THE EFFECT OF FOOD RATION ON GROWTH, REPRODUCTION AND MUSCLE LIPID
STORAGE IN MATURE FEMALE LAKE TROUT FROM DIFFERENT POPULATIONS
RAISED IN A COMMON ENVIRONMENT
By
Benjamin Felt

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
Fisheries and Wildlife – Master of Science
2013

ABSTRACT
THE EFFECT OF FOOD RATION ON GROWTH, REPRODUCTION AND MUSCLE LIPID
STORAGE IN MATURE FEMALE LAKE TROUT FROM DIFFERENT POPULATIONS
RAISED IN A COMMON ENVIRONMENT
By
Benjamin Felt
This study examined the differences in growth rates, percent lipid, and reproductive
output between hatchery and wild individuals from lake trout (Salvelinus namaycush) stocks that
show very different life history strategies in efforts to characterize energy allocation along life
history considerations, and responses to a food limitation stressor. This study monitored
individuals raised in a common environment for two years for growth rates, percent muscle lipid
storage and reproductive output under different feeding regimes from the Killala, Manitou,
Michipicoten, and Kingscote stocks and their wild counterparts. For the course of this study,
some fish were maintained on a high ration, some fish on a low ration, and some fish were
switched from a high ration to a low ration, and the remaining fish were switched from a low
ration to a high ration. Robust reaction norms between egg number and percent lipid were
identified within all stocks across rations within both years, but this relationship changed shape
between years. These results suggest that there are differences between lake trout stocks in the
range of phenotypic expressions, based on genetics, and that the relationship between lipids and
egg number may be predictable within each stock at a given age. This increased understanding
of stock-specific phenotypic expressions in limiting environments and energy allocation patterns,
in light of specific life history adaptations such as population mortality, may eventually be used
by managers to predict egg production, growth and storage of stocks in natural environments.

ACKNOWLEDGEMENTS

I begin by thanking my advisor, Dr. Cheryl Murphy, for her advice, guidance, and
patience throughout the duration of this project. I would also like to thank my committee
members Dr. Charles Krueger, Dr. James Bence, and Dr. Kim Scribner for their guidance and
advice on an earlier draft and presentation during a committee meeting. I thank the Ontario
Ministry of Natural Resources, specifically Bill Sloan, Scott Ferguson, Tara MacDonald, and Dr.
Chris Wilson for their help in the hatchery study in Codrington, Ontario. I would also like to
thank Brian Meredith, Paul Vieira, Lee Haslam, and Ryan Mills from the Ontario Ministry of
Natural Resources for their help with the field collections on Lake Superior. I would like to
thank my labmates Laura Claus, Ben Schmitt, and Dane Larsen for their assistance in processing
samples from the field. I would like to thank my wife, Christina, for all of her help and support
throughout these five years and also thank my parents and my brother Jeremy for giving me the
necessary support system to see this project through. Lastly, I would like to thank my Lord and
Savior Jesus Christ for being my source of strength and courage.

iii

TABLE OF CONTENTS

LIST OF TABLES ....................................................................................................................... vi
LIST OF FIGURES ...................................................................................................................... x
Chapter 1: Background Information and Conceptual Model .................................................. 1
INTRODUCTION ...................................................................................................................... 1
Study Rationale ....................................................................................................................... 1
Life History Theory and Optimal Energy Allocation .............................................................. 2
Seasonal Energy Allocation ................................................................................................ 5
Age-Specific Energy Allocation ......................................................................................... 7
Study Approach - Common Environment Experiment .......................................................... 11
WORKS CITED ....................................................................................................................... 14
Chapter 2: Characterization of Hatchery Fish in Context of Wild Counterparts .............. 21
INTRODUCTION .................................................................................................................... 21
Background of Hatchery Stocks ............................................................................................ 23
METHODS ............................................................................................................................... 26
Wild Collections .................................................................................................................... 26
Hatchery ................................................................................................................................ 26
Data Analyses ....................................................................................................................... 31
RESULTS ................................................................................................................................. 33
Characterizing the Wild Populations.................................................................................... 33
Fork Length ....................................................................................................................... 33
Percent Lipids ................................................................................................................... 37
Egg Size ............................................................................................................................ 41
Egg Number ...................................................................................................................... 45
Comparison of Field and Hatchery Data within Stocks ....................................................... 49
Comparison of Field and Hatchery Data between Stocks .................................................... 57
DISCUSSION ........................................................................................................................... 65
Stock Differences in Hatchery and Wild Fish ....................................................................... 65
Insight into Stock-specific Life Histories from Field and Hatchery Data ............................ 66
Stock-Specific Conceptual Models ........................................................................................ 68
Study Limitations .................................................................................................................. 70
Implications........................................................................................................................... 71
APPENDIX ............................................................................................................................... 73
WORKS CITED ....................................................................................................................... 86

iv

Chapter 3 – Life History Influence on the Energy Allocation towards Growth,
Reproduction and Storage in Common Environment Raised Hatchery Lake Trout Stocks
subjected to Food Limitation ..................................................................................................... 91
INTRODUCTION .................................................................................................................... 91
Seasonal Energy Allocation .................................................................................................. 92
Effects of Ration on Seasonal Growth and Percent Lipid .................................................... 93
Stock-specific Percent Lipid, Growth, and Reproductive Output ......................................... 95
Reaction Norms ..................................................................................................................... 96
METHODS ............................................................................................................................... 99
Hatchery Fish used in this Study .......................................................................................... 99
Treatments............................................................................................................................. 99
Measurements ..................................................................................................................... 103
Data Analyses ..................................................................................................................... 104
RESULTS ............................................................................................................................... 108
Effects of Age on Reproductive Output ............................................................................... 126
Reaction Norms ................................................................................................................... 131
DISCUSSION ......................................................................................................................... 144
Growth, Percent Lipid and Reproductive Output: General Ration-specific Responses ..... 145
Effects of Ration on Growth, Percent Lipid, and Reproductive Output: Stock Differences 146
Stock-Specific Conceptual Model Evaluation ..................................................................... 147
Seasonal Trends in Energy Allocation ................................................................................ 147
Annual Changes in Reproduction ....................................................................................... 149
Reaction Norms – Egg Number vs. Percent Lipids............................................................. 151
Study Limitations – Hatchery Design ................................................................................. 152
Implications......................................................................................................................... 154
Conclusions ......................................................................................................................... 155
APPENDIX ............................................................................................................................. 157
WORKS CITED ..................................................................................................................... 170

v

LIST OF TABLES

Table 1. Conceptual model of energy allocation in lake trout to reproduction, growth, and
storage and the hypothesized effects of stock-specific lifespans, local environmental conditions
(fishing pressure, predation, etc.), and diet on the amount of energy allocated to each energetic
category. .......................................................................................................................................... 5
Table 2. Conceptual model of energy allocation in lake trout to reproduction, growth, and
storage and the hypothesized effects of time of year on the amount of energy allocated to each
energetic category. .......................................................................................................................... 6
Table 3. Conceptual model of energy allocation in lake trout to reproduction, growth, and
storage and the hypothesized effects of age on the amount of energy allocated to each energetic
category. .......................................................................................................................................... 9
Table 4. Summary of initial fish densities and composition of lake trout stocks housed in a
common environment experiment in OMNR hatchery in Codrington, Ontario. .......................... 29
Table 5. Conceptual model of energy allocation in lake trout stocks to reproduction, growth, and
storage based on field data gathered and conceptual models summarized in Tables 1-3 (Chapter
1). .................................................................................................................................................. 70
Table 6. AIC values and Residual Sum of Squares (RSS) of linear, quadratic, logarithmic,
exponential, and hyperbolic relationships assigned to data age vs.: egg number, egg diameter,
length, and percent lipid in wild Manitou stock lake trout. .......................................................... 78
Table 7. AIC values and Residual Sum of Squares (RSS) of linear, quadratic, logarithmic,
exponential, and hyperbolic relationships assigned to data age vs.: egg number, egg diameter,
length, and percent lipid in wild Killala stock lake trout. ............................................................. 79
Table 8. AIC values and Residual Sum of Squares (RSS) of linear, quadratic, logarithmic,
exponential, and hyperbolic relationships assigned to data age vs.: egg number, egg diameter,
length, and percent lipid in wild Michipicoten stock lake trout. .................................................. 80
Table 9. AIC values and Residual Sum of Squares (RSS) of linear, quadratic, logarithmic,
exponential, and hyperbolic relationships assigned to data age vs.: egg number, egg diameter,
length, and percent lipid in wild Kingscote stock lake trout......................................................... 81
Table 10. Comparison of statistical significance of the fork length between stocks based on
ANOVA pairwise-comparison using a Tukey-Kramer adjustment. Field data are not included in
statistical comparison because value is a modeled value.............................................................. 82
Table 11. Comparison of statistical significance of the mean percent lipids between stocks based
on ANOVA pairwise-comparison using a Tukey-Kramer adjustment. Field data are not included
in statistical comparison because value is a modeled value. ........................................................ 83
vi

Table 12. Comparison of statistical significance of the egg size between stocks based on
ANOVA pairwise-comparison using a Tukey-Kramer adjustment. Field data are not included in
statistical comparison because value is a modeled value.............................................................. 84
Table 13. Comparison of statistical significance of the egg number between stocks based on
ANOVA pairwise-comparison using a Tukey-Kramer adjustment. Field data are not included in
statistical comparison because value is a modeled value.............................................................. 85
Table 14. Summary of hypotheses regarding responses of five lake trout stocks to common
environment experiment. .............................................................................................................. 98
Table 15. Summary of initial fish densities and composition of lake trout stocks housed in a
common environment experiment in OMNR hatchery in Codrington, Ontario. ........................ 102
Table 16. P-values (* = p<0.05, ** = p<0.01; + or – indicates direction of significant changes)
for repeated measures analysis for changes in egg size that occurred between the spawning
seasons of 2008 and 2009 of individuals from the Killala, 2000 Kingscote, 1999 Kingscote,
Manitou, and Michipicoten stocks of lake trout held in a common environment under multiple
feeding regimes (High = 0.5% of total tank biomass fed to tank daily; High-Low indicates a
switch from a high [0.5%] to a low [0.3%] ration level at the start of the study). ...................... 130
Table 17. P-values (* = p<0.05, ** = p<0.01; + or – indicates direction of significant changes)
for repeated measures analysis for changes in egg number that occurred between the spawning
seasons of 2008 and 2009 of individuals from the Killala, 2000 Kingscote, 1999 Kingscote,
Manitou, and Michipicoten stocks of lake trout held in a common environment under multiple
feeding regimes (High = 0.5% of total tank biomass fed to tank daily; High-Low indicates a
switch from a high [0.5%] to a low [0.3%] ration level at the start of the study). ...................... 130
Table 18. AIC differences between model of best fit (assigned based on AIC values) and an
alternate model describing the relationship between egg number and mean percent lipid of five
lake trout stocks housed in OMNR hatchery. Alternate model is model of best fit applied to
alternate data group (year or feeding group) to determine how much better a fit the original
model is than the model used to describe the relationship in the alternate ration group or study
year. Bolded values indicate AIC differences that exceed the threshold of 6, which indicates a
difference in the predictive ability between the original and alternate model for that given group.
..................................................................................................................................................... 143
Table 19. Summary of results found and how they support (or fail to support) initial hypotheses
regarding energy allocation of lake trout held in a common environment exposed to one of four
feeding regimes. .......................................................................................................................... 145
Table 20. Summary of significance of changes in percent lipid between July 11, 2008 and levels
during the spawn of five stocks of lake trout housed in a common garden environment in OMNR
hatchery. Lipid changes are recorded as not significant (NS), or positively significant based on
repeated measures analysis with α = 0.05. Ration levels were manipulated within each treatment
as follows High-High = maintained at a high [0.5% of body weight daily] ration throughout
study; High-Low = switched from a High [0.5% of body weight daily] to a low [0.3% of body
weight daily] ration, etc. Combined data are data combined from all treatments. .................... 158
vii

Table 21. Summary of significance of changes in percent lipid between July 27, 2009 and levels
during the spawn of five stocks of lake trout housed in a common garden environment in OMNR
hatchery. Lipid changes are recorded as not significant (NS), or negatively significant based on
repeated measures analysis with α = 0.05. Ration levels were manipulated within each treatment
as follows High-High = maintained at a high [0.5% of body weight daily] ration throughout
study; High-Low = switched from a High [0.5% of body weight daily] to a low [0.3% of body
weight daily] ration, etc. Combined data are data combined from all treatments. .................... 159
Table 22. Year 1 comparison of growth rates (daily growth rate between sampling dates,
mm/day) and December-spawn growth (change in fork length) in five stocks of lake trout housed
in a common environment hatchery, exposed to one of four feeding regimes. % difference
columns compare value of ration-specific growth in row to the high-high value. ..................... 160
Table 23. Year 2 comparison of growth rates (daily growth rate between sampling dates,
mm/day) and December-spawn growth (change in fork length)in five stocks of lake trout housed
in a common environment hatchery, exposed to one of four feeding regimes. % difference
columns compare value of ration-specific growth in row to the high-high value. ..................... 161
Table 24. AIC values of linear, hyperbolic, logarithmic, sigmoidal, and Michaelis-Menten
functions applied to the relationship between egg number and mean percent lipid of
Michipicoten lake trout housed in a common environment hatchery in Ontario. ...................... 162
Table 25. AIC values of linear, hyperbolic, logarithmic, sigmoidal, and Michaelis-Menten
functions applied to the relationship between egg number and mean percent lipid of Killala lake
trout housed in a common environment hatchery in Ontario. .................................................... 163
Table 26. AIC values of linear, hyperbolic, logarithmic, sigmoidal, and Michaelis-Menten
functions applied to the relationship between egg number and mean percent lipid of 1999 Year
class Kingscote lake trout housed in a common environment hatchery in Ontario. ................... 164
Table 27. AIC values of linear, hyperbolic, logarithmic, sigmoidal, and Michaelis-Menten
functions applied to the relationship between egg number and mean percent lipid of 2000 Year
class Kingscote lake trout housed in a common environment hatchery in Ontario. ................... 165
Table 28. AIC values of linear, hyperbolic, logarithmic, sigmoidal, and Michaelis-Menten
functions applied to the relationship between egg number and mean percent lipid of Manitou lake
trout housed in a common environment hatchery in Ontario. .................................................... 166
Table 29. AIC values of models of best fit, as calculated by AIC comparisons of MichaelisMenten, hyberbolic, linear, sigmoidal, and logarithmic function, and the alternate models which
represent the model of best fit applied to the alternate group (year or ration-group) in the
relationship between egg number and mean percent lipid of five stocks of lake trout housed in a
common environment experiment in an OMNR hatchery. The models of best fit and alternate
models were applied to the same data in each case to determine AIC differences. ................... 167
Table 30. P-values for 2008 hatchery data comparing residuals (absolute values) between ration
treatments based on sigmoidal function applied to 2008 lipid-fecundity data of five stocks of lake
trout held in a common garden experiment. ............................................................................... 168
viii

Table 31. P-values for 2009 hatchery data comparing residuals (absolute values) between ration
treatments based on sigmoidal function applied to 2009 lipid-fecundity data of five stocks of lake
trout held in a common garden experiment. ............................................................................... 169

ix

LIST OF FIGURES

Figure 1. General von Bertalanffy model of length vs. age in lake trout under natural conditions.
For interpretation of the references to color in this and all other figures, the reader is referred to
the electronic version of this thesis. ................................................................................................ 9
Figure 2. Hypothetical curve of potential relationship between age and reproductive investment
in lake trout stocks. ......................................................................................................................... 9
Figure 3. Map of locations of source populations for stocks used in a common environment
study in OMNR hatchery. ............................................................................................................. 24
Figure 4. Locations of lipid measurements taken on lake trout using non-invasive fat meter at
OMNR hatchery in Codrington, Ontario (Photograph Courtesy of Sara Smith). ......................... 30
Figure 5. Relationship between fork length and age from Michipicoten and Kingscote lake trout
stocks sampled in their natural environments. Line shows model of best fit to predict how length
changes as function of age, based on AIC values. ........................................................................ 34
Figure 6. Relationship between fork length and age from Killala and Manitou lake trout stocks
sampled in their natural environments. Line shows model of best fit to predict how length
changes as function of age, based on AIC values. ........................................................................ 35
Figure 7. Stock-specific modeled relationship between fork length and age of four lake trout
stocks in natural environments. The age range of the curve shown indicates the age range of
individuals sampled in the field. ................................................................................................... 36
Figure 8. Relationship between mean percent lipid and age from Michipicoten and Kingscote
lake trout stocks sampled in their natural environments. Line shows model of best fit to predict
how percent lipid changes as function of age. .............................................................................. 38
Figure 9. Relationship between mean percent lipid and age from Killala and Manitou lake trout
stocks sampled in their natural environments. Line shows model of best fit to predict how
percent lipid changes as function of age. ...................................................................................... 39
Figure 10. Stock-specific modeled relationship between mean percent lipid and age of four lake
trout stocks in natural environments. The age range of the curve shown indicates the age range
of individuals sampled in the field. ............................................................................................... 40
Figure 11. Relationship between egg diameter and age from Michipicoten and Kingscote lake
trout stocks sampled in their natural environments. Line shows regression of best fit based on
AIC values. ................................................................................................................................... 42
Figure 12. Relationship between egg diameter and age from Killala and Manitou lake trout
stocks sampled in their natural environments. Line shows regression of best fit based on AIC
values. ........................................................................................................................................... 43
x

Figure 13. Stock-specific modeled relationship between egg diameter and age of four lake trout
stocks in natural environments. The age range of the curve shown indicates the age range of
individuals sampled in the field. Lines show regression of best fit in each stock, based on AIC
values. ........................................................................................................................................... 44
Figure 14. Relationship between egg diameter and age from Michipicoten and Kingscote lake
trout stocks sampled in their natural environments. Lines show regression of best fit in each
stock, based on AIC values. .......................................................................................................... 46
Figure 15. Relationship between egg diameter and age from Killala and Manitou lake trout
stocks sampled in their natural environments. Lines show regression of best fit in each stock,
based on AIC values. .................................................................................................................... 47
Figure 16. Stock-specific modeled relationship between eggs/mm (number of egg produced /
length of individual) and age of four lake trout stocks in natural environments. The age range of
the curve shown indicates the age range of individuals sampled in the field. Lines show
regression of best fit in each stock, based on AIC values. ............................................................ 48
Figure 17. Comparison of fork length of four stocks of lake trout sampled in the wild to hatchery
individuals held in a common garden environment during the 2008 and 2009 spawning seasons.
The letters “A” and “B” above each data column indicate statistical groupings within a stock
based on Tukey’s test (p<0.05). Field values based on model and are not included in statistical
comparison. ................................................................................................................................... 50
Figure 18. Comparison of percent lipids of four stocks of lake trout sampled in the wild to
hatchery individuals held in a common garden environment during the 2008 and 2009 spawning
seasons. The letters “A” and “B” above each data column indicate statistical groupings within a
stock based on Tukey’s test (p<0.05). Field values based on model and are not included in
statistical comparison. ................................................................................................................... 52
Figure 19. Comparison of egg size of four stocks of lake trout sampled in the wild to hatchery
individuals held in a common garden environment during the 2008 and 2009 spawning seasons.
The letters “A” and “B” above each data column indicate statistical groupings within a stock
based on Tukey’s test (p<0.05). Field values based on model and are not included in statistical
comparison. ................................................................................................................................... 54
Figure 20. Comparison of egg number of four stocks of lake trout sampled in the wild to
hatchery individuals held in a common garden environment during the 2008 and 2009 spawning
seasons. The letters “A” and “B” above each data column indicate statistical groupings within a
stock based on Tukey’s test (p<0.05). Field values based on model and are not included in
statistical comparison. ................................................................................................................... 56
Figure 21. Comparison of Fork Length of four stocks of lake trout during: Year 1 of the hatchery
study, Year 2 of the hatchery study, and modeled field value at 8.5 years old. The letters “A” and
“B” above each data column indicate statistical groupings within a year based on Tukey’s test
(p<0.05). Field values based on model and are not included in statistical comparison. .............. 58

xi

Figure 22. Comparison of mean percent lipids of four stocks of lake trout during: Year 1 of the
hatchery study, Year 2 of the hatchery study, and modeled field value at 8.5 years old. The letters
“A” and “B” above each data column indicate statistical groupings within a year based on
Tukey’s test (p<0.05). Field values based on model and are not included in statistical
comparison. ................................................................................................................................... 60
Figure 23. Comparison of egg size of four stocks of lake trout during: Year 1 of the hatchery
study, Year 2 of the hatchery study, and modeled field value at 8.5 years old. The letters “A” and
“B” above each data column indicate statistical groupings within a year based on Tukey’s test
(p<0.05). Field values based on model and are not included in statistical comparison. .............. 62
Figure 24. Comparison of eggs number of four stocks of lake trout during: Year 1 of the hatchery
study, Year 2 of the hatchery study, and modeled field value at 8.5 years old. The letters “A” and
“B” above each data column indicate statistical groupings within a year based on Tukey’s test
(p<0.05). Field values based on model and are not included in statistical comparison. .............. 64
Figure 25. Figure showing relationship between egg number and fork length of spawning female
Michipicoten lake trout held in a common environment in OMNR hatchery in Codrington,
Ontario. Regression is linear, the linear equation fit to each relationship and corresponding r2
value are shown............................................................................................................................. 74
Figure 26. Figure showing relationship between egg number and fork length of spawning female
Killala lake trout held in a common environment in OMNR hatchery in Codrington, Ontario.
Regression is linear, the linear equation fit to each relationship and corresponding r2 value are
shown. ........................................................................................................................................... 75
Figure 27. Figure showing relationship between egg number and fork length of spawning female
Kingscote lake trout held in a common environment in OMNR hatchery in Codrington, Ontario.
Regression is linear, the linear equation fit to each relationship and corresponding r2 value are
shown. ........................................................................................................................................... 76
Figure 28. Figure showing relationship between egg number and fork length of spawning female
Manitou lake trout held in a common environment in OMNR hatchery in Codrington, Ontario.
Regression is linear, the linear equation fit to each relationship and corresponding r2 value are
shown. ........................................................................................................................................... 77
Figure 29. Locations of lipid measurements taken on lake trout using non-invasive fat meter at
OMNR hatchery in Codrington, Ontario (Photograph Courtesy of Sara Smith). ....................... 104
Figure 30. Change in growth rate (in mm/day) of Michipicoten stock housed in a common
environment experiment exposed to one of four feeding regimes in OMNR hatchery. ............. 109
Figure 31. Change in growth rate (in mm/day) of Killala stock housed in a common
environment experiment exposed to one of four feeding regimes in OMNR hatchery. ............. 110
Figure 32. Change in growth rate (in mm/day) of the 1999 Kingscote stock housed in a common
environment experiment exposed to one of four feeding regimes in OMNR hatchery. ............. 111
xii

Figure 33. Change in growth rate (in mm/day) of the 2000 Kingscote stock housed in a common
environment experiment exposed to one of four feeding regimes in OMNR hatchery. ............. 112
Figure 34. Change in growth rate (in mm/day) of the Manitou stock housed in a common
environment experiment exposed to one of four feeding regimes in OMNR hatchery. ............. 113
Figure 35. Growth rate of five stocks of lake trout housed in a common environment exposed to
one of four feeding rations. Standard error is shown above each bar. Statistical groupings within
a stock, based on an ANOVA pairwise comparison with a Tukey-Kramer adjustment are shown
above each bar............................................................................................................................. 114
Figure 36. Change in mean percent lipid of Michipicoten lake trout housed in a common
environment experiment exposed to one of four feeding regimes in OMNR hatchery. ............. 116
Figure 37. Change in mean percent lipid of Killala lake trout housed in a common environment
experiment exposed to one of two feeding regimes in OMNR hatchery. ................................... 117
Figure 38. Change in mean percent lipid of 1999 Kingscote lake trout housed in a common
environment experiment exposed to one of four feeding regimes in OMNR hatchery. ............. 118
Figure 39. Change in mean percent lipid of 2000 Kingscote lake trout housed in a common
environment experiment exposed to one of four feeding regimes in OMNR hatchery. ............. 119
Figure 40. Change in mean percent lipid of Manitou lake trout housed in a common
environment experiment exposed to one of four feeding regimes in OMNR hatchery. ............. 120
Figure 41. Mean spawning-time lipid levels of five stocks of lake trout housed in a common
environment exposed to one of four feeding regimes. Standard error is shown above each bar.
Statistical groupings within a stock, based on an ANOVA pairwise comparison with a TukeyKramer adjustment are shown above each bar. .......................................................................... 121
Figure 42. Mean egg diameter of five stocks of lake trout housed in a common environment
exposed to one of four feeding regimes in OMNR hatchery. Standard error is shown above each
bar. Statistical groupings within a stock, based on an ANOVA pairwise comparison with a
Tukey-Kramer adjustment are shown above each bar. ............................................................... 123
Figure 43. Egg number (corrected for length of spawning female) of five stocks of lake trout
housed in a common environment exposed to one of four feeding regimes. Standard error is
shown above each bar. Statistical groupings within a stock, based on an ANOVA pairwise
comparison with a Tukey-Kramer adjustment are shown above each bar. ................................ 125
Figure 44. Relationship between egg number as measured in eggs/mm and egg size of
Michipicoten lake trout housed under multiple feeding regimes in a common garden environment
in OMNR hatchery during 2008 spawning season. P values show the significance of changes
between Year 1 and Year 2 in eggs/mm and egg size. ............................................................... 127
Figure 45. Relationship between egg number as measured in eggs/mm and egg size of Killala
lake trout housed under multiple feeding regimes in a common garden environment in OMNR
xiii

hatchery during 2008 spawning season. P values show the significance of changes between Year
1 and Year 2 in eggs/mm and egg size. ...................................................................................... 127
Figure 46. Relationship between egg number as measured in eggs/mm and egg size of 1999
Kingscote lake trout housed under multiple feeding regimes in a common garden environment in
OMNR hatchery during 2008 spawning season. P values show the significance of changes
between Year 1 and Year 2 in eggs/mm and egg size. ............................................................... 128
Figure 47. Relationship between egg number as measured in eggs/mm and egg size of 2000
Kingscote lake trout housed under multiple feeding regimes in a common garden environment in
OMNR hatchery during 2008 spawning season. P values show the significance of changes
between Year 1 and Year 2 in eggs/mm and egg size. ............................................................... 128
Figure 48. Relationship between egg number as measured in eggs/mm and egg size of Manitou
lake trout housed under multiple feeding regimes in a common garden environment in OMNR
hatchery during 2008 spawning season. P values show the significance of changes between Year
1 and Year 2 in eggs/mm and egg size. ...................................................................................... 129
Figure 49. Relationships between percent lipid and egg number (in eggs/mm) of Michipicoten
lake trout housed in a common environment hatchery exposed to one of four feeding regimes.
Parameters used in each function included on each graph. Functions for each stock and year
were selected using AIC values of linear, hyperbolic, logarithmic, Michaelis-Menten and
sigmoidal function. ..................................................................................................................... 134
Figure 50, Relationships between percent lipid and egg number (in eggs/mm) of Killala lake
trout housed in a common environment hatchery exposed to one of four feeding regimes.
Parameters used in each function included on each graph. Functions for each stock and year
were selected using AIC values of linear, hyperbolic, logarithmic, Michaelis-Menten and
sigmoidal function. ..................................................................................................................... 135
Figure 51. Relationships between percent lipid and egg number (in eggs/mm) of 1999 Kingscote
lake trout housed in a common environment hatchery exposed to one of four feeding regimes.
Parameters used in each function included on each graph. Functions for each stock and year
were selected using AIC values of linear, hyperbolic, logarithmic, Michaelis-Menten and
sigmoidal function. ..................................................................................................................... 136
Figure 52. Relationships between percent lipid and egg number (in eggs/mm) of 2000 Kingscote
lake trout housed in a common environment hatchery exposed to one of four feeding regimes.
Parameters used in each function included on each graph. Functions for each stock and year
were selected using AIC values of linear, hyperbolic, logarithmic, Michaelis-Menten and
sigmoidal function. ..................................................................................................................... 137
Figure 53. Relationships between percent lipid and egg number (in eggs/mm) of Manitou lake
trout stocks housed in a common environment hatchery exposed to one of four feeding regimes.
Parameters used in each function included on each graph. Functions for each stock and year
were selected using AIC values of linear, hyperbolic, logarithmic, Michaelis-Menten and
sigmoidal function. ..................................................................................................................... 138
xiv

