{5:- J.
= 3:

.
3:.
3.1.

... 3 g to.

911V. :5... .2 3!:

2.»: M21133: 7.5....

. . . l» .5... 9

aix.1=‘hri:.§ Irv .
“'81:...1! z.
. i
3!:

7
:-
:5...
Ital-Iv. C!
72.13:... :J p“
in... .3 .. I.
1 .u 3.. 2:33.. I
.I: :1] l 2:.
.3 bl... .70 '
.3..::\\3I.:.
.3

IT‘SI.”

.rx.
3!:
l .

A...
.33.}... 3
I. 25s.. .3.

s
377.. It.
.3. :3
. . . i

 

Juana;

. .i a
pat?

.
,

l. . Kain»:

r

2. fit :.

11...!
. 1:. 3......

,a. v t'x.
5.5.35:
r. 1.... . c

.I. 93.3. 1.;

I, .

. .v::.

. . “31.7....
S

t...
w.

 

Ways

>~ LIBRARY
Michigan State
University

 

 

 

 

This is to certify that the

thesis entitled

Genetic Diversity and Interrelationships of Wild and
Hatchery Lake Trout in the Upper Great Lakes:
Inferences for Broodstock Management and
Development of Restoration Strategies

presented by

Kevin Scott Page

has been accepted towards fulfillment
of the requirements for

 

 

M.S. degree in Fish. & Wildl.

//,.,,. f/MZ

Major professor

 

Date ”-8-0,

 

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

PLACE IN RETURN Box to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 C‘JCIRCIDatoDue.p65-p.15

 

GENETIC DIVERSITY AND INTERRELATIONSHIPS OF WILD AND
HATCHERY LAKE TROUT IN THE UPPER GREAT LAKES: INFERENCES
FOR BROODSTOCK MANAGEMENT AND DEVELOPMENT OF
RESTORATION STRATEGIES
By

Kevin Scott Page

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF FISHERIES AND WILDLIFE
Department of Fisheries and Wildlife

2001

ABSTRACT
GENETIC DIVERSITY AND INTERRELATIONSHIPS OF WILD AND
HATCHERY LAKE TROUT IN THE UPPER GREAT LAKES: INFERENCES
FOR BROODSTOCK MANAGEMENT AND DEVELOPMENT OF
RESTORATION STRATEGIES
By

Kevin Scott Page

Lake trout (Salvelinus namaycush) populations were extirpated from most of the
upper Great Lakes due to overfishing and the invasion of the predatory sea lamprey
(Petromyzon marinas). A primary component of efforts to restore lake trout has
involved the stocking of juveniles from six lake trout broodstocks. Preservation of the
genetic diversity of these lake trout broodstocks and remaining wild populations is
considered an important component in lake trout hatchery and stocking programs. We
used molecular genetic techniques to l) evaluate how effective management practices
have been in preserving genetic diversity during the development and maintenance of
lake trout hatchery broodstocks and in production of offspring for release and 2) develop
a fundamental understanding of the levels and partitioning of genetic diversity among
hatchery broodstocks and remnant wild populations. We observed significant differences
in allele fiequencies among cohorts sampled at each stage of the broodstock program.
Most importantly, we present evidence of extremely low effective population sizes during
the production of lake trout juveniles. Analyses revealed that a significant portion of
genetic diversity in wild populations was structured based on basin of origin and among
lake trout morphotypes (bumpers, leans and siscowets). Recommendations related to

lake trout management are developed based on these findings.

ACKNOWLEDGEMENTS

Support for this project was provided by the Great Lakes Protection Fund, the
Great Lakes Fisheries Trust, and the Michigan Sea College Program. For their time and
effort collecting samples and providing valuable information on hatchery practices and
backgrounds of lake trout broodstocks, I would like to give a special thanks to the crews
of Iron River, Pendill’s Creek, and Jordan River National Fish Hatcheries, specifically,
Dale Bast, Faber Bland, Dave Huntly, Denise Johnston, Crystal LeGault, Dave Radloff
and Rick Westerhof. Without the support of these hatcheries, this project would not have
been possible. I would also like to thank David Anderson, Chuck Bronte, Mark Holey,
Jim Peck, Ruth Phillips, Mark Rude, and Steve Schram for providing the wild lake trout
samples used in this project. I am grateful to Darryl Bathe], Dave Blick, Michael
Donofrio, Randy Eshenroder, Mark Holey, John Huber and Roger Hugill for background
information they provided on lake trout broodstocks. Thanks to Mary Burnham-Curts,
Bryan Epperson, Mike Jones and Kim Scribner for their support, guidance and

encouragement.

iii

TABLE OF CONTENTS

LIST OF TABLES ............................................................................. Vi
LIST OF FIGURES ........................................................................... vii
INTRODUCTION .............................................................................. 1
CHAPTER 1

INFLUENCES OF HATCHERY PRACTICES ON THE PRESERVATION OF
GENETIC DIVERSITY IN LAKE TROUT HATCHERY BROODSTOCKS USED

FOR RESTORATION EFFORTS IN THE UPPER GREAT LAKES .................. 5
History of Hatchery Broodstocks .................................................... 6
Methods ............................................................................... 18

Stage 1: Broodstock Development ....................................... 18
Stage 2: Broodstock Maintenance and Production ..................... 24
Stage 3: Distribution and Stocking of Broodstock Juveniles .......... 30
Results .................................................................................. 3 1
Stage 1: Broodstock Development ........................................ 31
Stage 2: Broodstock Maintenance and Production ...................... 34
Stage 3: Distribution and Stocking of Broodstock Juveniles ......... 39
Discussion ............................................................................ 41
Role of Hatcheries in Lake Trout Restoration ........................... 42
Stage 1: Broodstock Development ....................................... 44
Stage 2: Broodstock Maintenance and Production ..................... 47
Stage 3: Distribution and Stocking of Broodstock Juveniles ......... 53
Summary .............................................................................. 55
CHAPTER 2

ASSESSING LEVELS AND PARTITIONING OF GENETIC DIVERSITY IN
WILD AND HATCHERY LAKE TROUT POPULATIONS: RELEVANCE FOR

LAKE TROUT MANAGEMENT AND RESTORATION ............................. 5 8
Methods ............................................................................... 64
Results ................................................................................. 69
Discussion ............................................................................. 72

Levels and Partitioning of Genetic Diversity ........................... 72

Management .................................................................. 77

Summary .............................................................................. 83
APPENDIX I

GUIDELINES EXPRESSED IN THE LITERATURE REGARDING HATCHERY
PRACTICES RELATED TO THE THREE STAGES OF HATCHERY
BROODSTOCK MANAGEMENT ........................................................ 84

APPENDIX 11
OVERALL RESULTS AND RECOMMENDATIONS FOR THE LAKE TROUT

HATCHERY PROGRAM ................................................................... 86
APPENDIX III

TABLES AND FIGURES ................................................................... 89
LITERATURE CITED ..................................................................... l l l

Table 1.

Table 2.

Table 3.

Table 4.

Table 5.

Table 6.

Table 7.

Table 8.

LIST OF TABLES

Comparisons of allele frequencies and measures of genetic diversities for three
wild populations and respective hatchery strains ..................................... 89
Intrastrain comparisons of allele frequencies and measures of genetic diversity
of the Marquette broodstock (SMD) ................................................... 92
Comparisons of allele frequencies and measures of genetic variability for
hatchery broodstocks and progeny ...................................................... 94

Hatchery strain comparisons of 1) total numbers of individuals spawned, 2) the
potential numbers of individuals contributing genes to the subsequent generation
after excising of eggs by hatcheries, 3) estimated effective number of breeders
for spawner lots evaluated (N), and 4) the realized contributers (Nb/N applied to
potential contributers) .................................................................... 96

Numbers and proportions of potential contributers representing the distribution
of lake trout (Salvelinus namaycush) broodstock egg production to various
rearing facilities and management programs .......................................... 97

Total egg numbers distributed to hatchery facilities and programs for three
hatchery broodstocks, and the proportion of total egg numbers attributed to
specific spawning dates and adult numbers spawned ................................ 98

Levels and partitioning of genetic diversity for populations of lake trout hatchery
broodstocks, wild lean lake trout, and wild lake trout populations and
morphotypes of Lake Superior .......................................................... 99

Summary of F -statistics partitioned genetic variation within and among hatchery
and wild populations of lake trout (Salvelinus namaycush) from the upper Great
Lakes ...................................................................................... 101

vi

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Figure 6.

Figure 7.

LIST OF FIGURES

Locations of hatchery and wild populations of lake trout (Salvelinus
namaycush) sampled. Numbers refer to the original source sites of broodstocks.
Italicized name represent wild source populations of associated broodstocks
sampled for genetic comparisons. 1.Lake trout stocked into Lewis Lake, WY
were collected from northern Lake Michigan. The Lewis Lake broodstock was
subsequently developed from these feral fish ................................................... 102

Stages during development of lake trout (Salvelinus namaycush) hatchery
broodstocks and juvenile production contributing to changes in levels of gene
diversity and relatedness ................................................................................... 103

Hypothetical spawning scenarios as advocated in the broodstock management
plan (1:1 matings; Holey 1997) and actual conditions realized at the hatcheries
(i.e., 5:5). In this simple example, we assume adult males and females are
unrelated and each female contributes two offspring (e.g., no variance in
reproductive success). 0 is the probability that any two alleles shared between
two individuals are inherited from a common ancestor (identical-by-descent).
0:0.25, offspring related at the level of full -siblings. 0:0.125, offspring are
related at the level of half-siblings. (a.) “01 1” represents offspring. Subscripts
designate the female that produced the offspring (i.e., female 1) and the
offspring number (i.e. offspring 1 of 2). (b.) Male 1 and 4 each contribute 20%
of the offspring (Male 1 offspring italicized). (c.) Male 2 contributes 60% of
the progeny (Male 2 offspring in bold). (d.) Male 3 and 5 do not contribute
offspring. Mating strategies should seek to minimize mean 0 among progeny
and maximize the proportion of unrelated offspring ........................................ 104

Frequency (Y -axis) distributions of par-wise coefficients of relatedness (X-axis)
for source populations and hatchery broodstocks ............................................ 105

Frequency (Y-axis) distributions of par-wise coefficients of relatedness (X-axis)
for intergenerational Marquette (SMD) hatchery broodstocks ........................ 106

Frequency (Y-axis) distributions of par-wise coefficients of relatedness (X-axis)
for broodtsock adults and juveniles .................................................................. 107

Neighbor-joining trees representing genetic divergences in hatchery (a.) and

wild (b.) lake trout (Salvelinus namaycush) populations based on Cavalli-Sforza
and Edwards chord distances (1967). Numbers represent bootstrap values. . . 108

vii

Figure 8. Neighbor-joining tree representing genetic divergence of remnant wild lake
trout (Salvelz'nus namaycush) populations based on Cavalli-Sforza and Edwards
chord distances (1967). The tree reveals that a comparatively small segment of
lake trout diversity (leans) is currently being utilized in the lake trout
broodstock program. All populations are from Lake Superior except for lean
lake trout from Parry Sound that are from Lake Huron waters of Georgian Bay,
Ontario ............................................................................................................ .109

Figure 9. Neighbor-joining tree representing genetic divergence of lake trout (Salvelinus
namaycush) hatchery broodstocks based on Cavalli-Sforza and Edwards
chord distances (1967). This tree illustrates two possible strategies for
maximizing genetic diversity in hatchery broodstocks. Strategy 1 advocates the
use of all existing hatchery broodstocks. Strategy 2 offers an alternative
strategy, advocating development of a composite Lake Michigan and Lake
Superior broodstock and maintaining the Seneca broodstock. Strategy 2 allows
for increased numbers of adults to be retained per broodstock within existing
hatchery facilities to increase effective population sizes while simultaneously
maintaining the single greatest source of genetic diversity (among lake
basins) .................................................................................... 110

viii

INTRODUCTION

Historically lake trout (Salvelinus namaycush) of the upper Great Lakes were
abundant and biologically diverse. The size of the Great Lakes basin, heterogeneous
nature of the lakes, and contributions from multiple isolated Pleistocene glacial refugia
(Wilson and Hebert 1996) promoted geographical and eco-phenotypic variation among
lake trout stocks (Brown et a1. 1981; MacLean et a1. 1981; Goodier 1981). Based on
phylogeographic evidence, lake trout are believed to have recolonized the Great Lakes
region fi‘om three discrete glacial refugia, Beringian, Mississippian, and Atlantic (Wilson
and Hebert 1996). It is likely that during long periods of isolation in these refugia lake
trout became diverged genetically and likely evolved locally adapted suites of traits
related to resident habitats. High levels of phenotypic diversity historically present in
lake trout in the upper Great Lakes may in part reflect the diversity of ancestral forms and
physiological features formerly isolated in glacial refugia (Wilson and Hebert 1996).

Discrete lake trout stocks were differentiated by spawning time and location, body
type and coloration, and occupancy of different water depths (Goodier 1981). Numerous
anecdotal accounts describe the diversity of lake trout populations in Lake Michigan,
Lake Huron, and Lake Superior and add testimony to the status of different lake trout
stocks to the Great Lakes fish community (Thomson 1883; Goode 1884). Three distinct
lake trout types were recognized from Lake Michigan: the shallow water “Mackinaw” or
lean trout, a deep-water “fat” or siscowet form, and a “Green Bay” variety of lake trout
that has been described in accounts of historical commercial fisheries (Brown et al.
1981). Variable phenotypes of lean trout, such as the “moss trout,” were also described

in Lake Michigan (Brown et al. 1981). As many as twelve phenotypes are believed to

have existed in Lake Huron alone (Eshenroder 1995) and as many or more were thought
to have been present in Lake Superior, particularly in waters around Isle Royale
(Rakestraw 1964). Goodier (1981) advocates evidence of at least four general types of
lake trout in Lake Superior; leans, siscowets, humpers, and half-breeds, noting evidence
of river-run lake trout in Canadian tributaries to Lake Superior.

The lean lake trout is a slender lake trout with a streamlined shape that inhabits
inshore waters (<70m). Lean lake trout spawn in shallow near-shore waters (<18m)
during the months of October and November (Goode 1884). The siscowet lake trout has
a more robust body and higher body fat content, and inhabits deeper offshore waters (70-
150m) compared to the lean lake trout (Goode 1884; Eshmeyer and Phillips 1965).
Siscowet populations have been observed in spawning condition at various times of the
year (Eshmeyer 1957; Bronte 1993). The humper lake trout resides near isolated offshore
reefs (or “humps”) commonly surrounded by water deeper than 100m (Rahrer 1965).
Phenotypically, humper lake trout are intermediates of leans and siscowets and possess
intermediate levels of body fat (Eshmeyer and Phillips 1965). Humper lake trout have a
restricted size distribution, mature at smaller sizes than leans and siscowets, and are long-
lived (Rahrer 1965; Bumham-Curtis and Bronte 1996). Humpers spawn in September,
sometimes as early as August (Rahrer 1965).

Goodier (1981), described an intermediate form of lake trout (“half-breeds”) that
was commonly captured at depths shallower than those inhabited by the siscowet lake
trout. Half-breeds possessed characteristics resembling those of both leans and siscowets

(Krueger and Ihssen 1995). Half-breeds are associated with specific areas of Lake

Superior consistent with areas of transitional depth (Goodier 1981), with sympatric or
spatially overlapping populations of leans and siscowets.

Phenotypic variation among lean, siscowet, and bumper lake trout could reflect
adaptations to local environmental regimes (i.e., water depth and temperature).
Experimental evidence suggests that variance in physical traits such as gas retention
(Ihssen and Tait 1974) and fat content (Eshmeyer and Phillips 1965), related to existence
at different water depth is heritable and not a plastic response to variable environments.
Genetic data collected from wild lake trout populations in Lake Superior reveals that
significant differences in allele frequency exist among populations within and between
lake trout phenotypes (review in Krueger and Ihssen 1995).

Through the 19505 and early 19603, a combination of habitat degradation from
pollution and eutrophication, overfishing, and the invasion of the sea lamprey decimated
lake trout populations throughout the upper Great Lakes (Cornelius et al. 1995; Elrod et
al. 1995; Eshenroder et a1. 1995; Eshmeyer 1957; Holey et al. 1995; Hansen et a1. 1995).
Wild lake trout populations were completely extirpated from Lake Michigan (Eshmeyer
1957) and US. waters of Lake Huron. Isolated remnant wild lean populations survived
in Georgian Bay of Lake Huron (Berst and Spangler 1973). Timely reductions in both
fishing intensity and sea lamprey abundance likely prevented extirpation of wild
populations in Lake Superior. Remnant wild populations exist around Isle Royale, the
Apostle Islands, Caribou Reef, and Stannard Rock (Figure 1; Rahrer 1965; Swanson and
Swedberg 1980; Curtis 1990; Hansen et a1. 1995). Of the historical diversity that once
existed, only a few recognized types (lean, siscowet, half-breed and humper) remain

(review in Krueger and Ihssen 1995), and only in Lake Superior.

In 1955, the Great Lakes Fishery Commission was established to facilitate efforts
to control the sea lamprey and restore lake trout populations in the Great Lakes. A major
emphasis of the lake trout restoration effort has been placed on stocking of offspring from
domestic lake trout hatchery strains (Fetterolf 1980). Currently six hatchery broodstocks
are maintained in the US. Fish and Wildlife Service hatchery system, and stocked in US.
waters of the upper Great Lakes. Broodstocks include the Isle Royale (SIW), Apostle
Islands (SAW), Marquette (SMD), Green Lake (GLW), Lewis Lake (LLW), and Seneca

Lake (SLW) broodstocks (Figure 1; Krueger and Ihssen 1995).

Chapter 1
INFLUENCES OF HATCHERY PRACTICES ON THE PRESERVATION OF

GENETIC DIVERSITY IN LAKE TROUT HATCHERY BROODSTOCKS USED FOR
RESTORATION EFFORTS IN THE UPPER GREAT LAKES

The overall goals of this chapter are: 1) To provide an overview of the history of
lake trout hatchery broodstock development and management and how management
practices have impacted genetic diversity of lake trout in hatcheries and of progeny
released into the Great Lakes, and 2) To examine specific areas of the broodstock
program (outlined in Figure 2) that can cause changes in gene frequency and levels of
genetic diversity and relatedness. We used molecular genetic techniques to evaluate the
effects of hatchery practices at three stages of the lake trout broodstock program. We
evaluated how effectively the genetic characteristics of wild populations were retained
during initial development of captive broodstocks (Stage 1). Hatchery broodstocks
should represent the genetic and ecological diversity present in wild populations. We
documented changes in allele frequencies and levels of genetic diversity that occurred as
successive generations of broodstocks were developed and during juvenile production
(Stage 2). Allele frequencies and levels of genetic diversity should not change
appreciably between broodstock generations. We examined several components of the
broodstock production program by evaluating broodstock spawning records,
documenting changes in allele fiequency and in levels of genetic diversity between adults
and progeny, and estimating the effective number of breeding adults (Nb). Allelic
frequencies and levels of genetic diversity should not change appreciably between adults

and juveniles. Broodstock management should maximize the number of adults

contributing progeny and minimize male and female reproductive variance to maximize
the effective number of breeders. Finally, we evaluated management practices related to
collection and distribution of fertilized gametes (Stage 3). Ideally, eggs and juveniles
should be collected, distributed, and released to maximize the number of adults spawned

across the entire spawning period.

History of Hatchery Broodstocks

Hatchery broodstocks maintained by the US. Fish and Wildlife Service constitute
a large portion of the remaining lake trout diversity in the upper Great Lakes (this study)
and form the basis for lake trout restoration in US. waters of Lake Michigan (Holey et al.
1995) and Lake Huron (Eshenroder et a1. 1995). Two of these broodstocks, LLW
(V isscher 1983) and GLW (Kincaid et a1. 1993), represent remnant genetic vestiges of
Lake Michigan populations. The SMD broodstock represents what remains of extirpated
near-shore lean populations in Michigan waters of southern Lake Superior (Krueger et a1.
1983). These broodstocks also may still possess genetically determined suites of co-
adapted traits (e.g., ecological and behavorial traits) characteristic of their respective
source populations (Krueger et a1. 1983). These co-adapted gene complexes are
important because they provide a blueprint for survival within the habitat that each source
population occupies. The historical background pertaining to establishment and
perpetuation of each lake trout broodstock provides important predictive potential of the
likelihood of changes in genetic characteristics of each broodstock and for identification

of factors that are of importance for the preservation of genetic diversity.

Apostle Island (SA W).- Collections were made from a wild lake trout population
from Gull Island Shoal in the Apostle Islands (eastern Lake Superior) over a five-year
period to produce five broodstock year classes (1986, 1987, 1996, 1998 and 1999). Wild
fish (2100), were spawned employing a male to female ratio of 2:1 and 1:1 to develop the
1986 and 1987 broodstock year classes. The 1993 and 1994 broodstock year classes
were developed from reciprocal crosses of the 1986 and 1987 broodstock year classes
from the matings of 138 individuals and greater than 120 individuals respectively (D.
Bast, Iron River National Fish Hatchery, personal communication; S. Schram, Wisconsin
Department of Natural Resources, personnel communication).

Green Lake (GLW: The Green Lake strain is one of two broodstocks (together
with the Lewis Lake broodstock) that originated from historic lake trout populations of
Lake Michigan. GLW is the oldest of the six domestic broodstocks and likely the most
influenced by hatchery practices. Between 1886 and 1944, Green Lake, Wisconsin was
stocked with lake trout obtained from spawning reefs in southern Lake Michigan and
Green Bay. In 1958 and 1959, Green Lake lake trout were spawned to form the original
Green Lake strain broodstock year classes 1959 and 1960. During 1958 and 1959, 187
(20 females and 167 males) and 309 (78 females and 231 males) individuals respectively
were spawned to create these two broodstock year classes. However, the genetic
integrity of these two broodstock year classes is suspect. Lake trout stocked into Green
Lake after 1952 were of Lake Superior origin and although identifiable by fin clips, some
individuals may have been spawned during the development of early broodstock year
classes. A large number of j uveniles stocked in 1944 may have originated from Lake

Superior populations and a large portion of this the 1944 year class likely contributed to

the 1959 and 1960 year classes. Additionally, the 1959 and 1960 broodstocks were
combined with males of Lake Superior origin to form a composite broodstock year class
referred to as the 1959/1960 Green Lake broodstock. Introduced Lake Superior males
were believed to count for as much as 10% of the Green Lake broodstock males (Krueger
et al. 1983). The1959/ 1960 Green Lake broodstock was used to create a 1965 year class
that was subsequently back-crossed with the 195 9/ 1960 parental broodstock to produce
the 1975 year class. The 1975 year class in turn was used to produced the 1984 year
class. The 1959/1960 Green Lake broodstock and the 1965 year class were eliminated in
1975. The 1975 Green Lake year class now resides in the Genoa National Fish Hatchery
(Coberly and Horall 1982; Krueger et a1. 1983).

During, 1970-1975, juveniles fiom crosses of the 1959/1960 Green Lake
broodstock were stocked in southern Lake Michigan. Individuals from the 1970-1975
production years were the primary source of the contemporary Green Lake strain
(Coberly and Horall 1982; Krueger et a1. 1983). Four Green Lake year classes, 1986,
1987, 1988, and 1989, were derived from the spawning of hatchery released lake trout
sampled fi'om Julian’s Reef and Black Can Reef in southern Lake Michigan. The 1986
broodstock year class was composed of separate broodstock lots, 86A and 86C.
Individuals spawned for each broodstock lot were 9 and 34 respectively. Eleven
individuals were spawned to create the 1987 broodstock year class. All year classes
except for 1987 and a small fraction of 1986 (N<100), were eventually lost due to
hatchery mechanical failures. As a result, the contemporary 1992 and 1993 Green Lake
year classes were derived from crosses of the remaining individuals of the 1986 and 1987

year classes, and the 1984 year class. Seven females from each year class were spawned

1:1 with males from each year class. Juveniles from each paired mating and within each
year class spawning combination were segregated, which effectively created 63 different
but related families. Juveniles were sampled equally across families to create the current
Green Lake 1992 and 1993 broodstock year classes (Kincaid et al. 1993).

Isle Royale (SIW).- The original Isle Royale broodstock was developed from a
wild population sampled from seven sites around Siscowet Bay, Isle Royale (northern
Lake Superior) (D. Bathel, Minnesota Department of Natural Resources, personal
communication). Four broodstock year classes, 1981, 1982, 1984 and 1986, were
developed from gametes taken from 56, 76, 126 and 54 individuals respectively. A male
to female spawning ratio of 2:1 or 1:1 was employed. A 1989 broodstock year class was
developed fiom crosses between 212-306 adults of the 1981 and 1982 year classes. A
1993 year class was developed from crosses of 192 adults from the 1982, 1984 and 1986
year classes. Approximately 371 and 469 individuals are maintained for the 1989 and
1993 year classes respectively (D. Bast, IRNFH, personal communication).

Lewis Lake (LL PW.- Lake trout from northern Lake Michigan were sampled in
1889 from several shallow reefs near Beaver Island. Of the nearly 5,000,000 eggs
collected, 100,000 were designated for stocking in Yellowstone National Park, Wyoming.
As few as 15-20 females could have contributed to this original egg take. Of greater than
42,000 fry, 12,010 were stocked into Lewis Lake. The remaining fry were stocked into
Shoshone Lake. In 1941 , an additional 5,890 fish of unknown origin were stocked into
Lewis Lake. The initial LLW broodstock was developed at the Jackson Hole National

Fish Hatchery (NFH) in 1983 (Krueger et a1. 1983; Visscher 1983). The 1989 and 1991

year classes at the Pendill’s Creek NFH were derived from 136 individuals of the captive
broodstock at Jackson Hole NFH (Kincaid et a1. 1994).

