Investigations into Using the Nematode ... - Springer Link

Arch. Environ. Contam. Toxicol. 33, 252–260 (1997)

A R C H I V E S O F

Environmental Contamination a n d Toxicology r 1997 Springer-Verlag New York Inc.

Investigations into Using the Nematode Caenorhabditis elegans for Municipal and Industrial Wastewater Toxicity Testing D. R. Hitchcock, M. C. Black, P. L. Williams Environmental Health Science, College of Agricultural and Environmental Sciences, University of Georgia, Athens, Georgia 30602, USA

Received: 21 September 1996/Accepted: 15 July 1997

Abstract. This investigative study assesses the ease and usefulness of the nematode Caenorhabditis elegans for identifying contributors to effluent toxicity within an industrial and municipal wastewater treatment plant (WWTP) system. Several different types of industries, including fiberglass manufacturing, paper packaging, and yarn dyeing, discharge effluent into the municipal wastewater treatment plant, which in turn discharges into a local creek. A major objective of this study was to identify primary sources of toxicity throughout the system with a nematode toxicity test. Twenty-four-hour composite water samples were taken periodically over a ten-month period at five strategic points within the system: (1) at the point of discharge at each of the three industries, (2) at the combined industrial influent of the wastewater treatment plant, (3) at the effluent of the WWTP, (4) upstream of the WWTP discharge, and (5) downstream of the WWTP discharge. Samples were analyzed for basic water chemistry, and each sample was tested for whole effluent toxicity using a 72-h nematode test with mortality as the end point. Results suggest that interactions between the wastewaters of certain industries may increase the overall nematode toxicity in the wastewater treatment facility’s composite influent and effluent. Nematode mortality trends indicate relatively high toxicity levels in wastewater entering the WWTP from contributing industries. High WWTP influent toxicity may potentially be due to varying flow rate ratios of industrial discharges, release of varying toxic constituents in wastewaters, and toxic interactions between chemical constituents of industrial wastewaters. The evaluation of toxicity within the treatment system may pinpoint locations where pollution prevention strategies may be implemented to reduce overall toxicity at the point of discharge.

Standardized methods have not been established for utilizing a Caenorhabditis elegans test in the toxicity assessment of wastewaters, but current published literature provides vital data and methodologies that bring the nematode toxicity test closer to standardization for use in effluent toxicity assessment. This literature includes the establishment of reference toxicant data

Correspondence to: P. L. Williams

for C. elegans (Cressman and Williams 1997), the creation of control charts for intralaboratory precision using the nematode toxicity test (Freeman et al. 1998), and the development of methods for using the C. elegans test in the ecotoxicological assessment of aquatic sediments (Traunspurger et al. 1997). A major benefit of the nematode toxicity test is its low cost. Using the methods presented herein, the total estimated supply and labor cost for conducting a test on 10 samples is approximately $75, assuming that a laboratory is properly equipped (e.g., centrifuge, dissecting microscope, incubators, autoclave, etc.) and that workers are paid a rate of $10/h (estimated to be 6 h of labor). The free-living bacterivorous nematode C. elegans has been demonstrated to embody many characteristics necessary for its use as an aquatic toxicity test organism. It inhabits the aquatic component of the soil environment, or interstitial water (Freckman 1988; Popham and Webster 1982). Because the nematode lives in the interstitial water between soil particles, toxicity studies using C. elegans can be conducted in aquatic media (Williams and Dusenbery 1990; Donkin and Williams 1995). The ability to culture age-synchronized populations of C. elegans with relative ease (Cox and Edgar 1981) is a benefit in any toxicity test. Test organisms ideally should be of the same size, age, health, and reproductive status prior to test initiation (US EPA 1993). The potential for genetic variability in nematode populations is minimized by the hermaphroditic nature of the worm (Brenner 1974). There is minimal variation in sensitivity within test populations, an important factor in the accuracy and precision of organismal toxicity tests (US EPA 1993). Concentration–response relationships have been established based on the mortality of the nematode after exposure to toxic chemicals, especially metals (Williams and Dusenbery 1988, 1990; Donkin and Williams 1995). C. elegans has been used to identify toxic constituents within a chemical mixture (Donkin et al. 1995) and can tolerate wide pH, salinity, and water hardness ranges (Khanna et al. 1997). Therefore, the nematode toxicity test may discriminate for toxicity based on an effluent’s toxic constituents, rather than water quality parameters. Thus, the whole effluent toxicity of a given water sample may be evaluated with minimal, if any, manipulation of that sample. Finally, nematodes are generally indigenous to the biofilms of sewage beds existing in wastewater treatment plants and are important to its proper functioning in wastewater treatment (Atlas and Bartha 1987).

