Metamorphosis in Chironomids,More than Mercury Supply, ControlsMethylmercury Transfer to Fish inHigh Arctic LakesJ O H N C H E T E L A T , M A R C A M Y O T , *L O U I S E C L O U T I E R , A N DA L E X A N D R E P O U L A I N †

Groupe de recherche interuniversitaire en limnologie (GRIL),Departement de sciences biologiques, Universite de Montreal,C.P. 6128, Succursale Centre-Ville, Pavillon Marie-Victorin,Montreal, Quebec, Canada, H3C 3J7

Received June 12, 2008. Revised manuscript receivedSeptember 18, 2008. Accepted October 10, 2008.

Lake-dwelling Arctic char (Salvelinus alpinus) are monitoredinternationally as a sentinel species for effects of atmosphericmercury deposition on Arctic fresh waters. We investigated thecontrol of mercury supply and biological processes on themethylmercury (MeHg) content of their main food, aquaticchironomids, in 22 lakes and ponds in the Canadian High Arctic.Total mercury (THg) concentrations in sediment (correctedfor organic matter content) increased with drainage basin size,suggesting a gradient in mercury loading among study sites.MeHg concentrations in sediment and water were lowand relatively uniform along this THg supply gradient, suggestingMeHg production in High Arctic lakes is weakly coupled toinorganic mercury supply. Metamorphosis was a key biologicalprocess that concentrated MeHg in adult chironomids1.7-2.9 times more than in immature stages. Drainage basinsize, environmental mercury levels, and habitat characteristicswere also significant factors but they explained less variationin chironomid MeHg concentration than their degree of maturity.Chironomid larvae, pupae, and adults are distinct mercurysources for fish, and we provide evidence from nitrogen stableisotopes and published feeding studies that suggest differentialconsumption of these stages may affect MeHg uptake byArctic char. We conclude that biological and food web processeshave a greater impact on MeHg transfer to fish thanatmospheric mercury deposition in High Arctic lakes.

IntroductionArctic char (Salvelinus alpinus), with its circumpolar dis-tribution, is used internationally as a sentinel species to detectspatial and temporal trends in the global transport of mercuryto Arctic fresh waters (1). This metal is a contaminant ofconcern in polar regions because atmospheric mercury isdeposited intensively on snow following polar sunrise (2),snow and melt waters often contain elevated concentrationsof mercury (3, 4), and high burdens have been reported intop predatory animals because of biomagnification (5). The

Arctic is also undergoing rapid environmental change fromclimate warming which is anticipated to impact the mercurycycle (6). Populations of landlocked char have been sampledin high latitude lakes of Canada (∼75°N) since the 1990s,and mercury levels in muscle tissue range an order ofmagnitude from 0.15-1.5 µg g-1 wet weight (7). Whilecannibal feeding by larger char is one factor that enhancesmercury bioaccumulation (7), the absence of informationon mercury in lower trophic levels may hinder the inter-pretation of spatial and temporal trends in food webbiomagnification of this metal.

Chironomids (Diptera: Chironomidae) play a key role inthe transfer of mercury to landlocked Arctic char in highlatitude lakes (>70°N) of Canada because they are a primaryfood source in many lakes. Few species can survive underthe extreme environmental conditions imposed on theseecosystems, and lake food webs are simple, consisting ofArctic char, chironomids, and chironomid food sources insediment (algae, bacteria, detrital organic matter) (8). Chi-ronomid production is far greater than that of zooplanktonor other macroinvertebrates in these lakes (9), making it themain source of food and as a consequence, the main vectorof methylmercury (MeHg) to landlocked char. These eco-systems differ from lower latitude Arctic lakes where zoop-lankton can be an important diet item for Arctic char (10).

Chironomids are holometabolous insects and pass through4 stages in their life cycle: eggs, larvae, pupae, and adults.Fertilized eggs are deposited on the water surface, sink tosediment and hatch within a month (9). Larvae grow slowlyin sediment, passing through 4 successive instars and typicallyrequire 2-3 years to reach maturity in the High Arctic (9, 11).In spring or summer of the final year in their life cycle, latefourth instar larvae develop wing pads, molt to pupae (2week inactive phase when adult structures are formed), andthen swim to the water surface where adults cast their pupalskin (exuvium) and emerge to mate.

In this study, we investigated the control of mercury supplyand biological processes that affect MeHg content of chi-ronomids in lakes and ponds of the Canadian High Arctic.We estimated mercury supply by measuring environmentalconcentrations and by comparing water bodies with differentdrainage basin size. Snowmelt is the main source of totalmercury (THg) to High Arctic lakes (12), and we hypothesizedthat sites with larger drainage basins receive greater loadingsof aqueous mercury originating from snow. Subsequent insitu methylation of inorganic Hg and MeHg uptake into thefood web should result in higher concentrations in chirono-mids. Sediment production of MeHg in Alaskan lakes ispositively related to drainage basin size because of theimportance of this inorganic mercury load for methylationrates (6). We sampled water bodies in barren polar desertsas well as a polar oasis with lush meadows to account forpotential influences of drainage basin type because theamount of organic matter transported from drainage basinscould affect the mercury cycle. Biological processes, namelygrowth and metamorphosis, were examined by comparingmercury accumulation in the different stages of chironomiddevelopment.

