EXTRINSIC AND INTRIN SIC FACTORS AFFECTIN G ADULT REPRODUCTIVE BIOLOGY AND EARLY LI FE STAGE DEVELOPMENT OF LAKE STURGEON By Kari Jean Dammerman A DISSERTATION Submitted to Michigan State University In partial fulfillment of the require ments For the degree of Zoology D octor of Philosophy Ecology, Evolutionary Biology and Behavior - Dual Major 2015 ABSTRACT EXTRINSIC AND INTRIN SIC FACTORS AFFECTIN G ADULT REPRODUCTIVE BIOLOGY AND EARLY LI FE STAGE DEVELOPMENT OF LAKE STURGEON By Kari Jean Dammerman Environmental change is occurring at an unprecedented rate . Consequently, alterations in environmental cues such as temperature, discharge, and depth influence the reproductive behavior of adult fishes. M aternal environmental effec ts including selection of mates, spawning site, and maternal yolk - sac provisioning determine the environmental influences offspring experience during critical early development and growth periods. Using genetic parentage analysis and experimentally - manipul ated environmental regimes, I examined the effects of adult reproductive behavior and altered environmental conditions on larval lake sturgeon ( Acipenser fulvescens ) development and behavior. Lake sturgeon are long - lived, iteroparous fishes that reside in freshwater, and migrate to riverine areas in early spring and summer to spawn. Analyses in my studies were conducted using a well - known adult population of approximately 1200 lake sturgeon from Black Lake, Michigan. First, I individually marked all spawnin g adults , examined the demographic and environmental factors affecting the timing of the reproduction , collected blood samples to determine female ovary/ egg quality, and used genetically - based parentage analysis to assign female reproductive success (RS) . The time females spent on the spawning grounds was dependent on temperature, discharge, number of males, size of males, an interaction between temperature and size as well as discharge and number of males . Female RS was dependent on the above factors as we ll as the OSR; however, the time females spent in the river had no effect on female RS or her ovary/egg quality indicating there does not appear to be a reproductive cost in being plastic in the timing and location of spawning. Next, I empirically tested w hether adult - selected and anthropogenically modified egg incubation conditions affect larval traits at hatch and until the timing of emergence by experimentally manipulating velocity and thermal regimes. A genotype - by - environment interaction was detected f or all traits at hatch, and the greatest range in thermal treatments. Additionally, egg incubation conditions and genetic (family) effects affected growth and behavior at th e timing of emergence indicating that conditions Lastly, I quantified microhabitat variation at female egg deposition sites , collected wild - produced eggs from the st ream, quantified offspring trait variation at hatch and four weeks post - emergence , and used genet ically - based parentage analysis to determine the influence of genetic (parentage) effects on offspring traits. Traits at hatch were influenced by enviro nmental conditions at egg deposition sites, but traits post - emergence were influenced by additive genetic effects and potential differences in feeding efficiency. Overall, f indings indicate how environmental variation affects adult reproductive success, ti ming of spawning, and larval trait variation during critical growth and development stages providing a greater understanding of how the phenotypic and genotypic variation within wild populations will be affected by environmental change. Copyri ght by KARI JEAN DAMMERMAN 2015 v A CKNOWLEDGEMENTS First and foremost, I want to thank my advisor, Kim Scribner. I came to graduate school straight from my undergraduate degree, and would not have been nearly as successful if it wa experiments in the river at midnight or assisting me in submitting a grant proposal five minutes before the deadline, Kim was always willing to lend a helping hand and made my success in graduate school a top priority. Thank you, Kim, for a wonderful five years! Juan Pedro Steibel, Gary Mittelbach, Edward Baker, and Janette Boughman. Thank you for offering endless advice on all aspects of my graduate career. Additionally, I owe so much to all members of the Scribner lab and field crews for their support in my research. I especially want to thank Jeannette Kanefsky for always providing much needed advice, and for being a grea want to thank Hope Draheim, Lisette Delgado, Shairah Abdul Razak, John Bauman, Nathan Barton, Annie VanSickle, Jimmy Garavaglia, and Adam Umstead for not only supporting me in all aspects of my resear ch, but for your friendship. I feel so lucky to have gotten to know all of you. Next, I want to thank my mother, sisters, and the Old Man. When I told you I was going to graduate school in Michigan to study sturgeon, you provided nothing but endless love a nd support. Thank you for listening to my numerous calls about my research, and for reassuring me during moments when I doubted myself. I love all of also vi to Arica for your love a nd support during my studies and research. Whether it was wading the river to find the sturgeon, assisting in counting newly hatched larvae, making me coffee during late night work sessions, or helping me find research papers when I w as on a tight deadline , you helped me succeed so much in graduate school . graduate school far from home, your friends become just like family as they support you through daily struggles, class and research complications, comprehensive exams, and night dinners, dance outings at the Green Door when Starfarm was playing, and your help one night in cutting and gluing screens on over 200 plastic beer cups for my research were some of the best times I had during my time in East Lansing. I will miss Comparative Anatomy group, f ellow docents from the Potter Park Zoo, the know all of you during our activities, and thank you for always helping me be successful. My last thank you goes to all of the organizations that provided funding support for my studies and research. The Department of Zoology (now Integrative Biology) and EEBB program at Michigan State University provided educational support, financial support including numerous scholarships and g rants, and leadership opportunities that enhanced my professional development. Thank you also to the International Association for Great Lakes Research, American Fisheries Society, Michigan Department of Natural Resources, Great Lakes Fishery Trust, Great Lakes Fishery Commission, and Michigan chapter of Sturgeon for Tomorrow for funding my research. vii TABLE OF CONTENTS LIST OF TABLES ................................ ................................ ................................ ............ ix LIST OF FIGURES ................................ ................................ ................................ .......... xi INTRODUCTION ................................ ................................ ................................ ............. 1 Chapter 1 ................................ ................................ ................................ ........................ 8 ASSOCIATIONS BETWEEN EGG QUALITY, BEHAVIO RAL PLASTICITY, AND REPRODU CTIVE SUCCESS OF SPA WNING FEMALE LAKE ST URGEON INTRODUCTION ................................ ................................ ................................ ....... 9 MATERIALS AND METHODS ................................ ................................ ................. 13 Study site ................................ ................................ ................................ ............ 13 Adult sampling ................................ ................................ ................................ .... 13 Blood collection and plasma hormone analysis ................................ .................. 14 Larval collection ................................ ................................ ................................ .. 16 Genetic analysis ................................ ................................ ................................ . 16 Parentage analysis ................................ ................................ ............................. 17 Statistica l analysis ................................ ................................ .............................. 18 RESULTS ................................ ................................ ................................ ................ 2 0 2012 adult data and parentage analysis ................................ ............................. 20 2013 adult data and parentage analysis ................................ ............................. 21 2013 plasma hormone analysis ................................ ................................ .......... 2 3 DISCUSSION ................................ ................................ ................................ ........... 2 4 APPENDIX ................................ ................................ ................................ ............... 29 LITERATURE CITED ................................ ................................ ............................... 36 C hapter 2 ................................ ................................ ................................ ...................... 43 GENETIC AND ENVIRONM ENTAL COMPONENTS OF PHENO TYPIC AND BEHAVIORAL TRAIT VAR IATION DURING LAKE S TURGEON EARLY ONTOGE NY I NTRODUCTION ................................ ................................ ................................ ..... 44 MATERIALS AND METHODS ................................ ................................ ................. 47 Study Site ................................ ................................ ................................ ........... 4 7 Sampling Locations and Gamete Collection ................................ ....................... 4 8 Field Experiment ................................ ................................ ................................ . 4 8 Velocity (Common Garden) Experiment ................................ ............................. 50 Statistical Analysis ................................ ................................ .............................. 51 RESULTS ................................ ................................ ................................ ................ 5 3 Field Experiment ................................ ................................ ................................ . 53 Velocity (Common Garden) Experiment ................................ ............................. 54 viii DISCUSSION ................................ ................................ ................................ ........... 55 APPENDIX ................................ ................................ ................................ ............... 62 LITERATURE CITED ................................ ................................ ............................... 70 C hapter 3 ................................ ................................ ................................ ...................... 78 INCREASING THERMAL R EGIMES REVEAL CRYPTI C GENETIC VARIATION DURING EARLY ONTOGEN ETIC STAGES OF LAKE STURGE ON INTRODUCTION ................................ ................................ ................................ ..... 78 MATERIALS AND METHODS ................................ ................................ ................. 82 Study site ................................ ................................ ................................ ............ 82 Fertilizations ................................ ................................ ................................ ....... 82 Thermal treatments ................................ ................................ ............................ 83 Larval traits ................................ ................................ ................................ ......... 83 Statistical analysis ................................ ................................ .............................. 85 RESULTS ................................ ................................ ................................ ................ 86 DISCUSSION ................................ ................................ ................................ ........... 88 Rearing temperatures influence phenotypic variation observed at hatch ........... 88 Fl uctuating and constant rearing temperatures influence the range of phenotypic variation observed in larval traits ................................ ................................ ........ 90 Rearing temperatures affect trait variation at a subsequent ontogenetic stage .. 91 Families differ in growth at a subsequent ontogenetic stage .............................. 93 Conclusions and future directions ................................ ................................ ....... 93 APPENDIX ................................ ................................ ................................ ............... 95 LITERATURE CITED ................................ ................................ ............................. 100 C h apter 4 ................................ ................................ ................................ .................... 10 7 EFFECTS OF PARENTAGE AND MICROHABITAT VAR IATION WITHIN ADULT - SELECTED SPAWNING SI TES ON LAKE STURGEON GROWTH DURING EA RLY LIFE STAGES INTRODUCTION ................................ ................................ ................................ ... 108 MATER IALS AND METHODS ................................ ................................ ............... 110 Study site ................................ ................................ ................................ .......... 110 Egg collection and microhabitat variables ................................ ........................ 111 Mean depth and discharge during the incubation period ................................ .. 112 Mean substrate size ................................ ................................ ......................... 113 Larval traits ................................ ................................ ................................ ....... 113 Genetic analysis ................................ ................................ ............................... 115 Parentage analysis ................................ ................................ ........................... 115 Statistical an alysis ................................ ................................ ............................ 116 RESULTS ................................ ................................ ................................ .............. 119 DISCUSSION ................................ ................................ ................................ ......... 121 APPENDIX ................................ ................................ ................................ ............. 127 LITERATURE CITED ................................ ................................ ............................. 135 CONCLUSION ................................ ................................ ................................ ............ 142 ix LIST OF TABLES Table 1.1 ................................ ................................ ................................ ....................... 30 Variable estimates and standard errors (SE) explaining changes in the expected log of mean female reproductive success (RS) and log o f mean retention time (RT). Estimates are provided for models of best fit for 2012 and 2013. Table 2.1 ................................ ................................ ................................ ....................... 63 Variance components and standard errors (SE) estimated from the model of best fit for the field and velocity experiments. P - values are presented from the likelihood ratio tests and heritability (h 2 ) estimates are provided for traits where only a significant family effect was detected. Table 2.2 ................................ ................................ ................................ ....................... 64 Daily mean, minimum, and maximum discharge levels (meters 3 /sec) during the egg incubation period at the experiment. Table 2.3 ................................ ................................ ................................ ....................... 65 Pairwise model comparisons for traits fit to a model including the interaction term and a reduced model without the interaction term. Log - likelihood values, chi - squ are values, p - values (P), and degrees of freedom (df) obtained from the likelihood ratio tests are presented where applicable. Table 2.4 ................................ ................................ ................................ ....................... 66 P - values computed from the likelihood ratio tests to determine the significance of family and site (or treatment) f or all traits from the field and velocity experiments. Values are not presented for models where a significant genotype - by - environment interaction was detected. Table 3.1 ................................ ................................ ................................ ....................... 96 Step - wise model comparisons and DIC - values used in model selection for the six la rval traits quantified at hatch and emergence in the experiment. Table 3.2 ................................ ................................ ................................ ....................... 97 Variance components and 95% highest posterior density (HPD) estimated from the models of best fit. Heritability (h 2 ) estimates are provided for traits where the model of best fi t included a significant family effect, but no family - by - treatment interaction. Table 4.1 ................................ ................................ ................................ ..................... 128 Mean (± Standard Error) values of the three microhabitat variables during the egg incubation period at the sixty - four sampling points where live lake sturgeon e ggs were collected. Depth and discharge during the egg incubation period were predicted using h ourly data logger measurements. x Table 4.2 ................................ ................................ ................................ ..................... 128 Mean ( ± Standard Error) and range in body size observed among larvae for the five measurement samples taken from ha tch (Age 1) to four - weeks post - emergence (Age 41). Table 4.3 ................................ ................................ ................................ ..................... 128 Coefficients, standard errors (SE), and p - values estimated from the models of best fit for the three traits measured at hatch. Table 4.4 ................................ ................................ ................................ ..................... 129 Pairwise model comparisons for the three hatch traits fit to models (M) including all three environmental variables and a reduced model. Log - likelihood values, chi - square values, p - values (P), and degrees of freedom (df) obtained from the likelihood ratio tests are presented. xi LIST OF FIGURES Figure 1.1 ................................ ................................ ................................ ...................... 31 The study site on the Upper Black River located in Michigan, USA. Shown are the location of the spawning sites and the larval collection site (a) . An enlarged view of the seven separate spawning locations where adults w ere sampled daily (b) . Figure 1.2 ................................ ................................ ................................ ...................... 32 The number of adult lake sturgeon present in the river along with mean daily water temperature (red) and river discharge (blue) durin g the adult spawning seasons in 2012 (a) and 2013 (b) . Figure 1.3 ................................ ................................ ................................ ...................... 33 Frequency distr ibution of spawning adult total length where males are gray bars and females are white bars for 2012 (a) and 2013 (b) . Figure 1.4 ................................ ................................ ................................ ...................... 34 Female plasma testosterone concentrations (ng/mL) and the number of days they were present on the spawning grounds (retent ion time). Figure 1.5 ................................ ................................ ................................ ...................... 34 Plasma estrogen level concentrations (ng/mL) for fish where sex was undetermined given the lack of expression of gametes. Figure 1.6 ................................ ................................ ................................ ...................... 35 The probability of females having normal ovaries estimated based on plasma testosterone leve ls and the number of days they were present on the spawning grounds (retention time). Figure 2.1 ................................ ................................ ................................ ...................... 67 The study location on the Upper Black River, the largest tributary of Black Lake, spawning site used during the 2011 spawning season Figure 2.2 ................................ ................................ ................................ ...................... 67 Four larval hatch traits A: body length (BL), B: body area (BA), C: yolk - sac area (YSA), D: head area (HA) (a) and one emergence trait, E: emergence body length(EBL), quantifie d in the experiment (b ) . The two additional emergence traits, time to emergence (ET) and total growth (TG), were quantified as the time (days) and growth (mm) from hatch to emergence, respectively. Scale bars (mm) are presented in the lower right corner F igure 2.3 ................................ ................................ ................................ ...................... 68 Total growth of lake sturgeon larvae from ha tch to emergence when eggs were xii genotype - by - environment (G - by - E) interaction was detected providing evidence that co nditions experienced during embryogenesis affect trait variation at a later ontogenetic stage (timing of emergence) Figure 2.4 ................................ ................................ ................................ ...................... 68 Reaction norm for body length at hatch for lake sturgeon larvae in the velocity (common garden) experiment. A significant gen otype - by - environment (G - by - E) interaction was velocity treatment Figure 3.1 ................................ ................................ ................................ ...................... 98 The study location on the Upper Black River, the largest tributary of Black Lake, Michigan sho wing the spawning locations where adult lake sturgeon were sampled during the 2012 spawning season Figure 3.2 ................................ ................................ ................................ ...................... 98 The four egg incubation temperatures (warm, cold, variable, and ambient) in which individuals were reared in during the 2012 spawning season Figure 3.3 ................................ ................................ ................................ ...................... 99 Thermal reaction norm for body length at hatch for lake sturgeon larvae subjected to the four thermal treatments. A significant genotype - by - environment (G - by - E) interaction was detected, and most of the variation among families was observed for larvae that hatched from eggs incubated in the warm (18 C) treatment Figure 3.4 ................................ ................................ ................................ ...................... 99 Thermal reaction norm for larval lake sturgeon growth measured from hatch to the timing of emergence. A significant family effect was detected (h 2 =0.33) indicating that differential size among families measured at hatch was p ersisting to a later ontogenetic s tage Figure 4.1 ................................ ................................ ................................ .................... 130 The study location on the Upper Black River, the largest tributary of Black Lake, Michigan showing the spawning site, Site B, where lake sturgeon eggs were collected along transect lines and the larval sampling location where wild - produced larvae were captured Figure 4.2 ................................ ................................ ................................ .................... 131 Schematic layout of the seven transect lines (A - G) and sampling points where the microhabitat variables were quantified and lake sturgeon eggs were collected. Bubbles indicate the number of live eggs collected at each transect sampling point. Figure 4.3 ................................ ................................ ................................ .................... 131 Individual growth trajectories quantified from hatch (Age 1 day) to four weeks post - xiii emergence (Age 41 days). Individual body size was measured at hatch, and once weekly between the timing of emergence (Age 21 days) and the end of the experiment (Age 41 days). Figure 4.4 ................................ ................................ ................................ .................... 132 Additive genetic variance ( ) and residual variance ( ) estimates quantified at each age in the random regression animal model. Figure 4.5 ................................ ................................ ................................ .................... 132 The narrow - sense heritability (h 2 ) estimates for body size quantified at each age the larvae were photographed. The h 2 esti mates were obtaining by dividing the additive genetic variance by the sum of the additive variance and residual variance. Figure 4.6 ................................ ................................ ................................ .................... 133 The relationship between the depth ratios calculated on May 8 th and May 18 th . The ratios were calculated as the point estimates of depth at each transect sampling point divided by the mean depth measurement from the data logger. The linear regression equation (y = 0.980x + 0.002, R 2 = 0.948) indicated the ratios were highly correlated between the two sampling dates. Fi gure 4.7 ................................ ................................ ................................ .................... 133 The relationship between the discharge ratios calculated on May 8 th and May 18 th . The ratios were calculated as the point estimates of discharge at each transect sampling point divided by the mean discharge measurement from the data logger. T he linear regression equation (y = 0.917x - 0.010, R 2 = 0.948) indicated the ratios were highly correlated between the two sampling dates. Figure 4.8 ................................ ................................ ................................ .................... 134 The relationship between mean substrate size when measuring rocks in three randomly selected grids versus the entire substrate photo. The linear regression equation ( y = 1.526x 1.315, R 2 = 0.954) indicated that measuring three random grids accurately estimated mean substrate size in the entire photo. 