Content uploaded by Bror Jonsson
Author content
All content in this area was uploaded by Bror Jonsson on Oct 18, 2017
Content may be subject to copyright.
An Experimental Study of the Reproductive Behaviour and Success of Farmed and Wild
Atlantic Salmon (Salmo salar)
Author(s): I. A. Fleming, B. Jonsson, M. R. Gross and A. Lamberg
Reviewed work(s):
Source:
Journal of Applied Ecology,
Vol. 33, No. 4 (Aug., 1996), pp. 893-905
Published by: British Ecological Society
Stable URL: http://www.jstor.org/stable/2404960 .
Accessed: 17/07/2012 14:23
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp
.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.
.
British Ecological Society is collaborating with JSTOR to digitize, preserve and extend access to Journal of
Applied Ecology.
http://www.jstor.org
Journal
of
Applied
Ecology
1996, 33,
893-905
An
experimental study
of the
reproductive behaviour
and
success of
farmed and wild
Atlantic salmon
(Salmo
salar)
I.A. FLEMING, B. JONSSON, M.R. GROSS* and
A. LAMBERG
Norwegian
Institute
for
Nature
Research,
Tungasletta 2,
N-7005
Trondheim,
Norway;
and *Department of
Zoology,
University of Toronto, Toronto,
Ontario, Canada M5S IAI
Summary
1. Escape of
cultured
organisms
into
natural ecosystems
may
threaten wild
popu-
lations both
ecologically and
genetically.
In
the aquaculture
industry, farmed Atlantic
salmon
(Salmo
salar
L.) often
escape
and enter the
spawning
grounds
of
wild
salmon.
We report experiments
to assess the
competitive
and
reproductive abilities of
fifth-
generation
farmed salmon and their
potential impacts upon
wild salmon.
2. The farmed and wild females had similar levels of
competitive
behaviour; however,
they
differed
in
reproductive
behaviour
and success. Farmed females
displayed
less
breeding behaviour,
constructed
fewer nests, retained
a greater weight
of eggs
unspawned,
were less efficient at
nest
covering,
incurred more nest
destruction,
and
suffered
greater egg mortality
than wild females. As a result,
farmed females had less
than one-third of the
reproductive success of wild females.
3. The farmed males
were even
less successful than the farmed females
in
competition
with the wild fish.
They
were
less
aggressive,
courted
less, partook
in
fewer
spawnings,
and achieved
only
an estimated one to
three
percentage
of the
reproductive
success
of the wild males.
4. The farmed males exhibited
inappropriate mating
behaviour,
that led to poor
fertilization
success,
even
in
the absence of
competition
with wild males.
5. Adult farmed
fish
are thus
likely
to be
relatively
unsuccessful
in
natural environ-
ments
due to a competitive and reproductive inferiority
apparently resulting
from
domestication.
Key-words:
artificial
culture,
captive breeding,
fish
farming, aquatic biodiversity,
breeding
success.
Journal
of Applied Ecology
(1996) 33,
893-905
Introduction
Biodiversity is threatened by
intentional and unin-
tentional releases of
cultured organisms
into
natural
ecosystems (e.g.
Allendorf
1983;
Hindar, Ryman
&
Utter
1991; Ledig 1992).
Animals
derived from cul-
tured
populations may
have both
ecological (e.g. com-
petition,
disease
introductions) and
genetic (e.g. loss
of local
adaptation, genetic
homogenization) impacts
on wild populations, but few data are available on
how cultured and wild animals interact.
Fishes, particularly salmon, are
among the most
intensely cultured organisms. Farmed salmon fre-
quently escape
in
large
numbers from
aquaculture
facilities
and
at
maturity, enter
nearby
rivers where
they may outnumber spawning populations of wild
salmon
(Gausen & Moen 1991; Webb et al. 1991).
Farmed salmon are artificially cultured throughout
their lives. Eggs are collected from broodstock
females, artificially fertilized
with sperm
from brood-
stock males and the offspring
are reared
in
freshwater
hatcheries
for
6 months to 1
year before transfer to
marine net pens, where they
are reared until harvest.
Captive rearing
conditions
combined
with
artificial
selection,
both intentional and unintentional, cause
farmed Atlantic salmon to diverge
from their wild
phenotype through environmental
and eventually,
evolutionary processes (e.g.
Cross & Challanain 1991;
Youngson et al. 1991; Fleming, Jonsson & Gross
1994). Moreover, farmed salmon are often derived
from non-indigenous sources
(e.g.
Cross & Challanain
1991; Gjedrem,
Gjoen & Gjerde 1991). Thus, the
escape of farmed salmon, which
occurs primarily dur-
ing the marine phase, raises both
ecological and gen-
? 1996 British
Ecological Society
893
894
Reproductive
success
offarmed
and wild
salmon
etic concerns for the existence
of wild salmon
(reviewed in Hindar et al. 1991; Hutchings 1991;
Waples
1991).
In
nature,
female Atlantic salmon
compete for ovi-
position
territories within
which they
may sequentially
construct
several nests to form a redd, i.e. area of
disturbed
gravel containing one or
more nests (White
1942; Crisp
& Carling 1989). Each nest contains a
portion
of
the female's
eggs
fertilized
by one
or
more
males.
Construction
of the
nest,
including its structure
and gravel
composition, will
determine embryo sur-
vival during
incubation (reviewed by
Chapman 1988).
Thus, female
reproductive success
is largely deter-
mined by territory access, nest quality, egg fer-
tilization and nest
survival.
Males do not
partake
in
nest
construction, but com-
pete
for
access to
ovipositing
females. The
complex
of
female and male
courting-
and
spawning-behaviour
is
important
for successful oviposition
and fertilization
(Jones 1959). While
females are sexually active
for
only
a few
days,
males
may
be active for a month
or
more
(Webb & Hawkins 1989).
This can result
in
extreme
male
biased operational sex ratios
(i.e. num-
ber
of
sexually
active males
relative
to
sexually active
females).
Thus,
there
may
be intense male compe-
tition,
with competitive ability
determining
male
breeding
success.
Successful
spawning by
farmed
Atlantic salmon has
been documented
(Lura & Sawgrov
1991;
Webb
et al.
1991,
1993a,b)
and evidence of
genetic
intermixing
with wild
populations
found
(Crozier
1993).
Evidence
suggests,
however,
that artificial culture
impairs
natu-
ral
reproductive success
in
salmonids reared
for
sea-
ranching
purposes (artificially
reared in fresh
water
and free-ranged
in the
ocean; Jonsson,
Jonsson &
Hansen 1990;
Leider et al. 1990;
Fleming
& Gross
1993).
It is thus
likely
that
escaped
farmed
salmon,
which
spend a greater proportion
of their life in
culture,
would therefore
incur further reduced
repro-
ductive
success.
Thus,
the
purpose
of
our
study
was to contrast the
breeding
behaviour of farmed and wild
salmon,
and
determine how farmed salmon interact
reproductively
with wild salmon. To do
this,
we
quantified
the
repro-
ductive
ability
of adult
farmed Atlantic
salmon
in
the
presence
and absence of
competition
with
wild
salmon;
and also examined the effects of
competition
with farmed salmon
on reproductive
behaviour and
success
of
wild
salmon.
Materials and
methods
STUDY SITE AND EXPERIMENTAL DESIGN
Experiments
were conducted
in 1990
at the
Norsk
Institute
for Naturforskning
(NINA) Research
Station in south-western
Norway (58059'N,
5058'E).
Four outdoor arenas provided circular stream
environments
that simulate natural breeding
con-
ditions
(Fig. 1). Each stream had
gravel
substrate suit-
able for
embryo
incubation (mean fredle index of
11 25 + 3-84
SD; Lotspeich & Everest
1981). Water
velocities
ranged between 4 and 32 cm
s-1 (measured
15 cm
above the gravel
substrate, every
meter through-
out
each stream),
encompassing velocities
spawning
salmon
occupy
in nature
(Heggberget et al. 1988).
Two floodlights, which
could be adjusted
by dimmer
switches, were directed at
each arena to
provide dim
light
for night-time
observation (x = 11-9
+ 3-6 lx at
water
surface; mid-day,
December
light
range:
1500
lx
[overcast]
to 14 000 lx
[direct sun]).
A 1
x 1
m
grid
of
strings suspended
over the arenas allowed observers
to record fish
positions
and nest locations. Arenas
were
similar,
with
no two differing
significantly
in
physical parameters
(Scheff& multiple
range tests:
P > 0.05).
