Latitudinal variation in growth and otolith-inferred field metabolic rates of Canadian young-of-the-year Arctic charr

Latitudinal variation in growth andotolith-inferred field metabolic rates of Canadianyoung-of-the-year Arctic charrR. Niloshini Sinnatamby1, J. Brian Dempson2, James D. Reist3, Michael Power11Department of Biology, University of Waterloo, Waterloo, ON Canada2Fisheries & Oceans Canada, Science Branch, St. John’s, NL Canada3Fisheries & Oceans Canada, Central & Arctic Region, Winnipeg, MB Canada

Accepted for publication June 30, 2014

Abstract – Countergradient variation (CGV) is defined as genetic variation that counteracts the negative influencesof the physical environment, minimising phenotypic variability along an environmental gradient. CGV is thought tohave relevance in predicting the response of organisms to climate variability and change. To test the hypothesis thatgrowth rate increased with latitude, consistent with CGV, young-of-the-year (YOY) Arctic charr, Salvelinus alpinus,were examined along a ~27° latitudinal gradient in central and eastern Canada. Growth rates were estimated fromfork lengths standardised by the thermal opportunity for growth based on experienced water temperatures derivedusing otolith oxygen stable isotopes. Results demonstrated patterns consistent with CGV, where northernpopulations demonstrated faster growth rates. A secondary aim was to test for similar geographical patterns inotolith-inferred metabolic rates, which reflect the energetic costs of standard metabolic rate (SMR) and otherprocesses such as feeding, locomotion, thermoregulation, reproduction and growth. Results demonstrated asignificant, positive relationship between otolith-inferred metabolic rate and latitude, which may reflect an increasein one, or a combination, of the above-noted physiological processes. The similar latitudinal pattern in growth andotolith-inferred metabolic rates suggests greater intake of food per unit of time by northern fish. The phenotypicvariation in physiological traits observed here demonstrates the significant adaptability of Arctic charr to differentthermal regimes with different growing season lengths. Determining the relative contributions of phenotypicplasticity and genetic variation to the observed latitudinal variation will be critical to predicting the responses ofArctic charr to climate change more accurately.

Key words: countergradient variation; growth rate; metabolic rate; stable isotopes; Salvelinus alpinus

Introduction

Of the approximately 99 freshwater and/or anadro-mous fish species found in the Arctic, Arctic charr,Salvelinus alpinus, is the only species with a mostlycomplete circumpolar distribution (Reist et al.2006a). Consistent with its northerly distribution, Arc-tic charr is also among the most cold-adapted fresh-water and/or anadromous fish species in the world(Johnson 1980; Baroudy & Elliott 1994a). Despitetheir association with cold environments, Arctic charrexhibit the widest natural distribution of all salmonids(Maitland 1995), extending from northern New

England (43°N) to the north coast of Ellesmere Island(83°N) in North America (Power et al. 2008). Under-standing whether, and how, Arctic charr populationsvary along a latitudinal gradient will therefore be vitalto predicting their potential responses to climatechange (Reist et al. 2006a,b). Furthermore, characte-rising the phenotypic and/or genotypic variabilityexhibited by Arctic charr over its geographic rangemay help to illustrate the adaptive capacity of the spe-cies and ultimately improve understanding of its sen-sitivity to climate change (Williams et al. 2008).Environmental change has been suggested to have

four possible outcomes for a species: (i) extinction or

Correspondence: R. N. Sinnatamby, Department of Biology, University of Waterloo, Waterloo, ON, Canada N2L 3G1. E-mail: [email protected]

doi: 10.1111/eff.12166 1

Ecology of Freshwater Fish 2014 � 2014 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd

ECOLOGY OFFRESHWATER FISH

extirpation, (ii) a shift in distributional range, (iii)genetic adaptation and/or (iv) compensation via phe-notypic plasticity (Fuller et al. 2010). Presently, theextent to which phenotypic plasticity contributes to aspecies’ niche utilisation and geographic distributionis not well understood (Ellers & Stuefer 2010; Fulleret al. 2010), even though phenotypic plasticity isthought to be the most likely initial response ofan individual organism to environmental change(Bradshaw & Holzapfel 2008) and is considered tobe more important than genetic adaptation in terms ofshort-term responses to climate change (Williamset al. 2008). The study of physiological traits overlarge geographic scales (i.e. macrophysiology) mayillustrate the physiological capacity of a species torespond to climate change (Chown & Gaston 2008;Williams et al. 2008) and may, by analogy, allowinference of likely responses to temporal climatechange (Baumann & Conover 2011).Countergradient variation (CGV), defined as

genetic variation that counteracts the negative influ-ences of the physical environment (e.g. lower temper-atures, shorter growing seasons and lower foodsupply), has been used to explain the phenotypic sim-ilarity often observed in traits across a latitudinalcline (Conover & Schultz 1995). The CGV growthrate hypothesis states that at similar temperatures,individuals from high-latitude environments willgrow faster than conspecifics from low-latitude envi-ronments (Levins 1969). CGV in growth rate hasbeen demonstrated in several taxa but is most com-monly observed in fish (Conover et al. 2009).Despite increasing numbers of examples of CGV(Conover et al. 2009), overall support for the CGVgrowth rate hypothesis remains uncertain and hasbeen mixed in salmonid species (e.g. Jensen et al.2000; Jonsson et al. 2001; Finstad et al. 2004; Nicola& Almod�ovar 2004; Forseth et al. 2009; Parra et al.2009; Logez & Pont 2011). A previous study on Arc-tic charr also demonstrated mixed results, reportinggrowth patterns consistent with CGV in Arctic charraged 6–12 years from ‘normal’ lacustrine populationsin eastern North America, but rather ambivalent sup-port for the hypothesis in anadromous populationsand no support for CGV in relictual ‘dwarf’ lacus-trine populations (Chavarie et al. 2010).Evidence of CGV in growth rate can be demon-

strated using common garden experiments (e.g.Conover & Present 1990; Nicieza et al. 1994), aswell as field-based studies on natural populations ifenvironmental differences among sites can be accountedfor (Conover et al. 2006). In fact, a comparative ana-lysis between experimental and field-based methodsto assess CGV in growth, conducted on striped bass(Morone saxatilis), demonstrated a high degree ofsimilarity in the growth-latitude slopes estimated

using both methods (Rypel 2014). Thermal opportu-nity for growth (TOG) and growing degree days(GDD), similar metrics based on air temperature data,have been used in previous studies on natural popula-tions to account for environmental differences amongsites and standardise growth (Power & McKinley1997; Chavarie et al. 2010; Rypel 2012). In the placeof air temperatures, otolith-inferred water tempera-tures based on oxygen stable isotopes (d18O) (e.g.Devereux 1967; Høie et al. 2004; Godiksen et al.2011) may be more representative of the temperaturesactually experienced by the fish.A number of broad-scale studies have also been con-

