INFLUENCE OF WATER QUALITY AND AGE ON NICKEL TOXICITY TO FATHEAD MINNOWS (PIMEPHALES PROMELAS)

Transcription

1 Environmental Toxicology and Chemistry, Vol. 23, No. 1, pp , SETAC Printed in the USA /04 $ INFLUENCE OF WATER QUALITY AND AGE ON NICKEL TOXICITY TO FATHEAD MINNOWS (PIMEPHALES PROMELAS) THAM CHUNG HOANG, JOSEPH R. TOMASSO, and STEPHEN J. KLAINE* Clemson University, Department of Environmental Toxicology, Pendleton, South Carolina 29670, USA Clemson University, Department of Aquaculture, Fisheries, Wildlife, Clemson, South Carolina 29631, USA (Received 7 January 2003; Accepted 13 May 2003) Abstract This research characterized the effects of water quality and organism age on the toxicity of nickel (Ni) to fathead minnows (Pimephales promelas) to facilitate the accurate development of site-specific water-quality criteria. Nickel sulfate hexahydrate (NiSO 4 6H 2 O) was used as the Ni source for performing acute toxicity tests (median lethal concentration after 96-h exposure [96-h LC50]) with -d-old and 28-d-old P. promelas under varying regimes of hardness, ph, alkalinity, and natural organic matter (NOM). The toxicity of Ni was inversely related to water hardness between hardness values of 20 and 150 mg/l (as CaCO 3 ). Below 30 mg/l alkalinity, Ni toxicity was related to alkalinity. The effect of ph was confounded by hardness and the presence of NOM. In the absence of NOM, the toxicity of Ni increased as ph increased at high hardness and alkalinity. In general, 28-d-old fish were less sensitive than -d-old fish to Ni. This lower sensitivity ranged from 12-fold at low hardness and alkalinity (20 and 4 mg/l, respectively) to 5-fold at high hardness and alkalinity (100 and 400 mg/l, respectively). The presence of NOM (10 mg/ L as dissolved organic carbon [DOC]) reduced Ni toxicity by up to 50%, but this effect appeared to be saturated above DOC at 5 mg/l. Incubating Ni with the NOM solution from 1 to 17 days had no effect on Ni toxicity. When using multivariate analysis, the 96-h LC50 for Ni was a function of fish age, alkalinity, hardness, and NOM (96-h LC (fish age) 0.005(alkalinity) 0.018(hardness) 0.138(DOC)). When using this model, we found a strong relationship between measured and predicted 96-h LC50 values (r ) throughout the treatment water qualities. The biotic ligand model (BLM) did not accurately predict Ni toxicity at high or low levels of alkalinity. Results of our research suggest that the BLM could be improved by considering NiCO 3 to be bioavailable. Keywords Nickel Toxicity Fathead minnow Natural organic matter INTRODUCTION Nickel (Ni) has many industrial uses including stainless steel production, electroplating, battery manufacturing, and coin production. Although these applications have provided many benefits, they have resulted in the redistribution of a significant amount of Ni into the environment. According to Nriagu [1], Ni emission to the atmosphere from anthropogenic sources was about kg, kg, and kg between 1951 and 1960, 1961 and 1970, and 1970 and 1980, respectively. Ultimately, this Ni falls to the land surface in rainwater and particles. The toxicity of Ni has been studied in rats and mice [2 4]. Although the toxicity of Ni to aquatic organisms has been examined [5 8], a systematic study on the influence of water quality on Ni toxicity has not been conducted. Recent studies suggest that Ni may be more toxic to larval fathead minnows as ph increases [5]. Data on the influence of alkalinity on the toxicity of Ni are lacking in the literature. Erickson et al. [9] reported that the dissolved copper 96-h median lethal concentration (LC50) for larval fathead minnows decreased with increasing alkalinity. Alkalinity may be important for Ni because the dominant species is NiCO 3 at ph 8 [10]. Many studies report that natural organic matter (NOM) influences bioavailability of metals such as copper and silver [9,11]. Again, no