Environmental Toxicology

Transcription

1 Environmental Toxicology and Chemistry, Vol. 22, No. 4, pp , SETAC Printed in the USA /03 $ Environmental Toxicology THE RELATIVE SENSITIVITY OF FOUR BENTHIC INVERTEBRATES TO METALS IN SPIKED-SEDIMENT EXPOSURES AND APPLICATION TO CONTAMINATED FIELD SEDIMENT DANIELLE MILANI,* TREFOR B. REYNOLDSON, UWE BORGMANN, and JUREK KOLASA National Water Research Institute, Burlington, Ontario, L7R 4A6 Canada McMaster University, Hamilton, Ontario, L8S 4K1 Canada (Received 6 February 2002; Accepted 26 August 2002) AbstractThe relative sensitivity of four benthic invertebrates (Hyalella azteca, Chironomus riparius, Hexagenia spp., and Tubifex tubifex) was determined for Cd, Cu, and Ni in water-only and in spiked-sediment exposures. Survival (median lethal concentrations [LC50s] and the concentrations estimated to be lethal to 25% of test organisms [LC25s]), and endpoints for growth and reproduction (mean inhibitory concentrations [IC25s]) were compared. The sensitivities differed depending on the species and metal, although some trends emerged. In water-only exposures, H. azteca is the most sensitive species to cadmium and nickel, with mean LC50s of and 3.6 mg/l, respectively; C. riparius is the most sensitive species to copper, with a mean LC50 of mg/l. In the spiked-sediment exposures, the order in decreasing sensitivity to copper is Hyalella Hexagenia Chironomus Tubifex for survival and growth/reproduction. For cadmium, the order in decreasing sensitivity is Hyalella Chironomus Hexagenia Tubifex, and for nickel is Hyalella K Hexagenia Chironomus Tubifex. Chironomus riparius and Hexagenia spp. survival can be used to distinguish between toxicity caused by different metals. Species test responses in field-collected sediment (Collingwood Harbour, ON, Canada) were examined in an attempt to determine the causative agent of toxicity throughout, using the established species sensitivities. Sediment toxicity was categorized first by comparing species responses to those established for a reference database. Test responses in the field-collected sediment do not support causality by Cu, a suspected toxicant based on comparison of sediment chemistry with sediment quality guidelines. KeywordsBenthic invertebrate Toxicity tests Relative sensitivity Metals INTRODUCTION Sediment toxicity tests using benthic invertebrates are an important line of evidence in the assessment of sediment quality. However, no one species is adequate for detection of potentially adverse effects of mixtures of contaminants [1]. Differences in species behavior, lifestyle, and physiology can contribute to different sensitivities to contaminants among species, and these differences can be contaminant specific [2 4]. The use of several species representing different sediment habitats, as well as different physiological endpoints, is therefore recommended for sediment toxicity evaluation [5,6]. Previous studies conducted at the National Water Research Institute ([NWRI], Burlington, ON, Canada) used four species of benthic invertebrates for the evaluation of sediment toxicity: The amphipod, Hyalella azteca; the midge, Chironomus riparius; the mayfly, Hexagenia spp.; and the oligochaete worm, Tubifex tubifex. These four species have been used in conjunction to evaluate toxicity of single chemicals and mixtures of chemicals in contaminated sediments [5,7 10]. Additionally, following the methods derived from the sediment quality triad [11], the relationship between the functional responses (survival, growth, and reproduction) of these four species in sediment toxicity tests, benthic community structure data, and key environmental variables have been used to develop numeric guidelines for nearshore sites of the Great Lakes [12,13]. Studies at NWRI revealed differential responses by the four species; however, the variables (contaminants) that modified the test responses were not positively identified. Knowledge * To whom correspondence may be addressed (danielle.milani@ec.gc.ca). of the species sensitivities to different compounds (i.e., organic and metal contaminants) could add insight to evaluating the source(s) of sediment toxicity. Furthermore, no studies have compared the relative sensitivities (using both sublethal and lethal responses) of these organisms, nor has there been an attempt to extrapolate results to field-collected sediment in an attempt to determine possible causes of toxicity. The current study aims to partially address these deficiencies. We compare lethal and sublethal endpoints to determine the relative sensitivity of these four species to cadmium, copper, and nickel in water-only and in spiked-sediment exposures. The species relative sensitivity to each metal is then used as a diagnostic tool for interpreting whether a metal of interest might be eliciting test responses in field-collected sediment contaminated with metals. Culture methods MATERIALS AND METHODS Test organisms were cultured at 23 1 C, with a 16:8-h light:dark photoperiod, and an illumination of 500 to 1,000 lux, with the exception of T. tubifex, which was cultured in the dark. Water used for culturing (culture water) was charcoalfiltered and aerated City of Burlington (ON, Canada) tap water (Lake Ontario). Water quality characteristics were: conductivity 273 to 347 S/cm; ph 7.5 to 8.5; hardness 120 to 140 mg/ L; alkalinity 75 to 100 mg/l; and chloride ion 22 to 27 mg/ L. Sediment used in the culturing of Hexagenia spp. and T. tubifex was acquired from Long Point Marsh (Lake Erie) and comprised, on average, 70.33% silt, 29.13% clay, 0.54% sand, and 8.1% average total organic carbon content. 845