We demonstrate that metamorphosis plays a dominantrole in controlling MeHg concentrations in chironomids andthat mercury supply is of secondary importance due to lowand relatively uniform environmental levels across sites.Different development stages of chironomids representdistinct sources of MeHg, and we provide evidence fromnitrogen isotope ratios and published diet studies of Arcticchar that different feeding on these stages likely affects MeHg

* Corresponding author phone: (514) 343-7496; fax: (514) 343-2293; e-mail: m.amyot@umontreal.ca.

† Present address: Departments of Biology and Earth and PlanetarySciences, Massachusetts Institute of Technology, Cambridge, MA02139.

Environ. Sci. Technol. 2008, 42, 9110–9115

9110 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 24, 2008 10.1021/es801619h CCC: $40.75 2008 American Chemical SocietyPublished on Web 11/11/2008

uptake by fish. These findings have fundamental implicationsfor the roles of atmospheric deposition and food webdynamics in the mercury cycle of High Arctic lakes.

Experimental SectionField Sampling. Water bodies were investigated on 3 islandsof the Canadian Arctic Archipelago (Cornwallis, Devon,Somerset, 74-76°N latitude) in July and August of 2005 and2006 (Figure 1). This High Arctic region has a severe climatewith cold temperatures (February mean)-33 °C, July mean) 4 °C), low annual precipitation (∼150 mm) and a veryshort ice-free season of approximately 6-8 weeks. Seventeenlakes and five ponds were sampled mostly located nearResolute Bay, Cornwallis Island (74°30′, 94°55′W) or a remotefield camp at Truelove Lowland, Devon Island (75°40′N,84°33′W). Truelove Lowland is a 43 km2 polar oasis with awarmer microclimate than Resolute (13), and drainage basinsare covered with lush meadows of sedge and moss. Incontrast, other study sites have polar desert drainage basinsof barren sedimentary rock with typically <5% plant cover(13). The study lakes had maximum depthsg7 m and all areknown to contain populations of landlocked Arctic charexcept one (Titti Lake). Ponds had maximum depths <3.5 mand were assumed to be fishless. Each site was sampled onone occasion except 4 lakes sampled in both 2005 and 2006(see Supporting Information, Table 1, for site descriptions).

Chironomids in High Arctic lakes are abundant at shallowdepths (∼2-4 m) (9, 14, 15) allowing them to emerge beforethe ice has completely melted. Mating occurs mainly onsubstrates (ice, water surface, shoreline rocks) rather thanby swarming because of nearly continuous winds (9). FromJuly to early August, adults and pupae were collected fromthree near-shore stations in each water body by removingindividuals from ice or water surfaces (occasionally shorelinerocks) using tweezers or a fine-mesh net. Larvae were alsocollected at three stations per lake with a 500 µm D-framenet by kick-sweeping sediment in water 0.5-1.25 m deep. Inaddition, larvae were occasionally collected from profundalsediment with an Ekman grab (n ) 5) or caught in off-shorewaters with a zooplankton net (n ) 4). Benthic primary

producers (macroscopic Nostoc spheres, filamentous algae)were collected from one to four stations in seven water bodies(n ) 15).

Chironomids were washed in ultrapure water, separatedby development stage (adult, pupa, larva), placed in plasticscintillation vials and frozen. Samples collected from the threestations per lake were treated as separate replicates. Adultsamples typically contained>100 individuals but pupae andlarvae samples contained less individuals (∼10, range )1->20). Fewer than three replicates were obtained for larvaeand pupae at some sites. Some individuals were preservedin concentrated ethanol for taxonomic identification. Thematurity of larvae varied from third or early fourth instarsthat needed an additional year of growth (9) to fourth instarsready to emerge that summer or next spring (11). Chironomidsubfamilies were pooled together, with the exception of larvaefrom the predatory subfamily Tanypodinae, which wereseparated from nonpredatory taxa when present at somepolar oasis sites.

Water was obtained as surface grabs from a near-shorestation (∼1 m depth) in each water body and also mid-depthat a pelagic station in 75% of the lakes (due to logisticconstraints). For total mercury (THg) and MeHg analyses,triplicate acid-cleaned bottles of amber glass or Teflon wererinsed with lake water prior to collection and preserved withHCl. Deep-water mercury samples were collected with acid-washed Teflon tubing and a peristaltic pump. Inflow watersof one polar oasis and two polar desert lakes were sampledon 1-4 occasions for mercury by surface grabs. Ancillarywater chemistry samples for anions (chloride, sulfate),dissolved organic carbon (DOC, 0.45 µm filtration), and totalphosphorus were collected in duplicate along with mercurysamples. Conductivity, pH, dissolved oxygen, and watertemperature were measured in situ with a YSI sonde. Sitewater quality is summarized in Supporting Information, Ta-ble 2.