Figure 54. Relationships between percent lipid and egg number (in eggs/mm) of Michipicoten
lake trout housed in a common environment hatchery exposed to one of four feeding regimes.
Parameters used in each function included on each graph. Functions for each stock and year
were selected using AIC values of linear, hyperbolic, logarithmic, Michaelis-Menten and
sigmoidal function. Ration-specific regressions are grouped by non-switchers (High-High and
Low-Low) and switchers (High-Low and Low-High). The solid line represents the regression
applied to non-switchers, the dotted line shows the regression applied to switchers. ................ 139
Figure 55. Relationships between percent lipid and egg number (in eggs/mm) of 1999 Kingscote
lake trout housed in a common environment hatchery exposed to one of four feeding regimes.
Parameters used in each function included on each graph. Functions for each stock and year
were selected using AIC values of linear, hyperbolic, logarithmic, Michaelis-Menten and
sigmoidal function. Ration-specific regressions are grouped by non-switchers (High-High and
Low-Low) and switchers (High-Low and Low-High). The solid line represents the regression
applied to non-switchers, the dotted line shows the regression applied to switchers. ................ 140
Figure 56. Relationships between percent lipid and egg number (in eggs/mm) of 2000 Kingscote
lake trout housed in a common environment hatchery exposed to one of four feeding regimes.
Parameters used in each function included on each graph. Functions for each stock and year
were selected using AIC values of linear, hyperbolic, logarithmic, Michaelis-Menten and
sigmoidal function. Ration-specific regressions are grouped by non-switchers (High-High and
Low-Low) and switchers (High-Low and Low-High). The solid line represents the regression
applied to non-switchers, the dotted line shows the regression applied to switchers. ................ 141
Figure 57. Relationships between percent lipid and egg number (in eggs/mm) of Manitou lake
trout housed in a common environment hatchery exposed to one of four feeding regimes.
Parameters used in each function included on each graph. Functions for each stock and year
were selected using AIC values of linear, hyperbolic, logarithmic, Michaelis-Menten and
sigmoidal function. Ration-specific regressions are grouped by non-switchers (High-High and
Low-Low) and switchers (High-Low and Low-High). The solid line represents the regression
applied to non-switchers, the dotted line shows the regression applied to switchers. ................ 142

xv

Chapter 1: Background Information and Conceptual Model
INTRODUCTION
Study Rationale
The allocation of energy to different physiological processes in individual lake
trout (Salvelinus namaycush) is determined by a combination of both environmental
conditions present, and the genetics of the individual; the underlying genetic makeup
(alleles) are usually selected for based on long-term environmental conditions in the local
environment (Comstock and Moll 1963). Local adaptations to prevailing environmental
conditions, as well as reproductive, behavioral, and geographic isolation allow separate
populations (or stocks) to genetically diverge and develop unique life history strategies
(Loftus 1958; Goodier 1981; Marsden et al 1995; Wilson and Mandrak 2004).
Optimally, these stock-specific life history strategies maximize lifetime reproductive
output in specific environments (Stearns 1989; Roff 1992). Energy allocation towards
physiological processes, such as growth and reproduction, are of interest for management
of this natural resource because both contribute to survival and abundance. In this thesis,
I explore the relationship between energy acquisition, growth, storage, and reproduction
in four lake trout stocks from different populations with varying life histories. These
stocks were raised in similar hatchery conditions to determine if such allocations fit into a
conceptual model that accommodates life history. More specifically, I explore how each
stock responds, in a common controlled environment, to a changed food ration and
evaluate how energy is reallocated towards growth, storage, and reproduction.

1

Life History Theory and Optimal Energy Allocation
Life history theory predicts that there will be intra-specific differences in energy
allocation towards maintenance, growth, and reproduction to maximize lifetime
reproductive output in a given environment (e.g. Jokela 1997; Heino and Kaitala 1999).
The general model used to express energetic demands in an individual at a given point in
time is:
G + R = C – (Rmet + F + U + A)
where G = growth, R = reproduction, C = consumption, R met = respiration (standard
metabolism), F = fecal egestion, U = excretion, and A = locomotion (Diana 1982).
All excess energy not accounted for in the equation is presumably allocated to
storage, in the form of lipid reserves. Energy is usually assumed to be first allocated
toward maintenance costs (standard metabolism, fecal egestion, excretion, and
locomotion), while the remaining energy is selectively allocated toward growth,
reproduction and storage (Kozlowski 1992). How the remainder of this energy is
allocated and what factors affect this allocation is one of the most important topics in life
history theory and optimal resource allocation theory (Stearns 1992, Jokela 1997). In this
study, I evaluate stock-specific energy allocation in lake trout to growth, reproduction,
and percent lipid which I assume is a measure of energy allocated toward storage.
. Storage of energy in the form of lipid reserves is widespread across species that
live in seasonably variable environments as these reserves serve as a way for individuals
to take advantage of times when food is abundant (Schultz and Conover 1997; Hurst
2007; Jonsson and Jonsson 2009). This stored energy can then be selectively allocated
towards growth, reproduction, or maintenance when food is more limiting and also
2

allows more flexibility in energy allocation priorities. This model of the role of storage
in energy allocation is summarized in a conceptual model in Jokela (1997) that explains
storage as a way to cope with seasonal differences in food consumption and to deal with
times when required energy does not reflect energetic demands.
Given the importance of storage discussed above, why not put all energy from
food consumption towards storage to maximize flexibility in energy allocation
throughout the season? Perhaps there are benefits to putting energy towards reproduction
and growth immediately. In iteroparous species, reproductive output is typically highly
correlated with spawner length (Roff 1992). This means that immediate allocation of
energy toward growth will increase maximum reproductive potential in following years,
likely at the expense of reproductive output in that year (Heino and Kaitala 1999; Jokela
1997). Although the goal is to maximize lifetime reproductive output, allocating all
energy toward growth would not be advantageous because maintaining critical levels of
storage is required for over-winter survival (Gardiner and Geddes 1980; Biro et al. 2004).
Lake trout inhabit a wide variety of environments including deep water lakes,
arctic lakes, and rivers across their wide geographic range that stretches across most of
Canada and the Northern United States (Scott and Crossman 1973). Lake trout stocks
show differences in life history strategies across their range which reflect local conditions
and theoretically maximize lifetime reproductive output of individuals in a given area. In
this study, I make predictions of how variations in lifespan, growth rate, and diet in the
wild influence the prioritization of allocation of energy to lipid reserves, growth, and
reproductive output. My ultimate goal is to apply a conceptual framework to make

3

predictions regarding the energy allocation of four lake trout stocks, relying on
information collected from a literature review and data gathered in the wild.
Stocks with long lifespans are expected to come from environments with low
mortality which is expected in systems that have no sea lamprey, low fishing mortality,
low predation rates on young lake trout as well as enough food resources. In these longlived stocks, fish will allocate a relatively low amount of energy to reproductive output
on an annual basis because they have a relatively long amount of time to maximize
lifetime reproductive output, and will prioritize energy allocation to growth in younger
years (Table 1; Williams 1966; Charnov 2005). Short-lived stocks are expected to come
from environments with high adult mortality (due to fishing pressure, sea lamprey, etc.)
or high predation rates on young lake trout. These short-lived stocks are expected to
prioritize allocation of energy to reproduction because they have a smaller amount of
time to maximize lifetime reproductive output (Table 1; Williams 1966; Charnov 2005).
Tracking growth rates of individuals in the field is logistically difficult, therefore
insight into stock-specific growth rates can be gained through analysis of length-at-age
curves. Differences in the slopes of these curves between stocks indicate differences in
growth rates between stocks. In stocks with steeper length-at-age curves, more rapid
growth is expected to be seen in stocks with high adult mortality (high fishing pressure,
sea lamprey), or gape-limited predators that prey on smaller lake trout (Table 1;
Hutchings 1993; Chivers et al. 2008). For the purposes of this study, growth is defined as
the change in fork length between two points in time.
In this study, diet is defined as the quantity and quality (energy density) of food
that stocks have access to in their natural environments. While diet is ultimately going to

4

drive all aspects of energy allocation in individuals because it determines energy intake, I
hypothesize that natural diet is going to be the most predictive of percent lipid. I predict
that stocks with a high quality diet in the wild (piscivorous, abundance of smaller bodied
fishes, little competition, etc.) will have higher percent lipid than those stocks that have
lower quality diets (lack of smaller bodied fishes, low productivity, high levels of
competition, etc; Table 1).
Table 1. Conceptual model of energy allocation in lake trout to reproduction, growth,
and storage and the hypothesized effects of stock-specific lifespans, local environmental
conditions (fishing pressure, predation, etc.), and diet on the amount of energy allocated
to each energetic category.
Energy Allocation
Energetic Category
Description
Priority
Short-lived (i.e., 10-20 Years)
High
Reproduction
Low (especially in
Long-lived (i.e., 20-40 Years)
younger ages)
Piscivorous, high productivity
High
Storage
Planktivorous, low productivity
Low
High fishing pressure, sea
High
lamprey presence, high predation
Growth Rate
Low fishing pressure, no sea
Low
lamprey, low predation
Seasonal Energy Allocation
Lake trout are an iteroparous species found in temperate zones (Wilson and
Mandrak 2004) and only have one spawning episode per year (Esteve et al. 2008; Muir et
al. 2012). This life history strategy is usually a result of adaptations to local
environments that select for low annual reproductive potential, and delayed maturation to
allow increased adult survival and reproductive success in sub-optimal environments
(Winemiller and Rose 1992; Shuter et al. 1998). How such life history adaptations would
refine my conceptual model is that I predict that part of the year will prioritize energy
allocation towards gonadal growth, where more energy will be shunted towards building

5

eggs, and towards spawning events (Quince et al 2008; Table 2). However, once the fish
spawns eggs, she will then spend energy on growth processes. Any excess energy
accumulated (in storage) will also be allocated towards somatic growth after spawning
(Quince et al 2008). Growth has been shown to be highly correlated with seasonal shifts
in food availability and reproductive demands (Mangel and Clark 1988; Houston and
McNamara 1999) in organisms, such as lake trout, that live in seasonally variable
environments (Power 2002). Growth is expected to slow over the winter and also during
the spawn because of these influences; however, excess storage can help jump start
growth and maintain fish over lean winter months when lipid levels are depleted (Swift
1955; Josrgensen et al. 2005) (Table 2). Lake trout start to build gonads in March
(Henderson and Wong 1998) and will put some energy towards lipids, but the majority of
energy is expected to go towards growth as gonads are still very small, especially in the
spring. This trend is expected to continue throughout the summer when metabolism and
food consumption is high, until periods of final gonad growth (late summer and Fall),
where most energy will be devoted towards final gonadal growth and very little energy
will be put towards somatic growth and storage (Table 2).
Table 2. Conceptual model of energy allocation in lake trout to reproduction, growth,
and storage and the hypothesized effects of time of year on the amount of energy
allocated to each energetic category.
Season
Growth
Storage
Reproduction
Converted to
Post-Spawn: Winter and Early Spring
High
High
Growth
Pre-Spawn: Spring and Early Summer
High
Low
Low
Summer
Medium
Medium
Medium
Converted to
Spawn
Low
High
Reproduction

6

Age-Specific Energy Allocation
Lake trout tend to be long-lived but have a wide range of maximum sizes across
their wide geographic range (McDermid et al. 2010) and will spawn over several years as
a form of bet hedging (Winemiller and Rose 1992; Donaldson-Matasci et al. 2008) The
relatively long life span of lake trout makes them a good candidate for life history studies
because it allows long-term monitoring of growth, storage, reproduction, and other
characteristics tied to life history strategy. As they age and approach the end of their
lifespan, they devote much less energy towards growth and storage and allocate more
energy to reproduction (Hutchings 1993; Quince et al. 2008a; Quince et al. 2008b). Most
growth trajectories of fish are modeled to simulate initial periods of rapid growth
followed by a slowing down which is attributed to slowing to reproduction, but rarely is
mechanistically modeled as such (Figure 1) (von Bertalanffy 1938; Cox 2010). Some
researchers also suggest that lake trout may also show evidence of senescence (Foster
1984; Lantry et al. 2008) and as a result, will show less reproduction as they enter old age
(Table 3). I build this effect of age into my conceptual model (Table 3). In my model,
based on other published models (e.g. Quince et al. 2008), I predict that immature fish
put more energy into growth and storage to maximize growth and survival during the
winter months. In immature fish, the relatively small amount of energy allocated to
reproduction is used for gonadal development, while energy allocated toward growth is
maximized.
Mature fish that have not yet reached the age of maximum reproduction (Table 3)
are expected to allocate energy equally between growth, storage and reproduction. These
fish are expected to put more energy towards reproduction than immature fish because
they are actively spawning, which reduces the energy available for growth and storage.
7

However, in these young spawning fish, energy allocated towards growth and lipids is
still critical, as the fish are in the period of rapid growth (Figure 1). Fish that are at the
age in which reproductive output is maximized are expected to put high amounts of
energy towards reproduction, at the expense of growth and storage (Figure 2). This age
also corresponds with the maximum point of the growth rate curve (Figure 1).
Mature fish that have passed the age at maximum reproduction (Table 3) and may
be in senescence are expected to show a decrease in energy allocation towards both
growth and reproduction, with increased amounts of energy being put towards lipid
reserves, representing a general inefficiency and shutting down associated with age
(Lantry et al. 2008; Gunderson and Dygert 1988). Basal metabolism and maintenance
costs are also increased in larger, older fish, especially in food-limited environments
(Paterson et al. 2005). This increased energy demand reduces the amount of energy
available to allocate towards growth, storage, and reproduction.
The age at which reproductive output is maximized was hypothesized to differ
between stocks with overall adult mortality for the population driving this life history
adaptation. Annual variability in environmental conditions that drive eventual year-class
success are variable from year to year so it is unlikely that these lifetime reproductive
maximum values are highly dependent on environmental conditions during the lifetime of
an individual fish. There are tradeoffs involved in the “decision” to spawn earlier in life
versus later in life. When individuals spawn early in life, they are sacrificing energy that
could have been put towards growth and reducing future survival and reproduction
(Williams 1966a; Williams 1966b). When fish delay maturation and spawn later in life,
they run the risk of not surviving long enough to spawn (Stearns 1989).

8

Table 3. Conceptual model of energy allocation in lake trout to reproduction, growth,
and storage and the hypothesized effects of age on the amount of energy allocated to each
energetic category.
Age
Growth
Storage
Reproduction
Immature Fish
High
Medium
Low
Young, Mature Fish
Medium
Medium
Medium
Mature Fish (Maximized Reproduction)
Low
Low
High
Old, Mature Fish
Low
Medium
Medium

Figure 1. General von Bertalanffy model of length vs. age in lake trout under natural
conditions. For interpretation of the references to color in this and all other figures, the
reader is referred to the electronic version of this thesis.

Figure 2. Hypothetical curve of potential relationship between age and reproductive
investment in lake trout stocks.
Because lake trout stocks are adapted to different environments, it is likely that
there are some genetic limitations for each stock in terms of their abilities to respond to
9

environmental conditions, which would indicate that the response to the environment has
a heritable component. However, there is strong evidence that there are general
environment-based responses that are similar between stocks (McDermid et al. 2007), as
a result of phenotypic plasticity. Phenotypic plasticity is the ability of an individual
organism to use a single, fixed genotype to express multiple phenotypes, based on
environmental conditions (Bradshaw 1965). It is critical to understand the stock-specific
abilities of lake trout to adjust their energy allocation patterns based on local conditions.
Lake trout stocks that overlap in the plasticity of a phenotypic expression in a given
environment will have the same ability to respond to, and succeed, in that specific
environment. However, understanding the limitation of each stock in their plasticity of
growth, storage, and reproduction will allow discovery of the non-overlapping potential
phenotypic expressions between stocks. These stock-specific differences in the range of
phenotypes will drive differences in reproductive performance and survival of these
stocks under specific environmental conditions, especially when resources are limiting.
The plasticity, or scope for adaptation, within lake trout stocks will determine how far
each stock can be pushed before growth or reproduction fall to the point where the stock
will not be successful, based on the prevailing environmental conditions that could be
altered due to a number of reasons, such as direct anthropogenic influences (e.g.,
harvesting), invasive species, or global climate change. Furthermore, establishing a link
between growth, storage, and reproduction will serve as a powerful tool for stock-specific
population models, as the environment and foraging success of an individual will directly
affect its reproductive output.

10

Because of the economic and ecological significance of lake trout in the Great
Lakes region, many management programs and research efforts have focused on their
restoration through stocking programs and investigation of the factors that led to their
decline. A more specific understanding of the scope for adaption in lake trout stocks will
allow managers to make more meaningful predictions of how lake trout will respond to
changes in their environment, whether it be an introduction to a new environment via
stocking, such as in rehabilitation programs, or changes in their existing environment
through introductions of non-native species, loss of critical habitat, changes in prey base,
or other human-induced changes.
Study Approach - Common Environment Experiment
In this study, a common environment experiment is used to evaluate the responses
of growth, storage, and reproduction of lake trout stocks to multiple feeding regimes.
The term “common environment” refers to a type of experiment in which geneticallydistinct organisms, or in this case fish stocks, are housed under common conditions in a
controlled environment to determine the phenotypic response of each respective stock.
Because reproductive traits in fish are influenced by both the environment and genetics,
common environment experiments allow me to determine the response of individuals to
changes in the environment, based on genetics (i.e., heritable constraints and phenotypic
plasticity) and the scope for adaptation of each stock. Common environment experiments
have been used to determine genetic differences in many organisms such as: amphibians
(Urban 2008), plants (Williams et al. 2008), insects (Armbuster and Conn 2006),
mammals (Dobson and Baudoin 2002), fish (Piche et al. 2008), and have been used to
track physiological (Strimbeck et al. 2007), morphological (Piche et al. 2008),
11

reproductive (Dobson and Baudoin 2002), and behavioral (Dobson and Baudoin 2002)
responses of these various species to environmental conditions.
In this study, lake trout from the Killala, Michipicoten, Manitou, and Kingscote
stocks were held in a common environment experiment for 2.5 years to see how energy
allocation towards growth, storage, and reproduction changes throughout the year and
across years. These stocks show different life history adaptations and I compared data
collected from the field to characterize the life history strategies employed by stocks.
Percent lipid, length, and fecundity (egg number and egg size) data taken from wild
individuals from the stocks studied in this experiment provide an overall view of each
stock’s life history in terms of life span, growth rate, and age-specific fecundity. This
allows me to speculate on how percent lipid, growth and reproduction will change in
hatchery fish as they age on year. This data informed my conceptual model of energy
allocation, and I manipulated the food ration of the hatchery fish and explored how
changing food ration affected storage (percent lipid), growth and reproductive output
(total, egg number and egg size) to see if patterns fit into what I would expect based on
their life history. I also compared the relationship between egg size and egg number
across the stocks in both the field and the hatchery to determine how this relationship is
affected by environmental conditions and fish age.
I explored whether physiological reaction norms, in the form of egg production in
response to food availability, changes as a function of stock. Specifically, can egg
production be predicted from a given percent lipid measurement? Reaction norms
examine the change in a variable in response to a range of environmental conditions. The
term “reaction norm” refers to all the possible phenotypes that can be potentially

12

expressed when an organism is exposed to a range of environmental conditions, or
alternatively, what changes occur in a phenotypic trait as environmental conditions (e.g.
food availability) change (Stearns and Koella 1986, Barot et al. 2004). The possible
existence of reaction norms was explored to determine the response of egg number to
percent lipid in four lake trout stocks to determine how individuals will respond when
energetic intake is altered (for example, when stocked into a new, and potentially food
limiting environment). Although referred to as a reaction norm in this study, I evaluated
multiple individuals from the same stock under different environmental conditions, rather
than evaluating a single organism under different environmental conditions which
constitutes a “true” reaction norm. A reaction norm exists when the relationship between
two variables is predictable and robust, meaning that fecundity can be reliably predicted
based on percent lipid, even after an environmental change occurs. Reaction norms can
be powerful predictive tools, and in this case, could be used to predict how fecundity of
multiple lake trout stocks with different life history adaptations responds to food
availability.

13

WORKS CITED

14

WORKS CITED

Armbuster, P. and J. E. Conn. 2006. Geographic Variation of Larval Growth in North
American Aedes albopictus (Diptera: Culicidae). Annals of the Entomological
Society of America 99: 1234-1243.
Barot, S., M. Heino, L. O’Brien, and U. Dieckmann. 2004. Long term trend in the
maturation norm of two cod stocks. Ecological Applications 14: 1257-1271.
Bradshaw, A.D. 1965. Evolutionary significance of phenotypic plasticity in plants. Adv.
Genet., 13: 115–155.
Charnov, E. L. 2005. Reproductive effort is inversely proportional to average adult life
span. Evolutionary Ecology Research ember 7: 1221-1222.
Chivers, D. P., Zhao, X., Brown, G. E., Marchant, T. A., and M. C. Ferrari. 2008.
Predator-induced changes in morphology of a prey fish: the effects of food level
and temporal frequency of predation risk. Evolutionary Ecology 22: 561-574.
Comstock, R.E., and R.H. Moll. 1963. Genotype-environment interactions. In: W.D.
Hanson and H.F. Robinson, editors, Statistical genetics and plant breeding.
National Academy of Sciences-National Research Council (U.S.) Publ. 982.
National Academies, Washington, DC. p. 164–196.
Cornelius, F. C., K. M. Muth and R. Kenyon. 1995. Lake Trout Rehabilitation in Lake
Erie: A Case History. Journal of Great Lakes Research 21:65-82.
Cox, B. 2010. Assessment of an invasive lake trout population in Swan Lake, Montana.
Master of Science Thesis, Montana State University.
Dobson, F.S., and C. Baudoin. 2002. Experimental tests of spatial association and
kinship in monogamous mice (Mus spicilegus) and polygynous mice (Mus
musculus domesticus). Canadian Journal of Zoology-Revue Canadienne de
Zoologie 80: 980-986.
Donaldson-Matasci, M. C., M. Lachmann, and C. T. Bergstrom. 2008. Phenotypic
diversity as an adaptation to environmental uncertainty. Evolutionary Ecology
Research 10: 493-515.
Edwards, C. J., R. A. Ryder, and T. R. Marshall. 1990. Using Lake Trout as a Surrogate
of Ecosystem Health for Oligotrophic Waters of the Great Lakes. Journal of
Great Lakes Research 16: 591-608.

