The Toxicity of Molybdate to Freshwater and Marine Organisms. III

Science of the Total Environment 609 (2017) 420–428

Contents lists available at ScienceDirect

Science of the Total Environment

journal homepage: www.elsevier.com/locate/scitotenv

The toxicity of molybdate to freshwater and marine organisms. III. Generating additional chronic toxicity data for the reﬁnement of safe environmental exposure concentrations in the US and Europe

D.G. Heijerick a,⁎,S.Careyb a ARCHE Consulting, Liefkensstraat 35d, 9032 Gent-Wondelgem, Belgium b International Molybdenum Association, 454–458 Chiswick High Road, London W4 5TT, United Kingdom

HIGHLIGHTS GRAPHICAL ABSTRACT

• Chronic ecotoxicity data for Mo were generated for a freshwater and marine species. • Data sets meet the US-EPA information requirements for deriving Final Chronic Values. • The US FCV-approach and European SSD-approach result in similar EQS- values for Mo.

article info abstract

Article history: The freshwater and marine long-term ecotoxicity datasets used in the European REACH registration dossiers for Received 27 March 2017 molybdenum and molybdenum compounds resulted in the derivation of a HC5,50%,freshwater (38.2 mg Mo/L) and Received in revised form 26 June 2017 HC5,50%,marine (5.70 mg Mo/L) by means of the statistical extrapolation method. Both datasets, however, did not Accepted 8 July 2017 meet the US-EPA information requirements for deriving Final Chronic Values (FCV) that were based on chronic Available online xxxx data. US-EPA compliance was achieved by generating chronic no-effect data for the freshwater benthic amphipod ﬁ Editor: Henner Hollert Hyalella azteca and the marine inland silverside sh Menidia beryllina, using sodium molybdate dihydrate as test substance. A 42d-EC10 of 44.6 mg Mo/L for reproduction was determined in a water-only exposure with H. azteca. Keywords: For M. beryllina, a 37d-NOEC of 139 mg mMo/L for standard length and blotted wet weight was found. Other end- Molybdate points (e.g., survival, hatching success) proved to be less sensitive. Data were added to the existing chronic tox- Marine assessment icity datasets, together with new long-term no-effect values that were identiﬁed in open literature for brown Freshwater assessment trout Salmo trutta, the marine alga Isochrysis galbana, the marine snail Nassarius dorsatus and the marine barnacle FCV derivation Amphibalanus amphitrite. The updated data sets resulted in a freshwater and marine HC5,50% of 35.7 and PNEC derivation 6.85 mg Mo/L, respectively. The same data sets were also used for the determination of US-EPA FCVs, where

the FVCfreshwater was 36.1 mg/L, and the FCVmarine was 3.85 mg Mo/L. As the Final Plant Values for both aquatic environments were higher than their respective FCVs, the Criterion Continuous Concentration (CCC) for molybdenum is equal to the FCV. © 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http:// creativecommons.org/licenses/by-nc-nd/4.0/).

⁎ Corresponding author at: Liefkensstraat 35 D, B9032 Gent-Wondelgem, Belgium. E-mail address: [email protected] (D.G. Heijerick).

http://dx.doi.org/10.1016/j.scitotenv.2017.07.070 0048-9697/© 2017 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). D.G. Heijerick, S. Carey / Science of the Total Environment 609 (2017) 420–428 421

1. Introduction MDR for the freshwater and marine environment and – where possible – allocates the species reported in Heijerick et al. (2012b) to these fami- The introduction of the EU-REACH legislation in 2006 (EC, 2006)re- lies. Based on the data in both tables, which include long-term no- quired the importers and manufacturers of high-tonnage chemicals effect levels for micro-alga, macro-alga and/or higher plants, two data (N1000 t/year) to submit a registration dossier for each of their sub- gaps were identified: a chronic endpoint for a benthic crustacean was stances. Conducting an environmental risk assessment on the basis of lacking for the freshwater environment, whereas the data gap for the mandatory data requirements (e.g. chronic toxicity data) was an inte- marine environment was a chronic endpoint for a second family of gral part of their registration obligation and is documented within the the Phylum “Chordata”. To ensure that water quality standards for Mo Chemical Safety Report (CSR) for each substance. in the United States are based on scientific relevant and robust data, The International Molybdenum Association (IMOA) coordinated the IMOA commissioned chronic toxicity tests for Mo with the freshwater development of such CSRs for eleven different molybdenum com- amphipod Hyalella azteca and the marine fish Menidia beryllina,hereby pounds, and commissioned an extensive research program for the as- addressing the identified data gaps with regard to FCV-derivation based sessment of long-term environmental effects of molybdate in the on chronic data. freshwater and marine aquatic compartment. The outcome of this re- This publication presents the outcome of these additional long-term search program was published by De Schamphelaere et al. (2010) and toxicity tests. These data, together with recently pubished data on Mo- Heijerick et al. (2012a). The statistical extrapolation methodology toxicity in the aquatic compartment (Lucas et al., 2017; Trenfield et al., used in Europe for deriving a safe concentration level for the freshwa- 2015, 2016; Van Dam et al., 2016), were added to the existing data ter/marine environment is based on a Species Sensitivity Distribution sets and used for a) setting freshwater and marine FCVs on one hand, (SSD) of at least 10 organisms covering 8 different taxonomic groups and b) to refine the freshwater and marine PNECs reported by (ECHA, 2008); this methodology was applied on the chronic data sets Heijerick et al. (2012a). for molybdate and resulted in a median 5% Hazardous Concentration