Marquette-Superior (SMD).- The Marquette broodstock represents extirpated lake
trout populations from in shore reefs in Keweenaw Bay, Lake Superior. From the 19603
through the 19803, the Marquette strain was the predominant strain stocked in Lake
Michigan. The initial Marquette broodstock was developed in 1948 from wild fish
captured in Marquette Harbor, Copper Harbor and an unidentified reef east of the
Keweenaw Peninsula. Crosses among individuals of this year class produced five
additional domestic year classes (1954, 1955, 1956 and 1957). A 1962 year class was
produced from crosses between adults from the 1954 and 1955 year classes. A 1963 year
class was produced from crosses among 1950’s year classes (Coberly and Horall 1982;
Krueger et al. 1983).

During the 1960’s the Marquette strain became partially integrated with lake trout
from the Green Lake (GLW) strain and hatchery fish derived from a wild remnant lake
trout population near Gull Island Shoal in the Apostle Islands (western Lake Superior).
Individuals from the 1959 GLW year class in addition to individuals from the 1955 and
1957 Marquette broodstock year classes were utilized to create the 1968 Marquette year
class. A 1956 year class, derived from Apostle Island lake trout, created a 1965 year
class that was directly integrated into the Marquette broodstock. Additionally, two
broodstock year classes, created purely from the 1970 and 1969 broodstock year classes,
were created during this time (Cobberly and Horall 1982; Krueger et a1. 1983).

Six additional year classes were created prior to the current Marquette broodstock.

These year classes include the 1974,1975,1976,1977,l978 and 1981. The 1976

10

broodstock was partially developed from the Apostle Island 1965 broodstock. The
history of the 1976 year class is of particular interest since crosses among adults within
this one year class were used to create the current 1988 year class. A 1987 year class is
also maintained (Kincaid et a1. 1994).

Seneca Lake (SLW: The Seneca Lake strain is the only strain originating from
outside of the upper Great Lakes basin. This broodstock was developed from deep—water
lake trout populations from Seneca Lake, New York. Evidence of successful
reproduction (Grewe et a1. 1994) and the propensity of Seneca Lake strain fish to avoid
sea lamprey predation (Eshenroder et a1. 1995), has made this strain an attractive option
for supplemental stocking in the upper Great Lakes.

In 1987 and 1992, respectively 137 (49 females and 88 males) and 315 (105
females and 210 males) wild lake trout from Seneca Lake were spawned to develop the
1987 and 1992 SLW year classes. Of the 107,000 eggs collected and fertilized fiom
adults in 1987, 60,000 survived to the fingerling stage, of which approximately 900 were
collected and maintained at Pendill’s Creek NFH as the 1987 SLW brodstock year class.
A total of 98,990 eggs were collected and fertilized from adult lake trout sampled in
1992. Of these 98,990 eggs, 70,000 survived to the fingerling stage. Approximately 500
fingerlings were collected to develop the 1992 SLW broodstock year class maintained at
Pendill’s Creek NFH (D. Blick, Allegheny NFH, personal communication).

Traverse Island (STW:- In 1996, 1997, and 1998, native lake lake trout from
near-shore waters around Traverse Island (Lake Superior) were sampled and spawned to
develop the STW broodstock. Traverse Island was one putative site sampled in the late

1940’s to develop the original SMD broodstock. The 1996 broodstock year class,

11

analyzed in this study, was developed from 128 males and 64 females spawned at a 2:1
ratio. Eggs from each 2:1 male to female mating (54 families, some sets lost) were
maintained segregated from one another. A total of 5,361 eggs were sampled fi'om these
families (100—131 from each family), of which, 4,088 were used to create the 1996
broodstock year class. This year class is maintained by Pendill’s Creek National Fish
Hatchery (M. Donofiio, Keweenaw Bay Indian Community Natural Resources
Department, personal communication).

The 1997 broodstock year class was developed from 24 females and an unknown
number of males sampled from Traverse Island waters in 1996. The spawning method
used to create this broodstock year class is unknown (M. Donofrio, Keweenaw Bay
Indian Community Natural Resources Department, personal communication). In 1997,
66 lake trout (33 females, 33 males) were collected and spawned 1:1 to create the 1998
STW broodstock year class. A total of 4,950 eggs were collected, 150 from each of 33
families. Of the resulting juveniles, 2,031 were designated for the 1998 STW broodstock
year class. This year class is maintained at Iron River National Fish Hatchery (Michael
Donofrio, Keweenaw Bay Indian Community Natural Resources Department, personal
communication).

In recognition of the genetic and ecological diversity represented by the different
lake trout broodstocks, the US. Fish and Wildlife Service recently adopted a hatchery
broodstock management plan (Holey 1997). The management plan recognizes the
importance of maintaining genetic diversity within and among existing broodstocks and
outlines hatchery practices that should be adopted to minimize loss of genetic diversity.

Included in the plan are guidelines for identification and sampling of source populations,

12

broodstock development and maintenance, and spawning. Recommendations outlined
include: 1) a minimum of 200 individuals sampled from source populations; 2) 1:1 paired
matings; 3) egg collections representative of numbers of individuals spawned and over
greater than 60% of the spawning period; and 4) broodstock production numbers
sufficient to minimize loss of genetic variation over three generations (i.e., 31% loss of
genetic variation or 150 individuals).

There is increased awareness of the degree to which hatchery management
practices can influence genetic characteristics of broodstocks and their progeny released
into natural environments (Allendorf and Ryman 1987, Waples et al. 1990, Waples
1991). The ability of hatcheries to conserve levels of genetic variation characteristic of
progenitor wild source populations hinges upon proactive hatchery management with
respect to several fundamental population genetics principles related to coancestry or
degree of inter-individual relatedness, inbreeding, genetic drift, and effective population
size.

Differences in genetic characteristics either between year classes or between
adults and progeny are a reflection of different spawning practices during production of
each cohort. Large breeding populations that were evaluated in this study were expected
to show minimal change in genetic characteristics across generations. Levels of
inbreeding and inter-individual relatedness, gene frequencies and genetic diversity (i.e.,
heterozygosity), and effective population sizes between lake trout cohorts were estimated
to measure the influences of the lake trout hatchery program on actual and potential

changes in these genetic characteristics.

13

Inbreeding results from matings between two related individuals. Inter-individual
relatedness or coancestry can accrue rapidly under circumstances routinely realized in
hatcheries (e.g., Figure 2). Inbreeding coefficients (F) of offspring are directly related to
the coancestry (0) of their parents. In the absence of immigration, as is typical for most
hatchery broodstocks, the accumulation of coancestry among individuals within groups is
affected by the manner in which male and female gametes are mixed as well as by the
means and variance in number of progeny produced by each mating (Chesser 1991a, b).
F and 0 will vary from cohort to cohort based on efficacy of spawning practices of each
spawning period. In closed populations, extreme values of coancestry are realized by
male polygamy (Chesser 1991 a, b), which is commonly realized in hatchery spawning
practices involving the pooling of gametes (Figure 3; Gharrett and Shirley 1985; Withler
1988). High coefficients of relationship among members of a domestic broodstock can
have profound impacts on levels of inbreeding and population fitness of progeny to be
released into the natural environment.

The. inbreeding coefficient, “F”, is a statistical measure of the rate of inbreeding
among individuals. In a finite population that is closed to immigration (e.g., a hatchery
broodstock), F will always increase. Inbreeding can be considered in the context of loss
of genetic variation in which increases in F over multiple generations can be compared to
expected levels of gene correlation within individuals (i.e., if all were mated at random).
Inbreeding affects genetic diversity by reducing heterozygosity in a population. The
number of homozygous individuals will increase with the number of individuals
possessing correlated genes (Busack and Currens 1995). As F increases across

generations, the fraction of heterozygotes in the population decreases (Kincaid 1983;

14

Tave 1993). Population fitness will decline as a result of decreased genetic variability
(Allendorf and Leary 1986) and inbreeding depression.

Inbreeding depression is the reduced “performance” measured for traits in a
population due to the expression of deleterious alleles (Kincaid 1983). The most
common traits affected by inbreeding depression are traits related to population
sustainability such as fecundity, gamete viability, juvenile size or other surrogate
measures of juvenile survival (Kincaid 1983). Numerous examples of reduced
population fitness related to inbreeding in aquaculture systems have been documented
(Kincaid 1983; Simon 1991; Tave 1993). Mating systems that avoid unequal sex ratios,
high variance in reproductive success, and consequently support large effective
population sizes will decrease the probability of incestual matings and minimize
inbreeding (Waldman and McKinnon 1993).

Genetic drift is the random change in gene frequencies over generations (Wright
1931). Drift is realized in all populations, but is most pronounced in populations of small
size. Genetic drift ultimately results in loss of genetic diversity and ultimately in fixation
of alleles. The probability that an allele becomes fixed or lost is dependent on its initial
fi'equency in a population. The rate at which alleles become fixed or lost is inversely
related to population size (Allendorf and Ferguson 1990).

An important principle to consider in the preservation of genetic diversity in
closed populations is the effective population size (Lande and Barrowclough 1987;
Nunney and Elam 1994). There are a numerous measures of effective population size
including variance effective population size, inbreeding effective population size, and the

coancestral effective population size. The variance effective population size (NC) is

15

defined as the number of individuals in a population that would give the same variance
(over time or between generations) in allele frequencies due to genetic drifi as in an ideal
population (i.e., equilibrium population, Wright 1931). Inbreeding and coancestral
effective population sizes are population sizes that would reflect the same rate of change
in heterozygosity and coancestry as an ideal population. Variance effective population
size estimate in this study is usually smaller than the actual number individuals in a
population (N) owing to reduced or fluctuating population size, variance in family size
(reproductive success), and unequal sex ratios (Lande and Barrowclough 1987; Frankham
1995)

Large effective population sizes are expected to minimize the effects of genetic
drift and inbreeding, thus conserving genetic diversity and maintaining population fitness.
Lower levels of genetic variation have been shown to be positively correlated with small
effective population size (review in Lande and Barrowclough 1987; F rankham 1996).
Genetic diversity (heterozygosity) is lost in a population under genetic drift at a rate equal
to 1/(2Nc) per generation. Decreased genetic diversity can lead to decreased fitness in the
form of smaller individual size, lower fecundity, smaller population size and increased
developmental deformities (Smith et a1. 1976; Allendorf and Leary 1986; Quattro and
Vrijenhoek 1989; Leberg 1990; Amos and Hoelzel 1992). Therefore, effective
population size is of importance for species of conservation or management interest.

The effective population size (Ne) differs from the effective number of breeders
(Nb) estimated in this study. The effective population size can be estimated directly using
ecological and demographic parameters of a population (N unney and Elam 1992), or by

independent means using temporal changes in allelic frequencies across generations

l6

(Waples 1991). Estimating Nc from ecological and demographic data for wild
populations is often difficult due to the paucity of background p0pulation data needed to
estimate parameters of ecological models (Nunney and Elam 1992). Even in captive
populations, pedigree and juvenile survival information may be unavailable or too costly
to obtain (Waples 1990). Therefore, indirect genetic methods have been advocated for
these situations. Estimation of effective population sizes using temporal variance in
allele frequencies for organisms with overlapping generations have been shown to be
biased (Jorde and Ryman 1995). The estimate of effective number of breeders (Nb)
utilized in this study is a proven method for estimating effective population sizes of
discrete, closed captive populations (Waples 1990).

In recognition of the importance of these genetic principles, recommendations
have been proposed to conserve genetic diversity in hatchery populations (see Appendix
I; Allendorf and Ryman 1987; Allendorf and Phelps 1980; Busack and Currens 1995;
Hynes et al. 1981; Kincaid 1995, Kreuger et al. 1981; Simon 1991; Kapuscinski and
Jacobson 1987; Waples et al. 1990; Gharrett and Shirley 1985; Withler 1988 ). Figure 2
illustrates the practices and issues that are involved in a hatchery broodstock program.
Genetic diversity may be progressively lost throughout a hatchery program, but diversity
may also be affected at points within discrete periods of the broodstock program. We
decompose the lake trout broodstock program into three major stages to better articulate
how genetic diversity may be affected by specific practices within the lake trout
broodstock program. These stages are: Stage 1 - the broodstock development stage in
which new broodstocks are developed by sampling fi’om source populations; Stage 2 - the

maintenance and production stage in which broodstocks are perpetuated and juveniles

l7

produced for stocking; Stage 3 - the collection and distribution of gametes and juveniles
in which fertilized eggs or juveniles are collected and distributed to rearing facilities and

specific stocking locations.

METHODS
Broodstocks are composed of several year classes that are defined by production
year. The broodstocks analyzed in this study were composites of multiple broodstock
year classes. To simplify statistical analyses and comparisons, we combined samples
across year classes based on their broodstock association. We will refer to these separate
year classes as simply the LLW, SIW, SAW, SLW, GLW, and SMD broodstocks. For
the year classes considered in this study for each broodstock, refer to the History of

Hatchery Broodstocks section of the introduction.

Stage 1: Broodstock Development

Sample Collection.- The LLW, SIW, and SAW hatchery broodstocks were
compared to their wild populations. Samples of the three captive hatchery broodstocks
were sampled by hatchery personnel during routine spawning events in the fall of 1998.
The LLW broodstock was sampled from Pendill’s Creek/Hiawatha National Forest Fish
Hatchery in Michigan and the SIW and SAW hatchery broodstocks were sampled from
the Iron River National Fish Hatchery in Wisconsin. Samples consisted of fin clips
(~1cm 2) removed from caudal fins and stored individually in 1.5 ml vials containing 1

m1 high salt buffer (4M urea, 0.2M NaCl, 0.1M Tris-HCL, 0.5% Sarcosine, lOmM

18

EDTA). Two hundred adults were sampled from the SAW, LLW, and SIW broodstocks.
Fin clips were stored at -20°C until analysis.

We also sampled the LLW, SIW, SAW hatchery broodstock source populations.
Lake trout representing the LLW source population (N=77) were sampled from Lewis
Lake, WY, by USFWS personnel at the Yellowstone Fisheries Resource Office from
Lewis Lake, Wyoming. Isle Royale wild lean lake trout, representing the source
population for the SIW broodstock (N=119), were sampled in the summer and fall 1995.
Wild lake trout sampled from Gull Island Shoal in the Apostle Islands, WI (N=68),
represent the source population for the SAW broodstock. Isle Royale and Lewis Lake
samples consisted of liver tissue preserved in ethanol. The Apostle Island wild samples
consisted of liver and scale tissue. Scale tissue was sampled in 1999 from archival
collections located at the Wisconsin Department of Natural Resources Bayfield station.
We used samples that were collected from the same or adjacent locales that were sampled
to developed the LLW, SAW, and SIW broodstocks.

DNA Extraction.- DNA extraction of liver and fin tissue was performed using a
proteinase K digestion and a modified Puregene extraction protocol (Gentra, Inc.,
Minneapolis, MN). DNA was resuspended in 50 ml of TE buffer (10 mM Tris-HCL, pH
8.0, 1 mM EDTA). Fluorometry was used to determine DNA concentrations. Prior to
flourometry, RNAse (2 ul of 20mg/ul stock) was added to each sample. One hundred
nanograms of DNA was used for each PCR reaction.

A Chelex procedure was utilized for DNA extraction from scale samples. Scales
(3-5 per individual) were added to 250 pl of a 5% Chelex and 10 mM Tris-HCL (pH 7.5-

8.0) suspension. Scales were digested overnight with 3 ul of proteinase K. Proteinase K

19

was subsequently inhibited by heat denaturation at 95°C for 5 minutes and samples were
centrifirged at 14000 rpm for 10 minutes, the resulting supernatant was removed and 2.5
ul of the supernatant was used for each PCR reaction.

Microsatellite Screening.- Microsatellite DNA markers possess qualities well
suited for population genetic analyses (Beaumont and Bruford 1999, Scribner and Pearce
2000). Microsatellite loci reside in noncoding regions of the genome and therefore are
not evolving under the constraints of natural selection, nor do they code for polygenic
traits (i.e., phenotypic and behavioral traits important for fitness). Microsatellite markers
are useful for analysis of the influences of hatchery practices on population genetics
dynamics (i.e., genetic drift and inbreeding). Allelic diversities at microsatellite loci are
typically greater than levels for other genetic markers, allowing greater sensitivity in
investigations of genetic drift and inbreeding. However, measures of gene diversity and
loss of diversity for loci underlying polygenic traits (i.e., spawning time, fecundity) will
not necessarily provide measures of diversity concordant with neutral markers (Hard
1995; Pfrender et al. 2000). Microsatellites are ideally suited to chronicle the
transmission of genes across generations. In addition, microsatellite markers are readily
amplified in vitro using polymerase chain reaction techniques, from small amounts of
tissue, and eliminating the need for invasive and destructive sampling. Microsatellite
markers used in this study were previously developed from brook trout (Salvelinus
fontanalis) (Sfol, Sf012 and Sf018; Angers et a1. 1995), sockeye salmon (Onchorynchus
nerka) (One/19 and 0ne,u10; Scribner et al. 1996), pink salmon (Onchorynchus

gorbuscha) (OgoIa and 0goIc; Olsen et al. 1998), bull trout (Salvelinus confluentus)

20

(Sea/119; Taylor et a1. 2001), and Atlantic sahnon (Salmo salar) (Ssa85; O’Reilly et al.
1996)

PCR reactions were performed in 25 pl volumes using conditions provided by the
respective authors. The PCR profile involved a single 2 min denaturing step at 94°C,
followed by 30 cycles of a 1 min denaturing step at 94°C, 1 min annealing step at the
appropriate temperature and 1min extension step at 72°C. PCR profiles for scale
extracted DNA required 35 cycles. Annealing temperatures were as follows: 60 °C for
Sfol and Sf012, 56 °C for Sf018 and Ssa85, 54 °C for 0ne,u9, 52 °C for 0goIa, 46 °C for

0nep10 and Scop19 and 48 °C for 0goIc. PCR products were screened using 6%
polyacrylamide vertical gels. Products were visualized by a Hitachi FMBIO 11 Multi-
View scanner and associated software. Microsatellite fragments were sized manually
using a 20 bp internal lane standard. Several individuals of known genotype served as
positive controls in each gel for standardization.

Statistical Analysis.- One of the primary goals in development of domestic
broodstocks is to minimize loss of genetic diversity and changes in allele frequency
(Holey 1997). A variety of measures were employed to evaluate changes in genetic
diversity between source and hatchery populations. Estimates of allele frequencies and
expected and observed heterozygotic diversity were performed using BIOSYS I
(Swofford and Selander 1981). Exact tests (Raymond and Rousset 1995a) of significance
of differences in allele frequency between wild source and hatchery broodstock samples
was performed using GENEPOP (V3.1b; Raymond and Rousset 1995b). The sequential
Bonferroni method was used to derive nominal significance levels for multiple testing

(Rice 1989). Allelic richness was calculated for each population using the program

21

CONTRIBUTE (Petit et a1 1997). Allelic richness provides a measure of the number
alleles per locus standardized for differences in population sample size.

Hatchery systems and management practices can affect gene correlations within
individuals (inbreeding coefficients, F) or among individuals (coancestry between
individuals, 0). An understanding of the magnitude of these effects is critical for
predicting change or rates of accrual of gene correlations and effective size for successive
generations of captive populations as well as of individuals introduced into natural
populations (Ryman and Laike 1992). Inbreeding coefficients relate the correlation of
genes within individuals to that expected if individuals had mated at random. The
expected level of correlation is one minus the total variation at the same levels. We
estimate expectations for fixation indices by calculation of observed and expected
heterozygosities [ F=l-H1 and 0=l-Hs] where F is the inbreeding coefficient, 0 is the
estimate of coancestry, and H1 and H5 are observed individual heterozyosity and mean
expected heterozygosity.

Estimates of coancestry (a measure of the proportion of genes shared between
individuals that are identical-by-descent) can be effectively estimated using the surrogate
coefficient of relatedness in the absence of pedigree information (rxy; Queller and
Goodknight 1989). The average coefficient of relatedness is an appropriate surrogate
measure for common pedigree measures (Blouin et al. 1995). The coefficient of

relatedness we used is an unbiased measure of degree of inter-individual relationship:

22

(1)

where px refers to the frequency of a given allele possessed by an individual “x ” (i.e. 0,
0.5, or 1), py refers to the frequency of the given allele within any individual “y” being
compared to individual “x”. The mean frequency of the given allele for the entire
population under question is p*. For each pair-wise estimate, the values are summed
over all loci, k = 1,. . .,nk, and alleles at each locus, a = 1,. . .,n,,. This statistic provides an
estimate of the proportion genes between individuals that are identical-by-descent.

Pair-wise estimates of coefficients of relatedness can range from —1 to 1. A value
of 1 represents complete correlation between two individuals and 0 represents no
correlation. Individuals with r,y 2 0.5 share alleles at levels equivalent to expectations for
firll siblings. Individuals correlated at the level of 0.25 are correlated at the level of half
siblings. Individuals with negative rxy values share fewer alleles than expected based on
the population average.

Pair-wise coefficients of relatedness values were calculated using KINSHIP 2.1
(Queller and Goodnight 1989). Individual pair-wise coefficients of relatedness were
summarized by calculating an average coefficient of relatedness for every source and
hatchery population. Calculating the average coefficients of relatedness for source
populations and hatchery broodstocks summarized individual pair-wise estimates. If
hatchery spawning methods are effective in equalizing contributions by adults to the next
generation, coefficients of relatedness should not differ significantly. An increase in

average coefficients of relatedness would suggest that disproportionately fewer adults

23

contributed to progeny, thus increasing coancestry among progeny. The significance of
differences in distributions of pair-wise coefficients of relatedness between the wild
source and complementary hatchery broodstocks were tested employing Mann-Whitney
U tests (SAS, 1999). The Mann-Whitney U test is a nonparametric test of the
significance of differences in location (i.e., the average) of non-normal distributions
(Blaisdell 1993).

The average coancestry coefficient does not provide a complete picture of the
levels of coefficients of relatedness. Another important consideration is the distribution
of the coefficient of relatedness values. Populations with similar average coefficients of
relatedness may possess important differences in distributions (i.e., skewness and
kurtosis, or frequency of individuals related at the full sibling level, 0.50). Frequency
histograms were developed to visually compare the skewness and kurtosis between
source and hatchery broodstock population estimates. This is important, because
distributions of coefficients of relatedness can have similar averages, but differ in the
distribution of the frequencies of pair-wise coefficients of relatedness estimates. The
frequency of inter-individual comparisons yielding rxy values that were significant at
P3005 consistent with the level of full siblings, was estimated by KIN SHIP 2.1 (Queller

and Goodnight 1989).

Stage 2: Broodstock Maintenance and Production
Consideration of the means by which broodstocks are perpetuated and juveniles
are produced for stocking is an important concern for the preservation of genetic

diversity. Any genetic differences derived from Stage 1 are likely to be accentuated

24

during the perpetuation of a broodstock if practices to minimize changes in genetic
diversity are not employed (Figure 2). This can result in changes in allele frequency and
loss of genetic variance in the broodstock compared to the wild source population. These
differences may subsequently be compounded in production juveniles or accentuated in
juveniles due to inefficient spawning practices (Figure 2). Juveniles stocked into the wild
are often the sole means of rehabilitation. Thus, it would be prudent to ensure that
juveniles reflect as much of the genetic diversity of wild source populations as possible.
To examine the effectiveness of management for preservation of genetic diversity
in lake trout broodstocks, we estimated allele frequencies and measures of genetic
variability for the SMD broodstock, where long-term multi-generational data were
available. The 1981 SMD broodstock year class, the current SMD broodstock (1987 and
1988 year classes), and the newly developed STW broodstock were the three SMD
generations evaluated. Conditions that commonly influence changes in genetic
variability across broodstock generations (i.e., numbers of adults spawned and
fertilization conditions) are those most likely to reduce the effective population size.
Issues pertaining to the production of j uveniles for restoration and enhancement
purposes are also investigated under Stage 2. Lake trout broodstocks are spawned
episodically throughout the spawning season. Lake trout at Pendill’s Creek NFH are
spawned using 5:5 male to female ratio. The gametes of five males are pooled and
combined with the pooled gametes of five females. A slightly different methodology is
used at the Iron River NFH where the gametes of two groups of five females are pooled
separately. Gametes from two separate groups of five males are also pooled separately.

Half the volume from one of the pools of male gametes is combined with one of the

25

female gamete pools. Half the volume of the other male gamete pool is combined with
the other female gamete pool. After a period of time (several minutes) the remaining
volumes of the male gamete pools are reciprocally added to the female gamete pools.
This effectively results in a 10:5 male to female mating system. Due to disparities in
adult male to female ratios, males are often used multiple times. In the Iron River
Hatchery males were rarely reused. In the Pendill’s Creek Hatchery males were reused
up to two times. Within the LLW, SLW and SMD broodstocks 77%, 56% and 15% of
the males were reused at least once. Eggs and juveniles developed by lake trout adults
spawned episodically throughout the spawning season were retained together by
spawning period through the swim-up stage until they were mixed with juveniles from
other spawning periods. For our analysis, we will refer to juveniles sampled during a
certain spawning period, and the adults that produced them, as “lots”.

During a spawning season, hatchery personnel will spawn ripe females and a
comparable number of males to fertilize the eggs. However, due to hatchery practices,
portions of a spawning adult “lot” may not contribute gametes to a subsequent juvenile
population. All eggs or juveniles of entire adult lots are at times discarded by hatchery
personnel due to disease or nonviability (i.e., green eggs). Nonviability of eggs of entire
spawning lots can be due to the lack of uniformity in egg maturation. The adults
associated with such removals are eliminated from the “potential contributers” to the
juvenile gene pool. Elimination of gametes or juveniles can reduce the effective
population size (and the representation of early, middle, and late spawners) irrespective

of spawning methodology or duration.