15

Eschmeyer, P.H., and A. M. Phillips Jr. 1965. Fat content of the flesh of siscowets and
lake trout from Lake Superior. Transactions of the American Fisheries Society
94: 62-74.
Eshenroder, R. L., N. R. Payne, J. E. Johnson, C. Bowen II and M. P. Ebener. 1995.
Lake Trout Rehabilitation in Lake Huron. Journal of Great Lakes Research
21:108-127.
Esteve M, McLennan D. A., and J. M. Gunn. Lake trout (Salvelinus namaycush)
spawning behaviour: the evolution of a new female strategy. Environmental
Biology of Fishes 2008: 69-76.
Fleming, I. A. and M. R. Gross. 1990. Latitudinal clines: a trade-off between egg
number and size in Pacific salmon. Ecology 71: 1-11.
Foster, N.R. 1984. Physiology and behavior. In Strategies for rehabilitation of lake trout
in the Great Lakes: proceedings of a conference on lake trout research, August
1983, R.L. Eshenroder, T.P. Poe, and C.H. Olver, eds., pp 22–27. Great Lakes
Fishery Commission Technical Report 40.
Goodier JL (1981) Native lake trout (Salvelinus namaycush) stocks in the Canadian
waters of Lake Superior prior to 1955. Can J Fish Aquat Sci 38:1724–1737.
Gregersen, F., Haugen, T. O., and N. Larsen. 2006. Egg size differentiation among
sympatric demes of brown trout: possible effects of density-dependent
interactions among fry. Ecology of Freshwater Fish 15: 237-246.
Gunderson, D. R., and P. H. Dygert. 1988. Reproductive effort as a predictor of natural
mortality rate. Journal Du Conseil 44: 200-209
Heino, M., and V. Kaitala. 1999. Evolution of resource allocation between growth and
reproduction in animals with indeterminate growth. Journal of Evolutionary
Biology 12: 423-429.
Henderson, B. A., and J. L. Wong. 1998. Control of lake trout reproduction: role of
lipids. Journal of fish biology, 52: 1078-1082.
Houston, A. and J. McNamara. 1999. Models of Adaptive Behaviour. New York, NY:
CambridgeUniversity Press.
Hutchings, J. A. 1993. Adapted life histories effected by age-specific survival and
growth rate. Ecology 74: 673-684.
Jastrebski, C. J., and Y. E. Morbey. 2009. Egg Size Variation in Lake Trout: Phenotype–
Habitat Correlations Show an Effect of Rearing Environment. Transactions of the
American Fisheries Society 138: 1342-1351.
16

Jokela, J. (1997). Optimal energy allocation tactics and indeterminate growth: Lifehistory evolution of long-lived bivalves. In Evolutionary Ecology of Freshwater
Animals (pp. 179-196). Birkhäuser Basel.
Jonsson, N. and B. Jonsson. 1999. Trade-off between egg mass and egg number in
brown trout. Journal of Fish Biology 55: 767-783.
Josrgensen, E. H., Johanse, J. H. S., and M. Jobling. 2005. Seasonal patterns of growth,
lipid deposition and lipid depletion in anadromous Arctic charr. Journal of Fish
Biology 51: 312-326.
Krueger, C. C., M. L. Jones and W. W. Taylor. 1995. Restoration of lake trout in the
Great Lakes: Challenges and strategies for future management. Journal of Great
Lakes Research 21: 547-558.
Lantry, B. F., O’Gorman, R., and L. S. Machut. 2008. Maternal Characteristics versus
Egg Size and Energy Density: Do Stocked Lake Trout in Lake Ontario
Experience Premature Reproductive Senescence. Journal of Great Lakes
Research 34: 661-674.
Lienesch, P. W., M. E. McDonald, A. E. Hershey, W. J. O’Brien, and N. D. Bettez.
2005. Effects of a whole-lake, experimental fertilization on lake trout in a small
oligotrophic arctic lake. Hydrobiologia 548: 51-56.
Lobon-Cervia, H., Utrilla, C., Rincon, P., and F. Amezcua. 2003. Environmentally
induced spatio-temporal variations in the fecundity of brown trout Salmo trutta
L.: trade-offs between egg size and number. Freshwater Biology 38: 277-288.
Loftus, K. H. 1958. Studies on River-Spawning Populations of Lake Trout in Eastern
Lake Superior. Transactions of the American Fisheries Society 87: 259-277.
Mangel, M. and C. W. Clark 1988. Dynamic Modeling in Behavioral Ecology. Princeton,
NJ: Princeton University Press.
Marsden J. E., Casselman J. M., and T. A.Edsall 1995. Lake trout spawning habitat in
the Great Lakes—a review of current knowledge. Journal of Great Lakes
Research 21(Supplement 1):487–497.
McDermid, J. L., P. E. Ihssen, W. N. Sloan, B. J. Shuter. 2007. Genetic and
Environmental Influences on Life History Traits in Lake Trout. Transactions of
the American Fisheries Society 136: 1018-1029.
McDermid, J. L., Shuter, B. J., and N. P. Lester. 2010. Life history differences parallel
environmental differences among North American lake trout (Salvelinus
namaycush) populations. Canadian Journal of Fisheries and Aquatic Sciences 67:
314-325.
17

Muir, A. M., Blackie, C. T., and C. C. Krueger. 2012. Lake charr Salvelinus namaycush
spawning behaviour: new field observations and a review of current knowledge.
Reviews in Fish Biology and Fisheries 22: 575-593.
Paterson, G., Huestis, S. Y., Whittle, D. M., Drouillard, K. G., and G. D. Haffner. 2005.
In situ measurement of tissue turnover and energy conversion efficiencies in lake
trout (Salvelinus namaycush) using a novel toxicokinetic approach. Canadian
Journal of Fisheries and Aquatic Sciences, 62: 464-471.
Piche, J., J.A. Hutchings, and W. Blanchard. 2008. Genetic variation in threshold
reaction norms for alternative reproductive tactics in male Atlantic salmon, Salmo
salar. Proceedings of the Royal Society B-Biological Sciences 275: 1571-1575.
Power, G. 2002. Charrs, glaciations and seasonal ice. Environmental Biology of Fishes
64: 17–35.
Quince, C., Abrams, P. A., Shuter, B. J., and N. P. Lester. 2008(b). Biphasic growth in
fish I: theoretical foundations. Journal of Theoretical Biology 254: 197-206.
Quince C, Shuter BJ, Abrams PA, Lester NP. 2008. Biphasic growth in fish II: empirical
assessment. J. Theor. Biol. 254(2): 207-214.
Roff, D. A. 1992. The evolution of life histories: theory and analysis. Chapman and
Hall, New York.
Scott, D. P. 1962. Effect of Food Quantity on Fecundity of Rainbow Trout, Salmo
gairdneri. Journal of the Fisheries Research Board of Canada 19: 715-731.
Scott, W. B. and E. J. Crossman. 1973. Freshwater fishes of Canada. Fisheries Research
Board of Canada. Bulletin 184. Ottawa, Ontario.
Simons, A. M. 2007. Selection for increased allocation to offspring number under
environmental unpredictability. Journal of evolutionary biology, 20: 813-817.
Scott, W. B., and E. J. Crossman. 1973. Freshwater fishes of Canada. Bulletin of the
Fisheries Research Board of Canada 184.
Shuter, B. J., M. L. Jones, R. M. Korver, and N. P. Lester. 1998. A general, life history
based model for regional management of fish stocks: The inland lake trout
(Salvelinus namaycush) fisheries of Ontario. Canadian Journal of Fisheries and
Aquatic Sciences 55: 2161-2177.

18

Smith, S. H. 1972. Factors of ecologic succession in oligotrophic fish communities of
the Laurentian Great Lakes. Journal of the Fisheries Research Board of Canada
29: 717-730.
Smith, C. C., and S. D. Fretwell. 1974. The Optimal Balance between Size and Number
of Offspring. The American Naturalist 108:499-506.
Stearns, S. C. 1989. Trade-Offs in Life-History Evolution. Functional Ecology 3: 259268.
Strimbeck, G. R., T. D. Kjellsen, P. G. Schaberg, and P. F. Murakami. 2007. Cold in the
common garden: comparative low-temperature tolerance of boreal and temperate
conifer foliage. Trees- Structure and Function 21: 557-567.
Swift, D. R. 1955. Seasonal Variations in the Growth Rate, Thyroid Gland Activity, and
Food Reserves of Brown Trout (Salmo trutta linn.). Journal of Experimental
Biology 32: 751-764.
Urban, M.C. 2008. Salamander evolution across a latitudinal cline in gape-limited
predation risk. Oikos 117: 1037-1049.
von Bertalanffy L. 1938. A quantitative theory of organic growth (inquiries on growth
laws. 11.). Human Biology 10: 181-213.
Wilberg, M. J., M. J. Hansen, and C. R. Bronte. 2003. Historic and Modern Abundance
of Wild Lean Lake Trout in Michigan Waters of Lake Superior: Implications for
Restoration Goals. North American Journal of Fisheries Management 23:100108.
Williams, G.C. 1966a. Adaptation and Natural Selection. Princeton University Press,
Princeton.
Williams, G.C. 1966b. Natural selection, the cost of reproduction, and a refinement of
Lack's principle. American Naturalist 100: 687-690.
Williams, J.L., H. Auge, and J.L. Maron. 2008. Different gardens, different results:
native and introduced populations exhibit contrasting phenotypes across common
gardens. Oecologia 157: 239-248.
Wilson, C. C., and N. E. Mandrak. 2004. Chapter two: History and evolution of lake
trout in Shield Lakes. Pages 21-35 in J. M. Gunn, R. J. Steedman, and R. A.
Ryder, editors. Boreal Shield Watersheds: Lake Trout Ecosystems in a Changing
Environment. CRC Press LLC.

19

Winemiller, K.O., and K.A. Rose. 1992. Patterns of life history diversification in North
American fishes: implications for population regulation. Canadian Journal of
Aquatic Sciences 49: 2196-2218.

20

Chapter 2: Characterization of Hatchery Fish in Context of Wild Counterparts
INTRODUCTION
Understanding how managed species, like lake trout, allocate energy towards
reproduction and growth is important for stock rehabilitation and restoration (Mason et
al. 1998). Many theoretical and empirical studies suggest that populations of organisms
will allocate energy towards growth and reproduction in a way that will maximize
reproductive output in face of their prevailing environmental conditions (Hutchings 1993;
Quince et al. 2008), resulting in heritable life history traits. The goal of my study is to
determine if particular stocks alter energy allocation towards growth, reproduction and
storage as part of their life history strategy, in both hatchery and field, because if so, then
I have the potential to develop powerful predictive tools that can be incorporated into
population models (Shuter et al. 1998; Quince et al. 2008). In this study, I tracked
growth, reproduction and storage of energy in lipid reserves in sexually mature female
lake trout (Ages 8 and 9 during Year 1 of the study) from four different populations that
were raised in a common environment. These hatchery fish were fed known rations (see
Chapter 3), and if the allocation patterns were completely plastic and dependent on the
environment, then I would expect every stock to show similar responses in growth,
reproduction and storage. However, my hypothesis is that some characteristics will be
heritable and I will see differences. Therefore, to make sense of any observed patterns
and to determine what factors could be contributing to life history adaptations, I first had
to determine how percent lipid, reproductive output, and fish length compare to their wild
counterparts, and how these characteristics vary with age. This chapter is a comparison
of the characteristics we can measure in wild fish related to life history (growth,

21

reproduction, percent lipid), to similar characteristics measured in hatchery fish, under
controlled conditions.
There are differences between hatchery-reared individuals and wild individuals
that make it difficult to predict responses of wild fish to changes in environmental
conditions from observation of hatchery-reared individuals. Several differences observed
between hatchery fish and their wild counterparts, that could be due to genetic or
environmental influences, include the following: the level of aggressive behaviors
(Deverill et al. 1999), morphology, osmoregulatory system performance (Hill et al.
2006), predator avoidance behavior (Beamish et al. 1992, Olla et al. 1994), and growth
(Hindar et al. 1991; Fleming et al. 1997; Fleming et al. 2002). These differences affect
the survival and reproduction of hatchery fish in the wild and impair my ability to
extrapolate hatchery studies to general conclusions about wild populations. The main
concern with stocking hatchery fish into wild environments for restoration purposes is
that hatchery fish come from different geographic areas and may not be adapted to the
new environment (Weber and Fausch 2003). In recent years, more care has been taken to
stock individuals into the wild that are of the same stock, or genetic strain, of the wild
fish that naturally occur in the area by spawning wild individuals to create progeny raised
in hatcheries. However, sometimes a particular genetic strain has been extirpated as was
the case with the native lake trout in Lake Michigan, and as a result, restoration is only
possible with a different strain (Dexter et al. 2010; Hanson et al. 2012).
Hatchery individuals may also substantially diverge from the genetics of their
wild counterparts if offspring of wild fish are not consistently brought into the hatchery
system. The progeny of the breeding hatchery fish will become more domesticated with

22

each generation, and will not have normal natural selection processes that determine
which individual survives and reproduces. Divergences between hatchery and wild
individuals may also occur due to conditions in the hatchery that are not consistent with
natural conditions, such as fish density, food availability, food type, and water quality
(Weber and Fausch 2003). While responses to these different conditions may be a result
of short-term phenotypic responses or long-term genetic adaptations, these differences in
environmental conditions affect the behavior and ultimately the probability of survival
and reproduction of these hatchery fish when they are stocked into natural waters (Olla et
al. 1994). The purpose of this chapter specifically is to compare the reproduction, lipid
storage and growth between hatchery and wild fish within four lake trout stocks and year
classes and to use what is known about the natural environment of each stock to explain
why these trends were observed.
Background of Hatchery Stocks
In this study I tracked 1999 and 2000 year-class lake trout from stocks derived
from 4 different environments. I studied lake trout from Michipicoten Island in Lake
Superior, lake trout from Killala Lake, which is a protected lake north of the Neys
Provincial Park in Ontario, lake trout from Kingscote Lake, a small, inland lake in
Algonguin Provincial Park in Ontario, Canada, and lake trout from Lake Manitou on
Manitoulin Island in Lake Huron (Figure 3). The fish available for study at the time of
this experiment were spawned at different times (1999 for Manitou, 2000 for
Michipicoten and Killala), however both 1999 and 2000 year classes were available for
Kingscote so I included them to tease out environment, tank effects and age differences.

23

Figure 3. Map of locations of source populations for stocks used in a common
environment study in OMNR hatchery.
The Michipicoten Island lake trout stock spawn in the shoals near Michipicoten
Island in the Canadian waters of Eastern Lake Superior. The Michipicoten stock grows
relatively slowly, is piscivorous, and has adapted to a Great Lakes system which has
smaller-bodied fish such as sculpins, lake whitefish, and rainbow smelt (Gamble et al.
2011a; Gamble et al. 2011b) available as a food source. This stock is being used for
restoration of native lake trout to the Canadian portion of Lake Huron, primarily in areas
where there is minimal sea lamprey abundance and low amounts of fishing pressure
(Ebener 1998).

24

The Killala stock are progeny from individuals from Killala Lake, Ontario, which
is a fish sanctuary for lake trout and fishing is not permitted. Killala lake trout are
relatively long-lived and are used for stocking of inland lakes (Ontario Ministry of
Natural Resources Policy Report).
The Kingscote stock are progeny of individuals from Kingscote Lake in southern
Algonquin Park, Ontario. The Kingscote lake trout are distinct in appearance because
they exhibit a uniform silver color and lack the characteristic markings observed on other
lake trout stocks (Wilson and Mandrak 2004). These differences in appearance may be
due to genetic differences, unique diets, or a combination of these two factors. Another
key characteristic is that the Kingscote stock is planktivorous in its natural habitat, due to
the lack of smaller-bodied forage fish (Martin and Olver 1976).
The Lake Manitou stock originated from an inland lake located on the Manitoulin
Island of Lake Huron (Ebener 1998). Lake Manitou lake trout are a short-lived, fastgrowing stock that have been used for restoration in Lake Huron since 1981 and have
shown higher survival rates when stocked in areas that have low fishing pressure and
limited sea lamprey abundance (Ontario Ministry of Natural Resources Policy Report
2005).

25

METHODS
Wild Collections
Lake trout from the Michipicoten, Killala, Kingscote, and Manitou stocks were
collected from their natural habitats during sampling trips with the Ontario Ministry of
Natural Resources (OMNR). The following stocks were sampled during the fall the
following years: Michipicoten (2008, Benjamin Felt and OMNR), Killala (2006, Cheryl
Murphy with OMNR), Kingscote (2006, Cheryl Murphy with OMNR), and Manitou
(2007, Cheryl Murphy with OMNR). Fish were captured using a combination of gill nets
and trap nets, with set durations and effort being specific to a site based on lake trout size
and abundance. Fish were euthanized immediately on site, using a concentrated dose of
MS-222. Mature females were scanned with a non-invasive fat meter to measure percent
lipids (described below), measured for length and weight, the ovary was removed and
stored for future enumeration and egg diameter measurements, and the otoliths were
removed for ageing by the OMNR.
Hatchery
Individually-marked lake trout from four stocks were held in a common garden
environment in an Ontario Ministry of Natural Resources (OMNR) hatchery in
Codrington, Ontario and were fed tank-specific rations to evaluate fecundity, growth, and
percent lipids in 2008 and 2009. Lake trout from each stock used in this study were
maintained separately in the hatchery early in life, but with similar densities, water
temperature, dissolved oxygen, and food regimes. As adults, the fish were individually
branded and moved to circular tanks that received high quality water at ambient
temperatures from a nearby stream. Fish were fed high quality salmonid brood pellets
26

formulated specifically for the OMNR by Corey Aquafeeds. The reared fish consisted of
individuals from the following stocks and year classes (with year class in parentheses):
Kingscote (1999, 2000), Michipicoten (2000), Manitou (1999), and Killala (2000). The
Killala, Manitou, and Michipicoten individuals used in this study were two generations
removed from the wild. Both year classes of the Kingscote stock used in this study were
first generation hatchery fish. Only fish from the 2000 Kingscote stock were used for the
purposes of this chapter; two year classes of the Kingscote stock allowed analysis of the
effects of age and ration in Chapter 3. Including both year classes in this chapter was
unnecessary because differences in percent lipid, growth and reproductive output
between the year classes in the ration combined data were negligible and comparison of
the 2000 Kingscote year class across two years allowed me to evaluate the effects of age
and tank effects in the next chapter.
Fish were maintained at similar densities in each tank during rearing and the
densities were continually adjusted due to minimal mortality rates of approximately 1-2
fish per tank per year. During the first two spawning seasons, about half of each stock
were maintained on a high ration (0.4% of tank biomass per day) or a low ration (0.2% of
tank biomass per day) to determine stock-specific patterns relating body percent lipid to
fecundity (Murphy et al., in prep). This study monitored reproductive investment and
percent lipids for an additional 2.5 years, with some fish remaining on a high food ration,
and others remaining on a low food ration. This study added an additional treatment by
creating “switchers”, or fish that were previously maintained on a low ration and
switched to a high, and vice versa. A minimum of eighteen fish from each stock were
maintained under a constant ration to track the effects of age alone on the relationship

27

between percent lipids and fecundity when held in a constant environment (Table 4). In
addition to allowing me to look at responses of stocks to changing ration levels, another
goal of manipulating ration levels in the hatchery was to represent variability in available
food that would somewhat mimic natural environments.

28

Table 4. Summary of initial fish densities and composition of lake trout stocks housed in a common environment experiment in
OMNR hatchery in Codrington, Ontario.
2000
2000
2000
1999
1999
Tank
Total 2000 Kingscote
Ration
Michipicoten
Killala
Slate
Manitou
Kingscote
No.
Fish
Male

1
2
3
4
5
6
7
8

High-High
High-Low
Low-Low
Low-High
High-High
High-Low
Low-Low
Low-High

30
31
28
28
29
30
30
30

Female

Male

Female

Male

Female

1
2
1
2

8
7
8
8

1
4
2
2

10
8
7
8

1
0
1
1

9
2
9
0

Female

2

6

2

Male

Female

Male

Female

2
3
2
2

29

Male

5
12
11
12
11

3
6
2
4

12
10
14
13

Fish were processed approximately every eight weeks to determine weight, total
length, fork length and fat content using a non-invasive fat meter (Distell Inc.). The fat
meter was calibrated prior to each measurement period and was used to measure the fat
content of each fish at four locations on the body (Figure 4). The fat meter uses lowpower microwaves to penetrate through the skin and measure the attenuation of the
microwaves by water in the flesh which is negatively correlated with fat content,
allowing a reliable lipid estimate (www.distell.com). The fat meter was calibrated in a
previous study by Claus (2011), and the following equation was used where Muscle Lipid
Concentration was a function of the raw fat meter reading:
(1.9436 + ln (Fat meter Reading)) – 9.5022

] x 100

Muscle Lipid Concentration = [e

2

(r = 0.6117)

Figure 4. Locations of lipid measurements taken on lake trout using non-invasive fat
meter at OMNR hatchery in Codrington, Ontario (Photograph Courtesy of Sara Smith).
During the Fall of 2008 and 2009, females were checked weekly to determine if
they had ovulated by lightly applying pressure to the urogenital region. If eggs were
easily extruded, the entire volume of eggs was stripped from the females into a
volumetric flask to determine total volume. One 30-mL egg sample was preserved in a
5% formalin solution to measure the diameter of the eggs at a future date using digital

30

microscope pictures of the eggs and analysis using Leica Application Suite ®. When
measuring eggs, thirty total eggs from each female were measured and each egg
measured six times for average egg diameter, to determine average egg size for a female.
Data Analyses
I standardized the number of eggs collected from each female by dividing by the
female’s length. The correlation between spawner size and fecundity is well documented
and corrections of fecundity based on spawner length are common (Blanchard et al. 2011;
Downs et al. 1997; Lobon-Cervia et al. 1997). For my analyses, egg production was
corrected for length of the spawner because of the strong linear correlation between these
two variables (Figure 27). Additionally, lipid levels were averaged across all lipid
measurement locations within each individual to approximate mean muscle lipid content
of the fish across the entire body.
To determine how length, percent lipid, egg size and egg number were influenced
by age, I modeled the best functions to describe these relationships using linear,
quadratic, hyperbolic, and logarithmic functions to choose the model that had the lowest
Akaike’s Information Criterion (AIC) value. AIC values were found using the following
formula:

 RSS 
AIC = n × ln
 + 2k
 n 
The parameter, k, represents the number of parameters within the model; the
parameter, n, represents the number of observations, and the parameter, RSS, represents
the Residual Sum of Squares (Burnham and Anderson 2002). I then used the model with
the lowest AIC value to estimate stock-specific length, percent lipid, egg size, and egg
31

number of wild individuals at age 8.5 years, which is the average age of hatchery fish
during the study. In cases where the linear model, or another model with fewer
parameters than the selected model, had an AIC difference of ≤ 2 (Burnham and
Anderson 2002) compared to the “best” model, the simpler model was selected. A
general guideline proposed in Burnham and Anderson (2002) is that models with a
difference of AIC less than or equal to two (AICInitial – AICAlternate ≤ 2) have strong
support to show there is no difference between models in terms of their ability to describe
the data (Burnham and Anderson 2002; Stylianou et al. 2013). Selection of simpler
models allows for a more straightforward comparison of metric values across stocks.
I grouped all the hatchery data across rations for the hatchery-field comparison
because variability in ration levels simulates differences in foraging success in the
hatchery and the exact ration that most closely approximates natural feeding conditions in
each stock is unknown. A post-hoc Tukey-Kramer test (α = 0.05) was used to determine
the significance of differences in length, mean percent lipid, egg size, and egg number
between hatchery years.

32

RESULTS
Characterizing the Wild Populations
Fork Length
A linear relationship was used to predict length at age for the wild caught fish in
all stocks based on AIC values for model selection (Figure 5, Table 6, Table 7, Table 8,
and Table 9). These stock-specific linear relationships in age vs. fork length were plotted
on one figure for comparison (Figure 7).

All stocks increased in length with age, but the

short-lived Kingscote and the long-lived Killalas were the slowest growing. The shortlived Manitou had a much higher rate of growth (steeper slope), while the long-lived
Michipicoten grew the fastest (Figure 5, Figure 6).

33

Figure 5. Relationship between fork length and age from Michipicoten and Kingscote
lake trout stocks sampled in their natural environments. Line shows model of best fit to
predict how length changes as function of age, based on AIC values.

34

Figure 6. Relationship between fork length and age from Killala and Manitou lake trout
stocks sampled in their natural environments. Line shows model of best fit to predict
how length changes as function of age, based on AIC values.

35

Figure 7. Stock-specific modeled relationship between fork length and age of four lake trout stocks in natural environments. The age
range of the curve shown indicates the age range of individuals sampled in the field.

36

Percent Lipids
Linear models were used to predict changes in percent lipid as a function of age in
the Michipicoten, Killala, and Manitou stocks. However, the Kingscote stock showed a
quadratic relationship between mean percent lipid and age because the AIC value showed
a difference of greater than 2, relative to the linear model (Figure 8, Figure 9, Table 6,
Table 7, Table 8, and Table 9). I created a generalized stock-specific figure that plotted
the lipids against age regression for all stocks on one graph to ease comparisons between
stocks (Figure 10). Both Kingscote and Killala had very low percent lipid values, across
all ages, while Manitou and Michipicoten had higher percent lipids, with maximum
values occurring at the youngest ages in the Manitou stock while Michipicoten mean
percent lipids showed a gradual increase with age. The stocks with low lipids did not
show any change with percent lipid across ages, except for a slight decline with age in the
Kingscotes. However, the two stocks with high percent lipids showed opposite patterns
with age. The long-lived Michipicoten stock gradually increased percent lipid with age,
while the short-lived Manitou showed a rapid decline in percent lipid with age (Figure
10).

37

Figure 8. Relationship between mean percent lipid and age from Michipicoten and
Kingscote lake trout stocks sampled in their natural environments. Line shows model of
best fit to predict how percent lipid changes as function of age.
38

Figure 9. Relationship between mean percent lipid and age from Killala and Manitou
lake trout stocks sampled in their natural environments. Line shows model of best fit to
predict how percent lipid changes as function of age.

39

Figure 10. Stock-specific modeled relationship between mean percent lipid and age of four lake trout stocks in natural environments.
The age range of the curve shown indicates the age range of individuals sampled in the field.

40

Egg Size
Linear relationships were used to model the relationship between age and egg
diameter (Figure 11, Figure 12) in all four stocks, based on AIC values (Figure 11, Figure
12, Table 6, Table 7, Table 8, and Table 9). All stocks showed an increase in egg
diameter as fish age (Figure 11, Figure 12). The Kingscote and Manitou stocks had the
smallest eggs and also showed a rapid increase in size with age (Figure 11, Figure 12). In
contrast, the Killala had the largest eggs, followed by the Michipicoten stock, but both
showed a more gradual increase in size with age (Figure 13).