ducted on the metabolic rates of ectotherms (e.g. Scho-lander et al. 1953; Bullock 1955; Wohlschlag 1960;Lindgren & Laurila 2009; Buckley et al. 2012). Physi-ologists typically measure standard metabolic rate(SMR), the metabolic rate of an inactive organism inthe laboratory, whereas field metabolic rate, the meta-bolic rate of free-living animals, reflects the averagedaily energy expenditure of an organism and incorpo-rates the energetic costs of SMR, thermoregulation,locomotion, feeding, reproduction and growth (Nagy1987). As the relative proportion of otolith carbonderived from its two isotopically distinct sources(ambient dissolved inorganic carbon and diet) is depen-dent on individual metabolic rate, otolith carbon stableisotope (d13C) values can be used to infer relative met-abolic rates (e.g. Schwarcz et al. 1998; Wurster & Patt-erson 2003; Wurster et al. 2005; Shephard et al.2007). Further, because otolith d13C values reflect lev-els of metabolically derived d13C in blood plasma(Sherwood & Rose 2003) and are metabolically inertonce precipitated into the otolith (Campana 1999), theotolith-inferred metabolic rate reflects the average fieldmetabolic rate of the organism over its lifetime.Here, we use information on differences among

young-of-the-year (YOY) Arctic charr growth ratesadjusted for the thermal opportunity for growth (sensuConover 1990) from fish sampled over a 27° latitudi-nal gradient in eastern and central Canada. The dataare used to test the hypothesis that growth ratesincrease with latitude in a manner consistent with thecountergradient variation hypothesis. As countergradi-ent variation is most commonly observed in physio-logical traits such as growth, metabolism, activity anddigestion (Conover et al. 2009), a secondary aim ofthe study was to estimate otolith-inferred field meta-bolic rates, which were obtainable for a subset of sam-ple locations, to assess whether a similar geographicmetabolic pattern could be observed.

Methods

Samples of YOY Arctic charr were obtained eitherdirectly by the authors or in conjunction with other

2

Sinnatamby et al.

sampling programmes from 23 sites across a latitudi-nal gradient between 56°N and 83°N in eastern andcentral Canada (Fig. 1, Table 1). Sampling occurredbetween July 20 and August 22 over a 4-year periodand included both freshwater resident and anadro-

mous populations (Table 2). Samples were capturedusing electrofishers and/or dip nets in nearshorefreshwater environments in water depths of <1 m.Whole otoliths were examined under a dissectingmicroscope to confirm that all individuals were YOY(Baker & Timmons 1991). Water samples were col-lected for d18O analysis (50 ml) coincident with allfish sampling events. Water samples were obtainedfor d13C analysis (250 ml) for 14 lacustrine sites(Table 3).Water samples were shipped to the University

of Waterloo Environmental Isotope Laboratory(UW-EIL) for analyses. Water d18O values weredetermined following procedures outlined in Epstein& Mayeda (1953) and Fritz et al. (1987) using anIsoPrime multiflow isotope ratio mass spectrometer(GV Instruments/Micromass, Manchester, UK). Resultswere corrected based on EIL-12 (sealed container oflaboratory water) calibrated to the international refer-ence materials for VSMOW (Vienna Standard MeanOcean Water) and SLAP (Standard Light AntarcticPrecipitation) from the International Atomic EnergyAgency (IAEA). Measured d18O isotope ratios wereconverted from VSMOW to VPDB (Vienna PeedeeBelemnite) following (Coplen et al. 1983):

d18OWaterðVPDBÞ ¼ 0:97002� d18OWaterðVSMOWÞ� 29:98

where d18OWater (VPDB) is the water d18O expressedwith respect to VPDB and d18OWater (VSMOW) is thewater d18O reported with respect to the internationalreporting scale for water (VSMOW). d13CDIC valueswere determined following procedures described in

Table 1. Sampled sites, corresponding map code and latitude andlongitude coordinates, by region.

Site Map Code Latitude (°N) Longitude (°W)

Labrador, NLKogluktokoluk Brook 1 56°170 62°090

Tom’s Pond 2 56°420 61°380

Ikarut River 3 58°090 63°030

Torr Bay Brook 4 58°270 62°490

Southwest Arm Brook 5 58°280 63°350

North Arm Brook 6 58°340 63°290

Baffin Island South, NUKatuqa, Pritzler Harbour 7 62°180 67°340

Qinngu, Blandford Bay 8 63°330 71°150

Qasigiat, Ptarmigan Fiord 9 64°380 66°190

Kent Peninsula Area, NULouis Lake 10 68°040 107°140

North Baffin Island, NULake 1 11 72°360 77°580

Cornwallis Island, NUResolute Lake 12 74°410 94°530

Char Lake 13 74°420 94°530

Small Lake 14 74°460 95°030

North Lake 15 74°460 95°050

Amituk Lake 16 75°030 93°530

Quttinirpaaq National Park, Ellesmere Island, NULake E 17 81°490 69°270

Lake Hazen 18 81°490 71°190

Lake G 19 81°490 69°160

Craig Lake 20 81°520 68°510

Lake C 21 82°050 68°240

Lake D 22 82°060 67°310

Clements Markham Lake 23 82°370 68°600

0 250 500 750

Kilometres

HudsonBay

Ellesm

ere Island

Ellesm

ere Island

Labrador Labrador

80°

60°

70°

80° 70° 60° 50° 40°90°100°

Baffin Island

Baffin Island

N

EW

S

456

Cornwallis IslandCornwallis Island

10

12

3

78

9

11

12-1516

2321 22

17-20

Kent PeninsulaKent Peninsula

Fig. 1. Map of sampling site locations.Numbers correspond to the map code asdescribed in Table 1.

3

Latitudinal variation in YOY Arctic charr

McCrea (1950) using a Gilson 222XL auto-samplercoupled to a MicroGas-IsoPrime Mass Spectrometer(GV Instruments/Micromass) with an analytical preci-sion of �0.2&.

All sampled YOY Arctic charr were frozen andshipped to the University of Waterloo or the Departmentof Fisheries and Oceans for processing. In mostcases, YOY Arctic charr fork lengths (FL, mm) were

Table 2. Mean and standard deviations of fork length (mm), otolith weight (mg), otolith-inferred temperature (°C) and TOG-adjusted growth rate (mm perdegree day) at each site. Year, sample size (n) and days used are also noted. Days used refers to the number of days between the beginning of the growingseason and capture date. TOG is the thermal opportunity for growth, which is estimated as mean otolith-inferred water temperature multiplied by the numberof days used.