(HC5,50%) values of 38.2 and 5.70 mg Mo/L for the freshwater and marine environment, respectively. The United States Environmental Protection 2. Material and methods Agency (US-EPA), however, applies an alternative methodology for the derivation of a safe concentration level, the so-called Final Acute Value The freshwater ampipod Hyalella azteca was selected for the gener- (FAV) and Final Chronic Value (FCV) (Stephen et al., 1985/2010). A ation of chronic ecotoxicity data for a benthic crustacean. This species major difference between both approaches is that the US-EPA calcula- has been widely used in laboratory sediment toxicity tests and bioaccu- tion method only takes into account a) the four lowest values of a mulation tests, and standard testing procedures have been developed dataset, and b) the number of data points in that data set. Typically, a for both sediment and water-only exposures. It has been shown to be FCV is derived by dividing the FAV by a Final Acute-to-Chronic Ratio among the more sensitive organisms when exposed to metal- (FACR). However, depending on the data that are available concerning contaminated sediments (Milani et al., 2003; Phipps et al., 1995). Simi- chronic toxicity to aquatic animals, the FCV can also be calculated in lar sensitivity was noted in water-only exposure, where H. azteca was the same manner as the FAV. In practice, the availability of reliable more sensitive to cadmium and nickel than three other benthic inverte- chronic data is rather limited for most substances, so the approach brates (Chironomus riparius, Hexagenia spp. and Tubifex tubifex)(Milani that uses the FAV as the starting point (i.e. FCV = FAV/FACR) is the et al., 2003). Thus H. azteca is suitable for assessing the long-term toxic- most used methodology for deriving an FCV. ity of molybdate to bentic crustaceans. In order to derive an FCV that is directly based on chronic no-effect The inland silverside Menidia beryillina is one of three species in the levels, the data set should be in line with the minimum data require- atherinid family that are amenable to laboratory culture, and one of four ments (MDRs) that have been set out by US-EPA (Stephen et al., 1985/ atherinid fish species used for chronic toxicity testing; it is recommend- 2010). This requires that the dataset includes at least acceptable chronic ed by US-EPA for assessing the acute and chronic toxicity of e.g. effluents values for the eight families that are also required for setting an FAV. in the estuarine and marine environments (US-EPA, 2002). The species Algal species and higher plants (e.g., green alga, duckweed) are not in- can be found along the North American Atlantic coast from Cape Cod cluded; therefore, no-effect levels for those taxonomic groups should (Massachusetts) all the way to Florida and west to Vera Cruz, Mexico be set aside when deriving the FCV. It should be noted that information (Johnson, 1975). The species tolerates a wide range of temperature for those taxonomic groups is needed when setting a Final Plant Value (2.9–32.5 °C) (Tagatz and Dudley, 1961; Smith, 1971) and salinity (0– (FPV); both FCV and FPV are considered for setting the Criterion Contin- 58‰)(Simmons, 1957; Renfro, 1960). It is an ecologically important uous Concentration (CCC). species that is part of the diet of several commercial species like the Both the freshwater and marine dataset for molybdate used by bluefish Pomatomus saltatrix,themackerelScomber scombrus,and Heijerick et al. (2012b), and likewise in the relevant molybdenum sub- striped bass Morone saxatilis (Bigelow and Schroeder, 1953). All these stances REACH dossiers, did not comply with the US-EPA 8-family factors make M. beryllina a suitable and relevant species for FCV setting guidelines. Consequently, neither a freshwater nor a marine FCV for mo- in the marine environment. lybdenum that would be acceptable for regulatory purposes in the Tests with the silverside fish Menidia beryllina and the amphipod United States could be derived with those existing datasets. Tables 1 Hyalella azteca were conducted with the test substance sodium molyb- fi and 2 list the eight species families that are speci ed by US-EPA as date dihydrate (Na2MoO4·2H20, Climax Molybdenum UK Ltd., Lot

Table 1 Taxonomic requirements for the derivation of a freshwater Final Acute/Chronic Value.

Freshwater families IMOA freshwater dataset Phylum Class/Family

Salmonidae family (Osteichtyes) Oncorhynchus mykiss Chordata Actinopterygii/Salmonidae Second family in Osteichtyes Pimephales promelas Chordata Actinopterygii/Cyprinidae Third family in phylum Chordata Xenopus laevis Chordata Amphibia/Pipidae Planktonic crustacean Daphnia magna, Ceriodaphnia dubia Arthropoda Branchiopoda/Daphniidae Benthic crustacean No data available for this requirement An aquatic insect Chironomus riparius Arthropoda Insecta/Chironomidae Family in phylum other than Chordata Brachionus calyciﬂorus Rotifera Monogononta/Brachionidae Family in any order of insect, or any phylum not already represented Lymnaea stagnalis Mollusca Gastropoda/Lymnaeidae 422 D.G. Heijerick, S. Carey / Science of the Total Environment 609 (2017) 420–428

Table 2 Taxonomic requirements for the derivation of a marine Final Acute/Chronic Value.

Freshwater families IMOA marine dataset Phylum Class/Family

Family in phylum Chordata Cyprinodon variegatus Chordata Actinopterygii/Cyprinodontidae Second family in phylum Chordata No data available for this requirement One family in phylum other than Arthropoda or Chordata Dendraster excentricus Echinodermata Echinoidea/Dendrasteridae One member of mysidae or Penaeidae Americamysis bahia Arthropoda Malacostraca/Mysidae Family not in phylum Chordata Acartia tonsa Arthropoda Maxillopoda/Acartiidae Second family not in phylum Chordata Strongylocentrotus purpuratus Echinodermata Echinoidea/Strongylocentrotidae Third family not in phylum Chordata Mytilus edulis Mollusca Bivalvia/Mytilidae Any other family Crassostrea gigas Mollusca Bivalvia/Ostreidae