41

Figure 11. Relationship between egg diameter and age from Michipicoten and Kingscote
lake trout stocks sampled in their natural environments. Line shows regression of best fit
based on AIC values.
42

Figure 12. Relationship between egg diameter and age from Killala and Manitou lake
trout stocks sampled in their natural environments. Line shows regression of best fit
based on AIC values.

43

Figure 13. Stock-specific modeled relationship between egg diameter and age of four lake trout stocks in natural environments. The
age range of the curve shown indicates the age range of individuals sampled in the field. Lines show regression of best fit in each
stock, based on AIC values.

44

Egg Number
Linear models were used to model the relationship between age and egg number
(in eggs/mm) in the Michipicoten, Kingscote, and Manitou stocks, based on AIC values.
However, a quadratic relationship was assigned to the age-egg number relationship in the
Killala stock because of a difference in AIC values that exceeded 2 between the quadratic
and linear models (Table 6, Table 7, Table 8, and Table 9). The stocks with the higher
percent lipids (Manitou and Michipicoten) had the highest egg number. There was an
increasing trend in egg number with age in the Michipicoten stock while the Manitou
stock showed a gradual decrease in egg production with age (Figure 11, Figure 12). The
Killala stock showed decreasing egg production as fish age while the Kingscote stock
showed increases in egg production with age. At the youngest ages, which are the most
comparable to the hatchery stocks, the Manitou and Michipicoten stocks appear to
produce the most number of eggs per mm of fish, followed by Killala and Kingscote,
respectively (Figure 14, Figure 15, Figure 16).

45

Figure 14. Relationship between egg diameter and age from Michipicoten and Kingscote
lake trout stocks sampled in their natural environments. Lines show regression of best fit
in each stock, based on AIC values.
46

Figure 15. Relationship between egg diameter and age from Killala and Manitou lake
trout stocks sampled in their natural environments. Lines show regression of best fit in
each stock, based on AIC values.
47

Figure 16. Stock-specific modeled relationship between eggs/mm (number of egg produced / length of individual) and age of four
lake trout stocks in natural environments. The age range of the curve shown indicates the age range of individuals sampled in the
field. Lines show regression of best fit in each stock, based on AIC values.

48

Comparison of Field and Hatchery Data within Stocks
To compare wild fish to hatchery fish, I used the modeled equations for predicting
length, percent lipid, egg size, and egg number by age to calculate these metrics for an
8.5-year-old fish. I compared these simulated values to what I observed in the Year 1 and
Year 2 of my hatchery study (Figure 17, Figure 18, Figure 19, and Figure 20). I also
determined if length, percent lipid, egg size and egg number differed between years of the
hatchery study.
Results suggest that my hatchery fish grew significantly each year (Figure 17),
and that the stocks that have low percent lipid in the wild (i.e., Kingscote and Killala)
were bigger in the hatchery than their wild counterpart, at the same age. The Manitou
and Michipicoten stocks were smaller in the hatchery than what I would expect to see
from a wild fish of the same age (Figure 17).

49

Figure 17. Comparison of fork length of four stocks of lake trout sampled in the wild to hatchery individuals held in a common
garden environment during the 2008 and 2009 spawning seasons. The letters “A” and “B” above each data column indicate statistical
groupings within a stock based on Tukey’s test (p<0.05). Field values based on model and are not included in statistical comparison.

50

The percent lipid in hatchery fish greatly differed from the percent lipid observed
in their wild counterparts in the Kingscote and Killala stocks. The hatchery-held
Kingscote and Killala had much higher percent lipid than the wild-caught individuals,
while the hatchery-held Michipicoten and Manitou stocks had slightly higher percent
lipid as wild fish at the same age (Figure 18).

51

Figure 18. Comparison of percent lipids of four stocks of lake trout sampled in the wild to hatchery individuals held in a common
garden environment during the 2008 and 2009 spawning seasons. The letters “A” and “B” above each data column indicate statistical
groupings within a stock based on Tukey’s test (p<0.05). Field values based on model and are not included in statistical comparison.
52

Egg size in hatchery fish was very comparable to the values measured/estimated
from wild fish. However, the long-lived stocks (i.e., Michipicoten and Killala), for which
I could only use estimated data for an age comparison, had smaller eggs in the hatchery
than what would be expected in the wild (Figure 19). The egg size in the hatchery
appeared to decrease with age for the Manitou, Kingscote and Killala stocks (Figure 19),
which is opposite of what I found in wild fish (Figure 13).

53

Figure 19. Comparison of egg size of four stocks of lake trout sampled in the wild to hatchery individuals held in a common garden
environment during the 2008 and 2009 spawning seasons. The letters “A” and “B” above each data column indicate statistical
groupings within a stock based on Tukey’s test (p<0.05). Field values based on model and are not included in statistical comparison.
54

Egg production was similar between hatchery fish and wild fish in the Manitou
and Killala stocks (Figure 20). However, wild Kingscotes had lower egg production than
seen in the hatchery while wild Michipicotens had increased egg production in the wild
relative to their hatchery counterparts.

55

Figure 20. Comparison of egg number of four stocks of lake trout sampled in the wild to hatchery individuals held in a common
garden environment during the 2008 and 2009 spawning seasons. The letters “A” and “B” above each data column indicate statistical
groupings within a stock based on Tukey’s test (p<0.05). Field values based on model and are not included in statistical comparison.

56

Comparison of Field and Hatchery Data between Stocks
Michipicoten fish were consistently larger than the other stocks, when compared
across stocks within years (Figure 21 and Table 10), which was consistent with modeled
predictions, in which I predicted Michipicoten to be the longest, followed by Manitou,
Killala, and Kingscote, respectively. However, the differences in length in the hatchery
fish from the other stocks did not differ much from each other.

57

Figure 21. Comparison of Fork Length of four stocks of lake trout during: Year 1 of the hatchery study, Year 2 of the hatchery study,
and modeled field value at 8.5 years old. The letters “A” and “B” above each data column indicate statistical groupings within a year
based on Tukey’s test (p<0.05). Field values based on model and are not included in statistical comparison.

58

Percent lipid differed significantly between stocks in the hatchery, with the fish
from stocks that show low percent lipid in the wild exhibiting the highest percent lipids in
Year 1 of the hatchery study (Figure 22 and Table 11). By Year 2, however, only the
Kingscote had significantly higher percent lipids than the other stocks. The percent lipid
observed in the hatchery fish did not compare well to wild counterparts, where the
Michipicoten and Manitou had the highest percent lipid, followed by the Killala stock,
and the Kingscote stock, respectively (Figure 22).

59

Figure 22. Comparison of mean percent lipids of four stocks of lake trout during: Year 1 of the hatchery study, Year 2 of the hatchery
study, and modeled field value at 8.5 years old. The letters “A” and “B” above each data column indicate statistical groupings within a
year based on Tukey’s test (p<0.05). Field values based on model and are not included in statistical comparison.
60

Egg sizes across stocks were consistent from Year 1 to Year 2 and were
comparable to that predicted from their wild counterparts (Figure 23 and Table 12) The
Kingscote were significantly smaller than most of the stocks (except Michipicoten) in
Year 1, and significantly smaller than all stocks in Year 2. The Killala consistently had
larger eggs than the other stocks in Year 1, and in Year 2 (except Michipicoten).

61

Figure 23. Comparison of egg size of four stocks of lake trout during: Year 1 of the hatchery study, Year 2 of the hatchery study, and
modeled field value at 8.5 years old. The letters “A” and “B” above each data column indicate statistical groupings within a year based
on Tukey’s test (p<0.05). Field values based on model and are not included in statistical comparison.
62

The number of eggs produced per millimeter (eggs/mm) of hatchery fish showed
a lot of variability, and there were no statistically significant differences between stocks
each year (Figure 24 and Table 13). The modeled field data suggested that the
Michipicoten and Manitou stocks should produce the most number of eggs, while the
Kingscote would produce the least. The hatchery data from Year 2 suggested a trend that
Manitou fish would produce more eggs/mm. The hatchery Kingscote, on the other hand,
did not show any indication that these stocks would produce the lowest number of
eggs/mm and seemed most divergent from their wild counterparts.

63

Figure 24. Comparison of eggs number of four stocks of lake trout during: Year 1 of the hatchery study, Year 2 of the hatchery study,
and modeled field value at 8.5 years old. The letters “A” and “B” above each data column indicate statistical groupings within a year
based on Tukey’s test (p<0.05). Field values based on model and are not included in statistical comparison.
64

DISCUSSION
Lake trout appear to allocate energy towards growth, reproduction and storage in
ways that may be explained by inherited life history traits, but the artificial environment
of the hatchery makes comparisons more complicated. The most interesting result from
this study is that each stock shows differences in some metrics when raised in common
environment, and these patterns can be explained, in part, by comparison to wild
counterparts.
Stock Differences in Hatchery and Wild Fish
The most conserved trait between wild and hatchery fish was egg size, while the
other traits showed variation between wild and hatchery fish. This is consistent with the
results found in Jastrebski and Morbey (2009) that show the heritability of egg size in
lake trout causes egg size to remain relatively stable in a wide variety of environments
within a given stock.
The more divergent of the two reproductive traits between the hatchery and wild
individuals appears to be egg number. This finding suggests that the local environment
(presumably ration), has a larger effect on egg number than egg size. This is consistent
with previous studies that have found egg number to be variable, depending on
environmental conditions (Scott 1962; Lobon-Cervia 2003).
Percent lipid showed a large difference between the wild and hatchery in the
Kingscote and Killala stocks which are adapted to low-energy diets. This suggests that
percent lipid of individuals is sensitive to local conditions and that environmental
differences between hatchery and wild conditions drive these differences, specifically
energy intake.
65

Growth, as measured by modeled fork length at age 8.5, also showed differences
between hatchery and wild which indicates it is influenced by hatchery conditions and is
likely dependent on food availability in the hatchery.
Insight into Stock-specific Life Histories from Field and Hatchery Data
Four distinct life history strategies resulted in some differences in percent lipid,
growth, and reproductive output between stocks in their natural environments as well as
when held in a common environment.
The wild fish showed differences in relative percent lipid between stocks (Figure
18). The Kingscote fish had lowest lipid levels in the field, and Killala had slightly more
lipid storage. The Michipicoten and Manitou fish had the highest lipid levels in the wild,
both of which were on par with lipid levels observed in the hatchery. These data,
combined with data showing similar growth in the Manitou and Michipicoten stocks
between the hatchery and field, suggest that energy intake in the hatchery is similar to
that in the natural environment for these two stocks.
Kingscote individuals in the hatchery had much higher levels of lipids than the
other stocks, followed by Killala. These two stocks that have limited access to energy
rich food sources in the wild and appear to be adapted to those environments because in
the hatchery, even though they are fed the same as the other stocks, they may be overfed
and put the excess into storage.
The percent lipid appeared to vary considerably between the hatchery and the
wild fish, and may be sensitive to ration. Percent lipid also appears to be influenced by
stock-specific food-conversion efficiencies. This finding shows that percent lipid is
sensitive to environmental conditions, and presumably food consumption, and that
66

changes in food availability in a given environment has the potential to influence the
amount of energy allocated to growth and reproduction because of the importance of lipid
reserves in the flexibility of energy allocation in an individual (Jokela 1997).
Within the hatchery, fork length was very consistent between stocks (Figure 17),
which suggests that there are not strong stock-specific differences in growth rates in a
common environment. However, there are large differences between stocks in length-atage in the wild fish which is likely a result of differences in energy intake based on the
natural environments of the stocks. Only the Manitou stock showed similar lengths in the
hatchery and field which suggests energy intake in the hatchery is on par with their
natural environment.
Egg size was more variable between stocks within the hatchery, but less variable
within a stock across environments (hatchery vs. field), which suggests that egg size has a
strong genetic component, while ration level and local environment play a lesser role
(Figure 19). This trend of consistent egg size within a lake trout stock across a range of
environments was also observed in Jastrebski and Morbey (2009). In my study, egg size
was consistent from Year 1 to Year 2 in the hatchery and was comparable to the wild
fish. However, the modeled estimate of the egg size of the long lived stocks,
Michipicoten and Killala, is larger than that observed in the hatchery. The size difference
could be due to the model extrapolating beyond the available data. Results from this
study confirm other studies that suggest egg size is an inherited trait, which is somewhat
conserved. In natural environments, there are a multitude of strong selection pressures on
egg size including predation, resource availability and oxygen levels (Chambers et al.

67

1989; Reznick et al. 2001; Einum et al. 2002) which would dictate such a rigid life
history trait.
Because egg size appeared to be fixed within a stock, the only other thing that can
vary substantially and adjust to a common environment situation is egg number. There
were no significant differences in egg number between the stocks in the hatchery (Figure
16). However, there seemed to be a trend to suggest that Manitou would produce the
most eggs in Year 2, which is consistent with the pattern observed in wild fish. The stock
that showed a highly divergent pattern in the hatchery compared to the wild was the
Kingscote stock. Kingscotes produced equal numbers of eggs as other stocks, but in the
field, they produce far fewer eggs. This result suggests that energy intake in the hatchery
was driving the egg production. This effect of ration on reproductive output is explored
further in Chapter 3.
Stock-Specific Conceptual Models
The stock-specific conceptual model (Table 5), based on the energy allocation
criteria presented in Chapter 1, would suggest that wild Michipicotens have higher lipid
reserves because they eat more energy-dense forage fish. The conceptual model also
predicts that Michipicotens will devote high amounts of energy towards growth because
of the relatively steep line in the length-at-age data from the wild (Figure 5, Figure 7).
Lastly, the model predicts a relatively low amount of energy allocated towards
reproduction because of the relatively long lifespan of the stock. These predictions,
based on environmental conditions and field data, are consistent with previous studies
that suggest a high investment to storage and growth in much larger environments (such
as the Great Lakes) where it may be beneficial to grow faster and bigger and have readily
68

available energy to survive sparse food or winter (Post and Parkinson 2001; Biro et al.
2004).
The conceptual model of energy allocation in the Killala stock (Table 5) predicted
that they would have a relatively low amount of energy put towards reproduction in a
given year because of their long lifespan. The Killalas are expected to have a low
amount of energy put towards growth because the length-at-age curve had a shallow
slope and also because of limited fishing pressure (OMNR Report). Lastly, the model
predicted that the Killala stock would have a low amount of energy allocated toward
reproduction because of the stock’s relatively long lifespan. In this case, the Killala stock
is expected to have low amounts of energy allocated to growth, storage, and reproduction
relative to the other stocks. Within the Killala stock, I predicted a relatively equal
allocation of energy toward each factor.
The conceptual model for the Kingscote stock (Table 5) predicted that they would
have a relatively high amount of energy put toward reproduction in a given year because
of their short-lived life history strategy. They were predicted to have a low amount of
energy allocated towards storage because of their planktivorous diet and were predicted
to allocate a small amount of energy to growth, based on the length-at-age data gathered.
The Manitou stock, a short-lived stock, derived from an ecologically rich aquatic
community with ample forage fish, was expected to feed on energy-dense food items
(Table 5). The stock is short-lived and has high natural rates of mortality in its natural
environment (Budd et al. 1968; Shuter et al. 1998) which, according to life history theory
(Murphy 1968; Budd et al. 1968) suggests that much energy will be devoted towards
reproduction in both hatchery and wild stocks so that reproduction is successful before

69

the end of life. Indeed, anecdotally, the Manitou strain is known in the hatchery system as
“egg machines”; however, the size of eggs produced by the stock is expected to be
smaller because of the tradeoff between egg size and egg number (Smith and Fretwell
1974). Also, because Lake Manitou has a rich community structure and presumably
high quality prey, they are expected to have a relatively high percent lipid.
Table 5. Conceptual model of energy allocation in lake trout stocks to reproduction,
growth, and storage based on field data gathered and conceptual models summarized in
Tables 1-3 (Chapter 1).
Stock
Growth Storage Reproduction
Michipicoten
High
High
Low
Killala
Low
Low
Low
Kingscote
Low
Low
High
Manitou
High
High
High
Study Limitations
While I found interesting results that could be related to life history, I only studied
stocks that were from divergent life histories. This study would be enhanced if I could
explore multiple stocks showing same life history strategy (i.e. long lived, low lipid) in a
common environment. Therefore, this study is more descriptive of these specific stocks
because there is little overlap in the life history strategies between these stocks.
An additional limitation of this study was the sample size and age range of wild
individuals collected that formed the basis of comparisons. It was difficult to acquire
samples of females encompassing a wide range of ages for the long-lived Michipicoten
and Killala stocks. In the case of the Michipicoten stock, there was a gap in data between
ages 15 and 25 which may have resulted in an inaccurate prediction of variables based on
age in intermediate-aged fish. There were no samples of individuals at younger ages in
the Michipicoten and Killala stocks which forced me to extrapolate beyond the range of
my wild datasets to predict percent lipid, growth, egg size, and egg number at age 8.5.
70

Also, the Kingscote stock had a gap from 8 to 11-years-old in the wild dataset. These
sample size and age range limitations in my wild dataset could be a result of a limited
number of individuals in certain size classes in a given stock, or the bias of gears used
when sampling lake trout with the OMNR.
One possible implication of this potential bias caused by limited sampled sizes
and missing ages in the field samples is seen in the effects of age on egg size between the
hatchery and the field data. Although it is possible that differences between the hatchery
and natural environments are affecting this change in egg size as fish age, it is also
possible that limited sample sizes in younger fish biased my models and resulted in
inaccurate predictions in the extrapolated data.
Implications
A comparison of hatchery and wild individuals from lake trout stocks with a wide
range of life histories permitted exploration into the differences between hatchery fish
and wild fish and what traits are most influenced by environmental differences. A
common environment resulted in similar length at age and egg number between stocks,
while egg size was relatively fixed within each stock at a given age, regardless of
environment. Percent muscle lipid storage showed differences between stocks and also
showed differences within stock between the hatchery and the wild. This indicates that
when fish from energy-limited environments in the wild are exposed to an energy-rich
environment, they take advantage of this environment by moving excess energy to lipid
reserves. Given that the stocks from energy-rich environments grew less than wild fish
when in a hatchery setting, these stocks are not saturated with food in the lab. However,
these stocks still showed allocation of energy to lipid storage. This suggests that there
71

may be a limit to what fraction of their overall intake of energy can be allocated to
growth and that “excess” energy will still be selectively allocated to storage despite a lack
of food. This result is consistent with other studies that show a tradeoff between growth
and storage and also predict that there will be allocation of energy to lipid reserves, even
in limiting environments, to maintain lipids at a minimum critical level to allow overwinter survival (Henderson and Wong 1998; Berg et al. 2011). Given the high
importance of lipids in over-winter survival, growth would then be more indicative of
limiting environments, which was seen in my study as evidenced by differences in
length-at-age between the hatchery and wild fish. The implications of this strategy of
maintaining critical lipid levels despite food abundance in a given year is that growth,
and the associated reproductive output in following years is going to be reduced due to
the energy allocated to lipid reserves for over-winter survival. In the next chapter, I
quantitatively explore the role of these lipid reserves on reproductive output within each
stock to determine the stock-specific responses to multiple rations, and in particular, to
explore the effect of low food ration on egg production, storage and growth to determine
if this stressor is mediated by life history.

72

APPENDIX

73

Figure 25. Figure showing relationship between egg number and fork length of spawning female Michipicoten lake trout held in a
common environment in OMNR hatchery in Codrington, Ontario. Regression is linear, the linear equation fit to each relationship and
2
corresponding r value are shown.

74

Figure 26. Figure showing relationship between egg number and fork length of spawning female Killala lake trout held in a common
environment in OMNR hatchery in Codrington, Ontario. Regression is linear, the linear equation fit to each relationship and
2
corresponding r value are shown.

75

Figure 27. Figure showing relationship between egg number and fork length of spawning female Kingscote lake trout held in a
common environment in OMNR hatchery in Codrington, Ontario. Regression is linear, the linear equation fit to each relationship and
2
corresponding r value are shown.

76

Figure 28. Figure showing relationship between egg number and fork length of spawning female Manitou lake trout held in a
common environment in OMNR hatchery in Codrington, Ontario. Regression is linear, the linear equation fit to each relationship and
2
corresponding r value are shown.

77

Table 6. AIC values and Residual Sum of Squares (RSS) of linear, quadratic,
logarithmic, exponential, and hyperbolic relationships assigned to data age vs.: egg
number, egg diameter, length, and percent lipid in wild Manitou stock lake trout.

Metric

Egg
Number
(Eggs/mm)

Egg
Diameter
(mm)

Length
(mm)

Percent
Lipid

Relationship
Linear
Quadratic
Logarithmic
Exponential
Hyperbolic
Linear
Quadratic
Logarithmic
Exponential
Hyperbolic
Linear
Quadratic
Logarithmic
Exponential
Hyperbolic
Linear
Quadratic
Logarithmic
Exponential
Hyperbolic

No. of
Parameters
(k)

Observations

2
3
2
1
2
2
3
2
1
2
2
3
2
1
2
2
3
2
1
2

35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35
35

No. of
(n)

78

Residual
Sum of
Squares
(RSS)
105.4089
99.3365
106.3809
560.3061
106.865
1.4299
1.3161
1.5184
154.4907
1.6025
25229.8815
25072.8706
26850.0535
10312561.5
28256.8704
205.6643
202.0699
200.1567
1121.8716
200.3428

Akaike
Information
Criterion
(AIC)
42.5875
42.5108
42.9087
99.0597
43.0676
-107.9210
-108.8236
-105.8192
53.9675
-103.9324
234.3153
236.0968
236.4936
442.7734
238.2810
65.9814
67.3643
65.0313
123.3592
65.0639

Table 7. AIC values and Residual Sum of Squares (RSS) of linear, quadratic,
logarithmic, exponential, and hyperbolic relationships assigned to data age vs.: egg
number, egg diameter, length, and percent lipid in wild Killala stock lake trout.

Metric

Relationship

Linear
Quadratic
Egg
Number Logarithmic
(Eggs/mm) Exponential
Hyperbolic
Linear
Quadratic
Egg
Diameter Logarithmic
(mm)
Exponential
Hyperbolic
Linear
Quadratic
Length
Logarithmic
(mm)
Exponential
Hyperbolic
Linear
Quadratic
Percent
Logarithmic
Lipid
Exponential
Hyperbolic

No. of
No. of
Parameters Observations
(k)
(n)
2
3
2
1
2
2
3
2
1
2
2
3
2
1
2
2
3
2
1
2

9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9
9

79

Residual
Sum of
Squares
(RSS)

Akaike
Information
Criterion
(AIC)

5.5226
3.2279
4.9491
19.3958
4.3835
0.419
0.4179
0.4206
45.8561
0.4234
10641.129
10541.8695
11056.6562
1197382.605
11569.2501
8.5225
8.5225
8.5418
23.6218
8.5659

-0.3954
-3.2285
-1.3822
8.9105
-2.4744
-23.6040
-21.6276
-23.5697
16.6546
-23.5100
67.6773
69.5930
68.0221
108.1858
68.4299
3.5094
5.5094
3.5297
10.6845
3.5551

Table 8. AIC values and Residual Sum of Squares (RSS) of linear, quadratic,
logarithmic, exponential, and hyperbolic relationships assigned to data age vs.: egg
number, egg diameter, length, and percent lipid in wild Michipicoten stock lake trout.

Metric

Relationship

Linear
Quadratic
Egg
Logarithmic
Number
(Eggs/mm) Exponential
Hyperbolic
Linear
Quadratic
Egg
Diameter Logarithmic
(mm)
Exponential
Hyperbolic
Linear
Quadratic
Length
Logarithmic
(mm)
Exponential
Hyperbolic
Linear
Quadratic
Percent
Logarithmic
Lipid
Exponential
Hyperbolic

No. of
parameters
(k)

No. of
observations
(n)

2
3
2
1
2
2
3
2
1
2
2
3
2
1
2
2
3
2
1
2

12
12
12
12
12
9
9
9
9
9
12
12
12
12
12
12
12
12
12
12

.

80

Residual
Sum of
Squares
(RSS)
153.9383
145.3113
153.0786
462.6192
152.711
1.0532
1.042
1.0673
45.9341
1.0839
72044.1391
63985.1691
70175.4461
4449190.291
69907.065
375.1968
312.9859
373.3259
571.3356
372.4094

Akaike
Information
Criterion
(AIC)
34.6197
35.9277
34.5525
45.8240
34.5237
-15.3085
-13.4047
-15.1888
16.6698
-15.0499
108.4015
108.9780
108.0862
155.8799
108.0402
45.3105
45.3350
45.2505
48.3568
45.2210

Table 9. AIC values and Residual Sum of Squares (RSS) of linear, quadratic,
logarithmic, exponential, and hyperbolic relationships assigned to data age vs.: egg
number, egg diameter, length, and percent lipid in wild Kingscote stock lake trout.

