REGULATORY

TOXICOLOGY

AND

PHARMACOLOGY

6, 1 l-23 (1986)

Evaluation of Acute Bioassays for Assessing Toxicity of Polychlorinated Biphenyl-Contaminated Soils Jo ELLEN HOSE,’ VANTUNA

Research

LINDAA.BARLOW,ANDSTANBENT

Group, Moore Laboratory Los Angeles, California

A. A.ELSEEWI,MARKCLIATH, Program

of Excellence

in Energy

Research.

of Zoology, 90041

Occidental

ANDMARGARET University

of California.

College.

RESKETO Riverside,

California

92521

AND COLLEEN Research

and Development,

Southern

Received

DOYLE

California

Edison,

Rosemead,

California

91770

July 30, 1985

Proposed State of California regulations use fish toxicity information as one criterion in municipal or industrial waste hazard evaluation. Static 96-hr bioassayswere performed using fathead minnows (Pimephales promelas), blacksmith (Chromis punctipinnis), and glass shrimp (Palaemonetes kadiakensis) exposed to soil experimentally contaminated with up to 500 ppm polychlorinated biphenyl (PCB) capacitor fluid added at a concentration of 500 mg liter-‘. Other bioassays were conducted with a 6-day mixing period prior to the bioassay or with acetone added to solubilize the PCBs. No mortality attributable to PCB toxicity was observed in definitive bioassays using the two fish and one invertebrate species. PCB levels leached from soil containing 500 ppm Aroclor 1242 ranged from ~0.6 to 3.4 ppb in freshwater tests to 3.5 ppb in seawater bioassays. Using these data as the basis for waste classification, soils contaminated with up to 500 ppb PCBs during capacitor spills would be designated nonhazardous. PCBs are known to be environmentally persistent and to bioaccumulate. Acute toxicity tests, therefore, do not adequately evaluate the general toxicity of PCB-contaminated soils. Hazardous waste regulations for hydrophobic compounds such as PCBs should instead be based upon chronic toxicity data and should also consider bioaccumulation potential. 0 1986 Academic press, Inc.

INTRODUCTION

Polychlorinated biphenyls (PCBs) are extensively used by electric power utilities as insulation in transformers and capacitors becauseof their excellent fireproof and di’ To whom correspondence should be sent. 11 0273-2300186 $3.00 Copyright 0 1986 by Academic Press, Inc. All rights of reproduction in any form reserved.

12

HOSE

ET

AL.

electric properties. Despite their use in closed systems, PCBs continue to enter the environment through accidental fires and spills, transportation spills, and leakage from obsolete equipment in landfills. Because PCBs are bioaccumulative, environmentally persistent, and potentially toxic, state and federal guidelines have been promulgated for the handling and disposal of PCB-contaminated wastes. Of these, the California Assessment Manual (CAM) for hazardous wastes proposed by the California Department of Health Services (1983) details a number of criteria by which a PCB-contaminated waste would be designated as hazardous. Some of these criteria are based on fish toxicity information; in particular, the soluble threshold limit concentration (STLC) of 5 ppm PCB and the total threshold limit concentration (TTLC) of 50 ppm. PCBcontaminated soils or leachates exceeding the TTLC or STLC values, respectively, would be considered hazardous on the basis of PCB concentration alone. Using the proposed 96hr aquatic bioassay method, soil with 96-hr L& values of ~500 mg liter-’ would be classified as hazardous. Waste classifications are then to be used to determine the method and location of waste disposal. In view of the estimated 17 million pounds of PCB-contaminated wastes entering landfills which were generated by phasing out of PCB equipment (Mackay et al., 1982) such regulations are of great importance to electric utilities’ waste management operations and to persons living near landfill sites. On first examination, the fish bioassay test would appear to be a useful indicator of environmental toxicity since a PCB waste or PCB residues leached from the waste could enter adjacent streams. There are, however, problems in using results of a shortterm bioassay to predict the environmental toxicity of hydrophobic, persistent compounds such as PCBs. The STLC of 5 ppm PCBs, which is based upon a 30-day LCsO of 0.05 ppm for rainbow trout (Mayer et al.. 1977) and an attenuation factor of 100, cannot be attained in an aqueous system without the use of solvents to increase PCB solubility from the solubility limit of approximately 56 ppb (Haque et al., 1974). Most published bioassays use solvents such as acetone to achieve PCB concentrations high enough to produce toxicity. In the CAM aquatic bioassay protocol (Department of Health Services, 1983) fish would be exposed to wastes in solution or suspension for 96 hr without the use of solvents. Few investigators have tested PCB-contaminated soils for toxicity using this direct method. In one such experiment, Nimmo et al. (197 1) found that soil containing 6 1 ppm Aroclor 1254 was not toxic to the shrimp, Penaeus duorurum, or crab, Ucu minux. Thus, the proposed CAM protocol may not adequately assay the toxicity of soils contaminated with PCBs. The objective of the research described here is to critically evaluate the aquatic bioassay proposed in the CAM (Department of Health Services, 1983) as a method of assessing the toxicity of soils contaminated with PCB capacitor fluid. Definitive bioassays were conducted with a freshwater and a marine fish species, fathead minnow (Pimephales promelas) and blacksmith (Chromis punctipinnis), respectively, using a California soil experimentally contaminated with Line Material Industries capacitor fluid (Aroclor 1242). Nominal PCB concentrations in soils were compared to actual levels of soluble PCBs. Results of other bioassays performed using modifications of the CAM procedure are presented. Also, bioassays were conducted with an invertebrate, glass shrimp (Palaemonetes kadiukensis), reported to be more sensitive to PCB toxicity than are fishes (Johnson and Finley, 1980).

