EFFECTS OF SPAWNING BEHAVIOR AND ENVIRONMENTAL FACTORS
ON ADULT REPRODUCTIVE ECOLOGY AND LARVAL DISPERSAL
OF LAKE STURGEON (ACIPENSER FULVESCENS)
By
Yen Thuy Duong

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment for the requirements
for the degree of
DOCTOR OF PHILOSHOPHY
Fisheries and Wildlife
Ecology, Evolutionary Biology and Behavior

2010

ABSTRACT
EFFECTS OF SPAWNING BEHAVIOR AND ENVIRONMENTAL FACTORS
ON ADULT REPRODUCTIVE ECOLOGY AND LARVAL DISPERSAL
OF LAKE STURGEON (ACIPENSER FULVESCENS)
By
Yen Thuy Duong
Adult reproductive behavior plays an important role in population growth and
genetic structuring, which is less well understood at the individual level for long-lived
fish species like lake sturgeon (Acipenser fulvescens). Using genetic determination of
parentage, I examined the effects of adult spawning behavior and environmental factors
on the species’ mating system, genetic structure within a population and on offspring
survival and timing to dispersal at the individual level. Analyses were based on adults
and larvae captured during 7 years (2001-2007) in the Upper Black River, Black Lake,
Michigan. First, I quantified the degree of temporal variation in aspects of the lake
sturgeon mating system. The species is polygamous, where males and females mate with
a large number of individuals. Mate number and reproductive success (RS, the number of
larvae) varied between sexes. However, the linear positive relationship between RS and
mate number was similar in males and females. Female RS and mate number was
independent of body size though RS and mate number for males slightly increased with
increasing body size. Females spawning at different times during the season and locations
did not differ in RS or mate number. Mating pairs between males and females that arrived
at spawning areas >8 days apart accounted for 30% of total mating pairs averaged across
years, suggesting that temporally discrete spawning groups identified based on direct
observations were not reproductively isolated. Second, I tested for evidence of genetic

differentiation among groups of adults spawning at different times in the season over 9
years (2001-2009). Although significant genetic differences were observed between early
and late groups of adults captured during ≥ 2 years, evidence of isolation by time was
lacking based on low variance in allele frequency (FST) between early and late spawning
groups and lack of correlation between pairwise relatedness and pairwise differences in
spawning time. Third, I estimated effective population size (Ne), effective number of
breeders (Nb) and degrees to which inter-annual variation in Nb was correlated with adult
demographic characteristics. Estimates of Nb and adult census sizes were fairly constant
among years. The Nb/N ratios varied 0.26 - 0.61 among years, consistent with low
2

standardized variance (variance/(mean) ) in RS across years. Ne per generation (132,
95% CI: 104-167) was close to the estimates of annual Nb (harmonic mean: 88). Fourth, I
evaluated the relative importance of environmental factors (water temperature and
discharge) and maternal effects (female body size, location and timing of spawning) on
the embryonic and larval developmental time until dispersal (ELDTUD). Both
environmental variables and maternal effects of female and spawning time were equally
important predictors of ELDTUD. Finally, I evaluated differences in relative larval loss
among females when larvae dispersed a short distance (from two sites 0.15 and 1.5 km
from spawning areas) during two consecutive nights. Larval collections were composed
of unequal proportions of offspring from females but there was no significant difference
in relative larval loss among females. Overall, the study greatly improves understanding
of adult reproductive ecology, the roles of adult spawning behavior and environmental
factors on timing to dispersal and survival of lake sturgeon at critical early life stages.

Copyright by
YEN THUY DUONG
2010

ACKNOWLEDGMENTS
I would like to express my sincere gratitude to those helped me with mentoring
and guidance during my degree. First, I would to thank my advisor, Dr. Kim Scribner.
Thank you for your encouragement and mentoring me, not only on dissertation research
but also on professional development. Without your support and help, my first stage of
moving to new fields of interest would have been much more challenging, I am very
grateful. I would also like to greatly thank my graduate committee, Drs. James Bence,
Daniel Hayes, Gary Mittelbach for their guidance and support.
I am grateful to the Vietnamese government’s “Project 322” for financial support
during four years of my studies. I would also like to thank MSU Graduate School,
Department for Fisheries and Wildlife, and Ecology, Evolutionary Biology and Behavior
Program for supporting additional fellowships and travel grants. My thanks are also to

Cantho University and my colleagues at College of Aquaculture and Fisheries for support
and sharing work at home. I greatly appreciate all people, especially Dr. Ed Baker, who
have been involved in lake sturgeon long-term projects from which I had samples for this
study. I also appreciate my technicians and lab-mates working at Molecular Ecology Lab,
especially Kristi Filcek, Jeannette Kanefsky, Jeannette McGuire, Christen David, James
Crossman, Patrick Forsythe, Hope Draheim, Jared Homola and John Bauman. Your help
and friendship are invaluable to me. During my time here at MSU, I received support
from staff/faculty in Department of Fisheries and Wildlife and friends in a variety of
ways. I cannot list you all, but your care, sharing and friendship to me is unforgettable.
I am deeply thankful to Dr. Chris Wheeler (College of Education) who helped
introduce me to MSU and whose family has always been supportive. My special thanks

v

to Dr. Phu Nguyen (Department of Forestry) and his family for lovely care,
encouragement and help. I felt warm and enjoyed family atmosphere whenever I visited
your home.
Finally I would like to thank my two families. Thank you, my two mothers and
siblings, for your unending love and support. I would also like to whisper my thanks to
my father who passed away six years ago but his love and advice is always in my heart.
Lastly but not the least, I am greatly indebted to my beloved husband and two sons, Ngoc
Son and Ngoc Phuc. My heart were as if broken when I left you. You have sacrificed
during my absence but always give me love and support. You are my life, my love and
my motivation.

vi

TABLE OF CONTENTS
LIST OF TABLES .......................................................................................................... IX
LIST OF FIGURES ........................................................................................................ XI
INTRODUCTION..............................................................................................................1
CHAPTER 1 .......................................................................................................................7
EFFECTIVE MATE NUMBER AND REPRODUCTIVE SUCCESS OF MALE AND
FEMALE LAKE STURGEON IN BLACK LAKE, MICHIGAN
Introduction ..............................................................................................................8
Methods..................................................................................................................12
Results ....................................................................................................................22
Discussion ..............................................................................................................26
CHAPTER 2 .....................................................................................................................46
POPULATION GENETIC STRUCTURE AND RELATEDNESS OF LAKE
STURGEON POPULATION IN BLACK LAKE, MICHIGAN
Introduction ............................................................................................................47
Methods..................................................................................................................50
Results ....................................................................................................................57
Discussion ..............................................................................................................59
CHAPTER 3 .....................................................................................................................70
EFFECTIVE POPULATION SIZE AND TEMPORAL VARIATION IN EFFECTIVE
NUMBER OF BREEDERS FOR THE LAKE STURGEON POPULATION IN BLACK
LAKE, MI
Introduction ............................................................................................................71
Methods..................................................................................................................77
Results ....................................................................................................................87
Discussion ..............................................................................................................88

vii

CHAPTER 4 .....................................................................................................................99
ENVIRONMENTAL AND MATERNAL EFFECTS ON EMBRYONIC AND LARVAL
DEVELOPMENTAL TIME UNTIL DISPERSAL OF LAKE STURGEON
Introduction ..........................................................................................................100
Methods................................................................................................................103
Results ..................................................................................................................113
Discussion ............................................................................................................117
CHAPTER 5 ...................................................................................................................128
RELATIVE LARVAL LOSS AMONG FEMALES DURING DISPERSAL OF LAKE
STURGEON LARVAE
Introduction ..........................................................................................................129
Methods................................................................................................................131
Results ..................................................................................................................137
Discussion ............................................................................................................139
REFERENCES ...............................................................................................................151

viii

LIST OF TABLES

Table 1.1 .......................................................................................................................... 35
Sample sizes of adults captured, adult sex ratios and the number of lake sturgeon larvae
captured and larvae genotyped in 7 consecutive years (2001 – 2007)
Table 1.2 .......................................................................................................................... 35
Summary of parentage assignment estimated for single programs and based on
concordance between two programs Pasos and Cervus
Table 1.3 .......................................................................................................................... 36
Percentage of mating pairs (inferred from parentage) within and between groups of adults
spawning at different times of spawning seasons (2001-2007) based on two ways groups
were classified. Frequency distributions of difference in spawning date between male and
female of a pair were compared between mating pairs and random pairs (all possible pairs
between males and females captured in a given year) using G-test (P-value provided)
Table 1.4 .......................................................................................................................... 36
Number of adults captured, fork length (mean ± SD (range)) and percentage of adults that
were assigned to larvae sampled in 2001 -2007. Difference in Lf/Lm between mating
pairs and random pairs was compared using G-test
Table 2.1 .......................................................................................................................... 66
Pairwise differences in genotypic frequency (Chi-square test, df = 24) and variance in
allele frequency (FST) among groups of individuals that were captured ≥ 2 years during
2001-2009
Table 2.2 .......................................................................................................................... 66
Mean (± SD) of estimated inter-individual relatedness within and between members of
early and late spawning groups based on adults captured ≥ 2 years, as indicated in Figure
2. 1
Table 2.3 .......................................................................................................................... 67
Differences in genotypic frequency (Chi-square test) and variance in allele frequency
(FST) between early and late spawning groups within each of 9 years 2001-2009
Table 2.4 .......................................................................................................................... 67
Results of Mantel tests of correlations between inter-individual relatedness (rxy) and
inter-individual differences in standardized spawning dates between individuals within a
year (2001-2009)

ix

Table 3.1 .......................................................................................................................... 96
Sex ratio and estimated individual reproductive success (RS) including mean, variance
and standardized variance (SDV) of females and males in 7 years 2001-2007
Table 3.2 .......................................................................................................................... 96
Effective number of breeders estimated using linkage disequilibrium (Nb) and
demographic methods (NbV), and the ratios of Nb and adult census size (Nb/N) in 20012007
Table 3.3 .......................................................................................................................... 96
Effective population size (mean and 95% confidence interval, CI) estimated using linkage
disequilibrium method based on adult genotypes from single year and pooled unique
individuals in 7 years (2001 -2007)
Table 4.1 ........................................................................................................................ 123
Mean (± 1 standard deviation) temperature at spawning and daily temperature over the
embryonic and larval developmental time until dispersal (ELDTUD) and cumulative
thermal units (CTU) of three larval groups illustrated in Fig. 4.3.
Table 4.2 ........................................................................................................................ 124
Model structure describing embryonic and larval developmental time until dispersal
(ELDTUD) and model selection criteria based on Akaike Information Criteria (AIC), AIC
difference (∆AIC) and Akaike weight (wi)
Table 4.3 ........................................................................................................................ 125
Coefficient estimates of fixed effects in the best model (model 8 in Table 2) representing
relationships between model parameters and embryonic and larval developmental time
until dispersal (ELDTUD).

x

LIST OF FIGURES
Figure 1.1 ......................................................................................................................... 37
Study site on the Upper Black River Michigan (MI, USA), showing positions of adult lake
sturgeon spawning areas and larval collection sites (a), and an enlarged view of the six
spawning areas (b)
Figure 1.2 ......................................................................................................................... 38
Number of adult males and females captured by day of the spawning season during 9
years (2001 -2009)
Figure 1.2 cont’d .............................................................................................................. 39
Figure 1.3 ......................................................................................................................... 40
Reproductive success (the number of offpsring; top) and effective mate number (EMN;
bottom) of males and females in each of 7 consecutive years (bars represent 1 standard
deviation, SD)
Figure 1.4 ......................................................................................................................... 41
Distribution of the number of offpsring assigned to each individual (a) and effective mate
number (b) of male and female lake sturgeon captured in 2007. Larvae were collected
during the time larvae dispersed approximately 1.5 km downstream from spawnign areas
(Figure 1.1a)
Figure 1.5 ......................................................................................................................... 42
Relationship between number of mates and number of offspring of males and females
captured in 7 years (2001-2007)
Figure 1.6 ......................................................................................................................... 42
Relationship between measures of reproductive success and body size of males and
females after adjusting for year effects
Figure 1.7 ......................................................................................................................... 43
Relationship between reproductive success (left) and the number of mates (right) of
females (expressed as natural logarithm) and standardized spawning dates across years,
after adjusting for year and year-spawning date interaction (slope ± SE for reproductive
success: 0.99 ± 0.53, P= 0.06; for the number of mates: 0.84 ± 0.49, P=0.08).
Figure 1.8 ......................................................................................................................... 43
Reproductive success of females spawning at different times (left) and locations (right,
bars represent the range excluding potential outliers) in 2007
Figure 1.9 ......................................................................................................................... 44
Frequencies of absolute pairwise difference in spawning date of random pairs and mating
pairs in 2001-2007

xi

Figure 1.10 ....................................................................................................................... 45
The ratios of fork length female to fork length male (Lf/Lm) of random pairs and mating
pairs in 2001 (top) and 2005 (bottom)
Figure 2.1 ......................................................................................................................... 68
Standardized spawning date (SSD) of 92 females (top) and 257 males (bottom) that respawned more than 1 time during 2001-2009. Individuals were ordered left to right by
ascending values of mean SSD
Figure 2.4 ......................................................................................................................... 69
Probability density estimation of estimated inter-individual genetic relatedness among
adults from early spawning groups (solid line) and from late spawning groups (dash line)
based on males and females that were captured ≥ 2 years during 2001-2009
Figure 3.1 ......................................................................................................................... 97
Total length (top) and age (bottom) distribution of male (n = 362) and female (n = 249)
lake sturgeon (Age data provided by Baker, unpublished)
Figure 3.2 ......................................................................................................................... 98
Total number of captured adults adjusted for sampling effort and effective number of
breeders estimated using linkage disequilibrium (Nb) (bars show 95% confidence
interval) and demographic methods (NbV) in 2001-2007
Figure 3.3 ......................................................................................................................... 98
Plots showing relationships between the ratio of effective number of breeders to adult
census size (Nb/N) and larval collection (left) and adult sex ratios (right) in 7 years (20012007). The points circled are data from 2002
Figure 4.1 ....................................................................................................................... 126
Number of adults captured (a) and number of larvae collected and those assigned to early,
middle or late female groups (b), together with mean daily water temperature and river
discharge during the periods from egg deposition until larval dispersal.
Figure 4.2 ....................................................................................................................... 127
Embryonic and larval developmental time until dispersal (ELDTUD) of three groups of
larvae whose maternal parents spawned during “early” (circle), “middle” (diamond) or
“late” (triangle ) periods of the spawning season. Darkened symbols indicate more than
one offspring characterized by the same ELDTUD and female spawning date
Figure 4.3 ....................................................................................................................... 127
Mean and standard deviation of embryonic and larval developmental time until dispersal
(ELDTUD) among siblings from the same female parent. Individual females were
ordered left to right by spawning date

xii

Figure 5.1 ....................................................................................................................... 145
The study site in the Upper Black River, Michigan, showing upstream and downstream
adult spawning locations (B and C, respectively); and upstream and downstream larval
collection sites (D1 and D2, respectively)
Figure 5.2 ....................................................................................................................... 146
Numbers of adults captured each day of the 2006 spawning season and total numbers of
offspring collected on two nights (May 28 and 29) that were assigned to captured females
using genetically determined parentage
Figure 5.3 ....................................................................................................................... 147
Hourly pattern of larval dispersal on 2 nights (May 28 and 29). Larval numbers are
cumulative totals over all females spawning at upstream (Site B) and downstream (Site
C) locations (Figure 5.1)
Figure 5.4 ....................................................................................................................... 148
Standardized offspring produced by each female estimated based on collections at two
larval sampling sites on two consecutive sampling nights. Female ID includes reference
to spawning locations (B vs. C) is presented in order of spawning date (May 1-15, Figure
5.2)
Figure 5.5 ....................................................................................................................... 149
Relative larval loss among females (bars represent 1SD between two nights). The order
of females was as the same as Figure 5. 4

xiii

INTRODUCTION

Reproductive variables including body size, spawning time and spawning location
together with environmental factors affect variation in recruitment of fish populations
(Van Winkle et al., 1997). Reproductive variables can vary among individuals within a
spawning group, among groups within a year, and among years. Most studies have
investigated how these variables affect recruitment variation at the population level.
Recently, with the availability of molecular genetic markers, these effects can be
analyzed at the individual level based on data obtained using genetic determination of
parentage (DeWoody and Avise, 2001; Garant and Kruuk, 2005; Pemberton, 2008). For
long-lived species with small population sizes, individual contributions to population
recruitment can affect population levels of genetic diversity and are important to the long
term viability of a population. However, the availability of such information is often
limited. I used genetic approaches, specifically parentage analysis based on microsatellite
loci, to study the effects of environmental factors and reproductive variables on adults’
mate number and reproductive success as well as on offspring’s timing to dispersal in a
threatened fish species, the lake sturgeon Acipenser fulvescens. The study was conducted
in Black Lake, Cheboygan County, Michigan, USA.
Lake sturgeon have an interesting life history characterized by extreme longevity
(potential > 100 years), iteroparity (Auer, 1999) and polygamous mating by both sexes
(Bruch and Binkowski, 2002). Males mature earlier (12-15 years) and spawn at shorter
inter-annual intervals (1-4 years) than females (> 18 years and every 2-7 years,
respectively) (Auer, 1999; Forsythe, 2010), resulting in male-biased sex ratios during the
reproductive period in many populations. Lake sturgeon spawn in groups where one
1

female can be surrounded by several males and one male can mate with multiple females
(Bruch and Binkowski, 2002). The spawning season usually begins at the end of April
and extends through early June during a period characterized by a wide range of
temperatures (9oC - 21oC, Bruch and Binkowski, 2002). Spawning activity occurs in
bimodal or multimodal peaks, typically 7-15 days apart (Forsythe, 2010). Adults have a
tendency to spawn at the same time in the season over multiple years (i.e., spawning time
is repeatable, Forsythe, 2010). Lake sturgeon do not provide parental care post-hatch,
which contributes to extremely high mortality during early life stages (Caroffino et al.,
2010; Forsythe, 2010). Low survival rates together with delayed maturity partly
contribute to low natural recruitment into lake sturgeon adult populations.
Numerous lake sturgeon populations have been severely depleted over the past
100 years throughout the Great Lakes (Houston, 1987; Hay-Chmielewski and Whelan,
1997). Therefore, lake sturgeon has been a species of conservation concern (Holey et al.,
2000; Peterson et al., 2007). Despite decades of conservation actions, abundance of many
lake sturgeon populations has remained low, in part due to anthropogenic factors
resulting in fragmented habitats and degraded water quality (Peterson et al., 2007).
Rehabilitation programs have been impeded by lack of knowledge regarding the species
reproductive ecology (Holey et al., 2000). Although the species has been studied
extensively regarding spawning habitat, population abundance and distribution,
knowledge with regards to factors contributing to behavior and survival during early life
stages and aspects of adult behavioral ecology during reproduction is less well
understood. Information with respect to the species mating system, effective population

2

size and roles of adult spawning behavior to offspring behavioral traits at the individual
level has been limited. These topics are difficult to study by direct observations alone.
Indirect methods, using information of pedigrees based on molecular genetic
markers, have been powerful tools and widely used to study evolutionary biology and
behavioral ecology (Avise, 2004; Garant and Kruuk, 2005; Pemberton, 2008).
Applications include estimation of effective population size, characterization of mating
systems and of parental effects on offspring phenotype, behavior and survival in natural
populations. Parentage analysis and genetic relatedness are important approaches that
facilitate pedigree reconstruction (Pemberton, 2008). Different statistical methods and
software employing multi-locus genotypes have been developed to identify parentage
(review in Jones et al., 2010) and sibship reconstructions (Wang, 2004; Wang, 2007a;
Wang and Santure, 2009a). In my dissertation research, I used genetic approaches based
on microsatellite markers to study the effects of reproductive variables (i.e., body size
and spawning behavior, which included choice of the time and location for spawning)
and environmental factors on adult reproductive ecology and larval dispersal of lake
sturgeon.
My dissertation is in five parts. Chapter 1 examines temporal and spatial variation
in the mating system of lake sturgeon across seven years (2001-2007). A mating system is
characterized by linear regression of the number of offspring (reproductive success) and
the number of mates (Bateman, 1948; Arnold and Duvall, 1994; Levitan and Petersen,
1995). Lake sturgeon have been known to be polygamous based on direct observations
(Bruch and Binkowski, 2002). However, the number of mates and reproductive success
can vary due to spatial and temporal synchrony of spawning, variation in adult

3

abundance, and operational sex ratios (Emlen and Oring, 1977; Arnold and Duvall, 1994;
Levitan and Petersen, 1995), which is known to vary year-by-year in lake sturgeon
populations (Auer, 1999; Forsythe, 2010). I quantified how the relationship of
reproductive success and mate number fluctuated across years. I also tested whether
reproductive success and mate number were independent of body size, spawning time
and spawning location. Findings under this objective are important to further studies
testing possibility of reproductive isolation by time within this lake sturgeon population.
If variation in individual reproductive success was a function of spawning time, which
has been shown to be repeatable (Forsythe, 2010), natural selection acting differently
during breeding seasons could result in reproductive isolation (Hendry et al., 1999;
Hendry and Day, 2005; Tomaiuolo et al., 2007). Estimating reproductive success also
provides parameters to facilitate estimation of effective number of breeding adults.
In chapter 2, I tested if levels of genetic differentiation between early and late
spawning adults existed as predicted based on observations of adult spawning time over 9
years (2001-2009). Forsythe (2010) suggested that the population was composed of
different reproductively isolated groups. This suggestion was based on the observations
that spawning activity was bimodal or multi-modal and that spawning time of individuals
repeatedly captured in two or more years was highly repeatable. Measures of genetic
differences and relatedness among groups spawning at different times within a year and
across years provided a quantitative means of evaluating this hypothesis.
In chapter 3, I estimated effective number of breeders (Nb), effective population
size (Ne), and degrees to which inter-annual variation in Nb correlated with adult
demographic characteristics. Ne and Nb are important parameters describing evolutionary
4

processes of finite populations (Palstra and Ruzzante, 2008; Charlesworth, 2009),
especially small and isolated populations such as the Black Lake lake sturgeon
population. In this population, larval collection, a surrogate measure of total annual larval
production, varied 40-fold across 7 years (2001-2007) but the number of adult was fairly
constant (115-223 individuals). Based on literature that Nb is sensitive to adult census
size (Frankham, 1995; Nunney, 1996) but varies less due to standardized variance (e.g.,
2

variance/(mean) ) in female fecundity (Nunney, 1996) , and given available information
of adults and larvae of the lake sturgeon population in Black Lake, I predicted that Nb and
Nb/N of the population would be relatively stable and would not vary proportionally with
the inter-annual variation of offspring collected at the larval stage. I also estimated
contemporary Ne and compared Ne and Nb. This study is one of only a few empirical
studies showing the relationship of Ne and Nb of a long-lived iteroparous species.
In chapter 4, I examined the relative importance of environmental factors (water
temperature and flow) and maternal effects (female body size, location and timing of
spawning) on embryonic and larval developmental time until dispersal (ELDTUD).
ELDTUD includes three periods: (i) embryonic development, (ii) yolk sac absorption;
and (iii) emergence and dispersal from spawning areas. The first two periods are likely
dependent on water temperature (Wang et al., 1985; Pepin et al., 1997; Kamler, 2002).
The third period may depend more on other environmental factors (e.g., river discharge,
food, and predators), larval age or size (Elliott, 1987; Day and Rowe, 2002) and larval
behavior (Shanks, 2009), lunar cycle effects on larval concealment (Hernandez-Leon
2008), and female spawning behavior (Copp et al., 2002; Hogan and Mora, 2005;
5

Shanks, 2009). If effects of environmental factors on ELDTUD outweighed maternal
effects, the latter would not significantly contribute to the explanation of variation in
ELDTUD. If these effects were relative important, they both would be in the best set of
models explaining ELDTUD.
The last chapter tests a hypothesis that environmental factors randomly affected
offspring survival among families. I examined whether larvae from different females
exposed to the same environmental conditions during dispersal differed in relative levels
of mortality. We estimated proportional contributions of females to larval lake sturgeon
collections and relative larval loss among females as larvae dispersed downstream during
two consecutive nights. Analyses were based on genetically determined parentage. Data
pertaining to inter-family variation in contributions to larval dispersal and the relative
roles of maternal and stream environmental effects on larval survivorship during
dispersal can lead to greater understanding of inter- and intra-annual variation in fish
recruitment.
Findings from this work will greatly improve understanding adult reproductive
ecology, roles of parental behavior (timing and location of spawning) under interactions
with environmental factors on timing to dispersal, and survival of lake sturgeon at critical
early life stages. The information will be also useful for restoration programs for this
threatened species.

6

Chapter 1
MATE NUMBER AND REPRODUCTIVE SUCCESS OF MALE AND FEMALE
LAKE STURGEON IN BLACK LAKE, MICHIGAN

ABSTRACT – Mating systems of polygamous fishes are dynamic and often difficult to
study based on direct observation alone. We used genetic parentage assignment to
examine temporal variation in the reproductive success and effective mate number in lake
sturgeon (Acipenser fulvescens) in a closed population in Black Lake, Michigan. Adults
(N= 611 individuals captured 1,024 times) and larvae (N=3,566) were sampled during the
entire spawning season during each of 7 consecutive years (2001-2007). Reproductive
success (RS) and effective mate number (EMN) differed between sexes and varied
greatly among individuals of each sex across years. However, in all years, we observed a
similar linear positive relationship between RS and EMN for males and females. RS and
EMN were independent of body size in females but RS and EMN of males increased with
increasing body size. No evidence of assortative mating by size was observed. Spawning
time and location did not affect female RS or mate number. Mating pairs between
individuals that arrived at spawning areas >8 days apart accounted for 30% of the total
number of mating pairs on average across years. These data suggest that direct
observations documenting the existence of discrete and temporal isolated spawning
groups inaccurately represent the degree of gene flow among spawning groups.
Collectively, data provide insights regarding factors including male mating behavior that
contribute to inter-individual variance in RS the degree of sub-population genetic
structure, and effective population size of this lake sturgeon population.

7

INTRODUCTION
Polygamous mating occurs in many taxa including mammals, birds, reptiles,
amphibians, fish and insects (Briton et al., 1994; Hernaman and Munday, 2007). Multiple
mating could be adaptive for both females and males (Verner and Willson, 1966). Males
that mate with a large number of females have more offspring than males that mate with
comparatively fewer females (Slagsvold and Lifjeld, 1994; Hasselquist, 1998). Multiple
matings by females may increase genetic diversity and genetic quality of offspring
(Jennions and Petrie, 2000; Neff and Pitcher, 2005). Non-genetic benefits such as
parental care, promoting egg maturation and oviposition, decreasing the risk of predation,
and increasing access to mates thereby increasing probabilities of egg fertilization could
be important factors favoring females that mate with multiple males (Reynolds, 1996;
Jennions and Petrie, 2000). For broadcast-spawning species, sperm limitation is the major
factor affecting female reproductive success (Levitan and Petersen, 1995; Franke et al.,
2002; Marshall and Evans, 2005).
In complex mating systems of polygamous mating fish species (Kokita and
Nakazono, 1998; Avise et al., 2002), inter-individual variation in mate number and
reproductive success (RS) can be affected by many factors. Within a population, the
number of mates can be altered by distribution and abundance of potential mates and
availability of essential resources (e.g., food and spawning locations) (Verner and
Willson, 1966; Emlen and Oring, 1977; Hernaman and Munday, 2007). The distribution
and abundance of one sex can vary over time and space (Shuster and Wade, 2003) due to
differences in timing of arrival between males and females (e.g., longnose filefish,
Oxymonacanthus longirostris (Kokita and Nakazono, 1998); steelhead Oncorhynchus

8

mykiss (Seamons et al., 2004)) or due to differences in spawning intervals between males
and females. As a result, prior to fertilization, temporal and spatial variation in sex ratios
and abundance can influence the number of mates and reproductive success.
During and after fertilization, factors that influence probabilities of egg and
offspring survival also affect adult reproductive success, including parental effects and
environmental conditions (Kamler, 2005). Maternal effects such as phenotypic traits (e.g.
body size) and female age on offspring growth and survival have been documented
(Chambers and Leggett, 1996; Trippel et al., 1997; Heins et al., 2004a; Kamler, 2005).
For example, in numerous fish species where larvae rely on endogenous yolk reserves for
growth and survival immediately after hatch, female body size is positively related to egg
size, larval size and larval resistance to starvation and predation (Kamler, 2005). Paternal
effects also contribute to offspring survival during embryo development stage (Trippel
and Neilson, 1992; Kamler, 2005). For instance, positive relationships between sperm
density and/or motility and egg fertilization rates were reported in several fish species
such as bluehead wrasse Thalassoma bifasciatum (Petersen et al., 2001) and bluegill
sunfish Lepomis macrochirus (Neff et al., 2003). Moreover, choice of spawning time and
spawning locations by females will determine the environmental conditions that eggs and
larvae initially experience (Einum and Fleming, 2000; Hendry and Day, 2005; Jorgensen
et al., 2008), and therefore can affect RS.
In broadcast-spawning species, the number of mates and reproductive success of
males and females indicated by fertilization success are strongly density-dependent
(Moller and Legendre, 2001; Rowe and Hutchings, 2003). When a population is at low
abundance, male RS will likely decrease due to reduced mating opportunities (Allee

9

effects) (Moller and Legendre, 2001; Rowe and Hutchings, 2003; Levitan, 2004). Female
RS is also expected to decrease due to sperm limitation (Levitan and Petersen, 1995;
Levitan, 2004; Marshall and Evans, 2005). At high population densities, female
reproductive success can also decrease due to polyspermy (eggs simultaneously fertilized
by more than one sperm become inviable) as a result of sperm competition (Franke et al.,
2002; Levitan, 2004). Furthermore, environmental factors, especially water current,
strongly influence fertilization rates of broadcast spawning species (Coma and Lasker,
1997; Petersen et al., 2001; Marshall and Evans, 2005). Spawning synchrony and
availability of mates are important for these species to increase fertilization rates and
ultimately reproductive success (Emlen and Oring, 1977; Levitan and Petersen, 1995;
Coma and Lasker, 1997).
Lake sturgeon Acipenser fulvescens is a broadcast-spawning species characterized
by extreme longevity (potential > 100 years), iteroparity (Auer, 1999) and polygamous
mating by both sexes (Bruch and Binkowski, 2002). Observational studies describing
spawning behavior of this species indicated that lake sturgeon spawn in groups where one
female can be surrounded by several males and each male can mate with multiple females
(Bruch and Binkowski, 2002). Spawning activity occurs during the day or night shortly
after ovulating females arrive at spawning sites. During a spawning bout, eggs and sperm
are released simultaneously into the water column. Egg deposition by each female
typically lasts only 8-12 hours (Bruch and Binkowski, 2002). Similar to other broadcastspawning species (DeWoody and Avise, 2001), lake sturgeon do not provide postovulatory parental care (Bruch and Binkowski 2002). Egg and larval lake sturgeon
experience high rates of mortality (e.g., survival rates from eggs to age-0 juvenile stage <

10

0.1%) (Caroffino et al., 2010; Forsythe, 2010). Extremely high mortality may result in
high variation in RS among individuals (Caroffino et al., 2010).
Other aspects of lake sturgeon reproductive ecology can contribute to the
variation in mating behavior over time and space. Males mature earlier (12-15 years) and
exhibit shorter inter-spawning intervals (1-4 years) than females (> 18 years and every 27 years, respectively) (Auer, 1999; Forsythe, 2010). Accordingly, sex ratios of spawning
groups of lake sturgeon are usually male-biased, and sex ratios can vary by year and
among spawning locations within a population (Auer, 1999). In addition, spawning
aggregations of lake sturgeon entering spawning areas during early and late times of the
spawning season correspond to different water temperatures and discharge. Adults
captured during >2 years exhibit high repeatability in spawning time (Forsythe, 2010).
Therefore, differences in spawning adult abundance, operational sex ratios, and
individual composition together with differences in environmental factors during a
spawning season may result in within-year variation in the number of mates and
reproductive success. Differences in spawning adult abundance and sex ratios
characterizing early and late spawning groups may be consistent across years, potentially
resulting in predictable variation in the number of mates and reproductive success across
years.
Mating systems of fish species are difficult to study using direct observations
alone, especially in species without parental care (DeWoody and Avise, 2001; Avise et
al., 2002; Serbezov et al., 2010), like lake sturgeon. DNA-based parentage inferences
provide an opportunity to increase understanding of reproductive strategies and aspects of
the mating system of fishes (DeWoody and Avise, 2001; Avise et al., 2002). Different