26

Sample Collection.- Information regarding spawning dates, numbers spawned,
and gamete take procedures were provided by Pendill’s Creek/Hiawatha and Iron River
National Fish Hatchery personnel. Information was also obtained regarding numbers of
juveniles produced, proportions distributed (sent to rearing stations or other programs),
and identification of spawning lots from which all juveniles were culled.

Hatchery broodstock adults and offspring were genotyped to estimate the
effective number of breeding adults. The LLW broodstock was not evaluated because
LLW juveniles were not available. Two hundred adults were sampled from the SAW,
GLW, and SLW broodstocks, and 166 from the SMD broodstock were sampled.

Juveniles from hatchery spawning periods were collected in the spring of 1999
and 2000 (SMD only). Juveniles were not random samples of the entire spawning
population but a random sample from each “lot” corresponding to an identifiable number
of adults spawned during a certain time period. Juveniles of the SMD broodstock were
the F1 progeny of the spawned SMD adults from 1999. All juveniles were collected as
swim-up fi'y fiom hatchery tanks using dip nets. Effort was made to limit sampling bias
by collecting equal numbers of juveniles from within tanks (head, middle, and foot) and
between tanks. Juveniles were stored whole in 95% ethanol at room temperature.
Juveniles of the SAW, SIW and GLW broodstocks were sampled from Iron River NFH.
Juveniles of the SMD, and SLW juveniles were collected fiom Jordan River National
Fish Hatchery, Michigan. Several hundred juveniles were collected from each strain with
the exception of the GLW juveniles of which 114 were collected.

Analysis of inter-generational changes in genetic characteristics was performed

using samples from the 1981 SMD broodstock year class, the current SMD broodstock,

27

and the newly developed STW broodstock. Thirty-eight liver tissue samples from the
1981 SMD broodstock year class were collected randomly from fish taken out of
production in 1998 and stored in 95% ethanol. The current SMD broodstock was
collected as described earlier, and fin clips from 60 individuals from the STW broodstock
were collected by Pendill’s Creek/Hiawatha NFH personnel in 1999, preserved in scale
envelopes, and stored at 0°C.

DNA Extraction.- Juvenile tails, removed posterior to the insertion of the anal fin,
were utilized for DNA extraction. The DNA extraction protocol for the adults and
juveniles in this analysis was performed as described under Stage 1.

Microsatellite Screening.- Three microsatellite loci, Sf018, Scoul9 and Ssa85
were utilized for estimation of the effective number of breeding adults (Nb). All nine loci
were employed for the intergenerational analysis of the SMD broodstock.

Statistical Analysis.- The fertilized eggs from each spawning lot (egg lot) were
maintained segregated from the eggs of all other spawning lots. Egg lots developed fi'om
adults spawned early or late in the spawning period may be “green” (females spawned
too early) or overly “ripe” (females spawned too late). These egg lots typically exhibit
low viability and are eliminated by hatchery personnel. This practice effectively removes
a portion of the adults as “potential contributers”. We documented the total number of
lake trout spawned and numbers of hatchery adults whose eggs were eliminated. We
subtracted the hatchery adults that failed to contribute to the juvenile population (due to
their eggs being eliminated) from the total number of individuals spawned. Remaining
individuals were designated “potential contributers”. A ratio of the potential contributers

to total number of lake trout adults spawned was calculated.

28

Genetic characteristics including observed (H0) and expected (H...)
heterozygosities, allele frequencies, allelic richness and tests of significance were
calculated as described in Stage 1. Average coefficients of relatedness (er) and
distributions of pair-wise estimates of coefficients of relatedness compared between
hatchery adults and subsequent juveniles in Stage 2 were calculated using three
microsatellite loci, SfoI8, Scop19, and Ssa85. Convergence of rxy to actual levels of
relatedness is realized when large numbers of loci are assayed (Blouin et al. 1996). We
based estimates on few loci. However, as Blouin et al. (1996) point out, use of relatively
few loci simply decreases statistical power (i.e., the ability to reject the null hypothesis of
no relationship between individuals when non-zero coefficients of relationship exist).
Further, Leberg (1992) and Spencer et al. (2000) have shown that relatively few loci can
be used to document even subtle demographic changes within populations.

If spawning methods were effective in equalizing contributions of spawning
adults to progeny (i.e., minimizing reproductive variance) than the effective number of
breeders should be approximately the actual numbers spawned. Estimations of N, were
performed on specific lots of adults and their corresponding offspring. The same three
loci, Sfol 8, Scop19 and Ssa85 were used to estimate the Nb. Allele frequencies for the
adult and juvenile populations were calculated. For each hatchery broodstock, 111 to 198
adults and 174 to 207 juveniles were screened per broodstock to minimize sampling
variance. Between 99 and 113 GLW adults and juveniles were screened.

Estimates of N, were made using temporal variance in allele frequency (Waples
1989; eqn 8). This method utilizes a standardized method for calculating allele variance

(Fc) between adult and juvenile samples for a single locus:

29

 

1
7(— ,-=1(xi — yi)/2 — xiyi (2)

where F, is the average variance in allele frequency over k alleles between two
generations or samples. The allele frequency of generation 0 at allele i is represented by
x and the allele frequency at generation t for allele i is represented by y.

The Waples model assumes that populations are closed to migration, selection and
mutation are minimal, and the population is randomly mating. Estimates of Nb are based

on the Plan I sampling methodology (Waples 1989):

t
b ._. 2[Fc-1/(ZSo)-1/(ZSt)+l/N] (3)

 

The total adult population of size N is sampled before reproduction (generation 0)
and then juveniles are sampled at generation t =1. The number sampled at generation 0 is
represented by So (adults) and the number sampled at generation t is S, (juveniles).
Confidence intervals (95%) for estimates of Nb were calculated by estimating the

confidence interval of the overall variance (Fe) for each broodstock (Waples 1989).

Stage 3: Distribution and Stocking of Broodstock Juveniles
The final stage of the lake trout broodstock program involves the collection,
distribution and stocking of juveniles. Ideally, juveniles collected from spawning lots,

produced across the spawning season, should be mixed prior to distribution to rearing

3O

facilities or stocking sites to provide equal representation of production from the entire
broodstock. Egg takes, distributions, and stocking practices that do not attempt to
equalize representation of the contributing adults across the entire spawning period will
likely fail to capture the genetic diversity present in the hatchery broodstock or the wild
source population. In addition, if progeny from different spawning lots are not mixed the
levels of coancestry among juveniles will be disproportionately high, increasing the
possibility of future inbreeding and potentially reduced population viability.

Data Collection and Analysis.- Information regarding egg distribution was
provided by Pendill’s Creek/Hiawatha National Forest and Iron River National Fish
Hatcheries. Egg take and distribution records of lake trout eggs, developed from
hatchery broodstock spawning sessions in 1998 and 1999 (SMD broodstock), were
evaluated for representation of spawning lots and potential effects on genetic diversity
and coancestry of eggs distributed. The numbers and proportions of j uveniles for each
spawning lot distributed to given rearing facilities were calculated. Ratios of numbers of
spawning adults associated with juveniles distributed to each rearing hatchery to the total
number of adults spawned at the hatchery were calculated. This provided a measure of
the proportion of the total broodstock that would be represented by the juveniles

distributed to each rearing station or stocking site.

RESULTS
Stage 1: Broodstock Development
Comparisons of Wild Source Populations to Captive Broodstocks.- Estimates of

observed (Ho) and expected (He) heterozygosities were similar in wild source populations

31

and hatchery broodstocks (Table 1). Differences in average observed heterozygosity (Ho)
between LLW, SIW, and SAW and hatchery broodstocks and their source populations
were 0.040, -0.010, and 0.037 respectively. Differences in average expected
heterozygosity between hatchery broodstocks and wild source populations were 0.026, -
0.017, and 0.024 for the LLW, SIW, and SAW broodstocks and associated wild source
populations (Table 1). Expected heterozygosities (H,)were higher than observed
heterozygosities for all populations, suggesting a modest, though non-significant
heterozygote deficiency.

Estimates of average allelic richness were similar between hatchery broodstocks
and wild source populations. However, all hatchery strains showed significant
differences in allelic frequencies from their source populations (Table 1) suggesting
genetic drift or non-random sampling of the wild populations. Across nine loci, alleles
differing 25% in frequency were found between all hatchery broodstocks and source
populations (LLW, 15 alleles; SIW, 9 alleles; SAW, 7 alleles). Differences in allele
frequencies were greatest between the LLW hatchery broodstock and the Lewis Lake
source population. Alleles 221 and 219 for the 0goI c locus differed in frequency by
21% and 14.2% respectively between the LLW hatchery broodstock and the Lewis Lake
wild source populations. The LLW hatchery broodstock and Lewis Lake wild source
populations differed by 14.2% for the 171 allele at the SfoI8 locus. Allele frequencies
differed significantly between the LLW hatchery strain and the Lewis Lake wild
population at the 0neu9 (P<0.05) locus, Scou19 and 0goIc loci (P<0.01) and at the
Ugo] a locus (P<0.001). After employing Bonferroni multiple test adjustments, the

differences at the 0ne,u9 locus were no longer significant. Based on exact tests, the SIW

32

hatchery broodstock differed nominally significantly from the Isle Royale lean population
at the Sf018, 0neu10, Scou19 (P<0.05) loci, but none of these differences were
significant after Bonferroni adjustments. Highly significant differences were found at the
SfoI8 (P<0.001) and Scou19 (P<0.01) loci between the SAW hatchery strain and the wild
Apostle Island population and differences at both of these loci remained significant after
Bonferroni adjustments. These results suggest that the individuals sampled from wild
populations to develop broodstocks were sufficient in preserving allelic and genotypic
diversity (Table 1). However, due to sampling over short time periods or of few
numbers, appreciable drift in allele frequencies were observed (Table 1).

If spawning regimes successfully captured the diversity present in adult
broodstocks and reproduction was equitably distributed across all breeding adults,
average r,‘y and distributions of r,,y estimates (reflecting levels of inter-individual
relatedness or coancestry) would not be expected to differ between source and hatchery
broodstock populations. Pair-wise estimates of relatedness (rxy) for individuals within
each source population and hatchery strain were calculated. Mean rxy estimates and
distributions of pair-wise rxy estimates were compared for hatchery and wild populations
(Figure 4). Frequency histograms (Figure 4) reveal that distributions of rxy values were
not found to differ significantly (P<0.05). Average r,y decreased for two comparisons
(Lewis Lake vs. LLW and Isle Royale vs. SIW) and remained constant for the other
(Apostle Island vs. SAW). The frequency of significant pair-wise estimates at the firll-
sibling level also decreased slightly between source populations and hatchery
broodstocks (Table 1). Qualitative inspection of distributions of rxy values between

broodstocks and wild source populations revealed a slight change in kurtosis (increased

33

peak of distribution for hatchery broodstocks), and a decrease in the frequency of extreme
r,y values and showing convergence on the mean rxy (Figure 4). Estimates of inbreeding
coefficients were similar for all broodstocks compared to wild source populations (Table
1). Estimates of inbreeding coefficients were not statistically significantly different from
zero.

Overall, use of few adults spawned from wild populations mean that each
broodstock was founded from a sample characterized by low variance effective
population size, potentially resulting in relatively high variance in allele frequency
between wild progenitor populations and hatchery broodstocks. Gene correlations do not

appear to have increased as a consequence of broodstock initiation.

Stage 2: Broodstock Maintenance and Production

Intergenerational Comparisons of Marquette Broodstocks.- The SMD81, current
SMD broodstock, and the preproduction STW broodstock showed little divergence in
allele frequencies (Table 2). The only significant differences in allelic frequencies were
found between SMD81 and STW broodstocks at the Sfol and Sf012 loci (P<0.05) and for
the SMD and STW comparison at the Sf012 locus (P<0.01). The difference between the
STW and SMD broodstocks at the Sf012 locus was the only significant difference
observed after Bonferroni adjustments. The average observed heterozygosity for the
SMD81 broodstock (0.353) was lower than that for the SMD (0.374) and STW (0.360)
broodstocks. Estimates of observed heterozygosity (Ho) were below expected
heterozygosities (H) for both SMD81 and STW broodstocks. Allelic richness estimates

were similar for the SMD and STW broodstocks (3.0), but both were higher than the

34

average allelic richness of the SMD81 broodstock (2.7). Inbreeding coefficients (F) of
the SMD81 and STW broodstocks were high (0.089 and 0.112 respectively) relative to
the inbreeding coefficient of the SMD broodstock (-0.002) but none were significant
(P<0.05). Allelic frequencies and measures of genetic diversity appear to have remained
relatively unchanged among these three broodstocks. However, we see some evidence
for increased levels of inbreeding. The STW broodstock has the highest inbreeding
coefficient of all three broodstocks (0.112).

Comparisons of coefficients of relationship between the three broodstocks reveal
no evidence for significant changes in means or distributions (Figure 5). Average
coefficients of relatedness for the SMD81, SMD, and STW broodstocks were 0.035,

—0.008 and —0.047 respectively. Estimates of the proportion of inter-individual
coefficients of relatedness consistent with full sibling relations (PS 0.05) were 0.079 for
the SMD81 broodstock, 0.061 for the SMD broodstock (0.061), and 0.068 for the STW
broodstock (Table 2). Qualitative inspection of distributions of rxy values between
SMD81, SMD, and STW revealed little differences in distributions based on degrees of
kurtosis or skewness (Figure 5).

Comparisons Between Broodstock Adults and Juveniles.- Significant differences
in allele frequencies were observed between adult and juvenile samples for a number of
the lake trout broodstocks (Table 3). The SAW adult and juvenile populations differed
significantly (P<0.05) at the SfoI8 locus. Allele frequencies of the SLW adult and
juvenile populations differ significantly (P<0.01) for the Sco,u19 loci. Highly significant
differences at the Scou19 and Ssa85 loci and significant differences at the Sf018 locus

were found between the GLW adults and juveniles. Allele frequencies at the SfoI8 locus

35

differed significantly for the SIW adults and juveniles. No significant differences in
allele fi'equency were found between the SMD adults and juveniles. After Bonferroni
corrections, only significant differences between adult and juveniles for the SLW and
GLW broodstocks were observed. Ten instances of allele frequency differences of
greater than 5% were documented. Estimates of allelic richness generally decreased
between adults and juveniles for the SMD, GLW, and SIW broodstocks. The GLW
broodstock showed the highest proclivity for large differences in allele frequency
between adults and juveniles (for each of the three loci sampled; Table 3). Slight
increases in allelic richness were observed for the SLW and SAW broodstocks.

Estimates of observed heterozygosities (Ho) were generally lower in juveniles
compared to adults. Only the SIW juvenile population exhibited a slight increase in
average heterozygosity. The GLW adults and juveniles showed the largest difference in
H0 estimates (8.3%). With the exception of SMD juveniles and SIW adults, sampled
populations showed heterozygote excess. Most adult and juvenile comparisons observed
decreased average expected (He) heterozygosities, except for the SLW and SIW adult and
juvenile comparisons. Estimates of inbreeding coefficients were higher in juveniles
compared to adults for four of the five broodstocks (Table 3), however, chi-square
analysis revealed that no estimates of inbreeding coefficients differed significantly from
zero (P<0.05).

Three hatchery strains, SMD, SAW and GLW exhibited a decrease in the average
coefficient of relatedness (er) from adult to juvenile populations (Figure 6). Only the
SLW and SMD strains showed an increase in average from adults to juveniles.

Significant differences (P<0.05) between the distributions of pair wise coefficients of

36

relatedness were observed in the SLW broodstock. Changes in the frequencies of
significant coefficients of relatedness at the full sibling level (er = 0.50) generally
decreased between adult and juvenile populations (Table 3); with the exception of the
SMD broodstock in which the frequency increased. The largest deviation in the
frequency of significant coefficients of relatedness at the full sibling level was observed
between the SAW broodstock and juveniles, 1.8%.

Qualitative inspection of distributions of r,,y values between the broodstock adults
and juveniles revealed little evidence of appreciable differences based on degrees of
kurtosis or skewness (Figure 6). We observed no consistent trend toward the increase in
levels of relatedness. Increases or decreases in rxy of juveniles over adults likely reflects
generational differences in total number of adults spawned or in the efficiency of the
spawning regime at the time broodstock adults and production juveniles were produced.

Effective Number of Breeders (Nb): Evaluation of hatchery records revealed that
the number of adults contributing offspring for some broodstocks was substantially lower
than the total number of adults actually spawned due to the elimination of entire lots of
fertilized gametes (Table 4). For example, entire lots of eggs from the GLW broodstock,
representing 236 adults, were discarded because eggs were of poor quality (eggs had not
fully matured and females spawned prematurely). Adults that created lots of juveniles
that were retained are considered “potential contributers” (Table 4). This analysis reveals
how the progression of gamete maturation of lake trout selected for spawning can
dramatically reduce the effective population size of hatchery broodstocks, independent of

issues related to spawning regimes.

37

Hatchery records indicated that in four of six broodstocks, some portion of the
total adult population spawned did not contribute to production (Table 4). For the SLW
and SMD broodstocks, eggs from all spawning lots (spawners) were used. The SAW and
SIW broodstocks experienced minor losses of egg lots, which occurred during the later
half of the spawning period. Ratios of total spawners to potential contributers for the
SAW and SIW broodstocks were 0.98 and 0.97 respectively. The ratios of total spawners
to potential contributers for the GLW and LLW broodstocks were substantially lower,
0.58 and 0.77 respectively. Elimination of eggs in the GLW broodstock occurred
episodically throughout the spawning period, while eggs were eliminated at the beginning
and end of the spawning period for the LLW broodstock

Not all breeding adults whose gametes were retained contributed to the
subsequent generation. The realized number of contributing adults depends on factors
that affect reproductive variance (e. g., sex ratios of spawned adults and pooling or
sequential mixing of gametes). Estimates of effective number of breeders (N) were
made using adult spawning lots. A spawning lot consisted of an identifiable group of
adults and associated juveniles. Estimates of N, were much smaller than the number
actually spawned (Table 4). The number of individuals in each hatchery broodstock
spawning lot ranged from 436-112. All broodstocks exhibited average effective breeder
numbers well below that of the number spawned (264-22). The ratio (Nb/N) of the
average effective breeder numbers to the total number of individuals in the evaluated
spawning lot (an estimate of spawning efficiency) were 0.61, 0.41, 0.28, 0.14, and 0.10
for the SAW, SMD, SIW, SLW, and GLW respectively. If the ratios estimated fi'om

spawning lots are applied to all potential breeding adults in Table 4, the total number of

38

realized contributers is 587, 35, 188, 130, and 107 for the SAW, GLW, SIW, SLW, and

SMD broodstocks respectively.

Stage 3: Distribution and Stocking of Broodstock Juveniles

Distribution of eggs from production facilities to rearing stations and
management projects for release are not representative of the entire number of potential
spawners (Table 5). For example, the SAW broodstock produced eggs that were
distributed to two hatchery facilities (Bayfield State Fish Hatchery, WI and Jordan River
NFH, MI), for a management program (fiy plant), and for retention at the Iron River
NFH. Of the 963 potential contributers (Table 5), Bayfield SFH received eggs produced
from 268 adults and Jordan River received eggs from 240 adults. These numbers
represent 28% and 25% percent of the potential contributers. Juveniles from this same
broodstock representing the contribution of only 96 adults (10%) were utilized for a fry
planting project. Juveniles representing approximately 436 adults (45% of the potential
contributers) were retained at Iron River NFH.

A similar situation exists for the SLW and SMD broodstocks at Pendill’s Creek
NFH. Of the 930 potential contributers for the SLW broodstock, Allegheny NFH, NY
received eggs from only 16 adults (2% of the potential spawners), Iron River NFH
received eggs from 426 adults (46%), and Jordan River NFH received eggs from 914
adults (98%). Eggs retained at Pendill’s Creek NFH represented 418 (47%) adults. For
the 260 potential contributers of the SMD broodstock, 148 (57%) adults contributed
juveniles to the Allegheny NFH, 78 (30%) contributed to juveniles sent to Iron River

NFH, and 112 (43%) contributed juveniles to Jordan River NFH. All juveniles for the

39

LLW strain were designated for management projects (astroturf). All eggs developed for
the GLW and SIW broodstocks were retained at the Iron River NFH.

Eggs distributed to facilities and projects also may only represent limited portions
of the spawning season. Hatchery adults are spawned at various times over a spawning
period that is typically 3-5 weeks long. Table 6 illustrates the number of adults spawned
per spawning period, the proportion of the total potential contributers that were spawned
on each given date, and the proportion of eggs contributed by adults spawned on each
date to the total egg numbers distributed to the various facilities and management
programs. For progeny to reflect the diversity in life history traits represented in the adult
broodstock (e. g., spawning time), eggs should be collected from throughout the spawning
period and in proportion to the number of adults spawned on each date. For the SLW
broodstock, Allegheny NFH received all eggs from the last spawning date which
realistically represented only 2% of the potential contributers. Eggs distributed to Jordan
River NFH and retained at Pendill’s Creek NFH were all collected on 10/14/98 and
10/15/1998, and represented 46% of the total potential contributers. Eggs sent to the
Jordan River NFH represented a more equitable cross section of the juveniles produced
across the spawning season. Similar results were observed for SMD and LLW
broodstocks. All eggs distributed to Iron River NFH for the SMD broodstock were from
the first spawning date and represented only 30% of the total potential contributers.
Eighty-two percent of the eggs distributed to Jordan River were collected from 78 adults
that represented 30% of the total potential spawning from one spawning date in the
middle of the spawning period. Eggs collected for Allegheny NFH were collected on two

dates in the middle of the spawning period and represented three separate spawning lots.

40

Two lots spawned on 10/15/99 and 10/20/99 represented 12% and 15% of the total
potential contributers, but contributed to 49% and 25% of the total eggs distributed to

Allegheny NFH in 1999.

DISCUSSION

Hatcheries have become a widely used tool for conservation and supplementation
of declining and endangered species (Anders 1998), including the mitigation of
decreasing fish populations, supplementation of populations, and restoration of extirpated
populations. Anders (1998) states that the role of hatcheries in conservation should be
“. . .to conserve wild fish populations along with their locally adapted gene pools and
characteristic phenotypes and behaviors.” This approach, referred to by Anders as
“conservation aquaculture”, should be employed as part of a more comprehensive
recovery program that includes habitat rehabilitation (Meffe 1995), and should not
emphasize quantity of fish produced over quality. However, selection forces that occur
within hatcheries and changes in genetic characteristics related to factors such as
inbreeding and genetic drift can lead to the loss of “locally adapted gene pools”.

Selection is a critical force to consider in the management of hatchery programs.
Campton (1995) defines artificial selection as selective forces that occur as a part of
normal operations in a hatchery system. Artificial selection includes advertent
(purposeful; Hynes et al. 1981) selection practices such as grading of fish, selecting for
larger fish, and selecting fish associated with a specific spawning time. Artificial
selection can lead to deviations in life history patterns such as alterations in spawning run

time and duration (F lagg et al. 1995). These selection pressures are often preventable

41

with minimal modification of hatchery routines. It is likely that hatcheries that employ
pervasive advertent selection pressures are not as concerned about preservation of genetic
diversity issues related to p0pulation genetics principles. For these activities, selection is
not Operating on the genotypes as we measured them in this study. Selection will operate
on individuals with reduced fitness as a result of inbreeding, however we do not measure
this. Other forms of selection in hatcheries are of importance as they relate to the degree
of domestication and abilities of offspring to survive in natural environments.
Domestication selection acts upon fish due simply to their existence in a hatchery
environment and is likely unpreventable (Campton 1995). This type of selection favors
adaptations conducive to grth and survival in a hatchery environment. Hatchery fish
tend to exhibit a greater propensity for capture, decreased tolerance for environmental
stochasticity, and decreased growth and stamina in natural environments (Hynes et al.
1981). Selection is more pronounced in populations retained for multiple generations in a
hatchery system. However, selection for individuals of higher fitness in hatchery

environments can occur rapidly (Crozier 1990; Doyle et al. 1995).

Role of Hatcheries in Lake Trout Restoration

Over thirty years of effort has failed to restore self-sustaining lake trout
populations to most of the upper Great Lakes (Selgeby 1995, Krueger et al. 1995). Lake
Superior is the only upper Great Lake that has shown significant recovery (Hansen et al.
1995). The existence of wild populations, lower fishing pressure, reduced pollution and
lower rates of sea lamprey induced mortality contributed to lake trout recovery (Hansen

et a1. 1995). Development of lake trout refirgia and effective sea lamprey control

42

measures have been instrumental in increasing natural reproduction in lakes Michigan
and Huron. However, levels of natural reproduction needed to establish self-sustaining
populations has not been realized (Eshenroder et a1 1995 and Holey et al. 1995).
Continued recruitment of sea lamprey from the St. Mary’s River (Eshenroder et al 1995;
Holey et al. 1995) and effects of early mortality syndrome (EMS) (F itzsirnons et al. 1999,
2001) identified in stocked lake trout, continue to compromise restoration efforts.

Due to the lack of natural recruitment in Lake Michigan and Lake Huron,
hatcheries will continue to play a critical role in the Great Lakes lake trout restoration
program. Until conditions are favorable for self-sustaining populations, hatcheries
provide a vital role as “gene banks” to preserve the genetic integrity of current and
historic lake trout populations. However, continued dependence on hatchery broodstock
production as the primary means of recruitment will require that ecological and genetic
diversity be maintained within lake trout hatchery broodstocks. In recognition of the
importance of preserving genetic diversity in hatchery broodstocks, a management plan
was developed for lake trout broodstocks in the federal hatchery system (Holey 1997).