data exist regarding the effect of NOM on Ni toxicity. A review of the literature suggests that organism age influences metal toxicity, but systematic studies on the influence of organism * To whom correspondence may be addressed (sklaine@clemson.edu). age on Ni toxicity to fathead minnows are lacking [5 7,12,13]. Finally, no research has been reported on the interactive effects of hardness, ph, alkalinity, NOM, and organism age on Ni toxicity. The objective of this research was to determine the effects of hardness, ph, alkalinity, NOM, and fish age individually and in combination on Ni toxicity to fathead minnows. MATERIALS AND METHODS Fish culture The fathead minnow culture was maintained in moderately hard water at the Institute for Environmental Toxicology, Clemson University (Pendleton, SC, USA) according to U.S. Environmental Protection Agency (U.S. EPA) methods [14]. Moderately hard water was prepared by adding reagent-grade salts to deionized water (Super-Q Millipore, Bedford, MA, USA). The culture room was maintained at 25 1 C with a photoperiod of 16:8 h light:dark. Fish were fed once daily with frozen brine shrimp (Artemia salina) or Tetramin food (TetraWerke, Melke, Germany). Fathead minnow eggs were incubated in treatment waters without Ni (described below) for hatching. Twenty-eight-day-old fathead minnows were obtained by raising larval fathead minnows in test water in the culture room. They were fed twice daily with newly hatched brine shrimp. Reference toxicity tests for -d-old fathead minnows were conducted monthly with copper sulfate hexahydrate. Toxicity tests The 96-h static-renewal tests were conducted according to U.S. EPA methods [14]. All test waters were prepared and 86

2 Factors influencing the toxicity of nickel to fathead minnows Environ. Toxicol. Chem. 23, aerated for at least 24 h before initiating a test. Tests were conducted in 600-ml graduated polyproplene beakers, each containing 250 ml of test medium. The beakers were randomly positioned at a temperature of 25 1 C, with a photoperiod of 16:8 h light:dark. Ten organisms were transferred to each beaker by means of a smooth glass tube (8 mm in diameter). Surviving fish in each test chamber were fed with 0.2 ml of newly hatched brine shrimp 2 h before water renewal at 48 h into the test. Dead fish were removed daily. Mortality was recorded daily until test termination. A preliminary set of experiments was conducted with - d-old fathead minnows by using Ni sulfate hexahydrate (NiSO 4 6H 2 O) and Ni chloride hexahydrate (NiCl 2 6H 2 O) to determine the influence of Ni salt on toxicity. Hardness, ph, and alkalinity of test waters were controlled by adjusting the concentrations of CaSO 4 2H 2 O, MgSO 4, NaHCO 3, KCl, and HNO 3. Natural organic matter was not added to test medium for the hardness, alkalinity, and ph experiments. A preliminary set of experiments was performed to determine the effect of Ni dissolvedorganic carbon (DOC) equilibration time on Ni toxicity to -d-old fathead minnows. Results indicated no difference in 96-h LC50 values for equilibration time up to 17 d before introducing fish. Hence, all solutions were equilibrated for 24 h before adding the fish. Experimental design All toxicity tests were conducted with five Ni levels and a negative control. Three replicates were used for each treatment level. The effect of hardness (mg/l, as CaCO 3 ) on the toxicity of Ni to -d-old fathead minnows was investigated by conducting 96-h static-renewal toxicity tests at hardness levels of 20, 50, 100, and 150 mg/l, at constant ph 7.0, and at constant alkalinity of 20 mg/l, as CaCO 3. To assess the effect of ph, two sets of 96-h static-renewal toxicity tests with -d-old fathead minnows were conducted under the same hardness and alkalinity and at different ph levels (6 9). One set was conducted with hardness and alkalinity of 20 mg/l each; the other was conducted