LOCAL ADAPTATION TO COLD TEMPERATURES BY LARVAL COHO
SALMON (ONCORHYNCHUS KISUTCH) FROM DIFFERENT POPULATIONS
THROUGHOUT BRITISH COLUMBIA
by
Kimberly Tuor
B.Sc., Environmental Science, Carleton University, 2012

THESIS SUBMITTED IN PARTIAL FULFILLMENT OF
THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE
IN
NATURAL RESOURCES AND ENVIRONMENTAL STUDIES
(BIOLOGY)

UNIVERSITY OF NORTHERN BRITISH COLUMBIA
November 2016
© Kimberly Mary Frances Tuor, 2016

ABSTRACT
The influence of environmental variables on larval development of coho salmon
(Oncorhynchus kisutch), with specific focus on the influence of near-freezing incubation
temperatures, was examined across populations within British Columbia. A survey across the
geographical distribution within British Columbia was conducted to determine the range and
variability of incubation temperatures experience by incubating coho salmon. Temperatures
throughout incubation differed significantly among locations, averaging approximately 1 °C
in colder interior locations and approximately 5 °C in warmer coastal locations.
Environmental variables influenced egg size, fecundity, female size and gonadal somatic
index, such that higher latitude of spawning grounds increased, larger systems decreased, and
increased temperatures experienced by a population increased the four life-history traits.
Suggesting significant effects of latitude of spawning grounds, size of spawning system and
temperatures experienced by a population on shaping patterns of reproductive investment. A
laboratory incubation study revealed no difference in survival and performance between
families from a southern and a northern population reared at near-freezing incubation
temperatures. These findings suggest plasticity in developmental processes of coho salmon,
as each population was successful across a wide range of temperatures, and in particular
developed successfully with minimal fitness effects at the extreme ranges of near-freezing
temperatures.

i

TABLE OF CONTENTS
ABSTRACT………………………………………………………………………………. i
TABLE OF CONTENTS………………………………………………………………… ii
LIST OF TABLES………………………………………………………………………... iv
LIST OF FIGURES………………………………………………………………………. vi
ACKNOLEDGEMENTS………………………………………………………………… viii
PROLOGUE……………………………………………………………………………… 1
CHAPTER 1……………………………………………………………………………… 5
ABSTRACT……………………………………………………………………………….. 5
INTRODUCTION…………………………………………………………………………. 6
METHODS………………………………………………………………………………… 8
Stream Temperature across British Columbia……………………………………………… 8
Statistical Analysis ……………………………………………………………………………… 10
RESULTS…………………………………………………………………………………. 11
DISCUSSION…………………………………………………………………………….. 16
CHAPTER 2………………………………………………………………………………. 21
ABSTRACT……………………………………………………………………………….. 21
INTRODUCTION…………………………………………………………………………. 22
METHODS………………………………………………………………………………… 24
Stocks used in the analysis and data collection……………………………………………… 24
Model Development……………………………………………………………………………… 27
Model Parameters……………………………………………………………………………… 27
Model Selection………………………………………………………………………………… 29
RESULTS………………………………………………………………………………….. 29
Egg Size…………………………………………………………………………………………… 29
Fecundity………………………………………………………………………………………… 30
Female Size……………………………………………………………………………………… 30
Gonadal Somatic Index………………………………………………………………………… 31
DISSCUSION……………………………………………………………………………... 36
Egg Size…………………………………………………………………………………………… 37
Fecundity………………………………………………………………………………………… 39
ii

Female Size……………………………………………………………………………………… 40
Gonadal Somatic Index………………………………………………………………………… 42
CONCLUSION……………………………………………………………………………. 43
CHAPTER 3……………………………………………………………………………… 44
ABSTRACT……………………………………………………………………………….. 44
INTRODUCTION…………………………………………………………………………. 45
METHODS………………………………………………………………………………… 47
Gamete collection and breeding design……………………………………………………… 47
Experimental Treatments……………………………………………………………………… 48
Size, Condition Factor and Survival………………………………………………………… 50
Oxygen Consumption…………………………………………………………………………… 50
Yolk Area Analysis……………………………………………………………………………… 52
Statistical Analysis……………………………………………………………………………… 53
RESULTS…………………………………………………………………………………. 54
Development Rate……………………………………………………………………………… 54
Size and Condition Factor..…………………………………………………………………… 54
Oxygen Consumption…………………………………………………………………………… 56
Yolk Area Analysis……………………………………………………………………………… 57
Survival…………………………………………………………………………………………… 57
DISCUSSION……………………………………………………………………………… 64
Development Rate…………………………………………………….………………………… 64
Size and Condition Factor ….………………………………………………………………… 66
Oxygen Consumption…………………………………………………………………………… 68
Yolk Area Analysis……………………………………………………………………… 69
Survival…………………………………………………………………………………………… 70
APPENDIX 3.1…………………………………………………………………………… 72
APPENDIX 3.2…………………………………………………………………………..... 75

EPILOGUE…………………………………………………………………. 80
REFERENCES……………………………………………………………… 84

iii

LIST OF TABLES
Table 1.1: Results from repeated measures analysis of variance testing whether average
intergravel temperature during the incubation period (November 22, 2012 – April 30,
2013) and the coldest portion of incubation (January) differed among spawning locations
of coho salmon in British Columbia. The variation is summarized based on the
difference between systems across British Columbia and the logger location within each
system (intergravel redd temperature and surface temperature), n = 5 pairs of surface and
intergravel loggers per system with a total n of 50.…………………………………… 15
Table 1.2: Summary of statistics comparing study systems determined by post hoc Tukey’s
Test. Systems that differed significantly are represented with an X for the average
intergravel temperature throughout incubation above the diagonal (November to April)
and a # for the coldest period of incubation below the diagonal (January). Populations
chosen are from four coastal regions with short migration distances but different
latitudes (Kanaka Creek (Kan) 49 ºN, Black Creek (Blk) 50 ºN, Bella Coola River (Bel)
52 ºN, and Kitimat River (Kit) 54 ºN), from three regions with intermediate migration
distances in central BC (Coldwater River (Col) 50 ºN, Nahatlach River (Nah) 50 ºN, and
Toboggan Creek (Tob) 55 ºN) and from three regions with long distance migrations
within the interior of BC (Eagle River (Eag) 51 ºN, Albreda River (Alb) and McKinley
Creek (McK) 52 ºN). See Figure 1 for locations.……………...……………………… 15
Table 2.1: Summary of the available data collected from each system for each life-history
trait (egg size, fecundity, female size at maturity, gonadal somatic index [GSI]). The
range of sample years and total number of years collected for each life-history
trait………………………………………………………………………...………..…. 27
Table 2.2: List of environmental variables used in analyses of variation in life-history traits
of coho salmon. ………………………………………………………………………. 28
Table 2.3: Summary of AICc ranking of candidate models for environmental variables
influencing egg size of coho salmon. Lat = Latitude, JanG = average temperature found
in each system throughout the coldest period of incubation (January), SysAvg = average
temperature found in each system throughout the incubation period from November to
April, YOH = total number of years an enhancement hatchery for coho salmon has been
in operation within each system, Mig = migration distance to each system from the
ocean, Size = size (using ranks) of each river, HeadW = type of headwater that feeds
each system…………………………………………………………………………... 32
Table 2.4: Summary of AICc ranking of candidate models for environmental variables
influencing fecundity of coho salmon. Lat = Latitude, JanG = average temperature found
in each system throughout the coldest period of incubation (January), SysAvg = average
temperature found in each system throughout the incubation period from November to
April, YOH = total number of years an enhancement hatchery for coho salmon has been
in operation within each system, Mig = migration distance to each system from the
iv

ocean, Size = size (using ranks) of each river, HeadW = type of headwater that feeds
each system…………………………………………………………………………… 33
Table 2.5: Summary of AICc ranking of candidate models for environmental variables
influencing female size of coho salmon. Lat = Latitude, SpTemp = average temperature
found in each system throughout the peak spawning period in November, SysAvg =
average temperature found in each system throughout the incubation period from
November to April, YOH = total number of years an enhancement hatchery for coho
salmon has been in operation within each system, Mig = migration distance to each
system from the ocean, Size = size (using ranks) of each river, HeadW = type of
headwater that feeds each system………………………………………………….….. 34
Table 2.6: Summary of AICc ranking of candidate models for environmental variables
influencing gonadal somatic index of coho salmon. Lat = Latitude, JanG = average
temperature found in each system throughout the coldest period of incubation (January),
SysAvg = average temperature found in each system throughout the incubation period
from November to April, YOH = total number of years an enhancement hatchery for
coho salmon has been in operation within each system, Mig = Mig distance to each
system from the ocean, Size = size (using ranks) of each river, HeadW = type of
headwater that feeds each system………………………………………....................... 35

v

LIST OF FIGURES
Figure P.1. The distribution of coho salmon (Oncorhynchus kisutch) throughout the Pacific
Rim (Sandercock1991)…………………………………………………………..….. 3
Figure 1.1: The distribution of coho salmon in British Columbia, represented as black circles
and the ten study systems, represented as red squares: four coastal short migration
distance populations (I-Kanaka Creek 49 ºN, J-Black Creek 50 ºN, C-Bella Coola River
52 ºN, and A-Kitimat River 54 ºN), three central BC populations with intermediate
migration distances (G-Coldwater River 50 ºN, H-Nahatlach River 50 ºN, and BToboggan Creek 55 ºN), and three interior BC populations with long-migration distances
(F-Eagle River 51 ºN, E-Albreda Creek and D-McKinley Creek 52 º………………... 9
Figure 1.2: The running average (over 2 weeks) of the intergravel maximum, average, and
minimum temperature regimes found across ten systems across BC where coho salmon
are found to spawn. The systems are ordered based on latitude from the most northern on
the top down to the most southern systems on the bottom...………………………….. 13
Figure 1.3: Average surface (open symbols) and intergravel (solid symbols) temperatures
throughout the incubation period (2012-2013) for coho salmon across all ten systems
measured in British Columbia. The average temperature throughout the incubation
period of November 22, 2012 to April 30, 2013 is represented as circles and the coldest
period throughout the incubation period (January) is represented as squares with
standard error bars. The systems are ordered based on latitude from the most northern to
the most southern systems (N = Northern latitude, Mid = Mid latitude, S = Southern
latitude and then ranked from interior to coastal within each of the latitudes (I = Interior,
C = Coastal)....……………………………………………………………………….. 14
Figure 2.1. Locations where data were collected for environmental variables and life-history
traits of coho salmon across British Columbia for this study. Location are labelled based
on the systems used throughout Chapter 1. The missing systems are due to unavailable
life-history data.…………………….............................................................................. 26
Figure 3.1: Temperature regimes followed throughout the incubation period for coho salmon
until button-up (red = warm, green = mid, and blue = cold treatment). The approximate
date of hatch (H) and button-up (B) are shown for each temperature regime………… 49
Figure 3.2: The required accumulated thermal units to reach hatch (a) and button-up (b) for
the different pure and hybrid families of coho salmon reared at cold, mid and warm
temperature regimes (see Figure 3.1). Families are represented as female X male pairs
from a northern population (N: Kitimat River) or a southern population (S: Kanaka
Creek). The upper case letters above each bar in figure a) represent the significant
difference between the three way interactions of treatment, maternal and paternal origin.
The letters above each bar in figure b) represent the significant differences determined
by the two-way interaction of treatment and maternal origin and the treatment and
paternal origin.………………………………………………………………………… 58
vi

Figure 3.3: Weight (a), length (b) and condition factor (c) at hatch for the different pure and
hybrid families of coho salmon reared at cold, mid and warm temperature regimes (see
Figure 3.1). Families are represented as female X male pairs from a northern population
(N: Kitimat River) or a southern population (S: Kanaka Creek). The upper case letters
above each bar represent the significant treatment and maternal origin interactions. Data
are shown as means + standard error. ………………………………………………… 59
Figure 3.4: Weight (a), length (b), and condition factor (c) at button-up for the different pure
and hybrid families reared at cold, mid and warm temperature regimes (Figure 3.1).
Refer to Figure 3.3 for detailed description of the legend. The upper case letter above
each bar represent the interaction effect of treatment and maternal origin and the
interaction effect of maternal and paternal origin. Figure a) has a three way interaction
effect of treatment, maternal origin and paternal origin and is described within the text.
Data are shown as means + standard error…………………………...……………..... 60
Figure 3.5: Resting metabolic rate at hatch (a) and button-up (b) for the different pure and
hybrid families reared at cold, mid and warm temperature regimes (Figure 3.1). Refer to
Figure 3.3 for detailed description of the legend. The upper case letters above each bar
represent a) the significant difference in treatment and b) the three-way interaction effect
of treatment, maternal origin and paternal origin. Data are shown as means + standard
error.…………………………………………………………………………………… 61
Figure 3.6: The yolk area (a) and standardized yolk area (b) at hatch for the different pure
and hybrid families reared at cold, mid and warm temperature regimes (Figure 3.1).
Refer to Figure 3.3 for detailed description of the legend. The solid line above each
treatment represents the significant difference from the treatments without a solid line.
The upper case letters above each bar represent the significant difference between
maternal origins. Data are shown as means + standard error.………………………… 62
Figure 3.7: The total percent survival for the different pure and hybrid families reared at
cold, mid and warm temperature regimes (Figure 3.1) from fertilization to button-up.
Refer to Figure 3.3 for detailed description of the legend. The upper case letters above
each bar represent the interaction effect of maternal and paternal origin. Data are shown
as means + standard error………………………………...…………………………… 63

vii

ACKNOWLEDGMENTS
Throughout my thesis I have received assistance from many groups and individuals
who have contributed to the overall success of this project. I would like to begin by thanking
my supervisor, Dr. Mark Shrimpton. I am deeply grateful for the support, guidance,
education, understanding and patience you gave me over my graduate studies. I could not
have completed my studies without these qualities in a mentor and supervisor. I would also
like to thank my graduate committee, Dr. Russ Dawson and Dr. Daniel Heath, for the
guidance and help over the years. I would also like to thank my external examiner Dr. Colin
Brauner. Funding for this project was provided to Dr. Mark Shrimpton by the Natural
Sciences and Engineering Research Council (NSERC) of Canada. I was also supported by a
scholarship from the Freshwater Fisheries Society of BC.
I would like to acknowledge Fisheries and Oceans Canada (DFO) for their assistance
and interest in my project. I would like to thank all of the DFO employees who provided
great field assistance and knowledge on coho salmon spawning site location and life-history
information during my field work. Thank you to Richard Bailey for his advice and input on
which systems to target throughout British Columbia. Thank you to everyone who provided
me with advice and assistance in the deployment of my temperature loggers within the
watersheds across British Columbia: David Nagtegaal and Peter Van Will for the Black
Creek System, Shane Kalyn and Kristin Singer for the Kanaka Creek System, Haakon
Hammer, Marshal Hans, and John Willis for the Bella Coola River System, Mike O’Neill for
the Toboggan Creek System, Markus Feldhoff for the Kitimat River System, Bruce
Whitehead and Jim Krivanek for the Eagle River System, Sherri Schmidt for the Nahatlatch
River System, Helen Olynyk for the Albreda River System, and Neil Todd and Jessica
Urquhart for the Coldwater River System. I would also like to thank the Kanaka Creek and
Kitimat River Hatcheries for providing me with the gametes needed for the breeding design
and hatchery assessment of temperatures on the development of coho salmon. I would also
like to acknowledge all of the employees with DFO who compiled and provided me with
life-history data from enhancement programs across British Columbia. Thank you to John
Willis and Haakon Hammer from the Snootli Creek Hatchery, Mike O’Neill from the
Toboggan Creek Hatchery, Markus Feldhoff from the Kitimat River Hatchery, Doug Turvey
from the Spius Creek Hatchery, Mourice Coulter-Boisvert and Darin McClain from the
Kanaka Creek Hatchery, and Ed Walls from the Big Qualicum River Hatchery.
From the University of Northern British Columbia, I would like to thank all of the
engineers and facilities workers for all their help in preparing the Aquatic Animal Holding
Facility and their assistance in the development of this facility for the assessment of
temperatures on the development of coho salmon. I would like to thank Dr. Chris Johnson
for help with statistical analysis and model development. I would like to thank all of the
students who assisted me with sampling throughout this experiment, Rick Elsner, Adam
O’Dell, Eric Vogt, Joe Strong, and Andrew McDermot-Fouts. I would also like to thank
Anne-Marie Flores and Dr. Mark Shrimpton for continuing the fish care and ensuring my
experiment continued smoothly while I took a short leave of absence, I could not have
continued this study without both of your support.
Finally I would like to thank my friends and family who supported me throughout my
studies and did not let me give up on my passion throughout the toughest times.
viii

PROLOGUE
Temperature profoundly affects how organisms function, and is particularly
important for poikilotherms. Temperature not only influences the attributes within a habitat
such as dissolved oxygen levels, conductivity and productivity, but also physiology and
survival of poikilotherms such as fish. Consequently, temperature is one of the leading
factors governing growth and development in fish (Fry 1971; Blaxter 1992). Effects of
temperature on poikilotherms are cumulative, as it has a direct effect on metabolic rate and
therefore their rate of development (Fry 1971; Brett 1995). Cold temperatures cause a
decrease in metabolic rates and development rates that in turn cause a decrease in multiple
measures of performance (Perry and Tufts 1998). Consequently, ectothermic animals have
evolved to cope with specific temperature regimes and relatively small temperature changes
can have measurable effects on community and population structure.
Adaptive evolution occurs when the genetic constitution of a population changes as a
consequence of natural selection (Merilä and Hendry 2014). Plasticity occurs when a given
genotype is unchanged, but adjust their phenotype due to the conditions experienced within
the environment (Merilä and Hendry 2014). When investigating the influence of temperature
on the physiology of poikilotherms such as metabolic rate and development, there are a
number of measures commonly used. The cumulative effect of temperature over time is
measured with the unit of accumulated thermal units which are calculated by the sum of the
average daily temperature experienced. Additionally, the temperature coefficient (Q10)
represents the factor by which the rate of a reaction increases for every ten degree rise in
temperature. It is useful as it may be used to infer mechanistic insight about the physiological
processes under investigation. The temperature coefficient is defined as:
1

10

𝑅2 (𝑇2 −𝑇1)
𝑄10 = ( )
𝑅1
where Q10 is the factor by which the reaction rate increases when the temperature is raised by
ten degrees. Q10 is a unitless quantity. R1 is the measured reaction rate at temperature
T1 (where T1 < T2). R2 is the measured reaction rate at temperature T2 (where T2 > T1).
For species with large geographic distributions, considerable differences in
temperature may be experienced, which have been shown to translate into differences in
thermal tolerance (Fangue et al. 2006). Such patterns, however, have not been adequately
determined. Adaptive phenotypic change through evolution or phenotypic plasticity may be
crucial for persistence of populations experiencing variable selective pressures. This is of
particular interest if demographic rescue from neighbouring populations is unlikely, such as
in salmonid fish, which have strong homing behaviour (Reed et al. 2011; Visser 2008).
Changing climates can result in loss of biodiversity (Thomas et al. 2004), shifts in
distribution (Brander 2003; Perry et al. 2005; Grebmeier et al. 2006), migration failure
(Farrell et al. 2008), and altered species productivity (Pörtner and Peck 2010), and
demonstrate the importance of temperature for animal function (Rosenzweig et al. 2008). An
understanding of the interaction between evolutionary and ecological processes, therefore, is
needed to determine the effect of anthropogenic disturbance and climate change on natural
populations (Reed et al. 2011).
Coho salmon (Oncorhynchus kisutch) have a large geographic distribution and spawn
around the Pacific Rim from California to Japan (Figure P.1; Shaul et al. 2007; Sandercock
1991). Moreover, throughout British Columbia (BC) coho salmon spawn in over 970 rivers
and small streams (Sandercock 1991). With such a large geographic distribution and with

2

spawning occurring along the coast as well as through the interior of BC (Figure P.1),
populations of coho salmon experience different ranges of temperatures during spawning and
incubation, and are expected to tolerate a wide range in temperatures. Coho salmon return to
their natal spawning grounds during the fall and spawn during fall and early winter, showing
a very narrow spawning window in comparison to other species of Pacific salmon. Spawners
migrate to small tributary streams, and even in coastal rivers coho salmon tend to spawn in
the upper reaches of a watershed (Shaul et al. 2007). Fry emerge from gravel the following
spring, approximately 4 to 7 months later (Murray and Beacham. 1988; Sandercock 1991;
DFO 2002). Fry emergence corresponds with periods of high discharge, and fry colonise
flooded habitats created by spring freshets (Sandercock 1991). Juveniles feed within
freshwater for one year and then migrate to the ocean as smolts (DFO 2005; Murray and
Beacham 1988). Studying early life-stage development is an important focus for
conservation and protection because environmental influences directly affect the
physiological function of developing coho salmon. Larval development is also of interest as
it is generally the time of highest mortality in salmonids (Quinn 2005).