One experiment
examined competition
between
farmed and
wild Atlantic salmon. Arenas 2
and
3
each
contained
six
female and six male
farmed,
and wild
fish
(Fig. 1). In a second
experiment
attempts
were
made to gather information that
could be used to
separate
effects
of
intergroup (farmed
vs.
wild)
com-
petition from
behavioural differences within
groups.
Twelve
female and 12
male wild fish were
placed in
Arena 1,
and 12 female
and 12
male farmed
fish in
Arena 4.
FISH GROUPS
The farmed Atlantic salmon were
fifth-generation
fish
from the breeding
programme
at Sunndalsora,
Norway.
'Sunndalsora
strain'
salmon are
widely
cul-
tured both
in
Norway,
where
they represent
over
80%
of farmed salmon
in the Norwegian
fish
farming
industry,
and in Great Britain. Salmon of the
1987
brood, originally
derived from
collections
of salmon
from 18
localities
in
1971
(Gjedrem
et al. 1991),
were
reared
in
sea
cages
at the Riska Fisk farm
(59'02'N,
5049'E)
and in
August
1990
transported
to the NINA
Research
Station,
where
they
were maintained
in a
72
m2
freshwater
pool until the start of
experiments
(described
in
Fleming
et al. 1994).
The
wild salmon were collected
in
a fish
trap
100
m
above the outlet of the River Imsa
during
their
spawn-
ing
ascent
in
the
period
July-October
1990. The
trap
was checked
daily
and fish were
transported
to the
Research
Station
where
they
were maintained
in a
72
m2
freshwater
pool until
the start of
experiments.
On 11
November 1990 all experimental
fish were
lightly
anaesthetized with
MS-222, length
measured
(total,
fork and postorbital-hypural),
weighed
and
tagged
with
uniquely
marked
3 4cm diameter disc
tags.
The farmed females
were
heavier,
but
not
sig-
nificantly longer
than the
wild
females
(Table 1).
The
farmed females were
also less variable in body
size
(weight: farmed CV = 19 1%, wild CV = 58 4%;
length: farmed CV = 8
5%, wild CV = 16
7%). The
farmed males were much
heavier and
longer than
wild
? 1996 British
Ecological Society,
Journal
of Applied
Ecology, 33, 893-905
895
LA. Fleming et al.
o.. 0~ ~ ~~~~~~~~~~0
0
outlet4 mv0 00 0
e oule
a t W
Fig. 1.
Experimental
arenas
used to simulate natural breeding
streams. Each
circular stream averaged
2 2 m in width,
was 21
m
in length (measured
as circumference
at mid stream) and provided
47m2
of
spawnable
area filled
with a 36cm
deep,
heterogeneous
mixture of gravel.
Constantly flowing water
from Lake Liavatn
at the headwaters
of the River Imsa was
supplied
to each arena
and water depth
was maintained
at 40 cm (arrows
indicate direction
of water flow). Arenas
were
observed
from a tower, 3 m above ground.
Nests made by farmed
(0) and wild females
(0) are
shown.
Table 1. Mean
body weight (total, g)
and
length
(postorbital-hypural, mm),
with standard deviation in
parentheses,
of farmed
and wild
Atlantic salmon
used
in the
experiments.
Statistics are
analyses
of variance of female
and
male
body
size
among
the
four
arenas and t-tests
of
differences between
populations
Effect
Arena Population
Sex Character Population Arena
1 Arena
2 Arena 3 Arena
4 Total F2,21 P t46 P
Female Weight Farmed - 3767 3535 3742 3696 0
20 0 823
- (414) (781) (818) (706)
2
27 0
030
Wild 2819 2722 3063 - 2856 0 06 0 939
(1902) (1655) (1426) - (1669)
Length Farmed - 559 563 547 554 0
26 0
776
_ (18) (38) (61) (47)
131 0199
Wild 526 519 536 - 527 0
05 0
951
(96) (93) (83) - (88)
Male Weight Farmed - 4036 4329 3835 4008 0
77 0
478
(436) (575) (995) (793) 11-12 <0 001
Wild 2000 2019 1662 - 1920 1
26 0 305
(429) (646) (269) - (466)
Length Farmed - 579 587 567 575 0 54 0 592
_ (21) (26) (50) (39) 976 <0001
Wild 485 482 458 - 477 1
87 0 179
(25) (39) (23) - (30)
896
Reproductive
success offarmed
and wild
salmon
males,
but the
degree
of
body
size variation
was
simi-
lar
(weight:
farmed
CV = 19
8%, wild CV = 24 2%;
length:
farmed
CV = 6
8%, wild CV = 6.3%). Male
and female
fish from
each source
were
sorted
by
body
size and
distributed
such
that there
were
no
significant
differences
among
arenas
(Table 1).
BEHAVIOURAL OBSERVATIONS
Arenas were monitored
continuously,
24 h day'1,
beginning
the
morning
of 12 November
1990.
Activity
within each arena
was recorded
for 30 min,
four
to
seven
times
each
24-h
period
with
at least 3
h between
successive
samples.
Fish
were
identified
individually
by
their tags
and
all instances
of
aggressive,
submis-
sive,
breeding
and cruising
behaviour
were
recorded.
Aggressive
behaviour
included
chasing
(attack
of an-
other
fish without reciprocation),
fighting
(reciprocal
attacks between two fish) and agonistic
display
(aggressive
posturing
towards
another
fish).
Three
forms
of
agonistic
display
were recognized:
(i) lateral
display,
which involved presentation
of the
lateral
side, with
dorsal and anal fins
extended,
to an
opponent;
(ii)
parallel
swimming,
where
two fish
swam
side by
side
for several
metres,
often
with dorsal
fins
erect;
and
(iii)
head-down display,
in which
the tail was
raised,
often breaking
the water
surface,
and the head
was
lowered towards
the
gravel.
Submissive
behaviour
was defined
as fleeing
from an aggressive
opponent.
Male breeding
behaviour
included
courting (attending
a nesting
female)
and quivering
(vibrating
the
body
next
to a female),
while
female breeding
behaviour
included
digging
(a series of body
flexures
while turned
to
one
side;
usually
associated
with
nest
construction)
and
being
courted
(attended
by
a courting
male).
The
final
category
of behaviour
was cruising
which
was
defined
as
swimming
over a
large
area without chasing
or
being pursued.
Onset
of female
breeding
was measured
as the time
from
introduction
into the
arena until
the female's
first oviposition
event,
and duration
as the time
from
the
female's
first to last
oviposition
event.
Male onset
of breeding
was measured
as the time
from
intro-
duction
into the
arena
until
his first observed courting,
and
duration
as time
from first to
last observed
court-
ing.
The frequency
with
which
each individual per-
formed
a given
behaviour
was calculated
as the sum
of the
average
number
of behavioural
acts
per
30
min
each day divided
by
the number
of
days
of obser-
vation. Observations
of general
activity
were
dis-
continued
on 20, 8, 12 and 8 December
for
fish in
Arenas 1, 2, 3 and 4, respectively.
These dates
coincided
with cessation
of
spawning
activity,
with
the
exception
of
Arena
1
in
which
one
female did
not
complete
spawning until
4
January.
Spawning
activity
was monitored
continuously,
24
h day-1,
and focal
data on the
activity of
nesting
females
and all courting
males
in spawning
aggre-
gations
were collected
with
the aid of video
cameras.
A theodolite
was used
to
map the
location
of
nests.
Focal observations
continued
until all spawning
activity
had ceased
on 4 January.
CONDITION AT DEATH
Moribund
fish
were
removed
from the
arenas
and
inspected
for
wounds
and fungal
infection.
Wounds
and fungus
were scored
on a scale from
0 to
3 based
on increasing
body coverage
(0 = none; 1
= small,
localized [<10%]; 2 = moderate; 3 = extensive
[
> 40%]). Gonads
were
removed and
the fish
weighed.
Total weights
of eggs
remaining
in females
were
re-
corded
to assess
the extent
of
breeding
failure.
The
heart
was
removed,
frozen
and
later dried
in a desic-
cating
oven
at 60'C for 48
h before
weighing.
Dry
heart mass relative to somatic
mass was used as a
measure
of physical
condition
(Farrell
et al. 1988;
Houlihan
et
al. 1988;
Jarvi
1990).
All
adults
that sur-
vived
the
breeding
season were
sacrificed on 13 Feb-
ruary
1991 and the
above measures
recorded.
Most
spent
adults
in
the wild
would have
been
expected
to
have
left the
river and returned
to the sea by
mid-
February
(Jonsson
et al. 1991).
