Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Thermal behaviour, survival, and reproductive success of adult Gates Creek sockeye salmon (Oncorhynchus… Minke-Martin, Vanessa 2016

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

Item Metadata

Download

Media
24-ubc_2016_september_minkemartin_vanessa.pdf [ 45.71MB ]
Metadata
JSON: 24-1.0306922.json
JSON-LD: 24-1.0306922-ld.json
RDF/XML (Pretty): 24-1.0306922-rdf.xml
RDF/JSON: 24-1.0306922-rdf.json
Turtle: 24-1.0306922-turtle.txt
N-Triples: 24-1.0306922-rdf-ntriples.txt
Original Record: 24-1.0306922-source.json
Full Text
24-1.0306922-fulltext.txt
Citation
24-1.0306922.ris

Full Text

THERMAL BEHAVIOUR, SURVIVAL, AND REPRODUCTIVE SUCCESS OF ADULT GATES CREEK SOCKEYE SALMON (ONCORHYNCHUS NERKA)  by VANESSA MINKE-MARTIN  B.K.I. (Honours), The University of Waterloo, 2012   A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF  THE REQUIREMENTS FOR THE DEGREE OF  MASTER OF SCIENCE in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Forestry)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  July 2016  © Vanessa Minke-Martin, 2016   ii Abstract Water temperature affects every aspect of the physiology of ectothermic fishes. Heterothermic stenotherms, like Pacific salmon, use behavioural thermoregulation to swim into more optimal water temperatures and out of less optimal ones. The range of preferred temperatures can coincide with the thermal optima of important physiological processes. During the reproductive migration, adult Pacific salmon partition limited endogenous energy and aerobic scope for activity to multiple activities, including swimming, recovery from physiological stress, maturation, and immune function. Although much is known about the effect of temperature on aspects of salmonid physiology from laboratory experiments, how free-swimming fish use thermal habitats to manipulate physiological processes is not well understood. Radio telemetry, archival temperature loggers, blood biopsy, and spawning assessments were used to evaluate the relative importance of physiology and migratory experience to thermal behaviour and metrics of fitness in the Gates Creek population of sockeye salmon in the Fraser River, British Columbia. A framework was developed, using relationships between temperature and routine oxygen consumption and aerobic scope for activity, to determine whether the thermal experience of tagged fish reflected minimization of energetic cost or maximization of aerobic scope. Nearly all male and female fish preferred temperatures several degrees below the optimum for aerobic scope (ToptAS; i.e. 16.4 ºC), but two groups of fish benefitted from use of temperatures near ToptAS. Later migrants used these temperatures, where the cost of transport is minimized, to swim quickly through one lake, and female sockeye salmon that spent a greater  iii proportion of the migration within a ToptAS window (13.4-19.5 ºC) lived longer on spawning grounds and had a lower probability of egg retention at death. In contrast, migrants that spent longer in the lakes occupied the cool water temperatures that I hypothesized would limit energy expenditure. Temperature preference was also related to the flow dynamics and water temperature that fish experienced during passage at the Seton Dam, which suggests that migrants use thermal habitats for recovery from anaerobic swimming. Future research can use existing relationships between temperature and physiological processes for greater insight into the ecological importance of fish temperature selection in thermally heterogeneous environments.   iv Preface This research was conducted as a component of a multi-year monitoring project on the effect of BC Hydro operations at Seton Dam and the Seton Generating Station on the migration behaviour and success of Pacific salmon that spawn in the Seton-Anderson watershed, near Lillooet, British Columbia. I held primary responsibility for the research design and experimental protocols, collection and analysis of data, and preparation of manuscripts. During the process, I received considerable logistical support from my colleagues, Matthew T. Casselman and Collin T. Middleton, and advice from my supervisor, Dr. Scott G. Hinch, supervisory committee member, David A. Patterson, and colleagues, Dr. Erika J. Eliason and Dr. Anthony P. Farrell. Individuals who made essential contributions to the development, experimentation, or manuscript preparation are listed as coauthors on manuscripts that will be submitted for publication. All capture, tagging, and handling procedures were approved by the University of British Columbia Animal Care Committee (certificate #A11-0125) and conducted in accordance with guidelines established by the Canadian Council on Animal Care.  Chapter 2: Effects of physiology and migration experience on survival, migration behaviour, and reproduction in female adult sockeye salmon in a regulated system Authors: Vanessa Minke-Martin, Scott G. Hinch, Douglas C. Braun, Nicholas J. Burnett, Matthew T. Casselman, Erika J. Eliason, Collin T. Middleton Acceptance: To be submitted for publication in July, 2016  v Comments: This study was conducted and written by VMM, under the supervision and guidance of SGH, with assistance from CTM, MTC, DCB, and NJB in conducting the fieldwork and analyses, and preparing the manuscript. EJE assisted with study development and manuscript preparation.  Chapter 3: Adult sockeye salmon temperature selection in thermally stratified lakes is not optimized for either energy conservation or aerobic capacity Authors: Vanessa Minke-Martin, Scott G. Hinch, Collin T. Middleton, Erika J. Eliason, Anthony P. Farrell Acceptance: To be submitted for publication in July, 2016 Comments: Data collection for this study was conducted by VMM and CTM, under the supervision and guidance of SGH. VMM analyzed the data and wrote the manuscript, with guidance and contributions from EJE and APF.     vi Table of Contents Abstract ............................................................................................................................... ii	Preface ................................................................................................................................ iv	Table of Contents ............................................................................................................... vi	List of Tables ..................................................................................................................... ix	List of Figures ..................................................................................................................... x	Acknowledgements ........................................................................................................... xv	Chapter 1: Introduction ....................................................................................................... 1	1.1 Fish physiology ......................................................................................................... 2	1.2 Swimming performance and recovery ...................................................................... 3	1.3 Maturation and reproduction ..................................................................................... 4	1.4 The sockeye salmon reproductive migration ............................................................ 6	1.5 Thesis overview and research objectives .................................................................. 8	Chapter 2: Effects of physiology and migration experience on survival, migration behaviour, and reproduction in female adult sockeye salmon in a regulated system ....... 11	2.1 Introduction ............................................................................................................. 11	2.2 Methods ................................................................................................................... 16	2.2.1 Fish collection and tagging ............................................................................... 16	2.2.2 Dam passage conditions ................................................................................... 20	2.2.3 Plasma physiology ............................................................................................ 21	2.2.4 Reproductive longevity and spawning assessments ......................................... 22	2.2.5 Temperature logging data and aerobic scope for activity ................................ 22	 vii 2.2.6 Statistical approach ........................................................................................... 23	2.3 Results ..................................................................................................................... 25	2.3.1 Receiver efficiency ........................................................................................... 26	2.3.2 Survival to spawning grounds (Model 1) ......................................................... 27	2.3.3 Migration duration (Model 2) ........................................................................... 27	2.3.4 Temperature experience in Seton and Anderson lakes .................................... 28	2.3.5 Reproductive longevity (Model 3) ................................................................... 28	2.3.6 Egg retention (Model 4) ................................................................................... 29	2.4 Discussion ............................................................................................................... 30	Chapter 3: Adult sockeye salmon temperature selection in thermally stratified lakes is not optimized for either energy conservation or aerobic capacity .......................................... 51	3.1 Introduction ............................................................................................................. 51	3.2 Methods ................................................................................................................... 55	3.2.1 Study System .................................................................................................... 55	3.2.2 Fish collection & tagging ................................................................................. 56	3.2.3 Fish temperature data ....................................................................................... 58	3.2.4 Physiological framework .................................................................................. 59	3.2.5 Environmental conditions ................................................................................. 60	3.2.6 Statistical approach ........................................................................................... 62	3.3 Results ..................................................................................................................... 64	3.3.1 Receiver efficiency ........................................................................................... 64	3.3.2 Dam passage ..................................................................................................... 65	3.3.3 Seton Lake ........................................................................................................ 65	 viii 3.3.4 Portage Creek ................................................................................................... 67!3.3.5 Anderson Lake ................................................................................................. 67!3.4 Discussion ............................................................................................................... 68!Chapter 4: Conclusion ....................................................................................................... 92!4.1 Metrics for evaluating temperature experience ....................................................... 93!4.2 Trade-offs between energy conservation and aerobic scope for activity ................ 94!4.3 Lethal and sub-lethal effects of stressors ............................................................ 96!4.4 Limitations to work and future research directions ............................................. 97!4.5 Management implications ................................................................................... 98!References ......................................................................................................................... 99!Appendix ......................................................................................................................... 116!      ix List of Tables Table 2.1 – Candidate model sets (models with cumulative weights ≥ 0.95) from multivariate statistical models of female sockeye migration survival and duration to spawning grounds, reproductive longevity on spawning grounds, and egg retention at death. .................................................................................................. 42	Table 2.2 – Comparison of tagging date, fork length (FL), plasma lactate, glucose, and estradiol concentrations, maximum Seton Dam temperature (Tdam) and discharge (sample sizes shown for ‘low’ and ‘high’ categories), proportion of time in the ToptAS window (i.e. 13.4-19.5°C), and arrival date at Gates Creek for female sockeye salmon tracked through Seton and Anderson lakes. The range (for tagging and arrival dates) or mean ± SE (for other variables) are shown for the fish included in statistical models of (1) survival to spawning grounds, (2) migration duration, (3) reproductive longevity, and (4) egg retention, along with the sample size (n). ............................................................................................... 44	Table 2.3 – Comparison of generalized linear models describing egg retention in female sockeye salmon (n=38), with Seton Dam discharge (categorical: low [0] or high [1]) and arrival date as single explanatory variables. ............................................ 50	Table 3.1 – Confidence model sets (models with cumulative Akaike weights ≥ 0.95) from multimodal inference of variables influencing adult sockeye salmon temperature mode in (1) Seton Lake (SL) and (2) Anderson Lake (AL), and global models of proportion of time at temperatures <10°C in (3) SL and (4) AL. ........ 82	Table 3.2 – Tagging date, GSE (MJ kg-1), migration duration (days), temperature mode (°C; i.e. most common temperature recorded by an individual fish), and proportion of time at temperatures <10°C (i.e. percentage of all temperature records from the respective lake) of Gates Creek sockeye salmon tagged in 2013 (n=27) and 2014 (n=87). Data are reported as mean ± standard error (range). .... 84	   x List of Figures Figure 2.1 – The Seton-Anderson watershed and its location in British Columbia, Canada (inset); Gates Creek and the artificial spawning channel (A); and the Seton Dam and fishway (B). R = radio receiver, P = pass-through PIT antenna, F = fish collection fence, D = Water Survey of Canada discharge gauge, W = spawning channel weir, white star = release site (2013 + 2014), filled star = release site (2013 only). Detections on telemetry receivers in A were grouped as ‘Gates Creek’ and in B as ‘Seton Dam’ for analysis. ....................................................... 40	Figure 2.2 – Fishway temperature (A) and Seton Dam discharge (B) during period when females were detected in the dam tailrace, August 9-September 5, 2013 (grey lines) and August 6-September 9, 2014 (black lines). ‘High’ discharge (i.e. altered flow conditions) occurred from August 8-19, 2014, and ‘low’ discharge (i.e. standard flow conditions) from August 6-7 and August 20-September 9, 2014. . 41	Figure 2.3 – Model-averaged standardized coefficients (mean=0, standard deviation=2) with 95% confidence intervals (CI) from models describing (A) survival, (B) migration duration, (C) reproductive longevity, and (D) egg retention. Filled circles indicate coefficients with CI that do not cross zero. ................................. 45	Figure 2.4 – Scatterplot of plasma glucose concentration at tagging vs. survival for females tagged in 2013 (grey triangles) and 2014 (black circles). Model 1 predictions shown for females that experienced low (solid line) and high (dashed line) discharge at the Seton Dam in 2013 (grey line) and 2014 (black line). High discharge (i.e. altered flow conditions) did not occur in 2013. ............................. 46	Figure 2.5 – Scatterplot of tagging date vs. migration duration for females tagged in 2013 (grey triangles) and 2014 (black circles). Model 2 predictions shown for females that experienced low (solid line) and high (dashed line) discharge at the Seton Dam in 2013 (grey line) and 2014 (black line). High discharge (i.e. altered flow conditions) did not occur in 2013. ........................................................................ 47	 xi Figure 2.6 – The thermal experience of two female sockeye salmon between Seton Dam and Gates Creek. These fish spent the highest (80%, left) and lowest (8%, right) proportion of time in the ToptAS window (13.4-19.5 °C; indicated by shaded region), of the females with recovered archival temperature loggers (n=38), during migration through Seton and Anderson lakes. Passage through Portage Creek is indicated by a dashed vertical line. The fish were tagged on Sept 6, 2014 (left) and Aug 17, 2014 (right), and one spawned successfully (left), while the other retained eggs at death (right). ...................................................................... 48	Figure 2.7 – Scatterplots of reproductive longevity vs. (A) arrival date and (B) proportion of time in ToptAS window (n=39) and egg retention vs. (C) arrival date and (D) proportion of time in ToptAS window (n=38) for female sockeye salmon tagged in 2014. Model predictions are shown for low (solid line) and high (dashed line) discharge at the Seton Dam in (A), (B), and (D). Boxplots (E) show reproductive longevity differed significantly for females that did (n=12) and did not (n=26) retain eggs (two-sample t-test: t31=-5.25, p<0.0001), with the mean and range (whiskers) shown. ................................................................................................. 49	Figure 3.1 – The Seton-Anderson watershed and its location in British Columbia, Canada (inset); Gates Creek and the artificial spawning channel (A); and the Seton Dam and fishway (B). The location of the Bridge River inflow to the northwest end of Seton Lake is indicated. R = radio receiver, P = pass-through PIT antenna, F = fish collection fence, D = Water Survey of Canada discharge gauge, W = spawning channel weir, white star = release site (2013 and 2014), filled star = release site (2013 only). Detections on telemetry receivers in A were grouped as ‘Gates Creek’ and in B as ‘Seton Dam’ for analysis. ........................................... 78	Figure 3.2 – Frequency histograms of the temperatures experienced in Seton and Anderson lakes by three sockeye salmon: a 2013 female (A; GSE=7.04 MJ kg-1) and male (B; GSE=4.99 7.04 MJ kg-1), and a 2014 female (C; GSE=4.45 MJ kg-1). Temperature bins are 1°C and right-closed, and the temperature mode in each lake is identified with a vertical dashed line. ........................................................ 79	 xii Figure 3.3 – The relative metabolic benefit of temperature to aerobic capacity (solid black and dotted line) and the maintenance cost of living (solid grey line) for Gates Creek sockeye salmon. Aerobic capacity values were determined from the equation, aerobic scope=-20.9124+3.8926*t-0.1184*t2 (Eliason et al. 2011), and normalized as a percentage of maximum aerobic scope (11.08 mg O2 kg-1 min-1). The dotted portion of the line indicates where the curve has been extrapolated beyond experimental oxygen consumption data. The optimum temperature for aerobic scope (ToptAS) is 16.4 °C (dashed black line), and critical temperatures (Tcrit; AAS=0) are 6.8 °C and 26.1 °C (zeros of the aerobic scope curve). Routine oxygen consumption (MO2routine) values were determined from the equation, MO2routine=1.7464*exp(0.0448*t) (Eliason et al. 2011), normalized as a percentage of MO2routine at 26.1 °C (5.62 mg O2 kg-1 min-1), and subtracted from 1, to generate the maintenance cost of living curve. At temperatures below the intersection of the two curves (i.e. 10 °C; dashed grey line), the relative benefit to maintenance costs exceeds the relative benefit to aerobic capacity. ..................... 80	Figure 3.4 – Relationships between (A) tagging date and gross somatic energy (GSE; r2=-0.54, p<0.0001), (B) tagging date and time in Seton Lake (days; r2=-0.51, p<0.0001), (C) tagging date and time in Anderson Lake (days; r2=-0.47, p<0.0001), (D) GSE and time in Seton Lake (r2=0.38, p<0.001), and (E) GSE and time in Anderson Lake (r2=0.35, p<0.001) for fish tagged in 2013 (grey triangles) and 2014 (black circles). ....................................................................................... 85	Figure 3.5 – The highest temperature (A) experienced by individual fish during Seton Dam passage in 2013 (grey triangles) and 2014 (black circles); hourly Seton Dam discharge (B) during period when fish were detected in the dam tailrace in 2013 (grey line) and 2014 (black line), with horizontal dashed line indicating period of altered Seton Dam flow conditions (i.e. ‘high’ discharge), Aug 9-19, 2014; the highest temperature (C) experienced by individual fish in Portage Creek in 2013 (grey circles) and 2014 (black circles). ................................................................. 86	 xiii Figure 3.6 – Distributions of the number of days spent in Seton and Anderson lakes by fish tagged in 2013 (grey bars; n=27) and 2014 (white bars; n=87). Time in Seton Lake differed significantly by year (Kolmogorov-Smirnov two-sided test, D=0.354, p-value=0.012), but time in Anderson Lake did not. ............................ 87	Figure 3.7 – Distributions of the temperature mode (i.e. most common temperature experienced) of individual tagged sockeye salmon in (A) Seton Lake and (B) Anderson Lake in 2013 and 2014. Temperature bins are 0.5 °C and right-closed, and the y-axis indicates the number of fish that had a temperature mode in a given bin. The distributions of preferred temperature differed significantly between lakes (two-sample Kolmogorov-Smirnov test, D=0.421, p<0.0001). ................... 88	Figure 3.8 – Scatterplot of temperature mode vs. (A) the highest temperature experienced at Seton Dam, (B) time in Seton Lake, and (C) time in Anderson Lake for sockeye salmon tagged in 2013 (grey triangles) and 2014 (black circles). Model 1 predictions shown for fish that experienced low (solid line) and high (dashed line) discharge at the Seton Dam (A, B); predictions did not differ by year. Model 2 predictions show for fish tagged in 2013 (grey line) and 2014 (black line; C); predictions did not differ by time in lake. ............................................................. 89	Figure 3.9 – Boxplots compare the proportion of time in Seton Lake (A, B) and Anderson Lake (D, E) that male and female sockeye salmon spent at temperatures below 10 °C. Scatterplots show percent time below 10 °C vs. number of days in Seton Lake (C) and Anderson Lake (F). Model 3 found a significant positive relationship between time in Seton Lake and proportion of time spent at temperatures below 10 °C (C). Model 4 found that females spent significantly more time below 10 °C in Anderson Lake in 2014 (E). ....................................... 90	Figure 3.10 – Standardized coefficients (mean=0, standard deviation=2) with 95% confidence intervals (CI) for models describing temperature mode in (A) Seton Lake and (B) Anderson Lake and use of thermal habitat <10 °C in (C) Seton Lake and (D) Anderson Lake. (A) and (B) are model-averaged coefficients, (C) and (D)  xiv are from global models. Filled circles indicate coefficients with CI that do not cross zero. ............................................................................................................. 91!Figure A.1 – Data from archival temperature loggers recovered from 114 sockeye salmon. Data between time of last telemetry detection at Seton Dam and first detection at Gates Creek are shown. Each day is indicated by a short tick on the x-axis, and the beginning of each week is labelled with the date. The vertical dashed line is the time of entry into Portage Creek; before the line, the fish is in Seton Lake and after, the fish is in Anderson Lake. Year, sex, and fish number are indicated. ............................................................................................................. 116!Figure A.2 – Histograms of temperature data, from each of 114 sockeye salmon in Seton Lake and Anderson Lake, binned by 0.5ºC and right-closed. The vertical dashed line indicates the temperature mode (i.e. most common temperature experienced) in each lake. ........................................................................................................ 132!    xv Acknowledgements Over these two and a half years, I have had the opportunity to work with and learn from many incredible scientists. To my supervisor, Scott Hinch; my committee, David Patterson, John Richardson and Mike Donaldson; and mentors, Erika Eliason, Tom Sullivan, and Tony Farrell: thank you for your guidance, wisdom, and difficult questions. To the Hinch lab, especially Collin Middleton, Nich Burnett, Nathan Furey, Art Bass, Nolan Bett, Natalie Sopinka, Matt Casselman, Matt Drenner, Andrew Lotto, Amy Teffer, Doug Braun, and Eduardo Martins: thank you for keeping things exciting, even when we were moving 12 volt batteries and sifting through R code. You taught me so much. Many others supported my work, including Jayme Hills, Kendra Robinson, and Taylor Nettles from DFO Environmental Watch; Roxx Ledoux, Jess Hopkins, Wes Payne, Avaleen Adolph, Alison James, and Bonnie Adolph from St’át’imc Eco-Resources; BC Hydro; InStream Fisheries Research Inc.; N’Quatqua First Nation; Ocean Tracking Network; American Fisheries Society; and NSERC. My friends and family members provided me with love, encouragement, and affirmation—some from very far away—during this often challenging time. Sande Minke, Rick Martin, Kieran Joyes, Liz MacInnis, Sherri Martin, Dan Shura, Noel Farrand, Corrin Whiteway, Paisley Cozzarin, Claire Scullion, Kyrie Vala-Webb, Bryson McLachlan, Carmen Ballard, Rachel Kennedy, Bridget Ellacott, Alison Myers, Linda Carson, and Andrew Hunter: I am grateful to have each of you in my life. To Michael Barden and Samantha Skelly: thank you for helping me finish this journey and begin a much longer one. Finally, thank you to Pamela, Sally, and all of the other sockeye salmon that sustained my research in life and in death.     1 Chapter 1: Introduction Teleost fish are ectothermic, poikilothermic vertebrates, with body temperatures that vary widely with the surrounding water temperature. However, the capacity for physiological processes and whole-animal performance are greatest for each species within a preferred range of temperatures, where the biochemical reactions underlying metabolism are optimized (Brett 1956, Swan 1974). One of the strategies that fish use to maintain physiological function in thermally heterogeneous environments is behavioural thermoregulation; fish manipulate body temperature by swimming into more optimal thermal habitats, and out of less optimal ones.  Temporal and spatial patterns of temperature dictate the global distributions of fish species. The strong influence of temperature on ecological range has been demonstrated across fish taxa (e.g. bluegill sunfish [Lepomis macrochirus] and largemouth bass [Micropterus salmoides], Block et al. 1984; striped bass [Morone saxatifis], Coutant and Carroll 1980; Pacific bluefin tuna [Thunnus thynnus orientalis], Kitagawa et al. 2000; and lake trout [Salvelinus namaycush], Snucins and Gunn 1995). As global temperature regimes change, fish will increasingly be exposed to suboptimal temperatures in their historical habitats. Already, populations are exhibiting range shifts, where organisms that are tracking preferred temperatures begin to occupy new habitats or geographic areas (e.g. Rhône River grayling [Thymallus thymallus], Daufresne et al. 2004; brown trout [Salmo trutta], Hari et al. 2006). However, for species that are tied to historical habitats by high fidelity to natal rearing areas, like Pacific salmon  2 (Oncorhynchus spp.; Burgner 1991), the abilities to regulate thermal exposure and cope with suboptimal temperatures are important for survival. This thesis evaluates the ways that habitat use in a thermally heterogeneous environment reflects prior experience and relates to subsequent fitness outcomes for a semelparous migratory fish, sockeye salmon (Oncorhynchus nerka).  1.1 Fish physiology  Water temperature affects all aspects of fish physiology and performance (Fry 1947, Brett 1971). The optimum temperatures that maximize the rates of different physiological processes (e.g. feeding, digestion, growth, active metabolism, swimming performance) may differ (Brett 1971). The preferred temperature, which fish occupy in laboratory thermal gradients, is hypothesized to support important physiological functions and life-history processes; in natural environments, preferred temperature also reflects food and habitat availability and inter- and intraspecific interactions (Fry 1947). Optimal and tolerable temperatures vary markedly between species, populations, sexes, and life stages (Elliott 1975, Spigarelli et al. 1983, Elliott and Hurley 1997, Stitt et al. 2014), and patterns of thermal habitat use may be cyclical, occurring on daily or seasonal scales. The rate of routine metabolism—the necessary biochemical processes that sustain life—increases exponentially with temperature, reducing capacity for additional activities, including swimming, feeding, immune function, and reproduction (Fry 1947, Eliason and Farrell 2016). Conversely, respiration rate is greatly reduced at low temperatures, limiting oxygen supply (Brett 1956). Aerobic scope for activity (the difference in routine and maximum oxygen consumption; Fry 1947) is a measure of  3 physiological capacity for oxygen-dependent processes. Aerobic scope is maximized at an optimum temperature (ToptAS) and declines as temperature increases and decreases from ToptAS (Fry 1947). In general, ectothermic organisms are thought to prefer temperatures close to, but below ToptAS, as super-optimal temperature exposure has greater negative consequences on many physiological processes compared to sub-optimal temperature exposure (Martin and Huey 2008, Asbury and Angilletta 2010).  