2 846 Environ. Toxicol. Chem. 22, 2003 D. Milani et al. Hyalella azteca was cultured in 20 to 25 plastic containers, each containing 1 L of culture water, 25 to 30 mature adults, and a gauze strip as substrate [14]. From 300 to 1,000 young/ week were produced. Chironomus riparius was cultured in 10-L covered aquaria, with silica sand used as substrate. Deposited egg masses were removed daily and were kept aside for testing purposes or used to initiate another culture [8]. Eggs of Hexagenia spp. were collected yearly from gravid females emerging from Lake St. Clair (ON, Canada) according to procedures described elsewhere [15]. Eggs were stored at 4 C to prevent hatching, with an allotment removed weekly for hatching. Hatched mayfly nymphs were added to culture sediment in 10-L aquaria and reared for six to seven weeks. Tubifex tubifex was cultured in sediment in plexiglas tanks. Each tank was inoculated with 200 full cocoons, and the animals were maintained to sexual maturity (seven to eight weeks) [16]. Water-only exposures (96 h) Water-only tests were conducted under static conditions for 96 h. Substrates used were: Nitex (B&SH Thompson, Scarborough, ON, Canada) screens for H. azteca [14], monolayer silica sand for C. riparius [17], and constructed glass tubes for Hexagenia spp. [18]. No substrate was used for T. tubifex, as they were shown to survive well without a substrate. Test beakers were supplemented with 2 mg crushed Nutrafin (Hagen Pet Foods, Waverly, NY, USA) fish flakes on day 0 and day 2. All tests passed an acceptability criterion of 90% control survival before being included in a data set. Each metal stock solution was prepared by dissolving reagent-grade cadmium (as CdCl 2 2½H 2 O), nickel (as NiCl 2 6H 2 O), or copper (as CuCl 2 2H 2 O) in Milli-Q water (Millipore, Bedford, MA, USA). Test beakers were not aerated during the test and were loosely covered with a plastic liner to minimize evaporation. Each test consisted of an unspiked control and five to six spiked concentrations, with one to six replicates per concentration. Sediment exposures The sediment used in the metal-spiked tests was collected from a reference site (303) ( N, W) located close to Long Point, Ontario (Lake Erie). Sediment was presieved (250- m mesh) to remove indigenous species. Sediment was spiked with the appropriate aliquot of metal and homogenized using a side-to-side shaker for 90 min. Overlying water used in the sediment exposures was the same water used for culturing purposes (characteristics described above). Test beakers were equilibrated for two weeks and aerated for one week prior to the start of the test. Sediment toxicity tests Test methods are described in Day et al. [8], Borgmann et al. [14], and Reynoldson et al. [16]. Water chemistry variables (ph, dissolved oxygen [mg/l], conductivity [ S/cm], temperature [ C], and total [ammonia and ammonium, mg/l]) were measured for each test in each replicate test beaker on day 0 (start of test) and at completion of the test (day 10 for C. riparius, day 21 for Hexagenia spp., and day 28 for H. azteca and T. tubifex). Tests were static and run under 16:8-h light: dark photoperiod with test beakers gently aerated to maintain between 50 and 100% saturation. Tests were conducted in 250- ml glass beakers (except for Hexagenia, in which 1-L glass jars were used), with a 1 to 1.5 ratio of sediment to overlying Fig. 1. Sampling locations in Collingwood Harbour (ON, Canada). water. Each test consisted of an unspiked control, a minimum of five test concentrations with two to five replicates per concentration depending on the amount of sediment available, and an additional beaker for chemical analysis. Each range of spiked concentrations was repeated a minimum of three times for each metal and for each species. For toxicity tests conducted with the field-collected sediment, five replicate beakers per test site were set up with an additional beaker for chemical analysis. Field-collected sedimentcollingwood Harbour Sediment was collected from six sites in Collingwood Harbour in July Three sites were selected in the east slip (C6, C7, and C8) and three sites in the west slip (C9, C10, and C11) (Fig. 1). Five field replicate samples were collected from each site using a miniponar sampler, placed in plastic bags, and stored on ice until return to the laboratory. Sediment was homogenized and wet sieved through a 250- m sieve prior to use in tests. A four to one ratio of culture water to sediment was used in the sieving process. The sieved sediment was allowed to settle for a minimum of 24 h, after which the water was decanted and used as the overlying water in the tests. Chemical analysis For the 96-h water-only tests, samples for chemical analyses were taken on day 0 prior to the introduction of the organisms. A sample from each concentration was poured into 20-ml scintillation vials and preserved with 2% nitric acid. Metal concentrations were determined by inductively coupled plasma optical emission spectroscopy (ICP-OES) (JY74 optical emission system) [19]. Water samples that fell below the method detection limit for the ICP were analyzed by atomic adsorption spectrophotometry (Varian SpectraAA-400, Varian Techtron, Mulgrave, Victoria, Australia) with Zeeman background cor-

3 Relative sensitivity of benthic invertebrates to metals Environ. Toxicol. Chem. 22, Table 1. Median lethal concentration (LC50) (geometric means) (mg/l) and range in parentheses for cadmium, nickel, and copper in 96-hour water-only exposures Metal Hyalella azteca Chironomus riparius Hexagenia spp. Tubifex tubifex Cadmium ( ) Nickel 3.62 ( ) Copper 0.21 ( ) ( ) 5.25 ( ) ( ) 7.82 ( ) a ( ) ( ) 0.87 ( ) ( ) 0.16 ( ) a Might not be accurate because of reductions in water hardness at the upper end of concentrations. rection [20]. For the sediment tests, bulk sediment, overlying water, and pore water was sampled from each chemistry beaker on day 0. The bulk sediment was centrifuged at 3,750 rpm at 4 C for 1 h to remove pore water. Overlying water and porewater samples were preserved with 2% nitric acid. Sediment samples were freeze-dried, ground, and homogenized, and metal concentration was determined by either ICP-OES or by AA spectroscopy. Sediment characterization Sediment characterization was performed by the Sedimentology Laboratory, NWRI (Burlington, ON, Canada). All sediments were analyzed for particle size and total organic carbon content. Particle size analysis was done following the procedures of Duncan and LaHaie (G. A. Duncan, G. G. LaHaie, National Water Research Institute, Burlington, ON, Canada, unpublished data). A homogenized sample of the sediment was dispersed in sodium metaphosphate and mixed for 15 min. The sample was then sieved through a 63- m screen. The residue on the sieve was