REPRODUCTIVE SUCCESS
Eggs
were
retrieved
from all nests that survived
the
spawning
period
(i.e. were
not
destroyed
by
females
digging
subsequent
nests).
A minimum
of
170
degree
days since oviposition
(days
x temperature
above
0C) were used
to
assess when
eggs
in each
nest could
be handled
(i.e.
'eyed';
Leitritz
1959),
and
nests were
excavated
between
21 February
and 6 March 1991.
After draining
the arenas,
the depth
of each
nest
from
the
gravel
surface
to
the centre
of the
egg pocket
was
measured.
Then an
open-ended
rectangular
steel box
was inserted
into
the
gravel,
the
egg
pocket
removed
and the
eggs
separated
from the
gravel
by
flushing
with
water.
Both live and dead eggs
were
counted,
and a sample
of each was
preserved
separately
in
4%
buffered
formalin.
Three months
after
preservation,
wet
weights
of 10 live
and 10 dead eggs
from each
sample
were measured
individually
to assess
egg
size.
The
gravel
surrounding
the
egg
pocket
was
analysed
following
Lotspeich
& Everest
(1981).
Spawning
records,
which
assigned
females to
nests,
were
cross-checked
using
the
egg
size
data.
Egg
size is a
reliable
means of
confirming
maternity
because
intra-
individual
variation
in
egg
size
is small
relative to
that
among
females
(coefficient
of variation
(CV) among
females
= 11
0%, within
female mean CV = 3
4 +
1
2% SD, n
= 41
females;
I.A. Fleming,
unpublished
data).
For each
female,
absolute
reproductive
success
was
accurately
measured
as the total number
of live
embryos
and
spawning
success
was defined
as
absolute
success
divided
by fecundity.
Initial
fecundity
was
esti-
? 1996
British
Ecological
Society,
Journal
of Applied
Ecology,
33,
893-905
897
IA. Fleming et al
mated for each female
from regressions
of unspawned
females (Imsa wild: fecundity
= 3 41 x [total
weight]0896;
R2
= 0
79, n
= 12, P < 0
001; Sunnd-
alsora
farmed:
fecundity
= 0 12 x [total
weight] 1341;
R2 = 0
98, n
= 5,
P < 0
001;
B.
Jonsson
& I.A. Flem-
ing,
unpublished data).
The reproductive
success of each male was cal-
culated
from an
estimate of his
proportional paternity
times
the number of live embryos in each nest he
spawned at or was suspected to have spawned at.
In the latter
case, the
spawning
was missed but
the
spawning aggregation
immediately
prior to the
spawning
had been observed.
Sixty-two
percentage
of
all spawnings
were observed
directly (109 of the 175
nests recovered with live
embryos). Male
reproductive
success was thus
analysed both including and
exclud-
ing
indirect
observations
(38%; see
Results).
If
more
than one male
partook
(13%, 14 of
109 direct obser-
vations),
or was suspected
to
have
possibly partaken
in
a spawning (74%, 49 of
66 indirect
observations),
a male's
paternity
was estimated from his rank
order
of nest
entry
at spawning,
or of
dominance
in the
spawning aggregation,
respectively. Each successive
male
entering,
or
ranked in the
dominance
hierarchy,
was assumed to obtain
50% the success of the male
ahead
of
him. This allocation was
made
on the
basis
of results
obtained in studies of chum salmon
(Oncorhynchus
keta
W.; Schroder
1982),
the
only
sal-
mon
species
for which such
data
appear
to be avail-
able. The overall results
were not sensitive to this
assumption,
because an alternative
assumption
of
equal paternity produced similar estimates
of
overall
male reproductive
success,
the
two estimates
being
highly correlated (R2
= 0
998, n
= 48, slope
/3
= 0971 + 0006 SE).
STATISTICAL ANALYSES
Heart and
body mass,
all
frequency data,
and all mea-
sures of male
reproductive
success were
logarithmic
transformed to meet
the
assumptions
of
analysis
of
variance.
Similarly,
all
percentage
data
were
arcsine,
square-root
transformed. Data that did not meet
requirements
for
parametric
analysis
were
analysed by
nonparametric
Mann-Whitney
U-tests or
Chi-square
tests. Individual fish were considered
replicates
within
arenas. Arena effects
were examined
by
comparing
replicate
arenas
2
and 3.
Adjustment
for
multiple
com-
parisons
was carried out
using sequential
Bonferroni
tests
(Rice 1989).
Results
BEHAVIOUR
When held in mixed
groups, wild and
farmed
females
did not differ
in either aggressive or
submissive
behav-
iour (Fig.
2). Possible differences
in the aggressive
behaviours
combat
(i.e. chasing
and fighting;
ANOVA
P = 0 933) and display (P = 0 339) were also exam-
ined, but
none were
found.
Aggression was directed
equally at
farmed and
wild female
opponents
(paired
t-test,
farmed:
tI = 1
01, P = 0 334; wild: t,1
= 1
34,
P = 0.208). Female aggression towards males,
however,
was directed
more often
at wild than
farmed
individuals (farmed females: t1,
= 431, P = 0001;
wild females: t,1
= 5 09, P < 0 001), reflecting
the
higher
frequency
of courting by wild than
farmed
males. Farmed females
were less active
than
wild
females in cruising
and digging,
and were
courted
less
often,
presumably
reflecting
their
lower levels of
digging
and associated
nesting
activity (Fig. 2). Only
the
frequency
of female
submissive behaviour differed
Females Males
03 P
=0-328 60 P <0-001
?~ 02 ; , 40
0 0
0-2 P=0-335 0*3
-<000
co c0 0o 0-
0.0 ~~~02
01 liE
co 0.2~~~~~ 0*
E 0
0
.10
0-0 0-0
0-15 P= 0-0106* 10 - Pc 0-696
CY 010
2 ) 02 -
C)0005 C)001
0 0.0
0F 6
.
P=0Feu yo 0a 2
0.2
0 0*0
0.6 p=.1*1.0 P<00
0*4
2 3 2 3
Arena Arena
Fig.
2. Frequency
of
female and male behaviours
in
farmed
and wild salmon
in Arenas 2 and 3. Data are means and
standard deviations
across arenas
adjusted
for the covariate
body size where
significant.
Probability
values are from
analyses
of variance or
covariance
testing for
behavioural
differences
between farmed
and wild fish. Significant
differ-
ences adjusted for
multiple
comparison by
sequential
Bon-
ferroni
tests are indicated
(*). Hatched
bars = farmed;
solid
bars = wild.
? 1996 British
Ecological
Society,
Journal
of Applied
Ecology,
33, 893-905
898
Reproductive
success
offarmed
and
wild salmon
between
Arenas
2 and 3 (P = 0.008)
and there
were
no significant
fish type-by-arena
interactions
(P > 0 05).
In contrast
to
farmed
females,
farmed
males
were
less
aggressive,
exhibiting
less combat
(P < 0
001)
and
display behaviour (P = 0.004) than wild males
(Fig.
2). Male aggression
was directed
more
often
at
wild than farmed
males (paired t-test,
farmed:
tj
I = 2 49, P = 0 030; wild: t1
I = 5 69, P < 0 001)
and
wild
males
exhibited
more
submissive
behaviour.
Farmed
males
exhibited
no significant
directionality
in their aggression towards females (t1
I = 137,
P = 0 199),
whereas
wild
males
were
more
aggressive
to wild than farmed
females (t1
I = 3 81, P = 0 003).
There
was no difference
between
wild
and farmed
males
in
cruising
behaviour,
which
tends
to
be
associ-
ated with searching
for mates. Farmed males,
however,
had difficulty
acquiring
access to mates,
showing
less
quivering
and courting
behaviour
than
wild
males
(Fig.
2). There were
no significant
arena
effects
or
fish type-by-arena
interactions
(P > 0 05).
There
were
only
minor
differences
in behaviours
displayed
by
wild
females
in mixed
groups
compared
to
those together
only
with
other
wild
fish (Table 2).
Similarly,
wild
males
did not alter
their
aggressive
and
submissive
behaviours
in the
presence
of
farmed
males,
but
did
increase
cruising,
courting
and
quiver-
ing
behaviour
when
held
in mixed
groups,
apparently
reflecting
their
superior
competitive
ability.
Farmed females
exhibited
a tendency
towards
reduced
aggressiveness,
particularly
combat
behav-
iour (P = 0
008), displayed
less digging
behaviour,
were
courted
less
often
and cruised
more frequently
when
wild
salmon
were
absent
(Table
2).
This may
be
a reflection
of
the
reduced
breeding
behaviour
dis-
played
by
farmed
relative
to
wild
males.
