U.S. Dept Commerce/NOAA/NMFS/NWFSC/Publications

NOAA Technical Memorandum NMFS-NWFSC-12

NOAA logo

NWFSC/NMFS identifier/logo

Volume I: Survey of Alaskan subsistence fish, marine mammal, and invertebrate samples collected 1989-91 for exposure to oil spilled from the Exxon Valdez

Usha Varanasi, Donald W. Brown, Tom Hom, Douglas G. Burrows, Catherine A. Sloan, L. Jay Field*, John E. Stein, Karen L. Tilbury, Bruce B. McCain, & Sin-Lam Chan

National Marine Fisheries Service
Northwest Fisheries Science Center
Coastal Zone and Estuarine Studies Division
2725 Montlake Blvd. E.
Seattle WA 98112-2097

*Hazardous Materials Reponse and Assessment Division Office of Resources Conservation and Assessment

October 1993

Ronald H. Brown, Secretary

National Oceanic and Atmospheric Administration
D. James Baker, Administrator

National Marine Fisheries Service
Rolland A. Schmitten, Assistant Administrator for Fisheries


Sample Collection
Fish and Shellfish
Marine Mammals
Bile Analyses
Analyses of Aromatic Compounds in Tissue (Edible Flesh)
Extraction of Aromatic Compounds
Fractionation of Extract to Isolate Aromatic Compounds
Aromatic Compound Determinations by Gas Chromatography/Mass Spectrometry-Sequenced Selected Ion Monitoring
Statistical Methods
Aromatic Compounds in Petroleum and Petroleum Products
Fluorescent Aromatic Compounds in Bile of Fish
Biliary FACsPHN in Salmon
Biliary FACsPHN in Bottomfish
Temporal Changes in Biliary FACs in Salmon
Aromatic Compounds in Edible Flesh of Fish
Fluorescent Aromatic Compounds in Bile of Marine Mammals
Aromatic Compounds in Edible Flesh of Marine Mammals
Aromatic Compounds in Invertebrates
Comparison of ACs in Molluscs by Area
Comparison of ACs in Molluscs by Species
Aromatic Compounds in Mussels
Aromatic Compounds in Butter Clams
Aromatic Compounds in Littleneck Clams
Aromatic Compounds in Chitons
Aromatic Compounds in Other Invertebrate Species and Miscellaneous Samples
Temporal Changes in the Concentrations of Aromatic Compounds in Molluscs
Patterns of Aromatic Compounds in Invertebrates

Executive Summary

The Exxon Valdez ran aground on Bligh Reef, Prince William Sound, Alaska on March 24, 1989, spilling millions of gallons of Prudhoe Bay crude oil (PBCO). During the weeks following the spill, large amounts of oil flowed towards southwestern Prince William Sound, and as a result, many shorelines were oiled. The spreading of spilled oil raised concerns of native Alaskans that their subsistence seafoods (fish, marine mammals, and invertebrate organisms) were contaminated by the spilled petroleum. At the request of native Alaskans, a study was conducted as a cooperative effort among NOAA, Exxon, and the Alaska Department of Fish and Game to assess the degree of contamination of subsistence organisms by PBCO. In this study, edible flesh of fish, marine mammals, and shellfish from 22 native subsistence food collection areas and from two reference areas (Angoon and Yakutat) were analyzed for aromatic compounds (ACs). Vertebrates can readily biotransform ACs to metabolites that are concentrated in bile for excretion. This process greatly limits the accumulation of ACs in tissues such as edible flesh. Thus, for fish and marine mammals, bile was first analyzed for the presence of fluorescent aromatic compounds (FACs) as an indication of exposure to petroleum.

Based on the concentrations of FACs in bile, it was evident that pink salmon, halibut, and Pacific cod from the Chenega area had been exposed to ACs during 1989, as were pink salmon from Tatitlek, Kodiak, and Old Harbor. The bile method was useful because it quickly identified those fish that were relatively unexposed to ACs and, therefore, of less immediate interest for analysis by the more detailed method for ACs in tissue.

As expected, most fish muscle samples were not contaminated with ACs (<10 ng/g). In fact, the highest concentration of ACs found in muscle samples of fish caught during this study was 100 ng/g in a pink salmon caught near Kodiak (city) in 1989. In contrast, two samples of smoked salmon obtained from Tatitlek and Old Harbor contained 23,000 and 8,100 ng/g ACs, respectively.

