Post on 29-Apr-2023
Relationship between Mercury Concentration and Growth Ratesfor Walleyes, Northern Pike, and Lake Trout from Quebec Lakes
MELYSSA LAVIGNE, MARC LUCOTTE,* AND SERGE PAQUET
Collaborative Mercury Research Network, Universite du Quebec a Montreal,C.P. 8888, Succursale Centre-Ville, Montreal, Quebec H3C 3P8, Canada
Abstract.—The relationship between mercury (Hg) concentrations in fish muscle and fish growth rates was
assessed for 54 walleye Sander vitreus, 52 northern pike Esox lucius, and 35 lake trout Salvelinus namaycush
populations throughout the Province of Quebec, Canada. We used the von Bertalanffy growth model to
estimate the ages of fish specimens for a given length, and Hg concentrations in fish specimens at
standardized length were determined via a quadratic regression model. Measured values of Hg concentrations
in walleyes, northern pike, and lake trout were then correlated to the estimated age at standardized length for
each population (375, 675, and 550 mm, respectively). A model-II regression was performed to describe the
existing relationships. Growth rates were positively related to Hg concentrations in walleyes and northern pike
(when three outliers were excluded), whereas no correlation was observed for lake trout. Our findings
demonstrate that slower-growing walleyes and northern pike have higher Hg concentrations at standardized
length. For these fish species, growth rate could be used as an integrated proxy to predict Hg concentration in
fish muscle on a regional scale. Our findings support the contention that biodilution can be an important factor
regulating mercury concentrations in fish. Thus, our findings suggest that proper control of fish growth rate
through fishing pressure, lake ecology, and watershed management could be used by fisheries management
authorities to minimize the toxic risk associated with Hg exposure from fish consumption.
As for many other regions in North America,
mercury levels in freshwater fish from Quebec
(Canada) often exceed the threshold value for human
consumption, especially in the case of predatory
species (Bodaly et al. 1993; Schetagne and Verdon
1999; Evans et al. 2005; Kamman et al. 2005; Lockhart
et al. 2005; Simoneau et al. 2005). Much work has
been accomplished to explain variability in Hg
concentrations in fish and to identify key factors
enabling adequate predictions of fish Hg levels in
specific lake environments. A scientific consensus is
currently emerging about the fact that such variability
is the result of complex interactions among numerous
environmental and biological factors (Parkman and
Meili 1993; Lucotte et al. 1999; Kainz et al. 2003;
Gorski et al. 2003; Montgomery et al. 2000; Wiener et
al. 2003). However, attempts to integrate these
interactions in order to predict specific fish Hg levels
and to identify Hg ‘‘hot spots’’ remain scarce,
particularly at a wide regional scale (EPRI 2003;
Roue-Le Gall et al. 2005; Surette 2005). In addition,
environmental factors have drawn much attention from
the scientific community (Cope et al. 1990; Bodaly et
al. 1993; Driscoll et al. 1994; Wiener et al. 2003).
Despite several studies showing relationships between
fish Hg concentration and length (Scott and Armstrong
1972; Olsson 1976; Schetagne and Verdon 1999) or
trophic level (Tremblay et al. 1998), biological factors,
such as fish growth rate, are generally overlooked.
Fish growth rate, which can be defined as the
temporal change in either fish weight or length (Ricker
1980; Weatherley and Gill 1987; Hopkins 1992), has
long been assumed to play a hypothetical role in the
biodilution of contaminants (de Freitas et al. 1974;
Norstrom et al. 1976). Such influence of growth rate on
Hg accumulation and hence Hg biodilution in fish
muscle was first tested in small Finnish and Swedish
oligotrophic lakes, where intensive fishing experiments
were reported to decrease Hg concentrations in the
remaining fish (Gothberg 1983; Jernelov and Lann
1973; Verta 1990). A similar experiment was later
applied to small lakes located in northern Quebec. In
two of these experimental lakes, Hg levels in walleye
Sander vitreus populations declined by 33%, and
growth rates of these populations increased by 35%(Doire 2003; Surette 2005). Decline in Hg levels was
not explained by changes in fish diet, structural
alterations of the trophic web (Doire 2003), a reduction
in methylmercury (MeHg) levels in forage fish, or by a
reduction in the whole-lake MeHg content following
fish removal (Surette 2005; Surette et al. 2005).
Through comparisons of Hg concentration in fish
before and after intensive fishing, Surette (2005)
hypothesized that biodilution would partly explain the
observed decrease in walleye Hg concentrations.
Numerous other studies using a variety of experimental
* Corresponding author: lucotte.marc_michel@uqam.ca
Received March 28, 2008; accepted April 26, 2010Published online October 25, 2010
1221
North American Journal of Fisheries Management 30:1221–1237, 2010� Copyright by the American Fisheries Society 2010DOI: 10.1577/M08-065.1
[Article]
approaches appear to corroborate this hypothesis
because they reported lower fish Hg levels in faster
growing populations (Rask et al. 1996; Doyon et al.
1998; Harris and Bodaly 1998; Stafford and Haines
2001; Essington and Houser 2003; Stafford et al. 2004;
Simoneau et al. 2005). However, most of the
aforementioned studies are either limited in terms of
number of lakes and fish species considered or are
vague about the actual contribution of growth rate to
biodilution. Simoneau et al. (2005) were the first to
point out growth rate as a key factor to explain
interlake fish Hg level variability; studying walleye
populations from 12 Quebec (Canada) lakes, these
authors reported that Hg levels in fish muscle at a
standardized length of 350 mm were significantly
correlated to estimated ages at this length.
In this study, we further explored the relationship
between Hg concentration and growth rate in fish
populations from 118 lakes in Quebec, Canada. Three
predator fish species that typically exhibit Hg levels
exceeding the Canadian threshold value for human
consumption (0.5 mg/kg of fish body weight) were
considered: walleyes, northern pike Esox lucius, and
lake trout Salvelinus namaycush. We specifically
aimed at verifying the existence of a positive
relationship between fish muscle Hg level at standard-
ized length (Lstd
), as estimated by quadratic regression
models (Tremblay et al. 1998) and growth indexes
obtained using the von Bertalanffy model (Ricker
1980). This approach presents the advantage of using
two simple models that can be applied to a wide series
of lakes. Compared with previous studies, our study
used a markedly broader data set of lakes over an
extended spatial area covering both the boreal and
temperate domains. Such a large spatial distribution of
lakes also allowed us to examine the relationships
between fish characteristics (Hg concentration and age
at Lstd
) and geographical as well as hydrographic
characteristics. Because the species we considered are
highly valued and frequently consumed by sport fishers
(Scott and Crossman 1974), we also discuss the
implications of our findings for the management of
wild freshwater fish resources.
Methods
Study area.—This study area covers 1,056,000 km2
within the Province of Quebec, stretching from 46–
588N and from 64–808W. The zone considered
encompasses both densely and sparsely populated
regions of Quebec, most of the hydrographic basins
of this province as well as many of its ecozones (Figure
1). A total of 118 natural lakes sustaining weak to high
fishing pressure were selected. The main characteristics
of the selected lakes are given in Lavigne (2007).
Database assemblage.—All data used in this study
were provided by five entities: the Collaborative
Mercury Research Network (COMERN); the Ministry
of Natural Resources and Wildlife of Quebec (MNRF);
the Ministry of Sustainable Development, Environ-
ment, and Parks of Quebec (MDDEP); the Energy
Corporation of James Bay (SEBJ); and the hydropower
corporation Hydro-Quebec (HQ).
Lake-based data were assembled for the three top
predator species—walleye, northern pike, and lake
trout—where fish length, age, and Hg concentration
(based on wet weight) were available for at least 10
specimens of a given species. From this data screening,
we obtained a database containing information on
5,207 walleyes (54 lakes), 2,419 northern pike (52
lakes), and 1,478 lake trout (35 lakes; see Appendix).
Fish sampling.—The fish sampling was conducted
by COMERN from 2001 to 2004, by MRNF from
1999 to 2004, by SEBJ from 1978 to 1985, and by HQ
from 1986 to 2001. Similar protocols were applied by
all for fish capture and sample collections. Experimen-
tal variable-mesh gill nets were either 46 or 61 m long
3 1.82 m high, with variable bar mesh sizes of either
2.5, 3.2, 3.8, 5.1, 6.4, 7.6, and 10.2 cm or 2.5, 3.2, 3.8,
5.1, 6.4, 7.6, 10.2, 12.7, and 15.2 cm. Total length
(mm), weight (g), sex, and maturity stage were
determined whenever possible. Aging structures such
as otoliths, scales, or bones were also collected for age
determination. A sample of fish muscle was taken from
the caudal region for Hg analysis. Bones and skin were
removed from muscle samples prior to preservation.