Behaviour
of
farmed
males
in the
absence
of
wild
fish
differed
little
from
that
in their presence
(Table 2), though
there
was
a non-significant
tendency
for
farmed
males
to
be
more
submissive
and
court
less
in
the
presence
of
wild
males.
HEART MASS, WOUNDING AND MORTALITY
Heart
mass
relative
to
somatic
mass
was smaller
for
farmed
than
for
wild
females
(ANCOVA F1 42 = 18
68,
P < 0 001;
allometric
coefficientj3
= 0O42,
P < 0
001),
but
did not differ
significantly
between
farmed
and
wild males (ANCOVA F1 41 = 2
03, P = 0 162; allo-
metric
coefficient
/3
= 037, P < 0 001).
In mixed
groups,
farmed
and
wild
females
incurred
similar degrees of wounding (Mann-Whitney
U
= 840, P = 0 148), but farmed
females
suffered
higher
mortality
(farmed:
42% died
before
13 Feb-
ruary
1991;
wild:
0%; x2 = 6 32,
d.f.
= 1,
P = 0
012).
Body
condition
of
farmed
males
appeared
to
deterior-
ate rapidly
as they
incurred
more
wounding
(Mann-
Whitney
U = 128
0, P < 0.001) and suffered
higher
mortality
than
wild
males (farmed:
100%;
wild:
42%;
X2= 9
88,
d.f.
= 1,
P = 0.002).
Wounding
rates
of wild salmon
were
unaffected
by
the
presence
of
farmed
salmon
(females:
Mann-
Whitney U = 72
0, P = 1
0; males: U = 83
5,
P = 0 38
1),
as was
mortality
(females:
0% irrespective
of
treatment;
males:
x2 = 075, d.f.
= 1,
P = 0
609).
Similarly,
the deterioration
of body condition
in
farmed
salmon, particularly
males,
was
unaffected
by
the
presence
of wild salmon,
as neither
wounding
(females:
Mann-Whitney
U = 78 0,
P = 0 623;
males:
U
= 79 0, P = 0.569) nor mortality (females:
Table
2. Comparison
of behaviours
of
wild and
farmed
salmon
when
held
as separate
groups
(Arenas
1
and
4,
respectively)
with
that
when they
were
held
in
mixed groups
(Arenas
2
and
3).
F statistics
and
probability
values
(P) are
from
analyses
of
variance
or covariance
(> indicates
variable
greater
in
separate
than
in
mixed
groups;
< indicates
the
vice
versa;
= indicates
no significant
effect
at P < 0
05; and [*] indicates
significant
difference
adjusted
for
multiple
comparison
by sequential
Bonferroni
tests)
Wild
fish Farmed
fish
Separate
vs. Separate
vs.
mixed Covariate mixed Covariate
body
weight body
weight
Variable Sex F P P F P P
Aggression F = 1-39 0-251a <0.001* < 5 09 0
035 0.003*
M = 4-21 0
050 0.007* = 1-31 0
265 -
Submission F < 5
30 0-031 - 1-60 0.220a -
M = 0-71 0 409 0.003* < 5 30 0-032 0-084
Cruising F = 007 0
790 <0.001* < 4 93 0
037 -
M < 9-68 0.005* 0-017 = 1-46 0-241 -
Digging F = 404 0 054 <0.001* < 20-97 <0.001* 0.001*
Courted F = 077 0 390 <0.001* < 14-42 0.001* 0-061
Courting M = 4-41 0
045 0-061 < 7 55 0
012 <0.001*
Quivering M < 7-42 0-012* 0-031 = 2-24 0-149 -
a Arenas
2 and
3 differed
significantly
(P < 0.05).
? 1996
British
Ecological
Society,
Journal
of Applied
Ecology,
33,
893-905
899
IA. Fleming
et al.
X2
= 0 75, d.f.
= 1, P = 0061; males: x2 = 0.0,
d.f.
= 1,
P = 1
0) differed
between
treatments.
FEMALE NESTING AND REPRODUCTIVE
SUCCESS
In mixed
groups
there
was
a non-significant
tendency
for
farmed
females
to
begin
breeding
slightly
earlier
than wild
females
(Fig.
3).
Farmed
females
made
fewer
nests
and
there was
a non-significant
trend
for
farmed
females
to
breed
for
a
slightly
shorter
period
than
wild
females.
The majority
of females
made
all
their
nests
within
a single
redd,
and
this
did
not
differ
between
farmed (58%) and wild females (67%) (X2 = 018,
d.f.
= 1, P = 0 675). Although farmed females con-
structed
fewer
nests
than
wild
females,
the
total
num-
ber of eggs recovered
per nest
did not differ
sig-
nificantly
(Fig.
3). Farmed females
seemed to be
poorer
at nest covering,
as they
dug
less
frequently
during
the
first
5
min
following
oviposition
and
took
_ 30 P= 0-020 - 10 P 0-022
12 p = -01
600 P =Q0-044
R 8 4Q
10
6u D 6
4~~~~~~~~~
2n
Fig.
3. Nesting
characteristics
of farmed
and
wild
female
sal-
mon
in Arenas
2 and
3.
Data are
means
and
standard
devi-
ations
across
arenas
adjusted
for
the
covariate
body size
where
significant.
Probability
values
are
derived
from
analy-
ses
of variance
or covariance.
Significant
differences
adjusted
for
multiple
comparison
by
sequential
Bonferroni
tests
are
indicated
(*). Hatched
bars
= farmed;
solid
bars
= wild.
longer
to cover
their
eggs.
There
were, however,
no
significant
differences
in the
depth
of
nests or nest
gravel
quality.
Farmed females retained significantly
greater
weight
of
unspawned
eggs
than
wild
females
(Fig.
4).
Nest
destruction,
resulting
from
reuse
of
the
nest
site
by
other
females,
also
occurred
more
often
to
farmed
than
wild
females
(farmed:
11
%; wild:
1 %; X2 = 8 95,
d.f. = 1,
P = 0.003). Survival of eggs in nests,
which
included
the
effects
of
poor
fertilization
and
mortality,
was
significantly
lower
for farmed
than
wild
females.
As
a consequence
of greater
egg
retention
and
poorer
survival
of
deposited
eggs,
the farmed
females
were
less
than
one-third
as successful
as wild
females
at
having
their
eggs
survive
through
incubation
(Fig.
4).
This resulted
in a considerably
lower
reproductive
success
for
farmed
females
(ANCOVA
F1,19
= 50 72,
P < 0
001; Fig.
5). There were
no significant
differ-
100 P
=0-003*
3, 80
c 60
0)
80 p
<0.001
~60
U)
40
> 40
U)
0
m- 20
0
1~60 P
<0.001*
50
a) Aren
O 40
C:)L 30
U)
O
20
CL0
0.
-a)
2 3
Arena
Fig.
4. Breeding
success
of farmed
and wild
female
salmon
in
Arenas
2
and 3.
Data are
means
and
standard
deviations
across
experiments.
Probability
values
are derived
from
a
Mann-Whitney U-test for analysis of eggs retained
(U = 114,
N = 24) and from
analyses
of variance
for
the
other
variables.
Significant
differences
adjusted
for
multiple
comparison
by sequential
Bonferroni
tests
are indicated
(*).
Hatched
bars
= farmed;
solid
bars
= wild.
? 1996
British
Ecological
Society,
Journal
of
Applied
Ecology,
33,
893-905
900
Reproductive
success
offarmed
and wild salmon
5
2
S^ 4 r =0813
o P<0.001
co x
3
co) .0
aE
0 2
E
D:- 1 */*i
o~~~~~~~D
**
0 1 2 3 4 5 6
Body weight
(kg)
Fig.
5.
Reproductive
success of
farmed (Arena 2
= 1,
Arena
3
= 0) and wild female
salmon
(Arena
2 =, Arena
3
= 0) in
competition. Solid line is
the relationship of repro-
ductive
success vs. body size for wild females
(y
= 0 81
1
*[weight, g]
- 743). The
relationship was non-
significant
for farmed
females (R2
= 0 123,
P = 0 262).
ences
between
Arenas
2
and 3,
and no significant
fish
type-by-arena
interactions (P > 0 05).
There was little
difference
in
nesting behaviour
of
wild females
irrespective
of
whether or not
farmed
salmon were
present (Table 3). There
were, however,
some
differences in
nest
characteristics,
as nests
were
found
at
greater
depths
and
there
was a
non-significant
trend for
females
to
construct
fewer
nests
when
farmed
salmon
were absent.