Sediment was collected by Ekman dredge from three near-shore stations (∼1-2 m depth) in 18 of the 22 water bodieswhere larvae were sampled. Each sample of sediment wastaken from the top 5 cm layer, avoiding sediment directly incontact with the Ekman, and frozen in plastic scintillationvials. In 2006 only, profundal sediment was collected in lakes(n ) 7) at 1-2 stationsg6.5 m deep using the same method.

Maximum water depth (Zmax) was obtained either fromthe literature (n ) 14 sites), from depth sounding alongtransects with a weighted line (n ) 4), or from the Zmax

observed during sampling (n ) 4). Surface area (SA) anddrainage basin area (DA) of water bodies were measured inSIGIS v2.60 software using Landsat-7 images and digital mapsof the Canadian National Topographic Data Base.

Laboratory Analyses. MeHg was extracted from lyoph-ilized and homogenized samples of chironomid (n ) 206),algae (n ) 15) or sediment (n ) 50) by digestion in 4 M HNO3

at 55 °C for 16 h. Some samples that contained small amountsof material (1-10 mg) were weighed intact on a microbalance(0.06 mg accuracy) and then homogenized directly in thedigestion vial with an acid-cleaned glass rod. Water samplesfor MeHg analysis were predistilled with additions of KCland H2SO4 to remove matrix interferences. Comparison ofdistillation and acid extraction methods for sediment MeHganalysis yielded similar results except for lower recovery byacid digestion on a few samples with high organic content.MeHg was extracted from those samples (n) 12) by aqueousdistillation as for water samples but with the additionalreagent CuSO4. MeHg extract was derivatized by aqueousethylation using NaBEt4, trapped with Tenax and measuredwith a Tekran 2500 cold vapor atomic fluorescence spec-trometer (CVAFS). MeHg burden (pg ind-1) was measuredfor a subset of chironomid samples (n ) 88) using mean

FIGURE 1. Locations of study areas on three islands in theCanadian Arctic Archipelago. The number of water bodiessampled is presented for each area.

VOL. 42, NO. 24, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 9111

body weights determined separately for each digestion(sample mass divided by the number of individuals).

THg concentration in sediment (n ) 62) and adultchironomids (n ) 74) was measured with a Direct MercuryAnalyzer (DMA-80) in which sample was combusted at 750°C and the mercury vapor was retained on a gold trap foranalysis by cold vapor atomic absorption spectrometry.Aqueous THg was determined on 50 mL samples by BrCloxidation, SnCl2 reduction, two-stage gold amalgamation andgas-phase detection with a Tekran 2600 CVAFS. Water DOCwas measured in a Pt-catalyzed Shimadzu TOC-5000 ana-lyzer. Organic matter (OM) content of sediment was deter-mined by loss on ignition (LOI) in an oven at 550 °C for 2 h.

Recoveries of certified reference materials for mercurywere 97 ( 11% of MeHg (n ) 133) and 106 ( 3% of THg (n) 19) in TORT-2 lobster hepatopancreas (National ResearchCouncil of Canada), 91 ( 17% of MeHg (n ) 23) in IAEA-405estuarine sediment (International Atomic Energy Agency),and 99 ( 6% of THg (n ) 16) in SO-2 soil (CANMET Mineraland Energy Technology). Analytical detection limits, esti-mated as 3 times the standard deviation (SD) of 10 blanks,were 0.04 ng MeHg g-1 for 250 mg of sediment and 0.05 ngTHg L-1 and 0.02 ng MeHg L-1 for 50 mL aliquots of water.

Stable isotope ratios of nitrogen and carbon were mea-sured in chironomids (larvae n ) 46, pupae n ) 19, adult n) 76) at the G.G. Hatch Stable Isotope Laboratory at theUniversity of Ottawa (Ottawa, Canada) on a DeltaPlus XPIsotope Ratio Mass Spectrometer. Stable isotope ratios wereexpressed in delta notation (δ) as the parts per thousand(‰) deviation from standards of atmospheric N2 gas for δ15Nand Vienna PeeDee Belemnite for δ13C. Analytical precisionwas e0.2‰.

Data Analysis. A general linear model (GLM) was used totest the effects of development stage (larvae, pupae, adults),mercury supply (DA:SA ratio, MeHg concentration in sedi-ment and water), and habitat charactersistics (DOC, watertemperature) on the MeHg concentration in chironomids.Four water bodies without sediment data were excluded fromthe model (n ) 183). Main effects were tested using type IIIsum-of-squares (SS), and effect sizes for factors in the modelwere evaluated with the partial ETA squared coefficient (effectSS divided by the effect SS plus error SS). Variables were logor arcsine transformed, if required, to meet assumptions ofstatistical tests and half the detection limit was used for valuesbelow detection. Means are presented ( 1 SD unlessotherwise specified.