1 INTRODUCTION Environmental change is occurring at an unprecedented rate (IPCC 2014). The ability of populations to adapt to changes is determined by levels of phenotypic variance given that it is the raw material for selection to act on (Warner et al. 2014). U nderstanding how phenotypic vari ation within wild populations is changing in response to environmental change has become a major goal for researchers ( Parmesan 2006; Hansen et al. 2012). Within populations, c hanges in phenotypic variation occur due to microevolutionary responses or pheno typic p lasticity (Gienapp et al. 2008). Plasticity includes changes in behavior, phenotype, or physiology th at occur without genetic changes (West - Eberhard 2003), whereas microevolutionary responses are genetic changes to selection (Réale et al. 2003). Th ere has been considerable debate about the role of plasticity in adaptive evolution given that it is said to constrain adaptations to environmental change by shielding genotypes from selection (Ghalambor et al. 2007). However, plasticity is generally viewe d as a quantitative trait with a heritable basis that is under selection (Via et al., 1995; Pigliucci 2001; Scheiner 2002; DeWitt and Scheiner 2005) . Additionally, plasticity is believed to produce novel variants on which natural selection - Eberhard 2003) indicating its importance to phenotypic trait evolution in wild populations . The ability of populations to respond to environmental change is dependent upon several characteristics including the presence of genetic varia tion in fitness - related traits (Anderson et al. 2012). Therefore, p artitioning sources of phenotypic variation into environmental and additive genetic components are necessary for making predictions about species responses ( Crespel et al. 2013; Warner et a l. 2014). Levels of additive 2 genetic variation indicate how a trait can evolve in response to selection (Wilson et al. evolutionary potential (Garant et al. 2008). Envir onmental effects includ ing genotype - by - environment (G - by - E) interactions illustrate the possible similarities and differences in phenotypic trait variation among related and unrelated individuals when plotted as a reaction norm (Byers 2008). Additionally, m aternal environment al effects which include choice of mates, selection of oviposition sites, and maternal provisioning determine the environmental conditions the offspring is exposed to thereby having a strong influence on offspring traits (Mousseau and F ox 1998; Räsänen and Kruuk 2007; Noble et al. 2013). An additional characteristic influencing the ability of organisms to respond to environmental cha nge is generation time (Anderson et al. 2012) . Evolutionary have been observed in species that are small - bodied, have s hort - generation times, and have large population sizes (Bradshaw and Holzapfel 2006). However, species with long - generation times, late sexual maturity, and small population sizes are limited by their inabi lity to keep pace with environmental change. Thus, long - lived species are expected to experience population declines due to a lagged - response to environmental variation or a lack of genetic variation in traits (Bradshaw and Holzapfel 2006; Skelly et al. 20 07). Alternatively, behavioural plasticity is predicted to be an adaptive strategy for long - lived species adjusting to environmental change (Refsnider and Janzen 2012). In long - lived species, behavioral p lasticity is effective for escaping evolutionary an d ecological traps (Schlaep f er et al. 2002). For example, plasti city in oviposition site 3 selection for oviparous species allows females to mate when environmental conditions are favorable for offspring growth and survival thereby positively influencing her reproductive success (Serbezov et al. 2010). The timing of female mating can have large influences on offspring trait variation given that a biotic conditions within mating sites affect developmental timing and growth ( Garant et al. 2003; Crozier et al. 20 08; Refsnider and Janzen 2010 ) . Additionally, mating conditions can have long - term effects on offspring traits given that conditions experienced during e arly embryonic development are known to have persisting effects on offspring phenotypic trait variation et al. 2002; Warner and Shine 2008; Warner et al. 2014). In lo ng - lived vertebrates , environmental variability has been shown to affect adult reproductive success and early survival of offspring (Thompson and Ollafson 2001) ; however, the relationship between the timing of spawning and adult reproductive success in stream fishes in particular is unclear (Anderson et al. 2012) . The focus of my dissertation re search is de termining how extrinsic (e.g. environmental variability and demographic characteristics) and intrinsic factors (e.g. genetic parentage effects, egg quality, variability in spawning date) a ffect female spawning behavior and early life stage development of l ake sturgeon ( Acipenser fulvescens ). Lake sturgeon are long - lived, iteroparous fish that migrate to riverine areas between mid - April and early June to spawn (Auer 1996; Forsythe et al. 2012). Females are behaviorally plastic in the timing and location of s pawning events, and will remain in the river for several days to weeks before spawning (Thiem et al. 2013). During spawning events, females release demersal eggs which are fertilized by sperm from 4 mul tiple males. Eggs become adhesive , sticking to the cobbl e and gravel substrate among the spawning grounds (Bruch and Binkowski 2002) , and incubate wi thout parental care until hatch (Duong et al. 2011). Hatched larvae remain at the spawning grounds, feeding on endogenous yolk - sac reserves before dispersing downs tream to begin exogenously feeding (Smith and King 2005). The timing of ontogenetic shifts during the early life stages are dependent on several abiotic conditions including temperature, substrate, depth, and discharge; however, temperature has the largest influence on the timing of development. Lake sturgeon populations have been numerically depressed due to overharvest, habitat loss and degradation, and limited recruitment (Hayes and Caroffino 2012). Long - term success of rehabilitation programs includes protection of spawning and rearing habitats (Hayes and Caroffino 2012), and identification of factors that affect growth and survival during critical early life stages. My research was conducted on the Black Lake sturgeon population located in Cheboygan Co unty, Michigan. The w adable conditions of the system allowed for experimental work within the river to examine adult spawning behavior and collection of wild - produced progeny, and the presence of a streamside rearing facility allowed for experimental work examining the influences of environmental variability on early life stage development . My dissertation is composed of four chapters. In Chapter 1, I individually marked all spawning adults and examined the demographic and environmental factors affecting t he timing of the reproduction in female lake sturgeon over two spawning seasons. Based on observational data and previous studies examining the time females sturgeon spend on the spawning grounds (Auer 1996; Paragamian and Kruse 2001; Webb et al. 5 2002), I hypothesized that the greater amount of time that females spent in the river, the less viable her eggs would become (due to an increased probability in atresia). I collected blood samples to determine female egg quality, and used genetically - based parentag e analysis to assign reproductive success to determine w hether delays in female spawning (i.e. longer time spent on the spawning grounds) resulted in lower reproductive success. The results from this study are currentl y in preparation for publication with the intent to submit the manuscript to the Journal of Fish Biology . In Chapter 2, I empirically tested whether adult choice of spawning location influenced larval traits at hatch , and whether the effects of egg incubation conditions persisted until a seque ntial ontogenetic stage (emergence) . I placed eggs of known genotype at an early season spawning site (where adults were spawning) and a late season spawning site (where adults were not spawning at time) . I predicted that adults were selecting spawning sit es that were conducive to offspring growth, and that larval trait variation would vary between the sites. Given the high discharge levels observed during the spawning season, I also experimentally tested whether velocity regimes experienced while in the eg g would affect trait variation at hatch and persist until influence of increased environmental variability in discharge due to an anthropogenic disturbance on larval trait var iation during early ontogenetic stages. Results from this study were published in Environmental Biology of Fishes in the March issue of 2015. In Chapter 3, I empirically tested the effects of four thermal treatments (two fluctuating and two constant) on la rval trait variation at hatch and at the timing of emergence for ten half - sibling families 6 expected disturbance in thermal regime due to increases in extreme weather events as predicted due to climate change. I predicted that phenotypic variation would be most variable in the fluctuating treatments due to a stress response, and that families would differ in phenotypic variation given the results from the previous study. Results from this study were submitted to Evolutionary Applications in May of 2015, and are currently under review. In Chapter 4 , I quantified microhabitat variation in water depth, discharge, and substrate size at female egg deposition sites within a spawning site, and collected wild - produced eg gs from the stream just prior to hatch. I hypothesized that individual phenotypic variation would differ among sampling points and among materna l groups given the significant family effects observed in the previous two studies. I quantified offspring trait variation at hatch and up to four weeks post - emergence in wild - produced larvae , and used genetically - based parentage analysis to construct a pedigree . I tested for the influence of genetic (parentage) effects which were predicted to vary with age on offsp ring trait variation from hatch until four weeks post - emergence . The results from this study are currently in preparation for publication. Findings from my dissertation work improve our understanding of how environmental variation affects adult reproducti ve success , timing of spawning, and larval trait variation during critical growth and development stages. Information provides a greater understanding of the critical habitat needs of the adults during spawning activities, an understanding of how thermal a nd discharge conditions affect larval p henotypic variation , and provides insight into the ability of the population to respond to 7 environmental change given the estimates of narrow - sense heritability (h 2 ) during the early ontogenetic stages for this long - l ived, threatened species. 8 Chapter 1 ASSOCIATIONS BETWEEN EGG QUALITY, BEHAVIO RAL PLASTICITY, AND REPRODUCTIVE SUCCESS OF SPAWNING FEMALE L AKE STURGEON ABSTRACT Reproductive timing is a key indicator of populati on resilience. Environmental conditions and traits such as age, size, and operational sex ratio have well - documented influences on spawning behavior and reproductive success. However, the effect of individual variation in the timing of female spawning on e gg quality and reproductive success is unclear. In 2012 and 2013, we captured adult lake sturgeon (Acipenser fulvescens, N=247 and 271, respectively) among seven spawning locations in the Black River, Michigan. We recorded body length, sex, and individuall y marked fish using PIT and Floy tags. Location, group size, operational sex ratio, river discharge, and water temperature were quantified daily at spawning sites. Plasma steroid concentrations of estradiol - amples taken daily from females in 2013 to verify sex and determine ovary quality. Wild - produced larvae were captured during annual drift surveys, and genetic parentage analysis was used to estimate individual reproductive success. Timing of female spawnin g was influenced by mean temperature, maximum discharge, group size, and average body length of males. Mean concentrations of estradiol - pre - spawning females was 6.81 ± 1.51 ng/mL and 55.9 ± 3.92 ng/mL, respectively. Testosterone con centrations varied among females, but egg quality did not vary due to the timing of spawning. Female reproductive success varied due to temperature, discharge, group size, and operational sex ratio. Results indicate that plasticity in female spawning time may affect egg quality, but female reproductive success is largely 9 influenced by environmental conditions and access to mates which are increasingly variable in riverine systems. INTRODUCTION The timing of reproduction is a key indicator of population res ilience to environmental variation in fishes (Lowerre - Barbieri et al. 2011). Although reproductive timing has a heritable basis (Neuheimer and MacKenzie 2014), individuals show considerable plasticity in behavior in response to current environmental condit ions (Warren and Morbey 2012; Mittelbach et al. 2014) as well as a lag effect to recent environmental changes (Forsythe et al. 2012). For example, females choose the timing and location of spawning by selecting environments that are conducive to offspring survival and growth. Behavioral plasticity in reproductive timing is advantageous for females given that offspring survival largely determines variability in female reproductive success (RS) whereas male RS is typically limited by access to mates (Bateman 1948; Andersson 1994; Serbezov et al. 2010) or mate quality (McGuire et al. 2013). Understanding factors that affect the timing female reproduction provide an understanding of population stability and persistence in response to environmental changes. In ad dition to environmental conditions, demographic characteristics and traits including age, body size, and operational sex ratio (OSR) have been shown to influence the timing of reproduction and individual RS (Jørgensen et al. 2006; Wright and Trippel 2009; Lowerre - Barbieri et al. 2011). Older, larger females typically spawn earlier (Wright and Trippel 2009) and produce higher quality eggs and larvae of larger body size than smaller females (Schreck et al. 2001; Olin et al. 2012). Large size can also be 10 advan tageous for males in establishing breeding territories (Anderson et al. 2010) or when females delay spawning in the presence of small males as an indirect form of female mate choice (Berejikian et al. 2000). Additionally, earlier arrival of males to spawni ng grounds can inflate OSRs (i.e. the number of mature males to the number of mature females) thereby increasing the intensity of intra - sexual competition and the variance in individual RS (Emlen and Oring 1977; Wade and Arnold 1980; Quinn 1999). Demograp hic characteristics and environmental conditions also affect the timing of reproduction and individual RS by influencing gonadal development. Females determine the optimal timing of ovulation based on exogenous cues such as water temperature and social in teractions within spawning sites (Lowerre - Barbieri et al. 2011; Morgan et al. 2013). In teleost fishes, egg deposition typically occurs within days of ovarian maturation and ovulation given that eggs only remain viable for a short period of time (Bobe et a l. 2008; Warren and Morbey 2012). However, fishes such as sturgeon can maintain mature ovaries for months before spawning (Webb et al. 1999). Prolonged periods between ovulation and egg deposition allow females to be more selective in their choice of mates and spawning locations (Brown - Peterson and Heins 2009); however, extended delays in spawning have been shown to decrease egg viability due to over - ripening (Johnston et al. 2008; Thorstad et al. 2008) and increase the probability of degeneration and reabs In sturgeon species, field and experimental studies have shown that discharge and thermal conditions within riverine spawning sites affect reproductive behavior, ovarian maturation, and egg via bility (Dettlaff et al. 1993; Forsythe et al. 2012). Delays in female spawning due to rapid fluctuations in discharge during the spawning period 11 have been observed in several species (Khoroshko 1972; Auer 1996; Paragamian and Kruse 2001), resulting in redu ced egg viability for females that remained on the spawning grounds until conditions appeared favorable (Khoroshko 1972). Additionally, elevated temperatures prior to spawning have resulted in higher incidences of atresia and spawning failure in cultured w hite sturgeon (Webb et al. 2001). Quantifying female ovary and egg quality has been difficult in wild populations of large, long - lived fishes such as sturgeon (Morgan et al. 2008), and has led to the development of several non - invasive methods for detectin g early signs of atresia (Talbott et al. 2011). One such method includes quantifying plasma concentrations of testosterone (T) and estradiol - - spawning females, but show significant decreases when fish become atretic (Webb et al. 2001). In this study, we examined factors affecting the timing of reproduction in female lake sturgeon (Acipenser fulvescens, Rafinesque 1817), and quantified plasma sex steroid concentration to determine female ovary quality. Lake sturgeon are ite roparous, broadcast spawning fish that migrate to riverine areas to spawn (Auer 1996). Males typically spawn every 2 years and females every 3 - 5 years with high repeatability in the timing of spawning (Forsythe et al. 2012; Thiem et al. 2013). Spawning is multi - modal with adults migrating to spawning grounds from late April to early June when water temperature is between 8.8 21.1ºC (Bruch and Binkowski 2002; Forsythe et al. 2012). Males arrive on the spawning grounds earlier than females, and begin spermi ating once females are present. Arrival date on the spawning grounds is generally believed to be the date of spawning for females (Bruch and Binkowski 2002), but females have been observed on the spawning grounds for 12 several days when environmental conditi ons are variable (Auer 1996). Sex of the adults is typically determined by the expulsion of gametes; however, females are often difficult to identify given that eggs are not readily visible until ovulation which has been shown to occur among a wide range o f temperatures (Bruch and Binkowski 2002). Alternatively, plasma steroid concentrations can be used to differentiate sex (Craig et al. 2009; Shaw et al. 2012) and determine spawning state when gametes are not observable (Thiem et al. 2013). Spawning bouts last approximately 1 - 2 minutes with females releasing adhesive, demersal eggs over cobble and gravel substrate (Bruch and Binkowski 2002). Eggs are fertilized by sperm from multiple males (Thiem et al. 2013), and incubate without parental care until hatch (Duong et al. 2011). Hatched larvae remain at the spawning grounds, feeding on endogenous yolk - sac reserves for approximately 6 - 15 days dependent on water temperature before dispersing downstream to begin exogenously feeding (Smith and King 2005). The ob jectives of this study were to: i) examine how environmental conditions (temperature and discharge) and demographic factors (body size, group size, OSR, and inter - spawning interval) within spawning sites affect the timing of female spawning, ii) determine whether female RS quantified using genetic parentage analysis varied due to the environmental and demographic factors within spawning sites, iii) quantify female ovary quality using plasma concentrations of testosterone, and iv) determine whether the amoun t of time females spend on the spawning grounds before spawning affects individual RS and egg quality. Knowledge on species reproductive biology is vital for understanding critical habitat needs and factors affecting population - levels of 13 recruitment for fi shes that have been over - exploited or are being threatened with habitat loss due to human activities and environmental change. MATERIALS AND METHODS Study site The study was conducted in 2012 and 2013 on the Upper Black River (UBR) in Cheboygan County, Mic higan (Fig.1a). Adult lake sturgeon migrate into the UBR and spawn among seven sites located along a 1.5km stretch of the river (Fig 1b). Spawning activities are dependent on both temperature and discharge, and typically begin when water temperature is app et al. 2012). Wadable, shallow spawning sites (~1 - 3 meters) of the UBR allow daily access to nearly all adults during the spawning season. Additionally, the presence of a streamside facility all ows for the collection and rearing of wild - caught, larval lake sturgeon. Adult sampling Adult lake sturgeon were captured daily among the seven spawning sites using long - handled dip - nets. Total length (cm) and site of capture were recorded for each indivi dual. Sex was determined by applying pressure to the abdomen to expel gametes or by examining the urogenital opening. Each fish was individually marked with T - bar anchor (floy) tags (Floy Tag, Inc.) of different color combinations at the base of the dorsal fin designating sex, identification number, and the time of entry in the river. Fish were checked for internal passive integrated transponder (PIT) tags from previous years located near the dorsal scutes using hand - held scanners. PIT tag retention is appr oximately 90 - 95% in sturgeon species (Forsythe et al. 2012; Hamel et al. 2012) 14 making them ideal tags for identifying individuals during long - term studies (Clugston 1996). Group size (total number of males and females in a spawning group) and operational sex ratio (number of males to females; Emlen and Oring 1977) were quantified daily at each spawning site. Spawning date for each female was assigned based on visual spawning evidence, or as the last date observed within the river given that females typical ly complete spawning within 12 hours (Bruch and Binkowski 2002) and then promptly leave the spawning grounds (Peterson et al. 2007). Retention time (RT, days) was quantified for each female as spawning date minus the day of first capture in the river. Male RT was not considered given that females decide the timing, location, and duration of spawning (Bruch and Binkowski 2002). Inter - spawning interval was calculated for each female as the average number of years between spawning using long - term data collecte d from 2001 - 2011 (data not shown), and used as a proxy for age. Traditional aging techniques using fin spines typically under - estimated age and body size is only reliable until a particular age (Shaw et al. 2012). Therefore, first - time spawning females wer e considered younger than females that had previously spawned. Water temperature and discharge were measured hourly at the spawning sites throughout the spawning season using Onset HOBO pressure loggers (Figure 2; Cape Cod, Massachusetts, USA). Blood coll ection and plasma hormone analysis In 2013, we collected blood from the caudal vein of thirty - eight females each day they were present in the river. Seventy - five total blood samples were collected using 3 - mL syringes and 22 gauge needles coated in 1000 uni ts/mL of heparin. Syringes were 15 placed on ice and transported back to the streamside facility. Blood was divided into two 1.5 mL microcentrifuge tubes and centrifuged at 1500g for 15 minutes. Plasma was extracted and placed in new microcentrifuge tubes. Tu bes were stored at - 20°C until hormone analysis. The steroids, T and E2, were extracted from plasma following the method of Fitzpatrick et al. (1987) for analysis by radioimmunoassay (RIA). Briefly, 100 µL of plasma were extracted twice with 2 mL of dieth yl ether. Tubes were vortexed vigorously with ether, and the aqueous phase was removed by snap - freezing in liquid nitrogen. Ether was allowed to evaporate overnight under a chemical hood; the next day extracted steroids were re - suspended in 1 mL of phospha te - buffered saline with gelatin (PBSG), and 10 or 50 µL were assayed for each steroid depending on the concentration of steroid in the sample. Recovery efficiencies for all steroids were determined by adding titrated steroids to tubes containing plasma (n= 4), which were extracted as described above. All steroid assay results were corrected for recovery. Recovery efficiencies for T varied between 89% and 94%, with a mean of 92%. Estradiol recovery efficiency was 71%. Plasma concentrations of T and E2 were me asured by RIA as described in Fitzpatrick et al. (1986) modified by Feist et al. (1990). All samples were analyzed in duplicate. A slightly more concentrated charcoal solution (6.25 g charcoal and 4.0 g dextran/L PBSG) was used for all assays to reduce non - specific binding. The intra - and inter - assay coefficients of variation for all assays were less than 5 and 10%, respectively. Steroid levels were validated by verifying that serial dilutions were parallel 16 to standard curves. The lower limit of detection w as 0.10 ng/mL for T and 0.16 ng/mL for E2. Larval collection Larval lake sturgeon were sampled approximately 2km downstream of the spawning sites (Fig. 1a). Sampling began 10 days after the first observed spawning date and was conducted at night when larva e disperse downstream, actively searching for food (Auer and Baker 2002). Larvae were collected using five D - frame nets placed evenly across the river. Nets were checked hourly from 2100 - 0200 hours. Captured larvae were transported to the streamside facili ty and reared separately by capture night for approximately three months until individuals were large enough for nonlethal sampling of fin clips for genetic analysis. Larvae that died during the rearing period were preserved in 95% ethanol and kept separat e by capture night for genetic analysis. Genetic analysis In 2012, we genotyped 10% of the dead and 10% of the live larvae collected per night during the annual drift survey (N=820) for parentage analysis. In 2013, we genotyped all live and dead larvae col lected during the drift survey (N=845) due to low number of collected individuals. DNA was extracted from larval fin clips using DNeasy(R) extraction kits (QIAGEN, Inc.), and quantified using a Nanodrop 1000 spectrophotometer (Thermo Fisher Scientific, Inc .). Larvae were genotyped at 12 microsatellite loci: AfuG68B (McQuown et al. 2002); Spl 120 (McQuown et al. 2000); Aox 27 (King et al. 2001); AfuG68, AfuG74, AfuG56, AfuG195, AfuG9, AfuG63, AfuG204, AfuG160, and AfuG112 (Welsh et al. 2003). Polymerase chai n reaction and genotyping were conducted using methods described in Duong et al. 2011. However, 17 cycle number was increased to 35 for all loci during the annealing phase and 1.5uL of forward and reverse primer were used to increase signals. Two experienced lab personnel assigned genotyping scores to all gels. A 10% subset of the samples were re - genotyped to empirically estimate genotyping error which was calculated as the number of allelic errors divided by the total number of alleles. Parentage analysis Ge notypic data was used to assign parentage using two software programs, CERVUS 3.0 (Kalinowski et al. 2007) and COLONY 2.0.5.8 (Jones and Wang 2010). CERVUS is the most commonly used parentage analysis program, and assigns the most likely parent as the indi vidual with the highest natural log of the likelihood - odds ratio (LOD) score. COLONY assigns offspring into full and half - sibling families before identifying the most likely candidate parent (Jones et al. 2010). Both programs use maximum likelihood and acc ount for empirical estimates of genotyping errors. Assignment of the most likely parent pair in our study was based on concordance between the two programs. In CERVUS, we simulated 10,000 offspring with a strict level of 95% and a relaxed level of 80% to d etermine the confidence in our parentage assignment based on population estimates of allelic frequencies. Parentage analysis in CERVUS and COLONY included 243 adults in 2012 and 262 adults in 2013, an 85% estimate of the proportion of adults captured (K. S cribner, unpublished data), and empirical genotyping error estimates of 0.97% and 0.54% for 2012 and 2013, respectively. Adults with the highest, positive LOD score in CERVUS and highest maximum likelihood score in COLONY were chosen as the most likely can didate parents. Reproductive success was 18 quantified as the total number of offspring assigned to each female. Male RS was not considered given that males are limited by the number of receptive females. Statistical analysis Statistical analyses were perform ed using the package, glmmADMB (Skaug et al. 2006) implemented in the statistical program, R (R Development Core Team). Data from 2012 and 2013 were analyzed separately given the differences observed in environmental conditions and proportion of larvae gen otyped. We tested for normality in all variables using Shapiro - Wilk tests. We used generalized linear models to determine how female retention time and reproductive success varied as a function of demographic characteristics and abiotic factors. Demographi c characteristics included female body length, female inter - spawning interval, male total length for males that were present with the female, mean group size when a female was present, and the operational sex ratio. Abiotic factors included the daily mean, minimum, maximum, and range in temperature and discharge observed from day of capture to day of spawning for each female. Range in temperature and discharge were quantified as the daily maximum minus the daily minimum. All explanatory variables were stand ardized prior to analysis to have a mean of zero and standard deviation of one to allow for direct comparisons between variables. We tested for multicollinearity between variables by fitting a full model including all demographic and abiotic factors, quant ifying the variance Minimum and maximum temperature were removed from the model due to evidence of multicollinearity (>7) and high correlation (>0.7) with mean tempera ture. Additionally, we removed mean and minimum discharge which showed high correlation 19 with maximum discharge. Maximum discharge was included in the model given that rapid increases in discharge have been shown to affect the timing of female reproduction (Paragamian and Kruse 2001; Auer 1996). For retention time, we fit a negative binomial regression with a log link function as: RT = mean temperature + range in temperature + maximum discharge + range in discharge + female total length + female inter - spawni ng interval + mean male total length + mean group size + mean operational sex ratio + possible two - way interactions of biological relevance (Burnham and Anderson 2002). We fit the same full model with individual RS as the response variable, but included RT as an additional explanatory variable. We used negative binomial error distributions in models instead of a Poisson given the over - dispersion in retention time and reproductive success. Likelihood ratio tests were used to determine the model of best fit. For variables that were significant predictors of RT, we also tested whether there were significant differences in the abiotic factors and demographic characteristics between females that stayed in the river for one day versus females that were observed in the river for greater than one. We used non - parametric Mann - Whitney U tests to test this given the non - normality observed in the significant explanatory variables. Quantified plasma concentrations of T and E2 were used to verify female sex assignment for sixteen fish that: i) had an inter - spawning interval of less than 3 years, ii) could not be identified based on spawning behavior or expression of gametes, or iii) were spawning for the first time. Female sex was verified using classification functions de veloped for white sturgeon (Acipenser transmontanus Richardson 1836) by Webb et al. (2002) and recently applied to lake sturgeon (Craig et al. 2009; Shaw et al. 2012; 20 Thiem et al. 2013). Concentrations of T were used to determine the probability of a femal e having normal versus atretic ovaries using the logistic regression developed by Talbott et al (2011). Application of the logistic regression in white sturgeon resulted in the correct assignment of 95% of fish with normal ovaries and 93% of fish with atre tic ovaries (Talbott et al. 2011) indicating that it is a reliable indicator of ovary and egg quality when observational data is unavailable. RESULTS 2012 adult data and parentage analysis In 2012, adult lake sturgeon were observed in the river from Apri l 16th until June 4th. The mean river water temperature on the first day adults were present was 13ºC and maximum discharge was approximately 7.0 m 3 /s ec (Fig. 2a). Females were on average larger than males with a mean (±SE) total length of 169.6 (±1.54) cm and range of 135 to 192cm (Fig. 3a). Males averaged 146.9 (±1.05) cm and range of 105 to 190cm. Group sizes in the river ranged from 1 to 29 fish with a mean of 10 fish per group when females were present. OSR was skewed towards males with an OSR of 184 m ales to 63 females for the entire season, and an average of 3 males per female at the spawning sites. OSR ranged from 1 to 9 males per each female during the spawning season. Average inter - spawning interval for the females marked in 2012 was 3 (±0.23) year s, but ranged from 0 to 10 years with 32% of the spawning females being first - time spawners. RT ranged from 1 to 15 days spent in the river. Mean RT was 2.9 (±0.41) days with 60% of females spending only 1 day in the river. Mean RT was negatively associate d with mean temperature, maximum discharge, mean group size, mean male 21 total length, and an interaction between maximum discharge and mean group size with the interaction term having the largest influence (Table 1). None of these variables differed for fem ales that stayed in the river one day versus those that stayed in the river greater than one day (Mann - Whitney U tests, p>0.05 in all tests) Eleven of the twelve loci genotyped for the 2012 adults were in Hardy - Weinberg equilibrium. Allelic diversity rang ed from 2 to 11 with a mean of 5.42 alleles per locus and an expected heterozygosity of 0.59. The combined non - exclusion probability for the parent pairs was 0.000057. Seventy - nine percent of the offspring were assigned a female parent with a 75% degree of concordance in female assignments between CERVUS and COLONY. All sixty - two females in 2012 were assigned at least 2 offspring during parentage analysis. Female RS ranged from 2 to 38 offspring with a mean of approximately 10 (±1.02) larvae assigned to eac h female. Mean RS was positively associated with mean temperature, maximum discharge, and mean group size, and negatively associated with mean OSR and an interaction between maximum discharge and mean group size. The interaction term had the largest influe nce on mean RS (Table 1). 