PCB TOXICITY

MATERIALS Experimental

AND

TESTING

13

METHODS

System

Tests were conducted in wooden-frame enclosures which supported two fiberglass troughs and which were covered with 12-mm-thick polyvinyl chloride (PVC) sheeting. Each enclosure had its own lighting system which consisted of two Vitalite bulbs per fiberglass trough. The photoperiod was 12 hr light and 12 hr dark. Enclosures were connected to a continuous exhaust system. The enclosures received air from a central area air compressor which was delivered to individual aquaria via PVC piping and flexible PVC air tubing. Each enclosure was fitted with a YSI Model 47 telethermometer with 12 probes for remote temperature sensing during bioassays. Fresh water for bioassays was obtained by softening Redondo Beach, California, city water in a Culligan Mark 59 automatic water conditioner followed by ultraviolet sterilization in an Aquafine Model SL- 1 sterilizer. This softened water had < 16 ppb residual chlorine, <2 ppb unionized ammonia, and pH values between 6.7 and 8.1. The resulting water was too soft (~20 ppm CaC03) and was diluted with very small amounts of Redondo Beach city water. Analyses of the bioassay water were negative for PCBs. Seawater used for bioassays was drawn from Southern California Edison’s Redondo Beach Generating Station Units 7 and 8 intake in Redondo Canyon, California. Routine pH values were between 6.7 and 8.1 and no ammonia or residual chlorine was detected. Experimental

Animals

Fathead minnows and glass shrimp were purchased from Fattig Fish Hatchery, Brady, Nebraska. Blacksmith were collected from Ring Harbor, California, by scuba divers using l-in-mesh monofilament gill nets. Water quality measurements in holding tanks were taken weekly using methods to be described later and were generally within acceptable ranges set by APHA (1975) and Department of Fish and Game (1976). Animals were acclimated to temperatures characteristic of their natural environments. Fish were fed daily ad Zibitum with Tetramin, live or frozen brine shrimp (Artemia salina), or trout chow. Some of the juvenile blacksmith collected during the warm summer months became afflicted with bacterial dermatitis or fin rot and severely diseased fish were removed. The remainder of the juveniles were treated with 20 ppm nitrofurazone on 2 consecutive days, 10 days before the bioassays were begun. Glass shrimp were fed brine shrimp nauplii ad libitum twice daily. Soil Samples Soil samples were experimentally contaminated with 1,2.5, 5, 10,25, 50, 100,250, or 500 ppm PCBs from Line Material Industries (LMI) capacitor oil or an Aroclor 1242 standard. The soil, Gila silty loam, contained 0.58% organic matter, 18.4% clay, 7.2 ppb DDE, and no detectable DDT. Subsamples were ground to pass a 425-pm screen. Aliquots (1000 g) were spread into a $- to $-in. layer and sprayed with appro-

14

HOSE

ET

AL.