Results and DiscussionOrganic Carbon and Drainage Basin Influences on MercurySupply. Concentrations of sediment mercury were very lowat our High Arctic sites (Table 1). THg and MeHg in near-shore sediment averaged 5.8 ( 2.8 ng g-1 and 0.13 ( 0.17 ngg-1, respectively, which is less than half of levels found insediment of lower latitude lakes (6, 16). Mercury concentra-tions in profundal sediment were higher but still compara-tively low (Table 1). Water-column concentrations of mercuryaveraged 0.62 ( 0.33 ng THg L-1 and 0.04 ( 0.03 ng MeHg

L-1, with slightly higher MeHg observed in 1 pond (Pond I:0.15 ng L-1). Our aqueous MeHg concentrations wereconsistent with other studies of lakes on Cornwallis (0.05 (0.01 ng L-1; (3)) and Ellesmere islands (0.04 ( 0.03 ng L-1;(4)).

Organic carbon was sparse in sediment (5 ( 8% LOI) andwater (1.7 ( 1.9 mg DOC L-1) and had a strong influence onthe environmental distribution of mercury. THg and MeHgconcentrations in sediment were positively related to OMcontent (linear regression r2 ) 0.55-0.65, Fg 74.4, p< 0.001,n ) 62, Supporting Information, Figure 1). Y-intercepts forthe linear regression models had low values, suggesting thatmost THg and MeHg was bound to OM rather than inorganicparticles (intercepts: 2.4 ng THg g-1, 0.02 ng MeHg g-1).Similarly, MeHg concentration in the water column wasslightly higher in more DOC-rich waters (linear regressionr2 ) 0.21, F ) 6.5, p ) 0.019, n ) 22).

Drainage basin size significantly affected the THg con-centration of sediment (Figure 2). THg in both near-shoreand deep-water sediment was higher at sites with largerdrainage basins, after correcting for the amount of OMpresent. This finding suggests that water bodies with largerdrainage basins received greater THg loads, most likely fromsurface runoff during spring freshet which is the principalsource of THg to High Arctic fresh waters (12). In contrast,MeHg in near-shore sediment (OM-corrected) was not relatedto inorganic mercury supply estimated either by sedimentTHg or drainage basin size (p ) 0.30, n ) 18). In profundalsediment, OM-corrected MeHg and THg contents weremarginally related (linear regression r2 ) 0.34, F ) 4.0, p )0.10, n ) 7) but drainage basin size had no effect on theformer (p ) 0.44, n ) 7).

We found consistently low levels of MeHg in water andsediment at sites along a gradient in THg supply suggestingin situ MeHg production was weakly coupled to inorganicmercury loading. Information is currently lacking on mercurymethylation rates in High Arctic lakes, and we suggest thatthe capacity of the sediment bacterial community to generateMeHg may be strongly limited by poor environmentalconditions for methylation rather than the availability ofinorganic mercury. Specifically, methylation rates may belimited by small pools of organic substrate in sediment, coldtemperatures, alkaline water (pH ∼8), and a well-mixed,oxygenated water column (6, 17, 18). Summer-time terrestrialinputs of aqueous MeHg may also be low as suggested byour limited sampling of inflows to a few lakes. TerrestrialDOC was high in seepage from meadows at the polar oasis(3.3-10.3 mg DOC L-1, n ) 3) and resulted in higher water-column DOC in the 8 water bodies there (3.7 ( 1.9 mg L-1)compared to polar desert sites (0.6 ( 0.4 mg L-1, n ) 14).However, surface seepage from the oasis had similar MeHgconcentrations (0.02-0.07 ng L-1, n) 3) to low-DOC inflowsfrom 2 polar desert drainage basins (0.02-0.07 ng L-1, n )8). Likewise, average water-column MeHg concentrationsdid not differ between water bodies at polar oasis and desertsites (t-test, t ) 1.6, p ) 0.13, n ) 22).

TABLE 1. Dry Weight Contents of OM, THg, and MeHg in Bulk Sediment from Near-Shore and Profundal Zones of Study Sites

near-shore sediment (n ) 18 lakes and ponds) profundal sediment (n ) 7 lakes)

characteristic mean range mean range

OM (% LOI) 3.7 1.3-12.2 17.1 2.0-31.5THg (ng g-1) 5.8 2.7-12.2 26.9 12.1-50.7THg (ng g OM-1) 206 93-529 257 91-710MeHg (ng g-1) 0.13 <0.04-0.64 0.52 0.13-1.09MeHg (ng g OM-1) 3.3 0.9-17.7 3.6 2.1-6.3% MeHg 2.1 0.5-8.7 1.9 0.9-3.3

9112 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 24, 2008

Bioaccumulation of MeHg in Chironomids. The MeHgconcentration in chironomids is primarily determined bythe concentration of MeHg in their diet, the amount of foodconsumed (resulting in MeHg uptake), and changes inbiomass during development (19). Potential diet sources,specifically algae and sediment OM, were investigated fortheir influence on MeHg concentrations in chironomid larvae.There was an almost significant association between MeHgconcentrations in algae and larvae (linear regression r2 )0.38, F)4.7, p)0.084, n)7 water bodies) but no relationshipwas found between MeHg concentrations in near-shoresediment (OM-corrected) and the larvae collected there (p) 0.31, n ) 18). Low interlake variation in MeHg content ofchironomid food sources (sediment ) 3.3 ( 3.8 ng g OM-1,benthic algae ) 2.5 ( 2.7 ng g-1) may explain the lack of astrong dietary influence.