dried and recorded as percent sand and gravel. The suspension that passed through the sieve was analyzed for percent silt and clay with a sedigraph analyzer. Total organic carbon was determined by drying a homogenized sediment sample for a minimum of 2 h, then burning 0.1 g of the dried sample for 250 s at 500 C. The percentage of organic carbon was determined by dividing the final weight of the sample by the initial weight (0.1 g) times 100. Statistical analysis For the spiked tests, median lethal concentrations (LC50s and LC25s) were computed using the trimmed Spearman Karber method [21]. The inhibition concentration (IC) estimate was performed at the 25% level on growth and reproductive endpoints of the spiked-sediment tests using the linear interpolation method with confidence intervals determined using the bootstrap method (random resampling of the test data with replacement) [22]. For field-collected sediment, species test responses were compared to acceptability criteria established for these four species for the Great Lakes area over a threeyear period [13]. Diagnostic capabilities of species relative sensitivity To discriminate between toxicity caused by specific metals, the geometric mean of the spiked-sediment LC25s and IC25s across the four species was determined. Each endpoint LC25 and IC25 was then divided by this geometric mean for each metal. For T. tubifex, only the most sensitive reproduction endpoint was used in the calculation (number of young produced per adult). Values 1 indicated the more sensitive endpoints (the smaller the number the more sensitive the endpoint), and values 1 indicated the less sensitive endpoints to each metal. For demonstration purposes, only the bulk sediment LC25s and IC25s were used in the calculations. Water-only exposures (96 h) RESULTS The LC50s for cadmium range from to 7.82 mg/l, a 602-fold difference between the species, which is the greatest range of the three tested metals. Cadmium is the most toxic (lowest LC50s) to two of the four species (C. riparius and H. azteca). The order in increasing LC50 is H. azteca C. riparius K T. tubifex K Hexagenia spp. (Table 1). Nickel is least toxic (highest LC50s) to all four species, with LC50s ranging from 3.62 to mg/l (21-fold difference). The order in increasing LC50 is H. azteca C. riparius T. tubifex Hexagenia spp. (Table 1). The LC50s for copper range from to 0.21 mg/l, a fivefold range in sensitivity between the species. Chironomus riparius is the most sensitive species to copper (lowest LC50). Copper is most toxic (lowest LC50s) to two of the four species (Hexagenia spp. and T. tubifex). The order in increasing LC50 is C. riparius Hexagenia spp. T. tubifex H. azteca (Table 1). Spiked-sediment exposures: Survival Cadmium LC50s for bulk sediment, overlying water, and pore water range from 33 to 815 g/g, from to 3.56 mg/ L, and from to 3.93 mg/l, respectively (Table 2). Cadmium is the most toxic to both H. azteca and C. riparius (LC50 ranges overlap). The LC50s are lowest in the overlying water fraction for H. azteca and C. riparius, whereas for T. tubifex and Hexagenia spp., overlying water and pore-water LC50s are similar (ranges overlap). The order in species sensitivity in increasing LC25/50 is H. azteca C. riparius K Hexagenia spp. T. tubifex. Nickel LC50s for bulk sediment, overlying water, and pore water range from 67 to 1,136 g/g, from 0.12 to mg/ L, and from 0.27 to mg/l, respectively (Table 2). Nickel is the least toxic metal to all species based on overlying water and pore-water LC50/25s, and the LC50/25s are lowest in the overlying water fraction for all species. The order in increasing LC25/50 is H. azteca K Hexagenia spp. C. riparius T. tubifex, different from the order observed in the water-only exposures. Thus, Hexagenia is more sensitive (lower LC25/ LC50) than Chironomus and Tubifex to nickel in sediment exposures, whereas it is the least sensitive organism (highest LC50) in water-only exposures (Table 1). Copper LC50s for bulk sediment, overlying water, and pore water range from 93 to 524 g/g, from to mg/l,

4 848 Environ. Toxicol. Chem. 22, 2003 D. Milani et al. Table 2. Geometric means of 25% and 50% median lethal concentrations (LC25s and LC50s, mg/kg or mg/l) and ranges in parentheses for cadmium, nickel, and copper in spiked-sediment exposures based on bulk sediment (BS), overlying water (OW), and pore-water (PW) metal concentrations Metal Hyalella azteca (28 d) Chironomus riparius (10 d) Hexagenia spp. (21 d) Tubifex tubifex (28 d) Cadmium Nickel Copper BS LC25 BS LC50 OW LC25 OW LC50 PW LC25 PW LC50 BS LC25 BS LC50 OW LC25 OW LC50 PW LC25 PW LC50 BS LC25 BS LC50 OW LC25 OW LC50 PW LC25 PW LC50 21 (16 32) 33 (28 44) ( ) ( ) ( ) ( ) 48 (43 57) 67 (62 74) ( ) 0.12 ( ) 0.17 ( ) 0.27 ( ) 81 (57 106) 128 ( ) ( ) ( ) ( ) ( ) 28 (26 30) 39 (36 46) ( ) ( ) ( ) ( ) 505 ( ) 665 ( ) 2.85 ( ) 9.89 ( ) 7.38 ( ) ( ) 265 ( ) 402 ( ) ( ) ( ) ( ) ( ) 560 ( ) 815 ( ) 0.48 ( ) 3.09 ( ) 0.79 ( ) 3.73 ( ) 324 ( ) 452 ( ) 2.10 ( ) 5.07 ( ) 4.02 ( ) 8.86 ( ) 60 (55 65) 93 (90 98) ( ) ( ) ( ) ( ) 600 ( ) 787 ( ) 1.14 ( ) 3.56 ( ) 1.15 ( ) 3.73 ( ) 918 ( ) 1136 ( ) ( ) ( ) ( ) ( ) 349 ( ) 524 ( ) ( ) ( ) ( ) ( ) and from to mg/l, respectively (Table 2). Copper is more toxic to Hexagenia spp. and T. tubifex (lowest LC25/ 50s) than cadmium and nickel. The order in increasing LC25/ 50 is H. azteca Hexagenia spp. C. riparius T. tubifex, different from the order observed in the water-only exposures. Thus, Hyalella is the least sensitive organism to copper (highest LC50) in the water-only exposures (Table 1), whereas it is the most sensitive organism ( Hexagenia) in the sediment exposures. The LC50/25s are lowest in the overlying water fraction for all species. Spiked-sediment exposures: Growth and reproduction Overall, growth and reproductive impairment occurs with increasing metal concentration, and in general, growth and reproduction endpoints are more sensitive than survival. In the bulk sediment fraction, Hyalella, Chironomus, and Hexagenia have a narrow range in cadmium IC25s for growth (10 to 16 g/g). The IC25s for Tubifex reproduction endpoints are higher than the growth endpoints, ranging from 301 to 769 g/g (Table 3). Overlying water