Site Year n Mean FL (�SD) Mean otolith weight (�SD)Mean otolith-inferredtemperature (�SD) Days used Mean growth rate (�SD)

Labrador, NLKogluktokoluk Brook 2008 7 46.2 (5.1) 0.1778 (0.0379) 11.3 (1.5) 116 0.035 (0.004)Tom’s Pond 2008 6 40.3 (4.8) 0.1185 (0.0248) 11.6 (2.2) 108 0.032 (0.004)Ikarut River 2009 10 28.8 (2.8) 0.0480 (0.0109) 8.2 (2.8) 60 0.059 (0.006)Torr Bay Brook 2008 9 35.5 (3.0) 0.0689 (0.0148) 8.4 (0.8) 93 0.045 (0.004)Southwest Arm Brook 2007 8 39.8 (5.7) 0.0784 (0.0284) 6.4 (2.9) 61 0.102 (0.015)North Arm Brook 2008 9 31.2 (4.3) 0.0617 (0.0231) 4.8 (1.9) 59 0.110 (0.015)

Baffin Island South, NUKatuqa, Pritzler Harbour 2007 8 29.9 (4.7) 0.0468 (0.0157) 3.7 (0.8) 75 0.109 (0.017)Qinngu, Blandford Bay 2007 10 34.7 (3.5) 0.0811 (0.0127) 5.0 (2.5) 72 0.096 (0.010)Qasigiat, Ptarmigan Fiord 2007 10 30.6 (2.7) 0.0302 (0.0054) 2.8 (0.7) 58 0.186 (0.016)

Kent Peninsula Area, NULouis Lake 2008 10 31.6 (3.8) 0.0321 (0.0101) 2.9 (1.1) 51–54 0.203 (0.023)

North Baffin Island, NULake 1 2010 10 34.8 (4.4) 0.0688 (0.0167) 7.3 (1.6) 61 0.078 (0.010)

Cornwallis Island, NUResolute Lake 2008 5 27.2 (1.3) 0.0206 (0.0023) 2.4 (1.1) 41 0.272 (0.012)Char Lake 2008 8 29.1 (2.3) 0.0326 (0.0059) 4.0 (1.1) 44 0.166 (0.013)Small Lake 2008 4 26.5 (1.1) 0.0185 (0.0013) 4.4 (2.9) 45 0.134 (0.005)North Lake 2008 7 22.6 (1.9) 0.0146 (0.0039) 2.3 (1.5) 45 0.222 (0.019)Amituk Lake 2008 7 28.2 (1.8) 0.0348 (0.0050) 2.7 (1.6) 46 0.226 (0.014)

Quttinirpaaq National Park, NULake E 2007 9 40.2 (2.3) 0.0684 (0.0145) 5.8 (2.0) 44 0.158 (0.009)Lake Hazen 2007 8 34.4 (5.5) 0.0387 (0.0264) 4.9 (0.9) 41–51 0.158 (0.012)Lake G 2007 5 30.6 (2.9) 0.0385 (0.0110) 4.2 (1.1) 42 0.172 (0.016)Craig Lake 2008 11 39.5 (2.1) 0.0680 (0.0150) 4.7 (1.4) 52 0.161 (0.009)Lake C 2008 10 39.0 (3.6) 0.0597 (0.0183) 2.7 (0.4) 51 0.283 (0.026)Lake D 2007 7 26.7 (5.1) 0.0291 (0.0270) 2.6 (1.3) 43 0.236 (0.045)Clements Markham Lake 2008 8 39.8 (4.7) 0.0712 (0.0197) 0.9 (0.9) 54 0.292 (0.034)

Table 3. Mean and standard deviation (in brackets) of muscle tissue carbon stable isotopes (d13CM) and estimated per cent of metabolically derived otolithcarbon (M), and carbon stable isotope value of ambient water (d13CDIC) collected during fish sampling.

Site Year n Mean d13CM (&) d13CDIC (&) Mean M

Labrador, NLTom’s Pond 2008 6 �24.50 (1.21) �10.91 15.7 (4.0)

Kent Peninsula Area, NULouis Lake 2008 10 �27.18 (1.54) �10.08 19.4 (6.4)

Cornwallis Island NUResolute Lake 2008 5 �23.56 (0.84) �0.28 21.0 (3.4)Char Lake 2008 8 �23.93 (0.67) 1.47 35.3 (2.9)Small Lake 2008 4 �23.61 (0.20) 0.06 21.3 (2.0)North Lake 2008 7 �23.73 (1.50) �2.33 27.7 (2.5)Amituk Lake 2008 7 �23.68 (1.39) �0.80 39.4 (3.8)

Quttinirpaaq National Park, NULake E 2007 9 �27.25 (0.39) �0.02 39.2 (3.7)Lake Hazen 2007 8 �24.02 (2.73) �1.43 34.4 (7.9)Lake G 2007 5 �25.17 (0.33) 0.79 35.7 (3.6)Craig Lake 2008 11 �26.92 (0.68) �1.93 41.3 (5.6)Lake C 2008 10 �28.56 (1.64) �2.79 37.5 (2.4)Lake D 2007 7 �27.52 (1.12) �2.69 24.9 (6.9)Clements Markham Lake 2008 8 �25.46 (0.37) 0.08 45.1 (3.7)

4

Sinnatamby et al.

measured in the field prior to freezing. In instanceswhen samples were obtained indirectly throughcollaborative sampling, and field FL measurementswere not taken, a linear regression relationship esti-mated from available field and thawed laboratorymeasurements was used to estimate field FLs. In thelaboratory, sagittal otoliths were removed from thefish, cleaned with distilled water and dried at ambientroom temperature (~21 °C). Prior to use in stable iso-tope analysis, otoliths were weighed (Table 2) on ahigh-precision ultra-microbalance (Mettler-Toledo,GmbH, Greifensee, Switzerland). Whole otolithswere then placed in small reaction glass vials andheated to 350 °C for 1 h to remove potential organiccontamination under a continuous flow of purifiedhelium (Epstein et al. 1953; Guiguer et al. 2003).Vials were flushed with injected purified helium for30 s immediately following roasting. Phosphoric acidwas added to each sample by injection, and sampleswere warmed to 90 °C for 1 h prior to analysis. Atthe UW-EIL, evolved CO2 gas from acid digestionwas drawn off and auto-injected into an IsoPrimemass spectrometer (GV Instruments/Micromass) withan analytical precision of � 0.2&. Machine accuracyduring analysis was assessed with the analysis ofcommercially available international calcite standardsNBS-18 (Carbonitite CaCO3: d13CVPDB = �5.01&,d18OVPDB = �23.00&), NBS-19 (TS-limestoneCaCO3: d13CVPDB = 1.95&, d18OVPDB = �2.20&),EIL-21 (Pine Point CaCO3: d13CPDB = �0.19,d18OPDB = �15.68) and EIL-23 (Fisher CaCO3:d13CPDB = �3.42, d18OPDB = �13.45) with the setof laboratory standards being analysed before andafter sample batches of 5–12 otoliths. Otolith d18Ovalues were converted to temperature estimates basedon a species-specific fractionation equation developedfor Arctic charr by Godiksen et al. (2010) as follows:

1000 ln a ¼ �89:90ð1000TK�1Þ2� 617:19ð1000TK�1Þ þ 1089:24

where

ln a ¼ d18OotolithðVPDBÞ þ 1000d18OwaterðVPDBÞ þ 1000

and TK is temperature (K).Dried and ground YOY tissue samples (minus

head and fins) were analysed for d13C at the UW-EIL, on a Delta Plus continuous flow isotope ratiomass spectrometer (Thermo Finnigan, Bremen, Ger-many) coupled to a Carlo Erba elemental analyzer(CHNS-O EA1108, Carlo Erba, Milan, Italy) with ananalytical precision of �0.2&. Precision was estab-lished through repeat analysis of the IAEA CH6 stan-dard. Study measurement variability was additionally

assessed by the repeat analysis of one in ten samples.All analytical results are quoted in standard delta (d)notation as parts per mil (&) with respect to VPDB(Craig 1957).As otolith d13C may include metabolically and

environmentally derived carbon, isotope mixing mod-els developed to estimate the per cent of otolith d13Cderived from metabolic sources were applied to theobtained stable isotope data (Kalish 1991a; McConn-aughey et al. 1997). The percentage of metaboliccarbon in the otolith is thought to reflect individualmetabolic rate where a higher metabolic rate resultsin a higher per cent of metabolically derived carbon(Wurster & Patterson 2003). As such, the per cent ofmetabolic otolith carbon may be used as an index ofindividual field metabolic rate. For sites where waterd13CDIC samples were collected, the mixing modeldeveloped by McConnaughey et al. (1997) was usedto estimate the per cent of metabolically derivedcarbon (M) in the otolith as:

M ¼ ðd13Co � ear�HCO�3Þ � d13CDIC

d13CM � d13CDIC� 100

where d13Co is the isotopic composition of the oto-lith, ear�HCO�

3is the fractionation factor between ara-

gonite and bicarbonate, d13CM is the isotopecomposition of metabolically derived carbon andd13CDIC is the isotope value for ambient water.Romanek et al. (1992) estimated ear�HCO�

3to equal

2.7. In the absence of direct isotope measurementsfor metabolically derived carbon, d13C from theYOY fish tissue was used, similar to methods used inprevious studies (e.g. Kalish 1991b; McConnaugheyet al. 1997; Gillikin et al. 2009).To standardise for environmental differences over

the sampled latitudinal gradient, and thus facilitategrowth rate comparisons, an index of the thermalopportunity for growth (TOG) was calculated foreach site. In place of air temperatures, which havebeen used in previous studies (Power & McKinley1997; Chavarie et al. 2010; Rypel 2012), TOG wasdefined as cumulative degree days calculated by mul-tiplying mean otolith-inferred water temperatures bythe number of days between the beginning of thegrowing season and the capture date (hereafter,termed days used). For sites that were sampled over arange of days (Table 2), the mean otolith-inferredwater temperature was multiplied by the days usednumber consistent with the capture date for eachsample. To determine the beginning of the growingseason date, mean daily air temperatures wereobtained from compiled airport weather station data(www.tutiempo.net, www.wunderground.com). Anymissing daily temperature data were linearly interpo-lated from existing data. An 11-point smoothing

5

Latitudinal variation in YOY Arctic charr

function was applied to the data, and the beginningof the growing season was defined as the date whensmoothed air temperatures reached and remained con-sistently above zero. A lower limit of 0 °C was usedto define the beginning of the growing season owingto evidence that charr can feed and grow at tempera-tures close to zero (Br€ann€as & Wiklund 1992; Larssonet al. 2005). A growth rate was estimated for eachfish as FL divided by the cumulative degree days(mm per degree day). FL was not standardised bylength at emergence as egg diameter, which hasbeen linked to length at emergence (e.g. Baroudy &Elliott 1994b), does not vary with latitude amongpopulations for Arctic charr in North America (Poweret al. 2005).Multiple regression was used to assess the relation-

ship between mean otolith weight and latitude whilecontrolling for differences in capture date (day of yearof capture, DYC). As sites were sampled over 4 years,a multiple regression was conducted with the dependentvariable, mean growth rate, versus both latitude andyear of capture as independent variables. A follow-upmultiple regression was used to assess the relationshipbetween mean growth rate and latitude with all yearscombined, while controlling for DYC, with partial cor-relation used to calculate the amount of explained varia-tion between growth rate and latitude while controllingfor DYC (Field 2005). Analyses involving metaboli-cally derived carbon (M) were conducted on lacustrinesites for which d13CDIC was measured (n = 14 of 18).Riverine sites (n = 5) were excluded owing to expecteddifferences in the metabolic rates of lacustrine and riv-erine individuals (Graham 1949; Facey & Grossman1990). Multiple regression was also used to assess therelationship between M and latitude while controllingfor variations in growth rate and DYC among sites,with partial correlation used to calculate the amount ofexplained variation between M and latitude while con-trolling for growth rate (Field 2005). Data normalityand variance homogeneity were assessed prior to allstatistical analyses using standardised testing proce-dures (Zar 2010). Estimated regression model residualswere similarly assessed. Cook’s distance statistic wasused to detect regression outliers and influential casesthat might unduly bias regression coefficient estimates(Field 2005). The maximal Type I error in all statisticalanalyses was set at a = 0.05. All statistical analyseswere conducted using SPSS version 17 (IBM Corp.,New York, NY, USA).

Results

Mean growth rates ranged between 0.032 mm perdegree day at Tom’s Pond (56°420N), Labrador, and0.292 mm per degree day at Clements MarkhamLake (82°370N), Ellesmere Island (Table 2).

Although the largest otoliths were found in the twomost southerly populations (Table 2; KogluktokolukBrook and Tom’s Pond), from 58°N, there was nosignificant change in otolith size with latitude whenthe effect of capture date (DYC) was removed(r2 = 0.01, P > 0.05). There was also no significantyear effect on the relationship between mean growthrate and latitude (year/t-stat = �0.632, P > 0.05).There was a significant positive linear relationshipbetween mean growth rate and latitude, with theeffect of day of year of capture (DYC) removed, inwhich 43.8% of the variance in mean growth ratewas explained by latitude (Table 4a, Fig. 2). Regres-sion diagnostics indicated normal, heteroscedasticresiduals with no overly influential cases (Cook’s dis-tance statistic: all D < 1).Mean otolith-inferred metabolic rates ranged from

15.7% at Tom’s Pond to 45.1% at Clements Mark-ham Lake (Table 3). A significant positive relation-ship was found between mean M (the per cent ofmetabolically derived carbon within the otolith) andlatitude, with the effect of DYC removed, where61.2% of the variation was explained (Table 4b,Fig. 3). When the effect of mean growth rate wasalso removed, the significant positive relationshipbetween M and latitude was maintained (r2 = 0.54,P < 0.02, Table 4c, Fig. 4). Both M regressions pro-duced normal, heteroscedastic residuals with no evi-dence of overly influential cases (Cook’s distancestatistic: all D < 1).

