Environ. Sci. Technol. 1997, 31, 84-88
Toxaphene in Great Lakes Fish: A Temporal, Spatial, and Trophic Study SUSAN T. GLASSMEYER,† DAVID S. DE VAULT,‡ TANYA R. MYERS,† AND R O N A L D A . H I T E S * ,† School of Public and Environmental Affairs and Department of Chemistry, Indiana University, Bloomington, Indiana 47405, and United States Environmental Protection Agency, Great Lakes National Program Office, 77 West Jackson Boulevard, Chicago, Illinois 60604
We report here on the concentrations of toxaphene, a complex mixture of hexa- to decachlorinated norbornanes and norbornenes, in Great Lakes’ lake trout and smelt sampled in 1982 and 1992. Lake trout from Lakes Superior, Michigan, and Huron had higher lipid-normalized toxaphene concentrations than the smelt, while in Lakes Ontario and Erie both trophic levels had about the same concentrations. Concentrations in both species declined between 1982 and 1992, with the exception of Lake Superior, where there was no significant difference. Except for the Lake Superior samples, these trends were expected because toxaphene was banned in the United States in 1982. The fact that the Lake Superior samples did not show lower concentrations over time suggests that there may be some lake-specific source that is continuing to load toxaphene into Lake Superior or that toxaphene is not being removed from Lake Superior as quickly as the other Great Lakes.
Introduction In the late 1940s, the Hercules Company introduced toxaphene in the United States as an insecticide. Hercules extracted crude R-pinene from pine stumps, using methyl isobutyl ketone, heat, and pressure. Isomerization of the R-pinene produced camphene, bornylene, and R-terpineol; the camphene was subsequently chlorinated to produce toxaphene (1). Since chlorination of camphene can take place to varying degrees and on various carbon atoms, there are 32 768 theoretically possible congeners (2), of which more than 600 have been found in commercial toxaphene (3). Due to the complexity of the mixture, less than 30 of these congeners have been isolated and structurally identified (4). The primary use of toxaphene, estimated at 67-90% of its total U.S. consumption (1, 5), was as an insecticide to treat cotton pests (6). The remainder was used as an insecticide and a herbicide on soybeans and peanuts, as well as to treat scabies on livestock, and as a piscicide to remove rough fish from lakes (1, 7, 8). The majority of this usage occurred in the southeastern United States; uses in the Midwest accounted for less than 1% of the total annual use (9). Toxaphene’s use was encouraged after the U.S. Environmental Protection Agency (EPA) banned DDT in 1972 (10). Like PCBs and other heavily used organochlorines, toxaphene is ubiquitous in the environment, and it has been * E-mail address:
[email protected]. † Indiana University. ‡ U.S. Environmental Protection Agency.
84
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 31, NO. 1, 1997
found in samples from such diverse areas as the Great Lakes region (11-14), the North Sea (15), Siberia (16), and the Nile River (17). The U.S. EPA canceled the registrations for most of toxaphene’s uses in 1982, but allowed the existing stocks to be used in limited circumstances until 1986 (18). Toxaphene has not been registered in Canada since 1983 (19). In addition to the U.S. and Canada, its use has been banned in England, Sweden, Finland, Denmark, France, Germany, Switzerland, Hungary, Italy, Egypt, India, China, and Algeria (20). Toxaphene, or a “toxaphene-like” mixture of chemicals, was first detected in Great Lakes fish (Lake Michigan Lake trout) in 1973 (21). Petty et al. (22) reported a mixture resembling toxaphene at wet weight concentrations ranging from 0.2 to 0.9 µg/g in several species of fish collected in the early 1980s from the Great Lakes, from rivers in the southeastern United States, and from Isle Royal in Lake Superior. The first unambiguous results were reported by Swackhamer et al. (23), who measured wet weight concentrations of 0.30.5 µg/g in carp, whitefish, and trout from Saginaw Bay and from Siskiwit Lake on Isle Royal in Lake Superior. Due the complex nature of toxaphene and the consequent analytical difficulties, there have been few comprehensive studies of its environmental trends and fate. Differences in the analytical methods and in reporting protocols have hindered comparability and have made analysis of concentration trends, which could then be used to gauge the effectiveness of the 1982 ban, difficult. Only recently have analytical methods been developed that provide analysis at the homologue level (23). In this study, we analyzed archived fish samples collected from each of the Great Lakes in 1982 and in 1992 to evaluate changes in toxaphene concentrations and composition at two trophic levels resulting from the 1982 ban.