IC25s for growth range from to mg/l and from 0.21 to 17.1 mg/l for reproduction endpoints. Pore-water IC25s for growth range from to mg/l and from 0.28 to 17.4 mg/l for reproduction endpoints. Cadmium is most toxic to Hyalella, Chironomus, and Hexagenia growth. The order in increasing IC25s is H. azteca C. riparius Hexagenia spp. K T. tubifex. Although this is the same order observed for survival, the LC25s for survival for Hexagenia are 27- to 209-fold higher than the LC25s for Hyalella (Table 2), whereas for growth, the IC25s for Hexagenia growth are only 1.4- to 9.6- fold higher than the IC25s for Hyalella growth. The IC25s are similar in the overlying water and pore-water fractions for all species. Based on the bulk sediment fraction, nickel IC25s for growth range from 40 to 146 g/g, overlying water IC25s range from 0.03 to 0.34 mg/l, and pore-water IC25s range from 0.11 to 1.16 mg/l (Table 3). Based on the water fractions, nickel is the least toxic (highest IC25s) to all four species. The IC25s for reproduction are higher than for growth, ranging from 408 to 669 g/g in the bulk sediment, from 6.5 to 26.0 mg/l in the overlying water, and from 10.2 to 31.9 mg/l in the pore water. The order in increasing IC25s is H. azteca Hexagenia spp. C. riparius T. tubifex, which is the same order observed for survival LC25s (Table 2). The IC25s are lowest in the overlying water fraction for all species. Bulk sediment copper IC25s for growth range from 38 to 78 g/g, for overlying water range from to mg/ L, and for pore water range from to mg/l (Table 3). The IC25s for reproduction are higher than for growth, ranging from 181 to 266 g/g in the bulk sediment, from to mg/l in the overlying water, and from 0.22 to 0.36 mg/l in the pore water. Of the three metals, copper is most toxic to Tubifex reproduction (lowest IC25s). The order in increasing sensitivity to copper is Hexagenia spp. H. azteca C. riparius T. tubifex, which is the same order observed for survival LC25s (Table 2). The IC25s are lowest in the overlying water fraction for all species. Field studycollingwood Harbour In general, test sediment consists mainly of silt (range 48 78%) and clay (range 12 22%) particles, with the exception of sites C8 and C9, which consist mainly of silt and sand (Table 4). Total organic carbon at all sites range from 1.4 to 2.3%. The Ontario sediment quality guidelines [23] for the measured metals, where available, are included for reference. Comparing the sediment metal concentrations to Ontario s sediment quality guidelines [23], sites C6, C7, and C8 are above the severe effect level (SEL) for zinc, and all sites are above the SEL for copper. Sites C8, C9, and C11 are above the SEL for lead. The lowest effect level (LEL) is exceeded for chromium, nickel, total organic carbon, total nitrogen, and total phosphorus at most sites; however, the reference mean also exceeds the LEL for these environmental variables. Data indicate potential toxicity to Hyalella and Chironomus survival at site C8 and potential toxicity to Hyalella growth at site C10 and to Chironomus growth at sites C6 and C8. No toxicity is evident for Hexagenia at any site (Table 5). Tubifex reproduction appears to be the most affected endpoint, with toxicity or potential toxicity observed for percent cocoon hatch and young production at all sites except C11.

5 Relative sensitivity of benthic invertebrates to metals Environ. Toxicol. Chem. 22, Table 3. Mean inhibitory concentrations (geometric means, mg/kg or mg/l) and ranges in parentheses for growth and reproduction endpoints based on bulk sediment (top), overlying water (middle), and pore-water (bottom) metal concentrations in spiked-sediment exposures Reproduction (Tubifex tubifex) Growth Number young per adult Number cocoons per adults Percent cocoons hatched Hyalella azteca Chironomus riparius Hexagenia spp. Metal 769 ( ) ( ) ( ) 467 ( ) 1.09 ( ) 1.21 ( ) 301 ( ) 0.21 ( ) 0.28 ( ) 14 (8 25) ( ) ( ) 16 (14 20) ( ) ( ) Cadmium 10 (6 18) ( ) ( ) 669 ( ) ( ) ( ) 451 ( ) 7.97 ( ) ( ) 408 ( ) 6.47 ( ) ( ) 83 (82 86) 0.13 ( ) 0.65 ( ) 146 (97 204) 0.34 ( ) 1.16 ( ) Nickel 40 (31 57) ( ) 0.11 ( ) 185 ( ) ( ) 0.22 ( ) 266 ( ) ( ) 0.36 ( ) 181 ( ) ( ) 0.22 ( ) 38 (31 49) ( ) ( ) 78 (35 143) ( ) ( ) Copper 76 (72 78) ( ) ( ) The sediment producing the greatest number of young/adult (C8) has the highest concentration of copper (5,150 g/g), zinc (16,000 g/g), chromium (103 g/g), and lead (1,540 g/ g) (Table 4). Diagnostic capabilities of species relative sensitivity For cadmium, the geometric mean of the individual LC25s and IC25s (bulk sediment) from Tables 2 and 3 (21, 28, 560, 600, 10, 16, 14, 301) is 58. For nickel and copper, the geometric mean is 186 and 108, respectively. Individual endpoint LC25s and IC25s divided by this geometric mean give the values shown in Figure 2. For example, the LC25 for Hyalella (21) when divided by 58 gives a value of The LC25s for Hexagenia shows the widest range in relative toxicity, ranging from 0.6 to 9.6. The LC25 for C. riparius also ranges from below 1 (0.5) to above 1 (2.7). However, all the other endpoints are consistently below 1 or consistently above 1. How these relative endpoint sensitivities can be applied to discriminating between metal toxicity in field sediment is explained by the following scenarios: If a sediment is lethal to H. azteca only, then toxicity could be due to nickel; if a sediment is lethal to H. azteca and C. riparius only, then toxicity could be due to cadmium; if a sediment is lethal to H. azteca and Hexagenia spp. only, then toxicity could be due to copper; and if a sediment is lethal to T. tubifex only, then toxicity is likely due to an undetermined contaminant. The application of this approach is shown using results from a study performed by Borgmann et al. [24]. Sediment toxicity tests were performed using the four species on sediments collected from various lakes in the Sudbury (ON, Canada) area, and tissue concentrations of various metals were measured for H. azteca after one-week exposures to the Sudbury sediment. Relating the tissue metal concentrations in Hyalella with body concentrations known to cause toxicity revealed that nickel was the only metal that accumulated