Samples of fish muscle were preserved frozen or
freeze-dried until being analyzed.
Age determination.—Age determination was per-
formed on aging structures using opercula, otoliths, and
(or) dorsal spines for walleyes (Campbell and Babaluk
1979; Pepin and Levesque 1985; Babaluk and Campbell
1987; Babaluk et al. 1993); cleithra or scales for
northern pike (Casselman 1974; Babaluk and Craig
1990; Laine et al. 1991); and otoliths for lake trout
(Dubois 1967; Lapointe and Clement 1985; Casselman
1987). All age estimations were performed at least
twice, typically by two different readers or until
agreement was reached on an age value. If disagreement
persisted on the age value after the third reading, the
structure was discarded and the age data were rejected.
Mercury analysis in fish muscle.—Total Hg concen-
tration in fish muscle was determined either by cold
vapor atomic fluorescence spectrometry (CVAFS;
typical detection limit of 1 g/L) or by cold vapor
atomic absorption spectrometry (CVAAS; typical
detection limit of 1 lg/L;) (Pichet et al. 1999). Details
on analytical procedures can be found elsewhere
(Tremblay et al. 1996; Pichet et al. 1999; CEAEQ
1222 LAVIGNE ET AL.
2003). The University of Quebec at Montreal, Mercury
Analytical Laboratory used CVAFS to perform anal-
yses on samples collected by COMERN, while the
Quebec Expert Center for Environmental Analyses
(Centre d’Expertise en Analyse Environnementale du
Quebec [CEAEQ]) analyzed samples collected by
MRNF using CVAAS (CEAEQ 2003). Fish caught
by SEBJ and HQ were analyzed by a series of
independent laboratories using standard CVAAS
protocol, and unidentified triplicates were introduced
at a rate of 10% to ensure quality control. All
participating laboratories followed rigorous quality
assurance–quality control procedures, including peri-
odical analytical blanks, sample replicates, certified
fish tissue standards and calibration curves. All
analytical facilities were also part of the interlaboratory
calibration program set for Hg by the Canadian Food
Inspection Agency. As a final validation procedure,
measured Hg values (mg/kg of wet weight) were
qualitatively tested using Hg level versus fish length
dispersion diagrams, and outlier samples were either
subjected to a new analysis or rejected. Sample
FIGURE 1.—Map of the Province of Quebec showing the locations of the lakes (black dots) selected to assess mercury
concentrations in fish muscle and fish growth rates and the 11 hydrological basins (00–10) in which the lakes lie; lake numbers
are those assigned by Lavigne (2007).
MERCURY AND FISH GROWTH IN QUEBEC LAKES 1223
rejection rates were 0.6% for walleyes, 1.4% for
northern pike, and 0.3% for lake trout.
Fish growth rate model.—The relationships between
fish length and age were determined using the von
Bertalanffy growth model (Ricker 1980):
LT ¼ L‘ð1� e�K½t�t0 �Þ; ð1Þ
where LT
is the total length (mm) of fish at time t(years; which represents the fish age as determined
with aging structures), L‘
is the asymptotic length
(mm), K is the growth coefficient, and t0
(years) is the
hypothetical fish age at length ¼ 0 mm. For each
combination of fish species and lake, we fitted the von
Bertalanffy growth model to age and total length data
using nonlinear regressions, for a total of 141 models.
Mercury concentrations versus fish length.—The
relationships between Hg concentration in fish muscle
and fish length were modeled using quadratic regres-
sions (Tremblay et al. 1996; 1998) of the form
½Hg� ¼ aþ bðLTCÞ þ cðLTC2Þ; ð2Þ
where [Hg] is the Hg concentration measured in fish
muscle (in mg/kg of wet body weight)and LTC is the
fish total length (mm). The intercept (a) as well as the
coefficients b and c were parameterized by applying
equation (2). The method used by Tremblay et al.
(1998) is robust enough to fit both linear and nonlinear
regressions. When the quadratic regression did not fit,
as indicated by a nonsignificant LTC2 parameter, a
simple linear regression was applied to describe the
fish population.
Fish standardized length.—For each species, an
average total length was calculated for each lake. Then,
for each species, we averaged the mean total lengths
across all lakes; we referred to this grand average as the
specific mean length. The mean specific length was
rounded to the nearest 25 mm to generate the
standardized length, Lstd
.
Correlation analysis.—Pearson’s product-moment
correlation coefficient analyses were used to assess
the significance of the relations between mercury
concentration at the standardized length ([Hg]Lstd
)
and age at the standardized length (AgeLstd
) for each
species. Data to be included in the Pearson analyses of
[Hg]Lstd
as a function of AgeLstd
, were selected
according to three criteria. (1) Lakes where the
regression coefficient (r2) was not statistically different
from a zero slope (P , 0.05) for one of the two models
were removed from further analysis. (2) The model’s
fit to the observed values was assessed visually using
dispersion diagrams and the model’s validity con-
firmed if the Lstd
was not located at one of the
extremities of the length distribution and the Lstd
was
not outside of the range of the model. (3) A
homogeneous data distribution around the Lstd
was
observed. Following these criteria, we kept 54 out of
58 lakes for walleyes, 52 of 70 for northern pike, and
35 of 50 for lake trout. Data exclusion from further
analysis was mainly due to fish length values below
Lstd
or insufficient numbers of fish.
A standard major axis model-II regression was then
applied to describe the relations (Legendre and
Legendre 1998). Model-II regression was preferred to
simple linear regression because [Hg]Lstd
and AgeLstd
are both affected by independent natural variations.
The standard major axis model-II was chosen among
the different types of model-II regressions considering
that this model was best adapted to properly describe
our data set, which was characterized by the following:
(1) correlations were statistically significant, (2) the
variables compared were not of the same physical
units, (3) the variances of the x-axis and y-axis were
comparable, and (4) the normality criterion for both
axis variables was met using Shapiro–Wilk W tests
(Legendre and Legendre 1998; Legendre 2001).
Furthermore, we tested the presence of outliers using
Mahalanobis distances on the dispersion diagrams of
the Pearson analyses between [Hg]Lstd
and AgeLstd
.
When we found outliers, we then double-checked with
jackknifed distances, which reduces distortion.
Relation of fish characteristics to latitude andhydrographic region.—Multiple correlation analysis
was applied to [Hg]Lstd
values, AgeLstd
values and
latitude of the lakes for each fish species. This type of
analysis allowed calculating a Pearson’s product-
moment correlation coefficient (rPearson
) indicative of
the presence and strength of a linear relation between
variables. Bonferroni corrections were applied on
threshold values because multiple testing was involved
(Legendre and Legendre 1998).
The selected lakes were sorted into four groups
according to both their ecological (ecozones) and
hydrographical (hydrographic regions) characteristics.
As the selected lakes were distributed over a very large
territory, including four ecozones, each one with is own
pedological structure that can influence the Hg
dynamics in the watershed. Furthermore, these lakes
spread over 10 different hydrographic regions (Figure
1; Marshall and Schut 1999; CEHQ 2005) and could
host distinct fish subpopulations, each with their own
genetic characteristics that potentially modulate their
growth rate or the Hg toxicokinetics. Group 1 includes
the St. Lawrence River hydrographic region (00) and
Chaleur Bay and Perce (01), which corresponds to the
Mixed Wood Plains ecozone. Group 2 comprises the
following hydrographic regions located in the southern
Boreal Shield ecozone: St. Lawrence Southeast (02),
1224 LAVIGNE ET AL.
St. Lawrence Southwest (03), Outaouais and Montreal
(04), St. Lawrence Northwest (05), Saguenay–Lake
Saint-Jean (06), and St. Lawrence Northeast (07).
Group 3 includes the northern hydrographic region of
Hannah and Rupert bays (08). Group 4 covers the two
hydrographic regions of James Bay: Hudson Bay (09)
and Ungava Bay (10), which are located in the far north
portion of Quebec’s territory, in the northern Boreal
Shield ecozone and in the Taiga Shield ecozone.
Considering that the statistical distribution of
[Hg]Lstd
and AgeLstd
values within fish subsets did
not meet the criteria for the application of analysis of
variance (ANOVA; Shapiro–Wilk W and Levene tests
for homocedasticity of variance), nonparametric Krus-
kall–Wallis tests were applied to determine significant
differences in mean [Hg]Lstd
and mean AgeLstd
across
the four lake groups. Then, Noether nonparametric
multiple comparison tests were performed to find
which particular groups differed from the others. We
set a ¼ 0.05 for all of our statistical testing.