Bile and tissue samples were collected from 33 harbor seals and 10 sea lions in 1989 and 1990. As with fish, the concentrations of FACs in bile of harbor seals varied considerably. Nine of the 12 bile samples with the highest concentrations of FACs were from animals collected in 1989 that were visibly oiled. With two minor exceptions, samples of muscle, liver, and kidney from harbor seals and sea lions, as well as blubber from sea lions, were not contaminated (<10 ng/g) with ACs. Samples of blubber from 12 of the harbor seals were minimally contaminated (10 to 99 ng/g), and samples of blubber from 4 harbor seals were moderately contaminated with ACs (100 to 1,000 ng/g).

Invertebrates from most of the sampling areas were not contaminated or were minimally contaminated by ACs (<100 ng/g). Therefore, results are presented for only those few stations where higher concentrations of ACs were found. Molluscs from some stations at the Chenega, Windy Bay, Kodiak, and Old Harbor sampling areas were moderately or highly contaminated (>100 ng/g) with ACs. For example, some of the mollusc samples from 4 of the 12 stations in the Chenega area (CHE1, CHE7, CHE10, CHE24) were moderately or highly contaminated with ACs. Mollusc samples from Windy Bay stations WNB1 and WNB3 contained concentrations of ACs as high as 18,000 ng/g; 34 of 106 invertebrate samples contained concentrations of ACs greater than 100 ng/g. The only two stations on Kodiak Island where mollusc samples had mean concentrations of ACs (by year for individual species) greater than 100 ng/g (moderately or highly contaminated) were KOD3 and OHA4. Most of the mollusc samples (26 of 30) from station KOD3 (located on Near Island about 1/4 mile from Kodiak's boat harbor) were moderately or heavily contaminated with ACs (>100 ng/g). Station OHA4 was adjacent to the village of Old Harbor near the boat harbor, and the concentrations of ACs in molluscs collected at this site in 1989 and 1990 were just within the moderately contaminated category or lower.

Aromatic compounds were present in molluscs at concentrations high enough to evaluate in terms of temporal trends only at some stations. For example, the concentrations of ACs declined significantly with time in mussels at: 1) Chenega stations CHE9 and CHE10 (1990 to 1991), and 2) at the combined Windy Bay stations WNB1/WNB3 (1989 to 1991). The degree of contamination of invertebrates from WNB1 and WNB3 stations varied with sampling year and by species. Specifically, some of the mussel samples from Windy Bay station WNB1 (1989) and from WNB3 (1990) were highly contaminated (>1,000 ng/g), whereas the concentrations of ACs in mussels from these stations in 1991 were minimally to moderately contaminated (10 ng/g to 1,000 ng/g). The decline in concentrations of ACs in these molluscs probably related to decreased exposure which resulted from weathering of the spilled oil at the particular stations. The concentrations of ACs in molluscs at some other stations did not decline significantly with time. For example, the concentrations of ACs in butter clams from Kodiak station KOD3 did not consistently decline over four sampling periods during 1990.