The critical temperatures (Tcrit) are high and low temperatures where aerobic scope is zero, activity relies on anaerobic metabolism, and survival is time-limited (Pörtner 2001, Pörtner and Knust 2007, Pörtner and Farrell 2008). Exposure to temperatures approaching Tcrit causes a stress response, a series of physiological and behavioural alterations, which re-establish homeostasis. The generalized stress response is characterized by changes in blood hormones and chemistry—including increases in catecholamine, cortisol, and metabolite concentrations and decreases in blood pH and oxygen—and to cardiorespiratory, reproductive, immune function (Wedemeyer et al. 1990, Bonga 1997, Barton 2002). Exposure to long lasting or multiple stressors causes a cumulative physiological response that may be lethal, if the fish is unable to regain homeostasis (Barton 2002). 1.2 Swimming performance and recovery During sustained periods of swimming, such as when foraging, moving between habitats, or maintaining body position, fish locomotion is fuelled by aerobic metabolism. Fish employ a swimming speed that minimizes the metabolic cost of transporting a given mass per unit time (i.e. cost of transport, COT; Weihs 1974, Webb 1995). As current velocity increases, upriver migrating fish, like Pacific salmon, must increase swimming  4 speed to make progress (Hinch and Rand 2000). Aerobic metabolism can support swimming speeds of up to 70-80% of maximum, but as swimming speed increases, anaerobic metabolism begins to contribute (Geist et al. 2003). Anaerobic (‘burst’) swimming provides rapid, but time-limited, movement, and is used for escaping predators, ambushing prey, and overcoming high flows (Webb 1995). However, anaerobic metabolism depletes cellular and blood oxygen stores and high-energy phosphates and causes the accumulation of lactate and H+ ions in tissues (Wood 1991), which can lead to lethal metabolic acidosis (Black 1958, Wood et al. 1983). During recovery, fish exert energy and consume large amounts of oxygen (i.e. excess post-exercise oxygen consumption [EPOC]; Gaesser and Brooks 1984, Wood 1991, Scarabello et al. 1992) to restore homeostasis. As water temperature approaches Tcrit and routine metabolic oxygen demand increases, anaerobic metabolism contributes an increasing proportion of the energy that fuels swimming, and fish become exhausted more rapidly. Because more oxygen is required for recovery, COT is higher at high temperatures. In addition, reduced aerobic scope slows recovery as temperature increases, which means that fish need more time to restore homeostasis. Nonetheless, alternating burst swimming with periods of coasting (when recovery can occur) reduces the net cost of transport of swimming at high velocities, by decreasing the amount of time that fish spend in high flow or high temperature environments (Weihs 1974, Martin et al. 2015). 1.3 Maturation and reproduction Environmental parameters, particularly temperature and photoperiod, are important proximal cues for maturation in temperate fish species (e.g. in salmonids, Bromage et al. 2001, Davies and Bromage 2002, Taranger et al. 2003). The rate of  5 vitellogenesis (i.e. oocyte development) in females increases with temperature, provided that the temperature is within the species’ preferred thermal range (Pankhurst and King 2010). However, it is unknown whether the optimum temperature for reproduction coincides with (Pörtner and Knust 2007, Pörtner and Farrell 2008, Eliason and Farrell 2016), or differs from (Macdonald et al. 2000), a species’ ToptAS.  The physiological stress response is associated with decreased concentrations of reproductive hormones (Pankhurst and Van Der Kraak 2000, Schreck et al. 2001). High water temperature and exhaustive exercise are stressors that have been linked to declines in salmonid reproductive performance (reviewed in Fenkes et al. 2015). Delayed maturation is a common response to thermal stress (e.g. in Chinook salmon, Kinnison et al. 2001; and pink salmon, Jensen et al. 2006, Jeffries et al. 2012), and thermal stress in adults has an inter-generational effect, reducing embryo and fry fitness (Van der Kraak and Pankhurst 1997, King et al. 2003). Anaerobic exercise, in response to high water velocity, may compromise reproduction by depleting energy necessary for maturation (Rand et al. 2006, Nadeau et al. 2010), suppressing maturation processes (Palstra and van den Thillart 2010), or requiring extended recovery, which delays arrival on spawning grounds (Martin et al. 2015).     Reproductive success depends on synchronizing maturation and arrival on spawning grounds with conspecifics, gaining access to mates, combining gametes, and defending fertilized eggs from disturbance (Mehranvar et al. 2004). Some species have energetically costly secondary sexual characteristics, such as the dorsal hump and kype of male Pacific salmon, which are important for obtaining mates (Quinn 2005), and females  6 generally invest more energy into gonad development than males (e.g. 11-20% vs. 1-4% in sockeye salmon, Gilhousen 1980, Hendry and Berg 1999).  1.4 The sockeye salmon reproductive migration   Anadromous fishes utilize energy and habitats in both freshwater and marine environments to maximize growth and reproductive output (Dingle 1996). Juveniles are born and reared in freshwater before migrating downstream to access marine feeding areas. After growing to an adequate size for maturation, adults swim back to freshwater to reproduce, with some populations traveling thousands of kilometres. Many salmonids are anadromous, although species differ in duration and extent of marine migrations and use of freshwater spawning habitats; some species spend the entire life cycle in freshwater (Rounsefell 1958). Pacific salmon are generally semelparous, and adults return to home streams to participate in a single, coordinated spawning event, followed by death. The reproductive life stage is accomplished in days or weeks, and individuals that do not survive to reach spawning grounds, or die on spawning grounds prior to completing reproduction, have zero lifetime fitness. Adults cease feeding in freshwater, using endogenous energy accrued during marine feeding to migrate. Nearly all stored energy is consumed by swimming activity during migration (Hendry and Berg 1999), and energy conserving behaviours are necessary to ensure fish arrive on spawning grounds with adequate energy for reproduction. Exposure to environmental (e.g. high temperature, river flow) and anthropogenic (e.g. hydropower development) stressors during migration requires costly physiological responses that further deplete somatic energy and jeopardize survival.   7 Adult Pacific salmon demonstrate behavioural thermoregulation throughout the migration, avoiding stressful temperatures and using thermal refugia where available. Populations with longer migrations enter freshwater prior to peak summer river temperatures and hold in cool waters prior to spawning, and populations with shorter migrations enter the river after temperatures have peaked, just prior to spawning (Hodgson and Quinn 2002). Thermal refugia are limited in the lower Fraser River (Donaldson et al. 2009), but many populations of sockeye salmon have access to cool temperatures in lakes near spawning grounds, where migrants reared as juveniles (Newell and Quinn 2005). Thermal experiences of adult sockeye salmon in natal lakes are thought to reflect energy conservation, stress recovery, and maturation processes (Newell and Quinn 2005, Roscoe et al. 2010a), and also affect disease progression (Wagner et al. 2005, Mathes et al. 2010, Miller et al. 2014). Temperature-limited aspects of cardiorespiratory function are thought to dictate thermal tolerance and migratory ability of adult sockeye salmon populations (Eliason et al. 2013b). Individuals from interior populations with wider and higher optimal ranges for aerobic scope continue to swim aerobically at higher temperatures than those from coastal populations, which tend to have lower optimal temperature ranges for aerobic scope, and experience anaerobiosis and mortality under the same conditions (Eliason et al. 2011, 2013a). Although studies on the effects of temperature on sockeye have been increasing, spawners have been one of the least studied life-stages (Martins et al. 2012a). Reproductive longevity is dependent on date of arrival on spawning grounds and somatic energy and maturation state of female fish (McPhee and Quinn 1998, Morbey and Ydenberg 2003, Hruska et al. 2011), and females that live longer deposit more of their  8 eggs (Hruska et al. 2011, Burnett et al. In press). Pre-spawning mortality, when females die on spawning grounds before spawning successfully, has been observed in sockeye salmon from Bristol Bay, Alaska, certain Fraser River populations, and other Pacific salmon populations (Burgner 1991). Pre-spawning mortality is associated with high temperature exposure during river migration (Gilhousen 1990), while holding in natal lakes (Quinn et al. 2007), and in spawning streams (West and Mason 1987). Research is needed to link stressors during the migratory experience to reproductive consequences on spawning grounds (Hruska et al. 2010). Sockeye salmon is one of the most abundant of the Pacific salmon species, with a freshwater distribution ranging from the Sacramento River (California) to Kotzebue Sound (Alaska), and from the Kuril Islands to the Anadyr River, in Russia (Quinn 2005). Sockeye salmon is also the most commercially valuable species and is important culturally to Pacific First Nations. As a result, its spawning migration is one of the most studied of Pacific salmon species (Hinch et al. 2006). In Canada, the Fraser River (British Columbia) is the largest producer of wild sockeye salmon. 1.5 Thesis overview and research objectives In this thesis, I study the relationships between physiological state, migratory challenges, thermal experience, and measures of fitness in sockeye salmon. Gates Creek sockeye salmon provide a unique opportunity to study these relationships, for several reasons. First, quantitative relationships between temperature and swimming performance and oxygen consumption have been determined for this population (Lee 2003a, 2003b, Eliason et al. 2011). Second, the Seton Dam is both an effective location for intercepting wild fish for study and a known anaerobic stressor for migrants (Burnett et al. 2014a,  9 2014b). Third, Seton and Anderson lakes are thermally stratified lakes, where sockeye salmon thermal behaviour may reflect stress recovery, energy conservation, and maturation processes (Roscoe et al. 2010a). Lastly, the artificial spawning channel in Gates Creek is an excellent location to assess metrics of female reproductive success. I employed biotelemetry (blood biopsy, archival temperature loggers, and radio telemetry) and spawning assessments to measure fish physiological condition, migratory behaviour, survival, and reproductive output. I then used multivariate statistical models to determine how physiology, dam passage, and lake behaviour affected thermal experience and fitness outcomes for the sockeye salmon that I tagged.  I had three main objectives for my research, and here I outline how they were achieved across two data chapters. The first objective was to characterize the thermal experience of sockeye salmon that successfully migrated through two thermally stratified lakes to reach spawning grounds. Specifically, in Chapter 2, I determined the proportion of the migration that female fish spent within three degrees of the optimum temperature for aerobic scope, where fish have at least 90% of maximum scope for activity. In Chapter 3, I developed a framework for assessing trade-offs between temperatures that reduce maintenance costs and temperatures that provide maximum physiological capacity. I then determined the preferred temperature of male and female fish and the proportion of time that migrants spent thermal refuging (i.e. at very cool temperatures, where maintenance costs are low). The second objective was to compare the relative importance of physiology and migratory experience to fitness for female fish. To assess multiple aspects of fitness, I quantified survival to spawning grounds, reproductive longevity on spawning grounds, and egg retention after death; the results are reported in  10 Chapter 2. In Chapter 3, I addressed the third objective, which was to compare the relative importance of physiology and migration experience to the thermal preference and refuging behaviour of male and female fish in two thermally stratified lakes. In Chapter 4, I provide a synthesis of my findings, discuss limitations of my work, and suggest possible extensions of these results to future research on behavioural thermoregulation in adult Pacific salmon.    11 Chapter 2: Effects of physiology and migration experience on survival, migration behaviour, and reproduction in female adult sockeye salmon in a regulated system 2.1 Introduction The success of a reproductive migration depends on interactions between the physiological state of an individual and the environmental conditions encountered (Nathan et al. 2008).  Individuals that reach breeding grounds are possibly in better physiological condition, encounter more benign environments, and/or respond more effectively to migratory challenges than individuals that die en route. Reproductive success further depends on individuals synchronizing maturation with arrival on breeding grounds (Dingle 1996). Adult Pacific salmon (Oncorhynchus spp.) have a single opportunity to reproduce before death, and unsuccessful migrants do not contribute genetically to the population (Dingle 1980). The energetic demands and historic environmental conditions of the adult migration have shaped the morphology (Crossin et al. 2004b), physiology (Crossin et al. 2004b, Eliason et al. 2011, 2013b), and behaviour (Hodgson and Quinn 2002, Crozier et al. 2008, Kovach et al. 2013) of Pacific salmon. Despite these adaptations, some populations exhibit high levels of en route and on spawning ground mortality (e.g. >90% in Fraser River sockeye salmon, reviewed in Hinch et al. 2012; >90% in Willamette River Chinook salmon [O. tshawytscha], Keefer et al. 2010, Benda et al. 2015). Peak summer river temperatures have steadily increased over the past 20 years in certain locations, and adult migrants now encounter temperatures 1.8 °C warmer in the  12 Fraser River (Patterson et al. 2007, Farrell et al. 2008) and 2.5 °C warmer in the Columbia River (Quinn and Adams 1996, Crozier et al. 2008). Exposure to high water temperature underlies or exacerbates many causes of migration mortality in Pacific salmon, because temperature is one of the most important factors affecting physiology and performance in salmonids (Fry 1947, Brett 1971, Payne et al. 2015). Spawning migrations are fuelled exclusively by stored energy, most of which is consumed by standard metabolism, swimming activity, and gonad maturation in years of typical migration conditions (Brett 1995, Hinch et al. 2006). At high temperatures, energy reserves can become exhausted and fish perish en route, particularly in long distance migrants (e.g. sockeye salmon, Rand and Hinch 1998, Macdonald et al. 2000, Crossin et al. 2004b, Rand et al. 2006).   High water temperature can also cause mortality via cardiorespiratory collapse (Eliason et al. 2013a). Aerobic scope for activity, the difference between maximum and standard metabolic rates, is greatest at an optimal temperature (ToptAS; Fry 1947, Pörtner and Knust 2007). Standard metabolism consumes an increasing proportion of available oxygen as temperature increases, limiting capacity for other activities, including swimming and maturation. At the critical temperature (Tcrit), where aerobic scope is zero, swimming activity is fueled by anaerobic metabolism (Pörtner and Knust 2007, Pörtner and Farrell 2008). Though anaerobic swimming is essential to the adult migration (e.g. for passing rapids and evading predators), it is costly and cannot be sustained for long periods of time (Lee 2003a, Geist et al. 2003, Martin et al. 2015). The collapse of aerobic scope is thought to be responsible for poor survival in some sockeye salmon spawning  13 migrations (e.g. Naughton et al. 2005, Keefer et al. 2008, Farrell et al. 2008, Mathes et al. 2010, Martins et al. 2011).  Exposure to environmental stressors can be identified by changes in the concentration of blood compounds stimulated by the physiological stress response (Bonga 1997). Plasma glucose, an energy substrate, increases to fuel aerobic swimming (Bonga 1997), although chronic stress can cause low levels if glycogen stores are limited (Hoar et al. 1992). Anaerobic swimming produces plasma lactate, which can cause lethal blood acidosis at high levels (Black 1958). Endogenous energy and oxygen are required to catabolize lactate (Gleeson 1996), and resting to recover can delay migration (Martin et al. 2015). Elevated plasma glucose and lactate concentrations have been correlated with migration failure in adult Fraser River sockeye salmon (e.g. Cooke et al. 2006a, 2006b, Crossin et al. 2009).  Migration through regulated rivers creates unique challenges to survival, because fish could encounter both high temperatures and high flows in dam tailraces, as they attempt to locate and enter fishways (Keefer and Caudill 2015). Both environmental and behavioural factors are known to influence migration success of Pacific salmon in regulated systems (e.g. sockeye salmon, Naughton et al. 2005, Keefer et al. 2008, Burnett et al. 2014a). However, few studies have related individual physiology to survival (except see Pon et al. 2009, Roscoe et al. 2010b), and none have examined how environmental factors and physiological condition influence reproductive outcomes for successful migrants. Salmonids regulate temperature exposure in thermally complex environments by occupying temperatures within a preferred range. Although limited in some areas of the  14 freshwater migration (e.g. the lower Fraser River, Donaldson et al. 2009), thermal refugia, including cool water tributaries (e.g. of the Columbia River, Hodgson and Quinn 2002, Goniea et al. 2006, Keefer et al. 2009, Keefer and Caudill 2015) and the hypolimnion of lakes and reservoirs, allow Pacific salmon to mitigate the negative effects of thermal stress. Indeed, sockeye salmon that occupy cool water in lakes close to spawning grounds conserve energy, reduce stress, and increase likelihood of survival to spawning grounds (Newell and Quinn 2005, Farrell et al. 2008, Mathes et al. 2010). Behavioural thermoregulation may be particularly important to females, as they allocate considerable energy to gamete development in the final stages of maturation (Rand and Hinch 1998); there is evidence that more mature females with lower levels of somatic energy occupy the coolest temperatures in natal lakes (Roscoe et al. 2010a). Because water temperature affects maturation rate (Pankhurst and King 2010), it has been proposed that prolonged exposure to temperatures near ToptAS may allow migrating females to mature most efficiently (Eliason and Farrell 2016); however, this has not be examined. Females appear to be more vulnerable to migratory challenges than males. In multiple studies of migrating adult Fraser River sockeye salmon, mortality was nearly twice as high in females than males at high temperatures (Crossin et al. 2008, Jeffries et al. 2012, Martins et al. 2012b). Smaller ventricular mass (Sandblom et al. 2009), higher plasma cortisol and glucose (Sandblom et al. 2009), greater reliance on anaerobic swimming (Burnett et al. 2014a), and greater susceptibility to holding and handling stress (Gale et al. 2014) are possible contributing factors.   15 Even females that overcome migratory challenges to reach spawning grounds may exhibit delayed effects on fitness. High rates of egg retention have been observed in populations of Pacific salmon exposed to high temperatures well below the lethal threshold en route to, or on, spawning grounds [e.g. 20% of females, West and Mason 1987; 0-90%, Gilhousen 1990 (but see Macdonald et al. 2000); 23-44%, Quinn et al. 2007]. There are multiple behavioural and physiological mechanisms that may contribute. First, arrival may be delayed, due to time spent recovering from stress, decreasing time on spawning grounds (i.e. ‘reproductive longevity,’ Morbey and Ydenberg 2003) required for reproductive behaviours (Hruska et al. 2011). Second, energy needed for maturation or reproductive behaviours may have been diverted to standard metabolism or recovery (Kinnison et al. 2001, Martin et al. 2015, Fenkes et al. 2015). Third, thermal stress can inhibit synthesis of the sex steroid 17-β estradiol and delay vitellogenesis (Pankhurst and King 2010, Jeffries et al. 2012). Fourth, anaerobic swimming can also suppress reproductive hormones and oocyte development (Kubokawa et al. 2001, Palstra and van den Thillart 2010, Hayashida et al. 2012). The relationships between arrival date, reproductive longevity, competitive ability on spawning grounds [including the importance of morphological traits (e.g. fork length) and indicators of stress (e.g. plasma lactate)], and spawning success have been studied in Pacific salmon (e.g. McPhee and Quinn 1998, Dickerson et al. 2002, Hruska et al. 2010, 2011), but research is needed to link migratory stressors to reproductive consequences for individual females (Hruska et al. 2010). The purpose of this study was to compare the relative importance of physiological condition and migratory experience to female sockeye salmon fitness in the final stages  16 of the reproductive migration. Specifically, I assessed the effects of indices of physiological stress (plasma glucose and lactate concentrations) and maturity (plasma 17-β estradiol concentration), and migratory experience (water temperature and discharge during Seton Dam passage, lake thermal experience) on four measures of fitness: survival to spawning grounds, migration duration, reproductive longevity, and egg retention. Although reproductive longevity had a strong negative relationship with egg retention in two previous studies of female sockeye salmon (Hruska et al. 2011, Burnett et al. In press), I assessed them as separate fitness outcomes to determine whether underlying physiological mechanisms differ. Since physiological stress has a negative influence on survival and maturation, I predicted that female fish that exhibited lower indices of stress (i.e. lower plasma glucose and lactate concentrations) and higher maturation (i.e. lower plasma 17-β estradiol concentration), and that experienced less stressful dam passage conditions (i.e. lower water temperature and discharge) would have higher survival, faster migration, greater reproductive longevity, and lower egg retention. Additionally, I predicted that female fish that spent a greater proportion of the lake migration at temperatures near ToptAS would have lower reproductive longevity and lower egg retention.  2.2 Methods 2.2.1 Fish collection and tagging The Seton Dam, in Lillooet, British Columbia, is ~3.5 km upstream of the Seton River-Fraser River confluence and ~364 km upstream from the Fraser River mouth. After negotiating the vertical-slot fishway at Seton Dam, adult sockeye salmon swim ~50 km  17 through Seton Lake, Portage Creek, and Anderson Lake to reach spawning areas in Gates Creek. Approximately 800 metres upstream of the creek mouth, a weir directs fish to a gate, which can be opened to guide fish into the natural creek upstream of the weir, or closed to load fish into a creek-fed artificial spawning channel (Figure 2.1). Female sockeye salmon from the Gates Creek population were captured in the Seton River using a full-spanning picket fence and trap installed ~400 metres downstream of Seton Dam (Figure 2.1). The fence was closed for fish collection 11 hours per day throughout the duration of the sockeye migration and left open each night. Between Aug 16-Sept 2, 2013 and Aug 5-Sept 7, 2014, 109 and 95 female sockeye salmon were collected for tagging, respectively.  Fish were netted from the trap box and placed in a flow-through holding pen in the Seton River for up to two hours before tagging. Individual fish were held supine, unanesthetized, in a padded, V-shaped trough with river water flowing continuously over the gills, following Cooke et al. (2005). Females were identified by secondary sexual characteristics (i.e. smaller kype and adipose fin, larger vent vs. male fish). Somatic lipid concentration was estimated using a handheld energy probe following Crossin and Hinch (2005; FM 692 Fish Fatmeter, Distell Inc., West Lothian, Scotland, UK). Energy probe readings were used to distinguish Gates Creek sockeye salmon from other populations with similar migration timing that are traveling to spawning grounds farther up the Fraser River. Migration distance and somatic energy density are positively related in Fraser River sockeye salmon populations (Crossin et al. 2004b). Individuals sampled at the fish fence with relatively high levels of lipids have been confirmed as strays from other populations using DNA population assignment (Casselman et al. 2012; Bett and Hinch  18 2015), and were released without tagging. Fork length (FL) was measured (to nearest 5 mm) and a 3-mL blood sample taken from the caudal vein with a heparinized vacutainer (22 gauge syringe; BD Canada, Mississauga, ON, CA). Blood samples were centrifuged for 6 minutes to isolate plasma, which was transferred to three 0.8-mL vials and stored in liquid nitrogen before transfer to a -80 °C freezer prior to laboratory analysis. Fish were tagged with a 32-mm half-duplex (HDX) passive integrated transponder (PIT) tag (Oregon RFID, Portland, OR, USA) in the dorsal musculature, a 12” spaghetti tag (Floy Manufacturing, Seattle, WA, USA) attached posterior to the dorsal fin (for visual identification on spawning grounds), and a uniquely coded radio transmitter (Pices5 model, 3s burst rate, Sigma Eight Inc., Newmarket, ON, Canada) inserted gastrically with a plastic plunger. Archival temperature loggers (iButton Thermochron model DS1921Z or DS1922L, Maxim Integrated, San Jose, CA, USA) were glued to the non-antenna end of radio transmitters and waterproofed (Plasti Dip International, St. Louis Park, MN, USA), as in Donaldson et al. (2009). Loggers archived data at 15-minute intervals and had resolutions of 1/16 °C (manufacturer stated accuracy: ± 0.5 °C, range: -10 °C to +65 °C; model DS1922L) or 1/8 °C (manufacturer stated accuracy: ± 1 °C, range: -5 °C to +26 °C; model DS1921Z). An adipose tissue punch was taken and stored in ethanol (for DNA population identification; Beacham et al. 2005) and the fish was photographed and returned to the holding pen for recovery. The time to tag each fish was not recorded in 2013, but likely equivalent to 2014 (mean ± SE: 4.7 ± 0.1 minutes, range: 2.7-9.1 minutes, n=95), as the procedure did not differ between years. To achieve the objectives of a parallel study, groups of ten to twelve fish were transferred to a 1,000-litre aerated transport tank and driven ~20 minutes to release locations on the west (N 50.6621, W  19 121.9184) and east (N 50.6611, W 121.9139) bank of the Fraser River, ~2.5 km downstream of the Seton River-Fraser River confluence. In 2014, fish were released from the west bank only, as survival to the Seton Dam and dam passage did not differ by release location in 2013 (Casselman et al. 2013). All capture, tagging, sampling, and transport procedures were approved by the University of British Columbia Animal Care Committee, following Canadian Council on Animal Care guidelines.  Stationary radio receivers (Orion, Sigma Eight Inc. or SRX-400, Lotek Wireless Inc., Newmarket, ON, Canada) with 3-element Yagi antenna were placed at the release location on the west bank of the Fraser River, at the Seton Dam (one directed at the tailrace, one directed at the fishway entrance), and at the entrance to the spawning channel in Gates Creek (Figure 2.1). During installation, the Seton Dam receivers were tested to ensure non-overlapping detection ranges. Pass-through PIT antennas (Oregon RFID) were installed inside the entrance and exit basins of the Seton Dam fishway, and the read range (~0.5 m) was tested daily. Two PIT receivers were installed on spawning grounds and tested weekly: a three-antenna array at the spawning channel entrance and a pass-through PIT antenna 120 m upstream of the mouth of Gates Creek in 2014 (Figure 2.1). PIT antenna construction is described in Burnett et al. (2014a). First and last detections were identified for each fish at each receiver from the raw telemetry data after screening for false detections. At the Seton Dam and Gates Creek, first and last detections were obtained by pooling all radio and PIT receivers in each location to account for possible missed detections at a single receiver (e.g., caused by turbulence in the dam tailrace and a power outage in D’Arcy in 2014).   