Assessing Wastewater Toxicity Using C. elegans

The system of concern in this study is an industrial park and wastewater treatment facility located in the city of Washington in Wilkes County, Georgia, USA, an area known for a high diversity of industrial activity. Industries considered in this study include a yarn dyeing facility (dyehouse), a fiberglass manufacturer, and a pulp processing facility. The effluents of all three industries combine prior to entrance into the Washington municipal wastewater treatment plant (WWTP), where this influent subsequently mixes with town sewage waste, undergoes biological treatment, and is discharged into a stream known as Rocky Creek. Mean flow rates of the dyehouse, fiberglass manufacturer, and pulp processing plant are 600,000– 700,000 gallons per day (gpd), 50,000 gpd, and 75,000–90,000 gpd, respectively [B. Rucker, Operations Management International (OMI), Washington, GA, personal communication]. In late 1992 and early 1993, a toxicity reduction evaluation (TRE), including a toxicity identification evaluation (TIE), was conducted on the discharge of the wastewater treatment plant into Rocky Creek. Results of this evaluation demonstrated that effluent toxicity is due primarily to high salt, chlorine, and metal (Cu and Zn) levels [BMI (Biological Monitoring, Inc.) 1994]. An industrial pretreatment program has been implemented by the city of Washington and employees of the WWTP (B. Rucker, OMI, Washington, GA, personal communication) to monitor the discharges of individual industries based on the National Pretreatment Program (NPP) guidelines (US GA Office 1989). Although the industries have been monitored on a categorical basis prescribed by NPP guidelines, the WWTP effluent still failed toxicity tests with the invertebrate Ceriodaphnia dubia in August 1995 (GA EPD 1995). The no-observedeffect concentration (NOEC) for reproduction equaled 60% effluent, and the lowest-observed-effect concentration (LOEC) for reproduction equaled 80% effluent. The instream wastewater concentration (IWC) was estimated to be 99.4% effluent under critical low flow conditions. Since the NOEC (60% effluent) was less than the IWC (99.4% effluent), instream chronic toxicity was predicted based on the daphnid tests. The objectives of this research were: (1) to utilize the nematode C. elegans to evaluate toxicities of aquatic samples collected from strategic locations in an industrial and municipal wastewater treatment discharge system; (2) to pinpoint the source(s) contributing to any toxicity in the composite effluent from the wastewater treatment plant; (3) to determine trends in toxicities, if any, at each sample location; and (4) to determine if interactions of mixed effluents contribute to any existing toxicity. In comparison to TIEs, which are generally performed on fractionated composite effluent samples from the WWTP discharge (US EPA 1991), this study incorporates the C. elegans 72-h aquatic toxicity test to determine relative toxicities by location for whole and mixed samples throughout the industrial and municipal wastewater discharge system. This research demonstrates the usefulness of the nematode aquatic toxicity test in assessing toxicity in effluent samples collected from the field.

Materials and Methods Experimental Design Our study consisted of a randomized complete block design in which treatments were considered to be nematode exposure to either a sample

253

from a specific location for the whole-effluent tests or to simulated samples created in the lab for mixed-sample tests. Nematode mortality data for each sample period were considered to be blocked. Eight to 12 nematodes were exposed to each treatment in each of six wells (replications), and our measurement end point was the percent of nematode mortality upon exposure to a given treatment. Although toxic effects are considered to be normally distributed within a population, our analyses within treatments were based on a binomial probability distribution, because when counting nematodes during the course of the study, an individual nematode was observed to be either alive or dead. Controls consisting of nematode exposure to K medium (Williams and Dusenbery 1990) were established for each test conducted, and a test was considered invalid if control mortality was greater than 10% (Williams and Dusenbery 1990; US EPA 1993).

Sampling Locations Water samples were taken periodically over ten months (October 1995 to July 1996) at five strategic points within the system (Figure 1): (1) at the point of discharge of the dyehouse, (2) before the combined industrial influent entered the WWTP, (3) at the effluent of the WWTP, (4) upstream of the WWTP discharge, and (5) downstream of the WWTP discharge. The dyehouse was targeted initially in this study based on its major contribution to influent into the WWTP (approximately 85%) (B. Rucker, OMI, Washington, GA, personal communication). Sampling of the two other industries (fiberglass manufacturer and pulp processing plant) on the industrial/WWTP influent line began in February 1996, after concluding that the dye wastewater was generally not toxic to the nematode relative to control mortality data (,10% mortality).

Sampling Methods Twenty-four-hour composite water samples were taken at the locations described previously. Portable automated water samplers (ISCO model 3700 Standard; Isco Company, Lincoln, NE) were used for upstream and downstream sampling. Influent and effluent wastewater samples were collected in a similar fashion by WWTP personnel (OMI), and effluent samples were provided by the appropriate industries. Grab samples were taken in place of composite samples in May and July 1996, due to difficulty in coordinating the sample collection. Samples were collected in 350-ml glass containers that were washed with 10% HNO3 prior to use. All samples were taken within 6 h of one another to ensure homogeneity in the relative composition of each sample between locations and throughout the treatment system. Water and wastewater samples were kept at 4°C throughout collection, transport, and storage until toxicity test initiation. Samples were not altered in any way before toxicity testing unless, as described below, mixtures were made from industrial samples in order to simulate influent proportions based on mean flow rates.

Simulated Influent Mixtures Industrial samples collected in May and July 1996 were not only tested individually for whole effluent toxicity, but were also mixed proportionally prior to toxicity testing in order to simulate the influent into the wastewater treatment plant. A ratio of 85:5:10 for the dyehouse, fiberglass plant, and pulp processing plant, respectively, was estimated to be the flow ratio exhibited into the WWTP (B. Rucker, OMI, Washington, GA, personal communication). Using May industrial effluent samples, mixtures were made to bracket this flow ratio. In July, the dyehouse was off-line at the time of sampling and did not contribute

254

D. R. Hitchcock et al.

alive, gentle probing with a sterile platinum wire was performed if necessary.