In contrast, concentrations of MeHg in chironomidsstrongly increased during growth and metamorphosis (Figure3, Table 2). Adults were 1.7 times more concentrated withMeHg than pupae and 2.9 times more concentrated thanimmature larvae when samples were averaged over all sites.MeHg was the predominant form of mercury in adultchironomids, averaging 82 ( 15% of THg concentrations (n) 74). We inferred changes in MeHg concentration duringchironomid development by comparing different individualsin each stage rather than following the same individualsthrough time. This approach may have introduced additionalvariation to the analysis due to differences in taxonomybetween stages or degree of maturity in the larvae. Never-theless, adults were consistently elevated in MeHg relativeto larvae in all 17 lakes and 5 ponds (Figure 4). Relative MeHgconcentrations in pupae were more variable. Interannual

variation in the MeHg content of larvae and adults from 4lakes was low for 2 consecutive sampling years (see theSupporting Information, Figure 2).

MeHg was more concentrated in adults than pupaebecause of a reduction in body weight during emergenceand mating with no associated loss in MeHg burden (one-way ANOVA, F ) 16.7, p < 0.001, n ) 88, Figure 3). About10% of pupal dry weight is lost when the pupal exuvium isshed (14), and this skin was found to be mercury-poor (1.0(0.5 pg MeHg, n)12). Since adults do not feed, an additional∼20-50% of body dry weight can be lost when energy reservesare consumed during mating (20). Larvae were least con-centrated with MeHg because they were still growing andhad the lowest burdens (Figure 3). This process is not uniqueto the High Arctic since other bioaccumulating compounds(e.g., organochlorines, cadmium) concentrate during meta-morphosis in chironomids of lower latitudes (21, 22).

A GLM identified the relative importance of biologicalprocesses, mercury supply, and habitat characteristics forMeHg bioaccumulation in chironomids (Table 2). Develop-ment stage explained the most variation in chironomid MeHgconcentration, and significant but secondary effects werefound for water temperature, MeHg concentrations insediment and water, DA:SA, and DOC (Table 2). The modelindicated that MeHg concentrations in chironomids wereinfluenced more by their biological development than byamong-site variation in mercury supply and habitat char-acteristics. The specific developmental effects were anincrease in MeHg accumulation during growth from im-mature larvae to pupae and a concentration of MeHg frompupae to adults due to loss of body mass (Figure 3).

Chironomids had lower MeHg content at sites withwarmer temperatures and higher DOC which were mainlyfound at the polar oasis (Table 2, and the SupportingInformation, Figure 3). DOC may have had an inhibitoryeffect on MeHg bioaccumulation by reducing mercurybioavailability in the water column (18). Taxonomic differ-ences could explain the observed pattern in MeHg ac-cumulation along a temperature gradient. The MeHg con-centration of adult chironomids was positively related to thepercent abundance of individuals from the subfamily Di-amesinae in samples from 2006 (linear regression r2 ) 0.22,F ) 5.3, p ) 0.038, n ) 16). Diamesinae prefer cold water(23), and their percent abundance at our study sites wasnegatively related to water temperature (linear regression r2

) 0.22, F ) 5.3, p ) 0.037, n ) 16). Adult chironomids fromNorth Lake (which had the highest MeHg content; Figure 4)were sorted into the 2 main subfamilies present (Ortho-cladinae and Diamesinae), and Diamesinae were on average

FIGURE 2. Relationship between drainage basin size and thedry weight concentration of sediment THg (corrected for OMcontent) in near-shore and profundal zones. Means ( 1standard error are presented for near-shore sediment (n ) 3replicates), whereas profundal sediment data are mostly singlesamples.

FIGURE 3. Comparison of MeHg concentrations and burdensbetween chironomid stages. Means ( 1 standard error withdifferent letters are statistically different (Bonferroni p < 0.05)and were tested with the GLM for concentrations (Table 2) anda one-way ANOVA for burdens (see text). Sample sizes are inparentheses.

VOL. 42, NO. 24, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 9113

2.7 times more concentrated with MeHg than Orthocladinae.However, sites where Diamesinae adults were abundant didnot always have high MeHg content (e.g., Char and Merettalakes) and it is unclear why this taxonomic group sometimesaccumulated more mercury.