Egg
retention
and
egg
survival
in
nests
were
unaffected
by the
presence of
farmed sal-
mon
(Table 3). Nest
destruction,
however,
increased
when
farmed females
were absent
(i.e. when
they were
replaced by an equal number
of wild females;
% = 18
89,
d.f.
= 1,
P < 0.001)
and thus wild
females
were less
reproductively
successful
(Table 3).
Farmed females
showed a non-significant
tendency
to
delay
breeding in
the
absence
of
wild salmon
(Table
3). Duration of
breeding,
and nest
depth and gravel
characteristics of
farmed females
were not
influenced
by the
presence
of wild fish.
Farmed females
made
fewer nests
in
the
absence of
wild
salmon,
but more
eggs were
deposited
in each nest
(Table 3). Repro-
ductive success of
farmed females
was dramatically
reduced
in the
absence of wild
fish.
Egg retention
increased
and fewer of the
eggs
spawned
survived.
In
the absence
of wild
males,
only 10% of the
nests
constructed
by
farmed
females
contained
eggs
with
live embryos,
while 98% of nests
contained live
embryos when wild males
were
present
(X2
= 74.93,
d.f.
= 1,
P < 0-001).
This
indicates
that
many
of the
ovipositions
made
by
farmed females
remained unfer-
tilized when
wild males
were absent. As a result,
the
reproductive success of
farmed females in
the
absence
Table 3. Comparison
of
nesting
characteristics and
reproductive
success
of wild and
farmed females
when held as separate
groups
(Arenas 1 and 4, respectively)
with that
when
they were
held in
mixed
groups
(Arenas
2 and 3). F statistics and
probability
values
(P) are
from
analyses of
variance or
covariance
(> indicates
variable
greater in
separate than
in mixed
groups;
< indicates the
vice
versa;
= indicates no
significant
effect at
P < 0
05;
and
[*]
indicates
significant
difference
adjusted
for
multiple
comparison by
sequential
Bonferroni
tests)
Wild
females Farmed
females
Separate
vs. Separate
vs.
mixed Covariate mixed Covariate
body weight body
weight
Variable F P P F P P
Nesting
characteristics
Onset of
breeding = 048 0
495 <0.001* 6 96 0
017 <0.003*
Duration
of
breeding = 002 0 874 <0.001* = 083 0
373
Numberofnests < 789 0011 <0.001* < 1701 <0.001* 0010
Eggs per nesta = 0
05 0 823 <0 001 > 11
51 0 003* <0 001*
Covering
boutsb = 147 0
243 - -
Time to cover
eggsb = 004 0 831 -
Nest depth > 13 64 0
001* 0
027 = 033 0 572 -
Gravel
quality = 0
05 0 822c - - 035 0 562 -
Reproductive
success
Eggs retained = 001 0
981 - > 19
41 <0 001* -
Survival of
deposited
eggs = 1
53 0
230 - < 92
51 <0.001* -
Spawning
success < 15 20 0.001* 0 001* < 64 15 <0.001* -
(eggs
deposited
+
survived/fecundity)
Reproductive
success < 7
71 0
011 <0.001* < 63
48 <0.001* -
a No nests
recovered for one
wild female in Arena
1
and three
farmed females in
Arena
4.
bNo nesting
covering
data for
four
wild
females in Arena
1
and
one
wild
female
in
each of
Arenas
2 and
3.
Nest
covering
was not
analysed
for farmed
females because
insufficient data
existed from Arena 4.
cArenas
2
and
3 differed
significantly
(P < 0-05)
for wild females.
I 1996 British
Ecological
Society,
Journal
of
Applied
Ecology,
33, 893-905
901
I.A. Fleming
et al.
of wild
fish was
less than
one-tenth
of
what
it was
in
the presence
of
wild
fish.
MALE BREEDING AND REPRODUCTIVE
SUCCESS
When
held
together,
farmed
and wild males
began
breeding
at about
the same
time,
but
the farmed
males
had a short
breeding
duration
(Fig.
6). This was a
consequence
of
the high
mortality
of farmed
males
during
the breeding
season. Farmed
males
took
part
in few spawnings
and
had lower
reproductive
success
than
wild
males
(ANCOVA, direct
+ indirect
spawning
observations: F1,19
= 8 67, P = 0
008; Fig. 7;
ANCOVA, direct observations
only: F1,19
= 41
52,
P < 0
001). There were no significant
differences
between
Arenas
2 and
3,
and
no
significant
fish
type-
by-arena
interactions
(P > 0
05).
Neither
the onset
nor the
duration
of spawning
activity by
wild
males was influenced
by
farmed
sal-
6
P= 0.059
5
m3 25
4
0
>
W 30
2 oo
qo'20*
10
0
30 P <0P 0.001
25 3
co20
- Arena
.0'
E 0
40
2
30
Fig. 6. Breeding
activity
of farmed
and
wild male
salmon
in
Arenas
2 and 3. Data are
means
and standard
deviations
across
experiments
adjusted
for the
covariate
body
size
where
significant.
Probability
values
are derived
from
analyses
of
variance
and covariance.
Significant
differences
adjusted
for
multiple
comparison
by sequential
Bonferroni
tests
are
indi-
cated
(*). Hatched
bars
= farmed;
solid
bars
= wild.
12
0
o U
x
W 10
N
.1
_
8
~ 0
0~~~~
E 2
. a) / r =0-451
a) P = 0-017
*"
O" 4
CL W ~ 0
rrE
en
a) 1 2 3 4 5
Body
weight
(kg)
Fig. 7. Reproductive
success
of farmed
(Arena
2 = D,
Arena
3 = 0) and
wild male
salmon
(Arena
2
= , Arena
3
= *)
in competition.
Solid
line
is the relationship
between
repro-
ductive success and body size for wild males
(y
= 4 96 x [weight]3
173),
which
is also significant
by
Spear-
man rank
correlation
(Rs
= 0
615,
P = 0 033).
The
relation-
ship
was non-significant
for farmed
males
(R2 = 0 180,
P =0 170).
mon (Table 4). Wild males did, however,
obtain
greater
numbers
of spawnings
and
had higher
repro-
ductive
success
with farmed
fish
present,
than
when
held in groups
comprised
solely
of wild
fish.
Farmed
males
were
not influenced
by
wild
males
in
their breeding
and reproductive
success
(Table 4).
Even
in the
absence
of
wild
males,
farmed
males
dis-
played inappropriate
reproductive
behaviours
and
only
in two
of six
spawnings
did
the
courting
farmed
male(s) enter
the
nest
and release
sperm
when
the
female
oviposited.
Discussion
Farmed
fish were
competitively
and reproductively
inferior
to
wild fish
(Figs
2-7). Body
size was a key
determinant
of
reproductive
success
in
wild,
but
not
farmed salmon
(Figs
5 and
7).
The farmed
fish,
males
and females
combined,
achieved
only 11-19% the
reproductive
success
of the
wild
fish when
in
compe-
tition.
Intergroup
competition,
however,
did not
reduce
the reproductive
ability
or success
of
either
the
farmed
or
wild
salmon
(Tables
2-4).
COMPETITIVE AND BREEDING BEHAVIOUR
While
there
were no evident
differences
in
expression
of aggressive
behaviour
of farmed
and wild
female
Atlantic
salmon,
farmed
males
were
distinctly
less
aggressive
than
wild
males.
Reduced
aggressiveness
of
cultured
relative
to wild
adult
males
has also been
observed
in sea-ranched
coho salmon
(Fleming
&
Gross 1992,
1993).
This response
to domestication
? 1996
British
Ecological
Society,
Journal
of
Applied
Ecology,
33,
893-905
902
Reproductive
success offarmed
and wild salmon
Table
4. Comparison
of breeding
of
wild and farmed
males
when held
as separate
groups
(Arenas
1 and 4, respectively)
with
that
when
they
were
held
in
mixed groups
(Arenas
2 and
3).
Fstatistics
and
probability
values
(P) are
from
analyses
of variance
or covariance
(> indicates
variable
is greater
in separate
than in mixed
groups;
< indicates
the vice versa;
= indicates
no
significant
effect
at P < 0 05;
and [*]
indicates significant
difference
adjusted
for multiple comparison
by sequential
Bonferroni
tests)
Wild males Farmed males
Separate
vs. Separate vs.
mixed Covariate mixed Covariate
body weight body weight
Variable F P P F P P
Onset
of
breeding activitya = 316 0 090 - 073 0 405 -
Duration
of breeding
activity = 407 0
056 - 085 0 366 -
Number
of
spawnings < 8 53 0.008* 0038 = 097 0 336 -
Estimated
reproductive
success
Direct spawning
observations
< 7
48 0.012* 0
225 = 100 0
328 -
Direct+indirect
observations
< 8 78 0
007* 0 148 = 010 0
754 -
a Two farmed
males
in Arena
2
and
one
in Arena 3 were
never observed expressing
breeding behaviour.
may involve both genetic and environmentally
induced
changes
as a consequence
of culture.
