Archives of Environmental Contamination and Toxicology

, Volume 50, Issue 3, pp 398–410

Occurrence of Perfluoroalkyl Surfactants in Water, Fish, and Birds from New York State

Authors

  • Ewan Sinclair
    • Wadsworth Center, New York State Department of Health, and Department of Environmental Health SciencesState University of New York at Albany
  • David T. Mayack
    • New York State Department of Environmental Conservation
  • Kenneth Roblee
    • New York State Department of Environmental Conservation
  • Nobuyoshi Yamashita
    • National Institute of Advanced Industrial Science and Technology (AIST)
    • Wadsworth Center, New York State Department of Health, and Department of Environmental Health SciencesState University of New York at Albany
Article

DOI: 10.1007/s00244-005-1188-z

Cite this article as:
Sinclair, E., Mayack, D.T., Roblee, K. et al. Arch Environ Contam Toxicol (2006) 50: 398. doi:10.1007/s00244-005-1188-z

Abstract

Concentrations of perfluorooctanesulfonate (PFOS) and several other perfluoroalkyl surfactants (PASs) were determined in nine major water bodies (n = 51) of New York State (NYS). These PASs were also measured in the livers of two species of sport fish (n = 66) from 20 inland lakes in NYS. Finally, perfluorinated compounds were measured in the livers of 10 species of waterfowl (n = 87) from the Niagara River region in NYS. PFOS, perfluorooctanoic acid (PFOA), and perfluorohexanesulfonate (PFHS) were ubiquitous in NYS waters. PFOA was typically found at higher concentrations than were PFOS and PFHS. Elevated concentrations of PFOS were found in surface waters of Lake Onondaga, and elevated concentrations of PFOA were found in the Hudson River. PFOS was the most abundant perfluorinated compound in all fish and bird samples. PFOS concentrations in the livers of fishes ranged from 9 to 315 ng/g wet weight. PFOS, PFOA, and PFOSA (perfluorooctanesulfonamide) concentrations in smallmouth and largemouth bass (taken together) caught in remote mountain lakes with no known point sources of PAS contamination were 14 to 207, < 1.5 to 6.1, and < 1.5 to 9.8 ng/g wet weight, respectively. PFOS concentrations in the livers of birds ranged from 11 to 882 ng/g wet weight. PFOS concentrations were 2.5-fold greater (p = 0.001) in piscivorous birds than in non-piscivorous birds. However, PFOA, PFOSA, and PFHS were not found in bird livers. Overall, average concentrations of PFOS in fish were 8850-fold greater than those in surface water. An average biomagnification factor of 8.9 was estimated for PFOS in common merganser relative to that in fish. This study highlights the significance of dietary fish in PFOS accumulation in the food chain. Furthermore, our results provide information on the distribution of PASs in natural waters, fish, and several bird species in NYS.

Perfluoroalkyl surfactants (PASs) are a group of fluorochemicals manufactured for their unique surface-active properties, and have found widespread use in many commercial and industrial applications (Kannan et al.2005). These fully fluorinated compounds are environmentally persistent, globally distributed, and bioaccumulative. Concern over these compounds has grown, as perfluorooctanesulfonate (PFOS), and perfluorooctanoic acid (PFOA), have been routinely measured in environmental matrices, wildlife, and human populations (Giesy and Kannan 2002).

Perfluorooctanesulfonyl fluoride (POSF)-based fluorochemicals have been manufactured by 3M since 1948 (Seacat et al.2002); PFOS is also the final metabolic degradation product of POSF-derived fluorochemicals (Kannan et al.2004). PFOA is an essential wetting agent in the production of polytetrafluoroethylene (PTFE) (Kennedy et al.2004), and there is evidence to suggest that PFOA is also a stable degradation product of fluorotelomer alcohols (FTOHs) (Ellis et al.2004; Wang et al.2005). Other perfluorinated contaminants, such as perfluorohexanesulfonate (PFHS) and perfluorooctanesulfonamide (PFOSA) are less frequently measured in biota than is PFOS, and are found typically at lower concentrations than is PFOS.

The stability of PASs contributes to their persistence in the environment. Of particular concern is their ability to bioaccumulate at successively higher trophic levels of a food chain. PFOS has been shown to bioconcentrate from water to benthic invertebrates by three orders of magnitude, and to bioaccumulate in top predators by 5–20-fold (Kannan et al.2005). Top predators, such as polar bears, mink, bald eagles, and river otters, accumulate PFOS in their livers at microgram per gram concentrations (Kannan et al.2001, 2002b).

In addition to concerns over the toxicological effects that PASs may exert on wildlife, there is a risk to human populations through dietary exposure. Fish can be a significant source of human dietary exposure to perfluorochemicals, because of these chemicals’ ubiquitous occurrence in aquatic environments. For the risk assessment of potential exposure of humans or wildlife to perfluoroalkyl surfactants via fish consumption, information on the concentrations of these compounds in sport fish and forage fish is needed. Because perfluorinated compounds accumulate preferentially in livers rather than in muscle tissues of fish (Giesy and Kannan 2001; Kannan et al.2005), human exposure via consumption of muscle tissues may be minimal. Because the analytical methods are not adequately characterized for the analysis of muscle tissue as they are for liver tissue, and because PASs preferentially accumulate in livers, we analyzed only liver tissues in this study. This should provide an indication of the magnitude of contamination in fish and an upper limit on human exposure via fish, for the study area. Liver concentrations also provide information on exposure of wildlife consumers of fish.

PFOS has been measured at nanogram per gram concentrations in the livers of fish and birds in North America (Kannan et al.2001), Korea and Japan (Kannan et al.2002a), and Japan (Taniyasu et al.2003). Very few studies have examined spatial distributions of PASs in the environment. In this study, we measured concentrations of PFOS, PFOA, PFHS, and PFOSA in surface waters from Lake Ontario, Lake Erie, the Finger Lakes, Lake Onondaga, the Erie Canal, the Hudson River, and Lake Champlain. The sites chosen represent the major water bodies of New York State (NYS). We also measured these perfluorinated compounds in the livers of two species of popular sport fish, smallmouth (Micropterus dolomieu) and largemouth bass (Micropterus salmoides). These fish were taken from 20 inland lakes that are popular with anglers. We also measured concentrations of PASs in 10 species of waterfowl that were collected from hunters or captured at sites on or near the Niagara River. These birds included both migratory and resident species, and they varied in dietary habits from the totally piscivorous common merganser (Mergusmerganser) to the highly herbivorous mallard (Anas platyrhynchos). The objectives of this study were to provide information on the regional distribution of PASs, the magnitude of contamination in NYS surface waters and biota, and the factors that affect PAS distribution in water, fish, and birds.