Effect of Metamorphosis on Chironomid δ15N Ratios.We observed differences in the δ15N ratios of chironomidstages associated with extensive restructuring of nitrogen-rich protein tissues during metamorphosis. The δ15N ratiosof larvae ranged from 1.2-6.8‰ among lakes indicatingdifferences in the baseline signature (see the SupportingInformation, Figure 4). On average (mean ( 1 SE), pupaewere enriched in δ15N by 2.1 ( 0.4‰ relative to larvae andemergent adults were enriched by an additional 1.0 ( 0.3‰in δ15N relative to pupae (t-tests, t g 3.6, p e 0.018, n )12-14 sites). Enrichment of δ15N during metamorphosisoccurs in a variety of insect groups without a change in diet(24, 25). Differences in δ13C between larvae and pupae werenot significantly different from zero (t-tests, p > 0.05),confirming that nitrogen isotope patterns in this study werelikely not due to dietary effects. Nitrogen isotope enrichmentoccurs because δ15N-depleted protein is preferentially catab-olized and subsequently eliminated in larval excrement ormeconium of emerging adults (24). Loss of δ15N-depletedexuvium could also result in enrichment during emergence(25).

These δ15N patterns have important implications for foodweb studies in High Arctic lakes. The difference in δ15Nbetween larvae and adults of ∼3‰ is close to the 3.4‰ δ15Nenrichment that occurs between a consumer and its diet(26). As a result, a char that mostly consumes larvae wouldbe depleted in δ15N by 2-3‰ compared to one that consumespupae and adults even though the two individuals have the

same trophic level. Nitrogen isotope ratios have been usedin feeding studies of Arctic char to identify fish withcannibalistic or chironomid diets (7, 27) and to infer trophiclevel for comparison of mercury between fish (7). A δ15Nrange of ∼5‰ is often observed between individual adultArctic char within a lake (7, 8, 27) when in theory, a maximumdifference of 3.4‰ should occur between fish that only eatinvertebrates and those that only eat fish. This broad rangein reported δ15N of Arctic char suggests that differentialconsumption of chironomid stages may commonly occurand affect the δ15N of these fish. Caution is required whenusing nitrogen isotope ratios to interpret trophic positionand dietary sources of MeHg.

Chironomid Stages as Distinct Sources of MeHg. Weestimated MeHg uptake by char feeding on different chi-ronomid stages. Average concentrations of each stage werecorrected for their nutritional quality using caloric content(larvae: 5 kcal g dw-1, pupae and adults: 5.5 kcal g dw-1;9, 13) and assuming an assimilation effeciency of 80% (19).We estimated a transfer to Arctic char of 6.5 ng MeHg kcal-1

of larvae consumed, 9.9 ng MeHg kcal-1 of pupae and 17.2ng MeHg kcal-1 of adult chironomid. As an example scenario,MeHg uptake from a diet of 25% larvae, 50% pupae and 25%adults would be 1.7 times higher than a diet of 100% larvae.

Arctic char may feed on a particular chironomid stagedepending on the size of the fish or the time of year. Forexample, in lakes on Ellesmere Island in the Arctic Archi-pelago, larger char fed selectively on pupae at the lake surfaceduring the period of emergence while smaller char (<20 cm)inhabited very shallow areas, feeding mostly on chironomidlarvae (28). Similarly, adult chironomids were an importantdiet item for Arctic char >17 cm in a Greenland lake(∼15-30% of gut contents) but were little consumed by fishless than that size (29). Chironomid pupae and adults areonly present during emergence from June to August, andlarvae are a more prominent diet item at other times of theyear (8, 10, 29). Differential consumption of chironomid stagesdepending on fish size and season likely affects the amountof MeHg bioaccumulated in Arctic char.

Implications for the Role of Atmospheric MercuryDeposition. This study provides strong evidence that drainagebasin inputs of THg, which are primarily from atmosphericdeposition (12), have less impact on mercury biomagnifi-cation than biological and food web processes in High Arcticfresh waters. Unique ecological conditions in this extremeenvironment may dampen the movement of mercury fromthe atmosphere to aquatic food webs. In spring, snowpackcan receive high loads of inorganic mercury during atmo-spheric depletion events although a significant fraction isreduced and returned to the atmosphere within several days(2). Preliminary findings in Arctic Alaska indicate thatincreased atmospheric deposition due to spring-time mer-cury depletion events has little impact on MeHg accumulationin Arctic food webs (30). Elevated levels of inorganic mercuryare found in High Arctic snowmelt (3, 4) but more than 50%of THg loads from meltwater flow under ice and exit the

TABLE 2. General Linear Model Explaining MeHg Concentrations of Chironomids in Relation to Their Development Stage, MercurySupply, and Habitat Characteristics (model r2 = 0.56, n = 183)

response variable factor effect effect sizea F p

log MeHgconcentration development stage + 0.44 68.0 <0.001water temperature - 0.07 13.2 <0.001log MeHgsediment

b + 0.04 8.2 0.005log MeHgwater + 0.04 6.5 0.012log DA:SA + 0.03 4.9 0.029log DOC - 0.02 4.3 0.040

a partial Eta squared coefficient. b Corrected for OM content.

FIGURE 4. MeHg concentrations of chironomid larvae, pupae,and adults at each of the study sites categorized according todrainage basin type. Means are ( 1 standard error.

9114 9 ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 42, NO. 24, 2008

outflow of lakes without mixing into the water column dueto differences in density (12). Low organic carbon in waterand sediment reduce mercury retention in lakes (31). In situMeHg production is probably limited more by extremeenvironmental conditions than inorganic mercury loading(as discussed earlier), and losses by photodegredation maybe high due very clear water (32). Future research shoulddetermine mercury methylation rates in High Arctic lakes.