Statistical software.—Quadratic regressions and von
Bertalanffy models were carried out using the SPSS
v13 software (SPSS 2004). Other statistical analyses
were performed using the JMP 5.0 software (SAS
Institute 2002), with the exception of the model-II
regression analysis that was conducted using the
model-II regression and permutation test software
(Legendre 2001).
Results
Data Range
Total Hg levels measured in fish muscle were highly
variable, ranging from 0.06 to 3.71 mg/kg of muscle
(wet weight) for walleyes, from 0.07 to 4.11 mg/kg for
northern pike, and from 0.05 to 6.22 mg/kg for lake
trout. Fish total length also varied within and among
species, ranging from 110 to 780 mm for walleyes,
from 150 to 1,170 mm for northern pike, and from 120
to 1,060 mm for lake trout. Finally, age ranges (years)
were 0–33 for walleyes, 0–24 for northern pike, and 1–
41 for lake trout.
Calculation of Standardized Length
Considering all selected lakes, average total lengths
(x) for walleye, northern pike and lake trout popula-
tions ranged between 290 and 500 mm (mean specific
length ¼ 384.4 mm), 515–904 mm (mean specific
length¼ 668.3 mm), and 340–700 mm (mean specific
length ¼ 561.5 mm), respectively. Based on mean
specific lengths, we then determined Lstd
to be 375 mm
for walleye, 675 mm for northern pike and 550 mm for
lake trout.
Growth Curves and Mercury Level versus Fish Length
When considering the whole data set, mean AgeLstd
was equal to 10.81 (SD, 0.95) for slow-growing
walleyes, 7.98 (SD, 1.30) for medium-growing wall-
eyes, and 3.40 (SD, 0.72) for fast-growing walleyes.
Means for northern pike were 9.74 (SD, 0.56; slow),
7.00 (SD, 0.60; medium) and 4.90 (0.79; fast). For the
lake trout, means were 22.51 (SD, 3.36; slow), 11.91
(SD, 1.82; medium), and 7.40 (SD, 1.28; fast).
The von Bertalanffy models indicated a fairly good
agreement between age and length, r2 values ranging
from 0.77 to 0.95 for the nine illustrative populations
shown in Table 1 (mean ¼ 0.81). In comparison, the
TABLE 1.—Parameters for the von Bertalanffy growth model (equation 1, where L‘¼ asymptotic length, K ¼ the growth
coefficient, and t0¼ hypothetical fish age at length ¼ 0 mm), age at standardized length (Age
Lstd), parameters for quadratic
regressions (equation 2, where a ¼ the intercept and b and c ¼ parameterized coefficients) and muscle mercury content at
standardized length ([Hg]Lstd
) for walleye, northern pike, and lake trout populations, illustrating three distinct growth patterns
(slow, medium, and fast).
Lakea N L‘
K (310�2) t0
AgeLstd
b (year) r2 a b c (310�3)[Hg]
Lstd
b
(mg 3 kg�1) r2
Walleye (375 mm)
Dana 32 889 3.98 �3.41 10.4 6 1.2 0.891 �924 5.72 9.29 1.22 6 0.26 0.751Mesgouez 27 891 5.42 �2.64 7.4 6 0.7 0.934 �318 1.82 4.48 0.36 6 0.05 0.917Saint-Francois 34 810 1.45 �1.52 2.8 6 0.5 0.876 �331 1.31 2.82 0.19 6 0.06 0.803
Northern pike (675 mm)
Pusticamica 27 2,565 2.59 �2.92 8.9 6 0.4 0.934 �734 2.78 �4.55 1.07 6 0.18 0.581Ouescapis 24 1,138 12.62 �1.09 6.0 6 0.4 0.939 �439 1.73 �1.57 0.73 6 0.09 0.759Waconichi 25 1,096 21.60 �0.86 3.6 6 0.4 0.892 �213 0.61 0.66 0.20 6 0.02 0.855
Lake trout (550 mm)
Amichinatwayach 44 1,166 3.67 �3.89 13.5 6 1.4 0.826 �1,026 2.73 5.52 0.48 6 0.10 0.786Tilly 40 1,433 3.57 �2.24 11.3 6 0.6 0.949 �733 2.03 0.007 0.39 6 0.10 0.820Chibougamau 110 844 8.33 �5.21 7.5 6 1.5 0.768 2,670 5.55 12.37 0.51 6 0.12 0.628
a Refer to Figure 1 and Table A.1 for lake identification and basic data.b Mean 6 95% confidence limits.
MERCURY AND FISH GROWTH IN QUEBEC LAKES 1225
quadratic models depicting the relationship between
fish Hg concentration ([Hg]Lstd
) and length were more
variable, r2 ranging from 0.58 to 0.92 (mean ¼ 0.70).
Table 1 clearly shows that higher [Hg]Lstd
values
correspond to slower growth rates and inversely.
Figure 2 illustrates the accumulation ([Hg]Lstd
versus
LT) and growth ([Hg]
Lstdversus Age
Lstd) curves for the
nine illustrative fish populations presented in Table 1.
Accumulation curves showed a fair segregation
between slow, medium, and fast growing populations
of walleyes and northern pike (Figure 2A, C), whereas
corresponding curves for lake trout exhibited large
regions of overlapping values of the three populations
(Figure 2E). Similarly, growth-curve segregation was
better in the case of walleyes and northern pike (Figure
2B, D) than it was for lake trout (Figure 2F). Curve
patterns also differed between walleyes and northern
pike (on the one hand) and lake trout (on the other
hand). Although the highest Hg accumulation is
associated with the slowest growth in walleyes and
northern pike, an inconsistent pattern is obtained in
lake trout, as shown for Lake Amichinatwayach, where
[Hg]Lstd
is intermediary when lake trout AgeLstd
is the
highest and there is a 95% CI overlap with the
medium- and fast-growing subsets.
Muscle Hg Levels Versus Growth
Figures 3, 4, and 5 present the dispersion diagrams
for the different estimated [Hg]Lstd
versusAgeLstd
values for all fish populations considered, as well as
model-II regressions when the linear relationship
(correlation coefficient) was statistically significant
(rPearson
6¼ 0 for P , 0.05). The correlations and
model-II regressions on the estimated values constitute
a simple way to illustrate the relationships found
between muscle Hg levels and growth rates for the
different fish populations from one lake to another.
Lakes situated on the left side of the three figures
exhibit fish with fast growth rates, whereas those to the
right side are those where fish are growing slowly.
There is a great variability in [Hg]Lstd
at a given
AgeLstd
for all three fish species considered, but on the
overall AgeLstd
variation scale, [Hg]Lstd
values are
fairly similar between species: 0.15–1.22 mg/kg for
walleyes, 0.20–1.62 mg/kg for northern pike, and
0.11–1.01 mg/kg for lake trout. Ranges for AgeLstd
also
greatly vary: 2.1–12.2 years for walleyes, 3.6–10.1
years for northern pike, and 5.5–14.2 years for lake
trout.
Standard major axis model-II regression provided a
functional equation with slope values that sensibly
differ between walleyes (0.09 mg/kg) and northern
pike (0.22 mg/kg; Figures 3, 4). This suggests higher
Hg concentrations for northern pike at any given
AgeLstd
. However, as shown on the dispersion diagram,
[Hg]Lstd
levels for northern pike were much more
variable than for walleyes. When applying the
Mahalanobis distance outlier test to exclude three lakes
exhibiting extreme values (Gaotanaga, Coutaceau, and
Dana) from the model, the regression for northern pike
was stronger. No significant relationship was found
between [Hg]Lstd
and AgeLstd
for lake trout even though
the ranges covered by these variables in that fish
population were quite similar to those of the two other
species (Figure 5).
Multiple Correlations for [Hg]Lstd
, AgeLstd
, Latitude,and Longitude
The moderate but still significant positive value of
rPearson
in Table 2 suggests that latitudinal distribution
of lakes influences the growth rate of walleyes.
Meanwhile, longitudinal distribution of lakes present
a lot more significant relationships with AgeLstd
for
walleyes and with [Hg]Lstd
and AgeLstd
for northern
pike, but the rPearson
values remained weak. On the
other hand, [Hg]Lstd
and AgeLstd
for lake trout
populations were not related to the latitudinal or
longitudinal positions.