Materials and Methods

Sample Collection

In July 2004, 51 surface-water grab-samples were collected in 500-mL polypropylene bottles, from nine major water bodies (Figure 1). These samples were refrigerated at 4°C until analysis. Thirty-eight smallmouth and 28 largemouth bass were collected by electroshocking or gill netting or angling, between January 2001 and September 2003, from 20 inland lakes in NYS (Figure 2). Liver samples were removed and stored in precleaned glass jars at −20°C until analysis.
https://static-content.springer.com/image/art%3A10.1007%2Fs00244-005-1188-z/MediaObjects/244_2005_1188_f1.jpg
Figure 1

Map showing PFOS and PFOA concentrations in New York State waters

https://static-content.springer.com/image/art%3A10.1007%2Fs00244-005-1188-z/MediaObjects/244_2005_1188_f2.jpg
Figure 2

Map showing liver PFOS concentration (ww; wet weight) in bass (smallmouth and largemouth) from New York State inland lakes

Liver samples of waterfowl belonging to 10 species were collected from cooperating hunters during the 1994/1995, 1995/1996, and 1999/2000 hunting seasons. With the exception of a ringed-neck duck (Aythya collaris), two hooded mergansers (Lophodytes cucullatus), and four mallards (Anas platyrhynchos) collected during October 1994, birds were collected during the last week of December or first week of January of each season. A black duck (Anas rubripes), surf scoter (Melanitta perpicillata), and two buffleheads (Bucephala albeola) were collected during the 1994/1995 season; an additional bufflehead was collected during the l999/2000 season. Totals of 20 common goldeneyes (Bucephala clangula) and 20 common mergansers (Mergus merganser) were collected during the 1995/1996 and 1999/2000 seasons. Two greater scaup (Aythya marlia) and six lesser scaup (Aythya affinis) were also collected during the 1995/1996 and 1999/2000 seasons, respectively. With the exception of hooded mergansers from the Tonawanda Wildlife Management Area, birds were collected from public or private blinds located on the Niagara River, which flows from Buffalo Harbor on Lake Erie to its mouth at Fort Niagara on Lake Ontario. Birds were collected from Fort Niagara State Park near the mouth of the river, upstream of Niagara Falls, from the east and west shores of Grand Island, Strawberry Island, and from the Buffalo Harbor area. In addition to mallards provided by hunters, we directly collected 27 mallards from local populations. During September 1994, eight immature (hatching-year) mallards were trapped in duck traps baited with corn at three sites on or near the Niagara River upstream of Niagara Falls; Robert Moses Parkway at the Adams Intake; the Little River near Riverside Drive, Cayuga Island; and Scajaquada Creek, Delaware Park, Buffalo. Three immature mallards were recovered from the Tonawanda Creek-Barge Canal after death, likely due to avian botulism, during October 1994. During their flightless period in July 1995, 16 adult mallards (after-hatching-year, males) were driven into pens at three sites on or near the Niagara River upstream of Niagara Falls: the Little River near Riverside Drive, Cayuga Island; Scajaquada Creek, Delaware Park, Buffalo; and Tonawanda Creek-Barge Canal, Tonawanda.

Birds were weighed, and total lengths were measured with head and tail extended. Length from the tip of the upper mandible to the end of the longest rectrix was determined. Gender and age of species other than common merganser were determined according to keys compiled by the Canadian Wildlife and U.S. Fish and Wildlife Services (1977); for common mergansers, we used keys developed by Anderson and Timken (1971). We categorized birds less than 1 year in age (hatching-year, second-year) as immature; however, those collected in December/January were categorized (after-hatching-year, after-second-year) as adult. The minimum age for adults collected in July was approximately 1.2 years; for those collected in December/January, it was approximately 1.6 years. For a number of species (e.g., scaups, mergansers, common goldeneye, bufflehead, scoters) that are not likely to breed until their second season (Bellrose 1980), the adult category may include birds that are not sexually mature. Livers were removed and homogenized using a homogenizer with a stainless-steel rotor-stator (Tissue-Tearor, Biospec Products, Inc.). Homogenized tissue was then stored in pre-cleaned glass jars (I-Chem) at −20°C until analysis.

Dietary Analysis

The gizzards and esophagi/proventriculi of 315 birds, collected at sites on or near the Niagara River between 1993 and 1996, were examined for dietary items. The esophagus and proventriculus were dissected separately from the gizzard to remove their contents. The contents of the esophagus/proventriculus were washed out into a small beaker using 70% ethanol. Material from the gizzard was washed out and passed through a 1-mm mesh sieve. Materials retained by the sieve were washed into a small beaker. Materials were sorted into eight categories: bivalve, univalve, mollusk, crayfish, fish, insect, plant, and unidentified. Bivalves included shell fragments from pelecypod mollusks (clams, mussels, etc.). Univalves included shell fragments from gastropod mollusks (snails, etc.). The category mollusk included shell fragments that could not be categorized further. Presence or absence of materials within each category was recorded for each esophagus/proventriculus or gizzard examined.

Chemical Analysis

Concentrations of PASs in surface waters were measured by a solid-phase extraction method described elsewhere (So et al.2004). To minimize particulates, all water samples were allowed to settle, and an aliquot of 200 mL was carefully decanted into a separate polypropylene bottle for extraction. Briefly, water samples were spiked with 5 ng of perfluorobutanesulfonate (PFBS) as an internal standard. The samples were then passed through preconditioned Oasis HLB (60 mg, 3 cc) cartridges (Waters Corporation, Milford, MA) at a rate of 1 drop/sec. A wash step of 20% methanol in water was applied, and the cartridges were dried completely under vacuum. The target compounds were eluted in 5 mL of methanol into a polypropylene tube and concentrated under nitrogen to a final volume of 1 mL. These extracts were filtered using a 0.2-μm nylon filter into an autosampler vial with polypropylene cap.

Concentrations of PASs in liver tissue were determined by the ion pairing liquid extraction method described elsewhere (Hansen et al.2001; Kannan et al.2001). Briefly, liver samples (1 g) were homogenized in 5 mL of Milli-Q water. A 1-mL aliquot of this homogenate was spiked with 5 ng of PFBS and 5 ng of 13C perfluorooctanoic acid (13C-PFOA) as internal standards. One milliliter of 0.5 M tetrabutylammonium hydrogen sulfate solution, 2 mL of sodium carbonate buffer (0.25 M, pH 10), and 5 mL methyl-tert-butyl ether (MTBE) were added to the sample. After shaking for 30 min, the organic layer was separated by centrifugation, and the extraction was repeated with a further 5 mL of MTBE. The extracts were combined and evaporated to dryness under a gentle flow of nitrogen, before being reconstituted in 1 mL of methanol and vortexed. The extract was filtered through a 0.2-μm nylon filter into an autosampler vial with polypropylene cap.