11

statistical methods employing multi-locus genotypes have been developed to provide
inferences about maternity, paternity (DeWoody and Avise, 2001; Jones et al., 2010) or
sibship reconstruction (Wang, 2004, 2007b; Wang and Santure, 2009b).
Our first objective was to estimate and compare variance in male and female RS
(the number of larvae assigned to each individual) across years and determine whether
RS was positively related to effective mate number, EMN (the number of mates which an
individual was assigned the same offspring with). The second objective was to examine
whether male and female RS and EMN varied as a function of body size, spawning time,
and spawning location. The magnitude of inter-individual variation in RS would provide
evidence for reinforcement of reproductive isolation among early and late spawning
groups (Hendry et al., 1999; Hendry and Day, 2005; Tomaiuolo et al., 2007). The last
objective was to test whether lake sturgeon exhibited random mating with regards to
body size and capture time.
METHODS
Study site
Our study was conducted in the Upper Black River (UBR), the largest tributary to
o

o

Black Lake, Michigan, USA (latitude 45 43’N, longitude 84 15’W; Figure 1.1a) and
only location used for spawning by lake sturgeon in the drainage. The lake sturgeon
population in Black Lake is isolated from other populations in adjacent lakes by dams
blocking immigration and emigration from Lake Huron (Smith and Baker, 2005). Adults
spawn over a 1.5 km section of the UBR. This section can be divided into six locations of
spawning activity that were used across years (Figure 1.1b; Forsythe, 2010). Shallow
spawning areas (~1 m in depth) and low turbidity allowed most adults to be observed and
12

captured (Crossman, 2008; Forsythe, 2010) and larvae dispersing from all spawning areas
to be collected (Smith and Baker, 2005).
Sample collection
We collected adults and larvae throughout the spawning season during 7
consecutive years (2001-2007). Sampling for adults was conducted by wading the entire
length of the stream encompassing all spawning sites one or more times per day during
the entire spawning seasons. We captured spawning adults using long-handled nets. A
2

dorsal fin clip (~ 1 cm ) was taken for genetic analysis. Sex of adults was determined by
extruding gametes, and all individuals were measured for weight (kg) and fork length
(cm). We also recorded date and location of capture. Based on observations that lake
sturgeon females spent only a few days on the spawning grounds (Forsythe, 2010), date
and location of capture were assumed to be the date and location of spawning for
females. For males, spawning behavior varied in response to the number and location of
spawning females (Forsythe 2010). Accordingly, the time and location of male capture
were not considered to accurately represent the time and location of spawning.
Larval sampling was conducted each year starting approximately 10 days after the
first spawning event and lasted for 25-40 days, until no larvae were captured for 2
consecutive nights. Larval sampling was conducted at night (2100 to 0200 hrs) when the
vast majority of larvae disperse (Auer and Baker, 2002). The sampling site was
approximately 2 km downstream from the spawning areas (Figure 1.1a). Five D-frame
larval nets were evenly spaced across the river channel (description in Smith and King,
2005) and were checked hourly from 2100-0200 hrs. Net locations remained consistent
throughout the sampling period. In some years (2005 - 2007), larvae collected were

13

transferred to a streamside hatchery where they were reared for several months. Before
the fish were released, a fin clip was collected from each individual for genetic analysis.
Mortalities during the rearing periods were also preserved in ethanol 95%. Samples used
for microsatellite genotyping were selected proportionally by capture night from the total
number of preserved fin clips and mortalities of each year. The proportion of the total
number of larvae captured for microsatellite genotyping varied across years (Table 1.1).
However, during each year the larvae genotyped represented proportionally a random
subset from the collection of entire larval collection for a given year.
Genetic analysis
DNA was extracted from fin clips using the QIAGEN DNeasy(R) kit (QIA Inc.).
DNA concentration was measured using a Nanodrop 1000 spectrophotometer (Thermo
Fisher Scientific Inc.). All samples were genotyped at 12 tetra-nucleotide microsatellite
loci including Spl 120 (McQuown et al., 2000); AfuG 68B (McQuown et al., 2002).
2002); Aox 27 (King et al., 2001); AfuG 68, AfuG 9; AfuG 63, AfuG 74, AfuG 112,
AfuG 56, AfuG 160, AfuG 195 and AfuG 204 (Welsh et al., 2003). Using polymerase
chain reaction (PCR), 100 ng DNA was amplified in 25 µl reaction mixtures containing
2.5 µl of 10 X PCR buffer (1 M tris-HCl, 1.5 M MgCl2, 1 M KCl, 10% gelatin, 10% NP40, and 10% triton X); additions of 1 mM MgCl2 (1.5 mM MgCl2 for AfuG 9) for all
reactions but no additional MgCl2 for AfuG 63, Aox 27 and AfuG 74; 0.8 mM deoxynucleotide-triphosphate (dNTP); 10pmol fluorescently labeled forward and unlabeled
reverse primers and 0.5 U of Taq polymerase.
All PCR reactions were conducted using a Robocyler 96 thermal cycler
o

(Stratogene). The PCR conditions were 94 C for 2 minutes, followed by 30 cycles of 1
14

o

minute for primer-specific annealing temperatures (48 C for AfuG 9, AfuG 63 and AfuG
o

o

o

o

112; 50 C for AfuG 74; 53 C for Aox 27; 56 C for AfuG 68 and AfuG 68B; 58 C for
o

o

AfuG 56, AfuG 160 and AfuG 195; and 62 C for Spl 120 and AfuG 204), 72 C for 1
o

minute, and the final extension for 2.5 minutes at 72 C. PCR products were run on 6%
denaturing polyacrylamide gels and visualized using a Hitachi FMBIO II scanner. Allele
sizes were determined independently by two experienced personnel based on
commercially available standards (MapMarker

TM

, Bioventures Inc.) and samples of

known genotype. Errors in genotyping were empirically checked by blindly regenotyping a random 10% of all samples. Reported genotyping error is the ratio between
observed number of allelic scoring differences and total number of alleles compared
(Bonin et al., 2004).
Parentage analysis
There are numerous programs that utilize multi-locus genetic data to estimate
parentage (e.g., CERVUS (Kalinowski et al., 2007); PASOS (Duchesne et al., 2005);
COLONY (Jones and Wang, 2010)), and the efficacy of use of different programs or
combinations of programs has been widely debated (Christie, 2010; Jones et al., 2010;
Walling et al., 2010). Use of multiple programs which are based on different statistical
properties to determine parentage has been advocated (Lee, 2008; Jones et al., 2010). We
used complimentary aspects of two programs, (1) the Parentage Allocation of Singles on
Open Systems (PASOS) program, version 1.0 (Duchesne et al., 2005) and (2) CERVUS
version 3.0 (Kalinowski et al., 2007) to conduct parentage analysis. Output of putative

15

parent-offspring allocations from the two programs were jointly used to increase parental
assignment accuracy.
We used program PASOS to estimate the proportion of adults captured and
parentage allocation correctness. PASOS can detect missing parents when they have not
been collected based on multi-locus genotypes of both parents and offspring. We set the
maximum number offset tolerance (MOT), the maximum number of offsets between a
parental and an offspring allele that PASOS accepts as possibly due to a scoring error, of
2 and the error model (0.002, 0.008, 0.98, 0.008, 0.002 – this is the assumed transmission
probabilities from a parental allele to an offspring allele with –2, –1, 0, +1, + 2 offsets
due to scoring errors) for simulations in PASOS (Duchesne et al., 2005). We conducted
simulations over 5 iterations of 1000 pseudo-offspring to estimate the allocation
correctness. Under the same MOT (MOT = 2) as conducting simulations, we used the
allocation function in PASOS to assign each offspring to two collected parents.
CERVUS is the most commonly used categorical parentage technique (Christie,
2010; Jones et al., 2010) and assigns progeny to a particular non-excluded parent based
on likelihood scores (LOD scores obtained by taking the natural log of the ratio of
likelihoods) derived from the genotypes of all offspring and parents sampled (Marshall et
al., 1998; Jones and Ardren, 2003; Kalinowski et al., 2007). Simulations conducted
within CERVUS to determine power and likelihood of assignment were based on
empirical estimates of population allele frequency, the proportion of adults captured
(PASOS output) and an empirical estimate of genotyping error (1.04%). Most likely
candidate parent pairs of offspring from the assignment output were accepted as “true
parent pairs” based on the criteria of ≥ 70% trio confidence with zero or 1 mismatch

16

between parents and offspring genotypes. Assignment outputs from the two programs
based on criteria above were then compared. Male parent-female parent-offspring triplets
that were consistently assigned by both programs were used in further analysis.
The allocation correctness (the probability that each allocation is correct) from
PASOS was almost constant across 7 years, ranging between 80 – 83%, and standard
deviation among iterations simulated was small (~1%). Meanwhile, CERVUS program
provides a confidence level for each allocation based on user-defined confidence levels
for cut off values. In order to choose a cut off value for 7- year data, we compared the
concordance in assignment between PASOS and three levels of confidence levels in
CERVUS including “the most likely”, 80 and 95% using parent and larval genotypes in
2001 as an example. The result showed the “most likely” confidence provided the highest
assignment rate and the percentage of assignment concordance (between PASOS and
CERVUS) was as high as the 95% confidence. Therefore, we used “the most likely”
criterion for other year data sets.
Statistical analysis
Effects of body size, spawning time and spawning location on reproductive success (RS)
and effective mate number (EMN)
We first tested whether RS and EMN were different between males and females
using ANOVA tests. Relative magnitude of variance compared to the mean of RS
between males and females was evaluated based on standardized variance
2

(variance/(mean) ; Nunny 1996). We used a Wilcoxon rank sum test to compare
standardized variance between males and females in 7 years. If RS and EMN differed
between sexes, further analysis was conducted separately for males and females except

17

for the test of relationship between RS and EMN (below). We used linear mixed models
and linear models to examine the relationship between RS and EMN and the relationship
of each of these variables with body size (i.e., fork length, FL), spawning time, and
spawning location for females. For males, effects of spawning time and location on RS
and EMN were not tested because of male behavior of moving different locations and
spending long periods of time in the river (Forsythe 2010).
In the models examining factors contributing to RS and EMN, the effect of years
in which sample sizes and parentage assignment rates were different was taken into
account and treated as a fixed effect. Adult individuals were treated as a random effect
because the same individual could be repeatedly sampled in several years. However,
random effect of repeated spawning individuals was not significant and therefore
individual was removed from all models that describe the relationship of RS (and the
number of mates) and body size, spawning time, spawning location and the number of
times of repeated spawning. Generally, models were as follows:
Yiyj = µ + αy + γXiyj + βyXiyj + εiyj

(1)
th

where Yiyj is the response variable (RS or EMN) of the i individual in year y and at the
th

value j of another predictor; µ is the population mean (overall intercept) of a response
variable; αy is offset to overall intercept due to year, γ is the overall slope, βy is the offset
to the slope for a particular year.
The standardized spawning date (SSD) of an adult in a given year is the number of days
its first capture was from the first day of the spawning season divided by the total
duration of spawning activity in that year. SSD accounts for different durations of

18

spawning activity across years (0 < SSD ≤ 1). Spawning locations were classified into 6
zones (Figure 1.1a; Forsythe 2010). All spawning zones were not used by adults in every
year. Because lake sturgeon had a tendency to use upstream zones at early spawning
season and downstream at late season (Forsythe, 2010), spawning location was treated as
a continuous variable to evaluate the relative effects of upstream compared to
downstream spawning location on the number of mates and RS. Effects of spawning
locations on RS and EMN could differ during a spawning season; therefore, we also
examined whether the slopes of the interaction between spawning date and spawning
differed among years (3-way interaction of year-spawning date-spawning location was
added in (1)).
We also tested whether the magnitude of difference in RS and EMN among
individuals captured during >2 years was consistent across year (2001-2007). We used
likelihood ratio tests (LRT) to compare the better of fit to the data between two models,
(i) a random intercept model:
Yiy = µ + βyXy + δi+ εiy

(2)

and (ii) a random slope model:
Yiy = µ + (βy + γi)Xy + εiy

(3)
th

where Yiy is the response variable (RS or EMN) of the i individual in year y; µ is the
2

overall intercept of a response variable; δ (in model 2) is random effect (δ ~ N(0, σ δ)) of
individuals that were captured during >2 years. βy is the overall slope relating RS (or
EMN) with year (Xy). In model (3), this slope could vary by the normally distributed
2

random effect γ (γ ~ N(0, σ γ )) of individuals captured during >2 years.
19

The better fit of model (3) compared to model (2) to the data would indicate
inconsistency of differences in reproductive success among individuals across years.
Values of RS and the number of mates have skewed distributions that violate an
assumption of normal distribution of model residuals and therefore can influence
inferences on linear relationship between two variables. We transformed the response
variables (Yln) by taking natural log of RS and the number of mates (Yln = Ln(Y) + 0.5).
Residuals of models with log-transformed data were nearly normally distributed. Most
analyses were consistent between transformed and original data. When the results were
different between transformed and original data, the significance of a predictor was
reported based on log-transformed data.
The relationship between RS and EMN of lake sturgeon as well as other
polygamous species was predicted to be positively correlated (Arnold and Duvall, 1994;
Andersson and Iwasa, 1996). We tested whether the slopes of this relationship differed
between females and males using the data set of two sexes combined and individuals with
zero offspring excluded. Similar to other analyses mentioned above, year effect was
included in the model:
RSiys = µ + αy + βs + γMiys + ηyMiys + δsMiys +λysMiys + εiys

(4)

where i is used as subscript for individual, y for year, and s for sex. RSiys is reproductive
th

success of the i individual from year y and sex s and Miysis the number of mates for that
individual. µ is the overall intercept, α is offset to overall intercept due to year, β is
offset to intercept due to sex, γ is the overall slope, η is the offset to the slope for a

20

particular year, δ is the offset to the slope for a particular sex, and λ is the offset to the
slope for the particular year and sex combination.
Testing assortative mating by size and by spawning time
We tested whether lake sturgeon had a tendency to choose mates of comparable
size (evidence for assortative mating by size) or closer in spawning time (evidence of
assortative mating by time) compared to random pairs. Random pairs included all
possible pairs between males and females captured each year (e.g., in 2001, there were
3,124 pairs from 71 males and 44 females). To test the null hypothesis of random mating
with respect to body size, we first computed the ratio of fork length between female and
male of each pair (Lf/Lm). We then compared the observed frequency distribution of
Lf/Lm from mating pairs inferred from parentage assignment with the null frequency
distribution of random pairs, using a G-test. We also tested the correlation of Lf/Lm and
reproductive success of males and females using Pearson correlations.
We also used a G-test to test other hypotheses of random mating with respect to
spawning time by comparing the observed frequency distributions of (absolute)
differences in capture dates between females and males of mating pairs (identified from
parentage assignment) with those from random pairs. In addition, we estimated the
proportion of mating between and within groups of adults based on their spawning dates.
We classified two individuals of a mating pair from within or between groups using two
ways. First, based on spawning intervals among groups within a year (Figure 1.2), an
individual was assigned as an “early” or “late” spawner. A mating pair was from a
“within group” if males and females were captured at similar times (e.g., early-early or

21

late-late). Secondly, based on differences in spawning dates between mates, we grouped a
mating pair as “within group” if absolute difference in spawning date < 8 days. In 20012005, it was clear that two individuals spawning at 8 days apart were from different
groups. In 2006-2007, although spawning activity was longer and comprised of less
discrete groups, we also used the same criterion (shorter or longer than 8 days difference
between two partners) as in the other years.
Because the time and location of the first capture of an individual was considered
to represent spawning time and spawning location and given our observations that
females spent few days at the spawning ground while male may remain for longer
periods, differences in spawning dates between female and males were based on
plasticity in male behavior.
RESULTS
Parentage assignment
Proportions of larvae assigned to captured adults from program CERVUS each
year (69.9% - 91.2%) were comparable to larvae assigned from PASOS (64.3 -82.3)
across 7 years (Table 1.2). Concordance in parentage assignment between the two
programs ranged from 69.2% to 85.2% of larvae assigned based on a single program
across 7 years. The percentage of larvae genotyped assigned concordantly from the two
programs in 7 years was 46.5% to 67.6%. A high proportion of captured males (57-94%)
and captured females (83-98%) was assigned offspring based on the two programs (Table
1.2).
Individual reproductive success and effective mate number across years
The number of offspring assigned (reproductive success, RS) to females (mean
±SE: 6.27 ± 0.94 across years) was significantly higher than that of males (mean ±SE:
22

2.75 ± 0.33) (F1, 1017 = 123, P <0.001, Figure 1.3, top). However, standardized variance
of RS was not significantly different between males and females (Wilcoxon rank sum
test, n= 7, P =0.8). Within a year and for each sex, RS and EMN were highly skewed and
varied widely among individuals as shown for 2007 (Figure 1.4).
Reproductive success for males and females varied across years (Figure 1.3, top).
Effects of “year” explained 23.6% variation in RS for females and 25.6% for males. We
evaluated relative success among adults captured more than 1 year to test whether adults
that had higher RS in a given year also had more offspring in other years. The interaction
between years and individuals resulted in better fit of model 3 compared to model 2
(LRT, df = 27, P <0.001), indicating that the difference in RS among individuals that
were captured during >2 years was not consistent across years. Results were comparable
for males and females.
Similar to results of variation in RS, variation in EMN was mainly explained by
the effect of year (30.3% and 33.4% for females and males, respectively, Figure 1.3,
bottom). The significant interaction between individuals and years (LRT, df = 27, P
<0.001) indicated that differences in EMN among individuals were not consistent across
years.
Relationship between reproductive success and effective mate number
Reproductive success of males and females was highly positively related to
effective mate number (P<0.001). There was no significant interaction of sex and EMN
on RS (F1,1003 = 0.1, P = 0.76), indicating the relationship of RS and EMN for males and
females was similar (slope ±SE: 1.38 ± 0.08 for female, 1.30 ± 0.08 for males) but
intercepts differed. For each sex, the slope of RS and EMN relationship varied by year
23

(Mate-Year interactions: F6, 302 = 14.9, P<0.001 for females, F6, 695 = 4.8, P<0.001 for
males). EMN, year and their interaction together explained 93.1% and 83.9% variation in
RS, in which EMN alone explained 90.2 and 82.6% variation in RS of females and males
respectively (Figure 1.5), indicating a strong relationship between RS and EMN.
Relationships of reproductive success and effective mate number and body size,
spawning time and spawning location
Relationships between EMN and other variables including body size, spawning
time and spawning location were similar to relationships of RS with these variables due
to EMN and RS were highly related. Effects of body size on RS and EMN did not vary
significantly among years in either sex (e.g., year-fork length interaction for female RS:
F6,302 = 0.7, P =0.7; for male RS: F6, 695 =1.7, P = 0.1). Larger adults had slightly higher
RS (Figure 1.6) but this relationship was not significant in females (both original and
transformed data, P > 0.11) but was significant for males (slope ± SE = 0.016 ± 0.009,
t1, 701 = 1.85, P =0.06, transformed data P <0.01).
Similarly, larger males had higher EMN (slope ± SE = 0.017 ± 0.006, t1, 701 =2.5,
P =0.02, transformed data P <0.01) than smaller males. For females, the positive
relationship of EMN and body size was not significant (slope ± SE = 0.04 ± 0.03,
t1, 308 =1.54, P = 0.12).
Female RS was not significantly related to either the 3-way interaction of
spawning time, spawning location and year (transformed data, F6, 287 =1.1, P=0.37) or the
interaction between spawning time and location (F1, 287 =2.8, P=0.09). Neither was the
interaction predictive of EMN. There were no general patterns of female RS and EMN
24

during the spawning seasons and among locations (main effects of spawning time and
location, all tests P > 0.10). Although the effects of spawning time and location on female
RS and EMN varied among year (e.g., year interaction with spawning time on EMN:
F6, 302 =6.4, P < 0.01 and with spawning location F6, 302=3.5, P < 0.01), RS and EMN
were not significantly different among females spawning at different times (Figure 1.7)
or locations in most years. Except in 2007, females spawning early and at upstream
locations had higher RS and EMN than females spawning late in the season (for both RS
and EMN: F1, 61 ~ 20, P<0.01) and at downstream (F1, 61 ~ 7, P<0.01), respectively
(Figure 1.8).
Assortative mating by captured time
To test whether adults were more likely to mate with individuals in the same
spawning groups relative to different groups, we compared the observed frequency
distribution of (absolute) differences in capture dates between partners of mating pairs
(inferred from parentage) to that of random pairs (random combination of males and
females captured). The results revealed that the two frequency distributions were
significantly different in 5 of 7 years studied (G-test, P <0.01, Table 1.3). Differences in
capture dates between males and females of mating pairs ranged 0 – 24 days (Figure 1.9).
Based on classifications of males and females into groups assigned based on date of
capture, mating occurred between individuals of the same within spawning groups on
average 69.4% of the time (range 53.6% – 92.3% over 7 years), and mating among
individuals from different spawning groups occurred on average 30.6% of the total
mating pairs each year (Table 1.3).

25

Assortative mating by size
Females are usually larger than males (Table 1.4). Therefore, the ratio of fork
length between females and males of mating pairs were greater than 1, ranging 1.12 –
1.23. The observed frequency distributions of Lf/Lm for mating pairs was not
significantly different from expected by chance in 5 of 7 years (G-test, Table 1.4),
indicating that there was no trend of nonrandom mating with respect to body size. In
2001 and 2005, where Lf/Lm of mating pairs was significantly higher than expected by
chance (P < 0.01 and P = 0.012, respectively; Figure 1.10), we tested whether this
difference was related to reproductive success of males and females. However, individual
reproductive success did not vary as a function of body size in either males (F1, 69 = 0.89,
P = 0.34 in 2001, and F1, 104 = 0.50, P = 0.48 in 2005) or females (F1, 42 = 0.003, P = 0.95
in 2001, and F1, 45 = 2.65, P = 0.11 in 2005), and thus was not correlated with Lf/Lm
(r = -0.002, df = 256, P = 0.97 for females, r = 0.02, P = 0.74 for males in 2001; r = 0.10,
df = 53, P = 0.47 for females, r = -0.07, P = 0.61 for males in 2005). Given the lack of
correlation between observed Lf/Lm of mating pairs and individual RS, differences in
frequencies of Lf/Lm between mating pairs random pairs did not provide evidence for
assortative mating.
DISCUSSION
Polygamous mating of lake sturgeon can vary temporally and spatially within a
population. Using DNA-based inference of parentage based on data from 7 years, we
could quantify the variation of and factors affecting reproductive success and mate
number. Our findings showed a similar strong positive relationship between RS and mate
26

number for males and females. Within each sex, the most reproductively successful
individuals in one year were not consistently more successful than other individuals. No
consistent patterns emerged suggesting that RS and mate number of individuals spawning
early in the season and at upstream locations differ from individuals spawning late in the
season and at downstream locations. We also found that mating could occur between
individuals from different spawning groups, indicating the Black Lake population was
not composed of reproductively isolated groups as previously thoughts (Forsythe, 2010).
Relationship between reproductive success and effective mate number
The relationship between RS and EMN in polygamous species was predicted to
be positively related in both sexes because of low reproductive investment (Arnold and
Duvall, 1994; Andersson and Iwasa, 1996). This prediction was observed in lake
sturgeon. For species with external fertilization like lake sturgeon, sperm availability
affects egg fertilization and thus greatly contributes to female reproductive success
(Levitan and Petersen, 1995; Franke et al., 2002; Levitan et al., 2004). Lake sturgeon
females release eggs as sperm is released from multiple males (Bruch and Binkowski,
2002). Females which spawn when surrounded by a large number of males are likely to
realize higher fertilization rates than females spawning in the presence of fewer males.
Meanwhile, males increased their reproductive success by mating with as many females
as possible (Levitan et al., 2004).
Based on Bateman’s principle (Bateman 1948), larger variance in RS and the
stronger relationship between RS and mate numbers would be expected in the sex that
invests less energy into reproduction compared to the other sex (Bateman, 1948; Arnold
and Duvall, 1994; Andersson and Iwasa, 1996). In lake sturgeon, we did not observe

27

different slopes of RS on EMN between males and females, which was consistent with
similar standardized variance of reproductive success between the sexes.
Factors affecting RS and EMN
Spawning synchrony and availability of mates in terms of abundance and
operational sex ratios are important factors contributing to fertilization success and thus
to the number of mates and reproductive success in both sexes of lake sturgeon as well as
other broadcast spawning species (Emlen and Oring, 1977; Arnold and Duvall, 1994;
Levitan and Petersen, 1995). Therefore, spawning time and spawning location could be
related to RS and EMN. Lake sturgeon spawning early in the season were usually more
numerous and spawned over longer periods compared to individuals spawning late in the
season (Figure 1.2). RS and EMN were expected to differ by spawning time and
spawning location. Specifically, RS and EMN were predicted to be higher for individuals
spawning early in the season and at upstream locations compared to individuals spawning
late in the season and downstream locations. This trend was observed only in 2007 but
was not consistent in other years.
In addition to differences in the number of adults and the length of spawning
periods between early and late portions of the season, environmental factors associated
with spawning time and spawning location of adults can affect the probability of survival
of fertilized eggs to larval stages, and thus could contribute to variation in adult RS and
EMN at different spawning locations and during a spawning season. Environmental
factors that can cause mortality of lake sturgeon in these early stages include biotic
factors such as predation, microbial infection, food availability, etc., and abiotic factors,

28

for example, water temperature and flow (direct or indirect effects via oxygen supply)
(Kempinger, 1988; Kamler, 1992; Caroffino et al., 2010; Forsythe, 2010).
Temperature and water flow in the Upper Black River, where lake sturgeon
spawn, follow a seasonal pattern in which lower temperature and higher flow occur early
in the spawning season compared to later in the season (Forsythe, 2010). Seasonal
patterns of the number of adults and two important environmental factors (temperature
and flow) was not reflected in variation in RS and mate numbers during the spawning
season, implying that RS and EMN likely result from environmental stochasticity, as
observed in other species (e.g, brown trout Salmo trutta (Serbezov et al., 2010).
Adult body size is another attribute that was predicted to contribute to variation in
RS and EMN. Similar to other fish species, body size and fecundity of lake sturgeon
females are positively related (Bruch et al., 2006). But high reproductive potential may
not alone predict high RS. We found that RS and EMN of lake sturgeon females were
independent of body size. A weak positive relationship between RS and body size was
observed in lake sturgeon males. There could be two possible explanations for this result.
First, larger males might have higher sperm quantity or/and quality (e.g., guppy Poecilia
reticulata (Skinner and Watt, 2007); lake whitefish Coregonus clupeaformis (Blukacz et
al., 2010),). Alternatively, larger males might compete better for mates, leading to higher
EMN and RS (e.g., in Atlantic cod Gadus morhua (Rowe et al., 2007); leopard gouper
Mycteroperca rosacea (Erisman et al., 2007)). We did not have evidence to distinguish
these or other possiblities. In addition, because the relationship of body size and RS as
well as EMN was weak in males and females given a large sample size across years, size
was not a dominat factor in determining RS.

29

However, based on spawning behavior of lake sturgeon and findings from other
similar group-spawning species we believe that male competition might be present
among lake sturgeon males. For such species, larger males often have a better ability to
position themselves in proximity to females during spawning, and thus may increase
EMN and also RS (Petersen and Warner, 1998; Bekkevold et al., 2002; Erisman et al.,
2007). In leopard groupers (Mycteroperca rosacea) for example, male - male competition
occurs, where dominant males who occupy the closest position to females can fertilize
more eggs and therefore increase higher reproductive success than peripheral males
(Erisman et al., 2007). In addition, male-biased of spawning groups observed in lake
sturgeon (Figure 1.2) suggested an opportunity for male competition (Sadovy et al.,
1994; Hall and Hanlon, 2002).
Assortative mating by size and by capture time
Weak assortative mating by size occurs commonly in many species (Blachford
and Agrawal, 2006). For broadcast spawning species, evidence of both nonrandom
mating (e.g., Atlantic cod Gadus morhua (Rowe et al., 2007)) and random mating (e.g.,
steelhead trout Oncorhynchus mykiss (Seamons et al., 2004) with respect to body size has
been reported. For lake sturgeon, we did not find correlations between male and female
body size for mated pairs. Although in two years (2001 and 2005), the ratios of female
fork length to male fork length (Lf/Lm) significantly differed between mating pairs
(inferred from parentage) and random pairs (all possible pairs of males and females
captured in a given years), the Lf/Lm from mating pairs was not correlated with male and
female reproductive success, indicating no evidence of assortative mating to increase
reproductive success. The difference in Lf/Lm between mating pairs and random pairs in
30

2001 and 2005 was not due to sex ratios or distribution of fork length of males and
females because sex ratios (Table 1.1) fork length of adults (Table 1.4) and in 2001 and
2005 were similar with the other years.
Mating among individuals might be not random with respect to arrival time,
especially for broadcast spawning species where spawning synchrony is an important
determinant of RS and mate number (Emlen and Oring, 1977; Levitan and Petersen,
1995). We found that for lake sturgeon, the majority (~70%) of mating pairs were from
males and females that were captured less than 8 days apart. Because ovipision of female
typically lasts for only 8 – 12 hours (Bruch and Binkowski, 2002) or a few days (by
observation of the time most females stayed in the spawning ground, Forsythe, 2010),
mating between members from different groups is likely due to male behavior. The date
of first capture was thus not representative of the spawning date for males. Males were
frequently observed to remain at the spawning areas over prolonged periods, potentially
increasing opportunities to mate with females arriving later in the season. We observed
that 14.5% of males were recaptured in 2007, 7-16 days following the first capture. In
other lake sturgeon populations, the same male behavior of prolonged occupancy of
spawning areas facilitating mating opportunities with multiple females was also observed
(Bruch and Binkowski, 2002).
Quantification of RS and EMN in this study was based on genetically identified
parentage at the larval stage and different sample sizes across years. Thus, the slopes of
RS on EMN may depend on sample sizes and life stages of offspring. Survival of
offspring from larvae to later stages (e.g., juvenile, (Nichols et al., 2003; Caroffino et al.,
2010)) can be very low, resulting in fewer offspring per family and EMN can be detected.

31

We acknowledged that the larger of sample sizes, the more offspring would be assigned
and therefore the higher EMN would be detected (Figure 1.3). Accordingly, slopes and
intercepts of the linear regression of RS on EMN for each sex would vary with different
sample sizes. However, the differences in slopes of RS on EMN between two sexes,
which imply levels of reproductive investment (Bateman, 1948; Arnold and Duvall,
1994; Andersson and Iwasa, 1996), were more interested. The relationship between RS
and EMN for males and females was strong and the same slopes for both sexes were
consistent even though sample size varied across years, indicating that estimates of RS
and EMN and their inter-relationship in males and females was not influenced by sample
sizes. Because the number of unassigned larvae varied across years (32 – 53%), results
could be biased. However, based on more relaxed criteria of parentage assignment in
which far greater proportions of larvae were assigned (60-80% across years), the general
results did not change (data not shown). Thus, we felt that the proportion of unassigned
larvae did not affect our conclusion.
Implications
Components of the mating system such as mate number, reproductive success,
and mating behavior play an important role in population growth (Rowe and Hutchings,
2003) and population genetic structure (Johannesen and Lubin, 1999; Fagan et al., 2010).
Two important implications can be drawn from our findings quantifying aspects of the
lake sturgeon mating system. First, RS among individuals was highly skewed, which
likely resulted in lower effective population size compared to population census size. The
magnitude of differences in RS among individuals captured during >2 years, however,
was inconsistent across years, indicating that cumulative differences among individuals

32

would be decrease disproportionally with mean lifetime RS. Thus, there is an advantage
of longevity and iteroparity with respect to maintaining higher effective population size.
Compared to other species, variation in RS of lake sturgeon was similar. As summarized
by Clutton-Brock (1988) and Nunney (1996), standardized variance in reproductive
success from several examples of insect, amphibians, birds, mammals ranged from 0.08 –
1.35 (mean ± SD: 0.44 ±0.32), compared to that range of lake sturgeon, from 0.38 – 3.08.
In fish (e.g., steelhead Oncorhynchus mykiss), variability index from several year-runs
ranged from 1.2 – 6.5 (Araki et al., 2007), compared to that of lake sturgeon from 1.1 10.8.
Second, male behavior leads to high incidence of mating between individuals
from different groups. Further, we found no significant effects of spawning time on RS
and EMN across years. These findings have important implications to the genetic
structure of this lake sturgeon population. Matings between individuals in different
spawning groups serves to maintain gene flow among groups spawning at different times,
thereby decreasing the potential for genetic differentiation among groups within this
population. Given that spawning time of lake sturgeon in this population is repeatable
(Forsythe, 2010), if matings between members of different groups were limited, and if RS
and EMN had consistently differed by spawning time, selection associated with
environmental conditions at the time of reproduction could reinforce temporal isolation
within this population (Hendry and Day, 2005). We found that no consistent trends of
differences in RS and MS among females spawning at different times during the season
across years were observed, supporting the idea that RS and EMN partly resulted from
stochasticity of environmental conditions (Serbezov et al., 2010).