Figure P.1. The distribution of coho salmon (Oncorhynchus kisutch) throughout the Pacific
Rim of North America and Asia (Sandercock 1991).
3

For my thesis I investigated the effect of temperature on a critical period of life
history and physiology of coho salmon – spawning and early juvenile development. Chapter
1 summarizes a survey of the temperatures experienced throughout incubation across the
geographical distribution of coho salmon within British Columbia. Chapter 2 investigated the
different life-history traits, and the environmental variables that affect them, found in
populations of coho salmon across their geographical distribution within BC. Specifically, I
examined environmental variables that may influence egg size, fecundity, female size and
reproductive investment (gonadal somatic index). From the environmental variables that I
found to influence life-history traits of coho salmon, I further examined incubation
temperature in Chapter 3. A controlled incubation study was conducted to test the effects of
incubation temperatures on the development of coho salmon from two populations, one from
northern BC and one from southern BC, with a specific focus on temperatures reaching nearfreezing during incubation. This research will provide a better understanding of how
temperature influences a species and how adaptive and plastic animals are to ranges in
temperature currently experienced. The potential response of fish to future changes in
climate are also of significant interest currently as temperatures are expected to change and
be less consistent from year to year – a concern for salmon found throughout watersheds
draining into the North Pacific Ocean.

4

CHAPTER 1
Variation in incubation temperature experienced by coho salmon (Oncorhynchus
kisutch) across their geographic range in British Columbia
Abstract
Coho salmon (Oncorhynchus kisutch) have a large geographic distribution throughout British
Columbia. It is expected that populations will experience highly variable temperatures
throughout their range, especially at the extremes of their distribution, but the temperature
regimes experienced by incubating coho salmon are not well understood. Currently, most
data on the conditions experienced by coho salmon throughout incubation have been
collected from surface waters, and not within the gravel environment of the redds. To gain a
better understanding of the temperature regimes experienced by coho salmon throughout the
incubation period, I surveyed both surface and intergravel temperatures within ten systems
across British Columbia. I found that incubation temperatures varied significantly across
British Columbia, which demonstrates the plasticity of this species as each population is
successful across such a wide range of temperatures. Moreover, all the populations spawn at
approximately the same time, despite the large geographic distances among them. Incubation
temperature within the gravel differed significantly from surface temperatures. This is an
important finding when attempting to extrapolate incubation temperature from surface water
temperature as even small differences can affect incubation and development because the
effects of temperature are cumulative. Considerable differences in accumulated thermal
units, therefore, will result from small differences in temperature. Consequently, it is
important to determine temperature regimes for each population, as future research needs to

5

be carried out on incubation temperatures within the redds themselves and how variation
influences larval development.
Introduction
Coho salmon (Oncorhynchus kisutch) have a large geographic distribution throughout
the north Pacific Rim. In North America, they can be found along the Pacific Coast from
California to Alaska and are found to spawn in over 970 rivers and small streams throughout
British Columbia (BC) (Sandercock 1991). With such a large geographic distribution, a wide
range of temperatures throughout spawning and incubation must be experienced by different
populations of coho salmon. Intergravel temperatures have been reported down to
approximately 3 ºC for south coastal populations in BC (Shepherd et al. 1986). Coho salmon
from an interior BC watershed, however, experience much lower temperatures during
incubation, which can approach freezing at the coldest part of winter (McRae 2009).
Habitat use for spawning is dependent on physical characteristics of the stream and
intergravel environment (McRae et al. 2012). An important factor for spawning site selection
of coho salmon is intergravel temperature (McRae et al. 2012). Temperature is of interest for
fisheries management as it has been found to be one of the leading factors in governing the
growth and development of fish (Fry 1971; Blaxter 1992). Temperature effects in
poikilotherms are cumulative as temperature has a direct effect on their metabolic rate, and
therefore rate of development (Fry 1971; Brett 1995). Cold temperatures cause a decrease in
metabolic and development rates that in turn cause a decrease in performance (Perry and
Tufts 1998). Consequently, ectothermic animals have evolved to cope with specific
temperature regimes, and relatively small temperature changes can have measurable effects

6

on community and population structure (Daufresne et al. 2009; Zuo et al. 2012; Paaijimans et
al. 2013).
Intergravel temperature within a stream is dependent on two factors: the infiltration of
stream water, and the upwelling of groundwater that together mix in the hyporheic zone.
Intergravel water temperatures can differ considerably from surface water temperatures, and
each system can vary based on the ratio of upwelling and downwelling (Freeze and Cherry
1979). Due to the relatively constant temperature of groundwater, intergravel temperatures
are generally found to be warmer in winter and cooler in summer than surface water.
Intergravel temperatures also show less response to diurnal heating than surface temperatures
and will have a lag effect as intergravel temperatures are buffered by the thermal mass of the
substrate and possible upwelling groundwater.
Little work has been done examining the range of temperatures that larval coho
salmon experience during incubation. In an earlier study, Murray and Beacham (1988) found
that survival rates of eggs and larvae from 13 populations of coho salmon from BC were
highest at approximately 4 ºC; complete mortality occurred above 14 ºC. The study by
Murray and Beacham (1988), however, may not represent the full range of temperatures that
coho salmon experienced during incubation in BC. McRae (2009) found that coho salmon
from an interior BC watershed experienced near-freezing temperatures of 0.3 ºC at the
coldest part of winter during incubation. Studies need to be expanded, therefore, to elucidate
the potential range of intergravel temperatures that coho salmon experience throughout their
distribution. Additionally, surface temperature is often used as a proxy for incubation
temperature, but the relationship between surface water temperature and redd temperature is
poorly understood. Thus, the objectives of this research were to 1) determine the range of

7

temperatures experienced throughout the incubation period by populations of larval coho
salmon across their distribution in BC, and 2) determine if surface temperature approximate
the intergravel temperatures experienced by coho during incubation.
Methods
Stream Temperature across British Columbia
Ten systems across BC were chosen to represent the range of locations where coho
salmon spawn (Figure 1.1). The systems were chosen based on their geographic position
(latitude and longitude) and migration distance in an attempt to characterize as much of the
distribution of coho salmon and potential physical environmental variability that exists
throughout the province. Populations chosen were from four coastal regions with short
migration distances but different latitudes (Kanaka Creek 49 ºN, Black Creek 50 ºN, Bella
Coola River 52 ºN, and Kitimat River 54 ºN), from three regions with intermediate migration
distances in central BC (Coldwater River 50 ºN, Nahatlach River 50 ºN, and Toboggan
Creek 55 ºN) and from three interior regions with long-distance migrations (Eagle River 51
ºN, Albreda River 52 ºN, and McKinley Creek 52 ºN).

8





)LJXUH7KHGLVWULEXWLRQRIFRKRVDOPRQLQ%ULWLVK&ROXPELDUHSUHVHQWHGDVEODFNFLUFOHV
DQGWKHWHQVWXG\V\VWHPVUHSUHVHQWHGDVUHGVTXDUHVIRXUFRDVWDOVKRUWPLJUDWLRQGLVWDQFH
SRSXODWLRQV ,.DQDND&UHHNž1-%ODFN&UHHNž1&%HOOD&RROD5LYHUž1DQG
$.LWLPDW5LYHUž1 WKUHHFHQWUDO%&SRSXODWLRQVZLWKLQWHUPHGLDWHPLJUDWLRQGLVWDQFHV
*&ROGZDWHU5LYHUž1+1DKDWODFK5LYHUž1DQG%7RERJJDQ&UHHNž1 DQG
WKUHHLQWHULRU%&SRSXODWLRQVZLWKORQJPLJUDWLRQGLVWDQFHV )(DJOH5LYHUž1($OEUHGD
&UHHNž1DQG'0F.LQOH\&UHHNž1 

)RUHDFKULYHUV\VWHPWHPSHUDWXUHORJJHUVZHUHSODFHGLQWKUHHGLIIHUHQWORFDWLRQV
ZLWKLQWKHV\VWHPZKHUHFRKRVDOPRQZHUHREVHUYHGWRVSDZQ6LWHVZHUHVHOHFWHGEDVHGRQ
KLVWRULFNQRZOHGJHRIVSDZQLQJORFDWLRQVSURYLGHGE\)LVKHULHVDQG2FHDQV&DQDGDDQG
HQKDQFHPHQWIDFLOLW\SHUVRQQHOGLUHFWREVHUYDWLRQRIWKHSUHVHQFHRIVSDZQLQJFRKRVDOPRQ
DWDORFDWLRQRUWKHSUHVHQFHRIUHGGVZKHUHFRKRVDOPRQKDGSUHYLRXVO\VSDZQHG)RUVRPH





of the systems the different locations were within the main stem river, but different reaches.
For other systems, the sample locations were located in different tributaries of the main stem
river.
Nine to twelve HOBO U-22 temperature loggers (Onset Computer Corporation,
Bourne MA) were deployed in each system in October and November prior to or during the
spawning period. Three intergravel temperature loggers and one surface temperature logger
were deployed at up to three sites within a system; sites were a minimum of three meters
apart to capture some of the potential variation in temperatures from different redds within
each system. Intergravel temperature loggers were buried approximately 25 cm upstream and
to the side of redds to minimize disturbance to eggs; 25 cm is the average depth of coho
salmon redds (Shephard et al. 1986; McRae et al. 2012). Surface temperature loggers were
placed on top of the gravel in a pool close to the intergravel loggers. Loggers recorded
temperature every hour for approximately six months throughout incubation from November
2012 to May 2013.
Statistical Analysis
Data collected from each logger were screened to remove data that indicated the
logger was not in the gravel or was out of the water. This could be seen as periods when
intergravel loggers had increased daily oscillations in temperature. Loggers that were out of
the water also exhibited increased daily oscillations and an overall decrease in temperatures.
Temperature regimes were analyzed using a repeated measure analysis of variance
(ANOVA) to assess differences in means of incubation temperatures among and within the
ten systems and their location in the system (gravel verses surface). Due to the requirement
for equal number of samples per system, a sample size of 5 was used (number of loggers

10

placed in the Nahatlatch River system). For systems with data from more than five
temperature loggers, five samples were chosen at random. Data from McRae (2009) was
used to supplement two additional sites within the McKinlay Creek samples. A single surface
temperature logger from each site within a system was paired with up to three intergravel
loggers. Little difference in temperature has previously been observed for surface water
temperature within a stream reach (McRae 2009; Williamson 2006). The assumption of
sphericity was not met, thus the measure of Greenhouse-Geisser was used. Two time periods
were used in the analysis; the temperature throughout the incubation period from November
22 to April 30, and the temperature during the coldest period of incubation – the month of
January. Significant temperature differences from intergravel loggers and surface loggers
between systems were also compared using post hoc Tukey test’s. Statistical analysis was
performed using SPSS statistical software (version IC 18; SPSS Inc., PASW Statistics,
Chicago, IL).
Results
The maximum, average and minimum temperature regimes for all ten systems
showed different trends over time (Figure 1.2). Most systems in my study were characterized
by a gradual decline in temperature until the coldest months of the winter (January and
February) and then a gradual increase in temperature towards the spring (Figure 1.2). The
rate of increase and decrease in temperatures, however, varied among systems. The duration
of the coldest temperatures also varied from a few weeks to more than a month (Figure 1.2).
Incubation temperatures for two of the systems were near-freezing throughout the
coldest period. McKinley Creek was the coldest system that coho salmon incubated in with
temperatures in all locations reaching 0.3 °C over the month of January. Toboggan Creek

11

reached a stable temperature of 0.7 °C early in incubation and maintained this temperature in
all locations until a warming period in late March (Figure 1.2). Coldwater, Eagle, Nahatlatch
and Bella Coola Rivers were found to be on average much warmer with coldest temperatures
only reaching approximately 3.5 °C. There was considerable variation in temperatures
among spawning sites within each of the warmest systems. Kanaka Creek was the warmest
system for the longest period of time, with a gradual decrease in temperature until a very
abrupt drop in temperature to 3 °C for a short period of time before an increase in
temperature.
Within each watershed intergravel temperatures were significantly warmer than
surface temperatures, for both the temperature throughout the entire incubation period (F1, 4 =
22.01, P = 0.009) and the temperature throughout the coldest period of incubation (F1, 4 =
28.96, P = 0.006) (Table 1.1; Figure 1.3). There was also a significant within-system effect
for both the temperature throughout the entire incubation period (F9, 44 = 24.353, P = 0.001)
and the temperature throughout the coldest period of incubation (F 9, 44 = 10.650, P < 0.001)
(Table 1.1). Additionally, there was a significant difference in incubation temperature among
the systems, with the southern coastal (Kanaka Creek, Black Creek, Nahatlatch River, Eagle
River and Bella Coola River) systems being significantly warmer than the northern coastal
and interior systems (Kitimat River, Toboggan Creek and McKinley Creek) for temperature
throughout the incubation period (F9, 44 = 1463 P < 0.001) and during the coldest part of
incubation (F9, 44 = 98468 P < 0.001) (Tables 1.1 and 1.2; Figure 1.3).

12

Figure 1.2: The running average (over 2 weeks) of the intergravel maximum, average, and
minimum temperature regimes found across ten systems across BC where coho salmon are
found to spawn. The systems are ordered based on latitude from the most northern on the top
down to the most southern systems on the bottom.
13




)LJXUH$YHUDJHVXUIDFH RSHQV\PEROV DQGLQWHUJUDYHO VROLGV\PEROV WHPSHUDWXUHV
WKURXJKRXWWKHLQFXEDWLRQSHULRG  IRUFRKRVDOPRQDFURVVDOOWHQV\VWHPV
PHDVXUHGLQ%ULWLVK&ROXPELD7KHDYHUDJHWHPSHUDWXUHWKURXJKRXWWKHLQFXEDWLRQSHULRGRI
1RYHPEHUWR$SULOLVUHSUHVHQWHGDVFLUFOHVDQGWKHFROGHVWSHULRG
WKURXJKRXWWKHLQFXEDWLRQSHULRG -DQXDU\ LVUHSUHVHQWHGDVVTXDUHVZLWKVWDQGDUGHUURUEDUV
7KHV\VWHPVDUHRUGHUHGEDVHGRQODWLWXGHIURPWKHPRVWQRUWKHUQWRWKHPRVWVRXWKHUQ
V\VWHPV 1 1RUWKHUQODWLWXGH0LG 0LGODWLWXGH6 6RXWKHUQODWLWXGH DQGWKHQUDQNHG
IURPLQWHULRUWRFRDVWDOZLWKLQHDFKRIWKHODWLWXGHV , ,QWHULRU& &RDVWDO 














Table 1.1: Results from repeated measures analysis of variance testing whether average
intergravel temperature during the incubation period (November 22, 2012 – April 30, 2013)
and the coldest portion of incubation (January) differed among spawning locations of coho
salmon in British Columbia. The variation is summarized based on the difference between
systems across British Columbia and the logger location within each system (intergravel redd
temperature and surface temperature), n = 5 pairs of surface and intergravel logger per
system with a total n of 50.
Incubation
Source of
SS
F
P
Temperature
Variation
Average
Within-Subject Effect
System
201.98
24.54
0.001
Logger location
3.31
22.01
0.009
System x Logger
2.307
0.983
0.414
location
Between-Subject
1032.36 1463.76 <0.001
Effect
Cold Period

Within-Subject Effect

Between-Subjects
Effect

System
Logger location
System x Logger
location

117.49
4.38
7.46

10.65
28.96
0.987

<0.001
0.006
0.396

339.84

98468

<0.001

Table 1.2: Summary of statistics comparing study systems determined by post hoc Tukey’s
Test. Systems that differed significantly are represented with an “X” for the average
intergravel temperature throughout incubation above the diagonal (November to April) and a
“#” for the coldest period of incubation below the diagonal (January). Populations chosen are
from four coastal regions with short migration distances but different latitudes (Kanaka
Creek (Kan) 49 ºN, Black Creek (Blk) 50 ºN, Bella Coola River (Bel) 52 ºN, and Kitimat
River (Kit) 54 ºN), from three regions with intermediate migration distances in central BC
(Coldwater River (Col) 50 ºN, Nahatlach River (Nah) 50 ºN, and Toboggan Creek (Tob) 55
ºN) and from three regions with long distance migrations within the interior of BC (Eagle
River (Eag) 51 ºN, Albreda River (Alb) and McKinley Creek (McK) 52 ºN). See Figure 1 for
locations.
System
McK Tob Alb Eag Nah Col
Kit
Bel
Bla
Kan
McK
Tob
Alb
Eag
Nah
Col
Kit
Bel
Bla
Kan