Discussion

Inferred growth rates of young-of-the-year Arcticcharr sampled across a ~27° latitudinal gradient ineastern and central Canada demonstrated patternsconsistent with countergradient variation. Once stan-dardised by the TOG using the mean otolith-inferredwater temperatures and growing season length experi-enced by fish at each site, YOY Arctic charr fromnorthern populations had higher growth rates (mmper degree day) than southern conspecifics. Otolith-inferred metabolic rates demonstrated a similar latitu-dinal pattern, with YOY Arctic charr from northernpopulations having higher indices of metabolic raterelative to those from southern populations.The latitudinal compensation in growth rate, dem-

onstrated in this study (Fig. 2), is consistent with theCGV hypothesis for growth rate (Levins 1969) andcontributes to the growing body of literature whichhas demonstrated faster growth in populations frommore northerly climates (e.g. Conover & Present1990; Nicieza et al. 1994; Conover et al. 2009;Rypel 2012), including that for Arctic charr (Chavar-ie et al. 2010). Despite the growing empirical supportfor the CGV hypothesis (Conover et al. 2009), its

6

Sinnatamby et al.

validity and ubiquity continues to be questioned(Clarke 1993), and its applicability to salmonidsremains uncertain, with study findings varying con-siderably (Nicieza et al. 1994; Jensen et al. 2000;

Jonsson et al. 2001; Nicola & Almod�ovar 2004;Blanck & Lamouroux 2007; Parra et al. 2009; Logez& Pont 2011). Potential reasons for contrasting evi-dence for CGV have been suggested in previousstudies. Chavarie et al. (2010) noted that mostinstances of weak or absent evidence of CGV haveresulted from assessments of European populations.As North America is characterised by a steeper latitu-dinal gradient in temperature and growing seasonlength, it is more likely to result in pronounced varia-tions in growth patterns consistent with CGV. Like-wise, Rypel (2012) demonstrated strong correlationsbetween growth rate, climate and latitude amongNorth American populations of northern pike, Esoxlucius, whereas similar relationships were noticeablyweaker among Eurasian populations. In addition,

Table 4. Regression results for multiple regression models relating (a)mean growth rate (mm per degree day) to latitude while controlling for dayof year of capture (DYC), (b) mean M to latitude while controlling for DYCand (c) mean M to latitude while controlling for DYC and growth rate.Partial correlation coefficients measure the strength of association betweenmean M and the listed explanatory variables when the effect of the othervariable has been controlled in the analysis. Results for the full regressionmodels are noted in footnotes below the table.

Unstandardisedcoefficients

t-statistic PPartialcorrelationB

Standarderror

(a) Mean growth rate vs. latitude + DYC*Constant 0.051 0.415 0.124 >0.05Latitude 0.005 0.001 3.947 0.001 0.662DYC �0.001 0.002 �0.784 >0.05 �0.173

(b) Mean M vs. latitude + DYC†

Constant �129.338 62.431 �2.072 >0.05Latitude 1.050 0.252 4.168 0.002 0.782DYC 0.379 0.254 1.491 >0.05 0.410

(c) Mean M vs. latitude + GR + DYC‡

Constant �129.400 65.392 �1.979 >0.05Latitude 1.077 0.312 3.455 0.006 0.738Growthrate

�5.320 32.625 �0.163 >0.05 �0.051

DYC 0.374 0.268 1.398 >0.05 0.404

DYC, day of year of capture.*r2 = 0.579, F2,20 = 13.726, P < 0.001.†r2 = 0.613, F2,11 = 8.713, P < 0.01.‡r2 = 0.614, F3,10 = 6.421, P < 0.02.

Fig. 2. Partial regression plot depicting the relationship betweenmean growth rate (mm per degree day) and latitude for sampledYOY Arctic charr (r2 = 0.44) once the effect of day of year ofcapture (DYC) was removed. The solid line depicts the estimatedlinear relationship between the plotted variables. Growth rateresiduals and latitude residuals are the residuals resulting fromsimple linear regression analysis with each variable regressedagainst day of year of capture (Moya-Lara~no & Corcobado 2008).

Fig. 3. Partial regression plot illustrating the mean per cent ofmetabolically derived carbon (M) versus latitude for sampledYOY Arctic charr (r2 = 0.61) once the effect of DYC wasremoved. The solid line depicts the estimated linear relationshipbetween the plotted variables.

Fig. 4. Partial regression plot depicting the relationship betweenmean M and latitude for sampled Arctic charr (r2 = 0.54) oncethe effects of growth rate and DYC were removed. M residualsand latitude residuals are the residuals resulting from multiple lin-ear regression analysis with each variable regressed against meangrowth rate and the DYC.

7

Latitudinal variation in YOY Arctic charr

variability in methods, such as the failure to accountfor environmental differences among sites in somestudies (e.g. Blanck & Lamouroux 2007; Parra et al.2009; Logez & Pont 2011) may have contributed tothe observed inconsistencies among study findings.Methods used in this study permitted the compari-

son of growth variation in natural populations basedon in situ derived measures (mm per degree day)directly influenced by natural conditions. Fuller et al.(2010), for example, noted that the potential ecologi-cal effects of climate change can only be predictedonce physiology at the organism level is understoodwithin its natural habitat. While experimentalapproaches, such as common garden experiments,should be used to confirm the presence of a geneticcomponent in CGV (Conover et al. 2006), the practi-calities of experimental logistics limit the applicationof the approach when attempting to test CGV using alarge number of populations from across a large lati-tudinal range that includes the high Arctic, as wasperformed here. Further, Finstad & Jonsson (2012)related incubation temperature to later growth perfor-mance in Atlantic salmon, Salmo salar, and specu-lated that their findings may have implications forcommon garden experiments that aim to assess CGVin growth. Specifically, the authors suggested thatunless the parents of study individuals are also raisedin common conditions, embryos originating fromcold environments may experience experimental tem-peratures as being increased compared to their natalenvironment, resulting in increased growth rates andvice versa for embryos that originate from warm natalenvironments (Finstad & Jonsson 2012).Similar to the latitudinal patterns observed in

growth rate, a significant positive linear relationshipwas observed between otolith-inferred metabolic rateand latitude. We recognise that the limited number oflacustrine sites sampled from the southern end of thelatitudinal gradient limits the confidence with whichthese findings can be considered; however, no siteswere found to be unduly influential on the regressionmodel. Otolith-inferred metabolic rates, as estimatedhere, are influenced by several processes typicallyperformed by free-living organisms, such as thermo-regulation, locomotion, feeding, aggression, repro-duction and growth (Nagy 1987). As such,interpretation of the relationship between M and lati-tude is complex, and when considered alone, theincreased field metabolic rate indices demonstratedby northern populations may be interpreted to reflectone, or a combination of these processes, as well asintrinsic differences in SMR.The comparable patterns in both inferred growth

and field metabolic rates, when considered together,are suggestive of increased food consumption (Beam-ish 1974) in high-latitude populations. Although

northern lakes are less productive overall (e.g. Short-reed & Stockner 1986; ACIA 2005), the large pulseof productivity that occurs in the Arctic summer as aconsequence of continuous daylight and warm tem-peratures (ACIA 2005) may result in larger amountsof available energy per individual fish during the lim-ited open water season and may provide the opportu-nity for higher relative consumption during that time(Jørgensen et al. 1993). Indeed, previous studies havelinked increased growth rate achieved by northernpopulations to increased foraging and consumption(Menidia menidia, Present & Conover 1992; Ranatemporaria, Lindgren & Laurila 2010; Oryzias lati-pes, Suzuki et al. 2010). The observed latitudinal pat-terns in M may also reflect higher intrinsic SMRs inhigh-latitude populations. Similar findings have beenreported for brown trout, Salmo trutta, with higherSMRs found in individuals from colder environments(�Alvarez et al. 2006). High SMRs may enhancegrowth rate when resources are abundant (Forsethet al. 1999; Millidine et al. 2009; Biro & Stamps2010), and Atlantic salmon juveniles with higherSMRs have been shown to extract and assimilatemore nutrients and process meals faster (Millidineet al. 2009). The faster processing and assimilationtimes associated with higher SMRs therefore mayadapt fish to take advantage of the large pulse ofsummer Arctic productivity.The phenotypic variation in inferred growth and