33

This study increases our understanding of the degree of variation in aspects of the
polygamous mating system of a broadcast-spawning species of conservation concern.
Estimates of the proportion of adults contributing offspring to the larval stage and
quantifying RS provides parameters necessary for the estimation of the effective number
of breeders (Nb) and helps explain temporal patterns of Nb. Plasticity in male behavior
with respect to duration of time spent in and near spawning areas is important for
interpretation of degree of sub-population structuring in this lake sturgeon population.

34

Table 1.1. Sample sizes of adults captured, adult sex ratios and the number of lake
sturgeon larvae captured and larvae genotyped in 2001 – 2007
Year
2001
2002
2003
2004
2005
2006
2007

Number of adults captured
Sex ratio
Male Female Total
(M:F)
71
44
115
1.61
71
33
104
2.24
80
40
120
2.00
76
25
101
3.04
106
47
153
2.26
162
63
225
2.57
143
63
206
2.27

Number of
larvae
collected
1,691
1,320
16,417
437
7,800
5,587
1,444

Larvae genotyped
(%) of
Number collection
553
32.7
136
10.3
488
3.0
241
55.2
362
4.6
342
6.1
1,444
100

Table 1.2. Summary of parentage assignment estimated for single programs and based on
concordance between two programs PASOS and CERVUS
% adults assigned
% offspring assigned by
Final assignment
Year
Concordance
% of
(%)*
Number larvae#
Male
Female PASOS CERVUS
2001
90.1
90.9
78.7
87.9
73.6
320
57.9
2002
76.1
93.9
79.4
78.7
85.2
92
67.6
2003
80.0
87.8
73.4
69.9
69.2
236
48.4
2004
71.1
88.0
64.3
78.4
72.3
112
46.5
2005
57.5
83.0
82.3
89.5
72.1
225
62.2
2006
65.4
98.4
80.1
91.2
73.0
212
62.0
2007
93.7
88.9
74.7
85.5
81.5
879
60.9
Note: (*) % concordance was calculated as the percentage of the female-male-offspring
triplets that were concordantly assigned by the two programs compared to the lower
number of larvae assigned from one program.
(#) % of larvae finally assigned concordantly from the two programs.

35

Table 1.3. Percentage of mating pairs (inferred from parentage) within and between
groups of adults spawning at different times of spawning seasons (2001-2007) based on
two ways groups were classified. Frequency distributions of difference in spawning date
between male and female of a pair were compared between mating pairs and random
pairs (all possible pairs between males and females captured in a given year) using G-test
(P-value provided)
a

b

By group
Within
Between
69.1
30.9
61.3
38.7
56.6
43.4
83.3
16.7
89.8
10.2
50.3
49.7
89.3
10.7

By spawning date
P-value
Within
Between
2001
64.5
35.5
< 0.01
2002
61.3
38.7
< 0.01
2003
56.6
43.4
0.65
2004
92.3
07.7
0.17
2005
89.8
10.2
< 0.01
2006
53.6
46.4
< 0.01
2007
67.4
32.6
< 0.01
Note: (a) based on discrete spawning intervals among groups within a year (Figure 2.2), an
individual was assigned as “early” or “late” spawner.
(b) based on differences in spawning dates between two partners of a pair, a
mating pair was a“within” group” if absolute difference in spawning date < 8 days.
Year

Table 1.4. Number of adults captured, fork length (mean ± SD (range)) and percentage of
adults that were assigned to larvae sampled in 2001 -2007. Difference in Lf/Lm between
mating pairs and random pairs was compared using G-test
Year
2001
2002
2003
2004
2005
2006
2007

Males captured
n
71
71
80
76
106
162
143

FL (cm)
139 ± 13
137 ± 11
136 ± 13
136 ± 14
131 ± 14
136 ± 15
135 ± 14

Female captured
n
44
33
40
25
47
63
63

FL (cm)
156 ± 10
153 ± 8.2
159 ± 10
156 ± 12
157 ± 13
160 ± 11
158 ± 10

36

Lf/Lm of a pair
Random pairs
1.16 ± 0.14
1.12 ± 0.12
1.19 ± 0.14
1.18 ±0.15
1.20 ± 0.16
1.19 ± 0.15
1.19 ± 0.15

Mating pairs
1.23 ± 0.14
1.12 ± 0.14
1.17 ± 0.13
1.18 ± 0.13
1.23 ± 0.15
1.18 ± 0.15
1.18 ± 0.14

P-value
<0.001
0.52
0.21
0.54
0.024
0.95
0.51

Easting

(a)

Canada

6

(b)

USA

5

Upper Black River

4

Northing

River Km 5

Black Lake

3
2
River Km 10
Larval capture area
Spawning area

1

N

0.0

1.0 km

0.0

0.1 km

Figure 1.1. Study site on the Upper Black River (Michigan, USA), showing positions of
adult lake sturgeon spawning areas and larval collection sites (a), and an enlarged view of
the six spawning areas (b). For interpretation of the references to color in this and all other
figures, the reader is referred to the electronic version of this dissertation.

37

Number of adults

Capture date during the spawning
season
Figure 1.2. Number of adult males and females captured by day of the spawning season
during 9 years (2001 -2009)

38

Number of adults

Capture date during the spawning
season
Figure 1.2. (cont’d)

39

Reproductive success

30
25

Female
Male

20
15
10
5
0
-5

Effective mate number

25
20
15
10
5
0
2001
-5

2002

2003

2004

2005

2006

2007

Year

Figure 1.3. Reproductive success (the number of offpsring; top) and effective mate
number (EMN; bottom) of males and females in each of 7 consecutive years (bars
represent 1 standard deviation, SD).

40

(a)

18

Female

Frequency

15

Male

12
9
6
3
0
0

4

8

12

16

20

24

28

32

36

40

44

48

52

Number of offpsring assigned to each adult

(b)

30

Frequency

25
20
15
10
5
0
0

4

8

12

16

20

24

28

32

Effective mate number of each adult

Figure 1.4. Distribution of the number of offpsring assigned to each individual (a) and
effective mate number (b) of male and female lake sturgeon captured in 2007. Larvae
were collected during the time larvae dispersed approximately 1.5 km downstream from
spawnign areas (Figure 1.1a).

41

Number of offspring per individual

50

Female
Male

40
30
20
10
0
0

5

10

15

20

25

30

Effective mate number per individual

Ln(number of offspring) per male

Ln(number of offspring) per female

Figure 1.5. Relationship between number of offspring and effective mate number of
males and females captured in 7 years (2001-2007)

4
3
2
1
0
-1
130

140

150

160

170

180

4
3
2
1
0
-1
100

Female fork length (cm)

120

140

160

Male fork length (cm)

Figure 1.6. Relationship between measures of reproductive success and body size of
males and females after adjusting for year effects.

42

Ln(number of mates) per female

Ln(number of offspring) per female

4
3
2
1
0
-1
0.0

0.2

0.4

0.6

0.8

1.0

3
2
1
0
-1
0.0

Standardized spawning date

0.2

0.4

0.6

0.8

1.0

Standardized spawning date

0.0

0.2

0.4

0.6

0.8

1.0

Standardized spawning date

4
3
2
1
0

Ln(number of offspring) per female

4
3
2
1
0

Ln(number of offspring) per female

Figure 1.7. Relationship between reproductive success (left) and effective mate number
number of mates (right) of females (expressed as natural logarithm) and standardized
spawning dates across years, after adjusting for year and year-spawning date interaction
(slope ± SE for reproductive success: 0.99 ± 0.53, P= 0.06; for effective mate number:
0.84 ± 0.49, P=0.08).

1

2

3

4

Spawning zone

Figure 1.8. Reproductive success of females spawning at different times (left) and
locations (right, bars represent the range excluding potential outliers) in 2007.

43

5

6

Frequency

0.50

2002

Random pairs

2001

0.40

Mating pairs

0.30
0.20
0.10
0.00
0
0.50

8

16

24

2003

2004

0.40
0.30
0.20
0.10
0.00
0.50

2005

0.40
0.30
0.20
0.10
0.00
0.50

2007

2006

0.40
0.30
0.20
0.10
0.00
0

8

16

0

24

8

16

24

Difference in spawning date of male and female

Figure 1.9. Frequencies of absolute pairwise difference in spawning date of random pairs
and mating pairs in 2001-2007

44

Frequency

0.30

Random pairs
Mating pairs

0.20
0.10
0.00
0.9

1.1

1.3

1.5

1.7

Lf/Lm (2001)

Frequency

0.30
0.20
0.10
0.00
0.8

1

1.2

1.4

1.6

1.8

Lf/Lm (2005)

Figure 1.10. The ratios of fork length female to fork length male (Lf/Lm) of random pairs
and mating pairs in 2001 and 2005.

45

Chapter 2
POPULATION GENETIC STRUCTURE AND RELATEDNESS
OF LAKE STURGEON POPULATION IN BLACK LAKE, MICHIGAN

ABSTRACT – Lake sturgeon (Acipenser fulvescens), a long-lived iteroparous fish
species, exhibit a multimodal temporal distribution of spawning activity, in which
individuals repeatedly spawn at the same time across spawning seasons. If repeatability
in spawning date has endured over generations, reproductive isolation within the
population may occur, leading to high genetic correlation among individuals within
relative to among members of different spawning groups. We tested this prediction for
adults captured in the same year and for adults that were captured during more than one
year over 2001-2009 in a closed population in Black Lake, Michigan, using genetic
relatedness rxy and variance in allele frequency (FST). We found that the population
exhibited a low level of genetic structure associated with spawning time, as evidenced by
significance of genetic differences and differences in mean relatedness between early and
late spawning groups of adults captured ≥ 2 times over 9 years. However, evidence of
isolation by time was lacking based on (i) small magnitude of FST and small differences
in rxy between groups of adults captured ≥ 2 years, (ii) no genetic differences among
groups within 5 of 7 years; and (iii) no relationship between inter-individual genetic
relatedness and inter-individual differences in spawning time. Based on male mating
behavior inferred from genetically identified parentage for 7 years (2001-2007), we
suggested that prolonged occupancy of spawning areas by males and plasticity in
individual spawning time increased mating opportunities among individuals from early
and late groups, thereby retarding temporal reproductive isolation.
46

INTRODUCTION
Groups of individuals within a population can become reproductively isolated and
genetically structured as a result of two major mechanisms including isolation by distance
(IBD) and isolation by time (IBT) (Hendry and Day, 2005; Maes et al., 2006; Doctor and
Quinn, 2009). Compared to IBD, which has been well-studied in species exhibiting
homing behavior to natal areas (Taylor, 1991; Quinn et al., 2000), IBT was less
commonly reported. Most empirical examples of IBT have been shown in plant and
salmonid species whose breeding times are heritable (Hendry and Day, 2005).
Heritability of spawning time can limit gene flow between individuals that spawn at
different times, especially early and late in the spawning season, leading to genetic
differentiation within a population even when spawning activity is unimodally distributed
(Hendry and Day, 2005; Doctor and Quinn, 2009). In addition, differences in spawning
time can result in nonrandom mating or temporal assortative mating (Fox, 2003; Weis
and Kossler, 2004; Devaux and Lande, 2008). Depending on the degree and duration that
reproductive isolation is maintained, increases in genetic relatedness among individuals
within groups can develop.
Degree of genetic structure within a population is influenced by interactions of
factors associating with spawning time such as natural selection (Hendry et al., 1999;
Hendry and Day, 2005; Casagrande et al., 2006; Tomaiuolo et al., 2007), spawning
behavior (Johannesen and Lubin, 1999; Fagan et al., 2010), and variation in individual
reproductive success (Maes et al., 2006; Pujolar et al., 2006). These factors can vary by
time and interact with environmental conditions at spawning time generating underlying
temporal genetic differences within populations (Calderon and Turon, 2010).

47

When environmental conditions associated with adult spawning time affect
offspring fitness (Einum and Fleming, 2000), selection regimes that vary during the
spawning season can result in differences in adaptation among temporal spawning groups
within a population (adaptation by time, Hendry et al., 1999; Hendry and Day, 2005). For
example, seasonal variation in reproductive success has been reported in numerous
species (e.g., birds (Verhulst et al., 1995); fish (Schultz, 1993)). Selection can also
reinforce differences in spawning time based on adaptations of individuals to conditions
at the time of spawning in sub-populations, thus further limiting effective gene flow
among subpopulations (Fraser and Bernatchez, 2005).
In addition, individual movements during the spawning season can result in
mating between individuals from different groups and likely plays an important role in
determine the degree of genetic population structure (Johannesen and Lubin, 1999; Fagan
et al., 2010). In mammals, male-mediated gene flow resulting in genetic structure within
and between populations has been well-documented (e.g., mammals (Crawford et al.,
2009; Pilot et al., 2010). In fish, especially for species with polygynous mating, male
fitness is limited by mate availability and males can employ different strategies to find
mates (Fagan et al., 2010). In brook charr Salvelinus fontinalis, for example, males
disperse among tributaries decreasing the degree of spatial isolation (Fraser et al., 2004).
Meanwhile, in other species such as lake sturgeon (Acipenser fulvescens), males can
remain at spawning locations, thereby increasing opportunities for mating with females
from different spawning groups (Bruch and Binkowski, 2002). Such male behavior
facilitates opportunities for gene flow among spawning groups spatially or temporally
distributed segregated.

48

While evidence of ITB has been reported in several fish species including
salmonids (review in Hendry and Day, 2005), Eels (Maes et al., 2006), there is a lack of
information on IBT in long-lived iteroparous species whose generation time can take one
to several decades. Heritability of spawning time in such species has not been available.
However, some species such as lake sturgeon Acipenser fulvesnces (Forsythe, 2010) have
been shown to spawn at similar times during multiple years. If repeatability of spawning
time in these iteroparous species has a genetic basis, IBT can exist in populations
composed of discrete multiple spawning groups. Such information is important for
conservation and management of threatened species with small population sizes because
limited gene flow together with genetic drift would further reduce genetic variation
within a small population.
The lake sturgeon have a unique life history characterized by extreme longevity
(potential > 100 years), iteroparity (Auer, 1999) and polygamous mating by both sexes
(Bruch and Binkowski, 2002). Males mature at younger ages (12 -15 yrs.) and spawn at
shorter intervals (1-3 yrs.) compared to females (18-27 yrs. and 3-7 yrs., respectively)
(Auer, 1999; Bruch, 1999; Forsythe, 2010). The species exhibits reproductive
characteristics which can lead to IBT and development of sub-population genetic
structure within the population. Spawning activity typically occurs in bimodal or
multimodal peaks. Like salmonids (review in Doctor and Quinn, 2009), lake sturgeon
individuals repeatedly spawn at similar times and locations (Auer, 1999; Forsythe, 2010).
Data collected over eight years in Black Lake, (MI) shown that repeatability (the ratio of
among-individual variance to the total variance for a given trait) in spawning time of
individuals captured in two or more years was high for both females (0.56) and males

49

(0.42) (Forsythe, 2010). These results suggest that different segments of the population
may be reproductively isolated as a function of spawning time. Moreover, effects of
environmental conditions (e.g., water temperature, discharge) which are correlated with
spawning time contribute to temporal variation in phenotypic traits of offspring and of
early and late spawning adults, possibly leading to different selection regimes through the
spawning season. Previous studies in this lake sturgeon population have documented
temporal variation in traits including egg incubation time, yolk-sac size at hatch
(Crossman, 2008), and embryonic and larval developmental time until dispersal (Duong
et al., in press), which may lead to reinforcement of difference in spawning time.
The goal of this study was determine whether levels of genetic differentiation
between early and late spawning adults in this population exited as predicted based on
high repeatability in spawning times observed over eight-year period in Black Lake,
Michigan. Our objectives were to (i) estimate genetic difference among adults from
different spawning groups within a year and across years; and (ii) quantify the
relationship between inter-individual genetic relatedness and inter-individual differences
in spawning time.
METHODS
Study site
The study was conducted in the Black Lake, Michigan. This lake sturgeon
population is isolated by dams blocking immigration and emigration from other
populations in adjacent lakes (Smith and King, 2005). Lake sturgeon adults spawn in the
Upper Black River (UBR), the largest tributary to Black Lake and the sole spawning
locations for the population. Adults spawn in multiple locations over a period of 19-42

50

days (Forsythe, 2010), extending from late April through early June (Figure 1.2).
Spawning occurs within a short 1.5 km-section of the river, about 9 km upstream from
the river mouth (Smith and King, 2005; Forsythe, 2010). The small, shallow and wadable
river section permits sampling of nearly all spawning adults each year.
Adult sampling
We sampled adults each year for 9 consecutive years (2001-2009). There were
1,448 captures of 760 individuals including 460 males and 300 females. During each
spawning season, we collected lake sturgeon adults daily by walking the entire spawning
area of the stream one or more times per day. Adults were captured as they arrived at the
spawning areas using long-handled nets. Upon capture, sex was determined and
biological information including, body weight (kg), fork length (cm), capture date and
2

location were recorded. A small fin clip (~1 cm ) was collected from each adult for
genetic analysis.
We used only information of date of capture to classify an individual as early or
late spawner in order to examine genetic differences among groups of adults (a group was
defined as those individuals that had similar dates of first capture) spawning at different
times. We have observed that lake sturgeon females spent one or few days at the
spawning areas whereas males remain on spawning areas for longer periods and earlyspawning males could leave the areas but return subsequently with late-spawning
females.
Microsatellite DNA analysis
DNA was extracted from fin clips of adult samples using a QIAGEN DNeasy(R)
kit (QIA Inc.). DNA concentration was measured using a Nanodrop 1000

51

spectrophotometer (Thermo Fisher Scientific Inc.). All samples were genotyped at 12
tetra-nucleotide microsatellite loci including Spl 120 (McQuown et al., 2000); AfuG 68B
(McQuown et al., 2002). 2002); Aox 27 (King et al., 2001); AfuG 68, AfuG 9; AfuG 63,
AfuG 74, AfuG 112, AfuG 56, AfuG 160, AfuG 195 and AfuG 204 (Welsh et al., 2003).
Using polymerase chain reaction (PCR), 100 ng DNA was amplified in 25 µl reaction
mixtures as described in the above references. All PCR reactions were conducted using a
Robocyler 96 thermal cycler (Stratagene). PCR products were run on 6% denaturing
polyacrylamide gels and visualized using a Hitachi FMBIO II scanner. Allele sizes were
scored independently by two experienced personnel. Genotyping errors checked by
blindly re-genotyping a random subset 10% of all samples.
Data analysis
A previous study showed that high repeatability in spawning times of lake
sturgeon was observed over eight-year period in Black Lake, Michigan (Forsythe, 2010).
If repeatability in spawning date has endured over generations, reproductive isolation
may lead to high genetic correlations among individuals within relative to among
members of different spawning groups (as evaluated by rxy) and accrual of variance in
allele frequency (FST). We tested this prediction for adults captured in the same single
years and for adults that were captured ≥ 2 years during 2001-2009.
Genetic differences among groups of adults spawning during early and late season
Analyses were based on adults captured in each of 9 years and for adults that were
captured ≥ 2 years over 9 consecutive years (2001-2009). For a single year, adults were
classified into early or late spawning groups based on discrete spawning dates between
groups (temporal spawning groups). In years 2006-2009, the interval among spawning
52

groups was short and there were more than two peaks of spawning activity. In these
years, spawning groups were classified as early, middle and late.
Over 9 years of sampling, there were a total of 349 individuals that were captured
≥ 2 years (92 females and 257 males). We classified an individual as an early, middle or
late spawners based on the mean of “standardized spawning dates” across years that the
individual was captured. The standardized spawning date (SSD) of an adult in a given
year is the number of days of first capture from the first day of the spawning season to the
day the individual was captured divided by the total duration of spawning activity in that
year (0 < SSD ≤ 1). SSD thus accounted for different durations of spawning activity
across years. The early spawning group was defined as individuals with 30% smallest
SSD, and late spawning group were composed of individuals with 30% highest SSD.
With this classification, the ranges (mean ± standard deviation) of SSD of early and late
groups did not overlap (Figure 2.1). Because spawning date of an individual could vary
across years, groups classified based on mean SSD better represent temporal spawning
groups of the whole population compared to groups in a single year.
We tested the null hypothesis that there was no genetic differentiation (i.e., degree
of population structuring) either among spawning groups of adults each year or among
groups of adults captured during ≥ 2 years. First, we tested whether genotypic
frequencies deviated from Hardy-Weinberg expectations for spawning adults each year.
Deviations were tested using a Fisher’s exact test in program GENEPOP 4.0 (Raymond
and Rousset, 1995; Rousset, 2008). The exact P-value of the tests were estimated based
on the Markov chain method (Guo and Thompson, 1992). The Markov chain method was
also used to estimate P-values for contingency tests of genotypic differences between

53

early and late spawning groups within a year. Intra-annual population structure measured
as the variance in allele frequency among individuals within and among groups (FIS and
FST, respectively) (Weir and Cockerham, 1984) was quantified using program
GENEPOP.
Relatedness
Estimates of genetic relatedness among individuals from the same and different
groups can be used to test whether spatial or temporal reproductive isolation has occurred
within a population (Rousset, 2000; Hendry and Day, 2005). Based on observations of
repeatable spawning times (Forsythe, 2010), we predicted that genetic relatedness among
individuals spawning at similar times would be greater than relatedness among
individuals spawning at different times.
Genetic relatedness (rxy) or the coefficient of relationship defined by Sewall
Wright is the probability that the alleles present at a random locus will be identical by
descent (Hamilton, 1963). More formally, rxy is a measure of genetic similarity between
two individuals relative to the relatedness between two random individuals from the same
population (Blouin, 2003). There are several commonly used estimators of relatedness,
which can be grouped into two classes, including maximum likelihood (ML) estimators
and various methods of moment (MOM) estimators (review in (Ritland, 2000; Van de
Casteele et al., 2001; Blouin, 2003; Csillery et al., 2006). The rank order of estimator
performance varies by research scenario, generally is population-specific, depending on
the frequency distribution of relatedness in the population and properties of markers used
(e.g., allele frequency distribution, number of allele per locus (Van de Casteele et al.,
2001; Csillery et al., 2006; Oliehock et al., 2006).
54

To choose a relatedness estimator for lake sturgeon population, we evaluated the
correlation of estimator-based rxy values and true values that are simulated from allele
frequencies of the population (Wang, 2010). We used program COANCESTRY (Wang,
2010), which allows simultaneous estimation of 2 ML estimators (Triadic likelihood
[TriML] (Wang, 2007b) and Dyadic likelihood [DyadML] (Milligan, 2003)) and 5
MOM estimators (proposed by Queller and Goodnight, 1989 [QG]; Lynch and Ritland
1999 [LynchR]; Wang, 2002; Lynch 1988 and Li et al., 1993; and Ritland 1996).
Simulations were conducted using 100 reference individuals as recommended (Wang,
2010) for each type of dyad relationships (e.g., unrelated, parent-offspring, ful-sib, halfsib, and first cousins). The QG, TriML and DyadicML were more highly correlated (0.74
– 0.75) with true values compared to other estimators (0.46 – 0.70), and they were highly
correlated to each other (0.89 – 0.99). Similarly, the QG-based estimates were also highly
correlated (0.81 – 0.86) with estimates based on Konovalov and Heg (Konovalov and
Heg, 2008a) and ML (Konovalov and Heg, 2008b) estimators using program
KINGROUP, version v2_09050 (Konovalov et al., 2004). Thus, we chose to use the
Queller and Goodnight (QG) method based on 1) the high degree of correlation with true
values and other estimators; 2) the high frequency of the QG method in the literature
(Csillery et al., 2006); and 3) that the QC method has been described as being robust in
different conditions (Van de Casteele et al., 2001). Queller and Goodnight relatedness
values were generated using the program RELATEDNESS 5.0 (Goodnight and Queller,
2001). Pairwise relatedness among spawning adults was estimated based of allele
frequencies of the whole population (each individual that spawned during 9 years, 20012009, was included, N = 722).

55

We estimated pairwise relatedness for adults in 9 single years and for adults
captured during ≥ 2 years. First, for single years, we examined the genetic relatedness
among individuals within a year in regard to their spawning time. Specifically, we used
the difference in SSD as a measure of the proximity in spawning time, and tested whether
individuals spawning at similar times were more closely related than individual spawning
at different times of a spawning season. Mantel tests implemented in PASSAGE
(Rosenberg, 2001) were conducted to test for significance of the correlation between
matrices of pairwise relatedness and pairwise difference in SSD, as suggested by Hendry
and Day (2005).
For 349 individuals captured during ≥ 2 years, we compared mean relatedness
within (early, middle and late) and between groups. Because each individual was used to
compute pairwise relatedness with multiple individuals, we used a permutation approach
to test for significance of differences in mean relatedness (rxy) within and between
spawning groups using an SAS-based program (A. Saxton, University of Tennessee,
unpublished software) (Ratnayeke et al., 2002). The program was set to compute the
observed difference between the means of the two groups first. Then, data from the two
groups were pooled, and the pooled data were randomly sub-sampled 1000 times. Each
time, the difference between the observed and theoretical mean calculated from the
pooled data was computed and the one-tailed probability that mean rxy of within groups
is greater than rxy between groups was tested. The absolute differences in distribution of
cumulative frequencies of r values between groups were tested using non-parametric
Kolmogorov-Smirnov tests conducted in R (R-Development-Core-Team, 2009).

56

RESULTS
Data from 9 years of adult capture and recapture show that adults consistently
spawn over a range of time (Figure 1.2) and that the time of individual spawning is
repeated across years. Degree of repeatability, as evidenced by the standard deviation of
standardized spawning date is particularly small in the earliest and latest spawning adults
(Figure 2.1). The possibilities of genetic differentiation within this population were
examined at two temporal scales, (i) among groups of adults captured in the same single
years and (ii) among groups of adults that were captured ≥ 2 years during 2001-2009.
Genetic differentiation and mean relatedness among groups of adults captured ≥ 2
years
Evidence for genetic differences among adults with regard to spawning time
(early vs. late) was examined across years. During 9 years, 349 individuals that were
captured ≥ 2 years were grouped as early-, middle- and late-repeated spawning groups.
The overall FST for these three groups of adults (both males and females) was low (mean
± SE = 0.004 ± 0.001) but was significant (P <0.01). Genotypic differences (chi-square
tests) and FST were significant between early- and middle-, and early- and late- repeated
spawning groups of males, and when both sexes were combined but was not significantly
different among groups of females. Meanwhile, there were no genetic differences
between middle- and late- spawning groups of males or females (Table 2.1).
We compared mean pairwise relatedness (rxy) among repeated spawning
individuals from within early-, middle- and late-spawning groups and among individuals
from different groups for females and males and when both sexes were combined. Mean
pairwise relatedness (rxy) of females was not significantly different among groups (P >
57

0.10). Mean rxy of males as well as estimates for all adults (both sexes) was significantly
higher in the early spawning groups compared to the middle or late spawning groups (P <
0.01). Mean rxy of individuals within middle and within late groups was similar and not
significantly different from mean rxy between groups, either sexes were separately
analyzed or when combined (P < 0.01) (Table 2.2).
We further compared the differences in frequency distribution of rxy values
between early spawning group (the group characterized by the highest mean rxy) and the
other spawning groups. The absolute differences in cumulative frequency distribution of
relatedness values from early spawning group compared to middle and late spawning
groups were 5.9% (P <0.01) and 4.4% (P <0.01), respectively. These differences were
small (Figure 2.2) and thus may not be biologically relevant.
Genetic differentiation and mean relatedness among groups of adults in single years
Within single years, genetic differences among groups were also evaluated based
on genotypic difference (Chi-square tests) and the significance of FST estimated. In years
2006 – 2009, spawning time was not clearly distinct among groups. At first, we divided 3
groups in each of these years and tested genetic differences among three groups. Because
there was no genetic difference between middle and late groups (non-significant
difference in genotypic frequencies between the two groups and non-significant
difference from zero of FST), individuals from middle and late groups were pooled as
“late” spawning group. Genetic differences between early and late adult groups were

58

inconsistent among years (Table 2.3). Estimates of FST between early and late spawning
groups was significant (P <0.01) only in 2007 (0.0048) and 2006 (0.0044).
To examine genetic relatedness among individuals in the same year, we tested the
significance of correlations between pairwise relatedness and pairwise difference in SSD
among individuals. A significantly negative correlation would indicate that individuals
spawning at a similar time were more genetically related than individuals spawning at
different times in a spawning season. However, correlations were not significantly
different from 0, indicating that pairwise relatedness was independent of pairwise
difference in spawning date among individuals captured within the same year (Table 2.4).
DISCUSSION
Limited gene flow among adults that spawn at different times can result in genetic
structure within populations (Hendry and Day, 2005; Doctor and Quinn, 2009). Genetic
differences among groups of adults spawning at different times if at low levels do not
provide evidence of IBT (Hendry and Day, 2005). Evidence of IBT would be stronger if
a low level of genetic differences among groups had a consistent pattern across years.
Our nine-year data on spawning time and genotypes for each individual of almost all lake
sturgeon adults in Black Lake population revealed important insights into the evidence
for IBT in this closed population.
We found that the lake sturgeon population in Black Lake exhibited a low level of
subpopulation structure associated with spawning time, as evidenced by significance of
genetic differences and differences in mean relatedness between early and late groups of
adults captured ≥ 2 years. However, there was lack of evidence of isolation by time in
this population given (i) low observed values of FST between early and late groups of
59

adults captured ≥ 2 years (across years); (ii) non-significance of intra-annual FST in 5 of 7
years; and (iv) no correlation between pairwise relatedness and pairwise difference in
spawning time.
Grouping an adult into early or late spawning groups based on mean standardized
spawning date better reflected the timing of spawning of each individual. Accordingly,
genetic differences were observed between early and late groups of adults captured ≥ 2
years but not between groups within a year suggesting that repeatability in spawning date
represented some level of heritability. Individuals that spawned early and late in the
seasons repeatedly across years are more likely to be progeny of adults from separate
groups from the parental generation that spawned early and late, respectively. Limited
gene flow could occur between groups of individuals spawning earliest and latest in the
season, while gene flow could be mediated by individuals spawning at middle season
and/or by individuals exhibiting greater plasticity in spawning time (based on larger
variation in standardized spawning date). As a result, genetic differentiation could be
initiated at very low levels and accumulate slowly over generations, resulting in the
heterogeneous genetic composition observed in the population at present. In a given year,
individual spawning time was plastic (as shown by standard deviation in SSD, Figure
2.1), thus a spawning group (i.e., early or late) could include individuals from different
groups in which they actually belonged to. Mixing individuals from different temporal
spawning groups that were genetically differentiated a low levels contributed to the nonsignificant genetic differences between early and late groups in a given year.
Levels of genetic differentiation within the lake sturgeon population could be
influenced by the behavior of males during spawning season, the plasticity in spawning

60

time of adults, and the comparatively short spawning season. These factors could vary
temporally in response to fluctuations environmental factors (Forsythe, 2010). In several
lake sturgeon populations (e.g., Black Lake, Michigan (Smith and Baker, 2005); Wolf
and upper Fox rivers, Wisconsin (Bruch and Binkowski, 2002), males have been
observed to spend long periods of time at the spawning areas, which could result in
matings with multiple females including females from early and late groups.
Concordantly, based on genetically identified parentage, we have found that mating of
individuals from different groups (early vs. late) comprised from 10 - 40% of total mating
pairs from 2001-2007 (see chapter 1). Therefore, gene flow among spawning groups
occurred frequently and varied yearly. Data revealed that “distinct” groups of adults
spawning early and late in the season based on direct observations were not
reproductively isolated. The behavior of males appears to play an important role in
retarding temporal reproductive isolation.
Plasticity in spawning time of repeated spawning individuals may also contribute
to the lack of evidence of reproductive isolation. Individuals that repeatedly spawn late in
the seasons had larger variation in spawning time compared to individuals spawning early
in the seasons (Figure 2.1). Plasticity in spawning time commonly observed in numerous
species (e.g., birds (Reed et al., 2009); fish (Valiente et al., 2010) has been reported as a
result of the interaction between age, experience and environment (Wilson et al., 2007).
In lake sturgeon, variation in spawning time was observed to be associated with interannual fluctuation of stream environmental factors such as water temperature and
discharge (Forsythe, 2010).