Distribution of [Hg]Lstd
and AgeLstd
by HydrographicRegions
Values of [Hg]Lstd
and AgeLstd
values were signif-
icantly lower for the walleyes living in the southern
hydrographic regions (Figure 6A, B). The highest
[Hg]Lstd
values for walleyes were found in the middle
eastern part of Quebec, and the highest AgeLstd
were
found from the middle to the northern parts of the
provincial territory. There was no statistical difference
for the distributions of [Hg]Lstd
and AgeLstd
between
hydrographic regions for northern pike (Figure 6C, D)
and lake trout (Figure 6E, F). These findings are in
agreement with the analysis of latitudinal correlations.
DiscussionMercury Variability in Muscle
Our results showed a high variability of Hg levels in
three piscivorous fish species, with concentrations
ranging over two orders of magnitude. Among the
9,104 fish specimens analyzed in this study, 50% were
above the 0.5 mg Hg/kg threshold value for human
consumption in Canada, the highest concentration was
6.00 mg Hg/kg of wet body weight. The [Hg]Lstd
values for the 118 lakes studied varied by a factor of up
to 10, and more than 60% of these lakes were hosting
individuals with [Hg]Lstd
over 0.5 mg Hg/kg. The
distribution of [Hg]Lstd
values from one lake group to
another also illustrates the extent of this variability,
considering that the only statistically significant
1226 LAVIGNE ET AL.
FIGURE 2.—Quadratic regressions (Hg levels versus total length; left panels) and von Bertalanffy growth curves (total length
versus age; right panels) for representative (A)–(B) walleyes, (C)–(D) northern pike, and (E)–(F) lake trout populations in Quebec.
For each combination of the three fish species and three growth rates (slow [diamonds], medium [times signs], and fast [circles]), one
representative population (lake; see Table 1) was chosen, namely, Dana (slow), Mesgouez (medium), and Saint-Francois (fast) in (A)
and (B); Pusticamica, Ouescapis, and Waconichi in (C) and (D); and Chibougamau, Tilly, and Amichinatwayach in (E) and (F).
MERCURY AND FISH GROWTH IN QUEBEC LAKES 1227
difference in [Hg]Lstd
regional distributions was found
between group 1 (St. Lawrence River – 00) and group
3 (Hannah Bay and Rupert Bay – 09), and only in the
case of walleye populations. It is rather surprising that
differences in [Hg]Lstd
cannot be further related to
geographic or hydrographic repartition of the lakes,
considering the large area covered by this study (800
km E-W by 1320 km S-N) and the fact that most fish
populations in a given hydrographic region have been
isolated from each other (Scott and Crossman 1974).
Variability in Growth Rates
High variability in the growth rate of different fish
populations of a given species has often been reported
in the literature (Pauly 1978; Weatherley and Gill
1987; Froese and Pauly 2000). Such variability is also
reflected in this study; that is, the observed distribution
of AgeLstd
varied between 7 and 10 years for each
species considered. The large variability in AgeLstd
observed for walleye populations is similar to that
reported by Simoneau et al. (2005). These authors also
reported fast growth rates for walleye populations, even
in northern locations, as well as high intraregional
growth rate variability. Despite the high variability
observed in walleye growth rates, there is a positive
correlation between AgeLstd
and latitude for the 54
lakes for which data are available (rPearson
¼ 0.5307, P, 0.0001; Table 2). This observation is in agreement
with the general trend in walleye growth rates, reported
to usually be slower in northern regions (Colby and
Nepszy 1981; Galarowicz and Wahl 2003). This trend
could also explain the significant difference observed
between the distributions of AgeL375
for group 1 (St
Lawrence River hydrographic region) and groups 3 and
4(northern hydrographic regions; Figure 6). Similarly,
the same trend could explain differences in the
distribution of [Hg]L375
between groups 1 and 3
(Figure 6), considering the strong and highly signifi-
FIGURE 3.—Estimated Hg levels at 375 mm standardized length ([Hg]L375
) versus estimated ages at that length (AgeL375
)
for walleyes.
1228 LAVIGNE ET AL.
cant correlation existing between [Hg]L375
and AgeL375
for walleyes (rPearson
¼ 0.7410, P , 0.0001; Table 2).
On the other hand, the absence of such a trend between
AgeLstd
and latitude or hydrographic regions for
northern pike and lake trout populations (Table 2;
Figure 6) probably reflects the high interlake variability
of growth rates we encountered. The latter observation
contradicts the generally accepted positive relations
reported between fish growth rates and the water
temperature (or the number of degree-days) and the
inverse relation between growth and latitude (Cassel-
man 1978; Pauly 1978; Colby and Nepszy 1981;
O’Connor et al. 1981; Weatherley and Gill 1987; Clapp
and Wahl 1996; Claramunt and Wahl 2000; Galar-
owicz and Wahl 2003). Both lower [Hg]L675
and lower
AgeL675
in more temperate climate could also be
explained by the influence of denser human popula-
tions and thus enhanced anthropogenic nutrient
loadings to lakes, which in turn enhances ecosystems
productivity and fish growth rates (Fortin et al. 1996).
Muscle Hg Levels and Growth Rates
The strong positive relationship we observed
between [Hg]L375
and AgeL375
for the 54 walleye
populations (r2¼ 0.549, P , 0.0001; Figure 3), which
suggest an influence of growth rate on Hg concentra-
tion in this fish species, is similar to that obtained by
Simoneau et al. (2005) between [Hg]L350
and AgeL350
for 12 walleye populations from Quebec (rPearson
¼0.9244, P , 0.0001). Likewise, using a bioenergetics
model, Harris and Bodaly (1998) estimated that 19% of
the variability in walleye muscle Hg levels from two
Ontario lakes could be related to growth rates.
We found a weaker relationship between Hg]L675
and AgeL675
for northern pike, even after removing
three outliers from the model. Elsewhere, Wren and
MacCrimmon (1986) reported higher levels of Hg in
FIGURE 4.— Estimated Hg levels at 675 mm standardized length ([Hg]L675
) versus estimated ages at that length (AgeL675
) for
northern pike. Open circles denote outliers identified with Mahalanobis distances on the correlation.
MERCURY AND FISH GROWTH IN QUEBEC LAKES 1229
FIGURE 5.—Estimated Hg levels at 550 mm standardized length ([Hg]L550
) versus estimated ages at that length (AgeL550
) for
lake trout.
TABLE 2.—Matrix of Pearson’s product-moment correlation coefficients for walleyes, northern pike, and lake trout from
selected lakes, where AgeLstd
is age at standardized length and [Hg]Lstd
is muscle mercury content at standardized length. Values
in bold italics are statistically different from 0; P-values are in parentheses. Asterisks denote statistical significance following
Bonferroni correction with the adjusted a ¼ 0.0167 (i.e., 0.05/3).
[Hg]Lstd
AgeLstd
Latitude Longitude
Walleye (N ¼ 54; std ¼ 375)
[Hg]Lstd
1.0000Age
Lstd0.7410 (,0.0001)* 1.0000
Latitude 0.2097 (0.1280) 0.5307 (,0.0001)* 1.0000Longitude �0.3877 (0.0038)* �0.3333 (0.0138)* �0.2095 (0.1285) 1.0000
Northern pike (N ¼ 52; std ¼ 675)
[Hg]Lstd
1.0000Age
Lstd0.5204 (0.0001)* 1.0000
Latitude �0.1603 (0.2562) 0.0806 (0.5702) 1.0000Longitude �0.4792 (0.0003)* �0.4815 (0.0003)* �0.0495 (0.7273) 1.0000
Lake trout (N ¼ 35; std ¼ 550)
[Hg]Lstd
1.0000Age
Lstd0.1837 (0.2907) 1.0000
Latitude 0.3284 (0.0541) 0.2094 (0.2273) 1.0000Longitude 0.3289 (0.0537) �0.1657 (0.3414) 0.0598 (0.7330) 1.0000
1230 LAVIGNE ET AL.
FIGURE 6.—Relationships between (1) estimated Hg levels and ages at standardized lengths for (A)–(B) walleyes, (C)–(D)northern pike, and (E)–(F) lake trout and (2) groupings according to watershed. Group 1 corresponds to watershed 01 in Figure
1; group 2 corresponds to watersheds 02–07; group 3 corresponds to watershed 08; and group 4 corresponds to watersheds 09
and 10; the number of fish populations in each group is given in parentheses.; outliers (circles) are identified in Figure 4. Plots
with the same letter are not significantly different.