field metabolic rates observed in YOY Arctic charracross eastern and central Canada demonstrates thesignificant ability of the species to utilise differentthermal regimes with different growing seasonlengths. Questions remain as to how Arctic charrmay ultimately respond to the rate of temperaturechange associated with climate warming (e.g. Reistet al. 2006a,b; Finstad & Hein 2012). Nevertheless,the marked phenotypic variation exhibited by the spe-cies suggests that phenotypic plasticity is likely toplay a key role in defining Arctic charr responses toclimate change and may ultimately facilitate geneticadaptation by buffering populations against theimmediate negative consequences of environmentalchange (Fuller et al. 2010; Fierst 2011). The capacityfor phenotypic plasticity, however, is not unlimited.Arctic charr populations from the southern extent ofthe range will face significant risks associated withrising temperatures and are more likely to face localextirpation if they are unable to behaviourally ther-moregulate and utilise cooler hypolimnetic waters,where available (e.g. Lehtonen 1996; Murdoch &Power 2013; Sinnatamby et al. 2013).The genetic component potentially underlying the

phenotypic patterns observed in this study could notbe verified or quantified using the field-based assess-ment employed here. In assessing the vulnerability of

8

Sinnatamby et al.

a species to climate change effects, the difficulty inseparating genetic from phenotypic components hasbeen identified as a challenge (Williams et al. 2008).Nevertheless, the implications of countergradient var-iation are thought to have relevance in predicting theresponse of organisms to projected climate warmingscenarios (Conover et al. 2009). Understanding thespecific contributions of phenotypic plasticity andgenetic variability to the observed latitudinal compen-sation gradient, however, will be critical to moreaccurately predicting the responses of Arctic charr toclimate change on various spatial and temporalscales, and facilitating the adaptive and sustainablemanagement of Arctic charr in many parts of its cur-rent geographic range.

Acknowledgements

The authors thank Milton Shears, Heidi Swanson, Tracey Loe-wen, Zoya Martin, Derek Muir, Nikolaus Gantner, GeoffPower, John Babaluk, Wendy Michaud and Shannon van derVelden for assistance with collection of field samples. Thanksalso to Parks Canada staff for base camp support in Quttinir-paaq National Park. Logistical support was provided by thePolar Continental Shelf Project. Financial support was pro-vided by an International Polar Year Grant to Reist, Powerand Dempson, ArcticNet and NSERC Discovery grants toPower, and an NSERC graduate scholarship to Sinnatamby.

References

ACIA 2005. Arctic climate impact assessment. New York,NY: Cambridge University Press. 1042 pp.

�Alvarez, D., Cano, J.M. & Nicieza, A.G. 2006. Microgeo-graphic variation in metabolic rate and energy storage ofbrown trout: countergradient selection or thermal sensitivity?Evolutionary Ecology 20: 345–363.

Baker, T.T. & Timmons, L.S. 1991. Precision of agesestimated from five bony structures of Arctic char (Salv-elinus alpinus) from the Wood River System, Alaska.Canadian Journal of Fisheries and Aquatic Sciences 48:1007–1014.

Baroudy, E. & Elliott, J.M. 1994a. The critical thermal limitsfor juvenile Arctic charr Salvelinus alpinus. Journal of FishBiology 45: 1041–1053.

Baroudy, E. & Elliott, J.M. 1994b. Racial differences in eggsand juveniles of Windermere charr, Salvelinus alpinus. Jour-nal of Fish Biology 45: 407–415.

Baumann, H. & Conover, D.O. 2011. Adaptation to climatechange: contrasting patterns of thermal-reaction-norm evolu-tion in Pacific versus Atlantic silversides. Proceedings of theRoyal Society B-Biological Sciences 278: 2265–2273.

Beamish, F.W.H. 1974. Apparent specific dynamic action oflargemouth bass, Micropterus salmoides. Journal of theFisheries Research Board of Canada 31: 1763–1769.

Biro, P.A. & Stamps, J.A. 2010. Do consistent individual dif-ferences in metabolic rate promote consistent individual dif-ferences in behavior? Trends in Ecology & Evolution 25:653–659.

Blanck, A. & Lamouroux, N. 2007. Large-scale intraspecificvariation in life-history traits of European freshwater fish.Journal of Biogeography 34: 862–875.

Bradshaw, W.E. & Holzapfel, C.M. 2008. Genetic response torapid climate change: it’s seasonal timing that matters.Molecular Ecology 17: 157–166.

Br€ann€as, E. & Wiklund, B.-S. 1992. Low temperature growthpotential of Arctic charr and rainbow trout. Nordic Journalof Freshwater Research 67: 77–81.

Buckley, L.B., Hurlbert, A.H. & Jetz, W. 2012. Broad-scaleecological implications of ectothermy and endothermy inchanging environments. Global Ecology and Biogeography21: 873–885.

Bullock, T.H. 1955. Compensation for temperature in themetabolism and activity of poikilotherms. Biological reviewsof the Cambridge Philosophical Society 30: 311–342.

Campana, S.E. 1999. Chemistry and composition of fish oto-liths: pathways, mechanisms and applications. Marine Ecol-ogy Progress Series 188: 263–297.

Chavarie, L., Dempson, J.B., Schwarz, C.J., Reist, J.D.,Power, G. & Power, M. 2010. Latitudinal variation ingrowth among Arctic charr in eastern North America: evi-dence for countergradient variation? Hydrobiologia 650:161–177.

Chown, S.L. & Gaston, K.J. 2008. Macrophysiology for achanging world. Proceedings of the Royal Society B-Biolog-ical Sciences 275: 1469–1478.

Clarke, A. 1993. Seasonal acclimatization and latitudinal com-pensation in metabolism: do they exist. Functional Ecology7: 139–149.

Conover, D.O. 1990. The relationship between capacity forgrowth and length of growing season: evidence for andimplications of countergradient variation. Transactions ofthe American Fisheries Society 119: 416–430.

Conover, D.O. & Present, T.M.C. 1990. Countergradient vari-ation in growth rate: compensation for length of the growingseason among Atlantic silversides from different latitudes.Oecologia 83: 316–324.

Conover, D.O. & Schultz, E.T. 1995. Phenotypic similarityand the evolutionary significance of countergradient varia-tion. Trends in Ecology & Evolution 10: 248–252.

Conover, D.O., Clarke, L.M., Munch, S.B. & Wagner, G.N.2006. Spatial and temporal scales of adaptive divergence inmarine fishes and the implications for conservation. Journalof Fish Biology 69 (Supplement C): 21–47.

Conover, D.O., Duffy, T.A. & Hice, L.A. 2009. The covari-ance between genetic and environmental influences acrossecological gradients: reassessing the evolutionary signifi-cance of countergradient and cogradient variation. Year inEvolutionary Biology 1168: 100–129.