priate quantities of PCB standard or LMI capacitor oil mixed with 30 ml of 1: 1 hexane:acetone solution. Two applications were made over the soil layer by spreading lo-ml aliquots on each application at a height of 1 in. Soil was rotated between applications. Two 5-ml volumes of a clean hexane:acetone solution were then applied in a similar manner to remove any PCB residues left in the aspirator. Solvent control samples were prepared by a similar procedure except that no PCB was added to the solvent solution prior to spraying. Ground soils served as control samples. Prepared soils were stored frozen until use. Under normal weather conditions, the solvent control soil sample had a moisture content of 10.1%. Aliquots of soil samples were analyzed for PCB content by gas chromatography (GC) using a Varian 3700 gas-liquid chromatograph equipped with an 63Ni electron capture detector and standard 6 ft X 2 mm i.d. glass columns. Typical GC conditions using 3% OV- 1 CW-HP 80/ 100 as the liquid phase were column temperatures from 180 to 200°C nitrogen as the carrier gas, and a flow rate of 10 ml/min. The minimum detectable amount for Aroclor 1254 was 0.24 ng (0.1 ppb) based on two to three times the instrumental noise level. PCB concentrations were an average of five peak values. Analyses showed that LMI capacitor oil contained essentially pure Aroclor 1242 using comparisons of peak retention times with known Aroclor standards. Chromatographic profiles of Aroclor 1242 and LMI capacitor oil with characteristic peaks used for quantification are shown in Fig. 1. Aqueous PCB Concentrations Ninety-six-hour water samples (500 ml) from definitive bioassays were successively extracted with three 50-ml portions of hexane. Extract volumes were then reduced by evaporation on a Brinkman rotating evaporator to 10 ml. Dissolved Aroclor 1242 concentrations were measured using gas chromatography as described above. Aroclor 1242 recoveries from fresh water and seawater were calculated from differences between standard solutions and hexane-extracted samples of known concentrations. Mean recoveries for fresh water and seawater were 86 and 80%, respectively. Experimental

Protocol

CAM bioassay protocol. The CAM (Department of Health Services, 1983) stipulates that the bioassay procedure must be that outlined in Standard Methods for the Examination of Water and Wastewater (APHA, 1975) or any other procedure accepted by the State of California. The protocol used for CAM bioassays was that of Guidelines for Performing Static Acute Fish Bioassays in Municipal and Industrial Wastewaters by the State of California Department of Fish and Game (1976). This protocol is based upon the method contained in APHA (1975). Fathead minnows and glass shrimp were tested in soft water (40 to 48 ppm CaC03 hardness) and blacksmith in ambient seawater of 32% salinity. Dissolved oxygen was generally maintained above 75% of saturation using aerators positioned such that as much PCB-contaminated sediment as possible was kept in suspension. Typical residual chlorine levels were 0.01 to 0.02 ppm (maximum of 0.03 ppm) and usual unionized ammonia values were below 3

F’CB TOXICITY

Aroclor

Line

Material

15

TESTING

1242

lndustrles

chlorptrlfos

15

10 TIK

5

0

(mln)

FIG. I. Gas-liquid chromatograms of Aroclor 1242 standard and line material industries capacitor oil. Experimental conditions are given in the text.

ppb (maximum of 7 ppb). Bioassay temperature ranges were 21 to 24°C for fathead minnows, 19 to 22°C for blacksmith, and 25 to 27°C for glass shrimp. For tests with fathead minnows, l-gal glass jars or 5-gal aquaria were used and for blacksmith, 15-gal aquaria. Aquaria were filled to a depth of at least 12 cm ( l-2 g fish liter-‘) and soil was added at 500 mg dry weight liter-‘. Ten fish were used per concentration. Fish were not fed 24 hr prior to or during the bioassay. Replicated tests were performed. Groups of five glass shrimp were tested in l-gal glass jars filled with 2 liters water (0.019 g body wt liter-‘). Soil was maintained in suspension by glass rod aerators provided a flow rate of approximately 90 ml min -‘. Three replicate tests were conducted. Each shrimp was fed approximately 10 brine shrimp nauplii daily during the test. Mortalities were recorded at least daily and weighed, measured, and frozen for future analysis. At the end of the bioassay, 500- or lOOO-ml water samples were taken in hexane-washed jars with Teflon-lined lids and stored at 0°C for PCB analysis. Survivors were examined for gross abnormalities, then weighed, measured, and frozen. Modified bioassuys. For some bioassays, CAM methods (Department of Health

16

HOSE

ET AL.