MERCURY AND FISH GROWTH IN QUEBEC LAKES 1231
the muscle of northern pike from Lake Tadenac
compared with Tadenac Bay in Ontario, whereas
growth rates were higher in the lake. However, many
fish species known as prey for northern pike were
reported to have higher Hg levels in Lake Tadenac than
in Tadenac Bay, and higher Hg levels in the lake
sediments were measured. Other intensive fishing
experiments held in small oligotrophic lakes of Sweden
and Finland led to a decrease muscle Hg levels in
northern pike and a concurrent increase in their growth
rates (Gothberg 1983; Verta 1990). However, even
though the Quebec intensive fishing experiment
successfully lowered northern pike Hg levels in two
of the five lakes under study, no concurrent change in
fish growth rates in the lakes was observed (Doire
2003; Surette 2005). This would indicate that in this
case biodilution was not responsible for the observed
decrease in northern pike Hg concentrations.
We determined that lake trout [Hg]L550
cannot be
explained by variations in AgeL550
. Another study
reported a similar lack of consistency between Hg
levels and growth rates for lake trout (Stafford and
Haines 2001). Inversely, Stafford et al. (2004) found a
significant negative relationship between the two
variables in Lake Flathead, Montana. However, these
two studies differed completely from ours in both the
experimental design and statistical treatment of the
data, and thus, results are hardly comparable. Evans et
al. (2005) reported that lake trout and walleyes from 20
lakes located in the Mackenzie River basin (Canada)
have higher mean age than those of the northern
Alberta, whereas measured mean Hg concentrations
were higher in the Mackenzie River basin (north)
compared with northern Alberta (north–south gradi-
ent). The authors propose that slower growth rates in
the north might be responsible for these observations.
Although biodilution applies to all fish species
(Jackson and Schindler 1996), this effect regarding
Hg concentrations in lake trout sampled in this study
appears masked by other factors affecting the accumu-
lation of the contaminant in the fish flesh.
Factors Affecting Growth Rate and Hg Accumulation
The strong relationships observed between [Hg]Lstd
and AgeLstd
for walleyes and northern pike seem to be
related to a biodilution process (Figures 3, 4). Data
presented here were gathered in situ. Thus, the
influences of numerous biotic and abiotic factors
affecting Hg behavior in aquatic environments and fish
growth rate were not segregated. One of (or a
combination of) these factors may have an impact on
Hg bioavailability, biomagnification, and bioaccumula-
tion, on the one hand, and on fish growth on the other
hand. A study by Roue-Le Gall et al. (2005) can be used
to illustrate the latter. Using a model based on simple
environmental parameters, these authors were able to
qualitatively predict the Hg accumulation patterns for
yellow perch Perca flavescens, walleyes, and northern
pike, and successfully ranked a set of 45 lakes from six
regions of the boreal forest (Quebec) from the highest to
the lowest Hg levels found in fish. The variables used by
the authors were (1) the ratio between primary watershed
area and lake area, (2) the ratio between drainage area
and lake area, (3) riparian wetland coverage, (4) land use
and vegetation coverage of the primary watershed, (5)
water quality variables (dissolved organic carbon
concentration, pH, chlorophyll a), and (6) sportfishing
intensity. The selected criteria thus excluded fish growth
rates, even if many of these factors may covary with or
against this parameter.
Given the complexity of the interactions between all
environmental factors having an impact on either the
Hg cycle in aquatic environments or its accumulation
in fish flesh, and given the variability exhibited by
these parameters from one lake to another, it becomes
clear that the modeling of Hg pathways from sources to
fish muscle will remain an intricate task. Our results
demonstrate that fish growth rate represents an
integrated parameter with the potential to help model
Hg accumulation in walleyes and northern pike on an
intraregional to interregional scale.
Conclusions
We developed an extensive database integrating
information from several sources on fish age, length,
and Hg concentrations for thousands of fish specimens
from three species commonly captured by sport fishers
in Quebec. More than 50% of the specimens had Hg
levels in their muscular tissues exceeding the threshold
mercury value of 0.5 mg/kg of fish wet body weight for
human consumption in Canada. More than 60% of the
lakes had fish at standardized mean length exceeding
this value. The use of a growth indicator—the
estimated age at standardized length (AgeLstd
)—
enabled us to explain more than 50% of the fish
muscle Hg concentration at standardized length for
walleyes, and between 25% and 50% of the variation
for northern pike populations living in more than 50
Quebec lakes. AgeLstd
values were also correlated with
latitude for walleye populations and with longitude for
northern pike. Lakes located in the northern part of
Quebec are consequently more likely to contain
walleyes with elevated muscle Hg levels. Contrary to
what was previously reported or suggested, the
variations in fish muscle Hg levels for lake trout
populations do not seem to be correlated to growth rate.
For this species, no correlation between AgeLstd
and
[Hg]Lstd
was observed. We hypothesize that Hg
1232 LAVIGNE ET AL.
biodilution following faster growth rate coupled to
multiple covarying factors affecting both Hg accumu-
lation in the fish and growth rate may explain lower Hg
levels for walleyes and northern pike. In this context,
growth rate represents a valid integrative indicator of
multiple environmental factors and varies in the
significance of its effect on Hg concentrations in
muscle tissues of theses species.
Quebec has one of the world’s greatest densities of
lakes (Downing et al. 2006). Considering the extensive
use of Quebec’s fish resources for commercial,
subsistence, and leisure, the management of human
health risks related to Hg in fish remains a critical
issue. Considering the high variability of Hg levels in
fish muscle reported here, proper Hg risk management
suggests extensive environmental monitoring efforts or
development of an integrated indicator enabling intra-
regional to interregional modeling of the Hg situation.
We consider that the simple approach applied in this
study allows for the achievement of such a task on
large geographic scales that would minimize sampling
efforts; however, this would involve accepting loss of
prediction preciseness compared with other sophisti-
cated modeling tools, such as the Onefish (Harris and
Bodaly 1998) or the Dynamic Mercury Cycling model
(EPRI 2002) models. Fish growth rates are often used
as indicators of the performance of fish population or
of the health of lake ecosystems. This information is
thus frequently available for numerous lakes (Munkit-
trick and Dixon 1989) and could be used for the
elaboration of regional recommendations on safe fish
consumption limits that would reflect the local Hg
situation better than the current national guidelines.
Growth rate may also be used as a preliminary
indicator to target hot spots where the regular
consumption of walleyes and northern pike may be
unsafe. Moreover, decision makers in the field of
fisheries management, being aware that fish growth
rates may be accelerated through a certain degree of
fishing pressure and control of lake ecology and
watershed management, may use the findings reported
in this paper to properly control predatory fish growth
rates in order to enhance Hg biodilution in fish of
edible interest for sport or subsistence fishers.
Acknowledgments
This research project was financially supported by
the Natural Sciences and Engineering Research
Council of Canada through a Research Network grant
administered by the Collaborative Mercury Research
Network. Our work would have been impossible
without the support from Hydro-Quebec, the Quebec
Ministry for Sustainable Development, Environment
and Parks, as well as from the Quebec Ministry for
Natural Resources and Fauna, all of whom coopera-
tively shared fish information, analysis, and occasion-
ally field operations with us. The authors would like to
thank the anonymous reviewers of this manuscript for
their helpful comments and constructive suggestions.
References
Babaluk, J. A., and J. S. Campbell. 1987. Preliminary results
of tetracycline labeling for validating annual growth
increments in opercula of walleye. North American
Journal of Fisheries Management 7:138–141.
Babaluk, J. A., and J. F. Craig. 1990. Tetracycline marking
studies with pike, Esox lucius. Aquaculture and Fisheries
Management 21:307–315.
Babaluk, J. A., J. F. Craig, and J. S. Campbell. 1993. Age and
growth estimation of walleye, Stizostedion vitreus, using
opercula. Canadian Manuscript Report of Fisheries and
Aquatic Sciences 2183.
Bodaly, R. A., J. W. M. Rudd, R. J. P. Fudge, and C. A. Kelly.
1993. Mercury concentrations in fish related to size of
remote Canadian Shield lakes. Canadian Journal of
Fisheries and Aquatic Sciences 50:980–987.
Campbell, J. S., and J. A. Babaluk. 1979. Age determination
of walleye, Stizostedion vitreus vitreus (Mitchill), based
on the examination of eight different structures. Canada
Fisheries and Marine Service Technical Report 849.
Casselman, J. M. 1974. Analysis of hard tissue of pike Esoxlucius L. with special reference to age and growth. Pages
13–27 in T. B. Bagenal, editor. Proceedings of an
international symposium on the ageing of fish. Unwin
Brothers, Farnham, UK.
Casselman, J. M. 1978. Effects of environmental factors on
growth, survival, activity, and exploitation of northern
pike. Pages 114–128 in R. L. Kendall, editor. Selected
coolwater fishes of North America. American Fisheries
Society, Special Publication 11, Bethesda, Maryland.