Experimental Section Sample Collection and Preparation. Archival fish samples collected in 1982 and 1992 were provided by the U.S. EPA and the U.S. National Biological Survey (NBS) though the Great Lakes Fish Contaminate Monitoring Program. Through this program, lake trout (Salvelinus namaycush) were collected at one site each on Lakes Superior, Huron, Michigan, and Ontario. Walleye (Stizostedion vitreum vitreum) were collected from Lake Erie. For the exact fish collection sites, see Figure 1. Rainbow smelt (Osmerus mordax) were collected by gill net from the same five sites in 1982 and from Lakes Superior, Huron, and Erie in 1994. The five sites were chosen since they are located at offshore fishing grounds that are not impacted by local sources. Following collection, the fish were weighed, measured for total length, placed in plastic bags, frozen, and shipped to the NBS Great Lakes Science Center, Ann Arbor, MI. At NBS, the fish were composited (five whole fish per sample), homogenized, and stored frozen at less than -30 °C in solventwashed jars with foil-lined screw caps until analyzed in this study. In order to reduce the effects of age and size variation on the composite samples, only fish of a narrow size range were used. Thus, all lake trout were between 600 and 700 mm; walleye were between 400 and 500 mm; and smelt were between 100 and 200 mm total length. Because the size-age relationship varies from lake to lake, the lake trout analyzed in this study range from approximately 4 years old in Lake Ontario to approximately 12 years old in Lake Superior. Smelt in the 100-200 mm size range are approximately the same age (1-2 years old) across the entire basin. The analytical method employed for toxaphene was based on that of Swackhamer et al. (23). A 10-g portion of the ground
S0013-936X(96)00150-2 CCC: $14.00
1996 American Chemical Society
FIGURE 1. Great Lakes fish sampling sites. fish tissue was blended with 80 g of sodium sulfate, loaded into a Soxhlet extractor, spiked with the internal standard, either 37Cl6-trans-nonachlor (EPA Repository, Research Triangle Park, NC) or polychlorinated biphenyl 204 (Ultra Scientific, North Kingston, RI), and extracted for 24 h with 1:1 acetone in hexane. Intralaboratory experiments have shown that values quantitated using PCB 204 and 37Cl6-transnonachlor are comparable, as long as interferences are sufficiently resolved. With every batch of 5-6 tissue samples, a procedural blank, consisting of sodium sulfate spiked with an internal standard, was prepared and similarly extracted. To ensure adequate recovery, a procedural blank spike, consisting of glass wool spiked with a known amount of toxaphene (Ultra Scientific, North Kingston, RI) was extracted with every other batch. With every other batch, a laboratory duplicate was extracted and analyzed to monitor precision. All of these control experiments gave acceptable results. The lipid concentrations were determined through gravimetric measurement. For each sample, not including the procedural blank and procedural blank spike samples, 200 µL of extract was pipetted (Drummond Microdispenser, Drummond Scientific, Broomall, PA) into each of three pans. The pans were loosely covered with aluminum foil and placed in a fume hood to dry overnight. The next day, the pans were reweighed, and the concentration of lipids was calculated. Samples were reduced in volume under a gentle stream of nitrogen (Pierce 18780 