20 Only females that successfully passed Seton Dam were included in the analyses. The 69 tagged females that did not pass Seton Dam (2013: 43, 2014: 26) were assessed in a parallel study (C.T. Middleton, University of British Columbia, unpublished data). The last detection at Seton Dam was considered the beginning of the in-lake migration. Survival was indicated by telemetry detection on either radio or PIT receivers at Gates Creek. Migration duration, or the time spent in Seton and Anderson lakes, was calculated as the difference in days between the last detection at Seton Dam and the first detection at Gates Creek.  In 2013, there were no fisheries openings during the study period. In 2014, fisheries were open from July 28-Oct 12 (Cindy Samaha, Fisheries and Oceans Canada, pers. comm.), and it is possible that tagged fish could have been captured and removed from the study area. To encourage return of radio transmitters and attached temperature loggers, information posters, which described the study, research partners (University of British Columbia, St’át’imc Eco-Resources, and BC Hydro), reward and contact information, were displayed in local grocery stores, gas stations, post offices, and on outdoor bulletin boards. Four tags were returned in 2014, and these fish were omitted from analyses.  2.2.2 Dam passage conditions A temperature data logger (TidBit v2; Onset Computer Corporation Inc., Bourne, MA, USA) in the top pool of the fishway recorded hourly Seton River water temperature throughout the study period. The highest hourly temperature recorded between the first and last detections of each fish at Seton Dam was used to characterize the temperature conditions experienced during dam passage.   21 Under standard operating procedure, Seton Dam spills water from a siphon directly adjacent to the fishway entrance (details in Burnett et al. 2014a, 2014b). Small fluctuations in discharge (<1 m3 s-1) in the area around the fishway entrance can increase reliance on anaerobic swimming among female fish, which was found to be associated with high mortality following dam passage (Burnett et al. 2014a). In 2014, the flow conditions in the Seton Dam tailrace were altered for a parallel study (Casselman et al. 2015). From Aug 8-19, 2014, water was spilled from a siphon approximately 10 m from the fishway entrance, which increased total Seton Dam discharge by up to 25% (Figure 2.2). For the purpose of this study, Seton Dam discharge was categorized as ‘low’ (i.e. standard flow conditions; until Aug 7 and from Aug 20-Sept 9, 2014) or ‘high’ (i.e. altered flow conditions; from Aug 8-19, 2014), because both discharge and tailrace flow dynamics differed substantially between the two flow conditions (Casselman et al. 2015). Individual fish were assigned to a discharge category (‘low’ or ‘high’) based on the date of first detection at Seton Dam. In both years of the study, a Water Survey of Canada gauge in the Seton River measured the total discharge from the Seton Dam (Figure 2.1). 2.2.3 Plasma physiology Glucose and lactate concentrations were measured from plasma samples in 2013 and 2014, using assays outlined in Roscoe and Hinch (2007). 17-β estradiol and testosterone (data not presented here) concentrations were measured from blood plasma samples, according to procedures outlined in Roscoe et al. (2010a). Relative concentrations of 17-β estradiol and testosterone were used to verify visual sex assignments made at tagging for all 2013 fish. Visual sex assignments were 92% accurate in 2013 (n=214 total fish tagged). Hormones were not used to verify sex in 2014, but four  22 visual sex assignments were changed based on recovered carcasses. Across years, females were more likely to be misidentified as males (two fish incorrectly called females, 20 fish incorrectly called males). All fish ultimately identified as female were included in analyses.   2.2.4 Reproductive longevity and spawning assessments Female carcasses containing radio transmitters were recovered from the spawning channel and off the weir in Gates Creek and assessed for egg retention daily. Reproductive longevity was calculated as the number of days between the first detection on the radio receiver or PIT antenna at the spawning channel (the arrival date) and the date each carcass was removed from the channel or creek. While it is not possible to assess male spawning success from carcasses (as males can retain ~50% of gonad size after spawning), spawning females typically expel most to all of their eggs (Brett 1995; D.A. Patterson, Fisheries and Oceans Canada, pers. comm.). Females were categorized as ‘retained eggs’ (i.e. unsuccessful spawner) if found with tight skeins or more than approximately 500 eggs in the body cavity (see example in Figure 2.3A). If all eggs were expelled, or <500 eggs remained in the body cavity, females were categorized as ‘did not retain eggs’ (i.e. successful spawner; Figure 2.3B), following Lingard et al. (2014). As very few tagged females were recovered in the spawning channel in 2013 (n=6, all ‘retained eggs’), only 2014 data were used in egg retention analyses.  2.2.5 Temperature logging data and aerobic scope for activity The maximum aerobic scope (11.08 mg O2 kg-1 min-1) and optimum temperature (i.e. ToptAS; 16.4 °C) for Gates Creek sockeye were previously derived from fish at rest  23 and swum maximally in a swim tunnel respirometer at a range of temperatures (Lee 2003a, Eliason et al. 2011). Analysis of survival rates across Fraser River sockeye populations led Eliason et al. (2011) to suggest that fish require temperatures within ~3 °C of ToptAS (90% of maximum aerobic scope) to complete migration. For Gates Creek sockeye, 90% of maximum corresponds to ≥9.98 mg O2 kg-1 min-1, or temperatures between 13.4 and 19.5 °C (hereafter, the ‘ToptAS window’). The temperatures recorded after the last detection at the Seton Dam and before the first detection at Gates Creek were selected, and the proportion of temperature readings within the ToptAS window was calculated for each fish. 2.2.6 Statistical approach The data were analyzed using four statistical models, with one for each of the response variables: survival, migration duration, reproductive longevity, and egg retention. Model 1 had the response variable of survival (1=detected at Gates Creek, 0=not detected) and was fitted with a generalized linear model (GLM; family: binomial; link: logit). Model 2 had the response variable of migration duration (d) and was fitted with a linear model. Both global models included the following explanatory variables: (1) year and (2) tagging date, to account for inter- and intra-annual differences in fish condition and migration experience; (3) FL (cm), to account for differences in fish size; (4) plasma lactate (mmol L-1) and (5) plasma glucose (mmol L-1), to test the effects of indices of physiological stress; (6) plasma 17-β estradiol (ng mL-1), to test the effect of maturity; and (7) Seton Dam discharge (low [0] or high [1]) and (8) Seton Dam water temperature (Tdam; °C), to test the effects of migratory stressors (Table 2.2).    24 Model 3 had the response variable of reproductive longevity (d) and was fitted with a linear model. Model 4 had the response variable of egg retention (1=retained eggs, 0=did not retain eggs) and was fitted with a GLM (family: binomial; link: logit). Both global models included the following explanatory variables: (1) FL; (2) plasma lactate; (3) plasma glucose; (4) plasma estradiol; (5) Seton Dam discharge; (6) Seton Dam water temperature (Tdam); (7) proportion of migration duration in ToptAS window, to test the effects of temperature experience in lakes; and (8) arrival date on spawning grounds, to account for intra-annual differences in fish condition and migration experience.  All variables were assessed for correlation and multicollinearity and excluded if r > |0.7| (Zuur et al. 2010), variance inflation factor > 4 (O’Brien 2007), or visual assessment of model residuals indicated that assumptions of normality, independence, or heteroscedasticity were violated (Zuur et al. 2010). Seton Dam discharge was collinear with arrival date (VIF > 4) in Model 4, so inclusion of one of the two variables was determined based on the individual GLM (response: egg retention, explanatory variable: Seton Dam discharge or arrival date) with the lower AICc score (Table 2.3). Reproductive longevity was not included in the egg retention model, as it does not provide a physiological explanation for spawning failure. Furthermore, inclusion in the egg retention model caused VIF > 6.  All statistical analyses were conducted with R software (version 3.1.3, R Core Team 2015). Candidate models were generated using the ‘MuMIn’ package from all combinations of variables in the global model and compared using AICc for small sample sizes (Barton 2012, Burnham and Anderson 2002). The candidate sets for Models 3 and 4 were limited to models with three or fewer explanatory variables to maintain a 10:1 ratio  25 of individuals to variables (Harrell 2001). To account for uncertainty, because top model weights were small (e.g. 0.27 for Model 1, Table 2.1), a confidence set of models with cumulative weights summed to ≥0.95 was averaged using the ‘natural average’ method to generate coefficient estimates and 95% confidence intervals for explanatory variables (Burnham and Anderson 2002, Grueber et al. 2011). To compare relative effect sizes within a given model, continuous variables were standardized, with a mean of 0 and standard deviation of 2 (Gelman 2008, Schielzeth 2010). Response variables were not transformed (Zuur et al. 2010). Fit of linear models was evaluated using adjusted-R2, and fit of GLMs was evaluated using pseudo-R2, following Hosmer and Lemeshow (1989). Two-sample t-tests were used to compare the reproductive longevity of females that did and did not retain eggs at death, and plasma estradiol concentrations between years. Statistical significance was assessed at alpha=0.05. Untransformed data are presented as mean ± SE, unless otherwise indicated.  2.3 Results Females that passed the Seton Dam were included in Model 1 (survival; n=135: 52 in 2013, 83 in 2014). Successful migrants to Gates Creek were included in Model 2 (migration duration; n=83: 23 in 2013, 60 in 2014). Females recovered in the spawning channel or off the Gates Creek weir in 2014 with thermal loggers (n=39) and spawning assessments (n=38) were included in Model 3 (reproductive longevity) and Model 4 (egg retention), respectively.  The variables included in each of the four models are summarized in Table 2.2 Plasma lactate, glucose, and estradiol concentrations, and FL did not show strong seasonal trends in either year, nor when years were pooled (all r<0.3). Mean  26 concentration of estradiol was higher in 2014 (7.32 ± 0.39 ng mL-1, range: 0.13-8.11 ng mL-1) than in 2013 (3.16 ± 0.36 ng mL-1, range: 0.28-16.08 ng mL-1; two-sample t-test: t129 = -7.81, p<0.0001), but glucose and lactate did not differ between years.  Fish spent 8.24 ± 1.23 hours (range: 0.3-121.2 hours) in the tailrace and fishway at Seton Dam (i.e. between first and last detections), but times were not correlated with tagging date or the maximum temperature or discharge category experienced during this period. Maximum fishway temperatures experienced (18.2 ± 0.1 °C, range: 13.8-20.4 °C; Figure 2.2A) were often above the optimal temperature for Gates Creek sockeye (16.4 °C), and decreased with tagging date in 2013 (r=-0.6), but not 2014. During standard flow conditions (i.e. ‘low’ discharge), mean hourly Seton Dam discharge was between 24.6-29.5 m3s-1 in 2013 and 25.4-27.6 m3 s-1 during Aug 8-19, 2014. ‘High’ discharge was between 31.0-33.1 m3 s-1 during Aug 20-Sept 7, 2014 (Fig 2.2B). Discharge decreased with tagging date in both years (r = -0.5 in 2013, r = -0.7 in 2014). 2.3.1 Receiver efficiency Detection efficiency of the radio receiver at the release location on the west bank of the Fraser River was high (99%, n=135). At the Seton Dam, detection efficiencies of individual receivers were moderate to high [radio receivers: tailrace (83%, n=135) and fishway (2014 only: 47%, n=83); PIT antennas: fishway entrance (93%, n=135) and exit (92%, n=135)], but all fish (n=135) had multiple detections on at least one receiver when data were pooled. At Gates Creek, receivers were also pooled, as no single receiver detected all fish [radio: channel entrance (98% of survivors, n=83); PIT antennas: lower Gates Creek (2014 only: 43%, n=60) and channel entrance (64%, n=83)]. Detection of a  27 fish on any receiver at Gates Creek mean that the individual was considered a successful migrant in Model 1 and included in Model 2.   2.3.2 Survival to spawning grounds (Model 1)  Plasma glucose and lactate concentrations, FL, and year formed the most parsimonious survival model, which accounted for 15% of variation in the data (Table 2.1). Lower plasma glucose concentrations at tagging were associated with higher survival to Gates Creek (Figure 2.3A; Figure 2.4), and glucose was the most important explanatory variable, appearing in all models in the confidence set (Table 2.1). Plasma glucose concentrations were 4.59 ± 0.08 mmol L-1 in females that reached Gates Creek and 5.32 ± 0.19 mmol L-1 in females that died en route. All other coefficient estimates were uncertain, as the confidence intervals (CI) crossed zero (Figure 2.3A). Plasma lactate concentration, FL, and year trended towards positive effects on survival (Figure 2.3A). Survival of tagged fish to spawning grounds was higher in 2014 (72% of fish, n=60) than in 2013 (44%, n=23; Figure 2.4), although the model-averaged coefficient for year was uncertain (Figure 2.3A). Plasma estradiol concentration, tagging date, and discharge were associated with increased survival, while the effect of maximum fishway temperature was approximately zero (Figure 2.3A). 2.3.3 Migration duration (Model 2) On average, females spent longer in Seton and Anderson lakes in 2014 (9.9 ± 0.7 days, range: 3.1-23.3 days; n=60) than in 2013 (6.4 ± 0.9 days, range: 2.4-23.8 days; n=23; Figure 2.5), and year was present in all models in the confidence set (Table 2.1). Females tagged earlier in the season took longer to reach spawning grounds (r=-0.6, years  28 pooled; Figure 2.5). Tagging date was present in all models in the candidate set (Table 2.1), and the effect of tagging date was twice as strong as the effect of year (Figure 2.3B). All other explanatory variables had uncertain coefficient estimates, with CI spanning zero (Figure 2.3B), and addition to the top-ranked model in the candidate set (i.e. tagging date + year; r2=0.33) described no additional variation in the data (Table 2.1). Discharge and plasma glucose concentration trended towards positive relationships with duration, while FL, plasma estradiol concentration, plasma lactate concentration, and maximum fishway temperature trended towards negative effects on duration (Figure 2.3B). 2.3.4 Temperature experience in Seton and Anderson lakes In 2014, recovered temperature loggers had archived a mean of 998 ± 102 temperature records (range: 228-3294 records; n=39) between the time of last telemetry detection at Seton Dam and first detection at Gates Creek. The mean lake temperatures experienced by individual females in 2014 ranged from 9.2-14.8 °C (n=39) and minimum and maximum temperatures ranged from 4.9 to 11.1 °C and 17.4 to 22.4 °C, respectively. On average, females spent 33% of the migration in the ToptAS window (between 13.4-19.5 °C; Table 2.1). Time in the ToptAS window ranged from 8% to 80% of the total time individuals spent in Seton and Anderson lakes (see examples in Figure 2.6). Temperatures >19.5 °C represented only 0.2% of records. 2.3.5 Reproductive longevity (Model 3) Radio-tagged females lived in the spawning channel for a mean of 10.8 ± 0.7 days (range: 4-22 days, n=39) in 2014. Time in the ToptAS window had a positive effect on longevity and appeared in all models in the confidence set (Table 2.1; Figure 2.3C).  29 Females that experienced high Seton Dam discharge lived fewer days on spawning grounds than females that experienced low discharge (Figure 2.3C), and discharge also appeared in all top models (Table 2.1). Females that arrived later lived fewer days on spawning grounds (Figure 2.3C). The top model, which included time in ToptAS window, Seton Dam discharge, and arrival date, accounted for 33% of variation in reproductive longevity (Table 2.2). According to model predictions, a female that arrived on September 2nd, the median arrival date, lived ~12.1 days, following exposure to low Seton Dam discharge, and ~8.5 days, following exposure to high discharge, given mean values of all other variables (Figure 2.7A). Females that spent 50% of time in the ToptAS window lived ~13.2 days in the spawning channel, following exposure to low discharge, and ~9.5 days, following exposure to high discharge, given mean values of all other variables (Figure 2.7B). The direction of effect was uncertain for all other variables. Plasma glucose concentration had negative trends, while plasma estradiol and lactate concentrations, FL, and maximum fishway temperature had positive trends with longevity (Figure 2.3C). 2.3.6 Egg retention (Model 4) Of the thirty-eight females assessed on spawning grounds in 2014, 32% (n=12) retained eggs at death and 68% (n=26) did not.  AICc scores from individual GLMs fitted to the egg retention data were compared for a model with Seton Dam discharge and a model with arrival date on spawning grounds as the only explanatory variable. The discharge model had more support from the data (AICc: 44) than the arrival date model (AICc: 51; Table 2.3), so Seton Dam discharge was included as a categorical variable in Model 4, and arrival date was  30 excluded. Both the earliest and latest arriving females successfully deposited eggs on spawning grounds (Figure 2.7C).  Females that spent a greater proportion of time in Seton and Anderson lakes in the ToptAS window were less likely to retain eggs, as were females that experienced low Seton Dam discharge (Figure 2.3D); both variables appeared in all models in the confidence set (Table 2.2). The averaged model predicted that a female fish had a 50% probability of retaining eggs following exposure to high discharge and ~30% of migration in the ToptAS window, or following exposure to low discharge and ~6.8% of migration in the ToptAS window, given mean values of all other variables (Figure 2.7D). The effects of all other variables were uncertain. Plasma glucose trended towards a negative relationship with egg retention (Figure 2.3D) and was also in the top-ranked model, which explained 33% of the variation in egg retention (Table 2.2). The effects of other explanatory variables were not statistically significant; plasma lactate and estradiol concentrations, FL, and maximum fishway temperature were all positively related to egg retention, while plasma lactate concentration was negatively related to egg retention (Figure 2.3D). Females that spawned successfully lived longer on spawning grounds (12.4 ± 0.8 days, range: 4-22 days, n=26) than females that retained eggs at death (6.8 ± 0.8 days, range: 4-11 days, n=12; two-sample t-test: t31=-5.25, p<0.0001; see Figure 2.7E). 2.4 Discussion I examined the relative importance of physiological condition and migratory experience to female sockeye salmon fitness in the final stages of reproductive migration. Explanatory variables differed in importance across the four response variables investigated here, which underlines the complexity of factors governing the final stages  31 of migration and spawning. Temporal and physiological factors were most important to the duration and success of migration, while migratory experience strongly affected longevity and egg retention on spawning grounds.  This is the first study to relate spawning success to use of a physiologically relevant temperature range (≈ ToptAS ± 3 °C) by wild adult Pacific salmon in the field environment. Gates Creek is one of a few populations of Pacific salmon with a known aerobic scope curve (Lee et al. 2003b; Eliason et al. 2011), which can be used to infer cardiac capacity and scope for activity from thermal experiences of individual fish. Females that spent a greater proportion of the migration through Seton and Anderson lakes in the ToptAS window (13.4-19.5 °C) had lower egg retention at death, which was consistent with my predictions. Females occupied the ToptAS window at different instances for varying durations, such that a variety of thermal experiences (i.e. 15-80% of time in the ToptAS window) were associated with successful spawning. My results provide further evidence that thermal experience in the weeks before spawning is important for successful reproduction in female Pacific salmon (e.g. in pink salmon [O. gorbuscha], Jensen et al. 2004, reviewed in McCullough et al. 2001). Studies on Atlantic salmon (Salmo salar) have demonstrated that females exposed to temperatures that cause hormone suppression and reduced egg size and survival (e.g. 22 °C) are able to recover, if subsequently exposed to cooler temperatures (e.g. 14 °C, King and Pankhurst 2003, King et al. 2007). Somatic energy conservation, physiological stress, reproductive maturation, olfactory homing, and disease are all thought to influence thermal behaviour of sockeye salmon during lake migration (e.g. Brett 1995, Newell and Quinn 2005, Wagner et al.  32 2005, Roscoe et al. 2010a). Populations with longer freshwater holding exhibit preference for cooler temperatures. Sockeye salmon that return to Copper Creek, enter freshwater >130 days before spawning, and bioenergetics models indicate temperatures <10 °C are required to conserve adequate energy for spawning (Katinic et al. 2014). Lake Washington sockeye salmon spend an average of 83 days in the lake and 92% of time between 9-11 °C (Newell and Quinn 2005). In comparison, Gates Creek sockeye salmon occupied Seton and Anderson lakes for much shorter periods (2-24 days, this study; 2-36 days, Roscoe et al. 2010a) at warmer and more variable average temperatures (9.2-14.8 °C, this study; 12.3-15.9 °C, Roscoe et al. 2010a). Cooler temperatures are available, so observed thermal behaviour does not reflect energy conservation alone (although females with lower somatic energy levels avoid the warmest temperatures; Roscoe et al. 2010a).  Females rarely occupied temperatures above 19.5 °C; instead they were observed at temperatures <13.4 °C during the proportion of time spent outside the ToptAS window. I expected that female fish that spent more time in the ToptAS window would have higher energy expenditure than females occupying cooler temperatures and exhibit reduced reproductive longevity, but more time in the ToptAS window was associated with greater reproductive longevity. There may be another physiological process underlying the associations between cooler lake temperature selection and lower longevity and higher egg retention. A possible explanation is that some tagged females were developing diseases and that disease processes or immune function lead to both cool thermal preference and reduced reproductive success. Water temperature increases virulence and replication rates in some pathogens (Holt et al. 1975, Udey et al. 1975), and use of cool lake temperatures may prevent pathogen-associated mortality (Wagner et al. 2005,  33 Mathes et al. 2010, Bradford et al. 2010). Although use of cool temperatures may allow diseased fish to survive to spawning grounds, my results suggest that this behaviour is associated with reduced reproductive success in female fish. It is yet to be examined whether Pacific salmon with higher pathogen loads or in more advanced stages of disease select cooler lake temperatures to inhibit pathogen replication and whether these fish also die prematurely on spawning grounds with retained eggs. In the present study, earlier arriving females lived longer on spawning grounds than late arrivals (Morbey and Ydenberg 2003, Dickerson et al. 2005). Longer-lived females were less likely to retain eggs at death, and the minimum longevity required for spawning was four days. Similar results were also observed in a study of female sockeye salmon behaviour on spawning grounds (Hruska et al. 2011). However, egg retention was not strongly related to arrival date, as both the earliest and latest arriving females spawned successfully. Arrival date was not examined in the global model, due to collinearity with Seton Dam flow. It is possible that all female fish arrived during peak spawning, so reproductive opportunities did not vary by arrival date. The absence of a strong effect of arrival date is unusual for patterns of egg retention, because pre-spawning mortality is thought to be more prevalent among early arriving female fish (Gilhousen 1990, Fukushima et al. 1998), but may be related to unusual migration timing in 2014. The first fish arrived at Gates Creek nine days later in 2014 than in 2013 (Lingard et al. 2014; S. Lingard, InStream Fisheries Research Inc., pers. comm.), so the early females in this study may have arrived on spawning grounds with greater reproductive preparedness.  Awareness of delayed or sublethal effects of migratory stressors is growing in research on Pacific salmon populations (e.g. Budy et al. 2002, Caudill et al. 2007,  34 Donaldson et al. 2012, Fenkes et al. 2015). High rates of pre-spawning mortality have been observed at the population level, following exposure to sub-lethal temperatures and discharge (West and Mason 1987, Gilhousen 1990, Quinn et al. 2007, Hinch et al. 2012). In this study, I found new evidence to support the hypothesis that exposure to a stressor can have a deleterious effect on a migrant’s fitness in a subsequent life stage. In 2014, there was an adjustment at Seton Dam to release water from a siphon ~10m from the fishway entrance, rather than directly adjacent to the fishway entrance, which is standard operating practice. The siphon change increased the water velocities in the tailrace, the area of the tailrace covered by turbulent flows, and the Seton Dam discharge by up to 25% (Casselman et al. 2015). Gates Creek sockeye salmon swimming behaviour during each of the flow conditions (i.e. ‘low’ and ‘high’ discharge) was examined in a parallel study (Burnett et al. In press). In this study, the dam passage experience was not important to survival of female fish to spawning grounds, and Model 1 predicted a slightly higher probability of survival for females that experienced high discharge (under altered tailrace flow conditions) in 2014. However, dam passage did affect reproduction in a subset of females. Female fish that experienced high discharge at Seton Dam had lower reproductive longevity and were more likely to retain eggs at death. Specifically, a female that experienced high Seton Dam discharge would have had to spend a greater proportion of time at ToptAS or arrive on spawning grounds 10-15 days earlier to achieve the same longevity as a female that experienced low discharge. The rate of egg retention in fish that experienced >30.0 m3s-1 was 57% (n=8), compared to 17% (n=4) that experienced <28.0 m3s-1. Exhaustion or energy depletion caused by high flows (Rand et al. 2006; Nadeau et al. 2010), cumulative stress (Gilhousen 1990), recovery time (Martin  35 et al. 2015), insufficient somatic energy (Dickhoff 1989, Crossin et al. 2004b, Hruska et al. 2010), suppression of maturation processes (Palstra and van den Thillart 2010), or a combination of multiple mechanisms (Fenkes et al. 2015), are possible explanations for the reduced reproductive longevity and spawning ability that I observed.   As predicted, females with lower plasma glucose concentrations at tagging exhibited higher survival following Seton Dam passage than females that died en route to Gates Creek. Fish mobilize glucose as an energy substrate for swimming, and high levels have been used as indicators of exhaustive exercise (Wood 1991, Kieffer 2000, Farrell et al. 2001a, 2001b). The highest concentrations measured in this study were similar to peak venous concentrations observed following maximal swimming (e.g. 7.5 ± 1.3 mmol L-1; Eliason et al. 2013a), indicating that some fish arrive in the Seton River in a state of physiological stress and may not have been able to recover from Seton Dam passage. This result is consistent with results of previous biotelemetry studies, which found higher plasma glucose concentrations in Pacific salmon that fail to reach spawning tributaries (Young et al. 2006, Roscoe et al. 2010b). Surviving sockeye salmon exposed to warmer than average Fraser River temperatures had similar plasma glucose concentrations (i.e. 4.68 ± 0.24 mmol L-1; Young et al. 2006) to those observed here.  