#
#

#
#

#

#

#

#

X
X
X

X
X

X
X

X

X
X
X

X

X
X

X
X
X

X
X
15

Discussion
This study showed that incubation temperature experienced by coho salmon in BC
differed significantly across populations (Table 1.1 and 1.2, Figure 1.3). Such differences
demonstrate the plasticity of this species, as all the populations spawn at approximately the
same time of year, despite the large geographic distance in locations and wide range of
temperatures. Temperatures in the hyporheic environment where the northern interior
populations spawn were near-freezing throughout the coldest period of incubation, January,
demonstrating an impressive physiological capacity of these fish (Figure 1.2 and 1.3).
Emergence time of fry is thought to differ across each of these systems as incubation periods
and growth rates of salmonids are dependent on the temperatures experienced within each
redd, specifically the number of accumulated thermal units (ATU) (Beacham and Murray
1990). Salmonid eggs incubating in warmer temperatures develop faster than those
incubating at cold temperatures as it takes longer to reach the required number of ATUs to
emerge from the redds. It is believed that, in general, populations of most species of salmon
spawn at specific times of the year to increase the chances of fry emergence coinciding with
the increased spring productivity (Burger et al. 1985). Populations spawning in colder
systems, therefore, will do so earlier to increase the length of incubation so development will
be completed for emergence at optimal spring productivity. Coho salmon, however, have a
relatively narrow spawning window compared to other salmonid species, with all populations
within BC spawning from October to December (Groot and Margolis 1991). Thus,
populations of coho salmon appear atypical compared to other species, as interior
populations are found to spawn when stream temperatures are already very cold, but at a
similar time of the year to southern populations. In contrast, other species of Pacific salmon

16

have a wider spawning period and interior populations often spawn earlier in the fall when
stream temperatures are warmer (Groot and Margolis 1991).
Intergravel water temperatures within each redd differed significantly from river
surface temperature, and were consistently warmer than surface temperatures throughout all
ten systems (Table 1.1, Figure 1.3). The effect of temperature on development of salmonids
is cumulative, resulting in considerable differences in ATU for very small differences in
water temperature. Thus, hatching and emergence dates could vary significantly with only a
small difference in temperature, especially for systems at near-freezing temperature where
developing embryos are likely to be highly sensitive to the temperature differences. For
example, 250 ATU are required to reach hatching, which takes 21 days at 0.5 °C, whereas
approximately 52 days are required at 0.2 °C. Consequently, surface water temperatures are
not likely to be a useful estimate of intergravel temperatures when analyzing incubation
conditions. Additionally, the temperature regimes found within the hyperheic zone are more
stable in comparison to the river surface temperatures, which have significant diurnal
variation as a result of several processes. The major factor influencing both air and stream
temperature is incoming solar radiation. Daily variation in stream temperature is related to
the amount of cover from riparian zones (Johnson and Jones 2000). Increased shading of a
stream is found to significantly decrease the maximum stream temperature by minimizing the
exposure of direct sunlight (Johnson 2004). Wind speed, relative humidity, subsurface
saturation of the bedrock and substrate all can influence stream temperature (Johnson 2003).
Stream temperatures can also be influenced by the amount of snowpack insulating the system
as well as anchor ice throughout the winter, which can freeze incubation sites (Cunjak and
Power 1986; Cunjak 1996). However, the hyperheic zone is found to be much more stable in

17

comparison as it is dependent on two factors: the infiltration of stream water and the
upwelling of groundwater that mix in the hyporheic zone. Intergravel water temperatures will
also differ considerably from surface water temperatures based on the ratio of upwelling and
downwelling water (Shepherd et al. 1986). Based on these patterns, Shephard et al. (1986)
also believed that intergravel temperatures would differ from surface temperatures, affecting
the precision of ATU and development rate predictions when relying on surface conditions.
Thus, it should not be assumed that intergravel redd environments can be characterized from
surface water variables – but knowledge of the hyporheic zone should be considered for fish
enhancement and habitat management projects.
Temperature within a system differed significantly (Figure 1.2, Table 1.1), suggesting
that the used spawning habitat selected within a system was not always consistent within a
population. A number of intergravel loggers in redds of coho salmon, however, were directly
above upwelling ground water, resulting in consistently warmer temperature (5 to 7 °C)
throughout the entire incubation period in comparison to other intergravel loggers deployed
in redds within the same system. Salmonids will often spawn in areas with discharging
ground water (Cunjak and Power 1986; Garrett et al. 1998), specifically when winter
incubation conditions are severe (Baxter and McPhail 1999). Offspring of salmonids found
spawning in locations with groundwater discharge have improved survival as redds
influenced by groundwater are not only warmer, but more stable than locations without
groundwater (Baxter and McPhail 1999; McRae et al. 2012). Groundwater, however, has
lower levels of dissolved oxygen (McRae et al. 2012), which is found to slow development
(Davis 1975). This suggests that enhanced development achieved by selecting warmer
temperatures may be offset by lower dissolved oxygen, which limits growth.

18

McRae et al. (2012), found a significant difference between used and unused
spawning habitat, suggesting that the small variation among sites within my study may
indicate selection of sites for spawning. Site selection would contribute to more uniform
offspring development, survival among individuals, and random survival within a
population. These factors are important for maintaining large effective population sizes
(Waples 1990a), which contributes to resiliency of a population to stochastic events that may
decrease numbers of fish. Waples (1990b) showed that low-frequency alleles are subject to
rapid extinction in Pacific salmon where effective population size is small and can lead to
inbreeding depression. Thus, populations with small effective sizes are at a much greater risk
for extinction (Newman and Pilson 1997), which may occur with little warning (Frankham
1995). Shrimpton and Heath (2003) demonstrated that available spawning habitat is
positively correlated with effective population size, suggesting that more area available for
spawning enhances population resilience. The relative effect on redd success is not known
from my study, but may lead to population-level effects on the effective size of breeding
stock. Temperature and site selection of individuals, therefore, can play a large role in
effective population size and should be considered in management practices.
Temperature is one of the most influential abiotic features affecting fish throughout
their life cycle. Stream temperatures have become a major issue and are at the centre of
policy debate, because elevated temperatures can negatively impact cold-water fish species,
such as threatened or endangered salmonids (Johnson 2003). Salmonids have evolved
temperature-specific life-history strategies to ensure that spawning occurs at a time of year
that will maximize incubation and emergence survival of their offspring (Quinn 2005). Thus,
an understanding of the full range of temperatures experienced by coho salmon throughout

19

incubation, and how populations have adapted and evolved throughout their distribution,
specifically for northern interior populations, is of interest both for management and
conservation initiatives. Future research should include long-term monitoring of intergravel
incubation temperatures experienced by coho salmon throughout a wider range of systems,
as my study only examined a single year and only 10 river systems. Stream temperatures can
vary from year to year, so a long-term monitoring study would increase our understanding of
temperature influences on development and survival of coho salmon. Also, quantifying
incubation temperatures within spawning sites and other apparently suitable, but unused,
sites within each system would provide a better understanding of whether coho salmon are
using spawning sites with specific temperature conditions to increase survival and
development throughout incubation, as suggested by McRae et al. (2012). Selecting for
specific spawning temperatures within a system could result in optimal emergence times in
relation to the spring freshet.

20

CHAPTER 2
Spatial and temporal variation in life-history strategies of coho salmon
(Oncorhynchus kisutch) populations throughout British Columbia
Abstract
Coho salmon (Oncorhynchus kisutch) have a narrow spawning window, from late October to
early December, throughout British Columbia. However, they have a very large geographic
distribution across British Columbia, and different environmental conditions are experienced
by populations throughout their range. Escapement for most populations is low in number
compared to other species of Pacific salmon, and so a better understanding of the
environmental variables influencing individual populations is needed to properly manage and
conserve this species. Here, I examined the effect of several environmental variables on egg
size, fecundity, female size and reproductive investment of populations of coho salmon from
across British Columbia. Egg size increased with higher latitude, warmer incubation
temperature, longer migration distance, more years of hatchery enhancement programs, and
smaller size of the spawning system. Both female size and fecundity increased with latitude,
warmer incubation temperature, and longer migration distance, but decreased with larger size
of spawning system. Gonadal somatic index decreased with higher latitude, warmer
incubation temperature and larger size of spawning system. Taken together, these results
reveal that latitude of spawning grounds, size of spawning system and temperatures
experienced by a population have a significant effect on shaping patterns of reproductive
investment. Thus, these three environmental variables should be considered when conserving
and developing management strategies for individual populations of coho salmon.

21

Introduction
Species with wide geographic distributions experience a range of physical factors, but
arguably temperature is the dominant factor for poikilotherms. I found that temperature
differed significantly throughout the geographical range within British Columbia that coho
salmon are known to spawn (Chapter 1). Understanding how individual organisms
successfully exploit habitats within their range is important to fully understand how to
manage the species not only at the population level, but throughout their geographic range.
This is especially important when working with and managing species of concern because
conservation efforts would benefit from a fuller understanding of the species' physiology,
ecology and life history. A species able to exploit a range of habitats throughout their
distribution may use different life-history traits and display different trade-offs to maximize
fitness (Lappalainen 2008). Specific differences in life-history traits among populations
during the spawning and incubation period are of particular interest in Pacific salmon
because the highest rates of mortality occur during the incubation and alevin stages (Groot
and Margolis 1991; Quinn 2005).
Svardson (1949) suggested that allocation of resources to egg production, defined as
the product of egg number and egg size, will be optimized but may vary due to local
selection pressures. There is, however, a trade-off between the size and number of eggs
produced, which precludes investing maximally in both traits simultaneously (Svardson
1949). The number of eggs, therefore, will vary in response to selective pressures both on
egg size and total investment in egg production (Fleming and Gross 1990). Previous work
has shown that fecundity of fish generally increases with latitude and increases with female
size (Leggett and Carscadden 1978; Beacham 1982; Fleming and Gross 1990; Beacham and

22

Morley 1985). Populations of coho salmon in Alaska have significantly higher fecundity than
lower latitude populations in British Columbia (BC) and California (Beacham 1982).
Fleming and Gross (1990) also found that total fecundity and egg size of Pacific salmon
increased with higher latitude and was independent of other influences such as competition
and migration distance.
Despite widespread distribution of coho salmon, most studies have focussed on
populations from the United States, the southern mainland of BC, and Vancouver Island.
While some studies have been conducted in Alaska, surprisingly these data were not
presented in several investigations (Beacham 1982; Fleming and Gross 1990; Beacham and
Murray 1993). Thus, data from northern and long-distance migrating interior populations of
coho salmon are needed to fully understand the variation in life-histories of coho salmon and
potential differences among populations.
Variation in life-history strategies among populations of coho salmon have been
linked to both spatial and temporal effects, but rarely have studies incorporated other
environmental factors that may influence life-history traits. Such variables include migration
distance, which is energetically costly for an individual, the temperature of the system, and
the size and type of the headwater system that feeds each system, as it will have an influence
on the temperature and productivity of each system. My objective was to understand how
life-history traits differ for populations of coho salmon, both spatially and temporally,
throughout British Columbia, and how these life-history traits relate to environmental
conditions found within the incubation habitat. To characterize habitat features that may
influence egg size, fecundity, female size, and gonadal somatic index, six environmental
variables were measured for locations throughout BC where coho salmon spawn. Spawning

23

systems were chosen based on available historic data which coincided with the systems in
Chapter 1. This research will provide a better understanding of how environmental variables
influence coho salmon populations across their distribution in BC, but also investigate how
environmental variables interact, which has not been addressed in previous studies.
Methods
Stocks used in the analysis and data collection
Life-history traits were compared for eight populations of coho salmon selected from
different regions across BC that represented a range of migration distances and potential
incubation conditions experienced by this species (Figure 2.1). Data were obtained from
Fisheries and Oceans Canada (FOC) or published sources (Beacham 1982; Fleming and
Gross 1900), and included female size (post orbital length), fecundity, and mean egg size.
Archived data were obtained from the Kitimat River Hatchery for the Kitimat population,
Snootli Creek Hatchery for the Bella Coola population, Spius Creek Hatchery for the Eagle
River and Coldwater River populations, Toboggan Creek Hatchery for the Toboggan Creek
population, Kanaka Creek Hatchery for the Kanaka Creek population, Quesnel River
Hatchery for the McKinley Creek population, and Big Qualicum Hatchery for Black Creek
population. Each population was assessed over multiple years (Table 2.1). Additional lifehistory data on mean fecundity and female size were also obtained from Beacham (1982) for
populations on Vancouver Island that were used in addition to the data from the Black Creek
(Table 2.1).
Data on egg size were available either as weight or volume measurements. I used egg
weight in my calculations and for consistency I transformed volume to weight using
Bonham’s (1976) equation:

24

𝑊𝑒𝑖𝑔ℎ𝑡

𝑔
𝑚𝑙
= 𝑣𝑜𝑙𝑢𝑚𝑒
𝑥 1.076
𝑒𝑔𝑔
𝑒𝑔𝑔

Gonadal somatic index (GSI), a ratio of gonad weight to somatic weight, was calculated by:

𝐺𝑆𝐼 =

𝑔
𝑊𝑒𝑖𝑔ℎ𝑡 𝑒𝑔𝑔 𝑥 𝐹𝑒𝑐𝑢𝑛𝑑𝑖𝑡𝑦
𝑇𝑜𝑡𝑎𝑙 𝐹𝑒𝑚𝑎𝑙𝑒 𝑊𝑒𝑖𝑔ℎ𝑡

Female weight, however, was not collected for each population, but instead data on length
were collected. Based on a condition factor of 1.06 g · cm–3 (CF; 100 W · L–3), determined
from Kitimat River population females (Chapter 3), length was used to calculate an
approximate total weight. A CF of 1.06 for mature female coho salmon falls within the range
reported by Scholz et al. (2011) for spawning female coho salmon in the Puget Sound
Lowlands.

25



)LJXUH/RFDWLRQVZKHUHGDWDZHUHFROOHFWHGIRUHQYLURQPHQWDOYDULDEOHVDQGOLIHKLVWRU\
WUDLWVRIFRKRVDOPRQDFURVV%ULWLVK&ROXPELDIRUWKLVVWXG\/RFDWLRQDUHODGOHGEDVHGRQWKH
V\VWHPVXVHGWKURXJKRXW&KDSWHU7KHPLVVLQJV\VWHPVDUHGXHWRODFNRIDYDLODEOHOLIH
KLVWRU\GDWD


7DEOH6XPPDU\RIWKHDYDLODEOHGDWDFROOHFWHGIURPHDFKV\VWHPIRUHDFKOLIHKLVWRU\
WUDLW HJJVL]HIHFXQGLW\IHPDOHVL]HDWPDWXULW\JRQDGDOVRPDWLFLQGH[>*6,@ 7KHUDQJHRI
VDPSOH\HDUVDQGWRWDOQXPEHURI\HDUVFROOHFWHGIRUHDFKOLIHKLVWRU\WUDLW
6\VWHP

/LIHKLVWRU\7UDLW Q 
<HDUV
(JJ )HFXQGLW\
)HPDOH
*6,

6L]H
6L]H
0F.LQOH\



1$
1$
7RERJJDQ





.LWLPDW





%ODFN
 1$


1$
&ROGZDWHU





(DJOH





%HOOD&RROD



1$
1$
.DQDND










Model Development
The effects of environmental variables on life-history traits were analyzed using
truncated regression models and negative binomial regression to determine variations across
each system. Truncated regression models were used to examine variation in egg size, female
size and gonadal somatic index as they are continuous data with the exclusion of values
below zero. Negative binomial regressions were used to explain the variation in fecundity as
the data were over-distributed count data. Variance inflation factors (VIF) were examined for
each independent variable to determine if they resulted in multicollinearity within models.
Variables with a VIF greater than 10 were considered to have a high degree of
multicollinearity and not used in the model with other highly collinear variables. Categorical
variables were adjusted to create a design matrix to correct for linear dependencies in each
model. System was used in each model as a random factor to account for non-independence
of data for each system.
Model Parameters
Candidate models were developed from a set of predictor variables recognized as
important for explaining variation in egg size, fecundity, female size, and gonadal somatic
index of coho salmon. A total of seven predictor variables were used: system temperature,
which was the average temperature (°C) throughout the entire incubation period; average
gravel temperature (°C) throughout the coldest period of incubation (January); average
temperature (°C) throughout the peak of spawning; latitude; migration distance from the
ocean (km); number of years of hatchery operation; system size (ranked as large or small
based on the size of the system where coho salmon are found to spawn); and type of
headwater influence (Table 2.2).

27

Table 2.2: List of environmental variables used in analyses of variation in life-history traits of coho salmon.
Model Parameter
Description
Categories
Reference
Temperature
January Temperature – average gravel
Continuous
Murray and Beacham 1988
temperature within each system throughout the
Shepherd et al. 1986
coldest period of incubation (average throughout
Clarke and Johnston 1999
January) determined in Chapter 1.
McRae et al. 2012

Latitude

Migration Distance
Stream Size
Headwater Type

System Average – average gravel temperature for
each system throughout the duration of
incubation from November - May determined in
Chapter 1.

Continuous

Murray and Beacham 1988
Clarke and Johnston 1999
McRae et al. 2012

Spawning Temperature – average river surface
temperature for each system throughout the peak
spawning period of November determined in
Chapter 1.
The average latitude measured across the three
study sites within each system studied in Chapter
1 were used as the latitudinal coordinate of each
system. Average latitude was calculated from
GPS coordinates taken at each site with an
accuracy of ± 3 meters.
The average migration distance in kilometers
traveled from the ocean to the spawning site in
each system.
The size of stream where coho were found to
spawn based on study sites used for the
temperature data collection in Chapter 1.
The source of headwaters that feeds into each
system.