with hardness and alkalinity of 100 mg/l each. To characterize the effect of alkalinity on Ni toxicity to -d-old fathead minnows, we used two sets of 96-h staticrenewal toxicity tests with ph and hardness levels constant while alkalinity was varied from 3 to 350 mg/l. One set was conducted with hardness of 20 mg/l and the other with hardness of 100 mg/l. To determine the effect of NOM on Ni toxicity to -dold fathead minnows, we conducted three sets of 96-h staticrenewal toxicity tests with different DOC concentrations (0, 2, 5, and 10 mg/l) under the same hardness, alkalinity, and ph. The first set was conducted with hardness of 20 mg/l, alkalinity of 4 mg/l, and ph 6. The second set was conducted with hardness of 50 mg/l, alkalinity of 20 mg/l, and ph 7. The third set was conducted with hardness of 100 mg/l, alkalinity of 400 mg/l, and ph 9. The organic matter used for this research was isolated from the Black River (Andrews, SC, USA) by a reverse osmosis membrane system with a membrane size of in (S4040, Osmonics, Minnetonka, MN, USA) and spiral-wound cross-flow module. To examine the effects of fish age on Ni toxicity, three sets of 96-h static-renewal toxicity tests with the same water quality (hardness, alkalinity, ph, and DOC) as the three sets described above were conducted with 28-d-old fathead minnows. To validate the interactive-effect model developed from the laboratory study, two tests were conducted with low water hardness, alkalinity, and ph with -d-old fathead minnows. One test was conducted with water collected from George Creek (Greenville, SC, USA); the other used laboratory water and matched water quality with the site water. Water chemistry Water hardness, ph, alkalinity, and dissolved oxygen were measured at the start and the end of every test in each treatment. Water samples were analyzed for total Ni, dissolved Ni, major anions, major cations, and DOC at the start and the end of each test. All water samples except those for DOC were placed in 15-ml graduated polyproplene tubes and acidified with concentrated HNO 3 for analyses of total and dissolved Ni. Dissolved Ni samples and DOC samples were filtered through m Gelman Nylon Mesh (Fisher Scientific, Fairlawn, NJ, USA) before acidification. The DOC samples were stored in 30-ml glass bottles. The ph was measured with a Thermo Orion model 525 meter (Orion Research, Beverly, MA, USA). Dissolved oxygen was measured with a YSI model 85 meter and probe (YSI, Yellow Springs, OH, USA). The concentration of dissolved oxygen in all test waters was approximately 8.0 mg/l. Hardness and alkalinity were determined by titrating with 0.01 M ethylenediaminetetraacetic acid and 0.02 N H 2 SO 4, respectively. Total and dissolved Ni were analyzed with a Perkin-Elmer atomic absorption spectrometer (model 800, Perkin-Elmer Instruments, Norwalk, CT, USA) equipped with an autosampler. Major anions and cations were analyzed by a Dionex IC model DX-500 (Dionex, Atlanta, GA, USA) with an autosampler and by inductively coupled plasma spectroscopy with a Thermo Jarrell-Ash model 61 E (Thermo Electron, Waltham, MA, USA). Dissolved organic carbon was measured with a Shimazu TOC-5000 equipped with a Shimadzu autosampler (Shimadzu Scientific Instruments, Columbia, MD, USA). The concentration of DOC in Milli-Q (Millipore) water was 0.5 to 0.7 mg C/L. Statistics Probit or Spearman Karber methods were used to calculate 96-h LC50 values from the mortality data. Regression (linear, nonlinear, and stepwise) analyses for determining the individual and multiple relationships between 96-h LC50 and hardness, ph, alkalinity, DOC, and fish age were performed by Microsoft Excel 98 (Redmond, WA, USA) and SPSS, Ver 10.1 (SPSS, Chicago, IL, USA), respectively. Analysis of covariance (ANCOVA) was conducted on life stage (organism age) results controlling for DOC on effects of hardness, alkalinity, ph, and age (SAS, Cary, NC, USA). Data met assumption of normality, independence, and homogeneity of variance, and did not require transformation for analysis of slope and intercept. A p 0.05 was considered significant. The biotic ligand model (BLM) [15] was used to predict 96-h LC50 values from measured water chemistry of the tests. RESULTS AND DISCUSSION Water quality and organism age affected the toxicity of Ni to fathead minnows (Table 1). The 96-h total Ni LC50 values and water chemistry measurements are shown in Table 1 (96- h total Ni and dissolved Ni LC50 values were similar; data not shown for dissolved Ni). At low hardness (21 mg/l), the 96-h LC50 values for Ni sulfate and Ni chloride were 0.36 and 0.40 mg Ni/L, respec-

3 88 Environ. Toxicol. Chem. 23, 2004 T.C. Hoang et al. Table 1. Water chemistry and 96-h median lethal concentration (LC50) values for Pimephales promelas exposed to Ni sulfate hexahydrate (all tests except tests 2 and 4; in which NiCl 2 6H 2 O was used). Water-quality factor are presented as mean standard deviation. Median lethal concentrations were calculated by using the Probit method or the Spearman Karber method Test 96-h LC50 (mg Ni/L) LC50 (95% CI) a Hardness (mg/l as CaCO 3 ) ph Alkalinity (mg/l as CaCO 3 ) DOC a (mg/l) Fish age (d) b 1.57 b c 2.93 c c c c a CI confidence interval; DOC dissolved organic carbon. b NiCl 2 6H 2 O. c Generated by the Spearman Karber method. tively. At high hardness (52 mg/l), the 96-h LC50 values for Ni sulfate and Ni chloride were 1.68 and 1.57 mg Ni/L, respectively. These results indicated that the chemical form of Ni did not significantly affect Ni toxicity (Table 2, text reference I). Because of chloride involvement in fish metabolism, Ni sulfate was used for toxicity testing. Monthly reference toxicity tests showed similar 96-h LC50 values ( g Cu/L) for -d-old fathead minnows. Reference toxicity tests were not conducted for the 28-d-old fathead minnows. However, the 96-h LC50 values for two subsets (Table 1, tests 39 42, 43 46) that were conducted with similar hardness, ph, alkalinity, and corresponding DOC con-

4 Factors influencing the toxicity of nickel to fathead minnows Environ. Toxicol. Chem. 23, Table 2. Models and statistics associated with Ni toxicity tests to fathead minnows (Pimephales promelas). Hardness and alkalinity values are expressed in mg/l as CaCO 3. Salts compared were Ni chloride and Ni sulfate. The p primary and p secondary are the probabilities that the slope of the model differed from 0 (except for Ni salt experiments) and the probability that the secondary treatments had an effect, respectively. Unless otherwise noted, organisms were d old at the start of the tests Text a reference Primary treatment Secondary treatment Test numbers (from Table 1) Model r 2 model p primary p secondary I Nickel salts Hardness 21 mg/l Hardness 52 mg/l 1, 2 3, b b II Hardness Alkalinity 20 mg/l 1, 3, 5, 6 y x III ph Hardness and alkalinity (20 mg/l, each) Hardness and alkalinity (100 mg/l, each) IV Alkalinity Hardness 20 mg/l Hardness 100 mg/l 1, 2, , 7 13, , y 0.031x y 0.649x y lnx y 0.002x for slope, for intercept for slope, for intercept V DOC c and fish age -d-old fish (low-ion water) de 28-d-old fish (low-ion water) de -d-old fish (medium-ion water) ef 28-d-old-fish (medium-ion water) ef -d-old fish (high-ion water) eg 28-d-old-fish (high-ion water) eg 17, y 0.008x y lnx y 0.024x y lnx y 0.127x y lnx for DOC for age for ions 1 for DOC and age for ion and age NS for DOC and ion NS for DOC, ion, and age a Information specifically referenced in the text. b Probability that one type of Ni salt has a treatment effect. c DOC dissolved organic carbon. d Hardness 20 mg/l as CaCO 3, alkalinity 4 mg/l as CaCO 3,pH 6.0. e Model is based on treatments with DOC 6.0 mg/l. f Hardness 50 mg/l as CaCO 3, alkalinity 20 mg/l as CaCO 3,pH 7.4. g Hardness 100 mg/l as CaCO 3, alkalinity 380 mg/l as CaCO 3,pH 8.7.