61

Spawning duration of lake sturgeon might also explain low levels of genetic
differentiation within the population and the lack of evidence of reproductive isolation.
The spawning season of lake sturgeon is short (19 -42 days in Blake Lake population,
similar to other populations (Bruch and Binkowski, 2002)). Short spawning duration and
thus a short temporal distance (the period separating spawning groups) provides
opportunities for gene flow within the population. Temporal distances of lake sturgeon
spawning groups were about 7- 15 days (4 days apart in some years). We have found
fewer mating pairs occurred when temporal distance between males and females
increased (see chapter 1). In many salmonids populations, temporal distance between
groups was about 1 month apart, and genetic difference was positively correlated with
temporal distance among groups (for review on salmonids, see Hendry and Day, 2005).
In two sockeye salmon populations in Tustumena Lake, Alaska, for example, genetic
differences between early and late groups were not significant (FST = 0.003) in one
population with temporal distance of 13-15 days, and significant (FST = 0.006) within the
other population with temporal distance between groups of 21-25 days (Woody et al.,
2000).
The values of FST found between early and late repeated spawning lake sturgeon
are comparably lower than FST reported from different groups within populations of
other species (e.g., sockeye salmon Oncorhynchus nerka (Fillatre et al., 2003); European
Eel Anguilla anguilla (Maes et al., 2006; Pujolar et al., 2006); red sea urchin
Strongylocentrotus franciscanus (Moberg and Burton, 2000); purple sea urchin,
Strongylocentrotus purpuratus (Edmands et al., 1996); Pacific oysters Crassostrea gigas

62

(Li and Hedgecock, 1998)). In sockeye salmon (Oncorhynchus nerka) in Klukshu River
(Yukon, Canada), for example, FST between runs (spawning groups) within years (19942000) were significantly greater than zero (range: 0.018–0.041) except one year FST
=0.004 (Fillatre et al., 2003).
In addition to testing genetic difference at the group-level, we also examined
pairwise relatedness as a measure of IBT at the individual level. Although mean
relatedness of repeated spawning individuals from the early group was significantly
higher than that of individuals from middle and late groups, the magnitude of differences
was small, and thus the differences were not likely biological relevant. Moreover, in each
year, estimates of pairwise relatedness among adults were found to be independent of
difference in spawning time. Therefore, at the individual level, there was no evidence of
high genetic correlations among individuals spawning at similar time during spawning
seasons.
There are several possible biological and statistical explanations for this result.
First, no correlation of pairwise relatedness and difference in spawning time despite
heritable spawning time could be due to polygamous mating system and straying of lake
sturgeon males. Polygamous mating helps reduce pairwise relatedness among adults
compared to monogamous mating with the same number of breeders (Woxvold and
Mulder, 2008) due to lower probability of shared alleles among siblings. Further, male
lake sturgeon prolong their stay in the spawning areas to increase opportunities to mate
with early and late females (see chapter 1), which increases gene flow between early and
late spawning groups. Mean pairwise relatedness in the population also depends on
effective population size, where small effective population size can increase the
63

probability of mating among closely relatedness individuals (Falconer, 1989; Launey et
al., 2001). With Ne estimated at ~132 (see chapter 3), inbreeding will increase only
slightly each generation. Finally, relatedness estimates have large variance (Ritland,
2000; Van de Casteele et al., 2001; Csillery et al., 2006), there would be low power to
detect significance of differences in mean relatedness among groups (Csillery et al.,
2006).
In sum, based on the low signal of genetic differentiation and difference in
relatedness between early and late groups of individuals captured ≥ 2 years, subpopulation structuring has developed but at a low level in this lake sturgeon population. If
spawning time is heritable, low genetic differentiation within this population can be
maintained despite the plasticity in spawning time of both males and females and male
behavior during the spawning season, which increases gene flow among spawning
groups.
ITB would evolve if gene flow among temporal spawning groups was limited and
if natural selection on adults and/or offspring acted differently during different periods of
the spawning season, initiating the process of temporal lineage segregation (Tomaiuolo et
al., 2007). In the lake sturgeon population in Black Lake, the conditions (i.e., limited
gene flow, differential selection) promoting ITB have not been observed. We have found
that the number of offspring produced did not vary as a function of spawning time,
indicating no differential selection across years during the spawning season (see chapter
1). Moreover, the time required for evolution of spatial reproductive isolation is highly
variable and usually takes long periods (Nei et al., 1983), from several decades
(introduced sockeye salmon in Lake Washington for 13 generations (Hendry et al.,

64

2000); Chinook salmon in Newzerland for 30 generations (Quinn et al., 2000) to
thousand years (Schluter, 1996; Taylor, 1999). The time required for evolution of
temporal reproductive isolation could take longer because factors such as environmental
conditions that contribute to temporal reproductive barriers are more dynamic.
Therefore, given the species’ prolonged inter-generation interval, genetic differentiation
could develop at low levels but would not exhibit temporal reproductive isolation for a
long period of time.

65

Table 2.1. Pairwise differences in genotypic frequency (Chi-square test, df = 24) and
variance in allele frequency (FST) among groups of individuals that were captured ≥ 2
years during 2001-2009
Female
Group

2

X

Male

Sexes combined

2

FST

X

FST

2

X

FST

Early-Middle

21.0

-0.0001

56.2***

0.0070**

56.7***

0.0053**

Early-Late

24.6

0.0019

57.0***

0.0065***

65.8***

0.0063***

Middle-Late
21.2 0.0008
22.2
-0.0009
22.0
0
Note: Asterisks indicate the significance level of a test (** when P <0.01, and *** when
P <0.001)

Table 2.2: Mean (± SD) of estimated inter-individual relatedness within and between
members of early and late spawning groups based on adults captured ≥ 2 years, as
indicated in Figure 2. 1
Spawning group

Male
b

Within early

0.0228 ± 0.2316

Within middle

-0.0107 ± 0.1965

Within late

-0.0001 ± 0.2171

Between groups

a

a

Female
a

0.0092 ± 0.2191
a

0.0152 ± 0.2123
a

-0.0193 ± 0.2118

a

a

Both sexes
b

0.0200 ± 0.2272
a

-0.0047 ± 0.2117
a

0.0019 ± 0.2136
a

-0.0011 ± 0.2087 -0.0092 ± 0.2083 -0.0028 ± 0.2105
Note: Values within a column with the same letter are not significantly different (P
>0.05), using permutation tests.

66

Table 2.3. Differences in genotypic frequency (Chi-square test) and variance in allele
frequency (FST) between early and late spawning groups within each of 9 years 20012009
Genotypic difference
Year

2

FST

X

P-value

2001

32.7

0.11

0.0069

2002

14.9

0.92

-0.0038

2003

37.2

0.042

0.0025

2004

18.8

0.76

-0.0013

2005

28.2

0.24

0.002

2006

42.9

0.01

0.0044*

2007

45.2

0.006

0.0048*

2008

23.3

0.49

0.0002

2009
34.5
0.08
0.0015
Note: An asterisk indicates the significance level (P <0.05) of a test

Table 2.4. Results of Mantel tests of correlations between inter-individual relatedness
(rxy) and inter-individual differences in standardized spawning dates between individuals
within a year (2001-2009)
Year

Correlation coefficient

P-value

2001

-0.0178

0.688

2002

0.0208

0.220

2003

-0.0214

0.984

2004

-0.0075

0.425

2005

-0.0151

0.519

2006

0.0040

0.810

2007

-0.0227

0.424

2008

0.0175

0.465

2009

-0.0002

0.838

67

Standardized spawnign date

1.2
Early
1.0

Middle
Late

0.8
0.6
0.4
0.2
0.0
0

20

40

60

80

100

Female

Standardized spawnign date

1.2
1.0
0.8
0.6
0.4
0.2
0.0
0

40

80

120

160

200

240

280

Male

Figure 2.1. Standardized spawning date (SSD) of 92 females (top) and 257 males
(bottom) that re-spawned more than 1 time during 2001-2009. Individuals were ordered
left to right by ascending values of mean SSD.

68

Figure 2.2. Probability density estimation of estimated inter-individual genetic
relatedness among adults from early spawning groups (solid line) and from late spawning
groups (dash line) based on males and females that were captured ≥ 2 years during 20012009.

69

Chapter 3
EFFECTIVE POPULATION SIZE AND TEMPORAL VARIATION
IN EFFECTIVE NUMBER OF BREEDERS FOR THE LAKE STURGEON
POPULATION IN BLACK LAKE, MI

ABSTRACT – Estimating effective population size (Ne) and quantifying temporal
variation in effective number of breeders (Nb) and the relationship between Nb and adult
census size (N) are important to predict genetic changes in small and isolated populations,
especially for threatened species like lake sturgeon (Acipenser fulvescens). Estimates of
Nb were computed using (1) demographic data inferred from genetically identified
parentage and (2) linkage disequilibrium (LD) based on 12 microsatellite loci for larval
samples collected in each of seven years, 2001-2007. The two methods yielded similar
estimates of Nb. Estimates of Nb were fairly constant as were adult census sizes among
years (mean of LD-based Nb ranged 63-126, 95% CI ranged 39-197). Low inter-annual
variation in Nb/N ratios (0.26 - 0.61) resulted from consistently low standardized variance
of relative reproductive success across years. Consistency of Nb/N was realized even
though the number of larvae collected, a surrogate measure of total annual larval
production, varied 40-fold across years (437 – 16,417). Ne per generation (mean: 132,
95% CI: 104-167) estimated based on a mixed sample of lake sturgeon adults captured
during 2001-2007 (N = 612) was close to annual estimates of Nb. Therefore, for lake
sturgeon generational estimates of Ne should not be estimated as a product of the
harmonic mean of effective number of breeders and generation time as has been

70

suggested for semelparous species. That approach would also probably be inappropriate
for other long-lived iteroparous species.
INTRODUCTION
Effective population size (Ne) is an important parameter describing evolutionary
processes of finite populations, and therefore has been an important conceptual focus in
evolutionary studies and in conservation biology (Palstra and Ruzzante, 2008;
Charlesworth, 2009; Waples, 2010). Ne allows predictions of generational rates of change
in allele frequencies and of inbreeding occurring in a finite population due to genetic drift
and interactive effects of drift with mutation, migration, recombination and selection
(Charlesworth, 2009). Since Frankham’s review (1995) revealing that the ratio of
effective to census population size (Ne/N) across 102 wild species was smaller (median of
0.1) than previously perceived, more attention has been focused on investigations of Ne
and causal factors affecting Ne for natural populations (Mackay, 2007). Small and
fragmented populations can be exposed to greater risk of extinction due to low Ne (Palstra
and Ruzzante, 2008). Therefore, estimation of Ne and Ne /N is especially important to
elucidate conservation issues for small and isolated populations.
Effective population size (Ne) as first defined by Wright (1931) is the number of
individuals in a theoretically “ideal” population that has the same magnitude of effects
due to genetic drift and inbreeding as a real population of size N (Nunney, 1993;
Charlesworth, 2009). In an ideal population, mating is assumed to be random and each
individual has an equal probability of contributing gametes to the next generation. In
natural populations, numerous factors can cause unequal probabilities of genetic
71

contributions among adults to progeny, resulting in smaller Ne than N (Frankham, 1995;
Charlesworth, 2009; Waples, 2010). Major factors that can lower Ne include fluctuations
in adult population size over time, skewed sex ratios and non-random variation in lifetime
reproductive success (Charlesworth, 2009) (Frankham, 1995; Charlesworth, 2009;
Waples, 2010). In addition, Ne is usually smaller than N in populations which are
structured by age or stage-classes, or which experience inbreeding, selection or limited
migration (Charlesworth, 2009). Population structuring which results from spatial and
temporal reproductive isolation among sub-populations (Hendry and Day, 2005; Doctor
and Quinn, 2009) can also reduce Ne (Chesser et al., 1993; Waples, 2010). Moreover,
biological characteristics of a species such as attributes of the mating system (Nunney,
1991, 1993; Pearse and Anderson, 2009), generation time, life span and levels of
iteroparity (Waples, 1990; Nunney, 1996) also influence Ne and Ne/N in different species
and populations.
Different methods have been used to estimate Ne. Because of the difficulties in
estimating demographic parameters necessary for estimation of Ne in natural populations,
indirect estimation of Ne using genetic data is commonly applied (Waples, 1989; Luikart
and Cornuet, 1999). Several approaches have been used including one – sample based
methods such as linkage disequilibrium (Hill, 1981; Waples, 1991), heterozygote excess
(Pudovkin et al., 1996), approximate Bayesian computation (Tallmon et al., 2008), and
sib-ship method (Wang, 2009). Two – sample based methods such as temporal variance
in allele frequency (Krimbas and Tsakas, 1971; Waples, 1989), and loss of

72

heterozygosity (Harris and Allendorf, 1989) (review in Luikart and Cornuet, 1999;
Nunney, 2000; Luikart et al., 2010) have also been employed. Genetic data can also be
used to determine parentage in natural populations (Pemberton, 2008), allowing
estimations of the mean and variance in male and female reproductive success that can be
used for estimation of Ne (Crow and Denniston, 1988; Caballero, 1994).
Most genetic estimators have been developed for semelparous species (e.g.,
Pacific salmon Oncorhynchus spp.) based on key assumptions of the Wright’s (1931)
ideal population, including population closure and discrete and non-overlapping
generations. For long-lived and iteroparous species, if samples are obtained from a single
cohort, standard methods for Ne estimation mentioned above can be used to estimate the
effective number of breeders (Nb, number of breeding adults that produced the cohort)
(review in Waples, 2010). However, Nb is difficult to relate to Ne in long-lived and
iteroparous species because breeders in a year include a mixture of adults from different
age classes including representatives from different generations (Luikart et al., 2010;
Waples, 2010). Long-term data from a closed population in which samples can be
collected from multiple cohorts may help resolve relationships between Ne and Nb for
long-lived and iteroparous species.
Lake sturgeon (Acipenser fulvescens) is an important species to study Ne,
relationship of Ne to Nb and ecological variables that affect effective sizes. The species is
a target of conservation (Holey et al., 2000; Peterson et al., 2007) because lake sturgeon
numbers have been severely depleted over the past 100 years throughout their native
range (Houston, 1987; Hay-Chmielewski and Whelan, 1997). Abundance of many lake
73

sturgeon populations has remained low due to anthropogenic factors resulting in
fragmented habitats and degraded water quality (Peterson et al., 2007). Aspects of the
species life history and reproductive ecology including longevity and low natural
recruitment rates of lake sturgeon also influence relationships among Ne, Nb and adult
census size.
Lake sturgeon are characterized by a long generation time (15-20 years), extreme
longevity (potential > 100 years), iteroparity (Auer, 1999) and polygamous mating by
both sexes (Bruch and Binkowski, 2002). Each of these characteristics can influence
Ne/N. In species with overlapping generations like lake sturgeon, the ratio of Ne/N is
expected to increase as generation time increases (Felsenst.J, 1971; Hill, 1972; Waples,
2010). Iteroparity of lake sturgeon can result in increasing variance in life time
reproductive success over time, which reduces Ne/N, if individuals with higher fecundity
are consistently more successful than other individuals in producing offspring that
survive to reach sexual maturity. In contrast, if differences in reproductive success among
individuals were not consistent across their lifetime, iteroparity would result in
comparatively higher Ne/N (Nunney, 1996; Turner et al., 2002). The broadcast spawning
behavior of lake sturgeon can also result in high variance in reproductive success. Lake
sturgeon exhibit aggregate mating, where negatively buoyant eggs and sperm are released
by groups of females and males over large areas of rock and gravel in the absence of nest
preparation and without post-ovulatory parental care. Because of the species' mating
behavior and exposure of eggs to environmental conditions and to predation, juvenile
lake sturgeon experience extremely high mortality early in life (Kempinger, 1988).

74

However, the prediction of high variance in reproductive success on Ne/N may be
countered by the species promiscuous mating system where males and females can have
multiple mates within a year (DeHaan, 2003; Duong, chapter 1). Promiscuity should
decrease male and female variance in lifetime reproductive success, acting in an opposite
manner compared to effects of variance in reproductive success on Ne/N (Nunney, 1993;
Sugg and Chesser, 1994; Pearse and Anderson, 2009).
Infrequent spawning intervals and repeatedly spawning at the same time and the
same location observed in lake sturgeon (Auer, 1999; Forsythe, 2010) can also influence
Ne and Ne/N. Lake sturgeon males mature at younger ages (12 -15 yrs.) and spawn at
shorter intervals (1-3 yrs.) compared to females (18-27 yrs. and 3-7 yrs., respectively)
(Auer, 1999; Bruch, 1999; Forsythe, 2010). Accordingly, many populations are malebiased and sex ratios can vary across years. In addition, the tendency of adults to spawn
at the same time and the same location, similar to other fish species (e.g., salmonids,
review in Doctor and Quinn 2009), can result in within-population structuring that
generally reduces Ne/N (Chesser et al., 1993; Caballero, 1994; Turner et al., 2002).
A small self-sustaining population of lake sturgeon in Black Lake, Michigan,
USA allows studies of Nb, Ne and the relationship between population effective size and
census size. A closed population helps avoid inaccuracy in Ne estimation due to
immigration (Waples, 2010). The Black Lake population has been isolated from gene
flow from other Great Lakes populations due to dam construction since 1903 (Baker and
Borgeson, 1999). The population has been well studied estimating adult numbers and
annual larval production as well as reproductive features including sex ratios and patterns
75

of spawning time and spawning location (DeHaan et al., 2006; Crossman, 2008;
Forsythe, 2010). The number of adults captured, when adjusted for inter-annual
variability in sampling effort, was similar across years. However, there was a high interannual variation in the number of larvae sampled, a surrogate of total larval production,
which varied by 40 fold over the 7 years of study. Long-term sampling of adults and
larvae in this population helps quantify demographic factors that affect inter-annual
variation in effective number of breeders. Moreover, large sample sizes of adults
collected across multiple years allows estimation of Ne and Ne/N, and comparison of the
estimates of Ne and Nb.
Our first objective was to estimate annual Nb of the lake sturgeon population in
Black Lake over 7 consecutive years in the current generation and qualitatively examine
the relationship of Nb and Nb/N to population demographic data. We proposed two
predictions of patterns of variation in Nb and Nb/N across 7 years. First, Nb and Nb/N of
the population were expected to be relatively stable and would not vary proportionally
with the inter-annual variability in larval collections. This prediction was based on
literature that Nb is sensitive to the number of census adults (Frankham, 1995; Nunney,
2

1996) but is expected to vary less due to standardized variance (e.g., variance/(mean) ) in
female fecundity (Nunney, 1996), and on our observation that the number of spawning
adults was relatively consistent across years. Alternatively, if inter-annual variation in
larval production reflected high standardized variance in reproductive success among
males and females (the number of offspring per individual), a large inter-annual variation
in standardized variance would result in fluctuation of Nb across years. The second
76

objective was to estimate contemporary Ne and compare values of Ne and annual Ne in
order to understand the relationship between Ne and Nb in a long-lived and iteroparous
species.
METHODS
Study site
The study was conducted in the Black Lake system in Michigan. Lake sturgeon
spawn only in the Upper Black River, the largest tributary to Black Lake (Smith and King
2005). Adults spawn in multiple locations over a period of 28-43 days (Forsythe, 2010),
extending from late April through early June. Spawning occurs within a shallow 1.5 kmsection of the river, about 9 km upstream from the river mouth (Smith and Baker, 2005;
Forsythe, 2010). The small, shallow and wadable river allows us to sample nearly all
spawning adults each year.
Sample collection
Adult and larvae samples were collected for 7 years (2001-2007, Table 1.1).
During each spawning season, we collected lake sturgeon adults daily by walking the
entire length of stream used for spawning one or more times per day and using longhandled nets to capture fish. Sex was determined by cloacal examination and extruding
gametes. Biological data including, body weight (kg), fork length (cm), and total length
2

were recorded along with capture date and location. A dorsal fin clip (~1 cm ) was taken
from each individual for genetic analysis to genetically assign parentage.
Larval sampling was conducted each year, starting 10 days after the first
spawning event and lasting for 25-40 days, until no larvae were captured for 2
consecutive nights. Five D-frame larval nets were spaced across the river channel at a site
77

about 2 km downstream from the spawning areas (Figure 1.1). We checked nets hourly,
from 2100 to 0200 hrs, when the vast majority of larvae dispersed (Auer and Baker,
2002). Sampling was consistent throughout the larval sampling period and among years.
In 2005 - 2007, larvae collected were transferred to a streamside hatchery where they
were reared for several months. Before the fish were released, a fin clip was collected
from each individual for genetic analysis. Fish that died in the hatchery were preserved in
95% ethanol. Samples used for microsatellite genotyping were taken proportionally from
the total preserved fin clips and mortalities from each year. The proportion of the total
larval catch sampled for microsatellite genotyping was not constant across years (Table
1.1). However, during each year the larvae genotyped represented proportionally a
random subset from the entire collection of larvae for a given year.
Genetic analysis
DNA was extracted from samples using the QIAGEN DNeasy(R) kit (QIAGEN
Inc.) and concentration was measured using a Nanodrop 1000 spectrophotometer
(Thermo Fisher Scientific Inc.). All samples were genotyped at 12 tetra-nucleotide
microsatellite loci including Spl 120 (McQuown et al., 2000); AfuG 68B (McQuown et
al., 2002). 2002); Aox 27 (King et al., 2001); AfuG 68, AfuG 9; AfuG 63, AfuG 74,
AfuG 112, AfuG 56, AfuG 160, AfuG 195 and AfuG 204 (Welsh et al., 2003). Using
polymerase chain reaction (PCR), 100 ng DNA was amplified in 25 µl reaction mixtures
as described in the above references. All PCR reactions were conducted using a
Robocyler 96 thermal cycler (Stratagene). Then, PCR products were run on 6%
denaturing polyacrylamide gels and visualized using a Hitachi FMBIO II scanner. Allele

78

sizes were scored independently by two experienced personnel. Genotyping errors
checked by blindly re-genotyping a random subset 10% of all samples.
Data analysis
Testing assumptions on marker loci used in the study
Most genetic marker-based methods (including linkage disequilibrium estimating
effective number of breeders and parentage analysis presented below) assume loci are
unlinked and under Hardy-Weinberg equilibrium (HWE). We tested these assumptions
for adult samples each year using program GENEPOP 4.0 (Raymond and Rousset, 1995;
Rousset, 2008). Deviations of observed genotypic frequencies from HWE were tested
using the exact test option. One (AfuG 68) of 12 loci departed from HWE, possibly due
to null alleles. We chose the log likelihood ratio statistic to test for linkage disequilibrium
among loci. No significant linkages among loci were detected.
Estimating effective number of breeders
Effective number of breeders was estimated using methods for calculating Ne
based on samples from the same cohort (Luikart et al., 2010). Then, Ne was replaced by
Nb, the number of breeders that produced progeny.
Performance of each Nb (and Ne) genetic estimator can vary among species and
markers used; therefore, using a combination of methods has been recommended (Luikart
et al., 2010; Waples, 2010). We used two methods: (1) linkage disequilibrium and (2)
estimation based on demographic data. The former estimates Nb based on genotype data
of offspring (single sample) without prior knowledge of parentage or demographic data.
Meanwhile, the latter requires information on the number of males and females that

79

produce offspring and estimates of mean and variance of the number of offspring per
individual (reproductive success), which was inferred based on genetically identified
parentage. Estimates of Nb based on demographic data may be overestimated if variance
in reproductive success is underestimated for natural populations (Turner et al., 2002).
Given the differences in methods used, similarities in Nb estimates based on two methods
(called Nb when based on LD and NbV when based on reproductive success) in each of 7
years should provide an indicator of reliability of each estimator.
To estimate Nb and calculate Nb/N each year (2001-2007), we used genotype data
from larvae randomly sampled from the entire larval collection each year (Table 1.1).
The number of adults captured each year (N) was adjusted based on catch per effort using
our records indicating that the effort (number of person-days sampling) in years 20042007 was approximately 1.5 times the effort expended in years 2001-2004. Therefore,
Nadjusted was estimated to be 1.5*Ncaptured in each of years 2001-2004.
Linkage disequilibrium (LD) method
The LD method of effective population size estimation has been the most
commonly used among single-sample estimators (England et al., 2006; Luikart et al.,
2010). Compared to the temporal method (two- sample estimator) that has been widely
applied, LD provides greater precision with microsatellite data (Waples, 2010). We used
the program LDNE (Waples and Do, 2008) to estimate the effective number of breeders
using multi-locus genotype data of larval lake sturgeon. LDNE was developed based on
the LD method proposed by Hill (1981) and adjusted for bias when the sample size is less
than the true unknown effective size (England et al., 2006; Waples, 2006; Waples and

80

Do, 2010). The method is based on the principle that in a finite population, genetic drift
generates the nonrandom association (linkage disequilibrium) among alleles at different
loci (Hill, 1981). This theoretical relationship between effective size and the correlation
of allele frequencies at pairs of loci is expressed as:
2

Nb = 1/[3(r -1/S)]

(1)
2

(Hill, 1981, modified by Waples, 2006), where r is a measure of linkage disequilibrium,
the mean squared correlation of allele frequencies at pairs of loci, and S is the number of
2

individuals sampled for genetic analysis. Several methods have been used to estimate r ,
in which Burrows’ Delta is the most commonly used (England et al., 2006; Waples,
2006; Waples and Do, 2010). LDNE uses the unbiased Burrow’s Delta estimator
2

(Waples, 2006) to compute r . The jackknife method implemented in LDNE used a
leave-one-out re-sampling procedure over loci to estimate 95% confidence interval for Nb
(Waples and Do, 2008).
Rare alleles can increase the precision but also upwardly bias Nb estimated using
the LD method due to the increased number of allelic combinations (Waples, 2006).
Simulations performed by Waples and Do (2010) have shown that a trade-off between
bias and precision of LD was realized when alleles with frequencies of less than 0.02
were excluded. We compared Nb estimates when alleles with frequencies less than 0.05,
0.02 and 0.01 were eliminated. We found that excluding alleles with frequencies less than
0.01 and 0.02 yielded similar estimates of Nb (the difference of the two estimates
averaged 1.4% compared to the estimates obtained with allele frequency ≤ 0.02), which

81

are 24.5% lower than Nb estimates when all alleles were included. Results of Nb reported
are based on LD estimates when allele frequencies less than 0.02 were excluded.
Estimation of Nb using reproductive success
To better interpret demographic factors (e.g., sex ratios and variance in reproductive
success) affecting Nb, we provided another estimate of Nb based on the mean and
variance in reproductive success as described by Crow and Denniston (1988) and
Caballero (1994). First, the effective numbers of males (Nbm) and females (Nbf) were
calculated as

kmNm – 1
Nbm =

(2)
km – 1 + Vm/km
kfNf – 1

and

Nbf =

(3)
kf – 1 + Vf/kf

respectively, where N is the number of males (Nm) and females (Nf) captured; k is the
mean number of offspring produced by males (km) or female (kf), and V is the variance in
the number of offspring for males (Vm) or females (Vf). The effective number of breeders
was then computed as:

Nb =

4NbmNbf

(4)

Nbm + Nbf

82

Mean and variance in reproductive success (the number of offspring assigned to each
parent) for males and females were estimated based on results from parentage assignment
using adult and offspring samples each of seven years 2001-2007.
We acknowledge that because larvae were sampled shortly after hatch (< 30 days,
Duong et al., in review) and given that significant mortality occurs between the time of
larval capture and the time larvae recruit to the adult breeding populations, estimates of
Nb for larval stages may differ from Nb that produce a cohort of adult recruits.
Parentage analysis - There are numerous programs that utilize multi-locus
genetic data to estimate parentage (e.g., CERVUS (Kalinowski et al., 2007); PASOS
(Duchesne et al., 2005); COLONY (Jones and Wang, 2010)), and the efficacy of
different programs or combinations of programs has been widely debated (Christie, 2010;
Jones et al., 2010; Walling et al., 2010). Use of multiple programs which are based on
different statistical properties to determine parentage has been advocated (Lee, 2008;
Jones et al., 2010). We used complimentary aspects of two programs, the Parentage
Allocation of Singles on Open Systems (PASOS) program, version 1.0 (Duchesne et al.,
2005) and CERVUS version 3.0 (Kalinowski et al., 2007) to conduct parentage analysis.
Output of putative parent-offspring allocations from the two programs were jointly used
to increase the allocation accuracy.
We used PASOS to estimate the proportion of adults captured and parentage
allocation correctness. PASOS can detect missing parents when they have not been
collected based on multi-locus genotypes of both parents and offspring. We set the
maximum number offset tolerance (MOT), the maximum number of offsets between a
parental and an offspring allele that PASOS accepts as possibly due to a scoring error

83

(Duchesne et al., 2005), of 2 and the error model (0.002, 0.008, 0.98, 0.008, 0.002) for
simulations in PASOS. We conducted simulations over five iterations of 1000 pseudooffspring to estimate the allocation correctness. Under the same MOT (MOT = 2) as
conducting simulations, we used the allocation function in PASOS to assign each
offspring to two collected parents and estimate the proportion of adults captured from the
natural population. The proportion of captured adults was used to parameterize analysis
in CERVUS program.
CERVUS is the most commonly used categorical parentage technique (Christie,
2010; Jones et al., 2010) and assigns progeny to a particular non-excluded parent based
on likelihood scores (LOD scores obtained by taking the natural log of the ratio of
likelihoods) derived from the genotypes of all offspring and parents sampled (Marshall et
al., 1998; Jones and Ardren, 2003; Kalinowski et al., 2007). Simulations conducted
within CERVUS to determine power and likelihood of assignment were based on
empirical estimates of population allele frequency, the proportion of adults captured
(PASOS output) and an empirical estimate of genotyping error (1.04%). Most likely
candidate parent pairs of offspring from the assignment output were accepted as “true
parent pairs” based on the criteria of ≥ 70% trio confidence with zero or 1 mismatch
between parents and offspring genotypes. Assignment outputs from the two programs
based on criteria above were then compared. Male parent-female parent-offspring triplets
that were consistent between the two programs were used for further analysis to
determine Nbm and Nbf.
The allocation correctness (the probability that each allocation is correct) from
PASOS was almost constant across 7 years, ranging 80 – 83%, and standard deviation
84

among iterations simulated was very small (~1%). Meanwhile, CERVUS program
provides a confidence level for each allocation based on user-define confidence levels. In
order to choose a confidence value for all 7 years of data, we compared the concordance
in assignment between PASOS and three confidence levels in CERVUS including “the
most likely”, 80 and 95% using parent and larval genotypes in 2001 as an example. The
result showed the “most likely” confidence provided the highest assignment rate with the
percentage of PASOS concordance as high as the 95% confidence. Therefore, we used
the criterion for other year data sets.
Relative reproductive success of males and females
Relative reproductive success used in this study was defined as the number of
offspring produced by each adult, which was the number of larvae genotyped and
assigned to each individual. Mean (km and kf) and variance (Vm and Vf) in reproductive
success among individuals were calculated for all adults captured in the same year.
Variation in Nb across years depends on differences in the magnitude of variance
compared to the mean of reproductive success among years (equation 2). Therefore, we
2

used standardized variance (variance/(mean) ) in reproductive success for males and
females to compare the magnitude of variance compared to the mean of reproductive
success among years. We tested the correlation of mean and variance in reproductive
success among years by using Spearman correlation (rho) because of small sample size (n
= 7). The tests were conducted using R (R-Development-Core-Team, 2009).

85

Estimating contemporary Ne
Relationships between Ne and Nb estimated using single-sample methods
including LD for a long-lived iteroparous species have rarely been investigated (Waples
and Do, 2010). Luikart et el. (2010) suggested that if samples included multi-age classes,
a LD-based estimate would provide values between Nb and Ne. Similarly, Waples and Do
(2010) proposed that if the number of age-classes represented in a sample is
approximately equal to the length of generation time, a LD-based estimate may
approximate Ne per generation. Our adult data meet this criterion. During 7 years (20012007), most adults in the population were captured, consisting of a wide range of ageclasses (15 – 52 yrs for males and 28-57 yrs for females, Figure 3.1, age was estimated
from age-length relationship provided by Baker, unpublished). The LD-based estimate
using genotypes of all adults captured in 2001-2007 (612 individuals genotyped) was
interpreted as Ne per generation. Because samples collected from a single year were large
(n =115 – 225, Table 1.1) and also contained the same range of age classes, single-year
samples could also be used to estimate Ne. The age range of adults captured in a single
year was as large as the entire adult sample captured in 7 years, the lower and upper
ranges for females 26 - 32 yrs. and 57 – 60 yrs., for males 13 – 17 yrs. and 51 – 59 yrs.,
respectively (based on age-length relationships provided by Baker, unpublished) . We
predicted that Ne estimated using adult samples from a single year would be similar to the
Ne estimated based on all adults over 7 years.