Continuous

Murray and Beacham 1988
McCullough 1999

Continuous

Leggett and Carscadden 1978
Beacham 1982
Fleming and Gross 1990
Chavaries et al. 2010

Continuous

Groot and Margolis 1991
Brett 1995

Small (1)
Big (2)

Rasenfeld et al. 2000

Snow melt (1)
Lake (2)

28

Model Selection
An information theoretic approach was used to evaluate competing models for
explaining the influence of environmental variables on life-history traits. Akaike’s
Information Criterion, corrected for small sample sizes (AICc), was used to rank models
within candidate model sets (Burnham and Anderson 2002). The model with the lowest AICc
in a candidate model set was considered the most parsimonious model, but those models
within 2 AICc units from the best approximating model were considered competitive
(Burnham and Anderson 2002). Akaike weights (ωi), the relative likelihood that the
candidate model is the best model given the data and the model set (Burnham and Anderson
2002), are reported for each model. I interpreted parameter estimates from a single top model
within the candidate model set when ωi of the top model exceeded 0.90; in cases where there
was model selection uncertainty (i.e., ωi < 0.90), I calculated model-averaged estimates of
parameter estimates using all models within the candidate model set with ∆AICc < 2
(Burnham and Anderson 2002). Model-averaging minimizes the effect of uninformative
parameters (confidence intervals that overlap 0) with little impact on the bias and precision
of the more useful parameter estimates (Arnold 2010). Statistical analysis was performed
using STATA statistical software (version IC 12; StataCorp, College Station, TX).
Results
Egg Size
The best approximating model for egg size variation suggested that latitude, January
incubation temperature, migration distance, number of years that hatchery enhancement
programs were run, and size of system were influential (Table 2.3). The second ranked
model was similar except that it did not include the number of years that hatchery

29

enhancement programs were run. The null model for egg size and all other life-history traits
were found to do very poorly. Model-averaged estimates suggested that egg size increased
with higher latitude (β = 12.16, SE = 4.22, 95% CI [3.84 to 20.37]), longer migration
distance (β = 0.19, SE = 0.03, 95% CI [0.11 to 0.26]), and with decreasing size of a system
(β = -55.17, SE = 10.81, 95% CI [-76.37 to -33.98]). Warmer January temperatures (β =
4.82, SE = 5.91, 95% CI [-6.77 to 16.41]) and the number of years of hatchery enhancement
to the system (β = 0.56, SE = 0.33, 95% CI [-0.09 to 1.22]), however, were uninformative as
the confidence intervals overlapped zero.
Fecundity
Environmental variables in the best approximating model explaining variation in
fecundity included latitude and January incubation temperature. The second ranked model
was similar, but also included size of system. The third competitive model included size of
system and migration distance in addition to latitude and January temperature (Table 2.4).
Model-averaged estimates indicated that latitude (β = 0.06, SE = 0.02, 95% CI [0.02 to
0.09]), January incubation temperature (β = 0.085, SE = 0.03, 95% CI [0.02 to 0.14]) had a
positive effect on fecundity. Migration distance (β = 0.001, SE = 0.0002, 95% CI [-0.0003
to 0.0006]) and size of system (β = -0.01, SE = 0.01, 95% CI [-0.028 to 0.007]), however,
were uninformative as the confidence intervals overlapped zero.
Female Size
For size of female coho salmon at maturity, the best approximating model included
latitude, spawning temperature throughout November, and size of system (Table 2.5). A
second competitive model was similar, but also included migration distance. Modelaveraging indicated that size of females increased with higher latitude (β = 10.44, SE = 0.51,

30

95% CI [9.38 to 11.38]) and spawning temperature (β = 7.33, SE = 7.33, 95% CI [6.08 to
8.57]). Migration distance (β = -0.002, SE = 0.002, 95% CI [-0.005 to 0.001]) and the size of
the river system (β = -18.10, SE = 0.96, 95% CI [-19.97 to -16.22]), however, were
uninformative as the confidence intervals overlapped zero.
Gonadal Somatic Index
The best approximating model for GSI included January incubation temperature and
size of system. The second competitive model was similar, but also included latitude (Table
2.6). Model-averaging suggested that all three variables were negatively related to GSI.
Gonadal somatic index was lower with warmer incubation temperatures (β = -5.93, SE =
1.15, 95% CI [-7.99 to -3.50]), and in larger systems (β = -4.48, SE = 0.83, 95% CI [-5.86 to
-2.61]). Latitude was uninformative as the confidences intervals overlapped zero (β = 0.08,
SE = 0.41, 95% CI [-0.87 to 0.72]).

31

Table 2.3: Summary of AICc ranking of candidate models for environmental variables
influencing egg size of coho salmon. Lat = Latitude, JanG = average temperature found in
each system throughout the coldest period of incubation (January), SysAvg = average
temperature found in each system throughout the incubation period from November to April,
YOH = total number of years an enhancement hatchery for coho salmon has been in
operation within each system, Mig = migration distance to each system from the ocean, Size
= size (using ranks) of each river, HeadW = type of headwater that feeds each system.
Model Parameters
K
AICc
∆AICc
ωi
Lat + JanG + Mig + YOH + Size
6
893.3
0.0
0.486
Lat + JanG + Mig + Size
5
893.7
0.3
0.413
Lat + JanG + Mig + YOH + HeadW
6
898.1
4.8
0.045
Lat + HeadW
3
900.2
6.8
0.016
Lat + JanG + Mig + HeadW
5
901.7
8.3
0.008
Lat
2
902.1
8.8
0.006
Lat + JanG + HeadW
4
902.2
8.8
0.006
Lat + Size
3
902.2
8.8
0.006
Lat + SysAvg
3
903.7
10.4
0.003
Lat + YOH
3
903.8
10.4
0.003
Lat + JanG
3
904.1
10.8
0.002
Lat + Mig
3
904.2
10.8
0.002
Lat + JanG + Size
4
904.3
11.0
0.002
Lat + JanG + YOH
4
905.0
11.6
0.001
Lat + JanG + Mig
4
905.9
12.5
0.001

32

Table 2.4: Summary of AICc ranking of candidate models for environmental variables
influencing fecundity of coho salmon. Lat = Latitude, JanG = average temperature found in
each system throughout the coldest period of incubation (January), SysAvg = average
temperature found in each system throughout the incubation period from November to April,
YOH = total number of years an enhancement hatchery for coho salmon has been in
operation within each system, Mig = migration distance to each system from the ocean, Size
= size (using ranks) of each river, HeadW = type of headwater that feeds each system.
Model Parameter
K
AICc
∆AICc
ωi
Lat + JanG
3
1589.3
0.0
0.275
Lat + JanG + Size
4
1590.7
1.4
0.137
Lat + JanG + Mig + Size
5
1590.8
1.5
0.130
Lat + JanG + YOH
4
1591.3
2.0
0.101
Lat + JanG + Mig + YOH + Size
6
1591.3
2.1
0.098
Lat + JanG + Mig
4
1591.5
2.2
0.093
Lat + JanG +HeadW
4
1591.5
2.2
0.093
Lat + JanG + Mig + YOH
5
1593.
4.1
0.035
Lat + JanG + Mig + YOH + Size + HeadW
7
1593.7
4.4
0.031
Lat + JanG + Mig + HeadW
5
1593.7
4.4
0.031
Lat + JanG + Mig + YOH + HeadW
6
1595.4
6.1
0.013
Lat + SysAvg
3
1596.4
7.1
0.008
Lat + YOH
3
1607.4
18.1
0.000
Lat + Mig
3
1614.3
25.0
0.000
Lat
2
1615.2
26.0
0.000

33

Table 2.5: Summary of AICc ranking of candidate models for environmental variables
influencing female size of coho salmon. Lat = Latitude, SpTemp = average temperature
found in each system throughout the peak spawning period in November, SysAvg = average
temperature found in each system throughout the incubation period from November to April,
YOH = total number of years an enhancement hatchery for coho salmon has been in
operation within each system, Mig = migration distance to each system from the ocean, Size
= size (using ranks) of each river, HeadW = type of headwater that feeds each system.
Model Parameters
K
AICc
∆AICc
ωi
Lat + SpTemp + Size
4
685.2
0.0
0.689
Lat + SpTemp + Mig + Size
5
687.4
2.2
0.230
Lat + Size
3
690.1
4.9
0.059
Lat + SpTemp + Mig
4
692.3
7.1
0.020
Lat + Mig
3
702.5
17.3
0.001
SpTemp + Size
3
702.9
17.7
0.000
Lat + SpTemp + YOH
4
704.3
19.1
0.000
0.000
Size
2
704.6
19.4
0.000
Lat + HeadW
3
708.1
22.9
0.000
Lat + SpTemp + HeadW
4
709.4
24.2
0.000
Lat + JanG
3
712.1
26.9
0.000
Lat + YOH
3
713.6
28.4
0.000
Lat
2
724.0
38.8
0.000
Lat + SysAvg
3
726.1
40.9
0.000
SpTemp + Mig
3
736.2
51.0

34

Table 2.6: Summary of AICc ranking of candidate models for environmental variables
influencing gonadal somatic index of coho salmon. Lat = Latitude, JanG = average
temperature found in each system throughout the coldest period of incubation (January),
SysAvg = average temperature found in each system throughout the incubation period from
November to April, YOH = total number of years an enhancement hatchery for coho salmon
has been in operation within each system, Mig = Mig distance to each system from the
ocean, Size = size (using ranks) of each river, HeadW = type of headwater that feeds each
system.
Model Parameters
K
AICc
∆AICc
ωi
JanG + Size
3
348.8
0.0
0.711
Lat + JanG + Size
4
351.1
2.2
0.230
Lat + Size
3
356.4
7.5
0.016
Lat + YOH
3
356.6
7.7
0.015
JanG + YOH
3
357.2
8.4
0.011
Lat + JanG + YOH
4
357.5
8.7
0.009
JanG + Mig
3
359.1
10.2
0.004
Lat + JanG + Mig
4
361.1
12.3
0.002
JanG + HeadW
3
362.0
13.5
0.001
JanG
2
362.7
13.8
0.001
Lat + JanG + HeadW
4
364.7
15.8
0.000
Lat + JanG
3
364.8
15.9
0.000
Lat + Mig
3
366.2
17.3
0.000

35

Discussion
This study represents the first examination of how environmental factors affect lifehistory variation in a wide-ranging species which incorporated not only physical variables
that fish experience during incubation across a range of latitudes, but also migration distance.
A number of previous studies have examined life-history variation among populations at
different latitudes, but the populations have been limited to coastal watersheds or short
migration distance into interior river systems (Fleming and Gross 1990; Beacham and
Murray 1993; Braun et al. 2013). In my study, data were collected from populations of coho
salmon that migrate approximately 20 km to stocks that spawn more than 750 km from the
ocean. Coho salmon are a good model organism for this study due to this wide geographic
distribution, but also their relatively narrow spawning window from late October to early
December. A general pattern observed was that females from populations in northern
systems were larger and more fecund with larger eggs compared to female spawners from
southern systems. Coho salmon from northern interior populations that migrate longer
distances were also found to have larger eggs than southern coastal stocks and are, therefore,
investing considerably more energy into not only somatic growth, but also reproductive
potential. Such a finding contrasts with the theory of reproductive investment trade-offs
presented in earlier literature for anadromous salmon – that with an increase in latitude
number of eggs increases, but egg size and total biomass of eggs produced decreases
(Beacham 1982; Fleming and Gross 1990; Beacham and Murray 1993). The variables that
affect the life-history traits examined in this study and potential reasons for the apparent
disparity from earlier work are explored below.

36

Egg Size
Egg size of coho salmon was found to increase in more northern populations, but also
with greater migration distance – the opposite trend presented in the studies by Fleming and
Gross (1990) and Beacham and Murray (1993). The geographic distribution in the study by
Fleming and Gross (1990) was relatively narrow and did not extend into the northern
distribution of coho salmon within BC compared to the geographical distribution that was
covered within my study. Beacham and Murray (1993) examined latitudes from 47° to 55°,
but did not incorporate migration distance in their analysis. My study examined latitudes
from 49° to 55° with migrations distances up to 700 km into the northern interior of British
Columbia. The earlier studies suggested egg size decreased with increased latitude as a tradeoff for an increase in the number of eggs produced allowing reproductive investment to be
similar across latitudes but increase offspring fitness. With fewer juvenile competitors,
increased spring productivity, and size selective gradients in higher latitudes it was suggested
that selection for larger egg size was reduced.
Fleming and Gross (1990) also suggest that the decrease in egg size at higher
latitudes is due to the decreased incubation temperatures, and the efficiency of conversion of
yolk to body tissues and offspring size relationships found by Johnston and McLay (1997).
Within my model egg size increased with an increase in temperature experienced throughout
the coldest period of incubation, but the 95% confidence intervals overlapped zero,
indicating incubation temperature throughout January on its own was not influential. It is
surprising that temperature did not influence egg size due to the direct effect of temperature
on metabolic rate and yolk conversion efficiency (Stickland et al. 1988; Johnston and McLay
1997; Kileen et al. 1999). Hyperstatic growth of white muscle fibers was found to decrease

37

with increased incubation temperatures, thus decreasing the efficiency of conversion of yolk
to body tissues (Stickland et al. 1988; Johnston and McLay 1997; Kileen et al. 1999). With
the decrease in conversion efficiency in warmer temperatures, salmon will produce smaller
alevin and fry for a given egg size at higher temperatures, thus larger eggs in warmer
incubation temperature will maximize overall growth (Stickland et al. 1988; Johnston and
McLay 1997). Murray and Beacham (1988) also found that northern and mainland stocks
incubated at low temperatures (1.5 to 2.0 °C) were found to have larger, heavier alevin and
fry than southern and coastal populations. My findings that temperature was not influential
may be due to how cold incubation conditions were for some of the populations of fish
examined. When temperatures reach near-freezing these trends may reverse as fish are found
to develop at rates faster than expected; less ATU are required to reach the same
development stage (Williamson 2006; Murray and Beacham 1988; Chapter 3). Metabolic
rates of fish incubating at near-freezing temperatures are not well understood. Development
at such cold temperatures, however, must be metabolically costly.
The fact that years of hatchery operation was in the top model suggest that this
variable is important, however, it also proved to be uninformative. The number of years of
hatchery enhancement within a system was previously shown to have a positive effect on egg
size. Fleming and Gross (1990) and Quinn et al. (2004) found coho salmon eggs were larger
for hatchery populations than wild populations. The data of Fleming and Gross (1990),
however, was later reanalyzed by Murray and Beacham (1993); addition of a regional
variable resulted in a lack of difference among populations. In contrast, Heath et al. (2003)
and Haring et al. (2016) found that enhancement programs relaxed selective pressure for

38

larger eggs in Chinook salmon, driving selection for decreased gonadal investment, increased
fecundity and smaller eggs.
My findings also suggest that females spawning in smaller systems have larger eggs –
opposite to previous studies. It was suggested that smaller systems will experience less
bedload movement resulting in smaller substrate as the lower flow will not wash out fine
sediments from among the larger gravel and boulders (Knighton 1998). Smaller eggs would
be selected for in systems with smaller gravel size, which have reduced oxygen level within
the gravel as they have reduced oxygen demand and more efficient surface-to-volume ratio
in comparison to larger eggs (Quinn et al. 1995; Rollinson and Hutchings 2011). Smaller
systems within this study were found to be warmer coastal systems with relativity good
spawning substrate, suggesting that the temperature influence and available nutrition on
emergence may have a bigger influence on egg size. Larger systems such as the Nahatlatch
River, Eagle River and Coldwater River had smaller gravel in areas where redds were
located, thus these finding may be correlated with the effects of gravel size on egg size.
Stream size, therefore, is a variable that is difficult to relate to life-history traits.
Fecundity
Fecundity increased with latitude and migration distance consistent with a number of
previous studies. The increase in fecundity with latitude has been suggested to be a
reproductive trade-off to produce more, smaller eggs rather than few, large eggs (Beacham
1982; Fleming and Gross 1990; Beacham and Murray 1993). This reproductive trade-off,
however, was not found among the coho salmon populations in my study, which have an
overall increase in reproductive investment, as the northern populations are found to have
increased fecundity and egg size with increased latitude – although latitude was included in

39

the model for GSI, surprisingly the effect was not found to be influential (see below).
Fecundity of sockeye salmon (Oncorhynchus nerka) has also been found to increase with
increased female size (Braun et al. 2013), consistent with my findings as large females were
found to spawn farther north and at longer migration distances.
Temperature had a positive influence on fecundity such that females whose offspring
had warmer incubation temperatures were more fecund. Systems with higher temperatures
were found along the coast in warmer, wetter climates, with longer growth seasons. Longer
growing seasons with higher productivity may support more individuals with less
competition for resources. The model also suggested that females spawning within smaller
systems had higher fecundity. Smaller rivers in my study were mostly found within the
coastal warmer systems. Conversely, systems with longer migration distances and colder
temperatures have longer and harsher winters resulting in shorter growth periods with lower
productivity. Lower fecundity may be offset by fewer, larger eggs to lower competition
within lower productivity systems and increase available energy reserves for initial
development. Although size of system and migration distance from the ocean have the
potential to affect fecundity, these two variables were found to be uninformative in my
model.
Female size
Female size and morphology have previously been shown to be influenced by total
reproductive investment (higher in larger females), migration distance (thinner, smaller fish
with longer caudle peduncle in streams with longer migration distance), and competition for
redd location on spawning grounds (more pronounced kype, larger body size, and brighter
colouration with increased competition) (Fleming and Gross 1989). Braun et al. (2013) found

40

that larger female sockeye salmon produced more large eggs than smaller females – an
overall increase in reproductive investment. Informative variables in my model were that
female size increased with latitude, increased with spawning temperature, and decreased with
size of system. The previous models for egg size and fecundity showed an increase in
fecundity and larger egg size with higher latitude, demonstrating increased reproductive
investment of larger females at higher latitudes. Smaller females also spawned within larger
systems, a pattern consistent with the previous models that less reproductive investment was
allocated by spawners in larger systems as both fecundity and egg size decreased. Van Den
Berghe and Gross (1989) found female size contributed to overall fitness in three ways,
increase initial biomass of egg production, the ability to acquire a high-quality territory for
redds, and success in nest defense. Thus, in smaller systems that may have limited spawning
habitat and higher competition, larger females may be more successful. The type of head
water influence to a system may also influence the productivity and temperature of a system
suggesting that females are larger in lake fed systems (Rasenfeld et al. 2000) – however,
system type was not in any of my top models.
My study used length of female as a measure of size due to data availability.
Although my model suggested that length of female decreased with increased migration
distance, the large confidence intervals suggest this variable was uninformative. Smaller
females have been found from stocks that migrate longer distances; Fleming and Gross
(1989) reported that both body depth and length of female coho salmon decreased with
increased migration distances. Taylor and McPhail (1985) found that interior Fraser juvenile
coho salmon have a more fusiform (streamlined) body shape, which may persist throughout
life into adult migration, in comparison to costal populations. Adults in my study from

41

interior populations with more energetically demanding migrations also may have distinctive
fusiform morphology as it is superior for sustained swimming compared to costal
populations having more robust bodies (Taylor and McPhail 1985; Fleming and Gross 1989).
To better understand the effect of environmental variables on female size condition factor
and overall morphology should be considered in the future where data are available.
Gonadal Somatic Index
Gonadal somatic index decreased for fish spawning in systems that were larger and
had warmer incubation temperatures. Potentially females from larger systems were allocating
a smaller proportion of their energy budget to reproduction and defending somatic size.
Larger systems could have higher discharge resulting in more difficult migration for an
individual. Larger systems may also have larger substrate due to increased bedload
movement with higher discharge. Thus, more somatic energy may be needed to ensure the
females are large enough to migrate and spawn in larger systems. Colder incubation
temperatures were also found for the northern interior systems where not only did the eggs
incubate at colder temperature, but temperature was colder for spawning and most likely for
migration. Braun et al. (2013) found that gonadal mass decreased with increased migration
difficulties for sockeye salmon – a finding consistent with my study as migration is more
difficult for adult salmon at warmer temperatures (Farrell et al. 2008). Additionally, less
energy is required for metabolic processes at colder temperature, thus there may be limited
benefit of producing large eggs, allowing females to allocate energy to other processes such
as producing more eggs and somatic growth (Stickland et al. 1988; Johnston and McLay
1997). Latitude was found to be uninformative within the top models for GSI, however, it is
a significant variable for all other life-history traits examined. With latitude significantly

42

influencing egg size, fecundity and female size, one would think it would also be a
significant environmental driver of GSI. Unfortunately, limited data were available for GSI.
An increased data set with increased latitudinal distribution may result in a significant
influence of latitude being detected.
Conclusion
Care must be taken when interpreting data as a few limitations do arise when working
with large data sets compiled from enhancement programs and previous studies. The
availability of life-history data was inconsistent across the populations of interest, resulting
in uneven representations of populations across each of the life-history models. Dependent
on the data collected by hatcheries in the past and the year sampling occurred within a
hatchery, data for specific life-history traits were not always available for each population or
for consecutive years, resulting in variation in sample sizes. However, a minimum of ten
years for each population was used whenever possible to create a robust data set. One area of
interest for further investigation would be to include the age of spawners in addition to the
life-history traits incorporated in the present study, as this may also vary within and among
populations due to environmental influences.
With these considerations in mind, the data suggests that based on geographic
distribution, three environmental variables have the most influential on reproductive
investment of coho salmon. The models reveal that latitude of spawning grounds, size of
spawning system and temperatures experienced by a population have a significant effect on
shaping patterns of reproductive investment throughout British Columbia. Thus, these three
environmental variables should be considered at a population level when conserving and
developing management strategies for individual populations.