Some
species
have
been
shown
to rapidly
adapt
genetically
to captivity
(Frankham
& Loebel 1992) and behav-
ioural
traits
may be
among
the first
traits to
respond
(Kohane & Parsons
1988).
In salmon
culture,
where
matings
are
determined
artificially,
adult
aggression
would
afford no apparent
reproductive
advantage
and
directed
selection
for
rapid
growth
may
result
in a
correlated response for reduced aggressiveness
(reviewed
in
Ruzzante
1994).
Thus,
a combination
of
artificial
and domestication
selection
may,
in part,
contribute
to differences
in aggressiveness
between
farmed
and wild males.
Reduced
aggressiveness
is also
likely to
be environ-
mentally
induced
and may
reflect
the
rapid
deterior-
ation in body condition
of
the farmed
fish, particularly
the
males,
which
incurred high
mortality,
wounding
and fungus
infection. Sexual maturation
in com-
bination with
chronic stress
typically
associated with
aquaculture
adversely
affect
the
condition
of salmon-
ids,
particularly
males,
increasing
susceptibility
to
dis-
ease and mortality
(Pickering
1993).
Observations
of
escaped farmed
salmon
having
higher
incidence
of
local scale loss
than wild
salmon
in
nature
(Webb
et
al. 1991)
further suggests
that the detrimental
effects
of
culture
on body
condition
persist
till
spawning.
Similar observations
have
been
made
of
sea-ranched
salmon,
which
incur more
extensive
wounding
during
the
breeding
season
than wild salmon (Jonsson
et al.
1990;
Fleming
& Gross
1993).
The increased wound-
ing
of farmed males
may
also be due
to their
apparent
inability,
or
unwillingness,
to
avoid
physical
contact
when attacked.
This cannot,
however,
be the full
explanation as wounding in farmed
males was
unaffected
by
levels
of
aggression
directed
at them,
which was
higher
in
the
presence
than absence
of
wild
salmon.
Differences
in
aggressiveness
between
farmed and
wild
males, but
not
females
may
reflect
differences
between
the sexes in intensity
of competition
for
breeding
resources
(cf.
Fleming
& Gross
1994;
Quinn
& Foote
1994).
Male
Atlantic salmon
compete
overtly
for access
to spawning
females,
while
females
appear
to
compete
more
subtly
for breeding
territories
(also
Jones 1959;
Webb
& Hawkins 1989).
This
pattern
is
similar
to that
observed
in adult coho
salmon,
where
sea-ranched
and
wild
males,
but
not females,
differed
in aggressiveness
(Fleming
& Gross
1993).
Competitive
differences
between
farmed
and wild
females were
subtle.
Reduced breeding
behaviour,
construction
of fewer nests
and
retention
of more eggs
unspawned
by
farmed
than wild females
may
have
resulted
from
competitive
inferiority.
Altered
behav-
iour
patterns
caused by domestication
and poorer
physical
condition
(relatively
small hearts;
also Gra-
ham
& Farrell 1992)
of farmed
than wild
females
may
have
also contributed
to
the differences.
Similarly,
a
combination
of these
factors and reduced
caudal fin
size,
the primary
fin used in
nest
construction,
due to
sea-pen
culturing
(Fleming
et al.,
in press)
may
explain
the
inefficient
covering
of nests
by
farmed
relative
to
wild females. Inefficient
nest covering probably
resulted
in
greater
egg
loss,
as fewer
eggs
of
farmed
females were
recovered
per nest
even though
they
partitioned
their
initial
fecundity
among
fewer nests
than wild
females.
Similarly,
inefficient
nest
covering
may explain
the observation
of
Lura,
Barlaup
& Se-
grov
(1993) that the
nests
of
a farmed
female
in the
River
Vosso,
Norway,
were
more variable
in
volume
and
contained
fewer
eggs
than
those of wild
females.
REPRODUCTIVE SUCCESS
There were clear
differences between
farmed and
wild
females
in
reproductive
success
with farmed
females
having
only
between 20 and 40% the
reproductive
success
of
wild
females. Several
factors contributed
to
? 1996 British
Ecological
Society,
Journal of
Applied
Ecology,
33,
893-905
903
IA. Fleming
et al.
these differences.
The retention
of more eggs
unspawned
by
farmed
than
wild
females
in the
com-
petitive
environment is a pattern
that has also been
observed
between
sea-ranched and wild salmonids
(Jonsson
et al. 1990;
Fleming
& Gross
1993).
Farmed
females
also
incurred
higher
levels of
nest
destruction
than wild females.
Nest destruction
through
nest
superimposition
may be
an
important cause
of
female
egg
mortality
in
Atlantic
salmon,
even in
seemingly
low
density
populations
(Webb
& Hawkins
1989;
Lura
& Sagrov 1991). Reduced
egg
survival
in nests of
farmed
relative
to wild females
might
have been
related to poorer
nest
covering
and/or
poorer
egg
quality (Srivastava & Brown
1991).
Farmed
males, more
so than
farmed
females,
were
reproductively
inferior
to wild fish.
Lack of
any
appar-
ent
differences in
reproductive success in
the
presence
and
absence of
wild
salmon
suggests
this was
not
due
to
competitive
inferiority
alone.
Rather,
inappropriate
reproductive behaviour and poor condition are
important
explanations.
Farmed males
when
courting
females often
failed to enter
nests to
fertilize the
eggs
during female oviposition. Inappropriate
repro-
ductive behaviour
of
farmed males
may
explain
Lura
& Segrov's (1991) observation
that
eggs
in several
nests
spawned
by
escaped
farmed female Atlantic
sal-
mon
were
unfertilized.
Farmed males in
our experi-
ments
were
estimated to have
attained
1-3% of the
reproductive
success
of
wild males.
Such loss of
breed-
ing
fitness
is known to occur
in other
species, even
where strict
breeding
programmes
to maintain
genetic
variability
are implemented,
due to rapid domes-
tication
(Loebel et al. 1992;
see also Lyles
& May
1987).
The
process
of
environmental and
genetic
adap-
tation, including behavioural and physiological
responses
to
captivity,
is
likely
accelerated
in
farmed
salmon,
where
breeding
is
determined
artificially.
It is possible that interpopulational
differences
unrelated to
artificial
culture
may
explain
the
present
results;
however,
this
seems
unlikely
given
the
mag-
nitude
of
differences
observed.
Our
experiments
were
designed
to be representative
of the
intentional or
unintentional
presence
of
farmed
salmon
in
Norwegian
rivers and thus farmed salmon
from the
principal
breeding
programme
of
farmed
salmon in Norway
were used.
Furthermore,
these
results
agree
with other
evidence that
suggests
captive
breeding
and
artificial
culture reduce
natural
reproductive
ability
of fish
(Jonsson
et
al. 1990;
Leider et
al. 1990;
Fleming
&
Gross
1993)
and
other
organisms
(Cade 1988;
Loebel
et
al. 1992).
IMPLICATIONS
Large numbers
of farmed
Atlantic salmon
escape;
minimum
estimates
suggest
they
compose
25-48%
of salmon
on the
feeding
grounds
in the
north-east
Atlantic
Ocean (Hansen, Jacobsen
& Lund 1993).
Some
of
these fish
enter
onto breeding
grounds
of
wild
populations
where it
is not
uncommon
for
them
to
outnumber
wild
spawners
by as much
as 3:1
(Lund,
Hansen
& 0kland 1994;
Lura & 0kland 1994).
Our
results
suggest
that farmed
salmonids,
artificially
reared to
maturity, will
have an
inferior
reproductive
ability
relative
to
their
wild
counterparts.
The
extent
of this
inferiority
is
likely to
be affected by
the
pro-
portion of a fish's
life, as well as the
number
of
gen-
erations
in
culture
(Fleming
et al. 1994).
For
instance,
the
reproductive
inferiority
shown
by
sea-ranched
sal-
monids
relative
to
their
wild
counterparts
(Leider
et
al. 1990;
Fleming
& Gross
1993; Jonsson & Fleming
1993)
is less
than that
shown
by
farmed salmon in
this
study.