Water Chemistry Water chemistry analyses for each sample were conducted upon toxicity test initiation. Sample temperature, conductivity, salinity, and total dissolved solids (TDS) were measured with a conductivity meter (model 130; Orion Co., Boston, MA). Total alkalinity (as mg/L CaCO3) and water hardness (as mg/L CaCO3) were measured using titration kits (Hach Chemical Co., Loveland, CO). pH was measured using a pH meter (model 720A; Orion Co.). Fig. 1. Schematic diagram of sampling locations surrounding the Washington, GA, wastewater treatment plant. Wide arrows indicate locations of sample collection

to influent into the WWTP (B. Rucker, OMI, Washington, GA, personal communication); therefore, some mixture ratios were adjusted to include only the fiberglass and pulp effluents. In addition, mixtures were prepared containing all industrial effluents for comparison with the May simulated influent mixtures.

Nematode Culturing Methods Nematode populations were cultured prior to sample collection using published techniques (Williams and Dusenbery 1990; Donkin and Williams 1995). The dauerlarva (dauer) alternative life-stage is exhibited by many free-living nematodes when subjected to environmental stresses such as starvation and overcrowding (Cassada and Russell 1975). The techniques used in culturing nematode populations for toxicity testing included culturing the dauer stock population. Prior to sample collection and subsequent test initiation, 100 to 200 dauers were transferred to agar plates containing an Escherichia coli (OP50 strain) (Brenner 1974) bacterial lawn, which allowed worms to develop out of their dauer stage to adulthood. In 2–3 days, eggs from resulting adult worms were isolated from agar plates using a Clorox solution (10% in deionized water) and subsequent rinsing with K medium. These eggs developed into the synchronized adult worm populations to be used in toxicity testing.

Toxicity Testing Methods Static nonrenewal toxicity tests were initiated within 36 h of sample collection. K medium was used as a control to ensure the health of the test population, and sterile handling procedures were followed while setting up the test. A nematode food source, consisting of pelleted E. coli (OP50 strain), was prepared as described by Donkin and Williams (1995) for each sample to be tested. Each pellet was centrifuged (3000 rpm, 7 min) from 30 ml of a saturated L broth culture (Donkin and Williams 1995), rinsed three times with K medium, and resuspended in 10 ml of sample. One milliliter of each sample/food mixture was pipetted into each of six wells of a Costar 3512 well dish plate (Corning Co., Kennebunk, ME). Subsequent to sample preparation, synchronized adult nematodes were rinsed and collected from culture plates 3–4 days after hatching. Eight to 12 worms were loaded into each of the six wells per sample (totaling approximately 60 worms per sample) using a glass capillary tube attached to a 10-µl micropipet tip. Once loaded, the worms were incubated at 20°C for 72 h, and survival counts were taken under a dissecting microscope every 24 h (6 1). To ensure that a worm was

Statistical Analyses Standard error within samples from given locations was calculated based on the binomial probability distribution (Anderson et al. 1994). Mortality data obtained periodically for each sample location were tested for homogeneity of variance using Bartlett’s test. Because the data were not normally distributed, a Kruskal-Wallis nonparametric analysis of variance (ANOVA) was used to determine if significant differences in means within periods existed, while a distribution-free multiple comparison test was used to determine where significant differences exist (Gad and Weil 1988). Friedman’s two-way analysis of variance by ranks test (Spence et al. 1968) was used to determine rank correlation coefficients by which consistency in ranking toxicity values by sample location was evaluated between sample periods. Coefficient of variation [CV 5 (standard deviation/grand mean) 3 100] values were calculated for percent nematode mortality data at each sample location across entire sample duration.

Results Water Chemistry Data Table 1 summarizes the water chemistry data for samples analyzed prior to toxicity test initiation. The ranges of pH and the ranges and means for conductivity, salinity, total dissolved solids (TDS), total alkalinity, and water hardness for each sample location are given. Values are reported only for sampling periods when control mortality was less than 10%: October and December 1995, and March, May, and July 1996.

Nematode Mortality Data Seventy-two-hour mortality data are presented by sample location for October and December 1995, and March, May, and July 1996 (Figure 2a–e). Toxicity data from certain sample periods were omitted due to an inability to count worms because of solids in some samples (February 1996) and nematode control mortality greater than 10% (January, April, and June 1996). Homogeneity of variance at sample locations between sample periods was tested using Bartlett’s test for homoscedasticity. Bartlett’s test demonstrated a significant (p , 0.05) difference (heterogeneity) in variances at all sample locations between sample periods. Therefore, periodic data for mortality at each sample location could not be statistically pooled to determine an overall toxicity rank based on location for the entire sample duration. However, mortality data within


255

Table 1. Water chemistry data for samples collected at Washington, GA, WWTP and vicinity for October and December 1995 and March, May, and July 1996

Sample Location

pH Range

Dyehouse effluent Fiberglass effluent Pulp effluent WWTP influent WWTP effluent Upstream Downstream

8.5–9.1 8.0–9.4 7.1–9.6 8.0–8.7 7.4–7.7 7.4–7.5 7.5–7.6

Conductivity (µS/cm)

Salinity (ppt)