#43006L, Mo 39.6% (theory 39.7%), stored in closed container at room The starting material for the test was b24 h post-fertilization eggs, temperature in a dry area). obtained from Aquatic Indicators Inc. (St. Augustine, Florida). Develop- ing embryos were incubated in retention baskets consisting of a glass Petri dish base (1.5 × 1.5 cm) with a nylon screen collar tall enough to 2.1. Long-term toxicity test with the freshwater benthic crustacean prevent the loss of animals. A light:dark cycle of 16:8 h with 30 min H. azteca transition period was applied during the test; maximum light intensity ranged between 432 and 454 lx. The test setup was shielded from sur- A 42 day water-only reproduction test was conducted at Fraunhofer rounding activities by curtains. Each dish contained twenty randomly Institute for Molecular Biology and Applied Ecology (IME; GLP Code selected embryos, counted on a daily basis during incubation and ARH-001/4-26/S), following the test principles of the US-EPA ringtest checked for potential fungal growth. When all living embryos had (US-EPA, 2000), Environment Canada (Government of Canada, 2013) hatched, the length of time to 95% hatch and hatchability were record- and ASTM (ASTM International, 2013) with modifications (Borgmann, ed. Day 0 post-hatch was based on ≥95% hatch in the control group. 1996). A detailed overview of methodology and results is available in All live fry were counted and released into their respective replicate the study report by Ziese et al. (2016). growth chambers on day three post-hatch. Food during the test Test organisms orginated from Freds Haustierzoo (Köln, Germany) consisted of rotifers (Brachionus sp.) and brine shrimp (Artemia sp.). and were bred in the laboratory of the Fraunhofer IME. Cultures were Test chambers were cleaned periodically during the test to remove maintained at 25 ± 2 °C, pH 6–8, and light:dark cycle of 16:8 h with waste material and uneaten food and to minimize biological growth light intensity of 500–1000 lx in “Hyalella” medium (1 mM CaCl ,0.01 2 on the sides and bottom of the test chamber. Survival was monitored NaBr,0.05mMKCl,1mMNaHCO ,0.25MgSO in activated charcoal fil- 3 4 daily by visually inspecting each test chamber, and any behavioral or tered drinking water). Amphipods were fed with ground TetraMin® physical changes were recorded, including abnormalities. The evaluated twice a week ad libitum (approx. 0.06 g per administration). toxicological endpoints in this test were: days to hatch start, days to A stock solution of molybdate, added as Na MoO ·2H 0, was freshly 2 4 2 hatch completion, egg hatchability, post-hatch fry survival, and growth. prepared in Hyalella medium before each renewal. A detailed summary Test solutions were analyzed for molybdenum with inductively of the test design and testing conditions is given in Table 3. coupled plasma mass spectrometry (ICP-MS, Agilent 7500A). A 10 mL Identification and removal of dead organisms occurred on a daily aliquot from each treatment was taken six days prior to initiation, and basis, and observed abnormalities in appearance and behaviour were on days 0, 7, 14, 21, 28, 35, and 37 (termination of the test). Samples recorded. Newborn amphipods per beaker were counted and removed were diluted as necessary with 2% HNO to provide final sample con- at each water renewal. At the end of the test period, the number of 3 centrations within the analytical standard concentration range. Tem- male amphipods was counted by visual inspection in each test vessel perature, pH, and dissolved oxygen concentration, and salinity were and total weight of adults per test vessel was measured as dry mass. measured in all replicates of the test substance treatments and control Dissolved molybdenum concentrations were measured in the test group at test initiation, weekly throughout the test, and at the end of media of each treatment before and after renewal. Equal volume sam- the exposure period. ples of aged test solution were taken from each replicate at each renewal and pooled together. Aliquots (20 mL) in 50 mL polyethylene vials, 2.3. Statistics were stabilized by adding 0.2 mL of 65% nitric acid (‘Rotipuran’ quality, Carl Roth, Karlsruhe, Germany) and stored in a refrigerator until analy- Arithmetic mean measured molybdenum concentrations were ap- sis. Molybdenum was determined by inductively-coupled plasma opti- plied for the effect evaluation with H. azteca. For biomass per adult spec- cal emission spectroscopy (Agilent 720 ICP-OES, Agilent Technologies, imen alive at the end of the exposure period, the total weight (dry mass) Waldbronn, Germany). Hardness, alkalinity, conductivity, pH and am- of each replicate was divided by the number of surviving adults. The monia content of the fresh test media were determined for each treat- number of offspring was calculated per replicate and divided by the ment at the start of the exposure period, between day 9 and day 11, total number of females per replicate. Determination of NOEC and and on day 28. These solution properties were also measured in the LOEC was based on ANOVA followed by Dunnett's or Williams' test or end-of- test media pooled for each molybdate concentration. an appropriate non-parametric test. An EC10 (+95% Confidence Inter- val) was only determined when a distinct concentration-response rela- 2.2. Conducting a long-term toxicity test with the inland silverside tionship was observed, using Probit-analysis assuming a log-normal M. beryllina distribution of the data. Statistical analyses on the M. beryllina results were performed using A 37 day (28 day-post-hatch) water-only growth test was conduct- SAS software, and inferences of statistical significance were based upon ed at ABC Laboratories following Good Laboratory Practice Standards a p b 0.05. The EC10 for all endpoints was estimated by fitting a four- (US-EPA, 1989), following an internal ZBC-protocol that was based on parameter logistic (sigmoid-shaped) model, with two fixed parameters US-EPA OPPTS guideline 850.1400 (US-EPA, 1996). A detailed overview (0 and 100% reduction), to the data with percent reduction for a given of methodology and results is available in the study report by Dinehart endpoint as the dependent variable and log molybdenum concentration (2013); a summary of the test design and testing conditions is given in as the independent variable. Testing for normality and homogeneity of Table 3. variance was done with the Shapiro-Wilk's (p b 0.01) and Levene's D.G. Heijerick, S. Carey / Science of the Total Environment 609 (2017) 420–428 423