My assessment of temperature effects was limited because only three females included in Model 1 experienced ≥20 °C in the Seton Dam tailrace. The earliest females tagged in 2013 experienced tailrace temperatures >21 °C (approaching Tcrit for Gates Creek sockeye, Lee et al. 2003a, Farrell et al. 2008) and had only 17% survival (n=6), but were omitted from analyses because they had different capture and release locations than the rest of the fish used in the study (they were captured in the fishway and released in  36 the lower Seton River), which could have confounded the effects of temperature (see Casselman et al. 2013). I could not assess the effect of lake thermal experience on survival, because archival temperature loggers were not recovered for fish that died en route. This is an important gap in current research, as it may reveal thermal behaviours associated with migration failure (Roscoe et al. 2010a, Keefer et al. 2015). Unsuccessful migrants may experience warmer temperatures than survivors when adult salmon are migrating near the limits of thermal tolerance (Keefer et al. 2008). Early migrants spent up to 20 days longer in Seton and Anderson lakes than females tagged later in the run. The negative relationship between tagging date and migration duration reflects the importance of coordinated arrival on spawning grounds for individuals from the same population. Environmental cues, like photoperiod and temperature are proximate drivers that synchronize final maturation and ovulation in salmonids (Bromage et al. 2001, Davies and Bromage 2002, Taranger et al. 2003). In Pacific salmon, the timing of peak spawning is under genetic control, which ensures suitable environmental conditions for reproduction, egg development, and juvenile emergence (reviewed in Burgner 1991, see also Brannon 1987), which show little variation in timing (e.g. in chum salmon [O. keta], Beacham and Murray 1986, Tallman and Healey 1991). It has long been noted that the majority of sockeye salmon in population arrive within one to two weeks and that a population’s migration and spawning dates vary little between years (Killick 1955). Unlike in the marine environment, where sockeye salmon migrate in a directed manner (Wilson et al. 2014), lake migrations are not likely unidirectional (Young and Woody 2007), and earlier tagged  37 Gates Creek females hold and circle more than later migrants, particularly in Anderson Lake (Roscoe et al. 2010a).  Females tagged in 2013 reached spawning grounds more quickly than females in 2014, which may potentially be attributed to several factors. First, a later mean tagging date in 2013 meant that the earliest migrants, presumably with the longest migration durations through Seton and Anderson lakes, were not sampled. Second, Fraser River temperatures were abnormally high in 2013 (July 22-Aug 12 daily means: 18.9-21.6 °C at Qualark; D.A. Patterson, Fisheries and Oceans Canada, unpublished data), and exposure to warm temperatures has previously been associated with faster migrations in sockeye salmon (Naughton et al. 2005, Keefer et al. 2008). Third, a small number of fish survived to spawning grounds in 2013 (n=23; only one with duration >15 days), as the earliest fish tagged in 2013 experienced high Fraser River temperatures and very low survival.  Although there was a tag recovery program in place during this study, wherein rewards were provided for returns from First Nations food fisheries, only four radio transmitters with attached archival temperature loggers (~1% of fish that passed the Seton Dam) were returned. I suspect that additional unreported harvest of migrating Gates Creek sockeye salmon occurred between the Seton Dam and spawning grounds during the study, although I have no direct means of accounting for this component of mortality in my models. In 2015, a year after the present study ended, a more comprehensive program was implemented for assessing unreported harvest. Mobile radio tracking in known locales of fishing harvest generated estimates of 5-10% of tagged fish being caught (Casselman et al. 2015). Although I cannot be certain whether this rate is applicable to 2013 or 2014, it would reflect a relatively small component of the total  38 mortality in this study. There is no reason to suspect that sockeye salmon captured in fisheries differed physiologically (see Cooke et al. 2009) or experienced different migration conditions compared to survivors, as harvest occurred throughout the entire Gates Creek run.  In lakes, female sockeye salmon thermal selection reflects trade-offs between temperatures that increase efficiency of physiological processes, including swimming, maturation, and stress recovery, and temperatures that conserve finite somatic energy and delay pathogen development. In this study, many females showed a preference for temperatures that fell below the ToptAS window (i.e. ~10-12 °C). Rates of decline in physiological performance are not equal as temperatures increase and decrease from ToptAS, and many ectotherms prefer temperatures below, but close to, the optimum (Martin & Huey 2008, Asbury & Angilletta 2010). Indeed, few females in this study occupied temperatures in the upper half of the ToptAS window (i.e. 16.5-19.5 °C) for long periods, and the warmest temperatures were experienced in Portage Creek (>20 °C), which has no opportunity for thermal refuge. Maturation processes likely influence the thermal preference of female sockeye salmon (e.g. Roscoe et al. 2010a), as migrating females invest a greater proportion of somatic energy into gonads than males (Gilhousen 1980, Hendry and Berg 1999). The optimum and upper threshold temperatures for vitellogenesis have not been defined for species of Pacific salmon, but the Topt likely differs from (Clark et al. 2011, 2013), and may be cooler than (Macdonald et al. 2000), the ToptAS. Female salmonids held at temperatures >18 °C exhibit suppressed plasma testosterone and estradiol, reduced egg size, and lower egg survival, relative to fish held at temperatures <15 °C (e.g. sockeye salmon, Macdonald et al. 2000; brown trout [Salmo  39 trutta], Campbell et al. 1994; Atlantic salmon, King et al. 2003), and quality of ovulated eggs declines quickly at temperatures >13 °C (e.g. in cutthroat trout [O. clarkii clarkii], Smith et al. 1983; in coho salmon [O. kisutch], Flett et al. 1996). Thermal selection may also fall outside of the ToptAS window because migrants do not require as much scope for swimming in natal lakes. Even the fastest females in this study swam slower [rate ≈21 km day-1 (50 km / 2.4 days)] than sockeye salmon tracked in the Fraser River (rate ≈17–40 km day-1, English et al. 2005), where fish must also overcome river flow. A lower threshold for aerobic scope (e.g. 80%) may be adequate to simultaneously migrate and mature in lakes. Additional research is required to identify thresholds of acute and cumulative stress for female Pacific salmon in the final stages of migration (Eliason and Farrell 2016) and to understand how migrants utilize thermal refugia for recovery. The results of this study demonstrate that migratory experience of individual females in the weeks prior to spawning are important to reproductive longevity and egg retention. I encourage biotelemetry studies of Pacific salmon to assess reproductive success on spawning grounds, in order to account for delayed effects of migratory experience on fitness.  40  Figure 2.1 – The Seton-Anderson watershed and its location in British Columbia, Canada (inset); Gates Creek and the artificial spawning channel (A); and the Seton Dam and fishway (B). R = radio receiver, P = pass-through PIT antenna, F = fish collection fence, D = Water Survey of Canada discharge gauge, W = spawning channel weir, white star = release site (2013 + 2014), filled star = release site (2013 only). Detections on telemetry receivers in A were grouped as ‘Gates Creek’ and in B as ‘Seton Dam’ for analysis. 41  Figure 2.2 – Fishway temperature (A) and Seton Dam discharge (B) during period when females were detected in the dam tailrace, August 9-September 5, 2013 (grey lines) and August 6-September 9, 2014 (black lines). ‘High’ discharge (i.e. altered flow conditions) occurred from August 8-19, 2014, and ‘low’ discharge (i.e. standard flow conditions) from August 6-7 and August 20-September 9, 2014. 131517192123Temperature (°C) AAug 06 Aug 13 Aug 20 Aug 27 Sep 03242628303234Discharge (m3 s-1) B 42 Table 2.1 – Candidate model sets (models with cumulative weights ≥ 0.95) from multivariate statistical models of female sockeye migration survival and duration to spawning grounds, reproductive longevity on spawning grounds, and egg retention at death.  Response Model logLik AICc ΔAICc wi r2 1. Survival Glucose + FL + Lactate + Year  -76.7 164 0.00 0.27 0.15  Survival Glucose + Lactate + Year -78.1 165 0.73 0.19 0.13  Survival Glucose + FL + Lactate  -78.3 165 1.20 0.15 0.13  Survival Glucose + FL + Lactate + Estradiol -77.3 165 1.25 0.15 0.14  Survival Glucose + FL + Year -78.5 165 1.59 0.12 0.13  Survival Glucose + FL + Estradiol -78.6 165 1.59 0.12 0.13  Survival Null: intercept -90.0 182 18.2 0.00 0.00 2.  Duration Tagging date + Year -234.3 477 0.00 0.34 0.33  Duration Tagging date + Year + Discharge -234.1 479 1.82 0.14 0.33  Duration Tagging date + Year + Estradiol -234.1 479 1.83 0.14 0.33  Duration Tagging date + Year + Lactate -234.1 479 1.91 0.13 0.33  Duration Tagging date + Year + Glucose -234.2 479 2.02 0.13 0.33  Duration Tagging date + Year + Tdam -234.2 479 2.03 0.13 0.33  Duration Null: intercept -252.2 509 31.5 0.00 0.00 3. Reproductive longevity Discharge + Arrival date + ToptAS -104.7 221 0.00 0.63 0.33  Reproductive longevity Discharge + Arrival date -108.0 225 3.86 0.09 0.19  Reproductive longevity Discharge + Arrival date + Estradiol  -106.7 225 3.93 0.09 0.22  Reproductive longevity Discharge + ToptAS -108.2 226 4.26 0.07 0.22  Reproductive longevity Discharge + ToptAS + Tdam  -107.0 226 4.61 0.06 0.24  Reproductive longevity ToptAS + Tdam + Lactate -107.1 226 4.68 0.06 0.26  Reproductive longevity Null: intercept -113.6 232 6.79 0.00 0.00    43  Response Model logLik AICc ΔAICc wi r2 4. Egg retention ToptAS + Discharge + Glucose -15.8 41 0.00 0.35 0.33  Egg retention ToptAS + Discharge -17.2 41 0.40 0.29 0.27  Egg retention ToptAS + Discharge + Lactate -17.0 43 2.46 0.10 0.28  Egg retention ToptAS + Discharge + Tdam -17.0 43 2.67 0.09 0.29  Egg retention ToptAS + Discharge + Estradiol  -17.2 44 2.79 0.09 0.28  Egg retention ToptAS + Discharge + FL -17.2 44 2.91 0.08 0.27  Egg retention Null: intercept -23.7 49 9.01 0.00 0.00 Note: logLik = model log likelihood, ΔAICc = difference in AICc from top model, wi is the AICc model weight, r2 = Hosmer-Lemeshow (1989) pseudo-r2 for models 1 and 4, adjusted-r2 for models 2 and 3.     44 Table 2.2 – Comparison of tagging date, fork length (FL), plasma lactate, glucose, and estradiol concentrations, maximum Seton Dam temperature (Tdam) and discharge (sample sizes shown for ‘low’ and ‘high’ categories), proportion of time in the ToptAS window (i.e. 13.4-19.5 °C), and arrival date at Gates Creek for female sockeye salmon tracked through Seton and Anderson lakes. The range (for tagging and arrival dates) or mean ± SE (for other variables) are shown for the fish included in statistical models of (1) survival to spawning grounds, (2) migration duration, (3) reproductive longevity, and (4) egg retention, along with the sample size (n).  Model response variable  Model explanatory variables  n  Tagging date FL  (cm) Lactate (mmol L-1) Glucose (mmol L-1) Estradiol  (ng mL-1) Tdam  (°C) Discharge (n) ToptAS window  Arrival date  Low High 1. Survival Aug 16- Sept 2, 2013 Aug 5- Sept 7, 2014 58.1 ± 0.28 4.97 ± 0.25 4.87 ± 0.09 5.72 ± 0.33 18.2 ± 0.09 108 27   135 2. Duration Aug 16- Sept 2, 2013 Aug 5- Sept 7, 2014 58.5 ± 0.37 5.54 ± 0.34 4.59 ± 0.08 6.40 ± 0.45 18.1 ± 0.12 63 20   83 3. Reproductive longevity Aug 11- Sept 6, 2014 59.3 ± 0.63 5.60 ± 0.40 4.51 ± 0.10 8.13 ± 0.65 18.0 ± 0.17 25 14 0.33 ± 0.03 Aug 22- Sept 12, 2014 39 4. Egg retention Aug 11- Sept 6, 2014 59.4 ± 0.65 5.63 ± 0.41 4.53 ± 0.10 8.11 ± 0.67 17.9 ± 0.17 24 14 0.33 ± 0.03 Aug 22- Sept 12, 2014 38 Note: Arrival date was omitted from model 4 due to multicollinearity with Discharge (see ‘Methods’). Data are presented for clarity. Sample sizes differ in models 3 and 4, as egg retention was not recorded for one female.  45  Figure 2.3 – Model-averaged standardized coefficients (mean=0, standard deviation=2) with 95% confidence intervals (CI) from models describing (A) survival, (B) migration duration, (C) reproductive longevity, and (D) egg retention. Filled circles indicate coefficients with CI that do not cross zero.    -2-1012Survival A-8-404Migrationduration B-8-4048ReproductivelongevityC-6-303Eggretention DYear TaggingdateFL Glucose Lactate Estradiol Discharge Tdam Time atToptASArrivaldateCoefficientsRelative effect size 46  Figure 2.4 – Scatterplot of plasma glucose concentration at tagging vs. survival for females tagged in 2013 (grey triangles) and 2014 (black circles). Model 1 predictions shown for females that experienced low (solid line) and high (dashed line) discharge at the Seton Dam in 2013 (grey line) and 2014 (black line). High discharge (i.e. altered flow conditions) did not occur in 2013.   low dischargehigh discharge4 6 8 100.00.20.40.60.81.0P(survival)Glucose (mmol L-1) 47  Figure 2.5 – Scatterplot of tagging date vs. migration duration for females tagged in 2013 (grey triangles) and 2014 (black circles). Model 2 predictions shown for females that experienced low (solid line) and high (dashed line) discharge at the Seton Dam in 2013 (grey line) and 2014 (black line). High discharge (i.e. altered flow conditions) did not occur in 2013.  low dischargehigh discharge215 225 235 245 2550510152025Duration (days)Tagging date 48  Figure 2.6 – The thermal experience of two female sockeye salmon between Seton Dam and Gates Creek. These fish spent the highest (80%, left) and lowest (8%, right) proportion of time in the ToptAS window (13.4-19.5 °C; indicated by shaded region), of the females with recovered archival temperature loggers (n=38), during migration through Seton and Anderson lakes. Passage through Portage Creek is indicated by a dashed vertical line. The fish were tagged on Sept 6, 2014 (left) and Aug 17, 2014 (right), and one spawned successfully (left), while the other retained eggs at death (right).Sep 09Sep 10Sep 11Sep 125101520Temperature (°C)Aug 19Aug 20Aug 21Aug 22Aug 23Aug 24Aug 25Aug 26Aug 27Aug 28Aug 29Aug 30Aug 31Sep 01Sep 02Sep 03 49  Figure 2.7 – Scatterplots of reproductive longevity vs. (A) arrival date and (B) proportion of time in ToptAS window (n=39) and egg retention vs. (C) arrival date and (D) proportion of time in ToptAS window (n=38) for female sockeye salmon tagged in 2014. Model predictions are shown for low (solid line) and high (dashed line) discharge at the Seton Dam in (A), (B), and (D). Boxplots (E) show reproductive longevity differed significantly for females that did (n=12) and did not (n=26) retain eggs (two-sample t-test: t31=-5.25, p<0.0001), with the mean and range (whiskers) shown.   low dischargehigh dischargeAAug 23 Sept 2 Sept 120510152025Longevity (days)B0.2 0.4 0.6 0.80510152025Aug 23 Sept 2 Sept 1201CArrival dateEgg retentionD0.2 0.4 0.6 0.80.00.20.40.60.81.0Time in ToptAS window0 15101520E*Longevity (days)Egg retention 50 Table 2.3 – Comparison of generalized linear models describing egg retention in female sockeye salmon (n=38), with Seton Dam discharge (categorical: low [0] or high [1]) and arrival date as single explanatory variables.    Response Model logLik AICc ΔAICc wi r2 4. Egg retention Discharge -20.0 44 0.00 0.91 0.15  Egg retention Arrival date -23.6 51 7.02 0.03 0.01  Egg retention Null: intercept -23.7 49 5.31 0.06 0.00  Note: Seton Dam discharge and arrival date were multicollinear, so the variable in the model with the lower AICc value was included in Model 4 (see ‘Methods’). logLik = model log likelihood, ΔAICc = difference in AICc from top model, wi is the AICc model weight, r2 = Hosmer-Lemeshow (1989) pseudo-r2.   51 Chapter 3: Adult sockeye salmon temperature selection in thermally stratified lakes is not optimized for either energy conservation or aerobic capacity 3.1 Introduction Within a thermally heterogeneous environment, ectothermic fishes preferentially occupy a relatively narrow range of temperatures (Magnuson et al. 1979), termed behavioural thermoregulation (Neill et al. 1972, Beitinger et al. 1975). All aspects of fish physiology and performance are temperature-dependent, and may differ in their thermal optima (Fry 1947, Brett 1971).  Patterns of temperature selection reflect trade-offs in capacity for different physiological processes (Cossins and Bowler 1987, Angilletta 2009, Farrell 2015).  Thermal tolerance is the range of temperatures that fish can withstand, in the absence of other environmental stressors (Fry 1947). Within this range are temperatures that result in maximum rates of physiological functions, including metabolic processes (e.g. routine and active metabolism, aerobic scope for activity); feeding, digestion, and growth; immune function; and swimming performance, all of which can be measured experimentally (e.g. Brett 1971). When temperature increases or decreases from the optimum, physiological rates and fish performance decrease. Thermal preference is the temperature that fish occupy in laboratory thermal gradients, thought to coincide with important physiological and life-history processes (Fry 1947). While much may be known about the thermal tolerance of a fish species, study of behavioural thermoregulation in wild, freely swimming fish is rare (Beitinger and Fitzpatrick 1979, Clark et al. 2013; but see Nowell et al. 2015, Harrison et al. 2016).   52 Anadromous Pacific salmon (Oncorhynchus spp.) are heterothermic stenotherms that spend multiple life stages in thermally heterogeneous habitats. As juveniles, species like sockeye salmon (Oncorhynchus nerka) rear for one to two years in freshwater lakes and experience estuarine environments en route to adult feeding areas in the ocean. During the reproductive migration back to natal streams, adults from some Pacific salmon populations are exposed to water temperatures approaching the upper limit of thermal tolerance, especially as rivers warm with climate change (Quinn and Adams 1996, Patterson et al. 2007, Hague et al. 2011). Adult Pacific salmon demonstrate behavioural thermoregulation throughout the migration. Migration strategies have evolved to avoid peak summer river temperatures; populations enter freshwater early and utilize thermal refugia during the migration, or enter freshwater after river temperatures have peaked (Hodgson and Quinn 2002). Once in rivers, some populations stop upstream movement when temperature surpasses a threshold (e.g. > 21 °C in Okanagan sockeye salmon, Major and Mighell 1967, Hyatt et al. 2003), or hold in coolwater tributaries (e.g. of the Columbia River, Hodgson and Quinn 2002, Goniea et al. 2006, Keefer et al. 2009, 2015). The hypolimnion of deep lakes and reservoirs also provide thermal refuge from high river temperatures. Many sockeye salmon populations have natal lakes near spawning grounds, where adults hold until spawning (e.g. Newell and Quinn 2005, Mathes et al. 2010, Katinic et al. 2015). Despite the wide range of habitats available in thermally stratified lakes, sockeye salmon may occupy a very precise range of temperatures (e.g. ~92% of time between 9-11 °C in Lake Washington, Newell and Quinn 2005).  A great deal is known about the temperature-dependence of Pacific salmon performance, particularly for sockeye salmon (e.g. Lee et al. 2003a, 2003b, Eliason et al. 2013a, 2013b). One well-studied property of the cardiorespiratory system is aerobic scope (the  53 difference between maximum and routine oxygen consumption rates; Fry 1947), a measure of cardiorespiratory capacity for activities, like swimming and reproductive maturation. Aerobic scope is maximized at an optimum temperature (ToptAS); above ToptAS, as the oxygen demands of routine metabolism increase, aerobic scope declines, and fish activity is increasingly powered by anaerobic metabolism. At the upper and lower critical temperatures (Tcrit), aerobic scope is zero and survival is time-limited (Pörtner 2001, Pörtner & Farrell 2008). Aerobic scope is hypothesized to limit thermal tolerance and underlie migratory performance of adult sockeye salmon in the Fraser River mainstem (Farrell et al. 2008, Eliason et al. 2011). However, thermal behaviour of adult Pacific salmon in thermally heterogeneous environments has not yet been interpreted using a physiological framework that considers the aerobic scope available to support activity at different temperatures. Thermal selection by adult Pacific salmon in lakes and reservoirs reflects trade-offs between a number of factors. Energy conservation in migrating adult salmon is likely the primary driver of behavioural thermoregulation, because like all anadromous migrants, Pacific salmon do not feed during freshwater spawning migrations, so somatic energy stores are limited. Energy conserving behaviours are adaptive traits; because Pacific salmon are semelparous, individuals that exhaust energy stores and perish en route to spawning grounds (e.g. after exposure to high temperatures or elevated river flows, Rand and Hinch 1998, Macdonald et al. 2000, Crossin et al. 2004, Rand et al. 2006), or that have insufficient energy for gamete development and spawning upon arrival (Fenke et al. 2015), will have zero lifetime fitness. The energetic costs associated with standard metabolism and swimming activity both increase with temperature (Brett 1995), so Pacific salmon likely use cool temperatures whenever available. En route to spawning grounds, Pacific salmon invest energy, and partition  54 aerobic scope, into a suite of other activities. First, swimming activity consumes more than 50% of total somatic energy for long distance migrants (Brett 1995, Crossin et al. 2004b), and net cost of transport (COT) is lowest at temperatures near ToptAS (Webb 1995, Lee et al. 2003a, Eliason et al. 2013b). While very cool temperatures minimize routine metabolic costs, they decrease tailbeat frequency and maximum swimming velocity (Brett 1995). In lakes and reservoirs, Pacific salmon can exhibit mean migration rates that are roughly equivalent to migration rates in mainstem river or ocean environments (e.g. sockeye salmon, Naughton et al. 2005, Roscoe et al. 2010a). However, when approaching spawning grounds, migrants sometimes mill and hold for long periods at tributary mouths (Newell and Quinn 2005, Young and Woody 2007, Roscoe et al. 2010a); this behaviour may be associated with a physiological requirement for cooler temperatures. Second, adults invest energy in primary and secondary maturation. The optimum temperature for maturation is unknown, but thought to be near ToptAS (Eliason and Farrell 2016); high temperatures can inhibit maturation (Pankhurst and King 2010, Jeffries et al. 2012, Fenkes et al. 2015). Female sockeye salmon invest considerably more somatic energy in gonads than males (Crossin et al. 2004b), which may underlie differences in thermal behaviour between sexes (Roscoe et al. 2010a). Third, migrants are intermittently recovering from physiological stress, caused by high temperature exposure and anaerobic activity. Following anaerobic swimming, a fish requires substantial oxygen and recovery time to re-establish homeostasis and restore oxygen stores (Wood et al. 1983, Gaesser and Brooks 1984, Wood 1991), which is hypothesized to be most efficient at temperatures near ToptAS (Lee et al. 2003b, Eliason and Farrell 2016). Lastly, Pacific salmon are exposed to a suite of freshwater pathogens en route to spawning grounds, and cool thermal experience is  55 thought to delay disease progression for migrants with high pathogen load (Farrell et al. 2008, Wagner et al. 2009).  The objective of this study was to determine whether thermal habitat use by adult migrating Gates Creek sockeye salmon reflected a trade-off between energy conservation (i.e. use of cool temperatures) and aerobic capacity (i.e. use of temperatures near ToptAS). I used radio telemetry and internal archival temperature loggers to track and record the temperature experience of adults as they passed through a fishway and two natal lakes en route to spawning grounds. I developed a framework for evaluating thermal selection using existing oxygen consumption and aerobic scope data for Gates Creek sockeye salmon. I predicted that sockeye salmon that had lower somatic energy or earlier migration timing would prefer cooler lake temperatures. As female sockeye salmon invest more somatic energy into reproductive maturation than males, I predicted that females would select cooler thermal habitats. Additionally, I predicted that fish that experienced more stressful migratory conditions (i.e. higher water temperatures, more challenging Seton Dam tailrace flows) would select temperatures closer to ToptAS for physiological recovery. 3.2 Methods 3.2.1 Study System Adult Gates Creek sockeye salmon swim ~315 kilometres up the Fraser River from the Pacific Ocean, arriving at the Seton River near the town of Lillooet, British Columbia, throughout the month of August. Five kilometres upstream of the Seton River-Fraser River confluence, migrants pass the Seton Dam via a vertical-slot fishway. Fish continue their migration through Seton Lake (~22 km), Portage Creek (~3 km), and Anderson Lake (~21 km) to reach spawning grounds (Figure 2.1). Seton and Anderson lakes have mean depths of 85 m  56 (maximum 151 m) and 140 m (maximum 215 m, Geen and Andrew 1961), respectively. Both are meso-oligotrophic and thermally stratified between May and October (Limnotek 2015). Seton Lake is cooler and more turbid than Anderson Lake, due to a hydroelectric diversion of glacial water from the adjacent Bridge River watershed that enters at the lake’s northwest end (Geen and Andrew 1961). Adults spawn in September in Gates Creek and a creek-fed artificial spawning channel near D’Arcy, British Columbia (Figure 3.1A).  3.2.2 Fish collection & tagging To capture fish, a full-spanning picket fence and fish trap were installed in the Seton River ~400 m downstream of the Seton Dam (Figure 3.1B). The fence was closed for 11 hours each day throughout the run and left open each night. From Aug 16-Sept 2, 2013, 51 males and 66 females were tagged, and from Aug 5-Sept 7, 2014, 71 males and 95 females were tagged. The tagging process followed Cooke et al. (2005). Fish were netted from the trap box into a holding pen with constantly flowing river water. Individual fish were netted from the pen and placed, supine and unanesthetized, in a foam-lined, V-shaped trough with river water pumped over the gills. A handheld energy probe was used to estimate somatic lipid concentration (FM 692 Fish Fatmeter, Distell Inc., West Lothian, Scotland, UK; see Crossin and Hinch 2005) and to distinguish Gates Creek sockeye salmon from co-migrating populations traveling to spawning grounds farther upstream in the Fraser River that may have strayed into the Seton River and been captured at the fence (further details in Casselman et al. 2012; Bett and Hinch 2015). For each Gates Creek fish, fork length (FL) was measured, sex was determined by secondary sexual characteristics, and a 3-mL blood sample was taken from the caudal vein for analyses not described in this paper. A 32-mm half-duplex (HDX) passive integrated transponder (PIT) tag (Oregon RFID, Portland, OR, USA) was inserted in the dorsal  57 musculature, and a radio transmitter (Pices5 model, 3s burst rate, Sigma Eight Inc., Newmarket, ON, Canada), which was assigned a unique code on one of ten radio frequencies, was inserted into the stomach with a plastic plunger. Each radio transmitter had an archival temperature logger (iButton Thermochron model DS1921Z or DS1922L, Maxim Integrated, San Jose, CA, USA) glued to the non-antenna end, following Donaldson et al. (2009). Fish were externally marked with a 12” spaghetti tag (Floy Manufacturing, Seattle, WA, USA) printed with a telephone number and the study name. For each fish, a sample of adipose fin was removed with a hole punch for DNA population identification (Beacham et al. 2005). Each fish was then photographed and returned to the holding pen. For a parallel study, which assessed behaviour and survival of tagged fish downstream of the Seton Dam (Casselman et al. 2014), fish were released on the west (N 50.6621, W 121.9184; in 2013 and 2014) and east (N 50.6611, W 121.9139; in 2013 only) banks of the Fraser River, ~2.5 km downstream of the Seton River confluence (Figure 3.1). Groups of 10-12 tagged fish were driven ~20 minutes to release sites in a 1,000-litre transport tank of aerated Seton River water. The University of British Columbia Animal Care Committee approved all capture, tagging, sampling, and transport procedures, under Canadian Council on Animal Care guidelines.  Tagged fish were detected on stationary radio receivers (Orion, Sigma Eight Inc. or SRX-400, Lotek Wireless Inc., Newmarket, ON, Canada, with 3-element Yagi antenna) at the tailrace and in the fishway entrance of Seton Dam, at the mouth of Portage Creek (in 2014 only), and at the entrance to the spawning channel in Gates Creek (n=1). Pass-through PIT antennas (Oregon RFID) were installed inside the entrance and exit basins of the Seton Dam fishway, 120 m upstream of the mouth of Gates Creek (in 2014 only), and in the entrance gate at the spawning channel (Figure 3.1). Detection ranges of radio receivers were tested at  58 installation, while PIT antennas were tested daily (at the Seton Dam) or weekly (in Gates Creek). Detections from radio and PIT receivers were pooled by location (i.e. Seton Dam, Portage Creek, Gates Creek), and the time of first and last detection at each location was identified to determine migration duration (in days) of each fish within Seton and Anderson lakes. See ‘Fish temperature data’ below for details on Portage Creek detections. 