43

CHAPTER 3
The effects of near-freezing incubation temperatures on development, metabolic rate
and yolk utilization in coho salmon (Oncorhynchus kisutch)
Abstract
Populations of coho salmon experience incubation temperatures that differ across their
distribution in British Columbia; from near-freezing for some populations to above 6 °C for
others. Incubation at near-freezing temperatures demonstrates an impressive physiological
capacity, and may represent local adaptation to increase survival. To investigate potential
adaptation to cold temperatures, an incubation study was conducted on a northern and a
southern population of coho salmon with pure and hybrid families incubated at three
different temperature regimes (cold reaching 0.5 °C, mid at 4.5 °C, and warm at 9 °C). The
near-freezing incubation temperatures resulted in less accumulated thermal units to reach
each development stage, lower condition factors, lower resting metabolic rate and more yolk
available for somatic growth than the warmer incubation temperatures. There was little
difference between the two populations, however, suggesting little local adaptation for
development at cold temperatures. Larval development, however, was strongly influenced by
maternal origin, which was associated with egg size; smaller eggs from the southern
population produced smaller offspring that developed faster than offspring from the northern
population with larger eggs. Coho salmon exhibit exceptional phenotypic plasticity as little
difference was found between families from the northern and southern populations.

44

Introduction
Temperature is one of the most influential variables governing reproductive
investment. In Chapter 2, I showed the importance of temperature for fecundity, female size
and gonadal somatic index, but temperature also governs growth, development and survival
of salmon at other stages of development. Even small changes have profound effects on the
development rate of fish as the effect of temperature is cumulative. This is particularly
evident for larval fish; intergravel temperature throughout early development of coho salmon
(Oncorhynchus kisutch) was found to be important for spawning site selection and
development (McRae et al. 2012; Chapter 2). For a species such as coho salmon with large
geographic distributions, considerable differences in temperature are experienced, yet the
effect of such differences are not clear. In the killifish (Fundulus herteroclitus), latitudinal
differences in populations have been shown to translate into differences in thermal tolerance
(Fangue et al. 2006). Adaptive phenotypic change or phenotypic plasticity may be crucial for
population persistence where selection pressures are altered. Use of spawning sites by
salmonids has been examined in two interior BC systems, McKinley Creek and Davis River.
These studies found that surface waters and even intergravel temperatures within redds of
coho salmon and bull trout (Salvelinus confluentus) were close to freezing for long periods
(McRae 2009; Williamson 2006). The development of embryos at such extreme temperatures
represents an impressive physiological capacity within a species and is not yet fully
understood. Fish that are adapted to cold temperatures have higher metabolic rates (Johnston
et al. 2000) and proportionally faster rates of development at a given temperature
(Williamson 2006). This increase in metabolism required for survival at extreme
temperatures near-freezing may come at a cost to other measures of performance.

45

Temperature tolerance of larval Pacific salmon has been examined previously
(Murray and Beachem 1988; Beacham and Murray 1990; Jonsson and Jonsson 2011;
Whitney at al. 2014). Murray and Beachem (1988) suggested that Pacific salmon appear to
adapt to specific spawning temperatures to maximize survival and size of fry (Murray and
Beachem 1988). Their study on coho salmon used 13 populations from BC, reared at five
different temperatures from 1.5 to 15 ºC, and showed survival was highest between 4 and 5
ºC and complete mortality occurred above 14 ºC. The study by Murray and Beachem (1988),
however, may not represent the full range of temperatures that coho salmon experience
during incubation in British Columbia. For example, McRae (2009) found that coho salmon
from an interior BC watershed experience near-freezing temperatures that averaged 0.3 ºC at
the coldest part of winter during incubation. Additionally, I showed that temperature regimes
experienced throughout incubation by coho salmon across BC differed significantly, ranging
from 0.1 to 5 ºC on average throughout the coldest period of incubation (Chapter 1).
Energy during larval development is limited by yolk supply, which must be
partitioned among metabolism, growth, development and activity (Callow 1985; Rombough
2006). Yolk protein serves two main functions: amino acids for tissue growth and energy for
catabolic processes (Heming and Buddington 1988). Energetic trade-offs experienced by
larval fish may arise when yolk reserves necessary for growth and development are used for
activity associated with movement and maintaining metabolic rates (Brett and Groves 1979;
Callow 1985). It is also important to understand how energy is allocated between somatic
growth and metabolic processes, as the rates of absorption of yolk are important determinants
of development, growth, and survival in larval fish (Heming and Buddington 1988).
Transforming yolk to body tissue as efficiently as possible should have fitness consequences

46

because larger larvae are expected to be stronger swimmers and less susceptible to predation
(Heming and Buddington, 1988). At colder temperatures, less energy is needed to maintain a
stable metabolic rate, resulting in more available energy to be allocated to somatic growth.
Thus, species at higher latitudes in colder environmental conditions will theoretically be able
to grow larger as there is more energy available for growth.
To gain a better understanding of physiological adaptations that may occur at
extremely cold temperatures experienced throughout incubation and early development, a
controlled incubation study was conducted. I examined the effect of incubation temperature
on the number of accumulated thermal units required for development, resting metabolic
rate, yolk availability for somatic growth, size of fish, and survival. To test whether local
adaptation had an effect on these variables, two populations were examined; one from
southern and one from northern BC. Further, both pure strain and reciprocal hybrid families
were created to assess maternal and paternal influences from both populations.
Methods
Gamete collection and breeding design
This study was conducted from November 2013 to June 2014. Two populations of
coho salmon with short migration distances (approximately 50 km), but from different
latitudes (a southern BC population, Kanaka Creek 49 ºN, and a northern BC population,
Kitimat River 54 ºN), were used to minimize the influence of migration distance on
development. Gametes from three males and three females were collected from both the
Kanaka Creek Hatchery and the Kitimat River Hatchery and transferred to the Aquatic
Animal Holding Facility at the University of Northern BC on November 22, 2013. Once at
the Aquatic Animal Holding Facility, eggs were fertilized to create a replicated two by two

47

design with pure and hybrid families to determine treatment effect on each family. One male
and one female from each population were crossed to create six pure and six hybrid families
(Appendix 3.1). Each family was divided into six groups, one for each temperature regime
and replicated twice within each temperature treatment.
Experimental treatments
Three incubation temperature regimes were used; a “cold” treatment which
simulated the coldest temperatures experienced by coho salmon throughout BC (McKinley
Creek), a “mid” treatment that was similar to the warmest temperatures (Kanaka Creek and
Bella Coola River; see Chapter 1), and a “warm” treatment which was increased at the same
rate that the cold treatment was decreased, until it reached close to the upper temperature
tolerance range for larval coho salmon (9 °C). Incubation commenced at the same
temperature (4.5 °C) for all three treatments to minimize stress on eggs when first transferred
to the incubators. Temperature was then gradually altered to reach the target temperatures for
each incubator (approximately 0.5 °C for the cold, 4.5 °C for the mid and 9 °C for the warm).
Temperatures were increased in the spring for the cold treatment to mimic the natural
warming of temperature regime of McKinley Creek (Figure 3.1).

48


)LJXUH7HPSHUDWXUHUHJLPHVIROORZHGWKURXJKRXWWKHLQFXEDWLRQSHULRGIRUFRKRVDOPRQ
XQWLOEXWWRQXS UHG ZDUPJUHHQ PLGDQGEOXH FROGWUHDWPHQW 7KHDSSUR[LPDWHGDWH
RIKDWFK + DQGEXWWRQXS % DUHVKRZQIRUHDFKWHPSHUDWXUHUHJLPH

7KUHH+HDWKVW\OHLQFXEDWRUVZHUHXVHGRQHIRUHDFKLQFXEDWLRQWHPSHUDWXUHUHJLPH
HDFKZLWKDODUJHSODVWLFEDUUHO / DVDKHDGZDWHUUHVHUYRLUFRQWDLQLQJD'RU'
UHIULJHUDWLRQXQLW )ULJLG8QLWV,QF7ROHGR2+ WRFRQWUROWHPSHUDWXUH3ULQFH*HRUJH
PXQLFLSDOZDWHUWUHDWHGZLWKDTXDULXPZDWHUFRQGLWLRQHUWRUHPRYHFKORULQHZDVXVHGLQ
HDFKUHVHUYRLUDQGLQWURGXFHGWRWKH+HDWKVWDFNVIURPWKHUHVHUYRLUEDUUHOE\JUDYLW\SDVVHG
WKURXJKWKHVWDFNDQGSXPSHGEDFNLQWRWKHUHVHUYRLUE\DQDTXDULXPSXPSSURYLGLQJ
FRQWLQXRXVFLUFXODWLRQRIZDWHUWKURXJKWKHV\VWHP(DFKLQFXEDWRUFRQWDLQHGIRXU+HDWK
WUD\VWKHWRSRQHWRFDWFKWKHRXWIORZIURPWKHUHVHUYRLUEHIRUHIORZLQJGRZQRYHUWKHHJJV
WKHPLGGOHWZRZHUHGLYLGHGLQWRVHFWLRQVHDFKKROGLQJGLIIHUHQWIDPLOLHVLQFXEDWHGDWWKH
VDPHWHPSHUDWXUHDQGWKHERWWRPWUD\SDVVHGWKHZDWHULQWRDVHFRQGUHVHUYRLUZKHUHLWZDV
SDVVHGWKURXJKDILOWHUDQGWKHQSXPSHGEDFNLQWRWKHLQLWLDOUHVHUYRLU
+HDWKVWDFNVZHUHFRYHUHGLQEODFNSODVWLFWREORFNDPELHQWOLJKWIURPWKHHJJVDQG
ODUYDH7HPSHUDWXUHGLVVROYHGR[\JHQS+DQGFRQGXFWLYLW\ZHUHPHDVXUHGGDLO\WRDVVHVV
ZDWHUTXDOLW\2QVHW8GDWDORJJHUVZHUHSODFHGLQHDFKUHVHUYRLUDQGZLWKLQHDFK+HDWK




tray to measure temperature hourly throughout the study to detect any potential positional
effects on the temperature experienced by the fish. Water was changed every 1-2 days (40 L
per stack, total volume of each system was 280 L) to maintain water quality.
Size, Condition Factor and Survival
Samples were collected from each of the families at three development stages: one
week after fertilization, hatch, and button-up. The hatching date was determined when 50%
of the replicate was hatched; the period from hatch of the first alevin to complete hatch for
the replicate did not differ among families within the same treatment, but it did vary among
treatments as eggs in the cold treatment were much slower to hatch. Button-up was
determined when the fish in each family had completely absorbed the yolk sack.
Accumulated thermal units (ATU) were calculated for each stage of development as the
number of days each family took to reach a particular development stage multiplied by the
mean daily temperature. Once the eggs had eyed, mortalities were removed daily to
determine survival rates for each stage of development. For development measurements, two
fertilized eggs were sampled from each family, while four to six samples (based on
population sizes of each replicate, as mortality rates were high for a few replicates) were
taken from each replicate in each treatment for the last two development stages. Length (if
possible) and weight was measured for each development stage, and a condition factor was
calculated for hatch and button-up as 100(W · L–3).
Oxygen Consumption
Oxygen consumption rates (µg O2 · h–1) were measured on three fish from each
replicate (six fish per family per treatment) at hatch and button-up for the middle and cold
treatments using intermittent flow respirometry. The respirometry equipment was not fully

50

set-up and running in time to collect measurements for the warm treatment groups. An eightchamber respirometer was used to measure six fish simultaneously, while two were left
empty for each trial and used as a control. Horizontal mini glass chambers (length 35mm, ID
9mm, CH10645, Loligo Systems, Tjele, Denmark) were fitted with mini fibre optic O2 sensor
spots (5 mm, X11050, Loligo System), which do not consume oxygen.
At the beginning of each set of trials, fish were placed into individual glass chambers
which were covered with black plastic to keep the chambers dark, and left to acclimatize and
reach a resting state for an hour before oxygen consumptions rates were measured. The
temperature bath and chambers were filled with water directly from the treatment incubator
with recirculating pumps to maintain the temperature as close as possible to the rearing
temperatures. The systems were filled prior to adding any fish to ensure the O2 spots were
completely wet before measurement. Two peristaltic pumps (PU10650, Loligo Systems)
were used to flush the chambers with fresh oxygenated water from the temperature bath and
two others used to create closed circuits with 1 mm (ID) tubing (TU11350, Loligo Systems).
The pumps were left running for an hour prior to the first measurement to allow the fish to
acclimatize to the chambers and reach resting metabolic rate. The flushing pumps were then
shut off and oxygen consumption measured until oxygen levels within the chambers reached
70% of saturation, at which time the flushing pumps were turned back on to exchange the
water within the chamber with 100% saturated water for 45 min. For each trial, fish were
measured until a minimum of two runs were recorded with a linear decrease in oxygen in
each chamber. Temperature and oxygen levels were recorded every 2 seconds using the
Witrox 4 oxygen analysers (OX11875, Loligo Systems) and fibre cables (OX11150, Loligo
Systems) which read the O2 sensor spots. Measurements were converted with the DAQ-M

51

instrument (AR12500, Loligo Systems) and AutoResp respirometry software (AR12600,
Loligo System). The length of time each test was run was dependent on the rate of oxygen
consumption within the chamber, and thus varied for each trial and treatment. Resting
metabolic rates were calculated as the decrease in oxygen in the chamber based on the
chamber volume and divided by the mass of the fish within the chamber. The temperature
coefficient (Q10) which represents the factor by which the rate of a reaction increases for
every ten degree rise in temperature was measured for each family between each treatment.
The temperature coefficient is defined as:
10

𝑅2 (𝑇2 −𝑇1)
𝑄10 = ( )
𝑅1
where Q10 is the factor by which the reaction rate increases when the temperature is raised by
ten degrees. Q10 is a unitless quantity. R1 is the measured reaction rate at temperature
T1 (where T1 < T2). R2 is the measured reaction rate at temperature T2 (where T2 > T1).
Yolk Area Analysis
To determine yolk area at hatch, images were taken of four to six fish from each
population within each treatment stages using a MEIJI microscope and Motic Image Plus
software. All images were taken at 1.4x magnification and the head was adjusted and
maintained at the highest point possible on the microscope to ensure consistent results. Total
length of fish, yolk area and yolk perimeter were measured in millimeters using the Motic
Image Plus software. Similar methods have used imaging software to measure yolk crosssectional area (Wells and Pinder 1996; Marty et al. 1997; Boucher 2012). Relative yolk area
was measured by dividing the yolk area at hatch by the average egg area of the population.

52

Statistical Analysis
The data for weight, length, condition factor, yolk area, resting metabolic rate, and
survival were collected using a partial factorial design with repeated measurements of each
population nested within each temperature treatment. I used a nested three-way analysis of
variance (ANOVA), with origin of maternal gamete and paternal gamete nested within
treatment, for all measurements except resting metabolic rate at hatch for which I used a
nested two-way ANOVA comparing the pure families. The two replicates for each family
within a temperature treatment were combined because there was no difference in the results
with all replicates analysed individually. Family groups were then used as replicates to test
for effects of maternal and paternal origin. Maternal and paternal effects were analysed for
each female and male cross however, families did not differ from one another, therefore for
all subsequent analysis maternal origin and paternal origin were analysed by population. All
assumptions for analysing each data set with a three-way ANOVA, (normal distribution,
homoscedasticity and independence of observations) were assessed before analysis (Gotelli
and Ellison 2004). For each measurement, the data met all the assumptions except variances
were heteroscedastic. Transformations of the data (log +1, ln, and square root) did not
achieve homoscedasticity. ANOVA, however, is robust to unequal variance when the sample
size is large (Zar 1996). Tukey post hoc tests and pairwise comparison with an adjustment
for unbalanced samples sizes were run to determine the specific significant differences
present in each model. Statistical analysis was preformed using STATA statistical software
(version IC 12; StataCorp, College Station, TX).