Separation of PASs was performed using an Agilent 1100 high performance liquid chromatograph (HPLC). Ten microliter of the extracts was injected onto a 50 × 2 mm (5 μm) Keystone Betasil C18 column (Thermo Electron, Bellefonte, PA). A gradient mobile phase of methanol and 2 mM ammonium acetate was used. At a flow rate of 300 μL/min, the mobile phase gradient was ramped from 10% to 25% methanol in 7 min and then to 100% methanol at 10 min, held at 100% methanol for 2 min, and then ramped down to 10% methanol. For quantitative analysis, the HPLC was interfaced with an Applied Biosystems API 2000 tandem mass spectrometer (MS/MS). The MS/MS was operated in electrospray ionization in the negative ion mode. Analyte ions were monitored using multiple reaction monitoring (MRM) mode. Parent and daughter ion transitions monitored for detection of PFOS, PFHS, PFBS, 13C-PFOA, PFOA, and PFOSA were 499 > 99, 399 > 80, 299 > 80, 370 >170, 369 > 169, and 498 > 78, respectively. Quantitation was performed using a linear regression fit analysis weighted 1/x of a single unextracted calibration curve. Seven-point calibration curves were produced from concentrations of 0.1 to 100 ng/mL. The coefficient of determination (r2) for each calibration was > 0.99. Quality-control standards were measured after every 10 samples, to check for instrumental drift. Analysis was stopped and a new calibration curve was run if the quality-control standard was not measured at ±30% of its theoretical value. PFOA was consistently found in procedural blanks and in methanol injections performed between samples. This background PFOA has been shown to be continuously leached from the HPLC system and concentrated on the head of the LC column during column equilibration periods (Tomy et al.2004). However, because this background PFOA signal is consistent, it can be subtracted from the calibration curves and samples. PFOA contamination can also be introduced from fluoropolymer-containing vial caps (Yamashita et al.2004). Polypropylene or aluminum foil caps were used in this study.

All procedural blank peak areas were less than half the determined limit of quantitation (LOQ) for each analyte. The LOQ was estimated as three times the lowest concentration point on the calibration curve, which is accurately measured within ±30% of its theoretical value. The LOQ was 1.5 ng/g for all analytes measured in fish livers, and 7.5 ng/g for analytes measured in bird livers. Matrix spikes were performed for both water and liver tissue samples. These matrices were spiked with 10 ng of each target analyte, and were passed through the whole analytical procedure. Recoveries were then corrected for endogenous analyte levels measured in the unspiked matrix. All analytes were recovered within 100% ±30% from water (Figure 3).
https://static-content.springer.com/image/art%3A10.1007%2Fs00244-005-1188-z/MediaObjects/244_2005_1188_f3.gif
Figure 3

Percent recovery, mean (standard deviation) of PASs from water and liver tissue matrix spikes

PFOA recoveries were highly variable in liver samples. Therefore, the reported values for PFOA in liver tissue are considered semiquantitative. Results were not corrected for recoveries. Concentrations are reported on a wet weight (wt) basis, unless specified otherwise.

Statistical Analyses

The distribution of PAS concentrations in surface waters, and in the livers of fish and birds, was assessed using Shapiro-Wilk’s W test. For normally distributed PAS concentrations, between-group differences were determined using two-tailed t-tests or analysis of variance (ANOVA) with Bonferroni post-hoc test. For PAS concentrations that were not normally distributed, between-group differences were determined using Kruskal-Wallis H tests and Mann-Whitney U-tests with Bonferroni correction. All statistical analyses were performed using SPSS 11.0 statistical package.

Results and Discussion

Surface Waters

PFOS was found at all but one of the surface-water sites at concentrations >0.8 ng/L (LOQ) (Figure 1, Table 1). Surface water concentrations did not follow a normal distribution. Log-transformation did not normalize the data and was therefore not applied. For comparisons, the samples were grouped according to the water body from which the sample was taken. With the exception of Lake Onondaga, PFOS concentrations in surface waters showed little variation and ranged from < 0.8 to 30 ng/L. PFOS concentrations in the finger Lakes were significantly lower than those measured in the surface waters of Lake Ontario (Table 2). Three water samples taken from Lake Onondaga in Syracuse contained 198 ng/L, 746 ng/L, and 1090 ng/L PFOS, the highest PFOS concentrations that have been measured in NYS waters. Lake Onondaga is a Superfund site and is influenced by sources from several industries located along the Lake. In addition, the Metropolitan Syracuse sewage treatment plant contributes 20% of the Lake’s annual inflow (Onondaga County Department of Water Environment Protection website; http://www.lake.onondaga.ny.us/ol11.htm), thus highlighting the importance of wastewater effluent as a potential source of PAS contamination. PFOS concentrations ranged from 3 to 30 ng/L in Lake Ontario, and from 3 to 7 ng/L in Lake Erie. These values are approximately fivefold lower than those previously reported for Lake Ontario and Lake Erie (Boulanger et al. 2004). PFOS contamination of NYS waters is comparable with that in Japan, where most waters contained < 2.5 to 21 ng/L PFOS (Taniyasu et al. 2003). With the exception of Lake Onondaga, NYS waters contained PFOS concentrations slightly greater than the 1.9 to 3.9 ng/L reported for certain rivers in Michigan (Kannan et al. 2005).
Table 1

Concentrations (ng/L) of PFOS, PFHS, PFOA, and PFOSA in surface waters of New York State

Location

n

PFOS

PFHS

PFOA

Lake Ontario

13

4.9 (2.9–30)

1.4 (1.2–2.8)

21 (18–34)

Niagara River

3

5.5 (3.3–6.7)

1.2 (1.2–1.4)

19 (18–22)

Lake Erie

3

3.0 (2.8–5.5)

1.2 (1.2–1.6)

15 (13–27)

Finger Lakes

13

1.6 (1.3–2.6)

0.9 (0.7–1.3)

14 (11–20)

Lake Onondaga

3

756 (198–1090)

7.4 (4.2–8.5)

49 (39–64)

Lake Oneida

1

3.5

0.9

19

Erie Canal

3

6.4 (5.7–13)

2.6 (2.5–5.6)

30 (25–59)

Hudson River

8

1.7 (1.5–3.4)

0.9 (0.7–1.6)

35 (22–173)

Lake Champlain

4

2.7 (0.8–7.7)

1.3 (0.5–1.6)

24 (10–46)

Field Blanks

2

<0.8

<0.5

<2.5

Values are median (min–max). PFOSA was not found at a detection limit of 2.5 ng/L in any sample.