tests (p b 0.01). Depending on the outcome, a parametric or non- para-

Table 3 metric ANOVA and Dunnett's test on the non-transformed or trans- Chronic testing conditions for the benthic crustacean H. azteca and the inland silverside M. formed data was conducted for determination of the NOEC and LOEC beryllina. for the different endpoints. Hyalella azteca Menidia beryllina 3. Results and discussion Testing facility Fraunhofer IME ABC Laboratories Inc. (Schmallenberg, Germany) (Columbia, Missouri, USA) Origin Freds Haustierzoo (Köln, Aquatic Indicators Inc. (St. 3.1. Chronic toxicity of molybdenum to H. azteca Germany) Augustine, Florida, USA) Culture/testing “Hyalella” medium (1 mM Artificial laboratory The mean measured concentrations of the freshly prepared test so- medium CaCl2, 0.01 NaBr, 0.05 mM seawater; mixture lutions were between 85% and 150% of nominal concentrations, and KCl, 1 mM NaHCO , 0.25 (hardness of 130–160 mg/L 3 remained stable during the incubation period. The arithmetic mean of mM MgSO4 in activated as CaCO3) of well charcoal filtered drinking water/dimineralized well the test molybdenum concentrations (Table 3). Physicochemical pa- water) water to which commercial rameters during the test period are also presented in Table 3. sea salt mix (Crystal Sea Some implausible effects on weight and offspring per female were Marinemix; Marine noted in the lowest concentration, but were not confirmed in the next Enterprises International, fi fi Inc. Baltimore, Maryland). two treatment levels. A reliable, signi cant t of the concentration- Prior to use, the medium is effect relationship could only be obtained by discarding the lowest con- heated and passed through centration from the analysis. No significant adverse effects on the sur- a UV sterilizer and a vival of H. azteca were observed in any of the treatments, and no sediment filter. clinical signs were observed for the surviving organisms. The Test duration 42 days 37 days (28 days post-hatch) NOECsurvival was therefore equal or higher than the highest test concen- a Endpoints Survival, growth , Survival, growth, hatching tration of 245.1 mg Mo/L. An EC10,survival of 254.0 mg Mo/L was reproductive performance performance estimated. b Tested lifestage 7 to 10 days old amphipods Newly fertilized eggs ( 24 Total weight of replicates was unaffected up to, and including a con- h post-fertilization) Tested concentrations Nominal test item: 0 Nominal as Mo: 0 (control) centration of 103.6 mg Mo/L (which is therefore the NOECweight); the (Control), 31.5, 63, 126, 62.5, 125, 250, 500, 1000 derived EC10 for the weight-endpoint was 171.9 mg Mo/L. The derived 252, 504, 1009 mg test mg Mo/L. NOECbiomass and EC10,biomass (adult growth) were 214.4 mg Mo/L and item/L Arithmetic mean 205.3 mg Mo/L. Adult reproduction was the most sensitive endpoint. Nominal as Mo: 0 (control), measured: control, 65.6, The estimated EC Was 44.6 mg Mo/L. Due to the larger variation 12.5, 25, 50, 100, 200, 400 139, 265, 532, and 1070 mg 10,reprod. mg Mo/L Mo/L. in reproduction among replicates for this endpoint, a statistically signif- Arithmetic mean measured icant effect was found only between the control and the highest test test item: Control, 32.5, concentration, hence resulting in a NOECreprod. that is about a factor of 95.2, 129.0, 261.1, 540.0, five higher than the EC , i.e. 214.4 mg Mo/L. 869.2 mg test item/L 10,reprod. Analytics Agilent 720 ICP-OES Agilent 7500A ICP-MS LOD: 0.197–0.345 MQL: 10 mg Mo/L 3.2. Chronic toxicity of molybdenum to M. beryllina μg/L/LOQ: 0.59–1.03 μg/L Recovery from Recovery: 100.3 ± 0.8% (n fortifications: 100–117% of The measured molybdenum concentrations in the different treat- = 11) nominal ments remained within ±20% of the nominal concentrations through- Recovery from fortifications: 98.9–104.7% out the exposure (Table 3). Physicochemical parameters are also Renewal frequency Three times/week Flow-through, proportional presented in Table 3. diluter system (cfr Mount Egg hatch began on day 5 in the control and all test substance treat- and Brungs, 1967) resulting ments, and Day 0 post-hatch (i.e., ≥95% hatch in the control treatment) in six volume additions per day was determined to be study day 9. Overall hatching success in the con- Number 84 trol was 93%, and ranged from 95 to 99% in the test substance treat- Replicates/treatment ments. Post-hatch survival (calculated as the percent of hatched fry # Individuals/replicate 10, transfer with a bore 20 embryos per replicate alive at test termination) was 93% in the control, and ranged from 87 Pasteur pipette to 97% in the test substance treatments. No statistically significant dif- Test conditions 250 mL covered glass 15 ∗ 22 ∗ 24 cm glass vessels, containing ±5 mL aquaria containing 500 mL ferences on hatch-related or survival endpoints were noted in any of clean sand; 200 mL test of test solution the test substance treatments as compared to the control, i.e., up to solution per glass vessel, no Temperature: 24.4–25.3 °C; 1070 mg Mo/L. – aeration pH 8.1 8.2; Oxygen: Growth of surviving fry was assessed at test termination (day 37; Temperature: 22.3–22.9 °C; 6.1–7.8 mg/L; salinity of pH 8.1–8.6; Oxygen: 19.4–20.9‰; light intensity 28 days post-hatch) through standard length and blotted wet weight 6.0–10.7 mg/L; light of 432–454 lx (16:8 L:D) measurements. Replicate values were pooled into a treatment mean intensity of 650–667 lx using the number of individuals present at the end of the test in (16:8 L:D) weighting the means. A summary of the mean standard length and Feeding Diatom Thalassiosira Starting at day 4: 14 mL mean blotted weight in the control and treatments is given in Table 4. weisflogii 1200TM® & frozen rotifers (Brachionus) TetraMin®: Starting at day 5: ad Wk 1: 0.5 mg/L & 0.25 mg/L libitum living rotifers Wk 2: 0.75 mg/L & 0.5 mg/L (Brachionus) Table 4 Wk 3: 1.0 mg/L & 1.0 mg/L Starting at day 12: ad Effect of Mo on the length and weight of the inland silverside M. beryllina. Wk 4: 1.5 mg/L & 1.5 mg/L libitum newly hatched Nominal Mo (mg Mo/L) Control 62.5 125 250 500 1000 Wk 5: 2.0 mg/L & 2.0 mg/L brine shrimp (Artemia) Wk 6: 2.5 mg/L & 2.5 mg/L Measured Mo (mg Mo/L) –a 65.6 139 265 532 1070 Feeding frequency At every renewal Addition of living food: 3 Mean standard length (mm) 14.1 14.2 14.1 13.5 12.9 12.2 times/day Mean blotted weight (g) 0.023 0.022 0.020 0.020 0.017 0.016

a Adult biomass at test termination. a Minimum Quantiﬁable Limit: 10 mg/L. 424 D.G. Heijerick, S. Carey / Science of the Total Environment 609 (2017) 420–428

The maximum dynamic biomass loading calculated at test termination 3.4. Derivation of safe concentration levels according to ECHA and US-EPA was 0.0142 g/L/day. There was a statistically significant reduction for standard length The data generated for H. azteca and M. beryllina (Ziese et al., 2016; and blotted wet weight in the 265, 532, and 1070 mg Mo/L test sub- Dinehart, 2013) and the data recently published in peer-reviewed sci- stance treatments when compared to the control (p b 0.05). The NOEC entific literature (Lucas et al., 2017; Trenfield et al., 2015, 2016; Van and LOEC for both standard length and blotted wet weight was 139 Dam et al., 2016) were added to the existing data set for Mo (Heijerick and 265 mg Mo/L, respectively. A reliable estimate of the EC10 for stan- et al., 2012b). An overview of all reliable chronic data is presented in dard length and blotted wet weight, however, could not be obtained; no Table 5. robust effect-concentration relationship could be determined due to the Under EU legislation, the Predicted No Effect Concentration (PNEC) lack of sufficiently severe effects at the highest test concentrations. The is one of the keystone values of a risk assessment. Sufficient high quality 37d-NOEC of 139 mg/L is considered as a reliable value that can be used data were already available for Mo to derive a freshwater and marine for risk assessment of molybdate in the marine environment. PNEC by means of the scientifically more robust statistical extrapolation concept (ECHA (European Chemicals Agency), 2008). The first step of

3.3. New long-term toxicity data for Mo this method entails the derivation of a HC5,50% (Hazardous Concentra- tion that adversely affects 5% of the species included in the dataset). A survey of peer-reviewed publications identified a number of re- This concentration represents the reference value for setting the final cent studies on the aquatic toxicity of Mo in the freshwater and marine PNEC; the latter is defined by dividing the HC5,50% by an assessment fac- environment. Lucas et al. (2017) studied the long-term effects of molyb- tor (AF) that is situated between 1 and 5. The value that is used as the AF date on the early-life stages of brown trout (Salmo trutta). No reliable depends on the outcome of an uncertainty analysis on the overall chronic data for this freshwater species had previously been published dataset which is assessed against the following criteria: in scientific literature. An 85-day development test was conducted • with sodium molybdate dihydrate as the test substance, following test- the overall quality of the database and the endpoints covered the di- ing methods based on Environment Canada procedures (APHA, 1992). versity and representativeness of the taxonomic groups covered by Average total molybdenum concentrations measured in test waters the database • during the test were used for the statistical calculations. No significant statistical uncertainties around the 5th percentile estimate, fl fi fi effects on survival or length were reported in any of the treatment con- e.g., re ected in the goodness-of- t or the size of con dence interval fi centrations: survival exceeded 98% in all treatments, and length aver- around the 5th percentile comparisons between eld and mesocosm fi aged 21 to 22 mm. These observations resulted in IC values of studies and the 5th percentile of mesocosm/ eld studies to evaluate 10 fi N1247 mg Mo/L for these endpoints. A significant reduction of wet the laboratory to eld extrapolation • comparison of the HC5,50% with unbounded NOECs weight was noted with increasing Mo-levels, resulting in an 85d-IC10 value for wet weight of 202.5 mg Mo/L. Using the quality criteria as outlined by Klimisch et al. (1997), this study is considered reliable and relevant when assessing long-term effects of molybdate in the aquatic Heijerick et al. (2012b) concluded that an AF of three was sufficient- environment, and the value of 202.5 mg Mo/L should therefore be ly protective for both the freshwater and marine environment. This taken account in the risk assessment of molybdate. value was also applied in the REACH registration dossiers for molybde-