Total Dissolved Solids (mg/L)

Alkalinity (as mg/L CaCO3 )

Hardness (as mg/L CaCO3 )

Range

Mean

Range

Mean

Range

Mean

Range

Mean

Range

Mean

1180–1599 67.3–72.1 99.8–121.8 157.9–1125 339–455 26.3–27.0 225–277.3

1390 74.7 110.8 641.5 397 26.6 251

0.5–0.7 0 0 0.0–0.4 0.0–0.4 0 0

0.6 0 0 0.2 0.2 0 0

1200–1600 68–82 100–122 158–1126 340–460 26–28 225–278

1400 75 112 645 400 27 251

100–304 34–98 98–160 68–240 52–120 18–22 34–108

207 68.7 130 175 94.5 20 62

32–112 50–74 34–138 20–88 44–70 30–36 32–48

70.5 60.7 110.7 46.5 54 33 38

Fig. 2. 72-h mortality data by sample location for each sample period. DH 5 dyehouse effluent; FG 5 fiberglass effluent; PP 5 pulp processing effluent; INF 5 wastewater treatment plant influent; EFF 5 wastewater treatment plant effluent; UP 5 upstream of WWTP discharge; DWN 5 downstream of WWTP discharge. Data for October and December 1995 and from March, May, and July 1996 are designated (a) through (e), respectively. Locations without bars represent zero percent nematode mortality. Error bars represent standard error based on the binomial probability distribution of the data (n 5 6)

each sample period, except March 1996, exhibited significant differences (p , 0.05) shown by using a Kruskal-Wallis nonparametric ANOVA to test for significant differences between mean mortality data for each location within each sample

period. Tests for multiple comparisons indicated which locations had significant mortality differences (p , 0.05) within each sample period. In October 1995, nematode mortality was significantly higher in the WWTP influent sample than in the

256

upstream water sample. In December 1995, mortality was significantly higher in the WWTP influent sample than in the downstream water sample and the dyehouse wastewater sample. In May 1996, mortality was significantly higher in the fiberglass wastewater sample than in the WWTP effluent sample, the pulp wastewater sample, and the dyehouse effluent sample. In July 1996, mortality was significantly higher in the WWTP influent sample than in the dyehouse wastewater sample and the upstream water sample. In order to evaluate variability in toxicity tests between sample periods at given sample locations, rankings based on mortality data at sample location were assigned for each sample period (Table 2). Because all three industries considered in this study were sampled only after February 1996, nonparametric methods were used to analyze the rank-order correlations separately for the five sampling locations (dyehouse effluent, WWTP influent, WWTP effluent, upstream and downstream of Rocky Creek) included in all five sample periods (n 5 5), and for all seven locations (including fiberglass and pulp processing industries) for three sample periods (n 5 3). Results of the Friedman two-way ANOVA by ranks, a rank test for two or more matched groups, indicated a lack of correlation of rank (a 5 0.05) in nematode mortality for samples collected and tested between the sample periods in March and July 1996, which included all industries, while ranks of toxicity were correlated in samples for the five sample periods when the fiberglass and pulp processing effluents were not sampled. Figure 3a and b represent mean 72-h mortality data by sample location for test results from October 1995 to July 1996 and from March to July 1996, respectively. Although means over the entire sample duration at locations could not be used in ranking, data by sample location were pooled to show the degree of variation by standard error. Figure 3a shows that mortality data from WWTP influent and WWTP effluent samples had the greatest degree of standard error for the five periods with valid data from October 1995 to July 1996. Figure 3b demonstrates that the fiberglass manufacturer samples had the highest standard error for March, May, and July 1996, relative to the other sample locations. Table 3 summarizes the results of CV calculations from mortality values in samples collected and tested over the October 1995 to July 1996 period (n 5 5) and the March to July 1996 period (n 5 3).

Simulated Influent Mixtures Figure 4a and b represent the 72-h mortality data from the simulated influent mixtures from May and July 1996, respectively. In both figures a dotted line indicates the nematode mortality for actual WWTP influent for each sample period. In May 1996 (Figure 4a) nematode mortality increased (40%) at a dyehouse–fiberglass–pulp effluent ratio of 75:12.5:12.5 and further increased (63.2%) at a ratio of 70:15:15. The percent nematode mortality for the WWTP influent sample for May 1996 was 23.1%. In July 1996 (Figure 4b) nematode mortality increased (27.1%) at a ratio of 0:30:70 for dyehouse–fiberglass– pulp effluent. The percent nematode mortality for July 1996 of the WWTP influent sample was 27.6%.


Discussion Water Chemistry Data Khanna et al. (1997) found that C. elegans can tolerate a pH range of 3.2–11.8 for 96 h in K medium and 3.4–11.7 for 96 h in moderately hard reconstituted water (US EPA 1993) without mortality. Nematodes survived salinities of 12.5 g/L NaCl and 9.2 g/L KCl in K medium, and 20.5 g/L NaCl and 18.5 g/L KCl in moderately hard reconstituted water (Khanna et al. 1997). The nematodes tolerated water hardnesses up to 140–145 mg/L CaCO3 (Khanna et al. 1997). All samples collected for this study exhibited pH, salinity, and water hardness levels within C. elegans tolerance ranges; therefore, any nematode mortality can not be directly attributed to these specific water-quality parameters (Table 1).