The goal of this project was to evaluate how effective the lake trout hatchery
broodstock program has been in preserving genetic diversity and in producing “high
quality” offspring for release into waters of the upper Great Lakes. Data were collected
to genetically characterize broodstocks and juveniles to evaluate program practices at
each of three stages. Generally, given that lake trout have been in the hatchery system a
relatively short period of time, the genetic diversity of adult broodstocks and juveniles are
high and consistent with genetic characteristics of source populations. Estimates of

observed heterozygosity, allelic richness, inbreeding coefficients, and coefficient of

43

relatedness do not differ appreciably between lake trout broodstocks fiom their source
populations. However, significant differences in allele frequencies were observed
between wild source and broodstock samples for the SAW and LLW broodstocks.
Intergenerational comparisons of the SMD broodstock revealed few differences in
measures of genetic diversity (i.e., allele frequencies, observed heterozygosity, allelic
richness), but showed increased levels of inbreeding (F) and relatedness (rxy). Most
hatchery broodstocks and juveniles produced for stocking showed a slight decrease in
diversity (i.e., observed heterozygosity and allelic richness). The GLW broodstock
exhibited significant differences in allele frequencies across all loci. The SLW
broodstock exhibited a significant increase in the coefficient of relatedness rxy between
adults and juveniles. Most striking were estimates of effective numbers of breeders,
which were severely depressed for all hatchery broodstocks. Nonrandom egg collection
and distribution precipitated further declines in the number and diversity of adults

contributing to juveniles stocked.

Stage 1: Development of Broodstocks

Measures of genetic diversity for hatchery broodstocks should not appreciably
differ from their wild source populations. Differences between wild source lake trout
populations and lake trout hatchery broodstocks were seen for observed and expected
heterozygosties, allele fi'equencies, and allelic richness. These differences were,
however, not extreme. The largest difference between observed and expected
heterozygosities of wild source and hatchery broodstock populations was only 4%.

Larger changes in allele frequencies (>10%) were observed (e. g., in the LLW broodstock

44

and its source population) however most differences in allele frequencies were under 5%.
There were no appreciable differences between sources and broodstocks in allelic
richness values, inbreeding coefficients, and coefficients of relatedness. This evidence
suggests that genetic diversity of wild source populations have been largely maintained in
the hatchery broodstocks.

The existence of only minor changes in measures genetic diversity does not
eliminate the need for concern. Broodstocks are relatively recently developed (1987-
1994). Broodstocks will likely be in production for many years and will be used to
perpetuate other broodstock year classes. Even small changes (i.e., generational changes
in allele frequencies) can be exacerbated over several generations due to genetic drift
(Allendorf and Ryman 1987). Significant differences between allele frequencies were
found between LLW and SAW wild source and hatchery broodstock populations. The
most, at 3 loci, in the LLW broodstock. Differences in allele frequencies are likely the
result of sampling and spawning of relatively few individuals from the wild source
populations. It is likely that small numbers of adults sampled and unequal sex ratios used
during spawning contributed to low founding populations that were not entirely
representative of the source populations.

Hatchery programs are not always successful in capturing the genetic variability
of source populations (Dodson et a1. 1998). Sufficient sampling of source populations
can be difficult and commonly only small numbers of fish are sampled, often from a
disproportionately small period of the spawning session (Allendorf and Ryman 1987).
Often, the sampling is dictated by the availability of firnds, time, manpower (Kerby and

Harrell 1990; Yeager et al. 1990) and source population abundance (Brown et al. 2000).

45

For example, white bass (Morone saxatilis) culture in the southeastern United States is a
costly venture that limits the sampling effort employed, and results in only a few captured
fish. In conjunction, white bass females are highly fecund, and only a few females 5 5,
are required to meet production goals in most southeastern states (Kerby and Harrell
1990). Variance in reproductive success among female white bass has resulted in
disproportionate contributions of parents to progeny. In 1990, as few as six females were
responsible for producing 50% of the progeny in a South Carolina white bass hatchery
(Secor et al. 1992). Sampling of wild populations by hatcheries may also be performed at
certain times of the spawning season in order to maximize effort. Salmonid populations
are typically sampled at the peak of the spawning run when fish are most plentiful (Hynes
et al. 1981).

Even if a substantial and representative number of fish are sampled from a source
population, spawning methods can reduce the realized effective number of contributing
adults. Brown et al. (2000) investigated a propagation-assisted restoration program for
American shad (Alosa sapidissima) in the James River, Virginia. Numbers of shad in
the James River were insufficient to develop a broodstock, consequently requiring a
broodstock to be developed from the nearby Pamunkey River. Brown et al. (2000) found
that although the shad population in Pamunkey River appeared to be sampled adequately,
the number of contributing adults may have been severely diminished by the spawning
methodology. Greater than 1,400 American shad from the Pamunkey River were
collected over the spawning season. Fish were spawned via the pooling of female
gametes and sequential addition of the sperm from multiple males (female to male sex

ratio averaged 1.5:1). Using an experimental family, Brown et al. (2000) discovered that

46

reproductive variance associated with this spawning method could have reduced the
effective population size by 88%.

Mueller (1995), describes high variance in adult contributions for endangered
razorback suckers (Xyrauchen texanus) in an enhancement project in the southwest
United States. Male to female ratios of approximately 2:1 were employed following
evidence that females commonly spawn with two males in the wild. Genetic analysis of a
sample of the offspring, revealed that over half the juveniles in the sample were offspring
of a single female. In addition, these fish were collected fi'om the beginning of the
spawning season and were likely not genetically representative of the entire population.

The broodstock development stage (Stagel) is the most influential stage of
management of genetic diversity (Tave 1993). Genetic differences between the source
population and new broodstock realized at this stage will potentially be perpetuated or
exacerbated by genetic drift related to low Ne (Allendorf and Phelps 1980). Several
studies report observed losses in genetic diversity and or differences in allele frequencies
between source populations and hatchery broodstocks, and between hatchery year
classes. The studies have commonly invoked founder effects and genetic drift as reasons
for these differences (Cross and King 1983, Allendorf and Phelps 1980, review in Utter

1991).

Stage 2: Broodstock Maintenance and Production
Comparisons among the three SMD broodstocks revealed that genetic diversity
has been maintained during successive generations. This presumably, is a result of more

appropriate spawning techniques that have been employed for broodstock perpetuation

47

(1 :1 matings and using fewer eggs per mating), compared to practices employed for
broodstock production. The STW broodstock exhibited significant differences in allele
frequencies at one locus with the SMD broodstock. STW also exhibited an increase in
the inbreeding coefficient estimate. The wild population from which the STW
broodstock was derived may be the product of the stocking of j uveniles from previous
SMD broodstocks. If broodstock production involves the use of unequal sex ratios and
gamete pooling techniques, these conditions, and others that will be described below, will
likely lead to low effective population sizes and juvenile populations that are not
representative of the parent broodstock. This is an important consideration for the
restoration effort if juveniles to be stocked back into the wild do not reflect their parental
broodstock or ultimately the broodstock’s source population.

The inbreeding coefficient (F) of the STW broodstock is the highest estimate
recorded in this study. The estimate F for STW is slightly larger than that of SMD81, but
far exceeds the inbreeding coefficient for SMD. One explanation for the larger estimate
in STW is that the juveniles that established and perpetuated the source population of the
STW broodstock were derived from small effective populations of highly related cohorts.
These high numbers of related individuals were more likely to interbreed, thus increasing
the inbreeding coefficient (F). An alternative explanation is that individuals that were
spawned to create the STW broodstock were sampled from multiple and genetically
differentiated populations. The pooling of progeny of spawning events in different
populations would create a deficiency in heterozygotes (Wahlund effect) resulting in a

higher estimate of the inbreeding coefficient.

48

Reductions in effective population size occurred at a multiple levels of the
broodstock maintenance and production stage for several broodstocks in this study.
Spawning records revealed that for some broodstocks, large numbers of adults were
removed from the pool of potential contributers due to the excision of entire lots of
juveniles. Juvenile lots were excised due to poor egg quality and were associated with
adults that were spawned at the beginning or the end of the spawning period, suggesting
that eggs were not ready (i.e., too green) or were no longer viable. In two cases, this
reduced the potential number of contributers by 41% for the GLW broodstock and 21%
for LLW broodstocks. This demonstrates that conditions unrelated to hatchery spawning
or sampling methods can also decrease the numbers of adults that may contribute to
subsequent generations.

Estimates of effective number of breeders revealed that several broodstocks show
a considerable decrease in the number of adults that contribute to subsequent generations.
The ratios of the number effective breeders to the total number spawned, or spawning
efficiency (Nb/N), ranged from 0.10 to 0.61. Low spawning efficiency is likely related to
particular spawning methodologies. SMD, SLW, and LLW broodstocks were spawned
by pooling gametes at a 5:5 male to female ratio and GLW, SIW, and SAW broodstocks
were spawned by pooling gametes at a 10:5 male to female ratio. For broodstocks at
Pendill’s Creek NFH, large disparities in the sex ratios require that multiple males be re-
used to spawn a larger number of females. Some males were reused up to three times in
the LLW broodstock. These methodologies likely increased the reproductive variance
between adults. High variances in reproductive success have previously been shown to

result from similar spawning techniques for salmonid species (Gharret and Shirley 1985;

49

Withler 1988). Genetic measures such as significant differences in allele frequencies in
the SLW broodstock and increased levels in the estimates of inbreeding are consistent
with these gamete pooling practices.

High reproductive variance leading to unequal contributions of adults to progeny
is the most common cause of low effective population size in fishes (Allendorf and
Ryman 1987; Busack and Currens 1995; Simon 1991; Hedgecock et al. 1992 and
citations therein). High variance in the contribution of males to subsequent generations
has been associated with spawning methods that involve the sequential addition or
pooling of male gametes to fertilize the eggs (Gharrett and Shirley 1985; Withler 1988).
Sequential and pooling methods are commonly employed for salmonid production and
are common in hatchery systems that possess limited numbers of males (e. g., Pacific
salmon hatcheries; Withler 1988). Kincaid (1995) surveyed 221 salmonid broodstocks in
the US. and found that 48% possessed unequal sex ratios and 34% possessed excess
numbers of females.

Lower effective population sizes are likely to increase the probability of the loss
of alleles and increases in variance in allele frequencies. The rate of change in allele
frequency and rate of loss of genetic variance is expected to be 1/2Ne per generation.
This translates into a larger rate of loss in allelic variation that could have otherwise been
avoided if more effective spawning measures had been employed. For example, applying
the efficiency ratio (Nb/N) to the total potential contributers for the GLW broodstock, the
effective number of breeders is 35 which translates into a 1.4% loss in variance per
generation. If the effective number of breeders equaled the potential contributers, loss of

allele variance would only be 0.1%. Lower effective numbers of breeders also increases

50

the chance that rare alleles (alleles with a frequency of s 1% in a population) will be lost
to genetic drift. Using the GLW broodstock as an example, the estimated Nb of 35 would
translate into a less than 55% probability of retaining a rare allele over one generation of
random spawning (review in Holey 1997 ). The chances of rare allele retention decreases
to less than 17% if the estimated N, for the GLW broodstock is applied over three
generations. However, if all potential spawners (GLW=346) contribute to the subsequent
generation, the probability of retention of rare alleles will increase to ~100% over three
generations (review in Holey 1997).

Differences in allele frequencies were observed between hatchery adults and
juveniles (Table 3). Significant differences in allele fiequencies were observed between
all comparisons except for the SMD broodstock. Allele fi'equencies differed significantly
at all loci for the GLW broodstock adults and juveniles, which is consistent with the
extremely low effective numbers of breeders.

No appreciable differences in genetic diversity measures were observed between
hatchery adults and juveniles. All juveniles, except the SIW juveniles, exhibited
decreased observed heterozygosities over that docmnented for adults. However, single
generational differences in measures of genetic diversity (numbers of alleles and
heterozygosity) were not large, and did not suggest that juveniles were no longer
representative of adult broodstocks. However, as stated before, differences are important
when compounded over multiple generations.

The SLW broodstock adults and juveniles differed significantly for the
distributions of coefficients of relatedness. This significance is likely related to the large

number of males reused (56%) during the spawning of the SLW broodstock. Males from

51

SMD (15%) and LLW (77%) broodstocks were also reused multiple times during a
spawning period. An increased level in the average coefficient of relatedness in progeny
suggest that on average, individuals within the juvenile population are more closely
related to each other than adults are related to other adults in the population. High
degrees of relatedness between individuals could lead to inbreeding if these juveniles
survive and reproduce themselves, which is the goal for lake trout restoration.

Estimates of effective population size are sufficiently high to prevent large
changes in genetic ‘diversity over a single generation, but do represent extremely low
spawning efficiencies that translates into inefficient use of these broodstocks and of
resources. A substantial amount of time, money, and manpower is required to develop,
maintain, and spawn individuals in these broodstocks. It is not in the best interests of the
lake trout hatchery program or the lake trout restoration program to expend considerable
effort and resources to maintain a large broodstock only to realize a 10% spawning
efficiency. Adoption of more effective spawning practices would be prudent. The use of
genetic markers would be an effective means of monitoring hatchery populations and
increasing their effectiveness and efficiency.

Broodstock management employed in Stage 2, during development of new
broodstock year classes and broodstock production, have also resulted in low effective
population sizes and genetic drift in other hatchery programs. Allendorf and Phelps
(1980) identified significant differences in allele frequencies related to genetic drift
between broodstock year classes in cutthroat trout. High reproductive variance and
gamete takes over a limited time period have been cited as causes for low effective

population sizes of broodstocks in a red sea brearn (Pagrus major) hatchery (Perez-

52

Enriquez et al. 1999). The Red Sea bream (P. major) hatchery program incorporates a
natural spawning system using 250 individuals at a sex ratio of 1:1. A limited gamete
take near the peak of egg production, in conjunction with high reproductive variance,
resulted in significant differences in allele frequencies between the broodstock and the
juvenile population and a low effective population size of 64. Similar patterns of genetic
differences that occur between source populations, broodstocks, and juvenile year classes
have been demonstrated for a number of other salmonid species (Allendorf and Ryman
1987)

Hatchery practices that increase adult reproductive variance by failing to spawn
all adults across the entire spawning period, discarding gametes, pooling gametes or
maintaining unequal sex ratios will lower effective population size and thereby increase
the likelihood of loss of genetic diversity, increase the likelihood of large generational

changes in gene frequency, and elevate levels of inbreeding over generations.

Stage 3: Distribution and Stocking

At Stage 3 of the broodstock program, great expenditures of time, human
resources and funding have been expended to develop and perpetuate broodstocks and
produce progeny to be stocked into US. waters of the upper Great Lakes. Measures
should be taken to ensure that eggs and/or juveniles are distributed to rearing stations and
stocking locations in a manner that accurately represents the entire broodstock from
which they were derived.

Analysis of hatchery distribution records revealed that effective population sizes

have been firrther compromised by disproportional distribution of gametes or progeny

53

(Table 5 and 6). Eggs produced at both Pendill’s Creek NFH and Iron River NFH are
distributed to several facilities and stocking programs throughout the Great Lakes.
However, these collections of eggs are frequently not representative of the spawned
population. For example, SAW broodstock eggs designated for a fry plant program were
derived fiom 96 adults spawned at the end of the spawning period that represented only
10% of the potential contributers. A majority of eggs (82%) collected from the SMD
broodstock were developed from adults spawned during the middle of the spawning
season from 78 adults which represented only 30% of the potential contributers.

If the spawning efficiency for each broodstock is applied to the numbers of adults
that were spawned on each of these dates, juveniles that were collected and distributed
from limited portions of the spawning season represent extremely low effective numbers
of breeders. For example, SLW eggs designated for distribution to the Allegheny NFH
were all collected from 16 individuals spawned at the end of the spawning season, which
represented only 2% of the potential contributers. The efficiency ratio for the SLW
broodstock (0.14) applied to these 16 individuals results in an effective number of
breeders of only 2 individuals! However, this is an extreme example and not all
distribution sites are this biased. The distribution of eggs from the SLW, SMD, and LLW
broodstocks to Jordan River, Allegheny, and a survival and imprinting enhancement
study (astroturf program), are examples of more representative collections of eggs. SLW
broodstock eggs distributed to Iron River NFH and retained at Pendill’s Creek NFH
likely represented most of the broodstock. Although these eggs were collected from a
single spawning date, the adults represented almost 50% of the total potential contributers

(426 individuals).

54

Stocking ”of juveniles or fertilized eggs derived from relatively few effective
breeding adults poses potential problems for lake trout restoration. Juveniles collected
from limited portions of the available spawners and from short intervals of the entire
spawning period will fail to represent the genetic and ecological diversity of their
progenitors.

Data presented in this study reveals low effective population sizes of broodstocks
maintained for upper Great Lakes lake trout restoration programs. Large numbers of
adults are maintained and spawned at the Pendill’s Creek and Iron River National Fish
Hatcheries (Table 4). However, the effective population numbers that contribute to
progeny stocked at specific locales are far fewer than the number initially spawned.
Managers need to re-evaluate the practices of excising entire lots of eggs due to “poor”
spawning quality and implement more efficient (e. g., 1:1 male to female) spawning
methods. The practice of nonrandom distribution of eggs should also be re-evaluated.

Results from Stage 2 revealed that overall levels of relatedness are not elevated in
juveniles over levels described for adults. However, if progeny from comparative few
adults are stocked, the average level of relatedness will be substantially higher than
expected. When these year classes are sexually mature, related individuals could spawn
resulting in elevated levels of inbreeding natural born progeny. This would not be

consistent with restoration of self-sustaining populations.
SUMMARY

Lake trout restoration efforts across the Great Lakes region rely heavily on

hatchery production. This is especially true for US. waters of Lake Michigan and Lake

55

Huron where wild lake trout populations have been extirpated. Hatchery production
goals (e.g., total juvenile production) should be balanced with goals of maximizing the
“quality” of progeny produced and stocked. Natural selection in the lake environment
will select for phenotypes/ genotypes of highest fitness. It would be prudent to stock
juveniles reflecting the greatest potential diversity available. Findings of this study offer
guidance to improve specific aspects of the lake trout broodstock program to produce
progeny that meet these goals (see Appendix II).

There appears to be no common pattern related to the overall genetic quality of
broodstocks over the entire federal lake trout hatchery program (Stage 1 through Stage 3).
At each stage we’ve shown examples of appreciable change in particular broodstocks at
some measures of genetic diversity, levels of inbreeding and coancestry, but not at others.
This reflects the fact that less than optimal practices are occurring at discrete points in the
broodstock program, but is not indicative of a general ineffectiveness of the program to
preserve genetic diversity over the entire program. Genetic diversity and allele
frequencies of lake trout of each generation reflect spawning events of a particular year.
Measures of genetic diversity reflect the processes by which matings were conducted.

Based on estimates of allelic diversity, allele frequencies, observed and expected
heterozygosity, inbreeding and relatedness, we found that lake trout broodstocks
examined in this study are largely representative of their source populations. Although
some differences were observed, we did not detect the extreme differences (i.e., in allele
frequencies and average heterozygosities) that would indicate that the genetic diversity of
the broodstocks fail to reflect those of their source populations. However, this does not

suggest that small differences do not justify some concern.

56

Allele frequencies will change and levels of inter-individual relatedness will
accrue over generations. Small genetic differences have the potential of being
exacerbated in closed p0pulations. Higher levels of relatedness among broodstock adults
in future year classes will result in elevated levels of inbreeding in production fish to be
stocked. This is especially likely in light of findings of low effective population sizes
(Table 4). Most broodstocks were established relatively recently. For example, the
oldest broodstock (SMD) (~ 7 generations) in production today are descendents of
broodstocks initially developed in 1949. It is conceivable that these broodstocks could be
perpetuated for several decades. Therefore, long-term planning is needed to preserve
genetic diversity over multiple generations.

The most significant findings of this study were of the low effective breeding
population size of production fish used for stocking. Although large numbers of adults
are maintained and spawned, a substantial number of adults do not effectively transmit
genes to the juveniles used for stocking. As an extreme example, we estimate that the
effective size of the SLW broodstock of 930 adults was reduced by 98.8% through
inefficient spawning methods and egg collection techniques to a realized effective
population size of 2 for a group of juveniles distributed to another hatchery. Hatcheries
have limited control over some activities. Reductions in effective population sizes due to
uncontrollable events such as the excising of egg lots due to poor quality or disease are
unavoidable. However, activities related to spawning methodologies, timing and
duration of egg takes, and egg collection and distribution should be managed to ensure

that juveniles stocked reflect the diversity present in broodstock adults.

57

Chapter 2
ASSESSING LEVELS AND PARTITIONING OF GENETIC DIV ERISTY IN WILD

AND HATCHERY LAKE TROUT POPULATIONS: RELEVANCE FOR LAKE
TROUT MANAGEMENT AND RESTORATION IN THE GREAT LAKES

Six hatchery broodstocks, the Seneca Lake (SLW), Lewis Lake (LLW),
Marquette (SMD), Green Lake (GLW), Apostle Island (SAW), and Isle Royale (SIW)
and wild lake trout populations of Lake Superior, represent the remaining stocks
available for restoration efforts in US. waters of the upper Great Lakes (Figure 1).
Broodstock selection was based on political considerations, traits of source populations,
source population availability, and a desire to maximize use of available genetic and
ecological diversity of lake trout populations still existing within the Great Lakes basin.
All broodstocks currently used for restoration efforts were developed from natural lean
lake trout populations. Lean lake trout were chosen for restoration efforts due to their
preference by sport and commercial fisherman (Krueger et al. 1983). Preference for lake
trout phenotypes of greatest recreational value curtailed development of broodstocks
from the firll complement of ecologically and phenotypically differentiated forms (e.g.,
siscowets and humpers) in the Great Lakes basin.

The SAW, SMD and SIW broodstocks were developed from Lake Superior
populations based on the availability of certain lake trout populations and a desire to
utilize native lake trout diversity (Lawrie and Rahrer 1973; Peck 1975; Lawrie 1978;
Swanson and Swedberg 1980, Krueger et al. 1983; G. Curtis, United States Geological
Survey, Great Lakes Science Center, unpublished data). The SMD broodstock was

opportunistically developed in 1948 from lake trout populations sampled near Marquette,

58

Michigan along the southern shore of Lake Superior. At the time the SMD broodstock
was being developed, southern Lake Superior lake trout populations had collapsed and
lake trout populations near Marquette were the only remaining populations available
(Lawrie and Rahrer 1973; Peck 1975; Lawrie 1978). The SMD lake trout broodstock is
the oldest of the hatchery broodstocks. The current broodstock year classes were
developed over fifty years from the original 1948 year class (Coberly and Horrall 1982;
Krueger et al. 1983; Kincaid et al. 1997).

The SAW and SIW broodstocks were derived in the middle 19908 from captive
populations collected from remnant wild lake trout populations from the Apostle Islands,
Wisconsin and Isle Royale, Michigan in Lake Superior, respectively. Similar to the SMD
broodstock, the SAW broodstock was developed opportunistically, but also with a desire
to utilize a native lake trout that had proven survivability during the collapse and
extirpation of other near-shore Lake Superior lake trout populations (Swanson and
Swedberg 1980; Krueger et al. 1983). Although lake trout populations inhabiting near-
shore waters around the Apostle Islands collapsed concomitant to other lake trout
populations of southern Lake Superior, vestigial lake trout remained and Apostle Islands
lake trout populations eventually rebounded (Swanson and Swedburg 1980; Schram et al.
1995). Lake trout populations around the Apostle Islands were also economically
feasible to sample given their proximity to ports and hatcheries. The SAW broodstock
was developed from reciprocal crosses between two captive year classes derived from
Apostle Island wild fish in 1985 and 1986 (D. Bast, USFWS, personal communication; S.

Schram, Wisconsin Department of Natural Resources, personal communication).

59

Populations sampled from locations around Siskiwit Bay, Isle Royale between
1981 and 1986 were the progenitors of the SIW broodstock (D. Bathel, MDNR, personal
communication). Multiple captive populations developed from five sampling years were
reciprocally crossed to produce the two year classes (1989 and 1993) evaluated in this
study (D. Bast, USFWS, personal communication). Isle Royale populations were
sampled to utilize native lake trout diversity in the lake trout restoration effort (Krueger et
al. 1983; Hansen et a1. 1995; G. Curtis, United States Geological Survey, Great Lakes
Science Center, unpublished data).

The GLW and LLW broodstocks represent the remaining vestiges of the genetic
diversity that existed in Lake Michigan. All wild lake trout populations in Lake
Michigan were extirpated. In an effort to develop broodstocks that reflected phenotypic
and behavioral characteristics of extirpated lake trout populations, feral lake trout from
Lewis Lake, WY and Green Lake, WI were sampled to develop the LLW and GLW
broodstocks (Coberly and Horell 1982; Krueger et al.1983; Visscher 1982; Kincaid
1993). The history of the GLW and LLW broodstocks can be traced back to egg
collections made in the late 1800’s from Lake Michigan (Coberly and Hall 1982; Krueger
et al. 1983; Visscher 1983; Kincaid et al. 1993). The GLW broodstock was developed
from adults collected fiom spawning populations in southern Lake Michigan. The
progenitors of the GLW broodstock were originally stocked in Green Lake, Wisconsin.
The initial GLW broodstock was developed in the late 1950’s fi'om Green Lake lake
trout. Juveniles from this broodstock were stocked into southern Lake Michigan during
the early and middle 1970’s. Due to hatchery logistical problems, the GLW broodstock

was retired in 1975 (Krueger et al. 1983). The GLW broodstock was subsequently

60

resurrected in the middle 19805 by sampling wild domestic and feral GLW lake trout in
southern Lake Michigan (Kincaid et al. 1993).