Services, 1983) were modified to determine if toxicity could be effected. In an effort to increase the amount of dissolved PCB, equilibrium periods were conducted before the assay in which soil suspensions were vigorously aerated inside covered l-gal glass jars for 6 days prior to the addition of the fish. Alternatively, small amounts (0.05%) of spectrograde acetone were added to 2 liters of soil suspension to increase PCB solubility. Water quality analyses. Water quality analyses were performed before the animals were introduced into the aquaria and pH, dissolved oxygen (DO), and temperature were measured at subsequent 24-hr intervals. Temperature was monitored remotely or by using hand-held mercury thermometers. DO and pH were measured using Orion specific ion electrodes, Models 9708 and 9105, respectively, attached to a portable Orion 407A ion meter. Water hardness and alkalinity or salinity, ammonia, and residual chlorine were measured in an aliquot of bioassay water. Calcium carbonate hardness was determined using a Hach HA-71A low range hardness kit and alkalinity with a Hach AL-AP titration kit. Salinity was measured with a hydrometer. Residual chlorine and total ammonia levels were monitored using Orion specific electrodes, Models 9770 and 9520, respectively, attached to an Orion Microprocessor Ionalyzer 90 1. Tissue PCB Analyses At the end of one bioassay, three fathead minnows representing the minimum, mean, and maximum weights were chosen from each treatment for whole body Aroclor 1242 analysis. Fish (0.1 to 1.2 g wet weight) were individually homogenized in 20 ml acetonitrile. The homogenizer and blade were rinsed twice with 20 ml acetonitrile. The rinsings were combined with the homogenate and allowed to stand overnight. Deionized water (12 ml) was added to the homogenate in a separatory funnel. The homogenate was extracted three times with 50 ml n-hexane. The extracts were combined and reduced to approximately 25 ml using a rotoevaporator. Aliquots of the extract were applied to a Tracer MT220 GC equipped with dual electron capture detectors and dual 6 ft X 2 mm i.d. glass columns which were packed with activated ( 1300°F for 4 hr) Florisil(80 to 100 mesh GasChrom-Q coated with 1.5% OV- 17 and 1.95% QF-1). The column temperature was 200°C and prepurified nitrogen was used as the carrier gas at a flow rate of 20 ml/min. Aroclor 1242 concentrations were determined against a U. S. Environmental Protection Agency standard using the largest peak value. RESULTS Fathead Minnows Bioassays were conducted with juvenile fathead minnows having mean standard lengths (SLs) of 33 and 37 mm and loading densities of 0.7 to 0.8 g liter-‘. Preliminary bioassays using the low concentrations series (1 to 25 ppm PCBs) failed to elicit toxicity, so definitive bioassays were performed with PCB concentrations from 50 to 500 ppm. No mortalities occurred (Table 1) and fish behavior was normal. Soluble PCB (Aroclor 1242) concentrations were very low for all treatment levels. In the 500 ppm PCB

F’CB

TOXICITY TABLE

SURVIVAL

1

RATES OF JUVENILE FATHEAD MINNOWS DURING A 96-hr STATIC BIOASSAY USING CONTAMINATED WITH LINE MATERIAL INDUSTRIES CAPACITOR OIL

PCB level of added soil (mm) Control Solvent SO 100 250 500

17

TESTING

control

Aroclor I242 concentration in solution (mb) NDb ND co.3 ~0.6

Nofe. Ten fish were tested per concentration. a Values are corrected for recovery of 86%. b ND = below the minimum detectable limit

Survival Replicate 100 100 100 100 100 100

1

(70) Replicate 100 100 100 100 100 100

2

GILA SOIL

Aroclor 1242 tissue burden (mm wet weight, X + SE) 0.78 0.53 8.32 53.37 60.09 43.61

-t 0.13 + 0.06 f 0.62 f 18.71 f 8.90 2 2.85

of 0.1 ppb.

treatment, 3.4 ppb PCB was detected and in the 250 ppm PCB treatment, only 0.6 ppb PCB. Dissolved PCB levels were otherwise undetectable. Whole body Aroclor 1242 concentrations of unexposed fish averaged less than 1 ppm wet weight PCBs. Tissue PCB concentrations in fish exposed to soil containing 50 ppm PCBs were higher than control levels, but not significantly so (analysis of variance, Student/Newman-Keuls test p = 0.05). Tissue PCB burdens were significantly elevated at the three highest doses, averaging from 44 to 60 ppm PCBs, but were statistically similar to each other. Because of concern that PCB concentrations may not have reached an equilibrium between the dissolved concentrations and molecules adsorbed onto the glass containers during the 96-hr test, a 6-day mixing period was performed before another bioassay was initiated. No mortalities were observed in these tests. Water from the 500 ppm PCB treatment contained 2.0 ppb Aroclor 1242, a value comparable to that obtained in the standard bioassay. Another method to increase PCB solubility was attempted. One milliliter of water was removed from each test aquarium and was replaced with 1 ml of acetone (0.05%) prior to the addition of the soil samples and the introduction of the minnows. Two controls using the solvent control soil (with and without the addition of acetone) were tested. Acetone was added to soil samples containing 50, 100,250, and 500 ppm LMI capacitor oil and a 500 ppm LMI capacitor oil sample was tested without the addition of acetone. At 96 hr, all fish were alive. Fathead minnow fry (~30 day old), averaging 0.1 g were tested. Two out of 10 fish died in the solvent control jar (one each on Days 1 and 3), and one mortality was observed in the 250 ppm PCB jar on Day 4. The mortalities were among the smallest fish in the bioassay, from 0.02 to 0.04 g. Chemical parameters were within acceptable limits during the test. Mortalities probably resulted from stress of maintaining themselves against the current necessary to keep soil in suspension, elevated temperatures (26 to 28”C), or a combination of both factors.