Nematode Mortality Data Nematode mortality above 10% occurred in controls during three of the nine periods of sample collection. This mortality may be attributed in part to an insufficient amount of food source (E. coli) provided for the nematode during sample preparation. In our study, a 1:3 volumetric ratio of sample to saturated L broth before centrifugation was used (10 ml sample to 30 ml L broth). However, recent literature suggests using only a 1:1 ratio (10 ml sample to 10 ml L-broth) for 72-h tests in aquatic media using photometric analysis to determine bacterial cell concentration, hence ensuring culture saturation (Traunspurger et al. 1997). In our study, photometric techniques were not used to ensure L-broth saturation. However, equal sample/ food source ratios were used for all tests, and for tests conducted during six of the nine periods of sample collection, control mortality remained below 10%. For these samples, the food source was apparently sufficient, and the corresponding tests were considered valid. Reference toxicants also can provide information on test failure (US EPA, 1993). At the start of this study, no reference toxicant data for C. elegans had been established, but a separate concurrent project within the laboratory did evaluate several potential reference toxicants (Cressman and Williams 1997). The data from that study demonstrate that there was no change in the health or sensitivity of the worms used (i.e., both studies used worms from the same culture). This finding further supports the lack of food as the cause of the high control mortality during some sampling periods. Upon initial sampling and subsequent toxicity testing during the months of October and December 1995, high mortality of C. elegans in the WWTP influent samples suggested high levels of toxicity (Figure 2a and b). Relative to influent toxicity, WWTP effluent samples consistently demonstrated reduced but still elevated (35–45%) toxicity based on nematode mortality after a 72-h exposure duration. However, the dyehouse, being a major contributor of flow into the WWTP (approximately 85%), showed low toxicity in comparison to the WWTP influent for these sample periods (Figure 2a and b). Therefore, another industry or industries on the WWTP influent line could have contributed to the high toxicity values in WWTP influent samples. Sample collection and subsequent toxicity testing


257

Table 2. Assigned toxicity ranks based on 72-h C. elegans mortality data for sample locations by sample perioda Sample Location Sample Period

Dyehouse

Fiberglass

Pulp

WWTP Influent

WWTP Effluent

Upstream

Downstream

October December March May July

4 ,10% ,10% ,10% ,10%

— — ,10% 1 3

— — ,10% ,10% 2

1 1 1 2 1

2 3 ,10% ,10% ,10%

,10% 2 ,10% 4 ,10%

3 4 ,10% 3 ,10%

1 5 highest toxicity. Designations of ,10% indicate that the nematode mortality percentage for a given sample was less than the upper limit for acceptable control mortality (10%) for test validity. —, no rank available due to lack of sample collection

a

Table 3. Means (µ), standard deviations (SD), standard errors (SE), and coefficients of variation (CV) of percent mortality data across sampling periods

Fig. 3. Mean 72-h mortality data by sample location from (a) October 1995 to July 1996 (n 5 5) and (b) March to July 1996 (n 5 3). Error bars represent standard errors calculated from standard deviations between sampling periods [SE 5 standard deviation/Ï (number of samples 2 1)]

from all industries on the WWTP influent line during March, May, and July 1996 suggested that individual industries do not appear to be the primary contributors to WWTP influent toxicity (Figure 2c–e), with the exception of the fiberglass manufacturer during the month of May 1996 (Figure 2d). Toxicity data from stream sampling appeared to show some consistency, with downstream samples being influenced by toxicity from WWTP effluent, especially in October and July (Figure 2a and e). Some upstream samples, especially in

Sample

µ

SD

SE

CV (%)

Dyehouse (n 5 5) (n 5 3) Fiberglass (n 5 3) Pulp processing (n 5 3) WWTP influent (n 5 5) WWTP effluent (n 5 5) Upstream (n 5 5) Downstream (n 5 5)

8.2 1.2 32.4 7.4 45.4 24.8 15.3 20.6

11.6 2.1 34.1 8.7 24.0 24.8 11.8 13.4

7.2 1.2 19.7 5.0 14.9 10.6 7.0 6.3

133.7 58.6 105.3 117.5 59.4 110.2 89.9 65.7

December, showed significant (p , 0.05) levels of mortality to the nematode (Figure 2b). The calculated rank correlation coefficient between the dyehouse effluent samples, WWTP influent and effluent samples, and the upstream and downstream samples of Rocky Creek for the sampling periods in October and December 1995 and March, May, and July 1996 indicated significant consistency (p , 0.05) in the rankings for these locations. However, when the calculated rank correlation coefficient included samples from all seven locations during March, May, and July 1996, the rankings were not consistent (p . 0.05). Perhaps collecting 24-h composite samples during May and July 1996, rather than grab samples, would have led to closer rank correlations for these sample periods. The collection of grab samples would not allow sample constituents to be relative over a given time period (e.g., 24 h), resulting in inconsistencies in wastewater constituents and relative toxicities between sample locations. Composite sampling provides a more average sample, which is not affected by temporary increases or decreases in toxic components to which a grab sample may be subjected (Bender 1986). Therefore, rank correlations between sampling periods may have been more comparable using composite sampling each time. When dealing with complex systems over time, such as natural aquatic systems or industrial wastewater systems, tremendous variability is involved. This variability provides strong evidence against the accuracy of the present ‘‘snapshot in time’’ approach to biomonitoring or toxicity testing of effluents incorporated into NPDES permit limitations (US EPA 1993). A high degree of variability in toxicity data between sampling periods for each sample was observed as indicated by CV values ranging from 59.4% (WWTP influent) to 133.7% (dyehouse effluent) for five periods of sampling (Table 3). For