Metric

Relationship

Linear
Quadratic
Egg
Logarithmic
Number
(Eggs/mm) Exponential
Hyperbolic
Linear
Quadratic
Egg
Diameter Logarithmic
(mm)
Exponential
Hyperbolic
Linear
Quadratic
Length
Logarithmic
(mm)
Exponential
Hyperbolic
Linear
Quadratic
Percent
Logarithmic
Lipid
Exponential
Hyperbolic

No. of
parameters
(k)

No. of
observations
(n)

2
3
2
1
2
2
3
2
1
2
2
3
2
1
2
2
3
2
1
2

15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15
15

81

Residual
Sum of
Squares
(RSS)
7.7512
6.5345
7.4473
12.5507
7.4031
0.4578
0.4483
0.4509
54.5326
0.4462
5269.627
5129.5
5194.624
1634210
5202.925
3.7303
2.6258
3.4656
26.7645
3.0286

Akaike
Information
Criterion
(AIC)
-5.9030
-6.4643
-6.5030
-0.6741
-6.5923
-48.3406
-46.6551
-48.5684
21.3612
-48.7256
91.9250
93.5207
91.7099
175.9793
91.7339
-16.8734
-20.1400
-17.9776
10.6854
-19.9992

Table 10. Comparison of statistical significance of the fork length between stocks based
on ANOVA pairwise-comparison using a Tukey-Kramer adjustment. Field data are not
included in statistical comparison because value is a modeled value.
Year 1 Fork Length
Michipicoten

Killala

Manitou

Kingscote

Michipicoten

n/a

0.0675

0.0003**

0.0334*

Killala

0.0675

n/a

0.6264

0.9774

Manitou

0.0003**

0.6264

n/a

0.1094

Kingscote

0.0334*

0.9774

0.1094

n/a

Killala

Manitou

Kingscote

0.0002** <.0001**

<.0001**

Year 2 Fork Length
Michipicoten
Michipicoten

n/a

Killala

0.0002**

n/a

0.8567

0.9029

Manitou

<.0001**

0.8567

n/a

0.1789

Kingscote

<.0001**

0.9029

0.1789

n/a

82

Table 11. Comparison of statistical significance of the mean percent lipids between
stocks based on ANOVA pairwise-comparison using a Tukey-Kramer adjustment. Field
data are not included in statistical comparison because value is a modeled value.
Year 1 Percent Lipids
Michipicoten

Killala

Manitou

Kingscote

Michipicoten

n/a

0.1576

0.9781

<.0001**

Killala

0.1576

n/a

0.0294*

0.3063

Manitou

0.9781

0.0294*

n/a

<.0001**

Kingscote

<.0001**

0.3063

<.0001**

n/a

Michipicoten

Killala

Manitou

Kingscote

Michipicoten

n/a

0.9536

0.3149

0.0033**

Killala

0.9536

n/a

0.1496

0.0642

Manitou

0.3149

0.1496

n/a

<.0001**

Kingscote

0.0033**

0.0642

<.0001**

n/a

Year 2 Percent Lipids

83

Table 12. Comparison of statistical significance of the egg size between stocks based on
ANOVA pairwise-comparison using a Tukey-Kramer adjustment. Field data are not
included in statistical comparison because value is a modeled value.
Year 1 Egg Size
Michipicoten

Killala

Manitou

Kingscote

Michipicoten

n/a

<.0001**

0.0132*

0.1007

Killala

<.0001**

n/a

0.0026**

<.0001**

Manitou

0.0132*

0.0026**

n/a

<.0001**

Kingscote

0.1007

<.0001** <.0001**

n/a

Year 2 Egg Size
Michipicoten

Killala

Manitou

Kingscote

Michipicoten

n/a

0.3214

0.1044

<.0001**

Killala

0.3214

n/a

0.0008**

<.0001**

Manitou

0.1044

0.0008**

n/a

<.0001**

Kingscote

<.0001**

<.0001** <.0001**

84

n/a

Table 13. Comparison of statistical significance of the egg number between stocks based
on ANOVA pairwise-comparison using a Tukey-Kramer adjustment. Field data are not
included in statistical comparison because value is a modeled value.
Year 1 Eggs/mm
Michipicoten Killala Manitou Kingscote
Michipicoten

n/a

0.9899

0.981

0.5648

Killala

0.9899

n/a

0.8881

0.8202

Manitou

0.981

0.8881

n/a

0.1255

Kingscote

0.5648

0.8202

0.1255

n/a

Year 2 Eggs/mm
Michipicoten Killala Manitou Kingscote
Michipicoten

n/a

0.9994

0.1203

0.0805

Killala

0.9994

n/a

0.2432

0.2002

Manitou

0.1203

0.2432

n/a

0.9999

Kingscote

0.0805

0.2002

0.9999

n/a

85

WORKS CITED

86

WORKS CITED

Beamish, R.J., Thomson, B.L., and G. A. McFarlane. 1992. Spiny dogfish predation on
chinook and coho salmon and the potential effects on hatchery-produced salmon.
Transactions of the American Fisheries Society 121: 444–455.
Biro, P. A., Morton, A. E., Post, J. R., and E. A. Parkinson 2004. Over-winter lipid
depletion and mortality of age-0 rainbow trout (Oncorhynchus mykiss). Canadian
Journal of Fisheries and Aquatic Sciences, 61: 1513-1519.
Blanchard, J. L., Frank, K. T, and J. E. Simon. Effects of condition on fecundity and total
egg production of eastern Scotian Shelf haddock (Melanogrammus aeglefinus).
Canadian Journal of Fisheries and Aquatic Sciences 60: 321-332.
Burnham, K. P., and D. R. Anderson. 2002. Model selection and multi-model inference:
a practical information-theoretic approach. Springer.
Budd, J.C., Fry, F.E.J., and J. B. Smith. 1968. Survival of marked lake trout in Lake
Manitou, Manitoulin Island, Ontario. Journal of the Fisheries Research Board of
Canada. 25: 2257–2268.
Chambers, R. C., Leggett, W. C., and J. A. Brown. 1989. Egg size, female effects, and
the correlations between early life history traits of capelin, Mallotus villosus: an
appraisal at the individual level. Fishery Bulletin, 87: 515-523.
Claus, L. C. 2011. Influence of life history traits on accumulation of polybrominated
diphenyl ethers and polychlorinated biphenyls in three lake trout populations from
Lake Superior. Master of Science Thesis, Michigan State University.
Deverill, J. I., Adams, C. E., and C. W. Bean. 1999. Prior residence, aggression and
territory acquisition in hatchery-reared and wild brown trout. Journal of Fish
Biology 55: 868-875.
Dexter, J. L., Jr., B. T. Eggold, T. K. Gorenflo, W. H. Horns, S. R. Robillard, and S. T.
Shipman. 2010. A fisheries management implementation strategy for the
rehabilitation of Lake Trout in Lake Michigan. Great Lakes Fishery Commission,
Lake Michigan Committee, Ann Arbor, Michigan.
Downs, C. C., White, R. G. and Shepard, B. B. 1997. Age at sexual maturity, sex ratio,
fecundity, and longevity of isolated headwater populations of westslope cutthroat
trout. North American Journal of Fisheries Management 17: 85-92.
Ebener, M. P. 1998. A lake trout rehabilitation guide for Lake Huron. Great
Lakes Fishery Commission 48 pages.
87

Einum, S., Hendry, A. P., and I. A. Fleming. 2002. Egg-size evolution in aquatic
environments: does oxygen availability constrain size?. Proceedings of the Royal
Society of London. Series B: Biological Sciences, 269: 2325-2330.
Eschmeyer, P.H., and A. M. Phillips Jr. 1965. Fat content of the flesh of siscowets and
lake trout from Lake Superior. Transactions of the American Fisheries Society
94: 62-74.
Fleming, I.A., T. Agustsson, B. Finstad, J. I. Johnsson, and B. T. Björnsson. 2002.
Effects of domestication on growth physiology and endocrinology of Atlantic
salmon (Salmo salar). Canadian Journal of Fisheries and Aquatic Sciences 59:
1323–1330.
Fleming, I.A., and Einum, S. 1997. Experimental tests of genetic divergence of farmed
from wild Atlantic salmon due to domestication. ICES J. Mar. Sci. 54: 1051–
1063.
Gamble, A. E., Hrabik, T. R., Stockwell, J. D., and D. L. Yule. 2011a. Trophic
connections in Lake Superior Part I: The offshore fish community . Journal of
Great Lakes Research 37: 541-549
Gamble, A. E., Hrabik, T. R., Stockwell, J. D., and D. L. Yule. 2011b. Trophic
connections in Lake Superior Part II: The nearshore fish community. Journal of
Great Lakes Research 37: 550-560.
Hanson, S. D., Holey, M. E., Treska, T. J., Bronte, C. R., T. H. Eggebraaten. 2012.
Evidence of Wild Juvenile Lake Trout Reproduction in Western Lake Michigan.
North American Journal of Fisheries Management 33:186-191.
Hill, M. S., Zydlewski, G. B., & Gale, W. L. 2006. Comparisons between hatchery and
wild steelhead trout (Oncorhynchus mykiss) smolts: physiology and habitat use.
Canadian Journal of Fisheries and Aquatic Sciences 63: 1627-1638.
Hindar, K., N. Ryman, and F. Utter. 1991. Genetic effects of cultured fish on natural fish
populations. Canadian Journal of Fisheries and Aquatic Sciences 48: 945–957.
Hutchings, J. A. 1993. Adapted life histories effected by age-specific survival and
growth rate. Ecology 74: 673-684.
Jastrebski, C. J., and Y. E. Morbey. 2009. Egg Size Variation in Lake Trout: Phenotype–
Habitat Correlations Show an Effect of Rearing Environment. Transactions of the
American Fisheries Society 138: 1342-1351.
Lobon-Cervia, H., Utrilla, C., Rincon, P., and F. Amezcua. 1997. Environmentally
induced spatio-temporal variations in the fecundity of brown trout Salmo trutta
L.: trade-offs between egg size and number. Freshwater Biology 38: 277-288.
88

Martin, N. V. and C. H. Olver. 1976. The distribution and characteristics of Ontario lake
trout lakes. Ontario Ministry of Natural Resources: Fish and Wildlife Research
Branch: Research Report Number 97.
Mason, D. M., Johnson, T. B., and J. F. Kitchell. 1998. Consequences of prey fish
community dynamics on lake trout (Salvelinus namaycush) foraging efficiency in
Lake Superior. Canadian Journal of Fisheries and Aquatic Sciences 55: 12731284.
Murphy, G. I. 1968. Pattern in Life History and the Environment. The American
Naturalist 102: 391-403.
Olla, B.L., Davis, M.W., and Ryer, C.H. 1994. Behavioural deficits in hatchery-reared
fish: potential effects on survival following release. Aquaculture and Fisheries
Management: 25 (Supplement 1): 19–34.
Ontario Ministry of Natural Resources. 2005. Crown Land Use Policy Atlas – Policy
Report. OMNR, Ontario, Canada.
Post, J. R., and E. A. Parkinson. 2001. Energy allocation strategy in young fish:
allometry and survival. Ecology 82: 1040-1051.
Quince C, Shuter BJ, Abrams PA, Lester NP. 2008. Biphasic growth in fish II: empirical
assessment. J. Theor. Biol. 254(2): 207-214.
Reznick, D., Butler I. V. M. J., and H. Rodd. 2001. Life history evolution in guppies.
VII. The comparative ecology of high and low predation environments. The
American Naturalist, 157(2), 126-140.
Scott, D. P. 1962. Effect of Food Quantity on Fecundity of Rainbow Trout, Salmo
gairdneri. Journal of the Fisheries Research Board of Canada 19: 715-731.
Shuter, B. J., M. L. Jones, R. M. Korver, and N. P. Lester. 1998. A general, life history
based model for regional management of fish stocks: The inland lake trout
(Salvelinus namaycush) fisheries of Ontario. Canadian Journal of Fisheries and
Aquatic Sciences 55: 2161-2177.
Smith, C. C., and S. D. Fretwell. 1974. The Optimal Balance between Size and Number
of Offspring. The American Naturalist 108:499-506.
Quince C, Shuter BJ, Abrams PA, Lester NP. 2008. Biphasic growth in fish II: empirical
assessment. J. Theor. Biol. 254(2): 207-214.

89

Weber, D. W. and K. D. Fausch. 2003. Interactions between hatchery and wild
salmonids in streams: differences in biology and evidence for competition.
Canadian Journal of Fisheries and Aquatic Sciences 60: 1018-1036.
Wilson, C.C. and D.O. Evans. 2010. Algonquin’s silver lake trout: highlighting the
history, habitat, and concerns for a unique biodiversity element. Aquatic Research
and Development Section. Ontario Ministry of Natural Resources. Peterborough,
Ontario.
Wilson, C. C., and N. E. Mandrak. 2004. Chapter two: History and evolution of lake
trout in Shield Lakes. Pages 21-35 in J. M. Gunn, R. J. Steedman, and R. A.
Ryder, editors. Boreal Shield Watersheds: Lake Trout Ecosystems in a Changing
Environment. CRC Press LLC.

90

Chapter 3 – Life History Influence on the Energy Allocation towards Growth,
Reproduction and Storage in Common Environment Raised Hatchery Lake Trout
Stocks subjected to Food Limitation
INTRODUCTION
Predicting how organisms can respond to stressors is very useful to any type of
management, and has been the focus of many studies (Frisch and Anderson 2000; Sloman
et al. 2001). However, in the past, individuals of a species have been managed as if all
populations respond in the same way to a stressor, despite intra-specific differences
shown in response to stressors (Schreck et al. 2001; Schreck 2010). Life history
adaptations have been thought to mediate responses to stressors, and indeed, studies on
species with different life histories suggest that this is true (Rose 2005; Spromberg and
Birge 2005). However, very rarely are populations within a species studied for their
response to stress. This study looks at how lake trout from different environments with
different life history strategies (Chapter 2) respond to food limitation.
Life history characteristics, such as age at maturity, size at maturity, and egg size
and number as a function of age are the evolutionary products of energetic tradeoffs that
that maximize lifetime reproductive success (Pianka 1976; Jennings and Beverton 1991;
Kozlowski 1996). Within a species, populations that face diverse environments can show
significant differences in life history characteristics and these differences can mediate
stress responses (Jennings and Beverton 1991; Shuter et al. 1998; Purchase and Brown
2001). In Chapters 1 and 2, I outlined how energy allocation can change throughout the
season and can be altered based on different pressures (mortality, density dependence,
etc.) and I assessed the status of field caught lake trout compared to hatchery lake trout.
In this chapter, I apply the conceptual model and information learned in Chapter 2 to
make predictions on how stocks explored in Chapter 2 would respond to a food limitation
91

stressor. I measured percent lipid, growth and reproductive output every two months
over the course of two years on individually identified female lake trout given different
rations, from four different lake trout stocks (Kingscote, Michipicoten, Killala and
Manitou – see Chapter 2) and determined if there were differences, and if such
differences could be predicted based on conceptual model built in Chapter 1 and Chapter
2.
Seasonal Energy Allocation
Iteroparous, long-lived fish from temperate regions do not allocate energy
equally to reproduction, growth and storage throughout the year; there tends to be
seasonal variation (Jokela 1997; Henderson and Wong 1998). Generally, during summer
months, metabolism and growth is high, due to increased prey availability and higher
temperatures (Chapter 1, Hill and Magnuson 1990). Iteroparous fish have also adapted
their timing of spawning events such that when the fry emerge, they will have ample food
available as presented in the match-mismatch hypothesis (Cushing 1969; Beauchamp et
al 2004). In lake trout, this occurs in the spring (Stauffer 1981; Marsden and Krueger
1991). During the winter months, growth and metabolism slows down considerably,
along with prey availability. Given these broad environmental constraints that are
universal to lake trout, I expected to see seasonal variation in energy allocation towards
growth, reproduction and lipid storage throughout the year, and that such allocations are
hard wired. Seasonal growth, lipid reserves, and survival have been linked to
reproduction in a variety of fish species (Heino and Kaitala 1999; Lambert and Dutil
2000; Hurst and Conover 2003). Therefore, when I measured lipids and growth on
individual lake trout every other month for two years, and determined reproductive
92

investment each year, I explored the hypothesis that individual fish will show differences
in growth and lipid allocation depending on where they are in gonadal growth phase, with
slower growth during period of gonadal growth and lipid deposition. More specifically,
from the energy allocation model in Chapter 1, I hypothesized that percent lipid would
increase until the spawn and would quickly decrease during the post-spawning season as
energy from these lipid reserves are moved towards increases in length from post-spawn
to the following summer. I expected every stock to show seasonal variation, but I
expected the Kingscote stock held in the hatchery to have the highest percent lipid during
the spawning season because of their low-energy diet/ high food conversion efficiency in
the wild and the excess food received in the hatchery. The piscivorous stocks held in the
hatchery were expected to follow similar seasonal patterns as the Kingscote, but have
lower percent lipids overall because they are adapted to a higher-energy diet in the wild
and have a lower food conversion efficiency.

Effects of Ration on Seasonal Growth and Percent Lipid
In this study I maintained four different lake trout stocks (Kingscote has two year
classes) in a common environment for two years and I manipulated food rations and
tracked seasonal lipid and growth and reproductive output. This study continued from a
previous study where half the individuals were maintained on a low food ration and half
were maintained on high food ration. For my study, I maintained a subset from each
stock on a high ration, and a subset on a low food ration. I also switched a subset of
individuals from each stock (except Killala because of limited supply of fish) from a
high ration to a low ration, and from a low ration to a high ration and followed these
“switchers” for two years to see how varying food abundance affects growth, lipid
93

storage and reproduction. I hypothesized that food ration will affect the maximum values
of percent lipid and growth rates, while not affecting the general seasonal patterns. I
expected individuals maintained on a high ration to have the highest lipid levels and
growth rates throughout the season, while the individuals maintained on a low ration to
have the lowest percent lipid and growth rates throughout the season, because of the
amount of available food. The fish that have switched rations are expected to have
intermediate percent lipids and growth relative to fish held at a constant ration (e.g.
High-Low fish will have lipids and growth levels higher than Low-Low individuals but
lower then High-High individuals). The effects of fish switching rations from a limited
ration to an increased ration was documented in Tian and Qin (2004), in which a
suspected increase in food-conversion efficiency allowed fish that were switched from a
Low to a High ration to surpass body weight of fish held at a high ration throughout the
study. Applying the concept of this observed compensatory growth to my study, I
expected that this increased ration in the Low-High treatment will allow fish from these
treatment to see increased growth and percent lipid over the Low-Low fish, and that these
values may approach or surpass the values of High-High fish, if two years is sufficient to
detect this hypothesized compensatory growth. In the case of the High-Low ration, it is
hypothesized that these individuals will show decreased growth relative to the High-High
ration because of energy intake and that these fish will have lipid levels and growth that
drop to near the levels of the Low-Low individuals because of the hypothesized decrease
in food conversion-efficiency created by exposure to high ration levels previously.

94

Stock-specific Percent Lipid, Growth, and Reproductive Output
I hypothesize that each stock will differ in magnitude on how they allocate energy
towards growth and lipids and reproductive output based on conceptual models presented
in Chapter 2 and information from the field. As summarized in the stock-specific
conceptual model in Chapter 2 (Table 5), I expect Manitou, a short-lived stock with high
lipids in the field, to have low lipid levels in the hatchery and have slow growth rates and
reproductive output. Adaptations related to the short lifespan of the Manitou stock were
expected to manifest in the hatchery individual as evidenced by prioritizing more energy
allocated towards growth and reproduction; I hypothesize this because they have adapted
to a short life span and want to maximize lifetime reproductive output before they die. In
the hatchery system, I expect them to be devoting all possible energy stores (lipids) to
reproduction and growth, and therefore they will have lower lipid levels than the other
stocks. I expected the Kingscote stock to have high lipid levels and growth rates in the
hatchery relative to other stocks because they are planktivorous in the wild and if adapted
to such a lifestyle, would presumably have higher energy-conversion efficiency (Table 5,
Chapter 2). This higher energy conversion efficiency should result in excess energy in
the hatchery to allocate towards lipids and growth. The Michipicoten stock was expected
to have moderate to fast growth rates and low reproductive output because of their longlived life history strategy in the field (Table 5, Chapter 2). Individuals from the
Michipicoten stock, on average, have more years to grow and reproduce, and tend to
maximize lifetime reproductive output later in life, so they are expected to have lower
energy allocated towards reproduction and more towards growth in the younger ages in
the hatchery. The Killala stock is piscivorous, but very slow growing in its natural

95

environment and is expected to have higher lipid levels in the hatchery than Michipicoten
and Manitou, but lower lipid levels relative to the Kingscote stock (Table 5, Chapter 2). I
expect that the stocks adapted to lower-energy diets in the wild (the Kingscote and the
Killala) will be the least sensitive to a change in ration while stocks that have abundant
food resources in their natural environment (Michipicoten and Manitou) will be more
sensitive to changes in ration because they are receiving less energy than that to which
they are adapted. The specific variables that show the highest sensitivity in each stock
will give insights into the relative priorities of each stock by showing which variables
they work to conserve while others are sacrificed. For example, the Kingscote have the
largest eggs and relatively smaller numbers of eggs. Furthermore, they do not show
indication of changing reproductive output over time (Chapter 2), so egg production may
be insensitive to ration.

Reaction Norms
I also wished to determine if energy allocation towards growth, reproduction and
percent lipid show relationships predicted by life history. This approach has implications
for management and I explored whether lipid storage would be predictive of reproductive
output. Such a relationship would have great utility, particularly if a non-invasive method
could be used to predict a usually lethal and labor intensive way of estimating fecundity.
I use the term “reaction norm”, which usually refers to the possible phenotypes that could
be expressed when an organism is exposed to a range of environmental conditions, and
also describes what changes occur in a phenotypic trait as environmental conditions, such
as food availability, change. (Stearns and Koella 1986; Barot et al. 2004). In this study, I
modify the term “reaction norm” to describe responses of a group (stocks of lake trout) to
96

a change in the environment, rather than evaluating the response of an individual across
varying environmental conditions. In this case, I use lipid measurement as a way to
measure the “environment” or food availability, and I explore whether lipid levels can
predict the number of eggs produced and if so, determine if the reaction norms are
specific to a particular stock.
To evaluate the potential for reaction norms, selected fish from each stock were
swapped from their initial ration level (high or low) to the alternate ration level to
determine if the relationship between lipids and fecundity changes when ration level is
manipulated. If a robust relationship is found to exist in this relationship between percent
lipid and egg number, this would provide a tool for managers to predict how a certain
stock’s reproductive output will change in a given environment, based on percent lipids
and associated food availability. I hypothesized that reaction norms would exist between
lipid levels and egg number for each stock within a given year across rations and that the
relationship would be less stable between years, particularly in short-lived stocks in
which annual changes in reproductive strategy are expected to be more pronounced
because they have less time to reach an age at which their reproductive output is
maximized
More specifically, I hypothesize that the Kingscote stock would have the least
sensitive relationship between lipids and egg number because they were overfed in the
hatchery study. I hypothesize that the Manitou and Michipicoten stocks would have the
most sensitive response to changes in ration because their percent lipid, and presumably
energy intake, most closely reflects natural conditions in the hatchery. The Michipicoten
stock was expected to have a sensitive relationship between lipid levels and egg number

97

because they are underfed in the hatchery relative to their natural energy intake. I
hypothesize that the Killala stock will have less sensitivity in the relationship between the
lipid levels and egg number than the Michipicoten and Manitou because the Killala stock
has relatively low lipid levels in the field, presumably due to competition for food
resources, and thus are also overfed in the hatchery (Table 5).
If these hypotheses regarding stock-specific energy allocation and reaction norms
are true, I am closer to predicting energy allocation towards growth and reproduction
based on population mortality and environment (available energy) using metrics such as
age and lipid measurements. Although the response of stocks in the hatchery to limiting
environments does not necessarily provide a perfect prediction on their performance in
the wild, these stock-specific responses of individuals to food limitation in a hatchery
setting can be used as a starting point for general predictions on how these fish would
respond to a new environment.
Table 14. Summary of hypotheses regarding responses of five lake trout stocks to
common environment experiment.
Hypothesis
Description
Individuals will differentially allocate energy towards growth and
Hypothesis #1 reproduction throughout the season based on where they are in their
gonadal growth phase.
Stocks will show significant differences in responses of percent lipid,
Hypothesis #2 growth, and reproductive output to ration based on Table 5 in Chapter
2.
Reaction norms exist between percent lipid and egg number within
Hypothesis #3 each stock within a year in the Michipicoten, Killala, and Manitou
stocks.
Reaction norms exist between years in the long-lived Michipicoten
Hypothesis #4
and Killala stocks.

98

METHODS
Hatchery Fish used in this Study
Lake trout used in the study were descendants of fish brought into the hatchery as
eyed eggs from wild collections performed by the Ontario Ministry of Natural Resources
(OMNR). The fish used in this study were from gametes spawned in 1999 (Kingscote
and Manitou) or 2000 (Michipicoten, Killala, and Kingscote). The Killala, Manitou, and
Michipicoten individuals used in this study were two generations removed from the wild.
Both year classes of the Kingscote stock used in this study were first generation hatchery
fish. The fish were reared at the OMNR fish culture station in Codrington, Ontario that
has high quality water from a nearby stream. Common conditions in the hatchery
included water temperature, oxygen, feeding schedule and ration (0.6% of total tank
biomass), size of holding tanks, and tank densities that were regularly adjusted for
minimal mortality rates of approximately 1-2 fish per tank per year. Males from each
stock and individuals from the 2000 Slate stock were housed in experimental tanks to
ensure that tank densities were similar at the start of the study, and so that females were
exposed to potential male priming pheromones (Stacey et al 2003) (Table 15). Prior to
sexual maturity in 2006, the individual stocks and year classes were kept in separate
holding tanks. Once they reached sexual maturity, they were branded to allow for
individual identification and transferred to eight 1,200-liter circular tanks and stocks were
mixed together (Table 15).
Treatments
From 2006 to 2007, ration level was manipulated for a study in which half the
fish received a low ration level (0.2% tank biomass) and the other half of the fish
99

received a high ration level (0.4% tank biomass) (Murphy et al in prep). From 2006 to
2008, individual fish were monitored for size, lipid content and reproductive output.
Then, at the beginning of this study in February 2008, the stocks were divided into 4
treatment groups with equal density (Table 15). One group from each stock was
randomly selected to be maintained at a low ration (0.3% biomass) and another group
from each stock was randomly selected to be maintained at a high ration (0.5% biomass).
An additional group was also randomly selected from each stock (except for Killala
because there were not enough fish) to be switched from a low ration feeding regime
(0.2% biomass) to the high ration feeding regime (0.5% tank biomass). The last group
was randomly selected to be switched from a high ration feeding regime to the low ration
feeding regime. The specific rations in the low and high treatments were changed
between studies to maintain health. A minimum of 16 fish from each stock were
maintained at an original ration (high or low), while the remaining fish were moved to the
alternate ration level (Table 15). These rations and tank configurations were maintained
for the entire duration of this study. Manipulation of ration level resulted in four
treatments; the terminology below will be used throughout Chapter 3:
1) High-High = Fish that were given the high ration (0.4%) prior to this study and
were also given the high ration (0.5%) during the current study (2008-2010).
2) High-Low = Fish that were given the high ration (0.4%) prior to this study and
were given the low ration (0.3%) during the current study (2008-2010).
3) Low – Low = Fish that were given the low (0.2%) ration prior to this study and
were also given the low ration (0.3%) during the current study (2008-2010).