The extent
of
residency
in
the
natural
environ-
ment
following
escape is thus
likely to be an
important
determinant of reproductive
ability once on the
spawning
grounds.
Even when reared
to maturity,
hybridization is
likely to occur
between
escaped
farmed
females
and
wild males
as found
in
the
present
study.
This
would
result
in
sex-biased
gene flow
between
cultured
and
wild fish
(also
Fleming
& Gross
1993). Given the
large
numbers
of farmed
salmon
escaping
and entering
spawning
grounds
of
wild
populations the
potential
for
gene
flow is
great.
Long-term
effects of
such
gene
flow are
unclear,
as little is known
about success
of
offspring
from
such
spawnings,
although
in
most
cases
we would expect
it to be lower than that of wild
offspring due
to
lack of local
adaptation.
Evidence of
lower fitness in
cultured/foreign and
hybrid
offspring
relative
to
that of
native
offspring
supports
this
con-
tention
(e.g. Reisenbichler & McIntyre
1977;
Lach-
ance & Magnan
1990;
Leider et
al. 1990;
Philipp
1991;
Reisenbichler cited
by
Emlen
1991;
Skaala 1992).
It
might
be
speculated
that the
productivity
of
wild
populations
could
be depressed
by
intrusions of
cul-
tured salmon that
resulted
in
ecological
interference,
including
mate and
territorial
competition. Our results
suggest
that the
effects
resulting
from
intrusions
by
cultured salmonids
will be sex
biased,
being generated
primarily
by
cultured
females,
and
such effects
would
be very
much
dependent
on timing
of
spawning.
If
farmed salmon
spawn
prior
to,
or
at the
same time as
wild
salmon,
as occurs in
parts
of
south-western Nor-
way
(also
Lura
& Swgrov
1993),
ecological
disruption
to
wild
populations
during
breeding
may
be
minimal.
Later
spawning
by
farmed
females
would,
however,
probably
result in
destruction of nests of
wild
females
(Lura & Swgrov
1991;
Webb et
al. 1991).
Thus,
the
spawning
time of
wild
relative
to
farmed
salmon
would
be an important
determinant of
susceptibility
of
wild
populations
to
ecological
interference.
Our
results also have
application
to
captive
breed-
ing
programmes
for
conservation and
reintroduction
of
species,
e.g.
Sacramento
River
winter-run chinook
salmon
Oncorhynchus
tshawytscha
(Hedrick
& Hedge-
cock 1994). To increase the success of such
programmes,
detrimental
effects
of captive-rearing
on
an organism's
phenotype
and
genotype,
including
its
?) 1996 British
Ecological
Society,
Journal
of
Applied
Ecology,
33, 893-905
904
Reproductive
success
offarmed
and
wild salmon
behavioural,
morphological
and
physiological
traits,
must
be minimized.
This may
be accomplished
by
keeping
the
number
of
generations
a species
needs
to
be in captivity
low and exposing
it to naturalistic
experiences
and
selection
during
this
time.
Acknowledgements
We thank
A. Braa, L. Korsnes,
J.
H. Pettersen,
I.
Uglem
and
the
staff
of the
NINA Research
Station
at
Ims
for their
assistance
with
the
experiments,
and
T.
Forseth,
K. Hindar
and
T. Jarvi
for
their
comments
on the
manuscript.
Financial
support
was provided
by grants
from
the
Norwegian
Fisheries
Research
Council
to Jonsson
and Fleming,
and the
National
Science
and
Engineering
Research
Council
of
Canada
to
M.R. Gross.
References
Allendorf,
F.W. (1983)
Conservation
biology
of fishes.
Con-
servation Biology,
2,
145-148.
Cade,
T.J.
(1988)
Using
science
and
technology
to
re-estab-
lish
species
lost
in nature. Biodiversity
(ed.
E. 0. Wilson),
pp.
279-288.
National
Academy
Press,
Washington.
Chapman,
D.W. (1988)
Critical
review
of
variables
used
to
define
effects
of
fines
in redds
of large
salmonids.
Trans-
actions
of
the
American
Fisheries
Society,
117,
1-21.
Crisp,
D.T. & Carling,
P.A. (1989)
Observations
on siting,
dimensions
and structure
of
salmonid
redds.
Journal
of
Fish
Biology,
34,
119-134.
Cross, T.F. & Challanain,
D.N. (1991) Genetic
charac-
terisation
of Atlantic
salmon
(Salmo
salar)
lines
farmed
in
Ireland.
Aquaculture,
98,
209-216.
Crozier, W.W. (1993) Evidence of genetic
interaction
between
escaped
farmed
salmon
and wild
Atlantic
salmon
(Salmo salar L.) in a Northern
Irish river.
Aquaculture,
113,
19-29.
Emlen, J.M.
(1991)
Heterosis
and
outbreeding
depression:
a
multi-locus
model and an application to salmon
production.
Fisheries
Research,
12,
187-212.
Farrell,
A.P., Hammons,
A.M., Graham,
M.S. & Tibbits,
G.F. (1988)
Cardiac
growth
in rainbow
trout,
Salmo
gaird-
neri.
Canadian
Journal
of
Zoology,
66,
2368-2373.
Fleming, I.A. & Gross,
M.R. (1992)
Reproductive
behavior
of hatchery
and
wild
coho
salmon
(Oncorhynchus
kisutch):
does
it
differ?
Aquaculture,
103,
101-121.
Fleming,
I.A. & Gross,
M.R. (1993) Breeding
success
of
hatchery
and wild
coho
salmon
(Oncorhynchus
kisutch)
in
competition.
Ecological
Applications,
3,
230-245.
Fleming,
I.A. & Gross,
M.R. (1994) Breeding
competition
in a
Pacific
salmon
(coho:
Oncorhynchus
kisutch):
measures
of natural
and
sexual
selection.
Evolution,
48,
637-657.
Fleming,
I.A.,
Jonsson,
B. & Gross,
M.R. (1994)
Phenotypic
divergence
of sea-ranched,
farmed
and wild
salmon.
Can-
adian
Journal of
Fisheries
and
Aquatic
Sciences,
51,
2808-
2824.
Frankham,
R. & Loebel,
D.A. (1992)
Modeling
problems
in
conservation
genetics
using captive Drosophila
popu-
lations:
rapid
genetic
adaptation
to captivity.
Zoo Biology,
11,
333-342.
Gausen,
D. & Moen,
V.
(1991)
Large-scale
escapes
of farmed
Atlantic
salmon (Salmo salar) into Norwegian
rivers
threaten
natural populations.
Canadian
Journal of
Fish-
eries
and
Aquatic
Sciences,
48,
426-428.
Gjedrem,
T.,
Gj0en,
H.M. & Gjerde,
B.
(1991)
Genetic
origin
of
Norwegian
farmed
Atlantic
salmon.
Aquaculture,
98,
41-50.
Graham,
M.S. & Farrell,
A.P. (1992) Environmental
influ-
ences on cardiovascular
variables in rainbow trout,
Oncorhynchus
mykiss
(Walbaum).
Journal
of Fish
Biology,
41,
851-858.
Hansen,
L.P., Jacobsen,
J.A.
& Lund, R.A. (1993) High
numbers
of farmed
Atlantic
salmon,
Salmo salar, L.,
observed
in oceanic
waters
north
of
the
Faroe Islands.
Aquaculture
and
Fisheries
Management,
24,
777-781.
Hedrick,
P.W.
& Hedgecock,
D. (1994)
Effective
population
size
in
winter-run
chinook
salmon.
Conservation
Biology,
8, 890-892.
Heggberget,
T.G., Haukeb0,
T.,
Mork,
J.
& Stahl,
G. (1988)
Temporal
and spatial
segregation
of spawning
in sym-
patric
populations
of
Atlantic
salmon,
Salmo salar,
L. &
brown
trout,
Salmo
trutta L. Journal
of
Fish
Biology,
33,
347-357.
Hindar,
K., Ryman,
N. & Utter,
F. (1991)
Genetic
effects
of
cultured
fish
on natural
fish populations.
Canadian
Journal
of
Fisheries
and
Aquatic
Sciences,
48, 945-957.
Houlihan,
D.F., Agnisola,
C., Lyndon,
A.R., Gray,
C. &
Hamilton,
N.M. (1988)
Protein synthesis
in
a fish heart:
responses
to increased
power
output.
Journal
of
Exper-
imental
Biology,
137,
565-587.
Hutchings,
J.A.
(1991) The threat
of extinction
to native
populations
experiencing
spawning
intrusions
by
cultured
Atlantic
salmon. Aquaculture,
98,
119-132.