Trenfield et al. (2015, 2016) investigated the long-term effects of Mo num and molybdenum compounds, resulting in a PNECfreshwater and and three other metals (Cu, Al, Ga) for two tropical marine species: the PNECmarine of 12.7 mg Mo/L and 1.9 mg Mo/L, respectively. snail Nassarius dorsatus and the microalga Isochrysis galbana. No stan- The addition of ecotoxicological information for new species to the dard guidelines could be followed as there is a lack of applicable test existing dataset altered the HC5,50% and the resulting PNEC. Given the methods for the trophic environment (Van Dam et al., 2008). The test high quality of the new data, and the fact that new taxonomic groups procedures used in these studies were in line with the conventional were introduced into both the freshwater and marine datasets – there- testing principles that are applied in ecotoxicology, including the quan- by increasing their representativeness for the actual ecosystem –,a tification of test concentrations via ICP-MS measurements. Both studies more stringent assessment factor on the HC5,50% when setting a PNEC concluded that Mo was the least toxic of the four metals tested. For I. value is not needed. galbana, no effects on growth rate were observed at the highest Mo con- The RIVM ETX 2.0 software program was used for the calculation of a centration (72 h–ErC10 N 10 mg Mo/L) across any of the three tempera- log- normal distribution through the toxicity data set. These log-normal tures that were tested (24 °C, 28 °C, 31 °C). Larvae of the snail N. dorsatus distributions are presented in Figs. 1 and 2. The HC5,50% and resulting were exposed to concentrations up to 7.0 mg Mo/L and their survival PNEC values associated with these distributions are given in Table 6. and growth (expressed as growth rate) were examined. Here too, no The revised PNEC for the freshwater compartment is 11.9 mg Mo/L adverse effects were observed (96 h-ErC10 N 7.0 mg Mo/L). which is marginally lower than the previous PNECfreshwater value of The same group of researchers also developed a novel marine chron- 12.7 mg Mo/L. For the marine environment an increase of the HC5,50% ic bioassay for the barnacle Amphibalanus amphitrite, and used the same by a factor of 1.2 is determined, resulting in a revised PNECmarine of four metals as reference toxicants (Van Dam et al., 2016). They con- 2.28 mg Mo/L. firmed the findings reported in Trenfield et al. (2015, 2016), i.e.; molyb- With the addition of the new species, the molybdenum chronic date being the least toxic of those four metals. No adverse effects on the ecotoxicity data are now compliant with the US-EPA requirements for successful transition of nauplii larvae into cyprids within 96 h were ob- deriving a freshwater and marine FCV using chronic data (Stephen served in any of the tested molybdate concentrations (ICP-MS mea- et al., 1985/2010). Because freshwater and saltwater have different sured), resulting in a 96 h-EC10 of N9.0 mg Mo/L. chemical compositions, and likewise because freshwater and saltwater Unbounded no-adverse effect concentrations are often not consid- (i.e., estuarine and true marine) species rarely inhabit both waters si- ered for setting safe concentration levels in the environment; the use multaneously, (US) National Guidelines provide separate species of such unbounded values (worst-case assumption on toxicity) may re- criteria for these two kinds of water. sult in overly conservative quality standards which in their turn could The first step for determining the FCV entails the calculation of a spe- trigger unnecessary risk management measures. They can, however, cies mean chronic value (SMCV) for each species for which at least one serve as supporting evidence to demonstrate that a derived quality chronic value is available. This is achieved by calculating the geometric standard – derived with bounded no-effect levels – is sufficiently pro- mean of all chronic values available for the species. This approach is sim- tective for those species with unbounded EC10-levels. ilar to the data treatment procedure that is applied in Europe, i.e. use of D.G. Heijerick, S. Carey / Science of the Total Environment 609 (2017) 420–428 425

Table 5 Reliable chronic no-effect concentrations of molybdate for freshwater and marine organisms.

Freshwater environment Marine environment

Species Reference value (EC10 or NOEC) Species Reference value (EC10 or NOEC) mg Mo/L mg Mo/L

Data taken from Heijerick et al. (2012b) Data taken from Heijerick et al. (2012b)

Oncorhynchus mykiss 43.3 Mytilus edulis 4.4 Pimephales promelas 60.2 Acartia tonsa 7.96 Ceriodaphnia dubia 63.0 Cyprinodon variegatus 84.1 Pseudokirchneriella subcapitata 74.3 Americamysis bahia 116 Daphnia magna 89.5 Phaeodactylum tricornutum 170 Xenopus laevis 115.9 Dendraster excentricus 233.6 Chironomus riparius 121.4 Strongylocentrotus purpuratus 325.8 Brachionus calyciﬂorus 193.6 Ceramium tenuicorne 641 Lymnaea stagnalis 221.3 Dunaliella tertiolecta 881 Lemna minor 241.5 Crassostrea gigas 1174

Freshwater environment Marine environment

Species Reference value (EC10 or NOEC) Species Reference value (EC10 or NOEC) mg Mo/L mg Mo/L

New data New data

Hyalella azteca 44.6 (Ziese et al., 2016) Menidia beryllina 139 (Dinehart, 2013)

Salmo trutta 202.5 (Lucas et al., 2017) Supportive information (unbounded EC10 values) Isochrysis galbana N10.0 (Trenfield et al., 2015) Nassarius dorsatus N7.0 (Trenfield et al., 2016) Amphibalanus amphitrite N9.0 (Van Dam et al., 2016) the geometric mean as the species reference value when more than one chronic limits, and their geometric mean is referred to as the maximum data point is available for a specific species. The second step is the calcu- acceptable toxicant concentration (MATC). Comparison of the individu- lation of appropriate genus mean chronic values (GMCVs). The use of al EC10sfromTable 4 with their respective MATC showed that the EC10 genus mean values prevents a dataset from being biased by an over- was the most sensitive parameter for most species (73% of the species). abundance of species in one or a few genera; in the case of Mo, each spe- The only species where the MATC was markedly lower than the EC10 cies belongs to a different genus, i.e., SMVC = GMVC for these data. was for the freshwater amphibian X. laevis. In this test the NOEC and

According to the US-EPA, a chronic value may be obtained by calcu- LOEC were respectively a factor of 10 and 3 below the EC10. The large lating the geometric mean of the lower and upper chronic limits from a difference between EC10 and NOEC is likely related to the low variation chronic test, or by analyzing chronic data using regression analysis among the control replicates in this test, resulting in statistical signiﬁ- (Stephen et al., 1985/2010). The guidance, however, does not specify cant differences at effect levels well below 10%. Under those conditions the percentage of effect (x%) that is considered relevant when applying a NOEC and LOEC reﬂect a mathematical effect rather than a “true” ad- the latter approach, i.e. which ECX% should be considered. De verse effect, and the EC10 should be used for assessment purposes. Schamphelaere et al. (2010) and Heijerick et al. (2010a) reported both Based on the comparison between MATC and EC10, it was decided that NOECs, LOECs, as well as EC10 levels that were obtained via regression more relevant and conservative FCVs would be obtained when using analysis. The NOECs and LOECs are equivalent to the lower and upper the EC10 (where available) as chronic value.