to a level causing toxicity in H. azteca. The toxicity tests from four of the Sudbury lakes showed a strong effect on Hyalella survival and growth and an effect on Hexagenia growth when compared to control values (Table 6). The relative sensitivities of the four species are, in fact, consistent with nickel toxicity (Fig. 2). An inconsistency with the findings, however, is that Chironomus growth is not affected by the Sudbury sediments, whereas Hexagenia survival is reduced below the fifth percentile for survival in uncontaminated Great Lakes sediments (84%) [5,24]. Chironomus growth should be a more sensitive endpoint than Hexagenia survival if nickel is the cause of toxicity (Fig. 2). This could be an indication of a minor contribution from another toxicant affecting Hexagenia survival. Copper, for instance, was also elevated in the Sudbury surface sediments when compared to reference lakes [24]. Water-only exposures (96 h) DISCUSSION Variations in test conditions employed in water-only exposures, age and condition of test organism, and factors affecting bioavailability (i.e., water hardness, ph, temperature, feeding) have resulted in a wide range of LC50s reported over the years for the same or closely related species of benthic invertebrates, making comparisons difficult [2,3,6,25 37]. Although the order in metal sensitivity between the species is in good agreement with previous studies examining metal sensitivity in the same or closely related species, actual LC50s

6 850 Environ. Toxicol. Chem. 22, 2003 D. Milani et al. Table 4. Physical and chemical characteristics of sediment collected from Collingwood Harbour (ON, Canada). Ontario s sediment quality guideline lowest level (LEL) and severe effect level (SEL) and the reference mean are included. Values exceeding the SEL are italicized Location Site Silt Clay Sand Cd Cr Cu Depth (m) Fe 2 O 3 MgO MnO Ni ( g/ g) Pb Total N Total organic C Total P Zn ph range Reference mean , C6 C7 C8 C9 C10 C , , ,160 2,020 1,390 1,340 2,040 1, , ,260 1,580 1,440 1,991 1,680 16, LEL a SEL a a Persaud et al. [23] , , are not. From previous studies, the LC50s for cadmium range (present study value in parentheses) from to 0.07 mg/ L for amphipods (0.013), from 1.2 to 300 mg/l for midges (0.02), from 0.5 to 28.0 mg/l for mayflies (7.8), and from 0.17 to 47.5 mg/l for tubificids (0.16). For copper, LC50s range from to 0.91 mg/l for amphipods (0.21), from 0.03 to 0.86 mg/l for midges (0.043), from 0.18 to 0.32 mg/ L for mayflies (0.073), and from 0.09 to 0.89 mg/l for tubificids (0.16). For nickel, LC50s range from 0.08 to 13 mg/l for amphipods (3.6), from 9 to 80 mg/l for midges (5), from 4 to 76 mg/l for mayflies (including the present study), and from 61 to 67 mg/l for tubificids (18). With the exception of the cadmium LC50 for C. riparius, LC50s from the present study fall within ranges observed from previous studies. It should be noted that the LC50 for nickel for Hexagenia might not be reliable because of the decrease in water hardness in the higher concentration treatments. The detoxification mechanisms of water hardness on metal toxicity are known [38]. The decrease in calcium and magnesium ions as a result of the dilution of the culture water with stock solution in the Hexagenia nickel series might have lead to a decrease in competition between nickel and these cations, subsequently increasing toxicity. The reported LC50 might then be an underestimation of the LC50, although other factors generally accompanying a reduction in hardness (i.e., lowering of ph) could also influence toxicity. Metal-specific sensitivities in the same or similar species in other studies generally agree with the present study. For example, examining T. tubifex sensitivity to metals in 96- h exposures, Khangarot [33] found copper more toxic to T. tubifex, followed by cadmium and nickel, and Reynoldson et al. [35] found copper more toxic than cadmium to T. tubifex. Warnick and Bell [25] found copper most toxic to the mayfly E. subvaria, followed by cadmium. In 10-d exposures, Phipps et al. [3] found H. azteca more sensitive to cadmium, followed by copper and nickel, and L. variegatus was more sensitive to copper, followed by cadmium and nickel. Finally, in 96-h exposures examining amphipod sensitivity to metals, Rehwoldt et al. [26] found cadmium to be more toxic to Gammarus sp., followed by copper and nickel. Differences in metal-specific sensitivities also exist, however. Brković-Popović and Popović [29] found cadmium more toxic to T. tubifex, followed by copper, and nickel in 48-h exposures, Borgmann et al. [4] found H. azteca more sensitive to nickel than copper in one-week exposures, and Rehwoldt et al. [26] found Chironomus sp. more sensitive to copper followed by cadmium and nickel. Spiked-sediment exposures The spiked-sediment tests indicate that the relative sensitivities of the four species are not accurately predicted from the 96-h water-only exposures. Hexagenia is less sensitive than Table 5. Mean survival, growth, and reproduction and standard deviations in parentheses of sediment collected from Collingwood Harbour (ON, Canada). Criteria for determining toxicity for nearshore sediments of the Great Lakes (Reynoldson et al. [13]) are included for reference Site Hyalella azteca Survival Growth Chironomus riparius Survival Growth Hexagenia spp. Survival Growth Survival Tubifex tubifex No. cocoons/ adult % Cocoons hatched No. young/ adult C6 C7 C8 C9 C10 C11 NT a PT b T c 80.0 (19.4) 77.3 (22.4) 66.7 (27.2) 78.3 (16.7) 95.0 (6.4) 92.0 (8.7) (0.16) 0.31 (0.09) 0.23 (0.21) 0.24 (0.007) 0.21 (0.09) 0.30 (0.15) (3.3) 85.3 (8.7) 61.7 (33.3) 70.0 (12.8) 85.3 (7.3) 90.7 (11.2) (0.04) 0.21 (0.03) 0.19 (0.09) 0.28 (0.09) 0.32 (0.002) 0.28 (0.04) (4.5) 96.0 (8.9) 98.0 (4.5) (0.79) 4.77 (1.55) 5.80 (2.27) 2.76 (1.80) 4.37 (1.56) 3.38 (1.75) (11.2) 95.0 (11.2) (0.7) 9.8 (1.6) 10.8 (0.6) 9.2 (0.6) 9.8 (0.9) 8.8 (1.4) (7.8) 16.0 (5.4) 29.2 (18.9) 22.0 (9.4) 16.0 (12.9) 38.5 (10.5) (2.2) 0.10 (0.1) 11.2 (4.9) 4.3 (5.6) 3.4 (2.9) 10.3 (3.5) a Nontoxic. The upper limit of the nontoxic category is set using 2 standard deviation of the mean and indicates excessive growth or reproduction. b PT potentially toxic range. c T toxic.