Coplen, T.B., Kendall, C. & Hopple, J. 1983. Comparison ofstable isotope reference samples. Nature 302: 236–238.

Craig, H. 1957. Isotopic standards for carbon and oxygen andcorrection factors for mass-spectrometric analysis of carbondioxide. Geochimica et Cosmochimica Acta 12: 133–149.

Devereux, I. 1967. Temperature measurements from oxygenisotope ratios of fish otoliths. Science 155: 1684–1685.

Ellers, J. & Stuefer, J.F. 2010. Frontiers in phenotypic plastic-ity research: new questions about mechanisms, inducedresponses and ecological impacts. Evolutionary Ecology 24:523–526.

9

Latitudinal variation in YOY Arctic charr

Epstein, S. & Mayeda, T. 1953. Variation of O18 content ofwaters from natural sources. Geochimica et CosmochimicaActa 4: 213–224.

Epstein, S., Buchsbaum, R., Lowenstam, H.A. & Urey, H.C.1953. Revised carbonate-water isotopic temperature scale.Bulletin of the Geological Society of America 64: 1315–1325.

Facey, D.E. & Grossman, G.D. 1990. The metabolic cost ofmaintaining position for four North American stream fishes:effects of season and velocity. Physiological Zoology 63:757–776.

Field, A. 2005. Discovering statistics using SPSS, 2nd edn.London: Sage Publications. 779 pp.

Fierst, J.L. 2011. A history of phenotypic plasticity acceleratesadaptation to a new environment. Journal of EvolutionaryBiology 24: 1992–2001.

Finstad, A.G. & Hein, C.L. 2012. Migrate or stay: terrestrialprimary productivity and climate drive anadromy in Arcticchar. Global Change Biology 18: 2487–2497.

Finstad, A.G. & Jonsson, B. 2012. Effect of incubation tem-perature on growth performance in Atlantic salmon. MarineEcology Progress Series 454: 75–82.

Finstad, A.G., Næsje, T.F. & Forseth, T. 2004. Seasonal varia-tion in the thermal performance of juvenile Atlantic salmon(Salmo salar). Freshwater Biology 49: 1459–1467.

Forseth, T., Næsje, T.F., Jonsson, B. & H�arsaker, K. 1999.Juvenile migration in brown trout: a consequence of ener-getic state. Journal of Animal Ecology 68: 783–793.

Forseth, T., Larsson, S., Jensen, A.J., Jonsson, B., Naslund, I.& Berglund, I. 2009. Thermal growth performance of juve-nile brown trout Salmo trutta: no support for thermal adapta-tion hypotheses. Journal of Fish Biology 74: 133–149.

Fritz, P., Drimmie, R.J., Frape, S.K. & O’Shea, O., 1987. Theisotopic composition of precipitation and groundwater inCanada. In: International Atomic Energy Agency, eds. Anon-ymous isotope techniques in water resources development.Vienna: International Atomic Energy Agency, pp. 539–550.

Fuller, A., Dawson, T., Helmuth, B., Hetem, R.S., Mitchell,D. & Maloney, S.K. 2010. Physiological mechanisms incoping with climate change. Physiological and BiochemicalZoology 83: 713–720.

Gillikin, D.P., Hutchinson, K.A. & Kumai, Y. 2009. Onto-genic increase of metabolic carbon in freshwater musselshells (Pyganodon cataracta). Journal of GeophysicalResearch-Biogeosciences 114: G01007.

Godiksen, J.A., Svenning, M.-A., Dempson, J.B., Marttila,M., Storm-Suke, A. & Power, M. 2010. Development of aspecies-specific fractionation equation for Arctic charr (Salv-elinus alpinus (L.)): an experimental approach. Hydrobiolo-gia 650: 67–77.

Godiksen, J.A., Svenning, M.-A., Sinnatamby, R.N., Demp-son, J.B., Borgstrøm, R. & Power, M. 2011. Stable isotope-based determinations of the average temperatures experi-enced by young-of-the-year Svalbard Arctic charr (Salveli-nus alpinus (L.)). Polar Biology 34: 591–596.

Graham, J.M. 1949. Some effects of temperature and oxygenpressure on the metabolism and activity of the speckledtrout, Salvelinus fontinalis. Canadian Journal of ResearchSeries D 27: 270–288.

Guiguer, K.R.R.A., Drimmie, R. & Power, M. 2003. Validat-ing methods for measuring d18O and d13C in otoliths from

freshwater fish. Rapid Communications in Mass Spectrome-try 17: 463–471.

Høie, H., Otterlei, E. & Folkvord, A. 2004. Temperature-dependent fractionation of stable oxygen isotopes in otolithsof juvenile cod (Gadus morhua L.). ICES Journal of MarineScience 61: 243–251.

Jensen, A.J., Forseth, T. & Johnsen, B.O. 2000. Latitudinalvariation in growth of young brown trout Salmo trutta. Jour-nal of Animal Ecology 69: 1010–1020.

Johnson, L., 1980. The Arctic charr, Salvelinus alpinus. In:Balon, E.K. ed. Charrs: salmonid fishes of the genus Salveli-nus. Volume 1: Perspectives in vertebrate science. Hingham,MA: Kluwer Boston Inc., pp. 15–92.

Jonsson, B., Forseth, T., Jensen, A.J. & Næsje, T.F. 2001.Thermal performance of juvenile Atlantic salmon, Salmo sa-lar L. Functional Ecology 15: 701–711.

Jørgensen, E.H., Christiansen, J.S. & Jobling, M. 1993.Effects of stocking density on food intake, growth perfor-mance and oxygen consumption in Arctic charr (Salvelinusalpinus). Aquaculture 110: 191–204.

Kalish, J.M. 1991a. 13C and 18O isotopic disequilibria in fishotoliths: metabolic and kinetic effects. Marine Ecology Pro-gress Series 75: 191–203.

Kalish, J.M. 1991b. Oxygen and carbon stable isotopes in theotoliths of wild and laboratory-reared Australian salmon (Ar-ripis trutta). Marine Biology 110: 37–47.

Larsson, S., Forseth, T., Berglund, I., Jensen, A.J., N€aslund,I., Elliott, J.M. & Jonsson, B. 2005. Thermal adaptation ofArctic charr: experimental studies of growth in eleven charrpopulations from Sweden, Norway and Britain. FreshwaterBiology 50: 353–368.

Lehtonen, H. 1996. Potential effects of global warming onnorthern European freshwater fish and fisheries. FisheriesManagement Ecology 3: 59–71.

Levins, R. 1969. Thermal acclimation and heat resistance inDrosophila species. The American Naturalist 103: 483–499.

Lindgren, B. & Laurila, A. 2009. Physiological variationalong a geographical gradient: is growth rate correlated withroutine metabolic rate in Rana temporaria tadpoles? Biolog-ical Journal of the Linnean Society 98: 217–224.

Lindgren, B. & Laurila, A. 2010. Are high-latitude individualssuperior competitors? A test with Rana temporaria tadpoles.Evolutionary Ecology 24: 115–131.