2013 adult data and parentage analysis In 2013, adult sturgeon were observed on the spawning grounds from May 1st until June 9th. Mean water temperature on the first day adults were present was 13ºC and maximum discharge was appro ximately 20.0 m 3 /s ec (Fig. 2b). Females were typically larger than males with a mean total length of 170.9 (±1.77) cm ranging from 141 to 198cm. Average male total length was 146.2 (±0.97) cm and ranged from 108 to 185 cm (Fig. 3b). Mean group size in the river was 10.9 (±0.67) fish per group and 22 ranged from 1 to 22 fish when females were present. The OSR for the entire season was 210 males to 61 females with an average of 3.4 males per each female on the spawning grounds. OSR ranged from 1 to 13 males per female during the spawning season. Average inter - spawning interval for the sixty - one females marked in 2013 was 3.4 (±0.4) years, and ranged from 0 to 11 years with 34% of the spawning females spawning for the first time. Mean RT was 3.5 (±0.57) days and ranged from 1 to 23 days spent in the river. Approximately 63% of females spent only 1 day in the river. Mean RT was negatively associated with mean temperature, maximum discharge, and mean group size, and positively associated with mean male total length, an interaction between mean temperature and mean male total length, and an interaction between maximum discharge and mean group size. Mean temperature had the largest influence on mean RT (Table 1). Maximum discharge (W = 303, p - value = 0.046), mean group size (W = 234.5, p - value = 0.003), and mean male total length (W = 591.5, p - value = 0.022) were significantly different for females that stayed in the river one day versus those that stayed in the river greater than one day. For females that stayed in the river for one day, maximum discharge was 2.3 m 3 /sec higher, mean group size had approximately 4.3 more fish, and mean male total length was 7.7cm smaller. Allelic diversity ranged from 2 to 11 with a mean of 5.33 alleles per locus and an expected heterozy gosity of 0.58 for the 12 loci genotyped for the 2013 adults. Eleven of the loci were in Hardy - Weinberg equilibrium. The combined non - exclusion probability for the parent pairs was 0.000066. Eighty - three percent of the offspring were assigned a female pare nt with an 80% degree of concordance in female assignments between 23 CERVUS and COLONY. Ninety - five percent of the spawning females were assigned at least one offspring. Female RS ranged from 3 to 37 offspring with a mean of approximately 11 (±1.12) larvae a ssigned to each female. Mean RS was positively associated with mean temperature and maximum discharge, and negatively associated with mean group size and an interaction between maximum discharge and mean group size. The interaction term had the largest inf luence on mean RS (Table 1). 2013 plasma hormone analysis On average, two blood samples were collected per female with some individuals being sampled up to nine times while on the spawning grounds. The average concentration of T in the seventy - five blood s amples was 55.9 ± 3.92 ng/mL. Samples showed significant variation in the concentration of T ranging from 0.29 to 136.94 ng/mL (Fig. 4). E2 concentrations averaged 6.81 ± 1.51 ng/mL and ranged from 0.05 to 18.26 ng/mL (Fig. 5). Of the sixteen fish with unv erified sex, fourteen were correctly assigned as female (~87.5%), one was determined to be male, and the sex of one fish could not be determined given that estrogen levels were too low to quantify. Based on testosterone concentrations, ninety - two percent of females had a high probability of having normal ovaries (Fig. 6) indicating ovary quality was not dependent on the number of days present on the spawning grounds. Two females showed signs of atretic ovaries and had testosterone levels less than 5 ng/mL on the day they spawned. One of the atretic females spent 13 days in the river and had high probabilities of normal ovaries (probability=0.88 - 098) until day nine when the probability dropped to 0.04 within 24 hours. The other atretic female was present on the spawning grounds for one day. 24 DISCUSSION Recent studies suggest that both demographic characteristics and environmental conditions within spawning sites can affect the timing of reproduction in fish (Wright and Trippel 2009; Lowerre - Barbieri et al. 20 11); however, our understanding of how the timing of reproduction affects individual reproductive success (RS) and egg quality is limited (Lowerre - Barbieri et al. 2011). In this study, we tracked individual fish movement during two spawning seasons to dete rmine the conditions that affect the number of days female lake sturgeon spend on the spawning grounds before spawning, and whether delays in spawning affect female ovary quality. Retention time (RT) was similar in both years with females averaging 2.9 to 3.5 days on the spawning grounds; however, females in 2013 stayed in the river approximately one week longer with one female remaining in the river for up to 23 days. Similarly, Thiem et al. (2013) observed lake sturgeon on the spawning grounds for an aver age of 5 days with some individuals staying for 27 days. In both years of the study, 60% - 63% of females were present on the spawning grounds for one day. In 2013, females that stayed in the river for only one day spawned at higher discharge levels when mea n group sizes were larger, but did not have higher RS given that RS was not dependent on the time females spent in the river. Remaining in the river for one day may be advantageous for adult lake sturgeon given that food availability and cover are typicall y limited on the spawning grounds (Auer 1996), and water depth decreases during the season. Mean female RT was significantly affected by both demographic and environmental conditions at the spawning sites. Male body size was a significant predictor of RT (Table 1) as average female RT decreased in 2012 as the average size 25 of males increased indicating that the presence of large males reduced the amount of times females spent in the river before spawning. Conversely, mean female RT in 2013 increased as the average size of males in the river increased due to an interaction between mean temperature and male body size. Female preference for large males has been well - documented in fishes given that large males are typically better competitors and produce larger offspring than small males (Dickerson et al. 2002; Dickerson et al. 2005). However, evidence of mate choice has been limited in sturgeon with studies documenting only subtle differences in spawning and reproductive success between males due to body size (D uong et al. 2010) or female preference for particular males (Kynard et al. 2011). Lack of evidence for female mate choice in lake sturgeon may also be related to size as females are on average larger than male conspecifics (Fig. 3). Male lake sturgeon mat ure earlier and spawn more frequently than females resulting in a greater number of males present on the spawning grounds than females (Peterson et al. 2007; Thiem et al. 2013). Given the polygamous mating behavior of sturgeon, mating with numerous males m aximizes female fertilization success and the genetic diversity of offspring (Bruch and Binkowski 2002), and has been shown to increase female RS (Duong et al. 2010). In our study, mean OSRs were 3:1 in 2012 and 3.4:1 in 2013 demonstrating the degree of ma le - bias on the spawning grounds. Our OSR estimates are comparable to ratios of 3:1 and 5.7:1 estimates in other wild populations of lake sturgeon (Auer 1999; Bruch and Binkowski 2002). The mean number of days female spent on the spawning grounds did not va ry due to OSR, but mean female RS was negatively associated with mean OSR in 2012 indicating that increases in the number of males did not increase female RS. 26 Water temperature and discharge are important cues directly affecting the timing of migration to spawning grounds and duration and location of spawning activities (Paragamian and Kruse 2001; Forsythe et al. 2012). In both years, fish arrived on the spawning grounds when mean water temperature was 13ºC which is within the optimal thermal range of 10 - 1 6ºC for spawning that has been observed in other studies (Auer 1996; Bruch and Binkowski 2002). In 2012 and 2013, increases in mean temperature reduced the mean amount of time females spent on the spawning grounds. Additionally, the largest influence on me an female RT in 2012 and female RS in 2012 and 2013 was the interaction between maximum discharge and group size. The negative association of the interaction term with female RT and RS indicates that increases in maximum discharge encountered by females wh ile in different groups reduced the amount of time females spent on the spawning grounds, and reduced mean female RS. Similarly, Paragamian and Kruse (2001) observed that females spent less time on the spawning grounds due to fluctuations in discharge whic h was predicted to reduce female RS and lead to lower levels of recruitment given that females spent less time on the spawning grounds. Quantification of plasma concentrations of T and E2 and use of discriminant classification functions provided a reliabl e means of verifying sex of spawning females in our study, and has been successfully applied in other lake sturgeon populations (Shaw et al. 2012; Thiem et al. 2013). Using the discriminant functions, we verified our female sex assignment based on cloacal morphology and determined that we correctly identified females 87.5% of the time. Results are similar to the 85% of female white sturgeon correctly assigned by Webb et al. (2002) during development of the 27 discriminant functions. Average T concentration in females was 55.9 ± 3.92 ng/mL and average E2 was 6.81 ± 1.51 ng/mL which was comparable to T and E2 concentrations quantified by Shaw et al. (2012) in spawning lake sturgeon. Delays in spawning have been shown to decrease fertilization rates and egg viabi lity due to over - ripening (Berejikian et al. 2000), and are predicted to reduce female RS. In our study, the time females spent on the spawning grounds did not significantly affect female RS. Additionally, the majority of females had a high probability of normal ovaries regardless of the time they spent on the spawning grounds (Fig. 6) suggesting that there is little to no reproductive cost for female lake sturgeon to wait on the spawning grounds until conditions appear favorable. However, the average T con centrations quantified in female lake sturgeon in the study was higher than white sturgeon T concentrations used to develop the threshold concentrations for the logistic regression used to determine the probability of ovaries. Additionally, one female in o ur study showed signs of atresia after spending nine days on the spawning grounds; due to time or due to stress associated with being captured and handled nine times whic h has been predicted to affect female sex steroid concentrations (Fuzzen et al. 2011; Thiem et al. 2013). Therefore, future studies should integrate histological samples or use of additional non - invasive techniques in addition to sex steroid concentration to verify early signs of atresia in lake sturgeon, and determine that atresia was not due to the stress from being captured multiple times. Female body size is typically correlated with offspring quality as large females typically produce higher quality eg gs and larvae. Additionally, larger older females are 28 hypothesized to spend less time on the spawning grounds and produce more viable eggs than younger, first time spawners although this has not been well - studied (Morgan et al. 2008). In our study, female RT and RS did not vary due to female body size or female inter - spawning interval. Several first - time spawning females spent the same amount of time on the spawning grounds as experienced, older females. Overall, our findings demonstrate the complexity of f actors affecting the timing of female spawning and illustrate why our understanding of reproductive timing in fishes has been limited (Lowerre - Barbieri et al. 2011). Future empirical work is needed that address the factors that affect female reproductive t iming and the probability of atresia, particularly due to inter - annual variability in environmental conditions. Knowledge within these areas is essential for developing effective management strategies that protect critical habitat, accurately predict popul ation levels of recruitment, and accurately predict how wild populations will be influenced by predicted environmental changes. 29 APPENDIX 30 Table 1. 1 Variable estimates and standard errors (SE) explaining c hanges in the expected log of mean female reproductive success (RS) and log of mean retention time (RT). Estimates are provided for models of best fit for 2012 and 2013. Trait Component Estimate SE (a) 2012 RT Mean Temperature - 0.32 0.10 Max Discharge - 0.04 0.14 Group Size - 0.08 0.15 Male Total Length - 0.19 0.11 Max Discharge*Group Size - 0.70 0.21 RS Mean Temperature 0.36 0.09 Max Discharge 0.09 0.09 Group Size 0.34 0.10 Operational Sex Ratio - 0.30 0.10 Max Discharge*Group Size - 0.51 0.13 (b) 2013 RT Mean Temperature - 0.52 0.13 Max Discharge - 0.22 0.13 Group Size - 0.47 0.14 Male Total Length 0.13 0.13 Mean Temperature*Male Total Length 0.29 0.12 Max Discharge*Group Size 0.40 0.14 RS Mean Temperature 0.22 0.09 Max Discharge 0. 07 0.09 Group Size - 0.02 0.09 Max Discharge*Group Size - 0.30 0.09 * Indicates an interaction between variables. 31 Figure 1.1 T he study site on the Upper Black River located in Michigan, USA. Shown are the location of the spawning sites and the larva l collection site (a). An enlarged view of the seven separate spawning locations where adults were sampled daily (b). 32 Figure 1. 2 The number of adult lake sturgeon present in the river along with mean daily water temperature (red) and river discharge (b lue) during the adult spawning seasons in 2012 (a) and 2013 (b). 33 Figure 1. 3 Frequency distribution of spawning adult total length where males are gray bars and females are white bars for 2012 (a) and 2013 (b). 34 Figure 1. 4 Female plasma testoster one concentrations (ng/mL) and the number of days they were present on the spawning grounds (retention time). Figure 1. 5 Plasma estrogen level concentrations (ng/mL) for fish where sex was undetermined given the lack of expression of gametes. 35 Fi gure 1. 6 The probability of females having normal ovaries estimated based on plasma testosterone levels and the number of days they were present on the spawning grounds (retention time). 36 LITERATURE CITED 37 LITERATURE CITED Andersson, M. (1994). Sexual Selection . Princeton University Press, Princeton. Anderson, J.H., Faulds, P.L., Atlas, W.I, Pess, G.R., & Quinn, T.P. (2010). Selection on breeding date and body size in colonizing coho salmon Oncorhyn chus kisutch. Molecular Ecology 19 :2562 - 2573 Auer, N. (1996). Response of spawning lake sturgeons to change in hydroelectric facility operation. Transactions of the American Fisheries Society 125 :66 - 77 Auer, N.A. (1999). Population characteristics and mo vement of lake sturgeon in the Sturgeon River and Lake Superior. Journal of Great Lakes Research 25(2) :282 - 293 Auer, N.A. & Baker, E.A. (2002). Duration and drift of larval lake sturgeon in the Sturgeon River, Michigan. Journal of Applied Ichthyology 18 :5 57 - 564 Bateman, A.J. (1948). Intrasexual selection in Drosophila. Heredity 2 :349 - 368 Berejikian, B.A., Tezak, E.P., & LaRae, A.L. (2000). Female mate choice and spawning behaviour of chinook salmon under experimental conditions. Journal of Fish Biology 5 7 :647 - 661 Bobe, J., Jalabert, B., & Fostier, A. (2008). Oogenesis: post - vitellogenic events leading to a fertilizable oocyte . Pages 1 36 in M. J. Rocha, A. Arukwe, B. G. Kapoor, editors. Fish reproduction. Science Publishers, Ensfield, New Hampshire. Bro wn - Peterson, N.J. & Heins, D.C. (2009). Interspawning interval of wild female three - spined stickleback Gasterosteus aculeatus in Alaska. Journal of Fish Biology 74 :2299 - 2312 Bruch, R.M. & Binkowski, F.P. (2002). Spawning behavior of lake sturgeon ( Acipens er fulvescens ). Journal of Applied Ichthyology 18 :570 - 579 Burnham, K. P., & Anderson, D.R. (2001). Kullback - Leiber information as a basis for strong inference in ecological studies. Wildlife Research 28 :111 119. Clugston, J.P. (1996). Retention of T - bar anchor tags and passive integrated transponder tags by gulf sturgeons. North American Journal of Fisheries Management 16 :682 - 685 Craig, J.M., Papoulias, D.M., Thomas, M.V., Annis, M.L., and Boase, J. Sex assignment of lake sturgeon ( Acipenser fulvescens ) based on plasma sex hormone and vitellogenin levels. Journal of Applied Ichtyhology 25(2) :60 - 67 38 Dettlaff, T. A., Ginsburg, A. S. & Shmalhausen, O. I. (1993). Sturgeon Fishes Developmental Biology and Aquaculture . Berlin. Dickerson, B.R., Quinn, T.P., & W ilson, M.F. (2002). Body size, arrival date, and reproductive success of pink salmon, Oncorhynchus gorbuscha . Ethology, Ecology, and Evolution 14 :29 - 44 Dickerson, B.R., Brinck, K.W., Willson, M.R., Bentzen, P., & Quinn, T.P. (2005). Relative importance of salmon body size and arrival time at breeding grounds to reproductive success. Ecology 86(2) :347 - 352 Duong, Y.T. (2010). Effects of spawning behavior and environmental factors on adult reproductive ecology and larval dispersal of Lake Sturgeon ( Acipenser fulvescens ). Ph.D. dissertation, Department of Zoology Michigan State University, East Lansing, Mich., USA. Duong, T.Y., Scribner, K.T., Crossman, J.A, Forsythe, P.S., & Baker, E.A. (2011). Environmental and maternal effects on embryonic and larval devel opmental time until dispersal of lake sturgeon (Acipenser fulvescens). Canadian Journal of Fisheries and Aquatic Sciences 68 :643 - 654 Emlen, S.T. & Oring, L.W. (1977). Ecology, sexual selection, and the evolution of mating systems. Science 197 : 215 - 223 Fei st, G., Schreck, C.B., Fitzpatrick, M.S., & Redding, J.M. (1990). Sex steroid profiles of coho salmon (Oncorhynchus kisutch) during early development and sexual differentiation. General and Comparative Endocrinology 80 :299 - 313. Fitzpatrick, M.S., Van De r Kraak, G., & Schreck, C.B. (1986). Profiles of plasma sex steroids and gonadotropin in coho salmon, Oncorhynchus kisutch , during final maturation. General and Comparative Endocrinology 62 :437 451. Fitzpatrick, M. S., Redding, J. M., Ratti, F. D., & Sch eck, C. B. (1987). Plasma testosterone predicts the ovulatory response of coho salmon (Oncorhynchus kisutch) to gonadotropin - releasing hormone analog. Canadian Journal of Fisheries and Aquatic Science 44 :1351 1357. Forsythe, P.S., Crossman, J.A., Bello, N .M., Baker, E.A., & Scribner, K.T. (2012). Individual - based analyses reveal high repeatability in timing and location of reproduction in lake sturgeon ( Acipenser fulvescens ). Canadian Journal of Fisheries and Aquatic Science 68 :1 - 13 39 Fuzzen, M., Bernier, N .J., & Van Der Kraak, G. (2011). Stress and reproduction . In Hormones and reproduction in vertebrates. Vol. 1. Fishes. D.O. Norris & K.H. Lopez, editors. Elsevier, London. pp. 103 117 Hamel, M.J., Hammen, J.J., & Pegg, M.A. (2012). Tag retention of t - bar anchor tags and passive integrated transponder tags in shovelnose sturgeon. North American Journal of Fisheries Management 32(3) :533 - 538 Johnston, T.A., Kaufman, S.D., Moles, M.D., Weigand, M.D., & Leggett, W.C. (2008). Effects of delayed fertilization on embryo viability in walleye Sander vitreus (Mitchill): the role of maternal effects. Journal of Fish Biology 72 :2634 - 2644 Jones, O. R. & Wang, J. L. (2010). COLONY: a program for parentage and sibship inference from multilocus genotype data. Molecular Ec ology Resources 10 :551 555. Jones, A. G., Small, C. M., Paczolt, K. A., & Ratterman, N. L. (2010). A practical guide to methods of parentage analysis. Molecular Ecology Resources 10 :6 30 Jørgensen, C., Dunlop, E.S., Opdal, A.F., Fiksen, Ø. (2008). The ev olution of spawning migrations: state dependence and fishing - induced changes. Ecology 89(12) :3436 - 3448 Kalinowski, S. T., Taper, M. L., & Marshall, T. C. (2007). Revising how the computer program CERVUS accommodates genotyping error increases success in p aternity assignment. Molecular Ecology 16 : 1099 - 1106 Khoroshko, P.N. (1972). The amount of water in the Volga Basin and its effect on the reproduction of sturgeons (Aipenseridae) under conditions of normal and regulated discharge. Journal of Ichthyology 12 :608 - 616 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. Conservation Genetics 2(2) : 103 119. Kynard, B., Pugh, D ., Parker, T., & Kieffer, M. (2011). Using a semi - natural stream to produce young sturgeons for conservation stocking: maintaining natural selection during spawning and rearing. Journal of Applied Ichthyology 27 :420 - 424 Lowerre - Barbieri, S.K., Ganias, K., Saborido - Rey, F., Murua, H., & Hunter, J.R. (2011). Reproductive timing in marine fishes: variability, temporal scales, and methods. Marine and Coastal Fisheries: Dynamics, Management, and Ecosystem Science 3 :71 - 91 Lubzens, E., Young, G. Bobe, J., & Cerd á, J. (2010). Oogenesis in teleosts: how fish eggs are formed. General and Comparative Endocrinology 165 :367 - 389 40 McGuire, J.M., Scribner, K.T., & Congdon, J.D. (2013). Spatial aspects of movements, mating patterns, and nest distributions influence gene fl ow among population subunits Conservation Genetics 14 :1029 - 1042 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. Transactions of the American Fisheries Society 129(6) :1380 1388. McQuown, E., Gall, G.A.E., & May, B. (2002). Characterization and inheritance of six microsatellite loci in lake sturgeon. Transactions of the American Fisheries Society 131(2) :299 - 307 Mittelbach, G.G., Ballew, N.G., & Kjelvik, M.K. (2014). Fish behavioral types and their ecological consequences. Canadian Journal of Fisheries and Aquatic Science 71 :927 - 944 Morgan, M.J. (2008). Integrati ng reproductive biology into scientific advice for fisheries management. Journal of Northwest Atlantic Fisheries Science 41 :37 - 51 Morgan, M.J., Wright, P.J., & Rideout, R.M. (2013). Effect of age and temperature on spawning time in two gadoid species. Fis heries Research 138 :42 - 51 Neuheimer, A.B. & MacKenzie, B.R. (2014). Explaining life history variation in a Ecology 95(12) :3364 - 3375 Olin, M., Jutila, J., Lehtonene, H., Vinni, M., Rvi, J.R., Estlander, S., Rask, M., Kuparinen, A., & Lappalainen, J. (2012). Importance of maternal size on the reproductive success of perch, Perca fluviatilis , in small forest lakes: implications for fisheries management. Fisheries Management and Ecology 19 :363 - 374 Paragamian, V.L. & Kruse, G. (2001). Kootenai River white sturgeon spawning migration behavior and a predictive model. North American Journal of Fisheries Management 21 :10 - 21 Peterson, D.L., Vecsei, P., & Jennings, C.A. (2007) Ecology and biology of the lake sturgeon: a syn thesis of current knowledge of a threatened North American Acipenseridae. Reviews in Fish Biology and Fisheries 17 :59 - 76 Quinn, T.P. (1999). Variation in pacific salmon reproductive behaviour associated with species, sex and levels of competition. Behavio ur 136 :179 - 204 Schreck, C.B., Contreras - Sanchez, W., & Fitzpatrick, M.S. (2001). Effects of stress on fish reproduction, gamete quality, and progeny. Aquaculture 197 :3 - 24 41 Serbezov, D., Bernatchez, L., E.M. Olsen, & Vøllestad, 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. Molecular Ecology 19 :3193 - 3205 Shaw, S.L., Chipps, S.R., Windels, S.K., Webb, M.A.H., McLeod, D.T., & Willis, D.W. (2012). Lake sturgeon population attributes and reproductive structure in the Namakan Reservoir, Minnesota and Ontario. Journal of Applied Ichthyology 28 :168 - 175 Skaug, H., Fournier, D. & Nielsen, A. (2006). glmmADMB: generalized linear mixed models using AD Model Builder. R package version 0.3. Smith, K.M. & King, D.K. (2005). Dynamics and extent of larval lake sturgeon Acipenser fulvescens drift in the Upper Black River, Michigan. Journal of Applied Ichthyology 21(3) :161 168 Talbott, M.J., Van Eenennaam, J.P., Linares - Casenave, J., Doroshov, S.I., Guy, C.S., Struggenegger, P., & Webb, M.A.H. (2011). Investigating the use of plasma testosterone and estradiol - sturgeon, Acipenser transmontanus . Aquaculture 315 : 283 - 289 Thiem, J.D., Dumont, H.P., Van Der Kra ak, G., & Cooke, S.J. (2013). Biology of lake sturgeon ( Acipenser fulvescens ) spawning below a dam on the Richelieu River, Quebec: behaviour, egg deposition, and endocrinology. Canadian Journal of Zoology 91 :175 - 186 Thorstad, E.B., Økland, F., Aerestrup, K., & Heggberget, T.G. (2008). Factors affecting the within - river spawning migration of Atlantic salmon, with emphasis on human impacts. Reviews in Fish Biology and Fisheries 8 :345 - 371 Wade, M.J. & Arnold, S.J. (1980). The intensity of sexual selection in relation to male sexual behaviour, female choice and sperm precedence. Animal Behaviour 28 :446 - 461 Warren, M.A. & Morbey, Y.E. (2012). Migration timing of female kokanee salmon Oncorhynchus nerka: diel patterns and effects of maturation state. Journal of Fish Biology 81 :1234 - 1247 Webb, M. A. H., Van Eenennaam, J.P., Doroshov, S.I., & Moberg, G.P. (1999). Preliminary observations on the effects of holding temperature on reproductive performance of female white sturgeon, Acipenser transmontanus Richardson. Aquaculture 176 :315 329. Webb, M.A.H., Van Eenennaam, J.P., Feist, G.W., Linares - Casenave, J., Fitzpatrick, M.S., Schreck, C.B., & Doroshov, S.J. (2001). Effects of thermal regime on ovarian 42 maturation and plasma sex steroids in farmed white sturgeon, Ac ipenser transmontanus . Aquaculture 201 :137 - 151 Webb, M.A.H., Feist, G.W., Foster, E.P., Schreck, C.B., & Fitzpatrick, M.S. (2002). Potential classification of sex and stage of gonadal maturiry of wild white sturgeon using blood plasma indicators. Transact ions of the American Fisheries Society 131 :132 - 142 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 . Molecular Ecological Not es 3(1) :47 55 Wright, P.J. & Trippel, E.A. (2009). Fishery - induced demographic changes in the timing of spawning: consequences for reproductive success. Fish and Fisheries 10 :289 - 304 43 Chapter 2 GENETIC AND ENVIRONM ENTAL COM PONENTS OF PHENOTYPI C AND BEHAVIORAL TRAIT VAR IATION DURING LAKE S TURGEON EARLY ONTOGE NY ABSTRACT Adults can plastically respond to environmental conditions by selecting breeding and egg incubation locations that affect offspring traits during embryoni c and larval development. Environmental conditions during incubation can also affect traits during later ontogenetic stages (i.e. ontogenetic contingency). Using a well - studied population of lake sturgeon ( Acipenser fulvescens ) from Black Lake, Michigan, e ggs were reared from full and half - sibling families at two spawning locations and in a common garden experiment consisting of three water velocity treatments: high, low, and variable. Larvae reared within the stream varied significantly between spawning si tes for traits quantified at hatch including body length, body area, and yolk - sac area. Estimates of trait heritabilities ranged from 0.42 - 0.48. Growth from hatch to 2 - 3 weeks post hatch when larvae emerged from the substrate to begin exogenously feeding v aried among families reared at the spawning locations due to a genotype - by - environment (G - by - E) interaction. In the common garden experiment, phenotypic variation among families was greatest for larvae reared under high velocity and a significant G - by - E wa s detected for body length, body area, and head area. Growth also varied among velocity treatments, but did not vary between families. Overall, results indicate that adult - selected spawning and rearing locations as well as genetic effects influence offspri ng phenotypic trait variation. Importantly, egg incubation conditions can also affect trait variation during sequential ontogenetic stages potentially affecting larval survival and population levels of recruitment. 