Progenitors of the LLW broodstock were developed from egg collections made in
1889 from the northern reaches of Lake Michigan that were subsequently stocked as fry
into Lewis Lake, Wyoming. Gamete collections of wild Lewis Lake lake trout in the
early 1980’s were used to develop the current LLW broodstock (V isscher 1982).

The SLW broodstock is the only broodstock in the upper Great Lakes derived
fiom a lake trout population outside the Great Lakes basin. The SLW broodstock was
developed from deep-water wild populations within Seneca Lake, New York (Krueger et
al. 1983; Eshenroder et al. 1995). Wild fish sampled and spawned in 1987 and 1992
from Seneca Lake were the basis for the current SLW broodstock (D. Blick, Allegheny
NFH, personal communication). The ability of Seneca Lake lake trout to avoid sea
lamprey predation during vulnerable (adult) life stages and the desire to utilize a deep
water variety of lake trout made lake trout of Seneca Lake an attractive choice for a
broodstock source population (Krueger et al. 1983).

Strategies for the stocking of lake trout juveniles derive from the six afore
mentioned broodstocks in the upper Great Lakes have emphasized stocking of fish from
multiple broodstocks and the need to correlate ecological and behavioral traits of
broodstocks to habitats of specific planting sites (Krueger et al. 1981; 1983; 1995).
Simultaneous use of multiple broodstocks for stocking at specific sites constitutes a
“lottery” method for selecting compatible broodstocks. The theory behind this stocking
strategy is to effectively offer the greatest diversity possible whereby those broodstocks

most suited to the habitat will prosper and reproduce (i.e., be selected for

61

and survive). Perpetuation of broodstocks that exhibit greater fitness (i.e., enhanced
survival and reproduction), will continue until broodstocks derived directly from self-
sustaining populations can be developed. Stocking of these broodstocks would be
continued until rehabilitation has been established. Due to the generation time of lake
trout (6-8 years) this method was deemed preferable to stocking one lake trout
broodstock at a time and assessing the success of each broodstock individually (Krueger
et al. 1981; 1983; 1995). Ecological and phenotypic characteristics of progenitor wild
stocks used to develop hatchery broodstocks have in part been used to design stocking
strategies. However, direct genetic characterization has yet to be employed and utilized
for stocking strategies in the upper Great Lakes.

In an effort to maximize the diversity of lake trout stocked into Lake Michigan,
numerous broodstocks have been selected for stocking (review in Holey 1995).
Broodstocks stocked into Lake Michigan included the GLW, LLW, Slvfl), SAW, and
SLW broodstocks. The Lake Michigan derived GLW and LLW broodstocks were an
intuitive selection for stocking in to Lake Michigan, whereas the SMD has historically
been the dominant broodstock stocked into Lake Michigan. Developed from ancestral
shallow water wild populations, in conjunction with its reputation for high rates of
survival, the SMD broodstock was an attractive option (Krueger et a1 1983). The SLW
broodstock provided a deep-water component to the cadre of broodstocks. In addition, in
an effort to develop a broodstock with both desirable hatchery and wild characteristics,
wild lake trout males from the Apostle Islands of Lake Superior were crossed with

females from the domesticated SMD broodstock (Krueger et al. 1983).

62

A similar strategy of multiple-broodstock stocking has been employed for Lake
Huron. The SLW, LLW, and SMD broodstocks are currently being experimentally
stocked onto specific sites of the Six Fathom Bank reef complex to evaluate relative
growth, survival, and reproduction (Eshenroder et al. 1995). Concurrently, all other lake
trout broodstocks are also being stocked elsewhere in the Lake Huron basin.

The marked lack of success in restoring viable and self-sustaining populations of
lake trout has elicited efforts to re-evaluate recovery programs and research needs.
Current stocking strategies have utilized indirect genetic considerations such as
broodstock source and environmental origin. Population genetics can provide inferences
into the extent of genetic differentiation between hatchery broodstocks and remnant wild
populations, predict the genetic implications for the simultaneous release of progeny from
multiple broodstocks, and allow development of protocols to promote preservation of the
genetic variation that remains in existing wild and domestic broodstocks. Restoration
efforts would be best based on biologically-sound criteria, founded on a greater
fundamental understanding of the relationship between genetic diversity of lake trout
broodstocks (both historical and contemporary) and extant remnant populations.

Krueger et al. (1989) employed protein-based molecular techniques to
characterize the genetic diversity of broodstocks stocked into Lake Ontario. Their work
has consequently led to the estimation of relative contribution of hatchery broodstocks to
naturally-produced progeny in Lake Ontario. Spatial diversity of wild lake trout
populations has been investigated in several other studies (Dehring et al. 1981; Ihssen et
al. 1988; Wilson and Hebert 1996). However, these studies were more descriptive in

nature and did not address genetic diversity of hatchery and wild populations together in

63

the context of management concerns. This project seeks to use molecular genetic
techniques to 1) develop background data pertaining to levels and partitioning of the
remaining genetic diversity of lake trout populations in the upper Great Lakes and 2)

discuss the application of this information to management and restoration.

METHODS

Sample Collection.- All hatchery broodstocks were sampled in the fall of 1998 by
hatchery personnel during routine spawning events. The Lewis Lake (LLW), Marquette
(SMD), and Seneca Lake (SLW) broodstocks were sampled from Pendill’s
Creek/Hiawatha National Forest Fish Hatchery in Michigan and the Isle Royale (SIW),
Apostle Island (SAW), and Green Lake (GLW) hatchery broodstocks were sampled from
the Iron River National Fish Hatchery in Wisconsin (Figure 1). Samples consisted of fin
clips (~1cm 2) removed from caudal fins and stored individually in high salt buffer (4M
urea, 0.2M NaCl, 0.1M Tris-HCL, 0.5% Sarcosine, 10mM EDTA). Two hundred adults
were sampled from the SLW, LLW, GLW, SIW, and SAW broodstocks, and 166 from
the SMD broodstock. Fin clips were stored at -20°C until analysis.

Liver tissue samples from the three wild lake trout morphotypes, lean, siscowet
and humper, were sampled from remnant wild populations at four locations across the
Lake Superior basin (Figure 1). Samples of lean, siscowet, and humper populations from
Isle Royale, M1 were taken in the summer and fall of 1995. Lean and siscowet lake trout
were also sampled in 1995 from the Apostle Islands, WI, and Stannard Rock, MI.
Apostle Island and Stannard Rock samples were supplemented with archival scale tissue

collected from the Wisconsin Department Natural Resources Bayfield field station.

Siscowet lake trout were sampled from Caribou Island, MI in 1995 and bumper lake trout
were sampled from Caribou Island in 1995 and 1998. Caribou Island 1998 bumper
samples consisted of fin tissue. Due to limited samples of siscowet lake trout directly
sampled from Caribou Island, siscowets from contiguous areas along the southeastern
shore of Lake Superior (between Grand Marias and Whitefish Point, MI) were included
with the Caribou Island samples. Lean lake trout from a remnant wild population in
Parry Sound of Georgian Bay, Lake Huron were sampled in the fall of 2000. Tissue
samples were preserved ethanol and scale tissue were preserved by dehydration and or
stored at -20°C

DNA Extraction.- DNA extraction of liver and fin tissue were performed using a
proteinase K digestion and a modified Puregene extraction protocol (Gentra, Inc.). DNA
was resuspended in 50 pl of TE buffer (10 mM Tris-HCL, pH 8.0, 1 mM EDTA).
Fluorometry was used to determine DNA concentrations. Prior to flourometry, RNAse (2
p1; 20mg/pl) was added to each sample. One hundred nanograms of DNA was used for
each PCR reaction.

A Chelex procedure was utilized for DNA extraction from scale samples. Scales
(3-5) were added to 250 pl of a 5% Chelex and 10 mM Tris-HCL (pH 7.5-8.0)
suspension. Scales were digested overnight with 3 ul of proteinase K. Proteinase K was
subsequently inhibited at 95 °C for 5 minutes and samples were centrifuged at 14000 rpm
for up to 10 minutes. The resulting supernatant was removed and 2.5 pl of supernatant
was used for each PCR reaction.

Microsatellite Screening- Nine polymorphic microsatellite markers were

screened including SfoI, Sf012 and SfoI8 (Angers et al. 1995), 0ne,u9, One/110 (Scribner

65

et al. 1996), 0gola and 0goIc (Olsen et al. 1998), Scoy19 (Taylor et al. 2001) and Ssa85
(O’Reilly et al. 1996). PCR reactions were performed in 25 pl volumes using
concentrations recommended by the respective authors. The PCR profile involved a
single 2 min denaturing step at 94°C, followed by 30 cycles of a 1 min denaturing step at
94°C, 1 min annealing step and 1min extension step at 72°C. PCR profiles for scale
extracted DNA required 35 cycles. Annealing temperatures were as follows: 60 °C for
Sfol and Sf012, 56 °C for Sf018 and Ssa85, 54 °C for 0nep9, 52 °C for 0g01a, 46 °C for

Onepl 0 and Scop19 and 48 °C for 0gol c. PCR products were screened using 6%
polyacrilamide vertical gels. Products were visualized by a Hitachi FMBIO 11 Multi-
View scanner and associated software. Microsatellite fragments were sized manually
using 20 bp internal lane standards. Several individuals of known genotype were also
used in each gel for standardization.

Statistical Analysis.— Allele frequencies were estimated for all populations using
BIOSYS I (Swofford and Selander 1981). Levels and partitioning of genetic diversity of
hatchery and wild populations was evaluated using the program CONTRIBUTE (Petit et
al. 1997). F statistics, used to measure the partitioning of allelic variance within and
among broodstocks and wild populations, were also calculated (Weir and Cockerham
1984)

The CONTRIBUTE program estimates the genetic diversity of a given population
based on its relative contribution to the overall genetic diversity of a cohort of
populations. For example, the program allows us to estimate the relative contribution of
each broodstock to total diversity when all broodstocks are evaluated simultaneously.

The contribution of a broodstock or wild population to the overall genetic diversity was

66

evaluated on two levels. First we evaluated the relative contribution each population to
total diversity based on population’s own diversity. Secondly we evaluated the relative
contribution of each population to total diversity based on a population’s divergence or
uniqueness from other populations. These distinctions are useful for the identification of
populations for conservation purposes. In the event that populations need to be
prioritized for conservation effort, populations can be evaluated not only on their
diversity, but also on their distinct genetic characteristics. The contribution of a
population, k, to the total diversity (diversity of all populations combined; Nei 1973) is
based on the population’s intrinsic diversity that is measured by estimates of the total
diversity (ht) without the kth population (n-l), the total mean diversity (h,) without the kth
population, and the mean genetic differentiation of the k population (G5,) from all other
populations. The relative contribution of a population, km, to overall diversity (Ct), is
decomposed into the contribution of the population based on its own diversity (Cs) and
the contribution to overall diversity based on an a population’s divergence or uniqueness
(Cd) from other populations.

Another means of estimating diversity is to measure allelic richness. Allelic
richness provides a measure of the number alleles per locus standardized for variable
sample sizes. The number of allelic states is dependent on sample size, and differences in
the average number of alleles observed between populations may be influenced by
differences in sample size among populations. Allelic richness is an important diversity
measure in that populations that do not retain high gene diversity (heterzygosity) may
however possess comparatively large numbers of alleles or unique alleles (Petit et al.

1996). Average allelic richness, r, was calculated for each population. Populations were

67

evaluated on their contributions to overall allelic richness (Cn), contribution to overall
allelic richness relative to intrinsic allelic richness (Cm), and contributions to overall
allelic richness relative to allelic divergence or uniqueness from other populations (Crd).
Program CONTRIBUTE was used to examine contributions to diversity using three
different population groupings. Analyses were performed separately for hatchery
broodstocks, wild lean populations, and all wild Lake Superior populations (all
morphotypes) in order to compare levels and partitioning of genetic diversity within and
between populations.

Composite estimates of F statistics that partitions variance in allele frequency into
components for all loci were derived for hatchery and wild populations using the program
FSTAT (vers 2.8; Goudet 1999). Allelic variance among individuals within populations
(f) and over all populations (F), and variance among populations (05,) was estimated. An
additional estimate of variance among morphotypes (0m) was calculated for wild
populations in Lake Superior. Hierarchical analysis of Lake Superior lake trout was
performed using the program GDA (Lewis and Zaykin 2001). Variances among wild
populations were calculated with and without the Parry Sound population in order to
elucidate genetic structure of lake trout populations within the Lake Superior basin (Table
8).

Genetic structure of hatchery broodstocks and wild populations was visualized
using Cavalli-Sforza and Edwards chord distance (1967) neighbor-joining trees. Cavalli-
Sforza and Edwards chord distances were estimated using BIOSYS I (Swofford and
Selander 1981). The Cavalli-Sforza and Edwards chord distance is a composite measure

of differences in allele frequency between each pair of populations, summed over all

68

alleles and all loci. Relationships among populations as summarized by genetic measures
were shown using neighbor-joining trees generated in the program MEGA (vers 2.1;
Kumar et al. 1993). The neighbor-joining method for generating a tree involves the use
of a phylogenetic algorithm that clusters populations based on the degree of genetic
divergence, while assuming that evolutionary change is not fixed among populations
(review in Avise 1994). This assumption allows branch length to vary, illustrating the

number of evolutionary changes (longer branches represent a greater number of changes).

RESULTS

Three hatchery broodstocks, LLW, SLW and SMD, exhibited the largest values
for genetic diversity and divergence in the CONTRIBUTE analysis (Table 7). The SLW
broodstock represented the largest values for estimated expected heterozygosity (hk=
0.449), relative differentiation or allelic variance of the k population (Gst= 0.110),
contribution to total diversity h, (Ct= 0.051), contribution to total diversity h, due to
intrinsic diversity (Cs= 0.014), contribution to total diversity based on genetic divergence
(Cd= 0.083), and the contribution to total allelic richness based on intrinsic allelic
richness (Cr,= 0.190). The SMD broodstock exhibited the lowest or moderate values for
most measures of diversity (i.e., lowest hk). However, the SMD broodstock was found to
exhibit the highest allelic diversity. Allelic richness for SMD (3.6) represented the
greatest contribution to total allelic richness based on Cm (0.107), and the greatest
contribution to total allelic richness due to the divergence or uniqueness of alleles. The
LLW broodstock was secondary to the SLW broodstock for diversity measures, but

exhibited the largest value for C, (0.059), relative contribution to total allelic richness.

69

The SLW and LLW broodstocks contributed the most to total broodstock diversity due to
their intrinsic variability (heterozygosity), while SMD contributed considerably to
broodstock total diversity due to its divergence or uniqueness (possession of discrete
alleles) in allelic richness.

Evaluation of wild lean populations revealed that certain populations contributed
disproportionately to allelic diversity and overall genetic variance. Lean lake trout from
Stannard Rock leans contributed the most to total diversity by exhibiting the largest
estimates for G5, (0.024), C, (0.037), Cs (0.023), and Crs (0.076), the greatest contribution
related to the intrinsic diversity of Stannard Rock leans. In contrast, lean lake trout from
Isle Royale contributed most to total diversity and allelic richness (r = 3.122) due to
genetic divergence and allelic uniqueness, [Cd (0.151) and Cd (-0.020) respectively].
Lean lake trout from Stannard Rock continue to contributed the most to total diversity
relative to their own diversity, [Ct (0.007), and Cs (0005)], and were the most divergent
relative to other wild populations (Gs, =0.031) when all wild populations are considered
together. Siscowet lake trout from Stannard Rock also represent high levels of diversity,
hk (0.437) and Cs (0.005), and high contribution to total allelic richness based on this
population’s own allelic richness, Crs (0.062). Isle Royal leans represent equal genetic
divergence to the Stannard Rock leans (0.031) and the greatest contribution to total
diversity due to genetic divergence (0.003). Caribou Island humpers were the most
diverse with respect to allelic richness (r = 3.3), Cu (0.016), and divergence Crd (0.045).

Analysis of genetic variance for hatchery and wild populations revealed that
variation is partitioned differently in each of the two groups (Table 8). Genetic variation

of hatchery populations is partitioned among hatchery broodstocks (mean 0,, = 0.058,

70

P<0.01), while for the wild populations, mean 0,, = 0.024 (P<0.01). This is not
surprising given that the wild populations are entirely from the Lake Superior Basin
while hatchery broodstocks include populations derived from lake basins other than Lake
Superior (e.g., the SLW broodstock from Seneca Lake, NY). Further, hierarchical
analyses for wild Lake Superior populations revealed that a significant portion of total
variation in allele frequency was partitioned among morphotypes (0m, = 0.029; P<0.01).
Some basin-dependent variation was evident given the increased variance observed
among the total wild populations when the Lake Huron Parry Sound population was
added to the analysis (0,, = 0.033; P<0.01).

Genetic structuring that corresponds to basin of origin is firrther revealed by
examination of genetic distances between both hatchery and wild populations (Figure 7).
Broodstocks were observed to cluster together based on the their lake origin (Figure 7a).
The Seneca Lake broodstock, the only broodstock developed from sources outside the
Great Lakes basin, was highly divergent from all other broodstocks. Structuring across
wild populations was most notable among populations from different basins, as the Parry
Sound population from Lake Huron was most genetically diverged from Lake Superior
populations. Within Lake Superior, among population genetic affinities were most
notably based on morphotypes irrespective of location of origin (Figure 7b); a result
consistent with greater overall variance attributed to differences among morphotypes
(humper, leans, siscowets; 0",, = 0.029; P<0.01) relative to variance among geographic

sampling locales for each morphotype.

71

DISCUSSION

Development of a fundamental understanding of the levels and partitioning of
genetic diversity of remaining wild and hatchery lake trout populations (Bumham-Curtis
1995; Hynes et al. 1981; Meffe 1995) should be a prerequisite for establishment of
restoration strategies. Although studies have characterized genetic diversity in wild
populations of the upper Great Lakes and broodstocks developed from them, there has
been a lack of integration of these two components. Herein we provide information on
levels of genetic diversity of wild and hatchery populations and how this diversity is
partitioned. These data have direct relevance for development of management strategies
primarily for lake trout, but applicable to other species for which supplementation is an

integral component of species recovery.

Levels and Partitioning of Genetic Diversity

The Seneca Lake (SLW) and Lewis Lake (LLW) broodstocks exhibited the
highest levels of diversity among lake trout hatchery broodstocks. Nearly all measures of
genetic diversity and divergence were highest in the Seneca Lake broodstock, which is
likely related to its location of origin. The Seneca Lake broodstock is the only
broodstock developed from fish that were completely segregated from upper Great Lakes
lake trout populations. Phylogeographic evidence suggests that lake trout populations of
the Great Lakes and eastern Canada originated from three separate Pleistocene glacial
refugia, (Berigian, Mississipian, and Atlantic; Wilson and Hebert 1996). As lake trout

reinvaded the Great Lakes region, the relative contributions of lake trout from these

72

various refugia could have contributed to the differences observed between Seneca Lake
lake trout (predominately originating from Atlantic refugia) and those developed from
within the upper Great Lakes basin (likely a mixture of all three refugia).

The Lewis Lake broodstock was developed from lake trout stocked into Lewis
Lake, Wyoming, originally from various locales in the northern Lake Michigan basin
plus some additional lake trout from an unknown origin (review in Grewe and Hebert
1988, and Visscher 1983). This may account for LLW’s high diversity and high
contribution to total allelic richness and allelic divergence or uniqueness (Cr, = 0.058, Cm
= 0.051; Table 7).

The Marquette broodstock represents the lowest genetic diversity of all
broodstocks, but exhibits the highest values of allelic richness and contribution to total
allelic richness based on its intrinsic allelic diversity. Low genetic diversity associated
with this broodstock is expected given the long history of domestication (Krueger
et al. 1983). The SMD broodstock was originally developed in 1949, prior to the collapse
of the fishery in that region of the lake. Subsequent broodstocks were perpetuated from
descendents of this original broodstock.

Allelic richness for the Marquette broodstock was unexpectedly high given this
broodstock’s degree of domestication (r = 3.6, C,, = 0.024; Table 7). Previous studies
have found comparatively high numbers of alleles per locus (Grewe and Hebert 1988 and
Ihssen et al. 1988) and mitochondrial DNA haplotypes for the Marquette broodstock and
wild lake trout populations sampled near the Marquette broodstock source population
(Wilson and Hebert 1996). Our findings corroborate those of earlier studies, suggesting

that the Marquette broodstock possesses a higher complement of unique alleles than other

73

broodstocks. In addition, lake trout from the Green Lake and Apostle Islands
broodstocks were added to the Marquette broodstocks in the middle and late 1960S
(Krueger et al. 1983), and potentially contributed to allelic diversity expressed in the
current Marquette broodstock.

Contribution to total diversity and allelic richness, in summary, can occur on two
levels: 1) contribution due to a population’s intrinsic diversity and allelic richness, and 2)
contribution due to a population’s genetic divergence or allelic divergence. Populations
therefore may contribute differently to overall diversity; both of these levels of diversity
should be considered in the development of management strategies (e.g., broodstock
development and stocking programs). The Seneca Lake (SLW) and Lewis Lake (LLW)
broodstocks contribute the most to total broodstock diversity based on their intrinsic
genetic diversity and divergence. The LLW broodstock contributes most to total allelic
richness, but the Marquette (SMD) broodstock represents the largest allelic richness value
due to its high Cm value (uniqueness).

Analyses of wild populations revealed that lean lake trout from Stannard Rock
contributed the most to total diversity based on that population’s intrinsic diversity for
comparisons between wild lean populations and all wild populations, while lean lake
trout from Isle Royale were the most divergent population of the lean populations.
Stannard Rock siscowets also exhibited high levels of diversity for comparisons among
all wild populations. Lean lake trout from Isle Royale and humper lake trout from
Caribou Island were most divergent of all wild populations.

Evidence suggests that hatchery broodstocks differ significantly in allele

frequency (mean 0,, = 0.058; Table 8) and that this diversity is related to lake basin of

74

origin (Figure 7a). The neighbor-joining tree derived with Cavalli-Sforza and Edwards
chord distances shows that hatchery broodstocks most similar to other broodstocks were
developed from within a given lake basin. As would be expected based on its origin, the
Seneca Lake broodstock was dissimilar to all other broodstocks. Present population
relationships based on basin of origin are similar to genetic relationships of historical
populations in the upper Great Lakes (Guinand et al. unpublished data). Consequently,
although these broodstocks have been exposed to hatchery and management
perturbations, they still retain a basin-dependent genetic signature.

Wild Lake Superior lake trout are less differentiated than the hatchery
broodstocks. Although wild populations differ significantly in allele frequency
(0,,=0.024; Table 8), more variation can be attributed to differences among wild lake
trout morphotypes (0m,=0.029; Table 8). This is supported by the Cavalli-Sforza and
Edwards chord distance neighbor-joining tree that reveals that wild Lake Superior lake
trout morphotypes are more similar within morphotype across locales throughout the
basin than to other morphotypes within the same location (Figure 7; b). Conversely,
Dehring et a1. (1981), in an allozyme analysis of lake trout morphotypes from similar
locations as this study, found that different morphotypes within a given location were
more genetically similar than similar morphotypes from other locales, suggesting a recent
divergence in lake trout. Our molecular microsatellite data suggests that morphotypes
had genetically diverged prior to recolonization of the upper Great Lakes or a higher rate
of historical gene flow among populations within morphotypes than between
morphotypes. Lake trout isolated into multiple refugia during the Pleistocene glaciation

would have developed ecological and behavioral adaptations to their resident habitats.

75

During recolonization, lake trout may have segregated based on habitats that paralleled
those in the glacial refugia.

In light of the significant evidence of genetic structuring among lake trout
morphotypes, management designed to restore wild lake trout would profit from
consideration of the natural fish community structure. Management policy should
recognize lake trout morphotypes as “distinct” units, similar to the level of Management
Units (MU; Moritz 1994a, b) utilized for a number of imperiled salmonid species (i.e.,
Pacific sahnonids; Waples 1991). Management units are homogeneous populations
where gene flow is restricted to allow for significant allele frequency divergence between
populations. MUs should be managed to consider these differences.

We present evidence of allelic divergence among lake trout morphotypes and
inferentially of restricted gene flow. Genetic variance of wild Lake Superior lake trout
was significantly partitioned between morphotypes (Table 8). Genetic structuring
(Figure 7b) of wild Lake Superior lake trout populations reveals that genetic affinities are
based on morphotype irrespective of location of origin (Figure 7b). For example,
Caribou Island siscowets are more similar to Isle Royale siscowets than they are to
Stannard Rock leans and humpers. This suggests that gene flow is restricted among
leans, siscowets, and humpers at a given location, most likely a result of spatial and
temporal segregation during spawning. As a result of this segregation, each morphotype
constitutes a significant independent portion of the overall genetic diversity of wild Lake
Superior lake trout. Genetic data suggests that lean, siscowet and humper lake trout
should not be managed inclusively as “lake trout,” but should be recognized as “distinct”

and managed under a fish community context.

76

In addition to genetic structure based on morphotype, wild populations are also
geographically structured by lake basin. Addition of the Parry Sound wild population of
Lake Huron to the Lake Superior populations increases the overall variance among wild
populations (0,,=0.033; Table 8) because it is genetically dissimilar from all Lake
Superior populations (Figure 7b). This complements results obtained from the hatchery
broodstock data analysis in which we found significant genetic structuring related to lake

basin origin (Figure 7a).