258


Table 4. Nematode percent mortality for individual industrial effluents, WWTP influent, and simulated influent mixtures for May and July 1996a Sample Period May July

M or M P M P

Pb

DH

FG

PP

2.6 — 1.7 —

80.0 — 15.5 —

2.6 — 19.6 —

WWTP Influent 23.1 6.5c 27.6 4.2c 18.4d

Mixture 90:5:5

80:10:10

70:15:15

60:20:20

0:50:50

0:30:70

2.4 6.5 1.6 3.3

2.4 10.3 3.3 4.9

63.2 14.2 1.5 6.4

— — 4.8 8.0

— — 5.2 17.6

— — 27.1 18.4

a DH 5 dyehouse effluent; FG 5 fiberglass effluent; PP 5 pulp processing effluent; mixture proportions are given as DH:FG:PP. —, toxicity value not available b Measured (M) or predicted (P) values. Predicted values calculated as: P 5 S(% mortality) 3 (flow proportion) i i i c Predicted value based on presumed 85:5:10 flow ratio d Predicted value based on presumed 0:30:70 flow ratio

the latter three months of industrial sampling (Table 3), CV values range from 58.6% (dyehouse effluent) to 117.5% (pulp processing effluent). Coefficient of variation values have been used to track the reproducibility of reference toxicants over time as a measure of assay precision and organism health. A recent study (SCTAG QAC 1995) assigned target CV values at 30% for toxicity tests with standard test organisms (e.g., Ceriodaphnia dubia, Pimephales promelas, and Selanastrum capricornutum). However, actual intralaboratory CV values from this study often exceeded 30%. Recently published CV values for reference toxicants in tests using C. elegans range from ,1 to 20% (Cressman and Williams 1997). Note that these CV values are for 24- and 48-h tests with and without food, and these tests are for single chemical compounds in standard media. The exposure media used by Cressman and Williams (1997) are much less complex in composition than the samples used in this study. Furthermore, in recent nematode control chart preparation, CV values for the 48-h tests with a food source can be as high as 40%; this variation was attributed primarily to food density issues (Freeman et al. 1998). Rather than tracking assay precision, CVs calculated for field-collected samples provide insight to the variability of effluent toxicity over time. Variability within the industrial wastewater treatment system may stem from the activities of industries, that is, whether or not they are online, and if so, the effluent flow rates and the kinds of chemicals being discharged by each industry on a given day. Due to the potential for variability in the toxic constituents of wastewaters, toxicity values are also expected to vary. Another source of variation in toxicity data, particularly in the WWTP effluent and downstream samples for this study, is the town wastewater influent into the WWTP (Figure 1), which enters the WWTP independently of the industrial influent and subsequently mixes within the WWTP. Small industrial and domestic wastewaters entering on the town influent line may have contributed toxic constituents to the WWTP effluent. Varying flow rates of town influent may also have contributed to substantial dilutions of industrial influent. However, sampling the town influent and the associated facilities on the town influent line was beyond the scope of this study. Moreover, nonpoint-source pollution, such as stormwater runoff from agricultural or landfill sites, was not considered in this study. Nematode mortality in some of the upstream samples may have stemmed from toxicants leached from the solid-waste landfill located upstream of the WWTP. During

storm events, this runoff may have contributed to toxicity in downstream samples, especially in May 1996 (Figure 2d); however, low upstream flow essentially eliminates the possibility of storm water runoff as a major contributor to toxicity in samples collected downstream of the receiving waters.

Simulated Influent Mixtures The combined industrial influent into the WWTP consistently had higher toxicity than would be expected based on individual industrial effluent toxicities (Figure 2a–e). To probe the cause of this discrepancy, toxicity tests were performed on simulated influent mixtures for two sampling periods, May and July 1996. Mixtures created from samples collected in May 1996 suggested that as the dyehouse proportion decreased, and the proportions from fiberglass manufacturing and pulp processing plants increased, the toxicity of the mixture increased, especially below an 80% contribution of dyehouse effluent (Figure 4a). Simulated influent mixtures for May that most closely resembled a dyehouse–fiberglass–pulp effluent ratio of 85:5:10 (based on presumed flow proportions) yielded toxicities that were 10-fold less than the actual whole influent toxicity (Figure 4a, Table 4). Mixtures containing higher ratios of the fiberglass and pulp effluents yielded toxicities that more closely mimicked the actual influent toxicity (Table 4). Therefore, the actual WWTP influent toxicity appeared to be driven by the percent contribution of fiberglass and pulp processing effluents. Additionally, results suggested that the presumed effluent flow ratio may have underestimated the contributions of the fiberglass and pulp industry inputs and overestimate the dyehouse contribution to the actual WWTP influent. The dyehouse was off-line during July, and therefore did not contribute to the WWTP influent sample collected during that sample period (B. Rucker, OMI, Washington, GA, personal communication). Toxicity test results of simulated influent mixtures from the two remaining industries indicated that as the fiberglass contribution to the mixture decreased, and the pulp processing plant contribution increased, the mortality of C. elegans increased. At a mixture ratio of 0:30:70 for dyehouse, fiberglass, and pulp processing, respectively, mortality of C. elegans (27.1%) appeared to coincide with the nematode mortality in the actual WWTP influent sample (27.6%) (Figure 4b, Table 4). These results appeared to coincide with presumed flow rates, where the fiberglass effluent to pulp processing effluent