44 INTRODUCTION Changes in environmental con ditions are increasingly affecting phenotypic variation in natural populations worldwide (Kopp and Matuszewki 2014; Teplitsky et al. 2008). Predicting organismal responses to environmental change can be achieved by quantifying the relative contributions of genetic and environmental effects on phenotypic variation (Crespel et al. 2013). Recent studies have led to greater understanding of the extent by which phenotypic trait variation in wild populations can evolve or respond plastically to changing environme nts (Stoks et al. 2014; Urban et al. 2014; Merilä and Hendry 2014; Merilä 2012; Hoffmann and Srgò 2011). Levels of additive genetic variance indicate how traits can evolve in response to selection (Wilson et al. 2005), and therefore provide an understandin evolutionary potential (Garant et al. 2008). Additionally, understanding the role of environmental effects including genotype - by - environment (G - by - E) interactions (Byers 2008) is necessary to avoid attributing phe notypic changes to genetic effects instead of phenotypic plasticity, a common goal in quantitative genetics (Gienapp et al. 2008; Chevin et al. 2010; Teplitsky et al. 2008). Plasticity includes changes in behavior, phenotype, or physiology that occur with out genetic changes (West - Eberhard 2003), and is commonly observed in response to environmental changes. For example, fishes can respond to variability in environmental conditions such as temperature, water velocity, and photoperiod through behavioral adju stments in the timing and location of reproduction. Behavioral plasticity including adult selection of mating locations and egg - incubation conditions can influence the likelihood of fertilization success and probabilities of offspring survival and growth 45 ( Refsnider and Janzen 2010; Ohba et al. 2012; Brodin et al. 2006). Examples of behavioral plasticity in adult fishes include delaying spawning (Warren et al. 2012; environmental c onditions at traditionally used areas are unfavorable (Gunn and Sein 2000). Little empirical work has examined how environmental conditions associated with the timing and location of adult spawning may affect offspring trait variation during incubation and subsequent ontogenetic stages. Given the temporal and spatial complexity of stream systems (Allan 2004; Fausch et al. 2002), fishes are well - suited for studies examining the effects of adult - selected spawning locations on offspring trait variation during early ontogenetic stages. Abiotic conditions within spawning sites affect offspring traits associated with survival such as time to hatch, body size, and behavior (Gupta et al. 2013; Crozier et al. 2008; Garant et al. 2003; Leach and Houde 1999). For exam ple, cool temperatures during egg incubation result in offspring of larger body size which is associated with swimming ability, growth rates, predation, and survival (Kingsolver and Huey 2008; Kamler 2005; Fuiman et al. 2005). Conditions experienced durin g egg incubation can also affect phenotypic traits during later ontogenetic stages. This phenomenon is referred to as 1998), and has been observed in numerous taxa includin g fishes (e.g., Melanogrammus aeglefinus , Martell et al. 2005; Oncorhynchus tshawytscha , Heath et al. 1993). Therefore, studies examining how environmental conditions experienced during egg incubation affect individual phenotypes during sequential ontogene tic stages are 46 important for predicting population - levels of phenotypic variation and recruitment (Xu et al. 2010). Stream discharge is one environmental variable that affects phenotypic and behavioral trait variation in fishes (Langerhans 2008). Spring ru noff, rainfall, and hydroelectric dams can cause variability in river discharge (Jonsson and Jonsson 2009; Mion et al. 1998; Skoglund et al. 2011). Changes in discharge can affect adult spawning decisions and potentially affect growth and survival of offsp ring during early life stages. For example, fluctuations in discharge can delay spawning, dislodge fish eggs from substrate or nest surfaces, affect movements of sediments that decrease oxygen levels in interstitial spaces where fish eggs incubate, and inf luence egg incubation time thereby affecting larval body size at the time of hatch (Auer 1996; Paragamian and Kruse 2001; Smith and Marsden 2009; Allan 2004). However, few studies have examined how variation in discharge and velocity conditions experienced during egg incubation will affect phenotypic trait variation across sequential ontogenetic stages. In this study, the genetic (family) and environmental (in - situ stream discharge and experimental velocity regimes) effects on larval phenotypic variation in lake sturgeon ( Acipenser fulvescens ) were quantified. Lake sturgeon are long - lived, iteroparous, broadcast spawning fish that engage in polygamous mating (Bruch and Binkowski 2002). Spawning occurs at multiple times and sites characterized by different th ermal and discharge regimes from late April to early June (Forsythe et al. 2011). During spawning events, adults extrude gametes over gravel and rocky substrate and provide no post - ovulatory parental care as eggs from numerous families incubate until hatch (Bruch and Binkowski 2002). The age and size 47 of larvae at hatch and the timing of emergence when larvae begin to exogenously feed and disperse downstream are dependent on the environmental conditions within spawning and egg rearing sites (Duong et al. 201 1). Discharge is one environmental condition associated with different lake sturgeon spawning and egg incubation habitats that varies due to seasonal effects (principally snow melt and precipitation) and the presence of hydroelectric dams on most Great Lak es tributaries used by spawning lake sturgeon (Peterson et al. 2007). The objectives of this study were to: 1) quantify the relative contributions of genotype (family), rearing environment, and their interaction on larval lake sturgeon phenotypes, 2) exami ne how discharge within adult - selected rearing environments and climate - induced variability in velocity regimes influences phenotypic trait variation, and 3) determine whether environmental conditions experienced during egg incubation affect phenotypic and behavioral trait variation at a sequential ontogenetic stage (emergence). Knowledge of how discharge and water velocity experienced during egg incubation affect larval phenotypic traits during sequential ontogenetic stages will provide a greater understan ding of habitat needs during critical early life stages of fishes and how incubation conditions affect larval survival and recruitment. MATERIALS AND METHODS Study Site The study was conducted on the Upper Black River (UBR), the largest tributary of Black Lake (~3500 hectares) located in northern Michigan, USA (Fig. 1). A well - studied population of approximately 1200 adult lake sturgeon (Pledger et al. 2013) resides in Black Lake which has been isolated from the Great Lakes since 1903 (Baker 48 and Borgeson 1 999; Smith and King 2005b). Shallow spawning areas (~1 - 3 meters) within the river allow access to adults for gamete collection and experimental work within the stream (Forsythe et al. 2013). Additionally, a streamside rearing facility at the Kleber Dam (Fi g. 1) provides a laboratory for concurrent experimental work (e.g. Hastings et al. 2013). Spawning Locations and Gamete Collection The 2011 spawning season provided the opportunity to examine how behavioral plasticity in adult spawning site selection in response to discharge conditions affected larval phenotypic trait variation. Uncharacteristically high rainfall and stream discharge eber Dam located several kilometers upstream (Fig. 1). Spawning below the dam has been observed in the past (Anderson 1984; Baker and Borgeson 1999); however, in most years spawning occurs in downstream e 2011 spawning season, spawning adults were sampled using long - handled dipnets by wading the stream one or more times per day from May 3 rd to June 3 rd . Eggs and sperm were hand - stripped from spawning adults by applying pressure from the anterior to the po sterior portion of the abdomen above the urogenital opening. Sperm was collected from males in 20 - mL syringes and stored in sterile plastic bags on ice. Eggs and any collected ovarian fluid were stored in sterile plastic bags surrounded by river water for transport. All fertilizations were conducted within 12 - hours of gamete collection. Field Experiment 49 On May 5 th and 6 th , four - hundred eggs from each of six females were placed on circular porous filter pads (7.79 cm 2 ; 3M Worldwide Inc, Buffing and Polishin g Pads) and fertilized with sperm from one male per female for a total of six full - sibling crosses. Filter pads were secured to the base of open - sided square iron rebar boxes (30.48cm width by 30.48 cm height) with flat bases to allow water passage over eg gs (Forsythe et al. 2013). Rebar boxes were secured to a rebar base that could withstand high water velocities and placed in the stream. In total, twelve boxes (two per family) were used. pawning site (Kleber allowing eggs to incubate under stream discharge conditions specific to each spawning site. Eggs were checked daily and dead eggs were removed to prevent f ungal infection. during early May; however, a linear regression was used to determine the relationship - May to August using discharge data obtained from the dam operator (Tower - Kleber Limited Partnership). The linear regression (y = 0.3541x + 0.5508, R 2 = 0.948) was used to predict disch arge at the (CTU) to estimate hatch date using the methods of Kempinger (1988). All su rviving eggs from both sites were collected one day prior to hatch on May 13 th and transferred to the streamside facility to keep families separate, prevent post - hatch movement, and 50 determine the influence of discharge conditions associated with spawning s ite on larval phenotypic traits. Larvae were anesthetized with tricaine methanesulfonate (MS - 222; 25 mg/mL) and photographed at hatch (Fig. 2a) using digital photography. Image J analysis software (Version 1.34, free - ware) was used to quantify four pheno typic traits at hatch from the digital photographs: body length (BL; mm), body area (BA; mm 2 ), yolk - sac area (YSA; mm 2 ), and head area (HA; mm 2 ; Fig. 2a). A subset of larvae from each family and incubation site were randomly assigned to individual plastic incubation chambers (12.7cm by 6.35cm) containing gravel substrate to provide cover during the period of endogenous feeding on yolk - sac reserves. Mesh siding on the chambers allowed for continual water passage (~0.1m/sec). Chambers were monitored multiple times daily until larvae depleted yolk - sac reserves and emerged from the substrate to begin the onset of exogenous feeding. Timing (days) to emergence was recorded. At emergence, larvae (Fig. 2b) were anesthetized and digitally photographed again to quanti fy three traits at emergence: time from hatch to emergence (ET; days), body length at emergence (EBL; mm), and growth (TG, changes in body length from hatch to emergence; mm). Velocity (Common Garden) Experiment On May 15 th and 16 th , six - hundred eggs from each of four females were fertilized with sperm from two males each to produce eight half - sibling crosses. One hundred eggs from each half - sib cross were allowed to adhere to screens mounted on 8cm by 4cm plexi - glass plates. After approximately 30 minutes (ample time allowing for eggs to adhere to the mesh given the adhesive glycoprotein layer of sturgeon eggs, Doering et 51 al. 2012), plates were secured onto six 2.70m long by 0.10m wide flumes (2 per treatment) made of PVC tubing and reared under one of thre e discharge treatments at the streamside facility. Velocity treatments included: low (constant 0.20 m/sec), high (constant 0.80 m/sec), and variable velocity (twelve hours at 0.20 m/sec and twelve hours of 0.80 m/sec). Flume velocity was estimated using a Marsh - McBirney flow meter. Low and high velocity treatments represented the natural range in velocity conditions (0.15 - 1.40m/sec) when lake sturgeon spawn and eggs have been observed on river substrate (Kempinger 1988; Hay - Chmielewski and Whelan 1997). T he variable treatment simulated a temporally - variable environment associated with diel changes in water velocity associated with flood events or anthropogenic disturbance associated with hydroelectric operations. Families were replicated among treatments and egg placement was randomized within each flume. CTU was estimated based on water temperature in the streamside facility using a YSI 5200 Recirculating System Monitor (Xylem, Inc.) to predict hatch date. Eggs were removed from flumes one day prior to hatch to keep families separate and so that phenotypes at hatch reflected velocity conditions experienced during incubation. The four hatch traits (body length, body area, yolk - sac area, and head area) and three emergence traits (timing of emergence, emergence body length, and growth) were quantified using the same methods as described in the field experiment. All research was conducted under animal use and care procedures approved by the Michigan State University Institutional Animal Care and Use Committee. Statistical Analysis 52 In the field experiment, linear mixed - effect models were used to test for family and spawning site effects and for G - by - E (i.e. family - by - site) interactions associated with larval phenotypic traits measured at hatch and emergence. All traits were analyzed and Bolker 2011) in the program, R version 2.13 (R Development Core Team). All models were fit using restricted maximum likelihood estimation (REML) usi ng a step - wise approach for model selection. In the field experiment, each trait was fit to the full model: [Eq 1.] Trait(y ijk i + F j + (FS) ij ijk where F ~ N(0, ), FS ~ N(0, ) , and N(0, ) with being the population mean, S i was the fixed effect for site, F j was a random effect for family, (FS) ij was the family - by - site interaction term treated as a ijk was the random residual error. The d ata was then fit to a model that excluded the interaction term: [Eq 2.] Trait(y ij i + F j ij where and . A likelihood ratio test was used for model comparison. Next, a likelihood ratio test was used to co mpare equation [2] to a simpler model: [Eq 3.] Trait(y ij i i to test for family effects. For each trait where equation [2] was the model of best fit, a point - estimate of the narrow - sense heritability (h 2 ) was computed. No confidence intervals co uld be estimated (see derivation in Online Resource 1). Lastly, a likelihood ratio test was used to test for effects of rearing site. The same analysis and model selection approach was used to analyze the velocity (common garden) experiment with treatment (low, variable, high velocity) as the fixed effect in all models instead of site. 53 R ESULTS Field Experiment characterized by different discharge conditions (Online Resource 2) for nine days and were collected at approximately 60 CTUs. Individuals began to hatch the following day when CTU was approximately 68 which is within the traditional range of CTUs at the time of hatch (58.1 to 71.4) estimated in previous lake sturgeon s tudies on the UBR (Smith and King 2005a). Twenty - eight individuals from each of the six families from both Families showed significant variation in three of the trai ts measured at hatch. Mean body length (BL) ranged from 11.84 (±1.24) to 13.41 (±0.85) mm while body area (BA) ranged from 22.90 (±3.73) to 28.98 (±4.73) mm 2 . Yolk - sac area (YSA) also varied between families and ranged from 6.73 (±1.08) to 8.20 (±1.08) mm. Phenotypic variance in all three traits varied among families with the percent of phenotypic variance explained by genetic differences (h 2 ) estimated as 0.48, 0.42, and 0.47 for BL, BA, and YSA, respectively (Table 1; Online Resource 1). The fourth hatch trait, mean head area (HA), showed little variation ranging from 0.98 (±0.15) to 1.20 (±0.16) mm 2 and no differences were detected between families. Additionally, no family - by - site (G - by - E) interaction or site effects were detected for any of the four trai ts measured at hatch (p>0.05; Online Resources 3 and 4). T wenty - four individuals from each family (N=12/site) were placed in individual incubation chambers and monitored until emergence. The number of larvae that emerged from the substrate and were photog raphed ranged from 15 to 22 individuals 54 per family. Variation in the timing of larval emergence (ET) was low, as most individuals emerged from substrate in 13 to 17 days. Larvae across both rearing sites varied in body size at emergence (EBL) from 17.48 ( ± 1.02) to 25.12 ( ± 1.36) mm; however, no significant effects were detected for ET or EBL across both sites (p>0.05; Online Resources 3 and 4). A significant family - by - site interaction for larval growth from hatch to emergence (TG) was detected (Table 1) as f amilies ranged from 10.47 ( ± 1.57) to 11.87 ( ± ± 2.28) to 10.89 ( ± - by - site interaction explained approximately 33% of the p henotypic variation observed in TG while family explained approximately 18%. Differences in TG among families and between sites are visualized in a reaction norm plot (Fig. 3). Velocity (Common Garden) Experiment Eggs were incubated under treatments simula ting different stream velocities likely experienced during the egg incubation period for seven days and were removed when CTU was approximately 58. Larvae began to hatch the following day when CTU was approximately 70 (within the traditional range of 58.1 to 71.4 CTUs at time of hatch, Smith and King 2005a). The number of larvae from each family that survived to hatch across all three treatments ranged from 13 to 46 individuals with the lowest number of hatched individuals from the High velocity treatment. A significant family - by - treatment effect was detected for BL, BA, and HA (Table 1) with the family - by - treatment interaction explaining approximately 13 - 18% of the phenotypic variation observed among the three traits. As indicated by the reaction norm (Fig. 4), BL was most variable in the High velocity treatment with a range of 10.47 to 13.67 (±0.64) mm versus the Low 55 velocity treatment where BL ranged from 11.47 to 12.52 (±0.68) mm. Body area (BA) and head area (HA) also showed variation across families and treatments ranging from 22.18 to 31.24 (±2.95) mm 2 and 0.98 to 1.21 (±0.14) mm 2 , respectively. Yolk - sac area (YSA) showed little variation across families and treatments ranging from 7.34 (±0.90) to 8.73 (±0.66) mm 2 and no significant effects were detecte d (Online Resources 3 and 4). The number of larvae that survived and emerged from the incubation chambers ranged from 12 to 37 individuals per family across the three treatments. There was little variance among treatments in ET which ranged from 13.75 to 14.25 (±1.29) days and no significant differences among families or treatments in EBL which ranged from 22.50 to 25.05 (±1.47) mm. Larvae varied in growth from hatch to emergence (TG) ranging from 10.49 to 14.33 (±2.09) mm across the three treatments. TG s ignificantly varied among treatments (p=0.040, p=0.012), but did not vary among families (Online Resource 4). D ISCUSSION In this study, the concurrent use of a natural stream and a common garden experiment allowed for determination of how potential discha rge and velocity regimes experienced during egg incubation and genetic (family) effects influenced larval phenotypic traits at hatch and emergence. In fishes, early life stages are the critical period in which mortality is highest (~99%) and growth and sur vival is affected by several abiotic and biotic factors (Chambers and Trippel 1997). Understanding factors that affect trait variation during this period is vital to predict population levels of recruitment. Environmental conditions experienced during egg incubation are known to 56 affect offspring traits associated with survival in numerous fishes. However, the lack of knowledge of how conditions experienced during egg incubation affect larval fish phenotypic traits during sequential ontogenetic stages is sur prising given the high variability of fluvial systems (Schumm 1988), particularly during and immediately following the spring spawning season. Results from the study showed that discharge (stream) and velocity (experimental) regimes experienced during egg incubation can influence offspring traits at later ontogenetic stages. In the field experiment, plasticity in simulated adult spawning site selection based on discharge had no effect on traits measured at hatch; however, differences in growth (TG) attribu ted to family - by - site (G - by - E) interactions were observed. The presence of G - by - E interactions indicate that certain genotypes are more sensitive to environmental influences, thereby resulting in a higher contribution of environmental effects on the observ ed phenotypic trait variation and less of a genetic effect (Falconer and Mackay 1996). Furthermore, differences in total growth among individuals due to egg rearing environment provides evidence of ontogenetic contingency. Our results emphasize the need to examine traits beyond the endogenous feeding stage in order to understand how rearing conditions will affect offspring traits that are likely associated with survival and population levels of recruitment (Burt et al. 2011). Stream velocity is one environm ental factor that varies spatially and temporally during the lake sturgeon spawning season (Forsythe et al. 2011) and was experimentally manipulated in the streamside rearing facility. Results revealed that there were significant differences in TG from hat ch to emergence between individuals 57 associated with the velocity regime experienced during egg incubation providing additional evidence of ontogenetic contingency. Lake sturgeon eggs have been observed in stream beds with velocities ranging from 0.15 to 1. 40 m/sec; however, populations in Michigan and Wisconsin typically experience stream velocities closer to 0.30m/sec (Kempinger 1988; Hay - Ch mielewski and Whelan 1997). Therefore, differences in growth occurred among individuals exposed to different velocity treatments representing a large range over which lake sturgeon spawn. No differences were observed in yolk - sac area (YSA) between lake sturgeon larvae reared under the different velocity treatments. Yolk is considered the primary source of energy as larv ae endogenously feed and is an important determinant of growth prior to exogenous feeding. However, growth can also be influenced by environmental variables as well maternal (e.g. egg size) and genetic effects (Kamler 2008). Differences observed in TG at t he timing of emergence may be attributed to genetic or maternal effects, explaining why no differences were observed in YSA at the time of hatch. Alternatively, differences in yolk - sac utilization based on larval body size may explain why TG differed in th e field experiment. For example, Hardy and Litvak (2004) found that larger shortnose sturgeon ( Acipenser brevirostrum ) larvae with higher metabolic demands may have experienced declines in yolk - sac utilization efficiency resulting in slower growth rates th an smaller Atlantic sturgeon ( Acipenser oxyrhynchus ) reared in the same environment. Discharge and velocity are two related environmental variables in riverine systems (Sloat and Hull 2004). Although different families were used, the lack of concordance between the field (discharge) and velocity experimental results suggest 58 that discharge regimes may not have been as variable between spawning sites as expected. Alternatively, discharge may not have been the only environmental condition affecting trait va riation at either spawning site. Thermal regimes are also known to vary across lake sturgeon spawning locations (Forsythe et al. 2011); however, data indicates that thermal regimes at the alternative and traditional spawning areas are highly correlated dur ing the lake sturgeon spawning season (data not shown). Further studies examining additional environmental conditions within spawning locations would be beneficial to understand how multi - variable incubation environments affect larval phenotypic variation across ontogenetic stages. In the field experiment, a significant effect of family on body length (BL), body area (BA), and yolk - sac area (YSA) at hatch was observed and approximately half of the inter - individual variation was attributed to genetic effects (Table 1). Heritability estimates (0.42 - 0.48) for the three traits in the study represented the upper - limit of a narrow - sense heritability estimate (similar to a broad - sense heritability estimate) given the inability to separate out additive genetic varia nce from other genetic effects. However, the estimates provide insight on the ability of the population to respond genetically (i.e. change in gene frequency) to changes in environmental conditions if trait expression is associated with differential surviv al. Therefore, differences in larval phenotypic traits at hatch such as larval body size may indicate a potential for differential survival between genotypes given that body size is associated with survival in many taxa (Fischer et al. 2011; DuRant et al. 2010; Einum and Fleming 1999). Significant differences in larval survivorship among families can lead to particular adult genotypes contributing more offspring to a given year class. For lake sturgeon with low 59 annual recruitment and long generation times, differential survival between members of different families can reduce cohort effective breeding population size and genetic diversity within the Black Lake population. Similar results have been observed in captive barramundi ( Lates calcarifer ) where diffe rential survival between larvae of different families led to reductions in effective population size and skewed reproductive contributions of families to the year class (Frost et al. 2006). Studies examining whether larval body size is associated with diff erential survival in age - 0 larval lake sturgeon would be beneficial. Spring flooding events and hydroelectric dam operations are two known causes of variability in discharge conditions and velocity rates in riverine systems (Jensen and Johnsen 1999; Lierm ann et al. 2012; Lytle and Poff 2004), and can alter larval rearing habitats and affect population levels of recruitment (Paragamian and Wakkinen 2011). Results from the velocity experiment indicate significant differences in body length (BL), body area (B A), and head area (HA) at the time of hatch due to family - by - treatment (G - by - E) interactions. Importantly, the largest variation in body size was observed among larvae of different families reared in the high velocity treatment (0.8m/s, Fig. 4) representin g the treatment of extreme environmental deviance. The large range in larval phenotypic variation associated with the high velocity treatment compared to the fluctuating variable velocity treatment may have resulted from the expression of cryptic genetic v ariation. Therefore, the high velocity experiment may have represented a stressful environment for lake sturgeon larvae, particularly on larvae from certain families (Fig. 4), given that cryptic genetic variation is often revealed in novel or stressful env ironments (Wund 2012; Charmantier and Garant 2005; McGuigan and Sgrò 2009). 