Management Considerations

Current stocking strategies within the upper Great Lakes have focused on a
“lottery” system that involves the stocking from multiple broodstocks (Krueger et al.
1981; Krueger et al. 1995; Holey et al. 1995; Eshenroder et al. 1995), relying on the lake
environment to naturally-select for adapted genotypes. However, the stocking of
individuals from multiple broodstocks at a single locale may result in a homogenization
of genetic diversity if progeny from stocked individuals of different broodstocks
interbreed. Evidence of homogenization of hatchery sahnon stocks of the Pacific
Northwest has resulted in the loss of the genetic distinctness between these stocks
(Waples et al. 1990; Waples 1991). Although the homogenization of these sahnon stocks
was related to egg and juvenile transfers and the development of hatchery stocks from
multiple lineages (Waples et al. 1990), it is not unlikely that a similar situation could
occur from the stocking of multiple lake trout broodstocks at the same locations.
Currently, the Seneca Lake broodstock, Lewis Lake broodstock, and the Marquette

broodstock are all simultaneously stocked in the same location in Lake Huron. These

77

broodstocks are likely the three most genetically divergent broodstocks. It would be
prudent to avoid stocking juveniles from such divergent broodstocks in identical locales
to reduce the potential for interbreeding and homogenization of lake trout stocked.

Wild populations enhanced with broodstocks developed from different lake basins
may lose their genetic distinctness. The detrimental effects (i.e., outbreeding depression;
Lynch 1997) of stocking hatchery fish within wild populations have been well
documented (Waples et a1. 1991; Waples 1990). As evidenced by this study, wild
populations and hatchery broodstocks are genetically structured by basin. Genetic
differentiation between the wild and hatchery populations may be exacerbated if the
hatchery lake trout are derived from a small effective population size. Ryman and Laikre
(1996) suggest that wild populations enhanced by large numbers of hatchery fish
developed from a low effective population size will decrease the effective population size
of the overall wild and hatchery populations. Small effective population sizes have been
associated with the contemporary production of lake trout juveniles (this study).
Fundamental knowledge of the levels and partitioning of genetic diversity of lake trout
populations available for restoration of the upper Great Lakes is crucial for effective
development of future management strategies.

Broodstock Retirement.- The number of broodstocks that can be maintained in a
hatchery system is dependent upon space, time, and resources. In light of the data
presented in this study, broodstocks that would be less desirable candidates for retirement
would be the Seneca Lake, Lewis Lake, and Marquette broodstocks. These three
broodstocks represent disproportionately high levels of genetic diversity and uniqueness

present across the six hatchery broodstocks. Additionally, given that there is evidence of

78

genetic relationships, based on phylogenetic trees (Figure 7a) consistent with the lake
basins from which the broodstocks were developed, it would be prudent to maintain
representatives of each basin among the broodstock choices. Phylogenetic trees have
been widely proposed as a measure for assessing distinctness among populations and
identifying populations as candidates for conservation (V an-Wright et al. 1991; Crozier
1992; Faith 1992).

It is not unlikely that a broodstock would have to be retired due to logistical
constraints such as money, hatchery space, or resources. Similar situations have occurred
historically in the lake trout hatchery system. In the mid 1960s, the Green Lake
broodstock was housed at the Marquette Hatchery in Michigan, in addition to
broodstocks of Lake Superior origin. The Green Lake broodstock was found to spawn
later than the Lake Superior broodstocks, causing difficulties with spawning and stocking
schedules. As a result, the Green Lake broodstock was discontinued in 1975 (Krueger et
a1. 1983), but later resurrected.

The Marquette broodstock is currently being phased out in favor of the newly
developed Traverse Island broodstock (STW). The Traverse Island broodstock was
developed from wild populations of lake trout sampled from locales that were putatively
used to develop the SMD broodstock. The Traverse Island broodstock exhibits lower
observed heterozygosity than the Marquette broodstock and high inbreeding coefficient,
suggesting an increase in homozygosity in this broodstock (this study).

Broodstock Management Strategies.- Historically, broodstocks from numerous
locales representing a variety of ecological and behavioral traits have been selected to be

stocked in the upper Great Lakes. By maximizing the variability of lake trout stocked, it

79

is hypothesized that lake trout most suited to current lake conditions will be selected for.
No direct genetic measure has been employed to match lake trout broodstocks to
complementary historical populations. Recent genetic characterization of historical
populations (Guinand et al. unpublished data) in conjunction with the information
presented in this project can be utilized to match broodstocks with proposed release sites,
providing an alternative proactive stocking strategy.

Interest has grown in the development of new broodstocks from other lake trout
morphotypes in addition to the lean morphotype that has been exclusively utilized
(Krueger et a1 1995). This project provides direct measure of genetic diversity of wild
populations that may be used as a measure for selecting wild populations as potential
candidates for new broodstock sources. Lean populations within Lake Superior represent
a small proportion of the overall available diversity (Figure 8). If it is desirable to
maximize diversity as a restoration goal, then the largest component of variation in Lake
Superior is partitioned among the lake trout morphotypes (Figure 7b). Plants of these
fish into comparable habitats in other Great Lakes would be desirable. Among the
siscowet and humper populations, the Stannard Rock siscowet population exhibits the
highest diversity (h, = 0.437, G,, = 0.024, C,= 0.005, C, = 0.005; Table 7). Isle Royale
and Caribou Island humper populations exhibit the highest measures in divergence of
allelic richness (Cm = 0.006; Table 7) and the Caribou Island humpers exhibit the greatest
allelic richness (r = 3.280, C,, = 0.016, Crs = 0.010; Table 7). Selection of these wild
populations as sources for firture siscowet and bumper broodstocks would offer the

greatest diversity.

80

Recent investigation of the lake trout hatchery broodstock program has revealed
that methodologies related to the spawning of adults and egg collection and distribution
result in extreme reductions in broodstock effective population size (this study).
Hundreds of lake trout adults are spawned for each lake trout broodstock during the
spawning period, but subsequent juveniles in many cases only represent a small fraction
of the number adults spawned (in an extreme case as low as 2%). Spawning practices
utilizing the pooling of gametes and the reuse of multiple males was cited as a major
contributor to low effective population size. Lake trout broodstocks are spawned
utilizing a 5:5 or 10:5 male to female ratio and broodstock sex ratios are unequal,
requiring the reuse of large fractions of males (as high as 77%) to spawn a greater
number of females (this study).

Given the small effective population sizes realized during broodstock production,
maintaining larger numbers of adults for each broodstock may be one means to increase
effective population sizes. However, increasing the number of adults in the hatchery
system will likely constitute the reduction in the number of broodstocks that can be
maintained due to limited space. We present evidence that suggests most differentiation
among lake trout broodstocks exist between broodstocks derived from different lake
basins (Lake Michigan, Lake Superior, and Seneca Lake) (Figure 7; a). Thus, if space
allocation is a limiting factor it may be justifiable to combine broodstocks developed
from the same basin to create two composite broodstocks, a Lake Michigan (GLW and
LLW), and Lake Superior (SAW, SMD, and SIW) broodstock. Development of a Lake
Michigan and a Lake Superior broodstock, while continuing to maintain the Seneca Lake

(SLW) broodstock, will allow the hatchery system to maintain a greater number of adults

81

for each broodstock while continuing to preserve most of the lake trout broodstock
genetic diversity (Figure 9).

Conservation.- Populations that represent the greatest diversity and possess
unique genetic characteristics should be given priority status for purposes of
conservation. The Isle Royale lean, Isle Royale bumper, Stannard Rock lean, Stannard
Rock siscowet, and Caribou Island bumper p0pulations would be the preferred
populations to capture a large portion of the diversity represented in Lake Superior. The
Parry Sound population should also be given priority status given its degree of genetic
differentiation from Lake Superior populations (Figure 7b). As reviewed by Petit et al.
(1997) prioritization should be made using both allelic richness and heterozygotic
diversity measures, along with relative contributions to total allelic richness and diversity
of a population based on intrinsic diversity and divergence from other populations. As
was observed in this study, populations can contribute little to overall diversity based on
heterozygosity, but may contribute significantly based on allelic richness (e. g., Marquette
broodstock). These populations are of importance because they possess a higher
diversity of alleles despite lower heterozyogitic diversity. In addition, populations may
possess low diversity overall, but still contribute substantially to total diversity based on
genetic divergence. The Isle Royale bumper population, for example, exhibited high
divergent values for allelic richness (Cm = 0.006; Table 7) despite a low allelic richness
(r = 2.9; Table 7) and low contributions to total allelic richness (C, = 0.000; Table 7).
Petit et al. (1997) describes populations with these characteristics as likely isolated
populations of limited gene flow and subjected to significant genetic drift, resulting in

low allelic diversity but high divergence. This data is consistent with mark-recapture

82

data of bumper lake trout (review in Dehring et al. 1981), corroborating evidence of
restricted range (low migration) of this lake trout morphotype (Bumham-Curtis and
Bronte 1996). Thus, bumper lake trout exhibit a high affinity for associated reefs,

resulting in limited gene flow between populations.

SUMMARY

Fundamental understanding of levels and partitioning of genetic diversity of lake
trout populations in the upper Great Lakes can help form the basis for restoration
management. It is prudent to gain a basic understanding of the genetic structure of what
existed previously and what is currently available if restoration goals to reconstitute
diversity to historical locales. This study revealed that genetic diversity of hatchery and
remnant wild populations is structured on multiple levels on the basis of morphotype and
among geographically structured populations. Management efforts should recognize this
diversity. Genetic data can be utilized to evaluate current management considerations,

and as a means of inference and evaluation of future management endeavors.

83

APPENDIX I

GUIDELINES EXPRESSED IN THE LITERATURE REGARDING HATCHERY
PRACTICES RELATED TO THE THREE STAGES OF HATCHERY BROODSTOCK
MANAGEMENT

Preservation of genetic diversity throughout all stages of a hatchery program is
important. Adaptability and the rate of evolutionary response are directly dependent on
population genetic diversity (Meffe 1995). Figure 2 presents critical issues for
preservation of genetic diversity during the three stages of broodstock development and
maintenance. Practices commonly employed in hatchery systems during all stages of
broodstock development, perpetuation and production are often contrary to what is
recommended for preserving genetic diversity. Some issues (i.e., population sampling)
are dependent on available resources (e.g., time, money, manpower), and accessibility to
wild sources. However, logistical considerations involving gamete takes (i.e., broodstock
sex ratios, spawning ratios, fertilization methods), and rearing and stocking methods can
be tailored to achieve management goals of maximizing retention of levels of genetic
diversity in the face of logistical constraints (Allendorf and Waples 1996). Below are
guidelines expressed in the literature regarding hatchery practices related to the three
stages of lake trout hatchery broodstock management investigated in this study.

Stage I .' Sampling of source populations for the development of broodstocks must
be designed to capture the genetic diversity characteristic of the source population. A
wild population should be sampled across the entire spawning period and include
sampling across the entire population (i.e., include subpopulations) (Hynes et al. 1981;
Krueger et al. 1981). Spawning methods that decrease variance in male and female
reproductive contributions will also serve to reduce relatedness among progeny and
increased effective population sizes are preferred (Gharret and Shirley 1985; Simon
1991; Kincaid 1995). An optimal spawning methodology would involve matings at a

male to female ratio of 1:1, and avoidance of pooling and sequential fertilization of
gametes (Gharrett and Shirley 1985; Withler 1988 Simon et a1 1991).

84

Stage 2: The creation of new broodstock year classes and production of progeny
for release should utilize the largest number of breeding adults possible. We recognize
that the numbers of individuals that contribute to subsequent broodstock year classes and
juveniles to be stocked can be compromised by a number of factors (Figure 2). It is
generally accepted that a broodstock of several hundred individuals is sufficient to
prevent significant deterioration of genetic diversity (Allendorf and Ryman 1987; review
in Kincaid 1995). Spawning methods equivalent to those outlined in Stage 1 should be
employed. Equal numbers of male and females for each broodstock should be
maintained and individuals should be spawned across the entire spawning period (Simon
1991)

Stage 3: Strategies employed to collect and disperse fertilized gametes and
stocking strategies should emphasize the preservation of genetic diversity (Figure 2)
(Kapuscinski and Jacobson 1987). Juveniles produced from each paired mating should
be combined immediately after fertilization and fertilized gametes should be dispersed
randomly to rearing and release locations. Consideration should be made with regard to
existing wild populations in the targeted stocking area as well as to the number of
hatchery strains being concurrently stocked (Waples 1991). Mixing of hatchery juveniles
with wild populations or mixing from multiple broodstocks may adversely effect the
population genetic constituency of wild populations (Ryman and Laikre 1991; Waples
1991; Tringali and Bert 1998).

85

APPENDIX II

OVERALL RESULTS AND RECOMMENDATIONS FOR THE LAKE TROUT
HATCHERY PROGRAM

Stage 1 -Overall, genetic characteristics of broodstocks reflect their wild source
populations. Few appreciable differences in measures of genetic diversity or allele
frequencies were observed. Broodstocks should be developed from multiple year classes
or lines using large numbers of adults, or from individuals derived from multiple source
populations of large N. These populations are less likely to be bottlenecked and exhibit
reduced variability and increased relatedness. If a broodstock is to be developed from
multiple year classes or lines, the size of each founding population should be large
enough to reflect the goal of the broodstock management plan. Spawn all fish using a 1:1
male to female ratio as recommended by the broodstock management plan. The greatest
source of low effective population size is high reproductive variance, typically due to
differentiated sperm penetrance and competition when male gametes are pooled.

Stage 2-Intergenerational comparisons revealed few differences in measures
genetic diversity. Evaluation of broodstock production revealed low estimates of the
effective number of breeders (Nb) related to elimination of egg lots and inefficient
spawning methods. Multiple year classes should be maintained and spawning should be
canied out by mating individuals between year classes. Spawning methods should
involve spawning all fish using a 1:1 male to female ratio as recommended by the
broodstock management plan. Adult sex ratios should be equalized to avoid reusing
males for spawning. This will serve to minimize the degree of relatedness among
progeny. Spawning numbers that include entire lots of spawners whose progeny were
culled or include reused spawners, are not representative of the true potential effective
population size and may not meet the requirements of the broodstock management plan.
If desired, practices that may improve the synchronicity of egg maturation include,
grading females based on ripeness to coordinate the spawning of females possessing the
same levels of egg maturation, and spawning over a larger spawning period but
increasing the number of spawning events to spawn more females at peak ripeness.

Stage3-Egg take and distribution methods frequently do not attempt to maximize
the total numbers of adults spawned and/or represent the entire spawning period.
Gametes from larger proportions of fish from throughout the spawning cycle should be
distributed and stocked when possible.

In general: The broodstock management plan (Holey 1997) provides a strong
foundation for the development and management of hatchery broodstocks. However, the
over reliance on genetic theory (i.e., 1/2Nc), assuming Ne relates simply to the total
numbers of adults spawned as an absolute measure of potential loss of genetic diversity
(i.e., heterozygosity), is not adequate. The composition of the entire broodstock of N

86

individuals (i.e., female to male ratios and year classes), the history of the individuals in
N and how gametes of individuals in N are combined can all effect genetic diversity.

87

APPENDIX III

TABLES AND FIGURES

88

Table 1. Comparisons of allele frequencies and measures of genetic diversities for three
wild populations and respective hatchery strains.

 

 

Population
Lewis Isle Apostle
Locus allele Lake LLW Royale SIW Island SAW

 

Source Broodstock

Source Broodstock

Source Broodstock

 

SfoI8 167
169
171
173
175
177
179
181
183
185
187
189
191

Sfol 108
110
116

0neu9 224
228
230
232
N

0neu10 170
174
178
N

0gola 142
144
146
148
150
152
154

0.000 0.000
0.000 0.000
0.508 0.366
0.000 0.000
0.016 0.004
0.000 0.000
0.008 0.009
0.361 0.451
0.057 0.112
0.033 0.045
0.016 0.013
0.000 0.000
0.000 0.000
61 1 12

P=0.]06'

0.000 0.000
0.979 0.974
0.021 0.026
4 7 76
P=1.000

0.000 0.000
0.992 0.934
0.008 0.046
0.000 0.020
66 76
P=0. 031

0.000 0.007
0.708 0.601
0.292 0.392
48 74
P=0. l 74

0.000 0.013
0.193 0.256
0.000 0.019
0.000 0.058
0.493 0.481
0.313 0.173
0.000 0.000
75 78
P=0. 000

0.009 0.000
0.000 0.010
0.536 0.510
0.018 0.019
0.009 0.055
0.000 0.000
0.018 0.000
0.345 0.271
0.000 0.010
0.009 0.039
0.055 0.081
0.000 0.006
0.000 0.000
55 [55

P=0.032

0.036 0.015
0.882 0.924
0.082 0.061
55 66
P=0.490

0.007 0.000
0.963 0.932
0.000 0.038
0.030 0.030
67 66
P=0. 078

0.000 0.000
0.731 0.846
0.269 0.154
52 65
P=0. 033

0.000 0.000
0.078 0.062
0.000 0.000
0.000 0.000
0.719 0.800
0.203 0.138
0.000 0.000
64 65
P=0. 295

89

0.000 0.000
0.008 0.000
0.562 0.562
0.015 0.008
0.008 0.044
0.008 0.000
0.000 0.000
0.308 0.228
0.000 0.062
0.008 0.003
0.085 0.083
0.000 0.008
0.000 0.003
65 I 93

P=0. 000

0.057 0.027
0.877 0.900
0.066 0.073
61 75
P=0. 483

0.000 0.000
0.955 0.927
0.000 0.053
0.045 0.020
33 75
P=0.063

0.000 0.000
0.902 0.807
0.098 0.193
46 75
P=0. 063

0.000 0.000
0.090 0.039
0.000 0.000
0.000 0.000
0.701 0.671
0.209 0.283
0.000 0.007
67 76
P=0. 105

Table l (cont’d)

 

 

 

 

 

Population
Lewis Isle Apostle
Locus allele Lake LLW Royale SIW Islands SAW
Source Broodstock Source Broodstock Source Broodstock
Scou19 159 0.000 0.000 0.007 0.015 0.000 0.000
161 0.128 0.057 0.100 0.039 0.111 0.174
163 0.000 0.000 0.000 0.003 0.016 0.000
165 0.020 0.039 0.029 0.018 0.016 0.013
167 0.027 0.022 0.000 0.000 0.016 0.000
169 0.000 0.000 0.000 0.000 0.016 0.000
171 0.176 0.250 0.300 0.352 0.278 0.265
173 0.020 0.000 0.021 0.048 0.000 0.020
175 0.527 0.478 0.429 0.473 0.468 0.465
177 0.068 0.061 0.029 0.018 0.024 0.040
179 0.034 0.092 0.079 0.027 0.056 0.020
181 0.000 0.000 0.007 0.006 0.000 0.000
N 74 114 70 I66 63 198
P=0.006 P=0.044 P=0.002
Ssa85 126 0.000 0.000 0.125 0.090 0.045 0.049
130 0.000 0.000 0.000 0.000 0.000 0.003
134 0.419 0.403 0.456 0.500 0.604 0.657
136 0.118 0.146 0.118 0.139 0.112 0.098
138 0.441 0.447 0.301 0.271 0.239 0.193
140 0.022 0.004 0.000 0.000 0.000 0.000
N 68 113 68 166 67 194
P=0.433 P=0.536 P=0. 726
Sf012 254 0.047 0.027 0.127 0.142 0.061 0.041
256 0.057 0.040 0.032 0.052 0.045 0.081
258 0.877 0.920 0.841 0.799 0.894 0.858
260 0.009 0.000 0.000 0.007 0.000 0.020
262 0.009 0.013 0.000 0.000 0.000 0.000
N 53 75 63 67 66 74
P=0.639 P=0.679 P=0.210
0golc 213 0.079 0.140 0.024 0.096 0.032 0.046
219 0.421 0.570 0.683 0.640 0.645 0.620
221 0.500 0.290 0.294 0.263 0.323 0.324
223 0.000 0.000 0.000 0.000 0.000 0.009
N 70 50 63 57 31 54
P=0.004 P=0.055 P=I.000
H02 0.396 0.436 0.380 0.370 0.355 0.392
H93 0.422 0.448 0.427 0.410 0.387 0.411
A4 3.0 3.1 3.2 3.3 3.1 3.2
F5 0.062 0.027 0.110 0.097 0.082 0.046
5.7," 0.006 -0002 0.013 -0003 -0009 -0009
5 0.064 0.057 0.100 0.072 0.080 0.051
U9 P=0.539 P=0.256 P=0.829

90

Table 1 (cont’d)

 

l. P values of exact test for significant differences in allele frequencies between source and
hatchery broodstock populations.

2. Observed heterozygosity.

3. Hardy-Weinberg expected heterozygosity (N ei 1978).

4. Allelic richness (average number of alleles standardized for sample size).

5. Wright’s (1951) inbreeding coefficient.

6. Average Coefficient of Relatedness (Queller and Goodnight 1989).

7. Proportion of coefficients of relatedness at the full sibling level (P<0.05).

8. Mann-Whitney U test for significance of difference in location (i.e., average) of the
distributions of coefficients of relatedness.

91

Table 2. Intrastrain comparisons of allele frequencies and measures of genetic diversity of the

 

 

 

 

Marquette broodstock (SMD).
Broodstocks
Locusl allele SMD812 SMD3 STW4 Locus allele SMD81 SMD STW
Sfol8 171 0.583 0.599 0.616 Scou19157 0.000 0.005 0.000
173 0.014 0.000 0.000 159 0.000 0.005 0.000
175 0.014 0.041 0.014 161 0.090 0.122 0.199
181 0.319 0.275 0.261 165 0.000 0.009 0.000
183 0.000 0.005 0.007 169 0.000 0.005 0.015
185 0.000 0.005 0.000 171 0.231 0.275 0.221
187 0.042 0.068 0.065 173 0.038 0.014 0.029
189 0.028 0.009 0.036 175 0.526 0.437 0.426
N 36 111 69 177 0.077 0.086 0.088
179 0.038 0.045 0.022
Sfol 108 0.000 0.040 0.058 N 39 111 68
110 0.936 0.905 0.819
116 0.064 0.056 0.123 Ssa85 126 0.013 0.018 0.036
N 39 63 69 134 0.613 0.694 0.667
136 0.050 0.063 0.043
Oneu9 224 0.000 0.008 0.000 138 0.325 0.225 0.254
228 0.943 0.903 0.956 N 40 111 69
230 0.043 0.065 0.022
232 0.014 0.016 0.000 Sf012 252 0.014 0.000 0.000
234 0.000 0.008 0.022 254 0.111 0.048 0.181
N 35 62 68 256 0.083 0.040 0.022
258 0.792 0.889 0.768
0neu10 166 0.000 0.000 0.008 260 0.000 0.016 0.029
174 0.865 0.893 0.818 262 0.000 0.008 0.000
178 0.135 0.107 0.174 N 36 63 69
N 37 56 66
0gola 144 0.150 0.087 0.101 0golc 213 0.076 0.059 0.059
150 0.738 0.762 0.775 219 0.682 0.686 0.743
152 0.112 0.151 0.109 221 0.242 0.255 0.169
154 0.000 0.000 0.014 223 0.000 0.000 0.029
N 40 63 69 N 33 51 68
H05 0.353 0.374 0.360
17,6 0.387 0.373 0.405
A7 2.7 3.0 3.0
F9 0.089 -0002 0.112
”‘1’? -0.008 -0017 0.003
s 0 0.079 0.061 0.068
U” P=0.484 P=0.34l”

92

Table 2 (cont’d)

 

1. Differences in allele frequencies observed between SMD81 and STW broodstocks for SfoI
(P=0.027) and Sf012 (P=0.039), and between SMD and STW broodstocks for Sf012 (P=0.002).

2. SMD81 represents 1981 broodstock yearclass.

3. SMD, represents current broodstock year classes, 1987 and 1988.

4. STW, represents new broodstock developed from SMD broodstock feral fish.

5. Observed heterozygosity.

6. Hardy-Weinberg expected heterozygosity (Nei 1978).

7. Allelic richness (average number of alleles, standardized for sample size).

8. Wright’s (1951) inbreeding coefficient.

9. Average Coefficient of Relatedness (Queller and Goodnight 1989).

10. Proportion of coefficients of relatedness at the full sibling level (P<0.05).

11. Mann-Whitney U test for significance of difference in location (i.e. average) of the
distributions of coefficients of relatedness.