86

RESULTS
Demographic parameters
Demographic parameters including adult sex ratios and male and female
reproductive success were estimated to examine the relationship of these parameters and
effective sizes. Sex ratios (males: females) were male-biased in all years, ranging from
1.61 -3.01. Results from parentage analysis revealed that high proportions of males (5794%) and females (83-98%) over 7 years were assigned offspring (Table 3.1). Estimated
mean and variance in reproductive success of males and females were highly positively
correlated (rho = 0.89, P = 0.012 for females and rho = 0.85, P = 0.016 for males, n = 7),
implying that both mean and variance in reproductive success were higher in years with
higher larval collection compared to years with lower larval collection. Standardized
variance across years varied 5.8-fold (between 0.38 – 2.20) and 6.2-fold (between 0.49 –
3.08) for females and males, respectively. Meanwhile, larval production varied
approximately 40-fold across years (437 -16,417).
Annual estimates of Nb (2001-2007)
Annual Nb estimates based on the LD method ranged from 47 to 167 (95% CI 39
– 425, Figure 3.2). Nb estimated in 2002 was unusually high (mean: 167, 95% CI: 95425) compared to other years. Except for 2002, the ratios of Nb to census adult numbers
N were relatively stable across years and averaged 0.47 (0.31 – 65) (Table 3.2).
Estimates of effective number of breeders using demographic data (NbV) were close to
but consistently lower than Nb estimated based on LD, ranging from 41 (in 2005) to 106
(in 2007) (Table 3.2, Figure 3.2). The ratios NbV/N were relatively constant, ranging from
87

0.27 to 0.55. The small range of Nb/N obtained from the two methods indicated similar
patterns of Nb and N across years. N and Nb varied 1.5-fold and 3.5-fold, respectively,
during 7 years. Both methods consistently revealed that Nb and Nb/N was smallest in
2005 compared to the other years. In 2002, LD-based estimate of Nb (167) was much
higher than that from demographic method (86), leading to an unusual ratio of Nb/N (>
1).
Effective population size per generation
Based on genotypes of all unique adults (612 individuals) collected in 7 years, Ne
estimate per generation was 132 (95% CI: 104 – 167) (Table 3.3). This estimate is similar
to the harmonic mean (133) of Ne estimated based on adults captured each year from
2001-2007 (Table 3. 1). Surprisingly, the estimate of Ne was close to annual Nb values.
The harmonic mean of Nb of the current population (82) was 38% smaller than Ne (Ne =
132). Given the number of adults (N) in the Blake Lake population was estimated as
approximately 750 - 1,100 (Baker, unpublished), the ratio Ne/N was approximately 0.12 0.18.
DISCUSSION
Using long-term data and applying a single sample method to estimate the
effective number of breeders (Nb) and effective population size (Ne), we quantified interannual variability of Nb and Nb/N over 7 years for a closed lake sturgeon population
characterized by high inter-annual variation in larval production. The number of

88

spawning adults each year and Nb were relatively stable although the number of larvae
collected varied 40-fold across years. Therefore, stability of annual Nb/N may be
2

explained by low yearly estimates of standardized variance (variance/(mean) ) in relative
reproductive success of lake sturgeon adults.
Relationship of Nb and aspects of the species biology
Effects of demographic factors such as reproductive success, sex ratios and adult
census size on Nb can differ among species (Osborne et al., 2010). The inverse
relationship between variance in reproductive success and Nb has led to the hypothesis
that high female fecundity and concomitantly high reproductive variance can result in
proportionally lower Nb/N (Hedgecock et al., 1992; Hedrick, 2005). Alternatively, Nb/N
should vary as a function of the standardized variance in relative reproductive success
(Nunney, 1996). Small and consistent standardized variance in relative reproductive
success resulted in relatively constant Nb/N over 7 years in this lake sturgeon population.
Given the fluctuation in numbers of spawning adults across years was small, years of
higher larval production were characterized by higher relative reproductive success and
also higher variance in relative reproductive success compared to years of lower larval
production. Estimates of mean and variance in individual reproductive success inferred
from parentage may not capture actual inter-annual trends in lake sturgeon reproduction
because unequal proportions of larvae collected were genotyped in each year. However,
standardized variance, which is independent from the mean and thus is not influenced by
sample size, can represent inter-annual variation in reproductive success. In addition,
estimates of standardized variance in reproductive success based on parentage were not
89

expected to differ from actual estimates based on production. This prediction was based
on the high observed correlation between mean and variance in reproductive success, and
on our sampling design that randomly sampled with proportional numbers from all larvae
collected. Differences in standardized variance in relative reproductive success among
years were small, resulting in small inter-annual variation in Nb. Therefore, variation in
relative reproductive success across years was not predictive of temporal variation in
Nb/N. Nunney (1996) demonstrated that among three sources of variation in relative
reproductive success including individual differences, seasonal differences and agerelated variation in relative reproductive success, individual differences most affected Ne
(or Nb), and age-related variation in relative reproductive success had minor effects on
Ne. Standardized variance of individual differences should be greater than 10 to reduce
Nb/N <0.1, and that value is unusual in natural populations (Nunney, 1996). In natural
populations of many taxa, standardized variance in female fecundity averaged 0.441 ±
0.322 (Clutton-Brock 1988). Similar values of standardized variance in relative
reproductive success were observed in our lake sturgeon population.
Lake sturgeon female are highly fecund (mean: 323,684 eggs per fish, range:
53,672 - 460,270 eggs, range of weight 21.32 - 37.65 kg, n =14 (Bruch et al., 2006)).
However, eggs and larvae experience extremely high rates of mortality (Caroffino et al.,
2010; Forsythe, 2010), including the period between hatch and the time larvae were
sampled in our study. Accordingly, the species may be susceptible to a “Hedgecock
Effect” (Hedrick, 2005), a phenomenon whereby species characterized by high fecundity
and extremely high variance in reproductive success due to progeny surviving from a few
90

parents will result in small within-generation Ne and a low Ne/N ratio (Hedgecock et al.,
1992). However, our data showed most adults captured each year contributed offspring to
larval collections. Therefore, the “Hedgecock Effect”, a scenario observed in some
marine fish (Hedrick, 2005; Hauser and Carvalho, 2008) and invertebrate species
(Levitan, 2005) may not occur commonly in lake sturgeon until the larval stage. Given
the species has a long generation time (~ 20 years), the “Hedgecock Effect” may be
observed at sexual maturity if survival rates vary widely among families. Information on
inter-family variation in survival from the larval stage to the adult stage is not available
for lake sturgeon.
In addition to variance in relative reproductive success, highly skewed sex ratios
can reduce Nb/N (Frankham, 1995). For lake sturgeon, however, the moderate interannual variation in sex ratios observed over 7 years (1.61 -3.04) did not contribute to
temporal variation in Nb/N (Figure 3.3, right). The non-effect of observed sex ratios on
Nb/N could be explained based on the species mating system. Lake sturgeon exhibit
polygyny and polyandry. The polygamous mating system reduces variance in
reproductive success in both males and females, and thus maintains moderate Nb/N and
compensates for the effect of sex ratios on Nb/N (Nunney, 1996; Pearse and Anderson,
2009).
Reliability of Nb estimates for the lake sturgeon population in Black Lake
Comparisons of different estimators provide better understanding of how different
factors contribute to Nb (Waples and Do, 2010), and allow evaluation of reliability of

91

each method (Beebee, 2009).The LD and demographic methods used in our study
yielded comparable estimates and a relatively consistent pattern of inter-annual variation
in Nb, indicating the estimates were reliable. Small confidence intervals of Nb estimated
using LD method are also indicators of precision of the estimates (Beebee, 2009).
The ranges of Nb estimated were relatively small compared to other studies on
different species using LD temporal methods (e.g., Atlantic salmon Salmo solar (Palstra
et al., 2009); Pecos bluntnose shiner Notropis simus pecosensis (Osborne et al., 2010)).
However, in some cases the LD method yielded Ne estimates of infinity for several
populations, and therefore the results were difficult to interpret (Palstra et al., 2009;
Osborne et al., 2010). Ne estimated using the LD method may be biased when sample
sizes are smaller than Ne (England et al., 2006; Waples, 2006; Waples and Do, 2010). A
higher range of confidence interval for Nb in 2002 compared to the other years (from
2001-2007) could be attributed to smaller sample size. Except for 2001, sample sizes in
the other years were large (Table 1.1) enough to minimize bias in Nb estimates.
Ratios of the effective number of breeders to adult census size for this population
are in the range commonly found in other species. Nunney (1993, 1996) reported that
Ne/N commonly varied between 0.25 and 0.75. In salmonids species (e.g., steelhead
trout), Ne/N ranged 0.1-0.3 (Heath et al 2002) or 0.2-0.4 (Araki et al., 2007). Metaanalyses from numerous species by Frankam (1995) and recently by Palstra and Ruzzante
(2008) showed that median of Ne/N was 0.11 and 0.14, respectively. Reviews of
theoretical and empirical studies from different species and characterized by different
92

mating systems suggest that a lower threshold of Ne be greater than 50 and the Ne/N be in
the range 0.3 – 0.5 to buffer from genetic and demographic stochasticity (Frankham,
1995; Palstra and Ruzzante, 2008).
Contemporary Ne and the relationship between Ne and Nb
The LD estimator of contemporary Ne derived using adult samples provides an
empirical example applying Waples and Do’s (2010) suggestion that a LD-based estimate
using a mixed sample including the number of age-classes as equal to generation time can
be interpreted as Ne per generation. Luikart et al. (2010) also suggested that LD-estimates
using samples composed of multiple cohorts represent values between Nb and Ne.
Although the conjecture of these authors has not been quantitatively evaluated (Waples,
2010), our 7-year dataset including a larger range of age classes than the generation time
of lake sturgeon would be expected to produce a reliable estimate of Ne. In addition, the
similarity in age ranges of samples pooled across years versus yearly samples yielded
similar estimates of Ne based on pooled samples and the harmonic mean of Ne based on a
single year sample.
The relationship of Nb and Ne is complicated for long-lived and iteroparous
species (Waples, 2010). In this study, we have found that Ne was comparable to Nb of a
single larval cohort. We proposed two alternative predictions based on this result. First,
the survival of larvae from the time sampled to the time larvae reach sexual maturity is
likely very low and non-random among families. In this case, Nb estimated for a cohort of
mature progeny could be much smaller than Nb estimated using samples at the larval
93

stage. If Nb/N was assumed to be similar to the observed Ne/N, approximately 0.12, the
observed estimated Ne was smaller than the product of assumed Nb (~ 20) and generation
time, as suggested for semelparous species with constant census sizes (Waples, 1990).
Second, if mortalities were random among families, the correlation of allele frequencies
at pairs of loci (r) based on genotypes of larvae would not vary when larvae mature. As a
result, the observed Nb would be stable at different stages of a cohort. In sum, both
scenarios suggested that Ne be smaller than the product of Nb and generation time. In
other iteroparous species, (e.g., natterjack toad (Bufo calamita)), Nb was also found to be
similar to Ne in several populations (Beebee, 2009).
Implications
Our long-term study in this unique Black Lake system allowed us to quantify
annual Nb from progeny that survived to the larval stage and Ne of adults. We were also
able to evaluate the relationship of Nb with demographic parameters estimated
concurrently over 7 years. The study has important implications for the conservation of
lake sturgeon and other fish species. Although the current census size of the population is
moderately large (upper limit ~1,100, Baker, unpublished), the ratio of Ne per generation
to census size is low (0.12-0.18), suggesting that the population should be closely
monitored to ensure retention of diversity. Comparisons of Ne and Nb suggest that for
lake sturgeon and perhaps other long-lived, iteroparous species, Ne may not be
approximated as the product of Nb and generation time (Waples 2010) as has been

94

demonstrated for semelparous salmon (Ne = gNb, (Waples, 1990). Estimates of Nb in this
study are close to Ne. Therefore, a supplemental breeding program for sturgeon species
should not divide the target Ne into an annual target (e.g., if target Ne =200 and
generation time g=20, target Nb =10) as suggested previously (Kincaid, 1999; Welsh et
al., 2010)Because Nb is positively correlated with adult breeding number, maintaining a
sufficient number of adults each year will likely minimize effects of other factors
reducing Nb and thus maintain genetic diversity of lake sturgeon populations.

95

Table 3.1. Sex ratio and estimated individual reproductive success (RS) including mean,
variance and standardized variance (SDV) of females and males in 7 years 2001-2007
Year
2001
2002
2003
2004
2005
2006
2007

% adults assigned
Female
Male
90
91
76
94
80
90
71
88
58
83
65
98
94
89

Mean
7.27
2.82
5.90
4.48
4.79
3.31
13.98

Female RS
Variance
35.04
3.03
32.96
22.68
50.43
12.88
149.8

SDV
0.66
0.38
0.95
1.13
2.20
1.17
0.77

Mean
4.51
1.31
2.95
1.47
2.12
1.31
6.16

Male RS
Variance
10.05
1.36
11.79
1.99
13.88
1.99
27.05

SDV
0.49
0.79
1.36
0.91
3.08
1.16
0.71

Table 3.2. Effective number of breeders estimated using linkage disequilibrium (Nb) and
demographic methods (NbV), and the ratios of Nb and adult census size (Nb/N) in 20012007
Number of
Nb
NbV/N
Nb/N
NbV
Year
adults*
2001
173
76
63
0.44
0.37
2002
156
86
167
0.56
1.07
2003
182
57
69
0.32
0.38
2004
152
43
77
0.28
0.51
2005
153
41
47
0.27
0.31
2006
225
105
147
0.46
0.65
2007
206
106
126
0.51
0.61
(*) Census adult numbers in 2001-2004 (Table 1.1) were adjusted based on sampling
effort (Nadjusted = 1.5 *Ncaptured).
Table 3.3. Contemporary effective population size (mean and 95% confidence interval,
CI) estimated using linkage disequilibrium method based on adult genotypes from single
year and pooled unique individuals in 7 years (2001 -2007).
Year

Sample size

Mean

2001
2002
2003
2004
2005
2006
2007
All samples pooled

114
103
119
100
152
224
205
612

110
122
154
208
126
129
119
132
96

95% CI
73 – 186
80 – 220
93 – 330
112 – 754
83 – 220
92 – 193
88 – 170
104 - 167

Number of observations

60

Female

50

Male

40
30
20
10
0
110

120

130

140

150

160

170

180

190

200

Number of observations

Total length range (cm)
80
60
40
20
0
15

25

35

45

55

65

Age range

Figure 3.1. Total length (top) and age (bottom) distribution of male (n = 362) and female
(n = 249) lake sturgeon (Age data provided by Baker, unpublished).

97

Size (individuals)

500

Total adults adjusted
Nb

400

NbV
300
200
100
0
2000

2001

2002

2003

2004

2005

2006

2007

Year

Figure 3.2. Total number of captured adults adjusted for sampling effort and effective
number of breeders estimated using linkage disequilibrium (Nb) (bars show 95%
confidence interval) and demographic methods (NbV) in 2001-2007.

1.0

0.8

0.8

Nb/N

Nb/N

1.0

0.6
0.4

0.6
0.4

0

5000

10000

15000

1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0

Larval collection

Sex ratios

Figure 3.3. Plots showing relationships between the ratio of effective number of breeders
to adult census size (Nb/N) and larval collection (left) and adult sex ratios (right) in 7
years (2001-2007). The points circled are data from 2002.

98

Chapter 4
ENVIRONMENTAL AND MATERNAL EFFECTS ON EMBRYONIC
AND LARVAL DEVELOPMENTAL TIME UNTIL DISPERSAL
OF LAKE STURGEON

ABSTRACT- For migratory fish like lake sturgeon (Acipenser fulvescens), the period
from egg deposition through embryonic and larval development until dispersal
(ELDTUD) contributes substantially to variation in lifetime survival at the individual
level and to population levels of recruitment. Using genetically determined parentage, we
examined the relative importance of stream environmental variables (e.g., temperature
and discharge) and maternal effects including female body size, spawning time and
location to ELDTUD on an individual basis. Lake sturgeon adults (n = 208) spawning in
the Upper Black River (Michigan, USA), and larvae (n = 1 444) dispersing downstream
were captured during the 2007 spawning season. We used generalized mixed models and
multi-model inference based on Kullback-Leibler information-theoretic criteria to
demonstrate that environmental variables and maternal effects of female and spawning
time were both important predictors of ELDTUD. Decreasing ELDTUD during the
season resulted from linearly increasing temperature and nonlinearly decreasing river
discharge. Spawning time and individual females explained a large proportion of
variation in ELDTUD. The individual-based approach used in this study provided precise
estimates of ELDTUD and also facilitated the partitioning of variation in ELDTUD of
larvae produced by the same female and among females spawning at different times and
different environmental conditions.

99

INTRODUCTION
Estimating the timing of development and of transitions between ontogenic events
during early life stages and determining the relative importance of factors effecting
developmental time under natural conditions are important subjects in fish biology.
Embryonic and larval developmental time until dispersal (ELDTUD) is a critical period
that exposes larvae to high risks of mortality due to biotic (e.g., predation and
competition (Garvey et al., 1994; Paradis et al., 1996), and abiotic factors such as
temperature, oxygen (Chandler and Bjornn, 1988; Einum and Fleming, 2000)). ELDTUD
may vary as a function of environmental factors (e.g., temperature, stream discharge, etc.)
and maternal effects (e.g., egg size) (Gillooly et al., 2002; Kamler, 2002; O'Connor et al.,
2007). Environmental conditions during embryonic and larval development are typically
selected by parents (Trippel et al., 1997; Kamler, 2002; Jorgensen et al., 2008).
Therefore, maternal effects, which occur when female phenotype or environments
experienced at the time and location of spawning influence offspring phenotypic traits
(Mousseau and Fox, 1998), and together with environmental variables associated with
spawning time and spawning location can collectively contribute to embryonic and larval
developmental time (Einum and Fleming, 2000; Kamler, 2002) and dispersal (e.g., in
aquatic invertebrates, Reitzel et al., 2004; in fish, Edwards et al., 2007).
Water temperature and stream discharge are two environmental factors that play
an important role in the timing of larval hatch and dispersal (Heggberget, 1988; Pepin et
al., 1997) (Heggberget, 1988; Pepin et al., 1997; Tetzlaff et al., 2005) and thus
ELDTUD. Metabolic rate increases with temperature (Gillooly et al., 2001), which
decreases the time required for incubation and yolk absorption. Temperature-dependent

100

developmental time has been reported in a variety of taxa including insects (Pritchard et
al., 1996; Johnson et al., 2007; Arbab et al., 2008), fish and invertebrates (Gillooly et al.,
2002; O'Connor et al., 2007), and amphibians (Gillooly et al., 2002). River discharge
indirectly influences developmental time via oxygen supply (Kamler, 2002), and directly
affects dispersal time by changing drifting and swimming speeds of larvae (Elliott, 1987;
Fausch et al., 2001; Siegel et al., 2003).
ELDTUD may also vary due to maternal effects including spawning time and
location, and effects associated with female body size or age. Spawning at specific times
and locations has been shown to dictate conditions for offspring development and
survival (Trippel et al., 1997; Jorgensen et al., 2008). Empirical and modeling studies
have shown that adult spawning time and location also affect timing of larval emergence
and dispersal (Einum and Fleming, 2000; Reitzel et al., 2004; Edwards et al., 2007).
Maternal effects on egg size, where larger females usually produce larger eggs, have been
observed in fish (Chambers and Leggett, 1996; Heath et al., 1999; Heins et al., 2004b).
Larger eggs may require longer time for embryonic development (Pepin et al., 1997;
Gillooly and Dodson, 2000). In addition, for many species, genetic differences among
adult groups spawning at different times (isolation by time) may contribute to variation in
phenotypic traits of offspring (Hendry et al., 2005).
Studies that simultaneously evaluate environmental and maternal effects may
better explain variation in ELDTUD than studies that focus on effects of one set of
variables alone. There is a lack of research that simultaneously quantifies effects of
environmental factors and maternal effects on ELDTUD at the individual-level under
natural conditions. Reasons for the lack of data include difficulties in identifying

101

genealogical relationships between adults and larvae and the need to collect a large
number of parents and offspring from natural populations. The problem can be overcome
using genetic makers to genetically determine parentage (Garant and Kruuk, 2005;
Pemberton, 2008) of a population reproducing in natural and accessible habitats.
In this study, we used genetically determined parentage to examine effects of
environmental factors and maternal effects on ELDTUD in a threatened fish species, the
lake sturgeon (Acipenser fulvescens). Lake sturgeon exhibit an aggregate mating behavior
where eggs and sperm are released by spawning adults over rock and gravel without nest
preparation or parental care (Bruch and Binkowski, 2002). Because of the species mating
behavior and exposure of eggs and post-hatch larvae to environmental conditions, larval
lake sturgeon experience extremely high mortality early in life (Kempinger, 1988;
Forsythe, 2010). Therefore, selection of spawning times and locations, which determines
environmental conditions that affect eggs and larvae (Mousseau and Fox, 1998), by adult
lake sturgeon could play important roles in offspring survival and ELDTUD. Newlyhatched larvae generally remain in the stream substrate until yolk-sac reserves are
depleted, and then individuals disperse downstream in the current at night (Auer and
Baker, 2002; Kynard and Parker, 2005; Smith and King, 2005). However, factors
underlying variation in ELDTUD at the individual level are still unknown.
Similar to heritable spawning time observed in many salmonid species (Siitonen
and Gall, 1989; Gall and Neira, 2004) repeatability of spawning time among adult groups
in iteroparous species may have a genetic basis. Repeatability has been used as a measure
of trait heritability (Boake, 1989). Data collected on spawning lake sturgeon in the Black
River, Michigan (USA) over eight years (2001-2008) revealed that repeatability of

102

spawning time for females and males who have been captured more than two occasions
was high (0.56 and 0.42, respectively, Forsythe 2010). We also observed that earlyspawning lake sturgeon females produced offspring that had longer incubation times,
larger body size, and larger yolk-sac reserves at hatch compared to offspring of latespawning females (Crossman, 2008), which might lead to longer period from egg
deposition to dispersal (ELDTUD). If groups of lake sturgeon spawning at different times
are genetically differentiated and if spawning time affects ELDTUD, genetic factors
could contribute to differences in ELDTUD.
The main objective of this study was to evaluate the relative importance of
environmental factors (water temperature and discharge) and maternal effects (female
body size, spawning time and location) and their interactions to ELDTUD of lake
sturgeon under natural conditions. We also tested whether different adult groups whose
offspring differed in ELDTUD were genetically differentiated. Additionally, we
quantified the degree of temperature-dependence (cumulative thermal units or CTU) for
ELDTUD to evaluate the practicality of this single variable as a predictor of the timing of
larval dispersal for lake sturgeon.
METHODS
Study site
Our study was conducted in Upper Black River (UBR), the largest tributary to
Black Lake, Michigan, USA (latitude 45o 43’N, longitude 84o 15’W; Figure 1.1a). The
lake sturgeon population in Black Lake is isolated from other populations in adjacent
lakes by dams blocking immigration and emigration from Lake Huron (Smith and King,
2005). Adults spawn over a 1.5 km-section of UBR. This section can be divided into six

103

locations of spawning activity that were used across years (Figure 1.1b). Shallow
spawning areas (~1 m deep) and low turbidity allowed most adults to be observed and
captured (Crossman, 2008; Forsythe, 2010) and larvae dispersing from all spawning areas
to be collected (Smith and King, 2005).
Sample collection
Sampling for adults was conducted daily in 2007 by wading through the entire
length of the stream encompassing all spawning sites one or more times per day during
the entire spawning season. We captured spawning adults (143 males and 63 females)
using long-handled nets. Sex of adults was determined by extruding gametes, and all
individuals were measured for weight (kg) and fork length (cm). We also recorded date
and location of capture, which were assumed to be the date and location of spawning
based on our observation that lake sturgeon females spent only a few days on the
2

spawning grounds (Forsythe, 2010). A dorsal fin clip (~1 cm ) was taken for genetic
analysis.
Larval sampling was conducted at night when the vast majority of larvae disperse
(Auer and Baker, 2002). The larval sampling site was about 1.5 km downstream from the
spawning areas (Figure 1.1a). Five D-frame larval nets were evenly spaced across the
river channel (description in Smith and King, 2005) and were checked hourly from 21000200 hrs. Net locations remained consistent throughout the sampling period. Sampling
began 10 days after the first spawning event was observed and continued until there were
two consecutive nights with no larvae captured. Larvae were transferred to a stream-side
hatchery and were reared in different tanks by sampling night (n = 31) until individuals
were large enough for dorsal fin clips to be collected non-lethally. Mortality during the

104

rearing period was recorded and dead larvae were preserved in 95% ethanol by sampling
night. In total, 1 444 tissue samples from dead and surviving larvae were available for
genetic analysis. Based on the estimates of stream velocity and the proportions of total
stream width sampled, juveniles collected were estimated to represent approximately
13% of total larval production (Smith and King, 2005).
Environmental data
Water temperature and river discharge data were collected continuously
throughout the season. Water temperatures were recorded hourly using HOBO® data
loggers (Onset Computer Corp.) placed at spawning areas. UBR discharge was predicted
based on current discharge data obtained from the United States Geological Survey
(USGS) gauging station on the Pigeon River, a nearby tributary of Mullet Lake,
Michigan. Using UBR and Pigeon River discharge data during April - June for 50 years
(1950-2000, USGS National Streamflow Information Program), we found that daily
discharge of the UBR could be predicted as a linear function of daily discharge from the
2

Pigeon River (r = 0.68, F1, 4 056 = 9 355, P < 0.001). We used this relationship to
estimate UBR daily discharge during 2007 because there were no discharge data from the
UBR that year.
Genetic analysis
DNA was extracted from fin clips using the QIAGEN DNeasy(R) kit (QIA Inc.).
DNA concentration was measured using a Nanodrop 1000 spectrophotometer (Thermo
Fisher Scientific Inc.). All samples were genotyped at 12 tetra-nucleotide microsatellite
loci including Spl 120 (McQuown et al., 2000); AfuG 68B (McQuown et al., 2002).
2002); Aox 27 (King et al., 2001); AfuG 68, AfuG 9; AfuG 63, AfuG 74, AfuG 112,

105

AfuG 56, AfuG 160, AfuG 195 and AfuG 204 (Welsh et al., 2003). Using polymerase
chain reaction (PCR), 100 ng DNA was amplified in 25 µl reaction mixtures containing
2.5 µl of 10 X PCR buffer (1 M tris-HCl, 1.5 M MgCl2, 1 M KCl, 10% gelatin, 10% NP40, and 10% triton X); additions of 1 mM MgCl2 (1.5 mM MgCl2 for AfuG 9) for all
reactions but no additional MgCl2 for AfuG 63, Aox 27 and AfuG 74; 0.8 mM deoxynucleotide-triphosphate (dNTP); 10pmol fluorescently labeled forward and unlabeled
reverse primers and 0.5 U of Taq polymerase.
All PCR reactions were conducted using a Robocyler 96 thermal cycler
(Stratogene). The PCR conditions were 94oC for 2 minutes, followed by 30 cycles of 1
minute for primer-specific annealing temperatures (48oC for AfuG 9, AfuG 63 and AfuG
112; 50oC for AfuG 74; 53oC for Aox 27; 56oC for AfuG 68 and AfuG 68B; 58oC for
AfuG 56, AfuG 160 and AfuG 195; and 62oC for Spl 120 and AfuG 204), 72oC for 1
minute, and the final extension for 2.5 minutes at 72oC. PCR products were run on 6%
denaturing polyacrylamide gels and visualized using a Hitachi FMBIO II scanner. Allele
sizes were determined independently by two experienced personnel based on
commercially available standards (MapMarker

TM

, Bioventures Inc.) and samples of

known genotype. Errors in genotyping were empirically checked by blindly regenotyping a random 10% of all samples. Reported genotyping error is the ratio between
observed number of allelic differences and total number of alleles compared (Bonin et
al., 2004).

106

Parentage analysis
There are numerous programs that utilize multi-locus genetic data to estimate parentage
(e.g., CERVUS (Kalinowski et al., 2007); PASOS (Duchesne et al., 2005); COLONY
(Jones and Wang, 2010)), and the efficacy of use of different programs or combinations
of programs has been widely debated (Christie, 2010; Jones et al., 2010; Walling et al.,
2010). Use of multiple programs which are based on different statistical properties to
determine parentage has been advocated (Lee, 2008; Jones et al., 2010). We used
complimentary aspects of two programs, the Parentage Allocation of Singles on Open
Systems (PASOS) program, version 1.0 (Duchesne et al., 2005) and CERVUS version
3.0 (Kalinowski et al., 2007) to conduct parentage analysis. Output of putative parentoffspring allocations from the two programs were jointly used to increase the proportion
of offspring allocated to parents and allocation accuracy.
We used PASOS to estimate the proportion of adults captured and parentage
allocation correctness. PASOS can detect missing parents when they have not been
collected based on multi-locus genotypes of both parents and offspring. We set the
maximum number offset tolerance (MOT), the maximum number of offsets between a
parental and an offspring allele that PASOS accepts as possibly due to a scoring error
(Duchesne et al., 2005), of 1 and the error model (0, 0.01, 0.98, 0.01, 0) for simulations
in PASOS. We conducted simulations over 5 iterations of 1000 pseudo-offspring to
estimate the allocation correctness. Under the restricted MOT (MOT of 1 is more
restricted than MOT of 2), we used the allocation function in PASOS to assign each
offspring to two collected parents and estimate the proportion of adults captured from the

107

natural population. The proportion of captured adults was used to parameterize analysis
in CERVUS program.
CERVUS is the most commonly used categorical parentage technique (Jones et
al. 2010; Christie 2010) and assigns progeny to a particular non-excluded parent based on
likelihood scores (LOD scores obtained by taking the natural log of the ratio of
likelihoods) derived from the genotypes of all offspring and parents sampled (Marshall et
al., 1998; Jones and Ardren, 2003; Kalinowski et al., 2007). Simulations conducted
within CERVUS to determine power and likelihood of assignment were based on
empirical estimates of population allele frequency, the proportion of adults captured
(PASOS output) and an empirical estimate of genotyping error (1.04%). Most likely
candidate parent pairs of offspring from the assignment output were accepted as “true
parent pairs” based on the criteria of ≥ 65% trio confidence with zero or 1 mismatch
between parents and offspring genotypes. Assignment outputs from the two programs
based on criteria above were then compared. Each male parent-female parent-offspring
triplet that was inconsistent between the two programs was evaluated based on biological
information (e.g., proximity of captured time and captured location of both parents).
Genetic differences and relatedness among spawning groups
Repeatability in spawning time and spawning location for individual lake
sturgeon males and females that have been captured 2 or more times (Forsythe, 2010)
suggests the possibility of genetic differentiation among spawning groups within the
UBR population, which may contribute to differences in ELDTUD. Due to higher
repeatability in individual spawning time than in spawning location (0.56 and 0.42 vs.
0.16 and 0.04 for females and males, respectively, Forsythe, 2010), we predicted that

108

genetic differences could be detectable among members of different adult groups
spawning at different times but not among groups spawning at different locations. We
first estimated genetic differentiation (FST) among the three temporal spawning groups
(early, middle and late) and among six spatial spawning groups of adults using the
method of Weir and Cockerham (1984) implemented in the program FSTAT 2.9.3
(Goudet, 2001). If adult groups could be genetically differentiated, we then compared
distributions and mean of inter-individual relatedness within and between genetically
different groups. Moment-based estimators for pair-wise relatedness, rxy (Queller and
Goodnight, 1989) based on 12 microsatellite-locus genotype were obtained using
KINGROUP v.2 program (Konovalov et al., 2004). Nonparametric KolmogorovSmirnov tests were used to compare distributions of rxy among adult groups. A
permutation approach was used to test for difference in mean relatedness within and
among spawning groups using a SAS-based program (A. Saxton, University of
Tennessee, unpublished software) (Ratnayeke et al., 2002). We also report allelic
diversity, heterozygosity, probabilities of false parental exclusion, and tests (with
Bonferroni correction) for deviations from Hardy-Weinberg equilibrium using CERVUS.
Data analysis
We defined the embryonic and larval developmental time until dispersal
(ELDTUD) as the number of days elapsed between the day of female capture and the day
of offspring capture as the larvae dispersed downstream from the spawning site (Figure
1.1a). Accordingly, ELDTUD included periods of incubation, yolk absorption in the
stream substrate, and migration over a 2-km distance.