60 Cryptic genetic variation is also revealed due to the presence of a G - by - E interaction (Rouzic and Carlborg 2007). Variation can be beneficial if genotypes produce phenotypes favo red under current environmental conditions (Candolin 2009). However, phenotypic variants often have lower survival and can be quickly eliminated by natural selection (Ghalambor et al. 2007). The presence of non - zero heritable traits in larval lake sturgeon and intra - and inter - annual variation in environmental conditions suggests different genotypes (e.g., offspring from different females) may be favored within and among years and would be represented in mixed - aged spawning groups in future years. However, observations from the study were only taken post - hatch up to the timing of emergence when mortality is highest and it is not known whether size differentials are maintained over longer periods. Further studies should examine the persistence of environmenta l effects to later ontogenetic stages and determine whether differences in phenotypic variation among families will lead to differential survival during early ontogeny. Changes in the distributions of phenotypic traits within populations have become a com mon occurrence as a response to environmental change; however the causes for these shifts are difficult to identify (Ozgul et al. 2010). Inter - individual variation in the degree of behavioral plasticity may be one explanation for within - population trait ch anges ( Dingemanse and Wolf 2013). Differences in growth between lake sturgeon use of different spawning locations can affect offspring size at later ontogenetic sta ges. Additionally, the timing of spawning within a season may affect offspring phenotypes as indicated by the differences in total growth observed between lake sturgeon larvae 61 reared in the different velocity regimes representing the range in velocity rate s during the entire spawning season. However, observations from the study did not include additional environmental variables such as temperature, substrate, and depth which vary among spawning locations during the spawning season. Further empirical work ad dressing the consequences of adult behavioral plasticity in response to annual and climate - induced variability in environmental conditions would be beneficial (Dingemanse and Wolf 2013), particularly in long - lived species where adaptive evolution cannot oc cur at the same pace as environmental changes (Refsnider and Janzen 2012). Knowledge within these research areas is essential to understand factors affecting larval survival and potential population levels of genetic diversity and recruitment. 62 APPENDIX 63 Table 2.1 Variance components and standard errors (SE) estimated from the model of best fit for the field and velocity experiments. P - values are presented from the likelihood ratio tests an d heritability (h 2 ) estimates are provided for traits where only a significant family effect was detected. Phenotypic Traits Component Var SE P - value (a) Field Experiment Body Length Family 0.30 0.03 0.008 Residual 0.94 0.05 h 2 0.48 -- Body A rea Family 4.62 0.12 0.028 Residual 17.05 0.23 h 2 0.42 -- Yolk - sac Area Family 0.27 0.03 0.006 Residual 0.87 0.05 h 2 0.47 -- Total Growth Family*Site 1.73 0.12 <0.001 Family 0.94 0.09 Residual 2.63 0.15 (b) Velocity Experiment Bod y Length Family*Treatment 0.13 0.02 0.003 Family 0.06 0.02 Residual 0.52 0.05 Body Area Family*Treatment 1.74 0.08 0.003 Family 0.78 0.06 Residual 8.68 0.19 Head Area Family*Treatment <0.01 <0.01 0.019 Family <0.01 <0.01 Residual 0.02 <0 .01 *Indicates an interaction between the components. 64 Table 2 . 2 Daily mean, minimum, and maximum discharge levels (meters 3 ment. Date Site Mean (m 3 /s) Min (m 3 /s) Max (m 3 /s) 5 - May Traditional Alternative 25.95 4.99 10.92 3.01 36.46 7.55 6 - May Traditional Alternative 23.40 4.66 10.38 2.89 34.48 6.07 7 - May Traditional Alternative 25.66 5.11 9.89 2.81 49.62 9.41 8 - May Traditi onal Alternative 26.33 5.07 10.04 2.82 33.18 5.90 9 - May Traditional Alternative 18.74 4.12 11.32 3.14 31.35 5.77 10 - May Traditional Alternative 20.30 4.34 10.53 2.97 31.28 5.82 11 - May Traditional Alternative 18.60 4.04 9.28 2.85 32.00 5.78 12 - May Tradi tional Alternative 19.79 4.27 12.68 3.26 28.76 5.38 13 - May Traditional Alternative 31.74 5.75 15.82 3.80 39.76 6.59 65 Table 2 . 3 Pairwise model comparisons for traits fit to a model including the interaction term and a reduced model w ithout the interaction term. Log - likelihood values, chi - square values, p - values (P), and degrees of freedom (df) obtained from the likelihood ratio tests are presented where applicable. Phenotypic Traits Model Comparison Log - likelihood Chi - square P df (a) Field Experiment Body Length Family + Site + Family*Site - 474.21 Family + Site - 474.21 <0.001 0.987 1 Body Area Family + Site + Family*Site - 960.48 Family + Site - 960.57 0.165 0.685 1 Yolk - sac Area Family + Site + Family*Site - 461.76 Family + Site - 461.76 <0.001 1.000 1 Head Area Family + Site + Family*Site - 285.29 Family + Site - 290.79 0.322 0.571 1 Emergence Time Family + Site + Family*Site - 220.97 Family + Site - 221.05 0.167 0.683 1 Emergence Body Length Family + Site + Family*Site - 211.05 Family + Site - 211.05 <0.001 1.000 1 Total Growth Family + Site + Family*Site - 227.71 Family + Site - 235.32 15.227 <0.001* 1 (b) Velocity Experiment Body Length Family + Site + Family*Site - 290.23 Family + Site - 294.68 8.891 0.003* 1 Body Area Family + Site + Family*Site - 642.65 Family + Site - 647.03 8.758 0.003* 1 Yolk - sac Area Family + Site + Family*Site - 315.11 Family + Site - 315.28 0.348 0.555 1 Head Area Family + Site + Family*Site 151.04 Family + Site 148.29 5.496 0.019* 1 Emergence Time Family + Site + Family*Site - 357.63 Family + Site - 357.63 <0.001 1.000 1 Emergence Body Length Family + Site + Family*Site - 451.39 Family + Site - 451.39 <0.001 1.000 1 Total Growth Family + Si te + Family*Site - 448.46 Family + Site - 448.46 <0.001 0.998 1 * Indicates a statistically significant difference (p<0.05) between models. 66 Table 2 . 4 P - values computed from the likelihood ratio tests to determine the significance of family and site (or treatment) for all traits from the field and velocity experiments. Values are not presented for models where a significant genotype - by - environment interaction was detected. Phenotypic Traits Family Site/Treatment (a) Field Experiment Body Leng th 0.008* 0.236 Body Area 0.028* 0.710 Yolk - sac Area 0.006* 0.556 Head Area 0.102 0.600 Timing of Emergence 0.456 0.076 Emergence Body Length 0.334 0.360 Total Growth -- -- (b) Velocity Experiment Body Length -- -- Body Area -- -- Yolk - sac Area 0.254 0.254; 0.352 Head Area -- -- Timing of Emergence 0.524 0.704; 0.238 Emergence Body Length 0.208 0.064; 0.28 Total Growth 0.438 0.040*; 0.012* *Indicates the component was statistically significant (p<0.05) and should not be removed from the mod el. 67 Figure 2.1 The study location on the Upper Black River, the largest tributary of Black Fig ure 2 .2 Four larval hatch traits A: body length (BL), B: body area (BA), C: yolk - sac area (YSA), D: head area (HA) (a) and one emergence trait, E: emergence body length(EBL), quantified in the experiment (b) . The two additional emergence traits, time to emerge nce (ET) and total growth (TG), were quantified as the time (days) and growth (mm) from hatch to emergence, respectively (b) . Scale bars (mm) are presented in the lower right corner 68 Fig ure 2. 3 Total growth of lake sturgeon larvae from hatch to emergenc e when eggs genotype - by - environment (G - by - E) interaction was detected providing evidence that conditions experienced during embryogenesis affect trait variation at a la ter ontogenetic stage (timing of emergence) Fig ure 2. 4 Reaction norm for body length at hatch for lake sturgeon larvae in the velocity (common garden) experiment. A significant genotype - by - environment (G - by - E) interaction was detected, and most of the variation was observed from larvae reared in 69 Heritability Derivation For each trait where equation [2] was the model of best fit in the paper, a point - estimate of the narrow - sense heritability (h 2 ) was computed using the fol lowing logic. Eq. [4] where var( is the additive variance and var( is the residual variance. Noti ce that the animal effect in this animal model can be decomposed to the average of Eq. [5] where we expect , while (Lynch and Walsh 1998). Going back to equation [2] in the paper, the family effect is just the average of the two parents and the residual variance contains the environmental residual of the animal mod el plus the Mendelian residual: Eq. [6] and Consequently, and then the narrow - sense heritability can be expresse d as: Eq. [7] Equation [7] represents the computational formula used when estimating heritability using variance components from equation 2 in the paper and does not al low for the estimation of confidence interv als for heritability estimates. 70 LITERATURE CITED 71 LITERATURE CITED Allan JD (2004) Landscapes and riverscapes: the influence of land use on stream ecosystems. Annu Rev Ecol Evol Syst 35:257 - 284 Anderson ER (1984) Artificial propagation of Lake sturgeon Acipenser fulvescens (Rafinesque), under hatchery conditions in Michigan. Michigan Department of Natural Resources Fisheries Division Research Report 1898, Ann Arbor . Auer NA (1996) Response of spawning lake sturgeons to change in hydroelectric facility operation. Trans Amer Fish Soc 125(1):66 - 77 Baker EA, Borgeson DJ (1999) Lake sturgeon abundance and harvest in Black Lake, Michigan, 1975 - 1999. N Am J Fish Man 19:1 080 - 1088 Bates D, Maechler M, Bolker B (2012) lme4: Linear mixed - effects models using S4 classes. Brodin T, Johansson F, Bergsten J (2006) Predator related oviposition site selection of aquatic beetles ( Hydroporus spp.) and effects on offspring life - hist ory. Freshw Biol 51:1277 - 1285 Bruch RM, Binkowski FP (2002) Spawning behavior of lake sturgeon ( Acipenser fulvescens ). J Appl Ichtyol 18:570 - 579 Burt JM, Hinch SG, Patterson DA (2011) The importance of parentage in assessing temperature effects on fish e arly life history: a review of the experimental literature. Rev Fish Biol Fisheries 21:377 - 406 Byers D (2008) Components of phenotypic variance. Nature Ed 1(1):161 Candolin U (2009) Population responses to anthropogenic disturbance: lessons from three - sp ined sticklebacks Gasterosteus aculeatus in eutrophic habitats. J Fish Biol 75:2108 - 2121 Chambers RC, Trippel EA (eds) (1997) Early life history and recruitment in fish populations. Chapman and Hall, London Charmantier A, Garant D (2005) Environmental qu ality and evolutionary potential: lessons from wild populations. Proc R Soc B 272:1415 - 1425 Chevin LM, Lande R, Mace GM (2010) Adaptation, plasticity, and extinction in a changing environment: towards a predictive theory. PLoS Biol 8(4):e1000357 72 Crespel A, Bernatchez L, Audet C, Garant D (2013) Strain specific genotype - environment interactions and evolutionary potential for body mass in Brook charr (Salvelinus fontinalis). G3: Genes, Genomes, Genetics 3(3):379 - 386 Crozier LG, Hendry AP, Lawson PW, Quinn TP, Mantua NJ, Battin J, Shaw RG, Huey RB (2008). Potential responses to climate change in organisms with complex life histories: evolution and plasticity in Pacific salmon. Evol Appl 1(2):252 - 270 Diggle PK (1994) The expression of andromonoecy in Solanum hirtum (Solanaceae): Phenotypic plasticity and ontogenetic contingency. Am J Bot 81(10):1354 - 1365 Dingemanse NJ, Wolf M (2013) Between - individual differences in behavioural plasticity within populations: causes and consequences. Anim Behav 85:1031 - 1039 Doering JA, Wiseman S, Beitel SC, Tendler BJ, Giesy JP, Hecker M (2012) Tissue specificity of aryl hydrocarbon receptor (AhR) mediated responses and relative sensitivity of white sturgeon ( Acipenser transmontanus ) to an AhR agonist. Aquat Toxicol 114 - 115:1 25 - 133 Duong TY, Scribner KT, Crossman JA, Forsythe PS, Baker EA (2011) Environmental and maternal effects on embryonic and larval developmental time until dispersal of lake sturgeon ( Acipenser fulvescens ). Can J Fish Aquat Sci 68:643 - 654 Durant SE, Hepp GR, Moore IT, Hopkins BC, Hopkins WA (2010) Slight differences in incubation temperature affect early growth and stress endocrinology of wood duck ( Aix sponsa ) ducklings. J Exp Biol 213:45 - 51 Einum S, Fleming I (1999) Maternal effects of egg size in brow n trout (Salmo trutta): norms of reaction to environmental quality. Proc Roy Soc London Ser B: Biol Sci 266(1433):2095 - 2100 Falconer DS, Mackay TFC (1996) Introduction to quantitative genetics (4 th ed). Longmans Green, Harlow, Essex. Fausch KD, Torgerse n CE, Baxter CV, Li HW (2002) Landscapes to riverscapes: bridging the gap between research and conservation of stream fishes. BioSci 52(6):483 - 498 Fischer B, Taborsky B, Kokko H (2011) How to balance the offspring quality quantity tradeoff when environme ntal cues are unreliable. Oikos 120:258 - 270 Forsythe PS, Crossman JA, Bello NM, Baker EA, Scribner KT (2011) Individual - based analyses reveal high repeatability in timing and location of reproduction in lake sturgeon ( Acipenser fulvescens ). Can J Fish Aqu at Sci 69(1):60 - 72 73 Forsythe PS, Scribner KT, Crossman JA, Ragavendran A, Baker EA (2013) Experimental assessment of the magnitude and sources of lake sturgeon egg mortality. Trans Am Fish Soc 142(4):1005 - 1011 Frost LA, Evans BS, Jerry DR (2006) Loss of g enetic diversity due to hatchery culture practices in barramundi ( Lates calcarifer ). Aquac 261:1056 - 1064 in fish larvae: variation with growth rate and the batch effect. Can J Fish Aquat Sci 62:1337 - 1349 Garant D, Dodson JJ, Bernatchez L (2003) Differential reproductive success and heritability of alternative reproductive tactics in wild Atlantic salmon ( Salmo salar L. ). Evol 57(5):1133 - 1141 Garant D, Hadfield JD, Kruuk LEB, Sheldon BC (2008) Stability of genetic variance and covariance for reproductive characters in the face of climate change in a wild bird population. Mol Ecol 17:179 - 188 Ghalambor CK, McKay JK, Carroll SP, Reznick DN (2007) Adaptive versus non - adaptive phe notypic plasticity and the potential for contemporary adaptation in new environments. Funct Ecol 21:394 - 407 Gienapp P, Teplitsky C, Alho JS, Mills JA, Merilä J (2008) Climate change and evolution: disentangling environmental and genetic responses. Mol Eco l 17:167 - 178 Gunn JM, Sein R (2000) Effects of forestry roads on reproductive habitat and exploitation of lake trout ( Salvelinus namaycush ) in three experimental lakes. Can J Fish Aquat Sci 57(Suppl.2):97 - 104 Gupta M, Pandey NN, Tiwari VK, Singh R, Kumar P, Halder RS, Ali S, Kumar S, Pandey A. (2013) Effects of different temperatures on incubation period, hatching rate, survival and growth of Labeo dyocheilus larvae. J Ecophysiol Occup Hlth 13(1,2):59 - 62 Hardy RS, Litvak MK (2004) Effects of temperature on the early development, growth, and survival of shortnose sturgeon, Acipenser brevirostrum , and Atlantic sturgeon, Acipenser oxyrhynchus , yolk - sac larvae. Environ Biol Fishes 70:145 - 154 Hastings RP, Bauman JM, Baker EA, Scribner KT (2013) Post - hatch dis persal of lake sturgeon ( Acipenser fulvescens , Rafinesque, 1817) yolk - sac larvae in relation to substrate in an artificial stream. J Appl Ichthyol 29(6):1208 - 1213 Hay - Chmielewski E, Whelan GE (1997) Lake sturgeon rehabilitation strategy. Michigan Departme nt of Natural Resources, Fisheries Division Special Report 18, Ann Arbor. 74 Heath DD, Fox CW, Heath JW (1999) Maternal effects on offspring size: variation through early development of chinook salmon. Evol 53(5):1605 - 1611 Hoffman AA, Sgro CM (2011) Climate change and evolutionary adaptation. Nat 470:479 - 485 Jensen AJ, Johnsen BO (1999) The functional relationship between peak spring floods and survival and growth of juvenile Atlantic salmon ( Salmo salar ) and Brown trout ( Salmo trutta ). Func Ecol 13:778 - 785 Jonsson B, Jonsson N (2009) A review of the likely effects of climate change on anadromous Atlantic salmon Salmo salar and brown trout Salmo trutta , with particular reference to water temperature and flow. J Fish Biol 75:2381 - 2447 Kamler E (2005) Parent egg progeny relationships in teleost fishes: an energetics perspective. Rev Fish Biol Fisheries 15:399 - 421 Kamler E (2008) Resource allocation in yolk - feeding fish. Rev Fish Biol Fisheries 18:143 - 200 Kempinger JJ (1988) Spawning and early life history o f Lake sturgeon in the Lake Winnebago system, Wisconsin. Am Fish Soc Symp 5:110 - 122 Kingsolver JG, Huey RB (2008) Size, temperature, and fitness: three rules. Evol Ecol Res 10:251 - 268 Kopp M, Matuszewski S (2014) Rapid evolution of quantitative traits: t heoretical perspectives. Evol Appl 7:169 - 191 Kruuk LEB, Slate J, Wilson AJ (2008) New answers for old questions: the evolutionary quantitative genetics of wild animal populations. Annu Rev Ecol Evol Syst 39:525 - 548 Langerhans RB (2008) Predictability of phenotypic differentiation across flow regimes in fishes. Integr Comp Biol 48(6):750 - 768 Leach SD, Houde ED (1999) Effects of environmental factors on survival, growth, and production of American shad larvae. Jour Fish Biol 54:767 - 786 Liermann CR, Nilsso n C, Robertson J, Ng RY (2012) Implications of dam obstruction for global freshwater fish diversity. BioSci 62(6):539 - 548 Lynch M, Walsh B (1998) Genetics and analysis of quantitative traits. Sinauer Associates, Sunderland. Lytle DA, Poff NL (2004) Adapt ation to natural flow regimes. TREE 19(2):94 - 100 75 Martell DJ, Kieffer JD, Trippel EA (2005) Effects of temperature during early life history on embryonic and larval development and growth in haddock. Jour Fish Biol 66:1558 - 1575 McGuigan K, Sgrò CM (2009) Evolutionary consequences of cryptic genetic variation. TREE 24(6):306 - 311 Merilä J (2012) Evolution in response to climate change: in pursuit of the missing evidence. Bio Essays 34:811 - 818 Merilä J, Hendry A (2014) Climate change, adaptation, and phenot ypic plasticity: the problem and the evidence. Evol Appl 7:1 - 14 Mion JB, Stein RA, Marschall EA (1998) River discharge drives survival of larval walleye. Ecol Appl 8(1):88 - 103 Ohba S, Ohtsuka M, Sunahara T, Sonoda Y, Kawashima E, Takagi M (2012) Differen tial responses to predator cues between two mosquito species breeding in different habitats. Ecol Ent 37:410 - 418 Ozgul A, Childs DZ, Oli MK, Armitage KB, Blumstein DT, Olson LE, Tuljapurkar S, Coulson T (2010) Coupled dynamics of body mass and population growth in response to environmental change. Nature 466(7305):482 - 485 Paragamian VL, Kruse G (2001) Kootenai river white sturgeon spawning migration behavior and a predictive model. N Am Jour Fish Man 21:10 - 21 Paragamian VL, Wakkinen VD (2011) White sturg eon spawning and discharge augmentation. Fish Man Ecol 18:314 - 321 Peterson DL, Vecsei P, Jennings CA (2007) Ecology and biology of the lake sturgeon: a synthesis of current knowledge of a threatened North American Acipenseridae . Rev Fish Biol Fisheries 17 :59 - 76 Pigliucci M (1998) Development phenotypic plasticity: where internal programming meets the external environment. Curr Opin Plant Biol 1(1):87 - 91 Pledger S, Baker E, Scribner K (2013) Breeding return times and abundance in capture - recapture models. Biometrics 69(4):991 - 1001 Reed TE, Wanless S, Harris MP, Frederiksen M, Kruuk LEB, Cunningham EJA (2006) Responding to environmental change: plastic responses vary little in a synchronous breeder. Proc R Soc B 273:2713 - 2719 76 Refsnider JM, Janzen FJ (201 0) Putting eggs in one basket: ecological and evolutionary hypotheses for variation in oviposition - site choice. Annu Rev Ecol Evol Syst 41:39 - 57 Refsnider JM, Janzen FJ (2012) Behavioural plasticity may compensate for climate change in a long - lived reptil e with temperature - dependent sex determination. Biol Cons 152:90 - 95 Rouzic AL, Carlborg O (2007) Evolutionary potential of hidden genetic variation. TREE 23(1):33 - 37 Schumm, SA (1988) Variability of the fluvial system in space and time. In: Rosswall T, W oodmansee RG, Risser PG (eds) Scales and global change: spatial and temporal variability in biospheric and geospheric processes. John Wiley, Chichester, pp 225 250 Skoglund H, Einum S, Robertsen G (2011) Competitive interactions shape offspring performanc e in relation to seasonal timing of emergence in Atlantic salmon. Jour Anim Ecol 80:365 - 374 Sloat JV, Hull M (2004) Computing discharge using the velocity - index method. SonTek Division Report 08 - 2004, San Diego. Smith KM, King DK (2005a) Dynamics and ex tent of larval lake sturgeon Acipenser fulvescens drift in the Upper Black River, Michigan. J Appl Ichthyol 21:161 - 168 Smith KM, King DK (2005b) Movement and habitat use of yearling and juvenile Lake sturgeon in Black Lake, Michigan. Trans Am Fish Soc 134 :1159 - 1172 Smith SJ, Marsden JA (2009) Factors affecting sea lamprey egg survival. N Am Jour Fish Man 29:859 - 868 Stoks R, Geerts AN, Meester LD (2014) Evolutionary and plastic responses of freshwater invertebrates to climate change: realized patterns and future potential. Evol Appl 7:42 - 55 change revisited: disentangling environmental and genetic responses in a wild bird population. Proc Nat Acad Sci 105(36):13492 - 134 96 Urban MC, Richardson JL, Freidenfelds NA (2014) Plasticity and genetic adaptation mediate amphibian and reptile responses to climate change. Evol Appl 7:88 - 103 Warren DR, Robinson JM, Josephson DC, Sheldon DR, Kraft CE (2012) Elevated summer temperatu res delay spawning and reduce red construction for resident brook trout ( Salvelinus fontinalis ). Glob Change Biol 18:1804 - 1811 77 West - Eberhard MJ (2003) Developmental plasticity and evolution. Oxford University Press, New York Wilson AJ, Kruuk LEB, Coltman DW (2005) Ontogenetic patterns in heritable variation for body size: using random regression models in a wild ungulate population. Am Nat 166(6):E177 - E192 Wund MA (2012) Assessing the impacts of phenotypic plasticity on evolution. Integr Comp Biol 52(1): 5 - 15 Xu CL, Letcher BH, Nislow KH (2010) Size - dependent survival of brook trout Salvelinus fontinalis in summer: effects of water temperature and stream flow. Jour Fish Biol 76:2342 - 2369 78 Chapter 3 INCREASING THERMAL R EGIME S REVEAL CRYPTIC GEN ETIC VARIATION DURING EARLY ONTOGEN ETIC STAGES OF LAKE STURGEON ABSTRACT Climate change is affecting thermal conditions worldwide. Understanding organismal responses associated with predicted changes are essential for predicting fut ure population persistence. However, few studies have examined the effects of both increased mean and variance in temperature on organismal traits, particularly during early life stages. Using lake sturgeon ( Acipenser fulvescens ) from Black Lake, MI, we te sted whether phenotypes and behavior would vary among families incubated in different C) and two fluctuating temperature treatments (10 - 19 C) representing temperatures experienced in the river and a simulated anthropogenic disturbance. Body length, body area, and yolk - sac are a were quantified at hatch. Traits quantified at emergence (i.e. the onset of exogenous feeding) included time to emergence, body length at emergence, and growth from hatch to emergence. A significant family - by - treatment interaction was detected for traits C) treatment. Families also varied in growth (h 2 =0.33) and the timing of emergence (h 2 =0.24). Results demonstrate that increases in mean thermal regimes may reveal cryptic genetic variation potentially leading to differential survival between genotypes thereby altering the genetic architecture of the population. INTRODUCTION Climate change and variability are altering thermal regime s on a global scale 79 (IPCC 2013). Within the next century, mean temperatures are predicted to increase by approximately 2 to 5 º C due to human influences (Estay et al. 2014). Additionally, increases in the magnitude and frequency of extreme climatic events are altering the variability in thermal patterns (Pincebourde et al. 2012; Bauerfeind and Fischer 2014). Thermal changes are known to directly affect biological functions in wild populations. For instance, previous studies have documented advances in the t iming of reproduction (Parmesan 2006), faster growth rates (Drinkwater 2005), and decreases in species abundance (Jonsson and Jonsson 2009) in response to increasing temperatures. Therefore, understanding the ecological and evolutionary responses of wild p opulations to predicted thermal changes due to climate change has become a major goal for ecologists and climate researchers (Walther et al. 2002; Hansen et al. 2012). Ectotherms are of particular conservation concern under predicted climatic changes give n that temperature directly affects their physiology including growth, reproduction, and locomotion (Deutsch et al. 2008; Jonsson and Jonsson 2009). For example, one well - documented trend among ectotherms is the temperature - size rule in which rearing tempe rature is inversely related to body size (Kingsolver and Huey 2008; Diamond and Kingsolver 2010). Individuals that are reared in warm temperatures grow faster, but are typically smaller than individuals reared in cold temperatures (Atkinson 1994; Forster e t al. 2011). Numerous studies have examined phenotypic trait changes in response to rearing temperatures in ectotherms given that changes in traits associated with fitness (i.e. body size) typically occur during early life stages which are characterized by high mortality, and are considered to be the most susceptible to thermal fluctuations (Jonsson and Jonsson 2009). Additionally, environmental 80 conditions experienced during early ontogeny have long - term consequences on individual fitness by affecting devel opmental rates, phenotypes, physiology, and behavior at later ontogenetic stages (i.e. ontogenetic contingency, Diggle 1994; Orizaola et al. 2010; Huey et al. 2012; Crespi and Warne 2013; Pittman et al. 2013; Dammerman et al. 2015). Changes in phenotypic a nd behavioral trait variation are commonly visualized as thermal reaction norms that are constructed by rearing individuals of known genotype over a range of temperatures that remain constant and quantifying the variation in phenotypes that are expressed ( Angilletta 2009). These plots have been useful for modeling the effects of climate change by exposing individuals to ecologically relevant changes in temperature and comparing their phenotypes to individuals reared under current thermal conditions in a lab oratory setting. However, few studies have reared individuals under fluctuating incubation temperatures (i.e. diurnal temperature changes) that mimic conditions encountered in the wild (Bauerfeind and Fischer 2014). Recent research has documented that rear ing individuals under fluctuating temperatures more accurately represents responses to climate change given that organisms encounter daily fluctuations in temperature (Niehaus et al. 2012; Bauerfeind and Fischer 2014), and temperature variability is predic ted to affect traits associated with fitness to a greater extent than mean temperature alone (Paaijmans et al. 2013). In this study, we reared different families of lake sturgeon ( Acipenser fulvescens ; Rafinesque 1817) under constant and fluctuating ther mal egg incubation conditions and quantified phenotypic and behavioral variation at hatch and at the timing of emergence. Lake sturgeon are long - lived ectotherms and a useful species for examining the effects 81 of thermal changes because spawning events, emb ryogenesis, and larval development occur across a range of temperatures. In the spring, adults migrate to riverine areas for spawning (Auer 1996). The timing of spawning is multi - modal, and has been observed when water temperature is between 8.8 C (Bruch and Binkowski 2002; Forsythe et al. 2012). During spawning events, females release demersal, adhesive eggs which are fertilized by multiple males (Thiem et al. 2013). Fertilized eggs from full and half - sibling families attach to the substrate and i ncubate without parental care under site - specific conditions until hatch (Duong et al. 2011). Larvae typically hatch within 5 - 14 days dependent on water temperature (Kempinger 1988; Smith and King 2005) then burrow into the substrate to endogenously feed o n yolk - sac reserves (Hastings et al. 2013). Yolk - sac utilization and the timing of emergence when larvae disperse downstream to begin exogenously feeding are dependent on several abiotic conditions within larval rearing sites including temperature (Duong e t al. 2011). Our first objective was to determine whether larval phenotypes and the timing of emergence would vary among individuals of different families experimentally reared in constant and fluctuating thermal incubation environments. We predicted that phenotypic variation within the fluctuating treatments would be greater than observed under constant treatments (Niehaus et al. 2012). Additionally, we predicted that phenotypes would vary among families given the large variance attributed to genetic (fam ily) effects quantified in a recent study on lake sturgeon responses to different river flow regimes (Dammerman et al. 2015). Our second objective was to assess whether thermal environments experienced during egg incubation would affect larval growth and b ehavior during subsequent ontogenetic stages. Quantifying trait changes conditional 82 on environments experienced during previous ontogenetic stages increases understanding of how fluctuations in local thermal regimes will affect phenotypic trait variation a nd survival of different genotypes during critical developmental stages of ectotherms. MATERIALS AND METHODS Study site During the 2012 spawning season, adult lake sturgeon were sampled daily using long - handled dip nets on the Upper Black River (UBR). The UBR is the largest tributary of Black Lake located in Cheboygan County, Michigan (Fig. 1; Smith and King 2005). Adults migrate into the UBR from late April to early June to spawn among shallow (~1 - 3 meters) rocky areas (Baker and Borgeson 1999; Forsythe e C. The wadable conditions of the stre am and presence of a streamside rearing facility provide the opportunity to collect gametes from spawning adults and conduct experimental temperature manipulations and monitoring of larval traits. Fertilizations Gametes were collected from five females and ten males on May 3 rd , 4 th , and 6 th of 2012. Eggs were collected from spawning females, placed in sealed plastic bags, and stored with river water to maintain eggs at ambient river temperature. Milt was collected using 20 - mL syringes and placed on ice. Gam etes were transported to the streamside rearing facility and fertilizations were conducted within twelve hours of collection. 