Casselman, J. M. 1987. Determination of age and growth. Pages
209–242 in A. H. Weatherley and H. S. Gill, editors. The
biology of fish growth. Academic Press, London.
CEAEQ (Centre d’Expertise en Analyse Environnementale du
Quebec). 2003. Determination du mercure dans les tissus
biologiques et les sediments: methode automatisee par
photometrie UV et par formation de vapeur. [Measure-
ments of mercury in biological samples and in sediments
by thermal breakdown: UV photometry assay.] Ministere
de l’Environnement du Quebec, MA 207-Hg 1.0, Quebec.
CEHQ (Centre d’Expertise Hydrique du Quebec). 2005.
Limites des bassins hydrographiques a l’echelle
1/250 000. [Hydrographic watershed limits at
1:250,000 scale.] (CD-ROM). Ministere du Developpe-
ment Durable, de l’Environnement et des Parcs, Quebec.
Clapp, D. F., and D. H. Wahl. 1996. Comparison of food
consumption, growth, and metabolism among muskel-
lunge: an investigation of population differentiation.
Transactions of the American Fisheries Society 125:402–
410.
Claramunt, R. M., and D. H. Wahl. 2000. The effects of abiotic
and biotic factors in determining larval fish growth rates: a
comparison across species and reservoirs. Transactions of
the American Fisheries Society 129:835–851.
Colby, P. J., and S. J. Nepszy. 1981. Variation among stocks
MERCURY AND FISH GROWTH IN QUEBEC LAKES 1233
of walleye (Stizostedion vitreum vitreum): management
implications. Canadian Journal of Fisheries and Aquatic
Sciences 38:1814–1831.
Cope, W. G., J. G. Wiener, and R. G. Rada. 1990. Mercury
accumulation in yellow perch in Wisconsin seepage
lakes: relation to lake characteristics. Environmental
Toxicology and Chemistry 9:931–940.
de Freitas, A. S. W., S. U. Qadri, and B. E. Case. 1974.
Origins and fate of mercury compounds in fish. Pages
31–36 in Proceedings of the International Conference on
Transport of Persistent Chemicals in Aquatic Ecosys-
tems. National Research Council of Canada, Ottawa.
Doire, J. 2003. Influence de peches intensives sur la
croissance et l’alimentation des poissons de lacs naturels
du nord du Quebec. [Influence of intensive fishing on the
growth and feeding of fish in natural lakes in northern
Quebec.] Master’s thesis. Universite du Quebec a
Montreal, Montreal.
Downing, J. A., Y. T. Prairie, J. J. Cole, C. M. Duarte, L. J.
Tranvik, R. G. Striegl, W. H. McDowell, P. Kortelainen,
N. F. Caraco, J. M. Melack, and J. J. Middelburg. 2006.
The global abundance and size distribution of lakes,
ponds, and impoundments. Limnology and Oceanogra-
phy 51:2388–2397.
Doyon, J. F., R. Schetagne, and R. Verdon. 1998. Different
mercury bioaccumulation rates between sympatric pop-
ulations of dwarf and normal lake whitefish (Coregonusclupeaformis) in the La Grande complex watershed,
James Bay, Quebec. Biogeochemistry 40:203–216.
Driscoll, C. T., C. Yan, C. L. Schoeld, R. Munson, and J.
Holsapple. 1994. The mercury cycle and fish in the
Adirondack lakes. Environmental Toxicology and Chem-
istry 28:136–143.
Dubois, A. 1967. Age et croissance du touladi (Salvelinusnamaycush) du lac Mistassini, Quebec. [Age and growth
of lake trout (Salvelinus namaycush) in Lake Mistassini,
Quebec.] Master’s thesis. Universite Laval Quebec,
Quebec.
EPRI (Electric Power Research Institute). 2002. Dynamic
mercury cycling model for Windows 98/NT/2000/
XPTM: a model for mercury cycling in lakes, D-DMCM
version 2.0. EPRI, Palo Alto, California.
EPRI (Electric Power Research Institute). 2003. Factors
affecting the predicted response of fish mercury concen-
trations to changes in mercury loading. EPRI, Final
Report 1005521, Palo Alto, California.
Essington, T. E., and J. N. Houser. 2003. The effect of whole-
lake nutrient enrichment on mercury concentration in
age-1 yellow perch. Transactions of the American
Fisheries Society 132:57–68.
Evans, M. S., W. L. Lockhart, L. Doetzel, G. Low, D. Muir,
K. Kidd, G. Stephens, and J. Delaronde. 2005. Elevated
mercury concentrations in fish in lakes in the Mackenzie
River basin: the role of physical, chemical, and biological
factors. Science of the Total Environment 351–352:479–
500.
Fortin, R., P. Dumont, and S. Guenette. 1996. Determinants of
growth and body condition of lake sturgeon (Acipenserfulvescens). Canadian Journal of Fisheries and Aquatic
Sciences 53:1150–1156.
Froese, R., and D. Pauly, editors. 2000. FishBase 2000:
concepts, design, and data sources. ICLARM (Interna-
tional Center for Living Aquatic Resources Manage-
ment), Manila, Philippines.
Galarowicz, T. L., and D. H. Wahl. 2003. Differences in
growth, consumption, and metabolism among walleye
from different latitudes. Transactions of the American
Fisheries Society 132:425–437.
Gorski, P. R., L. B. Cleckner, J. P. Hurley, M. E. Sierszen, and
D. E. Armstrong. 2003. Factors affecting enhanced
mercury bioaccumulation in inland lakes of Isle Royale
National Park, USA. Science of the Total Environment
304:327–348.
Gothberg, A. 1983. Intensive fishing: a way to reduce the
mercury level in fish. Ambio 12:259–261.
Harris, R. C., and R. A. Bodaly. 1998. Temperature, growth,
and dietary effects on fish mercury dynamics in two
Ontario lakes. Biogeochemistry 40:175–187.
Hopkins, K. D. 1992. Reporting fish growth: a review of the
basics. Journal of the World Aquaculture Society
23:173–179.
Jackson, L. J., and D. E. Schindler. 1996. Field estimates of
net trophic transfer of PCBs from prey fishes to Lake
Michigan salmonids. Environmental Sciences and Tech-
nology 30:1861–1865.
Jernelov, A., and H. Lann. 1973. Studies in Sweden on
feasibility of some methods for restoration of mercury-
contaminated bodies of water. Environmental Toxicolo-
gy and Chemistry 7:712–718.
Kainz, M., M. Lucotte, and C. C. Parrish. 2003. Relationships
between organic matter composition and methyl mercury
content of offshore and carbon-rich littoral sediments in
an oligotrophic lake. Canadian Journal of Fisheries and
Aquatic Sciences 60:888–896.
Kamman, N. C., N. M. Burgess, C. T. Driscoll, H. A.
Simonin, W. Goodale, J. Linehan, R. Estabrook, M.
Hutcheson, A. Major, A. M. Scheuhammer, and D. A.
Scruton. 2005. Mercury in freshwater fish of northeast
North America: a geographic perspective based on fish
tissue monitoring databases. Ecotoxicology 14:163–180.
Laine, A. O., W. T. Momot, and P. Ryan. 1991. Accuracy of
using scales and cleithra for aging northern pike from an
oligotrophic Ontario lake. North American Journal of
Fisheries Management 11:220–225.
Lapointe, M., and A. M. Clement. 1985. Determination de
l’age de touladis du lac des Trente et Un Milles et du lac
Simon Quebec. [Age determination of lake trout
specimens (Salvelinus namaycush) in Lake Trente et
Un Milles and Lake Simon, Quebec.] Ministere du
Loisir, de la Chasse et de la Peche, Hull, Quebec.
Lavigne, M. 2007. Le taux de croissance des poissons
predateurs: un indicateur pour integrer les facteurs
biologiques et environnementaux controlant les concen-
trations de mercure dans leur chair. [Growth rate of
piscivorous fish: a proxy to integrate biological and
environmental factors controlling mercury concentrations
in fish flesh.] Master’s thesis. Universite du Quebec a
Montreal, Montreal.
Legendre, P. 2001. Regression de modele II: guide. [Model-II
regression: user’s guide.] Departement de Sciences
Biologiques, Universite de Montreal, Montreal. Avail-
able: bio.umontreal.ca/casgrain/en/telecharger/index.
html#modeIIregression. (June 2010).
1234 LAVIGNE ET AL.
Legendre, P., and L. Legendre. 1998. Numerical ecology, 2nd
English edition. Elsevier, Amsterdam.