5 90 Environ. Toxicol. Chem. 23, 2004 T.C. Hoang et al. Fig. 1. Toxicity values as a function of NiCO 3 (upper panel) and total alkalinity (lower panel). 96-h LC50 median lethal concentration after 96-h exposure. centrations but different time (seven months) were similar (p 0.393). This indicated that the response of fish did not change significantly over time. At similar ph (7.24) and alkalinity levels (20 mg/l), the 96-h LC50 for Ni was directly related to water hardness, between 21 and 150 mg/l (Table 2, text reference II). Toxicity to -d-old fish decreased 10-fold when hardness was increased 7-fold. This result was in agreement with the results of Meyer et al. [7]. Results of this study were 10- to 27-fold lower than 96-h LC50 values reported by the U.S. EPA [16] ( mg Ni/L at hardness of 20 mg/l). However, our results were similar to the results of Pyle et al. [17] (0.45 mg Ni/L). Water quality and fish age used by Pyle et al. [17] were similar to those in our experiments. Results reported by the U.S. EPA [16] were obtained in similar water quality; however, the U.S. EPA did not report the age of the fish used in their experiments. Pyle et al. [17] also found that the 96-h LC50 was 0.50 and 2.27 mg Ni/L at hardness of 40 and 140 mg/l, respectively. However, based on the 96-h LC50 hardness relationship that we observed, we would have predicted the 96-h LC50 to be about 0.93 and 3.24 mg Ni/L at similar hardness levels, respectively. The 96-h LC50 value for Ni found in this research was 3.5 mg Ni/L at hardness of 150 mg/l (Table 1, test 6). A similar value was reported by Schubauer-Berigan et al. [5] (3.4 mg Ni/L) at higher hardness (300 mg/l). Examination of our data suggests a 96-h LC50 value closer to 7.0 mg Ni/L for 300 mg/l hardness. One reason for this discrepancy may have been the higher alkalinity used by Schubauer-Berigan et al. [5] (250 mg/l). Alkalinity in our hardness set was maintained at 20 mg/l. Toxicity of Ni to freshwater fish is influenced not only by hardness, but also by DOC, ph, alkalinity, and fish age. These parameters were not reported with the 96-h LC50 values of 4.6 to 9.8 mg Ni/L by the U.S. EPA [16]. At low levels of hardness and alkalinity (20 mg/l each), ph between 6.5 and 8.2 did not affect Ni toxicity to -d-old fish (Table 2, text reference III). However, increasing ph increased 96-h LC50 values at hardness and alkalinity of 100 mg/l (Table 2, text reference III). These results agreed with the results of Schubauer-Berigan et al. [5]. In contrast, the Windermere humic aqueous model (WHAM) predicted that Ni would be less bioavailable as ph increased [10]. Pyle et al. [17] found that at a hardness of 50 mg/l and an alkalinity of 32 mg/l, Ni toxicity increased slightly between ph 5.5 and 7.0 (96-h LC50 values of 0.69 and 0.54 mg Ni/L, respectively) and decreased between ph 7.0 and 8.5 (96-h LC50 values of 0.54 and 2.21 mg Ni/L, respectively). However, results from our study at higher and lower levels of hardness and alkalinity did not support this observation. At ph 6, the higher alkalinity level used by Pyle et al. [17] (32 mg/l) would have contributed to the increased toxicity (discussed below), which would have resulted in the lower reported 96-h LC50 values found by Pyle et al. [17] ( mg Ni/L). Results of ph experiments suggested a protective effect of hardness (Table 2, text reference III). At a hardness of 20 mg/l, the 96-h LC50 decreased from 0.81 to 0.32 mg Ni/L as alkalinity increased from 3 to 27 mg/ L (Table 2, text reference IV). This relationship is particularly important for waters in the southeastern United States, many of which have alkalinity values between 20 and 40 mg/l and hardness 20 mg/l. At a hardness of 100 mg/l, we found no effect of alkalinity on Ni toxicity to -d-old fish (Table 2, text reference IV). The Visual MINTEQ model (the developed model of MINTEQA2 and also WHAM) was used to predict Ni speciation from the water chemistry of each toxicity test ( amov.ce.kth.se/people/gustafjp/vminteq.htm). At low hardness levels, the shape of the plot of the 96-h LC50 as a function of NiCO 3 (Fig. 1) was similar to the plot of the 96-h LC50 as a function of alkalinity (Fig. 1), suggesting that NiCO 3 may be bioavailable. Results of alkalinity experiments indicated a protective effect of hardness to larval fathead minnows (Table 2, text reference IV). Dissolved organic carbon significantly affected the toxicity of Ni to -d-old fish exposed to Ni in water with high alkalinity, hardness, and ph, and Ni toxicity to 28-d-old fish in all three environments tested (Table 2, text reference V). The 96-h LC50 value was about 4-fold and 12-fold greater for 28- d-old fish at low and high DOC in all three environments tested, respectively (Table 1, tests 17, 25 54). When the six groups were considered in the multiway ANCOVA, DOC, ions, and age each significantly affected toxicity (Table 2, text reference V). Two significant s also were observed (DOC and fish age, and ion and fish age; Table 2, text reference V). The effects of NOM and ions (particularly hardness) on the toxicity of metals have been well documented elsewhere. The age-dependent tolerance of fish to Ni may be due to differences in metabolic rate between larval and juvenile members of the same species. Larvae, with a much higher metabolic rate, generally would be expected to be more sensitive to stressors, especially because respiratory toxicity has been suggested as the mechanism of Ni toxicity [18].