Table 2

Mann–Whitney U test comparisons of surface water PFOS, PFOA, and PFHS concentrations among New York State water bodies

 

Lake Ontario

Niagara River

Lake Erie

Finger Lakes

Lake Onondaga

Erie Canal

Hudson River

PFOS

       

  Niagara River

0.704

      

  Lake Erie

0.146

0.400

     

  Finger Lakes

0.001a

0.004

0.004

    

  Lake Onondaga

0.004

0.100

0.100

0.004

   

  Erie Canal

0.364

0.400

0.100

0.004

0.100

  

  Hudson River

0.001a

0.024

0.048

0.053

0.012

0.012

 

  Lake Champlain

0.130

0.400

0.857

0.202

0.057

0.229

0.570

PFOA

       

  Niagara River

0.189

      

  Lake Erie

0.296

0.700

     

  Finger Lakes

0.001a

0.018

0.233

    

  Lake Onondaga

0.004

0.100

0.100

0.004

   

  Erie Canal

0.039

1.000

0.200

0.004

0.400

  

  Hudson River

0.006

0.012

0.133

0.001a

0.776

1.000

 

  Lake Champlain

1.000

1.000

1.000

0.862

0.114

0.629

0.283

PFHS

       

  Niagara River

0.111

      

  Lake Erie

0.439

1.000

     

  Finger Lakes

0.001a

0.014

0.014

    

  Lake Onondaga

0.004

0.100

0.100

0.004

   

  Erie Canal

0.014

0.100

0.100

0.004

0.200

  

  Hudson River

0.003

0.085

0.048

0.697

0.012

0.012

 

  Lake Champlain

0.350

0.857

1.000

0.202

0.057

0.057

0.368

a Denotes a significant difference at α = 0.05 after Bonferroni correction (α = 0.05/28 = 0.002).

PFOA concentrations were > 2.5 ng/L (LOQ) in all water samples, and were greater than PFOS concentrations in most locations (Figure 1, Table 1). Surface water concentrations of PFOA ranged from 10 to 173 ng/L. PFOA concentrations were lowest in the Finger Lakes, and greatest in the Hudson River (Table 2).

PFOA concentrations ranged from 20 to 34 ng/L in Lake Ontario, and from 20 to 27 ng/L in Lake Erie. These values are approximately half those previously measured in Lake Ontario and Lake Erie (Boulanger et al. 2004). Although PFOS concentrations were low in the upper Hudson River water samples, concentrations of PFOA (20 to 173 ng/L) were elevated. This suggests that the most significant source of PAS to the upper Hudson River is enriched in PFOA relative to PFOS. In contrast, Lake Onondaga in Syracuse contained concentrations of PFOS that were elevated relative to the PFOA concentrations. This pattern may be indicative of different sources of PAS contamination. Specifically, PFOA contamination, in the absence of PFOS contamination, suggests sources arising from the use of fluoropolymers, or telomer alcohols. The use of POSF-based fluorochemicals in surface treatments or firefighting foams would explain elevated PFOS contamination. The concentrations of PFOA measured in Western NYS waters were slightly greater than the 15 ng/L PFOA measured in the Raisin River in Michigan (Kannan et al. 2005).

PFHS was measured in all but one sample at > 0.5 ng/L. Concentrations of PFHS were consistently low and ranged from 0.7 to 2.0 ng/L, with the exception of Lake Onondaga and the Erie Canal, which contained 4.2 to 8.5 ng/L and 2.5 to 5.6 ng/L, respectively. PFOSA was not found at > 2.5 ng/L in any water sample.

Overall, PFOS and PFOA were ubiquitous in NYS surface waters at low nanograms per liter concentrations. Lake Onondaga was particularly highly contaminated with PFOS, whereas Eastern NYS waters, including the Erie Canal, Lake Champlain, Lake Onondaga, and the Hudson River, contained elevated PFOA concentrations; such a pattern may be explained by a greater industrial use of fluoropolymers and telomer-alcohols in this region.

Fish

PFOS was the most abundant perfluorinated compound in fish liver, and was measured in all 66 of the fish liver samples analyzed. PFOS concentrations in all of the fish liver samples ranged from 9 to 315 ng/g wet wt (Tables 3a and 3b, Figure 2). When log-transformed (base e), PFOS concentrations, in both bass species, were found to be normally distributed (Shapiro-Wilk, p > 0.05). For statistical comparisons, liver PFOS concentrations were log-transformed. The livers of largemouth bass and smallmouth bass contained PFOS concentrations of 9 to 315 ng/g and 10 to 120 ng/g wet wt, respectively.
Table 3a

Concentrations (ng/g wet weight) of PFOS, PFOA, and PFOSA in the livers of smallmouth bass (SMB) and largemouth bass (LMB) from remote inland lakes (Adirondack) in New York State

Location (county)

Fish species (n)

Year caught

Gender

 

Body weight (g)

Length (cm)

PFOS

PFOA

PFOSA

Soft Maple Dam (Lewis)

SMB (3)

2003

3F

Mean

706

385

79 (30)

3.1 (18)

6.1 (2.8)

    

Median

  

73

2.4

5.5

    

Range

511–833

340–415

41–114

<1.5–5.2

3.1–10.0

Effley Falls (Lewis)

SMB (3)

2003

3F

Mean

792

398

66 (30)

3.2 (0.8)

3.7 (1.5)

    

Median

  

61

3.0

3.3

    

Range

750–825

383–410

42–109

2.0–3.9

1.6–5.4

Meacham Lake (Franklin)

SMB (3)

2001

3F

Mean

1250

438

32 (12)

3.5 (0.9)

4.8 (0.9)

    

Median

  

29

3.4

4.8

    

Range

900–1450

383–476

16–47

2.4–4.6

3.8–6.1

Polliwog Pond (Franklin)

SMB (3)

2002

3F

Mean

835

382

40 (12)

3.9 (0.8)

5.6 (1.1)

    

Median

  

39

3.8

5.5

    

Range

538–1058

344–412

32–57

3.1–5.0

4.0–6.4

Tupper Lake (Franklin)

SMB (3)

2001

3M

Mean

601

361

98 (28)

1.5 (1.0)

4.0 (1.9)

    

Median

  

93

1.2

3.6

    

Range

404–800

323–393

58–120

<1.5–2.9

2.6–6.7

Star Lake (St Lawrence)

SMB (2)

2003

2F

Mean

194

253

57 (12)

2.8 (2.0)

4.5 (0.8)

    

Median

  

56

1.9

4.4

    

Range

185–202

245–260

45–69

<1.5–4.8

3.6–5.3

 

LMB (1)

2003

F

Mean

285

291

207

2.3

7.6

    

Median

  

207

2.3

7.6

    

Range

     

Sylvia Lake (Jefferson)

SMB (3)

2003

2F, 1M

Mean

1528

448

38 (27)

2.1 (2.2)

14 (0.9)

    

Median

  

30

1.4

1.3

    

Range

334–2200

306–542

14–75

<1.5–4.9

<1.5–1.9

Payne Lake (Jefferson)

LMB (3)

2003

2F, 1M

Mean

309

287

30 (8)

5.2 (0.9)

<1.5

    

Median

  

29

5.1

 
    

Range

271–355

272–296

21–41

4.0–6.1

<1.5

M, male; F, female. Values in parentheses indicate standard deviation. PFHS was not detected in any fish at a detection limit of 1.5 ng/g wet weight.