100

4) Low – High = Fish that were given the low ration (0.2%) prior to this study and
were given the high ration (0.5%) during the current study (2008-2010).

101

Table 15. Summary of initial fish densities and composition of lake trout stocks housed in a common environment experiment in
OMNR hatchery in Codrington, Ontario.
Tank
No.

Ration

Total
Fish

1
2
3
4
5
6
7
8

High-High
High-Low
Low-Low
Low-High
High-High
High-Low
Low-Low
Low-High

30
31
28
28
29
30
30
30

2000 Kingscote

2000
Michipicoten

2000
Killala

Male

Female

Male

Female

Male

Female

1
2
1
2

8
7
8
8

1
4
2
2

10
8
7
8

1
0
1
1

9
2
9
0

2000
Slate

1999
Manitou

Female

2

6

2

Male

Female

Male

Female

2
3
2
2

102

Male

1999
Kingscote

5
12
11
12
11

3
6
2
4

12
10
14
13

Groups of fish were switched to the alternate ration to simulate variability in
allocation patterns and to determine if patterns in percent lipid, growth and reproductive
output remained similar to patterns shown in low and high rations.
Measurements
Every 8 weeks, fish were briefly anaesthetized using MS-222 at a low level
sufficient to minimize fish stress during handling. Once fish were anesthetized, I
measured length, weight and percent lipid. A non-invasive fat meter was used to measure
percent lipid in individuals at four locations across the body and get the average lipid
level between the four locations (Figure 29). The fat meter uses low power microwaves
to penetrate the skin and measure the attenuation of the microwaves by water in the flesh,
which is negatively correlated with fat content, allowing for a reliable lipid estimate
(www.distell.com). This relatively new technology allowed me to track lipid levels of
individuals through time which is not possible using more traditional, lethal methods of
2

lipid analysis. The fat meter was calibrated in a previous study (r = 0.6117; Claus 2011),
and the reading corresponded to an actual lipid level percentage using the following
equation:
(1.9436 + ln (Fat meter Reading)) – 9.5022

] x 100

Muscle Lipid Concentration = [e

103

Figure 29. Locations of lipid measurements taken on lake trout using non-invasive fat
meter at OMNR hatchery in Codrington, Ontario (Photograph Courtesy of Sara Smith).
During the spawning season (approximately early September through early
November) of 2008 and 2009, females were stripped of eggs when ovulated to determine
the total volume of eggs spawned and when fish were ripe. Ovulation of eggs was
assessed weekly during the Fall seasons by gently applying pressure to the urogenital
region. If eggs were easily extruded, the female was stripped and the entire volume of
eggs was recorded. A 30-mL sample of eggs was then separated and stored in a 5%
formalin solution to enumerate and measure at a future date. The total number of eggs in
the 30-ml sample was counted and the number was scaled up to the total volume of the
spawn to estimate total egg production for each female. Also, thirty eggs were randomly
selected from each 30-mL sample to determine egg size by photographing each egg using
a digital microscope and measurement software (Leica Application Suite®). The
diameter of each egg was determined by averaging six measurements that were taken
along each egg. Multiple measurements were used to account for irregularities in egg
size that occurred after preservation.
Data Analyses
Percent lipid, growth and reproductive output were calculated. The percent lipid
from each location on the fish (Figure 29) was averaged to calculate a mean percent
104

muscle lipid for each fish. Growth was calculated as the change in fork length (in
millimeters) between a given sampling date and the previous sampling date. In some
cases, growth between sampling dates was a small negative value; these data were
changed to a value of zero.
The number of eggs per female was standardized by dividing by the fork length to
get a value in eggs/mm because of the strong relationship between spawner length and
egg number (Chapter 2).
The daily growth rate and percent lipid were calculated for all rations within each
stock and within each year; the significance of differences in percent lipid between July
and spawning season were calculated using a repeated measures analysis. The daily
growth rate was calculated as the average daily increase in fork length (in mm) that
occurred between sampling dates. In the case of the growth rates going into the spawn,
the average annual spawning date for each stock was used as the comparison date for
growth calculations. I compared the growth rate (between December and the proceeding
spawn) within stock between rations, within year using an ANOVA with Tukey-Kramer
post-hoc comparisons. I also compared the mean percent lipid (spawning time percent
lipid), average egg size, and number of eggs/mm within stocks between rations and
within year using an ANOVA with Tukey-Kramer post-hoc comparison.
Paired t-tests were used to compare egg size and egg number values between
hatchery years in the combined-ration data within each stock. The same method was
used to determine the significance of ration-specific changes in egg size and egg number
between years.

105

I applied Akaike’s Information Criterion (AIC) to choose the model of best fit to
describe the relationship between eggs/mm and mean percent lipid within each stock.
The following types of relationships were analyzed to determine the model of best fit:
Michaelis-Menten, Sigmoidal, Hyperbolic, Linear, and Logarithmic. The chosen model
was the model with the lowest AIC value. Models were generated for each stock within
each year (all rations combined) and additionally within each year on a ration-specific
basis. In the case of the ration-specific analysis, specific regressions were applied to data
from those fish whose rations which switched ration level (“switchers”, i.e. High-Low
and Low-High) and the fish that remained at a constant ration throughout the study
(“non-switchers” i.e. High-High and Low-Low).
The parameter values in each regression, or model, applied to the data were
estimated using the regression wizard in Sigmaplot 12.0 and SAS 9.0. To determine if
reaction norms could potentially exist, I compared the AIC values between the model of
best fit with an alternate model which was generated using data from the alternate year or
ration group. Differences in AIC values were calculated between the model of best fit
and alternate model to determine if there is substantial support to suggest the initial
model is a better predictive model than the alternate model. Similar AIC values suggest
the existence of reaction norms because a change in ration or year does not change the
relationship between egg number and mean percent lipid. Specifically, for each stock, I
applied the model of best fit from Year 1’s data to the data in Year 2, and vice versa, and
compared the AIC values between years based on this common regression that was
applied to each year’s data. An identical procedure was completed to compare ration
groups, or the switchers and non-switchers. Within each stock, the regression assigned to

106

the non-switchers was applied to the data in the switchers, and vice versa, to determine
the difference in AIC values.
Although the comparison of differences in AIC values is not a significance test,
these comparisons can provide substantial support for the initial model being the best
predictor model over the alternate model (Anderson et al. 2001; Bence and Hayes 2004).
While there are varying approaches in the evaluation of AIC difference, a general
guideline proposed in Burnham and Anderson (2002) is that models with a difference of
AIC less than or equal to two (AICInitial – AICAlternate ≤ 2) have strong support to show
there is no difference between the alternate model the model of best fit in terms of their
ability to describe the data (Burnham and Anderson 2002; Stylianou et al. 2013). AIC
differences in the range of 4 to 7 (4 ≤ AICInitial – AICAlternate ≤ 7) have considerably
less support to show the alternate model equals the best model in its predictive ability
(Burnham and Anderson 2002). Lastly, AIC differences of greater than 10 (10 ≤
AICInitial – AICAlternate) have essentially no support to show that the alternate model fits
the data as well as the original model (Burnham and Anderson 2002). Although there is
no consensus on these cutoff values to definitively say when a model outperforms
another, a conservative value was put forth in Stylianou et al. 2013 which states models
must have a difference of 6 in AIC values to say that they differ. In this case, based on
the general guidelines put forth in Burnham and Anderson 2002 and the specific cutoff
used in Stylianou et al. 2013, I assigned a minimum AIC difference of 6 to say that two
model differ in their predictive ability.

107

RESULTS
Seasonal growth across sampling dates for the two years was variable, but every
stock showed decreases in growth during the spawning season followed by increases in
growth rates after the spawn, with the highest growth rates observed in July in Year 1 and
May in Year 2 (Figure 30 - Figure 35). Michipicoten showed the highest growth rates,
followed by Kingscote, then Killala and finally Manitou (Figure 30 - Figure 35). Growth
rates between rations showed a consistent pattern between stocks. For both years, fish
maintained at the highest food ration had the highest growth rates, while the fish
maintained at lowest ration had the lowest growth rates (Figure 30 - Figure 35). In Year 1
of the study, the switchers had intermediate growth rates, with the fish switched from the
high to the low ration growing faster than the fish switched from the low to high ration,
although these differences were not significant (Figure 35). By Year 2, the switchers had
recovered growth rates to match the fish in their new ration group (Figure 35, Table 22,
and Table 23). Growth rates exceeded the High-High ration, but not significantly so, in
the Michipicoten fish that were switched from the low to high ration by Year 2.

108

Figure 30. Change in growth rate (in mm/day) of Michipicoten stock housed in a common environment experiment exposed to one of
four feeding regimes in OMNR hatchery.

109

Figure 31. Change in growth rate (in mm/day) of Killala stock housed in a common environment experiment exposed to one of four
feeding regimes in OMNR hatchery.

110

Figure 32. Change in growth rate (in mm/day) of the 1999 Kingscote stock housed in a common environment experiment exposed to
one of four feeding regimes in OMNR hatchery.

111

Figure 33. Change in growth rate (in mm/day) of the 2000 Kingscote stock housed in a common environment experiment exposed to
one of four feeding regimes in OMNR hatchery.

112

Figure 34. Change in growth rate (in mm/day) of the Manitou stock housed in a common environment experiment exposed to one of
four feeding regimes in OMNR hatchery.

113

Figure 35. Growth rate of five stocks of lake trout housed in a common environment
exposed to one of four feeding rations. Standard error is shown above each bar.
Statistical groupings within a stock, based on an ANOVA pairwise comparison with a
Tukey-Kramer adjustment are shown above each bar.
114

General trends in the seasonal percent lipid data of Year 1 show increases in
percent lipid level up until the spawn, after which percent lipid decreases into the spring
(Figure 36– Figure 40). Year 2 of the seasonal percent lipid data showed a much more
constant percent lipid level within a ration of a given stock throughout the entire season
(Figure 36– Figure 40). Almost all the ration-specific comparisons showed a significant
lipid increase in Year 1 of the hatchery study; only 5 were not significant (Table 20).
Year 2 of the hatchery study was different, however, with no significant increase
observed in lipids during the course of the year in when rations were combined, and only
the High-High ration of the Killala comparisons was significant, out of 18 total
comparisons (Table 21).
Kingscote (both year classes) had the highest percent lipid overall (Figure 41),
followed by the Killala, then Michipicoten and finally the Manitou, and this order was
maintained over both years. The fish that were maintained on a low ration had decreased
percent lipid in all stocks but the reduction was smallest in the Kingscote stocks and the
same trend was seen in both years. The stocks responded differently to the switched
ration. In the first year, lipid levels in the Kingscotes that were switched from low-high
rebounded to levels observed in the high ration, whereas it took until the second year for
the Michipicoten and Manitou in the low-high ration to rebound to the same level as the
High-High ration. The fish switched from high-low ration did show a significant
decrease in lipid level to match the fish maintained on a low ration in Year 1 of the 2000
Kingscote stock, Year 2 of the 1999 Kingscote stock, Year 2 of the Manitou stock, and
Year 2 of the Michipicoten stock (Figure 37, Table 20, Table 21).

115

Figure 36. Change in mean percent lipid of Michipicoten lake trout housed in a common environment experiment exposed to one of
four feeding regimes in OMNR hatchery.

116

Figure 37. Change in mean percent lipid of Killala lake trout housed in a common environment experiment exposed to one of two
feeding regimes in OMNR hatchery.

117

Figure 38. Change in mean percent lipid of 1999 Kingscote lake trout housed in a common environment experiment exposed to one
of four feeding regimes in OMNR hatchery.

118

Figure 39. Change in mean percent lipid of 2000 Kingscote lake trout housed in a common environment experiment exposed to one
of four feeding regimes in OMNR hatchery.

119

Figure 40. Change in mean percent lipid of Manitou lake trout housed in a common environment experiment exposed to one of four
feeding regimes in OMNR hatchery.

120

Figure 41. Mean spawning-time lipid levels of five stocks of lake trout housed in a
common environment exposed to one of four feeding regimes. Standard error is shown
above each bar. Statistical groupings within a stock, based on an ANOVA pairwise
comparison with a Tukey-Kramer adjustment are shown above each bar.
121

Ration-specific egg size did not follow any clear trend within a stock in either
year. The only clear trend that was shown was a consistent decrease in egg size between
years within a stock in a given ration (Figure 42). Comparison of egg size across stocks
within rations showed that, in general, the Killala stock had the largest eggs, followed
closely by the Manitou stock. In all cases but the Low-High ration of Year 1, the
Kingscotes had the smallest eggs across rations. The Michipicoten egg size decreased in
the low-high ration both years, but this decrease was only significant in Year 1.

122

Figure 42. Mean egg diameter of five stocks of lake trout housed in a common
environment exposed to one of four feeding regimes in OMNR hatchery. Standard error
is shown above each bar. Statistical groupings within a stock, based on an ANOVA
pairwise comparison with a Tukey-Kramer adjustment are shown above each bar.
123

Ration-specific egg number was more variable than egg size as a function of
ration. Generally, as the ration was increased, so was the egg number with the
“switchers” having intermediate values, but there are small deviations from that pattern in
most stocks (Figure 43). The 2000 Kingscote stock had the highest egg number generally.
Like growth and percent lipid, the Low-High stock showed a rebound in egg number
between years in all stocks. Comparison of egg number across stocks within rations
suggested that the Kingscote was the most resilient to low food ration and had the lowest
response to reductions in egg number in both years. The Killala stock significantly
reduced their egg production in the low food ration. The Manitou and Michipicoten fish
switched from a low to high ration in the first year had similar egg production to those
maintained in the low ration, while the fish from these stocks switched from the high to
low ration still produced eggs at similar quantity to the fish maintained on the high ration.
However, the switchers from the Manitou and Michipicoten stocks diverged in Year 2:
the switchers from the Michipicoten stocks mimicked egg production from the fish in
their switched ration, while the Manitou fish that were switched from high to a low ration
still maintained high egg production (Figure 43).

124

Figure 43. Egg number (corrected for length of spawning female) of five stocks of lake
trout housed in a common environment exposed to one of four feeding regimes. Standard
error is shown above each bar. Statistical groupings within a stock, based on an ANOVA
pairwise comparison with a Tukey-Kramer adjustment are shown above each bar.
125

Effects of Age on Reproductive Output
There was a negative correlation between egg size and corrected egg number
(eggs/mm) in all stocks (Figure 44 - Figure 48). The decrease in egg size between 2008
and 2009 was significant in the Manitou, 2000 Kingscote, 1999 Kingscote, and Killala
stocks (Table 16 and Figure 48). Changes in egg size between the two years in the
Michipicoten stocks was not significant (p = 0.4046). The increases in egg number
(eggs/mm) between 2008 and 2009 were significant only in the 2000 Kingscote and
Manitou stocks (Figure 44 - Figure 48, Table 17). Overall, 3 of the 4 stocks (Kingscote,
Killala, and Manitou) showed significant changes in egg size between the two years in
the ration-combined data while two of the stocks (Manitou and 2000 Kingscote) showed
significant changes in egg number.
Analysis of ration-specific changes in egg size between years showed significant
decreases in 4 of 5 stocks in the High-High treatment: 2000 Kingscote, 1999 Kingscote,
Manitou, and Michipicoten (Table 16). In the High-Low treatment, two of the four
stocks showed significant decreases in egg size: 2000 Kingscote and Manitou. Three of
five stocks showed significant decreases in egg size in the Low-Low treatment: Killala,
2000 Kingscote, and Manitou. Lastly, only the Manitou stock showed significant
decreases in egg size in the Low-High treatment. Overall, 10 of the 18 possible rationspecific changes in egg size between years were significant.
Ration-specific repeated measures analysis of changes in egg number between
years showed positive increases in egg number (eggs/mm) in the Low-High rations in
three of the four stocks analyzed: Manitou, 1999 Kingscote, and 2000 Kingscote (Table
17). The other three rations showed no significant differences in egg number. Overall, 3
of the 18 possible ration-specific changes in egg number between years were significant.
126

Figure 44. Relationship between egg number as measured in eggs/mm and egg size of
Michipicoten lake trout housed under multiple feeding regimes in a common garden
environment in OMNR hatchery during 2008 spawning season. P values show the
significance of changes between Year 1 and Year 2 in eggs/mm and egg size.

Figure 45. Relationship between egg number as measured in eggs/mm and egg size of
Killala lake trout housed under multiple feeding regimes in a common garden
environment in OMNR hatchery during 2008 spawning season. P values show the
significance of changes between Year 1 and Year 2 in eggs/mm and egg size.
127

Figure 46. Relationship between egg number as measured in eggs/mm and egg size of
1999 Kingscote lake trout housed under multiple feeding regimes in a common garden
environment in OMNR hatchery during 2008 spawning season. P values show the
significance of changes between Year 1 and Year 2 in eggs/mm and egg size.

Figure 47. Relationship between egg number as measured in eggs/mm and egg size of
2000 Kingscote lake trout housed under multiple feeding regimes in a common garden
environment in OMNR hatchery during 2008 spawning season. P values show the
significance of changes between Year 1 and Year 2 in eggs/mm and egg size.
128

Figure 48. Relationship between egg number as measured in eggs/mm and egg size of
Manitou lake trout housed under multiple feeding regimes in a common garden
environment in OMNR hatchery during 2008 spawning season. P values show the
significance of changes between Year 1 and Year 2 in eggs/mm and egg size.

129

Table 16. P-values (* = p<0.05, ** = p<0.01; + or – indicates direction of significant
changes) for repeated measures analysis for changes in egg size that occurred between
the spawning seasons of 2008 and 2009 of individuals from the Killala, 2000 Kingscote,
1999 Kingscote, Manitou, and Michipicoten stocks of lake trout held in a common
environment under multiple feeding regimes (High = 0.5% of total tank biomass fed to
tank daily; High-Low indicates a switch from a high [0.5%] to a low [0.3%] ration level
at the start of the study).
Stock

All Rations

High-High

High-Low

Low-Low

Low-High

Killala

<0.0005**(-)

0.0665

n/a

0.013* (-)

n/a

2000 Kingscote

<0.0001**(-)

0.0032**(-)

0.0047** (-)

0.0022** (-)

0.0511

1999 Kingscote

0.0059** (-)

0.0152* (-)

0.2748

0.0763

0.5015

Manitou

<.0001** (-)

.0001** (-)

0.0114* (-)

0.0006** (-)

.0116* (-)

0.4046

0.0185* (-)

0.2029

0.2819

0.3259

Michipicoten

Table 17. P-values (* = p<0.05, ** = p<0.01; + or – indicates direction of significant
changes) for repeated measures analysis for changes in egg number that occurred
between the spawning seasons of 2008 and 2009 of individuals from the Killala, 2000
Kingscote, 1999 Kingscote, Manitou, and Michipicoten stocks of lake trout held in a
common environment under multiple feeding regimes (High = 0.5% of total tank biomass
fed to tank daily; High-Low indicates a switch from a high [0.5%] to a low [0.3%] ration
level at the start of the study).
Stock

All Rations

High-High

High-Low

Low-Low

Low-High

0.724

0.5579

n/a

0.4571

n/a

2000 Kingscote

0.0023** (+)

0.0725

0.0886

0.8385

0.0184* (+)

1999 Kingscote

0.6481

0.8301

0.6074

0.0695

0.0311* (+)

0.0109* (+)

0.1617

0.2438

0.8363

0.0017** (+)

0.9249

0.6872

0.156

0.3652

0.1023

Killala

Manitou
Michipicoten

130

Reaction Norms
The modeled relationship between egg number and mean percent lipid, based on
AIC values, for all stocks and for both years showed a general increasing trend in egg
number with mean percent lipid (Figure 49 - Figure 53). Unique regressions for ration
groups (non-switchers vs. switchers) were plotted to visually show differences between
the models assigned to each ration group within each stock and study year (Figure 54 Figure 57). Each stock showed a different shape and this shape differed between years,
with the relationship getting steeper in Year 2 (except for the Michipicoten stock which
had a similar curve as Year 1). The Michipicoten relationship showed a linear, gently
increasing relationship between percent lipid and eggs/mm in year 1 and a hyperbolic
increase in Year 2. The Killala stock showed a gently increasing hyperbolic curve in
Year 1 which slowed at higher lipid levels. However, the Year 2 function in the Killala
stock showed an increasing linear trend with a higher rate of increase at higher percent
lipids than in Year 1. The 1999 Kingscote stock showed a sigmoidal relationship
between egg production and percent lipid and the sharp increase in egg production at low
percent lipid is likely a result of two highly influential data points at low percent lipids.
In Year 2, the Kingscote stock showed an increasing linear trend in egg number with
percent lipid. The 2000 Kingscote stock had a similar pattern as the 1999 Kingscote,
with a relatively flat linear relationship in Year 1 and a steeper linear increase in Year 2.
Lastly, the Manitou stock showed a shallow hyperbolic increase in egg production with
increasing percent lipid in Year 1 with the rate of increase slowing at higher percent
lipids. In Year 2, the linear function applied to the data was steeper than in the previous
year (Figure 49 - Figure 53, Table 24, Table 25, Table 26, Table 27, and Table 28).

131

Comparison of original and alternate models using an AIC difference cutoff of 6,
revealed that differences between models in the comparison across years, in the
Michipicoten and Killala stocks did not provide substantial evidence support to suggest
the original model fits better than the model from the alternate year. These data suggest
that the relationship between egg number and mean percent lipid are relatively stable
across years in the Killala and Michipicoten stocks (Table 18). Conversely, AIC
differences greater than 6 in the 1999 Kingscote, 2000 Kingscote, and Manitou stocks
provide strong evidence to suggest that the original model is better than the alternate
model and that the relationship between egg number and mean percent lipid changed
between years. Across all stocks, 6 of the 10 (60%) total year-specific comparisons
provided strong evidence to suggest the original model better describes the data than the
alternate.
Comparisons of original and alternate models in the ration groups in Year 1
revealed that four of the eight (50%) total comparisons across stocks that exceeded an
AIC difference of 6. The following stocks (and ration group within that stock) had high
enough AIC differences to suggest a difference in the predictive ability of the original
and null models: 1999 Kingscote (non-switchers and switchers), 2000 Kingscote (nonswitchers), and Manitou (non-switchers). In these cases, robust relationships were not
observed between egg number and mean percent lipid in the comparisons across the
ration groups within a given year.
Comparisons of original and alternate models between ration-groups in Year 2
revealed that two of the eight (25%) total comparisons had AIC differences that exceeded
the cutoff of 6. The following stocks (and ration group) showed differences in the

132

predictive ability of the original and alternate model: 1999 Kingscote (switcher) and 2000
Kingscote (non-switcher). In these two cases, there was not a robust relationship
between egg number and mean percent lipid between the ration groups. Across both
years, 6 of the 16 (37.5%) total ration-group comparisons provided strong evidence to
suggest the original model better describes the data than the alternate.

133

Figure 49. Relationships between percent lipid and egg number (in eggs/mm) of Michipicoten lake trout housed in a common
environment hatchery exposed to one of four feeding regimes. Parameters used in each function included on each graph. Functions
for each stock and year were selected using AIC values of linear, hyperbolic, logarithmic, Michaelis-Menten and sigmoidal function.

134

Figure 50, Relationships between percent lipid and egg number (in eggs/mm) of Killala lake trout housed in a common environment
hatchery exposed to one of four feeding regimes. Parameters used in each function included on each graph. Functions for each stock
and year were selected using AIC values of linear, hyperbolic, logarithmic, Michaelis-Menten and sigmoidal function.

135

Figure 51. Relationships between percent lipid and egg number (in eggs/mm) of 1999 Kingscote lake trout housed in a common
environment hatchery exposed to one of four feeding regimes. Parameters used in each function included on each graph. Functions
for each stock and year were selected using AIC values of linear, hyperbolic, logarithmic, Michaelis-Menten and sigmoidal function.

136

Figure 52. Relationships between percent lipid and egg number (in eggs/mm) of 2000 Kingscote lake trout housed in a common
environment hatchery exposed to one of four feeding regimes. Parameters used in each function included on each graph. Functions
for each stock and year were selected using AIC values of linear, hyperbolic, logarithmic, Michaelis-Menten and sigmoidal function.

137

Figure 53. Relationships between percent lipid and egg number (in eggs/mm) of Manitou lake trout stocks housed in a common
environment hatchery exposed to one of four feeding regimes. Parameters used in each function included on each graph. Functions
for each stock and year were selected using AIC values of linear, hyperbolic, logarithmic, Michaelis-Menten and sigmoidal function.

138

Figure 54. Relationships between percent lipid and egg number (in eggs/mm) of Michipicoten lake trout housed in a common
environment hatchery exposed to one of four feeding regimes. Parameters used in each function included on each graph. Functions
for each stock and year were selected using AIC values of linear, hyperbolic, logarithmic, Michaelis-Menten and sigmoidal function.
Ration-specific regressions are grouped by non-switchers (High-High and Low-Low) and switchers (High-Low and Low-High). The
solid line represents the regression applied to non-switchers, the dotted line shows the regression applied to switchers.

139

Figure 55. Relationships between percent lipid and egg number (in eggs/mm) of 1999 Kingscote lake trout housed in a common
environment hatchery exposed to one of four feeding regimes. Parameters used in each function included on each graph. Functions
for each stock and year were selected using AIC values of linear, hyperbolic, logarithmic, Michaelis-Menten and sigmoidal function.
Ration-specific regressions are grouped by non-switchers (High-High and Low-Low) and switchers (High-Low and Low-High). The
solid line represents the regression applied to non-switchers, the dotted line shows the regression applied to switchers.