The relative concentrations of the hundreds of different ACs in various petroleums and petroleum products can vary considerably. The patterns of these concentrations can be useful for purposes of comparison. The patterns of some ACs (phenanthrenes and dibenzothiophenes) in selected mollusc samples from Chenega area stations CHE1 and CHE10 and Windy Bay stations WNB1 and WNB3 were similar to that of weathered PBCO. Because the overall patterns of ACs in molluscs did not exactly match that of PBCO, other observations were also important in considering sources. For example, following the spill, oil was observed in the area of station CHE1 and at CHE10, a tar mat about 1 m wide extended the length of the beach at the high tide line. Also, stations WNB1 and WNB3 were observed to be moderately to heavily oiled. Thus, based on the patterns of ACs in molluscs and the known proximity of the spilled oil to these areas, oil from the Exxon Valdez may have been the source of ACs in mollusc samples from CHE1, and most likely was the source of ACs in mollusc samples from CHE10, WNB1, and WNB3. The patterns of ACs in selected samples from Kodiak Island stations KOD3 and OHA4 were also similar to those for mollusc samples from CHE1, CHE10, WNB1, and WNB3. However, the presence of naphthalenes in the patterns was more prominent in the samples from KOD3 and OHA4 than in the Chenega and Windy Bay samples. This finding implies exposure of the KOD3 and OHA4 molluscs to a less weathered source of ACs. Therefore, the ACs in molluscs from KOD3 and OHA4 are suspected to be from a local continuing source of petroleum. Additional support for this conclusion includes: 1) KOD3 and OHA4 were near active boat harbors which could be a source of ACs, 2) the spilled oil was not observed to impact these areas, and 3) the concentrations in molluscs at KOD3 did not continually decline over four samplings during 1990. Based on patterns, PBCO was probably a minor source of the ACs in mollusc samples from station CHE7. The pattern of ACs implied that the source of these compounds in selected molluscs from CHE7 was due to exposure to creosote (perhaps from the creosoted pilings located near the sampling station) and/or products of combustion. Based on the patterns of ACs in selected samples, PBCO was probably not a source, or only a minor source of the ACs in molluscs from Tatitlek station T1. More likely, the source of the ACs in mollusc samples from T1 was products of combustion processes.

Interestingly, of the ACs found in fish muscle, unsubstituted ACs predominated, which was probably due to the more rapid metabolism of alkylated ACs than of unsubstituted ACs by fish liver. Conversely, molluscs, which have little ability to metabolize ACs, had both alkylated and unsubstituted ACs, and the patterns of ACs in molluscs more closely resembled that for petroleum components. Furthermore, in the blubber samples from harbor seals with elevated concentrations of ACs, the concentrations of alkyl-substituted ACs were similar to or greater than the concentrations of the corresponding nonsubstituted ACs. This pattern is similar to what was found in PBCO and molluscs, and generally the opposite of what was observed in fish.

In conclusion, the finding of elevated concentrations of FACs in some bile samples from fish and marine mammals was clear evidence of their exposure to petroleum. Generally, ACs were not found in muscle tissue of fish, harbor seals, and sea lions. Some harbor seal blubber samples did contain ACs; however, the concentration of ACs in most blubber samples was less than 100 ng/g. Smoked salmon contained higher concentrations of ACs (8,000 to 20,000 ng/g) than any of the untreated subsistence samples. The concentrations of ACs were less than 100 ng/g in approximately 90% of the more than 1,000 mollusc samples from 80 sampling beaches. The concentrations of ACs were elevated in some mollusc samples (as high as 18,000 ng/g), and the concentrations of ACs exceeded 1,000 ng/g in 24 samples.

The results to date provide important information on the level of contamination of subsistence fish, shellfish, and marine mammals from fishing areas of native Alaskan villages in and near Prince William Sound. In an advisory opinion, the Food and Drug Administration has indicated that little risk is involved in the consumption of the nonsmoked subsistence foods studied. Subsistence food gatherers were advised not to collect or consume food if oil was observed to be present. The results also show that in future oil spills, shellfish tissues should be given the highest priority for analysis, whereas rapid screening of bile from fish and marine mammals should be sufficient to provide information on level of exposure.

Hardcopies of NOAA Technical Memorandums are available from:

Paper copies vary in price

Return to the Table of Contents


The Exxon Valdez ran aground on Bligh Reef, Prince William Sound (PWS), Alaska, on March 24, 1989, spilling more than 10 million gallons of Prudhoe Bay crude oil (PBCO). During the 2 weeks following the spill, large amounts of floating oil spread to southwest PWS, resulting in heavy oiling of several shorelines (Galt et al. 1991). As the PBCO spilled from the Exxon Valdez and spread from PWS along the outer Kenai Peninsula, shoreline oiling was variable. Shorelines of some eastward-facing bays along the Gulf of Alaska such as Windy Bay received moderate to heavy oiling. By the time the oil reached Kodiak Island and the Alaska Peninsula, shoreline oiling was generally lighter because the oil was in the form of widely scattered patches of tar balls and mousse. The village of Tatitlek, which is only a few miles east of where the spill occurred, did not seem to receive any direct shoreline oiling as the oil moved south and west.

