rutherford, D. T., Bailey, D. & Jakubowski, M . 1998. Assessment of sockeye salmon production in Babine Lake, British Columbia with forecast for 1998. Canadian Technical Report Fisheries and Aquatic Science 2241:50 p. Yanagimachi, R., Cherr, G. N . , Pillai, M . C. & Baldwin, J. D. 1992. Factors controlling sperm entry into the micropyles of salmonid and herring eggs. Development, Growth and Differentiation 34:447-461. Young, K . A . 1999. Environmental correlates of male life history variation among coho salmon populations from two Oregon coastal basins. Transactions of the American Fisheries Society 128:1-16. Yund, P. O. & McCartney, M . A. 1994. Male reproductive success in sessile invertebrates: competition for fertilizations. Ecology 75:2151-2167. Zheng, W., Strobeck, C. & Stacey, N . 1997. The steroid phermone 4-pregnen-17oj,20/3-diol-3-one increases fertility and paternity in goldfish. Journal of Experimental Biology 200:2833-2840. 146 Appendix A A Model of Fertilization Success in Spawning Events In this model I consider spawning events occurring in the same manner as described qualitatively in the introduction of Chapter 3. In this model, males enter the nest and begin spawning one at a time. Each male that participates in spawning fertilizes fewer eggs than the male(s) that start before him but more than the male(s) that start after him. There are two size/age classes of males and the proportion of eggs fertilized by each individual is determined by a success factor (F) such that the proportion of eggs fertilized by a given male is equal to his success factor divided by the sum of the success factors of all males that participate. The magnitude and units of this success factor are not important as it only represents the potential success of a male relative to other males. The only variables that influence F in this model are age class and spawning order. The large male that begins to spawn first has a success factor Fn and the success factor of each subsequent large male is reduced by a proportion a (where 0 < a < 1) so that F12 = aFn, Ft3 = aFi2, etc. For example, if a = 0.5 then each male's success factor is 50% of the male that started spawning before. Similarly, the small male that starts spawning first has a success factor F s i and success factors of subsequent small males are reduced by a proportion b (where 0 < b < 1) so that Fs2 = bFsi,Fs3 = 6JPS2, etc. The total number of males participating in spawning is N and the proportion comprised by small males is p. The total success factor for large males is: FL = Fll-raFll+a2Fl2 + --- + a«1-p)N-1)Fl{{1_p)N) (l-p)JV = Fn £ a '" 1 (A. l) 147 The total success factor for small males is: FS = Fsl + bFsl + b2Fs2 + --- + b^N^Fs{pN) pN = F ^ c V - 1 (A.2) Given equations A.1 and A.2, we can calculate the following: FL Prop, sired by all large males: L = —— — (A.3) b L + b S F S Prop, sired by all small males: S = -=-- =-= (A.4) * FL + FS Mean prop, sired by individual large males: L = ^ — — ^ (A.5) S Mean prop, sired by individual small males: S = —— (A.6) g Relative success of small males: Sr = y (A.7) Note that relative success can be calculated for large males as well (Lr), but this is just the inverse of Sr. Therefore, only one of these values is necessary to examine FDS. If relative success is calculated when small males are at low frequency (yielding 5 r l ) and when small males are at high frequency (yielding SV2), a frequency-dependence index can be calculated: FBI = | i 148 The value of FDI reveals the magnitude and direction of frequency-dependence where FDI = 1 indicates no frequency-dependence, FDI > 1 indicates negative frequency-dependence, and FDI < 1 indicates positive frequency-dependence. I now use these equations to examine the effect of the number of males on the value of FDI. First, consider a spawning event with one small male and two large males: N = 3, p = l / 3 FL = Fa(l + a) FS = Fsl q _ 2F,,i Drl - Fll(l+a) Next, consider a spawning event with two small males and one large male: JV = 3, p = 2/3 FL = F a FS = Fsl(l + b) q _ F 3l (1+6) Comparing these two outcomes gives the frequency-dependence index: F D I = "t = (TT i^Tft) < A- 9 ) Now consider a greater number of males - two small males and four large males: N = 6, p = l / 3 FL = Fn(l + a + a2 + a 3) F 5 = F s l ( l + cb) c _ 2F„i(l+fr) ° r 3 F(1(l+a+a2-t-a3) • • 149 Consider six males comprised of four small males and two large males: AT = 6 p = 2/3 FL = Fn(l + a) FS = Fsl(l + b + b2 + b3) q _ Fsl(l+b+b2+b3) Jr4 - 2F,i(l+a) The frequency-dependence index with six males present is: ™ = ( i ^ b j < A- 1 0> Equations A.9 and A. 10 can be used to compare FDI between spawnings with six males and with three males for various values of a and b (Fig. A . l ) . Four conclusions can be drawn from theses results: 1. The model resulted in negative frequency-dependence (FDI > 1) for both levels of male number and for all values of a and b. 2. FDI is inversely related to both a and b. 3. The number of males influences the shape of the relationship between FDI and a. 4. FDI increases with the number of males present. 