Table 3b

Concentrations (ng/g wet weight) of PFOS, PFOA, and PFOSA in the livers of smallmouth bass (SMB) and largemouth bass (LMB) from several inland lakes in New York State

Location (county)

Fish species (n)

Year caught

Gender

 

Body weight (g)

Length (cm)

PFOS

PFOA

PFOSA

Otscgo Lake(Otsego)

SMB (4)

2001

2F, 2M

Mean

890

386

22 (8)

2.8 (0.8)

2.3 (1.6)

    

Median

  

20

2.7

1.8

    

Range

630–1210

360–422

10–29

1.5–3.7

<1.5–4.9

Goodyear Lake(Otsego)

LMB (3)

2001

3F

Mean

1560

454

38 (25)

4.6 (0.7)

<1.5

    

Median

  

28

4.6

<1.5

    

Range

1490–1660

445–466

9–71

3.8–5.5

 
 

SMB (3)

2001

3F

Mean

1053

417

46 (18)

2.9 (1.8)

<1.5

    

Median

  

43

2.2

<1.5

    

Range

 

850–1210

390–431

31–71

<1.5–5.2

Canadarago(Otsego)

LMB (3)

2001

3F

Mean

1177

404

58 (49)

5.3 (0.2)

<1.5

    

Median

  

39

5.3

 
    

Range

640–1870

347–474

14–126

5.0–5.5

<1.5

Canada Lake(Fulton)

SMB (3)

2003

3M

Mean

835

377

76 (15)

4.5 (2.2)

9.4 (1.9)

    

Median

  

75

3.7

9.2

    

Range

400–1093

294–435

58–95

1.7–7.7

6.9–11.4

 

SMB (2)

2001

2F

Mean

1221

444

56 (19)

2.0 (1.3)

5.5 (0.0)

    

Median

  

55

1.6

5.5

    

Range

1165–1276

432–455

39–77

<1.5–3.3

5.5

Willis Lake(Hamilton)

LMB (1)

2001

1F

Mean

1193

390

23

5.6

<1.5

    

Median

  

23

5.6

<1.5

    

Range

     

Rock Pond(Hamilton)

LMB (3)

2003

3F

Mean

1317

436

46 (11)

3.0 (1.6)

<1.5

    

Median

  

45

2.3

<1.5

    

Range

750–1800

370–483

32–58

<1.5–4.3

 

Dunham Reservoir(Rensselear)

LMB (3)

2003

2F, 1M

Mean

653

356

21 (7)

4.5 (0.8)

<1.5

    

Median

  

20

4.4

 
    

Range

580–760

331–377

16–32

3.3–5.2

<1.5

 

SMB (3)

2003

3M

Mean

837

387

70 (29)

3.7 (0.2)

2.4 (0.4)

    

Median

  

64

3.7

2.3

    

Range

651–1094

364 –429

33–104

3.5–3.9

1.8–2.8

Cuba Lake(Allegany)

LMB (1)

2003

M

Mean

758

345

16

4.4

<1.5

    

Median

  

16

4.4

<1.5

    

Range

     

Lake Huntington(Sullivan)

LMB (3)

2003

1F, 2M

Mean

323

293

102 (28)

4.5 (0.7)

<1.5

    

Median

  

99

4.9

 
    

Range

296–351

289–298

79–142

4.3–5.2

<1.5

Rio Reservoir(Sullivan)

SMB (3)

2003

3F

Mean

926

406

93 (6)

3.5 (0.4)

8.7 (1.1)

    

Median

  

93

3.5

8.6

    

Range

794–1089

390–431

86–98

3.1–4.0

7.7–10.3

Loch Sheldrake(Sullivan)

LMB (3)

2003

3F

Mean

1310

412

56 (16)

4.9 (0.4)

2.2 (0.8)

    

Median

  

54

4.9

2.1

    

Range

1150–1420

398–429

45–78

4.3–5.2

1.7–3.3

Swan Pond(Suffolk)

LMB (4)

2003

4M

Mean

357

295

282 (35)

4.3 (1.4)

2.9 (16)

    

Median

  

279

4.1

2.4

    

Range

171–526

225–335

224–315

3.4–6.7

<1.5–5.3

M, male; F, female. Values in parentheses indicate standard deviation. PFHS was not detected in any fish at a detection limit of 1.5 ng/g wet weight.

These concentrations of PFOS were comparable to those reported in the livers of two Great Lakes species, chinook salmon and lake whitefish, which measured 32 to 173 ng/g and 33 to 81 ng/g, respectively (Kannan et al.2005). The range of PFOS concentrations in our study is also comparable to the 159 ng/g and 309 ng/g PFOS concentrations reported in the livers of two largemouth bass from Lake Biwa, Japan (Taniyasu et al.2003). PFOS was reported in the muscle tissue of smallmouth bass from the Raisin River, Michigan, at 2 to 41 ng/g wet wt. However, because perfluorinated alkylated surfactants preferentially concentrate in liver tissue, a direct comparison between liver and muscle cannot be made (Kannan et al.2005); generally, concentrations in muscle tissues are expected to be much lower than in liver.

The concentrations of PFOS in fish livers showed some spatial variation (Figure 2). A difference in sources of PAS contamination to these lakes is most likely an important source of the variation. For example, the highest concentrations of PFOS were found in the livers of largemouth bass taken from Swan Pond. The livers of fish from Swan Pond contained significantly higher PFOS concentrations than did livers of fish from all other lakes [analysis of variance (ANOVA) with Bonferroni post-hoc test, p < 0.05] except Tupper Lake, Star Lake, Rio Reservoir, Soft Maple Dam, Lake Huntington, and Canada Lake. Swan Pond is located near an urban area (Long Island near New York City) and is likely to be strongly influenced by industrial and commercial discharges.

In order to assess the effects of body weight, length, gender, species, or temporal changes on the accumulation of PASs, we must minimize spatial influences. Tupper Lake, Meacham Lake, Polliwog Pond, Star Lake, Sylvia Lake, Effley Falls, Payne Lake, and Soft Maple Dam, are all remote lakes located in the Adirondack Mountain region. Because these lakes are rural, and not likely influenced by any point sources of PAS contamination, the concentrations of PASs in fish from these lakes should reflect background levels in NYS. Comparison of concentrations of PASs found in the livers of fish from these remote lakes will allow us to more clearly examine the effects of species, gender, and physical parameters, with minimal influence from changes in point sources of exposure.