3.2.3 Fish temperature data Archival temperature loggers were programmed to record temperature data at 15-minute intervals and had resolutions of 1/16 °C (accuracy: ± 0.5 °C, range: -10 °C to +65 °C; model DS1922L) or 1/8 °C (accuracy: ± 1 °C, range: -5 °C to +26 °C; model DS1921Z). Through a combination of carcass collection and manual radio tracking in the Seton River, Gates Creek and the artificial spawning channel, 133 out of 283 radio tags with temperature loggers were recovered. Thirteen temperature loggers were recovered from carcasses in the Seton River below the dam; these fish did not experience lake temperatures and will not be discussed further. Four fish that died following dam passage were removed from the sample due to incomplete lake temperature data: one logger was recovered in Seton Lake, two from fish that were captured in Portage Creek, and one in Anderson Lake. Two additional males were removed from the dataset because the fat probe malfunctioned during tagging. Temperature data were trimmed to exclude records from before the date and time of last telemetry detection at Seton Dam and after the date and time of first telemetry detection at Gates Creek. The date and time of last detection on the radio receiver at the mouth of Portage Creek in 2014 was used to distinguish Seton Lake and Anderson Lake temperature data. There was a consistent Portage Creek visual signature in the temperature data (see Figure A1.1 in Appendix 1), and among 2014 fish, the warmest temperature experienced coincided roughly  59 with the date and time of the last Portage Creek telemetry detection for all but five fish (n=84). The highest recorded temperature was used as an estimate of Portage Creek entry for 2013 fish (and one 2014 fish that was not detected on the receiver); the estimate was checked for proximity to the Portage Creek visual signature, which was consistent across years. The estimated date and time were then used to partition temperature data between lakes and to calculate migration duration within lakes.  All temperature records were rounded to the nearest 1/10 °C. Histograms with 1 °C, right-closed bins were plotted for all temperature data from each fish in each lake and visually inspected (examples in Figure 3.2; data for all fish shown in Figure A1.2). To assess temperature preference, the mode (i.e. most common) temperature value was determined for each fish in Seton Lake and in Anderson Lake using R (version 3.2.0, R Core Team 2015). If a fish had more than one mode per lake (e.g. Fish 756 recorded 11.8 °C and 12.1 °C an equal number of times in Anderson Lake), the mean of these values was used.  3.2.4 Physiological framework In order to determine the relative benefit of a given temperature for tagged fish, I developed a framework that considers the trade-off between aerobic capacity and maintenance costs of living for Gates Creek sockeye salmon (Figure 3.3).  The oxygen consumption of adult Gates Creek sockeye salmon (n=37) was previously measured in a swim tunnel respirometer, across a range of water temperatures (i.e. 10.9-22.7 °C) by Lee et al. (2003a), and curves were fit to the data by Eliason et al. (2011). I used the second-order polynomial relating absolute aerobic scope (AAS) to temperature (t), AAS=-20.9124+3.8926*t-0.1184*t2 and the exponential curve for routine oxygen consumption (MO2routine), MO2routine=1.7464*exp(0.0448*t) (Eliason et al. 2011), to predict AAS and  60 MO2routine at temperatures between 4.8-26.1 °C. To normalize the benefit of aerobic scope, I set the numerical value of AAS at ToptAS (i.e. 11.08 mg O2 kg-1 min-1) to 1. Absolute aerobic scope is 0 at Tcrit (i.e. 6.8 °C, 26.1 °C; Eliason et al. 2011), and intermediary AAS values were expressed as a proportion of maximum AAS. To normalize the value of MO2routine, I calculated MO2routine at temperatures from 5.0-26.1 °C, using the equation from Eliason et al. (2011), and divided all values by 5.62 mg O2 kg-1 min-1 (i.e. MO2routine at 26.1 °C). To represent the high cost of routine metabolism at high temperatures, I created the inverse of the curve presented in Eliason et al. (2011) by subtracting all MO2routine values from 1. At temperatures near ToptAS (16.4 °C), the relative benefit of aerobic scope exceeds the relative benefit of routine metabolism. Both aerobic scope and routine metabolism have zero benefit at the upper Tcrit, because survival is time-limited. Temperatures below ~10 °C, where the relative benefit of aerobic scope falls below MO2routine (Figure 3.3), confer greater benefit for routine metabolism because maintenance costs are low.  To assess the use of thermal habitat below 10°C, I calculated the proportion of time in each lake that each fish spent at temperatures <10°C [i.e. # of records <10 °C in Seton Lake / total # of temperature records in Seton Lake]. 3.2.5 Environmental conditions For each fish, the temperature and flow regime during detection at the Seton Dam were used to characterize the dam passage experience. Exposure to high temperatures and complex flow dynamics in the Seton Dam tailrace has been associated with sockeye salmon mortality following fishway passage (Roscoe et al. 2010b Burnett et al. 2014a, 2014b). From Aug 8-19, 2014, the flow conditions at Seton Dam were altered to study the effect of dam operations on fish passage (Burnett et al. In press). Rather than spilling water from a siphon directly beside  61 the fishway entrance, as was done in 2013, water was spilled from a siphon approximately 10 m from the fishway entrance, into the middle of the dam tailrace. Under standard flow conditions, high encounter velocities and complex, turbulent flows surrounded the fishway entrance. Under altered flow conditions, encounter velocities were lower around the fishway entrance, and reverse flow fields may have helped guide fish towards the entrance (Casselman et al. 2015). The operational change substantially altered the tailrace flow dynamics (Casselman et al. 2015) and increased total Seton Dam discharge by up to 25% (measured on the Water Survey of Canada gauge in the Seton River; Figure 3.1B). Tailrace flow dynamics affect swimming effort and contribution of anaerobic metabolism for salmon locating fishway entrances (Burnett et al. 2014a), so tagged fish were assigned to a category (0=‘low’ discharge [i.e. standard flow conditions]; 1=‘high’ discharge [i.e. altered flow conditions]), based on the date of first detection at Seton Dam in 2014. All fish tagged in 2013 were assigned to category 0. The highest temperature recorded on a fish’s archival temperature logger between first and last detections at Seton Dam was also used as a proxy for dam passage stress.  The highest temperature recorded on each fish’s archival temperature logger, within 24 hours following the last telemetry detection at Portage Creek, was used to characterize the thermal experience in Portage Creek.  The locations of the thermoclines in Seton and Anderson lakes, in August and September 2014, were measured by Limnotek Research and Development Inc. Continuous temperature at depth profiles were collected from six stations along a longitudinal transect of Seton Lake and two stations in the eastern half of Anderson Lake, on Aug 20 and Sept 24, 2014, using a SBE 19plus V2 CTD profiler (Sea-Bird Electronics, Bellevue, WA, USA), and  62 data were interpolated between stations and sampling dates. Station locations are described in Limnotek (2015).  3.2.6 Statistical approach The temperature data were analyzed using four statistical models, with one for each of the response variables (preferred temperature, proportion of time at temperatures below 10 °C) in each lake. Models 1 and 2 had the response variable of preferred temperature (i.e. mode, °C) in Seton Lake and Anderson Lake, respectively. Both global models included the explanatory variables (1) year, to account for inter-annual differences in fish condition and migration experience; (2) sex (male [0], female [1]), to test the effect of sex; (3) GSE (MJ kg-1), to test the effect of somatic energy density; and (4) time in the respective lake (d), to test the effect of migration duration. Model 1 included three additional variables: (5) tagging date, to account for differences in fish condition and migration experience with run timing; and (6) Seton Dam discharge (low [0] or high [1]) and (7) maximum Seton Dam water temperature (Tdam, °C), to test the effects of migratory stressors. Model 2 included one additional variable: (4) maximum Portage Creek temperature (TPC, °C), to test the effect of a potential migratory stressor. Tagging date was omitted from Model 2 because it was highly correlated with Portage Creek temperatures. Models 1 and 2 were fitted with linear models. However, visual assessment of model residuals revealed violation of heteroscedasticity in model 2, so a generalized least squares (GLS) model was fitted, and variance structure selection was conducted, using AIC, on residuals of parameters contributing to heteroscedasticity (i.e. sex; Zuur et al. 2009)).   Models 3 and 4 had the response variable of proportion of time at temperatures <10 °C in Seton Lake and Anderson Lake, respectively. Model 3 included the same explanatory variables as Model 1, and Model 4 included the same explanatory variables as Model 2 (Table  63 3.1). A generalized linear model (GLM; family: Gamma, link=log) was fitted to each dataset, using Tweedie model distributions for positive, continuous, zero-inflated response variables (Dunn 2014). All variables were assessed for correlation (all r < |0.7|, Zuur et al. 2010) and multicollinearity (all variance inflation factors < 4, O’Brien 2007). Continuous variables were standardized, with a mean of 0 and standard deviation of 2 (Gelman 2008, Schielzeth 2010). Response variables were not transformed (Zuur et al. 2010). Residuals from each model were visually inspected for normality, independence, and heteroscedasticity, following Zuur et al. (2009). For Models 1 and 2, candidate models were generated using the ‘MuMIn’ package from all combinations of variables in the global model and compared using AICc for small sample sizes (Burnham and Anderson 2002, Barton 2012). The lowest AICc score identifies the most parsimonious model, while models with ΔAICc (i.e. difference in AICc scores) of <2, 4-7, and >10 have substantial, considerably less, and virtually no support from the data, respectively (Burnham and Anderson 2002). The AICc weight (wi) is the probability that a given model is the most parsimonious description of the data (Burnham and Anderson 2002). Top model weights were small (e.g. 0.36 for Model 1; Table 3.1), so the models with cumulative weights summed to ≥0.95 were used to form the confidence set. The confidence set of models was averaged using the ‘natural average’ method to generate coefficient estimates and 95% confidence intervals for explanatory variables (Burnham and Anderson 2002, Grueber et al. 2011). Global models are presented for Models 3 and 4. Adjusted r2 values are presented for the confidence sets for Models 1 and 2, and model fit for Models 3 and 4 was assessed with the Hosmer-Lemeshow pseudo r2 (Table 3.1).   64 A two-sample t-test was used to compare GSE between years, and two-sample Kolmogorov-Smirnov tests were used to compare non-normal distributions of mode temperatures and lake residence times. Raw data are presented as mean ± SE, and statistical significance was evaluated at 0.05. All statistical analyses were conducted with R software (version 3.2.0, R Core Team 2015). 3.3 Results I analyzed temperature data from 114 sockeye salmon that migrated successfully to Gates Creek: 27 fish tagged in 2013 (11 males, 16 females) and 87 tagged in 2014 (40 males, 47 females). Temperature loggers recovered from tagged fish had an average (± SE) of 903 ± 49 temperature records (range: 228-3294 records) archived in Seton and Anderson lakes.  The mean tagging dates were Aug 21, 2013 and Aug 20, 2014 (Table 3.2). Earlier migrants arrived in the Seton River with higher somatic energy stores than later migrants, such that GSE declined with tagging date in both years (2013: r=-0.53, p=0.004; 2014: r=-0.6, p<0.0001; Figure 3.4A). Fish arrived with significantly higher somatic energy stores in 2013 (mean ± SE: 6.1 ± 0.1 MJ kg-1) than in 2014 (5.7 ± 0.1 MJ kg-1; two-sample t-test, t41=3.12, p=0.003). Migrants with higher somatic energy took longer to migrate through Seton Lake (r2=0.38, p<0.001; Figure 3.4D) and Anderson Lake (r2=0.35, p<0.001; Figure 3.4E).  3.3.1 Receiver efficiency At the Seton Dam, detection efficiencies of individual receivers were moderate to high [radio receivers: tailrace (87%, n=120) and fishway (2014 only: 45%, n=89); PIT antennas: fishway entrance (94%, n=120) and exit (92%, n=120)], but all fish (n=120) had multiple detections on at least one receiver when data were pooled. The detection efficiency of the  65 Portage Creek receiver was 99% (2014 only: n=88). At Gates Creek, detection efficiencies of receivers varied widely [radio: channel entrance (96%, n=116); PIT antennas: lower Gates Creek (2014 only: 35%, n=85) and channel entrance (66%, n=116)]. One fish that survived to Gates Creek was not detected on any receiver, so date and time of arrival was determined from temperature data.  3.3.2 Dam passage During Seton Dam passage (i.e. the period between first and last detections on telemetry receivers at the Seton Dam), fish experienced water temperatures between 14.8-20.5 °C (Figure 3.5A), according to recovered temperature loggers. Water temperature at Seton Dam was negatively correlated with Julian date in 2013 (r=-0.6, p<0.001), but not significantly in 2014 (r=-0.2, p=0.106), and fish experienced high temperatures (≈20 °C) in both years.  All 2013 fish (n=27) and 59 of the 2014 fish passed Seton Dam under standard flow conditions, where water was released from a siphon directly adjacent to the fishway. Under standard flow conditions (i.e. ‘low’ discharge), hourly discharge was 24.6-29.5 m3s-1 in 2013 and 25.4-27.6 m3 s-1 from Aug 20-Sept 7, 2014 (Figure 3.5B). Twenty-eight of the 2014 fish experienced altered flow conditions (i.e. ‘high’ discharge; 31.0-33.1 m3 s-1), when water was released from a siphon ~10 m from the fishway entrance, from Aug 8-19, 2014 (Figure 3.5B). Discharge from Seton Dam was negatively correlated with Julian date in both years (2013: r=-0.83, p<0.0001; 2014: r=-0.65, p<0.0001). 3.3.3 Seton Lake In August and September 2014, the depth and temperature of the Seton Lake thermocline were 20-25 metres and 9-12 °C, respectively. Surface temperatures were  66 approximately 22 °C and 16 °C in August and September, respectively. Below ~30 metres, water was < 6 °C (Limnotek 2015). Sockeye salmon migrated through Seton Lake more quickly in 2013 (2.7 ± 0.4 days, range: 0.8-9.7 days) than in 2014 (4.6 ± 0.4 days, range: 0.5-15.4 days; Kolmogorov-Smirnov two-sided test, D=0.354, p-value=0.012; Figure 3.6A). In 2013, mean migration rates were 13.7 ± 1.6 km day-1 (range: 2.3-28.2 km day-1), compared to 10.0 ± 1.1 km day-1 (range: 1.4-42.5 km day-1) in 2014. Fish that were tagged earlier in the run spent more time in Seton Lake than fish tagged later (r=-0.5, p<0.0001; Figure 3.4B). In 2013, the temperature preference (i.e. temperature mode) of individual fish in Seton Lake was between 6.2 °C and 18.6 °C. In 2014, fish preferred temperatures between 5.1 °C and 19.4 °C (Table 3.2). The distribution of temperature modes in Seton Lake was bimodal, with peaks at 10.5-11.5 °C and 15-16 °C (Figure 3.7A), and use of these temperatures was related to the dam passage experience and time spent in the lake. Migrants that experienced ‘low’ discharge at Seton Dam preferred cooler temperatures than fish that experienced ‘high’ discharge. Fish that experienced higher temperatures during dam passage had cooler temperature modes (Figure 3.8A). Longer migration duration in Seton Lake was associated with cooler preferred temperatures, as well as a greater proportion of time spent below 10 °C (Figure 3.8B; Figure 3.9C). In 2013, 52% (n=14) of migrants occupied temperatures below 10 °C for 2.5-67.5% of time in Seton Lake. In 2014, 66% (n=57) of fish spent time below 10 °C, and the duration of use ranged from 0.4-92.1% of the migration through Seton Lake (Table 3.2; Figure 3.9). All models in the confidence set for Model 1 included time in Seton Lake and both dam passage parameters, and were weakly supported by the data (i.e. ΔAICc <4). The top model  67 included no other variables (adjusted R2 = 0.36; Table 3.1), and the direction of the effect of all other explanatory variables was uncertain, as confidence intervals crossed zero (Figure 3.10A). In Model 3, only time in Seton Lake had a significant effect on the proportion of time that fish spent below 10 °C (Figure 3.10C).  3.3.4 Portage Creek Sockeye salmon entered Portage Creek between Aug 15-Sept 10 in 2014, based on telemetry detections on the Portage Creek receiver. Estimates of Portage Creek entry for 2013 fish, based on archival temperature logger data, were between Aug 12-Sept 9.  Fish experienced temperatures between 18.7-21.0 °C in Portage Creek in 2013 and between 15.6-22.7 °C in 2014 (Figure 3.5C), and fish that were tagged earlier experienced warmer temperatures than fish tagged later, across years (r=-0.8, p<0.0001).  3.3.5 Anderson Lake In Anderson Lake, the depth and temperature of the thermocline were approximately 10-20 metres and 10-17 °C, in August, and 15-20 metres and 11-16 °C, in September. Surface temperatures were approximately 22 °C and 16 °C in August and September, respectively, and water temperature was <6 °C at depths greater than ~40 metres (Limnotek 2015). There was no difference in the time that fish spent in Anderson Lake in 2013 and 2014 (two-sample Kolmogorov-Smirnov test, D=0.119, p=0.933; Figure 3.6B). In 2013, fish migrated through the lake in an average of 5.1 ± 0.5 days (range: 1.3-13.6 days; corresponding migration rates: 6.5 ± 0.8 km day-1, range: 1.8-18.5 km day-1). In 2014, the mean time spent in Anderson Lake was 5.0 ± 0.3 days (range: 1.0-17.0 days; corresponding migration rates: 7.1 ±  68 0.5 km day-1, range: 1.4-23.5 km day-1). Like in Seton Lake, fish that were tagged earlier in the run spent more time in Anderson Lake than fish tagged later (r=-0.5, p<0.0001; Figure 3.4C). Tagged fish exhibited a strong preference for temperatures between 10.5-12 °C in Anderson Lake, and the distribution of temperature modes differed significantly from the distribution in Seton Lake (two-sample Kolmogorov-Smirnov test, D=0.421, p<0.0001; Figure 3.7). Fish preferred cooler temperatures in Anderson Lake in 2013 (range: 5.1-12.1 °C) than in 2014 (range: 4.9-19.6 °C; Table 3.2; Figure 3.8C). All models in the confidence set for Model 2 included year, and the top three models had considerable support from the data (i.e. ΔAICc < 2; Table 3.1). The directions of the effects of all other explanatory variables in the average model were uncertain (Figure 3.10B).  Use of thermal habitats below 10 °C in Anderson Lake differed between years (Figure 3.10D). All of the fish tagged in 2013 occupied temperatures below 10 °C, and use was between 8.9% and 84.4% of the total time that individuals spent in the lake. In 2014, 67% of fish occupied temperatures below 10 °C; these fish spent between 0.2% and 78.7% of time in Anderson Lake at these temperatures (Table 3.2). Female fish spent a greater proportion of time at temperatures below 10 °C than did males (Figure 3.9D,E; Figure 3.10D).  3.4 Discussion This is the first study to interpret temperature selection by adult Pacific salmon in thermally heterogeneous environments in the context of the aerobic scope available at particular temperatures. Gates Creek is one of few Pacific salmon populations with known relationships between routine metabolic oxygen consumption, aerobic scope, and temperature (Lee et al. 2003a, Eliason et al. 2011), which we used to infer routine metabolic cost and capacity for aerobic activity from 114 recovered archival temperature loggers. Thermal  69 selection reflects trade-offs between temperatures that maximize different physiological processes (Brett 1995). In fishes, the adaptive importance of patterns of thermal behaviour (e.g. diel vertical migration) is generally attributed to growth maximization (i.e. partition time between habitats where metabolic costs are minimized and feeding and assimilation rates are maximized, Brett 1971) and mortality minimization (i.e. avoid prolonged exposure to stressful conditions or habitats with increased probability of predation, Clark and Levy 1988, Bevelhimer and Adams 1993). The ecological constraints differ for adult Pacific salmon in freshwater, because individuals are not feeding and somatic energy is limited. Predation risk is also limited in Seton and Anderson lakes. I created a physiological framework to identify temperatures that conferred a greater relative metabolic benefit to aerobic processes, such as swimming activity, reproductive maturation, and recovery from stress, and those where the costs of routine metabolism were minimized. At temperatures near 16.4 °C (i.e. ToptAS), Gates Creek sockeye salmon have the greatest scope for oxygen-dependent physiological functions and between 13.4-19.5 °C, they have at least 90% of maximum aerobic scope (Eliason et al. 2011). Conversely, at temperatures below 10 °C, the relative benefit of energy conservation exceeds the capacity for aerobic activity, and fish have only 60% of maximum aerobic scope available.  The sockeye salmon tagged in this study differed in thermal preference between individuals and lakes. There was much greater variation exhibited in Seton Lake than in Anderson Lake. In Seton Lake, the distribution of mode temperatures was bimodal, with approximately equal numbers of fish preferring temperatures ~10.5-11.5 °C and ~15-16 °C. In Anderson Lake, the vast majority of fish had a temperature mode of ~10.5-12 °C. Thermal preference in Seton and Anderson lakes suggest that fish were ‘trading off’ between cooler  70 temperatures that prioritize energy conservation and temperatures near ToptAS that maximize aerobic scope, as preference for temperatures that optimize either metabolic process was not common. The vast majority of thermal habitat in Seton and Anderson lakes is at temperatures below 10 °C in August and September, yet the preferred temperature of relatively few fish was in the hypolimnion of the lakes. Conversely, only a small proportion of fish had a temperature mode near ToptAS, and this was very rare in Anderson Lake. Adult sockeye salmon in lakes frequently use temperatures that are cooler than optimum, but not the coolest available (e.g. Lake Washington: mode 9-11 °C, Newell and Quinn 2005; Gates Creek: mean 12.3-15.9 °C, Roscoe et al. 2010; Shuswap complex: mean 11-15 °C, D.A. Patterson, Fisheries and Oceans Canada, unpublished data). Preference for lower than optimum temperatures has also been observed in lake resident salmonid species (e.g. lake trout [Salvelinus namaycush], Bergstedt et al. 2003, Mackenzie-Grieve and Post 2006), and is common when food is limited (e.g. juvenile sockeye salmon, Biette and Geen 1980; lake trout, Mac 1985).  In Seton Lake, thermal preference was influenced by metrics of dam passage experience. Sockeye salmon that passed Seton Dam when discharge was high (i.e. during Aug 8-19, 2014, when flow conditions were altered) preferred warmer temperatures in Seton Lake than fish that experienced low discharge (i.e. in 2013 and during Aug 20-Sept 7, 2014). Exposure to warmer water temperatures in the tailrace and fishway during dam passage was associated with preference for cooler temperatures in Seton Lake. A temperature mode below 10 °C was rare among fish that experienced ‘high’ discharge, but a number of fish that experienced ‘low’ discharge and water temperatures >18 °C, concurrently, preferred temperatures <10 °C in Seton Lake. Under standard flow conditions (i.e. ‘low’ discharge), Gates Creek sockeye salmon require anaerobic (‘burst’) swimming to overcome high velocities  71 and complex flow dynamics near the entrance to the Seton Dam fishway (Burnett et al. 2014a, 2014b). Greater reliance on anaerobic swimming is associated with increased probability of dam passage, but also increased probability of post-dam passage mortality, particularly when water temperatures are high (Burnett et al. 2014b). In a parallel study comparing swimming effort under the two Seton Dam flow conditions, Burnett et al. (In press) found that releasing water ~10 m from the fishway entrance reduced sockeye salmon swimming speeds in the tailrace and recovery time at the fishway exit, following dam passage. Although they were associated with higher discharge, the altered flow conditions created reverse flow fields that may have helped guide fish towards the fishway entrance (Casselman et al. 2015; Burnett et al. In press). Following anaerobic activity, Pacific salmon require substantial oxygen and recovery time to re-establish biochemical, ionic and osmotic homeostasis and restore oxygen stores, high energy phosphates and glycogen (Wood et al. 1983, Gaesser and Brooks 1984, Wood 1991). Ambient water temperature affects the rate of post-exercise oxygen consumption (EPOC; Lee et al. 2003b), but the temperature that maximizes the rate of recovery from anaerobic effort is unknown (Eliason and Farrell 2016). Fish may seek very cool temperatures (e.g. 5-10 °C, this study) following anaerobic effort to rapidly decrease routine metabolic costs and begin the process of physiological recovery; although, recovery is thought to be slow at cool temperatures. I could not identify a thermal behaviour that was most effective for recovering from physiological stress, because the swimming effort of individual fish in the dam tailrace was not measured. It remains unknown whether a particular thermal experience was associated with in-lake mortality, because archival temperature loggers were not recovered from fish that died in Seton and Anderson lakes.   72 Preference for very cool temperatures following dam passage may be a behavioural response to cumulative stress in fish with relatively poor physiological condition (e.g. prior injury, high pathogen load, progressive disease state). Thermal refuge is hypothesized to slow pathogen replication and delay disease progression (Wagner et al. 2005, Mathes et al. 2010, Bradford et al. 2010), although thermal preference of diseased sockeye salmon has not been well studied.  Preference for temperatures near ToptAS was strongly associated with short migration duration in Seton Lake (i.e. <3 days). The fish that transited the lake quickest were tagged later in the migration, an indication that sockeye salmon coordinate arrival on spawning grounds with conspecifics by migrating at different rates. In still and slow moving water, fish swim at metabolically optimal speeds [≈1 body length per second (BL s-1)], such that COT per unit body mass per unit distance is minimized (Webb 1995, Hinch and Rand 2000, Lee 2003b, Drenner et al. 2012). Gates Creek sockeye salmon are among several upper Fraser River stocks with challenging migrations that are able to maintain low COT across a broad range of swimming speeds, from 1.0 BL s-1 to approximately 70-80% of the critical swimming speed (Lee et al. 2003a, Eliason et al. 2013a). For Gates Creek sockeye salmon, this range corresponds to approximately 1.0-1.7 BL s-1 (i.e. 80% of 2.08 BL s-1, the critical swimming speed), and the temperature optimum for critical swimming speed is approximately equal to ToptAS (Lee et al. 2003a). At temperatures near ToptAS, fish have a greater proportion of maximum aerobic scope for active migration. However, Gates Creek sockeye salmon may not require much scope to migrate through lakes. The COT of swimming speeds between 1.0-1.7 BL s-1 is a small proportion of maximum aerobic scope (≈2-3%) for Gates Creek sockeye salmon, according to relationships between net COT and swimming speed determined by Lee  73 et al. (2003b). This suggests that migrants can sustain metabolically optimal swimming speeds across a range of temperatures that were available in Seton and Anderson lakes. And although sockeye salmon can exhibit mean ground speeds in reservoirs and lakes (e.g. Naughton et al. 2005, Roscoe et al. 2010a) that are roughly equivalent to migration rates in the mainstem of the Fraser River (e.g. English et al. 2005), upriver migrating fish swim at higher swimming speeds to overcome higher encounter velocities. The migration rates in this study were minimum estimates, generated from residence times in each lake. In their study of Gates Creek sockeye salmon, Roscoe et al. (2010a) distinguished between directed migration and circling or holding behaviour and found that mean migration rates (± SE) were nearly twice as fast through Seton Lake (38.4 ± 2.4 km day-1) as through Anderson Lake (17.0 ± 2.2 km day-1), and early migrants, especially females, held for longer periods in Anderson Lake, near the Gates Creek outflow. In this study, early migrants spent the most time in Anderson Lake, but males spent slightly longer than female fish, on average, and I was not able to distinguish between directed swimming and holding using my array of telemetry receivers. Consistent with predictions, Gates Creek sockeye salmon with earlier migration timing used cooler temperatures, although this was only apparent in Seton Lake. Tagged fish that spent longer in Seton Lake occupied temperatures below 10 °C for a greater proportion of time, which would have reduced the metabolic costs of persisting until peak spawning. In Anderson Lake, migrants spent a greater proportion of time below 10 °C, regardless of migration timing, or duration of residency. Use of cool temperatures by the majority of fish may explain why there was no effect of exposure to high Portage Creek temperatures on Anderson Lake thermal experience. The negative relationship between lake residence time and preferred temperature also exists across sockeye salmon populations; preferred temperature  74 decreases to compensate for energy allocated to persistence as residence time increases. Bioenergetic models predicted that temperatures below 10 °C were required for Copper Creek sockeye salmon to survive in freshwater for 130 days (Katinic et al. 2015), and fish in Lake Washington occupied 9-11 °C for over 80 days (Newell and Quinn 2005). In Kamloops Lake, sockeye salmon from the Shuswap population complex use similar mean temperatures to Gates Creek sockeye salmon (i.e. ~11-15 °C) over similar residence times (i.e. 2-23 days, D.A. Patterson, Fisheries and Oceans Canada, unpublished data).  