150 Figure A . l : Index of frequency-dependence versus a (reduction in success for large males) in the spawning model. Line 1 is for six males present and b = 0.2. Line 2 is for three males present and b = 0.2. Line 3 is for six males present and 6 = 0.8. Line 4 is for three males present and b = 0.8. 151 Appendix B Tables for Chapter 4 Table B . l : Linear regressions of three morphological measurements versus hypural length. A l l variables were log-transformed for the analyses. Variable Slope Intercept r 2 N F P Body Weight 3.273 -5.382 0.990 174 16390 < 0.0001 Hump size 1.238 -1.456 0.969 175 5433 < 0.0001 Upper jaw length 1.521 -2.130 0.961 175 4231 < 0.0001 Table B.2: Correlations between morphological variables for 174 males. Residuals were calculated as deviations from the reduced major axis regressions. Hypural length Residual weight Residual hump size Residual weight 0.000 Residual hump size 0.007 0.545* Residual upper jaw length 0.005 0.457* 0.543* * P < 0.001 Table B.3: Loadings from principal components analysis of morphological measurements for 174 males. The proportion of variance explained by each component was 0.508, 0.250, 0.136, and 0.106 respectively. Variable P C I PC2 PC3 PC4 Hypural length 0.001 0.999 0.007 -0.014 Residual weight 0.567 -0.002 -0.703 -0.429 Residual hump size 0.598 0.011 -0.007 0.801 Residual upper jaw length 0.566 -0.012 0.711 -0.417 152 Table B.4: Analysis of variance on the number of aggressive acts of alpha males in first and last observations. Observation Source df SS F P First OSR 1 33.492 10.362 0.005 Age composition 1 12.033 3.723 0.072 OSR x composition 1 31.708 9.810 0.006 Error 16 51.717 Last OSR 1 0.210 0.010 0.920 Age composition 1 20.650 1.0150 0.323 OSR x composition* 1 6.710 0.330 0.570 Error 27 549.270 * Removing the interaction term did not influence significance of the main effects. Table B.5: Analysis of variance on the number of courtship behaviours (square-root transformed) of alpha males in first and last observations. Removing the interaction term did not influence significance of main effects in either model. Observation Source df SS F P First OSR 1 0.029 0.047 0.831 Age composition 1 0.371 0.607 0.448 OSR x composition 1 0.179 0.292 0.597 Error 16 9.7984 Last OSR 1 3.076 2.269 0.144 Age composition 1 1.394 1.028 0.320 OSR x composition 1 1.946 1.436 0.241 Error 27 36.602 153 Table B.6: Analysis of variance on the number of satellite intrusions (square-root transformed) in first and last observations. The interaction term was not significant in either model (first: df = 1,SS = 0.207, F = 0.342, P = 0.567; last: df = 1, SS = 0.022, F = 0.034, P = 0.854). Observation Source df SS F P First OSR 1 4.037 6.948 0.017 Age composition 1 1.747 3.006 0.101 Error 17 9.878 Last OSR 1 1.421 2.266 0.143 Age composition 1 2.312 3.688 0.065 Error 28 17.553 Table B.7: A N O V A to test the effects of OSR and age group composition on time to spawn. The interaction between OSR and age composition was not significant and was removed from the model (Fh30 = 1.071, P = 0.309). IfouTce df SS F P' OSR 1 9.906 3.769 0.061 Age composition 1 24.545 9.339 0.005 Error 31 81.474 Table B.8: A N O V A to test the effects of OSR and age group composition on female spawning duration. The interaction between OSR and age composition was not significant and was removed from the model (Fi^o — 0.162, P = 0.691). Source df SS F P OSR 1 1.667 0.548 0.467 Age composition 1 17.761 5.839 0.025 Error 20 63.362 154 Table B.9: A N O V A to test the effects of OSR and age group composition on the time from the start of the female gape to the start of her first dig. The interaction between OSR and age composition was not significant and was removed from the model (1*1,25 = 0.022, P = 0.883). Source df SS F P OSR 1 2.545 0.487 0.492 Age composition 1 12.326 2.358 0.137 Error 26 135.901 Table B.10: A N O V A table for the effects of male age and spawning participation on testes depletion. The analysis was done as a linear mixed-effect model (Pinheiro Sz Bates 2000) with trial number included as a random factor. The analysis included 77 males in 19 trials. Source df F P Male age 1,55 0.203 0.654 Participation 1,55 7.397 0.009 Age x participation 1,55 2.841 0.098 Table B . l l : A N O V A to test for effects of OSR and age group composition on alpha male fertilization success. Alpha male success was arcsine squareroot transformed for the analysis. Source df SS F P OSR 1 0.00004 0.0003 0.987 Age composition 1 0.140 0.865 0.361 OSR x composition 1 1.000 6.161 0.020 Error 26 4.22 155 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0090552/manifest

Comment

Related Items