6 Factors influencing the toxicity of nickel to fathead minnows Environ. Toxicol. Chem. 23, Table 3. Coefficients for regression analysis (r ). Input variables were hardness, alkalinity, ph, dissolved organic carbon (DOC), and fish age (data from Table 1, except tests 55 and 56) Coefficients a Unstandardized Standardized Model B SE t p Intercept Fish age Alkalinity Hardness DOC a The dependent variable was the 96-h median lethal concentration. A more important question related to the larval fish is why DOC did not affect Ni toxicity in two of the three environments tested. No explanation is obvious. Explanations based on differences between larvae and juveniles in metabolic rate, gill surface area, and mechanisms and points of respiration do not account for the differential effect of NOM on Ni toxicity in water of varying hardness and alkalinity. Median lethal concentrations (tests 1 54) reported in Table 1 were used in stepwise regression analysis to determine interactive effects of hardness, ph, alkalinity, DOC, and fish age on Ni toxicity. The 96-h LC50 was a function of fish age, alkalinity, hardness, and DOC. The model was 96-h LC (fish age) 0.005(alkalinity) 0.018(hardness) 0.138(DOC) and explained 94% of the variation (Table 3). The units for 96-h LC50 were mg Ni/L, fish age was days, alkalinity and hardness were mg/l as CaCO 3, and DOC was mg C/L. This model was used to predict 96-h LC50 values from fish age, alkalinity, hardness, and DOC (Fig. 2). We found a strong relationship between the measured and predicted 96- h LC50 values (r ). Fig. 2. Correlation between measured and regression-predicted 96-h LC50 values for fathead minnows. The two bounding dashed lines represent 95% confidence interval (96-h LC (fish age) 0.005(alkalinity) 0.018(hardness) 0.138(DOC)). 96-h LC50 median lethal concentration after 96-h exposure; DOC dissolved organic carbon. Units for fish age days. Fig. 3. Predicted versus measured 96-h LC50 values for fathead minnows (FM) by the biotic ligand model (BLM). The two dotted lines represent a range of two times on either side of the prediction. 96-h LC50 median lethal concentration after 96-h exposure. The BLM also was used to predict Ni toxicity from the water quality of test 1 to 54 in Table 1 (Fig. 3). The predictions were 28% and 15% above and below the acceptable range of uncertainty (defined as twofold by BLM developers), respectively. Water of high and low alkalinity accounted for these cases, respectively. The BLM predicts Ni toxicity based on Ni bioavailability, using the assumption that only ionic Ni (Ni 2 ) is toxic to fish. The WHAM was used to predict Ni speciation in the BLM. Based on the WHAM, ionic Ni significantly decreased and NiCO 3 appears to be the dominant species at ph 8 [10]. Therefore, the BLM predicts that increasing alkalinity decreases Ni toxicity. However, our research showed that increasing alkalinity at alkalinity 30 mg/l increased Ni toxicity (Table 2, text reference IV). This result might explain why the 96-h LC50 values predicted by the BLM were greater and less than the measured values for the high- and low-alkalinity cases, respectively. Two toxicity tests (Table 1, tests 55, 56) were conducted with -d-old fathead minnows to validate the regression model. Test 55 was conducted with site water (George Creek), and test 56 was conducted with laboratory water adjusted to match site water with respect to hardness, ph, and alkalinity. The predictions by the regression model for test 55 and 56 (96-h LC and mg Ni/L, respectively) were 6% and 64% less than the measured values (measured 96-h LC and mg Ni/L, respectively). In other words, the regression model overpredicted the toxicity. However, the BLM underpredicted the toxicity response. The BLM predicted 67% and 28% greater than the measured values for test 55 and 56, respectively (BLM predicted 96-h LC and mg Ni/L, respectively). The regression model included DOC as a factor. However, test 56 was conducted in laboratory water with no added NOM. This would explain why the regression model did not predict Ni toxicity well for test 56. The effect of alkalinity at low alkalinity levels (Ni toxicity increased when alkalinity increased) would explain the higher 96-h LC50 values predicted by the BLM, compared to the measured values. We conclude that Ni toxicity to fathead minnows was influenced by fish age, NOM, and water-quality characteristics such as hardness, alkalinity, and ph. Fish age was the strongest factor influencing Ni toxicity; young fish were much more sensitive than older fish. These results can be used to refine