The spreading of spilled oil raised concerns among native Alaskans that their subsistence seafood (fish, marine mammals, and invertebrate organisms) was contaminated by the spilled petroleum. In response to their concerns, NOAA entered into memoranda of understanding (MOU) with Exxon and subsequently with the Alaska Department of Fish and Game (ADF&G) to analyze subsistence seafoods sampled from sites used by native Alaskans, and from reference areas for selected aromatic compounds (ACs) found in spilled oil (Varanasi et al. 1990).

Petroleum, such as PBCO, contains many hundreds of organic compounds, generally classified into groups including aliphatic and polycyclic aromatic hydrocarbons and compounds that contain sulfur, oxygen, and nitrogen. Aromatic hydrocarbons and dibenzothiophenes comprise the group of ACs that are known for their toxicity, and these compounds can be monitored in biota as indicators of exposure to petroleum-related ACs (Table 1). Previous laboratory studies have shown that fish and mammals efficiently biotransform ACs to polar derivatives that are concentrated in their bile for excretion (Statham et al. 1976, Varanasi and Gmur 1981, Stein et al. 1984, Varanasi et al. 1989b). As a result of this biotransformation, ACs may not readily accumulate in the edible flesh of fish or marine mammals. Thus, to quickly measure the exposure of fish and marine mammals to ACs, a sensitive method for screening bile for the presence of fluorescent aromatic compounds (FACs) which are characteristic of petroleum ACs was utilized. This semiquantitative procedure for measuring FACs utilizes high performance liquid chromatography (HPLC) with fluorescence detection (Krahn et al. 1984, 1986a). This method has been employed previously to evaluate exposure of fish to ACs from an oil spill on the Columbia River (Krahn et al. 1986b) and in a variety of other laboratory and field studies. The results of biliary FACs analyses were used to prioritize corresponding samples of edible fish tissue for analyses of individual ACs.

Individual ACs in selected tissue samples from invertebrates, fish, marine mammals, and other tissue types (e.g., herring roe on kelp) were analyzed using gas chromatography/mass spectrometry (GC/MS). This procedure is much more labor and time intensive than the measurement of FACs in bile. However, recent important improvements in, and automation of, the extract cleanup procedure (Krahn et al. 1988a) and the use of a special sequenced selected-ion monitoring (SSIM) GC/MS program (Burrows et al. 1990) enabled us to provide high quality analytical data for low concentrations (<1 ng/g wet weight) of ACs in tissue samples quickly and efficiently. The GC/MS method also detects sulfur-containing ACs such as the dibenzothiophenes, which can be useful markers for oil contamination.

In addition to organic compounds, petroleum contains certain metals (e.g., nickel and vanadium). Metals were not analyzed as part of the study because of very low concentrations of these metals relative to hydrocarbon components in PBCO and the lack of information about the bioavailability processes of metals in various marine biota (including those sampled in this study).

This report presents the results of analyses of bile for FACs and tissue samples for ACs from subsistence species collected in 1989, 1990, and 1991. The results were evaluated both in terms of geographical distribution and temporal changes.

This report is divided into two volumes, with the second volume containing supplemental data too voluminous to be included as Appendices to the primary report. It was felt that because everyone reading the primary report might not be interested in the supplementary data, it would be preferable to publish it as a separate volume available on request. Volume II, Supplemental Information Concerning a Survey of Alaskan Subsistence Fish, Marine Mammal, and Invertebrate Samples Collected 1989-91 for Exposure to Oil Spilled from the Exxon Valdez, includes the following sections: A - Concentrations of Metabolites of Fluorescent Aromatic Compounds in Fish Bile; B - Concentrations of Aromatic Compounds in Edible Tissue of Fish; C- Concentrations of Metabolites of Fluorescent Aromatic Compounds in Bile from Marine Mammals; D - Concentrations of Aromatic Compounds in Edible Tissues of Marine Mammal Samples; E - Concentrations of Aromatic Compounds in Invertebrate Samples; and F - Quality Assurance.


Following the spill in March 1989, the goal was to organize and implement a program to determine if subsistence seafoods were contaminated by the petroleum. In 1990, the program was expanded to include sampling more areas, especially for shellfish, and to collect more samples in selected sampling areas. In 1991, the study focused exclusively on shellfish stations in the Chenega Bay sampling area and Windy Bay islands. The reasons for this exclusive focus were that the southwestern area of PWS in the general vicinity of the village of Chenega Bay was potentially the most impacted by oil from the spill, several stations in the Chenega sampling area showed elevated concentrations in 1990, and the Windy Bay islands were the most heavily-oiled stations in the study and were important for assessing temporal trends.