HOSE ET AL.

18 Blacksmith

Definitive bioassayswere completed using juvenile blacksmith; the average SLs of the fish were 44 and 46 mm in the first and second replicates, respectively. Loading rates were 0.5 to 0.6 g liter-‘. Mortalities occurred in aquaria containing 50 (10% dead), 250 (15% dead), and 500 (5% dead) ppm PCB (Table 2). The dead fish were among the smallest fish tested, ranging from 38 to 42 mm SL. Of the fish which died in the first replicate, all three had bacterial lesions present at the termination of the assay.Survivors from the tanks in which the mortalities occurred also had bacterial lesions and infections were confined only to those tanks. Thus the mortalities could not be directly attributed to PCB toxicity. Low dissolved PCB levels were present during the blacksmith bioassays;only 3.5 (250 ppm PCBs in soil) and 3.1 ppb (500 ppm PCBs in soil) Aroclor 1242 were measured. Aqueous PCB levels in 50 and 100 ppm PCB treatments were not detectable. Glass Shrimp Three replicate assayswere conducted using five juvenile-adult shrimp per concentration. Glass shrimp averaged 17 mm in length and 38 mg wet weight. Shrimp were fed daily during the test and were in direct contact with the sediment for the majority of the test, swimming occasionally into the water column. Behavior appeared normal. Three mortalities were observed, one each in tanks containing control (10% dead), solvent control (7% dead), and 500 ppm PCBs (7% dead). One shrimp in the control tank (replicate 2) molted on Day 2 and was subsequently cannibalized. A shrimp in the solvent control aquarium (replicate 3) was missing at the termination of the assay and was presumed dead before being cannibalized. The mortality in the 500 ppm PCB tank was missing on Day 2 of the assay(replicate 1) (Table 3). The third bioassaywas continued for a total of 168 hr with no increase in mortality. Only 1.5 ppb Aroclor 1242 was measurable in the 500 ppm PCB treatment; solvent control and 50 ppm PCB treatments did not contain detectable PCBs. TABLE 2 SURVIVAL

RATES OF JUVENILE CONTAMINATED

BLACKSMITH DURING WITH LINE MATERIAL

A 96-hr STATIC BIOASSAYUSING GILA INDUSTRIES CAPACITOR OIL

SOIL

PCB level of added soil (mm)

PCB concentration in solution (ppb)

Replicate 1

Replicate 2

Control Solvent control 50 100 250 500

NDb ND 3.5 3.1

100 100 90 100 80 100

100 100 90 100 90 90

Survival (%)

Note. Ten fish were tested per concentration. ’ Values are corrected for recovery of 86%. b ND = below the minimum detectable limit of 0.1 ppb.

PCB

TOXICITY TABLE

SURVIVAL

3

RATES OF JUVENILE GLASS SHRIMP DURING A DE~NITIVE 96-hr STATIC BIOASSAY USING GILA SOIL CONTAMINATED WITH LINE MATERIAL INDUSTRIES CAPACITOR OIL

PCB level of added soil (wm) Control Solvent 50 500

19

TESTING

Survival PCB concentration in solution (ppb)

Replicate

ND< 1.5

control

1

Replicate

-b 100 100 80

Note. Five shrimp were tested per concentration. ’ Values are corrected for recovery of 80%. b Treatment was not performed. ’ ND = below the minimum detectable amount

(%) 2

Replicate

80 100 100 100

3

100 80 100 100

Mean 90 93 100 92

of 0.1 ppb.

DISCUSSION Sediment contaminated with up to 500 ppm Aroclor 1242 from Line Material Industries capacitor oil was not toxic to either fish species or glass shrimp under CAM (Department of Health Services, 1983) static exposure conditions for 96 hr. Similarly, exposure to sediment containing up to 500 ppm of an Aroclor 1242 standard was not lethal to the fish species tested (data not shown). These results are consistent with those reported by Halter and Johnson (1977) who exposed fathead minnows to soil containing from 10 to 500 ppm Aroclor 1254 for 32 days with no significant mortality. More recent research has focused on the toxicity of soluble PCBs to aquatic organisms. Representative 4- to ‘I-day LCsos of Aroclor 1242 for fish are 67 to 125 ppb PCBs (Tables 4 and 5). In general, more heavily chlorinated PCB formulations have greater toxicity (Drill et al., 1982); however, Aroclor 1242 is more toxic to rainbow trout than is Aroclor 1254 (Johnson and Finley 1980). Certain invertebrates, such as crustaceans, appear to be more sensitive to PCB toxicity than are fish (Table 5). Soluble PCB levels generated under CAM (Department of Health Services, 1983) conditions in the static tests were quite low, with sediment loads of at least 250 ppm PCBs necessary to yield detectable water concentrations. In freshwater tests, from 1.5 to 3.4 ppb Aroclor 1242 was measured in 500 ppm LMI capacitor oil treatments.