83 Approximately 200 eggs from each female were placed on 1 - mm mesh screens within polymerized vinyl chloride couplings (31.90cm 2 ). C ouplings were used to keep families create half - sibling families. Fertilized eggs were widely distributed within the couplings and left undisturbed for 30 minutes to allow a dhesion to the mesh screens. Thermal treatments Fertilized eggs were incubated in heathtrays where temperature was controlled using heating and cooling units to produce two constant and two fluctuating thermal treatments (Fig. 2). Constant treatments, col C), represented mean temperatures that characterized the early and late season spawning periods, respectively (Forsythe et al. 2012). Fluctuating treatments consisted of ambient river temperatures (natural fluctuations of 1 - 3 deg rees daily in river temperature) and a variable treatment simulating an anthropogenic disturbance with a range of 1 - 9 degrees of change per day. Water temperature within each treatment was measured daily using Onset HOBO pressure loggers (Cape Cod, Massach usetts, USA). Fertilized eggs were slowly acclimated to treatment temperatures (two - degree change per hour) prior to placement in heathtrays. Families were replicated among treatments and dead eggs were removed daily to prevent fungal infections. Larval tr aits Thirty larvae from each family within each treatment (N=1200 total) were anesthetized with tricaine methanesulfonate (MS - 222; 25 mg/mL) and photographed at hatch using a digital camera. Photographs were analyzed using Image J analysis software (Versi on 1.34, free - ware) to quantify three phenotypic traits: body length (mm), 84 body area (mm 2 ), and yolk - sac area (mm 2 ; Dammerman et al. 2015). Eighteen individuals were subsampled from five of the half - sibling families within each treatment (N=360 total) and randomly assigned to individual plastic incubation chambers (12.7cm by 6.35cm) to monitor until the timing of emergence. Incubation chambers contained gravel substrate to provide cover during the endogenous feeding period, ambient river water, and mesh sid ing to allow for continual water passage (~0.1m/sec). Chambers were randomly placed across three fiberglass raceways (3.7m by 0.67m) with respect to family and thermal incubation treatment to prevent any influence of chamber location on the timing of emerg ence. Incubation chambers were monitored daily until larvae emerged from the substrate with depleted yolk - sac reserves to begin exogenously feeding. Larvae were anesthetized and photographed again to quantify body length at emergence (mm), growth from hatc h to emergence (mm), and the time from hatch to emergence (days). Individuals within the thermal treatments were predicted to hatch at different times given that stream temperatures experienced during the endogenous feeding stage directly affect yolk - sac u tilization and timing of emergence in fishes including sturgeon (Hardy and Litvak 2004; Duong et al. 2011). Therefore, we tested the influence of the ambient river temperature experienced while in the incubation chambers on the three traits measured at the time of emergence. Temperature data was collected hourly within the streamside facility using a YSI 5200 Recirculating System Monitor (Xylem, Inc.). For each individual, we calculated the mean temperature (MeanT) and range in temperature (daily maximum mi nus minimum, RangeT) experienced during the endogenous feeding period in the incubation chamber. 85 All research was conducted under animal use and care procedures approved by the Michigan State University Institutional Animal Care and Use Committee. Statis tical analysis Statistical analyses were performed using the program, R (R Development Core Team) and the Bayesian inference package, MCMCglmm (Hadfield 2010). General linear mixed - effect models were used to test the contribution of Family, Treatment, and a Family - by - Treatment interaction on the larval traits measured at hatch. Traits were analyzed separately by fitting the full model: Eq. (1) ijk was the random residual error. Treatment was fit as a fixed effect. Family and the Family - by - Treatment interaction terms were fit as random ef fects where , , and . Models were parameterized in terms of Family effects as opposed to fitting an animal - specific random effect (i.e. animal model, Wilson et al. 2010). Given the simple pedigree structure and the possible presence of a genotype - by - environment interaction, this parameterization was equivalent and simpler to implement. Traits quantified at the time of emergence were analyzed using the full model with the addition of MeanT and RangeT experienced while in the incubation chamber as fixed effects. The correlation the variables in the same model. All models included a burn - in of 10,000, thinning interval of 100, and a total of 200,000 iterations which were chosen after examining diagnostic plots when using 86 default starting values in the MCMCglmm package. Variances of random effects were giv en an inverse - gamma distribution with a scale and shape of 0.001 which results in a non - informative prior. A sensitivity analysis conducted by fitting models with a proper prior attributing a higher degree of the phenotypic variance to genetic effects reve aled there was little effect of the priors on the posterior distribution. Model selection was based on Deviance Information Criterion (DIC) which incorporates model fit and complexity based on an expected deviance parameter and the effective number of para meters estimated in the model (Spiegelhalter et al. 2002). We compared DIC estimates for the full model against all possible simpler models. The model with the lowest DIC estimate was selected as the model of best fit. When models were within two DIC value s of each other, the simplest and less complex model was chosen as the model of best fit (Spiegelhalter et al. 2002). Parameter estimates were calculated as the modes of the posterior distributions estimated from the model of best fit. The 95% highest post erior density (HPD) was also estimated for each parameter. For traits where the model of best fit included a Family effect but no Family - by - Treatment interaction, the mode of the posterior distribution of the narrow - sense heritability (h 2 ) and 95% HPD inte rvals were computed. Heritability estimates reported were obtained using the expression derived in Dammerman et al. 2015 (see Appendix 1). RESULTS Mean body length at hatch ranged from 10.19 (±0.09) to 13.84 (±0.09) mm among families across the four treat ments. A large range in phenotypic variation among families was observed within the fluctuating treatment groups; however, the greatest range was observed in the warm thermal treatment with Families D and H differing 87 approximately 3 - mm in size (Fig. 3). Th e cold treatment group showed the least amount of variation with families varying up to 1.2 - mm in size. A significant Family - by - Treatment interaction was detected for body length (Table 1), and explained approximately 39% of the phenotypic variance observe d (Table 2). Mean body area and mean yolk - sac area at hatch also varied significantly among families across treatments ranging from 16.82 (±0.17) to 28.46 (±0.32) mm 2 and 6.91 (±0.11) to 8.11 (±0.14) mm 2 , respectively. Family - by - Treatment interactions were significant for both traits (Table 1), explaining approximately 43% of the variation observed in body area and 13% of the variation observed in yolk - sac area (Table 2). Approximately 87% of the larvae from the five half - sibling families emerged from the substrate to begin exogenously feeding. Individuals began to emerge approximately 10 days after being placed in the incubation chambers. Families showed significant variation in the mean time spent in the incubation chambers ranging from 12.67 (±0.34) to 1 4.21 (±0.26) days; however, no Family - by - Treatment interaction or effect of the temperatures experienced while in the incubation chambers (MeanT and RangeT) were detected for the timing of emergence (Table 1). The mode of the heritability (h 2 ) distribution explaining the similarity in the timing of emergence between larvae from the same family and 95% HPD were estimated as 0.24 (0.002, 0.706; Table 2). Families also showed significant variation in growth from hatch to emergence ranging from 10.97 (±0.35) t o 15.39 (±0.47) mm among families. Individuals that were incubated as eggs in the two fluctuating treatments differed approximately 1.3 to 1.7 - mm in growth (Fig. 4). Similar to the traits measured at hatch, the largest range in mean growth among families w as observed for individuals that were reared as eggs in the 88 warm treatment with families differing approximately 2.7 - mm in size (Fig. 4). No Family - by - Treatment interaction or effect of temperatures experienced while in the incubation chambers (MeanT and R angeT) were detected (Table 1). The mode of the h 2 distribution for growth and 95% HPD were estimated as 0.33 (0.021, 0.904; Table 2). Variation in mean body size at the timing of emergence was low among families ranging from 24.57 (±0.20) to 25.45 (±0.16) mm, and no significant effects were detected (Table 1). DISCUSSION The early life stages of many ectothermic species are characterized by high levels of mortality (Fuiman and Werner 2002; Vitt and Caldwell 2014). Accordingly, it is during these early ont ogenetic stages that selection may have a considerable effect. Increased temperatures encountered during early ontogeny are known to affect phenotypic and behavioral traits in ectothermic species (Atkinson 1996; Angilletta et al. 2004). However, there is a limited understanding of how thermal variability during egg incubation will affect trait variation at hatch and during sequential ontogenetic stages. Empirical work quantifying the effects of both increased mean and variability in temperature are vital to predict the ecological and evolutionary responses of populations to future climatic regimes. In this study, we experimentally manipulated egg incubation temperatures in a common garden experiment and demonstrated large effects of thermal regimes, genetic (family) effects, and their interaction on phenotypic trait variation during consecutive early life stages. Rearing temperatures influence phenotypic variation observed at hatch 89 Rearing temperature had a significant effect on the range of phenotypic varia C (Bruch a ther mal treatment may have been due to the presence of cryptic genetic variation revealed under stressful thermal conditions. Cryptic genetic variation is not commonly observed under typically encountered conditions, and is often exposed in stressful environme nts (Badyaev et al. 2005; Ledón - Rettig et al. 2014). Results indicate that increases in mean thermal regimes (not just thermal variability) may reveal cryptic genetic variation expressed as differences among families thereby increasing phenotypic variation and the likelihood of exposing variant phenotypes to selection. Cryptic variation can also be exposed when a genotype - by - environment interaction is present (Schlichting 2008; Paaby and Rockman 2014). In our common garden experiments, Family - by - Treatment interactions detected for body length, body area, and yolk - sac area measured at hatch indicate that families were responding differently to the same thermal incubation environments. A large proportion of the observed phenotypic variation in the traits (~13 - 43%) was explained by the Family - by - Treatment interaction. Other studies (Beacham 1988 and 1990), found that genotype - by - temperature interactions explained a substantial portion of observed variation in morphometric traits in Oncorhynchus gorbuscha and On corhynchus keta after rearing 90 individuals under different thermal regimes indicating that water temperature has a significant effect on the genotypic expression of traits during early development. Additionally, the cryptic genetic variation observed in phe notypic traits at hatch may lead to differential survival among genotypes if selection favors larger individuals. Therefore, consideration of parentage, interaction effects, and cryptic genetic variation is important to understand how predicted thermal cha nges will affect phenotypic variation within wild populations. Fluctuating and constant rearing temperatures influence the range of phenotypic variation observed in larval traits In the variable and ambient treatments where temperatures fluctuated throug hout the day, variation in body length at hatch among families was over 1.5 times as large as variation observed in the constant cold treatment (Fig. 3). Although diurnal temperature fluctuations were much larger in the variable than the ambient treatment, families showed similar responses in the ranges of phenotypic variation observed in the two fluctuating treatments. Results indicate that rearing individuals at a constant temperature reflecting the mean thermal regime observed in the wild will not accura tely represent the extent of phenotypic variation that would be observed and on which selection would act on in wild - reared larvae when even modest daily temperature fluctuations occur. Similarly, Niehaus et al. (2012) found that growth and developmental r ates of striped marsh frogs ( Limnodynastes peronii ) reared in fluctuating thermal treatments were continually under - predicted based on reaction norms constructed from individuals reared at constant rearing temperatures. Therefore, rearing individuals under fluctuating treatments provides a more accurate representation of phenotypic trait 91 variance and any potential genotype - by - environment interactions expected in the wild in response to climatic changes. Rearing temperatures affect trait variation at a subse quent ontogenetic stage Thermal conditions experienced during embryogenesis are known to affect phenotypic traits associated with survival such as shape, color, behavior, and size at developmental stages beyond hatch in several taxa of ectotherms (e.g. f ishes, Martell et al. 2005; reptiles, Goodman 2008; amphibians, Orizaola et al. 2010). In our study, larvae reared in the fluctuating thermal treatments showed lower overall growth than those reared in the constant treatments indicating that egg incubation temperature affects trait variation at a subsequent ontogenetic stage (Fig. 4). Additionally, families within the warm treatment (Fig. 4) show the greatest range in growth indicating that differences observed among families in body size at hatch are persi sting to a later ontogenetic stage. The persistence of a large range of variation in growth in the warm treatment and presence of non - zero heritability (h 2 = 0.33) for growth indicates that thermally - induced, cryptic variation could potentially be maintain ed across ontogenetic stages. Given that body size during early life stages is associated with survival with larger individuals typically having lower levels of mortality (Brown and Shine 2004; Fischer et al. 2011; Perez and Munch 2011), differential grow th among families may lead to changes in the genetic architecture of the population if selection favors genotypes that produce larger offspring during critical development periods. In our thermally - induced phenotypic variation and survival in the population given that we only measured 92 individuals to emergence and that the long generation time of lake sturgeon requires long - term monitoring. However, the early life stages of fishes are where mortality is highest (>99%; Chambers and Trippel 1997), and thus, selection operating on phenotypic variation is likely effective at altering population genetic architecture through different (family - specific) survival. Several abiotic and biotic factors affect the timing of transition between early ontogenetic stages (Day and Rowe 2002). In larval fishes, the timing of emergence is typically under stabilizing selection (Crozier et al. 2008) and can be dependent on heritable variation, maternal effects, an d/or environmental conditions such as temperature (Curry et al. 1995; Einum and Fleming 2000; Skoglund et al. 2011). In our experiment, there were no effects of thermal regimes experienced while in the incubation chambers. Alternatively, approximately 24% of the variation observed in the timing of emergence was attributed to differences among families. Heritability estimates for behavioral traits are rare in wild populations of fishes even though several behaviors are believed to have a heritable component (Mittelbach et al. 2014). Our estimate of h 2 =0.24 for emergence time is comparable to other reported h 2 estimates of 0.15 - 0.20 for behavioral traits in fishes (Carlson and Seamons 2008; Chervet et al. 2011). However, our estimate represented the upper - limi t of the narrow - sense heritability (resembling heritability in the broad - sense) given that we were unable to separate out additive genetic variance from other genetic effects. Despite the limitation, our h 2 estimate provides insight on the potential change in emergence time and ability of the population to respond genetically (i.e. changes in gene frequency) to changes in thermal regimes. 93 Families differ in growth at a subsequent ontogenetic stage We observed no differences in body size at the time of emerg ence although families differed in the timing of emergence and growth from hatch to emergence. One explanation is that families that had smaller body sizes at hatch may have remained in the substrate longer to allow more time to grow thereby emerging from the substrate at the same size as individuals that spent less time in the substrate and were larger at hatch. Alternatively, differences in growth and time spent in the substrate may be attributed to differences among families in size of endogenous yolk re serves. Differences between maternal provisioning of endogenous yolk - sac reserves have been well - documented in ectotherms where offspring with larger yolk reserves typically have higher survival and grow to a larger body size (Kamler 2005; Dziminski and Ro berts 2006; Gagliano and McCormick 2007). In our experiment, we observed differences in yolk - sac area at hatch due to a Family - by - Treatment interaction. Therefore, differences in yolk sac area among families at hatch may have led to differential growth and time spent in the substrate feeding on endogenous yolk - sac reserves until the timing of emergence. Conclu sions and future directions Our findings have important implications regarding the impact of predicted thermal changes due to climate change on pheno typic variation in ectothermic species. Understanding how cryptic genetic variation is maintained across sequential ontogenetic stages in wild populations is vital to understanding how populations can respond to environmental change (Ledón - Rettig et al. 20 14). Additionally, reaction norms constructed from rearing individuals at constant temperatures are not reliable 94 indicators of response under variable conditions. Therefore, researchers may wish to rally occurring (e.g. diel) varying temperature into experimental designs for treatments (Paaijmans et al. 2013). Given our findings on behavioral and phenotypic traits that are tied to survival, further empirical work addressing how incubation conditions and differences among families affect trait variation and survival at later ontogenetic stages would be beneficial to predict changes in population levels of genetic diversity due to environmental perturbations. Knowledge within these research areas is ess ential to understand the ecological responses and evolutionary potential of a population to the impacts of climate change. 95 APPENDIX 96 Table 3.1 Step - wise model comparisons and DIC - values used in model selection for the six larval traits quantified at hatch and emergence in the experiment. Larval Trait Model DIC (a) Hatch Body Length Family + Family*Treatment 2739.86* Family 3135.40 3334. 08 3637.86 Body Area Family + Family*Treatment 5687.97* Family 6123.70 6290.38 6821.93 Yolk - sac Area Family + Family*Treatment 2668.96* F amily 2779.48 2824.39 3045.09 (b) Emergence Time to Emergence Family + Family*Treatment 1219.22 Family 1218.40 Family 1218.43 Family 1217.83* 1233.12 1372.54 Emergence Body Length Family + Family*Treatment 1125.42 Family 1125.11 t + MeanT + Family 1126.24 Family 1124.31 1128.53 1124.33* Total Growth Family + Family*Treatment 1146.03 Family 1145.74 + Treatment + MeanT + Family 1143.58 Family 1141.91* 1166.68 1207.01 *Indicates the model of best fit. 97 Table 3.2 Variance components and 95% highest posterior density (HPD) estimated from the models of best fit. Heritability (h 2 ) estimates are provided for traits where the model of best fit included a significant family effect, but no family - by - treatment interaction. Phenotypic Traits Component Var ±95% HPD Body Length Family*Treatment 0.35 (0.00, 0 .26) Family <0.01 (0.20, 0.62) Residual 0.56 (0.51, 0.60) Body Area Family*Treatment 4.85 (0.00, 2.06) Family 0.02 (2.85, 8.10) Residual 6.56 (6.01, 7.04) Yolk - sac Area Family*Treatment 0.08 (0.00, 0.05) Family <0.01 (0.05, 0.15) Residual 0.5 3 (0.49, 0.57) Time to Emergence Family 0.12 (0.00, 1.60) Residual 3.21 (2.69, 3.69) h 2 0.24 (0.00, 0.71) Total Growth Family 0.19 (0.02, 1.81) Residual 2.37 (2.08, 2.86) h 2 0.33 (0.02, 0.90) *Indicates an interaction between the components 98 Figure 3.1 The study location on the Upper Black River, the largest tributary of Black Lake, Michigan showing the spawning locations where adult lake sturgeon were sampled during the 2012 spawning season Figure 3.2 The four egg incubation temperatures (warm, cold, variable, and ambient) in which individuals were reared in during the 2012 spawning season 99 Figure 3.3 Thermal reaction norm for body length at hatch for lake sturgeon larvae subjected to the four thermal treatments. A significant genotyp - - - - Figure 3.4 Thermal reaction norm for larval lake sturgeon growth measu red from hatch to the timing of emergence. A significant family effect was detected (h2=0.33) indicating that differential size among families measured at hatch was persisting to a later ontogenetic stage 100 LITERATURE CITED 101 LITERATURE CITED Angilletta, M.J., P.H. Niewiarowski, A.E. Dunham, A.D. Leaché, and W.P. Porter. 2004. - history perspective with sceloporine lizards. American Naturalist 164(6) :E168 - E183 A ngilletta, M.J. 2009. Thermal adaptation: a theoretical and empirical synthesis. New York, NY: Oxford University Press Atkinson, D. 1994. Temperature and organism size a biological law for ectotherms? Advances in Ecological Research 25 :1 - 58 Atkinson, D . 1996. Ectotherm life history responses to developmental temperature. Pages 183 - 204 in I.A. Johnston and A.F. Bennett, eds. Animals and temperature: Phenotypic and evolutionary adaptation. Cambridge University Press, Cambridge. Auer, N. 1996. Response o f spawning lake sturgeons to change in hydroelectric facility operation. Transactions of the American Fisheries Society 125 :66 - 77 Badyaev, A. 2005. Stress - induced variation in evolution: from behavioural plasticity to genetic assimilation. Proceedings of the Royal Society B 272 :877 - 886 Baker, E.A. and D.J. Borgeson. 1999. Lake sturgeon abundance and harvest in Black Lake, Michigan, 1975 - 1999. North American Journal of Fisheries Management 19 :1080 - 1088 Bauerfeind, S.S. and K. Fischer. 2014. Simulating cli mate change: temperature extremes but not means diminish performance in a widespread butterfly. Population Ecology 56 :239 - 250 Beacham, T.D. 1988. A genetic analysis of early development in pink ( Oncorhynchus gorbuscha ) and chum salmon ( Oncorhynchus keta ) at three different temperatures. Genome 30 :89 - 96 Beacham, T.D. 1990. A genetic analysis of meristic and morphometric variation in chum salmon ( Oncorhynchus keta ) at three different temperatures. Canadian Journal of Zoology 68 :225 - 229 Brown, G.P. and R. S hine. 2004. Maternal nest - site choice and offspring fitness in a tropical snake ( Tropidonophis mairii , Colubridae). Ecology 85(6) :1627 - 1634 Bruch, R.M. and F.P. Binkowski. 2002. Spawning behavior of lake sturgeon ( Acipenser fulvescens ). Journal of Applied Ichthyology 18 :570 - 579 102 Carlson, S.M. and T.R. Seamons. 2008. A review of quantitative genetic components of fitness in salmonids: implications for adaptation to future change. Evolutionary Applications 1 :222 - 238 Chambers, E.A. and R.C. Trippel, eds. 199 7. Early life history and recruitment in fish populations. Chapman and Hall, London Chervet, N., M. Zöttl, R. Schürch, M. Taborsky, and D. Heg. 2011. Repeatability and heritability of behavioural types in a social cichlid. International Journal of Evoluti onary Biology 2011 :1 - 15 Crespi, E.J. and R.W. Warne. 2013. Environmental conditions experienced during the tadpole stage alter post - metamorphic glucocorticoid response to stress in an amphibian. Integrative and Comparative Biology 53(6) :989 - 1001 Crozier, L.G., A.P. Hendry, W. Lawson, T.P. Quinn, N.J. Mantua, J. Battin, R.G. Shaw, and R.B. Huey. 2008. Potential responses to climate change in organisms with complex life histories: evolution and plasticity in Pacific salmon. Evolutionary Applications 1(2) :25 2 - 270 Curry, R.A., D.L.G. Noakes, and G.E. Morgan. 1995. Groundwater and the incubation and emergence of brook trout ( Salvelinus fontinalis ). Canadian Journal of Fisheries and Aquatic Sciences 52 :1741 - 1749 Dammerman, K.J., J.P. Steibel, and K.T. Scribner . 2015. Genetic and environmental components of phenotypic and behavioral trait variation during lake sturgeon ( Acipenser fulvescens ) early ontogeny. Environmental Biology of Fishes 98 :1659 - 1670 Day, T. and L. Rowe. 2002. Developmental thresholds and the evolution of reaction norms for age and size at life - history transitions. American Naturalist 159(4) :338 - 350 Diamond, S.E. and J.G. Kingsolver. 2010. Environmental dependence of thermal reaction norms: host plant quality can reverse the temperature - size r ule. American Naturalist 175(1) :1 - 10 Diggle, P.K. 1994. The expression of andromonoecy in Solanum hirtum (Solanaceae): phenotypic plasticity and ontogenetic contingency. American Journal of Botany 81(10) :1354 - 1365 Deutsch, C.A., J.J. Tewksbury, R.B. Huey , K.S. Sheldon, C.K. Ghalambor, D.C. Haaks, and P.R. Martin. 2008. Impacts of climate warming on terrestrial ectotherms across latitude. 105(18) :6668 - 6672 Drinkwater, K.F. 2005. The response of Atlantic cod (Gadus morhua) to future climate change. ICES Jo urnal of Marine Science 62 :1327 - 1337 103 Duong, T.Y., K.T. Scribner, J.A. Crossman, P.S. Forsythe, and E.A. Baker. 2011. Environmental and maternal effects on embryonic and larval developmental time until dispersal of lake sturgeon ( Acipenser fulvescens ). Can adian Journal of Fisheries and Aquatic Sciences 68 :643 - 654 Dziminski, M.A. and J.D. Roberts. 2006. Fitness consequences of variable maternal provisioning in quacking frogs ( Crinia georgiana ). Journal of Evolutionary Biology 19 :144 - 155 Einum, S. and I.A. Fleming. 2000. Selection against late emergence and small offspring in Atlantic salmon ( Salmo salar ). Evolution 54(2) :628 - 639 Estay, S.A., M. Lima, and F. Bozinovic. 2014. The role of temperature variability on insect performance and population dynamics i n a warming world. Oikos 123 :131 - 140 Fischer, B., B. Taborsky, and H. Kokko. 2011. How to balance the offspring quality - quantity tradeoff when environmental cues are unreliable. Oikos 120 :258 - 270 Forster, J., A.G. Hirst, and D. Atkinson. 2011. How do org anisms change size with changing temperature? The importance of reproductive method and ontogenetic timing. Functional Ecology 25 :1024 - 1031 Forsythe, P.S., J.A. Crossman, N.M. Bello, E.A. Baker, and K.T. Scribner. 