Lockhart, W. L., G. A. Stern, G. Low, M. Hendzel, G. Boila,
P. Roach, M. S. Evans, B. N. Billeck, J. DeLaronde,
S. K. Friesen, B. Kidd, S. Atkins, D. C. G. Muir, M.
Stoddart, G. Stephens, S. Stephenson, S. Harbicht, N.
Snowshoe, B. Grey, S. Thompson, and N. DeGraff.
2005. A history of total mercury in edible muscle of fish
from lakes in northern Canada. Science of the Total
Environment 351–352:427–463.
Lucotte, M., R. Schetagne, N. Therien, C. Langlois, and A.
Tremblay. 1999. Mercury in the biogeochemical cycle:
natural environments and hydroelectric reservoirs of
northern Quebec. Springer- Verlag, New York.
Marshall, I. B., and P. H. Schut. 1999. A national ecological
framework for Canada. Agriculture and Agri-food
Canada. Available: sis.agr.gc.ca/cansis/nsdb/ecostrat/
intro.html. (February 2007).
Montgomery, S., M. Lucotte, and L. Cournoyer. 2000. The
use of stable carbon isotopes to evaluate the importance
of fine suspended particulate matter in the transfer of
methylmercury to biota in boreal flooded environments.
Science of the Total Environment 261:33–41.
Munkittrick, K. R., and D. G. Dixon. 1989. A holistic approach
to ecosystem health assessment using fish population
characteristics. Hydrobiologia 188–189:123–135.
Norstrom, R. J., A. E. McKinnon, and A. S. W. de Freitas.
1976. A bioenergetics-based model for pollutant accu-
mulation by fish: simulation of PCB and methylmercury
residue levels in Ottawa River yellow perch. Journal of
the Fisheries Research Board of Canada 33:248–267.
O’Connor, D. V., D. V. Rottiers, and W. H. Berlin. 1981.
Food consumption, growth rate, conversion efficiency,
and proximate composition of yearling lake trout. U.S.
Fish and Wildlife Service, Great Lakes Fishery Labora-
tory, Administrative Report 81-5, Ann Arbor, Michigan.
Olsson, M. 1976. Mercury level as a function of size and age
in northern pike, one and five years after the mercury ban
in Sweden. Ambio 5:73–76.
Parkman, H., and M. Meili. 1993. Mercury in macroinverte-
brates from Swedish forest lakes: influence of lake type,
habitat, life cycle, and food quality. Canadian Journal of
Fisheries and Aquatic Sciences 50:521–534.
Pauly, D. 1978. A preliminary compilation of fish length
growth parameters. Berichte des Instituts fur Meere-
skunde an der Universitat Kiel, Number 55, Kiel,
Germany.
Pepin, S., and F. Levesque. 1985. Technique de determination
de l’age des dores applicables aux populations de cette
espece au Quebec. [Walleye age determination technique
applicable to fish populations of that species in Quebec.]
Ministere du Loisir, de la Chasse et de la Peche, Quebec.
Pichet, P., K. Morrison, I. Rheault, and A. Tremblay. 1999.
Analysis of total mercury and methylmercury in
environmental samples. Pages 41–52 in M. Lucotte, R.
Schetagne, N. Therien, C. Langlois, and A. Tremblay,
editors. Mercury in the biogeochemical cycle: natural
environments and hydroelectric reservoirs of northern
Quebec. Springer-Verlag, Berlin.
Rask, M., M. Jarvinen, K. Kuoppamaki, and H. Poysa. 1996.
Limnological responses to the collapse of the perch
population in a small lake. Annales Zoologici Fenneci
33:517–524.
Ricker, W. E. 1980. Calcul et interpretation des statistiques
biologiques des populations de poissons. [Computation
and interpretation of biological statistics of fish popula-
tions.] Fisheries Research Board of Canada Bulletin 191F.
Roue-Le Gall, A., M. Lucotte, J. Carreau, R. Canuel, and E.
Garcia. 2005. Development of an ecosystem sensitivity
model regarding mercury levels in fish using a preference
modeling methodology: application to the Canadian
boreal system. Environmental Toxicology and Chemistry
39:9412–9423.
SAS Institute. 2002. JMP version 5.0. SAS Institute, Cary,
North Carolina.
Schetagne, R., and R. Verdon. 1999. Mercury in fish of
natural lakes of northern Quebec. Pages 115–130 in M.
Lucotte, R. Schetagne, N. Therien, C. Langlois, and A.
Tremblay, editors. Mercury in the biogeochemical cycle:
natural environments and hydroelectric reservoirs of
northern Quebec. Springer-Verlag, Berlin.
Scott, D. P., and F. A. J. Armstrong. 1972. Mercury
concentration in relation to size in several species of
freshwater fishes from Manitoba and northwestern
Ontario. Journal of the Fisheries Research Board of
Canada 29:1685–1690.
Scott, W. B., and E. J. Crossman. 1974. Poissons d’eau douce
du Canada. [Freshwater fishes of Canada.] Office des
Recherches sur les Pecheries du Canada, Bulletin 184,
Ottawa.
Simoneau, M., M. Lucotte, S. Garceau, and D. Laliberte.
2005. Fish growth rates modulate mercury concentrations
in walleye (Sander vitreus) from eastern Canadian lakes.
Environmental Research 98:73–82.
SPSS. 2004. SPSS version 13.0. SPSS, Chicago.
Stafford, C. P., B. Hansen, and J. A. Stanford. 2004. Mercury
in fishes and their diet items from Flathead Lake,
Montana. Transactions of the American Fisheries Society
133:349–357.
Stafford, C. P., and T. A. Haines. 2001. Mercury contamination
and growth rate in two piscivorous populations. Environ-
mental Toxicology and Chemistry 20:2099–2101.
Surette, C. 2005. Effets des peches intensives sur les
concentrations de mercure dans les poissons de lacs
naturels du nord quebecois. [Influence of intensive
fishing on mercury concentrations in fish from natural
lakes of northern Quebec.] Doctoral dissertation. Uni-
versite du Quebec a Montreal, Montreal.
Surette, C., M. Lucotte, and A. Tremblay. 2005. Influence of
intensive fishing on the partitioning of mercury and
methylmercury in three lakes of northern Quebec.
Science of the Total Environment 368:248–261.
Tremblay, G., J. F. Doyon, and R. Schetagne. 1996. Reseau de
suivi environnemental du complexe La Grande. Demarche
methodologique relative au suivi des teneurs en mercure
des poissons. [The Grande River environmental monitoring
network. Methodological aspects relative to the monitoring
of fish mercury concentrations.] GENIVAR and Hydro-
Quebec, Joint Technical Report, Montreal.
Tremblay, G., P. Legendre, J. F. Doyon, R. Verdon, and R.
Schetagne. 1998. The use of polynomial regression
analysis with indicator variables for interpretation of
mercury in fish data. Biogeochemistry 40:189–201.
MERCURY AND FISH GROWTH IN QUEBEC LAKES 1235
Verta, M. 1990. Changes in fish mercury concentrations in an
intensively fished lake. Canadian Journal of Fisheries and
Aquatic Sciences 47:1888–1897.
Weatherley, A. H., and H. S. Gill. 1987. The biology of fish
growth. Academic Press, London.
Wiener, J. G., D. P Krabbenhoft, G. H. Heinz, and A. M.
Scheuhammer. 2003. Ecotoxicology of mercury. Pages
409–464 in D. J. Hoffman, B. A. Rattner, G. A.
Burton Jr., and J. Cairns Jr., editors. Handbook of
ecotoxicology, 2nd edition. CRC Press, Boca Raton,
Florida.
Wren, C. D., and H. R. MacCrimmon. 1986. Comparative
bioaccumulation of mercury in two adjacent freshwater
ecosystems. Water Research 20:763–770.
Appendix: Basic Data
TABLE A.1.—Basic data and sources for lakes selected to examine the relationship between mercury concentrations in fish
muscle and fish growth rates, including 54 walleye (Wal; 5,207 individuals), 52 northern pike (NP; 2,419), and 35 lake trout (LT;
1,478) populations (lakes).