TABLE FIVE-DAY

LCso VALUES OF VARIOUS UNDER

Aroclor

PCB FORMULATIONS

FLOW-THROUGH

formulation

Johnson

FOR JUVENILE

RAINBOW

CONDITIONS

LGo

1242 1248 1254 1266 Note. Data are from

4

(mb)

67 54 142 >232 and Finley

(1980).

TROUT

HOSE ET AL.

20

TABLE

5

RELATIVE SENSITIVITY OF FISH AND INVERTEBRATESTO AROCLORS 1242 AND 1254 Aroclor formulation

Go (ppb)

Test organism

1242

Rainbow trout Bluegill Amphipod (Gammarus pseudolimnaeus) Crayfish (Oronectes nais)

1254

Rainbow trout Bluegill Glass shrimp Amphipod (G. pseudolimnaeus) Crayfish (0. nais)

Conditions

61 125 10 30

5-day, flow-through 5-day. flow-through 4-day, flow-through T-day

142 2740 3 2400 100

5-day, flow-through 4-day 7-day, flow-through 4day 4-day

Note. Data are from Johnson and Finley (1980).

Using sediment with 500 ppm Aroclor 1254, Halter and Johnson (1977) measured PCB water levels of 0.6 ppb at time 0 and 7.6 ppb at equilibrium (Day 8). Water used in their tests was hard (329 ppm CaC03) and temperatures were slightly lower. According to CAM specifications, a maximum soil concentration of 500 mg dry weight liter-’ was used while in Halter and Johnson’s experiment 66.7 g liter-’ was tested. The much higher soil concentration in their study was probably responsible for the higher dissolved PCB levels but it is important to note that this produced only a relatively slight increase in dissolved PCB levels. Our results shown that soluble PCB concentrations generated during the seawater bioasssay were similar to those measured during freshwater tests. Some of the PCB soil burdens used in this study (50 to 100 ppm dry weight) are representative of those found in polluted aquatic environments (Table 6). Point estimates of PCB concentrations in Hudson River sediments, the most contaminated water system in the Atlantic area, reach 140 ppm (NAS, 1979). The higher concentrations of 250 and 500 ppm PCBs which were tested are levels which may be present following spillage of PCB transformer or capicitor fluid on land. Mackay et al. ( 1982) estimated that PCB levels following a spill would reach a concentration of a few percent TABLE

6

REPRESENTATIVEENVIRONMENTAL LEVELS OF PCBs IN WATER AND AQUATIC SEDIMENTS Source

PCB concentration

Reference

Marine sediment, Florida Marine sediment, Santa Monica Bay Marine sediment, Meditemurean Sea Freshwater sediment, estimated mean Seawater, deep Mediterranean Sea Seawater, surface California Fresh water, Great Lakes Fresh water, estimated mean

61 ppm drywt 1.6 ppm wet wt O.lppmdrywt 0.06 ppm 0.8 x lo-’ ppb 1.1-5.9 ppb 0.8-37 ppb 1.I ppb

Nimmo et al., 197 1 Schafer et al., 1982 Fowler et al., 1978 Mackay et al., 1982 Elder et al., 1976 NAS, 1979 NAS, 1979 Mackay et al., 1982