Details of protocols for the field sampling, chemical analyses, and statistical evaluation are outlined below.

Sample Collection

Samples of marine biota were collected under two programs using similar protocols. Exxon sponsored collections by its contractor, Dames and Moore, in 1989, 1990, and 1991. The ADF&G sponsored collections in 1990 and 1991. A NOAA Hazardous Materials Response and Assessment Division representative also participated in the sample collections. Invertebrate samples were collected over a 3-year period with at least three collection cycles each year. Fish and marine mammal samples were collected in 1989 and 1990. In most cases, the sampling areas and stations were selected by area residents because they were important subsistence collection areas.

Fish and Shellfish

In 1989, comparable numbers of shellfish and fish samples were collected from each village. The target invertebrate species included mussels (Mytilus edulis), butter clams (Saxidomus giganteus), littleneck clams (Protothaca staminea), and chitons (order Neoloricata). Fish species included pink salmon (Oncorhynchus gorbuscha), coho salmon (O. kisutch), sockeye salmon (O. nerka) chum salmon (O. keta), chinook salmon (O. tshawytscha), halibut (Hippoglossus stenolepis), and Pacific cod (Gadus macrocephalus). Marine mammals, including harbor seals (Phoca vitulina) and Steller sea lions (Eumetopias jubatus), were also sampled. Three sampling cycles were conducted during the summer of 1989 (July, August, and September) in subsistence areas associated with villages in PWS (Tatitlek and Chenega Bay); Lower Cook Inlet (including the Outer Kenai Peninsula villages of Port Graham and English Bay, Kasitsna Bay, and Windy Bay); and Kodiak Island (city of Kodiak, Chiniak, Larsen Bay, Karluk, Akhiok, Old Harbor, Ouzinkie, and Port Lions) (Fig.1, Fig. 2, Fig. 3, Fig. 4, Fig. 5, Fig. 6, Fig. 7, Fig. 8, Fig. 9, Fig.10, Fig. 11, Fig. 12, Fig. 13).

In 1990, all of the above-mentioned areas were sampled at least once and the number of samples for each species from each sampling area was increased compared to 1989. In addition, samples were collected from other areas in southwestern PWS near the village of Chenega Bay and from areas on the Kenai Peninsula as requested by villagers (Port Chatham, Sadie Cove, Point Bede, Port Dick, Chugach Bay) (Fig. 4, Fig. 5, Fig. 6), and from four village areas on the Alaska Peninsula (Chignik, Ivanof Bay, Kashvik, and Perryville) (Fig. 1, Fig. 14, Fig. 15). The focus during the second year was on collecting larger numbers of invertebrate species at each sampling station.

In 1991, the project focused primarily on areas in southwestern PWS and Lower Cook Inlet (areas near Chenega Bay and Windy Bay). Some shorelines in both of these areas were heavily oiled in 1989 and thus would be areas at which temporal changes in levels of ACs in invertebrate subsistence species could possibly be evaluated over the 3-year sampling period.

Selection of sample stations and the subsistence resources to be sampled was done in cooperation with village representatives. Sampling at each village area usually included two beaches for intertidal invertebrates. Salmon and bottomfish were collected from stations designated by village fishermen as ones generally fished. Hence, sample collection represented the approach of subsistence collectors rather than random or statistical sampling. No attempt was made to avoid or seek out oiled areas for sampling of fish or invertebrates. If oil was observed on a beach or in the vicinity of a sample collection area, it was noted in the sampling log. Since the habitat requirements vary for different species, individual stations sometimes encompassed a relatively large beach area over a range of tidal elevations in order to collect several species. For example, mussels were collected from higher tidal elevations than clams, while chitons were collected from rocky, lower intertidal areas.