Fig. 1. Species sensitivity distribution for Mo in the aquatic freshwater environment. 426 D.G. Heijerick, S. Carey / Science of the Total Environment 609 (2017) 420–428

Fig. 2. Species sensitivity distribution for Mo in the aquatic marine environment.

Quantiﬁcation of the freshwater and marine FCV requires the four The freshwater data set includes 10 species, and the critical chronic lowest genus mean values for each compartment, as well as the total values are 43.2, 44.6, 60.2 and 63 mg Mo/L (values for O. mykiss, number of data points in each dataset (algae and plants not included). H. azteca, P. promelas, C. dubia, respectively). Using these values results

The FCV is calculated with the equations in a FCVfreshwater of 36.1 mg Mo/L. The marine data set also contains 10 species, and the critical chronic R values are 4.4, N7.0, 7.96 and N9.0 mg Mo/L (values for M. edulis, N. P ¼ ðÞN þ 1 dorsatus, A. tonsa and A. amphitrite, respectively). It should be noted that unbounded “greater than” NOECs are considered in the United PðÞðÞ∑ 2 States when setting a water quality criterion such as the FCV, whereas ðÞ2 − ln GMCV ln GMCV this strategy is less common in European environmental legislation – S2 ¼ 0 14 P pffiffiffi 2 especially when the unbounded value is one of the drivers in the assess- P B P C ðÞP −@ A ment. The rationale behind the use of such unbounded values In the 4 United States is that their rejection would unnecessarily lower the Final Chronic Value by eliminating chronic values for resistant species. P P pffiffiffi Hence, provided the tests were conducted properly, chronic values re- ðÞln GMCV −S P L ¼ ported as “greater than” values should be used. 4 pffiffiffiffiffiffiffiffiffiffi 4. Discussion A ¼ S 0:05 þ L Both the freshwater and marine FCV are considered reliable as their FCV ¼ eA underlying datasets meet the quality criteria that have been outlined by US-EPA (Stephen et al., 1985/2010). Long-term data are available for with P the cumulative probability, R the rank, N the total number of data each of the required taxonomic species. These data were generated in points, and the summation symbol applicable on the four lowest values, well-documented studies published in peer-reviewed journals, follow- e.g. Σ(P) is the sum of the cumulative ranking of the four lowest data ing the principles of internationally accepted standard guidelines. points. Where more than one value was available for a species, the min-max

Table 6 Summary of water quality criteria for molybdenum according to European and United States calculation methodologies.

Freshwater environment Marine environment

2012 data seta Current data set 2012 data seta Current data set

HC5,50% 38.2 mg Mo/L (95% CL: 18.7–57.3 mg 35.7 mg Mo/L (95% CL: 18.6–52.9 mg 5.70 mg Mo/L (95% CL: 0.58–21.0 mg 6.85 mg Mo/L (95% CL: 0.92–22.1 mg Mo/L) Mo/L) Mo/L) Mo/L) PNEC (AF = 3) 12.7 mg Mo/L 11.9 mg Mo/L 1.90 mg Mo/L 2.28 mg Mo/L FCV 34.8 mg Mo/L (incomplete dataset) 36.1 mg Mo/L 0.75 mg Mo/L (incomplete dataset) 3.85 mg Mo/L FPV 74.3 mg Mo/L 74.3 mg Mo/L 274 mg Mo/L N10 mg Mo/L

HC5,50%: median Hazard Concentration affecting 5% of the ecosystem. PNEC: Predicted No Effect Concentration. FCV: Final Chronic Value. FPV: Final Plant Value. a Data used in Heijerick et al., 2012. D.G. Heijerick, S. Carey / Science of the Total Environment 609 (2017) 420–428 427 range was well within a factor of 10. The difference between the four The currently applied approach for setting water quality standards lowest GMCVs was also well below a factor of 10 (factor of 1.46 and for Mo does not take the potential impact of the composition of the 2.04 for the freshwater and marine GMCVs, respectively). The derived test medium on Mo-toxicity into account. Numerous publications FCVs were lower than all the GMCVs that were used in the calculations, have demonstrated the importance of parameters such as pH and hard- and are thus considered to be protective for all the species that were ness on the bioavailability and toxicity of metals; pH, for instance, can covered in the datasets. Finally, there is no indication that any of the alter the speciation of a metal (e.g. signiﬁcant formation of hydroxides GMCVs should be considered as an outlier. and carbonates at alkaline pH-levels) whereas Ca- and Mg-ions can