7 Relative sensitivity of benthic invertebrates to metals Environ. Toxicol. Chem. 22, two species (54 g/l), suggesting exposure route differences and possible geochemical influences of the sediment [39]. Fig. 2. Bulk sediment median lethal concentrations of 25% (LC25s) and mean inhibitory concentrations of 25% (IC25s) for each species divided by the geometric mean across all endpoints for each metal. Bars below the dotted line ( 1) indicate the more sensitive endpoints; bars above the dotted line indicate the less sensitive endpoints. Tubifex to cadmium and nickel in the water-only exposures, yet the two show similar sensitivity or Hexagenia shows greater sensitivity to cadmium and nickel in the spiked-sediment tests. Chironomus is the most sensitive organism to copper in the water-only exposures but is less sensitive than both Hyalella and Hexagenia to copper in the spiked-sediment exposures. Chironomus and Hyalella also show comparable sensitivity to nickel in the water-only exposures, but Chironomus is far less sensitive than Hyalella ( 10-fold difference in LC50s in all test fractions) in the spiked-sediment exposures. Finally, Hyalella is the least sensitive organism to copper in the 96-h water-only exposures, yet is most sensitive ( Hexagenia) in the spiked-sediment tests. West et al. [39] also found differences in the order of species sensitivity in water-only versus sediment exposures. In determining the relative sensitivities of H. azteca, C. tentans, and L. variegatus in 10-d water-only exposures and in coppercontaminated sediments from the Keweenaw waterway, similar LC50s were found for L. variegatus and H. azteca in the waterborne tests (35 and 31 g/l, respectively). Exposure in field-collected sediments, however, resulted in reduced survival in Hyalella in 8 of 11 sites, whereas survival was unaffected for L. variegatus in all sites. Additionally, Chironomus survival was reduced in 7 of the 11 field sites, yet in the waterborne exposures, the LC50 was higher than the other Table 6. Mean survival and growth (mg wet wt) or reproduction in sediment collected from four lakes in the Sudbury (ON, Canada) area and control values (Borgmann et al. [24]). Toxicity is italicized Species Hyalella azteca 2 Chironomus riparius 79 Hexagenia spp. 75 Tubifex tubifex 94 Survival Lakes Control Growth (mg wet wt) Lakes Control Reproduction (young/adult) Lakes Control Endpoint sensitivity A comparison of LC25s and IC25s shows Hyalella survival and growth to be similarly sensitive indicators of chronic toxicity for the three metals tested. Similar results were found by Borgmann et al. [14], who reported that cadmium caused neither mortality nor induced a growth effect in Hyalella. This might indicate that, for Hyalella, the mode of toxicity for these three metals could be the same and that survival might be sufficient to infer cadmium, copper, and nickel toxicity. Chironomus growth is a more sensitive endpoint than survival for all metals, as evidenced by the lower IC25s; yet, in the dissolved phases, LC25 and IC25 values are close. That Chironomus is least sensitive to nickel (as well as the other three species) in both exposure types (water-only and spikedsediment tests) might suggest that the specific nature of the metal is important in explaining species sensitivities. For instance, copper and cadmium exhibit characteristics of group B metals (highly reactive, lack of specific binding to organic ligands, form strong covalent bonds), and group B metals tend to be persistent and toxic in small amounts [40]. Hexagenia growth is more sensitive than survival for all the metals tested. Copper, however, exerts a greater effect on Hexagenia survival than the other two metals. Hexagenia growth is similarly sensitive to Hyalella growth for cadmium (ranges overlap). Day et al. [8] also found Hexagenia survival and growth a sensitive indicator of chronic toxicity (more sensitive than Hyalella), reporting a greater than twofold difference in IC25s between Hyalella and Hexagenia in tributyltinspiked-sediment. Tubifex were quite tolerant of cadmium and nickel compared to the other three species (evidenced by the higher chronic LC25s). This is not surprising if considering the life stage of the organisms used in each test. The Tubifex test employs the organism in the adult stage, whereas for the other three tests, immature organisms are used. Copper, however, exerts a much stronger effect on Tubifex survival, and when looking at the dissolved phases, it is comparable to the other three species, especially Chironomus. Compared to the growth effects of the other three species, reproduction is also a less sensitive indicator of chronic toxicity, but again for copper, in the dissolved phases, differences are generally not as great as for cadmium and nickel. Tubifex reproduction has been documented to be a more sensitive endpoint than survival and a sensitive indicator of chronic toxicity [16]. Reynoldson et al. [12] reported Tubifex reproduction to be the most sensitive endpoint tested in contaminated sediments (more sensitive than growth), although the causative agent(s) of toxicity were not specifically identified. Differences in copper and nickel sensitivities observed for C. riparius in the water-only versus the spiked-sediment tests might occur because of poor survival in the water-only tests. Despite the use of a substrate and added food, low control survival required the water-only tests to be repeated several times. Also, the use of first instar organisms made it difficult to recover the animals after 96 h because of their small size. Control survival was not a problem in the spiked-sediment tests run over 10 d. It is possible that abrasion by silica sand on the first instar chironomids could have lead to an increase in mortality, also reported by others [41]. Pascoe et al. [42] also reported high mortality in first instar C. riparius in their