The Canadian Arctic Archipelago is undergoing rapidenvironmental change as a result of climate warming, andthe cycling of mercury is expected to increase because ofgreater watershed loadings to lakes, scavenging by algae anddetritial OM, and enhanced MeHg production in sediment(6, 31). A warmer Arctic will also have consequences for foodwebs by stimulating growth rates and altering the speciescomposition of aquatic communities. Interactions betweenmercury supply and food web processes will ultimatelydetermine future biomagnification of MeHg to Arctic char.

AcknowledgmentsThis work was funded by NSERC and FQRNT postgraduatescholarships to J.C. and NSERC Discovery, FQRNT-Equipe,Science Horizon, and Northern Supplement grants to M.A.Logistical support from the Polar Continental Shelf Projectto M.A. is gratefully acknowledged. J.C. and M.A. receivedNSTP funds from INAC. Lake and drainage basin areas weredetermined by Jean Daoust. We thank Marie-Claude Turmeland Richard Carignan for ancillary water quality analyses,Edenise Garcia and Pilipoosie Iqaluk for assistance in thefield, and Dominic Belanger for laboratory assistance. Thisis PCSP contribution number 04908.

Supporting Information AvailableDetails on location, size, and water quality of study sites,organic matter influence on sediment mercury, interannualand drainage basin type differences in chironomid MeHg,δ15N ratios of chironomids (PDF). This material is availablefree of charge via Internet at http://pubs.acs.org.

Literature Cited(1) AMAP Assessment 2002: Heavy Metals in the Arctic; Arctic

Monitoring and Assessment Programme: Oslo, Norway, 2005.(2) Steffen, A.; Douglas, T.; Amyot, M.; Ariyas, P.; Aspmo, K.; Berg,

T.; Bottenheim, S.; Brooks, S.; Cobbett, F.; Dastoor, A. et al. Asynthesis of atmospheric mercury depletion event chemistry inthe atmosphere and snow. Atmos. Chem. Phys. 2008, 8, 1445–1482.

(3) Loseto, L. L.; Lean, D. R. S.; Siciliano, S. D. Snowmelt sourcesof methylmercury to High Arctic ecosystems. Environ. Sci.Technol. 2004, 38, 3004–3010.

(4) St. Louis, V. L.; Sharp, M. J.; Steffen, A.; May, A.; Barker, J.; Kirk,J. L.; Kelly, D. J. A.; Arnott, S. E.; Keatley, B.; Smol, J. P. Somesources and sinks of monomethyl and inorganic mercury onEllesmere Island in the Canadian High Arctic. Environ. Sci.Technol. 2005, 39, 2686–2701.

(5) Braune, B.; Muir, D.; DeMarch, B.; Gamberg, M.; Poole, K.; Currie,R.; Dodd, M.; Duschenko, W.; Eamer, J.; Elkin, B.; Evans, M.;Grundy, S.; Hebert, C.; Johnstone, R.; Kidd, K.; Koenig, B.;Lockhart, L.; Marshall, H.; Reimer, K.; Sanderson, J.; Shutt, L.Spatial and temporal trends of contaminants in Canadian Arcticfreshwater and terrestrial ecosystems: A review. Sci. TotalEnviron. 1999, 230, 145–207.

(6) Hammerschmidt, C. R.; Fitzgerald, W. F.; Lamborg, C. H.;Balcom, P. H.; Tseng, C. M. Biochemical cycling of methylm-ercury in lakes and tundra watersheds of Arctic Alaska. Environ.Sci. Technol. 2006, 40, 1204–1211.

(7) Muir, D.; Wang, X.; Bright, D.; Lockhart, L.; Kock, G. Spatial andtemporal trends of mercury and other metals in landlockedchar from lakes in the Canadian Arctic archipelago. Sci. TotalEnviron. 2005, 351-352, 464–478.

(8) Hobson, K. A.; Welch, H. E. Cannibalism and trophic structurein a high Arctic lake: insights from stable-isotope analysis. Can.J. Fish. Aquat. Sci. 1995, 52, 1195–1201.

(9) Welch, H. E. Ecology of Chironomidae (Diptera) in a polar lake.J. Fish. Res. Board Can. 1976, 33, 227–247.

(10) Johnson, L. The Arctic Charr. In Charrs: Salmonid Fishes of theGenus Salvelinus; Balon, E. K., Ed.; Dr. W. Junk: The Hague, TheNetherlands, 1980.

(11) Oliver, D. R. Adaptations of Arctic Chironomidae. Ann. Zool.Fenn. 1968, 5, 111–118.

(12) Semkin, R. G.; Mierle, G.; Neureuther, R. J. Hydrochemistry andmercury cycling in a High Arctic watershed. Sci. Total Environ.2005, 342, 199–221.

(13) Bliss, L. C. Introduction. In Truelove Lowland, Devon Island,Canada: A High Arctic Ecosystem; Bliss, L. C., Ed.; University ofAlberta Press: Edmonton, AB, Canada, 1977.