In most cases, an individual from each village participated in sample collections conducted in areas adjacent to his/her village, except for marine mammal samples. Field collection protocols were followed to minimize the possibility of external contamination of samples. Tools used in sampling (e.g., shovels) were washed with soap and water and a new pair of surgical gloves was worn between sample collections. Intertidal shellfish were collected by hand or shovel. A minimum of four to six individual animals of each invertebrate species was collected at each station as one sample. Gear used in the fish collections included hook-and-line and gill nets for salmon, and hook-and-line and longline for halibut and bottomfish. Bile taken from fish was placed in solvent rinsed 4-mL vials and frozen as soon as possible for analysis for FACs. Fillets or whole fish were double wrapped in aluminum foil (which had been pre-baked at 350 °F for 1 hour), placed in Zip-Loc® freezer bags and placed in ice coolers. Samples were shipped frozen and stored at -80°C at the laboratory until analyzed. Bile and tissue samples from fish were sampled and maintained as individual samples, whereas the invertebrate organisms that comprised a sample were packaged together in the field. Chain-of-custody forms were used and were signed by both parties when samples were transferred from sample collectors to laboratory personnel.

The number of fish and mollusc samples collected by site and year is shown in Table 2.

Marine Mammals

Marine mammals were collected under two programs with different objectives. Samples collected by ADF&G were from animals and areas with expected high exposure to spilled oil. Samples were collected from animals harvested by subsistence hunters in 1990 by Dr. Paul Becker of NOAA as part of the Marine Mammal Tissue Archival Program, Office of Protected Resources. Harbor seals and Steller sea lions that were collected both by ADF&G and NOAA in 1989 and 1990 were from sites that were oiled during the Exxon Valdez spill or that were adjacent to oiled areas. Samples were collected from 33 harbor seals and 10 Steller sea lions. Chain-of-custody procedures and bile and tissue sample procedures were the same as for fish samples (described above). Bile from 29 harbor seals and the 10 sea lions was analyzed for FACs, and samples from all the harbor seals and sea lions were analyzed for ACs in tissues.

Bile Analyses

The concentrations of FACs in bile were determined using a Waters HPLC equipped with a Perkin-Elmer HC-ODS/ PAH column (0.26 X 25 cm), an automatic injector, and Perkin-Elmer model 40 fluorescence (UV-F) detectors connected in series (Krahn et al. 1988b). Thawed bile was injected directly into the HPLC and eluted through the column using a linear gradient from 100% solvent A (water containing 5 mg/L of acetic acid) to 100% solvent B (methanol) over 15 minutes. The flow rate was 1 mL/min and the column temperature was 50°C. All solvents were degassed with helium. The UV-F responses were recorded at the wavelength pairs for naphthalene (NPH) and phenanthrene (PHN), prominent constituents of ACs in PBCO (Table 1). The fluorescence of NPH metabolites was monitored using excitation and emission wavelength pairs of 290 and 335 nm, respectively. Fluorescence of PHN metabolites was monitored using excitation and emission wavelength pairs of 260 and 380 nm, respectively.

The total integrated area from each detector was then converted to corresponding equivalents of either NPH or PHN standards that would give the same integrated response. Concentrations of FACs in bile are reported on the basis of bile weight and biliary protein. The levels of protein in bile samples were determined by the method of Lowry et al. (1951).

Analyses of Aromatic Compounds in Tissue (Edible Flesh)

The results of the bile analyses were used to estimate the exposure of fish to ACs and to rank the exposure as low, medium, or high. Fish samples were then selected for analysis of edible flesh for the ACs found in PBCO by the more quantitative and costly GC/MS technique. Edible flesh samples from the same fish species from a sampling area were analyzed as individual samples or as composites of individuals which had similar FACs levels in their bile. Samples of flesh from fish (collected in 1989) with relatively low levels of bile FACs were generally analyzed as composites. Some of the samples of fish collected in 1990 that had relatively low levels of FACs in their bile were not analyzed. Tissue samples from marine mammals were not composited for analysis.

Edible flesh of fish, marine mammals, and invertebrate samples were analyzed for the ACs listed in Table 1 using the procedures of Krahn et al. (1988a) and Burrows et al. (1990). Summaries of the analytical protocols are given below and consisted of four major steps: a) extraction; b) HPLC cleanup; c) analyte determination by GC/MS; and d) quality assurance. In addition, samples of petroleum and related products were weighed, dissolved in methylene chloride, and analyzed, starting with the HPLC cleanup.