140

Figure 56. Relationships between percent lipid and egg number (in eggs/mm) of 2000 Kingscote lake trout housed in a common
environment hatchery exposed to one of four feeding regimes. Parameters used in each function included on each graph. Functions
for each stock and year were selected using AIC values of linear, hyperbolic, logarithmic, Michaelis-Menten and sigmoidal function.
Ration-specific regressions are grouped by non-switchers (High-High and Low-Low) and switchers (High-Low and Low-High). The
solid line represents the regression applied to non-switchers, the dotted line shows the regression applied to switchers.

141

Figure 57. Relationships between percent lipid and egg number (in eggs/mm) of Manitou lake trout housed in a common environment
hatchery exposed to one of four feeding regimes. Parameters used in each function included on each graph. Functions for each stock
and year were selected using AIC values of linear, hyperbolic, logarithmic, Michaelis-Menten and sigmoidal function. Ration-specific
regressions are grouped by non-switchers (High-High and Low-Low) and switchers (High-Low and Low-High). The solid line
represents the regression applied to non-switchers, the dotted line shows the regression applied to switchers.

142

Table 18. AIC differences between model of best fit (assigned based on AIC values) and
an alternate model describing the relationship between egg number and mean percent
lipid of five lake trout stocks housed in OMNR hatchery. Alternate model is model of
best fit applied to alternate data group (year or feeding group) to determine how much
better a fit the original model is than the model used to describe the relationship in the
alternate ration group or study year. Bolded values indicate AIC differences that exceed
the threshold of 6, which indicates a difference in the predictive ability between the
original and alternate model for that given group.
Stock
Michipicoten
Killala
1999 Kingscote
2000 Kingscote
Manitou

Year 1 Data AIC Difference
3.1112
2.2401
14.0221
38.9710
24.4733

Year 2 Data AIC Difference
1.2281
1.2011
9.1408
20.4130
13.6536

Stock

Year 1 Non-Switcher Data AIC
Difference
3.2226
18.7383
17.3210
7.0939

Year 1 Switcher Data AIC
Difference
5.7248
14.2875
5.1830
5.6611

Year 2 Non-Switcher Data AIC
Difference
4.1306
4.8189
7.4080
3.7698

Year 2 Switcher Data AIC
Difference
1.01794
11.1447
1.3258
3.6117

Michipicoten
1999 Kingscote
2000 Kingscote
Manitou
Stock
Michipicoten
1999 Kingscote
2000 Kingscote
Manitou

143

DISCUSSION
Life history has a subtle influence on growth, reproductive output and lipid
storage in four different stocks of lake trout, and these influences were observable when I
imposed a food limitation and changed the feeding regime. Overall, the most important
findings from this study was that I found the long-lived Michipicoten stock maintained
high levels of growth following a reduction in ration, while the short-lived Manitou stock
maintained egg production within a ration across years, at the expense of growth and
percent lipid. This provided evidence to support the initial hypothesis of stock-specific
energy allocation and stock-specific effects in the response of energy allocation to ration
(Table 19) and also suggests that the Michipicotens prioritize energy allocation to growth
while the Manitous prioritize energy allocation to reproduction (Chapter 2). The
Michipicoten stock also will shift their energy allocation between egg size and egg
number when switched from a low to high feeding regime. Robust reaction norms were
found to exist in the majority (62.5%) of relationships between lipid levels and egg
number within a given year across rations in each stock while these relationships were
less predictable across years, as evidence by only 40% of the relationships remaining the
same between years within each stock in the combined-ration data. These data provided
evidence to support the hypothesis of inter-annual reaction norms within the long-lived
stocks and inter-ration reaction norms within each year for each stock (Table 19). Lastly,
seasonality was observed in percent lipid and growth rates in each stock which provides
evidence to support the initial hypothesis that predicted seasonality in energy allocation,
depending on proximity to the spawning season (Table 19). However, percent lipids
remained high throughout the summer right up until the spawn, which was unexpected.
144

Table 19. Summary of results found and how they support (or fail to support) initial
hypotheses regarding energy allocation of lake trout held in a common environment
exposed to one of four feeding regimes.
Hypothesis
Description
Result
Supported by data somewhat
Individuals will differentially allocate
– seasonality in lipids and
Hypothesis energy towards growth and reproduction
growth was observed but did
#1
throughout the season based on where they
not follow exact pattern
are in their gonadal growth phase.
predicted
Stocks will show significant differences in Supported by data
Hypothesis
responses of percent lipid, growth, and
#2
reproductive output to ration.
Reaction norms exist between percent lipid Supported by data somewhat
Hypothesis and egg number within each stock within a (in 62.5 % of cases)
#3
year in the Michipicoten, Killala, and
Manitou stocks.
Hypothesis Reaction norms exist between years in the
Supported by data
#4
long-lived Michipicoten and Killala stocks.
Growth, Percent Lipid and Reproductive Output: General Ration-specific
Responses
Trends in growth rate and mean percent lipid generally followed the ration group.
The individuals switched from the Low to High rations showed a rapid rebound in growth
and mean percent lipid to levels that were similar to those maintained on high rations
This rapid rebounding of lipids and growth has been documented in other studies that
show periods of increased growth and percent lipid increases relative to fish maintained
at a high ration, presumably due to increased food conversion efficiency following a
period of food deprivation (Tian and Qin 2004; Alvarez and Nicieza 2005). In contrast,
individuals switched from High to Low rations showed just slight reductions in percent
lipid in Year 2 relative to the High-High ration but the differences were still not
significant in any stock. This suggests that there is a tendency to maintain lipid reserves
once they are established, despite a reduction in ration. This result is inconsistent with
previous studies that suggest lipid levels will be rapidly depleted in fish exposed to a
reduction in ration, particularly during the winter when resources are most limiting
145

(Thompson et al. 2001; Biro et al. 2004). However, it is possible that the decreased food
ration did not emulate winter conditions adequately. These lipid reserves are critical for
survival, particularly during the winter months (Thompson et al. 2001; Biro et al. 2004).
Egg size was relatively fixed within a given stock, providing evidence that egg
size is the least sensitive variable to ration. This is supported by a previous study
conducted on lake trout stocks that document the predictability of egg size within stocks
across different environments (Jastrebski and Morbey 2009; Myers et al. 2011).
Egg number was more sensitive to ration, as was growth and lipid, but egg
number did not rebound as quickly when switched from low to high ration. The egg
number of the Manitou stock is relatively resilient to a decrease in ration while the
Michipicoten stock’s egg number has the most sensitive response to a ration change,
suggesting that the short lived stock prioritized reproductive output. Food limitation has
been shown to influence only egg number and not egg size in other salmonids such as
rainbow trout (Scott 1962). Scott (1962) showed that starvation was highly associated
with the degree of follicular atresia and the breakdown of ovarian follicles.
Effects of Ration on Growth, Percent Lipid, and Reproductive Output: Stock
Differences
Intraspecific differences in energy allocation and responses to food limitation in
common environments have been shown in numerous previous studies and have been
hypothesized to reflect differences in energy intake, metabolic costs, season, age, and
reproductive status of stocks (Yibo and Jianking 1990; Hutchings et al. 1999; Robards et
al. 1999). In this study, the long-lived stocks were most sensitive to a ration change as
manifested in growth rates and percent lipids while the egg production of the long-lived
stocks was less sensitive to ration.
146

Stock-Specific Conceptual Model Evaluation
In general, the lake trout stocks followed predictions of energy allocation from
the conceptual models in Chapter 1 and Chapter 2 that made predictions regarding the
effects of stock lifespan, environmental conditions (fishing pressure, sea lamprey
mortality, etc.), and food quality and quantity on energy allocation (Table 5). One
deviation from my conceptual model was the lack of sensitivity of percent lipid of the
Kingscotes in the hatchery. Based on my conceptual model, the storage of energy in lipid
reserves is the lowest priority for the Kingscotes which should have been reflected in
highly variable lipid levels between rations. The other deviation from the conceptual
models was the sensitivity of growth to a reduction in ration in the Manitou stock which
suggests a decreased emphasis on growth in limiting environments.
Seasonal Trends in Energy Allocation
In Chapter 1, I created conceptual models to predict the effects of time of year on
the allocation of energy in individual lake trout. This conceptual model predicted
increasing allocation of energy towards reproduction as the spawning season approached
while allocation to growth would be maximized following the spawning season. One
deviation from the conceptual model of seasonal energy allocation was the increasing
levels of mean percent lipid through the summer period, with maximum values
coinciding with the spawn. This increase in percent lipids in the summer period is not
consistent with the results of Henderson and Wong (1998) which showed decreasing
percent lipids prior to the spawn, followed by increased percent lipids post-spawn in lake
trout sampled from the South Bay of Lake Huron. The first possible explanation for this
difference between studies is that my study tracked lipids in individual fish, rather than
147

measuring lipids from a homogenized group. Although the difference in methods alone
would not fully account for a complete reversal in the patterns observed between the
studies, group-based data may have biases related to the sampling methods in the field,
such as differences in size distribution or age distribution in individuals sampled on a
given date. Secondly, the Henderson and Wong (1998) study involved the use of field
fish to show seasonal lipid patterns while I evaluated seasonal percent lipids in hatchery
individuals. Finally, it is possible that the lipid allocation patterns I observed in the
hatchery are a result of artificial conditions (e.g., food availability, fish density, water
temperature) and that environmental conditions would dictate entirely different lipid
patterns.
Many studies have focused on the effects of ration level on growth of hatchery
fish, particularly for applications in improving hatchery efficiency (e.g. Johnsen et al.
2013; Flores and Vergara 2012; Flores et al. 2012). In general, increased food rations
result in increased growth in hatchery fish, although some negative consequences such as
increased waste (Flores and Vergara 2012), decreased food conversion efficiency
(Imsland and Gunnarsson 2011), and associated increased costs, limit the benefits of
over-feeding in production hatcheries. Based on these potential negative effects of overfeeding, much research has been done to determine the optimal diet type and composition
(e.g. Figueiredo-Silva et al. 2013; Zhang et al. 2012), amount (Laporte and Trushenski
2012; Luo et al. 2006), and feeding frequency (Johnsen et al. 2013; Hafs et al. 2012;
Flores and Vergara 2012) to maximize production per dollar spent on feed. In the current
study, I focused on growth rate as it relates to ration level and the tradeoffs that exist in
energy allocation towards growth, reproduction and percent lipids. The growth rates and

148

percent lipids in this study were inversely related, with a significant decline occurring in
growth during the spawning season, followed by a post-spawn increase. During this
post-spawn period, what is likely happening is that energy reserves in lipids are being
moved towards increases in length which results in decreased storage. The significance of
this trend is that lipid availability is what is likely driving increases in growth that occur
post-spawn. This hypothesized re-allocation of energy from lipid reserves is supported
by previous life history theory studies that predict that the role of lipid reserves is to
increase flexibility in their energy allocation by allowing re-allocation of energy from
lipid reserves to growth or reproduction, depending on seasonal needs (Jokela 1997;
Naesje et al. 2006; Berg et al. 2009).
While these hatchery data do not allow me to make direct comparisons to wild
scenarios because of differences in the hatchery and the wild, the seasonal data showing
allocation toward percent lipid and growth do provide strong evidence that there is a
connection between percent lipids and growth and that a strong tradeoff exists in the
energy allocation towards these variables, as seen in Henderson and Wong (1998).
Annual Changes in Reproduction
In the previous section, I discussed the effects of time of year on the energy
allocation of individual lake trout to growth, percent lipid and reproductive output. In
Chapter 1, I also presented a conceptual model to predict the effects of age on energy
allocation. Egg size and egg number in fish are a function of maternal effects such as
spawner age, size, and growth history and are also a function of environmental conditions
(Bergenius et al. 2002; Meekan and Fortier 1996; Miller et al. 1995). Two general
models have been put forward in the literature to describe effects of spawner age and size
149

on egg size have been shown in the literature: 1) egg size increases with spawner age and
size, (Gregersen et al. 2006; Berkeley et al. 2004) and 2) egg size is maximized at an
intermediate age, with young and old spawners having smaller, lower-quality eggs (Wang
et al. 2012; Kamler 2005). In either case, egg size is a very important measure of
reproductive output because it has been linked with growth and survival rates of larvae
(Berkeley et al. 2004). In the current study, the annual differences in egg size and egg
number were evaluated to determine the ration-specific effects of age on reproductive
output.
Within a treatment, egg size was more variable between years than was egg
number. This is consistent with previous studies that show the strong influence of
maternal effects, and specifically age, on egg size (e.g. Berkeley et al. (2004); Kamler
(2005); Gregersen et al. (2006); Wang et al. (2012)). Egg number was less variable across
years, within a treatment, because fish were exposed to the same ration level between
years and egg number is more reflective of environmental conditions, as seen in Scott
(1962).
One unexpected result that is not consistent with previous studies was the
decrease in egg size between years in all cases where significant changes occurred.
Numerous studies have shown that egg size typically increases with spawner age (e.g.
Pitman 1979; Sargent et al. 1987; Hislop 1988). This decrease in egg size suggests that,
in general, reproductive output was reduced between years due to the lack of as many
significant changes in egg number. Based on the age of these fish in the hatchery, the
conceptual model predicted reproductive output would increase because individuals are
all relatively young and should still be approaching the age where reproductive output is

150

maximized. This result may be an artifact of unnatural hatchery conditions in terms of
fish densities, temperatures, food intake, or environmental predictability. A second factor
may be that increases in egg number, although mostly not significant, were increasing
overall reproductive output of individuals through switching to a higher number of
smaller eggs. Modeled field data in Chapter 2 predicted an increase in egg size as fish
age which further suggests that the decrease in egg size between years in the hatchery is
an artifact of hatchery conditions.
Reaction Norms – Egg Number vs. Percent Lipids
Reaction norms have been shown to exist in multiple fish species in previous
studies in a multitude of measurements including growth rates, survival, and body size.
However, these relationships have been shown to be very sensitive and can change based
on introgression, domestication in hatchery fish, and environmental stressors (Darwish
and Hutchings 2009; Solberg et al. 2013). Percent lipids were shown to be positively
correlated with egg production in lake trout in a previous study that sought to determine
the role of ration on spawning timing and the role that food limitation plays in the
“decision” of females to spawn in a given year (Henderson and Wong 1998). In the
current study, I found predictable relationships between egg number and lipids in both
years across rations and, to a lesser extent, in between years.
Only two of the four stocks, the long-lived Michipicotens and Killalas, showed
robust relationships between lipids and egg number that held for both years, which
allowed rough estimates of fecundity based on non-invasive lipid measurements. These
data are consistent with life history theory that predicts annual changes in reproduction of
long-lived stocks will be less pronounced than those changes seen in shorter-lived stocks.
151

This is expected to be the case because long-lived stocks have a longer amount of time to
reach their age at maximized reproductive output, resulting in smaller changes between
years (Wang et al. 2012; Kamler 2005). This result shows that the relationship between
percent lipid and egg number changes as fish age, making prediction of egg number
based on percent lipids less reliable, particularly in short-lived stocks.
Reaction norms were found to exist in the majority of the comparisons of ration
groups. The two year classes of Kingscote showed relatively unstable relationships
between mean percent lipid and egg number as evidenced by five of the eight total
comparisons showing differences between the original and alternate model which may be
due to the unrealistically high ration they receive in the hatchery. In the Michipicoten
and Manitou stocks, only one of the comparisons revealed a difference between the
original and alternate model. This information provides evidence to suggest that when
exposed to a natural range of feeding levels, lake trout mean percent lipid can be used to
accurately predict egg production in a given age-class.
Study Limitations – Hatchery Design
The design of the common environment experiment in that the hatchery was such
that every stock exposed to a given feeding regime was contained in one tank, which
could be considered pseudoreplication (Hurlbert 1984). However, given the available
hatchery space, the number of fish available and logistics of the study, it was not possible
to set up additional tanks that would provide true replicates. Although, it is not true
replication, I did have two year classes of Kingscote that were in separate tanks and they
showed similar responses between year classes even though they were spread out across
all eight tanks. Also, the general trends across tanks were similar. For example, the
152

Kingscote stock had the highest lipid levels throughout all the tanks which suggests a
common pattern across all the tanks. Prior to the start of the experiment, I tried to
account for this lack of true replication by including two stocks within a tank from the
same year class. When these stocks are housed in a tank together, any differences are
more likely due to genetic divergences rather than tank effects, particularly because there
were no observable differences in behavior. Future projects should focus on increased
tank numbers to accommodate traditional statistical analyses. One last limitation of this
study was the differences in how many generations each stock had been in the hatchery.
Although the stocks that had been in the hatchery the longest were only two generations
removed from their wild ancestors, domestication has been shown to occur in fish that are
two generations removed from the wild (Reisenbichler and Rubin 1999; Reisenbichler et
al. 2003). Domestication of fish in the hatchery has been shown to result in changes in a
multitude of factors including growth (Solberg et al. 2013), fat reserves (Bronson 1984;
Gross 1998), and reproduction (Belyaev 1979; Fleming et al. 1996) relative to wild fish.
These potential differences in the amount of domestication between hatchery stocks
could influence energy allocation patterns in unpredictable ways as it was not possible to
disentangle the effects of life history from the effects of domestication in this study.
When possible, future studies involving stock comparisons should use fish that were
brought into the hatchery system at the same time. Additionally, when comparing
hatchery and wild fish, it would be ideal to examine individuals in the hatchery that have
had a minimal number of generations in the hatchery.

153

Implications
Understanding the stock-specific responses of lake trout to a change in their
environment gives us a predictive tool that estimates responses of individuals to a change
in prey availability. Predator-prey relationships in natural systems are dynamic and the
stochastic nature of prey availability has been documented in response to predators in a
number of piscivorous fish species, including lake trout (Kitchell and Crowder 1986; Eby
et al. 1995). These predator-prey dynamics and associated reductions in a preferred prey
species can results in niche shifts, decreased growth rates, and increased levels of
competition for remaining prey (Kitchell and Crowder 1986). Given that reductions in
prey in the natural environment are likely due to trophic shifts and anthropogenic
stressors, understanding how these stocks respond to reductions in energy intake in a
controlled environment gives some insights into how growth, percent lipid, and
reproductive output will be affected. For the case of the Killala stock, growth, lipids, and
reproductive output were all lower in the fish maintained at the low ration suggesting that
a reduction in ration would negatively affect energy allocation to all three variables. For
the Michipicoten stock, a reduction in energy intake affected growth and egg number,
with a greater change in the switchers in Year 2. This suggests that growth and egg
number would suffer when prey is limiting and that the effects of this reduction would be
even more pronounced if it lasted for multiple years in the Michipicoten stock. The
Manitou stock was relatively resistant to changes in egg number across all rations, at the
expense of percent lipid and growth rates in the low rations. This potentially means that
the Manitou stock would be the ideal stock for introduction to a limiting environment
because reproductive output is the least sensitive to ration of all the stocks studied.
However, it is also possible that the corresponding reduced growth and percent lipid in
154

limiting environments could reduce survival or reduce reproductive output in the future.
Finally, the Kingscote stock would be expected to have the largest decrease in growth of
all the variables based on my results in the situation that their natural environment
became even more limiting. These predictions of stock-specific changes in growth,
percent lipid, and reproductive output can have significant impacts on wild populations in
a situation where prey availability is reduced. Any reductions in egg production based on
a trophic change will directly affect future generations; however, reductions in growth
and percent lipid also have the possibility of reducing survival of individuals which will
also negatively impact the population.
Conclusions
This study demonstrated the importance of ration, and how changes in ration
levels determine growth rates, percent lipid, and reproductive output in different stocks.
The conceptual models presented in Chapters 1 and 2 predicted energy allocation of
individuals based on life history strategy, time of year, and fish age. Robust reaction
norms provided reliable predictions of fecundity within both years in the Michipicoten
and Manitou stocks but were only effective at predicting the following year’s fecundity in
the long-lived Michipicoten and Killala stocks. This likely reflects that the relationship
between these variables changes from year to year and depends, to some extent, on the
age of individuals and the mortality schedule to which they are adapted. Reproductive
output was evaluated as a function of both egg size and egg number. Egg number was
more variable across treatments than was egg size and egg size was more variable across
years (within treatments) than was egg number. This suggests that egg number is more
reflective of environmental conditions and food availability, while egg size is more fixed
155

and changes through time as a year-class ages. This general understanding of egg
number within a year class gives fishery managers insights into the food consumption by
individual fish and also allows managers to predict how a year-class will respond to a
year with limited food availability.

156

APPENDIX

157

Table 20. Summary of significance of changes in percent lipid between July 11, 2008
and levels during the spawn of five stocks of lake trout housed in a common garden
environment in OMNR hatchery. Lipid changes are recorded as not significant (NS), or
positively significant based on repeated measures analysis with α = 0.05. Ration levels
were manipulated within each treatment as follows High-High = maintained at a high
[0.5% of body weight daily] ration throughout study; High-Low = switched from a High
[0.5% of body weight daily] to a low [0.3% of body weight daily] ration, etc. Combined
data are data combined from all treatments.
Stock

Ration
p-value
Percent Lipid Change
High-High
0.000553
Positive
Killala
Low-Low
0.0840
NS
All Rations
0.00250
Positive
High-High
0.118
NS
High-Low
0.000481
Positive
2000 Kingscote Low-Low
0.0254
Positive
Low-High
0.000291
Positive
All Rations 0.000000623
Positive
High-High
0.0000873
Positive
High-Low 0.000000866
Positive
1999 Kingscote Low-Low
0.0000797
Positive
Low-High
0.0000419
Positive
All Rations 1.785E-013
Positive
High-High
0.00378
Positive
High-Low
0.00225
Positive
Manitou
Low-Low
0.0317
Positive
Low-High
0.148
NS
All Rations 0.0000819
Positive
High-High 0.00000130
Positive
High-Low
0.00987
Positive
Michipicoten
Low-Low
0.495
NS
Low-High
0.137
NS
All Rations
0.00120
Positive

158

Table 21. Summary of significance of changes in percent lipid between July 27, 2009
and levels during the spawn of five stocks of lake trout housed in a common garden
environment in OMNR hatchery. Lipid changes are recorded as not significant (NS), or
negatively significant based on repeated measures analysis with α = 0.05. Ration levels
were manipulated within each treatment as follows High-High = maintained at a high
[0.5% of body weight daily] ration throughout study; High-Low = switched from a High
[0.5% of body weight daily] to a low [0.3% of body weight daily] ration, etc. Combined
data are data combined from all treatments.
Stock

Ration
p-value Percent Lipid Change
High-High 1.66E-9
Positive
Killala
Low-Low
0.983
NS
All Rations 0.939
NS
High-High
0.689
NS
High-Low
0.777
NS
2000 Kingscote
Low-Low
0.670
NS
Low-High
0.343
NS
All Rations 0.945
NS
High-High
0.510
NS
High-Low
0.951
NS
1999 Kingscote
Low-Low
0.327
NS
Low-High
0.195
NS
All Rations 0.841
NS
High-High
0.857
NS
High-Low
0.339
NS
Manitou
Low-Low
0.841
NS
Low-High
0.536
NS
All Rations 0.775
NS
High-High
0.791
NS
High-Low
0.390
NS
Michipicoten
Low-Low
0.514
NS
Low-High
0.260
NS
All Rations 0.623
NS

159

Table 22. Year 1 comparison of growth rates (daily growth rate between sampling dates,
mm/day) and December-spawn growth (change in fork length) in five stocks of lake trout
housed in a common environment hatchery, exposed to one of four feeding regimes. %
difference columns compare value of ration-specific growth in row to the high-high
value.

Stock

Michipicoten

Killala
1999
Kingscote

2000 Kingscote

Manitou

Ration
H-H
H-L
L-L
L-H
H-H
L-L
H-H
H-L
L-L
L-H
H-H
H-L
L-L
L-H
H-H
H-L
L-L
L-H

Max Daily
Growth
Rate
0.1365
0.0843
0.0548
0.1181
0.1823
0.0486
0.2305
0.1688
0.0938
0.131
0.2129
0.1629
0.1289
0.2168
0.1534
0.1136
0.0765
0.0895

%
Difference

DecemberSpawn Growth

%
Difference

0.00%
-38.24%
-59.85%
-13.48%
0.00%
-73.34%
0.00%
-26.77%
-59.31%
-43.17%
0.00%
-23.49%
-39.46%
1.83%
0.00%
-25.95%
-50.13%
-41.66%

51.667
30.2
17.33
21.75
39.78
13.56
43
29.8
18.54
25.64
36.875
34.286
20.25
36.625
31.25
25.5
9.8
16.8

0.00%
-41.55%
-66.46%
-57.90%
0.00%
-65.91%
0.00%
-30.70%
-56.88%
-40.37%
0.00%
-7.02%
-45.08%
-0.68%
0.00%
-18.40%
-68.64%
-46.24%

160

Table 23. Year 2 comparison of growth rates (daily growth rate between sampling dates,
mm/day) and December-spawn growth (change in fork length)in five stocks of lake trout
housed in a common environment hatchery, exposed to one of four feeding regimes. %
difference columns compare value of ration-specific growth in row to the high-high
value.
Stock

Michipicoten

Killala

1999 Kingscote

2000 Kingscote

Manitou

Ration
H-H
H-L
L-L
L-H
H-H
L-L
H-H
H-L
L-L
L-H
H-H
H-L
L-L
L-H
H-H
H-L
L-L
L-H

Max Daily
Growth Rate
0.1482
0.0775
0.0867
0.1958
0.1392
0.0602
0.1396
0.0639
0.0686
0.1196
0.0994
0.0809
0.0783
0.1491
0.1365
0.0843
0.0548
0.1181

%
Difference
0.00%
-47.71%
-41.50%
32.12%
0.00%
-56.75%
0.00%
-54.23%
-50.86%
-14.33%
0.00%
-18.61%
-21.23%
50.00%
0.00%
-38.24%
-59.85%
-13.48%

161

DecemberSpawn Growth
36.5
18
18
46
33.75
23
32.182
14.5
18.77
36.67
38.6
20.86
17
36.38
30.75
15.5
9.667
28.8

%
Difference
0.00%
-50.68%
-50.68%
26.03%
0.00%
-31.85%
0.00%
-54.94%
-41.68%
13.95%
0.00%
-45.96%
-55.96%
-5.75%
0.00%
-49.59%
-68.56%
-6.34%

Table 24. AIC values of linear, hyperbolic, logarithmic, sigmoidal, and MichaelisMenten functions applied to the relationship between egg number and mean percent lipid
of Michipicoten lake trout housed in a common environment hatchery in Ontario.