Reacti-vap Evaporation Unit, Rockford, IL) or diluted until a maximum lipid concentration of 100 mg/mL was achieved. The majority of the lipids were removed using a gel permeation chromatography system consisting of a glass column (2.5 × 1000 cm, YMC Inc., Wilmington, NC) packed with porous styrene divinylbenzene copolymer beads (Bio-Beads, SX-3 or SX-8, Bio-Rad Laboratories, Hercules, CA). The solvent was 3:2 cyclohexane in dichloromethane. The toxaphene fraction was solvent exchanged into hexane through rotary evaporation (Buchi Rotovapor-R, Flavil, Switzerland) and subjected to further chromatographic cleanup on 1% water-deactivated silica (100-200 mesh grade, Davidson Chemical, Baltimore, MD) in a 1.5 × 30 cm column (Indiana University Glass Shop). Four solvent fractions were collected: hexane, 1:9 dichloromethane in hexane, dichloromethane, and methanol. The first three fractions were combined, solvent exchanged into hexane though rotary evaporation, and reduced to 200 µL under a steady stream of nitrogen in preparation for analysis by electron capture, negative ionization, gas chromatographic mass spectrometry (GC/MS). Analysis by Negative Ionization, Electron Capture GC/ MS. A Hewlett Packard 5989A mass spectrometer was used to analyze the samples. The samples were injected into a Hewlett Packard 5890 Series II gas chromatograph containing
a 30 m DB-5MS column (film thickness 0.25 µm, 250 µm i.d., J&W Scientific, Folsom, CA). Helium was used as the carrier gas at a linear velocity of 25 cm/s. The 1-µL injections were made in the splitless mode, with a vent time of 1.9 min. The injection port temperature was maintained at 285 °C to ensure complete volatilization of the sample. The temperature program for the column began with a 1 min hold at 40 °C, followed by a 10 °C/min ramp to 200 °C, a 1.5 °C/min ramp to 230 °C, and a 10 °C/min ramp to 300 °C, which was held for 5 min. After eluting from the column, the effluent was carried through a 300 °C transfer line into the ion source of the mass spectrometer, which was held at 125 °C. Methane was used as the reagent gas in the ion source; its pressure was maintained at 0.43 Torr. The electron capture GC/MS analysis procedure, using selected ion monitoring, was developed by Swackhamer et al. (23). The only notable difference in our procedure is that quantitation was based on the M- ions (quantitation ion m/z 344, confirmation ion m/z 342) for the hexachlorinated homologues; as in the earlier paper, the (M - Cl)- ions of the hepta- to decachlorinated norbornanes and norbornenes were monitored. Four time windows, each monitoring a subset of the ions, were used to increase sensitivity relative to monitoring all of the ions all of the time. At the beginning of each day of analysis and after every 4-5 sample injections, a relative response factor standard was injected into the GC/MS system. This standard, with its known concentration of 37Cl6-trans-nonachlor and toxaphene allowed accurate quantitation of the toxaphene present in the sample. With every batch, a standard consisting of known concentrations of five individual toxaphene congeners isolated by Hainzl et al. (4) (Dr. Ehrenstorfer GmbH, Augsburg, Germany) as well as the internal standard was also analyzed; this allowed us to quantitate response factors for these five congeners. With every other batch, a 50 pg/µL instrument detection limit standard was analyzed to ensure consistent performance.