12. Difference between distributions of the SMD81 and STW broodstocks was insignificant
(P=0.892).

93

 

 

 

 

 

 

nmod Rod «3 .o md— .o mvod oNod oNod oood vmod nvod a:
mood w d od Nmod wo— .o god ovod m d od mvod owod owod n:
momd mnvd onmd momd wmvd movd hmmd Smd wbvd nmvd m:
mmod mvod oood o_od Edd oNod oo_.o Sod mood Sod mn—
de Nmmd wand ohmd ummd momd Rmd mdvd MKNd mbmd 2L
oood oood Sod oNod Nood mood oood wood oood mood oo—
oood oood oood mood oood oood oood god oood oood no—
omod Sod oood oood NSd m_od o_od oood oood oood mo—
o_od mood oood oood oood oood oood oood oood oood mo—
nood omod ovod m2 .o E ~ .o E.— .o SNd ommd mm d .o 2 _.o S:
NNod m _o.o oood oood oood oood oood oood oood mood on—
oood oood oood oood oood oood oood oood oood mood hm _ 338%
mmmdum om o .oum mmodum N: dnm 13 out
9: m2 «3 do vow m2 vow w: va N: 2
oood oood oood oood mood mood oood oood oood oood 3—
wood oood oood n d od oNod wood oood oood mood oood dw—
o_ _.o Sod o— _.o ovod _o_d mwod mood oood oSd wood hw—
oNod dmod oood oood Sod mood oood oood wood wood mm—
m _o.o oood oood oood oNod mood mood oood wood wood mfi
onmd Kmd vomd owed NmNd wmmd wood omod :Nd mhmd 5—
oood oood oood oood Nood oood oood oood oood oood a:
ovod mmod wood wood Sod vvod memo vomd mnod god m:
oood Sod mNod mNod wood wood oNod NNod mood oood m:
and and and movd nde Nomd mend wind nmmd and :—
oood o_od oood oood oood oood oood oood oood oood of «53%.
.82; 333‘ 482:. 33.5, .52... 333‘ .52: $3.5 .55.. 333‘ 30:.“ «:84
Em 340 >9<m 34m QEm
mmz :33 no: mmZ x080 Magoo
8.88955
.zaowoa 98 385303 beacon: Bo b:3mtm> uwoeow oo 8.5308 98 885395 20:“ do mnemEQEoU .m 2an

94

$85338 oo $56808 he Ecosofiflo

05 oo Aowfiga red 5502 E 885$? we ooaaocawmm 8o 68 D $833-:52 .

.Amodvmv _o>o_ wad—em =3 05 on 326058 he 858808 .3 5.58qu .

.Aowa 5ng 98 8:25; wagons—om oo 8205000 owfio>< .

588E666 38655 c 3: 5.583 .

.Aofim 3958 So oufiofioesm £233 oo Sofia: owfiuio magnet ozo=< .

.Awho— BZV atmowzneaon @2398 wu3£o3->o82 .

56096922 34,830 .
Econ—soon xooumooefi 400:8“:

NMV‘OOFM

98 08:8 52809 86833.“ £33 5 mooqouobfi ogocfiwa 8o 68 ~85 oo 83.; Q .—

 

 

oNn duk ooo .on& R. o dum woo duk wmm .oul ob
Sod mood Nwod owod wNod owod Nmod wood wood mwod who
oood Sod- wood- oood mood- mood- hood Nod- Sod- “odd 0 a
omod- oNod Sod- do“ .o- _ 5d- mwod- w_o.o- owod- mood mmod- no
o.w no o.w o.w o.o no o.w 5m o.w o.o cw
Nood Nmod m _ o.o owod mhmd owod momd wmmd w _o.o wood not
mwod owod w_o.o hood ommd wood oood wwmd wNod wmod we:
3%.on ooodnm o3. out moodum Smduk
m3 oo~ as o2 ooN w: wow u: w: ~: 2
med :Nd and omwd 5w _ .o me .o oood ommd woNd mNNd mm—
om _.o om _ .o hmod owod wood wood N. _ o.o oood wmod mood on—
ohwd oood oood momd hwod hood wnwd ohwd thd wood ww—
oood oood oood oood oood mood oood wood oood oood om ~
oood oood nmod mood mmod owod mood oood hmod w d o.o o3 35mm.
mmo out ooodnm 34 out Nooduk oo~ .onk
SN oo~ m: NoN Rom my: wow m: 3% N: 2
oood oood oood oood Nood mood oood oood oood oood m2
oood oood oood oood oood oood oood oood oood oood $—

 

6.583 m use.

95

£09385 £000 Bo $0395.80 338a 05 9 2175 55850 $.33me warn—n50 3 doom—30.00 ._

 

 

 

- - .. - - - - - and «mm oww 3.3
nod «fimm dwd ow mw on w ~_ ~ oo._ oom oom 92m
o2 own—.5 Ed no 2. mm on www do; omo omo 34m
mm: ommdm wmd No— - mm o2 wwm 3d oho Noo 35
mm _w.m o_d mm 3 E wm wNN and own mwm 340
gm _wm.mw Gd wow mwm mfl wk. omw 3d moo odd 3<m
.88388 6 $3 222 5z .02 38m 288 38m 9: 3 88 5.23880 86:3an seam
:38, 0:3an 3808a :33.
$0305 oo “38:2 0>o00bm flou— wwm we wfimmoxm «mom

 

.Afiousowneon. fioqaoa 8 Bonus 217C 80:55:00 3538 05 Aw 98 A70 ooze—$6 32 08.22% 8o E0305 he
Sofia: 0>o00b0 “003850 An $30.33: .3 wwwo oo wEmeo Boa soofioeom E03038 05 2 008w wet—£580

$030363 oe 838:: 3883 05 AN 4009580 £3232: me 338:: .88 2 oo 885928 58% boneam .w 030...

96

.8833 .35on 3 $02 magnum no @853 wwwm .m
.39» 05 mo 8:5 9593 S E x83. 9 BEG E Erna—ow wwwo mo wanes: .N

dam—man m3“ 8 Sta wounnmbmfi «on 825 ammo Mon? ESE—Ego 13:88.3 £3852 ._

.33 E @0260:— 803 £033

wwwo xoogmvoofi 92m 05 8m E85 mom: 5 3:3an 3:65 we :83 2: fl couusga wwu =< ._ 8 88 .8 >9: 80938.“
58033800 .32 magnum Bob mwwo mo counnEmB 25 E 3126 2 26 #023005 :80 .8.“ Spas: “£82558

5958 3883 05 3:3 8: >5: £098an :08 B.“ mEaumoa c5 motion.“ $88 9.83.580 inn—80a .«o 335:: .88. “302

 

 

soda; - .. - - - - Nwm >23

- - - 5.82 _ 808$ - :2va 8m 9%

- - €3va 32:8 Gwefiv - @982 o8 25m

- - A8438 - - - - one am

- - A333 - - - - 2% >50

- 82% $.80? Gmeovm - :38me - 8a 3%
”.583 News Q0539; 52 52 $5 52 323580 £8§85

E

 

88“on

Std 53—. 83d :8— 22.3mm

zaonwo=< 3:88a—

 

movaonm b03033

 

.mfiflwoa EquwmcmE 98 33:8.“ was": 9883 9 noted—wen wmo 388955 SaauAaEuc

mazzméumv Bob 83 .3 5955mm. 05 33:08am: $835.50 3888 .3 80.5389 98 $3552 .m oEaH

97

.8— :ogmam some 8 conga 8 Am 055.5 232 :23 £0305 mo 83:5:
about“. was—a8 588:8 88588: E mo=_u> (:2: 3:3an 9 3:553 wwwo :38 mo 5:3qu .v
.838 w5§«% 05 :8 3:3an flop—:3: :38 05 .3 8:98:95 .m
326 53w a 8 3:3an :39 96: >9: :8? 28 55 89:8 83 wfigam ha 356% «53:32 .N

 

 

 

 

 

 

 

 

 

 

 

 

.83 wfigam .—
8 .o 86 a; $3 :3 8 32 3:8: 8332
A28 2 :8 a: .8 :28 E8 wwwm 33 ad
a a. 8 «a 3
32.88 523 _ 323

Ram H .o - 883° - - 83.2 .05: 568:
- - - - £88.: 832 55: 8:
- 6:25 3836 2:83 - “8.2 82:22
58 £8 8M8 5.8 8M8 wwwm ESP :5

3. mm m» mm M:

85:8 85>: SEE

- 888: - - - - 898m :85 9:68,:
- 8825 £82 .o c an; $88... «586 8:33 53: =86:
- 888.: - - - - 833 :52 no:
58: - - - - - 80.8 5&5?
no.8 688 30.8 95.8 5.8 M8:8 8»: 33. 3.5

2 8.. mm on NE N3

gas 524.2: 8:88 $883
dogamm

£382: :23 98 no.3 magnum 2.38% 9 359.53 E385: wwo 38 .«o ”83:30:“ 05 v5
.mxooumcoos >853: 3:: :8 magma: 98 82:8.“ >558: 8 :32:me flop—:3: wwo 30H .0 03m...

98

 

 

..o.o wood o.o.o mood. mmod- owod- o.o.o o.w.o oovd Smd owod. $83. 2.8%.
.mod wuod- mood- m.o.o mmod Rod vmod «dd 8.3 omvd o.w.m V.03. Squaw
..od- .Nod o.o.o .m . .o S. d- wood Sod oNvd o.wd .oNd NN . .m 0.9.3. o.w.
20 go to 6 d “u so 5 x. v... «3% as. 2.3
wood- w.od m.od oood- wood- ~.o.o- owod omvd m .vd o.w.o movd 92m. o2.2.3. o.w.
.Nod owod- .ood ..od- oood ..o.o- .vod wmvd o.w.o o.w.o 3:: $50. 8.3 58w
mmod- 38d oood wood- omod- wmod- owod mmvd oovd mnmd .omd 52m. 8.3.832
Sod- m.o.o .ood ~.od- mood- m .od- wmod envd o.w.o ..v.o 3.0m 32m. 98.... 2.83.
o~od ~3d- .Nod- wmod Sod .mod o. .d movd omvd ovvd mnmd C515. 8.3 809%
.mod hood mmod .ood m.o.o Sod mood oovd omvd wvvd onmd C5... 8.3 £38.
:20 2.6 no ”no no 0.0 no v... m: m... 2%? £8885 @222.

 

.3596 8.“. ..o wagoneg. .88 88.3.2.2. Bob 8.“. 2.3 can
.25.. 83 :8. 2.3 .mxuofiooofi bones: Bob 83. ..o 88.3.2.2. 8.. 3.83... ozoaow ..o wficougn .88 £88. .n 0.8....

99

88888.8 88 38883.8 8.8.3 98. o. 88.8 8.8.8.. 8.8.3 .808 8 8.3.8.8. 8. 8.8 .8 8.88.8880 ...

.8888 8.8.3 88 M8. 8 86 888.8... 8.8.3 .88 8 8.3.8.8. 8. 8.8 .8 8.88.880 .o.
888.8... 8.8.3 .88 8 8.3.8.8. 8. 8.8 .8 8.88.880 .o

88888.88 8 888838 8.8. o. 88 3.88.8 .88 8 8.3.8.8. 8. 8.8 .8 8.88.880 .w
5.88.8 E8 M8. 8 .888. 8.838 .88 8 8.3.8.8. 8. 8.8 .8 8.88.880 .n

5.9.8.8 .88 8 8.3.8.8. 8. 8.8 .8 8.88.880 d

88.3.8.8. 88.8 :8... 8.3.8.8. 8. 8.8 .8 88983.8 3.3.8 ow38o>< .m

88.3.8.8. 8. 8.. 88...? 3.88.8 .88.. .w

88.3.8.8. 8. 8.. 88.8.3 5.838 .382 .m

5.89.8888. 888.88 .8 8388mm. .N
.oN.m 8.988 88:85 8.8 m. 2 88.3 .ZN a88.88 8.83.83. .53. .3 8 8.8. V 88.8.. o..o..< ..

 

 

oood o.o.o o.od oood .ood- .ood- mNod mmwd .-.w.o o.w.o ow~.m 88.8... 88.8 88.30
oood- mood .ood- Nood- mood- wood- w.o.o mmwd m.wd N.w.o mm..m 88.88% 883.8. 88.30
mood Rod- mood .ood- wood wood wmod Rwd owwd nmwd wwdd 883885 8.83. 83.3%
.-oo.o- o.od- mood- mood wood wood .mod owwd ond omwd 3.8 .88.-. 8.83. .8835
mood- .ood- mood- .ood- mood- wood- oNod mmwd m.w.o nowd 3.2 883886 83.8. 8.888...
mood Nood- mood- .ood oood- wood- omod m.w.o wowd nwmd wood :38. .833. 8.888...
oood oood- oood .ood- oood .ood- owod wmwd w.w.o Swd wowd 88.8.: 89.3. 8.8.
mood- wood wood. .ood- mood .ood owod .mwd .Nwd mmwd mm ..m 88.88% 8.9.3. 8.8.
o.o.o- .ood dood- mood mood oood .mod wmwd wad owwd mmo.m 88.. 0.83. 8.8.

28 am to 8d 6 Ho .8 8.. m8. 8... .88. so: 23 2E

 

6.8.88. . 858

100

Table 8. Summary of F -statistics partitioned genetic variation within and among hatchery
and wild populations of lake trout (Salvelinus namaycush) from the upper Great

 

 

 

 

Lakes.

Populations Loci F l f 2 053 0M4

Hatchery Sf018 -0.012 -0.086 0.068" -
Sfol 0073* -0.226 0.094" -
0neu9 0.055 0.051 0.004" -
OneuI 0 -0.054 -0.1 16 0.055” -
0g01a 0086* -0.029 0.111" -
Scou19 0.023 -0.005 0.028" -
Ssa85 0.006 -0.056 0.059” -
Sf012 0.014 -0.020 0.033" -
0golc 0.538“ 0501* 0.074” -
Mean 0.081" 0.002 0.058"
95% CI (0.000 , 0.259) (-0.062 , 0.205) (0.043 , 0.082)

Wild Sf018 0.091" 0.028 0.065M 0.086"

(Lake Superior) Sfol 0.115'" 0.080“ 0.039'" 0.050'Ml
0neu9 0.015 -0.002 0.017“ 0.019“
0neu10 -0.038 -0.051 0.013 ** 0.011
0g01a 0.043 0.030 0.014" 0.015"
Scou19 0.046“ 0.034 0.013“ 0.018"
Ssa85 -0.026 -0.045 0.019M 0.024“I
Sf012 0.022 0.016 0.007 0.007
0g01c 0.657“ 0.647M 0.029" 0.031"
Mean 0.103" 0.082" 0.024" 0.029“
95% CI (0.011 , 0.312) (-0.010 , 0.296) (0.014 , 0.040) (0.017 , 0.051)
All wnd5 0.078“ 0.103" 0033'"
95% CI (0.012 , 0.312) (-0.009 , 0.291) (0.017 , 0.042)

 

Note: *P<0.05; "P<0.01

1. Allelic variance among all individuals F it.

2. Allelic variance among individuals within populations Fig.

3. Allelic variance among populations within each morphotype (leans,
bumpers, siscowets).

4. Allelic variance among lake trout morphotypes.

5. All wild includes Parry Sound population from Lake Huron.

101

.nmw 35w 305 Bob @0335“. 3303335 ma? xooamuoen 8:3 £304 2F .SNMEBE 3.3 5058: Set
36028 20? >3 .oxaq £33 85 coy—09m Bob 875 A .mQOmEana otoaow Sm c0388 flooumgos
33083 mo 983.238 858 2153828“ 25: ENE—m: .mxoSmwooE .3 mega 858 3538 05 8

.692 E02852 .voanam QQBAQEa: «.3533qu Sch 0%: mo mucus—anon 25 ER boss“: mo $8333 ._ 8:me

          

A9 xooumwoofi 8:3 520
A8 xoouuoos 233 2325
A S xooumcoos ofixom 03
a?

63085
:84

33.3%

     

6V xoowmocoum 88352
Amvxoouuoofi BEA £33

9; xoouwoohm moocow _ S 33% Magma IV

 

   

33
5983.“
59:5: «oBooflm 83086
63085 :84 Suva

 

333%

102

32 03.5223: .2 M32 .3 388m .3 W22 25: 33 320 .8 :22 .38»: 2.82 3.5 .2 M82

:23: 33 38m .2 332 .3 3 was: .2 N82 82830 .2 H32 3:82 .2 W32 8250 23 3.3:: .2232 .3 3 8%: .2 N82
.3 3 883E .: .82 3382 .2 622 :23: E3 .332 .o .222 3:23 .m .32 33.3 23 3.35 .3 H22 .38»: 33 388:... .0
N32 33.: 33 3.38:... .3 £2 .3 3 3335— .v 52 .3 3 32.33 .3. .82 com 93 32.5 .N 322 23550 3393: 38.32 .2

 

 

 

 

U

 

 

 

 

+

 

A
A

00:002.: U

 

 

 

 

 

8.2.2.2.2.m::00:8:02 23:03:95me 02:02:.
0322:2300..— :0m 032:2 2.2:0nm:_e:00 :Rbm
£.2.%:0zm_::0m wfiummxm 2822300; wfixoofi
.33: 2.339283 33.332

3.3—233.0% :0 wE—oom 2.”.mmmvofi0z :033N520m
223—030:3 2.».nemoumm mfifigam

0__3:00:D 230:0M x0m 30063005

22: Exam 2.263056% 0:00:52

nzmfim 8.0322352 203552

tdasiufifigm :0 w:=00m 3.000502 gunfimtom
_~.:.z.a_3=_00:3 3.330033% mgmqm
3.3330032 30038.8 mound x0m

2.2.0302 22:0>m wfiafiam

2.. 3:2:m 0.3003523 23:52

2.2302 3.2... 029:3 gouflaomnsm

+

 

00200005 bosofiIU

 

 

 

 

+ .

 

O... :U

.m 0w8m

N 0w8m

2 09am

 

mw:E0m E0523: :2 302.30%
b88800 0:056:00

3080352 3:: 2:0EQO_0>0Q
0.0000008m 0: 2:05:03” m0:mm~

208%ng 0:3 2:08:0_0>0D
20080005 E m0w8m

 

38:30:20: 3:0 b28033 0:0w .20 m_0>0_ E m0m§:0 02 32:55:00 830208: 0=:0>:.a 0:3
30036005 20:80: 338.850: 3220303 “:0: 003 m0 #:0832205: w:_.:6 mowfim .N 0325

103

Spawning Scenario: 1:1 male to female matings.

 

 

 

 

 

 

Male 1 2 3 4 5
Female 1 2 3 4 5
011 012 021 022 031 032 041 042 051 052

o] 1“ - e 0 o 0 0 0 0 o o
012 - O O O O 0 0 O 0
021 - 6 O O O O 0 0
022 - O O O O 0 O
031 - 9 O O 0 0
O 32 - 0 0 0 0
041 - 9 0 0
042 - O 0
051 - 6
052 -

Mean 6:0.028

Offspring Unrelated=89%

Spawning Scenario: 5:5 male to female matings, gametes pooled, unequal reproductive variance.

Male 1b 2c 3d 4h 5d

Female 1 2 3 4 5

0n 012 021 022 031 032 041 042 051 052

On - 6/2 6/2 0 9/2 6/2 0 O 9/2 9/2
012 - 0 0 O O 6/2 0 O O
021 - 9/2 6/2 9/2 0 0 6/2 9/2
022 - 0 O O 9/2 0 0
031 - 9 O O 9/2 9/2
032 - 0 O 9/2 9/2
041 - 9/ 2 0 0
042 - 0 O
051 - 9
052 -

Mean 9:0.061

Offspring Unrelated=55%

 

 

Figure 3. Hypothetical spawning scenarios as advocated in the broodstock management plan (1:1 matings;
Holey 1997) and actual conditions realized at the hatcheries (i.e., 5:5). In this simple example,
we assume adult males and females are unrelated and each female contributes two offspring
(e.g., no variance in reproductive success). 9 is the probability that any two alleles shared
between two individuals are inherited from a common ancestor (identical-by-descent). 9=O.25,
offspring related at the level of full -siblings. 6=O. 125, offspring are related at the level of half-
siblings. (a) “01 1” represents offspring. Subscripts designate the female that produced the
offspring (i.e., female 1) and the offspring number (i.e. offspring 1 of 2). (b.) Male 1 and 4 each
contribute 20% of the offspring (Male 1 offspring italicized). (c.) Male 2 contributes 60% of
the progeny (Male 2 offspring in bold). (d.) Male 3 and 5 do not contribute offspring. Mating
strategies should seek to minimize mean 9 among progeny and maximize the proportion of
unrelated offspring.

104

.mv—ooumcoofi bong“: EH
macaw—anon 02:8 How @530 305822 mo 853508 837.23 no 30:35va @536 55:35 .v oSwE

 

1! ll l.|l.l.|||||llullg

océ
mod

o_.o
- .- 2.0
0N6
mmd

 

°\° _

253m 02 n 35 293% 03 n d
82.: 228.3 u 22m $52 0:89... u 3
8:3 £304 H 3.5 E .834 £33“ £304

 

 

 

 

 

 

 

85
ll , 25
z 3.3. o; .x.
lmtsoq. it 26
. . o3
Qmwduk wwd
mood- H 344 dood H £304 ”:32

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

105

.mo—ooamcoos bogus: 2:sz 0:03.82
353805385 co.“ $38-5 $05058 mo 356508 337:3 .«o acorn—swam? E535 xoaosvoum .m oSwE

 

A 1. O .0 v .0
9

.r
O

 

 

 

 

zooumuoofi wad—fl 0.0.52“; H 3km
moan—o 30% #0066095 hwm— fig wwo— H gm
ago 3% #82805 SE n 59%

 

 

 

 

 

 

 

 

 

 

 

 

 

3.5 I
9% I

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

NVM..Q”& . . . WN.C WNd . __

y Sod- H DEM .mood .HEQEM ”502

 

106

$2825“ 98 $33 M08383 .8“ Guava $053.30.. go $56508 bugs—Rm mo $835388 Ammxa$v 5593“,.“ .o onE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3:53;.
Bowl
95de
. . . and
r woo. - n moragzh woo o n 250 back;
1 cod
$2333.. H MAN”
0
Alivll>film1I 26 A.
u emnsum o2
. mud

Tl l. Holocd H may—magshJood- n 35 ”5%

 

.QH .

 

 

 

 

 

 

 

 

r- Bod- u mozco>=hmoow¥ n 3<m ”sale:

   

 

 

83A 580 u 340
253% 2m— " Em
$52 0:83. u 3%
8:3 «Baum H 34m
030332 n 92m

 

 

 

 

 

 

 

 

 

 

flag“
7 lliilllllilli i4
cod
30:32:.- mod a m
§ 35' 25 \o H
m _.o
«Sena 62 “
Bod u moreoéwsd- u 25m ”5%,? - H.
j
Wwwwaogal

 

aim-
and"

m.

 

I Bob- "Mm—€33. .506 "gm :5on a

 

 

 

 

  

 

 

 

107

.83: nabfioon 88292 E3852 Aug: moo—5.0.6 Bone 3.338 van «Ecumazgao no woman macaw—anon
33985:: maéoéamv Bob 83 A8 25 can Adv E383 5 mooaowuo>€ ocean» magnesia moon mafiorgnwmoz .5 oEwE

cam. ocaom Ema

 

39:3: o.o»om o.w.

 

59:3: 9.5m. :ontmo

 

 

 

 

cam. 2.3m. 2302

 

 

 

 

com. 288”". o.w. _ g _ ll
. _ mm

:8. xoom 2255 cc .2

 

 

.mzoomfi o.w»om o.w. , .

62686 0:22 2.82 _ om—

 

 

108

 

 

 

$2683 39m. 8950 3

 

638$.» xoom Emccflm mm

xooamuoos 9.3 moocom

 

 

 

 

 

 

 

 

x0238... 9.3 :85
xooumcooz 9.3 $26.. on
xooumooob Ego”. o.w. — _
xooamuoos $830.55.
cm
1 CV .m

x0988... 95.2 2.82

 

633:0 Sam 5%800 mo mafia? seam BRA 80¢ 03 35

950m Em 80¢ 39: 83 SB— .8m :3er 859m 834 89a 08 macaw—anon =< .83on £98005

So: 83 05 5 RES: wfion huaobso mm $50: 35.3% Bob 83 mo :88me Exam bozfiaqfioo

a 35 33>»: out 2? Anna: 32866 Bozo 3833— ?E «Eommgzgao no woman £83238
3.39853: .353333 Bob 83 E; “5:82 we 3533:. navy—ow $35852 ooh 330.33an Z .w oSwE

cam. venom Ema

 

.maEzc $.96“. 2m.

 

.383: 95.2 3850 L

 

 

:52 95.2 ¢=m0a<
35222 :3. 2.23”. 2m. _ III
c_ 2.5: , :3. :03. 23:05
62683 2.361 2m. 1
$2685 956. mzmoad. I
338$» new? :ontmo _

638mm xoom Emccmuw

 

 

 

T

 

 

 

 

 

 

 

109

.Amfifin 00—0— w0080v 5000030 ouocow .«0 000000 0000000w 0.900

05 9005508 5000053386 0.23 000.0 02003000 0280000 0000000. 00 0055000 55:80: waummxo

0023 0.0900005 000 000.0000 09 00 02000 .00 00008:: 00000.80. .8.“ 030:0 N 330.0% 0.0900005

000:0m 05 $058508 0:0 0.0900005 00.096 00.04 0:0 0035.2 00.04 03009000 0 .00 000800.050

$200960 .3800...“ 03005003 :0 0000.00 N 3000bm 0300000005 b05000: means :0 .00 00:

05 00000260 fl $000.0m 0030000005 E0500: 5 3.00030 0.000% 3.0.8.088 00.“ 00.30050 03.0000

030 0000003: 005 05H A303 00000008 000:0 000030m 0:0 000050.300 :0 .0000 8.0900003
b05000 3.300950: 30.03303 80b 00.0. .00 0000m00>=0 0.000% muggy—000 005 magovuonnwfiz .0 00%..»

0.0900005 00.0.. 000c0w

 

x
0.0900005 00.0.. 0005
0.0200005 00.0.. 0.30...
0.0900005 0.03m. 0.0.

 

 

 

 

0.0900005 0.003905.

><><><><><><

 

0.0900005 000.0. 0.0009...

 

N >m0u0bw .. 30.0.5

3.0.020. 0N_E_x0_2 I .000 0008000005.

110

LITERATURE CITED

Allendorf, F. W. and S. R. Phelps. 1980. Loss of genetic variation in a hatchery stock of
cutthroat trout. Transactions of the American Fisheries Society 1092537-543.