The FCVfreshwater is marginally higher than the preliminary compete with other cations for binding (and uptake) to aquatic organ- FCVfreshwater of 34.8 mg Mo/L that was estimated in Heijerick et al. isms. These kind of interactions, however, are not expected to have a (2012b), but is almost identical to the HC5,50% of 35.7 mg Mo/L that is de- significant impact on Mo-toxicity. Under physiological conditions (pH 2− termined for the updated dataset. The FCVmarine is a factor of five higher N 6.5) the sole molybdenum species is the molybdate anion, [MoO4] than the preliminary FCVmarine of 0.75 mg Mo/L as derived in Heijerick (Cruywagen, 2000; Cruywagen et al., 2002), and therefore any pH- et al. (2012b). The difference with the current marine HC5,50% of related impact on toxicity due to changes in Mo-speciation is not ex- 6.85 mg Mo/L is about a factor of two. pected. Moreover, as Mo is present as an anion, any competition with Once an FCV is calculated for a substance, the final step is to deter- cations for binding to the same site can also be excluded. An electrostat- mine the Criterion Continuous Concentration (CCC). This reference ic interaction between cations and the molybdate-ion (formation of an value is intended to be a good estimate of a threshold of unacceptable ion-pair) was studied by Torres et al. (2016). At a Mo-concentration effect. If maintained continuously, any concentration above the CCC is that was typical for uncontaminated freshwater (Mototal =1 expected to cause an unacceptable effect, but it does not imply that ×10−7 M) and Ca- and Mg levels that represent typical concentrations 2+ 2+ any concentration of a substance above its CCC necessarily causes an un- in groundwater (Ca total =13.6mM;Mg total = 0.092 mM), about acceptable effect. The aquatic environment is expected to tolerate such 30% of the molybdate formed an electrostatic binding with Ca- and exceedances if the magnitude and duration of the exceedances above Mg-ions. Molybdate concentrations at EC10/NOEC levels are several or- the CCC is appropriately limited, and there are sufficient periods for re- ders of magnitude higher (range: 4.59 × 10−5 M–1.22 × 10−2 M) covery during which the concentration remains below the CCC. It stands than the concentration that was tested by Torres et al. (2016), whereas to reason that the higher the concentration is above the CCC, the shorter the hardness of the test media was up to one order of magnitude lower the period of time it can be tolerated. The CCC represents the lowest than the typical groundwater concentrations that were used by Torres value of three parameters: the FCV, the Final Plant Value (FPV), and et al. (2016) in their calculations. It is therefore expected that most of the Final Residue Value (FRV). the molybdate in the test media will be present as “free” molybdate The relevance of the latter value, which is determined on the basis of (i.e., no ion-pair formation). In addition, removal of molybdate from bioconcentration factors, is questionable for essential metals like the water column by biological uptake will shift the equilibrium to- molybdenum. Interpreting measured BCF values for metals can be wards the release of molybdate from the Ca-molybdate system. There- complicated as biota regulate internal concentrations of metals fore not expected that the natural variation of Ca and Mg in the through (1) active regulation, (2) storage, or (3) a combination of ac- aquatic environment will not have a significant effect on the bioavail- tive regulation and storage over a wide range of environmental ex- ability and toxicity of the molybdate anion. posure conditions (Regoli et al., 2012). Metals that are biologically essential (e.g., copper, zinc, and also molybdenum) are actively reg- 5. Conclusions ulated in organisms that require them. When environmental concentrations are low, high BCF values may be expected as a natural New, reliable chronic data were generated for H. azteca and consequence of active metal uptake to meet nutritional require- M. beryllina. The amphipod was among the most sensitive freshwater ments. On the other hand, when an organism is exposed to an envi- species to molybdate. Using all the new freshwater toxicity information ronment containing excess quantities, three main strategies are used on Mo, the previously reported HC5,50% of 38.2 mg Mo/L decreased to to counter potential toxicity: reduced intake, enhanced excretion, or 35.7 mg Mo/L. For the marine environment, the addition of the no- sequestration of the metals within tissues to render them non-toxic effect concentration for the inland silverside caused an increase of the (Baker, 1981). Literature reviews on laboratory and field data show HC5,50% from 5.70 mg Mo/L to 6.85 mg Mo/L. The new data permitted that, for both essential and non-essential metals, bioaccumulation the calculation of a freshwater and marine FVC/CCC that was compliant and the BCF are inversely related to the exposure concentration in with the US-EPA information requirements, i.e., 36.1 mg Mo/L and most aquatic organisms (McGeer et al., 2003; DeForest et al., 2007), 3.85 mg Mo/L, respectively. Despite the differences in EU- and USA- ‘ ’ and thus the concept of a constant BCF value is not applicable. Con- methodology with regard to the derivation of a safe concentration sequently the derivation of a FRV is not justified for an essential level, the freshwater HC5,50% and FVC/CCC are almost identical. A differ- metal like molybdenum. ence of a factor of two is noted between both approaches for the marine The FPV is equivalent to the lowest result from a test with an impor- environment. The fact that the EU-methodology does not take un- tant aquatic plant species in which the concentrations of a substance is bounded values into account, may account for the observed difference. measured and the endpoint is biologically important; the US-EPA guidance refers to 96 h-algal tests or tests with an aquatic vascular plan. The References freshwater and marine datasets do include chronic values for algae and aquatic plants (Table 5) but the test duration in the algal tests is 72 h. American Public Health Association (APHA), 1992. Standard Methods for the Examination This deviation from the recommended 96 h period is not expected to of Water and Wastewater. 18th edition. American Public Health Association, 18th have a major effect on the calculated chronic values; results for both ex- edition. ASTM International, 2013. Standard test method for measuring the toxicity of sediment- posure periods are, for instance, accepted for environmental classifica- associated contaminants with freshwater invertebrates (ASTM E1706-05(2010)). tion purposes (GHS, CLP). The two EC10s that can be considered for ASTM International Book of Standard Volume 11.06. ASTM, West Conshohocken, PA, USA. the FPVfreshwater are 74.3 and 241.5 mg Mo/L. Four relevant values are — fi N Baker, A.J.M., 1981. Accumulators and excluders strategies in the response of plants to identi ed for the marine environment and range between 10 and heavy metals. J. Plant Nutr. 3, 643–654. 881 mg Mo/L. The lowest values were generated for the freshwater Bigelow, H.B., Schroeder, W.C., 1953. Fishes of the Gulf of Maine. U.S. Fish Wildl. Serv. Fish. green algae P. subcapitata and the marine algae I. galbana, and were Bull. 53, 1–577. Borgmann, U., 1996. Systematic analysis of aqueous ion requirements of Hyalella azteca:a taken forward as FPVfreshwater and FPVmarine, respectively (Table 6). Since standard artificial medium including the essential bromide ion. Arch. Environ. the FVC is lower than the FPV, the former values represent the CCCs. Contam. Toxicol. 30, 356–363. 428 D.G. Heijerick, S. Carey / Science of the Total Environment 609 (2017) 420–428