8 852 Environ. Toxicol. Chem. 22, 2003 D. Milani et al. control treatment, in which shredded filter paper was used as a substrate. If Chironomus experienced higher mortality in exposure treatments using aquarium gravel as a substrate, the LC50s could be underestimated. Diagnostic capabilities of species relative sensitivity Because contaminated field sediment is more often composed of a mixture of metals, a method to discriminate between toxicity caused by specific metals would be useful. The scenario described above or in the results describes the diagnostic capabilities using the survival (LC25) endpoints. The chronic endpoints (IC25s) can also be used to further confirm or check the cause of toxicity. For example, toxicity to Hyalella, Chironomus, and Hexagenia growth will always be greater than to Tubifex reproduction if toxicity is attributed to cadmium, nickel or copper. Although this diagnostic capability is shown using the bulk sediment fraction, it can also be applied to the overlying water and pore-water LC25s and IC25s. The comparison (Fig. 2) shows that C. riparius and Hexagenia spp. mortality are the only discriminatory endpoints across the three metals, since they showed the widest range in relative toxicity. All other endpoints provide little discriminatory power. Application to field-collected sediment Collingwood Harbour, previously listed as an area of concern, was chosen for the field evaluation because it represents an area affected by metals from the shipbuilding industry and contains elevated levels of copper [12]. Previous research on Collingwood Harbour has compared metal concentrations in sediments with Ontario s chemical sediment quality criteria. Chemical concentrations were high enough that the potential for biological effects could not be discounted. However, no adverse biological responses were evident in the benthic community structure, and laboratory toxicity tests indicated a problem only with T. tubifex reproduction, even though sediment contaminant concentrations exceeded the severe effect levels at certain sites. Overall, species responses to Collingwood Harbour sediments do not appear to reflect copper toxicity solely. Survival and growth in Hexagenia are the most sensitive indicators of copper toxicity (lowest values) (Fig. 2). Toxicity to Hyalella survival and growth and Chironomus growth can also occur (values 1). Site C8 showed potential toxicity to Hyalella and Chironomus (survival), whereas sediment from sites C6 and C8 showed potential toxicity to Chironomus (growth) (Table 5). These sites (C8 and C6) have the highest (5,150 g/g) and second highest (723 g/ g) copper concentration, respectively (Table 4). In the absence of other data, one might speculate that copper is the cause of toxicity at these sites. Hexagenia growth in the Collingwood Harbour sediments (Table 5) exceeds the upper range in the nontoxic category in sites C6 and C8, which would not be expected with copper toxicity. Additionally, Tubifex is the least sensitive species to copper in the spikedsediment exposures, yet toxicity to Tubifex reproduction is evident in sites C6, C7, C9, and C10. This implies that the causative agent of toxicity in the Collingwood Harbour sediments is not copper (or copper alone). The high copper concentration in the sediments, which is well above the SEL for copper in three of the six sites, is likely not bioavailable to the organisms. It is possible that either another metal is responsible for toxicity to Tubifex (zinc and lead have concentrations above the SELs in most sites), a combination of metals is responsible (additive or synergistic effects), or possibly that some unmeasured contaminant is causing the observed toxicity. Because the species sensitivities are not known for zinc or lead or for metal mixtures, no conclusions can be drawn. That percent cocoon hatch and young production are affected in five of the six sites (not adult survival and cocoon production) suggests that the mechanism of toxicity affects embryo development. Similar results were found by Reynoldson et al. [12] from these sites and also indicate that the toxicant could be acquired in the dissolved phase because embryonic development occurs inside the cocoon with food supplied by a yolk sac. It is also possible that food organisms (i.e., species that Hyalella, Chironomus, Hexagenia, and Tubifex feed on) might be contaminated with metals under field conditions, which might contribute to toxicity beyond what was observed in the lab. The presence of different kinds of food organisms at different sites could, perhaps, vary the amount of metal ingested by the different test species. Therefore, if ingestion via food is an important source of metal to an organism, it could affect the relative sensitivity of the different species to different metals at different sites under field conditions, thereby making it more difficult to interpret relative toxicity in sediments from different sites. Further studies with Tubifex would be useful to establish its sensitivity to other metals such as zinc and lead, as well as metal mixtures and other classes of compounds (organic contaminants). The data from Collingwood Harbour illustrate the importance of multiple single-species testing since the use of only one known sensitive species (i.e., H. azteca) could lead to erroneous conclusions regarding the presence of sediment toxicity in this particular case. Because organic contaminant analyses were not performed on the Collingwood Harbour sediment and because no information is currently available on the sensitivity of invertebrates to other metals in the Collingwood Harbour sediments in high concentrations, such as zinc and lead, or on the sensitivity to metal mixtures, no strong conclusions can be drawn regarding the cause of toxicity at these sites. Interactive effects among metals might be an important factor at the Collingwood Harbour sites since they contain several metals together in high concentrations, making it difficult to determine or distinguish between metal toxicity. CONCLUSIONS Species in the 96-h water-only exposures exhibited a wide range of sensitivity among the three metals, and this is most pronounced for the mayfly Hexagenia spp. Hyalella azteca is the most sensitive species to cadmium, and H. azteca and C. riparius are the most sensitive species to nickel. Chironomus riparius is the most sensitive species to copper; however, problems associated with using first instar chironomids in water-only tests could have resulted in an overestimation of toxicity. The order of species sensitivity in the spiked-sediment exposures follow that observed in the water-only exposures for cadmium, but do not for nickel and copper. The causes of the difference remain poorly understood. With respect to survival, H. azteca is the most sensitive species to cadmium and nickel, and Hexagenia spp. and H. azteca are the most sensitive species to copper. Although growth and reproduction are more sensitive indicators of toxicity than survival, they are not dis-