In these remote lakes, PFOS was measured from 14 to 207 ng/g, wet wt, in the livers of fish (Table 3a). There was no significant difference in PFOS concentrations between smallmouth bass and largemouth bass (p = 0.886). This is expected, because the dietary habits of the two species are similar, with small fishes, crayfish, and insects as preferred prey items. Because so few largemouth bass are represented in these remote lakes (n = 4), further comparisons have been restricted to smallmouth bass (n = 20). No significant difference in PFOS concentrations was found between 2001 and 2003 (Figure 4). Hepatic concentrations of PFOS were lower in the livers of female smallmouth bass than in the livers of males (Figure 4). Large concentrations of PFOS in whitefish and brown trout fish eggs have recently been reported, providing evidence of oviparous elimination in female fish (Kannan et al.2005). However, the present study was not specifically designed to test such a relationship, and the subset analyzed here included only a limited number of smallmouth bass (n = 20). Liver concentrations of PFOS were significantly inversely correlated with body weight and body length in smallmouth bass (Table 4). This is in contrast with the concentrations of organochlorine compounds, which were positively correlated with body weight in smallmouth bass (Henry et al.1998). However, unlike organochlorines, which associate with lipids, perfluorinated surfactants associate strongly with proteins, and their residence in biota is prolonged through enterohepatic circulation (Kannan et al.2002b). The observed negative correlation of hepatic PFOS concentration with body weight and length could be explained by growth dilution, in which the rate of tissue elaboration exceeds the rate of PFOS accumulation.
https://static-content.springer.com/image/art%3A10.1007%2Fs00244-005-1188-z/MediaObjects/244_2005_1188_f4.gif
Figure 4

Gender- and time-related differences in PFOS, PFOA, and PFOSA in the livers of smallmouth bass taken from remote lakes. Of the 20 smallmouth bass, 3 were caught in 2002 and were excluded from temporal comparisons. Asterisk denotes a significant difference at α = 0.05

Table 4

Regressions of hepatic PFOS, PFOA, and PFOSA with body weight and length of smallmouth bass taken from remote lakes

PAS

Dependent variable

Regression coefficient

Intercept

Coefficient of determination R2

Observed probability (p)

Ln (PFOS)

Body weight

−0.651

4.56

0.424

0.002a

 

Body length

−0.513

5.58

0.263

0.021a

Ln (PFOA)

Body weight

−0.011

1.29

0.001

0.971

 

Body length

−0.086

1.45

0.007

0.771

Ln (PFOSA)

Body weight

−0.359

1.65

0.129

0.120

 

Body length

−0.234

2.02

0.055

0.320

Notes: Body weights ranged from 185 to 2200 g (median = 800 g). Body lengths ranged from 245 to 542 mm (median = 392 mm).

a Denotes a significant regression at α = 0.05.

PFOA was present in 90% of the fish liver samples, at concentrations ranging from <1.5 to 7.7 ng/g wet wt (Tables 3a and 3b). When log-transformed (base e), PFOA concentrations, in both bass species, were found to be normally distributed (Shapiro-Wilk, p > 0.05). For statistical comparisons, liver PFOA concentrations were therefore log-transformed. PFOA concentrations ranged from < 1.5 to 6.7 ng/g and from < 1.5 to 7.7 ng/g wet wt in largemouth bass and smallmouth bass, respectively. No significant spatial variation was observed in fish taken from different lakes (ANOVA with Bonferroni post-hoc tests, p > 0.05). Within the remote lakes, concentrations of PFOA ranged from < 1.5 to 5.2 ng/g and from 2.3 to 6.1 ng/g wet wt in the livers of smallmouth bass and largemouth bass, respectively (Table 3a). No species-related difference in PFOA accumulation was observed (p = 0.143). Within smallmouth bass, no temporal or gender-related differences in PFOA were observed (Figure 4). The accumulation of PFOA was not affected by either body weight or body length (Table 5).
Table 5

Concentrations of PFOS (ng/g wet weight) in the liver of several species of waterfowl collected from the Niagara River region, New York

     

PFOS

Bird species

n

Gender

Body weight (g)

Length (cm)

Mean

Median

Common merganser (Mergus merganser)

20

6F, 14M

1706 [1252–2020]

64 [54–78]

441 (154)

409 [146–715]

Hooded merganser (Lophodytes Cucullatus)

2

1F, 1M

585 [540–630]

44 [44–44]

35 (24)

26 [11–60]

Bufflehead (Bucephala albeola)

3

3M

502 [465–530]

36 [35–36]

635 (281)

550 [242–882]

Mallard (Anas platyrhynchos)

31

3F, 27M

1237 [1010–1510]

56 [50–61]

172 (124)

130 [28–425]

Surf scoter (Melanitta perspicillata)

1

1M

1170

49

28 (0)

28

Black duck (Anas rubripes)

1

1M

1420

56

204 (0)

204

Common goldeneye (Bucephala clangula)

20

5F, 15M

1040 [784–1204]

45 [41–48]

204 (119)

176 [73–505]

Greater scaup (Aythya marila)

2

1F, 1M

1180 [1140–1240]

46 [45–46]

82 (24)

79 [59–106]

Lesser scaup (Aythya affinis)

6

4F, 2M

896 [756–1125]

41 [39–43]

148 (65)

131 [49–240]

Ring-neck (Aythya collaris)

1

1M

520

41

16 (0)

16

M, male; F, female. Values in parentheses indicate standard deviation, whereas those in brackets indicate the range. PFHS, PFOA, and PFOSA were not detected in any bird at >7.5 ng/g wet weight.

PFOSA was present in 62% of the fish liver samples at concentrations ranging from <1.5 to 11.4 ng/g wet wt (Tables 3a and 3b). PFOSA concentrations ranged from < 1.5 to 7.6 ng/g and < 1.5 to 11.4 ng/g wet weight in largemouth bass and smallmouth bass, respectively. PFOSA liver concentrations were significantly (p < 0.05; ANOVA with Bonferroni post-hoc tests) higher in fish from Rio Reservoir and Canada Lake, than in fish from Otsego Lake, Cuba Lake, Loch Sheldrake, Sylvia Lake, and Swan Pond. PFOSA was not found in any fish from Lake Huntington, Dunham Reservoir, Payne Lake, Rock Pond, Goodyear Lake, or Lake Canadarago. PFOSA was measured in the livers of 79% of the fish taken from remote lakes and ranged from < 1.5 to 9.8 ng/g and < 1.5 to 7.6 ng/g, in smallmouth bass and largemouth bass, respectively (Table 3a). Largemouth bass contained significantly lower concentrations of PFOSA (p = 0.03) than did smallmouth bass. Within smallmouth bass, no temporal or gender-related differences in PFOSA were observed (Figure 4). PFOSA was not correlated with body weight or length in either fish species (Table 4). PFHS was not measured at > 1.5 ng/g (LOQ) in any of the fish.

Birds

PFOS was the most abundant perfluorinated compound measured in livers from waterfowl. All bird livers contained PFOS, at concentrations ranging from 11 to 882 ng/g wet wt (Table 5). The greatest hepatic concentrations of PFOS were found in bufflehead and common mergansers. The lowest concentrations of PFOS were found in the livers of the ring-neck, surf scoter, and two hooded mergansers. When log-transformed, PFOS concentrations in mallard, common goldeneye, and common merganser were found to be normally distributed (Shapiro-Wilk test, p > 0.05). Therefore, PFOS concentrations were log-transformed for statistical comparisons. The sample sizes for the remaining species of bird were too small to be tested for normality, and we assumed a log-normal distribution of PFOS concentrations.