Cruywagen, J.J., 2000. Protonation, oligomerization, and condensation reactions of Regoli, L., Van Tilburg, W., Heijerick, D., Stubblefield, W., Carey, S., 2012. The vanadate(V), molybdate (VI), and tungstate (VI). Adv. Inorg. Chem. 49, 127–182. bioconcentration and bioaccumulation factors for molybdenum in the aquatic envi- Cruywagen, J.J., Draaijer, A.G., Heyns, J.B.B., Rohwer, E.A., 2002. Molybdenum (VI) equilib- ronment from natural environmental concentrations up to the toxicity boundary. ria in different ionic media. Formation constants and thermodynamic quantities. Sci. Total Environ. 435–436, 96–106. Inorg. Chim. Acta 31, 322–329. Renfro, W.C., 1960. Salinity relations of some fishes in the Aransas River. Tulane Stud. De Schamphelaere, K.A.C., Stubblefield, W., Rodriguez, P., Vleminckx, Janssen, C.R., 2010. Zool. 8, 83–91. The chronic toxicity of molybdate to freshwater organisms. I. Generating reliable ef- Simmons, E.G., 1957. Ecological survey of the Upper Laguna Madre of Texas. Publ. Inst. fects data. Sci. Total Environ. 408, 5362–5371. Mar. Sci. 4 (2), 156–200. DeForest, D.K., Brix, K.V., Adams, W.J., 2007. Assessing metal bioaccumulation in aquatic Smith, B.A., 1971. An Ecological Study of the Delaware River in the Vicinity of Artificial Is- environments. The inverse relationship between bioaccumulation factors, trophic land. Part V. The Fish of Four Low-salinity Tidal Tributaries of the Delaware River Es- transfer factors and exposure concentrations. Aquat. Toxicol. 84 (2), 236–246. tuary. Progress Report to Public Service Electric and Gas CoIchthyological Assoc., Dinehart, S., 2013. Sodium molybdate: early life-stage toxicity test with the inland silver- Ithaca, NY (291 pp). side, Menidia beryllina, une flow-through conditions. Study Report for the Interna- Stephen CE, Mount DI, Hansen DJ, Gentile JR, Chapman GA, Brungs WA. Guidelines for De- tional Molybdenum Association, ABC-Study No. 69114. ABC Laboratories Inc, riving Numerical National Water Quality Criteria for the Protection Of Aquatic Organ- Columbia, Missouri, USA (79p). isms and Their Uses. US-EPA PB85-227049, United States Environmental Protection EC (European Commission), 2006. Regulation (EC) No. 1907/2006 of the European Parlia- Agency, Office of Research and Development, 1985 (2010 electronic version), (59p). ment and of the Council of 18 December 2006 Concerning the Registration, Evalua- Tagatz, M.E., Dudley, D.L., 1961. Seasonal Occurrence of Marine Fishes in Four Shore Hab- tion, Authorisation and Restriction of Chemical Substances (REACH), Establishing a itats Near Beaufort, N.C., 1957–1960. U.S. Fish. Wildl. Serv. Sc. Rept. Fish 390 (19 pp). European Chemicals Agency, Amending Directive 1999/45/EC and Repealing Council Torres, J., Tissot, F., Santos, P., Ferrari, C., Kremer, C., Kremer, E., 2016. Interactinos of Regulation (EEC) No. 793/93 and Commission Regulation (EC) No. 1488/94 as well as W(VI) and Mo(VI) oxyanions with metal cations in natural waters. J. Solut. Chem. Council Directive 76/769/EEC and Commission Directives 91/155/EEC, 93/67/EEC, 93/ 45, 1598–1611. 105/EC and 2000/21/EC. Trenfield, M.A., van Dam, J.W., Harford, A.J., Parry, D., Streten, C., Gibb, K., van Dam, R.A., ECHA (European Chemicals Agency), 2008. Guidance in Information Requirements and 2015. Aluminium, gallium, and molybdenum toxicity to the tropical marine Chemical Safety Assessment. Chapter R.10: Characterisation of Dose (Concentra- microalga Isochrysis galbana. Environ. Toxicol. Chem. 34 (8), 1833–1840. tion)–Response for Environment (65 pp). Trenfield, M.A., van Dam, J.W., Harford, A.J., Parry, D., Streten, C., Gibb, K., van Dam, R.A., Government of Canada, 2013. Environment Canada: Biological Test Method: Test for Sur- 2016. A chronic toxicity test for the tropical marine snail Nassarius Dorsatus to assess vival and Growth in Sediment and Water Using the Freshwater Amphipod Hyalella the toxicity of copper, aluminium, gallium and molybdenum. Environ. Toxicol. Chem. azteca. EPS 1/RM/33. Second edition. 35 (7), 1788–1795. Heijerick, D.G., Regoli, L., Stubblefield, W., 2010. Thechronictoxicityofmolybdatetoma- US-EPA, United States Environmental Protection Agency, 1989. Pesticide programs; good rine organisms. I. Generating reliable effects data. Sci. Total Environ. 430, 260–269. laboratory practice standards; final rule (40 CFR, part 160). Fed. Regist. 54 (158), Heijerick, D.G., Regoli, L., Carey, S., 2012. The toxicity of molybdate to freshwater and ma- 34067–34074. rine organisms. II. Effects assessment of molybdate in the aquatic environment under US-EPA, United States Environmental Protection Agency. 1996. Ecological Effects Test REACH. Sci. Total Environ. 435–436, 179–187. Guidelines, OPPTS 850.1400, Fish Early-Life Stage Toxicity Test, 1996:(13p). Johnson, M.S., 1975. Biochemical systematics of the atherinid genus Menidia. Copeia 4, US-EPA, United States Environmental Protection Agency, 2000. Methods for Measuring 662–691. the Toxicity and Bioaccumulation of Sediment-associated Contaminants With Fresh- Klimisch, H.J., Andreae, E., Tillmann, U., 1997. A systematic approach for evaluating the water Invertebrate. 2nd edition. EPA/600/R-99/064. Duluth, MN and Washington DC. quality of experimental and ecotoxicological data. Regul. Toxicol. Pharmacol. 25, 1–5. US-EPA, United States Environmental Protection Agency, 2002. Method 1006.0: Inland Lucas, B.T., Quinteros, C., Burnett-Seidel, C., Elphick, J.R., 2017. An Evaluation of Molybde- Silverside, Menidia beryllina, Larval Survival and Growth; Chronix Toxicity. num Toxicity to the Oligochaete, Tubifex tubifex, and Early-life Stages of Brown Trout, Van Dam, R.A., Harford, A.J., Houston, M.A., Hogan, A.C., Negri, A.P., 2008. Tropical marine Salmo trutta. Bull. Environ. Contam. Toxico. 98 (6), 747–752. toxicity testing Australia: a review and recommendations. Australas. J. Ecotoxicol. 14, McGeer, J.C., Brix, K.V., Skeaff, J.M., DeForest, D.K., Brigham, S.I., Adams, W.J., et al., 2003. 55–. 88 Inverse relationship between bioconcentration factor and exposure concentration Van Dam, J.W., Trenfield, M.A., Harries, S.J., Streten, C., Harford, A.J., Parry, D., van Dam, for metals: implications for hazard assessment of metals in aquatic environments. R.A., 2016. A novel bioassay using the barnacle Amphibalanus amphitrite to evaluate Environ. Toxicol. Chem. 22 (5), 1017–1037. chronic effects of aluminium, gallium and molybdenum in tropical marine receiving Milani, D., Reynoldson, T.B., Borgmann, U., Kolasa, J., 2003. The relative sensitivity of four environments. Mar. Pollut. Bull. 112 (1–2), 427–435. benthic invertebrates to metals in spiked-sediment exposures and application to con- Ziese, C., Klawonn, T., Schäfers, C., 2016. Effect of Sodium Molybdate Dihydrate on the Re- taminated field sediment. Environ. Toxicol. 22 (4), 845–854. production of Hyalella azteca. Study Report for the International Molybdenum Asso- Mount, D.I., Brungs, W.A., 1967. A simplified dosing apparatus for fish toxicological stud- ciation on Hyalella azteca, 42 Day Reproduction Test (Water-only) Under Semi-static ies. Water Res. 1, 21–29. Conditions, ARH-001/4-26/S. Fraunhofer Institute for Molecular Biology and Applied Phipps, G.L., Mattson, V.R., Ankley, G.T., 1995. Relative sensitivity of three freshwater Ecology (IME), Schmallenberg, Germany (61p). benthis macroinvertebrates to ten contaminants. Arch. Environ. Contam. Toxicol. 28, 281–286.