2012. Individual - based analyses reveal hi gh repeatability in timing and location of reproduction in lake sturgeon ( Acipenser fulvescens ). Canadian Journal of Fisheries and Aquatic Science 68 :1 - 13 Fuiman, L.A. and R.G. Werner, eds. 2002. Fishery science: the unique contributions of early life sta ges. Blackwell Publishing, Oxford Gagliano, M. and M.I. McCormick. 2007. Maternal condition influences phenotypic selection on offspring. Journal of Animal Ecology 76 :174 - 182 Goodman, R.M. 2008. Latent effects of egg incubation temperature on growth in t he lizard Anolis carolinensis . Journal of Experimental Zoology 309(9) :525 - 533 Hadfield, J.D. 2010. MCMC methods for multi response generalised linear mixed models: the MCMCglmm R package. Journal of Statistical Software 33(2) :1 - 22 Hansen, M.M., I. Olivie ri, D.M. Waller, and E.E. Nielsen. 2012. Monitoring adaptive genetic responses to environmental change. Molecular Ecology 21 :1311 - 1329 Hardy, R.S. and M.K. Litvak. 2004. Effects of temperature on the early development, growth, and survival of shortnose st urgeon, Acipenser brevirostrum , and Atlantic 104 sturgeon, Acipenser oxyrhynchus , yolk - sac larvae. Environmental Biology of Fishes 70 :145 - 154 Hastings, R.P., J.M. Bauman, E.A. Baker, and K.T. Scribner. 2013. Post - hatch dispersal of lake sturgeon ( Acipenser fu lvescens , Rafinesque, 1817) yolk - sac larvae in relation to substrate in an artificial stream. Journal of Applied Ichthyology 29(6) :1208 - 1213 Huey, R.B., M.R. Kearney, A. Krockenberger, J.A.M. Holtum, M. Jess, and S.E. Williams. 2012. Predicting organismal vulnerability to climate warming: roles of behaviour, physiology, and adaptation. Philosophical Transactions of the Royal Society Biological Sciences 367 :1665 - 1679 Intergovernmental Panel on Climate Change. 2013. Climate change 2013: the physical science basis. Contribution of working group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge, United Kingdom and New York Jonsson, B. and N. Jonsson. 2009. A review of the likely effects of climate change on anadromous Atlantic salmon Salmo salar and brown trout Salmo trutta , with particular reference to water temperature and flow. Journal of Fish Biology 75 :2381 - 2447 Kamler, E. 2005. Parent - egg - progeny relationships in teleost fishes: an e nergetics perspective. Reviews in Fish Biology and Fisheries 15 :399 - 421 Kempinger, J.J. 1988. Spawning and early life history of Lake sturgeon in the Lake Winnebago system, Wisconsin. American Fisheries Society Symposium 5 :110 - 122 Kingsolver, J.G. and R. B. Huey. 2008. Size, temperature, and fitness: three rules. Evolutionary Ecology Research 10 :251 - 268 Ledón - Rettig, C.C., D.W. Pfennig, A.J. Chunco, and I. Dworkin. 2014. Cryptic genetic variation in natural populations: a predictive framework. Integrative and Comparative Biology 54(5) :783 - 793 Martell, D.J., J.D. Kieffer, and E.A. Trippel. 2005. Effects of temperature during early life history on embryonic and larval development and growth in haddock. Journal of Fish Biology 66(6) :1558 - 1575 Mittelbach, G. G., N.G. Ballew, and M.K. Kjelvik. 2014. Fish behavioral types and their ecological consequences. Canadian Journal of Fisheries and Aquatic Science 71 :927 - 944 Niehaus, A.C., M.J. Angilletta, M.W. Sears, C.E. Franklin, and R.S. Wilson. 2012. Predicting the physiological performance of ectotherms in fluctuating thermal environments. Journal of Experimental Biology 215 :694 - 701 105 Orizaola, G., E. Dahl, and A. Laurila. 2010. Compensating for delayed hatching across consecutive life - history stages in an amphibian . Oikos 119 :980 - 987 substrate. Nature Reviews Genetics 15 :247 - 258 Paaijmans, K.P., R.L. Heinig, R.A. Seliga, J.I. Blanford, S. Blanford, C.C. Mudrock, and M.B. Thomas. 2013 . Temperature variation makes ectotherms more sensitive to climate change. Global Change Biology 19(8) :2373 - 2380 Parmesan, C. 2006. Ecological and evolutionary responses to recent climate change. Annual Review of Ecology, Evolution, and Systematics 37 :637 - 669 Perez, K.O. and S.B. Munch. 2011. Extreme selection on size in the early lives of fish. Evolution 64(8) :2450 - 2457 Pincebourde, S., E. Sanford, J. Casas, and B. Helmuth. 2012. Temporal coincidence of environmental stress events modulates predation ra tes. Ecology Letters 15 :680 - 688 Pittman, K., M. Yúfera, M. Pavlidis, A.J. Geffen, W. Koven, L. Ribeiro, J.L. Zambonino - Infante, and A. Tandler. 2013. Fantastically plastic: fish larvae equipped for a new world. Reviews in Aquaculture 5(s1) :S224 - 267 Schli chting, C.D. 2008. Hidden reaction norms, cryptic genetic variation, and evolvability. Annals of the New York Academy of Sciences 1113 :187 - 203 Skoglund, H., S. Einum, and G. Robertsen. 2011. Competitive interactions shape offspring performance in relation to seasonal timing of emergence in Atlantic salmon. Journal of Animal Ecology 80 :365 - 374 Smith, K.M. and D.K. King. 2005. Dynamics and extent of larval lake sturgeon Acipenser fulvescens drift in the Upper Black River, Michigan. Journal of Applied Ichthy ology 21 :161 - 168 Spiegelhalter, D.J., N.G. Best, B.P. Carlin, and A. van der Linde. 2002. Bayesian measures of model complexity and fit. Journal of the Royal Statistical Society: Series B 64(4) :583 - 639 Thiem, J.D., H.P. Dumont, G. Van Der Kraak, and S.J. Cooke. 2013. Biology of lake sturgeon (Acipenser fulvescens) spawning below a dam on the Richelieu River, Quebec: behaviour, egg deposition, and endocrinology. Canadian Journal of Zoology 91 :175 - 186 106 Vitt, L.J. and J.P. Caldwell. 2014. Herpetology: An int roductory biology of amphibians and reptiles (4 th ed.). Academic Press, London Walther, G.R., E. Post, P. Convey, A. Menzel, C. Parmesan, T.J.C. Beebee, J.M. Fromentin, O. Hoegh - Guldberg, and F. Bairlein. 2002. Ecological responses to recent climate chang e. Nature 416 :389 - 395 Wilson, A.J., D. Réale, M.N. Clements, M.M. Morrissey, E. Postma, C.A. Walling, L.E.B. Animal Ecology 79 : 13 - 26 107 Chapter 4 EFFECTS OF PARENTAGE AND MICROHABITAT VARIATI ON WITHIN ADULT - SELECTED SPAWNING SI TES ON LAKE STURGEON GROWTH DURING EA RLY LIFE STAGES ABSTRACT - Understanding the impact of environmental variation on organismal traits during early life stages is essential for predicting changes i n populations. Maternal effects including oviposition site selection are one source of variation. Female choice of sites can vary based on microhabitat variation which directly affects larval development and survival. However, the persistence of microhabit at effects across sequential early ontogenetic stages is limited. In 2013, we collected fertilized lake sturgeon (Acipenser fulvescens) eggs from the Black River, Michigan, just prior to hatch from an adult - selected spawning location, and quantified three microhabitat variables (water depth, discharge, and substrate size). Larval body length, body area, and yolk - sac area were quantified at hatch. Body size was measured for an additional four weeks post - emergence. Genetic - based parentage analysis was conduct ed using fin clips. The traits at hatch varied due to the microhabitat variables, but not due to additive genetic effects. Growth significantly varied among larvae with the greatest range in body size observed at 41 days post - hatch. Additive genetic varian ce covaried with age. Narrow - sense heritability estimates for body length ranged from 0.45 to 0.64. Results demonstrate that female - selected, microhabitat variation influenced offspring phenotypic variation at hatch, but had no influence on body length at sequential ontogenetic stages. Additionally, the additive genetic effects explained roughly half the variation observed in 108 body size, but individuals showed considerable variation in growth which may be due to differences in the degree of plasticity among individuals. INTRODUCTION Knowledge on how environmental variation affects organismal trait variation during early life stages is essential for predicting changes in populations due to potential threats including habitat degradation and environmental chang e (Monaghan 2008; Aubry et al. 2013). Maternal effects including choice of mates, selection of oviposition sites, and maternal provisioning are well - documented sources of variation that have pervasive effects on phenotypic variation in offspring (Mousseau and Fox 1998; Räsänen and Kruuk 2007). Female selection of oviposition site determines the abiotic conditions offspring will encounter during development (Rudolf and Rödel 2005; Lowerri - Barbieri 2011). Thus, females are expected to select oviposition sites that are conducive to offspring growth and survival in order to maximize their own reproductive success (Kern et al. 2013). However, several factors have been shown to influence female choice of oviposition site including proximity of mates, predation ris k to offspring, micro - habitat variation, and habitat stability (Refsnider and Janzen 2010; Silberbush and Blaustein 2011). Therefore, studies examining conditions at oviposition sites provide insight on environmental cues influencing adult decisions, and t he effect of those decisions on larval growth and survival (Lowerri - Barbieri 2011; Eilers et al. 2013). In stream systems, female oviposition site selection is influenced by abiotic conditions that can vary spatially and temporally (Reich and Downes 2003), and potentially on micro - geographic scales (<1m 2 ; Merten et al. 2013). These abiotic conditions include substrate type and size (Van Liefferinge et al. 2005), water velocity 109 (Lancaster 2010), water depth (Reich and Downers 2003), and temperature (Torgerse n et al. 1999) which are known to directly affect development and survival during early life stages (Refsnider and Janzen 2010). For example, interstitial spaces between substrate provide cover and protection from predators for newly hatched larvae that ar e highly vulnerable to predation (Bunn and Arthington 2002; Boucher et al. 2014). Thus, identifying variation in stream microhabitat conditions is essential for understanding abiotic habitat needs, and factors affecting potential population levels of recru itment in stream species (McAdam 2011). In stream fishes, spatial and seasonal variation in stream conditions cause adults to frequently shift their use of different microhabitats (Kanno et al. 2012), thereby exposing offspring to a wide variety of rearin g conditions. In polygamous species, spatial variation in microhabitat use during spawning events results in larvae of the same and different genotype to be reared under different environmental regimes, even within a small geographic area (e.g., within the same spawning site). These environmental regimes are known to have persistent effects on trait variation in fishes (Martell et al. 2005; Martell and Kieffer 2007) by influencing growth and behavior during individuals of the same and differing genotype (Dammerman et al. 2015). Additionally, the relative contribution of both environmental and additive genetic effects can vary among life stages (Houle 1998). Therefore, understandi ng the effects of microhabitat rearing environments on larval phenotypic variation and survival requires determination of genetic parentage and quantification of traits across sequential ontogenetic stages in order to accurately predict potential populatio n levels of recruitment. 110 In this study, we examined the influence of additive genetic effects and microhabitat variation on phenotypic variation during early ontogenetic stages of lake sturgeon ( Acipenser fulvescens ). Lake sturgeon are long - lived, iteropar ous fish, that have been numerically depressed due to overharvest, habitat loss and degradation, and limited recruitment (Hayes and Caroffino 2012). Adults migrate to riverine areas to spawn in polygamous groups, and provide no post - ovulatory care as eggs incubate under site - specific conditions (Smith and King 2005; Duong et al. 2011). Similar to other fishes, mortality is high (~99%) during the egg and larval period (Forsythe et al. 2013). Long - term success of restoration programs includes protection of sp awning and rearing habitats (Hayes and Caroffino 2012), and identification of factors that affect growth and survival during critical early life stages. Genetic (parentage) effects as well as egg incubation conditions are known to affect phenotypic trait v ariation and behavior associated with survival at hatch and at the timing of emergence (Dammerman et al. 2015; Dammerman et al., In Review ). However, knowledge of how traits vary following emergence due to additive genetic effects and stream microhabitat v ariables is limited. The objectives of this study were to: i) quantify microhabitat variation within an adult - selected rearing habitat, ii) determine whether microhabitat conditions affects larval trait variation at hatch, and iii) quantify the influence o f additive genetic effects and environmental variation on post - emergence growth. MATERIALS AND METHODS Study site The study was conducted in 2013 on the Upper Black River (UBR), the largest tributary of Black Lake. Black Lake is a 3500 hectare lake locate d in Cheboygan 111 County, Michigan (Fig. 1) that supports a well - studied population of approximately 1200 adult lake sturgeon (Pledger et al. 2013). Adult sturgeon migrate to the UBR from late April to early June to spawn among a 1.5 - km stretch of the river ( Spawning sites are characterized by shallow water depth (~1 - 3 meters) co ntaining cobble, pebble, and gravel substrate which provide cover for newly - hatched larvae (Forsythe et al. 2012). The wadable conditions of the stream and presence of a streamside facility provide the opportunity to collect eggs from the substrate and rea r individuals in a captive setting to monitor traits across sequential ontogenetic stages. Egg collection and microhabitat variables In 2013, adult lake sturgeon were monitored daily by wading the UBR at least once per day from May 1 st to June 9 th . Once s pawning behavior and deposited eggs were observed on at a commonly utilized spawning site referred to as Site B (Duong et al. 2011), temperature was monitored daily using an Onset HOBO pressure logger ( Cape Cod, Massachusetts, USA) to calculate cumulative thermal units (CTU) to estimate egg hatch date using the methods of Kempinger (1988). On May 8 th , one day prior to estimated egg hatch date, seven transect lines (A - G) were placed at 5 - meter intervals along the stream bank parallel to the river at Site B ( Fig. 2). Abiotic microhabitat variables including substrate size, water depth, and water discharge were sampled at 1 - m intervals along each transect line (N=91 sampling points). Sampling along transect lines was conducted using a steel apparatus which con tained a pole, a 0.216 x 0.279m frame attached to the bottom of the pole, and a 112 Canon Powershot D10 underwater camera mounted 0.5m above the frame. The presence of the steel frame allowed for a consistent river area to be sampled, and the camera height of 0.5 - m allowed the area being photographed to remain at a constant 0.053m 2 (0.201 x 0.262m). At each sampling point, a photo was taken to quantify mean substrate size at a later date. Point estimates of water depth (m) and water velocity (m 2 /sec) at each po int were measured using a wading rod and Marsh McBirney flow meter. Discharge (m 3 /sec) at each sampling point was calculated by multiplying velocity point measurements by the 1 - m river area. Temperature was not measured in the study given that it was confo unded by the time of sampling. River depth (m) and discharge (m 3 /sec) were simultaneously collected hourly upstream of the spawning site during the entire spawning season using a staff gauge and a HOBO pressure logger . Sturgeon eggs were collected at each sampling point by disturbing the substrate within the steel frame for ten seconds upstream of a triangular kick net. Live eggs (Fig. 2) were transported to the streamside facility and kept separate by sampling point until hatch. Dead eggs were discarded to prevent microbial infection. Mean depth and discharge during the incubation period Water depth (m) and water discharge (m 3 /sec) at each transect sampling point were measured again on May 18 th . Water depth and discharge measurements collected at each tra nsect point were compared to the mean depth and mean discharge collected by the data logger on the same date by taking a ratio of the point estimate divided by the logger measurement. Ratios between the two dates were compared using linear regressions. The calculated regression for water depth (y = 0.980x + 0.002, R 2 = 0.948; Fig. S1) and regression for water discharge (y = 0.917x - 0.010, R 2 = 0.948; Fig. S2) 113 indicated that ratios between the two sampling points were highly correlated. Therefore, we estima ted mean water depth and mean river discharge at each transect point during the entire incubation period by multiplying the mean depth and discharge collected daily from the data logger by the mean ratio calculated at each transect point. Mean substrate size Substrate photos collected at each transect point were analyzed using Image J software (Version 1.34, free - ware). The axis with the greatest linear distance was measured for each particle, and recorded as the particle length (mm). Given the large n umber of particles in each photo, a preliminary analysis was conducted by placing a 3x3 grid over the image. Grids measured 0.067m wide by 0.087m long. We subsampled 10% of the photos from transect sampling points where eggs were collected, and measured th e length of all particles in the photo. Then, particles within three randomly - selected grids within each photo were measured. Measurements within the entire photo and the grids were compared using a linear regression (y = 1.526x 1.315, R 2 = 0.954; Fig. S 3). Three - randomly sampled grids were measured for all remaining photos, and the regression was used to estimate mean substrate size (mm) in the entire photo. Particles within each photo were classified based on the Wentworth Classification System as modif ied by Cummins where size categories included: sand/fine particles (>2mm), gravel (2 - 16mm), pebble (16 - 64mm), cobble (64 - 256mm), and boulders (>256mm; Wentworth 1922; Cummins 1962). Larval traits At hatch, three - hundred and fifty - nine larvae were subset f rom the 64 transect points where eggs were collected. Larvae were anesthetized with tricaine 114 methanesulfonate (MS - 222; 25 mg/mL) and photographed using digital photography. Image J software was used to quantify body length (mm), body area (mm 2 ), and yolk - s ac area (mm 2 ; Dammerman et al. 2015). Larvae were randomized with respect to transect point and assigned to modified, liter - sized milk jugs (15.2 x 6.8cm) which served as individual incubation chambers. Chambers contained gravel substrate to provide cover during endogenous feeding, mesh covering, and independent water sources allowing for continual water movement (~0.1m/sec) within chambers. Chambers were monitored multiple times daily until larvae emerged from the substrate with depleted yolk - sac reserves to begin exogenously feeding. Due to algal accumulation on screens, one fish independently emerged from the substrate on May 21 st before substrate was removed and chambers were cleaned on May 25 th . Dead larvae (N=25) due to natural mortality within the in cubation chambers were discarded due to microbial infection. All surviving larvae were anesthetized and photographed again on May 28 th to determine post - emergence body length (mm). Larvae remained in the chambers without substrate and were photographed wee kly to quantify body size during the critical growth and development period. Larvae were photographed five times between hatch and 41 days post - hatch until individuals grew to the size of the incubation chambers when a fin - clip could be taken non - lethally for genetic analysis. During rearing, larvae were fed brine shrimp nauplii ( Artemia spp) three times daily to satiation for three weeks until transitioning to processed frozen blood worms (chironomid midge larvae) for the remainder of the rearing period. L arvae that died during the rearing period (N=42) were kept separate and preserved in 95% ethanol for genetic analysis. 115 Genetic analysis DNA was extracted from all larval fin clips using DNeasy(R) extraction kits (QIAGEN, Inc.), and quantified using a Nanod rop 1000 spectrophotometer (Thermo Fisher Scientific, Inc.). Larvae were genotyped at 12 microsatellite loci: AfuG68B (McQuown et al. 2002); AfuG68, AfuG74, AfuG9, AfuG63, AfuG56, AfuG195, AfuG160, AfuG112, AfuG204 (Welsh et al. 2003); Spl 120 (McQuown et al. 2000); and Aox 27 (King et al. 2001). Polymerase chain reaction and genotyping were conducted using methods described in Duong et al. 2011. Due to low DNA concentration, 1.5uL of forward and reverse primer were used to increase signals and cycle number was increased to 35 for all loci during the annealing phase. All genotyping scores were assigned by two experienced lab personnel, and a 10% subset of the samples were re - genotyped to empirically estimate genotyping error. Genotyping error was calculated as the number of allelic errors divided by the total number of alleles that were genotyped. Parentage analysis Parentage analysis was conducted using the software programs, CERVUS 3.0 (Kalinowski et al. 2007) and COLONY 2.0.5.8 (Jones and Wang 2010). In C ERVUS, our simulation included 10,000 offspring with a strict level of 95% and a relaxed level of 80% to determine the confidence in parentage assignment based on the population estimates of allelic frequencies. Our parentage analysis included all 262 adul ts from the 2013 spawning run, an 85% estimate of the proportion of adults captured (K. Scribner, unpublished data), and an empirical genotyping error estimate of 0.40%. Adults with the highest maximum likelihood score in COLONY and highest, positive LOD s core in 116 CERVUS were assigned as the most likely candidate parents. Assignment of the most likely parental pair in our study was based on concordance between the two programs. Statistical analysis Statistical analyses were performed using the statistical pr ogram, R (v. 3.1.2, R Development Core Team). We used the package, glmmADMB (Skaug et al. 2006) to determine whether the number of live lake sturgeon eggs collected at each transect point was dependent on the three microhabitat variables. We tested for mul ticollinearity between the three variables by quantifying the variance inflation factor (vif) using the between pairs of variables. Correlation estimates and vif values w ere low between the three variables (r < 0.29, vif < 1.22) so all three were included in the model. We fit a negative binomial regression with a log link function as: [Eq 1] Linear mixed - effect models were used to test for the effects of the environmental covariates and genetic (additive) e ffect associated with larval phenotypic traits measured at hatch. The three traits at hatch (BL, BA, and YSA) were analyzed McCullagh 2014). Models were fit using restricted m aximum likelihood estimation. Traits were fit to the full model: [Eq 2] 1 17 where microhabitat variables were fit as fixed effects, Animal refers to the additive genetic effect w ith ) where the additive variance ( ) is estimated based on the comparison of phenotypes among relatives denoted by the additive . Traits were then fit to the simpler model: [Eq 3] and all subsequent simpler models in a step - wise manner by eliminating one variable at a time. Likelihood - ratio tests were used to determine the model of best fit. We tested for spatial autocorrelation among the transect sampling points given that ecological variables can be correlated at small (<1m) spatial scal es (Legendre 1993), and this spatial autocorrelation can influence trait variation for individuals reared in close proximity. First, we fit the three traits measured at hatch and the four additional measurements of BL to a model including the three environ mental covariates [Eq 3] with each trait fit separately. We extracted the residuals from each model, and plotted semi - - variograms depict the similarity between variances (e.g. residuals) ba sed on spatial distance, and are used to determine the correlation structure present in the data (Dormann et al. 2007). Visualizations of the semi - variogram plots and correlation curves indicated there was no evidence of correlation between the residuals b ased on the distance between them. Therefore, no spatial effect terms were included in the analysis of the larval traits. To determine how additive variance associated with BL varied over time, we used a multi - step approach. First, we fit each of the five measurements of BL separately to the full animal model with the three environmental covariates [Eq 2]. We used 118 likelihood ratio tests to determine there was no influence of the environmental covariates on BL except at hatch so we dropped them from the mod el. Next, we fit each of the five measurements of BL to an animal model with the population mean as a fixed effect: [Eq 4] We used the results from these models to investigate patterns of heteroskedastic ity. We also investigated the pattern of phenotypic correlation between repeated measures. Next, we fit a random regression animal model (Meyer and Hill 1997; Schaeffer 2004) on a quadratic polynomial of Age, and used the residual variance heterogeneity pa ttern estimated from the individual time point analysis to account for heterogeneous environmental variances. The resulting model was: [Eq 5] where Age was fit as a fixed effect, refers to the random regression coefficients for the additive genetic effect as a function of age de scribed by an intercept, slope (linear), output from the model weighted by the residual variances estimated from [Eq 4] resulting in five estimates of heterogeneous residual variances. We used the weighted model given the lack of convergence in a more complex model that allowed additive genetic and permanent environmental effects to vary with age. Lastly, we estimated the same model in [Eq 5] without heterogeneous variances, and compared the go odness - of - fit of between the model and the weighted model based on Akaike Information Criterion (AIC) values. The narrow - sense heritability (h 2 ) for BL was estimated at each Age by taking the additive 119 genetic variance divided by the sum of the additive an d residual variances estimated from the random regression animal model. For larvae that were assigned a mother (N=224), we tested for the presence of a maternal effect by fitting the body size of the female (FemaleTL) as an additional random effect in [Eq 2] for the three traits measured hatch, and in [Eq 4] for each measurement of BL where FemaleTL ). In all models, variance in the trait explained by the maternal effect was negligible ranging from 0 - 1.7%. Therefore, we did not fit a random regression model that included the maternal effect term. RESULTS Adult lake sturgeon were observ ed spawning at Site B on May 3 rd , and fertilized eggs were observed on the substrate on May 5th. Mean water temperatures were high during early May averaging between 15 - 17 º C resulting in a shorter egg incubation time. CTU values calculated from daily mean water temperature indicated that eggs were expected to hatch on May 9 th . Fertilized eggs incubated under site - specific conditions from May 3 rd May 8th until being collected along the transect lines. Eggs were collected from Site B when CTU was approximat ely 66.5 which is within the traditional range of CTUs at time of hatch (58.1 to 71.4) estimated in previous studies on the UBR (Smith and King 2005). Five - hundred and forty - one live lake sturgeon eggs were collected from 64 of the sampling points along th e transect lines at Site B. Most eggs were collected near the center of the river, and ranged from 0 - 73 per transect point (Fig. 2). The largest proportion (54%) of the total number of live eggs were collected from transect lines C and D. 120 The number of li ve eggs collected was positively associated with mean water discharge and mean substrate size at each transect sampling point with discharge having the largest effect. Mean (± SE) water discharge for all transect points at Site B was 0.69 (± 0.06) m 3 /sec, and mean substrate size 11.74 (± 0.70) mm. Gravel was the most prevalent substrate type at Site B, but size varied among transect lines with some points contained predominantly pebble substrate (Table 1). Most eggs were collected near the center of the riv er where water was deeper, and the largest number of eggs were collected from transect lines C and D (Fig. 2). Mean depth among all transect points was 0.75 (± 0.02) m, and also varied among transect lines (Table 1), but did not affect the number of live e ggs collected at each sampling point ( ). The mean proportion of larvae sub - sampled from each transect sampling point at hatch, and measured in the experiment was 78.8%. Larvae showed considerable variation in the three traits measured at hatch. Mean larval BL at hat ch was 13.02 (±0.05) mm, and ranged from 9.80 to 16.18 mm. Mean BA was 28.17 (±0.22) mm 2 , and ranged from 17.37 to 41.10 mm 2 . Both BL and BA at hatch were dependent on the three microhabitat variables at each transect point. BL and BA were negatively assoc iated with depth and substrate size, but positively associated with discharge (Table 3). Depth had the largest influence