Unexpectedly, energy density was not an important predictor of thermal behaviour in any model in this study, despite evidence that fish were using temperatures that provide a greater benefit to energy conservation than aerobic processes. Indeed, energy may not have been limiting for the fish tagged in 2013 and 2014. On average, individuals had higher somatic energy levels at tagging (5.5-6.1 MJ kg-1) than migrants from other Fraser River sockeye salmon populations have at spawning ground arrival (≈4.0-5.0 MJ kg-1; Crossin et al. 2004b), and considerably more than the approximate energetic threshold to sustain life (i.e. 4.0 MJ kg-1, Crossin et al. 2003, 2004b, Hruska et al. 2010). The earliest fish arriving in the Seton River had the highest levels of somatic energy, so the effect of GSE might have been clearer if early, middle, and late migrants, with similar somatic energy levels, were compared. Additional analysis is required to determine the amount of somatic energy required to complete the migration to Gates Creek; however, results suggest that fish had adequate somatic energy and thermal refuging behaviour was not necessary for survival. In years that fish arrive at the Seton Dam with lower somatic energy levels (i.e. following poor marine feeding conditions, Crossin et al. 2004a), I predict that between-individual differences in thermal behaviour will be directly related to GSE. In this study, there were a number of differences between fish tagged in 2013  75 and 2014. In 2013, fish had higher GSE values, migrated more quickly through Seton Lake, and preferred cooler temperatures in Anderson Lake. All fish tagged in 2013 occupied temperatures less than 10 °C, and the average time below 10 °C was 38.6 ± 6.0 hours (range: 1-120 hours). The range of tagging dates differed slightly between years, and the difference was exacerbated by high mortality among tagged fish in 2013. As a result, the fish from which archival temperature loggers were recovered were not distributed equally across the run in both years, and both somatic energy level and migration duration vary strongly with run timing. Additionally, migrants in the two years experienced different migration conditions. Temperatures were extremely high in 2013, both in the Fraser River mainstem (>18 °C at Qualark, D.A. Patterson, Fisheries and Oceans Canada, unpublished data) and in the Seton River during tagging (>18 °C, Casselman et al. 2014), which may have caused faster migration rates (Naughton et al. 2005, Keefer et al. 2008) and greater preference for cool temperatures in Anderson Lake.  Female Gates Creek sockeye salmon spent a greater proportion of time in Anderson Lake at temperatures below 10 °C than males, which was consistent with my prediction. Female Pacific salmon allocate considerable energy to gamete development during the freshwater migration (Rand and Hinch 1998, Crossin et al. 2004b), yet reproductive success is strongly related to reproductive longevity (i.e. an individual’s lifetime on spawning grounds; Hruska et al. 2011; Burnett et al. In press), so females may use cool temperatures to conserve energy for survival and reproductive behaviours on spawning grounds. Water temperature also affects the rate of vitellogenesis (Pankhurst and King 2010). The optimum temperature for vitellogenesis is thought to be at (Eliason and Farrell 2016) or slightly cooler than ToptAS  76 (Macdonald et al. 2000). Mature females that prefer cool temperatures (e.g. Roscoe et al. 2010a) may be attempting to prevent over-development of oocytes (Flett et al. 1996).  The migration temperatures that sockeye salmon selected corresponded to depths of approximately 10-30 m in both lakes and were strongly associated with the thermocline. Seton Lake has a slightly deeper thermocline, due to the inflow of water from the glacial Bridge River at the northwest end of the lake. Fish that used temperatures of ~15-16 °C in Seton Lake swam at depths of ~10-15 m, which was just above the thermocline. Fish that used temperatures of ~10.5-12 °C traversed the thermocline, at depths of ~20-25 m and 15-20 m in Seton and Anderson lakes, respectively. Association with the thermocline may provide olfactory cues, as sockeye salmon navigate to spawning grounds. In coastal and lake environments, sensory impairment of Atlantic salmon (Salmo salar) causes vertical migrations throughout the water column, whereas unimpaired fish prefer thermally stratified areas and display neuronal activity indicative of olfactory discrimination (Westerberg 1982, Døving et al. 1985). Marine migration depth of adult sockeye salmon homing to the Fraser River has also been attributed to stratification of olfactory cues in the water column (Quinn 1988, Drenner et al. 2014).  During the reproductive migration semelparous Pacific salmon partition available aerobic scope to swimming activity, navigation behaviours, reproductive maturation, stress responses, and immune function, yet the thermal optima for many of these processes are unknown (Clark et al. 2013, Eliason and Farrell 2016). Sockeye salmon thermal preference in the present study was related to prior migratory stressors, including exposure to high temperature and high, complex flows, and immediately prior to spawning, females spent a greater proportion of time at very cool temperatures. These results underscore the importance  77 of thermally heterogeneous environments for multiple important physiological processes in Pacific salmon en route to spawning grounds. As summer water temperatures in the Fraser River increase over the next several decades (Morrison et al. 2002, Hague et al. 2011), thermoregulatory behaviours will be increasingly important to the continued persistence of sockeye salmon in the watershed.   78  Figure 3.1 – The Seton-Anderson watershed and its location in British Columbia, Canada (inset); Gates Creek and the artificial spawning channel (A); and the Seton Dam and fishway (B). The location of the Bridge River inflow to the northwest end of Seton Lake is indicated. R = radio receiver, P = pass-through PIT antenna, F = fish collection fence, D = Water Survey of Canada discharge gauge, W = spawning channel weir, white star = release site (2013 and 2014), filled star = release site (2013 only). Detections on telemetry receivers in A were grouped as ‘Gates Creek’ and in B as ‘Seton Dam’ for analysis.   Fraser&Riverkilometres0&&&2&&&4RRFSeton&RiverSeton&DamBmetres0&&&&&&20&&&&&40fishwayPPGates&CreekPPRspawning&channelA metresW0&&100&&200RRDBridge&RiverinflowSeton&LakeAnderson&LakePortage&Creek SetonRiver 79  Figure 3.2 – Frequency histograms of the temperatures experienced in Seton and Anderson lakes by three sockeye salmon: a 2013 female (A; GSE=7.04 MJ kg-1) and male (B; GSE=4.99 7.04 MJ kg-1), and a 2014 female (C; GSE=4.45 MJ kg-1). Temperature bins are 1°C and right-closed, and the temperature mode in each lake is identified with a vertical dashed line.   subset(temp, Lake == "Seton" & FishID == 7145)$TempFrequency0 5 10 15 20 25050100150Seton Lakesubset(temp, Lake == "Anderson" & FishID == 7145)$TempFrequency0 5 10 15 20 25050100150200250300 AAnderson Lakesubset(temp, Lake == "Seton" & FishID == 7148)$TempFrequency0 5 10 15 20 2501020304050Number of temperature recordssubset(temp, Lake == "Anderson" & FishID == 7148)$TempFrequency0 5 10 15 20 25050100150200 Bsubset(temp, Lake == "Seton" & FishID == 397)$TempFrequency0 5 10 15 20 25010203040subset(temp, Lake == "Anderson" & FishID == 397)$TempFrequency0 5 10 15 20 25020406080 CTemperature (°C) 80    5 10 15 20 25 3000.20.40.60.81Temperature (°C)Relative metabolic benefit 81 Figure 3.3 – The relative metabolic benefit of temperature to aerobic capacity (solid black and dotted line) and the maintenance cost of living (solid grey line) for Gates Creek sockeye salmon. Aerobic capacity values were determined from the equation, aerobic scope=-20.9124+3.8926*t-0.1184*t2 (Eliason et al. 2011), and normalized as a percentage of maximum aerobic scope (11.08 mg O2 kg-1 min-1). The dotted portion of the line indicates where the curve has been extrapolated beyond experimental oxygen consumption data. The optimum temperature for aerobic scope (ToptAS) is 16.4 °C (dashed black line), and critical temperatures (Tcrit; AAS=0) are 6.8 °C and 26.1 °C (zeros of the aerobic scope curve). Routine oxygen consumption (MO2routine) values were determined from the equation, MO2routine=1.7464*exp(0.0448*t) (Eliason et al. 2011), normalized as a percentage of MO2routine at 26.1 °C (5.62 mg O2 kg-1 min-1), and subtracted from 1, to generate the maintenance cost of living curve. At temperatures below the intersection of the two curves (i.e. 10 °C; dashed grey line), the relative benefit to maintenance costs exceeds the relative benefit to aerobic capacity.   82 Table 3.1 – Confidence model sets (models with cumulative Akaike weights ≥ 0.95) from multimodal inference of variables influencing adult sockeye salmon temperature mode in (1) Seton Lake (SL) and (2) Anderson Lake (AL), and global models of proportion of time at temperatures <10 °C in (3) SL and (4) AL.  Response Model K logLik AICc ΔAICc wi r2 1. Temperature mode,  Seton Lake time in SL + Tdam + Discharge 3 -273.0 557 0.00 0.36 0.34  time in SL + Tdam + Discharge + GSE 4 -272.6 558 1.47 0.18 0.34  time in SL + Tdam + Discharge + tagging date 4 -272.7 558 1.73 0.15 0.33  time in SL + Tdam + Discharge + sex 4 -273.0 559 2.16 0.12 0.33  time in SL + Tdam + Discharge + year 4 -273.0 559 2.23 0.12 0.33  time in SL + Tdam + Discharge + GSE + sex 5 -272.5 560 3.46 0.07 0.33  Null: intercept 0 -298.0 600 44.0 0.00 0.00 2.  Temperature mode, Anderson Lake year + sex 2 -243.2 497 0.00 0.31 0.15  year  3 -244.6 498 0.57 0.24 0.13  year + sex + time in AL 1 -243.1 499 1.96 0.12 0.15  year + sex + TPC 3 -243.1 499 2.03 0.11 0.15  year + sex + GSE 3 -243.1 499 2.05 0.11 0.15  year + GSE 2 -244.2 499 2.08 0.11 0.14  Null: intercept 0 -249.3 505 51.3 0.00 0.00          3. Proportion of Seton Lake migration at temperatures <10°C Global: year + tagging date + sex + GSE + Discharge + Tdam + time in SL 7 NA NA NA NA 0.40  Null: intercept 0 NA NA NA NA 0.00 4. Proportion of Anderson Lake migration at temperatures <10°C Global: year + sex + GSE + TPC + time in AL  5 NA NA NA NA 0.18  Null: intercept 0 NA NA NA NA 0.00 Note: K = number of parameters in the model, logLik = model log likelihood, ΔAICc = difference in AICc from top model, wi is the AICc model weight, Tdam = highest temperature experienced at Seton Dam, TPC = highest temperature experienced in Portage Creek,  83 Discharge = Seton Dam discharge experienced (categorical: 0=‘low’, 1=‘high’), GSE = gross somatic energy, r2 = adjusted-r2 for models 1 and 2, Hosmer-Lemeshow pseudo-r2 for models 3 and 4.     84 Table 3.2 – Tagging date, GSE (MJ kg-1), migration duration (days), temperature mode (°C; i.e. most common temperature recorded by an individual fish), and proportion of time at temperatures <10 °C (i.e. percentage of all temperature records from the respective lake) of Gates Creek sockeye salmon tagged in 2013 (n=27) and 2014 (n=87). Data are reported as mean ± standard error (range). Year Tagging date Sex GSE Duration (days) Temperature mode (°C) % Records <10°C  Seton  Lake Anderson Lake Seton  Lake Anderson  Lake Seton  Lake Anderson  Lake 2013 Aug 16- Sept 2  male (n=11) 6.0 ± 0.1 (5.0-6.6) 2.3 ± 0.5 (0.8-5.2) 6.2 ± 0.7 (3.1-9.4) 13.9 ± 1.3 (6.4-18.6) 10.5 ± 0.5(6.6-12.1) 15.2 ± 7%  (0-65%) 32.9 ± 7%  (1-64%)   female (n=16) 6.1 ± 0.2 (5.1-7.2) 3.1 ± 0.7 (0.8-9.7) 4.3 ± 0.5 (1.0-13.7) 12.7 ± 0.9 (6.2-16.6) 10.0 ± 0.5 (5.1-11.9) 21.3 ± 6%  (0-67%) 41.2 ± 6% (1-84%) 2014 Aug 7- Sept 7 male (n=40) 5.5 ± 0.1 (4.6-6.6) 4.6 ± 0.5 (0.5-13.2) 4.8 ± 0.4 (1.0-13.7) 12.5 ± 0.5 (5.8-18.2) 11.9 ± 0.3 (5.5-17.3) 13.7 ± 4%  (0-92%)  8.8 ± 2% (0-56%)   female (n=47) 5.9 ± 0.1 (4.5-7.1) 4.7 ± 0.5 (0.6-15.4) 5.1 ± 0.4 (1.1-17.0) 12.9 ± 0.5 (5.1-19.4) 11.3 ± 0.4 (4.9-19.6) 13.0 ± 3%  (0-61%) 19.6 ± 4% (0-78.7%) Overall   5.8 ± 0.1 (4.5-7.2) 4.2 ± 0.3 (0.5-15.4) 5.0 ± 0.3  (1.0-17.0) 12.9 ± 0.3 (5.1-19.4) 11.3 ± 0.2 (4.9-19.6) 14.6 ± 2% (0-92%) 20.1 ± 2% (0-84%) Note: GSE = gross somatic energy. Seton Lake migration distance ≈ 22 km, Anderson Lake migration distance (includes Portage Creek) ≈ 24 km. Archival temperature loggers recorded data at 15-minute intervals.  85  Figure 3.4 – Relationships between (A) tagging date and gross somatic energy (GSE; r2=-0.54, p<0.0001), (B) tagging date and time in Seton Lake (days; r2=-0.51, p<0.0001), (C) tagging date and time in Anderson Lake (days; r2=-0.47, p<0.0001), (D) GSE and time in Seton Lake (r2=0.38, p<0.001), and (E) GSE and time in Anderson Lake (r2=0.35, p<0.001) for fish tagged in 2013 (grey triangles) and 2014 (black circles).   subset(set, Year == 2013)$Daysubset(set, Year == 2013)$GSE A4.55.05.56.06.57.0GSE (MJ kg-1)IndexNULLsubset(set, Year == 2013)$Daysubset(set, Year == 2013)$SL_days B0.02.55.07.510.012.515.0Seton Lakesubset(set, Year == 2013)$GSEsubset(set, Year == 2013)$SL_days D220 230 240 250subset(set, Year == 2013)$Daysubset(set, Year == 2013)$AL_daysC0.02.55.07.510.012.515.0Tagging dateAnderson LakeTime in lake (days)subset(set, Year == 2013)$GSEsubset(set, Year == 2013)$AL_days E4.5 5.0 5.5 6.0 6.5 7.0GSE (MJ kg-1) 86  Figure 3.5 – The highest temperature (A) experienced by individual fish during Seton Dam passage in 2013 (grey triangles) and 2014 (black circles); hourly Seton Dam discharge (B) during period when fish were detected in the dam tailrace in 2013 (grey line) and 2014 (black line), with horizontal dashed line indicating period of altered Seton Dam flow conditions (i.e. ‘high’ discharge), Aug 9-19, 2014; the highest temperature (C) experienced by individual fish in Portage Creek in 2013 (grey circles) and 2014 (black circles).   1416182022Temperature (°C) A242628303234Discharge (m3 s-1) B1517192123Temperature (°C) CAug 13 Aug 20 Aug 27 Sep 03 Sep 10 87  Figure 3.6 – Distributions of the number of days spent in Seton and Anderson lakes by fish tagged in 2013 (grey bars; n=27) and 2014 (white bars; n=87). Time in Seton Lake differed significantly by year (Kolmogorov-Smirnov two-sided test, D=0.354, p-value=0.012), but time in Anderson Lake did not.  0 5 10 15024681012Seton Lake0 5 10 15024681012Anderson Lake0 5 10 1502468101214Number of fish0 5 10 1502468101214Time in lake (days) 88   Figure 3.7 – Distributions of the temperature mode (i.e. most common temperature experienced) of individual tagged sockeye salmon in (A) Seton Lake and (B) Anderson Lake in 2013 and 2014. Temperature bins are 0.5°C and right-closed, and the y-axis indicates the number of fish that had a temperature mode in a given bin. The distributions of preferred temperature differed significantly between lakes (two-sample Kolmogorov-Smirnov test, D=0.421, p<0.0001).0 5 10 15 20 2505101520 Seton Lake0 5 10 15 20 2505101520 Anderson LakeNumber of fishTemperature mode (°C) 89  Figure 3.8 – Scatterplot of temperature mode vs. (A) the highest temperature experienced at Seton Dam, (B) time in Seton Lake, and (C) time in Anderson Lake for sockeye salmon tagged in 2013 (grey triangles) and 2014 (black circles). Model 1 predictions shown for fish that experienced low (solid line) and high (dashed line) discharge at the Seton Dam (A, B); predictions did not differ by year. Model 2 predictions show for fish tagged in 2013 (grey line) and 2014 (black line; C); predictions did not differ by time in lake. 15 16 17 18 19 205101520 APredicted temperature mode (°C)Tdam (°C)low dischargehigh discharge0 3 6 9 12 15BTime in lake (days)Seton Lake201320140 3 6 9 12 15CTime in lake (days)Anderson Lake 90  Figure 3.9 – Boxplots compare the proportion of time in Seton Lake (A, B) and Anderson Lake (D, E) that male and female sockeye salmon spent at temperatures below 10 °C. Scatterplots show percent time below 10 °C vs. number of days in Seton Lake (C) and Anderson Lake (F). Model 3 found a significant positive relationship between time in Seton Lake and proportion of time spent at temperatures below 10 °C (C). Model 4 found that females spent significantly more time below 10 °C in Anderson Lake in 2014 (E).   0.00.20.40.60.81.0male femaleSeton LakeA0.00.20.40.60.81.0male femaleB0 3 6 9 12 150.00.20.40.60.81.0 C0.00.20.40.60.81.0male femaleAnderson LakeD2013Time <10°C (%)0.00.20.40.60.81.0male femaleE*20140 3 6 9 12 150.00.20.40.60.81.0 FTime in lake (days) 91  Figure 3.10 – Standardized coefficients (mean=0, standard deviation=2) with 95% confidence intervals (CI) for models describing temperature mode in (A) Seton Lake and (B) Anderson Lake and use of thermal habitat <10°C in (C) Seton Lake and (D) Anderson Lake. (A) and (B) are model-averaged coefficients, (C) and (D) are from global models. Filled circles indicate coefficients with CI that do not cross zero.    -4-2024Seton Lake A-1012Anderson Lake B-2024Seton Lake C-2-1012Anderson Lake Dyear taggingdatesex GSE discharge Tdam Tpc SL(days)AL(days)CoefficientsRelative effect sizeTemperature modeTime <10 (°C) 92 Chapter 4: Conclusion Sockeye salmon complete reproductive migrations that are remarkable in their geographic distance and precision. Successful migrants are able to allocate limited somatic energy to multiple processes, including aerobic swimming, recovery from anaerobic activity, reproductive maturation, and immune function. However, migrants perish, both en route to and on spawning grounds, after exposure to environmental stressors, including high water temperature and complex, high velocity flows. My thesis examined the relative importance of physiological state and migratory experience to behaviour, survival, and reproductive success in a population of sockeye salmon impacted by a hydroelectric dam. The Gates Creek sockeye salmon population presents a unique opportunity to study thermal behaviour of wild, freely swimming fish in the context of population-specific relationships between temperature, oxygen consumption, and aerobic scope (Lee et al. 2003a, 2003b, Eliason et al. 2011). Aerobic scope is an important property of the cardiorespiratory system, because it limits a fish’s capacity for simultaneous physiological processes; yet, very little is known about how fish use temperature to manipulate scope in natural environments, or how fish partition available scope between oxygen-dependent processes (Clark et al. 2013, Eliason and Farrell 2016).  In Chapter 2, I characterized the thermal experience of female fish in Seton and Anderson lakes, by calculating the proportion of the migration that female fish spent at temperatures that provide at least 90% of maximum scope for activity. I then assessed the relative importance of physiological state and migratory experience to survival, migration duration, reproductive longevity, and egg retention. In Chapter 3, I devised a framework to assess how temperature selection reflects physiological trade-offs between somatic energy  93 conservation and aerobic capacity for activity. I subsequently identified the preferred temperature of male and female sockeye salmon in Seton and Anderson lakes, as well as the proportion of time that fish spent at temperatures below 10 ºC (i.e. with less than 60% of maximum scope for activity). Lastly, I compared the relative importance of physiological state and migratory experience to each of these temperature metrics.   4.1 Metrics for evaluating temperature experience In my thesis, I developed three metrics for summarizing thermal experience. With these metrics, I was able to identify correlations with prior migratory experience or subsequent fitness outcomes, which may be indicative of the physiological processes underlying temperature selection.  First, I determined the proportion of the migration that female fish spent in an optimum temperature window, which corresponded to ToptAS ± 3 ºC (i.e. 13.4-19.5 ºC for Gates Creek sockeye salmon). I observed that thermal experience in the days and weeks before spawning is important to reproductive success in wild female fish, which supports existing data from aquaculture and controlled experiments (see Pankhurst and King 2010). Greater use of the ToptAS window may facilitate maturation processes that enable successful reproduction, or use of temperatures outside the window (i.e. <13.4 ºC) may be indicative of a compromised condition, such as high pathogen load or disease, which underlies both cooler thermal preference and pre-spawning mortality.  The second temperature metric, the proportion of migration spent at temperatures <10 ºC, was based on the physiological trade-off framework that I developed. In this framework, 10 ºC was a threshold where the relative benefit of temperature switched from energy conservation (<10 ºC) to aerobic capacity (>10 ºC). Use of temperatures <10 ºC is an  94 indication of maintenance cost minimization during holding behaviour, which is common in sockeye salmon waiting to enter spawning streams (e.g. Newell and Quinn 2005, Mathes et al. 2010, Roscoe et al. 2010a). However, the extent to which fish occupy this thermal habitat and the proximity of use to spawning grounds have important implications for physiological capacity. For example, females that spend a greater proportion of time at temperatures <10 ºC (i.e. with less than 60% of maximum aerobic scope) have reduced spawning success (A.L. Bass, University of British Columbia, unpublished data), which could be caused by insufficient oxygen for gonad perfusion during vitellogenesis (E.J. Eliason, University of British Columbia, pers. comm.).  Third, I used the temperature mode (i.e. most common temperature value) of individual fish to identify the temperatures that fish preferentially occupied in Seton and Anderson lakes. I observed that fish manipulate body temperature in response to exposure to stressful conditions. Migrants that experienced both the standard flow conditions at Seton Dam (known to cause anaerobiosis and subsequent mortality in Gates Creek sockeye; Burnett et al. 2014a) and high water temperatures preferred cooler temperatures in Seton Lake, which demonstrates that fish can recover effectively from anaerobic effort, even at temperatures where fish have low aerobic scope.  4.2 Trade-offs between energy conservation and aerobic scope for activity Migration duration through the Seton-Anderson watershed was strongly influenced by migration timing, both overall and within each lake. Fish that arrived and were tagged earlier in the Seton River spent longer in lakes than later migrants, with the result that fish roughly coordinated their arrival on spawning grounds. Time in Seton Lake was an important correlate of thermal preference. Preference for ToptAS appeared to be a strategy used by later migrants to  95 optimize swimming ability for fast migration through Seton Lake (Lee et al. 2003a, 2003b). Conversely, fish that spent longer in Seton Lake spent a greater proportion of time at temperatures below 10 ºC. The inverse relationship between lake residence time and temperature preference appears to be scalable to the population level, as sockeye salmon populations with longer residency in freshwater lakes require cooler temperatures (e.g. Newell and Quinn 2005, Katinic et al. 2015). The different migratory strategies that I observed in earlier and later migrants may be related to relative somatic energy levels, as there was a strong negative relationship between tagging date and somatic energy density. Interestingly, I did not observe a direct effect of somatic energy level on temperature experience in any of my statistical models; this may be an indication that somatic energy was not low enough for the fish in my study, which all survived to spawning grounds, to be exceptionally energy conserving in their thermal behaviour.  My results suggest that migratory strategy is related to the cumulative experience of adult sockeye salmon, including factors affecting migrants prior to arrival in the Seton-Anderson watershed, which I did not include in my analyses. In both years of tagging, water temperature in the Fraser River (measured at Qualark) exceeded 20 °C during the period when Gates Creek sockeye salmon were migrating upriver (Casselman et al. 2013, 2014). The 2013 migrants are likely to have experienced greater cumulative thermal stress, because Fraser River temperatures were higher and remained elevated longer (i.e. >18 °C for three weeks, D.A. Patterson, Fisheries and Oceans Canada, unpublished data) and fish were additionally exposed to temperatures >18 °C in the Seton River during tagging, transport, and release (Casselman et al. 2013). Mortality was much higher in 2013 fish, and those that did survive to Gates Creek exhibited a more marked trade-off in thermal selection between Seton and Anderson lakes than  96 2014 fish. Many of the sockeye salmon tagged in 2013 migrated quickly through Seton Lake at temperatures close to ToptAS, and then occupied Anderson Lake for a longer period, at cooler temperatures (i.e. all fish used temperatures <10 °C). To reduce the duration of exposure to stressful conditions, some populations of sockeye salmon migrate faster at high temperatures (Naughton et al. 2005, Keefer et al. 2008), and 2013 fish may have exhibited this strategy.  4.3 Lethal and sub-lethal effects of stressors Many scientific evaluations of fishways at hydroelectric dams do not consider the effects on survival or fitness that fish may experience after successful passage (Roscoe and Hinch 2009). Previous research on the Seton Dam fishway identified post-passage mortality in Seton and Anderson lakes (Roscoe et al. 2010b, Burnett et al. 2014a), apparently caused by failure to recover from the anaerobic effort required to cross the dam tailrace and locate the fishway entrance (Burnett et al. 2014b). Between Aug 8-19, 2014, the flow dynamics in the Seton Dam tailrace were altered by releasing water from a siphon ~10 m away from the fishway entrance (i.e. altered flow conditions), rather than directly adjacent to the entrance (i.e. routine flow conditions), which also increased total discharge. In my thesis, I observed effects on survival, behaviour, and reproduction. Female fish that experienced altered Seton Dam flow conditions (i.e. ‘high’ discharge) had higher survival to Gates Creek, although the difference was not significant. Male and female sockeye salmon that passed the Seton Dam when water temperature was high, or under the routine flow conditions (i.e. ‘low’ discharge), showed a preference for cooler temperatures in Seton Lake than fish that experienced cooler dam passage temperatures and altered flow conditions (i.e. ‘high’ discharge). The optimum temperature for recovery from anaerobic effort has not been conclusively determined for sockeye salmon, although there is some suggestion that it is close to ToptAS (Lee et al. 2003b),  97 where fish have the greatest available oxygen. Stressed fish that select cool temperatures would reduce routine metabolic oxygen consumption, but also respiration rate (Brett 1956), and presumably take longer to recover than fish that select temperatures near ToptAS. A fish’s recovery strategy is likely to depend on the severity and duration of the stressor, as well as the individual’s physiological condition and migration timing. On spawning grounds, female fish exhibited a delayed effect of dam passage: the female fish that passed Seton Dam during the period of altered flow conditions (i.e. ‘high’ discharge) had lower reproductive longevity in Gates Creek and were more likely to die with retained eggs than individuals that experienced routine flow conditions (i.e. ‘low’ discharge). There are multiple possible mechanisms for the delayed effect of dam passage on spawning, including energy depletion caused by high discharge (Rand et al. 2006, Nadeau et al. 2010), extended anaerobic recovery (Martin et al. 2015), cumulative stress (Gilhousen 1990), suppression of maturation (Palstra et al. 2010), or a combination (Fenkes et al. 2015). Carryover effects have been observed in other migratory species, where there is a physiological conflict between completing migration and preparing for reproduction (e.g. in macaroni penguins [Eudyptes chrysolophus], Crossin et al. 2010). 4.4 Limitations to work and future research directions Research on adult Pacific salmon migrations has identified numerous physiological adaptations and behavioural strategies that conserve energy (Hodgson and Quinn 2002, Hinch and Rand 2000, Crossin et al. 2004b, Newell and Quinn 2005). My work has contributed to the study of thermoregulatory behaviour by interpreting thermal selection in the context of physiological trade-offs between energy conservation and aerobic capacity. However, the physiological pathways that allocate oxygen to multiple simultaneous processes are complex. To secure lifetime fitness, adult sockeye salmon that are staging in natal lakes prior to  98 spawning may need to further develop gonads and secondary sexual characters, recover from anaerobic activity, and combat pathogens as the fish’s immune function declines with senescence. All of these activities require energy and the use of more optimal water temperatures would facilitate the efficient use of limited endogenous energy reserves. Further experimentation is required to determine whether thermal optima differ for these physiological processes, and if staging migrants alter their thermal behaviour accordingly.  