53

Results
Development Rate
All progeny in the mid and warm treatments required significantly more ATU to
reach hatch than the cold treatment (F2, 35 = 3438, P < 0.001; Figure 3.2a). There was a
significant maternal (F1, 35 = 576, P < 0.001) and paternal (F1, 35 = 7.48, P = 0.012) origin
effect, and a maternal by paternal interaction (F1, 35 = 231, P < 0.005) effect on the number of
ATU to reach hatch. Additionally, there was a significant interaction effect between all three
variables (treatment, maternal origin and paternal origin) on the number of ATU required to
reach hatch (F2, 35 = 4.95, P = 0.01). The progeny of the pure northern origin fish required
significantly more ATU to reach hatch than the progeny from the southern maternal origin
within all three treatments. Summary tables for all 3-way ANOVA results are presented in
Appendix 3.2.
The number of ATU required to reach button-up was significantly greater with
increased temperature treatment (F2, 35 = 14020, P < 0.001; Figure 3.2b). At button-up,
however, there was no longer a significant interaction between all three variables (F2, 35 =
1.31, P = 0.28), but there was a significant interaction between treatment and maternal origin
(F2, 35 = 50.91, P < 0.001), as well as treatment and paternal origin (F2, 35 = 4.18, P = 0.03).
The progeny of the northern maternal origin required significantly more ATU to reach hatch
than the progeny from the southern maternal origin within the cold and mid treatments
(Figure 3.2b).
Size and Condition Factor
The weight of fish at hatch was strongly dependent on maternal origin (F1, 364 =
222.35, P < 0.0001) and treatment (F2, 364 = 7.04, P < 0.001). The progeny of the northern

54

females were significantly heavier than the progeny of the southern females within all three
temperatures (Figure 3.3a). Families reared in the mid temperature regime were significantly
larger than families reared in cold and warm temperature regimes (Figure 3.3a). The effect of
maternal origin by treatment was significant for length (F2, 364 = 10.79, P < 0.0001) and
condition factor (F2, 364 = 10.00, P < 0.0001). Progeny of southern females reared in the cold
temperature regime were significantly shorter than fish reared at the mid and warm
temperature regimes (Figure 3.3b). There was less difference for progeny of northern
females, however, the fish reared at the mid temperature regime were significantly longer
than cold and warm treatments. Condition factor also differed with treatment; the
significantly shorter progeny from northern females reared at cold and warm temperatures
resulted in significantly higher condition factor for these groups (Figure 3.3c).
The relationships found at button-up, however, differed from the findings at hatch.
The interaction between all three variables (treatment, maternal origin and paternal origin)
for weight at button-up, however, was not significant (F2, 666 = 2.96, P = 0.053). There was a
significant interaction between maternal and paternal origin for weight (F1,666 = 16.47, P <
0.0001), such that the progeny of females of the northern origin were heavier than from
females of the southern origin, and a paternal effect with hybrid families for both north and
south origins were lighter than the pure families (Figure 3.4a). The effect of treatment was
also significant for weight at button-up (F2, 666 = 26.76, P < 0.0001). Families reared in the
warm temperature regime were larger than families reared in cold temperature regime and
the families reared in the mid temperature regime were intermediate (Figure 3.4a). At buttonup there was a significant interaction between maternal and paternal origin for length (F1, 666
= 7.95, P = 0.005) and also a significant interaction between treatment and maternal origin

55

for length (F2, 666 = 8.01, P < 0.0001); progeny of northern females reared at the coldest
temperature regime were the longest (Figure 3.4b). The paternal effect was due to the pure
strain families being heavier and longer than the hybrid families (Figure 3.4). For condition
factor, there was also a significant treatment by maternal effect (F2, 666 = 6.80, P < 0.002). At
button-up families incubated within the cold treatment were lighter but longer than families
incubated within the mid and warm treatments which resulted in the lowest condition factors
(Figure 3.4). Although, families from warm treatment groups at button-up were heaviest and
exhibited the highest condition factor, fish within the cold treatment were lighter, but longer
and exhibited the lowest condition factor at button-up; opposite to the findings at hatch
(Figure 3.3c and 3.4c).
Oxygen Consumption
Resting metabolic rate was measured at hatch for the cold and mid temperature
treatment groups. Temperature for the cold group was approximately 0.5 ºC and resting
metabolic rate was significantly lower than the mid temperature treatment at 4.5 ºC (F1, 279 =
186.43, P < 0.0001; Figure 3.5a). The Q10 values at hatch between the mid (4.5 ºC) and the
cold (0.5 ºC) treatments was 8.4 for the Kitimat Pure families and 8.2 for the Kanaka Pure
families. There was no maternal or paternal origin effect on resting metabolic rate at hatch.
Interestingly, at button-up there was a significant three-way interaction between treatment,
female, and male origin (F1, 326 = 16.51, P < 0.0001, Figure 3.5b). The northern hybrid family
within the mid treatment had a significantly higher metabolic rate and was the only family
found to be significantly different than all other families within both treatments. Despite the
different thermal histories experienced by the two groups, there was minimal treatment effect

56

on the resting metabolic rate as oxygen consumption was measured at similar temperatures
for each treatment (Figure 3.5b).
Yolk Area
The yolk area at hatch was found to differ significantly between treatments (F2, 315 =
19.72, P < 0.0001), being significantly smaller in the warm treatment than the cold and mid
treatments (Figure 3.6a). Yolk area was also strongly dependent on the maternal origin (F1,
315 = 328.74, P < 0.0001) as the progeny of females from the northern stock, both pure and

hybrid families, had significantly larger yolks than the individuals with a southern maternal
gamete. Standardizing yolk area based on initial egg size, however, the maternal effect was
no longer significant. There was still a significant treatment effect (F2, 315 = 19.11, P <
0.00001) with the warm treatments having significantly smaller yolk area than the cold and
mid (Figure 3.6b).
Survival
Temperature had no effect on overall survival. Survival for all families in all three
treatments was over 50%. There was a significant maternal by paternal interaction on total
percent survival (F1, 71 = 7.90, P < 0.006; Figure 3.7), such that the progeny of the southern
paternal origin had significantly lower overall survival than the progeny of the northern
paternal origin. Across all three treatments the northern hybrid families had the lowest
overall percent survival. Both the southern pure and hybrid families had very high variation
in survival among replicates (Figure 3.7). Timing of mortalities, however, differed across the
treatments, with the most mortalities occurring between hatch and button-up within the warm
treatments for all four families. Most mortalities in the cold temperature treatments occurred
prior to hatch for all families.

57

Figure 3.2: The required accumulated thermal units to reach hatch (a) and button-up (b) for
the different pure and hybrid families of coho salmon reared at cold, mid and warm
temperature regimes (see Figure 3.1). Families are represented as female X male pairs from a
northern population (N: Kitimat River) or a southern population (S: Kanaka Creek). The
upper-case letters above each bar in figure a) represent the significant difference between the
three way interactions of treatment, maternal and paternal origin. The letters above each bar
in figure b) represent the significant differences determined by the two-way interaction of
treatment and maternal origin and the treatment and paternal origin.
58

Figure 3.3: Weight (a), length (b) and condition factor (c) at hatch for the different pure and
hybrid families of coho salmon reared at cold, mid and warm temperature regimes (see
Figure 3.1). Families are represented as female X male pairs from a northern population (N:
Kitimat River) or a southern population (S: Kanaka Creek). The upper-case letters above
each bar represent the significant treatment and maternal origin interactions. Data are shown
as means + standard error.

59

Figure 3.4: Weight (a), length (b), and condition factor (c) at button-up for the different pure
and hybrid families reared at cold, mid and warm temperature regimes (Figure 3.1). Refer to
Figure 3.3 for detailed description of the legend. The upper-case letter above each bar
represent the interaction effect of treatment and maternal origin and the interaction effect of
maternal and paternal origin. Figure a) has a three-way interaction effect of treatment,
maternal origin and paternal origin and is described within the text. Data are shown as means
+ standard error.

60

Figure 3.5: Resting metabolic rate at hatch (a) and button-up (b) for the different pure and
hybrid families reared at cold (0.5 ºC at hatch, 4.5 ºC at button-up) and mid (4.5 ºC at hatch
and button-up) temperature regimes (Figure 3.1). Refer to Figure 3.3 for detailed description
of the legend. The upper-case letters above each bar represent a) the significant difference in
treatment and b) the three-way interaction effect of treatment, maternal origin and paternal
origin. Data are shown as means + standard error.

61

Figure
3.6: The yolk area (a) and standardized yolk area (b) at hatch for the different pure and
hybrid families reared at cold, mid and warm temperature regimes (Figure 3.1). Refer to
Figure 3.3 for detailed description of the legend. The solid line above each treatment
represents the significant difference from the treatments without a solid line. The upper-case
letters above each bar represent the significant difference between maternal origins. Data are
shown as means + standard error.

62

Figure 3.7: The total percent survival for the different pure and hybrid families reared at
cold, mid and warm temperature regimes (Figure 3.1) from fertilization to button-up. Refer
to Figure 3.3 for detailed description of the legend. The upper-case letters above each bar
represent the interaction effect of maternal and paternal origin. Data are shown as means +
standard error.

63

Discussion
This study is one of the first to examine the effects of near-freezing temperatures
during incubation on the development of coho salmon. A number of previous studies have
examined the effect of temperature on the development of Pacific salmon including coho
salmon, but temperatures below 1.5 ºC have previously not been achieved (Heming 1982;
Murray and Beacham 1988; Beacham and Murray 1989; Beacham and Murray 1990;
Johnston et al. 2000; Whitney et al. 2014). In my study, incubation temperature regimes
near-freezing were found to have a more profound effect on the development of coho salmon
at hatch than the warmer incubation temperature regimes. Development, resting metabolic
rate and yolk absorption of fish incubating in the cold treatment at hatch differed
significantly from the warmer treatments. However, these differences decreased as the fish
reached button-up and the cold treatment temperature regimes were increased gradually to 5
ºC. Additionally, survival was similar and above 50% across all three treatments,
demonstrating the plasticity of this species and the ability to develop and mature successfully
despite the thermal history experienced by the fish.
Larval development was also found to be strongly influenced by the maternal origin
but not as strongly by paternal origin. Maternal effects on early development are associated
with the maternal investment in egg size such that smaller eggs will produce smaller
offspring that develop faster than offspring from larger eggs (Heming 1982; Beacham 1988;
Murray and McPhail 1988; Heath et al. 1999).
Development Rate
Eggs in the cold treatment required two months longer to reach hatch and three months
longer to reach button-up than in the warm treatment (Figure 3.2). The eggs in the cold

64

treatment, however, required significantly less ATU (approximately half) to reach hatch and
button-up in comparison to the mid and warm treatments (Figure 3.2). More rapid
morphological development at warmer temperatures, however, required a greater number of
ATU, which is well documented in the literature (Heming 1982; Beacham and Murray 1985;
Murray and Beacham 1986; Murray and McPhail 1988; Whitney et al. 2014). With colder
systems typically having a longer winter period with decreased temperatures and later spring
freshet, productivity and food availability would consequently not be available until later in
the spring/summer compared to warmer coastal systems. Thus, a longer development period
and later emergence would align with these natural processes and increase the survival
prospects of juveniles. The date of hatch and fry emergence was found to be dependent on
the environment during incubation (Beacham 1988) such that fry from smaller eggs emerge
sooner than those from larger eggs (Beacham and Murray 1985; Heath et al. 1999). This
trend was also present within my study as the families with the southern maternal origin and
smaller eggs reached button-up earlier than the families with the northern maternal origin
and larger eggs in all three treatments.
The rate of development and efficiency of converting yolk into somatic growth may
not have been consistent throughout the development period for the cold treatment.
Individuals at hatch were found to be significantly smaller than the individuals in the warm
and mid treatment. However, size differences among the temperature treatments at button-up
were reduced or the opposite trends were detected. Length at button-up within the cold
treatment was greater than the warm treatment, which may be because alevin use their yolk
sacs less for growth and more for basal metabolism (Rombough 1994; Kamler 2008;
Whitney et al. 2014). This increase in size of fish within the cold treatment may also suggest

65

that as temperatures increase throughout the spring, development may accelerate resulting in
a compensation for slow growth in the cold treatment and ultimately less difference in size
among the treatment temperatures
Size and Condition Factor
Size at hatch was strongly influenced by an interaction between treatment and
maternal origin, while at button-up the interaction between treatment and maternal origin
was present, a significant interaction between maternal origin and paternal origin was also
found (Figure 3.3 and 3.4). Within the literature incubation temperatures have been found to
have a significant effect on the size of Pacific salmon throughout early development
(Peterson 1977; Beacham 1985; Murray and McPhail 1988; Murray et al. 1990), a finding
consistent with my study. Lower incubation temperatures produce larger alevin and fry
(Peterson 1977; Beacham 1985; Murray and McPhail 1988; Murray et al. 1990). For coho
salmon, alevin and fry were largest at 2 ºC (Murray and McPhail 1988), while Murray et al.
(1990) found fry to be largest at mid temperatures ranging from 4-8 ºC. I found that fish were
longer and slightly heavier at hatch in the mid treatment at 4.5 ºC (Figure 3.3). At button-up,
however, this pattern changed with fish in the cold treatment being the longest and fish in the
warm treatment being the heaviest (Figure 3.4).
Maternal effects on progeny body size and early development are well documented in
the literature (Heming 1982; Beacham 1988; Murray and McPhail 1988; Heath et al. 1999).
For Chinook salmon (Oncorhynchus tshawytscha), however, this maternal effect is only
present during early development prior to emergence, and decreases post-hatch, becoming
negative at the beginning of emergence when paternal effects become dominant (Heath et al.
1999). Beacham (1988) also found that size of fry of pink (O. gorbuscha) and chum (O. keta)

66

salmon was less influenced by maternal egg size than alevin size characters, and the hatch
time was dependent on the environmental conditions during incubation. In my study,
maternal origin had a significant effect on progeny size at hatch, but at button-up a
significant interaction between the paternal effect and the maternal effect was present. This
transition between a maternal effect throughout early development and the importance of egg
size, to a paternal effect later in development also occurs in coho salmon populations. For
Chinook salmon, Heath et al. (1999) suggested that the negative maternal effect observed at
emergence is a result of different size eggs having different hatch dates and growth rates,
such that progeny from smaller eggs hatch earlier and grew faster than progeny from larger
eggs. Two mechanisms were suggested to explain these patterns: differential feeding
behavior, such that smaller progeny from smaller eggs will attempt to compensate for their
small size and feed more aggressively, and/or progeny from small eggs emerge earlier than
those from large eggs and have longer exogenous feeding time. Both mechanisms would lead
to increased risk of predation for progeny from small eggs, but these would overall gain a
growth advantage (Heath et al. 1999). My fish were not fed exogenously, but a paternal
effect was found for NxN and NxS families; it is possible that if I had continued the
experiment longer a greater paternal effect would have been observed.
Condition factor can be used to determine fitness of fish. At hatch fish from the cold
and warm treatments had higher condition factors compared to the mid treatment, and there
was a significant maternal effect as progeny of northern maternal origin had higher condition
factor at hatch and button-up (Figure 3.3 and 3.4). The maternal effect was likely due to the
larger initial size of eggs of the northern female. Condition factor at button-up was more

67

consistent across all three temperatures, although fish from the cold treatment had
significantly lower condition factor than the other treatments.
Oxygen Consumption
Temperature has a doubling effect on metabolic rate when body mass is taken into
account; metabolic rate increases with increased temperature in teleost species (Clarke and
Johnston 1999) including Pacific salmon such that metabolic rate increased 2-3 (Q10) times
with every 10 ºC increase in temperature (Brett 1971; Brett 1995; Perry and Tufts 1998).
Resting metabolic rate at hatch within the cold treatment was significantly lower than within
the mid treatment as expected (Figure 3.5). At colder temperatures there is a decrease in
utilisation of energy towards maintaining a stable metabolic rate, resulting in more available
energy to be allocated for somatic growth (Heming 1982). The Q10 value was found to be
three times more the expected trend in the literature with a Q10 value of approximately 8 for
all families. Such a finding would suggest that the expected increase in reaction of 2-3 with
every 10 ºC increase is lost when the individuals are incubated at the extreme ranges in
temperature.
At button-up, the pattern of differences in metabolic rate among families within the
cold and mid treatment groups was quite complex (Figure 3.5). For the most part, however,
resting metabolic rate did not differ between families or the cold and mid temperature
treatments. The temperature regimes at button-up for the cold treatment had reached 5 ºC
following the natural spring warming. Consequently, resting metabolic rate was not
influenced by the thermal history experienced by an individual, just the temperature at the
time of measurement. Resting metabolic rates have been found to increase significantly
between hatch and emergence when temperatures remained consistent throughout incubation

68

(Rombough 2011), which was also found in this study as metabolic rate increased at buttonup in the mid treatment which had a constant temperature throughout incubation. Although
prior thermal history did not have an effect on metabolic rate, differences in condition factor
and size of fish at hatch and button-up from the different temperature treatments might
reflect prior rates of oxygen consumption. Weight of fish was lowest from the coldest
temperature treatment, suggesting that less maternal energy stores were used for somatic
growth in this group. Development at temperatures close to freezing potentially was less
efficient metabolically, reducing the number of ATU required for growth.
Yolk Absorption
Yolk area at hatch was smallest for all four family groups reared in the warm treatment
indicating that at warmer temperatures yolk was absorbed at a faster rate than at colder
temperatures (Figure 3.6). Beacham and Murray (1989) found that yolk weight at hatch
increased with decreasing temperatures. Heming (1983) found that the duration of yolk
absorption and energy available for somatic growth was reduced at higher incubation
temperatures, consistent with the metabolic rate measurements conducted at hatch in my
experiment. This finding suggests that more energy was used to maintain higher metabolic
rates in the warm treatment than the cold treatment. Although size at hatch did not differ in
my study, smaller yolk area for the individuals within the warm treatment suggests that less
of the yolk used was allocated to somatic growth and more used for routine basal metabolism
(Beacham and Murray 1989; Rombough 1994; Kamler 2008).
The amount of yolk present is determined ultimately by ovum size (Heming 1982).
Yolk area in my study was affected significantly by the maternal stock as the individuals
with northern maternal gametes had significantly more yolk present at hatch than progeny

69

from southern females (Figure 3.6). After relative yolk area was calculated whereby yolk
area was scaled to the initial egg size of each female, the maternal effect was no longer
present. Therefore, maternal effects are likely due to differences in initial investment in egg
size, and the rate of yolk absorption may be similar among families within the same
treatment. This trend was also found by Heming (1982) as efficiency and rate of yolk
absorption was dependent largely on rearing temperature after measurements were scaled to
initial egg size.
Survival
Despite all of the differences found in development across the three treatments, there
were no significant differences in overall survival among the treatments, demonstrating the
plasticity and resiliency of this species to extreme temperatures (Figure 3.7). Additionally,
survival was greater than 60% for all families within the cold treatment, also demonstrating
the physiological capacity of coho salmon to survive a wide range of incubation
temperatures, including near-freezing temperatures. Survival of most Pacific salmon are
found to decrease as temperatures reach 3 ºC or less and increase with increased temperature
until an upper thermal limit of approximately 12 ºC (Murray and McPhail 1998; Beacham
1990), but for coho salmon survival was above 80% at temperatures below 2 ºC (Beacham
1988; Murray and McPhail 1988). The paternal origin of progeny had a significant effect on
survival as the progeny with the southern paternal origin had significantly lower survival
across each treatment (Figure 3.7), suggesting that the genetic contributions from southern
male were weaker despite the maternal origin or temperature treatment. Paternal origin was
found to influence survival post emergence (Heath et al. 1999) and influence stress response
and resistance to pathogens (Nadeau et al. 2009). Interestingly, an influence of paternal effect

70

has been linked to behaviour; the offspring of less aggressive males have been found to have
higher survival (Peterson and Järvi 2007). Additionally, cryptic female choice has been
found in salmonid populations, such that eggs differentially influence the fertilization success
of sperm from different males (Rosengrave et al. 2016). My results demonstrate that this may
be occurring as the southern males had lower fertilization success and lower overall survival
– but this was particularly dramatic for the pairings with northern females.
Other than the difference in egg size that resulted in significant maternal effects in my
study, there was remarkably little difference between families from the northern or southern
populations in the variables measured. Paternal effects were seen in the sizes of some of the
families at button-up, but these were subtle. Interestingly, the progeny from the northern
populations did not show significantly improved performance in the cold temperature
regime. Such a finding would suggest little evidence of local adaptation – rather exceptional
phenotypic plasticity within this species is occurring.