The birds analyzed in this study were collected during two hunting seasons, 1994/1996 and 1999/2000. Before making any comparisons between species collected from different years, it was important that we assess the temporal differences in PAS concentrations. Of the 10 species of birds studied, only bufflehead, common goldeneye, and common merganser were collected during both periods. None of these three species showed significant temporal changes in PFOS concentrations (Figure 5). This comparison is most meaningful for the common merganser, which is represented equally in the two periods. The similarity in concentrations of PFOS found in common mergansers captured during the two periods provides evidence that this species experienced comparable exposures to PFOS in 1994 and in 2000. This suggests that any observed differences in PFOS concentrations between the bird species measured in this study are not significantly affected by temporal changes in PFOS exposure. Differences in PFOS concentrations observed between species may be better explained in terms of dietary and migratory behaviors.
https://static-content.springer.com/image/art%3A10.1007%2Fs00244-005-1188-z/MediaObjects/244_2005_1188_f5.gif
Figure 5

Age, gender, and time-related differences in hepatic PFOS concentrations for common merganser, common goldeneye, and mallards. Asterisk denotes a significant difference at α = 0.05

The birds analyzed in this study were examined to assess the dietary behavior of each species. This method provides information on the dietary behavior of each bird in the few days before its capture, and this information is combined with previous documentation of the dietary habits for these birds. Common mergansers are known to be piscivorous (Bellrose 1980); this was confirmed in our examination (Table 6). Hooded mergansers are also piscivorous birds and, although only two were examined here, fish was found in both birds. Mallards are considered non-piscivorous, and plant matter and mollusks appeared to dominate the diet of the 53 birds examined. Only one black duck was examined. Black ducks are known to have a diet similar to mallards, with plant matter as the preferred diet. Common goldeneyes were found to feed on plant matter, mollusks, and crayfish. Both the greater and lesser scaups were found to have been feeding mainly on mollusks and plant matter. Only one bufflehead was examined. Buffleheads are known to feed heavily on animal matter such as mollusks, crayfish, and fish, with fish being a preferred prey item during winter periods (Bellrose 1980). Ring-necks and surf scoters are deep-water benthivores with a diet similar to that of scaups. Interestingly, surf scoters winter in coastal regions and are not common in the Niagara region. Opportunistic feeding may explain why fish was found in this typically non-piscivorous bird.
Table 6

The percentage of esophagi/proventriculi (Eso/Pro) and gizzards that contain dietary items, and the percentage of item-containing organs that contain a specific category of dietary item, for waterfowl collected within the vicinity of the Niagara River, 1993/1996

 

C. Goldeneye

Mallard

C. Merganser

H. Merganser

Bufflehead

 

Gizzard

Eso/Pro

Gizzard

Eso/Pro

Gizzard

Eso/Pro

Gizzard

Eso/Pro

Gizzard

Eso/Pro

Organs examined % with items

160

157

52

53

17

18

2

2

1

1

 

92

29

100

38

76

30

50

100

0

100

Dietary item %

          

  Bivalve

31

20

33

25

6

0

0

0

0

0

  Univalve

38

7

31

20

0

0

0

0

0

0

  Mollusk

16

22

0

20

0

0

0

0

0

0

  Crayfish

55

47

0

0

0

0

0

0

0

0

  Insect

2

0

0

0

0

0

0

0

0

0

  Fish

1

2

0

0

35

100

100

100

0

0

  Plant

78

64

100

100

65

40

0

0

0

100

  Unidentified

10

7

0

0

12

0

0

0

0

0

 

Lesser Scaup

Greater Scaup

Ringneck

S. Scoter

Black Duck

 

Gizzard

Eso/Pro

Gizzard

Eso/Pro

Gizzard

Eso/Pro

Gizzard

Eso/Pro

Gizzard

Eso/Pro

Organs examined

14

13

61

61

1

0

1

1

1

0

% with items

93

15

97

11

100

0

100

100

100

0

Dietary Item %a

          

  Bivalve

69

50

30

0

0

N/A

100

100

100

N/A

  Univalve

100

50

90

83

0

N/A

0

0

100

N/A

  Molluskb

8

0

3

17

0

N/A

0

0

0

N/A

  Crayfish

0

0

0

0

0

N/A

0

0

0

N/A

  Insect

0

0

2

0

0

N/A

0

0

0

N/A

  Fish

0

0

0

0

 

N/A

0

100

0

N/A

  Plant

77

50

39

83

0

N/A

100

0

100

N/A

  Unidentified

15

0

36

0

100

N/A

0

0

0

N/A

a Percentages may sum to greater than 100% because an organ may contain items from more than one category.

b Includes those items that were identified as mollusks but could not be further categorized as bivalve (pelecypod mollusk) or univalve (gastropod mollusk).

Grouped as piscivorous birds, the common merganser, hooded merganser, and bufflehead contained significantly (p = 0.001) greater PFOS concentrations, by a factor of 2.5, than did the non-piscivorous birds (i.e., all other species sampled). The hooded mergansers contained low PFOS concentrations, from 11 to 60 ng/g, wet wt, which was unexpected for a piscivorous bird. However, hooded mergansers were collected from a wildlife refuge remote from the Niagara River. It is possible that in this area, the hooded mergansers were less exposed to PFOS than if they had been collected from the Niagara River. PFOS concentrations were 2.4-fold greater (p = 0.001) in the common merganser than in the common goldeneye. However, this difference may be overestimated, because the common mergansers do not migrate far from the Niagara region, whereas the common goldeneye migrate from the Canadian Northwest Territories to winter on the Niagara River. Common goldeneye may thus be influenced by a Canadian diet, less contaminated with perfluorinated compounds. This difference may also relate to dietary differences between these two species. PFOS concentrations were 4.2-fold greater (p = 0.001) in the livers of common merganser than in the livers of mallard. Both of these species are local to the Niagara River region, so this comparison provides a more accurate reflection of the influence of fish, in the diet of common merganser, on PFOS levels. PFOS concentrations found in the livers of common goldeneye were not significantly different from those found in mallard (p = 0.112).

Significantly greater concentrations of PFOS were found in the livers of adult mallards than in livers of juvenile mallards (Figure 5). This suggests accumulation of PFOS over time. In contrast, concentrations of PFOS were not significantly different between juvenile and adult birds of common goldeneyes and common mergansers. Lack of consistent, gender-related difference in accumulation of perfluorochemicals in birds could be explained by excretion and elimination of these compounds (Kannan et al.2005).