Allendorf, F. W. and R. F. Leary. 1986. Heterozygosity and fitness in natural populations
of animals. Pages 57-76 in ME. Soulé, editor. Conservation biology, the
science of scarcity and diversity. Sinauer Associates, Inc., Sunderland, MA.

Allendorf, F. W. and N. Ryman. 1987. Genetic management of hatchery stocks. Pages
141-159 in N. Ryman and F. Utter, editors. Population genetics and fishery
management. University of Washington Press. Seattle, WA.

Allendorf, F. W. and M. M. Ferguson. 1990. Genetics. Pages 35-63 in CB. Schreck and
P. B. Moyle, editors. Methods for fish biology. American Fisheries Society,
Bethesda, MD.

Allendorf, F. W. and R. S. Waples. 1996. Conservation and genetics of salmonid fishes.
Pages 238-280 in J .C. Avise and J .L. Hamrick, editors. Conservation Genetics:
Case Histories From Nature. Chapman & Hall, New York, NY.

Amos, B. and A. R. Hoelzel. 1992. Applications of molecular genetic techniques to the
conservation of small populations. Biological Conservation 61:133-144.

Anders, P. J. 1998. Concervation aquaculture and endangered Species: Can objective
science prevail over risk anxiety? Fisheries 23(11):28-31.

Angers, B., L. Bernatchez, A. Angers, and L. Desgroseillers. 1995. Specific
microsatellite loci for brook charr reveal strong population subdivision on a
microgeographic scale. Journal of Fish Biology 47(Supplement A): 177-185.

Avise, J. C. 1994. Molecular markers, natural history and evolution. Chapman & Hall,
New York, NY.

Beaumont, M. A. and M. W. Bruford. 1999. Microsatellites in conservation genetics.
Pages 165-182 in DB. Goldstein and C. Schlotterer Microsatellites, Evolution
and applications. Oxford University Press, Inc., New York, NY.

Berst, A. H. and G. R. Spangler. 1973. Lake Huron, the ecology of the fish community
and man’s effects on it. Great Lakes Fishery Commission, Technical Report,
Number 21.

Blaisdell, EA. 1993. Statistics in practice. Saunders College Publishing, New York, NY.

111

Blouin, M. S., M. Parsons, V. Lacaille, and S. Lotz. 1996. Use of microsattelite loci to
classify individuals by relatedness. Molecular Ecology 5:393-401.

Bronte, C. R. 1993. Evidence of spring spawning lake trout in Lake Superior. Journal
of Great Lakes Research 19:625-629.

Bronte, C. R., J. H. Selgeby, J. H. Saylor, G.S. Miller, and N. R. Foster. 1995. Hatching,
dispersal, and bathymetric disribution of age-O wild lake trout at Gull Island Shoal
complex, Lake Superior. Journal of Great Lakes Research. 21 supp.1:233-245.

Brown, Jr., E. H., G. W. Eck, N. R. Foster, R. M. Horrall, and C. E. Coberly. 1981.
Historical evidence for discrete stocks of lake trout (Salvelinus namaycush) in

Lake Michigan. Canadian Journal of Fisheries and Aquatic Science 38: 1747-
1758.

Brown, B. L., T. P. Gunter, J. M. Waters, and J. M. Epifanio. 2000. Evaluating genetic
diversity associated with propagation — assisted restoration of American shad.
Conservation Biology l4(1):294-303.

Bumham-Curtis, M. K. 1993. Intralacustrine speciation of Salvelinus namaycush: an
investigation of genetic and morphological variation and evolution of lake trout
in the Laurentian Great Lakes. Doctoral dissertation. University of Michigan,
Ann Arbor, Michigan.

Bumham-Curtis, M. K., and eight coauthors. 1995. Genetic startegies for lake trout

rehabilitation: a synthesis. Journal of Great Lakes Research. 21 (supplement 1):
477-486.

Bumham-Curtis, M. K., and C. R. Bronte. 1996. Otoliths reveal a diverse age structure
for humper lake trout in Lake Superior. Transactions of the American Fisheries
Society 125:844-851.

Busack, C. A. and K. P. Currens. 1995. Genetic risks and hazards in hatchery operations:
fimdamental concepts and issues. American Fisheries Society Symposium
15:71-80.

Cavalli-Sforza, L. L. and A. W. F. Edwards. 1967. Phylogentic analysis: Models
and estimation procedures. Evolution 21:550-570.

Campton, D. E. 1995. Genetic effects of hatchery fish on wild populations of Pacific
sahnon and steelhead: What do we really know? Pages 337-353 in H. L. Scramm,
Jr. and R. G. Piper, editors. Uses and effects of cultured fishes in aquatic
ecosystems. American Fisheries Society, Symposium 15, Bethesda, Maryland.

112

Chesser, R. K. 1991a. Gene diversity and female philopatry. Genetics 127: 437-447.

Chesser, R. K. 1991b. Influence of gene flow and breeding structure on gene diversity
within populations. Genetics 129: 573-583.

Coberly, C. E. and R. M. Horrall. 1982. A strategy for re-establishing self-sustaining
lake trout stocks in Illinois waters of Lake Michigan. Institute for Environmental
Studies, Marine Studies Center, University of Wisconsin, Madison, Illinois
Coastal Zone Management Developemnt Project, Report Number 42.

Cornelius, F. C., K. M. Muth, and R. Kenyon. 1995. Lake trout rehibilitation in
Lake Erie: a case history. Journal of Great Lakes Research 21(supplement 1):
65-82.

Cross, T. F. and J. King. 1983. Genetic effects of hatchery rearing in Atlantic salmon.
Aquaculture 33:33-40.

Crozier, R. H. Genetic diversity and the agony of choice. Biological Conservation. 61 :1-
10.

Crozier, W. W. 1998. Genetic implications of hatchery rearing in Atlantic salmon: effects
of rearing environment on genetic composition. Journal of Fish Biology 52: 1014-
1 025.

Curtis, G. L. 1990. Recovery of an offshore lake trout (Salvelinus namycush) population
in eastern Lake Superior. Journal of Great Lakes Research 16:279-287.

Dehring, T. R., A. F. Brown, C. H. Daugherty, and S. R. Phelps. 1981. Survay of genetic
variation among eastern Lake Superior lake trout. Canadian Journal of
Fisheries and Aquatic Sciences 38: 1738-1746.

Dodson, J. J .,and nine coauthors. 1998. Elements in the development of conservation
plans for Atlantic salmon (Salmo salar). Canadian Journal of Fisheries and
Aquatic Sciences 55(Supplement 1):312-323.

Doyle, R. W., C. Herbinger, C. T. Taggert, and S. Lochman. 1995. Use of DNA
microsatellite polymorphism to analyze genetic correlations between hatchery and
natural fitness. Pages 205-211 in H. L. Scramm, Jr. and R. G. Piper, editors. Uses
and effects of cultured fishes in aquatic ecosystems. American Fisheries Society,
Symposium 15, Bethesda, Maryland.

Elrod, J ., R. O'Gorman, C. P. Schneider, T. H. Eckert, T. Schaner, J. N. Bowlby, and

L. P. Schleen. Lake trout rehabilitation in Lake Ontario. Journal of Great Lakes
Research 21(Supplement 1):83-107.

113

Eshenroder, R. L., E. J. Crossman, G. K. Meffe, C. H. Olver, and E. P. Pister. 1995. Lake
trout rehabilitation in the Great Lakes: an evolutionary, ecological, and ethical
perspective. Journal of Great Lakes Research 21(Supplement 1):518-529.

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 Resaerch
21(Supplement l):108-127.

Eshmeyer, P. H. 1957. Contributions to the study of subpopulations of fishes. US. Fish
and Wildlife Service Special Scientific Report on Fisheries No. 208:128.

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:
67-74.

Faith, D. P. 1992. Conservation evaluation and phylogenetic diversity. Biological
Conservation 61 :1 1-15.

Fetterolf, C. M. 1980. Why a Great Lakes Fishery Commission and why a Sea Lamprey

International Symposium. Canadian Journal of Fisheries and Aquatic sciences 37:
1588- 1 593.

F itzsimons, J. D., S. B. Brown, D. C. Honeywell, and J. G. Hnath. 1999. A review of
early mortality syndrome (EMS) in Great Lakes salmonids: Relationship with
thiamine deficiency. Ambio 28(1):9-15.

Fitzsimons, J. D., L. Vandenbyllaardt, and S. B. Brown. 2001. The use of thiamine and
thiamine antagonists to investigate the etiology of early mortality syndrome in
lake trout (Salvelinus namaycush). Aqautic Toxicology 52(3-4):229-239.

Fiumera, A. C., L. Wu, P. G. Parker, and P. A. Fuerst. 1999. Effective population size in
the Captive breeding program of the Lake Victoria cichlid Paralabidochromis
chilotes. Zoo Biology 18: 215-222.

Flagg, T.A., C. V. W. Mahnken, and K. A. Johnson. 1995. Captive broodstocks for
recovery of
Snake River sockeye salmon. American Fisheries Society Symposium 15: 81-90.

Frankham, Richard. 1996. Relationship of genetic variation to population size in wildlife.
Conservation Biology. 10(6): 1500-1508.

Gharrett, A. J. and S. M. Shirley. 1985. A genetic examination of spawning
methodology in a salmon hatchery. Aquaculture 47: 245-256.

Gilpin, M. E. and M. E. Soulé. 1986. Minimum viable populations: processes of species

extinction. Pages 19-34 in ME Soulé, editor. Conservation biology: the science
of scarcity and diversity. Sinauer Associates, Sunderland, MA.

114

Goode, G. B. 1884. The fisheries and fishery industries of the United States. Section 1:
Natural history of useful aquatic animals. Report to the US. Commission on Fish
and Fisheries, Washington, DC.

Goodier, J. L. 1981. Native lake trout (Salvelinus namaycush) stocks in the Canadian
waters of Lake Superior prior to 1955. Canadian Journal of Fisheries and Aquatic
Sciences 38(12):]724-1737.

Goudet, J. 2000. FSTAT, a program to estimate and test gene diversities and fixation
indices (version 2.9.1). Available from http://www.unil.ch/izea/softwares/
fstat.html. Updated from Goudet (1995)

Grewe, P. M. and D. N. Hebert. 1988. Mitochondrial DNA diveristy among broodstocks
of the lake trout, Salvelinus namaycush. Canadian Journal of Fisheries and
Aquatic Sciences 45:21 14-2122.

Grewe, P. M., C. C. Krueger, and J. E. Marsden. 1994. Hatchery origins of naturally
produced lake trout fry captured in Lake Ontario: Temporal and spatial
variability based on allozyme and mitochondrial DNA data. Transactions of the
American Fisheries Society 123:309-320.

Hansen, M. J ., and eleven coauthors. 1995. Lake trout (Salvelinus namaycush)
populations in Lake Superior and their restoration. Journal of Great Lakes
Research 21 (Supplement 1): 152-175.

Hard, J. J. 1995. A quantitative genetic perspective on the conseravtion of intraspective
diversity. Pages 304-328 in J. L. Nielsen, editor. Evolution and the aquatic
ecosystem: Defining unique units in population conservation. American
Fisheries Society Symposium 17, Bethesda, MD.

Hedgecock, D., V. Chow, and R. S. Waples. Effective population numbers of shellfish
broodstocks estimated from temporal variance in allelic frequencies. Aquaculture
108:215-232.

Holey, M. E., and eight coauthors. 1995. Progress toward lake trout restoration in
Lake Michigan. Journal of Great Lakes Research. 21(Supplement 1): 128-151.

Holey, M. E. 1997. Broodstock management plan for wild lake trout and brook trout
in the Great Lakes. US. Fish and Wildlife Service, Region 3.

Hynes, J. D., E. H. Brown Jr., J. H. Helle, N. Ryman, and D. A. Webster. 1981.

Guidelines for the culture of fish stocks for resource management. Canadian
Journal of Fisheries and Aquatic Sciences 38:1867-1876.

115

Ihssen, P. H. and J. S. Tait. 1974. Genetic differentiation in retention of swimbladder gas
between two populations of lake trout (Salvelinus namaycush). Journal of
Fisheries Research Board of Canada. 31 :1351-1354

Ihssen, P. E., J. M. Casselman, G. W. Martin, and R. B. Phillips. 1988. Biochemical
genetic differentiation of lake trout (Salvelinus namaycush) stocks of the Great
Lakes region. Canadian Journal of Fisheries and Aquatic Sciences 45:1018-1029.

Jorde, P. E. and N. Ryman. 1995. Temporal allele frequency change and estimation of
effective size in populations with overlapping generations. Genetics 139:1077-
1090.

Kapuscinski, A. R. and L. D. Jacobson. 1987. Genetic guidelines for fisheries
management. University of Minnesota, Minnesota Sea Grant Publication 17.

Kerby, J. H. and R. M. Harrell. 1990. Hybridization, genetic manipulation, and gene
pool conservation of striped bass. Pages 159-190 in R. M. Harrell, J. H. Kerby,
and RV. Minton, editors. Culture of striped bass and its hybrids. Striped Bass
Committee Southern Division American Fisheries Society. Bethesda, MD.

Kincaid, H. L. 1983. Inbreeding in fish populations used for aquaculture. Aquaculture
33: 2 1 5-227.

Kincaid, H. L., C. C. Krueger, and B. May. 1993. Preservation of genetic variation in the
green lake strain lake trout derived from remnant domestic and feral
populations. North American Journal of Fisheries Management 13: 318-325.

Kincaid, H. L., M. J. Gray, L. J. Mengel, and S. Brimm. 1997. National fish strain
registry — trout. U. S. Fish and Wildlife Service and US. Geological Survey,
(electronic database) Washington, C.C.

Kincaid, H. L. 1995. An evaluation of inbreeding and effective population size in
Sahnonid broodstocks in federal and state hatcheries. American Fisheries Society
Symposium 15:193-204.

Krueger, C. C., A. J. Gharrett, T. R. Dehring, and F. W. Allendorf. 1981. Genetic aspects
of fisheries rehabilitation programs. Canadian Journal of Fisheries and Aquatic
Sciences 38:1877-1881.

Krueger, C. C., R. M. Horrall, and H. Gruenthal. 1983. Strategy for the use of lake trout
strains in Lake Michigan. Report of the Genetics Subcommittee to the Lake Trout
Technical Committee for Lake Michigan of the Great Lakes Fishery Commission.

Krueger, C. C., J. E. Marsden, H. L. Kincaid, and B. May. 1989. Genetic Differentiation

among Lake Trout Strains Stocked into Lake Ontario. Transactions of the
American Fisheries Society 118:317-330.

116

Krueger, C. C. and P. E. Ihssen. 1995. Review of genetics of lake trout in the
Great Lakes: History, molecular genetics, physiology, strain comparisons, and
restoration management. Journal of Great Lakes Research 21(Supplement 1):348-
363.

Kumar, S., K. Tamura, I. B. J akobsen, and M. Nei (2001) MEGA2: Molecular
Evolutionary Genetics Analysis software, Arizona State University, Tempe,
Arizona, USA.

Lawrie, A. H. 1978. The fish community of Lake Superior. Journal of Great Lakes
Research 4:513-549.

Lawrie, A. H., and J. F. Rahrer. 1973. Lake Superior: A case history of the lake and its
fishes.Great Lakes Fisheries Commission, Technical Report 19, Ann Arbor, MI.

Lande, R. and G. F. Barrowclough. 1987. Effective population size, genetic variation
and their use in population management. Pages 87-123 in ME. Soulé, editor.
Viable populations for conservation. Cambridge University Press, New York,
NY.

Leberg, P. L. 1990. Influence of genetic variability on population growth: implications
for conservation. Journal of Fish Biology 37(Supplement A): 193-195.

Leberg, P. L. Effects of population bottlenecks on genetic diversity as measured by
allozyme electrophoresis. Evolution 46(2):477-494.

Lewis, P. O., and Zaykin, D. 2001. Genetic data analysis: Computer program for the
analysis of allelic data. Version 1.0 (dl6c). Free program distributed by the
authors over the intemet from http://lewis.eeb.uconn.edu/lewishome/sofiware.html.

Lynch, M. 1997. Inbreeding depression and outbreeding depression. Pages 59-67 in
W. S. Grant, editor. Genetic effects of strying of non-native hatchery fish into
natural populations.

MacLean, J. A., D. 0. Evans, N. V. Martin, and R. L. DesJardine. 1981. Survival,
growth, spawning distribution, and movements of introduced and native lake trout

(Salvelinus namaycush) in two inland Ontario Lakes. Canadian Journal of
Fisheries and Aquatic Sciences 38:1685-1700.

Meffe, Gary K. 1995. Genetic and ecological guidelines for species reintroduction
programs: application to Great Lakes fishes. Journal of Great Lakes Research 21
(Supplement 1):3-9.

Moritz, C. 1994a. Application of mitochondrial DNA analysis in conservation: a critical
review. Molecular Ecology 3 :401-41 1.

117

 

Moritz, C. 1994b. Defining “evolutionary significant units” for conservation. Trends in
Ecology and Evolution 92373-375.

Mueller, G. 1995. A program for maintaining the razorback sucker in Lake Mohave.
American Fisheries society Symposium 15: 127-135.

National Research Council. 1996. Upstream: Salmon and society in the Pacific
Northwest. National Academy Press, Washington, DC.

Nunney, L. and D. R. Elam. 1994. Estimating the effective population size of conserved
populations. Conservation Biology 8(1): 175-184.

Olsen, J. B., P. Bentzen, and J. E. Seeb. 1998. Characterization of seven microsatellite
loci derived from pink salmon. Molecular Ecology 7: 1087-1089.

O’Reilly, P. T., L. C. Hamilton, S. K. McConnell, and J. W. Wright. 1996. Rapid analysis
of genetic variation in Atlantic salmon (Salmo salar) by PCR multiplexing of
dinucleotide and tetranucleotide microsatellite. Canadian Journal of Fisheries and
Aquatic Sciences 53:2292-2298.

Peck, J. W. 1975. Brief life history accounts of five commercial salmonid fishes in Lake
Superior. Michigan Department of Natural Resources, Fisheries Research
Publication 1821, Lansing, MI.

Petit, R. J ., A. E. Mousadik, and O. Pons. 1998. Identifying populations for conservation
on the basis of genetic markers. Conservation Biology 12(4):844-855.

Perez-Enriquez, R., M. Takagi, and N. Taniguchi. 1999. Genetic variability and pedigree
tracing of hatchery — reared stock of red sea bream (Pragus major) used for stock
enhancement, based on microsatellite DNA markers. Aquaculture 173:413-123.

Pfrender, M. E., and five coauthors. 2000. Lack of concordance between genetic diversity

estimates at the molecular and quantitative — trait levels. Conservation Genetics
1:263-269.

Philippart, J. C. 1995. Is captive breeding an effective solution for the preservation of
endemic species? Biological Conservation 72:281-295.

Quattro, J. M. and R. C. Vrijenhoek. 1989. Fitness differences among remnant
populations of the endangered Sonoram topminnow. Science 245:976-978.

Queller, D. C. and K. F. Goodnight. 1989. Estimating relatedness using genetic markers.
Evolution 43:258-275.

118

Rahrer, J. F. 1965. Age, growth, maturity, and fecundity of "bumper" lake trout,
Isle Royale, Lake Superior. Transactions of the American Fisheries Society
94:75-83.

Raymond, M., and F. Rousset. 1995a. An exact test for population differentiation.
Evolution, 49: 1280-1283.

Raymond, M., and F. Rousset. 1995b. GENEPOP (version 1.2): population genetics
software for exact tests and ecumenicism. Journal of Heredity 86:248-249.

Rice, W. R. 1989. Analyzing tables of statistical tests. Evolution 43(1):223-225.

Rees, R. A. and R. M. Harrell. 1990. Artificial spawning and fry production of striped
bass and hybrids. Pages 43-72 in R. M. Harrell, J. H. Kerby, and R. V. Minton,
editors. Culture of striped bass and its hybrids. Striped Bass Committee Southern
Division American Fisheries Society. Bethesda, MD.

Ryman, N. 1991. Conservation Genetics Considerations in Fishery Management. Journal
of Fish Biology 39(Supplement A):211-224.

Ryman, N. and L. Laikre. 1991. Effects of supportive breeding on the genetically
effective population size. Conservation Biology 5(3):325-329.

Soulé, M. E. 1985. What is conservation biology? Bioscience 35 2727-734.

Schram, S. T., J. H. Selgeby, C. R. Bronte, and B. L. Swanson. Population recovery and
natural recruitment of lake trout at Gull Island shoal, Lake Superior 1964-1992.

Scribner, K. T., J. R. Gust, and R. L. Fields. 1996. Isloation and characterization of novel
microsatellite loci: cross-species amplification and population genetic
applications. Canadian Journal of Fisheries and Aquatic Sciences 53:685-693.

Secor, D. H., J. M. Dean, T. .A. Curtis, and F. W. Sessions. 1992. Effect of female size
and propagation methods on larval production at a South Carolina striped bass

(Morone saxatilis) Hatchery. Canadian Journal of Fisheries and Aquatic Sciences
49:1778-1787.

Selgeby, J. H. 1995. Introduction to the proceedings of the 1994 international conference
on restoration of lake trout in the Laurentian Great Lakes. Journal of Great Lakes
Research 2 1 (Supplement 1): 1-2.

Simon, R. C. 1991. Management techniques to minimize the loss of genetic variability

in hatchery fish populations. Pages 487-494 in J. Colt and R. J. White, editors.
Fisheries Bioengineering Symposium 10, Bethesda, Maryland.

119

Spencer, C. C., J. E. Neigel, and P. L. Leberg. 2000. Experimental evaluation of the
usefulness of microsatellite DNA for detecting demographic bottlenecks.
Molecular Ecology 9(10):]517-1528.

Swanson, B. L. and D. V. Swedberg. 1980. Decline and recovery of the Lake Superior
Gull Island Reef lake trout (Salvelinus namaycush) population and the role of sea
lamprey (Petromyzon marinas) predation. Canadian Journal of Fisheries and
Aquatic Sciences 37:2074-2080.

Swofford, D. L. and R. B. Selander. 1981. BIOSYS-1: a FORTRAN program for the
comprehensive analysis of electrophoretic data in population genetics and
systematics. Journal of Heredity 72:281-283.

Taylor, E. B., Z. Redenbach, A. B. Costello, S. M. Pollard, and C. J. Pacas. 2001. Nested
analysis of genetic diversity in northwestern North American char, Dolly Varden
(Slavelinus malma) and bull trout (Salvelinus confluentus). Canadian Journal of
Fisheries and Aquatic Sciences 58(2):406-420.

Rakestraw, L. 1968. Commercial fishing on Isle Royale. Isle Royale National History
Assoiation, MI. Pg. 24.

Ryman, N. 1991. Conservation genetics considerations in fishery management. Journal of
Fish Biology 39(Supplement A):211-224.

Tave, D. 1993. Genetics for Fish Hatchery Managers. Van Nostrand Reinhold, New
York, NY. Pg. 312.

Thomson, J. 1883. A trout trip to St. Ignace Island. In C. F. Orvis and A. N. Cheney
editors. Fishing with the fly: sketches by lovers of the art with illustrations of
standard flies. C. F. Orvis, Manchester, UT. Pg. 299.

Tringali, M. D. and T. M. Bert. 1998. Risk to genetic effective population size should be

an important consideration in fish stock-enhancement programs. Bulletin of
Marine Science 62:641-659.

Utter, F. 1991. Biochemical genetics and fishery management: An historical
perspective. Journal of Fish Biology 39(SupplementA):1-20.

Van-Wright, R. I., C. J. Humphries, and P. H. Williams. 1991. What to protect —
systematics and the agony of choice. Biological Conservation 55:235-254.

Visscher, L. 1983. Lewis Lake lake trout. US. Department of the Interior, Fisheries and
Wildlife Service.

120

Waldman, B., and J. S. McKinnon. 1993. Inbreeding and outbreeding in fishes,
amphibians, and reptiles. Pages 250-282 in N. W. Thomhill, editor. The natural
history of inbreeding and outbreeding. University of Chicago Press, Chicago.

Waples, R. S. 1989a. A generalized approach for estimating effective population size
from temporal changes in allele frequency. Genetics 121 :379-391.

Waples, R. S. 1989b. Temporal variation in allele frequencies: Testing the right
hypothesis. Evolution 43(6): 1236-125 1 .

Waples, R. S., G. A. Winans, F. M. Utter, and C. Mahnken. 1990. Genetic approaches to
the management of Pacific sahnon. Fisheries 15(5): 19-25.

Waples, R. S. 1991. Genetic interactions between hatchery and wild populations—
lessons from the Pacific Northwest. Canadian Journal of Fisheries and Aquatic
Sciences 48(Supplement 1): 124-133.

Weir, B. S. and C. C. Cockerham. 1984. Estimating F-statistics for the analysis of
population structure. Evolution 38:1358-1370.

Wilson, C. C. and P. D. N. Hebert. 1996. Phylogeographic origins of lake trout
(Salvelinus namaycush) in eastern North America. Canadian Journal of Fisheries
and Aquatic Sciences 53:2764-2775.

Withler, R. E. 1988. Genetic consequences of fertilizing chinook salmon
(Oncorhynchus tshawytscha) eggs with pooled milt. Aquaculture 68:15-25.

Wright, S. 1931. Evolution in mendelian populations. Genetics 16:97-159.

Yeager, D. M., J. E. Van Tassel, and C. M. Wooley. 1990. Collection, transportation, and
handling of striped bass brood stock. Pages 29—42 in R. M. Harrel, J. H. Kerby,
and R. V. Minton, editors. Culture and propagation of striped bass and its hybrids.
Striped Bass Committee, Southern Division, American Fisheries Society,
Bethesda, Maryland.

121

"ll11117111111111.1111