PCB TOXICITY

TESTING

21

(1% equals 10,000 ppm). Dissolved PCB levels generated in this study of ~3.5 ppb represent levels commonly measured in California coastal waters during the early 1970s and in major U. S. rivers (Table 6). A 96-hr exposure of fathead minnows to soil PCB levels between 50 and 500 ppm, which would be designated nonhazardous using the CAM (Department of Health Services, 1983) aquatic bioassay criterion, resulted in tissue levels exceeding the recent U. S. Food and Drug Administration action limit of 2 ppm. These tissue concentrations are 2 to 10 times higher than those reported for minnows similarly exposed to Aroclor 1254 adsorbed onto soil (Halter and Johnson, 1977). Laboratory experiments using young striped bass (Morone saxatilis) have demonstrated that accumulation resulted in part from desorption of sediment-bound PCBs and that a dietary component is also involved (Califano et al., 1982; Pizza and O’Connor, 1983). Our observations that PCB tissue burdens in fathead minnows reach a threshold at soil concentrations of 2100 ppm and that dissolved PCBs are undetectable using sediment-bound PCB levels of 100 ppm suggest that direct exposure to contaminated particles is the more important. Halter and Johnson (1977) showed that direct contact to sediment-sorbed PCBs greatly enchances bioaccumulation and suggest that the typical fathead minnow behavior of substrate grazing was responsible for the unexplained increase. Our results underscore the need for caution in the routine application of mathematical models of toxicant bioconcentration without supporting observations on species-specific feeding and behavior patterns. While acute exposure to sediments containing 50 to 500 ppm Aroclor 1242 PCBs and to dissolved PCB concentrations of up to 3.5 ppb was not toxic to the animals tested, chronic toxic effects have been reported following exposure of aquatic biota to similar PCB levels (NAS, 1979). Exposure to sediment containing 30 ppm Aroclor 1254 for 90 days was lethal to marine worms (Fowler ef al., 1978). Reproduction and growth of fathead minnows was affected at concentrations of 2.2 ppb Aroclor 1248 and >1.8 ppb Aroclor 1254 (Nebeker et al., 1974). The low PCB residues leached from highly contaminated sediments may deleteriously affect long-term processes of aquatic organisms such as growth and reproduction. Jenkins et al. ( 1982) and Perkins et al. (1982) have reported that white croaker (GenyonevMus heatus), which live off Palos Verdes, California, an area characterized by high chlorinated organic hydrocarbon levels (sediment PCB concentrations up to 10.1 ppm), have liver changes which correlate with sediment chlorinated hydrocarbon concentrations. Using several of the CAM (Department of Health Services, 1983) criteria (bioaccumulation, environmental persistence, and a PCB concentration greater than the limit of 50 ppm TTLC), most of the soils tested would be classified as toxic. None, however, demonstrated acute toxicity using the CAM aquatic bioassay procedure with the suggested fish species. Moreover, several modifications which could be included in the procedure with a minimum of labor or expense (such as testing a reportedly more sensitive species or life history stage or increasing solubility by chemical or physical methods) did not alter the observed lack of toxicity. Many other environmentally significant pollutants such as high-molecular-weight petroleum hydrocarbons, organochlorine pesticides, and wood-preserving chemicals are by virtue of their chemical structure bioaccumulative and environmentally persistent like PCBs. More specific examples include aromatic hydrocarbons like anthracene and benzo[a]pyrene which are generated during the production of fossil and

22

HOSE ET AL.

synthetic fuels and pesticides or herbicides such as Kelthane, DDT, dieldrin, Mirex, and chlorophenols. These compounds possesshigh octanol:water partition coefficients, a measurement of hydrophobicity, as well as high soil adsorption coefficients (McCall et al., 1980). Thus, binding of these hydrophobic compounds to soils, particularly those of low organic carbon content, tends to mitigate toxicity since very low (usually ppb) levels of toxicant are leached from soils under environmental conditions and adsorbed compounds may also have reduced bioavailability (McCarthy, 1983). Many of these chemicals, for example, DDT and benzo[a]pyrene, have very low acute toxicity, and short-term tests such as the 96-hr fish bioassay described in the CAM (Department of Health Services, 1983) will not adequately evaluate their general toxicity. As in the case of PCB-contaminated soils, chronic toxicity testing appears to be a more realistic assessment of general toxicity of hydrophobic chemicals adsorbed onto soils. STLC and TTLC determinations proposed in the CAM could be more accurate predictors of general toxicity if limit values are based upon valid toxicity studies. CONCLUSIONS Toxicity information derived from static 96-hr aquatic bioassays do not provide an accurate assessment of the toxicity of hydrophobic compounds such as PCBs. Chronic toxicity studies and estimations of bioaccumulation potential appear to be more appropriate indictors of the toxicity of these compounds and hazardous waste regulations should be based upon these data. ACKNOWLEDGMENTS This work was supported by the Southern California Edison Company under Contract C-305-2903. Tissue PCB analysis was performed by Richard Gossett of the Southern California Coastal Water Research Project. We thank Waheedah Muhammad for typing the manuscript.