A limitation of my research approach is that I was unable to assess the importance of thermal experience to survival to spawning grounds, as archival temperature loggers could not be recovered from fish that died in Seton and Anderson lakes. Use of a different technology, such as acoustic temperature transmitters (e.g. Mathes et al. 2010), would help to disentangle the importance of thermal selection for recovery, following dam passage.  4.5 Management implications In regulated rivers, adult Pacific salmon are often exposed to stressful migratory conditions (Naughton et al. 2005, Keefer et al. 2008, Keefer and Caudill 2015), which may be fatal for a portion of the population. Managers in some systems have the ability to manipulate the water temperatures and flow velocities that fish encounter, thereby reducing the physiological stress and energetic expenditure of migrants and increasing rates of survival to spawning grounds (e.g. Burnett et al. In press). The water temperatures that adult sockeye salmon occupy in natal lakes are responsive to prior migratory experience and important to reproductive success for female fish. As migrants are exposed to increasingly stressful temperatures caused by climate warming, the ability of fish to facilitate recovery and prepare for spawning using thermal refugia will becomes even more important.    99 References Angilletta, M. J. 2009. Thermal Adaptation: A Theoretical and Empirical Synthesis. Oxford University Press, Oxford. Asbury, D.A., and Angilletta, M.J., Jr. 2010. Thermodynamic Effects on the Evolution of Performance Curves. American Naturalist 176(2): E40–E49.  Barton, B.A. 2002. Stress in Fishes: A Diversity of Responses with Particular Reference to Changes in Circulating Corticosteroids. Integrative and Comparative Biology 42: 517–525.  Barton K. 2012. MuMIn: multi-model inference. R package, version 1.7.11. http://CRAN.R-project.org/packagepMuMIn.  Beacham, T.D., and Murray, C.B. 1986. Comparative Developmental Biology of Chum Salmon (Oncorhynchus keta) from the Fraser River, British Columbia. Canadian Journal of Fisheries and Aquatic Sciences 43: 252–262.  Beacham, T.D., Candy, J.R., McIntosh, B., MacConnachie, C., Tabata, A., Kaukinen, K.H., Deng, L., Miller, K.M., Withler, R.E., and Varnavskaya, N. 2005. Estimation of Stock Composition and Individual Identification of Sockeye Salmon on a Pacific Rim Basis Using Microsatellite and Major Histocompatibility Complex Variation. Transactions of the American Fisheries Society 134(5): 1124–1146.  Beitinger, T.L., and Fitzpatrick, L.C. 1979. Physiological and Ecological Correlates of Preferred Temperature in Fish. American Zoologist 19(1): 319–329.   Beitinger, T.L., Magnuson, J.J., Neill, W.H., and Shaffer, W.R. 1975. Behavioural thermoregulation and activity patterns in the green sunfish, Lepomis cyanellus. Animal Behaviour 23: 222–229.  Benda, S.E., Naughton, G.P., Caudill, C.C., Kent, M.L., and Schreck, C.B. 2015. Cool, Pathogen-Free Refuge Lowers Pathogen-Associated Prespawn Mortality of Willamette River Chinook Salmon. Transactions of the American Fisheries Society 144(6): 1159–1172.   Bergstedt, R.A., Argyle, R.L., Seelye, J.G., Scribner, K.T., and Curtis, G.L. 2003. In Situ Determination of the Annual Thermal Habitat Use by Lake Trout (Salvelinus namaycush) in Lake Huron. Journal of Great Lakes Research 29(S1): 347–361.   Bett, N.N., and Hinch, S.G. 2015. Attraction of migrating adult sockeye salmon to conspecifics in the absence of natal chemical cues. Behavioral Ecology 26(4): 1180–1187.   Bevelhimer, M.S., and Adams, S.M. 1993. A Bioenergetics Analysis of Diel Vertical  100 Migration by Kokanee Salmon, Oncorhynchus nerka. Canadian Journal of Fisheries and Aquatic Sciences 50: 2336–2349.  Biette, R.M. and Geen, G.H. 1980. Growth of underyearling salmon, Oncorhynchus nerka, under constant and cyclic temperatures in relation to live zooplankton ration size. Canadian Journal of Fisheries and Aquatic Sciences 37: 203–210.  Black, E.C. 1958. Hyperactivity as a Lethal Factor in Fish. Journal of the Fisheries Research Board of Canada 15(4): 573–586.  Block, C.J., Spotila, J.R., Standora, E.A., and Gibbons, J.W. 1984. Behavioral thermoregulation of largemouth bass, Micropterus salmoides, and bluegill, Lepomis macrochirus, in a nuclear reactor cooling reservoir. Environmental Biology of Fishes 2(1): 41–52.  Bonga, S. 1997. The Stress Response in Fish. Physiological Reviews 77(3): 591–625.  Bradford, M.J., Lovy, J., and Patterson, D.A. 2010. Infection of gill and kidney of Fraser River sockeye salmon, Oncorhynchus nerka (Walbaum), by Parvicapsula minibicornis and its effect on host physiology. Journal of Fish Diseases 33(9): 769–779.   Brannon, E.L. 1987. Mechanisms stabilizing salmonid fry emergence timing. Canadian Special Publication of Fisheries and Aquatic Sciences 96: 120–124.  Brett, J.R. 1956. Some Principles in the Thermal Requirements of Fishes. The Quarterly Review of Biology 31(2): 75–87.   Brett, J.R. 1971. Energetic Responses of Salmon to Temperature. A Study of Some Thermal Relations in the Physiology and Freshwater Ecology of Sockeye Salmon (Oncorhynchus nerka). American Zoologist 11(1): 99–113.  Brett, J. R. 1995. Energetics. In Physiological Ecology of Pacific Salmon. Groot, C., Margolis, L. and Clarke, W. C., eds. University of British Columbia Press, Vancouver. pp. 1–68.  Bromage, N., Porter, M., and Randall, C. 2001. The environmental regulation of maturation in farmed finfish with special reference to the role of photoperiod and melatonin. Aquaculture 197: 63–98.  Budy, P., Thiede, G.P., Bouwes, N., Petrosky, C.E., and Schaller, H. 2002. Evidence linking delayed mortality of Snake River salmon to their earlier hydrosystem experience. North American Journal of Fisheries Management 22: 35–51.  Burgner, R.L. 1991. Life history of sockeye salmon. In Pacific Salmon Life Histories. Groot, C. and Margolis, L., eds. University of British Columbia Press, Vancouver. pp. 1–117.  Burnett, N.J., Hinch, S.G., Bett, N.N., Braun, D.C., Casselman, M.T., Cooke, S.J., Gelchu, A., Lingard, S.L., Middleton, C.T., Minke-Martin, V., and White, C. In press. Improving wild  101 sockeye salmon migration success through a large-scale management experiment. River Research and Applications.  Burnett, N.J., Hinch, S.G., Braun, D.C., Casselman, M.T., Middleton, C.T., Wilson, S.M., and Cooke, S.J. 2014a. Burst Swimming in Areas of High Flow: Delayed Consequences of Anaerobiosis in Wild Adult Sockeye Salmon. Physiological and Biochemical Zoology 87(5): 587–598.   Burnett, N.J., Hinch, S.G., Donaldson, M.R., Furey, N.B., Patterson, D.A., Roscoe, D.W., and Cooke, S.J. 2014b. Alterations to dam-spill discharge influence sex-specific activity, behaviour and passage success of migrating adult sockeye salmon. Ecohydrology. doi: 10.1002/eco.1440.  Burnham, K.P., and Anderson, D.R. 2002. Model Selection and Multimodel Inference: A Practical Information-Theoretic Approach. Springer, New York.  Campbell, P.M., Pottinger, T.G., and Sumpter, J.P. 1994. Preliminary evidence that chronic confinement stress reduces the quality of gametes produced by brown and rainbow trout. Aquaculture 120: 151–169.  Casselman MT, Burnett NJ, Bett NN, McCubbing D, Hinch SG. 2012. BRGMON-14 Effectiveness of Cayoosh flow dilution, dam operation, and fishway passage on delay and survival of upstream migration of salmon in the Seton-Anderson watershed. Annual Report – 2012. Report prepared for St’át’imc Government Services and BC Hydro. The University of British Columbia, Vancouver, BC.  Casselman, MT, Burnett, NJ, Bett, NN, Middleton, CT, Martins, EG, Braun, DC, McCubbing, D, Hinch, SG. 2013. BRGMON-14 Effectiveness of Cayoosh flow dilution, dam operation, and fishway passage on delay and survival of upstream migration of salmon in the Seton-Anderson watershed. Annual Report – 2013. Report prepared for St’át’imc Government Services and BC Hydro. The University of British Columbia, Vancouver, BC.  Casselman, M.T., N.J. Burnett, N.N. Bett, C.T. Middleton, V. Minke-Martin, D.C. Braun, D. McCubbing, and Hinch, S.G. 2015. BRGMON-14 Effectiveness of Cayoosh Flow Dilution, Dam Operation, and Fishway Passage on Delay and Survival of Upstream Migration of Salmon in the Seton-Anderson Watershed. Annual Report – 2014. Report prepared for St’át’imc Government Services and BC Hydro. The University of British Columbia, Vancouver, BC.  Caudill, C.C., Daigle, W.R., Keefer, M.L., Boggs, C.T., Jepson, M.A., Burke, B.J., Zabel, R.W., Bjornn, T.C., and Peery, C.A. 2007. Slow dam passage in adult Columbia River salmonids associated with unsuccessful migration: delayed negative effects of passage obstacles or condition-dependent mortality? Canadian Journal of Fisheries and Aquatic Sciences 64(7): 979–995.   Clark, C.W., and Levy, D.A. 1988. Diel Vertical Migrations by Juvenile Sockeye Salmon and the Antipredation Window. American Naturalist 131(2): 271–290.  102  Clark, T.D., Jeffries, K.M., Hinch, S.G., and Farrell, A.P. 2011. Exceptional aerobic scope and cardiovascular performance of pink salmon (Oncorhynchus gorbuscha) may underlie resilience in a warming climate. Journal of Experimental Biology 214(18): 3074–3081.   Clark, T.D., Sandblom, E., and Jutfelt, F. 2013. Aerobic scope measurements of fishes in an era of climate change: respirometry, relevance and recommendations. Journal of Experimental Biology 216(15): 2771–2782.  Cooke, S.J., Crossin, G.T., Patterson, D.A., English, K.K., Hinch, S.G., Young, J.L., Alexander, R.F., Healey, M.C., Van Der Kraak, G., and Farrell, A.P. 2005. Coupling non-invasive physiological assessments with telemetry to understand inter-individual variation in behaviour and survivorship of sockeye salmon: development and validation of a technique. Journal of Fish Biology 67(5): 1342–1358.   Cooke, S.J., Donaldson, M.R., Hinch, S.G., Crossin, G.T., Patterson, D.A., Hanson, K.C., English, K.K., Shrimpton, J.M., and Farrell, A.P. 2009. Is fishing selective for physiological and energetic characteristics in migratory adult sockeye salmon? Evolutionary Applications 2(3): 299–311.   Cooke, S.J., Hinch, S.G., Crossin, G.T., Patterson, D.A., English, K.K., Healey, M.C., Shrimpton, J.M., Van Der Kraak, G., and Farrell, A.P. 2006a. Mechanistic Basis of Individual Mortality in Pacific Salmon During Spawning Migrations. Ecology 87(6): 1575–1586.  Cooke, S.J., Hinch, S.G., Crossin, G.T., Patterson, D.A., English, K.K., Shrimpton, J.M., Van Der Kraak, G., and Farrell, A.P. 2006b. Physiology of individual late-run Fraser River sockeye salmon (Oncorhynchus nerka) sampled in the ocean correlates with fate during spawning migration. Canadian Journal of Fisheries and Aquatic Sciences 63(7): 1469–1480.   Cossins, A.R., and Bowler, K. 1987. Temperature Biology of Animals. Chapman and Hall, London.  Coutant, C.C. and Carroll, D.S. 1980. Temperatures occupied by ten ultrasonic-tagged striped bass in freshwater lakes. Transactions of the American Fisheries Society 109, 195–202.  Crossin, G.T., and Hinch, S.G. 2005. A nonlethal, rapid method for assessing the somatic energy content of migrating adult Pacific salmon. Transactions of the American Fisheries Society 134(1): 184–191.   Crossin, G.T., Hinch, S.G., Cooke, S.J., Cooperman, M.S., Patterson, D.A., Welch, D.W., Hanson, K.C., Olsson, I., English, K.K., and Farrell, A.P. 2009. Mechanisms Influencing the Timing and Success of Reproductive Migration in a Capital Breeding Semelparous Fish Species, the Sockeye Salmon. Physiological and Biochemical Zoology 82(6): 635–652.    103 Crossin, G.T., Hinch, S.G., Cooke, S.J., Welch, D.W., Patterson, D.A., Jones, S.R.M., Lotto, A.G., Leggatt, R.A., Mathes, M.T., Shrimpton, J.M., Van Der Kraak, G., and Farrell, A.P. 2008. Exposure to high temperature influences the behaviour, physiology, and survival of sockeye salmon during spawning migration. Canadian Journal of Zoology 86(2): 127–140.   Crossin, G.T., Hinch, S.G., Farrell, A.P., Higgs, D.A., and Healey, M.C. 2004a. Somatic energy of sockeye salmon Oncorhynchus nerka at the onset of upriver migration: a comparison among ocean climate regimes. Fisheries Oceanography 13(5): 345–349.   Crossin, G.T., Hinch, S.G., Farrell, A.P., Higgs, D.A., Lotto, A.G., Oakes, J.D., and Healey, M.C. 2004b. Energetics and morphology of sockeye salmon: effects of upriver migratory distance and elevation. Journal of Fish Biology 65: 788–810.   Crossin, G.T., Hinch, S.G., Farrell, A.P., Whelly, M.P., and Healey, M.C. 2003. Pink salmon (Oncorhynchus gorbuscha) migratory energetics: response to migratory difficulty and comparisons with sockeye salmon (Oncorhynchus nerka). Canadian Journal of Zoology 81(12): 1986–1995.  Crossin, G.T., Trathan, P.N., Phillips, R.A., Dawson, A., Le Bouard, F., and Williams, T.D. 2010. A carryover effect of migration underlies individual variation in reproductive readiness and extreme egg size dimorphism in macaroni penguins. American Naturalist 176(3): 357–366.   Crozier, L.G., Hendry, A.P., Lawson, P.W., Quinn, T.P., Mantua, N.J., Battin, J., Shaw, R.G., and Huey, R.B. 2008. Potential responses to climate change in organisms with complex life histories: evolution and plasticity in Pacific salmon. Evolutionary Applications 1(2): 252–270. Blackwell Publishing Ltd.   Daufresne, M., Roger, M.C., Capra, H., and Lamouroux, N. 2004. Long-term changes within the invertebrate and fish communities of the Upper Rhone River: effects of climatic factors. Global Change Biology 10(1): 124–140.   Davies, B., and Bromage, N. 2002. The effects of fluctuating seasonal and constant water temperatures on the photoperiodic advancement of reproduction in female rainbow trout, Oncorhynchus mykiss. Aquaculture 205: 183–200.  Dickerson, B.R., Quinn, T.P., and Willson, M.F. 2002. Body size, arrival date, and reproductive success of pink salmon, Oncorhynchus gorbuscha. Ethology Ecology & Evolution 14(1): 29–44.  Dickhoff W.W. 1989. Salmonids and annual fishes: death after sex. In Development, Maturation, and Senescence of Neuroendocrine Systems: A Comparative Approach. Schreibman, M.P. and Scanes, C.G., eds. Academic Press, New York. pp. 253–266.  Dingle, H. 1980. The ecology and evolution of migration. In Animal Migration, Orientation, and Navigation. Gauthreaux, S.A., ed. Academic Press, New York. pp.1-101.   104 Dingle, H. 1996. Migration: The Biology of Life on the Move. Oxford University Press, New York.  Donaldson, M.R., Cooke, S.J., Patterson, D.A., Hinch, S.G., Robichaud, D., Hanson, K.C., Olsson, I., Crossin, G.T., English, K.K., and Farrell, A.P. 2009. Limited behavioural thermoregulation by adult upriver-migrating sockeye salmon (Oncorhynchus nerka) in the Lower Fraser River, British Columbia. Canadian Journal of Zoology 87(6): 480–490.   Donaldson, M.R., Hinch, S.G., Raby, G.D., Patterson, D.A., Farrell, A.P., and Cooke, S.J. 2012. Population-specific consequences of fisheries-related stressors on adult sockeye salmon. Physiological and Biochemical Zoology 85(6): 729–739.  Drenner, S.M., Clark, T.D., Whitney, C.K., Martins, E.G., Cooke, S.J., and Hinch, S.G. 2012. A Synthesis of Tagging Studies Examining the Behaviour and Survival of Anadromous Salmonids in Marine Environments. PLoS ONE 7(3): e31311–13.   Drenner, S.M., Hinch, S.G., Martins, E.G., Robichaud, D., Clark, T.D., Thompson, L.A., Patterson, D.A., Cooke, S.J., and Thomson, R.E. 2014. Variable thermal experience and diel thermal patterns of homing sockeye salmon in coastal marine waters. Marine Ecology Progress Series 496: 109–124.   Dunn, P.K. 2014. Package “tweedie.” R package, version 2.2.1. http://www.r-project.org/package=tweedie. : 1–21.  Døving, K.B., Westerberg, H., and Johnsen, P.B. 1985. Role of Olfaction in the Behavioral and Neuronal Responses of Atlantic Salmon, Salmo salar, to Hydrographic Stratification. Canadian Journal of Fisheries and Aquatic Sciences 42: 1658–1667.  Eliason, E.J., and Farrell, A.P. 2016. Oxygen uptake in Pacific salmon Oncorhynchus spp.: when ecology and physiology meet. Journal of Fish Biology 88(1): 359–388.   Eliason, E.J., Clark, T.D., Hague, M.J., Hanson, L.M., Gallagher, Z.S., Jeffries, K.M., Gale, M.K., Patterson, D.A., Hinch, S.G., and Farrell, A.P. 2011. Differences in Thermal Tolerance Among Sockeye Salmon Populations. Science 332(6025): 109–112.   Eliason, E.J., Clark, T.D., Hinch, S.G., and Farrell, A.P. 2013a. Cardiorespiratory collapse at high temperature in swimming adult sockeye salmon. Conservation Physiology 1(1): 1–19.   Eliason, E.J., Clark, T.D., Hinch, S.G., and Farrell, A.P. 2013b. Cardiorespiratory performance and blood chemistry during swimming and recovery in three populations of elite swimmers: Adult sockeye salmon. Comparative Biochemistry and Physiology, Part A 166(2): 385–397.   Elliott, J.M. 1975. The Growth Rate of Brown Trout (Salmo trutta L.) Fed on Maximum Rations. Journal of Animal Ecology 44(3): 805–821.  Elliott, J.M., and Hurley, M.A. 1997. A functional model for maximum growth of Atlantic  105 Salmon parr, Salmo salar, from two populations in northwest England. Functional Ecology 11: 592–603.  English, K.K., Koski, W.R., Sliwinski, C., Blakley, A., Cass, A., and Woodey, J.C. 2005. Migration Timing and River Survival of Late-Run Fraser River Sockeye Salmon Estimated Using Radiotelemetry Techniques. Transactions of the American Fisheries Society 134(5): 1342-1365.   Farrell, A.P. 2015. Pragmatic perspective on aerobic scope: peaking, plummeting, pejus and apportioning. Journal of Fish Biology 88(1): 322–343.   Farrell, A.P., Gallaugher, P.E., and Routledge, R. 2001a. Rapid recovery of exhausted adult coho salmon after commercial capture by troll fishing. Canadian Journal of Fisheries and Aquatic Sciences 58(12): 2319–2324.   Farrell, A.P., Gallaugher, P.E., Fraser, J., Pike, D., Bowering, P., Hadwin, A.K.M., Parkhouse, W., and Routledge, R. 2001b. Successful recovery of the physiological status of coho salmon on board a commercial gillnet vessel by means of a newly designed revival box. Canadian Journal of Fisheries and Aquatic Sciences 58(10): 1932–1946.  Farrell, A.P., Hinch, S.G., Cooke, S.J., Patterson, D.A., Crossin, G.T., Lapointe, M., and Mathes, M.T. 2008. Pacific salmon in hot water: Applying aerobic scope models and biotelemetry to predict the success of spawning migrations. Physiological and Biochemical Zoology 81(6): 697–709.   Fenkes, M., Shiels, H.A., Fitzpatrick, J.L., and Nudds, R.L. 2015. The potential impacts of migratory difficulty, including warmer waters and altered flow conditions, on the reproductive success of salmonid fishes. Comparative Biochemistry and Physiology, Part A. doi: 10.1016/j.cbpa.2015.11.012.   Flett, P.A., Munkittrick, K.R., Van Der Kraak, G., and Leatherland, J.F. 1996. Overripening as the cause of low survival to hatch in Lake Erie coho salmon (Oncorhynchus kisutch) embryos. Canadian Journal of Zoology 74: 851–857.  Fry, F.E.J. 1947. Effects of the environment on animal activity. Publications of the Ontario Fisheries Research Laboratory 68: 1–63.  Fukushima, M., Quinn, T.P., and Smoker, W.W. 1998. Estimation of eggs lost from superimposed pink salmon (Oncorhynchus gorbuscha) redds. Canadian Journal of Fisheries and Aquatic Sciences 55: 618–625.  Gaesser, G.A., and Brooks, G.A. 1984. Metabolic bases of excess post-exercise oxygen consumption: a review. Medicine and Science in Sports and Exercise 16(1): 29–43.  Gale, M.K., Hinch, S.G., Cooke, S.J., Donaldson, M.R., Eliason, E.J., Jeffries, K.M., Martins, E.G., and Patterson, D.A. 2014. Observable impairments predict mortality of captured and released sockeye salmon at various temperatures. Conservation Physiology 2(1). doi:  106 10.1093/conphys/cou029.  Geen, G.H., and Andrew, F.J. 1961. Limnological Changes in Seton Lake Resulting from Hydroelectric Diversions. International Pacific Salmon Fisheries Commission Progress Report No. 8.  Geist, D.R., Brown, R.S., Cullinan, V.I., Mesa, M.G., VanderKooi, S.P., and McKinstry, C.A. 2003. Relationships between metabolic rate, muscle electromyograms and swim performance of adult Chinook salmon. Journal of Fish Biology 63(4): 970–989.   Gelman, A. 2008. Scaling regression inputs by dividing by two standard deviations. Statistics in Medicine 27(15): 2865–2873.   Gilhousen, P. 1980. Energy sources and expenditures in Fraser River sockeye salmon during their spawning migration. International Pacific Salmon Fisheries Commission Bulletin No. 22.  Gilhousen, P. 1990. Prespawning mortalities of sockeye salmon in the Fraser River system and possible causal factors. International Pacific Salmon Fisheries Commission Bulletin No. 26.  Gleeson, T.T. 1996. Post-exercise lactate metabolism: A Comparative Review of Sites, Pathways, and Regulation. Annual Review of Physiology 58: 565–581.  Goniea, T.M., Keefer, M.L., Bjornn, T.C., Peery, C.A., Bennett, D.H., and Stuehrenberg, L.C. 2006. Behavioral Thermoregulation and Slowed Migration by Adult Fall Chinook Salmon in Response to High Columbia River Water Temperatures. Transactions of the American Fisheries Society 135(2): 408–419.  Grueber, C.E., Nakagawa, S., Laws, R.J., and Jamieson, I.G. 2011. Multimodel inference in ecology and evolution: challenges and solutions. Journal of Evolutionary Biology 24(4): 699–711.   Hague, M.J., Ferrari, M.R., Miller, J.R., Patterson, D.A., Russell, G.L., Farrell, A.P., and Hinch, S.G. 2011. Modelling the future hydroclimatology of the lower Fraser River and its impacts on the spawning migration survival of sockeye salmon. Global Change Biology 17(1): 87–98.  Hari, R.E., Livingstone, D.M., Siber, R., Burkhardt-Holm, P., and Guttinger, H. 2006. Consequences of climatic change for water temperature and brown trout populations in Alpine rivers and streams. Global Change Biology 12: 10–26.  Harrell, F. 2001. Regression Modeling Strategies: With Applications to Linear Models, Logistic and Ordinal Regression, and Survival Analysis. Springer.  Harrison, P.M., Gutowsky, L.F.G., Martins, E.G., Patterson, D.A., Cooke, S.J., and Power, M. 2016. Temporal plasticity in thermal-habitat selection of burbot Lota lota a diel-migrating  107 winter-specialist. Journal of Fish Biology 88(6): 2111–2129.  Hendry, A.P., and Berg, O.K. 1999. Secondary sexual characters, energy use, senescence, and the cost of reproduction in sockeye salmon. Canadian Journal of Zoology 77: 1663–1675.  Hinch, S.G., and Rand, P.S. 2000. Optimal swimming speeds and forward-assisted propulsion: energy-conserving behaviours of upriver-migrating adult salmon. Canadian Journal of Fisheries and Aquatic Sciences 57: 2470–2478.  Hinch, S.G., Cooke, S.J., Healey, M.C., and Farrell, A.P. 2006. Behavioural Physiology of Fish Migrations: Salmon as a model approach. In Fish Physiology Volume 24: Behaviour and Physiology of Fish. Sloman, K., Balshine, S., and Wilson, R., eds. Elsevier Press. pp. 239-295.  Hinch, S.G., Cooke, S.J., Farrell, A.P., Miller, K.M., Lapointe, M., and Patterson, D.A. 2012. Dead fish swimming: a review of research on the early migration and high premature mortality in adult Fraser River sockeye salmon Oncorhynchus nerka. Journal of Fish Biology 81: 576–599.   Hoar, W., Randall, D., and Farrell, A., eds. 1992. Fish Physiology, Volume XII, Part A, The Cardiovascular System. Academic Press, Inc., San Diego.  Hodgson, S., and Quinn, T.P. 2002. The timing of adult sockeye salmon migration into fresh water: adaptations by populations to prevailing thermal regimes. Canadian Journal of Zoology 80: 542–555.  Holt, R.A., Sanders, J.E., Zinn, J.L., Fryer, J.L., and Pilcher, K.S. 1975. Relation of water temperature to Flexibacter columnaris infection in Steelhead Trout (Salmo gairdneri), Coho (Oncorhynchus kisutch) and Chinook (O. tshawytscha) Salmon. Journal of the Fisheries Research Board of Canada 32: 1553–1559.  Hosmer, D.W. and Lemeshow, S. 1989. Applied logistic regression. Wiley, New York.  Hruska, K.A., Hinch, S.G., Healey, M.C., Patterson, D.A., Larsson, S., and Farrell, A.P. 2010. Influences of sex and activity level on physiological changes in individual adult sockeye salmon during rapid senescence. Physiological and Biochemical Zoology 83(4): 663–676.   Hruska, K.A., Hinch, S.G., Patterson, D.A., and Healey, M.C. 2011. Egg retention in relation to arrival timing and reproductive longevity in female sockeye salmon (Oncorhynchus nerka). Canadian Journal of Fisheries and Aquatic Sciences 68(2): 250–259.  Hyatt, K.D., Stockwell, M.M., and Rankin, D.P. 2003. Impact and Adaptation Responses of Okanagan River Sockeye Salmon (Oncorhynchus nerka) to Climate Variation and Change Effects During Freshwater Migration: Stock Restoration and Fisheries Management Implications. Canadian Water Resources Journal 28(4): 689–713.   Jeffries, K.M., Hinch, S.G., Martins, E.G., Clark, T.D., Lotto, A.G., Patterson, D.A., Cooke,  108 S.J., Farrell, A.P., and Miller, K.M. 2012. Sex and proximity to reproductive maturity influence the survival, final maturation, and blood physiology of Pacific salmon when exposed to high temperature during a simulated migration. Physiological and Biochemical Zoology. 85(1): 62–73.   Jensen, J.O.T., McLean, W.E., Sweeten, T., Damon, W., and Berg, C. 2006. Puntledge River High Temperature Study: Influence of High Water Temperature on Adult Summer Chinook Salmon (Oncorhynchus tshawytscha) in 2004 and 2005. Canadian Technical Report of Fisheries and Aquatic Sciences 2662: vii–47p.  Katinic, P.J., Patterson, D.A., and Ydenberg, R.C. 2014. Thermal regime, predation danger and the early marine exit of sockeye salmon Oncorhynchus nerka. Journal of Fish Biology 86(1): 276–287.   Keefer, M.L., and Caudill, C.C. 2015. Estimating thermal exposure of adult summer steelhead and fall Chinook salmon migrating in a warm impounded river. Ecology of Freshwater Fish. doi: 10.1111/eff.12238.  Keefer, M.L., Clabough, T.S., Jepson, M.A., Naughton, G.P., Blubaugh, T.J., Joosten, D.C., and Caudill, C.C. 2015. Thermal exposure of adult Chinook salmon in the Willamette River basin. Journal of Thermal Biology 48: 11–20.   Keefer, M.L., Peery, C.A., and Heinrich, M.J. 2008. Temperature‐mediated en route migration mortality and travel rates of endangered Snake River sockeye salmon. Ecology of Freshwater Fish 17(1): 136–145.  Keefer, M.L., Peery, C.A., and High, B. 2009. Behavioral thermoregulation and associated mortality trade-offs in migrating adult steelhead (Oncorhynchus mykiss): variability among sympatric populations. Canadian Journal of Fisheries and Aquatic Sciences 66(10): 1734–1747.   Keefer, M.L., Taylor, G.A., Garletts, D.F., Gauthier, G.A., Pierce, T.M., and Caudill, C.C. 2010. Prespawn mortality in adult spring Chinook salmon outplanted above barrier dams. Ecology of Freshwater Fish 19(3): 361–372.  Kieffer, J.D. 2000. Limits to exhaustive exercise in fish. Comparative Biochemistry and Physiology, Part A 126: 161–179.  Killick, S.R. 1955. The chronological order of Fraser River sockeye salmon during migration, spawning and death. International Pacific Salmon Fisheries Commission Bulletin No. 7.  King, H.R., and Pankhurst, N.W. 2003. Ovarian growth and plasma sex steroid and vitellogenin profiles during vitellogenesis in Tasmanian female Atlantic salmon (Salmo salar). Aquaculture 219(1-4): 797–813.   King, H.R., Pankhurst, N.W., and Watts, M. 2007. Reproductive sensitivity to elevated water temperatures in female Atlantic salmon is heightened at certain stages of vitellogenesis.  109 Journal of Fish Biology 70(1): 190–205.   King, H.R., Pankhurst, N.W., Watts, M., and Pankhurst, P.M. 2003. Effect of elevated summer temperatures on gonadal steroid production, vitellogenesis and egg quality in female Atlantic salmon. Journal of Fish Biology 63(1): 153–167.   Kinnison, M.T., Unwin, M.J., Hendry, A.P., and Quinn, T.P. 2001. Migratory Costs and the Evolution of Egg Size and Number in Introduced and Indigenous Salmon Populations. Evolution 55(8): 1656–1667.  Kitagawa, T., Nakata, H., Kimura, S., Itoh, T., Tsuji, S., and Nitta, A. 2000. Effect of ambient temperature on the vertical distribution and movement of Pacific bluefin tuna Thunnus thynnus orientalis. Marine Ecology Progress Series 206: 251–260.  Kovach, R.P., Joyce, J.E., Echave, J.D., Lindberg, M.S., and Tallmon, D.A. 2013. Earlier Migration Timing, Decreasing Phenotypic Variation, and Biocomplexity in Multiple Salmonid Species. PLoS ONE 8(1): e53807–10.   Kubokawa, K., Yoshioka, M., and Iwata, M. 2001. Sex-specific Cortisol and Sex Steroids Responses in Stressed Sockeye Salmon during Spawning Period. Zoological Science 18(7): 947–954.   Lee, C.G. 2003a. The effect of temperature on swimming performance and oxygen consumption in adult sockeye (Oncorhynchus nerka) and coho (O. kisutch) salmon stocks. Journal of Experimental Biology 206(18): 3239–3251.   Lee, C.G. 2003b. Excess post-exercise oxygen consumption in adult sockeye (Oncorhynchus nerka) and coho (O. kisutch) salmon following critical speed swimming. Journal of Experimental Biology 206(18): 3253–3260.   Limnotek. 2015. Seton Lake Aquatic Productivity Monitoring: Progress in 2014. Report prepared for St’át’imc Government Services and BC Hydro. Limnotek Research and Consulting Ltd., Vancouver, BC.  Lingard, S.L., Ladell, J.J., and McCubbing, D.J.F. 2014. Gates Creek Adult Sockeye Escapement, Fall 2013. Report prepared for St’át’imc Government Services and BC Hydro. InStream Fisheries Research Inc., Burnaby, BC.  Mac, M.J. 1985. Effects of ration size on preferred temperature of lake charr Salvelinus namaycush. Environmental Biology of Fishes 14(2/3): 227–231.  Macdonald, J.S., Foreman, M.G.G., Farrell, A.P., Williams, I.V., Grout, J., Cass, A., Woodey, J.C., Enzenhofer, H., Clarke, W.C., Houtman, R., Donaldson, E.M., and Barnes, D. 2000. The Influence of Extreme Water Temperatures on Migrating Fraser River Sockeye Salmon (Oncorhynchus nerka) During the 1998 Spawning Season. Canadian Technical Report of Fisheries and Aquatic Sciences No. 2326.   