Extraction of Aromatic Compounds

The edible tissues of invertebrates that comprised a sample were homogenized and a 5-g portion of the homogenized tissue was analyzed for ACs. Fish and marine mammal tissue samples were taken for analysis by cutting back the surface layer of tissue and then taking 5 g of tissue that had not been in contact with the foil wrapping. When tissue samples from more than one fish were to be composited for analysis, the individual 5-g samples were combined and homogenized and a 3- to 5-g portion of the homogenate was used for analysis for ACs. The 3- to 5-g sample of homogenate was added to a centrifuge tube containing sodium sulfate and methylene chloride. The method internal standards (surrogate standards) for ACs were added and the mixture macerated with a Tekmar Tissumizer®. The resulting extract was filtered through a column of silica and alumina and the extract concentrated to 1 mL for cleanup by HPLC.

Fractionation of Extract to Isolate Aromatic Compounds

The ACs were isolated using a Spectra-Physics (Mountain View, California) model 8800 HPLC equipped with an ultraviolet detector (254 nm), an automatic injector, and a fraction collector. Two 22.5 x 250-mm stainless-steel preparatory size columns containing Phenogel 100-Å size-exclusion packing (Phenomenex, Rancho Palos Verdes, California) were used in series with a 2-uM Rheodyne model 7302 filter and a 7.8 x 50-mm guard column containing the same Phenogel packing. The HPLC pre-column and column were connected to a six-port valve that allowed the guard column to be backflushed to remove extraneous materials after cleanup of a set of samples (n ~ 10).

Methylene chloride was used as the solvent and was pumped at a flow rate of 7 mL/min for 20 minutes at ambient temperature. The HPLC solvent was degassed by bubbling helium through it. The helium was delivered via a regulator equipped with a stainless-steel diaphragm and was passed through an in-line charcoal filter (200-cc hydrocarbon trap, Alltech Assoc., Deerfield, Illinois), then through a high-purity heated trap (Supelco Inc., Bellefonte, Pennsylvania), and an oxygen indicating trap (Alltech Assoc.) to eliminate compounds which could be transferred to the HPLC solvent by the helium.

A 250-µL portion of a 1 mL extract was injected onto the HPLC column and the fraction containing the ACs collected according to Krahn et al. (1988b). The solvent in the HPLC fraction was exchanged into hexane as the volume was reduced by evaporation to approximately 1 mL. Further evaporation, and then the addition of standards, brought the final volume to ca. 120 µL for analysis by capillary column GC with mass spectrometric quantification.

Aromatic Compound Determinations by Gas Chromatography/Mass Spectrometry-Sequenced Selected Ion Monitoring

The ACs were determined according to MacLeod et al. (1985) by GC/MS quantification as outlined by Burrows et al. (1990). A 30-m x 0.25-mm DB-5 capillary column (J & W Scientific) was used in a Hewlett-Packard (HP) model 5890 GC. The GC sample (3 µL) was injected splitless, and the split valve opened after 30 seconds (split ratio 20:1). The oven temperature of 50°C was held for 1 minute and then programmed to increase at 4°C/min to 300°C, where the temperature was held for 10 minutes. The GC/MS analyses were accomplished using either a Finnigan Incos 50B or HP 5970 MSD mass spectrometer with the appropriate data system and an HP autosampler. A SSIM scan descriptor that included 10 groups of about 20 ions each was used.

Statistical Methods

Results of chemical analyses of samples of fish and shellfish from stations within each sampling area were compared using GT2 comparison-interval graphs (in cases where an analyte was not detected, it was assigned a value of zero for statistical treatment). The GT2 comparison intervals are similar in appearance to confidence intervals, but they actually present the results of analysis of variance followed by a multiple-range test. The GT2 comparison interval for each mean is calculated based on the number of samples for that mean, the variability about that mean, and the number of means being compared (Gabriel 1978, Sokal and Rohlf 1981).

The advantage of using comparison intervals is that they provide a graphical way of depicting significant differences between means (e.g., when comparison intervals overlap they are not significantly different (p <= 0.05)). When the number of samples for the means of interest are unequal, it is desirable to use the GT2 method (Gabriel 1978, Sokal and Rohlf 1981) for calculating the comparison intervals because the span of a comparison interval depends on the within-group variability in the entire data set and on the number of samples in each category.