109

We calculated cumulative thermal units (CTU) and examined the importance of
factors affecting ELDTUD. CTU (degree-day in Celsius) was calculated using the
method of Kempinger (1988). CTU for each larvae was estimated as the sum of daily
temperature (adjusted by substracting a constant 5.8 oC) during the period from egg
deposition until dispersal.
We used analysis of variance (ANOVA) and pairwise t tests with unequal variance
to test CTU differences among offspring produced by females that spawned at different
times (“early”, “middle” and “late”). Females were grouped into three categories based
on observed discontinuities in spawning dates among adult groups. Difference in CTU
among larval groups would indicate other factors beside temperature affecting ELDTUD
(Kamler, 2002).
We used generalized linear and nonlinear mixed effect models to simultaneously
examine environmental factors and maternal effects affecting ELDTUD. Environmental
factors include mean daily water temperature and daily river discharge that larvae
experienced during the period from egg deposition until capture. We defined maternal
effects to include female spawning date (Julian date), spawning location (n = 6, Figure
1.1b) and body size (i.e., fork length) and a random effect of females. Females were
treated as a random effect to account for non-independence among offspring of the same
female.
We initially fit data to the full model including fixed effects of water temperature
(Temj), river discharge (Disl), Julian spawning date (Datem), spawning location (Locn)
and fork length (Leno) and possible two-way interactions of fixed effects and random
effect of females (Femu) using generalized linear mixed effect models. Two-way
110

interactions of fixed effects were examined based on biological relevance as
recommended by Burnham and Anderson (2002). Accordingly, the effect of female body
size on ELDTUD would not be expected to vary as a function of the time or location of
spawning. Therefore, we excluded interactions with female body size from the initial full
model. We also tested for multi-collinearity among independent variables of the model
by examining the variance inflation factor (VIF) from the linear model with all fixed
effects without interactions and the Pearson correlation of variable pairs (Graham, 2003).
There was evidence of multi-collinearity between water temperature and Julian date (VIF
of temperature and Julian date was 10.3 and 7.8, respectively; Pearson correlation, r =
0.91, n = 1154, P < 0.001). To isolate confounded effects of these two explanatory
variables, we first assumed one variable was more explanatory of observed variance. We
then replaced the less important variable by its residuals from the regression against the
more important variable (Graham, 1997, 2003); review in (Heikkinen et al., 2006). We
reasoned that Julian date representing the conditions including temperature at spawning
would be a more important explanatory variable than mean water temperature. The
unique contribution of Julian date was disassociated from the shared effects of
temperature by adding temperature residuals (T.resj) from regression of water
temperature (dependent variable) against Julian date (independent variable) in the full
model. We evaluated normality of the data by examining residual plots of response and
predicted variables. Based on the plot showing the nonlinear relationship between
ELDTUD and river discharge, the quadratic term of river discharge was treated as a fixed
effect and was also added to the model:

111

2

ELDTUD = µ +T.resj + Disl + Dis l + Datem + Locn + Leno + Dis*Datelm + Dis*Locln
+ Date*Locmn + Femu + εijlmnou

(1)

where µ is the overall mean, (*) represents the interaction of two variables; and εijlmnou is
th

the residual error for each larva i . In the above model, the random effect of females
(Femu) was assumed to be normally distributed.
Model selection for the best model from a set of all possible combinations of
environmental factors and maternal effects in (1) was based on Akaike Information
Criteria (AIC) (Burnham and Anderson, 2002). The inclusion of females as a random
effect in the full model was examined first, followed by selecting fixed effects of
interactions and main effects (Ngo and Brand, 1997). The random effect was included if
the AIC difference (∆AIC) of the full model without and with random effect of females
was greater than 4 indicating the model without random effect was less supported by the
data (Burnham and Anderson, 2002). We used restricted maximum likelihood (REML) to
fit models differing by the presence of the random effect and maximum likelihood (ML)
for competing models differing in fixed effects (Pinheiro and Bates, 2000). The relative
importance of environmental factors or maternal effects in ELDTUD was assessed based
th

on ∆AIC (∆i =AIC of the i model - smallest AIC) and Akaike weights (wi = exp (-∆i/2)).
Variables with higher sum wi (importance weight) from all models containing the
variables in question are more important (Burnham and Anderson, 2002). We also used a
confidence set for Kullback-Leibler best models (Burnham and Anderson, 2002) to
provide the best explanation of variation in ELDTUD. The 95% confidence set represents
112

the subset of models with the sum of wi ≥ 0.95. All statistical analyses were conducted
using R (R-Development-Core-Team, 2009).
RESULTS
Adult capture, larval collection and parentage analysis
The 2007 spawning season extended from 23 April to 1 June. Timing of spawning
activity was multimodal with three distinct peaks in spawning (Figure 4.1a). More
individuals were observed spawning at the beginning (“early” adults, n = 108 individuals)
and middle of May (“middle” adults, n = 74) as opposed to later in May (“late” adults, n
= 24). Larvae dispersed in two peaks 15 to 20 days following adult spawning (Figure
o

4.1b). Water temperatures increased throughout the spawning season, from 10.6 C to
o

3

22 C. River discharge (range 5.71 to 9.61 m /s) was higher and more variable during
adult spawning than during the larval dispersal period.
Estimates of allele diversity (range from 2 to 11, average 5.3 alleles per locus) and
expected heterozygosity (0.59) of lake sturgeon adults were moderate. Genotype
frequencies from 11 of 12 loci conformed to Hardy Weinberg expectations for all adults
sampled.
Of the total number of captured adults (143 males and 63 females) PASOS
estimated that 89.2% of females and 87.2% of males contributed to the larvae collected.
The assignment rate of offspring to two collected parents was 74.7% (1 079 parentsoffspring allocations) and mean allocation correctness (± standard deviation) was
estimated to be 79.1 ± 0.6%. The assignment rate from CERVUS was 85.5% (1 234
allocations). Comparison of parent-offspring triplets assigned from the two programs

113

revealed assignment consistency to be 78.6%. Exclusion of larvae with allocation
discrepancies between the two programs did not affect results (data not shown). Each
parent-offspring triplet in PASOS parental allocations was also one of the most likely
triplets with positive LOD scores found in CERVUS. Based on concordant assignments
from both programs, 1 154 offspring (79.9% of larvae collected) were assigned to
collected parents including 137 male parents (94.5% of males captured) and 58 female
parents (92.1% of females captured). The average number of offspring collected per male
was 8.4 ± 6.5 and per female was 19.9 ± 14.3.
Embryonic and larval developmental time until dispersal
ELDTUD varied from 4-36 days (mean ± SD = 20.3 ± 5.7, n = 1 154). ELDTUD
of larvae from the same female parent (the time period from the first to the last offspring
collected) was 11.6 ± 4.0 days (n = 58 female parents). ELDTUD was significantly
different among larvae whose female parents spawned during different segments (early,
middle or late) of the breeding season (F2,96.3 = 655, P < 0.001) (Table 4.1). Cumulative
thermal units (CTU) averaged 201 ± 53 degree - days and also differed among larval
groups (Table 4.1).
Certain combinations of both environmental variables and maternal effects better
explained variability in ELDTUD than did variables within only one or the other of the
variable categories (Table 4.2). ELDTUD was best explained by model 8 (wi = 0.66;
Table 4.2) including the random effect of females and fixed effects of temperature
residuals, river discharge, quadratic effect of discharge, Julian spawning date, spawning
location, and the discharge and spawning date interaction. Fixed effects in the model
explained most of variation in ELDTUD (88.0%). Of the remaining variation
114

unexplained by the fixed effects, the random effect representing differences among
females (variance 8.77, confidence intervals, CI = 6.05 – 12.71) accounted for 94.5% (CI
= 92.8% – 95.8%), which is much greater than differences within females (i.e., residual
variance) in average ELDTUD.
Based on Kullback-Leibler criteria, river discharge and spawning date associated
with temperature were equally important predictors of ELDTUD. Importance weights for
all these factors (sum of wi for models containing these factors) were close to 1 (Table
4.2). ELDTUD was non-linearly related to river discharge. Coefficient estimates of the
linear term (slope ± SE = - 143 ± 3.0) and quadratic term (slope ± SE = 2.58 ± 0.16) of
river discharge in the final model (Table 4.3) indicate the nonlinear decrease in ELDTUD
with increasing river discharge. Larvae produced by females spawning early in the season
when water temperature was comparatively low (14 – 15oC) dispersed after longer
periods of time compared to larvae produced by females spawning later in the year (slope
± SE = -6.40 ± 0.12, t1, 56 = - 51.9, P < 0.001). However, the magnitude of effects of
spawning date and river discharge on ELDTUD depended on the interaction between
these variables (Table 4.3). Spawning date effects on ELDTUD included the effect of
water temperature (Pearson correlation, r = 0.91, n = 1154, P < 0.001). Julian date alone
explained 55.8% of the variation in ELDTUD, compared to 20.9% variation explained by
2

water temperature (based on coefficient of determination, r , from linear models of
ELDTUD with each predictor of Julian date and temperature, respectively). Nevertheless,
temperature residuals increased model fit. The full model with temperature residuals had
an AIC score lower compared to the same model without temperature residuals

115

(difference in AIC = 6), indicating the importance of temperature when its effect was
disassociated from the shared effect of spawning date on ELDTUD.
Spawning location and female body size were less important predictors of
ELDTUD (importance weights were 0.30 and 0.15, respectively; Table 4.2). When the
fixed effect of spawning date and random effect of females were included in the model to
account for variation in ELDTUD of larvae from different families, spawning location
and female body size on ELDTUD no longer improved model predicting (comparing AIC
values of model 9 and 10, Table 4.2).
The confidence set for Kullback-Leibler best models included models 8, 7 and 5
(sum wi = 0.95). These nested models supported the hypothesis that a combination of
environmental factors and maternal effects best explained variation in ELDTUD. There
was no evidence (wi ~ 0) supporting the hypotheses that either environmental factors or
reproductive variables alone provided the best prediction of ELDTUD of lake sturgeon.
Genetic differences and relatedness among spawning groups
Genetic differences (FST) were not significant among females and adult groups
spawning at different locations, (95% confidence interval of overall FST [bootstrapping
over loci] for only females ranged from 0.000 to 0.016, and for all adults from 0.000 to
0.002). Genetic differences were observed among adult groups (males and females)
spawning at different times but not among different spawning groups of female. Pairwise
FST between early and middle adults groups (0.007) was significant (alpha = 0.05, after
standard Bonferroni corrections, P-value adjusted = 0.016). However, no genetic
differences were observed between early-late (FST = 0.002) and middle-late (FST =
116

0.006) adult groups (P = 0.17 and 0.20, respectively). Mean inter-individual relatedness
(rxy) among adults of the middle spawning group (mean ± SD; 0.034 ± 0.213) was
significantly higher than rxy among members of the early spawning group (-0.015 ±
0.223) and late spawning group (-0.017 ± 0.209). Distributions of cumulative expected
frequencies of rxy values were significantly different between middle adult group
compared to early and late groups (differences in distribution, Kolmogorov-Smirnov
tests, 9.8 and 12.1%, respectively, P < 0.001). No differences in mean (P = 0.45) and
distribution (P = 0.81) of rxy were observed between early and late groups. Mean rxy
estimated within and between groups of adults were also not significantly different (P =
0.08).
DISCUSSION
Using individual-based measures of embryonic and larval developmental time
until dispersal (ELDTUD) based on genetic determination of parentage, we quantified
variation in offspring ELDTUD within and among females and evaluated the relative
contributions of environmental and maternal factors affecting ELDTUD. Variation in
ELDTUD was lower among larvae produced by the same female relative to those
produced by different females that spawned during the same period of the spawning
season (early, middle and late groups). Similarly, variation in ELDTUD was lower among
larvae produced by females spawning within the same group relative to those produced
by females that spawned at different times.
Non-genetic components of maternal effects (e.g., spawning time and spawning
location) confounded by environmental factors affected ELDTUD. Using statistical

117

approaches to account for correlations among variables based on Kullback-Leibler multimodel inference (Burnham and Anderson, 2002; Johnson and Omland, 2004), we
disassociated confounding effects due to colinearity among predictor variables. Maternal
effects of female and spawning date, and environmental factors (water temperature and
discharge) were in the confidence set for Kullback-Leibler best models indicating that
these variables were of comparable importance to ELDTUD of larval lake sturgeon.
Temperatures during spawning and embryonic and larval development can affect
offspring traits, including body size (Fox and Czesak, 2000) and developmental time
(Gillooly and Dodson, 2000; Stillwell and Fox, 2005). Although temperature impacts
larval developmental time in many species (Gillooly et al., 2001; O'Connor et al., 2007)
inter-individual variation in developmental time may be substantial. In our study, mean
daily temperature only explained 20.9% of variation in ELDTUD of larval lake sturgeon.
The degree of temperature-dependency of ELDTUD could differ during the three
consecutive developmental periods represented by this composite variable including (i)
embryonic development, (ii) yolk sac absorption; and (iii) emergence and dispersal from
spawning areas. The first two periods are likely more dependent on water temperature
(Wang et al., 1985; Pepin et al., 1997; Kamler, 2002). The third period may depend more
on other environmental factors (e.g., river discharge, food, and predators), larval age or
size (Elliott, 1987; Day and Rowe, 2002) and behavior (Shanks, 2009), lunar cycle effects
on larval concealment (Hernandez-Leon, 2008), and female spawning behavior (Copp et
al., 2002; Hogan and Mora, 2005; Shanks, 2009).
Another environmental factor, river discharge, has been shown to affect timing of
larval emergence (Fausch et al., 2001) and movements (Elliott, 1987; Siegel et al., 2003),

118

and therefore was predicted to affect ELDTUD. Fausch et al. (2001) found that
emergence of rainbow trout (Oncorhychus mykiss) fry occurred during the period of high
water velocity. After emergence, larvae can be dislodged from substrate by strong
currents (salmonids (Elliott, 1987); pallid sturgeon Scaphirhynchus albus, (Kynard et al.,
2007)). However, for some species such as reef fishes, larvae can actively adjust their
swimming speed. At high river discharge, larvae swim more slowly (Hogan and Mora
2005). Similarly, lake sturgeon larvae might exhibit considerable behavioral plasticity by
remaining in substrates longer during times of high river discharge.
The effect of female spawning date on ELDTUD could be due to different
environmental conditions (e.g., water temperature and discharge) at spawning and also
different adult groups. In many fish species, adult groups spawning early or late in the
season might differ in maternal effects (Chambers, 1997; Champer, 1997; Einum and
Fleming, 2000). This is also the case for lake sturgeon, evidenced by highly significant
differences in ELDTUD among the three larval groups (Table 4.1 and Figure 4.2) and
greater among-female than within female variation in ELDTUD (Figure 4.3). Adult
groups that spawned at different times exhibited the low level of genetic differentiation,
providing a certain degree of support for the hypothesis that maternal effects on
ELDTUD could be due in part to genetic factors. Higher FST among temporal spawning
groups than spatial spawning groups is consistent with stronger effects of spawning time
compared to spawning location on ELDTUD, which provides further evidence of a
genetic component that may partly contribute to differences in ELDTUD. However,
given the low variance in frequency of alleles at the microsatellite loci used, gene flow
among members of different spawning groups is likely. Further investigation of the

119

concordance among measures of inter-group variance in additive and neutral genetic
traits is warranted.
The importance of spawning date in determining larval developmental and
dispersal time observed in lake sturgeon was also reported in marine fish and
invertebrates (Reitzel et al., 2004). Edwards et al. (2007) used 2-dimensional dispersal
kernels to examine factors affecting larval dispersal and found that spawning time and
location might be more important than larval behavior in determining larval dispersal
time. In our study, the effect of spawning location on larval dispersal time was less
important than spawning date. The relatively small spawning areas of the lake sturgeon
population covering 1.5 km of UBR may explain comparatively minor effects of
spawning locations on ELDTUD.
Incorporating both ecological and genetic data in this study provided a useful tool
to explore degrees of, and factors contributing to inter-individual variation in ELDTUD
of lake sturgeon. Individual-level measurement of ELDTUD based on the knowledge of
parentage was more informative compared to population-based observations. For
example, our results of CTU influences on ELDTUD were different from those reported
by Smith and King (2005) for the same lake sturgeon population. Smith and King (2005)
calculated CTU based on observations from the day of adult spawning to the day of peak
larval drift. They reported that CTU for two or three larval groups from each spawning
season from the years 2000 through 2002 ranged from 136.2 – 181.2. In their study, CTU
and time to dispersal of the early larval group were lower than for the late larval group in
the same year, which is opposite of our findings. Water temperature is generally colder
early in the spawning season. Consequently, the time from egg deposition to dispersal of

120

larvae would be expected to be longer due to temperature-dependent development
(O'Connor et al., 2007). Importantly, population-based approaches could not detect
important relationships between developmental time at early life stages and spawning
behavior of individual female parents. Population-based estimates of CTU might
introduce considerable bias in estimating ELDTUD in the absence of genetic data that
conclusively ties larvae to a specific date and location of a spawning event.
Several factors should be considered when interpreting the data. First, the
characterizations of spawning time and spawning locations were based on the time and
location of adult capture. This assumption could lead to decreased precision in calculating
ELDTUD. However, daily surveys of the spawning sites suggest that few females spent
more than 2 or 3 days in the spawning areas (Forsythe, 2010), indicating minor effects on
ELDTUD given the duration of spawning activity. Second, data were collected from a
single year, while inter-annual variation in environmental conditions can be large and
different groups of adults spawn in different years (Forsythe, 2010). Based on Akaike
weights of the best model set, we believe that the relative importance of the variables in
our model will not vary substantially, though ELDTUD observed as a function of these
variables may vary depending on inter-annual variation in environmental conditions
realized each spawning season. Finally, the dispersing larvae were collected within 2 km
of the spawning area and numbers captured during this critical life stage may not reflect
the numbers of larvae that disperse over different ELDTUD periods and survive.
Future research on costs of timing of dispersal is necessary and can be
accommodated using genetic methodologies as employed in this study. Rates of mortality
may vary as a function of larval body size, whereby selection acts to select individuals

121

that disperse at different times and concomitantly at different sizes (Sogard, 1997).
Predation acting on different size classes of juveniles is likely a significant source of
mortality (Mittelbach and Persson, 1998). Spatially complex rearing habitats are expected
to contribute to variation in larval growth and dispersal time. Costs of dispersal are
expected to be high, and can increase if larvae disperse at certain times or developmental
stages in response to stream conditions. Further analyses are warranted to address the
costs of dispersal time in terms of probability of mortality during this period.
Our findings have important implications for early life history studies and for
species conservation. CTU and timing of adult spawning events should be simultaneously
taken into account when predicting larval ELDTUD. Individual-based estimates revealed
different levels of variation in ELDTUD within and between female families and among
female groups. Such variation likely provides conditions for natural selection to act and
may be of importance for retention of levels of viability at the population level. Effects of
environmental factors did not outweigh maternal effects including spawning time on
ELDTUD. Accordingly, we suggest that the maternal effects may counter the effects of
the environmental variables on development, performance, and survival of larval lake
sturgeon as well as other fish species.

122

Table 4.1. Mean (± 1 standard deviation) temperature at spawning and daily temperature
over the embryonic and larval developmental time until dispersal (ELDTUD) and
cumulative thermal units (CTU) of three larval groups illustrated in Figure 4.2.
Larval group
Number of
Spawning
Mean daily ELDTUD CTU
(Female
observations temperature temperature (days)
(degreespawning date)
(oC)
(oC)
days)
Early
792
12.9 ± 1.0
15.4 ± 0.5
22.7 ± 4.6 219 ± 48
(4/25 – 5/6)
Middle
329
16.9 ± 0.8
16.6 ± 0.3
15.4 ± 3.4 168 ± 41
(5/10 – 5/13)
Late
33
15.9 ± 1.3
18.3 ± 0.7
9.24 ± 2.7 114 ± 33
(5/20 – 5/27)
Note: All variables differed significantly among periods (P < 0.001).

123

Table 4.2. Model structure describing embryonic and larval developmental time until dispersal (ELDTUD) and model selection
criteria based on Akaike Information Criteria (AIC), AIC difference (∆AIC) and Akaike weight (wi).
No.
1.
2.
3.
4.
5.
6.
7.
8.
9.

ELDTUD model

k

Including only maternal effects
10. Datem + Locn + Leno + Date*Locmn + Femu
11. Datem + Femu
Including only environmental variables
12. Temj + Disl + Dis2l + Tem*Disjl
13. Temj + Disl + Tem*Disjl

24

14
13
13

2 868
4 678
3 061
2 849
2 844
3 980
2 844

0.00
0.00
0.00
0.02
0.15
0.00
0.15

2 841
2 847

27
1 837
220
7
3
1 139
3
0
6

6 249
6 232

3 408
3 391

0.00
0.00

6

2

T.resj + Disl + Dis l + Datem + Dis*Datelm + Femu
2
Disl + Dis l + Datem + Dis*Datelm + Femu

wi

15
4

2

T.resj + Disl + Dis l + Datem+ Locn + Leno + Dis*Datelm + Femu
2
T.resj + Disl + Dis l + Datem + Locn + Leno + Femu
2
T.resj + Disl + Dis l + Datem + Locn + Dis*Datelm + Femu

∆ AIC

8
7

Including environmental and maternal effects
2
T.resj + Disl + Dis l + Datem + Locn + Leno + Dis*Datelm + Dis*Locln + Date*Locmn + Femu
2
T.resj + Disl + Dis l + Datem+ Locn + Leno + Dis*Datelm + Dis*Locln + Date*Locmn
T.resj + Disl + Datem + Locn + Leno + Dis*Datelm + Dis*Locln + Date*Locmn + Femu
2
T.resj + Disl + Dis l + Datem + Locn + Leno + Dis*Datelm + Date*Locmn + Femu

AIC

6 310

3 469

0.00

23
23
19

0.66
0.02

5 6 309 3 468 0.00
Note: Models include different combinations of fixed effects of temperature residuals (T.resj, from temperature [Temj] regression
against Julian date [Datem]), river discharge (Disl), location (Locn), fork length (Leno), and interactions of discharge-Julian date
(Dis*Datelm), discharge-location (Dis*Locln) and Julian date-location (Date*Locmn), and random effect of females (Femu). k =
total number of parameters including intercept and error terms. Model 1 and 2 was fit by restricted maximum likelihood (REML)
for testing the importance of random effect.
124

Table 4.3. Coefficient estimates of fixed effects in the best model (model 8 in Table 4.2)
representing relationships between model parameters and embryonic and larval
developmental time until dispersal (ELDTUD).
Coefficient
a
b
SE
denDF
Parameters
estimate
t-value
P-value
Intercept

1 072

18.6

1 092

57.6

<0.001

Temperature residual

-0.43

0.16

1 092

-2.62

<0.01

Discharge

-143

3.0

1 092

-48.1

<0.001

Discharge

2.58

0.16

1 092

16.06

<0.001

Julian date

-6.40

0.12

56

-51.87

<0.001

Discharge*Julian date

0.72

0.02

1 092

44.08

<0.001

2

a

Standard error
Denominator degrees of freedom. The numerator degrees of freedom for all parameters
are equal to 1.
b

125

Females
Temperature

Males
Discharge

35

25

Number of adults

30

20

25
20

15

15

10

10
5

5
0
22 April 29 April

Temperature (oC)
Discharge (m3/s)

(a)

0
6 May

13 May 20 May 27 May

Capture date
(b)
250

Early
Late
Discharge

Middle
Collected larvae
Temperature

25

150

15

100

10
5

0
15 May

Temperature (oC)
Discharge (m3/s)

20

50

Number of larvae

200

0
22 May

29 May

Night of larval sampling
Figure 4. 1. Number of adults captured (a) and number of larvae collected and those
assigned to early, middle or late female groups (b), together with mean daily water
temperature and river discharge during the periods from egg deposition until larval
dispersal.

126

40
ELDTUD (days)

35
30
25
20
15
10
5
0
24 April

1 May

8 May

15 May

22 May

29 May

Female spawning date
Figure 4.2. Embryonic and larval developmental time until dispersal (ELDTUD) of three
groups of larvae whose maternal parents spawned during “early” (circle), “middle”
(diamond) or “late” (triangle ) periods of the spawning season. Darkened symbols
indicate more than one offspring characterized by the same ELDTUD and female
spawning date.
35

ELDTUD (days)

30
25
20
15
10
5
0
0

10

20

30

40

50

60

Individual female
Figure 4.3. Mean and standard deviation of embryonic and larval developmental time
until dispersal (ELDTUD) among siblings from the same female parent. Individual
females were ordered left to right by spawning date.

127

Chapter 5
RELATIVE LARVAL LOSS AMONG FEMALES DURING DISPERSAL
OF LARVAL LAKE STURGEON

ABSTRACT - Mortality that occurs during larval dispersal as a consequence of
environmental, maternal, and genetic effects and their interactions can affect annual
recruitment in fish populations. We studied larval lake sturgeon (Acipenser fulvescens)
drift to examine whether larvae from different females exposed to the same
environmental conditions during dispersal differed in relative levels of mortality. We
estimated proportional contributions of females to larval collections and relative larval
loss among females as larvae dispersed downstream between two sampling sites during
two consecutive nights based on genetically determined parentage. Larval collections
were composed of unequal proportions of offspring from different females that spawned
at upstream and downstream locations (~0.8 km apart). Hourly dispersal patterns of
larvae produced from females spawning at both locations were similar, with the largest
number of larvae observed during 2200-2300 hrs. Estimated relative larval loss did not
differ significantly among females as larvae were sampled at two sites approximately
0.15 and 1.5 km from the last section downstream of spawning locations. High inter- and
intra-female variation in larval contributions and relative larval loss between nights may
be a common feature of lake sturgeon and other migratory fish species, and likely is a
source of inter- annual and intra-annual variation in fish recruitment.

128

INTRODUCTION
Mortality during early life stages is a major factor affecting recruitment in fish
populations (review in Pepin 2009). Quantifying mortality rates of fish at early life stages
is important to understand fish population dynamics (Keesing and Halford, 1992; Pepin,
2009). During the larval stage, high mortality occurs at critical periods, for example,
when larvae switch from endogenous feeding (yolk absorption) to exogenous feeding
(Kamler, 1992; Fuiman and Higgs, 1997), and when individuals disperse from spawning
locations to nursing grounds (Bartsch et al., 1989; Kamler, 1992; Fuiman and Higgs,
1997). In many fish species, these two events occur at or near the same time, as larvae
often begin feeding prior to or during dispersal (e.g., lake sturgeon Acipenser fulvescens
(Auer and Baker, 2002)). Therefore, larvae become most susceptible to mortality during
the initial period of dispersal when larvae emerge and disperse from the spawning ground
(Einum and Fleming, 2000).
Exogenous and endogenous factors can contribute to larval mortality during
dispersal (Kamler, 1992; Cowen and Sponaugle, 2009). Major exogenous sources include
biotic (e.g., predation, food availability) and abiotic factors (e.g., extreme temperatures,
stream flow) (Kamler, 1992; Mion et al., 1998; Cowen and Sponaugle, 2009).
Endogenous factors such as egg size, egg energy content (yolk size), larval size, etc.,
involve female phenotypes as maternal effects (e.g., size, age) (Hutchings, 1991; Keckeis
et al., 2000; Kamler, 2005). Spawning time and location, which are also classified as
maternal effects (Mousseau and Fox, 1998) also determines the environmental conditions
that developing eggs and larvae are exposed to after egg deposition (Trippel et al., 1997;
Hendry and Day, 2005; Jorgensen et al., 2008). Therefore, maternal effects can influence

129

probabilities of survival at embryonic, larval and juvenile stages (Kamler, 1992;
Chambers, 1997). Previous studies have emphasized that exogenous factors are dominant
sources of larval mortality (Kamler, 1992). Recently, increasing emphasis has been
placed on the importance of larval behavior and genetic and maternal effects, which can
lead to different probabilities of individual survival even when exposed to similar
environmental conditions (Heath and Gallego, 1997; Bolnick et al., 2003; Leis, 2006;
Fiksen et al., 2007; Clobert et al., 2009). When larvae are exposed to a common
environment, if maternal effects are weak, and differences in probability of individual
survival are small compared to strong effects of external environmental factors, the
relative rates of offspring mortality among females would be expected to be similar. The
opposite prediction would lead to greater inter-family variation compared to intra-family
variation, which are two important components of selection, in larval survival. We tested
these predictions using a well-studied population of lake sturgeon (Acipenser fulvescens).
Lake sturgeon are characterized by a long life span, delayed maturity, iteroparity
and high fecundity but low annual recruitment due to high mortality during early life
stages (Peterson et al., 2007; Forsythe, 2010). Lake sturgeon do not provide postovulatory parental care for eggs or larvae (Bruch and Binkowski, 2002). Deposited eggs
adhere to stream substrates and hatch following 5 to 11 days of incubation depending on
water temperature (Auer and Baker, 2002; Smith and King, 2005). Newly-hatched larvae
remain in the stream substrates until yolk-sac reserves are depleted. Larvae are negatively
phototactic and disperse downstream at night (Auer and Baker, 2002; Kynard and Parker,
2005; Smith and King, 2005). This nocturnal behavior is considered to be an adaptive
means of predator avoidance (Auer and Baker, 2002), and has been observed in other

130

sturgeon species (Kynard and Parker, 2005) and other fishes (Crisp and Hurley, 1991;
Bradford and Taylor, 1997). Previous studies have investigated spatial and temporal
patterns in abundance of dispersing lake sturgeon larvae in relation to environmental
conditions (e.g., water temperature and water flow), revealing that dispersing larvae were
distributed in a non-uniform manner across the width of stream and position in the water
column, and over different hours of the night (Auer and Baker, 2002; Smith and King,
2005). Little information is available about the contributions of different females to
dispersing larvae each night. Neither are quantitative estimates available regarding
relative larval loss among females during the larval dispersal period for many fish species
including lake sturgeon.
The objectives of this study were to (i) quantify proportional contributions of
different females to dispersing larvae collected during consecutive nights; (ii)
characterize hourly dispersal patterns of larvae from females spawning at different
locations; and (iii) quantify relative larval loss among females and among female groups
spawning at different locations. Data pertaining to inter-family variation in contributions
to larval dispersal and the relative roles of maternal and stream environmental effects on
larval survivorship during dispersal can lead to greater understanding of inter-annual and
intra-annual variation in fish recruitment.
METHODS
Study site
A well-studied population of lake sturgeon located in Black Lake, northern
Michigan was used for this study. In this system, the Upper Black River (UBR) is the
largest tributary to Black Lake and the sole location used for spawning (Figure 5. 1). The

131

Black Lake population is isolated by dams blocking immigration and emigration from
other populations (Smith and King, 2005). Adults spawn over ~ 1 km-section of the
UBR, classified for the purposes of this study into two areas referred to as upstream (site
B) and downstream (site C), which are utilized consistently by spawning females in all
years (Forsythe, 2010).
Sampling design, sample collection and measurement of environmental factors
Lake sturgeon adults were sampled daily as they arrived at spawning locations
during the 2006 spawning season that occurred from April 20 to May 26. We walked the
entire spawning area one or more times per day using long handled nets to capture
spawning adults. Biological information including sex, capture date and location was
2

recorded for all individuals. A dorsal fin clip (~1 cm ) was taken from each individual for
genetic analysis.
Larvae were sampled at 2 sites located 0.15 (D1) and 1.5 (D2) km downstream
from the most downstream zone where adults spawned (Figure 5. 1). Larval sampling
was conducted during two nights, May 28 and 29, 2006. At each site, five D-frame larval
nets were placed across the river channel (description in Smith and King, 2005) and
checked hourly from 2200 to 0200 hrs. Larvae captured were kept separate by hour of
capture and site and were reared in a streamside hatchery until they were large enough
(for 3 months) for fin clips to be taken. Then, 30% of all the fish were sub-sampled for
genetic analysis. Samples genotyped included 831 larvae including 283 and 145 larvae
collected at sites D1 and D2 on May 28, and 271 and 132 larvae at sites D1 and D2 on
May 29, respectively.