Jarvi,
T. (1990) The effects
of male
dominance,
secondary
sexual
characteristics
and female
mate
choice
on the
mat-
ing
success
of male
Atlantic
salmon
Salmo
salar.
Ethology,
84,
123-132.
Jones,
J.W.
(1959)
The Salmon.
Collins,
London.
Jonsson,
B. & Fleming,
I.A. (1993) Enhancement
of wild
salmon populations.
Human Impact
on Self-Recruiting
Populations
(ed. G. Sundnes),
pp. 209-242.
Tapir
Press,
Trondheim.
Jonsson,
B.,
Jonsson,
N. & Hansen,
L.P.
(1990)
Does
juvenile
experience
affect
migration
and
spawning
of adult
Atlantic
salmon?
Behavioural
Ecology
and
Sociobiology,
26,
225-
230.
Jonsson,
B.,
Jonsson,
N. & Hansen,
L.P. (1991)
Differences
in life history
and migratory
behaviour
between
wild
and
hatchery-reared
Atlantic
salmon
in nature.
Aquaculture,
98,
69-78.
Kohane,
M.J.
& Parsons,
P.A. (1988)
Domestication:
evolu-
tionary
change
under
stress.
Evolutionary
Biology,
23,
31-
48.
Lachance,
S. & Magnan,
P. (1990)
Performance
of
domestic,
hybrid,
and
wild
strains
of brook
trout,
Salvelinus fonti-
nalis,
after
stocking:
the
impact
of intra-
and
interspecific
competition.
Canadian
Journal
of
Fisheries
and
Aquatic
Sciences,
47,
2278-2284.
Ledig,
F.T. (1992) Human
impacts
on genetic
diversity
in
forest
ecosystems.
Oikos,
63,
87-108.
Leider,
S.A.,
Hulett,
P.L.,
Loch,
J.J. & Chilcote,
M.W.
(1990)
Electrophoretic
comparison
of
the
reproductive
success
of
naturally
spawning
transplanted
and wild
steelhead
trout
through
the
returning
adult
stage.
Aquaculture,
88,
239-
252.
Leitritz,
E. (1959) Trout and salmon
culture.
California
Department
of
Fish
and
Game,
Fish
Bulletin,
107,
1-169.
Loebel,
D.A., Nurthen,
R.K., Frankham,
R.,
Briscoe,
D.A.
& Craven,
D. (1992)
Modeling problems
in conservation
genetics
using captive Drosophila
populations:
conse-
quences of equalizing founder representation.
Zoo
Biology,
11,
319-332.
Lotspeich,
F.B. & Everest,
F.H. (1981) A new method
for
reporting
and
interpreting
textural
composition
of
spawn-
ing gravel.
United
States
Department
of Agriculture
and
Forestry
Service
Research
Note PNW-369.
Pacific
North-
?) 1996
British
Ecological
Society,
Journal
of Applied
Ecology,
33, 893-905
905
I.A. Fleming et al.
west Forest
and Range
Experimental
Station,
Corvallis,
OR.
Lund, R.A., Hansen,
L.P. & Okland, F. (1994) Escaped
farmed
salmon and
geographical zones
established for
wild
fish
protection. NINA Oppdragsmelding,
Trondheim,
303,
15
pp. [In
Norwegian
with
English
Abstract.]
Lura, H., Barlaup, B.T. & Sxgrov, H. (1993) Spawning
behaviour of
a farmed
escaped female Atlantic
salmon
(Salmo
salar).
Journal
of
Fish
Biology,
42,
311-313.
Lura,
H. & Okland, F. (1994) Content
of
synthetic
astax-
anthin in
escaped farmed
Atlantic
salmon, Salmo
salar,
L.,
ascending
Norwegian rivers.
Fisheries
Management
and
Ecology,
1,
205-216.
Lura,
H. & Sxgrov, H. (1991)
Documentation of
successful
spawning
of escaped farmed
female
Atlantic
salmon,
Salmo
salar,
in
Norwegian
rivers.
Aquaculture, 98, 151-
159.
Lura,
H. & Sxgrov,
H.
(1993)
Timing
of
spawning
in
cultured
and
wild Atlantic
salmon
(Salmo
salar) and
brown trout
(Salmo trutta)
in the
River
Vosso, Norway.
Ecology of
Freshwater
Fish, 2,
167-172.
Lyles, A.M. & May, R.M. (1987) Problems in
leaving the
ark.
Nature,
326,
245-246.
Philipp, D.P. (1991)
Genetic
implications of
introducing Flo-
rida largemouth
bass, Micropterus
salmoides
floridanus.
Canadian
Journal
of
Fisheries and Aquatic
Sciences, 48
(Suppl.
1), 58-65.
Pickering,
A.D. (1993) Endocrine-induced
pathology
in
stressed
salmonid fish.
Fisheries
Research, 17,
35-50.
Quinn,
T.P. & Foote,
C.J.
(1994) The effects of
body
size
and
sexual
dimorphism
on
the
reproductive
behaviour of
sockeye
salmon,
Oncorhynchus
nerka.
Animal
Behavior,
48,
751-761.
Reisenbichler,
R.R. & McIntyre,
J.D.
(1977)
Genetic differ-
ences in
growth
and
survival of
juvenile
hatchery
and
wild
steelhead
trout,
Salmo
gairdneri.
Journal
of
the
Fisheries
Research
Board
of
Canada,
34, 123-128.
Rice,
W.R. (1989)
Analyzing
tables
of
statistical tests.
Evolu-
tion, 43,
223-225.
Ruzzante, D.E. (1994) Domestication
effects on aggressive
and
schooling
behavior in
fish.
Aquaculture,
120,
1-24.
Schroder,
S.L. (1982) The influence of intrasexual
com-
petition on
the
distribution
of
chum
salmon in
an
exper-
imental
stream.
Salmon and Trout
Migratory
Behavior
Symposium
(eds E. L. Brannon
& E. 0. Salo), pp. 275-
285.
University
of
Washington
Press,
Seattle.
Skaala, 0. (1992) Genetic
variation
in
brown
trout
Salmo
trutta,
L. & application
of
genetic
markers in
studies on
geneflowfrom
culturedpopulations. PhD thesis,
University
of
Bergen.
Srivastava,
R.K. & Brown,
J.A. (1991) The biochemical
characteristics and
hatching
performance of
cultured
and
wild Atlantic
salmon
(Salmo
salar)
eggs. Canadian
Journal
of
Zoology,
69, 2436-2441.
Waples, R.S. (1991) Genetic
interactions
between
hatchery
and wild
salmonids: lessons from
the Pacific
Northwest.
Canadian
Journal
of
Fisheries and Aquatic
Sciences, 48
(Suppl.
1),
124-133.
Webb,
J.H.
& Hawkins,
A.D. (1989) The movements and
spawning
behaviour of
adult salmon in the
Girnock
Burn,
a tributary
of
the
Aberdeenshire
Dee (1986)
Scottish Fish-
eries
Research
Report,
40,
1-42.
Webb,
J.H.,
Hay,
D.W., Cunningham,
P.D. & Youngson,
A.F. (1991)
The spawning
behaviour of
escaped farmed
and
wild
adult Atlantic
salmon
(Salmo salary
L.)
in
a northern
Scottish
river.
Aquaculture,
98,
97-110.
Webb,
J.H.,
McLaren,
I.S., Donaghy,
M.J.
& Youngson,
A.F. (1993a)
Spawning
of
farmed
Atlantic
salmon,
Salmo
salar,
L.,
in
the second
year
after
their
escape.
Aquaculture
and
Fisheries
Management,
24,
557-561.
Webb,
J.H.,
Youngson,
A.F., Thompson,
C.E., Hay,
D.W.,
Donaghy, M.J. & McLaren, I.S. (1993b) Spawning
of
escaped
farmed
Atlantic
salmon, Salmo
salar, L.,
in
western
and northern
Scottish
rivers:
egg
deposition
by
females.
Aquaculture and
Fisheries
Management,
24, 663-670.
White,
H.C. (1942) Atlantic salmon redds and artificial
spawning
beds. Journal
of
the
Fisheries Research Board
of
Canada, 6,
37-44.
Youngson,
A.F.,
Martin,
S.A.M.,
Jordan, W.C. & Verspoor,
E. (1991)
Genetic
protein
variation
in
Atlantic salmon in
Scotland:
comparison
of wild and farmed
fish.
Aqua-
culture, 98,
231-242.
Received 16 September
1994; revision
received
13 January
1995
? 1996
British
Ecological
Society,
Journal of
Applied
Ecology,
33,
893-905