71

$33(1',;



)LJXUH$/D\RXWRIWKHEUHHGLQJGHVLJQWKDWZDVXVHGWRVWXG\WKHHIIHFWRIWHPSHUDWXUHRQ
WKHGHYHORSPHQWRIODUYDOFRKRVDOPRQ7KUHHPDOHVDQGWKUHHIHPDOHVZHUHFROOHFWHGIURP
WZRSRSXODWLRQVZLWKVKRUWPLJUDWLRQGLVWDQFHVWRVSDZQLQJORFDWLRQVEXWIURPGLIIHUHQW
ODWLWXGHV2QHSRSXODWLRQZDVIURPDVRXWKHUQV\VWHP .DQDND RI%&ZLWKZDUPLQFXEDWLRQ
FRQGLWLRQVUHSUHVHQWHGE\WKHUHGKLJKOLJKWHGEORFNVDQGRQHSRSXODWLRQZDVIURPD
QRUWKHUQV\VWHP .LWLPDW ZLWKFROGLQFXEDWLRQFRQGLWLRQVUHSUHVHQWHGZLWKWKHEOXH
KLJKOLJKWHGEORFNV6L[SXUHDQGVL[K\EULGIDPLOLHVZHUHFUHDWHGE\IHUWLOL]LQJRQHIHPDOHE\
RQHPDOHIURPRQHRUERWKSRSXODWLRQV)DPLOLHVFUHDWHGDUHVKRZQZLWKWKHZKLWHEORFNV






NxN

NxS

SxN

SxS

F3

F2

F1

Figure A.2: Layout of the top Heath tray with the first replicate of each population of coho
salmon in each of the three incubators at Hatch. F1, F2, and F3 refer to the different families;
placement for families in the Heath trays differed to limit any positional effect on the
variables measured. N (northern) and S (southern) represent the population of each female x
male crossed within each cell.

73

NxN

NxS

SxN

SxS

F2

F1

F3

Figure A.3: Layout of the bottom Heath tray with the second replicate of each population of
coho slamon in each of the three incubators at Button-up. F1, F2, and F3 refer to the different
families; placement for families in the Heath trays differed to limit any positional effect on
the variables measured. N (northern) and S (southern) represent the population of each
female x male crossed within each cell.

74

APPENDIX 3.2
Development Rate
Table A3.1: Summary statistics for accumulated thermal units required to reach hatch.
ATU
SS
df
MS
F
Sig
Model
137866.4
11
12533.31
613.01
0.00001
Treatment
Maternal
Paternal
Treat*Mat
Treat*Pat
Mat*Pat
Treat*Mat*Pat

136583.1
575.66
148.602
24.958
106.177
231.335
196.589

2
1
1
2
2
1
2

68291.57
575.666
148.602
12.479
53.088
231.335
98.294

Residual
Total

476.69
138343.166

24
35

19.862
3952.66

3438.25
28.98
7.48
0.63
2.67
11.65
4.95

0.00001
0.00001
0.0115
0.5421
0.0895
0.0023
0.0159

Table A3.2: Summary statistics for accumulated thermal units required to reach button-up.
ATU
SS
df
MS
F
Sig
Model
347080.35
11
31552.75
2578.78
0.00001
Treatment
Maternal
Paternal
Treat*Mat
Treat*Pat
Mat*Pat
Treat*Mat*Pat

343120.89
2490.91
105.21
12445.70
102.33
5.877
9.40

2
1
1
2
2
1
2

171560.45
2490.91
105.22
622.85
51.17
5.87
4.71

Residual
Total

293.65
347374.01

24
35

12.23
9924.97

14021.49
203.58
8.60
50.91
4.18
0.48
0.38

0.00001
0.00001
0.0073
0.00001
0.0277
0.4949
0.6851

75

Size and Condition Factor
Table A3.3: Summary statistics for the weight at hatch.
Weight
SS
df
MS
Model
0.1223
36
0.0111

F
26.43

Sig
0.00001

Treatment
Maternal
Paternal
Treat*Mat
Treat*Pat
Mat*Pat
Treat*Mat*Pat

0.05921
0.09354
0.000096
0.0023
9.25-06
0.000336
0.00099

2
1
1
2
2
1
2

0.00296
0.09354
0.000096
0.00119
4.69-06
0.000336
0.000495

7.04
222.35
0.23
2.83
0.01
0.80
1.18

0.0010
0.00001
0.6328
0.0604
0.9891
0.3721
0.3090

Residual
Total

0.14809
0.27043

352
363

0.00042
0.000744

Table A3.4: Summary statistics for the length at hatch.
Length
SS
df
MS
Model
587.0475
11
53.368

F
50.57

Sig
0.0001

Treatment
Maternal
Paternal
Treat*Mat
Treat*Pat
Mat*Pat
Treat*Mat*Pat

396.36
84.851
2.1444
22.731
1.570
0.9432
0.7027

2
1
1
2
2
1
2

198.1848
84.851
2.1444
11.3565
0.7851
0.9432
0.3513

187.78
80.40
2.03
10.76
0.74
0.89
0.33

0.0001
0.0001
0.1549
0.0001
0.4760
0.3451
0.7171

Residual
Total

372.5625
959.612

353
364

1.0554
2.6362

Table A3.5: Summary statistics for the condition factor at hatch.
Length
SS
df
MS
F
Model
2.285
11
0.207794
8.54

Sig
0.0001

Treatment
Maternal
Paternal
Treat*Mat
Treat*Pat
Mat*Pat
Treat*Mat*Pat

0.5238
0.9629
0.03712
0.4865
0.05744
0.0094
0.7717

2
1
1
2
2
1
2

0.2619
0.962921
0.00371
0.2432
0.0287
0.0094
0.3858

0.0001
0.0001
0.6962
0.0001
0.3084
0.5326
0.2061

Residual
Total

8.5848
10.870

353
364

0.02431
0.2986

10.77
39.59
0.15
10.00
1.18
0.39
1.59

76

Table A3.6: Summary statistics for the weight at Button-up.
Length
SS
df
MS
Model
0.43871
11
0.03988

F
50.55

Sig
0.0001

Treatment
Maternal
Paternal
Treat*Mat
Treat*Pat
Mat*Pat
Treat*Mat*Pat

0.04223
0.2676
0.0032
0.00438
0.0023
0.1299
0.00467

2
1
1
2
2
1
2

0.02111
0.2676
0.00328
0.00219
0.00119
0.01299
0.00233

26.76
339.19
4.316
2.78
1.52
16.47
2.96

0.0001
0.0001
0.0418
0.0629
0.2205
0.0001
0.0525

Residual
Total

0.5168
0.9555

655
666

0.00078
0.0014

Table A3.7: Summary statistic for the length at button-up
Length
SS
df
MS
Model
418.293
11
38.0266

F
41.59

Sig
0.00001

Treatment
Maternal
Paternal
Treat*Mat
Treat*Pat
Mat*Pat
Treat*Mat*Pat

159.25
212.15
1.1567
14.656
3.9477
7.2667
1.3016

2
1
1
2
2
1
2

79.628
212.15
1.156
7.328
1.973
7.266
0.6508

87.09
232.03
1.27
8.01
2.16
7.95
0.71

0.00001
0.00001
0.2611
0.0004
0.1163
0.0050
0.4912

Residual
Total

579.68
997.97

634
645

0.9143
1.5472

Table A3.8: Summary statistics for condition factor at Button-up.
Length
SS
df
MS
F
Model
2.4208
11
0.2200
29.93

Sig
0.0001

Treatment
Maternal
Paternal
Treat*Mat
Treat*Pat
Mat*Pat
Treat*Mat*Pat

0.8592
0.8254
0.0062
0.0999
0.2562
0.119
0.0182

2
1
1
2
2
1
2

0.4296
0.8254
0.0062
0.0499
0.1281
0.1193
0.0091

0.0001
0.0001
0.3562
0.0012
0.1759
0.0001
0.2898

Residual
Total

4.661
7.0825

634
645

0.0073
0.1098

58.43
112.27
0.85
6.80
1.74
16.23
1.24

77

Oxygen Consumption
Table A3.9: Summary statistics of the two-way ANOVA for the metabolic rate at Hatch.
Length
SS
df
MS
F
Sig
Model
148796.26
3
49598.75
55.09
0.0001
Treatment
Population
Treat*Pop

144355.78
14.05
1747.26

1
1
1

Residual
Total

175578.25
324374.52

195
198

144355.78
14.05
1747.26

160.32
0.02
1.94

0.0001
0.9007
0.1652

Table A3.10: Summary statistics of the three-way ANOVA for the metabolic rate at Buttonup.
Length
SS
df
MS
F
Sig
Model
134333
7
19190
5.93
0.00001
Treatment
Maternal
Paternal
Treat*Mat
Treat*Pat
Mat*Pat
Treat*Mat*Pat

4508.09
37437
81.56
13680
5204.23
9864.88
53429

1
1
1
1
1
1
1

4508
37437
81.56
13680
5204
9864
53429

Residual
Total

1032508
1166841

319
326

32326.7
3579.23

1.39
11.57
0.03
4.23
1.61
3.05
16.51

0.22388
0.0008
0.8740
0.0406
0.2057
0.0818
0.0001

78

Yolk Area
Table A3.11: Summary statistic for the yolk area at hatch.
Yolk
SS
df
MS
Model
735.017
11
68.455

F
36.01

Sig
0.00001

Treatment
Maternal
Paternal
Treat*Mat
Treat*Pat
Mat*Pat
Treat*Mat*Pat

74.966
624.97
3.186
7.135
4.376
1.393
6.047

2
1
1
2
2
1
2

37.483
624.97
3.1863
3.5679
2.1880
1.393
3.023

19.72
328.74
1.68
1.88
1.15
0.73
1.59

0.00001
0.00001
0.1964
0.1549
0.3177
0.3927
0.2055

Residual
Total

577.938914
1330.95

304
315

1.9011
4.2252

Table A3.12: Summary statistic for the standardized yolk area at hatch.
Yolk
SS
df
MS
F
Model
0.1730
11
0.0157
5.23

Sig
0.00001

Treatment
Maternal
Paternal
Treat*Mat
Treat*Pat
Mat*Pat
Treat*Mat*Pat

0.1149
0.0020
0.0052
0.0126
0.0058
0.0024
0.0084

2
1
1
2
2
1
2

0.0574
0.0020
0.0052
0.0063
0.0029
0.0024
0.0042

19.11
0.68
1.73
2.10
0.98
0.83
1.40

0.00001
0.4111
0.1893
0.1248
0.3772
0.3627
0.2489

Residual
Total

0.9142
1.0873

304
315

0.0030
0.0034

Survival
Table A3.13: Summary statistics for the total percent survival.
Survival
SS
df
MS
Model
11189
11
1017.21

F
3.74

Sig
0.0004

Treatment
Maternal
Paternal
Treat*Mat
Treat*Pat
Mat*Pat
Treat*Mat*Pat

310.50
278.40
7360.32
527.52
280.34
2147.98
284.32

2
1
1
2
2
1
2

155.25
278.40
7360.32
263.76
140.17
2147.98
142.16

0.57
1.02
27.08
0.97
0.52
7.90
0.52

0.5679
0.3156
0.00001
0.3848
0.5997
0.0067
0.5954

Residual
Total

16310
27499

60
71

271.84
387.32
79

EPILOGUE
This work expands the current knowledge of the importance of temperature on
development of larval coho salmon and the environmental variables they experience
throughout the incubation period. Coho salmon are of significant importance within BC, both
economically and culturally. Interior Fraser coho salmon are currently listed as endangered
by the Committee on the Status of Endangered Wildlife in Canada, but not currently listed by
Species at Risk Act. As incubation is a critical stage in development for Pacific salmon, I
would expect strong selective pressure at this life-history stage. The wide range of
temperatures experienced by larval coho salmon across BC during the incubation period,
some very cold and near freezing, would suggest that temperature is an important constraint
on performance. Population genetics also suggests that local adaptation is likely to maximize
performance for fish at the extreme limit of their range.
The importance of temperature on the early life history and development of coho
salmon was confirmed in my studies, but I also showed that other variables influence lifehistory variation – latitude of spawning grounds, size of spawning system, and migration
distance (Chapter 2). A number of previous studies have examined life-history variation
among populations that differ in latitude, but the populations have been limited to coastal
watersheds or those with short migration distances into interior river systems (Fleming and
Gross 1990; Beacham and Murray 1993; Braun et al. 2013), and additionally they did not
incorporate incubation temperatures and other environmental variables experienced. Thus, by
creating models with a number of variables that influence early life-history traits I have
helped focus management on the most significant variables.

80

With temperature being so important for early life history, governing the growth and
development in fish (Chapter 2; Fry 1971; Blaxter 1992), and having direct effect on
metabolic rate and therefore rate of development (Chapter 3; Fry 1971; Brett 1995),
determining the range in temperatures experienced by coho salmon is of great interest.
Chapter 1 demonstrated how temperatures experienced within the redd and river systems
vary significantly throughout British Columbia. Additionally Chapter 3 demonstrated that
small differences in temperature can have significant effects on early development.
Temperature regimes ranged from 1 to 5 °C on average throughout incubation and from 0.1
to 4 °C during the coldest period of incubation, confirming the capacity of coho salmon to
tolerate a wide range of temperatures. My findings demonstrate the importance of
understanding the conditions present within the redds themselves throughout incubation at a
population level and not just surface water conditions. The effect of temperature is
cumulative with even very small differences in temperature having profound effect on the
ATU required, resulting in considerable differences in the development of salmonids. Thus,
hatching and emergence dates could vary significantly with only a small difference in
temperature, especially for systems at near-freezing temperature where developing embryos
are likely to be highly sensitive to the temperature differences.
To better understand the importance of temperature on early development of coho
salmon and the adaptations that enhance their survival at near-freezing incubation
temperatures, a controlled incubation experiment measuring early development differences
between pure and hybrid families of two populations of coastal coho salmon was conducted.
The results demonstrated that rate of development, growth, yolk utilization and metabolic
rates were all significantly affected by temperature, specifically when temperatures approach

81

freezing. Near-freezing incubation temperatures resulted in less accumulated thermal units to
reach each development stage, lower condition factors, lower resting metabolic rates and
more yolk available for somatic growth than the warmer incubation temperatures.
Additionally, the incubation experiment also demonstrated that temperature has more of an
influence on early development of coho salmon compared to maternal and paternal genetic
effects. Maternal and paternal origin did, however, have a significant effect on the variables
measured, but compared to temperature these genetic influences were subtle. Maternal
effects were related to the differences in egg size between the northern or southern
populations. Paternal effects were more subtle and only seen for some families at button-up.
The progeny from the northern populations also did not show significant improvement in
performance in the cold treatments regime as predicted. Thus, the results from this
experiment suggest little local adaptation on development at cold temperatures – rather
exceptional phenotypic plasticity within this species is occurring.
To properly conserve and manage coho salmon, a better understanding of the
temperature regimes experienced by different populations is needed, particularly at a finer
scale within streams and at the extreme ranges in distribution. Future research could expand
my work and continue to measure metabolic rates and development throughout later life
stages such as parr and smolting to gain a better understanding of how populations adapt to
different temperature regimes and how it affects overall fitness of these later life stages.
Incubating at near-freezing temperatures, or cold-adapted populations incubating within
warm temperature regimes, may not have a significant effect on early development but may
reduce the overall fitness and success throughout smolting and maturation. Additionally,
conducting more frequent sampling of yolk and metabolic rate at a larger range of

82

temperatures will allow a better understanding of how energy is being allocated and yolk is
being used by larval fish from these populations within these near-freezing temperatures, and
the effects on overall fitness. Previous studies have examined these processes (Heming 1982;
Beacham and Murray 1989; Rombough 1994; Kamler 2008), but not at the near-freezing
temperatures experienced by coho salmon at the limits of their range. My expectation was
that the northern population, which is routinely exposed to cold temperatures during larval
development, should have performed better than the southern population. Although there was
a strong maternal effect, the lack of a paternal effect would argue that adaptation to rearing at
cold temperatures was not strongly developed in the northern population. In contrast, my
results show tremendous phenotypic plasticity exists within a single species that enable it to
thrive under a wide range of temperatures during larval development.

83

REFERENCES
Baxter JS, McPhail JD. 1999. The influence of reed site selection, groundwater upwelling,
and over-winter incubation temperature on survival of bull trout (Salvelinus
confluentus) from egg to alevin. Can J Zool 77: 1233-1239.
Beacham TD. 1982. Fecundity of coho salmon (Oncorhynchus kisutch) and chum salmon (O.
keta) in the northeast Pacific Ocean. Can J Zool 60: 1463-1469.
Beacham TD, Morley WRB. 1985. Effect of egg size on incubation time and alvin and fry
size in chum salmon (Oncorhynchus keta) and coho salmon (Oncorhynchus kisutch).
Can J Zool 63: 847-850.
Beacham TD, Murray CB. 1985. Effect of female size, egg size and water temperature on
developmental biology of chum salmon (Oncorhynchus keta) from the Nitinat River,
British Columbia. Can J Aquat Sci 42: 1755-1765.
Beacham TD, Withler FC, Morley WRB. 1985. Effect of egg size on incubation time and
alevin and fry size in chum salmon (Oncorhynchus keta) and coho salmon
(Oncorhynchus kisutch). Can J Zool 63: 847-850.
Beacham TD, Murray CB. 1987. Adaptive variation in body size, age, morphology, egg size
and developmental biology of chum salmon (Oncorhynchus keta). J Fish Biol 28: 233246.
Beacham TD. 1988. A genetic analysis of early development in pink (Oncorhynchus
gorbuscha) and chum salmon (Oncorhynchus keta) at three different temperatures.
Genome 30: 89- 96
Beacham TD, Murray CB. 1989. Variation in developmental biology of sockeye salmon
(Oncorhynchus nerka) and chinook salmon (Oncorhynchus tshawytscha) in British
Columbia. Can J Zool 67: 2081-2089.
Beacham TD, Murray CB. 1990. Temperature, egg size and development of embryos and
alevins of five species of Pacific salmon: a comparative analysis. Trans Am Fish Soc
119: 927-945.
Beacham TD, Murray CB. 1993. Fecundity and egg size variation in North American Pacific
salmon. J Fish Biol 42: 485-508.
Blaxter JHS. 1992. The effect of temperature on larval fishes. Netherlands J Zool 42: 336357.
Bonham K. 1976. Specific gravity of salmonid eggs. Trans Am Fish Soc 105: 351-352.