259

respectively, implying a synergistic interactive effect between industrial effluents on nematode mortality. In July (Figure 4b, Table 4), when the dyehouse effluent was included in the simulated influent mixtures, toxicity was insignificant based on the 10% benchmark mortality at all ratios (measured and predicted), probably due to dilution of the more toxic constituents. When the simulated mixture included only fiberglass and pulp effluents at a 1:2 ratio, mortality was observed. Again, measured mortality was higher than predicted mortality, but to a lesser extent than observed in May. However, there may be some interactive effects between effluents, because the pulp processing effluent did not cause nematode mortality as high as the simulated influent mixture mortality (0:30:70). In comparing simulated influent mixture results from the samples collected in May and July 1996, variability in industrial effluent flow rates and in toxic constituents present in effluents appeared to lead to variation in lethal effects, while actual WWTP influent lethality was comparable between the two sample periods (Figure 4a and b). In May, the WWTP influent toxicity may have been driven by synergistic interactions between effluents, while in July, toxicity appeared to be more additive. This research further demonstrates the difficulty in assessing the actual toxicity present in complex aquatic and wastewater systems based on periodical toxicity testing.

Conclusions

Fig. 4. Measured mortality data and predicted mortality values for bracketed simulated effluent mixtures based on flow rate proportions in May 1996 (a) and July 1996 (b). The dotted line represents the percent mortality of C. elegans in the actual wastewater treatment influent sample taken during the same period

flow ratio is approximately 1:2 (B. Rucker, OMI, Washington, GA, personal communication). In order to discern nematode mortality differences between individual industrial effluents, actual WWTP influent, and simulated influent mixtures, mortality was predicted using data from individual contributors, assuming that additive toxicity occurred between the industrial effluents. Predicted mortality percentages for simulated influent mixtures were calculated from May and July toxicity data using the following formula: Pi 5

o (% mortality ) 3 (% flow ) i

i

where Pi is the predicted additive percent mortality; % mortalityi is the nematode mortality percentage for each industrial effluent (i); and % flowi is the flow contribution from each industrial effluent (i). This formula provides an indication of an additive lethal contribution from each industry based on proportional flow contribution. Subsequently, predicted mortality percentages were compared to actual measured nematode mortality percentages (Figure 4a and b). In May (Figure 4a), actual mortality data for simulated influent mixtures did not exceed predicted additive mortality (63.2% measured versus 14.2% predicted) until the mixture ratio was 70:15:15 for dyehouse, fiberglass, pulp effluents,

Wastewaters containing toxic constituents that enter a wastewater treatment plant may potentially alter biological processes necessary to treat wastewater, thus impeding the efficiency of the overall treatment process. Our research demonstrates that toxicity trends based on relative toxicity values between locations in an industrial and municipal wastewater treatment system can give an indication of contributors to overall toxicity in WWTP influent and subsequently WWTP effluent. This research was conducted based on the effluent-testing principle that the most effort must be put forth where the largest differences lie (Mount 1986). The unknown representativeness of the samples collected in this study had a tremendous effect on the evaluation of potential toxicity at each location for each sampling period. Factors contributing to the ambiguity of toxic constituents present in the samples include sample collection techniques, varying industrial flow rates and wastewater types, and variability within the natural system itself; these factors contribute to the difficulty in consistently assessing environmental hazard due to a high degree of variability in wastewater toxicity over time. Considering the number and frequency of toxicity tests that were performed in this study, the nematode toxicity test is relatively easy and inexpensive, a beneficial quality in assessing environmental hazards in wastewaters. By using the 72-h C. elegans toxicity test with mortality as an end point, trends were established for periodic toxicity data at locations within an industrial and municipal wastewater system. From these trends, toxicity levels found in simulated wastewater mixtures can give an indication of the periodic industrial contribution to toxicity levels of composite wastewater entering a WWTP based on industrial influent flow rate. Pinpointing areas of toxicity within a system can aid industries and wastewater treatment personnel in taking measures to reduce overall toxicity in effluents, therefore complying with regulations established in NPDES permits. Furthermore,

260

establishing limitations on industrial flow rate based on toxicity data may provide regulators with a better means of preventing potential chemical interactions between the various complex types of wastewaters.

Acknowledgments. The authors acknowledge the State of Georgia’s Consortium for Competiveness in Apparel, Carpet and Textile Industries (CCACTI) and the University of Georgia Agricultural Experiment Station for financial support. The nematode strain used in this study was provided by the Caenorhabditis Genetics Center, which is funded by the NIH National Center for Research Resources (NCRR). The authors would also like to thank Charles P. ‘‘Chip’’ Cressman III, Chris Tatara, and Mark Freeman for lab assistance, and Charles Chafin, Brenda Rucker, John Scott, and Kelly Wilson, as well as Terry Bridges, David Chafin and Julie Chastain, for their help in sample collection.