Group

Year 1,
All
Rations

Year 2,
All
Rations

Year 1,
Switchers

Year 1,
NonSwitchers

Year 2,
Switchers

Year 2,
NonSwitchers

Relationship

No. of
Parameters
(k)

No. of
Observations
(n)

Residual
Sum of
Squares
(RSS)

Akaike
Information
Criterion
(AIC)

Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten
Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten
Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten
Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten
Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten
Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten

2
2
2
3
2
2
2
2
3
2
2
2
2
3
2
2
2
2
3
2
2
2
2
3
2
2
2
2
3
2

21
21
21
21
21
25
25
25
25
25
9
9
9
9
9
12
12
12
12
12
10
10
10
10
10
15
15
15
15
15

17.4833
18.8087
18.9058
17.1675
18.9058
131.5277
127.3083
128.8843
130.0384
128.4000
3.1158
3.0670
3.0880
3.1271
3.0880
12.9210
14.8584
14.6950
12.3231
14.6950
60.7910
60.6164
60.4096
61.5126
60.4096
63.9234
61.8395
62.8968
65.2343
62.8968

0.1512
1.6857
1.7939
1.7683
1.7939
45.5085
44.6934
45.0010
47.2239
44.9069
-5.5467
-5.6887
-5.6272
-3.5142
-5.6273
4.8874
6.5639
6.4312
6.3188
20.8464
22.0486
22.0198
21.9856
24.1666
21.9856
25.7445
25.2474
25.5017
28.0490
25.5017

162

Table 25. AIC values of linear, hyperbolic, logarithmic, sigmoidal, and MichaelisMenten functions applied to the relationship between egg number and mean percent lipid
of Killala lake trout housed in a common environment hatchery in Ontario.

Group

Year 1,
All
Rations

Year 2,
All
Rations

Relationship

No. of
Parameters
(k)

No. of
Observations
(n)

Residual
Sum of
Squares
(RSS)

Akaike
Information
Criterion
(AIC)

Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten
Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten

2
2
2
3
2
2
2
2
3
2

19
19
19
19
19
18
18
18
18
18

36.4578
34.2970
34.1326
33.2732
34.1326
32.3531
34.6681
34.4181
31.8365
43.6984

16.3826
15.2218
15.1304
16.6459
15.1305
14.5541
15.7981
15.6678
16.2643
19.9649

163

Table 26. AIC values of linear, hyperbolic, logarithmic, sigmoidal, and MichaelisMenten functions applied to the relationship between egg number and mean percent lipid
of 1999 Year class Kingscote lake trout housed in a common environment hatchery in
Ontario.

Group

Year 1,
All
Rations

Year 2,
All
Rations

Year 1,
Switchers

Year 1,
NonSwitchers

Year 2,
Switchers

Year 2,
NonSwitchers

Relationship

No. of
Parameters
(k)

No. of
Observations
(n)

Residual
Sum of
Squares
(RSS)

Akaike
Information
Criterion
(AIC)

Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten
Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten
Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten
Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten
Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten
Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten

2
2
2
3
2
2
2
2
3
2
2
2
2
3
2
2
2
2
3
2
2
2
2
3
2
2
2
2
3
2

46
46
46
46
46
45
45
45
45
45
21
21
21
21
21
25
25
25
25
25
21
21
21
21
21
24
24
24
24
24

99.1985
98.7019
97.9769
84.4786
97.9769
107.6139
108.0412
108.3411
107.3965
108.3000
47.2195
45.6034
44.4187
37.5113
44.4187
34.8705
35.5427
35.9675
34.7432
35.9675
31.5784
31.2686
31.2272
31.1979
31.2272
68.2676
70.6104
70.7726
67.2793
70.7726

39.3502
39.1193
38.7802
33.9614
38.7802
43.2349
43.4133
43.5380
45.1439
43.5209
21.0160
20.2847
19.7319
18.1825
19.7319
12.3191
12.7964
13.0935
14.2277
13.0935
12.5670
12.3599
12.3321
14.3124
12.3321
29.0891
29.8990
29.9540
30.7392
29.9540

164

Table 27. AIC values of linear, hyperbolic, logarithmic, sigmoidal, and MichaelisMenten functions applied to the relationship between egg number and mean percent lipid
of 2000 Year class Kingscote lake trout housed in a common environment hatchery in
Ontario.

Group

Year 1,
All
Rations

Year 2,
All
Rations

Year 1,
Switchers

Year 1,
NonSwitchers

Year 2,
Switchers

Year 2,
NonSwitchers

Relationship

No. of
Parameters
(k)

No. of
Observations
(n)

Residual
Sum of
Squares
(RSS)

Akaike
Information
Criterion
(AIC)

Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten
Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten
Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten
Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten
Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten
Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten

2
2
2
3
2
2
2
2
3
2
2
2
2
3
2
2
2
2
3
2
2
2
2
3
2
2
2
2
3
2

29
29
29
29
29
27
27
27
27
27
15
15
15
15
15
14
14
14
14
14
15
15
15
15
15
12
12
12
12
12

47.5423
47.6245
47.7139
47.5375
47.7139
36.1960
37.2474
37.5353
35.9898
37.5353
12.9319
12.7325
12.7487
12.5472
12.7487
30.9364
30.8620
30.8077
31.0387
30.8077
25.0761
25.2157
25.1732
24.9499
25.1732
9.5636
10.2680
10.5903
9.3117
10.5903

18.3354
18.3855
18.4399
20.3325
18.4399
11.9140
12.6871
12.8950
13.7597
12.8950
1.7747
1.5416
1.5607
3.3217
1.5607
15.1003
15.0666
15.0419
17.1465
15.0419
11.7080
11.7913
11.7660
13.6323
11.7659
1.2767
2.1295
2.5004
2.9564
2.5004

165

Table 28. AIC values of linear, hyperbolic, logarithmic, sigmoidal, and MichaelisMenten functions applied to the relationship between egg number and mean percent lipid
of Manitou lake trout housed in a common environment hatchery in Ontario.

Group

Year 1,
All
Rations

Year 2,
All
Rations

Year 1,
Switchers

Year 1,
NonSwitchers

Year 2,
Switchers

Year 2,
NonSwitchers

Relationship

No. of
Parameters
(k)

No. of
Observations
(n)

Residual
Sum of
Squares
(RSS)

Akaike
Information
Criterion
(AIC)

Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten
Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten
Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten
Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten
Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten
Linear
Logarithmic
Hyperbolic
Sigmoidal
Michaelis - Menten

2
2
2
3
2
2
2
2
3
2
2
2
2
3
2
2
2
2
3
2
2
2
2
3
2
2
2
2
3
2

43
43
43
43
43
39
39
39
39
39
21
21
21
21
21
22
22
22
22
22
18
18
18
18
18
21
21
21
21
21

122.4853
121.3740
120.1531
118.8509
120.2000
201.0749
204.3550
206.7124
198.8060
222.4000
67.3583
64.1144
63.3285
60.2421
63.3285
47.3867
48.7971
48.8508
47.2918
48.8508
75.9975
74.7863
75.3779
75.1482
76.7054
116.7690
120.8377
121.9949
115.3620
122.0000

49.0120
48.6201
48.1854
49.7168
48.2021
67.9645
68.5956
69.0429
69.5219
71.8957
28.4756
27.4391
27.1801
28.1308
27.1801
20.8806
21.5258
21.5500
22.8365
21.5500
29.9259
29.6367
29.7786
31.7236
30.0928
40.0292
40.7484
40.9486
41.7746
40.9495

166

Table 29. AIC values of models of best fit, as calculated by AIC comparisons of
Michaelis-Menten, hyberbolic, linear, sigmoidal, and logarithmic function, and the
alternate models which represent the model of best fit applied to the alternate group (year
or ration-group) in the relationship between egg number and mean percent lipid of five
stocks of lake trout housed in a common environment experiment in an OMNR hatchery.
The models of best fit and alternate models were applied to the same data in each case to
determine AIC differences.
Stock

Year 1 AIC

Michipicoten
Killala
1999 Kingscote
2000 Kingscote
Manitou

0.1512
15.1304
33.9614
18.3354
48.1854

Year 1 Data with
Year 2 Regression
AIC
3.2624
17.3706
47.9834
57.3064
72.6587

Stock

Year 1 Non
Switcher
AIC

Year 1 Non Switcher
Data with Switcher
Regression AIC

Year 1
Switcher
AIC

Michipicoten
1999 Kings
2000 Kingscote
Manitou

4.8874
12.3191
15.0419
20.8806

8.1100
31.0574
32.3629
27.9745

-5.6887
18.1825
1.5416
27.1801

Stock

Year 2 Non
Switcher
AIC

Year 2 Non Switcher
Data with Switcher
Regression AIC

Year 2
Switcher
AIC

Michipicoten
1999 Kingscote
2000 Kingscote
Manitou

25.2474
29.0891
1.2767
40.0292

29.3780
33.9081
8.6847
43.7990

21.9856
12.3321
11.7080
29.6367

167

Year 2 AIC
44.6934
14.5541
43.2349
11.9140
67.9645

Year 2 Data with
Year 1 Regression
AIC
45.9215
15.7552
52.3757
32.3270
81.6180
Year 1 Switcher
Data with Non
Switcher Regression
AIC
0.0361
32.4700
6.7246
32.8411
Year 2 Switcher
Data with Non
Switcher Regression
AIC
23.0036
23.4768
13.0338
33.2484

Table 30. P-values for 2008 hatchery data comparing residuals (absolute values) between
ration treatments based on sigmoidal function applied to 2008 lipid-fecundity data of five
stocks of lake trout held in a common garden experiment.
Killala
High-High

High-High

High-Low
n/a

High-Low

n/a

Low-Low

0.5749

n/a

Low-High
2000 Kingscote
High-High

n/a
High-High

n/a
High-Low
0.4570

High-Low

0.4570

Low-Low

0.4807

0.9997

Low-High
1999 Kingscote
High-High

0.8475
High-High

0.8757
High-Low
0.9923

High-Low

0.9923

Low-Low

0.8221

0.7027

Low-High
Manitou
High-High

0.2558
High-High

0.2001
High-Low
0.5683

High-Low

0.5683

Low-Low

0.9996

0.6441

Low-High
Michipicoten
High-High

0.4922
High-High

0.9994
High-Low
0.5941

High-Low

0.5941

Low-Low

0.6408

0.9988

Low-High

0.9992

0.6431

Low-Low
0.5749

Low-High
n/a

n/a

n/a
n/a

n/a
Low-Low
0.4807

Low-High
0.8475

0.9997

0.8737
0.9004

0.9004
Low-Low
0.8221

Low-High
0.2558

0.7027

0.2001
0.7572

0.7572
Low-Low
0.9996

Low-High
0.4922

0.6441

0.9994
0.5686
Low-High
0.9997

0.9988

168

0.5686
Low-Low
0.6408

0.6431
0.6924

0.6924

Table 31. P-values for 2009 hatchery data comparing residuals (absolute values) between
ration treatments based on sigmoidal function applied to 2009 lipid-fecundity data of five
stocks of lake trout held in a common garden experiment.
Killala
High-High

High-High

High-Low
n/a

High-Low

n/a

Low-Low

0.2491

n/a

Low-High
2000 Kingscote
High-High

n/a
High-High

n/a
High-Low
0.1576

High-Low

0.1576

Low-Low

0.9285

0.1177

Low-High
1999 Kingscote
High-High

0.1264
High-High

0.8163
High-Low
0.0919

High-Low

0.0919

Low-Low

0.0654

0.9985

Low-High
Manitou
High-High

0.0021
High-High

0.2893
High-Low
0.3367

High-Low

0.3367

Low-Low

0.1041

0.0064

Low-High
Michipicoten
High-High

0.4297
High-High

0.0723
High-Low
0.6752

High-Low

0.6752

Low-Low

0.2236

0.1752

Low-High

0.0782

0.4367

Low-Low
0.2491

Low-High
n/a

n/a

n/a
n/a

n/a
Low-Low
0.9285

Low-High
0.1264

0.1177

0.8163
0.0981

0.0981
Low-Low
0.0654

Low-High
0.0021

0.9985

0.2893
0.2696

0.2696
Low-Low
0.1041

Low-High
0.4297

0.0064

0.0723
0.3865
Low-High
0.0782

0.1752

169

0.3865
Low-Low
0.2236

0.4367
0.0366

0.0366

WORKS CITED

170

WORKS CITED

Álvarez, D., and A. G. Nicieza. 2005. Compensatory response ‘defends’ energy levels
but not growth trajectories in brown trout, Salmo trutta L. Proceedings of the
Royal Society B: Biological Sciences, 272: 601-607.
Anderson, D. R., Link, W. A., Johnson, D. H., and K. P. Burnham. 2001. Suggestions for
presenting the results of data analyses. The Journal of Wildlife Management 65:
373-378.
Barot, S., M. Heino, L. O’Brien, and U. Dieckmann. 2004. Long term trend in the
maturation norm of two cod stocks. Ecological Applications 14: 1257-1271.
Belyaev, D. K. 1979. Destabilizing selection as a factor in domestication. Journal of
Heredity 70: 301–308.
Berkeley, S. A., Chapman, C., and S. M. Sogard. 2004. Maternal age as a determinant of
larval growth and survival in a marine fish, Sebastes melanops. Ecology: 12581264.
Beauchamp, D. A., Sergeant C. J., Mazur, M. M., Scheuerell, J. M., Schindler, D. E., and
M. D. Scheuerell. 2004. Spatial-temporal dynamics of early feeding demand and
food supply for sockeye salmon fry in LakeWashington. Transactions of the
American Fisheries Society 133:1014–32.
Bronson, F. H. 1984. Energy allocation and reproductive development in wild and
domestic mice. Biology of Reproduction 31: 83–88.
Burnham, K. P., and D. R. Anderson. 2002. Model selection and multi-model inference:
a practical information-theoretic approach. Springer.
Claus, L. C. 2011. Influence of life history traits on accumulation of polybrominated
diphenyl ethers and polychlorinated biphenyls in three lake trout populations from
Lake Superior. Master of Science Thesis, Michigan State University.
Cushing, D. H. 1969. The regularity of the spawning season of some fishes. J Cons Int
Explor Mer 33:81–92.
Darwish, T. L., and J. A. Hutchings 2009. Genetic variability in reaction norms between
farmed and wild backcrosses of Atlantic salmon (Salmo salar). Canadian Journal
of Fisheries and Aquatic Sciences 66: 83-90.

171

Eby, L. A., Rudstam, L. G., & Kitchell, J. F. (1995). Predator responses to prey
population dynamics: an empirical analysis based on lake trout growth rates.
Canadian Journal of Fisheries and Aquatic Sciences, 52: 1564-1571.
Ferreri, C. P., & Taylor, W. W. (1996). Compensation in individual growth rates and its
influence on lake trout population dynamics in the Michigan waters of Lake
Superior. Journal of Fish Biology, 49(5), 763-777.
Figueiredo-Silva, A. C., Saravanan, S., Schrama, J. W., Panserat, S., Kaushik, S., &
Geurden, I. 2012. A comparative study of the metabolic response in rainbow trout
and Nile tilapia to changes in dietary macronutrient composition. The British
journal of nutrition: 1-11.
Fleming I. A., Jonsson B., Gross M. R., and A. Lamberg. 1996. An experimental study of
the reproductive behaviour and success of farmed and wild Atlantic salmon
(Salmo salar). Journal of Applied Ecology 33: 893–905.
Flores, H., & Vergara, A. (2012). Effect of reducing the feeding frequency on the
survival, growth, conversion, and feeding behavior of juvenile Atlantic salmon
Salmo salar (Linnaeus, 1758): an experience at the productive level. Latin
American Journal of Aquatic Research 40: 536-544.
Frisch, A. J., & Anderson, T. A. (2000). The response of coral trout (Plectropomus
leopardus) to capture, handling and transport and shallow water stress. Fish
Physiology and Biochemistry, 23(1), 23-34.
Gregersen, F., Haugen, T. O., and N. Larsen. 2006. Egg size differentiation among
sympatric demes of brown trout: possible effects of density-dependent
interactions among fry. Ecology of Freshwater Fish 15: 237-246.
Gross, M. R. 1998. One species with two biologies: Atlantic salmon (Salmo salar) in the
wild and in aquaculture. Canadian Journal of Fisheries and Aquatic Sciences 55:
131–144.
Hafs, A. W., Mazik, P. M., Kenney, P. B., and J. T Silverstein. 2012. Impact of carbon
dioxide level, water velocity, strain, and feeding regimen on growth and fillet
attributes of cultured rainbow trout Oncorhynchus mykiss. Aquaculture 350: 4653.
Heino, M., and V. Kaitala. 1999. Evolution of resource allocation between growth and
reproduction in animals with indeterminate growth. Journal of Evolutionary
Biology 12: 423-429.
Henderson, B. A., and J. L. Wong. 1998. Control of lake trout reproduction: role of
lipids. Journal of fish biology, 52: 1078-1082.

172

Hill, D. K., and J. J. Magnuson, J. J. 1990. Potential effects of global climate warming
on the growth and prey consumption of Great Lakes fish. Transactions of the
American Fisheries Society 119: 265-275.
Hislop, J. R. 1988. The influence of maternal length and age on the size and weight of the
eggs and the relative fecundity of the haddock, Melanogrammus aeglefinus, in
British waters. Journal of Fish Biology 32: 923-930.
Hurlbert, S. H. 1984. Pseudoreplication and the design of ecological field experiments.
Ecological monographs: 54:187-211.
Hurst, T. P., and D. O. Conover. 2003. Seasonal and interannual variation in the
allometry of energy allocation in juvenile striped bass. Ecology 84: 3360-3369.
Hutchings, J. A., Picklet, A., McGregor-Shaw, C. R. and L. Poirier. 1999. Influence of
sex, body size and reproduction on overwinter lipid depletion in brook trout.
Journal of Fish Biology 55: 1020–1028.
Imsland, A. K., Foss, A., Folkvord, A., Stefansson, S. O., and T. M. Jonassen. 2005. The
interrelation between temperature regimes and fish size in juvenile Atlantic cod
(Gadus morhua): effects on growth and feed conversion efficiency. Fish
Physiology and Biochemistry, 31(4), 347-361.
Jastrebski, C. J., and Y. E. Morbey. 2009. Egg Size Variation in Lake Trout: PhenotypeHabitat Correlations Show an Effect of Rearing Environment. Transactions of the
American Fisheries Society 138: 1342-1351.
Jennings, S. and R. J. H. Beverton. 1991. Intraspecific variation in the life history tactics
of Atlantic herring (Clupea harengus L.) stocks. ICES Journal of Marine Sciences
48: 117-125.
Johnsen, C. A., Hagen, Ø., Solberg, C., Björnsson, B. T. H., Jönsson, E., Johansen, S. J.
S., and E. Å. Bendiksen. 2013. Seasonal changes in muscle structure and flesh
quality of 0+ and 1+ Atlantic salmon (Salmo salar L.): impact of feeding regime
and possible roles of ghrelin. Aquaculture Nutrition 19:15-34.
Kamler, E. 2005. Parent–egg–progeny relationships in teleost fishes: an energetics
perspective. Reviews in Fish Biology and Fisheries 15: 399-421.
Kitchell, J. F., and L. B. Crowder. 1986. Predator-prey interactions in Lake Michigan:
model predictions and recent dynamics. Environmental Biology of Fishes 16:
205-211.
Kozlowski, J. 1996. Optimal Allocation of Resources Explains Interspecific Life-History
Patterns in Animals with Indeterminate Growth. Proceedings of the Royal Society
B 263: 559-566.
173

Lambert, Y., and J. D. Dutil 2000. Energetic consequences of reproduction in Atlantic
cod (Gadus morhua) in relation to spawning level of somatic energy reserves.
Canadian Journal of Fisheries and Aquatic Sciences 57: 815-825.
Laporte, J., and J. Trushenski. 2012. Production performance, stress tolerance and
intestinal integrity of sunshine bass fed increasing levels of soybean meal. Journal
of Animal Physiology and Animal Nutrition 96: 513-526.
Marsden, J. E., and C. C. Krueger. 1991. Spawning by hatchery-origin lake trout
(Salvelinus namaycush) in Lake Ontario: data from egg collections, substrate
analysis, and diver observations. Canadian Journal of Fisheries and Aquatic
Sciences 48: 2377-2384.
Myers, J. M., Park, L. K., Neely, K. G., Swanson, P., Elz, A. E., and J. J. Hard. 2011.
Feeding Ration, Genetics, and Reproductive Traits in Female Coho Salmon: Is
Bigger Better?. Journal of the World Aquaculture Society 42: 812-823.
Næsje, T. F., Thorstad, E. B., Forseth, T., Aursand, M., Saksgard, R. and A. G. Finstad.
2006. Lipid class content as an indicator of critical periods for survival in
juvenile Atlantic salmon (Salmo salar). Ecology of Freshwater Fish 15: 572–577.
Pianka, E. R. 1976. Natural Selection of Optimal Reproductive Tactics. American
Zoologist 16: 775-78.
Pitman, R. W. 1979. Effects of female age and egg size on growth and mortality in
rainbow trout. The Progressive Fish Culturist 41: 202-204.
Purchase, C. F., and J. A. Brown. 2001. Stock-specific changes in growth rates, food
conversion efficiencies, and energy allocation in response to temperature change
in juvenile Atlantic cod. Journal of Fish Biology 58: 36-52.
Reisenbichler, R. R. and S.P. Rubin. 1999. Genetic changes from artificial propagation of
Pacific salmon affect the productivity and viability of supplemented populations.
ICES Journal of Marine Science 56: 459-466.
Reisenbichler, R.R., Utter, F. M., and C.C. Krueger. 2003. Genetic concepts and
uncertainties in restoring fish populations and species. In: Strategies for Restoring
River Ecosystems: Sources of Variability and Uncertainty in Natural and
Managed Systems: 149-183.
Robards, M. D., Anthony, J. A., Rose, A. G. and J. F. Piatt, J. F. 1999. Changes in
proximate
composition and somatic energy content for Pacific sand lance (Ammodytes
hexapterus) from Kachemak Bay, Alaska relative to maturity and season. Journal
of Experimental Biology and Ecology 242, 245–258.
174

Sargent, R. C., Taylor, P. D., and M. R. Gross. 1987. Parental care and the evolution of
egg size in fishes. American Naturalist 129: 32-46.
Schreck, C. B., Contreras-Sanchez, W., and M. S. Fitzpatrick. 2001. Effects of stress on
fish reproduction, gamete quality, and progeny. Aquaculture 197: 3-24.
Schreck, C. B. 2010. Stress and fish reproduction: the roles of allostasis and hormesis.
General and comparative endocrinology 165: 549-556.
Shuter, B. J., M. L. Jones, R. M. Korver, and N. P. Lester. 1998. A general, life history
based model for regional management of fish stocks: The inland lake trout
(Salvelinus namaycush) fisheries of Ontario. Canadian Journal of Fisheries and
Aquatic Sciences 55: 2161-2177.
Sloman, K. A., Metcalfe, N. B., Taylor, A. C., & Gilmour, K. M. (2001). Plasma cortisol
concentrations before and after social stress in rainbow trout and brown trout.
Physiological and Biochemical Zoology, 74: 383-389.
Solberg, M. F., Skaala, Ø., Nilsen, F., and K. A. Glover. 2013. Does domestication cause
changes in growth reaction norms? A study of farmed, wild and hybrid atlantic
Salmon families exposed to environmental stress. PloS one, 8(1), e54469.
Spromberg, J. A., and W. J. Birge. 2005. Modeling the effects of chronic toxicity on fish
populations: The influence of life history strategies. Environmental Toxicology
and Chemistry 24: 1532-1540.
Stauffer, T. M. 1981. Collecting gear for lake trout eggs and fry. The Progressive FishCulturist, 43: 186-193.
Stearns, S.C., and J.C. Koella. 1986. The Evolution of Phenotypic Plasticity in LifeHistory Traits: Predictions of Reaction Norms for Age and Size at Maturity.
Evolution 40: 893-913.
Thompson, J. M., Bergersen, E. P., Carlson, C. A., and L. R. Kaeding. 1991. Role of
Size, Condition, and Lipid Content in the Overwinter Survival of Age− 0
Colorado Squawfish. Transactions of the American Fisheries Society, 120(3),
346-353.
Tian, X., and J. G. Qin. 2004. Effects of previous ration restriction on compensatory
growth in barramundi Lates calcarifer. Aquaculture 235: 273-283.

175

Wang, H. Y., D. W. Einhouse, D. G. Fielder, L. G. Rudstam, C. S. Vandergoot, A. J.
VanDeValk, T. G. Zorn, and T. O. Hook. 2012. Maternal and stock effects on
egg-size variation among walleye Sander vitreus stocks from the Great Lakes
region. Journal of Great Lakes Research 38: 477-489.
Yibo, C., and L. Jiankang. 1990. Comparison of energy budget among six teleosts—IV.
Individual differences in growth and energy budget. Comparative Biochemistry
and Physiology Part A: Physiology 97: 551-554.
Zhang, Y., Øverland, M., Shearer, K. D., Sørensen, M., Mydland, L. T., and T.
Storebakken. 2012. Optimizing plant protein combinations in fish meal-free diets
for rainbow trout Oncorhynchus mykiss by a mixture model. Aquaculture 360: 25

176