Logez, M. & Pont, D. 2011. Variation of brown trout Salmotrutta young-of-the-year growth along environmental gradi-ents in Europe. Journal of Fish Biology 78: 1269–1276.

Maitland, P.S. 1995. World status and conservation of theArctic charr, Salvelinus alpinus (L.). Nordic Journal ofFreshwater Research 71: 113–127.

McConnaughey, T.A., Burdett, J., Whelan, J.F. & Paull, C.K.1997. Carbon isotopes in biological carbonates: respirationand photosynthesis. Geochimica et Cosmochimica Acta 61:611–622.

McCrea, J.M. 1950. On the isotopic chemistry of carbonatesand a paleotemperature scale. Journal of Chemical Physiol-ogy 18: 849–857.

Millidine, K.J., Armstrong, J.D. & Metcalfe, N.B. 2009.Juvenile salmon with high standard metabolic rates havehigher energy costs but can process meals faster. Proceed-ings of the Royal Society B-Biological Sciences 276:2103–2108.

10

Sinnatamby et al.

Moya-Lara~no, J. & Corcobado, G. 2008. Plotting partial corre-lation and regression in ecological studies. Web Ecology 8:35–46.

Murdoch, A. & Power, M. 2013. The effect of lake morphom-etry on thermal habitat use and growth in Arctic charr popu-lations: implications for understanding climate-changeimpacts. Ecology of Freshwater Fish 22: 453–466.

Nagy, K.A. 1987. Field metabolic rate and food requirementscaling in mammals and birds. Ecological Monographs 57:111–128.

Nicieza, A.G., Reyes-Gavil�an, F.G. & Bra~na, F. 1994. Differ-entiation in juvenile growth and bimodality patterns betweennorthern and southern populations of Atlantic salmon (Salmosalar L.). Canadian Journal of Zoology 72: 1603–1610.

Nicola, G.G. & Almod�ovar, A. 2004. Growth pattern ofstream-dwelling brown trout under contrasting thermal con-ditions. Transactions of the American Fisheries Society 133:66–78.

Parra, I., Almod�ovar, A., Nicola, G.G. & Elvira, B. 2009. Lat-itudinal and altitudinal growth patterns of brown trout Salmotrutta at different spatial scales. Journal of Fish Biology 74:2355–2373.

Power, M. & McKinley, R.S. 1997. Latitudinal variation inlake sturgeon size as related to the thermal opportunity forgrowth. Transactions of the American Fisheries Society 126:549–558.

Power, M., Dempson, J.B., Reist, J.D., Schwarz, C.J. &Power, G. 2005. Latitudinal variation in fecundity amongArctic charr populations in eastern North America. Journalof Fish Biology 67: 255–273.

Power, M., Reist, J.D. & Dempson, J.B. 2008. Fish in highlatitude arctic lakes. In: Vincent, W.F. & Laybourn-Parry, J.,eds. Polar lakes and rivers: limnology of Arctic and Antarc-tic aquatic ecosystems. Oxford: Oxford University Press, pp.249–268.

Present, T.M.C. & Conover, D.O. 1992. Physiological basis oflatitudinal growth differences in Menidia menidia: variationin consumption or efficiency. Functional Ecology 6: 23–31.

Reist, J.D., Wrona, F.J., Prowse, T.D., Power, M., Dempson,J.B., Beamish, R.J., King, J.R., Carmichael, T.J. & Saw-atzky, C.D. 2006a. General effects of climate change onArctic fishes and fish populations. Ambio 35: 370–380.

Reist, J.D., Wrona, F.J., Prowse, T.D., Power, M., Dempson,J.B., King, J.R. & Beamish, R.J. 2006b. An overview ofeffects of climate change on selected Arctic freshwater andanadromous fishes. Ambio 35: 381–387.

Romanek, C.S., Grossman, E.L. & Morse, J.W. 1992. Carbonisotopic fractionation in synthetic aragonite and calcite:effects of temperature and precipitation. Geochimica et Cos-mochimica Acta 56: 419–430.

Rypel, A.L. 2012. Meta-analysis of growth rates for a circum-polar fish, the northern pike (Esox lucius), with emphasis oneffects of continent, climate and latitude. Ecology of Fresh-water Fish 21: 521–532.

Rypel, A.L. 2014. Concordant estimates of countergradientgrowth variation in striped bass (Morone saxatilis) usingcomparative life-history data. Canadian Journal of Fisheriesand Aquatic Sciences 69: 1261–1265.

Scholander, P.F., Flagg, W., Walters, V. & Irving, L. 1953.Climatic adaptation in Arctic and tropical poikilotherms.Physiological Zoology 26: 67–92.

Schwarcz, H.P., Gao, Y., Campana, S.E., Browne, D., Knyf,M. & Brand, U. 1998. Stable carbon isotope variations in ot-oliths of Atlantic cod (Gadus morhua). Canadian Journal ofFisheries and Aquatic Sciences 55: 1798–1806.

Shephard, S., Trueman, C., Rickaby, R. & Rogan, E. 2007.Juvenile life history of NE Atlantic orange roughy from oto-lith stable isotopes. Deep-Sea Research I 54: 1221–1230.

Sherwood, G.D. & Rose, G.A. 2003. Influence of swimmingform on otolith d13C in marine fish. Marine Ecology Pro-gress Series 258: 283–289.

Shortreed, K.S. & Stockner, J.G. 1986. Trophic status of 19sub-Arctic lakes in the Yukon Territory. Canadian Journalof Fisheries and Aquatic Sciences 43: 797–805.

Sinnatamby, R.N., Shears, M., Dempson, J.B. & Power, M.2013. Thermal habitat use and growth in young-of-the-yearArctic charr from proximal fluvial and lacustrine populationsin Labador, Canada. Journal of Thermal Biology 38: 493–501.

Suzuki, Y., Miyake, T. & Yamahira, K. 2010. An acquisitiontrade-off with fast growth in a fish, the medaka Oryzias lati-pes: why do low-latitude ectotherms grow more slowly?Evolutionary Ecology 24: 749–759.

Williams, S.E., Shoo, L.P., Isaac, J.L., Hoffmann, A.A. &Langham, G. 2008. Towards an integrated framework forassessing the vulnerability of species to climate change.PLoS Biology 6: 2621–2626.

Wohlschlag, D.E. 1960. Metabolism of an Antarctic fish andthe phenomenon of cold adaptation. Ecology 41: 287–292.

Wurster, C.M. & Patterson, W.P. 2003. Metabolic rate of lateHolocene freshwater fish: evidence from d13C values of oto-liths. Paleobiology 29: 492–505.

Wurster, C.M., Patterson, W.P., Stewart, D.J., Bowlby, J.N. &Stewart, T.J. 2005. Thermal histories, stress, and metabolicrates of chinook salmon (Oncorhynchus tshawytscha) inLake Ontario: evidence from intra-otolith stable isotopeanalyses. Canadian Journal of Fisheries and Aquatic Sci-ences 62: 700–713.

Zar, J.H. 2010. Biostatistical analysis, 5th edn. Upper SaddleRiver, NJ: Prentice Hall. 663 pp.

11

Latitudinal variation in YOY Arctic charr