(14) Welch, H. E. Emergence of Chironomidae (Diptera) from CharLake, Resolute, Northwest Territories. Can. J. Zool. 1973, 51,1113–1123.

(15) Minns, C. K. Limnology of some lakes in Truelove Lowland. InTruelove Lowland, Devon Island, Canada: A High Arctic Eco-system; Bliss, L. C., Ed.; University of Alberta Press: Edmonton,AB, Canada, 1977.

(16) Watras, C. J.; Back, R. C.; Halvorsen, S.; Hudson, R. J. M.; Morisson,K. A.; Wente, S. P. Bioaccumulation of mercury in pelagic freshwaterfood webs. Sci. Total Environ. 1998, 219, 183–208.

(17) Benoit, J.; Gilmour, C.; Heyes, A.; Mason, R. P.; Miller, C.Geochemical and Biological Controls Over MethylmercuryProduction and Degradation in Aquatic Ecosystems. In Bio-geochemistry of Environmentally Important Trace Elements;Chai, Y., Braids, O. C., Eds.; ACS Symposium Series No. 835;American Chemical Society: Washington, D.C., 2003.

(18) Kelly, C. A.; Rudd, J. W. M.; Holoka, M. H. Effect of pH on mercuryuptake by an aquatic bacterium: implications for Hg cycling.Environ. Sci. Technol. 2003, 37, 2941–2946.

(19) Trudel, M.; Rasmussen, J. B. Predicting mercury concentration infish using mass balance models. Ecol. Appl. 2001, 11, 517–529.

(20) Hamburger, K.; Lindegaard, C.; Dall, P. C. The role of glycogenduring the ontogenesis of Chironomus anthracinus (Chirono-midae, Diptera). Hydrobiologia 1996, 318, 51–59.

(21) Currie, R. S.; Fairchild, W. L.; Muir, D. C. G. Remobilization andexport of cadmium from lake sediments by emerging insects.Environ. Toxicol. Chem. 1997, 16, 2333–2338.

(22) Bartrons, M.; Grimalt, J. O.; Catalan, J. Concentration changesof organochlorine compounds and polybromodiphenyl ethersduring metamorphosis of aquatic insects. Environ. Sci. Technol.2007, 41, 6137–6141.

(23) Barley, E. M.; Walker, I. R.; Kurek, J.; Cwynar, L. C.; Mathewes,R. W.; Gajewski, K.; Finney, B. P. A northwest North Americantraining set: distribution of freshwater midges in relation to airtemperature and lake depth. J. Paleolimnol. 2006, 36, 295–314.

(24) Tibbets, T. M.; Wheeless, L. A.; Martinez del Rio, C. Isotopicenrichment without change in diet: an ontogenetic shift in δ15Nduring insect metamorphosis. Funct. Ecol. 2008, 22, 109–113.

(25) Doi, H.; Kikuchi, E.; Takagi, S.; Shikano, S. Changes in carbonand nitrogen stable isotopes of chironomid larvae during growth,starvation and metamorphosis. Rapid Commun. Mass Spectrom.2007, 21, 997–1002.

(26) Vander Zanden, M. J.; Rasmussen, J. B. Variation in δ15N andδ13C trophic fractionation: Implications for aquatic food webstudies. Limnol. Oceanogr. 2001, 46, 2061–2066.

(27) Guiguer, K. R. R. A.; Reist, J. D.; Power, M.; Babaluk, J. A. Usingstable isotopes to confirm the trophic ecology of Arctic charrmorphotypes from Lake Hazen, Nunavut, Canada. J. Fish Biol.2002, 60, 348–362.

(28) Parker, H. H.; Johnson, L. Population structure, ecologicalsegregation and reproduction in non-anadromous Arctic charr,Salvelinus alpinus (L), in four unexploited lakes in the Canadianhigh Arctic. J. Fish Biol. 1991, 38, 123–147.

(29) Riget, F. F.; Nygaard, K. H.; Christensen, B. Population structure,ecological segregation, and reproduction in a population ofArctic char (Salvelinus alpinus) from Lake Taseruaq, Greenland.Can. J. Fish. Aquat. Sci. 1986, 43, 985–992.

(30) Hammerschmidt, C. R.; Fitzgerald, W. F. Methylmercury in arcticAlaskan mosquitoes: implications for impact of atmosphericmercury depletion events. Environ. Chem. 2008, 5, 127–130.

(31) Outridge, P. M.; Sanei, H.; Stern, G. A.; Hamilton, P. B.; Goodarzi,F. Evidence for control of mercury accumulation rates in CanadianHigh Arctic lake sediments by variations of aquatic primaryproductivity. Environ. Sci. Technol. 2007, 41, 5259–5265.

(32) Hammerschmidt, C. R.; Fitzgerald, W. F. Photodecompositionof methylmercury in an Arctic Alaskan lake. Environ. Sci. Technol.2006, 40, 1212–1216.

ES801619H

VOL. 42, NO. 24, 2008 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 9115