9 Relative sensitivity of benthic invertebrates to metals Environ. Toxicol. Chem. 22, criminatory endpoints because of the similar responses of the species to the three metals. For instance, Hyalella growth is always more affected than Tubifex reproduction, and a growth reduction in Chironomus and Hexagenia occurs for all three metals. Because of the wide range in response of Chironomus and Hexagenia survival to the three metals, these two endpoints could be used to discriminate between cadmium, copper, and nickel toxicity in sediments. Examination of species responses in Collingwood Harbour sediment revealed that although copper is present in high concentrations, it is not the likely cause of toxicity in contaminated field-collected sediment. Other metals are present in high concentrations (i.e., lead, zinc), but because species sensitivities to these metals (and to metal mixtures and organic contaminants) are unknown, the cause of toxicity in Collingwood Harbour sediments remains undetermined. AcknowledgementThanks to P.L. Gillis, R. Fletcher, R.S. Kirby, S.P. Thompson, P. Hoekstra, L.C. Diener, W. Norwood, J. Rajkumar, T. Pascoe, C. Logan, and M. Mawhinney. REFERENCES 1. Giesy JP, Rosiu CJ, Graney RL Benthic invertebrate bioassays with toxic sediment and pore water. Environ Toxicol Chem 9: Hickey CW, Martin ML Relative sensitivity of five benthic invertebrate species to reference toxicants and resin-acid contaminated sediments. Environ Toxicol Chem 14: Phipps GL, Mattson VR, Ankley GT Relative sensitivity of three freshwater benthic macroinvertebrates to ten contaminants. Arch Environ Contam Toxicol 28: Borgmann U, Cheam V, Norwood WP, Lechner J Toxicity and bioaccumulation of thallium in Hyalella azteca, with comparison to other metals and prediction of environmental impact. Environ Pollut 99: Day KE, Dutka BJ, Kwan KK, Batista N, Reynoldson TB, Metcalfe-Smith JL Correlations between solid-phase microbial screening assays, whole sediment toxicity tests with macroinvertebrates and in situ benthic community structure. J Gt Lakes Res 21: Suedel BC, Deaver E, Rodgers JH Jr Experimental factors that may affect toxicity of aqueous and sediment-bound copper to freshwater organisms. Arch Environ Contam Toxicol 30: Day KE, Kirby RS, Reynoldson TB The effect of manipulations of freshwater sediments on responses of benthic invertebrates in whole-sediment toxicity tests. Environ Toxicol Chem 14: Day KE, Maguire RJ, Milani D, Batchelor SP Toxicity of tributyltin to four species of freshwater benthic invertebrates using spiked sediment bioassays. Water Qual Res J Can 33: Cheam V, Reynoldson T, Garbai G, Rajkumar J, Milani D Impacts of coal mines and power plants across Canada. II. Metals, organics and toxicity in sediments. Water Qual Res J Can 35: Rochfort Q, Grapentine L, Marsalek J, Brownlee B, Reynoldson T, Thompson S, Milani D, Logan C Using benthic assessment techniques to determine combined sewer overflow and stormwater impacts in the aquatic ecosystem. Water Qual Res J Can 35: Chapman PM Sediment quality criteria from the sediment quality triad: An example. Environ Toxicol Chem 5: Reynoldson TB, Bailey RC, Day KE, Norris RH Biological guidelines for freshwater sediment based on benthic assessment of sediment (the BEAST) using a multivariate approach for predicting biological state. Aust J Ecol 20: Reynoldson TB, Day KE, Pascoe T The development of the BEAST: A predictive approach for assessing sediment quality in the North American Great Lakes. In Wright JF, Sutcliffe DW, Furse MT, eds, Assessing the Biological Quality of Fresh Waters. RIVPACS and Other Techniques. Freshwater Biological Association, Ambleside, Cumbria, UK, pp Borgmann U, Ralph KM, Norwood WP Toxicity test procedures for Hyalella azteca, and toxicity of cadmium and pentachlorophenol to H. azteca, Gammarus fasciatus, and Daphnia magna. Arch Environ Contam Toxicol 18: Hanes EC, Ciborowski JJH Effects of density and food limitation on size variation and mortality of larval Hexagenia rigida (Ephemeroptera: Ephemeridae). Can J Zool 70: Reynoldson TB, Thompson SP, Bamsey JL A sediment bioassay using the tubificid oligochaete worm Tubifex tubifex. Environ Toxicol Chem 10: American Society for Testing and Materials Standard guide for conducting acute toxicity tests with fishes, macroinvertebrates, and amphipods. E a. In Annual Book of ASTM Standards, Vol Philadelphia, PA, pp Henry MG, Chester DN, Mauck WL Role of artificial burrows in Hexagenia toxicity tests: Recommendations for protocol development. Environ Toxicol Chem 5: McLaren JW, Berman SS, Boyko VJ, Russell DS Simultaneous determinations of major, minor and trace elements in marine sediments by inductively coupled plasma atomic emission spectrometry. Anal Chem 53: Agemian H, Chau ASY A study of different analytical extraction methods for nondetrital heavy metals in aquatic sediments. Arch Environ Contam Toxicol 6: Hamilton MA, Russo RC, Thurston RV Trimmed Spearman Karber method for estimating median lethal concentrations in toxicity bioassays. Environ Sci Technol 11: Norberg-King TJ A linear interpolation method for sublethal toxicity: The inhibition concentration (ICp) approach (Ver 2.0). Technical Report U.S. Environmental Protection Agency, Duluth, MN. 23. Persaud D, Jaagumagi R, Hayton A Guidelines for the protection and management of aquatic sediment quality in Ontario. ISBN Final Report. Ontario Ministry of the Environment, Rexdale, ON, Canada. 24. Borgmann U, Norwood WP, Reynoldson TB, Rosa F Identifying cause in sediment assessments: Bioavailability and the sediment quality triad. Can J Fish Aquat Sci 58: Warnick SL, Bell HL The acute toxicity of some heavy metals to different species of aquatic insects. J Water Pollut Control Fed 41: Rehwoldt R, Lasko L, Shaw C, Wirhowski E The acute toxicity of some heavy metal ions toward benthic organisms. Bull Environ Contam Toxicol 10: Clubb RW, Gaufin AR, Lords JL Acute cadmium toxicity studies upon nine species of aquatic insects. Environ Res 9: Nehring RB Aquatic insects as biological monitors of heavy metal pollution. Bull Environ Contam Toxicol 15: Brković-Popović I, Popović M Effects of heavy metals on survival and respiration rate of tubificid worms: Part IEffects on survival. Environ Pollut 13: Chapman PM, Farrel MO, Brinkhurst RO Relative tolerances of selected aquatic oligochaetes to individual pollutants and environmental factors. Aquat Toxicol 2: Williams KA, Green DWJ, Pascoe D Studies on the acute toxicity of pollutants to freshwater macroinvertebrates 1. Cadmium. Arch Hydrobiol 102: Powlesland C, George J Acute and chronic toxicity of nickel to larvae of Chironomus riparius (Meigen). Environ Pollut 42: Khangarot BS Toxicity of metals to a freshwater tubificid worm Tubifex tubifex (Muller). Bull Environ Contam Toxicol 46: Collyard SA, Ankley GT, Hoke RA, Goldenstein T Influence of age on the relative sensitivity of Hyalella azteca to diazinon, alkylphenol ethoxylates, copper, cadmium, and zinc. Arch Environ Contam Toxicol 26: Reynoldson TB, Rodriguez P, Madrid MM A comparison of reproduction, growth and acute toxicity in two populations of Tubifex tubifex (Muller, 1774) from the North American Great Lakes and northern Spain. Hydrobiologia 334: Milani D, Day KE, McLeay DJ, Kirby RS Recent intraand interlaboratory studies related to the development of Environment Canada s biological test methods for measuring sediment toxicity using freshwater amphipods (Hyalella azteca) or midge