84

Boucher MA. 2012. The effects of substrate rearing on the growth, development, and
survival of larval white sturgeon (Acipenser transmontanus) during early ontogeny.
M.Sc. thesis. University of Northern British Columbia, Prince George, BC.
Brander KM, Blom G, Borges MF, Erzini K, Henderson G, MacKenzie BR, Mendes H,
Ribeiro J, Santos AMP, Toresen R. 2003. Changes in fish distribution in the eastern
North Atlantic: Are we seeing a coherent response to changing temperature? ICES
Marine Science Symposia, 219: 261-270.
Braun DC, Patterson DA, Reynolds JD. 2013. Maternal and environmental influences on egg
size and juvenile life-history traits in Pacific salmon. Ecol Evol 3: 1727-1740.
Brett JR. 1971. Energetic responses of salmon to temperature: a study of some thermal
relations in the physiology and freshwater ecology of sockeye salmon (Oncorhynchus
nerka). Am Zool 11: 99-113.
Brett JR, Groves TDD. 1979. Physiological Energetics. In Fish Physiology. Vol. 8.
Bioenergetics and Growth. (Hoar, W.S. and Randall, D.J., eds), pp. 280–344. New
York: Academic Press.
Brett JR. 1995. Energetics in Physiological Ecology of Pacific Salmon. (Eds.) Groot, C.,
Margolis, L., and Clarke, W.C. UBC Press, Vancouver. pp. 3-68
Burger CV, Wilmot RL, Wangaard DB. 1985. Comparison of spawning areas and times for
two runs of chinook salmon (Oncorhynchus tshawyscha) in the Kenai River, Alaska.
Can J Fish Aquat Sci 42: 693-700.
Burnham KP, Anderson DR. 1998. Model selection and inference: a practical information
theoretic approach. Springer, New York.
Burnham KP, Anderson DR. 2004. Multimodel inference: Understanding AIC and BIC in
model selection. Sociol Methods Res 33: 261-304.
Callow P. 1985. Adaptive aspects of energy allocation. In Fish Energetics, (Tytler, P. and
Callow, P., eds), pp. 13–32. Baltimore: Johns Hopkins University Press.
Chavarie L, Demprson JB, Schwarz, Reist JD, Power G, Power M. 2010. Latitudianl
variation in growth among Arctic charr in eastern North America: evidence for
countergradient variation? Hydrobiologia 650:161-177.
Clark A, Fraser KPP. 2004. Why does metabolism scale with temperature? Functional Ecol
16: 243-251.
Clarke A, Johnston NM. 1999. Scaling of metabolic rate with body mass and temperature in
teleost fish. J Anim Ecol 68: 893-905.

85

Cunjak RA. 1996. Winter habitat of selected stream fishes and potential impacts from landuse activity. Can J Fish Aquat Sci 53: 267-282.
Cunjak RA, Power G. 1986. Winter habitat utilization by resident brook trout (Salvelinus
fontinalis) and brown trout (Salmo trutta), Can J Fish Aquat Sci 43:1970-1981.
Daufersene M, Lengfellner K, Sommer U. 2009. Global warming benefits the small in
aquatic ecosystms. Proc Natl Acad Sci USA 106: 12788-12793.
Davis JC. 1975. Waterborne dissolved oxygen requirements and criteria with particular
emphasis on the Canadian environment. NRCC No. 14100. National Research Council
of Canada, Associate Committee on Scientific Criteria for Environmental Quality,
Ottawa.
DFO. 2002. Interior Fraser River Coho Salmon. DFO 2005 Science Stock Report D6-08
(2002).
DFO. 2005. Recovery Assessment Report for Interior Fraser Coho Salmon (Oncorhynchus
kisutch). DFO Can. Sci. Advis. Sec. Sci. Advis. Rep. 2005/061.
Fangue NA, Hafmeister M, Schulte PM. 2006. Interspecific variation in thermal tolerance
and heat shock protein gene expression in common killifish, Fundulus heteroclitus. J
Exp Biol 209: 2859-1872.
Farrell AP, Hinch SG, Cooke SJ, Patterson DA, Crossin GT, Lapointe M, Mathes MT. 2008.
Pacific Salmon in hot water: applying aerobic scope models and biotelemetry to predict
the success of spawning migrations. Phys Bioc Zool 81:697-708.
Fleming IA, Gross MR. 1989. Evolution of adult female life history and morphology in
Pacific salmon (Coho: Oncorhynchus kisutch). Evolution 43: 141-157.
Fleming IA, Gross MR. 1990. Latitudinal Clines: A trade-off between egg number and size
in Pacific salmon. Ecology 71: 1-11.
Frankham R. 1995. Inbreeding and extinction: a threshold effect. Cons Biol 9: 792–799.
Freeze RA, Cherry JA. 1979. Groundwater. Prentice-Hall, Englewood Cliffs, NJ.
Fry FEJ. 1971. The effect of environmental factors on the physiology of fish in Fish
Physiology. Vol. 6. Environmental relations and behavior. (Eds.) Hoar, W.S. and
Randall, D.J. Academic Press, New York. pp. 1-98
Garrett JW, Bennett DH, Frost FO, Thurow RF. 1998. Enhanced incubation success for
kokanee spawning in groundwater upwelling sites in small Idaho Stream. N Am J Fish
Manage 18: 925-930.

86

Gotelli NJ, Ellison AM. 2004. A primer of ecological statistics. Sinauer Associates Inc.
Massachesetts. pp.180-189
Grebmeier JM, Overland JE, Moore SE, Farley EV, Carmack EC, Cooper LW, Frey KE,
Helle JH, McLaughlin FA, McNutt SL. 2006. A major ecosystem shift in the Northern
Bering Sea. Science. 11:1461-1464.
Groot C, Margolis L. 1991. Pacific Salmon Life Histories. University of British Columbia
Press, Vancouver.
Haring MH, Johnston TA, Wiegand MD, Fisk At, Pitcher TE. 2016. Differences in egg
quantity and quality among hatchery – and wild-origin chinook salmon (Oncorhynchus
tshawytscha). Can J Fish Aquat Sci 73: 737-746.
Heath DD, Heath JW, Bryden CA, Johnson RM, Fox CW. 2003. Rapid evolution of egg size
in captive salmon. Science 299: 1738-1740.
Heath DD, Fox CW, Heath JW. 1999. Maternal effect on offspring size: variation through
early development of chinook salmon. Evolution 53: 1605-1611.
Heming TA. 1982. Effect of temperature on utilization of yolk by chinook salmon
(Oncorhynchus tshawytscha) eggs and alevins. Can J Fish Aquat Sci 39: 184-190.
Heming TA, Buddington RK. 1988. Yolk absorption in embryonic and larval fishes. In: Fish
Physiology (eds WS Hoar and DJ Randall). XIA: Egg and larvae, pp 408-446.
Academic Press, San Diego
Johnston IA, McLay HA. 1997. Temperature and family effects on muscle cellularity at
hatch and first feeding in Atlantic salmon (Salmo salar L.). Can J Zool 75: 64-74.
Johnson SL, Jones JA. 2000. Stream temperature responses to forest harvest and debris flows
in western Cascades, Oregon. Can J Fish Aquat Sci 57: 30-39.
Johnson, SL. 2003. Stream temperature: scaling of observations and issues for modelling.
Hydrol Processes 17: 497-499.
Johnson, SL. 2004. Factors influencing stream temperature in small streams: substrate effects
and shading experiment. Can J Fish Aquat Sci 61: 913-923.
Johnston IA, McLay HA. 1997. Temperature and family effects on muscle cellularity at
hatch and first feeding in Atlantic salmon. Can J Zool 75: 64-74.
Johnston IA, McLay HA, Abercromby M, Robins D. 2000. Phenotypic plasticity of early
myogenesis and satellite cell numbers in Atlantic salmon spawning in upland and
lowland tributaries of a river system. J Exp Biol. 203: 2539-2552.

87

Jonsson B, Jonsson N. 2011. Development and growth. Fish Fisheries Ser 33: 1-21.
Kalmer E. 2008. Resource allocation in yolk-feeding fish. Rev Fish Biol Fisheries 18: 143200.
Kileen J, McLay HA, Johnston IA. 1999. Development in Salmo trutta at different
Temperatures, with a Quantitative scoring methods for intraspecific comparison. J Fish
Biol 55: 382-494
Knighton D. 1998. Fluvial Forms and Processes: A New Perspective. Hodder Arnold
Publication. pages 80-95; 107-136.
Leggett WC, Carscadden JE. 1978. Latitudinal variation in reproductive characteristics of
American shad (Alosa sapidissima): evidence for population specific life history
strategies in fish. J Fish Res Board Can 35: 1469-1478.
Lappalainen J, Tarkan AS, Harrod C. 2008. A meta-analysis of latitudinal variations in lifehistory traits of roach, Rutilus rutilus, over its geographical range: linear or non-linear
relationships? Freshwater Biol 53: 1491-1501.
Marty GD, Short JW, Dambach DM, Willits NH, Heintz RA, Rice SD, Stegeman JJ, Hinton
DE. 1997. Ascites, premature emergence, increased gonadal cell apoptosis, and
cytochrome P4501A induction in pink salmon larvae continuously exposed to oilcontaminated gravel during development. Can J Zool 75: 989-1007.
McCullough DA. 1999. A review and synthesis of effects of alterations to the water
temperature regime on freshwater life stages of salmonids, with special reference to
chinook salmon. Region 10 Water Resources Assessment Report No. 910-R-99-010,
United States EPA, Seattle.
McRae CJ. 2009. Spawning Site Selection and the Influence of Incubation Environment on
Larval Success in Interior Fraser Coho. MSc Thesis, University of Northern British
Columbia, Prince George BC.
McRae CJ, Warren KD, Shrimpton JM. 2012. Spawning site selection and the influence of
incubation environment on larval success in Interior Fraser Coho. Endang Species Res
16: 249-260.
Merilä J, Hendry AP. 2014. Climate change, adaptation, and phenotypic plasticity: the
problem and the evidence. Evol App. 1: 1-14.
Murray CB, Beacham TD. 1988. Influence of parental stock and incubation temperature on
the early development of coho salmon (Oncorhynchus kisutch) in British Columbia.
Can J Zool 68: 347-358.

88

Murray CB, Beacham TD, McPhail JD. 1990. Influence of paternal stock and incubation
temperature on the early development of coho salmon (Oncorhynchus kisutch) in
British Columbia. Can J Zool 68: 347-358.
Murray CB, McPhail JD. 1988. Effect of incubation temperature on the development of five
species of Pacific salmon (Oncorhynchus) embryos and alevins. Can J Zool 66: 266273.
Nadeau PS, Hinch SG, Pon LB, Patterson DA. 2009. Presistent paternal effect on survival
and size, but not burst swimming performance of juvenile sockeye salmon
Oncorhynchus nerka. J Fish Biol 75: 538-551.
Newman D, Pilson D. 1997. Increased probability of extinction due to decreased genetic
effective population size: experimental populations of Clarkia pulchella. Evolution 51:
354–362.
Paaijmans KP, Heining RL, Seliga RA, Blandford JI, Blandford S, Murdock CC, Thomas
MB. 2013. Temperature variation makes ectotherms more sensitive to climate change.
Glob Chang Biol 19: 2373-2380.
Perry SF, Tufts BL. 1998. Fish Respiration. Fish Physiology. Volume 17. Academic Press.
Perry AL, Low PJ, Ellis JR, Reynolds JD. 2005. Climate change and distribution shifts in
marine fishes. Science 308: 1912-1915.
Peterson RH, Spinney HCE, Sreedharan A. 1977. Development of Atlantic salmon (Salmo
salar) eggs and alevins under varied temperature regimes. J Fish Res Board Can 34:
31-43.
Petersson E, Järvi T. 2007. Characteristics of brown trout males influence growth and
survival of their offspring. J Fish Biol 71: 493-509.
Pörtner HO, Peck MA. 2010. Climate change effects on fishes and fisheries: towards a
cause-and-effect understanding. J Fish Biol 77: 1745-1779.
Quinn TP. 2005. The Behavior and Ecology of Pacific Salmon and Trout. Seattle: University
of Washington Press.
Quinn TP, Hendry AP, Wetzel LA. 1995. The influence of life history trade-offs and the size
of incubation gravels on egg size variation in sockeye salmon (Oncorhynchus nerka).
Oikos. 74: 425-438.
Quinn TP, Vøllestad LA, Peterson J. Gallucci V. 2004. Influence of freshwater and marine
growth on the egg size-egg number trade-off in coho and chinook salmon. Trans Am
Fish Soc 133: 55-65.

89

Rasenfeld J, Oirter M, Parkinson E. 2000. Habitat factors affecting the abundance and
distribution of juvenile cutthroat trout (Oncorhynchus clarki) and coho salmon
(Oncorhynchus kisutch). Can. J Fish Aquat Sci 57: 766-774.
Reed TE, Schindler DE, Hague MJ, Patterson DA, Meir E, Waples RS, Hinch SG. 2011.
Time to evolve? Potential evolutionary responses of Fraser River sockeye salmon to
climate change and effects on persistence. PLoS One 6: e20380.
Doi:10.1371/journal.pone.0020380.
Rollinson N and Hutchings JA. 2011. Why does egg size of salmonds increase with the mean
size of population spawning gravel? Can J Fish Aquat Sci 68:1307-1315.
Rombough PJ. 1994. Energy partitioning during fish development: assistive or compensatory
allocation of energy to support growth? Functional Ecol 8: 178-186.
Rombough PJ. 2006. Developmental costs and partitioning of metabolic energy. In:
Comparative Developmental Physiology Contributions, Tools and Trends. (Eds.)
Warburton, S.J., Burggren, W.W., Pelster, B., Reiber, C.L. and Spicer, J. pp. 99-123.
Oxford University Press, Oxford.
Rombough P. 2011. The energetics of embryonic growth. Respir Physiol Neurobiol 178: 2229.
Rosenfeld J, Porter M, Parkinson E. 2000. Habitat factors affecting the abundance and
distribution of juvenile cutthroat trout (Oncorhynchus clarki) and coho salmon
(Oncorhynchus kisutch). Can J Fish Aquat Sci 57: 766-774.
Rosengrave P, Montgomerie R, Gemmell N. 2016. Cryptic female choice enhances
fertilization success and embryo survival in Chinook Salmon. Proceedings of the Royal
Society B: Biological Sciences 283: 20160001
Rosenzweig C, Karoly D, Vicarelli M, Neofotis P, Wu Q, Casassa G, et al. 2008. Attributing
physical and biological impacts to anthropogenic climate change. Nature 453:
doi:10.1038/nature06937
Sandercock FK. 1991. Life history of coho salmon (Oncorhynchus kisutch). Pp. 397-445 In
C. Groot and L. Margolis (Eds.), Pacific salmon life histories. Vancouver: University
of British Columbia Press.
Scholz NL, Myers MS, McCarthy SG, Labenia JS, McIntyre JK, Ylitalo GM, Rhodes LD,
Laetz CA, Stehr CM, French BL, McMillan B, Wilson D, Reed L, Lynch KD, Damm
S, Davis JW, Collier TK. 2011. Recurrent die-offs of adult coho salmon returning to
spawn in Puget Sound Iowland urban streams. PLoS One. 6:e28013.
Shaul L, Weitkamp L, Simpson K, Sawada J. 2007. Trends in abundance and size of coho
salmon in the Pacific Rim. North Pacific Anadromous Fish Commission 4: 93-104.
90

Shepherd BG, Hartman GF, Wilson WJ. 1986. Relationships between stream and intergravel
temperatures in coastal drainages, and some implications for fisheries workers. Can J
Fish Aquat Sci 43: 1818-1822.
Shrimpton JM, Heath DD. 2003. Census vs. effective population size in chinook salmon:
Large- and small-scale environmental perturbation effect. Mol Ecol 12: 2571-2583.
Smith K, Lavis ME. 1975. Environmental influences on the temperature of a small upland
stream. Oikos 26: 228--236.
Shultz ET, Conover DO. 1997. Latitudinal differences in somatic energy storage: Adaptive
responses to seasonality in an estuarine fish (Atherinidae: Menidia menidia). Oecologia
109: 516-529.
Stearns SC. 1983. A natural experiment in life-history evolution: field data on the
introduction of mosquitofish (Gambusia affinis) to Hawaii. Evolution 37: 601--617.
Stickland NC, White RN, Mescall PE, Crook AR, Thorpe JE. 1998. The effect of
temperature on myogenesis in the embryonic development of Atlantic salmon (Salmo
salar L.). Analytical Embryol 178: 253-257.
Sullivan K, Adams TA. 1990. The physics of stream heating: an analysis of temperature
patterns in stream environments based on physical principles and field data.
Weyerheauser Technical Report. 044–5002·90/2. Technology Center, Tacoma, WA
98477.
Svardson 1949G. 1949. Natural selection and egg number in fish. Institute of Freshwater
Research, Drottningholm Report 29:115-122.
Taylor EB, McPhail JD. 1985. Variation in Body Morphology among British Columbia
populations of coho salmon, Oncorhynchus kisutch. Can J Fish Aquat Sci 42: Can J
Fish Aquat Sci 42:2020-2028.
Thomas CD, Cameron A, Green RE, Bakkenes M, Beaumont LJ, Collingham YC, Erasum
BFN, Siqueira MF, Grainger A, Hannah L, Hughes L, Huntley B, van Jaarsveld AS,
Midgley GF, Miles L, Ortega-Huerta MA, Peterson AT, Phillips OL, Williams SE.
2004. Extinction risk from climate change. Nature 427: 145-148.
Van den Berghe EP, Gross MR. 1989. Natural selection resulting from female breeding
competition in a Pacific Salmon (Coho: Oncorhynchus kisutch). Evolution 48: 125140.
Visser ME. 2008. Keeping up with a warming world; assessing the rate of adaptation to
climate change. Proc Roy Soc B 275: 649-659.

91

Waples RS. 1990a. Conservation genetics of Pacific salmon. III. Estimating effective
population size. J Heredity 81: 277-289.
Waples RS. 1990b. Conservation genetics of Pacific salmon. II. Effective population size and
the rate of loss of genetic variability. J Heredity 81: 267-276.
Wells PR, Pinder AW. 1996. The respiratory development of Atlantic salmon I.
Morphometry of gills, yolk sac and body surface. J Exp Biol 199: 2125-2736.
Whitney CK, Hinch SG, Patterson DA. 2014. Population origin and watertemperauter affect
development timing in embryonic sockeye salmon. Trans Am Fish Soc 143: 13161329.
Whittingham MJ, Stephens PA, Bradbury RB, Freckleton. 2006. Why do we still use
stepwise modeling in ecology and behavior? J Animal Ecol 75: 1182-1189.
Williamson CJ. 2006. Characteration of spawning habitat, incubation environment and early
growth and development in Bull Trout (Salvelinus confluentus) from pristine streams
of Northern British Columbia. MSc Thesis, University of Northern British Columbia,
Prince George BC.
Zar JH, 1996. Biostatistical Analysis (3rd ed.) Upper Saddle River, NJ: Prentice-Hall, P. 128
Zuo W, Mouses ME, West GB, Hou C, Brown JH. 2012. A general model for effects of
temperature on ectotherms ontogenetic growth and development. Proc Biol Sci 279:
1840-1846.

92