A significant gender-related difference in PFOS concentrations was observed only in common mergansers, for which concentrations were greater in the livers of males than females (Figure 5). As discussed earlier, females have been shown to have an additional elimination route for PFOS, through oviparous transfer (Giesy and Kannan 2001; Kannan et al.2005). PFOS oviparous transfer has been shown to occur in mallard in vivo studies (Newsted et al.2004). Physical dimorphism between male and female common mergansers is greater than that for other waterfowl species, suggesting that greater physiological differences (e.g., metabolism) may exist between sexes of common mergansers but not species with less pronounced dimorphism. Lack of gender-related difference in PFOS concentrations in species other than common merganser is not clear.

Liver concentrations of PFOS were significantly positively correlated with body weight in the common merganser, but not in the mallard or the common goldeneye (Figure 6). Liver PFOS concentrations were not significantly correlated with body length in any of these birds.
https://static-content.springer.com/image/art%3A10.1007%2Fs00244-005-1188-z/MediaObjects/244_2005_1188_f6.gif
Figure 6

Regressions of hepatic PFOS with body weight and length for mallard, common merganser, and common goldeneye. Asterisk denotes a significant regression at α = 0.05

The concentrations of PFOS measured in the birds in this study were comparable those reported elsewhere. PFOS was found in the liver of a mallard from Japan at 493 ng/g, and also in pintail ducks from Japan at levels of 239 to 497 ng/g (Taniyasu et al.2003). A similar range of 170 to 650 ng/g PFOS was reported in the livers of common cormorants from Japan (Kannan et al. 2002).

The measured concentrations of PFOS in the livers of the mallards in our study (28 to 425; median =130 ng/g) were two orders of magnitude lower than the 21.4 μg/g threshold concentration for PFOS in the livers of mallards determined in an acute in vivo study (Newsted et al.2004). An acute-to-chronic application factor (AF) of 0.1 is often used to account for the effects of long-term exposure to environmental contaminants in the wild (Newell et al.1987). The concentrations of PFOS measured in our mallards were 1.6-fold lower than the AF-adjusted threshold of 2.1 μg/g. At concentrations exceeding this threshold, mallards would be expected to show reductions in body weight. PFOA, PFOSA, and PFHS were not measured at more than 7.5 ng/g in any of the birds.

Bioconcentration and Biomagnification

Although water, fish, and birds were collected from different areas and at different times, a general estimation of bioconcentration factor (BCF) and biomagnification factor (BMP) could be made; such information is necessary for risk assessments. PFOS was measured at levels ranging from 3 to 7 ng/L (median = 5.5 ng/L) in the waters of the Niagara region. Hepatic concentrations of PFOS in smallmouth and largemouth bass taken from remote NYS lakes ranged from 14 to 207 ng/g (median = 46 ng/g); PFOS concentration in the Niagara River fish is assumed to be in this range. Based on this, a BCF of 8850 between the livers of fish and the surface waters could be estimated. This compares well with the average BCF of 8540 reported for Japanese coastal fish (Taniyasu et al.2003).

As discussed earlier, we have shown that temporal effects have a minimal influence on the levels of PFOS in fish and birds analyzed in this study (Figures 4 and 5). For an estimation of the BMF of PASs in birds relative to fish, spatial variations can be minimized by comparison of concentrations found in the livers of common mergansers, a species limited in migration. Again, based on the smallmouth bass and largemouth bass taken from remote lakes, we can estimate PFOS concentrations in fish to be at least 14 to 207 ng/g (median = 46 ng/g) in the livers of fish from the Niagara River. PFOS was measured ranging from 146 to 715 ng/g (median = 409) in the livers of common mergansers in the Niagara region. Based on this, an average BMF of 8.9 for PFOS could be estimated in common mergansers, relative to fish. A BMF of 9.0 was reported for glaucous gulls:arctic cod (Tomy et al.2004). A four-to-five fold accumulation of PFOS was calculated for bald eagle:fish from Michigan (Kannan et al.2005). These results highlight the significance of PFOS accumulation from dietary fish. The bird–fish comparison is limited to mergansers that consume bass of the size collected in this study.

PFOA was measured at levels ranging from 18 to 22 ng/L (mean = 19 ng/L) in the waters of the Niagara region. Based on the concentrations in smallmouth bass and largemouth bass taken from remote lakes, background PFOA concentrations in the livers of fish from the Niagara River are expected to be from < 1.6 to 6.1 ng/g (median = 3.5 ng/g). A BCF of 184 can be estimated for PFOA in the livers of fish. A BCF of 4 to 27 was reported for PFOA in rainbow trout exposed under laboratory conditions (Martin et al.2003). PFOA was not present in the livers of any of the birds, a fact that indicates the lower bioaccumulation potential of PFOA than PFOS.

Conclusions

In summary, our study of the distribution of PFOS, PFOA, PFOSA, and PFHS in surface waters of NYS highlights that Lake Onondaga contains elevated levels of PFOS, whereas the Hudson River contains elevated levels of PFOA. Both PFOS and PFOA were ubiquitous in New York waters; however, PFOA was typically found at higher concentrations. Background concentrations in the livers of bass (smallmouth and largemouth taken together) from remote lakes, with no point sources of PAS contamination, were determined to be 14 to 207 ng/g, < 1.5 to 6.1 ng/g, and < 1.5 to 9.8 ng/g, wet wt, for PFOS, PFOA, and PFOSA, respectively. Only PFOS was found in the livers of birds. Concentrations of PFOS were not significantly different in the livers of birds collected in 1994/1996, and in 1999/2000. Significantly higher concentrations of PFOS were found in the livers of piscivorous birds than in non-piscivorous birds (p = 0.001). An average BCF of 8850 was estimated for PFOS accumulation in fish relative to the PFOS concentration in ambient waters. An average BMF of 8.9 was estimated for the accumulation of PFOS in common merganser:fish. This degree of bioconcentration highlights the significance of dietary fish in PFOS accumulation in the food chain.

Acknowledgments

The authors wish to thank Jefferey Loukmas, New York State Department of Environmental Conservation, for providing fish samples and information regarding fish. We thank Erik Latremore and Dustin Edwards for collecting fish and processing tissues and Larry Skinner and Howard Simonin (all with NYSDEC) for assistance with the fish collections. Fish samples were collected as part of a New York State Energy and Research and Development Authority (NYSERDA) grant. Funding for the collection of waterfowl and dietary analysis was provided under Federal Aid for the Restoration of Wildlife to New York State, Project WE-173-G. We thank T. Martin, assisted by J. Curtiss, T. Forti, C. Brown, M. Kandel, J. Rogers and D. Seyler in the field and G. Kimber, M. Levendusky and C. Dean in the laboratory, for collection and processing of waterfowl during 1994-1996. We thank A. Bathrick, K. Hellijas, and K. Geesler for the processing of waterfowl collected during 1999/2000; M. A. Ellis for diet analysis; and S. Fonda, N. Wright, B. Bidwell, and S. Reese for data management. We thank all waterfowl hunters who donated birds, and J. Daniels, who volunteered to solicit birds from hunters and to maintain duck traps for the study.

Copyright information

© Springer Science+Business Media, Inc. 2006