on both traits. YSA varied among larvae with a mean of 7.95 (±0.06) mm 2 , and ranged from 5.13 to 12.42 mm 2 . YSA was also dependent on mean depth and mean discharge at the transect sampling points, but did not vary due to mean substrate size (Table S1). YSA was negatively associated with mean depth, and positively associated with mean discharge which had the largest effect (Table 3). The 121 additive genetic effect was not significant for all three traits (p>0.05) when fitting the linear mixed models. In the parentage analysis of the larvae, mean expected heterozygosity for the 12 loci genotyped was 0.576. Allelic diversity ranged from 2 to 11 with a mean of 5.25 alleles per locus. Eleven of the loci were in Hardy - Weinberg equilibrium. The combined non - exclusion probability for the parent pairs was 0.00008456. Ninety percent (N=274) of the offspring that were genotyped were assigned at least on e parent with a 75% degree of concordance in parental assignments between CERVUS and COLONY. Forty females were assigned offspring with a mean of 5.60 (± 1.03) offspring per female; however, 51% of the assigned larvae came from seven females. Sixty - one mal es were assigned offspring with a mean of 4.30 (± 0.46) offspring. Seven males were assigned 10 or more offspring, and contributed approximately 31% of the total offspring. Post - hatch survival in the incubation chambers was approximately 93%. Mean growth f rom hatch (Age 1 days) to four weeks post - emergence (Age 41 days) was 28.19 (±0.31) mm, but individuals showed considerable variation in size (Table 2). The largest range in BL was observed at 41 days post - hatch as the largest and smallest individuals show ed a 28.08 mm difference in BL (Fig. 3). The heterogeneous variance model was a better fit model (AIC = 5496.1) over the model that did not account for heterogeneous variances (AIC = 5646.0). Additive genetic variance estimates obtained from the random reg ression animal model increased with age (Fig. 4). The narrow - sense heritability (h 2 ) for BL also varied with age, and ranged from 0.45 to 0.64 (Fig. 5). DISCUSSION 122 Microhabitat conditions within adult selected oviposition sites directly affect offspring g rowth and development in a variety of species (Refsnider and Janzen 2010); however, knowledge of the persistent effects of these environmental variables on larval trait variation across sequential ontogenetic stages is limited. In this study, quantificatio n of microhabitat variables within the stream, rearing of wild - produced larvae, and genetically - based parentage analysis allowed for determination of how phenotypic trait variation across early life stages was affected by adult - selected microhabitats and a dditive genetic effects. Depth, discharge, and substrate size were quantified within adult - selected microhabitats given their known influence on lake sturgeon spawning site selection and reproductive success (Bruch and Binkowski 2002; Chiotti et al. 2008). The three microhabitat variables had significant effects on the number of live eggs collected from the stream, and on the three larval traits measured at hatch: body length (BL), body area (BA), and yolk - sac area (YSA). The number of live lake sturgeon e ggs collected from the stream was positively associated with mean discharge and mean substrate size at each transect sampling point with discharge having the largest effect. Adequate discharge levels (0.15 to 1.40 m/s; Hay - Ch mielewski and Whelan 1997) are vital for adequate mixing of sperm and eggs during broadcast spawning events. Additionally, high discharge levels can dislodge fish eggs from the substrate surface. The largest number of live eggs were collected from transect lines C and D. These transect lines had moderate discharge, depth, and substrate size, and may have represented an optimal rearing microhabitat rearing conditions for the eggs. However, 123 certain eggs were from multiple females except for the larvae that were genotyped in the study. Water depth was negatively associated with the three larval traits. Deeper waters are typically associated with higher levels of dissolved oxygen given the lower water tempe ratures which slow down metabolic rates (Guppy and Withers 1999). Therefore, subtle differences in water depth between transect sampling points may have slowed embryo and subsequently larval growth and development. Similarly, Baltz et al. (1998) found a po sitive association between deeper water and growth in juvenile red drums ( Sciaenops ocellatus ) in nursery habitats presumably due to temperature and dissolved oxygen concentrations. Discharge was positively associated with the three larval sturgeon traits. Reduced water flows cause sediment deposition (Kempinger 1988), but increases in flow may keep the adhesive surface of the incubating sturgeon eggs clear of sediment, debris and microbial communities (Fujimoto et al. 2013) thereby positively influencing d evelopment within the egg and larval phenotypes at hatch. Body length of brown trout ( Salmo trutta ) smolts have also been positively associated with increasing discharge regimes (Jonsson et al. 2001) indicating the importance of discharge on larval develop ment of stream fishes. Increases in mean substrate sizes negatively influenced the three traits at hatch. Substrate is important for the early life stages of sturgeon given that the fertilized, adhesive eggs adhere to the substrate surface during incubati on (Zubair et al. 2012), and newly hatched larvae immediately burrow into the interstitial spaces between the substrate to escape predation (Hastings et al. 2013). Therefore, extreme substrate sizes are detrimental to lake sturgeon development given that s mall sand and fine particles 124 can cause egg mortality by coating the egg surface, and large substrate sizes such as boulders provide little interstitial space to protect eggs during incubation and during the period of endogenous feeding by larvae. Preferenc e for spawning sites containing gravel substrate has been well - documented in sturgeon (McAdam 2011; Boucher et al. 2014) and other stream fishes (Gerkens and Thiel 2001), and was the predominant substrate type documented within the microhabitats in the stu dy. Growth during early life stages of fish is predicted to vary due to intrinsic and extrinsic factors including genetic and environmental effects (Kamler 2008). In the study, the three microhabitat variables had no effect on post - emergence growth. Alter natively, a significant proportion of the phenotypic variation observed in larval growth was due to additive genetic effects (Fig. 5). We observed an initial increase in h 2 after larvae came out of incubation chambers. Increasing h 2 with age is concordant with traits expressed later in life will always show more variance due to the cumulative effects of allelic variants (Houle 1998) and any new gene expression (Wilson et al. 2005). In our study, the decrease in additive variance after 25 days could be due either to increased environmental variance or due to the 42 offspring that died during the rearing period which further reduced our sibling groups thereby reduced our estimat e of the degree of resemblance. Larval growth trajectories varied significantly among individuals with the range in BL increasing with increasing age (Fig. 3). Although it is difficult to distringuish, larval growth curves showed heterogeneous slopes as in dividuals that were large at hatch did not remain hatch during the entire experiment. 125 Individual differences in growth trajectories have been documented in several vertebrate species (Nussey et al. 2007) including fishes (Dower et al. 2009). BL narrow - sen se heritability (h 2 ) estimates which document the degree of resembles among related individuals (Wilson et al. 2010) also varied across age (Fig. 5). We had high estimates of 0.45 to 0.64 are comparable to the heritability estimates of 0.42 for body length estimated at hatch and 0.33 for growth from hatch to emergence estimated in previous studies on the Black Lake sturgeon population (Dammerman et al. 2015; Dammerman et al., In Review ). The growth plot of the larvae demonstrated that individuals within the same maternal half - sibling groups also significantly varied in size. Several studies have reported individual - level variation size, based on differences in the degree of phenotypic plasticity in response to environmental variation (Evans et al. 2010). One explanation is that individual differences in growth among larvae in our study could be differences in feeding efficiency. For example, lipid storage capacity increases with body size in juvenile lake sturgeon resulting in differential growth among indivi duals (Volkman et al. 2004). Therefore, larvae that grew larger in our study may have utilized the hatchery diet more effectively thus growing to a larger body length. Our study has important implications for lake sturgeon rehabilitation strategies. Protec tion of spawning and rearing habitats requires maintaining moderate flows, adequate water depth, and the presence of suitable substrate types. All three of these abiotic conditions can be drastically altered by habitat loss, and the increase in extreme wea ther events predicted to occur due to climate change. Additionally, variation in individual growth trajectories among and within different maternal sibling groups can influence growth in the hatchery, and the size of offspring that are being stocked back 126 i nto the wild population. Size is an important indicator of the over - winter survival of stocked individuals given that larger individuals are better at escaping predators and better foragers (Dower et al. 2009). Therefore, optimizing growth in the hatchery is essential for designing effective stocking strategies. Additional observations on larval phenotypic traits at multiple spawning sites would benefit this study by allowing for the variation in permanent environmental effects during early life stages t o be modeled. Using weighted regressions, we were able to model heterogeneous variances in this study and confirm it was a better fit than the model without heterogeneous variances; however, a random regression animal model incorporating the change in addi tive variance and permanent environmental effects would provide a deeper understanding of all the factors affecting traits associated with survival during sequential early ontogenetic stages. Additional future work in the stream could include examining bio tic microhabitat variables that influence offspring development such as the spatial distribution of aquatic macro - invertebrates which are important prey species for juvenile lake sturgeon (Peterson et al. 2007). Knowledge on all stream conditions affecting larval and juvenile growth and development allows for the development of more effective lake sturgeon rehabilitation strategies. 127 APPENDIX 128 Table 4.1 Mean (± Standard Error) values of the three microhabi tat variables during the egg incubation period at the sixty - four sampling points where live lake sturgeon eggs were collected. Depth and discharge during the egg incubation period were predicted using hourly data logger measurements. Transect Line # Sampl ing Points Mean Depth (m) Mean Discharge (m 3 /sec) Mean Substrate Size (mm) A 6 1.23 (0.03) 0.75 (0.04) 13.99 (2.51) B 12 0.90 (0.09) 0.68 (0.08) 12.22 (1.65) C 11 0.80 (0.10) 0.73 (0.08) 9.51 (1.01) D 12 0.69 (0.09) 0.68 (0.10) 12.94 (1.39) E 10 0.75 (0.06) 0.82 (0.09) 10.40 (1.55) F 10 0.64 (0.06) 0.75 (0.09) 12.34 (2.56) G 3 0.63 (0.14) 0.88 (0.21) 10.58 (2.16) Table 4.2 Mean ( ± Standard Error) and range in body size observed among larvae for the five measurement samples taken from hatch (Age 1 ) to four - weeks post - emergence (Age 41). Age (days) Mean (mm) Minimum (mm) Maximum (mm) 1 13.02 (0.05) 9.80 16.18 20 22.73 (0.08) 19.19 27.02 27 27.72 (0.13) 19.82 33.97 34 33.67 (0.22) 23.88 44.53 41 41.29 (0.32) 25.43 53.51 Table 4.3 Coefficient s, standard errors (SE), and p - values estimated from the models of best fit for the three traits measured at hatch. Hatch Traits Fixed Effect Estimate SE P - value Body Length Intercept 13.79 0.34 <0.001 Mean Depth - 1.07 0.29 <0.001 Mean Discharge 0.76 0.28 0.007 Mean Substrate - 0.04 0.01 0.001 Body Area Intercept 30.57 1.43 <0.001 Mean Depth - 3.40 1.25 0.007 Mean Discharge 3.29 1.18 0.006 Mean Substrate - 0.18 0.05 <0.001 Yolk - sac Area Intercept 7.36 0.36 <0.001 Mean Depth - 0.01 0.35 0.99 Mean Discharge 0.71 0.33 0.04 129 Table 4 . 4 Pairwise model comparisons for the three hatch traits fit to models (M) including all three environmental variables and a reduced model. Log - likelihood values, chi - square values, p - values (P), and degrees of freed om (df) obtained from the likelihood ratio tests are presented. Hatch Traits Model Log - likelihood Chi - square df P Body Length M1: Mean Depth + Mean Discharge + Mean Substrate Size - 364.46 M2: Mean Depth + Mean Discharge - 369.62 10.310 1 0.001* Body Area M1: Mean Depth + Mean Discharge + Mean Substrate Size - 761.96 M2: Mean Depth + Mean Discharge - 768.23 12.552 1 <0.001* Yolk - sac Area M1: Mean Depth + Mean Discharge + Mean Substrate Size - 414.81 M2: Mean Depth + Mean Discharge - 414.92 0.224 1 0.6358 * Indicates a statistically significant difference (p<0.05) between models. 130 Figure 4. 1 The study location on the Upper Black River, the largest tributary of Black Lake, Michigan showing the spawning site, Site B, where lake sturgeon egg s were collected along transect lines and the larval sampling location where wild - produced larvae were captured 131 Figure 4. 2 Schematic layout of the seven transect lines (A - G) and sampling points where the microhabitat variables were quantified and la ke sturgeon eggs were collected. Bubbles indicate the number of live eggs collected at each transect sampling point. Figure 4. 3 Individual growth trajectories quantified from hatch (Age 1 day) to four weeks post - emergence (Age 41 days). Individual body size was measured at hatch, and once weekly between the timing of emergence (Age 21 days) and the end of the experiment (Age 41 days). 132 Figure 4. 4 Additive genetic variance ( ) and residual variance ( ) estimates quantified at each age in the random regression animal model. Figure 4. 5 The narrow - sense heritability (h 2 ) estimates for body size quantified at each age the larvae were photographed. The h 2 esti mates were obtaining by dividing the additive genetic variance by the sum of the additive variance and residual variance. 133 Figure 4.6 The relationship between the depth ratios calculated on May 8 th and May 18 th . The ratios were calculated as the poin t estimates of depth at each transect sampling point divided by the mean depth measurement from the data logger. The linear regression equation (y = 0.980x + 0.002, R 2 = 0.948) indicated the ratios were highly correlated between the two sampling dates. Figure 4.7 The relationship between the discharge ratios calculated on May 8 th and May 18 th . The ratios were calculated as the point estimates of discharge at each transect sampling point divided by the mean discharge measurement from the data logger. T he linear regression equation (y = 0.917x - 0.010, R 2 = 0.948) indicated the ratios were highly correlated between the two sampling dates. 134 Figure 4.8 The relationship between mean substrate size when measuring rocks in three randomly selected grids ve rsus the entire substrate photo. The linear regression equation ( y = 1.526x 1.315, R 2 = 0.954) indicated that measuring three random grids accurately estimated mean substrate size in the entire photo. 135 LITER ATURE CITED 136 LITERATURE CITED Aubry, L.M., R.F. Rockwell, E.G. Cooch, R.W. Brook. C.P.H. Mulder, and D.N. Koons. 2013. Climate change, phenology, and habitat degradation: drivers of gosling body condition and juvenile survival in less er snow geese. Global Change Biology 19:149 - 160 Auer, N. 1996. Response of spawning lake sturgeons to change in hydroelectric facility operation. Transactions of the American Fisheries Society 125:66 - 77 Barriga, J.P., N.A. Espinós, J.M. Chiarello - Sosa, a nd M.A. Battini. 2013. The importance of substrate size and interstitial space in the microhabitat selection by the stream - dwelling catfish Hatcheria macraei (Actinopterygii, Trichomycteridae). Hydrobiologia 705(1):191 - 206 Boucher, M.A., S.O. McAdam, and J.M. Shrimpton. 2014. The effect of temperature and substrate on the growth, development, and survival of larval white sturgeon. Aquaculture 430:139 - 148 Brommer, J.E., K. Rattiste, and A.J. Wilson. 2008. Exploring plasticity in the wild: laying date - tempe rature reaction norms in the common gull Larus canus . Proceedings Royal Society B 275:687 - 693 Bruch, R.M. and F.P. Binkowski. 2002. Spawning behavior of lake sturgeon ( Acipenser fulvescens ). Journal of Applied Ichthyology 18:570 - 579 Bunn, S.E. and A.H. A rthington. 2002. Basic principles and ecological consequences of altered flow regimes for aquatic biodiversity. Environmental Management 30(4):492 - 507 Clifford, D. and P. McCullagh. 2014. The regress package R package version 1.3 - 14. Cummins, K.W. 1962. An evaluation of some techniques for the collection and analysis of benthic samples with special emphasis on lotic waters. American Midland Naturalist 67:477 504 Dammerman, K.J., J.P. Steibel, and K.T. Scribner. 2015. Genetic and environmental components of phenotypic and behavioral trait variation during lake sturgeon (Acipenser fulvescens) early ontogeny. Environmental Biology of Fishes 98:1659 - 1670 Diggle, P.K. 1994. The expression of andromonoecy in Solanum hirtum (Solanaceae): Phenotypic plasticity a nd ontogenetic contingency. American Journal of Botany 81(10):1354 - 1365 137 Dormann, C.F., J.M. McPherson, M.B. Araujo, R.Bivand, J. Bolliger, G.Carl, R.G. Davies, A. Hirzel, W. Jetz, W.D. Kissling, I. Kuhn, R. Ohlemuller, P. R. Peres - Neto, B. Reineking, B. Schroder, F.M. Schurr, and R. Wilson. 2007. Methods to account for spatial autocorrelation in the analysis of species distributional data: a review. Ecography 30:609 - 628 Dower, J.E., P. Pepin, and G. Cheon Kim. 2009. Covariation in feeding success, size - a t - age and growth in larval radiated shanny (Ulvaria subbifucata): insights based on individuals. Journal of Plankton Research 31(3):235 - 247 Duong, T.Y., K.T. Scribner, J.A. Crossman, P.S. Forsythe, and E.A. Baker. 2011. Environmental and maternal effects on embryonic and larval developmental time until dispersal of lake sturgeon ( Acipenser fulvescens ). Canadian Journal of Fisheries and Aquatic Science 68:643 - 654 Eilers, S., L.B. Pettersson, and E. Öckinger. 2013. Micro - climate determines oviposition site selection and abundance in the butterfly Pyrgus armoricanus at its northern range margin. Ecological Entomology 38:183 - 192 Evans, M.L., B.D. Neff, and D.D. Heath. 2010. Quantitative genetics and translocations experiments reveal genotype - by - environment ef fects on juvenile life - hisory traits in two populations of Chinook salmon (Oncohynchus tshawytscha). Journal of Evolutionary Biology 23:687 - 698 Forsythe, P.S., J.A. Crossman, N.M. Bello, E.A. Baker, and K.T. Scribner. 2012. Individual - based analyses revea l high repeatability in timing and location of reproduction in lake sturgeon ( Acipenser fulvescens ). Canadian Journal of Fisheries and Aquatic Science 68:1 - 13 Forsythe, P.S., K.T. Scribner, J.A. Crossman, A. Ragavendran, and E.A. Baker. 2013. Experimental assessment of the magnitude and sources of lake sturgeon egg mortality. Transactions of the American Fisheries Society 142:1005 - 1011 Fox, J. and S. Weisberg. 2011. An R Companion to Applied Regression, Second Edition. Thousand Oaks CA: Sage. Gates, K. K., C.C. Vaughn, and J.P. Julian. 2015. Developing environmental flow recommendations for freshwater mussels using the biological traits of species guilds. Freshwater Biology 60(4):620 - 635 Gerkens, M. and R. Thiel. 2001. Habitat use of Age - 0 Twaite shad ( alosa Fallax Lacepede, 1803) in the tidal freshwater region of the Elbe River, Germany. 138 Hastings, R.P., J.M. Bauman, E.A. Baker, and K.T. Scribner. 2013. Post - hatch dispersal of lake sturgeon ( Acipenser fulvescens , Rafinesque, 1817) yolk - sac larvae in rel ation to substrate in an artificial stream. Journal of Applied Ichthyology 29(6):1208 - 1213 Michigan Department of Natural Resources, Fisheries Division Special Report 62, Lansing Houle, D. 1998. How should we explain variation in the genetic variance of traits? 102/103:241 - 253 Jones, O. R. and J.L. Wang. 2010. COLONY: a program for parentage and sibship inference from multilocus genotype data. Molecular Ecology Resou rces 10, 551 555. Kalinowski, S. T., M.L. Taper, and T.C. Marshall. 2007. Revising how the computer program CERVUS accommodates genotyping error increases success in paternity assignment. Molecular Ecology 16, 1099 - 1106 Kanno, Y., C.U. Schmidt, S.B. Cook , and H.T. Mattingly. 2012. Variation in microhabitat use of the threatened spotfin chub (Erimonax monachus) among stream sites and seasons. Ecology of Freshwater Fish 21:363 - 374 Kempinger, J.J. 1988. Spawning and early life history of Lake sturgeon in th e Lake Winnebago system, Wisconsin. American Fisheries Society Symposium 5:110 - 122 Kern, M.M., A.A. Nassar, J.C. Guzy, and M.E. Dorcas. 2013. Oviposition site selection by spotted salmanders ( Ambystoma maculatum ) in an isolated wetland. Journal of Herpeto logy 47(3):445 - 449 King, T.L., B.A. Lubinski, A.P. Spidle. 2001. Microsatellite DNA variation in Atlantic sturgeon ( Acipenser oxyrinchus ) and cross - species amplification in the Acipenseridae. Conservation Genetics 2:103 119 Lancaster, J., B.J. Downes, an d A. Arnold. 2010. Oviposition site selectivity of some stream - dwelling caddisflies. Hydrobiologia 652:165 - 178 Legendre, P. 1993. Spatial autocorrelation: trouble or new paradigm? Ecology 74(6):1659 - 1673 Lowerre - Barbieri, S.K., K. Ganias, F. Saborido - Rey , H Murua, and J.R. Hunter. 2011. Reproductive timing in marine fishes: variability, temporal scales, and methods. Marine and Coastal Fisheries: Dynamics, Management, and Ecosystem Science 3(1):71 - 91 139 Martell, D.J., J.D. Kieffer, E.A. Trippel EA. 2005. Eff ects of temperature during early life history on embryonic and larval development and growth in haddock. Jour Fish Biol 66:1558 - 1575 Martell, D.J. and J.D. Kieffer. 2007. Persistent effects of incubation temperature on muscle development in larval haddock (Melanogrammus aeglefinus L.) 210:1170 - 1182 McAdam, S.O. 2011. Effects of substrate condition on habitat use and survival by white sturgeon (Acipenser transmontanus) larvae and potential implications for recruitment. Canadian Journal of Fisheries and Aqu atic Sciences 68:812 - 822 McQuown, E.C., B.L. Sloss, R.J. Sheehan, J. Rodzen, G.J. Tranah, and B. May. 2000. Microsatellite analysis of genetic variation in sturgeon: new primer sequences for Scaphirhynchus and Acipenser. Transactions of the American Fishe ries Society 129:1380 1388 McQuown, E.C., G.A.E. Gall, and B. May. 2002. Characterization and inheritance of six microsatellite loci in lake sturgeon. Transactions of the American Fisheries Society 131:299 307 Merten, E.C., Z.R. Snobl, and T.A. Wellnitz. 2014. Microhabitat influences on stream insect emergence. Aquatic Science 76:165 - 172 Meyer, K. and W.G. Hill. 1997. Estimation of genetic and phenotypic covariance for longitudinal data by restricted maximum likelihood. Livestock Production Science 47:18 5 - 200 Monaghan, P. 2008. Early growth conditions, phenotypic development and environmental change. Philosophical Transactions of the Royal Society B 363:1635 - 1645 Pledger, S., E.A. Baker, and K.T. Scribner. 2013. Breeding return times and abundance in ca pture - recapture models. Biometrics 69(4):991 - 1001 Räsänen,K. and L.E.B. Kruuk. 2007. Maternal effects and evolution at ecological time - scales. Functional Ecology 21:408 - 421 Refsnider, J.M. and F.J. Janzen. 2010. Putting eggs in one basket: ecological and evolutionary hypotheses for variation in oviposition - site choice. Annual Review of Ecology, Evolution, and Systematics 41:39 - 57 Reich, P. and B. J. Downes. 2003. The distribution of aquatic invertebrate egg masses in relation to physical characteristics of oviposition sites at two Victorian upland streams. Freshwater Biology 48:1497 - 1513 140 Ribeiro Jr., P.J. and P.J. Diggle. 2001. geoR: a package for geostatistical analysis. R - News 1(2):ISSN 1609 - 3631 Rudolf, V.H.W. and M.O. Rödel. 2005. Oviposition site selection in a complex and variable environment: the role of habitat quality and conspecific cues. Oecologia 142:316 - 325 Schaeffer, L.R. 2004. Application of random regression models in animal breeding. Livestock Production Science 86:35 - 45 Silberbush, A . and L. Blaustein. 2011. Mosquito females quantify the risk of predation to their progeny when selecting an oviposition site. Functional Ecology 25:1091 - 1095 Skaug, H., Fournier, D. & Nielsen, A. 2006. glmmADMB: generalized linear mixed models using AD M odel Builder. R package version 0.3. Smith, K.M. and D.K. King. 2005. Dynamics and extent of larval lake sturgeon Acipenser fulvescens drift in the Upper Black River, Michigan. Journal of Applied Ichthyology 21:161 - 168 Thiem, J.D., D. Hatin, P. Dumont, G . Van Der Kraak, and S.J. Cooke. 2013. Biology of lake sturgeon (Acipenser fulvescens) spawning below a dam on the Richelieu River, Quebec: behavior, egg deposition, and endocrinology. Torgersen, C.E., D.M. Price, H.W. Li, and B.A. McIntosh. 1999. Multisc ale thermal refugia and stream habitat associations of Chinook salmon in northeastern Oregon. Ecological Applications 9(1):301 - 319 Van Liefferinge, C., P. Seeuws, P. Meire and R.F. Verheyen. 2005. Microhabitat use and preferences of the endangered Cottus gobio in the River Voer, Belgium. Journal of Fish Biology 67: 897 909. Volkman, E.T., K.L. Pangle, D.A. Rajchel, and T.M. Suton. 2004. Hatchery performance attributes of juvenile lake sturgeon fed two natural food types. North American Journal of Aquacult ure 66(2):105 - 112 Welsh, A.B., M. Blumberg, and B. May. 2003. Identification of microsatellite loci in lake sturgeon, Acipenser fulvescens , and their variability in green sturgeon, A. medirostris . Molecular Ecology Notes 3:47 55 Wentworth, C. K. 1922. A scale of grade and class terms for clastic sediments. Journal of Geology 30:377 392 141 Wilson, A.J., L.E.B. Kruuk, and D.W. Coltman. 2005. Ontogenetic patterns in heritable variation for body size: using random regression models in a wild ungulate population . The American Naturalist 166(6):177 - 192 142 CONCLUSION Res earch res ults from this dissertation demonstrate the importance of environmental variability on adult, egg, and larval riverine fishes such as lake sturgeon . The work on adult spawning behavior in Chapter 1 indicates that in order to develop a self - sustaining population with yearly stocking efforts, the maintenance of adequate abiotic and demographic cues are essential. Temperature and discharge are vital cue s used by adults for determining the timing and location of spawning, and subtle increases in both cues due to spring runoff or the release of water from the hydroelectric dam greatly influence adult spawning behavior. Therefore, great effort should be mad e to maintain adequate flows and temperatures that represent a run - of - the - river scenario during the spawning and post - spawning months to allow greater spawning opportunities and provide critical habitat for developing embryos. In Chapters 2 - trated that the environmental conditions experienced while in the egg largely affects the growth, development, and behavior of larval sturgeon. Therefore, understanding how conditions affect offspring traits within the population requires monitoring across all early life stages and not just post - hatch or during the juvenile stage in order to determine the factors governing recruitment in this wild population. Additionally, this research shows that h igh discharge levels and high water temperatures result in a greater range in phenotypic variation during these early life stages indicating a greater potential for differential survival among diffe rent family groups. M aintenance of adequate conditions in the wake of climate change and through 143 the dam operation is essential to maintain the genetic diversity of this closed lake sturgeon population. Several research subjects within this dissertation can be built upon to understand the role of environmental variability and genetic (family) effects on riverine fish be havior and development. One research subject includes continual monitoring of spawning females to determine whether the probability of atresia increases as climates continue to change. Understanding factors governing individual timing of spawning has been limited, but is a growing field given the importance of numerating spawning population sizes for conservation efforts. Additional research includes examining the influ ence of size differences and swimming abilities on juven ile sturgeon survival. My researc h has shown that there are demonstrable size differences among individuals largely du e to additive genetic effects; h owever, additional factors within a captive setting and in the river including food availability can largely influence offspring growth and survival, particularly during the first over - winter period.