References Anderson DR, Sweeney DJ, Williams TA (1994) Introduction to statistics: Concepts and applications, 3rd. ed. West Publishing Co., Minneapolis, Minnesota Atlas RM, Bartha R (1987) Microbial ecology: Fundamentals and applications, 2nd ed. Benjamin/Cummings Publishing Company, Menlo Park, California Bender ES (1986) Effluent sampling for biological effects testing. In: Bergman HL, Kimerle RA, Maki AW (eds) Environmental hazard assessment of effluents, SETAC special publication series. Pergamon Press, New York, pp 81–91 Biological Monitoring, Inc. (1994) Summary of toxicity reduction evaluation for Rocky Creek Wastewater treatment plant, City of Washington, GA: Toxicity and chemistry results. Blacksburg, Virginia Brenner S (1974) The genetics of Caenorhabditis elegans. Genetics 77:71–94 Cassada RC, Russell RL (1975) The dauerlarva, a post-embryonic developmental variant of the nematode Caenorhabditis elegans. Dev Biol 46:326–342 Cox G, Edgar R (1981) Cuticle of Caenorhabditis elegans: Its isolation and partial characterization. J Cell Biol 90:7–17 Cressman CP, Williams PL (1997) Reference toxicants for aquatic toxicity testing using Caenorhabditis elegans. In: Dwyer FJ, Doane TR, Hinman ML (eds) Environmental toxicology and risk assessment: Modeling and risk assessment, 6th vol, ASTM STP 1317. American Society for Testing and Materials, Philadelphia, pp 518–532 Donkin SG, Williams PL (1995) Influence of developmental stage, salts and food presence on various end points using Caenorhabditis elegans for aquatic toxicity testing. Environ Toxicol Chem 14:2139–2147 Donkin SG, Eiteman MA, Williams PL (1995) Toxicity of glucosinolates and their enzymatic decomposition products to Caenorhabditis elegans. J Nematol 27:258–262 Freckman DW (1988) Bacterivorous nematodes and organic-matter decomposition. Agric Ecosyst Environ 24:195–217


Freeman MN, Marse TJ, Williams PL (1998) Establishment of quality assurance procedures for aquatic toxicity testing with the nematode Caenorhabditis elegans. In: Little EE, DeLonay AJ, Greenberg BM (eds) Environmental toxicology and risk assessment, 7th vol, ASTM STP 1333. American Society for Testing and Materials, Philadelphia, in press Gad S, Weil CS (1988) Statistics and experimental design, 2nd. ed. The Telford Press, Caldwell, New Jersey, pp 43–69 GA EPD (Georgia Environmental Protection Division) (1995) Biomonitoring test results for the City of Washington: Rocky Creek WPCP (water pollution control plant). NPDES permit no. GA0031101. Georgia Department of Natural Resources (DNR), Atlanta, Georgia Khanna N, Cressman CP, Tatara CP, Williams PL (1997) Tolerance of Caenorhabditis elegans to pH, salinity, and hardness in aquatic media. Arch Environ Contam Toxicol 32:110–114 Mount DI (1986) Principles and concepts of effluent testing. In: Bergman HL, Kimerle RA, Maki AW (eds) Environmental hazard assessment of effluents, SETAC special publication series. Pergamon Press, New York, pp 61–65 Popham JD, Webster JM (1982) Ultrastructural changes in Caenorhabditis elegans (Nematoda) caused by toxic levels of mercury and silver. Ecotoxicol Environ Saf 6:183–189 SCTAG QAC (Southern California Toxicity Assessment Group Quality Assurance Committee) (1995) Evaluation of submitted reference toxicant data and preliminary recommendations of toxicant and concentration series for the EPA freshwater chronic protocols (draft). A reference toxicant standardization and use in toxicity testing first report Spence JT, Underwood BJ, Duncan CP, Cotton JW (1968) Elementary statistics, 2nd ed. Appleton-Century-Crofts, New York, pp 218– 220 Traunspurger W, Haitzer M, Ho¨ss S, Beier S, Ahlf W, Steinberg C (1997) Ecotoxicological assessment of aquatic sediments with Caenorhabditis elegans (Nematoda): A method for testing liquid medium and whole-sediment samples. Environ Toxicol Chem 16:245–250 USEPA (US Environmental Protection Agency) (1991) Methods for aquatic toxicity identification evaluations: Phase I toxicity characterization procedures, 2nd ed. Office of Research and Development, Cincinnati, Ohio. EPA/600/3-91/034 USEPA (US Environmental Protection Agency) (1993) Methods for measuring the acute toxicity of effluents and receiving waters to freshwater and marine organisms, 4th ed. Office of Research and Development, Cincinnati, Ohio. EPA/600/4-90/027F USGAO (US General Accounting Office) (1989) Water pollution: Improving monitoring and enforcement needed for toxic pollutants entering sewers. Washington, DC. GAO/RCED-89-101 Williams PL, Dusenbery DB (1988) Using the nematode Caenorhabditis elegans to predict mammalian acute lethality to metallic salts. Toxicol Ind Health 4:469–478 Williams PL, Dusenbery DB (1990) Aquatic toxicity testing using the nematode Caenorhabditis elegans. Environ Toxicol Chem 9:1285– 1290