132

We measured water temperature and water velocity every day of the spawning
season through the larval sampling period. Water temperature was recorded hourly using
HOBO data loggers (Oneset Computer Corp.) at three locations: the upstream spawning
location, upstream of, and downstream of larval collection sites. Three water velocity
readings (left, center, right) were taken in front of the drift net openings (0.6 of the stream
depth measured downward from the surface) with a Marsh–McBirney Flo-Mate 2000
(Marsh–McBirney Inc., Fredrick, MD, USA). Measurements of water velocity were also
recorded from the entire cross-sectional area of the channel at 1m intervals across a
standardized transect. In this study, we reported water temperature and water velocity in
two days of larval sampling. Average (±SD) temperatures and velocity among locations
measured on May 28 and 29 were 20.33± 0.69ºC and 20.72 ± 0.05ºC; and 0.36± 0.11 m/s
and 0.47 ± 0.12 m/s, respectively.
Genetic analysis
DNA was extracted from larval tissue samples using a QIAGEN DNeasy(R) kit
(QIA Inc.). DNA concentration was measured using a Nanodrop 1000 spectrophotometer
(Thermo Fisher Scientific Inc.). All samples were genotyped at 12 tetra-nucleotide
microsatellite loci including Spl 120 (McQuown et al., 2000); AfuG 68B (McQuown et
al., 2002). 2002); Aox 27 (King et al., 2001); AfuG 68, AfuG 9; AfuG 63, AfuG 74,
AfuG 112, AfuG 56, AfuG 160, AfuG 195 and AfuG 204 (Welsh et al., 2003). Using
polymerase chain reaction (PCR), 100 ng DNA was amplified in 25 µl reaction mixtures
as described in the above references. All PCR reactions were conducted using a
Robocyler 96 thermal cycler (Stratagene). PCR products were run on 6% denaturing
polyacrylamide gels and visualized using a Hitachi FMBIO II scanner. Allele sizes were

133

scored independently by two experienced personnel. Genotyping errors were checked by
blindly re-genotyping a random subset of 10% of all samples.
Parentage analysis
The number of candidate parents affects the success rate of and confidence in
parentage assignment (Marshall et al., 1998). We used a subset of all captured spawning
adults for parentage analysis based on empirical estimates of the cumulative temperature
units from the period from egg deposition to larval dispersal (Smith and King, 2005;
Duong et al., in press), and estimated spawning date (based on date of capture) of adults.
Using these criteria larvae collected on May 28-29 were expected to be offspring of
adults captured from May 1 to May 15 (35 females and 91 males, Figure 5. 2). The
offspring of these adults were predicted to disperse on average 17 days following the
female spawning date under mean water temperature (from May 1, the first day of
spawning by these adults, to May 29, the last day of collecting larvae) of 14.7±2.4⁰C.
Meanwhile, offspring of females spawning earlier (from April 20-24) were predicted to
disperse approximately19 days post-spawning, and offspring of females spawning later
(May-25-26) predictably dispersed 13 days post-spawning. Thus, at the time of larval
sampling, larvae from early spawning females had already dispersed, and larvae from late
spawning adults had not hatched or dispersed yet.
Two software programs, the Parentage Allocation of Singles on Open Systems
(PASOS), version 1.0 (Duchesne et al., 2005) and CERVUS 3.0 (Kalinowski et al., 2007)
were used for parentage analysis. Parental assignments for individual larvae from the two
programs were compared, and larvae that were concordantly assigned from the two
programs were used for further analysis.

134

We used two functions in PASOS including the allocation function for assigning
offspring to collected adults and estimating missing parents that were not collected, and
the simulation function for estimating the probability that any assignment is correct
(correctness probability). Simulations to determine accuracy (correctness) of parentage
assignments were conducted over 5 iterations using 1,000 simulated offspring. Using
CERVUS, we first simulated 10,000 offspring to determine the likelihood of assignment
and confidence levels of parental assignment based on population estimates of allele
frequencies (N=126 spawning adults), captured adult proportions estimated from PASOS,
and the empirical estimate of genotyping errors (0.67%). From the assignment output, a
“true parent pair” of an offspring was chosen from the most likely candidate parents. We
then compared the concordance of parental assignment outputs from the two programs.
Inconsistent assignments of offspring to two collected parents between the two programs
were evaluated based on biological information (e.g., proximity of captured time and
location of both parents).
Data analysis
We first tested the hypothesis that females contributed the same proportion of
larvae to collections at both sampling sites (upstream D1 and downstream D2) during
each of 2 nights (May 28 and 29). Proportional contributions of larvae from each female
were estimated as the ratio of the number of offspring per female to the total number of
offspring collected at each larval sampling site each night (hereafter called standardized
offspring). We used a general linear mixed model approach to quantify differences in
standardized offspring among females as a fixed effect at two larval sampling sites and at
two nights. Because sites and nights of sampling were not independent, they were treated

135

as random effects (Burnham and Anderson, 2002; Bolker et al., 2009). We used
likelihood ratio tests to determine significance of random effects and Wald tests for fixed
effects.
A mixed model was also used to compare proportions of larvae from the upstream
and downstream spawning locations, which contributed to the total number of larvae
collected at sites D1 and D2. In this model, spawning location was treated as a fixed
effect and female as a random effect.
Hourly dispersal patterns of larvae from females spawning at upstream (site B)
and downstream (site C) across five consecutive hours of sampling at site D1 during each
of two nights were also characterized. We used nonparametric Mann-Whitney Wilcoxon
tests to compare the number and proportion of larvae collected by hours from two female
groups spawning at upstream and downstream sites.
We then estimated and compared the relative loss of larvae among females
between upstream and downstream collection sites. Relative loss by a female was defined
as the ratio of the number of larvae that were collected at the downstream site and
assigned to each female to the total number of the same female’s offspring collected at
both upstream and downstream sites (D1 and D2, respectively). Differences in relative
loss among females were tested using a general linear mixed model including the fixed
effect of female and the random effect of sampling dates. Similarly, relative loss for
female groups spawning at upstream and downstream locations was also compared with a
mixed model including spawning locations as a fixed effect and females as random
effects.

136

RESULTS
Parentage assignment
Based on program PASOS, 741 larvae (89% of the total number of larvae
genotyped) were assigned to two collected parents with allocation correctness estimated
to be 0.82±0.01. Male and female adults captured were estimated (by the program) to
represent 89.3 and 91.8%, respectively, of the total number of adults estimated to have
contributed to the larvae sampled. The assignment rate obtained from CERVUS was
91.8%. Comparing parental assignment outputs from the two programs revealed that
72.4% (602 larvae) of all sampled larvae were concordantly assigned to the same two
collected parents. Based on biological information regarding timing of spawning
observed for putative male-female pairs to evaluate discordance in assignment of the two
programs, we finally selected 738 larvae for further analysis.
Hourly pattern of larval dispersal
The number of larvae collected revealed an hourly pattern of larval dispersal from
each spawning area. Hourly patterns were consistent across the 2 nights and between
female spawning locations (Figure 5. 3). Based on stream velocity (0.36 and 0.47 m/s on
May 28 and 29, respectively), the time necessary to disperse passively from the upstream
spawning location (site B) to the upstream larval collection site (D1) was about 35-45
minutes. Genetic data revealed that when checking nets hourly, larvae from both
spawning sites were present in the upstream larval collection. We observed an
approximately one hour lag between the peaks of larvae captured at upstream and
downstream locations. The highest number of larvae from both spawning sites was
observed from 2200-2300 hrs at site D1 and 2300-2400 hours at site D2. The number and

137

proportion of larvae collected by hour at both larval collection sites were not significantly
different between the two groups of females that spawned at locations B (upstream) and
C (downstream) (Wilcoxon tests, P > 0.5 for all tests).
Proportional contributions of females to larval collections at two sampling sites and
nights
The number of larvae captured during 2 nights and assigned to a female varied
from 1 to 140 individuals (mean = 21, median = 11, mode = 5). Standardized offspring
also varied significantly among females (F34, 104 = 32.0, P < 0.001). Four females
consistently had higher standardized offspring at both sampling locations and dates
(Figure 5. 4). Random effects of sampling night and sampling site were not significant
(LRT ≈ 0, df =1, P = 0.99). Therefore, the rank order of standardized offspring among
females was consistent between two nights and also between the two larval sampling
sites.
The number of larvae produced by females from the upstream location B
comprised a higher proportion of larvae captured relative to larvae from females
spawning at the downstream spawning location C (F1, 137 = 14.4, P < 0.01). However, in
terms of larval proportion (standardized offspring) contributed by each female,
standardized offspring per female spawning at location B was not significantly higher
than that of females spawning at location C (F1, 104 = 1.90, P = 0.17). There was a large
variation in standardized offspring among females within each spawning location.
Relative larval loss
Differences in the number of offspring between upstream and downstream larval
sampling sites provided a measure of relative larval loss among females between the
138

sampling sites. Relative larval loss was not significantly different among females (F34, 34
= 1.36, P = 0.19). This result indicates that females with higher numbers of offspring lost
proportionally more offspring than did females contributing comparatively fewer
offspring, when larvae from all females dispersed through a common environment from
site D1 to site D2. However, large variation in relative larval loss within female (the same
female between nights) and among females was observed (Figure 5. 5). Partitioning
components of variance in relative larval loss revealed that a small proportion (3%) of
total variance was attributed to random effects of sampling nights (across females), and a
larger proportion (97%) was attributed to variance in relative loss of the same female in
two nights (residual error). When females were grouped by spawning site, estimated
relative larval loss from females spawning at the upstream location (0.25±0.20) was not
significantly different from females spawning at the downstream location (0.29±0.22)
(F1, 34 = 0.55, P = 0.46). The non-significance in relative loss of females from two
spawning locations was likely attributed to the large degree of variation among females
within each spawning location.
DISCUSSION
Differences in trait expression among individual larvae, which affect survival, can
occur as a result of maternal and genetic effects, even when individuals are exposed to
the same environmental conditions (Heath and Gallego, 1997; Bolnick et al., 2003; Leis,
2006; Fiksen et al., 2007; Clobert et al., 2009). Further, inter- and intra-family variation
in phenotypes at larval stages can affect traits and survival at later life stages, providing
the opportunity for the evolution of maternal traits and behavior (e.g., spawning time) in
natural populations (Einum and Fleming, 2000). One important question is whether
139

larvae produced by different females are differentially susceptible to mortality when
exposed to a common environment, for example, when larvae disperse from spawning
grounds to areas utilized during subsequent early life stages. Such a question cannot be
addressed without genetic data or other measures establishing pedigree relationships and
estimates of reproductive success and survival. Variation in relative larval mortality
among females would imply that maternal effects and individual variability outweigh the
effects of stream environmental factors.
Using genetically determined parentage, we quantified the intra- and inter-female
variability of larval loss during dispersal through a common natural stream environment
over a distance of 1.5 km during two consecutive nights. A considerable portion of
mortality during early life stages occurs at emergence and during the initial period of
dispersal (Einum and Fleming, 2000). Therefore, the time and distance over which we
examined larval loss was predicted to be an important period when larvae experienced
high mortality. Results showed that lake sturgeon females differed in both absolute
numbers and relative proportions of offspring contributed to larval collections each night.
However, no significant difference in relative larvae loss among females or between
groups of females spawning at upstream and downstream locations was detected. This
result indicates stronger effects of exogenous factors (e.g., predation, food availability,
temperature, water flow, etc.,) than endogenous factors (e.g. maternal effects and larval
behavior) on mortality of larvae from all females and spawning locations. However,
variation in relative larval loss between the two sampling nights for each female (within
female) was as high as variation among females (Figure 5. 5). Such large intra-female
variation can mask the significant effects of inter-female variability in larval loss. On the

140

other hand, daily mortality at larval stages can vary greatly, and is a common
phenomenon across species (e.g., walleye Stizostedion vitreum (Mion et al., 1998); Pepin
2009). Accordingly, the large degree of intra-and inter-female variation in larval loss
observed in lake sturgeon may be a normal feature in stream environments.
Estimates of larval loss between two sampling sites can likely be attributed to
natural mortality (e.g., due to starvation or predation), differences in capture efficiencies
at the two larval sampling sites or differences in movement behavior among larvae. We
believe the latter two hypotheses are unlikely. First, if we assumed that all larvae had an
equal probability of capture, differences in capture efficiencies between sampling sites
would not bias estimates of relative larval loss among females. Second, it is possible that
downstream movement may not be consistent among larvae (Siegel et al., 2003; Leis,
2006; Shanks, 2009). For example, some larvae could settle in substrates between the two
sampling sites or slow their movement. However, based on our observation of stream
conditions over the stream area surveyed and consistent patterns of larval dispersal by
hour each night (Figure 5. 3), variation in larval movement behavior was likely minor and
did not contribute to differences in relative larval loss among females. In sum, the lack of
differences in relative larval loss among females was likely due to similar susceptibility
to natural mortality during the two nights.
The main sources of natural mortality during the initial period of larval dispersal
can be predation, starvation, and physical injure due to water velocity (Kamler, 1992;
Mion et al., 1998; Cowen and Sponaugle, 2009). Water velocity (Mion et al., 1998) and
other exogenous abiotic factors such as oxygen concentration and temperature (Kamler,
2005) are unlikely to cause larval loss in this study because the stream conditions were

141

characterized by intermediate levels of water velocities (0.36 – 0.47 m/s) and stable
temperatures (20 – 21ºC) during two nights. Predation is one of the most significant
causes of larvae mortality during dispersal for numerous fish species (Paradis et al.,
1996; Cowen and Sponaugle, 2009). Predators of lake sturgeon larvae in the Upper Black
River included crayfish (Orconectes rusticus) and rock bass (Ambloplites rupestris)
(Crossman, 2008; Forsythe, 2010). Larvae are not able to escape predators during
dispersal until the juvenile stages where the total length is > 40 cm (Auer and Baker,
2002). Therefore, the effects of predation on larval mortality are expected to be random
among females. Starvation can cause substantial mortality of larvae in several fish
species (e.g., trout Salmo trutta (Elliott, 1986); plaice Pleuronectes platessa (Fox et al.,
2007)). In lake sturgeon, larvae emerge from the substrate and disperse after depleting
yolk-sac reserves (Auer and Baker, 2002; Kynard and Parker, 2005; Smith and King,
2005). Accordingly, starvation likely results in mortality during the initial period of larval
dispersal. Given starvation mortality may be related to egg size and egg energy content
(Keckeis et al., 2000), which are components of maternal effects (Mousseau and Fox,
1998; Kamler, 2005), larger inter-female variation compared to intra-female variation in
relative larval loss would be expected if starvation or decreased energy levels due to food
limitation was a main source of mortality. However, our study showed that intra- and
inter-female variation were approximately equal, suggesting predation mortality was
dominant over starvation mortality.
Biologists have examined whether larvae dispersing in groups consisted of
offspring from different families, in order to understand schooling behavior of fish during
dispersal (Avise et al., 2002). We found that collections of dispersing lake sturgeon

142

larvae each night were comprised of offspring from the majority of females sampled in
spawning locations. However, proportional contributions from different females varied
greatly (Figure 5. 4). Different proportional contributions of females to larval drift could
be explained due to variability in female fecundity (Bruch et al., 2006), hatching rate
(Nichols et al., 2003; Crossman, 2008), and survival from egg deposition to larval
dispersal (Caroffino et al., 2010). These factors could result in different reproductive
success of individual females. Although this study was not designed to disentangle the
causes of differences in proportional contribution of females to larval collections, a large
range in the numbers of offspring captured and the consistent magnitude of differences in
standardized offspring among females between two sampling nights is indicative of high
variation in reproductive success among females.
Our study also revealed similar hourly dispersal patterns of larvae produced from
females spawning at upstream and downstream locations in two nights, with the peak of
larval drift occurring between 2200 and 2300 hrs. Under average water velocities of 0.36
– 0.47 m/s, similar patterns of dispersal across hours (nets were checked hourly) observed
in larvae from two spawning locations implies that movements of larvae downstream
over the distance investigated were likely passive. These patterns were observed
consistently for 2 consecutive nights in this study and also in other studies conducted in
different years (e.g., from 2000-2003, Smith and King 2005).
The short sampling period might have limited our ability to detect differences in
relative larval loss among females. We did not collect larvae produced from females
representing the entire spawning period. Given that daily mortality of larvae within and
among families is highly variable (Mion et al., 1998; Pepin, 2009), many days of

143

sampling may reduce intra-family variation and thus allow detection of inter-family
variation in larval loss. In addition, because larvae produced from the same spawning
event may disperse over many days (Smith and King, 2005; Duong et al., in press)
proportional contributions of larvae from females and relative loss interpreted based on
two sampling nights and a short distance may not reflect differences in numbers of
offspring produced by females and relative larval loss among females over the entire
period of larval dispersal.
Further studies that use genetic markers to quantify relative larval loss by families
could profitably explore the effects of timing, duration and distance of dispersal on larval
loss. Severe environmental conditions such as high river discharge, high predation
pressure, extreme temperatures, and other exogenous factors may accentuate differences
in rates of mortality among families.

144

N

Black Lake
Upper Black River

.000000

D2
D1

C
B
0

0.5

1

2

Kilometer
s

Figure 5.1. The study site in the Upper Black River, Michigan, showing upstream and
downstream adult spawning locations (B and C, respectively); and upstream and
downstream larval collection sites (D1 and D2, respectively).

145

Number of males and females

40

Female
Male
Offpsring

30

200
150

20

100

10

50

0

Number of females' offspring

250

50

0

20-Apr

27-Apr

4-May

11-May

18-May

25-May

Adult capture date
Figure 5. 2. Numbers of adults captured each day of the 2006 spawning season and total
number of offspring collected on two nights (May 28 and 29) that were assigned to
captured females using genetically determined parentage.

146

Number of larvae collected and
assigned

80
70
60

Site B-May 28

50

Site C-May 28

40

Site B-May 29

30

Site C-May 29

20
10
0
22:00

23:00

0:00

1:00

2:00

Sampling time
Figure 5.3. Hourly pattern of larval dispersal based on collections at sampling site D1 on
2 nights (May 28 and 29). Larval numbers are cumulative totals over all females
spawning at upstream (Site B) and downstream (Site C) locations (Figure 5. 1).

147

Standardized offpsring

0.2

May 28

0.15

Location D1
Location D2

0.1
0.05

C077
B078
C080
C081
C087
C090
C091
C094
C096
C099
C100
C110
C114
C125
C126
C133
C134
C135
C146
C147
C150
C153
C161
C165
C170
B175
B176
B179
B180
C185
B186
B187
C189
B196
B203

0

Female ID
Standardized offpsring

0.25

May 29

0.2
0.15
0.1
0.05

C077
B078
C080
C081
C087
C090
C091
C094
C096
C099
C100
C110
C114
C125
C126
C133
C134
C135
C146
C147
C150
C153
C161
C165
C170
B175
B176
B179
B180
C185
B186
B187
C189
B196
B203

0

Female ID
Figure 5.4. Standardized offspring produced by each female estimated based on
collections at two larval sampling sites on two consecutive sampling nights. Female ID
includes reference to spawning locations (B vs. C) and are presented in order of spawning
date (May 1-15, Figure 5. 2).

148

Relativelarval loss

1.2
1.0
0.8
0.6
0.4
0.2
0.0
-0.2
0

5

10

15
20
Female

25

30

35

Figure 5. 5. Relative larval loss among females (bars represent 1SD between two nights).
The order of females was as the same as Figure 5. 4.

149

REFERENCES

150

REFERENCES
Andersson, M., Iwasa, Y., 1996. Sexual selection. Trends Ecol Evol 11, A53-A58.
Araki, H., Waples, R.S., Ardren, W.R., Cooper, B., Blouin, M.S., 2007. Effective
population size of steelhead trout: influence of variance in reproductive success,
hatchery programs, and genetic compensation between life-history forms. Mol Ecol
16, 953-966.
Arbab, A., Kontodimas, D.C., McNeill, M.R., 2008. Modeling Embryo Development of
Sitona discoideus Gyllenhal (Coleoptera: Curculionidae) Under Constant
Temperature. Environmental Entomology 37, 1381-1388.
Arnold, S.J., Duvall, D., 1994. Animal Mating Systems - a Synthesis Based on Selection
Theory. Am Nat 143, 317-348.
Auer, N.A., 1999. Population characteristics and movements of Lake Sturgeon in the
Sturgeon River and Lake Superior. J Great Lakes Res 25, 282-293.
Auer, N.A., Baker, E.A., 2002. Duration and drift of larval lake sturgeon in the Sturgeon
River, Michigan. J Appl Ichthyol 18, 557-564.
Avise, J.C., 2004. Molecular markers, natural history, and evolution. Sinauer Associates,
Sunderland, Mass.
Avise, J.C., Jones, A.G., Walker, D., DeWoody, J.A., Collaborators, 2002. Genetic
mating systems and reproductive natural histories of fishes: Lessons for ecology and
evolution. Annu Rev Genet 36, 19-45.
Baker, E.A., Borgeson, D.J., 1999. Lake sturgeon abundance and harvest in Black Lake,
Michigan, 1975-1999. N Am J Fish Manage 19, 1080-1088.
Bartsch, J., Brander, K., Heath, M., Munk, P., Richardson, K., Svendsen, E., 1989.
Modeling the advection of herring larvae in the north-sea. Natrue, 632-636.
Bateman, A.J., 1948. Intra-Sexual Selection in Drosophila. Heredity 2, 349-368.
Beebee, T.J.C., 2009. A comparison of single-sample effective size estimators using
empirical toad (Bufo calamita) population data: genetic compensation and population
size-genetic diversity correlations. Mol Ecol 18, 4790-4797.
Bekkevold, D., Hansen, M.M., Loeschcke, V., 2002. Male reproductive competition in
spawning aggregations of cod (Gadus morhua, L.). Mol Ecol 11, 91-102.
Blachford, A., Agrawal, A.F., 2006. Assortative mating for fitness and the evolution of
recombination. Evolution 60, 1337-1343.

151

Blouin, M.S., 2003. DNA-based methods for pedigree reconstruction and kinship
analysis in natural populations. Trends Ecol Evol 18, 503-511.
Blukacz, E.A., Koops, M.A., Sutton, T.M., Arts, M.T., Fitzsimons, J.D., Muir, A.M.,
Claramunt, R.M., Johnson, T.B., Kinnunen, R.E., Ebener, M.P., Suski, C., Burness,
G., 2010. Linking lake whitefish (Coregonus clupeaformis) condition with male
gamete quality and quantity. J Great Lakes Res 36, 78-83.
Boake, C.R.B., 1989. Repeatability - Its Role in Evolutionary Studies of MatingBehavior. Evol Ecol 3, 173-182.
Bolker, B.M., Brooks, M.E., Clark, C.J., Geange, S.W., Poulsen, J.R., Stevens, M.H.H.,
White, J.S.S., 2009. Generalized linear mixed models: a practical guide for ecology
and evolution. Trends Ecol Evol 24, 127-135.
Bolnick, D.I., Svanback, R., Fordyce, J.A., Yang, L.H., Davis, J.M., Hulsey, C.D.,
Forister, M.L., 2003. The ecology of individuals: Incidence and implications of
individual specialization. Am Nat 161, 1-28.
Bonin, A., Bellemain, E., Eidesen, P.B., Pompanon, F., Brochmann, C., Taberlet, P.,
2004. How to track and assess genotyping errors in population genetics studies. Mol
Ecol 13, 3261-3273.
Bradford, M., Taylor, G., 1997. Individual variation in dispersal behaviour of newly
emerged Chinook salmon (Oncorhynchus tshawytscha) from the upper Fraser River,
British Columbia. Can J Fish Aquat Sci, 1585-1592.
Briton, J., Nurthen, R.K., Briscoe, D.A., Frankham, R., 1994. Modelling problems in
conservation genetics using Drosophila: Consequences of harems. Biol Conserv 69,
267-275.
Bruch, R.M., 1999. Management of lake sturgeon on the Winnebago System - long term
impacts of harvest and regulations on population structure. J. Appl. Ichthyol.-Z.
Angew. Ichthyol. 15, 142-152.
Bruch, R.M., Binkowski, F.P., 2002. Spawning behavior of lake sturgeon (Acipenser
fulvescens). J Appl Ichthyol 18, 570-579.
Bruch, R.M., Miller, G., Hansen, M.J., 2006. Fecundity of lake sturgeon (Acipenser
fulvescens, Rafinesque) in Lake Winnebago, Wisconsin, USA. J Appl Ichthyol 22,
116-118.
Burnham, K.P., Anderson, D.R., 2002. Model selection and multimodel inference : a
practical information-theoretic approach. Springer, New York.
Caballero, A., 1994. Developments in the prediction of effective population-size.
Heredity 73, 657-679.

152

Calderon, I., Turon, X., 2010. Temporal genetic variability in the Mediterranean common
sea urchin Paracentrotus lividus. Mar Ecol-Prog Ser 408, 149-159.
Caroffino, D.C., Sutton, T.M., Elliott, R.F., Donofrio, M.C., 2010. Early life stage
mortality rates of lake sturgeon in the Peshtigo River, Wisconsin. N Am J Fish
Manage 30, 295-304.
Casagrande, S., Dell'Omo, G., Costantini, D., Tagliavini, J., 2006. Genetic differences
between early- and late-breeding Eurasian kestrels. Evol Ecol Res 8, 1029-1038.
Chambers, R.C., 1997. Environmental influences on egg and propagule sizes in marine
fishes. In: Chambers, R.C., Trippel, E.A. (Eds.), Early life history and recruitment in
fish populations. Chapman and Hall.
Chambers, R.C., Leggett, W.C., 1996. Maternal influences on variation in egg sizes in
temperate marine fishes. Am Zool 36, 180-196.
Champer, C., 1997. Environmental influences on egg and propagule sizes in marine
fishes. In: Chambers, R.C., Trippel, E.A. (Eds.), Early life history and recruitment in
fish populations. Chapman and Hall.
Chandler, G.L., Bjornn, T.C., 1988. Abundance, growth, and interactions of juvenile
steelhead relative to time of emergence. T Am Fish Soc 117, 432-443.
Charlesworth, B., 2009. Effective population size and patterns of molecular evolution and
variation. Nat Rev Genet 10, 195-205.
Chesser, R.K., Rhodes, O.E., Sugg, D.W., Schnabel, A., 1993. Effective sizes for
subdivided populations. Genetics 135, 1221-1232.
Christie, M.R., 2010. Parentage in natural populations: novel methods to detect parentoffspring pairs in large data sets. Mol Ecol Resour 10, 115-128.
Clobert, J., Le Galliard, J., Cote, J., Meylan, S., Massot, M., 2009. Informed dispersal,
heterogeneity in animal dispersal syndromes and the dynamics of spatially structured
populations. Ecology Letters, 197-209.
Coma, R., Lasker, H.R., 1997. Small-scale heterogeneity of fertilization success in a
broadcast spawning octocoral. J Exp Mar Biol Ecol 214, 107-120.
Copp, G.H., Faulkner, H., Doherty, S., Watkins, M.S., Majecki, J., 2002. Diel drift
behaviour of fish eggs and larvae, in particular barbel, Barbus barbus (L.), in an
English chalk stream. Fish Manage Ecol 9, 95-103.
Cowen, R., Sponaugle, S., 2009. Larval dispersal and marine population connectivity.
Ann Rev Mar Science 1, 443-466.

153

Crawford, J.C., Liu, Z.W., Nelson, T.A., Nielsen, C.K., Bloomquist, C.K., 2009. Genetic
population structure within and between beaver (Castor canadensis) populations in
Illinois. J Mammal 90, 373-379.
Crisp, D.T., Hurley, M.A., 1991. Stream channel experiments on downstream movement
of recently emerged trout, Salmo-trutta L and Salmon, S. salar L .1. Effect of 4
different water velocity treatments upon dispersal rate. J Fish Biol 39, 347-361.
Crossman, J.A., 2008. Evaluating collection, rearing, and stocking methods for lake
sturgeon (Acipenser fulvescens) restoration programs in the Great Lakes. Department
of Fisheries and Wildlife. Michigan State University, East Lansing, MI.
Crow, J.F., Denniston, C., 1988. Inbreeding and variance effective population numbers.
Evolution 42, 482-495.
Csillery, K., Johnson, T., Beraldi, D., Clutton-Brock, T., Coltman, D., Hansson, B.,
Spong, G., Pemberton, J.M., 2006. Performance of marker-based relatedness
estimators in natural populations of outbred vertebrates. Genetics 173, 2091-2101.
Day, T., Rowe, L., 2002. Developmental thresholds and the evolution of reaction norms
for age and size at life-history transitions. Am Nat 159, 338-350.
Dehaan, P.W., 2003. Demographic and life history characteristics of remnant lake
sturgeon populations in the upper great lakes basin: Inference based on genetic
analyses. Michigan State University. Master's thesis.
DeHaan, P.W., Libants, S.V., Elliott, R.F., Scribner, K.T., 2006. Genetic population
structure of remnant lake sturgeon populations in the upper Great Lakes basin. T Am
Fish Soc 135, 1478-1492.
Devaux, C., Lande, R., 2008. Incipient allochronic speciation due to non-selective
assortative mating by flowering time, mutation and genetic drift. P R Soc B 275,
2723-2732.
DeWoody, J.A., Avise, J.C., 2001. Genetic perspectives on the natural history of fish
mating systems. J Hered 92, 167-172.
Doctor, K.K., Quinn, T.P., 2009. Potential for adaptation-by-time in sockeye salmon
(Oncorhynchus nerka): the interactions of body size and in-stream reproductive life
span with date of arrival and breeding location. Can J Zool 87, 708-717.
Duchesne, P., Castric, T., Bernatchez, L., 2005. PASOS (Parental allocation of singles in
open systems): a computer program for individual parental allocation with missing
parents. Mol Ecol Notes 5, 701-704.
Duong, T.Y., Scribner, K.T., Crossman, J., Forsythe, P., Baker, E., in press.
Environmental and maternal effects on embryonic and larval developmental time
until dispersal of lake sturgeon.
154

Edmands, S., Moberg, P.E., Burton, R.S., 1996. Allozyme and mitochondrial DNA
evidence of population subdivision in the purple sea urchin Strongylocentrotus
purpuratus. Mar Biol 126, 443-450.
Edwards, K.P., Hare, J.A., Werner, F.E., Seim, H., 2007. Using 2-dimensional dispersal
kernels to identify the dominant influences on larval dispersal on continental shelves.
Mar Ecol-Prog Ser 352, 77-87.
Einum, S., Fleming, I.A., 2000. Selection against late emergence and small offspring in
Atlantic salmon (Salmo salar). Evolution 54, 628-639.
Elliott, J.M., 1986. Spatial distribution and behavioral movements of migratory trout
Salmo trutta in a lake district stream. J Anim Ecol 55, 907-922.
Elliott, J.M., 1987. The distances traveled by downstream-moving trout fry, Salmo trutta,
in a Lake District Stream. Freshw Biol 17, 491-499.
Emlen, S.T., Oring, L.W., 1977. Ecology, sexual selection, and evolution of mating
systems. Science 197, 215-223.
England, P.R., Cornuet, J.M., Berthier, P., Tallmon, D.A., Luikart, G., 2006. Estimating
effective population size from linkage disequilibrium: severe bias in small samples.
Conserv Genet 7, 303-308.
Erisman, B.E., Buckhorn, M.L., Hastings, P.A., 2007. Spawning patterns in the leopard
grouper, Mycteroperca rosacea, in comparison with other aggregating groupers. Mar
Biol 151, 1849-1861.
Fagan, W.F., Cosner, C., Larsen, E.A., Calabrese, J.M., 2010. Reproductive asynchrony
in spatial population models: how mating behavior can modulate allee effects arising
from isolation in both space and time. Am Nat 175, 362-373.
Falconer, D.S., 1989. Introduction to quantitative genetics. Longman, London.
Fausch, K.D., Taniguchi, Y., Nakano, S., Grossman, G.D., Townsend, C.R., 2001. Flood
disturbance regimes influence rainbow trout invasion success among five holarctic
regions. Ecol Appl 11, 1438-1455.
Felsenst.J, 1971. Inbreeding and variance effective numbers in populations with
overlapping generations. Genetics 68, 581-&.
Fiksen, O., Jorgensen, C., Kristiansen, T., Vikebo, F., Huse, G., 2007. Linking
behavioural ecology and oceanography: larval behaviour determines growth,
mortality and dispersal. Mar Ecol-Prog Ser, 195-205.
Fillatre, E.K., Etherton, P., Heath, D.D., 2003. Bimodal run distribution in a northern
population of sockeye salmon (Oncorhynchus nerka): life history and genetic analysis
on a temporal scale. Mol Ecol 12, 1793-1805.
155

Forsythe, P.S., 2010. Evaluation of exogenous variables affecting spawning timing,
natural egg deposition and mortality during the early life stages of lake sturgeon.
Department of Zoology. Michigan State University, East Lansing, MI.
Fox, C.J., Geffen, A.J., Taylor, N., Davison, P., Rossetti, H., Nash, R.D.M., 2007. Birthdate selection in early life stages of plaice Pleuronectes platessa in the eastern Irish
Sea (British Isles). Mar Ecol-Prog Ser 345, 255-269.
Fox, C.W., Czesak, M.E., 2000. Evolutionary ecology of progeny size in arthropods. Ann
Rev Entom 45, 341-369.
Fox, G.A., 2003. Assortative mating and plant phenology: evolutionary and practical
consequences. Evol Ecol Res 5, 1-18.
Franke, E.S., Babcock, R.C., Styan, C.A., 2002. Sexual conflict and polyspermy under
sperm-limited conditions: In situ evidence from field simulations with the freespawning marine echinoid Evechinus chloroticus. Am Nat 160, 485-496.
Frankham, R., 1995. Effective population-size adult-population size ratios in wildlife - A
review. Gen Res 66, 95-107.
Fraser, D.J., Bernatchez, L., 2005. Adaptive migratory divergence among sympatric
brook charr populations. Evolution 59, 611-624.
Fraser, D.J., Lippe, C., Bernatchez, L., 2004. Consequences of unequal population size,
asymmetric gene flow and sex-biased dispersal on population structure in brook charr
(Salvelinus fontinalis). Mol Ecol 13, 67-80.
Fuiman, L.A., Higgs, D.M., 1997. Ontogeny, growth and the recruitment process. In:
Chambers, R.C., Trippel, E.A. (Eds.), Early life history and recruitment in fish
populations. Chapman and Hall, pp. 225–249.
Gall, G.A.E., Neira, R., 2004. Genetic analysis of female reproduction traits of fanned
coho salmon (Oncorhyncus kisutch). Aquaculture 234, 143-154.
Garant, D., Kruuk, L.E.B., 2005. How to use molecular marker data to measure
evolutionary parameters in wild populations. Mol Ecol 14, 1843-1859.
Garvey, J.E., Stein, R.A., Thomas, H.M., 1994. Assessing how fish predation and
interspecific prey competition influence a crayfish assemblage. Ecology 75, 532-547.
Gillooly, J.F., Brown, J.H., West, G.B., Savage, V.M., Charnov, E.L., 2001. Effects of
size and temperature on metabolic rate. Science 293, 2248-2251.
Gillooly, J.F., Charnov, E.L., West, G.B., Savage, V.M., Brown, J.H., 2002. Effects of
size and temperature on developmental time. Nature 417, 70-73.