110 Mackenzie-Grieve, J.L., and Post, J.R. 2006. Thermal Habitat Use by Lake Trout in Two Contrasting Yukon Territory Lakes. Transactions of the American Fisheries Society 135(3): 727–738.   Magnuson, J.J., Crowder, L.B., and Medvick, P.A. 1979. Temperature as an Ecological Resource. American Zoologist 19(1): 331–343.  Major, R.L., and Mighell, J.L. 1967. Influence of Rocky Reach Dam and the temperature of the Okanogan River on the upstream migration of sockeye salmon. Fishery Bulletin 66(1): 131–147.  Martin, B.T., Nisbet, R.M., Pike, A., Michel, C.J., and Danner, E.M. 2015. Sport science for salmon and other species: ecological consequences of metabolic power constraints. Ecology Letters 18(6): 535–544. doi: 10.1111/ele.12433.  Martin, T.L., and Huey, R.B. 2008. Why “Suboptimal” Is Optimal: Jensen’s Inequality and Ectotherm Thermal Preferences. American Naturalist 171(3): E102–E118.   Martins, E.G., Hinch, S.G., Cooke, S.J., and Patterson, D.A. 2012a. Climate effects on growth, phenology, and survival of sockeye salmon (Oncorhynchus nerka): a synthesis of the current state of knowledge and future research directions. Reviews in Fisheries Biology and Fisheries 22(4): 887–914.   Martins, E.G., Hinch, S.G., Patterson, D.A., Hague, M.J., Cooke, S.J., Miller, K.M., Lapointe, M.F., English, K.K., and Farrell, A.P. 2011. Effects of river temperature and climate warming on stock-specific survival of adult migrating Fraser River sockeye salmon (Oncorhynchus nerka). Global Change Biology 17(1): 99–114.   Martins, E.G., Hinch, S.G., Patterson, D.A., Hague, M.J., Cooke, S.J., Miller, K.M., Robichaud, D., English, K.K., Farrell, A.P., and Jonsson, B. 2012b. High river temperature reduces survival of sockeye salmon (Oncorhynchus nerka) approaching spawning grounds and exacerbates female mortality. Canadian Journal of Fisheries and Aquatic Sciences 69(2): 330–342.  Mathes, M.T., Hinch, S.G., Cooke, S.J., Crossin, G.T., Patterson, D.A., Lotto, A.G., and Farrell, A.P. 2010. Effect of water temperature, timing, physiological condition, and lake thermal refugia on migrating adult Weaver Creek sockeye salmon (Oncorhynchus nerka). Canadian Journal of Fisheries and Aquatic Sciences 67(1): 70–84.   McCullough, D., Spalding, S., Sturdevant, D., and Hicks, M. 2001. Summary of technical literature examining the physiological effects of temperature on salmonids. United States Environmental Protection Agency, Seattle.   McPhee, M.V., and Quinn, T.P. 1998. Factors affecting the duration of nest defense and reproductive lifespan of female sockeye salmon, Oncorhynchus nerka. Environmental Biology of Fishes 51(4): 369–375.    111 Mehranvar, L., Healey, M.C., Farrell, A.P., and Hinch, S.G. 2004. Social versus genetic measures of reproductive success in sockeye salmon, Oncorhynchus nerka. Evoluntionary Ecology Research 6: 1167–1181.  Miller, K.M., Teffer, A., Tucker, S., Li, S., Schulze, A.D., Trudel, M., Juanes, F., Tabata, A., Kaukinen, K.H., Ginther, N.G., Ming, T.J., Cooke, S.J., Hipfner, J.M., Patterson, D.A., and Hinch, S.G. 2014. Infectious disease, shifting climates, and opportunistic predators: cumulative factors potentially impacting wild salmon declines. Evolutionary Applications 7(7): 812–855.   Morbey, Y.E., and Ydenberg, R.C. 2003. Timing Games in the Reproductive Phenology of Female Pacific Salmon (Oncorhynchus spp.). American Naturalist 161(2): 284–298.   Morrison, J., Quick, M.C., and Foreman, M.G.G. 2002. Climate change in the Fraser River watershed: flow and temperature projections. Journal of Hydrology 263: 230–244.  Nadeau, P.S., Hinch, S.G., Hruska, K.A., Pon, L.B., and Patterson, D.A. 2010. The effects of experimental energy depletion on the physiological condition and survival of adult sockeye salmon (Oncorhynchus nerka) during spawning migration. Environmental Biology of Fishes 88(3): 241–251.   Nathan, R., Getz, W.M., Revilla, E., Holyoak, M., Kadmon, R., Saltz, D., and Smouse, P.E. 2008. A movement ecology paradigm for unifying organismal movement research. Proceedings of the National Academy of Sciences 105(49): 19052–19059.  Naughton, G.P., Caudill, C.C., Keefer, M.L., Bjornn, T.C., Stuehrenberg, L.C., and Peery, C.A. 2005. Late-season mortality during migration of radio-tagged adult sockeye salmon (Oncorhynchus nerka) in the Columbia River. Canadian Journal of Fisheries and Aquatic Sciences 62(1): 30–47.   Neill, W.H., Magnuson, J.J., and Chipman, G.G. 1972. Behavioral Thermoregulation by Fishes: A New Experimental Approach. Science 176: 1443–1445.  Newell, J.C., and Quinn, T.P. 2005. Behavioral thermoregulation by maturing adult sockeye salmon (Oncorhynchus nerka) in a stratified lake prior to spawning. Canadian Journal of Zoology 83(9): 1232–1239.   Nowell, L.B., Brownscombe, J.W., Gutowsky, L.F.G., Murchie, K.J., Suski, C.D., Danylchuk, A.J., Shultz, A., and Cooke, S.J. 2015. Swimming energetics and thermal ecology of adult bonefish (Albula vulpes): a combined laboratory and field study in Eleuthera, The Bahamas. Environmental Biology of Fishes 98(11): 2133–2146.   O’Brien, R.M. 2007. A Caution Regarding Rules of Thumb for Variance Inflation Factors. Quality and Quantity 41(5): 673–690.   Palstra, A.P., and van den Thillart, G.E.E.J.M. 2010. Swimming physiology of European silver eels (Anguilla anguilla L.): energetic costs and effects on sexual maturation and  112 reproduction. Fish Physiology and Biochemistry 36(3): 297–322.   Pankhurst, N.W., and King, H.R. 2010. Temperature and salmonid reproduction: implications for aquaculture. Journal of Fish Biology 76(1): 69–85.   Pankhurst, N.W., and Van Der Kraak, G. 2000. Evidence That Acute Stress Inhibits Ovarian Steroidogenesis in Rainbow Trout in Vivo, through the Action of Cortisol. General and Comparative Endocrinology 117(2): 225–237.  Patterson, D.A., Macdonald, J.S., Skibo, K.M, Barnes, D.P., Guthrie, I. and Hills, J. 2007. Reconstructing the Summer Thermal History for the Lower Fraser River, 1941 to 2006, and Implications for Adult Sockeye Salmon (Oncorhynchus nerka) Spawning Migration. Canadian Technical Report of Fisheries and Aquatic Sciences No. 2724.  Payne, N.L., Smith, J.A., van der Meulen, D.E., Taylor, M.D., Watanabe, Y.Y., Takahashi, A., Marzullo, T.A., Gray, C.A., Cadiou, G., and Suthers, I.M. 2015. Temperature-dependence of fish performance in the wild: links with species biogeography and physiological thermal tolerance. Functional Ecology doi: 10.1111/1365-2435.12618.  Pon, L.B., Hinch, S.G., Cooke, S.J., Patterson, D.A., and Farrell, A.P. 2009. A Comparison of the Physiological Condition, and Fishway Passage Time and Success of Migrant Adult Sockeye Salmon at Seton River Dam, British Columbia, under Three Operational Water Discharge Rates. North American Journal of Fisheries Management 29(5): 1195–1205.   Pörtner, H.O. 2001. Climate change and temperature-dependent biogeography: oxygen limitation of thermal tolerance in animals. Naturwissenschaften 88(4): 137–146.   Pörtner, H.O., and Farrell, A.P. 2008. Physiology and climate change. Science 322(5902): 690–692.   Pörtner, H.O., and Knust, R. 2007. Climate Change Affects Marine Fishes Through the Oxygen Limitation of Thermal Tolerance. Science 315(5808): 92–95.   Quinn, T.P. 1988. Estimated swimming speeds of migrating adult sockeye salmon. Canadian Journal of Zoology 66: 2160–2163.  Quinn, T.P., and Adams, D.J. 1996. Environmental Changes Affecting the Migratory Timing of American Shad and Sockeye Salmon. Ecology 77(4): 1151–1162.  Quinn, T.P. 2005. The Behavior and Ecology of Pacific Salmon and Trout. The American Fisheries Society: Bethesda, Maryland.  Quinn, T.P., Eggers, D.M., Clark, J.H., and Rich, H.B., Jr. 2007. Density, climate, and the processes of prespawning mortality and egg retention in Pacific salmon (Oncorhynchus spp.). Canadian Journal of Fisheries and Aquatic Sciences 64(3): 574–582.  Rand, P.S., and Hinch, S.G. 1998. Swim speeds and energy use of upriver-migrating sockeye  113 salmon (Oncorhynchus nerka): simulating metabolic power and assessing risk of energy depletion. Canadian Journal of Fisheries and Aquatic Sciences 55: 1832–1841.  Rand, P.S., Hinch, S.G., Morrison, J., Foreman, M.G.G., MacNutt, M.J., Macdonald, J.S., Healey, M.C., Farrell, A.P., and Higgs, D.A. 2006. Effects of River Discharge, Temperature, and Future Climates on Energetics and Mortality of Adult Migrating Fraser River Sockeye Salmon. Transactions of the American Fisheries Society 135(3): 655–667.   Roscoe, D.W. and Hinch, S.G. 2007. Fishway passage, water diversion and warming temperatures: factors limiting successful spawning migration of Seton-Anderson watershed sockeye salmon. Final Report for the Bridge Coastal Restoration Program, Project. University of British Columbia, Vancouver, BC.  Roscoe, D.W., Hinch, S.G., Cooke, S.J., and Patterson, D.A. 2010a. Behaviour and thermal experience of adult sockeye salmon migrating through stratified lakes near spawning grounds: the roles of reproductive and energetic states. Ecology of Freshwater Fish 19(1): 51–62.   Roscoe, D.W., Hinch, S.G., Cooke, S.J., and Patterson, D.A. 2010b. Fishway passage and post-passage mortality of up-river migrating sockeye salmon in the Seton River, British Columbia. 27(6): 693–705.   Rounsefell, G.A. 1958. Anadromy in North American Salmonidae. U.S. Fish and Wildlife Service Bulletin No. 58: 171–185.   Sandblom, E., Clark, T.D., Hinch, S.G., and Farrell, A.P. 2009. Sex-specific differences in cardiac control and hematology of sockeye salmon (Oncorhynchus nerka) approaching their spawning grounds. American Journal of Physiology: Regulatory, Integrative and Comparative Physiology 297(4): 1136–1143.   Scarabello, M., Heigenhauser, G.J.F., and Wood, C.M. 1992. Gas exchange, metabolite status and excess post-exercise oxygen consumption after repetitive bouts of exhaustive exercise in juvenile rainbow trout. Journal of Experimental Biology 167: 155–169.  Schielzeth, H. 2010. Simple means to improve the interpretability of regression coefficients. Methods in Ecology and Evolution 1(2): 103–113.   Schmidt-Nielsen, K. 1997. Animal Physiology: Adaptation and Environment. Cambridge University Press, Cambridge.  Schreck, C.B., Contreras-Sanchez, W., and Fitzpatrick, M.S. 2001. Effects of stress on fish reproduction, gamete quality, and progeny. Aquaculture 197: 3–24.  Smith, C.E., Dwyer, W.P., Piper, R.G. 1983. Effect of water temperature on egg quality of cutthroat trout. The Progressive Fish-Culturalist 45(3): 176–178.  Snucins, E.J. and Gunn, J.M. 1995. Coping with a warm environment: Behavioral  114 thermoregulation by Lake Trout. Transactions of the American Fisheries Society 124(1): 118–123.  Spigarelli, S.A., Romberg, G.P., Prepejchal, W., and Thommes, M.M. 1974. Body-temperature characteristics of fish at a thermal discharge on Lake Michigan. In Thermal Ecology. Gibbons, J.W. and Sharitz, R.R., eds. U.S. Atomic Energy Commission, Oakridge. pp. 119–132.  Stitt, B.C., Burness, G., Burgomaster, K.A., Currie, S., McDermid, J.L., and Wilson, C.C. 2014. Intraspecific Variation in Thermal Tolerance and Acclimation Capacity in Brook Trout (Salvelinus fontinalis): Physiological Implications for Climate Change. Physiological and Biochemical Zoology 87(1): 15–29.   Swan, H. 1974. Thermoregulation and Bioenergetics. Elsevier, Amsterdam.  Tallman, R.F., and Healey, M.C. 1991. Phenotypic Differentiation in Seasonal Ecotypes of Chum Salmon, Oncorhynchus keta. Canadian Journal of Fisheries and Aquatic Sciences 48: 661–671.  Taranger, G.L., Vikingstad, E., Klenke, U., Mayer, I., Stefansson, S.O., Norberg, B., Hansen, T., Zohar, Y., and Andersson, E. 2003. Effects of photoperiod, temperature and GnRHa treatment on the reproductive physiology of Atlantic salmon (Salmo salar L.) broodstock. Fish Physiology and Biochemistry 28: 403–406.  Udey, L.R., Fryer, J.L., and Pilcher, K.S. 1975. Relation of water temperature to Ceratomyxosis in Rainbow Trout (Salmo gairdneri) and Coho Salmon (Oncorhynchus kisutch). Journal of the Fisheries Research Board of Canada 32: 1545–1551.  Van Der Kraak, G. and Pankhurst, N.W. 1997. Temperature effects on the reproductive performance of fish. In Global Warming: Implications for Freshwater and Marine Fish. Wood, C.M. and McDonald, D.G., eds. Cambridge University Press, Cambridge.  Wagner, G.N., Hinch, S.G., and Kuchel, L.J. 2005. Metabolic rates and swimming performance of adult Fraser River sockeye salmon (Oncorhynchus nerka) after a controlled infection with Parvicapsula minibicornis. Canadian Journal of Fisheries and Aquatic Sciences 62(9): 2124–2133.  Webb P.W. 1995. Locomotion. In Physiological ecology of Pacific salmon. Groot, C., Margolis, L., and Clarke, W.C., eds. University of British Columbia Press, Vancouver. pp. 69–99.  Wedemeyer, G.A., Barton, B.A., and McLeay, D.J. 1990. Stress and acclimation. In Methods for fish biology. Schreck, C.B. and Moyle, P.B., eds. American Fisheries Society, Bethesda. pp. 451–489.    Weihs, D. 1974. Energetic Advantages of Burst Swimming of Fish. Journal of Theoretical Biology 48: 215–229.  115  West, C.J. and Mason, J.C. 1987. Evaluation of sockeye salmon Oncorhynchus nerka production from the Babine Lake Development Project. In Sockeye salmon (Oncorhynchus nerka) population biology and future management. Smith, H.D., Margolis, L., and Wood, C.C., eds. Canadian Special Publications of Fisheries and Aquatic Sciences. pp. 176–190.  Westerberg, H. 1982. Ultrasonic Tracking of Atlantic Salmon (Salmo salar L.) - II. Swimming depth and temperature stratification. Drottingholm Report No. 60: 102–120.  Wilson, S.M., Hinch, S.G., Drenner, S.M., Martins, E.G., Furey, N.B., Patterson, D.A., Welch, D.W., and Cooke, S.J. 2014. Coastal marine and in-river migration behaviour of adult sockeye salmon en route to spawning grounds. Marine Ecology Progress Series 496: 71–84.  Wood, C.M. 1991. Acid-base and ion balance, metabolism, and their interactions, after exhaustive exercise in fish. Journal of Experimental Biology 160: 285–308.  Wood, C.M., Turner, J.D., and Graham, M.S. 1983. Why do fish die after severe exercise? Journal of Fish Biology 22(2): 189–201.   Young, D.B., and Woody, C.A. 2007. Dynamic in-lake spawning migrations by female sockeye salmon. Ecology of Freshwater Fish 16(2): 155–164.   Young, J.L., Hinch, S.G., Cooke, S.J., Crossin, G.T., Patterson, D.A., Farrell, A.P., Van Der Kraak, G., Lotto, A.G., Lister, A., Healey, M.C., and English, K.K. 2006. Physiological and energetic correlates of en route mortality for abnormally early migrating adult sockeye salmon (Oncorhynchus nerka) in the Thompson River, British Columbia. Canadian Journal of Fisheries and Aquatic Sciences 63: 1067–1077.  Zuur, A.F., Ieno, E.N., and Elphick, C.S. 2010. A protocol for data exploration to avoid common statistical problems. Methods in Ecology and Evolution 1(1): 3–14.   Zuur, A.F., Ieno, E.N., Walker, N.J., Saveliev, A.A., and Smith, G.M. 2009. Mixed Effects Models and Extensions in Ecology with R. Springer, New York.   116 Appendix Figure A.1 – Data from archival temperature loggers recovered from 114 sockeye salmon. Data between time of last telemetry detection at Seton Dam and first detection at Gates Creek are shown. Each day is indicated by a short tick on the x-axis, and the beginning of each week is labelled with the date. The vertical dashed line is the time of entry into Portage Creek; before the line, the fish is in Seton Lake and after, the fish is in Anderson Lake. Year, sex, and fish number are indicated. 	 117  09/0409/1151015202013male# 3408/1308/2008/2751015202014female# 6709/0551015202013female# 7208/1308/2051015202014male# 8908/1308/2008/2751015202014male# 9708/1508/2208/2951015202014male# 9908/1508/2208/2951015202014female# 11108/1308/2008/2751015202014male# 117	 118     08/1508/2208/2951015202014male# 12908/1908/2609/0251015202014male# 17408/2408/3151015202014female# 19508/1308/2008/2751015202014female# 21808/1608/2308/3051015202014female# 22808/1608/2351015202014male# 23108/1851015202014female# 24608/2308/3051015202014male# 267	 119     08/2551015202014female# 27508/2551015202014female# 27908/2609/0251015202014female# 28208/2751015202014male# 28608/2909/0551015202014male# 29108/1508/2251015202014female# 33108/1808/2509/0151015202014male# 33708/1908/2609/0251015202014female# 344	 120     08/1908/2651015202014male# 34508/2108/2809/0451015202014female# 34709/0751015202014male# 37609/0851015202014female# 39708/1908/2651015202014female# 45108/2008/2709/0309/1051015202014female# 45408/3009/0651015202014female# 49008/3051015202014female# 495	 121     08/0808/1551015202014male# 51608/1508/2208/2951015202014female# 52208/1808/2509/0151015202014male# 52608/2709/0351015202014female# 57308/2751015202014female# 59008/2851015202014female# 59408/1408/2108/2851015202014female# 62108/1851015202014female# 651	 122     08/1908/2651015202014male# 66508/2308/3051015202014male# 67808/2509/0151015202014male# 68408/1808/2551015202013female# 69309/0551015202014female# 69908/1808/2509/0151015202014female# 74208/2108/2851015202014female# 75108/2108/2851015202014female# 756	 123     08/2451015202014male# 77008/2709/0351015202014male# 79408/1608/2351015202014male# 81708/1808/2551015202014female# 81808/2108/2851015202014female# 82008/2208/2951015202014male# 82508/2509/0151015202014male# 82708/2709/0351015202014female# 837	 124  08/2851015202014male# 83808/3051015202014male# 84209/0751015202014female# 84708/1751015202013female# 89908/1608/2351015202014female# 91308/1851015202014female# 91808/2108/2851015202014male# 92408/2208/2951015202014male# 934	 125    08/2751015202014male# 94208/2809/0451015202014male# 96308/3109/0709/1451015202014male# 97409/0751015202014male# 98608/1408/2108/2851015202014male# 101908/1408/2108/2851015202014female# 102008/1608/2351015202014female# 102108/2251015202014female# 1031	 126     08/2408/3151015202014male# 103808/2651015202014male# 104708/2651015202014male# 105008/2951015202014female# 106608/2351015202014female# 110108/2308/3051015202014male# 110508/2651015202014female# 111008/2509/0151015202014male# 1111	 127     08/2408/3109/0751015202014female# 111408/2609/0251015202014female# 111808/2709/0351015202014female# 112608/2809/0451015202014male# 113009/0151015202014female# 113709/0751015202014female# 115208/2451015202013male# 117709/0409/1151015202013female# 1211	 128     09/0951015202014female# 211008/2451015202013male# 318808/1808/2551015202013female# 410108/2108/2851015202013male# 411808/2251015202013male# 415609/0251015202013female# 420208/3151015202014female# 510208/2251015202013female# 5106	 129     09/0851015202014male# 510809/0851015202014male# 511208/2408/3151015202013male# 515408/2451015202013male# 517309/0251015202013female# 520808/1908/2651015202013female# 610808/2108/2851015202013male# 614108/2951015202014female# 7115	 130    08/2308/3009/0651015202013female# 714508/2308/3051015202013male# 714808/2351015202013female# 814408/2451015202013female# 815308/2509/0151015202013female# 818208/2509/0151015202013male# 819208/2551015202013female# 819408/2351015202013female# 9165	 131    08/2108/2851015202013male# 1011309/0351015202013female# 10212 132 Figure A.2 – Histograms of temperature data, from each of 114 sockeye salmon in Seton Lake and Anderson Lake, binned by 0.5ºC and right-closed. The vertical dashed line indicates the temperature mode (i.e. most common temperature experienced) in each lake.    5 10 15 20 25050100150200250300 # 34Seton5 10 15 20 25050100150200250300 # 34Anderson5 10 15 20 25050100150200250300 # 67Seton5 10 15 20 25050100150200250300 # 67Anderson5 10 15 20 25050100150200250300 # 72Seton5 10 15 20 25050100150200250300 # 72Anderson5 10 15 20 25050100150200250300 # 89Seton5 10 15 20 25050100150200250300 # 89Anderson	 133    5 10 15 20 25050100150200250300 # 97Seton5 10 15 20 25050100150200250300 # 97Anderson5 10 15 20 25050100150200250300 # 99Seton5 10 15 20 25050100150200250300 # 99Anderson5 10 15 20 25050100150200250300 # 111Seton5 10 15 20 25050100150200250300 # 111Anderson5 10 15 20 25050100150200250300 # 117Seton5 10 15 20 25050100150200250300 # 117Anderson	 134    5 10 15 20 25050100150200250300 # 129Seton5 10 15 20 25050100150200250300 # 129Anderson5 10 15 20 25050100150200250300 # 174Seton5 10 15 20 25050100150200250300 # 174Anderson5 10 15 20 25050100150200250300 # 195Seton5 10 15 20 25050100150200250300 # 195Anderson5 10 15 20 25050100150200250300 # 218Seton5 10 15 20 25050100150200250300 # 218Anderson	 135    5 10 15 20 25050100150200250300 # 228Seton5 10 15 20 25050100150200250300 # 228Anderson5 10 15 20 25050100150200250300 # 231Seton5 10 15 20 25050100150200250300 # 231Anderson5 10 15 20 25050100150200250300 # 246Seton5 10 15 20 25050100150200250300 # 246Anderson5 10 15 20 25050100150200250300 # 267Seton5 10 15 20 25050100150200250300 # 267Anderson	 136    5 10 15 20 25050100150200250300 # 275Seton5 10 15 20 25050100150200250300 # 275Anderson5 10 15 20 25050100150200250300 # 279Seton5 10 15 20 25050100150200250300 # 279Anderson5 10 15 20 25050100150200250300 # 282Seton5 10 15 20 25050100150200250300 # 282Anderson5 10 15 20 25050100150200250300 # 286Seton5 10 15 20 25050100150200250300 # 286Anderson	 137    5 10 15 20 25050100150200250300 # 291Seton5 10 15 20 25050100150200250300 # 291Anderson5 10 15 20 25050100150200250300 # 331Seton5 10 15 20 25050100150200250300 # 331Anderson5 10 15 20 25050100150200250300 # 337Seton5 10 15 20 25050100150200250300 # 337Anderson5 10 15 20 25050100150200250300 # 344Seton5 10 15 20 25050100150200250300 # 344Anderson	 138    5 10 15 20 25050100150200250300 # 345Seton5 10 15 20 25050100150200250300 # 345Anderson5 10 15 20 25050100150200250300 # 347Seton5 10 15 20 25050100150200250300 # 347Anderson5 10 15 20 25050100150200250300 # 376Seton5 10 15 20 25050100150200250300 # 376Anderson5 10 15 20 25050100150200250300 # 397Seton5 10 15 20 25050100150200250300 # 397Anderson	 139    5 10 15 20 25050100150200250300 # 451Seton5 10 15 20 25050100150200250300 # 451Anderson5 10 15 20 25050100150200250300 # 454Seton5 10 15 20 25050100150200250300 # 454Anderson5 10 15 20 25050100150200250300 # 490Seton5 10 15 20 25050100150200250300 # 490Anderson5 10 15 20 25050100150200250300 # 495Seton5 10 15 20 25050100150200250300 # 495Anderson	 140    5 10 15 20 25050100150200250300 # 516Seton5 10 15 20 25050100150200250300 # 516Anderson5 10 15 20 25050100150200250300 # 522Seton5 10 15 20 25050100150200250300 # 522Anderson5 10 15 20 25050100150200250300 # 526Seton5 10 15 20 25050100150200250300 # 526Anderson5 10 15 20 25050100150200250300 # 573Seton5 10 15 20 25050100150200250300 # 573Anderson	 141    5 10 15 20 25050100150200250300 # 590Seton5 10 15 20 25050100150200250300 # 590Anderson5 10 15 20 25050100150200250300 # 594Seton5 10 15 20 25050100150200250300 # 594Anderson5 10 15 20 25050100150200250300 # 621Seton5 10 15 20 25050100150200250300 # 621Anderson5 10 15 20 25050100150200250300 # 651Seton5 10 15 20 25050100150200250300 # 651Anderson	 142    5 10 15 20 25050100150200250300 # 665Seton5 10 15 20 25050100150200250300 # 665Anderson5 10 15 20 25050100150200250300 # 678Seton5 10 15 20 25050100150200250300 # 678Anderson5 10 15 20 25050100150200250300 # 684Seton5 10 15 20 25050100150200250300 # 684Anderson5 10 15 20 25050100150200250300 # 693Seton5 10 15 20 25050100150200250300 # 693Anderson	 143    5 10 15 20 25050100150200250300 # 699Seton5 10 15 20 25050100150200250300 # 699Anderson5 10 15 20 25050100150200250300 # 742Seton5 10 15 20 25050100150200250300 # 742Anderson5 10 15 20 25050100150200250300 # 751Seton5 10 15 20 25050100150200250300 # 751Anderson5 10 15 20 25050100150200250300 # 756Seton5 10 15 20 25050100150200250300 # 756Anderson	 144    5 10 15 20 25050100150200250300 # 770Seton5 10 15 20 25050100150200250300 # 770Anderson5 10 15 20 25050100150200250300 # 794Seton5 10 15 20 25050100150200250300 # 794Anderson5 10 15 20 25050100150200250300 # 817Seton5 10 15 20 25050100150200250300 # 817Anderson5 10 15 20 25050100150200250300 # 818Seton5 10 15 20 25050100150200250300 # 818Anderson	 145    5 10 15 20 25050100150200250300 # 820Seton5 10 15 20 25050100150200250300 # 820Anderson5 10 15 20 25050100150200250300 # 825Seton5 10 15 20 25050100150200250300 # 825Anderson5 10 15 20 25050100150200250300 # 827Seton5 10 15 20 25050100150200250300 # 827Anderson5 10 15 20 25050100150200250300 # 837Seton5 10 15 20 25050100150200250300 # 837Anderson	 146    5 10 15 20 25050100150200250300 # 838Seton5 10 15 20 25050100150200250300 # 838Anderson5 10 15 20 25050100150200250300 # 842Seton5 10 15 20 25050100150200250300 # 842Anderson5 10 15 20 25050100150200250300 # 847Seton5 10 15 20 25050100150200250300 # 847Anderson5 10 15 20 25050100150200250300 # 899Seton5 10 15 20 25050100150200250300 # 899Anderson	 147    5 10 15 20 25050100150200250300 # 913Seton5 10 15 20 25050100150200250300 # 913Anderson5 10 15 20 25050100150200250300 # 918Seton5 10 15 20 25050100150200250300 # 918Anderson5 10 15 20 25050100150200250300 # 924Seton5 10 15 20 25050100150200250300 # 924Anderson5 10 15 20 25050100150200250300 # 934Seton5 10 15 20 25050100150200250300 # 934Anderson	 148    5 10 15 20 25050100150200250300 # 942Seton5 10 15 20 25050100150200250300 # 942Anderson5 10 15 20 25050100150200250300 # 963Seton5 10 15 20 25050100150200250300 # 963Anderson5 10 15 20 25050100150200250300 # 974Seton5 10 15 20 25050100150200250300 # 974Anderson5 10 15 20 25050100150200250300 # 986Seton5 10 15 20 25050100150200250300 # 986Anderson	 149    5 10 15 20 25050100150200250300 # 1019Seton5 10 15 20 25050100150200250300 # 1019Anderson5 10 15 20 25050100150200250300 # 1020Seton5 10 15 20 25050100150200250300 # 1020Anderson5 10 15 20 25050100150200250300 # 1021Seton5 10 15 20 25050100150200250300 # 1021Anderson5 10 15 20 25050100150200250300 # 1031Seton5 10 15 20 25050100150200250300 # 1031Anderson	 150    5 10 15 20 25050100150200250300 # 1038Seton5 10 15 20 25050100150200250300 # 1038Anderson5 10 15 20 25050100150200250300 # 1047Seton5 10 15 20 25050100150200250300 # 1047Anderson5 10 15 20 25050100150200250300 # 1050Seton5 10 15 20 25050100150200250300 # 1050Anderson5 10 15 20 25050100150200250300 # 1066Seton5 10 15 20 25050100150200250300 # 1066Anderson	 151   5 10 15 20 25050100150200250300 # 1101Seton5 10 15 20 25050100150200250300 # 1101Anderson5 10 15 20 25050100150200250300 # 1105Seton5 10 15 20 25050100150200250300 # 1105Anderson5 10 15 20 25050100150200250300 # 1110Seton5 10 15 20 25050100150200250300 # 1110Anderson5 10 15 20 25050100150200250300 # 1111Seton5 10 15 20 25050100150200250300 # 1111Anderson	 152 5 10 15 20 25050100150200250300 # 1114Seton5 10 15 20 25050100150200250300 # 1114Anderson5 10 15 20 25050100150200250300 # 1118Seton5 10 15 20 25050100150200250300 # 1118Anderson5 10 15 20 25050100150200250300 # 1126Seton5 10 15 20 25050100150200250300 # 1126Anderson5 10 15 20 25050100150200250300 # 1130Seton5 10 15 20 25050100150200250300 # 1130Anderson	 153    5 10 15 20 25050100150200250300 # 1137Seton5 10 15 20 25050100150200250300 # 1137Anderson5 10 15 20 25050100150200250300 # 1152Seton5 10 15 20 25050100150200250300 # 1152Anderson5 10 15 20 25050100150200250300 # 1177Seton5 10 15 20 25050100150200250300 # 1177Anderson5 10 15 20 25050100150200250300 # 1211Seton5 10 15 20 25050100150200250300 # 1211Anderson	 154    5 10 15 20 25050100150200250300 # 2110Seton5 10 15 20 25050100150200250300 # 2110Anderson5 10 15 20 25050100150200250300 # 3188Seton5 10 15 20 25050100150200250300 # 3188Anderson5 10 15 20 25050100150200250300 # 4101Seton5 10 15 20 25050100150200250300 # 4101Anderson5 10 15 20 25050100150200250300 # 4118Seton5 10 15 20 25050100150200250300 # 4118Anderson	 155    5 10 15 20 25050100150200250300 # 4156Seton5 10 15 20 25050100150200250300 # 4156Anderson5 10 15 20 25050100150200250300 # 4202Seton5 10 15 20 25050100150200250300 # 4202Anderson5 10 15 20 25050100150200250300 # 5102Seton5 10 15 20 25050100150200250300 # 5102Anderson5 10 15 20 25050100150200250300 # 5106Seton5 10 15 20 25050100150200250300 # 5106Anderson	 156    5 10 15 20 25050100150200250300 # 5108Seton5 10 15 20 25050100150200250300 # 5108Anderson5 10 15 20 25050100150200250300 # 5112Seton5 10 15 20 25050100150200250300 # 5112Anderson5 10 15 20 25050100150200250300 # 5154Seton5 10 15 20 25050100150200250300 # 5154Anderson5 10 15 20 25050100150200250300 # 5173Seton5 10 15 20 25050100150200250300 # 5173Anderson	 157    5 10 15 20 25050100150200250300 # 5208Seton5 10 15 20 25050100150200250300 # 5208Anderson5 10 15 20 25050100150200250300 # 6108Seton5 10 15 20 25050100150200250300 # 6108Anderson5 10 15 20 25050100150200250300 # 6141Seton5 10 15 20 25050100150200250300 # 6141Anderson5 10 15 20 25050100150200250300 # 7115Seton5 10 15 20 25050100150200250300 # 7115Anderson	 158    5 10 15 20 25050100150200250300 # 7145Seton5 10 15 20 25050100150200250300 # 7145Anderson5 10 15 20 25050100150200250300 # 7148Seton5 10 15 20 25050100150200250300 # 7148Anderson5 10 15 20 25050100150200250300 # 8144Seton5 10 15 20 25050100150200250300 # 8144Anderson5 10 15 20 25050100150200250300 # 8153Seton5 10 15 20 25050100150200250300 # 8153Anderson	 159    5 10 15 20 25050100150200250300 # 8182Seton5 10 15 20 25050100150200250300 # 8182Anderson5 10 15 20 25050100150200250300 # 8192Seton5 10 15 20 25050100150200250300 # 8192Anderson5 10 15 20 25050100150200250300 # 8194Seton5 10 15 20 25050100150200250300 # 8194Anderson5 10 15 20 25050100150200250300 # 9165Seton5 10 15 20 25050100150200250300 # 9165Anderson	 160   5 10 15 20 25050100150200250300 # 10113Seton5 10 15 20 25050100150200250300 # 10113Anderson5 10 15 20 25050100150200250300 # 10212Seton5 10 15 20 25050100150200250300 # 10212Anderson

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0306922/manifest

Comment

Related Items