7 92 Environ. Toxicol. Chem. 23, 2004 T.C. Hoang et al. predictive models, such as the BLM, to more accurately predict site-specific water-quality criteria. However, given the common practice of using larval fathead minnow toxicity tests to characterize metal toxicity, attention must be given to the difference reported here between larval and juvenile responses to Ni in the presence of NOM. Acknowledgement We thank Adam Ryan, Eric Van Genderen, K. Benjamin Wu, John Smink, Milton Taylor, Shawn Young, and J. Jeff Isely for their help. We acknowledge the Nickel Producers Environmental Research Association for funding this research. REFERENCES 1. Nriagu JO Global cycle and properties of nickel. In Nriagu JO, ed, Nickel in the Environment. John Wiley, New York, NY, USA, pp Stoner GD, Shimkin MB, Troxell MC, Thompson TL, Terry LS Tests for carcinogenicity of metallic compounds by the pulmonary tumor response in strain A mice. Cancer Res 36: Kasprzak KS, Diwan BA, Konishi N, Misra M, Rice JM Initiation by nickel acetate and promotion by sodium barbital of renal cortical epithelial tumors in male F344 rats. Carcinogenesis 11: Diwan BA, Kasprzak KS, Rice JM Transplacental carcinogenic effects of nickel (II) acetate in the renal cortex, renal pelvis and adenohypophysis in F344/NCr rats. Carcinogenesis 13: Schubauer-Berigan MK, Dierkes JR, Monson PD, Ankley GT ph-dependent toxicity of Cd, Cu, Ni, Pb and Zn to Ceriodaphnia dubia, Pimephales promelas, Hyalella azteca, and Lumbriculus variegates. Environ Toxicol Chem 12: Segner HD, Lenz WH, Schüürmann G Cytoxicity of metals toward rainbow trout R1 cell line. Environ Toxicol Water Qual 9: Meyer JS, Santore RC, Bobbitt JP, DeBrey LD, Boese CJ, Paquin PR, Allen HE, Bergman HL, Di Toro DM Binding of nickel and copper to fish gills predicts toxicity when water hardness varies, but free ion activity does not. Environ Sci Technol 33: Kszos LA, Stewart AJ, Taylor PA An evaluation of nickel toxicity to Ceriodaphnia dubia and Daphnia magna in a contaminated stream and in laboratory tests. Environ Toxicol Chem 11: Erickson JR, Duane AB, Vincent RM, Henry PN, Edward NL The effects of water chemistry on the toxicity of copper to fathead minnows. Environ Toxicol Chem 15: Water Environment Research Foundation Phase I development of a biotic ligand model for nickel. HydroQual Project WERF0040. HydroQual, Mahwah, NJ, USA. 11. Karen JD, Ownby DR, Forsythe L, Bills TP, La Point TW, Cobb GB, Klaine SJ Influence of water quality on silver toxicity to rainbow trout (Oncorhynchus mykiss), fathead minnows (Pimephales promelas), and water fleas (Daphnia magna). Environ Toxicol Chem 18: Pickering QH Chronic toxicity of nickel to fathead minnow. Water Pollut Control Fed 46: Eisler R Nickel hazards to fish, wildlife, and invertebrates: A synoptic review. USGS/BRD/BSR U.S. Department of the Interior, U.S. Geological Survey, Patuxent Wildlife Research Center, Laurel, MD. 14. Weber CI Method for measuring the acute toxicity of effluents and receiving waters to freshwater and marine organisms. EPA 600-/4-90/027. U.S. Environmental Protection Agency, Cincinnati, OH. 15. Di Toro DM, Allen HE, Bergman HL, Meyer JS, Santore RC, Paquin P The Biotic Ligand Model; A Computational Approach for Assessing the Ecological Effects of Copper and Other Metals in Aquatic Systems. International Copper, New York, NY, USA. 16. U.S. Environmental Protection Agency Ambient water quality criteria for nickel. EPA 440/ Cincinnati, OH. 17. Pyle GG, Swanson SM, Lehmkuhl DM The influence of water hardness, ph, and suspended solids on nickel toxicity to larval fathead minnows (Pimephales promelas). Water Air Soil Pollut 133: Pane EF, Richards JG, Wood CM Acute waterborne nickel toxicity in rainbow trout (Oncorhynchus mykiss) occurs by a respiratory rather than ionoregulatory mechanism. Aquat Toxicol 63:65 82.