Results and Discussion Lake trout (walleye in Lake Erie) were available for each lake and time period. Smelt were available from each lake in 1982, but only from Lakes Superior, Michigan, and Ontario in 1994. The mean lengths, weights, percent lipid, and mean lipid and wet weight-normalized total toxaphene concentrations in fish measured in this study are given in Table 1. The wet weights, lipid weights, wet weight-normalized concentrations, lipid-normalized concentrations, homologue distributions, and specific congener concentrations are given as Supporting Information (see paragraph at end of paper). The means and 90% confidence limits of the lipid-normalized concentrations are also plotted in Figure 2. The toxaphene concentrations in lake trout reported in this study exceed the concentrations of other persistent organochlorine compounds in trout. Newsome and Andrews (13) reported average wet weight concentrations of total PCBs to be 0.63 µg/g, total DDT to be 0.05 µg/g, dieldrin to be 0.04 µg/g, oxychlordane to be 0.02 µg/g, and total HCH to be 0.02 µg/g in commercial trout fillets. De Vault et al. (24) reported wet weight concentrations of several organochlorine compounds in whole lake trout composites from 1990 to be in the following ranges: PCBs, 0.45-2.7 µg/g; total DDT, 0.18-1.4 µg/g; dieldrin, 0.04-0.18 µg/g; oxychlordane, 0.01-0.04 µg/g (the low end of each of these ranges is from Lake Superior, while the high end is from Lake Michigan). Compare these concentrations with an average wet weight toxaphene level of about 4 µg/g that we measured in Great Lakes trout. Toxaphene seems to be one of the most prevalent organochlorine compounds in Great Lakes fish, particularly in Lake Superior. The data from this study agree well with those of other recent studies. De Vault et al. (24) reported wet weight
VOL. 31, NO. 1, 1997 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
85
TABLE 1. Average Length, Weight, Percent Lipid, Mean Lipid-Normalized Concentration (µg of Toxaphene/g of Lipid) with Standard Error, Mean Wet Weight Concentration (µg of Toxaphene/g wet weight) with Standard Error, and Degree of Chlorination Index Valuea with Standard Errorb length (mm)
weight (g)
% lipid
concn (µg/g of lipid)
Superior lake trout Michigan lake trout Huron lake trout Erie walleye Ontario lake trout Superior smelt Michigan smelt Huron smelt Erie smelt Ontario smelt
632 608 630 453 618 162 163 162 161 168
2634 2432 2612 899 2868 23.9 27.2 21.0 23.4 25.8
17.8 18.2 17.5 11.6 20.6 10.2 7.4 6.3 6.3 6.6
28 27 30 2.2 24 10 10 13 2.5 11
Superior lake trout (excluding outlier) Michigan lake trout Huron lake trout Erie walleye Ontario lake trout Superior smelt Michigan smelt Ontario smelt
622
2504
18.9
659 628 488 627 159 165 158
2850 2558 1130 2632 22.7 28.5 26.1
18.2 18.1 9.0 18.7 5.2 5.5 6.5
concn std error
concn (µg/g wet wt)
concn std error
DCI
DCI std error
1982 5 3 3 0.2 7 0.1 1 1 0.2 1
4.9 5.0 5.2 0.25 4.5 0.41 0.74 0.83 0.16 0.72
0.9 0.9 0.3 0.03 0.8 0.02 0.09 0.03 0.01 0.07
0.49 0.68 0.55 1.40 0.53 0.44 0.49 0.38 1.46 0.48
0.04 0.16 0.11 0.18 0.04 0.04 0.05 0.03 0.07 0.07
6.7 (4.9) 1.5 2.4 0.13 0.54 0.16 0.059 0.067
2.1 (1.4) 0.3 0.5 0.02 0.2 0.04 0.006 0.02
0.41
0.11
1.29 0.19 2.22 1.10 0.27 1.15 0.86
0.12 0.01 0.30 0.14 0.02 0.13 0.10
1992/1994 35 (25) 7.6 13 1.4 2.8 3.1 1.1 1.1
11 (5.6) 2 3 0.1 0.1 0.8 0.1 0.4
a The DCI is the sum of the concentrations of the hexa- plus heptachlorinated toxaphene homologues divided by the sum of the octa- plus nonachlorinated homologues. b Laboratory duplicates were not used in the calculations.
DDT, chlorinated dioxins, and chlorinated furans in lake trout (24, 25), in coho salmon (26), and in water (27). It is possible that the similarity in fish tissue residues in 1982 resulted from similar water column concentrations at that time, but data are not available to further examine this suggestion.
FIGURE 2. Mean concentrations of toxaphene in lake trout, walleye, and smelt in the Great Lakes in 1982 and 1992/1994. Error bars represent the 90% confidence level of the average. toxaphene concentrations ranging from