REFERENCES American Public Health Association (APHA) (1975). Standard Methods for the Examination of Water and Wastewater. 14th ed. APHA. CALIFANO, R. J., O’CONNOR, J. M., AND HERNANDEZ, J. A. (1982). Aquat. Toxicol. 2, 187-204. Department of Fish and Game (1976). Guidelines for Performing Static Acute Toxicity Fish Bioassays in Municipal and Industrial Wastewaters. State of California State Water Resources Control Board, Dept. of Fish and Game, Sacramento. Department of Health Services (1983). Criteria for ZdentiJcation of Hazardous and Extremely Hazardous Wastes: Proposed Revisions to Title 22. State of California Department of Health Services, Sacramento. Drill, Friess, Hayes, Loomis and Schaffer, Inc. (1982). Comments and Studies on the Use of Polychlorinated Biphenyls in Response to an Order of the U. S. Court ofAppeals for the District of Columbia Circuit, Vol. II, Potential Health purities. Arlington,

Eflects

in the Human

from

Exposure

to Polychlorinated

Biphenyls

and Related

Zm-

Va. ELDER, D. L., PARSI, P., AND HARVEY, G. R. (1976). Polychlorinated biphenyls in seawater, sediment and over ocean air of the Mediterranean. In Activities of the International Laboratory ofMarine Radioactivity, I976 Report, pp. 136- 151. International Atomic Energy Agency, Vienna. FOWLER, S. W., POLIKARPOV, G. G., ELDER, D. L., PARSI, P., AND VILLENEUVE, J.-P. (1978). Mar. Biol. 48, 303-309. HALTER, M. T., AND JOHNSON, H. E. (1977). A model system to study the desorption and biological

PCB TOXICITY

TESTING

23

availability of PCB in hydrosoils. In Aquatic Toxicology and Hazard Evaluation (F. L. Mayer and J. L. Hamelink, eds.), ASTM STP634, pp. 178-195. Amer. Sot. for Testing and Materials, Philadelphia. HAQUE, R., SCHMEDDING, D. W., AND FREED, V. H. (1974). Environ. Sci. Technol. 8, 139-142. HARVEY, G. R., AND STEINHAUER, G. W. (1976). Biogeochemistry of PCB and DDT in the North Atlantic. In Environmental Biogeochemistry Vol. 1, Carbon, Nitrogen, Phosphorus, S&r and Selenium Cycles (J. 0. Nriagu, ed.), pp. 203-22 1. Ann Arbor Science, Ann Arbor, Mich. JENKINS, K. D., BROWN, D. A., HERSHELMAN, G. P., AND MEYER, W. C. (1982). Contaminants in white croakers Genyonemus lineatus (Ayres, 1855) from the Southern California Bight. I. Trace metal detoxiIication/toxiIication. In Physiological Mechanisms of Marine Pollutant Toxicity (W. B. Vemberg et al., eds.), pp. 177- 195. Academic Press, New York. JOHNSON,W. W., AND FINLEY, M. T. (1980). Handbook ofAcute Toxicity of Chemicals to Fish and Aquatic Invertebrates: 196.5-1978,

Summaries

of Toxicity

Tests Conducted

at Columbia

National

Fisheries

Research

Laboratory,

Fish and Wildlife Service Resource Publ. 137. U. S. Dept. of the Interior, Washington, D. C. MACKAY, D., AND Institute for Environmental Studies (1982). Comments and Studies on the Use of Polychlorinated Biphenyls in Response to an Order of the U. S. Court of Appeals for the District of Columbia Circuit, Vol. IV, Environmental Pathways of Polychlorinated Biphenyls. University of Toronto, Toronto. MAYER, F. L., MEHRLE, P. M., AND SANDERS,H. 0. (1977). Arch. Environ. Contam. Toxicol. 5, 501-5 11.

MCCALL, P. J., SWANN, R. L., LASKOWSKI, D. A., UNGER, M., VRONA, S. A., AND DISBURGER, H. J. (1980). Bull. Environ. Contam. Toxicol. 24, 190-195. MCCARTHY, J. F. (1983). Arch. Environ. Contam. Toxicol. 12, 559-568. National Academy of Sciences (NAS) (1979). Polychlorinated Biphenyls. NEBEKER, A. V., PUGLISI, F. A., AND DEFOE, D. L. (1974). Trans. Amer. Fish. Sot., 562-568. NIMMO, D. R., WILSON, P. D., BLACKMAN, R. R., AND WILSON, A. J. (1971). Nature (London) 231, 5052.

PERKINS, E. M., BROWN, D. A., AND JENKINS, K. D. (I 982). Contaminants in white croakers Genyonemus linatus (Ayres, 1855) from the Southern California Bight. III. Histopathology. In Physiological Mechanisms of Marine Pollutant Toxicity (W. B. Vemberg et al., eds.), pp. 2 15-23 1. Academic Press, New York. PIZZA. J. C., AND O’CONNOR, J. M. (1983). Aquat. Toxicol. 3, 3 13-327. SCHAFER,H. A., HERSHELMAN, G. P., YOUNG, D. R., AND MEARNS, A. J. (1982). Contaminants in ocean food webs. In Coastal Water Research Project Biennial Report 1981-82 (W. Bascom, ed.), pp. 17-28. Southern California Coastal Water Research Project, Long Beach.