Lake
RegionaNumber of
angling events Year
Number of fish Coordinates
Data sourcebName Number Wal NP LT Latitude (N) Longitude (W)
Achikunipis 1 09 1 1990 15 56844 007 00 73840 015 00 1Achiyaskunapiskuch 2 09 1 1990 24 19 52831 000 00 75812 000 00 1Alegrain 3 09 1 1990 16 56821 000 00 73817 000 00 1Amichinatwayach 4 09 1 1989 44 54817 000 00 73805 000 00 1Anonyme A 5 09 1 1988–1990 24 13 54844 050 00 76802 003 00 1Archipel Saint-Pierre 6 00 1 2003 175 46805 001 00 73801 015 00 3Au Brochet 7 09 1 1992 23 49836 046 00 69835 033 00 1Au Goeland 8 08 2 1988–1990 79 53 49847 000 00 76848 000 00 1Aux Cedres 9 07 1 1992 30 48850 000 00 69807 000 00 1Aux Dores 10 08 1 2001 66 23 58 49852 000 00 74820 000 00 3Aux Sangsues 11 04 1 2001 202 46828 015 00 77856 000 00 3–4Bienville 12 09 1 1989 46 55805 000 00 72840 000 00 1Bob-Grant 13 05 1 1990 20 47845 000 00 73831 000 00 1Brehrad 14 05 1 1991 42 47852 000 00 73849 000 00 1Bruce 15 09 1 1988 22 53811 000 00 77855 000 00 1Cecile 16 05 1 1990 24 49807 005 00 74802 022 00 1Chaumont 17 09 1 1987 10 53823 000 00 70835 000 00 1Chibougamau 18 08 4 1999–2002 145 53 110 49850 000 00 74815 000 00 1Chicobi 19 08 1 2004 114 48851 043 00 78831 000 00 3Clarkie 20 09 1 1990 30 18 52814 000 00 75830 000 00 1Conn 21 09 1 1988 30 52835 000 00 77836 000 00 1Corvette 22 09 1 1989 16 53825 000 00 74803 000 00 1Cosnier 23 08 1 2004 251 50854 000 00 72843 000 00 3Coutaceau 24 09 1 1978 25 26 53833 000 00 76835 000 00 2Craven 25 09 1 1988 23 54820 000 00 76856 000 00 1Dana 26 08 3 1979–1991 32 43 50853 000 00 77820 000 00 1David 27 04 1 2004 10 46815 030 00 74852 039 00 3De Ganne 28 09 1 1990 15 55838 000 00 76812 000 00 1De l’Est 29 02 1 2004 17 47811 011 00 69833 041 00 3Des Vœux 31 09 7 1987–1999 53 268 53856 015 00 72837 035 00 1Deschamps 32 08 1 1990 25 13 48835 012 00 75831 018 00 1Desjardins 33 04 1 2001 223 47817 000 00 78814 000 00 3–4Dissimieux 34 07 1 1992 23 49852 000 00 60848 000 00 1Du Tast 35 08 1 1979 12 51800 000 00 77822 000 00 1Duparquet 36 08 2 2002–2003 258 48828 009 00 79816 019 00 3–4Duxbury 37 09 1 1988 29 52827 000 00 77828 000 00 1Earhart 38 04 1 2004 16 46805 000 00 75826 000 00 3Eau Claire 39 08 2 1989–1990 63 50810 000 00 75810 000 00 1En Cœur 40 05 1 2004 32 45858 005 00 74800 041 00 3Evans 41 08 3 1979–1991 73 95 50855 000 00 77800 000 00 1–2Faguy 42 05 1 1990 28 22 48833 000 00 74805 000 00 1Fregate 43 09 1 1989 16 53812 000 00 74845 000 00 1Fressel 44 09 1 1989 24 55827 000 00 75812 000 00 1Gaillarbois 45 07 1 1997 29 52800 000 00 67827 000 00 1Gaotanaga 46 04 1 1990 14 11 47838 000 00 77835 022 00 1Giffard 47 08 3 1979–1991 20 50 51808 000 00 76855 000 00 1–2Grasset 48 08 1 1990 29 28 49856 033 00 78809 043 00 1Hazeur 49 10 7 1987–1999 202 54858 000 00 69812 000 00 1Herve 50 09 1 1987 10 54827 000 00 71814 000 00 1
1236 LAVIGNE ET AL.
TABLE A.1.—Continued.
Lake
RegionaNumber of
angling events Year
Number of fish Coordinates
Data sourcebName Number Wal NP LT Latitude (N) Longitude (W)
Jolliet 51 08 2 1979–1990 27 44 51833000 00 76854 000 00 1Julian 52 09 1 1988 26 54826000 00 77855 000 00 1Kawayapiskach 53 09 1 1989 28 52837000 00 78830 000 00 1Kowskatehkakmow 54 09 6 1990–2000 179 110 53827000 00 77827 000 00 1Largarde 55 04 1 2001 211 46816038 00 77826 053 00 3, 4Lamain 56 09 1 1990 17 55854000 00 75806 000 00 1Laraire 57 09 1 1989 15 55827000 00 72857 000 00 1Le Royer 58 08 1 2004 30 50 49835030 00 74827 030 00 4Lessard 59 08 1 1988 31 49832000 00 75839 000 00 1Letemplier 60 07 1 1988 17 49827022 00 68847 001 00 1Loups Marins 61 10 2 1988–1989 18 56830030 00 65803 030 00 1Maicasagi 62 08 1 1979 30 49857020 00 76833 012 00 2Malartic 63 08 2 2002–2003 320 48815052 00 78806 018 00 3, 4Matagami 64 08 3 1979–1991 62 121 49853000 00 77830 000 00 1, 2Maulnier 65 09 1 1990 19 56836052 00 73859 025 00 1McNab 66 09 1 1988 15 52853000 00 77827 000 00 1Megiscane 67 08 1 2004 228 48835007 00 75851 057 00 3, 4Mesgouez 68 08 2 1979–1990 27 26 50825039 00 75805 028 00 1, 2Midway 69 07 1 1992 30 52828000 00 67802 000 00 1Morpain 70 09 1 1989 30 55800000 00 74818 000 00 1Mureau 71 05 2 1989–1990 22 54855000 00 73813 000 00 1Nemenjiche 72 08 1 2004 26 34 49822000 00 74826 000 00 4Nemiscau amont 73 08 1 1979 16 51826000 00 76843 000 00 2Nemiscau 74 08 1 1990–1991 44 20 51826000 00 76843 000 00 1Obatogamau 75 08 2 2001–2002 81 62 49835037 00 74827 020 00 3Ouescapis 76 08 1 1990 27 24 50814000 00 77800 000 00 1Pamigamachi 77 09 1 1988 19 54810000 00 77828 000 00 1Petit Loups Marins 78 09 2 1989–1990 44 56807000 00 73815 000 00 1Pikutamaw 79 09 9 1984–2000 281 245 52834000 00 77805 000 00 1, 2Poncheville 80 08 2 1979–1991 43 58 50810000 00 76855 000 00 1, 2Preissac 81 04 1 2002 634 48820000 00 78820 000 00 3, 4Pusticamica 82 08 2 1988–1989 45 27 49821000 00 76823 000 00 1Quenonisca 83 08 3 1979–1991 29 53 50836000 00 76833 000 00 1, 2Rinfret 84 05 1 1990 25 48824000 00 73840 000 00 1Rodayer 85 08 1 1990 19 20 50852000 00 77842 000 00 1Roggan 86 09 1 1988 22 54808000 00 77849 000 00 1Saint-Francois 87 03 1 2004 34 114 45809025 00 74822 022 00 3Saint-Pierre 88 00 1 2002 495 320 46816010 00 72837 048 00 3Sans nom 89 07 1 1996 11 51808000 00 67800 000 00 1Serigny 90 10 4 1993–1999 17 138 55822000 00 69838 000 00 1Six-Milles 91 04 1 2004 125 46826037 00 78801 050 00 3, 4Soscumica 92 08 1 1990 24 53 50815000 00 77827 000 00 2Theodat 93 08 1 1990 24 12 50855000 00 76810 000 00 1Tilly 94 09 2 1987–1989 39 53855000 00 73858 000 00 1Vaulezar 95 09 1 1993 30 19 54833000 00 71850 000 00 1Vieux-Comptoir 96 09 1 1988 14 52847000 00 77833 000 00 1Village nord 97 09 1 1990 21 52812042 00 75817 050 00 1Waconichi 98 08 2 2000–2001 21 25 30 50808000 00 74800 000 00 3Waswanipi 99 08 4 1979–1990 111 91 49834000 00 76829 000 00 1, 2Yasinski 100 09 1 1988 24 53816041 00 77834 059 00 1
a Hydrographic region as defined by the Ministry of Sustainable Development, Environment, and Parks (Figure 1).b (1) Hydro Quebec, (2) Bay James Energy Society, (3) Ministry of Sustainable Development, Environment, and Parks, and (4) Collaborative
Mercury Research Network.
MERCURY AND FISH GROWTH IN QUEBEC LAKES 1237