156

Gillooly, J.F., Dodson, S.I., 2000. The relationship of egg size and incubation
temperature to embryonic development time in univoltine and multivoltine aquatic
insects. Freshw Biol 44, 595-604.
Goodnight, K.F., Queller, D.C., 2001. Relatedness 5.0.8. Available at:
http://www.gsoftnet.us/GSoft.html., Rice University.
Goudet, J., 2001. FSTAT, a program to estimate and test gene diversities and fixation
indices (version 2.9.3). Available from
http://www.unil.ch/izea/softwares/fstat.html. Updated from Goudet (1995).
Graham, M.H., 1997. Factors determining the upper limit of giant kelp, Macrocystis
pyrifera Agardh, along the Monterey Peninsula, central California, USA. J Exp Mar
Biol Ecol 218, 127-149.
Graham, M.H., 2003. Confronting multicollinearity in ecological multiple regression.
Ecology 84, 2809-2815.
Guo, S.W., Thompson, E.A., 1992. A Monte-Carlo method for combined segregation and
linkage analysis. Am J Hum Genet 51, 1111-1126.
Hall, K.C., Hanlon, R.T., 2002. Principal features of the mating system of a large
spawning aggregation of the giant Australian cuttlefish Sepia apama (Mollusca :
Cephalopoda). Mar Biol 140, 533-545.
Hamilton, W.D., 1963. Evolution of Altruistic Behavior. Am Nat 97, 354-&.
Harris, R.B., Allendorf, F.W., 1989. Genetically Effective population-size of large
mammals - an assessment of estimators. Conserv Biol 3, 181-191.
Hasselquist, D., 1998. Polygyny in great reed warblers: A long-term study of factors
contributing to male fitness. Ecology 79, 2376-2390.
Hauser, L., Carvalho, G.R., 2008. Paradigm shifts in marine fisheries genetics: ugly
hypotheses slain by beautiful facts. Fish Fish 9, 333-362.
Hay-Chmielewski, E.M., Whelan, G.E., 1997. Lake sturgeon rehabilitation strategy.
Michigan Department of Natural Resources Fisheries Division, Ann Arbor, MI.
Heath, D.D., Fox, C.W., Heath, J.W., Qk, 1999. Maternal effects on offspring size:
Variation through early development of chinook salmon. Evolution 53, 1605-1611.
Heath, M., Gallego, A., 1997. From the biology of the individual to the dynamics of the
population: bridging the gap in fish early life studies. J Fish Biol 51, 1-29.
Hedgecock, D., Chow, V., Waples, R.S., 1992. Effective population numbers of shellfish
broodstocks estimated from temporal variance in allelic frequencies. Aquaculture
108, 215-232.
157

Hedrick, P., 2005. Large variance in reproductive success and the Ne/N ratio. Evolution
59, 1596-1599.
Heggberget, T.G., 1988. Timing of spawning in Norwegian Atlantic salmon (Salmo
salar). Can J Fish Aquat Sci 45, 845-849.
Heikkinen, R.K., Luoto, M., Araujo, M.B., Virkkala, R., Thuiller, W., Sykes, M.T., 2006.
Methods and uncertainties in bioclimatic envelope modelling under climate change.
Prog Phys Geog 30, 751-777.
Heins, D.C., Baker, J.A., Guill, J.M., 2004a. Seasonal and interannual components of
intrapopulation variation in clutch size and egg size of a darter. Ecol Freshw Fish 13,
258-265.
Heins, D.C., Baker, J.A., Guill, J.M., So, 2004b. Seasonal and interannual components of
intrapopulation variation in clutch size and egg size of a darter. Ecol Freshw Fish 13,
258-265.
Hendry, A.P., Berg, O.K., Quinn, T.P., 1999. Condition dependence and adaptation-bytime: breeding date, life history, and energy allocation within a population of salmon.
Oikos 85, 499-514.
Hendry, A.P., Day, T., 2005. Population structure attributable to reproductive time:
isolation by time and adaptation by time. Mol Ecol 14, 901-916.
Hendry, A.P., Day, T., Mj, 2005. Population structure attributable to reproductive time:
isolation by time and adaptation by time. Mol Ecol 14, 901-916.
Hendry, A.P., Wenburg, J.K., Bentzen, P., Volk, E.C., Quinn, T.P., 2000. Rapid
evolution of reproductive isolation in the wild: Evidence from introduced salmon.
Science (Washington D C) 290, 516-518.
Hernaman, V., Munday, P.L., 2007. Evolution of mating systems in coral reef gobies and
constraints on mating system plasticity. Coral Reefs 26, 585-595.
Hernandez-Leon, S., 2008. Natural variability of fisheries and lunar illumination: a
hypothesis. Fish Fish 9, 138-154.
Hill, W.G., 1972. Effective Size of Populations with Overlapping Generations. Theor
Popul Biol 3, 278-&.
Hill, W.G., 1981. Estimation of Effective Population-Size from Data on Linkage
Disequilibrium. Genetical Research 38, 209-216.
Hogan, J.D., Mora, C., 2005. Experimental analysis of the contribution of swimming and
drifting to the displacement of reef fish larvae. Mar Biol 147, 1213-1220.

158

Holey, M., Baker, E.A., Thuemler, T.F., Elliott, R.F., 2000. Research and assessment
needs to restore Lake Sturgeon in the Great Lakes: results of a workshop sponsored
by the Great Lakes Fishery Trust., Lansing, MI.
Houston, J.J., 1987. Status of the lake sturgeon, Acipenser fulvescens, in Canada. Can
Field Nat 101, 171-185.
Hutchings, J.A., 1991. Fitness Consequences of variation in egg size and food abundance
in brook trout Salvelinus fontinalis. Evolution 45, 1162-1168.
Jennions, M.D., Petrie, M., 2000. Why do females mate multiply? A review of the
genetic benefits. Biol Rev 75, 21-64.
Johannesen, J., Lubin, Y., 1999. Group founding and breeding structure in the subsocial
spider Stegodyphus lineatus (Eresidae). Heredity 82, 677-686.
Johnson, J.B., Omland, K.S., 2004. Model selection in ecology and evolution. Trends
Ecol Evol 19, 101-108.
Johnson, S.N., Zhang, X.X., Crawford, J.W., Gregory, P.J., Young, I.M., 2007. Egg
hatching and survival time of soil-dwelling insect larvae: A partial differential
equation model and experimental validation. Ecol Modelling 202, 493-502.
Jones, A.G., Ardren, W.R., 2003. Methods of parentage analysis in natural populations.
Mol Ecol 12, 2511-2523.
Jones, A.G., Small, C.M., Paczolt, K.A., Ratterman, N.L., 2010. A practical guide to
methods of parentage analysis. Mol Ecol Resour 10, 6-30.
Jones, O.R., Wang, J.L., 2010. COLONY: a program for parentage and sibship inference
from multilocus genotype data. Mol Ecol Resour 10, 551-555.
Jorgensen, C., Dunlop, E.S., Opdal, A.F., Fiksen, O., 2008. The evolution of spawning
migrations: state dependence and fishing-induced changes. Ecology 89, 3436-3448.
Kalinowski, S.T., Taper, M.L., Marshall, T.C., 2007. Revising how the computer
program CERVUS accommodates genotyping error increases success in paternity
assignment. Mol Ecol 16, 1099-1106.
Kamler, E., 1992. Early life history of fish : an energetics approach. Van Nostrand
Reinhold [distributor], London Chapman & Hall ; New York, NY.
Kamler, E., 2002. Ontogeny of yolk-feeding fish: an ecological perspective. Rev Fish
Biol Fisher 12, 79-103.
Kamler, E., 2005. Parent-egg-progeny relationships in teleost fishes: an energetics
perspective. Rev Fish Biol Fisher 15, 399-421.

159

Keckeis, H., Bauer-Nemeschkal, E., Menshutkin, V.V., Nemeschkal, H.L., Kamler, E.,
2000. Effects of female attributes and egg properties on offspring viability in a
rheophilic cyprinid, Chondrostoma nasus. Can J Fish Aquat Sci 57, 789-796.
Keesing, J.K., Halford, A.R., 1992. Importance of Postsettlement Processes for the
Population-Dynamics of Acanthaster planci (L). Australian Journal of Marine and
Freshwater Research 43, 635-651.
Kempinger, J.J., 1988. Spawning and early life history of lake sturgeon in the Lake
Winnebago system. American Fisheries Society Symposium 5, 110-112.
Kincaid, H., 1999. Genetic Considerations for captive breeding programs to preserve
genetic variability. Great Lakes lake sturgeon genetics: Status, needs, and
standardization, Chicago, Illinois.
King, T.L., Lubinski, B.A., Spidle, A.P., 2001. Microsatellite DNA variation in Atlantic
sturgeon (Acipenser oxyrinchus oxyrinchus) and cross-species amplification in the
Acipenseridae. Conserv Genet 2, 103-119.
Kokita, T., Nakazono, A., 1998. Plasticity in the mating system of the longnose filefish,
Oxymonacanthus longirostris, in relation to mate availability. J Ethol 16, 81-89.
Konovalov, D.A., Heg, D., 2008a. Estimation of population allele frequencies from small
samples containing multiple generations. The 6th Asia-Pacific Bioinformatics
Conference, Kyoto, Japan, pp. 321-331.
Konovalov, D.A., Heg, D., 2008b. A maximum-likelihood relatedness estimator allowing
for negative relatedness values. Mol Ecol Resour 8, 256-263.
Konovalov, D.A., Manning, C., Henshaw, M.T., 2004. KINGROUP: a program for
pedigree relationship reconstruction and kin group assignments using genetic
markers. Mol Ecol Notes 4, 779-782.
Krimbas, C.B., Tsakas, S., 1971. Genetics of Dacus oleae .5. Changes of esterase
polymorphism in a natural population following insecticide control-selection or drift.
Evolution 25, 454-&.
Kynard, B., Parker, E., 2005. Ontogenetic behavior and dispersal of Sacramento River
white sturgeon, Acipenser transmontanus, with a note on body color. Environ Biol
Fish 74, 19-30.
Kynard, B., Parker, E., Pugh, D., Parker, T., 2007. Use of laboratory studies to develop a
dispersal model for Missouri River pallid sturgeon early life intervals. J Appl Ichthyol
23, 365-374.
Launey, S., Barre, M., Gerard, A., Naciri-Graven, Y., 2001. Population bottleneck and
effective size in Bonamia ostreae-resistant populations of Ostrea edulis as inferred by
microsatellite markers. Gen Res 78, 259-270.
160

Lee, P.L.M., 2008. Molecular ecology of marine turtles: New approaches and future
directions. J Exp Mar Biol Ecol 356, 25-42.
Leis, J.M., 2006. Are larvae of demersal fishes plankton or nekton?. Adv Mar Biol, Vol
51, pp. 57-141.
Levitan, D.R., 2004. Density-dependent sexual selection in external fertilizers: Variances
in male and female fertilization success along the continuum from sperm limitation to
sexual conflict in the sea urchin Strongylocentrotus franciscanus. Am Nat 164, 298309.
Levitan, D.R., 2005. The distribution of male and female reproductive success in a
broadcast spawning marine invertebrate. Integr Comp Biol 45, 848-855.
Levitan, D.R., Fukami, H., Jara, J., Kline, D., McGovern, T.M., McGhee, K.E., Swanson,
C.A., Knowlton, N., 2004. Mechanisms of reproductive isolation among sympatric
broadcast-spawning corals of the Montastraea annularis species complex. Evolution
58, 308-323.
Levitan, D.R., Petersen, C., 1995. Sperm Limitation in the Sea. Trends Ecol Evol 10,
228-231.
Li, G., Hedgecock, D., 1998. Genetic heterogeneity, detected by PCR-SSCP, among
samples of larval Pacific oysters (Crassostrea gigas) supports the hypothesis of large
variance in reproductive success. Can J Fish Aquat Sci 55, 1025-1033.
Luikart, G., Cornuet, J.M., 1999. Estimating the effective number of breeders from
heterozygote excess in progeny. Genetics 151, 1211-1216.
Luikart, G., Ryman, N., Tallmon, D.A., Schwartz, M.K., Allendorf, F.W., 2010.
Estimation of census and effective population sizes: the increasing usefulness of
DNA-based approaches. Conserv Genet 11, 355-373.
Mackay, T.F.C., 2007. Wild populations are smaller than we think: a commentary on
'Effective population size/adult population size ratios in wildlife: a review' by Richard
Frankham. Genet Res 89, 489-489.
Maes, G.E., Pujolar, J.M., Hellemans, B., Volckaert, F.A.M., 2006. Evidence for
isolation by time in the European eel (Anguilla anguilla L.). Mol Ecol 15, 2095-2107.
Marshall, D.J., Evans, J.P., 2005. Does egg competition occur in marine broadcastspawners? J Evolution Biol 18, 1244-1252.
Marshall, T.C., Slate, J., Kruuk, L.E.B., Pemberton, J.M., 1998. Statistical confidence for
likelihood-based paternity inference in natural populations. Mol Ecol 7, 639-655.
McQuown, E., Gall, G.A.E., May, B., 2002. Characterization and inheritance of six
microsatellite loci in lake sturgeon. T Am Fish Soc 131, 299-307.
161

McQuown, E.C., Sloss, B.L., Sheehan, R.J., Rodzen, J., Tranah, G.J., May, B., 2000.
Microsatellite analysis of genetic variation in sturgeon: New primer sequences for
Scaphirhynchus and Acipenser. T Am Fish Soc 129, 1380-1388.
Milligan, B.G., 2003. Maximum-likelihood estimation of relatedness. Genetics 163,
1153-1167.
Mion, J., Stein, R., Marschall, E., 1998. River discharge drives survival of larval walleye.
Ecol Appl, 88-103.
Mittelbach, G.G., Persson, L., 1998. The ontogeny of piscivory and its ecological
consequences. Can J Fish Aquat Sci 55, 1454-1465.
Moberg, P.E., Burton, R.S., 2000. Genetic heterogeneity among adult and recruit red sea
urchins, Strongylocentrotus franciscanus. Mar Biol 136, 773-784.
Moller, A.P., Legendre, S., 2001. Allee effect, sexual selection and demographic
stochasticity. Oikos 92, 27-34.
Mousseau, T.A., Fox, C.W., 1998. Maternal effects as adaptations. Oxford University
Press, New York.
Neff, B.D., Fu, P., Gross, M.R., 2003. Sperm investment and alternative mating tactics in
bluegill sunfish (Lepomis macrochirus). Behav Ecol 14, 634-641.
Neff, B.D., Pitcher, T.E., 2005. Genetic quality and sexual selection: an integrated
framework for good genes and compatible genes. Mol Ecol 14, 19-38.
Nei, M., Maruyama, T., Wu, C.I., 1983. Models of Evolution of Reproductive Isolation.
Genetics 103, 557-579.
Ngo, L., Brand, R., 1997. Model selection in linear mixed effects models using SAS proc
mixed. Proceedings of the 22nd Annual SAS Users Group International Conference,
SAS Institute. Cary, North Carolina, pp. 1335-1340.
Nichols, S.J., Kennedy, G., Crawford, E., Allen, J., French, J., Black, G., Blouin, M.,
Hickey, J., Chernyak, S., Haas, R., Thomas, M., 2003. Assessment of lake sturgeon
(Acipenser fulvescens) spawning efforts in the lower St. Clair River, Michigan. J
Great Lakes Res 29, 383-391.
Nunney, L., 1991. The Influence of Age structure and fecundity on effective population
size. P Roy Soc Lond B Bio 246, 71-76.
Nunney, L., 1993. The influence of mating system and overlapping generations on
effective population size. Evolution 47, 1329-1341.
Nunney, L., 1996. The influence of variation in female fecundity on effective population
size. Biol J Linn Soc 59, 411-425.
162

O'Connor, M.I., Bruno, J.F., Gaines, S.D., Halpern, B.S., Lester, S.E., Kinlan, B.P.,
Weiss, J.M., Eb, 2007. Temperature control of larval dispersal and the implications
for marine ecology, evolution, and conservation. P Natl Acad Sci USA 104, 12661271.
Oliehock, P.A., Windig, J.J., van Arendonk, J.A.M., Bijma, P., 2006. Estimating
relatedness between individuals in general populations with a focus on their use in
conservation programs. Genetics 173, 483-496.
Osborne, M.J., Davenport, S.R., Hoagstrom, C.W., Turner, T.F., 2010. Genetic effective
size, Ne, tracks density in a small freshwater cyprinid, Pecos bluntnose shiner
(Notropis simus pecosensis). Mol Ecol 19, 2832-2844.
Palstra, F.P., O'Connell, M.F., Ruzzante, D.E., 2009. Age structure, changing
demography and effective population size in atlantic salmon (Salmo salar). Genetics
182, 1233-1249.
Palstra, F.P., Ruzzante, D.E., 2008. Genetic estimates of contemporary effective
population size: what can they tell us about the importance of genetic stochasticity for
wild population persistence? Mol Ecol 17, 3428-3447.
Paradis, A., Pepin, P., Brown, J., 1996. Vulnerability of fish eggs and larvae to predation:
Review of the influence of the relative size of prey and predator. Can J Fish Aquat
Sci, 1226-1235.
Pearse, D.E., Anderson, E.C., 2009. Multiple paternity increases effective population
size. Mol Ecol 18, 3124-3127.
Pemberton, J.M., 2008. Wild pedigrees: the way forward. P R Soc B 275, 613-621.
Pepin, P., 2009. The Impacts of Environmental Change and Ecosystem Structure on the
Early Life Stages of Fish: A Perspective on Establishing. In: Beamish, R.J.,
Rothschild, B.J. (Eds.), The Future of Fisheries Science in North America. Springer
Netherlands, pp. 255-274.
Pepin, P., Orr, D.C., Anderson, J.T., 1997. Time to hatch and larval size in relation to
temperature and egg size in Atlantic cod (Gadus morhua). Can J Fish Aquat Sci 54,
2-10.
Petersen, C.W., Warner, R.R., 1998. Sperm competition in Fishes. In: Birkhead, T.,
Moller, A. (Eds.), Sperm competition and sexual selection. Academic, London, pp.
435–463.
Petersen, C.W., Warner, R.R., Shapiro, D.Y., Marconato, A., 2001. Components of
fertilization success in the bluehead wrasse, Thalassoma bifasciatum. Behav Ecol 12,
237-245.

163

Peterson, D.L., Vecsei, P., Jennings, C.A., 2007. Ecology and biology of the lake
sturgeon: a synthesis of current knowledge of a threatened North American
Acipenseridde. Rev Fish Biol Fisher 17, 59-76.
Pilot, M., Dahlheim, M.E., Hoelzel, A.R., 2010. Social cohesion among kin, gene flow
without dispersal and the evolution of population genetic structure in the killer whale
(Orcinus orca). J Evolution Biol 23, 20-31.
Pinheiro, J.C., Bates, D.M., 2000. Mixed-effects models in S and S-PLUS. Springer, New
York.
Pritchard, G., Harder, L.D., Mutch, R.A., 1996. Development of aquatic insect eggs in
relation to temperature and strategies for dealing with different thermal environments.
Biol J Linn Soc 58, 221-244.
Pudovkin, A.I., Zaykin, D.V., Hedgecock, D., 1996. On the potential for estimating the
effective number of breeders from heterozygote-excess in progeny. Genetics 144,
383-387.
Pujolar, J.M., Maes, G.E., Volckaert, F.A.M., 2006. Genetic patchiness among recruits in
the European eel Anguilla anguilla. Mar Ecol-Prog Ser 307, 209-217.
Queller, D.C., Goodnight, K.F., 1989. Estimating relatedness using genetic markers.
Evolution 43, 258-275.
Quinn, T., Unwin, M., Kinnison, M., 2000. Evolution of temporal isolation in the wild:
Genetic divergence in timing of migration and breeding by introduced chinook
salmon populations. Evolution, 1372-1385.
R-Development-Core-Team, 2009. R: A language and environment for statistical
computing. R Foundation for Statistical Computing, Vienna, Austria. ISBN 3900051-07-0, URL http://www.R-project.org.
Ratnayeke, S., Tuskan, G.A., Pelton, M.R., 2002. Genetic relatedness and female spatial
organization in a solitary carnivore, the raccoon, Procyon lotor. Mol Ecol 11, 11151124.
Raymond, M., Rousset, F., 1995. Genepop (Version-1.2) - Population genetics software
for exact tests and ecumenicism. J Hered 86, 248-249.
Reed, T.E., Warzybok, P., Wilson, A.J., Bradley, R.W., Wanless, S., Sydeman, W.J.,
2009. Timing is everything: flexible phenology and shifting selection in a colonial
seabird. J Anim Ecol 78, 376-387.
Reitzel, A.M., Miner, B.G., McEdward, L.R., 2004. Relationships between spawning
date and larval development time for benthic marine invertebrates: a modeling
approach. Mar Ecol-Prog Ser 280, 13-23.

164

Reynolds, J.D., 1996. Animal breeding systems. Trends Ecol Evol 11, A68-A72.
Ritland, K., 2000. Marker-inferred relatedness as a tool for detecting heritability in
nature. MOLECULAR ECOLOGY 9, 1195-1204.
Rosenberg, M.S., 2001. PASSaGE: Pattern Analysis, Spatial Statistics, and Geographic
Exegesis. Version xxxx. Department of Biology, Arizona State University, Tempe,
AZ.
Rousset, F., 2000. Genetic differentiation between individuals. J Evolution Biol 13, 5862.
Rousset, F., 2008. GENEPOP ' 007: a complete re-implementation of the GENEPOP
software for Windows and Linux. Mol Ecol Resour 8, 103-106.
Rowe, S., Hutchings, J.A., 2003. Mating systems and the conservation of commercially
exploited marine fish. Trends Ecol Evol 18, 567-572.
Rowe, S., Hutchings, J.A., Skjaeraasen, J.E., 2007. Nonrandom mating in a broadcast
spawner: mate size influences reproductive success in Atlantic cod (Gadus morhua).
Can J Fish Aquat Sci 64, 219-226.
Sadovy, Y., Rosario, A., Roman, A., 1994. Reproduction in an Aggregating Grouper, the
Red Hind, Epinephelus guttatus. Environ Biol Fish 41, 269-286.
Schluter, D., 1996. Ecological speciation in postglacial fishes. Philos T Roy Soc B 351,
807-814.
Schultz, E.T., 1993. The effect of birth date on fitness of female dwarf perch,
Micrometrus minimus (Perciformes, Embiotocidae). Evolution 47, 520-539.
Seamons, T.R., Bentzen, P., Quinn, T.P., 2004. The mating system of steelhead,
Oncorhynchus mykiss, inferred by molecular analysis of parents and progeny.
Environ Biol Fish 69, 333-344.
Serbezov, D., Bernatchez, L., Olsen, E.M., Vollestad, L.A., 2010. Mating patterns and
determinants of individual reproductive success in brown trout (Salmo trutta)
revealed by parentage analysis of an entire stream living population. Mol Ecol 19,
3193-3205.
Shanks, A., 2009. Pelagic Larval Duration and Dispersal Distance Revisited. Biological
Bulletin 216, 373-385.
Shuster, S.M., Wade, M.J., 2003. Mating systems and strategies. Princeton University
Press, Princeton, N.J.
Siegel, D.A., Kinlan, B.P., Gaylord, B., Gaines, S.D., 2003. Lagrangian descriptions of
marine larval dispersion. Mar Ecol-Prog Ser 260, 83-96.
165

Siitonen, L., Gall, G.A.E., 1989. Response to Selection for early spawn date in rainbow
trout, Salmo gairdneri. Aquaculture 78, 153-161.
Skinner, A.M.J., Watt, P.J., 2007. Phenotypic correlates of spermatozoon quality in the
guppy, Poecilia reticulata. Behav Ecol 18, 47-52.
Slagsvold, T., Lifjeld, J.T., 1994. Polygyny in Birds - the role of competition between
females for male parental care. Am Nat 143, 59-94.
Smith, K.M., Baker, E.A., 2005. Characteristics of spawning lake sturgeon in the Upper
Black River, Michigan. N Am J Fish Manage 25, 301-307.
Smith, K.M., King, D.K., 2005. Dynamics and extent of larval lake sturgeon Acipenser
fulvescens drift in the Upper Black River, Michigan. J Appl Ichthyol 21, 161-168.
Sogard, S.M., 1997. Size-selective mortality in the juvenile stage of teleost fishes: A
review. B Mar Sci 60, 1129-1157.
Stillwell, R.C., Fox, C.W., 2005. Complex patterns of phenotypic plasticity: Interactive
effects of temperature during rearing and oviposition. Ecology 86, 924-934.
Sugg, D.W., Chesser, R.K., 1994. Effective Population sizes with multiple paternity.
Genetics 137, 1147-1155.
Tallmon, D.A., Koyuk, A., Luikart, G., Beaumont, M.A., 2008. ONeSAMP: a program to
estimate effective population size using approximate Bayesian computation. Mol Ecol
Resour 8, 299-301.
Taylor, E.B., 1991. A Review of local adaptation in salmonidae, with particular reference
to pacific and Atlantic salmon. Aquaculture 98, 185-207.
Taylor, E.B., 1999. Species pairs of north temperate freshwater fishes: Evolution,
taxonomy, and conservation. Rev Fish Biol Fisher 9, 299-324.
Tetzlaff, D., Soulsby, C., Youngson, A.F., Gibbins, C., Bacon, P.J., Malcolm, I.A.,
Langan, S., 2005. Variability in stream discharge and temperature: a preliminary
assessment of the implications for juvenile and spawning Atlantic salmon. Hydrol
Earth Syst Sci 9, 193-208.
Tomaiuolo, M., Hansen, T.F., Levitan, D.R., 2007. A theoretical investigation of
sympatric evolution of temporal reproductive isolation as illustrated by marine
broadcast spawners. Evolution 61, 2584-2595.
Trippel, E.A., Kjesbu, O.S., Solemial, P., 1997. Effects of adult age and size structure on
reproductive output of marine fishes. In: Chambers, R.C., Trippel, E.A. (Eds.), Early
life history and recruitment in fish populations. Chapman and Hall, pp. 31-62.

166

Trippel, E.A., Neilson, J.D., 1992. Fertility and sperm quality of virgin and repeatspawning atlantic cod (Gadus morhua) and associated hatching success. Can J Fish
Aquat Sci 49, 2118-2127.
Turner, T.F., Wares, J.P., Gold, J.R., 2002. Genetic effective size is three orders of
magnitude smaller than adult census size in an abundant, estuarine-dependent marine
fish (Sciaenops ocellatus). Genetics 162, 1329-1339.
Valiente, A.G., Juanes, F., Nunez, P., Garcia-Vazquez, E., 2010. Brown trout (Salmo
trutta) invasiveness: plasticity in life-history is more important than genetic
variability. Biol Invasions 12, 451-462.
Van de Casteele, T., Galbusera, P., Matthysen, E., 2001. A comparison of microsatellitebased pairwise relatedness estimators. Mol Ecol 10, 1539-1549.
Verhulst, S., Vanbalen, J.H., Tinbergen, J.M., 1995. Seasonal Decline in Reproductive
Success of the Great Tit - Variation in Time or Quality. Ecology 76, 2392-2403.
Verner, J., Willson, M.F., 1966. Influence of habitats on mating systems of north
american passerine birds. Ecology 47, 143-&.
Walling, C.A., Pemberton, J.M., Hadfield, J.D., Kruuk, L.E.B., 2010. Comparing
parentage inference software: reanalysis of a red deer pedigree. Mol Ecol 19, 19141928.
Wang, J., 2007a. Parentage and sibship exclusions: higher statistical power with more
family members. Heredity 99, 205-217.
Wang, J., 2007b. Triadic IBD coefficients and applications to estimating pairwise
relatedness. Genetical Research 89, 135-153.
Wang, J., 2010. COANCESTRY: A program for simulating, estimating and analysing
relatedness and inbreeding coefficients. Mol Ecol Resour (in print).
Wang, J., Santure, A.W., 2009a. Parentage and Sibship Inference From Multilocus
Genotype Data Under Polygamy. Genetics 181, 1579-1594.
Wang, J.L., 2004. Sibship reconstruction from genetic data with typing errors. Genetics
166, 1963-1979.
Wang, J.L., 2009. A new method for estimating effective population sizes from a single
sample of multilocus genotypes. Mol Ecol 18, 2148-2164.
Wang, J.L., Santure, A.W., 2009b. Parentage and Sibship inference from multilocus
genotype data under polygamy. Genetics 181, 1579-1594.

167

Wang, Y.L., Binkowski, F.P., Doroshov, S.I., 1985. Effect of Temperature on Early
Development of White and Lake Sturgeon, Acipenser transmontanus and Acipenser
fulvescens. Environ Biol Fish 14, 43-50.
Waples, R.S., 1989. A Generalized-Approach for Estimating Effective Population-Size
from Temporal Changes in Allele Frequency. Genetics 121, 379-391.
Waples, R.S., 1990. Conservation genetics of Pacific salmon .2. Effective population size
and the rate of loss of genetic variability. J Hered 81, 267-276.
Waples, R.S., 1991. Genetic interactions between hatchery and wild salmonids - Lessons
from the Pacific-Northwest. Can J Fish Aquat Sci 48, 124-133.
Waples, R.S., 2006. A bias correction for estimates of effective population size based on
linkage disequilibrium at unlinked gene loci. Conserv Genet 7, 167-184.
Waples, R.S., 2010. Spatial-temporal stratifications in natural populations and how they
affect understanding and estimation of effective population size. Mol Ecol Resour 10,
785–796.
Waples, R.S., Do, C., 2008. LDNE: a program for estimating effective population size
from data on linkage disequilibrium. Mol Ecol Resour 8, 753-756.
Waples, R.S., Do, C., 2010. Linkage disequilibrium estimates of contemporary Ne using
highly variable genetic markers: a largely untapped resource for applied conservation
and evolution. Evol Appl 3, 244-262.
Weir, B.S., Cockerham, C.C., 1984. Estimating F-Statistics for the analysis of population
structure. Evolution 38, 1358-1370.
Weis, A.E., Kossler, T.M., 2004. Genetic variation in flowering time induces
phenological assortative mating: Quantitative genetic methods applied to Brassica
rapa. Am J Bot 91, 825-836.
Welsh, A.B., Blumberg, M., May, B., 2003. Identification of microsatellite loci in lake
sturgeon, Acipenser fulvescens, and their variability in green sturgeon, A medirostris.
Mol Ecol Notes 3, 47-55.
Welsh, A.B., Elliott, R.F., Scribner, K.T., Quinlan, H.R., Baker, E.A., Eggold, B.T.,
Holtgren, J.M., Krueger, C.C., May, B., 2010. Genetic guidelines for the stocking of
lake sturgeon (Acipenser fulvescens) in the Great Lakes Basin. Great Lakes Fishery
Commission.
Wilson, S., Norris, D.R., Wilson, A.G., Arcese, P., 2007. Breeding experience and
population density affect the ability of a songbird to respond to future climate
variation. P R Soc B 274, 2539-2545.

168

Winkle, W. V., Shuter, B. J., Holcomb, B. D., Jager, H. I., Tyler, J. A. & Whitaker, S. Y.,
1997. Regulation of energy acquisition and allocation to respiration, growth and
reproduction: simulation model and example using rainbow trout. In: Chambers,
R.C., Trippel, E.A. (Eds.), Early life history and recruitment in fish populations.
Chapman and Hall, pp. 103-137.
Woody, C.A., Olsen, J., Reynolds, J., Bentzen, P., 2000. Temporal variation in
phenotypic and genotypic traits in two sockeye salmon populations, Tustumena Lake,
Alaska. T Am Fish Soc 129, 1031-1043.
Woxvold, I.A., Mulder, R.A., 2008. Mixed mating strategies in cooperatively breeding
apostlebirds Struthidea cinerea. J Avian Biol 39, 50-56.

169